The coeruleospinal noradrenergic neurons: Anatomical and electrophysiological studies in the rat

Brain Research, 189 (1980) 121-133 © Elsevier/North-Holland Biomedical Press

121

T H E C O E R U L E O S P I N A L N O R A D R E N E R G I C NEURONS: A N A T O M I C A L A N D E L E C T R O P H Y S I O L O G I C A L STUDIES IN T H E RAT

PATRICE G. GUYENET Department of Pharmacology, University o/" Virginia School of Medicine, 1300 Jefferson Park Avenue, Charlottesville, I/a. 22908 (U.S.A.)

(Accepted September 27th, 1979) Key words: locus coeruleus -- spinal projection -- HRP -- electrophysiology

SUMMARY The neuroanatomical location and pharmacological sensitivity of coeruleospinal neurons were studied with a combination of retrograde tracing experiments and single unit recording. Coeruleospinal neurons were multipolar, medium-sized cells and were found in the ventral division of the locus coeruleus and in the locus subcoeruleus. In the locus coeruleus proper, they presumably corresponded to the large cells of the ventral division defined in previous Golgi studies. Coeruleospinal cells were identified by antidromic stimulation from the cervical spinal cord. Their firing rate was slow and regular, their conduction velocity characteristic of unmyelinated fibers (0,65 m/sec). The method o f antidromic stimulation also revealed that coeruleospinal neurones possess an anteriorly directed collateral traveling in the midbrain reticular formation outside the main noradrenergic dorsal bundle. These neurones were strongly inhibited by the iontophoretic application of morphine, noradrenaline, clonidine, GABA and excited by ACh. Although the coeruleo- and subcoeruleospinal neurones are clearly a group of cells distinct from the coeruleocortical projection, their electrophysiological and pharmacological properties are essentially identical.

INTRODUCTION The presence of catecholaminergic fibers and terminals in the spinal cord was described by Carlsson et al. a, but the contribution of the locus coeruleus to the noradrenergic innervation of this structure, mainly of its ventral horn, has been recognized only recently1,13,sz,a5. Despite an early claim that all locus coeruleus neurones might send an axonal collateral into the spinal cord 31, several recent studies

122 using the retrograde transport of horseradish peroxidase (HRP) have suggested that the spinal projection originates from the ventral division of the locus coeruleus and from the nucleus subcoeruleusZ2,'~6,27,33,3s. Interestingly, these two areas also contain the cells of origin of the so-called 'subcoeruleus projection' which innervates the periventricular hypothalamus and the preoptic area28, z2. These neurons as well as coeruleospinal neurons appear significantly larger than coeruleocortical neurons in histofluorescence32, HRP as and Golgi studies 49. These peculiarities suggested that the cells belonging to the subcoeruleus complex (ventral division of the locus coeruleus and nucleus subcoeruleus) might have different physiological properties than the coeruleocortical neurons. The present study was therefore initiated to identify coeruleospinal neurons electrophysiologically and to determine their pharmacological sensitivity to a number of neurotransmitters and pharmacological agents. The location of these cells was reinvestigated with the retrograde transport of HRP to clarify some remaining inconsistencies and to provide a useful parallel with the electrophysiological data. MATERIAL AND METHODS All experiments were made with male Sprague-Dawley 250-300 g rats.

Retrograde tracing experiments The rats were anesthetized with sodium pentobarbital (50 mg/kg i.p.). Horseradish peroxidase (Sigma VI, 50 ~ solution in distilled water) was pressure-injected in the spinal cord at the cervical (C6-C7) or lumbar level using a glass micropipette (tip 80 /zm approximately). The tips of the pipettes were aimed at the ventrolateral funiculus on one side where coeruleospinal axons are known to travel 32. The injections intentionally created some tissue damage at the tip in order to label as many locus coeruleus cells as possible via the axonal uptake of the enzyme. Each animal received a total of 2/zl in 3 separate deposits 0.2 mm apart and was sacrificed 40 h later. Two histochemical procedures were used, the Hanker-Yates and the tetramethylbenzidine method of Mesulam za,29. In both cases the animals were perfused intracardially with 1 liter of a solution of 1.25~ glutaraldehyde and i~o paraformaldehyde in 0.1 M sodium phosphate buffer, pH 7.4; following one night in cold 10 ~o sucrose (in sodium phosphate buffer, pH 7.4) frontal 40/~m sections were cut on a freezing microtome. The following steps were performed according to the original articles quoted above.

