Delineation of Vascular Disruption and Investigation of a Bioengineered ZFP-VEGF Gene Therapy Following

Traumatic Spinal Cord Injury

by

Sarah A. Figley

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy

Institute of Medical Science University of Toronto

© Copyright by Sarah A. Figley (2013)

ii

Delineation of Vascular Disruption and Investigation of a Bioengineered ZFP-VEGF Gene Therapy Following Traumatic

Spinal Cord Injury

Sarah A. Figley

Doctor of Philosophy

Institute of Medical Science University of Toronto

2013

Abstract

Background: Traumatic spinal cord injury (SCI) results in vascular disruption which appears to

contribute to the pathobiology of SCI. Vascular endothelial growth factor (VEGF) is known for

vascular development and repair, and more recently for its neuroprotective properties. Given this,

I investigated the temporal-spatial changes to the spinal vasculature, as well as examined the role

of VEGF as a therapeutic approach for SCI.

Hypothesis: It is hypothesized that clip-compression injury will result in significant vascular

changes, and that ZFP-VEGF gene therapy will enhance molecular and functional recovery

following spinal cord injury.

Methods: Briefly, female Wistar rats received a two-level laminectomy and a 35g clip-

compression injury at T6-T7 for 1 minute. Control animals received a laminectomy only. AdV-

ZFP-VEGF or AdV-eGFP was administered 24 hour post-injury by intraspinal injection. For

molecular and vascular analysis, tissues were extracted at various time points between 1 hour and

14 days post-SCI. For behavioural experiments animals were studied for 8 consecutive weeks.

Results: I have shown that vasculature undergoes structural and functional changes, which occur

as early as 1 hour following SCI. Although endogenous improvement is observed, SCI results in

permanent vascular damage. Animals receiving AdV-ZFP-VEGF treatment had increased levels

of VEGF mRNA and protein. AdV-ZFP-VEGF resulted in neuroprotection, as observed by

increased NF200 protein and NeuN counts, and decreased TUNEL staining. Animals treated

with AdV-ZFP-VEGF also showed an increased number of newly formed vessels (angiogenesis),

iii

as well as an increase in total number of vessels. Moreover, animals treated with AdV-ZFP-

VEGF showed significant increases in hindlimb weight support and reduction neuropathic pain.

Conclusions: I have characterized the dramatic temporal-spatial changes which occur in the

spinal vasculature following SCI. Additionally, I have demonstrated that AdV-ZFP-VEGF

administration results in beneficial molecular and functional outcomes. Overall, the results of

this study indicate that AdV-ZFP-VEGF administration can be delivered in a clinically relevant

time-window following SCI (24 hours) and provide significant molecular and neurobehavioural

benefits, by acting through angiogenic and neuroprotective mechanisms.

iv

Contributions

For the research described in this thesis, I maintain primary accountability for all results and

interpretation of such results. I was principally responsible for the experimental design and

execution of the project, including surgical procedures, animal care/ethics, behavioural testing,

tissue processing, data collection/quantification, data analysis, data interpretation, and writing of

this document. However, I acknowledge that I received substantial assistance from lab members

who are experts in certain scientific areas, and I was responsible for training a number of

research students who significantly contributed to the collection of the data displayed in this

thesis. Therefore, I wish to formally recognize the scientific contributions of these individuals

for their involvement in this research.

Dr. Kajana Satkunendrarajah collected and analyzed all electrophysiological data from

animals at 8 weeks following injury. Additionally, Kajana was involved in the interpretation and

discussion of the electrophysiology data.

Spyros Karadimas assisted with Catwalk™ data collection for long-term behavioural

experiments. Spyros was solely responsible for paw/print identification and data analysis of the

fore-limb and hindlimb gait data acquired by the Catwalk™ system.

Ramak Khosravi, Michelle Legasto, Christine Tseng, Sofia Khan, Eunice Cho, and Peter

Fettes were involved in histological sectioning of spinal cord tissue, HE/LFB staining,

immunohistochemical staining for RECA-1, Ki67, FITC-LEA, TUNEL, and NeuN,

quantification of immunohistochemistry, and lesion/tissue sparing analysis for 8 week post-SCI

tissues. Students also assisted in capturing microscopic images from spinal cord tissue, some of

which are used in this thesis.

v

Sangamo Biosciences (Drs. Kaye Spratt, Gary Lee, Dale Ando, Martin Giedlin and Richard

Surosky) – This project was made possible by financial, technical, and laboratory reagent

contributions from Sangamo Biosciences Inc. located in California, USA. Sangamo Biosciences

Inc. invented and patented the ZFP-VEGF gene therapy technology. Additionally, scientists

from Sangamo Biosciences Inc. generated all viral constructs used in the following experiments,

including viral development, production, and titer quantification. Sangamo also provided

scientific review of the project and manuscripts.

Yang Liu and Dr. Sherri Steele provided extensive surgical and behavioural assessment

training to me during my research.

Behzad Azad provided substantial post-operative animal care for animals in all experiments

throughout my entire thesis. He was also involved in long-term behavioural assessments.

vi

Acknowledgments

In many ways, this (albeit small) section of my thesis, is the most important of all. The people

who have been part of my “Ph.D. journey” have helped me maintain my sanity, stay focused,

grow, learn, have fun, stay active and become a fantastic scientist (ok… I might be biased on

that, but we can all agree that I’ve become a bit of a scientist). This thesis encompasses over five

amazing years of laughter, blood, sweat and tears (not too many tears), and I could not have done

it alone – nor would I have wanted to. To all the people who been involved along the way – in

any way, big or small – I cannot thank you enough. In no particular order, I would like to say a

special thank you to the following people:

To Dr. Fehlings: Thank you for taking me on as a graduate student. Although I came into the

lab with no neuroscience experience/knowledge, you believed in me. I have enjoyed my time as

a member of your research group, and I want to thank you for this wonderful opportunity. Thank

you for your mentorship and allowing me to work and learn independently. Academically, your

lab offered a rich learning environment; however, it was the people in the lab (some who have

turned into life-long friends) that truly made the experience.

To my Program Advisory Committee – Dr. Eubanks, Dr. Koeberle and Dr. Marsden:

Thank you for your mentorship, guidance and support. Thank you for forcing me to “think

outside the box”, and achieve my full potential as a scientist.

To my parents – Don and Shirley: The words “thank you” do not begin to describe my

gratitude for everything you have given, taught, or shared with me. Thank you for your support,

your love and your guidance. Thank you for teaching me the value of hard work, but also for

vii

encouraging me to have fun, experience new things, and enjoy life outside of work. I couldn’t

have done this degree without you guys. I love you.

To Chase: you are an inspiration to any scientist. Your enthusiasm and dedication to

understanding science (and REALLY getting it, not just pretending to get it) is a quality that I

admire. You have a passion for science and learning that is contagious, which is partially why I

like spending so much time with you . Although PhD’s run in the family, I can honestly say

that starting grad school and pursuing a PhD was definitely a “side effect” of striving to be like

you – perhaps this is the highest form of flattery. Thank you for all your supportive and

motivating talks over the years… It was truly a blessing to have someone who was so

considerate and empathetic.

To Alisha: you are amazing and one of my best friends! In addition to all the good times and

laughs we have shared together, you have also kept me grounded on my grad school journey.

You have been so kind and encouraging at every step of the way. You keep telling me, “You can

have anything you want… just work your butt off for it!” This is exactly what I needed to hear.

Thank you for being the wisest and most motivating younger sibling.

To Jessica: thanks for everything you’ve taught me about life and the search for internal

happiness. As someone who “dances to the beat of their own drum”, you have inspired me to

seek passion, love, and genuine enjoyment in things that are different than from what people

expected for us. You have the greatest compassion for this world and every living thing on this

planet, and everyone who has ever met you should feel so lucky to have had you in their

company. Thanks for all the laughs, cries, debates, adventures, dancing, and baking fiascos…

Here’s hoping that in the years to come we share a million more memories together! xo

viii

To my Grandparents - Lorne, Jo, Fred and June: Thank you for your love, hugs, and

childhood memories. Thanks for raising your kids right… They turned out to be the best parents

you could ask for.

To Dionne, Ashley and Lindsay: You three lovely ladies are the best friends a girl could ask

for! Thank you for helping me through my education and my life! You always seem to give the

best advice, or know the right thing to say. Without you, my journey through life, school and

Toronto would have been terrible… Although, I think we can all agree that it might have been

less expensive if you weren’t around . It hasn’t been an easy five years apart, and I appreciate

your love, patience, guidance and faith in our friendship. As promised, I’m finally returning

home so we can be together like old times. Let the fun begin!

To John Weil: Thank you for all the stories you shared, and for everything you ever taught me.

Your love and commitment to science was inspirational, and it is partly because of you that I

made it to where I am today. You were a dear friend and I will miss you.

To Rosemary Marchant: Living in Toronto, I have very much enjoyed getting to know you as

an adult. You have been so kind, generous, supportive and helpful in all of my endeavours

(academic, or not). Knowing you exposed me to many of my favourite Toronto memories:

Massey College, and all of the wonderful friends that I made there. You are a gracious host, and

a loving soul. Thank you for everything.

To Massih: Graduate school would have been so different without you, and I am forever grateful

that our paths crossed! You are a great friend, a wonderful scientist, and someone who I admire

and respect. Thanks for all the social outings, the debates, the coffee breaks, and the late nights

at the lab… Having you around made my PhD infinitely more enjoyable!

ix

To Alex Weber: You and I can go for months without talking, and just pick up right where we

left off… That’s how I know we’ll be friends for a long time! Thanks for being such a positive

part of my life and always making me smile.

To the Fehlings’ Team – Sherri Steele, James Rowland, James Austin, Kajana

Satkunendrarajah, Hoang Nguyen, Nicole Forgione, Hamideh Emrani, Ryan Salewski, Jared

Wilcox, Spyros Karadimas, Alex Laliberte, Sukhvinder Kalsi-Ryan, Stuart Faulkner, Allyson

Tighe, Amy Lem, Rahki Sharma, Yang Liu, Behzad Azad, and Jian Wang: Thank you for

teaching me everything I know! You’ve been so helpful and compassionate along the way, and

I am forever grateful.

To Jennifer Jin, Eva Dias and Sara Züger: You’re my original lab/science friends! I’m so

lucky to have maintained great friendships with you over the years. Thanks for all your support

and words of kindness along the way.

To Shelby (and the Sebestyen-McLean clan): You are my oldest childhood friend, and we’ve

shared so many memories together! It’s hard to believe how far we’ve come since those care-

free summers at the lake, but I’m so happy that you’ve been around for the journey of “growing

up”! Whenever I need to be cheered up, I know I can always count on you for a laugh. xo

To Matt Hassler: You’re like a second brother to me; always providing the right combination of

pestering mixed with love. I enjoyed our journey through grad school together… It’s always nice

to have someone to commiserate with . Thanks for all the laughs, support and hugs over the

years, and hopefully we’ll live closer together someday soon!

To Eunice Cho, Sofia Khan, Ramak Khosravi, Michelle Legasto, Christine Tseng and Peter

Fettes: THANK YOU!!! I truly could not have accomplished everything in my research

x

without your help, and I am forever in your debt. Thank you for being so dedicated and

interested in science.

To Massey College and John Fraser: Thank you for being my “home away from home”.

Massey College holds many of my fondest memories of living in Toronto, and I cannot begin to

express my gratitude for the opportunity to be part of this wonderful community! Master Fraser,

you have built and maintained an environment that is all-inclusive, genuine, and diverse, and I

will always remember and cherish the time I spend at Massey. I hope to be back for a visit soon!

To the Chapman Family: All I can say is that I wish I would have met you 5 years ago! The

time we have spent together has been amazing and my only regret is that we didn’t have more of

it. You are a lovely, fun, athletic, and kind family, and I thank you for letting me be a part of it.

Thanks for your motivation and support along this journey… It wouldn’t have been the same

without you!

To Scott and Paulette Batson: You are two of the nicest people in this world!! Thank you for

accepting me into your family with such open arms, and taking care of me over the years. It’s

difficult to express the extent of my gratitude in words, but just know that I am forever thankful

for all your kindness and the time we shared (and hopefully will continue to share) together.

To Morgan: Thank you for being such an amazing friend (in fact, you’re more like a sister).

You’re so many things: smart, kind, pretty, graceful, fun, spontaneous, loveable, genuine,

dedicated, passionate, reliable, honest and caring… And you have been each of these things

to/for me. Most of all, you’re always the “voice of reason”, so thanks for keeping me in line!

We’ve made some great memories over the years, and I’m really looking forward to whatever

crazy idea(s) we come up with next

xi

To Laura, Glynn and Kate: I wish we had met in my first years of being in Toronto, instead of

the last few… But life isn’t always fair. Thanks for making so many good memories filled with

uncontrollable laughter and inside jokes, and being supportive through the bad ones. When I’m

out riding (not hills, because those aren’t in the prairies) I’ll be thinking of you!

To Ben: You are one of my best friends, and one of my favourite people in this world. I respect

you, and have always valued your opinions and friendship. You have taught me many things

about the world and about myself, which have shaped me (I hope) into a better person. For this,

I am eternally thankful. You’re one of the most dedicated, driven people I know (sometimes too

much for your own good), but you somehow remain calm and poised while doing an amazing job

– this is to be admired. Your dedication and perseverance are contagious, and these qualities

helped get me to where I am today. To say, “I couldn’t have done it without you”, would be a bit

much (I probably could have ), but this PhD and the adventures of my life certainly wouldn’t

have been as much fun without you! And now, you can finally stop asking “how’s your thesis

coming along?”… Because, here it is! Xo

To John Mayer: Your songs have kept me company (and awake) for many long, late nights at

the lab. Thank you for making music that is beautiful.

xii

Abbreviations

AAV Adeno-Associated Virus

AdV Adenovirus

AdV-ZFP-VEGF Adenovirus Synthesizing a VEGF Specific Zinc-Finger Protein

AIB α-Aminoisobutyric Acid

AJ Adherens Junction

Akt A Serine/Threonine-specific Protein Kinase

Ang Angiopoietin

ASIA American Spinal Injury Association

ATP Adenosine Triphosphate

BBB Blood Brain Barrier

BBB scale Basso, Beattie, Bresnahan Locomotor Rating Scale

BCFB Blood-Cerebrospinal Fluid Barrier

BDNF Brain Derived Neurotrophic Factor

BSCB Blood-Spinal Cord Barrier

Ca2+ Calcium

Caspase Cysteine Protease, Which Cleaves at an Aspartate Residue

CNS Central Nervous System

COX-1/COX-2 Cyclooxygenase-1 or -2

CPG Central Pattern Generator

CsA Cyclosporin-A

CSF Cerebrospinal Fluid

CSPG Chondroitin Sulfate Proteoglycan

xiii

Da Dalton

DMF N,N'-dimethylformamide

DNA Deoxyribonucleic Acid

DRG Dorsal Root Ganglion

EB Evans Blue

ECF Extracellular Fluid

ECM Extracellular Matrix

EGF Epidermal Growth Factor

Erk Extracellular Signal-regulated Kinases

HAMC Hyaluronic Acid Methyl Cellulose

HRP Horseradish Peroxidase

IGF-1 Insulin-like Growth Factor 1

IHC Immunohistochemistry

IL Interleukin

K+ Potassium

KDa Kilodalton

LFB Luxol Fast Blue

LIF Leukemia Inducing Factor

MAPK Mitogen Activated Protein Kinase

MMP Matrix Metalloproteinase

MPSS Methylprednisolone

MPTP Mitochondrial Permeability Transition Pore

MW Molecular Weight

Na+ Sodium

xiv

NASCIS National Acute Spinal Cord Injury Study

NVU Neurovascular Unit

NFkB Nuclear Factor Kappa Light Polypeptide Gene Enhancer in B-

Cells

NGF Nerve Growth Factor

NO Nitrous Oxide

NP Neuropilin

NPC Neural Precursor Cell

OPC Oligodendrocyte Precursor Cell

p38 p38 MAPK

PAGE Poly Acrylamide Gel Electrophoresis

PBS Phosphate Buffered Saline

PDGF Platelet Derived Growth Factor

PECAM-1 Platelet Endothelial Cell Adhesion Molecule 1

PEG Polyethylene Glycol

PFU Plaque Forming Units

PGE2 Prostaglandin E2

PI3K Phosphatidylinositol 3-Kinase

PLGF Placental Growth Factor

PTS Post-Traumatic Syringomyelia

PVS Perivascular Space

qRT-PCR Quantitative Real-Time Polymerase Chain Reaction

Ras Small GTPase Involved in Cell Signaling

ROS Reactive Oxygen Species

xv

SAS Subarachnoid Space

SCI Spinal Cord Injury

TBI Traumatic Brain Injury

TGF-β Tissue Growth Factor Beta

TJ Tight Junction

TNF-α Tumor Necrosis Factor Alpha

TNFR Tumor Necrosis Factor Receptor

TUNEL Deoxynucleotide Transferase dUTP Nick End-Labeling

VE-Cadherin Vascular Endothelial Cadherin

VEGF Vascular Endothelial Growth Factor

VEGF-NPC VEGF Over-expressing Neural Precursor Cell

VEGFR Vascular Endothelial Growth Factor Receptor

VHL Von Hippel-Lindau

ZFP Zinc-Finger Protein

ZO-1/ ZO-2/ ZO-3 Zona Occludens

xvi

Table of Contents

Abstract ........................................................................................................................................... ii

Contributions .................................................................................................................................. iv

Acknowledgments .......................................................................................................................... vi

Abbreviations ................................................................................................................................ xii

Table of Contents ......................................................................................................................... xvi

List of Figures ............................................................................................................................. xxii

List of Tables .............................................................................................................................. xxv

List of Appendices ..................................................................................................................... xxvi

Chapter 1 ....................................................................................................................................... 1

1 Introduction ................................................................................................................................ 1

1.1 Overview of Spinal Cord Injury ......................................................................................... 1

1.2 Anatomy of the Spinal Cord ............................................................................................... 2

1.2.1 Neuroanatomy of the Spinal Cord .......................................................................... 2

1.2.2 The Meninges .......................................................................................................... 6

1.2.3 Vascular Organization and Blood Flow .................................................................. 7

1.2.4 Cerebrospinal Fluid ............................................................................................... 12

1.2.5 The Blood-Spinal Cord and Cerebrospinal Fluid Barriers ................................... 12

1.3 Epidemiology of Spinal Cord Injury ................................................................................. 18

1.4 Primary and Secondary Injury .......................................................................................... 20

1.5 Animal Models of Spinal Cord Injury .............................................................................. 21

1.6 Neurological Function Following Spinal Cord Injury ...................................................... 23

1.6.1 Motor Function ..................................................................................................... 23

1.6.2 Sensory Function .................................................................................................. 25

1.6.3 Autonomic Function ............................................................................................. 25

xvii

1.7 Secondary Injury ............................................................................................................... 26

1.7.1 Summary of Secondary Injury Progression .......................................................... 27

1.7.2 The Acutely Injured Spinal Cord .......................................................................... 29

1.7.3 The Sub-Acutely Injured Spinal Cord .................................................................. 37

1.7.4 The Intermediate Phase ......................................................................................... 42

1.7.5 The Chronically Injured Spinal Cord .................................................................... 42

1.8 Non-Traumatic Causes of Spinal Cord Injury .................................................................. 44

1.9 Therapeutic Targets .......................................................................................................... 45

1.9.1 Cell Transplantation .............................................................................................. 45

1.9.2 Molecular and Neuroprotective Therapies ............................................................ 48

1.9.3 Rehabilitation Therapies ....................................................................................... 50

1.10 Summary of Spinal Cord Injury ........................................................................................ 51

1.11 Vascular Growth and Development .................................................................................. 52

1.11.1 Vasculogenesis ...................................................................................................... 53

1.11.2 Angiogenesis ......................................................................................................... 56

1.12 Vascular Endothelial Growth Factor ................................................................................ 59

1.12.1 Molecular Biology of VEGF ................................................................................ 59

1.12.2 VEGF Receptors and VEGF Signaling ................................................................. 61

1.12.3 Regulation of VEGF ............................................................................................. 65

1.12.4 VEGF in Models of Neurotrauma ......................................................................... 67

1.12.5 Angiogenesis Following Injury ............................................................................. 69

1.13 Gene Therapy .................................................................................................................... 70

1.14 ZFP-VEGF Technology and Production of VEGF ........................................................... 72

Chapter 2 ..................................................................................................................................... 78

2 General Methods ...................................................................................................................... 78

2.1 Animal Model of SCI and Intraspinal Injections .............................................................. 78

xviii

2.2 Viral Vector Constructs .................................................................................................... 80

2.3 Western Blotting ............................................................................................................... 81

2.4 Evans Blue: Blood-Spinal Cord Barrier Disruption ......................................................... 83

2.5 Histochemistry .................................................................................................................. 83

2.5.1 Histological Processing ......................................................................................... 83

2.5.2 Immunohistochemistry ......................................................................................... 84

2.5.3 Quantification of Blood Vessels ........................................................................... 85

2.5.4 Identification of Functional Blood Vessels ........................................................... 86

2.5.5 Quantification of Apoptosis .................................................................................. 88

2.5.6 Quantification of Neurons ..................................................................................... 88

2.5.7 Quantification of Angiogenesis ............................................................................ 88

2.5.8 Assessment of Tissue Sparing and Cavity Formation .......................................... 89

2.6 Behavioural Testing .......................................................................................................... 89

2.6.1 Open-Field Locomotor Scoring ............................................................................ 89

2.6.2 Automated Gait Analysis (Catwalk™) ................................................................. 90

2.6.3 Neuropathic Pain: Von Frey Filaments ................................................................. 91

2.7 Electrophysiology ............................................................................................................. 92

2.7.1 Motor Evoked Potentials ....................................................................................... 92

2.7.2 H-Reflex ................................................................................................................ 93

2.8 Statistical Analysis ............................................................................................................ 93

Chapter 3 ..................................................................................................................................... 95

3 Characterization of Vascular Disruption and Blood-Spinal Cord Barrier Permeability Following Traumatic Spinal Cord Injury ................................................................................. 95

3.1 Abstract ............................................................................................................................. 95

3.2 Introduction ....................................................................................................................... 96

3.3 Objective ......................................................................................................................... 100

xix

3.4 Hypothesis ....................................................................................................................... 100

3.5 Specific Aims .................................................................................................................. 100

3.6 Methods ........................................................................................................................... 101

3.7 Results ............................................................................................................................. 106

3.7.1 BSCB permeability following SCI ..................................................................... 106

3.7.2 Spatial-temporal disruption of the vasculature ................................................... 108

3.7.3 Endogenous Angiogenesis Occurs Following SCI ............................................. 117

3.8 Discussion ....................................................................................................................... 119

3.9 Conclusions ..................................................................................................................... 122

Chapter 4 ................................................................................................................................... 124

4 Delayed AdV-ZFP-VEGF Administration Provides Neuroprotection and Promotes Angiogenesis Post-SCI........................................................................................................... 124

4.1 Abstract ........................................................................................................................... 124

4.2 Introduction ..................................................................................................................... 125

4.3 Objective ......................................................................................................................... 127

4.4 Hypothesis ....................................................................................................................... 127

4.5 Specific Aims .................................................................................................................. 127

4.6 Methods ........................................................................................................................... 128

4.7 Results ............................................................................................................................. 135

4.7.1 AdV-ZFP-VEGF Delivery into the Injured Spinal Cord .................................... 135

4.7.2 VEGF mRNA and protein expression is increased following 24 hour delayed AdV-ZFP-VEGF administration ......................................................................... 138

4.7.3 Apoptosis is reduced in animals treated with AdV-ZFP-VEGF 24 hours post-SCI ...................................................................................................................... 140

4.7.4 24 hour delayed AdV-ZFP-VEGF administration provides neuroprotection ..... 143

4.7.5 24 hour delayed AdV-ZFP-VEGF administration results in an increased number of vessels ................................................................................................ 147

4.7.6 AdV-ZFP-VEGF promotes angiogenesis ........................................................... 149

xx

4.8 Discussion ....................................................................................................................... 151

4.9 Conclusions ..................................................................................................................... 153

Chapter 5 ................................................................................................................................... 154

5 AdV-ZFP-VEGF Results in Functional Improvements and Reduced Allodynia Following SCI.......................................................................................................................................... 154

5.1 Abstract ........................................................................................................................... 154

5.2 Introduction ..................................................................................................................... 155

5.3 Objective ......................................................................................................................... 157

5.4 Hypothesis ....................................................................................................................... 157

5.5 Specific Aims .................................................................................................................. 157

5.6 Methods ........................................................................................................................... 158

5.7 Results ............................................................................................................................. 166

5.7.1 AdV-ZFP-VEGF results in functional improvement .......................................... 166

5.7.2 AdV-ZFP-VEGF does not result in improved BBB scores ................................ 170

5.7.3 Delayed AdV-ZFP-VEGF administration does not improve motor evoked potentials or H-reflex following SCI .................................................................. 172

5.7.4 AdV-ZFP-VEGF administration significantly reduces allodynia ....................... 174

5.7.5 AdV-ZFP-VEGF treatment results in spared grey matter, but not white matter tissue at 8 weeks post-SCI .................................................................................. 177

5.8 Discussion ....................................................................................................................... 180

5.9 Conclusions ..................................................................................................................... 183

Chapter 6 ................................................................................................................................... 185

6 General Discussion and Future Directions ............................................................................. 185

6.1 Potential mechanisms of VEGF-A treatment ................................................................. 186

6.2 Vasculature damage plays an important role in SCI ....................................................... 188

6.3 Targeting the neurovascular niche .................................................................................. 189

6.4 Advantages of AdV-ZFP-VEGF compared to other VEGF therapies ............................ 190

xxi

6.5 Potential disadvantages of VEGF therapies .................................................................... 191

6.6 Comparison of Results to Other SCI Therapies .............................................................. 192

6.6.1 AdV-ZFP-VEGF: Immediate vs. 24-hour Administration ................................. 192

6.6.2 Comparison to Other Vascular Therapies ........................................................... 196

6.6.3 Comparison to Other Neuroprotective Therapies ............................................... 197

6.7 Future Research .............................................................................................................. 199

6.7.1 Investigating the Glial Scar and Inflammation ................................................... 199

6.7.2 Alternative ZFP-VEGF Delivery ........................................................................ 201

6.7.3 Elucidating the Functional and Sensory Benefits of AdV-ZFP-VEGF .............. 206

6.7.4 Imaging Vascular Changes Using a Spinal Cord Window Chamber ................. 206

6.7.5 Further Investigation of the Blood-Spinal Cord Barrier ..................................... 207

6.7.6 Multiple Angiogenic Factors as a Potential Treatment Option .......................... 209

6.8 Final Conclusions ............................................................................................................ 210

References ................................................................................................................................... 211

xxii

List of Figures

Figure 1. The arrangement of the spinal cord segments ................................................................. 5

Figure 2. The meninges of the spinal cord ..................................................................................... 7

Figure 3. The vascular supply of the spinal cord ......................................................................... 11

Figure 4. The blood-spinal cord barrier and neurovascular unit .................................................. 14

Figure 5. Causes of spinal cord injury ......................................................................................... 19

Figure 6. Schematic of secondary injury pathophysiology .......................................................... 28

Figure 7. Temporal progression of spinal cord injury pathophysiology ...................................... 29

Figure 8. Vascular development .................................................................................................. 55

Figure 9. VEGF gene and splice isoforms ................................................................................... 60

Figure 10. VEGF Receptors ....................................................................................................... 62

Figure 11. VEGFR-2 signaling ................................................................................................... 63

Figure 12. Hypoxic regulation of VEGF .................................................................................... 66

Figure 13. ZFP-VEGF technology .............................................................................................. 73

Figure 14. Model of spinal cord injury and intraspinal injections .............................................. 80

Figure 15. ZFP-VEGF expression cassette ................................................................................. 81

Figure 16. Schematic of immunohistochemistry quantification ................................................. 86

Figure 17. Femoral vein injections ............................................................................................. 87

Figure 18. Blood-spinal cord barrier permeability following traumatic SCI ............................ 107

xxiii

Figure 19. Spatial-temporal disruption of the spinal cord vasculature following clip-compression

injury .......................................................................................................................................... 110

Figure 20. Vascular disruption of the grey matter following traumatic SCI ............................ 115

Figure 21. Vascular disruption of the white matter following traumatic SCI .......................... 116

Figure 22. Endogenous angiogenic response after traumatic thoracic SCI .............................. 118

Figure 23. Transduction of AdV-eGFP into the spinal cord ..................................................... 137

Figure 24. Evaluation of AdV-ZFP-VEGF gene transfer ......................................................... 137

Figure 25. AdV-ZFP-VEGF increases VEGF mRNA and protein ........................................... 140

Figure 26. AdV-ZFP-VEGF administration reduces apoptosis after SCI . ............................... 143

Figure 27. AdV-ZFP-VEGF administration attenuated axonal degradation ............................ 144

Figure 28. AdV-ZFP-VEGF administration results in increased neuron sparing post-SCI ..... 147

Figure 29. AdV-ZFP-VEGF results in increased vessel counts ............................................... 149

Figure 30. AdV-ZFP-VEGF promotes angiogenesis at 5 days post-SCI ................................. 151

Figure 31. AdV-ZFP-VEGF improves hindlimb weight support ............................................. 168

Figure 32. Forelimb and Hindlimb locomotion is improved by AdV-ZFP-VEGF administration

..................................................................................................................................................... 170

Figure 33. AdV-ZFP-VEGF does not improve open-field walking (BBB) scores following SCI

..................................................................................................................................................... 171

Figure 34. Electrophysiological assessment following AdV-ZFP-VEGF administration ........ 174

Figure 35. AdV-ZFP-VEGF significantly reduces allodynia at 8 weeks post-SCI .................. 177

Figure 36. Tissue sparing quantification at 8 weeks post-SCI .................................................. 180

xxiv

Figure 37. 24 hour delayed AAV-ZFP-VEGF administration does not result in beneficial

outcomes following SCI ............................................................................................................ 196

xxv

List of Tables

Table 1. American Spinal Injury Association (ASIA) Impairment Scores ................................ 24

Table 2. Factors involved in angiogenesis .................................................................................. 53

Table 3. Antibodies used in Western Blot analysis .................................................................... 82

Table 4. Antibodies used in immunohistochemistry ................................................................... 85

Table 5. Animals used in Chapter 3 experiments ..................................................................... 105

Table 6. Spatial and temporal data from FITC-LEA and RECA-1 analysis ............................ 111

Table 7. Animals used in Chapter 4 experiments ..................................................................... 134

Table 8. Animals used in Chapter 5 experiments ..................................................................... 165

xxvi

List of Appendices

Appendix 1. Immediate Administration of AdV-ZFP-VEGF Following SCI .......................... 240

Appendix 2. A Novel Spinal Cord Window Chamber For In Vivo Imaging ............................ 250

Chapter 1

1 Introduction

1.1 Overview of Spinal Cord Injury

Spinal cord injury (SCI) – which can result from sudden trauma or progressive spinal cord

diseases – is a devastating event. Patients suffering from SCI experience significant functional

and sensory deficits, as well as emotional, social and financial burdens. Additionally,

individuals with spinal cord injury have an increased risk for other health conditions, including

cardiovascular complications, deep vein thrombosis, osteoporosis, pressure ulcers, autonomic

dysreflexia and the development of neuropathic pain [1]. In North America, it is estimated that

approximately 1.5 million individuals are currently living with SCI, with over 12,000 new cases

occurring each year [2].

By convention, spinal cord injuries can be divided into two events: a primary and a secondary

injury, which refer to the physical injury and the subsequent physiological cascade, respectively

[3]. It is well established that pathophysiological processes which occur in the secondary injury

phase are largely responsible for exacerbating the initial damage. These pathological processes

– inflammation, ischemia, lipid peroxidation, production of free radicals, disruption of ion

channels, necrosis and programmed cell death – are rapidly initiated in response to the primary

injury and, for the most part, are inhibitory towards endogenous regeneration and repair

mechanisms [4, 5]. The spatial and temporal dynamics of these secondary mediators are

fundamental to SCI pathophysiology and will be a discussed in more detail later.

1

Clinical evaluation and treatment of human patients is complicated by the heterogeneous nature

of SCI, which results in variable outcomes for each individual injury. In recognizing the drastic

variations of human injuries, the scientific community has been driven to develop and investigate

a variety of different animal models. More specific animal models may allow for better

translational success, since very few studies that have shown pre-clinical promise result in

similar efficacy upon translation into the clinic.

In the past decade, significant advances have been made which have notably contributed to

understanding the complexity of SCI pathophysiology. However, despite advances in pre-

hospital care, medical and surgical management and rehabilitation approaches, many patients

with acute SCI still experience substantial neurological disability. Intensive efforts are currently

underway to develop effective neuroprotective and neuroregenerative strategies. This chapter

aims to summarize the pathophysiological mechanisms initiated as a result of SCI, and in

addition, briefly highlight some potential molecular targets/processes which may benefit from

therapeutic intervention.

1.2 Anatomy of the Spinal Cord

1.2.1 Neuroanatomy of the Spinal Cord

The spinal cord is a long, thin, tubular bundle – consisting of nerves and neuroglia – that extends

42 – 48 centimeters (on average in humans) distally inside the vertebral column from the

brainstem, specifically the medulla oblongata [6]. The cord measures between 0.6 cm and 1.3

cm in width, with the broadest areas located at the cervical and lumbar enlargements. The spinal

2

cord performs three major functions: 1) Acts as a channel for transmitting descending motor

information, 2) Acts as a conduit for ascending sensory information, and 3) Serves as a central

hub for coordinating reflexes, via numerous central pattern generators [7].

The spinal cord is arranged into two distinct regions: white and grey matter. In cross-section,

the white matter is found laterally surrounding a central butterfly-shaped grey matter [8]. The

outer white matter contains myelinated axons, comprising sensory and motor tracts; whereas the

grey matter consists of motorneuron, interneuron, and sensory neuron cell bodies, which are

arranged into functional groups of cells called nuclei. The grey matter surrounds the central

canal, which carries cerebrospinal fluid (CSF) to the spinal cord as an anatomical extension of

the ventricles in the brain [8].

Within the spinal cord, there are a total of 31 segments, which act to synchronize all peripheral

sensory and motor functions (Figure 1). Similar to the vertebral column, the spinal cord is sub-

divided into segments – 8 cervical, 12 thoracic, 5 lumbar, 5 sacral and 1 coccygeal – however,

these do not directly correspond to the vertebral segments in the adult, especially below the

lumbar region (Figure 1) [6]. At each segmental level, left and right pairs of spinal roots, from

ventral (motor; exiting the cord) and dorsal (sensory; entering the cord) rootlets, combine to form

spinal nerves on both sides of the spinal cord. The spinal nerves are labeled according to the

level they emerge from the vertebral canal. C1-7 nerves emerge above their respective vertebrae,

and C8 emerges between the seventh cervical and first thoracic vertebrae. The subsequent

thoracic, lumbar and sacral nerves emerge below the level of their respective vertebrae,

eventually resulting in the cauda equina at the most caudal end of the spinal cord. The brachial

plexus innervates the upper limbs, and is controlled by sensory input and motor output in the

3

cervical enlargement, spanning C4 to T1 spinal segments [6]. The lumbar enlargement is an

analogous structure to the cervical enlargement, which is located between L1 and S3 spinal

segments. It is innervated by the lumbosacral plexus, which is responsible for movement and

sensation in the lower limbs [6].

4

Figure 1. The arrangement of the spinal cord segments. The spinal cord is divided into

cervical, thoracic, lumbar and sacral regions. Each segment is responsible for innervating

5

distinct areas of the body, with individual spinal nerves projecting to specific targets. Spinal

nerves exit the spinal column below their corresponding vertebral segment, with the exception of

the cervical nerves (since there are only 7 vertebrae and 8 nerves). The spinal cord ends at L1-

L2 where a dense network of lumbar and sacral spinal projections (called the cauda equina) fills

the spinal column. Anatomical enlargements in the cord occur in the cervical (C4-T1) and in the

lumbar (cord level: L1-S3; vertebral level: T9-T12), which correspond to an increased in the

number of motor neurons found in these areas. Figure drawn by Sarah A. Figley.

1.2.2 The Meninges

Between the vertebral column and the cord, the spinal cord is surrounded by fatty tissue,

lympathics, and a network of thin-walled vessels (creating the epidural venous plexus), which

comprise the epidural space [9]. Around the cord, spinal meninges form three protective layers

around the spinal cord (Figure 2) [6]. The outermost layer, the dura mater (literally meaning

“tough mother”), provides protection to the spinal cord and maintains its structure. The middle

layer, the arachnoid mater (named for its spiderweb-like appearance), provides a cushion by

retaining the CSF in the subarachnoid space (SAS). The deepest layer, the pia mater (meaning

“tender mother), forms a fragile layer of fibroblasts around the spinal cord that tightly follows

the contour of the spinal cord. The pia mater acts as the barrier between the CSF and the

vascular supply, directly associating with the glial limitans: a collection of astrocytic end feet

that provide a physical and immunological barrier into the CNS, and is a major component of the

blood-spinal cord barrier (BSCB) [10-12].

6

Figure 2. The meninges of the spinal cord. The spinal cord in surrounded by three protective

layers: pia, arachnoid, and dura mater. These layers provide structural support, as well as

protective cushioning to the spinal cord. The vasculature runs along the pia mater, branching

inward along the spinal axis at various levels. Figure drawn by Sarah A. Figley.

1.2.3 Vascular Organization and Blood Flow

The blood supply of the spinal cord is provided by three major arteries (2 posterior spinal

arteries, and 1 anterior spinal artery), and by a number of peripheral arteries

(medullary/segmental arteries) that feed the longitudinal spinal arteries at various segmental

levels (Figure 3) [6]. Posterior arteries supply the dorsal 1/3rd of the cord, while anterior arteries

supply the other 2/3rds of the spinal cord – namely the grey matter, and the ventral white matter.

The major spinal arteries are supplied by the aorta, which branches into vertebral arteries, and

eventually provide blood to the posterior and anterior spinal arteries. The spinal arteries are

located in the subarachnoid space, and branch into the cord along the rostrocaudal axis. The

7

blood flows caudally through these arteries (derived from the posterior cerebral circulation);

however, this blood supply is insufficient to maintain the vascular demands of the spinal cord

below the cervical regions, therefore, posterior and anterior radicular arteries directly enter the

spinal cord along the nerve roots, bypassing the major dorsal and ventral arteries. These

intercostal and lumbar radicular arteries, which arise from the aorta, provide large anastomoses

around the cord (termed the “vasocorona” – the connection of posterior and anterior blood

supplies) and supplement the blood flow to the spinal cord, particularly the lateral white matter

[13]. The medullary arteries are irregular and sporadically distributed along the spinal axis;

however, they are more prominent where blood supply demands are increased – namely the

cervical and lumbar enlargements. The grey matter of the spinal cord is highly vascularized

(approximately four-times more than the white matter), due to the metabolic demands of the

neuronal cell bodies. Therefore, grey matter is particularly susceptible to vascular and ischemic

injury [14].

Other areas of the spinal cord that are highly vulnerable to ischemia are watershed zones [15]. In

the spinal cord, watershed zones are located in the mid-thoracic region [16]. Watershed zones

are vascularized by two arterial supplies stemming from the most distal branches of large

vessels. These blood supplies do not overlap or connect, but rather provide blood to the same

areas via two distinct sources. The dual vascular supply protects these areas in the event of an

arterial blockage (presumably the other vessel is not occluded and will continue to deliver

blood); however, in the case of systemic hypotension, these areas experience hypoxia/ischemia

and are susceptible to neuronal damage.

8

The venous system of the spinal cord closely parallels the arterial supply, with capillaries and

veins directly adjacent to arterial vessels [17]. The spinal veins create and extensive network and

drain via anterior and posterior radicular veins, which lead to the internal (epidural space) and

external (vertebral surface) vertebral venous plexus via vertebral veins. The venous system of

the spinal cord has no valves; therefore, blood is free to enter the systemic venous circulation.

Ultimately, the blood of the spinal cord either i) flows superiorly inside the vertebral canal and

connects to the veins of the skull, or ii) enters the inferior vena cava and is transported directly to

the heart. The venous transport route depends on the spinal level.

9

10

Figure 3. The vascular supply of the spinal cord. A) Posterior and anterior views of the spinal

cord vascular network. Spinal arteries, which start at the base of the skull, run rostrocaudal

along the spinal cord. Segmental/medullary arteries, which branch from the aorta, provide

additional blood to the spinal cord since the spinal arteries are insufficient alone. B) Transverse

section of the vertebral column and the aortic vascular supply and venous drainage. The aorta

branches at various spinal levels to supplement blood flow to the cord, via segmental arteries.

