Morphine, a naturally occurring alkaloid, is the prototype of opiate analgesics. Although morphine acts through opiate receptors (1), it is unlikely that all clinically observed side effects, e.g., nausea and emesis, are exclusively dependent on this type of signal transduction. Neurophysiological and pharmacological evidence has shown that 5-HT, in addition to opioids, plays an important role in signal transmission at various levels of the pain pathway (2). At the spinal level, intrathecal administration of 5-HT produces an antinociceptive effect via spinal 5-HT3 receptors (3), which are ligand-gated ion channels, forming a cation-permeable membrane channel. 5-HT3 receptors are also involved in the mediation of acute chemo-inflammatory pain (4). Intrathecal administration of the 5-HT3 receptor agonist, 2-methylserotonin, produced dose-related analgesia against formalin-induced chemo-inflammatory pain in male rats (5).
Within the central nervous system, 5-HT3 receptors are found primarily in the limbic system, brainstem, and spinal cord. In the spinal cord, 5-HT3 receptors have been localized in the superficial laminae of the dorsal horn (6) and are prominently expressed in a variety of peripheral ganglia, including a subset of neurons in the dorsal root ganglion (7).
The 5-HT3 receptor is directly and indirectly linked to several physiologic and pathologic processes, including gastrointestinal motility, visceral pain, nausea, and emesis (8).
The 5-HT3 receptor in the area postrema triggers emesis (9), and 5-HT3 receptor antagonists such as ondansetron are clinically used to prevent nausea associated with chemotherapy and general anesthesia (4). A variety of drugs affect the 5-HT3 receptor apart from their primary action at other receptors. For example, cannabinoids (10) also act on 5-HT3 receptors and it has been suggested that these effects may be linked to in vivo effects such as analgesia and antiemesis.
More recently, morphine has been shown to act as a specific competitive antagonist at rat 5-HT3 receptors (11). Therefore it was the aim of this study to investigate whether morphine also has a specific effect on human 5-HT3 receptors, to characterize it, and to elucidate if there are differences between human and rat 5-HT3 receptors.
HEK 293 cells were grown as monolayers on culture plates (NUNC, Wiesbaden, Germany) in DMEM Nutrient Mix F12 (1:1; v v−1) medium containing 10% heat inactivated fetal calf serum, penicillin (100 I.U./mL), streptomycin (100 μg/mL), geneticine (0.75 μg/mL), and glutamine (292 μg/mL). The cells were cultured at 37°C in a humidified atmosphere (5% CO2). For electrophysiological experiments cells were transferred to 35 mm Petri dishes (NUNC). They were used 2–6 days after transfer, before the cell layer became confluent.
For stable transfection 20% confluent HEK 293 cells were transfected by the modified calcium phosphate method with the human 5-HT3A receptor cDNA subcloned into the mammalian expression vector pcDNA3 (Invitrogen, Karlsruhe, Germany) under the control of the human cytomegalovirus promoter. Two days after transfection, cells were selected by the addition of geneticine (800 μg/mL) to the culture medium. The medium was changed every 2 days. After occurrence of single cell colonies these were separated by use of cloning cylinders (Sigma-Aldrich, Munich, Germany) and further subcultured in 24 well plates (Falcon, BD Biosciences, Heidelberg, Germany) until confluence. Approximately 20 to 40 colonies from each transfection experiment were tested for stable expression of the particular cDNA by 5-HT-induced [14C]-guanidinium influx and binding of the selective 5-HT3 receptor antagonist [3H]-GR65630. Colonies with the highest receptor expression were used for further experiments. Although most physiological 5-HT3A receptors are not exclusively composed of this subunit, the homopentameric 5-HT3A receptor is very suitable for mechanistic studies because of its homogeneous composition.
Currents through the 5-HT3 receptor were measured with the patch-clamp technique. A patch-clamp amplifier (EPC-7; List Electronic, Darmstadt, Germany) was used with the output filter set between 65 and 500 Hz (sampling rate 125–1000 Hz) and pClamp software (Axon Instruments, Foster City, CA, U.S.A.). 5-HT (3 μM) was applied for 60 s at −30 or −60 mV in a standard voltage-clamp experiment in whole cell configuration. Capacitative transients and series resistance were measured and compensated using the internal compensation circuitry of the amplifier. A series resistance compensation of up to 70% was used.
