The roles of reactive gliosis and mitosis on tropic factor production in traumatized nervous system tissue

Proximal stumps of rat sciatic nerves were attached to inlet ends of Y-shaped silastic implants and offered a 'choice' of growing toward an Elvax pellet containing homogenate from previously crushed optic nerve which had been exposed to saline or cytosine arabinofuranoside (AraC). Previous studies indicate that AraC administration inhibits reactive gliosis in crushed optic nerve. Preferential or exclusive growth of axons occurred in implant forks attached to pellets containing saline- (vs drug-) exposed optic nerve homogenates. In contrast, inhibition of Schwann cell mitosis had no discernible effect on tropic factor production in distal stumps of transected sciatic nerves. Tropic activity of homogenates from cultures containing reactive-like astrocytes was nearly 4 times higher than homogenates not containing these cells. Results suggest a possible link between formation of hypertrophic reactive astrocytes and production of neurotropic factor.     

Tropic factors in reactive mammalian central nervous system tissue

Previous studies suggest that distal stumps of transected peripheral nerves contain diffusible factors which can attract/support axonal regeneration over distances of several mm in vivo. The present experiments were undertaken to determine if this is so for distal regions of traumatized central (i.e., optic) nerves. Proximal stumps of transected rat sciatic nerves were inserted into the single inlet ends of 6 mm long Y-shaped Silastic implants. Alternative ‘lures’ were attached to the paired outlets, the ability of these lures to attract/support regeneration of nerve fibers in their associated forks assessed 3.5 weeks postoperatively. Exclusive or preferential growth of nerve fibers occurred in implant forks associated with optic nerve grafts, of Elvax pellets containing homogenate obtained from previously crushed (reactive) optic nerves. Grafts of tendon, as well as homogenate from unoperated optic nerve had no effect.Results suggest that, with repect to the assay used, degenerating optic nerve tissue contains factor(s) which can attract/support regenerating nerve fibers.     

Specificity in mammalian peripheral nerve regeneration at the level of the nerve trunk

Previous studies indicate that distal stumps of transected rat peripheral nerves secrete 'tropic' factors which can attract/support axonal regeneration over distances of several mm in vivo. The present study was undertaken in order to determine if there is specificity of neurotropic interaction at the level of the nerve trunk. Proximal stumps of transected peroneal or tibial nerves were inserted into the single inlet end of Y-shaped Silastic implants and offered alternative 'lures' at the paired outlet ends (specifically, grafts of peroneal vs tibial distal stump tissue). Several weeks later, the overwhelming majority of preparations showed exclusive growth of nerve fibers in implant forks attached to 'native' (originally associated) nerve stumps. Inversion of the distal stump grafts (such that the proximal stump was facing an analogous native distal stump, but a different region of it) diminished the frequency and extent of native preference. Taken together, data suggest the possibility that there can be a specificity of nerve regeneration at the level of the nerve trunk.     

Regeneration of transected rat sciatic nerves after using isolated nerve fragments as distal inserts in silicone tubes

We determined blood vessel and perineurial fascicle densities as well as axonal numbers in regenerated rat sciatic nerves 8 weeks after the nerves had been transected, the proximal stumps placed into the proximal ends of silicone tubes, and isolated fragments of nerve placed into the distal ends of the same tubes. The data are compared with data from the normal nerve and from regeneration in a similar paradigm in which the distal stumps were used as the inserts into the distal end of the silicone tubes. A major difference between the two regeneration paradigms was that axons were discouraged from reaching the periphery when the distal insert was an isolated fragment and encouraged to reach the periphery when the distal insert was the distal stump. We found that fascicle and blood vessel densities were greater than normal but less than with the distal stump as the distal insert. Thus we concluded that the nature of the distal insert had a bearing on how many vessels and perineurial fascicles were formed during regeneration in these conditions. Myelinated axon numbers did not differ in the two conditions whereas there were more unmyelinated axons with the isolated distal stump as the distal insert. Thus at this regeneration time the numbers of myelinated axons were not as dependent on the nature of the distal insert as were the numbers of unmyelinated axons. Finally the length of the gap had a great influence on the numbers of axons that regenerated.(ABSTRACT TRUNCATED AT 250 WORDS)     

Nerve Regeneration in Predegenerated Basal Lamina Graft: The Effect of Duration of Predegeneration on Axonal Extension

