UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The effects of impact depth and velocity on spinal cord injury severity Lam, Cameron Jonathan 2013

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2013_fall_lam_cameron.pdf [ 4.16MB ]
Metadata
JSON: 24-1.0071997.json
JSON-LD: 24-1.0071997-ld.json
RDF/XML (Pretty): 24-1.0071997-rdf.xml
RDF/JSON: 24-1.0071997-rdf.json
Turtle: 24-1.0071997-turtle.txt
N-Triples: 24-1.0071997-rdf-ntriples.txt
Original Record: 24-1.0071997-source.json
Full Text
24-1.0071997-fulltext.txt
Citation
24-1.0071997.ris

Full Text

THE EFFECTS OF IMPACT DEPTH AND VELOCITY ON SPINAL CORD INJURY SEVERITY  by CAMERON JONATHAN LAM B.A.Sc, The University of Toronto, 2010  A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF  MASTER OF APPLIED SCIENCE  in  THE FACULTY OF GRADUATE STUDIES (Biomedical Engineering)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  June 2013  © Cameron Jonathan Lam, 2013  ABSTRACT Currently a treatment for spinal cord injury (SCI) remains elusive to clinicians and researchers. This is, in part, due to variation between primary injury mechanisms and diversity of mechanical impact factors such as impact velocity, depth, force, and acceleration. This research examines both the individual and combined effects of impact velocity and depth on the cervical spinal cord and also aims to understand the contribution of the energy applied, not only the impact factors. In this study, contusion spinal cord injuries were induced in 54 male, Sprague-Dawley rats at impact speeds of 8 mm/s, 80 mm/s, or 800 mm/s with displacements of 0.9 mm or 1.5 mm. Animals recovered for seven days followed by behavioural assessment and examination of the spinal cord tissue for demyelination and tissue sparing at 1 mm intervals ±3 mm rostrocaudally to the epicentre. In parallel, a finite element model of the rat spinal cord was used to examine the resulting maximum principal strains in the spinal cord during impact. Impact depth was a consistent factor in qualifying axonal damage in the spinal cord, tissue sparing, and resulting behavioural deficit.  Increased impact velocity resulted in  significantly different impact energies and measureable outcomes at the 1.5 mm impact depth, but not the 0.9 mm impact depth, identifying threshold interactions between the two factors. The difference of injury severity to velocity at different impact depths identifies the existence of threshold interactions between the two impact factors. Linear correlation analysis with finite element analysis (FEA) strain showed significant (  ) correlations with axonal damage in the ventral (  regions of the spinal cord and with white matter (  ) and lateral (  )  ) and grey matter (  )  sparing. Non-parametric correlation analysis identified strong correlations between grey and white matter strain with open field behavioural scores (  ).  The results shown by this work extend the research identifying significant correlation between maximum principal strain and neurologic tissue damage. Furthermore, a relationship between the impact depth and velocity of injury demonstrated a more rate sensitive response of the spinal cord at the 1.5 mm impact depth than at the 0.9 mm impact depth.  ii  PREFACE Under the guidance of my supervisor, Dr. Thomas Oxland, I have identified and developed the experimental design for the research presented in this thesis. The finite element model used in this thesis was originally developed by Colin Russell which I modified, in part, to match the experimental injuries analyzed.  A trained microsurgeon, at the International  Collaboration on Repair Discoveries (ICORD) research centre, Dr. Jie Liu, performed all surgically related procedures described in the this thesis.  I conducted the majority of all  experimental procedures including characterizing the initial injury, monitoring animals, behavioural analysis, tissue preparation, staining and imaging, and computational contusion simulations. I also completed final analysis and statistical tests of the spinal cord tissue and behavioural results. Ethical approval for animal research was granted by the University of British Columbia Office of Research Services Animal Care Committee under certificate number A07-0379 (Biomech Behaviour of the SC) and A11-0379 (Effects of Impact Velocity and Impact Depth on Spinal Cord Injury).  iii  TABLE OF CONTENTS  Abstract .................................................................................................................... ii Preface ..................................................................................................................... iii Table of Contents ................................................................................................... iv List of Tables ......................................................................................................... vii List of Figures ....................................................................................................... viii Acknowledgements................................................................................................ xii Chapter 1 Introduction............................................................................................1 1.1  Overview and motivation .............................................................................................. 1  1.2  Anatomy of the spine and spinal cord .......................................................................... 2  1.2.1  Anatomical planes ....................................................................................................... 2  1.2.2  Anatomy of the spinal column .................................................................................... 3  1.2.3  Anatomy of the spinal cord ......................................................................................... 5  1.3  Overview of spinal cord injury mechanisms................................................................ 9  1.3.1  Contusion models........................................................................................................ 9  1.3.2  Dislocation and distraction models ........................................................................... 14  1.3.3  Transection and clip compression models ................................................................ 16  1.4  Spinal cord injury biomechanics ................................................................................ 16  1.4.1  Material properties of the spinal cord ....................................................................... 16  1.4.2  Impact mechanics and kinetics ................................................................................. 20  1.5  FE models ...................................................................................................................... 22  1.5.1 1.6  Strain theory .............................................................................................................. 24  Project objectives and scope ........................................................................................ 25  1.6.1  Objectives ................................................................................................................. 25  1.6.2  Scope ......................................................................................................................... 26  iv  Table of Contents  Chapter 2 Methods.................................................................................................27 2.1  Experimental overview ................................................................................................ 27  2.2  UBC impactor characterization .................................................................................. 29  2.2.1  Static loading ............................................................................................................ 29  2.2.2  Dynamic loading ....................................................................................................... 30  2.3  Survival study ............................................................................................................... 31  2.3.1  Kinetic analysis ......................................................................................................... 35  2.3.2  Behavioural analysis and correlations ...................................................................... 36  2.3.2.1  Rearing and grooming ..................................................................................... 36  2.3.2.2  Open field behaviour ....................................................................................... 38  2.3.3  2.4  Histological analysis ................................................................................................. 40  2.3.3.1  Luxol fast blue ................................................................................................. 40  2.3.3.2  Demyelination ................................................................................................. 42  FEA strain computation .............................................................................................. 45  2.4.1  Geometry and material properties ............................................................................. 45  2.4.2  Contusion injury simulation ...................................................................................... 47  2.4.3  FEA optimization ...................................................................................................... 49  2.5  Correlations .................................................................................................................. 49  Chapter 3 Results ...................................................................................................51 3.1  FEA strain computation .............................................................................................. 51  3.1.1  FEA optimization ...................................................................................................... 51  3.1.2  Contusion injury simulation ...................................................................................... 54  3.1.3  Principal strain distribution ....................................................................................... 56  3.2  Survival study ............................................................................................................... 57  3.2.1  Impact mechanics...................................................................................................... 57  3.2.2  Behavioural analysis ................................................................................................. 60  3.2.2.1  Rearing and grooming ..................................................................................... 60 v  Table of Contents 3.2.2.2  Open field behaviour ....................................................................................... 62  3.2.2.3  Correlations ..................................................................................................... 63  3.2.3  Histological analysis ................................................................................................. 65  3.2.3.1  Luxol fast blue ................................................................................................. 65  3.2.3.2  Demyelination ................................................................................................. 66  3.2.3.3  Correlations ..................................................................................................... 68  Chapter 4 Discussion .............................................................................................72 4.1  Effects of impact velocity and depth on secondary injury ....................................... 72  4.2  Contusion simulation ................................................................................................... 76  4.3  Limitations .................................................................................................................... 77  4.3.1  Experimental model .................................................................................................. 77  4.3.2  Computational model ................................................................................................ 79  Chapter 5 Conclusion ............................................................................................81 5.1  Conclusions ................................................................................................................... 81  5.2  Contributions ................................................................................................................ 82  5.3  Recommendations for future work ............................................................................. 83  5.4  Concluding statement .................................................................................................. 85  References ...............................................................................................................86 Appendix A : UBC Machine Accuracy Testing ..................................................91 Appendix B : Primary Injury Pilot Study ...........................................................95 B.1  Methods ......................................................................................................................... 95  B.1.1 B.2  Extravasation analysis ............................................................................................... 96  Results ........................................................................................................................... 97  Appendix C : Monitoring and Scoring Sheets ..................................................100 Appendix D : Staining Protocols ........................................................................104  vi  LIST OF TABLES  Table 2.1: Experimental injury group design ............................................................................... 32 Table 2.2: UBC Multi-mechanism injury system compensated load parameters ......................... 34 Table 2.3: Articular movement scoring ........................................................................................ 38 Table 2.4: Paw placement/orientation scoring .............................................................................. 39 Table 2.5: Material properties of the spinal cord and dura ........................................................... 46 Table 3.1: FEA simulation and experimental injury force and displacement comparison ........... 54 Table 3.2: Experimental contusion injury parameters .................................................................. 58 Table 3.3: Correlation coefficients for mechanical impact factors with behavioural assessment scores .......................................................................................................................... 63 Table 3.4: Correlation coefficients for maximum principal strain with open field behavioural scores .......................................................................................................................... 64 Table 3.5: Correlation coefficients for mechanical impact factors with LFB/MBP analysis ....... 69 Table 3.6: Correlation coefficients for maximum principal strain with LFB/MBP histology...... 70 Table 3.7: Spearman rank correlation coefficients for open field scores with LFB and MBP histology ..................................................................................................................... 71 Table A.1: 22.5 N load cell error .................................................................................................. 91 Table A.2: 500 g accelerometer error ........................................................................................... 92 Table A.3: Instron dynamic impact parameter comparison (1.5 mm displacement, 80 mm/s) .... 93 Table A.4: Instron dynamic impact parameter comparison (1.5 mm displacement, 100 mm/s) .. 94 Table B.1: Total haemorrhage volume immediately following contusion injury ......................... 98  vii  LIST OF FIGURES  Figure 1.1: Anatomical orientation (a) in humans (adapted from Wikimedia Commons [15]) and (b) the rat ...................................................................................................................... 3 Figure 1.2: (a) Anterior view of the human spinal column. (b) Superior view of a typical human vertebra......................................................................................................................... 4 Figure 1.3: Diagram of the rat vertebral column (illustration adapted from Sebek [19]) ............... 5 Figure 1.4: Anatomy of the spinal cord .......................................................................................... 6 Figure 1.5: (a) Illustration of oligodendrocytes and neurons. (b) Cross-section illustration of the myelin sheath ............................................................................................................... 7 Figure 1.6: Divisions of spinal cord white matter........................................................................... 7 Figure 1.7: (a) Major ascending tracts (blue) in the rat spinal cord. (b) Major descending tracts (red) in the rat spinal cord ............................................................................................ 8 Figure 1.8: Summary of maximum displacement vs. impact velocity in recent rat contusion injury studies. ............................................................................................................. 14 Figure 1.9: Strain response of (a) an elastic material and (b) a viscoelastic material................... 17 Figure 1.10: Typical stress-strain response for a (a) linear elastic material and (b) a nonlinear viscoelastic material. .................................................................................................. 19 Figure 1.11: Deformation pattern of a gel filled tube to a contusion type impact using an ink tracer........................................................................................................................... 21 Figure 1.12: Example of contusion injuries with (a) different impact depths and velocities and roughly similar area under the curve.......................................................................... 22 Figure 1.13: FEA model of the rat cervical spine in PAM-CRASH developed by Russell et al. [68] ............................................................................................................................. 23 Figure 1.14: Correlations between maximum principal strain and axonal membrane permeability in contusion (left, blue) and dislocation (right, red), using pooled data from all four white matter regions (A and B), and data from the ventral grey matter (C and D). .. 24 Figure 1.15: 3D strain geometry for a finite element in the x-axis only. ...................................... 25 Figure 2.1: UBC Multi-mechanism injury system set-up ............................................................. 29 Figure 2.2: Inertial load compensation following a blank air impact ........................................... 31 viii  List of Figures Figure 2.3: A typical force vs. impactor displacement curve recorded from a C5/C6 midline contusion injury.......................................................................................................... 35 Figure 2.4: Scoring categories for grooming analysis .................................................................. 37 Figure 2.5: Open field used for behavioural assessment .............................................................. 39 Figure 2.6: LFB stained cross-section sectioned into intact white matter (red), damaged white matter (yellow), intact grey matter (blue), and damaged grey matter (green) ........... 41 Figure 2.7: Stitched and immunostained cross-section at the injury epicentre at 20x magnification ............................................................................................................. 43 Figure 2.8: Systematic random sampling of the ventral region with the 75x75 µm overlay at 40x .................................................................................................................................... 44 Figure 2.9: 25x25 µm axon counting region of interest. Examples of healthy axons (yellow), pathological axons (purple), demyelinated axons (teal), and swollen axons (white) are shown. .................................................................................................................. 44 Figure 2.10: C3-C6 segment of the rat spine used to simulate experimental contusion injuries .. 48 Figure 2.11: FEM cross-section regions of interest: lateral (yellow), ventro-lateral (blue), ventromedial (red), dorsal (green), and grey matter (purple). The grey and pink elements represent the grey and white matter respectively which were not part of any defined region (A) Cord regions used for comparison with experimental data spaced at 1 mm intervals (B)................................................................................................................ 48 Figure 3.1: Element size (top) and maximum principal strain distributions (bottom) in the original model (A) and the medium dura mesh (B). .................................................. 52 Figure 3.2: Comparison of impactor tip and maximum contact force between the fine and medium mesh applied to the dura. The displacement of all three versions of the model was identical. ................................................................................................... 52 Figure 3.3: Element size (top) and maximum principal strain distributions (bottom) in (a) the original model, (b) medium grey/white matter mesh, and (c) refined grey/white matter mesh. ............................................................................................................... 53 Figure 3.4: Comparison of impactor tip and maximum contact force between the fine, medium, and refined medium mesh applied to the grey and white matter parts ...................... 54 Figure 3.5: Experimental displacements and contact forces with the dura plotted alongside the simulated displacement and force traces. ................................................................... 55 ix  List of Figures Figure 3.6: FEA distribution of maximum principal strain direction during a 1.5 mm impact depth at 800 mm/s simulation. ................................................................................... 56 Figure 3.7: Average maximum principal strain distributions shown rostrocaudally at 1 mm intervals for each injury group ................................................................................... 57 Figure 3.8: Representative force, displacement, and velocity curves for each injury group: (a,c,e) 0.9 mm impact depth injuries, (b,d,f) 1.5 mm impact depth injuries. ............. 59 Figure 3.9: Average energy applied to the spinal cord for each injury group. ............................. 60 Figure 3.10: Rearing test showing initial contact using only the left paw (a), right paw (b), or both (c) for weight support. (d) Shows the proportion of paw preference during initial rearing contact between all injury groups at baseline and seven days postinjury. ......................................................................................................................... 61 Figure 3.11: Grooming test at baseline and seven days post-injury. ............................................ 62 Figure 3.12: Open field behavioural results. ................................................................................. 63 Figure 3.13: Correlations of impact velocity and depth with behavioural scores ........................ 64 Figure 3.14: Correlations of maximum principal strain grey matter and pooled white matter maximum principal strain in the cord simulated by the FEA. ................................... 65 Figure 3.15: (a) Luxol fast blue stained spinal cord tissue at 1 mm increments........................... 66 Figure 3.16: The percentage of damaged axons in the ventral region of the spinal cord tissue. .. 67 Figure 3.17: The percentage of damaged axons in the lateral region of the spinal cord tissue. ... 68 Figure 3.18: Correlations of impact velocity and depth with MBP/LFB histological results ...... 69 Figure 3.19: Correlations of maximum principal strain with (a) % spared white and (b) % spared grey matter determined by a LFB stain. Maximum principal strain correlations with % of damaged or demyelinated axons are shown in the (c) ventral region and (d) lateral region of the spinal cord.................................................................................. 70 Figure 3.20: Scatterplot Spearman rank correlations of open field behavioural and grooming scores with MBP/LFB histology ................................................................................ 71 Figure A.1: Load reported vs. calibrated load applied (22.5 N load cell) .................................... 91 Figure A.2: Acceleration reported vs. calculated acceleration (500 g accelerometer) ................. 92 Figure A.3: DynaMight Instron vs. UBC Machine impact characterization (1.5 mm displacement, 80 mm/s) ............................................................................................. 93  x  List of Figures Figure A.4: DynaMight Instron vs. UBC Machine impact characterization (1.5 mm displacement, 100 mm/s) ........................................................................................... 94 Figure B.1: Fully stitched spinal cord section (Red: Rabbit anti-RBC, Green: Alexa Fluor 488labelled BSA) ............................................................................................................. 97 Figure B.2: Representative images of haemorrhage volume immediately following injury ........ 98 Figure B.3: Force vs. time trace during impact of all injury groups. ............................................ 