Spinal projections from the periaqueductal grey and dorsal raphe in the rat, cat and monkey

Spinal projections from the periaqueductal grey and dorsal raphe in the rat, cat and monkey

Neuroscience Printed 0306-4522/82/l 12769-08SO3.00/0 Pergamon Press Ltd 0 1982 IBRO Vol. 7, No. 11, pp. 2769 to 2776, 1982 in Great Britain SPINAL...

1MB Sizes 0 Downloads 81 Views

Neuroscience Printed

0306-4522/82/l 12769-08SO3.00/0 Pergamon Press Ltd 0 1982 IBRO

Vol. 7, No. 11, pp. 2769 to 2776, 1982

in Great Britain

SPINAL PROJECTIONS FROM THE PERIAQUEDUCTAL GREY AND DORSAL RAPHE IN THE RAT, CAT AND MONKEY P. W. hkNTYH*t

and M.

PESCHANSKIt$

*Biomedical Research Division, Ames Research Center, NASA, Mogett Field, California 94035, U.S.A.; tDepartment of Anatomy, UCSF Medical School, San Francisco, California 94143, U.S.A.; $UnitC de Recherches de Neurophysiologie Pharmacologique, INSERM U. 161 2 rue d’Altsia. Paris, France Abstract-There is considerable evidence that the periaqueductal grey and the dorsal raphe contribute to an endogenous analgesia system and to the regulation of a wide variety of other responses, many of which involve spinal sites of action. To map the areas of the periaqueductal grey and dorsal raphe which contain neurons that project to the spinal cord, wheat germ agglutinin conjugated to horseradish peroxidase was injected into hemisected spinal cords in rat, cat, and monkey. After cervical or lumbar injections labelled neurons were found in the periaqueductal grey and dorsal raphe in all species examined. In the rat, labelling of the dorsal raphe is sparse but numerous labelled neurons are present in the mid and rostra1 periaqueductal grey. In the cat, the number of retrogradely-labelled neurons in both the dorsal raphe and the periaqueductal grey are considerable. In the monkey, like the rat, the labelling in the dorsal raphe was light but numerous labelled neurons were present in the periaqueductal grey and the adjacent nucleus cuneiformis. Injections into the lumbar spinal cord produced the same pattern of labelling as seen after cervical level injections with approximately 40% fewer labelled cells in all areas. Thus, while each species had a similar pattern of spinal projections from the periaqueductal grey and dorsal raphe, quantitative differences were evident among the species examined. These results suggest that the number of periaqueductal grey and dorsal raphe neurons projecting to the spinal cord in the rat, cat and monkey are considerably more numerous than previously reported and that the effects described during the stimulation of these regions could be, at least partly, due to the involvement of these direct pathways.

Behavioral and physiological studies have shown that stimulation of the midbrain periaqueductal grey (PAG) region can influence a wide variety of visceral responses including: sexual function,” motility of the gut,= tumor formation22 and analgesia.20 Several studies have suggested that the anatomical substrate for this descending control is made by PAG-medullary connections1~2~8 and it is hypothesized that some of these medullary neurons which receive a projection from the PAG in turn project to the spinal cord where they exert their effects.’ Thus the PAG’s influence on the spinal cord is thought to necessarily involve a medullary link. Recent studies, however, have suggested that at least in the rat’ and monkey4 a few PAG-spinal projection neurons exist. In the cat, this PAG-spinal projection was not found to exist2S11 although a small dorsal raphe-spinal projection to the second cervical level has been described in the cat.24 This dorsal raphe-spinal projection has not been observed in the Please address all correspondence to: Dr P. Mantyh, MRC Neurochemical Pharmacology Unit, Medical Research Council Centre, Medical School, Hills Road, Cambridge CB2-2AH, U.K. Abbreviations: PAG, periaqueductal grey; TMB. tetramethyl benzidine; WGA-HRP, wheat germ agglutininhorseradish peroxidase.