We used predegenerated acellular grafts to bridge proximal and distal stumps of transected nerves and studied how the duration of predegeneration might affect axonal regeneration. Predegenerated acellular grafts were prepared by transecting the tibial nerve of donor rats and, after a period of degeneration, freeze-thawing a 40-mm long segment of the distal stump. Five degeneration periods were used: 0 days (for fresh grafts), 3 days, 1 week, 4 weeks, and 8 weeks. Fresh cellular grafts not treated with freeze-thawing were also used for comparison. Each graft was then transplanted to an isogeneic recipient rat, in which it was used to bridge the proximal stump of the transected left tibial nerve and the distal stamp of the transected right tibial nerve. Six weeks were allowed for the regeneration of axons in all grafts. The regeneration was then assessed by studying transverse sections of the grafts, to determine the maximum length that the axons had regenerated, and the packing density of axons (percentage of sampled areas occupied by axons). The results show that axons had grown to the maximum length in the 4-week predegenerated grafts, and had the highest packing density in the 1-week predegenerated grafts. Regeneration in the fresh acellular (0-day predegenerated) and 8-week predegenerated grafts, especially the latter, was poor. We examine the results with reference to time-dependent events of Wallerian degeneration and propose that there are beneficial effects of multiple factors on the grafts during the first 4 weeks of predegeneration, causing a slow but significant improvement in their capability to support axonal growth. The subsequent rapid deterioration of such capability may be related to structural changes in the extracellular scaffold.     

Axonal regeneration across transected mammalian spinal cords: An electron microscopic study of delayed microsurgical nerve grafting

A microneurosurgical technique is reported in which a spinal cord gap in the dog caused by transection 1 week previously was grafted with autogenous sciatic nerve segments. Electron microscopic studies disclosed that successful axonal regeneration had bridged the gap between transected spinal cord stumps via the grafted nerve. Several factors demonstrated that the regenerated axons crossing the grafted nerves were of spinal cord origin. Among several barriers which tended to block the advancement of regenerating axons was a glial basement membrane which formed at each end of the spinal cord stumps after spinal cord transection. Delayed nerve grafting resulted in delayed formation of the glial basement membrane, thus leading to successful axonal regeneration across the spinal cord gap.     

Prevention of gliotic scar formation by NeuroGel allows partial endogenous repair of transected cat spinal cord

Spinal cords of adult cats were transected and subsequently reconnected with the biocompatible porous poly (N-[2-hydroxypropyl] methacrylamide) hydrogel, NeuroGel. Tissue repair was examined at various time points from 6-21 months post reconstructive surgery. We examined two typical phenomena, astrogliosis and scar formation, in spines reconstructed with the gel and compared them to those from transected non-reconstructed spines. Confocal examination with double immunostaining for glial fibrillary acidic protein (GFAP) and myelin basic protein (MBP) showed that the interface formed between the hydrogel and the spine stumps did prevent scar formation and only a moderate gliosis was observed. The gel implant provided an adequate environment for growth of myelinated fibers and we saw angiogenesis within the gel. Electron microscopy showed that regenerating axons were myelinated by Schwann cells rather than oligodendrocytes. Moreover, the presence of the gel implant lead to a considerable reduction in damage to distal caudal portions of the spine as assessed by the presence of more intact myelinated fibers and a reduction of myelin degradation. Neurologic assessments of hindlimb movement at various times confirmed that spinal cord reconstruction was not only structural but also functional. We conclude that NeuroGel lead to functional recovery by providing a favorable substrate for regeneration of transected spinal cord, reducing glial scar formation and allowing angiogenesis.     

Local administration of vasoactive intestinal peptide after nerve transection accelerates early myelination and growth of regenerating axons

