- Email: [email protected]
Effects of motor versus sensory nerve grafts on peripheral nerve regeneration
Experimental Neurology 190 (2004) 347 – 355 www.elsevier.com/locate/yexnr
Effect of motor versus sensory nerve grafts on peripheral nerve regeneration Chris M. Nicholsa, Michael J. Brennerb, Ida K. Foxa, Thomas H. Tunga, Daniel A. Huntera, Susan R. Rickmana, Susan E. Mackinnona,* a
Division of Plastic and Reconstructive Surgery, Department of Surgery, Washington University School of Medicine, Saint Louis, MO 63110, United States b Department of Otolaryngology-Head and Neck Surgery, Washington University School of Medicine, Saint Louis, MO 63110, United States Received 21 May 2004; revised 5 August 2004; accepted 11 August 2004 Available online 30 September 2004
Abstract Autologous nerve grafting is the current standard of care for nerve injuries resulting in a nerve gap. This treatment requires the use of sensory grafts to reconstruct motor defects, but the consequences of mismatches between graft and native nerve are unknown. Motor pathways have been shown to preferentially support motoneuron regeneration. Functional outcome of motor nerve reconstruction depends on the magnitude, rate, and precision of end organ reinnervation. This study examined the role of pathway type on regeneration across a mixed nerve defect. Thirty-six Lewis rats underwent tibial nerve transection and received isogeneic motor, sensory or mixed nerve grafts. Histomorphometry of the regenerating nerves at 3 weeks demonstrated robust nerve regeneration through both motor and mixed nerve grafts. In contrast, poor nerve regeneration was seen through sensory nerve grafts, with significantly decreased nerve fiber count, percent nerve, and nerve density when compared with mixed and motor groups ( P b 0.05). These data suggest that use of motor or mixed nerve grafts, rather than sensory nerve grafts, will optimize regeneration across mixed nerve defects. D 2004 Elsevier Inc. All rights reserved. Keywords: Nerve regeneration; Tibial nerve; Motor neurons; Sensory neurons; Preferential motor reinnervation
Introduction Peripheral nerve injuries account for 2.8% of traumatic injuries, 65% of which occur in the upper extremity, most often involving the radial nerve. These injuries are most common in males, aged 18–35, and are often the result of motor vehicle or industrial accidents (Kreiger et al., 1981; Nolte et al., 1989). When a peripheral nerve injury is associated with a loss of tissue, a nerve gap will result. The current standard of care for nerve gap injuries is reconstruction with autologous sensory nerve grafts.
* Corresponding author. Division of Plastic and Reconstructive Surgery, Department of Surgery, Washington University School of Medicine, 660 South Euclid Avenue, Campus Box 8238, Saint Louis, MO 63110. Fax: +1 314 362 4536. E-mail address: [email protected] (S.E. Mackinnon). 0014-4886/$ - see front matter D 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.expneurol.2004.08.003
Other techniques under investigation include synthetic nerve conduits (Archibald et al., 1991; Doolabh et al., 1996) and nerve allografts (Mackinnon et al., 1984, 1985, 2001). However, despite significant advances in microsurgical technique, full functional recovery after peripheral nerve injury is seldom achieved (Mackinnon and Dellon, 1988; Mackinnon et al., 2001). Several prior investigations have demonstrated that after peripheral nerve injury, motor axons will preferentially reinnervate motor rather than sensory pathways (Brushart, 1988, 1993; Brushart et al., 1998; Le et al., 2001; Madison et al., 1999). Brushart et al. (1998) showed that this preferential motor reinnervation (PMR) was due to a selective pruning process (Brushart, 1988, 1993). Motor neurites will initially progress at random down any Schwann cell basal lamina (SCBL) tubes available at the transection site. However, as they grow, axon sprouts that interact with motor pathways (Brushart,
348
C.M. Nichols et al. / Experimental Neurology 190 (2004) 347–355
1993; Brushart et al., 1998) or end organs (Brushart, 1988; Madison et al., 1996) are maintained in favor of those erroneously growing within sensory channels. Motor axons that are correctly matched with a motor pathway receive positive trop(h)ic support and consume a greater portion of the parent neuron’s resources; sister neurites that are in mismatched pathways are thereby effectively bstarvedQ (Brushart, 1993). This is congruent with the bsibling neurite modelQ described by Devor and Schneider (1975). PMR is thought to be due to collaboration between the regenerating axons and the appropriate pathway type. For example, in motor nerves, the L2/HNK-1 carbohydrate epitope, a growth-promoting molecule, is produced by Schwann cells in significant quantities only when motor axons interact with motor pathways (Martini and Schachner, 1986, 1988; Martini et al., 1992, 1994). This nerve to pathway synergy is not unique to motor axons; sequential double labeling studies have shown that sensory as well as motor neurons display preferential regeneration (Madison et al., 1996). This process of preferential reinnervation is beneficial to the regenerating nerve in that it increases the specificity of axonal regrowth to the correct target organ and minimizes the amount of neuronal resources wasted on erroneous, non-functional connections. It also limits dysfunctional, mismatched neurite projections which occupy SCBL tubes and block the regeneration of the appropriate neurites to the proper end organ connections (Brushart, 1993). Although primarily studied in rodents (Brushart, 1988, 1993; Robinson and Madison, 2003), PMR has been shown to occur in primates as well, which suggests that this trait is conserved in humans (Madison et al., 1999). Preferential reinnervation is described in the setting where both motor and sensory SCBL tubes are available distally for regrowth. However, in most clinical nerve graft reconstructions, autologous sensory nerve grafts are often used to repair motor nerve deficits. In these cases, the regenerating motor neurites are forced to grow in sensory SCBL tubes by default. They must grow without support from the graft pathway until they can be drescuedT by contact with the native nerve tissue distal to the graft (Martini et al., 1994). This precludes any benefit that would be gained from preferential trop(h)ic interactions between motor neurites and motor pathways. It is unclear whether the improper pathway may have a deleterious effect on the rate or extent of axonal regrowth. In a motor injury both the time to neurotization and the number of nerve/end plate connections made will effect the ultimate functional result; therefore, pathway contributions have a direct impact on quality of motor recovery. The goal of this study is to investigate the effect of motor versus sensory nerve grafts on nerve regeneration. The regeneration of the rat tibial nerve through motor, sensory, and mixed nerve grafts was examined and compared via functional and histomorphometric analysis.
Experimental procedures Experimental animals Adult male Lewis rats (Harlan Sprague–Dawley, Indianapolis, IN), each weighing 300 to 400 g, were used in this study. All surgical procedures, experimental manipulations, and peri-operative care measures were carried out in strict accordance with National Institutes of Health guidelines and were approved by the Washington University institutional Animal Studies Committee. All animals were housed in a central animal facility, given a rodent diet (PicoLab Rodent Diet 20 #5053, PMI Nutrition International) and water ad libitum. Animals were promptly returned to the animal facility following surgical procedures and monitored for weight loss, infection, or other impairment. Experimental design Thirty-six animals were randomly assigned to three graft recipient groups: mixed, sensory, or motor. Additional 12 animals were used as nerve donors. Group I animals received 5 mm reversed tibial nerve autografts, group II animals received 5 mm sensory (femoral cutaneous branch) isografts, and group III animals received 5 mm motor (femoral motor branch to quadriceps) isografts. Size disparity between sensory and motor nerves was compensated for by constructing a three cable sensory graft and a two cable motor graft. This provided grafts of comparable diameter and area. Functional analysis was performed by serial walking track analysis at 0, 2, and 3 weeks post graft. Animals were sacrificed 3 weeks for histomorphometric analysis of nerve tissue and harvest of gastrocnemius muscles for wet muscle mass analysis to compare degree of muscle wasting. Surgeries For all surgical procedures, general anesthesia was achieved by subcutaneous injection of 75 mg/kg ketamine hydrochloride (Fort Dodge Animal Health, Fort Dodge, IA) and 0.5 mg/kg medetomidine hydrochloride (Orion Corporation, Espoo, Finland). Animals were shaved and then prepped with betadine solution. Surgeries were carried out under aseptic conditions. Donor nerves were harvested through a longitudinal groin incision to exposed the femoral neurovascular bundle. Under 16 magnification, using a Wild M651 operating microscope (Leica Microsystems, Deerfield, IL), the saphenous nerve was traced back to the proximal femoral trunk and where the motor branch of the femoral nerve was identified branching to the quadriceps muscles. Both branches were neurolyzed proximally and harvested by sharp transection with microscissors distally and proximally with each nerve under gentle traction. Under 25–40 magnification the nerves were measured and cut into 5
C.M. Nichols et al. / Experimental Neurology 190 (2004) 347–355
349
optimally. All incisions were closed in layers with 4–0 vicryl muscular sutures and 4–0 nylon skin sutures. Donor animals were sacrificed with intracardiac injection of Euthasol (Delmarva Laboratories, Des Moines, IA) while still under general anesthesia, immediately after graft harvest. Experimental animals were reversed with antipamezole HCl (Pfizer Animal Health, Exton, PA), allowed to recover in a heated area and then returned to the animal housing facility. At the 3-week endpoint, the animals were re anesthetized, and nerve harvests were performed by reopening the prior muscle splitting incision. The nerve graft and a portion of native nerve both proximally and distally were harvested. The specimen was marked with a proximal marking stitch and placed in 3% Glutaraldehyde at 48C to before histomorphometric analysis. Muscle mass Gastrocnemius muscle harvests were performed through a longitudinal incision on the posterior aspect of the leg. The overlying skin was dissected free of the gastrocnemius muscle. Using 6–10 magnification, the gastrocnemius was isolated and freed along the muscular tissue planes from insertion to origin. It was removed en block and immediately weighed. The contralateral, uninjured muscle was also harvested to control for variation in size between individual rats. Walking tracks Fig. 1. This photomicrograph demonstrates the differences in size between the saphenous and quadriceps nerve. The femoral nerve sensory (saphenous) branches (a) have an average an area of 81,106 Am2 (this yields a total 3 cable graft area of 243,319 Am2), while the femoral motor (quadriceps) branches (b) have an average area of 140,229 Am2 (this yields a total 2 cable graft area of 280,457 Am2). Thus, the use of three cables of sensory nerve and two cables of motor nerve produces grafts of most equivalent area. Scale bar = 5 Am.
