Dendritic organization of the human spinal cord: The motoneurons

The dendritic organization of motoneurons was analyzed with the Golgi stain and a morphometric method in the immature and adult human spinal cord. Each motoneuronal column was found to be characterized by a specific orientation of dendritic trees and by a distinct pattern of dendritic bundling. Ventromedial motoneurons have a pyramidal dendritic tree with numerous, short longitudinal branches and elongated dorsal branches. The latter form thick bundles oriented toward the ventral gray commissure. Longitudinal dendrites form a narrow-meshed dendritic plexus, containing abundant microbundles. Motoneurons of the ventromedial column have fewer primary dendrites and a lower ramification index than other motoneurons. Central motoneurons are predominantly oriented longitudinally. The meshes of the rostrocaudal dendritic plexus are looser and the microbundles are finer. Most transverse dendrites run laterally and participate in dendritic bundles which penetrate into the ventrolateral funiculus. The rostrocaudal dendritic domain of ventrolateral motoneurons is the largest dendritic domain of all spinal neurons. The longitudinal dendritic network contains fine microbundles and appears wide-meshed. Transverse dendrites form lateral or media dendritic bundles depending upon the position of their perikaryon. Dorsolateral motoneurons differ from other motoneurons by their multipolar organization with a slight preponderance of dorsoventral dendritic spread. Rudimentary lateral dendrite bundles are restricted to marginal neurons. The longitudinal plexuses of motoneuronal dendrites and the verticotrans verse dendrite bundles of the ventromedial column are well developed in the 26–28-week-old fetus. In contrast, the horizontotransverse dendrite bundles of central and ventrolateral motoneurons can only be recognized from 36 weeks on. The possible specific functions of the various types of dendrites bundles are examined and a laminar dendroarchitectonic schema of the human cord is proposed.     

Changes in synaptic inputs to sympathetic preganglionic neurons after spinal cord injury

Spinal cord injury (SCI) leads to plastic changes in organization that impact significantly on central nervous control of arterial pressure. SCI causes hypotension and autonomic dysreflexia, an episodic hypertension induced by spinal reflexes. Sympathetic preganglionic neurons (SPNs) respond to SCI by retracting and then regrowing their dendrites within 2 weeks of injury. We examined changes in synaptic input to SPNs during this time by comparing the density and amino acid content of synaptic input to choline acetyltransferase (ChAT)-immunoreactive SPNs in the eighth thoracic spinal cord segment (T8) in unoperated rats and in rats at 3 days or at 14 days after spinal cord transection at T4. Postembedding immunogold labeling demonstrated immunoreactivity for glutamate or gamma-aminobutyric acid (GABA) within presynaptic profiles. We counted the number of presynaptic inputs to measured lengths of SPN somatic and dendritic membrane and identified the amino acid in each input. We also assessed gross changes in the morphology of SPNs using retrograde labeling with cholera toxin B and light microscopy to determine the structural changes that were present at the time of evaluation of synaptic density and amino acid content. At 3 days after SCI, we found that retrogradely labeled SPNs had shrunken somata and greatly shortened dendrites. Synaptic density (inputs per 10-microm membrane) decreased on ChAT-immunoreactive somata by 34% but increased on dendrites by 66%. Almost half of the inputs to SPNs lacked amino acids. By 14 days, the density of synaptic inputs to dendrites and somata decreased by 50% and 70%, respectively, concurrent with dendrite regrowth. The proportion of glutamatergic inputs to SPNs in spinal cord-transected rats ( approximately 40%) was less than that in unoperated rats, whereas the GABAergic proportion (60-68%) increased. In summary, SPNs participate in vasomotor control after SCI despite profound denervation. An altered balance of excitatory and inhibitory inputs may explain injury-induced hypotension.     