Our goal was to determine whether local injections of vasoactive intestinal peptide (VIP) promote early stages of regeneration after nerve transection. Sciatic nerves were transected bilaterally in 2 groups of 10 adult mice. In the first group, 15 microg (20 microL) of VIP were injected twice daily into the gap between transected ends of the right sciatic nerve for 7 days (4 mice) or 14 days (6 mice). The same number of mice in the second group received placebo injections (20 microL of 0.9% sterile saline) in the same site, twice daily, for the same periods. After 7 days, axon sizes, relationships with Schwann cells and degree of myelination were compared in electron micrographs of transversely sectioned distal ends of proximal stumps. Fourteen days after transection, light and electron microscopy were used to compare and measure axons and myelin sheaths in the transection gap, 2-mm distal to the ends of proximal stumps. Distal ends of VIP-treated proximal stumps contained larger axons 7 days after transection. More axons were in 1:1 relationships with Schwann cells and some of them were surrounded by thin myelin sheaths. In placebo-treated proximal stumps, axons were smaller, few were in 1:1 relationships with Schwann cells and no myelin sheaths were observed. In VIP-treated transection gaps, measurements 14 days after transection showed that larger axons were more numerous and their myelin sheaths were thicker. Our results suggest that in this nerve transection model, local administration of VIP promotes and accelerates early myelination and growth of regenerating axons.     

Mouse sciatic nerve regeneration through semipermeable tubes: A quantitative model

The regeneration of transected mouse sciatic nerves using semipermeable acrylic copolymer tubes to enclose both stumps has been qualitatively assessed from 1 to 30 weeks post-operative. Quantitative morphometric analysis of electron micrograph montages of complete transverse sections of the segment regenerated between stumps has permitted determinations of the percents of total area occupied by the various tissue constituents—blood vessels, epineurium, perineurium, endoneurium, myelinate d axon/Schwann cell units, and unmyelinated axon/Schwann cell units. Significant differences were found in the total cross-sectional area of segments regenerated through tubes of 1.0 mm versus 0.5 mm internal diameters. Segments regenerated with the distal stump inserted in the tube contained significantly greater percentages of neural units and were significantly larger at 8 weeks post-operative compared to segments regenerated for 9-10 weeks with the distal stump avulsed. The morphometric method permits rapid quantitation of sizeable electron micrograph montages which at 1300 × permit all types of tissue components, including the unmyelinated axons, to be visualized.     

Olivocerebellar Axon Regeneration and Target Reinnervation Following Dissociated Schwann Cell Grafts in Surgically Injured Cerebella of Adult Rats

The ability of Schwann cells to induce the regeneration of severed olivocerebellar and Purkinje cell axons across an injury up to their deafferented targets was tested by transplanting freshly dissociated cells from newborn rat sciatic nerves into surgically lesioned adult cerebella. The grafted glial cells consistently filled the lesion gap and migrated into the host parenchyma. Transected olivocerebellar axons vigorously regenerated into the graft, where their growth pattern and direction followed the arrangement of Schwann cell bundles. Although some of these axons terminated within the transplant, many of them rejoined the cerebellar parenchyma beyond the lesion. Here, their fate depended on the territory encountered. No growth occurred in the white matter. Numerous fibres penetrated into the granular layer and formed terminal branches that remained confined within this layer. A few of them, however, regenerated up to the molecular layer and formed climbing fibres on Purkinje cell dendrites. By contrast, the growth of transected Purkinje cell axons into the grafts was very poor. These results underscore the different intrinsic responsiveness of Purkinje cell and olivocerebellar axons to the growth-promoting action of Schwann cells, and show that the development and outcome of the regenerative phenomena is strongly conditioned by the spatial organization and specific features of the environmental cues encountered by the outgrowing axons along the course they follow. However, Schwann cells effectively bridge the lesion gap, induce the regeneration of olivocerebellar axons, and direct their growth up to the deafferented host cortex, where some of them succeed in reinnervating their natural targets.     

Temporal changes of neuronotrophic activities accumulating in vivo within nerve regeneration chambers

