Development of immunoreactivity for calcitonin gene-related peptide, substance p and glutamate in primary sensory neurons, and for serotonin in the spinal cord of fetal sheep

Neuroscience Vol. 54,No. I, Printed in Great Britain

pp. 239-252, 1993

0306-4522/93 $6.00+0.00

Pergamon Press Ltd 0 1993IBRO

DEVELOPMENT OF IMMUNOREACTIVITY FOR CALCITONIN GENE-RELATED PEPTIDE, SUBSTANCE P AND GLUTAMATE IN PRIMARY SENSORY NEURONS, AND FOR SEROTONIN IN THE SPINAL CORD OF FETAL SHEEP I. NITMS and S. I&ES* Department of Physiology, Monash University, Clayton, Victoria 3168, Australia Abstract-In this study we have described the ontogeny of immunoreactivity for calcitonin gene-related peptide, substance P and glutamate in primary sensory neurons, and for serotonin in the sacral spinal cord, of fetal sheep (n = 37) from 56 to 140 days of gestation (term = 146 days). A few fine, varicose fibres immunoreactive for calcitonin gene-related peptide were present in Lissauer’s tract, the dorsolateral funiculus and in laminae I and V in the dorsal horn of the spinal cord at 5661 days of gestation. At this age, two groups of intensely staining immunoreactive cells were present in the motoneuron pool in laminae VIII and IX in the ventral horn of the spinal cord. By 77 days, immunoreactive fibres were also present in laminae II and X. With advancing gestational age, an increase in the intensity of staining was observed throughout the cord to term, with the exception of laminae VIII and IX, where a decrease was seen. Intense staining of cells in the motoneuron pool was evident until c. 128 days, after which time staining became very faint. Fine fibres immunoreactive for substance P were present in Lissauer’s tract and lamina I of the spinal cord at 5661 days of gestation. They were also present throughout laminae IV-VI and X as well as throughout the entire ventral horn. Immunoreactive fibres in lamina II were evident by 77 days. The staining increased in density but remained similar in distribution with increasing gestational age to term in the dorsal horn, but decreased markedly in the ventral horn. Cells immunoreactive for substance P were evident from 56 days, particularly on the border of laminae II and III, until late in gestation. Ultrastructural studies showed that axon terminals immunoreactive for calcitonin gene-related peptide and for substance P were present in lamina I by 61 days. Immunoreactivity for glutamate was evident at 83 days in dorsal root fibres and also in lamina I and II, where it was more prominent in cells than in fibres. At all ages examined, the dorsal horn stained more intensely than the ventral horn. Immunoreactivity for glutamate and neuropeptides appeared in the cells and fibres of dorsal root ganglia at 97-100 days. In the skin, immunoreactivity for calcitonin gene-related peptide and substance P was present at 85days, some time after its appearance in the cord. Fibres immunoreactive for serotonin appeared in lamina I, at the neck of the dorsal horn and in the ventral horn at 83 days of gestation. Few immunoreactive fibres were seen at any stage during gestation in laminae II and III. Staining increased in laminae IV-X during gestation, but there was eventually a slight decrease in the ventral horn prior to term. These studies have shown that neurochemicals thought to be involved in the transmission and modification of somatosenory information in the spinal cord are present in fetal sheep from early in gestation, with neuropeptides appearing before glutamate and serotonin.

and functional apparatus necessary for the ability to perceive somatosensory information during fetal life. Recent work on fetal rats in vivo has shown that dorsal horn cells in the spinal cord first respond to electrical stimulation of the skin of the hindlimb in late gestation at embryonic day (E) 17, although responses to natural cutaneous stimulation are not evoked until E19.14 The neuropeptides calcitonin gene-related peptide (CGRP) and substance P (SP), which are contained in small diameter A delta and C fibre afferents37,38and thought to be concerned with neurotransmission or modulation of cutaneous input from the skin to the spinal cord,‘3,42are first detected in the region of termination of these fibres in the dorsal horn (laminae I, II and V) at El&18.35.48 It is not yet certain how the timing of the functional and neurochemical development of sensory pathways

There is now considerable interest in understanding the events which occur prenatally in the development of synaptic connections in sensory systems and the factors which might influence this development. In this study, it has been of particular interest to us to determine when the cutaneous pathways from the skin to the spinal cord and brain develop in order to determine when a fetus has the minimum structural *To whom correspondence should be addressed. Present address: Department ofAnatomy and Cell Biology, University of Melbourne, Parkville, Victoria 3052, Australia. Abbreviations : CGRP, calcitonin gene-related peptide; CGRP-LI. CGRP-like immunoreactivitv: DRG. dorsal root ganglion; E, embryonic day; GLU, glutamate; GLU-LI, GLU-like immunoreactivity; 5-HT, serotonin; 5-HT-LI, 5-HT-like immunoreactivity; SP, substance P; SP-LI, SP-like immunoreactivity. 239

240

I NI rsos and S. Rtts

in the rat, a short gestation species (term = 22 days) relates to the events in longer gestation specks. including man. Therefore. we have studied these aspects of development in the cutaneous afferent pathways from the hindlimb to the spinal cord in fetal sheep from mid-gestation to just prior to term (term = 146 days). The long gestation period of the sheep also offers the advantage of a clearer temporal resolution of developmental events than is found in shorter gestation species. In this paper we report the results of the neurochemical study: the functional results will be reported separately. Jmmunocytochemical techniques have been used at the light and ultrastructural levels to study the ontogeny of immunoreactivity for CGRP and SP fol the reasons stated above, and for glutamate (GLU), since there are several lines of evidence to support its role as a neurotransmitter for at least some of the primary afferent fibers.‘“.47 The processing of sensory information, particularly from nociceptors, can be modified in the dorsal horn by activity in the descending fibres from the brainstem.‘” Serotonin (5-HT) is thought to be one of the neurotransmitters in this pathway,2,4h so we have also studied the time course of the appearance of immunoreactivity for 5-HT. This study therefore aims to provide new information on the ontogeny of putative neurotransmitters and/or neuromodulators involved in the transmission and modification of somatosensory information in the fetal sheep spinal cord. EXPERIMENTAL

PROCEDURES

Tissues used in this study were obtained from fetal sheep (Border Leicester cross, n = 37) aged between 56 and 140 days of gestation (term = 146 days) at the conclusion of electrophysioiogical experiments performed to examine the functional development of cutaneous afferent pathways. Both the ewe and fetus were given an overdose of pentobarbitone (130 mg/kg) and the fetus was then perfused via the descending aorta with physiological saline followed by 4% paraformaldehyde for light microscopy and 4% paraformaldehyde and 0.3% glutaraldehyde for electron microscopy, in 0.1 M phosphate buffer ([pH 7.4). The first sacral segment of the spinal cord and associated dorsal root ganglia (DRG) and skin from the region above the right hind hoof were removed and placed in 4% paraformaldehyde in buffer overnight. Spinal cord sections for electron microscopy were left overnight in 4% paraformaldehyde and 0.3% glutaraldehyde in 0. I M phosphate buffer. Tissue was then placed in 20% sucrose in 0.1 M phosphate buffer at 4‘;C for 2-3 h prior to sectioning for

