Exp Brain Res (1992) 92:310-317
Experimental BrainResearch 9
Springer-Verlag 1992
An analysis of adrenergic influences on the sural-gastrocnemius reflex of the decerebrated rabbit J. Harris and R . W . Clarke
Department of Physiology and Environmental Science, University of Nottingham, Sutton Bonington Campus, Loughborough, Leicestershire LE12 5RD, UK Received March 26, 1992 / Accepted July 6, 1992
S u m m a r y . The sural-gastrocnemius reflex was observed
in decerebrated rabbits during intrathecal application of four ~-adrenoceptor antagonists. Idazoxan and yohimbine, which are antagonists at the a2-receptor, caused facilitation of the reflex, although idazoxan was more potent and produced a larger overall increase in the reflex response. However, when given after yohimbine, idazoxan elicited no further increase in reflex responses. The differences between the two drugs may result from the interaction of yohimbine with receptors for 5-hydroxytryptamine. The selective ~l-receptor antagonist prazosin had no consistent effects when given alone, but reduced the facilitatory effects of idazoxan. The putative selective post-junctional a2-receptor blocker SK&F 104078 had no significant effects when given alone, nor did it influence the facilitatory action of a subsequent dose of idazoxan. Section of the spinal cord in the presence of idazoxan always caused a decrease in gastrocnemius responses to sural nerve stimulation. These data show that the facilitatory effects of idazoxan are almost certainly mediated at the spinal cord and that they do not involve blockade of ~l-receptors. It appears that idazoxan acts by blockade of adrenergic descending inhibition in combination with increased descending facilitation. The inhibition is probably mediated through noradrenaline acting at ~2-receptors, and the facilitation may be the result of release of noradrenaline (acting at ~l-receptors) and 5-hydroxytryptamine in the spinal cord. Key words: Spinal reflex Noradrenaline - I d a z o x a n D e s c e n d i n g inhibition and facilitation Rabbit
Introduction
In the decerebrated rabbit, electrical stimulation of the cutaneous sural nerve elicits a short-latency polysynaptic Correspondence to." J. Harris
reflex in the ipsilateral ankle extensor, gastrocnemius medialis (GM). This reflex is subject to tonic inhibitory influences arising from the brain stem and from within the spinal cord. Previous work from this laboratory has provided some pointers to the identity of the neurotransmitters involved in controlling this reflex pathway. In spinal rabbits, the sural-GM reflex is powerfully enhanced by the opioid antagonist naloxone but is only weakly increased by the ~2-receptor antagonist idazoxan, whereas in animals with an intact spinal cord, idazoxan enhances the reflex to a much greater extent than does naloxone (Clarke et al. 1988a). These findings indicate that the sural-GM reflex is suppressed by opioidergic transmitters released within the spinal cord, and by a pathway of supraspinal origin in which noradrenaline has an important role. In the studies outlined above, drugs were given intravenously. It was not therefore possible to be certain of the sites of action of agents producing effects in nonspinal preparations. Furthermore, the effective doses of idazoxan were rather large (1-2 mg/kg), which raised the possibility that the actions of this drug may not have been mediated entirely through blockade of a2adrenoceptors. Another problem is that idazoxan is known to bind with high affinity to non-adrenergic imidazoline binding sites (NAIBS), which are found in many tissues including rabbit brain (Hamilton et al. 1988). The present experiments were designed to answer the questions of receptor selectivity and sites of action of idazoxan in facilitating the suraLGM reflex. In this study drugs have been delivered directly to the lumbosacral spinal cord via a subdural (intrathecal, i.t.) cannula to attempt to limit their range of action to the spinal cord. Four drugs have been used in this way to assess the receptor selectivity of idazoxan: idazoxan itself; the selective al-adrenoceptor antagonist prazosin; the nonimidazoline a2-adrenoceptor antagonist yohimbine; and the putative post-junctional c~z-receptor antagonist SK&F 104078 (Ruffolo et al. 1987). Some of these data have appeared as an abstract (Harris et al. 1990).
311 Material and methods
2.5-100 pl to give cumulative doses of 13, 26, 128, 256 and 512 nmol. After each intrathecal injection, the cannula was flushed with 60 ~tl of Ringer's solution. In some experiments, 60 gl of vehicle and a 60-gl flush of Ringer's solution was injected before any drug was given, to test for any effects of the vehicle on the sural-GM reflex. After intrathecal administration of drugs and in the 21 animals which received no drug treatment, the spinal cord was transected by two adjacent dorsoventral hemisections separated by an interval of 24 min. On post-mortem examination, after exposure of the lumbosacral spinal cord, 60 gl of pontamine sky blue followed by 60 gl of Ringer's solution were injected through the i.t. cannula to assess the spread of drugs around the spinal cord. This indicated that most of the spinal cord caudal to L2 was bathed in drug solutions.
Surgical procedures Experiments were performed on 63 New Zealand Red (NZR) and NZR x New Zealand White rabbits, of either sex, weighing between 1.94 and 3.46 kg. Anaesthesia was induced by an intravenous injection of 10-20 mg/kg methohexitone sodium (Brietal; Eli Lilly) which was supplemented to effect. The trachea was exposed and cannulated so that anaesthesia could be continued with a mixture of halothane (ICI or May and Baker) in nitrous oxide and oxygen (70:30%). One carotid artery was cannulated to record arterial blood pressure and the other carotid artery was ligated. For the purpose of drug administration, two cannulae were inserted into one jugular vein. The electrocardiogram (ECG) was recorded via an intraoesophageal probe. A laminectomy was performed between spinal segments T 12 and L2 and a cannula inserted beneath the dura so that the tip lay at or near the L7 and S1 spinal segments. The position of this cannula was verified on post-mortem examination. All rabbits were decerebrated by suction to pre- (n = 56) or midcollicular (n = 7) levels. The core temperature of the rabbit was recorded with a rectal probe and maintained between 37.5 and 38.5~ C using a thermostatically controlled heating blanket. The left hindlimb was clamped rigidly at the knee and ankle, and the sciatic nerve exposed by section and retraction of the overlying muscle groups. The resultant pool was filled with warmed paraffin oil (37~ C). The sural nerve was dissected free, cut, and placed over paired platinum stimulating electrodes. A further portion of the sural nerve was freed proximal to the stimulating electrodes to allow recording of the afferent volley. The GM nerve was dissected free, cut, desheathed and the central end placed over a pair of platinum recording electrodes. When surgery was completed, anaesthesia was discontinued and the animal was paralysed with gallamine triethiodide (Flaxedil; May and Baker; 4 mg/kg i.v. initially). Ventilation was continued using a Starling ideal pump with the stroke volume adjusted to give an end-tidal CO2 of 3-5%. A period of at least 1 h was left between cessation of halothane anaesthesia and the beginning of recording.
