REGENERATION ACROSS COLD PRESERVED PERIPHERALNERVE ALLOGRAFTS
PETER J. EVANS, M.D., Ph.D., 1* SUSAN E. MACKINNON, M.D., 2
RAJIV MIDHA, M.D., M.Sc., 3 JUDITH A. WADE, B.Sc., 3
DANIEL A. HUNTER, R.T., 2 YASUSHI NAKAO, M.D., Ph.D., 2 and
GREGORY M.T. HARE, M.D., Ph.D.3
The feasibility of peripheral nerve allograft pretreatment uti-lizing cold storage (5°C in the University of Wisconsin ColdStorage Solution) or freeze-thawing to prevent rejection wasinvestigated. Regeneration across cold-stored (3 or 5 weeks)or freeze-thawed (FT), 3.0-cm sciatic nerve allografts werecompared to fresh auto- and allografts in an inbred rat model.At 16-week post-engraftment, only FT allografts appearedsimilar to autografts on gross inspection; FT grafts were nei-ther shrunken nor adherent to the surrounding tissue as seenin the other allograft groups. Qualitatively, the pattern of re-generation in the graft segments of the fresh allograft and toa lesser extent of pretreated allografts was inferior to that ofautografts as evidenced by a disruption in the perineurium,more extrafascicular axons, smaller and fewer myelinatedaxons, increased intrafascicular collagen deposition, and thepersistence of perineurial cell compartmentation and perivas-cular infiltrates. Distal to these grafts, the regeneration be-came more homogenous between groups, although areas ofongoing Wallerian degeneration, new regeneration as well ascompartmentation, were more prevalent in fresh and pre-
treated allografts. Although the number of myelinated fibreswas equivalent to autografts, the fibre diameters, the numberof large diameter fibres, and the G-ratio were significantlydecreased in the allograft groups, which, in part, accountedfor the significant decrease in conduction velocity in the3-week stored and fresh allograft, and the slight decrease inthe 5-week stored and FT allograft groups. There was a smallreturn in the Sciatic Function Index towards normal, but noconsistent differences between groups were found. Pro-longed cold storage and freeze-thawing of nerve allograftsresulted in regeneration that was better than fresh allografts,but inferior to autografts. With the concomitant use of hostimmunosuppression or other immunotherapies, these stor-age techniques can provide a means of transporting nerveallografts between medical centres and for converting urgentinto elective procedures.
Conventional management of the peripheral nerve gap isgrafting with a limited number of “expendable,” small di-ameter, cutaneous nerve autografts (e.g., sural nerve). Al-ternatively, cadaveric nerve allografts would provide a vastsupply of nerve graft material for reconstructive proceduresand obviate problems of donor site morbidity (residualnumbness, neuroma formation, and a notable scar).
Our laboratory has extensively investigated methods ofgraft pretreatment1–8 and recipient immunosuppression9–11
with the aim of facilitating allograft tissue acceptance andsubsequent regeneration. Regeneration across nerve allo-
grafts in recipients immunosuppressed with cyclosporin Awas equivalent to autografts in both rat12–18 and primatemodels.19–22 In the future, a necessary adjunct to clinicalnerve allografting is the ability to store harvested nervesduring transport. Furthermore, prolonged storage would in-crease access and allow time for testing of donor tissue andfacilitate elective planning and less costly reconstructiveprocedures. Short-term (days to weeks) storage might beachieved by use of a standard transplant solution and long-term (weeks to months) by deep freezing.
Recently, we investigated nerve graft storage in an au-tograft model to avoid the variables of an allograft response.Regeneration across autografts following storage at 5°C for3 weeks in the Belzer/University of Wisconsin Cold StorageSolution (UWCSS) was equivalent to fresh nerve auto-grafts.8 Using a rabbit model, Gutmann and Sanders23,24
and Sanders and Young25 found that short (2 cm) nerveallografts stored for 7–21 days in Ringer’s solution at 2°Cexhibited a reduced cellular infiltrate compared to fresh al-lografts and regeneration similar to autografts. Based onthese findings with allografts and our encouraging resultswith autografts, we assessed nerve allograft storage in the
1Department of Orthopaedic Surgery, Johns Hopkins University School ofMedicine, Baltimore, Maryland
2Division of Plastic Surgery, Washington University School of Medicine, St.Louis, Missouri
3Department of Surgery, University of Toronto, Toronto, Ontario, Canada
Grant sponsor: Medical Research Council of Canada; Grant sponsor: Cana-dian Orthopaedic Research and Education Foundation; Grant sponsor: NIH;Grant number: NS34406-01.
*Correspondence to: Dr. Peter J. Evans, Department of Orthopaedic Surgery,Johns Hopkins Bayview Medical Center, 4940 Eastern Avenue, Baltimore, MD21224.
UWCSS using an inbred rodent model, not as an adjunct toimmunosuppression, but rather as a method of graft pre-treatment.
Nerve allograft pretreatment by repeated cycles of deepfreezing and thawing to render grafts acellular was firstdescribed in 1954 by Sanders26 and is of current inter-est.27–33However, the latter authors predegenerated the do-nor nerves for 1–6 weeks in vivo prior to harvesting andfreeze-thawing. Predegeneration fails to render allograftsnonimmunogenic1 does not enhance regeneration25,34,35andhas no clinical application. Nevertheless, freeze-thawingmay provide both a method of reducing the immunogenicityof the nerve allograft and provide a means of long-termgraft storage. However, regeneration across freeze-thawed(FT) allografts was not assessed quantitatively. Further-more, recent studies have demonstrated reduced regenera-tion across acellular vs. fresh autografts.28,36,37
In the present study, we quantitatively assessed whetherregeneration across 3.0-cm sciatic nerve allografts that werepretreated by either cold storage for 3 or 5 weeks in theUWCSS or freeze-thawed, would be comparable to freshautografts and superior to fresh allografts in an inbred ratmodel.
MATERIALS AND METHODS
Animal Model
Inbred male Lewis (RT1l) rats (Harlan Sprague Dawley,Indianapolis, IN), 200–250 g, were used as recipients ofeither syngeneic (Lewis, n4 24) or allogeneic ACI (RT1a)3-cm sciatic nerve grafts. Allogeneic nerve grafts weregrafted either fresh (n4 26) or pretreated by storage for 3weeks (n4 24) or 5 weeks (n4 25) or by freeze-thawing(n 4 25). In addition, age- and weight-matched Lewis ratsserved as normal controls (n4 6). Nerve graft recipientswere acclimatized prior to surgical procedures, housed inflat bottom cages post-operatively, and allowed standard ratchow and water ad libitum. All experiments were performedin accordance to the standards of the Animal Care Commit-tee at the University of Toronto.
