Brain Research, 589 (1992) 217-224
217
© 1992 Elsevier Science Publishers B.V. All rights reserved 0006-8993/92/$05.00
BRES 18001
Axonal regeneration after spinal cord transection and reconstruction H a r r y S. G o l d s m i t h
a
a n d J.C. de la T o r r e
b
a Boston UniversitySchool of Medicine, Department of Surgery, Boston, MA 02118 (USA) and b Division of Neurosurgery, Universityof Ottawa, Faculty of Medicine, Ottawa, Ont. (Canada) (Accepted 24 March 1992)
Key. words: Axonal regeneration; Surgical reconstruction; Cord transection; Matrix bridge
Following complete transection of the spinal cord, cats were separated into 2 groups to undergo: (i) surgical reconstruction of the disconnected cord using a neuroactive agent mixed into a collagen matrix bridge and omental transposition and (ii) cord transection-only. After 90 days, animals were killed and the brain and spinal cord were removed for immunohistochemistry. Two weeks prior to sacrifice, spinal cord blood flows were measured and the retrograde axonal tracer Fluoro-Gold was injected below the transection site. Gross inspection of the spinal cords at autopsy showed excellent integration and continuity of the collagen matrix bridge with the proximal-distal stumps in the surgical reconstruction group. In the transection-only group, the proximal-distal stumps were connected by a fibrotic, often tapered in the middle, tissue bridge. Results show that omental transposition in the surgical reconstruction group increased spinal cord blood flow by 58% when compared to transection-only animals. Fiuoro-Gold was found in mesencephalic and brainstem catecholaminergic and cholinergic neurons known to send axons to the spinal cord. Immunohistochemical staining with antibodies against catecholamine synthesizing enzymes tyrosine hydroxylase (TH) and dopamine-/3-hydro~lase (DBH) showed that surgical reconstruction treated cat cords but not transection-only, developed dense bundles of dopaminergic and noradrenergic fibers which were present in the collagen matrix bridge and in the distal spinal cord. Extension of these catecholaminergic fibers in surgical reconstruction treated cats showed maximal outgrowth of 90 mm below the transection site when the neuroactive agent 4-aminopyridine was mixed into the collagen matrix, in addition, the synaptogenic marker synaptophysin (SYN) was observed on preganglionic sympathetic neurons in association with dopaminergic- and noradrenergic-containing varicosities distal to the collagen matrix bridge, an indication that neo-synaptic contacts may have been made on these previously denervated neurons. No TH, DBH or SYN was observed below the transection site in transection-only cats. These findings indicate that surgical reconstruction treated cords can develop dense supraspinal fiber outgrowth across a treated collagen matrix bridge fed by an omental blood supply and that these fibers may have made neo-synaptic contacts with n.npmpriate distal spinal cord target tissue,
INTRODUCTION
Functional regeneration of mammalian spinal cord fibers after total cord transection remains as elusive today as when it was described by the Egyptians 5,0~ years ago 2. A number of crucial physic.al and chemical obstacles have been reported to impede successful regeneration of axotomized central spinal fibers t'4,s'7,9,2~. Although successful reconnection of lesioned peripheral nerve fibers is possible in most species, including mammals, regeneration of descending spinal pathways originating in the brain (supraspinal fibers) is limited to lower non-mammalian animals such as fish and frogs 2~'22. We first reported that limited outgrowth of supraspinal catecholaminergic fibers could be induced
using a collagen matrix to bridge the transected spinal cord stumps in rats ~'s. Subsequent studies in cats showed that placing a lengthened omental pedicle ~4't5 on the surface of the collagen matrix provided increased blood flow to the transection site and surrounding host tissue and promoted robust supraspinal axon regeneration into the collagen bridge and nearby the distal cord stump 9'rap. More recent studies by other investigators have confirmed the usefulness of collagen matrix when used as a bridge or support frame for cellular, neuritic and vascular outgrowth in CNS tissue n.m3,2°'27. In this study, we attempted to create an even richer physiological microenvironment for the transected descending spinal fibers most of which originate in the brainstem (supraspinal), by adding a neuroactive agent to the connecting collagen bridge and by
Correspondence: J.C. de la Torre, Division of Neurosurgew, !,rniversity of Ottawa, 451 Smyth Road, Ottawa, Ont. KIH 8M5, Canada. Fax: (1) (613) 738-3191.
218 shortening the
interstump gap
distance from 6 mm to 3
mm.
MATERIALS AND METHODS Twelve anesthetized female adult cats were subjected to a sterile laminectomy between "Is and T~o. After removing the dura at the site, the exposed cord was completely transected at T,~ and verified by microscopy. Before and after transection, the cord was bathed in ice-cold physiological saline for 10 min to reduce tissue metabolic activity, facilitate cord cutting and promote hemostasis. Cats undergoing surgical reconstruction had sterile collagen matrix (Collagen Corp., Paid Alto, CA) dispensed into the 3 mm gap created between the transected cord stumps. This collagen matrix is fluid at 4°C and hardens into a gel by heat polymerization at body temperature within 40 mint(). The collagen matrix is derived from purified pepsin, solubilized bovine collagen and contains 35 mg active collagen/ml in phosphate buffer at pH 7.6. it is obtained and dispensed in sterile condition. One of the following neurotrophic-like substances was mixed into the collagen matrix gel before it hardened: 4-aminopyridine (5/zg), laminin (5 p.g), O-gila maturation factor (10 ~.1), omental lipid angiogenic factor (5 ttg) and physiologic saline (5 p.I). Two cats underwent spinal cord transection only and received no collagen matrix, omentum or neurotrophic-like treatment. Two cats did not undergo any surgery and were kept as intact controls. In group surgical reconstruction cats, a 4 cm left subcostal laparotomy incision was made. The omentum was then gently pulled out of the abdominal cavity, lengthened into a pedicle while maintaining its blood supply intact, tunnelled subcutaneously to the transected cord and placed on the dorsal surface of the hardened collagen matrix where it was anchored with several sutures to the edges of the cut dura I°'ts. All incisions were closed in layers and animals were allowed to recover for 90 days. On day 76 after cord transection, all cats were anesthetized and spinal cord blood flows were recorded using the hydrogen washout method I°, Insulated platinum-iridium microelectrodes were inserted into the dorsal gray column I cm caudal to the collagen matrix distal cord stump junction. A reference Ag/AgCI electrode was ph,ced in a subcutaneous pouch. Cats were exposed to 5% hydrogen gas an0 baseline spinal cord blood flow were recorded using the initial slope technique 18, Following spinal cord blood flow measurements and without disturbing the recording electrodes, surgical reconstruction cats had the pedicled omentum clamped above its point of contact with the collagen matrix and additional blood flows were taken. After spinal cord blood flow measurements, a 2% aqueous solution of the retrograde axonal tracer Fluoro-Gold (5 p.I), was injected into each column of the cord approximately 2 cm distal to the collagen matrix. The active constituent of Fluoro-Gold responsible for its retrograde labeling of cells is a 280 Da hydroxystilbamidine and its mechanism of uptake has recently been reviewed "~s,Cats were allowed to recover and 14 days later were deeply anesthetized and killed by perfusing transcardially with I I of heparinized saline followed by 2 I of 4% paraformaldehyde in 0.1 M phosphate buffer adjusted to pH 7.4. The brain and spinal cord were removed and cut coronally into 2-3 cm tissue blocks. The blocks were incubated for 6 h in phosphate buffer and tissues were transferred to a ]0% sucrose solution in 0.1 M phosphate buffered saline and stored at 4°C for a minimum of 14 h.
Sagittal cryostat sections of individual brain and spinal cord blocks were taken and stained immunohistochemically using antiserum against: (i) tyrosine hydroxylase (TH, I : 1000), (ii) dopamine/3-hydroxylase (DBH, I: 1000), (iii) phenylethanolamine-N.methyltransferase (PNMT, !:1000), (iv) serotonin-like (5-HT, 1:1000), (v) glial fibrillaryacidic protein (GFAP, I : 400), (vi) synaptophysin (SYN, I : 10). Antiserum for TH, DBH, PNMT, Serotonin-like, GFAP and NPY were obtained from Eugene Tech (Allendale, NJ), antiserum for synaptophysin from Boehringer (Montreal, Quebec) and FluoroGold from Fluorochrome Inc. (Englewood, CO).