Table

1. Summary

of gestational

CGRP Spinal

Light micro.wop~ Free $outing immunohistocllemistr~. Frozen sections (50 rm) of the spinal cord (transverse, S, segment), DRG (longitudinal, S, segment) and skin (150 pm. transverse) were cut and collected in buffer. Sections of spinal cord were reacted for SP-like immunoreactivity (SP-LI), CGRP-LI. GLU-LI and 5-HT-LT; of DRG for SP-LI, CGRP-LI and GLU-LI; and skin for SP-LI and CGRP-LI. The avidin-biotin peroxidase complex (Vector laboratories) was the detection mechanism used in this study. Antisera to SP (monoclonal rat) were purchased from Sera-lab; CGRP (polyclonal rabbit) was purchased from Amersham; s-H7 (polyclonal rabbit) was purchased from Eugene Tech and GLU (polyclonal mouse) was kindly donated by Dr Al Beitz, University of Minnesota, U.S.A. The production and characterization of this antisera, which was raised against glutaraldehyde-fixed glutamate, has been described by McDonald c/ 01.~~ Sections were first placed in O..?‘% H,02 in methanol for 20 min to block for endogenous peroxidase activity. Free-floating sections were then placed m IO%, normal goat serum in 0.1 M phosphate buffer (pH 7.4) for 30 min. The dilutions used for the primary antisera were: SP (1: 500), CGRP (1 : 2OOO), GLU (I : 10,000) and 5-HT (I : 1000). The primary antisera were prepared using 0.1 M phosphate buffer/2% normal goat serum;O.3”/o Triton X-100 as diluent. The sections were placed in the appropriate primary antisera and left overnight at 4 C on a vibrating platform. Sections were then incubated for 45 min in biotinylated secondary antibodies diluted to I :400 in 0.1 M phosphate buffer/2% normal goat serum: antirabbit immunoglobulin for CGRP and 5-HT. anti-rat immunoglohulin for SP and anti-mouse immunoglobulin for GLU. After the incubation, the sections were placed in a I : I : 200 dilution of avidin- biotin peroxidase complex for a further 45 min. Sections were then placed in 0.05% diaminobenzadine solution for IOmin, after which tmlc 30 p L of stock (30%) H,Oz was added for c. 6 min. Sectlons were mounted, dried overnight and then the peroxidasc reaction was intensified by immersing the slides in O.Ol”C 0~0, m 0. I M phosphate bufTer for 30 s. The sections were then dehydrated, cleared. coverslipped and viewed with both light-and dark-field illumination. Table I 15 a summary of the number of animals and gestational ages examined for each antibody in the spinal cord. DRG and skin. Control rqwrimmts. Control experiments included the omission of the primary antisera, and its replacement with either 0.1 M phosphate buffer or non-immune serum. Preabsorption controls with an excess of the appropriate antigen were also carried out for SP (200 pg/ml) and CGRP (20pg/ml) overnight at 4°C. In both of these procedures immunoreactive staining failed to occur.

ages examined

SP

for specific antibodies Glutamate

56, 61, 68, 77,* 89, 97. 100, llO,* 120,*, 128. 140, 28 PN, adult

56, 61, 68, 77,t 100, 110,* 114,t 120,t 134, 140,* 28 PN, adult

76, 83, 92, 97,* ilO,* 115, 128, 140, 28 PN, adult

DRG

83, 97, 100, 110,’ 128, 140

80, 100, I IO.* 114.t 120,* 140*

83, 92, 97,* 1 IO. 115. 128. 140

Skin

76, 85. 97. 100, 110, 120, 128, 140

Unless

cord

immunohistochemical stammg at the light mIcroscope /c\cl In addition, tissue from one lamb (28 days postnatal) and one adult sheep was also collected and prepared ah described above. The purpose of including these animal5 wa:, 10 determine whether or not the levels of neurochemicals altered significantly after birth.

otherwise

stated

one animal

120,*

5-HT 83, 97,* lo&* 140, 28 PN

76, 80, 85. 100, 106, 110,t 120.t 134. 140 was examined

at each of these ages. *n = 2: in = 3; PN, postnatal

day

115, 128.

Ontogeny of neurochemicals in primary sensory neurons Electron microscopy Pre-embeddtig ~m~~h~to~hem~try. Sections of spinal cord at 61 (n = 1) and 140 days (n = 1) of gestation were cut at 5Opm on an Oxford Vibratome and were then washed for 30 min in 1% sodium borohydride in distilled water. This was followed by a 2 h wash in 0.1 M phosuhate buffer (DH 7.4) with chantzea of buffer everv 20 min. Sections were-tben t&sferred tcYa vial containingc. 0.7 ml of 10% sucrose in 0.1 M phosphate buffer. The vial was then dipped in liqnid nitrogen and allowed to thaw at room temperature.. This procedure was repeated two or three times to enhance the penetration of immunoreagents. The sucrose was then removed by three 15 min washes in 0.1 M phosphate buffer prior to continuing with the immunohistochemicai procedure for SP or CGRP immunoreactivity described for the light microscope level with the following exceptions: (i) no Triton X-100 was used; (ii) no blocking for ~dogenous peroxidase activity was carried out. At the compietion of the immunohistochemicai procedure described above, sections of spinal cord were washed for 45 min in 0.1 M phosphate buffer and then placed in 1% OSO, in 0.1 M phosphate buffer for 1.5 h. After the incubation, sections were washed in buffer, dehydrated in increasing concentrations of alcohol and then placed in propylene oxide. Spinal cord sections were then placed in a 1: 1 mixture of Araidite and propylene oxide overnight. The sections were embedded in Araidite placed between two glass sides and the resin was then cured in a WC oven for 24 h. Medial and lateral areas of dorsal horn were then cut and secured onto Araidite blocks. Semithin sections of dorsal horn were cut on a Reichert-Jung Uitracut microtome to determine when tissue containing immunor~ctive product was visible. Ultrathin sections of silver interference were then cut and colfected on 200-mesh grids, stained with uranyi acetate and lead citrate, and viewed with a Joel 100 s electron microscope. Morphometric analysis Light microscopy. At 110, 120 (128 for GLU) and 140 days of gestation, transverse sections of DRG immunoreactive for SP, CGRP and GLU were selected for quantitative analysis. Two animals were examined at each age for SP and CGRP and one for GLU. From each animal, two sections were assessed-one from the centre of the ganglion and one from the periphery. The following measurements were made with a Zeiss MOP-I image anaiyser on each section: (1) The area of ail immunorea~tive *iis was measured in a cross-section (300-400 ceils) and their mean diameter calculated. (2) The number of ceils immunoreactive for each substance was counted and expressed as a percentage of the total number of ceils per section. Results are presented as means & S.E.M. BeWon mjcroscop~.Synaptic terminals immuno~ctive for SP or CGRP at 61 and 140 days were counted in iamina I of the dorsal horn. For each animal, three serial sections of spinal cord were collected at c. 1.5pm intervals and counts were made in an area of 1.1 x 10”flm* (area of a 200.mesh grid square) in a randomly selected lateral and medial region of the iamina per section. Only those iabeiied terminals where it was possible to clearly resolve the pre- and postsynaptic membranes and a synaptic cleft were counted. A mean estimate of the iabeiied terminals (SP or CGRP) was obtained for each animal and the figure converted to terminai!/mm.2 Statistics An analysis of variance was performed on quantitative data in order to determine if the described change in cell diameter and percentage of ceils in the DRG immunoreative