Nerve stimulation and recording Reflexes were evoked by electrical stimulation of the sural nerve with square-wave pulses of 0.1 ms duration, applied at a frequency of 1 Hz, and of a strength sufficient to recruit all myelinated axons (i.e. 20-80 times threshold). This was confirmed by observation of the afferent volley recorded from the nerve. Reflexes were recorded from the GM muscle nerve which was crushed between the electrodes to give a monophasic signal. Stimuli were applied in blocks of 12 at 2-min intervals and, to allow for wind-up in the reflex pathway (see Catley et al. 1983), the responses to stimuli 4-11 in each block were averaged and integrated by computer.
Statistical analysis In each experiment reflex responses were normalized to the mean of the pre-drug control level. Pooled data from many experiments are expressed as median percentages, and scatter of results is indicated by the range. Statistical analysis of the results was performed using Wilcoxon's paired ranks test; Mann-Whitney U-test; Kruskal-Wallis ANOVA; and Fisher's exact probability test, assuming a significance level of 0.05 throughout.
Results The responses o f G M to sural nerve s t i m u l a t i o n were extremely v a r i a b l e b e t w e e n a n i m a l s , as r e p o r t e d previously (Clarke et al. 1988a). F o r the a n i m a l s which were to receive i d a z o x a n , the m e d i a n p r e - d r u g reflexes were 76 g V - m s (range 11 1072 g V . m s ) a n d the corres p o n d i n g values for the other d r u g t r e a t m e n t g r o u p s were: y o h i m b i n e , 132 g V . m s (range 20 1883 ~tV-ms); prazosin, 91 g V - m s (range 13-578 p V . m s ) ; a n d S K & F 104078, 41 g V . m s (range 2 6 - 1 1 9 2 g V - m s ) respectively. There were n o significant differences b e t w e e n a n y o f these g r o u p s ( K r u s k a l - W a l l i s test, P > 0 . 2 ) . 0 IDAZOXAN
BOO
PRAZOSIN [] SK&FIO407B 9 YOHII~BINE
o/ a
I
500
[1_ z
400 5OO
Drug administration J ta_ Ld
The following drugs were used: idazoxan hydrochloride (a gift of Dr. S.L. Dickinson, Reckitt and Colman) dissolved in Ringer's solution (8.3 mmol/1) and administered intrathecally in volumes of 2.5-100 gl to give cumulative doses of 21, 42, 208, 415 and 813 nmol; prazosin hydrochloride (a gift of Dr. R. Packard, Pfizer) dissolved in 0.4% dimethyl sulphoxide (DMSO) in 5% D-glucose solution (4.8 mmol/1) and administered intrathecally in volumes of 2.5 100 gl to give cumulative doses of 12, 24, 119, 238 and 476 nmol; SK&F 104078 maleate (a gift of SmithKline Beecham) dissolved in 2% DMSO in Ringer's solution (5.1 mmol/l) and administered intrathecally in volumes of 2.5 100 gl to give cumulative doses of 13, 25, 126, 253 and 505 nmol; and yohimbine hydrochloride (Sigma) dissolved in 1% dimethyl formamide (DMF) in 5 % D-glucose solution (5.1 mmol/1) and administered intrathecally in volumes of
700
~:
200
', 0.5). Figure 6 shows the effects of spinalization in individual idazoxan-treated rabbits and those which were spinalized without prior drug treatment. It also includes, for comparison, data from other experiments in which spinalization was carried out in the presence of saturating intravenous doses of naloxone alone, and naloxone with idazoxan (from
DRUG
I
I
I
NLX
IDZ
IDZ
ONLY
+NLX
ONLY
Fig. 6. The effect of spinalization on the sural-gastrocnemius reflex
in individual rabbits in the presence of different drug milieux, in this case with the response normalized with respect to the post-drug (i.e. immediate pre-spinalization) level. The figures show, from right to left, the effects of spinal section in untreated animals (NO DRUG, n= 21); in the presence of naloxone 388.5 gg/kg i.v. (NLX ONLY, n= 12); in the presence of naloxone 250 or 388.5 Mg/kg i.v. with idazoxan 2 or 3 mg/kg i.v., (IDZ+ NLX, n = 14); and in the presence ofidazoxan 1517 nmol i.t. (IDZ ONLY, n = 11). Only the data from this last column were obtained in the present study: the figures for untreated animals are from a separate set of experiments carried out contemporaneously with these, but yet to be published, and the data with naloxone are from Clarke et al. (1988a). Each point is the pooled mean of responses recorded over a 24-min period after spinalization Clarke et al. 1988a). When the spinal cord was sectioned in the absence of drug treatment, the s u r a l - G M reflex usually increased in size, although it decreased in 7 of 21 animals. The outcome of spinalization in the absence of drug treatment was significantly different from spinal section in the presence of idazoxan (Fisher's exact probability = 0.0002); naloxone (P = 0.027); or idazoxan with naloxone (P=0.04). There was also a significant difference between spinalization in the presence o f naloxone compared with naloxone plus idazoxan ( P = 0.0003).
Discussion
The effects of intrathecal admin&tration of a-receptor antagonists Intrathecal administration of nanomolar quantities of idazoxan caused a sixfold increase in G M responses to sural nerve stimulation. This effect was almost certainly due to a direct action of the drug at the spinal cord, as intravenous administration of similar doses results in much smaller increases in the s u r a l - G M reflex (Clarke et al. 1988a). Idazoxan is a highly selective %-adrenoceptor antagonist with some 40-fold higher potency at a2 than at al receptors (Doxey et al. 1983; Megens et al. 1986) and no appreciable affinity at receptors for other known
315 transmitters (Pettibone et al. 1986; although, see below). The fact that prazosin, a highly selective al-receptor antagonist (see Wilson et al. 1991), had no significant effects on the sural-GM reflex when given alone, is good evidence that the facilitatory action of idazoxan is not due to the result of blockade of 0h-receptors. On the basis of these few facts, one might conclude that idazoxan causes facilitation of the reflex by blockade of inhibitory cz2-receptors in the spinal cord. Furthermore, the fact that the cz2-agonist clonidine suppresses reflexes in spinal rabbits (Clarke et al. 1988a), indicates that inhibitory adrenoceptors are post-synaptic to adrenergic terminals. The receptors involved do not appear to be the same as the post-junctional a2 sites found in the periphery, as SK&F 104078, a putative selective antagonist at these receptors (Ruffolo et al. 1987), had no effect on the sural-GM reflex. However, this drug also binds with moderate affinity to 5-HT~A and 5-HT2 receptors (Kilpatrick et al. 1989), which complicates interpretation of the effects of this material. It is not yet possible to say with certainty that all of the effects of idazoxan are mediated entirely through cz2-receptors, as the present data have not eliminated the possibility that some of the actions of this drug might be mediated by actions at non-adrenergic imidazoline binding sites (see Introduction). The functions of these sites have yet to be defined. Yohimbine, long promoted as a selective cz2-receptor antagonist (see Goldberg and Robertson 1983), does not bind to NAIBS. In the present study this drug did facilitate the sural-GM reflex, although it was some 15 times less potent than idazoxan and had a much lower peak effect. The fact that both idazoxan and yohimbine facilitated GM reflex responses indicates that blockade of cz2-receptors is one aspect of the facilitatory action of idazoxan, but the disparities between the potencies and ceiling effects of idazoxan and yohimbine cannot be explained only by actions at a2-receptor sites (see Doxey et al. 1983; Megens et al. 1986; Pettibone et al. 1986). There are two possible explanations for the differences between the two drugs: either the effects of idazoxan are due in part to an action at NAIBS or yohimbine has actions at other receptors which restrict the expression of the cz2-blocking action of this drug. The fact that idazoxan had no effect after yohimbine suggests that the latter explanation is correct. Although yohimbine has often been used as a selective cz2-adrenoceptor antagonist, it has a range of actions at receptors for 5-HT. Yohimbine or its enantiomers have been shown to bind to 5-HT1A (Convents et al. 1989) and 5-HT2 (see Wilson et al. 1991) receptors and, most curiously, yohimbine is an agonist at the 5-HT terminal autoreceptor (presumed to be of the 5-HT~D type) in the rabbit (Feuerstein et al. 1985; see also Hoyer et al. 1990). Before further considering the significance of these facts, it is necessary to examine the complexities of the idazoxaninduced facilitation of the sural-GM reflex.
