- Email: [email protected]
Antinociceptive role of 5-HT1A receptors in rat spinal cord
British Journal of Anaesthesia 88 (5): 679±84 (2002)
LABORATORY INVESTIGATIONS Antinociceptive role of 5-HT1A receptors in rat spinal cord R. Nadeson and C. S. Goodchild* Monash University Department of Anaesthesia, Monash Medical Centre, Level 5 Block E, Clayton, Victoria 3168, Australia *Corresponding author Background. Intrathecal administration of 5-hydroxytryptamine (5-HT) is antinociceptive to noxious heat and electrical stimuli. The contributions of different receptor subtypes to the antinociceptive effects of 5-HT are controversial. The main reasons for this are the poor receptor subtype selectivity of some agonist drugs and the dif®culty of restricting drug action to the spinal cord in some experimental paradigms. This study investigated the roles of different 5-HT receptor subtypes involved in the spinal cord control of the nociception produced by these two nociceptive testing paradigms. Methods. Tail-¯ick latency and electric current threshold for nociception were measured in an acute pain model that allowed the study of the antinociceptive effects of intrathecally administered drugs that were due to actions of these drugs at spinal cord receptors. Experiments were performed in male Wistar rats with chronically implanted lumbar subarachnoid catheters. Dose±response curves for spinally mediated antinociceptive effects of agonists selective for 5-HT receptor subtypes were constructed. Results. The 5-HT1 agonist 1-(3-chlorophenyl)-piperazine dihydrochloride caused a dosedependent antinociceptive effect, measured by both nociceptive tests. However, 8-hydroxyDPAT (selective 5-HT1A agonist) produced antinociception assessed by electric current but not tail ¯ick. A 5-HT1A-selective antagonist, 4-[3-(benzotriazol-1-yl)propyl]-1-(2-methoxyphenyl)-piperazine, reversed the antinociception in the electrical test produced by both of these agonists but the tail-¯ick latency effects after intrathecal 1-(3-chlorophenyl)-piperazine were not suppressed by this antagonist. Conclusions. We conclude that 5-HT1A receptors in the spinal cord are involved in the nociceptive mechanisms assessed by noxious electrical stimuli. Other 5-HT1 receptors (non 5-HT1A receptors) are involved in the spinally mediated antinociception assessed by thermal noxious stimuli. Br J Anaesth 2002; 88: 679±84 Keywords: rat; serotonin (5-hydroxytryptamine); spinal cord; sensory block; pain Accepted for publication: December 21, 2001
The antinociceptive effects of spinally administered 5hydroxytryptamine (5-HT) are well established.1±3 Intrathecally administered 5-HT mimics the effect of electrical activity of tracts descending from neurones located in brainstem raphe nuclei, electrical stimulation of which also produces antinociceptive effects in single dorsal horn neurones and awake animals.4 5 The 5-HT antagonist methysergide has been shown to reverse the antinociception produced by spinal 5-HT and to produce hyperalgesia in rats
when used alone intrathecally.3 6±10 These latter experiments suggest that 5-HT receptors are involved in spinal antinociception. More recent experiments have shown the involvement of subtypes of 5-HT receptors in the spinal cord control of nociception. In mice, a 5-HT2 agonist given intrathecally produced biting and scratching behaviour, whereas a 5-HT1A agonist caused antinociception, assessed by tail-¯ick latency (TFL).11 Receptors for 5-HT1A have also been implicated in correlates of nociception in single
Ó The Board of Management and Trustees of the British Journal of Anaesthesia 2002
Nadeson and Goodchild
neurones in the spinal cord dorsal horn.12 Others have concluded, from experiments in which a selective 5-HT1A antagonist (WAY-100635) was given intrathecally, that 5-HT1A receptors in the spinal cord are responsible for the bulbospinally mediated antinociceptive effects of fentanyl administered intracereboventricularly.13 Much experimental work has been published in this area, but the question of which subtypes of 5-HT receptor are involved, particularly in the spinal cord control of nociception, is still controversial. This controversy stems from the lack of selectivity of the drugs for the receptors used in some studies and from the inability of some experimental paradigms to limit intrathecal drug action to the spinal cord. Thus, incorrect conclusions about the subtype of 5-HT receptors involved in the spinal cord control of nociception could be made if the drug administered has activity at other 5-HT receptors at the same dose, or if the drug spreads away from the spinal cord site of injection to sites with 5-HT receptors in the brain that then mediate the antinociceptive effects. This has clear clinical implications for the development of a drug targeting spinal cord receptors for analgesia, perhaps by spinal injection. It has been shown that intrathecal injection of 5-HT in such a way that its actions are con®ned to the spinal cord causes antinociception, as revealed by the response to noxious electrical stimuli [electric current threshold (ECT) test] and by TFL.14 Methysergide reverses these antinociceptive effects, indicating the involvement of spinal cord 5-HT receptors. The investigators went on to show, in selective antagonist and cross-tolerance experiments, that the antinociception revealed by the noxious electrical stimulation paradigm (but not TFL) involved spinal cord m opioid and GABAA receptors.15 16 Because different nociceptive testing paradigms were used, it is uncertain what 5-HT receptor subtypes were involved in the m opioid and GABAA receptor interaction. The purpose of the present study was to investigate the involvement of 5-HT1 receptor subtypes in the spinal cord control of nociception, assessed in the ECT and TFL paradigms. The response to noxious electrical stimulation and TFL were measured in a model that allowed the study of drug actions that were con®ned to the caudal segments of the spinal cord. Drugs selective for different 5-HT1 receptor subtypes were injected intrathecally into conscious rats with chronically implanted lumbar subarachnoid catheters. Antinociceptive effects were assessed with noxious heat and electric current. Antagonism of these effects was investigated by intrathecal injection of a selective 5-HT1A antagonist.
Methods Experiments were performed on male Wistar rats (weight 185612 g (mean6SEM)). In all experiments, we followed international guidelines for the care of laboratory and experimental animals and guidelines for the investigation of
pain in laboratory animals.17 This work was carried out with permission of the Monash University Standing Committee on Ethics in Animal Experimentation (SCEAE Project No. 93017).