99  xi  ACKNOWLEDGEMENTS I offer my enduring gratitude to Dr. Thomas Oxland and Dr. Wolfram Tetzlaff for their support along this path in a new and multidisciplinary field. Tom has provided me with guidance, optimism, and created a comfortable research environment for me while Wolf has met my modest neuroscience background with patience and enthusiasm to teach.  Many thanks to Dr. Jie Liu, Peggy Assinck, Colin Russell, Tim Bhatnagar, Kinon Chen, Hannah Gustafson, Dr. Peter Cripton, and Stephen Mattucci for their help, in one way or another, over the years.  I owe special thanks to my wife, Miriam, and to my family for their encouragement and moral support while completing this work, which I could not have done without.  xii  Chapter 1 Introduction 1.1 Overview and motivation Traumatic spinal cord injury (SCI) occurs when the spinal column experiences failure most commonly caused by motor vehicle collisions, falls, sports related accidents, work related accidents, and violence [1, 2, 3]. These injuries often result in long term disability and have incidence rates estimated at 1,785 per year in Canada and 12,000 per year in the United States [2, 4]. As a result these injuries create health care costs of an average of $1.1 million to $3.5 million USD over the lifetime of an individual, which does not include the costs associated with loss of productivity [5]. Much current research has focused on the understanding of the mechanism of spinal cord injury and the resulting effects. Failure of the spinal column can cause a variety of different mechanisms of primary injury. Burst fractures, one of the most commonly seen injuries, result from compression of spinal cord causing the vertebral bodies to fracture and impinge on the spinal cord [1, 3], an injury mechanism known as a contusion.  Bone  fragments from fracture cause this contusion, and ongoing compression, of the spinal cord. The movement of one vertebra anteriorly or posteriorly to the adjacent one, thus also the corresponding regions of the spinal cord, causes a dislocation injury. Stretching of the spinal cord is known as a distraction injury mechanism and, while less common than contusion or dislocation injuries, still results in severe neurologic deficit.  Each of these injury  mechanisms creates a unique pattern of damage to the spinal cord and pathophysiological processes, which we do not yet fully understand [3, 6, 7]. In addition to the different mechanisms which cause SCI, the population which sustains the injury is diverse in age, gender, ethnicity, and social or economic status. These injuries also span different severities and level of injury, potentially occurring higher or lower in the spinal cord. The disregard of this fact in many clinical trials has been suggested as a potential reason for the failure to develop effective clinical therapies [8]. More recently, the importance of the mechanism of primary injury is being explored as an important factor 1  Chapter 1 Introduction to being able to understand SCI. Each of these injuries may be a result of a variety of spinal cord injury mechanisms, cord compression depths, and impact velocities; it is important to examine these effects on injury to help guide the development of targeted treatments. SCI is characterized by an initial primary (immediate) injury causing cell death around the injury site. Over the next few minutes to days a cascade of biological responses cause the formation of a lesion, or glial scar, around the injury site which creates secondary injuries and further degeneration [3, 9]. Many of the cells which are lost in the white matter are oligodendrocytes and result in the demyelination of nearby axons. This demyelination in the spinal cord has been suggested to contribute to physiological impairment [10, 11, 12, 13]. Contusion injuries have shown more severe and extensive demyelination than transection injuries and the lesion spans a much larger area [14]. The variation of injury velocity and impact depth in contusion injuries has been thought to lead to differences in injury morphology. This thesis examines these mechanical effects independently and interactively in the rat model to attempt to describe relative importance of these parameters in determining injury severity. Within the Orthopaedic and Injury Biomechanics Group (OIBG) at the University of British Columbia this work continues research in both primary and secondary response to SCI after mechanically controlled injuries and in computational models. The motivation of this study is to combine both the experimental and computational work to attempt to understand how mechanical injury factors result in a range of functional outcomes, biological response in the spinal cord, and mechanical strain patterns in the spinal cord after secondary injury.  1.2 Anatomy of the spine and spinal cord 1.2.1  Anatomical planes The human and quadruped animal body can be split up by three orthogonal planes as  shown in Figure 1.1. In humans, the frontal or coronal plane divides the body into anterior (towards the front of the body) and posterior (towards the back of the body) directions. The transverse plane separates the body into superior/cranial (towards the head) and inferior/caudal (towards the feet) directions. In both humans and animals the body can be  2  Chapter 1 Introduction separated into left and right by the sagittal plane where lateral describes a direction away from the plane and medial describes a direction towards it. Animals can also be defined by a transverse plane which divides the body into rostral (towards the head) and caudal (towards the tail) directions. The frontal plane in quadrupeds separates the body in dorsal (towards the back) and ventral (towards the front) directions.  Figure 1.1: Anatomical orientation (a) in humans (adapted from Wikimedia Commons [15]) and (b) the rat (Adapted From Wingerd [16]. Reprinted with permission of The Johns Hopkins University Press.)  1.2.2  Anatomy of the spinal column The vertebral column is structurally complex, both geometrically and mechanically,  and surrounds the spinal cord to protect it from trauma. These bones, called vertebrae, travel axially along the body and have soft tissue intervertebral discs found between them which help support weight and permit a significant range of motion.  Each vertebra is also  connected to the adjacent vertebrae through a system of ligaments. The human vertebral column can be separated into five different regions: cervical, thoracic, lumbar, sacral, and coccygeal with seven, twelve, five, five, and four vertebrae in each region respectively for a total of 33 vertebrae (Figure 1.2A).  Each of these vertebrae can be separated into the  following major components: a vertebral body, a vertebral arch, three processes for muscle attachment (one spinous and two transverse), and four articular processes which form joints between vertebrae (Figure 1.2B).  3  Chapter 1 Introduction  Figure 1.2: (a) Anterior view of the human spinal column. (b) Superior view of a typical human vertebra. (Adapted From Rizzo [17]. Delmar's Fundamentals of Anatomy and Physiology, 1E. © 2001 Delmar Learning, a part of CengageLearning, Inc. Reproduced by permission. www.cengage.com/permissions [17])  The human cervical spine shares many commonalities with its rat counterpart, which is the one of the most commonly used animal models in spinal cord injury research [18]. Figure 1.3 shows the regions of the rat’s vertebral column.  4  Chapter 1 Introduction  Figure 1.3: Diagram of the rat vertebral column (illustration adapted from Sebek [19])  1.2.3  Anatomy of the spinal cord The spinal cord travels along the spinal column enclosed by the vertebral body and  vertebral arch known as the vertebral foramen/canal and is covered by three membrane layers called the pia mater, arachnoid mater, and dura mater (Figure 1.4). Directly sheathing the cord and its vessels is the pia mater which is surrounded by a subarachnoid space containing the cerebrospinal fluid (CSF). Superior to this layer is the arachnoid mater followed by the outermost layer, the dura mater, which is in immediate contact with the vertebral foramen. Between each vertebral level there is a pair of nerve roots which carry signals to and from the central nervous system (CNS) to the peripheral nervous system (PNS). The spinal cord is made up of two types of tissue, grey matter, found in a butterfly formation at the centre of a cross-section of the cord, and the surrounding white matter (Figure 1.4). The grey matter contains a combination of nerve cells and neuroglial cells (support cells) while the white matter is mostly made up of axons that transmit the signals between neurons. Grey matter is primarily responsible for the processing and control of signals at a specific vertebral level, while the white matter carries signals longer distances along the spinal cord to the brain and brain stem. Disruption of the grey matter at a given level usually results in localized deficiencies and defects; however, damage of the white matter prevents signal transmission at all levels below the level of injury.  5  Chapter 1 Introduction  Figure 1.4: Anatomy of the spinal cord (From Rizzo [17]. Delmar's Fundamentals of Anatomy and Physiology, 1E. © 2001 Delmar Learning, a part of Cengage Learning, Inc. Reproduced by permission. www.cengage.com/permissions)  Neuronal cell bodies contained in the grey matter communicate via electrochemical signals, known as action potentials, to other neurons through a system of branched dendrites in response to a given stimulus. In contrast, white matter primarily consists of axons which carry these action potentials long distances throughout the body. These axons are analogous to insulated wires carrying electricity through them and are encased by a material known as myelin.  This myelin acts as an insulator and capacitor to help increase the signal  transmission speed and prevent the electrical current from leaving the axon. There are periodic gaps in this insulation along the length of the axon, called nodes of Ranvier, which contain voltage gated ion channels to allow the propagation of action potentials by altering the voltage across the axon membrane (Figure 1.5B). Oligodendrocytes are a type of cell located in the CNS responsible for creating these myelin sheaths that wrap the axons and can form many segments of myelin (Figure 1.5A). If these cells become damaged or are unable to generate these myelin sheaths it will inhibit the ability to send signals down the length of the spinal cord below that level. 6  Chapter 1 Introduction  Figure 1.5: (a) Illustration of oligodendrocytes and neurons. (b) Cross-section illustration of the myelin sheath ((a) From Rizzo [17] Delmar's Fundamentals of Anatomy and Physiology, 1E. © 2001 Delmar Learning, a part of Cengage Learning, Inc. Reproduced by permission. www.cengage.com/permissions)  The grey matter is separated into three separate regions, the dorsal/posterior and ventral/anterior horns of the butterfly shape, and the intermediate zone found in between the two (Figure 1.4). Figure 1.6 displays the division of white matter into three bundles of nerve fibres bound together, called funiculi: the dorsal funiculus, the lateral funiculus, and the ventral funiculus. Each of these regions contains ascending and descending tracts made up of bundles of axons that connect different parts of the CNS (Figure 1.7). Tract locations are approximations based on pathological, clinical, and animal studies.  Figure 1.6: Divisions of spinal cord white matter  The dorsal funiculus contains mostly ascending tracts which are all ipsilateral, while the lateral funiculus consists of both ascending and descending tracts. In the cervical region 7  Chapter 1 Introduction of the spine these ascending tracts in the dorsal funiculus are most concerned with the conscious awareness of movement and position of upper limbs along with the sensation of light touch and vibration. Axons that carry the lower limb proprioception modality in this region travel only as far as the thoracic region of the cord. Continuing upward, the pathway for proprioception in the lower limbs is located in the lateral funiculus. In humans, the corticospinal tract is a relatively large group of axons which mainly control the ability for voluntary, discrete, and skilled movements and has been thought to be important to motor control in non-primate mammals. The rubrospinal tract is rudimentary in humans; however, in rodents this tract is substantial and is also responsible for large muscle movement and fine motor control of the upper limbs.  Axons in the reticulospinal tract are involved in  movement-related activities and autonomic functions [20]. The ventral funiculus mainly contains descending tracts with the medial and lateral vestibulospinal tracts also involved in controlling movement-related activities and coordination of limbs [21, 22].  Figure 1.7: (a) Major ascending tracts (blue) in the rat spinal cord. (b) Major descending tracts (red) in the rat spinal cord (adapted from Watson [22])  8  Chapter 1 Introduction Lesions in the spinal cord caused by trauma can disrupt both the ascending and descending tracts and cause impairment if specific tracts become damaged. The organization and cross-sectional area of both the grey and white matter vary along the cranial-caudal length of the cord, with the area of the white matter decreasing caudally as there are fewer ascending and descending axons. The cervical spinal cord consists of the largest range of tracts responsible for function and can serve both the upper and lower body, including control of breathing and voluntary movement; injury at the cervical level is often considered to be the most potentially disabling.  1.3 Overview of spinal cord injury mechanisms There are various different mechanisms of SCI which can be caused by multiple combinations of forces and moments applied to the spine during an injury event which are mechanically distinct from one another [6, 23]. If the spinal column becomes disrupted by these forces it can cause damage to the spinal cord resulting in injury. These injuries can be classified into groups based on their cause and injury patterns and a brief overview of some of the most prominent injury mechanisms will be covered in this section.  1.3.1  Contusion models Vertebral burst fractures are one of the most common fracture types in clinical SCI  and account for approximately 30-50% of vertebral column injuries in human adults [1, 3]. These burst fractures result in small bony fractures of the vertebral column which eject into the spinal cord, a primary injury mechanism known as a contusion, causing damage. Immediate injury results in damage to the vascular structure and haemorrhaging of the grey and white matter around the impact site causing tissue necrosis and formation of a central cavity encapsulated by a glial scar in the spinal cord. Additionally, a few axons around the injury epicentre are severed and are lost as a result of the primary injury.  Cascading  biological responses result in the migration of macrophages, followed by neutrophils that both clear the cellular debris around the lesion. This process creates enzymes and free radicals that damage healthy tissue and expand the lesion [24]. Neuronal apoptosis, an actively controlled form of cell death, of oligodendrocytes peaks at 7 to 8 days post-injury causing the demyelination of axons surrounding the lesion and further neurologic impairment 9  Chapter 1 Introduction [25, 26].  A better understanding of the relationship between the mechanical injury  parameters and the  resulting biological  responses,  such as  demyelination  and  pathophysiology of the axons in the spinal cord which ultimately affect physical deficit, may help elucidate the mechanics of SCI. Creating a reliable and reproducible injury which simulates a burst fracture is difficult given the variability in vertebrae geometry, internal stresses, and other factors between cases of SCI which would determine the speed and direction of bone fragment ejection. In order to study this injury, the use of an impactor at a given velocity, depth, or force to injure the spinal cord has been used to remove the fracturing of the vertebrae from experimental design. The contusion mechanism currently being used consists of a rapid compression of the spinal cord from the dorsal surface in an anterior-posterior direction. Although clinically these injuries occur on the anterior surface of the spinal cord, it has proven difficult to access the anterior surface with an impactor head without creating additional complications for the animals. In an attempt to mimic the blunt force created by a contusion injury Allen developed a basic weight drop model in 1911 [27], whereby an impactor of known weight followed a guided release from a known height. Given the clinical relevance of this model other researchers created their own weight drop models for use on different animals models [28, 29, 30, 31]. However, the caveat to using this model is that the control of the injury severity is determined by the height by which the impactor is dropped from. An advantage of this model lies in its low cost and the simplicity of the single g-cm control parameter defined as the product of the impactor mass and drop height. However, as this parameter does not provide a meaningful physical representation of injury parameters, the definition remains abstract. Increasing this control factor will result in an increase of impact velocity, but also intrinsically the energy applied to the cord, and thus the impact depth and impact force of the injury. As a result, the impact velocity, depth, and force of the weight drop model are coupled together. Simplistically, considering the acceleration of an object falling from a given height it can be shown that increasing the height of the drop will result in a square root increase of velocity. Increasing drop height will also increase the energy available to do work on the cord resulting in an increase in the force and depth of injury.  10  Chapter 1 Introduction √  √  Where,  Since these early contusion models were developed, improvements have been made to contemporary impactors which allow for more precise and reproducible control of injury parameters, such as impact velocity, depth, and force. A general consensus of the use of a 10 g impactor mass with a diameter of 2.5 mm has been used in more recent studies [30, 32, 33]. In rat spinal cord injury, typical drop heights of 12.5, 25, and 50 mm have been used by Basso et al. to show graded histological and behavioural outcomes [34]. The experimental contusion impact model has developed over time since Allen’s original weight drop model in 1911, ultimately attempting to achieve the most reliable form of injury parameter control and measureable outcomes. Allen’s weight drop model provided a crude method, by today’s standards, to create similar injuries in an animal model which were defined by the mechanics of a falling object from a predetermined height. With current technology, specifically control devices, the development of negative feedback controlled force, displacement, and velocity impactors have provided more precise control over these older models. The use of impactors with definable impact parameters has allowed for more accurate prediction and observation of the impactor and cord kinetics from contusion SCI. The most widely used contusion injury impactor design is the New York University (NYU) impactor [18, 33, 34]. This device improves on the basic principles of Allen’s weight drop model and uses a round tipped 10 g rod dropped from a specified height, ranging from 6.25-75 mm, on to the dorsal surface of the rat spinal cord resulting in a range of chronic injuries [30]. The NYU impactor allows for the measurement of impact velocity, impact depth, cord compression rate, and forces applied to the cord. This device was used with the rat model and produced reproducible functional deficits measured and defined by Basso et al. at different severity levels [34]. However, all weight drop contusion models retain the inherent coupling between drop height, impact velocity, impact depth, and contact force, 11  Chapter 1 Introduction preventing the ability to control velocity and impact depth independently discussed below in chapter 1.4. A slightly different take on the rat contusion model, the Ohio State University (OSU) impactor, was first used in 1987 and employed a computer controlled mechanism allowing for feedback of impactor displacement during injury [35, 36]. Control by an electromagnetic motor allowed for much more precise prediction and execution of impact depth in injuries and measured both force and displacement using two separate transducer systems. The use of this device allows for the examination of the effect of physical parameters aside from the g-cm unit defined by weight drop models. Sparrey et al. used a modified OSU impactor to examine the contribution of impact velocity at similar impact depths on injury severity, concluding that velocity greatly affected the white matter injury magnitude [7]. This result suggested the importance of the viscoelastic response of the spinal cord. With both the OSU and the NYU impactors having undergone revisions since the original conception of either device there is no clear superiority of one over the other. While the previously discussed impactor devices have been the most widely used in the field there are other alternatives also being used to study contusion injuries such as the Infinite Horizons (IH) force-controlled impactor [37, 38] and custom-built pneumatic impactors [39, 40, 41]. Scheff et al. showed that increased contusion force at the thoracic level of the adult rat resulted in less spared tissue which correlated well with locomotor ability [37]. Velocity was not recorded in this experiment, but the data did show increased impact depth with the increased force. Anderson et al. also used the IH device to create mild (200-230 kdyn) and moderate (250-290 kdyn) midline contusion injuries in female SpragueDawley rats [38]. Neither impact depth nor velocity was reported in this study, but graded tissue damage and behavioural deficits were observed. As both of these impact factors can affect the resulting force, it is difficult to determine the individual effect of each. Kearney et. al employed a custom, stroke constrained pneumatic impactor and modeled contusion injuries in a ferret model varying both velocity and the relative compression of the spinal cord independently [39].  The stroke of the impactor was  controlled by a mechanical stop to prevent motion, which allowed for precise control of the impact depth, but did not allow for accurate force or energy measurements. The ability to study contusion injuries across a wide range of controlled velocities and depths by Kearney 12  Chapter 1 Introduction et al. showed evidence that a viscous criterion, the product of impact velocity and relative cord compression, performed well as a predictor of injury severity.  Choo et al. have  developed their own electromechanically controlled impactor device capable of creating the three main types of SCI mechanisms including contusion, distraction, and dislocation and allowing for the independent control of impact velocity and impact depth [6]. The design of modern impactor devices allow the force, displacement, and velocity to be monitored during injury. Experimental injury conditions studied in rodent models are summarized below in Figure 1.8. Comparisons between these studies are not easily made as each experimental design differs from most others. In many contusion injury models, the coupling between the impact velocity and impact depth has created a concentration of the current data along a linear corridor. The interactive effects of the impact factors will alter the impact energy applied to the cord and may play a role in determining injury severity. Some studies have used a different strain of rat, gender, impactor, or loading method leaving the need for a study across different velocities and displacements to elucidate their individual and interactive effects on injury severity. Furthermore, these studies do not address the concept of the energy applied to the cord by the impact as a result of the differing impact depth and velocity. This gap also leaves the question of whether physical and biological outcomes in SCI are dependent purely on the energy applied to the cord on impact, how the impact occurred, or some combination of the two.  13  Chapter 1 Introduction  Figure 1.8: Summary of maximum displacement vs. impact velocity in recent rat contusion injury studies. A linear pattern highlighted by the red corridor is clear due to the coupling of the impact velocity and depth in many contusion models. Injury severity is relative within each individual study [6, 7, 30, 33, 34, 36, 37, 42, 43, 44, 45, 46, 47]  1.3.2  Dislocation and distraction models Spinal  column  fracture  dislocation,  where  one  vertebral  body  moves  anteriorly/posteriorly relative to the adjacent vertebrae, causes spinal cord shearing and has been shown to contribute to approximately 30-40% of SCI in humans [1, 3]. Despite the large contribution to current SCI in humans, the application of this injury mechanism in an animal model was not developed until recently, due to the difficulty in creating reproducible injuries. In the dislocation injury model, the displacement and velocity of the vertebrae are controlled, rather than the movement of the cord itself. Due to this fact, it is more difficult to control injury severity as some energy is absorbed by the intervertebral discs and other components of the spinal column. Insult to the spinal cord by dislocation results in vascular damage to the spinal cord, as in the contusion injury; however the morphology of the  14  Chapter 1 Introduction haemorrhage has been shown to be diffuse and cause considerable axonal damage in the tensile regions of the cord [6, 48]. Fiford et al. [48] examined the primary injury effects of controlled vertebral displacement and velocity in lateral dislocations in the rat. This model did not require the spinal cord to be exposed and left the vertebral column fully intact and showed an ability to produce a graduated injury response by varying the dislocation parameters. Following this, Choo et al. [6, 49] developed a multi-mechanism injury device capable of creating anteriorposterior dislocations injuries in the rat cervical spine. The device was used to examine both the primary and secondary effects and showed a visible difference in lesion morphology, acute haemorrhage, and membrane compromise, with the contusion injury mechanism, using this model. Dislocation injuries show similar formation of a lesion at the injury epicentre with necrosis focused in the grey matter. However, the membrane compromise was visible over several vertebral levels rostrally and caudally. Clarke et al. also studied the age effects of the spinal cord in vertebral dislocation and simultaneously showed that, as seen clinically, anterior-posterior dislocation is noticeably more severe than lateral dislocation [50, 51]. Distraction injuries occur in the human population occur when the spinal cord is stretched, often resulting from hyper flexion or hyper extension of the neck. These injuries make up roughly 5% of human SCI injuries [3] and have been less prominently studied due to the smaller incident rates and the difficulty to create a consistent model. The distraction injury model, like dislocation, controls the displacement and velocity of the vertebrae as opposed to direct application to the cord. Spinal cord distraction has been shown to also cause ischemia and the cascade of biological responses that follow, similar to the other two mechanisms previously mentioned; however, the lesion morphology differs and the biomechanics of this injury mechanism remain unclear [52, 53].  