Deoxygenated blood from the cord enters a local venous plexus, which eventually connects to

the systemic circulation. The area highlighted in yellow is expanded in detail in Panel “C”. C)

Vasculature of the spinal cord shown in cross-section (enlarged from Panel “B”). Posterior

arteries supply blood to the dorsal 1/3rd of the spinal cord, and anterior arteries are responsible

for providing blood to the central and ventral areas. The vasocorona is the connection between

posterior and anterior arteries and is predominantly responsible for vascularizing the lateral white

matter. Spinal veins are distributed in a similar arrangement to the arteries. Veins transport

deoxygenated blood to a number of spinal plexuses, eventually shunting the blood into the

systemic circulation via the inferior vena cava or cranial veins. Figure drawn by Sarah A.

Figley.

On an interesting note, the cardiac cycle and the arterial pulsations result in rhythmic anterior-

posterior oscillations of the spinal cord [18, 19]. Moreover, cerebrospinal fluid (CSF) flow is

also linked to persistent vascular pulsations and the cardiac cycle [20, 21].

11

1.2.4 Cerebrospinal Fluid

Cerebrospinal fluid (CSF) has two main functions: 1) to act as a protective cushion to the CNS,

and 2) to act as a pseudo lymphatic system. Daily, the human body produces approximately 500

mL of CSF, mainly via the choroid plexus of the lateral, third and fourth ventricles of the brain.

CSF flows from the third and fourth ventricles, posteriolaterally through the SAS of the spinal

cord and back up towards the brain ventrally. Flow is generated by arterial pulsations,

respiratory function and ependymal cilia movement [22-24]. In animals, there is CSF flow from

the SAS into the spinal cord parenchyma via perivascular spaces (PVS), and finally flow into the

central canal to integrate extracellular fluid (ECF) and CSF, ultimately transferring solutes

(electrolytes, proteins, etc.) [25]. Although this flow/exchange has not been confirmed in

humans, this may be a possible mechanism. The central canal itself is not associated with

longitudinal (rostral-caudal) flow of CSF as it is a non-continuous structure [26]. Water and

blood solutes are filtered by the choroid plexus, using various transporters and channels to

secrete water and the various electrolytes that comprise the CSF [27]. CSF is reabsorbed into the

venous system at the arachnoid granulations (by villi) in the sinuses of the brain [24].

1.2.5 The Blood-Spinal Cord and Cerebrospinal Fluid Barriers

1.2.5.1 The Cerebrospinal Fluid Barrier

As arteries enter the spinal cord from the SAS, they form arterioles which eventually branch to

form capillaries. These arterioles are surrounded by the pia and glia limitans; however, the pia is

eventually lost as the artery gets smaller and capillaries are formed. The space between the

12

blood vessels and neuronal/glial cells is filled with interstitial fluid (ISF), and is commonly

referred to as the Virchow-Robin space (VRS) [28]. The VRS is continuous with the

subarachnoid and sub-pial spaces, and is strictly limited to the arterial vasculature, as it is not

present around veins. The blood cerebrospinal fluid barrier (BCFB) is represented as the region

where the CSF is in direct contact with endothelial cells via the VRS.

The VRS plays an integral role in regulating fluid movement and drainage within the CNS [29].

Virchow-Robin spaces absorb fluids from neurons and glia and drain excess fluid into the

cervical lymph nodes. As described by the “tide hypothesis”, fluid flow between the VRS and

SAS is thought to be driven by the cardiac cycle [30]. Importantly, the VRS also functions as

part of the blood-brain barrier (BBB) and blood-spinal cord barrier (BSCB) and aids in

immunoregulation. [11, 31]. It is hypothesized that the VRS is where the flow of CSF and

extracellular fluid (ECF) occurs between the central canal, spinal cord parenchyma and SAS

[25].

1.2.5.2 The Blood-Spinal Cord Barrier

The blood-spinal cord barrier (BSCB) acts as a physical and biological barrier in the separation

of circulating blood and ECF within the CNS [11, 31]. Most importantly, the roles of the

BSCB/BBB are to: 1) protect the CNS from foreign pathogens, and drugs, 2) protect the CNS

from drastic hormone or neurotransmitter changes, and 3) maintain homeostasis for optimal

nervous system function and synaptic activity. Present only at the capillary level, the BSCB is

comprised of many components, including endothelial cells, astrocytic endfeet, pericytes and a

basement membrane, which collectively work to control the passage of fluid, ions, molecules and

13

cells (Figure 4). The BSCB is controlled by local interactions with neurons and pericytes,

forming what is known as the “neurovascular unit” (NVU) [32]. The coupling of these cellular

components is complex, and not completely understood. However, it is known that the NVU

responds to the energy and oxygen needs of local neurons – which directly innervate endothelial

cells – by making spatial and temporal adjustments to the blood supply [33].

Figure 4. The blood-spinal cord barrier and neurovascular unit. A) Schematic

representation of the cellular components of the BSCB/NVU. Endothelial cells are surrounded

by pericytes and astrocytes, as well as innervated by neurons. B) Homeostasis of larger

molecules is controlled by specialized transmembrane transporters located within endothelial

cells (glucose transporters, amino acid transporters, P-glycoprotein transporters (P-gp)). C)

14

Tight junctions and adherens junctions are present between endothelial cells, and at endothelial-

pericyte interfaces. These include, VE-cadherin, α/β-catenins, JAMs, claudins, ZOs, and

occludin. Image has been modified from [34-36] (Figure permissions requested).

1.2.5.2.1 Endothelial cells

Endothelial cells are responsible for the highly regulated selectivity of the BSCB. The

endothelial cells of CNS vessels are “linked” by tight junctions (TJ) and adherens junctions (AJ)

which restrict the flow of large hydrophilic solutes (i.e. bacteria, viruses, proteins), while

allowing diffusion of smaller hydrophobic molecules (i.e. oxygen, hormones) via paracellular

transport (transport between cells) (Figure 4) [31]. TJs occur closer to the apical surface,

whereas AJs occur toward the basal surface of endothelial cells [37]. Tight junction proteins

include occludins, claudins, junctional adhesion molecule (JAM). Adherence junction proteins

include cadherins (vascular endothelial: VE-cadherin) and catenins (α, β, or δ-catenin). Both TJ

and AJs are transmembrane proteins, which connect the actin cytoskeletons of adjacent cells

together via peripheral associated proteins (mainly by zonula occludens 1, 2 or 3 (ZO-1, ZO-2,

ZO-3)) [38, 39]. In addition to TJ and AJs, endothelial cells also contain a number of transporter

molecules – P-glycoprotein 1(P-gp), glucose transporters, amino-acid transporters – which

regulate transcellular flux (transport across the cell membrane).

1.2.5.2.2 Astrocytic Processes

Astrocytic endfeet (also known as the “glial limitans”) form an enclosure around perivascular

spaces (VRS) of CNS capillaries, and predominantly provide biochemical support rather than a

15

physical barrier [37]. Although these astrocytes act partially as a physical separation between

the immune-privileged CNS and the peripheral circulation, they are connected by gap junctions

allowing molecules to freely pass through. The high expression of aquaporins and potassium

channels within the astrocytic endfeet of the BSCB/BBB are involved in regulating volume,

hormone levels, and ion concentration in the CNS, ultimately maintaining homeostasis of the

perivascular environment [40]. Additionally, astrocytes are critical for proper neuronal function

and are therefore an essential component of the NVU. Moreover, emerging evidence suggests

that these glial cells are a key participant in coordinating neurovascular coupling and can, in fact,

signal the vascular smooth muscle cells of blood vessels to regulate blood flow in response to

rapid changes in local neuronal activity [41]. Finally, deficiencies in neuron-astrocyte

interactions have been linked to the development of neurodegenerative diseases and improper

BSCB/BBB function [42].

1.2.5.2.3 The Basement Membrane

Structural proteins (elastin and collagen), specialized proteins (laminin and fibronectin) and a

number of proteoglycans are integrated into protein layers, which form the extracellular matrix

(ECM) around the CNS microvessels [43]. This ECM is called the basement membrane (or

basal lamina) is considered an essential component of the BSCB and is located around

endothelial cells and pericytes. Physically it does not prevent the diffusion of smaller

molecules; however, the presence of the basement membrane provides structural support to the

microvasculature, ultimately stabilizing the BSCB [44]. The role of the basement membrane has

also been shown to regulate the proliferation and differentiation of BSCB cells (such as

pericytes), therefore, keeping the barrier in a established, mature form [44].

16

1.2.5.2.4 Pericytes

As previously mentioned, pericytes are physically separated from endothelial cells by the basal

lamina [45]. Pericytes, which are small vessel wall-associated cells, communicate with

endothelial cells via gap junctions and soluble factors. Until fairly recently, the role of pericytes

remained largely unknown; however, recent research shows that pericytes are a vital cell for

BSCB development, maintenance and functional regulation [45]. It has been shown that

pericytes regulate endothelial cell function in the BSCB by promoting the formation of tight

junctions, and controlling vesicle trafficking. Since pericytes are contractile cells, they also play

a key role in blood flow and molecular influx into the CNS tissue [33, 45]. Regulating influx, is

an important function of pericytes as it prevents certain molecules or cells (i.e. large plasma

proteins or inflammatory mediators) from entering the CNS, therefore maintaining homeostasis

for the surrounding CNS cells. Moreover, pericytes may also act to inhibit immune cells, reduce

vascular permeability, phagocytose cells and debris, and promote angiogenesis [46, 47]. Finally,

pericyte function/dysfunction has been linked to the onset and progression of neurodegenerative

diseases, further indicating that they play an essential role in maintaining and regulating the

BSCB/BBB in vivo [33].

1.2.5.2.5 Differences Between BBB and BSCB

Generally, it was assumed that the BSCB was an anatomical and functional extension of the

BBB. While this is remains partially true, recent research has elucidated a number of structural

differences between the two barrier systems, which may account for their functionality and roles

in pathological diseases [48]. Most notably, it appears that the BSCB may be a more permeable,

17

less selective barrier: as indicated by decreased TJ and AJ proteins, decreased transporter

molecules, and increased permeability to vascular tracers [31, 48].

1.3 Epidemiology of Spinal Cord Injury

The spinal cord is normally protected by the vertebral column; however, physical disruption

from trauma often alters the position and structural integrity of the vertebra, leaving the spinal

cord vulnerable. Traumatic forces, such as the ones sustained in sports injuries, traffic accidents,

and diving into shallow water, cause vertebral column complications associated with SCI (Figure

5) [2]. Although blunt injury is the predominant cause of SCI, other penetrating trauma due to

knife or gunshot wounds occur in a significant percentage of cases [49]. Approximately 55% of

SCIs occur at the cervical level (C1 to C7-T1), and thoracic (T1 to T11), thoracolumbar (T11-

T12 to L1-L2) and lumbosacral (L2 to S5) injuries each account for approximately 15% of SCI

[2].

18

Figure 5. Causes of spinal cord injury. Graphical representation of the cause and prevalence

of SCI in Canada. Chart was created based on 2006 data from the Rick Hansen Foundation

(Rick Hansen Spinal Cord Injury Registry) [50].

SCI can result from shearing, stretching, laceration and, in very rare cases, transection of the

cord; however, the most common injuries are the result of contusive and compressive forces [2].

Worldwide, SCI occurs with an estimated annual incidence of 15–40 cases per million. Within

the USA, new studies suggest that 1.275 million individuals are currently living with SCI – an

estimate which is drastically larger than previous estimates of 280,000 individuals – and over

19

10,000 new injuries occur annually [51, 52]. Within Canada, it is estimated that 40,000

individuals are currently living with SCI, with approximately 1,000 new injuries occurring each

year. Depending on the age of the patient, severity, and level of SCI, the lifetime cost of health

care and other injury-related expenses ranges from $1.25 million to $25 million [53].

Perhaps one of the most devastating statistics of SCI is that it predominantly occurs in young,

healthy individuals – mostly between 15 and 34 years of age – and statistically, males are almost

four-times more likely to incur a spinal injury compared to females [54]. In the past 30 years

there has been a slight shift in SCI demographics, where the average age of injury has increased

from 28.7 to 37.6 years and elderly individuals (those > 60 years of age) now account for 10% of

SCI cases (an increase from 4.7%) [55]. Traumatic SCI can result from a variety of different

causes; however, the most common causes of SCI are motor vehicle accidents, falls, work-

related injuries and sports-related injuries (Figure 5).

1.4 Primary and Secondary Injury

By convention, spinal cord injuries can be divided into two events: a primary injury and a

secondary injury, which refer to the physical injury and the subsequent physiological cascade,

respectively [3]. Research has shown that secondary injury is responsible for a large portion of

damage and degeneration that is associated with SCI, including inflammation, ischemia, lipid

peroxidation, production of free radicals, disruption of ion channels, necrosis and programmed

cell death [5, 56, 57]. Additionally, drastic changes occur in the spinal microvascular structure

and function following SCI, including reduction in blood flow, hemorrhage, systemic

hypotension, loss of microcirculation, disruption of the blood-spinal cord barrier (BSCB) and

20

loss of structural organization [3, 58]. While these secondary events are responsible for the

majority of the extensive damage that results following injury, these pathways also present ideal

target environments for therapeutic intervention.

In mammals, a cystic cavity forms at the injury epicenter and spreads mediolaterally and

rostrocaudally from the injury site over time, resulting in substantial functional and

morphological alteration. Infiltrating inflammatory cells are present within the cavity, along with

myelin debris and axons in various degrees of demyelination [59, 60]. Typically, a subpial rim

of tissue survives the injury and contains axons also in varying states of myelination [61, 62].

Astrocytes proliferate and surround the cavity in an attempt to attenuate the spread of the lesion,

forming a glial scar [63, 64]; however, this astrogliosis also produces a physical and chemical

barrier which is inhibitory to axonal regeneration. A fibrous scar consisting of collagen and

various inhibitory extracellular matrix (ECM) molecules is deposited within and surrounding the

lesion. Wallarian degeneration of axons towards their cell bodies and away from the epicenter is

a common fate of severed axons [65]. Severed axonal ends distal to the injury site degenerate

along with disrupted myelin and are broken down, eventually being phagocytosed by

macrophages. A chronic snapshot of the injury demonstrates a cystic cavity containing

vascular/glial bundles, regenerated nerve roots, collagenous fibers and astrocytes [66].

1.5 Animal Models of Spinal Cord Injury

Many small animal models have been developed to investigate the pathophysiology and

functional recovery of SCI; however, due to the heterogeneity of the human condition, no one

21

model precisely mimics the complete pathobiological spectrum. Animal models – usually rats,

mice or other small mammals – include spinal cord compression by forceps or modified

aneurysm clips, balloon compression, weight drop devices/contusion, hemi or full transection

injuries, and chemically-induced SCI [67]. It has been shown in rats that neurological

impairment increases relative to both the time of spinal compression and the force of trauma

[68]. Similarly, it has been suggested that reducing the time of spinal compression in human

patients, by early surgical intervention and spinal decompression, may result in better functional

outcomes for patients [69]. Cells of the CNS, particularly neurons and their axons, are prone to

shearing and compressive forces and physical damage results in rapid cell dysfunction and death

[70]. Although individual animal models fail to address all of the issues present in human SCI,

each of the models exhibit at least one of the pathophysiological consequences. Rat

contusion/compression models accurately mimic vascular damage and disruption to the blood –

spinal cord barrier (BSCB), cyst or cavity formation, neuronal loss and demyelination, and

functional and sensory loss [71, 72].

Although not a perfect design, clip compression models offer a number of advantages over other

models in more closely representing the human condition. In contrast to forcep-crush models,

the clip compression models provide a more consistent, and reliable application of force,

resulting in a more reproducible and constant model of injury. In comparison to weight drop

models, clip compression injuries afford two distinct benefits: i) clip injuries result in anterior

and posterior spinal cord damage (where weight drop models primarily injure the dorsal surface),

and ii) clip compression injuries restrict blood flow and create an ischemic environment, which

is a key component of SCI in human injuries. While, hemi or full transection models are ideal

for regenerative studies, these models lack a strong clinical significance due to a lack of physical

22

trauma, inflammatory response, vascular disruption and scar formation. Lastly, chemically-

induced SCI models have validity and are appropriate for specific use; however, again in contrast

to clip compression models (and similarly the human condition), chemical-induced SCI lacks

physical trauma (mechanical disruption) to the tissues, which in turn, results in a diminished

pathophysiological response. For the reasons mentioned above, our laboratory has utilized the

clip compression model of SCI, as we believe it accurately mimics human SCI pathology [73-

76].

1.6 Neurological Function Following Spinal Cord Injury

1.6.1 Motor Function

A devastating consequence of spinal cord injury is paralysis due to damaged axons and neurons

in motor pathways at or above the level of injury. Various animal models have been developed to

attempt to mimic the motor deficits seen in clinical cases of SCI. Thoracic injuries are well-

characterized in their functional deficits, and result in significant paralysis of the hindlimbs.

Cervical models of SCI – although not as well-characterized as thoracic models – can display

decreases in respiratory function and forelimb deficits depending on location and severity of

injury. Clinically, spinal cord injuries are divided into one of two general categories: complete

or incomplete, referring to total loss of motor and sensory function below the injury site, or

partial motor and/or sensory loss, respectively. The American Spinal Injury Association (ASIA)

has developed a much more descriptive evaluation, called the ASIA Impairment Scale, to more

accurately assess motor and sensory deficits in humans (Table 1) [77].

23

Table 1. American Spinal Injury Association (ASIA) Impairment Scores [78].

Classification Definition

A Complete: No motor of sensory function is preserved in the sacral segments S4-

S5.

B Incomplete: Sensory but not motor function is preserved below the neurological

level and includes the sacral segments S4-S5.

C Incomplete: Motor function is preserved below the neurological level. More than

half of the key muscles below the neurological level have a muscle grade < 3.

D Incomplete: Motor function is preserved below the neurological level. At least

half of the key muscles below the neurological level have a muscle grade > 3.

E Normal: Both motor and sensory function are normal.

Motor impairment following SCI is due to both upper and lower motor neuron damage. Loss of

lower motor neurons in the anterior/ventral horn results in paralysis of muscles at the level of

injury. Axons from upper motor neurons which would normally pass through the injury site

(such as ones from the corticospinal tract) are also damaged, and as a result, efferent input to

muscles below the level of injury is also impaired [6]. Typically, a sub-pial rim of upper motor

axons survive the injury, however they are left in varying states of demyelination [74]. These

axons transverse the lesion site and can supply somewhat limited motor information to muscles

below – the quality of which is dependent on the number and myelination state of the axons.

Since a significant proportion of all injuries occur in the cervical level (approximately 50%),

patients experience loss of control to muscles of the upper limbs and diaphragm (in high cervical

24

injuries), and most patients experience some functional deficit in their trunk and lower limbs

[54].

1.6.2 Sensory Function

Sensory information pertaining to pain and temperature is collected from specialized receptors in

the periphery and ascends via the spinothalamic tract to the brain where it is processed [6]. First

order neurons entering the CNS must ascend or descend one or two vertebral levels before

synapsing on second order sensory neurons in the dorsal/posterior horn, which then decussate

and continue rostrally to the brain. Conversely, first order neurons found in the posterior

column-medial lemniscus pathway, which transmit fine touch and vibration information, enter

the spinal cord and travel rostrally towards the brain prior to decussating in the medulla. Injury

to first and second order spinothalamic neurons, or first order neurons from the medial lemniscus

pathway, interrupts sensory information processing at and below the level of injury and prevents

normal signal transmission to the brain. Miscommunication in sensory pathways can result in

severe complications for patients suffering from SCI. Development of neuropathic pain occurs

in many patients, and although the exact mechanism is unknown, it is hypothesized that it caused

by misguided axonal sprouting or abnormal sodium channel excitability in sensory neurons [79].

Neuropathic pain will be discussed in more detail in subsequent paragraphs.

1.6.3 Autonomic Function

As described above, SCI results in motor or sensory loss as a result of disrupted neural pathways.

Unfortunately, damaged neural circuits exhibit more than motor and sensory dysfunction. SCI

results in miscommunication between higher centers of the hypothalamus and limbic system, as

25

well as the various effector organs of the autonomic nervous system [80]. Preganglionic cell

bodies of the sympathetic system are distributed in the intermediolateral horn of the gray matter

between T1 and L2, whereas preganglionic cell bodies of the parasympathetic system are

situated in the brain stem and sacral levels of the spinal cord. Similar to motor and sensory

deficits, the autonomic dysfunction is dependent upon the level of injury. Regulation of cardiac

output, vascular tone, and respiration is controlled between T1-T4, therefore cervical injuries

often exhibit respiratory or cardiovascular complications [6]. Gastrointestinal and other organs,

including sexual organs, are controlled in lower segments of the spinal cord: T5-L2. In the

uninjured spinal cord, spinal sympathetic interneurons remain rather inactive; however,

following SCI they are activated and are thought to contribute to the impaired autonomic

function. A common autonomic condition in patients with injuries at or above T6 is autonomic

dysreflexia [81]. This condition is caused by excessive afferent stimuli, most frequently from

paroxysmal hypertension (spikes in blood pressure, often reading over 200 mmHg) as a result of

bladder or bowel distension. These episodic increases in blood pressure manifests into

undesirable side-effects, such as headaches, blurred vision, anxiety, and severe sweating [82].

1.7 Secondary Injury

Many factors dictate the patholophysiology of SCI, including the force, severity and location of

the initial injury. Ultimately, these variables are responsible for the extent of molecular damage,

tissue loss and functional deficit. The primary injury, caused by mechanical force, stimulates a

complex series of systemic, cellular and molecular cascades that expand the lesion from the

initial injury into surrounding white and grey matter, consequently increasing the extent of tissue

loss. Secondary injury is primarily defined by detrimental pathophysiological responses;

26

however, endogenous repair and regenerative mechanisms are employed during the secondary

phase of injury in an attempt to minimize the extent of the lesion and reunite damaged neural

circuits.

1.7.1 Summary of Secondary Injury Progression

Mammals such as rats and humans (but not mice) develop a fluid filled cystic cavity at the injury

epicenter following traumatic SCI. Over time, this cavity displaces cellular structures as it

expands in a rostrocaudal manner from the epicenter, producing significant functional and

structural alterations. Infiltrating macrophages, lymphocytes, and activated microglia are present

within the cavity, along with granular myelin debris and axons in varying states of demyelination

As mentioned earlier, a sub-pial rim of tissue often survives the injury and consists of axons

which exist in various myelinated states [61]. In an endogenous effort to restrict the progression

of the cystic cavity, astrocytes are recruited and proliferate; a process termed astrogliosis [63].

These astrocytes eventually surround the injury cavity, and in addition to their physical barrier,

they express inhibitory molecules which results in a chemical barrier hindering future attempts of

axonal regeneration. Damaged axons undergo Wallerian degeneration towards their cell bodies

and away from the epicenter, whereas severed axonal ends distal to the injury site degenerate

along with disrupted myelin and are eventually phagocytosed by macrophages [65]. The chronic

phase of SCI is characterized by the presence of a cystic cavity containing vascular/glial bundles,

regenerated nerve roots, collagenous fibers and astrocytes [66]. Refer to Figure 6 and Figure 7,

which provide an overview of pathophysiological events initiated by traumatic SCI.

27

Figure 6. Schematic of secondary injury pathophysiology. Representation and connection

between the physiological processes that are initiated following spinal cord trauma. Figure was

28

created by Sarah A. Figley, using information from multiple sources [2, 83, 84]. Figure appears

in The Cervical Spine: 5th Edition (2012) [85] (Figure permission requested).

Figure 7. Temporal progression of spinal cord injury pathophysiology. Chart shows the

time-course of cellular and systemic events following traumatic spinal cord injury in humans.

Figure has been modified from Rowland et al. [86]. Figure created by Sarah A. Figley, and

appears in The Cervical Spine: 5th Edition (2012) [85] (Figure permission requested).

1.7.2 The Acutely Injured Spinal Cord

Primary physical damage and death of neural cells onsets the acute secondary phase of SCI.

Typically, the acute phase represents the first 24-48 hours following injury [84, 86]. This phase

29

is characterized by vascular dysfunction, energy and ion imbalances, excitotoxicity, and early

inflammatory events that lead to necrotic and to a lesser extent apoptotic cell death.

Almost immediately following injury, edema and hemorrhage, which correlate to injury severity,

result in ischemic zones and produce necrotic cell death [3, 5]. Microglia, responding to by-

products of necrosis (such as DNA, ATP, K+), become activated and secrete inflammatory

cytokines that actively recruit systemic inflammatory cells.

The immediate acute injury, is defined as the initial 2 hours post injury [86]. In this phase,

spinal shock is initiated below the level of injury and function is immediately lost [87].

Additionally, during this time neurons and glia that have survived but previously sustained

damage from the initial injury, hang in the balance of survival and necrotic cell death. Typically,

the gross histology of the acutely injured spinal cord has not been significantly altered and may

appear normal with MR imaging [88].

1.7.2.1 Vascular Damage Following SCI

Following injury, drastic changes are observed in the spinal microvascular structure and function

[83, 89], and histological studies have shown that the areas with significant vascular damage

coincide with the areas of severe neuronal loss [71]. Since the grey matter of the spinal cord

contains approximately four times more vessels than the white matter, the grey matter is

particularly susceptible to vascular injury [90]. Vascular changes include reduction in blood

flow, hemorrhage, systemic hypotension, loss of microcirculation, disruption of the blood-spinal

cord barrier (BSCB) and loss of structural organization [3, 91]. Vascular disruption, particularly

30

vasospasm, impaired autoregulation and loss of microcirculation, are observed almost

immediately following injury and have been shown to significantly contribute to the ischemic

pathology. Vasospasm occurring after SCI can be initiated by the injury itself or by the release

of vasoactive factors (such as nitric oxide or histamine) [92]. Ischemia, resulting from

disturbances to the spinal cord blood flow (SCBF) is apparent in almost all cases of human and

animal models of SCI and it has been shown that there is a correlation between injury severity

and SCBF [5, 93]. Following injury, fragile microvascular networks are prone to ischemia since

local spinal cord blood pressure is diminished and is further exacerbated by systemic

hypotension [94, 95].

1.7.2.1.1 Blood-Spinal Cord Barrier Disruption

The blood-spinal cord barrier (BSCB) serves as a semi-permeable barrier, separating the

systemic blood and the CSF, and maintaining homeostasis [31]. The function of this barrier is to

regulate the transport of molecules and cells in and out of the CNS – protecting the nervous

system from toxins, viruses and bacteria. Structurally, astrocytic end-feet and pericytes surround

the endothelial cells of vessels. Tight-junctions between endothelial cells and transmembrane

transport proteins (found within endothelial cells) are the key regulators of molecular influx [96].

Vascular disruption is a hallmark of SCI [66]. Mechanical damage from the primary injury leads

to vessel rupture and hemorrhaging, which eventually subsides as bleeding is controlled by

homeostatic responses. However, degradation of endothelial tight junction proteins,

disappearance of astrocytic end feet – due to astrocyte cell death – and cytokine effects on

endothelial cells cause further BSCB compromise. The resultant ‘leaky vessels’ allow the

31

passage of cellular and molecular inflammatory mediators from the blood into the spinal cord

parenchyma, propagating the initial damage and contributing to the secondary injury and

subsequent pathophysiology [14, 66]. Importantly, the compromised BSCB also contributes to

the formation or progression of edema following trauma. BSCB permeability post-injury peaks

at 24 hours; however, it remains compromised long after the initial mechanical damage. Studies

have shown that the BSCB remains dysfunctional and disorganized for approximately 2 weeks

following injury, although a recent study by Popovich et al. implies that BSCB disruption

following SCI may be a biphasic occurrence, showing increased permeability again at 28 days

post-injury [66, 97]. Inflammatory cytokines (such as IL-1beta and TNF-alpha), and other

signaling molecules (ROS, NO, and histamine) are known to contribute to this prolonged

permeability, whereas angiogenic factors (such as VEGF and matrix metalloproteinases

(MMPs)) are involved in vascular remodeling and BSCB repair [98, 99].

1.7.2.2 Energy, Ion and Glutamate Imbalances

Dysfunction in metabolic homeostasis, involving imbalances in Na+, K+, Ca2+, and glutamate, is

well documented and causes impairment and cell death post-injury. Following injury, axonal

concentrations of Na+ and Ca2+ are significantly increased as a result of ion pump failure,

inactivation of ion channels, reverse function of ion exchangers and depolarization of

membranes [100, 101]. Astrocytes and oligodendroctyes also experience increased intracellular

levels of Ca2+ following SCI. Through L-type and N-type calcium channels, along with excess

glutamate signaling (via metabotropic and ionotropic glutamate receptors), elevated

concentrations of Ca2+ may contribute to exacerbation of white matter injury [102, 103].

Impaired glutamate reuptake by astrocytes is a consequence of glutamate transporter

32

dysfunction, cell death, and glutamate release from glia, neurons and axons via reversal of Na+-

dependent glutamate transport, whereby each event contributes to the excessive extracellular

glutamate [102]. As early as 3 hours post-SCI, increases in extracellular glutamate is observed,

leading to alterations in glial and axonal function and gray matter neuronal cell death [104, 105].

Changes in acute energy metabolism following spinal cord injury is characterized by diminution

of ATP, decreased glucose and increased lactate/pyruvate ratios (indicative of hypoxia) [106].

The ensuing deficits in energy metabolism are undoubtedly due to hypoperfusion/ischemia

mediated decreases in oxygen and loss of glucose availability to cells.

1.7.2.2.1 Intracellular Consequences of Acute Excessive Calcium Concentrations

Unnecessary accumulation of intracellular Ca2+ results in axonal degradation and neuronal cell

death by activation of protein kinases, proteases, and mitochondrial dysfunction. Calpains, a

class of Ca2+ dependent proteases, are triggered acutely following SCI due to accrual of

intracellular Ca2+. Calpains are known to degrade cytoskeletal proteins, such as neurofilaments

and microtubules, which notably disrupts axonal integrity and function [107, 108].

Additionally, excessive intracellular calcium levels are detrimental to mitochondria, since

increased Ca2+ stimulates the production of reactive oxygen species (ROS) in neurons and glia.

The production of ROS results in oxidative stress and intensifies neural damage.

1.7.2.3 Oxidative Stress

As previously mentioned, an increased production of ROS occurs following SCI due to

metabolic imbalances and excess intracellular Ca2+, leading to mitochondrial dysfunction and the

33

production of unchecked ROS [109]. Experimental methods have determined that peak ROS

production occurs 12 hours following injury, with levels remaining elevated until 4-5 weeks

post-injury, at which point they appear to return to normal levels [110, 111]. It has been well

characterized that ROS initiate necrotic cell death; however, studies have found that brief

oxidative stress can cause apoptosis of oligodendrocytes and neurons [112]. ROS produced by

mitochondria include superoxide (O2-) and hydrogen peroxide (H2O2), and if left unneutralized,

O2- freely reacts with nitric oxide (NO) to produce peroxinitrite (-ONOO), one of the most

reactive and detrimental free radicals known. When generation of ROS increases above the anti-

oxidative capacities of cells – as is the case of mitochondrial dysfunction – these reactive

molecules can damage proteins, DNA, and lipids. Neutrophil infiltration and rupture has also

been recognized as a detrimental source of ROS post-injury through their respiratory (oxidative)

burst [113].

1.7.2.4 Inflammation

Inflammation is a mechanism through which circulating leukocytes (neutrophils, monocytes/

macrophages, T cells), soluble factors (cytokines, chemokines, complement, lipid byproducts),

and resident microglia attempt to restore tissue homeostasis. The inflammatory response initiated

following SCI involves a complex interaction between systemic and local factors. Inflammation

post-injury has demonstrated both beneficial and detrimental aspects, including removal of

cellular debris and propagation of secondary damage, respectively. Profound differences in

inflammatory responses of animal models have been observed between species (even between

individual strains of rats). In line with this, inflammatory responses in humans also differ from

34

observations from disease models, rendering direct comparisons between the two difficult [111,

114].

Within hours of SCI, activated microglia are recruited due to local vascular disruption, loss of

tissue homeostasis and necrotic by-products (ATP, DNA, extracellular K+). In the process of

activation, microglia undergo morphological conversion from ramified to amoeboid and

subsequently release cytokines (TNF-alpha, IFN-gamma, IL-6, IL-1beta) and nitric oxide (NO).

These cytokines are potent molecules which recruit systemic inflammatory cells, modulate local

protein expression, and result in neurotoxicity and myelin damage [115, 116]. Neutrophils are

the first systemic immune cells to respond to the site of injury, and although they are initially

observed within hours after injury, it is not until 24-48 hours following injury that they are most

abundant [113, 117-119]. Neutrophils also secrete matrix metalloproteinases (MMPs) which,

upon activation, can cleave endothelial tight junction proteins, leading to increased BSCB

permeability. Additionally, neutrophils are a source of myeloperoxidase, which is implicated as

an early cause of ROS lipid peroxidation and the subsequent weakening of cell membranes

[119]. Blocking neutrophil extravasation following experimental SCI has produced both

detrimental and beneficial results, thus it is still debated whether the presence of neutrophils is

good or bad following SCI [120, 121].

1.7.2.5 Cell Death: Necrosis, Apoptosis and Oncosis

Cell death in the broader sense can be looked at as a continuum. Whereas exclusively necrotic or

apoptotic cell death can be observed from certain stimuli, studies suggest that the intensity of the

cellular insult can dictate the ensuing phenotype [122]. Necrotic cell death requires no energy

35

and typically results in intracellular contents being released into the ECM causing an

inflammatory reaction. Apoptosis, on the other hand, requires energy and results in the

formation of apoptotic bodies (which contain intracellular contents) that are phagocytosed

without exacerbating inflammation. Oncosis is another form of cell death, whereby the cell or

cellular components (usually the mitrochondria or nucleus) swell as a result of ionic pump

malfunction; eventually bursting the cell [123]. Oncosis is most commonly initiated by ischemic

insults, and usually takes 24 hours to onset, whereas apoptosis can be initiated much quicker

(within minutes). Also, a distinct difference between oncosis and apoptosis that oncosis typically

provokes an inflammatory response surrounding the dying cells, while inflammation is minimal

following apoptosis [123]. Unfortunately, limited evidence exists to support apoptosis in

neurons following human SCI, even though apoptosis is clearly evident acutely in many animal

models [124, 125]. Moreover, no limited evidence of oncosis following SCI has been

demonstrated; however, it has been noted in stroke, liver failure, and cardiac insult [126]. In

most cases, acute cell death resulting from SCI is necrotic, whereas delayed cell death is

predominantly driven by apoptotic pathways. Necrosis can result from physical membrane

damage, chemical membrane damage (lipid peroxidation), and intracellular ROS excess due to

ion imbalances and energy depletion. [72, 127].

1.7.2.6 Demyelination

Although all cells are vulnerable under hypoxic-ischemic conditions, oligodendrocytes exhibit a

particular susceptibility to low blood/oxygen environments [128-130]. As such, damaged or

affected oligodendrocytes die early on following SCI, resulting in axonal demyelination.

Endogenous efforts for remyelination have been observed; however, the spinal cord lacks a

36

sufficient population of oligodendrocyte precursors, therefore the majority of lost cells are not

replaced and permanent neurological deficits remain [131].

Animal models of SCI have shown that surviving axons are left in varied stages of

demyelination, which results in aberrant signal conduction [132]. This pathophysiological

phenomenon contributes to part of the overall functional deficit following injury. These

surviving axons are an attractive therapeutic target, as increasing their axonal conductance

through remyelination could restore function. Many studies have employed the use of stem cells

to promote remyelination of axons, although it is now known that stem cells can have other

beneficial effects, not related to their myelinating properties, such as neurotrophic factor

secretion, which may account for their effective therapeutic use [133, 134]. Although

demyelination is a hallmark of neuronal dysfunction in animal models, studies describing SCI in

clinical patients have suggested that demyelination might not be as prominent in humans.

1.7.3 The Sub-Acutely Injured Spinal Cord

The sub-acute phase of SCI extends between two days to two weeks post-injury in rodent models

(Figure 7). Conversely, it has been suggested that the sub-acute phase in human SCI occurs in a

much more delayed fashion, occurring between two weeks and 6 months. This phase is

characterized by a plethora of molecular events, including continued inflammatory response,

reactive astrogliosis, remodeling of the extracellular matrix (ECM), delayed cell death and

progressive axonal demyelination/degeneration. In parallel with the aforementioned processes –

which encompass the detrimental effects of SCI – endogenous rescue and repair mechanisms are

initiated in the sub-acute phase of injury. Endogenous progenitor cell proliferation, removal of

37

cellular debris, angiogenesis, and control of the cavity size by astrocytes are all processes

designed to minimize the damage from the injury.

1.7.3.1 Inflammation

Following the initial inflammatory response by neutrophils and microglia, which occurs almost

immediately post-injury, blood derived monocytes/macrophages are recruited within 2-3 days

and can remain activated for several weeks [135]. Upon activation, macrophages become almost

indistinguishable from resident microglia, as they assume a similar morphology and cytokine

expression profile. It is still unclear whether the presence of microglia in the subacute phase of

injury is advantageous or deleterious to the progression of SCI pathophysiology [136, 137].

Rather than manipulating the presence of microglia/macrophages, the most promising

therapeutic avenue towards creating a more permissive post injury inflammatory landscape

revolves around modulating the temporal gene expression of these cells. It has been reported that

these cells are capable of adopting an M1 (inflammatory) or M2 (beneficial) phenotype [136].

Although detrimental in the initial stages of activation, microglia contribute to the secretion of

growth factors and neurotrophins and are able to clear the injury site of dead tissue and cellular

debris by phagocytosis, and these characteristics make them and integral part of wound healing

regeneration [138, 139].

In addition to neutrophils and microglia/macrophages, other inflammatory cells are also recruited

to the site of injury. The presence of T-lymphocytes in the spinal cord is maximally observed

between 3 and 7 days following injury. T-lymphocytes are employed in response to the

cytokine/chemokine signals produced by activated microglia and macrophages [140]. By

38

dictating the secretion of pro- and anti- inflammatory cytokines, T-lymphocytes are able to direct

macrophage/microglial activity. Moreover, antigen-independent T-cells can be recruited to the

site of injury – through cytokine signaling from CNS-specific T-cells – and these cells secrete

various trophic factors (such as IGF-1 and BDNF) which are important for future regeneration

and growth [141, 142].

1.7.3.2 Sub-Acute Cell Death and Axonal Degeneration

A multitude of extracellular and intracellular events are able to initiate apoptosis in the subacute

phase of SCI, including removal of trophic factors, increases in inflammatory mediators, death

receptor activation, and DNA damage [128]. The execution of apoptosis is highly varied

depending on the cell type and the type of cell death signal it receives.

Clinically relevant animal models of SCI have identified neuron and oligodendrocyte apoptosis

as a significant event in the injury pathophysiology through the use of TUNEL and caspase-3

staining [143, 144]. Activation of caspase-3 and caspase-8 temporally coincide to apoptosis

after SCI [145]. Caspase-3 activation, in both neurons and oligodendrocytes, has been observed

as early as 4 hours and up to 8 days after experimental SCI. Studies have also suggested that

oligodendrocyte apoptosis and axonal demyelination are mechanistically linked [72]. The spatial

distribution of caspase-3 activation has been observed in the injury epicenter as well as in the

surrounding penumbra. Following SCI, there is also an increase in cytochrome c, which may be

responsible for activation of cell death pathways [144]. This increase is observed in neuron

within several hours following SCI, whereas it is not observed until several days in

oligodendrocytes.

39

1.7.3.3 The Glial Scar

The severity and type of injury dictate the magnitude of glial scarring observed following SCI.

Transection injury models produce drastically different scarring patterns compared to contusion

or compression injuries [146]. In rats and humans, surviving astrocytes respond to the site of

injury where they proliferate and become activated, eventually surrounding the cystic cavity to

prevent it from expanding. As mentioned earlier, the phenomenon of astrocytes forming a

“heteromorphic network” is commonly referred to as astrogliosis or glial scarring [147].

Although the formation of the glial scar limits the exacerbation of damage and cavity formation,

the presence of astrocytes creates a chemical and physical barrier that is inhibitory for

endogenous or therapeutically initiated axonal regeneration. Astrocytes express and secrete

chondroitin sulfate proteoglycans (CSPGs) and other inhibitory molecules which result in growth

cone collapse and dystrophic end bulb formation in neurons [148]. Astrogliosis is prominent in

rodent models of SCI; however, it is not as pronounced in human SCI [149]. Therefore, SCI

therapies that specifically target glial scarring, such as ChABC, may show promising results in

animal models but could have limited results when translated into clinical trials [150].

Spinal cord injury results in profound changes in the ECM. The fibrous scar is predominantly

composed of collagen IV. Although collagen IV is not inhibitory itself, but is described as ‘sticky’

and is known to binds other ECM molecules [146]. Collagen IV and laminin expression are

upregulated following injury and can be associated with scar formation in rats and humans along

with fibronectin [151, 152]. In rats, laminin expression remains upregulated into chronic phases of

the injury, whereas collagen IV decreases chronically, but it does not return to basal levels.

