Search

We use opossums Monodelphis domestica to study the development of mammalian motor systems. The immature forelimbs of the newborn perform rhythmic and alternating movements that are likely under spinal control. The hindlimbs start moving in the second week. Chemical synapses are scant in the spinal enlargements of neonatal opossums and the presence of electrochemical synapses has not been evaluated in this species or in other marsupials. As a first step aiming at evaluating the existence of such synapses in the neonatal spinal cord, we have investigated the presence of the exclusively neuronal gap junction protein connexin36 (Cx36) by immunohistochemistry in light microscopy. At birth, Cx36 immunoreactivity is moderate in the presumptive gray matter in both enlargements. Thereafter, it decreases gradually, except in the superficial dorsal horn where it increases to a plateau between P10 and P20. Cx36 labeling is detected in the presumptive white matter at birth, but then decreases except in the dorsal part of the lateral funiculus, where it is dense between P10 and P20. Cx36 has become virtually undetectable by P52. The presence of Cx36 in the spinal enlargements of postnatal opossums suggests that neurons might be linked by gap junctions at a time when chemical synapses are only beginning to form. The greater abundance of Cx36 observed transiently in the superficial dorsal horn suggests a stronger involvement of this protein in spinal sensory systems than in direct motor control of the limbs.

Using immunohistochemistry in light microscopy, the myelin basic protein and proteolipid protein were localized on sections of the spinal cord enlargements of opossums, Monodelphis domestica, to determine the timecourse of myelinogenesis therein and compare it with other events of motor systems development. Additional tissue not processed for immunohistochemistry was prepared for transmission electron microscopy. No immunolabeling for either protein occurred on spinal sections from the newborn opossum, but in electron microscopy occasional fibers surrounded by loose, irregular membranous rings were seen on the outskirts of the ventral horn. Immunolabeling was detected first in the brachial enlargement during the second week, presumably on motoneuronal, vestibular and reticular axons. The areas of the dorsal columns, other spino-encephalic, reticulospinal and propriospinal projections became labeled in the third week, and the area of rubrospinal axons at 4 weeks. In the brachial gray matter, immunolabeling appeared along ventrodorsal and lateromedial gradients from the fourth to seventh weeks. Labeling developed similarly in the white and gray matter of the lumbosacral enlargement, but 3–5 days later than at brachial levels. Labeling intensity in the white and gray matter increased until at least 4 months, but remained light in laminae I–III. Thus, myelinogenesis in the spinal cord enlargements of the opossum is protracted and follows general rostrocaudal, ventrodorsal and lateromedial sequences. It occurs later than synaptogenesis at comparable levels of the cord, but earlier than myelinogenesis in the corresponding ventral and dorsal roots. Spinal myelinogenesis correlates with the development of sensorimotor reflexes, weight support and quadrupedal locomotion.

Synaptic proteins were localized in light microscopy on sections of the brachial and lumbosacral enlargements of the spinal cord of postnatal opossums, Monodelphis domestica, to determine whether their expression correlates with the development of major motor pathways and simple motor behaviors. The tissues were fixed, cryoprotected, frozen, cut in 15-micrometer sections, and processed immunohistochemically using antibodies against synaptophysin, synaptotagmin-I, or SNAP-25. Immunolabeling was observed in the presumptive white matter before the presumptive gray matter, suggesting that the proteins are evidenced in growing axons before the onset of synaptogenesis, and it was observed in presumed propriospinal axons before most presumed descending axons of supraspinal origin. In the newborn opossum, the immunolabeling was scant in the gray matter and was limited to the periphery of the ventral horn, and indeed few synapses were seen in electron microscopy in nonexperimental material. Labeling increased in intensity and spread throughout the gray matter until 5-7 weeks, when it was no longer found in the white matter and resembled the adult pattern of labeling. Considering the location and relative intensity of the immunolabeling for the three proteins over time in the two enlargements, synaptogenesis occurs according to three general gradients: rostrocaudal, ventrodorsal, and lateromedial. These gradients match those of spinal cord and limb development, and of the growth of descending axons into the cord. Synaptogenesis is most intense when the spinal sensorimotor reflexes begin to be expressed.

Dextran-amines were used as retrograde tracers to investigate the organization of cortical projections to different cytoarchitectonic regions of the pontomedullary reticular formation of the cat. Injections into the nucleus reticularis pontis oralis resulted in labelling of neurones in the proreus cortex and area 6a|iB of the premotor cortex, with little labelling in the motor cortex (area 4). This labelling was predominantly ipsilateral to the injection site. In contrast, injections into the nucleus reticularis pontis caudalis (NRPc), nucleus reticularis gigantocellularis (NRGc), and nucleus reticularis magnocellularis (NRMc) resulted in bilateral labelling—primarily in areas 6aβ, 6aγ, and in the rostromedial region of area 4—with little labelling in the proreus cortex. In general, the cortical projections to the caudal NRGc and the NRMc were larger than those to the NRPc. More than 25% of the total projections to each of the latter three reticular regions arose from the medial part of area 4. Labelling in the hindlimb regions of area 4 was largest following the NRMc injections and smallest after injections in the NRPc. The projections to the NRPc originated from more medial parts of areas 4 and 6 than did the projections to the caudal region of the NRGc. These results suggest that areas 4 and 6 may be able to differentially activate different regions of the pontomedullary reticular formation depending on the movement that is made and perhaps also on the context of that movement. J. Comp. Neurol. 388:228–249, 1997. © 1997 Wiley-Liss, Inc.