Maiman et al. used  fluoroscopy to estimate the amount of coupling between the movement of the vertebral column and that of the spinal cord during distraction and showed that the movement of the spinal cord was always less than the vertebral column [54]. In addition to the dislocation model, Choo et al. [6, 49, 55] also created a distraction injury model which displayed the potential presence of mechanism specific injury regions.  15  Chapter 1 Introduction  1.3.3  Transection and clip compression models Transection injuries are characterized by a complete disruption of the spinal cord  ensuring the completeness of the injury and allowing for specific control of function loss based on the level and depth of the cut. This injury model is frequently used in research for its ease of implementation and ability to predict the resulting injury. These types of injuries generally occur in the human population as a result of violence, such as gunshot injuries to the spinal cord, but represent only a very small portion of SCI seen clinically [9, 18]. Transection injuries can either be complete, where the entire cross-section of the spinal cord is severed, or partial, where specific tracts of the cord are cut. This type of injury model has proved to be useful to ensure the completeness of an injury or allow specific targeting and identification of important regions of the spinal cord. Siegenthaler et al. examined the myelination pathology using both the transection and contusion injury model showing more widespread demyelination resulting from contusion among other substantial differences [14]. SCI in humans often results in ongoing spinal cord compression due to lasting spinal column displacement. The contusion models discussed in section 1.3.1 create a single, rapid impact to the cord and do not simulate the residual compression. In 1978, Rivlin and Tator [56] developed a clip compression model whereby the spinal cord was compressed by a modified aneurysm clip for a given amount of time. This model highlighted the importance of compression duration on SCI severity. Both the duration of compression and the force applied by the clip are the factors which determine the injury severity in clip compression. Although this model has been useful in determining the importance of decompression timing and potential treatment therapy it does not address the rapid impact of SCI and is difficult to compare dynamically to the more common contusion model.  1.4 Spinal cord injury biomechanics 1.4.1  Material properties of the spinal cord The macroscopic behaviour and mechanical response of spinal cord tissue are still  being closely studied to characterize the material. Both the grey matter and the white matter of the spinal cord are composite materials made up of different proportions of neuronal cells,  16  Chapter 1 Introduction axon tracts, and vasculature which contribute to its holistic response to deformation and force. A material which returns to its original form after experiencing strain is known as an elastic material and is independent of time (Figure 1.9a) and a viscous material resists strain (the ratio of the change in a dimension over the original dimension) or deformation at a given stress creating a time dependent component.  The characterization of a linear elastic  material’s deformation response is quantified by Young’s modulus, an intrinsic and geometrically independent material property, described by the ratio of stress (force per unit area) to strain. If the ratio of stress over strain is constant over a range of strains it is known as a linear material. Viscous deformations display a response to the rate of deformation in addition to the applied stress. Given the presence of fluid components in the spinal cord it is characterized by both elastic and viscous response components and is classified as a viscoelastic material. The response of viscoelastic materials (Figure 1.9b) can also be either linear, where the function can be separated in both an elastic and viscoelastic portion, or nonlinear, where this function is not separable into its respective parts. Hyperelasticity of the spinal cord is an example of a nonlinear response which is defined by increasing stiffness with increasing strain rate. Therefore we can say that a material displays linear or nonlinear and elastic, viscous, or viscoelastic properties.  Figure 1.9: Strain response of (a) an elastic material and (b) a viscoelastic material  17  Chapter 1 Introduction Material deformation can be either elastic which are recoverable as stored energy or plastic which can only be recovered if more energy is applied to the system. Viscous materials also display deformation rate dependence where larger rates characteristically result in more energy loss.  The time dependence of the material response emphasizes the  importance of understanding impact velocity as a useful impact factor in determining the spinal cord response, both biologically and mechanically. In linear elastic materials all energy applied upon impact is returned during unloading. Due to the viscoelastic nature of the spinal cord, this is not the case and a component of that energy is absorbed by the cord itself during loading and unloading, resulting in a phenomenon known as hysteresis. This can be seen in a typical stress-strain curve following impact of the spinal cord compared to a purely elastic material (Figure 1.10). Linear viscoelastic models yield the following stress and strain equations as a function of time: ( ) ( )  ( ) ( )  ∫ ( ∫  (  ) ̇( ) ) ̇( )  Where, ( ) ( )  ( ) ( )  Creep is the response where, over time, tissue strain will increase when subjected to a constant stress, while relaxation displays a decrease in tissue stress over time subjected to a constant strain.  The creep and relaxation functions are important in determining the  mechanical response of the cord and define how energy is lost during a loading cycle.  18  Chapter 1 Introduction  Figure 1.10: Typical stress-strain response for a (a) linear elastic material and (b) a nonlinear viscoelastic material. The area between the load and unloading curves is the amount of energy lost during loading. Hysteresis shown in (b) is one manifestation of viscoelasticity.  Little is known about the biomechanical properties of the spinal cord and attempts to measure the overall behaviour of the material, which represents the aggregate response of the separate cellular elements, have resulted in a variety of different results. Difficulty with the accuracy and repeatability of measuring the material properties of the spinal cord has prevented the agreement on a single set of definitive values. Many studies have investigated simple uniaxial tensile tests on the spinal cords of various animals and human cadavers at different rates to define the stress-strain response of the tissue. Studies have shown that the preconditioning factors of repeated strain or stress relaxation cycles can affect the mechanical response of the spinal cord and other soft tissue during testing [57, 58]. Fiford and Bilston applied a range of strain, from 2% to 5%, in vitro to a rat spinal cord over a period of at least 30 minutes to observe the stress relaxation over time [59, 60]. Each specimen displayed hyperelasticity and a viscoelastic stress-strain response. Clarke et al. repeated a similar test applying a 2% to 5% strain range at three different strain rates, but focusing on the age dependence of the strain response in the rat spinal cord [61]. Although age of the animal was shown to have an effect on the magnitude of the stress-strain response of the spinal cord, the patterns of stress relaxation remained similar and increasing strain rate increased material stiffness. These studies have exhibited the effect on stiffness that control of the velocity of deformation can have on the response of the spinal cord. In addition to the mechanical response of the spinal cord as a whole, the white and grey matter each have a unique composition of components that determine their properties. Ichihara et al. attempted to research the mechanical difference in vitro between the two 19  Chapter 1 Introduction materials using a bovine spinal cords [62]. For both the grey and white matter, a nonlinear region followed by a linear region of the stress-strain response was observed before failure. The grey matter exhibited higher stress and earlier failure than the white matter in the linear region of the curve before rupture suggesting an explanation as to why cavitation after SCI typically begins in the central region of the cord. The inability to reach a general consensus for the material properties of the spinal cord, due to testing in vivo versus in vitro; across different species; and between grey and white matter has proven problematic; however, a general nonlinear viscoelastic, time dependent, response of the spinal cord has been shown to define mechanical deformation of the cord.  1.4.2  Impact mechanics and kinetics Contusion injuries have previously reported concentrated damage around the  epicentre of injury and spread outward within a few millimeters [42]. This pattern of damage and location of maximum displacement of the cord was studied by Blight and Decrescito in 1986 using an animal model and a gel filled tube [63]. A primary area of tissue displacement was shown in the central region where the spinal cord lesion typically begins. While this does not fully explain the phenomenon behind secondary injury in the spinal cord concentrated mechanical strain patterns were observed (Figure 1.11). The composite neural tissue is typically considered incompressible and compression in the dorsal-ventral direction results in strain both rostrocaudally and laterally. Generally, contusion injuries result in a central region of large displacement and a concentration of high strain around the epicentre is seen. These strains elicit a biological cascade of events discussed in section 1.3.1 causing secondary injury. Various different kinetic and kinematic characteristics in soft tissue injury have been investigated as predictors to injury severity. Viano and Lau [64] classified a tolerance criterion for determining soft tissue damage based on tissue compression and velocity of deformation of the chest, highlighting the importance of incorporating material viscoelasticity when considering injury risk. Kearney et al. also observed the effects of spinal cord contusion depth and velocity in ferrets and found a similar correlation with a viscous criterion [39].  Specifically, the product of impact velocity and percentage  compression of the cord diameter (achieved by either a large impact depth at low velocity or 20  Chapter 1 Introduction a smaller impact depth at high velocity) was shown to result in similar somatosensory evoked potentials, used as a measure of functional ability.  Figure 1.11: Deformation pattern of a gel filled tube to a contusion type impact using an ink tracer. Central regions show areas of maximum displacement rostrocaudally (Adapted from Blight [63]).  Impact velocity and depth can also be seen as an indicator of the energy dissipated causing the deformation of the tissue. By changing either of these parameters the amount of energy applied to the spinal cord is altered. The total energy applied to the spinal cord, rather than only the impact factors alone could also be useful in determining the damage to the spinal cord to help strengthen and explain the observations seen by Kearney [39] and Viano and Lau [64]. Other studies have also investigated the relationship of impact velocity or depth in animal contusion injuries with varying, and some contradictory, results.  Maikos and  Shreiber saw stronger correlations to the rate of spinal cord compression rather than depth of injury of extravasation volume immediately following impact using the NYU impactor [33]. Drop heights of 12.5 mm, 25 mm, and 50 mm were used resulting in impact velocities of 489 mm/s, 690 mm/s, and 974 mm/s coupled with impact depths of 1.65 mm, 2.14 mm, and 2.88 mm respectively. Kim et al. observed an opposing result and found no significant decrease in functional assessment of mice at a constant depth (0.8 mm) with increasing impact velocity (100, 200, and 400 mm/s) [65]. Pearse et al. showed an increase on axon  21  Chapter 1 Introduction demyelination and lower motor function with increasing impact depth (0.80, 0.95, and 1.1 mm) at constant velocity (~100 mm/s) [47]. Sparrey et al. concluded that increasing contusion velocity from 3 mm/s to 300 mm/s shows an effect on the distribution of haemorrhage and damage of axons in the spinal cord at a depth of 0.9 mm and 1.03 mm respectively. Basso et al. saw reduction in locomotor outcomes following increasing drop height (6.25, 12.5, 25, and 50 mm), and impact velocity and depth by definition, using the NYU impactor [34]. Difficulty has been encountered when trying to compare these various studies with one another as each used unique experimental procedures such as: impactor device; animal model; and the variation of either impact velocity, depth, or both. In some of these studies the impact velocity and depth are naturally coupled and an increase in one parameter is often accompanied by an increase in the other which poses a problem in isolating which, if any, factor remains the driving component in determining injury severity. This leaves a need to study injuries where both the impact velocity and depth are modified independently of one another to help define the relative importance of each (Figure 1.12a,b).  Figure 1.12: Example of contusion injuries with (a) different impact depths and velocities and roughly similar area under the curve, which quantifies the energy applied to the system. (b) Contusion injuries where the impact velocity is increased over a constant set of impact depths.  1.5 FE models In addition to experimental studies, much research has focused on modelling these injuries in a computational environment in order to help determine the stress and strain patterns in the spinal cord itself during injury and reduce the amount of live or cadaveric 22  Chapter 1 Introduction specimens required for experimental research. Understanding and determining the material properties of the spinal cord experimentally is an essential component to be able to simulate SCI accurately. Greaves et al. created a human finite element model (FEM) of the cervical human spine using ANSYS to simulate contusion, distraction, and dislocations [66]. Greaves’ model showed distinct strain patterns between each of these three mechanisms suggesting further research into this area.  However, with little experimental data for  comparison, validation of human models remains difficult. The use of animal models, rats in particular, provide a much larger pool of data to draw from for validation due to their relatively low cost and ease of accessibility. A three-dimensional finite element model of the thoracic rat spine simulating a weight-drop experiment was created by Maikos et al. using previous experimental data to determine the material properties of the spinal cord [67]. Russell et al. also developed a dynamic injury model of the rat cervical spine using the FE software PAM-CRASH (Version 7.5, ESI Group, Paris, France) (Figure 1.13) [68]. The geometry of the model was created using high resolution magnetic resonance image and imported into PAM-CRASH. Material properties for the spinal cord and dura were adapted from an earlier study conducted by Maikos et al. [67]. As an extension of the results reported by Greaves, Russell found correlations between maximum principal strain and tissue damage reported from experimental data in both contusion and dislocation injuries (Figure 1.14).  Figure 1.13: FEA model of the rat cervical spine in PAM-CRASH developed by Russell et al. [68]  23  Chapter 1 Introduction  Figure 1.14: Correlations between maximum principal strain and axonal membrane permeability in contusion (left, blue) and dislocation (right, red), using pooled data from all four white matter regions (A and B), and data from the ventral grey matter (C and D). Linear regression lines of best fit are plotted as solid lines, along with 95% confidence intervals as dashed lines. Pooled white matter regions show significant (p < 0.0001) and similar correlations between maximum principal strain and tissue damage for contusion (R2 = 0.86) and dislocation (R2 = 0.52) mechanisms. Correlations within the grey matter are also strong (R2 = 0.93), and the slope of the tissue damage versus strain regression is steeper compared to the pooled white matter result, especially for dislocation. [Figure and caption text taken from Russell et al. [68]]  1.5.1  Strain theory Strain provides a measurement of the deformation of a material over its original  shape. In computational finite element analysis an entire structure can be divided into smaller elements which make up the entire form. The use of these finite elements allows the numerical calculation of strain in each element to provide a precise measurement of strain in the material as a whole. These individual element strains describe the deformation of each small element that is not a result of rigid body motion (such as translation or rotation). Strain is defined by both a magnitude and direction which determines the dimension of deformation. In a three dimensional object this strain can have three components of normal  24  Chapter 1 Introduction strain (perpendicular to the face of an element - Figure 1.15a) and six components of shear strain (along the fact of an element - Figure 1.15b,c).  Figure 1.15: 3D strain geometry for a finite element in the x-axis only. (a) Normal strain in the in the x-axis, (b) shear strain with respect to the y-axis, and (c) shear strain with respect to the z-axis.  For each strain a unique set of axes can be defined where the magnitude of all shear strain values is zero and only the normal strains remain. The three orthogonal strains are referred to as principal strains and their axes as the directions of principal strain. Maximum principal strain,  , also called first principal strain is defined as the largest of the three  principal strains and is often a positive value, indicating tension. Minimum principal strain, , is the smallest and typically negative value, indicating compression. In spinal cord injury much of the soft tissue components, such the axons which run down the length of the cord, are more likely to fail in tension rather than compression or shear [69] and this is somewhat analogous to pushing vs. pulling on a rope. For these reasons, it is expected that maximum principal strain will provide the strongest correlation to tissue damage.  1.6 Project objectives and scope 1.6.1  Objectives Previous work in this field of study has primarily focused on either the effects of  impact depth or impact velocity on SCI, but has yet to closely examine the potential interactive effects of the two and the total energy applied to the cord due to varying impact mechanics. Attempting to vary these impact factors while keeping the energy applied to the  25  Chapter 1 Introduction spinal cord similar has an inherent difficulty with the prediction of the energy itself; applied impact energy may be a function of the spinal cord properties and injury mechanics. By controlling the applied energy with different impact depths and velocities an understanding of the importance of how an injury occurs may become more apparent. The primary goals of this study were to:   Determine the contribution and effects of impact depth and maximum impact velocity in a rat contusion SCI at seven days post injury on behavioural recovery and axonal demyelination;    Examine the effect of different mechanical impact factors which have similar total applied energy to the spinal cord;    Examine the correlations between the maximum principal strain (through finite element analysis (FEA) and computational modeling), behavioural recovery, and axonal demyelination in the spinal cord. From a more clinical perspective, it is hoped that this work might contribute to the  understanding of contusion SCI as a whole and help identify target regions of the spinal cord which may benefit from specific treatment plans.  1.6.2  Scope The focus of this thesis was to carefully examine the relationship between injury  depth and maximum impact velocity in contusion SCI with respect to behavioural outcomes, spinal cord pathology, and correlation to predicted strain patterns. Choo et al. developed a multimechanism injury system used in this research which provided the capability to control both the impact depth and velocity of injury simultaneously [6, 49, 55]. It was hypothesized that these different injury parameters have a significant effect on secondary damage caused by the strain on, rather than the severing of, axons contained in the spinal cord. Much research has been done examining the effects of impact velocity and depth separately; however, the importance of this work lies in the ability to control both the velocity and depth individually and the inclusion of behavioural analysis and computational FEA. This research adds additional complexity by attempting to elucidate correlations between injury kinetics and its longer term effects on the resulting injury severity. 26  Chapter 2 Methods 2.1 Experimental overview To evaluate the effects of both the impact velocity and impact depth of a contusion SCI a custom electromagnetic displacement controlled impactor was used to control the injuries. This impactor was developed and validated by Choo et al. [6] in 2007 and will be referred to as the UBC multi-mechanism injury system (Figure 2.1). The design of the machine was based around an electromagnetic linear actuator (TestBench ELF LM-1, Bose Corporation, Eden Prairie, MN) with a nominal stroke of ±6 mm. This actuator was mounted to a rotary axis on a radially translating arm which was secured to a large motorized z-axis (Linear Stage 2DB160UBW-SL, Thomson Industries, Ronkonkoma, NY; Servo Motor, BSM63N-375AA, Baldor, Fort Smith, AR; Controller FlexDriveII, Baldor, For Smith, AR). This design allowed for coarse and fine positioning of the impactor tip before injury. The specimen platform consisted of a stereotaxic frame (Model 900, David Kopf Instruments, Tujunga, CA) mounted to an x-y table (2.54 mm, 0.1 in lead, pre-loaded antibacklash nut, Thomson). A custom damped vibration table was used (78-111-02DR-SPECIAL, TMC, Peabody, MA) as the base of the system. The electromagnetic actuator was controlled by WinTest software (v2.56, EnduraTEC, Penetanguishene, ON, Canada) under displacement feedback control. Force feedback was also an option to control the actuator, but not used for this work. Displacement of the impactor was measured using a linear variable differential transformer (Model MHR250, Schaevitz Sensors, Hampshire, UK) mounted inside the body of the actuator. A 22.5 N load cell (Model 31, Honeywell-Sensotec, Columbus, OH) was used to measure impact forces, although other interchangeable load cells were available with different resolutions. To measure the acceleration of the impactor a 500 g accelerometer (355B02, PCB Piezotronics, Depew, NY) was used and determined inertial compensation during the dynamic loading of the spinal cord.  27  Chapter 2 Methods Before beginning experimentation, the UBC multi-mechanism injury system was re-characterized to ensure that the load cell and accelerometer were working as designed and that minimal noise had been introduced into the system that would affect the data. The system was examined in both static and dynamic loading to ensure proper calibration of the components. A pilot study (Appendix B) was first conducted consisting of 15 male SpragueDawley animals placed in four different contusion injury groups and a control group to examine primary injury response. Each animal received either a deep (2.0 mm) or shallow (0.7 mm) impact at a fast (800 mm/s) or slow (80 mm/s) velocity. The purpose of this pilot was to determine the immediate difference in primary injury, measured by haemorrhage volume in the spinal cord between varying impact factors and determine if either one had an obviously greater effect than the other and to help define the maximum injury parameters to be used for the main survival study. The aim of the survival study was to examine the effects of these injury parameters on secondary injury after a period of seven days post-injury. A larger experimental matrix was chosen to thoroughly examine the effects of the impact velocity and depth effects with three different speeds (8, 80, and 800 mm/s) at two depths (0.9 mm and 1.5 mm) for a total of 6 injury groups and one surgical sham control group. Behavioral tests were employed in addition to the tissue analysis following injury to quantify the injury deficit in multiple ways and examine any potential correlations. Finite element analysis was conducted in parallel with the contusion survival study. This required modification of the FEM created by Russell et al. [68] discussed briefly in section 1.5. The increased duration of impact due to the slower injury groups required re-evaluation and optimization of the model to reduce the computational costs without losing considerable accuracy. The maximum principal strain measurements output from the FEA were also used to correlate with both the behavioural and tissue analysis obtained experimentally.  