British Journal of Anaesthesia 1997; 79: 352 356 Actions of tramadol, its enantiomers and principal metabolite, O-desmethyltramadol, on serotonin (5-HT) efflux and uptake in the rat dorsal raphe nucleus T. A. BAMIGBADE, C. DAVIDSON, R. M. LANGFORD AND J. A. STAMFORD Summary Tramadol is an atypical centrally acting analgesic agent with relatively weak opioid receptor affinity in comparison with its antinociceptive efficacy. Evidence suggests that block of monoamine uptake may contribute to its analgesic actions. Therefore, we have examined the actions of ( )-tramadol, ( )-tramadol, ( )-tramadol and O- desmethyltramadol (M1 metabolite) on electrically evoked 5-HT efflux and uptake in the dorsal raphe nucleus (DRN) brain slice, measured by fast cyclic voltammetry. Racemic tramadol and its ( )- enantiomer (both 5 mol litre 1 ) significantly blocked DRN 5-HT uptake (both P 0.05) and increased stimulated 5-HT efflux (P 0.01 ( )-tramadol; P 0.05 ( )-tramadol). The ( )- enantiomer and metabolite, O-desmethyltramadol, were inactive at the concentration tested (5 mol litre 1 ). For both ( )-tramadol and the ( )- enantiomer, the action on 5-HT efflux preceded an effect on 5-HT uptake, suggesting that uptake block was not the cause of the increased 5-HT efflux and that tramadol might therefore have a direct 5-HT releasing action. This activity, at clinically relevant concentrations, may help to explain the antinociceptive efficacy of tramadol despite weak opioid receptor affinity and adds to evidence that tramadol exerts actions on central monoaminergic systems that may contribute to its analgesic effect. (Br. J. Anaesth. 1997; 79: 352 356). Key words Analgesics opioid, tramadol. Serotonin (5-hydroxytryptamine). Model, rat. Measurement techniques, voltammetry. Tramadol is a centrally acting analgesic agent, effective in the treatment of moderate to moderately severe pain with a relatively low addiction potential. 1 A striking feature of the pharmacology of tramadol is its relatively weak opioid receptor affinity in comparison with its antinociceptive efficacy, indicative of the involvement of other mechanisms. Interestingly, its antinociceptive action in some animal models is only partially antagonized by naloxone, 2 strongly suggesting a non-opioid component to its analgesic potency. Furthermore, tramadol has an active metabolite, O-desmethyltramadol. 3 Analgesia can be achieved both centrally and peripherally by interference with a variety of neurotransmitter systems. In particular, the control of pain is subject to descending modulation by brainstem cell groups such as the locus coeruleus/subcoeruleus and raphe complex. 4 These nuclei contain mainly noradrenaline and serotonin (5-HT), respectively. Drugs that block noradrenaline uptake (such as desipramine) or that stimulate 2 adrenoceptors (e.g. clonidine) are useful adjuncts to standard analgesic therapy in intractable pain and are also analgesic in their own right. Tramadol has been shown to block noradrenaline uptake in cortical synaptosomes and brain slices 5 thereby increasing noradrenaline efflux. 6 This and the fact that 2 adrenoceptor blockade reduces its analgesic efficacy 7 8 indicates that enhancement of noradrenaline function contributes significantly to its analgesic profile. The involvement of 5-HT in the descending control of pain is also widely recognized 4 and tramadol has been shown to enhance basal efflux of [ 3 H]-5-HT in frontocortical synaptosomes. 9 Tramadol also blocks 5-HT uptake 10 and it is thus possible that the effects of tramadol on cortical 5-HT efflux are mediated via uptake blockade. However, the synaptosome is such a highly reduced system, devoid of the transmitter interactions found in vivo, that it is not easy to draw conclusions on the rela tionship between 5-HT efflux a nd upta ke inhibition. Therefore, we chose to examine the actions of tramadol on 5-HT efflux and uptake in the dorsal raphe nucleus (DRN) brain slice, for four reasons. First, the dorsal raphe is part of the brainstem raphe complex and has been shown to have a role in the descending control of pain. 11 Second, the brain slice retains much of the local synaptic integrity and interactions of the whole animal while being easier to manipulate. Third, drug concentrations may be controlled precisely in vitro, thus allowing the elimination of pharmacokinetic considerations that TIMOTHY A. BAMIGBADE, RICHARD M. LANGFORD, Department of Anaesthesia, St Bartholomew s Hospital, West Smithfield, London EC1A 7BE. COLIN DAVIDSON, JONATHAN A. STAMFORD, Anaesthetics Unit (Neurotransmission Laboratory), Royal London Hospital, Whitechapel, London E1 1BB, UK. Accepted for publication: April 2, 1997. Correspondence to J. A. S.
