|USFDC Home | USF Electronic Theses and Dissertations||| RSS|
This item is only available as the following downloads:
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001491309
007 cr mnu|||uuuuu
008 041102s2004 flua sbm s000|0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0000467
Rosenthal, Oren D.
Peripheral nerve repair using biomaterial nerve guides containing guidance channels
h [electronic resource] /
by Oren D. Rosenthal.
[Tampa, Fla.] :
University of South Florida,
Thesis (Ph.D.)--University of South Florida, 2004.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
Title from PDF of title page.
Document formatted into pages; contains 187 pages.
ABSTRACT: Traumatic injuries to peripheral nerves often leave gaps that cannot be repaired by direct suture methods. In such instances, repair with a tubular nerve guide, allows connection of the nerve ends, provides directional guidance, and concentrates endogenous trophic factors for regenerating axons. We hypothesized that collagen nerve guides containing longitudinally oriented channels would further improve the outcome of nerve repair by increasing the surface area available for cell migration. We restored the continuity of a 10mm peripheral nerve gap (rat sciatic nerve) by suturing the nerve stumps into a type I collagen nerve guide (1.5 mm ID), which contained longitudinal channels. Two different channel designs were tested. They were compared to empty nerve guides and autografts. One channel design contained five longitudinally-oriented collagen microtubes (0.4 mm ID) and the other contained 32 longitudinally-oriented collagen filaments (90 micro m diameter).Nerve regeneration was examined at 6 weeks and 12 weeks post repair by a determination of the number and diameter of myelinated axons in the middle sections of the nerve guides. Sciatic function Indices were calculated from walking tracks and static stance images, and electrophysiological assessments were performed. Compound muscle action potentials of the gastrocnemius and intrinsic muscles of the foot were recorded from animals in each group at 12 weeks, indicating that axons regenerated through the nerve repair site, into the distal nerve stump, and successfully reinnervated peripheral targets. At 6 weeks, there was no significant difference between the mean number of myelinated axons with the mid sections of the 3 types of nerve guides (P = 0.488). At 12 weeks, the nerve guide that contained 5 microtubes within its lumen had significantly more axons than the nerve guide that contained 32 filaments in its lumen (P = 0.008).The mean myelinated axon number in the microtube group is larger than the empty nerve guide group but this difference was not statistically significance (P < 0.05). Autografts at both 6 and 12 weeks had significantly more myelinated axons in the mid section of the repair site than either of the nerve guide repairs at the respective time points (P < 0.05).
Adviser: Liuzzi, Francis J.
t USF Electronic Theses and Dissertations.
Peripheral Nerve Repair Using Biomateria l Nerve Guides Containing Guidance Channels by Oren D. Rosenthal A dissertation submitted in partial fulfillment of the requirement s for the degree of Doctor of Philosophy Department of Anatomy College of Medicine University of South Florida Major Professor: Fran cis J. Liuzzi, Ph.D. Paula Bickford, Ph.D. Kendall F. Morris, Ph.D. Michael Nolan, Ph.D. Samuel Saporta, Ph.D. Date of Approval: August 20, 2004 Keywords: Collagen, Injury, Entubul ation, Autogra ft, Regeneration Copyright 2004, Oren D. Rosenthal
Dedication I owe the greatest gratitude to my wife Jeri. Not only was she steadfast in her personal support and belief in me, she took ca re of the most essential duties in our household, the care of our children, while I pursued the Ph.D. Without her, I would not have been able to complete this goal. She did all this, and excelled in her full time job as a physical therapis t. My daughter Emily and son Jeffrey, with their wonderful smiles, lifted my spirit s and made me realize that all the hard work would be worth it.
Acknowledgments I would like to thank my committee me mbers, Paula Bickford, Ph.D., Kendall F. Morris, Ph.D., Michael No lan, Ph.D., Samuel Saporta, Ph.D. for their interest in the dissertation project. I would espec ially like to thank my major professor, Francis J. Liuzzi, Ph.D., for accepting me into his laboratory, providing me with guidance, and giving me the confidence to complete this research. He has been a mentor and a friend. I would like to recognize and thank all members of the Department of Anatomy faculty who hav e contributed to my education at the College of Medicine. I would also like to thank Daniel Greenwald, M.D. for taking the time to serve as the chairperson of t he dissertation defense.
i Table of Contents List of Tables ........................................................................................................iv List of Figures .......................................................................................................v Abstract ..............................................................................................................viii Preface.................................................................................................................x Chapter 1: Resear ch Object ives..........................................................................1 Aims of Re search......................................................................................1 Main Objective...........................................................................................2 Specific Aims.............................................................................................3 Chapter 2: Backgroun d & Signifi cance................................................................5 Introducti on................................................................................................5 Peripheral Nerv e Inju ry..............................................................................6 Peripheral Nerv e Repa ir............................................................................9 Entubulation Ne rve Repai r.......................................................................12 Biomaterials For Nerve R epair.................................................................14 Sciatic Nerve Gr oss Anat omy..................................................................16 Microanatomy of Peri pheral Ne rve..........................................................20 Nerve Conduction In Myelinated Axons ...................................................23 Summary.................................................................................................24 Chapter 3: Materi als And Me thods ....................................................................28 Animals....................................................................................................28 Nerve Guide Fabric ation and De sign.......................................................29 Nerve Re pair............................................................................................33 Functional A ssessment ............................................................................37 Walking Tr acks..............................................................................37 Static St ance.................................................................................40 Electrophysi ology.....................................................................................41 General Electrophysiol ogy Princi ples............................................42 Animal Preparation fo r Electrophysi ology.....................................45 Method 1: Compound Nerve Action Pot entials .............................45 Method 2: Compound Muscle Action Pot entials............................47 Method 3: Motor Unit Num ber Estimation (MUNE).......................49 Method 4: Compound Muscle Action Pot entials............................52
ii Tissue Isol ation........................................................................................53 Morphometric Analysi s.............................................................................54 Chapter 4: Results.............................................................................................58 Introducti on..............................................................................................58 General Obse rvati ons..............................................................................58 Morpholog y..............................................................................................61 Nerve Regeneration in Autogr afts.................................................63 Empty Collagen Ne rve Gui des......................................................69 Nerve Guides with Longit udinal Microt ube Channel s....................74 Nerve Guides with Longitudi nal Filament s Channel s....................83 Myelinated Axon Counts and St atistical A nalyses...................................91 Nerve Guides (CT, MF MT) at 6 Weeks.......................................97 Nerve Guides Verses Aut ografts at 6 Weeks................................98 Nerve guides at 6 Weeks Ve rses Normal Control.........................98 Nerve Guides (CT, MF MT) at 12 Weeks.....................................99 Nerve Guides Verses Aut ografts at 12 Weeks............................100 Nerve guides at 12 Weeks ve rses Normal C ontrol......................100 Analysis of Groups Over Time....................................................100 Nerve Repair Versus Normal Control Across Time Periods........101 Post Hoc A nalyses ......................................................................102 Myelinated Axon Diameters ...................................................................103 Electrophysiol ogy...................................................................................109 Compound Nerve Acti on Potentia ls............................................109 Compound Muscle Action Pot entials and An atomy.....................111 Motor Unit Number Es timation (M UNE).................................................121 Functional Asse ssments........................................................................122 Sciatic Functi on Index .................................................................122 Static Stance Analysis .................................................................125 Chapter 5: Discussion .....................................................................................131 Myelinated Axon Number .......................................................................136 Variabilit y...............................................................................................138 Axon Diamete r.......................................................................................141 Functional Re covery..............................................................................143 Electrophysiol ogy...................................................................................146 Nerve Action Po tentials...............................................................147 Compound Muscle Acti on Potentia ls...........................................150 Compound Muscle Action Potent ials and Anatomy.....................151 Nerve Conduction Velocity..........................................................153 MUNE.........................................................................................155 Final Electrophysiological Protocol for the phase III rat study.....156 Chapter 6: C onclusion.....................................................................................159
iii Referenc es.......................................................................................................162 About the Aut hor.....................................................................................E nd Page
iv List of Tables Table 1 Nerve Repair Groups and Time Po ints........................................29 Table 2 Myelinated Axon Co unts Raw Data..............................................95 Table 3 Myelinated Axon Diamet ers at 6 W eeks....................................106 Table 4 Myelinated Axon Diamet ers at 12 Weeks..................................106 Table 5 Myelinated Axon Diameters Over Time CT Group.....................108 Table 6 Nerve Conduction Velocities 12 Weeks Post Repair ..................120 Table 7 Mean Sciatic Function Index From Walki ng Tracks...................123 Table 8 Mean Static Stance Index (Perce nt of Normal Toe Spread)......126 Table 9 24 Week CT Static Stance Index Axon Number, Foot CMAP...130
v List of Figures Figure 1 Diagram of Multicent er Projec t......................................................xiii Figure 2 Normal Nerve & Walleri an Degenera tion........................................7 Figure 3 Sciatic and Spinal Nerve Disse ction..............................................17 Figure 4 Sciatic Nerve Dissectio n...............................................................18 Figure 5 Normal Gross Anatomy Ra t Sciatic Nerve....................................19 Figure 6 Normal Rat Sciatic Ne rve Microanat omy......................................22 Figure 7 Normal Sciatic Nerve Hi gh Power Micr ograph..............................23 Figure 8 Three Dimensional Schematic of Collagen Nerv e Guides............31 Figure 9 Cross Section Schematic Di agram of Nerv e Gui des.....................32 Figure 10 Surgical Exposure of t he Rat Sciati c Nerve..................................34 Figure 11 Schematic Diagram of Nerve Tr ansection & Entubulation Repair.35 Figure 12 Schematic Diagram of Ent ubulation Nerv e Repai r........................35 Figure 13 Calipers for Measuri ng Sciatic Nerve............................................36 Figure 14 Walking Tracks For Scia tic Function Index...................................39 Figure 15 Enlargement Of Normal Rat Hind Paw Print.................................39 Figure 16 Image Static Stance Photography Box..........................................41 Figure 17 Hook Electrodes for Elevat ing the Sciati c Nerve...........................46 Figure 18 Electrophysiology Equipment wit h Surgical Mi croscope...............48 Figure 19 Needle El ectrode...........................................................................48 Figure 20 Cathode Stimulator Held By Mani pulator ......................................50 Figure 21 Stimulating Electrode Proxim al to Nerv e Guide............................50 Figure 22 Montage of Nerv e Guide...............................................................55 Figure 23 Axon Counting of Normal Ne rve....................................................56 Figure 24 Normal Nerve and Regenerated Ne rve In Colla gen Tube.............62 Figure 25 Suture Resorption By Giant Cells ..................................................63 Figure 26 Autograft Nerv e Repai rs................................................................64 Figure 27 Nerve Regeneration in a 3 Week Aut ograft...................................65 Figure 28 Nerve Autograft 6 W eeks Post Repair ..........................................66 Figure 29 Autograft Nerve Repair 12 W eeks Post Impl antation....................67 Figure 30 Autograft With Ax onal E scape ......................................................68 Figure 31 Low Power 6, 12, 24 Week CT Nerve Guide Sections..................70 Figure 32 Regenerated Nerve in Collagen Gu ide.........................................71 Figure 33 Minifascicles in Regenerated Nerve.............................................71 Figure 34 Regeneration in CT at 24 w eeks...................................................72 Figure 35 Failed Regeneration in CT Group.................................................73 Figure 36 Microtube With Mye linated Ax ons.................................................75 Figure 37 Regenerated Nerve with Plasma Cells inside Microtube...............76 Figure 38 Low Power MT Nerve Gui des at 6 & 12 Weeks............................77
vi Figure 39 Axons Regenerating In and Between Mi crotubes.........................78 Figure 40 MT Nerve Guide at 12 Wee ks Fasicular Mo rphology....................79 Figure 41 Variable Resorption of Microt ubes Proximal to Distal...................80 Figure 42 Microtube Vari ability ......................................................................81 Figure 43 Failed MT Nerv e Guide.................................................................82 Figure 44 Nerve Guides with Filam ents at 6 & 12 Week s.............................84 Figure 45 Nerve Guide With Filam ents.........................................................85 Figure 46 Myelinated Axons Growing Outs ide Region Of F ilaments.............86 Figure 47 Nerve Guide Image Demonstrati ng 32 Separate Filaments..........88 Figure 48 Nerve Guide Filaments And M any Axons At 6 Weeks..................88 Figure 49 Nerve Guide With Filaments and Few Axons at 6 Weeks.............89 Figure 50 Aggregated Filaments & Failed Regener ation..............................90 Figure 51 Small Number of Axons In between F ilament s.............................91 Figure 52 Nerve Repair Mont ages................................................................92 Figure 53 Normal Control Montage...............................................................92 Figure 54 Partially Counted Nerve Gu ide......................................................93 Figure 55 Number of Myelinated Axons After Post Ne rve Repai r.................94 Figure 56 Axon Diameters of No rmal Contro l..............................................103 Figure 57 Regenerated Nerve, Myelinated ve rsus Unmyelinated Axons....104 Figure 58 Histogram of Myelinated Axon Diameters at 6 Weeks................105 Figure 59 Histogram of Myelinated Axon Di ameters at 12 Weeks..............105 Figure 60 Myelinated Axon Diameters Ov er Time CT Group......................108 Figure 61 Compound Nerve Action Pot entials CT Group............................110 Figure 62 Compound Nerve Action Pot entials MT Group...........................111 Figure 63 Normal Gastrocnemius CMAP & Biceps Fe moris.......................115 Figure 64 Gastrocnemius CMAP Cut Sciatic Ne rve....................................116 Figure 65 Electrophysiology Expe riment Tw o.............................................116 Figure 66 Anatomical Study & Nerve Branc h to Biceps Femoris................117 Figure 67 Anatomical Study Rat Hin dlimb Muscula ture..............................117 Figure 68 CMAPs From Gastrocnemius & Biceps fe moris.........................118 Figure 69 Gastrocnemius CMAP and G2 Electrode Posi tion......................119 Figure 70 Shift of CMAP Latency Fr om S1 to S2........................................119 Figure 71 CMAP in 1st Metatarsal Spac e in 12 Week MT Group...............120 Figure 72 Late CMAP in Foot ......................................................................121 Figure 73 Motor Unit Number Es timation (M UNE)......................................122 Figure 74 Normal & 1 Week Wa lking Tra cks...............................................123 Figure 75 Walking Track From Rat With Flexed Toes.................................124 Figure 76 Walking Track From CT Group at 12 W eeks..............................124 Figure 77 Setting Scale For Toe Spread Measuremens With ImageJ........126 Figure 78 Typical Foot Posture 1 Week Af ter Nerve Injury & Repair..........127 Figure 79 Static Stance Photo Mo vement of Digit....................................... 127 Figure 80 Digital Flexion Abnormalit y in Operat ed Limb .............................128 Figure 81 Autograft With Return Of T oes Spread At 12 Weeks..................128 Figure 82 Toes Spread Return (MT Repai r Group at 12 Weeks)................129
vii Figure 83 SSI Plotted Agains t Axon Number.............................................130
viii Peripheral Nerve Repair Using Biomateria l Nerve Guides Containing Guidance Channels Oren D. Rosenthal ABSTRACT Traumatic injuries to peripheral nerves often leave gaps that cannot be repaired by direct suture methods. In such instances repair with a tubular nerve guide, allows connection of the nerve ends provides directional guidance, and concentrates endogenous trophic factor s for regenerating axons. We hypothesized that collagen nerve guides containing longitudinally oriented channels would further improve the outco me of nerve repair by increasing the surface area available for cell migration. We restored the continuity of a 10mm peripheral nerve gap (rat sciatic nerve) by suturing the nerve stumps into a type I collagen nerve guide (1.5 mm ID), whic h contained longitudinal channels. Two different channel designs were tested. They were compared to empty nerve guides and autografts. One channel desi gn contained five longitudinally-oriented collagen microtubes (0.4 mm ID) and t he other contained 32 longitudinallyoriented collagen filaments (90 m diameter). Nerv e regeneration was examined at 6 weeks and 12 weeks post repair by a determination of the number and diameter of myelinated axons in the mi ddle sections of the nerve guides. Sciatic
ix function Indices were calculated from wa lking tracks and static stance images, and electrophysiological asse ssments were performed. Compound muscle action potentials of the gastrocnemius and intrinsic muscles of the foot were recorded fr om animals in each group at 12 weeks, indicating that axons regenerated through th e nerve repair site, into the distal nerve stump, and successfully reinnervated per ipheral targets. At 6 weeks, there was no significant difference between t he mean number of myelinated axons with the mid sections of the 3 types of nerve guides (P = 0.488). At 12 weeks, the nerve guide that contained 5 microtubes within its lumen had significantly more axons than the nerve guide that contained 32 filaments in its lumen (P = 0.008). The mean myelinated axon number in t he microtube group is larger than the empty nerve guide group but th is difference was not statistically significance (P < 0.05). Autografts at both 6 and 12 weeks had signi ficantly more myelinated axons in the mid section of the repair site than either of the nerve guide repairs at the respective time points (P < 0.05).
x Preface The research reported herein is a component of one phase of a multiphase, multicenter study. The overall goal of this multicenter study is to design, manufacture, and test biomaterial nerve guides, which can eventually be used in human patients to repair traumatic peripheral nerve gap injuries. ShuTung Li, Ph.D. is the principle investigat or for research funded under the National Institute of Health (NIH) RO1 grant. Dr. Li is the founder and Chief Executive Officer of the biotechnology firm Collagen Matr ix Inc., which is located in Franklin Lakes, New Jersey. The research collabora tors are Francis J. Liuzzi, Ph.D. from the University of South Florida, Colle ge of Medicine, Anatomy Department, and Roger D. Madison, Ph.D. from the Nerve Regeneration Laboratory at Duke University, Medical Center and Veterans Affairs Medical Cent er, Durham, North Carolina. The project scope is briefly summari zed below. Collagen Matrix Inc. designed and fabricated seve ral different prototypes of bovine type I collagen nerve guides. To thoroughly evaluate t he effectiveness of these biomaterial nerve guides, they were tested in cell culture first, and then used to repair a sciatic nerve injury in the rat. The ner ve guides will eventually be used to repair a median nerve injury in the forearm of t he primate. The labor atory of Dr. Liuzzi
xi at the University of South Florida is responsible for in vitro (cell culture) and in vivo (rat) experiments using the nerve guides fabricated by Collagen Matrix Inc. Following the in vivo screening in the rat a final nerve guide prototype will be chosen for evaluation in pr imates at the Duke University Laboratory. The in vivo study at USF is composed of three phases. Each phase follows the same experimental design. The difference between the three phases is in the design parameters of the nerve guides. The goal of these experiments is to determine which nerve guide most favorably affects nerve regeneration when implanted to bridge a ten-millimeter nerve gap in the transected rat sciatic nerve. The first phase compared collagen tubes of two different permeabilities. The second phase evaluated collagen nerve guides that contain longitudinal channels within the nerve guide lumen. Two different designs with channels were evaluated. The third phase of the rat study evaluates collagen nerve guides that incorporate laminin, basic fibr oblast growth factor (bFGF), insulin-like growth factor II (IGFII), either alone or in combination. This final rat study will also include nerve guides that comb ine these molecules with the optimal guidance channel design chosen from the second phase of the rat study. One of these designs will be the final prototy pe used for evaluation in the primate. The dissertation research carried out by Oren D. Rosenthal, M.P.T., graduate student, in the laboratory of Fr ancis Liuzzi, Ph.D., pertains to the in vivo evaluation of nerve guides with longitudinal guidance channels (phase II). We
xii hypothesized that increasing the surface ar ea of a resorbable biomaterial nerve guide; by incorporating longitudinal guiding channels within its lumen, will improve axonal regeneration and functional recovery as compared to an empty tube. Although integrally involved in the entire in vivo study (phase I, II, & III) at the University of South Florida, the aut hor has focused the dissertation research on the effect of repairing a peripheral ner ve gap injury with a biomaterial nerve guide containing intra-luminal guidance channels. Nerve guide permeability, growth factors, and cellular adhesion molecu les, as they relate to peripheral nerve regeneration, are discussed in t he background and discussion sections to provide the reader with a brief history of peripheral nerve repair and its possible future directions.
xiii Figure 1 Diagram of Multicenter Project The dissertation research (enclosed in dotted box) is comprised of in vivo evaluation of type 1 collagen nerve guides containing longitudinal guidance channels for the repair of a 10 mm gap in the rat sciatic nerve.
1 Chapter 1: Research Objectives Aims of Research When an individual suffers a peripheral nerve transection injury, the nerve must be surgically repaired to insure optimal functional recovery (Archibald, Krarup et al. 1991; Evans 2000). When ex tensive tissue damage prevents direct union of the cut nerve ends, an autograft is often harvested from the leg to bridge the gap (Evans 2000). Most often, the su ral nerve, a large sensory nerve, serves as the graft (Evans 2001). Although this practice compromises sensation in the region of the donor site for the benefit of reestablishing motor and sensory function in a denervated region, it is just ified when the goal is reinnervation of a region critical for normal daily function, such as the hand. According to the 9th revision of the International Classifica tion of Diseases (ICD-9), patients received over 50,000 peripheral nerve surgeries in 1995. Biomaterial guide tubes can be used to bridge nerve gaps, but nerve regeneration through a biomater ial tube, to-date, has not surpassed that through an autograft (Keeley, Atagi et al. 1993; Ch amberlain, Yannas et al. 1998; Evans, Brandt et al. 1999; Evans 2001). Because of this, most clinicians choose autografts to repair peripheral nerve injuries in their patients. Although autografts currently provide the greatest functional restoration to a denervated region, the risk of sensory deficits, infection, and t he formation of a painful neuroma at the
2 autograft donor site are always present (Keeley, Atagi et al. 1993; Evans 2001). Bridging a nerve gap with a bi omaterial guide tube elimi nates all of these side effects. Consequently, there is a need to improve the efficacy of biomaterial grafts so they become a viable alter native to the autograft (Evans 2000). Although many investigators hav e added exogenous neurotrophins and improved nerve regeneration through guide tubes in animal models, few have explored the effect of diffe rent guide tube architectures (Fine, Decosterd et al. 2002; Xu, Yee et al. 2003). In addition, many of the neur otrophin delivery methods are impractical and have not been tested in humans (Evans 2000; Lundborg 2000; Lundborg 2002) We believe that the la ck of focus on guide tube architecture represented a significant gap in the literature t hat we addressed in this study. Only recently have researc hers noticed the effect that increased surface area of guide tubes has on axon regeneration (Arai, Lundborg et al. 2000; Hadlock, Sundback et al. 2001). We now possess the technology to incorporate these concepts into the design and fabrication of a biomaterial nerve guide. Main Objective We examined axonal r egeneration and functional ner ve recovery following repair of a 10 mm peripheral nerve gap wit h bovine type I collagen nerve guides that contain longitudinally oriented collagen guidance channels within their lumen. We collected data (outlined in the specific aims) at 6 and 12 weeks post
3 nerve transection/repair to analyze and compare nerve regeneration in nerve guides with guidance channels, nerve gui des without guidance channels, and autografts. We hypothesized that increas ing the surface ar ea of a resorbable biomaterial nerve guide, by incorporati ng longitudinal guidi ng channels within its lumen, will improve axonal r egeneration and functional recovery as compared to an empty tube. Specific Aims The overall goal of the this study was to compare the efficacy of nerve guides with guidance channels of two differ ent designs in the repair of a 10 mm gap in the rat sciatic nerve. One contai ned longitudinally oriented microtubes within its lumen; the other c ontained longitudinally oriented filaments. They were compared to both nerve guides without l ongitudinal channels and to the gold standard, autograft repair. Aim 1: To quantify and compare by morphometri c analyses, axonal regeneration through the different nerve repairs by measuring mean myelinated axonal diameter and by counting the number of myelinated axons in the middle segment of the nerve guide tube or autograft.
4 Aim 2: To examine by electrophysiological anal yses, the effect of each type of nerve repair on facilitating axon elongati on into the distal nerve stumps and reestablishing synaptic connec tions with denervated muscle. We analyzed the evoked compound muscle ac tion potential (CMAP) in the gastrocnemius (a muscle that is distal to the repair site and is normally innervated by the tibial br anch of the sciatic nerve). Aim 3: To examine by behavior analysis, the ef fect of each type of nerve repair on the gross motor function of the rei nnervated limb by using walking track and static stance footprint analyses. Th is included calculating a sciatic function index (SFI) and a static sciati c index (SSI) to compare functional recovery of the sciatic nerve followi ng repair with different nerve guides.