241

for SP and CGRP increased significantly with increasing gestational age.

RESULTS Light microscopy Calcitonin gene-related peptic&: Spinal cord. At 56 days of gestation, fine immunoreactive fibres were

present in the tract of Lissauer and in the dorsolateral funiculus (Fig. 1A). There was also evidence of a bundle of varicose fibres running completely around the edge of the dorsal horn in lamina I. By 61 days a few fibres were seen to emerge from this bundle and course through the dorsal horn to lamina V, where they branched to form a fme reticulated plexus in the lateral part of the layer. At this age, immunoreactive fibres were absent from all other regions of the spinal cord. In laminae VIII and IX of the ventral horn, however, there were two groups of large, intensely immunoreactive cells in the motoneuron pool. By 77 days, there was evidence of immunoreactive varicose fibres and terminals extending down into the central region of lamina i1, predominantly the outer layer (lamina II,). By 97-100 days, the staining in lamina V had become more pronounced and immunoreactive fibres now extended around the medial edge of Iamina I to terminate in lamina X. From 120 days to term, the intensity of staining in all laminae increased, most noticeably in lamina II, where it now extended around into the lateral and medial regions of the full depth of this layer. In the lamb, 28 days postnatal, there was a slight increase in the density of terminals in al1 the laminae and this then remained constant into adulthood. Intensely staining neurons were largely concentrated in the lateral motoneuron pools up to 128 days (Fig. iB), after which time the staining gradually became very faint. In the postnatal lamb and the adult, only one or two neurons in the motoneuron pool were immunoreactive at this level of the spinal cord. It is of interest to note that between 110 and 128 days a small, discrete nucleus of medium-sized neurons on either sidC of the central canal, in laminae X, was immunoreactive. Figure 2 is a diagrammatic representation of the distribution of CGRP-LI in the spinal cord from 56 to 140 days. Dorsal root ganglia. CeH bodies and fibres in the DRG failed to stain at 83 days, but by 97 days they displayed faint immunoreactivity for CGRP (Fig, IC). The intensity of staining and the number of immunoreactive c&s appeared to increase with gestational age up to term and this was confirmed by quantitative analysis. Statistical analysis (compa~son of two proportions) of the percentage of cells immunoreactive for CGRP showed that there was no significant difference (P>O.S) between central and peripheral sections of the DRG, indicating that there was a random distribution of these &Is throughout the ganglion. Data from both sections were combined and a mean value per cross-section calcutated for

A

Ontogeny of neurochemicals in primary sensory neurons

each animal. At 110 days, 27 f 1% of ganglion cells showed CGRP-LI and this increased to 31 + 2% at 120 days (Fig. 1D) and to 37 f 2% at 140 days. At 110 days, the mean diameter of immunoreactive cells was 22.0 f 0.7 pm, which increased to 26.1 f 0.9 pm at 120 days and 30.6 f 0.2 pm at 140 days. Analysis of variance showed significant differences among the percentage of immunoreactive cells (P ~0.01) and mean diameters (P
243

the ventral horn, Cells immunoreactive for SP were present at 56 days in lamina I, along the border of laminae II and III, and laterally in laminae V-VII (Fig. 3A). By 77 days, these cells on the border of laminae II and III had formed into a distinct, strongly staining band. The intensity of staining increased in lamina I and fine staining now extended throughout lamina II, particularly in the central region. Staining remained intense in laminae IV-VII and X, with profuse but less dense staining in laminae VIII and IX. With increasing gestational age from 100 days to term, there was a steady increase in the density of reactive terminals and fibres in laminae I, II and X (Fib. 3B). Concomitan~y, there was a decrease in the density of reactive fibres in the ventral horn in laminae VI-IX. This trend continued in the postnatal lamb and then remained constant into adulthood. The staining of immunoreactive cells in laminae V-VII was no longer evident after 77 days, while cells in all other laminae were noticeably less intense by 100 days and disappeared completely after 120 days. The dist~bution of SP-LI in the spinal cord from 56 to 140 days is shown diagrammatically in Fig. 2. Dorsal root ganglia. In the DRG, SP-LI could not be detected at 80 days of gestation, but was observed in small cell bodies and fibres at 100 days (Fig. 3C). The number of cell bodies exhibiting SP-LI appeared to increase with gestational age and this was confirmed by counting the number of SP-LI cells at various gestational ages. As for CGRP, immunoreactive cells were distributed‘ throughout the DRG and were not restricted to any particular part of the ganglia. As there was no significant difference (P >0.5) in the percentage of immunoreactive cells per cross-section between the two areas of ganglia (central and peripheral) used in the morphometric analysis, the results obtained from these two areas were combined to obtain the following results. At 1IO days of gestation, 21 + 1% of all DRG cells stained for SP, and this increased to 25 + 1% at 120 days and 29 i 1% at 140 days. At 110 days of gestation, the mean diameter of immunoreactive ganglion cells was 22.4 f 0.5 pm, at 120 days 23.9 + 0.8 pm and at 140 days 29.8 &-1.1 pm. Analysis of variance shows significant differences among the percentage of immunoreactive cells (PcO.01) and mean diameters (P <0.05) at these three developmental ages.