rat’*” or monkey4vio It is possible that these reported species differences reflect the sensitivities of the various chromagens employed and we have, therefore, reinvestigated the presence of the PAG and dorsal raphe spinal connections using the tracer wheat germ agglutinin conjugated to horseradish peroxidase (WGA-HRP) and tetramethyl benzidine (TMB) as the chromagen. We have used the retrograde tracer WGA-HRP because it is estimated to be up to forty times as avid as horseradish peroxidase HRP alone in the retrograde direction.’ TMB has been employed because of the superior sensitivity exhibited by this chromagen when compared to previously used chromagens such as diaminobenzidine. l7

EXPERIMENTAL

PROCEDURES

Twenty-six rats, five cats, and five primates (4 squirrel monkeys [snimiri sciureus] and one baboon [papio papio]) were used in this study. The animals were anesthetized with either choral hydrate (400 mg/kg i.p. for rats) or pentobarbital (40mg/kg i.v. for cats and monkeys) and the appropriate cervical or lumbar lamina exposed. For all animals, this exposure was either at C7-C8 or L2-L3. After the exposure was made a laminectomy was performed, the dura reflected and the cord hemisected. After the hemisection a 10% solution of wheat germ agglutinin conjugated horseradish peroxidase (Sigma No. L-9008) to

2769

‘770

P. W. .Mantyh and M. Peschanski

(WGA-HRP) was injected into the spinal cord 1 mm rostral to the hemisection. In the rat 0.02---l~1 was injected; in the cat and monkeys l-5 1’1was injected. By hemisecting the cord and then injecting. we hoped to ensure that all the descending fibers on the hemisected side were exposed to the WCA-HRP tracer. One to five days after the injection of WGA-HRP. the animals were anesthetized and perfused transcardially with 3 0.1 M phosphate-buffered saline solution (pH 7.4) followed by a solution of 0.1 M phosphate-buffered (pH 7.4) fixative. -‘” 11glutaraldehyde and I”, paraformaldehyde After this fixative a 0.1 M phosphate bu&r solution fpH 7.4) containing 5”, sucrose was perfused through the vasculature to remove any excess fixative. The brain and spinal cord were then removed and carried through graded sucrose solutions to cryoprotect the brain. Serial 50pm frozen sections were collected and every other section was reacted for the presence of WGA-HRP using the TMB procedure.‘” Every other reacted section was then counterstained with neutral red. Agi reacted sections were examincd using both dark- and light-tieid microscopy. Each labelled cell was plotted onto a projection of the midbrain that was made with the aid of a overhead projector. Schematic reconstructions were then made in appropriate sections from the atlases of Pellegrino, Pelfegrino & Cushman’* for the Tat. k-man3 for the cat and Emmers & Akert” for the monkey.

RESULTS The results of these experiments are illustrated in Figs 1, 2 and 3 for the rat, cat and monkey, respectively. The dots represent those neurons IabeIled after an ipsilateral hemisection and subsequent injection of WGA-HRP on the right side at C7. The number of dots in each transverse section in the reconstructions represents all the cells observed in one representative 50itm section. In the cases with lumbar levef injections essentially the same pattern of Iabelling were observed but there was approximateiy 407; fewer cells in all labelled regions. It should be emphasized that we shall report. illustrate and discuss only the cells labelled in the dorsal raphe. PAG and nucleus cuneifkjrmis. Large numbers of Iabelled cells in the midbrain and pons *were also observed in the EdingerWestphal. Insterstitial nucleus of Cajal, Darkschewitsch. locus coeruleus and red nucleus but in general our results confirmed previous reports’.4”” 14~24~25 and therefore need not be reported here again.