Idazoxan and descendin9 facilitation In an earlier paper we hypothesized that the major effect of idazoxan on the sural-GM reflex was due to the
blockade of adrenergic descending inhibition (Clarke et al. 1988a). Results obtained in the present experiments show this conclusion to be far too simplistic. When the spinal cord was sectioned in the presence of idazoxan alone, the reflex always decreased in size. This shows that the drug not only blocks inhibition in the spinal cord, it also reveals or potentiates descending facilitation of the sural-GM pathway. The interpretation of this finding is complicated by the fact that opioid peptides are tonically active in the spinal cord. In spinal rabbits, the sural-GM reflex is powerfully enhanced by the opioid antagonist naloxone (Clarke and Ford 1987), but this effect is much weaker in non-spinal animals (Clarke et al. 1988a). It is possible that the level of opioidergic tone is lower in non-spinal animals and that spinal section increases activity in spinal opioidergic neurones. If this were the case, one might expect cord section in untreated rabbits to have resulted only in decreases in the size of reflexes, which was clearly not the case. Furthermore, although spinal section in the presence of naloxone alone always results in an increase in the sural-GM reflex, if the cord is cut in the presence of naloxone and idazoxan together, the reflex usually decreases in size (Clarke et al. 1988a). Taken together, these data show that idazoxan really does facilitate reflexes by blockade of descending inhibition and revealing, or augmenting, descending facilitation. The present experiments provide some indicators to the identity of the transmitters involved in descending facilitation. High doses of prazosin alone resulted in depression of the sural-GM reflex in some rabbits, and the same drug reduced the facilitatory effects of idazoxan whether given before or after the ~2-antagonist. Thus, in addition to having an inhibitory influence over transmission through the sural-GM pathway by an action at a2-receptors (probably in the dorsal horn; see Fleetwood-Walker et al. 1985), noradrenaline may have a facilitatory role, acting through al-receptors. This is not a new idea: prazosin has been shown to reduce excitatory influences from the brain on reflex pathways in rat (Palmeri and Wiesendanger 1990; Tanabe et al. 1990) and cat (Lai et al. 1989) and noradrenaline-induced depolarization of motoneurones in neonatal rat spinal cord is known to be mediated by an o~l-receptor (Connell et al. 1989). The general import of this body of work is that noradrenaline facilitates motor function via an action at al-receptors probably located in the ventral horn (see also White and Neuman 1980; Bell and Matsumiya 1981). It is likely that 5-HT also is involved in descending facilitation of reflex function as it has similar actions to noradrenaline on ventral horn cells (White and Neuman 1980; Bell and Matsumiya 1981; Connell and Wallis 1988; Jackson and White 1990). As it is known that noradrenaline can modulate the release of 5-HT from slices of rabbit hippocampus (Feuerstein et al. 1985) and caudate nucleus (Feuerstein et al. 1992), one might speculate that idazoxan promotes descending facilitation of reflexes by increasing the release of noradrenaline and 5-HT in the ventral horn.
316
The modes of action of idazoxan and yohimbine To summarize, we propose, on the basis of the current evidence, that idazoxan facilitates the s u r a l - G M reflex by acting in three ways: (1) blockade of post-synaptic Ctz-receptors in the dorsal horn, resulting in the attenuation of adrenergic descending inhibition; (2) blockade of pre-synaptic a2-autoreceptors on the terminals of adrenergic axons in the ventral horn, resulting in increased adrenergic descending facilitation (mediated through al-receptors); and, most tentatively, (3) blockade of c~/-receptors on the terminals of 5-HT-ergic axons in the ventral horn, resulting in increased 5-HT-ergic descending facilitation. Yohimbine, being an ~2-antagonist and an agonist at the 5 - H T - a u t o r e c e p t o r , would have effect 1 and p r o b a b l y also 2, but would, if anything, reduce 5 - H T release and not have effect 3. This would explain why yohimbine produced less facilitation than idazoxan and why idazoxan produced no further increases in the reflex when given after yohimbine, although one has to bear in mind the possibility that these effects m a y be due to interaction o f this drug with other types of receptor (Convents et al. 1989). In support of our view we have recently shown that cyanopindolol, another agonist at the rabbit 5 - H T terminal autoreceptor (see H o y e r et al. 1990; Feuerstein et al. 1992), also blocks the facilitatory effects of idazoxan (Harris et al. 1992). The. apparently self-limiting actions of yohimbine m e a n that the possibility that imidazoline binding sites play a part in the actions of idazoxan cannot be eliminated. There are available drugs which are highly selective for the a2-receptor but have no affinity for N A I B S (e.g. see Langin et al. 1990); these will f o r m the basis of further investigations into this problem. The present results underline the complex nature of descending control of spinal function. It is clear that descending pathways which tonically modulate spinal reflexes in decerebrate animals can be facilitatory or excitatory and that the effects of spinal section or cooling are the net result of removing a range of influences. An i m p o r t a n t future step must be to identify definitively those sites in the spinal cord at which facilitation and inhibition are effected, and to define the amine receptor subtypes involved at each location (e.g. Bras et al. 1989, 1990). The key to this must be the use of highly selective receptor antagonists and agonists, with the fullest possible knowledge of the range of actions and limitations of each drug used.