Fig. 1. (A) At 56 days of gestation, CGRP-LI fibres in the sacral (Sl) spinal cord are seen in the tract of Lissauer (arrow), the dorsofateral funiculus and lamina I. Medial and lateral motoneuron pools in the ventral horn are strongly immunoreactive. (3) At 128 days, immunoreactive fibres have extended into laminae II, V and X and all laminae stain intensely. Intensely staining neurons are seen in the lateral motoneuron pool. Scale bar = 400 pm for A and B. (C) In the DRG, cell bodies and fibres (arrows) show CGRP-LI at 97 days. (D) At 120 days, their number has increased. Scale bar = 100 pm for C and D. (E) Wool follicle (W) innervation (dark-field illumination) was seen at 100 days with single varicose CGRP-LI fibres (NF, arrows) innervating wool follicles. (F) At 120 days, the innervation was more complex. Scale bar = 50 ym for E and F. (G) Simple CGRP-LI free nerve endings (dark-field illumination) were seen at 110 days. (H) At 128 days, the te~ination of these fibres has become more complex. Scale bar = 100 pm for G and H. E, epidermis.

I. NITSOSand S. REES

244 SP-Ll

CGRP-LI

77 D

12OD

2140

D

1400

Fig. 2. A diagrammatic representation of the distribution of SF-L1 and CGRP-LI in the fetal sheep sacraf (SI) spinal cord between 56 and 140 days of gestation. The major areas of immunoreactivity in the dorsal and ventral horns are represented by circles, the larger circles representing more intense staining than the smaller ones. Immunoreactive cells within the spinal cord are represented by triangles, with the intensely staining cells being represented by tilled triangles, while the open triangles represent cells staining less intensely. Skin. At 85 days of gestation, faintly staining varicose fibres exhibiting SP-LI were first observed growing across the dermis towards the epidermis. Large granular cells, similar to those which stained positively for CGRP, also stained for SP and were observed in close proximity to the growing fibres. At 100 days, fibres could clearly be observed growing along blood vesseh before branching off to innervate

wool follicles (Fig. 3E). The follicle innervation became more extensive with increasing gestational age, as illustrated in Fig. 3F for 134 days. Simple, free nerve endings exhibiting SP-LI were first seen innervating the epidermis at 110 days of gestation. These fibres frequently penetrated to within a few micrometres of the surface of the epidermal layer (Fig. 3G). With increasing gestational age, the nerve endings

Fig. 3. (A) The distribution of SP-LI in the sacral (Sl) spinal cord is present at 56 days of gestation in Lissauer’s tract (arrow), the dorsolateral funiculus, in laminae I, IV-X and in cells in lamina I, on the border of larninae II and III (arrows) and in laminae V-VII. (8) At 134 days, the distribution remains the same but the staining intensity has increased and the cells no longer exhibit SP-LI. Scale bar = 200 pm for A and B. (C) In the DRG, cell bodies and fibres (arrows) exhibited SP-LI at 100 days. (D) At 120 days the number of cell bodies and fibres has increased. Scale bar = 100 pm for C and D. (E) Wool follicle (W) innervation (dark-field illumination) was seen at 100 days with single varicose fibres (NF) inne~ating wool follicles. Large granular SP-LI cells (short arrows) were observed in close proximity to fibres. (F) At 134 days, the innervation of wool follicles has become more complex. Scale bar = 50 pm for E. and F. (G) At 110 days (dark-field illumination), simple free nerve endings innervate the epidermis (E). (H) at 120 days, the termination of these fibres has become more complex. Scale bar = 100 ym for G and H. 245

246

I. NITSOS and S. REES

became more complex, with most fibres branching extensively within the epidermis, as illustrated in Fig. 3H at 120 days. In general, the distribution and density of fibres immunoreactive for SP were similar to those for CGRP at each of the gestational ages examined. Glutamate: Spinal cord. Immunoreactivity for GLU was not observed until 83 days of gestation. At this age, dorsal root fibres exhibited GLU-LI and there was sparse staining of cells in laminae I and

II of the dorsal horn and of a few fibrcs running through these laminae. With increasing gcstational age (97 days), the number of immunoreactive cells in laminae I and II increased to extent shown in Fig. 4A and bundles of immunoreactive fibres were observed coursing through laminae I--III. Axons throughout the fibre tracts in the white matter exhibited GLU-LI as did the nuclei of motoneurons in the ventral horn and fibres which formed a meshwork around them. With increasing gestational age, the intensity of stain-

Fig. 4. (A) In the spinal cord. GLU-LI can be seen in cells and fibres (arrows) in laminae 1 and II of the dorsal horn at 97 days of gestation. Scale bar = IOOpm. (B) At 128 days, GLU-LI was present in both small and large cell bodies (arrows) in the DRG. Scale bar = IOOpm. (C) At 108 days, 5-HT-LI fibres and occasionally cells (dark-field illumination) were found throughout the spinal cord, with intense staining of fibres in laminae III-X. In the ventral horn (vh), a meshwork offibres surrounded motoneurons (arrows). dh, dorsal horn. Scale bar = 250pm. (D) At higher magnification, fibres (thick arrows) and cells (thin arrows) in the dorsal horn can be seen in laminae 1 and II. Scale bar = 200 pm. Note that blocking for endogenous peroxidase activity results in the white matter appearing refractile in dark-field illumination.

Ontogeny of neur~hemi~ls

ing increased slightly and cells showing GLU-LI were now seen throughout the entire cross-section of spinal cord. Overall, however, iaminae I-III consistently stained more intensely for GLU-LI than did the other laminae. At no stage did fibres in the dorsal horn stain as prominently or as intensely for glutamate as they did for the neuropeptides. Dorsal root ganglia. In the DRG, GLU-LI was first observed in small and large cell bodies and in fibres at 97 days of gestation. Again the quantitative analysis showed that there was a similar ~st~bution of immunoreactive cells in a cross-section of DRG from the central and peripheral regions of the ganglion; hence the data were combined. The number of cell bodies exhibiting GLU-LI increased with gestational age from 30 f 3% at 110 days to 46 + 2% at 128 days and 49 f 3% at 140 days. The diameters of these immunoreactive cells ranged from 15-75 pm at 128 days (Fig. 4B) to 2@-90pm at 140 days. At this age (140 days), 10% of immunoreactive cells could be classified as small cells (~30 pm), 40% as medium-sized (30-50pm) and 50% as largesized cells (> 50 pm). The number of dorsal root fibres stoning positively for GLU-LI increased with gestational age. Serotonin: Spinal cord. The first evidence of fibres immunoreactive for serotonin was seen at 83 days of gestation. One or two fragments of varicose fibres were observed in the dorsal columns and in each of laminae I, V and VII-X. At 97 days, there was a marked increase in the number of positive fibres in all of the above mentioned laminae except lamina I. Staining now extended into lamina IV but there were still only a few fibres in laminae II and III. By 108 days, the staining in laminae IV-X had further increased in intensity and there was a profuse meshwork of fibres around motoneurons in laminae VIII and IX (Fig. 4C). A few immunoreactive cells were observed in laminae I and II (Fig. 4D). Staining in laminae II and III remained sparse, although it did increase marginally in density from 128 days to term. Over this period there was a slight decrease in the staining in the ventral horn. The distribution of immunoreactive fibres was similar in the postnatal lamb with further increases in intensity observed in laminae I and V. Electron microscopy ~ubs#u~ce P. At 61 days of gestation, round to oblong-shah axon terminals containing small spherical vesicles immunoreactive for SP were observed in lamina I of the dorsal horn predominantly synapsing on dendrites (Fig. SA), although occasionally axosomatic synapses were seen. Initially, axon terminals were c. 0.6 pm in their longest diameter (Fig. SA), increasing to c. 2.2pm by 140 days (Fig. SB). By this stage, terminals made synaptic contacts with several dendrites. Quantitative analysis showed that 76 + 15 immunoreactive terminals/mm’ were present at 61 days, and this increased to