Surgical Procedures
All procedures were performed under ketamine (100mg/kg IM, Rogarsetic, Rogar-STB, Montreal, Quebec,Canada) and acepromazine maleate (10 mg/kg IM, Atravet,Ayerst Laboratories, Montreal, Quebec, Canada) anaesthe-sia. Following exposure, 3 cm of sciatic nerve was har-vested by transection proximal to the sciatic notch and distalat the site of trifurcation. Allogeneic grafts were randomizedto fresh, storage, or FT groups. Subsequently, the right sci-atic nerve in the recipient was exposed and sharply tran-sected in the mid-thigh proximal to its trifurcation. Thedonor nerve graft was reversed longitudinally 180°, inter-posed and microneurosurgically repaired to the recipient
nerve with 10-0 epineurial sutures (Dermalon, Davis andGeck, American Cyanamid Company, Danbury, CT). Theleft sciatic nerve was left non-operated and served as acontrol for walking track analysis. At endpoint (16 weeks)electrophysiologic studies, the animals were anesthetized,paralyzed with tubocurare (0.3 mg/kg IP, Burroughs Well-come Inc., Kirkland, Quebec, Canada), tracheotomized, andthen ventilated with room air supplemented with oxygen.An EKG monitor (Model 78213C, Hewlett-Packard, An-dover, MA) was used to follow the heart rate throughout theprocedure.
Storage Parameters
Immediately following harvest, the nerve grafts werestored in sterile 25-mm2 flasks (Becton Dickinson, Missis-sauga, Ontario, Canada) containing 15 ml of the UWCSS(DuPont, Scarborough, Ontario, Canada) supplementedwith 1% Penicillin-Streptomycin (100 U/ml Penicillin and100 mg/ml Streptomycin, GIBCO, Burlington, Ontario,Canada), 40 U/L Humulin-Regular (Eli Lilly & Co., India-napolis, IN), and 16 mg/L dexamethasone (OrganonTeknika, Toronto, Ontario, Canada) prepared fresh on theday of use. Allografts were stored in a controlled environ-ment at 5°C for either 3 weeks (3-wk) or 5 weeks (5-wk).FT allografts were prepared by placing fresh grafts ontocopper plates and freezing them on dry ice (approximately−40°C) for 4 minutes, then thawing them for 5 minutes atroom temperature and then repeating the cycle a total of 5times.27,28,30,31,33In the syngeneic (autograft) and alloge-neic (FRESH) control groups, nerves were immediatelygrafted and no pretreatment was performed.
Assessment of Nerve Regeneration
Walking track analysis. A previously described sciaticfunction index (SFI) was used to serially evaluate globalhind limb function by assessing walking track pat-terns.8,38–40 Lewis rats were chosen as recipients becausethey do not engage in autotomy of their grafted limb.41
Briefly, the walking track consists of an 8.2 × 42 cm trackdarkened at one end with a length of exposed X-ray filmplaced on the bottom of the track. The hind feet of the ratwere placed in a Petri dish containing X-ray film developerand then the rat was allowed to walk down the track. Thehind footprints appeared quickly as the developer reactedwith the film. Measurements of footprints from walkingtracks were used to calculate a sciatic function index asdescribed previously.8,38–40
Electrophysiologic evaluation. Electrophysiologic re-cordings across the nerve grafts were made using a com-puter-assisted electromyographic machine (Advantage,Clark-Davis Medical Inc., London, Ontario, Canada). In asuitably prepared animal, the sciatic nerve was exposed bothabove the sciatic notch, 1.5–2.0 cm proximal to the proxi-mal repair site (stimulation site), and distally, 0.5–1.0 cm
116 Evans et al.
beyond the distal repair site (recording site), carefullyavoiding the graft site so as not to disturb any tenuousregenerated fibres. Bipolar hooked platinum stimulating andrecording electrodes were placed on the proximal and distalparts of the sciatic nerve, respectively, providing a 3.0–6.5-cm inter electrode distance. A “ground” was placed inmuscle midway between the stimulating and recording elec-trodes. A constant current stimulator provided a stimulusduration of 0.05 ms. After determining a supramaximal re-sponse, a minimum of four and a maximum of ten directcompound nerve action potentials per nerve were recordedand averaged. The compound nerve action potential(CNAP) conduction velocity was taken from the leadingedge (shortest latency) of the dominant negative wave andrepresents the maximum CNAP conduction velocity. Tem-perature was monitored and maintained at approximately37°C using a heating lamp and cellophane was used toprevent drying.Histologic evaluation. Following electrophysiologic re-cordings, the entire sciatic nerve including the proximalsegment, nerve graft, and distal segment, was removed. Theproximal host segment and proximal half of the graft wasfixed by immersion in 10% w/v buffered formalin at pH 7.0and then embedded in paraffin (Fig. 1). Longitudinal sec-tions of 5 mm in thickness were made through the proximalsuture line and cross-sections were made approximately 10mm proximal (host) and distal (donor graft) to the sutureline and all sections were stained with Masson’s trichromestain.
The distal half of the graft and distal host segment werefixed by immersion in a 3% v/v phosphate buffered glutar-aldehyde solution for 24 hours at pH 7.20–7.40 (MarivacEM Supplies, Halifax, Nova Scotia, Canada). The tissuewas post-fixed in 1% w/v phosphate buffered osmium te-troxide for 12 hours, serially ethanol dehydrated, and infil-trated and embedded in Araldite 502 epoxy resin (CIBA-Geigy Canada Ltd., Dorval, Quebec, Canada). Cross-
sections were made approximately 10 mm proximal (donorgraft) and distal (host) to the distal suture line. A 1% w/vtoluidine blue solution was used to stain 1-mm-thick sec-tions for light microscopy. Ultrathin sections (800 mm)were cut with an LKB III ultramicrotome (LKB-ProdukterA.B., Bromma, Sweden), placed on 300-mesh copper/rhodium grids, stained with 5% w/v uranyl acetate in 70%methanol for 5 minutes, dried, and stained with lead citratefor 5 minutes and dried, and examined on a Zeiss 10a trans-mission electron microscope (Zeiss, Oberkochen, West Ger-many).Morphologic evaluation. Morphologic evaluation wascarried out on the myelinated fibres in the host sciatic nervesegment at least 0.5 cm distal to the graft. Measurementswere made at 1,000× magnification using computer-automated morphometry software (Leco 2005, St. Joseph,MI) and based on a minimum of 500 myelinated fibresmeasured in 5 representative fields (centre, 3, 6, 9, and 12o’clock) of the tibial branch of the sciatic nerve. Smallestmean sieve axon and fibre (axon + myelin) diameter, myelindiameter, G-ratio (axon/fibre diameter), fibre density (num-ber/mm2 sampled), and the total number of fibres (number/sciatic nerve area) that regenerated across the grafts werecalculated.
Statistical Analysis
An overall comparison between experimental groupswas performed by a one-way analysis of variance (1W-ANOVA) and, if significant (two-tailedP < 0.05), post-hocpairwise comparisons between individual groups were madeusing the Tukey’s test. Post-hoc comparisons to normalcontrols and nerve graft controls (autograft) were performedusing the Dunnett’s test for control groups. An analysis ofthe presence or absence of a measurable walking track wasassessed by the Chi-square test. Both Tukey’s and Dunnett’scorrected for multiple comparisons, and for all statistics anoverall two-tailed alpha level of 0.05 was considered statis-tically significant. Data in text and tables are expressed asmean ± standard deviation and in graphs as mean ± standarderror.