Acknowledgements. This work was funded by the Agricultural and Food Research Council. John Harris is a Science and Engineering Research Council scholar. We are grateful to Caroline Northway and Irka Zajac-Galij for expert technical assistance, and to Reckitt and Colman, Pfizer and SmithKline Beecham for the supply of drugs.
References Bell JA, Matsumiya T (1981) Inhibitory effects of dorsal horn and excitant effects of ventral horn intraspinal microinjections of norepinephrine and serotonin in the cat. Life Sci 29:1507-1514
Bras H, Cavallari P, Jankowska E, McCrea D (1989) Comparison of effects of monoamines on transmission in spinal pathways from group I and II muscle afferents in the cat. Exp Brain Res 76: 27-37 Bras H, Jankowska E, Noga B, Skoog B (1990) Comparison of effects of various types of NA and 5-HT agonists on transmission from group-II muscle afferents in the cat. Eur J Neurosci2:1029-1039 Catley DM, Clarke RW, Pascoe JE (1983) Naloxone enhancement of spinal reflexes in the rabbit. J Physiol (Lond) 339: 63-72 Clarke RW, Ford TW (1987) The contributions of ~t-, 6- and ~-opioid receptors to the actions of endogenous opioids on spinal reflexes in the rabbit. Br J Pharmacol 91 : 579-589 Clarke RW, Ford TW, Taylor JS (1988a) Adrenergic and opioidergic modulation of a spinal reflex in the rabbit. J Physiol (Lond) 404: 407-417 Clarke RW, Ford TW, Harris J, Taylor JS (1988b) The effects of direct spinal application of idazoxan and prazosin on a spinal reflex in the decerebrated rabbit. Neurosci Lett Suppl 32:$32 Connell LA, Wallis DI (1988) Response to 5-hydroxytryptamine evoked in the hemisected spinal cord of the neonate rat. Br J Pharmacol 94: 1101-1114 Connell LA, Majid A, Wallis DI (1989) Involvement of a 1adrenoceptors in the depolarizing but not the hyperpolarizing responses of motorneurones in the neonate rat to noradrenaline. Neuropharmacology 28:1399-1404 Convents A, De Kayser J, De Backer JP, Vauquelin G (1989) [3H]Rauwolscine labels c%-adrenoceptors and 5t-ITIAreceptors in human cerebral cortex. Eur J Pharmacol 159:307-310 Doxey JC, Roach AG, Strachan DA, Virdee NK (1983) Effects of RX 781094 and yohimbine on the responses to UK 14304 at various ct2-adrenoceptors in rats. Br J Pharmacol 79:311P Feuerstein TJ, Hertting G, Jackisch R (1985) Endogenous noradrenaline as modulator of hippocampal serotonin (5-HT)release. Dual effects of yohimbine, rauwolscine and corynanthine as a-adrenoceptor antagonists and 5-HT-receptor agonists. Naunyn Schmiedebergs Arch Pharmacol 329:216-221 Feuerstein TJ, Lupp A, Hertting G (1992) Quantitative evaluation of autoinhibitory feedback of release of 5-HT in the caudate nucleus of the rabbit where an endogenous tone on a2 receptors does not exist. Neuropharmacology 31 : 15-23 Fleetwood-Walker SM, Mitchell R, Hope P J, Molony V, Iggo A (1985) An Ctz-receptormediates the selective inhibition by noradrenaline of nociceptive responses of identified dorsal horn neurones. Brain Res 334:243-254 Goldberg MR, Robertson D (1983) Yohimbine: a pharmacological probe for the study of the cz2-adrenoceptor. Pharmacol Rev 35:143-180 Hamilton CA, Reid JL, Yakubu MA (1988) [3H]-Yohimbine and [3H]-idazoxan bind to different sites in rabbit forebrain. Eur J Pharmacol 146:345-349 Harris J, Ford TW, Clarke RW (1990) The effects of intrathecal application of c~-adrenoceptor antagonists on the sural-gastrocnemius reflex in the decerebrated rabbit (abstract). J Physiol (Lond) 429 : 130P Harris J, White DP, Clarke RW (1992) The roles of 5-HT in modulating a spinal reflex in the decerebrated rabbit. Neurosei Lett Suppl 42: S 13 Hoyer D, Schoeffter P, Waeber C, Palacios JM (1990) Serotonin 5-HT1D receptors. Ann NY Acad Sci 600:168-181 Jackson DA, White SR (1990) Receptor subtypes mediating facilitation by serotonin of excitability of spinal motoneurons. Neuropharmacology 29 : 787-797 Kilpatrick AT, Brown CM, MacKinnon AC, Spedding M (1989) The az-adrenoceptor antagonist SK&F 104078 has high affinity for 5-HT1A and 5-I-IT2 receptors. Eur J Pharmacol 166:315-318 Lai Y-Y, Strahlendorf I-IK, Fung SJ, Barnes CD (1989) The actions of two monoamines on spinal motoneurons from stimulation of the locus coeruleus in the cat. Brain Res 484:268-272
317 Langin D, Paris H, Dauzats M, Lafontan M (1990) Discrimination between alpha-2 adrenoceptors and [3H]idazoxan-labelled nonadrenergic sites in rabbit white fat cells. Eur J Pharmacol 188:261-272 Megens AAHP, Leysen JE, Awouters FHL, Niemegeers CJE (1986) Further validation of in vivo and in vitro pharmacological procedures for assessing the ch/ct2 selectivity of test compounds : (1) ct-adrenoceptor antagonists. Eur J Pharmacol 129:49-55 Palmeri A, Wiesendanger M (1990) Concomitant depression of locus coeruleus neurons and of flexor reflexes by an a2-adrenergic agonist in rats: a possible mechanism for an ct2-mediated muscle relaxation. Neuroscience 34:177-187 Pettibone D J, Clineschmidt BV, Lotti VJ, Martin GE, Huff JR, Randall WC, Vacca J, Baldwin JJ (1986) L 654,284 a new potent and selective a2-adrenoceptor antagonist. Naunyn Schmiedebergs Arch Pharmacol 333:110 116
Ruffolo RR Jr, Sulpizio AC, Nichols A J, DeMarinis RM, Hieble JP (1987) Pharmacologic differentiation between pre- and postjunctional az-adrenoceptors by SK&F 104078. Naunyn Schmiedeberg Arch Pharmacol 336:415-418 Tanabe M, Ono H, Fukuda H (1990) Spinal alpha-1-adrenoceptors and alpha-2-adrenoceptors mediate facilitation and inhibition of spinal motor transmission, respectively. Jpn J Pharmacol 54:6%77 White SR, Neuman RS (1980) Facilitation of spinal motoneurone excitability by 5-hydroxytryptamine and noradrenaline. Brain Res 188:119-127 Wilson VG, Brown CH, McGrath JC (1991) Are there more than two types of ct-adrenoceptors involved in physiological responses ? Exp Physiol 76: 317-346
Experimental BrainResearch 9
Springer-Verlag 1992
An analysis of adrenergic influences on the sural-gastrocnemius reflex of the decerebrated rabbit J. Harris and R . W . Clarke
Department of Physiology and Environmental Science, University of Nottingham, Sutton Bonington Campus, Loughborough, Leicestershire LE12 5RD, UK Received March 26, 1992 / Accepted July 6, 1992
S u m m a r y . The sural-gastrocnemius reflex was observed