We have quantified the number and size of myelinated fibers of the ventral and dorsal roots of selected segments that innervate the forelimbs (C8) and hindlimbs (L4) in the developing opossum, Monodelphis domestica. The gray opossum was chosen because it is born very immature and its somatomotor development occurs almost entirely postnatally. After aldehyde fixation, osmium postfixation, and resin embedding, the roots were cut transversely (1.5 pm), stained with toluidine blue, and observed and photographed by using light microscopy. The counts and measurements were made with a digitizing table. Myelination of the C8 and L4 roots begins during the second week of life and occurs according to two gradients: rostrocaudal and ventrodorsal. The number of myelinated fibers in these roots increased over approximately 7 weeks after which an excess, compared with their adult value, was recorded during the following weeks in three of the four roots. The supernumerary myelinated fibers are presumed to be collaterals. The fibers increased in diameter until at least 98 days. The classification according to size for the ventral roots (alpha and gamma) became evident in the fourth week, but that in types I, II, and III for the dorsal roots was never clear. There was no significant difference in the number and size distribution of myelinated fibers between sexes until late in development. The fibers innervating the limbs thus become myelinated postnatally in the opossum, a process that occurs over a protracted period and that continues after sensorimotor reflexes and locomotion appear adult-like.

The development of spontaneous locomotor behaviors was studied in the opossum Monodelphis domestica. The newborn opossum performs alternate, rhythmic movements with its forelimbs to crawl on the mother's belly where it attaches to a nipple, and its hindlimbs are little more than embryonic buds. The forelimbs retain the above movements for about 3 weeks, while the hindlimbs begin to move late in the second week. When detached from the nipple at 2-3 weeks, the pup can support its weight on the forelimbs and pivot around its hindquarter. Around the fourth week, the young can detach from the mother, its hindlimbs can support weight and linear locomotion appears, but the four limbs are not well coordinated. However, it can swim with coordinated movements of all limbs. Coordination when walking appears around the sixth week. During development, the duration of the step cycle decreases significantly. The durations of the stance and swing phases of the step cycle decrease in absolute terms, but swing increases as a percentage of the step cycle. The results are discussed in relation to the development of nervous and skeletomuscular components as well as sensorimotor reflexes.

Like other marsupials, the opossum Monodelphis domestica is born very immature and crawls, unaided by the mother, from the urogenital opening to a nipple where it attaches and pursues its development. If the alternate, rhythmic movements of the forelimbs which allow this locomotion are generated by the developing spinal motor networks, sensory information is nonetheless needed to guide the newborn to a nipple. Behavioral, anatomical and physiological studies suggest that the auditory and the visual systems are insufficiently developed in newborn opossums to influence spinal motor centers, while the vestibular, trigeminal, and olfactory systems are likelier candidates. The trigeminal, vestibular and olfactory regions of the brain were electrically stimulated to test their relative effectiveness at eliciting forelimb movement in newborn opossums, using in vitro preparations of brain-spinal cord with the limbs attached. The minimal stimulation of the cervical spinal cord needed to induce forelimb movement was considered as threshold (T). Stimulations of the trigeminal ganglion (5G) at ∼2 T and of the vestibular complex at ∼20 T could induce the same movement, and were not statistically different, in contrast to the ∼600 T necessary for the olfactory bulb (OB). Neurofilament-200 immunohistochemistry and retrograde tracing with Texas-Red conjugated Dextran Amines were used to study trigeminal innervation of the facial skin and pathways by which trigeminal inputs may be relayed to the spinal cord. Numerous nerve fibers were observed in the snout dermis, especially in the maxillary region, but also elsewhere in the head skin. Some 5G cells project to the upper spinal cord, but more project to the caudal medulla where they could contact secondary trigeminal neurons or reticular cells projecting to the spinal cord. These results support a significant influence of the trigeminal and the vestibular systems, but not of olfaction, on forelimb movement of neonatal opossums.

The development of a number of sensorimotor reflexes was studied in the Brazilian opossum Monodelphis domestica. At birth, an opossum's forelimbs execute rhythmic, alternate movemen

The origin of the long descending propriospinal (LDP) projections have been studied in adult and developing opossums, Monodelphis domestica. This species has been chosen because of the considerable immaturity of the hindlimbs at birth, the postnatal appearance of their motility and the late development of coordination between them and the forelimbs. Neuroanatomical tracing has indicated that some LDP projections form postnatally. The ones present at birth arise from the regions of the cord where they are the most numerous in the adult opossum, presumptive laminae VII and VIII of the brachial enlargement. Subsequently, LDP projections arise from neurons located in adjacent laminae (IV to VI and IX and X) and at more rostral cervical levels. The origin of LDP projections in the adult opossums generally matches that reported for other mammals. These long propriospinal projections are in place well before the behavioral appearance of coordination between the hindlimbs and the forelimbs, but the timing of their synaptogenesis is not yet known.

A recently proposed four-parameter generalization of the Gompertz curve (Jolicoeur et al., 1992a) is used to analyze chronological growth and allometry in the stature, head circumference, and brain weight of 101 male and 95 female fetuses and children autopsied at the Sainte-Justine Hospital for children in Montreal. Although asymptotic stature cannot be determined accurately because the sample does not include adults, the data nevertheless suggest that head circumference and brain weight terminate most of their growth much earlier than stature. This conclusion is also supported by the fact that, whereas the log-log relationship between head circumference and brain weight is almost straight and closely approximates simple allometry, the log-log relationships between stature, on one hand, and head circumference or brain weight, on the other hand, are markedly curved, indicating complex allometry. Applied to 78 males and 73 females whose total age is between 20 weeks and 1.5 years, the confidence intervals of the slope of the major axis of logarithmically transformed data indicate that the allometry exponent of brain weight relative to head circumference is somewhat higher than 3.0, which may reflect changes in head proportions. © 1993 Wiley-Liss, Inc.