Ultrastructural analysis of choline acetyltransferase-immunoreactive sympathetic preganglionic neurons and their dendritic bundles in rat thoracic spinal cord

We have used a monoclonal antibody against choline acetyltransferase (ChAT) to aid in the identification of sympathetic preganglionic neurons (SPNs) and to examine their ultrastructure in rat thoracic spinal cord. The clusters of ChAT-immunoreactive (ChAT-IR) preganglionic cell bodies and their distinctive bundles of dendrites give rise to a ladder-like appearance in horizontal light microscopic sections. This organization also produced a characteristic appearance of preganglionic neuropil when viewed electron microscopically. The intermediolateral (IML) nucleus contained numerous rostrocaudally oriented ChAT-IR dendrites. In addition, mediolaterally oriented ChAT-IR dendrites extended between the IML and the central autonomic region. Both the ChAT-IR dendrites and somata of preganglionic neurons were intimately associated with astroglial processes. The cell bodies were typically covered over a large proportion of their surface by a thin astrocytic sheath, and this was associated with a paucity of axon terminals forming axosomatic synapses. Instead, the vast majority of synapses upon SPNs were of the axodendritic type. The most frequently observed type of axon terminal contained numerous round, clear vesicles along with several dense-core vesicles (DCVs). In addition, some boutons contained a predominance of DCVs. Serial section analysis revealed that these apparently diverse morphological classes of synaptic boutons may simply represent variability of structure throughout a single terminal, with a greater proportion of DCVs being located distal to the synaptic specialization and a greater number of round, clear vesicles being present adjacent to the synapse. Analysis of the dendritic bundles revealed that individual dendrites, like the cell bodies, were often isolated from each other and the surrounding neuropil by astrocytic processes. This arrangement usually was interrupted only at regions of synaptic contact where astrocytic processes surrounded the synaptic complex as a whole. Thus, astroglial ensheathment of SPNs seems designed to minimize cross-talk between the bundled dendrites, as well as to isolate or segregate the diverse afferent inputs known to impinge on these cells.     

N-methyl-D-aspartate-type glutamate receptors are found in post-synaptic targets of adrenergic terminals in the thoracic spinal cord

Adrenergic (C1) neurons in the rostral ventrolateral medulla (RVL) are sympathoexcitatory and project directly to sympathetic preganglionic neurons (SPNs) in the thoracic spinal cord. C1 neurons also contain glutamate which may mediate the excitatory effects of RVL stimulation on SPNs through the N-methyl-D-aspartate (NMDA)-type receptor. Dual-labeling immunocytochemistry, combined with electron microscopy, was used to determine if NMDA receptors are located post-synaptic to adrenergic terminals in the spinal cord. Adrenergic terminals were labeled using an antibody directed against phenylethanolamine-N-methyl transferase (PNMT) and the NMDA receptor was identified using an antibody directed against the R1 subunit of the receptor (NMDAR1). NMDAR1 was found primarily in large and small dendrites and a few perikarya. The presence of NMDAR1 in the dendritic targets of PNMT-containing terminals was quantified in spinal cords sectioned either horizontally or coronally. PNMT-labeled terminals formed asymmetric synapses on dendrites containing immunogold labeling for NMDAR1, but NMDAR1 was more often detected in the targets of PNMT terminals when spinal cords were sectioned horizontally (59%) rather than coronally (28%). This difference in prevalence of NMDAR1 in targets of PNMT terminals is likely due to the preferential orientation of SPN dendrites in the horizontal plane, since longer dendritic shafts were visible in horizontal sections. When NMDAR1 was present in the dendritic targets of many adrenergic terminals, it was usually located at sites distal to the adrenergic input. We conclude that NMDA receptor ligands are likely to modulate the activity of dendritic targets of adrenergic terminals in the spinal cord, but are not closely associated with adrenergic synaptic input.     

The vegetative network in the thoracolumbar spinal cord of the guinea pig: a comparison of the distribution of AChE-enzyme activity and choline acetyltransferase-like immunoreactivity