Inhibition of neurite outgrowth is a result of CSPGs (such as NG2), tenascin, myelin associated

40

glycoprotein (MAG), oligodendrocyte myelin glycoprotein (OMgp), brevican, versican, and Nogo

A, B & C expression in the glial scar [148].

Research has shown that reactive astrocytes contribute to the production of CSPGs following SCI,

and that CSPGs, specifically NG2, are upregulated by 24 hrs after injury and peak expression

occurs at 7 days post-injury[153, 154].

Myelin associated inhibitors such as Nogo, OMgp and MAG bind the Nogo-66 receptor (NgR)

found on neurons. Receptor binding stimulates the downstream Rho/Rock pathway, which

results in decreased growth cone mobility and growth cone collapse. Recently, it was

discovered that PTPsigma, a transmembrane tyrosine phosphatase, was expressed on neurons

and acts as a receptor for CPSGs, which also signal through the Rho/Rock pathway and inhibit

axonal regeneration [155].

1.7.3.4 Progenitor Cell Proliferation

Stem/progenitor cells have been identified in the central canal adult mammalian spinal cord and

have been shown to proliferate extensively following SCI [156, 157]. These cells differentiate

into glia, mainly astrocytes, as endogenous neurogenesis is generally not seen in the spinal cord.

NG2 is a CSPG that is expressed on a subpopulation of progenitor cells and macrophages

following injury [153, 158]. NG2+ progenitors have been described as having the capacity to

differentiate into astrocytes and oligodendrocytes following trauma, with cues for progenitor

differentiation coming from changes in post-injury niches [159]. Additionally, vascular injury

(and changes in VEGF, FGF or EPO expression) can signal the recruitment of endothelial

41

progenitor cells to the site of insult, and results in migration, differentiation, and maturation of

these cells in an endogenous effort to repair the vasculature [160].

1.7.4 The Intermediate Phase

The intermediate phase of SCI occurs between 2 weeks to 6 months post injury in the human

condition (Figure 7). Glial and fibrous scarring progresses, severed axons continue to degenerate

in peri-lesional areas, endogenous axonal sprouting occurs, macrophages remain present and

active in the lesion, and endogenous efforts of remyelination are observed. Macrophages

continue to phagocytose debris from apoptotic cells, degenerating axons and myelin breakdown.

In the intermediate phase of rat SCI, research suggests that endogenous axonal regeneration

exists since axonal sprouting in both corticospinal and reticulospinal tracts has been observed

[161]. Peripheral Schwann cells, as well as oligodendrocyte precursor cells (OPCs), have been

shown to remyelinate and restore some axonal function following spinal cord injury [162].

1.7.5 The Chronically Injured Spinal Cord

The chronic phase of SCI is typically defined as 6 months post-injury humans, and

approximately 6 weeks in rats (Figure 7). Wallerian degeneration of severed axons towards their

cell bodies continues and developing neuropathic pain can be incapacitating. In the chronic

phase of injury, the initial site of injury is characterized by a cystic cavity transversed by

vascular-glial bundles with regenerated nerve roots [163]. What's more, astrocyte and collagen

fibers extend through the lesion and surround the cyst. It is believed that the lesion, after 1-2

years, will cease to progress and continuing deficits are stabilized.

42

1.7.5.1 Post-Traumatic Syringomyelia

Post-traumatic syringomyelia (PTS) is characterized by the formation of a fluid-filled cavity

anytime up to 30 years following injury, and is anatomically observed by imaging in

approximately 21-28% of human SCI patients [164]. However, symptomatic syringomyelia is

seen in only 1-9% of patients: making it a relatively uncommon post-injury complication [165,

166]. Nevertheless, for individuals affected, the side-effects can be undesirable. Common side-

effects of PTS include progressive and asymmetrical weakness, increased spasticity, and

segmental pain or sensory loss due to compressive/pressure injury of spinalthalamic pathways at

or above the level of injury [167]. A syrinx is often formed as a result of increased CSF pressure

from arachnoid lesions or cord compressions, and ultimately leads to increased inflow of CSF

[168]. Animal studies have also shown that even small areas of subarachnoid scarring can result

in fluid flow alteration and increased pressure in the subarachnoid space (SAS) [169]. CSF flow

in humans with PTS is also associated with blockage of the SAS by the progression of arachnoid

scarring [170].

1.7.5.2 Neuropathic Pain

Development of neuropathic pain is dependent on location of the injury site and the surrounding

neural pathways. Clinically, neuropathic pain is divided into three areas which help to describe

the location of the pain: “above-level”, “at-level” and “below-level”. Chronic astrocyte and

microglial activation produce factors that result in hyperexcitability of neurons in distal regions

of the dorsal/ventral horns, with respect to the epicenter [171, 172]. Patients may also develop

mechanical and/or thermal allodynia, which causes previously innocuous stimuli to feel noxious.

In rats, neuropathic pain develops approximately 4 weeks post-injury and depends on injury

43

intensity [173]. In humans, it is estimated that the number of patients exhibiting neuropathic

pain is as high as 58% in some patient populations of SCI [174]. Consistent with the knowledge

that astrocytes and microglia are highly active in neuropathic pain, therapies that inhibit or

modulate astrocyte, microglial/macrophage activation have shown a reduced incidence of

neuropathic pain in animal models of SCI [175, 176].

1.8 Non-Traumatic Causes of Spinal Cord Injury

Non-traumatic spinal cord injury results in slow, prolonged damage to the spinal cord, rather

than a one-time blunt trauma [177]. With a prevalence estimated at near two-times greater than

traumatic SCI, and considering it occurs in elderly individuals, the diagnosis and treatment of

non-traumatic injuries is an important research area [178]. Similar to traumatic injuries, damage

can occur at any location along the axis of the spinal cord and the functional deficits are dictated

by the level of injury. Common causes of non-traumatic injury include systemic hypotension,

cardiac arrest or stroke, primary or secondary neoplasms, spinal infection, multiple sclerosis,

cervical spondylotic myelopathy, amyotrophic lateral sclerosis, and birth defects, including

cerebral palsy and neural tube deficits [177, 179]. Due to the slow progression of non-traumatic

SCI, the pathophysiology shows a drastically different spatial-temporal profile compared to

traumatic SCI [180]. Moreover, the slow onset of molecular and anatomical alteration allows for

endogenous compensatory mechanisms to be initiated, which often reduce functional deficits and

lead to better patient prognosis. Unfortunately, non-traumatic SCI results in gradual functional

losses, which in turn, results in a delayed diagnosis and treatment of the problem. It is important

that diagnostic methods be improved, as these tools could greatly enhance the early detection of

non-traumatic SCI and ultimately improve the patient outcomes.

44

1.9 Therapeutic Targets

Elucidating pathways and mechanisms is critical to understanding the progression of diseases

and trauma; however, the ultimate goal is to develop therapies which will yield positive

functional outcomes. Although some aspects of SCI pathophysiology still remain undefined

(such as the precise spatio-temporal distribution of secondary injury), there is a clear

understanding of spinal cord anatomy and neural systems, the relationship between primary and

secondary, and many key regulators involved in cell death and degeneration. With this

knowledge, researchers have started to develop therapeutic interventions targeting specific

secondary cascades (see Figure 6 and Figure 7).

In therapeutic development, it is pivotal that treatments are unequivocally safe and effective, and

it is now required that preclinical studies are conducted in an independent laboratory to ensure

the data are reproducible. Researchers have experimented with cell transplant therapies,

bioengineered therapies, molecular therapies and rehabilitation therapies [181]. Some promising

research from each type of therapeutic intervention will be highlighted.

1.9.1 Cell Transplantation

Overall, cell and/ or tissue transplantation after SCI aims to: 1) transverse any cysts or cavities;

2) restore dead or damaged cells; and 3) create an environment which is conducive to axonal

regeneration.

Individually, peripheral nerve grafts have been shown to support growth of various axonal types,

with the exception of supraspinal axons [182]. However, in combination with various therapies

45

(including anti-inflammatory drugs and acidic fibroblast growth factor), peripheral graft were

able to promote recovery with regeneration of supraspinal axons into, through and beyond grafts

[183]. Using a contusion model of SCI, researchers implanted Schwann cells and observed a

reduction in cavity size, and observed that some spinal axons extended into grafts, and many

were remyelinated [184]. Functional hindlimb recovery was also reported in this study.

Stem cells and their progenitors have been a focus in regenerative medicine for many years.

Embryonic stem cells (ESCs), adult stem cells, adult neural precursors (aNPCs), and induced

pluripotent stem cells (iPSCs) all have a regenerative capacity that could be harnessed to treat

SCI; however, each cell type carries both advantages and disadvantages for their use in a clinical

setting. ESCs have the ability to differentiate into many cell types, indefinitely self-renew and

differentiate into any cell type, which makes them a powerful regenerative tool. However, ESCs

are harvested from human a fetus, which makes their use highly controversial. Moreover, since

ESCs infinitely self-renew and can differentiate into any cell type, transplantation of these cells

could produce excessive cell proliferation and tumor formation, or produce unwanted cell types

at the site of transplantation, respectively.

The most successful approach of embryonic CNS-derived stem cells in neurotrauma is using

progenitor cells that have been pre-differentiated to a neural lineage prior to transplantation,

since this restricts the cell types that are able to be generated in vivo. Neural progenitors have

been transplanted into immunosuppressed mice and non-human primates following spinal cord

contusion, and in both experiments the transplanted cells showed acceptable survival and

differentiated into cells displaying oligodendrocyte and neuronal markers. Moreover, these

experiments reported locomotor improvements as a result of neural progenitor transplantation

46

[185]. Adult neural precursor cells (NPCs) are now being investigated for CNS repair, since

there are fewer ethical issues associated with their use. Mouse brain-derived adult NPCs have

been transplanted, in combination with growth factors, into the injured spinal cord of adult rats.

NPCs transplanted 2 weeks post-injury displayed survival, migration, and integration in the host

spinal cord tissue, and were able to generate mature oligodendrocytes that resulted in the

remyelination of injured axons, and promoted some functional recovery [186].

Pre-clinical data have shown that animals treated with human ESCs (hESCs) show promising

improvement in functional recovery following SCI [187]. After observing such promising pre-

clinical data, further studies were conducted to characterize the safety and efficacy of these

hESCs. The Geron trial – which was originally approved by the FDA, but then halted due to

concerns of abnormal cyst formation – was later re-initiated and approved for phase I clinical

trials in the United States of America, using human ESCs in patients with SCI. In October

2010, the first patient in Geron’s clinical trial was treated. However, in November 2011, the

phase I trial was closed due to financial costs of the clinical trial, and indicated that of the

patients treated to date, the therapy showed no signs of being effective [188].

In December 2010, StemCells Inc., an American-based company, started a phase I/II clinical trial

using human neural stem cells (HuCNS-SCs) in chronic spinal cord injury, with the intention of

showing efficacy and safety of HuCNS-SCs in human patients [189]. This on-going trial is

taking place in Switzerland, out of the University of Zurich. Patients were initially recruited

from across Europe; however, as of December 2011, StemCells Inc. has opened up recruitment

of patients to Canada and the USA [190]. The trial will include thoracic SCI patients categorized

as either ASIA A or B.

47

Recently, hESCs have also been approved for a clinical trial in Scotland, which focuses on

treating stroke patients [191]. To date, ReNeuron reports no cell-related adverse effects in the

five stroke patients treated, and the company remains optimistic that long-term follow-ups will

show beneficial results. Although this clinical trial will not include SCI patients, the results of

this trial will be very interesting and relevant, since stroke and SCI share many common

pathophysiological mechanisms.

TotipotentRX Cell Therapy, a company based out of the USA and India, has recently launched a

phase I/II clinical trial using autologus bone marrow stem cells [192]. The trial was initiated in

December 2011, although no patients have been enrolled yet. The trial plans to enroll a total of

15 SCI patients with ASIA A, B or C injuries that are below a C4 level. The study will take

place in India, and to date no feasibility or safety data from the study has been published.

1.9.2 Molecular and Neuroprotective Therapies

SCI initiates many molecular processes which have both beneficial and detrimental effects. In

theory, molecular therapies aim to: 1) protect neurons from cell death, therefore reducing tissue

loss; 2) mediate the inflammatory response, 3) promote axonal growth; and 4) restore blood flow

or reduce vascular damage [193].

Neuroprotective therapies are attractive, since they target neuronal and glial survival following

injury. Several studies have reported that intravenous administration of minocycline reduces cell

death and advances hindlimb function in rodent models of SCI [194]. Methylprednisolone, the

only commonly used neuroprotective treatment for acute SCI, has been shown to be

48

neuroprotective with administered within 8 hours of the primary injury; however, the clinical

enthusiasm for methylprednisolone is diminished by its adverse side effects, including increased

rates of infection and impaired wound healing [193]. Riluzole, a calcium-channel blocker, has

shown substantial promise as a neuroprotective therapy. Studies have shown that riluzole is

capable of sparing motorneurons [195], and provides functional and histological improvements

[196, 197]. Other molecules, including estrogen, cyclosporin-A, erythropoietin, have been

shown to reduce cell death and be neuroprotective following traumatic SCI [198, 199].

Additionally, granulocyte colony-stimulating factor (G-CSF) has shown beneficial outcomes,

when coupled with bone marrow-derived stem cell transplantation [200, 201]. Finally,

polyethylene glycol (PEG) has also been associated with both cellular and functional

improvements, which has been attributed to its ability to restore membrane disruption, and

therefore reducing reactive oxygen species and subsequent lipid peroxidation [202-204].

Both systemic and localized hypothermia have been investigated as possible neuroprotective

therapies following spinal cord injury [205, 206]. From a biological view, cooling the injured

tissues would decrease enzymatic activity, and reduce cellular energy needs: overall reducing

and slowing the secondary pathophysiology following SCI. Mixed results have been observed

between preclinical rodent models and human trials, although it is still generally believed that

hypothermia may be a promising therapy for treating neurotrauma; resulting in improved

functional and anatomical outcomes [207-210].

Rescue and regeneration of the microvasculature within the epicenter and penumbra remains a

largely unexplored yet may be a promising therapeutic route to facilitate tissue sparing and

49

functional recovery following SCI. Although vascular endothelial growth factor (VEGF) is

predominantly recognized for its effects on vascular development, it is now accepted as having

cell survival, proliferative, and migratory properties [211]. VEGF supports the “neurovascular

niche”, as it appears to play important roles in both vascular and neural development, bridging

both endogenous systems. Emerging evidence suggests that administration of VEGF following

SCI results in improved vascular networks, reduced cell death and beneficial functional recovery

[212].

The early expression of pro-inflammatory cytokines and chemokines following SCI may

represent an important therapeutic target [117]. Studies have shown that reducing the host

inflammatory response, by modulating factors such as IL-10 and MMP-9, is beneficial and

results in a better outcome following SCI; however, other studies have indicated that obliterating

reactive astrocytes in the injured spinal cord results in detrimental molecular and functional

effects [213, 214]. Clearly, immune-modulating therapies have potential, but further research

will need to elucidate the complexity of immune regulation, as well as the spatio-temporal

distribution of the inflammatory response, in order to accurately intervene and effectively control

the inflammatory response following SCI.

1.9.3 Rehabilitation Therapies

Improvements in locomotor function have been reported in mammals with both incomplete and

complete SCI following exercise or rehabilitation [215]. Locomotor training has been shown to

enhance the ability of many spinally transected mammals to walk on a treadmill in the presence

of body-weight support [216]. Since the spinal circuitry below the lesion site does not become

50

silent following SCI, it often maintains active and functional neuronal properties, which are able

to react to input from below the level of the injury. These circuits are capable of generating

oscillating, coordinated motor patterns and have demonstrated neural plasticity [217].

Combinatorial research is becoming increasingly popular, and many studies are looking at the

combined effects of rehabilitation with molecular strategies for promoting CNS axon

regeneration and recovery of limb function.

Many on-going clinical trials for SCI examine aspects of rehabilitation, including upper-

extremity exercise, body-weight-supported treadmill training, robotic or manually assisted

training, and/or functional electrical stimulation (FES). Studies have shown that, with assisted

weight-support, locomotor training enhances the ability of humans with neurologically complete

SCI to walk on a treadmill. While this is a promising advance for the field of SCI, this

rehabilitation strategy does not result in unaided walking for patients with neurologically

complete SCI [218]. Significant improvements in health, including improved cardiovascular

function and reductions in spasticity, bone loss and bladder/bowel complications have been

observed in patients receiving rehabilitation following injury [219].

1.10 Summary of Spinal Cord Injury

The temporal and spatial progression of SCI results in considerable morphological and functional

damage. Homeostatic regulation of the uninjured CNS is very precise and highly complex, and

disruption from trauma initiates a sophisticated cascade of events. Mechanisms – including

inflammation, apoptosis, necrosis, scarring, and axonal degeneration – are rapidly instigated as a

reaction to the primary injury in an attempt to control and minimize the damage. Although each

51

of these processes has beneficial intent, they also, unfortunately, exacerbate the initial damage

and create an inhibitory milieu which prevents endogenous efforts of repair, regeneration and

remyelination.

In the past few decades, significant advances have contributed to understanding SCI

pathophysiology; however, a full appreciation for the complexity of temporal and spatial

synchronization is yet to be unveiled. Future research will need to further characterize and

develop animal models, which more closely mimic the diverse progression and outcomes that are

observed in human SCI. Moreover, there is an explicit need to define the timeline of

inflammation, cell death, glial response, and regenerative mechanisms following SCI. Even

though these secondary events are responsible for the much of the damage following injury, they

conversely offer potential targets for therapeutic intervention.

1.11 Vascular Growth and Development

Angiogenesis is the growth of new blood vessels from pre-existing blood vessels; essentially, the

sprouting and branching of vessels [220]. This process is distinctly different from

vasculogenesis, which is the formation of blood vessels de novo [221]. While both processes are

pivotal vascular mechanisms, vasculogenesis occurs primarily in development and

embryogenesis, while angiogenesis occurs during development and adulthood – including during

female reproduction, wound healing and the onset of disease states (particularly in the transition

of tumors from dormant to malignant). Both processes are tightly regulated (except in disease

states) to maintain a delicate balance of pro-angiogenic and anti-angiogenic factors (see Table 2).

52

Table 2. Factors involved in angiogenesis.

Pro-Angiogenic Factors Anti-Angiogenic Factors

Angiogenin Angioarrestin/ Arrestin

Angiopoietin-1/ Angiopoietin-2 Angiostatin

Fibroblast growth factors: acidic (aFGF) and

basic (bFGF)

Endostatin (collagen XVIII fragment)

Follistatin Fibronectin fragment

Granulocyte colony-stimulating factor (G-

CSF)

Heparinases/ Heparin hexasaccharide

fragment

Hepatocyte growth factor (HGF) Human chorionic gonadotropin (hCG)

Interleukin-8 (IL-8) Interferon alpha/beta/gamma

Leptin Interferon inducible protein (IP-10)

Placental growth factor Interleukin-12 (IL-12)

Platelet-derived growth factor (PDGF) Metalloproteinase inhibitors (TIMPs)

Pleiotrophin (PTN) Placental ribonuclease inhibitor

Progranulin Plasminogen activator inhibitor

Proliferin Platelet factor-4 (PF4)

Transforming growth factor-alpha (TGF-

alpha)

Prolactin 16kD fragment

Transforming growth factor-beta (TGF-

beta)

Transforming growth factor-beta (TGF-

beta)

Tumor necrosis factor-alpha (TNF-alpha) Soluble Fms-like tyrosine kinase-1 (sFlt-1)

Vascular endothelial growth factor (VEGF)/

Vascular permeability factor (VPF)

Vasculostatin

1.11.1 Vasculogenesis

Vasculogenesis is a complex and dynamic series of processes, controlled by the cell–

extracellular matrix (ECM) and specific cell–cell interactions in the presence of growth factors

53

[221]. The hemangioblast, is a common progenitor of both the endothelial and hematopoietic

cell, that is responsible for the initiating and expanding the original vascular network [222].

Hemangioblasts aggregate in the embryonic yolk sac, and based on location, two separate cell

populations are formed. Outer cells form endothelial cells, while inner cells develop into

hematopoietic precursors. Factors such as, vascular endothelial growth factor (VEGF), VEGF

receptor 2 (VEGFR-2) and basic fibroblast growth factor (bFGF) influence angioblast

differentiation, but on the contrary, VEGFR-1 inhibits differentiation [223, 224].

There are four definitive stages associated with vasculogenesis [221, 225]. First, the angioblast

is created from the mesoderm. Second, angioblasts form endothelial cells, which assemble into

vascular structures (“blood islands”). Third, lumen is developed as part of more defined and

mature vessels. And finally, the vascular structures are organized and connected to form a

continuous vascular network. Afterwards, angiogenesis occurs to further expand the vasculature

and reach more distant tissues (Figure 8).

54

Figure 8. Vascular development. A) Hemangioblasts/angioblasts are formed from the

mesoderm germ layer. B) Vasculogenesis occurs from a collection of angioblasts and

endothelial cells, to form the first vascular networks. C) Angiogenesis occurs, by a variety of

mechanisms, to increase the local blood supply or extend the blood supply to more distal targets.

Figure taken from Hendrix et al. [226]. (Figure permission requested).

Originally, vasculogenesis was thought to only exist in embryonic development. While

vasculogenesis still predominantly occurs in development, recent research has shown that adults

are also able to exhibit this form of vascular growth [227]. Key evidence to support this

demonstrates the presence of circulating endothelial cells and endothelial precursor cells in

postnatal animals, as well as a deeper understanding toward the mechanisms that control blood

vessel formation in tumor growth.

55

1.11.2 Angiogenesis

Angiogenesis is driven by the release of angiogenic factors, but mediated by the simultaneous

expression of anti-angiogenic factors (Table 2). The release of these factors is often from

injured, diseased, or developing tissues, which are signaling the need for a greater blood supply

due to a lack of oxygen or nutrients.

Angiogenesis can occur by a number of mechanisms: sprouting, bridging or intussusception,

although sprouting is (by far) the most extensively studied [222, 226]. Vasodilation, triggered by

the presence of nitric oxide (NO), is the initiating step in angiogenesis [220, 224]. In response to

localized VEGF, vascular permeability is increased, which allows plasma proteins to leak out

and create a temporary scaffold for endothelial cells that migrate to the angiogenic site.

Increased permeability is mediated by the spatial rearrangement of platelet endothelial cell

adhesion molecule 1 (PECAM-1) and vascular endothelial (VE)−cadherin , and the formation of

vascular fenestrations. Permeability is a highly regulated and important process. Although

permeability promotes and drives angiogenesis, excessive vascular leakage can result in

circulatory collapse, hypertension, edema or metastasis [220, 221]. Excessive vascular leakage

is restricted by the expression of angiopoietin 1(Ang1), a ligand of the endothelial TIE-2

receptor, which is released as an inhibitor of vascular permeability to constrict pre-existing

vessels [228].

In order for angiogenesis to occur, the mature vasculature needs to be destabilized, which allows

resident endothelial cells to emigrate. Ang2, an inhibitor of TIE-2 signaling, is involved in the

destabilizing process and is responsible for detaching smooth muscle cells and loosening the

56

matrix [229, 230]. Proteinases belonging to the plasminogen activator, matrix metalloproteinase

(MMP), or heparanase families promote angiogenesis by activating/releasing growth factors

(bFGF, VEGF) that are stored within the extracellular matrix [231]. Additionally, these

proteinases stimulate angiogenesis by contributing to the physical degradation of the matrix,

further destabilizing the original vessels.

When the vasculature has been reverted to an immature state, proliferating endothelial cells are

able to migrate to distal sites. Growth factors, such as VEGF and placental growth factor

(PLGF), along with their receptors, stimulate endothelial cell recruitment, proliferation and

organization [220]. Ang1 is a chemoattractant for endothelial cells, induces sprouting, and

potentiates VEGF; however, it is unable to induce endothelial proliferation [229, 232]. In

contrast to VEGF, Ang1 cannot initiate endothelial network organization, rather it stabilizes the

vascular networks initiated by VEGF. This is likely to occur by Ang1 stimulating the cell-cell

interactions between endothelial and peri-endothelial cells, suggesting that Ang1 acts in response

to VEGF or in the latter stages of angiogenesis. In the presence of VEGF, Ang2 is also pro-

angiogenic. However, the function of Ang2 is still being elucidated, since recent research

indicates that overexpression of Ang2 in tumors suppresses their growth [233]. Moreover, recent

studies have shown that low levels of phosphorylated TIE2 receptors have been observed in the

dormant vasculature of adults, which may suggest that TIE2 plays a role in basal vascular

maintenance [233].

Angiogenic sprouting is restricted by a delicate balance of activators and inhibitors. Angiogenic

inhibitors, which suppress the proliferation or migration of endothelial cells, include angiostatin,

endostatin, transforming growth factor-beta (TGF-β), interferon-α/β/γ, leukemia inhibitory factor

57

(LIF) and platelet factor 4 [220, 222]. Activators may be cytokines or proteins, many of which

are involved in cell-cell or cell-matrix interactions, induce proliferation and migration of

endothelial cells. Activators include angiopoietin, nitric oxide, tumor necrosis factor-alpha

(TNF-α), platelet derived growth factor (PDGF), VEGF/VPF, and fibroblast growth factors

(aFGF/bFGF). Some of these factors – PDGF, TNF-α, nitric oxide, and TGFβ – are not essential

for embryonic vascular development, but have been shown affect pathological angiogenesis

and/or improve angiogenesis following exogenous administration [220, 221]. The following

molecules have also been associated with angiogenesis following exogenous delivery; however,

their role in endogenous angiogenic signaling still remains uncertain: erythropoietin, IGF-1,

neuropeptide-Y, leptin, epidermal growth factor, hepatocyte growth factor, interleukin hormones

and chemokines.

When endothelial cells have been assembled in new vessels, they become quiescent and can

survive for many years [220]. However, endothelial apoptosis occurs as a natural mechanism of

vessel regression in the retina and ovary after birth, and is the common intention of exogenous

angiogenic inhibitors. The balance between apoptosis and survival is regulated by vascular and

metabolic demands of the tissue. Endothelial survival factors (including VEGF, Ang1 and αvβ3)

activate survival pathways (PI3-kinase/Akt, Bcl-2, surviving), but more importantly, they are

also capable of suppressing cell death factors/pathways (p53 and Bax) [234].

Generally, endothelial cells accumulate as solid cords, and establish lumen formation at a later

time point. Vessels can increase their diameter or length by intercalation or tapering of

endothelial cells, as well as by fusing pre-existing vessels. Varying VEGF isoforms play

different roles on the size and formation of new vessels, where VEGF189 decreases luminal

58

diameter, compared to isoforms VEGF121, VEGF165 which promote lumen formation and

increase vessel length [235, 236]. Ang1 and integrins (αvβ3 or α5) also promote the formation of

lumen or increase vessel maturity [237]. In contrast, thrombospondin-1 is an endogenous

inhibitor of lumen formation [238].

To date, many questions still remain about the spatial cues that guide endothelial cells into

correct patterns and three-dimensional networks. Ultimately, a greater understanding of these

mechanisms would allow for more targeted and specific therapeutic angiogenesis.

1.12 Vascular Endothelial Growth Factor

1.12.1 Molecular Biology of VEGF

It is well known that vascular endothelial growth factor (VEGF) has a multitude of cellular

functions, including vascular endothelial proliferation and differentiation, embryonic

development, maintenance and repair of blood vessels, and angiogenesis [239-242].

VEGF (also known as VEGF-A, as it was discovered first) belongs to a sub-family of growth

factors, specifically the family of platelet-derived growth factor (PDGF). This group of growth

factors includes VEGF-A, -B, -C, -D, -E, -F, and placental growth factor (PLGF). VEGF-A

consists of 16,272 base pairs and is located on the human chromosome 6p12 [243]. Twelve

different splice isoforms – from a single gene with 8 exons – are expressed as a result of

alternative splicing events in exons 6, 7 and 8 (Figure 9) [244]. Splice isoforms are named

according to the number of amino acids present, and it should be noted that rodent isoforms

contain one less amino acid (e.g. VEGF164 would be the rodent equivalent to human VEGF165).

59

The most common and well characterized forms in the CNS are VEGF121, VEGF165 and

VEGF189 – although isoforms 111, 145, 148, 183, 206 exist – and these isoforms vary with

respect to their solubility and affinity [239, 243]. Two VEGF sub-families are produced as a

result of variable splice sites found in exon 8. Isoforms produced using the proximal splice site

are pro-angiogenic, whereas proteins formed by the distal splice site are known to be anti-

angiogenic (and are denoted VEGFXXX B) [236].

Figure 9. VEGF gene and splice isoforms. A) The VEGF gene is composed of 8 exons, which

are responsible for determining the receptor binding specificity, affinity and dimerization. B)

Isoforms are named based on their amino acid length (i.e. VEGF164 would have 164 amino acid

residues). The number of amino acids found in the various exons are displayed. C) The

diversity of VEGF results from differential splicing of VEGF mRNA, forming both pro-

angiogenic and anti-angiogenic proteins. All pro-angiogenic isoforms contain exon 8a, whereas

60

exon 8b is responsible for anti-angiogenic properties. Figure created by Sarah A. Figley: adapted

from Harper and Bates, 2008 [243].

1.12.2 VEGF Receptors and VEGF Signaling

VEGF receptor expression is, for the most part, limited to vascular endothelial cells [239, 245,

246]. Therefore, proliferation induced by VEGF is predominantly restricted to endothelial cells.

Although VEGF is mitogenic for other cell types – such as lymphocytes and Schwann cells –

alternative functions result from VEGF binding to other non-endothelial cells, notably the

induction of monocyte migration [247-251].

Signal transduction of VEGF is mediated through three tyrosine kinase receptors: VEGFR-1 (Flt-

1), VEGFR-2 (Flk-1/KDR), VEGFR-3 (Flt-4) (Figure 10), and two non-tyrosine kinase-type co-

receptors neuropilin-1 (NP-1) and neuropilin-2 (NP-2) [252]. VEGFR-1 is believed to be

involved in hematopoietic events, and VEGFR-3 is primarily responsible for the lymphatic

endothelium [253]. VEGFR-2 is likely the most characterized and studied receptor, and is

thought to be responsible for the neuroprotective and proliferative properties, by signaling

through the PI3K/Akt and MEK/Erk pathways, respectively (Figure 11) [247, 254].

61

Figure 10. VEGF Receptors. VEGF receptors (VEGFRs) are transmembrane tyrosine kinase

receptors, with the exception of VEGFR-1, which can be cleaved into a soluble form and act as a

decoy receptor. Each receptor binds VEGF; however, they have varying affinities for the

different VEGF proteins (i.e. VEGF-A does not bind VEGFR-3). VEGF ligand binding results

in receptor dimerization and activation via transphosphorylation. EC = endothelial cell. LEC =

lymphatic endothelial cell. Figure taken from Hoeben et al. [255] (Figure permission requested).

62

Figure 11. VEGFR-2 signaling. The majority of the angiogenic effects exhibited by VEGF

occur through VEGFR-2 signaling, and are predominantly restricted to endothelial cells.

VEGFR-2 mediates cell survival (through Akt/Caspase), vascular permeability (via eNOS), cell

migration (via PI3K/p38MAPK), and cell proliferation (through Ras/Erk). Figure taken from

Cross et al. [253] (Figure permission requested).

VEGFR-2 (200–230-kDa) acts as a high-affinity receptor for VEGF-A, and is expressed in both

vascular endothelial and lymphatic endothelial cell, although VEGFR-2 expression has also been

observed in several other cell types such as hematopoietic stem cells [256]. Knockout

experiments demonstrate that VEGFR-2 expression is critical for vascular development and

Vegfr-2 -/- embryos die by embryonic day 8.5–9.5, show signs of endothelial and hematopoietic

precursor defects [257].

63

VEGFR-2, as opposed to the other VEGF receptors, is largely responsible for endothelial

proliferation, survival, migration and vascular permeability. VEGFR-2, stimulates proliferation

through activation of the extracellular regulated kinase (Erk) pathway, ultimately leading to an

increase in gene transcription [258]. Regulation of cell survival is controlled by the Akt/PKB

pathway, which inhibits pro-apoptotic pathways such as B-cell lymphoma 2 (Bcl-2)-associated

death promoter homologue (BAD) and Caspase-9 [259]. Furthermore, the Akt/PKB pathway

also induces endothelial nitric oxide synthase (eNOS) expression, which generates NO and

results in an increase in vascular permeability and cellular migration [260, 261]. VEGFR-2

signaling mediates actin reorganization and cell migration via p38 mitogen-activated protein

kinase (MAPK) and focal adhesion kinase (FAK) [262, 263].

Several other important intracellular signaling molecules are activated by VEGFR-2, notably

Src. It is unknown how VEGFR-2 interacts with Src or what the downstream signaling role is;

however, mice mutants lacking Src family members – specifically Src and Yes – have increased

vascular permeability [264]. Additionally, VEGFR-2 function is altered by co-receptors such as

heparan sulfated proteoglycans. In a functional VEGF–VEGFR-2 complex, neuropilins –

ubiquitous membrane-bound molecules also involved in axon guidance by binding to the

semaphorin family members – are present [246, 265]. These co-receptors interact with certain

VEGF isoforms and VEGFR-2, and although there is currently no evidence that neuropilins have

the capacity to signaling in endothelial cells, it is thought that neuropilins act by stabilizing the

VEGF– VEGFR-2 complex [266].

64

1.12.3 Regulation of VEGF

1.12.3.1 Hypoxic Regulation of VEGF

Endogenously, VEGF is regulated in a complex fashion by several mechanisms; however,

activation of hypoxia-inducible factors would be the primary regulator [267]. Hypoxic events

have been shown to trigger rapid VEGF gene expression by increasing transcription, stabilizing

mRNA and promoting preferential translation [268-270]. Transcription factors, such as hypoxia-

inducible factor 1 (HIF-1) and hypoxia-inducible factor 2 (HIF-2), activate transcription by

binding the hypoxia response element (HRE) in the 5’ flanking region of the VEGF gene. The

HIF-DNA binding complex is formed by a heterodimer of α- and β-subunits [271]. Under

normoxic conditions α-subunits are unstable and are targeted for proteosomal destruction, which

limits the formation of HIF-DNA binding complexes [272]. Degradation of the α-subunit is

dependent on the von Hippel-Lindau (VHL) tumor suppressor, which is recognized by a

ubiquitin ligase and results in ubiquitin-dependent proteolysis of HIF-α [273, 274]. In hypoxic

cells, HIF-α degradation is suppressed due to the functional loss of pVHL. This results in

elevated HIF DNA binding complexes, which ultimately leads to transcriptional activation of

target genes – such as VEGF (Figure 12).

65

Figure 12. Hypoxic regulation of VEGF. HIF-1, the most important transcription factor for

hypoxic-regulated genes, is a heterodimer consisting of HIF-1α and HIF-1β. HIF-1β is

constitutively expressed and stable; however, HIF-1a is stabilized by hypoxic conditions. Under

normoxia, HIF-1α is hydroxylated by prolyl hydroxylase. This hydroxylation signals Von

Hippel Lindau proteins (pVHL) to bind, resulting in ubiquitination (ub) and subsequent HIF-1α

protein degradation. Figure created by Sarah A. Figley.

1.12.3.2 Cytokine and Hormonal Regulation of VEGF

Estrogens have been shown to stimulate VEGF transcription, as well as stabilize VEGF mRNA,

therefore extending the half-life of the transcripts [241, 275]. Examination of the 5′ regulatory

regions of VEGF have not found to any estrogen response elements; however, the regulatory

regions contain several AP-1 and Sp1 sites, which have been shown to control estrogen

66

expression and function. The mechanisms of which are still to be determined [275]. Moreover,

reports have indicated that progestins are able to increase VEGF-A expression in the human

uterus and in human breast cancer cells, by promoting transcriptional activation of the VEGF

gene [276, 277].

In addition to hormonal regulation, VEGF/VEGFR expression and activity has been shown to be

regulated by a variety of cytokines, including tumor necrosis factor-alpha (TNF-α), tissue growth

factor-β (TGF-β), epidermal growth factor (EGF), platelet-derived growth factor (PDGF), IL-1α,

IL-1β, IL-6, IGF-1, cyclooxygenase (COX)-1 and -2 enzymes, and Prostaglandin E2 (PGE2)

[255]. Although the role of each molecule on VEGF regulation is not entirely clear, it appears

that VEGF/VEGFR is mediated at both the transcription and post-transcription levels by a

number of mechanisms in vivo.

1.12.4 VEGF in Models of Neurotrauma

Recent work has shown that VEGF has neurotrophic, neuroprotective, and neuroproliferative

effects, and that the expression of VEGF and its receptors during hypoxia/ischemia, brain, and

spinal cord injury are increased [211, 212, 240, 278-282].

Immediately following CNS trauma, a series of cellular events are triggered, which target

vascular repair and enclose the injury area [283, 284]. Revascularization and repair of the blood-

brain barrier re-establish trophic and metabolic support to the injured tissue [91]. VEGF, which

is involved in revascularization following injury, is upregulated during many pathological events

67

in the CNS, including ischemia [278, 285-287], brain contusion [281], and spinal cord injury

[279, 288, 289].

In the uninjured adult CNS, VEGF expression is limited to the cerebellar granule cells, choroid

plexus, and area postrema, and VEGF receptor expression tends to be very low [265, 290, 291].

However, following CNS injury, VEGF is upregulated and is involved in post-traumatic

angiogenesis, via endothelial VEGFR-2 signaling [278, 281, 285-287, 291-293]. VEGF protein

expression is upregulated in both astroglia and inflammatory cells surrounding the injury site

[279, 281, 288, 289, 294]. Additionally, it has been shown in models of ischemia that neurons

can express VEGF [295]. Following traumatic insults, the primary VEGF receptors present a

particular cellular distribution – VEGFR-2 receptors are observed to be upregulated in neurons,

whereas VEGFR-1 receptors are upregulated almost entirely in reactive astrocytes [292, 293,

296].

Due to the important pleiotropic functions of VEGF, it has been a popular research topic in

recent years, specifically in neurotrauma research. Recently is has been reported that VEGF-

overexpressing transgenic mice show enhanced post-ischemic neurogenesis and neuromigration

[297]. Additionally, in a weight-drop SCI model, rats treated with VEGF165 showed significantly

improved behaviour after SCI, notable repair of blood vessels and reduced apoptosis [298].

However, the previously described approaches using VEGF-A have relied on the introduction of

a single splice isoform of VEGF-A (VEGF165), which may not result in optimal neuroprotective

or angiogenic effects.

68

1.12.5 Angiogenesis Following Injury

As previously discussed, SCI often results in significant vascular damage. Vessels are disrupted

at the macroscopic and microscopic levels, with vessels being physically crushed and/or severed

and by notable alterations to the function and arrangement of the BSCB, respectively.

Angiogenesis, the production of blood vessels from pre-existing vessels, is highly regulated and

essential for the repair and remodeling of tissues following injury [222]. It has been previously

reported that in a number of SCI models, there is a relationship between blood vessel density and

improvements in recovery [299]. Similarly, other CNS injury models also display a correlation

between the formation of new vessels and recovery [300]. Substantial trophic support is

provided by CNS microvessels and these microvessels are crucial for tissue survival [301, 302].

Furthermore, some studies have demonstrated that regenerating axons have a tendency to grow

along blood vessels, using them as a scaffolding pathway for regeneration [303].

To an extent, angiogenesis is observed following traumatic injury; however, the endogenous

efforts are insufficient to fully repair the damage [89]. These studies have determined that there

is an angiogenic response to SCI, and it occurs within the first week following injury. However,

the progression of endogenous revascularization noticeably diminishes with the simultaneous

onset of significant histopathology and pathophysiology. Therefore, new therapies designed to

limit vascular damage, improve vessel density and/or restore blood flow to the injured cord may

be promising for spinal cord repair and recovery. Moreover, therapies administered before 7

days (even between 3 and 7 days), as suggested by Loy and colleagues in 2002 might best

facilitate tissue sparing and regeneration [151].

69

1.13 Gene Therapy

Gene therapy can be described as introducing a gene into an organism to either replace a

defective or non-functional gene, or to regulate the expression of a gene [304]. This objective

can be accomplished by a number of techniques – viral delivery, gene silencing, lipoplexes –

each of which have advantages and disadvantages. Gene therapy is an attractive therapeutic

option, as it has the potential to be customized and highly specific; however, limited success has

been achieved in translating gene therapies due to immune and inflammatory responses,

targeting issues, gene control, and the complexity of multi-gene disorders. Viral techniques have

become popular methods of introducing genes into a biological system, as they have proven to be

efficient and effective; however, there are a number of factors contribute to the effectiveness of

in vivo gene therapy. Viral vectors, such as adenoviruses (AdV) or adeno-associated viruses

(AAV), vary in their biological characteristics, including their ability to elicit an immune

response within a host system, the rate at which they produce the transgene and finally the cell

types in which they transduce.