Actions of tramadol on 5-HT in the rat dorsal raphe nucleus 353 might confound data interpretation. Finally, the consequences of metabolism to active (or inactive) metabolites in the whole animal are excluded; this is particularly pertinent for tramadol. Materials and methods Fast cyclic voltammetry (FCV) at carbon fibre microelectrodes was used to determine the effects of tramadol, its enantiomers and principal metabolite on electrically stimulated 5-HT efflux and uptake in rat DRN brain slices. No Home Office-regulated procedures on live animals were performed. PREPARATION OF SLICES OF THE RAT DORSAL RAPHE NUCLEUS Male Wistar rats (150 25 g) were stunned and killed by rapid cervical dislocation. No prior anaesthesia was administered. The brain was excised rapidly and irrigated with ice-cold ( 1 to 1 C) artificial cerebrospinal fluid (acsf). A Campden 752M vibratome was used to prepare 350- m thick brainstem slices ( 1.0 to 1.3 mm vs the interaural line 12 ) containing the DRN (seen as a dove-shaped translucent area, ventral to the cerebral aqueduct). The brain slice was secured in a 1-ml volume superfusion-type brain slice tissue bath 13 by a nylon mesh drawn over a stainless steel grid. The internal temperature of the chamber was maintained at 32 C and the slice was superfused with oxygenated (95% oxygen-5% carbon dioxide) acsf at 1 ml min 1 throughout the experiment. Artificial CSF consisted of NaCl 124 mmol litre 1, NaHCO 3 25 mmol litre 1 ( )-glucose 11 mmol litre 1, KCl 2 mmol litre 1 CaCl 2 2 mmol litre 1, MgSO 4 2 mmol litre 1 and KH 2 PO 4 1.25 mmol litre 1. MEASUREMENT OF 5-HT EFFLUX AND REUPTAKE BY FAST CYCLIC VOLTAMMETRY A glass-coated carbon fibre (8 m in diameter, 50 m in length) recording electrode 14 was inserted 80 m below the surface of the slice, approximately 200 m from a bipolar tungsten stimulating electrode (A-M Systems, Seattle, USA). Figure 1A shows the basic set-up. The auxiliary (a 0.5-mm diameter stainless steel wire) and reference (Ag/AgCl cylinder) electrodes were positioned at a convenient location in the chamber. Qua ntita tive rea l time efflux of 5-HT wa s measured using FCV, as previously described. 15 An input voltage (1.5 cycles of a triangular waveform, 1.0 to 1.4 V vs Ag/AgCl, scan rate of 480 V s 1 ) was applied to the potentiostat (Millar Voltammetric Analyser, PD Systems, West Molesey, UK) every 500 ms. The working electrode was disconnected between scans. The current output of the carbon fibre microelectrode was displayed on a digital storage oscilloscope (Nicolet 310DD). Background (charging) current signals before a stimulation were saved and subtracted from those obtained after a stimulus. Subtraction of one current signal from the other yielded the current caused by 5-HT oxidation and its subsequent reduction. A sample and hold circuit monitored the current at the oxidation potential for 5-HT ( 570 m V vs Ag/AgCl). Its output was displayed on a chart recorder and stored on a microcomputer using CED (Cambridge Electronic Design) Chart software. ELECTRICAL STIMULATION OF 5-HT RELEASE All electrical stimulations were generated with standard Neurolog modules and applied to the stimulating electrode via an NL 800 constant current isolator. The slice was allowed to equilibrate in the chamber for at least 1 h before stimulation was conducted. 5-HT efflux was evoked by trains consisting of 20 pulses (0.1 ms duration, 10 ma constant current at 100 Hz) applied every 10 min. On each stimulation, two variables were recorded: 5-HT efflux and uptake (see fig. 1B). 5-HT efflux was taken as the peak extracellular 5-HT concentration attained after stimulation. On cessation of stimulation, 5-HT was removed from the extracellular space by uptake. The half-time (T 1/2 ) of 5-HT uptake (the time taken for the extracellular 5-HT concentration to decrease to half of the peak concentration attained) was used as an (reciprocal) estimate of the rate of 5-HT uptake. EXPERIMENTAL PROCEDURE After three stable consecutive 5-HT efflux events were obtained (a control period of 30 min), the agent to be studied (at a concentration of 5 mol litre 1 ) was added to the superfusate for 2 h. Five treatment groups were compared: ( )-tramadol, ( )- tramadol, ( )-tramadol, O-desmethyltramadol (M1 