Histochemically the acetylcholinesterase (AChE) enzyme activity and immunocytochemically the choline acetyltransferase-like immunoreactivity (ChAT-LI) were located in the components of the vegetative network of the thoracolumbar spinal cord of the guinea pig. Both reaction products showed an identical distribution among the preganglionic sympathetic cells and the processes of the vegetative network, namely: cells of the nucleus intermediolateralis pars principalis (ILp), the nucleus intermediolateralis pars funicularis (ILf), the nucleus intercalatus spinalis (IC) and the nucleus intercalatus paraependymalis (ICpe; terminology according to Petras and Cummings 1972). In longitudinal horizontal sections through the intermediate zone of the thoracolumbar spinal cord both AChE-positive- and ChAT-like immunoreactive nerve fibers were organized into two longitudinal lateral fascicles (FLL), two longitudinal medial fascicles (FLM) as well as oblique and transverse bundles that interconnect repeatedly the autonomic cell groups of this zone along the spinal cord and contribute to the ladder-like shape of the vegetative network. The ChAT-like immunostaining of the vegetative network showed that the dendrites of the ILp cells are oriented mainly in a rostrocaudal, but also in a mediolateral direction. Similar orientation of the dendrites was observed for the ICpe cell groups of the thoracolumbar intermediate zone. Thus it is evident that ILp cell bodies and dendrites are involved in the formation of the FLL, whereas the ICpe cells and their dendrites--of the FLM. The IC cells send their dendrites towards both the ILp and ICpe cells and build up together with dendrites of the ILp and ICpe cells the transverse and oblique interconnecting bundles. The vegetative network is strongly developed within the intermediate zone of T1-T4 (mostly T3) and T7-T8 segments of the spinal cord. In these segments a greater variety of interconnections between the preganglionic sympathetic cell groups which are constituents of the network are also revealed. In the remaining segments the vegetative network is more poor developed. In the first two lumbar segments the distance between the interconnecting bundles in the rostrocaudal direction diminishes to 100 microns in contrast to 300-500 microns within the upper segments. The results obtained reveal that the cholinergic preganglionic sympathetic nuclei of the intermediate zone of the thoracolumbar spinal cord together with their dendrites represent the basis (frame) of the ladder-like vegetative network to which join in addition different peptidergic fibers of supraspinal, peripheral and propriospinal origin.     

N-methyl- -aspartate-type glutamate receptors are found in post-synaptic targets of adrenergic terminals in the thoracic spinal cord

Adrenergic (C1) neurons in the rostral ventrolateral medulla (RVL) are sympathoexcitatory and project directly to sympathetic preganglionic neurons (SPNs) in the thoracic spinal cord. C1 neurons also contain glutamate which may mediate the excitatory effects of RVL stimulation on SPNs through the N-methyl- -aspartate (NMDA)-type receptor. Dual-labeling immunocytochemistry, combined with electron microscopy, was used to determine if NMDA receptors are located post-synaptic to adrenergic terminals in the spinal cord. Adrenergic terminals were labeled using an antibody directed against phenylethanolamine-N-methyl transferase (PNMT) and the NMDA receptor was identified using an antibody directed against the R1 subunit of the receptor (NMDAR1). NMDAR1 was found primarily in large and small dendrites and a few perikarya. The presence of NMDAR1 in the dendritic targets of PNMT-containing terminals was quantified in spinal cords sectioned either horizontally or coronally. PNMT-labeled terminals formed asymmetric synapses on dendrites containing immunogold labeling for NMDAR1, but NMDAR1 was more often detected in the targets of PNMT terminals when spinal cords were sectioned horizontally (59%) rather than coronally (28%). This difference in prevalence of NMDAR1 in targets of PNMT terminals is likely due to the preferential orientation of SPN dendrites in the horizontal plane, since longer dendritic shafts were visible in horizontal sections. When NMDAR1 was present in the dendritic targets of many adrenergic terminals, it was usually located at sites distal to the adrenergic input. We conclude that NMDA receptor ligands are likely to modulate the activity of dendritic targets of adrenergic terminals in the spinal cord, but are not closely associated with adrenergic synaptic input.     

Phenylethanolamine N-methyltransferase-containing terminals synapse directly on sympathetic preganglionic neurons in the rat

The ultrastructural morphology as well as neuronal and glial associations of phenylethanolamine N-methyltransferase (PNMT)-containing terminals in the intermediolateral cell column (IML) of the thoracic spinal cord were examined in the rat utilizing the peroxidase-antiperoxidase (PAP) method. The PNMT-immunoreactive terminals were 0.5-1.4 micron in diameter and contained a few mitochondria, a large population of small clear vesicles and from 1 to 6 large dense-core vesicles. The terminals formed synapses primarily with dendrites. The type of axodendritic association (i.e. symmetric or asymmetric) varied with the size of the dendrite, such that the majority of synapses on large dendrites were symmetric and those on smaller dendrites and dendritic spines were asymmetric. Moreover, most of the synaptic associations of PNMT-containing terminals were with the smaller dendritic processes. Many of the PNMT-labeled terminals, as well as their postsynaptic targets, were closely invested with, or apposed to fibrous astrocytic processes. In a subsequent set of experiments, we combined immunoautoradiographic labeling for PNMT with horseradish peroxidase (HRP) retrograde identification of sympathetic preganglionic neurons (SPNs) in the IML to determine whether or not SPNs receive direct synaptic input from the adrenergic terminals. In these sections, PNMT-containing terminals directly synapsed on the HRP-containing (i.e. retrogradely labeled SPNs) perikarya and dendrites. The axosomatic synapses observed between PNMT-labeled terminals and SPN perikarya were exclusively symmetric; whereas the type of axodendritic association varied depending upon the size of the dendrite such that the majority were asymmetric. The findings provide ultrastructural evidence that in the rat IML, adrenergic (i.e. PNMT-containing) terminals (1) may be either excitatory (asymmetric) or inhibitory (symmetric) depending on their site of termination and (2) can influence sympathetic nerve discharge through a direct effect on the SPN cell membrane.     