28  Chapter 2 Methods  Figure 2.1: UBC Multi-mechanism injury system set-up  2.2 UBC impactor characterization 2.2.1  Static loading A calibrated brass weight set was used to test the response of the 22.5 N bonded foil  strain gauge load cell in both compression and tension. To test the compressive response of the load cell, the weights were placed on the load cell in increasing mass from 0.02 kg to 2 kg, while connected to the data acquisition computer used to control the UBC impactor. To test the load cell in tension the impactor was replaced with an eye hook and the calibrated masses were hung from the impactor set-up in increasing mass from 0.02 kg to 2 kg. The root mean square accuracy for the load cell, used as a measure of average error [70], was 0.10925 N or 0.49% of the full scale reading (Figure A.1).  29  Chapter 2 Methods  2.2.2  Dynamic loading Characterization of the dynamic components of the multi-mechanism injury system  was required before proceeding with experimental use. The accelerometer was characterized using a measure of root mean squared error compared to the response from the LVDT to determine any discrepancies.  Air only impacts were also conducted to determine and  compensate for inertial load on the system due to the weight of the impactor and mounting column. Throughout all testing, data were collected from the load cell, accelerometer, and LVDT so that a fast Fourier transform analysis could be used to determine the presence of any noise in the system. Given that the UBC multi-mechanism injury system was a custom design, an Instron materials testing machine capable of creating similar impact velocities below 100 mm/s was used to compare dynamic response. Each system was set up using the same load cell, impactor tip, and known rubber material for impact. Maximum force, depth, and velocity were recorded and analyzed for differences between the two systems. The first step after examining the static loading response was to study the time dependent response of the system under dynamic loading conditions similar to a real contusion injury experiment. Before any load was applied to the system the accelerometer was characterized by cycling the actuator through the full range of motion (±6 mm) at 1, 5, 10, and 20 Hz and comparing the reported response of the 500 g accelerometer to the second derivative of the linear variable differential transformer. Transducers were sampled at 4 kHz during impactor motion. The root mean square accuracy of the accelerometer was 0.19535 g or 0.08% error of the full scale reading (Figure A.2). Air only impacts at each velocity and depth, were conducted to determine the inertial forces on the load cell [71]. A linear equation including both the force recorded from the load cell and the acceleration from the accelerometer was used to calculate the compensated load caused only by the impact in the form of:  Where a and b are constants, load is measured in N, and acceleration is measured in g. The constants were adjusted through repeated air impact tests with different impact parameters to reduce the inertial forces to less than 10% of the expected maximum forces (Figure 2.2). Maximum forces during contusion injuries were estimated to be in the order of  30  Chapter 2 Methods a few Newtons based on previous studies conducted by our research group [6, 7, 49, 55]. In addition to determining the inertial forces, the data from the blank air impacts was discretized into the frequency domain using a fast Fourier transform algorithm. The resulting data displayed peak frequencies at 660 Hz and 1100 Hz in these blank air impacts to help determine regions of electrical or mechanical noise which could be filtered out.  A  zero-phase Butterworth low-pass filter with a cut off frequency of 630 Hz was applied to the force and displacement data to attenuate the noise.  Figure 2.2: Inertial load compensation following a blank air impact  In order to characterize the system during impact, the displacement and force data were recorded during two different impacts to a depth of 1.5 mm, one at 80 mm/s and the other at 100 mm/s on a cylindrical rubber blank (40A hardness, McMaster-Carr, Santa Fe Springs, CA). The force vs. time and displacement vs. time traces were recorded and compared to an identical impact set up on a standard materials testing device (Model 8841, Instron, Canton, MA) using the same impactor tip and rubber blank at the same velocity (Figure A.3, Figure A.4).  2.3 Survival study Resulting data from the pilot injury study (Appendix B) guided the selection of injury parameters during the experimental design for the longer term, seven day survival study to examine the secondary injury effects of impact velocity and displacement on contusion injury severity. Fifty four male Sprague-Dawley rats were separated equally into six different 31  Chapter 2 Methods injury groups (n = 8 for all groups) and one surgical sham group (n = 6). Each injury group received a maximum impact velocity of 8 mm/s, 80 mm/s, or 800 mm/s and a maximum impact depth of 0.9 mm or 1.5 mm (Table 2.1). Injury depths were chosen based on the pilot study which showed characteristics of bone impact beyond 1.5 mm and an injury that was considered too light below 0.9 mm, which might have failed to yield any useful distinctions from the surgical sham. The maximum impact velocity was chosen based on previous work by Choo et al. [6, 49, 55] and is similar in scale believed to occur in humans [72, 73]. Table 2.1: Experimental injury group design  The mean weight ± SD of the animals used in this study was 337 ± 17 g. Animals were brought in to our vivarium and group housed in enriched, custom-designed environments with food and water freely available one week before injury to allow them to familiarize themselves with their new housing. Animals were monitored carefully during the one week recovery period post-injury before behavioral and tissue analysis. Eighteen animals were brought in every two weeks over a total of six weeks; each animal was randomized into one of the six injury groups or sham group and housed with three other animals per cage. To ensure any effects of animal lot were accounted for, the number of animals in each injury group was spread out evenly over each grouping of eighteen animals.  Animals were  individually handled gently for fifteen minutes daily to familiarize them with being handled by humans to reduce stress both prior to and post-injury.  Pre-surgical In order to provide a baseline behavioural assessment each rat was placed in an open field for four minutes and scored by two independent researchers using a modified Basso, Beattie, Bresnahan (BBB) scale [74]. This also served as a way for the animals to familiarize themselves with the open field before injury. Each animal was also placed in a transparent 32  Chapter 2 Methods cylinder and allowed to explore the environment freely. Data were recorded on video for 15 minutes to assess both the rearing and grooming behaviour of each animal pre-injury using previously defined scoring systems [75, 76, 77]. All behavioural assessment techniques are discussed in detail in section 2.3.1.  Surgical Aerosolized isoflurane was used to deeply anaesthetize animals (5% for induction) and then placed in a nose cone at 1-2% to maintain a surgical plane of anaesthesia (foot pinch and corneal reflexes absent). Subcutaneous injection of buprenorphine (0.03 mg/kg) was given to mitigate any possibility of acute pain. Luke warm Ringer’s lactate solution was provided subcutaneously to each animal to prevent dehydration and Lacrilube ophthalmic ointment was also applied to prevent drying during anaesthesia. Heart rate and blood oxygen levels were monitored using a pulse oximeter throughout the surgical procedure. Following anaesthetization to a surgical plane, the spinal column was exposed from C2 to C7 dorsally by a trained microsurgeon at the International Collaboration on Repair Discoveries (ICORD) research centre, Dr. J.L. A laminectomy was performed between the C5 and C6 vertebral levels to expose the spinal cord and dura. Custom stainless steel vertebral clamps were attached rigidly to hold both the C5 and C6 vertebrae in place. These clamps were designed and characterized by Choo et al. [55] and exhibited a stiffness of 83.6 ± 18.9 N/mm with a failure load at 64.7 ± 10.2 N when secured to the spinal column, significantly higher than the loading conditions expected in this study. At a nominal force of 2 N expected from contusion injuries the movement in the vertebrae, the region most likely responsible for any compliance in the system, was shown to be 0.03 ± 0.01 mm. Once the clamps were attached the animal was moved to the UBC multi-mechanism test system’s stereotaxic frame. The 2 mm diameter impactor head was lowered in 50 μm increments until a touch force of 0.03 N was reached, dimpling the surface of the dura to achieve a consistent datum for all injuries [71]. The impactor was then retracted 6 mm above the dura before accelerating downward to impact the cord using the pre-determined depth and speed. Immediately following injury the impactor head was retracted 6 mm above the dural surface with no dwell time specified.  33  Chapter 2 Methods Each individual impact test used WinTest software to record the force, displacement, and acceleration traces sampled at 4000 Hz. Inertial forces were compensated for each injury group using the same blank air impact method used in the characterization of the system described in section 2.2.2. Table 2.2 shows the compensation parameters used to account for the inertial forces. Table 2.2: UBC Multi-mechanism injury system compensated load parameters  Injury Parameters  a (-)  b (N·s2/m)  0.9 mm, 8 mm/s 0.9 mm, 80 mm/s 0.9 mm, 800 mm/s 1.5 mm, 8 mm/s 1.5 mm, 80 mm/s 1.5 mm, 800 mm/s  1 1 1 1 1 1  0.94 0.94 0.98 0.94 0.94 0.94  The skin and muscle at the injury site was sutured over the incision to allow healing of the wound and avoid infections. Animals were allowed to recover from injury in an incubator to help regulate their body temperature.  Post-Surgical For seven days following injury each animal was monitored four times daily to ensure proper health. Weight loss, appearance, and other stress indicators were recorded and scored on a scale of 20 for animals at each check (Appendix C); a total score of 20 required consultation with the veterinarian and potentially euthanizing the animal. Animals were expected to be tetraplegic following injury for the first three days and were given buprenorphine (0.03 mg/kg), to mitigate any pain; 10 mL of Ringer’s lactate three times daily; and separated into individual cages during this period. After a recovery period of seven days each animal was re-assessed in both the open field and cylinder rearing behavioural tests before tissue analysis. All animals were perfused with an intracardial needle with 250 mL of phosphate buffered saline (PBS) and followed with 500 mL of 4% paraformaldehyde in order to fix the tissues. The spinal cord was then carefully removed from the animal approximately ±5 mm rostral and caudal around injury epicenter and fixed overnight in 4% paraformaldehyde. Each cord was cryoprotected in increasing concentrations of sucrose solutions (12%, 18%, 34  Chapter 2 Methods and 24%) for 12 hours at each grade and frozen in isopentane before being cyrosectioned in 20 μm increments in the transverse plane (cross-sectional) in ten serial sets.  2.3.1  Kinetic analysis In order to assess the total energy applied to the spinal cord in each individual impact  and injury group the integral of the force and displacement data were approximated in MATLAB (vR2010b, The MathWorks, Natick, MA) using the trapezoidal method, trapz function. A typical force vs. displacement curve is shown in Figure 2.3. This method approximates the region under a given curve, ( ), by calculating the area sum of trapezoids between multiple points equally spaced along the curve: ∫  ( )  ∑(  )( (  )  (  ))  This method tends to overestimate the true value of the integral if the second derivative is positive and underestimate the true value if the second derivative is negative. However, given that each force vs. displacement curve was similarly shaped this estimation can be considered consistent between impacts and provide a reliable relative measurement. During retraction of the impactor a small negative force was sometimes recorded which may have been a result of the inertial compensation, but did not have a physical manifestation and was set to a value of 0. The energy recovered during retraction of the impactor was subtracted from the total energy applied to the cord during impact to calculate the energy absorbed by the cord.  Figure 2.3: A typical force vs. impactor displacement curve recorded from a C5/C6 midline contusion injury. The small force recorded at 0 mm impactor displacement is caused by contact with the dura before impact with the cord itself. (a) The red hatched area denotes the total energy applied to the spinal cord and (b) the blue hatched area displays the total energy absorbed by the cord.  35  Chapter 2 Methods  2.3.2  Behavioural analysis and correlations One of the main focuses of this study was to observe and identify the secondary  injury effects of impact velocity and depth on spinal cord injury. Evaluating the behaviour of each animal after a seven day recovery period post-injury allowed the assessment of secondary injury effects. Many methods of behavioural analysis have been standardized for this purpose and for this study the open field, rearing, and grooming tests were selected. These tests primarily focus on the assessment of forelimb motor control where the most noticeable deficit should be visible given the cervical level of injury in the spinal cord. All behavioural tests were conducted pre-injury, to obtain a baseline measurement, and seven days post-injury without pre-training. Scoring sheets for rearing, grooming, and open field analysis are shown in Appendix C.  2.3.2.1 Rearing and grooming The rearing test was used to determine the preferential use of one forelimb to another due to injury [75, 78, 79]. To evaluate forelimb use each animal was placed in a 30 cm tall, 20 cm diameter transparent cylinder and recorded while freely exploring its surroundings for 15 minutes without outside influence. Two mirrors were set up behind the cylinder to ensure that all angles were viewable by the camera when the animal was rearing. Each instance an animal stood up on its hindlimbs and used the left, the right, or both forelimbs to touch the cylinder wall the initial contact forelimb was recorded. If both forepaws were used to bear weight within 0.1 seconds of each other a score of “both” was recorded. Subsequent lateral wall touches were all recorded until the animal returned to resting position on all four limbs. Only the first 10 rears were recorded to standardize the total number of initial wall touches to observe any preferential use of either forelimb. If animals did not complete at least 10 rears the data were not included in the final dataset. Video analysis was conducted by an observer blinded to the injury groups. Each animal was also assessed for forelimb grooming function using a scoring system described by both Gensel et al. and Bertelli [76, 77]. This simple test allowed for easy assessment of damage to the brachial plexus, a network of nerve fibres running through C5-C8 which affect sensation and movement of the upper limb. Grooming was assessed by 36  Chapter 2 Methods recording the animals for 15 minutes using the same data captured for the rearing analysis. A typical grooming sequence was considered to include the licking of the forepaws and washing of the face followed by grooming of the face, licking of the body, and scratching with the hindlimbs [76, 80]. The scoring system used is shown in Figure 2.4. The maximum score for each forelimb, right and left, was recorded for each grooming activity and averaged. Where it was required, a frame by frame analysis of the recording was used to identify the maximum score received by the animal. Analysis was conducted by an observer blinded to the injury groups to remove observer bias.  Figure 2.4: Scoring categories for grooming analysis: 0) Animal unable to groom 1) Animal’s forepaw able to touch the mouth 2) Animal’s forepaw able to touch between the nose and eyes 3) Animal’s forepaw able to touch the eyes 4) Animal’s forepaw able to contact the front of the ears 5) Animal’s forepaw able to reach behind the ears. (Illustration adapted from Gensel [76])  A difference of proportions statistical comparison test with an alpha value of 0.05 was used to determine significance between paw preferences during the rearing behavioural testing within an injury group. Comparisons were measured at baseline and seven days post injury to identify any changes following injury. Grooming scores were evaluated using a non-parametric Kruskal-Wallis analysis of variance (ANOVA) and followed with a multiple comparisons of mean ranks for all groups. A non-parametric test was chosen as the scores were based on a categorical scale and does not require the same assumption of normality. Each score was given a rank from 1 to N based on the highest score recorded regardless of injury group. Ties were resolved by assigning an average value between the total number scores that had the same value. Ranks were then compared across groups to determine significance based on an alpha value of 0.05. A total of 16 individual group comparisons, between groups with similar impact depth; impact velocity; and impact energy, were done with a Mann-Whitney U test using a Bonferroni correction bringing the corrected alpha value to  .  37  Chapter 2 Methods  2.3.2.2 Open field behaviour It is generally accepted that the BBB rating scale does not properly evaluate the resulting behavioural deficits in cervical spinal cord injuries [81]. The more sensitive open field evaluation tool developed by Martinez et al. was modified and used for this study [74]. In addition to scoring the forelimbs and hindlimbs separately, each individual side (left and right) was scored since each animal received a midline injury. Animals were handled frequently for the one week acclimatization period to reduce stress during behavioural evaluation and to increase the reliability of the measurements. The open field enclosure was made of transparent plastic in a rectangular shape (120 cm by 150 cm) shown in Figure 2.5. Animals were gently placed in the open field one at a time for a period of 4 minutes and their articulation movements of the forelimbs, weight support, paw stepping, digit position, running coordination, and tail position were evaluated and scored described in detail below. If the animal did not move for a period of 20 seconds or longer it was picked up and placed at the centre of the open field. Two examiners, blinded to the injury each animal received, scored the behaviour for each animal by consulting with each other to complete each category. All categories were scored separately with a maximum total score of 72. Forelimb and hindlimb articular movement was evaluated by the examiners during spontaneous locomotion. For each joint, including the shoulder; elbow; and wrist as well as the hip; knee; and ankle, a score was given as shown in Table 2.3.  Score  Table 2.3: Articular movement scoring Observation  0 1 2  Absent articulation Less than 50% of maximum range of motion More than 50% of maximum range of motion  38  Chapter 2 Methods  Figure 2.5: Open field used for behavioural assessment  Weight support was observed as the ability of the animal to place a portion of its total weight on the limb being evaluated. Each score was recorded as present (score of 1) or absent (score of 0). Weight support was scored separately during stationary stance and during locomotion. A score of 0 in the stationary stance would automatically result in a score of 0 for locomotion. The ability to actively support weight was required to evaluate paw stepping. Paw stepping and movement was evaluated during locomotion by observing paw placement during both initial contact and lift off during the gait cycle. During locomotion either the dorsal or plantar surface of the paw can contact the ground with dorsal contact considered to be abnormal. If more than 50% of paw placement was contacted the dorsal surface a score of 0 was recorded.  When plantar surface contact was more frequent,  deviations from parallel placement in line with the joints determined score (Table 2.4).  Score  Table 2.4: Paw placement/orientation scoring Observation  0 1 2  More than 50% of contact with the dorsal surface Internal or external paw deviation Parallel with body axes  During the swing phase of the gait any jerky movement was given a score of 1 and considered abnormal. Movement was recorded as normal (score of 2) when the three joints moved in together in synchronized normal way.  39  Chapter 2 Methods If more than four consecutive steps were observed and the animal displayed the ability to support its weight then the forelimb-hindlimb coordination was also evaluated. A run was considered coordinated when the forelimb and hindlimb alternated in movement in placement. Each run was tallied as either coordinated or uncoordinated and the total score was described as being absent (score = 0), occasional (<50% of runs, score = 1), frequent (50-90% of the time, score = 2) or consistent (>90% of runs, score = 3). An elevated tail is considered a reliable indicator of stability during movement and was given a score of 1 out of 1. After injury, some animals held their tail closely to the ground (score = 0) suggesting behavioural deficit. Total scores (/72), forelimb scores (/36), and hindlimb scores (/36) were compared using a non-parametric Kruskal-Wallis ANOVA comparison of ranks for statistical significance. Any scores which displayed a p-value of less than 0.05 were further evaluated using multiple comparisons of mean ranks. A total of 16 individual group comparisons, between groups with similar impact depth; impact velocity; and impact energy, were done with a Mann-Whitney U test using a Bonferroni correction bringing the corrected alpha value to  2.3.3  .  Histological analysis  2.3.3.1 Luxol fast blue Following serial sectioning of the spinal cord tissue one set of slides was stained with luxol fast blue (LFB) for white and grey matter tissue sparing analysis. The luxol fast blue stain is commonly used to view myelin using light microscopy and stains the lipoproteins in the myelin sheath a blue colour. Slides were thawed and dipped into a 1% LFB solution overnight at 60oC. Lithium carbonate was used to differentiate the slides to remove the stain from the non-myelinated regions of the spinal cord and finally washed in increasing concentrations of ethanol (protocol in Appendix D). All images were captured on an AxioPlan2 microscope (Carl Zeiss, Thornwood, NY) using a 2.5x objective. Images were captured through the attached monochrome camera (Retiga Exi, QImaging, Burnaby, BC) using Northern Eclipse acquisition software (Empix Imaging, Mississauga, ON).  Visual basic script was created to control the motorized 40  Chapter 2 Methods scanning stage (Scan 100x100, Marzhauser, Wetzlar-Steindoft, Germany, MAC5000, controller, Ludl, Hawthorne, NY) and AxioPlan2 to incrementally move the stage and capture the entire cross-section. Each set of serial images were stitched together using Adobe Photoshop CS3 (v10.0) using the photomerge function to represent the entire section. Images were captured with 1 mm spacing between each section from 3 mm rostral to 3 mm caudal of the injury epicentre. Each cross-sectional image was sectioned into damaged white matter, damaged grey matter, intact white matter, and intact grey matter in Photoshop using the magic wand tool to select areas of similar pixel intensity (Figure 2.6). In some sections the tissue was so damaged that an estimation of the grey and white matter boundary was hand drawn based on non-damaged tissue in the sham spinal cords. After being sectioned into these four regions each image was converted to an 8-bit image and measured for total area using a constant threshold of 215/255 using ImageJ (v1.47c, National Institutes of Health, Bethesda, Maryland, USA). Total % grey matter spared and total % white matter spared were recorded as the measured areas normalized to the total area of grey or white matter in the section being analyzed [82]. Some sections displayed a cutting artifact which caused a fold in the tissue. These areas were accounted for by doubling the area of the folded portion of the section to capture both the visible and underlying tissue. Arcsine transformations were applied to the data prior to statistical testing to normalize proportional data (% tissue) which displayed a binomial distribution. Sectioning using the magic wand tool was compared to completely manual sectioning and found to have a root mean square (RMS) difference of 6.49%  Figure 2.6: LFB stained cross-section sectioned into intact white matter (red), damaged white matter (yellow), intact grey matter (blue), and damaged grey matter (green)  41  Chapter 2 Methods A two-way ANOVA test was carried out to compare the effects of impact velocity and impact depth on tissue sparing. An alpha level of 0.05 was used for the ANOVA and an unequal N Tukey post-hoc test followed if significance was found.  