Adenoviruses (AdV) have been well characterized since their discovery in 1953 [305]. To date,

there are approximately 50 adenovirus serotypes, which are classified into subgroups A-F [306].

Adenoviruses are medium-sized (approximately 100 nm), naked viruses with a linear double-

stranded DNA genome. Due to the size of the virus, the length of the inserted gene is limited,

which eliminates AdV gene therapy as a possibility for some genetic conditions. These viruses

remain transient in the cell, which allows them to use the cellular machinery of the host to

replicate rapidly; however, transient expression results in an evoked immune response, and a

limited time-window for gene expression (generally 10 days). For the later reason, this type of

70

gene therapy is ineffective for genetic conditions that require continuous or long-term treatment.

The immune response that is initiated by AdV vectors is also a concern for its use as a potential

gene therapy tool, although the effects can sometimes be mediated by systemic immune

suppression (i.e. cyclosporin A) [307, 308]. Adenovirus serotype 5 (AdV5), which will be used

in our experiments, has been shown to transduce a wide-variety of cell types, including neurons,

astrocytes and oligodendrocytes [309].

Adeno-associated viruses (AAV) are considered to be small viruses, which can transduce both

dividing and non-dividing cells and stably incorporate into specific sites in the host genome

[310]. The predictable integration of AAV makes these viruses more desirable in medical

research, since other vectors, such as retroviruses, are prone to random insertion and subsequent

mutagenesis [311]. Typically, AAV's evoke a very low – often non-detectable – immunogenic

response when administered into a host system [308, 312]. This, as well as the capacity to

transduce quiescent cells, demonstrates an obvious advantage of AAV for human gene therapy,

compared to AdV therapies.

Conversely, AAV systems do present some disadvantages. AAV vectors have a limited cloning

capacity and most therapeutic genes require the complete replacement of the virus' 4.8 kilobase

genome. Therefore, large genes are not suitable for use in a standard AAV vector. Serotype 2

(AAV2) has been the most extensively examined so far and it appears that AAV2 display

preferential transduction of neurons, skeletal muscle, vascular smooth muscle and hepatocytes

[313-315]. Integration into the host chromosome requires time, therefore the kinetics of gene

expression for AAV gene therapy is usually considerably slower than AdV gene therapy [312].

For therapies requiring rapid changes in gene expression, AAV viruses may not be ideal.

71

1.14 ZFP-VEGF Technology and Production of VEGF

ZFP-VEGF technology – a viral vector encoding a zinc-finger transcription factor protein (ZFP),

which activates endogenous VEGF-A expression – has been previously used to demonstrate that

expression in vivo leads to induced expression of VEGF-A protein, stimulated angiogenesis, and

accelerated wound healing [316, 317]. Evidence for a potentially therapeutic biophysiologic

effect of ZFPs has also been reported in animal models of hindlimb ischemia [53, 318, 319] and

diabetes [320, 321]. Here, we propose that AdV-ZFP-VEGF be administered in a delayed

fashion following SCI. This unique method aims to promote upregulation of endogenous

mechanisms – therefore all splice isoforms will be produced – which should mimic physiological

VEGF function (Figure 13).

72

Figure 13. ZFP-VEGF technology. Adenoviral (AdV) and Adeno-associated virus (AAV)

vectors have been designed. Vectors transducer the cell in vivo, and generate bio-engineered

zinc-finger protein (ZFP) transcription factors (AdV-ZFP-VEGF). AdV-ZFP-VEGF binds to

the endogenous gene with high affinity and specificity. The transcriptional activator (TF) drives

transcription of the endogenous VEGF gene, resulting in increased mRNA and protein. This

ZFP-VEGF approach is advantageous since it promotes expression of the endogenous gene,

resulting in the appropriate balance of the multiple VEGF isoforms, which are required for

proper VEGF function. Figure modified from I. Siddiq [322] (Figure permission requested).

73

Rationale

This research aims to investigate two important areas in spinal cord injury (SCI) research. First,

we aim to characterize the vascular disruption following traumatic SCI, since understanding the

vascular profiles may provide interesting and specific therapeutic targets. Secondly, we aim to

use a ZFP-VEGF gene therapy in an animal model of thoracic SCI, and examine the

neuroprotective and vascular effects, at both molecular and functional levels.

The blood-spinal cord barrier (BSCB) plays an important role in maintaining homeostasis in the

central nervous system (CNS), by regulating the transport of molecules and cells across its

barrier [11, 31]. Following injury, the BSCB and the vasculature are significantly disrupted,

which results in increased permeability, decreased immunological protection, and dysregulation

of vascular homeostasis. While a disordered BSCB has been observed following injury, and is

known to propagate pathophysiological mechanisms, it may also presents a unique opportunity

to deliver therapeutics to CNS tissues, which normally (or at later times following injury) cannot

enter due to the intact BSCB. Since SCI is a dynamic injury, exhibiting temporal and spatial

changes, it is important to investigate these alternations in order to identify potential therapeutic

targets. Identifying anatomical/molecular targets, as well as a time-window for regeneration will

prove highly beneficial for administering and designing vascular therapies. In the present study,

we aim to examine the temporal profile of vascular and BSCB damage along the rostrocaudal

axis of the thoracic spinal cord in a clip-compression model. Although other studies have

examined vascular and BSCB disruption post-SCI, our study will be the first to investigate both

of these aspects in a clip-compression model, where data extends past 24 hours. Moreover, this

74

will be the first study to examine the endogenous angiogenic response in a model of clip-

compression SCI.

Secondly, the rationale for studying SCI is that it is a devastating injury that primarily affects

young individuals, therefore significantly reducing their quality of life for many years. It has

been estimated that at least 10,000 North Americans will suffer a SCI each year, as it is

estimated that acute traumatic spinal cord injury occurs with an annual incidence rate of 15–40

persons per million [2]. At present, there are no universally accepted treatments for this

debilitating neurological condition.

SCI results in the initiation of multiple secondary injury cascades, one of which is the disruption

of spinal cord blood flow and the onset of spinal cord ischemia. Vascular changes include

reduction in blood flow, hemorrhage, systemic hypotension, loss of microcirculation, disruption

of the blood-spinal cord barrier (BSCB) and loss of structural organization [58, 323].

Approaches that address the onset and downstream consequences of the ischemic injury are

attractive treatment options for patients with SCI. Moreover, therapies specifically targeting

vascular damage, vessel density and restoration of blood flow to the injured spinal cord may

provide an opportunity for spinal cord repair and recovery. Recent reports have shown

significant correlations between blood vessel density and improvements in recovery following

CNS trauma [324-326].

Rescue and regeneration of the microvasculature within the epicenter and penumbra remains

largely unexplored yet may be a promising therapeutic route to facilitate tissue sparing and

functional recovery following SCI. It has been shown that substantial trophic support is

75

provided by CNS microvessels [302] and that microvessels are critical for tissue survival [301,

327, 328]. In this research we target the spinal vasculature using VEGF, a very prominent

angiogenic mediator in development, which is responsible for endothelial proliferation, survival,

migration and vascular permeability. By inducing VEGF expression in vivo, we are hopeful that

increased VEGF following SCI will result in beneficial effects following SCI. Although many

angiogenic factors exist, VEGF is often regarded as the most important [255]. VEGF, delivered

in a multitude of ways, has shown very promising results towards vasculature and

neuroprotection in models of both neurotrauma and neurological disease, further supporting our

choice to use VEGF as gene therapy [319, 329-333]. Our novel approach shows more promise

than previous attempts to target angiogenesis as a therapeutic target because endogenous VEGF-

A expression is upregulated by the production of a specific zinc-finger protein. This form of

gene therapy mimics physiological VEGF production, which should result in the production of

all VEGF isoforms in the injured spinal cord, a necessary component for proper and functional

angiogenesis. Therefore, we believe that ZFP-VEGF gene therapy presents an advantage over

previous VEGF therapies, which introduce a single VEGF isoform (predominantly VEGF165 has

been used) into the CNS.

Previously, we have reported that immediate AdV-ZFP-VEGF administration results in

increased vessels, decreased cell death, and improved functional recovery following injury (See

Appendix 1) [212]. While these data are important in providing a “proof-of-concept” for the use

of AdV-ZFP-VEGF in vivo, the administration immediately following SCI has limited

therapeutic relevance. In the present study, we will investigate the delivery of AdV-ZFP-VEGF

at 24 hours post-injury to determine if the beneficial effects are preserved with a delayed

administration of the treatment.

76

Overarching Hypothesis

It is hypothesized that significant vascular disruption will occur following a clip-compression

model of spinal cord injury, and that using a ZFP-VEGF gene therapy will enhance molecular

and functional recovery following spinal cord injury through neuroprotective and angiogenic

mechanisms.

Statement of Objectives

1. Characterize the structural and functional changes to the spinal cord vasculature

following moderately-severe SCI

2. Identify potential vascular and neuronal mechanisms of AdV-ZFP-VEGF in vivo

3. Examine the acute therapeutic effects of delayed AdV-ZFP-VEGF administration

following acute spinal cord injury, specifically angiogenic and neuroprotective effects

4. Determine the neurobehavioural effects of delayed AdV-ZFP-VEGF administration post-

spinal cord injury

77

Chapter 2

2 General Methods

2.1 Animal Model of SCI and Intraspinal Injections

All animal protocols (979.31 and 891.9) and procedures were approved by the Animal Care

Committee at the University Health Network, Toronto, ON, Canada.

Animals were subject to a contusive-compressive spinal cord injury using a modified aneurysm

clip, which has been extensively characterized by our laboratory and previously described [74].

Briefly, adult female Wistar rats (250-300g; Charles River, Montreal, Canada) were deeply

anesthetized using 4% isoflurane, and were sedated for the remainder of the surgery under 2%

isoflurane. Animals received a two-level laminectomy of mid-thoracic vertebral segments T6-

T7. A modified clip calibrated to a closing force of 35g was applied extradurally to the cord for 1

minute and then removed (Figure 14). The animals were divided into four groups in a

randomized and “blinded” manner, (1) Sham control group (laminectomy only – no SCI), (2)

Non-injected injured control group (laminectomy and SCI – no injection), (3) AdV -ZFP-VEGF

treatment group, and (4) AdV-eGFP control group. Using a stereotaxic frame and glass capillary

needle (tip diameter 60 µm) connected to a Hamilton microsyringe, a total of 5x108 viral plaque

forming units (PFU) were injected into the dorsal spinal cord 24 hours post-SCI. Four 2.5 μl (10

μl total) intraspinal injections were made bilaterally at 2mm rostral and caudal of the injury site

(Figure 14). Injections were 1mm lateral from the midline and 1mm deep into the spinal cord.

The injection rate is 0.60 µl/min and when the injection was completed, the capillary needle was

78

left in the cord for at least 1 min to allow diffusion of the virus from the injection site and to

prevent back-flow. The incision was closed in layers using standard silk sutures and animals

were given a single dose of buprenorphine (0.05 mg/kg). Animals were allowed to recover in

their cage under a heat-lamp and, subsequently, were housed in a temperature-controlled warm

room (26°C) with free access to food and water. Animals were given buprenorphine (0.05

mg/kg) every 12 hours for 48 hours following surgery, and their bladders were manually voided

three times daily. A subcutaneous injection of 10mg/kg of cyclosporin-A was administered daily

starting 24 hours prior to the SCI until the end of the experiments for immunosuppression. For

histological and protein analysis, n=3-10/group were used. For long-term and behavioural

analysis, n=12/group were used. Final animal values reported in each study vary from the

original values due to mortalities, and are outlined in detail within each chapter.

79

Figure 14. Model of spinal cord injury and intraspinal injections. A) A modified aneurysm

clip with a 35g closing force is applied to the spinal cord for 1 minute, creating a moderately-

severe contusion-compression injury. A two-level laminectomy is performed at T6-T7 to expose

the spinal cord. B) Intraspinal injections are given at four locations surrounding the injury site:

two rostral and two caudal. Injections are approximately 2 mm from the injury site, 1 mm from

the spinal midline, and 1 mm deep into the spinal cord.

2.2 Viral Vector Constructs

The VEGF-A-activating ZFP and controls were provided in viral vectors by Sangamo

BioSciences (Pt. Richmond, CA) and have been previously described [321, 334]. The ZFP-

VEGF expression cassette is illustrated in Figure 15. The VEGF-A-activating ZFP (32E-p65) –

referred to as ZFP-VEGF for the remainder of the manuscript – is a 378 amino acid multi-

domain protein that is composed of three functional regions: (1) the nuclear localization signal

(NLS) of the large T-antigen of SV40, (2) a designed 3-finger zinc-fingered protein (32E) that

binds to a 9 base-pair target DNA sequence (GGGGGTGAC) present in the human VEGF-A

promoter region and (3) the transactivation domain from the p65 subunit of human NFκB, which

is identical to VZ+434, subcloned into pVAX1 (Invitrogen, San Diego, CA) with expression

driven by the human cytomegalovirus (CMV) promoter. Adenoviral (Ad5-32Ep65 or Ad5-

eGFP) vectors, referred to as AdV-ZFP-VEGF and AdV-eGFP, respectively, were packaged by

transfecting T-REx-293 cells (Invitrogen, San Diego, CA). T-REx-293 cells in ten-stack cell

factories were inoculated with Ad vectors at a multiplicity of infection (MOI) of 50 to 100

particles per cell. When adenoviral mediated cytopathy effect (CPE) was observed, cells were

harvested and lysed by three cycles of freezing and thawing. Crude lysates were clarified by

80

centrifugation, and 293 cells were seeded at 4x107 PFU and grown 3 days prior to transfection.

The calcium phosphate method was used for transfection. Infectious titers of the Ad vectors were

quantified using the Adeno-X Rapid Titer kit (Clontech, Mountain View, CA).

Figure 15. ZFP-VEGF expression cassette. CMV pro – cytomegalovirus promoter/enhancer;

NLS – nuclear localization sequence; VEGF-ZFP – engineered VEGF transcriptional activator;

NF-қB p65 AD – transactivation domain from the p65 subunit of human NF-κB; bGH pA –

bovine growth hormone polyadenylation sequence. Arrow indicates transcription initiation site.

Control viruses Ad-DsRed and AAV-GFP have been designed with both VEGF-ZFP and NF-қB

p65 domains deleted, and either DsRed or GFP domains inserted, respectively.

2.3 Western Blotting

Following deep inhalational anesthetic (isoflurane), animals were sacrificed at five or ten days

post-SCI and a 5 mm length of the spinal cord centered at the injury site was extracted. Samples

were mechanically homogenized in 400 μl of homogenization buffer (0.1M Tris, 0.5M EDTA,

0.1% SDS, 1M DTT solution, 100mM PMSF, 1.7mg/ml aprotinin, 1mM pepstatin, 10mM

leupeptin) and centrifuged at 15,000 rpm for 10 minutes at 4 °C. Supernatants were extracted

and used for western blot analysis, where 20 μg of protein was loaded into 7.5% or 12%

polyacrylamide gels (Bio-Rad, Mississauga, Canada). Membranes were probed with either

81

monoclonal anti-NF200 antibody (1:2000; Sigma, Oakville, Canada), rabbit IgG anti-VEGF-A

antibody (1:100; Santa Cruz Biotechnology, Santa Cruz, CA), or rabbit IgG anti-NFκBp65

(1:1000; Santa Cruz Biotechnology, Santa Cruz, CA). NFκBp65 rabbit polyclonal antibody

(1:500; Abcam, Toronto, ON, Canada) was used to recognize the p65 activation domain in the

ZFP-VEGF treated animals. Primary antibodies were labelled with horseradish peroxidase-

conjugated secondary antibodies (goat anti-mouse/rabbit IgG, 1:3000; Jackson Immuno Research

Laboratories, West Grove, PA), and bands were imaged using an enhanced chemiluminescence

(ECL) detection system (Perkin Elmer, Woodbridge, Canada). Mouse monoclonal, beta-actin

(1:500; Chemicon International, Inc., Temecula, CA) was immunoblotted as a loading control.

Quality One detection software (Bio-Rad Laboratories, Hercules, CA) was used for integrated

optical density (OD) analysis.

Table 3. Antibodies used in Western Blot analysis.

Antibody Specificity Source Company Catalog # Working Dilution

anti-NF200 Axons Mouse Sigma N0142 1:2000

anti-VEGF-A VEGF Rabbit Santa Cruz sc-152 1:100

anti-NFκBp65 p65 fragment of

NFkB protein

Rabbit Abcam ab31481 1:500

β-actin Actin Mouse Chemicon MAB1501 1:500

82

2.4 Evans Blue: Blood-Spinal Cord Barrier Disruption

Animals were injected with 1 mL of 2% Evans Blue (EB) into the tail vein [335, 336]. EB was

allowed to circulate for 20-30 minutes and then the animals were transcardially perfused with

saline. One cm of the spinal cord surrounding the injury site was extracted, weighed, and snap-

frozen in dry ice. Samples were then homogenized in 400 μl of N,N’-dimethylformamide

(DMF) and incubated at 50°C for 72 hours. Samples were centrifuged at 18,000 rpm for 30

minutes. The supernatant was collected, aliquoted into a 48 well glass plate, and colorimetric

measurements were performed using a Perkin Elmer Victor3 1420 spectrophotometer at the

absorption maximum for EB (620 nm). Samples were normalized to the original sample weight,

and EB concentration was calculated based on a standard curve of EB in DMF (data reported as

EB per spinal cord weight: μg/g).

2.5 Histochemistry

2.5.1 Histological Processing

Following deep inhalational anesthetic (isoflurane), animals were transcardially perfused with

4% paraformaldehyde (PFA) in 0.1 M PBS. Then, the tissues were cryoprotected in 20% sucrose

in PBS. A 10 mm (1 cm) length of the spinal cord centered at the injury site was fixed in tissue-

embedding medium. The tissue segment was snap frozen on dry ice and sectioned on a cryostat

at a thickness of 14 μm. Serial spinal cord sections at 500 μm intervals were stained with myelin-

selective pigment Luxol Fast Blue (LFB) and the cellular stain Hematoxylin-Eosin (HE) to

identify the injury epicenter. Tissue sections showing the largest cystic cavity and greatest

83

demyelination were taken to represent the injury epicenter. Rostrocaudal spinal cord maps were

created by calculating the distance from the epicenter to the corresponding tissue sections.

2.5.2 Immunohistochemistry

The following primary antibodies were used: mouse anti-NeuN (1:500; Chemicon International,

Inc., Temecula, CA, USA) for neurons, mouse anti-GFAP (1:500; Chemicon International, Inc.,

Temecula, CA, USA) for astrocytes, mouse anti-APC (CC1, 1:100; Calbiochem, San Diego, CA,

USA) for oligodendrocytes, mouse anti-RECA-1 (1:25; Serotec Inc., Raleigh, NC, USA) for

endothelial cells, and rabbit anti-Ki67 (1:1000; Abcam, Toronto, ON, Canada) for cell

proliferation (See Table 4). The sections were rinsed three times in PBS after primary antibody

incubation and incubated with either fluorescent Alexa 568, 647 or 488 goat anti-mouse/rabbit

secondary antibody (1:400; Invitrogen, Burlington, Canada) for 1 hour. The sections were rinsed

three times with PBS and cover slipped with Mowiol mounting medium containing DAPI

(Vector Laboratories, Inc., Burlingame, CA) to counterstain the nuclei. Immunohistochemistry

controls were completed exactly as described above; however, the primary antibody was omitted

from the protocol and only the secondary antibody was added. The images were taken using

either a Zeiss 510 laser confocal microscope (Carl Zeiss Canada, Toronto, ON, Canada) or a

Leica MZ FLIII epifluorescence microscope (Leica Microsystems, Richmond Hill, ON, Canada).

84

Table 4. Antibodies used in immunohistochemistry.

Antibody Specificity Source Company Catalog # Working Dilution

anti-NeuN Neuron Mouse Chemicon MAB377 1:500

anti-GFAP Astrocyte Mouse Chemicon MAB3402 1:500

anti-APC Oligodendrocyte Mouse Calbiochem OP80 1:100

anti-RECA-1 Endothelial cell Mouse Serotec MCA970R 1:25

anti-SMI-71 Blood-spinal

cord barrier

Mouse Calbiochem NE1026 1:1000

anti-Ki67 Proliferation Rabbit Abcam ab16667 1:1000

2.5.3 Quantification of Blood Vessels

Tissue sections – taken from animals sacrificed 10 days post-SCI – were used for

immunofluorescence studies with a monoclonal antibody specific for RECA-1 (Rat Endothelial

Cell Antibody). As shown in Figure 16, the counting of vessels was performed on 4 selected

fields (ventral horn, dorsal horn, left and right lateral columns) in each section under 25X

magnification (0.14 mm2). These areas were selected to provide a representative quantification

of the whole cord, as vascularity varies throughout the grey and white matter. The number of

RECA-1-positive vessels was calculated at 2 mm and 4 mm, both rostral and caudal from the

epicenter, for each animal.

85

Figure 16. Schematic of immunohistochemistry quantification. Four areas of the spinal cord

were selected (2 white matter, 2 grey matter) under 25X magnification. Values from each cord

were either reported as, (i) pooled values or (ii) separated into grey and white matter values.

2.5.4 Identification of Functional Blood Vessels

At 1 hour, 4 hours, and 1, 3, 5, 7, 10, 14 days post-injury animals were sedated with inhalational

anesthetic (isoflurane) and the right femoral vein was surgically exposed (Figure 17). 0.5 mg of

FITC-LEA diluted in saline to a final volume of 1 mL was injected into the femoral vein and was

allowed to circulate for 20-30 minutes (FITC-conjugated Lycopersicon esculentum agglutinin;

Cat # L0401, Sigma, Oakville, ON, Canada). Animals were transcardially perfused with 4%

paraformaldehyde and tissues were fixed and processed as described above. Tissue sections at

the injury epicenter, and 1000 μm, 500 μm, and 250 μm (rostral and caudal) from the epicenter

were used to assess the spatial distribution of vascular damage. Tissue sections were stained

with RECA-1 (1:25; Serotec Inc., Raleigh, NC, USA) to identify endothelial cells/blood vessels.

86

Vessels which were co-labelled with FITC-LEA and RECA-1 were identified as “functional”,

since the presence of FITC-LEA indicated these vessels were connected to the systemic

vasculature and exhibited a perfusion state. Vascular quantification was performed on four

selected fields (ventral horn, dorsal horn, left and right lateral columns) in each section under

25X magnification (0.14 mm2) (Figure 16). The data are presented, separating the white matter

and grey matter, since these regions showed distinct variation in their vascular responses.

Figure 17. Femoral vein injections. Injections of FITC-LEA were delivered into the right

femoral vein of the animals. FITC-LEA was allowed to circulate for 20-30 minutes prior to

sacrifice.

87

2.5.5 Quantification of Apoptosis

An in situ terminal-deoxy-transferase mediated dUTP nick end-labeling (TUNEL) apoptosis kit

(Chemicon International, Inc., Temecula, CA) was used to label apoptotic cells in tissues

extracted from animals 5 days post-SCI. TUNEL staining was completed as described in the

manufacturer’s instructions. The numbers of TUNEL positive nuclei were counted at the

epicenter, as well as at 1, 2 and 3 mm (rostral and caudal) from the injury epicenter. In each

tissue section, the whole section was counted to include all apoptotic nuclei visible. The entire

cord was quantified, as apoptosis may not be evenly distributed throughout the cord, and

sampling four sections (as per Figure 16), may not provide the most representative data for cell

death analysis.

2.5.6 Quantification of Neurons

Tissue sections – taken from animals sacrificed 5 days post-SCI – were used for

immunofluorescence studies with a monoclonal antibody specific for NeuN (neuronal nuclei).

Neuron quantification was conducted under 25X magnification (0.14 mm2), and all NeuN-

positive cells were counted. NeuN is a nuclear antibody, therefore neuronal quantification was

only carried out in the grey matter. The number of NeuN-positive cells was calculated at 1, 2

and 3 mm, both rostral and caudal from the epicenter, as well as at the epicenter.

2.5.7 Quantification of Angiogenesis

Tissue sections – taken from animals sacrificed 5 days post-SCI – were used to quantify

angiogenesis following SCI and AdV-ZFP-VEGF administration. Angiogenesis was calculated

88

as vessels co-labelled with RECA-1 and Ki67 (cellular proliferation). As shown in Figure 16, the

counting of angiogenesis was performed on 4 selected fields (ventral horn, dorsal horn, left and

right lateral columns) in each section under 25X magnification (0.14 mm2). The number of

angiogenic vessels were calculated at 1, 2 and 3 mm, both rostral and caudal from the epicenter,

for each animal. Rostral and caudal values were pooled for each distance.

2.5.8 Assessment of Tissue Sparing and Cavity Formation

Tissue sparing and cavity formation was analyzed 8 weeks after SCI, at the center of the lesion, 2

mm above and 2 mm below the epicenter. Sections were stained with LFB-HE. The

measurements were carried out on coded slides using StereoInvestigator software (MBF

Bioscience, Williston, VT). Cross-sectional residual tissue and cavity areas were normalized

with respect to total cross-sectional area and the areas were calculated every 500 µm within the

rostrocaudal boundaries of the injury site.

2.6 Behavioural Testing

2.6.1 Open-Field Locomotor Scoring

Locomotor recovery of the animals was assessed by two independent observers using the 21

point Basso, Beattie, and Bresnahan (BBB) open field locomotor score [337] from 1 to 8 weeks

after SCI. The BBB scale was used to assess hindlimb locomotor recovery including joint

movements, stepping ability, coordination, and trunk stability. Testing was done every week on a

blinded basis and the duration of each session was 4 minutes per rat. Scores were averaged

89

across both the right and left hindlimbs to arrive at a final motor recovery score for each week of

testing.

2.6.2 Automated Gait Analysis (Catwalk™)

Gait analysis was performed using the Catwalk™ system (Noldus Information Technology,

Wageningen, Netherlands) as described [338, 339]. In short, the system consists of a horizontal

glass plate and video capturing equipment placed underneath and connected to a PC. In our

work, for correct analysis of the gait adaptations to the chronic compression, after

standardization of the crossing speed, the following criteria concerning walkway crossing were

used: (1) the rat needed to cross the walkway, without any interruption, and (2) a minimum of

three correct crossings per animal were required. Files were collected and analyzed using the

Catwalk™ program, version 7.1. Individual digital prints were manually labeled by one observer

blinded to groups. With the Catwalk™, a vast variety of static and dynamic gait parameters can

be measured during spontaneous locomotion. In the present study, we examined the following

parameters, most of which have been studied in human CSM gait analysis:

• Forelimb stride length (expressed in mm): distance between two consecutive forelimb

paw placements

• Hindlimb print area (expressed in mm2)

• Hindlimb print width (expressed in mm)

• Hindlimb print length (expressed in mm)

• Hindlimb swing speed (expressed in pixels/sec): is the speed of the paw during the swing

phase (the duration of no paw contact with the glass plate during a step cycle).

90

Before surgery, animals were acclimated and trained to the walking apparatus following the

method describing by Gensel et al. [340].

2.6.3 Neuropathic Pain: Von Frey Filaments

At level mechanical allodynia was determined at 4 weeks and 8 weeks post-SCI using 2 g and 4

g von Frey monofilaments as previously described [173]. Animals were acclimatized for 30

minutes in an isolated room for 30 minutes prior to pain testing. The von Frey monofilament

was applied to the dorsal skin surrounding the incision/injury site 10 times and animals’

behavioural response to each was recorded. An adverse response to the application of the

monofilament (determined in advance of experiments) included vocalization, licking, biting and

immediate movement to the other side of the cage. The proportion of rats to exhibit allodynia in

each group is reported, and an increased number of responses was associated with the

development of at-level mechanical allodynia. Below-level mechanical allodynia was determined

by quantifying the pain threshold of the hindpaws. Animals were placed in stance on a raised

grid, allowing von Frey filaments to be applied to the plantar surface of the hindpaw. Increasing

monofilaments were used (2, 4, 8, 10, 16, 21, and 26 g) until the animal displayed an adverse

response (as described above). The weight of the von Frey filament that elicited the response

was recorded as the pain threshold value, with lower threshold values indicating increased

sensitivity to mechanical stimuli (and perhaps the development of mechanical allodynia).

Finally, below-level thermal allodynia was assessed using the tail flick method. A 50°C thermal

stimulus was applied to the distal portion of the animals’ tail by a Tail Flick Analgesia Meter

(IITC Inc. Life Science, Woodland Hills, California, USA), and the time for the animal to

91

remove its tail from the stimulus was recorded. The latency time is graphed for each treatment

group, and decreased latency times were associated with the development of thermal allodynia.

2.7 Electrophysiology

2.7.1 Motor Evoked Potentials

Motor evoked potentials (MEPs): In addition to the behavioural assessements, MEPs were

recorded in vivo to assess the physiological integrity of spinal cord. This approach has

been extensively used in our laboratory in rodent models of SCI [5, 341, 342]. In vivo recordings

of motor evoked potentials were recorded from the each of the treatment and control groups at 8

weeks post-injury. For MEPs, rats were under light isoflourane anaesthesia (<1%),

and recordings were obtained from hindlimb biceps femoris muscle. Stainless steel subdermal

needle electrodes were inserted into the muscle. Recordings were acquired using Keypoint

Portable (Dantec Biomed, Denmark). A reference electrode was placed under the skin between

the recording and stimulating electrodes. Stimulation was applied to the midline of the cervical

spinal cord (0.13 Hz; 0.1 ms; 2 mA; 200 sweeps). The amplitude was determined by the

difference between the positive peak and negative peak. Latency was calculated as the time from

the start of the stimulus artifact to the first prominent peak. For individual rats, the average of

peak amplitude and latency was averaged from 200 sweeps and analyses was undertaken by

ANOVA.

92

2.7.2 H-Reflex

The Hoffmann reflex is one of the most studied reflexes in humans and is the electrical analogue

of the monosynaptic stretch reflex. The H-reflex is evoked is evoked by low-intensity electrical

stimulation of the afferent nerve, rather than a mechanical stretch of the muscle spindle, that

results in monosynaptic excitation of alpha-motorneurons. H-reflex can be used as a tool to

study spasticity and short- and long-term plasticity of the nervous system. Recording electrodes

were placed two centimeters apart in the mid-calf region and the posterior tibial nerve was

stimulated in the popliteal fossa using a 0.1 ms duration square wave pulse at a frequency of 1

Hz. The rats were tested for maximal plantar H-reflex / maximal plantar M-response (H /M)

ratios to determine the excitability of the reflex. The recordings were filtered between 10-10000

Hz.

2.8 Statistical Analysis

Data were analyzed with SigmaPlot software (Systat Software Inc., San Jose, California, USA).

For data that investigated the percentage of cells, the data were subjected to an arcsine

transformation prior to statistical analysis to attain a more normal distribution. For comparison

of groups sampled at various distances from the injury site (TUNEL, RECA-1, NeuN), a two-

way analysis of variance (ANOVA) with repeated measures was used, followed by the post-hoc

Holm-Sidak test. For comparisons of multiple groups at a single time point (Western blotting,

BBB, Catwalk™, Electrophysiology), a one-way ANOVA was performed, followed by the post-

hoc Holm-Sidak test.

93

The Holm-Sidak post-hoc was used, as it is recommended as the best multiple comparisons test

following an ANOVA [343, 344]. The Holm-Sidak test is more sensitive and powerful

compared to Bonferroni or Tukey post-hoc tests, therefore it is more likely to detect all

significant results and increases the probability of not committing type II errors (reduces the

chance of rejecting something that is true).

In all figures, the mean value ± SEM are used to describe the results. Statistical significance was

accepted for p values of <0.05.

94

Chapter 3

3 Characterization of Vascular Disruption and Blood-Spinal Cord Barrier Permeability Following Traumatic Spinal Cord Injury

3.1 Abstract

Spinal cord injury (SCI) can be divided into a primary and secondary injury, which refer to the

initial mechanical trauma and later cascade of pathophysiological damage, respectively.

Importantly, vascular changes following injury – such as increased vascular permeability and

disruption to the blood-spinal cord barrier (BSCB) – appear to contribute to the progressive

pathophysiology of SCI, although much remains to be learned about this key mechanism. Using

a clip-compression thoracic SCI model, I characterized the vascular damage and disruption of the

BSCB with the aim of delineating these vascular changes. Female Wistar rats (300-350g)

received a 35 g clip-compression injury at T6-T7. Animals were sacrificed at 1 hour, 4 hours,

and 1, 3, 5, 7, 10, 14 days post-injury. Prior to sacrifice, animals were injected with vascular

tracing dyes: 2% Evans Blue (EB) or FITC-LEA to assess BSCB integrity or vascular

architecture, respectively. Immunohistochemistry was used to verify vascular tracing data.

Spectrophotometry of weight normalized EB showed a dramatic increase in BSCB disruption at

1, 3, and 5 days post-injury compared to uninjured controls (p < 0.01). FITC-LEA identified

functional vasculature was reduced by 24 hours up to 14 days after injury. Similarly, RECA-1

immunohistochemistry showed a significant decrease in the number of vessels observed at 2 and

4 mm from the lesion epicenter at 24 hours post-injury compared to uninjured animals (p < 0.01),

with endogenous re-vascularization showing a slight increase in vessel counts by 10 days post-

95

injury. Separation of the white matter and grey matter quantification showed that grey matter

vessels are more susceptible to SCI, compared to white matter vasculature. Finally, I observed

endogenous angiogenesis following SCI. The angiogenic response spanned between 3 and 7

days post-injury, although maximal endothelial cell proliferation was observed at day 5. These

data indicate that BSCB disruption and endogenous re-vascularization occur at specific time-

points following injury, which may be important for developing effective therapeutic

interventions for SCI.

3.2 Introduction

In general, traumatic spinal cord injury (SCI) results in drastic alterations in spinal cord blood

flow and can cause systemic hypotension. However, at the cellular level, SCI pathology results

in rapid, permanent changes to the structure and function of the microvessels [83, 89, 345]. This

includes loss of microcirculation, disruption to the blood-spinal cord barrier (BSCB), loss of

structural organization, endothelial cell death and vascular remodeling [58, 83]. These changes

have more widespread effects, since vascular damage assists in spreading and enhancing the

secondary injury cascades following SCI. Notably, the breakdown of the BSCB increases the

inflammatory response (allowing inflammatory cells to enter the injury site). Moreover, the

death of endothelial cells, severed vascular networks and ischemia result in apoptosis and cell

death of other CNS cells, since they cannot survive without an adequate blood supply [11, 125,

129, 130].

The major form of vascular regeneration that occurs in injuries/wound-healing is angiogenesis

(although post-natal vasculogenesis has been shown to occur) [227]. Angiogenesis, defined as

96

the production of blood vessels from pre-existing vessels, is a complex process and critical for

the remodeling and survival of tissues following injury [222]. Previous SCI research has

confirmed that a relationship exists between blood vessel density and improved functional

outcomes, therefore, sparing or regenerating vasculature post-injury would be a desirable

outcome [324-326, 346]. These findings are further supported by studies that have demonstrated

that CNS microvessels provide trophic support, and are essential for survival of localized tissue

[291, 302]. Recently, Bearden et al. showed that regenerating axons have a tendency to grow

along blood vessels, suggesting that the vasculature may act as a scaffold and provide guidance

for axonal sprouting following injury [347].

Angiogenesis does occur following injury; however, unfortunately the endogenous mechanisms

cannot provide a sufficient amount of repair. This leaves the injury site in a constant state of

hypoxia-ischemia, leading to further cell death. It has been proposed that following SCI, the

angiogenic response occurs within the first week post-injury, with signs of endogenous

revasularization disappearing shortly after that [89, 348]. Logically, novel vascular therapies

would aim to target early vascular mechanisms (between 3 and 7 days) to maximally improve the

local vasculature and blood supply, in turn, reducing the amount of cell loss and neurological

deficits. Moreover, residual vessels or newly generated vessels that are damaged and/or

immature present an opportunity for adverse physiological events; notably providing a direct

route to the lesion site for circulating inflammatory mediators. Since the BSCB is disrupted, this

allows a large influx of inflammatory cells to enter into an otherwise immune-protected CNS.

97

Previous characterizations of the BSCB and vascular disruption

Previous groups have characterized and investigated many of the cellular events following SCI

[129, 349-351]. Additionally, previous work has been published which describes some of the

vascular changes that result from SCI, including reduction in blood flow, disruption of the blood-

spinal cord barrier (BSCB) and loss of structural organization [58, 83, 89, 97, 323, 345, 352,

353]. However, our study investigates a number of novel aspects compared to other SCI

vascular studies: 1) In a model of clip-compression injury, BSCB permeability has not been

examined. 2) BSCB disruption and blood vessels have been examined separately in previous

studies. In my research I examine both aspects of vascular injury, ideally providing a more

detailed assessment of the damage following SCI. 3) In clip-compression injury, vascular

studies have not extended past 24 hours. 4) The clip-compression injury offers an alternative

SCI model for research, and may exhibit varied vascular profiles compared to contusion or

transection models. 5) The endogenous angiogenic response following clip-compression injury

has not been examined.

To the best of our knowledge, previous studies have only examined the cellular changes in that

occur in “mild” spinal cord injury; the NYU impactor model has been commonly used [66, 97,

129, 354, 355], or in transection models [352, 353]. Other studies that have used the clip-

compression model have examined blood flow, hemorrhage, blood gases; however, none have

investigated the cellular profiles, nor the angiogenic response. Moreover, the majority of these

studies have examined only acute vascular changes (≤ 24 hours), whereas we have examined up

to 14 days post-injury. As previously mentioned, the clip-compression model of SCI (as used by

the Fehlings’ laboratory) presents an alternative SCI model that offers differences compared to

other models, ultimately helping to mimic the diversity and heterogeneity observed in the clinic.

98

The clip-compression model has a number of key advantages over other models of SCI, which

may more accurately portray the human condition. Of note, the clip-compression injury results

in both dorsal and ventral damage of the cord, and the clip-compression model also creates

temporary ischemia and impaired blood flow, which contribute to secondary pathology of the

injury, and are commonly observed in man.

This study aims to examine the vascular changes that result from a moderately-severe injury,

and determine the temporal and spatial profile of these changes. In order to critically assess the

outcomes of any therapeutic intervention (and in my studies, a ZFP-VEGF gene therapy), it is

imperative that we first understand the vascular changes involved in the clip-compression model

of SCI. Although previous research has investigated the vascular changes of mild spinal cord

injuries, the amount of damage and disruption to microvascular structures is anticipated to be

more extensive and will therefore result in different characteristics and outcomes within the

injury epicentre and penumbra. This research presents novelty in describing the vascular

changes in an alternate, well-utilized model of SCI: the clip-compression model. Additionally,

this study employs some novel techniques for assessing vasculature, specifically highlighting the

differences in perfused and non-perfused vessels. Importantly, the results of this study provide

key insights into the dynamic vascular alterations and endogenous repair that occur following

SCI. Take together, these data may help elucidate ideal time-points and spatial areas that could

maximize the effectiveness of therapeutics aiming to target the spinal vasculature following

injury.

99

3.3 Objective

Characterize the structural and functional changes to the spinal cord vasculature following

moderately-severe clip-compression spinal cord injury.

3.4 Hypothesis

It is hypothesized that a moderately-severe SCI will result in significant disruption to the BSCB,

and a significant loss of structure and function to the vasculature at the epicenter and adjacent

spinal levels.

3.5 Specific Aims

1. Determine the temporal profile of BSCB disruption following SCI.

2. Determine the temporal progression of vascular damage following SCI.

3. Examine the spatial distribution of vascular damage following SCI, specifically the

rostrocaudal distribution and the grey vs. white matter changes.

4. Investigate the spatio-temporal progression of endogenous angiogenesis of the spinal

cord post-SCI.

100

3.6 Methods

Animal Model of SCI

All animal protocols (979.31 and 891.9) and procedures were approved by the Animal Care

Committee at the University Health Network, Toronto, ON, Canada.

Animals were subject to a contusive-compressive spinal cord injury using a modified aneurysm

clip, which has been extensively characterized by the Fehlings’ laboratory and previously

described [74]. Briefly, adult female Wistar rats (250-300g; Charles River, Montreal, Canada)

were deeply anesthetized using 4% isoflurane, and were sedated for the remainder of the surgery

under 2% isoflurane. Animals received a two-level laminectomy of mid-thoracic vertebral

segments T6-T7. A modified clip calibrated to a closing force of 35g was applied extradurally to

the cord for 1 minute and then removed (Figure 14). The incision was closed in layers using

standard silk sutures and animals were given a single dose of buprenorphine (0.05 mg/kg).

Animals were allowed to recover in their cage under a heat-lamp and, subsequently, were housed

in a temperature-controlled warm room (26°C) with free access to food and water. Animals were

given buprenorphine (0.05 mg/kg) every 12 hours for 48 hours following surgery, and their

bladders were manually voided three times daily. A subcutaneous injection of 10mg/kg of

cyclosporin-A was administered daily starting 24 hours prior to the SCI until the end of the

experiments for immunosuppression. The number of animals used in each experiment is

outlined in Table 5. Final animal numbers in each group are slightly varied due to unexpected

mortalities during the experiments.