in primary sensory neurons

1614 f 102 immunoreactive days (Fig. 53).

247

te~inals/mm2

at 140

Calcitonin gene-related peptide. Round to oblongshaped axon terminals containing small spherical vesicles immunoreactive for CGRP were observed in lamina I of the spinal cord at 61 days of gestation (Fig. SC). At this age, terminals were c. 0.7pm in their longest diameter, increasing to 2.1 pm at 140 days. Terminals also contained large immunoreactive granular vesicles in addition to the small spherical vesicles at this age. At 61 days, there were 121 &-19 immunoreactive te~i~Is~~~ and this increased to 1280 5 74 immunoreactive te~inals~~2 at 140 days. An axon terminal at 140 days is illustrated in Fig. 5D. The most abundant type of immunoreactive terminal was axodendritic; however, some immunoreactive axosomatic terminals were also observed. DISCUSSION

The purpose of this immunohistochemical study has been to comprehensively describe the ontogeny of immunoreactivity for the neuropeptides CGRP and SP, and for the exci~to~ amino acid GLU in sacral primary sensory neurons in the DRG and in the terminal fields of their central (spinal cord) and peripheral (skin) projections, in fetal sheep. These neurochemicals have all been implicated in neurotransmission or neuromodulation of cutaneous information in the cord. We have also described the appearance of S-HT in the spinal cord since it is known that serotonergic axons, which originate in brainstem nuclei, have an inhibitory influence on sensory systems at the spinal level. This is a companion study to an electrophysiological assessment of the functional development of some of the cutaneous afferent pathways (tactile and nociceptive) during gestation. Correlation of information from both these sources will add to our growing understanding of how functional, sensory systems emerge during fetal life. Spinal cord

We have shown that in fetal sheep, very few fibres immunoreactive for CGRP appear in lamina I and for SP in laminae I and II, at 56 days of gestation. Cell bodies immunoreactive for CGRP are present in the motoneuron pool in the ventral horn at this age and for SP in both the dorsal and ventral horns. There are relatively few studies describing the ontogeny of immunoreacti~ty for CGRP and SP in the developing spinal cord in other species. CGRP is first detected in the motoneuron pool in the ventral horn at El5 and in the dorsal horn at El7 in the rat,35 and in the human at week 6 in the ventral horn and week 10 in the dorsal horn.35 For SP, it has been reported that immunoreactive fibres first appear in the dorsal horn of the mouse at E16,4” in the rat at E1648 to Elg3’ and in the human at 11 weeks of fibres in the ventral gestation. 7*35Immunoreactive

24x

I. Ntrsos

and S. REM

Fig. 5. Electron micrographs of CGRP-LIand SP-LI-containing terminals in lamina 1 of the dorsal horn of the spinal cord. (A) SP-LI axodendritic synapses (arrow) containing small spherical vesicles were preseni at 61 days of gestation. (B) At 140 days, SP-LI terminals were larger and more numerous and made axodendritic synapses (arrows). (C) CGRP-LI was also seen at 61 days in axon terminals which contained small spherical vesicles and made axodendritic synapses (arrows). (D) At 140 days, axodendritic synapses (arrows) were larger and more numerous, and contained small spherical and large granular vesicles. Scale bar = 0.5 /Irn for A-D.

Ontogeny of neurochemicais in primary sensory neurons

funiculus are present earlier in development at El4 in the rat4* and mouse.45 Comparison of fetal sheep spinal cord sections with those from rat and human35 suggests that a similar stage of spinal cord development is reached in the rat at E17, the sheep at about day 56 and in the human at week 8 of gestation. This would suggest, therefore, that the two peptides examined in this study appear in the dorsal horn at a similar stage of spinal cord development in the rat and sheep, but reiatively later in humans. Although it is known that peptides remain stable for up to 87 h in post mortem tissue,22 it cannot be ruled out that loss of immunoreactivity due to delayed fixation could contribute to the later appearance in humans. In terms of gestational period, the first appearance of these peptides occurs at 28% of gestation in man, 38% in sheep and 79% in rat. In human and sheep, therefore, peptides are present for an extensive period of the intrauterine development of these pathways. Whether or not this is particularly si~ifi~nt in terms of their growth and differentiation in utero, as opposed to the postnatal growth of these pathways in ahricial species such as the rat and mouse, remains to be tested. There is increasing evidence, however, that neuropeptides exert a widespread trophic action on aspects of neural development, 24 For example, it has been shown in vitro that SP stimulates neurite outgrowth in embryonic DRb3 and that motoneuron-derived CGRP appears to have a trophic action on acetylcholine receptor synthesis.44 The prominent transient expression of neuropeptides which we observed from early in fetal life in cells of the dorsal horn and inte~ediate gray matter (for SP) and motoneurons (for CGRP), and also reported for other species,3s might be a further indication that these substances play an important role in the normal growth and differentiation of the nervous system during fetal development. This initial role of peptides might gradually by replaced by their role as neurotransmitters or modulators.35*45,48 The depletion of immunoreactivity of CGRP* and the marked decrease for SP*’ in the region of termination of primary afferent fibres in the dorsal horn following dorsal rhizotomy in the cat and rat suggests that CGRP and SP are notably concentrated in the central branches of primary aBerent fibres. The appearance of these peptides in the dorsaf horn could therefore be taken to signal the approximate time of arrival of primary afferents in the cord. To demonstrate this more directly, however, we are currently undertaking a study with the anterograde tracer biocytin (Sigma). To determine whether the earliest fibres (56-61 days) to enter the cord in this study were forming synaptic terminals, we localized CGRP and SP immunoreactivity at the ultrastructural level in axon terminals in lamina I. Although sparse, Iabelled terminals making well defined synapses with the dendrites of cells in the dorsal horn were present at this age.