Catheter implantation Lumbar intrathecal catheters (dead space volume 3±5 ml) were implanted under halothane anaesthesia and aseptic surgical conditions as described previously.18 Portex catheters were introduced into the lumbar intrathecal space to lie next to the most caudal segments of the spinal cord, responsible for innervation of the tail and hind legs. The catheters were tunneled subcutaneously to an exit wound at the base of the neck. The exact volume of the catheter was measured before insertion, thus allowing accurate dosing of drugs injected through the catheter into the cerebrospinal ¯uid (CSF). Correct catheter placement was con®rmed by injecting 2% lidocaine solution 10 ml down the catheter 10 min after recovery from general anaesthesia. The catheter was judged to be intrathecal if paralysis and dragging of the hind legs occurred within 30 s of this injection. All animals failing this test and those with neurological de®cit after surgery were rejected from further study. The lidocaine test was also performed after each experiment to con®rm that all drugs injected down the intrathecal catheters were injected into the lumbar intrathecal space. The rats were rested for 24 h before further experiments were performed. Only one experiment per day up to a maximum of ®ve was performed on each animal.
Nociceptive tests Two nociceptive tests were used in all experiments, the ECT test and the TFL test, as described previously.18 The electrical test allows demonstration of drug action at the level of the spinal cord after lumbar intrathecal injection, i.e. the measurement of an increase in nociceptive threshold in the tail with no change in the neck threshold measurements. Such a differential block proves that the effects of the drug are con®ned to the caudal segments of the spinal cord responsible for tail and hind limb innervation. A rise in TFL in the same experiment in which such a differential effect is demonstrated must also therefore be due to action of the drug on the caudal segments of the spinal cord. Both of these tests were applied in sequence to each rat every 5 min. Electric current threshold test
Each rat was placed in a darkened restrainer and electrical stimulating electrodes were applied to the skin of the tail and neck. Brief bursts of electric current (0.5 s train, 2 ms pulse width, 50 Hz, 0±10 mA) were delivered in turn to each set of electrodes (tail ®rst, 15 s before the neck) every 5 min. The minimum current necessary to evoke an obvious nociceptive response (a squeak or sharp withdrawal), de®ned by the `up±down' method, was measured at each
680
Spinal cord 5-HT1A receptors and antinociception
skin site every 5 min until three stable consecutive readings had been obtained. The intrathecal drug was then given and measurements continued at each skin site every 5 min for a further 45±50 min. In order to standardize results obtained at each skin site, the response to intrathecal drug was calculated as a ratio with respect to the mean of the three stable readings taken before intrathecal drug injection, as described previously.18 These values for responses to each dose of drug were combined and expressed as the mean (SEM) on agonist dose±response curves. Tail-¯ick latency test
This was performed every 5 min immediately before the electrical measurement in the tail. The heat from a 150 watt projector bulb was directed onto the blackened tip of the tail and the power adjusted so that the rat ¯icked its tail away from the noxious stimulus in approximately 3 s. At this intensity of stimulation, three stable control (pre-drug injection) readings were obtained and then the test drug was injected. TFL measurement continued every 5 min thereafter for 30±45 min. The response to each dose of drug was calculated as a percentage of maximum possible effect (MPE): %MPE
TFL post-drug ÿ TFL pre-drug 100: cut-off time ÿ TFL pre-drug
Antagonist dose±response relationships Doses of agonist drugs were chosen from their dose±response curves for the experiments with antagonists. These were the roughly equieffective doses that produced a nearmaximal response for tail ECT. This dose was 0.1 mmol for 8-hydroxy-DPAT hydrobromide and 0.5 mmol for 1-(3chlorophenyl)-piperazine dihydrochloride. The agonist was given intrathecally alone at the beginning and at the end of a series of once-daily experiments in each rat. In the intervening days the same dose of the same agonist was given intrathecally combined with a range of doses (0.001±0.05 mmol) of a 5-HT1A-selective antagonist [4-[3(benzotriazol-1-yl)propyl]-1-(2-methoxyphenyl)-piperazine] dissolved in 6% dextrose. Twenty experiments were performed in nine rats using 8-hydroxy-DPAT hydrobromide (n=4±5 observations per dose of antagonist). Eleven rats were used in 27 experiments with 1-(3-chlorophenyl)piperazine dihydrochloride (n=4±5 observations per dose of antagonist). Antinociceptive effects were measured using ECT in the tail and neck and TFL. The pre-series and postseries measurements of the antinociceptive effect for agonist alone in each rat were then averaged (R). The percentage suppression of agonist effect by each dose of antagonist was calculated using the following equation, described previously:18 % suppression
A cut-off time of 10 s was used. Values were combined for each drug dose, expressed as mean (SEM) and used to produce dose±response curves.
Agonist dose±response relationships A range of doses (0.005±5.0 mmol) of 1-(3-chlorophenyl)piperazine dihydrochloride (non-selective agonist of 5HT1)19 20 was given intrathecally to six rats with chronically implanted subarachnoid catheters. Sixteen experiments were performed (n=4 per dose of agonist). Similarly, a range of doses (0.01±0.5 mmol) of 8-hydroxy-2-dipropylaminotetralin hydrobromide (a 5-HT1A-selective agonist)21 22 was given intrathecally to 15 rats with chronically implanted subarachnoid catheters. Twenty-three experiments were performed (n=4±6 per dose of agonist). The drugs were dissolved in 6% dextrose solution to make the injectate hyperbaric compared with CSF. Doses were delivered into the lumbosacral CSF in 5 ml volumes while the rat was held at a 15° inclined plane, head up. This was done to restrict the spread of drug to the caudal segments of the spinal cord, as described previously.23 The dextrose solution has been shown not to affect nociceptive thresholds when given alone.18 Once-daily injections of agonist drugs were given on up to ®ve occasions to individual rats. This has been shown in previous work on intrathecal 5-HT not to cause tolerance or tachyphylaxis.15 16
Rÿr 10 Rÿ1
where r is the individual response to agonist mixed with a particular dose of antagonist and R the mean response to agonist given intrathecally alone. The values of percentage suppression were calculated for both ECT and TFL tests. Values were combined to produce means and SEM, which were then plotted as dose±response relationships for suppression of the antinociceptive effects of each agonist.
Results Positive lidocaine tests were obtained in each rat and after each experiment that contributed values to the dose±response relationships reported in this paper. Thus, all agonist and antagonist drugs injected into intrathecal catheters were delivered to the lumbosacral intrathecal space. Both agonist drugs caused dose-related increases in tail ECT with no change in neck ECT. Thus, both drugs caused antinociception by an action in the caudal segments of the spinal cord responsible for tail innervation (Figs 1 and 2). The non-selective 5-HT1 agonist 1-(3-chlorophenyl)piperazine dihydrochloride caused simultaneous doserelated antinociceptive effects, as assessed with the TFL test (Fig. 1). By contrast, the selective 5-HT1A agonist 8hydroxy-DPAT hydrobromide caused no such TFL effects, the antinociception being con®ned to ECT (Fig. 2).
