THE JOURNAL OF COMPARATIVE NEUROLOGY 318:l-17 (1992)

Serotoninergic, Noradrenergic, and Peptidergic Innervation of OnuFs Nucleus of Normal and Transected Spinal Cords of Baboons (Papiopapio) N. RAJAOFETRA, J.-G. PASSAGIA, L. MARLIER, P. POULAT, F. PELLAS, F. SANDILLON, B. VERSCHUERE, D. GOUY, M. GEFFARD, AND A. PRIVAT INSERM U.336 (DPVSNkEPHE, 34095 Montpellier (N.R., L.M., P.P., F.P., F.S., A.P.), Service de Neurochirurgie, HBpital La Tronche, 38043 Grenoble (J.-G.P.),Centre de Recherches Sanofi Clin-Midy,34080 Montpellier (B.V.,D.G.), and Laboratoire d'Immunologie, CJF 88-13 INSERM, BP 66, Universite de Bordeaux 11, 33076 Bordeaux (M.G.), Cedex France

ABSTRACT We have investigated with light and electron microscope immunocytochemistry the aminergic and peptidergic innervation of Onuf s nucleus in adult baboons. This nucleus, located in the ventrolateral part of the sacral spinal cord (S, and SJ, is considered to control urethral and anal sphincters and penile muscles. By comparison of intact and transected spinal cords, we have found that serotoninergic innervation has two origins: first, supraspinal, innervating the whole nucleus, with a possible predominance in the dorsal half; and second, intraspinal, corresponding to the ventral half of the nucleus. Thyrotropin-releasing hormone innervation appears largely coincident with serotonin, both in intact and transected spinal cords. Noradrenaline is exclusively of supraspinal origin, as attested by its disappearance below the level of the section. Substance P, calcitonin gene-related peptide, and Leu- and Met-enkephalin, which profusely innervate Onuf's nucleus, are on the contrary not affected by the transection. They most likely originate from the cord itself or the dorsal root ganglia. Thus, Onuf s nucleus innervation in the baboon arises both from supraspinal and intraspinal sources. The present study provides an anatomical basis for both voluntary and reflex controls of excretory and sexual functions in a primate. The same neurotransmitter (serotonin) according to its cell origin and discrete topography could exert different influences upon the same effector system. Key words: serotonin, noradrenaline,peptides, transection, primate

Onuf s nucleus, first described in the human spinal cord by Onufrowicz (1900), is a group of neurons located in the sacral ventral horn (S, and S,) (Schrqjder, '81; Gibson et al., '84). Retrograde tracing techniques have shown that the homologue of this nucleus in animals innervates the perineal striated muscles and external urethral and anal sphincters (Sato et al., '78; Kuzahara, '79; Nagashima et al., '79; Kuzahara et al., '80; De Groat et al., '81; Roppolo et al., '85; McKenna and Nadelhaft, '86) and a similarity has been suggested between the functions of Onuf s nucleus in animals and humans (Onufrowicz, 1900; Schr@der,'81). However, there is no direct experimental evidence to demonstrate the exact function of Onuf's nucleus. Clinicopathological studies have shown that Onuf s nucleus is spared, and vesicorectal functions are preserved in patients with amyotrophic lateral sclerosis (Mannen et al., '77, '82; Sung, '82; o 1992 WILEY-LISS, INC.

Oyanagi et al., '83; Konno et al., '86). In contrast, in Shy-Drager syndrome (characterized by a lesion of the autonomic centers), this nucleus is affected by a loss of neurons that is correlated with disturbed vesicorectal functions (Sung et al., '79; Konno et al., '86). Recently, Kojima et al. ('89) reported that Onufs nucleus escaped inflammatory degeneration due to poliovirus infection. Conversely, several studies in experimental animals indicated that neurons of Onufs nucleus are somatic instead of autonomic (Sato et al., '78; Kuzahara et al., '80). Lastly, Yamamoto et al. ('78) suggested that this nucleus contained a mixed neuronal population. In the course of a transplantation experiment in the spinal cord of adult baboons (Rajaofetra and Privat, ,881, we found that, after spinal cord Accepted November 15,1991.

2

N. RAJAOFETRA ET AL.

transection, the distal part of the cord and especially the sacral motoneuron area presented a significant serotoninergic innervation by intrinsic serotonin (5-HT) neurons, thus differing from rodent spinal cord where most if not all serotoninergic afferents are of supraspinal origin. Following these results, the present study was undertaken in order to examine with immunocytochemistry the serotoninergic, noradrenergic, and peptidergic innervations of Onufs nucleus in the intact and transected spinal cords of adult baboons (Papiopupio). Indeed, the contribution of supraspinal vs. intraspinal influences upon micturition (Kuru, '65; Maggi and Meli, '86) is still a matter of controversy, and both the origin and the topography of aminergic and peptidergic afferents are relevant to this issue.

MATERIALS AND METHODS Surgical procedures Seven male baboons (Pupio pupio) provided by Sanofi Recherche-farm (France) weighing between 7 and 10 kg were used in this study. Five of them were subjected to complete spinal cord transection and the remaining two animals were kept intact for control without any manipulation. In the former, following premedication with Vetranquil (0.75 mg/kg) and intravenous pentobarbital anesthesia (30 mg/kg), a thoracic (T,) laminectomy was performed under sterile conditions. The spinal cord was completely transected, including the dura mater at T,T, cord levels. No spinal cord tissue was removed. After hemostasis, a retraction of several millimeters (3-5 mm) of the spinal stumps occurred spontaneously, thus testifying for complete section. The back musculature and the skin were then sutured, and the animals were placed in a special heated box until full recovery from anesthesia. After surgery, the animals were kept under standard welfare conditions in individual cages by Sanofi Recherchefarm. They routinely received prophylacticdoses of antibiotics (penicillin-streptomycin 20,000 U/kg, i.m.1 for 1week, and urinary retention during the first week was counteracted by manual compression on the bladder; their diet consisted mainly of fruit and vegetables.