Most behavioral studies on the ontogeny of sensorimotor reflexes and locomotion were done in quadrupedal species with equally developed forelimbs (FL) and hindlimbs (HL). In contrast, the Mongolian gerbil has long and strong HL but relatively small FL, indicating their differential use for locomotion. We have used the gerbil to study the ontogeny of a number of reflexes and locomotor acts to see if their sequence of appearance, their maturation, as well as their intensity of expression, differed from some other mammals. The following sequence was observed: forward FL hopping, FL grasp, forward HL hopping, surface body righting, chin tactile placing, lateral FL hopping, lateral HL hopping, medial FL hopping, medial HL hopping, visual placing, air body righting, and FL and HL tactile placing. When comparing FL and HL, a given reflex does not necessarily appear earlier in the limb that expresses it more strongly. The results are discussed in relation to the development of the central nervous system, the limb, and the locomotor behavior.

The ontogeny of a number of sensorimotor reflexes has been studied in the Mongolian gerbil. In contrast to a number of other mammals, the gerbil has relatively long and strong hindlimbs but small forelimbs, indicating their different importance for a number of locomotor acts, and during the developmental period studied, the hindlimbs grow at a much faster rate than the forelimbs. The following sequence of appearance and maturation of the reflexes was observed: rooting, forelimb hopping, surface body righting, forelimb grasp, hindlimb hopping, chin tactile placing, visual placing, air body righting and, at the same time, forelimb and hindlimb tactile placing. This sequence concords with the general gradient of development and maturation of the spinal and brain centers subserving these reflexes, as evaluated from Nissl preparations. The results indicate that there is no clearcut rostro-caudal gradient of postnatal maturation of the spinal cord and the spinally mediated reflexes, but that there is a general caudo-rostral gradient of brain maturation and of the brain-mediated reflexes. Comparisons with other mammals are made.

The development of limb motor innervation was studied in the opossum Monodelphis domestica, a marsupial born with immature mobile forelimbs and immobile hindlimbs. Choline acetyltransferase (ChAT), the synthesis enzyme of acetylcholine, was evidenced on sections of the spinal enlargements, and the protein that transports acetylcholine (VAChT) on limb sections. In newborn, ChAT immunolabeling occurred in small, undifferentiated neurons of the ventral horn, presumably motoneurons, and intermediate and dorsal gray matter, and in the presumptive white matter, all less abundant at lumbosacral than brachial levels. Scant immunolabeling for VAChT marked small terminal-looking profiles, presumably growth cones or immature neuromuscular junctions, decreasing proximodistally in each limb and being less abundant in hindlimbs than forelimbs; it was absent distally in the foot where no muscle tissue was formed. ChAT labeling disappeared from the white matter within 1 week while cholinergic neurons increased in number and size. Motoneurons segregated in a medial and lateral group by 4–5 weeks. VAChT-labeled profiles increased in number and size and they flattened along a proximodistal gradient within each limb, but later in the hindlimbs than in the forelimbs. Labeling appeared in distal foot muscle at 1 week. The density, size, and shape of terminals became comparable in all segments of a given limb by 3–4 weeks. Their number and size increased, and by 8 weeks, they clustered in 3 or 4 along muscle fibers. Thus, limb motor innervation develops largely postnatally in the opossum, along rostrocaudal and proximodistal gradients. Its timecourse is compared to the development of motor behaviors.

Vesicular acetylcholine transporter (VAChT) was revealed immunohistochemically in light microscopy on hindlimb sections of developing opossums, Monodelphis domestica. In the immobile hindlimbs of the newborn, which comprise cartilaginous bones and loose, unstriated myofibers, scant immunolabeled nerve segments and small spherical terminals, presumably growth cones or immature neuromuscular junctions, are found in the muscle tissue of the thigh, leg and proximal foot, decreasing in number and size proximodistally. When the hindlimbs start moving at 1 week, terminals are more numerous and larger, still decreasing proximodistally, and occur in the newly formed interosseous foot muscles. At 4 weeks, when the hindlimbs start supporting weight and quadrupedal locomotion appears, terminals are more numerous, flattened and in comparable size and density in thigh, leg and foot muscles. By 7 weeks, large and completely flat terminals are observed in groups of 3 to 4 at regular intervals along muscle fibers. VAChT expression develops largely postnatally in the opossum hindlimbs, along a proximodistal gradient that parallels somatic and reflex development.

The opossum Monodelphis domestica is a marsupial born considerably immature 14-15 days after conception. It is possible to study postnatally, in this species, almost the entire development of its motor behaviors as well as of the nerve centers involved in their control. The lumbosacral spinal cord of the newborn comprises a thick ventricular zone containing mitotic figures, an intermediate zone of small and undifferentiated cells, and a thin marginal zone. The hindlimbs are little more than embryonic buds. The presumptive bones consist of cartilageneous or mesenchymal condensations and the presumptive muscles of immature myofibers mixed and surrounded with mesenchyme. Cholinergic fibers from lumbosacral motoneurons are already seen among the myofibers, but most of hindlimb motor innervation develops postnatally. The long descending and ascending projection systems connecting the lumbosacral enlargement to the cervical cord and the encephalon also form largely postnatally, but lateral vestibular and medullary reticular axons are present in the lumbosacral cord at birth. Synaptogenesis in the lumbosacral enlargement occurs largely postnatally, according to a general outside-in gradient, and the earliest evidence for it is on lateral motoneurons. Myelinogenesis therein is even later. These observations on neural development are correlated with observations on the development of simple reflex behaviors and locomotion.