Electrophysiological recordings The rats were initially anesthetized with chloral hydrate (400 mg/kg i.p.), then supplementary doses of the anesthetic were given intravenously via the tail vein during the course of the experiment. Body temperature was maintained at 37 °C with a feedback controlled warming pad. A 2.5 mm burr hole was drilled in the occipital bone at the approximate coordinates of the locus coeruleus and the dura was slit just posterior to the sinus. Electrodes were lowered vertically 4 mm into the brain behind the sinus which was then gently pushed forward by the shaft of the electrodes as

123 required to obtain a correct vertical stereotaxic placement. This procedure combined the accuracy of vertical electrode placements and the preservation of the occipital sinus. Two types of electrodes were used: single-barrel glass electrodes (4-8 M f~) filled with a solution of 2 % fast green in 2 M NaC1 or 5-barrel glass electrodes whose recording central barrels were filled with the same solution. In the latter case, 3 side barrels contained one of the following solutions: morphine sulfate 50 mM, pH 4.0; acetylcholine 0.2 M, pH 4.0; noradrenaline bitartrate 0.1 M, pH 4.0; clonidine HCI, 20 mM in NaC1 160 mM, pH 4.0, GABA 10 mM in NaCI 160 mM, pH 4.0. The fourth barrel contained 4M NaC1 and was used for automatic current balancing. Electrical signals were amplified using conventional equipment, filtered and displayed on an oscilloscope. The use of a rate-averaging computer permitted to chart integrated rate-histograms of the neuronal firing. A 4-channel unit equipped with automatic current balancing was used for iontophoresis. All these techniques have been previously described 7,2°. The locus coeruleus was found at a depth of 5.8-6.5 mm below the pial surface of the cerebellum, 1.5-2.5 mm behind the lambdoid suture and 1-1.3 mm lateral to the midline. As described earlier, the nucleus is characterized by a high density of units firing at a regular low rate (0.1-5 per sec) in the chloral hydrate anesthetized ratt0,11. It is located just medial to the mesencephalic nucleus of the fifth nerve which fires characteristic high frequency bursts in synchrony with passive movements of the lower jaw and above an area of the reticular formation characterized by low amplitude rhythmic unit activity. The central gray area medial to the locus coeruleus displays little if any recordable activity. These landmarks make the localization of the locus coeruleus rather straightforward. In addition, we also confirmed the fact that stimulation of the dorsal noradrenergic bundle produces in the locus coeruleus a negative field potential (up to 1 mV) synchronous with the antidromic driving of the neurones 24. The field potential is best recorded at the level of the posterior two-thirds of the nucleus and is not recorded in adjacent nuclei. At the end of each experiment fast green was iontophoresed for the final identification of the recording site (20 #A negative DC for 40 min). In 8 animals, a single green spot was used for the histological localization of two recorded cells; the site of the last recorded cell corresponded to the dye spot and the position of the first was charted using the relative stereotaxic coordinates of the two cells as well as the vertical distance from the first cell to the bottom of the nucleus locus coeruleus.

Brain stimulation Bipolar concentric electrodes (Kopf SNE I00) were placed in the dorsal noradrenergic bundle (coordinates anterior 2.0, lateral 0.7, depth 6.1 below the dura) and/or in the cervical spinal cord (level C6-C7), 2 mm below the dorsal surface and 0.5 mm away from the midline (approximate location of the ventrolateral funiculus). Constant current square-wave pulses were delivered with a stimulator coupled with stimulus isolation units (W.P.I. 302-T). At the end of the experiments, a lesion was made at the stimulating site by passing 100 #A DC for 20 sec.

Histological procedures The animals were perfused with 10 % buffered formalin. The brain and spinal

/

V

/.'

A':

i/'

"--

t'/

Fig. 1. Location of coeruleospinal neurons. A. frontal section corresponding to level 2 of Fig. ID. Labeled cells are present in the ventral part of the locus coeruleus and in the nucleus subcoeruleus. (Hanker -Yates method). B: this section corresponds to level 4 of Fig 1D (Hanker-Yates method). C: level 3 of Fig 1 D (Mesulam technique). Labeled cells are located in the lower third of the locus coeruleus whose extent {s recognizable at its darker color. D : schematic drawing of the location of coeruleospinal neurons in 5 consecutive frontal sections 160 # m apart. E: typical location of a coeruleospinal neuron recorded electrophysiologically. The fast green spot (circled) is located at the bottom of the nucleus (cresyl violet and neutral red stain). Abbreviations: CG, griseum centrale; GF, genu nervi facialis; LC, locus coeruleus; LsC, nucleus subcoeruleus; Mv, nucleus motorius nervi trigemini; PCS, pedunculus cerebellaris superior; Vm, nucleus mesencephalicus nervi trigemini.