Morphology of sympathetic preganglionic neurons in the thoracic spinal cord of the cat: an intracellular horseradish peroxidase study

Horseradish peroxidase was intracellularly injected into sympathetic preganglionic neurons (SPN) of the third thoracic segment in cats. Seven neurons were reconstructed from serial horizontal or parasagittal sections of the spinal cord. The cell bodies of all neurons were located in the n. intermediolateralis pars principalis (ILp). They were spindle-shaped with the long axis in craniocaudal direction or large and multipolar or small and oval in shape. Preferentially on the cranial and caudal pole of the cell body, five to eight primary dendrites arose from the cell body. Dendritic branches were traced to their terminations at distances up to 1,330 microns from the cell body. The dendritic fields of all SPNs were strictly oriented in the longitudinal direction with a total length of 1,500-2,540 microns. The cranial and caudal dendritic fields were about equal in length but, with one exception, the degree of branching was always greater in the cranial than in the caudal dendritic field. The dendritic fields of all SPNs were primarily restricted to the ILp. In the mediolateral direction it extended from 130 to 360 microns and in the dorsoventral direction from 50 to 180 microns. Only rarely, a higher-order dendrite left the boundaries of the ILp and projected dorsolaterally or laterally into the white matter or ventromedially or medially into the adjacent n. intercalatus. All dendrites showed various forms of spines. At a distance of 132-437 microns from the cell body the axon arose as a direct extension of a process which closely resembled a primary or second-order dendrite. The axons projected ventrally and mostly caudally along the lateral border of the gray matter until they turned laterally at the end of the ventral horn. No axon collaterals were observed.     

Phenylethanolamine N-methyltransferase-containing terminals synapse directly on sympathetic preganglionic neurons in the rat

The ultrastructural morphology as well as neuronal and glial associations of phenylethanolamine N-methyltransferase (PNMT)-containing terminals in the intermediolateral cell column (IML) of the thoracic spinal cord were examined in the rat utilizing the peroxidase-antiperoxidase (PAP) method. The PNMT-immunoreactive terminals were 0.5–1.4 μm in diameter and contained a few mitochondria, a large population of small clear vesicles and from 1 to 6 large dense-core vesicles. The terminals formed synapses primarily with dendrites. The type of axodendritic association (i.e. symmetric or asymmetric) varied with the size of the dendrite, such that the majority of synapses on large dendrites were symmetric and those on smaller dendrites and dendritic spines were asymmetric. Moreover, most of the synaptic associations of PNMT-containing terminals were with the smaller dendritic processes. Many of the PNMT-labeled terminals, as well as their postsynaptic targets, were closely invested with, or apposed to fibrous astrocytic processes. In a subsequent set of experiments, we combined immunoautoradiographic labeling for PNMT with horseradish peroxidase (HRP) retrograde identification of sympathetic preganglionic neurons (SPNs) in the IML to determine whether or not SPNs receive direct synaptic input from the adrenergic terminals. In these sections, PNMT-containing terminals directly synapsed on the HRP-containing (i.e. retrogradely labeled SPNs) perikarya and dendrites. The axosomatic synapses observed between PNMT-labeled terminals and SPN perikarya were exclusively symmetric: whereas the type of axodendritic association varied depending upon the size of the dendrite such that the majority were asymmetric. The findings provide ultrastructurral evidence that in the rat IML, adrenergic (i.e. PNMT-containing) terminals (1) may be either excitatory (asymmetric) or inhibitory (symmetric) depending on their site of termination and (2) can influence sympathetic nerve discharge through a direct effect on the SPN cell membrane.     