The presence of synaptophysin in the lumbosacral enlargement of developing opossums, Monodelphis domestica, was studied immunohistochemically at the light microscopic level. In newborn, synaptophysin-labeling was observed in the presumptive white matter, presumably in growing axons, and was scant in the ventrolateral gray matter. Over the next 3 weeks the labeling filled the gray matter following a general ventrodorsal gradient. Labeling was found in the white matter until the fifth week. Synaptogenesis in the lumbosacral enlargement of the opossum thus occurs mostly postnatally, when many descending axons have already reached that level. It is particularly intense in the ventral horn when the hindlimbs begin to move, and in the dorsal horn when sensorimotor reflexes can be elicited.

The present study was designed to determine whether the transmission of cutaneous afferent information from the limbs to the medullary reticular formation is phasically modulated during locomotion. Experiments were carried out in three chronically prepared, intact cats in which nerve cuff electrodes were placed, bilaterally, on the superficial radial and the superficial peroneal nerves. Thirty-seven reticulospinal neurones (RSNs) were identified by stimulation of their axons in the lumbar spinal cord (L2); 29 of 37 of these were recorded with the cat at rest, 28 of 37 during locomotion and 20 of 37 both at rest and during locomotion. Low-threshold stimulation of the cutaneous nerves evoked excitatory responses in the majority of RSNs both at rest and during locomotion. In the 28 of 37 RSNs recorded during locomotion, it was possible to record the evoked response to stimulation of all four limb nerves, giving a total of 184 tested cases [RSNs tested x number of nerves stimulated x phase of stimulation (swing or stance)]. The responses of most RSNs to cutaneous stimulation were modulated in a phase-dependent manner during locomotion. The maximal responses in most, but not all, cases were obtained during the swing phase of the limb that was stimulated and were largely independent of the discharge pattern of the cell. We interpret this result as indicating that the efficacy of transmission of the afferent information is determined more by the excitability of the spinal relay neurones than by the level of excitability of the RSNs in the brainstem. It is suggested that the base discharge pattern of RSNs might be largely determined by their central afferent input, while peripheral afferent inputs would primarily serve to modify the RSN discharge pattern in response to perturbations.

The gray short-tailed opossum Monodelphis domestica is born in a very immature state (eyes and ears closed, budlike hindlimbs, etc.) 14 days post-coitum, but it can locomote with its forelimbs from the mother's genital aperture to a nipple to which it attaches. The forelimb movements allowing this behavior may be the expression of central pattern generators in the spinal cord, but sensory clues must guide it. One such system may be the vestibular system, which senses linear and angular acceleration and has a strong influence on posture and balance at rest and during locomotion in adult animals. Using the neuronal tracer DiI, we have looked at the vestibular primary afferents and the vestibular nuclear projections to the cervical spinal cord in newborn and postnatal opossums, as well as in the adult animal. The projections in the adult opossum conform to those described for other mammals. Fibers of the vestibular portion of the eighth nerve distribute to all four vestibular nuclei and toward the cerebellar primordium on the day of birth. In addition, some of the fibers project to the contralateral vestibular ganglion, a projection that is not found in the adult opossum. Projections from the lateral, medial and inferior vestibular nuclei to the cervical cord are also present in the newborn. Although we cannot exclude the possibility that chemical and/or tactile guidance is used for directing the movements of the newborn opossum, our results support the hypothesis that the vestibular system may be directly involved in the control of these movements. However, not all components of the system are equally developed at birth, and the circuit from the utricle to the lateral vestibular nucleus and from the latter to the cervical cord may be better formed than that from the semicircular canals to the medial and inferior vestibular nuclei to the cervical cord.

A quantitative light-microscopic analysis of the ventral and dorsal roots at the C8 and L4 segments of the spinal cord was made in the opossum, Monodelphis domestica, to evaluate the number of myelinated fibers and their class distribution, and will serve as a baseline to a study of myelinogenesis in that species. In male opossums, the C8 ventral root comprises an average of 595 myelinated fibers (70.3% a:29.7%y), and the dorsal root 1,124 fibers (29.4% type I: 41.2% type 11:29.4% type III). The L4 ventral root has an average of 831 fibers (57.9% a:42.1%y), and the dorsal root 2,079 fibers (17.4% type 1:44.7 type II: 37.9% type III). The females have less fibers but their size and class proportions are comparable to those of the males. These data are discussed in relation to peripheral innervation and are compared to those reported for the rat and the cat.

Retrograde and orthograde transport techniques show that the nucleus reticularis gigantocellularis pars ventralis and the nucleus reticularis gigantocellularis project the entire length of the spinal cord. Double-labelling methods show that some of the neurons in each area innervate both cervical and lumbar levels. There is evidence, however, that neurons in the lateral part of the nucleus reticularis gigantocellularis pars ventralis and the dorsal extreme of the nucleus reticularis gigantocellularis project mainly to cervical and thoracic levels. The autoradiographic method shows that the above nuclei supply direct innervation to somatic and autonomic motor columns as well as to laminae V–VIII and X. The nucleus reticularis gigantocellularis pars ventralis provides additional projections to lamina I and the outer part of lamina II. Several areas of the medullary reticular formation project mainly, and in some cases exclusively, to cervical and thoracic levels. These areas include the nucleus reticularis parvocellularis, the nucleus reticularis lateralis, the nucleus retrofacialis, the nucleus ambiguus, the nucleus lateralis reticularis, caudal parts of the nuclei reticularis medullae oblongatee dorsalis and ventralis, and the nucleus supraspinalis. Autoradiographic experiments reveal that neurons in the ventrolateral medulla, particularly rostrally (the nucleus reticularis lateralis and neurons related to the nucleus lateralis reticularis), innervate sympathetic nuclei. Our results indicate that spinal projections from bulbar areas of the reticular formation are more complicated than previously supposed. Axons from separate areas project to different spinal levels and in some cases to different nuclear targets. These data are in conformity with the evolving concept of reticular heterogeneity.