RESULTS
Gross Appearance of Grafts
The surrounding tissue was easily dissected off all au-tograft and FT nerve allografts on exploration at 16 weeks.They appeared similar in colour to the host nerve and wereapproximately the same length compared to their originallength. In contrast, all FRESH, 3-wk and 5-wk grafts werefirmly adherent to the surrounding tissue and there wasprominent scarring, especially at each suture line. Thesegrafts were yellowish in colour and markedly shrunken.Furthermore, some FRESH grafts appeared almost com-pletely resorbed and there was new regeneration that had
Figure 1. Pictorial of histologic technique. Each nerve was removeden bloc including the proximal, graft, and distal segments. Half wasfixed in formalin for Masson’s trichrome staining (light microscopy)and the other half in glutaraldehyde for toluidine blue (light micros-copy) or uranyl lead citrate (electron microscopy) staining. Histologicsections were made in either a cross-sectional (XS) or longitudinal(LS) fashion.
Banking Nerve Allografts 117
grown around the original graft and down to the distal su-ture line.
Walking Track Analysis
General observations. Immediately following grafting,all animals were unable to spread their toes. When theywalked on all feet, they were unable to walk solely on theirfoot pads, but instead walked on their foot pads and theproximal part of their foot. Therefore, their footprints be-came narrowed and longer than preoperatively. With time,they again were able to spread their toes and walk on theirfoot pads, although to varying degrees. In addition, all ani-mals developed a fixed ankle joint flexion deformity. At 12weeks, the walking tracks were unmeasurable in 13 animals(three autograft, five FRESH, three 3-wk, one FT, one5-wk), which was not statistically significant (P 4 0.34,Chi-square). At 16 weeks, 19 animals were unmeasurable(seven autograft, five FRESH, one 3-wk, five FT, one5-wk), and this resulted in a trend towards significance (P4 0.06, Chi-square). Unmeasurable tracks were due to ei-ther marked toe-flexion deformity or ambulation on the dor-sum of the foot.SFI. The changes in SFI are depicted in Figure 2. Postop-eratively, the mean SFI declined in all groups to a levelbetween -83 and -91 at 2 weeks. The SFI increased mar-ginally in all groups over the following 14 weeks. The dif-ference in SFI between groups was found to be highly vari-able and no consistent differences were seen. Therefore,further statistical analysis was not deemed useful.
Electrophysiology
The distal CNAP amplitude (mV) was highly variable(Fig. 3), but the autograft group (94.6 ± 91.6) was signifi-cantly greater than the FRESH graft group (38.6 ± 50.8) (P
< 0.05, Dunnett’s,P 4 0.02, 1W-ANOVA). The CNAPnegative peak area (mVms) was also highly variable and nosignificant differences were detected between groups (P 40.053, 1W-ANOVA). The difference in CNAP conductionvelocity (m/s) between groups was highly significant (P <0.0001, 1W-ANOVA) (Fig. 4). The conduction velocityacross autografts (34.8 ± 13.2) was greater than all othergraft groups, but only significantly greater than the FRESHand 3-wk graft group (16.8 ± 15.8 and 17.6 ± 13.7) (P <0.01, Dunnett’s). The conduction velocity across the FT(28.6 ± 8.0) and 5-wk (28.0 ± 14.6) stored grafts weresignificantly greater than across the FRESH and 3-wkstored grafts (P < 0.05, Tukey’s). The CNAP amplitude(2,827.8 ± 891.1), negative peak area (1,695.5 ± 455.7), andconduction velocity (65.8 ± 15.0) in the normal controlgroup were significantly greater than in all graft groups (P< 0.01, Dunnett’s,P < 0.0001, 1W-ANOVA). All otherbetween group comparisons were not statistically signifi-cant.
Histology
Graft segments. Despite the gross appearance of rejec-tion in the FRESH and stored allograft groups, there wasevidence of regeneration in almost all specimens by light(Figs. 5 and 6) and electron microscopy (Figs. 7–10). Whilethe FT allografts appeared grossly similar to autografts, thepattern of axonal regeneration in the graft segments wasinferior to autografts and more similar to FRESH and storedallografts. All allografts appeared to be at an earlier stage of
Figure 2. Walking track analysis in stored nerve allograft recipients.The sciatic function index (SFI) recovered very little, was highly vari-able, and no consistent trends were found. Data are expressed asmean ± SEM.
Figure 3. CNAP amplitude across stored nerve allografts. The am-plitude of the CNAP in the normal (N) group was significantly greater(A) than all graft groups (P < 0.01, Dunnett’s, P < 0.0001, 1W-ANOVA) and significantly greater (B) in the autograft (AUTO) com-pared to the FRESH (C) allograft group (P < 0.05, Dunnett’s, P =0.02, 1W-ANOVA). There was no significant difference between the3-wk, 5-wk, and FT groups compared to either the autograft or theFRESH graft groups. Data are expressed as mean ± SEM.
118 Evans et al.
regeneration, evidenced by the presence of compartmenta-tion, smaller diameter, and fewer myelinated fibres. Fur-thermore, there was evidence of ongoing cellular infiltratesby lymphocytes and thickened, abnormal vessels in all al-lograft groups. In the FRESH, and to a slightly lesser extentin the FT and stored allografts, the perineurium was oftendisrupted with extrafascicular regeneration noted; collagendeposition and less uniform regeneration occurred withinthe original fascicles (compare in Fig. 11).Distal segments. Qualitatively, there was excellent re-generation in all graft groups on light and electron micros-copy (Figs. 12 and 13). There was evidence of ongoingdegeneration adjacent to areas of regeneration, although thiswas more apparent in the allograft groups. Similar to thegraft segments, there were areas of compartmentation in theallograft groups, but not in the autograft group. Regenera-tion in all groups was noted to predominantly occur downoriginal SCBL tubes (Fig. 13). Several collapsed SCBLtubes were noted in all groups.
Morphometry
Morphometric measures were performed on myelinatedfibres in the distal host tibial fascicle at least 5 mm distal tothe distal suture line (Table 1). In general, compared tonormals, in all graft groups there were less fibres, the axonand the fibre diameter and the percentage of large diameterfibres (>7 mm) were significantly less (P < 0.01, Dunnett’s,P < 0.0001, 1W-ANOVA), which corresponded to a sig-nificantly greater fibre density (P < 0.01, Dunnett’s,P 40.002, 1W-ANOVA). Although the myelin thickness was
significantly less in the nerve auto- and allograft groups, theG-ratio (axon/fibre diameter) was also significantly less,indicating a relatively increased myelination (P < 0.01,Dunnett’s,P < 0.0001, 1W-ANOVA).
There was no statistically significant difference betweenauto- and allograft groups in the total number of myelinatedfibres per mm2 or in the fibre density (number/mm2
sampled) of tibial nerve (P 4 0.20, 1W-ANOVA, Table 1).Axon and fibre diameters (1.76 ± 0.16 and 3.4 ± 0.18,respectively) and the percentage of large diameter fibres (>7mm) in the autograft group were significantly greater thanall allograft groups (P < 0.01, Dunnett’s,P < 0.005, 1W-ANOVA) (Fig. 14). Comparisons between graft groups re-vealed no significant differences in myelin thickness, butthe G-ratio in the autograft group (0.54 ± 0.04) was signifi-cantly greater (increased myelination) compared to theFRESH and 3-wk group (P < 0.05 andP < 0.01, respec-tively, Dunnett’s,P 4 0.004, 1W-ANOVA). In summary,all graft groups had aberrant morphometric parameters, withthe autograft group being closest to normal.