Serial sections from each block were also processed for FluoroGold, Nissl and Palmgren silver impregnation stains, lmmunoreaction of cord sections was done according to the peroxidase-antiperoxidase and fluoresceine isothiocyanate methods using antibodies to the above enzymes33. When used in serial sections, antiserum against TH, DBH and PNMT can differentiate contributions from dopaminergic, noradrenergic and adrenergic immunoreactive fibers. In addition, a number of sections were double-immunostained with TH or DBH using the peroxidase-antiperoxidase PAP technique and with the synaptogenic marker, synaptophysin (SYN), using the fluoresceine isothiocyanate method 33. RESULTS Neurologic examination performed 75 days after spinal cord transection by an observer blind to treatments showed one cat given 4-aminopyridine and another given laminin developing f o r e - h i n d l i m b coordinated locomotion when supported by the tail. We have not observed this type of locomotion in our experience with a large series of cord transected cats during related studies. The coordinated locomotion in these 2 cats became visibly impaired but not lost after 5 / z l of Fluoro-Gold was injected below the collagen matrix bridge for labelling purposes. This volume of FluoroGold is sufficient to cause focal axotomy and limited tissue necrosis. The remaining cats in control and experimental groups failed to show coordinated locomotor activity. Mild to no atrophy and lack of spasticity of hindlimb muscles was noted in all cats with surgical reconstruction. By contrast, cats with cord transection only, showed spastic extension of their hind legs accompanied by severe atrophy, of their thigh muscles. In group surgical reconstruction, the mean spinal cord blood flow measured 1 cm distal to the collagen matrix bridge was increased by the omental pedicle by 58% with a mean blood flow value of 43:1:8 m l / 1 0 0 g t i s s u e / m i n . When the omental pedicle blood supply to the collagen matrix bridge was temporarily occluded, blood flows dropped to a mean of 18:1:5 ml. Normal intact cat spinal cord blood flow in this region measured 51 :t: 4 ml which agrees with previous values s. Transection-only cats showed a mean spinal cord blood flow value of 11 :t: 4 ml, 1 cm caudal to the transection site. Histological examination of the distal cord below the transection was p e r f o r m e d on a blind to t r e a t m e n t basis and revealed 6 consistent findings in all group surgical reconstruction cats: (i) presence of dense bundles of T N and DBN fibers within and distal to the collagen matrix bridge (Fig. 1C, D); (ii) invasion of T N and D B H fibers into distal cord gray-white m a t t e r pathways (Fig. 1E); (iii) T H - D B H fiber sprouting for long anatomic distances distal to the collagen matrix bridge (Fig. IF); (iv) a p p e a r a n c e of large-size T N and
219 DBH ‘terminal staining dots’ on the somatic surface of adrenoceptive neurons (cells with catecholaminergic receptors) (Fig. 2); (v) large-size synaptophysin (SYN) varicosities on the somatic surface of preganglionic sympathetic neurons located distal to collagen matrix bridge (Fig. 3) and (vi) accumulation of Fluoro-Gold particles within brainstem neurons (Fig. 4). I’mmunohistochemical density of monoaminergic axons and other fibers (determined by antisera staining against TH, DBH, PNMT, S-HT, GFAP, NPY and SYN) was closely examined in 3 spinal cord regions using longitudinal sections: (i) proximal collagen matrix bridge junction (or proximal-transection site junction in transected-only cats); (ii) collagen matrix bridge (and
transection site); (iii) distal cord to collagen matrix bridge (and distal to transection site). Examination of proximal collagen matrix junction showed near normal density of TH, DBH, PNMT and S-HT immunoreactive fibers in reconstructed cords but a considerable reduction of these fibers in the same region of cord transected-only animals. Immunofluorescence staining against GFAP showed minimal or no difference in the number of reactive astrocytes when surgical reconstruction treated cords were compared to normal cord tissue; however, intense staining of GFAP fibers was observed in sections from cord transectedonly animals at both proximal and distal junctions of the transection site. This finding indicates that reactive
Fig. 1. lmmunohistochemicalmicrographsof tyrosine hydroxylase (TH) immunoreactive fibers sectioned longitudinally from spinal cord of animalssubjectedto complete cord transection and surgical reconstruction (see text for details). TH positive fibers (seen as chains of small dots) indicate the presence of supraspinal (originating in brain) dopaminergic and noradrenergic axon outgrowth seen at (a) 200 pm proximal to collagen matrix bridge; (b) proximal junction to collagen matrix bridge, (c) center of collagen matrix bridge; (d) distal junction to collagen matrix bridge; (e) 250 pm distal to collagen matrix bridge, and (f) 90 mm distal to collagen matrix bridge (central canal * is seen below). Dopamine-@-hydroxylase (DBH) immunostaining (not shown) revealed a similar pattern to TH immunoreaction but the density of positive fibers in collagen matrix bridge and distal cord was approximately half of that seen after TH staining. Note very fine TH-positive varicosities surrounding proximal spinal neurons (a) slightly increasing their size as they approximate and enter collagen matrix (CM) bridge (b,c) and exit to distal cord (d-f). Larger black dots (d,e) are blood cells reacting to peroxidase stain. Bar = 25 pm.
221 astrocytosis is attenuated in collagen-omentum treated cords but not following cord transection-only. The reason for this phenomenon remains unclear although we have observed this lack of gliosis in our previous series of cats similarly subjected to surgical reconstruction of the spinal cord after transection m°. It should be noted that reactive astrocytes resulting from spinal cord injury may inhibit axonal regeneration in mammals 2s,3m through a neuronoglial interactive process 29. There were no PNMT or 5-HT fibers present within the collagen matrix bridge or at the transection site, nor in the distal cord of any animal. Within the collagen matrix bridge, dense TH-DBH bundles of fibers were present in an estimated "2:1 ratio. The same TH-DBH fiber density ratio was observed in the distal cord-collagen matrix bridge junction with these bundles invading gray and white matter pathways where catecholaminergic tracts are normally found (Fig. 1). No TH or DBH fibers were found at the tissue bridge site which developed in transected-only controls. The 'fibrotic bridge' forming between the proximal-distal stumps in transection-only cats was composed of collagenous connective tissue derived mostly from proliferative fibroblasts, gliai cells and macrophages. The longest distal cord outgrowth of TH-DBH immunoreactive fibers was observed in 2 cats subjected to surgical reconstruction whose collagen matrix bridge was treated with 4-aminopyridine or with laminin. THDBH fibers extended maximally 90 mm with 4-aminopyridine treatment and 76 mm with laminin treatment (Fig. 1). This rate of fiber outgrowth (approximately 1 mm/day) is comparable to axon regeneration 22. The average maximal TH-DBH outgrowth for all surgical reconstruction animals treated with collagen matrix omentai pedicle neuroactive agent, was 66 mm when measured from the collagen matrix bridge-distal cord junction. Treatment of the collagen matrix in surgical reconstruction cats with gila maturation factor or lipid angiogenic factor resembled saline treatment; maximal
TH-DBH distal outgrowth in these groups averaged 55 mm as measured from the collagen matrix bridge-distal cord junction. Transection-only cords showed no TH-DBH fiber outgrowth below the transection site. The relative contribution of dopaminergic and noradrenergic fiber outgrowth distal to the collagen matrix bridge after surgical reconstruction was determined by serial sections stained with TH or DBH antiserum which are the respective catecholamine-synthesizing enzymes to dopamine and noradrenaline. We found that dopaminergic and noradrenergic fibers appeared equally distributed within the distal cord intermediolateral region but dopaminergic fibers were more abundant in the lateral funiculus and in the area surrounding the central canal. Distal to the collagen matrix bridge, dopaminergic and noradrenergic 'terminal staining dots' (characterized by large-size TH-DBH varicosities) were observed on the somatic surface of gray and white matter adrenoceptive neurons. These terminal staining dots were consistently noted on preganglionic sympathetic neurons and on their processes in the region of the intermediolateral nucleus. Dopaminergic terminal staining dots were also noted on lateral funiculus preganglionic sympathetic neurons in the white matter. A small number of gray matter dorsal and ventral horn neurons contained both dopaminergic and noradrenergic terminal staining dots. Double immunostaining showed that SYN varicosities were often but not always found adjacent to the TH or DBH terminal staining dots when these were detected on the surface of adrenoceptive elements distal to the collagen matrix bridge (Fig. 2). Although SYN is normally present in spinal cord gray but not white matter 2~, reconstructed cord tissue showed many large-size SYN varicosities often in contact with adrenoceptive neurons such as preganglionic sympathetic neurons, in both gray and white matter (Fig. 3). Other atypical SYN varicosities appeared
Fig. 2. Double immunostaining of (a) synaptophysin (SYN) and (b) tyrosine hydroxylase (TH) immunoreactive fibers using immunohistofluorescence and immunohistochemical techniques of distal cord tissue in cats subjected to transection and surgical reconstruction of the spinal cord. Atypically large SYN varicosities are seen surrounding an adrenoceptive cell and its process (a,b: 9 ). Terminal staining dots are seen on several adrenoceptive cells (b: 1'1') in association with SYN varicosities (a: 1`1`). Large black stain on top of micrographs is a small tear in the section. Normal (above the transection site) punctate size, homogenous SYN varicosities (c) and TH-positive fibers (d) are shown for comparison surrounding a large, ventral horn motor neuron (N). Bar = 25/zm. Fig. 3. lmmunofluorescence micrograph of large-size synaptophysin (SYN) varicosities (1') on the somatic surface of several preganglionic sympathetic neurons located distally to collagen matrix bridge in surgically reconstructed spinal cord. Region is the lateral funiculus of the white matter. Bar -- 50 ~m. Fig. 4. Photomicrograph demonstrating intraojtoplasmic accumulation of the retrograde axonal tracer Fluoro-Gold in cells of the locus coeruleus (ventromedial to the motor trigeminal tract) in cat pons. Fluoro-Gold was injected 14 days before in the distal spinal cord of a cat subjected to surgical reconstruction plus 4-aminopyridine treatment. Bar = 25/zm.