mm cables. To facilitate handling, cables were attached together at each end with one epineurial 11–0 nylon rosette stitch. Due to the differences in nerve diameters and crosssectional area (Brushart and Seiler, 1987), three cables were used in the sensory grafts, while two cables were used in the motor grafts (Fig. 1). In recipients, the right tibial nerve was exposed through a gluteal muscle splitting incision. The tibial nerve was divided sharply 1 cm distal to the sciatic trifurcation, and the 5-mm motor or sensory cable grafts were interposed using an epineurial repair at 25–40 magnification with 11–0 nylon microsuture. One to two sutures were placed per cable and nerve fascicles were optimally aligned at each end. In group I, animals underwent identical preparation and exposure of the tibial nerve. A 5-mm segment of tibial nerve in the same location was then excised, reversed, and grafted back into the tibial nerve. Four 10–0 nylon epineurial sutures were placed for each anastomosis and fascicles were again aligned
Walking tracks were conducted as previously described (Bain et al., 1989; Brown et al., 1989). Briefly, the walking track consists of an 8.2 42 cm track darkened at one end with a length of exposed X-ray film placed on the bottom of the track. The hind feet of the rat were coated with X-ray film developer and then the rat was allowed to walk down the track. The hind footprints appeared as the developer reacted with the film. Measurements of footprints from walking tracks were used to calculate a tibial function index as described previously (Bain et al., 1989; de Medinaceli et al., 1982). Histomorphometry Tibial nerves were harvested and fixed in a cold, buffered 3% glutaraldehyde solution, post-fixed with 1% osmium tetroxide, dehydrated in ethanol, and then embedded in Araldite 502 (Polysciences Inc., Warrington, PA). Onemicrometer-thick cross-sections were obtained from the suture line and segments 3 to 5 mm distally, using an LKB III Ultramicrotome (LKB-Produkter A.B., Broma, Sweden). Light microscopy was used to evaluate toluidine blue-stained cross-sections for the quality and quantity of regenerated nerve fibers, preservation of nerve architecture, the degree of myelination, and the presence of Wallerian degeneration. For
350
C.M. Nichols et al. / Experimental Neurology 190 (2004) 347–355
Table 1 Results of histomorphometric analysis of nerves distal to the grafted segments Mixed Fiber number Nerve density Percent nerve Fiber width Percent debris
1193.9 2171.7 1.59 2.57 11.79
Sensory F F F F F
977.2 1769.8 1.44 0.24 5.17
405.9 767.8 0.536 2.37 12.02
F F F F F
Motor 454.1 747.5 0.563 0.43 3.69
1074.1 2359.6 1.87 2.67 11.74
F F F F F
657.1 2131.2 1.90 0.17 3.88
Nerve density is given in fibers/mm2, fiber width is given in Am, all other values are reported as counts or percentages.
electron microscopy, ultrathin sections were cut in an LKB III ultramicrotome (LKB-Produkter A.B.), stained with uranyl acetate-lead acetate and examined on a Zeiss 902 electron microscope (Zeiss Instruments, Chicago IL).
Microscopic images were examined using a digital image-analysis system linked to morphometric software (Leco Instruments, St. Joseph, MI). This system displayed the microscope image on a video monitor calibrated to 0.125 mm/pixel. Total fascicular area and total fiber number were measured through computer analysis of digitized information based on gray and white scales. At 1000 magnification, six randomly selected fields per nerve, or a minimum of 500 myelinated fibers, were measured to determine axon width, fiber diameter, and myelin width. These primary measurements were then used to calculate percentage of neural tissue (100 neural area/intrafascicular area), percentage of neural debris (100 neural debris/intrafascicular area), total number of myelinated fibers, and nerve fiber density (fibers/mm2).
Fig. 2. Histomorphometric comparisons of nerve regeneration across motor, sensory, and mixed nerve graft groups. Comparable nerve regeneration is noted in motor and mixed nerve graft groups. The sensory nerve graft group exhibited significantly decreased nerve fiber count, percent nerve, and nerve density when compared with mixed and motor groups ( P b 0.05). Asterisks denote statistically significant differences.
C.M. Nichols et al. / Experimental Neurology 190 (2004) 347–355
Statistical analysis All data are presented as mean F standard deviation. Statistical analyses were performed using Statistica (version 5.1; Stat Soft Inc., Tulsa, Okla.). Overall analysis of differences between the means for each group was performed using an ANOVA. For histomorphometric data, a post-hoc Newman-Keuls test was then used to compare significant differences between groups. Statistical significance was set at p b 0.05.
351
ance of the nerve segments distal to the grafts reveals clearly poorer regeneration in the sensory group (Fig. 4). Electron microscopy sections from the motor and sensory nerve grafts (groups 2 and 3) are shown in Fig. 5. This analysis was performed after the light histomorphometry to determine whether differences in motor versus sensory grafts could were accounted for by unmyelinated axons not discernible by light microscopy. Electron microscopy showed similar prevalence of unmyelinated fibers in the two groups. Walking track and muscle mass analysis
Results Histology and histomorphometry Quantitative analysis of the nerve tissue harvested from each group revealed significant differences in multiple histomorphometric parameters. (Results are summarized in Table 1 and Fig. 2). The most striking differences were seen in the total number of fibers, percent nerve and nerve density analyses, which showed much poorer regeneration in the sensory graft group. A closer analysis of overall fiber width distribution correlated with percent nerve shows fewer large fibers regenerating in the sensory nerve graft group when compared with the motor and mixed groups (Fig. 3). Qualitative examination of distal nerve sections shows good early regeneration in the motor and mixed nerve graft groups, with high numbers of immature regenerating fibers. This is in contrast to the sensory nerve graft group, which shows much less robust nerve regeneration with many fewer regenerating fibers evident. Overall, the histological appear-
Walking track analysis at both the 2- and 3-week time points showed paw-print lengthening consistent with tibial nerve injury. No significant recovery of print length factor was noted in any of the groups by the 3-week endpoint. Analysis of gastrocnemius muscle mass revealed a dramatic drop in mass of the denervated gastrocnemius in all groups. No inter-group differences in muscle mass ratio were detected.
Discussion A crucial factor in optimizing muscular function after denervating injury is minimizing the time required for a regenerating nerve to reach a target muscle. Muscle undergoes a progressive decrease in recovery capability proportional to the duration of denervation (Finkelstein et al., 1993; Gordon and Stein, 1982). After prolonged denervation, functional recovery of muscle may be precluded entirely
Fig. 3. This histogram shows percent of each nerve stratified for specific fiber widths: 1–2, 2–3, 3–4, 5–6, and 6–7 Am. Stratified comparison shows that the motor and mixed graft groups are composed of a greater percentage of fibers of larger diameter when compared with the sensory group. This analysis is based on fiber populations in the regenerating nerve distal to the graft. Thus, after crossing the grafts, the fiber distribution is skewed toward smaller fibers in the sensory graft groups in comparison to the motor and mixed groups.
352
C.M. Nichols et al. / Experimental Neurology 190 (2004) 347–355
Seiler, 1987; Brushart et al., 1983). These effects may involve selective upregulation of specific factors within motor and sensory pathways (Brushart et al., 1998; Martini and Schachner, 1986, 1988; Martini et al., 1994). This study demonstrates that regeneration of mixed nerves through pure sensory nerve graft pathways is inferior to that occurring through motor or mixed nerve pathways. In this experiment, the use of two motor cable grafts to match the 3 sensory cable grafts warrants special consideration. This approach allowed for significantly closer
Fig. 4. Photomicrographs of representative histologic cross-sections of nerves 3 mm distal to the nerve grafts. Robust nerve regeneration is noted in both the mixed (a) and motor (b) nerve groups, with multiple regenerating nerve fibers present. In comparison, nerve regeneration distal to the sensory (c) nerve grafts is relatively poor. Scale bar = 5 Am.