681
Nadeson and Goodchild
Fig 1 Dose±response curves for the antinociceptive effects of the nonselective 5-HT1 agonist 1-(3-chlorophenyl)-piperazine dihydrochloride in the ECT and TFL tests. Intrathecal 1-(3-chlorophenyl)-piperazine dihydrochloride produced dose-related antinociception in both tests. ECT (neck) values did not increase after intrathecal injections of 1-(3chlorophenyl)-piperazine dihydrochloride, indicating that the drug was con®ned in its actions to the caudal segments of the spinal cord responsible for tail innervation. Data points are means and bars are SEM (n=4).
Figure 3 shows dose±response curves for the suppression of tail ECT and TFL antinociceptive effects of 1-(3chlorophenyl)-piperazine dihydrochloride (Fig. 3A) and 8hydroxy-DPAT hydrobromide (Fig. 3B) by the antagonist 4[3-(benzotriazol-1-yl)propyl]-1-(2-methoxyphenyl)-piperazine. This selective 5-HT1A antagonist caused dose-related antagonism of the antinociception assessed by the ECT test for both agonist drugs. However, the TFL effects of the nonselective 5-HT1 receptor agonist 1-(3-chlorophenyl)-piperazine dihydrochloride were unaffected by intrathecal injection of the selective 5-HT1A antagonist (Fig. 3A).
Discussion Many studies have reported antinociception after intrathecal administration of serotonergic agonists.1±3 11 12 24±26 Many of these involved injections of drugs via catheters reported to be intrathecal but no proof of catheter position and site of drug action at the level of the spinal cord was given. Although conclusions about spinal cord 5-HT receptors were drawn, the antinociceptive effects observed could have been due to actions of the drugs on the brain after spread via CSF. The study reported in this paper concentrated on the restriction of drug actions to the spinal cord and the provision of support for such con®nement. The experiments were performed in rats with positive lidocaine tests, and antinociceptive effects occurred only in the tail, not the
Fig 2 Dose±response curves for the antinociceptive effects of the selective 5-HT1A agonist 8-hydroxy-DPAT in the ECT and TFL tests. Intrathecal 8-hydroxy-DPAT caused dose-related antinociception in the ECT test only. TFL was not affected by intrathecal injections of 8hydroxy-DPAT. ECT (neck) values did not increase after intrathecal injections of 8-hydroxy-DPAT, indicating that the drug was con®ned in its actions to the caudal segments of the spinal cord responsible for tail innervation. Data points are means and bars are SEM (n=4±6).
neck. It has been shown previously in this model that intrathecally administered 5-HT caused its ECT antinociceptive effects by involvement of spinal cord GABA and m opioid receptors but no such interactions occurred for the TFL effects.16 23 The results of the experiments reported in this paper extend these ®ndings. Again, there was a difference in the mechanisms of antinociception revealed by the TFL and ECT tests. The ECT test reveals antinociception involving 5-HT1A receptors, whereas the TFL test involves other 5-HT1 (but not 5-HT1A) receptors. One study in mice used the same 5-HT1A-selective agonist as that used in the present study (8-hydroxyDPAT).11 The investigators concluded that 5-HT1A receptors in the spinal cord were responsible for TFL antinociception. This result is contrary to the ®ndings of the experiments reported in this paper. This may represent a supraspinal effect of the drug. Many of these drugs are very soluble in water and thus spread easily throughout the CSF. The investigators made no measurements, such as a comparison of rostral and caudal dermatome nociceptive thresholds, that would indicate con®nement of drug action to the caudal segments of the spinal cord. The importance of distinguishing spinal and supraspinal effects was highlighted recently in a study of the involvement of 5-HT1A receptors in opioid-induced bulbospinal inhibition.13 It was shown that fentanyl injected into CSF in the brain inhibited sural±gastrocnemius re¯ex responses. Intrathecal administration of the selective 5-HT1A receptor antagonist WAY-100635 signi®cantly reduced this inhibition. In a separate group of experiments, i.v. fentanyl
682
Spinal cord 5-HT1A receptors and antinociception
Fig 3 Dose±response curves for reversal of the antinociceptive effects of (A) 1-(3-chlorophenyl)-piperazine dihydrochloride (non-selective 5-HT1 agonist) and (B) 8-hydroxy-DPAT (selective 5-HT1A agonist) by the selective 5-HT1A antagonist 4-[3-(benzotriazol-1-yl)propyl]-1-(2-methoxyphenyl)piperazine in the ECT and TFL tests. The antagonist suppressed the antinociceptive effects of both agonists in the ECT test in a dose-related manner. [Antinociception revealed by TFL after intrathecal injections of the non-selective 5-HT1 agonist 1-(3-chlorophenyl)-piperazine dihydrochloride was not affected by intrathecal injections of the antagonist 4-[3-(benzotriazol-1-yl)propyl]-1-(2-methoxyphenyl)-piperazine.] ECT (neck) values did not increase after intrathecal injections of either agonist in these experiments, indicating that the agonist drugs were con®ned in their actions to the caudal segments of the spinal cord responsible for tail innervation. Points shown are means and bars are SEM (n=4±5).
depressed the sural±gastrocnemius re¯ex. This inhibition was not affected by intrathecal WAY-100635 (100 mg), but combined administration of the 5-HT1A antagonist with an a2 adrenoceptor antagonist (RX 821002) signi®cantly reduced the effects of i.v. fentanyl. They concluded that the bulbospinal and direct spinal actions of fentanyl may occur together to produce overall inhibition of the re¯ex.13 Other studies have claimed interaction of serotonergic receptors with other neurotransmitter systems in the spinal cord. It was reported that a 5-HT3 receptor selective agonist (2-methyl serotonin) caused antinociceptive effects that involved spinal opioid and GABA receptor systems.27 When this drug was injected intracerebroventricularly no antinociception was observed. It was therefore concluded that the antinociception after intrathecal injection was a spinal cord effect. However, 2-methyl serotonin does have af®nity for 5-HT1 receptors.28 Selectivity of drugs for 5-HT receptor subtypes is often not high. Frequently, pKi values for a drug binding to different receptor subtypes differ by one log unit or less.28 29 Thus, Giordano27 may have been describing the system described in this paperÐthe activation of 5-HT1A receptors in the spinal cord. However, more con®dence may be placed in the ®ndings of the present paper because the 5-HT1A-selective drugs that were used have pKi values for the 5-HT1A receptor subtype that are two or three log units greater than values for other receptors.28 29 Clearly, the involvement of 5-HT1 receptors other than 5-
HT1A receptors and of 5-HT2 and 5-HT3 receptors needs to be investigated because of the many reports of antinociceptive effects after intrathecal drugs reported to be selective for these 5-HT receptor subtypes. Furthermore, more detailed dose±response studies must await the arrival of more selective agonist and antagonist drugs that are also water-soluble, so that it will be possible to inject these drugs intrathecally and to study actions con®ned to the spinal cord. These issues are clearly important if there is to be development of a drug designed to target the spinal cord 5-HT receptors involved in antinociception. For example, an intrathecal drug selective for a 5-HT receptor subtype that is only involved in antinociception in the brain will work only if a high dose of the drug is given intrathecally and the drug subsequently spreads to the brain. Such an approach to therapy will inevitably lead to side-effects. We conclude that 5-HT1A receptors in the spinal cord are involved in nociceptive mechanisms that can be assessed by measuring responses to noxious electrical stimuli. Other 5HT1 receptors (non-5-HT1A receptors) are involved in spinally mediated antinociception that can be assessed by measuring responses to thermal noxious stimuli.