222 densely distributed in the proximal and distal cord junctions to the collagen matrix bridge. These atypical SYN varicosities extended 7-15 mm into the distal cord where they eventually returned to their normal size and appearance. Large-size SYN varicosities are described here as 'atypical' because in normal spinal cord they are punctate in size and homogenous in appearance 2s (Fig. 2). Accumulation of the retrograde axonai marker Fluoro-Gold in the brain following its distal spinal cord injection 76 days after spinal cord surgical reconstruction was noted in cerebral areas known to contribute direct descending spinal cord pathways. Fluoro-Gold was observed within cells and processes of the locus coeruleus, subcoeruleus, Kolliker-Fuse and scattered regions in the medulla, including lateral reticular nucleus and an area lateral to the inferior olive (Fig. 4). These brainstem nuclei are known to contain noradrenergic cell bodies which project to the spinal cord and make synaptic contacts with adrenoceptive cells including preganglionic sympathetic neurons and ventral/dorsal horn cells mL~7'26'3°'32'36. Fluoro-Gold was also observed within neurons known to send cholinergic fibers to the spinal cord 23. Cholinergic neurons accumulating Fluoro-Gold included the red nucleus and scattered arcas of the ponto-mesencephalic reticular formation particularly gigantoceilular neurons. No Fluoro-Gold particles were found in the raphe complex (spinal source of serotoninergic neurons) or rostral medulla CI-C2 cell groups (spinal source of adrenergic neurons). No Fluoro-Gold was observed in the brainstem of transection-only control cats. DISCUSSION We had previously shown that surgically reconstructed transected cat spinal cord can develop limited catecholaminergic outgrowth distal to the transection sitC '~°. In those previous experiments, the transected cord stumps were bridged with a collagen matrix which received added blood supply from a pedicled omental graft ~'~°. We observed in those preliminary studies that supraspinai catecholaminergic fibers crossed the collagen matrix bridge and continued distally for a maximal distance of 9.6 mm during the 12 week recovery period. in addition, spinal cord blood flow in the distal host tissue bordering the transection site was increased by a mean of 53% after omental grafting as compared to cats receiving a collagen matrix bridge but no omenturn Hi. Moreover, cats treated either with collagen matrix only or with omental pedicle only, showed no catecholaminergic outgrowth distal to the transection site "~.
We also observed that somatosensory evoked potentials were present in 4 of 4 cord transected cats subjected to collagen matrix-omentum graft when a stimulus was applied below the lesion site ~°. Cats with a positive somatosensory evoked potentials also demonstrated short distance retrograde axonal transport of HRP in the proximal cord when the marker was injected distal to the collagen matrix bridge. The present study is the logical continuation of our previous experiments and an attempt to clarify some conditions governing central regeneration after axotomy. Moreover, one of the primary goals of the present investigation was to further improve distal cord supraspinal regeneration, specifically maximal distance outgrowth, axonal density and appropriate target tissue reconnection caudal to the transection site. The findings from this study indicate that considerable regeneration of supraspinal dopaminergic and noradrenergic fibers develop distal to the transected spinal cord following surgical reconstruction using an omental graft on a collagen matrix bridge treated with a neuroactive agent. Our reason for adding neuroprotective or neuronotrophic-like compounds was 2-fold. First, we wanted to evaluate the effects of these compounds on axonai density outgrowth within the collagen matrix; second, we wondered whether 'fixed-exposure' of these agents suspended in the solidified gel matrix could enhance the elongation of neuritic outgrowth coursing distal to the bridge. We observed both events when collagen matrix was treated with 4aminopyridine, a potassium channel blocker reported to improve conduction in demyelinated nerves a4 and laminin, a basement membrane protein known to stimulate central neurite outgrowth from neurons transplanted to rat brain 37. It is not clear from the present results whether the improved elongation and increased density of supraspinal fibers caudal to the collagen matrix bridge or the apparent neosynaptic contacts made by the regenerated fibers on distal cord tissue was due to: (a) a reduced bridge gap (from 6 to 3 mm), (b) the addition of a neuroactive agent to the collagen matrix, (c) improved surgical technique, (d) post-surgical stabilization of the collagen matrix bridge or (e) a combination of the above. However, 4 morpholegic findings appear to support the conclusion that significant improvement from prior experiments was achieved: (i) presence of dense bundles of dopaminergic and noradrenergic fibers within the collagen matrix bridge which extended caudally an average of 66 mm and a maximal distance of 90 mm (Fig. 1); (ii) atypical distribution of synaptophysin varicosities anatomically adjacent to the terminal staining dots (Fig. 2A,B); (iii) large-size TH-DBH 'terminal staining dots' on the
223 somatic surface and processes of distal adrenoceptive neurons (Fig. 3); (iv) retrograde transport of FluoroGold particles from the distal cord into cerebral neurons known to send dopaminergic and noradrenergic axons to the spinal cord (Fig. 4). The distribution and size of the SYN varicosities seen associated with TH-DBH terminal staining dots on the surface of preganglionic sympathetic neurons in surgically reconstructed cords suggests that synaptic remodelling of regenerating axons, specifically dopaminergic/noradrenergic fibers, but possibly other fibers too, developed on adrenoceptive target tissues distal to the collagen matrix bridge. However, this observation requires immunoelectron microscopic confirmation demonstrating that the TH and DBH terminal staining dots are indeed making synaptic contact with adrenoceptive neurons in distal cord. tissue. Chiba et al.3 have described cholinergic-dopaminergic-noradrenergic synaptic boutons on preganglionic sympathetic neurons and determined that preterminal catecholaminergic varicosities were about 0.08/zm in diameter while synaptic terminal varicosities (terminal staining dots) ranged from 0.5 to 1.5 /zm in diameter. Chiba's 3 findings are in agreement with Hagihira et ai. ~6 who observed noradrenergic synaptic terminal varicosities on adrenoceptive neurons in normal rat cord; the size of these varicosities is the same as the TH-DBH terminal staining dots we found distal to the collagen matrix bridge. An explanation concerning the significance or large-size SYN/TH varicosities on adrenoceptive elements scattered distal to the collagen matrix bridge may help clarify the present findings. SYN is a 38 kDa calcium-binding glycoprotein found in synaptic vesicle membrane and appears to be a good marker for synaptogenesis I~. Since SYN is also present in synaptic vesicles of growth cones and immature axons 24 but is not normally found in cord white matter 2s it is reasonable to suspect that the presence of atypically large-size SYN varicosities noted in gray-white matter distal to the collagen matrix bridge are an expression of 'synaptic remodelling' generated by the supraspinai outgrowth, particularly because these SYN varicosities were observed associated with TH-DBH 'terminal staining dots'. It should be noted moreover that atypical SYN varicosities were not observed in cats with transection-only caudal to the transection site and were only sporadically present at the proximal-transection site junction where abortive axonal sprouting normally occurs.