(Gutmann, 1948). Regenerating neurons interact closely with the pathways through which they regenerate. This interaction creates a microenvironment that nurtures regenerating nerve sprouts and increases the specificity and accuracy with which they return to end organ targets (Brushart, 1988; Brushart and
Fig. 5. Electron Micrograph (Uranyl Acetate–Lead Citrate Stain, 11,520 magnification): (a) Group II (sensory nerve graft). Myelinated and unmyelinated fibers have regenerated into the distal nerve. Wallerian degeneration has resolved and neural architecture has been restored. (b) Group III (motor nerve graft). The prevalence of unmyelinated fibers and overall ultrastructural appearance is similar to that noted in (a).
C.M. Nichols et al. / Experimental Neurology 190 (2004) 347–355
matching of cross-sectional area than in prior comparisons of motor/sensory grafting (Ghalib et al., 2001) and also allowed for comparison against a mixed (tibial nerve) graft. But, a roughly 13% difference in cross-sectional area persisted favoring the motor group over the sensory group. Furthermore, it has been shown in a mouse model that surgical adaptions, such as use of fibrin sealant versus suture, can influence preferential motor neuronal regeneration (Robinson and Madison, 2003). Fortunately, several factors mitigate concerns regarding a systematic bias in favor of the motor neuron graft group. First, the histomorophometry data on percent neural tissue and nerve density are normalized values that account for differences in cross-sectional area. Similarly, mean fiber width is an average among all measured fibers. Second, the disparity in cross-sectional area between motor and sensory groups was small in magnitude relative to differences in total fiber count in favor of the motor group. Third, despite the smaller caliber of sensory grafts, each sensory graft contains a greater number of Schwann cell basal lamina tubes than each motor graft. Therefore, the sensory graft group had a potential advantage in terms of Schwann cell tubes available for regeneration into grafts. Fourth, the thinner caliber of the three sensory graft cables corresponds to greater surface area for diffusion of metabolites during the period of neovascularization. Last, concerns regarding surgical adaption in the mouse, which is a particularly technically challenging model, are likely to be less applicable to rats, where cable grafting was straightforward and did not prove problematic from a technical standpoint. Standard deviations in quantitative analyses were typical for histomorphometry work on peripheral nerve regeneration. The timing of this experiment was designed based on extensive prior experience with rat peripheral nerve regeneration (Mackinnon et al., 1991). Although the 3week endpoint allowed optimization for the detection of histomorphometric differences, this came at the expense of functional analyses. At the 3-week time point the regenerating nerve front has just begun to cross the distal end of the graft. Studying regeneration at 3 weeks allows one to discern differences in the rate and magnitude of regeneration between groups. Due to the limitations of the small animal model, it is difficult to simultaneously demonstrate histomorphometric and functional differences. As a general rule, functional recovery lags behind histomorphometric evidence of regeneration (Jensen et al., 2004; Jost et al., 2000; Sunderland, in press). Failure to demonstrate differences on walking track or wet muscle mass analyses was not particularly surprising, given that the regenerating nerves had probably not reached the muscle at this early time point. A future goal of research in this area will be assessment of functional recovery. It is possible that at later time points, when functional outcomes are more readily demonstrable, histomorphometric differences may be less apparent.
353
As the preceding discussion suggests, timing is crucial in small animal research. The superior neuroregenerative capacity inherent in the rodent peripheral nervous system may cause negative controls to become positive at later time points. Because of this concern, caution must be taken in applying findings in small animal models to clinically significant injuries. Nonetheless, these studies provide the groundwork for large animal studies and eventual clinical application. Differences in nerve regeneration that may be detected only transiently in a small animal model (Calvert et al., 2001; Jensen et al., 2004; Myckatyn and Mackinnon, 2004) may be long-lasting in larger animals (Brenner et al., in press(a,b); Jensen et al., submitted for publication). Furthermore, even in the absence of a long-lasting difference in nerve regeneration, an acceleration of regeneration is clinically important due to the significant functional morbidity associated with prolonged muscle denervation. The comparatively poor regeneration in the sensory group may reflect the predilection of motor neurons to regenerate through motor pathways (PMR). Motor fibers are generally larger than sensory fibers, but increased fiber width may also reflect greater maturity nerve fibers (Evans et al., 1994). In our study, the presence of increased fiber widths in motor grafts versus sensory grafts may signify greater maturity, predominance of motor fibers, or both. Future work involving formal retrograde labeling studies is needed to distinguish between these possibilities by determining the modality of fibers regeneration though grafts. After evaluating light microscopy findings, we considered the possibility that differences in nerve regeneration between the motor versus sensory groups arose largely from differences in myelination. Specifically, were a disproportionate number of unmyelinated (and hence uncounted) fibers present in the sensory graft group? To address this question, electron microscopy was performed on nerve sections. Upon review of these sections, it was readily apparent that the difference between motor and sensory graft groups could not be explained by myelination patterns alone. The number of unmyelinated fibers found in nerves was generally proportional to the number of myelinated fibers in both groups. These findings served to confirm the validity of light microscopy-based histomorphometry data. The findings of this study are in general agreement with related work in this area. For example, one prior study has shown that motor neurite diameters and myelination increase if they are grown in motor rather than sensory pathways (Ghalib et al., 2001). However, the effects of neuron-pathway synergy on rate and magnitude of regeneration have not been previously investigated. In the current study, interactions between nerve and pathway were examined by comparing regeneration through motor versus sensory or mixed pathways. Histomorphometric analysis revealed that sensory pathways showed inferior regeneration in comparison to both the motor and mixed pathways. These results were borne
354
C.M. Nichols et al. / Experimental Neurology 190 (2004) 347–355
out in analysis of fiber density, percent nerve and total nerve fiber counts. Analysis of overall fiber distribution also shows that more large fibers were present distally in the motor and mixed graft groups (Fig. 3). This is consistent with a greater rate of regeneration through these grafts. A possible explanation for this phenomenon is the absence of motor elements in the pure sensory pathways. Both motor and sensory modalities are present in the mixed and the motor nerve grafts; this is due to the fact that the motor grafts contain sensory axons in the form of motor afferent fibers (serving muscle stretch receptors, Golgi tendon organs, etc.). Thus, given that motor associated pathways produce motor-specific factors (Ghalib et al., 2001; Martini and Schachner, 1986, 1988; Martini et al., 1992, 1994; Nieke and Schachner, 1985), it is likely that these have a trophic influence on regenerating motor neurites. This growth advantage is denied to motor neurites that are forced to grow in a pure sensory milieu. The key element in this interaction is the Schwann cell. Schwann cells are the myelinating glial cells of peripheral nerves. Following peripheral nerve injury, they de-differentiate to secrete survival and growth-promoting factors for the regenerating neurites (Bunge, 1994; Feneley et al., 1991). The specificity of motor Schwann cells to produce neurotrophins in response to motor neurites has been demonstrated in vitro and in vivo (Martini et al., 1992, 1994). It is likely that the contribution of these cells is what confers the growth advantage to the mixed modality pathways. Alternatively, given that motor nerves tend to have more myelination, it is possible that the motor nerve grafts represent a more Schwann cell-rich milieu for regeneration. These finding of the present study have direct implications to clinical nerve grafting. The existing gold standard is to utilize autogenous sensory nerve grafts for reconstruction of nerve gap injuries. Results of the present study suggest that clinical outcomes might be improved by finding alternatives to sensory grafts in the reconstruction of mixed nerves. However, there are relatively few expendable motor nerves suitable for graft material in the human body. Perhaps, a more feasible alternative lies in matching strategies that make use of exogenous graft material. This could be done in the form of allograft or biosynthetic composite graft material. In these cases, application of the results of this study to conduit or allograft technology could increase the efficacy of these modes of reconstruction. Previous studies have shown that motor neurons have inferior regeneration through synthetic nerve grafts (Madorsky et al., 1998). The use of Schwann cells and Schwann cell neurotrophins to augment nerve allografts and conduits is currently under investigation (Guenard et al., 1992; Lee et al., 2003; Mosahebi et al., 2002). With this new data showing that modality matching improves growth of peripheral nerves, research efforts can now be directed at
not only augmenting graft scaffolds with Schwann cells, but with cells derived from nerves of the appropriately matched phenotype, thereby equally enhancing the growth of all regenerating neurites.
Acknowledgment The authors would like to thank the American Society for Surgery of the Hand for supporting this research.