Identification of sympathetic preganglionic neurons controlling the small intestine in hamsters using a recombinant herpes simplex virus type-1

Sympathetic preganglionic neurons (SPNs) may be organized topographically within the spinal cord for selective control of visceral organs. We used a recombinant herpes simplex virus type-1 (rHSV-1) to identify SPNs innervating the small intestine in hamsters. These SPNs were distributed bilaterally in the cord from the fifth thoracic spinal segment to the second lumbar segment, but predominantly in thoracic segments 5–10. They had morphology similar to that of renal and adrenal SPNs infected with HSV-1. The majority of intestinal SPNs were found in the intermediolateral cell column, with a few located in the lateral funiculus. The SPNs labelled following duodenal injection of rHSV-1 were in the same spinal segments as the SPNs labelled following jejunal or ileal injections, suggesting lack of a relation between target topography and the topographic organization of these neurons. In addition, intestinal SPNs were located in the same spinal segments, and autonomic nuclei as renal and adrenal SPNs suggesting that SPNs controlling the abdominal viscera are not organized viscerotopically for discrete control of different organs. © 1997 Elsevier Science B.V. All rights reserved.     

Postnatal development of cat hind limb motoneurons. I: Changes in length, branching structure, and spatial distribution of dendrites of cat triceps surae motoneurons

The postnatal development of length, branching structure, and spatial distribution of dendrites of triceps surae motoneurons, intracellularly stained with horseradish peroxidase, was studied from birth up to 44-46 days of postnatal (d.p.n.) age in kittens and compared with corresponding data from adult cats. The number of dendrites of a triceps surae motoneuron was about 12, and the arborization of each dendrite generated an average of 12-15 terminal branches. There was no net change in the number of dendrites of a neuron or in the degree of branching of the dendrites despite the occurrence of both a transient remodeling of the dendritic branching structure and changes of the spatial distribution of the dendritic branches during postnatal development. The perisomatic territory in the transverse plane occupied by the dendritic branches of a motoneuron increased in parallel with the overall growth of the spinal cord. Thus, the relative size of the dendritic territory in this plane was kept almost constant, whereas dendritic branches projecting in the rostrocaudal direction grew much faster than the spinal cord and also became more numerous. At birth the rostro-caudal dendritic span of individual motoneurons bridged 1:6 to 1:5 of the L7 spinal cord segment length; this figure was 1:3 at 22-24 d.p.n. Hence, in this direction, the growing dendritic branches invaded novel dendritic territories. The change in dendritic branch length from birth to 6 weeks of age corresponded to an average growth rate of 2 to 4 microns per dendritic branch and day, which implies that the total increase in length of the dendrites of a neuron could amount to 1 mm/day. The increase in branch length did not occur in a uniform or random manner; instead, it followed a spatiotemporal pattern with three phases: From birth to 22-24 d.p.n., growth was particularly prominent in greater than or equal to 3rd order preterminal and 2nd through 6th order terminal branches. From 22-24 to 44-46 d.p.n., a large increase in branch length confined to terminal branches of greater than or equal to 3rd branch orders was observed. As indicated by topological analysis, this length increase was probably due in part to a resorption of peripheral dendritic branches during this stage of development. From 44-46 d.p.n. to maturity, the increase of dendritic branch length was restricted to preterminal branches of low (less than or equal to 4th) branch order.(ABSTRACT TRUNCATED AT 400 WORDS)     

Morphology of cat phrenic motoneurons as revealed by intracellular injection of horseradish peroxidase

The morphology of phrenic motoneurons (PMs) of adult cat was examined by utilizing the technique of intracellular injection of horseradish peroxidase. Twenty-one cells were reconstructed from serial sections in transverse, sagittal, and horizontal planes. The cell bodies were ellipsoid, with the major diameter oriented parallel to the longitudinal axis of the spinal cord. The dendrites of PMs are not distributed in a radially symmetric fashion, but rather project to four separate fields. The field containing the greatest number of dendrites extends rostrocaudally within the phrenic motor column. This collection of dendrites forms a rostrocaudal bundle in which the dendrites from neighboring PMs lie in close association with one another. The remaining dendrites project dorsolaterally, dorsomedially, and to a lesser extent, ventrally. The dorsolaterally directed dendrites form bundles upon entering the lateral funiculus with the dendrites from other PMs. Several of the dorsomedially directed dendrites cross to the contralateral spinal cord via the anterior commissure or central gray. A wide variety of dendritic spines and appendages was observed. There were no instances in which axon collaterals were observed for the 11 well-stained axons examined. The length of the initial segment of the axon was a function of the distance of the cell body from the ventral funiculus.     