References
683
1 Headley PM, Duggan AW, Griersmith BT. Selective reduction by noradrenaline and 5-hydroxytryptamine of nociceptive
Nadeson and Goodchild
2 3 4 5
6 7 8
9 10 11 12 13 14 15
responses in cat dorsal horn neurones. Brain Res 1978; 145: 185±9 Wang JK. Antinociceptive effect of intrathecally administered serotonin. Anesthesiology 1977; 47: 269±71 Yaksh TL, Wilson PR. Spinal serotonin terminal systems mediate antinociception. J Pharmacol Exp Ther 1979; 208: 446±53 Watkins LR, Cobelli DA, Faris P, Aceto MD, Mayer DJ. Opiate vs non-opiate footshock-induced analgesia (FSIA): the body region shocked is a critical factor. Brain Res 1982; 242: 299±308 Rivot JP, Chaouch A, Besson JM. Nucleus raphe magnus modulation of response of rat dorsal horn neurons to unmyelinated ®ber inputs: partial involvement of serotonergic pathways. J Neurophysiol 1980; 44: 1039±57 Glaum SR, Proud®t HK, Anderson EG. 5-HT3 receptors modulate spinal nociceptive re¯exes. Brain Res 1990; 510: 12±16 Holohean AM, Hackman JC, Davidoff RA. Modulation of frog spinal cord interneuronal activity by activation of 5-HT3 receptors. Brain Res 1995; 704: 184±90 Mjellem N, Lund A, Eide PK, Storkson R, Tjolsen A. The role of 5-HT1A and 5-HT1B receptors in spinal nociceptive transmission and in the modulation of NMDA induced behaviour. Neuroreport 1992; 3: 1061±4 Proud®t HK, Hammond DL. Alterations in nociceptive threshold and morphine-induced analgesia produced by intrathecally administered amine antagonists. Brain Res 1981; 218: 393±9 Zemlan FP, Kow LM, Pfaff DW. Spinal serotonin (5-HT) receptor subtypes and nociception. J Pharmacol Exp Ther 1983; 226: 477±85 Eide PK, Hole K. Different role of 5-HT1A and 5-HT2 receptors in spinal cord in the control of nociceptive responsiveness. Neuropharmacology 1991; 30: 727±31 Gjerstad J, Tjolsen A, Hole K. The effect of 5-HT1A receptor stimulation on nociceptive dorsal horn neurones in rats. Eur J Pharmacol 1996; 318: 315±21 Clarke RW, Ward RE. The role of 5-HT1A-receptors in fentanylinduced bulbospinal inhibition of a spinal withdrawal re¯ex in the rabbit. Pain 2000; 85: 239±45 Guo Z, Gent JP, Goodchild CS. Intrathecal 5-hydroxytryptamine produces analgesia by activating spinal cord opioid pathways. Pain 1990; 43 Suppl 5: S103 Goodchild CS, Guo Z, Freeman J, Gent JP. 5-HT spinal antinociception involves mu opioid receptors: cross tolerance and antagonist studies. Br J Anaesth 1997; 78: 563±9
684
16 Nadeson R, Guo Z, Porter V, Gent JP, Goodchild CS. GABAA receptors and spinally-mediated antinociception in rats. J Pharmacol Exp Ther 1996; 278: 620±6 17 Zimmerman M. Ethical guidelines for investigation of experimental pain on conscious animals. Pain 1983; 16: 109±10 18 Edwards M, Serrao JM, Gent JP, Goodchild CS. On the mechanism by which midazolam causes spinally mediated analgesia. Anesthesiology 1990; 73: 273±7 19 Glennon RA. Central serotonin receptors as targets for drug research. J Med Chem 1987; 30: 1±12 20 Lucki I, Ward HR, Frazer A. Effect of 1-(m-chlorophenyl)piperazine and 1-(m-tri¯uoromethylphenyl)piperazine on locomotor activity. J Pharmacol Exp Ther 1989; 249: 155±64 21 Corn®eld LJ, Lambert G, Arvidsson LE, et al. Intrinsic activity of enantiomers of 8-hydroxy-2-(di-n-propylamino)tetralin and its analogs at 5-hydroxytryptamine1A receptors that are negatively coupled to adenylate cyclase. Mol Pharmacol 1991; 39: 780±7 22 Bjork L, Hook BB, Nelson DL, Anden NE, Hacksell U. Resolved N,N-dialkylated 2-amino-8-hydroxytetralins: stereoselective interactions with 5-HT1A receptors in the brain. J Med Chem 1989; 32: 779±83 23 Goodchild CS, Guo Z, Freeman J, Gent JP. 5-HT spinal antinociception involves mu opioid receptors: cross tolerance and antagonist studies. Br J Anaesth 1997; 78: 563±9 24 Bardin L, Bardin M, Lavarenne J, Eschalier A. Effect of intrathecal serotonin on nociception in rats: in¯uence of the pain test used. Exp Brain Res 1997; 113: 81±7 25 FuÈrst S. Transmitters involved in antinociception in the spinal cord. Brain Res Bull 1999; 48: 129±41 26 Zhang Z-H, Yang S-W, Chen J-Y, Xie Y-F, Qiao J-T, Dafny N. Interaction of serotonin and norepinephrine in spinal antinociception. Brain Res Bull 1995; 38: 167±71 27 Giordano J. Analgesic pro®le of centrally administered 2methylserotonin against acute pain in rats. Eur J Pharmacol 1991; 199: 233±6 28 Hoyer D. 5-Hydroxytryptamine receptors and effector coupling mechanisms in peripheral tissues. In: Fozard J, ed. Peripheral Actions of 5-HT. Oxford: Oxford University Press, 1989; 72±99 29 Baxter G, Kennett G, Blaney F, Blackburn T. 5-HT2 receptor subtypes: a family re-united? Trends Pharmacol Sci 1995; 16: 105±10
LABORATORY INVESTIGATIONS Antinociceptive role of 5-HT1A receptors in rat spinal cord R. Nadeson and C. S. Goodchild* Monash University Department of Anaesthesia, Monash Medical Centre, Level 5 Block E, Clayton, Victoria 3168, Australia *Corresponding author Background. Intrathecal administration of 5-hydroxytryptamine (5-HT) is antinociceptive to noxious heat and electrical stimuli. The contributions of different receptor subtypes to the antinociceptive effects of 5-HT are controversial. The main reasons for this are the poor receptor subtype selectivity of some agonist drugs and the dif®culty of restricting drug action to the spinal cord in some experimental paradigms. This study investigated the roles of different 5-HT receptor subtypes involved in the spinal cord control of the nociception produced by these two nociceptive testing paradigms. Methods. Tail-¯ick latency and electric current threshold for nociception were measured in an acute pain model that allowed the study of the antinociceptive effects of intrathecally administered drugs that were due to actions of these drugs at spinal cord receptors. Experiments were performed in male Wistar rats with chronically implanted lumbar subarachnoid catheters. Dose±response curves for spinally mediated antinociceptive effects of agonists selective for 5-HT receptor subtypes were constructed. Results. The 5-HT1 agonist 1-(3-chlorophenyl)-piperazine dihydrochloride caused a dosedependent antinociceptive effect, measured by both nociceptive tests. However, 8-hydroxyDPAT (selective 5-HT1A agonist) produced antinociception assessed by electric current but not tail ¯ick. A 5-HT1A-selective antagonist, 4-[3-(benzotriazol-1-yl)propyl]-1-(2-methoxyphenyl)-piperazine, reversed the antinociception in the electrical test produced by both of these agonists but the tail-¯ick latency effects after intrathecal 1-(3-chlorophenyl)-piperazine were not suppressed by this antagonist. Conclusions. We conclude that 5-HT1A receptors in the spinal cord are involved in the nociceptive mechanisms assessed by noxious electrical stimuli. Other 5-HT1 receptors (non 5-HT1A receptors) are involved in the spinally mediated antinociception assessed by thermal noxious stimuli. Br J Anaesth 2002; 88: 679±84 Keywords: rat; serotonin (5-hydroxytryptamine); spinal cord; sensory block; pain Accepted for publication: December 21, 2001