References Archibald, S.J., Krarup, C., Shefner, J., Li, S.T., Madison, R.D., 1991. A collagen-based nerve guide conduit for peripheral nerve repair: an electrophysiological study of nerve regeneration in rodents and nonhuman primates. J. Comp. Neurol. 306, 685 – 696. Bain, J.R., Mackinnon, S.E., Hunter, D.A., 1989. Functional evaluation of complete sciatic, peroneal, and posterior tibial nerve lesions in the rat. Plast. Reconstr. Surg. 83, 129 – 138. Brenner, M.J., Jensen, J.N., Lowe III, J.B., Myckatyn, T.M., Fox, I.K., Hunter, D.A., Mohanakumar, T., Mackinnon, S.E., in press a. Anti-CD40 ligand antibody permits regeneration through peripheral nerve allografts in a non-human primate model. Plast. Reconstr. Surg. Brenner, M.J., Lowe, I., J.B., Mackinnon, S.E., Darcy, M., Duncan, J., Fox, I.K., Jaramillo, A., Hunter, D.A., Mohanakumar, T., in press b. Effects of Schwann cells and donor antigen on regeneration through long nerve allografts. Microsurgery. Brown, C.J., Mackinnon, S.E., Evans, P.J., Bain, J.R., Makino, A.P., Hunter, D.A., Hare, G.M., 1989. Self-evaluation of walking-track measurement using a sciatic function index. Microsurgery 10, 226 – 235. Brushart, T.M., 1988. Preferential reinnervation of motor nerves by regenerating motor axons. J. Neurosci. 8, 1026 – 1031. Brushart, T.M., 1993. Motor axons preferentially reinnervate motor pathways. J. Neurosci. 13, 2730 – 2738. Brushart, T.M., Seiler, W.A., 1987. Selective reinnervation of distal motor stumps by peripheral motor axons. Exp. Neurol. 97, 289 – 300. Brushart, T.M., Tarlov, E.C., Mesulam, M.M., 1983. Specificity of muscle reinnervation after epineurial and individual fascicular suture of the rat sciatic nerve. J. Hand Surg. [Am.] 8, 248 – 253. Brushart, T.M., Gerber, J., Kessens, P., Chen, Y.G., Royall, R.M., 1998. Contributions of pathway and neuron to preferential motor reinnervation. J. Neurosci. 18, 8674 – 8681. Bunge, R.P., 1994. The role of the Schwann cell in trophic support and regeneration. J. Neurol. 242, S19 – S21. Calvert, G.T., Doolabh, V.B., Grand, A.G., Hunter, D.A., Mackinnon, S.E., 2001. Rat-strain differences in recovery following peripheral-nerve allotransplantation. J. Reconstr. Microsurg. 17, 185 – 191. de Medinaceli, L., Freed, W.J., Wyatt, R.J., 1982. An index of the functional condition of rat sciatic nerve based on measurements made from walking tracks. Exp. Neurol. 77, 634 – 643. Devor, M., Schneider, G., 1975. Neuroanatomical plasticity: the principle of conservation of total axonal arborization. In: Vital-Durand, F., Jeannerod, M. (Eds.), Asp. Neural Plast., vol. 43. INSERM Colloquia, Paris, pp. 191 – 200. Doolabh, V.B., Hertl, M.C., Mackinnon, S.E., 1996. The role of conduits in nerve repair: a review. Rev. Neurosci. 7, 47 – 84. Evans, P.J., Midha, R., Mackinnon, S.E., 1994. The peripheral nerve allograft: a comprehensive review of regeneration and neuroimmunology. Prog. Neurobiol. 43, 187 – 233. Feneley, M.R., Fawcett, J.W., Keynes, R.J., 1991. The role of Schwann cells in the regeneration of peripheral nerve axons through muscle basal lamina grafts. Exp. Neurol. 114, 275 – 285.
C.M. Nichols et al. / Experimental Neurology 190 (2004) 347–355 Finkelstein, D.I., Dooley, P.C., Luff, A.R., 1993. Recovery of muscle after different periods of denervation and treatments. Muscle Nerve 16, 769 – 777. Ghalib, N., Houst’ava, L., Haninec, P., Dubovy, P., 2001. Morphometric analysis of early regeneration of motor axons through motor and cutaneous nerve grafts. Ann. Anat. 183, 363 – 368. Gordon, T., Stein, R.B., 1982. Time course and extent of recovery in reinnervated motor units of cat triceps surae muscles. J. Physiol. 323, 307 – 323. Guenard, V., Kleitman, N., Morrissey, T.K., Bunge, R.P., Aebischer, P., 1992. Syngeneic Schwann cells derived from adult nerves seeded in semipermeable guidance channels enhance peripheral nerve regeneration. J. Neurosci. 12, 3310 – 3320. Gutmann, E., 1948. Effect of delay of innervation on recovery of muscle after nerve lesions. J. Neurophysiol. 11, 279 – 294. Jensen, J.N., Tung, T.H.H., Mackinnon, S.E., Brenner, M.J., Hunter, D.A., 2004. Use of anti-CD40 ligand monoclonal antibody as anti-rejection therapy in a murine peripheral nerve allograft model. Microsurgery 24, 309–315. Jensen, J.N., Brenner, M.J., Tung, T.H.H., Hunter, D.A., Mackinnon, S.E., submitted for publication. FK506 accelerates peripheral nerve regene-ration through long grafts in inbred swine. Ann. Plast. Surg. Jost S.C., Doolabh, D.V., Mackinnon, S.E., Lee, M., Hunter, D., 2000. Acceleration of peripheral nerve regeneration following FK506 administration. Restor. Neurol. Neurosci. 17, 39 – 44. Kreiger, N., Kelsey, J.L., Harris, C., Pastides, H., 1981. Injuries to the upper extremity: patterns of occurrence. Clin. Plast. Surg. 8, 13 – 19. Le, T.B., Aszmann, O., Chen, Y.G., Royall, R.M., Brushart, T.M., 2001. Effects of pathway and neuronal aging on the specificity of motor axon regeneration. Exp. Neurol. 167, 126 – 132. Lee, A.C., Yu, V.M., Lowe III, J.B., Brenner, M.J., Hunter, D.A., Mackinnon, S.E., Sakiyama-Elbert, S.E., 2003. Controlled release of nerve growth factor enhances sciatic nerve regeneration. Exp. Neurol. 184, 295 – 303. Mackinnon, S.E., Dellon, A. Lee., 1988. Surgery of the Peripheral Nerve. Thieme Medical Publishers, New York. Mackinnon, S.E., Hudson, A.R., Falk, R.E., Kline, D., Hunter, D., 1984. Peripheral nerve allograft: an assessment of regeneration across pretreated nerve allografts. Neurosurgery 15, 690 – 693. Mackinnon, S.E., Hudson, A.R., Falk, R.E., Hunter, D.A., 1985. The nerve allograft response—An experimental model in the rat. Ann. Plast. Surg. 14, 334 – 339. Mackinnon, S.E., Dellon, A.L., O’Brien, J.P., 1991. Changes in nerve fiber numbers distal to a nerve repair in the rat sciatic nerve model. Muscle Nerve 14, 1116 – 1122.
355
Mackinnon, S.E., Doolabh, V.B., Novak, C.B., Trulock, E.P., 2001. Clinical outcome following nerve allograft transplantation. Plast. Reconstr. Surg. 107, 1419 – 1429. Madison, R.D., Archibald, S.J., Brushart, T.M., 1996. Reinnervation accuracy of the rat femoral nerve by motor and sensory neurons. J. Neurosci. 16, 5698 – 5703. Madison, R.D., Archibald, S.J., Lacin, R., Krarup, C., 1999. Factors contributing to preferential motor reinnervation in the primate peripheral nervous system. J. Neurosci. 19, 11007 – 11016. Madorsky, S.J., Swett, J.E., Crumley, R.L., 1998. Motor versus sensory neuron regeneration through collagen tubules. Plast. Reconstr. Surg. 102, 430 – 436. Martini, R., Schachner, M., 1986. Immunoelectron microscopic localization of neural cell adhesion molecules (L1, N-CAM, and MAG) and their shared carbohydrate epitope and myelin basic protein in developing sciatic nerve. J. Cell Biol. 103, 2439 – 2448. Martini, R., Schachner, M., 1988. Immunoelectron microscopic localization of neural cell adhesion molecules (L1, N-CAM, and myelin-associated glycoprotein) in regenerating adult mouse sciatic nerve. J. Cell Biol. 106, 1735 – 1746. Martini, R., Xin, Y., Schmitz, B., Schachner, M., 1992. The L2/HNK-1 carbohydrate epitope is involved in the preferential outgrowth of motor neurons on ventral roots and motor nerves. Eur. J. Neurosci. 4, 628 – 639. Martini, R., Schachner, M., Brushart, T.M., 1994. The L2/HNK-1 carbohydrate is preferentially expressed by previously motor axonassociated Schwann cells in reinnervated peripheral nerves. J. Neurosci. 14, 7180 – 7191. Mosahebi, A., Fuller, P., Wiberg, M., Terenghi, G., 2002. Effect of allogeneic Schwann cell transplantation on peripheral nerve regeneration. Exp. Neurol. 173, 213 – 223. Myckatyn, T.M., Mackinnon, S.E., 2004. A review of research endeavors to optimize peripheral nerve reconstruction. Neurol. Res. 26, 124 – 138. Nieke, J., Schachner, M., 1985. Expression of the neural cell adhesion molecules L1 and N-CAM and their common carbohydrate epitope L2/ HNK-1 during development and after transection of the mouse sciatic nerve. Differentiation 30, 141 – 151. Nolte, C., Schachner, M., Martini, R., 1989. Immunocytochemical localization of the neural cell adhesion molecules L1, N-CAM, and J1 in pacinian corpuscles of the mouse during development, in the adult and during regeneration. J. Neurocytol. 18, 795 – 808. Robinson, G.A., Madison, R.D., 2003. Preferential motor reinnervation in the mouse: comparison of femoral nerve repair using a fibrin sealant or suture. Muscle Nerve 28, 227 – 231. Sunderland, I.R.P., Brenner, M.J., Singham, J., Hunter, D.A., Mackinnon, S.E., in press. Effect of tension on nerve regeneration in rat sciatic nerve transection model. Ann. Plast. Surg.
