Transsynaptic degeneration of motoneurones caudal to spinal cord lesions

Brain Research Bulkfin, Vol. 22, pp. 39-45. 0 Pergamon

Press plc, 1989. Printed

0361-9230/89

in the U.S.A.

$3.00 + .OO

Transsynaptic Degeneration of Motoneurones Caudal to Spinal Cord Lesions E. EIDELBERG,’

L. H. NGUYEN,

R. POLICH

AND J. G. WALDEN

Division of Neurosurgery, Audie L. Murphy Veterans Administration Hospital and Division of Neurosurgery, The University of Texas Health Science Center, San Antonio, TX 78284

EIDELBERG, E., L. H. NGUYEN, R. POLICH AND J. G. WALDEN. Transsynaptic degeneration of motoneurones caudal to spinal cord lesions. BRAIN RES BULL 22(l) 39-45, 1989. -We studied the effects of complete transversal section of the spinal cord, at T8-10, in adult rats, upon the number and morphology of identified motoneurones in lumbar segments L4 and L5. In observations by light and electron microscopy many lumbar motoneurones had structural abnormalities when the interval between surgery and perfusion ranged between a few hours and one week. We found also that as many as 25% of the motoneurones distal to a cord transection disappeared as a consequence of the lesions. We did not find comparable changes in the spinal cord at C6 after transection at TS-10. Complete removal of the cerebellum did not reduce the lumbar motoneurone counts. Bilateral ablation of the “motor” cortex did cause a reduction of motoneurone counts at L4-5; these animals showed normal or near normal spontaneous locomotor activity beginning a few days after the lesion was placed. Motoneurone counts were significantly reduced after partial cord lesions that spared the dorsal funiculi (where the corticospinal tract travels in the rat), but in this case the rats were paraplegic as a result of the lesion. Cord transection at 7 days of postnatal age resulted in reduced motoneurone counts when the rats reached adulthood. Intraspinal or subarachnoid administration of colchicine led to reduced motoneurone counts. Prolonged infusion of a GABA agonist, muscimol, into the lumbar CSF did not prevent the loss of did not motoneurones produced by cord transection. Pretreatment of animals with a Ca *+ channel blocker (nimodipine) prevent the effects of cord transection. We conclude that lesions that destroy descending spinal axons may lead to transsynaptic degeneration of motoneurones, perhaps by interruption of axonal transport of putative “neuronotrophic factors,” but other explanations cannot he ruled out at this time. Rats

Spinal

shock

Transsynaptic

degeneration

Motoneurones

SEVERE injury to the spinal cord has certain transient effects upon the function of neurones caudal to the lesion, in addition to paralysis and anesthesia. Spinal shock (6, 17, 22) is prominent among the transient effects. Others, less widely known, are muscular atrophy and supersensitivity to acetylcholine (3,29). The atrophy in spinal rats is found in muscle fibers of all histochemical types. It is first found a few days after the cord injury, peaks at 3 weeks, and reverses gradually in the course of the subsequent 6-8 weeks (21,22). Supersensitivity to intraarterial acetylcholine follows a similar time course (3,29). It has been proposed that muscle atrophy after “upper motoneurone” lesions is due to transsynaptic degeneration of spinal motoneurones, but direct evidence for this proposal is not available in the literature (2, 5, 19). Other investigators favor disuse as explanation for muscle atrophy in spinal animals (22); however, since disuse is a permanent consequence of spinal transection in adult rats, it is not a likely explanation for concurrent “spontaneous” recovery from the atrophy. We report the results of a series of studies, in rats, directed ‘Requests for reprints should be addressed to Eduardo Eidelberg, Science Center, 7703 Floyd Curl Drive, San Antonio, TX 78284.

