A quantitative analysis of frog optic nerve regeneration: is retrograde ganglion cell death or collateral axonal loss related to selective reinnervation?

The present study was designed to assess whether axon collateral formation and loss or retrograde cell death contribute to selective reinnervation during optic nerve regeneration in the frog, Rana pipiens. The right optic nerve was crushed in 18 frogs, and samples were taken near the optic disc (retinal segment) and near the optic chiasm (brain segment). These samples were studied quantitatively with the electron microscope at various postoperative survival times (1, 2, 3, 4, 6, 12 weeks, 6 months, 1 year; N = 2). The number and size of axons in each segment were estimated from a series of electron micrographs taken at intervals across the transverse extent of each nerve and compared with normal nerves (N = 4). Results show that there are 5.3 +/- 1.8 X 10(5) (S.D.) unmyelinated and 2.3 +/- .5 X 10(4) myelinated axons in the normal nerve. One week post-crush (p.c.) there is a 27% decrease in the number of axons in the retinal segment (4.1 +/- 1.4 X 10(5)), indicating early retrograde axonal loss. As expected, there is a greater loss of axons at this time in the brain segment (3.0 +/- 1.3 X 10(5)). Between 2 and 6 weeks p.c. the number of axons increases in the retinal segment to over twice the normal number of axons increases in the retinal segment to over twice the normal number (12.3 +/- 3.8 X 10(5)) and to over four times this number in the brain segment (20.0 +/- 3.0 X 10(5)), showing collateral axon formation results from this injury. A large loss in the number of axons occurs in both nerve segments between 6 and 12 weeks p.c. (4.3 +/- 1.5 X 10(5)) and an additional loss at 20 weeks p.c. (2.2 +/- .98 X 10(5)). Subsequently, the number remains constant, approximately 40% of normal. Visual recovery was seen in the two frogs tested one year after optic nerve crush that were used for optic axon counts. Autoradiography in these same animals showed the optic nerve projections normally seen after regeneration. Besides axonal loss, our results also indicate that the size of both myelinated and unmyelinated axons is significantly above normal at chronic postoperative periods. This increase in axonal size is interpreted to be related to the increased territory each remaining optic axon must fill to restore the optic projections.(ABSTRACT TRUNCATED AT 400 WORDS)

Tests of the regenerative capacity of tectal efferent axons in the frog, Rana pipiens.

Experiments were designed to determine if neurons of the ranid optic tectum, a major target of the optic nerve, possess the same regenerative potential as optic axons. Normal tectal efferent (TE) projections were reexamined by using the anterograde transport of 3H-proline and autoradiography (n = 18), bulk-filling damaged TE axons with horseradish peroxidase (HRP; n = 18) and anterogradely transporting wheat germ agglutinin-HRP (n = 8) to label TE axons. Results were similar to reports that used degeneration methods (Rubinson: Brain Behav. Evol. 1:529-561, '68; Lazar: Acta. Biol. Hung. 20:171-183, '69). Following a brainstem hemisection just caudal to the nucleus isthmi (1-20 weeks), the ipsilateral descending TE pathway was autoradiographically examined (n = 20). While all other TE projections appeared normal, there was no detectable ipsilateral descending projection beyond the lesion site. Ascending TE axons were cut at the anterior tectal border by hemisecting the left diencephalon (LDH)--a lesion that also cuts optic axons projecting to the left tectum. There was no indication of TE axonal regeneration with the aid of autoradiography or HRP histochemistry 1-30 weeks postlesion (n = 48) even when the medial diencephalon was intentionally left intact (n = 4). However, in all four cases examined, optic axons regenerated following the same LDH where TE axonal regeneration failed (also see Stelzner, Lyon, and Strauss: Anat. Rec. 205:191A-192A, '83). Local effects of LDH should be similar for both the cut optic and cut TE axons. Other factors were tested that may contribute to the lack of TE axonal regeneration. Our results indicate that optic regeneration itself (n = 8), postaxotomy retrograde cell death of TE neurons (n = 6), deafferentation of the tectum of optic axons, and potential sprouting within tectal targets by intact contralateral TE axons (n = 10) are not critical factors aborting TE axonal regeneration. TE axons filled with HRP at chronic periods after LDH (n = 4) terminate anomalously near the LDH border. Many of these endings are similar to reactive endings or terminal clubs seen after axonal injury in the mammalian CNS. Our results suggest that this disparity in regenerative ability of optic and TE axons may be related to a difference in the responsive ability of these cell types to initiate or maintain axonal elongation after axotomy within the amphibian CNS environment.