Effect of motor versus sensory nerve grafts on peripheral nerve regeneration Chris M. Nicholsa, Michael J. Brennerb, Ida K. Foxa, Thomas H. Tunga, Daniel A. Huntera, Susan R. Rickmana, Susan E. Mackinnona,* a
Division of Plastic and Reconstructive Surgery, Department of Surgery, Washington University School of Medicine, Saint Louis, MO 63110, United States b Department of Otolaryngology-Head and Neck Surgery, Washington University School of Medicine, Saint Louis, MO 63110, United States Received 21 May 2004; revised 5 August 2004; accepted 11 August 2004 Available online 30 September 2004
Abstract Autologous nerve grafting is the current standard of care for nerve injuries resulting in a nerve gap. This treatment requires the use of sensory grafts to reconstruct motor defects, but the consequences of mismatches between graft and native nerve are unknown. Motor pathways have been shown to preferentially support motoneuron regeneration. Functional outcome of motor nerve reconstruction depends on the magnitude, rate, and precision of end organ reinnervation. This study examined the role of pathway type on regeneration across a mixed nerve defect. Thirty-six Lewis rats underwent tibial nerve transection and received isogeneic motor, sensory or mixed nerve grafts. Histomorphometry of the regenerating nerves at 3 weeks demonstrated robust nerve regeneration through both motor and mixed nerve grafts. In contrast, poor nerve regeneration was seen through sensory nerve grafts, with significantly decreased nerve fiber count, percent nerve, and nerve density when compared with mixed and motor groups ( P b 0.05). These data suggest that use of motor or mixed nerve grafts, rather than sensory nerve grafts, will optimize regeneration across mixed nerve defects. D 2004 Elsevier Inc. All rights reserved. Keywords: Nerve regeneration; Tibial nerve; Motor neurons; Sensory neurons; Preferential motor reinnervation
Introduction Peripheral nerve injuries account for 2.8% of traumatic injuries, 65% of which occur in the upper extremity, most often involving the radial nerve. These injuries are most common in males, aged 18–35, and are often the result of motor vehicle or industrial accidents (Kreiger et al., 1981; Nolte et al., 1989). When a peripheral nerve injury is associated with a loss of tissue, a nerve gap will result. The current standard of care for nerve gap injuries is reconstruction with autologous sensory nerve grafts.
* Corresponding author. Division of Plastic and Reconstructive Surgery, Department of Surgery, Washington University School of Medicine, 660 South Euclid Avenue, Campus Box 8238, Saint Louis, MO 63110. Fax: +1 314 362 4536. E-mail address: [email protected] (S.E. Mackinnon). 0014-4886/$ - see front matter D 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.expneurol.2004.08.003
Other techniques under investigation include synthetic nerve conduits (Archibald et al., 1991; Doolabh et al., 1996) and nerve allografts (Mackinnon et al., 1984, 1985, 2001). However, despite significant advances in microsurgical technique, full functional recovery after peripheral nerve injury is seldom achieved (Mackinnon and Dellon, 1988; Mackinnon et al., 2001). Several prior investigations have demonstrated that after peripheral nerve injury, motor axons will preferentially reinnervate motor rather than sensory pathways (Brushart, 1988, 1993; Brushart et al., 1998; Le et al., 2001; Madison et al., 1999). Brushart et al. (1998) showed that this preferential motor reinnervation (PMR) was due to a selective pruning process (Brushart, 1988, 1993). Motor neurites will initially progress at random down any Schwann cell basal lamina (SCBL) tubes available at the transection site. However, as they grow, axon sprouts that interact with motor pathways (Brushart,
348
C.M. Nichols et al. / Experimental Neurology 190 (2004) 347–355
1993; Brushart et al., 1998) or end organs (Brushart, 1988; Madison et al., 1996) are maintained in favor of those erroneously growing within sensory channels. Motor axons that are correctly matched with a motor pathway receive positive trop(h)ic support and consume a greater portion of the parent neuron’s resources; sister neurites that are in mismatched pathways are thereby effectively bstarvedQ (Brushart, 1993). This is congruent with the bsibling neurite modelQ described by Devor and Schneider (1975). PMR is thought to be due to collaboration between the regenerating axons and the appropriate pathway type. For example, in motor nerves, the L2/HNK-1 carbohydrate epitope, a growth-promoting molecule, is produced by Schwann cells in significant quantities only when motor axons interact with motor pathways (Martini and Schachner, 1986, 1988; Martini et al., 1992, 1994). This nerve to pathway synergy is not unique to motor axons; sequential double labeling studies have shown that sensory as well as motor neurons display preferential regeneration (Madison et al., 1996). This process of preferential reinnervation is beneficial to the regenerating nerve in that it increases the specificity of axonal regrowth to the correct target organ and minimizes the amount of neuronal resources wasted on erroneous, non-functional connections. It also limits dysfunctional, mismatched neurite projections which occupy SCBL tubes and block the regeneration of the appropriate neurites to the proper end organ connections (Brushart, 1993). Although primarily studied in rodents (Brushart, 1988, 1993; Robinson and Madison, 2003), PMR has been shown to occur in primates as well, which suggests that this trait is conserved in humans (Madison et al., 1999). Preferential reinnervation is described in the setting where both motor and sensory SCBL tubes are available distally for regrowth. However, in most clinical nerve graft reconstructions, autologous sensory nerve grafts are often used to repair motor nerve deficits. In these cases, the regenerating motor neurites are forced to grow in sensory SCBL tubes by default. They must grow without support from the graft pathway until they can be drescuedT by contact with the native nerve tissue distal to the graft (Martini et al., 1994). This precludes any benefit that would be gained from preferential trop(h)ic interactions between motor neurites and motor pathways. It is unclear whether the improper pathway may have a deleterious effect on the rate or extent of axonal regrowth. In a motor injury both the time to neurotization and the number of nerve/end plate connections made will effect the ultimate functional result; therefore, pathway contributions have a direct impact on quality of motor recovery. The goal of this study is to investigate the effect of motor versus sensory nerve grafts on nerve regeneration. The regeneration of the rat tibial nerve through motor, sensory, and mixed nerve grafts was examined and compared via functional and histomorphometric analysis.
Experimental procedures Experimental animals Adult male Lewis rats (Harlan Sprague–Dawley, Indianapolis, IN), each weighing 300 to 400 g, were used in this study. All surgical procedures, experimental manipulations, and peri-operative care measures were carried out in strict accordance with National Institutes of Health guidelines and were approved by the Washington University institutional Animal Studies Committee. All animals were housed in a central animal facility, given a rodent diet (PicoLab Rodent Diet 20 #5053, PMI Nutrition International) and water ad libitum. Animals were promptly returned to the animal facility following surgical procedures and monitored for weight loss, infection, or other impairment. Experimental design Thirty-six animals were randomly assigned to three graft recipient groups: mixed, sensory, or motor. Additional 12 animals were used as nerve donors. Group I animals received 5 mm reversed tibial nerve autografts, group II animals received 5 mm sensory (femoral cutaneous branch) isografts, and group III animals received 5 mm motor (femoral motor branch to quadriceps) isografts. Size disparity between sensory and motor nerves was compensated for by constructing a three cable sensory graft and a two cable motor graft. This provided grafts of comparable diameter and area. Functional analysis was performed by serial walking track analysis at 0, 2, and 3 weeks post graft. Animals were sacrificed 3 weeks for histomorphometric analysis of nerve tissue and harvest of gastrocnemius muscles for wet muscle mass analysis to compare degree of muscle wasting. Surgeries For all surgical procedures, general anesthesia was achieved by subcutaneous injection of 75 mg/kg ketamine hydrochloride (Fort Dodge Animal Health, Fort Dodge, IA) and 0.5 mg/kg medetomidine hydrochloride (Orion Corporation, Espoo, Finland). Animals were shaved and then prepped with betadine solution. Surgeries were carried out under aseptic conditions. Donor nerves were harvested through a longitudinal groin incision to exposed the femoral neurovascular bundle. Under 16 magnification, using a Wild M651 operating microscope (Leica Microsystems, Deerfield, IL), the saphenous nerve was traced back to the proximal femoral trunk and where the motor branch of the femoral nerve was identified branching to the quadriceps muscles. Both branches were neurolyzed proximally and harvested by sharp transection with microscissors distally and proximally with each nerve under gentle traction. Under 25–40 magnification the nerves were measured and cut into 5
C.M. Nichols et al. / Experimental Neurology 190 (2004) 347–355
349
optimally. All incisions were closed in layers with 4–0 vicryl muscular sutures and 4–0 nylon skin sutures. Donor animals were sacrificed with intracardiac injection of Euthasol (Delmarva Laboratories, Des Moines, IA) while still under general anesthesia, immediately after graft harvest. Experimental animals were reversed with antipamezole HCl (Pfizer Animal Health, Exton, PA), allowed to recover in a heated area and then returned to the animal housing facility. At the 3-week endpoint, the animals were re anesthetized, and nerve harvests were performed by reopening the prior muscle splitting incision. The nerve graft and a portion of native nerve both proximally and distally were harvested. The specimen was marked with a proximal marking stitch and placed in 3% Glutaraldehyde at 48C to before histomorphometric analysis. Muscle mass Gastrocnemius muscle harvests were performed through a longitudinal incision on the posterior aspect of the leg. The overlying skin was dissected free of the gastrocnemius muscle. Using 6–10 magnification, the gastrocnemius was isolated and freed along the muscular tissue planes from insertion to origin. It was removed en block and immediately weighed. The contralateral, uninjured muscle was also harvested to control for variation in size between individual rats. Walking tracks Fig. 1. This photomicrograph demonstrates the differences in size between the saphenous and quadriceps nerve. The femoral nerve sensory (saphenous) branches (a) have an average an area of 81,106 Am2 (this yields a total 3 cable graft area of 243,319 Am2), while the femoral motor (quadriceps) branches (b) have an average area of 140,229 Am2 (this yields a total 2 cable graft area of 280,457 Am2). Thus, the use of three cables of sensory nerve and two cables of motor nerve produces grafts of most equivalent area. Scale bar = 5 Am.