Neuronotrophic

factors

at testing the transsynaptic degeneration hypothesis. We relied primarily upon counting the motoneurones of the lumbar enlargement, at L4-5, after lesions of the neuraxis (27). Complementary observations were made by light and electron microscopy. METHOD

We did most of the experiments on adult Sprague-Dawley male rats weighing between 200-250 g. One study required that rats be spinalized 7 days after birth and studied histologically at 50-60 days of age. The surgery in adult rats was carried out with aseptic technique, and under general anesthesia by RAK, the initial dose being 1.2 ml/kg, IM. This agent is a mixture of Rompun@, 180 mg, Acepromazine@, 30 mg, and Vetalar@ 900 mg, in 21.0 ml of vehicle. The newborn rats were anesthetized by cooling on ice (30,32). Postoperative measures included penicillin for 5 days, twice-daily bladder expression, hand feeding, and padding of the cage bottom with shredded paper. The

M.D.,

39

Division

of Neurosurgery

(Surgery),

The University

of Texas Health

40

EIDELBERG,

neonatal rats were returned to their mothers after waking from the surgery. The spinal cord was transected via a thoracic laminectomy at T8-10. The exposed spinal cord was gently lifted with a nerve hook and sectioned with a scalpel. The passage of the hook across the cut insured that the lesion was complete. Partial lesions of the spinal cord, at the same level, were produced by bilateral section of the ventral and lateral funiculi, made with a sharp blade introduced obliquely near the dorsal root entry zone. This spared the pyramidal tracts, since they are contained in the dorsal funiculi in the rat (1). The motor cortex (9) was ablated bilaterally by suction through parasaggital craniectomies. The cerebellum was removed by suction through an occipital craniectomy. All the lesions were verified by histological means at the end of the experiments. For this, the animals were deeply anesthetized and then perfused via the left ventricle of the heart with 4% paraformaldehyde in buffered saline. The lesion sites were blocked, postfixed, embedded in paraffin, cut serially, and stained with H&E. The segments C6, L4 and L5 of the spinal cord were removed, postfixed, embedded in paraffin and cut transversally at 10 pm; every 10th section was saved and stained with cresyl violet. We scanned each section at 100X magnification, using an eyepiece micrometer. We recorded the number, in both sides of each section, of large (~20 pm lesser diameter) cells, containing a single nucleolus, and located in the ventral horn. We counted all the identified motoneurones in 48 sections/block; (C6= 1 block; L4 + LS= 1 block). Since we read l/lOth of the sections the counts in the figures represent approximately 10% of the population under study. Double counting was unlikely because of the 90 pm spacing between consecutive sections counted (14). There were at least 4 rats in each group. The mean counts per block were compared across groups by ANOVA followed by paired t-tests. Significance was set at p=O.O5. In one experiment the triceps surae motoneurones were labeled retrogradely, by injecting HRP into the corresponding muscle nerves (0.2 ~1 of 30% solution), and the rats were perfused 2 days later. Blocks from L2-L5 were cut serially and the sections were developed by the TMB method (23). We compared the size spectrum (the crossectional area of the soma), of HRP motoneurones in intact rats versus rats transected at T8-10 six days before the label was introduced. Electron microscopic studies were carried out in rats prepared separately from the groups described above. These animals were either intact or their spinal cords had been transected at T8-10,4 hours, 24 hours, or 4, 8, or 15 days prior to perfusion. After fixation the L5 segment was removed and cut at 50 pm with a vibratome. These sections were postfixed in 2% osmium tetroxide, dehydrated and embedded in plastic for EM. After ultrathin sectioning the specimens were poststained with uranyl acetate and lead citrate. The sections were viewed and photographed on a Philips 301 E.M. RESULTS

Effects of Complete Transection of the Spinal Cord, in Adult Rats, Upon the Number and Morphology of Lumbar Motoneurones Complete transection at T8-10 resulted in a significant decrease in the number of L4-5 motoneurones, as defined by standard criteria (see the Method section). This effect was detectable from the first day after transection. It did not grow with time thereafter, since there were no significant differences among groups of rats sacrificed at different intervals after transection (Fig. 1A). There were no significant differences between the motoneurone counts at C6 from intact controls and from