The 2:1 density ratio difference observed between dopamine and noradrenaline is of interest since noradrenaline content in the spinal cord is normally 10
times that of dopamine ~l. The 2:1 ratio between the fibers implies that dopaminergic axon sprouting is extremely active in comparison to noradrenergic outgrowth in this experiment. Although these preliminary findings may have considerable clinical implications, our results should be interpreted with caution until further work can determine the exact conditions governing the supraspinal outgrowth seen in the present model. Acknowledgements. This study was supported by the International Spinal Research Trust (UK). The authors thank Teresa Fortin and Julia Katnick for excellent technical assistance and Dr. Ramon Lira for supplying the/3-glia maturation factor. REFERENCES 1 Berry, M., Regeneration in the central nervous system, Recent Ado. Neuropathol., 67 (1979) 111-132. 2 Breasted, J.H., The Edwin Smith Surgical Papyrus, 2 Vols., University of Chicago Press, 1930. 3 Chiba, T. and Masuko, S., Synaptic structure of the monoamine and peptide nerve terminals in the intermediolateral nucleus of the guinea pig thoracic spinal cord,/. Comp. Neurol., 262 (1987) 242-255. 4 de la Torre, J.C., Spinal cord injury: a review of basic and applied research, Spine, 6 (1981) 315-355. 5 de la Torre, J.C., Spinal cord injury models, Prog. Neurobiol., 22 (1984) 289-344. 6 de la Torre, J.C., Catecholamine fiber regeneration across a collagen bioimplant after spinal cord transection, Brain Res. Bull., 9 (1982) 545-552. 7 de la Torre, J.C., Experimental problems in spinal cord neural reconstruction, in G. Gilad, A. Gorio and G. Kreutzberg (Eds.), Process of Recocery from Neural Trauma, Springer, New York, 1986, 317-325. 8 de la Torre, J.C., Hill, P.K., Gonzalez, M. and Parker, J., Transected spinal cord axons regenerate into a collagen bioimplant. In Nercous System Regeneration, Liss, New York, 1983, 515-520. 9 de la Torre, J.C. and Goldsmith, H.S., Increased blood flow enhances axon regeneration after spinal transection, Neurosci. Lett., 94 (1988) 269-273. 10 de la Torre, J.C. and Goldsmith, H.S., Collagen-omental graft in experimental spinal cord transection, Acta Neurochir., 102 (1990) 152-163. 11 Fleetwood-Walker, S.M. and Coote, J.H., Contribution of noradrenaline, dopamine and adrenaline-containing axons to their innervation of different regions of the spinal cord of the cat, Brain Res., 206 (1991) 95-106. 12 Gelderd, J.B., Evaluation of blood vessel and neurite growth into a collagen matrix placed within a surgically created gap in rat spinal cord, Brain Res., 511 (1990)80-92. 13 Gelderd. J.B. and Quarles, J., A preliminary study of homotopic fetal cortical and spinal co-transplants in adult rats, Brain Res. Bull., 25 (1990) 35-48. 14 Goldsmith, H.S. Steward, E., Chert, W.F. and DuckeR, S., Application of the intact omentum to normal and traumatized spinal cord. In C.C. Kao, A. Bunge, P. Reirer (Eds.), Spinal Reconstruction, Raven, New York, 1983, 235-243. 15 Goldsmith, H.S., Duckett, S. and Chert, W., Spinal cord vascularization by intact omentum, Am. J. Surg., 129 (1975) 262-265. 16 Hagihira, S., Senba, E., Yoshida, S., Tohyama, M. and Yoshiya, I., Fine structure of noradrenergic terminals and their synapses in the rat spinal dorsal horn: an immunohistochemical study, Brain Res., 526 (1990) 73-80. 17 H6kfelt, T., Phillipson, O. and Goldstein, M., Evidence for a dopaminergic pathway in the rat descending from All cell group to the spinal cord, Acta Physiol. Scand., 107 (1979) 393-395.
224 18 lngvar, D.H. and Lassen, N.A., Regional blood flow of the cerebral cortex determined by krypton 85, Acta PhysioL Scand., 54 (1962), 325-338. 19 Jahne, R., Schiebler, W., Ouimet, C. and Greengard, P., A 3,000-dalton membrane protein (p 38) present in synaptic vesicles, Proc. Natl. Acad. Sci. USA, 82 (1985) 4137-4141. 20 Kawaja, M.D. and Gage, F., Nerce Growth Factor Producing Primao' Fibroblasts Within a Collagen Matr& Serce as Conductice Encironment for Axon Regrowth in Vico, Abstracts Third IBRO World Congress of Neuroscience, Montreal, Canada, August 4-9, 1991, p. 21. 21 Kiernan, J.A., Hypotheses concerned with axonal regeneration in the mammalian nervous system, Biol. Rel'.o 54 (1979) 155-197. 22 Kiernan, J., An explanation of axonal regeneration in peripheral nerves and its failure in the central nervous system, Med, Hypothese, 4 (1978) 15-26. 23 Kimura. H., McGeer, P.L., Pang, J. and McGeer, E.G., The central cholinergic system studied by choline acetyltransferase immunohistochemistry in the cat, J. Comp. NeuroL, 200 (1981) 151-2111. 24 Leclerc. N., Beesley, P., Brown, I. et al.. Synaptophysin expression during synaptogenesis in the rat cerebellar cortex, J. Comp. Neurol., 280 (1989) 197-212. 25 Liuzzi, F.J. and Lasek, R.J., Astrocytes block axonal regeneration in mammals by activating physiological stop pathway, Science, 237 (1987) 642-645. 26 Magnusson. T.. Effect of chronic transection on dopamine, noradrenaline and 5-hydroxytryptamine in the rat spinal cord, Naunyn.Schmiedeherg ~" Arch. Pathol. Pharnmcol., 278 (1973) 1322. 27 Marchand, R. and Woerly, S., Transected spinal cords grafted with in situ self-assembled collagen matrices, Neuroscience, 36 (1990) 45-611.
28 Miller, D.C., Koslow, M., Budzilovich, G. and Burstein, D., Synaptophysin: a sensitive and specific marker for ganglion cells in central neurons system neoplasms, Human Pathol., 21 (1990) 271-276. 29 Moonen, G., Rogister, B., Leprince, P., Rigo, J.M., Delree, P., Lefebre, P. and Schoenen, J., Neurono-glial interactions and neural plasticity, Prog. Brain Res., 86 (1990) 63-73. 30 Petras, J.M. and Cummings, J.F., Autonomic neurons in the spinal cord of the rhesus monkey: a correlation of the findings of cytoarchitectonics and sympathectomy with fiber degeneration following dorsal and hizotomy, J. Comp. Neuroi., 146 (1~72) 189-218. 31 Preddy, R. and Malhotra, S., Reactive astrocytes in lesioned rat spinal cord: effect of neural transplants, Brain Res. Bull., 22 (1989) 81-87. 32 Ross, R.A. and Reis, DJ., Effects of lesions of locus coeruleus on regional distribution of dopamine /3-hydroxylase activity in rat brain, Brain Res., 73 (1974) 161-166. 33 Sternberger, L.A., Immunocytochemistry, 3rd edn., Wiley, New York, 1986. 34 Targ, E.F. and Kocsis, J., 4-Aminopyridine leads to restoration of conduction in demyelinated rat sciatic nerve, Brain Res., 328 (1985) 358-361. 35 Wessendorf, M.W., Fluoro-Gold: composition and mechanism of uptake, Brain Res,, 553 (1991) 135-148. 36 Westlund, K.N., Bowker, R., Zeigler, M. and Coulter, J., Noradrenergic projections to the spinal cord, Brain Res., 263 (1983) 15-1. 37 Zhou, F.C. and Azmitia E., Laminin facilitates and guides fiber outgrowth of transplanted neurons in adult brain, J. Chem. Neuroanat., i (1988) 133-146.