in decerebrated rabbits during intrathecal application of four ~-adrenoceptor antagonists. Idazoxan and yohimbine, which are antagonists at the a2-receptor, caused facilitation of the reflex, although idazoxan was more potent and produced a larger overall increase in the reflex response. However, when given after yohimbine, idazoxan elicited no further increase in reflex responses. The differences between the two drugs may result from the interaction of yohimbine with receptors for 5-hydroxytryptamine. The selective ~l-receptor antagonist prazosin had no consistent effects when given alone, but reduced the facilitatory effects of idazoxan. The putative selective post-junctional a2-receptor blocker SK&F 104078 had no significant effects when given alone, nor did it influence the facilitatory action of a subsequent dose of idazoxan. Section of the spinal cord in the presence of idazoxan always caused a decrease in gastrocnemius responses to sural nerve stimulation. These data show that the facilitatory effects of idazoxan are almost certainly mediated at the spinal cord and that they do not involve blockade of ~l-receptors. It appears that idazoxan acts by blockade of adrenergic descending inhibition in combination with increased descending facilitation. The inhibition is probably mediated through noradrenaline acting at ~2-receptors, and the facilitation may be the result of release of noradrenaline (acting at ~l-receptors) and 5-hydroxytryptamine in the spinal cord. Key words: Spinal reflex Noradrenaline - I d a z o x a n D e s c e n d i n g inhibition and facilitation Rabbit
Introduction
In the decerebrated rabbit, electrical stimulation of the cutaneous sural nerve elicits a short-latency polysynaptic Correspondence to." J. Harris
reflex in the ipsilateral ankle extensor, gastrocnemius medialis (GM). This reflex is subject to tonic inhibitory influences arising from the brain stem and from within the spinal cord. Previous work from this laboratory has provided some pointers to the identity of the neurotransmitters involved in controlling this reflex pathway. In spinal rabbits, the sural-GM reflex is powerfully enhanced by the opioid antagonist naloxone but is only weakly increased by the ~2-receptor antagonist idazoxan, whereas in animals with an intact spinal cord, idazoxan enhances the reflex to a much greater extent than does naloxone (Clarke et al. 1988a). These findings indicate that the sural-GM reflex is suppressed by opioidergic transmitters released within the spinal cord, and by a pathway of supraspinal origin in which noradrenaline has an important role. In the studies outlined above, drugs were given intravenously. It was not therefore possible to be certain of the sites of action of agents producing effects in nonspinal preparations. Furthermore, the effective doses of idazoxan were rather large (1-2 mg/kg), which raised the possibility that the actions of this drug may not have been mediated entirely through blockade of a2adrenoceptors. Another problem is that idazoxan is known to bind with high affinity to non-adrenergic imidazoline binding sites (NAIBS), which are found in many tissues including rabbit brain (Hamilton et al. 1988). The present experiments were designed to answer the questions of receptor selectivity and sites of action of idazoxan in facilitating the suraLGM reflex. In this study drugs have been delivered directly to the lumbosacral spinal cord via a subdural (intrathecal, i.t.) cannula to attempt to limit their range of action to the spinal cord. Four drugs have been used in this way to assess the receptor selectivity of idazoxan: idazoxan itself; the selective al-adrenoceptor antagonist prazosin; the nonimidazoline a2-adrenoceptor antagonist yohimbine; and the putative post-junctional c~z-receptor antagonist SK&F 104078 (Ruffolo et al. 1987). Some of these data have appeared as an abstract (Harris et al. 1990).
311 Material and methods
2.5-100 pl to give cumulative doses of 13, 26, 128, 256 and 512 nmol. After each intrathecal injection, the cannula was flushed with 60 ~tl of Ringer's solution. In some experiments, 60 gl of vehicle and a 60-gl flush of Ringer's solution was injected before any drug was given, to test for any effects of the vehicle on the sural-GM reflex. After intrathecal administration of drugs and in the 21 animals which received no drug treatment, the spinal cord was transected by two adjacent dorsoventral hemisections separated by an interval of 24 min. On post-mortem examination, after exposure of the lumbosacral spinal cord, 60 gl of pontamine sky blue followed by 60 gl of Ringer's solution were injected through the i.t. cannula to assess the spread of drugs around the spinal cord. This indicated that most of the spinal cord caudal to L2 was bathed in drug solutions.
Surgical procedures Experiments were performed on 63 New Zealand Red (NZR) and NZR x New Zealand White rabbits, of either sex, weighing between 1.94 and 3.46 kg. Anaesthesia was induced by an intravenous injection of 10-20 mg/kg methohexitone sodium (Brietal; Eli Lilly) which was supplemented to effect. The trachea was exposed and cannulated so that anaesthesia could be continued with a mixture of halothane (ICI or May and Baker) in nitrous oxide and oxygen (70:30%). One carotid artery was cannulated to record arterial blood pressure and the other carotid artery was ligated. For the purpose of drug administration, two cannulae were inserted into one jugular vein. The electrocardiogram (ECG) was recorded via an intraoesophageal probe. A laminectomy was performed between spinal segments T 12 and L2 and a cannula inserted beneath the dura so that the tip lay at or near the L7 and S1 spinal segments. The position of this cannula was verified on post-mortem examination. All rabbits were decerebrated by suction to pre- (n = 56) or midcollicular (n = 7) levels. The core temperature of the rabbit was recorded with a rectal probe and maintained between 37.5 and 38.5~ C using a thermostatically controlled heating blanket. The left hindlimb was clamped rigidly at the knee and ankle, and the sciatic nerve exposed by section and retraction of the overlying muscle groups. The resultant pool was filled with warmed paraffin oil (37~ C). The sural nerve was dissected free, cut, and placed over paired platinum stimulating electrodes. A further portion of the sural nerve was freed proximal to the stimulating electrodes to allow recording of the afferent volley. The GM nerve was dissected free, cut, desheathed and the central end placed over a pair of platinum recording electrodes. When surgery was completed, anaesthesia was discontinued and the animal was paralysed with gallamine triethiodide (Flaxedil; May and Baker; 4 mg/kg i.v. initially). Ventilation was continued using a Starling ideal pump with the stroke volume adjusted to give an end-tidal CO2 of 3-5%. A period of at least 1 h was left between cessation of halothane anaesthesia and the beginning of recording.