NGUYEN, POLICH AND WALDEN

adult rats with their spinal cord transected 8 days before at T8TlO (Fig. 1B). To investigate the possibility that the loss of motoneurones might be related to their relative size, we compared the size distribution histograms of triceps surae motoneurones labeled with HRP. Figure 2 shows that there was apparently no selective loss of cells of a particular size class. We studied, by light and electron microscopy, the appearance of motoneurones at L4-5 in intact rats, and in animals whose spinal cord was transected 4 hours, or 1, 4, 8 and 15 days prior to perfusion. At 4 hours the light microscope observations did not show any obvious changes compared to the controls. The EM did show aggregation of mitochondria near the nucleus in many of the motoneurones observed. At 24 hours many of the motoneurone nuclei were eccentric, mitochondria were clumped together, and many cells appeared to have lost the normal canalicular structure of the endoplasmic reticulum, as illustrated in Fig. 3. None of these abnormalities were observed in rats perfused at 8 or 15 days posttransection. Effects of Complete Transection of the Cord, at Midthoracic Levels, Carried out at 7 Days Postnatal Age The rats in this group were perfused at 50-60 days of age. Most of them exhibited the motor activity in the hindlimbs described by Stelzner and his colleagues (30,32): they stepped when suspended in air or on the ground, and were capable of at least some hindquarter support. The cell counts at C6 and at L4-5 were significantly lower than in nonoperated controls (Fig. 1A and B), and lower also than in rats transected when adult. Effects of Lesions Other Than Complete Transection of the Spinal Cord As shown in Fig. lD, total cerebellectomy-which produced profound ataxia and virtual immobility-had no detectable effect upon the cell counts at L4-5 (interval between surgery and perfusion=8 days). By contrast, bilateral ablation of the “motor” cortex, as defined by Hall and Lindholm (9), caused significant loss of motoneurones in the lumbar cord 8 days later. These animals had recovered substantial locomotor activity in all 4 limbs within the first 4 days after surgery. Partial, bilateral, lesions of the thoracic cord, sparing the dorsal funiculi, left the rats as completely paraplegic as those with complete cord transection. The cell counts in this group were significantly lower than those obtained in the control group (Fig. 1D). Effects of Colchicine on Motoneurone

Counts

This experiment was addressed at finding out whether interruption of axonal transport in the spinal white matter would have consequences resembling those of complete transection. One group of rats was prepared by laminectomy at T8-10 followed by injection of a solution of colchicine (5.0 pg/pl) into the exposed spinal cord. These rats were perfused 8 days later. The cell counts at L4-L5 (Fig. 1C) showed a significant drop in motoneurone counts. Another group of rats was prepared by injecting the same amount of colchicine into the lumbar subarachnoid space via a laminectomy, using a soft catheter threaded into a nick in the dura. The motoneurone counts in these animals (perfused 8 days later) were significantly below the control data, both at L4-L5 and C6. Pretreatment

With a GABA Agonist (Muscimol)

This experiment was designed to test the hypothesis that the loss of motoneurones might be due to the interruption of tonic

TRANSSYNAPTIC

DEGENERATION

41

OF MOTONEURONES

D

FIG. 1. Each bar represents the mean (and SD.) number of motoneurones counted (i.e., l/10 of the population). A: Cell counts in controls, and in animals with spinal transection at TS-10, at stated intervals prior to perfusion. Bar on far right: comparable measurements made in rats spinalized at 7 days of postnatal age and perfused at 50-60 days. Asterisk indicates significant differences with controls. B: Similar observations as in A, but counts made in the cervical cord (C, segment). C: Effects of drugs upon cell counts at L4-5 in spinalized animals. D: Effects of cerebellectomy, motor cortex ablation, or partial cord sections on cell counts.

inhibitory influences (presumably mediated by GABA) by the lesions [cf. (28)]. Since GABA itself is not readily taken up into tissue from the blood stream or the CSF, we used its agonist, muscimol, delivered into the lumbar CSF space by a chronicallyimplanted minipump (Alzet 2001) coupled to a thin catheter (28). The combination used released 1.0 PI/hour of the muscimol solution (5.0 pg/pl) in Ringer%. Two days after the implantation of the pumps the animals were reanesthetized and their spinal cord was transected at T8-10; 8 days later they were perfused and motoneurone counts were made at L4-L5. This experiment (Fig. 1C) showed that the counts made in the “treated” (muscimol) group were not significantly different

from those made in control rats implanted containing Ringer’s solution alone.

with minipumps

Effects of Nimodipidine Upon Loss of Motoneurones After Transection It has been proposed that the mechanism of neuronal death following trauma, anoxia, etc. involves irruption of Ca2+ into nerve cells and consequent activation of self-destructive enzymes (24). We wanted to test whether pretreatment with the Ca2+ channel blocker nimodipine would attenuate the consequences of subsequent spinal cord transection. One group of

42

EIDELBERG,

30

CONTROL

!