217
© 1992 Elsevier Science Publishers B.V. All rights reserved 0006-8993/92/$05.00
BRES 18001
Axonal regeneration after spinal cord transection and reconstruction H a r r y S. G o l d s m i t h
a
a n d J.C. de la T o r r e
b
a Boston UniversitySchool of Medicine, Department of Surgery, Boston, MA 02118 (USA) and b Division of Neurosurgery, Universityof Ottawa, Faculty of Medicine, Ottawa, Ont. (Canada) (Accepted 24 March 1992)
Key. words: Axonal regeneration; Surgical reconstruction; Cord transection; Matrix bridge
Following complete transection of the spinal cord, cats were separated into 2 groups to undergo: (i) surgical reconstruction of the disconnected cord using a neuroactive agent mixed into a collagen matrix bridge and omental transposition and (ii) cord transection-only. After 90 days, animals were killed and the brain and spinal cord were removed for immunohistochemistry. Two weeks prior to sacrifice, spinal cord blood flows were measured and the retrograde axonal tracer Fluoro-Gold was injected below the transection site. Gross inspection of the spinal cords at autopsy showed excellent integration and continuity of the collagen matrix bridge with the proximal-distal stumps in the surgical reconstruction group. In the transection-only group, the proximal-distal stumps were connected by a fibrotic, often tapered in the middle, tissue bridge. Results show that omental transposition in the surgical reconstruction group increased spinal cord blood flow by 58% when compared to transection-only animals. Fiuoro-Gold was found in mesencephalic and brainstem catecholaminergic and cholinergic neurons known to send axons to the spinal cord. Immunohistochemical staining with antibodies against catecholamine synthesizing enzymes tyrosine hydroxylase (TH) and dopamine-/3-hydro~lase (DBH) showed that surgical reconstruction treated cat cords but not transection-only, developed dense bundles of dopaminergic and noradrenergic fibers which were present in the collagen matrix bridge and in the distal spinal cord. Extension of these catecholaminergic fibers in surgical reconstruction treated cats showed maximal outgrowth of 90 mm below the transection site when the neuroactive agent 4-aminopyridine was mixed into the collagen matrix, in addition, the synaptogenic marker synaptophysin (SYN) was observed on preganglionic sympathetic neurons in association with dopaminergic- and noradrenergic-containing varicosities distal to the collagen matrix bridge, an indication that neo-synaptic contacts may have been made on these previously denervated neurons. No TH, DBH or SYN was observed below the transection site in transection-only cats. These findings indicate that surgical reconstruction treated cords can develop dense supraspinal fiber outgrowth across a treated collagen matrix bridge fed by an omental blood supply and that these fibers may have made neo-synaptic contacts with n.npmpriate distal spinal cord target tissue,
INTRODUCTION
Functional regeneration of mammalian spinal cord fibers after total cord transection remains as elusive today as when it was described by the Egyptians 5,0~ years ago 2. A number of crucial physic.al and chemical obstacles have been reported to impede successful regeneration of axotomized central spinal fibers t'4,s'7,9,2~. Although successful reconnection of lesioned peripheral nerve fibers is possible in most species, including mammals, regeneration of descending spinal pathways originating in the brain (supraspinal fibers) is limited to lower non-mammalian animals such as fish and frogs 2~'22. We first reported that limited outgrowth of supraspinal catecholaminergic fibers could be induced
using a collagen matrix to bridge the transected spinal cord stumps in rats ~'s. Subsequent studies in cats showed that placing a lengthened omental pedicle ~4't5 on the surface of the collagen matrix provided increased blood flow to the transection site and surrounding host tissue and promoted robust supraspinal axon regeneration into the collagen bridge and nearby the distal cord stump 9'rap. More recent studies by other investigators have confirmed the usefulness of collagen matrix when used as a bridge or support frame for cellular, neuritic and vascular outgrowth in CNS tissue n.m3,2°'27. In this study, we attempted to create an even richer physiological microenvironment for the transected descending spinal fibers most of which originate in the brainstem (supraspinal), by adding a neuroactive agent to the connecting collagen bridge and by
Correspondence: J.C. de la Torre, Division of Neurosurgew, !,rniversity of Ottawa, 451 Smyth Road, Ottawa, Ont. KIH 8M5, Canada. Fax: (1) (613) 738-3191.
218 shortening the
interstump gap
distance from 6 mm to 3
mm.
MATERIALS AND METHODS Twelve anesthetized female adult cats were subjected to a sterile laminectomy between "Is and T~o. After removing the dura at the site, the exposed cord was completely transected at T,~ and verified by microscopy. Before and after transection, the cord was bathed in ice-cold physiological saline for 10 min to reduce tissue metabolic activity, facilitate cord cutting and promote hemostasis. Cats undergoing surgical reconstruction had sterile collagen matrix (Collagen Corp., Paid Alto, CA) dispensed into the 3 mm gap created between the transected cord stumps. This collagen matrix is fluid at 4°C and hardens into a gel by heat polymerization at body temperature within 40 mint(). The collagen matrix is derived from purified pepsin, solubilized bovine collagen and contains 35 mg active collagen/ml in phosphate buffer at pH 7.6. it is obtained and dispensed in sterile condition. One of the following neurotrophic-like substances was mixed into the collagen matrix gel before it hardened: 4-aminopyridine (5/zg), laminin (5 p.g), O-gila maturation factor (10 ~.1), omental lipid angiogenic factor (5 ttg) and physiologic saline (5 p.I). Two cats underwent spinal cord transection only and received no collagen matrix, omentum or neurotrophic-like treatment. Two cats did not undergo any surgery and were kept as intact controls. In group surgical reconstruction cats, a 4 cm left subcostal laparotomy incision was made. The omentum was then gently pulled out of the abdominal cavity, lengthened into a pedicle while maintaining its blood supply intact, tunnelled subcutaneously to the transected cord and placed on the dorsal surface of the hardened collagen matrix where it was anchored with several sutures to the edges of the cut dura I°'ts. All incisions were closed in layers and animals were allowed to recover for 90 days. On day 76 after cord transection, all cats were anesthetized and spinal cord blood flows were recorded using the hydrogen washout method I°, Insulated platinum-iridium microelectrodes were inserted into the dorsal gray column I cm caudal to the collagen matrix distal cord stump junction. A reference Ag/AgCI electrode was ph,ced in a subcutaneous pouch. Cats were exposed to 5% hydrogen gas an0 baseline spinal cord blood flow were recorded using the initial slope technique 18, Following spinal cord blood flow measurements and without disturbing the recording electrodes, surgical reconstruction cats had the pedicled omentum clamped above its point of contact with the collagen matrix and additional blood flows were taken. After spinal cord blood flow measurements, a 2% aqueous solution of the retrograde axonal tracer Fluoro-Gold (5 p.I), was injected into each column of the cord approximately 2 cm distal to the collagen matrix. The active constituent of Fluoro-Gold responsible for its retrograde labeling of cells is a 280 Da hydroxystilbamidine and its mechanism of uptake has recently been reviewed "~s,Cats were allowed to recover and 14 days later were deeply anesthetized and killed by perfusing transcardially with I I of heparinized saline followed by 2 I of 4% paraformaldehyde in 0.1 M phosphate buffer adjusted to pH 7.4. The brain and spinal cord were removed and cut coronally into 2-3 cm tissue blocks. The blocks were incubated for 6 h in phosphate buffer and tissues were transferred to a ]0% sucrose solution in 0.1 M phosphate buffered saline and stored at 4°C for a minimum of 14 h.
Sagittal cryostat sections of individual brain and spinal cord blocks were taken and stained immunohistochemically using antiserum against: (i) tyrosine hydroxylase (TH, I : 1000), (ii) dopamine/3-hydroxylase (DBH, I: 1000), (iii) phenylethanolamine-N.methyltransferase (PNMT, !:1000), (iv) serotonin-like (5-HT, 1:1000), (v) glial fibrillaryacidic protein (GFAP, I : 400), (vi) synaptophysin (SYN, I : 10). Antiserum for TH, DBH, PNMT, Serotonin-like, GFAP and NPY were obtained from Eugene Tech (Allendale, NJ), antiserum for synaptophysin from Boehringer (Montreal, Quebec) and FluoroGold from Fluorochrome Inc. (Englewood, CO).