5 Chapter 2: Background & Significance Introduction When a peripheral nerve is severed dur ing trauma, the victim suffers loss of motor and sensory function distal to the injury (Sunderland 1991). Although axons will elongate from t he proximal nerve stump fo llowing nerve transection injury, the likelihood that a majority of the axons will successfully find and enter the distal nerve stump, and subsequently re innervate appropriate target tissues is small (Sunderland 1991). Therefore, su rgical intervention is necessary to reunite the cut ends of the nerve (Evans 2000; Evans 2001). When this is not possible because the trauma has created a gap, a nerve aut ograft is harvested to bridge the nerve gap (Evans 2001). Fr equently the sural nerve, a pure sensory nerve from the lower extr emity, is chosen for the repair (Evans 2000; Evans 2001). An acceptable trade off exists, wher e the patient is wil ling to sacrifice sensation in a region of the lower leg and foot for a chance of functional recovery in a more important region such as the hand. A major disadvantage of the autograft repair method is the limited am ount of tissue available to use for grafting. Another problem is that of size mismatch, where the donor autograft is significantly smaller in diameter than the nerve that requires repair.
6 Peripheral Nerve Injury Following a peripheral nerve crush or transection injury, axons and myelin distal to the injury fragment and degenerate (see figure 2), (Waller 1850; Lubinska 1977; Stoll, Griffin et al. 1989). Waller (1850) first described degenerative changes in transected di stal nerve segments of frog glossopharyngeal and hypoglossal nerve s (Waller 1850). These changes are now commonly referred to as Wallerian deg eneration. It is now known that calcium-dependent proteases, at least in part, are responsible for degeneration of the axonal cytoskeleton (George, Gla ss et al. 1995). Schwann cells in the proximal and distal nerv e stumps of a transected nerve proliferate (Ramon Y Cajal 1928). Schwann cells and macrophages phagocytose myelin and degenerated axon fragments of the distal nerve segm ent (Ramon Y Cajal 1928; Stoll, Griffin et al. 1989). The same pheno menon also occurs for a short distance in the proximal nerve stum p. This is an essential step as there are myelin proteins that are inhibitory to axonal regeneration both in t he peripheral nervous system (PNS), and in the central nerv ous system (CNS) (Grados-Munro and Fournier 2003; Schwab 2004)
7 Figure 2 Normal Nerve & Wallerian Degeneration Photomicrographs (100x oil objective lens) of to luidine blue stained 2 m plastic sections. Normal sciatic nerve (A) shows large myelinat ed axons (solid white arrows), and nuclei of Schwann cells (open white arrow). Many small unmyelinated axons are visible between the myelinated axons. Image (B) is from a nerve segm ent distal to an autograft repair at 3 weeks showing Wallerian degeneration (solid white ar rows), nuclei of macrophages/Schwann cells (open black arrow) phagocytosing whirls of my elin debris, mast cell in center of image, empty/collapsed basal lamina endoneurial tubes alread y cleared of myelin (solid black arrows). These are spaces formerly occupied by myelinat ed axons. There is no evidence of regeneration in this section of nerve, no myelinated or unmye linated axons have entered this region of the sciatic nerve at this time.
8 One difference between the PNS and CNS, regarding axonal regeneration, is that myelin is efficiently removed from the lesion in the PNS by the many macrophages that ent er the injury site from the blood (Stoll, Griffin et al. 1989; Stoll, Trapp et al. 1989). In the CNS, the influx of blood-derived phagocytotic cells that can clear myelin from degenerating axons is far less and myelin remains in the CNS for months after injury. In the PNS, mast cells in the region surrounding the injury degranulate. They release histamine, which dramatically increases vascular permeab ility and facilitates entry of monocytederived macrophages to the injury site (E sposito, De Santis et al. 2002). Additionally, mast cell numbers dramatic ally increase in the injury site and contribute to nerve regeneration by rel easing nerve growth factor (NGF), and vascular endothelial growth factor (VEGF) (Zochodne and Cheng 2000; Norrby 2002; Rosenstein and Krum 2004). Mast cells and Schwann cells both release tumor necrosis factor alpha (TNF), which causes monocytes to increase expression of adhesion molecule s, allowing them to form attachments to vessel endothelium and to extravasate to the injury site. After debris has been cleared from a per ipheral nerve lesion, capillaries and endoneurial tubes lined with Schwann cells remain in the distal nerve stump (Ramon Y Cajal 1928). Schwann cells from both the distal and proximal stumps proliferate and migrate across the lesion. These cells facilitate axon elongation across the lesion by secreting neurotrophi c factors, which are chemoattractants for axonal growth cones, and by depositing ex tracellular matrix molecules, which
9 offer contact guidance for growth cone f ilopodia (Rogers, Letourneau et al. 1983; Fornaro, Tos et al. 2001; Goldberg 2003). Eventually, Schwann cells invest regenerating axons and elaborate new my elin (Ramon Y Cajal 1928). Functional reinnervation of target tiss ues is more complete following nerve crush injuries than after transection injuries (Hare, Evans et al. 1992). In crush injury the connective tissues of the nerve may remain intact and provide channels that guide axon elongation to per ipheral targets (Ide, Tohyama et al. 1983; Ramon Y Cajal 1928). Although no ne rve graft or guidetube can direct axons precisely to their former endoneuria l tubes, the addition of longitudinal channels in a biomaterial nerve guide dram atically increases the surface area for axon growth cone adhesion during el ongation (Ngo, Waggoner et al. 2003). These channels may also provide improv ed directional guidance when compared to an empty nerve guide. We used type I collagen guide tubes that have type I collagen longitudinal channels in the current study. Peripheral Nerve Repair Direct suture of severed nerve offers the best functional outcome as long as the final repair is tension free at the suture line (Terzis, Faibisoff et al. 1975; Evans 2000; Evans 2001). Immobilizing join ts adjacent to the repair, in either excessive flexion or extension, to mi nimize tension to a repaired nerve, is counterproductive since nerve tension will be greater once movement is reinitiated. During the period of imm obility there would likely be fibrotic
10 adherence of the repair site to the underly ing structures and thus normal nerve gliding with joint movem ent would be restricted to a greater degree than would occur if movement was allowed (Sunderland 1991). Prolonged joint immobilization surrounding denervated limb regions may also lead to articular, periarticular fibrosis, and adhesion forma tion, which limits functional recovery even if regenerating axons reach their di stal targets (Sunder land 1991). In addition to the increased risk of tension, a possible complication of the direct suture method is the failure of axons to penetrate the scar that forms at the suture line (Sunderland 1991). Using a nerve graft or guide tube to bridge a large gap can prevent excessive tension on the repair and improve the functional outcome (Terzis, Faibiso ff et al. 1975). Using autologous tissues to repair lar ge nerve gaps (vein, muscle, tendon, artery, or nerve) affords the benefit of immunological acceptance, but unfortunately, requires the sa crifice of healthy donor tissues (Brandt, Dahlin et al. 1999; Battiston, Tos et al. 2000; Geuna, Tos et al. 2000 ; Fornaro, Tos et al. 2001; Rodrigues Ade and Silva 2001). Cadav eric nerve allografts eliminate the side effects associated with surgical har vesting of nerve aut ografts, but outcomes are not as good (Trumble and Shon 2000; Ev ans 2001). Allografts may also require immunosuppression, which cannot be justified for nonlife threatening peripheral nerve repair (Trumble and Shon 2000; Evans 2001). This has lead to the design and fabrication of synthetic and biomaterial nerve guides for the repair of transection injuries with large gaps (Madison, da Silva et al. 1987; Keeley,
11 Nguyen et al. 1991; Chamberlain, Yannas et al. 1998; Hadlock, Sundback et al. 2001; Young, Wiberg et al. 2002; Ahm ed, Underwood et al. 2003; Ngo, Waggoner et al. 2003). Surgical nerve repair using guide tubes to join proximal and distal ends of a transected nerve provides directiona l guidance for Schwann cells and axon growth cones. In addition, it may c oncentrate essential trophic factors for elongating axons. This technique is frequent ly referred to as entubulation repair (Madison, da Silva et al. 1987; Madison, Da Silva et al. 1988; Keeley, Atagi et al. 1993). Additional variables that may e ffect regeneration thro ugh guide tubes are permeability, bioresorption, immunoreac tivity, kink-resistance, presence of growth factors, vascularization, ce ll adhesive properties, and presence of Schwann cells (Evans, Brandt et al. 1999; For naro, Tos et al. 2001). With this in mind, a myriad of nerve grafting and guide tube technologies have been explored (Godard, Coulon et al. 1984; Glasby, G schmeissner et al. 1986; Archibald, Krarup et al. 1991; Keeley, Nguyen et al. 1991; Evans, Brandt et al. 1999; Hadlock, Sundback et al. 2001; Hashimoto, Suzuki et al. 2002; Young, Wiberg et al. 2002; Ahmed, Underwood et al. 2003; Ngo, Waggoner et al. 2003). What follows is a brief review of the relevant research of entubulation nerve repair.
12 Entubulation Nerve Repair Bridging a 4 mm nerve gap with a colla gen nerve guide restores nerve function as effectively as an autogr aft. This was demonstrated by electrophysiology of target muscle and sensory nerves in rats and Macaca fascicularis monkeys (Archibald, Krarup et al. 1991). Nerve gaps of 10 mm or larger, however, provide a greater cha llenge to regenerating axons in a guide tube and this challenge is a driving forc e for the multitude of designs and tissues employed as nerve guides. Ceballos (1999) demonstrated that a collagen tube filled with collagen fibrils suspended in a gel and aligned longitudinally by a magnetic field, improved peripheral nerve regeneration across a si x mm gap in the mouse (Ceballos, Navarro et al. 1999). However, when t he collagen was cross-linked with ribose to improve stability, nerve regeneration was inhibited. Consequently, the degree of cross-linking and the specific cross-lin king agent are importa nt variables that affect nerve regeneration th rough fabricated gui de tubes (Itoh, Takakuda et al. 2002). In another study, parallel longitudinally oriented fi brils of collagen matrix inside a silicone tube facilitated axon regeneration be tter than a random orientation of collagen matrix, as evi denced by an earlier and more complete target reinnervation (Verdu, Labrador et al. 2002). Aria (2000) was able to successfully bridge a 15 mm gap in the rat sciatic nerve by including 7 resorbable synthetic filaments inside a silicone tube (Arai, Lundborg et al. 2000). He
13 repeated this with 5 different filament materials and all supported regeneration better than empty tubes. This result is significant because it is widely reported that repair of rat sciatic nerve gaps gr eater than 10 mm with empty tubes, fail. Nerve guide tubes packed with poly(Llactide) microfilaments facilitated regeneration in the rat sciat ic nerve at 10 weeks following entubulation repair better than empty tubes (Ngo, Waggoner et al. 2003). However, these guide tubes were silicone and therefore not reso rbable, unlike the collagen tubes in the current study (Chamberlain, Yannas et al. 1998) This is an important distinction. When silicone tubes are used for nerve guides, they eventually cause constriction of the tissue cable, which can lead to ischemia and failure of functional reinnervation of target tiss ues (Chamberlain, Yannas et al. 1998). Nerve repair with silicone tubes requires a second surgery to remove the tubes after axons have bridged the nerve gap. In a long term study of entubulation repair of the rat sciatic nerve, collagen tubes filled with a collagenglycosaminoglycan (collagenGAG) matrix, which was designed to degrade in six weeks, had more myelinated axons gro wing through them t han empty tubes did at both the 30 and 60-week time points (Chamberlain, Yannas et al. 1998). Silicone tubes filled with GAG-collagen matrix tubes also proved beneficial for axon regeneration when compared to em pty tubes (Chamberlain, Yannas et al. 1998). Although a collagen tube filled with collagen-GAG matrix provides a porous and resorbable substrate for ax on regeneration (Chamberlain, Yannas et
14 al. 1998), this intraluminal material is unl ikely to decrease the risk of nerve guide collapse. Biomaterials For Nerve Repair High tensile strength is a characteristic of type I collagen that makes it an ideal biomaterial for the fabrication of nerve guide tubes (D aamen, van Moerkerk et al. 2003). Collagen, found in the extrac ellular matrix of many tissues, is composed of a parallel arrangement of collagen fibers (Boot-Handford and Tuckwell 2003). These fibers are composed of fibrils, which in turn are composed of a triple helix of polypeptides (t ropocollagen) (Boot-H andford and Tuckwell 2003). Tropocollagen molecules are cova lently cross-linked to one another in the extracellular environment, giving colla gen fibrils their high tensile strength (Daamen, van Moerkerk et al. 2003). When purified collagen fibers are formed into tubes in the laboratory, they are chemically cross-linked to increase stability and tensile strength (Berglund, Mohseni et al. 2003; Charulatha and Ra jaram 2003). The improved strength of the collagen tubes allows suture fixati on during nerve repair. Varying the degree of cross-linking not only affects the mechani cal properties of t he formed tube, but also alters in vivo degradation (Charulatha and Rajaram 2003). Tubes with a higher degree of cross-linking resist degradation (Charulatha and Rajaram 2003). The degree of chemical cro ss-linking and the wall thickne ss of the guide tube are
15 parameters that were tightly controlled during the f abrication of collagen tubes for this study. A major advantage of collagen over synt hetic materials for fabrication of nerve guides is biocompatibil ity. Collagen is an integral structural protein of the extracellular matrix, and is found in all tissues (Boot-Handford and Tuckwell 2003). When bovine type I collagen vascular stents were implanted in rabbits it proved biocompatible (Cloft, Kallmes et al. 2000). In another study, bovine collagen sponges that were implanted subc utaneously in rats for three months did not elicit antibody production to fore ign material (Ansel me, Bacques et al. 1990). In addition to the support provided by animal studies, type I bovine collagen has been safely used in humans to improve healing of leg ulcers and pressure sores (Beghe, Menicagli et al. 1992). Since type I collagen protein is highly conserved among vertebrate specie s, immune rejection of purified bovine collagen tissue grafts implanted in humans is unlikely (Boot-Handford and Tuckwell 2003). Moreover, macrophages po ssess the necessary mechanisms to phagocytose and enzymatically digest colla gen. Consequently, resorption and remodeling of implanted collagen nerve guides can take place with minimal inflammation. By contrast, synthetic nerve guides made from poly L-lactide, and those made from Poly-hydroxybutyric acid (P HB), demonstrate delayed resorption and significant inflammation, as evidenced by the presence of giant cells surrounding
16 residual material at 24 weeks post impl antation (Borkenhagen, Stoll et al. 1998; Ngo, Waggoner et al. 2003). When silicone t ubes are used as nerve guides they eventually constrict the regenerating axons and myofibroblasts take residence in a continuous layer around the tubes (Cham berlain, Yannas et al. 1998). Clearly, collagen is a superior material for implant ation in human pati ents. It is readily available from bovine tendon and is an exce llent biomaterial for fabrication of nerve guide tubes with guidance channel s (Archibald, Krarup et al. 1991). Sciatic Nerve Gross Anatomy For the rat sciatic nerve to serve as a model for peripheral nerve injury and repair, a thorough understand ing of its anatomy is es sential. Greene gives a comprehensive anatomical report of the rat hind limb anatomy in his 1968 text Anatomy of the Rat The cell bodies of motor neur ons with axons that travel through the sciatic nerve and innervate skele tal muscle of the lower limb reside in the ventral horn of the spinal cord. The cell bodies of sensory neurons with axons that travel through the sciatic nerve reside in the dorsal root ganglia, which lie just outside of the spinal cord. They send central axons into the dorsal horn of the spinal cord and peripheral axons into sp inal nerves. Segmental spinal nerves from the caudal spinal co rd divide into dorsal and ventral primary rami, which form the lumbosacral plexus. In addition to the motor and sensory nerve fibers, autonomic axons that innervate blood vessels and skin travel in the sciatic nerve.
17 In an anatomical study of twenty f our adult Sprague-Dawley rats, the L4 & L5 spinal nerves joined to form the sciat ic nerve in all animals studied (Asato, Butler et al. 2000). The L6 spinal nerve joined the sciat ic nerve in 54% of the rats, while in the rest of the animals t he L6 spinal nerve ran along with the sciatic nerve but never merged with it. The L3 sp inal nerve joined the sciatic nerve in 25% of the rats studied. It is presumed that the largest contribution to the sciatic nerve in the Lewis rat (strain used in t he present study) is via the L4 & L5 spinal nerves (figure 3). Figure 3 Sciatic and Spinal Nerve Dissection Dissection of a Lewis rat following formalin fixation. The left laminae of the lumbar vertebra removed to expose the L3-L6 dorsal root ganglia a nd spinal nerves. A small slip of nervous tissue (black arrow farthest to the left) connects the L3 spinal nerve to the L4 spinal nerve. Solid black arrows from left to right point to the L3, L4, L5, and L6 spinal nerves respectively that all appear to contribute to the sciatic nerv e (white arrow) in this specimen.
18 The most proximal location of the sciatic nerve is in the pelvis. It travels through the sciatic notch and enters the glut eal region. The sciatic nerve has two main divisions that are held together by connective tissue ensheathment. These divisions, the tibial and peroneal nerve s, can be divided by dissection but normally appear as one nerve in the prox imal to mid thigh. The tibial and peroneal nerves are the terminal branches of the sciatic nerve and are named the same: common peroneal nerve and the ti bial nerve (Greene 1968). There are many muscular and sensory branches giv en off by the sciatic nerve along its course. These branches will not be reviewed here. Figure 4 Sciatic Nerve Dissection The sciatic nerve (solid white arrow) after it emerges from the sciatic notch continues through gluteal region where it gives off a muscular bran ch (open black arrow) to the biceps femoris from its tibial division. The sciatic nerve continues distally though the thigh (open white arrow). The tibial nerve (solid arrowhead) travels between t he lateral (black arrow) and medial gastrocnemius muscle heads.
19 Figure 5 Normal Gross Anatomy Rat Sciatic Nerve (A) High magnification view of the sciatic ner ve from the boxed region of image (B), which was captured during the terminal surgical exposure of the normal sciatic nerve. The sciatic nerve emerges in the top half of the image in the gluteal region and gives rise to a large muscular branch (open arrow) that innervates the biceps femo ris. The sciatic nerve (solid arrow) continues down the thigh where the tibial and common peron eal divisions of the sciatic nerve can be seen in the thigh through the translucent connec tive tissue that ensheathes both nerves. The common peroneal and tibial nerves di verge at the popliteal fossa. The tibial nerve remains deep and medi al, while the common peroneal takes a more lateral course in the lower leg. The tibial nerve travels between the two heads of the gastrocnemius muscle where it gives off three muscular branches. The first of these innervates the latera l head of the gastrocnemius, the soleus, and the plantaris muscles. The second branch of the tibial nerve, in this region, innervates the medial gastrocnemius. T he last of the three muscular branches
20 runs to the deep posterior compartment and innervates all of the deep muscles of the posterior lower leg: t he flexor hallucis longus, the tibialis posterior, and the flexor digitorum longus muscles The tibial nerve continues distally until it divides into the medial and lateral plantar nerves, which innervate the intrinsic muscles of the foot and sensory rec eptors in the digits. The common peroneal nerve divides in to deep and superficial branches. The superficial peroneal nerve innervate s the peroneal muscles, which are lateral to the fibula. The deep peroneal nerve inner vates the tibialis anterior muscle and the digital extensors. Its terminal br anches supply the dorsum of the foot and end in the second interdigital space (Greene 1968). All of the above muscles are denerva ted when the sciatic nerve is transected in the thigh. They do not recover function unless axons from the proximal sciatic nerve stump rege nerate and reach the formerly denervated neuromuscular junctions. Microanatomy of Peripheral Nerve Individual axons of peripheral ner ves are surrounded by a connective tissue termed the endoneurium, which contai ns small diameter collagen fibers (Ross, Romrell et al. 1995). The peri neurium surrounds bundles of axons, and delineates fascicles (Ross, Romrell et al 1995). Peripheral nerves contain one or more fascicles, which are contai ned within a connective tissue termed the epineurium (Ross, Romrell et al. 1995). Of the three connective tissues, the
21 epineurium contains the largest diameter collagen fibers. Mast cells are found in the spaces bound by the epineurium and perineurium, and can be easily identified by their large cytopl asmic metachroma tic granules. Schwann cells, from neural crest, are predominant cell type in the endoneurial compartment, and are responsib le for ensheathing axons of the PNS. Although a single Schwann cell can envelope many small diameter unmyelinated axons, a single Schwann ce ll can only wrap a myelin sheath around a single large diameter axon. When Schwann cells wrap their plasma membranes successively around large ax ons, most of the Schwann cell cytoplasm is squeezed from the concentric layers, leaving a sphingomyelin rich sheath termed the myelin sheath or the sheath of Schwann. The larger diameter axons receive more myelin wrappings, leaving them with a thicker myelin sheath. A myelin sheath from a single Schwann cell may occupy approximately 1 mm of the axons lengt h. The axonal area between two successive myelin sheaths is called a node of Ranvier (Ross, Romrell et al. 1995).