Plasticity of the corticospinal tract following midthoracic spinal injury in the postnatal rat.

Rats received a midthoracic spinal cord "overhemisection" including right hemicord and left dorsal funiculus at birth (neonatal operates, N = 15) or 21 days of age (weanling operates, N = 14). In a second experiment neonatal (N = 6), 6-day (N = 3), and 12-day (N = 7) rats sustained a right sensorimotor cortex (SmI) ablation to destroy the left corticospinal tract (CST) at the same time as the spinal injury (double lesion operates). Later (3-12 months) injections of 3H-proline and autoradiography were used to label the left or right CST. The results of the first experiment showed that most right CST axons failed to grow around the spinal lesion in neonatal operates (N = 9). There was an increase in the density of label, mainly to CST projection areas, in a 1-mm zone rostral to the lesion. However, left CST axons bypassed the lesion by growing through the intact tissue in neonatal operates (N = 6). These displaced axons were consistently located within the dorsal portion of the lateral funiculus (dLF) and remained within that location caudal to the lesion, an area normally containing only a few CST axons. In spite of this abnormal position, these axons terminated bilaterally throughout the remainder of the cord in normal CST sites. In weanling operates, CST axons severed by the lesion did not regenerate around the lesion site. An increased density of label over the few spared axons within the left dLF and in CST projection zones immediately caudal to the lesion site suggested axonal sprouting by these axons. The results of the second experiment showed that the lack of growth of right CST axons around this injury in neonatal operates was, at least partially, due to an interaction with left CST axons. In neonatal double lesion operates, right CST axons grew around the spinal injury for a varying distance within the left dLF and distributed bilaterally to normal CST sites. The number of right CST axons bypassing the lesion was related to the configuration of the lesion site. A smaller number of right CST axons bypassed the lesion in 6-day double lesion operates and most terminated within 2-3 mm of the lesion site. Right CST axons failed to grow around this injury in 12-day double lesion operates.(ABSTRACT TRUNCATED AT 400 WORDS)

The aberrant retino-retinal projection during optic nerve regeneration in the frog. I. Time course of formation and cells of origin.

We have reported previously that during optic nerve regeneration in Rana pipiens, axons are misrouted into the opposite nerve and retina. In the present investigation we have examined the time course of formation of these "misrouted" axons and their cells of origin. The right eye of 31 frogs was injected with 3H-proline at various times after right optic nerve crush. In every frog examined 2 weeks and later after nerve crush, the distribution of autoradiographic label indicated that axons from the right eye had grown into the left optic nerve at the chiasm. The amount of label increased from 2 weeks to reach a maximum at 6 weeks where the entire left nerve was filled with silver grains. At 5 to 6 weeks after crush, labeled axons were found within the ganglion cell fiber layer (GCFL) of the retina of the opposite eye for a maximum distance of 2.3 mm from the optic disc. In frogs examined at intervals later than 6 weeks after crush, the amount of label within the left eye and nerve progressively decreased, indicating a gradual disappearance of the misrouted axons. Studies using anterograde transport of horseradish peroxidase (HRP) after nerve injection confirmed these autoradiographic findings. The position of ganglion cells in the right eye whose axons were misrouted to the left eye was determined by retrograde transport of HRP. Five or 6 weeks after crushing the right optic nerve, the left eye was injected with HRP and labeled ganglion cells were found throughout the right eye retina. The largest percentage of labeled cells was found within the ventral half of the retina, particularly within the temporal quadrant, and nearly all of the labeled cells were found in more peripheral portions of the retina. Since few retino-retinal axons are found during normal development, the present results show that the factors guiding regenerating axons in the adult frog differ substantially from those present during development.