2.3.3.2 Demyelination A second set of slides was immunostained with chicken anti-myelin based protein (MBP), mouse anti-neurofilament-200 (NF-200), mouse anti-beta III tubulin, and rabbit anti-glial fibrilliary acidic protein (GFAP) to detect demyelination of axons surrounding the injury epicentre. Standard immunohistochemistry staining protocols (Appendix D) were followed to label the tissue with the previously mentioned markers.  All slides were thawed  from frozen at room temperature for 30 minutes before rehydrating for 10 minutes in 0.01 M PBS then delipidized in graded concentrations of ethanol. The tissue was blocked for 30 minutes in 10% normal donkey serum before a 2 hour incubation at room temperature in the primary antibody dilution (1:200 chicken anti-MBP, 1:500 mouse anti-NF-200, 1:500 mouse anti-beta III tubulin, 1:1000 rabbit anti-GFAP) in 0.01 M PBS containing 0.1% Triton X-100. Following incubation, slides were washed for 5 minutes in 0.01 M PBS three times and incubated in the secondary antibody dilution (1:200 Alexa Fluor-594 donkey anti-mouse, 1:200 Alexa Fluor-488 donkey anti-chicken, 1:200 Alexa Fluor-405 donkey anti-rabbit) overnight in a light opaque container. Sections were then washed for 5 minutes in 0.01 M PBS and mounted in Fluoromount-G (Southern Biotechnology, Birmingham, AL) to help prevent photobleaching and covered with glass coverslips (No 1.5, VWR, Radnor, PA). All images were acquired using a Zeiss microscope (Axio Observer Z1, Carl Zeiss, Thornwood, NY) with a 40x objective and equipped with a motorized scanning stage (MS2000, Applied Scientific Instrumentation, Eugene, Oregon). Exposure settings for image capture were automatically optimized through the Zen software (Blue v1.0.1.0, Carl Zeiss Microscopy GmbH, Jena, Germany) for each section as pixel intensity or area was not a factor in identifying the final outcome measure. Sections were imaged with a spacing of 1 mm from 3 mm rostral to 3 mm caudal.  Systematic random sampling through the  cross-sectional images through the ventral and lateral white funiculi regions was used to image the white matter (Figure 2.7, Figure 2.8).  42  Chapter 2 Methods A 75x75 µm grid was overlaid on top of the image and sampling frames were selected at every fourth 75x75 µm square until the entire region of interest was sampled. Images were not sampled at the edge of the tissue or on the grey and white matter border (Figure 2.8). Within each 75x75 µm square a smaller, 25x25 µm, region of interest at the centre was used to identify and count all demyelinated, swollen, and pathological axons (Figure 2.9). A demyelinated axon was classified if most of its myelin sheath was missing, while pathological axons were defined as axons which were no longer tightly wrapped by myelin. Swollen axons, which could indicate the broken or severed end of axons, were quantified if the total area of the axon occupied more than half of the 25x25 µm region of interest.  Figure 2.7: Stitched and immunostained cross-section at the injury epicentre at 20x magnification (Green: MBP, Red: NF-200/Beta-III tubulin). Both ventral (outlined in white; shown in Figure 2.8) and lateral (outlined in yellow) regions were imaged at 40x for analysis.  43  Chapter 2 Methods  Figure 2.8: Systematic random sampling of the ventral region with the 75x75 µm overlay at 40x. Squares shaded in grey mark tissue edge and grey matter border that were not sampled. (Green: MBP, Red: NF200/Beta-III tubulin)  Figure 2.9: 25x25 µm axon counting region of interest. Examples of healthy axons (yellow), pathological axons (purple), demyelinated axons (teal), and swollen axons (white) are shown.  44  Chapter 2 Methods Cells which shared a border on the top or far right lines of the region of interest were included in the total count while any cell which cross the far left or bottom line were not counted as part of the total. A two-way analysis of variance (ANOVA) test was carried out to compare the effects of impact velocity and impact depth on demyelination as an indicator of secondary injury. An alpha level of 0.05 was used for the ANOVA and an unequal N Tukey post-hoc test followed if significance was found.  2.4 FEA strain computation 2.4.1  Geometry and material properties The finite element model of the rat cervical spine developed by Russell [68] was used  to compute the maximum principal strain in specific regions of the cord during a contusion injury. Hexahedral three dimensional elements were used for the grey matter, white matter, and dura mater while tetrahedral elements were used for the vertebrae and intervertebral discs. A total of 235,734 three dimensional elements made up the C3-C6 segment of the spine and 18,972 smooth particles were created to mimic the CSF surrounding the cord. Russell et al. used a linear approximation of the velocity profile to control the impactor in their studies; however, for this study the impactor velocity during the survival study was used as input to the FEM to control the speed and depth of the impactor tip for each injury group. This provided an additional step towards replicating the experimental injuries exactly within the PAM-CRASH (Version 7.5, ESI Group, Paris, France) software to match not only the maximum impact velocity and depth, but also the acceleration. Loading and boundary conditions for contusion injuries were modelled to replicate the experimental set-up. All vertebrae and intervertebral discs were created as rigid bodies and the clamped vertebrae around the injury epicentre were fixed in PAM-CRASH. The rostral and caudal ends of the dura were constrained thoroughly preventing motion along the axis of the cord. The model developed by Russell et al. used in the computational portion of this study used an Ogden model of hyperelasticity defined by:  45  Chapter 2 Methods (  )  ∑  (  )  Where, is the strain energy density in terms of the principal stretches ( is the order of the model, nonlinearity, and  ),  are the material constants describing hyperelastic  defines the material’s shear modulus  To model the viscoelastic response of the material a Prony series exponential decay model was used: ( )  ⁄  ∑  Where, is the steady-state shear modulus, component, and  is the order of the model,  is the time decay constant.  By noting that ( )  ∑ ( )  Where  is the kth shear modulus  (  this equation can also be written as: ∑  (  ⁄  ))  . The material properties used by the computational model in this study and developed  by Russell et al. are shown below in Table 2.5. Table 2.5: Material properties of the spinal cord and dura used by Russell et. al [68]  Tissue Grey/White matter  Dura  Ogden constants μ = 40.04 kPa α = 4.7 ν = 0.45 μ = 207.41 kPa α = 16.2 ν = 0.45  Prony series constants g1 = 0.5282 τ1 = 8 ms g2 = 0.3018 τ2 = 150 ms g1 = 0.3182 τ1 = 9 ms g2 = 0.1238 τ2 = 81 ms g2 = 0.0997 τ2 = 564 ms g2 = 0.0997 τ2 = 4.69 s  46  Chapter 2 Methods  2.4.2  Contusion injury simulation During experimental injury, impactor displacement was recorded by the WinTest  software along with force data. One representative injury, with approximately average peak force and displacement traces from each of the six injury groups, was chosen for use as the input into the FEA.  Indenter motion was velocity controlled over time based on the  experimental injuries.  Defining indenter velocity at each instance in time also strictly  enforced the displacement of the impact given that the initial displacement, velocity, and acceleration were zero. The indenter was normal to and located at the surface of the dura. This initial condition did not account for the 0.03 N touch force used in the experimental injuries to determine the surface of the cord. To correct for this, the initial impact velocity recorded experimentally for each group was applied for the short duration of time between contact with the dura and the spinal cord. Experimental contusion injuries were conducted at the C5/C6 vertebral level; however, the FEM placed the indenter at the C4/C5 vertebral level (Figure 2.10). Russell et al. originally reduced the complexity of the model to include only four vertebrae from C3 to C6 surrounding an impact epicentre of C4/C5 to reduce computational cost [68]. Due to the scope of this project there was not enough time to go back to the original full scale model and reduce the simplified model down to include C4 to C7 followed by full validation. Material properties in the FEM were applied uniformly to each part, such as the grey matter, white matter, and dura. As such, shifting the injury epicentre one level caudally would not greatly alter the strain pattern at each distance relative to the impact site. Elements were removed from C4 and C5 to mimic the laminectomy performed prior to the experimental impact injuries.  47  Chapter 2 Methods  Figure 2.10: C3-C6 segment of the rat spine used to simulate experimental contusion injuries  Specific regions of interest quantified by Choo et al. and Russell et al. were modified and outlined in the model to reflect the dorsal, lateral, and ventral regions of the white and grey matter (Figure 2.11a) [6, 68]. Each cross-section was spaced at 1 mm increments rostral-caudally from the injury epicentre (Figure 2.11b). Maximum principal strain was recorded in each of these regions during contusion simulation. The average of the maximum strain values for each element in a region was compared with the demyelination, tissue sparing, and behavioural deficit recorded experimentally.  Figure 2.11: FEM cross-section regions of interest: lateral (yellow), ventro-lateral (blue), ventro-medial (red), dorsal (green), and grey matter (purple). The grey and pink elements represent the grey and white matter respectively which were not part of any defined region (A) Cord regions used for comparison with experimental data spaced at 1 mm intervals (B)  48  Chapter 2 Methods  2.4.3  FEA optimization Relative to the contusion injuries simulated by Russell et al., the duration of impact in  these experiments was more than 30x longer for the slowest injury which created a heavy increase in computational cost. For the simulations to be completed within a reasonable time frame the FEM was closely examined for areas where optimizations could be made without sacrificing accuracy. A standard model for comparison with changes to the model was created by applying the survival study impact conditions to Russell’s cervical spine model. Each change to the model was made incrementally to attempt to locate the computational bottleneck and then compared with the original results. First, the elements in the dura were increased to roughly 8 times the volume of the original elements creating a coarse mesh. The re-meshing of the dura also required that the volume of the smooth particles, simulating the CSF, be remodelled using Hypermesh (Altair Engineering, Troy, MI). In addition, the grey and white matter parts were smoothed and re-meshed with a coarser mesh size with an edge length of approximately 0.3 mm, while the original model contained elements with an edge length of roughly 0.15 mm. A model was also created by refining the elements near the impactor only by splitting the hexahedral elements. The interface between the two mesh sizes was bridged by splitting elements into three smaller hexahedral elements and merging the connecting surface nodes. Given the difficult nature of quantitatively comparing two models with different element and node locations the indenter displacement, maximum contact force, and qualitative strain distribution patterns were used for comparison. Some inaccuracy was expected to be introduced into the model by changing the element size and model configuration, but the purpose was to retain similar strain distribution patterns while reducing computational load. Once the calculation time was significantly reduced without altering the maximum strain values by more than 10% in the grey and white matter compared to the standard model, the changes were implemented into the final model before simulating injury conditions.  2.5 Correlations Following complete data collection from the behavioural analysis, histology, and FEA the outcome measures were examined for correlations between each other and with the 49  Chapter 2 Methods impact factors.  Non-parametric Spearman rank order tests were used to determine the  relationships between the open field, grooming, and rearing scores with impact velocity and depth, maximum principal strain, tissue sparing, and percentage of damaged axons. For the normally distributed data sets a linear correlation analysis was employed to identify correlations between the tissue sparing and percentage of damaged axons with the mechanical impact factors and the regional maximum principal strains.  Correlation  coefficient values were recorded for all comparisons and were reported along with complete scatterplot graphs to visualize the data pool.  50  Chapter 3 Results 3.1 FEA strain computation 3.1.1  FEA optimization During simulation, PAM-CRASH consistently reported elements located in the dura  requiring the smallest time step for calculation which determined the minimum time step for the model as a whole, increasing total computational time. This limitation was used as the first area to attempt to make modifications to decrease simulation time. Figure 3.1 shows the relative dura element sizes and qualitative strain distribution patterns in the cord used for comparison.  Both dura mesh sizes showed similar first principal strain patterns and  maximum values in the grey and white matter tissue. Plots of the impactor displacement and dura contact force were created in PAM-CRASH for the original model, medium dura mesh, and medium dura mesh with the new CSF volume (Figure 3.2). Impactor displacement did not vary between models and had a maximum displacement of -1.278 mm for the test simulation. The CSF volumes did not change significantly between the two models and only removed 175 smooth particles which had initial positions outside the coarse mesh dura. The maximum impact force for both the medium mesh dura with new and old CSF volumes was 0.750 N compared to the 0.699 N in the original model, approximately a 7% increase from the original contact force. The total computational time required to complete this simulation was reduced by approximately 63% requiring a total of 9 hours and 11 minutes compared to 24 hours and 34 minutes for the original simulation.  51  Chapter 3 Results  Figure 3.1: Element size (top) and maximum principal strain distributions (bottom) in the original model (A) and the medium dura mesh (B). Strain distribution patterns in the cord (with the dura hidden) show little difference between models.  Figure 3.2: Comparison of impactor tip and maximum contact force between the fine and medium mesh applied to the dura. The displacement of all three versions of the model was identical.  Following the reduction in computational time from the modification of the element size of the dura, the grey and white matters of the spinal cord were also increased in element size. The general pattern of strain distribution between the two models was similar; however, there was a notable reduction of the maximum principal strain directly underneath the impactor. An attempt to correct this by refining the element size near the impact area showed improvement near the impactor (Figure 3.3c). Impactor displacement did not vary between any of the models and had a minimum value of -0.981 mm (Figure 3.4). The maximum impact force in the original model, medium grey/white matter model, and refined medium model was 0.260 N, 0.240 N, and 0.223 N respectively. The increase in mesh size on the entire grey and white matter parts resulted in a reduction of computational time from 52  Chapter 3 Results the 9 hours and 11 minutes previously seen to 1 hour and 55 minutes, a 79% total reduction in computational time.  However, this reduction in computational time changed the  distribution of the primary measurement of principal strain directly around the impactor, seen qualitatively. Refining the mesh around the impactor saw a reduction of computational time to 4 hours and 36 minutes, a 50% reduction, but displayed continuity issues around the fine and coarse mesh interface. The final model to be used was the medium mesh dura with the new CSF volume calculation included. Both the element sizes of the grey and white matters were kept identical to the original model. The increased grey/white matter mesh size model created a divergence from the original strain distribution pattern and was not used in the final version of the model.  Figure 3.3: Element size (top) and maximum principal strain distributions (bottom) in (a) the original model, (b) medium grey/white matter mesh, and (c) refined grey/white matter mesh. Strain distribution patterns show distribution in the cord with the dura hidden  53  Chapter 3 Results  Figure 3.4: Comparison of impactor tip and maximum contact force between the fine, medium, and refined medium mesh applied to the grey and white matter parts  3.1.2  Contusion injury simulation The optimized finite element model was used to simulate all injury groups in the  survival study. Corresponding experimental displacement and force profiles were compared with the FEA results using similar simulation corridors of ±2 SD used by Russell et al. [68] in the validation data (Figure 3.5). Simulation displacements displayed a similar trace and peak value as the data recorded experimentally. The simulated contact forces also showed similar trace patterns as the experimental data, but had lower peak values. There was also a slight lag, between 1 ms to 1.5 ms, in the simulated force data compared to the experimental results recorded by the UBC multi-mechanism injury system. Comparisons of simulation and experimental data are summarized below in Table 3.1. Table 3.1: FEA simulation and experimental injury force and displacement comparison Impact Parameters  Peak FEA Displacement (mm)  Peak Experimental Displacement (mm)  Peak FEA Force (N)  Peak Experimental Force (N)  0.9 mm-8 mm/s  0.8898  0.896 (0.002)  0.2717  0.637 (0.106)  0.9 mm-80 mm/s  0.8915  0.880 (0.003)  0.3739  0.794 (0.105)  0.9 mm-800 mm/s  0.9800  0.888 (0.003)  0.5130  0.903 (0.199)  1.5 mm-8 mm/s  1.5508  1.471 (0.032)  0.9087  1.201 (0.112)  1.5 mm-80 mm/s  1.4579  1.459 (0.006)  1.1515  1.770 (0.232)  1.5 mm-800 mm/s  1.5744  1.461 (0.006)  1.6189  2.134 (0.266)  Experimental values shown are mean (±SD)  54  Chapter 3 Results  Figure 3.5: Experimental displacements and contact forces with the dura plotted alongside the simulated displacement and force traces. (a, b, c, g, h, i) Simulated impactor displacement for all injury groups matches closely with the mean experimental displacement. The 800 mm/s impact velocities show a slight deviation away from the experimental displacement, but follow the same pattern. (d, e, f, j, k l) The simulated contusion forces follow the same patterns as the recorded by the load cell experimentally. Peak forces are lower in the finite element model and lag behind the experimental data by approximately 1 to 1.5 ms in all injury groups.  55  Chapter 3 Results  3.1.3  Principal strain distribution The maximum principal strain in each element was the primary output of the FEA  simulation.  Figure 3.6 shows the parasagittal strain distribution pattern of a  1.5 mm-800 mm/s injury of the contusion injury along the length of the cord.  A  cross-section at the epicentre of injury shows the strain patterns in each region of interest described in section 2.4.2. Contusion injuries showed a uniform distribution of strain around the epicentre of injury spreading out as far as 3 mm both rostral and caudal.  The  cross-sectional distribution of strain also shows a mirrored strain pattern around the centreline.  Figure 3.6: FEA distribution of maximum principal strain direction during a 1.5 mm impact depth at 800 mm/s simulation. Distribution of strain is relatively uniform both rostral (left) and caudal (right) to the epicentre. High strain is localized in the dorsal region of the spinal cord that dissipates along the length of the cord away from the epicentre. Moderately high strain values can be seen in the ventral region caused by contact with the anterior side of the vertebrae. The strain pattern displays similarity to the displacement pattern shown in Figure 1.11.  The regional averages of maximum principal strains are plotted in Figure 3.7. The 1.5 mm impact depth groups showed zones of relatively high maximum principal strain (>0.25) in both the dorsal and ventral columns which was not seen in the 0.9 mm impact depth group. Strain in the 0.9 mm impact depth groups were <50% of the corresponding 1.5 mm impact at the same velocity between 1 mm rostral and caudal around the impact site. This difference in strain was not shown between groups with increasing impact velocity at the same impact depth. The zone of high strain was centralized around 1 mm rostral and caudal to the epicentre and dropped off outside of this region.  56  Chapter 3 Results  Figure 3.7: Average maximum principal strain distributions shown rostrocaudally at 1 mm intervals for each injury group (blue: 0.9 mm-8mm/s, red: 0.9 mm-80mm/s, green: 0.9mm-800mm/s, purple: 1.5 mm-8mm/, teal: 1.5 mm-80mm/s, orange: 1.5 mm-800mm/s). Principal strain in the (a) dorsal column, (b) grey matter, (c-d) lateral white funiculus, and (e-f) ventral white funiculus are represented in the FEM. Negative distance denotes the rostral direction while positive values denote the caudal direction. High strains at the epicentre significantly decrease beyond 1 mm rostrocaudal to the epicentre at both injury depths. Within this zone of high strain the average maximum principal strain shows more sensitivity to impact depth than impact velocity.  3.2 3.2.1  Survival study Impact mechanics The mechanical data and loads were recorded for each contusion impact and are  shown in Table 3.2 and Figure 3.8. Some animals experienced complications during the 57  Chapter 3 Results surgical laminectomy or in the 7 day survival period and were euthanized, resulting in fewer animals in those groups. Little relative variation was seen in the impact depth with a standard deviation of depth less than 0.3% of the average. Impact velocity showed slightly more variability in the 8 mm/s impacts with a standard deviation of approximately 12% of the average. Both the 80 mm/s and 800 mm/s impacts had a standard deviation of less than 2% of the average. Given the increase in magnitude between impact velocity groups the absolute variation of the slow impact group was deemed acceptable. Two animals, one from the 1.5 mm-8 mm/s group and one from the 1.5 mm-800 mm/s group, saw abnormally high forces during impact that may have been due to contact with the lateral wall of the vertebral canal, but recorded normal impact depth and velocity. These data points were determined to be outliers based on the interquartile range of the force data and excluded from the dataset. The energy applied to the spinal cord itself is a function of both the spinal cord displacement and the force applied and therefore could not be independently controlled from the impact factors. No significant difference in energy applied to the spinal cord (p > 0.05) was observed between groups I & II, II & III, and III & IV (Figure 3.9). With the energy applied to the spinal cord in the 0.9 mm-800 mm/s group and 1.5 mm-8 mm/s group falling within a similar range, both behavioural deficit and demyelination could be compared with different mechanical impact factors at roughly similar levels of impact energy. Table 3.2: Experimental contusion injury parameters Group  Impact Parameters  n  Impact Depth (mm)  Impact Velocity (mm/s)  Force (N)  Energy Applied (mJ)  I  0.9 mm-8 mm/s  8  0.896 (0.002)  8.797 (1.003)  0.637 (0.106)  0.250 (0.038)  II  0.9 mm-80 mm/s  7  0.880 (0.003)  83.800 (1.112)  0.794 (0.105)  0.356 (0.050)  III  0.9 mm-800 mm/s  8  0.888 (0.003)  838.272 (2.921)  0.903 (0.199)  0.506 (0.088)  IV  1.5 mm-8 mm/s  6  1.471 (0.032)  8.263 (0.928)  1.201 (0.112)  0.711 (0.074)  V  1.5 mm-80 mm/s  7  1.459 (0.006)  83.868 (1.396)  1.770 (0.232)  1.187 (0.180)  VI  1.5 mm-800 mm/s  7  1.461 (0.006)  840.555 (1.696)  2.134 (0.266)  1.775 (0.209)  VII  Surgical Sham  5  N/A  N/A  N/A  N/A  Values shown are mean (±SD)  58  Chapter 3 Results  Figure 3.8: Representative force, displacement, and velocity curves for each injury group: (a,c,e) 0.9 mm impact depth injuries, (b,d,f) 1.5 mm impact depth injuries. Green lines indicate the datum point (on the surface of the cord), red lines indicate the point of maximum impact depth before retraction of the impactor. A slight increase in force can be seen before reaching 0 mm displacement caused by impact with the dura.  59  Chapter 3 Results Total Energy Applied to the Spinal Cord Mean  Mean±SD  2.4 2.2 2.0 1.8 1.6  Energy (mJ)  1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 8 mm/s  800 mm/s 80 mm/s  Depth: 0.9 mm  8 mm/s  800 mm/s 80 mm/s  Depth: 1.5 mm  Figure 3.9: Average energy applied to the spinal cord for each injury group. All groups showed significant difference in applied energy (p > 0.05) except the 0.9 mm-80 mm/s and 0.9 mm-8 mm/s (p = 0.831) or the 0.9 mm-800 mm/s (p = 0.533) groups. The 0.9 mm-800 mm/s injury group did not show significantly different energy applied to the cord compared to the 1.5 mm-8 mm/s group (p = 0.097).  