125

Fig. 2. Effect of spinal cord stimulation on the activity of coeruleospinal neurons. A: seven consecutive sweeps illustrate the antidromic activation of a coeruleospinal neuron. Note A-B inflexions, constant latency of the response and occurrence of collisions with spontaneous spikes (latency of antidromic spike = 33 msec, spike amplitude 0.4 mV). In this and following illustrations, positivity is upwards. B: Collisions between spontaneous spikes (s) and antidromic spikes (a). The stimulus (arrow) was triggered 30 msec (top, arrow) and 28.5 msec after a spontaneous spike (3 superimposed sweeps). C and D: two records of the same cell with different time scales and stimulus intensities. Repetitive stimuli applied with just above threshold intensity (0.25 mA, 0.2 msec duration) demonstrate that the cell is antidromically activated (D upper trace, 5 sweeps). Raising the stimulus intensity to 0.32 mA (D lower trace) results in orthodromic driving and disapFearanee of all antidromic spikes except one (10 sweeps). The orthodromic driving (C) is followed by a pause of long duration (40 sweeps). Sweep duration is 1 sec in C and 50 msec in D. Spike amplitude is 0.6 inV. a, antidromic spike; s, spontaneous spike; arrow = stimulus. c o r d were sliced with a freezing m i c r o t o m e (40 btm). The sections were stained with a c o m b i n a t i o n o f cresylviolet a n d n e u t r a l red a n d the l o c a t i o n o f the fast green spot (recording site) a n d the lesions ( s t i m u l a t i o n sites) were r e c o r d e d 7. RESULTS

Mapping of coeruleospinal neurons with horseradish peroxidase U n i l a t e r a l multiple injections o f H R P were m a d e in the v e n t r o l a t e r a l funiculus o f the cervical spinal c o r d (8 rats) in o r d e r to label as m a n y n e u r o n e s as possible and thus to d e t e r m i n e the t o p o g r a p h y o f coeruleospinal neurons. C o n t r o l unilateral injections were m a d e in the p a r i e t a l a n d occipital cortices for c o m p a r a t i v e p u r p o s e s (3 injections p e r a n i m a l t o t a l i n g 2/~1 o f H R P in 2 rats). T h e H a n k e r - Y a t e s (8 animals) o r the M e s u l a m p r o t o c o l (2 animals) were used to visualize the r e t r o g r a d e l y t r a n s p o r t e d

126

Fig. 3. Refractory period of coeruleospinal and other locus coeruleus neurons. A: neuron activated antidromically at low intensity from the dorsal noradrenergic bundle (50/tA, 0.2 msec). Top traces show A-B break and failure of B segment (calibration 0.2 mV, 5 msec). Bottom traces show A segment of antidromic spike following high stimulation rate (calibration 0.2 mV, 2 msec). B: neuron activated antidromically from spinal cord. Note A-B break, high following frequency of A spike (each trace 3 sweeps, calibration 0.2 mV, 5 msec). enzyme. Although the second technique usually revealed more cells than the first, the overall labeling pattern was identical following all spinal injections. Labeled cells were found almost exclusively in the ventral third of the locus coeruleus along almost its whole extent except the very anterior pole (Fig. I A, B, C and D). Labeled cells were also found in the locus subcoeruleus. All labeled cells were multipolar and significantly larger than those present in the two animals which had received the cortical injections. Coeruleocortical neurons were found in the dorsal division of the locus coeruleus along its whole extent. Up to 180 labeled cells were counted after one of the spinal injections of HRP, using the most sensitive technique (64 in the nucleus subcoeruleus and 116 in the ventral division of the locus coeruleus). This figure was obtained by counting all the labeled cells in each 40 # m section throughout the nucleus and may be slightly larger than the actual number of cells due to occasional double-counting.