The antinociceptive effects of spinally administered 5hydroxytryptamine (5-HT) are well established.1±3 Intrathecally administered 5-HT mimics the effect of electrical activity of tracts descending from neurones located in brainstem raphe nuclei, electrical stimulation of which also produces antinociceptive effects in single dorsal horn neurones and awake animals.4 5 The 5-HT antagonist methysergide has been shown to reverse the antinociception produced by spinal 5-HT and to produce hyperalgesia in rats
when used alone intrathecally.3 6±10 These latter experiments suggest that 5-HT receptors are involved in spinal antinociception. More recent experiments have shown the involvement of subtypes of 5-HT receptors in the spinal cord control of nociception. In mice, a 5-HT2 agonist given intrathecally produced biting and scratching behaviour, whereas a 5-HT1A agonist caused antinociception, assessed by tail-¯ick latency (TFL).11 Receptors for 5-HT1A have also been implicated in correlates of nociception in single
Ó The Board of Management and Trustees of the British Journal of Anaesthesia 2002
Nadeson and Goodchild
neurones in the spinal cord dorsal horn.12 Others have concluded, from experiments in which a selective 5-HT1A antagonist (WAY-100635) was given intrathecally, that 5-HT1A receptors in the spinal cord are responsible for the bulbospinally mediated antinociceptive effects of fentanyl administered intracereboventricularly.13 Much experimental work has been published in this area, but the question of which subtypes of 5-HT receptor are involved, particularly in the spinal cord control of nociception, is still controversial. This controversy stems from the lack of selectivity of the drugs for the receptors used in some studies and from the inability of some experimental paradigms to limit intrathecal drug action to the spinal cord. Thus, incorrect conclusions about the subtype of 5-HT receptors involved in the spinal cord control of nociception could be made if the drug administered has activity at other 5-HT receptors at the same dose, or if the drug spreads away from the spinal cord site of injection to sites with 5-HT receptors in the brain that then mediate the antinociceptive effects. This has clear clinical implications for the development of a drug targeting spinal cord receptors for analgesia, perhaps by spinal injection. It has been shown that intrathecal injection of 5-HT in such a way that its actions are con®ned to the spinal cord causes antinociception, as revealed by the response to noxious electrical stimuli [electric current threshold (ECT) test] and by TFL.14 Methysergide reverses these antinociceptive effects, indicating the involvement of spinal cord 5-HT receptors. The investigators went on to show, in selective antagonist and cross-tolerance experiments, that the antinociception revealed by the noxious electrical stimulation paradigm (but not TFL) involved spinal cord m opioid and GABAA receptors.15 16 Because different nociceptive testing paradigms were used, it is uncertain what 5-HT receptor subtypes were involved in the m opioid and GABAA receptor interaction. The purpose of the present study was to investigate the involvement of 5-HT1 receptor subtypes in the spinal cord control of nociception, assessed in the ECT and TFL paradigms. The response to noxious electrical stimulation and TFL were measured in a model that allowed the study of drug actions that were con®ned to the caudal segments of the spinal cord. Drugs selective for different 5-HT1 receptor subtypes were injected intrathecally into conscious rats with chronically implanted lumbar subarachnoid catheters. Antinociceptive effects were assessed with noxious heat and electric current. Antagonism of these effects was investigated by intrathecal injection of a selective 5-HT1A antagonist.
Methods Experiments were performed on male Wistar rats (weight 185612 g (mean6SEM)). In all experiments, we followed international guidelines for the care of laboratory and experimental animals and guidelines for the investigation of
pain in laboratory animals.17 This work was carried out with permission of the Monash University Standing Committee on Ethics in Animal Experimentation (SCEAE Project No. 93017).
Catheter implantation Lumbar intrathecal catheters (dead space volume 3±5 ml) were implanted under halothane anaesthesia and aseptic surgical conditions as described previously.18 Portex catheters were introduced into the lumbar intrathecal space to lie next to the most caudal segments of the spinal cord, responsible for innervation of the tail and hind legs. The catheters were tunneled subcutaneously to an exit wound at the base of the neck. The exact volume of the catheter was measured before insertion, thus allowing accurate dosing of drugs injected through the catheter into the cerebrospinal ¯uid (CSF). Correct catheter placement was con®rmed by injecting 2% lidocaine solution 10 ml down the catheter 10 min after recovery from general anaesthesia. The catheter was judged to be intrathecal if paralysis and dragging of the hind legs occurred within 30 s of this injection. All animals failing this test and those with neurological de®cit after surgery were rejected from further study. The lidocaine test was also performed after each experiment to con®rm that all drugs injected down the intrathecal catheters were injected into the lumbar intrathecal space. The rats were rested for 24 h before further experiments were performed. Only one experiment per day up to a maximum of ®ve was performed on each animal.