enkephalins (Enk) (Cambridge) and thyrotropin-releasing hormone (TRH) (Oliveret al., '74). After treatment with trypsin-EDTA for 5 minutes and sodium borohydride (0.038% for 5-HT and NA, 1% for TRH) for 10 minutes, the sections were incubated in the primary antiserum (5-HT 1:20,000; NA 1:15,000; SP 1:4,000; TRH 1:10,000; CGRP 1:4,000; Met-Enk 1:10,000; Leu-Enk 1:8,000)with 1%of nonspecific goat serum (NSS) for 24 hours at 4°C in Tris-SMB.For light microscopy, 0.1% Triton X-100 was added during incubation. After rinses in 50 mM of Tris-saline buffer, sections were successively incubated for 1 hour with goat anti-rabbit IgG (Nordica) antiserum, and rabbit peroxidase-antiperoxidase(Dako) at a dilution of 1 : l O O with 1 : l O O of NSS in Tris-saline buffer. Under visual control, the peroxidase deposit was revealed with 0.05%diaminobenzidine and 0.01%hydrogen peroxide in the same buffer (Sternberger et al., '70). Finally, sections were transferred to 60 mM phosphate buffer, mounted on glass slides, dried overnight at 50"C, cleared in xylene, and mounted with Giirr-DPX. For electron microscopy, sections from the seven animals were selected for optimal signal/background ratio and reacted with each of the antiserum. They were postfixed for 40 minutes in osmium tetroxyde (1%osmium, 100 mM phosphate buffer), dehydrated in graded series of ethanol concentrations, infiltrated with araldite, and then flatembedded, remounted on araldite-filled gelatin capsules, and resectioned in a parallel plane with an LKB ultramicrotome. The ultrathin sections were collected on 200 mesh grids. They were then contrasted with uranyl acetate and examined with a Jeol2000 electron microscope. For routine light microscopy, some sections were stained with cresyl violet (Nissl stain).

RESULTS With a Nissl stain, Onufs nucleus appears as a small circular group of neurons, located in the sacral (S,-S,) ventrolateral horn (Fig. 1).We shall now describe, with the use of immunocytochemicalmethods, 5-HT, NA, TRH, SP, CGRP, and Enk immunoreactive fibers and terminals in Onufs nucleus with light and electron microscopy in intact and transected spinal cords.

Sacrifice and immunocytochemical processing

Light microscopy

Experimental animals were sacrificed 2 months after spinal transection, and controls were processed in the same way. A tracheotomy was performed for artificial ventilation on the animals under sodium pentobarbital anesthesia, followed by a thoracotomy. Following intracardiac injections of 5 ml heparin (5,000 U/ml) and 20 ml sodium nitrite (l%), the animals were perfused through the heart first with 800 ml to 1000 ml glutaraldehyde (0.5%) in cacodylic acid 50 mM-sodium metabisulfite (SMB) 50 mM buffer at pH 7.5, followed by 8-10 liters of fixative composed of glutaraldehyde (5%)in the same buffer. The spinal cords were then removed and immersed in the same fixative for 24 hours at 4°C. On all seven animals, cross sections (50 km thick) of the spinal cords at the sacral 6,-S,) level were performed with a Vibratome in Tris (50 mM)-SMB (50 mM) buffer at pH 7.6. Then the sections were treated for immunocytochemical detection of 5-HT and noradrenaline (NA)(Geffardet al., '85); for calcitonin gene-related peptide (CGRP) (Amersham); and for substance P (SP) and Leu- and Met-

Serotonin (5-HT). The most salient feature of the baboon spinal cord related to serotonin is the abundance of

Fig. 1. A vibratome cross section of a baboon spinal cord with Nissl stain, showing Onuf s nucleus (arrows) and the dorsolateral nucleus (DL), which are located in the ventral horn, at sacral level. Scale bar = 1 mm. Fig. 2. A cross section of a control spinal cord immunostained for 5-HT, showing the characteristic distribution of 5-HT immunoreactivity around the central canal (CC) and 5-HT intraspinal neurons, which are located in its ventral part and along the ventral sulcus (arrow). Scale bar = 80 pm. Fig. 3. A cross section at the sacral level, showing 5-HT intraspinal neurons in spinal cords transected 2 months prior to immunostaining. CC, central canal. Scale bar = 80 pm. Fig. 4. High magnification of fusiform bipolar and multipolar 5-HT intraspinal neurons. Scale bar = 40 pm.

Figures 1 4

4

intrinsic 5-HT neurons. A preliminary report (Rajaofetra and Privat, '88) provided the first immunocytochemical demonstration of intrinsic 5-HT neurons in the adult baboon spinal cord. These cells were found in both normal and L-tryptophan/MAO inhibitor-loaded baboons at all levels of the spinal cord. In this study, we focused our attention on the sacral level, where Onufs nucleus is present. At this level, 5-HT perikarya were located mostly ventrally to the central canal (CC)and along the ventral sulcus of the cord (Fig. 2). Most of these intrinsic 5-HT neurons are small and fusiform; a few cells are larger or polygonal. In Onufs nucleus, 5-HT immunoreactive fibers were distributed diffusely, with a particular concentration in close apposition to the neurons of this region (Fig. 5). Moreover, the highest concentration of 5-HT immunoreactive fibers at the sacral level was also detected around the CC (layer X of Rexed, '52) and mostly in medial and dorsolateral areas of ventral horn, and the superficial part of the dorsal horn (layer I). I n transected spinal cords, intrinsic spinal 5-HT neurons similar to those seen in the intact animal (Figs. 3, 4) appeared with light microscopy. After transection, immunoreactive fibers were concentrated in two locations: firstly, around the CC (Fig. 31, and they were seen to extend through ependymal cells. Secondly, 5-HT fibers and terminals were highly concentrated in the ventral part of the Onufs nucleus where they appeared denser than in the intact animal (Fig. 6); in the dorsal part, 5-HT fibers were scanty. Noradrenaline (NA). In the intact animal, noradrenergic fibers were mostly concentrated in the ventral part of Onuf s nucleus (Fig. 7). No intrinsic spinal noradrenergic neurons were detected with our technique. After spinal cord transection, no N A immunoreactive fibers were observed in the cord below the transection (Fig. 8).

Thyrotropin-releasing hormone (TRH). At the light microscopic level, TRH immunoreactive fibers appeared concentrated in Onufs nucleus, the remainder of the sacral ventral horn being devoid of immunoreactivity (Fig. 9). After transection, TRH immunoreactive fibers and terminals were present only in the ventral part of Onuf s nucleus (Fig. 10). Substance P (SP). In the intact animal, a dense network of varicose SP axons was observed in Onuf s nucleus (Fig. 11).In our experimental conditions (no pretreatment with colchicine), scanty or no SP cell bodies were detected ventrally and dorsally to the CC of the sacral cord. After spinal cord transection, the distribution of SP immunoreactive fibers was similar to that of the intact animal (Fig. 12). Calcitonin gene-related peptide (CGRP). With light microscopy, CGRP immunoreactive profiles appear scanty in Onufs nucleus. (Fig. 13). We have not observed any CGRP-labeled perikarya, contrary to the case of other nuclei of the anterior horn. After spinal cord transection, immunodetection of CGRP disclosed large cells bodies in medial and dorsolateral groups of motoneurons, but none in Onufs nucleus, where only immunoreactive fibers were detected (Fig. 14). Leu- and Met-enkephalins (Enk). In control animals, Enk immunoreactive fibers densely innervated Onufs nucleus (Figs. 15, 17).