22 Figure 6 Normal Rat Sciatic Nerve Microanatomy (A) Image (10x objective lens) of 2 micrometer cr oss section of normal rat sciatic nerve stained with toluidine blue. Nearly the ent ire cross sectional area of the sciatic nerve is shown. The dense collagen epineurium encircles the nerve and a circular cross section is noted. A large venule can easily be distinguished in (B) at (40x objective). A B
23 Figure 7 Normal Sciatic Nerve High Power Micrograph Many small unmyelinated axons are visible betw een the larger myelinated axons (solid black arrows). The thick dark blue border that encircl es these large axons is the myelin sheath. It appears as one thick layer, multiple layers of t he myelin sheath can be resolved with the electron microscope. In the center of these large axons, very small dark staining bodies are discernable. From studies using electron microscopy, thes e are known to be mitochondria. Schwann cell nuclei (white arrows) can be identified in close proximity to the axonal myelin sheath. Some axons appear to have a double ring of myelin, this is a region of Schwann cell cytoplasm either perinodal or at a Schmidt-Lanterman cleft. In the upper right corner of the image, collagen bundles of the epineurium are visible along with a perineurial cell nucleus (open black arrow). Nerve Conduction In Myelinated Axons Voltage-gated sodium channels are hi ghly concentrated at the nodes of Ranvier. When they reach threshold pot ential, these channels open and sodium rapidly flows into the axon down an elec trochemical gradient. This depolarization creates the voltage differential that forces the current to fl ow, and propagates the
24 nerve action potential along the axon (Oh 1993). Voltage gated potassium channels allow potassium to flow out of the axon down a electrochemical gradient hyperpolarizing the axonal memb rane. The sodium potassium pump and potassium leak channels are responsible for establishing the resting membrane potential of -75mV. Mye lin insulates axons between nodes, and permits current to flow easily from node to node with out signal loss. This saltatory nerve conduction greatly increases conduction ve locity (Daube 1996). Another factor influencing conduction velocity is axon diameter; larger axons have less axoplasmic electrical resistance so they are faster conducting (Daube 1996). The largest axons with the thickest my elin sheath are therefore the fastest conducting axons. Thus, the number of myelinated axons and the size distribution of myelinated axons in t he distal stump of a regenerating nerve provide 2 indices by which to evaluate nerve regeneration (Bain, Mackinnon et al. 1988; Keeley, Atagi et al. 1993; Chamberla in, Yannas et al. 1998; Evans, Brandt et al. 1999; Hashimoto, Suzuki et al. 2002). The number of myelinated axons in the nerve guide tube was assessed in this study. Summary In the future, nerve repair shoul d combine the concepts that have successfully improved nerve regenerati on, such as those containing longitudinally oriented fibers, constructed of a biomaterial that is both resorbable over time and kink-resistant for the first several wee ks post-implantation. Type I
25 collagen tubes packed with longitudinally oriented collagen mi crofilaments or packed with longitudinally oriented collagen microtubes, were tested in this study. We hypothesized this architecture would optimize the surface area available for cell migration and axon elongation, and that it would prevent collapse of the guide tube during the early phas e of regeneration. Both of these factors should improve axon elongation and navigation into the distal nerve stump when incorporated in nerve guides us ed to repair gap injuries. Numerous investigators have dem onstrated that adding exogenous neurotrophins improves nerve regeneration through guide tubes (Utley, Lewin et al. 1996; Santos, Rodrigo et al. 1998; Yin, Kemp et al. 2001; Fansa, Schneider et al. 2002; Fine, Decosterd et al. 2002). T here are, however, questions as to which single neurotrophin or combinations of neurotrophins provide optimal nerve regeneration. Additionally, the opt imal delivery method of exogenous neurotrophins has not been determined (Evans 2000; Lundborg 2000; Lundborg 2002). Another technology recently employed to improve axon regeneration across long nerve gaps is the use of cellu lar grafts (Fansa, Schneider et al. 2001; Hadlock, Sundback et al. 2001; Mosahebi Fuller et al. 2002). Although Schwann cell-laden grafts may aid regenerati on by providing trophic support and extracellular matrix molecules, there is currently no practical way to deliver this technology in a clinical setting (Mosahebi, Fuller et al. 2002). What is needed is
26 a ready off the shelf nerve guide that pr ovides the clinician and patient a viable alternative to the autograft. Our goal in this study was to optim ize nerve guide architecture for entubulation repair before we proceeded with the inclusion of trophic factors or adhesion molecules in our design. We believe that once nerve guide architecture is optimized it will provide a better platform from which to study the benefits of exogenous growth factors on nerve regeneration. We used histological, electrophysiol ogical, and functional assessments to provide evidence of nerve regeneration in this study. It is imperative to understand how regenerating axons inte ract with collagen guidance channels and to determine if more axons grow through nerve guides with guidance channels than grow through empty nerve guides. We designed histological assessments to answer these questions. Many investigators studying nerve regeneration through nerve guides have successfully used axon counting and axon diameter measurements to assess nerve regeneration (Bain, Mackinnon et al. 1988; Keeley, Atagi et al. 1993; Cham berlain, Yannas et al. 1998; Evans, Brandt et al. 1999; Hashimoto, Suzuki et al. 2002). If axons grow through the nerve guide, but do not enter the distal nerve stump and eventually reinnervate muscle, it cannot be considered a successful nerve repair. Therefore, electr ophysiological assessments, including measurement of latency and amplitude of evoked compound muscle action potentials (CMAPs), and calculations of conduction velocity, are frequently used
27 to assess nerve regeneration following nerve repair. Such methods can confirm muscle reinnervation (Bain, Mackinnon et al. 1988; Archibald, Krarup et al. 1991; Keeley, Atagi et al. 1993; Chamberlain, Yannas et al. 1998; Valero-Cabre, Tsironis et al. 2001). Additional elec trophysiological methods provide an estimate of the number of motor units in a muscle following nerve injury (1996; Krarup, Archibald et al. 2002). We em ployed these methods in the present study. It must be noted that regener ating axons sprout multiple growth cones and frequently form simultaneous synaptic c onnections with agoni st and antagonist muscles (Valero-Cabre, Tsironis et al. 2001). Functional reorganization of newly re-afferented cortical and cerebellar r egions challenge motor control (Lundborg 2000). The need to assess motor function as a measure of nerve regeneration cannot be overstated, as it is the ultimate goal of nerve repair to restore function. This study incorporated two sciatic func tion indices that have been used by many investigators who have studied muscle function recovery following nerve transection and entubulation repair (de Medinaceli, Freed et al. 1982; Bain, Mackinnon et al. 1988; Bain, Mackinnon et al. 1989; Hare, Evans et al. 1992; Keeley, Atagi et al. 1993; Evans, Brandt et al. 1999; Evans, Brandt et al. 2000).
28 Chapter 3: Materials And Methods Animals Ninety 250g male Lewis rats were used in this study. The Lewis rat was chosen because this strain displays autophagia of the denervated limb less frequently than other rat species (Hare, Ev ans et al. 1992; Chamberlain, Yannas et al. 1998). The number of animals chosen is based on the minimum number necessary to apply statistical analysis that will discriminate between groups. The animals were housed in the fully accredited animal facility in the University of South Florida Division Comparative Medi cine. There are four fulltime licensed veterinarians in the Division of Comparat ive Medicine to offer any veterinary care if the need had arisen. Animals were hous ed in pairs, and provided with food and water ad libitum. They were on a 12-hour light/dark cycle. Rats were randomly assigned to each group. Twenty rats were designated to receive autograft nerve repair (AG), twenty rats were designated to receive empty collagen nerve guides (CT), twenty rats we re designated to ra ts received nerve guides with filaments inside (MF), and tw enty rats were designated to receive nerve guides that contained microtubes in side (MT). For each type of nerve repair (AG, CT, MF, MT) there were two time points: 6 weeks and 12 weeks.
29 The rats in each repair group were divide d equally to each time point so there were initially 10 rats per group at each time point. During the course of the study five additional rats were added, 2 rats were added to the 12 week MT group and 3 rats were added to the MF group. The CT group also contained a 24-week time point with 5 animals. Table 1 summarize s the number of rats in each nerve repair group and each time point. Table 1 Nerve Repair Groups and Time Points Nerve Repair Type Survival TimeNumber of Rats per group Autograft (AG) 6 weeks 10 Empty Collagen Tube (CT) 6 weeks 10 Collagen tube with filaments (MF) 6 weeks 10 Collagen tube with microtubes (MT)6 weeks 10 Autograft (AG) 12 weeks 10 Empty Collagen Tube (CT) 12 weeks 10 Collagen tube with microtubes (MT)12 weeks 12 Collagen tube with filament (MF) 12 weeks 13 Empty Collagen Tube (CT) 24 weeks 5 Total Number of Animals 90 Nerve Guide Fabrication and Design All type I collagen used in this st udy was derived from bovine tendon collagen. Collagen Matrix Inc. was respons ible for the purification of bovine type I collagen and fabrication of the nerve gui des according to the following methods and design specifications. A fixed dry weig ht (5.0 % of the collagen dry weight) of heparin (Diosynth, Chicago, IL) wa s added to a collagen and lactic acid dispersion. The collagen and heparin dispersion was homogenized and the solution was brought to the isoelectric point of collagen (pH 5.0). This coacervated (quantitatively co-precipitat ed) the collagen-heparin dispersion. All
30 components of the nerve guides were f abricated with this type I collagen and heparin composite material. The heparin mo lecule was incorporated into the collagen in order to bind and provide a sl ow release for both exogenous bFGF and laminin for the Phase III rat sciatic nerve study. Heparin was used in this study to avoid it being a confounding vari able when attempting to systematically determine the effect of bFGF and laminin in phase III. All components of the nerve guides in this study (phase II) were fabricated with the collagen-heparin composite. Hereafter, t he collagen-heparin composite is simply referred to as collagen. The coacervated fibers were s pun onto a rotating mandrel and then freeze-dried. The outer collagen tube which contains the guidance channels within it, has a diameter of 1.5 mm and 0.1 mm wall thickness. Bovine type I collagen filaments of 50-100 m diameters were produced by adjusting the extrusion nozzle. Microtubes with an i nner diameter of 0.2-0.6 mm were produced in the same fashion as the out er collagen tubes. All outer collagen tubes and microtubes were mechanically cr imped, leaving longitudinal folds to improve resistance to kinking. The t ubes were cross-linked with formaldehyde gas to improve tensile strengt h and to control the rate of in vivo degradation. They were targeted to degrade within 3 months. The collagen filaments and guide tubes are cross linked in the same manner as the outer tubes. Prior studies on crosslinking of purified bovine collagen and in vivo resorption have been carried out by Dr. Li (Cloft, Ka llmes et al. 2000). The length of the
31 microtubes and filaments ar e 10 mm. When assembled and packed longitudinally into the outer 14mm length tube there will be 2mm region on each end of the guide tube that does not cont ain guidance channels. This empty region allows for insertion of the sciatic nerve stumps during the nerve repair procedure. Nerve guides that contain microtubes as guidanc e channels, contained 5 microtubes in the lumen of the larger guide tube. Nerv e guides with filaments, contained 32 collagen filaments in the lumen of the outer guide tube. In the tubes that contain filaments, the guidance c hannels are formed by the s paces that exist between the solid collagen filaments. Figure 8 pr ovides a three dimensional visualization (not drawn to scale) of the three nerve guides that were used in this study. Figure 9 provides a cross-sectional schematic diagram of the three nerve guides. Figure 8 Three Dimensional Schematic of Collagen Nerve Guides Schematic diagram (not drawn to scale) of the th ree designs of collagen nerve guides used in this study. (Top left) depicts the nerve guide with 5 hollow collagen microtubes with in the lumen of an outer collagen nerve guide. (Top right) depicts t he nerve guide with 32 solid collagen filaments with in the outer collagen nerve guide. In this des ign, the microguidance channels are provided by the space between the solid filaments. (Bo ttom center) is the empty collagen nerve guide.
32 Figure 9 Cross Section Schematic Diagram of Nerve Guides Schematic diagrams (not drawn to scale) of t he cross-section of the three collagen nerve guides used in this study. All dimensions given are fr om dry nerve guides. The outer tube wall thickness is 0.02mm and the inner diameter is1.5mm for all three nerve guides. The nerve guide in the middle contains 5 collagen microtubes with a wall thickness of 0.02mm and an inner diameter of 0.4mm. The nerve guide on the right contains 32 solid collagen filaments, each with a diameter of 90 m. The theoretical maximum surface area as calculated from dry dimensions: Collagen tube (CT) = 17.1mm, Collagen tube with micr otubes (MT) = 179.2 mm, collagen tube with filaments (MF) = 147.6 mm. Residual formaldehyde was less than 04% of the nerve guide by weight following the crosslinking step (communi cation Jenssen, 2004). Nerve guides were packaged and sterilized by gamma irradiation. Although in many animal studies, the aldehyde cross linking step is accepted as a sterilization procedure, the food and drug administration (FDA) mandates that implants for human use are more stringently sterilized. Since the ul timate goal of this study is to provide a product that can be used in humans, it is designed and tested as such. This step has more significance for the phase III rat study, which incorporates growth factors and laminin into the nerve guides. Irradiation affects t he activity of these molecules, so their concentrations must be set initially to allow for this loss of activity. They undergo testing in cell cult ure prior to implantation in the rat.
33 Nerve Repair The rats were anesthetized with an intr aperitoneal injection of a mixture of ketamine HCl (90 mg/kg) and xylazine HCl (10 mg/kg), the hindquarters shaved on the right side, scrubbed with Betadine, and draped with a sterile towel while in the left side lying position. The right sciatic nerve was exposed through a longitudinal muscle splitting incision in t he mid-thigh and disse cted free from the underlying muscle bed (figure 10). T he nerve was further anesthetized with several drops of 2% lidocaine (Butler). The sciatic nerve was then transected with sharp scissors at the mid-thigh leve l and a 2 mm segment resected. The proximal nerve stump was inserted 2mm into the lumen of one end of a 14 mm long collagen guide tube and fixed in place wit h a single sling stitch (7.0 Vicryl). Next, the distal nerve stump was insert ed into the other end of the nerve guide and sutured in the same fashion. Remova l of a 2 mm segment of the rat sciatic nerve provides an acceptable balance between the nerve slack lost due to retraction of the proximal and distal sciatic nerve stumps, which occurs immediately upon transection, and the length gained by the addition of the nerve guide, which contains the 10 mm gap lesion within its lumen. In this model, this procedure eliminates excess ive tension on the newly repaired nerve. Figure 11 provides an schematic illustration of the su rgery. Polyglycolic acid (Vicryl) suture was chosen for the repair since it is re sorbable and causes minimal inflammation. A small gauge needle and suture were chos en to minimize trauma to the sciatic
34 nerve stumps. The muscle borders we re approximated and then sutured with 3.0 Vicryl. Next, the skin incision was closed with stainless steal staples (Autoclip). Figure 10 Surgical Exposure of the Rat Sciatic Nerve (A) Rat sciatic nerve viewed through the muscle splitting incision at the mid thigh of the hind limb. (B) Magnified view from enclosed region in (A).
35 Figure 11 Schematic Diagram of Nerve Transection & Entubulation Repair (A) Sciatic nerve, (B) Nerve transection, and resect ion, (C) & (D) entubulation repair. The nerve was transected at mid thigh and 2 mm segment re sected. The collagen nerve guide was placed into the surgical field and aligned with the proximal sciatic nerve. A single suture was place through the nerve guide, then through the proximal sciatic nerve, and again through the nerve guide. The suture was pulled taught to pull the nerve into the guide tube. The suture was tied to secure the repair while maintaining the patency of the nerve guide. This process is repeated on the distal nerve stump to complete the repair. Figure 12 Schematic Diagram of Entubulation Nerve Repair Image (A) is of an entubulation nerve repair of the rat sciatic nerve in vivo Image (B) illustrates the nerve guide as if it were transparent, allowing visualization of the final 10 mm nerve gap at the conclusion of the repair.
36 Figure 13 Calipers for Measuring Sciatic Nerve Vernier calipers used for all measurements during the initial surgical procedure and the terminal electrophysiological procedure. After completion of the surgery, the animal was observed and its body temperature maintained with a thermostat ic heating pad until it awoke and moved about the cage. To minimize post operat ive pain, rats received buprenorphine (Buprenex, 0.1-0.5 mg/kg) vi a intramuscular injection twice daily for three days following surgery. They received biweek ly sessions of passive range of motion to the right hip, knee, and ankle to minimi ze joint contracture in the right hindlimb throughout the six or twelve week time points. The rats were housed two per cage, kept on a 12-hour li ght/dark cycle, provided with standard rat chow and water ad libitum following the surgery. We remo ved the skin staples seven days after the nerve repair surgery.
37 Functional Assessment Walking Tracks A sciatic function index (SFI) wa s calculated from walking track measurements (de Medinaceli, Freed et al. 1982). The rats hind paws were dipped into black non-toxic water-soluble ink (Evans, Brandt et al. 1999). The rats walked across a clean sheet of white paper that had been place in the walking chamber (10 cm wi de x 85 cm long), leaving f ootprints from the hind limbs. After several practice sessions the rats became accustomed to the walking track and walked directly to the ot her end of the chamber into a dark box. The paper sheet with ink footprints was left to dry before measurements were taken. Print length (PL) was measured from the heel print to t he farthest mark left by the toes. Toes spread (TS) was m easured from the wides t prints made from the first and fifth toes. Intermediate toe spread (ITS) was measured from the middle of the second to forth toe. The largest value, during each walk, for each parameter was recorded. Walking tra cks were taken prior to surgery and at weeks 1, 3, 6, 9, and 12 following surgery. The use of walking tracks to assess sciatic nerve function after injury is a widely used method and was first implemented by De Mendinaceli in 1982 (de Medinaceli, Freed et al. 1982). De Mendi nacelis formula for SFI includes a measurement of step length from one foot to the opposite foot. Bain (1989) modified De Mendinacelisl SFI formula to exclude the distance to opposite foot measurement, as it is believed to be an in consistent indicator of changed sciatic
38 nerve function following injury (Bain, Mackinnon et al. 1989). This modified formula was used for this study. Formula: SFI = -38.3 ((EPL-NPL)/NPL) + 109.5((EPS-NTS)/NTS) + 13.3 ((EIPNIT)/NIT) 8.8 This formula contains several correct ion factors that cause the toe spread measurements to most heavily reflect in the final SFI. The formula creates an index and during normal sciatic function, the calculated value is approximately zero. Following sciatic nerve transecti on there is complete conduction block to the previously innervated muscles. T he print length measurement increases because the rat is unable to actively plant ar flex the ankle and support weight on its toes (de Medinaceli, Freed et al 1982). The toe spread measurements decrease because the rat is unable to abduc t the toes. The in trinsic muscles of the foot that abduct the toes have their innervations or iginating in the sciatic nerve (Greene 1968; Woodburne and Burkel 1988). The calculated value of the SFI under these conditions is approximatel y 100. As reinnervation of target muscle takes place the SFI value increases from 100 back toward zero, thus providing a functional index of sciatic nerve regeneratio n (Evans, Brandt et al. 1999).
39 Figure 14 Walking Tracks For Sciatic Function Index (A) Walking track lined with white paper used to record the rats hind limb footprints, which enables calculation of the sciatic function index (SFI ). (B) Prints taken from a rat prior to sciatic nerve injury. Figure 15 Enlargement Of Normal Rat Hind Paw Print Enlargement of a normal rat hind paw demonstrating the three measurements obtained for calculation of the SFI. Intermediate toe sp read (ITS) is the distance between the 2nd and 4th toes. Toes spread (TS) is the distance between the 1st and 5th toes. Print length (PL) is the distance from the furthest mark left by the heel to the tip of the 3d toe. These values were recorded for the experimental hind paw prints, the contralateral paw, and were entered into a formula to calculate the SFI.
40 Static Stance A second index, percent of pre-inju ry toe spread, was calculated from static stance photographs. The rats we re placed in a clear Plexiglas box (20x20x20 cm) supported on a frame above a 45 degree angled mirror. This apparatus allowed a photograph to be taken of the plantar aspect of the feet. A ruler with a millimeter scale was included in the image to allow calibration of the software (ImageJ, NIH) measuring tool. The digital photographs allowed a more accurate measurement of toe spread than could be obtained by measuring the ink walking tracks. Static stance phot otgraphs and measurements were taken prior to surgery, one week following surger y, and at the final six or twelve week time point. A static stance sciatic index was calculated by dividing the toe spread at the terminal time point into the pre injury toe spread to obtain percent of normal toe spread. Bervar (2000) used a video camera to obtain similar images and found that using this simple formula did not decrease accuracy significantly for determining sciatic ner ve function (Bervar 2000).
41 Figure 16 Image Static Stance Photography Box (A) Frame, clear box, and mirror set to a 45-degr ee incline, was used to capture images of the plantar surface of the rats hind paws. (B) Im age obtained after a rat was placed in the box and a camera was placed on a tripod and aimed at the mirr or. A ruler with a millimeter scale is included in the image. When the digital image was impo rted into the computer and subsequently loaded into the image analysis software (NIH, ImageJ) t he ruler was used to set the scale, enabling accurate measurements of both intermediate toe spread and 1st to 5th toe spread. Electrophysiology Great efforts were made at the onse t of this study to delineate an electrophysiological protocol that would provide objective quantitative data, which would allow testing for statistically signifi cant differences in functional recovery between groups. It was realized at seve ral points during the study that our electrophysiology protocols were not goi ng to give us the quantitative data on which we had originally planned. After consultation with by Drs. Liuzzi, Madison,
42 and Li, decisions were made to modify the electrophysiology methods several times during the course of this study. We started the study recording evoked compound nerve action potentials directly fr om the sciatic nerve distal to the repair. We then changed the protocol to record compound muscle action potentials from muscles innervated by t he sciatic nerve. There were several derivations of muscle recording techniques. With each change in protocol, we believe the accuracy and relevance of the data improved. However, this came at the cost of inadequate and nonuniform sampling of the different r epair groups and time points needed for statistical analysis. We accepted this trade off, during the guidance channel (phase II) rat study, because we wanted the most refined methods prior to entering the growth factor and adhesive mo lecule phase (phase III) of the study at USF. The results of these differ ent methods and the discussion of their limitations are elaborated upon in the respective secti ons of the dissertation. General Electrophysiology Principles When either the nerve or muscle depol arize in the region perpendicular to the active extracellular re cording electrodes the electrical potential becomes more negative relative to the reference electrode (Oh 1993). This is represented by a negative deflection on t he computer monitor at the positive electrode. However, the cathode is used as the acti ve recording electrode, so the initial deflection on the computer monitor appear s in the positive direction (Misulis
43 1997). By convention, it is referred to as the negative deflectio n of the compound muscle action potential (Misulis 1997). Latency, measured in ms, from th e stimulus to the initial negative deflection from baseline of the supramaximal CMAP is t he time it takes for the fastest conducting axons to conduct the action potentials from the stimulating electrodes, down the sciatic nerve, and to depolarize muscle tissue under the recording electrodes (Oh 1993). Estimate s of nerve conduction velocity (NCV) can be calculated by dividing the distanc e in millimeters from the stimulating electrode on the sciatic nerve to t he active recording electrode in the gastrocnemius by the latency of the CMAP. Latency is defined either as the time interval between the stimulus and the peak of the first negative deflection of the CMAP, or as the time interval between t he stimulus and the first deviation from baseline of the CMAP. Finally, the amplitude, measured in mV, of the supramaximal CMAP was noted. Although latency or latency divided by distance, provide an estimate of conduction velocity, a more accurate method for evaluating nerve conduction velocity was employed near the end of this study. By first stimulating the sciatic nerve proximal to the repai r site (S1), and recording supramaximal CMAPs from the gastrocnemius muscle, and then stimulat ing the sciatic nerve distal to the repair site (S2), while recording fr om the same gastrocnemius recording electrode, two latency values are derived. In normal nerve, the latency is shorter from the S2 site than S1 because the nerve conducts over a shorter distance to
44 the recording electrode in the gastrocnem ius muscle. The distance from each stimulation site to the active recordi ng electrode in the gastrocnemius muscle is measured with calipers and recorded. Final ly, NCV across the nerve repair site is calculated by taking the differenc e in the two distance measurements and dividing it by the difference in the latency values. To verify that the recorded signals are indeed derived from the nerve of interest, some investigator s cut the nerve at the end of the experiment (Krarup, Archibald et al. 2002). If a stimulus no longer produces a CMAP following nerve transection, there is confirmation that t he previously recorded signal originated from the nerve of interest. While this pr ovides a control, it prevents fixation by perfusion and the possibility to te st for correlations between the electrophysiological results and the histologic al results in the same animal. In this study we perfused the rats to optimize ti ssue fixation for morphological study. This is an important design element of this study. One important consideration in this study was the possibility that while inserting needle electrodes into the gastrocnemius in animals with muscle atrophy, the needle would pass into deeper pos terior tibial muscles. Therefore, every attempt was made to prevent this from occurring. It is im portant to realize, however, that all deep muscles below the knee have their innervation originating in the sciatic nerve (Greene 1968; Woodburne and Burkel 1988). This means that all CMAPs recorded from this region, following repa ir of the sciatic nerve, should represent successful reinnerv ation via the nerve graft.
45 Animal Preparation for Electrophysiology At 6 or 12 weeks post nerve guide impl antation, rats were anesthetized with an intraperitoneal injection of a mixture of Ketamine (90 mg/kg) and Xylazine (8 mg/kg). Anesthesia was maintained with supplemental injections. The animal was prepared for surgery as previously described and body temperature was maintained at 37C on a thermostatic pad. The right sciatic nerve was reexposed through a longitudinal muscle spli tting incision in the mid-thigh. The nerve repair site was typically surrounded by fibrous adhesions, but immediately proximal to this region normal nerve was identified. Recording electrodes were connected to a bio amplifier (AD/Inst ruments) which was wired to a signal integration unit (AD/Instruments PowerL ab 4/SP). This equipment displays the recorded signal on a computer monitor. The electrophysiological methods evolved during the course of this st udy and each method is described below in the order they were implemented. Method 1: Compound Nerv e Action Potentials The first electrophysiology method used in this study was the recording of the sciatic compound nerve ac tion potential (CNAP). The nerve was exposed as described above, and carefully mobilized from the underlying tissue interface before it was suspended on two pairs of hook electrodes (Figure 17). The stimulating electrode pair was place prox imal to the nerve repair site with the
46 distal hook set as the cathode. The pai r of recording electrodes was placed immediately distal to the repair site. With this set up, only nerve conduction across the repair site was recorded. A piece of Parafilm was cut and placed between the nerve and the underlying muscle to facilitate isolation of the stimulus and recording electrodes, and to prevent conduction though ionic fluids present in the surgical area. We slowly increas ed the stimulus intensity from zero and noted both the stimulus leve l in milliamperes (mA) at which a visible muscle twitch below the knee could first be obser ved and the lowest stimulation level at which a CNAP was recorded. This was performed on the contralateral normal sciatic nerve in the same manner. Figure 17 Hook Electrodes for Elevating the Sciatic Nerve
47 Method 2: Compound Muscl e Action Potentials All of the remaining derivations in the electrophysiology methods used in this study involved recording CMAPs from the gastrocnemius muscle. Initially, we maintained the same stimulation electrode setup as was used for the recording of nerve action potentials. We el evated the sciatic nerve, proximal to the repair site, with a pair of platinum hook electrodes. A monopolar stainless steel uncoated needle electrode was insert ed percutaneously in the midpoint of the right lateral gastrocnemius muscle, 30 mm proximal to the calcaneus. This electrode was the cathode and served as the active recording electrode. A second monopolar needle electrode was insert ed 5 mm distal to the cathode and was positioned subcutaneously over the gas trocnemius. A second pair of needle recording electrodes was inserted into the interosseous muscle of the third metatarsal space of the right foot. A common ground for each pair of recording electrodes was inserted in the anterior th igh of the same limb. Each pair of recording electrodes was connected to a s eparate bio amplifier (AD/Instruments) which was attached to a signal integration unit (AD/Instruments PowerLab 4/SP). The signal was displayed on a computer monitor and recorded (figure 18).