Serial sections from each block were also processed for FluoroGold, Nissl and Palmgren silver impregnation stains, lmmunoreaction of cord sections was done according to the peroxidase-antiperoxidase and fluoresceine isothiocyanate methods using antibodies to the above enzymes33. When used in serial sections, antiserum against TH, DBH and PNMT can differentiate contributions from dopaminergic, noradrenergic and adrenergic immunoreactive fibers. In addition, a number of sections were double-immunostained with TH or DBH using the peroxidase-antiperoxidase PAP technique and with the synaptogenic marker, synaptophysin (SYN), using the fluoresceine isothiocyanate method 33. RESULTS Neurologic examination performed 75 days after spinal cord transection by an observer blind to treatments showed one cat given 4-aminopyridine and another given laminin developing f o r e - h i n d l i m b coordinated locomotion when supported by the tail. We have not observed this type of locomotion in our experience with a large series of cord transected cats during related studies. The coordinated locomotion in these 2 cats became visibly impaired but not lost after 5 / z l of Fluoro-Gold was injected below the collagen matrix bridge for labelling purposes. This volume of FluoroGold is sufficient to cause focal axotomy and limited tissue necrosis. The remaining cats in control and experimental groups failed to show coordinated locomotor activity. Mild to no atrophy and lack of spasticity of hindlimb muscles was noted in all cats with surgical reconstruction. By contrast, cats with cord transection only, showed spastic extension of their hind legs accompanied by severe atrophy, of their thigh muscles. In group surgical reconstruction, the mean spinal cord blood flow measured 1 cm distal to the collagen matrix bridge was increased by the omental pedicle by 58% with a mean blood flow value of 43:1:8 m l / 1 0 0 g t i s s u e / m i n . When the omental pedicle blood supply to the collagen matrix bridge was temporarily occluded, blood flows dropped to a mean of 18:1:5 ml. Normal intact cat spinal cord blood flow in this region measured 51 :t: 4 ml which agrees with previous values s. Transection-only cats showed a mean spinal cord blood flow value of 11 :t: 4 ml, 1 cm caudal to the transection site. Histological examination of the distal cord below the transection was p e r f o r m e d on a blind to t r e a t m e n t basis and revealed 6 consistent findings in all group surgical reconstruction cats: (i) presence of dense bundles of T N and DBN fibers within and distal to the collagen matrix bridge (Fig. 1C, D); (ii) invasion of T N and D B H fibers into distal cord gray-white m a t t e r pathways (Fig. 1E); (iii) T H - D B H fiber sprouting for long anatomic distances distal to the collagen matrix bridge (Fig. IF); (iv) a p p e a r a n c e of large-size T N and
219 DBH ‘terminal staining dots’ on the somatic surface of adrenoceptive neurons (cells with catecholaminergic receptors) (Fig. 2); (v) large-size synaptophysin (SYN) varicosities on the somatic surface of preganglionic sympathetic neurons located distal to collagen matrix bridge (Fig. 3) and (vi) accumulation of Fluoro-Gold particles within brainstem neurons (Fig. 4). I’mmunohistochemical density of monoaminergic axons and other fibers (determined by antisera staining against TH, DBH, PNMT, S-HT, GFAP, NPY and SYN) was closely examined in 3 spinal cord regions using longitudinal sections: (i) proximal collagen matrix bridge junction (or proximal-transection site junction in transected-only cats); (ii) collagen matrix bridge (and
transection site); (iii) distal cord to collagen matrix bridge (and distal to transection site). Examination of proximal collagen matrix junction showed near normal density of TH, DBH, PNMT and S-HT immunoreactive fibers in reconstructed cords but a considerable reduction of these fibers in the same region of cord transected-only animals. Immunofluorescence staining against GFAP showed minimal or no difference in the number of reactive astrocytes when surgical reconstruction treated cords were compared to normal cord tissue; however, intense staining of GFAP fibers was observed in sections from cord transectedonly animals at both proximal and distal junctions of the transection site. This finding indicates that reactive
Fig. 1. lmmunohistochemicalmicrographsof tyrosine hydroxylase (TH) immunoreactive fibers sectioned longitudinally from spinal cord of animalssubjectedto complete cord transection and surgical reconstruction (see text for details). TH positive fibers (seen as chains of small dots) indicate the presence of supraspinal (originating in brain) dopaminergic and noradrenergic axon outgrowth seen at (a) 200 pm proximal to collagen matrix bridge; (b) proximal junction to collagen matrix bridge, (c) center of collagen matrix bridge; (d) distal junction to collagen matrix bridge; (e) 250 pm distal to collagen matrix bridge, and (f) 90 mm distal to collagen matrix bridge (central canal * is seen below). Dopamine-@-hydroxylase (DBH) immunostaining (not shown) revealed a similar pattern to TH immunoreaction but the density of positive fibers in collagen matrix bridge and distal cord was approximately half of that seen after TH staining. Note very fine TH-positive varicosities surrounding proximal spinal neurons (a) slightly increasing their size as they approximate and enter collagen matrix (CM) bridge (b,c) and exit to distal cord (d-f). Larger black dots (d,e) are blood cells reacting to peroxidase stain. Bar = 25 pm.
221 astrocytosis is attenuated in collagen-omentum treated cords but not following cord transection-only. The reason for this phenomenon remains unclear although we have observed this lack of gliosis in our previous series of cats similarly subjected to surgical reconstruction of the spinal cord after transection m°. It should be noted that reactive astrocytes resulting from spinal cord injury may inhibit axonal regeneration in mammals 2s,3m through a neuronoglial interactive process 29. There were no PNMT or 5-HT fibers present within the collagen matrix bridge or at the transection site, nor in the distal cord of any animal. Within the collagen matrix bridge, dense TH-DBH bundles of fibers were present in an estimated "2:1 ratio. The same TH-DBH fiber density ratio was observed in the distal cord-collagen matrix bridge junction with these bundles invading gray and white matter pathways where catecholaminergic tracts are normally found (Fig. 1). No TH or DBH fibers were found at the tissue bridge site which developed in transected-only controls. The 'fibrotic bridge' forming between the proximal-distal stumps in transection-only cats was composed of collagenous connective tissue derived mostly from proliferative fibroblasts, gliai cells and macrophages. The longest distal cord outgrowth of TH-DBH immunoreactive fibers was observed in 2 cats subjected to surgical reconstruction whose collagen matrix bridge was treated with 4-aminopyridine or with laminin. THDBH fibers extended maximally 90 mm with 4-aminopyridine treatment and 76 mm with laminin treatment (Fig. 1). This rate of fiber outgrowth (approximately 1 mm/day) is comparable to axon regeneration 22. The average maximal TH-DBH outgrowth for all surgical reconstruction animals treated with collagen matrix omentai pedicle neuroactive agent, was 66 mm when measured from the collagen matrix bridge-distal cord junction. Treatment of the collagen matrix in surgical reconstruction cats with gila maturation factor or lipid angiogenic factor resembled saline treatment; maximal
TH-DBH distal outgrowth in these groups averaged 55 mm as measured from the collagen matrix bridge-distal cord junction. Transection-only cords showed no TH-DBH fiber outgrowth below the transection site. The relative contribution of dopaminergic and noradrenergic fiber outgrowth distal to the collagen matrix bridge after surgical reconstruction was determined by serial sections stained with TH or DBH antiserum which are the respective catecholamine-synthesizing enzymes to dopamine and noradrenaline. We found that dopaminergic and noradrenergic fibers appeared equally distributed within the distal cord intermediolateral region but dopaminergic fibers were more abundant in the lateral funiculus and in the area surrounding the central canal. Distal to the collagen matrix bridge, dopaminergic and noradrenergic 'terminal staining dots' (characterized by large-size TH-DBH varicosities) were observed on the somatic surface of gray and white matter adrenoceptive neurons. These terminal staining dots were consistently noted on preganglionic sympathetic neurons and on their processes in the region of the intermediolateral nucleus. Dopaminergic terminal staining dots were also noted on lateral funiculus preganglionic sympathetic neurons in the white matter. A small number of gray matter dorsal and ventral horn neurons contained both dopaminergic and noradrenergic terminal staining dots. Double immunostaining showed that SYN varicosities were often but not always found adjacent to the TH or DBH terminal staining dots when these were detected on the surface of adrenoceptive elements distal to the collagen matrix bridge (Fig. 2). Although SYN is normally present in spinal cord gray but not white matter 2~, reconstructed cord tissue showed many large-size SYN varicosities often in contact with adrenoceptive neurons such as preganglionic sympathetic neurons, in both gray and white matter (Fig. 3). Other atypical SYN varicosities appeared
Fig. 2. Double immunostaining of (a) synaptophysin (SYN) and (b) tyrosine hydroxylase (TH) immunoreactive fibers using immunohistofluorescence and immunohistochemical techniques of distal cord tissue in cats subjected to transection and surgical reconstruction of the spinal cord. Atypically large SYN varicosities are seen surrounding an adrenoceptive cell and its process (a,b: 9 ). Terminal staining dots are seen on several adrenoceptive cells (b: 1'1') in association with SYN varicosities (a: 1`1`). Large black stain on top of micrographs is a small tear in the section. Normal (above the transection site) punctate size, homogenous SYN varicosities (c) and TH-positive fibers (d) are shown for comparison surrounding a large, ventral horn motor neuron (N). Bar = 25/zm. Fig. 3. lmmunofluorescence micrograph of large-size synaptophysin (SYN) varicosities (1') on the somatic surface of several preganglionic sympathetic neurons located distally to collagen matrix bridge in surgically reconstructed spinal cord. Region is the lateral funiculus of the white matter. Bar -- 50 ~m. Fig. 4. Photomicrograph demonstrating intraojtoplasmic accumulation of the retrograde axonal tracer Fluoro-Gold in cells of the locus coeruleus (ventromedial to the motor trigeminal tract) in cat pons. Fluoro-Gold was injected 14 days before in the distal spinal cord of a cat subjected to surgical reconstruction plus 4-aminopyridine treatment. Bar = 25/zm.