N. RAJAOFETRA ET AL. After spinal cord transection, a high concentration of Enk immunoreactivities was still found around the neurons in Onuf s nucleus (Figs. 16, 18).

Electron microscopy Serotonin (5-HT).With the electron microscope, intrinsic 5-HT neurons were readily identified, and generally appeared as small bipolar cells. Their cytoplasm contained an abundance of endoplasmic reticulum, a well-developed Golgi apparatus, and many mitochondria (Fig. 19).Usually, small glial cells were found in apposition to 5-HT neurons. These neurons received afferent boutons on their somata (Fig. 20) and their main dendritic trunks and spines (Fig. 21). The unlabeled axonal terminals presynaptic to intrinsic 5-HT neurons contained numerous small spherical vesicles mixed with a few flattened ones. The contacts were generally asymmetric (type I) showing a conspicuous postsynaptic differentiation. In Onuf s nucleus, medium-sizedimmunoreactive varicosities were in synaptic contact with large and medium-sized dendrites, and with perikarya. The majority of 5-HT axonal varicosities contained densely packed, small spherical synaptic vesicles, a few dense-core vesicles, and many mitochondria. They established numerous symmetrical (type 11) synaptic contacts with large and medium (1pm and more) dendrites (Fig. 22) and cell bodies (Fig. 23). Rare type I synapses were observed between 5-HT terminal boutons and small dendrites or spines (Fig. 24). I n the transected spinal cord, no morphological differences were seen for intrinsic 5-HT neurons vs. control. With the electron microscope, 5-HT immunoreactive profiles appeared in Onuf s nucleus as large and medium varicosities containing small, round, and dense synaptic vesicles that most often established symmetrical type I1 synaptic contacts with large and medium-sized dendrites (Fig. 251, exhibiting only a modest postsynaptic density. In Onuf's nucleus, few 5-HT terminals establishing type I synapses were encountered (Fig. 26). Also a few axosomatic type I1 contacts were found as in the intact animal (Fig. 23). Noradrenaline. With the electron microscope, most immunoreactiveprofiles appeared as large varicositiescontaining dense and round vesicles, a large number of mito-

Figs. 5 and 6. Cross sections through Onufs nucleus, showing 5-HT immunoreactive fibers in intact animal (Fig. 5) and after spinal cord transection (Fig. 6). Note the high concentration of 5-HT fibers in the ventral part of Onuf s nucleus in transected spinal cord, whereas the dorsal part is almost completely depleted. Scale bars = 80 pm. Figs. 7 and 8. Distribution of NA immunoreactivity in Onuf's nucleus. NA fibers are highly concentrated in the ventral part of Onuf s nucleus in the intact animal (Fig. 7). After spinal cord transection, OnuPs nucleus is devoid of imrnunoreactivity (Fig. 8). Scale bars = 80 Pm.

Figs. 9 and 10. Cross sections through Onufs nucleus, showing

TRH immunoreactivity in intact (Fig. 9) and transected spinal cords (Fig. 10).Arrow points to a cell body surrounded by TRH fibers. After transection, TRH irnmunoreactivity is only detected in ventral part of Onuf s nucleus. Scale bars = 80 pm. Figs. 11and 12. Cross sections showing dense SP immunoreactivity in Onufs nucleus, in control (Fig. 11) and transected (Fig. 12)spinal cords. Scale bars = 80 pm.

Figures 5-12

ONUF'S NUCLEUS IN BABOON chondria, but few dense-cored vesicles. They established symmetrical type I1 and asymmetrical type I synaptic contacts with small-, medium-, and large-sized dendrites (Figs. 27, 28). Symmetrical (type 11) axosomatic contacts were frequently found (Fig. 28). Thyrotropln-releasing hormone (TRH). With the electron microscope, TRH axon terminals were seen to form synaptic contacts in Onuf s nucleus. The majority of small and medium axon terminals containing pleomorphic vesicles and many mitochondria were observed to contact medium and large dendrites (Fig. 29), making type I and type I1 synapses. They established symmetrical type I1 synapses with large neuronal somata (Fig. 30). After spinal cord transectwn, only axodendritic synapses were detected. TRH immunoreactive profiles were filled with small clear and round vesicles intermixed with many dense-cored vesicles, and exhibited type I and type I1 synaptic contacts with medium and large dendritic processes (Fig. 31). Substance P (SP). With the electron microscope, we could see in Onuf s nucleus a large number of SP varicosities containing mostly clear, round vesicles, and a few flattened ones, which established types I and I1 synapses with medium and large dendrites (Fig. 32). Few densecored vesicles were present in these terminals. Numerous axosomatic contacts were also observed (Fig. 33); these axon terminals containing dense-cored and small clear vesicles contributed type I1 synapses in Onuf's nucleus. A few type I axosomatic synapses were also detected. After spinal cord transection, SP immunoreactive profiles appeared similar to those of the control. The majority of SP immunoreactive axon terminals exhibit asymmetrical type I synaptic contacts with large dendrites and spines (Fig. 34).The cell bodies of Onuf s nucleus were also contacted by SP immunoreactive s o n s with symmetrical type I1 and asymmetrical type I synapses, as shown in Figure 35. Calcitonin gene-related peptide (CGRP). With the electron microscope, CGRP immunoreactive profiles appeared as large- and medium-sized varicosities (1 pm and more), which most often made type I synaptic contacts with largeand medium-sized dendrites (Fig.36);type I1 synapses were less frequent. Axosomatic synapses were rare. &er, spinal cord transection, CGRP immunoreactive profiles exhibited large and medium Varicosities containing round, elongated, and densely packed vesicles that were most often in asymmetrical synaptic contact (type I) with large- and medium-sized dendrites (Fig. 37). The postsynaptic densities frequently exhibited an array of subjunctional dense bodies. Very few symmetrical (type 11)synapses were encountered (Fig. 38) and neuronal somata were only occasionally contacted by CGRP immunoreactive profiles. Leu- and Met-enkephalins (Enk). Leu- and Metenkephalins presented the same ultrastructural features: Enk immunoreactive fibers appeared as medium and large varicositiescontaining small and round vesicles and a large

7

number of dense-cored vesicles. Enk terminals exhibited mostly type I1 (Figs. 39,40) axodendritic synapses, and a very few asymmetrical type I contacts (Fig. 41).In addition, (Figs. 39, 421, cell bodies of Onufs nucleus were seen contacted by large Enk immunoreactive axons, exhibiting symmetrical type I1 axosomaticsynapses. After transection of the spinal cord, most Enk immunoreactive profiles appeared as large and medium varicosities containing small round and dense vesicles making type I1 synaptic contacts with medium and large dendrites as in the intact animal. Sometimes, they were seen in contact with dendritic spines as shown in Figure 43,and less frequently with cell bodies (Fig. 44). A small number of E n k immunoreactive axons were seen making type I synapses, with a prominent postsynaptic density (Fig. 45).