3.2.2  Behavioural analysis  3.2.2.1 Rearing and grooming During baseline testing animals displayed no preference (p > 0.05) of either left or right paw to support weight during the initial vertical exploration of the cylinder, but generally used either one paw for initial contact over the use of both. Following injury, no change was observed in paw preference during rearing in any of the injury groups or the control group (Figure 3.10e). However, none of the animals in the 1.5 mm-800 mm/s injury group were able or willing to perform the minimum 10 rears post-injury and were not included in the comparison. As the only animals which displayed the inability or reluctance to perform 10 rears, this suggests a significant level of injury not seen by any of the other groups as a result of the combined increased impact depth and impact velocity. These results confirm the nature of the contusion injuries given to each animal as a midline injury and was not predicted to produce a more prominent deficit on either the left or right side.  60  Chapter 3 Results  Figure 3.10: Rearing test showing initial contact using only the left paw (a), right paw (b), or both (c) for weight support. (d) Shows the proportion of paw preference during initial rearing contact between all injury groups at baseline and seven days post-injury. No preference for right or left paw was seen at baseline and did not change post-injury. Animals which were did not complete the minimum 10 rears were excluded from the dataset. Data shown are mean + SD (negative SD is not shown for clarity).  The grooming score results, based on the same video data, showed no difference between groups at baseline on average grooming score (Figure 3.11).  Multiple group  comparisons showed a significant decrease in grooming ability with the 1.5 mm-800 mm/s group to all other groups (p < 0.003125) at seven days post-injury. No effects of injury impact depth or impact velocity on grooming ability between other injured or control groups were observed.  61  Chapter 3 Results  Figure 3.11: Grooming test at baseline and seven days post-injury. After injury only the 1.5 mm-800 mm/s injury group showed a significant decrease in the grooming score (*p < 0.003125). Data shown are mean ± SD.  3.2.2.2 Open field behaviour Scores were compiled as a hindlimb score (/36), forelimb score (/36), and a combined total score (/72). All animals received perfect baseline scores showing no difference. At seven days post-injury animals that received a 1.5 mm depth injury saw a significant reduction in hindlimb, forelimb, and total score (p < 0.007), which was not shown by the 0.9 mm depth injury regardless of impact velocity. Post-injury group comparison showed a significant reduction of total score (p < 0.003125) when impact depth was increased from 0.9 mm to 1.5 mm above an 80 mm/s impact. Increased impact velocity to 800 mm/s also displayed a significant (p < 0.003125) reduction of total score in the open field at the 1.5 mm impact depth (Figure 3.12a).  Injury groups that had a similar range of impact energy  displayed no change in total score; however, the 0.9 mm-800 mm/s and 1.5 mm-8 mm/s injury groups displayed a trend towards increased deficit.  This trend led to further  investigation of forelimb function and displayed noticeable reduction of forelimb score (p = 0.003743) between these groups (Figure 3.12b).  62  Chapter 3 Results  Figure 3.12: Open field behavioural results. (a) Following injury, a significant reduction in total score was observed when impact depth was increased from 0.9 mm to 1.5 mm above 80 mm/s impact velocity (*p < 0.0030125). (b) Analysis of forelimb scores showed a large reduction in score between the two groups with similar applied impact energy and different impact depths (#p = 0.003743). Data shown are mean ± SD.  3.2.2.3 Correlations Correlation analysis of the behavioural open field and grooming scores showed significant (p < 0.05) relationship with both the velocity and depth of impact. Impact depth displayed considerably better correlation with post-injury forelimb ( (  ), total (  ), and grooming scores (  ), hindlimb  ) than impact velocity,  with increasing depth resulting in lower overall scores. No correlations were seen between post-injury forelimb score and impact velocity. All other correlations with impact velocity were in the low to moderate range (0.3 to 0.49). Table 3.3 summarizes all impact factor correlation data. Table 3.3: Correlation coefficients for mechanical impact factors with behavioural assessment scores  7 day post injury Forelimb score Hindlimb score Total score Grooming score  Impact Velocity Spearman Rank p-value Correlation Coefficient, rs 0.07 N/A 0.04 -0.33 0.02 -0.38 0.005 -0.44  p-value < 0.001 < 0.001 < 0.001 0.001  Impact Depth Spearman Rank Correlation Coefficient, rs -0.70 -0.66 -0.66 -0.49  63  Chapter 3 Results  Figure 3.13: Correlations of impact velocity and depth with behavioural scores  Comparison of the relationship between the open field behavioural analysis and FEM maximum principal strain showed strong non-parametric correlations. For the pooled white matter strain, significant (p < 0.05) correlations were shown with post-injury total score (  ), forelimb score (  ), and hindlimb score (  ), with lower  scores with increasing strain (Figure 3.14, Table 3.4). They grey matter strain displayed significant (p < 0.05) correlations with both post-injury total score ( score (  ) and forelimb  ). No correlation existed between the grey matter strain and post-injury  hindlimb score which could be explained given that cervical spinal cord injuries mainly result in forelimb deficit and the functional responsibility of the grey matter tissue to control of signal at a specific vertebral level only. Additionally, the grooming scores showed no correlation with either white or grey matter strain. Table 3.4: Correlation coefficients for maximum principal strain with open field behavioural scores  7 day post injury Total score Forelimb score Hindlimb score  Pooled White Matter Strain Spearman Rank p-value Correlation Coefficient, rs 0.005 -0.94 0.049 -0.81 0.005 -0.94  Grey Matter Strain Spearman Rank p-value Correlation Coefficient, rs 0.005 -0.94 0.036 -0.84 0.072 N/A  64  Chapter 3 Results  Figure 3.14: Correlations of maximum principal strain grey matter and pooled white matter maximum principal strain in the cord simulated by the FEA.  3.2.3  Histological analysis  3.2.3.1 Luxol fast blue The percentage of tissue sparing at the epicentre and each 1 mm increment rostral and caudal was analyzed using a standard ratio of the total area of intact grey or white matter versus total tissue. Figure 3.15a shows representative sections of damaged tissue stained with luxol fast blue. A multifactorial ANOVA revealed significant main effects of impact depth (p << 0.01), impact velocity (p = 0.012), and depth*velocity (p = 0.0055) at the epicentre of injury on white matter sparing (Figure 3.15b,c). Grey matter sparing only displayed significant effects of impact depth (p << 0.01). The 1.5 mm impact depth injury groups showed less white matter tissue sparing than the control animals up to ±2 mm from the injury epicentre; the 0.9 mm impact depth displayed less tissue sparing up to ±1 mm from the epicentre.  65  Chapter 3 Results  Figure 3.15: (a) Luxol fast blue stained spinal cord tissue at 1 mm increments. The area percentages of (b) white matter spared and (c) grey matter spared were calculated as the ratio of intact tissue versus total tissue. *p < 0.05 compared to the 0.9 mm impact depth injury at the same impact velocity; #p < 0.05 compared to the 8 mm/s impact velocity at the same impact depth; +p < 0.05 compared to the 800 mm/s impact velocity at the same impact depth; $p < 0.05 between the 1.5 mm-8 mm/s and 0.9 mm-800 mm/s groups (similar impact energy range). Data shown are group means ± SD.  A post-hoc unequal N Tukey test of the grey matter tissue sparing showed no effect of impact velocity between groups. Increased impact velocity by two orders of magnitude, from 8 mm/s to 800 mm/s, displayed a significant reduction in the amount of white matter spared for the 1.5 mm impact depth only (p = 0.024). Impact velocity quickly showed a diminishing significant effect on the amount of tissue spared further from injury epicentre than impact depth.  3.2.3.2 Demyelination The percentage of damaged and demyelinated axons at each 1 mm interval from the injury epicentre, in both the ventral and lateral regions of the spinal cord, were analyzed using the ratio of damaged axons to total axons in the tissue described in section 2.3.3.2. For the ventral region of the spinal cord a multifactorial ANOVA revealed the significant main effects of impact depth (p << 0.01), impact velocity (p << 0.01), and depth*velocity 66  Chapter 3 Results (p = 0.0077) at the epicentre. These factors remained significant at each interval from the epicentre to 3 mm rostral and caudal affecting the percentage of damaged and demyelinated axons. Results plotted in Figure 3.16 show that at the epicentre of injury, increased impact depth, regardless of impact velocity, resulted in a significant increase in the percentage of damaged axons (p < 0.01). Increased impact velocity was only a significant effect at the 1.5 mm impact depth (p < 0.01) or when increased from 8 mm/s to 800 mm/s at the 0.9 mm impact depth (p < 0.01). The 1.5 mm-800 mm/s group also showed an increase in percentage of damaged axons over the 0.9 mm-800 mm/s group (p < 0.01) up to ± 2 mm from epicentre although the applied energy was within a similar range. Impact velocity displayed a diminishing effect on the amount of damaged axons further from injury epicentre than impact depth in the ventral region.  Figure 3.16: The percentage of damaged axons in the ventral region of the spinal cord tissue. *p < 0.01 compared to the 0.9 mm impact depth injury at the same impact velocity; #p < 0.05 compared to the 8 mm/s impact velocity at the same impact depth; +p < 0.05 compared to the 800 mm/s impact velocity at the same impact depth; $p < 0.05 between the 1.5 mm-8 mm/s and 0.9 mm-800 mm/s groups (similar impact energy range). Data shown are group means ± SD.  In the lateral region of the spinal cord a multifactorial ANOVA revealed the significant main effects of impact depth (p << 0.01), impact velocity (p << 0.01), and depth*velocity (p << 0.01) at the injury epicentre. The interactive effect of depth*velocity disappeared at ±1 mm and impact velocity had no significant effect on the percentage of damaged axons beyond ±2 mm. At the injury epicentre the percentage of damaged axons in 67  Chapter 3 Results the lateral region of the spinal cord was significantly increased by both increasing impact velocity (p < 0.05) and impact depth (p < 0.001) (Figure 3.17). Comparison between the 1.5 mm-8 mm/s and 0.9 mm-800 mm/s groups showed no statistical increase of damaged axons with a similar range of impact energy applied to the cord. At ±1 mm from the epicentre the effect of increased impact depth on the percentage of demyelinated axons disappeared. Similarly, the effect of increased impact velocity was reduced and only saw significance between the 1.5 mm group from 8 mm/s to 80 mm/s at -1 mm (p = 0.047). In contrast to the results seen in the ventral region, neither increased impact depth nor impact velocity significantly increased the percentage of damaged axons beyond ± 1 mm in the lateral white matter.  Figure 3.17: The percentage of damaged axons in the lateral region of the spinal cord tissue. *p < 0.01 compared to the 0.9 mm impact depth injury at the same impact velocity; #p < 0.05 compared to the 8 mm/s impact velocity at the same impact depth; +p < 0.05 compared to the 800 mm/s impact velocity at the same impact depth. Data shown are group means ± SD.  3.2.3.3 Correlations Linear correlation analysis of the MBP and LFB analysis showed significant (p < 0.05) correlations with depth of impact. Greater impact depth resulted in decreased healthy tissue sparing in both the white and grey matter regions (  ,  respectively). Increased impact depth also corresponded well with more axonal damage in both the ventral and lateral regions (  , respectively).  Impact 68  Chapter 3 Results velocity also showed significant correlation with axonal damage, but not with tissue sparing. Increased impact velocity resulted in increasing axonal damage in the ventral and lateral regions of the white matter (  , respectively). Table 3.5 provides a  summary of all correlation coefficients between the mechanical impact factors and histological analysis. Table 3.5: Correlation coefficients for mechanical impact factors with LFB/MBP analysis  7 day post injury Ventral % damaged axons Lateral % damaged axons % White matter spared % Grey matter spared  Impact Velocity Correlation p-value coefficient, R2  p-value  Impact Depth Correlation coefficient, R2  < 0.01  0.17  < 0.01  0.83  < 0.01  0.47  < 0.01  0.46  0.12  N/A  < 0.01  0.81  0.80  N/A  < 0.01  0.79  Impact Factors vs. MBP/LFB Histology Impact Velocity  Impact Depth  Ventral % Damaged  Lateral % Damaged  % White matter spared  % Grey matter spared  Figure 3.18: Correlations of impact velocity and depth with MBP/LFB histological results  All white matter regions in the model (ventro-medial, ventro-lateral, left lateral, right lateral, and dorsal) were collected and averaged which showed significant (p << 0.001) correlation with white matter sparing (  , Figure 3.19a), pooled from all  69  Chapter 3 Results experimental injury groups. Less tissue spared with increased strain in the ventral grey matter also showed significant (p << 0.001) correlation (  , Figure 3.19b).  In addition to the correlations with white and grey matter sparing, a measure of the whole grey/white matter damage, were the strong correlations with the region specific percentage of axons damaged or demyelinated. Increased strain was strongly tied to an increased percentage of axons becoming damaged in the ventral ( and lateral (  , Figure 3.19c)  , Figure 3.19d) regions of the spinal cord. Table 3.6 provides a  summary for all the FEA correlation data. Table 3.6: Correlation coefficients for maximum principal strain with LFB/MBP histology  p-value  Correlation coefficient, R2  % White matter spared  9.87E-22  0.90  % Grey matter spared  7.29E-14  0.76  Ventral % damaged axons  5.78E-19  0.86  Lateral % damaged axons  3.66E-13  0.74  Figure 3.19: Correlations of maximum principal strain with (a) % spared white and (b) % spared grey matter determined by a LFB stain. Maximum principal strain correlations with % of damaged or demyelinated axons are shown in the (c) ventral region and (d) lateral region of the spinal cord.  70  Chapter 3 Results There were also significant (p << 0.001), moderate to strong correlations between the percentages of damaged axons and white/grey matter sparing with post-injury open field and grooming scores (Figure 3.20, Table 3.7). An increase in the percentage of damaged axons in either the lateral or ventral region resulted in a decrease in the score received during open field evaluation and grooming. Similarly a decrease in the amount of healthy grey or white matter tissue spared resulted in a lower score during the open field and grooming evaluation at seven days post-injury. Table 3.7: Spearman rank correlation coefficients for open field scores with LFB and MBP histology  Spearman Rank Correlation Coefficient, rs  Post-Injury Forelimb Score Post-Injury Hindlimb Score Post-Injury Total Score Post-Injury Grooming Score Average All p << 0.001  Ventral % Damaged  Lateral % Damaged  % WM Spared  % GM Spared  -0.800917  -0.728445  0.830523  0.713485  -0.770400  -0.706333  0.798563  0.698189  -0.797984  -0.738210  0.809568  0.700991  -0.671192  -0.614715  0.723938  0.589917  Figure 3.20: Scatterplot Spearman rank correlations of open field behavioural and grooming scores with MBP/LFB histology  71  Chapter 4 Discussion 4.1 Effects of impact velocity and depth on secondary injury This experimental design was created to examine both the individual and interactive effects of a large range of contusion impact velocities and depths in SCI at seven days post-injury. Much of the existing research has not decoupled the control of these mechanical impact factors independently. This has resulted in a concentrated set of results along a linear relationship, between impact velocity and depth, which is difficult to compare across studies. The effects of these individual parameters have been demonstrated by this research and by others to be important to understanding the biomechanics of contusion SCI [33, 57, 59, 60]. This research is the first study to control both impact velocity and depth independently within the same experimental matrix, while also intending to apply similar impact energy with different impact factors. An understanding of the response of the spinal cord to these types of injuries might help eventually identify areas of concern or improve targeted treatments clinically. In the behavioural and histological assessment of the six injury groups, both the increased impact velocity and impact depth showed significant increase in damage to the spinal cord. Specifically, the interactive effects between the two impact factors play an important role in tissue sparing and axon demyelination within ±1 mm of the injury epicentre (Figure 3.15b,c; Figure 3.16; Figure 3.17). Kim et al. [65] found that, in mice, there was no effect of impact velocity given a fixed impact depth in contusion injury, which was only observed in this study at the low impact depth of 0.9 mm. Interactive effects between these two parameters are shown to be important here; the depth of injury in the previously mentioned study may not have been great enough to elicit a significant difference between increased impact velocities. The idea of this interaction between mechanical measurements of soft tissue injury was first studied by Viano and Lau [64] and extended to the spinal cord by Kearney in 1988 [39]. It was shown that at low impact velocity the impact depth acts as  72  Chapter 4 Discussion the primary mechanical effect on the injury severity of contusion SCI, but with increasing velocity the viscous response of the cord becomes more important. Here, similar results are observed with impact depth appearing to be a constant predictor of injury severity, while velocity becomes more important only once a threshold depth has been achieved. However, fundamental mechanics equations show that adding more energy that can do work on an object should result in higher displacement, or strain in the case of spinal cord injury, as a result. This increased strain has been linked to increasing functional deficit in spinal cord injury [67, 68, 83, 84]. The importance of the viscous response here can be identified by the results of the energy measurements of each injury group, where increased impact depth results in a significant increase in the energy applied to the cord, while the increased velocity creates a similar increase in applied energy only at the 1.5 mm impact depth (Figure 3.9). There appears to be a point in the combination between impact velocity and depth where relative increases in either will significantly increase the energy applied to the spinal cord and therefore the severity of injury. Below this threshold, the viscous response of the cord appears to be less important and severity is simply defined by the depth of the impact. Based on the results of this study, this threshold lies somewhere between the 0.9 mm-800 mm/s and 1.5 mm-8 mm/s injury combination in these animals. Unfortunately it is not possible to determine which factor, or combination of both, defines this point based on the experimental design. This study compared the results of secondary injury in contusion SCI through behavioural analysis, demyelination of axons in the spinal cord, tissue sparing, and principal strain at a time point post-injury where the earliest stages of secondary injury effects could be observed [25, 26]. Results from the rearing behavioural analysis showed no measureable amount of deficit of one forelimb over the other in any injury group including the control. This test was originally developed to detect differences in unilateral contusion injuries to the spinal cord [76]; however, its use in this study was to ensure that no preferential damage to either the left or right side of the cord was seen by an injury group. Animals were required to attempt at least 10 rears to be included in the dataset over a period of 15 minutes. While there were no measureable side-to-side differences between groups, no animals in the 1.5 mm-800 mm/s injury group attempted 10 rears. Qualitatively, compared to baseline and other injury groups, the animals in this group displayed noticeably reduced willingness and 73  Chapter 4 Discussion interest to explore their surroundings, which may be an indication of the severity of their injury not measured by the rearing analysis. The grooming test allowed for quick evaluation of damage to the brachial plexus, with resulting impairment of forelimb sensation, and movement for each injury group. Results from the grooming assessment of injured animals showed similar findings from the rearing test.  Based on multiple comparisons between post-injury scores only the  1.5 mm-800 mm/s group showed a significant decrease in the scores associated with grooming. With respect to injury parameters, an increase in impact depth from 0.9 mm to 1.5 mm resulted in a clear trend towards forelimb deficit. Increasing velocity was also an important factor in determining this deficit, but was only apparent at the 1.5 mm injury depth. Compared to baseline and other injury groups, the animals in this group also displayed reduced willingness or ability to groom themselves, which may be explained by similar inhibition to sensory inputs and motivational dispositions recorded by Berntson et al. in decerebrated rats [80]. Open field behaviour analysis showed the first statistically significant results of the interactive effects between impact velocity and depth and hinted at the energy effects of impact. Only groups that received injuries at 1.5 mm showed significant reduction in score from their baseline measurements while changes in velocity had no effect. However, injury groups that experienced an impact velocity of 80 mm/s or greater saw significantly reduced motor function, coordination, and balance, among other measurement categories between groups, when impact depth was also increased from 0.9 mm to 1.5 mm. Also, at the 1.5 mm impact depth, increasing impact velocity to 800 mm/s greatly reduced total score and forelimb score which suggests a potential threshold level of velocity at a given depth causing severe behavioural deficit [39, 64].  The 0.9 mm-800 mm/s and 1.5 mm-8 mm/s injury  groups displayed trends in total score and specifically in forelimb function, suggesting that, although a similar energy range was applied to the cord, there were differences in the behavioural deficit measurements. A likely explanation for this observation is the greater strain, seen in the 1.5 mm impact, experienced by axons in the ventral vestibulospinal and reticulospinal tracts causing demyelination and affecting movement related activity.  In  contrast, the injury groups with a similar energy range at the 0.9 mm depth did not show any  74  Chapter 4 Discussion difference in open field scoring displaying the interactive effects between the two mechanical injury parameters. In the current study, spinal cord tissue analysis supplemented the behavioural observations to provide additional explanation for deficits or lack thereof.  Holistic white  and grey matter damage observed with the LFB stain showed distinctly different contributions to tissue damage from increasing impact velocity or depth. Although impact depth, velocity, and the interaction between the two showed significant effects on the amount of white matter spared at the epicentre, the interactive and velocity effects very quickly dropped off further away from the epicentre. Given the decreasing displacement, or strain, in the sections further from the epicentre it should be expected that impact velocity becomes less important as we are moving further form the interaction effect threshold. In the grey matter, it was immediately evident that impact depth was the only factor which suggested influence on the amount of tissue spared in that region. Again, similar impact energy groups between the 0.9 mm-800 mm/s and 1.5 mm-8 mm/s impact groups displayed a largely different amount of tissue spared in both the grey and white matter regions, but not in the 0.9 mm injury groups with similar applied energy, as suggested by the behavioural results. In the white matter, the demyelination of axons in the ventral and lateral regions causes the impairment of signal transmission in the CNS [11, 12, 13, 85]. Demyelination of the axons in the spinal cord has been considered a major contributor to functional loss following SCI due to the impaired signal conduction and action potential propagation [86]. This study suggests that increased impact velocity and depth result in significant and widespread demyelination following contusion injuries. Again, as shown by the behavioural and LFB analysis, the increase of impact depth consistently increased the amount of demyelination of axons in the ventral region of the spinal cord. No noticeable similarities were observed between the 0.9 mm-800 mm/s and 1.5 mm-8mm/s groups with a similar range of applied energy. However, change in impact velocity did remain important in the 1.5 mm impact depth groups and when increased from 8 mm/s to 800 mm/s in the 0.9 impact depth groups. Although no specifically defined threshold was shown, these results reinforce the idea that some threshold impact velocities or depths exist, which result in graded histological outcomes [34, 39, 64].  