The aberrant retino-retinal projection during optic nerve regeneration in the frog. II. Anterograde labeling with horseradish peroxidase.

Previous experiments have shown that a substantial number of regenerating optic axons in adult frogs (Rana pipiens) are misrouted into the opposite optic nerve and retina during early stages of regeneration. This projection is maximal at 5 and 6 weeks after optic nerve crush. To further characterize this anomalous projection, small quantities of horseradish peroxidase (HRP) were injected into the right eye or right optic nerve 5 or 6 weeks after right optic nerve crush. Twenty-four hours later the animals were killed and regenerating axons anterogradely filled with HRP were reacted with the tetramethyl-benzidine method or a diaminobenzidine-CoCl2 method. Serial reconstruction tracing the course of individual axons through the optic chiasm showed that few of the axons projecting into the opposite optic nerve were collaterals of axons projecting centrally. Instead, the majority of labeled axons misdirected into the opposite nerve or contributing to an expanded projection into the ipsilateral optic tract turned out of the chiasm without branching. Many of the labeled regenerating axons had unusual trajectories within the chiasm, making abrupt turns or changing their direction of growth. Most of the axons misrouted into the opposite nerve came from portions of the chiasm nearest to the nerve of other eye. In three of eight frogs with an intact optic nerve, a small number of HRP-labeled axons were found in the left nerve after right nerve injection, but there was no indication that these axons reached the left eye. The results from this investigation suggest that the most parsimonious explanation for the chiasmal misrouting of regenerating frog optic axons is that axons are mechanically deflected into inappropriate pathways.

The aberrant retino-retinal projection during optic nerve regeneration in the frog. III. Effects of crushing both nerves.

Previous reports from this laboratory have shown that a substantial number of optic axons are misrouted after optic nerve regeneration in the adult frog, Rana pipiens. Regenerating axons from a crushed optic nerve are distributed throughout the opposite nerve. In this study, we report the effect of crushing both optic nerves (double crush) on the pattern and degree of axonal misrouting. In 28 frogs both optic nerves were crushed at the same time (simultaneous double crush) and animals survived for varying periods before the right eye was injected with 3H-proline and the brain processed for autoradiography 24 hours later. In every frog with postoperative survivals longer than 2 weeks, labeled axons from the right eye were found in the left optic nerve. However, when compared to the amount of label seen in frogs in which only the right optic nerve was crushed (single crush) there was substantially less label within the left nerve of frogs after crushing both nerves. Label was also found only at the edge of the left nerve in material from double crush frogs, unlike that found after single crush. In four of six frogs where the left nerve was crushed 1 week after the right nerve (delayed double crush), the proximal end of the left nerve was completely filled with label, but more distally, label was found only along the edge of this nerve. Although fewer optic axons were labeled in the opposite optic nerve of double crush frogs, label did extend to the optic disc of that eye. However, label was not apparent in the ganglion cell fiber layer of the opposite eye. Instead, it was confined to the edge of the optic disc in a space apparently associated with papilledema resulting from crushing the optic nerve of that eye. In six frogs the retina of the left eye was removed at the same time the right optic nerve was crushed. Labeled axons of the right eye filled the left optic nerve to the retina-less shell of the left eye. Thus, these data show that the amount and distribution of axonal misrouting into the opposite optic nerve during optic nerve regeneration is affected by intact or regenerating optic axons from the other eye.

Anatomical studies of dorsal column axons and dorsal root ganglion cells after spinal cord injury in the newborn rat.