101

Evans Blue: Blood-Spinal Cord Barrier Disruption

Animals were injected with 1 mL of 2% Evans Blue (EB) into the tail vein [335, 336]. EB was

allowed to circulate for 20-30 minutes and then the animals were transcardially perfused with

saline. One cm of the spinal cord surrounding the injury site was extracted, weighed, and snap-

frozen in dry ice. Samples were then homogenized in 400 μl of N,N’-dimethylformamide

(DMF) and incubated at 50°C for 72 hours. Samples were centrifuged at 18,000 rpm for 30

minutes. The supernatant was collected, aliquoted into a 48 well glass plate, and colorimetric

measurements were performed using a Perkin Elmer Victor3 1420 spectrophotometer at the

absorption maximum for EB (620 nm). Samples were normalized to the original sample weight,

and EB concentration was calculated based on a standard curve of EB in DMF (data reported as

EB per spinal cord weight: μg/g).

Histochemistry

Histological Processing. Following deep inhalational anesthetic (isoflurane), animals were

transcardially perfused with 4% paraformaldehyde (PFA) in 0.1 M PBS. Then, the tissues were

cryoprotected in 20% sucrose in PBS. A 10 mm (1 cm) length of the spinal cord centered at the

injury site was fixed in tissue-embedding medium. The tissue segment was snap frozen on dry

ice and sectioned on a cryostat at a thickness of 14 μm. Serial spinal cord sections at 500 μm

intervals were stained with myelin-selective pigment Luxol Fast Blue (LFB) and the cellular

stain Hematoxylin-Eosin (HE) to identify the injury epicenter. Tissue sections showing the

largest cystic cavity and greatest demyelination were taken to represent the injury epicenter.

Rostrocaudal spinal cord maps were created by calculating the distance from the epicenter to the

corresponding tissue sections.

102

Immunohistochemistry. Mouse anti-RECA-1 (1:25; Serotec Inc., Raleigh, NC, USA) and rabbit

anti-Ki67 (1:1000; Abcam, Toronto, ON, Canada) were used to stain endothelial cells and

proliferating cells, respectively (See Table 4). The sections were rinsed three times in PBS after

primary antibody incubation and incubated with either fluorescent Alexa 568, 647 or 488 goat

anti-mouse/rabbit secondary antibody (1:400; Invitrogen, Burlington, Canada) for 1 hour. The

sections were rinsed three times with PBS and cover slipped with Mowiol mounting medium

containing DAPI (Vector Laboratories, Inc., Burlingame, CA) to counterstain the nuclei. The

images were taken using a Leica MZ FLIII epifluorescence microscope (Leica Microsystems,

Richmond Hill, ON, Canada).

Identification of Blood Vessels. At 1 hour, 4 hours, and 1, 3, 5, 7, 10, 14 days post-injury

animals were sedated with inhalational anesthetic (isoflurane) and the right femoral vein was

surgically exposed (Figure 17). 0.5 mg of FITC-LEA diluted in saline to a final volume of 1 mL

was injected into the femoral vein and was allowed to circulate for 20-30 minutes (FITC-

conjugated Lycopersicon esculentum agglutinin; Cat # L0401, Sigma, Oakville, ON, Canada).

Animals were transcardially perfused with 4% paraformaldehyde and tissues were fixed and

processed as described above. Tissue sections at the injury epicenter, and 1000 μm, 500 μm, and

250 μm (rostral and caudal) from the epicenter were used to assess the spatial distribution of

vascular damage. Tissue sections were stained with RECA-1 (1:25; Serotec Inc., Raleigh, NC,

USA) to identify endothelial cells/blood vessels. Vessels which were co-labelled with FITC-

LEA and RECA-1 were identified as “functional”, since the presence of FITC-LEA indicated

these vessels were connected to the systemic vasculature and exhibited a perfusion state.

Vascular quantification was performed on four selected fields (ventral horn, dorsal horn, left and

right lateral columns) in each section under 25X magnification (0.14 mm2) (Figure 16). The data

103

are presented, separating the white matter and grey matter, since these regions showed distinct

variation in their vascular responses.

Quantification of Angiogenesis. Tissue sections – taken from animals sacrificed 1 hour, 4 hours,

and 1, 3, 5, 7, 10, 14 days post-injury – were used to quantify angiogenesis following SCI.

Angiogenesis was calculated as vessels co-labelled with RECA-1 and Ki67 (cellular

proliferation). As shown in Figure 16, the counting of angiogenesis was performed on 4 selected

fields (ventral horn, dorsal horn, left and right lateral columns) in each section under 25X

magnification (0.14 mm2). The number of angiogenic vessels was calculated at 500 μm, 1, 2 and

3 mm, both rostral and caudal from the epicenter, for each animal. Rostral and caudal values

were pooled for each distance.

Statistical Analysis

Data were analyzed with SigmaPlot software (Systat Software Inc., San Jose, California, USA).

For data that investigated the percentage of cells, the data were subjected to an arcsine

transformation prior to statistical analysis to attain a more normal distribution. For comparison

of groups sampled at various distances from the injury site I used two-way analysis of variance

(ANOVA) with repeated measures, followed by the post-hoc Holm-Sidak test. In all figures, the

mean value ± SEM are used to describe the results. Statistical significance was accepted for p

values of <0.05.

104

Table 5. Animals used in Chapter 3 experiments.

Experiment Group Original Animal # Final Animal #

Evans Blue Sham 6 6 (Figure 18) 1 hour 10 8

4 hour 10 8

24 hour 10 8

3 day 10 10

5 day 10 10

7 day 10 8

10 day 10 10

14 day 10 9

RECA/FITC-LEA Sham 5 5 (Figure 19, 20, 21; Table 5) 1 hour 5 5

4 hour 5 5

24 hour 5 5

3 day 5 4

5 day 5 5

7 day 5 5

10 day 5 4

14 day 5 5

Ki67/RECA-1 Sham 5 5 (Figure 22) 1 hour 5 5

4 hour 5 5

24 hour 5 5

3 day 5 4

5 day 5 5

7 day 5 5

10 day 5 4

14 day 5 5

105

3.7 Results

3.7.1 BSCB permeability following SCI

A key component in understanding the disruption of the vasculature following SCI is to

investigate the vascular/BSCB permeability. The opening of the BSCB results in negative events

that propagate SCI pathology – including inflammation, edema, hemorrhage and homeostatic

dysregulation – however, we may be able to take advantage of BSCB breakdown, since it may

allow for certain drugs/therapies to enter the spinal cord, which may otherwise be excluded from

the highly regulated CNS environment. We hypothesized that elucidating the BSCB

permeability following SCI may offer important information that may identify the most optimal

time-window for therapeutic intervention. Here, we have examined the temporal changes to

BSCB permeability: from 1 hour post-injury to 14 days post-injury. We observe that the BSCB

is disrupted very early on (as early as 1 hour following injury), maximally disrupted at 24 hours

post-SCI, and appears to be restored by 14 days (Figure 18). Therefore, we suggest that there is

a relatively large time-window to exploit on the disruption of the BSCB. In a model of clip-

compression injury, therapies administered between 1 hour and 5 days (and most notably at 24

hours post-injury), may have an added advantage in reaching their CNS targets.

106

Figure 18. Blood-spinal cord barrier permeability following traumatic SCI. Evans Blue

extravasation between 1 hour and 14 days post-SCI. Evans Blue (EB) extravasation following

SCI was quantified by spectrophotometry at 630 nm. EB concentrations were calculated from a

standard curve using concentrations between 0 and 50 ng. Tissue samples were normalized to

their wet weight. Values are shown as the mean ± SEM . N-values are displayed in the figure

legend. One-way ANOVA, Dunn’s post-hoc. * p < 0.05, compared to Sham.

107

3.7.2 Spatial-temporal disruption of the vasculature

The next objective was to delineate the distribution of the vascular damage following clip-

compression injury. Not surprisingly, I detected a significant disruption to the localized

vasculature (Figures 19 and 20). The epicenter of the injury is most significantly disturbed, with

the adjacent sections exhibiting less damage. RECA-1 (Figure 19A) and FITC-LEA (Figure

19B) counts are both significantly below sham animals at all time points, spanning 500 µm

rostrocaudal (p < 0.001). Interestingly, the vascular changes appear to stay relatively confined,

spanning approximately 2 mm rostrocaudal from the injury epicenter (Figure 19, Table 6).

Adjacent sections to the epicenter appear to show vascular recovery (particularly more distal

areas) in comparison to the lesion epicenter, which remains significantly altered following injury.

Overall, it the penumbra of vascular damage remains relatively small (spanning approximately 1

mm by 14 days), suggesting that vascular therapies may only need to target a localized area to be

effective.

To assess the functionality of the vasculature following injury, I compared the number of vessels

marked by in vivo tracer (FITC-LEA) and ex vivo histological RECA-1 staining (Figure 19). I

observed that RECA-1 counts were higher than FITC-LEA-positive vessels, indicating that

although vascular components may be preserved post-SCI, the perfusion of vessels is

diminished. Vessels that are RECA-1-positive, but not FITC-LEA-postive, may be physically

blocked (by blood clots), disconnected from the blood supply (as a result of mechanical trauma),

or new vessels that have not yet been connected to the vascular network (immature or angiogenic

vessels).

108

109

Figure 19. Spatial-temporal disruption of the spinal cord vasculature following clip-

compression injury. A) Spatio-temporal comparison of RECA-1 quantification between sham

animals and injured animals. Animals were sacrificed at 1 hour, 4 hours, 1, 3, 5, 7, 10 and 14

days post-SCI and tissues were examined at the epicenter, 250 μm, 500 μm, and 1000 μm rostral

and caudal to the injury epicenter. Significant disruption of the vasculature was observed as

early as 1 hour post-injury, with the epicenter and adjacent areas being most significantly

affected. B) Spatio-temporal comparison of FITC-LEA quantification between sham animals

and injured animals. FITC-LEA was allowed to circulate for approximately 20 minutes prior to

animal sacrifice (tagging the vasculature internally) and indicated the vessels which had

maintained blood perfusion following injury. Animals were sacrificed at 1 hour, 4 hours, 1, 3, 5,

7, 10 and 14 days post-SCI and tissues were examined at the epicenter, 250 μm, 500 μm, and

1000 μm rostral and caudal to the injury epicenter. C) Representative images taken from the

dorsal grey matter at 500 μm rostral to the epicenter at 10 days post-injury. White arrowheads

indicate vessels that are labeled with RECA-1, but not FITC-LEA. Lack of double-labeling

indicates that vascular structures (i.e. endothelial cells) are present; however, the blood vessel did

not have an active perfusion state. ** p < 0.001. Scale bar = 100 μm. Green = FITC-LEA. Red

= RECA-1. Blue = DAPI. n= 4-5 animals/time point (see Table 5).

110

Table 6. Spatial and temporal data from FITC-LEA and RECA-1 analysis.

Time Point

FITC-LEA + Vessel Counts

RECA-1 + Vessel Counts

FITC vs. RECA Counts

(%)

% FITC/RECA in Grey Matter

% FITC/RECA in White Matter

Sham Animals 265 ± 8.8 269 ± 11.4 98.7 98.5 99.6

1 hour post-injury

1000 µm Caudal 179 ± 19.3 195 ± 16.2 91.9 92.8 89.0

500 µm Caudal 141 ± 14.8 202 ± 34.7 69.8 68.3 76.1

250 µm Caudal 131 ± 2.8 217 ± 32.3 60.5 60.2 61.8

Epicenter 63 ± 14.4 138 ± 19.8 45.8 46.7 42.5

250 µm Rostral 94 ± 12.9 183 ± 34.5 51.2 44.2 74.6

500 µm Rostral 114 ± 29.9 190 ± 15.3 60.0 58.3 65.4

1000 µm Rostral 165 ± 15.3 198 ± 8.1 83.4 83.0 85.3

4 hours post-injury

1000 µm Caudal 125 ± 20.9 150 ± 28.9 83.6 78.9 94.5

500 µm Caudal 83 ± 15.5 125 ± 7.0 66.7 59.4 87.6

250 µm Caudal 68 ± 6.2 122 ± 18.2 55.2 46.9 80.8

Epicenter 27 ± 8.1 81 ± 5.7 33.8 27.0 49.0

250 µm Rostral 87 ± 21.6 130 ± 33.3 66.8 57.5 89.5

500 µm Rostral 93 ± 22.7 114 ± 26.7 81.7 80.7 83.7

1000 µm Rostral 110 ± 22.4 141 ± 24.9 78.3 73.5 89.1

24 hours post-injury

1000 µm Caudal 144 ± 13.4 208 ± 12.4 68.7 68.1 70.9

500 µm Caudal 129 ± 16.3 191 ± 21.0 67.3 68.6 64.1

250 µm Caudal 112 ± 7.3 188 ± 13.3 59.2 54.0 74.9

Epicenter 57 ± 5.9 135 ± 26.9 41.9 23.4 78.5

250 µm Rostral 138 ± 13.2 223 ± 33.0 61.8 55.2 80.8

500 µm Rostral 149 ± 21.5 197 ± 26.3 75.5 72.5 85.6

111

1000 µm Rostral 200 ± 36.9 225 ± 30.6 88.9 88.9 88.8

3 days post-injury

1000 µm Caudal 191 ± 22.4 218 ± 10.2 87.2 85.5 92.2

500 µm Caudal 194 ± 44.0 215 ± 44.2 89.8 87.4 97.0

250 µm Caudal 131 ± 15.4 160 ± 31.5 81.9 78.9 89.0

Epicenter 112 ± 15.3 143 ± 15.5 78.4 73.3 88.8

250 µm Rostral 106 ± 14.8 139 ± 21.8 76.7 70.6 91.0

500 µm Rostral 121 ± 25.9 141 ± 23.4 85.8 81.1 95.2

1000 µm Rostral 185 ± 20.4 211 ± 21.0 87.8 85.9 94.6

5 days post-injury

1000 µm Caudal 163 ± 21.2 234 ± 24.7 69.7 68.6 74.0

500 µm Caudal 117 ± 25.4 183 ± 27.3 63.8 60.6 72.8

250 µm Caudal 106 ± 9.8 195 ± 26.8 54.4 48.5 77.3

Epicenter 72 ± 16.0 124 ± 16.3 58.1 53.0 68.5

250 µm Rostral 93 ± 19.1 158 ± 18.5 58.8 49.0 86.3

500 µm Rostral 121 ± 28.9 189 ± 18.7 63.8 58.9 76.3

1000 µm Rostral 174 ± 28.2 232 ± 42.7 75.0 68.0 94.5

7 days post-injury

1000 µm Caudal 228 ± 12.8 250 ± 12.7 91.2 90.9 92.2

500 µm Caudal 213 ± 14.4 243 ± 15.2 87.4 86.0 90.9

250 µm Caudal 153 ± 4.8 207 ± 8.7 74.0 66.7 93.0

Epicenter 83 ± 9.0 158 ± 20.8 52.7 35.2 85.5

250 µm Rostral 81 ± 15.1 124 ± 8.7 65.0 51.7 90.6

500 µm Rostral 137 ± 16.7 169 ± 13.3 81.0 77.2 91.6

1000 µm Rostral 241 ± 30.1 251 ± 31.5 95.8 94.8 99.2

10 days post-injury

1000 µm Caudal 191 ± 8.7 209 ± 11.3 91.4 90.7 94.4

112

500 µm Caudal 132 ± 24.7 141 ± 21.3 93.8 93.2 95.5

250 µm Caudal 86 ± 14.1 112 ± 6.6 77.4 71.8 88.5

Epicenter 49 ± 1.8 76 ± 4.2 64.7 49.2 89.7

250 µm Rostral 141 ± 26.9 153 ± 26.9 92.1 90.1 97.6

500 µm Rostral 187 ± 27.2 203 ± 26.9 92.0 91.1 96.4

1000 µm Rostral 231 ± 20.6 236 ± 23.1 97.2 96.4 100.0

14 days post-injury

1000 µm Caudal 248 ± 8.1 269 ± 10.8 92.0 91.3 94.6

500 µm Caudal 207 ± 21.3 227 ± 23.1 91.1 90.1 94.8

250 µm Caudal 213 ± 15.5 230 ± 16.6 92.5 91.8 94.5

Epicenter 148 ± 20.8 171 ± 31.5 86.5 84.6 93.0

250 µm Rostral 196 ± 13.0 217 ± 11.6 90.5 89.1 94.1

500 µm Rostral 216 ± 13.6 231 ± 13.7 93.5 94.8 90.2

1000 µm Rostral 252 ± 19.6 263 ± 19.6 95.8 95.9 95.4

Table 6. Spatial and temporal data from FITC-LEA and RECA-1 analysis. Values are

provided from the data presented in Figure 19, Figure 20 and Figure 21. The table shows FITC-

LEA and RECA-1 counts within the spinal cord (columns 1 and 2, respectively), and displays

ratios of FITC-LEA/RECA-1 staining observed overall (column 3), in the grey matter (column 4)

and in the white matter (column 5). Values are provided for each time point between 1 hour and

14 days post-SCI, and the spatial quantification within each time point is also shown (between

1000 µm rostral and 1000 µm caudal to the injury site). Data are presented as either the mean ±

SEM (columns 1 and 2), or as percentages (columns 3-5). n= 4-5 animals/time point (see Table

5).

113

Since the vasculature of the grey and white matter varies drastically in an uninjured spinal cord,

it was of interest to investigate the two regions separately to determine if one area was more

vulnerable to vascular disruption. The results indicate that both grey and white matter

vasculature is disrupted following SCI; however, I observe that the grey matter vessels (Figure

20, Table 6) are more susceptible to interruptions in blood flow following SCI, compared to

white matter vessels (Figure 21, Table 6). Moreover, I observe that the white matter vasculature

endogenously recovers to over 90% of an uninjured cord by 5 days post-injury, and restoration of

vascular flow occurs much quicker (over 70% of vessels have a perfusion state by 24 hours post-

SCI). In the white matter, points between 500 µm rostral and 500 µm caudal are significantly

disrupted at 1 hour following injury, and only the epicenter is significantly disrupted at 4 hours

post-SCI (p < 0.001). In the grey matter, the vasculature recovers much slower and does not

recover to the same extent. Between 500 µm rostral and 500 µm caudal, a significant disruption

is observed from 1 hour and extends to 10 days post-SCI (p < 0.001). By 10 days post-SCI, less

than 50% of the vessels are perfused; however, by 14 days an average of 84% of vessels show a

perfusion state. Here I report the percentage of FITC-LEA labeled vessels relative to the number

RECA-1-positive vessels, indicating a ratio of perfused (or “functional”) of the vessels. To my

knowledge no previous studies have specifically distinguished between white and grey matter

vascular disruption in a detailed spatio-temporal analysis, and these findings have important

clinical and therapeutic implications, which will be discussed later in the discussion.

114

Figure 20. Vascular disruption of the grey matter following traumatic SCI. Spatio-

temporal quantification of vessel perfusion in the grey matter. Data shown are the % of FITC-

LEA vessels compared to RECA-1-positive vessels, at various distances from the epicenter (0 to

1000 μm rostral and caudal), as well as various time-points following injury (1 hour to 14 days

post-SCI). A drastic reduction in vascular perfusion is observed as early as 1 hour following

injury, and a partial restoration is observed by 14 days post-SCI. The epicenter and areas

directly adjacent to the epicenter show the most disruption, whereas more distal sections appear

less affected. n= 4-5 animals/time point (see Table 5). ** p < 0.001.

115

Figure 21. Vascular disruption of the white matter following traumatic SCI. Spatio-

temporal quantification of vessel perfusion in the white matter. Data shown are the % of FITC-

LEA vessels compared to RECA-1-positive vessels, at various distances from the epicenter (0 to

1000 μm rostral and caudal), as well as various time-points following injury (1 hour to 14 days

post-SCI). A drastic reduction in vascular perfusion is observed as early as 1 hour following

injury; however, the vasculature appears 95% restored by 24 hours post-SCI. The epicenter and

areas directly adjacent to the epicenter show the most disruption, whereas more distal sections

appear less affected. Compared to the grey matter vasculature, vascular disruption in the white

116

matter is confined to a narrower rostrocaudal distribution. n= 4-5 animals/time point (see Table

5). ** p < 0.001.

3.7.3 Endogenous Angiogenesis Occurs Following SCI

Very few studies have examined the endogenous angiogenic response following SCI [58, 89,

149]. Moreover, none of them have investigated such a response using a model of clip-

compression injury, or extensively examined the spatio-temporal angiogenic response, as

described in this research. In these studies, I observed a dynamic angiogenic response initiated

following a moderately-severe clip-compression injury. I quantified proliferating endothelial

cells (RECA-1/Ki67) to assess endogenous angiogenesis. Results showed angiogenesis

occurring as early as 3 days following injury and ending by 7 days, with maximal angiogenesis

occurring at 5 days post-SCI (Figure 22) (p < 0.001). At the peak of angiogenesis (5 days post-

SCI), I noted that 1 mm distal to the epicenter showed the most proliferating endothelial cells,

with approximately 15% of vessels marked as Ki67-positive (p < 0.001).

117

Figure 22. Endogenous angiogenic response after traumatic thoracic SCI. A)

Representative Ki67/RECA-1 double-label imaging. Image taken at 1000 μm caudal from an

animal sacrificed at 5 days post-SCI. B) Spatial and temporal quantification of the angiogenic

response following SCI. Uninjured spinal cord tissue showed relatively low basal levels of

endothelial cell proliferation (1%) compared to days 3, 5, and 7 following injury (6-15%).

Maximal proliferation was observed at 5 days post-SCI, which angiogenesis occurring most

notably around 1000 μm distal to the injury site. The epicenter did exhibit some vascular

proliferation; however, due to a diminished number of vessels present, the ratio of regenerating

118

endothelial cells was notably less than more distal areas. Scale bar = 100 μm. n= 4-5

animals/time point (see Table 5). * p < 0.001.

3.8 Discussion

In the present study, I aimed to identify some of the major vascular alterations that occur

following spinal cord injury. I examined the temporal and spatial loss of vasculature, as well as

the perfusion of the vasculature and the vascular permeability. I observed that the BSCB is

significantly disrupted early on, but appears to partially recover by 14 days post-SCI.

Additionally, I noted that significant vascular loss and function occurs, and grey matter vessels in

particular are more affected and less likely to recover following SCI. This is consistent with

previous research done using a clip-compression model [356, 357]. Lastly, I showed that an

endogenous angiogenic response does occur in the spinal cord following clip-compression

injury, and consistent with other reports, I observe maximal angiogenesis between 3 and 7 days

post-injury [89].

This study shows that the BSCB is disrupted as early on as 1 hour post-injury, and remains open

until 5 days post-injury, with maximum permeability observed at 24 hours post-SCI (Figure 18).

Therefore, it appears that there is a relatively large time-window to exploit on the disruption of

the BSCB. In a model of clip-compression injury, therapies administered between 1 hour and 5

days (and most notably at 24 hours post-injury), may have an added advantage in reaching their

CNS targets; potentially making them more effective compared to an administration time where

an intact BSCB exists.

119

The time-course of BSCB dysfunction from a clip-compression model of SCI (as shown by my

research) coincide with previous reports by Noble and Wrathall, and Popovich et al., which used

weight-drop or transection models of SCI [66, 97, 352, 353]. In this study, I used Evans Blue as

a marker for spinal cord vascular permeability, which has benefits and disadvantages for

assessing BSCB disruption. Evans Blue is a convenient, simple dye to administer and quantify;

however, it binds to serum albumin making it a 70 kDa protein, which is considered a large

molecule [335, 336]. The previous studies have used horseradish peroxidase (HRP), which is

approximately 45 kDa, or α-aminoisobutyric acid (AIB), which is only 0.1 kDa. In comparison,

Evans Blue assays may be less sensitive to detecting minor disruptions to the BSCB, as it is too

large to leak out. Therefore, it may be beneficial to repeat this study using a smaller vascular

tracer to improve the detection of vascular permeability. With that said, Noble and Wrathall

used HRP as an in vivo marker, and showed restoration of the BSCB by 14 days post-SCI [97].

An additional caveat of using in vivo tracers to detect BSCB disruption is the inherent issue of

reduced or obstructed blood flow following injury. By using a circulating vascular marker,

results may be under-represented since the tracer may be physically restricted from reaching

certain vascular networks (by blood clots and/or broken/severed vessels). In future studies,

histological examination of BSCB proteins and structure may be a complimentary approach in

assessing vascular permeability.

BSCB disruption following injury is a double-edged sword. On one hand, an increase in

vascular permeability creates an ideal opportunity for the influx of inflammatory mediators and

proteins not usually permitted in the CNS. Conversely, a compromised BSCB provides a unique

opportunity for therapeutic intervention, as getting drugs/molecules/etc. into the highly regulated

CNS normally presents a substantial challenge. In the model of clip-compression SCI, I observe

120

the BSCB remains significantly disrupted for up to 5 days following injury, which is a clinically

relevant therapeutic window for administering treatments.

Grey matter vessels are typically smaller vessels or capillaries, therefore it is not surprising that

they (being the most distal structures) would be more affected. However, diminished grey matter

perfusion results in hypoxia-ischemia to neuronal cell bodies, and if a decreased blood flow

persists, neurons will undergo cell death. From a therapeutic and regenerative medicine

prospective, it is therefore critical to re-vasularize the grey matter of the cord following trauma.

More importantly, it appears that the extent of the vascular damage is restricted to a relatively

small area (spanning a total of 2 mm rostral and caudal), therefore effective vascular therapies

that are administered locally are likely to provide adequate repair.

In the present study, I investigated the perfusion of the vasculature using an in vivo FITC-LEA

dye. I believed that distinguishing “functional” vessels from other vessels was an important area

of research, especially since it has not been specifically addressed in previous characterizations

of vascular damage post-SCI. In comparing FITC-LEA and RECA-1 counts, I noted a decreased

number of FITC-LEA vessels, suggesting a reduced number of vessels connected to the blood

stream. Although this may not be a surprising result, it is nevertheless, an important finding. In

many studies (including the ones in Chapter 4 of this thesis), vessels and vascular regeneration

are often quantified by histological analysis. Histological assessment (i.e. RECA-1) quantifies

all vascular structure without distinguishing if the vessel had an active perfusion state. FITC-

LEA marks the luminal surface of endothelial cells, which have a hemodynamic state. With this,

it is important to note that FITC-LEA may misrepresent the number of “functional” vessels, as

some newly formed vessels (depending on lumen size) may be plasma-perfused without

121

supporting cellular perfusion, and therefore are not truly acting as “functional” vessels. While

this technique provides some useful information, based on the results of this study we

recommend that future research be cautious when interpreting the results since the values may be

over-estimated.

Considering the extent of research dedicated to promoting vascular regeneration over the past

decade, it was interesting to note that very few studies had focused on characterizing the

vasculature of the injury itself, and no studies have been completed using the clip-compression

model. In order to investigate the effectiveness of vascular therapies, it seems logical that I first

understand the endogenous events, which could be used as a baseline for subsequent outcomes.

With that, I believe that this research significantly contributes to the field of vascular

regeneration following spinal cord injury, and will aid in designing and administering more

effective vascular therapies for neurotrauma.

3.9 Conclusions

As observed by Evans Blue, the BSCB exhibits FITC-LEA identified functional vasculature was

reduced by 24 hours up to 14 days after injury. Similarly, RECA-1 immunohistochemistry

showed a significant decrease in the number of vessels observed at 2 and 4 mm from the lesion

epicenter at 24 hours post-injury compared to uninjured animals (p < 0.01), with endogenous re-

vascularization showing a slight increase in vessel counts by 10 days post-injury. Separation of

the white matter and grey matter quantification showed that grey matter vessels are more

susceptible to SCI, compared to white matter vasculature, suggesting that vascular therapies

122

should strive to target smaller, deeper vascular repair within the spinal cord. Finally, I observed

endogenous angiogenesis following SCI. The angiogenic response spanned between 3 and 7

days post-injury, although maximal endothelial cell proliferation was observed at day 5. These

data, consistent with other research in the field, suggests that angiogenic therapies should be

administered before 5 days post-injury to complement and enhance endogenous

revascularization.

123

Chapter 4

4 Delayed AdV-ZFP-VEGF Administration Provides Neuroprotection and Promotes Angiogenesis Post-SCI

4.1 Abstract

Following spinal cord injury (SCI) there are drastic changes that occur in the spinal

microvasculature, including ischemia, hemorrhage, endothelial cell death and blood-spinal cord

barrier disruption. Vascular endothelial growth factor-A (VEGF-A) is a pleiotropic factor

recognized for its pro-angiogenic properties; however, VEGF has recently been shown to

provide neuroprotection. It was hypothesized that delivery of AdV-ZFP-VEGF – an adenovirus

that generates a bio-engineered zinc-finger transcription factor that promotes endogenous VEGF-

A expression – would result in angiogenesis, neuroprotection and functional recovery following

SCI. This novel VEGF gene therapy induces the endogenous production of multiple VEGF-A

isoforms; a critical factor for proper vascular development and repair. Briefly, female Wistar

rats – under cyclosporin-A immunosuppression – received a 35g clip-compression injury and

were administered AdV-ZFP-VEGF or AdV-eGFP at 24 hours post-SCI. Tissues were extracted

at 3, 5 or 10 days post-SCI. qRT-PCR and Western Blot analysis of VEGF-A mRNA and

protein, showed significant increases in VEGF-A expression in AdV-ZFP-VEGF treated animals

(p<0.02). Analysis of NF200, TUNEL, RECA-1 and Ki67 indicated that AdV-ZFP-VEGF

increased axonal preservation (p<0.05), reduced cell death (p<0.01), increased blood vessels

(p<0.01), and increased angiogenesis (p<0.001) respectively. Overall, the results of this study

124

indicate that AdV-ZFP-VEGF administration can be delivered in a clinically relevant time-

window following SCI (24 hours) and provide significant molecular and functional benefits.

4.2 Introduction

In North America, it is estimated that approximately 1.5 million individuals are currently living

with SCI, with over 12,000 new cases occurring each year [2]. Spinal cord injury is divided into

two events, to separate the physical and the cellular pathologies. The primary injury, is

associated with the initial mechanical trauma that the cord undergoes, whereas the secondary

injury refers to the physiological cascade that propagates from 1 minute to 6 months following

the initial injury [3]. Although the primary injury is responsible for triggering all of the

downstream events, it is widely accepted that the processes that take place in the “secondary

injury” phase are predominantly responsible for a significant portion of the damage and

degeneration that is associated with SCI, including inflammation, ischemia, lipid peroxidation,

production of free radicals, disruption of ion channels, necrosis and programmed cell death [5,

56, 57]. Moreover, radical alterations to the spinal microvascular architecture and function occur

following SCI and contribute to the secondary injury. Reduction in blood flow, hemorrhage,

systemic hypotension, loss of microcirculation, disruption of the blood-spinal cord barrier

(BSCB) and loss of structural organization, ultimately enhance the cellular damage post-injury

[3, 58]. Despite the fact that these secondary events are responsible for the majority of the

damage associated with SCI, many of these pathways alternatively provide an opportunity to

target with therapeutic interventions.

125

Recently, research has given much attention to therapies designed at repairing or minimizing

vascular damage following injury. Angiogenic factors, such as vascular endothelial growth

factor (VEGF)-A, are known to promote the proliferation of endothelial cells and initiate

angiogenesis [241]. Emerging evidence suggests that VEGF-A (which will be referred to as

VEGF) also has neurotrophic, neuroprotective, and neuroproliferative effects [240]. VEGF is a

homodimeric glycoprotein that is expressed as multiple splice variants encoded by a single gene;

however, VEGF signals as a homo- or heterodimer via VEGF receptors (VEGFRs) [244]. The

predominant isoforms in the central nervous system are VEGF121, VEGF165 and VEGF189.

Studies have demonstrated that VEGF and its receptors are upregulated during and after

hypoxic/ischemic injury to the brain and spinal cord, which suggests that VEGF likely plays a

neuroprotective (or beneficial) role in these pathophysiological processes.

Previously described approaches using VEGF have relied on the introduction of a single splice

isoform of VEGF-A (VEGF165), which may not result in optimal neuroprotective or angiogenic

effects. In this study, I utilized novel ZFP-VEGF technology – a viral vector encoding a zinc-

finger transcription factor protein (ZFP), which activates endogenous VEGF-A expression to

produce multiple splice isoforms of VEGF – which has previously demonstrated induced

expression of VEGF-A protein, increase vascular counts and significant functional recovery

following SCI [212]. Although our research group has already shown beneficial effects of AdV-

ZFP-VEGF when administered immediately following SCI as a proof-of-concept, the current

study aims to investigate a clinically-relevant administration of AdV-ZFP-VEGF by delaying

administration by 24 hours post-SCI.

126

4.3 Objective

Examine the acute and cellular therapeutic effects of delayed AdV-ZFP-VEGF administration

following spinal cord injury.

4.4 Hypothesis

It was hypothesized that ZFP-VEGF gene therapy will enhance molecular recovery following

spinal cord injury through neuroprotective and angiogenic mechanisms.

4.5 Specific Aims

1. Confirm in vivo transduction of AdV-eGFP into the spinal cord.

2. Examine the specific cell types transduced by AdV-eGFP.

3. Assess the acute neuroprotective effects of AdV-ZFP-VEGF following SCI.

4. Investigate the in vivo vascular and angiogenic effects of AdV-ZFP-VEGF administration

post-SCI.

127

4.6 Methods

All animal experiments were conducted with approval from the Animal Care Committee,

University Health Network (Toronto, Canada).

Viral Vector Constructs

The VEGF-A-activating ZFP and controls were provided in viral vectors by Sangamo

BioSciences (Pt. Richmond, CA) and have been previously described [321, 334]. The VEGF-A-

activating ZFP (32E-p65) – referred to as AdV-ZFP-VEGF – is a 378 amino acid multi-domain

protein that is composed of three functional regions (Figure 15): (1) the nuclear localization

signal (NLS) of the large T-antigen of SV40, (2) a designed 3-finger zinc-fingered protein (32E)

that binds to a 9 base-pair target DNA sequence (GGGGGTGAC) present in the human VEGF-A

promoter region and (3) the transactivation domain from the p65 subunit of human NFκB, which

is identical to VZ+434, subcloned into pVAX1 (Invitrogen, San Diego, CA) with expression

driven by the human cytomegalovirus (CMV) promoter. Adenoviral (Ad5-32Ep65 or Ad5-

eGFP) vectors, referred to as AdV-ZFP-VEGF and AdV-eGFP, respectively, were packaged by

transfecting T-REx-293 cells (Invitrogen, San Diego, CA). T-REx-293 cells in ten-stack cell

factories were inoculated with Ad vectors at a multiplicity of infection (MOI) of 50 to 100

particles per cell. When adenoviral mediated cytopathy effect (CPE) was observed, cells were

harvested and lysed by three cycles of freezing and thawing. Crude lysates were clarified by

centrifugation, and 293 cells were seeded at 4x107 PFU and grown 3 days prior to transfection.

The calcium phosphate method was used for transfection. Infectious titers of the Ad vectors were

quantified using the Adeno-X Rapid Titer kit (Clontech, Mountain View, CA).

128

SCI and Intraspinal Microinjection

Animals were subject to a compressive spinal cord injury using a modified aneurysm clip, which

has been extensively characterized by the Fehlings’ laboratory and previously described [358].

Briefly, adult female Wistar rats (250-300g; Charles River, Montreal, Canada) were deeply

anesthetized using 4% isoflurane, and were sedated for the remainder of the surgery under 2%

isoflurane. Animals received a two-level laminectomy of mid-thoracic vertebral segments T6-

T7. A modified clip calibrated to a closing force of 35g was applied extradurally to the cord for 1

minute and then removed (Figure 14). The animals were divided into four groups in a

randomized and “blinded” manner, (1) Sham control group (laminectomy only – no SCI), (2)

Non-injected injured control group (laminectomy and SCI – no injection), (3) AdV -ZFP-VEGF

treatment group, and (4) AdV-eGFP control group. Using a stereotaxic frame and glass capillary

needle (tip diameter 60 µm) connected to a Hamilton microsyringe, a total of 5x108 viral plaque

forming units (PFU) were injected into the dorsal spinal cord 24 hours post-SCI. Four 2.5 μl (10

μl total) intraspinal injections were made bilaterally at 2mm rostral and caudal of the injury site

(Figure 14). Injections were 1mm lateral from the midline and 1mm deep into the spinal cord.

The injection rate is 0.60 µl/min and when the injection was completed, the capillary needle was

left in the cord for at least 1 min to allow diffusion of the virus from the injection site and to

prevent back-flow. The incision was closed in layers using standard silk sutures and animals

were given a single dose of buprenorphine (0.05 mg/kg). Animals were allowed to recover in

their cage under a heat-lamp and, subsequently, were housed in a temperature-controlled warm

room (26°C) with free access to food and water. Animals were given buprenorphine (0.05

mg/kg) every 12 hours for 48 hours following surgery, and their bladders were manually voided

three times daily. A subcutaneous injection of 10mg/kg of cyclosporin-A was administered daily

starting 24 hours prior to the SCI until the end of the experiments for immunosuppression. A

129

subcutaneous injection of 10mg/kg of cyclosporin-A was administered daily starting 24 hours

prior to the SCI until the end of the experiments for immunosuppression. The number of animals

used in each experiment is outlined in Table 7. Final animal numbers in each group are slightly

varied due to unexpected mortalities during the experiments.

Western Blotting

Following deep inhalational anesthetic (isoflurane), animals were sacrificed at five or ten days

post-SCI and a 5 mm length of the spinal cord centered at the injury site was extracted. Samples

were mechanically homogenized in 400 μl of homogenization buffer (0.1M Tris, 0.5M EDTA,

0.1% SDS, 1M DTT solution, 100mM PMSF, 1.7mg/ml aprotinin, 1mM pepstatin, 10mM

leupeptin) and centrifuged at 15,000 rpm for 10 minutes at 4 °C. Supernatants were extracted

and used for western blot analysis, where 20 μg of protein was loaded into 7.5% or 12%

polyacrylamide gels (Bio-Rad, Mississauga, Canada). Membranes were probed with either

monoclonal anti-NF200 antibody (1:2000; Sigma, Oakville, Canada), rabbit IgG anti-VEGF-A

antibody (1:100; Santa Cruz Biotechnology, Santa Cruz, CA), or rabbit IgG anti-NFκBp65

(1:1000; Santa Cruz Biotechnology, Santa Cruz, CA). NFκBp65 rabbit polyclonal antibody was

used to recognize the p65 activation domain in the ZFP-VEGF treated animals. Primary

antibodies were labelled with horseradish peroxidase-conjugated secondary antibodies (goat anti-

mouse/rabbit IgG, 1:3000; Jackson Immuno Research Laboratories, West Grove, PA), and bands

were imaged using an enhanced chemiluminescence (ECL) detection system (Perkin Elmer,

Woodbridge, Canada). Mouse monoclonal, beta-actin (Chemicon International, Inc., Temecula,

CA) was immunoblotted as a loading control. Quality One detection software (Bio-Rad

Laboratories, Hercules, CA) was used for integrated optical density (OD) analysis.

130

Histochemistry

Histological Processing. Five or ten days post-SCI, following deep inhalational anesthetic

(isoflurane), animals were transcardially perfused with 4% paraformaldehyde (PFA) in 0.1 M

PBS. Then, the tissues were cryoprotected in 20% sucrose in PBS. A 10 mm length of the spinal

cord centered at the injury site was fixed in tissue-embedding medium. The tissue segment was

snap frozen on dry ice and sectioned on a cryostat at a thickness of 14 μm. Serial spinal cord

sections at 500 μm intervals were stained with myelin-selective pigment luxol fast blue (LFB)

and the cellular stain hematoxylin-eosin (HE) to identify the injury epicenter. Tissue sections

showing the largest cystic cavity and greatest demyelination were taken to represent the injury

epicenter.

Immunohistochemistry. The following primary antibodies were used: mouse anti-NeuN (1:500;

Chemicon International, Inc., Temecula, CA) for neurons, mouse anti-GFAP (1:500; Chemicon

International, Inc., Temecula, CA) for astrocytes, mouse anti-APC (CC1, 1:100; Calbiochem,

San Diego, CA) for oligodendrocytes, and mouse anti-RECA-1 (1:25; Serotec Inc., Raleigh, NC)

for endothelial cells. The sections were rinsed three times in PBS after primary antibody

incubation and incubated with either fluorescent Alexa 568, 647 or 488 goat anti-mouse/rabbit

secondary antibody (1:400; Invitrogen, Burlington, Canada) for 1 hour. The sections were rinsed

three times with PBS and cover slipped with Mowiol mounting medium containing DAPI

(Vector Laboratories, Inc., Burlingame, CA) to counterstain the nuclei. The images were taken

using a Zeiss 510 laser confocal microscope.