In the lateral white matter region, both the effect of impact velocity 75  Chapter 4 Discussion and depth were diminished, although still significant, at the epicentre of injury. Almost immediately at ±1 mm rostro-caudal from the impact site neither depth or velocity effects on axon damage were seen; however, a trend is visible showing that increasing depth results in a larger increase of demyelination up to and including ±1 mm from the epicentre. Interestingly, even at the injury site, the groups with a similar energy range showed similar demyelination patterns, suggesting that an area of the cord not in direct contact with the impactor may be more affected by the incoming energy to the cord rather than just the mechanical injury parameters; this could be explained by the lower overall strains seen in the lateral region significantly reducing the energy applied to the lateral region of the spinal cord. Qualitatively, the LFB images show that even a relatively high percentage of spared tissue in the ventral and lateral white matter regions still result in behavioural deficit. This suggests that the physiological damage threshold of the spinal cord is much lower than the structural threshold similarly observed by Bain et al. [87]. Demyelination of the axons in these regions shown by the MBP analysis indicates that myelin repair and regenerative strategies, targeted to these areas and within close proximity to the injury epicentre, could be important in restoring some motor function given that the structure the spinal cord remains intact.  4.2 Contusion simulation In addition to observing the effects of mechanical factors on injury severity in SCI, the examination of the mechanical response of the spinal cord, such as impact strain, was displayed as a predictor for biological damage and behavioural deficit. Overall, the response of the FEM has demonstrated its biofidelity simulating contusion injuries over a range of impact velocities and depths.  Contact forces, although slightly lower than shown  experimentally, were within realistic values with nearly identical displacement and velocity profiles. The strain patterns and range of maximum principal strain values calculated by this model displayed similarity to both the cervical cord results seen by Russell et al. [68] and thoracic cord model developed by Maikos et al [67]. Results from this work show very strong correlations between not only maximum principal strain and spinal cord tissue damage, supporting previous research [67, 83, 87], but also behavioural deficit strengthening the validity and potential utility of this finite element 76  Chapter 4 Discussion model. These correlations did not specifically identify any tolerance threshold of tissue damage or behavioural deficit; however, the strain was only greater than the 0.14 tolerance threshold determined by Bain et al. [83] for the 1.5 mm impact depth groups within ±1 mm of the injury epicentre where the correlations to axonal damage were the strongest.  4.3 Limitations Limitations of the current work, in both the experimental and computational models of spinal injury, are identified here and may be possible areas for improvement in future studies.  4.3.1  Experimental model  Contusion impact model A common limitation of displacement controlled contusion impact models is the requirement to find a common initial position to define the datum. This study, along with others [6, 37, 71], used a small contact force of 0.03 N as the starting point for all impacts. The measured force, at the resolution provided by the load cell, did fluctuate after reaching the initial 0.03 N reading due to the viscoelasticity of the cord, response of the dura, and the presence of the CSF. While variations in the forces were small, altering the datum could result in a larger or smaller absolute cord compression depth than in other contusions of the same injury group. Without real-time and simple imaging techniques to allow a view of the spinal cord and impactor position, force control was considered the best solution. During injury it was also assumed that the clamps, and thus the vertebral bodies, were rigid. If any relative motion existed between the vertebrae and the spinal cord which was not measured then the impact depth of the cord cannot be known. The original design of the clamps was created to reduce this potential motion, but short of measuring the actual position of the impactor tip, spinal cord, and vertebrae in real time it is not possible to know the exact motion of all bodies. Clamps were secured as rigidly as possible to the stereotaxic frame and attached firmly to the C5/C6 vertebrae to reduce the potential for relative motion. Contusion injuries in this study were midline C5/C6 cervical spine impacts and while both the behavioural and histological results agree with this, mediolateral and rostrocaudal alignment were done by hand.  Markings on the clamp, in addition to consultation with 77  Chapter 4 Discussion Dr. J.L., identified the centreline rostrocaudally and mediolaterally on the spinal cord while adjusting impactor position. A mechanical system, or software imaging overlay could be employed in the future if more precise impact locations are required. Finally, the contusion injuries applied in this work impacted the dorsal side of the spinal cord. Clinically, these injuries occur on the ventral, or anterior, side of the cord as a result of a bone fragment fracture imprinting on the cord. Ventral contusion injuries have not been modelled to date given the difficulty of accessing the spinal cord through the anterior side of the animal without damaging other internal organs. The strain patterns and resulting tissue damage presented in this thesis are a representation of dorsal impacts only.  Behavioural analysis The behavioural analysis of each animal during was obtained at both a baseline time point and at seven days post-injury. However, the baseline measurements for all tests were the animals’ first exposure to each testing environment. Due to the limitations in project time, project scope, and the simplicity of the behavioural tests chosen, pre-training for these tests was excluded. The use of pre-training for these tests can reduce animal stress which may affect the histological outcomes or recovery time or skew the data. For more complex tasks that evaluate behavioural deficit, such as the walkway, grip strength, ladder, and staircase tests this may be considered less acceptable [74, 75, 76, 88]. During the seven day recovery period post-injury animals were monitored four times daily to ensure that animals were comfortable and had food and water freely available to them. Several times throughout the week some animals had either removed their sutures or had them come loose on their own. The wounds had to be re-sutured to prevent infection and allow proper healing. In each instance an animal would have to have their wound re-sutured which required another round of anaesthesia prior to surgical intervention. While every attempt was made to keep the wound site sterile, every post-injury procedure introduced a chance of infection which could have caused differences in the spinal cord tissue. The short survival period post-injury allowed the initial analysis of behaviour after trauma, but did not monitor the long term recovery patterns. Differences in forelimb motor function deficit, specifically digit dexterity and overground locomotion as a result of the damage to the rubrospinal [89] and reticulospinal tracts [90, 91] in the lateral and ventral 78  Chapter 4 Discussion funiculi, may have become more apparent over time. Most animals observed significant behavioural recovery following the first three days post-injury and received near perfect scores at seven days, with the exception of some of the most severe injury groups. Either the modified open field behavioural scoring system used was not sensitive enough to the injuries given, or a larger range of injury severity could have been used to help identify differences. This was somewhat indicated by the pilot study and injury parameters were increased from the initial study, but it was not apparent until the main experiment was in progress and could not be modified.  Histological analysis Sectioning of the spinal cord tissue was done while completely frozen with a 20 μm thickness between each cross-section. Some sections were lost during the slide mounting or immunostaining phase of analysis as a result of poor tissue quality following injury. Loss of these sections was unavoidable, but in cases where those sections marked the 1 mm offsets between observations, the next available section had to be used. This may have caused slight differences in the analysis of the tissue, although in each case a standard of moving to the section further from the epicentre was used to minimize compounding errors in section location. The staining techniques and protocols (Appendix D) used in this research were well defined and strictly adhered to; however, a degree of observer subjectivity could be applied during the analysis of the tissue. A single observer, blinded to the injury groups conducted the analysis with intermittent confirmation by a secondary observer to remove any effect of observer bias.  4.3.2  Computational model The computational model used in the current study was a modification of the rat  cervical spine developed by Russell et al and inherently retains the same limitations, primarily in the definition of the material properties. Material properties were initially determined by a hyperviscoelastic model and adjusted to comply with experimental behaviour. This material model could be improved with further validation with whole cord behaviour to fully simulate the response of the spinal cord. Particularly, the hyperelastic and 79  Chapter 4 Discussion viscoelastic constants were developed using data based on the experimental strain of rat spinal cord tissue of only up to 5% and the cord was assumed to be isotropic and homogeneous [59, 68]. Although these values provided a concrete starting point for the development of the material model, more detailed characterization of the rat spinal cord experimentally, along with more complex, fully nonlinear, viscoelastic models may be necessary to simulate a more accurate response. This model also did not include structural failure for the grey or white matter which may create different strain patterns in the tissue. Given the strong correlations between the FEA results and experimental tissue damage despite their differences in impact kinetics shown in Figure 3.5, a sensitivity analysis of strain to changes of the material properties could be warranted. This would help identify the dependency of the results on the geometry of the model and resulting impact or on the material properties. The observations from this sensitivity analysis could also help focus the priorities, in either material properties or more accurate physical representation of the spine, for future development of the computational model. The results from the computational portion of this work were also limited to only the correlations with the magnitude of maximum principal strain due to the time constraints. There may be merit in examining not only this parameter, but the other principal strain values, such as minimum principal strain, to look at the relationship with compressive strain. Perhaps the most identifiable limitation to the computational model was the position of the impactor on the spinal cord compared to the experimental set up. The simulated contusion injuries occurred at the C4/C5 level of the spinal cord corresponding to previous experimental protocols used by Choo [6, 49, 55]. Experimental long term cervical spinal cord injuries have since been moved to the C5/C6 level to reduce the respiratory issues that were observed with injuries at C4/C5. The changes to the experimental protocol were not reflected in the finite element model due to geometrical restrictions and remained outside the scope of the current work. Although this created a discrepancy between the simulated and experimental injuries, the material properties of the grey and white matter in the model used for this research were both isotropic and homogenous. Additionally, the nonlinear and viscoelastic properties of the cord were not dependent on the location of the contusion on the spinal cord and for these reasons the potential differences in the pattern and values of maximum principal strain at C4/C5 and C5/C6 were considered to be negligible. 80  Chapter 5 Conclusion 5.1 Conclusions Based on the findings from the current work, both the impact velocity and impact depth of contusion type spinal cord injuries are differently important factors when considering secondary injury effects.  The depth of contusion injuries appears to be a  consistent factor in qualifying the amount of axonal damage in the spinal cord and the resulting deficit. This is especially evident at the lower impact depth where the speed of impact had to be increased by two orders of magnitude to potentially elicit significantly different injury severities. Increased impact depth resulted in a decrease of healthy tissue sparing and an increase of axonal damage within that tissue over a larger area. Increased impact velocity typically only showed significance within ±1 mm rostrocaudally from the epicentre. Linear correlation with the two controlled impact factors confirmed a better or equal relationship with impact depth over velocity in all measures of spinal cord injury. The open field, grooming, and rearing behavioural tests used in this study supplemented the spinal cord tissue analysis and helped tie the corresponding tissue damage to physical outcomes based on the mechanical impact factors. These tests were found to be less sensitive to changes in the impact velocity and depth than the histological measures, but showed significantly decreased functional ability between the 0.9 mm and 1.5 mm impact depths and with increased impact velocity at the 1.5 mm depth only. These observations strongly reinforce the existence of an impact velocity and depth tolerance criterion and support the results shown by Kearney and Viano et al. [39, 64]. The inclusion and examination of the applied impact energy to the cord provides explanation for the determination of the criterion where the interactive effects of impact depth and velocity begin to have significant effect on outcome. The increase of impact depth at a given impact velocity in this work resulted in a significant increase in the impact energy while the incremental increases of velocity provided a smaller increase in energy at the smaller impact depth. Once the impact depth was increased to 1.5 mm, presumably beyond some threshold 81  Chapter 5 Conclusion interaction point with velocity, changes in velocity resulted in significantly different impact energies and measureable outcomes (axonal damage, tissue sparing, and forelimb function) simultaneously. A finite element model was modified and used to simulate all experimental contusion injury conditions used in this work to determine the maximum principal strain within the spinal cord and its relationship to tissue damage. Optimization and modification of an existing model was required to reduce the computational costs associated with the increased impact duration of the slow velocity impacts. Principal strains were calculated for specific regions of interested defined by Choo et al. at 1 mm intervals from the injury epicentre and compared to experimental behavioural and histological data collected simultaneously. Linear correlation analysis with FEA strain showed significant ( with axonal damage in the ventral ( cord and with white matter (  ) and lateral ( ) and grey matter (  ) and strong correlations ) regions of the spinal ) sparing. In addition to  the measured tissue damage, nonlinear correlation analysis identified strong correlations between both grey or white matter strain and open field behavioural scores ( ), but not with the grooming or rearing results. The finite element simulation indicated areas of high strain below the impactor tip which correlated strongly to the total amount of tissue destroyed and cell death. Strain patterns and approximate values in the white and grey matter matched very well with the models developed by Fiford et al. and Russell et al [59, 68]. No specific principal strain tolerance criterion, used to define physiological damage to the cord, was explicitly discovered as a result of the FEA. However, there were similar trends compared to axonal damage of the optic nerve, studied by Bain et al., which showed a threshold strain of 0.14 for morphological damage to the white matter [83].  5.2 Contributions This thesis attempts to elucidate the relationship between the biomechanics of SCI and the resulting biological damage. It is hoped that this work helps to identify important factors that contribute to SCI for further research and injury classification that could aid in targeted therapies.  82  Chapter 5 Conclusion The current work adds to contemporary animal model research of contusion SCI allowing identification of the individual impact factors on both the biological and physical outcomes. Specifically, the recognition of the importance of impact velocity and impact depth on secondary injury effects is clearly shown within the same experimental design, something which has not been closely studied in previous work. This study also attempts to explain the contributions of these impact factors to injury severity by examining the resulting total energy applied to the spinal cord and the resulting effects by controlling the applied impact energy under different impact velocity and depth. In addition to showing correlation between maximum principal strain with axonal damage and tissue sparing, this is first study that I am aware of that attempts to directly relate spinal cord strain with simple behavioural deficit measurements. Furthermore distinctly different strain patterns were identified between the 1.5 mm and 0.9 mm impact depths while increased impact velocity showed much more subtle differences.  The results from the  computational modelling portion of this work further validate, and potentially extend, the utility of the model.  5.3 Recommendations for future work The results and conclusions drawn from this work may be used by future research to advance the current knowledge of SCI biomechanics. In order to fully develop a threshold of injury severity based on impact velocity and depth a regression model could be used to quantify the absolute effect of each factor on contusion injuries.  Employing a larger  experimental design covering a larger span of impact factors with only one or two subjects per group could help identify the incremental effects of increasing impact velocity, depth, or energy and define regression model parameters. The initial secondary injury effects were studied in this thesis; however, no plan to view the recovery from of these injuries was completed. A longer survival study taken over 3-6 months would elucidate the recovery of animals and remyelination of axons over time. Given the behavioural results of the animals from this work, more severe injuries or a more sensitive rating scale may be required to view significant long term deficit and recovery patterns. As improved hand and arm functionality is a priority for patients with cervical SCI,  83  Chapter 5 Conclusion more complex behavioural studies could be employed to view targeted forelimb deficit resulting from different spinal cord injury impact parameters. As well as studying the effect of impact velocity on spinal cord injury future research should also include the consideration of impact dwell time on behavioural and biological outcomes as residual compression is often a clinical factor in human SCI. The understanding of the effect of impact dwell or time before decompression could help identify critical or optimal treatment timelines in human SCI where these methods are currently used. Further work in the future could also expand the current work by examining additional markers for axonal damage including neurofilament degradation and β-amyloid precursor protein (β-APP) or other commonly used markers of axonal pathology to study the regional strain differences caused by varied impact factors. Although outside the scope of this thesis, additional spinal cord tissue was collected following injury for this purpose and available to stain for others markers of injury or apoptosis. In addition to examining the effects of impact velocity and depth in contusion injuries, all the above recommendations and this work should be expanded to include other common injury mechanisms such as dislocations and distractions. The inclusion of other injury mechanisms will help identify if the behaviour of the spinal cord displays preferential response to increased impact depth/displacement or velocity with different loading conditions or if a general response is apparent. Both dislocation and distraction injury mechanisms are currently modelled using vertebral displacement as a control factor, but direct manipulation of the spinal cord itself may be required to create repeatable injuries. Computationally, the results from the FEA should also be expanded to include not only maximum principal strain, but also minimum principal strain and measures of principal stresses. Although maximum principal strain has shown correlation with injury severity, the direction of strain may also be important to determine given the anisotropic nature of the spinal cord. Future work in this direction should also look into incorporating anisotropic properties of both the grey and white matter in the FEA. Finally, from a computational perspective the data collected from this study could be used, in addition to future experimental work, to refine the material properties of the spinal cord. There has also yet to be any experimental studies which examine clinically relevant strain and strain rate values to determine the properties of the spinal cord. The majority of 84  Chapter 5 Conclusion spinal cord tissue properties have been identified using simple tensile testing, but the inclusion of this realistic injury deformation to contusions could vastly improve the simulated response. Furthermore, there still remains room for improvement in the geometry of the finite element model itself. Identification and inclusion of the nerve roots and realistic ligaments are important in both dislocation and distraction of the spinal cord and would affect the resulting strain distribution. These additional geometrical features may require a higher resolution MRI to identify and incorporate them into the existing model.  5.4 Concluding statement Spinal cord injury in the human population can be caused by a variety of factors which result in failure of the spinal column and damage of the spinal cord. In a very broad sense this thesis attempts to understand and define a small portion of the biomechanical inhomogeneity seen clinically, specifically of the impact velocity and depth of contusion injuries.  The results shown by this work identified distinct secondary injury effect  contributions of both impact velocity and depth to specific regions of the spinal cord following injury. These findings may better identify potential regions of the spinal cord for clinical therapies. Though not an original objective of this study, trends in the results suggest the existence of threshold strains and velocity/depth tolerance criterion which could assist in injury classification. Moreover, the correlation between the computational model strain and the experimental data suggest continued improvement of these models could reduce the amount of experimental analysis required to fully understand the biomechanics of SCI and eventually be incorporated into strategies for injury prediction and prevention development.  85  References [1] [2] [3] [4] [5] [6]  [7] [8] [9] [10]  [11]  [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23]  G. E. Pickett, et al., "Epidemiology of traumatic spinal cord injury in Canada," Spine, vol. 31, pp. 799-805, 2006. "Spinal Cord Injury Facts and Figures at a Glance," National Spinal Cord Injury Statistical Center at the University of Alabama at Birmingham, Technical Report, 2012. L. H. Sekhon and M. G. Fehlings, "Epidemiology, demographics, and pathophysiology of acute spinal cord injury," Spine, vol. 26, pp. S2-12, 2001. V. K. Noonan, et al., "Incidence and prevalence of spinal cord injury in Canada: a national perspective," Neuroepidemiology, vol. 38, pp. 219-26, 2012. Y. Cao, et al., "Lifetime Direct Costs After Spinal Cord Injury," Topics in Spinal Cord Injury Rehabilitation, vol. 16, pp. 10-16, 2011. A. M. Choo, et al., "Contusion, dislocation, and distraction: primary hemorrhage and membrane permeability in distinct mechanisms of spinal cord injury," Journal of Neurosurgery.Spine, vol. 6, pp. 255-266, 2007. C. J. Sparrey, et al., "The distribution of tissue damage in the spinal cord is influenced by the contusion velocity," Spine, vol. 33, pp. E812-9, 2008. C. H. Tator, "Review of treatment trials in human spinal cord injury: issues, difficulties, and recommendations," Neurosurgery, vol. 59, pp. 957-82; discussion 982-7, 2006. M. D. Norenberg, et al., "The pathology of human spinal cord injury: defining the problems," Journal of Neurotrauma, vol. 21, pp. 429-440, 2004. A. R. Blight, "Delayed demyelination and macrophage invasion: a candidate for secondary cell damage in spinal cord injury," Central Nervous System Trauma : journal of the American Paralysis Association, vol. 2, pp. 299-315, 1985. T. Goto and Y. Hoshino, "Electrophysiological, histological, and behavioral studies in a cat with acute compression of the spinal cord," Journal of Orthopaedic Science : Official Journal of the Japanese Orthopaedic Association, vol. 6, pp. 59-67, 2001. S. G. Waxman, "Demyelination in spinal cord injury," Journal of the Neurological Sciences, vol. 91, pp. 1-14, 1989. M. O. Totoiu and H. S. Keirstead, "Spinal cord injury is accompanied by chronic progressive demyelination," The Journal of Comparative Neurology, vol. 486, pp. 373-383, 2005. M. M. Siegenthaler, et al., "The extent of myelin pathology differs following contusion and transection spinal cord injury," Journal of Neurotrauma, vol. 24, pp. 1631-1646, 2007. "Body Planes", (2006). [Online Website]. Available: http://upload.wikimedia.org/wikipedia/commons/3/34/BodyPlanes.jpg [Accessed: Dec 21 2012]. B. D. Wingerd and G. Stein, Rat dissection manual. Baltimore, Md.: Johns Hopkins Univ. Pr., 1988. D. C. Rizzo, Fundamentals of Anatomy and Physiology, 1st ed. Clifton Park, NY, USA: Delmar Cengage Learning, 2001. B. K. Kwon, et al., "Animal models used in spinal cord regeneration research," Spine, vol. 27, pp. 1504-1510, 2002. F. Sebek, "Rat Skeleton". [Online Website]. Available: http://bclee.net/skeleton.html [Accessed: Jan 8 2013]. G. Paxinos, The Rat Nervous Sytem, 2nd ed.: Acad. Press, 1995. M. L. Barr and J. A. Kiernan, The human nervous system: an anatomical viewpoint: Lippincott, 1988. C. Watson, et al., The Spinal Cord: A Christopher and Dana Reeve Foundation Text and Atlas: Elsevier Science, 2009. R. Braakman and L. Penning, "Mechanisms of injury to the cervical cord," Paraplegia, vol. 10, pp. 314-20, Feb 1973.  