Ml

30

1

TRANSECT

NGUYEN,

a0

HR9

1s

POLICH

AND WALDEN

TRANSECT HRlO

26

1000 AREA

2000

lyml

30

3000

4000

AREA

lpml

5000

so00

7000

500

1500250035004500 AREA

TRANSECT

JO

HRll

lpml

TRANSECT

HR12

2, I

500

1500

2500 AREA

3500

4500

5500

500

1500

2500 AREA

Ipml

3500

4500

500

5500

1500

I~ml

2500 AREA

3500

4500

5500

6500

HR12

HR13

lpml

30

I

TRANSECT

HR13

2500

5500

25

500

1500

AREA

3500

4500

6500

tiR1

IIRZ

HRlO

HRll

Ipml

FIG. 2. Distribution of triceps surae motoneurones by size. Controls: no preliminary surgery. “Transect” rats spinalized at T8-10 eight days before perfusion. Bottom right histogram: mean size of motoneurones +S.D. in each rat.

rats (N=5) received a dose of 5.0 mg/kg of nimodipine dissolved in polyethylene glycol400. The controls (n=5) received solvent alone; both were injected via the tail vein, taking precautions to avoid photochemical degradation of the drug. The injections preceded spinalization at T8-10 by 40-45’. There were no significant differences in the cell counts in the experimental and control groups following perfusion, 8 days after transection (Fig. 1C). DISCUSSION

The main finding of this study is that transection of the spinal cord, at the thoracic level, led to a significant reduction of the number of lumbar motoneurones. We found no changes in the cervical enlargement. Morphological observations in lumbar motoneurones, carried out at varying intervals of time after spinal cord transection at T8-10, suggest that many lumbar

motoneurones may have been affected reversibly. Sections studied a week or longer after spinalization appeared to be structurally normal by LM and EM criteria. Since we counted only motoneurones for these studies, we do not know if similar changes developed in Renshaw cells, tract cells, interneurones, etc. Besides the ease and reliability of counting large motoneurones, the main reason for concentrating upon this cell type was the observation by Solandt and Magladery (29), that lesions of the “upper motoneurone” produce muscular atrophy and supersensitivity to acetylcholine. Others have shown more recently, muscle fiber type transformations (3, 4, 15, 18, 21, 29). These facts suggested to us that “upper motoneurone” lesions may act on muscles via transsynaptic changes in the spinal motoneurones (19). After finding that the population of lumbar motoneurones is significantly affected by rostra1 spinal cord transection, we endeavored to find out whether there is a preference for larger

TRANSSYNAPTIC

DEGENERATION

24 hr TRANS

CONT.

FIG. 3. Appearance of motoneurones prior to perfusion (C, D).

43

OF MOTONEURONES

in control

group

(A, B), compared

to an L5 motoneurone

from a rat with cord transection

at T,, 24 hours

EIDELBERG,

44

or smaller motoneurones. It is well known that the smaller alpha motoneurones tend to fire tonically, innervate slow-twitch muscle fibers, and are more readily activated by reflex inputs than the large, phasic, motoneurones that innervate fast-twitch muscle [Henneman’s “size principle” (1 i)] . Our observations on the size distribution of retrogradely filled (with HRP) motoneurones in intact and in spinal rats suggest that size alone may not be a major determinant. This is not an entirely reliable test of the question, however, because the measurements were not corrected for artifactual cell shrinkage. It seems unlikely that the low cell counts after transection could be accounted for by cell atrophy: At least in the soleus pool the motoneurones increase in soma volume after thoracic-levef cord tr~section (Eidelberg eb at., Sot. Neurosci. Abstr., 1988). The size distribution histograms of the triceps pool show no consistent shift to the left (Fig. 2). Still, this possibility requires further investigation. It seems worth mentioning that motoneurones belonging to the same pools may differ in their susceptibility to disease or to toxic agents. For example, Swash et al. reported recently that human motoneurones are differenti~ly affected in amyotrophic lateral sclerosis (31), and Koh et al. (15) have proposed that the relative resistance of neurones to the toxic actions of quinolinate is closely linked to their content in NADPH-diaphorase. Stelzner and his collaborators (30,32) have shown that transection of the spinal cord in newborn rats does not prevent the development of some locomotor function in the affected hindlimbs, compared to the minimal or nil recovery observed in rats whose spinal cord was transected after weaning. We did not describe our behavioral observations in these rats, since they replicated the findings of Weber and Stelzner. We were particularly interested in testing the hypothesis that the functional advantage found in rats operated early in life compared to adult operates could be attributed to survival of larger numbers of motoneurones. This was not the case, since the motoneurone counts at C6 and L4-5 were significantly lower in the rats operated early after birth than in those whose spinal cord was sectioned later. The finding that transection of the spinal cord led to reduced lumbar motoneurone counts raised the possibility that