48 Figure 18 Electrophysiology Equipment with Surgical Microscope Figure 19 Needle Electrode 10 mm 30 gauge stainless steel monopolar needle electrode used for recording compound muscle action potentials from the gastrocnemius muscle and intrinsic muscles of the rat hind paw. A notch filter for 60 Hz interferen ce was used during recording along with the high pass filter setting of 10 Hz and low pass filter of 5 kHz. A 20 ms sweep was recorded for each 0.10 ms electrical st imulus to the sciatic nerve. As the
49 amplitude of the CMAP incr eased with each increase in stimulus intensity, the recording range was increased to maintain the peak of the CMAP on the screen. This is comparable to reducing the gain in an analog system, which prevents loss of data referred to as clip ping. The range of 0.01 mA to 3.0 mA was explored to insure reaching supramaximal int ensity (Daube 1996). The supramaximal stimulus is defined as the stimulus above wh ich there is no further increase in the ampitude of the CMAP (Oh 1993; Misulis 1997). Abov e 2-3 mA of stimulus intensity, volume conduction can occu r and thus, any CMAPs recorded are not physiologically relevant (Personal communication, R. Madison, 2003). Method 3: Motor Unit Nu mber Estimation (MUNE) Upon the implementation of the motor unit num ber estimation (MUNE) technique, we modified the stimulatory electrode setup by discontinuing the use of the hook electrodes for stimulating the sciatic nerve. The new cathodal stimulating electrode was a platinum wir e shaped into an inverted Y. It was lowered onto the nerve proximal to the repair site, positioned to maximize contact with the nerve, and held in place by a sma ll manipulator. The anodal stimulating electrode was placed deep into the gluteal muscle adjacent to the sciatic nerve approximately 10 mm proximal to the ca thode. The recording electrode was inserted percutaneously into the gastrocnem ius muscle as previously described. The filters were set as in the previous muscle recording method.
50 Figure 20 Cathode Stimulator Held By Manipulator Manipulator used to hold the cathode electrod e over the top of the sciatic nerve during the electrophysiological assessments. (B) magni fied view of the platinum wire electrode used to stimulate the sciatic nerve prox imal to the region that was repaired. Figure 21 Stimulating Electrode Proximal to Nerve Guide The intensity of the stimulus was increased in small increments until a small CMAP appeared on the monitor. The stimulus was increased further, in small increments, and five traces were recorded at each stimulus intensity. Between each stimulus, a 1-second delay was provided to prevent fatigue.
51 To facilitate analysis, an overlay of all recorded CMAPs was displayed on the monitor to allow the identification of separate and unique waveforms. As the CMAP amplitude increased with increased st imulus intensity it represents the summation of CMAPS previously identified at lower stimulus intensities (1996; Krarup, Archibald et al. 2002). For any CMAP between threshold and the supramaximal stimulus, subtracting t he amplitude of the preceding CMAP gives the approximated amplitude si ze of a motor unit. Th is was performed for 5-10 distinct CMAPs above threshold. The diffe rences in amplitudes of the individual motor CMAPs were then averaged to obtai n a more accurate representation of the average motor unit amplitude. This number was then divided into the amplitude of the supramaximal CMAP to arrive at an esti mate of the total number of motor units (McComas, Fawcett et al 1971; Daube, Gooch et al. 2000; Krarup, Archibald et al. 2002). This method has been termed the statistical MUNE method since an average is calculated (Daube 1995; Shefner 2001). The statistical motor unit number esti mation method has been proven more reproducible than the multiple point method, which involves moving the stimulating electrode along the nerve to obtain different wave forms (Daube, Gooch et al. 2000). When t he incremental method was compared with multipoint methods in transgenic ALS mouse, t he results demonstrated .95 correlation (Shefner, Cudkowicz et al. 2002). MUNE has been successfully employed in the study of peripheral nerve regeneration fo llowing entubulation repa ir of the median nerve using collagen tubes in adult monk eys (Krarup, Archibald et al. 2002).
52 The stimulus isolation unit (World Pr ecision Instruments) uses a 5V pulse from the PowerLab main unit to trigger a constant current output with resolution of .01mA through the 10mA range. Preliminary studies performed in our laboratory have validated that this unit has the sensitiv ity to adjust current in increments necessary to resolve sma ll quantal increases in CMAP amplitudes. This is necessary for MUNE. Please see the discussion section for an elaboration of the limitations of MUNE in this study. Method 4: Compound Muscl e Action Potentials The final method used in this study was the recording of the supramaximal CMAPs of the gastrocnemius and intrinsic muscles of the foot using the same setup as before, but with the additional step of transe cting and dissecting out the nerve to the biceps femoris muscle that branches from the tibial division of the sciatic nerve proximal to the repair site. The reason for removing this nerve is elaborated upon in the results and discussion sections. In addition, the sciatic nerve was stimulated distal to the repai r (S2) and CMAPs were recorded from the same gastrocnemius muscle recordi ng electrodes. Following recording of the gastrocnemius muscle CMAPs from the two stimulating sites described (S1 and S2), the distance from the nerve stim ulation sites to the active recording electrode was measured with calipers and re corded. Nerve conduction velocity was calculated from the difference betw een the distance measurements divided by the difference in the CMAP lat encies from each stimulation site.
53 Tissue Isolation Following electrophysiological assessm ent, animals were euthanitized by intraperitoneal injection of pentobarbital (150mg/kg). The animals were transcardially perfused with phosphate buffe red saline (pH 7.2) followed by 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.2). All attempts were made to minimize stress and discomfort of the animals prior to euthanasia. These methods of euthanasia are consistent wit h the recommendations of the Panel on Euthanasia of the Am erican Veterinary Medical Associ ation. The animal protocol for the proposed study was reviewed by the University of Sout h Florida Animal Care and Use Committee and received approval #1820. After perfusion the nerve guide was disse cted free, divided into proximal, middle, and distal segments (each 3.0 mm in length), post-fixed in 1% percent glutaraldehyde and 2% percent paraf ormaldehyde in 0.1 M phosphate buffer (pH7.2) for 24 hours followed by 1 hour in 2% osmium tetroxide (Electron Microscopy Sciences). One 3 mm section of the contralateral mid-thigh sciatic nerve was processed in the same m anner for a normal control. The nerve specimens were embedded in Araldite/E mbed-812 (Electron Microscopy Sciences). Two-micrometer transverse se ctions were cut from the tissue blocks, floated onto glass slides (Superfrost charged), stained with toluidine blue, mounted (Permount), and cover slipped. Toluidine blue staining of peripheral nervous tissue enables visualization of myelinated axons, blood vessels, mast
54 cells, macrophages, and other cell types (Chamberlain, Yannas et al. 1998; Hashimoto, Suzuki et al. 2002). Morphometric Analysis Partially overlapping images of the entire cross section of the nerve guide/sciatic nerve were captured using a digital camera and 40x objective lens (Qcapture software, Olympus Q-color ca mera, Olympus BX41 microscope). The images were montaged using commercially available software (Adobe Photoshop Elements 2.0) and saved as one image of the entire cross section of the sciatic nerve (figure 22).
55 Figure 22 Montage of Nerve Guide Twenty-four partially overlapping images (40x objective lens) digitally captured (Qcapture software, Olympus Q-color camera, Olympus BX41 microscope) from the nerve guide of an animal that underwent entubulation repair of the sciatic nerve with an empty type I collagen nerve guide 6 weeks prior to tissue harvesting. The images were montaged (Adobe Photoshop Elements 2.0) and saved as one image that incorp orated the entire cross section of the sciatic nerve. Although not appreciated, due to the size of the figure above, this montage contains sufficient resolution for the operator to zoom in and identify individual myelinated axons. The montages varied in size according to the cross sectional area of the regenerated nerve. Every myelinated axon in the images were counted (ImageJ). The image was displayed on the computer monitor and the operator moved a cross hair over an axon using the computer mouse and clicked the mouse (figure 23). The software creates a running tally and this process was repeated until all myelinated axons in the image are counted. The 40x objective lens provides magnification necessary to correctly identify myelinated axons and a field of view that is amenable to creati ng a montage of the entir e cross section.
56 Nonparametric statistical analyses were per formed, using Sigma Stat software to test for significant differences among groups. Figure 23 Axon Counting of Normal Nerve Micrograph of normal rat sciatic nerve shows t he method of counting myelinated axons using ImageJ software. A crosshair is moved with the mouse over a myelinated axon and when clicked it leaves a colored dot in the middle of the ax on. This prevents recounting of any axon in the image. This micrograph captured through a (100x oil objective lens) is shown for illustration of the method we will use to count axons. The act ual axon counting was performed on montages of images since they cover a larger field of view. Axon diameters were meas ured (ImageJ) from thr ee (100x oil objective) fields digitally captured from the sa me sections chosen for myelinated axon counting. In ImageJ the measuring scale is set by photographing an image of a micrometer with the same microscope lens and camera parameters that are used for photographing the nerve images, and then drawing a line with the computer mouse and line tool across a known dist ance on the micrometer image. This calibrates the pixel dimensions of the di gital image to micrometers. Regions with
57 a high density of myelinated axons were selected for measurement. Magnification provided by the 100x oil objective lens facilitates accurate measurement of axon diameters. A stra ight line across the sh ortest distance of the axon cross section, ex cluding the myelin, was selected for measurement of each axon. Only axons with myelin per imeters entirely in the images were measured. Axons with double myelin ri ngs, such as occurs when SchmidtLanterman clefts were in the plane of section, were excl uded from measurement to facilitate a more accurate evaluation of axon diameters.
58 Chapter 4: Results Introduction The results section is organized as follows: First, general observations are discussed. Following the secti on on general observations, a section on descriptive morphology of both the autograft and entubulation repairs is presented. It starts with findings that ar e generalized to all the nerve repairs in this study and then each repair type are discussed individually. Following this section on morphology, the results from myelinated axon counts and diameter measurements are presented. Next, repr esentative electroph ysiological data are presented. Lastly, the resu lts of the walking track and static stance analysis are presented. General Observations None of the rats through the cour se of this study demonstrated autophagia. Many pr esented with elongated toenails an d differing amounts of phalangeal flexion abnormalitie s of the experimental limb. No animals that survived the initial surgery were lost dur ing the remainder of the study, and those
59 that were lost due to surgical complicati ons were replaced in order to achieve a minimum of ten rats per group. Several animals were noted to have mild heel ulcerations on the experiment al limb. These lesions healed without complication or intervention. All rats presented with si gnificant ankle contra ctures that limited full plantar flexion of the experimental limb. At the six-week time point, the proximal and distal ends of the entubulation repair site were readily identified during the surgical exposure, whereas at the twelve-week time point this was more difficult. At si x weeks, the intramuscular suture of the biceps femoris muscle had not been fully resorbed and was quite evident. However, at twelve weeks, t here was no grossly detectable trace of suture. To accurately identify the nerve r epair site at 12 weeks we relied on anatomic landmarks such as the nerve branch from the tibial division of the sciatic nerve and the underlying gluteal musculature, to identify the collagen guide tube prior to dissection. At both the 6 and 12 week time points, the region of nerve injury and repair was covered wit h fibrous adhesions. The repairs that went 12 weeks appeared to have more adhesions, thus challenging the identification of the exact si te of the nerve guide. The autograft repairs were covered with fibrous adhesions, but less so than the entubulation repairs with collagen ne rve guides. After careful dissection to free the nerve repair site from t he adhesions, the 12-week autografts appeared
60 grossly as normal nerve. The junctions of the nerve graft with the proximal and distal sciatic nerve were i ndiscernible in most cases. The fibrous white adhesions tethered t he nerve repair to the surrounding muscular interfaces. In contrast, the nerve prior to injury was surrounded by a translucent areolar connective tissue inte rface, which allows the sciatic nerve mobility independent of the surrounding muscle. This tissue is necessary to prevent traction of the nerve by allowing for the nerves longitudinal excursion during flexion and extension of the limb join ts. This translucent tissue is referred to as the mesoneurium by some surgeons and investigators, however, this terminology is not universally accept ed (Mackinnon and Dellon 1988). The fact that this flexible connecti ve tissue, which is adherent to, but distinctly different from the epineurium, is presen t, in our opinion is indis putable. This change from a translucent mobile connective tissue prior to injury, to an adherent fibrous opaque tissue in the weeks following repair is one of the most obvious changes in the gross anatomy noted during surgical inspection. Another observation made was that once adhesions were care fully dissected from the nerve repair site, a swelling at the proximal union of the sciatic nerve and the nerve guide could be frequently identified. The nerve to the biceps femoris muscl e, which branches from the tibial division of the sciatic nerve proximal to the repair site, wa s most often adherent to the repair site. This has great relevance to the electrophysiological assessment. Since the stimulatory elec trode was placed proximal to the nerve
61 repair, it stimulated the nerve to the bi ceps femoris along with the sciatic nerve unless the nerve to the biceps femoris muscle was carefully dissected free from this site and transected. This is elaborated on in the section devoted to electrophysiology results and in the discussion section. Morphology In the region of the gap repair, all the successfully regenerated nerves contained myelinated axons of smaller diam eter than normal nerv es (figure 24). The regenerating axons had a thinner myel in investment at all time points studied. There appeared to be a greater number of bot h mast cells and blood vessels in the regenerating nerves when compared to normal control nerves. Monocyte derived macrophages aggregated by the suture site, their cell membranes fused, and they became gi ant cells (figure 25), which phagocytose the polyglycolic acid suture. In contra st to the six week time point, where histological evidence remained of the ner ve guide suture fixation, by twelve weeks, suture material had been fully resorbed.
62 Figure 24 Normal Nerve and Regenerated Nerve In Collagen Tube (A) Normal sciatic nerve image (B) section of the distal segment of the nerve guide without microguidance channels harvested 12 weeks post entubulat ion repair nerve. Note the significant number of cells, myelinated axons, and blood ve ssels in the regenerating nerve. Since the myelinated axons in the guide tube do not grow in a homogeneous fashion throughout the entire cross section of the nerve guide, they were a ll counted to ensure an accurate account of the nerve regeneration. The normal nerve is less cellular, has a more homogeneous distribution of myelinated axons, contains axons of larger diam eter, and contains axons that are surrounded with a thicker myelin investment, as compared to the regenerated nerve.
63 Figure 25 Suture Resorption By Giant Cells Region of nerve guide 6 weeks post nerve repair in the process of resorption by macrophages, which fuse and become giant cells (arrow). Nerve Regeneration in Autografts In the mid section of nerve specimens harvested from rats that received autograft nerve repairs, there were as many as three mast cells in a single 1000x field (Figure 28). Myelin debr is from axons that were present in the nerve graft prior to harvesting was still present at 6 weeks post injury/repair and was in the process of removal by Schwann ce ll and macrophage phagocytosis. At 12 weeks post implantation, the majority of myelin debris from the donor nerve had been cleared. In addition, there was more myelinated axons and larger myelinated axons at th is later time point.
64 Figure 26 Autograft Nerve Repairs Low power images of mid section of autogra ft nerve repairs (A) 6 weeks, (B) 12 weeks
65 Figure 27 Nerve Regeneration in a 3 Week Autograft (A) Low power (B) high power views of the midsection of an autograft 3 weeks post nerve repair, showing small myelinated axons (black arrows ) regenerating along Schwann cell basal lamina tubes (asterisk) that have been cleared of myelin. Ovoids of degenerating myelin remain in many of the tubes (white arrows). (C) Low power, (D) high power views of the sciatic nerve distal to the repair site show a similar distribution of em pty Schwann cell tubes and those with degenerated myelin to (A & B), however, no regenerating axon s can be detected in this distal region of the nerve with this staining preparation.
66 Figure 28 Nerve Autograft 6 Weeks Post Repair Photomicrograph (100x oil objective lens) from mid section of autograft repair at 6 weeks. Myelin debris from axons that were present in the nerve graft prior to harvesting are still present at 6 weeks post injury/repair (black open arrows). Based on morphology and their relationship with degenerating axons, nuclei of a Schwann cell or macrophage (white circle) is identified and is presumably in the process of removing the myelin debris via phagocytosis. Nuclei of Schwann cells can be identified by their cl ose proximity to newly regenerated axons, which are identified by their smaller than normal diameter and thin mye lin investment (white open arrows). Their dark staining cytoplasmic granules easily identify mast ce lls (solid white arrows) present in this section.
67 Figure 29 Autograft Nerve Repair 12 Weeks Post Implantation (B) High magnification view of boxed region in (A). This specimen had 22,590 myelinated axons counted from a montage of this section. Note the extremely high density and uniform distribution of myelinated axons. Axonal escape occurred at the prox imal junction between autograft and sciatic nerve, as evidenced by the pres ence of many myelinated axons growing
68 outside the confines of the autograft epi neurium. This was noted in several specimens harvested from the middl e of the graft (figure 30). Figure 30 Autograft With Axonal Escape (B) High magnification view of enclosed region in (A) showing many regenerated axons (asterisk) that grew outside the confines of the autograft epineurium. AG with 13,582 myelinated axons at 6 weeks. Image demonstrates axonal escape with AG repair
69 Empty Collagen Nerve Guides The majority of empty collagen nerv e guides had patent lumens at six, twelve, and twenty four weeks post impl antation. Some had a flattened cross section, which was noted both grossly and hi stologically. Over time the collagen nerve guide wall became more infiltrat ed with cells, however, the approximate wall thickness did not appear to change to a large degree (figur e 31). At the six week time point the nerve guide lum en was more sparsely occupied by regenerated nerve (axons and Schwann cells) than at twelve weeks. By 24 weeks, the area occupied by axons appear ed greater than at 6 and 12 weeks, filling the nerve guide lumen. A frequent observation was the presence of mini fascicles in regenerated nerve se ctions (figures 32, 33).
70 Figure 31 Low Power 6, 12, 24 Week CT Nerve Guide Sections Low power images of mid section of empty nerve guides (A) 6 weeks, (B) 12 weeks, and (C) 24 weeks.
71 Figure 32 Regenerated Nerve in Collagen Guide Distal segment of the collagen guide tube 12 weeks po st repair with (B). Note the significant number of cells, myelinated axons, and blood vessels in the regenerating nerve. Figure 33 Minifascicles in Regenerated Nerve High power view (100x oil objective lens) of regenerated nerve in at 12 weeks demonstrating minifascicles (open arrows), and a perineurial ce ll (black arrowhead) bottom left region of image.
72 Figure 34 Regeneration in CT at 24 weeks (A), (B), (C), low, medium, high power views respectively from a CT nerve guide specimen harvested at 24 weeks post implantation. This particular specimen had the highest number of myelinated axons (24,515) in the mid-section of the nerve guide of all entubulation repairs performed in this study. It also had more myeli nated axons than the majority of the 12-week AG repairs. The myelinated axons are well distribut ed throughout the entire cross-section of the nerve guide lumen.
73 Figure 35 Failed Regeneration in CT Group Middle section of a nerve guide 6 weeks post implantation with a flattened morphology. The lumen has been reduced, but not obliterated. No myelinated axons were identified in this specimen.
74 Nerve Guides with Longitudinal Microtube Channels Regenerating axons grew through and between longitudinal collagen microtubes within the lumen of the lar ger collagen nerve guide. Morphology in these regenerated nerve guides was highly variable. The outer guide tube wall in some specimens was infiltrated by cells and appeared less dense than some microtubes (figure 36). Within a single nerve guide, the degree of resorption of the individual microtubes varied, as well as the axonal growth through the microtubes (figure 41, 42). Variation also existed between specimens. The number of microtubes that c ould be identified hi stologically, at 6 and 12 weeks, was on occasion, less than 5, even though we confirmed their presence prior to surgical implantation.
75 Figure 36 Microtube With Myelinated Axons (A) Low and (B) High power micrographs of the distal segment of a nerve guide with microtube guidance channels 12 weeks after injury and repair. Note myelinated axons in a microtube that has not degraded. The original outer collagen tube has been infiltrated by cells and has been remodeled to a greater degree than the collagen microtubes. In one particular specimen, the presence of a population of plasma cells was noted (figure 37). This was the ex ception, and was not observed in other
76 specimens. Minifasicles were observ ed in the regenerating nerves in some specimens of the MT group (figure 40). Figure 37 Regenerated Nerve with Plasma Cells inside Microtube (A) Low power view of few regenerating axons in a microtube within a nerve guide at 12 weeks. (B) High power view from same specimen rev ealing a mini-fascicle containing myelinated and unmyelinated axons. Outside the borders of the fa scicle are plasma cells (open arrows), which are identified by their clock faced pattern of dark staining heterochromatin and basophilic cytoplasm. A blood vessel containing erythrocytes and a leukocyte is present in the image. The image contains a mast cell (white arrow).
77 Figure 38 Low Power MT Nerve Guides at 6 & 12 Weeks (A) Sample middle sections from 6 week MT group. (B) Sample middle sections form the12 week MT group
78 Figure 39 Axons Regenerating In and Between Microtubes (A) MT nerve guide specimen harvested 12 weeks po st implantation. (C) magnified view of enclosed region in (A) demonstrates dense axon growth in and between (arrow) two microtubes. (B) High magnification (100x oil objective l ens) showing well myelinated axons from this specimen.
79 Figure 40 MT Nerve Guide at 12 Weeks Fasicular Morphology A12 week MT specimen with 9994 myelinated ax ons and a highly fascicular morphology. Perineurial cells (arrow). Variability in nerve guide resorption al so existed among different sections of the same nerve guide. In one particula r specimen from the 6 week MT group, the proximal section had evidence of only 2 microtubes, and the middle section had only 4 microtubes. The distal secti on had 3 intact micr otubes, 1 microtube that was largely resorbed and at the same time partially collapsed, and a region where a tube may have exist ed but remnants were not det ectable (figure 41). In a failed repair from the 6 week MT group, the microtubes were highly infiltrated
80 by cells, and highly resorbed. They also contained fibrotic tissue without the presence of axons (figure 43). Figure 41 Variable Resorption of Microtubes Proximal to Distal (A) proximal, (B) middle, (C) distal sections of a 6 week MT demonstrating variability of microtube morphology. Upon implantation there were 5 collagen microtubes, however, at the six week time point there are only 2 identifiable micr otube present proximal se ction, 4 in the middle section, 4 in the distal (one collapsed, arrow). In the distal section, the wall thickness of each microtube is different. There are many axons in the proximal section. The middle section has only 275 myelinated axons.
81 Figure 42 Microtube Variability (A) Nerve guide section from 12 week MT group. (B) Nerve guide section from 6 week MT group. These two specimens although from two different ti me points show remarkably similar microtube morphology; each of the 5 microt ubes are visible within the outer guide tube and have a variable wall thickness. This particular 12-week spec imen was one with very successful regeneration (12,905 myelinated axons in midsection of nerve guide), where as this 6 week specimen had particularly poor regeneration (186 myelinat ed axons in mid-section of nerve guide).
82 Figure 43 Failed MT Nerve Guide This specimen had no visible axons in the nerve guide at 6 weeks post nerve repair.
83 Nerve Guides with Longitudinal Filaments Channels Axonal regeneration was highly vari able at both 6 and 12 weeks in nerve repairs with nerve guides containing longitudi nal collagen filaments. Intraluminal filament position was highly variable and there was little resorption or cellular infiltration of the solid co llagen filaments (fi gure 45). In some specimens axons and cells grew between the filaments (fi gure 45), but in others axons grew in regions outside a grouping of filaments (figure 46).
84 Figure 44 Nerve Guides with Filaments at 6 & 12 Weeks Low power images of mid section of nerve guides with longitudinal filaments (A) 6 weeks, (B) 12 weeks.