The response of dorsal column axons was studied after neonatal spinal overhemisection injury (right hemicord and left doral funiculus). Rat pups (N = 11) received this spinal lesion at the C2 level within 30 hours after birth. The cauda equina was exposed 3 months later in one group of chronic operates (N = 5) and in a group of normal adults (N = 2), and all spinal roots from L5 caudally were cut bilaterally; 4 days later the spinal cord and medulla were processed for Fink-Heimer impregnation of degenerating axons and terminals. In a second group of chronic operates (N = 6) and normal adult controls (N = 4) the left sciatic nerve was injected with a cholera toxin-HRP conjugate (C-HRP), followed by a 2-3 day transganglionic transport period, and then the spinal cord and medulla were processed with tetramethylbenzidine histochemistry. Both control groups have a consistent dense projection in topographically adjacent regions of the dorsal funiculus and gracile nucleus. However, there is no sign of axonal growth around the lesion in either group of chronic experimental operates. Instead, there is a decreased density of projection within the dorsal funiculus near the lesion site. Many remaining C-HRP labeled axons in the experimental operates have abnormal, thick varicosities and swollen axonal endings (5-10 microns x 10-30 microns) within the dorsal funiculus through several spinal segments caudal to the lesion. Ultrastructural analysis of the dorsal funiculus in three other chronic experimental operates reveals the presence of numerous vesicle filled axonal profiles and reactive endings which appear similar to the C-HRP labeled structures. Transganglionic labeling after C-HRP sciatic nerve injections (N = 4) and retrograde labeling of L4, L5 dorsal root ganglion neurons after fast blue injections of the gracile nucleus (N = 6) both suggest that all dorsal column axons project to the gracile nucleus in the newborn rat. Dorsal root ganglion (DRG) cell survival following the neonatal overhemisection injury was also examined in the L4 and L5 DRG. DRG neurons that project to the gracile nucleus were prelabeled by injecting fast blue into this nucleus at birth two days prior to the cervical overhemisection spinal injury. Both normal littermates (N = 9) and spinally injured animals (N = 12) were examined after postinjection survival periods of 10 or 22 days.(ABSTRACT TRUNCATED AT 400 WORDS)

Increases in collateral axonal growth rostral to a thoracic hemisection in neonatal and weanling rat.

The spinal cords of newborn and weanling rats were hemisected at the mid-thoracic level. Control studies revealed that Fink-Heimer positive debris was absent in the gray matter at three months postoperative. The remaining animals were given a second lesion, a high cervical spinal hemisection, at five to seven months after the original thoracic hemisection. The pattern of degeneration rostal to the thoracic lesion was compared with similar regions of the spinal cord from animals receiving only a cervical hemisection at the adult stage. In neither experimental group of doubly hemisected rats was there any degeneration observed below the thoracic lesion site, even though no glial or connective tissue scar had formed in animals originally operated at birth. Thus no regeneration had occurred. At least one segment above the initial hemisection: 1. the majority of degenerating axons were localized toward the lateral edge of the spinal cord, especially in the doubly lesioned neonatal group; 2. the erae of ipsilateral white matter was reduced more in the neonatal than the weanling operates; 3. there was an upward shift in axonal diameter of ipsilateral fibers in both the region of the rubrospinal tract and the ventrolateral portion of the lateral funiculus of the doubly hemisected rats when compared with the cervically lesioned controls; 4. a significantly greater amount of degeneration was present in lamina VII of Rexed in both the neonatal and weanling experimental operates (p less than 0.05 weanling; p less than 0.001 neonate); 5. no mean difference in area was seen between the ipsilateral and contralateral gray matter in any group for the segments of the spinal cord in which the judgements and measurements were taken. These data suggest that there has been sprouting of axons from descending nerve tracts rostral to the thoracic lesion in both the neonatal and weanling experimental groups. The question remains whether the sprouting of descending nerve tracts is from collateral of axons which normally project rostral to the thoracic hemisection and are not cut by the thoracic lesion (collateral sprouting) or from collaterals of lesioned axons (regenerative sprouting). Present evidence favors collateral sprouting, expecially in the neonatal operate where much retrograde cell death appears to have taken place.

Changes in the magnocellular portion of the red nucleus following thoracic hemisection in the neonatal and adult rat.