Ontogeny of adrenergic fibers in rat spinal cord in relationship to adrenal preganglionic neurons

Adrenergic neurons in the C1 cell group in the rostral ventrolateral medulla oblongata contain epinephrine, as well as its biosynthetic enzyme, phenylethanolamine N-methyltransferase (PNMT). These neurons send axons to regions of the central nervous system known to regulate autonomic function, including the sympathetic preganglionic nuclei of thoracic and upper lumbar spinal cord. Previous studies have shown that PNMT is expressed in neurons located in the medulla oblongata on embryonic day 14; however, the development of the projections from these cells has not been studied. With the aid of high-performance liquid chromatography (HPLC) to determine levels of catecholamines and immunocytochemistry to demonstrate PNMT, the ontogeny of the adrenergic bulbospinal pathway in the embryonic, postnatal, and adult rat has been studied. In addition, the relationship between PNMT-immunoreactive (IR) fibers and retrogradely labeled sympathetic preganglionic neurons projecting to adrenal medulla are described. PNMT-IR fibers were first observed in the caudal medulla oblongata and lateral funiculus of spinal cord on gestational day 15(E15). On E16, PNMT-IR fibers in the thoracic spinal cord were observed in the intermediate gray matter at the level of the lateral horn. Epinephrine was measureable in spinal cord on E20. Both the density of PNMT-IR fibers and the levels of epinephrine increased to a maximum during the second postnatal week and then declined to adult levels. These observations suggest that a period of adrenergic hyperinnervation of spinal sympathetic nuclei occurs during the neonatal period. PNMT-IR terminals in spinal cord were observed, primarily, although not exclusively, in sympathetic nuclei of thoracic cord and parasympathetic nuclei of upper sacral cord. Adrenergic fibers in the intermediolateral nucleus (IML) and the central autonomic nucleus (CAN) dorsal to the central canal were particularly dense during the second postnatal week in both midthoracic and upper sacral segments. In the neonate, a "ladder-like" pattern of PNMT-IR fiber staining was observed which represented transverse fiber bundles connecting IML with CAN and extensive longitundinal fiber bundles along the border of the funiculus in IML. At all spinal levels, adrenergic fibers were also observed adjacent to the ependyma dorsal or lateral to the central canal. The relationship between adrenal preganglionic neurons and PNMT-IR fibers in IML was examined on postnatal days 4, 15, and 60. With retrograde labeling from adrenal medulla, it was demonstrated that PNMT-IR fibers are associated with adrenal preganglionic neurons throughout postnatal development.(ABSTRACT TRUNCATED AT 400 WORDS)     

Dynamic transformation of Bergmann glial fibers proceeds in correlation with dendritic outgrowth and synapse formation of cerebellar Purkinje cells

Bergmann glia (BG) are unipolar cerebellar astrocytes, whose radial (or Bergmann) fibers associate with developing granule cells and mature Purkinje cells (PCs). In the present study, we investigated the morphodifferentiation of BG by immunohistochemistry for glutamate transporter GLAST and electron microscopy. GLAST was expressed widely in cerebellar radial glia/astrocytes during fetal and neonatal periods and became concentrated in BG postnatally. During the second postnatal week when PC dendrites grow actively, GLAST immunostaining revealed dynamic cytologic changes in Bergmann fibers in a deep-to-superficial gradient; Bergmann fibers traversing the external granular layer were stained as rod-like fibers, whereas in the molecular layer, the rod-like pattern was gradually replaced with a reticular meshwork. At postnatal day 10, the superficial rod-like domain was composed of glial fibrillary acidic protein (GFAP)-positive/GLAST-positive straight fibers, forming cytoplasmic swellings and short filopodia. Along this domain, the tip of growing PC dendrites ascended vertically and entered the base of the external granular layer. The deeper reticular domain of Bergmann fibers was characterized by active expansion of GFAP-negative/GLAST-positive lamellate processes, which surrounded PC synapses almost completely. Therefore, the transformation of Bergmann fibers proceeds in correlation with dendritic differentiation of PCs. The intimate PC-BG relationships during cerebellar development raise the possibility that a preexisting glial shaft could serve as a structural substrate that directs dendritic outgrowth toward the pial surface, whereas the successive formation of a reticular glial meshwork should lead to structural maturation of newly formed PC synapses.     