mm cables. To facilitate handling, cables were attached together at each end with one epineurial 11–0 nylon rosette stitch. Due to the differences in nerve diameters and crosssectional area (Brushart and Seiler, 1987), three cables were used in the sensory grafts, while two cables were used in the motor grafts (Fig. 1). In recipients, the right tibial nerve was exposed through a gluteal muscle splitting incision. The tibial nerve was divided sharply 1 cm distal to the sciatic trifurcation, and the 5-mm motor or sensory cable grafts were interposed using an epineurial repair at 25–40 magnification with 11–0 nylon microsuture. One to two sutures were placed per cable and nerve fascicles were optimally aligned at each end. In group I, animals underwent identical preparation and exposure of the tibial nerve. A 5-mm segment of tibial nerve in the same location was then excised, reversed, and grafted back into the tibial nerve. Four 10–0 nylon epineurial sutures were placed for each anastomosis and fascicles were again aligned
Walking tracks were conducted as previously described (Bain et al., 1989; Brown et al., 1989). Briefly, the walking track consists of an 8.2 42 cm track darkened at one end with a length of exposed X-ray film placed on the bottom of the track. The hind feet of the rat were coated with X-ray film developer and then the rat was allowed to walk down the track. The hind footprints appeared as the developer reacted with the film. Measurements of footprints from walking tracks were used to calculate a tibial function index as described previously (Bain et al., 1989; de Medinaceli et al., 1982). Histomorphometry Tibial nerves were harvested and fixed in a cold, buffered 3% glutaraldehyde solution, post-fixed with 1% osmium tetroxide, dehydrated in ethanol, and then embedded in Araldite 502 (Polysciences Inc., Warrington, PA). Onemicrometer-thick cross-sections were obtained from the suture line and segments 3 to 5 mm distally, using an LKB III Ultramicrotome (LKB-Produkter A.B., Broma, Sweden). Light microscopy was used to evaluate toluidine blue-stained cross-sections for the quality and quantity of regenerated nerve fibers, preservation of nerve architecture, the degree of myelination, and the presence of Wallerian degeneration. For
350
C.M. Nichols et al. / Experimental Neurology 190 (2004) 347–355
Table 1 Results of histomorphometric analysis of nerves distal to the grafted segments Mixed Fiber number Nerve density Percent nerve Fiber width Percent debris
1193.9 2171.7 1.59 2.57 11.79
Sensory F F F F F
977.2 1769.8 1.44 0.24 5.17
405.9 767.8 0.536 2.37 12.02
F F F F F
Motor 454.1 747.5 0.563 0.43 3.69
1074.1 2359.6 1.87 2.67 11.74
F F F F F
657.1 2131.2 1.90 0.17 3.88
Nerve density is given in fibers/mm2, fiber width is given in Am, all other values are reported as counts or percentages.
electron microscopy, ultrathin sections were cut in an LKB III ultramicrotome (LKB-Produkter A.B.), stained with uranyl acetate-lead acetate and examined on a Zeiss 902 electron microscope (Zeiss Instruments, Chicago IL).
Microscopic images were examined using a digital image-analysis system linked to morphometric software (Leco Instruments, St. Joseph, MI). This system displayed the microscope image on a video monitor calibrated to 0.125 mm/pixel. Total fascicular area and total fiber number were measured through computer analysis of digitized information based on gray and white scales. At 1000 magnification, six randomly selected fields per nerve, or a minimum of 500 myelinated fibers, were measured to determine axon width, fiber diameter, and myelin width. These primary measurements were then used to calculate percentage of neural tissue (100 neural area/intrafascicular area), percentage of neural debris (100 neural debris/intrafascicular area), total number of myelinated fibers, and nerve fiber density (fibers/mm2).
Fig. 2. Histomorphometric comparisons of nerve regeneration across motor, sensory, and mixed nerve graft groups. Comparable nerve regeneration is noted in motor and mixed nerve graft groups. The sensory nerve graft group exhibited significantly decreased nerve fiber count, percent nerve, and nerve density when compared with mixed and motor groups ( P b 0.05). Asterisks denote statistically significant differences.
C.M. Nichols et al. / Experimental Neurology 190 (2004) 347–355
Statistical analysis All data are presented as mean F standard deviation. Statistical analyses were performed using Statistica (version 5.1; Stat Soft Inc., Tulsa, Okla.). Overall analysis of differences between the means for each group was performed using an ANOVA. For histomorphometric data, a post-hoc Newman-Keuls test was then used to compare significant differences between groups. Statistical significance was set at p b 0.05.
351
ance of the nerve segments distal to the grafts reveals clearly poorer regeneration in the sensory group (Fig. 4). Electron microscopy sections from the motor and sensory nerve grafts (groups 2 and 3) are shown in Fig. 5. This analysis was performed after the light histomorphometry to determine whether differences in motor versus sensory grafts could were accounted for by unmyelinated axons not discernible by light microscopy. Electron microscopy showed similar prevalence of unmyelinated fibers in the two groups. Walking track and muscle mass analysis
Results Histology and histomorphometry Quantitative analysis of the nerve tissue harvested from each group revealed significant differences in multiple histomorphometric parameters. (Results are summarized in Table 1 and Fig. 2). The most striking differences were seen in the total number of fibers, percent nerve and nerve density analyses, which showed much poorer regeneration in the sensory graft group. A closer analysis of overall fiber width distribution correlated with percent nerve shows fewer large fibers regenerating in the sensory nerve graft group when compared with the motor and mixed groups (Fig. 3). Qualitative examination of distal nerve sections shows good early regeneration in the motor and mixed nerve graft groups, with high numbers of immature regenerating fibers. This is in contrast to the sensory nerve graft group, which shows much less robust nerve regeneration with many fewer regenerating fibers evident. Overall, the histological appear-
Walking track analysis at both the 2- and 3-week time points showed paw-print lengthening consistent with tibial nerve injury. No significant recovery of print length factor was noted in any of the groups by the 3-week endpoint. Analysis of gastrocnemius muscle mass revealed a dramatic drop in mass of the denervated gastrocnemius in all groups. No inter-group differences in muscle mass ratio were detected.
Discussion A crucial factor in optimizing muscular function after denervating injury is minimizing the time required for a regenerating nerve to reach a target muscle. Muscle undergoes a progressive decrease in recovery capability proportional to the duration of denervation (Finkelstein et al., 1993; Gordon and Stein, 1982). After prolonged denervation, functional recovery of muscle may be precluded entirely
Fig. 3. This histogram shows percent of each nerve stratified for specific fiber widths: 1–2, 2–3, 3–4, 5–6, and 6–7 Am. Stratified comparison shows that the motor and mixed graft groups are composed of a greater percentage of fibers of larger diameter when compared with the sensory group. This analysis is based on fiber populations in the regenerating nerve distal to the graft. Thus, after crossing the grafts, the fiber distribution is skewed toward smaller fibers in the sensory graft groups in comparison to the motor and mixed groups.
352
C.M. Nichols et al. / Experimental Neurology 190 (2004) 347–355
Seiler, 1987; Brushart et al., 1983). These effects may involve selective upregulation of specific factors within motor and sensory pathways (Brushart et al., 1998; Martini and Schachner, 1986, 1988; Martini et al., 1994). This study demonstrates that regeneration of mixed nerves through pure sensory nerve graft pathways is inferior to that occurring through motor or mixed nerve pathways. In this experiment, the use of two motor cable grafts to match the 3 sensory cable grafts warrants special consideration. This approach allowed for significantly closer
Fig. 4. Photomicrographs of representative histologic cross-sections of nerves 3 mm distal to the nerve grafts. Robust nerve regeneration is noted in both the mixed (a) and motor (b) nerve groups, with multiple regenerating nerve fibers present. In comparison, nerve regeneration distal to the sensory (c) nerve grafts is relatively poor. Scale bar = 5 Am.