NGUYEN, POLICH AND WALDEN

specific descending pathways might be critical for this effect, since certain cortical lesions in humans produce contralateral muscle atrophy [cf. (6)]. We sought to mimic this situation by ablation of the motor cortex. As noted in Results (Fig. lD), there was a significant drop in motoneurone counts as a result of these lesions. However, sparing the pyramidal tract as the only surviving motor pathway did not prevent similar ceil losses. The loss in number of motoneurones and the seemingly transient morphological changes we described, may be due to transsynaptic degeneration, since none of the lesions of the neuraxis could have caused retrograde degeneration of the motoneurones. It is not known whether this phenomenon can be mediated via several synaptic relays. We tested, in the experiments with colchicine, whether the degeneration of motoneurones could be attributed to interruption of axonal transport (8) or nerve impulse traffic. The results give some support to the axonal transport alternative, although it is not known to what extent this agent interferes with action potential conduction in spinal cord tracts. We turn now to the initial reason for this study, which was to account for the muscle changes observed after spinal cord transection. The nature and timing of the motoneurone changes agree with the hypothesis that the mucular atrophy is the consequence of degenerative changes in the motoneurones. The same applies to recovery from the atrophy, which may be accounted, in part, by functional restitution of surviving spinal cord motoneurones, and possibly by reinnervation by peripheral collaterat sprouting (10). A fascinating phenomenon seen in some muscles, especially soleus, after spinal cord transection, is the almost total replacement of slow-twitch (I) by fast-twitch (HA) fibers (4, 18, 21). This transformation seems to be permanent (15,16); it is not easily explained by collateral reinnervation from fast-twitch axons from adjoining muscle nerves, since in that case the consequent fiber composition of soleus shouid resemble the mixed pattern seen in neighboring muscles, which it does not. Perhaps one consequence of transsynaptic degeneration of the motoneurones is a change in the expression of genes controlling the synthesis of different myosins.

REFERENCES 1. Arias, M. J.; Lopez, J. M.; Cilveti, A. Morphometry of the pyramidal tract in the spinal cord of normal rats. Surg. Neural. 28: 38-40; 1987. 2. Barron, D. H. Structural changes in anterior horn cells following central lesions. Proc. Sot. Exp. Bioi. Med. 30:1327-1329; 1933. 3. Cannon, W. B.; Rosenblueth, A. The supersensitivity of denervated structures. New York: Macmillan Co.; 1949:116. 4. Cope, T. C.; Bodine, S. C.; Fournier, M.; Edgerton, V. R. Soleus motor units in chronic spinal transected cats: physiological and morphological alterations. J. Neurophysiol. 55: 1202-1220; 1986. 5. Cowan, W. M . Anterograde and retrograde transneuronal degeneration in the central and peripheral nervous system. In: Nauta, W. J. H.; Ebbeson, S. 0. E., eds. Contemporary research methods in neuroanatomy. New York: Springer-Verlag; 1970:217-251. 6. Critchley, M. The parietal lobes. Hafner Publ. 1966:163-171. 7. Eidelberg, E. Consequences of spinal cord lesions upon motor function, with special reference to locomotor activity. Prog. Neurobiol. 17:185-202; 1981. 8. Fink, B. R.; Byers, M. T.; Middaugh, M. E. Dynamics of colchitine effects on rapid transport and axonal morphology. Brain Res. 561299-311; 1973. 9. Hall, R. D.; Lindholm, E. P. Organization of motor and sensory neocortex in the albino rat. Brain Res. 66:23-38; 1974.