Nerve stimulation and recording Reflexes were evoked by electrical stimulation of the sural nerve with square-wave pulses of 0.1 ms duration, applied at a frequency of 1 Hz, and of a strength sufficient to recruit all myelinated axons (i.e. 20-80 times threshold). This was confirmed by observation of the afferent volley recorded from the nerve. Reflexes were recorded from the GM muscle nerve which was crushed between the electrodes to give a monophasic signal. Stimuli were applied in blocks of 12 at 2-min intervals and, to allow for wind-up in the reflex pathway (see Catley et al. 1983), the responses to stimuli 4-11 in each block were averaged and integrated by computer.
Statistical analysis In each experiment reflex responses were normalized to the mean of the pre-drug control level. Pooled data from many experiments are expressed as median percentages, and scatter of results is indicated by the range. Statistical analysis of the results was performed using Wilcoxon's paired ranks test; Mann-Whitney U-test; Kruskal-Wallis ANOVA; and Fisher's exact probability test, assuming a significance level of 0.05 throughout.
Results The responses o f G M to sural nerve s t i m u l a t i o n were extremely v a r i a b l e b e t w e e n a n i m a l s , as r e p o r t e d previously (Clarke et al. 1988a). F o r the a n i m a l s which were to receive i d a z o x a n , the m e d i a n p r e - d r u g reflexes were 76 g V - m s (range 11 1072 g V . m s ) a n d the corres p o n d i n g values for the other d r u g t r e a t m e n t g r o u p s were: y o h i m b i n e , 132 g V . m s (range 20 1883 ~tV-ms); prazosin, 91 g V - m s (range 13-578 p V . m s ) ; a n d S K & F 104078, 41 g V . m s (range 2 6 - 1 1 9 2 g V - m s ) respectively. There were n o significant differences b e t w e e n a n y o f these g r o u p s ( K r u s k a l - W a l l i s test, P > 0 . 2 ) . 0 IDAZOXAN
BOO
PRAZOSIN [] SK&FIO407B 9 YOHII~BINE
o/ a
I
500
[1_ z
400 5OO
Drug administration J ta_ Ld
The following drugs were used: idazoxan hydrochloride (a gift of Dr. S.L. Dickinson, Reckitt and Colman) dissolved in Ringer's solution (8.3 mmol/1) and administered intrathecally in volumes of 2.5-100 gl to give cumulative doses of 21, 42, 208, 415 and 813 nmol; prazosin hydrochloride (a gift of Dr. R. Packard, Pfizer) dissolved in 0.4% dimethyl sulphoxide (DMSO) in 5% D-glucose solution (4.8 mmol/1) and administered intrathecally in volumes of 2.5 100 gl to give cumulative doses of 12, 24, 119, 238 and 476 nmol; SK&F 104078 maleate (a gift of SmithKline Beecham) dissolved in 2% DMSO in Ringer's solution (5.1 mmol/l) and administered intrathecally in volumes of 2.5 100 gl to give cumulative doses of 13, 25, 126, 253 and 505 nmol; and yohimbine hydrochloride (Sigma) dissolved in 1% dimethyl formamide (DMF) in 5 % D-glucose solution (5.1 mmol/1) and administered intrathecally in volumes of
700
~:
200
', 0.5). Figure 6 shows the effects of spinalization in individual idazoxan-treated rabbits and those which were spinalized without prior drug treatment. It also includes, for comparison, data from other experiments in which spinalization was carried out in the presence of saturating intravenous doses of naloxone alone, and naloxone with idazoxan (from
DRUG
I
I
I
NLX
IDZ
IDZ
ONLY
+NLX
ONLY
Fig. 6. The effect of spinalization on the sural-gastrocnemius reflex
in individual rabbits in the presence of different drug milieux, in this case with the response normalized with respect to the post-drug (i.e. immediate pre-spinalization) level. The figures show, from right to left, the effects of spinal section in untreated animals (NO DRUG, n= 21); in the presence of naloxone 388.5 gg/kg i.v. (NLX ONLY, n= 12); in the presence of naloxone 250 or 388.5 Mg/kg i.v. with idazoxan 2 or 3 mg/kg i.v., (IDZ+ NLX, n = 14); and in the presence ofidazoxan 1517 nmol i.t. (IDZ ONLY, n = 11). Only the data from this last column were obtained in the present study: the figures for untreated animals are from a separate set of experiments carried out contemporaneously with these, but yet to be published, and the data with naloxone are from Clarke et al. (1988a). Each point is the pooled mean of responses recorded over a 24-min period after spinalization Clarke et al. 1988a). When the spinal cord was sectioned in the absence of drug treatment, the s u r a l - G M reflex usually increased in size, although it decreased in 7 of 21 animals. The outcome of spinalization in the absence of drug treatment was significantly different from spinal section in the presence of idazoxan (Fisher's exact probability = 0.0002); naloxone (P = 0.027); or idazoxan with naloxone (P=0.04). There was also a significant difference between spinalization in the presence o f naloxone compared with naloxone plus idazoxan ( P = 0.0003).
Discussion
The effects of intrathecal admin&tration of a-receptor antagonists Intrathecal administration of nanomolar quantities of idazoxan caused a sixfold increase in G M responses to sural nerve stimulation. This effect was almost certainly due to a direct action of the drug at the spinal cord, as intravenous administration of similar doses results in much smaller increases in the s u r a l - G M reflex (Clarke et al. 1988a). Idazoxan is a highly selective %-adrenoceptor antagonist with some 40-fold higher potency at a2 than at al receptors (Doxey et al. 1983; Megens et al. 1986) and no appreciable affinity at receptors for other known
315 transmitters (Pettibone et al. 1986; although, see below). The fact that prazosin, a highly selective al-receptor antagonist (see Wilson et al. 1991), had no significant effects on the sural-GM reflex when given alone, is good evidence that the facilitatory action of idazoxan is not due to the result of blockade of 0h-receptors. On the basis of these few facts, one might conclude that idazoxan causes facilitation of the reflex by blockade of inhibitory cz2-receptors in the spinal cord. Furthermore, the fact that the cz2-agonist clonidine suppresses reflexes in spinal rabbits (Clarke et al. 1988a), indicates that inhibitory adrenoceptors are post-synaptic to adrenergic terminals. The receptors involved do not appear to be the same as the post-junctional a2 sites found in the periphery, as SK&F 104078, a putative selective antagonist at these receptors (Ruffolo et al. 1987), had no effect on the sural-GM reflex. However, this drug also binds with moderate affinity to 5-HT~A and 5-HT2 receptors (Kilpatrick et al. 1989), which complicates interpretation of the effects of this material. It is not yet possible to say with certainty that all of the effects of idazoxan are mediated entirely through cz2-receptors, as the present data have not eliminated the possibility that some of the actions of this drug might be mediated by actions at non-adrenergic imidazoline binding sites (see Introduction). The functions of these sites have yet to be defined. Yohimbine, long promoted as a selective cz2-receptor antagonist (see Goldberg and Robertson 1983), does not bind to NAIBS. In the present study this drug did facilitate the sural-GM reflex, although it was some 15 times less potent than idazoxan and had a much lower peak effect. The fact that both idazoxan and yohimbine facilitated GM reflex responses indicates that blockade of cz2-receptors is one aspect of the facilitatory action of idazoxan, but the disparities between the potencies and ceiling effects of idazoxan and yohimbine cannot be explained only by actions at a2-receptor sites (see Doxey et al. 1983; Megens et al. 1986; Pettibone et al. 1986). There are two possible explanations for the differences between the two drugs: either the effects of idazoxan are due in part to an action at NAIBS or yohimbine has actions at other receptors which restrict the expression of the cz2-blocking action of this drug. The fact that idazoxan had no effect after yohimbine suggests that the latter explanation is correct. Although yohimbine has often been used as a selective cz2-adrenoceptor antagonist, it has a range of actions at receptors for 5-HT. Yohimbine or its enantiomers have been shown to bind to 5-HT1A (Convents et al. 1989) and 5-HT2 (see Wilson et al. 1991) receptors and, most curiously, yohimbine is an agonist at the 5-HT terminal autoreceptor (presumed to be of the 5-HT~D type) in the rabbit (Feuerstein et al. 1985; see also Hoyer et al. 1990). Before further considering the significance of these facts, it is necessary to examine the complexities of the idazoxaninduced facilitation of the sural-GM reflex.