Nociceptive tests Two nociceptive tests were used in all experiments, the ECT test and the TFL test, as described previously.18 The electrical test allows demonstration of drug action at the level of the spinal cord after lumbar intrathecal injection, i.e. the measurement of an increase in nociceptive threshold in the tail with no change in the neck threshold measurements. Such a differential block proves that the effects of the drug are con®ned to the caudal segments of the spinal cord responsible for tail and hind limb innervation. A rise in TFL in the same experiment in which such a differential effect is demonstrated must also therefore be due to action of the drug on the caudal segments of the spinal cord. Both of these tests were applied in sequence to each rat every 5 min. Electric current threshold test
Each rat was placed in a darkened restrainer and electrical stimulating electrodes were applied to the skin of the tail and neck. Brief bursts of electric current (0.5 s train, 2 ms pulse width, 50 Hz, 0±10 mA) were delivered in turn to each set of electrodes (tail ®rst, 15 s before the neck) every 5 min. The minimum current necessary to evoke an obvious nociceptive response (a squeak or sharp withdrawal), de®ned by the `up±down' method, was measured at each
680
Spinal cord 5-HT1A receptors and antinociception
skin site every 5 min until three stable consecutive readings had been obtained. The intrathecal drug was then given and measurements continued at each skin site every 5 min for a further 45±50 min. In order to standardize results obtained at each skin site, the response to intrathecal drug was calculated as a ratio with respect to the mean of the three stable readings taken before intrathecal drug injection, as described previously.18 These values for responses to each dose of drug were combined and expressed as the mean (SEM) on agonist dose±response curves. Tail-¯ick latency test
This was performed every 5 min immediately before the electrical measurement in the tail. The heat from a 150 watt projector bulb was directed onto the blackened tip of the tail and the power adjusted so that the rat ¯icked its tail away from the noxious stimulus in approximately 3 s. At this intensity of stimulation, three stable control (pre-drug injection) readings were obtained and then the test drug was injected. TFL measurement continued every 5 min thereafter for 30±45 min. The response to each dose of drug was calculated as a percentage of maximum possible effect (MPE): %MPE
TFL post-drug ÿ TFL pre-drug 100: cut-off time ÿ TFL pre-drug
Antagonist dose±response relationships Doses of agonist drugs were chosen from their dose±response curves for the experiments with antagonists. These were the roughly equieffective doses that produced a nearmaximal response for tail ECT. This dose was 0.1 mmol for 8-hydroxy-DPAT hydrobromide and 0.5 mmol for 1-(3chlorophenyl)-piperazine dihydrochloride. The agonist was given intrathecally alone at the beginning and at the end of a series of once-daily experiments in each rat. In the intervening days the same dose of the same agonist was given intrathecally combined with a range of doses (0.001±0.05 mmol) of a 5-HT1A-selective antagonist [4-[3(benzotriazol-1-yl)propyl]-1-(2-methoxyphenyl)-piperazine] dissolved in 6% dextrose. Twenty experiments were performed in nine rats using 8-hydroxy-DPAT hydrobromide (n=4±5 observations per dose of antagonist). Eleven rats were used in 27 experiments with 1-(3-chlorophenyl)piperazine dihydrochloride (n=4±5 observations per dose of antagonist). Antinociceptive effects were measured using ECT in the tail and neck and TFL. The pre-series and postseries measurements of the antinociceptive effect for agonist alone in each rat were then averaged (R). The percentage suppression of agonist effect by each dose of antagonist was calculated using the following equation, described previously:18 % suppression
A cut-off time of 10 s was used. Values were combined for each drug dose, expressed as mean (SEM) and used to produce dose±response curves.
Agonist dose±response relationships A range of doses (0.005±5.0 mmol) of 1-(3-chlorophenyl)piperazine dihydrochloride (non-selective agonist of 5HT1)19 20 was given intrathecally to six rats with chronically implanted subarachnoid catheters. Sixteen experiments were performed (n=4 per dose of agonist). Similarly, a range of doses (0.01±0.5 mmol) of 8-hydroxy-2-dipropylaminotetralin hydrobromide (a 5-HT1A-selective agonist)21 22 was given intrathecally to 15 rats with chronically implanted subarachnoid catheters. Twenty-three experiments were performed (n=4±6 per dose of agonist). The drugs were dissolved in 6% dextrose solution to make the injectate hyperbaric compared with CSF. Doses were delivered into the lumbosacral CSF in 5 ml volumes while the rat was held at a 15° inclined plane, head up. This was done to restrict the spread of drug to the caudal segments of the spinal cord, as described previously.23 The dextrose solution has been shown not to affect nociceptive thresholds when given alone.18 Once-daily injections of agonist drugs were given on up to ®ve occasions to individual rats. This has been shown in previous work on intrathecal 5-HT not to cause tolerance or tachyphylaxis.15 16
Rÿr 10 Rÿ1
where r is the individual response to agonist mixed with a particular dose of antagonist and R the mean response to agonist given intrathecally alone. The values of percentage suppression were calculated for both ECT and TFL tests. Values were combined to produce means and SEM, which were then plotted as dose±response relationships for suppression of the antinociceptive effects of each agonist.
Results Positive lidocaine tests were obtained in each rat and after each experiment that contributed values to the dose±response relationships reported in this paper. Thus, all agonist and antagonist drugs injected into intrathecal catheters were delivered to the lumbosacral intrathecal space. Both agonist drugs caused dose-related increases in tail ECT with no change in neck ECT. Thus, both drugs caused antinociception by an action in the caudal segments of the spinal cord responsible for tail innervation (Figs 1 and 2). The non-selective 5-HT1 agonist 1-(3-chlorophenyl)piperazine dihydrochloride caused simultaneous doserelated antinociceptive effects, as assessed with the TFL test (Fig. 1). By contrast, the selective 5-HT1A agonist 8hydroxy-DPAT hydrobromide caused no such TFL effects, the antinociception being con®ned to ECT (Fig. 2).