Enkephalin-immunoreactive interneurons extensively innervate sympathetic preganglionic neurons regulating the pelvic viscera

Enkephalin (ENK)-immunoreactive (IR) axons occur in regions containing spinal autonomic neurons and endogenous opiates contribute to spinal regulation of bladder function. To identify possible spinal sites of opiate action, we used immunocytochemistry for ENK with retrograde tracing from the major pelvic ganglion (MPG), a key location for postganglionic neurons controlling pelvic viscera, with cholera toxin B subunit (CTB) or CTB-horseradish peroxidase (CTB-HRP). We compared the relationship of ENK-IR axons with sympathetic preganglionic neurons (SPNs) projecting to the MPG between intact spinal cords and cords with 2- or 11-week complete transections between thoracic segments 4 and 5. By light microscopy, sections of intact cord showed dense networks of ENK-IR axons surrounding CTB-IR SPNs in the intermediolateral cell column (IML), intercalated nucleus, and central autonomic area of lower thoracic and upper lumbar cord. This staining pattern was similar in rats with 2- or 11-week transections. Ultrastructurally, ENK-IR axons formed synapses on SPNs in all three autonomic subnuclei of intact cord. In the IML, ENK-IR varicosities contributed 52% of the synapses on the somata of MPG-projecting SPNs. In 2-week transected cord, synapses from ENK-IR axons persisted on SPNs and the proportion of input to IML SPNs had increased to 67%, probably reflecting loss of supraspinal input. These results suggest that endogenous opioids could play a major role in controlling sympathetic outflow to the bladder through a direct action on SPNs. The persistence of the dense ENK innervation after complete cord transection indicates that the ENK-IR input to SPNs arises predominantly from intraspinal sources.     

Spatiotemporal distribution of neuronal calcium sensor-1 in the developing rat spinal cord

The present study revealed the localization of neuronal calcium sensor (NCS)-1 immunoreactivity (IR) in the developing rat spinal cord. The NCS-1 IR first appeared at embryonic day 12 in the peripheral nerves and their somata. Intense NCS-1 IR was expressed in ascending and descending tracts in the white matter during the late prenatal period, which gradually decreased to the faint level during postnatal development. Intense NCS-1 IR was colocalized with growth associated protein (GAP)-43 IR in the marginal zone and with the glutamate-aspartate transporter (GLAST) IR in the radial processes traversing the marginal zone. In the adult rat white matter, radially oriented astrocytes and astrocytes in the glia limitans were double-labeled for NCS-1 and glial fibrillary acidic protein (GFAP), whereas small dots on finger-like dendritic projections were double-labeled for NCS-1 and synaptophysin. In the developing gray matter, the NCS-1 IR appeared at embryonic day 12 and gradually increased in the neuronal somata and neuropil, reaching a plateau after the end of the 4th postnatal week. The small dots in neuropil were colabeled for NCS-1 and GFAP or NCS-1 and synaptophysin in the adult rat gray matter. These results strongly suggest that NCS-1 is involved in axogenesis and synaptogenesis in the developing rat spinal cord. NCS-1 can serve as a Ca(2+)-sensor not only in neurons but also in radial glial cells or even in radially oriented astrocytes in the adult rat spinal cord.     

Motoneuron morphology in the dorsolateral nucleus of the rat spinal cord: Normal development and androgenic regulation