The spinal cords of newborn (0-3 day old) and adult rats were mid-thoracically hemisected. Ninety days later a glial and connective tissue scar had formed at the lesion site in the adult hemisected rats while in neonatally lesioned animals only normal appearing regions of the contralateral spinal cord were found in the area of hemisection. Comparisons of the magnocellular portions of the red nucleus (MPRN) revealed a decrease in cell number in the MPRN contralateral (C-MPRN) to the spinal lesion. However, only in the newborn operates was there massive cell loss accompanied by reduction in area and change in shape of the nucleus. These changes were most obvious in the caudal and ventrolateral portions of the C-MPRN. Pooled data from each group of operates indicated that significantly more cells were lost in the C-MPRN in the newborn than in the adult operates (p less than 0.01). Neurons of the C-MPRN which are known to project to the lower cervical and upper thoracic segments of the spinal cord (Brown, '74; Gwyn, '71) remained undamaged after the mid-thoracic hemisection in both groups. However, neurons of this region were enlarged in both groups when compared to a similar region of the ipsilateral MPRN. These neurons were found to be more enlarged in the newborn than in the adult operates (p less than 0.01). This result indicates that massive retrograde cell death takes place after a mid-thoracic hemisection in the neonatal rat. The retrograde degeneration of axotomized neurons partially may explain why CNS regeneration is not found in the immature mammal even though many of the factors thought to limit regeneration in the adult mammal may not be apparent. The increase in cell size of C-MPRN neurons which remain in the neonatal animals after mid-thoracic hemisection may be related to the increase in axonal size found in the region of the rubrospinal tract rostral to the thoracic lesion reported earlier (Prendergast and Stelzner, '76a). Both the increase in axonal and perikaryal size are hypothesized to be related to the increased distribution of supraspinal axons found in the gray matter rostral to a hemisection of the neonatal rat spinal cord.

Attempts to facilitate dorsal column axonal regeneration in a neonatal spinal environment.

The response to injury of ascending collaterals of dorsal root axons within the dorsal column (DC) was studied after neonatal spinal overhemisection (OH) made at different levels of the spinal cord. The transganglionic tracer, cholera toxin conjugated to horseradish peroxidase, and the anterograde tracer, biotinylated dextran amine, were used to label dorsal root ganglion cells with peripheral axons contributing to the sciatic nerve. There was no indication of a regenerative attempt by DC axons at acute survival times (3 days and later) after cervical injury, replicating previous work done at chronic survival periods (Lahr and Stelzner [1990] J. Comp. Neurol. 293:377-398). There was also no evidence of DC regeneration after lumbar OH injury even though immunohistochemical studies using the oligodendrocyte markers Rip and myelin basic protein showed few oligodendrocytes in the gracile fasciculus at lumbar levels at birth. Therefore, the lack of myelin in the dorsal funiculus at lumbar levels does not enhance the growth of neonatally axotomized DC axons. In addition, DC axons did not regenerate when presented with fetal spinal tissue implanted into thoracic OH lesions, even though positive control experiments showed that segmental dorsal root axons containing calcition gene-related peptide and corticospinal axons grew into these implants, replicating previous work of others. When a thoracic OH lesion, with or without a fetal spinal implant, was combined with sciatic nerve injury to attempt to stimulate an intracellular regenerative response of DRG neurons, again, no evidence of DC axonal regeneration was detected. Quantitative studies of the L4 and L5 dorsal root ganglia (DRG) showed that OH injury did not result in DRG neuronal loss. However, sciatic nerve injury did result in significant post-axotomy retrograde cell loss of DRG neurons, even in groups receiving thoracic embryonic spinal implants, and is one explanation for the minimal effect of sciatic nerve injury on DC regeneration. Although fetal tissue did not appear to rescue a significant number of DRG neurons, the quantitative analysis showed an enlargement of the largest class of DRG neuron, the class that contributes to the DC projection, in all groups receiving fetal tissue implants. This apparent trophic effect did not affect DC regeneration or neuronal survival after peripheral axotomy. Further studies are needed to determine why DC axons do not regenerate in a neonatal spinal environment or within fetal tissue implants, especially because previous work by others in both the developing and adult spinal cord shows that dorsal root axons will grow within the same type of fetal spinal implant.