DISCUSSION With light and electron microscope immunocytochemistry, serotoninergic and peptidergic profiles were detected in Onuf s nucleus in the intact and the transected spinal cord of baboons. Contrary to the other transmitters studied, no noradrenaline immunoreactive fibers were observed in Onuf s nucleus after transection.

5-HT innervation In this study, the use of spinal transections showed that 5-HTspinal cord innervation has two origins in the adult baboon: A first group of fibers originates from a supraspinal population of brain stem 5-HT neurons, most probably from the caudal raphe nuclei, as reported by Azmitia and Gannon in Mama fascicularis ('86);these fibers disappear from the cord after transection. Other fibers originate from intraspinal 5-HT neurons (LaMotte et al., '82; Bowker, ,861, which are described in the present study as the one source of Onuf s nucleus innervation that remains after complete spinal cord tR3INdXOl-L

In both control and transected spinal cords, as stated previously (Rajaofetraand Privat, '881, L-tryptophan/MAO loading was not necessary to detect intrinsic 5-HT perikarya. Contrary to the case of the rat (Newton et al., '86; Newton and Hamill, '88), intrinsic 5-HT neurons are numerous in the baboon, and they are located essentially in the ventral and lateral parts of the CC in Rexed's lamina X ('52) (see LaMotte et al., '82; Bowker, '861, although some were observed outside this area along the ventral SUICUS, and occasionallyat the base of the dorsal horn at sacral level, in transected cords. No 5-HTneurons were detected dorsally to the central canal. In this study, a systematic quantitative survey of these cells was not performed. However, we did not observe a marked difference between cell densities at cervical and thoracolumbar levels. In transected cords, as in controls, 5-HT neurons were quite numerous at sacral levels. These Figs. 13 and 14. Cross sections showing CGRP immunoreactive results are in agreement with those of Bowker ('86). On the fibers in Onufs nucleus and CGRP motoneurons in dorsolateral other hand, in the same species, LaMotte et al. ('82) nucleus (arrows) of aacral ventral horns in control (Fig. 13) and reported greater numbers of intrinsic 5-HT neurons at transected (Fig.14)cords. Scale bars = 80 pm. cervical level than in the more caudal thoracolumbar levels. The ontogeny of intrmpinal 5-HT neurons in primates Figs.15-18. Cross sections showing dense Leu- (Fig.16,161and Met-enkephalinergic (Figs. 17, 18)innervation in the Onufs nucleus has not been thoroughly studied, except by Bowker ('86) in of control (Figs. 15,17) and trans& (Figs. 16, 18) spinal cords. Scale the late term baboon fetal spinal cord. Their midline and ventral position suggests a continuity with raphe nuclei of bars = 80 Fm.

Fig. 19. A fusiform 5-HT intraspinal neuron as seen with EM immunocytochemistry, containing a notched nucleus, many mitochondria, and a prominent Golgi zone (arrow). A small glial cell is in apposition to this neuron. Scale bar = 2 pm.

Figs. 20 and 21. AfTerent synapses to 5-HT immunoreactive neurons: a nonimmunoreactive bouton containing round and elongated small vesicles makes a symmetrical type I1 synapse with a 5-HT immunoreactive soma (Fig. 20) and a large bouton makes an asymmetrical synapse with a dendritic spine. Scale bar = 500 nm in Figure 20 and 200 nm in Figure 21.

ONUF'S NUCLEUS IN BABOON

9

the caudal rhombencephalon, constituting the so-called In the rat medullary raphe nuclei, it has been reported median paracore (Nieuwenhuys, '85). that TRH immunoreactivity was present in 5-HT neurons Numerous studies have described the presence of spinal (Johansson et al., '81). In the spinal cord, TRH, SP, and 5-HT neurons in a variety of species, and suggestions have 5-HT were all reduced by 5-HT neurotoxins (Gilbert et al., been made concerning their possible functions (LaMotte et '82; Marsden et al., '82), suggesting their possible coexistal., '82; Ritchie and Leonard, '82; DiTirro et al., '83; Ritchie ence. Moreover, in the rat, Barbeau and BBdard ('81) et al., '84; Harris-Warrick et al., '85; Van Dongen et al., suggested that TRH seemed to have 5-HT releasing proper'85). For instance, Ritchie et al. ('84), considering the ties in the spinal cord. In agreement with the results of distribution of these neurons and their processes around Felten et al. ('74) and Hubbard and DiCarol('74) in squirrel the CC and along the ventral sulcus, suggested their monkey, and of Azmitia and Gannon ('86) in Macaca involvement in the secretion or control of cerebrospinal fascicularis, we did not observe any major difference befluid. Their influence on ventral horn neurons and particu- tween the distribution and the morphology of 5-HT neurons in medullary raphe between the rat and the baboon. larly Onuf s nucleus has never been studied. After spinal transection, we found that 5-HT innervation Thus, we can hypothesize that descending TRH innervain Onufs nucleus was modified; only the ventral part still tion to the spinal cord also arises from the medullary raphe. Regarding the local origin of spinal TRH innervation, contained immunoreactive profiles suggesting a predominant supraspinal origin for the innervation of the dorsal particularly that of Onuf s nucleus, no TRH immunoreacpart. This projection appeared denser than in the control. tive cell bodies were ever detected below the transection (for Conversely,5-HT immunoreactive perikarya were not mod- ethical consideration, animals were not treated with colchiified in numbers or in morphology in transected cords. With cine). However, the similarity between TRH and 5-HT EM observation, in this nucleus, numerous synaptic con- innervations suggested their coexistence in the same termitacts were detected on the dendrites and fewer on the soma nals in Onuf s nucleus. of Onufs neurons in control and transected spinal cords. SP innervation We could not find any qualitative difference between 5-HT serotoninergic synapses in control and transected animals In this study, we found many SP immunoreactive fibers in the ventral part of Onuf s nucleus. Most of them were and terminals distributed a t the sacral level, particularly type I1 symmetrical synapses, and involved dendrites rather within Onuf s nucleus. After spinal transection we did not than cell bodies. observe any major decrease of SP immunoreactivity. This The exact significance of this dorsoventral partition is indicates an intraspinal origin for the majority of SP fibers, obscure. Worth noting, however, is the coincidence of this as in the cat (where they contain also enkephalins) (Tashiro partition with the topography of TRH innervation in the et al., '89) and in the rat (see Micevych, '86; Uda et al., '85, transected cord, and of NA innervation in the intact cord. '86). In the absence of colchicinetreatment, we only detected a Noradrenergic innervation few SP cells, around the CC. In the cat, numerous reports The noradrenergic innervation of Onuf s nucleus is the mentioned the presence of SP cell bodies in the spinal cord only one that disappears completely after spinal cord (Naftchi et al., '78; Tashiro et al., '89). In Onuf s nucleus, Tashiro et al. ('89) have observed that transection. We have not found any evidence of intrinsic noradrenergic perikarya in the lumbosacral cord. The only a small fraction of SP fibers, also showing 5-HT immunorereport of intraspinal NA neurons is that of Wolters et al. activity, have a supraspinal origin. In the rat spinal cord, ('89) claiming, with the use of an anti-NA antibody, the the coexistence of SP and 5-HT in the ventral horn was presence of immunoreactive cells in layers I and I1 of the rat studied by our laboratory in the course of transplantation dorsal horn at lower lumbar levels. We did not find any such studies; we could observe with the electron microscope this cells in the intact adult rat spinal cord, nor did we find NA coexistence in the terminals of transplanted 5-HT neurons immunoreactive profiles after transection (Yakovleff et al., derived from Bl-B2 groups (Rajaofetra et al., '91).With a '89) in accordance with previous descriptions (Nygren and different immunocytochemical method, Foster et al. ('85) Olson, '77; Westlund et al., '81, '83, '84; Fritschy et d., '87; also revealed this coexistence. Finally, this colocalization Fritschy and Grzanna, '90).Disappearance of this innerva- was indirectly confirmed by the depletion of SP immunoretion after transection indicates their exclusively supraspi- activity after 5-HT neurotoxin treatment. Indeed, Emson nal origin. The coincidence of this disappearance from the ventral half of the nucleus, with the increased density of 5-HT innervation in the same area, could suggest that 5-HT fibers have replaced the degenerated NA endings. Figs. 22-24 (see next page). Electron micrographs showing 5-HT