85 Figure 45 Nerve Guide With Filaments (A) Low, (B) medium, and (C) high power images of a cross section of a nerve guide that contained filaments, 12 weeks after injury and r epair. (B) High power image from enclosed region in image (A). The collagen filaments show li ttle cell infiltration or degradation (black arrows in image B). Note myelinated axons (black arrows in image C) in the spaces in between the collagen filaments that have not degraded and the significant number of blood vessels (open white arrows) that have infiltrat ed the nerve guide. The nucleus of an endothelial cell is bulging into the vessel lumen. Some of the blood vessel s have infiltrated the collagen filaments but most occupy the space outside t he filaments. Although there is a f ilament in this image (B) that is surrounded by myelinated axons, this was not t he trend. Most filam ents aggregated and axons grew around them.
86 Figure 46 Myelinated Axons Growing Outside Region Of Filaments (A) Low power image of a nerve guide middle sect ion with filaments from rat 171 at 6 weeks demonstrates that filaments did not attract axons to grow in the channels between them and that a very large population of axons exists outside t he region of the filaments. In this specimen there exists 2292 myelinated axons. (B) High power im age from the boxed region enclosed in (A).
87 There were several specimens in which the individual filaments maintained an equal distribution within the outer nerve guide, thus providing the maximal surface area permitted by this des ign (figure 47). These specimens did not have the greatest number of axons. In fact, certain specimens in which the filaments aggregated, and axons grew in r egions outside the filaments, had the greatest axon counts of the MF groups. This was not a trend. There were specimens with aggregated fila ments that had little or no axons. In some, the aggregated filaments occluded the area avail able for axon growth while in others small numbers of axons grew between t he filaments (figures 50, 51). In specimens that had evenly dispersed filame nts, the filaments were not equal in diameter. Some appear appr oximately twice as large as others do (figure 47).
88 Figure 47 Nerve Guide Image Demonstrating 32 Separate Filaments Nerve guide section containing 32 distinct and separate filament s. The myelinated axon count was 1681. Though this nerve guide contained f ilament of even spacing, regeneration was marginal. Figure 48 Nerve Guide Filaments And Many Axons At 6 Weeks This specimen had the greatest number myelinated ax ons (3334) of all of the MF nerve guides at 6 weeks. Some filaments maintained a fair deg ree of separation, and there were axons growing between them, while other filaments contacted each other.
89 Figure 49 Nerve Guide With Filaments and Few Axons at 6 Weeks (A) Montage from 6 week MF specimen. The fila ments are separate and distinct. (B) reveals poor regeneration, few myelinated axons (black arrows) present in spaces between filaments. Total number of myelinated axons in this specimen is 393.
90 Figure 50 Aggregated Filaments & Failed Regeneration In this 12-week MF specimen, the filaments appear to have swollen and aggregated obliterating space for axon growth. There are no axons in this section and little cellular infiltration of the filaments has taken place.
91 Figure 51 Small Number of Axons In between Filaments At 6 weeks post implantation this MF nerve guide specimen has few myelinated and unmyelinated axons growing axons between filaments Myelinated Axon Counts and Statistical Analyses The montage technology worked successfully. It allowed us to count myelinated axons from the entire cross se ction of the nerve specimen. Although labor intensive, it proved critical for providing accurate evaluation of nerve regeneration through nerve guides with such varied morphology. Sample montage images from each nerve repai r group and the normal control are displayed in figures 52 and 53.
92 Figure 52 Nerve Repair Montages Montages from each nerve repair group. (A) AG, (B) CT, (C) MT, (D) MF. Figure 53 Normal Control Montage
93 Figure 54 Partially Counted Nerve Guide (A) Partially counted nerve area from a montage of a nerve guide with microtube channels 12 weeks after repair. Counted, myelinated axons ar e marked by dots. (B) Magnified view of region enclosed in (A).
94 Descriptive statistics of mye linated axon count are discussed and summarized in Table number 2 and Figur e number 55 below. Following these are the results from the infe rential statistical analyses. Figure 55 Number of Myelinated Axons After Post Nerve Repair The mean number of myelinated axons in the middl e section of the repair site increased from 6 weeks to 12 weeks in each repair group except MF. Differences between groups are considered significant with P < 0.05. At 6 weeks, AG has significantly more axons than CT, MF, and MT. There is no significant difference in the number of myelinated axons between CT, MT, and MF. At 12 weeks, AG has significantly more myelinated axons than CT, MT, and MF. Both CT and MT have significantly more myelinated axons than MF. There is no significant difference between the MT and CT group at 12 weeks. CT group at 24 weeks is not significantly different from the AG at 12 weeks. At both 6 and 12 weeks, only the AG had significantly more myelinated axons than the normal control.
95Table 2 Myelinated Axon Counts Raw Data 6 week 6 week 6 week 6 week 12 week 12 week 12 week 12 week 24 week Normal Control AG CT MF MT AG CT MF MT CT NC 9,201 2,354 0 2,17212,712 9,3840 4,52122,515 6,031 12,694 0 3,3340 9,1331,2720 744 21,486 4,279 13,582 36 82 222 13,7154,444103 3,4779,812 7,430 11,168 485 159 0 22,5904,28010 5,0556,459 5,920 8,393 5,108 393 0 14,2773, 2131,05011,923 1,756 7,640 0 2,048 913 3,95121,4122,9474,7899,994 7,318 10,106 135 1,6810 14,0261,7550 2,115 7,234 9,881 527 220 275 13,6674,174255 24 7,852 10,031 1,084 2,292193 9,5382,1792,696575 7,687 10,425 0 1,632186 9,842464 4,957 7,327 0 12,905 1,520166 935 At 6 weeks, the mean number of myelinated axon s in the midsection of the AG repair was 9548. There was one complete failure (0 myelinated axons) out of 10 animals (n = 10). The AG repair at 6 weeks had myelinated axon counts ranging from 0 to 13,582. O ne specimen had a 0 axon count. The specimen containing the fewest number of myelinated axons in the 6-week AG group, had 8393 myelinated axons. The m ean number of myelinated axons in the midsection of the entubulation nerve gui de repairs (CT, MF, MT) was 1,178, 1,051, and 700 respectively (n = 10 in each group). In the 6-week time point, the myelinated axon counts ranged from 0 to 51 08 in the CT group, 0 to 3334 in the MF group, and 0 to 3951 in the MT group. Each of the 6-week entubulation repair groups had at least 3 animals that contained fewer than 100 axons in the
96 mid-section of their nerve guides, and each group had at least one complete failure (0 myelinated axons). In the MT group only 2/10 (20%) had myelinated axon counts greater than 2000, and the rest, 8/10 (80 %), had axon counts less than 300. In fact, 4/10 (40%) had 0 mye linated axons in the MT group at the 6 week time point. In the MF group 6/10 (60%) had counts great er than 300, 2/10 (20%) had counts great er than 2000, At 12 weeks, there were no failure s among the AG group. The mean myelinated axon number was 14,563, and the number of myelinated axons ranged from 9,133 to 22,590 (n = 10). There were no obvious failures in the 12 week CT group, which had a mean of 4, 349 myelinated axons and counts that ranged from 1,272 to 9,842 (n = 10), and the 12 week MF group had a mean of 909 with a range from 0 to 4,789 (n = 10) The 12 week MT group had a mean myelinated axon count of 4,705 with count s that ranged from 24 to 12,905. In the CT group that went 24 weeks the mean myelinated axon count was 12,406 with a range from 1,756 to 22,515. The two highest counts in the 24 week CT group (22,515 and 21, 486) were almost identical to the 2 highest counts in the 12 week AG group (22,590 and 21,412). Althoug h the AG repairs more consistently contained greater num bers of regenerated axons at 12 weeks, the empty collagen tube was capable of not only sust aining regeneration across a 10 mm gap but at a longer time point (24 weeks for CT verses 12 weeks for AG) regeneration was similar to the autograft, based on myelinated axon number.
97 Nerve Guides (CT, MF, MT) at 6 Weeks Analyses of data included all specimens (i.e. specimens with 0 axon counts and extremely low counts, such as those with <100 axons). Myelinated axon counts from the mid se ction of the nerve guides in the 3 entubulation repair groups (CT, MF, MT) failed tests for no rmality (P=<0.001). Kruskal-Wallis One Way Analysis of Variance (ANOVA) on R anks (a nonparametric comparisons test for 3 or more groups) was performed on the results of the myelinated axon counts. The null hypothesis was that the axon counts were not drawn from populations with different medians. The alternative hypothesis was that the myelinated axon counts were drawn from populations with different medians. More specifically, the number of mye linated axons in the nerve guides with guidance channels would be gr eater than that of the em pty collagen nerve guide. We failed to reject the null hypothesis. There was no statistically significant difference in the number of myelinat ed axon between the three entubulation repair groups at six weeks (P = 0.488). Mann-Whitney Rank Sum Tests were also performed to compare the myelinated axon counts from each micr oguidance channel nerve guide repairs to the empty collagen tube (CT) independently. First, the six-week axon counts from the CT group were compared to the MT group. The difference in the median between the CT group and the MT was not statistically significant (P = 0.344). Next the CT group was compar ed to the MF group and there was no
98 statistically significant difference (P = 0.970) At 6 weeks there is not a significant difference in the number of myeli nated axons between the MT and MF group (Mann-Whitney Rank Sum Test, P = 0.307). Nerve Guides Verses Autografts at 6 Weeks Kruskal-Wallis One Way ANOVA on R anks followed by Dunns Method of Multiple Comparisons versus Control was used to test for signi ficant differences in myelinated axon counts from all entubulation nerve repair groups and the autografts at 6 weeks. The ANOVA on r anks revealed a statistically significant difference (P = 0.001). The multiple co mparisons test revealed statistically significant differences between each of the entubulation repairs and the autograft repairs at 6 weeks (P < 0.05). All the entubulation repairs with collagen nerve guides (CT, MF, MT) had significantly fewer myelinat ed axons than the AG group at 6 weeks. Nerve guides at 6 Weeks Verses Normal Control Kruskal-Wallis One Way Analysis of Variance on Ranks was used to test for significant differences in myel inated axon counts between the entubulation repair groups and the normal sciatic nerve control. There was a statistically significant difference between these repai r groups and the normal control (P = 0.001). Dunns method of Multiple Co mparisons versus the Normal Control group was performed to isolate the groups t hat differed significantly. Each of the
99 6 week entubulation repair groups (CT, MT, MF) had myelinated axon counts were significantly less than the normal control axon counts (P < 0.05). Nerve Guides (CT, MF, MT) at 12 Weeks As in the 6-week time point, the nu ll hypothesis was that there is no difference in the number of myelinated axons in the mid section of the nerve guide between the different entubulation repa irs. The alternative hypothesis was that the number of mye linated axons in the nerve guides with guidance channels would be greater than that of the empty collagen nerve guides. The myelinated axon count s from the mid section of the nerve guides in the 3, 12-week entubulation repair groups (C T, MF, MT) failed tests for normality (P = < 0.001). Kruskal-Wallis One Way Analysis of Variance (ANOVA) on Ranks was performed on the results of the my elinated axon counts. There was a statistically significant difference betw een groups (P = 0.002). Dunns Method of Multiple Comparisons procedure follow ed. Both the CT and MT groups have significantly more myelinated axons than the MF group at 12 weeks (P < 0.05). The mean myelinated axon number in the MT group is larger than the CT group but this difference was not statistically significant. Additionally, a Mann-Whitney test was run to compare the 12-week MT group to the 12-week MF group. The MT group had significantly more myeli nated axons than the MF group (P = 0.008).
100 Nerve Guides Verses Autografts at 12 Weeks Kruskal-Wallis One Way ANOVA on Ran ks was used to test for significant differences between the number of myeli nated axons in the mid section of the entubulation nerve guide repairs and the AG nerve repairs at 12 weeks. The differences in median values between grou ps are significantly difference (P = 0.001). Dunns Method of Multiple Com parisons versus Control was used to isolate differences between gr oups with the autograft assi gned as the control. All of the entubulation repairs (CT, MF, MT) h ad significantly fewer myelinated axons in the mid section of the repair si te than did the AG repairs (P < 0.05). Nerve guides at 12 Weeks verses Normal Control Kruskal-Wallis One Way ANOVA on Ran ks was used to test for significant differences between the number of myelinat ed axons in the all the entubulation nerve repairs at 12 weeks as compared to the normal control sciatic nerve. There were significant differences (P = < 0.001). Dunns Method of Multiple Comparisons verses Control Group followe d. Both the CT and MF groups had fewer myelinated axons than the normal control sciatic nerve (P < 0.05). Analysis of Groups Over Time There was a trend of increasing number of myelinated axons group over time. In order to test if a significant difference in groups existed over time
101 Kruskal-Wallis One Way ANOVA on Ran ks was performed on the CT groups of 6, 12, and 24 weeks. This was the only r epair type with a 24 week survival time (n=5). The difference in m edians was statistically signif icant (P = 0.002). Dunns Method of Multiple Compar isons revealed a statistically significant difference between the 6 week CT group and the 12 w eek CT group, the 6 week CT group and the 24 week CT group (P < 0.05). No statistically significant difference existed between the 12 week CT group and the 24 week CT group. Each nerve repair group was analyzed across the 6 week and 12 week time periods to see if there was a statis tically significant increase in the number of myelinated axons with in each group ov er time. The 12 week CT group had significantly more myelinated axons than the 6 week CT group (Mann-Whitney Rank Sum Test, P = 0.006). The 12 week MT group had significantly more myelinated axons than the 6 week MT group (Mann-Whitney Rank Sum Test, P = 0.008). The difference between the 6 and 12 week MF groups was not significant (P = 0.438). The 12 week AG group had significantly more myelinated axons than the 6 week AG group (t -test, P = 0.018). The Bonferroni approximation to correct for multiple comparisons was not indicated since results were reported for within group comparisons across time only. Nerve Repair Versus Normal Control Across Time Periods Both the 6 week and 12 week AG gr oups had significantly greater number of myelinated axons than the normal c ontrol (Kruskal-Wallis One Way ANOVA on
102 Ranks, P < 0.05). The only entubulation repair to have a mean number of myelinated axons greater than the normal control was the 24 week CT group. This difference, however, did not reach statistically significance (Mann-Whitney Rank Sum Test, P = 0.358). In other words, by 24 weeks the number of myelinated axons in the CT group was not significantly differ ent from the normal control. Post Hoc Analyses Both the data from the 6 week groups and from the 12 week groups were subjected to further analysis using different cut offs for inclusion in the statistics, based on the notion that incl uding the failures may somehow skew the results and cause us to miss an important trend. Cut offs of 100, 500, 1500, and 2000 axons were tested. Only nerve guides with axon counts greater than those cut offs were included in the statistical comparisons. There was essentially no change in the results when using variable cut offs for statistical analyses.
103 Myelinated Axon Diameters Images captured with the digital ca mera and 100x oil objective lens provided the resolution necessary for meas urements of axonal diameter. ImageJ software leaves a line and number designat ion on each measured axon (figure 56). This feature enabled us to be certain that each axon was measured but once. The toluidine blue stained, osmica ted, sections enabled us to readily discern myelinated from unmyelinated axons in regenerating nerves (figure 57). Tables 3 and 4 summarize the diamet er results, and they are displayed graphically in Figures 58 and 59. Figure 56 Axon Diameters of Normal Control Cropped image (100x oil objective lens) of normal nerve. The axonal diameter was measured from all the myelinated axons (black arrows) fr om digital images captured through the 100x oil objective lens excluding any axons that are not completely within the field, and any axons that have double myelin rings (white arrows), which oc cur at Schmidt-Lanterman clefts. Non-circular and distorted axons were measured across their shor test diameter to avoid inflating estimates of axon size.
104 Figure 57 Regenerated Nerve, Myelinated versus Unmyelinated Axons Cropped image (100x oil objective) from a nerve guide specimen (6 week MT group), which shows the contrast of myelinat ed (black stars) and unmyelinated (white stars) axons. This contrast was sufficient for counting and meas uring myelinated axons in regenerated nerve sections.
105 Myelinated Axon Diameters at 6 Weeks0% 10% 20% 30% 40% 50% 60%0.00.9 1.01.9 2.02.9 3.03.9 4.04.9 5.05.9 6.06.9 7.07.9 8.08.9 9.09.9 10.0+MicrometersPercent of Total Normal AG CT MT MF Figure 58 Histogram of Myelinated Axon Diameters at 6 Weeks Myelinated Axon Diameters at 12 Weeks0% 10% 20% 30% 40% 50% 60%0.00.9 1.01.9 2.02.9 3.03.9 4.04.9 5.05.9 6.06.9 7.07.9 8.08.9 9.09.9 10.0+MicrometersPercent of Total Normal AG CT MT MF Figure 59 Histogram of Myelinated Axon Diameters at 12 Weeks
106Table 3 Myelinated Axon Diameters at 6 Weeks Micrometers Normal AG CT MT MF 0.0-0.9 5% 28% 30% 27% 51% 1.0-1.9 19% 46% 44% 48% 36% 2.0-2.9 17% 20% 18% 19% 9% 3.0-3.9 15% 4% 7% 5% 3% 4.0-4.9 16% 1% 1% 1% 0% 5.0-5.9 12% 0% 0% 0% 0% 6.0-6.9 8% 0% 0% 0% 0% 7.0-7.9 5% 0% 0% 0% 0% 8.0-8.9 2% 0% 0% 0% 0% 9.0-9.9 1% 0% 0% 0% 0% 10.0+ 0% 0% 0% 0% 0% Mean 3.83 m 1.53 m 1. 54 m 1.56 m 1.21 m Table 4 Myelinated Axon Diameters at 12 Weeks Micrometers Normal AG CT MT MF 0.0 0.9 4.9% 38.1% 28.1% 38.6% 57% 1.01 .9 18.7% 40.3% 47.0% 39.1% 34% 2.0-2.9 16.7% 14.7% 16.9% 15.1% 7% 3.0-3.9 15.2% 4.3% 4.8% 4.7% 2% 4.0-4.9 15.8% 1.5% 1.9% 1.7% 0% 5.0-5.9 12.3% 0.7% 0.7% 0.7% 0% 6.0-6.9 8.3% 0.3% 0.0% 0.1% 0% 7.0-7.9 5.2% 0.0% 0.0% 0.0% 0% 8.0-8.9 2.1% 0.0% 0.0% 0.0% 0% 9.0-9.9 0.6% 0.0% 0.0% 0.0% 0 % 10.0+ 0.2% 0.0% 0.0% 0.0% 0% Mean 3.83 m 1.40 m 1. 56 m 1.42 m 0.99 m Regenerated nerves in all repair groups at both the 6 and 12 week time points, contained myelinated axons with a mean diameter significantly less than the normal control sciatic nerve (P < 0.05, multiple comparisons versus control, Dunnetts method). There was no signifi cant difference in the mean axon diameter between repair groups at the 6 we ek time point (P = 0.298, power is
107 0.100, One Way ANOVA) or 12 week time point (P = 0.063, power is 0.402, One Way ANOVA). All the nerve repair groups had sim ilar percentages myelinated axons in the each diameter range (Figures 58, 59). At least 70% of the myelinated axons in the 12-week repair groups were less t han 2.0 m (Figure 59). This was in contrast to normal nerve, which has less than 24% of myelinated axons less than 2.0 m. Each repair group except the MF group has less than 1% of myelinated axons greater than 5.0 m at 12 wee ks post repair. The MF group had no myelinated axons 4.0 m or greater in diameter. In the CT group, there was an increase in mean myelinated axon number over time, as previously stated; however, the percentage of axons in each di ameter range is similar (Figure 60). The exception was at 24 weeks, there were 0.3% of the myeli nated axons in the 6.0 to 6.9 m range while at earlier time points all axons were smaller in diameter than 6.0 m in the CT repair group. The AG had 0.3% of the myelinated axons in the 6.0 to 6.9 m at the 12-week time point.
108 Myelinated Axon Diameters Over Time In CT Group0.0% 5.0% 10.0% 15.0% 20.0% 25.0% 30.0% 35.0% 40.0% 45.0% 50.0%0.0-0.91.0-1.92.0-2.93.0-3.94.0-4.95.0-5.96.0-6.97.0-7.98.0-8.99.0-9.910.0+MicrometersPercent Normal 6 wk 12 wk 24 wk Figure 60 Myelinated Axon Diameters Over Time CT Group Table 5 Myelinated Axon Diameters Over Time CT Group Diameter Normal Control 6 wk CT 12 wk CT 24 wk CT 0.0-0.9 4.9% 30.4% 28.1% 43.0% 1.0-1.9 18.7% 43.9% 47.0% 37.9% 2.0-2.9 16.7% 17.8% 16.9% 12.5% 3.0-3.9 15.2% 6.8% 4.8% 4.0% 4.0-4.9 15.8% 1.0% 1.9% 1.9% 5.0-5.9 12.3% 0.1% 0.7% 0.4% 6.0-6.9 8.3% 0.0% 0.0% 0.3% 7.0-7.9 5.2% 0.0% 0.0% 0.0% 8.0-8.9 2.1% 0.0% 0.0% 0.0% 9.0-9.9 0.6% 0.0% 0.0% 0.0% 10.0+ 0.2% 0.0% 0.0% 0.0%
109 Electrophysiology Axons grew across the repair site in specimens from each group as evidenced by the recording of extracellu lar nerve action potential recordings of the sciatic nerve distal to the repair. In addition, there was electrophysiological evidence from each group that axons not only crossed the repair site but also grew back to reinnervate intrinsic muscles of the foot. Nerve conduction velocity across the repairs site was calculat ed in several specimens and found to be significantly lower than normal. The pr esence of a clean CMAP shift following a change in stimulus location, which is necessary for calculation of NCV, in general, was more consistently recorded from animals that received autograft repair than entubulation repair As stated previously, several elec trophysiological evaluation methods were used to determine return of function in different nerve repair groups. Below, are representative data from each method along with the results of experiments that allowed us to understand how errant action potentials cause interference patterns during gastrocnemius muscle recordings. Compound Nerve Action Potentials Figures 61 and 62 are images of the computer screen from electrophysiological experiments in which evoked CNAPs were recorded with two sets of hook electrodes placed on either si de of the repair site. In each pair of images, the top image was deriv ed from the sciatic nerve 12 weeks post repair,
110 and the bottom image was derived from the normal contralateral sciatic nerve. These images demonstrate conduction of nerve action potentials across the repair site. In each exam ple, the magnitude of the action potential from the repaired nerve was less t han that of the contrala teral normal nerve. Unfortunately, in many cases, further analysis was not possibl e due to the size and proximity of the stimul us artifact, which obscured a portion of the CNAP. Figure 61 Compound Nerve Action Potentials CT Group Representative Electrophysiological Trace. (A) No rmal sciatic nerve extracellular action potential tracing. (B) Recording from nerve repaired with empty nerve guide (CT) 12 weeks prior. The stimulus artifact is large (solid arrow) and obscures a portion of the nerve action potential (open arrow). This specimen had 9,384 myelinated axon s counted from the middle of the repair site.