The baseline control response to 3 μM 5-HT was measured before and after the drug application. To exclude rundown effects, the mean value of these two agonist applications was taken as the average control current. Each current trace was measured three times and averaged to reduce noise effects. A washout time of at least 90 s was allowed for recovery of the receptors from desensitization.
The solutions were applied via a perfusion pipette positioned close to the cell. The solution exchange rate was monitored by “open pipette” experiments. The external solution applied to the patch had the following composition (mM): NaCl 150, KCl 5.6, CaCl2 1.8, MgCl2 1.0, HEPES 10, pH 7.4. Patch pipettes with resistances of 1.5–3 MΩ were filled with “intracellular” solution containing (mM): KCl 140, EGTA 10, MgCl2 5, HEPES 10, adjusted to a pH of 7.4. To minimize loss of drug from interaction with plastic material, drug solutions were stored in glass reservoirs and Teflon tubing was used. The recordings were performed at room temperature (21 ± 1°C).
The currents were digitized with an interface (Digidata 1200; Axon Instruments, Molecular Devices Corporation) and stored on an IBM 586-compatible PC. Data analysis was performed with pClampR 6/8 software (Axon Instruments). GraphPad PrismR 3.03 software (GraphPad, San Diego, CA) was used to create graphics. The concentration-response curves for 5-HT, and analogously for the morphine effect, were fitted by the Hill equation,
i: immediate peak current as fraction of the maximal (control) current,
c: 5-HT concentration,
n: Hill coefficientEC50: 5-HT concentration inducing the half-maximal effect.
To calculate a KB value for morphine at human 5-HT3A receptors, the EC50 values of 5-HT in the absence and in the presence of morphine (1 μM) were applied to the Schild Equation:
The time course of agonist-induced activation and desensitization f(t) were fitted either separately using a single exponential function, or simultaneously with a bi-exponential function (according pClamp 6/8, Axon Instruments):
5-hydroxytryptamine (creatinine sulfate) was obtained from Sigma (München, Germany). Morphine (sulfate) was obtained from Mundipharma (Limburg, Germany) and naloxone (hydrochloride) was obtained from Ratiopharm (Ulm, Germany). Drug solutions were prepared daily from aqueous stock solutions (10–50 mM).
At negative membrane potentials the application of 5-HT to single HEK 293 cells induced inward currents (Fig. 1A) in a concentration-dependent manner. Current amplitudes between 500 and 5000 pA were observed in whole-cell patches at −30 mV. The specific human 5-HT3A receptor antagonist ondansetron (0.3 nM applied before and during the agonist) reversibly inhibited the 5-HT (3 μM)-induced peak currents (Fig. 2A). The 5-HT concentration-response curve was measured and fitted to the Hill equation yielding an EC50 value for 5-HT of 2 μM and a Hill coefficient of 2.2 (Fig. 1B).
Morphine inhibited 5-HT (3 μM, ≈EC70)-induced peak currents concentration-dependently and reversibly (Fig. 2B) with an IC50 value of 1.1 μM and a Hill coefficient of 1.2 (Fig. 2C). In these experiments, the opioid was present continuously for 1 min before and also for the duration of the 5-HT stimulus (equilibrium application). In addition to suppressing peak currents, morphine decreased the activation and inactivation time courses of the currents in a concentration–dependent manner: e.g., 1 μM morphine increased the activation time constant (τon) by a factor of 2.3 and the inactivation time constant (τoff) by a factor of 2.6 (Fig. 3; paired Student's t-test, P < 0.05). At morphine concentrations larger than 1 μM the inactivation time constant could not be determined because the activation process was not complete even after 10 s of simultaneous 5-HT and morphine application. Morphine (0.1–10 μM) applied alone, to patches that were sensitive to 5-HT, did not evoke any current.
The morphine effect was attenuated by increasing 5-HT concentrations from 1 μM to 100 μM 5-HT and the agonist concentration-response curve was shifted to the right in the presence of 1 μM morphine (Figs. 4 and 5). Applying the EC50 values of 5-HT in the absence and presence of 1 μM morphine (1.97 and 3.2 μM, respectively) to the Schild Equation, a KB value for morphine of 1.6 μM could be calculated. This value is close to the IC50 value (1.1 μM). In addition, the slowing of current kinetics by morphine is attenuated by increasing 5-HT concentrations.