Electrophysiological identification of coeruleospinal neurons A total of 33 units in the locus coeruleus (25 animals) could be activated with a fixed latency (means ~ S.D. : 38 -k 8 msec, range 28-52) when the spinal cord was stimulated at the level of the sixth or seventh cervical vertebra (0.2 msec pulses;

127

Fig. 4. Collaterals of coeruleospinal neurons. A: constant latency response of a locus coeruleus neuron upon stimulation of both the dorsal bundle (top trace, calibration 1 msec 0.2 mV, 3 sweeps) and the spinal cord (bottom trace, calibration 5 msec, 0.2 mV, 3 sweeps). B- collissions between spontaneous (s) and evoked spikes (a), during spinal cord stimulation (arrow) indicate that evoked spike is antidromic (3 sweeps, calibration 10 msec, 0.2 mV). C" same cell is driven from the dorsal bundle (calibration 0.2 mV, 2 msec). Collisions indicate that evoked spike (a) is antidromic. threshold 50-400/zA). Based on a straight line distance of 23-25 mm, a conduction velocity of 0.65 m/sec could be estimated (range 0.45-0.86). Subsequent histological determination of the recording site (fast green spot) revealed that 30 units were in the ventral division of the locus coeruleus (Fig. 1E), and only 3 in the locus subcoeruleus. Their location coincided exactly with the topography of coeruleospinal neurons obtained by retrograde labeling with H R P (Fig. 1E). The presence of collisions between spontaneous spikes and these fixed-latency action potentials suggested that the latter were generated antidromically (Fig. 2A). The presumed antidromic spikes displayed a clear A-B inflexion (Fig. 2D) indicating that the recorded action potentials were of somadendritic and not axonal origin. This point will be confirmed below when the results of iontophoretic experiments are described. In order to substantiate further the fact that the fixed latency responses obtained during the stimulation of the spinal cord were antidromically rather than synaptically mediated, a quantitative analysis of the time relation between the minimum collision interval c, refractory period r, and latency 1, was performed in 10 ceils using the criteria described by Darian-Smith et al. 14 and Fuller and Schlag is. Indeed, only in the case of an antidromic activation, should the relation c ---- 1 d- r be verified; c was determined

128 by triggering the stimulus (1.5 times threshold) at different intervals following a spontaneous spike and measuring the delay producing 50 ~o failure of the presumed antidromic invasion (Fig. 2B); r is the minimum interval between two propagated spikes. Its determination requires a brief comment. Indeed, when paired shocks (l.5 times threshold) were applied, A-B dissociation of the second spike occurred (Fig. 3B). The refractory period of the B spike was long and variable (means ::c S.D. : 20 :::t: 12 msec, n = 4). The A spike (Fig. 3B), however, could follow at stimulus intervals close to 2 msec (means ± S.D. 2.1 ± 0.2 msec, n -- 4). Using the refractory period of the A spike as the value of r, the relationship c ~ r + 1 was verified in each case within 0.5 msec or 1.3 ~ of the conduction velocity, thus meeting the most stringent requirements of the collision test. No difference was noted between the conduction velocity of neurons found in the locus coeruleus (30 cells) and the locus subcoeruleus (3 cells). The firing rate of coeruleospinal neurons was low and regular (means ± S.D. 1.2 ~ 0.9 spikes/sec; range 0-4.5/sec, n ~ 25). The occasional antidromic activation of a silent cell suggests that an unknown proportion of the neurons did not fire spontaneously. Identified coeruleospinal neurons as well as other locus coeruleus neurons responded to a nociceptive stimulus (toe-pinch) or to the electrical stimulation of the spinal cord by a brief burst of activity followed by a pause of long duration (Fig. 2C). The latencies of the orthodromic responses following spinal cord stimulation were variable (12-35 msec) but always shorter than that of the antidromic activation (Fig. 3D). Consequently, antidromic activation could only be observed when its threshold was lower than that for orthodromic activation (Fig. 2D), since the reverse situation would have resulted in the systematic extinction of the antidromic spikes by collision. Since the threshold range for antidromic and orthodromic activation of locus coeruleus neurons were overlapping, it is reasonable to assume that a number of coeruleospinal neurons were recorded but could not be identified as such because of the extinction of antidromic spikes by the orthodromic driving. This phenomenon may explain in part why only 3.5~'o of all recorded locus coeruleus neurons could be identified as projecting to the spinal cord while the HRP data indicate that a far greater proportion of the cells belong to this pathway.

Collaterals of coeruleospinal neurons In 6 animals, a second bipolar stimulating electrode was located in the dorsal noradrenergic bundle in order to determine whether coeruleospinal neurons possess collaterals traveling through this area of the midbrain reticular formation. In agreement with previous reports 2,24 stimulation of the dorsal bundle (0.2-0.5 msec, 50-500/~A) resulted in a negative going field potential restricted to the posterior twothirds of the nucleus and coincidental with the antidromic activation of up to 90 0/,, of locus coeruleus neurons. The threshold for eliciting antidromic responses was very significantly lower in the anterior and dorsal part of the nucleus than in the ventral and posterior aspects. The following characteristics of these neurons were indistinguishable from those of identified coeruleospinal neurons: conduction velocity im ~ S.D. : 0.53 ± 0.1 m/sec, n = 23), A-B break upon repetitive stimulation, refractory periods

129 5

10

10

NE ACh

',

"O == O

i

10

C

t

~

it

10

NE

'

J

' 5

10

lO ACh

M

ACh

--

q.)