111 Figure 62 Compound Nerve Action Potentials MT Group Representative Electrophysiological Trace. (A) No rmal sciatic nerve extracellular action potential tracing. (B) Recording from a nerve repaired with a MT nerve guide 12 weeks prior. The stimulus artifact is large (solid arrow) and obscures a portion of the nerve action potential (open arrow). This specimen had 4,521 myelinated axons count ed from the middle of the repair site. Compound Muscle Action Potentials and Anatomy Figures 63, 64, and 65 illustrate t he results from two experiments performed on a normal rat to determine the electrophysiological contribution to
112 the gastrocnemius muscle recordings of a lar ge branch of the tibial division of the sciatic nerve, which we now know i nnervates the biceps femoris muscle (accessory head, anterior head & posteri or head). Two experiments were performed and are described below. It is noteworthy that t he biceps femoris muscle is superficial to and complete ly covers the lateral head of the gastrocnemius in the Lewis rat. Thes e experiments demonstrated that when uncoated needle recording electrodes were inserted percutaneously into the normal gastrocnemius muscle, by passing the needle through the biceps femoris, and the sciatic nerve was stimulated in gl uteal region, the recorded CMAPs were a summation of both the gastrocnemius and the biceps femoris. Stated differently, the distinct gastrocnemius CMAP was obscured by combining with biceps femoris CMAP. The CMAPs from both muscles are in phase because the stimuli reaching each muscle were traveling through a normal nerve, which have approximately the same conduction velocity. Figure 63 shows the results of exper iment 1 (normal left hind limb), in which the sciatic nerve was stimulated wit h the platinum wire electrode in the exposed gluteal region. Recordings we re taken from the gastrocnemius muscle and 3d metatarsal space prior to and afte r transection of the nerve branching from the tibial division of the sciatic ner ve, proximal to where the nerve repair site would be in an operated rat, dur ing stimulation of the sciat ic nerve in the gluteal region. The amplitude of the gast rocnemius CMAPs dropped by approximately half when the sciatic nerve branch to t he biceps femoris was cut while the
113 amplitude of the CMAPs recorded from t he 3d metatarsal space did not change in magnitude. This result indicated t hat the muscle innervated by this sciatic nerve branch was contributing to the magnitude of the am plitude of the electrodes that were targeted for the gas trocnemius muscle. Recordings from muscle in the foot were not affected by the nerve that branched from the sciatic nerve proximal to the precis e site that is transected a nd repaired in the animals in this study; they were innervated by ner ves that branched from the sciatic nerve distal to the repair site. Next, the sciatic nerve was transected in the same location that is used in the rats that receive entubulation repair. The nerve was stimulated and CMAPs were recorded after transection of the sciatic nerve Experiment 2 (normal right hind limb): Recordings were taken from the gastrocnemius muscle and from the 3d meta tarsal space during stimulation of the sciatic nerve in the exposed gluteal r egion. The sciatic nerve was transected at the same location as in animals that entubulation repair was performed. This left the nerve branching from t he tibial division of the sciatic nerve, intact since it was proximal to the lesion site. When the sciatic nerve was stimulated in the gluteal region, the gastrocnemius recordi ng electrodes record a large CMAP, but the electrodes in the 3d metatarsal space record none. If the electrodes that were set up to record gastrocnemius CM APs was only recording potentials from the gastrocnemius muscle there should have been no CMAPs following transection of the sciatic nerve. Ou r conclusion is, the CMAP must have been
114 recording from another muscle. The likel y contributor is the biceps femoris, which lies superficial to the gastrocnemius on its lateral aspect. It is this muscle that the uncoated electrode passed through in order to reach the gastrocnemius. When the nerve branching from the tibial division of the sciatic nerve is subsequently cut, there is no longer a re corded CMAP from this electrode set up. The results of these two experimen ts systematically demonstrate the contribution of the proximal tibial divi sion of the sciatic nerve to the biceps femoris muscle in the normal rat dur ing percutaneous bare needle electrode recordings of the gastrocnemius with the G2 electrode placed subcutaneous in the region of the gastrocnemius and sciat ic nerve stimulation in the gluteal region.
115 Figure 63 Normal Gastrocnemius CMAP & Biceps Femoris Experiment #1 Normal Rat Left Hindlimb. Overla y of CMAPS prior to (solid arrow) and after (open arrow) transection of proximal tibial divi sion of the sciatic nerve with percutaneous needle electrode insertion to gastrocnemius muscle. Pr ior to cutting tibial nerve: the CMAP has a peak of 23.28mV occurring at 5.15ms, after the ner ve is cut the CMAP peak is 12.95mV peak at 5.30ms. Notice the deviation from baseline (3.4ms) and the peak of the CMAP under both conditions occurs at approximately the same time (latency) in these normally innervated muscles. There is not a similar change in peak amplitude on lower channel (3d metatarsal space) because the tibial nerve, a terminal division of the sciatic nerve originates distal to the region used for the entubulation repair groups, continues as the late ral plantar nerve as it enters the foot and innervates the musculature of the foot
116 Figure 64 Gastrocnemius CMAP Cut Sciatic Nerve (A) CMAP of normal gastrocnemius (arrow) top tr ace) and 3d metatarsal space (bottom trace) after transection of the proximal tibial division of the sciatic nerve. The arrow head points to the stimulus artifact. (B) Absence of CMAPs on both channels recorded immediately after transection of the sciatic nerve at the site us ed in animals that receive entubulation repair. Figure 65 Electrophysiology Experiment Two The Sciatic nerve is cut first, and then the nerve to the biceps femoris is cut in a normal Lewis rat. (A) Before cutting sciatic nerve, a CMAP (top chan nel) is recorded from gastrocnemius electrodes and from the 3d metatarsal space (bottom channel). (B) After sciatic nerve is cut (distal to branching of the nerve to the bi ceps femoris muscle). Note ther e is still a recorded CMAP on the top channel even after sciatic n. is cut, however there is no CMAP recorded on bottom channel (3d metatarsal space.) (C) Recordings taken afte r the nerve to the biceps femoris is cut showing that are no longer CMAPs.
117 Figure 66 Anatomical Study & Nerve Branch to Biceps Femoris Rat sciatic nerve in gluteal region and posterior th igh (solid arrow) displaying a large nerve (open arrow) that branches from its tibial division. This nerve innervates the biceps femoris. Figure 67 Anatomical Study Rat Hindlimb Musculature This Figure shows the relationship of the gastr ocnemius (solid black arrow) and the biceps femoris (open arrow). The biceps femoris muscle is transected and reflected to display the gastrocnemius. Prior to the dissection, the biceps femoris completely covered the lateral head of the gastrocnemius. The gastrocnemius muscle (lateral and medial heads) are innervated by muscular branches (white arro w) of the tibial nerve.
118 Figure 68 CMAPs From Gastrocnemius & Biceps femoris These recordings are from the specimen from t he 12week CT group with the greatest number of myelinated axons (9,842) in the nerve guide midsection. CMAPs recorded from the gastrocnemius prior to (A) and after (B) cutting the nerve to biceps femoris. (A) CMAPs recorded from percutaneous needle electrodes in gastrocne mius muscle image prior to nerve transection (showing 2 peaks: 1st (solid arrow) with a latency to p eak of 3.124ms, 4.58mV max, amplitude 9.986mV. Second peak (open arrow) in the top im age has a latency of 6.3, 3.36mV max, 4.52mV Amplitude (Max-Min) (B) is a CMAP after nerve to biceps femoris was cut. Latency to peak is 6.2ms, 2.75mV max, 4.13mV amplitude. (C) Is the overlay of both recordings (A) and (B) showing that the second peak is from the gastrocnem ius muscle. The regenerated axons have reinervated the muscle and the latency is longer relative to the peak from the biceps femoris peak.
119 Figure 69 Gastrocnemius CMAP and G2 Electrode Position This Figure illustrates the affect of G2 elec trode placement on amplitude in a specimen from the 12 week CT group that contained the greatest numbe r of myelinated axons (9,842) in the nerve guide midsection. It is an overlay of several CMAPs recorded with the G2 electrode placed subcutaneously over the gastrocnemius (sma ller amplitude, open arrow), and several CMAPs recorded with the G2 electrode place subcutaneously over the lateral malleolus (larger amplitude, solid arrow). Figure 70 Shift of CMAP Latency From S1 to S2 CMAPs recorded from the gastrocnemius muscle of a specimen from the 12 week CT group. The shift from S1 to S2 shows the decrease in latency when the stimulating electrode is moved from the sciatic nerve in the gluteal region (S1) to the sciatic nerve in the thigh (S2) distal to the nerve repair. Measurements of the inter-electr ode distances were recorded and used to calculate conduction velocity across the repair site. S1 latency to peak = 6.28ms, S1 to G1 distance = 51mm. S2 latency to peak = 5.125ms, S2 to G1 distance = 29mm. NCV = 19.1 m/s. This specimen had 9,842 myelinated axons in the midsection of the repair.
120 Table 6 Nerve Conduction Velocities 12 Weeks Post Repair Normal Control AG CT MF MT 52 m/s 8 (n=7) 20 m/s 6 (n=2) 17 m/s 2 (n=3) Figure 71 CMAP in 1st Metatarsal Space in 12 Week MT Group On the top channel the 1st large CMAP (open arrow) is likely fr om the normally innervated biceps femoris and was recorded unintentionally. One or both of the later peaks (striped arrows) on this same channel are likely CMAPs recorded from the gastrocnemius muscle. Their latencys to peak amplitude (8.5 & 11.4ms) and the magnitude of their amplitudes are more indicative of CMAPs from immature regenerated nerve. Init ially the recording electrode was in the 3d metatarsal space, and no CMAP was recorded upon st imulation of the sciatic nerve in the gluteal region. Because of the observation of active 1st toe abduction/adduction by this animal when it was in the static stance photography box, a needle electrode was inserted into the 1st metarsal space and a CMAP (solid black arrow on bottom cha nnel) was successfully recorded. This is the rat from the MT group with the greatest number of myelinated axons (12, 905) in the mid-section of the repair.
121 Figure 72 Late CMAP in Foot R86 12 wk MF group had only 103 myelinated axon s counted in mid-section of nerve guide yet there is a CMAP (solid black arrow) in the foot. It has a long latency (36.4ms) indicative of poorly myelinated axons, which have reduced conduction velo cities relative to normal. The CMAP is of small amplitude (58.4V) indicating few and small motor units, which is in turn indicative of few axons reaching and forming functional synaptic c onnections with this muscle group. The top channel scale (open arrow pointing to large CMAP) is in millivolts where as the bottom is in microvolts to enable viewing this small amplitude CMAP recorded from the foot. Motor Unit Number Estimation (MUNE) Recordings were taken in manner conducive to MUNE analysis, however this method was discontinued and the anal ysis was not performed. The reasons for this are elaborated upon in the discussi on. Figure 74 is simply provided to demonstrate how the examiner would di splay recorded CMAPs prior to the analysis described in the methods section.
122 Figure 73 Motor Unit Number Estimation (MUNE) Overlay of multiple compound muscle action poten tials (CMAPs) recorded following electrical stimulation (in incremental intensit ies) of the sciatic nerve proximal to the repair site. The traces on the top channel were recorded with monopolar needle electrodes inserted into the gastrocnemius of an animal in the 12-week MF gr oup. The trace in the lower portion of the Figure shows small CMAPs in the intrinsic muscles of the foot in the same animal, indicating successful reinnervation to the intrinsic muscles of the foot. Functional Assessments Sciatic Function Index At 12 weeks there is a significant di fference in the SFI between groups (P = 0.040), however the power of the test is only 0.484, which is well below the desired value of 0.800. T here is no significant difference among these groups (P
123 < 0.050) when pairwise multiple com parisons procedures are performed (Student-Newman-Keuls Method). Table 7 Mean Sciatic Function Index From Walking Tracks AG CT MT MF SFI prior to nerve injury/repair -5.6 -8.6 -7.79 -7.67 SFI at wk 1 -84.22 -83.69 -80.55 -80.62 SFI at wk 6 -82.32 -79.85 -76.69 -82.19 SFI at wk 12 -78.17 -77.06 -78.57 -86.60 Figure 74 Normal & 1 Week Walking Tracks (A) In the normal rat each toe leaves a distinct ink print. The toes are abducted and the body weight is distributed among all of the toes duri ng walking. (B) Walking track from an animal 1 week after nerve injury and repair shows a cons istent full print length and a reduced toe spread relative to normal. Distinct toe prints cannot be delineated because the ink runs together when the animals toes are in full adduction.
124 Figure 75 Walking Track From Rat With Flexed Toes (A) Walking tracks from a rat from 12 weeks a fter autograft repair. (B) Magnified view of enclosed region in (A). This rat maintained its toes in full flexion both in static standing and during walking. Note that toe spread cannot be measured from the operated hindlimb and reduced print length observed is an artifact due to the toes being curled under the forefoot. This specimen had 13,715 myelinated axons in the middle of the repair site. Figure 76 Walking Track From CT Group at 12 Weeks (B) & (C) are from the enclosed regions in (A). This particular specimen had 9842 myelinated axons in the middle of the repair site, indicating a good probability of functional return, yet this is difficult to gauge from these prints because the ink is not uniformly distributed on the paper.
125 Static Stance Analysis The static stance box allo wed us to observe the functional result of nerve injury and regeneration from a unique v antage point. Because of noting 1st toe abduction in one particular specimen (Figure 79) we decided to modify the recording of CMAPs in t he feet to include the 1st and 2nd metatarsal spaces instead of only the 3d metatarsal space. The results of the SSI are summarized in table 8.
126 Figure 77 Setting Scale For Toe Spread Measuremens With ImageJ Screen captures of digital image taken from st atic stance photography box and imported into ImageJ software. This rat is one from the CT group that went 24 weeks. (A ) Shows calibration of scale from metric scaled ruler, which is inclu ded in the image. A line is drawn across a 50 mm (encircled region) of the ruler and entered as a kn own distance (arrow) that equals the computer derived distance of 346.00 pixels. (B ) Shows measurement of the normal 1-5th toes spread, (C) demonstrates measurement of intermediate toe spread in the operated limb. Table 8 Mean Static Stance Index (Percent of Normal Toe Spread) AG CT MT MF 6wk 51.20 43.60 39.32 37.45 12wk 47.32 61.61 42.36 37.97 24wk 61.43
127 Figure 78 Typical Foot Posture 1 Week After Nerve Injury & Repair Unoperated foot shows normal abducted positio n of each digit, where as the operated foot (arrow) shows foot posture without toes in addu ction at 1-week post nerve injury repair. Figure 79 Static Stance Photo Movement of Digit Two images of the same rat taken moments apar t. This rat had sciatic nerve repair with a guide tube with guidance channels 12 weeks prior to these images being captured. The image on the left shows the1st toe (arrow) is not abducted. Image on the right shows the 1st toe (arrow) in abduction. This rat actively abducted its toe when it stood on its hind limbs. This movement widened the animals base of support to fa cilitate balance, demonstrating a coordinated movement from a reinnervated muscle group.
128 Figure 80 Digital Flexion Abnormality in Operated Limb Example of an animal with toe flexion abnormality in the operated limb (AG group at 12 weeks). Images on bottom of Figure are close up views of the image above, (A & C) the rat is in quadrupedal stance, (B & D) it is standing on it s hind limbs only. This foot posture prevents accurate toe spread measurements from either wa lking tracks or static stance photos. This particular rat had 13,715 myelinated axons in t he middle of the autograft at 12 weeks. Figure 81 Autograft With Return Of Toes Spread At 12 Weeks All images from same animal, (A) & (C) in quadrupedal stance, (B) & (D) in bipedal stance. This rat received autograft repair 12 weeks prior to these images being taken and shows excellent return of toe abduction.
129 Figure 82 Toes Spread Return (MT Repair Group at 12 Weeks) All images from same animal, (A) & (C) in quadrupedal stance, (B) & (D) in bipedal stance. This rat received MT nerve guide 12 weeks prior to thes e images being taken. There is evidence of return of toe spread that increases somewhat when the animal transfers more weight onto the hind limbs in bipedal stance. This particular rat had 12,905 myelinated axons in the mid-section of the nerve guide at this time point. This was the highest axon number in the MT group. The five rats in the 24-week CT group were chosen to examine if a relationship existed between the number of axons counted and the score from the SSI using 1-5th toes spread measurements. Amongst these five animals there appeared to be was no correlation (summa rized in table 9 and Figure 83). The specimen with the highest number of myelinated ax ons had flexion deformities of all the digits of the expe rimental hindfoot (Figure 80), which made accurate measurement of toes spread impo ssible. The specimen with the fewest myelinated axons had the largest 1-5th toe spread. This large toe spread may have represented contracture and not acti ve toe abduction. The three rats with
130 intermediate axon numbers demonstrat ed no correlation with the toe spread measured from stat ic stance images. Table 9 24 Week CT Static Stance Index, Axon Number, Foot CMAP Rat# Group Weeks Axons % normal 1-5th TS CMAP in foot 1 CT 24 22,515 Deformity Yes 2 CT 24 21,486 53.04 Yes 3 CT 24 9,812 32.03 Yes 4 CT 24 6,459 63.62 Yes 5 CT 24 1,756 97.04 Yes Figure 83 SSI Plotted Against Axon Number
131 Chapter 5: Discussion Autografts are the preferred method of repair for gap injuries and our data verified their efficacy. In the pres ent study, the mean num ber of myelinated axons in the middle of the autograft incr eased by 45% from t he 6 to the 12 week time point, and at both time points the myelinated axon number was much greater than the normal contro l. These data are in contra st to Chamberlain et al., who, using similar histological as sessment methods, found no significant difference between myelinated axon counts at 6 weeks, 30 weeks, and 60 weeks in Lewis rats that underwent a 10 mm aut ograft repair (Chamberlain, Yannas et al. 1998; Chamberlain, Yannas et al. 1998). In that study, the orientation of the autograft was not reversed, unlike the pr esent study, because of the desire to optimize fascicular alignment. This may account for their autograft reaching a greater number of my elinated axons than ours at t he 6 week time point and then reaching a plateau. In newly regenerated peripheral nerve many more axons occupy the nerve segment distal to the injury site than occupy the nerve segment proximal to the injury site. This is true for both aut ograft and entubulation repair of peripheral nerve and occurs because the parent axon br anches into several neurites, each of which may elongate and take a different path (Dohm, Streppel et al. 2000). If
132 a neurite reestablishes an appropriate perip heral contact, in time, the axons supernumerary branches are eliminat ed (Mackinnon, Dellon et al. 1991). The autografts in this study, and prev ious animal studies, represent an idealized nerve graft in many ways. They are not directly comparable to clinical autografts. The size of the autograft is a perfect match in these studies, and a second surgical site is unnecessary to obtai n the graft. Unlike the clinical nerve autograft, which may reside out of the body for 90 minutes prior to implantation, our autografts do not leave the surgical si te. The presumption is that in our autograft model the majority of Schwann ce lls in the autograft survive, and in the clinical autograft the majority of Schwann cells die. It is the presence of live Schwann cells that give the autografts in this study a major advantage over the entubulation repair groups. Schwann cells provide an endogenous source of growth factors, NGF, BDNF, CNTF, which enhance both neuronal survival and growth (Ho, Coan et al. 1998; Lee, Yu et al. 2003). Several studies of nerve gap repair have shown that living Schwann cell laden grafts are superior to acellular grafts at promoting nerve regeneration (Evans 2000; Hadlock, Sundback et al. 2001). Although the present design of micr otube or filament guidance channels provide greater surface area for cell mi gration than empty tubes they cannot compare to the much greater surface area provided by an autograft. The remnant basal lamina Schwann cell tubes of the autograft not only provide a much greater surface area t han the fabricated nerve guides provide, but they are
133 the ideal substrate for neur ite growth cone adhesion an d elongation. Even when an autograft is repeatedly frozen and thawed, killing the resident Schwann cells, the basal lamina supports axon growth and Schwann cell migratio n (Ide 1983). The basal lamina contains laminin, ty pe collagen IV, fibronectin, and heparin sulfate proteoglycans (Laur ie, Leblond et al. 1982), and each of these macro molecules has a proven ability to promot e neurite outgrowth from dorsal root ganglia, in vitro in the presence of nerve growth factor (Chernousov, Stahl et al. 2001). Moreover, these molecules enhance regeneration across nerve gaps when they are added to the lumens of ner ve guides (Madison, Da Silva et al. 1988; Woolley, Hollowell et al. 1990; Zhang, Oswald et al. 2003). Schwann cells that have lost contact with axons, like those in the autograft at the time of initial implantation, undergo prolifer ation. Proliferation and migration is far greater when Schwann ce lls adhere to laminin or fibronectin as compared to type I collagen, and both of these molecules are present in the Schwann cell basal laminae of th e autograft (Ahmed and Brown 1999) (Vleggeert-Lankamp, Pego et al. 2004). When fibronectin is manipulated into a fibrillar macromolecule, it accelerates Schwann cell migration and directional guidance via alignment and bi nding of cytoskeletal F-ac tin filaments (Ahmed and Brown 1999). When autografts cannot be used, a biomaterial graft may be a good alternative (Keeley, Atagi et al. 1993). In addition to avoiding the comorbidity of the harvest site necessitated by aut ograft repair, entubul ation repair using
134 biomaterial tubes offers a mechanical feature, a cuff to surround the nerve stumps, which appears to decrease the ax onal escape that can occur at the autograft suture line. The composition and architecture of the biomaterial guide should contribute to the degree of regeneration and functional recovery (Madison, Da Silva et al. 1988). Many design cues can be derived by analyzing features of the autograft. An overall goal of the projec t (phases I, II, and III) was to systematically determine the effect s of the nerve guide parameters of permeability, architecture, adhesive molecu les, and growth promoting molecules on regeneration following entubu lation repair of a peripheral nerve gap injury, in order to build the final prototype for repair of a median nerve lesion in the monkey. The focus of the present study was studying the effects of different nerve guide architectures. The present study demonstrated that axons grow through type I collagen nerve guides that contain guidance channel s, as evidenced by the histological data. Regenerating axons reestablish synap tic contact with peripheral targets as evidenced by the presence of evoked gastrocnemius muscle CMAPs. There is return of motor function in the limb that underwent nerve repair as demonstrated by the partial return of toe spread in some animals. Although the data demonstrated that nerve guides with longitudinal channels could be used for the repair of peripheral nerve gap injuries, t hey were not significantly better than empty tubes.
135 Heparin was incorporated into the ner ve guides to provide a binding site for bFGF and laminin for the Phase III rat in vivo study on the effect of bioactive molecules on nerve regeneration. To ensure that inclusion of exogenous bioactive molecules were the only variables between phase II and phase III, heparin was incorporated into the type I collagen matrix of the nerve guides fabricated for this study ( phase II, guidance channels). The effect on nerve regeneration of type I collagen (5.0% heparin) nerve guide as compared to a pure type I collagen nerve guide was not tested in this study. Although it is known that specific heparin binding pept ides benefit aspects of nerve regeneration, it is unclear how heparin alone effects regeneration in collagen nerve guides. The presenc e of heparin may be beneficial to regeneration in that if ma y temporarily bind soluble peptide growth factors released by Schwann and mast ce lls, and aid nerve regeneration by concentrating these factors, which may othe rwise diffuse out of the tubular nerve guide. For FGF to bind to its high affinity re ceptor it must first be bound to either heparin or heparan sulfate, and FGF-heparin complex is more resistant to degradation than FGF alone (Tyrre ll, Ishihara et al. 1993; Is hihara, Shaklee et al. 1994). We are not the first to use heparin as a delivery molecule. Recently, heparin was used successfully to bind and deliver a variety of peptides, both growth factors and adhesive molecules, to enhance embryonic chick dorsal root ganglia neurite outgrowth in a three-dim ensional matrix (Sakiyama, Schense et
136 al. 1999). It also has been used to improve regeneration across a 13 mm rat sciatic nerve gap (Lee, Yu et al. 2003) Myelinated Axon Number At the 12 week time point recovery of MT group was clearly superior to the MF group. Although the m ean number of myelinated axons in the mid section of the nerve guide was higher in the MT gr oup (4705) than the CT group (4349) this difference did not reach statis tical significance. In fact, 80% of the CT group had axon counts greater than 2000 compared to only 67% in the MT group. None of the rats from the 12 week CT group had axon counts greater than 10,000 (n = 10) while in the MT group 1 rat had a myelinated axon count of 11,923, and another had a count of 12,905. These num bers are within the range of those observed for the AG group. If endogenous fibrin matrix from the w ound site fills the lumen of both a nerve guide with microtubes and one withou t microtubes there may be a wash out effect of the greate r surface area provided by the collagen microtubes. In type I collagen/heparin composite nerve gui des, Schwann cells migrate into the graft ahead of neurite growth cones, and lay down an extracellular matrix that the growth cones can more easily advanc e upon (Madison and Archibald 1994). Schwann cells may migrate through the fibr in matrix and not see the specific advantage of the surface area provided by the biomateria l nerve guide. It may be that the main advantage of a nerve guide with type I collagen guide tubes is
137 structural stability and not su rface area for cell migration, in that it may prevent collapse of the guide tube dur ing the early critical r egeneration phase. Although the empty nerve guides may have appeared more flattened than the guides with microtubes, they did not collapse to a degr ee that their lumens were obliterated and therefore the benefit of the microtubes could not be revealed over a gap length of only 10 mm. Perhaps, they will provide more obvi ous benefit to a longer gap repair, where an empty guide tube may be more susc eptible to kinking. This may be especially true if nerve repair is perform ed near or across joints. The elbow provides an example of a vulnerable peripheral nerve site (Mackinnon and Dellon 1988; Hunter, Schneider et al. 1997). The ulnar nerve at the elbow has no muscular covering and lies in the boney groove between the medial epicondyle of the humerus and the olecrano n of the ulna. If the ulnar nerve is injured at this site and is subsequently repaired, it ma y benefit from a nerve guide that has specific structural architecture to prevent lumen collapse. Under these conditions, the true benefits of this architecture may rev eal itself. This is unlike the present model where the nerve guide is placed in the rats mid thigh, away from forces that would cause it to kink and obliterate the lumen. This location is also well protected by t he overlying hamstring muscles unlike the superficial location of the ulnar nerve at the elbow. In the we ll protected sciatic nerve environment of the rat mid-th igh, factors other than architecture, such as growth
138 factors and adhesive substrate likely ar e the predominant factors influencing successful regeneration ac ross a nerve deficit. Variability The nerve guides rapidly absorb saline, which is applied to rehydrate the nerve guides prior to implantation. This c auses swelling, which in the case of the microtubes appears to press them against each other and the outer tube in a manner that is beneficial in that it prevents the microt ubes from moving laterally and the patency of the tubes is maintain ed. The collagen f ilaments however, swell, aggregate, and do not appear to offe r a benefit to maintaining the patency of the nerve guide lumen. There was considerably less within group variability in the results of the myelinated axon counts in the CT group when compared to either nerve guide group with guidance channels. Perhaps more cumulative variability exists with the more elements that are added to a nerve guide. Since there were more parts in the nerve guides with guidance channel s as compared to the empty nerve guides (32 more parts in the MF group, and 5 more in the MT group) there is likely greater variability in the nerve gui des before they are ever put in the animals, variability that exists simply due to the manufacturing process. In this study, every attempt wa s made to minimize manufacturing variability by utilizing strict standards ( outlined in the methods section), which
139 included using nerve guides from the sa me lot (manufactured on the same day without altering machine calibration settings). Another reason for variability in the re sults of the axon counts from nerve guides with guidance channels may be the movement of the guidance channels within the lumen of the outer nerve guide. This is not tightly controlled once the nerve guides have been implanted. Several times during the initial surgeries, immediately prior to implantat ion, filaments or microtubes partially slid out of the outer nerve guide. We were able to sli de them back in without difficulty prior to completion of the surgeries. No nerve guides in which microt ubes or filaments slid completely out during preparation fo r implantation were used in the study. They were discarded immediately and a st erile replacement was used instead. If the proximal and distal sciatic nerve stumps are not sutured so that they are in direct contact with the guidanc e channels (MT or MF), longitudinal movement of the guidance chan nels could theoretically o ccur. If longitudinal movement of the guidance c hannels did occur, it could very well be non-uniform. Some filaments or microtubes may slip proximally while others within the same nerve guide would occupy a position more distal in the nerve guide. Although this is theoretically possible, every attempt was made to prevent this from occurring during this study. Since the outer nerve guide becomes translucent when hydrated, it is possible to see t he position of the nerve stump within the outer nerve guide; we were thus able to ensure the nerve end contacted the guidance channels prior to securing it with suture. Additionally, the 10 mm
140 intraluminal nerve gap was confirmed by measurement with a vernier caliper via the translucent outer nerve guide. The only movement that is poorly contro lled is the lateral movement of the solid collagen filaments. The filaments fail to remain uniformly distributed within the lumen of the outer nerve guide. In several specimens, 32 individual filaments could be identified while in many others the filaments ag gregated, thus occluding the spaces between the filaments and obl iterating the theoretical guidance channels. The microtubes appear to remain in their position wi th respect to the outer guide tube better t han the filaments. We were unable to calculate correlation coefficients between the myelinated axon counts and electrophysi ology results due to insufficient sampling with electrophysiology. Many of the inferential statistical analysis suffered from limited power due to both small sample size, non-equal variance within groups, and lack of normality. It ma y be that a larger sample size would have solved some of these problems. Al though there was a larger than expected variance in the axon counts it is unlikely due to surgical error. Since in the combined AG groups there was only a 5% (1 out of 20) failure rate and in the 12 week CT group all rats had axon counts greater than 1200, it is fa ir to state that the variability resulted from the incl usion of type 1 co llagen filaments or microtubes in the nerve guides. It cannot be overstated that a det ailed understanding of the gross anatomy of the animal that is used for the model is as essential as the microanatomy.