Quantification of Blood Vessels. Tissue sections – taken from animals sacrificed 10 days post-

SCI – were used for immunofluorescence studies with a monoclonal antibody specific for

RECA-1 (Rat Endothelial Cell Antibody). As shown in Figure 16, the counting of vessels was

131

performed on 4 selected fields (ventral horn, dorsal horn, left and right lateral columns) in each

section under 25X magnification (0.14 mm2). The number of RECA-1-positive vessels was

calculated at 2 mm and 4 mm, both rostral and caudal from the epicenter, for each animal.

Quantification of Angiogenesis. Tissue sections – taken from animals sacrificed 5 days post-SCI

– were used to quantify angiogenesis following SCI and AdV-ZFP-VEGF administration.

Angiogenesis was calculated as vessels co-labelled with RECA-1 and Ki67 (cellular

proliferation). As shown in Figure 16, the counting of angiogenesis was performed on 4 selected

fields (ventral horn, dorsal horn, left and right lateral columns) in each section under 25X

magnification (0.14 mm2). The number of angiogenic vessels was calculated at 1, 2 and 3 mm,

both rostral and caudal from the epicenter, for each animal. Rostral and caudal values were

pooled for each distance.

Quantification of Apoptosis. An in situ terminal-deoxy-transferase mediated dUTP nick end-

labeling (TUNEL) apoptosis kit (Chemicon International, Inc., Temecula, CA) was used to label

apoptotic cells in tissues extracted from animals 5 days post-SCI. TUNEL staining was

completed as described in the manufacturer’s instructions. The numbers of TUNEL positive

nuclei were counted at the epicenter, as well as at 1, 2 and 3 mm (rostral and caudal) from the

injury epicenter. In each tissue section, the whole section was counted to include all apoptotic

nuclei visible.

Quantification of Neurons. Tissue sections – taken from animals sacrificed 5 days post-SCI –

were used for immunofluorescence studies with a monoclonal antibody specific for NeuN

(Neuronal Nuclei). Neuron quantification was conducted only in the grey matter under 25X

magnification (0.14 mm2), and all cells were counted. The number of NeuN-positive cells was

132

calculated at 1, 2 and 3 mm, both rostral and caudal from the epicenter, as well as at the

epicenter.

Statistical Analysis

Data were analyzed with SigmaPlot software (Systat Software Inc., San Jose, California, USA).

For data that investigated the percentage of cells, the data were subjected to an arcsine

transformation prior to statistical analysis to attain a more normal distribution. For comparison

of groups sampled at various distances from the injury site I used two-way analysis of variance

(ANOVA) with repeated measures, followed by the post-hoc Holm-Sidak test. In all figures, the

mean value ± SEM are used to describe the results. Statistical significance was accepted for p

values of <0.05.

133

Table 7. Animals used in Chapter 4 experiments.

Experiment Group Original Animal

# Final Animal

#

AdV Transduction All groups 5 5 (Figure 23)

NFkBp65 protein AdV-ZFP-VEGF 3 3

(Figure 24) AdV-eGFP 3 3

VEGF mRNA Sham 4 4 (Figure 25A) Injured Control 5 4

AdV-eGFP 5 5

AdV-ZFP-VEGF 5 5

VEGF protein Sham 4 4 (Figure 25B and 25C) Injured Control 5 4

AdV-eGFP 5 4

AdV-ZFP-VEGF 5 5

TUNEL, RECA-1, Ki67/RECA-1 Sham 4 4 (Figure 26, Figure 29, Figure 30) Injured Control 5 4

AdV-eGFP 5 5

AdV-ZFP-VEGF 5 5

NF200 protein Sham 5 5 (Figure 27) Injured Control 5 4

AdV-eGFP 5 4

AdV-ZFP-VEGF 5 4

NeuN Counts Sham 5 5 (Figure 28) Injured Control 5 4

AdV-eGFP 5 5

AdV-ZFP-VEGF 5 5

134

4.7 Results

4.7.1 AdV-ZFP-VEGF Delivery into the Injured Spinal Cord

To evaluate the transduction efficiency of the adenoviral constructs in vivo, AdV-eGFP was

injected into animals at 24 hours post-SCI. The AdV-eGFP fluorescent signal was detected in

both the white and grey matter of the injured spinal cord five days after SCI (Figure 23). Figures

24B and 24C demonstrate eGFP expression in neurons, astrocytes, endothelial cells and

oligodendrocytes, indicating successful adenoviral transduction into each cell type. Further

quantification of co-labelled cells showed that AdV vector non-preferentially transduces all cell

types (Neurons – 30.0% ± 3.6%, Oligodendrocytes – 26.9% ± 4.2%, Astrocytes – 21.4% ± 2.9%,

Endothelial cells – 17.2% ± 3.3%). Since the AdV-ZFP-VEGF construct contains the p65

subunit of the human NFκB transcription factor as the activation domain [321], I was able to

confirm delivery of AdV-ZFP-VEGF by immunoblotting using an NFκB p65 antibody to detect

the presence of the transcription factor (Figure 24). As a positive control, HEK293 cells were

transduced with ZFP-VEGF and cell lysates were processed for immunoblotting using the same

NFκB p65 antibody (data not shown). These results demonstrate the successful delivery of a

localized gene therapy to the injured spinal cord.

135

136

Figure 23. Transduction of AdV-eGFP into the spinal cord. A) Photomicrographs showing a

transverse section of rat spinal cord obtained adjacent to the injury site 10 days after spinal cord

injury and AdV-eGFP injection. eGFP signal was detected in both the gray matter and white

matter. B) High-power (63X) confocal images show that the AdV-eGFP vector transfected

neurons (NeuN), astrocytes (GFAP), oligodendrocytes (CC1) and endothelial cells (RECA-1).

Neurons, astrocytes, and endothelial cells are taken from the grey matter, and oligodendrocytes

are taken from the white matter. C) Bar graph displays quantification of transduced cell types ±

SEM, as identified by the cell-specific markers NeuN, GFAP, RECA-1 and CC1. Scale bar:

1000 μm for A; 100 μm for B. n=5 animals/antibody (see Table 7).

Figure 24. Evaluation of AdV-ZFP-VEGF gene transfer. Western blot showed that the

NFκB p65 rabbit polyclonal antibody recognizes the p65 activation domain in the AdV-ZFP-

VEGF treated animals. The higher molecular weight bands are endogenous NFκBp65 fragments,

which are also recognized by the antibody; however, these bands are present in both the control

and treatment groups. The lower band (arrow) corresponds to the AdV-ZFP-VEGF and was only

present in the treated animals. Lower panel shows actin expression as a protein control. n=3

animals/group (see Table 7).

137

4.7.2 VEGF mRNA and protein expression is increased following 24 hour delayed AdV-ZFP-VEGF administration

Animals were sacrificed 5 days post-SCI and mRNA expression levels of three predominant

VEGF isoforms found in the CNS – VEGF120, VEGF164 and VEGF188 – were measured by

quantitative real-time PCR (qRT-PCR). Figure 25A shows that 24 hour delayed administration

of AdV-ZFP-VEGF resulted in significant increases in VEGF mRNA of isoforms 120 (p <

0.001), 164 (p < 0.001), but not isoform 188, when compared with AdV-eGFP control animals

and injured control animals. VEGF-A protein expression was assessed at 10 days following SCI

by Western blot using anti-VEGF antibodies, which detect the 42kDa and 21kDa bands and are

recommended for the detection of the 189, 165 and 121 amino acid splice variants of VEGF. In

Figures 25B and 25C, I show that the 42 kDa VEGF-dimer protein was significantly increased

by approximately 2.5-fold in AdV-ZFP-VEGF treated animals versus AdV-eGFP and injured

control groups (p < 0.02). VEGF protein was increased by approximately 1.8-fold in AdV-ZFP-

VEGF treated animals compared to sham animals (p < 0.05). Previous studies using AdV-ZFP-

VEGF have shown increases in VEGF mRNA and protein levels [212, 316, 318, 319].

Consistent with these studies, my results confirm that AdV-ZFP-VEGF increases both mRNA

and protein levels of VEGF in the spinal cord following 24 hour delayed administration.

138

139

Figure 25. AdV-ZFP-VEGF increases VEGF mRNA and protein. A) VEGF mRNA levels

encoding for VEGF120, VEGF164 and VEGF188 isoforms were measured by quantitative real-time

PCR at 5 days post-SCI. The bar graph illustrates that administration of ZFP-VEGF resulted in

an increase of VEGF mRNA compared with AdV-eGFP and SCI injured control groups.

Relative mRNA levels are expressed as the mean ± SEM, n = 4/ sham and injured control

groups, n = 5/ AdV-eGFP and AdV-ZFP-VEGF groups (see Table 7). One-way ANOVA

(Holm-Sidak) was completed individually for each isoform **p < 0.001, *p < 0.01. B) Western

blot showing administration of AdV-ZFP-VEGF resulted in increased VEGF-A protein levels at

10 days post-SCI, and C) Quantification shows a significant increase in VEGF-A 42 kD protein

in AdV-ZFP-VEGF treated animals compared with control groups. Optical density (OD) of

VEGF-A was normalized to actin. Data are presented as mean ± SEM, n = 4/sham, injured

control and AdV-eGFP treated groups and n = 5/AdV-ZFP-VEGF treated group (see Table 7).

One-way ANOVA (Holm-Sidak) **p < 0.02, *p < 0.05.

4.7.3 Apoptosis is reduced in animals treated with AdV-ZFP-VEGF 24 hours post-SCI

The Fehlings’ laboratory has previously shown that apoptotic cell death occurs as early as 6

hours following SCI and persists until 14 days post injury [73]. To assess the effects of AdV-

ZFP-VEGF treatment on apoptotic cell death, in situ terminal-deoxy-transferase mediated dUTP

nick end-labeling (TUNEL) staining was performed 5 days after injury (Figure 26). TUNEL-

positive cells were found evenly distributed through the gray and white matter in the injured

spinal cord, with the greatest apoptosis observed near the injury epicenter. TUNEL-stained

nuclei were counted at the injury epicenter, and at 1, 2, and 3 mm from the injury epicenter both

140

rostral and caudal to the lesion site, but rostral and caudal values were pooled. Figure 26B

shows that AdV-ZFP-VEGF treatment was associated with a significant reduction in the number

of TUNEL-positive cells rostral and caudal from the injury epicenter (p < 0.01).

141

142

Figure 26. AdV-ZFP-VEGF administration reduces apoptosis after SCI. A) Representative

sections taken 2 mm rostral to the epicenter from animals sacrificed at 5 days post-SCI and tissue

processed with TUNEL staining; scale 200 μm. An overall reduction of TUNEL-positive cells

was observed in the AdV-ZFP-VEGF treated group. B) Bar graph shows quantification of the

TUNEL-positive cell counts at 5 days after SCI (pooled values from rostral and caudal counts).

There was a significant decrease in TUNEL-positive cells in the AdV-ZFP-VEGF treatment

group versus control injured groups. Values are mean ± SEM, n = 4/ sham and injured control

groups, n = 5/ AdV-eGFP and AdV-ZFP-VEGF groups (see Table 7). Two-way ANOVA

(Holm-Sidak), * p < 0.01.

4.7.4 24 hour delayed AdV-ZFP-VEGF administration provides neuroprotection

Neurofilament protein (NF200), a hallmark protein lost following neurodegeneration, was

quantified in the injured region of the cord to assess the neuroprotective effects of AdV-ZFP-

VEGF after SCI. Previous research from Dr. Fehlings’ group demonstrated a significant loss of

NF200 after SCI [75, 108]. As shown in Figure 27, the amount of NF200 protein was

significantly increased by approximately 2-fold at 10 days following SCI in animals treated with

AdV-ZFP-VEGF versus control animals (p < 0.05).

143

Figure 27. AdV-ZFP-VEGF administration attenuated axonal degradation. A) Western

blot indicates that administration of AdV-ZFP-VEGF resulted in a significant attenuation of

NF200 degradation 10 days after injury. Lower panel shows actin protein control. B) Relative

OD value of controls versus AdV-ZFP-VEGF treated animals. Significant NF200 sparing was

observed in AdV-ZFP-VEGF-treated animals compared to control groups at 10 days after injury,

although all injured groups showed significant NF200 loss following SCI. Optical density of

NF200 was normalized to actin. Bar graph shows mean OD values ± SEM; n = 5/sham, n =

144

4/Injured Control, AdV-eGFP and AdV-ZFP-VEGF groups (see Table 7). One-way ANOVA

(Holm-Sidak), * p < 0.05.

To further assess the neuroprotective effects of AdV-ZFP-VEGF following SCI, I quantified

spared neurons 5 days after injury. NeuN, which recognizes neuronal cell bodies, was used to

identify neurons in cross-sections of spinal cord tissue. Figure 28B demonstrates that AdV-ZFP-

VEGF treatment results in a significant sparing of neurons both rostral and caudal to the injury

epicenter, when compared to injured control (p < 0.05) and AdV-eGFP (p < 0.001) animals.

145

146

Figure 28. AdV-ZFP-VEGF administration results in increased neuron sparing post-SCI.

A) Representative sections taken 2 mm rostral to the epicenter from AdV-ZFP-VEGF treated and

AdV-eGFP treated animals immunostained with NeuN at 5 days after SCI; scale 200 μm. A

greater number of NeuN-positive cells was observed in animals treated with AdV-ZFP-VEGF.

B) Bar graph shows quantification of the NeuN-positive cell counts at 5 days after SCI. There

was an overall significant preservation of neurons in AdV-ZFP-VEGF treated animals, when

quantified by a Two-way ANOVA (treatment and distance from injury). Values are mean ±

SEM, n = 4/ injured control group, n = 5/ sham, AdV-eGFP and AdV-ZFP-VEGF groups (see

Table 7). Two-way ANOVA (Holm-Sidak), * p < 0.02.

4.7.5 24 hour delayed AdV-ZFP-VEGF administration results in an increased number of vessels

In order to quantify the vascular response to ZFP-VEGF, I conducted immunostaining with

RECA-1, a monoclonal antibody specific for endothelial cells, at 10 days following SCI. The

severity of the compression injury resulted in considerable disruption to the spinal cord

vasculature at the injury epicentre, thus I was unable to quantify the epicenter accurately.

Therefore, I assessed spinal cord tissue sections at 2 mm and 4 mm – both caudal and rostral –

from the lesion epicenter (Figure 30A). Figures 29B and 29C show that AdV-ZFP-VEGF

administration markedly increases the number of RECA-1-positive vessels both rostral and

caudal, when compared to control animals (p < 0.01). These results are consistent with previous

findings from the Fehlings’ laboratory in studies administering AdV-ZFP-VEGF immediately

following injury [212].

147

148

Figure 29. AdV-ZFP-VEGF results in increased vessel counts. A) Illustration of the area of

spinal cord areas used for RECA-1 counting (2 grey matter areas, 2 white matter areas). B)

Representative sections taken 2 mm rostral to the epicenter from a AdV-ZFP-VEGF treated and

AdV-eGFP control animal respectively immunostained with RECA-1 at 10 days after SCI; scale

100 μm. An increased number of vessels were observed in the AdV-ZFP-VEGF treated group.

C) Bar graph illustrating the RECA-1 positive cell counts 10 days after SCI. AdV-ZFP-VEGF

administration resulted in a significant increase in vascular counts (2 mm and 4 mm away from

the epicenter) as compared with the control group. Data are presented as mean ± SEM, n = 4/

sham and injured control groups, n = 5/ AdV-eGFP and AdV-ZFP-VEGF groups (see Table 7).

Two-way ANOVA (Holm-Sidak) * p < 0.01.

4.7.6 AdV-ZFP-VEGF promotes angiogenesis

To investigate some of the potential mechanisms of AdV-ZFP-VEGF action, I examined the

effects of 24 hour delayed AdV-ZFP-VEGF administration on endothelial cell proliferation. One

of the most characterized roles of VEGF is promoting angiogenesis in both embryonic

development and wound healing [235], therefore I aimed to study if AdV-ZFP-VEGF

administration would further promote angiogenesis. Tissues co-labeled with RECA-1 and Ki67

at 5 days following SCI indicated that AdV-ZFP-VEGF administration increased angiogenesis

by approximately 10% (p <0.001) (Figure 30). These results indicate that AdV-ZFP-VEGF

administration, which results in an increase in VEGF expression, ultimately promotes angiogenic

pathways following SCI. Other research has suggested that VEGF administration results in

angiogenesis; however, these studies simply show an increase in the number of vessels present.

Here, I demonstrate that VEGF increases endothelial cell proliferation in vivo following SCI, and

149

to my knowledge, this is the first study to show that an increase in vessels (as previously shown

in Figure 29) may be attributable to angiogenesis.

150

Figure 30. AdV-ZFP-VEGF promotes angiogenesis at 5 days post-SCI. A) Representative

image from an ADV-ZFP-VEGF treated animal at 5 days post-injury. Image was taken at 25X

at 2 mm from the epicenter. Scale bar = 100 µm. B) Angiogenesis was assessed by quantifying

Ki67/RECA-1 co-labeled vessels. Data are presented at the percentage of RECA-1+ vessels that

were also Ki67+, with an overall average increase of 10% vascular proliferation observed in the

animals receiving AdV-ZFP-VEGF administration. n = 4/sham and injured control groups, n =

5/AdV-eGFP and AdV-ZFP-VEGF groups (see Table 7). Data were analyzed by performing an

arcsine transformation of the values, and conducting a two-way ANOVA with Holm-Sidak post-

hoc testing. ** p < 0.001.

4.8 Discussion

In the current research, I chose to investigate the efficacy of 24 hour delayed AdV-ZFP-VEGF

administration, which presents a clinically relevant therapeutic window. This form of gene

therapy mimics physiological VEGF production, which should result in the production of all

VEGF isoforms in the injured spinal cord, a necessary component for proper and functional

angiogenesis. Here I observe that 24 hour delayed administration of AdV-ZFP-VEGF results

increased VEGF mRNA and protein levels, an increased number of vessels, enhanced vascular

proliferation (as observed by RECA-1/Ki67 staining) and improved neuroprotection – as

observed by increased NeuN counts and spared NF200.

In previous research that has used VEGF following SCI, authors have observed varying results.

Choi et al. used a hypoxia-inducible VEGF-A expression system to treat rats with SCI and

observed neuroprotective effects and enhanced VEGF-A expression [359]. Another group used

151

an adenovirus coding for VEGF165, delivered via matrigel, in a partial spinal cord transection

model. They observed a significant increase in vessel volume and a reduction in the retrograde

degeneration of corticospinal tract axons [360]. However, Benton et al. [354] reported an

exacerbation of lesion size and increased inflammation after the delivery of 2 μg of recombinant

VEGF165 directly into the contused spinal cord 3 days post SCI. This study highlights several

factors which are likely to be critical in the successful application of VEGF-A as a therapy for

SCI. The method of VEGF delivery is likely a critical factor. In this study, I injected the ZFP-

VEGF adjacent to the injury epicenter as the peri-lesional ischemic penumbra is likely the zone,

which would benefit the most from approaches to enhance angiogenesis. Moreover, the delivery

technique, using a ZFP-VEGF gene therapy, has the ability to upregulate several isoforms of

VEGF-A, mimicking endogenous expression. In contrast, most other research has focused on

the delivery of a single VEGF isoforms.

Previously, our lab has shown that immediate administration of AdV-ZFP-VEGF following SCI

resulted in neuroprotection, increased vascular counts and improved functional recovery [212].

These promising results encouraged us to investigate a more feasible time-window for clinical

intervention: 24 hours post-injury administration of AdV-ZFP-VEGF. Moreover, administration

24 hours following injury aims to target a few important pathophysiological events post-SCI,

particularly vascular damage and apoptosis. In a model of spinal cord contusion, Ling and Liu

show that TUNEL-positive cells are maximally seen at 48 hours following injury in both the

grey and white matter [129]. Similarly, Crowe et al. demonstrated that maximal apoptosis was

observed at 48 hours following contusion injury, with apoptosis identified between 6 hours and 3

weeks [351]. Liu et al. showed that following contusion injury TUNEL-positive neurons were

observed between 4-24 hours, whereas TUNEL-positive glia were seen between 4 hours and 14

152

days with maximal numbers observed at 24 hours, although another peak of TUNEL-positive

glia were observed at 7 days post-injury [130]. With respect to vascular targets, research

suggests that angiogenic therapies should be administered to target endogenous vascular repair,

which occurs between 3 and 7 days following injury [58, 151]. Therefore, by administrating

AdV-ZFP-VEGF 24 hours following injury, I aim to target and reduce apoptosis as well as

enhance vascular regeneration. My data, showing reduced TUNEL and increased endothelial

cell proliferation, suggest that AdV-ZFP-VEGF is in fact capable of both neuroprotection and

angiogenesis. Although I have not investigated the detailed mechanisms or signaling pathways

of AdV-ZFP-VEGF in vivo, collectively data provide strong evidence that increasing VEGF

following injury may be beneficial and may stimulate cell survival and angiogenic pathways.

4.9 Conclusions

The present data demonstrate that, similar to immediate administration of AdV-ZFP-VEGF,

treated animals show increased VEGF mRNA and protein levels, increased vascular counts,

increased neuroprotection and reduced apoptosis. Overall, the administration of AdV-ZFP-

VEGF shows promise as a therapeutic treatment for SCI, and these findings suggest that AdV-

ZFP-VEGF treatment can be delayed up to 24 hours following injury, which presents a feasible

time-window for clinical intervention. To the best of my knowledge, this research is the first to

investigate the delayed administration of AdV-ZFP-VEGF in a model of SCI. Based on the

beneficial effects observed in a variety of cell populations, I believe AdV-ZFP-VEGF

administration further supports the use of VEGF as a potential candidate for neurotrauma

treatments.

153

Chapter 5

5 AdV-ZFP-VEGF Results in Functional Improvements and Reduced Allodynia Following SCI

5.1 Abstract

For decades, spinal cord injury research has looked for a cure; however, there are still limited

therapeutic options. Spinal cord injury often results in permanent functional deficits, which

drastically alter the quality of life of an individual. To this end, the scientific community

continues to explore novel therapeutics for SCI treatments. Previously, I have investigated

vascular endothelial growth factor-A (VEGF-A) as a potential therapy for SCI, and have shown

positive outcomes at the cellular and molecular level; however, the current research aimed to

elucidate if the effects of AdV-ZFP-VEGF observed in previous studies can be translated into

functional and behavioural improvements. Briefly, female Wistar rats – under cyclosporin-A

immunosuppression – received a 35g clip-compression injury and were administered AdV-ZFP-

VEGF or AdV-eGFP at 24 hours post-SCI. Animals were subject to weekly behavioural testing

(using BBB open-field scoring) for 8 consecutive weeks by two-blinded observers. On weeks 4,

6, and 8 post-injury, animals were tested for both mechanical and thermal allodynia, and

performed Catwalk™ testing. Animals were sacrificed at 8 weeks post-SCI, and tissue was

collected for histological assessment. Catwalk™ analysis showed AdV-ZFP-VEGF treatment

improves hindlimb weight support (p<0.05) and increases hindlimb swing speed (p<0.02) when

compared to control animals. Interestingly, animals treated with AdV-ZFP-VEGF also showed

improvements in forelimb function, specifically forelimb stride length (p<0.007). Finally, AdV-

ZFP-VEGF administration provided a significant reduction in allodynia, both at and below the

154

level of injury (p<0.01). The results of this study suggest that 24 hour delayed AdV-ZFP-VEGF

administration is may be an effective therapy following SCI, as it results in improved hindlimb

function and decreased allodynia.

5.2 Introduction

Perhaps the most devastating outcomes of spinal cord injury are paralysis and neuropathic pain.

Paralysis is caused by damaged axons and neurons in motor pathways at or above the level of

injury. Many models of SCI have been used to model the physical deficits post-injury, and

thoracic injuries are among the best-characterized for the targets loss of hindlimb function.

Motor impairment following SCI results from damage to and/or loss of both upper and lower

motor neurons. Injury to first and second order spinothalamic neurons, or first order neurons

from the medial lemniscus pathway, interrupts sensory information processing at and below the

level of injury and prevents normal signal transmission to the brain. Miscommunication in

sensory pathways can result in severe complications for patients suffering from SCI.

Development of neuropathic pain occurs in many patients, and although the exact mechanism is

unknown, it is hypothesized that it caused by misguided axonal sprouting or abnormal sodium

channel excitability in sensory neurons [79]. Neuropathic pain will be discussed in more detail

in subsequent paragraphs.

Development of neuropathic pain is dependent on location of the injury site and the surrounding

neural pathways. Clinically, neuropathic pain is divided into three areas which help to describe

the location of the pain: “above-level”, “at-level” and “below-level”. Chronic astrocyte and

microglial activation produce factors that result in hyperexcitability of neurons in distal regions

155

of the dorsal/ventral horns, with respect to the epicenter [171, 172]. Patients may also develop

mechanical and/or thermal allodynia, which causes previously innocuous stimuli to feel noxious.

In rats, neuropathic pain develops approximately 4 weeks post-injury and depends on injury

intensity [173]. In humans, it is estimated that the number of patients exhibiting neuropathic

pain is as high as 58% in some patient populations of SCI [174]. Consistent with the knowledge

that astrocytes and microglia are highly active in neuropathic pain, therapies that inhibit or

modulate astrocyte, microglial/macrophage activation have shown a reduced incidence of

neuropathic pain in animal models of SCI [175, 176].

In the current study, I aimed to elucidate the effects of 24 hour delayed AdV-ZFP-VEGF

delivery on neurobehavioural and functional outcomes. Investigating the functional effects of

AdV-ZFP-VEGF seemed worthwhile, considering our previous data (Chapter 4, Appendix 1),

and consistent with previous studies I do observed significant improvements in hindlimb weight

support (p<0.05) and marked reductions in mechanical and thermal allodynia (p<0.01). These

results provide important and meaningful pre-clinical outcomes, as decreased weight support and

the development of neuropathic pain are major issues in patients following traumatic SCI. From

the aspect of recovery and treatments for SCI, any therapy that improved hindlimb function

and/or reduced pain has the potential to significantly alter the quality of life in SCI patients. To

my knowledge, this is the first report that uses AdV-ZFP-VEGF in a delayed fashion which has

shown significant neurobehavioural improvements.

156

5.3 Objective

Based on the previous data suggesting beneficial outcomes (at cellular/molecular levels)

following AdV-ZFP-VEGF administration, I aim to investigate the effects of ZFP-VEGF gene

therapy at chronic time-points following SCI; specifically focusing on neurobehavioural and

neuroanatomical outcomes.

5.4 Hypothesis

It is hypothesized that the acute effects of ZFP-VEGF gene therapy observed in previous

research will translate into functional and behavioural improvements following spinal cord

injury.

5.5 Specific Aims

1. Assess the neurobehavioural effects of AdV-ZFP-VEGF post-SCI on hindlimb

locomotion and weight support.

2. Determine if AdV-ZFP-VEGF administration alters the development of neuropathic pain

following SCI.

3. Assess the neuroanatomical (tissue sparing) effects of AdV-ZFP-VEGF following SCI.

157

5.6 Methods

All animal experiments were conducted with approval from the Animal Care Committee,

University Health Network (Toronto, Canada).

Viral Vector Constructs

The VEGF-A-activating ZFP and controls were provided in viral vectors by Sangamo

BioSciences (Pt. Richmond, CA) and have been previously described [321, 334]. The VEGF-A-

activating ZFP (32E-p65) – referred to as AdV-ZFP-VEGF – is a 378 amino acid multi-domain

protein that is composed of three functional regions (Figure 15): (1) the nuclear localization

signal (NLS) of the large T-antigen of SV40, (2) a designed 3-finger zinc-fingered protein (32E)

that binds to a 9 base-pair target DNA sequence (GGGGGTGAC) present in the human VEGF-A

promoter region and (3) the transactivation domain from the p65 subunit of human NFκB, which

is identical to VZ+434, subcloned into pVAX1 (Invitrogen, San Diego, CA) with expression

driven by the human cytomegalovirus (CMV) promoter. Adenoviral (Ad5-32Ep65 or Ad5-

eGFP) vectors, referred to as AdV-ZFP-VEGF and AdV-eGFP, respectively, were packaged by

transfecting T-REx-293 cells (Invitrogen, San Diego, CA). T-REx-293 cells in ten-stack cell

factories were inoculated with Ad vectors at a multiplicity of infection (MOI) of 50 to 100

particles per cell. When adenoviral mediated cytopathy effect (CPE) was observed, cells were

harvested and lysed by three cycles of freezing and thawing. Crude lysates were clarified by

centrifugation, and 293 cells were seeded at 4x107 PFU and grown 3 days prior to transfection.

The calcium phosphate method was used for transfection. Infectious titers of the Ad vectors were

quantified using the Adeno-X Rapid Titer kit (Clontech, Mountain View, CA).

158

SCI and Intraspinal Microinjection

Animals were subject to a compressive spinal cord injury using a modified aneurysm clip, which

has been extensively characterized by the Fehlings’ laboratory and previously described [358].

Briefly, adult female Wistar rats (250-300g; Charles River, Montreal, Canada) were deeply

anesthetized using 4% isoflurane, and were sedated for the remainder of the surgery under 2%

isoflurane. Animals received a two-level laminectomy of mid-thoracic vertebral segments T6-

T7. A modified clip calibrated to a closing force of 35g was applied extradurally to the cord for 1

minute and then removed (Figure 14). The animals were divided into four groups in a

randomized and “blinded” manner, (1) Sham control group (laminectomy only – no SCI), (2)

Non-injected injured control group (laminectomy and SCI – no injection), (3) AdV -ZFP-VEGF

treatment group, and (4) AdV-eGFP control group. Using a stereotaxic frame and glass capillary

needle (tip diameter 60 µm) connected to a Hamilton microsyringe, a total of 5x108 viral plaque

forming units (PFU) were injected into the dorsal spinal cord 24 hours post-SCI. Four 2.5 μl (10

μl total) intraspinal injections were made bilaterally at 2mm rostral and caudal of the injury site

(Figure 14). Injections were 1mm lateral from the midline and 1mm deep into the spinal cord.

The injection rate is 0.60 µl/min and when the injection was completed, the capillary needle was

left in the cord for at least 1 min to allow diffusion of the virus from the injection site and to

prevent back-flow. The incision was closed in layers using standard silk sutures and animals

were given a single dose of buprenorphine (0.05 mg/kg). Animals were allowed to recover in

their cage under a heat-lamp and, subsequently, were housed in a temperature-controlled warm

room (26°C) with free access to food and water. Animals were given buprenorphine (0.05

mg/kg) every 12 hours for 48 hours following surgery, and their bladders were manually voided

three times daily. A subcutaneous injection of 10mg/kg of cyclosporin-A was administered daily

starting 24 hours prior to the SCI until the end of the experiments for immunosuppression. The

159

number of animals used in each experiment is outlined in Table 8. Final animal numbers in each

group are slightly varied due to unexpected mortalities during the experiments.

Histochemistry

Histological Processing. 8 weeks post-SCI, following deep inhalational anesthetic (isoflurane),

animals were transcardially perfused with 4% paraformaldehyde (PFA) in 0.1 M PBS. Then, the

tissues were cryoprotected in 20% sucrose in PBS. A 10 mm length of the spinal cord centered

at the injury site was fixed in tissue-embedding medium. The tissue segment was snap frozen on

dry ice and sectioned on a cryostat at a thickness of 14 μm. Serial spinal cord sections at 500 μm

intervals were stained with myelin-selective pigment luxol fast blue (LFB) and the cellular stain

hematoxylin-eosin (HE) to identify the injury epicenter. Tissue sections showing the largest

cystic cavity and greatest demyelination were taken to represent the injury epicenter.

Assessment of Tissue Sparing and Cavity Formation. Tissue sparing and cavity formation was

analyzed 8 weeks after SCI, at the center of the lesion, 2 mm above and 2 mm below the

epicenter. Sections were stained with LFB-HE. The measurements were carried out on coded

slides using StereoInvestigator software (MBF Bioscience, Williston, VT). Cross-sectional

residual tissue and cavity areas were normalized with respect to total cross-sectional area and the

areas were calculated every 500 µm within the rostrocaudal boundaries of the injury site.

Behavioural Testing

Open-field Locomotor Scoring. Locomotor recovery of the animals was assessed by two

independent observers using the 21 point Basso, Beattie, and Bresnahan (BBB) open field

locomotor score [337] from 1 to 8 weeks after SCI. The BBB scale was used to assess hindlimb

160

locomotor recovery including joint movements, stepping ability, coordination, and trunk

stability. Testing was done every week on a blinded basis and the duration of each session was 4

min per rat. Scores were averaged across both the right and left hindlimbs to arrive at a final

motor recovery score for each week of testing.

Automated Gait Analysis (Catwalk™). Gait analysis was performed using the Catwalk™ system

(Noldus Information Technology, Wageningen, Netherlands) as described [338, 339]. In short,

the system consists of a horizontal glass plate and video capturing equipment placed underneath

and connected to a PC. In our work, for correct analysis of the gait adaptations to the chronic

compression, after standardization of the crossing speed, the following criteria concerning

walkway crossing were used: (1) the rat needed to cross the walkway, without any interruption

(2) a minimum of three correct crossings per animal were required. Files were collected and

analyzed using the Catwalk™ program, version 7.1. Individual digital prints were manually

labeled by one observer blinded to groups. With the Catwalk™, a vast variety of static and

dynamic gait parameters can be measured during spontaneous locomotion. In the present study, I

examined the following parameters, most of which have been studied in human CSM gait

analysis:

• forelimb stride length (expressed in mm): distance between two consequtive forelimb

paw placements

• hindlimb print area (expressed in mm2)

• hindlimb print width (expressed in mm)

• hindlimb print length (expressed in mm)

• hindlimb swing speed (expressed in pixels/sec): is the speed of the paw during the swing

phase (the duration of no paw contact with the glass plate during a step cycle).

161

Before surgery, animals were acclimated and trained to the walking apparatus following the

method describing by Gensel et al. [340].

Neuropathic Pain. At-level mechanical allodynia was determined at 4 weeks and 8 weeks post-

SCI using 2 g and 4 g von Frey monofilaments as previously described [173]. Animals were

acclimatized for 30 minutes in an isolated room for 30 minutes prior to pain testing. The von

Frey monofilament was applied to the dorsal skin surrounding the incision/injury site 10 times

and animals’ behavioural response to each was recorded. An adverse response to the application

of the monofilament (determined in advance of experiments) included vocalization, licking,

biting and immediate movement to the other side of the cage. The proportion of rats to exhibit

allodynia in each group is reported, and an increased number of responses was associated with

the development of at-level mechanical allodynia. Below-level mechanical allodynia was

determined by quantifying the pain threshold of the hindpaws. Animals were placed in stance on

a raised grid, allowing von Frey filaments to be applied to the plantar surface of the hindpaw.

Increasing monofilaments were used (2, 4, 8, 10, 16, 21, and 26 g) until the animal displayed an

adverse response (as described above). The weight of the von Frey filament that elicited the

response was recorded as the pain threshold value, with lower threshold values indicating

increased sensitivity to mechanical stimuli (and perhaps the development of mechanical

allodynia). Finally, below-level thermal allodynia was assessed using the tail flick method. A

50°C thermal stimulus was applied to the distal portion of the animals’ tail by a Tail Flick

Analgesia Meter (IITC Inc. Life Science, Woodland Hills, California, USA), and the time for the

animal to remove its tail from the stimulus was recorded. The latency time is graphed for each

treatment group, and decreased latency times were associated with the development of thermal

allodynia.

162

Electrophysiology

Motor Evoked Potentials. Motor evoked potential recordings (MEPs): In addition to

the behavioural assessements, MEPs were recorded in vivo to assess the physiological integrity

of spinal cord. This approach has been extensively used in our laboratory in rodent models of

SCI. In vivo recordings of motor evoked potentials were recorded from the each of the treatment

and control groups at 8 weeks post-injury. For MEPs, rats were under light isoflourane

anaesthesia (<1%), and recordings were obtained from hindlimb biceps femoris muscle.

Stainless steel subdermal needle electrodes were inserted into the muscle. Recordings were

acquired using Keypoint Portable (Dantec Biomed, Denmark). A reference electrode was placed

under the skin between the recording and stimulating electrodes. Stimulation was applied to the

midline of the cervical spinal cord using a silver ball electrode (0.13 Hz; 0.1 ms; 2 mA; 200

sweeps). The interlaminar ligaments were removed and a small amount of bone was removed

from the vertebra (not a full laminectomy, just enough to create a space for the electrode to reach

the cervical cord). The amplitude was determined by the difference between the positive peak

and negative peak. Latency was calculated as the time from the start of the stimulus artifact to

the first prominent peak. For individual rats, the average of peak amplitude and latency

was averaged from 200 sweeps and analyses was undertaken by ANOVA.

H-Reflex. The Hoffmann reflex is one of the most studied reflexes in humans and is the electrical

analogue of the monosynaptic stretch reflex. The H-reflex is evoked is evoked by low-intensity

electrical stimulation of the afferent nerve, rather than a mechanical stretch of the

muscle spindle, that results in monosynaptic excitation of alpha-motorneurons. H-reflex can be

used as a tool (in combination with other outcome measures) to examine spasticity and short-

and long-term plasticity. Recording electrodes were placed two centimeters apart in the mid-calf

163

region and the posterior tibial nerve was stimulated in the popliteal fossa using a 0.1 ms

duration square wave pulse at a frequency of 1 Hz. The rats were tested for maximal plantar H-

reflex / maximal plantar M-response (H /M) ratios to determine the excitability of the reflex.

The recordings were filtered between 10-10000 Hz.

Statistical Analysis

Data were analyzed with SigmaPlot software (Systat Software Inc., San Jose, California, USA).

For comparison of groups sampled at various distances from the injury site (TUNEL, RECA-1,

NeuN), I used two-way analysis of variance (ANOVA) with repeated measures, followed by the

post-hoc Holm-Sidak test. For comparisons of multiple groups at a single time point (Western

blotting, BBB, Catwalk™, Electrophysiology), I performed a One-way ANOVA, followed by

the post-hoc Holm-Sidak test. In all figures, the mean value ± SEM are used to describe the

results. Statistical significance was accepted for p values of <0.05.

164

Table 8. Animals used in Chapter 5 experiments.

Experiment Group Original Animal # Final Animal #

CatWalk™ Sham 5 5 (Figure 31, Figure 32) Injured Control 12 5

AdV-eGFP 12 5

AdV-ZFP-VEGF 12 5

BBB Scoring, Tissue Sparing Sham 5 5 Neuropathic Pain Injured Control 12 10

(Figure 33, Figure 35, Figure 36) AdV-eGFP 12 10

AdV-ZFP-VEGF 12 8

Electrophysiology Sham 5 5 (Figure 34) Injured Control 12 6

AdV-eGFP 12 6

AdV-ZFP-VEGF 12 6

165

5.7 Results

5.7.1 AdV-ZFP-VEGF results in functional improvement

At the cellular/molecular level, I have previously observed that AdV-ZFP-VEGF results in

beneficial effects. However, in order to assess the viability of any therapy, these effects must be

translated into functional gains. In this study, I assessed hindlimb function using open-field BBB

scoring and Catwalk™, between 1-8 weeks following clip-compression SCI. Analysis of

Catwalk™ data showed that animals treated with AdV-ZFP-VEGF had significantly improved

hindlimb weight support (p<0.05) (Figure 31), hindlimb swing speed (p<0.02) (Figure 32), and

forelimb stride length (p<0.02) compared to all other injured control groups. Although AdV-

ZFP-VEGF animals still perform significantly worse than uninjured controls, the functional

recovery observed by Catwalk™ are promising. Enhancements in hindlimb weight support and

overall gait (hindlimb swing speed, and forelimb stride length) are important changes that may

improve the quality of life of individuals suffering with SCI.

166

167

Figure 31. AdV-ZFP-VEGF improves hindlimb weight support. Catwalk™ gait analysis

was used to assess hindlimb weight support. Animals were assessed every week between 4-8

weeks, and each animal performed a standardized Catwalk™ run. A blinded observer analyzed

the data. A) Paw area, B) Paw width, and C) Paw length. Data presented are the mean ± SEM, n

= 5/group (the best five animals were selected and quantified from each experimental group), at

8 weeks following SCI. One-way ANOVA (Holm-Sidak). * p<0.05, ** p<0.005.

168

169

Figure 32. Forelimb and Hindlimb locomotion is improved by AdV-ZFP-VEGF

administration. A) Catwalk™ gait analysis was used to assess hindlimb swing speed. Animals

were assessed every week between 4-8 weeks, and each animal performed a standardized

Catwalk™ run. A blinded observer analyzed the data. Data presented are the mean ± SEM, n =

5/group, at 8 weeks following SCI. One-way ANOVA (Holm-Sidak). * p<0.02. B) Catwalk™

gait analysis was used to assess forelimb stride length. Animals were assessed every week

between 4-8 weeks, and each animal performed a standardized Catwalk™ run. A blinded

observer analyzed the data. Data presented are the mean ± SEM, n = 5/group, at 8 weeks

following SCI. One-way ANOVA (Holm-Sidak). * p<0.02.