Rrrr Labelted cells could be observed in the most caudaf aspect of the PAG at the level of the locus coeruleus. At this level small, scattered, fusiform- and stellateshaped neurons could be seen in the PAG. Labelled dorsal raphe neurons are also found, primarily within the pars medialis; only a few were located in the pars later&is. Further rostral, at the ‘ievef of the inferior and superior colliculus labelled neurons ccuid be observed scattered throughout the PAG (Fig. lC), with

most of the labelled PAG neurons being present in the ventral two-thirds of the structure. lin addition to

the scattered cells a group of very well labelled cells stretched in a band-like fashion from the inner onethird of the ventral aspect of the PAG well out into the adjacent nucleus cuneiformis. This band of labelled cells is composed of the moderate-sized fusiform and stellate cells interspersed among large muttipolar cells that were very well tabelled. At the levei of the posterior commissure the band of labelled cells spanning the PAG and the nucleus cuneiformis is no longer present but large numbers of well Iabelled cells of heterogeneous morphology were present in the PAG [Fig. IA). Occasionally, small clusters of morphologically similar neurons could be observed in the ventral aspect of the PAG at this level (Fig. 1B). Scattered labelled neurons continued to be present in the rostra1 periventricular grey region, where they appeared to be continuous with the tabelied c&s in caudaf posterior h~othaIamic region. C’crt The pattern of labelling observed in the cat. compared to that found in the rat, was strikingly similar with onIy quantitatively differences apparent. Many more Iabelled neurons were obserxled in both the pars medialis and pars lateralis subdivisions of the cat dorsal raphe (Fig. 2C). Further rostral, at the intercollucular and superior colliculus levels, the cat has both large and small neurons labelled which form a band of cells which extends from the inner one-third of the PAG lateralty out into the adjacent nucleus cuneiformis (Figs 2A and 3). Further rostra1 at the posterior commissure and periventricular grey levels, the labelling was qualitatively similar to the rat but with only half as many Iabelled neurons.

The pattern of labelling observed in the monkey was similar to both the cat and rat; again quantitative differences were found. The relatively light pattern of labelling in the dorsal raphe was similar to that seen in the rat. At the intercollicular and superior collic&us levels the band like pattern of labelled cells in the PAG and adjacent nucleus cuneiformis (Figs 3B and C) was again present. At the posterior commissure levels the pattern of labelling was similar to the cat where moderate numbers of labelled neurons were observed. No marked differences in the lOCatiOn of Iabeiled neurons were noted between the baboon and the squirrel monkeys. Approximately 802, of the I&belled PAG-spinal neurons were ipsilaterai to the injection site in ail of the animals. The dorsal raphe-spinal projection appears to project bilaterally, although the pars later&is division had slightly fewer cells contralateral to the injection. In examining the differences between cervical and lumbar injections, it appears that similar areas in the PAG and dorsal raphe are labelled after injection at either level although in lumbar injections approximately 40:+, fewer cells were labelled.

Fig. 1. On the left is a reconstruction in the frontal plane of the location of labelled neurons in the periaqueductal grey (PAG) and dorsal horn (DR) of the rat. In all the following schematics the data is presented from one representative animal with hemisection and injection of WGA-HRP at C7 on the right side of the spinal cord. The dots represent the average number of neurons recorded in one 50pm section. It should be emphasized that we have illustrated only those neurons that were labelled in the DR. PAG and nucleus cuneiformis. On the right are photomicrographs from rats with cervical injections of WGA-HRP as described above. (A) Labelling in the rostra1 PAG at the level of the third nucleus. Corresponds to level of section lb. Magnification x 11.5. (B) Labelling in the rostra1 PAG at the third nucleus-posterior commissure level. Note the characteristic cluster of labelled neurons in the rodent PAG at the level of the posterior commissure. Corresponds to level of section la. Magnification x 120. (C) Retrogradely-labelled cells in the ventral PAG at the level of the rostra1 inferior colliculus. Corresponds to level of section le. Magnification x 1 IO. Fig. 2. On the left is a reconstruction of the location of retrogradely-labelled neurons in the periaqueductal grey and dorsal horn of the cat in the frontal plane of section. See Fig. 1 for explanation. On the right are photomicrographs from cat cases with cervical injections of WGA-HRP. (A) A dark-field photomicrograph of retrogradely-labelled neurons in the PAG at the level of the third nucleus and mid superior colliculus. Note that labelled cells extend from the inner one-third of the PAG into the adjacent nucleus cuneiformis. Corresponds to level of section 2c. Magnification x 30. (B) A dark-field photomicrograph showing retrogradely-labelled neurons at the PAG-nucleus cuneiformis border. Note that the labelled neurons appear to form a band of cells that spans the intervening mesencephalic V border. Corresponds to the level of section 2d. Magnification x 100. (C) A dark-field photomicrograph showing labelled neurons in the pars medialis division of the DR. Corresponds to the level of section 2e. Magnification x 85.