The presence of neuronotrophic factors (NTFs) in noninjured sciatic nerve extract and the course of their accumulation from 3 h to 30 days after nerve transection was examined. Rat sciatic nerves were transected and their proximal and distal stumps sutured into the openings of cylindrical silicone chambers leaving a 10-mm interstump gap. Previous studies had shown that regeneration occurs in chambers containing both stumps but is absent in chambers lacking the distal stump. Chambers became completely filled with fluid 10 to 12 h after implantation. Fluid from chambers without nerve stumps (open-ended) implanted adjacent to nerve-containing chambers had markedly lower trophic activities than those containing one or both stumps. In fluid collected from chambers containing both proximal and distal nerve stumps, the highest titers of NTFs directed to sensory neurons were measured at 3 h posttransection whereas the highest titers of NTFs directed to sympathetic and spinal cord neurons were detected at 1 and 3 days, respectively. Chambers containing only the proximal or only the distal stumps showed similar temporal dynamics for sensory and sympathetic NTFs. Sensory and sympathetic neuronotrophic activity in extracts of proximal and distal stumps followed a similar temporal course to those in chamber fluid. Extracts of nonlesion nerve segments 5 mm from the transection site contained higher sensory and lower sympathetic trophic activity than extracts including the transection site. Spinal cord activity was undetectable in all extracts. Antiserum to nerve growth factor had no effect on fluid or extracts containing high sensory or sympathetic activities. These observations suggested that (i) some NTFs may be present in normal nerves and others may be synthesized or accumulated in response to nerve injury, (ii) sensory, sympathetic, and spinal cord NTFs are separate agents and immunochemically distinct from nerve growth factor, (iii) NTFs predominantly originate from nerve stumps rather than from surrounding fluid, and (iv) proximal and distal nerve stumps accumulate and release NTFs at similar rates.     

Leukemia inhibitory factor enhances the regeneration of transected rat sciatic nerve and the function of reinnervated muscle

The cytokine leukemia inhibitory factor (LIF) favors the survival and growth of axons in vitro and in vivo. Fibronectin has been shown to enhance nerve regeneration when added in combination with various growth factors including LIF. The goal of this study was to evaluate the effect of LIF plus fibronectin on the regeneration of transected nerve and functional recovery of reinnervated skeletal muscle, in one experimental model of peripheral nerve repair, at two recovery times. The rat sciatic nerve was cut at mid-thigh level and a silicone cuff containing either saline (control), LIF, or LIF plus fibronectin (L + F) was used to bridge the proximal and distal nerve stumps leaving a 1 cm gap between them. Rats were then explored at 6 or 12 weeks following the initial surgery. Regenerating nerves were assessed by measuring the diameter of myelinated axons, conduction velocity, and number of myelinated fibers. Muscle reinnervation was assessed by measuring muscle mass, force of contraction, and histologically for changes in muscle fiber type (type I and type II). In this report we demonstrate that at 6 weeks there were significant increases in 1) nerve conduction velocity, 2) myelinated axon diameter, and 3) number of myelinated axons over that of control (saline-treated) animals. Both LIF groups demonstrated a shift in type II muscle fiber area compared to saline-treated controls, with the L + F group having a significant increase in muscle mass. At 12 weeks there was an improved recovery over and above that demonstrated at 6 weeks. Muscle mass was 65% and 42% greater than control for LIF and L + F, respectively. Force of contraction, conduction velocity, myelinated fiber number, and diameter were also significantly greater for both LIF- and L + F- treated rats than saline-treated rats. These results demonstrate that LIF significantly improves the regeneration of damaged peripheral nerves and the preservation of muscle viability, resulting in greatly enhanced recovery of skeletal muscle function. J. Neurosci. Res. 47:208–215, 1997. © 1997 Wiley-Liss, Inc.     

Current strategies in peripheral nerve repair

The paper presents main research trends focused on peripheral nerves regeneration. Modern surgery techniques in humans require dissection of the nerve bridges, leading to extra mutilation of a patient. The recovery of motor and sensory functions is also unsatisfactory. Therefore scientists are looking for new materials and techniques capable of replacing injured nerve trunks as well as of improving the reinnervation of target tissues. In this kind of experiments chambers joining proximal and distal stumps of transected peripheral nerve are very useful. Such chambers may also contain growth-promoting factors or glial cells supporting the regrowth. Progress in molecular biology and biomaterials engineering will be probably the most critical step to therapeutic intervention in nerve regeneration.     

Electromyographic evaluation of a novel surgical preparation to enhance nerve-muscle specificity that follows mammalian peripheral nerve trunk transection