5.7.2 AdV-ZFP-VEGF does not result in improved BBB scores

Interestingly, AdV-ZFP-VEGF treated animals did not show improved BBB scores, compared in

injured control animals, although they did perform better than AdV-eGFP injected animals

(p<0.01) (Figure 33). Despite my observations of significant recovery shown by Catwalk™, the

BBB scores between injured control and AdV-ZFP-VEGF animals are virtually identical at 8

weeks post-injury (BBB scores ≈ 9). In the discussion I will provide a more detailed explanation

that may validate these findings; however, the discrepancy between BBB and Catwalk™ data

may be due to the more qualitative nature of the BBB, as opposed to the quantitative gait

analysis software.

170

Figure 33. AdV-ZFP-VEGF does not improve open-field walking (BBB) scores following

SCI. Open-field locomotion was assessed using the 21-point BBB scale. Animals were assessed

weekly for 8 weeks following injury by blinded observers. The left and right limbs were scored

individually, but the data presented are the average between left and right hindlimb recovery.

n=5-10/group (see Table 8). ** p<0.001, * p<0.01.

171

5.7.3 Delayed AdV-ZFP-VEGF administration does not improve motor evoked potentials or H-reflex following SCI

To further examine the functional changes I performed in vivo electrophysiology on the

hindlimbs of animals at 8 weeks post-SCI. My data indicates that although AdV-ZFP-VEGF

treated animals show an improved gait via Catwalk™ analysis, I did not observe any significant

improvements in axonal conduction in the hindlimbs, as assessed by motor evoked potential

recordings (Figure 34A and 34B). I also examined the H-reflex (H/M ratios) following SCI as a

measure of spasticity, and observed no significant electrophysiological differences between

groups, although AdV-eGFP treated animals demonstrated worse electrophysiological outcomes

(Figure 34B).

172

173

Figure 34. Electrophysiological assessment following AdV-ZFP-VEGF administration. A)

Representative tracings of MEP’s recorded from the hindlimb at 8 weeks post-injury. B) MEP

quantification. Recordings were obtained from hindlimb biceps femoris. Stimulation was

applied to the midline of the cervical spinal cord (0.13 Hz; 0.1 ms; 2 mA; 200 sweeps). Latency

was calculated as the time from the start of the stimulus artifact to the first prominent peak.

AdV-ZFP-VEGF did not result in improved MEP’s. C) H-Reflex quantification. Recording

electrodes were placed two centimeters apart in the mid-calf region and the posterior tibial nerve

was stimulated in the popliteal fossa using a 0.1 ms duration square wave pulse at a frequency of

1 Hz. The rats were tested for maximal plantar H-reflex / maximal plantar M-response (H /M)

ratios to determine the excitability of the reflex. AdV-ZFP-VEGF administration did not

significantly alter the H/M ratio. n=5-6/group (animals were randomly selected from a larger

population of animals within the experimental group).

5.7.4 AdV-ZFP-VEGF administration significantly reduces allodynia

A devastating post-injury condition is neuropathic pain, which affects a significant portion of

SCI patients [174, 361]. In this study I aimed to investigate the development of neuropathic

pain in AdV-ZFP-VEGF treated animals: hopeful that I would observe no increases in pain

unlike the resent report by Nesic et al. [362]. Animals were tested for thermal and mechanical

allodynia at 4 and 8 weeks following SCI, and here I observe that animals receiving AdV-ZFP-

VEGF gene therapy have a significant reduction in allodynia, for both at-level and below-level

pain, at 8 weeks post-injury (Figure 35). Testing with calibrated von Frey filaments around the

lesion site (on the dorsal skin) showed AdV-ZFP-VEGF animals to have a significant reduction

in at-level mechanical allodynia (p<0.005). An increasing application of von Frey filaments to

174

the plantar surface of the hindlimbs demonstrated a marked reduction in below-level alloydnia,

compared to injured control (p<0.05) and AdV-eGFP treated animals (p<0.005). Furthermore, I

examined below-level thermal allodynia, and observed a significant increase in pain tolerance

(increased response time) in animals receiving AdV-ZFP-VEGF (p<0.05).

175

176

Figure 35. AdV-ZFP-VEGF significantly reduces allodynia at 8 weeks post-SCI.

Mechanical and thermal allodynia, measures of neuropathic pain, were monitored with von Frey

monofilaments. Error bars represent SEM. n=5-10 per group (see Table 8). A) At-level

mechanical allodynia. Animals were assessed with 2 g or 4 g von Frey monofilaments around

the dorsal incision (above T6-T7 laminectomy and injury). Data are expressed as the average

number of adverse reactions out of 10 applications of the monofilament. There was an overall

treatment effect with AdV-ZFP-VEGF using the 2 g and 4 g monofilaments at 4 weeks and 8

weeks post-injury (ANOVA, * p < 0.05. B) Below-level mechanical allodynia. Animals were

subject to increasing von Frey filaments (2 g – 26 g), and the when they elicited a response, this

value was taken as the pain threshold value. Data are reported as the average threshold for each

group. AdV-ZFP-VEGF increased hindlimb threshold compared to other injured groups (* p <

0.05, ** p < 0.005). C) Below-level thermal allodynia. A 50°C thermal stimulus was applied to

the distal tip of the tail. The data shown are the average time it took for the animals to withdrawl

their tail from the stimulus (“tail flick”). Shorter response times indicate a decreased pain

threshold. Animals treated with AdV-ZFP-VEGF showed an increased tolerance/threshold to

thermal stimuli at 8 weeks post-injury compared to other injured groups (* p < 0.05).

5.7.5 AdV-ZFP-VEGF treatment results in spared grey matter, but not white matter tissue at 8 weeks post-SCI

Eight weeks after SCI, spinal cord cross-sections were stained serially with LFB-HE.

Measurements of tissue sparing were calculated using StereoInvestigator software, and are

expressed as the average cross-section area. Spinal cords from AdV-ZFP-VEGF treated rats did

not show evidence of white matter tissue sparing compared to control injured animals (Figure

177

36A); however, AdV-ZFP-VEGF administration exhibited an overall increase in residual grey

matter in (sections spanning 2 mm rostral and 2 mm caudal to the injury epicenter) when

compared to tissue sections from AdV-GFP and injured control rats (Figure 36B; ** p < 0.001).

178

179

Figure 36. Tissue sparing quantification at 8 weeks post-SCI. A) Residual white matter

quantification. B) Residual grey matter quantification. AdV-ZFP-VEGF improves spinal cord

grey matter preservation. C) Representative sections are shown from each group. Sections

shown are taken 1 mm rostral to the epicenter at 8 weeks after SCI. AdV-ZFP-VEGF treated

spinal cord exhibited a larger extent of grey matter spared tissue (** p < 0.001). Data are mean ±

SEM values (n = 5-10/group, see Table 8).

5.8 Discussion

In the present study, I investigated the effects of 24 hour delayed AdV-ZFP-VEGF on the

functional recovery and neuroanatomical preservation following thoracic SCI. Catwalk™

analysis demonstrated that animals treated with AdV-ZFP-VEGF showed improved hindlimb

gait and weight support, as well as improved forelimb gait, even though no differences were

observed by BBB testing or electrophysiological assessment. Moreover, results indicated that

AdV-ZFP-VEGF drastically reduced the development of thermal and mechanical allodynia in

animals at 8 weeks post-injury, suggesting that AdV-ZFP-VEGF may prevent the

development/onset of neuropathic pain. Lastly, results showed that AdV-ZFP-VEGF spared a

significant amount of grey matter tissue compared to other injured groups.

Despite extensive research efforts, there is still no therapy for spinal cord injury (SCI). With

that, our laboratory and many others believe that continuing SCI and neurotrauma research is

critical. Further research will contribute to our understanding of SCI pathophysiology and help

us identify potential therapeutic agents which may be suitable for clinical translation. To

accurately mimic patient treatment post-SCI, developing therapies must consider the reality of

180

delayed surgical or pharmaceutical intervention. The literature suggests that early intervention

following SCI, either surgically or pharmaceutically, is beneficial and that longer wait times

result in lesser outcomes and a poorer long-term prognosis [363, 364]. Ideally, patients would

receive immediate treatment post-SCI; however, realistically, patients are more likely to be

admitted to a trauma center and treated within 8-24 hours. In this research I used a 24 hour time

point to assess if AdV-ZFP-VEGF would be a viable treatment option for SCI.

The Basso Beattie Bresnahan (BBB) scoring scale to assess hindlimb deficits in thoracic SCI has

been, and continues to be the “gold standard” for functional assessment [337]. The scoring

system evaluates the hindlimb joint movement and the hind-paw orientation/stepping, provides a

general indication of the locomotor capabilities of the animal, and establishes if the animal can

weight-bear. The major shortcomings of the BBB are two-fold. First, although the BBB is to be

conducted by blinded observers, the system is still highly subjective to human errors. Secondly

– and perhaps the most confounding factor – the BBB is a qualitative system: simply indicating

if the animal is competent of defined movements, providing a relatively subjective score of how

much or how well an animal can perform a task (occasional, frequent or consistent). For

detecting major functional differences in animals, the BBB scoring scale is highly effective and

easy to conduct; however, more subtle differences between treatment groups may not be

observed by BBB assessment. Additionally, since the BBB scale is not a linear relationship

between the numerical value and the functional gains associated with them, teasing out

meaningful results can become a challenge. In this research, I used both the BBB and the

Catwalk™ gait analysis software to assess functional recovery post-injury. While scoring

animals using BBB, I noted differences subtle differences between animals (some of the moved

more normally, and with greater consistency); however, these variations were not strong enough

181

to increase their BBB score. Overall, I observed no differences in BBB scores between groups.

On the other hand, Catwalk™ data indicate that AdV-ZFP-VEGF treated animals have

significant improvements in hindlimb weight support, and hindlimb swing speed. Although the

BBB is a valid, widely used method of behavioural evaluation, the Catwalk™ is a more sensitive

and quantitative outcome measure, which may reveal understated changes in recovery. From a

clinical perspective, improvements in hindlimb weight support and gait may have an important

impact on the mobility and independence of an injured individual. Interestingly, Catwalk™

analysis also revealed that AdV-ZFP-VEGF animals showed improved forelimb stride length,

suggesting that AdV-ZFP-VEGF could potentially enhance hindlimb-forelimb coordination;

although with BBB scores of 9, we did not observe hindlimb-forelimb coordination in any

injured animal group. In the histological examination of grey and white matter post-SCI, I did

not investigate sparing of specific pathways or specific neuronal phenotypes (i.e. interneurons vs.

motor neurons); however, improvements in both hindlimb and forelimb kinetics could suggest

that AdV-ZFP-VEGF may spare propriospinal interneurons, which are located at the grey-white

matter interface and are involved in coordination of limb movements [365]. Future experiments

involving AdV-ZFP-VEGF should aim to investigate the effects of AdV-ZFP-VEGF on

interneuron sparing/survival, since these cells have been attributed to regulating central pattern

generators (CPGs) and should therefore be of interest for promoting locomotor recovery

following SCI.

In compliment to our previous data showing AdV-ZFP-VEGF spares neurons (Chapter 4), in this

study I quantified residual tissue at 8 weeks post-SCI and my data shows that delayed AdV-ZFP-

VEGF administration results in improved grey, but not white matter. Sparing grey matter at T6-

T7 is likely to result in improved trunk/abdominal stability, which may account for some of the

182

behavioural improvements I observed. Taken together, these results suggest that AdV-ZFP-

VEGF likely acts by promoting survival of neuronal cell bodies. The functional outcomes

observed in this study – improved gait, hindlimb weight support and decreased pain – would be

consistent with previous research demonstrating improved propriospinal tracts (limb

coordination) [366], reticulospinal tracts (locomotion and weight-bearing stepping) [367], and

spinothalamic tracts (neuropathic pain) [368, 369] following traumatic SCI; all pathways which

surround the grey-white matter interface. A recent paper by Nesic et al. showed that delivery of

VEGF165 resulted in the development of neuropathic pain [362]. Based on this, I thought it was

important to assess neuropathic pain following injections of AdV-ZFP-VEGF, considering

increased pain would be a detrimental side-effect of any SCI treatment. In contrast to the Nesic

et al. study, I showed a marked reduction in allodynia following ZFP-VEGF administration.

Although the exact mechanisms of these findings are unknown, it is potentially due to the

delivery of multiple VEGF isoforms that leads to a better outcome. The sparing of spinal tracts

were not specifically identified; however, increased residual grey matter likely contributes to

improved pain processing pathways (interneurons) and a decrease in aberrant pain [370-372].

Future studies are required to investigate the exact mechanisms of the attenuated allodynia

observed following AdV-ZFP-VEGF administration.

5.9 Conclusions

I believe this is the first report to investigate the functional outcomes following delayed AdV-

ZFP-VEGF administration. My previous findings showed cellular and vascular improvements

following AdV-ZFP-VEGF delivery, and in the current research I demonstrate that these

183

beneficial effects can be translated into improved functional recovery, as well as attenuated

allodynia following SCI. Collectively, these data suggest that targeting vascular and

neuroprotective mechanisms by AdV-ZFP-VEGF administration may be a viable treatment for

spinal cord injury. Moreover, the delayed administration of this therapy enhances its potential to

be used clinically.

184

Chapter 6

6 General Discussion and Future Directions

I have demonstrated that there is rapid, widespread damage to the vasculature following

traumatic spinal cord injury. Although conceptually this is not surprising or a new idea; until

now, no research has examined the temporal and spatial profiles of vascular damage in a

contusive-compressive model of SCI. Previous research has shown that the vasculature is

disrupted following SCI; however, this is the first study to collectively demonstrate detailed

quantitative data about vessel loss, deceased vascular perfusion, differences between grey and

white matter vascular damage, and endogenous vascular proliferation after clip-compression

SCI. To some extent, the spinal cord initiates an endogenous effort to repair and revascularize

the injured area; however, this does not completely restore the vascular network and the vascular

damage may propagate additional secondary injury mechanisms – ultimately exacerbating the

damage. Based on the data I have collected, taken with other research dedicated to spinal

vasculature, I believe that the vasculature is an important therapeutic target following SCI.

I have shown that AdV-ZFP-VEGF administration can be delayed 24 hours following spinal

cord injury, and still provide beneficial effects. To date, these studies are the first to use AdV-

ZFP-VEGF in a delayed fashion, and one of few studies that have used any form of VEGF

therapy at a delayed point post-injury [282, 298]. I observe significant improvements at the

cellular and molecular levels, including an increased number of vessels, reduced apoptosis and

increased neurons. Additionally, I observe improved functional benefits in animals treated with

AdV-ZFP-VEGF, including increased hindlimb weight support and significant reductions in

185

neuropathic pain. Collectively, the data suggests that: (i) administration of AdV-ZFP-VEGF

results in an increase in VEGF, which may elicit effects through cell survival and angiogenic

mechanisms, (ii) AdV-ZFP-VEGF has a therapeutic time window following SCI which extends

at least until 24 hours following injury, and (iii) vascular damage following neurotrauma may be

an effective therapeutic target.

6.1 Potential mechanisms of VEGF-A treatment

VEGF, although predominantly known for its vascular properties, is now recognized as a multi-

functional signaling protein [211, 235]. Although it has been the focus of recent neurotrauma

research, its specific roles following VEGF administration and SCI are not completely

understood.

Mechanistically, VEGF acts as a ligand for three trans-membrane tyrosine kinase receptors

(VEGFR-1, VEGFR-2, and VEGFR-3) [239]. While all receptors play important roles in vivo, it

is thought that VEGFR-2 is responsible for signaling cascades such as angiogenesis,

proliferation, cell survival, cell migration, and changes in vascular permeability – many of which

are initiated following spinal cord injury [253].

Previous studies have indicated that VEGF-A, VEGFR-1 and VEGFR-2 are upregulated

following neurotrauma: perhaps as an endogenous attempt to stimulate cell survival and

regeneration [281, 286, 288, 373]. Research has shown that VEGF-dependent cell survival is

controlled by the Akt/PKB pathway, which inhibits pro-apoptotic pathways such as BAD and

186

Caspase-9 and activates some anti-apoptotic proteins, including Bcl-2, XIAP, and A1 [258, 259,

374]. Additionally, this research demonstrated that VEGF stimulates proliferation through

activation of the extracellular regulated kinase (Erk) pathway, ultimately leading to an increase

in gene transcription.

Stimulation of the Akt/PKB pathway via VEGFR-2 can also induce endothelial nitric oxide

synthase (eNOS) expression. This signalling results in the subsequent generation of NO and

increases in vascular permeability and cellular migration [260, 261]. Additionally, p38 mitogen-

activated protein kinase (p38MAPK) is also signaled through VEGFR-2 and has been shown to

mediate actin re-organization and cell migration [262]. Moreover, VEGF-A is able to induce

the migration and differentiation of endogenous hematopoietic stem cells – a property which

may be very important for repair and regeneration following SCI [211, 239].

It is clear that VEGF-A acts on various pathways and many different cell types. Although I may

not be able to selectively control which pathways are activated by AdV-ZFP-VEGF delivery, it

appears that overall upregulation of VEGF is beneficial following SCI. Here I have

demonstrated that AdV-ZFP-VEGF targets neurons and vasculature, which appears to result in

cellular and functional improvements. AdV-ZFP-VEGF administration at 24 hours post-injury is

likely acting in two separate time-windows. In the early phase (24 – 48 hours), VEGF appears to

target cell survival mechanisms (as observed by an increase in NeuN and RECA-1 – Figure 28

and 29). While in the later phase (72 hours – 7 days), increased VEGF levels are likely directed

towards revascularization mechanisms (as observed by an increase in angiogenesis at 5 days –

Figure 30). I believe that because of VEGF-A’s multifaceted properties in vivo, it is an ideal

candidate for potential SCI therapy. Moreover, using a gene therapy delivery method (such as

187

AdV-ZFP-VEGF), the prolonged upregulation of VEGF between 2-7 days post-injury allows it

to target a number of cellular processes, thereby increasing its therapeutic effectiveness.

6.2 Vasculature damage plays an important role in SCI

SCI results in significant vascular damage, including disruption of spinal cord blood flow, the

onset of spinal cord ischemia, hemorrhage, edema and breakdown of the blood-spinal cord

barrier (BSCB). These vascular changes encompass many of the earliest pathological processes

following SCI, therefore therapies aimed directly at the vascular disruption or the ensuing

downstream consequences of vascular injury are highly attractive. In theory, rapid vascular

repair following injury will likely result in the most favourable outcomes. Promoting repair and

regeneration of vascular structures would mediate ischemia, hemorrhage and further edema by

restoring proper blood-flow and stopping leaky vessels. Moreover, restoring the proper structure

and function of the BSCB would likely reduce the influx of inflammatory cells into the spinal

cord, thereby reducing the damage caused by reactive microglia [136, 137].

In the present study, I examined the effects of AdV-ZFP-VEGF – a ZFP transcription factor

designed to increase the expression of all major VEGF-A isoforms – in a well-characterized

model of compressive SCI. AdV-ZFP-VEGF has been previously used in rabbit hindlimb,

mouse ear and rat spinal cord tissues with all results showing increases in endogenous VEGF-A

expression and beneficial vascular outcomes [212, 316, 318, 319, 321]. Likewise, the data

demonstrates that VEGF-A was upregulated at both the mRNA and protein levels after AdV-

ZFP-VEGF treatment. Notably, I observed significant increases in VEGF120 and VEGF164

isoforms. VEGF121 and VEGF165 – which are the human homologues of VEGF120 and VEGF164

188

– are the predominant isoforms of VEGF-A expressed in the human central nervous system and

appear to be the key players in angiogenesis in the spinal cord [375]. Further studies are required

to determine whether AdV-ZFP-VEGF results in improved blood-flow or BSCB repair;

however, the current data suggests that AdV-ZFP-VEGF has positive effects on the

microvasculature following SCI, since AdV-ZFP-VEGF treated animals have a greater number

of blood vessels 10 days following SCI compared to control animals.

6.3 Targeting the neurovascular niche

Therapies specifically targeting vascular damage, vessel density and restoration of blood flow to

the injured spinal cord may provide an opportunity for spinal cord repair and recovery [376].

Recent reports have shown significant correlations between blood vessel density and

improvements in recovery following CNS trauma [324-326, 377]. Rescue and regeneration of

the microvasculature within the epicenter and penumbra remains largely unexplored, yet may be

a promising therapeutic route to facilitate tissue sparing and functional recovery following SCI.

It has been shown that substantial trophic support is provided by CNS microvessels [302] and

that microvessels are critical for tissue survival [291].

VEGF supports the “neurovascular niche”, as it appears to play important roles in both vascular

and neural development, bridging both endogenous systems. Expression of VEGF-R’s have

been observed in many cell types, including neurons, microglia/macrophages, endothelial cells,

smooth muscle cells and astrocytes [281, 285, 286, 288, 292, 293]. Through interactions with

co-receptors, Neuropilins, VEGF is able to influence the function and development of neural

cells, which may be a key role for VEGF therapies following neurotrauma [246, 378].

189

Additionally, studies have shown that regenerating axons have a tendency to grow along blood

vessels, therefore promoting vascular growth following injury may provide scaffolding for

regenerating axons [332, 347].

6.4 Advantages of AdV-ZFP-VEGF compared to other VEGF therapies

Due to the important pleiotropic functions of VEGF, it has been a popular research topic in

recent years, specifically in neurotrauma research [279, 285, 286, 289, 294, 359, 373].

Complementary to this research, recent publications have also examined the role of VEGF-A in

models of SCI and shown positive results. In a weight-drop SCI model, rats treated with

VEGF165 showed significantly improved behaviour after SCI, notable repair of blood vessels and

reduced apoptosis [298]. Choi et al. induced VEGF-A expression using a hypoxia-inducible

system to treat SCI and demonstrated increased VEGF expression and observed neuroprotective

effects [359]. Similarly, Facchiano et al. used an adenovirus coding for VEGF165 in a partial

spinal cord transection model and detected a significant increase in vessel volume and a

reduction of corticospinal tract axon degeneration [360]. Lastly, another study reported that

delayed administration of recombinant human VEGF165 (48 hours post-ischemia) to ischemic

rats enhanced angiogenesis and significantly improved neurological recovery [379, 380].

However, the previously described approaches using VEGF-A have relied on the introduction of

a single splice isoform of VEGF-A (VEGF165), which may not result in optimal neuroprotective

or angiogenic effects. Research has shown that the administration of a single VEGF isoform

(typically VEGF164/165) results in improper vascular regeneration and repair, creating leaky

vessels which can further contribute to the SCI pathophysiology [354, 379].

190

ZFP-VEGF technology – a viral vector encoding a zinc-finger transcription factor protein (ZFP),

which activates endogenous VEGF-A expression – has been previously used to demonstrate that

expression in vivo leads to induced expression of VEGF-A protein, stimulated angiogenesis, and

accelerated wound healing [316]. I believe that this novel approach presents advantages over

previous attempts to use VEGF-A as an angiogenic target because AdV-ZFP-VEGF upregulates

endogenous VEGF-A expression, thereby mimicing physiological VEGF production and the

expression of multiple VEGF isoforms in the injured spinal cord – a requirement for proper

vascular development and functional angiogenesis [381].

6.5 Potential disadvantages of VEGF therapies

Although VEGF has many desirable attributes for neuroprotection and vascular repair, it is

important to recognize that some of these attributes have the potential to be deleterious and

exacerbate damage following SCI. In development or maintenance of vascular structures, VEGF

stimulates angiogenesis by signaling matrix-metalloproteinases (MMPs) to breakdown the BSCB

and matrix in order to make way for new vascular sprouts. However, following injury, greater

amounts of VEGF are released from surrounding cells and vascular remodeling is quickly

initiated, which leads to a rapid hyperpermeability of the local vessels. This increase in

permeability may contribute to an increased inflammatory response or increased edema

following CNS injury. In particular, disruption of the BSCB following injury presents an entry

route for inflammatory mediators to enter the CNS without resistance. A previous study reported

that VEGF is able to promote monocyte migration in vitro and that administration of VEGF

therapies may contribute to inflammatory responses following injury [382]. Although I observed

no increased inflammatory response in AdV-ZFP-VEGF animals compared to other injured

191

control groups at 10 days following injury (data not shown), it is possible that VEGF therapies

exacerbate early inflammation.

Varying studies report that 26-96% of human patients experience neuropathic pain following SCI

[361]. Research has identified VEGF as one of the potential factors involved in the development

of neuropathic pain; however, it is still unclear if VEGF plays a beneficial or detrimental role.

Schratzberger et al., found that intramuscular injections of VEGF improved vascularity, blood

flow and peripheral nerve function in a rabbit model of diabetic neuropathy [383]. Since it is

believed that diabetic neuropathy is caused from microvascular ischemia, their findings

reasonably support the use of VEGF for the treatment of neuropathies. Conversely, Nesic et al.

recently showed that VEGF administration into the spinal cord resulted in an increased number

of animals displaying neuropathic pain, as well as an increase in myelinated dorsal horn neurons,

suggesting that VEGF results in non-specific axonal sprouting [362].

Regardless of whether VEGF therapies result in favourable or damaging outcomes, it is most

important for future research to be aware of potential pitfalls of VEGF administration and to

consider the implications they may have on the bench-to-bedside translation of these therapies.

6.6 Comparison of Results to Other SCI Therapies

6.6.1 AdV-ZFP-VEGF: Immediate vs. 24-hour Administration

Previously, our lab examined AdV-ZFP-VEGF administration when delivered immediately

following SCI (See Appendix 1). In the previous study, we aimed to investigate if AdV-ZFP-

192

VEGF functioned in vivo, and resulted in beneficial outcome. In the present study, the main

objective was to investigate if AdV-ZFP-VEGF administration could be delayed, while retaining

its protective properties that we previously observed.

When comparing data between 0 hour and 24 hour administration, I observed very similar VEGF

mRNA and protein expression profiles. VEGF120 and VEGF164 mRNA are increased

approximately 2-fold over uninjured animals, and VEGF protein is also increase by

approximately 2-fold.

Following 24 hour administration of AdV-ZFP-VEGF, I observed a 15-25% increase in RECA-1

cells, whereas following 0 hour administration (our previous research) we observed closer to a

35% increase in RECA-1-positive vessels. For future research, this may be an important

consideration for determining the optimal delivery time. Obviously, therapies cannot be

administered immediately post-injury; however, this suggests that an earlier intervention would

associate with an improved vascular recovery (by approximately 10%).

In comparing 0 and 24 hour delivery time-points, I also observe very similar results for NF200

preservation and apoptosis. NF200 quantification at both time-points demonstrates a 2-fold

increase in NF200 expression compared to other injured groups. TUNEL quantification

following 0 hour administration results in an average of 50% reduction in apoptotic cells,

whereas, following 24 hour administration I observe an average of 30% reduction of apoptosis.

Again, although delivery at 24 hours results in a significant decrease in cell death, comparison of

the data from two administrations suggests that an earlier delivery may provide a more

substantial neuroprotective effect.

193

Evaluation of functional data between 0 hour and 24 hour administration, is not directly

comparable since the study conducted by Liu et al. uses AAV vectors for the long-term

experiments (AAV-ZFP-VEGF and AAV-DsRed) [212]. Regardless, for both studies, we

observe that VEGF treated animals improve to 8-9 on the BBB scale and control animals (eGFP

or DsRed) achieve a BBB score of 6. The consistency of these results highlights the reliability

and reproducibility of both the clip-compression injury model, as well as the delivery/injection

technique across individual researchers and laboratory conditions. Additionally, it may be

possible for both AdV and AAV vectors to improve function; however, as seen in Figure 37, I do

not show improvement in animal locomotion when AAV-ZFP-VEGF is delivered 24 hours post-

SCI. This may be due to the delay of onset of AAV vector kinetics in vivo, which is discussed in

more detail in the discussion below (see Section 6.7.2). Overall, between 0 hour and 24 hour

ZFP-VEGF administration, we observe strikingly similar results; however, the current research

has included additional experimental groups (namely an “injured control” group which does not

receive intraspinal injections), which slightly alters our interpretation of the BBB data. Here we

observe the same differences in BBB scores between ZFP-VEGF and eGFP treated animals

(comparing 0 and 24 hour time points), except now we observe AdV-ZFP-VEGF and Injured

Control animals display similar BBB scores. By using Catwalk™ analysis, we are able to show

that AdV-ZFP-VEGF animals are actually performing better than Injured Control animals;

however, I believe that including an “injured control” group is an important experimental

control, since ultimately you should be able to answer the question: Is administering a therapy

better than doing nothing?

194

195

Figure 37. 24 hour delayed AAV-ZFP-VEGF administration does not result in beneficial

outcomes following SCI. A) BBB open-field locomotor scores. Animals were scored weekly

by two blinded observers to assess himb-limb locomotion and functional recovery. No

differences in BBB scores were observed between the three injured groups. B) Residual tissue

sparing quantification. HE/LFB staining was performed on tissue sections at 6 weeks post-

injury, and residual tissue area was quantified. Values were normalized to the tissue area of an

uninjured animal, and the graph displays data as a percentage of sham tissue. Among the injured

animals, groups did not show significant differences in spared tissue.

6.6.2 Comparison to Other Vascular Therapies

To assess the effectiveness of a therapy, it must be compared to other similar studies. Multiple

research groups have investigated the role of vascular factors following neurotrauma; however,

very few have applied a vascular therapy as a potential treatment for spinal cord injury. Benton

et al. have administered VEGF165 at 3 days post-injury in a model of contusive SCI, and have

concluded that this therapy exacerbated vascular permeability and lesion volume [354]. In

comparison to AdV-ZFP-VEGF administration that upregulates multiple gene isoforms,

VEGF165 delivery alone is not likely to result in proper vascular formation, which could account

for increase vascular permeability. Moreover, delivery at 3 days post-injury might be too

delayed to provide beneficial effects, effectively missing the therapeutic window.

Bakshi et al. implanted a non-biodegradable scaffold into the spinal cord following transection

injury to provide a physical matrix for vascular regeneration [384]. The authors observed

196

significant angiogenesis in animals with the scaffold; however, the experiments did not include

assessment of functional recovery.

Another study conducted by Kim et al. examined transplanting neural precursor cells (NPC),

which were engineered to over-express VEGF (VEGF-NPC) [385]. This study used a contusion

model that resulted in a comparable injury severity to our model of clip-compression, and

applied a combinatorial approach to VEGF treatment. The data are very promising. In

summary, the results showed a relatively similar increase in expression of VEGF levels

compared to the ZFP-VEGF method of gene activation (Figure 25). However, the additional

transplantation of NPCs resulted in proliferation of the cells into NG2+ cells, ultimately creating

new oligodendrocytes and resulted in replacement of lost/damaged white matter. They also

showed the VEGF-NPC treatment resulted in angiogenesis; the improvements were analogous to

my observations (Figure 30). Functionally, the VEGF-NPC animals improved significantly

compared to other controls, which the authors predominantly attribute to the cell replacement

characteristics of the therapy, but indicate that VEGF over-expression likely played a key role in

cell survival and proliferation of the NPC transplanted cells. VEGF-NPC animals have

improved BBB scores compared to the BBB scores observed following AdV-ZFP-VEGF

administration (Figure 33).

6.6.3 Comparison to Other Neuroprotective Therapies

Decades of research have been dedicated to finding a neuroprotective therapy for spinal cord

injury and countless cellular and drug therapies have been examined. Here I aim to put my data

in perspective against some of the key therapeutics that have shown encouraging results as a

197

treatment for SCI. Overall, the results displayed following delayed AdV-ZFP-VEGF

administration are comparable, if not more robust than other published data in the field.

Riluzole

Riluzole was designed as a Na+ channel blocker, to attenuate pathophysiology associated with

cellular toxicity and subsequent degeneration [196]. In previous research, Schwartz and Fehlings

have shown riluzole to have beneficial effects following traumatic SCI [197]. They show

between 0-25% of grey matter spared by riluzole delivery, depending on the distance from the

lesion epicenter. In contrast, I observed between 18-50% grey matter sparing following AdV-

ZFP-VEGF administration, depending on the distance from the epicenter (Figure 36). Schwartz

and Fehlings demonstrate a significant improvement in BBB scores as a result of riluzole

administration (4 points), whereas my research shows no notable improvements in BBB scores

(Figure 33).

Minocycline

Stirling et al. use minocycline as an anti-inflammatory and neuroprotective agent, since previous

research has shown it to exhibit inhibition of microglial activation and inhibition of caspases,

respectively [386]. In this study, they demonstrate minocycline to have a 30% reduction in

apoptosis, which is not significantly varied from my results, which showed a reduction of

apoptosis by approximately 20% (Figure 26). In this study, they used footprint analysis as an

outcome measure for functional recovery, which is not directly comparable to the methods used

in this thesis; however, they observed an increase in function of approximately 40%, which is a

robust effect, yet less than my observations of close to 50-70% increases in hindlimb recovery as

detected by Catwalk™ analysis (Figure 31 and 32).

198

Methylprednisolone

Weaver et al. used a 35g clip-compression model of SCI at T12 to investigate the effects of

methylprednisolone on SCI recovery [387]. Methylprednisolone is a clinically accepted

treatment for human cases of SCI, and is involved in reducing inflammation and provides anti-

oxidation by reducing lipid peroxidation. In this study, Weaver et al. display that animals treated

with methylprednisolone recover to a maximum BBB score of 8, which is approximately what I

observe following AdV-ZFP-VEGF, albeit the injury model used in my studies was at T6-T7

(Figure 33). Methylprednisolone appears to reduce neuropathic pain; however, the data observed

by Weaver et al. is not as significant as the findings following AdV-ZFP-VEGF (Figure 35), and

their research only indicates that methylprednisolone improves at-level pain and not below-level

pain. Lastly, Weaver et al. demonstrate that methylprednisolone results in significant sparing of

neurofilament compared to injured control animals. In my studies using AdV-ZFP-VEGF I

assess NF200 and tissue sparing, and although these are similar outcome measures, the data

cannot be directly compared. By indirect comparison, it appears that methylprednisolone results

in less significant sparing as compared to AdV-ZFP-VEGF treatment.

6.7 Future Research

6.7.1 Investigating the Glial Scar and Inflammation

Following SCI, a large influx of inflammatory mediators enters the injury site [135]. Early

inflammation is considered to be a beneficial process, since it involves the phagocytosis of

cellular debris and dead/dying cells, which is helpful to “clear” the injury site of unwanted

material [111]. However, persisting inflammation results in the formation of a glial scar around

199

the injury site, which creates a restrictive environment and inhibits endogenous regenerative

processes [115]. Moreover, inflammation may be the source of neuropathic pain, therefore

determining the correct way to modulate inflammatory responses may be key for managing

aberrant pain [115, 171].

The inflammatory response following injury is greatly facilitated and enhanced by the

breakdown of the BSCB [115]. BSCB breakdown triggers a post-traumatic inflammatory

response, which involves the invasion of macrophages and neutrophils. Moreover, both injured

endothelial and glial cells release vasoactive substances (i.e. reactive oxygen molecules, nitric

oxide and histamines) that boost spinal cord blood flow and aid in plasma-derived molecules

entering the cord [118, 388]. Vascular injury – a major factor of the secondary injury

pathophysiology – therefore plays a vital role in initiating and regulating post-traumatic

inflammation. Studies have shown that BSCB/BBB breakdown is maximal 1 day post-CNS

injury, which is consistent with our observations following SCI (Chapter 3) [118, 388].

Moreover, other studies have examined the extent of BSCB breakdown, and demonstrated that

the disruption extends along the axis of the injured spinal cord and, therefore disruption is not

exclusively observed at the site of injury [97, 118, 389]. These findings are also consistent with

our results, as we showed vascular disruption up to 1 mm distal (rostral and caudal) to the injury

epicenter (Figure 19). In the studies presented in this thesis we only examined the BSCB

disruption as it pertained to vascular permeability; however, future studies may wish to critically

examine the role of the BSCB and its involvement in the post-traumatic inflammatory response.

Considering VEGF plays important roles in many in vivo pathways, including vascular

permeability and cellular recruitment [244, 253], it may be interesting to investigate the

200

inflammatory role of AdV-ZFP-VEGF delivery in the spinal cord. In Chapter 3, I define the

spatial and temporal events of vascular damage as it pertains to the injury (no treatment);

however, now that these data are available as a comparison, future studies may wish to examine

how AdV-ZFP-VEGF administration alters BSCB permeability and quantify the acute

inflammation following injection. Results from the 8 week experiments demonstrate spared grey

matter and significant reductions in pain, therefore it seems plausible that in this model of SCI

and vascular gene therapy, VEGF may play a role in mediating neuroinflammation following

injury. For future research, it would be interesting to know: Does AdV-ZFP-VEGF alter glial

scar formation? Choi et al. have examined spinal cord tissue following injury, and

demonstrated that endogenous VEGF plays a role in macrophage/microglia recruitment and

astroglial response [373]. Therefore, it is reasonable to hypothesize that AdV-ZFP-VEGF may

play a role in the inflammatory response and scar formation, although based on the results that I

have shown (displaying beneficial outcomes from AdV-ZFP-VEGF), it is likely that ZFP-VEGF

is modulating or minimizing these events.

6.7.2 Alternative ZFP-VEGF Delivery

Understandably, therapies for spinal cord injury should aim to be minimally invasive. As a

proof-of-concept, I have chosen to administer the AdV-eGFP or AdV-ZFP-VEGF directly into

the spinal cord surrounding the injury site; however, I recognize that this delivery mechanism

may not be ideal. Although not perfect, direct injection into the spinal cord has some advantages

and has been widely used for the delivery of therapeutics and stem cells [76, 390, 391]. Firstly,

this method allows specific and localized delivery of the ZFP-VEGF gene therapy. I have

selected two injections sites two-millimeters rostral and caudal to the injury site to target the

201

penumbra of the injury – a site which is more likely to be rescued by a delayed therapeutic

intervention compared to the injury epicenter. Additionally, direct injections result in rapid

delivery of the therapy, since it is not required to migrate or circulate before reaching the target

tissue. In AdV-eGFP treated groups, I observed decreased NeuN counts, increased TUNEL-

positive cells and a decrease in RECA-1-positive vessels compared to animals which received

only a compression injury (injured control group). These deficits are likely attributable (or at

least in part) to additional damage caused by the intraspinal injections. Importantly, AdV-ZFP-

VEGF treated animals were able to overcome these additional deficits and still show significant

improvement compared to injured control animals. Future experiments will investigate

alternative delivery methods for AdV-ZFP-VEGF, since I hypothesize that VEGF-treated

animals may display even greater histological improvements over control animals if AdV-ZFP-

VEGF were to be administered in a less invasive manner.

One approach I investigated was the use of adeno-associated viruses (AAV) to deliver ZFP-

VEGF in vivo. AAV vectors have some advantages over AdV vectors; namely they evoke a

negligible immune response in the host system [314, 392, 393]. This, obviously, would be a

major benefit for any CNS therapy. However, AAV vectors contain a few characteristics that

make them less than ideal for treating SCI. First, AAV vectors are slow to initiate expression of

their genetic material. In contrast to AdV vectors that stay episomal, AAV vectors integrate into

the host DNA prior to replication, therefore they show a delayed onset of their genes [306, 315].

AAV vectors generally begin expression by 5 days following injection; however, since I was

delaying the administration of AAV-ZFP-VEGF to 24 hours post-SCI, VEGF upregulation may

not have started until 6 days after the injury. In a preliminary experiment, I observed that AAV-

ZFP-VEGF did not provide beneficial effects following SCI (Figure 37). I believe this is due to

202

the delayed onset of VEGF expression, and thus the therapeutic window for intervention was

missed. Secondly, AAV vector integration into the host chromosomes has the potential to

produce genetic mutations and errors. This is a concern for delivering gene therapies, since you

would not want to deliver a treatment at the expense of disrupting the reading frame of the host

DNA. Lastly, integration of AAV vectors into the host can result in permanent genetic

modifications. For this reason, AAV vectors may be ideal for treating hereditary conditions, but

for treating rapid, traumatic injuries, it may be unnecessary and possibly harmful to have long-

term overexpression of a target gene. In particular, long-term overexpression of VEGF may lead

to detrimental effects. In fact, it has been well-established that VEGF plays a key role in tumor

growth and tumor angiogenesis [394].