The retrograde transport of fluorescent markers has been combined with the glyoxylic acid and Falck-Hillarp techniques to identify the origin of monoamine axons within the spinal cord of the North American opossum. Catecholamine axons arise from neurons located within the ventrolateral medulla, dorsal to the superior olivary complex, within the dorsolateral and rostrolateral pons and within the periventricular nuclei of the hypothalamus. Such neurons are most numerous within the dorsolateral pons where they are found dorsal and lateral to the motor trigeminal nucleus, within the nucleus locus coeruleus pars alpha and adjacent reticular formation as well as within the ventral part of the nucleus locus coeruleus. Neurons containing the fluorescent marker and catecholamines were interspersed with others containing only the injected marker with the possible exception of the nucleus locus coeruleus. Spinal axons of the indoleamine type arise from neurons within the nuclei pallidus, obscurus and magnus raphae, the nucleus reticularis gigantocellularis, the nucleus reticularis gigantocellularis pars ventralis, the nucleus reticularis pontis pars ventralis and the nucleus dorsalis raphae. The latter nucleus only innervates rostral cervical levels. Most of the above areas also contain many non-indoleamine neurons which were labelled by the injected marker. This was particularly true of the nucleus magnus raphae and the adjacent nucleus reticularis pontis pars ventralis after injections of fluorescent markers into the superficial dorsal horn.

The North American opossum is born 12 days after conception and is therefore available for experimental manipulation in an immature state. We have used the opossum to study the growth of cortical axons into the brainstem and spinal cord and have obtained evidence that such growth occurs in an orderly fashion. Cortical axons reach the ventral mesencephalon 12 days after birth and some of them have grown into the caudal medulla where they decussate by 23 days. At the latter stage immature cortical axons also distribute to the midbrain tegmentum, the basilar pons, the inferior olive and the hilum of the nucleus cuneatus. Cortical axons first enter the spinal cord about 30 days after birth where they are present in the white matter before growing into the dorsal horn. The forelimb placing reaction does not develop until well after cortical axons have reached cervical levels. Axons from the cerebral cortex grow into the spinal cord before there is evidence for cortical innervation of either the red nucleus or the bulbar reticular formation and well before pyramidal cells of the neocortex are mature. The relatively late development of corticospinal and corticobulbar systems contrasts markedly with the early growth of bulbospinal axons.

The North American opossum is born in a very immature state, 12 days after conception, and climbs into an external pouch where it remains attached to a nipple for an extended period of time. We have taken advantage of the opossum's embryology to study the development of brainstem and cerebellar projections to the diencephalon as well as the timing of diencephalic projections to somatosensory motor areas of neocortex. The techniques employed included immunocytochemistry for serotonin, the retrograde and orthograde transport of wheat germ agglutinin conjugated to horseradish peroxidase, and the selective impregnation of degenerating axons. Our results suggest that serotoninergic axons, presumably from the dorsal raphe and superior central nuclei, are present in the diencephalon at birth. Axons from the bulbar reticular formation, the vestibular complex, the trigeminal sensory nuclei, and the dorsal column nuclei reach at least mesencephalic (and probably diencephalic) levels by postnatal day (PND) 3, whereas those from the cerebellar nuclei may not grow into comparable levels until PND 5. The dorsal column and cerebellar nuclei innervate the ventral nuclei of the thalamus by estimated postnatal day (EPND) 17 and all of the diencephalic nuclei supplied in the adult animal by EPND 26. Diencephalic axons enter ventrolateral (face) areas of presumptive somatosensory motor cortex by PND 12, but do not reach dorsomedial (limb) regions until EPND 21. At both ages, diencephalic axons are limited to the cortical subplate and marginal zone; they do not innervate an identifiable internal granular layer until considerably later. Our results suggest that axons from the brainstem and cerebellum grow into the diencephalon early in development, but that they do not influence the cerebral cortex until relatively late. When the results of the present study are compared with those reported previously on the development of ascending spinal (Martin et al., '83) and corticofugal (Martin et al., '80; Cabana and Martin, '85b, c) projections, it appears that specific components of major somatosensory and motor circuits develop according to different timetables.

The North American opossum does not have a corpus callosum; neocortical commissural axons are contained entirely within the anterior commissure. We have used axonal transport techniques to study the origin and distribution of commissural axons from somatic motor-sensory cortex in developing and adult opossums. Neocortical axons grow into the anterior commissure by postnatal day (PND) 12, the contralateral external capsule by approximately PND 19, the area deep to the contralateral homotypic cortex by approximately PND 26 and the cortex proper by approximately PND 35. Commissural neurons were first demonstrated at about PND 26, when they form a fairly continuous band in the cortical subplate (presumptive layers V-VI). By at least PND 37, commissural neurons are also present in layers II and III, where they form a continuous band, and in layer IV, where they are sparse. In older pouch young and adult opossums the bands of commissural neurons, especially in layers V-VI, are interrupted, and commissural neurons are rare in layer IV. In general, commissural axons in both pouch-young and adult opossums innervate areas containing commissural neurons as well as layer I. In the acallosal opossum as well as in the callosal rat, the development of commissural connections from somatic motor-sensory cortex is characterized by pauses during the growth of axons into the opposite cortex, by a general inside-out-gradient, and by a transition from continuous bands to patchy, radial columns of commissural neurons and axons. This suggests that similar mechanisms govern the formation of commissural connections in the two species.