TRH innervation In the control spinal cord, TRH fibers and terminals were diffusely present and intensely stained in Onuf s nucleus, as already described in rhesus monkeys (Lechan et al., '84). After thoracic spinal transection, TRH immunoreactivity was only detected in the ventral part of Onufs nucleus. It therefore appears that the topography of this innervation and its reaction to injury are similar to those of 5-HT. In addition, the ultrastructural characteristics of 5-HT and TRH boutons are also similar. These findings indicate that, as for 5-HT, TRH innervation of Onufs nucleus has a double origin: supraspinal and intraspinal.

varicosities in Onufs nucleus of control spinal cord. Figure 22: A medium 5-HT bouton containing round and small vesicles makes a symmetrical (type XI) synapse with a large dendrite. Scale bar = 500 nm. Figure 2 3 A large 5-HT bouton containing round, small, and clear vesicles and many mitochondria establishes symmetrical (type 11) synaptic contact with a cell body of Onuf s nucleus. Scale bar = 500 nm. Figure 2 4 A 5-HT bouton makes an asymmetrical (typical type I) synapse with a small dendrite.Scale bar = 1 km. Figs. 25 and 26 (next page). Electron micrographs showing 5-HT varicosities in Onuf s nucleus 60 days after spinal cord transection. Figure 2 6 A large unlabeled dendrite is contacted by a large 5-HT bouton making a symmetrical SYMPS~ (type 11). Scale bar = 1 km. Figure 26: A large dendrite is contacted by a 5-HT bouton with evidence of asymmetrical (type I) synapse. Scale bar = 1 pm.

Figures 22-26

Figs. 27 and 28. Electron micrographs showing NA immunoreactive boutons in synaptical contact in Onufs nucleus. Figure 27: An NA immunoreactive bouton containing many mitochondria and round vesicles makes a typical asymmetrical synapse on a large dendrite. Scale bar = 500 nm. Figure 28: Numerous NA boutons containing many mitochondria establish symmetrical (type 11) synapses with a cell body and dendrites. Scale bar = 1 pm.

Figs. 29 and 30. Electron micrographs of control spinal cord, showing TRH varicosities that make an asymmetrical (type I) synapse on a large dendrite (Fig.29) and symmetrical (type 11) synapse on a cell body (Fig. 30).Scale bar = 500 nm in Figure 29 and 1 pm in Figure 30. Fig. S1. A TRH bouton containing many dense and small and round vesicles makes an asymmetrical (type I) synapse with a large dendrite, 60 days after spinal cord transection. Scale bar = 500 nm.

12

Figs. 32 and 33. Electron micrographs showing SP immunoreactive varicosities in Onufs nucleus of the control spinal cord. They establish asymmetrical and symmetrical synapses with large dendrites (Fig.32) and several symmetrical contacts with a cell body (Fig. 33). Scale bar = 500 nm in Figure 32 and 1 Fm in Figure 33.

N. RAJAOFETRA ET AL.

Figs. 34 and 35. After spinal cord transection, SP boutons establish axodendritic synapses (Fig. 34) and an asymmetrical axosomatic synapse (Fig. 35) with a conspicuous postsynaptic density. Scale bars = 1 Fm.

ONUF'S NUCLEUS IN BABOON

13

Fig. 36. A large calcitonin gene-related peptide (CGRP) bouton making asymmetrical (type I) synapse with a large nonimmunoreactive dendrite in Onuf s nucleus of intact animal. Scale bar = 1 ym. Figs. 37 and 38. After spinal cord transection, large CGRP varicosities containing small and clear vesicles make asymmetrical (type I) (Fig. 37) and symmetrical (type 11) (Fig. 38) synapses with medium and large dendrites. Scale bars = 1 ym in Figure 37 and 500 nm in Figure 38.

and Gilbert ('79) showed a severe loss of SP and of 5-HT uptake after 5-HT neurotoxin, and concluded that both SP and 5-HT may be stored in the same spinal cord fibers originating from caudal raphe nuclei. However, contrary to TRH immunoreactivity, which closely follows the topography and evolution of 5-HT, SP in Onufs nucleus was only minimally affected by the transection, indicating a nonsupraspinal origin, at least partially independent of that of 5-HT intrinsic neurons, which project only to the ventral part of the nucleus. The primary sensory fibers from dorsal root ganglia are a likely origin for these fibers (Hokfelt et al., '75; Takahashi and Otsuka, '75; Naftchi et al., '78).