DISCUSSION
The availability of nerve allografts for the repair of largeperipheral nerve gaps would be of great benefit to patientsand reconstructive surgeons. Since the first reported nerveallograft in 1885 by Albert,42 numerous methods of graftpretreatment and host immunosuppression have been at-tempted to avoid or overcome the host immune response(for a review see4,27). Only host immunosuppression withcyclosporin A was quantitatively assessed and demonstratedto prevent rejection and result in regeneration equivalent toautografts in rat12–18and primate models.19–22Pretreatmentby freeze-thawing was similarly found to prevent rejectionand allow similar regeneration to autografts as assessed byqualitative histology.27,29–33In a rabbit model, Sanders andYoung25 and Gutmann and Sanders23,24stored nerves for 7,14, or 21 days in Ringer’s solution at 2°C prior to grafting.They noted a marked decrease in viability of Schwann cells(SCs) with longer storage and a subsequently decreasedlymphocytic infiltrate and better nerve regeneration in the14- and 21-day stored grafts in comparison to fresh allo-grafts. The function in the stored allograft groups was onlyslightly less than that in autografts.
The relative merits of freeze-thawing and graft storagecannot be ascertained from the above studies, which utilizedeither outbred animal models,23–25 non-quantitative out-come parameters,27,30–33 short nerve allograft segments(0.5–2.0 cm),23–25,27,29,30or lacked autograft controls.30–33
In most studies, sample size was relatively small. To mini-mize these deficiencies, we transplanted a 3-cm nerve allo-graft in an inbred rat model, utilizing both qualitative his-tology and quantitative outcome parameters (histomor-
Figure 4. CNAP conduction velocity (CV) across stored nerve allo-grafts. The difference in CV between groups was highly significant (P< 0.0001, 1W-ANOVA). The CV across the normal (N) group wassignificantly greater than all graft groups (P < 0.01, Dunnett’s, P <0.0001, 1W-ANOVA). The CV across the fresh and 3-wk allograftgroups was significantly slower than the autograft, 5-wk, and FTallograft groups (P < 0.05, Tukey’s). Values with different letters aresignificantly different from each other. Data are expressed as mean± SEM. Reproduced from Evans et al.4 with permission of the pub-lisher.
Banking Nerve Allografts 119
phometry, electrophysiology, and a serial walking trackanalysis).
Recently, in an ovine model, at 7 days post-engraftmentthe cellular infiltrate into 3-wk stored, as compared to freshallografts, was noted to be reduced using an 111-indiumlabeled lymphocyte migration assay.43 The amount of lym-phocyte infiltrate in response to FT allografts was similar to
autografts. The reduced cellular infiltrate suggested a de-creased and absent host immune response to 3-wk storedand FT allografts, respectively. Cold preservation delayedor prevented the infiltration of activated lymphocytes (CD5and class II MHC staining positive).44 However, while gen-eralized (early) and activated (late) lymphocyte migrationinto the graft site was profoundly reduced over a prolonged
Figure 5. Autograft nerve; graft segment 16 weeks after grafting. There is uniform regeneration of well-myelinated axons and few focal areasof continued Wallerian degeneration. Blood vessels appear normal (arrow). Toluidine blue, light microscopy, ×540. Scale bar = 30 µm.
Figure 6. Five-week stored graft; graft segment 16 weeks after grafting. There is evidence of good myelination and continued Walleriandegeneration. Note also the distinct compartmentation (arrows). Toluidine blue, light microscopy, ×540. Scale bar = 30 µm.
120 Evans et al.
time (3–4 weeks), quantitative cellular outputs and pheno-typic characteristics of efferent lymphocytes from regionallymph nodes revealed the allo response was still present.44
In an identical rat study to the present, we have demon-strated that at least some SCs and fibroblasts remain viableafter 3-wk, but not 5-wk of graft storage or freeze-thawing.6,45 Donor-specific antibody (humoral response)
was present as early as 2 weeks post-engraftment in recipi-ents of FRESH and 3-wk allografts, but it was detected laterat 4 weeks in 5-wk allograft recipients and not at all in FTallografts. This may represent an altered kinetics of immu-nogenicity to 5-wk and FT allografts, possibly due to theabsence of viable cells within these grafts.6 Structures pro-posed as antigenic component(s) in the nerve allograft in-clude myelin46,47 perineurial cells and fibroblasts,48,49 butdata is minimal.50 The critical immunogenic componentmay be an antigen presenting cell (APC) such as Schwanncells, endothelial cells, or perivascular macrophage-likecells that can express major histocompatibility antigens(MHC). For a complete discussion of nerve allograft im-munogenicity see refs.4,6
Repetitive freeze-thawing kills all cells within a nervegraft, but leaves the perineurial and endoneurial (SCBL)connective tissue structures and at least one extracellularmatrix component (laminin) of SCBL intact, as assessed bylight and electron microscopy.6 Freeze-thawing drasticallyreduces early host cellular infiltrates into nerve allo-grafts6,27,29–33,43and prevents any detectable early or latehumoral response.6 Basal lamina is the basement membranethat is secreted by and surrounds SCs, delimiting each en-doneurial tube, in the peripheral nerve.51 During regenera-tion, axons extended along the inner aspect of the SCBL,6,52
Figure 7. Autograft nerve; graft segment 16 weeks after grafting.Excellent regeneration of myelinated and unmyelinated axons andcomplete dissolution of the few former compartments (arrows) isnoted with a small remnant of a perineurial cell (lower arrow).Schwann cells can be seen around myelinated and unmyelinatedaxons. There is normal appearing collagen between fibres. Uranyllead citrate, electron microscopy, ×4,800. Scale bar = 5 µm.
Figure 8. Fresh nerve allograft; graft segment 16 weeks after graft-ing. Numerous, but smaller diameter-regenerating myelinated andunmyelinated axons associated with Schwann cells. Evidence ofthickened, abnormal blood vessel (V) and compartmentation (arrow-head). Uranyl lead citrate, electron microscopy, ×4,800. Scale bar =5 µm.
Figure 9. Five-week stored nerve allograft; graft segment 16 weeksafter grafting. Numerous unmyelinated and large diameter, but thinlymyelinated axons are associated with Schwann cells in a large com-partment delimited by perineurial cells (arrows). There is evidence ofcellular debris and a thickened and abnormal appearing perineurium(P) and epineurium (E). Uranyl lead citrate, electron microscopy,×3,000. Scale bar = 5 µm.
Banking Nerve Allografts 121
probably because of their rich laminin content,6,27,53whichis a strong neurite promoting factor in vitro (i.e., to promoteand guide axonal elongation).51,54–56Successful regenera-tion across acellular scaffolds of nerve27 and muscle57–64
can be severely decreased by pretreatment of nerve53 andmuscle65 with antibodies to laminin. Acellular grafts, how-ever, lack SCs, which play a critical role in regeneration bysupplying “neurotrophic” factors, such as nerve growth fac-tor,66–68 to the regenerating neurites. Additionally, the SCitself confers neurite promoting effects,55 which may haveits molecular basis on ligand receptor interactions betweenthe axon and the SC surface.69,70 The lack of SCs in acel-lular grafts and their finite migratory ability was identifiedas limiting the extent28,37,71 and speed72 of regenerationseen with acellular as opposed to cellular autografts. Wehave also shown delayed early (2 and 4 week) regenerationacross FT allografts,6 which may indicate that acellulargrafts lack sufficient trophic support for normal rates ofgrowth.