681
Nadeson and Goodchild
Fig 1 Dose±response curves for the antinociceptive effects of the nonselective 5-HT1 agonist 1-(3-chlorophenyl)-piperazine dihydrochloride in the ECT and TFL tests. Intrathecal 1-(3-chlorophenyl)-piperazine dihydrochloride produced dose-related antinociception in both tests. ECT (neck) values did not increase after intrathecal injections of 1-(3chlorophenyl)-piperazine dihydrochloride, indicating that the drug was con®ned in its actions to the caudal segments of the spinal cord responsible for tail innervation. Data points are means and bars are SEM (n=4).
Figure 3 shows dose±response curves for the suppression of tail ECT and TFL antinociceptive effects of 1-(3chlorophenyl)-piperazine dihydrochloride (Fig. 3A) and 8hydroxy-DPAT hydrobromide (Fig. 3B) by the antagonist 4[3-(benzotriazol-1-yl)propyl]-1-(2-methoxyphenyl)-piperazine. This selective 5-HT1A antagonist caused dose-related antagonism of the antinociception assessed by the ECT test for both agonist drugs. However, the TFL effects of the nonselective 5-HT1 receptor agonist 1-(3-chlorophenyl)-piperazine dihydrochloride were unaffected by intrathecal injection of the selective 5-HT1A antagonist (Fig. 3A).
Discussion Many studies have reported antinociception after intrathecal administration of serotonergic agonists.1±3 11 12 24±26 Many of these involved injections of drugs via catheters reported to be intrathecal but no proof of catheter position and site of drug action at the level of the spinal cord was given. Although conclusions about spinal cord 5-HT receptors were drawn, the antinociceptive effects observed could have been due to actions of the drugs on the brain after spread via CSF. The study reported in this paper concentrated on the restriction of drug actions to the spinal cord and the provision of support for such con®nement. The experiments were performed in rats with positive lidocaine tests, and antinociceptive effects occurred only in the tail, not the
Fig 2 Dose±response curves for the antinociceptive effects of the selective 5-HT1A agonist 8-hydroxy-DPAT in the ECT and TFL tests. Intrathecal 8-hydroxy-DPAT caused dose-related antinociception in the ECT test only. TFL was not affected by intrathecal injections of 8hydroxy-DPAT. ECT (neck) values did not increase after intrathecal injections of 8-hydroxy-DPAT, indicating that the drug was con®ned in its actions to the caudal segments of the spinal cord responsible for tail innervation. Data points are means and bars are SEM (n=4±6).
neck. It has been shown previously in this model that intrathecally administered 5-HT caused its ECT antinociceptive effects by involvement of spinal cord GABA and m opioid receptors but no such interactions occurred for the TFL effects.16 23 The results of the experiments reported in this paper extend these ®ndings. Again, there was a difference in the mechanisms of antinociception revealed by the TFL and ECT tests. The ECT test reveals antinociception involving 5-HT1A receptors, whereas the TFL test involves other 5-HT1 (but not 5-HT1A) receptors. One study in mice used the same 5-HT1A-selective agonist as that used in the present study (8-hydroxyDPAT).11 The investigators concluded that 5-HT1A receptors in the spinal cord were responsible for TFL antinociception. This result is contrary to the ®ndings of the experiments reported in this paper. This may represent a supraspinal effect of the drug. Many of these drugs are very soluble in water and thus spread easily throughout the CSF. The investigators made no measurements, such as a comparison of rostral and caudal dermatome nociceptive thresholds, that would indicate con®nement of drug action to the caudal segments of the spinal cord. The importance of distinguishing spinal and supraspinal effects was highlighted recently in a study of the involvement of 5-HT1A receptors in opioid-induced bulbospinal inhibition.13 It was shown that fentanyl injected into CSF in the brain inhibited sural±gastrocnemius re¯ex responses. Intrathecal administration of the selective 5-HT1A receptor antagonist WAY-100635 signi®cantly reduced this inhibition. In a separate group of experiments, i.v. fentanyl
682
Spinal cord 5-HT1A receptors and antinociception
Fig 3 Dose±response curves for reversal of the antinociceptive effects of (A) 1-(3-chlorophenyl)-piperazine dihydrochloride (non-selective 5-HT1 agonist) and (B) 8-hydroxy-DPAT (selective 5-HT1A agonist) by the selective 5-HT1A antagonist 4-[3-(benzotriazol-1-yl)propyl]-1-(2-methoxyphenyl)piperazine in the ECT and TFL tests. The antagonist suppressed the antinociceptive effects of both agonists in the ECT test in a dose-related manner. [Antinociception revealed by TFL after intrathecal injections of the non-selective 5-HT1 agonist 1-(3-chlorophenyl)-piperazine dihydrochloride was not affected by intrathecal injections of the antagonist 4-[3-(benzotriazol-1-yl)propyl]-1-(2-methoxyphenyl)-piperazine.] ECT (neck) values did not increase after intrathecal injections of either agonist in these experiments, indicating that the agonist drugs were con®ned in their actions to the caudal segments of the spinal cord responsible for tail innervation. Points shown are means and bars are SEM (n=4±5).
depressed the sural±gastrocnemius re¯ex. This inhibition was not affected by intrathecal WAY-100635 (100 mg), but combined administration of the 5-HT1A antagonist with an a2 adrenoceptor antagonist (RX 821002) signi®cantly reduced the effects of i.v. fentanyl. They concluded that the bulbospinal and direct spinal actions of fentanyl may occur together to produce overall inhibition of the re¯ex.13 Other studies have claimed interaction of serotonergic receptors with other neurotransmitter systems in the spinal cord. It was reported that a 5-HT3 receptor selective agonist (2-methyl serotonin) caused antinociceptive effects that involved spinal opioid and GABA receptor systems.27 When this drug was injected intracerebroventricularly no antinociception was observed. It was therefore concluded that the antinociception after intrathecal injection was a spinal cord effect. However, 2-methyl serotonin does have af®nity for 5-HT1 receptors.28 Selectivity of drugs for 5-HT receptor subtypes is often not high. Frequently, pKi values for a drug binding to different receptor subtypes differ by one log unit or less.28 29 Thus, Giordano27 may have been describing the system described in this paperÐthe activation of 5-HT1A receptors in the spinal cord. However, more con®dence may be placed in the ®ndings of the present paper because the 5-HT1A-selective drugs that were used have pKi values for the 5-HT1A receptor subtype that are two or three log units greater than values for other receptors.28 29 Clearly, the involvement of 5-HT1 receptors other than 5-
HT1A receptors and of 5-HT2 and 5-HT3 receptors needs to be investigated because of the many reports of antinociceptive effects after intrathecal drugs reported to be selective for these 5-HT receptor subtypes. Furthermore, more detailed dose±response studies must await the arrival of more selective agonist and antagonist drugs that are also water-soluble, so that it will be possible to inject these drugs intrathecally and to study actions con®ned to the spinal cord. These issues are clearly important if there is to be development of a drug designed to target the spinal cord 5-HT receptors involved in antinociception. For example, an intrathecal drug selective for a 5-HT receptor subtype that is only involved in antinociception in the brain will work only if a high dose of the drug is given intrathecally and the drug subsequently spreads to the brain. Such an approach to therapy will inevitably lead to side-effects. We conclude that 5-HT1A receptors in the spinal cord are involved in nociceptive mechanisms that can be assessed by measuring responses to noxious electrical stimuli. Other 5HT1 receptors (non-5-HT1A receptors) are involved in spinally mediated antinociception that can be assessed by measuring responses to thermal noxious stimuli.