249

The mature pattern and density of immunoreactivity for SP in sheep were similar to those described for other species such as the rat,’ cat,26 monkey’* and man.‘l CGRP immunoreactivity also has a similar distribution to that described for several species.” The only notable difference is that for the rat, McNeil1 et al.” reported a mediolaterai difference in the distribution of CGRP-LI in lamina II,, which is not seen in the sheep or indeed reported for any other species, Althou~ the mature pattern and ~s~bution of these peptides are not reached until well into the third trimester of gestation in both sheep and human,” the intensity of staining, at least in the sheep, is quite marked from 100 days onwards. We have evidence from our electrophysiological studies (Rees, Rawson and Nitsos, unpublished observations) that cells in the dorsal horn respond to peripheral stimuli even earlier than this, at 82 days of gestation, about the time of the first appearance of immunoreactivity for GLU. Although GLU-LI has been described previously in the dorsal horn of the adult rat,” this appears to be the first description of the ontogeny of GLU-LI in the fetal spinal cord; therefore, comparisons with other species cannot yet be made. The appearance of S-HT, one of the neurotransmitters involved in the inhibition of sensory neurons, is an important aspect of the study of the development of pathways concerned with sensory processing. As for other species,s,35 immunoreactivity for 5-HT in fetal sheep was found in the dorsal and ventral horns, in the case of the sheep most densely in laminae IV-X. Immunoreactivity for this amine first appeared at 83 days, which is considerably later than for neuro~ptides (56 days). This result is in agreement with reports in other species, where 5-HT does not in fact appear in the spinal cord until just prior to or after birth. In the rat, fibres immunoreactive for 5-HT appeared at E2135 or in the neonatal period’ with adult density and distribution being reached by postnatal day 14. Similarly, 5-HT has not been detected in human fetal cord and is thought to develop some time after the fjrst six postnatal weeks.35In the rat, the appearance of 5-HT parallels the physiological development of the inhibitory pathway.i5 If this is also the case for the sheep, appearance of 5-HT in the second trimester of gestation would suggest a functional inhibitor system at the time of birth. It should be pointed out that the appearance of amines well after neuropeptides in the cord could be due to lack of species cross-reactivity of the antibodies in very young fetuses the concentration of the antibody being below the level of detection by immunohistochemical techniques or loss of small molecules before fixation. As all fetuses were successfully perfused and fixed immediately upon conclusion of the experiment, fixation problems are unlikely. The other possibilities cannot be totally ruled out, but the fact that a similar result has been found in several other species makes it likely that 5-HT appears well after peptides in the spinal cord.

250

I. NIISOS and S. Rrrrs

Since it has been shown for the rat and human that the developmental profiles for CGRP, SP and 5-HT are similar at all levels of the cord,‘,‘5 it is unlikely that the timetable which we describe for the sacral region in the sheep would be substantially different from that seen at higher levels of the cord. Dorsul root ganglia

cord.‘“.“.‘h Furthermore, it IS also known that 5-HT is co-localized with CGRP’ and with SP,” ” ” mainly within fibres projecting to the ventral horn. There are now many examples of how co-existing transmitters and peptides might interact and it is clear that they can have both synergistic and antagonistic effects, with peptides usually playing the auxiliary role.”

Skin The delay in the time of onset of immunoreactivity for the neuropeptides and GLU in the DRG comIn sheep, the first appearance of peptides in the pared with the spinal cord probably represents the peripheral endings of primary afferent fibres in the skin at 85 days is delayed compared to their appearperiod required for accumulation of these substances ance in the central endings at 56 days. Previous to detectable levels in the DRG cell bodies. Blocking studies have reported the presence of SPJ7.‘“.?’ and axoplasmic transport with colchicine pretreatment2* CGRP’7.“‘,‘9 in nerve fibres in the skin. but the might have resulted in the earlier detection of these ontogeny of these substances in skin has only been substances but was not feasible in this study. described briefly for the rat.35 Marti et u/.‘~ noted, as In this and other studies,20.‘5.48 SP has been localwe report in this study, that peptides were not present ized in small-sized DRG cells (< 30 pm). CGRP has in the skin until some time after there was evidence been detected in small-sized cells in this study and in that axons had innervated the dermis. Furthermore. small- to medium-sized (30-50 Lirn) cells in most this was one (for CGRP) to five days (for SP) after studies on the rat, ” although there are also reports the appearance of these substances in fibres and of its presence in large DRG cells (> 5Opm) in terminals in the dorsal horn. the rat.2a.32,34 GLU was present predominantly in The reason for this delayed appearance is not medium- and large-sized cells in fetal sheep. Howcertain, but could be partly due to slower axonal ever, in the adult rat, some studies have reported that transport in the peripheral branch of the developing GLU is found primarily in small cellsi while sensory fibre or possibly to the peripheral release of others have shown that it is present in medium to peptides. The identity of the large granular cells, large cel1s.5’ immunoreactive for CGRP and SP, situated in close This study supports work on the ontogeny of proximity to the nerve libres early in fetal life is CGRP and SP in the chick DRG,6 which also showed not known, although they have the appearance of that the number of cells expressing these peptides macrophages. It is of interest that similar cells have increased with gestational age. In the present study, been described in the vicinity of growing axons in from 110 to 140 days, the percentages of DRG cells developing human skin.23 expressing CGRP increased significantly from 26 to 37%, for SP from 20 to 29% and for GLU from 30 to 49%. As the total number of ganglion cells pet CONCLUSION longitudinal section is actually decreasing over this period, the percentages will be an underestimate of We have shown that fibres and axon terminals the total number of cells containing these substances immunoreactive for CGRP and SP appear in the fetal per ganglion. As in other species, CGRP is present in sheep spinal cord at the same relative time of cord greater percentage of cells than SP. In general, SP development as the rat, but slightly earlier than is reported to occur in c. 20% of DRG cells17.25,‘K.34 for man. However, for both man and sheep this and CGRP in 4&50%.20,‘8.32.34In the DRG of the rat, is early in fetal life. Neuropeptides might initially 45%5’ to 70% of neurons4 have been reported to be play a trophic role in axonal development and synapimmunoreactive for GLU. tic maturation before converting to their role in As co-localization experiments were not performed sensory transmission. Evidence from our physiologiwe cannot comment on the co-existence of these cal studies indicates that cells in the dorsal horn substances in primary sensory neurons in sheep. first respond to cutaneous input several weeks after However, the similarity in the distribution of the the first appearance of neuropeptides, and that fibres immunoreactive for CGRP and SP in the this neural activity coincides with the appearance dorsal horn of the spinal cord would suggest that of immunoreactivity for GLU. Immunoreactivity co-localization does occur, as has been demonstrated for 5-HT also appears at this age, suggesting that in varicosities within the substantia gelatinosa of the descending pathways have developed. frogs0 and the rat,49 and within DRG neurons in all other species examined so far.‘R,2”,32.34Recently, Acknowledgements-We are grateful to Mrs Jane Ng and MS Sarah Spencer for their excellent technical assistance, to evidence for the co-existence of GLU and SP’” and Dr Al Beitz, University of Minnesota, U.S.A. for donating for SP, GLU and CGRP within the same axon the antiserum to glutamate, and Dr John Rawson and terminals in rat spinal cord has been reported,4’ Prof. G. D. Thorburn for the discussion of the protect. strengthening the hypothesis that these substances This work was supported by a grant from the National Health and Medical Research Council of Australia. interact in sensory transmission in the spinal

Ontogeny of nemochemicals in primary sensory neurons

251

REFERENCES I. Arvidsson U., Schalling M., Cullheim S., Ulfiake

2. 3. 4. 5. 6.