Fig. 3. On the left is a reconstruction in the frontal plane of the location of retrogradely-labelled periaqueductal grey and dorsal horn neurons in the squirrel monkey. See Fig. 1 for explanation. On the right are photomicrographs from a baboon (Fig. 3A) and from a squirrel monkey (Figs 38 and C) after a cervical hemisection and injection of WGA-HRP. (A) A light-field photomicrograph of retrogradelylabelled cells in a baboon PAG at mid-superior colliculus level. In the baboon the labelled neurons were located in similar areas of the brain as was found in the squirrel monkey. Corresponds to the level of section 3d. Magnification 40x. (B) A light-field photomicrograph showing labelled neurons at the PAG-nucleus cunieformis area. Note that the labelled neurons appear as an uninterrupted band between the two regions. Corresponds to the level of section 3e. Magnification x 60. (C) A light-held photomicrograph of Fig. 3B at higher magnification. Note that the labelled cells display a variety of shapes and sizes. Corresponds to the level of section 3e. Magnification x 120.

Abhwiations III IV APT BC BIC C CUN DBC DTN IC EW H HIPT L LC

used in figures

third nucleus fourth nucleus pretectal area brachium conjunctivum brachium of the inferior colliculus caudate nucleus nucleus cuneiformis decussation of the brachium conjunctivum dorsal tegmental nucleus inferior colliculus nucleus of Edinger-Westphal habenula habenulo-interpeduncular tract nucleus linnearis locus coeruleus

LGB MGB MD MLF NST PAG PC PULV PVG RN SC SM SN VPL VTN

271

I

lateral geniculate body medial geniculate body nucleus medialis dorsalis thalami medial longitudinal fasciculus nucleus trigemini sensorius periaqueductal grey posterior commissure pulvinar periventricular grey red nucleus superior colliculus stria medullaris substantia nigra ventral posterior lateral nucleus thalami ventral tegmental nucleus

Fig. 2712

1

Fig. 2 2773

Spinal projections from the periaqueductal grey and dorsal raphe

DISCUSSION These results emphasize that the PAG, generally thought to have a sjgnifi~ant projection only as far caudal as the medulla, has a considerable projection to the lumbar spinal cord in the rat, cat, and monkey. Also interesting is the projection from the dorsal raphe to the spinal cord in all species examined here. Thus, in the three species examined both the PAGspinal and dorsal raphe-spinal projections appear to be phylogenetically stable with only quantitative differences. Previous reports have suggested that few if any neurons in the PAG and dorsal raphe project to the spinal cord in the rat,‘z4 cat2s11*24and monkey.‘*” In contrast the present report has found that significant numbers of PAG and dorsal raphe neurons project to the spinal cord in al1 three species. In the previous studies, a variety of relatively insensitive chromagens were employed to detect the retrogradely transported HRP, whereas in the present report we have uniformly used a relatively more sensitive tracer and chromagen in all three species examined. Thus, the differences between previous reports and the present one appear to be due to the more sensitive tracer and chromagen employed in this study that were not available when the previous studies were perfarmed. After spinal injections, the one area most consistently and heavily labeiled was the lateral aspect of the PAG and the adjacent tegmentum (nucleus cuneiformis). Previously, we have noted that these two adjacent regions appear to have a similar cytology,i5 as revealed by the Golgi technique and on this basis, have suggested that these two regions may be functionally related. In the present study, it appears that these neurons which have similar morphology also have a similar projection to the spinal cord. These results further suggest that these two adjacent areas may participate in a similar function. The fact that the PAG does have a significant direct projection to the spinal cord provides an additional pathway by which the PAG can exert control over the