When the opioid-antagonist naloxone (3 μM) was applied 1 min before and also simultaneously during the agonist pulse (3 μM 5-HT), naloxone caused a moderate but significant (P < 0.01) reduction of the peak current by 17%. However, when applied together, naloxone (3 μM) did not increase but slightly attenuated the pronounced inhibition by morphine (1 μM) described above (Fig. 5).
The present investigation addresses the question whether morphine has specific effects on human 5-HT3A receptors at clinical concentrations. First, however, to validate the whole cell preparation for this purpose, we measured the 5-HT concentration-response curve in whole cells and compared our results with existing data from whole cell patches (12). The EC50 value of 5-HT reported here is in agreement with previous results for whole cell measurements (13). The specific 5-HT3 antagonist, ondansetron, inhibits the 5-HT-evoked currents at nanomolar concentrations, which proves that the measured currents are exclusively mediated by 5-HT3 receptors (Fig. 2B).
The concentration–dependent inhibition by morphine of 5-HT-induced currents (IC50 = 1.1 μM) at least partially involves a specific and competitive mechanism. First, the inhibition was attenuated when the agonist concentration was increased, suggesting competition between morphine and 5-HT. This is further supported by the observation that morphine concentration-dependently slows the activation and inactivation kinetics of the 5-HT-induced currents. Second, the potency of morphine observed here is high when compared to its relative low lipophilicity (logP = 0.89), which argues against a pure “unspecific” mechanism. Third, although the specific μ-receptor antagonist naloxone slightly inhibited 5-HT-induced currents, it did not further enlarge but attenuated the inhibition by morphine. This is consistent with the assumption that both structurally related drugs compete for one common binding site at the 5-HT3A receptor rather than act independently (i.e., additive) from each other. A competitive mechanism with similar potency was also found for rat 5-HT3 receptors (11). However, the moderate discrepancy between the measured IC50 of 1.1 μM and the calculated KB value of 1.6 μM suggests an additional inhibitory (noncompetitive) component, which manifests itself as a reduction in efficacy of 5-HT (Fig. 4).
In the present study, morphine reduced the 5-HT-induced early peak current. After this inhibition, the currents were enhanced by a slowing of the inactivation (see traces in Fig. 2B). This long lasting enhancing effect could correspond to an emetic action of morphine. Assuming that the suppression of 5-HT3 receptors has an antiemetic effect whereas enhancement of 5-HT3 receptors promotes emesis, these results are initially surprising, but there are reports that clinically morphine can have both an emetic as well as an antiemetic effect (14). For instance, it has been shown, in dogs, that administration of methylnaltrexone blocked the peripheral emetic effect of morphine and unmasked a central antiemetic effect of morphine (15). Therefore we assume that the peripheral emetic effect might be evoked via opioid receptors, but in the central nervous system an inhibition of 5-HT3 receptors might contribute to an antiemetic effect. The clinically observed emetic action of morphine may be related to the higher potency of morphine at opioid-receptors than at 5-HT3 receptors.
Concentrations of morphine, which almost completely inhibited 5-HT3 receptors in the present study (≈10 μM), are clinically meaningful, in contrast to IV administration observed in the cerebrospinal fluid during epidural or intrathecal administration (4–10 μM) (16). Thus, if one assumes that 5-HT3 receptors are already completely blocked by morphine after subarachnoid injection, then the administration of additional 5-HT3 receptor antagonists would not be expected to be a meaningful therapeutic strategy to further reduce emesis. In agreement with this reasoning, a previous study reported that the specific 5-HT3 receptor antagonist, tropisetron, did not significantly reduce nausea and vomiting after subarachnoid injection of bupivacaine and morphine (17).
Several studies support the idea that, in addition to the opioid system, the serotonergic system is responsible for modulation of nociception (18). The interaction between opioid and serotonergic pathways is further supported by the observation that enkephalin-containing interneurons in the rat spinal trigeminal nucleus and in the rat spinal cord superficial dorsal horn are innervated by 5-HT-containing fibers (19). Furthermore, enkephalinergic inhibitory neurons appear to mediate serotonin-induced spinal analgesia, at least in part, via activation of 5-HT3 receptors expressed in enkephalinergic neurons (20). Thus the present study adds to the existing evidence of interactions between the opioid and the serotonergic pathways.
Besides direct interactions of morphine with serotonergic receptors, indirect effects could be important: actions of morphine on the reuptake, release or metabolism of 5-HT could also contribute to its emetogenic activity by changes of the free 5-HT concentration and will be the subject of further studies.