Idazoxan and descendin9 facilitation In an earlier paper we hypothesized that the major effect of idazoxan on the sural-GM reflex was due to the
blockade of adrenergic descending inhibition (Clarke et al. 1988a). Results obtained in the present experiments show this conclusion to be far too simplistic. When the spinal cord was sectioned in the presence of idazoxan alone, the reflex always decreased in size. This shows that the drug not only blocks inhibition in the spinal cord, it also reveals or potentiates descending facilitation of the sural-GM pathway. The interpretation of this finding is complicated by the fact that opioid peptides are tonically active in the spinal cord. In spinal rabbits, the sural-GM reflex is powerfully enhanced by the opioid antagonist naloxone (Clarke and Ford 1987), but this effect is much weaker in non-spinal animals (Clarke et al. 1988a). It is possible that the level of opioidergic tone is lower in non-spinal animals and that spinal section increases activity in spinal opioidergic neurones. If this were the case, one might expect cord section in untreated rabbits to have resulted only in decreases in the size of reflexes, which was clearly not the case. Furthermore, although spinal section in the presence of naloxone alone always results in an increase in the sural-GM reflex, if the cord is cut in the presence of naloxone and idazoxan together, the reflex usually decreases in size (Clarke et al. 1988a). Taken together, these data show that idazoxan really does facilitate reflexes by blockade of descending inhibition and revealing, or augmenting, descending facilitation. The present experiments provide some indicators to the identity of the transmitters involved in descending facilitation. High doses of prazosin alone resulted in depression of the sural-GM reflex in some rabbits, and the same drug reduced the facilitatory effects of idazoxan whether given before or after the ~2-antagonist. Thus, in addition to having an inhibitory influence over transmission through the sural-GM pathway by an action at a2-receptors (probably in the dorsal horn; see Fleetwood-Walker et al. 1985), noradrenaline may have a facilitatory role, acting through al-receptors. This is not a new idea: prazosin has been shown to reduce excitatory influences from the brain on reflex pathways in rat (Palmeri and Wiesendanger 1990; Tanabe et al. 1990) and cat (Lai et al. 1989) and noradrenaline-induced depolarization of motoneurones in neonatal rat spinal cord is known to be mediated by an o~l-receptor (Connell et al. 1989). The general import of this body of work is that noradrenaline facilitates motor function via an action at al-receptors probably located in the ventral horn (see also White and Neuman 1980; Bell and Matsumiya 1981). It is likely that 5-HT also is involved in descending facilitation of reflex function as it has similar actions to noradrenaline on ventral horn cells (White and Neuman 1980; Bell and Matsumiya 1981; Connell and Wallis 1988; Jackson and White 1990). As it is known that noradrenaline can modulate the release of 5-HT from slices of rabbit hippocampus (Feuerstein et al. 1985) and caudate nucleus (Feuerstein et al. 1992), one might speculate that idazoxan promotes descending facilitation of reflexes by increasing the release of noradrenaline and 5-HT in the ventral horn.
316
The modes of action of idazoxan and yohimbine To summarize, we propose, on the basis of the current evidence, that idazoxan facilitates the s u r a l - G M reflex by acting in three ways: (1) blockade of post-synaptic Ctz-receptors in the dorsal horn, resulting in the attenuation of adrenergic descending inhibition; (2) blockade of pre-synaptic a2-autoreceptors on the terminals of adrenergic axons in the ventral horn, resulting in increased adrenergic descending facilitation (mediated through al-receptors); and, most tentatively, (3) blockade of c~/-receptors on the terminals of 5-HT-ergic axons in the ventral horn, resulting in increased 5-HT-ergic descending facilitation. Yohimbine, being an ~2-antagonist and an agonist at the 5 - H T - a u t o r e c e p t o r , would have effect 1 and p r o b a b l y also 2, but would, if anything, reduce 5 - H T release and not have effect 3. This would explain why yohimbine produced less facilitation than idazoxan and why idazoxan produced no further increases in the reflex when given after yohimbine, although one has to bear in mind the possibility that these effects m a y be due to interaction o f this drug with other types of receptor (Convents et al. 1989). In support of our view we have recently shown that cyanopindolol, another agonist at the rabbit 5 - H T terminal autoreceptor (see H o y e r et al. 1990; Feuerstein et al. 1992), also blocks the facilitatory effects of idazoxan (Harris et al. 1992). The. apparently self-limiting actions of yohimbine m e a n that the possibility that imidazoline binding sites play a part in the actions of idazoxan cannot be eliminated. There are available drugs which are highly selective for the a2-receptor but have no affinity for N A I B S (e.g. see Langin et al. 1990); these will f o r m the basis of further investigations into this problem. The present results underline the complex nature of descending control of spinal function. It is clear that descending pathways which tonically modulate spinal reflexes in decerebrate animals can be facilitatory or excitatory and that the effects of spinal section or cooling are the net result of removing a range of influences. An i m p o r t a n t future step must be to identify definitively those sites in the spinal cord at which facilitation and inhibition are effected, and to define the amine receptor subtypes involved at each location (e.g. Bras et al. 1989, 1990). The key to this must be the use of highly selective receptor antagonists and agonists, with the fullest possible knowledge of the range of actions and limitations of each drug used.