10 z--

10

AC_..h N E

ACh

m

m

I0 M

e~

'

t

t

Fig. 5. Pharmacological sensitivity of coeruleospinal neurons. All three cells were antidromically identified as coeruleospinal neurons. Each bar represents the cell firing integrated over a 10 sec period. The numbers refer to the intensity of the iontophoretic current in hA. Lower left and lower right represent two typical time-courses for the cell recovery following the application of morphine. Note the long duration of the inhibition by clonidine as compared to noradrenaline. of the A and B components of the antidromic spikes (means ± S.D. ----2.6 q- 0.6 msec, n -- 9; 32 4- 25 msec, n -- 39). These characteristics are illustrated in Fig. 3A. The average firing rate of the locus coeruleus neurons that were antidromically activated from the dorsal bundle was significantly higher, however, than that of identified coeruleospinal neurons (means 4- S.D. -- 2.6 :k 1.2 spikes per sec, n ---- 24 vs 1.2 40.97 spikes/sec, n -- 25). Four identified coeruleospinal neurons in a total sample of 8 were also antidromically activated by stimulating the dorsal bundle (Fig. 4). The threshold for eliciting the response was 0.3 mA and upwards. Four cells could not be antidromically activated even with currents as high as 3 mA. These experiments suggest that the forward-oriented collaterals ofcoeruleospinal neurons travel some distance away from the center of the dorsal bundle. In order to exclude the possibility of a sampling bias, in a control series of experiments two rats were implanted with an electrode in the cerebellar cortex and one in the dorsal bundle. Four cells were found to be antidromically activated from the cerebellum and all of them were also driven antidromically from the dorsal bundle. In this case, however, the threshold for antidromic activation from a dorsal bundle placement was at the lower end of the spectrum (50-200 pA) and all 4 cells were located in the dorsal division of the nucleus.

Phamacology of coeruleospinal neurons Various neurotransmitters and pharmacological agents that were previously

130 shown to be active on locus coeruleus neurons of undetermined projection were applied via iontophoresis at the vicinity of antidromically identified coeruleospinal neurons. These neurons were excited by ACh (7 cells tested), inhibited by norepinephrine (5 cells), morphine (6 cells), clonidine (4 cells) and GABA (1 cell), delivered with current intensities between 5 and 15 nA (representative examples are given in Fig. 5). DISCUSSION The present mapping of coeruleospinal neurones is in complete agreement with previous reportsSZ, 8s and provides further evidence that the cortical and spinal projections of this nucleus originate from distinct although slightly overlapping parts of the nucleus. In the experiments where the most sensitive technique was used, up to 180 labeled cells were counted ipsilateral to the injection site; 64 in the locus subcoeruleus and 116 in the ventral part of the locus coeruleus proper. All labeled cells were multipolar and those located in the locus coeruleus proper fitted the description given by Swanson 42 of the larger neurones belonging to his cytoarchitectonically defined ventral division. Thus, in the rat, the ventral division of the posterior twothirds of the locus coeruleus (see Swanson a2, Fig. 1) and the locus subcoeruleus are presumably a continuum of neurones, a large proportion of which project to the spinal cord. Most of these neurones are thought to contain noradrenalinela, 3s. Swanson estimates at 210 the number of large cells of the ventral division of the locus coeruleus. Up to 55 ~ of these could be labeled with HRP in our experiments, therefore, it is possible that all of them are coeruleospinal neurones; indeed, despite the large size of our HRP injections and their location in the ventrolateral funiculus every neurone could not possibly have taken up the enzyme. Despite the differences in the morphology and anatomical localization of coeruleospinal and other locus coeruleus neurones, the present study emphasizes the fact that their electrophysiological and pharmacological properties are very similar. The estimated conduction velocity of coeruleospinal neurones was 0.5-0.9 m/sec, a value in close agreement with that of the anteriorly oriented projections ~,16,3°. This slow conduction velocity is identical to that of other monoaminergic neurones in the CNS like Ag-A10 dopaminergic neurones 20 or serotoninergic neurons in the dorsal raphe 45. It is also similar to that of the noradrenergic postganglionic sympathetic neurones s. A possible exception may be the bulbospinal serotoninergic neurons projecting from the medullary raphe to the spinal cord which may have a higher conduction velocity17,46. However, the serotoninergic nature of the slow-conducting raphespinal neurones has not been demonstrated unambiguously. Both identified coeruleospinal and other locus coeruleus neurones are activated orthodromically by stimulating the cervical spinal cord or by applying a nociceptive stimulus 4z. This activation is followed by a pause of long duration which in the case of neurones antidromicaUy identified by stimulating the dorsal noradrenergic bundle has been attributed to collateral inhibition2,11. Since coeruleospinal neurons exhibit the same characteristics and are also sensitive to noradrenaline, they may be subject to the same inhibitory mechanism.