The rat lumbar spinal cord contains two sexually dimorphic motor nuclei, the spinal nucleus of the bulbocavernosus (SNB), and the dorsolateral nucleus (DLN). These motor nuclei innervate anatomically distinct perineal muscles that are involved in functionally distinct copulatory reflexes. The motoneurons in the SNB and DLN have different dendritic morphologies. The dendrites of motoneurons in the medially positioned SNB have a radial, overlapping arrangement, whereas the dendrites of the laterally positioned DLN have a bipolar and strictly unilateral organization. During development, SNB motoneuron dendrites grow exuberantly and then retract to their mature lengths. In this experiment we determined whether the adult difference in SNB and DLN motoneuron morphology was reflected in different patterns of dendritic growth during normal development. Furthermore, the development of both these nuclei is under androgenic control. In the absence of androgens, SNB dendrites fail to grow; testosterone replacement supports normal dendritic growth. Thus, we also examined the development of DLN dendrites for similar evidence of androgenic regulation. By using cholera toxin-horseradish peroxidase (BHRP) to label motoneurons retrogradely, we measured the morphology of DLN motoneurons in normal males, and in castrates treated with testosterone or oil/blank implants at postnatal day (P) 7, P28, P49, and P70. Our results demonstrate that in contrast to the biphasic pattern of dendritic development in the SNB, dendritic growth in the DLN was monotonic; the dendritic length of motoneurons increased more than 500% between P7 and P70. However, as in the SNB, development of DLN motoneuron morphology is androgen-dependent. In castrates treated with oil/blank implants, DLN somal and dendritic growth were greatly attenuated compared to those of normal or testosterone-treated males. Thus, while androgens are clearly necessary for the growth of motoneurons in both the SNB and DLN, their different developmental patterns suggest that other factors must be involved in regulating this growth. © 1993 Wiley-Liss, Inc.     

Purkinje cell dendrites grow in alignment with Bergmann glia

The pattern of growth of Purkinje cell dendrites has been analyzed and related to their interactions with Bergmann glial radial processes. In cerebellar slice cultures from mice expressing green fluorescent protein (GFP) under the glial fibrillary acidic protein (GFAP) promoter, Purkinje cells were transfected and imaged with two-photon microscopy over 2 days. We report that while the Purkinje cell dendritic tree grows, individual dendrites increase or decrease in length. Importantly, we demonstrate that vertical growth of Purkinje cell dendrites occurs primarily in alignment with radial glial processes. These findings suggest that radial glial processes provide a structural substrate for the directional growth of Purkinje cell dendrites, thus influencing the shape of the dendritic tree.     

The development of radial glia and radial dendrites during barrel formation in mouse somatosensory cortex

The development of the mouse barrel field (the mystacial whisker representation in SI cortex) was examined using immunocytochemical probes for radial glia and neuronal dendrites. The maturing dendrites were revealed using antibodies against microtubule-associated protein 2 (MAP2) and the radial glia were demonstrated with a recently described monoclonal antibody, RC2. By postnatal day 7 both antibodies clearly demonstrated a non-uniform distribution of dendrites and glia that was unique to layer IV of the barrel field. Both MAP2-immunoreactive dendrites and RC2-immunoreactive radial glial fibers were dense near the walls (sides and septae) of barrels than near the hollows (centers) of barrels. In contrast, in other cortical regions, radial glia and dendrites did not appear obviously patterned. Not until postnatal day 4 did the pattern of both radial glial fibers and apical dendrites begin to emerge in a barrel-like distribution. We conclude that the non-uniform distribution of radially oriented dendrites and radial glial fibers appears with a similar developmental time course to that described for the appearance of the cellular barrels themselves.     

Postnatal development of cell columns and their associated dendritic bundles in the lumbosacral spinal cord of the rat. I. The ventrolateral cell column

The postnatal development of the ventrolateral dendrite bundle (LDB) in the rat lumbosacral cord was studied quantitatively with Golgi-Cox impregnation. At birth, motoneuronal perikarya and their dendrites were not fully developed, and had not begun to form bundles; varicose dendritic shafts radiated symmetrically from motoneurons. Dendrites contained numerous spines and growth cones. At 5 days, dendritic shafts began to arrange themselves longitudinally, and by 10 days of age, dendrite bundling was apparent. Dendritic growth and bundling appeared complete by two months of age. LDB formation was a dynamic process; rapid dendritic growth occurred in discrete phases with brief intervals of slower dendritic development between them. The mean number of secondary and tertiary dendrites, and the mean branch length of all orders progressively increased. Motoneurons of the LDB primarily innervate the pelvic musculature. Selective horizontal orientation of dendrites into discrete compact bundles suggests that the LDB may serve as a specialized receiving and integrating system for autonomic control over excretory and reproductive functions. It is interesting to note that in patients suffering from amyotrophic lateral sclerosis, motoneurons in the LDB are resistant to destruction. This finding suggests that motoneurons in the LDB may express unique features that protect them from certain disease processes. A better understanding of the developmental anatomical, physiological and biochemical properties of the LDB may provide insight into the treatment of patients with disease processes involving spinal cord and brainstem lower motoneurons.