metabolite) and control. Controls received acsf. Group sizes were: ( )-tramadol (n 5 slices), ( )-tramadol (n 6), ( )-tramadol (n 6), O- desmethyltramadol (n 7) and control (n 5). DRUGS ( )-Tramadol, ( )-tramadol, ( )-tramadol and O- desmethyltramadol were gifts from Grünenthal GmbH (Germany). Stock solutions of each drug were prepared in distilled water. Subsequent dilutions were made in acsf. STATISTICAL ANALYSIS All drug effects on 5-HT efflux and reuptake were plotted against time. The area under the curve was calculated for each individual experiment and means (SEM) were determined for each series of experiments. The data for ( )-tramadol, ( )-tramadol, ( )-tramadol and O-desmethyltramadol were tested for statistical significance compared with the control group using one-way analysis of variance (ANOVA) with post hoc application of Dunnett s test. Results Electrical stimulation (20 pulses, 0.1 ms, 100 Hz) in the DRN evoked efflux of 5-HT that was detected at
354 British Journal of Anaesthesia Figure 2 Effects of tramadol, its enantiomers and metabolite on stimulated 5-HT efflux in the rat dorsal raphe nucleus slice. Data are expressed as changes in the area under the 5-HT efflux vs time curve, relative to the pre-drug period. All values are mean (SEM) (n 5 7). *P 0.05, **P 0.01 vs control (Dunnett s test). Figure 1 A: Schematic representation of the study showing implantation of the stimulating and working (carbon fibre) electrodes in the brainstem slice. The bird-shaped structure is the dorsal raphe nucleus. B: A typical sample-and-hold record of 5-HT efflux and reuptake after local stimulation in the dorsal raphe nucleus. an adjacent carbon fibre microelectrode. Figure 1 shows a typical rapid 5-HT efflux and reuptake profile after local stimulation in the DRN. Figure 2 shows the effects of tramadol, its enantiomers and main metabolite on stimulated 5-HT efflux. Data are expressed as mean (SEM) post-drug incremental area under the curve. ( )-Tramadol and its ( )-enantiomer significantly increased 5-HT efflux relative to acsf controls (P 0.05 and P 0.01, respectively; Dunnett s test). The ( )- enantiomer and O-desmethyltramadol had no significant effect compared with control. Figure 3 shows the effects of tra ma dol, its enantiomers and metabolite on 5-HT uptake T1/2. As before, data are expressed as mean (SEM) postdrug incremental area under the curve. Similarly, the racemic mixture and ( )-enantiomer significantly slowed 5-HT reuptake (both P 0.01 vs control; Dunnett s test). O-desmethyltramadol and the ( )- enantiomer were inactive. Figure 4 shows the time course of the effects of the active drugs, ( )-tramadol and ( )-tramadol, on stimulated 5-HT efflux and uptake T 1/2. For both Figure 3 Effects of tramadol, its enantiomers and metabolite on 5-HT uptake half-time (T 1/2 ) in the rat dorsal raphe nucleus slice. Data are expressed as changes in the area under the 5-HT uptake T 1/2 vs time curve, relative to the pre-drug period. All values are mean (SEM) (n 5 7). **P 0.01 vs control (Dunnett s test). ( )-tramadol and ( )-tramadol, there was a temporal dissociation between their effects on 5-HT efflux and uptake. In both cases, the drug effects on 5-HT efflux occurred before changes in 5-HT reuptake T 1/2. Discussion Release of neurotransmitter from isolated tissues maintained in vitro has long been a popular means of evaluating monoaminergic uptake mechanisms. 16 However, problems with this technique have been highlighted regarding its applicability to neurotransmitter release within small anatomical regions in the brain. Furthermore, the neurotransmitter pool is frequently labelled with tritium which assumes, probably incorrectly, that release of the radiolabel accurately reflects true neurotransmitter release. 17 We have therefore used FCV at carbon fibre microelectrodes to study the effects of racemic tramadol, its enantiomers and the major human metabolite, O-desmethyltramadol, on endogenous 5-HT neurotransmission in the DRN, a key nucleus in the regulation of afferent nociceptive transmission. An FCV scan takes only milliseconds and can be repeated many times per second, thus giving real time detection of neurotransmitter. This high