131 Both coeruleospinal and other locus coeruleus neurones have the same refractory period (2 msec) and the following frequency of the B component of antidromic spikes is low (see also ref. 43). This following frequency is highly dependent on the anesthetic used, chloral hydrate, and ketamine producing failure of the B spike at much lower frequency than a-chloralose or sodium pentobarbital (results not shown). Locus coeruleus neurones possess highly collateralized axonal processesa, 40, 42,44. The present study provides electrophysiological evidence that coeruleospinal neurones have a collateral directed towards the forebrain and preliminary evidence that these axons do not travel in the dorsal noradrenergic bundle. The presence of an anteriorly oriented collateral is demonstrated by the fact that 50~ of the identified coeruleospinal neurones sampled in this study could also be activated antidromically by stimulating an area of the midbrain corresponding to the location of the dorsal noradrenergic bundle zS. In agreement with previous studies z,11, up to 90 ~ of the cells in the locus coeruleus could be antidromically activated from these electrode placements with low (50-200/~A) intensities. The fact that the cells located in the posterior and ventral portions of the nucleus required in general higher intensities to elicit antidromic spiking suggests the possibility that to the topography of locus coeruleus cell bodies corresponds a precise organization of noradrenergic fibers in the dorsal bundle. The ascending axons of the 'subcoeruleus projection' are supposed to travel in the ventral noradrenergic bundle ~ or somewhere inbetween the dorsal and ventral noradrenergic bundles in an intermediate pathway2s. Since coeruleospinal neurones required the highest current intensities for eliciting an antidromic response from the dorsal bundle, their anteriorly oriented collaterals probably travel in this intermediate tract or in the ventral noradrenergic pathway. The neurones of the subcoeruleus projection innervate mainly the periventricular hypothalamus and preoptic area2S,3L Whether these and the coeruleospinal neurones are identical remains still to be demonstrated. Coeruleospinal neurones possess the same sensitivity to neurotransmitters and pharmacological agents as other locus coeruleus cells. Indeed, they are excited by ACh and inhibited by GABA, clonidine, noradrenaline and morphineS,10,21. Whether the coeruleospinal projection plays a contributive role in the antinociceptive or motor effects of morphine remains to be established. The anatomical connections of this group of cells are not inconsistent with this possibility12,34,36. Moreover, manipulations of the locus coeruleus as a whole have resulted in alterations of the nociceptive threshold in the rat6,89,41. Because of its preferential innervation of the ventral horn and the importance of catecholamines in spinal reflexes4,15,1a,a7, the coeruleospinal projection could be involved in the effect of narcotic drugs on motor reflexes associated with nociception. ACKNOWLEDGEMENTS This research was supported in part by a Research Starter Grant from the Pharmaceutical Manufacturer's Association Foundation.

132 C l o n i d i n e was a gift f r o m Dr. G a r b u s ( B o e h r i n g e r I n g e l h e i m ) . T h e a n a t o m i c a l e x p e r i m e n t were c a r r i e d o u t in the l a b o r a t o r y o f Dr. O. S t e w a r d ( U n i v e r s i t y o f Virginia, D e p a r t m e n t o f N e u r o s u r g e r y ) .