Synaptogenesis in the intermediate gray region of the lumbar spinal cord in the postnatal rat.

Mid-thoracic spinal cord transection produces dramatically different behavioral results depending upon a rat's age at the time of surgery. The present study was initiated to determine whether the synaptic development in the gray matter of the normal, developing spinal cord differs before and after the period when maximal behavioral recovery occurs. The L6 segments from 10 groups of animals, 0--30 days of age, taken at 3 day intervals (4 animals/group) were studied by light microscopy. Areal measurements of the gray matter were made using an integrating x-y tablet interfaced to a computer. Cell size, cell density and area of neuropil were evaluated in the lateral portions of the intermediate gray matter, laminae VI and VII. Electron microscopic analyses of synaptogenesis were performed on material from the same region in animals 3, 12, 15, 21 and 30 days old using similar morphometric methods while taking note of vesicle, junctional, and mitochondrial morphology. A 60% increase in area of neuropil paralleled a linear increase, of comparable magnitude, in area of the gray matter until 15 days of age when both curves reached plateau. Neuronal perikaryal size remained constant (congruent to 200 sq. microns in plane of nucleolus) throughout development and so could not have contributed to the increase in area of gray matter. Areal measurements of the size and counts of the number of vesicle containing profiles demonstrated a 50% increase in density of axon terminals between 3 and 12 days of age and a steady decline thereafter. The size of vesicle-containing profiles in laminae VI and VII remained constant at a small value (congruent to 0.35 sq microns) until 12 days of age, showed rapid growth to 0.54 sq. microns between 12 and 15 days of age, followed by a more moderate increase in sectional area after 15 days. These results suggest that during the period when recovery of function follows spinal injury, synaptogenesis in the intermediate gray region of the lumbar spinal cord proceeds rapidly, while at stages when little recovery of function follows spinal transection, synaptogenesis is essentially complete.

Neocortical projections of the rat anterior commissure.

The posterior limb of the rat anterior commissure (ACp) was studied in order to better define a temporal cortex in this species. Two methods were used: (1) the anterior commissure (AC) was destroyed on one side of the midline, and the resulting anterograde degeneration was traced with the Fink-Heimer stain; and (2) the corpus callosum and hippocampal commissure were severed, but the AC was left intact. [3H]-Leucine and proline were injected into one hemisphere, and the transported label was traced into the contralateral hemisphere with autoradiography. ACp was found to project to the contralateral cortex along the rhinal sulcus. The pyriform cortex received a projection along the entire length of the sulcus. There was also a distinct projection to neocortex on the lateral surface above the rhinal sulcus which appears to be analogous to the temporal cortex projection of ACp in other species. This finding of a neocortical projection of the ACp in the rat is consistent with observations that have been made on other mammals.

Lack of intralaminar sprouting of retinal axons in monkey LGN.

In the dorsal lateral geniculate nucleus (LGN) of the adult cat there is no evidence for translaminar sprouting of retinal axons to fill sites freed of retinal endings from the other eye. We tested the possibility that retinal axons will sprout to fill denervated retinal sites within laminae of the monkey LGN. In 4 monkeys, retinal ganglion cell axons from either the upper or lower half of the retina were destroyed. To maximize the potential for sprouting in the LGN, on one side of the brain the LGN cells to which the remaining retinal axons normally project were removed by ablation of the appropriate portion of the striate cortex. Three months later the eye receiving the retinal lesion was injected with [3H]proline and the retinal projection to the LGN on both sides of the brain was studied using autoradiography. We found no evidence of intralaminar sprouting of retinal axons either in the normal LGN or in the LGN in which the usual targets of retinal axons had been removed.

Behavioral effects of spinal cord transection in the developing rat.