2775

spinal cord. In relation to central grey stimulationproduced analgesia and opiate analgesia, this hypothesis is consistent with the fact that even when both serotonin and noradrenergic antagonists are applied to the cord, thus blocking many of the medullaryspinal interactions, a component of central grey opiate analgesia can still be generated in the rat.25 Since the rat PAG contains neither noradrenergic nor serotoninergic somas5 the direct PAG-spinal projections could provide part of the remaining descending inhibition. Furthermore in cat, which appears to have the heaviest dorsal raphe-spinal projection among the species examined here, part of midbrain stimulationproduced analgesia could be mediated by serotonin which is known to be concentrated in some dorsal raphe neurons.’ Previous reports in the cat have suggested that, unlike the rat and monkey, it is the dorsal raphe and not the PAG that is the most eflicacious midbrain site for generating stimulation-pr~~Iced analgesia. Whether it is the relatively strong dorsal raphe-spinal projection in the cat, compared to the sparse dorsal raphe-spinal projection seen in the monkey and rat, that underlies these previous reported differences is still an open question. In conclusion these studies have shown that in the rat, cat and monkey cytologically similar cells with a comparable distribution in the PAG and dorsal raphe project to the spinal cord. Since the PAG and dorsal raphe are known to be involved in a variety of visceral activities including regulation of the gut, sexual behavior and an endogenous analgesia, it is suggested that these PAG-spinal and dorsal raphe-spinal projecting neurons be placed alongside the PAG-medullary-spinal connections in regulating these responses. Ackno~v[~~gentents-We would like to thank Dr H. J. Ralston III for providing some of the support needed for the

study, Drs Basbaum and Mehler for invaluable suggestions and criticisms and to D. Akers for excellent photographic assistance. P.W.M. is a National Research Council Fellow. (Supported by NASA task 199-05-02-07 to W. R. Mehler and NS-11614 to H. J. Ralston.)

REFERENCES 1, Abols 1. A. & Basbaum A. I. (1981) Afferent connections of the rostra1 medulla of the cat: a neural substrate for mjdbrain-meduflary interactions in the modulation of pain. J. camp. !?euror’.201, 285-299. 2. Basbaum A. I. & Fields H. L. (1979) The origin of descending pathways in the dorsolateral funiculus of the spinal cord of the cat and the rat: further studies on the anatomy of pain modulation. J. camp. Neural. 187, 513-531. 3. Berman A. L. (1968) The Brainstem of the Cat, Cytoarchitectonic Atlas with Stereotaxic Coordinates University of Wisconsin Press, Madison. 4. Castiglioni A. J., Gallaway M. C. & Couher J. D. (1978) Spinal projections from the midbrain in monkey. J. romp. Neural. 178, 329-346. 5. Dahlstrom A. & Fuxe K. (1964) Evidence for the existence of monoamine-containing neurons in the central nervous system--L Demonstration of monoamines in the cell bodies of brainstem neurons. Acta physiol. stand. 62, Suppt 232, l-S5 6.

Emmers R. 52 Akert R. (1963)A Stereotaxic At& oj’the Bruin of&e Squirrel Monkey (Saimiri sciureus) University of

Wisconsin Press, Madison. 7. Fieids H. L. & Basbaum I. A. (1978) Brainstem control of spinal pain transmission neurons. A. Rev. Physiol. 40, 193-221.