REFERENCES 1 Ader, J. P., Postema, F. and Korf.J., Contribution of the locus coeruleus to the adrenergicinnervation of the rat spinal cord, J. Neural Trans., 44 (1979) 159-173. 2 Aghajanian, G. K., Cedarbaum, J. M. and Wang, R. Y., Evidence for norepinephrine-mediated collateral inhibition of locus coeruleus neurons, Brain Research, 136 (1977) 570-577. 3 Amaral, D. G. and Sinnamon, H. M., The locus coeruleus: neurobiology of a central noradrenergic nucleus, Progr. NeurobioL, 9 (1977) 147-196. 4 Anden, N. E., Jukes, M. G. M. and Lundberg, The effect of L-DOPA on the spinal cord. 2: a pharmacological analysis, Acta physiol, scand., 67 (1966) 387-397. 5 Bird, S. J. and Kuhar, M. J., Iontophoretic application of opiates to the locus coeruleus, Brain Research, 122 (1977) 523-533. 6 Bodnar, R. J., Ackermann, R. F., Kelly, D. D. and Glusman, M., Elevations in nociceptive thresholds following locus coeruleus lesions, Brain Res. Bull., 3 (1978) 125-130. 7 Bunney, B. S., Walters, J. R., Roth, R. H. and Aghajanian, G. K., Dopaminergic neurons: effect of antipsychotic drugs and amphetamine on single cell activity, J. Pharmacol. exp. Ther., 185 (1973) 560-571. 8 Cabot, J. B. and Cohen, D. H., Avian sympathetic cardiac fibers and their cells of origin: anatomical and electrophysiological characteristics, Brain Research, 131 (1977) 73-87. 9 Carlsson, A., Falck, B., Fuxe, K. and Hillarp, N. A. Cellular localization of monoamines in the spinal cord, Acta physiol, scand., 60 (1964), 112-119. 10 Cedarbaum, J. M. and Aghajanian, G. K., Catecholamine receptors on locus coeruleus neurons: pharmacological characterization, Europ. J. Pharmaeol., 44 (1977) 375-385. 1 l Cedarbaum, J. M. and Aghajanian, G. K., Activation of locus coeruleus neurons by peripheral stimuli: modulation by a collateral inhibitory mechanism, Life Sci., 23 (1978) 1383-1392. 12 Cedarbaum, J. M. and Aghajanian, G. K., Afferent projections to the rat locus coeruleus as determined by a retrograde tracing technique, J. comp. Neurol., 178 (1978), 1-16. 13 Commissiong, J. W., Hellstrom, S. O. and Neff, N. H., A new projection from locus coeruleus to the spinal ventral columns: histochemical and biochemical evidence, Brain Research, 148 (1978) 207-213. 14 Darian-Smith, I., Phillips, G. and Ryan, R. D., Functional organization in the trigeminal main sensory and rostral spinal nuclei of the cat, J. Physiol. (Lond.), 168 (1963) 129-146. 15 Donaldson, I. MacG., Dolphin, A., Jenner, P., Mardsen, C. D. and Pycock, C., The involvement of noradrenaline in motor activity as shown by retational behavior after unilateral lesions of the locus coeruleus, Brain, 99 (1976) 427-266. 16 Faiers, A. A. and Mogenson, G. J., Electrophysiological identification of neurons in locus coeruleus, Exp. Neurol., 53 (1976) 254-266. 17 Fields, H. L. and Anderson, S. D., Evidence that raphe-spinal neurons mediate o!ciate and midbrain stimulation produced analgesias, Pain, 5 (1978) 333-349. 18 Fuller, J. H. and Schlag, J. D., Determination of antidromic excitation by the collision test: problems of interpretation, Brain Research, 112 (1976) 283-298. 19 Grossman, W., Burna, I. and Nell, T., The effect of reserpine and DOPA on reflex activity in the rat spinal cord, Exp. Brain Res., 22 (1975) 351-361. 20 Guyenet, P. G. and Aghajanian, G. K., Antidromic identification of dopaminergic and other output neurons of the rat substantia nigra, Brain Research, 150 (1978) 69-84. 21 Guyenet, P. G. and Aghajanian, G. K., ACh, substance P and metenkephalin in the locus coeruleus: pharmacological evidence for independent sites of action, Europ. J. Pharmaeol., 53 (1979) 319-328. 22 Hancock, M. B. and Fougerousse, C. L., Spinal projections from the nucleus locus coeruleus and nucleus subcoeruleus in the cat and monkey as demonstrated by the retrograde transport of horseradish peroxidase, Brain Res. Bull., 1 (1976) 229-234. 23 Hanker, J. S., Yates, P. E., Metz, C. B. and Rustioni, A., A new specific, sensitive and noncarcinogenic reagent for the demonstration of horseradish peroxidase, Histoehem. J., 9 (1977} 789-792.