The results presented here suggest that the human 5-HT3 receptor is a sensitive and specific target for morphine and, thus, that opioid actions affecting 5-HT3 receptors may be relevant for emesis and analgesia.
We thank Dr. M. Brüss for the supply of HEK 293 cells, stably transfected with the cDNA of the human 5-HT3A receptor.
1. Gomes I, Gupta A, Filipovska J, et al. A role for heterodimerization of mu and delta opiate receptors in enhancing morphine analgesia. Proc Natl Acad Sci U S A 2004;101:5135–9.
2. Kesim M, Duman EN, Kadioglu M, et al. The different roles of 5-HT(2)
receptors on antinociceptive effect of paroxetine in chemical stimuli in mice. J Pharmacol Sci 2005;97:61–6.
3. Hamon M, Gallissot MC, Menard F, et al. 5-HT3
receptor binding sites are on capsaicin-sensitive fibres in the rat spinal cord. Eur J Pharmacol 1989;164:315–22.
4. Greenshaw AJ. Differential effects of ondansetron, haloperidol and clozapine on electrical self-stimulation of the ventral tegmental area. Behav Pharmacol 1993;4:479–85.
5. Giordano J. Analgesic profile of centrally administered 2-methylserotonin against acute pain in rats. Eur J Pharmacol 1991;199:233–6.
6. Glaum SR, Proudfit HK, Anderson EG. Reversal of the antinociceptive effects of intrathecally administered serotonin in the rat by a selective 5-HT3
receptor antagonist. Neurosci Lett 1988;95:313–7.
7. Tecott LH, Maricq AV, Julius D. Nervous system distribution of the serotonin 5-HT3
receptor mRNA. Proc Natl Acad Sci U S A 1993;90:1430–4.
8. Gyermek L. 5-HT3
receptors: pharmacologic and therapeutic aspects. J Clin Pharmacol 1995;35:845–55.
9. Karim F, Roerig SC, Saphier D. Role of 5-hydroxytryptamine3
) antagonists in the prevention of emesis caused by anticancer therapy. Biochem Pharmacol 1996;52:685–92.
10. Barann M, Molderings G, Bruss M, et al. Direct inhibition by cannabinoids of human 5-HT3A
receptors: probable involvement of an allosteric modulatory site. Br J Pharmacol 2002;137:589–96.
11. Fan P. Nonopioid mechanism of morphine modulation of the activation of 5-hydroxytryptamine type 3 receptors. Mol Pharmacol 1995;47:491–5.
12. Brown AM, Hope AG, Lambert JJ, Peters JA. Ion permeation and conduction in a human recombinant 5-HT3
receptor subunit (h5-HT3A
). J Physiol 1998;507:653–65.
13. Sessoms-Sikes JS, Hamilton ME, Liu LX, et al. A mutation in transmembrane domain II of the 5-hydroxytryptamine(3A)
receptor stabilizes channel opening and alters alcohol modulatory actions. J Pharmacol Exp Ther 2003;306:595–604.
14. Krueger H, Eddy N, Sumwalt M. The pharmacology of the opium alkaloids. Public Health Rep 1940;704:165.
15. Foss JF, Yuan CS, Roizen MF, Goldberg LI. Prevention of apomorphine- or cisplatin-induced emesis in the dog by a combination of methylnaltrexone and morphine. Cancer Chemother Pharmacol 1998;42:287–91.
16. Nordberg G, Hedner T, Mellstrand T, Dahlstrom B. Pharmacokinetic aspects of intrathecal morphine analgesia. Anesthesiology 1984;60:448–54.
17. Pitkanen MT, Niemi L, Tuominen MK, Rosenberg PH. Effect of tropisetron, a 5-HT3
receptor antagonist, on analgesia and nausea after intrathecal morphine. Br J Anaesth 1993;71:681–4.
18. Gupta S, Vaswani R, Verma D. The 5HT3
receptors and the descending nociceptive pathway: a review. Middle East J Anesthesiol 1999;15:247–58.
19. Nikfar S, Abdollahi M, Kebriaeezadeh A, Chitsaz M. Effect of granisetron on morphine-induced analgesia in mice by formalin test. Gen Pharmacol 1998;31:55–8.
© 2006 International Anesthesia Research Society
20. Tsuchiya M, Yamazaki H, Hori Y. Enkephalinergic neurons express 5-HT3
receptors in the spinal cord dorsal horn: single cell RT-PCR analysis. Neuroreport 1999;10:2749–53.