At 16 weeks post engraftment, regeneration through anddistal to nerve autografts and all allograft groups was found.The MHC disparity between donors and recipients in thepresent study was previously shown by histology to result inearly rejection of fresh allografts.1,2,6,9–11,16,18,73Despite anappearance of rejection on gross inspection at 16 weeks inthe present study, equivalent numbers (and density) of my-elinated fibres regenerated across FRESH allografts com-pared to autografts. This was seen previously in a short (2cm) rat tibial nerve allograft model.18 However, the mean
fibre diameter in the FRESH allograft group was signifi-cantly smaller, and resulted in a slower CNAP conductionvelocity74,75as compared to the autograft group. Similarly,there were fewer large diameter (>7 mm) myelinated fibresin the FRESH allograft group, which may account, in part,for the significantly smaller CNAP amplitude found com-pared to autografts.76 Therefore, despite statistically equiva-lent numbers of myelinated fibres, the electrophysiologicfunction of FRESH nerve grafts was inferior to autografts,which may have its basis in the reduced fibre diameterspresent distally.
Many authors have noted that following proximal nervetransection, there is a transient (7–14 day) increase,77,78thena decrease in fibre diameter in the distal nerve segment thatis progressive with increasing periods of denervation.77–81
The transient increase in SCBL tube diameter is attributedto myelin ovoid formation and SC hypertrophy and hyper-plasia. Decreased cellular contents from metabolic degrada-tion of axon and myelin debris during Wallerian degenera-tion may account for the progressive decrease and eventualcollapse of SCBL tubes.78 Failing reinnervation, fibre di-ameter decreased rapidly and significantly by 4 weeks78 andan 80–90% reduction in the diameter of the largest fibresoccurred after 89 days of denervation.77 After 4 weeks ormore of denervation, progressively increased discontinuity,dispersion, and partial disappearance of SCBL occurred.78
Following both immediate neurorrhaphy or nerve grafting,axons reenter the distal stump and regenerate down col-lapsed SCBL tubes,82 but the incomplete return in fibrediameter seen may compromise subsequent conductive83
and functional24 properties. We documented regenerationboth down original SCBL tubes in the donor graft6 and inthe distal host segments (present study) in all groups, as wellas within newly formed SCBL external to the original fas-cicles, resulting in fibres smaller in diameter than seen innormal nerve. Following a delayed neurorrhaphy, SCBLtube diameter is reduced more than after an immediate re-pair.82 In the present study, the host immune response mayhave delayed regeneration across FRESH nerve allografts,further reducing the fibre diameter and subsequent conduc-tion velocity. This is consistent with our previous findingsof marked lymphocyte infiltration and Wallerian degenera-tion with no detectable regeneration in the mid-graft seg-ments at 2 and 4 weeks post grafting in FRESH allo-grafts.6,16
In the present study, endpoint (16 week) histomorphom-etry demonstrated that the total number and density of my-elinated fibres that regenerated across 3-wk and 5-wk storednerves was statistically equivalent to autografts. However,the distal myelinated axon and fibre diameters were signifi-cantly smaller distal to both 3-wk and 5-wk stored nerves;correspondingly, the maximal CNAP conduction velocitieswere significantly decreased across 3-wk (but not signifi-cantly across 5-wk) stored nerves compared to autografts.
Figure 10. Freeze-thawed nerve allograft; graft segment 16 weeksafter grafting. Numerous unmyelinated and large diameter, but thinlymyelinated axons associated with Schwann cells in a large compart-ment, which appears to have grown through a common donor basallamina and is delimited by a perineurial cell (P). Fibroblasts (F) andmacrophages surround the compartment. Uranyl lead citrate, elec-tron microscopy, ×3,780. Scale bar = 5 µm.
122 Evans et al.
There was no significant difference in the number of my-elinated fibres >7 mm in diameter in the 3-wk and 5-wkgroups compared to autografts and this, in part, may reflectthe similar CNAP amplitude recordings found in thesegroups. Therefore, not due to a lack of myelinated fibres,but rather to the reduced fibre diameters, the electrophysi-ologic function of FRESH and 3-wk (but not 5-wk) storedgrafts was inferior to autografts. Delayed regenerationacross 3-wk and 5-wk grafts due to allorecognition andrejection6 may be responsible for the smaller fibre diametersseen.
Similar to FRESH and stored nerve allografts, a delay inregeneration across FT allografts may in part account for thesmaller myelinated axon and fibre diameters seen distally.The delay in regeneration may have been a result of delayed
revascularization, the lack of neurotrophic SCs, or allograftrejection. Although there was little evidence of an earlycellular or early or late humoral response to FT allografts inprevious work,6 there was evidence of lymphocytic infil-trates at 16 weeks in the graft segments in the present study,which may indicate that these grafts are weakly immuno-genic. Darcy84 demonstrated that FT submaxillary glandallografts undergo rapid cellular infiltration by lympho-cytes, only if hosts were previously sensitized. The sensiti-zation was with either fresh or FT grafts, indicating a weak,but persistent immunogenic component to submandibulargland allografts following freeze-thawing, which was ableto induce immunologic memory in hosts.
In the present study, the delayed regeneration, smallermyelinated fibre diameters, and subsequent CNAP conduc-
Figure 11. Graft segments 16 weeks post-engraftment. Masson’s trichrome stain for collagen (green), cytoplasmic structures (red), andnuclei (blue or purple). Cross-sections of autograft nerve 16 weeks (A) weeks post-engraftment. The collagen structure appears wellmaintained following grafting, with some intrafascicular scarring noted. Cytoplasmic staining clearly indicates the presence of axons (red).Numerous cells (presumably mostly Schwann cells and fibroblasts) are contained almost entirely intrafascicular. Red blood cells can be seenin normal appearing vessels. There was a marked disruption in the epineurium (arrow, A,D) and perineurium (p) in the fresh (B) and 5-weekstored (C) allograft nerves, but somewhat less in the freeze-thawed group (D). There was heavy deposition of collagen in the intrafascicular,perineurial, and extrafascicular areas in fresh and to a lesser extent in pretreated allografts. Regeneration was often extrafascicular (EF),especially in fresh allografts. Not all areas of the nerves were filled by axons indicating non uniform regeneration. Cell infiltrates were notedin all allograft groups. Light microscopy ×540. Scale bar = 30 µm.
Banking Nerve Allografts 123
tion velocity and amplitude in FT allografts was similar toautografts and 5-wk stored allografts. Myelin thickness wasequivalent in all grafted groups, but the G-ratio (axon/fibrediameter) was significantly less in the FRESH and 3-wkgroups compared to autografts, indicating relatively in-creased myelination. However, G-ratio can be expected todecrease precipitously even in normal nerve as axon diam-eters decrease less than 2 mm as seen in all graft groups inthis study.75 It appears that other quantitative or qualitativefactors not measured in the present study play a role in thesuperior maximal CNAP conduction velocity measured inthe 5-wk and FT groups compared to the FRESH and 3-wkstored groups.