CGRP innervation In this study, immunodetection of CGRP in control and transected spinal cords disclosed large immunoreactive cell bodies in the medial and dorsolateral groups of motoneurons, but none in Onuf s nucleus, where only immunoreactive terminals were present. These findings suggest that Onuf s neurons are different from most motoneurons of the anterior horn, in the baboon cord. Recent work (Feuerstein, personal communication) has shown that in the rat most of the motoneurons appeared CGRP and C U T immunoreactive, whereas in Onufs nucleus, neurons were only C U T

Figures 39-45

15

ONUF'S NUCLEUS IN BABOON positive. Furthermore, in amyotrophic lateral sclerosis, Onufs nucleus neurons were spared. In contrast, in Fabry's disease, neuronal loss was observed within this cell group. On the other hand, Kuzahara et al. ('80)reported that Onufs nucleus was somatic. The second finding is that CGRP innervation of Onuf s nucleus is not modified by transedion: synaptic contacts detected on dendrites and soma were still present, suggesting that they do not originate in supraspinal sources. A possible source is that of primary sensory fibers from dorsal root ganglia. It can be hypothesized that at least part of SP and CGRP varicosities are the same, i.e., that the two peptides coexist in terminals located in Onufs nucleus. Indeed, in both cases most of the synapses are type I contacts involving large dendrites. These boutons could correspond to a monosynaptic reflex arc involving Onuf s nucleus.

Enkephalinergic innervation We found profuse Met and Leu-Enk immunoreactive fibers and terminals in Onuf s nucleus of the control spinal cord. After spinal transection, no changes of enkephalinergic distribution were detected. These results are consistent with previous observations in the rat by Romagnano et al. ('87); transections and rhizotomy did not induce changes of enkephalinergic innervation at the sacral level. In the cat, no changes of enkephalinergic distribution were observed after hemicordotomy (Tashiro et al., '89).It was also reported that 30-50% of SP terminals in Onufs nucleus exhibited enkephalin immunoreactivity, and originated mainly from lamina X at the sacral level (see Konishi et al., '85). The results of the present study are consistent with the idea that enkephalin immunoreactive fibers and terminals in Onuf s nucleus originate from an intraspinal source. Zn conclusion, the innervation of Onuf s nucleus in the spinal cord, as seen from the present study, can be divided into four groups, which, however, are not totally independent of each other: 1. A monoaminergic innervation made of noradrenergic and serotoninergic fibers. All noradrenergic fibers originate from supraspinal sources, whereas serotoninergic fibers have a dual origin: those that innervate the dorsal part of the nucleus are essentially supraspinal, while

those that innervate its ventral part originate from intraspinal serotoninergic neurons. 2. TRH innervation that is very similar to that of 5-HT, and follows a similar fate after transection. Our hypothesis is that TRH is mostly colocalized in 5-HT axons, whether they originate from supra- or intraspinal sources. 3. A peptidergic innervation by substance P and CGRP containing fibers, which is not substantially modified by transection, and which could originate massively from dorsal root ganglia neurons. 4. A peptidergic innervation by Leu and Met-enkephalin fibers, which would originate mainly from neurons intrinsic to the cord. This sharp delineation does not exclude other associations, such as for instance that of SP and 5-HT, but their generally different distribution after transection excludes a major coincidence. The prominent role of Onuf s nucleus in the control of vesical and anal sphincters would be under the influence of a complex array of afferents: besides the voluntary control of the pyramidal system, which was outside the scope of this study, it appears that local, intraspinal systems may have a major influence. A characteristic feature of the baboon is the outstanding development of an intrinsic serotoninergic system, of which Onufs nucleus is a major target. After spinal transection, these animals, contrary to other species, such as the rat, rapidly regained a reflex control of urination and defecation. This could be ascribed at least in part to an intrinsic serotoninergic control. It is worth noting that such control was partially restored in the rat by transplantation of embryonic serotoninergic neurons below a complete transection of the cord (Privat et al., '88).

ACKNOWLEDGMENTS We are indebted to the technicians of Centre de Toxicologie de Sanofi-Recherchefor their excellent technical contributions during the course of this study for monkey care. We wish to thank J.R. Teilhac for high quality photographical assistance. This work was supported by grants from IRME, the D. Heumann Fund for Spinal Cord Research, and AFM.

LITERATURE CITED Figs. 39 and 40. In the control spinal cord, numerous Leu- (Fig. 39) and Met- (Fig. 40) Enk boutons containing small and round vesicles make mainly symmetrical (type 11) synapses with medium and large dendrites in Onufs nucleus. Scale bars = 1wm. Fig. 41. A conspicuous asymmetrical (type I) synapse between a large Met-Enk bouton and a large unlabeled dendrite, with an array of subjunctional dense bodies in Onuf s nucleus of control spinal cord. Scalebar=Ipm. Fig. 42. In the control spinal cord, Met-Enk boutons make symmetrical contact synapses with an immunoreactive cell body. Scale bar = 500 nm. Figs. 43-45. Electron micrograph showing Enk varicosities in Onuf s nucleus 60 days after spinal cord transection. Figure 43: A Leu-Enk bouton makes a symmetrical synapse with a dendritic spine. Scale bar = 1 pm. Figure 44: A large Met-Enk bouton making a symmetrical (type 11) synapse with a neuron soma (N). Scale bar = 1 pm. Figure 45: A large unlabeled dendrite is contacted by a Leu-Enk bouton, making an asymmetrical synaptic contact. Scale bar = 500 nm.

Azmitia, E.C., and P.J. Gannon (1986) The primate serotonergic system: A review of human and animal studies and a report onMacacafascicularis. Adv. Neurol. 43:407-468. Barbeau, H., and P. BQdard(1981) Similar motor effects of 5-HT and TRH in rats following chronic spinal transection and 5,7-dihydroxytryptamine injections.Neuropharmacology20:477481. Bowker, R.M. (1986) Intrinsic 5HT-immunoreactive neurons in the spinal cord of fetal non-human primate. Dev. Brain Res. 28:137-143. De Groat, W.C., I. Nadelhaft, R.J. Miline, A.M. Booth, C. Morgan, and K. Thor (1981) Organisationof the sacral parasympatheticreflex pathways to the urinary bladder and large intestine. J. Auton. Nerv. Syst. S135-160. DiTirro, F.J., G.F. Martin, and R.H. Ho (1983) A developmental study of substance P, somatostatin, enkephalin and serotonin immunoreactive elements in the spinal cord of the North American opossum. J. Camp. Neurol. 213:241-261. Emson, P.C., and R.F.T. Gilbert (1979) Time course of degeneration of bulbo-spinal 5-HT/SP neurones after 5,7-dihydroxytryptamine.Prw. B.P.S. 17th-19th.279. Felten, D.L., A.M. Laties, and M.B. Carpenter (1974)Monoamine-containing cell bodies in the squirrel monkey brain. Am. J. Anat. 139:153-166. Foster, G.A.,M. Schultzber, F.H.Gage, A. Bjorklund, T. Hokfelt, H. Nornes, A.C. Cuello, A.A.J. Verhofstad, and T.J. Visser (1985) Transmitter