Albino rats, 0, 9, 12, 15, 18, 21 or greater than 90 days of age, were given a mid-thoracic spinal cord transection. Evaluation of responses of the hindlimbs to a variety of behavioral tasks was begun on the day of surgery and at intervals throughout the postoperative survival period (up to 300 days). Two investigators, independently and without knowledge of the animals' ages or survival times, rated the response data. Histological study showed all transections to be complete. Large differences in behavior are observed when animals trasected at the neonatal stage (0-4 days of age) are compared with animals transected at the weanling stage (21-26 days of age)37. Results of the present investigation indicate a critical period near 15 days of age; animals lesioned prior to this age (0, 9, 12 days of age) show response development and recovery similar to the neonatally lesioned animal, whereas those animals lesioned at a later age (18, 21, greater than 90 days of age) show little recovery and are behaviorally similar to the weanling transected animal. In animals lesioned prior to the fifteenth postnatal day, postural responses appear depressed for a brief period but recover rapidly while most responses of animals in the older groups are depressed for longer periods and never attain the degree of recovery characteristic of the neonatally transected animal. Finally, like the neonatally transected animal, rats lesioned on the ninth and twelfth postnatal day develop certain responses at appropriate times relative to normal response development. If, however, these responses are mature and supraspinal control is present at the time of lesioning, they appear to be permanently depressed and fail to recover.

A comparison of the effect of mid-thoracic spinal hemisection in the neonatal or weanling rat on the distribution and density of dorsal root axons in the lumbosacral spinal cord of the adult.

Transecting the thoracic spinal cord of the rat has markedly different effects on behavioral responses of the hindlimbs if the lesion is made at the neonatal or weanling stage of development. The present investigation tested the possibility that the behavioral differences were related to a difference in the distribution or density of dorsal root connections in the lumbosacral spinal cord. In order to use each animal as its own control the distribution and density of dorsal root axons was compared on the two sides of the L5-S1 segments of the lumbosacral spinal cord in adult rats given a mid-thoracic spinal hemisection at the neonatal or weanling stage of development. Comparing the experimental (initially hemisected side) and control sides of the cord, we found no evidence for a change in the distribution of dorsal root axons. The distribution of Fink-Heimer stained degeneration 4--6 days after bilateral spinal root section was virtually identical on the two sides of the cord from animals hemisected at either stage. However, in rats spinally hemisected at the neonatal stage (n = 8), a significantly greater density of dorsal root degeneration was found within the intermediate nucleus of Cajal (INC) on the experimental side using coded material and a blind analysis. No difference in the density of dorsal root degeneration was detected in the group of rats spinally hemisected at the weanling stage (n = 6). Controls indicated that the increased density of degeneration was not due to compression resulting from shrinkage of the INC or to degeneration remaining from the initial hemisection. We conclude that the increased amount of argyrophilia within the INC of neonatally hemisected rats is due to an increased density of dorsal root axons in this zone. This result supports the hypothesis that the behavioral differences found when comparing animals transected at the neonatal or weanling stages of development are related to an increased number of dorsal root connections within the lumbosacral spinal cord.

NMDA antagonism during development extends sparing of hindlimb function to older spinally transected rats.

Hindlimb weight support and bipedal stepping occur after spinal cord transection in neonatal rats (birth to 12 days of age) while the same lesion in 15-day and older animals results in permanent loss of these responses. Some compensatory change in lumbar spinal circuitry must occur after spinal transection in young animals subserving these hindlimb behaviors. In contrast, animals just a few days older are incapable of such compensatory responses. We have examined the hypothesis that neural activity leads to the postnatal loss of plasticity in spinal circuitry. We find that antagonism of the N-methyl-D-aspartate (NMDA) subtype of glutamate receptor with MK-801 in young animals extends the sparing of hindlimb function after spinal transection to older animals. This effect is not due to a non-specific depression of all exciatory drive to motor neurons since Ia to motor neurons synaptic transmission through non-NMDA receptors is preserved during MK-801 treatment. Acute administration of MK-801 at the time of spinal transection or chronic administration of MK-801 after postnatal day 17 has no effect on recovery of hindlimb function after spinal transection. These results highlight the importance of NMDA receptor activation in spinal circuit maturation.