10. Hatcher, D. D.; Luff, A. R.; Westerman, R. A.; Finkelstein, D. I. Contractile properties of cat motor units enlarged by motoneurone sprouting. Exp. Brain Res. 60:%X%593; 1985. It. Henneman, E.; Somjen, Ci.; Carpenter, D. D. 0. Functional significance of cell size in spinal motoneurons. J. Neurophysiol. 28: 599-620; 1965. 12. Hultborn, H.; Malmsten, J. Changes in segmental reflexes following chronic spinal hemisection in the cat. 1. Increased monosynaptic and polysynaptic ventral root discharges. Acta Physiol. Stand. 119:405-422; 1983. 13. Koh, J. Y.; Peters, S.; Choi, D. W. Neurons containing NADPHdiaphorase are selectively resistant to quinoiinate toxicity. Science 234:73-76; 1986. 14. Konigsmark, B. W. Methods for the counting of neurones. In: Nauta, W. J. H.; Ebbeson, S. 0. E., eds. Contemporary research methods in neuroanatomy. New York: Springer-Verlag; 1970:315340. 15. Lieber, R. L.; Johansson, C. B.; Vahlsing, H. L.; Hargens, A. R.; Feringa, E. R. Long term effects of spinal cord transection on fast and slow rat skeletal muscie. 1. Contractile properties. Exp. Neurol. 91:435-448; 1986. 16. Lieber, R. L.; Friden, J. 0.; Hargens, A. R.; Feringa, E. R. Long term effects of spinal cord transection on fast and slow rat skele-

45

TRANSSYNAPTIC DEGENERATION OF MOTONEURONES

17. 18.

19.

20. 21.

22.

23. 24.

tal muscle. II. Morphometric properties. Exp. Neurol. 91:435-448; 1986. Little, J. W. Serial recording of reflexes after feline spinal cord transection. Exp. Neurol. 93:510-521; 1986. Lofton, S. J.; Munson, J. B.; Sypert, G. W.; Foehring, R. C. Altered distribution of motor unit types following spinal cord transection. Sot. Neurosci. Abstr. 8:966; 1982. McComas, A. F.; Sica, R. E. P.; Upton, A. R. M.; Aguilera, N. Functional changes in motoneurones of hemiparetic patients. J. Neurol. Neurosurg. Psychiatry 36:183-193; 1973. McCouch, F. P.; Liu, C. N.; Chambers, W. W. Descending tracts and spinal shock in the monkey. Brain 89:359-376; 1966. Maxwell, L. C.; Eidelberg, E. Transient atrophy of skeletal muscle fibers after spinal cord injury. Sot. Neurosci. Abstr. 12:1083; 1986. Mayer, R. F.; Burke, R. E.; Toop, J.; Walmsley, B.; Hodgson, J. A. The effect of spinal cord transection on motor units in cat medial gastrocnemius muscle. Muscle Nerve 7:23-31; 1984. Mesulam, M. M. Tracing neural connections with horseradish peroxidase. New York: J. Wiley and Sons; 1-151; 1982. Meyer, F. B.; Anderson, R. E.; Yaksh, T. L.; Sundt, T. M., Jr. Effect of nimodipine on intracellular brain pH, cortical blood flow,

25. 27.

28.

29.

30.

31.

32.

and EEG in experimental focal cerebral ischemia. J. Neurosurg. 64:617-626; 1986. Munson, J. B.; Foehring, R. C.; Lofton, S. A.; Zengel, J. E.; Sypert, G. W. Plasticity of medial gastrocnemius motor units following cordotomy in the cat. J. Neurophysiol. 55:619-634; 1986. Oppenheim, R. W. The absence of significant postnatal motoneuron death in the brachial and lumbar spinal cord of the rat. J. Comp. Neurol. 246:281-286; 1986. Saji, M.; Reis, D. J. Delayed transneuronal death of substantia nigra neurons prevented by gamma-aminobutyric acid agonist. Science 235:66-69; 1987. Solandt, D. Y.; Magladery, J. W. A comparison of the effects of upper and lower motoneurone lesions on skeletal muscle. J. Neurophysiol. 5:373-380; 1942. Stelzner, D. J.; Eichler, W. B.; Weber, E. D. Effects of spinal transection in neonatal and weanling rats: survival of function. Exp. Neurol. 46:156-177; 1975. Swash, M.; Leader, M.; Brown, K.; Swettenham, K. W. Focal loss of anterior horn cells in the cervical cord in motor neuron disease. Brain 109:939-952; 1986. Weber, E. D.; Stelzner, D. J. Behavioral effects of spinal cord transection in the developing rat. Brain Res. 125:241-255; 1977.