16

N. RAJAOFETRA ET AL.

expression and morphologicaldevelopment of embryonic medullary and mesencephalic raphe neurons after transplantation to the adult rat central nervous system. I. Grafts to the spinal cord. Exp. Brain Res. 60t427-444. Fritschy, J.M., and R. Grzanna (1990) Distribution of locus coeruleus axons within the rat brainstem demonstrated by Phaseolus vulgaris leucoagglutinin anterograde tracing in combination with dopamine-beta-hydroxyylase immunofhmrescence.J. Comp. Neurol. 293:61&631. fitschy, J.M., W.E. Lyons, C.A. Mullen, B.E. Kosofsky, M.E. Molliver, and R. Grzanna (1987) Distribution of locus coeruleus axons in the rat spinal cord: A combined anterograde transport and immunohistochemical study. Brain Res. 437:176180. Geffard, M., A.M. Henrick-Roek, J. Dulluc, and P. Seguela (1985) Antisera against small neurotransmitter-lie molecules. Neurochem. Int. X403413. Gibson, S.J., S.R. Bloom, and J.M. Polak (1984) A novel substance P pathway linking the dorsal and ventral horn in the upper lumbar segments of the rat spinal cord. Brain Res. 301:243-251. Gilbert, R.F.T., P.C. Emson, S.P. Hunt, G.W. Bennet, C.A. Marsden, B.E.B. Sandberg, H.W.M. Steinbusch, and A.A.J. Verhofstad (1982) The effects of monoamine neurotoxins on peptides in the rat spinal cord. Neuroscience 7:69-87. Harris-Warrick, RM.,J.C. McF'hee, and J.A. Filler (1985) Distribution of serotonergic neurons and processes in the lamprey spinal cord. Neuroscience 14:1127-1 140. Hokfelt, T., J.O. Kellerth, G. Nilsson, and B. Pernow (1975) Experimental immunohistochemical studies on the localization and distribution of substance P in cat primary sensory neurons. Brain Res. 100:235-252. Hubbard, J.E., and V. DiCarol (1974) Fluorescence histochemistry of monoamine-containing cell bodies in the brain stem of the squirrel monkey (Saimiri sciureus). 11. Serotonin-containing groups. J. Comp. Neurol. 153:385-398. Johansson, O., T. Hokfelt, B. Pernow, S.L. Jeffcoate, N. White, H.W.M. Steinbusch, A.A.J. Verhofstad, P.C. Emson, and E. Spindel (1981) Immunohistochemical support for three putative transmitters in one neuron: Coexistence of 5-hydroxytryptamine, substance P- and thyrotropin-releasing hormone-like immunoreactivity in medullary neurons projecting to the spinal cord. Neuroscience 6:1857-1881. Kojima, H., Y. Furuta, M. Fujiti, Y. Fujioka, and K. Nagashima (1989) Onuf s motoneuron is resistant to poliovirus. J. Neurol. Sci. 93335-92. Konishi, A., K. Itoh, T. Sugimoto, Y. Yasui, T. Kaneko, M. Takada, and N. Mizuno (1985) Leucineenkephalin-like immunoreactive afferent fibers to pudendal motoneurons in the cat. Neurosci. Lett. 6It109-113. Konno, H.,T. Yamamoto, Y. Iwasaki, and H. Iizuka (1986) Shy-Drager syndrome and amyotrophic lateral sclerosis. J. Neurol. Sci. 73tl93-204. Kuru, M.(1965)Nervous control of micturition. Physiol. Rev.4.5425494. Kuzahara, S. (1979) Localization of sacral spinal neurons innervating the perineal muscles-a HRP study in dogs. Clin. Neurol. (Jpn.) 19:704711. Kuzahara, S.,I. Kanazawa, and T. Nakanishi (1980) Topographic localisation of Onuf s nucleus neurons innervatingthe rectal and vesical striated sphincter muscles: A retrograde fluorescent double labelling study in cat and dog. Neurosci. Lett. 16t713-723. LaMotte, C.C., D.R. Johns, andN.C. De Lanerolle (1982) Immunohistochemical evidence of indolamine neurons in monkey spinal cord. J. Comp. New]. 206:35%370. Lechan, R.M., S.B. Snapper, S. Jacobson, and I.M.D. Jackson (1984) The distribution of thyrotropin-releasing hormone (TRH) in the Rhesus monkey spinal cord. Peptides 5(Suppl.1):186-194. Maggi, C.A., and A. Meli (1986)The role of neuropeptides in the regulation of micturition reflex. J. Auton. Pharmacol. 6t133-162. Mannen, T., I. Makoto, Y. Toyokura, and K. Nagaehima (1977) Preservation of a certain motoneuron group of the sacral cord in amyotrophic lateral sclerosis: Its clinical significance. J. Neurol. Neurosurg. Psychiatry 40~464-469. Mannen, T., M. Iwata, Y. Toyokura, and K. Nagashima (1982) The Onufs nucleus and the external and sphincter muscles in amyotrophic lateral sclerosis and Shy-Drager syndrome. Acta Neuropathol. (Berl.) 58:255260.

Marsden,C.A.,G.W.Bennet, J. Irons, R.F.T.Gilbert,andP.C. Emson(1982) Localization and release of 5-hydroxytryptamine, thyrotropin-releasing hormone and substance P in rat ventral spinal cord. Comp. Biochem. Physiol. 72Ct263-270. McKenna, K.E., and I. Nadelhaft (1986) The organisation of the pudendal nerve in the male and female rat.J. Comp. Neurol. 24R532-547. Micevych, P.E., A. Coquelin, and A.P. Arnold (1986) Immunohistochemical