It has been argued that CNAP conduction velocity andamplitude predominantly reflect only the large (>7 mm)diameter myelinated sensory fibres and do not necessarilycorrelate with clinical recovery in patients85,86 or animalstudies.87,88 Therefore, recovery of integrated motor andsensory behaviour following sciatic nerve grafting was as-sessed by a standardized walking track technique.38–40,89,90
The differences in the SFI between groups was not consis-tently better or worse in any group, indicating the extremevariability in the measures in the present and past8 studies.
The clinical use of nerve allografts has been virtuallycondemned since the early 1970s,4 but recently with the useof cyclosporin A immunosuppression, successful regenera-tion across a fresh ten-cable 23-cm sciatic nerve allograftwas reported.13 Donor allograft harvesting, tissue typing,and transport was performed with great speed and expense
because the fate of stored grafts was uncertain. The presentstudy was designed to study the potential utility of short(UWCSS) and long-term (FT) techniques of allograft stor-age as a method of pretreatment (vs. preservation) in orderto obviate the need for host immunosuppression and itspotential side effects. Regeneration through long allograftsmay become dependent upon SC survival, which remainsnear normal for up to 1 week in storage.6,45Short-term graftstorage (<3 weeks) was an ineffective means of graft pre-treatment most likely due to the presence of viable antigeniccells (SCs). The potential usefulness of nonimmunogenicacellular (FT) allografts for repairing large peripheral nervedefects is uncertain in view of the delayed regenerationseen. The paramount dilemma in nerve allografting thatremains to be solved is the ability to independently controlthe rival roles that SCs possess in regeneration (required)and the immune response (undesirable). At present, success-ful clinical nerve allografting requires concomitant use ofhost immunosuppression or other immunotherapies.4 Usedin this manner, graft storage is feasible and can provide anample supply of allogeneic nerve tissue, for graft transportbetween medical centres and for converting urgent intoelective procedures. Further research into the mechanismsof revascularization, SC repopulation, and axonal regenera-tion through long segments of stored and FT allografts isneeded.
Figure 13. Fresh nerve allograft; distal host segment 16 weeks aftergrafting. Relatively immature myelinated and unmyelinated axonsand accompanying Schwann cells can be seen regenerating throughhost basal lamina tubes (arrow). Uranyl lead citrate, electron micros-copy, ×6,000. Scale bar = 5 µm.
Figure 12. Three-week stored nerve allograft; distal host segment16 weeks after grafting. Evidence of excellent regeneration of my-elinated and unmyelinated axons associated with Schwann cells, butthe persistence of regenerating compartments and surrounding peri-cytes is seen. Uranyl lead citrate, electron microscopy, ×3,780.Scale bar = 5 µm.
124 Evans et al.
REFERENCES
1. Mackinnon SE, Hudson AR, Falk RE, Kline D, Hunter DA. Peripheralnerve allograft: an immunological assessment of pretreatment meth-ods. Neurosurgery 1984;14:167–171.
2. Mackinnon SE, Hudson AR, Falk RE, Kline D, Hunter DA. Peripheralnerve allograft: an assessment of regeneration across pretreated nerveallografts. Neurosurgery 1984;15:690–693.
3. Evans PJ, Awerbuck DC, Mackinnon SE, Hudson AR, McKee NH.Preservation of peripheral nerve grafts. Surg Forum 1990;41:522–524.
4. Evans PJ, Midha R, Mackinnon SE. The peripheral nerve allograft: acomprehensive review of regeneration and neuroimmunology. ProgNeurobiol 1994;43:187–233.
5. Evans PJ, Awerbuck DC, Mackinnon SE, Wade JA, McKee NH. Iso-metric contractile function following nerve grafting: a study of graftstorage. Muscle Nerve 1994;17:1190–1200.
6. Evans PJ, Mackinnon SE, Levi ADO, Hunter DA, Wade JA, Nakao Y,Midha R. Cold preserved nerve allografts: changes in basement mem-brane, viability, immunogenicity and regeneration. Muscle Nerve1997.
7. Carvalho A, Evans PJ, Mackinnon SE, McKee NH. Isometric contrac-tile function of slow and fast twitch muscles distal to cold preservednerve allografts. J Reconstr Microsurg 1997.
8. Evans PJ, Mackinnon SE, Best TJ, Wade JA, Awerbuck DC, MakinoAP, Hunter DA, Midha R. Regeneration across preserved peripheralnerve grafts. Muscle Nerve 1995;18:1128–1138.
9. Mackinnon SE, Hudson AR, Falk RE, Hunter DA. The nerve allograftresponse: an experimental model in the rat. Ann Plast Surg 1985;14:334–339.
10. Mackinnon SE, Hudson AR, Bain JR, Falk RE, Hunter DA. The pe-ripheral nerve allograft: an assessment of regeneration in the immu-nosuppressed host. Plast Reconstr Surg 1987;79:436–444.
11. Bain JR, Mackinnon SE, Hudson AR, Falk RE, Falk JA, Hunter DA.The peripheral nerve allograft: a dose-response curve in the rat immu-nosuppressed with cyclosporin A. Plast Reconstr Surg 1988;82:447–455.
12. Mackinnon SE, Midha R, Bain J, Hunter D, Wade J. An assessment ofregeneration across nerve allografts in rats receiving short courses ofCyclosporin A immunosuppression. Neuroscience 1992;46:585–593.
13. Mackinnon SE, Hudson AR. Clinical application of peripheral nervetransplantation. Plast Reconstr Surg 1992;90:695–699.
14. Bain JR, Mackinnon SE, Hudson AR, Falk RE, Hunter DA. Evaluationof nerve regeneration across nerve allografts in rats immunosuppressedwith cyclosporin A. Surg Forum 1987;38:515–517.
15. Bain JR, Mackinnon SE, Hudson AR, Falk RE, Falk JA, Hunter DA.The peripheral nerve allograft: an assessment of regeneration acrossnerve allografts in rats immunosuppressed with cyclosporin A. PlastReconstr Surg 1988;82:1052–1066.
16. Midha R, Mackinnon SE, Evans PJ, Hunter DA, Becker LE. Rejectionand regeneration through peripheral nerve allografts: immunoperoxi-dase studies with laminin, S100 protein and neurofilament antisera.Restor Neurol Neurosci 1994;7:45–57.
17. Midha R, Mackinnon SE, Evans PJ, Best TJ, Wong PY. Subcutaneousinjection of oral cyclosporin A solution. Microsurgery 1992;13:92–94.
18. Midha R, Mackinnon SE, Evans PJ, Best TJ, Hare GMT, Hunter DA,Falk-Wade JA. Comparision of regeneration across nerve allograftswith temporary or continuous Cyclosporin A immunosuppression. JNeurosurg 1993;78:90–100.
19. Bain JR, Mackinnon SE, Hudson AR, Falk JA, Hunter DA, Makino A.Preliminary report on peripheral nerve allografting in the primate im-munosuppressed with cyclosporin A. Transpl Proc 1989;21:695–697.
20. Bain JR, Mackinnon SE, Hudson AR, Wade JA, Evans PJ, MakinoAP, Hunter DA. The peripheral nerve allograft in the primate immu-nosuppressed with cyclosporin A. Part 1: Histological and electro-physiological assessment. Plast Reconstr Surg 1992;90:1036–1046.