Previous studies indicated that axons from proximal stumps of transected peripheral nerves "prefer" to grow through Silastic tubes attached to their native (originally associated) rather than foreign (not originally associated) distal stumps. We determined whether or not this specificity is expressed at the level of the neuromuscular junction. Proximal stumps of transected rat sciatic nerves (peroneal and tibial branches) were attached to single inlet ends of 6-mm-long, Y-shape Silastic implants. One outlet was attached to the distal peroneal and the other to the distal tibial stump. Ten weeks later, innervation of the anterior tibialis and interosseous muscles (normally innervated predominantly by peroneal and tibial nerve fibers, respectively) was assessed by measuring compound muscle action potential amplitudes and latencies that follow supramaximal peroneal and tibial nerve stimulation. Results showed higher amplitudes in anterior tibialis muscle, induced by "native" peroneal (vs. tibial) stimulation in four of five animals, and higher amplitudes in interosseous muscles after "native" tibial (vs. peroneal) stimulation in all cases examined. Preparations in which bridges between proximal and distal nerve stumps were bridged with unbranched tubes showed random patterns of muscle innervation. The results suggest that if allowed to express "specificity" at the level of nerve trunk transection, regenerating mammalian peripheral axons can grow into, and form functional connection with, native (vs. foreign) muscle groups. This finding has possible clinical significance.     

Reactions of unmyelinated nerve fibers to injury. An ultrastructural study

Reactions of unmyelinated nerves to injury were studied in the distal stumps of rabbit anterior mesenteric nerves following transection. These nerves, chosen because they are almost exclusively unmyelinated, were examined by phase contrast and electron microscopy at intervals from 12 h to 2 weeks after transection. Swollen axons containing mitochondria and other organelles were prominent in the proximal few mm of the distal stump of anterior mesenteric nerve trunks during the first 4 days after transection. As early as 6 days after injury, regenerative changes consisting of numerous small axons with an increased axon-Schwann cell ratio were observed; there was little trace of degenerating axons, or their debris. Thus the capacity of unmyelinated nerve fibers for rapid regeneration has been demonstrated. It is anticipated that this delineation of reactions in unmyelinated nerves will contribute to a greater understanding of functional and morphologic abnormalities in disorders of peripheral nerves.     

The numbers of unmyelinated and myelinated axons in normal and regenerated rat saphenous nerves

Counts have been made of the numbers of unmyelinated and myelinated axons in the proximal and distal stumps of regenerated rat saphenous nerves and from equivalent sites in normal nerves. In the proximal part of normal nerves there were averages of 1 045 myelinated axons and 4 160 unmyelinated ones. Regenerated nerves contained the same number of myelinated axons in their proximal stumps but there was a 40% reduction in the unmyelinated axon count. In the distal stumps of these nerves the myelinated axon count had increased by an average of 620; this comes about because some regenerated myelinated axons support more than one process in the distal stump. In contrast, the number of unmyelinated axons was reduced further, from a mean of 2 476 in the proximal stump to one of 2 219.

The sizes of Schwann cell units in the normal and regenerated nerves were also noted. Schwann cell units in the proximal and distal stumps of the regenerated nerves were smaller than those in the normal ones.

These changes associated with unmyelinated axons in regenerated nerves are likely to contribute to the sensory, vasomotor and sudomotor abnormalities that sometimes occur after peripheral nerve injury and regeneration.     

Sciatic nerve regeneration after autologous sural nerve transplantation in the rat

The present study is concerned with the question as to whether the size of a nerve used as a transplant to bridge a gap between the stumps of transected nerves has a bearing on the number of axons and the cytological structure of the regenerate. The paradigm is rat sciatic nerve transection with 8 mm of nerve removed with the stumps placed in a silicone tube and two strands of the smaller sural nerve used as bridging transplants. The comparisons are with previously published results where the transplant, which is the removed piece of sciatic nerve, is exactly matched in size and with no transplant in the same regeneration paradigm. One surprising finding is that the size of the transplant does not seem to determine the size of the regenerated nerve. The cytological structure of the regenerated nerve is related to the size of the transplant, however, in that the proportion of axons that regenerate inside and outside the transplanted perineurial tubes differs in relation to the size of the transplant. In addition, although there is an increase in the number of blood vessels in all of these paradigms, the greatest increase is with the sural nerve transplants. The key finding in the study, however, is the similarity in numbers of regenerated axons in the gap, distal stump and tributary nerves when regeneration after sciatic nerve transplantation is compared with regeneration after sural nerve transplantation. Thus, notwithstanding the cytologic differences of the two types of regenerate, regenerated axon numbers are approximately the same. The conclusion is that the size of the transplant determines neither the size of the regenerate nor the numbers of regenerated axons in this paradigm. On the assumption that regeneration is better when axonal numbers are closer to normal, the non-matched sural nerve transplant is approximately equal to the matched sciatic nerve transplant and both are superior to the regeneration that takes place in the absence of a transplant in this paradigm.     