Actions of tramadol on 5-HT in the rat dorsal raphe nucleus 355 Figure 4 A: Time course of effects of ( )-tramadol 5 mol litre on stimulated 5-HT efflux and 5-HT uptake halftime (T 1/2 ) in the rat dorsal raphe nucleus slice. B: Time course of effects of (+)-tramadol 5 mol litre on stimulated 5-HT efflux and 5-HT uptake half-time (T 1/2 ) in the rat dorsal raphe nucleus slice. All data are mean (SEM) (n 6). temporal resolution allows the relationship between 5-HT efflux and uptake to be determined precisely (see fig. 1). In each case the drugs were applied at a concentration of 5 mol litre 1. This approximates to the plasma concentration of tramadol after a therapeutically effective dose. 18 Both ( )-tramadol and the ( )-enantiomer significantly enhanced stimulated 5-HT efflux in the DRN at this concentration. Neither ( )-tramadol nor the M1 metabolite, O-desmethyltramadol, had any effect (fig. 2). A qualitatively similar picture was obtained for the effects on 5-HT uptake (fig. 3). Again, both the ( )-enantiomer and the racemate were effective while the metabolite was devoid of action. Of interest, ( )-tramadol delayed 5-HT uptake in some slices although the group data were not statistically significant. These findings are broadly consistent with the literature. Tramadol inhibited 5-HT reuptake, the ( )-enantiomer being the active form. 9 The hint of an effect on uptake with ( )-tramadol also supports the findings of other studies, using synaptosomes and radiolabelled neurotransmitter, indicating that it has approximately 25% of the activity of the ( )-enantiomer on the serotoninergic system. 9 De-methylation of tramadol resulted in a marked reduction of activity on 5-HT reuptake inhibition. This is interesting for two reasons. First, the M1 metabolite is an effective analgesic and probably contributes significantly to the clinical profile of tramadol 1, presumably via opioid receptors. 3 Second, it is clear that the O-methyl function of the parent drug is therefore a structural determinant 5-HT uptake block by tramadol. As only those drugs that blocked 5-HT uptake the ( )-enantiomer and the racemate also increased 5-HT efflux, there is a prima facie case that uptake block is the cause of increased 5-HT efflux. However, such a conclusion is erroneous as the increase in 5-HT efflux after tramadol preceded reduced uptake (fig. 4). For both ( )-tramadol and its ( )-enantiomer, the effects on 5-HT efflux were essentially maximal on the first post-drug stimulation. Effects on 5-HT uptake T 1/2 did not reach maximum until considerably later. Furthermore, with paroxetine, the effect on 5-HT reuptake is much greater than that on 5-HT efflux because of activation of negative feedback mechanisms reducing efflux. 15 Conversely, the effects of ( )-tramadol on 5-HT efflux and reuptake were of a similar magnitude. The rapid effect of racemic tramadol and its ( )-enantiomer on 5-HT efflux cannot be caused by 5-HT 1 autoreceptor block as the choice of shortduration stimulation variables precludes activation of the autoreceptor during the stimulation train. 19 Furthermore, we have shown previously that the potent 5-HT 1A antagonist WAY 100135 did not increase 5-HT efflux in this experimental procedure. 15 Hence it appears that tramadol may have a direct releasing action. Other studies have shown that the tramadol-induced increase in 5-HT efflux does not occur if 5-HT reuptake sites are blocked by 6-nitroquipazine, before tramadol administration. 9
356 British Journal of Anaesthesia This suggests that tramadol can compete with 5-HT for the uptake mechanism and, when inside the presynaptic terminal, increase stimulation-evoked 5-HT release. A direct effect on 5-HT release, dissociated from reuptake inhibition, may be important to the multimodal analgesic actions of tramadol. It is unlikely that opioid receptor activation is responsible for the 5-HT releasing action observed with ( )-tramadol and the racemate as the M1 metabolite had no 5-HT releasing action despite being a more potent opioid than the parent drug. 3 In conclusion, both racemic tramadol and its ( )-enantiomer blocked DRN 5-HT uptake and increased stimulated 5-HT efflux. The temporal dissociation between these actions suggests that uptake blockade was not the cause of the increased 5-HT efflux and that tramadol might therefore