133 24 Huang, Y. H. and Maas, J. W. (1976), Stimulation of the dorsal noradrenergic bundle and field potentials in the locus coeruleus, Brain Research, 115 (1976) 91-94. 25 Jones, B. E. and Moore, R. Y., Ascending projections of the locus coeruleus in the rat. II. Autoradiographic study, Brain Research, 127 (1977) 23-53. 26 Kneisley, L. W., Biber, M. P. and LaVail, J. H., A study of the origin of brain stem projections to monkey spinal cord using the retrograde transport method, Exp. Neurol., 60 (1978) 116-139. 27 Kuypers, H. G. J. M. and Maisky, V. A., Retrograde axonal transport of horseradish peroxidase from spinal cord to brain stem cell groups in the cat, Neurosci. Lett., 1 (1975) 9-14. 28 Maeda, T. and Shimizu, N., Projections ascendantes du locus coeruleus et d'autres neurones aminergiques pontiques an niveau du prosencephale de rat, Brain Research, 36 (1972), 19-35. 29 Mesulam, M. M.,TehamethylbenzidineforHRPhistochemistry:anon-carcinogenicbluereaction product with superior sensitivity for visualizing neural afferents and efferents, J. Histochem. Cytochem., 26 (1978) 106-117. 30 Nakamura, S. and Iwama, K., Antidromic activation of the rat locus coeruleus neurons from hippocampus, cerebral and cerebellar cortices Brain Research, 99 (1975) 372-376. 31 Nygren, L. G. and Olson, L., A new major projection from locus coeruleus: the main source of noradrenergic nerve terminals in the ventral and dorsal columns of the spinal cord, Brain Research, 132 (1977) 85-93. 32 OIson, L. and Fuxe, K., Further mapping out of central noradrenergic neuronal systems: projections of the subcoeruleus area, Brain Research, 43 (1972) 289-295. 33 Pearson, J. A., The coeruleo-spinal pathway: an anatomical, neurochemical and electrophysiological study, Neurosci. Abstr., (1978) 1831. 34 Pickel, V. M., Joh, T. H., Reis, D. J., Leeman, S. E. and Miller, R. J., Electron microscopic localization of substance P and enkephalin in axon terminals related to dendrites of catecholaminergic neurons, Brain Research, 160 (1979) 387-400. 35 Ross, R. A. and Reis, D. J., Effects of lesions of locus coeruleus on regional distribution of dopamine-fl-hydroxylase activity in rat brain, Brain Research, 73 (1974) 161-166. 36 Sakai, K., Touret, M., Salvert, D., Leger, L. and Jouvet, M., Afferent projections to the cat locus coeruleus as visualized by the horseradish peroxidase technique, Brain Research, 119 (1977) 21-41. 37 Sastry, B. S. R. and Sinclair, J. G., Tonic inhibitory influence of a supraspinal monoaminergic system on recurrent inhibition of an extensor monosynaptic reflex, Brain Research, 117 (1976) 69-76. 38 Satoh, K. Tohyama, M., Yamamoto, K., Sakumoto, T. and Shimizu, N., Noradrenaline innervation of the spinal cord studied by the horseradish peroxidase method combined with monoamineoxidase staining, Exp. Brain Res., 30 (1977) 175-186. 39 Segal, M. and Sandberg, D., Analgesia produced by electrical stimulation of catecholamine nuclei in the rat brain, Brain Research, 123 (1977) 369-372. 40 Shimizu, N. and lmamoto, K., Fine structure of the locus coeruleus in the rat, Arch. Histol. Jap., 31 (1970) 229-246. 41 Slater, P. and Blundell, C., The effects of a permanent and selective depletion of brain catecholamines on the antinociceptive action of morphine, Naunyn's Schmiedeberg's Arch. exp. Path. Pharmak., 305 (1978) 227-232. 42 Swanson, L. W., The locus coeruleus: a cytoarchitectonic, Golgi and immunohistochemical study in the albino rat, Brain Research, 110 (1976) 39-56. 43 Takigawa, M. and Mogenson, G. J., A study of inputs to antidromically identified neurons of the locus coeruleus, Brain Research, 135 (1977) 217-230. 44 Tohyama, M., Maeda, T. and Shimizu, M., Detailed noradrenaline pathways of locus coeruleus neurons to the celebral cortex with use of 6-hydroxy-DOPA, Brain Research, 79 (1974) 139-144. 45 Wang, R. Y. and Aghajanian, G. K., Inhibition of neurons in the amygdala by dorsal raphe stimulation: mediation through a direct serotoninergic pathway, Brain Research, 120 (1977) 85-102. 46 West, D. C. and Wolstencroft, J. H., Facilitation of raphe spinal and other bulbar raphe neurons by stimulation of the sensorimotor cortex, J. Physiol. (Land.), 277 (1978) 47.