21. Fish JS, Bain JR, McKee NH, Mackinnon SE. Recovery of contractilefunction in nerve allografted, cyclosporin A-treated primates. SurgForum 1989;40:600–602.
22. Fish JS, Bain JR, McKee NH, Mackinnon SE, Hudson AR, Wade JA,Makino AP, Hunter DA. The peripheral nerve allograft in the primate
Table 1. Morphometric Assessment of the Myelinated Fibres in the Tibial Fascicle Distal to Control and Stored Nerve Allografts(Mean ± SD)
*P < 0.01 vs. all other groups, Dunnett’s.**P < 0.01 vs. all allograft groups, Dunnett’s.
***P < 0.01 vs. all other graft groups, Dunnett’s.****P < 0.05 vs. FRESH and P < 0.01 vs. 3-wk group, Dunnett’s.
Figure 14. Myelinated fibre diameters distal to nerve autografts andstored allografts. Fibre diameters were significantly decreased in allgrafted nerve groups compared to normal (N) nerve (P < 0.01, Dun-nett’s, P < 0.0001, 1W-ANOVA). Fibre diameters in the autograft(AUTO) group were significantly greater than the fresh and pre-treated allograft groups (P < 0.01, Dunnett’s, P < 0.0001, 1W-ANOVA). Groups with different letters are significantly different fromeach other. Data are expressed as mean ± SEM.
immunosuppressed with cyclosporin A. Part II. Functional evaluationof reinnervated muscle. Plast Reconstr Surg 1992;90:1047–1052.
23. Gutmann E, Sanders FK. Functional recovery following nerve graftsand other types of nerve bridge. Brain 1942;65:373–407.
24. Gutmann E, Sanders FK. Recovery of fibre numbers and diameters inthe regeneration of peripheral nerves. J Physiol (Lond) 1943;101:489–518.
25. Sanders FK, Young JZ. The degeneration and re-innervation of graftednerves. J Anat 1942;76:143–166.
26. Sanders FK. The preservation of nerve grafts. Ciba Foundation Sym-posia. In: Wolstenholme GEW, Cameron MP, editors. Preservationand transplantation of normal tissues. London, UK:J & A ChurchillLtd.; 1959. p 179–189.
27. Ide C, Osawa T, Tohkyama K. Nerve regeneration through allogeneicnerve grafts, with special reference to Schwann cell basal lamina. ProgNeurobiol 1990;34:1–38.
28. Gulati AK. Evaluation of acellular and cellular nerve grafts in repair ofrat peripheral nerve. J Neurosurg 1988;68:117–123.
29. Gulati AK, Cole GP. Nerve graft immunogenecity as a factor deter-mining axonal regeneration in the rat. J Neurosurg 1990;72:114–122.
30. Osawa T, Ide C, Tohyama K. Nerve regeneration through allogenicnerve grafts in mice. Arch Histol Jpn 1986;49:69–81.
31. Osawa T, Tohyama K, Ide C. Allogeneic nerve grafts in the rat, withspecial reference to Schwann cell basal lamina in nerve regeneration.J Neurocytol 1990;19:833–849.
32. Tohyama K, Ide C, Osawa T. Nerve regeneration through cryotreatedallogeneic grafts in the rabbit. Acta Neuropathol 1990;80:138–144.
33. Tajima K, Tohyama K, Ide C, Abe M, Morioka I. Regenerationthrough nerve allografts in the cynomolgous monkey (Macaca fascicu-laris). J Bone Joint Surg 1991;73A:172–185.
34. Bentley FH, Hill M. Nerve grafting. Br J Surg 1936;24:368–387.35. Bunnell S. Transplantation of nerves. In: Peer LA, editor. Transplan-
tation of tissues. Baltimore: Williams & Wilkins Co.; 1959. p 247–268.
36. Hall SM. The effect of inhibiting Schwann cell mitosis on the re-innervation of acellular autografts in the peripheral nervous system ofthe mouse. Neuropathol Appl Neurobiol 1986;12:401–414.
37. Nadim W, Anderson PN, Turmaine M. The role of Schwann cells andbasal lamina tubes in the regeneration of axons through long lengths offreeze-killed nerve grafts. Neuropathol Appl Neuobiol 1990;16:411–421.
38. Bain JR, Mackinnon SE, Hunter DA. Functional evaluation of com-plete sciatic, peroneal, and posterior tibial nerve lesions in the rat. PlastReconstr Surg 1989;83:129–138.
39. Hare GMT, Evans PJ, Mackinnon SE, Hunter DA, Best TJ. Walkingtrack analysis: a long-term assessment of peripheral nerve recovery.Plast Reconstr Surg 1992;89:251–258.
40. Hare GMT, Evans PJ, Mackinnon SE, Best TJ, Midha R, Szalai JP,Hunter DA. Walking track analysis: utilization of individual footprintparameters. Ann Plast Surg 1993;30:147–153.
41. Carr MM, Best TJ, Mackinnon SE, Evans PJ. Strain differences inautotomy in rats undergoing sciatic nerve transection or repair. AnnPlast Surg 1992;28:538–544.
42. Albert E. Einige operationen an nerven. Wien Med Presse 1885;26:1285–1288.
43. Hare GMT, Evans PJ, Mackinnon SE, Nakao Y, Midha R, Wade JA,Hunter DA, Hay JB. Effect of cold preservation on lymphocyte mi-gration into peripheral nerve allografts in sheep. Transplantation 1993;56:154–162.
44. Hare GMT, Evans PJ, Mackinnon SE, Wade JA, Young AJ, Hay JB.Phenotypic analysis of migrant, efferent lymphocytes after implanta-tion of cold preserved, peripheral nerve allografts. J Neuroimmunol1995;56:9–16.
45. Levi ADO, Evans PJ, Mackinnon SE, Bunge RP. Cold storage ofperipheral nerves: an in vitro assay of cell viability and function. Glia1994;10:121–131.
46. Das Gupta TK. Mechanism of rejection of peripheral nerve allografts.Surg Gyn Obstet 1967;125:1058–1068.
47. Pollard JD, Gye RS, Mcleod JG. An assessment of immunosuppres-sive agents in experimental peripheral nerve transplantation. Surg GynObstet 1971;132:839–845.
48. Levinthal R, Brown J, Rand RW. Fascicular nerve allograft evaluation.Part 1: Comparison with autografts by light microscopy. J Neurosurg1978;48:423–427.
49. Levinthal R, Brown J, Rand RW. Fascicular nerve allograft. Part 2:Comparison with whole-nerve allograft by light microscopy. J Neu-rosurg 1978;48:428–433.
50. Pollard JD, McLeod JG. Fresh and predegenerate nerve allografts andisografts in Tremble mice. Muscle Nerve 1981;4:274–281.
51. Bunge MB, Bunge RP, Kleitman N, Dean AC. Role of peripheralnerve extracellular matrix in Schwann cell function and neurite regen-eration. Dev Neurosci 1989;11:348–360.
52. Ide C, Tohkyama K, Yokota R, Nitatori T, Onodera S. Schwann cellbasal lamina and nerve regeneration. Brain Res 1983;288:61–75.