141 Some shortcomings of this study came about because of a less than perfect understanding of the rat anatom y at the onset of the study. This affected the electrophysiological assessment in particular. Investigators that studied nerve regeneration in the rodent have often found that findings of electrophysiology, functional measures, and histological assessments are not correlated (Munro, Szalai et al. 1998). These methods evaluate separate aspects of the changes that take pl ace after nerve injury and repair. For this study (phase II), the sci atic nerve distal to the repair was not routinely harvested for evaluation of myelinated axon number. For phase III we have incorporated harvesting sciatic nerve se ctions from the region distal to the nerve repair to ensure that number of myelinated axons that have entered the distal nerve stump can be evaluated morphologically, and not solely by extrapolation from electr ophysiological data. Axon Diameter Although the autograft repairs had mo re myelinated axons than the other repair groups, the percentage of myeli nated axons with diam eters between 2.05.0 m is almost identical for the AG, CT, and MT groups at the 12-week time point. Chamberlain et al. found axon di ameters centered ar ound a mode of 2.5 m in collagen tubes with a collagen-GAG co mplex used for repair of 10 mm rat sciatic nerve gap, at both 30 and 60 wee ks post repair (Chamberlain, Yannas et al. 1998). These data are similar to our results and the diamet ers of regenerated
142 myelinated axons at these time points are significantly less than the normal control (Keeley, Atagi et al. 1993). Mackinno n et al. found that for direct suture repair of transected rat sciatic nerve, ax on diameters were 3. 1 m at 1 month and 3.4 m at 3 months pos t up with the mean diameter only increasing to 4.7 m by 24 months (Mackinnon, Dellon et al. 1991). Norma l axon diameter averaged 6.5 m. Axon diamet er is one morphological par ameter that correlates with an aspect of functional recovery nerve conduction velocity, but not necessarily with functional gait par ameters (Dellon and Mackinnon 1989). The light microscopy methods employ ed in this study were designed to provide estimates of axon diameters that could discriminate differences between groups if they were present. It would r equire electron microscopy to resolve the axonal-myelin borders to a sufficient degree to provide conclusive data regarding the diameter of r egenerated or normal myelinated axons. The methods worked well for this part of the st udy (phase II) and may be a valuab le tool for phase III to demonstrate differences in myelinated axon diameter, which may occur when growth factors are included in the nerve guides. Ho et al. demonstrated increased axonal diam eter, neurofilament gene expression, and nerve conduction velocity following rat sciati c nerve repair with collagen nerve guides supplemented with the growth factors CTNF, BDNF, or both, as compared to tubes without these molecule s (Ho, Coan et al. 1998).
143 Functional Recovery After sciatic nerve injury and repair, rats walk by using their hip flexors and quadriceps to lift and advance the affected limb, and although muscles supplied by the sciatic nerve are denervated immedi ately after nerve transection the rats use the limb for support. In contrast to normal rats, which support their weight on their digits during walking, the bulk of the rats weight supported by the experimental hindlimb is focused directly on the heel. We believe that when the rat walks with the insensate hindlimb, t he disproportionate am ount of weight on the heel leads to ulceration, periarticular inflammation, and eventually fibrosis of the ankle joint. It is the probable cause of the limited ankle movement in this study, and is in agreement with Sunderl ands work (Sunderland 1991). During long periods of denervation, fibr osis of myocytes is a cause of limited joint motion that severely impacts functional recovery, however, this does not occur until 12-24 months post denerva tion and therefore is not a cause for the rats limited ankle motion in this study (Mackinnon and Dellon 1988; Sunderland 1991). This ankle joint contra cture profoundly limits the recovery of function, independent of muscle reinnervation Therefore, the relevance of the SFI calculated from walking tracks coul d be questioned because it incorporates a print length measurement, which is direct ly related to the animals ability to plantar flex the ankle (S henaq, Shenaq et al. 1989). The print length is small when the ra t supports weight on the digits with the ankle plantar flexed by the gastrocnem ius and soleus muscles. This elevates
144 the heel so no true heel print is left duri ng normal walking. It is only when the normal rat slows down to inspect its env ironment or rises up on the hindlimbs that the heel makes full contact wit h the supporting surface. When the gastrocnemius, soleus, and the rest of the muscles below the knee are denervated, as they are in this study, the rat cannot plantar flex its ankle against the load imposed by the rest of the body. The print length is therefore large in the denervated state and small in t he normally innerva ted state. When reinnervation takes place muscl e atrophy will start to reverse, however, these muscles will be unable to overcome the force of the rats body weight during stance phase of walking fo r quite some time. Even after early reinnervation, the degree of ankle joint contracture can continue to increase because the animal is unable to move the limb through normal physiological motion during functional weightbearing activi ties such as walking. Consequently, even if reinnervation takes place after a su ccessful nerve repair, the contracture that develops during the period of nerve re generation is irreversible; and this is what influences the SFI. The SFI may be useful for examining recovery of function following crush injury to t he sciatic nerve since the period of convalescence is markedly shorter and fewer complications arise (Bervar 2000), but for 10 mm gap repair it does not correla te with myelinated or unmyelinated axon number (Shenaq, Shenaq et al. 1989). Investigators have reported finding di gital contractures following sciatic nerve transection and repair. We noted t oe flexion abnormalities, however, they
145 were eliminated when the animal was put under anesthesia in preparation for electrophysiological assessment. This indicated to us that these flexion abnormalities were from an imbalance of muscle tone between the flexors and extensors and not joint flexion contractur es. This can occur due to an imbalance in the reinnervation of digital flexor and extensor muscles. Another possibility is that there is imbalance of muscle tone because of poor modulation by the CNS after axons have innervated muscles antagonist ic to their original role. In other words, some of the axons that previous ly innervated the extensor muscles have sent sprouts to the flexor compartm ent, and the higher motor centers cannot coordinate the contraction of opposing muscles. In fact the animals that most often were noted to have these toe flex ion abnormalities were the animals with the highest number of regenerat ed myelinated axons in the mid section of the repair site. They were the autografts. Shenaq et al. found that the clinical observations of improvement, decreased foot dragging and increased we ightbearing in some rats that underwent repair of 10 mm sciatic ner ve gaps, was not correlated with an improved SFI, and that these particular animals had a SFI that was no different than those with clinical observations of complete dysfunction (Shenaq, Shenaq et al. 1989). In addition, the SFI did not appear to correlate with axon counts. Unlike the toe flexion abnormalities in the present study, the limited ankle motion noted in all rats was because of tr ue joint contracture; it remained when
146 we attempted passive mobilization of the limb while the animal was under anesthesia (Dellon and Mackinnon 1989). The successful functional outcomes following nerve transection injury rely on two dependent, but distinct factors: surg ical nerve repair and rehabilitation. Optimal functional recovery will not o ccur without appropriate rehabilitation therapy, even when experienced surgeons use the most advanced nerve repair technologies. Optimal functional recovery will also not occur without a successful surgical nerve repair, no matter how much rehabilitation takes pl ace. Our current model is designed to best address surgic al repair strategies for nerve gap injuries. Although these functional m easures were chosen as one of the outcome assessments, perhaps they are better suited for studying rehabilitation strategies. It may be useful to draw distinctions, fo r research purposes, between successful surgical repair and successful functional recovery, following nerve transection injury. This is an academic distinction however, since patients who have suffered a nerve transection injury are only interested in functional improvement. Electrophysiology One of several factors t hat drove the evolution of the electrophysiological methods was the inconsistency of the resu lts during the earlier pa rt of the study. This inconsistency appeared to be due to some factor other than the expected variability of nerve regeneration. It was suspected, then proven through
147 experimentation, that our early results of recordings from the gastrocnemius muscle with needle electrodes were c onfounded by the inclusion of muscle action potentials from the normally innervat ed biceps femoris muscle. In the final protocol, when the nerve to the biceps femoris muscle was severed and removed, the electrophysiology results were more consistent and more easily interpreted. Below is a critical review of t he electrophysiology methods employed during this study. This discussion focuses not on the differences in the entubulation repair groups based on electr ophysiology results, but on the advantages and disadvantages of each method used for this model of studying peripheral nerve injury and repair. Nerve Action Potentials Initially evaluation of nerve acti on potentials was believed to be the simplest and most reliable method to determine the degree of functional nerve regeneration across the gap repai r. We believed we woul d obtain data that could be quantified and allow comparison between the entubulation repair groups as well as allow reference to the normal control and AG group. Additionally, we chose this method because the results were not influenced by secondary events, such as synapse formation at neuromuscu lar junctions, and axon misdirection to inappropriate targets. Both of these events potentially complicate interpretation of electrophysiological recordings from muscle.
148 One of the param eters recorded was the minimu m stimulus intensity at which a visible muscle twitch occurred. We initially thought this may be valuable information as better regeneration shoul d require less stimulus intensity to cause a visible muscle contraction. It was later realized that t he non-specificity of the muscle twitch was a major problem. When the sciatic nerve was stimulated in the gluteal region, the ent ire hind limb frequently moved viol ently in extension. The fact that the entire limb moved was not noteworthy since the muscles performing that function were likely not dener vated by sciatic nerve transection in the mid-thigh. Limb extension was pr esumably caused by contraction of the hamstrings and possibly the gluteal muscu lature. In addition, the entire limb movement made observation of the ankle plantar or dorsi flexion, which would be relevant movement and indicative of reco very of function, impossible to isolate visually. The gastrocnemius and tibialis anterior muscles, which cause plantar flexion and dorsi flexion respectively, are i nnervated distal to the repair site and movement here would be the indicative of regenerating axons reestablishing their peripheral synaptic contacts. Numerous fibrous adhesions and blood vessels made increasing the size of the surgical exposure challenging. On many occasions, our surgical exposure provided insufficient space to place either the recording or stimulating pair of electrode pairs on the actual sciatic nerve (proximal or distal) without placing one of the electrodes on the collagen nerve guide itself. We were not satisfied with this, but at that time we believed we were unable to enlarge the exposure without
149 either damaging the sciatic nerve or caus ing increased blood loss to the animal. The problem with contacting the collagen nerve guide instead of pure nerve is that although some resorption and remodeli ng of the collagen tube took place by 6 or 12 weeks, it is not normal nerve. The electrical impedance of the nerve guide, in all likelihood, is sufficiently different than nerve tissue and may affect the recordings. Another problem with this setup, was the recording electrodes placed immediately distal to the nerve gui de were approximatel y 14 mm from the stimulation electrode. Over th is short distance there is li ttle drop in the size of the stimulus artifact that is recorded simu ltaneously with the nerve action potentials. Thus, the recorded stimulus artifact was very large relative to the nerve action potential and in most case s obscured a large part of the recorded nerve action potential. This made determinations of latency or amplitude extremely difficult and in most cases impossible. The resu lts derived from this electrophysiology setup could only indicate whether there was nerve conduction across the gap. Lastly, prolonged elevation of the sci atic nerve on the hook electrodes may have caused a temporary ischemia and dehydration that accounted for changes that we observed during long ex periments. Often the action potential would diminish over time during an experi ment. On occasion, simply rehydrating the nerve with sterile salin e would restore the amplit ude. When the nerve was temporarily taken off the hook electrodes and then replaced a moment later, the evoked action potential would return. Kno wing this, we could not use amplitude
150 measurements to discriminate quality of regeneration between groups, even if the CNAP was clearly visible outside t he shadow of the stimulus artifact. Recording of evoked action potential across a nerve lesion is appropriate for a quick screening of whether axons crossed the injury. In fact, this is often how surgeons evaluate whether to perform surgical nerve repair. If an injury leaves the outer connective tissues of t he nerve intact and the patients complaint is loss of function or pain, as can occu r for example with a neuroma incontinuity, and nerve action potentials can be recor ded across the lesion, the decision is typically made to perform no furt her surgical intervention. Compound Muscle Action Potentials When the electrophysiology protocol was changed from recording of sciatic nerve action potentials to re cording CMAPs from the gastrocnemius muscle we were interested in obtaining quant ifiable data. Data, such as latency, amplitude, and estimates of nerve conduction velocity, which may discriminate the quality of regeneration bet ween groups, were all of interest. Recording of CMAPs from muscles that are denervat ed by the sciatic nerve lesion offers information about whether regenerating axons negotiated through the nerve guides and into the distal nerve stump and reestablished synaptic contact with a muscular target.
151 Compound Muscle Action Potentials and Anatomy It was essential for the remainder of the project that the cathodestimulating electrode made direct contact with a region of normal nerve proximal to the repair site. This involved expa nding the surgical incision and dissection proximally. We were careful to prevent compression of the sciatic nerve and this procedure became routine. Because of this enlarged exposure we were able confirm that a nerve previously observed, was branching from the sciatic nerve. This nerve was incorrectly identified as the posterior femoral cutaneous nerve of the thigh, due to its similar relationship of this nerve to the sciatic nerve in the human. Later in the project, after a more thorough anatomical study was undertaken and the anatomy was correlated with electrophysiology findings, this nerve was determined to be a muscular br anch from the tibial division of the sciatic nerve, which branches proximal to t he sciatic nerve injury and repair site. This proximal nerve branch descends deepl y and travels distally to innervate the biceps femoris. The rat anatomic text we consulted did not clearly illustrate this anatomy, the Figures were sc hematic in nature (Greene 1968). The tibial nerve emerges distal to our repair site and any axons present in it following the nerve repair represent regenerating axons. This anatomical configuration is similar to the human. The tibial ner ve gives off branches that innervate the gastrocnemius muscle. A major anatomical difference between the rat hind limb and the human is that in the rat, the gastrocnemius muscle is completely covered on its lateral aspect by the biceps femoris muscles anterior
152 and posterior heads. Thus, when a needle electrode is inserted into the lateral gastrocnemius muscle it must pass thr ough the biceps femoris muscle. If recording electrodes are not insulated, such that the entire needle shaft is conductive, the electrodes will record from the biceps femoris muscle as well as the gastrocnemius muscle when the sciatic nerve is stimulated in the gluteal region. The electrodes in this study were solid stainless steel, not insulated, and therefore capable of record ing from the biceps femoris muscle as well as the gastrocnemius muscle. Unlike the gastr ocnemius, which is a muscle that had been denervated by the sciatic nerve lesion, and may or may not have been reinervated by regenerating axons, the bi ceps femoris was innervated by normal nerve that was undistur bed during the sciatic ner ve lesion repair. These findings confounded the analysis of the recordings from the gastrocnemius. One could mistakenly assume successful regeneration when there was none, or incorrectly use the latency/amplitude of the normal biceps femoris CMAP instead of a smaller delayed gastrocnemius CMAP. We demonstrated this during experiments w here CMAPs where recorded from the gastrocnemius before and after the nerve to the biceps femoris was severed. A thorough study of the rat anatom y is critical prior to in itiating a peripheral nerve regeneration study such as this, and that assumptions based upon knowledge of human anatomy are inadequate. One additional factor that may influence CMAP recordings from the gastrocnemius muscle is that since the bi ceps femoris muscle is incised during
153 the surgical nerve repair there exists a local region of denervated biceps femoris muscle. This region would be attractive to regenerating axons of motor neurons, and during several terminal exposures fo r electrophysiology, axons that may have escaped the repair were noted in the vicinity. These may have erroneously reinnervated the incised regi on of the biceps femoris, and once again contributed unusual CMAPs to the electrodes targeted for the gastrocnemius muscle. In light of the above discussion, the ti bialis anterior muscle is perhaps, a better choice for percutaneous needle el ectrode recording following a sciatic nerve lesion than the gastrocnemius muscle. After all, there are no muscles superficial to the tibialis anterior muscl e in the rat and it is easily palpated. Following a sciatic nerve lesion, CMAP recordi ngs from the tibialis anterior reflect axons regeneration through the common per oneal (fibular) nerve into the deep peroneal nerve, and reflect the reformation of synapses with the tibialis anterior muscle. Nerve Conduction Velocity Regarding methods of obtaining estima tes of nerve conduction velocity, there is error when attempting to meas ure the distance between the stimulating electrode and the recording electrode in t he gastrocnemius muscle. First, since the nerve crosses posterior to the knee jo int axis it is slack when the knee is flexed. Even when the rats knee is extended prior to the measurement, the nerve does not take a perfectly straight course toward its entry into the
154 gastrocnemius. Second, since the nerve glides distally under the stimulating electrode during passive limb extension, it can be difficult to identify the exact measurement location. A measurement error of a fe w millimeters could greatly affect the resultant calculations. The part of the recording needle that is most accessible for measurement is the par t of the needle shaft that remains superficial to the skin. The needle how ever, was often inserted obliquely on route to the gastrocnemius muscle so t hat the position of the needle tip may differ significantly from that of the needle shaft at the skin interface. Palpating the needle tip through the tissues, where it rested in the gastrocnemius muscle, was performed in this study to facilitat e the more accurate measurements. On the other hand, if one does not measure the distance and only uses latency as the comparison between groups, one must be perfectly consistent with anatomical land marks used for placement of the electrodes to ensure the same inter-electrode distance between animals. This is not easily accomplished with the percutaneous needle-recording electrod e, and could have been facilitated by a surgical exposure of the muscle in question. A problem with using CMAP amplit ude measurements to test for differences between groups is that they can be greatly affect ed by the placement of the G2 (reference) electrode. In ear ly experiments, the G2 electrode was place subcutaneously, but it may have been too close to the muscle of interest. If the G2 recording elec trode picked up potentials fr om the biceps femoris
155 muscle, it may have increased CMAP amp litudes in the normal controls, or created a second peak in the recordings of the experimental limb. Although conceptually this principle was understood, we had difficulty in finding an appropriate location for the G2 electrode. One configuration often sited in texts of electrophysiology is t hat of placing the G2 electrode over the tendon of the tested muscle (Oh 1993). It was diffi cult to insert the needle in the region of the Achilles tendon, and limb movement upon stimulation often dislodged the needle from this location. T he final location of the G2 electrode was subcutaneous over the lateral malleolus. This gave us results that were more consistent. MUNE Although the methods employed fo r MUNE did not preclude more traditional assessments of latency and ampl itude, they took longer than previous experiments. This ultimately led to the decision to collect normal control data from the contralateral limb on only some animals in order to keep up with the schedule that had already been set by the nerve repair dates. There is little report of using MUNE for the study ner ve regeneration following transection of peripheral nerve. It is typically used to assess motor unit dropout with progressive degenerative diseases such as ALS. Although one of the research collaborators, R. Madison, has used M UNE for the study of peripheral nerve regeneration in the monkey the method was validated by using histological assessment of neuromuscular junctions. There was no such plan for validating
156 this method in the rat sciatic nerve le sion model using the gastrocnemius muscle for MUNE. It did not seem appropriate to assume that si nce MUNE is valid in the monkey thenar group following a median nerve lesion and repair that MUNE would be valid in the rat gastrocnemius fo llowing sciatic nerve lesion and repair. There are great differences in t he median nerve lesion model in the monkey and the sciatic nerve lesion model in the rat. Alt hough both peripheral nerves, the median nerve at the wrist innervates muscles a relatively short distance from the lesion site, with no mo tor branches to muscle groups that antagonize the thenar group. A mid-thigh sciatic transec tion is a more proximal nerve lesion in the hindlimb relative to the median nerve transection in the wrist. Therefore, regenerat ing axons that cross the r epaired sciatic nerve gap have many more branch points to navigate befor e reaching target tissues than axons crossing the repaired median nerve gap. Th is greatly increases the likelihood of axonal misdirection, and can compromise function recovery even when axons successfully regenerate across the repaired site. In addition, since all of the CMAPs recorded during this time period of the study contained the interference pattern of the biceps femoris, CMAPs analysis was extremely difficult if not impossible. The research group, collect ively, decided to discontinue MUNE and concentrate on simple recordings of CMAPs from relevant muscle groups. Final Electrophysiological Protoc ol for the phase III rat study The electrophysiological protocol evolv ed during this study (phase II). In order to ensure that CMAPs were recorded from the gastrocnemius muscle
157 several final modifications were te sted on a normal animal, and were then implemented for the phase III rat study. To eliminate doubt about the location of the recording needle electrode insert ed percutaneously and targeted for the gastrocnemius muscle, an open procedure was implemented. We now surgically expose the gastrocnemius during the termi nal electrophysiology assessment. An insulated/coated needle elec trode with only a small conduc tive area at the tip is used. Although these additional procedures theoretically eliminate the possibility of errant recordings from the biceps femoris muscle, we c ontinue to sever and dissect free the nerve to biceps femoris muscle to prevent any biceps femoris potentials from bleeding over to the needles inserted into the gastrocnemius muscle. This could theoretically occur if a high intensity stimulus is given and the recording amplifier gain is turned up to pick up small amplitude potentials. We stimulate both proximal and distal to the repa ir site to obtain latency values that can be used to calculate NCV across the repair site. To improve accuracy of the distance measurement betw een the stimulating and recording electrodes, a small suture is tied at both the proximal and distal locations of the stimulus electrodes. This also facilitates the electrodes being placed in the same location if the procedure must be repeated during the experi ment. Vernier calipers are used to measure from the sutures to the shaft of the needle recording electrode, which remains in the gastrocnemius muscle until measurements are completed. The knee joint is passively extended to end range. This pulls the sciatic nerve taught
158 and facilitates a measurement that is as close to a linear as is possible in vivo Following the gastrocnemius muscle record ings and measurements, we record CMAPs from the tibialis anterior muscle as well as the intrinsic muscles of the foot. We sample several metatarsal spaces. We are satisfied with this final protocol in moving into the final phase of the rat study.