86  References [24] [25] [26] [27]  [28] [29] [30] [31] [32] [33] [34]  [35] [36] [37] [38] [39] [40] [41]  [42] [43] [44] [45] [46]  L. Schnell, et al., "Acute inflammatory responses to mechanical lesions in the CNS: differences between brain and spinal cord," Eur J Neurosci, vol. 11, pp. 3648-58, Oct 1999. X. Z. Liu, et al., "Neuronal and glial apoptosis after traumatic spinal cord injury," J Neurosci, vol. 17, pp. 5395-406, Jul 15 1997. S. L. Shuman, et al., "Apoptosis of microglia and oligodendrocytes after spinal cord contusion in rats," J Neurosci Res, vol. 50, pp. 798-808, Dec 1 1997. A. Allen, "Surgery of experimental lesion of spinal cord equivalent to crush injury of fracture dislocation of spinal column: A preliminary report," Journal of the American Medical Association, vol. LVII, pp. 878-880, 1911. L. B. Jakeman, et al., "Traumatic spinal cord injury produced by controlled contusion in mouse," J Neurotrauma, vol. 17, pp. 299-319, Apr 2000. T. B. Ducker and H. F. Hamit, "Experimental treatments of acute spinal cord injury," J Neurosurg, vol. 30, pp. 693-7, Jun 1969. J. A. Gruner, "A monitored contusion model of spinal cord injury in the rat," Journal of Neurotrauma, vol. 9, pp. 123-6; discussion 126-8, 1992. F. C. Wagner, Jr., et al., "Pathological changes from acute to chronic in experimental spinal cord trauma," J Neurosurg, vol. 48, pp. 92-8, Jan 1978. M. S. Beattie, et al., "Endogenous repair after spinal cord contusion injuries in the rat," Exp Neurol, vol. 148, pp. 453-63, Dec 1997. J. T. Maikos and D. I. Shreiber, "Immediate damage to the blood-spinal cord barrier due to mechanical trauma," Journal of Neurotrauma, vol. 24, pp. 492-507, 2007. D. M. Basso, et al., "Graded Histological and Locomotor Outcomes after Spinal Cord Contusion Using the NYU Weight-Drop Device versus Transection," Experimental Neurology, vol. 139, pp. 244-256, 1996. D. H. Noyes, "Correlation between parameters of spinal cord impact and resultant injury," Experimental neurology, vol. 95, pp. 535-547, 1987. J. C. Bresnahan, et al., "A behavioral and anatomical analysis of spinal cord injury produced by a feedback-controlled impaction device," Experimental Neurology, vol. 95, pp. 548-570, 1987. S. W. Scheff, et al., "Experimental modeling of spinal cord injury: characterization of a forcedefined injury device," Journal of Neurotrauma, vol. 20, pp. 179-193, 2003. K. D. Anderson, et al., "Bilateral cervical contusion spinal cord injury in rats," Experimental Neurology, vol. 220, pp. 9-22, 2009. P. A. Kearney, et al., "Interaction of contact velocity and cord compression in determining the severity of spinal cord injury," Journal of Neurotrauma, vol. 5, pp. 187-208, 1988. T. E. Anderson, "A controlled pneumatic technique for experimental spinal cord contusion," J Neurosci Methods, vol. 6, pp. 327-33, Nov 1982. Y. P. Zhang, et al., "Spinal cord contusion based on precise vertebral stabilization and tissue displacement measured by combined assessment to discriminate small functional differences," J Neurotrauma, vol. 25, pp. 1227-40, Oct 2008. J. C. Bresnahan, et al., "Three-dimensional computer-assisted analysis of graded contusion lesions in the spinal cord of the rat," Journal of Neurotrauma, vol. 8, pp. 91-101, 1991. S. K. Somerson and B. T. Stokes, "Functional analysis of an electromechanical spinal cord injury device," Experimental Neurology, vol. 96, pp. 82-96, 1987. B. T. Stokes and P. J. Reier, "Fetal grafts alter chronic behavioral outcome after contusion damage to the adult rat spinal cord," Experimental Neurology, vol. 116, pp. 1-12, 1992. D. L. Behrmann, et al., "A comparison of YM-14673, U-50488H, and nalmefene after spinal cord injury in the rat," Experimental Neurology, vol. 119, pp. 258-267, 1993. D. L. Behrmann, et al., "Modeling of Acute Spinal Cord Injury in the Rat: Neuroprotection and Enhanced Recovery with Methylprednisolone, U-74006F and YM-14673," Experimental Neurology, vol. 126, pp. 61-75, 1994.  87  References [47]  [48] [49] [50]  [51]  [52] [53] [54] [55]  [56] [57] [58] [59] [60] [61] [62] [63]  [64] [65] [66]  [67] [68]  D. D. Pearse, et al., "Histopathological and behavioral characterization of a novel cervical spinal cord displacement contusion injury in the rat," Journal of Neurotrauma, vol. 22, pp. 680-702, 2005. R. J. Fiford, et al., "A vertebral dislocation model of spinal cord injury in rats," Journal of Neurotrauma, vol. 21, pp. 451-458, 2004. A. M. Choo, et al., "Secondary pathology following contusion, dislocation, and distraction spinal cord injuries," Experimental Neurology, vol. 212, pp. 490-506, 2008. E. C. Clarke and L. E. Bilston, "Contrasting biomechanics and neuropathology of spinal cord injury in neonatal and adult rats following vertebral dislocation," Journal of Neurotrauma, vol. 25, pp. 817-832, 2008. E. C. Clarke, et al., "Anterior fracture-dislocation is more severe than lateral: a biomechanical and neuropathological comparison in rat thoracolumbar spine," J Neurotrauma, vol. 25, pp. 37183, Apr 2008. E. J. Dolan, et al., "The effect of spinal distraction on regional spinal cord blood flow in cats," J Neurosurg, vol. 53, pp. 756-64, Dec 1980. R. J. Dumont, et al., "Acute spinal cord injury, part I: pathophysiologic mechanisms," Clin Neuropharmacol, vol. 24, pp. 254-64, Sep-Oct 2001. D. J. Maiman, et al., "Cord/spine motion in experimental spinal cord injury," J Spinal Disord, vol. 2, pp. 14-9, Mar 1989. A. M. Choo, et al., "Modeling spinal cord contusion, dislocation, and distraction: characterization of vertebral clamps, injury severities, and node of Ranvier deformations," J Neurosci Methods, vol. 181, pp. 6-17, Jun 30 2009. A. S. Rivlin and C. H. Tator, "Effect of duration of acute spinal cord compression in a new acute cord injury model in the rat," Surg Neurol, vol. 10, pp. 38-43, Jul 1978. S. Cheng, et al., "The effects of preconditioning strain on measured tissue properties," J Biomech, vol. 42, pp. 1360-2, Jun 19 2009. S. L. Woo, "Mechanical properties of tendons and ligaments. I. Quasi-static and nonlinear viscoelastic properties," Biorheology, vol. 19, pp. 385-96, 1982. R. J. Fiford and L. E. Bilston, "The mechanical properties of rat spinal cord in vitro," J Biomech, vol. 38, pp. 1509-15, Jul 2005. L. E. Bilston and L. E. Thibault, "The mechanical properties of the human cervical spinal cord in vitro," Ann Biomed Eng, vol. 24, pp. 67-74, Jan-Feb 1996. E. C. Clarke, et al., "The mechanical properties of neonatal rat spinal cord in vitro, and comparisons with adult," J Biomech, vol. 42, pp. 1397-402, Jul 22 2009. K. Ichihara, et al., "Gray matter of the bovine cervical spinal cord is mechanically more rigid and fragile than the white matter," J Neurotrauma, vol. 18, pp. 361-7, Mar 2001. A. R. Blight and V. Decrescito, "Morphometric analysis of experimental spinal cord injury in the cat: the relation of injury intensity to survival of myelinated axons," Neuroscience, vol. 19, pp. 321-341, 1986. D. C. Viano and I. V. Lau, "A viscous tolerance criterion for soft tissue injury assessment," Journal of Biomechanics, vol. 21, pp. 387-399, 1988. J. H. Kim, et al., "Impact speed does not determine severity of spinal cord injury in mice with fixed impact displacement," J Neurotrauma, vol. 26, pp. 1395-404, Aug 2009. C. Y. Greaves, et al., "A three-dimensional finite element model of the cervical spine with spinal cord: an investigation of three injury mechanisms," Ann Biomed Eng, vol. 36, pp. 396-405, Mar 2008. J. T. Maikos, et al., "Finite element analysis of spinal cord injury in the rat," Journal of Neurotrauma, vol. 25, pp. 795-816, 2008. C. M. Russell, et al., "Maximum principal strain correlates with spinal cord tissue damage in contusion and dislocation injuries in the rat cervical spine," J Neurotrauma, vol. 29, pp. 1574-85, May 20 2012.  88  References [69] [70]  [71] [72] [73] [74] [75] [76] [77]  [78] [79]  [80] [81] [82] [83]  [84] [85]  [86] [87]  [88] [89]  A. Blight, "Mechanical factors in experimental spinal cord injury," J Am Paraplegia Soc, vol. 11, pp. 26-34, Jul-Oct 1988. A. M. Choo and T. R. Oxland, "Improved RSA accuracy with DLT and balanced calibration marker distributions with an assessment of initial-calibration," J Biomech, vol. 36, pp. 259-64, Feb 2003. B. T. Stokes, et al., "An electromechanical spinal injury technique with dynamic sensitivity," Journal of Neurotrauma, vol. 9, pp. 187-195, 1992. M. M. Panjabi, et al., "Dynamic canal encroachment during thoracolumbar burst fractures," Journal of Spinal Disorders, vol. 8, pp. 39-48, 1995. R. W. Nightingale, et al., "Experimental impact injury to the cervical spine: relating motion of the head and the mechanism of injury," J Bone Joint Surg Am, vol. 78, pp. 412-21, Mar 1996. M. Martinez, et al., "A new rating scale for open-field evaluation of behavioral recovery after cervical spinal cord injury in rats," J Neurotrauma, vol. 26, pp. 1043-53, Jul 2009. M. L. Starkey, et al., "Assessing behavioural function following a pyramidotomy lesion of the corticospinal tract in adult mice," Exp Neurol, vol. 195, pp. 524-39, Oct 2005. J. C. Gensel, et al., "Behavioral and histological characterization of unilateral cervical spinal cord contusion injury in rats," J Neurotrauma, vol. 23, pp. 36-54, Jan 2006. J. A. Bertelli and J. C. Mira, "Behavioral evaluating methods in the objective clinical assessment of motor function after experimental brachial plexus reconstruction in the rat," J Neurosci Methods, vol. 46, pp. 203-8, Mar 1993. T. Schallert, et al., "Recovery of function after brain damage: severe and chronic disruption by diazepam," Brain Res, vol. 379, pp. 104-11, Jul 30 1986. Y. Liu, et al., "Transplants of fibroblasts genetically modified to express BDNF promote regeneration of adult rat rubrospinal axons and recovery of forelimb function," J Neurosci, vol. 19, pp. 4370-87, Jun 1 1999. G. G. Berntson, et al., "Brainstem systems and grooming behaviors," Ann N Y Acad Sci, vol. 525, pp. 350-62, 1988. A. A. Webb and G. D. Muir, "Compensatory locomotor adjustments of rats with cervical or thoracic spinal cord hemisections," J Neurotrauma, vol. 19, pp. 239-56, Feb 2002. K. Gale, et al., "Spinal cord contusion in the rat: behavioral analysis of functional neurologic impairment," Exp Neurol, vol. 88, pp. 123-34, Apr 1985. A. C. Bain and D. F. Meaney, "Tissue-level thresholds for axonal damage in an experimental model of central nervous system white matter injury," J Biomech Eng, vol. 122, pp. 615-22, Dec 2000. Q. Zhu, et al., "Predicting unconsciousness from a pediatric brain injury threshold," Dev Neurosci, vol. 28, pp. 388-95, 2006. D. M. McTigue, et al., "Neurotrophin-3 and brain-derived neurotrophic factor induce oligodendrocyte proliferation and myelination of regenerating axons in the contused adult rat spinal cord," The Journal of Neuroscience : The Official Journal of the Society for Neuroscience, vol. 18, pp. 5354-5365, 1998. R. Nashmi and M. G. Fehlings, "Changes in axonal physiology and morphology after chronic compressive injury of the rat thoracic spinal cord," Neuroscience, vol. 104, pp. 235-251, 2001. A. C. Bain, et al., "Dynamic stretch correlates to both morphological abnormalities and electrophysiological impairment in a model of traumatic axonal injury," J Neurotrauma, vol. 18, pp. 499-511, May 2001. K. A. Irvine, et al., "A novel method for assessing proximal and distal forelimb function in the rat: the Irvine, Beatties and Bresnahan (IBB) forelimb scale," J Vis Exp, 2010. G. W. Schrimsher and P. J. Reier, "Forelimb motor performance following dorsal column, dorsolateral funiculi, or ventrolateral funiculi lesions of the cervical spinal cord in the rat," Exp Neurol, vol. 120, pp. 264-76, Apr 1993.  89  References [90] [91] [92]  D. N. Loy, et al., "Functional redundancy of ventral spinal locomotor pathways," J Neurosci, vol. 22, pp. 315-23, Jan 1 2002. P. Schucht, et al., "Anatomical correlates of locomotor recovery following dorsal and ventral lesions of the rat spinal cord," Exp Neurol, vol. 176, pp. 143-53, Jul 2002. C. V. Howard and M. G. Reed, Unbiased Stereology: A Practical Guide: Bios, 1998.  90  Appendix A: UBC Machine Accuracy Testing  Figure A.1: Load reported vs. calibrated load applied (22.5 N load cell) Table A.1: 22.5 N load cell error  Root Mean Square Error (RMSE): Percent Error of Full-Scale: Slope: y-intercept:  0.10925 N 0.49% 0.9925 0.0254  91  Appendix A UBC Machine Testing  Figure A.2: Acceleration reported vs. calculated acceleration (500 g accelerometer) Table A.2: 500 g accelerometer error  Root Mean Square Error (RMSE): Percent Error of Full-Scale: Slope: y-intercept:  0.19535 g 0.04% 1.1013 0.0277  92  Appendix A UBC Machine Testing Data sampled for both the UBC multi-mechanism injury system and the DynaMight Instron was sampled at 4000 Hz. Force data for the UBC multi-mechanism injury system was filtered with a Butterworth low-pass filter with a cut off frequency of 630 Hz to remove the noise resulting from the natural frequency of the load cell.  1.5 mm Displacement, 80 mm/s Comparison  Displacement (mm)  30 20 10 0  UBC Machine DynaMight Instron  Contact Force (N)  -10 0 -2 -4 -6 -8  UBC Machine DynaMight Instron 0  0.1  0.2  0.3  0.4  0.5  0.6  0.7  0.8  0.9  1  Time (s) Figure A.3: DynaMight Instron vs. UBC Machine impact characterization (1.5 mm displacement, 80 mm/s) Table A.3: Instron dynamic impact parameter comparison (1.5 mm displacement, 80 mm/s)  1.5 mm, 80 mm/s Max Depth (mm) Max Force (N) Average Speed (mm/s) UBC Machine 1.497 6.602 85.1622 DynaMight 1.49145 6.096453 81.9455 % Error 0.37% 8.29% 3.93%  93  Appendix A UBC Machine Testing  1.5 mm Displacement, 100 mm/s Comparison  Displacement (mm)  40 30 20 10 0  UBC Machine DynaMight Instron  -10  Contact Force (N)  0 -2 -4 -6 -8  UBC Machine DynaMight Instron 0  0.1  0.2  0.3  0.4  0.5  0.6  0.7  0.8  0.9  1  Time (s) Figure A.4: DynaMight Instron vs. UBC Machine impact characterization (1.5 mm displacement, 100 mm/s) Table A.4: Instron dynamic impact parameter comparison (1.5 mm displacement, 100 mm/s)  1.5 mm, 100 mm/s Max Depth (mm) Max Force (N) Average Speed (mm/s) UBC Machine 1.453 6.646 105.2461 DynaMight 1.48355 6.09852 102.8173 % Error 2.06% 8.98% 2.36%  94  Appendix B: Primary Injury Pilot Study B.1 Methods All methods and procedures were approved by the University of British Columbia’s Office of Research Services Animal Care Committee. Fifteen male Sprague-Dawley rats were separated into four different contusion injury groups at two different impact velocities and impact depths (n = 3 at 2.0 mm, 80 mm/s; n = 3 at 2.0 mm, 800 mm/s; n = 3 at 0.7 mm, 80 mm/s; n = 3 at 0.7 mm, 800 mm/s) and one surgical sham group (n = 3). The mean ± SD weight of the animals used in this study was 291 ± 13.1 g. Animals were brought in to our vivarium and group housed in enriched, custom-designed environments with food and water one week before injury. All animals were deeply anaesthetized with isoflurane (5% for induction) and then placed in a nose cone at 1-2% to maintain a surgical plane of anaesthesia (foot pinch and corneal reflexes absent). Subcutaneous injection of buprenorphine (0.03 mg/kg) was given to mitigate any possibility of acute pain. At approximately 10 minutes prior to injury Alex Fluor 488-labeled bovine serum albumin (0.325 mg diluted in 200 μL phosphate buffered saline, Molecular Probes, Eugene, OR) was injected intravenously through the tail vein used to detect the extravasation volume of plasma immediately following injury [33]. The dura of the spinal cord was exposed dorsally by laminectomy, performed by a trained surgeon, between the C5 and C6 level. Once the spinal cord was exposed the C5 and C6 vertebrae were held as a single unit by custom designed vertebral clamps [55]. The 2 mm spherical impactor was lowered in 50 μm increments on to the spinal cord until an initial force of 0.03 N was recorded to provide a datum point for all animals and dimple the surface of the dura [47, 71]. Once the touch force was reached the impactor was retracted away from the surface of the dura and then accelerated towards the cord by a displacement controlled actuator to the specified maximum velocity and depth. In order to compensate for the inertial forces created by the impactor on the load cell a 500 g accelerometer was attached to the impactor column to determine these forces during blank air impact tests. The blood oxygen level and heart rate were monitored for each animal by a pulse oximeter during the injury. Animals were kept deeply anaesthetized for 5 minutes after the injury to allow for haemorrhaging around the injury site [33] before being euthanized with an overdose of 5% chloral hydrate intraperitoneally injected. 95  Appendix B Primary Injury Pilot Study All animals were perfused with an intracardial needle with 250 mL of phosphate buffered saline (PBS) and followed with 500 mL of 4% paraformaldehyde in order to fix the tissues. The spinal cord was then carefully removed from the animal approximately ±5 mm rostral and caudal around injury epicenter and fixed overnight in 4% paraformaldehyde. Each cord was then cryoprotected in increasing sucrose solutions (12%, 18%, and 24%) for 12 hours at each grade and frozen in isopentane before being cyrosectioned in 20 μm increments in the frontal plane into two serial sets.  B.1.1 Extravasation analysis The Alexa Fluor 488 labelled tracer contained in the tissue was immediately visible using epifluorescent microscopy and one set of slides of frozen spinal cord tissue, from all animals, was stained with a rabbit anti-red blood cell (RBC) antibody (Fitzgerald Industries International, Acton, MA) to estimate haemorrhage volume around the injury site.  Standard  immunohistochemistry staining protocols (Appendix D) were followed to label the tissue with the previously mentioned markers.  All slides were thawed from frozen at room temperature for  30 minutes before rehydrating for 10 minutes in 0.01 M PBS. The tissue was blocked for 30 minutes in 10% normal donkey serum before 2 hour incubation at room temperature in the primary antibody dilution (1:400 rabbit anti-RBC) in 0.01 M PBS containing 0.1% Triton X-100. Following incubation, slides were washed for 5 minutes in 0.01 M PBS three times and incubated in Alexa Fluor-594 donkey anti-rabbit secondary antibody (1:200 dilution in 0.01 M PBS, Jackson ImmunoResearch Laboratories, West Grove, PA) overnight in a light opaque container.  Sections were then washed for 5 minutes in 0.01 M PBS and mounted in  Fluoromount-G (Southern Biotechnology, Birmingham, AL) to help prevent photobleaching and covered with glass coverslips (No 1.5, VWR, Radnor, PA). For the extravasation analysis of both the BSA tracer and the rabbit anti-RBC marker, sections were examined on an AxioPlan2 microscope (Carl Zeiss, Thornwood, NY) using a 2.5x objective.  Images were captured through the attached monochrome camera (Retiga Exi,  QImaging, Burnaby, BC) using Northern Eclipse acquisition software (Empix Imaging, Mississauga, ON). Visual basic script was created to control the motorized scanning stage (Scan 100x100, Marzhauser, Wetzlar-Steindoft, Germany, MAC5000, controller, Ludl, Hawthorne, NY) and AxioPlan2 to incrementally move the sections through 3 mm rostral to 3 mm caudal of 96  Appendix B Primary Injury Pilot Study the epicentre of injury and capture all fluorescent markers in each colour channel. Each set of serial images were stitched together using Adobe Photoshop CS3 (v10.0) using the photomerge function to represent the entire section (Figure B.1).  Figure B.1: Fully stitched spinal cord section (Red: Rabbit anti-RBC, Green: Alexa Fluor 488-labelled BSA)  Extravasation volumes of both markers were calculated using the Cavaleri method [92] with a known spacing of 40 μm between sections. Area calculations on sections of each fluorescent marker were analyzed by converting each image to 8-bit grey scale and applying an adaptive threshold in ImageJ (v1.47c, Research Services Branch, National Institute of Mental Health, Bethesda, Maryland, USA) to create a black and white mask of the stained area.  Before  proceeding with analysis the investigator was blinded by renaming and randomizing the images using custom scripts written in Matlab.  B.2 Results Extravasation volume of each injury group in the pilot study did not produce any significant data due to the loss of tissue due to staining and slide mounting issues. However, the resulting issues from these protocols were addressed and modified to avoid complication in 97  Appendix B Primary Injury Pilot Study future studies. Although no significant conclusions could be drawn from the data, trends were visible in the haemorrhage volume which directed the impact parameter choices for the main study. There appeared to be no discernible differences between increased impact velocity groups on haemorrhage volume, but large increases when increasing impact depth (Table B.1). Table B.1: Total haemorrhage volume immediately following contusion injury  Number of Animals Average Hemorrhage Volume (mm3)  80 mm/s, 2.0 mm depth  800 mm/s, 2.0 mm depth  80 mm/s, 0.7 mm depth  800 mm/s, 0.7 mm depth  3  1  1  3  3.1665  2.386  0.64  0.506  Figure B.2: Representative images of haemorrhage volume immediately following injury (red: rabbit anti-RBC, green: Alexa Fluor 488-labelled BSA). Scale bars show 1 mm.  The force and displacement traces from the resulting injuries provided information regarding the duration to peak displacement and maximum force values of each injury group. Beyond an impact depth of 1.6 mm noticeable spikes in the force data, up to 7 times the expected values were seen and though to be a result of contact with the vertebrae (Figure B.3). As a result the maximum impact depth planned for the main survival contusion injuries was reduced to 1.5 mm.  98  Appendix B Primary Injury Pilot Study  Figure B.3: Force vs. time trace during impact of all injury groups. Large increases forces were observed in both the 2.0 mm depth injuries.  99  Appendix C: Monitoring and Scoring Sheets Rearing/Grooming Scoring Sheet  100  Open Field Scoring Sheet  Appendix C Monitoring and Scoring Sheets  101  Post-Injury Monitoring Sheet  Appendix C Monitoring and Scoring Sheets  102  Appendix C Monitoring and Scoring Sheets  103  Appendix D: Staining Protocols Immunohistochemical Staining Protocol Use Robertsons’s 0.01 M PBS for using   Stock solution is at 0.1 M    Dilute 1 in 10 to make 0.01M PBS    PBST is 0.01 M PBS with 0.1% Triton X-100  All steps done at room temperature 1. Thaw slides  30 minutes (minimum)  2. Apply PAP pen around slide edge 3. Rehydrate in 0.01 M PBS  10 minutes  4. Delipidization (Only when staining with myelin based protein) a. Use 50, 70, 90, 95, and 100% ethanol bottles b. 50%  70%  90%  95%  100%  95%  90%  70%  50% for 2 minutes at each concentration c. Wash slides in 0.01 M PBS  3 x 5 minutes  5. Block with 10% normal donkey serum  30 minutes  6. Apply primary antibodies diluted in PBST to slides  Overnight  7. Wash slides in 0.01 M PBS  3 x 5 minutes  8. Apply secondary antibodies diluted in PBST to slides  2 hours  9. Wash slides in 0.01 M PBS  3 x 5 minutes  10. Coverslip with Fluromount-G  104  Appendix D Staining Protocols Luxol Fast Blue Staining Protocol     1. 2. 3. 4. 5. 6. 7.  1% Luxol Fast Blue Solution (LFB) o Luxol Fast Blue (Solvent Blue 38) 10 g o 95% EtOH 800 mL o Glacial acetic acid (add in fumehood) 5 mL o Top up to 1 L with 95% EtOH o Filter into brown bottle o Stain can be reused. Filter again before use if unused for a long time. 0.05% Lithium Carbonate (for differentiation) o Lithium Carbonate 0.5 g o dH2O 800 mL o Top up to 1 L with dH2O o Do not reuse. Make fresh. Thaw slides 0.01 PBS x 2 50% EtOH 70% EtOH 90% EtOH 95% EtOH x 2 1% LFB solution o Seal jars with parafilm or a Ziploc bag.  8. 95% EtOH x 2 9. dH2O  1 hr 10 min each 2 min 2 min 2 min 2 min each Overnight at 60oC  2 min 2 min  10. 11. 12. 13.  0.05% Lithium Carbonate 5 sec, with light agitation 70% EtOH x 2 10 sec each, with light agitation dH2O rinse until dye stop coming out Repeat steps 10-12 until good differentiation o White matter will be darker than grey matter o Careful not to over-differentiate. If this happens, go back to step 7  14. 15. 16. 17. 18. 19.  50% EtOH 70% EtOH 90% EtOH 100% EtOH x 2 Xyelene x 2 (in fumehood, separate jars) Coverslip with SHUR/mount (in fumehood)  2 min 2 min 2 min 2 min each 2 min each  105  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0071997/manifest

Comment

Related Items