53. Wang G-Y, Hirai K-I, Shimada H. The role of laminin, a componentof Schwann cell basal lamina, in rat sciatic nerve regeneration withantiserum-treated nerve grafts. Brain Res 1992;570:116–125.
54. Rogers SL, Letourneau PC, Palm SL, McCarthy J, Furcht LT. Neuriteextension by peripheral and central nervous system neurons in re-sponse to substratum-bound fibronectin and laminin. Dev Biol 1983;98:212–220.
55. Ard MD, Bunge RP, Bunge MB. Comparision of the Schwann cellsurface and Schwann cell extracellular matrix as promoters of neuriteoutgrowth. J Neurocytol 1987;16:539–555.
56. Kleinman HK, Sephel GC, Tashiro K, Weeks BS, Burrous BA, AdlerSH, Yamada Y, Martin GA. Laminin in neuronal development. AnnNY Acad Sci 1990;580:302–310.
57. Ide C. Nerve regeneration through the basal lamina scaffold of theskeletal muscle. Neurosci Res 1984;1:379–391.
58. Glasby MA, Gschmeissner SE, Huang CL-H, de Souza BA. Degen-erated muscle grafts used for peripheral nerve repair in primates. JHand Surg 1986;11-B:347–351.
59. Glasby MA, Gilmour JA, Gschmeissner SE, Hems TEJ, Myles LM.The repair of large peripheral nerves using skeletal muscle autografts:a comparison with cable grafts in the sheep femoral nerve. Br J PlastSurg 1990;43:169–178.
60. Glasby MA. Interposed muscle grafts in nerve repair in the hand: Anexperimental basis for future use. World J Surg 1991;15:501–510.
61. Pereira JH, Bowden REM, Gattuso JM, Norris RW. Comparison ofresults of repair of digital nerves by denatured muscle grafts andend-to-end sutures. J Hand Surg 1991;16B:519–523.
62. Feneley MR, Fawcett JW, Keyes RJ. The role of Schwann cells in theregeneration of peripheral nerve axons through muscle basal laminagrafts. Exp Neurol 1991;114:275–285.
63. Jiming K, Shizhen Z, Bo S, Shengxiu Z. Experimental study of bridg-ing the peripheral nerve gap with skeletal muscle. Microsurgery 1986;7:183–189.
64. Gattuso JM, Glasby MA, Gschmeissner SE. Recovery of peripheralnerves after surgical repair with treated muscle grafts (2) Morphomet-ric assessment. Neuro Orthop 1988;6:1–6.
65. Bryan DJ, Miller RA, Costas PD, Wang K-K, Seckel BR. Immuno-cytochemistry of skeletal muscle basal lamina grafts in nerve regen-eration. Plast Reconstr Surg 1993;92:927–940.
66. Campenot RC. Local control of neurite development by nerve growthfactor. Proc Natl Acad Sci 1977;74:4516–4519.
67. Gundarsen RW. Sensory neurite growth cone guidance by substrateadsorbed nerve growth factor. J Neurosci Res 1985;13:199–212.
68. Levi-Montalcini R. The nerve growth factor 35 years later. Science1987;237:1154–1162.
69. Taniuchi M, Clark HB, Johnson EMJ. Induction of nerve growth factorreceptor in Schwann cells after axotomy. Proc Natl Acad Sci USA1986;83:4094–4098.
70. Taniuchi M, Clark HB, Schweitzer JB, Johnson EMJ. Expression ofnerve growth factor receptors by Schwann cells of axotomized periph-eral nerve: ultrastructural location, suppression by axonal contact, andbinding properties. J Neurosci 1988;8:664–681.
71. Hall SM. Regeneration in cellular and acellular autografts in the pe-ripheral nervous system. Neuropathol Appl Neurobiol 1986;12:27–46.
72. Sketelj J, Bresjanac M, Popovic M. Rapid growth of regeneratingaxons across the segments of sciatic nerve devoid of Schwann cells. JNeurosci Res 1989;24:153–162.
73. Mackinnon SE, Hudson AR, Falk RE, Bilbao J, Hunter DA. Nerve
allograft response: a quantitative immunological study. Neurosurgery1982;10:61–69.
74. Gasser HS, Grundfest H. Axon diameter in relation to the spike di-mensions and the conduction velocity in mammalian fibers. Am JPhysiol 1939;127:393–414.
75. Berry CM, Grundfest H, Hinsey JC. The electrical activity of regen-erating nerves in the cat. J Neurophysiol 1944;7:103–115.
76. Buchthal F, Rosenfalck A, Behse F. In: Sensory potentials of normaland diseased nerves. Dyck PJ, Thomas PK, Lambert EH, Bunge R,editors. Peripheral neuropathy. Toronto: W. B. Saunders Company;1984. p 980–1015.
77. Sunderland S, Bradley KC. Endoneurial tube shrinkage in the distalsegment of a severed nerve. J Comp Neurol 1950;93:411–420.
78. Giannini C, Dyck PJ. The fate of Schwann cell basement membrane inpermanently transected nerves. J Neuropathol Exp Neurol 1990;49:550–563.
79. Aberchrombie M, Johnson ML. Quantitative histology of Walleriandegeneration. I. Nuclear population in rabbit sciatic nerve. J Anat1946;80:37–50.
80. Roytta M, Salonen V. Long-term endoneurial changes after nervetransection. Acta Neuropathol 1988;76:35–45.
81. Salonen V, Peltonen J, Roytta M, Virtanen I. Laminin in traumitizedperipheral nerve: basement membrane changes during degenerationand regeneration. J Neurocytol 1987;16:713–720.
82. Holmes W, Young JZ. Nerve regeneration after immediate and de-layed suture. J Anat 1942;77:63–96.
83. Sanders FK, Young JZ. Role of the peripheral stump in the control offibre diameter in regenerating nerves. J Physiol 1944;103:119–136.
84. Darcy DA. The reaction of the rabbit to frozen homografts. J Path Bact1950;70:143–156.
85. Almquist E, Eeg-Olofsson O. Sensory-nerve-conduction velocity andtwo point discrimination in sutured nerves. J Bone Joint Surg 1970;52A:791–796.
86. Jabaley ME, Burns JE, Orcutt BS, Bryant WM. Comparision of his-tologic and functional recovery after peripheral nerve repair. J HandSurg 1976;1:119–130.
87. Dellon AL, Mackinnon SE. Selection of the appropriate parameter tomeasure neural regeneration. Ann Plast Surg 1989;23:197–202.
88. Goldberg NH, Deshpande SS, May CS, Toy FK, Smoot RT, MeshulCK, Alburquerque EX. Disparity between neurophysiologic measure-ments and clinical reality following periheral nerve transection andmicroneurorrhapy. Surg Forum 1984;35:608–610.
89. Brown CJ, Mackinnon SE, Evans PJ, Bain JR, Makino AP, HunterDA, Hare GMT. Self-evaluation of walking track measurement usinga sciatic function index. Microsurgery 1989;10:226–235.
90. Brown CJ, Evans PJ, Mackinnon SE, Bain JR, Makino AP, HunterDA, Hare G. Inter- and intraobserver realibility of walking-trackanalysis to assess sciatic nerve function in rats. Microsurgery 1991;12:76–79.