7116

P. W. Mantyh and M. Peschanski

x. Gallagher D. W. & Pert A. (1978) Afferents to brainstem nuclei (brainstem raphe n., n. reticularis pontis caudalis, and n. gigantocellularis) in the rat as demonstrated by microiontophoretically-applied horseradish peroxidase. Brain Res. 144, 257 -275. 9. Gonatas N. K.. Harper C.. Mizutani T. & Gonatas J. 0. (1979) Superior sensitivity of conjugates of horseradish peroxidase with wheat germ agglutinin for studies of retrograde axonal transport. J. ~~sr[~~~~t~.C~~~~~?~~~I. 27, 728 -734. 10. Kneisley L. S., Biber M. P. & LaVail J. II. (1978) A study of the origin of brainstem projections to monkey spinal cord using the retrograde transport method. Erpl Neural. 60, 116139. 11. Kuypers H. G. J. M. & Maisky V. A. (1975) Retrograde axonal transport of horseradish peroxidase from spinal cord to brainstem cell groups in the cat. Neuroscience Letters I, 9-14. 12. Leichnetz G. R., Watkins L., Griffin G.. Murfin R. & Mayer D. J. (1978) The projections from nucleus raphe magnus and other brainstem nuclei to the spinal cord in the rat: a study using the HRP blue reaction. ~e~{~~,s~~e~~~ Letrers 8, If9 124. 13. Loewy A. D. & Saper C. B. (1978) Edinger-Westphal nucleus: projections to the brainstem and spinal cord in the cat. Bwirz Rrs. 150, I- 27. 14. Loewy A. D., Saper C. B. & Yamondis N. D. (1978) Re-evaluation of the efferent projections of the Edinger-Westphal nucleus in the cat. Brain Rex 141, 1533159. 15. Mantyh P. W. (1982) The midbrain periaqueductal grey of the rat, cat and monkey. A Nissl, Weil and Golgi analysis. J. (onip. .lintroi. 204, 349.-363. 16. Mesulam M. M. (1978) Tetramethyl benzidine for horseradish peroxidase neurohistochemistry : a noncarcinogenic blue reaction-product with superior sensitivity for visualizing neural afferents and efferents. J. Histockem. Cytorhem. 26, 106117. 17. Mesulam M. M. & Rosene D. L. (1979) Sensitivity in horseradish peroxidase neurohistochemistry: a comparative and quantitative analysis of nine methods. J. Histochem. Cytochem. 27, 763-773. 18. Olivcras J. L., Besson J. M.. Guilbaud G. & Liebeskind J. 0. (1974) Behavioral and electrophysiological evidence of pain inhibition from midbrain stimulation in the cat. Expl Brain Res. 20, 3244. 19. Peflegrino L. J,. Pellegrino A. S. & Cushman A. J. (1979) A Srereotaxic Arlus o$the Ruf Brain Plenum Press, New York. 20. Reynolds D. V. (1969) Surgery in the rat during electrical analgesia induced by focal brain stimulation. Science, N.Y. 164, 444445. 21. Sakuma Y. & Pfaff D. W. (1979) Mesencephalic mechanisms for integration of female reproduction behavior in the rat. AM. J. Physiol. 237, R2855290. 22. Simon R. H.. Lovett E. J.. Tomaszek D. & Lundy J. (1980) Electrical stimulation of the midbrain mediates metastdtic tumor growth. Science. N.Y. 209, 1132 1133. 23 Skultety F. M. (1959) Relation of periaqueductal grey matter to stomach and bladder motility. Neurology 9, 190.~197. 24 Tohyama M., Sakai K.. Touret M., Salvert D. & Jouvet M. (1979) Spinal projections from the lower brainstem in the cat as demonstrated by the horseradish peroxidase technique- -11. Projections from the dorsolateral pontine tegmenturn and raphe nuclei. Brain Rex 176, 215-231. 25 Toyoshima K.. Kawana E. & Sakai H. (1980) On the neuronal origin of the afferents to the ciliary ganglion in the cat. Bruin Res. 185, 67.. lb. 25 Yaksh T. L. (1978) Direct evidence that spinal serotonin and noradrenaline terminals mediate spinal antinociceptive effects of morphine in the periaqueductal grey. Bruin Res. 160, 180-185. (Accepted

6 July 1982)