222 densely distributed in the proximal and distal cord junctions to the collagen matrix bridge. These atypical SYN varicosities extended 7-15 mm into the distal cord where they eventually returned to their normal size and appearance. Large-size SYN varicosities are described here as 'atypical' because in normal spinal cord they are punctate in size and homogenous in appearance 2s (Fig. 2). Accumulation of the retrograde axonai marker Fluoro-Gold in the brain following its distal spinal cord injection 76 days after spinal cord surgical reconstruction was noted in cerebral areas known to contribute direct descending spinal cord pathways. Fluoro-Gold was observed within cells and processes of the locus coeruleus, subcoeruleus, Kolliker-Fuse and scattered regions in the medulla, including lateral reticular nucleus and an area lateral to the inferior olive (Fig. 4). These brainstem nuclei are known to contain noradrenergic cell bodies which project to the spinal cord and make synaptic contacts with adrenoceptive cells including preganglionic sympathetic neurons and ventral/dorsal horn cells mL~7'26'3°'32'36. Fluoro-Gold was also observed within neurons known to send cholinergic fibers to the spinal cord 23. Cholinergic neurons accumulating Fluoro-Gold included the red nucleus and scattered arcas of the ponto-mesencephalic reticular formation particularly gigantoceilular neurons. No Fluoro-Gold particles were found in the raphe complex (spinal source of serotoninergic neurons) or rostral medulla CI-C2 cell groups (spinal source of adrenergic neurons). No Fluoro-Gold was observed in the brainstem of transection-only control cats. DISCUSSION We had previously shown that surgically reconstructed transected cat spinal cord can develop limited catecholaminergic outgrowth distal to the transection sitC '~°. In those previous experiments, the transected cord stumps were bridged with a collagen matrix which received added blood supply from a pedicled omental graft ~'~°. We observed in those preliminary studies that supraspinai catecholaminergic fibers crossed the collagen matrix bridge and continued distally for a maximal distance of 9.6 mm during the 12 week recovery period. in addition, spinal cord blood flow in the distal host tissue bordering the transection site was increased by a mean of 53% after omental grafting as compared to cats receiving a collagen matrix bridge but no omenturn Hi. Moreover, cats treated either with collagen matrix only or with omental pedicle only, showed no catecholaminergic outgrowth distal to the transection site "~.
We also observed that somatosensory evoked potentials were present in 4 of 4 cord transected cats subjected to collagen matrix-omentum graft when a stimulus was applied below the lesion site ~°. Cats with a positive somatosensory evoked potentials also demonstrated short distance retrograde axonal transport of HRP in the proximal cord when the marker was injected distal to the collagen matrix bridge. The present study is the logical continuation of our previous experiments and an attempt to clarify some conditions governing central regeneration after axotomy. Moreover, one of the primary goals of the present investigation was to further improve distal cord supraspinal regeneration, specifically maximal distance outgrowth, axonal density and appropriate target tissue reconnection caudal to the transection site. The findings from this study indicate that considerable regeneration of supraspinal dopaminergic and noradrenergic fibers develop distal to the transected spinal cord following surgical reconstruction using an omental graft on a collagen matrix bridge treated with a neuroactive agent. Our reason for adding neuroprotective or neuronotrophic-like compounds was 2-fold. First, we wanted to evaluate the effects of these compounds on axonai density outgrowth within the collagen matrix; second, we wondered whether 'fixed-exposure' of these agents suspended in the solidified gel matrix could enhance the elongation of neuritic outgrowth coursing distal to the bridge. We observed both events when collagen matrix was treated with 4aminopyridine, a potassium channel blocker reported to improve conduction in demyelinated nerves a4 and laminin, a basement membrane protein known to stimulate central neurite outgrowth from neurons transplanted to rat brain 37. It is not clear from the present results whether the improved elongation and increased density of supraspinal fibers caudal to the collagen matrix bridge or the apparent neosynaptic contacts made by the regenerated fibers on distal cord tissue was due to: (a) a reduced bridge gap (from 6 to 3 mm), (b) the addition of a neuroactive agent to the collagen matrix, (c) improved surgical technique, (d) post-surgical stabilization of the collagen matrix bridge or (e) a combination of the above. However, 4 morpholegic findings appear to support the conclusion that significant improvement from prior experiments was achieved: (i) presence of dense bundles of dopaminergic and noradrenergic fibers within the collagen matrix bridge which extended caudally an average of 66 mm and a maximal distance of 90 mm (Fig. 1); (ii) atypical distribution of synaptophysin varicosities anatomically adjacent to the terminal staining dots (Fig. 2A,B); (iii) large-size TH-DBH 'terminal staining dots' on the
223 somatic surface and processes of distal adrenoceptive neurons (Fig. 3); (iv) retrograde transport of FluoroGold particles from the distal cord into cerebral neurons known to send dopaminergic and noradrenergic axons to the spinal cord (Fig. 4). The distribution and size of the SYN varicosities seen associated with TH-DBH terminal staining dots on the surface of preganglionic sympathetic neurons in surgically reconstructed cords suggests that synaptic remodelling of regenerating axons, specifically dopaminergic/noradrenergic fibers, but possibly other fibers too, developed on adrenoceptive target tissues distal to the collagen matrix bridge. However, this observation requires immunoelectron microscopic confirmation demonstrating that the TH and DBH terminal staining dots are indeed making synaptic contact with adrenoceptive neurons in distal cord. tissue. Chiba et al.3 have described cholinergic-dopaminergic-noradrenergic synaptic boutons on preganglionic sympathetic neurons and determined that preterminal catecholaminergic varicosities were about 0.08/zm in diameter while synaptic terminal varicosities (terminal staining dots) ranged from 0.5 to 1.5 /zm in diameter. Chiba's 3 findings are in agreement with Hagihira et ai. ~6 who observed noradrenergic synaptic terminal varicosities on adrenoceptive neurons in normal rat cord; the size of these varicosities is the same as the TH-DBH terminal staining dots we found distal to the collagen matrix bridge. An explanation concerning the significance or large-size SYN/TH varicosities on adrenoceptive elements scattered distal to the collagen matrix bridge may help clarify the present findings. SYN is a 38 kDa calcium-binding glycoprotein found in synaptic vesicle membrane and appears to be a good marker for synaptogenesis I~. Since SYN is also present in synaptic vesicles of growth cones and immature axons 24 but is not normally found in cord white matter 2s it is reasonable to suspect that the presence of atypically large-size SYN varicosities noted in gray-white matter distal to the collagen matrix bridge are an expression of 'synaptic remodelling' generated by the supraspinai outgrowth, particularly because these SYN varicosities were observed associated with TH-DBH 'terminal staining dots'. It should be noted moreover that atypical SYN varicosities were not observed in cats with transection-only caudal to the transection site and were only sporadically present at the proximal-transection site junction where abortive axonal sprouting normally occurs.