have a direct 5-HT releasing action. It should nevertheless be remembered that we have only examined a single, albeit clinically relevant, concentration of tramadol and it would be interesting to examine such actions over an extended concentration range. Nevertheless, the activity of such mechanisms at clinically pertinent concentrations may help explain the antinociceptive efficacy of tramadol despite weak opioid receptor affinity. This study adds to evidence that tramadol exerts actions on central monoaminergic systems and that this mechanism may contribute to its analgesic effect. Acknowledgements We thank Grüenthal GmbH (Germany) for gifts of ( )-tramadol, ( )-tramadol, ( )-tramadol and O- desmethyltramadol. This research was supported, in part, by Searle. We thank Paul Phillips for graphical assistance. References 1. Gibson TP. Pharmacokinetics, efficacy and safety of analgesia with a focus on tramadol HCl. American Journal of Medicine 1996; 101: 47S 53S. 2. Raffa RB, Friderichs E, Reimann W, Shank RP, Codd EE, Vaught JL. Opioid and non-opioid components independently contribute to the mechanism of action of tramadol, an atypical opioid analgesic. Journal of Pharmacology and Experimental Therapeutics 1992; 260: 275 285. 3. Dayer P, Collart L, Desmeules J. The pharmacology of tramadol. Drugs 1994; 47: 3 7. 4. Stamford JA. Descending control of pain. British Journal of Anaesthesia 1995; 75: 217 227. 5. Reimann W, Hennies HH. Inhibition of spinal noradrenaline uptake in rats by the centrally acting analgesic tramadol. Biochemical Pharmacology 1994; 47: 2289 2293. 6. Driessen B, Reimann W, Giertz H. Effects of the central a na lgesic tra ma dol on the upta ke a nd relea se of noradrenaline and dopamine in vitro. British Journal of Pharmacology 1993; 108: 806 811. 7. Kayser V, Besson JM, Guilbaud G. Evidence for a noradrenergic component in the antinociceptive effect of the analgesic tramadol in an animal model of clinical pain, the arthritic rat. European Journal of Pharmacology 1992; 224: 83 88. 8. Desmeules JA, Piguet V, Collart, L, Dayer P. Contribution of monoaminergic modulation to the analgesic effect of tramadol. British Journal of Clinical Pharmacology 1996; 41: 7 12. 9. Driessen B, Reimann W. Interaction of the central analgesic tramadol with the uptake and release of 5-hydroxytryptamine in the rat brain in vitro. British Journal of Pharmacology 1992; 105: 147 151. 10. Raffa RB, Friderichs E, Reimann W, Shank RP, Codd EE, Va ught JL, Ja coby HI, Selve N. Complementa ry a nd synergistic antinociceptive interaction between the enantiomers of tramadol. Journal of Pharmacology and Experimental Therapeutics 1993; 267: 331 340. 11. Oliveras JL, Guilbaud G, Besson JM. A map of serotonergic structures involved in stimulation producing analgesia in unrestrained freely moving cats. Brain Research 1979; 164: 317 322. 12. Pa xinos G, Wa tson C. The Rat Brain in Stereotaxic Coordinates. London: Academic Press, 1986. 13. Richards CD, Tegg WJB. A superfusion chamber suitable for maintaining mammalian brain tissue slices for electrical recording. British Journal of Pharmacology 1977; 59: 526P. 14. Armstrong James M, Millar J. Carbon fibre microelectrodes. Journal of Neuroscience Methods 1979; 1: 279 287. 15. Davidson C, Stamford JA. The effect of paroxetine on 5-HT efflux in the rat dorsal raphe nucleus is potentiated by both 5-HT 1a and 5-HT 1b/d receptor antagonists. Neuroscience Letters 1995; 188: 41 44. 16. Farnebo LO, Hamberger B. Drug induced changes in the release of H monoamines from field stimulated rat brain slices. Acta Physiologica Scandinavica 1971; 371: 35 44. 17. Herdon H, Strupish J, Nahorski S. Differences between the release of radiolabelled and endogenous dopamine from superfused rat brain slices: Effects of depolarising stimuli, amphetamine and synthesis inhibition. Brain Research 1985; 348: 309 320. 18. Jellinek H, Haumer H, Grubhofer G, Klappacher G, Jenny T, Weindlmayr-Goettel M. Tramadol in postoperative pain therapy. Patient-controlled analgesia versus continuous infusion. Anaesthesist 1990; 39: 513 20. 19. Singer EA. Transmitter release from brain slices elicited by single pulses: a powerful method to study presynaptic mechanisms. Trends in Pharmacological Sciences 1988; 9: 274 276.