Neuritic growth maintained near the lesion site long after spinal cord transection in the newborn rat.

The spinal cord of neonatal and weanling rats was mid-thoracically transected. Either 3 or 6 months later the borders of the lesion site were studied using electron microscopy. No sign of axonal regeneration through the lesion site was found in either group, even though the glial reaction was minimal in neonatal operates. In both groups of operates, reactive axonal endings, presumed to result from the original surgery, and neuritic growth were found in a reactive zone on both sides of the lesion site. We conclude that the potential for axonal growth (regeneration or generation) is maintained at the borders of the lesion in both groups of operates.

Cell survival or cell death: differential vulnerability of long descending and thoracic propriospinal neurons to low thoracic axotomy in the adult rat.

Previous studies show that most short thoracic propriospinal (TPS; T5-T7) and long descending propriospinal tract (LDPT; C4-C6) neurons are lost following low-thoracic spinal cord contusion injury (cSCI), as assessed by retrograde labeling with fluorogold (FG). Gene microarray and terminal deoxynucleotidyl transferase dUTP nick end (TUNEL)/caspase-3 immunolabeling indicate that post-axotomy cell death may be responsible for the observed decrease in number of labeled TPS neurons post cSCI. Yet, no indications of post-axotomy cell death are evident within LDPT neurons following the same injury. The present experiments were devised to understand this difference. We assessed the number and size of LDPT and TPS neurons at different time points, retrogradely labeling these neurons with FG prior to delivering a moderate low-thoracic cSCI or after they were axotomized by a complete low-thoracic spinal transection. Counts of FG-filled TPS and LDPT cells indicate a large loss of both neuronal populations 2 weeks post cSCI. Propriospinal neurons in other animals were retrogradely labeled with dextran tetramethyl rhodamine prior to cSCI and tissue was processed for detection of TUNEL- or caspase-3-positive profiles at chronic times post injury. Our overall findings confirm that cell death post injury is the major factor responsible for the loss of TPS neurons during the acute period post cSCI, and that little post-axotomy cell death occurs in LDPT neurons during the first 2 months after the same injury. After chronic axotomy retrograde transport is impaired in LDPT neurons, but can be reinitiated by re-axotomy. Our results also indicate that FG is metabolized/lost from retrogradely labeled neurons at increasing survival times, and that this process appears to be accelerated by injury.

Loss of propriospinal neurons after spinal contusion injury as assessed by retrograde labeling.

We studied the number, location and size of long descending propriospinal tract neurons (LDPT), located in the cervical enlargement (C3-C6 spinal levels), and short thoracic propriospinal neurons (TPS), located in mid-thoracic spinal cord (T5-T7 spinal levels), 2, 6 and 16 weeks following a moderate low thoracic (T9) spinal cord contusion injury (SCI; 25 mm weight drop) and subsequent injections of fluorogold into the upper lumbosacral enlargement (L2-L4 spinal levels). Retrograde labeling showed that approximately 23% of LDPT and 10% of TPS neurons were labeled 2 weeks after SCI, relative to uninjured animals. No additional significant decrease in number of labeled LDPT and TPS cells was found at the later time points examined, indicating that the maximal loss of propriospinal neurons in these two subpopulations occurs within the first 2 weeks post-SCI. The distribution of labeled cells post-moderate SCI was similar to normal in terms of their location within the gray matter. However, there was a significant change in the size (cross sectional area) of labeled neurons following injury, relative to uninjured controls, indicating a loss in the number of the largest class of propriospinal neurons. Interestingly, the number of labeled LDPT and TPS neurons was not significantly different following different injury severities. Although the rostro-caudal extent of the lesion site expanded between 2 and 16 weeks following injury, there was no significant difference in the number of propriospinal neurons that could be retrogradely labeled at these time points. Possible reasons for these findings are discussed.