Studies using axonal transport techniques on the North American opossum show that rubral neurons innervating the cervical cord are not distinctly separated from those which project to lumbar levels. This absence of clear rubrospinal somatotopy contrasts with that described for the placental mammals studied to date. Use of fluorescent markers in double-labelling experiments shows that most rubral neurons in the opossum still innervate either the cervical or lumbar enlargement alone, but that some supply collaterals to both levels.

The origin of brain stem neurons giving rise to axons innervating the thoracic cord has been determined in a developmental series of pouch-young opossums. The first spinal axons identified (17 days after conception, postnatal day 5) arise from the medullary and pontine reticular formation, certain raphe nuclei, the vestibular nuclei and the coeruleus complex. The contribution of the red nucleus before postnatal day 10 is not certain. The brain stem sensory relay nuclei and the hypothalamus do not project spinalwards until much later in development.

The use of fluorescent markers in double-labeling experiments reveals the presence of reticular and raphe neurons in the opossum's brainstem which innervate both the cervical and lumbar enlargements of the spinal cord by way of axonal collaterals. Such neurons were mixed with those innervating either the cervical or lumbar enlargement alone and were found within the nuclei reticularis medullae oblongatae dorsalis and ventralis, the nucleus reticularis gigantocellularis, the nucleus reticularis gigantocellularis pars ventralis, the nucleus reticularis pontis and the nuclei obscurus and magnus raphae. In some nuclei over 50% of the neurons projecting to the cervical enlargement also innervate lumbar levels.

The orthograde transport of horseradish peroxidase conjugated to wheat germ agglutinin (WGA-HRP) has been used to study the distribution of corticospinal axons in adult and pouch-young opossums. As predicted from the results of degeneration and autoradiographic experiments, injections of WGA-HRP into limb areas of somatic motor-sensory cortex labeled axons in the dorsal and lateral funiculi of the cervical and upper thoracic spinal cord which could be traced to dense terminal zones in laminae III–VI. In addition, we obtained evidence for the presence of a few cortical axons in the ventral white matter and for innervation of the medial part of laminae I and II, laminae VII and VIII and lamina X. A few cortical axons are even present in lamina IX.

We have studied the early development of subcortical projections to presumptive somatic sensory-motor areas of neocortex in the North American opossum Didelphis virginiana. The opossum is born in a very immature state, 12-13 days after conception, and climbs into an external pouch where it is available for experimental manipulation. Using the retrograde transport of wheat germ agglutinin conjugated to horseradish peroxidase, we have obtained evidence that axons from the dorsal raphe and superior central nuclei, the substantia nigra, the locus coeruleus and the parabrachial nuclei reach presumptive somatic sensory-motor areas of neocortex by at least postnatal day (PND) 10. Axons showing serotonin-like immunoreactivity, presumably from the dorsal raphe and/or superior central nuclei, and axons containing tyrosine hydroxylase immunoreactivity, presumably from the substantia nigra and/or locus coeruleus, are present in the same areas at birth or shortly thereafter. Thalamic axons do not grow into comparable areas of neocortex until after PND 10. Such axons reach the subplate region of ventrolateral neocortex first and then proceed dorsomedially; by estimated PD (EPND) 21, they are present in presumptive layers I, V and VI, but they do not innervate an identified layer IV until EPND 48. The developmental sequences suggested by our study are compared with those reported for other species and are discussed in light of their importance in the formation of major sensory and motor circuits.

We have employed axonal transport and degeneration techniques to study the development of major rubral connections in the North American opossum. Opposums were chosen for study because they are born. 12 d after conception and have a protracted postnatal development. Our results suggest that: (1) The red nucleus innervates the spinal cord early in development, well before the somatic motor-sensory cortex (Cabana and Martin, 1984); (2) the red nucleus projects to the spinal cord before it receives substantial projections from the cerebellum or cerebral cortex; and (3) projections from the cerebellum reach the red nucleus significantly earlier than those from the cerebral cortex.

The Falck-Hillarp technique, serotonin (5-HT) immunohistochemistry and the retrograde transport of horseradish peroxidase (HRP) were utilized to investigate the development of raphe-spinal connections in the pouch-young opossum. The brainstem raphe and adjacent reticular formation contain 5-HT immunoreactive neurons in the newborn opossum (12 days after conception) and processes from these cells can be visualized in the marginal zone of the spinal cord. Between eight and 15 days after birth 5-HT immunoreactive varicosities begin to grow into the presumptive deep layers of the dorsal horn, the intermediolateral cell column and the ventral horn. In the latter region some of them approximate presumed motor neurons. Between 40–50 days after birth 5-HT immunoreactive varicosities appear in presumptive laminae I and II of the dorsal horn.