7. 8. 9. 10.

11, 12. 13. 14.

15. 16. 17.

B., Terenius L., Verhofstad A. and Hokfelt T. (1990) Evidence of coexistence between calcitonin gene-related peptide and serotonin in the bulbospinal pathway in the monkey. Brain Res. 532, 41-51. Basbaum A. I. and Fields H. L. (1984) Endogenous pain control systems: brainstem spinal pathways and endorphin circuitry. A. Rev. Neurosci. 7, 309-338. Battaglia G. and Rustioni A. (1988) Coexistence of glutamate and substance P in dorsal root ganglion neurons of the rat and monkey. J. camp. Neural. 277, 302-312. Battaglia G., Rustioni A., Altschuler R. A. and Petrusz P. (1987) In Fine Afirent Nerve Fibers and Pain (eds Schmidt R. F., Schaible H. G. and Vahle-Hinz C.), pp. 78-84. VCH, Weinheim, F.R.G. Bregman B. S. (1987) ~velopmental of serotonin immunoreactivity in the rat spinal cord and its plasticity after neonatal spinal cord lesions. &vi Brain Res. 34, 245-263. Castrignano F., de Stefano M. E., Leone F., Mulatero B., Tata A. M., Faso10 A. and Augusti-Tocco G. (1990) Ontogeny of acetylcholinesterase, substance P and calcitonin gene-related peptide-like immunoreactivity in chick dorsal root ganglia. Neuroscience 34, 499-510. Charnay Y., Paulin C., Chayvialle J.-A. and Dubois P. M. (1983) Distribution of substance P-like immunoreactivity in the spinal cord and dorsal root ganglia of the human foetus and infant. Neuroscience 10, 41-55. Chung K., Lee W. T. and Carlton S. M. (1988) The effect of dorsal rhizotomy and spinal cord isolation on calcitonin gene-related peptide-labelled terminals in the rat lumbar dorsal horn. Neurosci. Lert. 90, 27-32. Cue110A. C., Galfre G. and Milstein C. (1979) Detection of substance P in the central nervous system by a monoclonal antibody. Proc. tram. Acad. Sci. U.S.A. 76, 3532-3536. DeBiasi S. and Rustioni A. (1988) Glutamate and substance P coexist in primary afferent terminals in the superficial laminae of the spinal cord. Proc. narn. Aead. Sci. U.S.A. 85, 7820-7824. De Lanerolle N. C. and La Motte C. C. (1982) The human spinal cord: substance P and me~ionine-enk~halin immunoreacti~ty. J. Neurosci. 2, 1369-1386. DiFiglia M., Aronin N. and Leeman S. E. (1982) Light micro~opic and ultrastructural localization of immunoreactive substance P in the dorsal horn of monkey spinal cord. Neuroscience 7, 1127-I 139. Duggan A. W., Hendry I. A., Morton C. R., Hutchinson W. D. and Zhao Z. Q. (1988) Cutaneous stimuli releasing immunoreactive substance P in the dorsal horn of the cat. Brain Res. 451, 261-273. Fitzgerald M. (1990) A physiological study of the prenatal development of cutaneous sensory inputs to dorsal horn cells in the rat. J. Physiol. 432, 473-182. Fitzgerald M. (1991) The developmental neurobiology of pain. In Pain Research and Clinical Management (eds Bond M. R., Charlton J. E. and Woolf C. J.), Vol. 4, Chapter 29. Elsevier, Amsterdam. Fitzgerald M. and Koltzenburg M. (1986) The functional development of descending inhibitory pathways in the dorsolateral funiculus of the newborn rat spinal cord. Deul Brain Res. 24, 261-270. Gibbins I. L., Furness J. B. and Costa M. (1987) Pathway-specific patterns of the coexistence of substance P, calcitonin gene-related peptide, cholecystokinin and dynorphin in neurons of the dorsal root ganglia of the guinea pig. Cell Tiss.

Res. 248,417437. 18. Gibbins I. L., Furness J. B., Costa M., MacIntyre I., Hillyard C. J. and Girgis S. (198.5) Co-localisation

19. 20.

21. 22. 23. 24. 25. 26. 27.

28. 29. 30.

31. 32.