(Gutmann, 1948). Regenerating neurons interact closely with the pathways through which they regenerate. This interaction creates a microenvironment that nurtures regenerating nerve sprouts and increases the specificity and accuracy with which they return to end organ targets (Brushart, 1988; Brushart and
Fig. 5. Electron Micrograph (Uranyl Acetate–Lead Citrate Stain, 11,520 magnification): (a) Group II (sensory nerve graft). Myelinated and unmyelinated fibers have regenerated into the distal nerve. Wallerian degeneration has resolved and neural architecture has been restored. (b) Group III (motor nerve graft). The prevalence of unmyelinated fibers and overall ultrastructural appearance is similar to that noted in (a).
C.M. Nichols et al. / Experimental Neurology 190 (2004) 347–355
matching of cross-sectional area than in prior comparisons of motor/sensory grafting (Ghalib et al., 2001) and also allowed for comparison against a mixed (tibial nerve) graft. But, a roughly 13% difference in cross-sectional area persisted favoring the motor group over the sensory group. Furthermore, it has been shown in a mouse model that surgical adaptions, such as use of fibrin sealant versus suture, can influence preferential motor neuronal regeneration (Robinson and Madison, 2003). Fortunately, several factors mitigate concerns regarding a systematic bias in favor of the motor neuron graft group. First, the histomorophometry data on percent neural tissue and nerve density are normalized values that account for differences in cross-sectional area. Similarly, mean fiber width is an average among all measured fibers. Second, the disparity in cross-sectional area between motor and sensory groups was small in magnitude relative to differences in total fiber count in favor of the motor group. Third, despite the smaller caliber of sensory grafts, each sensory graft contains a greater number of Schwann cell basal lamina tubes than each motor graft. Therefore, the sensory graft group had a potential advantage in terms of Schwann cell tubes available for regeneration into grafts. Fourth, the thinner caliber of the three sensory graft cables corresponds to greater surface area for diffusion of metabolites during the period of neovascularization. Last, concerns regarding surgical adaption in the mouse, which is a particularly technically challenging model, are likely to be less applicable to rats, where cable grafting was straightforward and did not prove problematic from a technical standpoint. Standard deviations in quantitative analyses were typical for histomorphometry work on peripheral nerve regeneration. The timing of this experiment was designed based on extensive prior experience with rat peripheral nerve regeneration (Mackinnon et al., 1991). Although the 3week endpoint allowed optimization for the detection of histomorphometric differences, this came at the expense of functional analyses. At the 3-week time point the regenerating nerve front has just begun to cross the distal end of the graft. Studying regeneration at 3 weeks allows one to discern differences in the rate and magnitude of regeneration between groups. Due to the limitations of the small animal model, it is difficult to simultaneously demonstrate histomorphometric and functional differences. As a general rule, functional recovery lags behind histomorphometric evidence of regeneration (Jensen et al., 2004; Jost et al., 2000; Sunderland, in press). Failure to demonstrate differences on walking track or wet muscle mass analyses was not particularly surprising, given that the regenerating nerves had probably not reached the muscle at this early time point. A future goal of research in this area will be assessment of functional recovery. It is possible that at later time points, when functional outcomes are more readily demonstrable, histomorphometric differences may be less apparent.
353
As the preceding discussion suggests, timing is crucial in small animal research. The superior neuroregenerative capacity inherent in the rodent peripheral nervous system may cause negative controls to become positive at later time points. Because of this concern, caution must be taken in applying findings in small animal models to clinically significant injuries. Nonetheless, these studies provide the groundwork for large animal studies and eventual clinical application. Differences in nerve regeneration that may be detected only transiently in a small animal model (Calvert et al., 2001; Jensen et al., 2004; Myckatyn and Mackinnon, 2004) may be long-lasting in larger animals (Brenner et al., in press(a,b); Jensen et al., submitted for publication). Furthermore, even in the absence of a long-lasting difference in nerve regeneration, an acceleration of regeneration is clinically important due to the significant functional morbidity associated with prolonged muscle denervation. The comparatively poor regeneration in the sensory group may reflect the predilection of motor neurons to regenerate through motor pathways (PMR). Motor fibers are generally larger than sensory fibers, but increased fiber width may also reflect greater maturity nerve fibers (Evans et al., 1994). In our study, the presence of increased fiber widths in motor grafts versus sensory grafts may signify greater maturity, predominance of motor fibers, or both. Future work involving formal retrograde labeling studies is needed to distinguish between these possibilities by determining the modality of fibers regeneration though grafts. After evaluating light microscopy findings, we considered the possibility that differences in nerve regeneration between the motor versus sensory groups arose largely from differences in myelination. Specifically, were a disproportionate number of unmyelinated (and hence uncounted) fibers present in the sensory graft group? To address this question, electron microscopy was performed on nerve sections. Upon review of these sections, it was readily apparent that the difference between motor and sensory graft groups could not be explained by myelination patterns alone. The number of unmyelinated fibers found in nerves was generally proportional to the number of myelinated fibers in both groups. These findings served to confirm the validity of light microscopy-based histomorphometry data. The findings of this study are in general agreement with related work in this area. For example, one prior study has shown that motor neurite diameters and myelination increase if they are grown in motor rather than sensory pathways (Ghalib et al., 2001). However, the effects of neuron-pathway synergy on rate and magnitude of regeneration have not been previously investigated. In the current study, interactions between nerve and pathway were examined by comparing regeneration through motor versus sensory or mixed pathways. Histomorphometric analysis revealed that sensory pathways showed inferior regeneration in comparison to both the motor and mixed pathways. These results were borne
354
C.M. Nichols et al. / Experimental Neurology 190 (2004) 347–355
out in analysis of fiber density, percent nerve and total nerve fiber counts. Analysis of overall fiber distribution also shows that more large fibers were present distally in the motor and mixed graft groups (Fig. 3). This is consistent with a greater rate of regeneration through these grafts. A possible explanation for this phenomenon is the absence of motor elements in the pure sensory pathways. Both motor and sensory modalities are present in the mixed and the motor nerve grafts; this is due to the fact that the motor grafts contain sensory axons in the form of motor afferent fibers (serving muscle stretch receptors, Golgi tendon organs, etc.). Thus, given that motor associated pathways produce motor-specific factors (Ghalib et al., 2001; Martini and Schachner, 1986, 1988; Martini et al., 1992, 1994; Nieke and Schachner, 1985), it is likely that these have a trophic influence on regenerating motor neurites. This growth advantage is denied to motor neurites that are forced to grow in a pure sensory milieu. The key element in this interaction is the Schwann cell. Schwann cells are the myelinating glial cells of peripheral nerves. Following peripheral nerve injury, they de-differentiate to secrete survival and growth-promoting factors for the regenerating neurites (Bunge, 1994; Feneley et al., 1991). The specificity of motor Schwann cells to produce neurotrophins in response to motor neurites has been demonstrated in vitro and in vivo (Martini et al., 1992, 1994). It is likely that the contribution of these cells is what confers the growth advantage to the mixed modality pathways. Alternatively, given that motor nerves tend to have more myelination, it is possible that the motor nerve grafts represent a more Schwann cell-rich milieu for regeneration. These finding of the present study have direct implications to clinical nerve grafting. The existing gold standard is to utilize autogenous sensory nerve grafts for reconstruction of nerve gap injuries. Results of the present study suggest that clinical outcomes might be improved by finding alternatives to sensory grafts in the reconstruction of mixed nerves. However, there are relatively few expendable motor nerves suitable for graft material in the human body. Perhaps, a more feasible alternative lies in matching strategies that make use of exogenous graft material. This could be done in the form of allograft or biosynthetic composite graft material. In these cases, application of the results of this study to conduit or allograft technology could increase the efficacy of these modes of reconstruction. Previous studies have shown that motor neurons have inferior regeneration through synthetic nerve grafts (Madorsky et al., 1998). The use of Schwann cells and Schwann cell neurotrophins to augment nerve allografts and conduits is currently under investigation (Guenard et al., 1992; Lee et al., 2003; Mosahebi et al., 2002). With this new data showing that modality matching improves growth of peripheral nerves, research efforts can now be directed at
not only augmenting graft scaffolds with Schwann cells, but with cells derived from nerves of the appropriately matched phenotype, thereby equally enhancing the growth of all regenerating neurites.
Acknowledgment The authors would like to thank the American Society for Surgery of the Hand for supporting this research.