We have employed the orthograde transport of wheat germ agglutinin conjugated to horseradish peroxidase to study the organization of rubrospinal connections in adult and pouch young opossums. Our results suggest that: (1) in the adult opossum rubrospinal axons are distributed more widely than suggested by previous studies; (2) rubrospinal projections are formed postnatally in the opossum, but much earlier than corticospinal connections; (3) rubrospinal axons do not grow synchronously, as a massive bundle following a few leading axons, but by addition of axons over a protracted period of time; and (4) the growth of rubral axons into the spinal gray matter follows a predictable rostral to caudal gradient as well as a proximal to distal one relative to the tract. Rubrospinal development is discussed in light of the growth of cerebellar and cortical axons into the red nucleus and the development of motor function.

Retrograde transport studies revealed that the nuclei pallidus, obscurus, and magnus raphae as well as the adjacent reticular formation innervate the spinal cord in the opossum. HRP-lesion experiments showed that a relatively large number of neurons within the nucleus obscurus raphae and closely adjacent areas of the nucleus reticularis gigantocellularis project through the ventrolateral white matter and that many cells within the nucleus magnus raphae, the nucleus reticularis gigantocellularis pars ventralis, and the nucleus reticularis pontis pars ventralis contribute axons to the dorsal half of the lateral funiculi. Neurons within the rostral pole of the nucleus magnus raphae and the adjacent nucleus reticularis pontis pars ventralis may project exclusively through the latter route. Each of the above-mentioned raphe and reticular nuclei contain nonindolaminergic as well as indolaminergic neurons (Crutcher and Humbertson, 1978). When True-Blue was injected into the spinal cord and the brain processed for monoamine histofluorescence evidence for True-Blue was found in neurons of both types. Injections of 3H-leucine centered within the nuclei pallidus and obscurus raphae and/or the closely adjacent nucleus reticularis gigantocellularis labeled axons within autonomic nuclei and laminae IV-X. Labeled axons were particularly numerous within the intermediolateral cell column and within laminae IX and X. Injections of the caudoventral part of the nucleus magnus raphae or the adjacent nucleus reticualris gigantocellularis pars ventralis labeled axons in the same areas as well as within laminae I-III. When the injection was placed within the rostal part of the nucleus magnus raphae or the adjacent nucleus reticularis pontis pars ventralis axons were labeled within laminae I-III and external zones of laminae IV-VII, but not within lamina IX. The immunohistofluorescence method revealed evidence for indolaminergic axons in each of the spinal areas labeled by injections of 3H-leucine into the raphe and adjacent reticular formation. They were particularly abundant within the intermediolateral cell column and within laminae IX and X. These data indicate that raphe spinal systems are chemically and connectionally heterogeneous.

Rubral neurons innervating the spinal cord, the lateral brainstem and the inferior olivary nucleus have been identified in the North American opossum by the retrograde transport of horseradish peroxidase and fluorescent markers. Neurons which project to the spinal cord are found mainly in caudal and rostroventral parts of the red nucleus. In contrast, rubral neurons which innervate the facial nucleus and lateral areas of the medulla are most numerous rostrodorsally. There is overlap in the location of rubral neurons which project to the spinal cord and lateral brainstem, however and; double-labelling techniques show that some rubral neurons provide collaterals to both areas. Neurons which innervate the inferior olive are located primarily in rostromedial portions of the red nucleus with some overlap with those innervating the spinal cord. Few rubral neurons provide collaterals to both the spinal cord and inferior olive.

We have employed axonal transport and degeneration techniques to study the development of selected rubral connections in the North American opossum. Opossums were chosen for study because they are born in an immature state, 12 days after conception, and have a lengthy postnatal development. The results of our studies suggest that: (1) the red nucleus innervates the spinal cord early in development, but not as early as some areas of the brainstem; (2) rubrospinal development occurs postnatally in the opossum; (3) rubrospinal axons do not grow synchronously into the spinal cord, but are added over time; (4) rubrospinal development follows rough rostral to caudal and lateral to medial gradients; (5) the red nucleus is innervated by the cerebellum well before it receives projections from the cerebral cortex; and (6) cortical axons do not grow into the red nucleus until after rubrospinal axons have reached most of their adult targets.

Publisher Summary The chapter reviews the organization of reticular and raphe projections to the opossum's spinal cord and describes possible neurotransmitters or modulators employed by some of them. The opossum's central nervous system is relatively generalized in comparison with that of most mammals. Corticospinal tracts exist, but they are limited and can only be traced to laminae III–VI of the cervical cord. In contrast, axons from the red nucleus and vestibular complex, as well as from several reticular and raphe nuclei, project the length of the spinal cord. Reticular and raphe neurons projecting to different levels of the spinal cord are identified by the horseradish peroxidase (HRP) technique. In a study described in the chapter, HRP was injected into the desired level of the cord where it was incorporated by axons and transported retrogradely to the parent cell bodies. The presence of HRP in reticular and raphe cells was revealed by several histochemical techniques.

The development of projections from somatic motor-sensory areas of neocortex to the diencephalon and brainstem was studied by using the orthograde transport of wheat germ agglutinin conjugated to horseradish peroxidase (WGA-HRP) in a series of pouch-young opossums. The opossum was chosen for study because it is born in a very immature state, 12 days after conception, and has a protracted postnatal development. Cortical axons form a cerebral peduncle by at least postnatal day (PD) 10, a medullary pyramid by estimated PD (EPD) 17, a pyramidal decussation by EPD 26, and reach the first cervical segment of the spinal cord by EPD 29. Cortical axons innervate diencephalic nuclei and perhaps the substantia nigra by EPD 17, but do not grow into more caudal brainstem nuclei until EPD 26. The first brainstem areas innervated by cortical axons are the mesencephalic and rostral pontine tegmentum and parts of the pontine gray adjacent to the pyramidal tract (EPD 29). By EPD 31, cortical axons project to additional areas of the pontine gray, the gigantocellular reticular formation, the medial accessory olive, and the cuneate nucleus. Cortical innervation of the red nucleus and superior colliculus begins at EPD 31 but is not well developed until EPD 35. Cortical axons do not innervate the parvicellular reticular formation or the sensory trigeminal nuclei until EPD 35. Evidence for transient cerebrocerebellar axons was also found.