of calcitonin gene-related peptide-like immunoreactivity with substance P in cutaneous, vascular and visceral sensory neurones of guinea pigs. Neurosci. L&t. 57, 125-130. Gibbins I. L., Watta~how D. and Coventry B. (1987) Two immunohist~hemically identified populations of calcitonin gene-related peptide (CGRF)-immunoreactive axons in human skin. Brain Res. 414, 143-148. Gibson S. J., Polak J. M., Bloom S. R., Sabate I. M., Mulderry P. M., Ghatei M. A., McGregor G. P., Morrison J. F. B., Kelly J. S., Evans R. M. and Rosenfeld M. G. (1984) Calcitonin gene-related peptide immunoreactivity in the spinal cord of man and eight other species. J. Neurosci. 4, 3101-31 Il. Gilbert R. F. T., Emson P. C., Hunt S. P., Bennett G. W., Marsden C. A., Sandberg 8. E. E., Steinbusch H. W. M. and Verhofstad A. A. J. (1982) The effects of monoamine neurotoxins on peptides in the rat spinal cord. Neuroscience 7, 69-87. Helme R. D. and White D. M. (1983) Immunocytochemical localisation of substance P in postmortem rat and human spinal cord. J. Neuropath. exp. Neural. 42, QQ--105. Hogg 1. D. (1941) Sensory nerves and associated structures in the skin of human fetuses of 8 to 14 weeks of menstrual age correlated with functional capability. J. camp. Neural. 75, 371410. Hokfelt T. (1991) Neuropeptides in perspective: the last ten years. Neuron 7, 867-879. Hokfelt T., Johansson O., Ljungdahl A., Lundberg J. M. and Schultzberg M. (1980) Peptidergic neurones. Nature 284, 515.-521. Hokfelt T., Kellerth J. O., Nilsson G. and Pernow B. (1975) Ex~~mental immunohist~hemi~l studies on localization and distribution of substance P in cat primary sensory neurons. Bruin Res. 100, 235-252. Hokfelt T., Ljungdahl A., Steinbusch H., Verhofstad A. N., Nilsson G., Brodin E., Pernow 8. and Goldstein M. (1978) Immunohistochemical evidence of substance P-like immunoreactivity in some 5-hydroxytryptamine containing neurons in the rat central nervous system. Neuroscience 3, 517-538. Inagaki S. and Kito S. (1986) Peptides in the peripheral nervous system. Prog. Brain Res. 66, 269-316. Ishida-Yamamoto A., Senba E. and Tohyama M. (1989) Distribution of fine structure of calcitonin gene-related peptide like immunoreactive fibres in the rat skin. Brain Res. 491, 93-101. Jesse1 T. M., Yoshiko K. and Jahr C. E. (1986) Amino acid receptor mediated transmission of primary afferent synapses in rat spinal cord. J. exp. Biol. 124, 239-258. Jonakait G. M., Ni Li, Walker P. D. and Hart R. P. (1991) Development of substance P (SP)-containing cells in the central nervous system: consequences of neurotransmitter co-localisation. Prog. Neurobiol. 36, l-2 1. Ju G., Hokfelt T., Brodin E., Fahrenkrug J. Fischer J. A., Frey P., Elde R. P. and Brown J. C. (1987) Primary sensory neurons of the rat showing calcitonin gene-related peptide immuno~a~tivity and their relation to substance P-, somatostatin-, galanin-, vasoactive intestinal polypeptide-, and ~hol~ystokinin-immunoreactive ganglion cells. Celi Tiss. Res. 247. 417431.

32

I. NI~SOS and S. REES

33. Kangrga I., Larew J. S. A. and Randic N. (1990) The effects of substance P and calcitornn gene-relateo peptide on the efflux of endogenous glutamate and aspartate from rat spinal dorsal horn in vitro. Neuro.cci. Lr//. 108, I.5 Ihtl. 34. Lee Y., Takami K., Kawai Y., Girgis S., Hillyard C. J.. MacIntyre I., Emson P. C. and Tohyama M. (1985) Distribution of calcitonin gene-related peptide in the rat peripheral nervous system with reference to its coexistence with sulxtnncc P. Neuroscience 15, 122771231. 35. Marti E., Gibson S. J., Polak J. M., Facer P., Springall D. R., Van Aswegen G., Aitchison M. and Koltzenburg M (1987) Ontogeny of peptide- and amine-containing neurones in motor, sensory, and autonomic regions of rat and human spinal cord, dorsal root ganglia, and rat skin. J. romp. Neural. 266, 332-359. 36. Mayer C. and Zieglgansberger W. (1987) Peptidergic neurotransmission in the dorsal horn of the rat spinal cord. In Fine ,4fferent Nerve Fibres and Pain (eds Schmidt R. F.. Schaible H. G. and Vahle-Hinz C.). pp. 249 262. VCH Press. Weinheim. 37. McCarthy P. W. and Lawson S. N. (19X9) Cell type and conduction velocity of rat primary sensory neurons with substance P-like immunoreactivity. Neuroscience 28, 745 753. 38. McCarthy P. W. and Lawson S. N. (1990) Cell type and conduction velocity of rat primary sensory neurons with calcitonin gene-related peptide-like immunoreactivity. Neuroscience 34, 623632. 39. McDonald A. J., Beitz A. J.. Larson A. A., Kuriyama A. A., Sellitto C. and Mad1 J. E. (1989) Co-localization of glutamate and tubulin in putative excitatory neurons of the hippocampus and amygdala: an immunohistochemical study using monoclonal antibodies. Neuroscience 30, 405-42 1. 40. McNeil1 D. L.. Coggeshall R. E. and Carlton S. M. (1988) A light and microscopic study of calcitonin gene-related peptide in the spinal cord of the rat. E.xp/ Neural. 99, 6999708. 41. Merighi A., Polak J. M. and Theodosis D. T. (1991) Ultrastructural visualization of glutamate and aspartate immunoreactivities in the rat dorsal horn. with special reference to the co-localization of glutamate, substance P and calcitonin gene-related peptide. Neumcciencr 40, 67--X0. 42. Morton C. R. and Hutchinson W. D. (1989) Release of sensory neuropeptides in the spinal cord: studies with calcitonin gene-related peptide and galanin. Neuroscience 31, 807 -8 15. 43. Narumi S. and Fujita T. (1978) Stimulatory effects of substance P and nerve growth factor (NGF) on neurite outgrowth in embryonic chick dorsal root ganglia. hieuropkart7luco/o~~ 17, 73.-76. 44. New H. V. and Mudge A. W. (1986) Calcitonin gene-related peptide regulates muscle acetylcholine receptor synthesis. Nature 323, 80998 I 1. 45. Ni L. and Jonakait G. M. (1988) Development of substance P-containing neurons in the central nervous system in mice: an immunohistochemical study. J. camp. Neural. 275, 493.-510. 46. Proudfit H. K. and Hammond D. L. (1983) Alterations in nociceptive threshold and morphine induced analgesia produced by intrathecally administered amine antagonists. Brain Res. 218, 393-399. 47. Schneider S. P. and Per1 E. R. (1988) Comparison of primary afferent and glutamate excitation of neurones in the mammalian spinal dorsal horn. J. Neurosci. 8, 2062-2073. 4X. Senba E., Shiosaka S.. Hara Y., lnagaki S., Sakanaka M., Takatsuki K., Kawai Y. and Tohyama M. (1982) Ontogeny of the peptidergic system in the rat spinal cord: immunohistochemical analysis. J. camp. Neural. 208, 54.~66. 49. Tuchsherer M. M. and Seybold V. S. (1989) a quantitative study of the co-existence of peptides in varicosities within the superficial laminae of the dorsal horn of the rat spinal cord. J. Neurosci. 9, 195- 205. 50. Venesio T., Mulatero B. and Faso10 A. (1987) Coexistence of substance P and calcitonin gene-related peptide in the frog spinal cord. Nrurosci. Let/. 80, 246 250. 51. Wanaka A., Shiotani Y.. Kiyama H.. Matsuyama T., Kamada T., Shiosaka S. and Tohyama M. (1987) Glutamate-like immunoreactive structures in primary sensory neurons in the rat detected by a specific antiserum against glutamate. E.xpl Brain Rm. 65, 691 -694. (Accepted 23 October 1992)