159 Chapter 6: Conclusion When comparing the two designs of ner ve guides with guidance channels, the results from the 12-week myelinated axon counts clearly show the superiority of the MT design over the MF design. However, when microtube guidance channels are constructed with type I co llagen without the addition of bioactive molecules (as they were in this phase of the study) the addi tional surface area provided by the microtubes confers no cl ear benefit to nerve regeneration over that of the empty type I collagen t ube for the entubulatio n repair of a 10 mm peripheral nerve gap in the rat sciatic nerve. As previously stated, the presence of channels (microtubes or filaments) in the nerve guides increased the variabili ty of the number of myelinated axons growing through the nerve guide. In a peripheral nerve repair model where compression becomes a pivotal factor in fluencing the success of a nerve graft, microtubes may provide the essential stru ctural support to maintain the patency of the nerve guide lumen and thus prove to be an essential component. The most consistent results from ent ubulation repair occurred in the group without longitudinal channels. Although some of the specimens in the MT group exceeded the highest axon numbers of the CT group, there were considerably more poor results. Neither of these tw o nerve guides could match the success of
160 the autograft when compared at the same time point. The autograft appears to accelerate regeneration across the nerve gap relative to the nerve guides. Cross-linked type I collagen may be the optimal material for the outer nerve guide due to its high tensile strength (Berglund, Mohseni et al. 2003). It allows the nerve guide to accept a suture without tearing, and has the benefit that it is resorbable over time. It is no t, however, the optimal substrate for cell migration (Hashimoto, Suzuki et al. 2002; Vleggeert-Lankamp, Pego et al. 2004). In contrast to type I collagen, type IV co llagen is a major component of the basal lamina that axons grow upon. It is lami nin that has binding sites both for collagen and cellular adhesion molecules in the axonal growth cone (Rogers, Letourneau et al. 1983). Incorporating adhesive molecule s such as laminin to the surfaces of the collagen guidance channels may dramatic ally improve their effectiveness in promoting nerve regenerati on. (Rogers, Letourneau et al 1983; Verdu, Labrador et al. 2002) An improved study design aimed to st rictly test the hypothesis that increased surface area impr oves nerve regeneration ac ross a biomaterial nerve guide, might include nerve guides with lami nin on their interior but no guidance channels, verses nerve guides that contain laminin both in their interior but also contain laminin coated mi crotube guidance channels. This would provide a comparison of two nerve guides, each with a different architecture and surface area, but both coated with a molecu le known to promote cell adhesion and migration. Studies underwa y at this time in our laboratory are designed to
161 assess the effect of the inclusion of laminin in microtube guidance channels on nerve regeneration. The cumulative effect of the longi tudinal channels, cellular adhesive molecules, and growth fact ors, may be greater than si mply the increase in surface area. It may provide a synergi stic benefit that allows greater regeneration than just grow th factors and adhesive molecules in combination, since it provides increased surface ar ea for seeding these molecules. These questions remain to be answered by future studies. Further study into the complete in vivo resorption of the collagen tubes with and without guidance channels is necessary as well. A final note is that despite the va riability in the regenerated myelinated axon numbers, based on histological result s, CMAPs were recorded at 12 weeks post nerve repair from t he feet of rats in each group, demonstrating the extraordinary ability of the PNS to regenerate. In the case when an autograft repair is not possible, biomaterial ner ve guides may be used to successfully bridge a gap in a severed peripheral nerve.
162 References Ahmed, Z. and R. A. Br own (1999). "Adhesion, a lignment, and migration of cultured Schwann cells on ultrathin fibronectin fibres." Cell Motil Cytoskeleton 42 (4): 331-43. Ahmed, Z., S. Underwood, et al. ( 2003). "Nerve guide material made from fibronectin: assessment of in vitro properties." Tissue Eng 9 (2): 219-31. Anselme, K., C. Bacques, et al. (1990) "Tissue reaction to subcutaneous implantation of a collagen sponge. A histological, ultrastructural, and immunological study." J Biomed Mater Res 24 (6): 689-703. Arai, T., G. Lundborg, et al. (2000). "Bioartificial nerve graft for bridging extended nerve defects in rat sciatic nerve based on resorbable gui ding filaments." Scand J Plast Reconstr Surg Hand Surg 34 (2): 101-8. Archibald, S. J., C. Krarup, et al. (1991). "A collagen-bas ed nerve guide conduit for peripheral nerve repair: an el ectrophysiological study of nerve regeneration in rodent s and nonhuman primates." J Comp Neurol 306 (4): 685-96. Asato, F., M. Butler, et al. (2000). "Variation in rat sciatic nerve anatomy: implications for a rat model of neur opathic pain." J Peripher Nerv Syst 5 (1): 19-21. Bain, J. R., S. E. Mackinnon, et al. (1988). "The peripheral nerve allograft: an assessment of regeneration across nerve allografts in rats immunosuppressed with cyclosporin A." Plast Reconstr Surg 82 (6): 105266. Bain, J. R., S. E. Mackinnon, et al. (1989). "Functional evaluation of complete sciatic, peroneal, and posterior tibial nerve lesions in the rat." Plast Reconstr Surg 83 (1): 129-38. Battiston, B., P. Tos, et al. (2000). "Nerve repair by means of vein filled with muscle grafts I. Clinical results." Microsurgery 20 (1): 32-6.
163 Beghe, F., C. Menicagli, et al. (1992) "Lyophilized non-denatured type-I collagen (Condress) extracted from bovine Ach illes' tendon and suitable for clinical use." Int J Tissue React 14 Suppl : 11-9. Berglund, J. D., M. M. Mohseni, et al. (2003). "A biological hybrid model for collagen-based tissue engineered vascular constructs." Biomaterials 24 (7): 1241-54. Bervar, M. (2000). "Video analysis of standi ng--an alternative footprint analysis to assess functional loss following injury to the rat sciatic nerve." J Neurosci Methods 102 (2): 109-16. Boot-Handford, R. P. and D. S. Tuckwel l (2003). "Fibrillar collagen: the key to vertebrate evolution? A tale of molecular incest." Bioessays 25 (2): 142-51. Borkenhagen, M., R. C. Sto ll, et al. (1998). "In vi vo performance of a new biodegradable polyester urethane system used as a nerve guidance channel." Biomaterials 19 (23): 2155-65. Brandt, J., L. B. Dahlin, et al. (1999). "Autologous tendons used as grafts for bridging peripheral nerve defects." J Hand Surg [Br] 24 (3): 284-90. Ceballos, D., X. Navarro, et al. (1999). "M agnetically aligned coll agen gel filling a collagen nerve guide improves peripher al nerve regeneration." Exp Neurol 158 (2): 290-300. Chamberlain, L. J., I. V. Yannas, et al. (1998). "Early peripher al nerve healing in collagen and silicone tube implants: myofibroblasts and the cellular response." Biomaterials 19 (15): 1393-403. Chamberlain, L. J., I. V. Yannas, et al. (1998). "Collagen-GAG substrate enhances the quality of nerve regener ation through collagen tubes up to level of autograft." Exp Neurol 154 (2): 315-29. Charulatha, V. and A. Rajaram (2003). "Influence of different crosslinking treatments on the physi cal properties of collagen membranes." Biomaterials 24 (5): 759-67. Chernousov, M. A., R. C. Stahl, et al. (2001). "Schwann cell type V collagen inhibits axonal outgrowth and promotes Schwann cell migrat ion via distinct adhesive activities of the collagen and noncollagen domains." J Neurosci 21 (16): 6125-35. Cloft, H. J., D. F. Kallmes, et al (2000). "Bovine ty pe I collagen as an endovascular stent-graft mate rial: biocompatibility study in rabbits." Radiology 214 (2): 557-62.
164 Daamen, W. F., H. T. van Moerkerk, et al. (2003). "Preparatio n and evaluation of molecularly-defined collagen-elas tin-glycosaminoglycan scaffolds for tissue engineering." Biomaterials 24 (22): 4001-9. Daube, J. R. (1995). "Estim ating the number of motor units in a muscle." J Clin Neurophysiol 12 (6): 585-94. Daube, J. R. (1996). Compound Mu scle Action Potentials. Clinical Neurophysiology J. R. Daube. Philadelphia, F.A. Davis Company. 46: 199-234. Daube, J. R. (1996). Estimating The Number of Motor Units In a Muscle. Clinical Neurophysiology J. R. Daube. Philadelphia, F.A. Davis Company. 46: 301-311. Daube, J. R., C. Gooch, et al. (2000) "Motor unit number estimation (MUNE) with nerve conduction studies." Suppl Clin Neurophysiol 53 : 112-5. de Medinaceli, L., W. J. Fr eed, et al. (1982). "An index of the functional condition of rat sciatic nerve based on measur ements made from walking tracks." Exp Neurol 77 (3): 634-43. Dellon, A. L. and S. E. Mackinnon (1989). "S ciatic nerve regeneration in the rat. Validity of walking track assessment in the presence of chronic contractures." Microsurgery 10 (3): 220-5. Dellon, A. L. and S. E. Mackinnon (1989) "Selection of the appropriate parameter to measure neural regeneration." Ann Plast Surg 23 (3): 197202. Dohm, S., M. Streppel, et al. (2000). "Loc al application of extracellular matrix proteins fails to reduce the number of axonal branches after varying reconstructive surgery on rat facial nerve." Restor Neurol Neurosci 16 (2): 117-126. Esposito, B., A. De Santis, et al. (2002) "Mast cells in Wallerian degeneration: morphologic and ultrastructural changes." J Comp Neurol 445 (3): 199210. Evans, G. R. (2000). "Challenges to ner ve regeneration." Semin Surg Oncol 19 (3): 312-8. Evans, G. R. (2001). "Peripheral nerve in jury: a review and approach to tissue engineered constructs." Anat Rec 263 (4): 396-404.
165 Evans, G. R., K. Brandt, et al. (2000). "Clinical long-term in vivo evaluation of poly(L-lactic acid) porous conduits for peripheral nerve regeneration." J Biomater Sci Polym Ed 11 (8): 869-78. Evans, G. R., K. Brandt, et al. (1999). "In vivo evaluation of poly(L-lactic acid) porous conduits for peripheral nerve regeneration." Biomaterials 20 (12): 1109-15. Fansa, H., W. Schneider, et al. (2001). "Revasculari zation of tissue-engineered nerve grafts and invasion of macrophages." Tissue Eng 7 (5): 519-24. Fansa, H., W. Schneider, et al. (2002). "Influence of in sulin-like growth factor-I (IGF-I) on nerve autografts and tiss ue-engineered nerve grafts." Muscle Nerve 26 (1): 87-93. Fine, E. G., I. Decosterd, et al. ( 2002). "GDNF and NGF released by synthetic guidance channels support sciatic nerve regeneration across a long gap." Eur J Neurosci 15 (4): 589-601. Fornaro, M., P. Tos, et al. (2001). "Confocal imaging of Schwann-cell migration along muscle-vein combined grafts used to bridge nerve defects in the rat." Microsurgery 21 (4): 153-5. George, E. B., J. D. Gla ss, et al. (1995). "Axotomy -induced axonal degeneration is mediated by calcium influx thro ugh ion-specific channels." J Neurosci 15 (10): 6445-52. Geuna, S., P. Tos, et al. (2000). "Mor phological analysis of peripheral nerve regenerated by means of vein grafts fill ed with fresh skeletal muscle." Anat Embryol (Berl) 201 (6): 475-82. Glasby, M. A., S. Gschmei ssner, et al. (1986). "Regenerat ion of the sciatic nerve in rats. The effect of muscle basem ent membrane." J Bone Joint Surg Br 68 (5): 829-33. Godard, J., G. Coulon, et al. (1984). "[Nerve regeneration th rough a venous graft in the rat. Preliminary results]." Neurochirurgie 30 (6): 407-16. Goldberg, J. L. (2003). "How does an axon grow?" Genes Dev 17 (8): 941-58. Grados-Munro, E. M. and A. E. Fournier (2003). "Myelin-a ssociated inhibitors of axon regeneration." J Neurosci Res 74 (4): 479-85. Greene, E. C. (1968). Anatomy of the Rat New York, Hafner Publishing Company, Inc.
166 Hadlock, T. A., C. A. Sundback, et al (2001). "A new artificial nerve graft containing rolled Schwann cell monolayers." Microsurgery 21 (3): 96-101. Hare, G. M., P. J. Evans, et al. (1 992). "Walking track analysis: a long-term assessment of peripheral nerve re covery." Plast Reconstr Surg 89 (2): 251-8. Hashimoto, T., Y. Suzuki, et al. ( 2002). "Peripheral nerve regeneration through alginate gel: analysis of early outgrowth and late increase in diameter of regenerating axons." Exp Brain Res 146 (3): 356-68. Ho, P. R., G. M. Coan, et al. (1998). "Repair with collagen tubules linked with brain-derived neurotrophic fa ctor and ciliary neurotr ophic factor in a rat sciatic nerve injury model." Arch Otolaryngol Head Neck Surg 124 (7): 7616. Hunter, J. M., L. H. Schneider, et al (1997). Tendon And Nerve Surgery In The Hand: A Third Decade St. Louis, Mosby-Year Book, Inc. Ide, C. (1983). "Nerve r egeneration and Schwann cell basal lamina: observations of the long-term regenerat ion." Arch Histol Jpn 46 (2): 243-57. Ide, C., K. Tohyama, et al. (1983). "Schwann cell basal lamina and nerve regeneration." Brain Res 288 (1-2): 61-75. Ishihara, M., P. N. Shaklee, et al. (1994) "Structural features in heparin which modulate specific biological activities mediated by basic fibroblast growth factor." Glycobiology 4 (4): 451-8. Itoh, S., K. Takakuda, et al. (2002). "Eva luation of cross-linking procedures of collagen tubes used in peripheral nerve repair. Biomaterials 23 (23): 447581. Keeley, R., T. Atagi, et al. (1993). "P eripheral nerve regeneration across 14-mm gaps: a comparison of aut ograft and entubulation r epair methods in the rat." J Reconstr Microsurg 9 (5): 349-58; discussion 359-60. Keeley, R. D., K. D. Nguyen, et al. (1991). "The artificial nerve graft: a comparison of blended elastomer-h ydrogel with polyglycolic acid conduits." J Reconstr Microsurg 7 (2): 93-100. Krarup, C., S. J. Archibald, et al. (2002). "Factors t hat influence peripheral nerve regeneration: an electrophysiological st udy of the monkey median nerve." Ann Neurol 51 (1): 69-81.
167 Laurie, G. W., C. P. Leblond, et al. (1982). "Localizatio n of type IV collagen, laminin, heparan sulfate proteoglycan, and fibronectin to the basal lamina of basement membranes." J Cell Biol 95 (1): 340-4. Lee, A. C., V. M. Yu, et al. (2003). "Controlled release of nerve growth factor enhances sciatic nerve regeneration." Exp Neurol 184 (1): 295-303. Lubinska, L. (1977). "Early course of Wallerian degeneration in myelinated fibres of the rat phrenic nerve." Brain Res 130 (1): 47-63. Lundborg, G. (2000). "A 25-year perspecti ve of peripheral nerve surgery: evolving neuroscientific concepts and c linical significance." Journal of Hand Surgery 25 (3): 391-414. Lundborg, G. (2002). "Editorial: Enhancing posttraumatic nerve regeneration." Journal of the Peripheral Nervous System 7 : 139-140. Mackinnon, S. E. and A. L. Dellon (1988). Surgery Of T he Peripheral Nerve. New York, Thieme Medical Publishers, Inc. : 3-4. Mackinnon, S. E., A. L. De llon, et al. (1991) "Changes in nerve fiber numbers distal to a nerve repair in the ra t sciatic nerve model." Muscle Nerve 14 (11): 1116-22. Madison, R. D. and S. J. Archibald (1994). "Point sour ces of Schwann cells result in growth into a nerve entubulation repair site in the absence of axons: effects of freeze-t hawing." Exp Neurol 128 (2): 266-75. Madison, R. D., C. da S ilva, et al. (1987). "Peri pheral nerve regeneration with entubulation repair: com parison of biodegradeable nerve guides versus polyethylene tubes and the effects of a laminin-containing gel." Exp Neurol 95 (2): 378-90. Madison, R. D., C. F. Da Silva, et al. (1988). "Ent ubulation repair with protein additives increases the maximum nerve gap distance successfully bridged with tubular prostheses." Brain Res 447 (2): 325-34. McComas, A. J., P. R. Fawcett, et al. ( 1971). "Electrophysiologic al estimation of the number of motor units within a human muscle." J Neurol Neurosurg Psychiatry 34 (2): 121-31. Misulis, K. M. (1997). Essentia ls of Clinical Neurophysiology Boston, Butterworth-Heinemann.
168 Mosahebi, A., P. Fuller, et al. (2002) "Effect of allogeneic Schwann cell transplantation on peripheral ner ve regeneration." Exp Neurol 173 (2): 21323. Munro, C. A., J. P. Szal ai, et al. (1998). "Lack of association between outcome measures of nerve regeneration." Muscle Nerve 21 (8): 1095-7. Ngo, T. T., P. J. W aggoner, et al. (2003). "Poly( L-Lactide) microfilaments enhance peripheral nerve regeneration ac ross extended nerve lesions." J Neurosci Res 72 (2): 227-38. Norrby, K. (2002). "Mast ce lls and angiogenesis." Apmis 110 (5): 355-71. Oh, S. J. (1993). Clinical Electrom yography: Nerve Conduction Studies Baltimore, Willi ams & Wilkins. Ramon Y Cajal, S. (1928). Cajal's degener ation and regeneration of the nervous system New York, Oxford University Press. Rodrigues Ade, C. and M. D. Silva (2001) "Inside-out versus standard artery graft to repair a sensory ner ve in rats." Microsurgery 21 (3): 102-7. Rogers, S. L., P. C. Let ourneau, et al. (1983). "Neurit e extension by peripheral and central nervous system neurons in response to substratum-bound fibronectin and laminin." Dev Biol 98 (1): 212-20. Rosenstein, J. M. and J. M. Krum (2004). "New roles for VEGF in nervous tissue-beyond blood vessels." Exp Neurol 187 (2): 246-53. Ross, M. H., L. J. Romrell, et al. (1995) Histology A Text A nd Atlas. Baltimore, Williams & Wilkins : 256-300. Sakiyama, S. E., J. C. Schense, et al (1999). "Incorporation of heparin-binding peptides into fibrin gels enhances neurite extension: an example of designer matrices in tissue engineering." Faseb J 13 (15): 2214-24. Santos, X., J. Rodrigo, et al. (1998). "Evaluation of peripheral nerve regeneration by nerve growth factor locally administered with a novel system." J Neurosci Methods 85 (1): 119-27. Schwab, M. E. (2004). "Nogo and axon regenerati on." Curr Opin Neurobiol 14 (1): 118-24. Shefner, J. M. (2001). "M otor unit number estimati on in human neurological diseases and animal m odels." Clin Neurophysiol 112 (6): 955-64.
169 Shefner, J. M., M. E. Cudk owicz, et al. (2002). "Com parison of incremental with multipoint MUNE methods in transgenic ALS mice." Muscle Nerve 25 (1): 39-42. Shenaq, J. M., S. M. Shenaq, et al. (1989). "Reliability of sciatic function index in assessing nerve regeneration acro ss a 1 cm gap." Microsurgery 10 (3): 214-9. Stoll, G., J. W. Griffin, et al. (1989). "Wallerian degeneration in the peripheral nervous system: participation of both Schwann cells and macrophages in myelin degradation." J Neurocytol 18 (5): 671-83. Stoll, G., B. D. Trapp, et al. (1989) "Macrophage function during Wallerian degeneration of rat optic nerve: clear ance of degenerating myelin and Ia expression." J Neurosci 9 (7): 2327-35. Sunderland, S. S. (1991). Nerve Injuries and their Repair: A Critical Appraisal New York, Churchill Livingstone. Terzis, J., B. Faibisoff, et al. (1975). "The nerve gap: suture under tension vs. graft." Plast Reconstr Surg 56 (2): 166-70. Trumble, T. E. and F. G. Shon (2000). "T he physiology of nerve transplantation." Hand Clin 16 (1): 105-22. Tyrrell, D. J., M. Ishihara, et al. (1993). "Structure and biological activities of a heparin-derived hexasacchar ide with high affinity for basic fibroblast growth factor." J Biol Chem 268 (7): 4684-9. Utley, D. S., S. L. Lewin, et al. (1996) "Brain-derived neurotrophic factor and collagen tubulization enhance functional recovery after peripheral nerve transection and repair." Arch Otolaryngol Head Neck Surg 122 (4): 407-13. Valero-Cabre, A., K. Tsironis, et al. (2001). "Superior muscle reinnervation after autologous nerve graft or poly-L-lactide-epsiloncaprolactone (PLC) tube implantation in comparison to sili cone tube repair." J Neurosci Res 63 (2): 214-23. Verdu, E., R. O. Labrador, et al. ( 2002). "Alignment of collagen and laminincontaining gels improve nerve regenerat ion within silicone tubes." Restor Neurol Neurosci 20 (5): 169-79. Vleggeert-Lankamp, C. L., A. P. Pego, et al. (2004). "Adhesion and proliferation of human Schwann cells on adhesiv e coatings." Biomaterials 25 (14): 2741-51.
170 Waller, A. (1850). "Experiments on the Section of the Glossopharyngeal and Hypoglossal Nerves of the Frog, and Observations of the Alterations Produced Thereby in the Stru cture of Their Primitive Fibres." Philosophical Transactions of the Royal Society of London 140 : 423-429. Woodburne, R. T. and W. E. Burkel ( 1988). Essentials of Human Anatomy New York, Oxford University Press. Woolley, A. L., J. P. Hollo well, et al. (1990). "First pl ace--Resident Basic Science Award 1990. Fibronectin-laminin comb ination enhances peripheral nerve regeneration across long gaps." Otolaryngol Head Neck Surg 103 (4): 50918. Xu, X., W. C. Yee, et al. (2003). "Peripheral nerve regeneration with sustained release of poly(phosphoester) microenc apsulated nerve growth factor within nerve guide conduits." Biomaterials 24 (13): 2405-12. Yin, Q., G. J. Kemp, et al. (2001). "Neurotrophin-4 delivered by fibrin glue promotes peripheral nerve regeneration." Muscle Nerve 24 (3): 345-51. Young, R. C., M. Wiberg, et al. (2002) "Poly-3-hydroxybutyrate (PHB): a resorbable conduit for long-gap repair in peripheral nerves." Br J Plast Surg 55 (3): 235-40. Zhang, J., T. M. Oswald, et al. (2 003). "Enhancement of rat sciatic nerve regeneration by fibronectin and lamini n through a silicone chamber." J Reconstr Microsurg 19 (7): 467-72. Zochodne, D. W. and C. Cheng (2000). "N eurotrophins and other growth factors in the regenerative milieu of prox imal nerve stump tips." J Anat 196 ( Pt 2) : 279-83.
About the Author Oren D. Rosenthal received a Bachelor s Degree in Psychology from the University of Buffalo in 1989 and a Mast ers Degree in Physical Therapy from Rutgers University/Universit y of Medicine and Dentistry of New Jersey in 1995. After physical therapy school, Orens pos tgraduate continuing education focused on orthopedics and rehabilitat ion of the spine. He prac ticed physical therapy in a variety of clinical settings prior to r eentering graduate school in the fall of 2000. He enrolled in the Ph.D. program in M edical Sciences with a concentration in anatomy, at the University of South Flori da, College of Medicine. While in the Ph.D. program, Oren taught human anatomy to first year medical students and presented research at a nati onal Neuroscience meeting.