The 2:1 density ratio difference observed between dopamine and noradrenaline is of interest since noradrenaline content in the spinal cord is normally 10
times that of dopamine ~l. The 2:1 ratio between the fibers implies that dopaminergic axon sprouting is extremely active in comparison to noradrenergic outgrowth in this experiment. Although these preliminary findings may have considerable clinical implications, our results should be interpreted with caution until further work can determine the exact conditions governing the supraspinal outgrowth seen in the present model. Acknowledgements. This study was supported by the International Spinal Research Trust (UK). The authors thank Teresa Fortin and Julia Katnick for excellent technical assistance and Dr. Ramon Lira for supplying the/3-glia maturation factor. REFERENCES 1 Berry, M., Regeneration in the central nervous system, Recent Ado. Neuropathol., 67 (1979) 111-132. 2 Breasted, J.H., The Edwin Smith Surgical Papyrus, 2 Vols., University of Chicago Press, 1930. 3 Chiba, T. and Masuko, S., Synaptic structure of the monoamine and peptide nerve terminals in the intermediolateral nucleus of the guinea pig thoracic spinal cord,/. Comp. Neurol., 262 (1987) 242-255. 4 de la Torre, J.C., Spinal cord injury: a review of basic and applied research, Spine, 6 (1981) 315-355. 5 de la Torre, J.C., Spinal cord injury models, Prog. Neurobiol., 22 (1984) 289-344. 6 de la Torre, J.C., Catecholamine fiber regeneration across a collagen bioimplant after spinal cord transection, Brain Res. Bull., 9 (1982) 545-552. 7 de la Torre, J.C., Experimental problems in spinal cord neural reconstruction, in G. Gilad, A. Gorio and G. Kreutzberg (Eds.), Process of Recocery from Neural Trauma, Springer, New York, 1986, 317-325. 8 de la Torre, J.C., Hill, P.K., Gonzalez, M. and Parker, J., Transected spinal cord axons regenerate into a collagen bioimplant. In Nercous System Regeneration, Liss, New York, 1983, 515-520. 9 de la Torre, J.C. and Goldsmith, H.S., Increased blood flow enhances axon regeneration after spinal transection, Neurosci. Lett., 94 (1988) 269-273. 10 de la Torre, J.C. and Goldsmith, H.S., Collagen-omental graft in experimental spinal cord transection, Acta Neurochir., 102 (1990) 152-163. 11 Fleetwood-Walker, S.M. and Coote, J.H., Contribution of noradrenaline, dopamine and adrenaline-containing axons to their innervation of different regions of the spinal cord of the cat, Brain Res., 206 (1991) 95-106. 12 Gelderd, J.B., Evaluation of blood vessel and neurite growth into a collagen matrix placed within a surgically created gap in rat spinal cord, Brain Res., 511 (1990)80-92. 13 Gelderd. J.B. and Quarles, J., A preliminary study of homotopic fetal cortical and spinal co-transplants in adult rats, Brain Res. Bull., 25 (1990) 35-48. 14 Goldsmith, H.S. Steward, E., Chert, W.F. and DuckeR, S., Application of the intact omentum to normal and traumatized spinal cord. In C.C. Kao, A. Bunge, P. Reirer (Eds.), Spinal Reconstruction, Raven, New York, 1983, 235-243. 15 Goldsmith, H.S., Duckett, S. and Chert, W., Spinal cord vascularization by intact omentum, Am. J. Surg., 129 (1975) 262-265. 16 Hagihira, S., Senba, E., Yoshida, S., Tohyama, M. and Yoshiya, I., Fine structure of noradrenergic terminals and their synapses in the rat spinal dorsal horn: an immunohistochemical study, Brain Res., 526 (1990) 73-80. 17 H6kfelt, T., Phillipson, O. and Goldstein, M., Evidence for a dopaminergic pathway in the rat descending from All cell group to the spinal cord, Acta Physiol. Scand., 107 (1979) 393-395.
224 18 lngvar, D.H. and Lassen, N.A., Regional blood flow of the cerebral cortex determined by krypton 85, Acta PhysioL Scand., 54 (1962), 325-338. 19 Jahne, R., Schiebler, W., Ouimet, C. and Greengard, P., A 3,000-dalton membrane protein (p 38) present in synaptic vesicles, Proc. Natl. Acad. Sci. USA, 82 (1985) 4137-4141. 20 Kawaja, M.D. and Gage, F., Nerce Growth Factor Producing Primao' Fibroblasts Within a Collagen Matr& Serce as Conductice Encironment for Axon Regrowth in Vico, Abstracts Third IBRO World Congress of Neuroscience, Montreal, Canada, August 4-9, 1991, p. 21. 21 Kiernan, J.A., Hypotheses concerned with axonal regeneration in the mammalian nervous system, Biol. Rel'.o 54 (1979) 155-197. 22 Kiernan, J., An explanation of axonal regeneration in peripheral nerves and its failure in the central nervous system, Med, Hypothese, 4 (1978) 15-26. 23 Kimura. H., McGeer, P.L., Pang, J. and McGeer, E.G., The central cholinergic system studied by choline acetyltransferase immunohistochemistry in the cat, J. Comp. NeuroL, 200 (1981) 151-2111. 24 Leclerc. N., Beesley, P., Brown, I. et al.. Synaptophysin expression during synaptogenesis in the rat cerebellar cortex, J. Comp. Neurol., 280 (1989) 197-212. 25 Liuzzi, F.J. and Lasek, R.J., Astrocytes block axonal regeneration in mammals by activating physiological stop pathway, Science, 237 (1987) 642-645. 26 Magnusson. T.. Effect of chronic transection on dopamine, noradrenaline and 5-hydroxytryptamine in the rat spinal cord, Naunyn.Schmiedeherg ~" Arch. Pathol. Pharnmcol., 278 (1973) 1322. 27 Marchand, R. and Woerly, S., Transected spinal cords grafted with in situ self-assembled collagen matrices, Neuroscience, 36 (1990) 45-611.
28 Miller, D.C., Koslow, M., Budzilovich, G. and Burstein, D., Synaptophysin: a sensitive and specific marker for ganglion cells in central neurons system neoplasms, Human Pathol., 21 (1990) 271-276. 29 Moonen, G., Rogister, B., Leprince, P., Rigo, J.M., Delree, P., Lefebre, P. and Schoenen, J., Neurono-glial interactions and neural plasticity, Prog. Brain Res., 86 (1990) 63-73. 30 Petras, J.M. and Cummings, J.F., Autonomic neurons in the spinal cord of the rhesus monkey: a correlation of the findings of cytoarchitectonics and sympathectomy with fiber degeneration following dorsal and hizotomy, J. Comp. Neuroi., 146 (1~72) 189-218. 31 Preddy, R. and Malhotra, S., Reactive astrocytes in lesioned rat spinal cord: effect of neural transplants, Brain Res. Bull., 22 (1989) 81-87. 32 Ross, R.A. and Reis, DJ., Effects of lesions of locus coeruleus on regional distribution of dopamine /3-hydroxylase activity in rat brain, Brain Res., 73 (1974) 161-166. 33 Sternberger, L.A., Immunocytochemistry, 3rd edn., Wiley, New York, 1986. 34 Targ, E.F. and Kocsis, J., 4-Aminopyridine leads to restoration of conduction in demyelinated rat sciatic nerve, Brain Res., 328 (1985) 358-361. 35 Wessendorf, M.W., Fluoro-Gold: composition and mechanism of uptake, Brain Res,, 553 (1991) 135-148. 36 Westlund, K.N., Bowker, R., Zeigler, M. and Coulter, J., Noradrenergic projections to the spinal cord, Brain Res., 263 (1983) 15-1. 37 Zhou, F.C. and Azmitia E., Laminin facilitates and guides fiber outgrowth of transplanted neurons in adult brain, J. Chem. Neuroanat., i (1988) 133-146.