Oxidized galectin-1 stimulates the migration of Schwann cells from both proximal and distal stumps of transected nerves and promotes axonal regeneration after peripheral nerve injury

Oxidized galectin-1 has recently been identified as a key factor that plays important roles in initial axonal growth in injured peripheral nerves. The aim of this study was to investigate the effects of oxidized galectin-1 on regeneration of rat spinal nerves using acellular autografts (containing no viable cells) and allografts (containing no cell membranes) with special attention to the relationship between axonal regeneration and Schwann cell migration. Immunohistochemically, endogenous galectin-1 was expressed in dorsal root ganglion (DRG) neurons, spinal cord motoneurons, and axons and Schwann cells in normal sciatic nerves. Administration of oxidized recombinant human galectin-1 (rh-gal-lox, 5 ng/ml) in autograft model promoted axonal regeneration from motoneurons as well as from DRG neurons; this was confirmed by a fluorogold tracer study (p < 0.05). Anti-rh-gal-1 antibody (30 microg/ml) strongly inhibited axonal regrowth (p < 0.05). Pretreatment of allografts with rh-gal-lox stimulated the migration of Schwann cells not only from proximal stumps but also from distal stumps into the grafts, resulting in accelerated axonal regeneration (p < 0.05). Moreover, Schwann cell migration preceded the axonal growth in the presence of exogenous rh-gal-lox in the grafts. These results strongly suggest that local administration of exogenous rh-gal-lox promotes the migration of Schwann cells followed by axonal regeneration from both motor and sensory neurons, resulting in acceleration of neuronal repair. This technique may also be of value in the repair of human nerves.     

Specificity in regenerative outgrowth and target reinnervation by mammalian peripheral axons

There are indications that specific factors are present in the distal stump of transected nerves which preferentially attract axons of the corresponding proximal stump into the distal nerve stumps. However, the impact of these factors is unclear, since there is abundant evidence that numerous regenerating motor and sensory axons are topographically misdirected after nerve transection and repair. Topographic reinnervation is improved after fascicular repair of fasciculated nerves, and quite precise after nerve crush. The latter may not be true, however, for non-myelinated axons, which show a high degree of aberrant growth even after crush. In contrast, regenerative outgrowth appears to be topographically specific after neonatal nerve transection. Reinnervation of muscle fibers appears to be unspecific in adult mammals, but specific after neonatal injury under certain circumstances. Some preference for reinnervation of the appropriate sensory receptors seems to exist although this preference does not preclude reinnervation of receptors by 'foreign' sensory fibers. In conclusion, incorrect topographic and target reinnervation commonly occurs after peripheral regeneration in adult mammals, and most certainly explains some of the functional disturbances after peripheral nerve lesions. Topographic regeneration appears to be better after nerve injury in developing mammals indicating that mechanisms from the developmental period may persist and aid in accurate regenerative outgrowth.     

Bone marrow stromal cells and resorbable collagen guidance tubes enhance sciatic nerve regeneration in mice

We evaluated peripheral nerve regeneration using a tubular nerve guide of resorbable collagen filled with either bone marrow-derived cells (BMDCs) in Dulbecco's cell culture medium (DMEM) or with DMEM alone (control). The control group received just the culture medium (vehicle). The left sciatic nerves of ten isogenic mice were transected and the tubular nerve guides were sutured to the end of the proximal and distal nerve stumps. Motor function was tested at 2, 4 and 6 weeks after surgery using the walking track test. The pawprints were analyzed and the print lengths (PL) were measured to evaluate functional recovery. After 6 weeks, mice were anesthetized, perfused transcardially with fixative containing aldehydes, and the sciatic nerves and tubes were dissected and processed for scanning and transmission electron microscopy. Scanning electron microscopy of the collagen tube revealed that the tube wall became progressively thinner after surgery, proving that the tube can be resorbed in vivo. Quantitative analysis of the regenerating nerves showed that the number of myelinated fibers and the myelin area were significantly increased in the experimental group. Also, motor function recovery was faster in animals that received the cell grafts. These results indicate that the collagen tube filled with BMDCs provided an adequate and favorable environment for the growth and myelination of regenerating axons compared to the collagen tube alone.