distribution of substance P, serotonin, and methionine enkephalin in sexually dimorphic nuclei of the rat lumbar spinal cord. J. Comp. Neurol. 248:235-244. Naftchi, N.E., S.J. Abraham, H.M. St. Paul, E.W. Lowman, and W. Schlosser (1978) Localization and changes of substance P in spinal cord of paraplegic cats. Brain Res. 153t507-513. Nagashima, T., M. Beppu, M. Uono, and H. Yamanda (1979) Demonstration of n e u r o d localization in OnufrotRiez's group X in rabbit by double labeling method. Acta Histochem. Cytochem. 12:409A15. Newton, B.W., and R.W. Hamill (1988) The morphology and distribution of rat serotonergic intraspinal neurons: An immunohistochemical study. Brain Res. Bull. 20:349-360. Newton, B.W., B.E. Maley, and R.W. Hamill (1986) Immunohistochemical demonstration of serotonin neurons in the autonomic regions of the rat spinal cord. Brain Res. 376t155-163. Nieuwenhuys, R. (1985) 'New' entities in the central nervous system: The tpracrine?l core and its adjuncts. In: Chemoarchitecture of the brain. Berlin: Springer-Verlag,pp. 177-193. Nygren, L.G., and L. Olson (1977) 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 Res. 132t85-93. Oliver, C., R.L. Eskay, N. Ben-Johnathan, and J.C. Porter (1974) Distribution and concentration of TRH in the rat brain. Endocrinology 96340546. Onufrowicz, B. (1900) On the arrangement and formation of the cell groups of the sacral region of the sacral spinal cord in man. Arch. Neurol. Psychopathol. 3:387412. Oyanagi, M., T. Makifuchi, and F. Ikuta (1983) A topographic and quantitative study of neurons in human spinal gray matter, with special reference to their changes in amyotrophic lateral sclerosis. Biomed. Res. 4t211224. Privat, A., H. Mansour, and M. Geffard (1988) Transplantation of foetal serotonin neurons into the transected spinal cord of adult rats: Morphological development and functional influence. Prog. Brain Res. 78:155166. Rajaofetra, N., and A. Privat (1988) 5-HT innervation of the adult baboon spinal cord with a special emphasis on intrinsic neurons. Soc. Neuroaei. 48t190. (Abstract). Rajaofetra, N., P. Kachidian, L. Marlier, P. Poulat, N. Konig, M. GBffard, and A. Privat (1991) Electron microscopic detection of the axonal coexistence of serotonin and substance P in Bl-B2 raphe cells transplanted into the transected spinal cord of adult rat. Brain Res. 542159-162. Rexed, B. (1952) The cytoarchitectonic organisation of the spinal cord in the cat. J. Comp. Neurol. 96:415-495. Ritchie, T.C., and R.B. Leonard (1982) Immunocytochemicaldemonstration of serotonergic cells, terminals and axons in the spinal cord of the stingray, Dasyatis sabina. Brain Res. 240:334-337. Ritchie, T.C., L.J. Roos, B.J. Williams, and R.B. Leonard (1984) The descending and intrinsic serotoninergic innervation of an elasmobranch spinal cord. J. Comp. Neurol. 224t395-406. Romagnano, M A , J. Braiman,M. Loomis, and R.W. Hamill (1987)Enkephalin fibers in autonomic nuclear regions: Intraspinalvs supraspinal origin. J. Comp. Neurol. 266t319-331. Roppolo, J.R., I. Nadelhaft, and W.C. De Groat (1985) The organisation of pudendal motoneumns and primary afferent projections in the spinal cord of the rhesus monkey revealed by horseradish peroxidase. J. Comp. Neurol. 234t475-488. Sato, M., N. Mizuno, and A. Konishi (1978) Localisation of motoneurons innervating the pelvic muscles of the male rat. J. Comp. Neurol. 192t567-587.

Schwder, H.D. (1981) Onufs nucleus X: A morphologieal study of a human spinal nucleus. Anat. Embryol. 162443-453. Sternberger, L.A., P.H. Hardy, J.J. Cuculis, and H.G. Meyer (1970) The unlabeled antibody enzyme method of immunohistochemistry preparation and properties of soluble antigen-antibody complex horseradish antihorseradish peroxidase and its use in identification spirochetes. J. Histochem. Cytochem. 18t315-333. Sung, J.H. (1982) Autonomic neurons of the sacral spinal cord in amyotrophic lateral sclerosis, anterior poliomyelitis and neuronal intranuclear hyaline inclusion disease: Distribution of sacral autonomic neurons. Acta Neuropathol. 56:233-237. Sung, J.H., A.R. Mastri, and E. Segal (1979) Pathology of Shy-Drager syndrome. J. Neuropathol. Exp. Neurol. 38:353-368. Takahashi, T.,and M. Otsuka (1975) Regionaldistribution of substance P in the spinal cord and nerve mots of the cat and effeet of dorsal root section. Brain Res. 87:1-11

ONUFS NUCLEUS IN BABOON Tashiro, T., M.A. Ruda, T. Satoda, R. Matsushima, and N. Mizumo (1989) Convergence of serotonin-, enkephalin- and substance P-like immunoreactive afferent fibers on single pudendal motoneurons in Onuf s nucleus of the cat: A light microscope study combining the triple immunocytochemical staining technique with the retrograde HRP-tracing method. Brain Res. 481:392-398. Uda, K., H. Okamura, and Y. Ibata (1985) Immunocytochemical localization of substance P in the rat spinal cord with special reference to fibers within the ventral column of the rostra1 lumbar segments. Neurosci. Lett. 57:185-190. Uda, K., H. Okamura, F. Kawakami, andY. Ibata (1986) Sexual differences in the distribution of substance P immunoreactive fibers in the ventral horn of the rat lumbar spinal cord. Neurosci. Lett. 64:157-162. Van Dongen, P.A.M., T. Hokfelt, S. Grillner, A.A.J. Verhofstad, and H.W.M. Steinbusch (1985) Possible target neurons of 5-hydroxyttryptamine fibers in the lamprey spinal cord: Zmmunohistochemistry combined with intracellular staining with lucifer yellow. J. Comp. Neurol. 234:523635. Westlund, K.N., R.M. Bowker, M.G. Ziegler, and J.D. Coulter (1981) Origins

17 of spinal noradrenergic pathways demonstrated by anterograde transport of antibody to dopamine-p-hydroxylase.Neurosci. Lett. 25.243-249. Westlund, K.N., R.M. Bowker, M.G. Ziegler, and J.D. Coulter (1983) Noradrenergic projections to the spinal cord of the rat. Brain Res. 263: 15-31. Westlund, K.N., R.M. Bowker, M.J. Ziegler, and J.D. Coulter (1984) Origins and terminations of descending noradrenergic projections to the spinal cord of monkey. Brain Res. 292r1-16. Wolters, J.G., H.W.M. Steinbusch, and J.G.J.M. Bol(1989) The distribution of dopamine-versus noradrenaline-immunoreactive fibers and cell bodies in the spinal cord of the rat. SOC.Neurosci. 10.17:20 (Abstract). Yakovleff, A., A. Roby-Brami, B. Guezard, H. Mansour, B. Bussel, and A. Privat (1989) Locomotion in rats transplanted with noradrenergic neurons. Brain Res. Bull. 22115-121. Yamamoto, T., H. Satomi, H. Ise, H. Takamata, and K. Takahashi (1978) Sacral spinal innervations of the rectal and vesical smooth muscles and the sphinderic muscles a s demonstrated by the horseradish peroxidase method. Neurosci. Lett. 7.4-47.