The origin of descending pathways to thoracic and cervical levels of the spinal cord has been investigated with retrograde tracing techniques in a series of pouch young and adult opossums. The opossum was chosen because it is born in a very immature state, 12–13 days after conception, and has a protracted development in an external pouch. A few neurons in the pontine reticular formation and nucleus coeruleus were labeled by horseradish peroxidase (HRP) injections of the thoracic cord as early as postnatal day (PND) 3. By PND 5, similar injections labeled neurons in the same areas as well as in the medullary reticular formation, the raphe nuclei of the caudal pons and medulla, the spinal trigeminal nuclei, the vestibular complex, the accessory oculomotor nuclei and the interstitial nucleus of Cajal. When Nuclear Yellow (NY) was employed, neurons were also labeled in the red nucleus, the hypothalamus and possibly in the nucleus of the solitary tract. Regardless of the technique employed, neurons in the dorsal column nuclei were not labeled by thoracic injections until at least PND 14. Axons from the nucleus ambiguus, the fastigial and interposed nuclei of the cerebellum as well as the intermediate and deep layers of the superior colliculus reach cervical levels of the cord, where they are specifically targeted, by at least PND 17. They do not significantly overgrow those levels during development. Corticospinal axons are the last of the major descending pathways to innervate the spinal cord. Cortical neurons cannot be labeled by cervical injections of either HRP or NY until at least PND 30. Evidence for transient brainstem-spinal and corticospinal projections was obtained.

Publisher Summary This chapter discusses the development of descending spinal connections. The opossum is a good model for developmental studies. The development of the opossum's nervous system is quite protracted in comparison to that of placental mammals. The opossum's hindlimbs are little more than buds at birth and do not become motile for 7–10 days postnatally. The development of hindlimb motility and the influence of descending pathways on that motility can be divided into three stages: (1) stage I begins at birth (or conception) and ends when spontaneous hindlimb movements first appear (postnatal day 7–10), (2) stage II begins with the onset of hindlimb motility and ends approximately 20 days later (50–60 mm snout-rump length) when that motility can be altered by transecting the spinal cord at thoracic levels, and (3) stage III begins when hindlimb movements can first be affected by thoracic transection and continues until the animal can support its weight and ambulate (90–120 mm S-R length, estimated to have been 60–75 days in the pouch). Axons from the reticular formation, raphe nuclei, coeruleus complex, vestibular nuclei, and red nucleus invade most of the lumbosacral areas innervated in the adult animal before transecting them produces noticeable change in hindlimb motility.

The course and distribution of corticospinal axons have been traced in a series of pouch young and adult opossums using wheat germ agglutinin conjugated to horseradish peroxidase (WGA-HRP). Cortical axons enter the spinal cord approx. 28 days after birth (40 days after conception) and, at that time, are limited to those portions of the dorsal and lateral funiculi which contain them in the adult animal. Shortly thereafter, cortical axons are also found in regions of presumptive white matter where they are not seen in older pouch young or adult opossums. Those in the dorsal and lateral funiculi reach their caudal extent, the fourth thoracic segment, approx. 38 days after birth and do not significantly overgrow that level during development. Cortical axons grow into the gray matter exclusively from the dorsal and lateral funiculi and first innervate adjacent portions of laminae IV and V. They subsequently extend into laminae III and VI, then VII, VIII and X and finally, at the cervical enlargement, the medial edge of laminae I–II and lamina IX. There is a subsequent period of development during which the density of cortical innervation in all spinal laminae, particularly in the ventral horn, appears to exceed that in the adult opossum.

We have employed the retrograde transport of horseradish peroxidase alone or conjugated to wheat germ agglutinin, to label neurons within the medullary reticular formation which project to the spinal cord, the diencephalon and the cerebellum at different stages of development in the North American opossum. At selected ages, the fluorescent markers Fast Blue and Diamidino Yellow were also used in double-labeling experiments to determine if single neurons innervate both the spinal cord and diencephalon or the spinal cord and cerebellum, presumably via axonal collaterals. The opossum was employed because it is born in a very immature state, 12 days after conception, and is thus available for injections at early stages of development. At all ages studied, the location of retrograde labeling within the medullary reticular formation alter spinal, diencephalic or cerebellar placements of horseradish peroxidase or its conjugate appeared similar to that obtained in the adult animal. Such results suggest that the origin of projections from the medullary reticular formation to the areas injected is specified early in development. At some ages, however, the labeling density appeared greater than in the adult animal. When either Fast Blue or Diamidino Yellow was injected into the spinal cord and the other marker was placed into the diencephalon at such ages, relatively few neurons of the medullary reticular formation were double-labeled. When one marker was injected into the spinal cord and the other was placed within the cerebellum, no double-labeled neurons were found. These results indicate that at the ages studied, relatively few neurons of the medullary reticular formation provide collateral innervation to either the spinal cord and diencephalon or the spinal cord and cerebellum. Similar conclusions have been reached previously for the adult opossum. We have interpreted our results to suggest that the organization of reticular projections, at least to the areas injected, may not be shaped by the selective elimination of axonal collaterals as in certain other areas of the brain.