References Archibald, S.J., Krarup, C., Shefner, J., Li, S.T., Madison, R.D., 1991. A collagen-based nerve guide conduit for peripheral nerve repair: an electrophysiological study of nerve regeneration in rodents and nonhuman primates. J. Comp. Neurol. 306, 685 – 696. Bain, J.R., Mackinnon, S.E., Hunter, D.A., 1989. Functional evaluation of complete sciatic, peroneal, and posterior tibial nerve lesions in the rat. Plast. Reconstr. Surg. 83, 129 – 138. Brenner, M.J., Jensen, J.N., Lowe III, J.B., Myckatyn, T.M., Fox, I.K., Hunter, D.A., Mohanakumar, T., Mackinnon, S.E., in press a. Anti-CD40 ligand antibody permits regeneration through peripheral nerve allografts in a non-human primate model. Plast. Reconstr. Surg. Brenner, M.J., Lowe, I., J.B., Mackinnon, S.E., Darcy, M., Duncan, J., Fox, I.K., Jaramillo, A., Hunter, D.A., Mohanakumar, T., in press b. Effects of Schwann cells and donor antigen on regeneration through long nerve allografts. Microsurgery. Brown, C.J., Mackinnon, S.E., Evans, P.J., Bain, J.R., Makino, A.P., Hunter, D.A., Hare, G.M., 1989. Self-evaluation of walking-track measurement using a sciatic function index. Microsurgery 10, 226 – 235. Brushart, T.M., 1988. Preferential reinnervation of motor nerves by regenerating motor axons. J. Neurosci. 8, 1026 – 1031. Brushart, T.M., 1993. Motor axons preferentially reinnervate motor pathways. J. Neurosci. 13, 2730 – 2738. Brushart, T.M., Seiler, W.A., 1987. Selective reinnervation of distal motor stumps by peripheral motor axons. Exp. Neurol. 97, 289 – 300. Brushart, T.M., Tarlov, E.C., Mesulam, M.M., 1983. Specificity of muscle reinnervation after epineurial and individual fascicular suture of the rat sciatic nerve. J. Hand Surg. [Am.] 8, 248 – 253. Brushart, T.M., Gerber, J., Kessens, P., Chen, Y.G., Royall, R.M., 1998. Contributions of pathway and neuron to preferential motor reinnervation. J. Neurosci. 18, 8674 – 8681. Bunge, R.P., 1994. The role of the Schwann cell in trophic support and regeneration. J. Neurol. 242, S19 – S21. Calvert, G.T., Doolabh, V.B., Grand, A.G., Hunter, D.A., Mackinnon, S.E., 2001. Rat-strain differences in recovery following peripheral-nerve allotransplantation. J. Reconstr. Microsurg. 17, 185 – 191. de Medinaceli, L., Freed, W.J., Wyatt, R.J., 1982. An index of the functional condition of rat sciatic nerve based on measurements made from walking tracks. Exp. Neurol. 77, 634 – 643. Devor, M., Schneider, G., 1975. Neuroanatomical plasticity: the principle of conservation of total axonal arborization. In: Vital-Durand, F., Jeannerod, M. (Eds.), Asp. Neural Plast., vol. 43. INSERM Colloquia, Paris, pp. 191 – 200. Doolabh, V.B., Hertl, M.C., Mackinnon, S.E., 1996. The role of conduits in nerve repair: a review. Rev. Neurosci. 7, 47 – 84. Evans, P.J., Midha, R., Mackinnon, S.E., 1994. The peripheral nerve allograft: a comprehensive review of regeneration and neuroimmunology. Prog. Neurobiol. 43, 187 – 233. Feneley, M.R., Fawcett, J.W., Keynes, R.J., 1991. The role of Schwann cells in the regeneration of peripheral nerve axons through muscle basal lamina grafts. Exp. Neurol. 114, 275 – 285.
C.M. Nichols et al. / Experimental Neurology 190 (2004) 347–355 Finkelstein, D.I., Dooley, P.C., Luff, A.R., 1993. Recovery of muscle after different periods of denervation and treatments. Muscle Nerve 16, 769 – 777. Ghalib, N., Houst’ava, L., Haninec, P., Dubovy, P., 2001. Morphometric analysis of early regeneration of motor axons through motor and cutaneous nerve grafts. Ann. Anat. 183, 363 – 368. Gordon, T., Stein, R.B., 1982. Time course and extent of recovery in reinnervated motor units of cat triceps surae muscles. J. Physiol. 323, 307 – 323. Guenard, V., Kleitman, N., Morrissey, T.K., Bunge, R.P., Aebischer, P., 1992. Syngeneic Schwann cells derived from adult nerves seeded in semipermeable guidance channels enhance peripheral nerve regeneration. J. Neurosci. 12, 3310 – 3320. Gutmann, E., 1948. Effect of delay of innervation on recovery of muscle after nerve lesions. J. Neurophysiol. 11, 279 – 294. Jensen, J.N., Tung, T.H.H., Mackinnon, S.E., Brenner, M.J., Hunter, D.A., 2004. Use of anti-CD40 ligand monoclonal antibody as anti-rejection therapy in a murine peripheral nerve allograft model. Microsurgery 24, 309–315. Jensen, J.N., Brenner, M.J., Tung, T.H.H., Hunter, D.A., Mackinnon, S.E., submitted for publication. FK506 accelerates peripheral nerve regene-ration through long grafts in inbred swine. Ann. Plast. Surg. Jost S.C., Doolabh, D.V., Mackinnon, S.E., Lee, M., Hunter, D., 2000. Acceleration of peripheral nerve regeneration following FK506 administration. Restor. Neurol. Neurosci. 17, 39 – 44. Kreiger, N., Kelsey, J.L., Harris, C., Pastides, H., 1981. Injuries to the upper extremity: patterns of occurrence. Clin. Plast. Surg. 8, 13 – 19. Le, T.B., Aszmann, O., Chen, Y.G., Royall, R.M., Brushart, T.M., 2001. Effects of pathway and neuronal aging on the specificity of motor axon regeneration. Exp. Neurol. 167, 126 – 132. Lee, A.C., Yu, V.M., Lowe III, J.B., Brenner, M.J., Hunter, D.A., Mackinnon, S.E., Sakiyama-Elbert, S.E., 2003. Controlled release of nerve growth factor enhances sciatic nerve regeneration. Exp. Neurol. 184, 295 – 303. Mackinnon, S.E., Dellon, A. Lee., 1988. Surgery of the Peripheral Nerve. Thieme Medical Publishers, New York. Mackinnon, S.E., Hudson, A.R., Falk, R.E., Kline, D., Hunter, D., 1984. Peripheral nerve allograft: an assessment of regeneration across pretreated nerve allografts. Neurosurgery 15, 690 – 693. Mackinnon, S.E., Hudson, A.R., Falk, R.E., Hunter, D.A., 1985. The nerve allograft response—An experimental model in the rat. Ann. Plast. Surg. 14, 334 – 339. Mackinnon, S.E., Dellon, A.L., O’Brien, J.P., 1991. Changes in nerve fiber numbers distal to a nerve repair in the rat sciatic nerve model. Muscle Nerve 14, 1116 – 1122.
355
Mackinnon, S.E., Doolabh, V.B., Novak, C.B., Trulock, E.P., 2001. Clinical outcome following nerve allograft transplantation. Plast. Reconstr. Surg. 107, 1419 – 1429. Madison, R.D., Archibald, S.J., Brushart, T.M., 1996. Reinnervation accuracy of the rat femoral nerve by motor and sensory neurons. J. Neurosci. 16, 5698 – 5703. Madison, R.D., Archibald, S.J., Lacin, R., Krarup, C., 1999. Factors contributing to preferential motor reinnervation in the primate peripheral nervous system. J. Neurosci. 19, 11007 – 11016. Madorsky, S.J., Swett, J.E., Crumley, R.L., 1998. Motor versus sensory neuron regeneration through collagen tubules. Plast. Reconstr. Surg. 102, 430 – 436. Martini, R., Schachner, M., 1986. Immunoelectron microscopic localization of neural cell adhesion molecules (L1, N-CAM, and MAG) and their shared carbohydrate epitope and myelin basic protein in developing sciatic nerve. J. Cell Biol. 103, 2439 – 2448. Martini, R., Schachner, M., 1988. Immunoelectron microscopic localization of neural cell adhesion molecules (L1, N-CAM, and myelin-associated glycoprotein) in regenerating adult mouse sciatic nerve. J. Cell Biol. 106, 1735 – 1746. Martini, R., Xin, Y., Schmitz, B., Schachner, M., 1992. The L2/HNK-1 carbohydrate epitope is involved in the preferential outgrowth of motor neurons on ventral roots and motor nerves. Eur. J. Neurosci. 4, 628 – 639. Martini, R., Schachner, M., Brushart, T.M., 1994. The L2/HNK-1 carbohydrate is preferentially expressed by previously motor axonassociated Schwann cells in reinnervated peripheral nerves. J. Neurosci. 14, 7180 – 7191. Mosahebi, A., Fuller, P., Wiberg, M., Terenghi, G., 2002. Effect of allogeneic Schwann cell transplantation on peripheral nerve regeneration. Exp. Neurol. 173, 213 – 223. Myckatyn, T.M., Mackinnon, S.E., 2004. A review of research endeavors to optimize peripheral nerve reconstruction. Neurol. Res. 26, 124 – 138. Nieke, J., Schachner, M., 1985. Expression of the neural cell adhesion molecules L1 and N-CAM and their common carbohydrate epitope L2/ HNK-1 during development and after transection of the mouse sciatic nerve. Differentiation 30, 141 – 151. Nolte, C., Schachner, M., Martini, R., 1989. Immunocytochemical localization of the neural cell adhesion molecules L1, N-CAM, and J1 in pacinian corpuscles of the mouse during development, in the adult and during regeneration. J. Neurocytol. 18, 795 – 808. Robinson, G.A., Madison, R.D., 2003. Preferential motor reinnervation in the mouse: comparison of femoral nerve repair using a fibrin sealant or suture. Muscle Nerve 28, 227 – 231. Sunderland, I.R.P., Brenner, M.J., Singham, J., Hunter, D.A., Mackinnon, S.E., in press. Effect of tension on nerve regeneration in rat sciatic nerve transection model. Ann. Plast. Surg.