Acknowledgements. This work was funded by the Agricultural and Food Research Council. John Harris is a Science and Engineering Research Council scholar. We are grateful to Caroline Northway and Irka Zajac-Galij for expert technical assistance, and to Reckitt and Colman, Pfizer and SmithKline Beecham for the supply of drugs.
References Bell JA, Matsumiya T (1981) Inhibitory effects of dorsal horn and excitant effects of ventral horn intraspinal microinjections of norepinephrine and serotonin in the cat. Life Sci 29:1507-1514
Bras H, Cavallari P, Jankowska E, McCrea D (1989) Comparison of effects of monoamines on transmission in spinal pathways from group I and II muscle afferents in the cat. Exp Brain Res 76: 27-37 Bras H, Jankowska E, Noga B, Skoog B (1990) Comparison of effects of various types of NA and 5-HT agonists on transmission from group-II muscle afferents in the cat. Eur J Neurosci2:1029-1039 Catley DM, Clarke RW, Pascoe JE (1983) Naloxone enhancement of spinal reflexes in the rabbit. J Physiol (Lond) 339: 63-72 Clarke RW, Ford TW (1987) The contributions of ~t-, 6- and ~-opioid receptors to the actions of endogenous opioids on spinal reflexes in the rabbit. Br J Pharmacol 91 : 579-589 Clarke RW, Ford TW, Taylor JS (1988a) Adrenergic and opioidergic modulation of a spinal reflex in the rabbit. J Physiol (Lond) 404: 407-417 Clarke RW, Ford TW, Harris J, Taylor JS (1988b) The effects of direct spinal application of idazoxan and prazosin on a spinal reflex in the decerebrated rabbit. Neurosci Lett Suppl 32:$32 Connell LA, Wallis DI (1988) Response to 5-hydroxytryptamine evoked in the hemisected spinal cord of the neonate rat. Br J Pharmacol 94: 1101-1114 Connell LA, Majid A, Wallis DI (1989) Involvement of a 1adrenoceptors in the depolarizing but not the hyperpolarizing responses of motorneurones in the neonate rat to noradrenaline. Neuropharmacology 28:1399-1404 Convents A, De Kayser J, De Backer JP, Vauquelin G (1989) [3H]Rauwolscine labels c%-adrenoceptors and 5t-ITIAreceptors in human cerebral cortex. Eur J Pharmacol 159:307-310 Doxey JC, Roach AG, Strachan DA, Virdee NK (1983) Effects of RX 781094 and yohimbine on the responses to UK 14304 at various ct2-adrenoceptors in rats. Br J Pharmacol 79:311P Feuerstein TJ, Hertting G, Jackisch R (1985) Endogenous noradrenaline as modulator of hippocampal serotonin (5-HT)release. Dual effects of yohimbine, rauwolscine and corynanthine as a-adrenoceptor antagonists and 5-HT-receptor agonists. Naunyn Schmiedebergs Arch Pharmacol 329:216-221 Feuerstein TJ, Lupp A, Hertting G (1992) Quantitative evaluation of autoinhibitory feedback of release of 5-HT in the caudate nucleus of the rabbit where an endogenous tone on a2 receptors does not exist. Neuropharmacology 31 : 15-23 Fleetwood-Walker SM, Mitchell R, Hope P J, Molony V, Iggo A (1985) An Ctz-receptormediates the selective inhibition by noradrenaline of nociceptive responses of identified dorsal horn neurones. Brain Res 334:243-254 Goldberg MR, Robertson D (1983) Yohimbine: a pharmacological probe for the study of the cz2-adrenoceptor. Pharmacol Rev 35:143-180 Hamilton CA, Reid JL, Yakubu MA (1988) [3H]-Yohimbine and [3H]-idazoxan bind to different sites in rabbit forebrain. Eur J Pharmacol 146:345-349 Harris J, Ford TW, Clarke RW (1990) The effects of intrathecal application of c~-adrenoceptor antagonists on the sural-gastrocnemius reflex in the decerebrated rabbit (abstract). J Physiol (Lond) 429 : 130P Harris J, White DP, Clarke RW (1992) The roles of 5-HT in modulating a spinal reflex in the decerebrated rabbit. Neurosei Lett Suppl 42: S 13 Hoyer D, Schoeffter P, Waeber C, Palacios JM (1990) Serotonin 5-HT1D receptors. Ann NY Acad Sci 600:168-181 Jackson DA, White SR (1990) Receptor subtypes mediating facilitation by serotonin of excitability of spinal motoneurons. Neuropharmacology 29 : 787-797 Kilpatrick AT, Brown CM, MacKinnon AC, Spedding M (1989) The az-adrenoceptor antagonist SK&F 104078 has high affinity for 5-HT1A and 5-I-IT2 receptors. Eur J Pharmacol 166:315-318 Lai Y-Y, Strahlendorf I-IK, Fung SJ, Barnes CD (1989) The actions of two monoamines on spinal motoneurons from stimulation of the locus coeruleus in the cat. Brain Res 484:268-272
317 Langin D, Paris H, Dauzats M, Lafontan M (1990) Discrimination between alpha-2 adrenoceptors and [3H]idazoxan-labelled nonadrenergic sites in rabbit white fat cells. Eur J Pharmacol 188:261-272 Megens AAHP, Leysen JE, Awouters FHL, Niemegeers CJE (1986) Further validation of in vivo and in vitro pharmacological procedures for assessing the ch/ct2 selectivity of test compounds : (1) ct-adrenoceptor antagonists. Eur J Pharmacol 129:49-55 Palmeri A, Wiesendanger M (1990) Concomitant depression of locus coeruleus neurons and of flexor reflexes by an a2-adrenergic agonist in rats: a possible mechanism for an ct2-mediated muscle relaxation. Neuroscience 34:177-187 Pettibone D J, Clineschmidt BV, Lotti VJ, Martin GE, Huff JR, Randall WC, Vacca J, Baldwin JJ (1986) L 654,284 a new potent and selective a2-adrenoceptor antagonist. Naunyn Schmiedebergs Arch Pharmacol 333:110 116
Ruffolo RR Jr, Sulpizio AC, Nichols A J, DeMarinis RM, Hieble JP (1987) Pharmacologic differentiation between pre- and postjunctional az-adrenoceptors by SK&F 104078. Naunyn Schmiedeberg Arch Pharmacol 336:415-418 Tanabe M, Ono H, Fukuda H (1990) Spinal alpha-1-adrenoceptors and alpha-2-adrenoceptors mediate facilitation and inhibition of spinal motor transmission, respectively. Jpn J Pharmacol 54:6%77 White SR, Neuman RS (1980) Facilitation of spinal motoneurone excitability by 5-hydroxytryptamine and noradrenaline. Brain Res 188:119-127 Wilson VG, Brown CH, McGrath JC (1991) Are there more than two types of ct-adrenoceptors involved in physiological responses ? Exp Physiol 76: 317-346