References
683
1 Headley PM, Duggan AW, Griersmith BT. Selective reduction by noradrenaline and 5-hydroxytryptamine of nociceptive
Nadeson and Goodchild
2 3 4 5
6 7 8
9 10 11 12 13 14 15
responses in cat dorsal horn neurones. Brain Res 1978; 145: 185±9 Wang JK. Antinociceptive effect of intrathecally administered serotonin. Anesthesiology 1977; 47: 269±71 Yaksh TL, Wilson PR. Spinal serotonin terminal systems mediate antinociception. J Pharmacol Exp Ther 1979; 208: 446±53 Watkins LR, Cobelli DA, Faris P, Aceto MD, Mayer DJ. Opiate vs non-opiate footshock-induced analgesia (FSIA): the body region shocked is a critical factor. Brain Res 1982; 242: 299±308 Rivot JP, Chaouch A, Besson JM. Nucleus raphe magnus modulation of response of rat dorsal horn neurons to unmyelinated ®ber inputs: partial involvement of serotonergic pathways. J Neurophysiol 1980; 44: 1039±57 Glaum SR, Proud®t HK, Anderson EG. 5-HT3 receptors modulate spinal nociceptive re¯exes. Brain Res 1990; 510: 12±16 Holohean AM, Hackman JC, Davidoff RA. Modulation of frog spinal cord interneuronal activity by activation of 5-HT3 receptors. Brain Res 1995; 704: 184±90 Mjellem N, Lund A, Eide PK, Storkson R, Tjolsen A. The role of 5-HT1A and 5-HT1B receptors in spinal nociceptive transmission and in the modulation of NMDA induced behaviour. Neuroreport 1992; 3: 1061±4 Proud®t HK, Hammond DL. Alterations in nociceptive threshold and morphine-induced analgesia produced by intrathecally administered amine antagonists. Brain Res 1981; 218: 393±9 Zemlan FP, Kow LM, Pfaff DW. Spinal serotonin (5-HT) receptor subtypes and nociception. J Pharmacol Exp Ther 1983; 226: 477±85 Eide PK, Hole K. Different role of 5-HT1A and 5-HT2 receptors in spinal cord in the control of nociceptive responsiveness. Neuropharmacology 1991; 30: 727±31 Gjerstad J, Tjolsen A, Hole K. The effect of 5-HT1A receptor stimulation on nociceptive dorsal horn neurones in rats. Eur J Pharmacol 1996; 318: 315±21 Clarke RW, Ward RE. The role of 5-HT1A-receptors in fentanylinduced bulbospinal inhibition of a spinal withdrawal re¯ex in the rabbit. Pain 2000; 85: 239±45 Guo Z, Gent JP, Goodchild CS. Intrathecal 5-hydroxytryptamine produces analgesia by activating spinal cord opioid pathways. Pain 1990; 43 Suppl 5: S103 Goodchild CS, Guo Z, Freeman J, Gent JP. 5-HT spinal antinociception involves mu opioid receptors: cross tolerance and antagonist studies. Br J Anaesth 1997; 78: 563±9
684
16 Nadeson R, Guo Z, Porter V, Gent JP, Goodchild CS. GABAA receptors and spinally-mediated antinociception in rats. J Pharmacol Exp Ther 1996; 278: 620±6 17 Zimmerman M. Ethical guidelines for investigation of experimental pain on conscious animals. Pain 1983; 16: 109±10 18 Edwards M, Serrao JM, Gent JP, Goodchild CS. On the mechanism by which midazolam causes spinally mediated analgesia. Anesthesiology 1990; 73: 273±7 19 Glennon RA. Central serotonin receptors as targets for drug research. J Med Chem 1987; 30: 1±12 20 Lucki I, Ward HR, Frazer A. Effect of 1-(m-chlorophenyl)piperazine and 1-(m-tri¯uoromethylphenyl)piperazine on locomotor activity. J Pharmacol Exp Ther 1989; 249: 155±64 21 Corn®eld LJ, Lambert G, Arvidsson LE, et al. Intrinsic activity of enantiomers of 8-hydroxy-2-(di-n-propylamino)tetralin and its analogs at 5-hydroxytryptamine1A receptors that are negatively coupled to adenylate cyclase. Mol Pharmacol 1991; 39: 780±7 22 Bjork L, Hook BB, Nelson DL, Anden NE, Hacksell U. Resolved N,N-dialkylated 2-amino-8-hydroxytetralins: stereoselective interactions with 5-HT1A receptors in the brain. J Med Chem 1989; 32: 779±83 23 Goodchild CS, Guo Z, Freeman J, Gent JP. 5-HT spinal antinociception involves mu opioid receptors: cross tolerance and antagonist studies. Br J Anaesth 1997; 78: 563±9 24 Bardin L, Bardin M, Lavarenne J, Eschalier A. Effect of intrathecal serotonin on nociception in rats: in¯uence of the pain test used. Exp Brain Res 1997; 113: 81±7 25 FuÈrst S. Transmitters involved in antinociception in the spinal cord. Brain Res Bull 1999; 48: 129±41 26 Zhang Z-H, Yang S-W, Chen J-Y, Xie Y-F, Qiao J-T, Dafny N. Interaction of serotonin and norepinephrine in spinal antinociception. Brain Res Bull 1995; 38: 167±71 27 Giordano J. Analgesic pro®le of centrally administered 2methylserotonin against acute pain in rats. Eur J Pharmacol 1991; 199: 233±6 28 Hoyer D. 5-Hydroxytryptamine receptors and effector coupling mechanisms in peripheral tissues. In: Fozard J, ed. Peripheral Actions of 5-HT. Oxford: Oxford University Press, 1989; 72±99 29 Baxter G, Kennett G, Blaney F, Blackburn T. 5-HT2 receptor subtypes: a family re-united? Trends Pharmacol Sci 1995; 16: 105±10