Future experiments on this project should investigate the use of double-stranded AAV vectors,

which have similar expression kinetics to AdV vectors without evoking inflammation, or

alternatively, non-viral delivery systems such hyaluronic acid and methyl cellulose (HAMC),

which has been shown to be an effective delivery method and inherently neuroprotective [395,

396]. Utilizing HAMC to deliver VEGF has a number of advantages. First, HAMC is injected

into the subarachnoid space, whereas ZFP-VEGF has always been injected 1 mm deep into the

spinal cord. Although I have tried to make each injection minimally invasive, four injections

into the cord undoubtedly results in some physical trauma: adding to the extensive damage

already done by the injury itself. Applying a therapy to the dorsal surface without causing

additional damage the spinal cord would be ideal. Secondly, HAMC has been designed to

degrade over time, and with varying ratios of hyaluronic acid (HA) and methyl cellulose (MC)

the degradation kinetics can be modified. The degradation of HAMC provides a slow,

continuous release of any drug/therapy suspended in the HAMC, which may have therapeutic

203

benefits. Earlier, I discussed a major disadvantage of AAV vectors as resulting in long-term,

permanent delivery, but HAMC may present an option for the constant administration of

therapies over a short period of time (3-5 days post-injury).

6.7.2.1 Implications of AdV and Cyclosporin-A Administration

Using adenoviral vectors for gene therapy are a efficient, rapid, and easy method of inducing

gene expression; however, AdV constructs have inherent disadvantages. AdV vectors have been

shown to initiate an innate immune response in host tissues, activating a non-specific cascade of

cytokines, leukocytes, neutrophils and macrophages [306, 308]. This immune response enhances

the inflammation already occurring as a result of SCI, which may (or may not) exacerbate the

pathophysiology [115, 135]. Potentially, this immune response may stimulate the endogenous

angiogenic response, as inflammation and angiogenesis share common pathways in wound

healing mechanisms and in disease pathology [397-399]. Briefly, inflammation results in the

expression of cytokines and recruitment of macrophages/microglia (which in turn express more

cytokines). MMP’s are expressed and activated also. Endothelial cells respond to this cytokine

expression, and are recruited to the site of inflammation. MMP’s help to destabilize local

vascular structures (i.e. the basement membrane), which promotes branching and remodeling,

ultimately leading to angiogenesis. Based on this, it is therefore possible that using an AdV

vector to deliver ZFP-VEGF, may be stimulating endogenous angiogenic mechanisms.

However, considering AdV-eGFP treated animals show decreases in most outcome measures, it

is not anticipated that delivering an adenovirus significantly contributes to angiogenesis (via

enhanced inflammation) in AdV-ZFP-VEGF treated animals.

204

Aside from the inflammatory response, using a non-viral method of ZFP-VEGF delivery would

be advantageous, since in the current research animals receiving AdV-ZFP-VEGF also received

immune-suppression via cyclosporin-A (CsA). Cyclosporin-A is used in an attempt to reduce

the host response to the viral vector, although CsA most notably blocks T-cell function (adaptive

immunity), which is not the primary immune response initiated by AdV vectors [307]. The

implications of CsA use in vivo are two-fold: a) imunosuppression may diminish a significant

portion of the inflammation caused by the injury, some of which is beneficial and promotes

endogenous recovery (as discussed above) [111], and b) cyclosporin-A has been shown to have

neuroprotective properties [199, 400]. CsA as a neuroprotective agent is a relatively new

concept; however, it has been shown that administration of CsA may act in two potential

mechanisms. First, Diaz-Ruiz et al. have shown that cyclosporin-A decreases lipid peroxidation,

and results in improved functional outcomes [401]. Here they suggest that CsA acts to reduce

the humoral inflammatory response, which decreases the production of free radial/reactive

oxygen species (ROS), which ultimately reduces the lipid peroxidation and degradation of cell

membranes. A second mechanism has been proposed whereby CsA acts to reduce mitochondrial

dysfunction and leads to neuronal survival [402-405]. Following neurotrauma, excess Ca2+ is

released, and in an attempt to maintain homeostasis, mitochondria uptake Ca2+ via a

mitochondrial permeability transition pore (MPTP). Large amounts of Ca2+ inside the

mitochondria decrease the membrane potential, leading to the uncoupling of the electron

transport chain, production of ROS and subsequent production of ATP. This initiates cell death

via necrosis. Alternatively, high concentrations of intra-mitochondrial Ca2+ results in the

activation of caspases and cytochrome C, which are known initiators of apoptosis. CsA has been

shown to inhibit MPTP, which prevents Ca2+ influx, mitochondrial dysfunction, and ultimately

cell (neuronal) death [404, 405].

205

6.7.3 Elucidating the Functional and Sensory Benefits of AdV-ZFP-VEGF

Currently, I have hypothesized that the improved functional outcomes observed in animals

treated with AdV-ZFP-VEGF are likely a result of spared neurons (I observe increased grey

matter sparing), specifically those involved in propriospinal (limb coordination), reticulospinal

(locomotion and weight supported stepping), or spinothalamic (pain) tracts. However, in my

experiments, I only examined white matter and grey matter sparing and did not specifically

analyze individual tracts. In future research, anterograde or retrograde labeling experiments

should be conducted to further understand which ascending and descending pathways are

preserved as a result of AdV-ZFP-VEGF administration.

6.7.4 Imaging Vascular Changes Using a Spinal Cord Window Chamber

Imaging the spinal cord and its vasculature presents additional challenges compared to other

CNS targets. The spinal cord is located deep in the abdomen, making it physically difficult to

access, and the structural organization of the cord is such that the majority of vascular structures

are in the interior grey matter [8, 17]. These anatomical obstacles have made in vivo spinal cord

imaging difficult in the past, often resulting in multiple surgeries to expose the area of interested

for repeated or long-term imaging. To overcome some of these barriers, I was involved in

designing, developing and testing a spinal cord window chamber (SCWC) (See Appendix 2).

The SCWC is surgically implanted into the dorsal skin of an animal following a multi-level

laminectomy, and provides a clear optical window to conduct multimodal imaging in vivo. As

described by Figley et al., implantation of the device does not result in local inflammation or

functional impairment, and the device is compatible for fluorescent and confocal microscopy, as

206

well as Doppler, speckle-variance optical coherence tomography (svOCT) and photoacoustic

imaging. The device was generated in both stainless steel and polycarbonate designs, to allow

for a diverse application and data collection.

Now that the technology has been tested as a proof-of-concept project, future studies should use

this device to monitor vascular changes following traumatic SCI. Sophisticated software and

mathematical algorithms are available to quantify vessels and capillaries, and given the correct

equipment, deep grey matter vessels may be observed in the rat cord. There are two main

advantage to using the SCWC to assess changes following injury. First, each animal can serve as

its individual control, where you quantify the vasculature prior to injury and compare changes

within each animal. Second, from an ethical stand-point, this technology would drastically

reduce the number of animals required in an experiment. Repeated, in vivo imaging would allow

you to monitor the same animal at various time points, instead of sacrificing groups of animals at

designated time-points to examine the vasculature.

6.7.5 Further Investigation of the Blood-Spinal Cord Barrier

My research has provided an initial overview of vascular damage and BSCB disruption that

occurs as a result of traumatic SCI; however, this was by no means an exhaustive study. I

observed that the BSCB is significantly disrupted early on (by 1 hour post-SCI) and remains

significantly open for 5 days. These data provide an aspect of BSCB dysfunction, but what

changes occur at the cellular and molecular level remains unknown. Future studies should focus

on examining the expression and distribution of tight junction and adheren junction proteins to

determine what factors contribute to the BSCB “leakiness”. Furthermore, vascular architecture

207

should be investigated to determine the maturity of vascular structures, such as the extent of

basal lamina deposition, and the association of pericytes and astrocytic endfeet. These data

would provide insightful information and may identify key therapeutic targets that can be

addressed following SCI.

In the current study, I examined BSCB and vascular disruption at 1 and 4 hours, 3, 5, 7, 10, and

14 days post-injury. I selected these time-points as a detailed range of acute time-points;

however, chronic time-points were neglected. Popovich et al. have observed a bi-phasic opening

of the BSCB following SCI, suggesting that peaks of disruption occur at 24 hours and then again

between 14 and 28 days [66]. Future studies should investigate time-points extending past 14

days, to determine the long-term BSCB deficits that result from SCI. Additionally, as previously

suggested in the discussion of Chapter 3, future studies might also wish to consider using a

smaller vascular marker than Evans Blue dye, since its relatively large size may limit the

detection of vascular permeability.

Finally, future research should repeat the BSCB experiments with the addition of AdV-ZFP-

VEGF and AdV-eGFP treated groups. The aim of this project should be to determine what

effects AdV-ZFP-VEGF has on BSCB permeability, if any. The results of this study would be

interesting since varied results have been published on VEGF delivery and vascular

permeability, and both outcomes are plausible based on the diverse properties of VEGF. Patel et

al. published research indicating that VEGF administration improved acute BSCB permeability

[406], while others report VEGF-induced breakdown of the BSCB [288, 379].

208

6.7.6 Multiple Angiogenic Factors as a Potential Treatment Option

Many reports have suggested that a combinatorial approach will be the best chance at curing

spinal cord injury, as no one therapy will be able to adequately target the complex cellular

pathology [193, 407, 408]. It is first important to assess individual treatments in order to

understand their risks and benefits equally; however, once a treatment has been screened for

efficacy and safety, it only seems logical to pair with another therapy that has been shown to be

effective towards another aspect of SCI.

In the current research, I solely examined the administration of VEGF (via a bioengineered

transcription factor) with the aim to promote vascular repair following injury. VEGF, in itself,

could be classified as a “combination approach” since it is involved in diverse cellular pathways

in vivo: hence the basis of its appeal for a SCI therapy. I demonstrated that AdV-ZFP-VEGF

administration results in many beneficial outcomes, with no adverse effects noted. Although

VEGF plays a many important roles in angiogenesis and wound healing, the process of vascular

regeneration is elaborate and coordinated with numerous other factors required [222]. Therefore,

it could be assumed that delivering multiple angiogenic factors would only provide improved

revascularization, whether it increase the rate of growth, improve vessel maturation or promote

necessary branching. A few other researchers have investigated the possibility of combining

VEGF with alternate therapies. Lutton et al. explored the combination therapy of VEGF and

platelet derived growth factor (PDGF) and shown promising outcomes, compared to VEGF

treatment alone [409]. Another attractive research study examined the combination of VEGF

and neural precursor cells (NPC), whereby the NPCs were engineered to over-express and

secrete VEGF in vivo, and the results showed beneficial outcomes following SCI [385]. The

209

possibility of combining VEGF with other treatments is vast, and I believe that a VEGF

combination therapy has scientific merit and should be investigated in future studies.

6.8 Final Conclusions

I have studied the BSCB disruption, vascular damage and endogenous revascularization that

occur following SCI, which presents important temporal and spatial data. These results highlight

the maximal disruption and regenerative processes, and in turn, suggest cellular events and

locations within the cord that could be pharmacologically targeted to enhance recovery.

Maximal BSCB disruption occurs at 24 hours following SCI, therefore drug administration is

likely to be the most effective while the barrier is open. Endogenous angiogenesis occurs

between 3 and 7 days, with maximal endothelial cell proliferation occurring at 5 days post-

injury. With that, I suggest that vascular therapies be administered before 3 days in order to

amplify the endogenous repair processes.

I also studied the acute and long-term effects of 24 hour delayed administration of AdV-ZFP-

VEGF on vascular regeneration, neuroprotection and functional recovery. I observe a number of

beneficial outcomes as a result of AdV-ZFP-VEGF administration, indicating that this therapy

can be delayed to a clinically relevant time point (24 hours post-SCI) and still elicit positive

effects. In general, the data suggest that AdV-ZFP-VEGF could be a suitable candidate for SCI

treatment, although further characterization of the gene therapy may be required. Overall, I

believe that addressing the vascular damage following SCI is an important therapeutic target, and

that VEGF and other vascular therapies are likely to result in promising outcomes following

traumatic injury.

210

References

1. Haisma, J.A., et al., Complications following spinal cord injury: Occurrence and risk factors in a longitudinal study during and after inpatient rehabilitation. Journal of Rehabilitation Medicine, 2007. 39(5): p. 393-398.

2. Sekhon, L.H. and M.G. Fehlings, Epidemiology, demographics, and pathophysiology of acute spinal cord injury. Spine, 2001. 26(24 Suppl): p. S2-S12.

3. Tator, C.H. and M.G. Fehlings, Review of the secondary injury theory of acute spinal cord trauma with emphasis on vascular mechanisms. J Neurosurg, 1991. 75(1): p. 15-26.

4. Beattie, M.S., Inflammation and apoptosis: linked therapeutic targets in spinal cord injury. Trends in Molecular Medicine, 2004. 10(12): p. 580-583.

5. Fehlings, M.G., C.H. Tator, and R.D. Linden, The relationships among the severity of spinal cord injury, motor and somatosensory evoked potentials and spinal cord blood flow. Electroencephalogr Clin Neurophysiol, 1989. 74(4): p. 241-59.

6. Nieuwenhuys, R., J. Voogd, and C. van Huijzen, The Human Central Nervous System: A Synopsis and Atlas. 4th Edition ed2007, Berlin: Springer Berlin Heidelberg. 440.

7. Hooper, S.L., Central Pattern Generators, in eLS2001, John Wiley & Sons, Ltd.

8. Standring, S., Gray's Anatomy: The Anatomical Basis of Clinical Practic, Expert Consult. 40th ed, ed. S. Standring2008: Churchill Livingston Elsevier.

9. Richardson, J. and G.J. Groen, Applied epidural anatomy. Continuing Education in Anaesthesia, Critical Care & Pain, 2005. 5(3): p. 98-100.

10. Bechmann, I., I. Galea, and V.H. Perry, What is the blood brain barrier (not)? Trends in immunology, 2007. 28(1): p. 5-11.

11. Engelhardt, B. and C. Coisne, Fluids and barriers of the CNS establish immune privilege by confining immune surveillance to a two-walled castle moat surrounding the CNS castle. Fluids and Barriers of the CNS, 2011. 8(1): p. 4.

12. Nicholas, D.S. and R.O. Weller, The fine anatomy of the human spinal meninges. Journal of Neurosurgery, 1988. 69(2): p. 276-282.

13. Dommisse, G.F., The Blood Supply of the Spinal Cord: A Critical Vascular Zone in Spinal Surgery. J Bone Joint Surg Br, 1974. 56-B(2): p. 225-235.

14. Mautes, A.E., et al., Vascular Events After Spinal Cord Injury: Contribution to Secondary Pathogenesis. Physical Therapy, 2000. 80(7): p. 673-687.

211

15. Martirosyan, N.L., et al., Blood supply and vascular reactivity of the spinal cord under normal and pathological conditions. Journal of Neurosurgery: Spine, 2011. 15(3): p. 238-251.

16. Shamji, M.F., et al., Circulation of the spinal cord: an important consideration for thoracic surgeons. Ann Thorac Surg, 2003. 76(1): p. 315-321.

17. Goshgarian, H.G., Blood Supply of the Spinal Cord, in Spinal Cord Medicine: Principles and Practice, V.W. Lin, et al., Editors. 2003, Demos Medical Publishing: New York.

18. Jokich, P.M., J.M. Rubin, and G.J. Dohrmann, Intraoperative ultrasonic evaluation of spinal cord motion. Journal of Neurosurgery, 1984. 60(4): p. 707-711.

19. Mikulis, D.J., et al., Oscillatory motion of the normal cervical spinal cord. Radiology, 1994. 192(1): p. 117-121.

20. Feinberg, D.A. and A.S. Mark, Human brain motion and cerebrospinal fluid circulation demonstrated with MR velocity imaging. Radiology, 1987. 163(3): p. 793-799.

21. Enzmann, D.R. and N.J. Pelc, Normal flow patterns of intracranial and spinal cerebrospinal fluid defined with phase-contrast cine MR imaging. Radiology, 1991. 178(2): p. 467-474.

22. Henry-Feugeas, M.C., et al., Origin of subarachnoid cerebrospinal fluid pulsations: a phase-contrast MR analysis. Magn Reson Imaging, 2000. 18(4): p. 387-95.

23. Nitz, W.R., et al., Flow dynamics of cerebrospinal fluid: assessment with phase-contrast velocity MR imaging performed with retrospective cardiac gating. Radiology, 1992. 183(2): p. 395-405.

24. Segal, H.D.a.M.B., Physiology of the CSF and blood brain barriers1996, Boca Ranton: CRC Press.

25. Brodbelt, A.R., et al., Altered subarachnoid space compliance and fluid flow in an animal model of posttraumatic syringomyelia. Spine, 2003. 28(20): p. E413-9.

26. Milhorat, T.H., R.M. Kotzen, and A.P. Anzil, Stenosis of central canal of spinal cord in man: incidence and pathological findings in 232 autopsy cases. J Neurosurg, 1994. 80(4): p. 716-22.

27. Brown, P.D., et al., Molecular mechanisms of cerebrospinal fluid production. Neuroscience, 2004. 129(4): p. 957-70.

28. Jungreis, C.A., et al., Normal perivascular spaces mimicking lacunar infarction: MR imaging. Radiology, 1988. 169(1): p. 101-104.

29. Esiri, M.M. and D. Gay, Immunological and neuropathological significance of the Virchow-Robin space. Journal of the Neurological Sciences, 1990. 100(1–2): p. 3-8.

212

30. Agnati, L.F., et al., Energy gradients for the homeostatic control of brain ECF composition and for VT signal migration: introduction of the tide hypothesis. Journal of Neural Transmission, 2005. 112(1): p. 45-63.

31. Ballabh, P., A. Braun, and M. Nedergaard, The blood-brain barrier: an overview: Structure, regulation, and clinical implications. Neurobiology of Disease, 2004. 16(1): p. 1-13.

32. Lok, J., et al., Cell–cell Signaling in the Neurovascular Unit. Neurochemical Research, 2007. 32(12): p. 2032-2045.

33. Zlokovic, B.V., The Blood-Brain Barrier in Health and Chronic Neurodegenerative Disorders. Neuron, 2008. 57(2): p. 178-201.

34. Cecchelli, R., et al., Modelling of the blood-brain barrier in drug discovery and development. Nature Reviews Drug Discovery, 2007. 6(8): p. 650-661.

35. Francis, K., et al., Innate immunity and brain inflammation: the key role of complement. Expert Reviews in Molecular Medicine, 2003. 5(15): p. 1-19.

36. Förster, C., Tight junctions and the modulation of barrier function in disease. Histochemistry and Cell Biology, 2008. 130(1): p. 55-70.

37. Niessen, C.M., Tight Junctions/Adherens Junctions: Basic Structure and Function. Journal of Investigative Dermatology, 2007. 127(11): p. 2525-2532.

38. Stevenson, B.R., et al., Identification of ZO-1: a high molecular weight polypeptide associated with the tight junction (zonula occludens) in a variety of epithelia. The Journal of Cell Biology, 1986. 103(3): p. 755-766.

39. Schneeberger, E.E. and R.D. Lynch, The tight junction: a multifunctional complex. American Journal of Physiology - Cell Physiology, 2004. 286(6): p. C1213-C1228.

40. Abbott, N.J., L. Ronnback, and E. Hansson, Astrocyte-endothelial interactions at the blood-brain barrier. Nature Reviews Neuroscience, 2006. 7(1): p. 42-53.

41. Koehler, R.C., R.J. Roman, and D.R. Harder, Astrocytes and the regulation of cerebral blood flow. Trends in Neurosciences, 2009. 32(3): p. 160-169.

42. Salmina, A.B., Neuron-Glia Interactions as Therapeutic Targets in Neurodegeneration. Journal of Alzheimer's Disease, 2009. 16(3): p. 485-502.

43. Timpl, R., Structure and biological activity of basement membrane proteins. European Journal of Biochemistry, 1989. 180(3): p. 487-502.

44. Persidsky, Y., et al., Blood–brain Barrier: Structural Components and Function Under Physiologic and Pathologic Conditions. Journal of Neuroimmune Pharmacology, 2006. 1(3): p. 223-236.

213

45. Sá-Pereira, I., D. Brites, and M. Brito, Neurovascular Unit: a Focus on Pericytes. Molecular Neurobiology, 2012. 45(2): p. 327-347.

46. Balabanov, R., et al., CNS Microvascular Pericytes Express Macrophage-like Function, Cell Surface Integrin αM, and Macrophage Marker ED-2. Microvascular Research, 1996. 52(2): p. 127-142.

47. Abbott, N.J., et al., Structure and function of the blood–brain barrier. Neurobiology of Disease, 2010. 37(1): p. 13-25.

48. Bartanusz, V., et al., The blood–spinal cord barrier: Morphology and Clinical Implications. Annals of Neurology, 2011. 70(2): p. 194-206.

49. Burney, R.E., et al., Incidence, characteristics, and outcome of spinal cord injury at trauma centers in North America. Arch Surg, 1993. 128(5): p. 596-9.

50. Spinal Cord Injury Facts and Statistics, 2006, Rick Hansen Spinal Cord Injury Registry.

51. Spinal Cord Injury Facts and Figures at a Glance, 2010, National Spinal Cord Injury Statistical Center: Birmingham, Alabama.

52. One Degree of Separation: Paralysis and Spinal Cord Injury in the United States, 2009, Christopher and Dana Reeve Foundation.

53. Xie, D., et al., An engineered vascular endothelial growth factor-activating transcription factor induces therapeutic angiogenesis in ApoE knockout mice with hindlimb ischemia. J Vasc Surg, 2006. 44(1): p. 166-75.

54. McDonald, J.W. and C. Sadowsky, Spinal-cord injury. Lancet, 2002. 359(9304): p. 417-425.

55. Priebe, M.M., et al., Spinal Cord Injury Medicine. 6. Economic and Societal Issues in Spinal Cord Injury. Archives of Physical Medicine and Rehabilitation, 2007. 88(3): p. S84-S88.

56. Beattie, M.S., Inflammation and apoptosis: linked therapeutic targets in spinal cord injury. Trends Mol Med, 2004. 10(12): p. 580-3.

57. Leypold, B.G., et al., The impact of methylprednisolone on lesion severity following spinal cord injury. Spine, 2007. 32(3): p. 373-8; discussion 379-81.

58. Benton, R.L., et al., Griffonia simplicifolia isolectin B4 identifies a specific subpopulation of angiogenic blood vessels following contusive spinal cord injury in the adult mouse. J Comp Neurol, 2008. 507(1): p. 1031-52.

59. Waxman, S.G., Demyelination of spinal cord injury. Journal of Neuroscience, 1989. 91: p. 1-14.

214

60. Blight, A.R., Delayed demyelination and macrophage invasion: a candidate for secondary cell damage in spinal cord injury. Cent Nerv Syst Trauma, 1985. 2(4): p. 299-315.

61. Nashmi, R. and M.G. Fehlings, Changes in axonal physiology and morphology after chronic compressive injury of the rat thoracic spinal cord. Neuroscience, 2001. 104(1): p. 235-51.

62. Blight, A.R. and V. Decrescito, Morphometric analysis of experimental spinal cord injury in the cat: the relation of injury intensity to survival of myelinated axons. Neuroscience, 1986. 19(1): p. 321-41.

63. Faulkner, J.R., et al., Reactive astrocytes protect tissue and preserve function after spinal cord injury. J Neurosci, 2004. 24(9): p. 2143-55.

64. Preston, E., J. Webster, and D. Small, Characteristics of sustained blood-brain barrier opening and tissue injury in a model for focal trauma in the rat. J Neurotrauma, 2001. 18(1): p. 83-92.

65. Ehlers, M.D., Deconstructing the axon: Wallerian degeneration and the ubiquitin-proteasome system. Trends Neurosci, 2004. 27(1): p. 3-6.

66. Popovich, P.G., et al., A quantitative spatial analysis of the blood-spinal cord barrier. I. Permeability changes after experimental spinal contusion injury. Exp Neurol, 1996. 142(2): p. 258-75.

67. Akhtar, A.Z., J.J. Pippin, and C.B. Sandusky, Animal models in spinal cord injury: a review. Reviews in the neurosciences, 2008. 19(1): p. 47-60.

68. Nystrom, B. and J.E. Berglund, Spinal cord restitution following compression injuries in rats. Acta Neurol Scand, 1988. 78(6): p. 467-72.

69. Tator, C.H., et al., Current use and timing of spinal surgery for management of acute spinal cord injury in North America: results of a retrospective multicenter study. Neurosurgical focus, 1999. 6(1): p. Article 2.

70. LaPlaca, M.C., et al., CNS injury biomechanics and experimental models. Prog Brain Res, 2007. 161: p. 13-26.

71. Maikos, J.T. and D.I. Shreiber, Immediate damage to the blood-spinal cord barrier due to mechanical trauma. J Neurotrauma, 2007. 24(3): p. 492-507.

72. Totoiu, M.O. and H.S. Keirstead, Spinal cord injury is accompanied by chronic progressive demyelination. J Comp Neurol, 2005. 486(4): p. 373-83.

73. Casha, S., W.R. Yu, and M.G. Fehlings, Oligodendroglial apoptosis occurs along degenerating axons and is associated with FAS and p75 expression following spinal cord injury in the rat. Neuroscience, 2001. 103(1): p. 203-18.

215

74. Fehlings, M.G. and C.H. Tator, The relationships among the severity of spinal cord injury, residual neurological function, axon counts, and counts of retrogradely labeled neurons after experimental spinal cord injury. Experimental Neurology, 1995. 132(2): p. 220-228.

75. Karimi-Abdolrezaee, S., E. Eftekharpour, and M.G. Fehlings, Temporal and spatial patterns of Kv1.1 and Kv1.2 protein and gene expression in spinal cord white matter after acute and chronic spinal cord injury in rats: implications for axonal pathophysiology after neurotrauma. Eur J Neurosci, 2004. 19(3): p. 577-89.

76. Karimi-Abdolrezaee, S., et al., Delayed Transplantation of Adult Neural Precursor Cells Promotes Remyelination and Functional Neurological Recovery after Spinal Cord Injury. The Journal of Neuroscience, 2006. 26(13): p. 3377-3389.

77. Ditunno, J.F., Jr., et al., The international standards booklet for neurological and functional classification of spinal cord injury. American Spinal Injury Association. Paraplegia, 1994. 32(2): p. 70-80.

78. El Masry, W.S., et al., Validation of the American Spinal Injury Association (ASIA) Motor Score and the National Acute Spinal Cord Injury Study (NASCIS) Motor Score. Spine, 1996. 21(5): p. 614-619.

79. Waxman, S.G., et al., Sodium channels, excitability of primary sensory neurons, and the molecular basis of pain. Muscle & Nerve, 1999. 22(9): p. 1177-1187.

80. Weaver, L.C. and C. Polosa, Progress in Brain Research: Autonomic Dysfunction After Spinal Cord Injury. Vol. Volume 152. 2005: Elsevier. 472.

81. Schramm, L.P., Spinal sympathetic interneurons: their identification and roles after spinal cord injury. Prog Brain Res, 2006. 152: p. 27-37.

82. Weaver, L.C., et al., Autonomic dysreflexia after spinal cord injury: central mechanisms and strategies for prevention, in Progress in Brain Research, C.W. Lynne and P. Canio, Editors. 2006, Elsevier. p. 245-263.

83. Tator, C.H. and I. Koyanagi, Vascular mechanisms in the pathophysiology of human spinal cord injury. Journal of Neurosurgery, 1997. 86(3): p. 483-492.

84. Dumont, R.J., et al., Acute Spinal Cord Injury, Part I: Pathophysiologic Mechanisms. Clinical Neuropharmacology, 2001. 24(5): p. 254-264.

85. Benzel, E.C., The Cervical Spine. 5th ed2012. 1666.

86. Rowland, J.W., et al., Current status of acute spinal cord injury pathophysiology and emerging therapies: promise on the horizon. Neurosurg Focus, 2008. 25(5): p. E2.

87. Ditunno, J.F., et al., Spinal shock revisited: a four-phase model. Spinal Cord, 2004. 42(7): p. 383-95.

216

88. Aoyama, T., et al., Ultra-early MRI showing no abnormality in a fall victim presenting with tetraparesis. Spinal Cord, 2007. 45(10): p. 695-9.

89. Loy, D.N., et al., Temporal progression of angiogenesis and basal lamina deposition after contusive spinal cord injury in the adult rat. The Journal of Comparative Neurology, 2002. 445(4): p. 308-324.

90. Brockstein, B., L. Johns, and B. Gewertz, Blood supply to the spinal cord: Anatomic and physiologic correlations. Annals of Vascular Surgery, 1994. 8(4): p. 394-399.

91. Whetstone, W.D., et al., Blood-Spinal Cord Barrier After Spinal Cord Injury: Relation to Revascularization and Wound Healing. Journal of Neuroscience Research, 2003. 74(2): p. 227–239.

92. Anthes, D.L., E. Theriault, and C.H. Tator, Ultrastructural evidence for arteriolar vasospasm after spinal cord trauma. Neurosurgery, 1996. 39(4): p. 804-14.

93. Sharma, H.S., Pathophysiology of the blood-spinal cord barrier in traumatic injury, in The blood-spinal cord and brain barriers in health and disease, H.S. Sharma and J. Westman, Editors. 2004, Elsevier Academic Press: San Diego. p. 437-518.

94. Guha, A., C.H. Tator, and J. Rochon, Spinal cord blood flow and systemic blood pressure after experimental spinal cord injury in rats. Stroke, 1989. 20(3): p. 372-7.

95. Guha, A. and C.H. Tator, Acute cardiovascular effects of experimental spinal cord injury. J Trauma, 1988. 28(4): p. 481-90.

96. Smith, Q.R., A Review of Blood-Brain Barrier Transport Techniques, in The Blood-Brain Barrier: Biology and Research Protocols2003. p. 193-208.

97. Noble, L.J. and J.R. Wrathall, Distribution and time course of protein extravasation in the rat spinal cord after contusive injury. Brain Research, 1989. 482(1): p. 57-66.

98. Weis, S.M. and D.A. Cheresh, Pathophysiological consequences of VEGF-induced vascular permeability. Nature, 2005. 437(7058): p. 497-504.

99. Galis, Z.S. and J.J. Khatri, Matrix Metalloproteinases in Vascular Remodeling and Atherogenesis: The Good, the Bad, and the Ugly. Circ Res, 2002. 90(3): p. 251-262.

100. Stys, P.K. and R.M. Lopachin, Mechanisms of calcium and sodium fluxes in anoxic myelinated central nervous system axons. Neuroscience, 1998. 82(1): p. 21-32.

101. Agrawal, S.K. and M.G. Fehlings, Mechanisms of secondary injury to spinal cord axons in vitro: role of Na+, Na(+)-K(+)-ATPase, the Na(+)-H+ exchanger, and the Na(+)-Ca2+ exchanger. J Neurosci, 1996. 16(2): p. 545-52.

102. Park, E., A.A. Velumian, and M.G. Fehlings, The role of excitotoxicity in secondary mechanisms of spinal cord injury: a review with an emphasis on the implications for white matter degeneration. J Neurotrauma, 2004. 21(6): p. 754-74.

217

103. Agrawal, S.K., R. Nashmi, and M.G. Fehlings, Role of L- and N-type calcium channels in the pathophysiology of traumatic spinal cord white matter injury. Neuroscience, 2000. 99(1): p. 179-88.

104. McAdoo, D.J., et al., Changes in amino acid concentrations over time and space around an impact injury and their diffusion through the rat spinal cord. Exp Neurol, 1999. 159(2): p. 538-44.

105. Liu, D., et al., Neurotoxicity of glutamate at the concentration released upon spinal cord injury. Neuroscience, 1999. 93(4): p. 1383-9.

106. Anderson, D.K., et al., Spinal cord energy metabolism following compression trauma to the feline spinal cord. J Neurosurg, 1980. 53(3): p. 375-80.

107. Banik, N.L., et al., Increased calpain content and progressive degradation of neurofilament protein in spinal cord injury. Brain Res, 1997. 752(1-2): p. 301-6.

108. Schumacher, P.A., R.G. Siman, and M.G. Fehlings, Pretreatment with calpain inhibitor CEP-4143 inhibits calpain I activation and cytoskeletal degradation, improves neurological function, and enhances axonal survival after traumatic spinal cord injury. J Neurochem, 2000. 74(4): p. 1646-55.

109. Luo, J., et al., Detection of reactive oxygen species by flow cytometry after spinal cord injury. J Neurosci Methods, 2002. 120(1): p. 105-12.

110. Xiong, Y., A.G. Rabchevsky, and E.D. Hall, Role of peroxynitrite in secondary oxidative damage after spinal cord injury. J Neurochem, 2007. 100(3): p. 639-49.

111. Donnelly, D.J. and P.G. Popovich, Inflammation and its role in neuroprotection, axonal regeneration and functional recovery after spinal cord injury. Exp Neurol, 2008. 209(2): p. 378-88.

112. Bao, F. and D. Liu, Peroxynitrite generated in the rat spinal cord induces apoptotic cell death and activates caspase-3. Neuroscience, 2003. 116(1): p. 59-70.

113. Carlson, S.L., et al., Acute inflammatory response in spinal cord following impact injury. Exp Neurol, 1998. 151(1): p. 77-88.

114. Popovich, P.G., P. Wei, and B.T. Stokes, Cellular inflammatory response after spinal cord injury in Sprague-Dawley and Lewis rats. J Comp Neurol, 1997. 377(3): p. 443-64.

115. Hausmann, O.N., Post-traumatic inflammation following spinal cord injury. Spinal Cord, 2003. 41(7): p. 369-78.

116. Watanabe, T., et al., Differential activation of microglia after experimental spinal cord injury. J Neurotrauma, 1999. 16(3): p. 255-65.

117. Fleming, J.C., et al., The cellular inflammatory response in human spinal cords after injury. Brain, 2006. 129(Pt 12): p. 3249-69.

218

118. Schnell, L., et al., Acute inflammatory responses to mechanical lesions in the CNS: differences between brain and spinal cord. European Journal of Neuroscience, 1999. 11(10): p. 3648-3658.

119. Taoka, Y., et al., Role of neutrophils in spinal cord injury in the rat. Neuroscience, 1997. 79(4): p. 1177-82.

120. Saville, L.R., et al., A monoclonal antibody to CD11d reduces the inflammatory infiltrate into the injured spinal cord: a potential neuroprotective treatment. J Neuroimmunol, 2004. 156(1-2): p. 42-57.

121. Stirling, D.P., et al., Depletion of Ly6G/Gr-1 leukocytes after spinal cord injury in mice alters wound healing and worsens neurological outcome. J Neurosci, 2009. 29(3): p. 753-64.

122. Bonfoco, E., et al., Apoptosis and necrosis: two distinct events induced, respectively, by mild and intense insults with N-methyl-D-aspartate or nitric oxide/superoxide in cortical cell cultures. Proc Natl Acad Sci U S A, 1995. 92(16): p. 7162-6.

123. Majno, G. and I. Joris, Apoptosis, oncosis, and necrosis. An overview of cell death. Am J Pathol, 1995. 146(1): p. 3-15.

124. Keane, R.W., et al., Apoptotic and anti-apoptotic mechanisms following spinal cord injury. J Neuropathol Exp Neurol, 2001. 60(5): p. 422-9.

125. Lou, J., et al., Apoptosis as a mechanism of neuronal cell death following acute experimental spinal cord injury. Spinal Cord, 1998. 36(10): p. 683-90.

126. Weerasinghe, P. and L.M. Buja, Oncosis: An important non-apoptotic mode of cell death. Experimental and Molecular Pathology, 2012. 93(3): p. 302-308.

127. Crowe, M.J., et al., Apoptosis and delayed degeneration after spinal cord injury in rats and monkeys. Nat Med, 1997. 3(1): p. 73-6.

128. Ellis, R.E., J.Y. Yuan, and H.R. Horvitz, Mechanisms and functions of cell death. Annu Rev Cell Biol, 1991. 7: p. 663-98.

129. Ling, X. and D. Liu, Temporal and spatial profiles of cell loss after spinal cord injury: Reduction by a metalloporphyrin. Journal of Neuroscience Research, 2007. 85(10): p. 2175-2185.

130. Liu, X.Z., et al., Neuronal and Glial Apoptosis after Traumatic Spinal Cord Injury. The Journal of Neuroscience, 1997. 17(14): p. 5395-5406.

131. Levine, J.M., R. Reynolds, and J.W. Fawcett, The oligodendrocyte precursor cell in health and disease. Trends in Neurosciences, 2001. 24(1): p. 39-47.

219

132. Karimi-Abdolrezaee, S., et al., Delayed transplantation of adult neural precursor cells promotes remyelination and functional neurological recovery after spinal cord injury. J Neurosci, 2006. 26(13): p. 3377-89.

133. Keirstead, H.S., Human embryonic stem cell-derived oligodendrocyte progenitor cell transplants remyelinate and restore locomotion after spinal cord injury. J. Neurosci., 2005. 25: p. 4694-4705.

134. Xiao, M., et al., Human adult olfactory neural progenitors rescue axotomized rodent rubrospinal neurons and promote functional recovery. Exp Neurol, 2005. 194(1): p. 12-30.

135. Blight, A.R., Macrophages and inflammatory damage in spinal cord injury. J Neurotrauma, 1992. 9 Suppl 1: p. S83-91.

136. Kigerl, K.A., et al., Identification of two distinct macrophage subsets with divergent effects causing either neurotoxicity or regeneration in the injured mouse spinal cord. J Neurosci, 2009. 29(43): p. 13435-44.

137. Mabon, P.J., L.C. Weaver, and G.A. Dekaban, Inhibition of monocyte/macrophage migration to a spinal cord injury site by an antibody to the integrin alphaD: a potential new anti-inflammatory treatment. Exp Neurol, 2000. 166(1): p. 52-64.

138. Jones, T.B., E.E. McDaniel, and P.G. Popovich, Inflammatory-mediated injury and repair in the traumatically injured spinal cord. Curr Pharm Des, 2005. 11(10): p. 1223-36.

139. Popovich, P.G., The neuropathological and behavioral consequences of intraspinal microglial/macrophage activation. J. Neuropathol. Exp. Neurol., 2002. 61: p. 623-633.

140. Popovich, P.G., Immunological regulation of neuronal degeneration and regeneration in the injured spinal cord. Prog Brain Res, 2000. 128: p. 43-58.

141. Moalem, G., et al., Production of neurotrophins by activated T cells: implications for neuroprotective autoimmunity. J Autoimmun, 2000. 15(3): p. 331-45.

142. Schwartz, M., et al., Innate and adaptive immune responses can be beneficial for CNS repair. Trends Neurosci, 1999. 22(7): p. 295-9.

143. Liu, X.Z., et al., Neuronal and glial apoptosis after traumatic spinal cord injury. J Neurosci, 1997. 17(14): p. 5395-406.

144. Springer, J.E., R.D. Azbill, and P.E. Knapp, Activation of the caspase-3 apoptotic cascade in traumatic spinal cord injury. Nat Med, 1999. 5(8): p. 943-6.

145. McEwen, M.L. and J.E. Springer, A mapping study of caspase-3 activation following acute spinal cord contusion in rats. J Histochem Cytochem, 2005. 53(7): p. 809-19.

220

146. Brazda, N. and H.W. Muller, Pharmacological modification of the extracellular matrix to promote regeneration of the injured brain and spinal cord. Prog Brain Res, 2009. 175: p. 269-81.

147. Kakulas, B.A., Neuropathology: the foundation for new treatments in spinal cord injury. Spinal Cord, 2004. 42(10): p. 549-63.

148. Jones, L.L., R.U. Margolis, and M.H. Tuszynski, The chondroitin sulfate proteoglycans neurocan, brevican, phosphacan, and versican are differentially regulated following spinal cord injury. Exp Neurol, 2003. 182(2): p. 399-411.

149. Hagg, T. and M. Oudega, Degenerative and spontaneous regenerative processes after spinal cord injury. J Neurotrauma, 2006. 23(3-4): p. 264-80.

150. Bradbury, E.J., Chondroitinase ABC promotes functional recovery after spinal cord injury. Nature, 2002. 416: p. 636-640.

151. Loy, D.N., et al., Temporal progression of angiogenesis and basal lamina deposition after contusive spinal cord injury in the adult rat. J Comp Neurol, 2002. 445(4): p. 308-24.

152. Buss, A., et al., Growth-modulating molecules are associated with invading Schwann cells and not astrocytes in human traumatic spinal cord injury. Brain, 2007. 130(Pt 4): p. 940-53.

153. Jones, L.L., et al., NG2 is a major chondroitin sulfate proteoglycan produced after spinal cord injury and is expressed by macrophages and oligodendrocyte progenitors. J Neurosci, 2002. 22(7): p. 2792-803.

154. Sandvig, A., et al., Myelin-, reactive glia-, and scar-derived CNS axon growth inhibitors: expression, receptor signaling, and correlation with axon regeneration. Glia, 2004. 46(3): p. 225-51.

155. Shen, Y., et al., PTPsigma is a receptor for chondroitin sulfate proteoglycan, an inhibitor of neural regeneration. Science, 2009. 326(5952): p. 592-6.

156. Weiss, S., et al., Multipotent CNS stem cells are present in the adult mammalian spinal cord and ventricular neuroaxis. J Neurosci, 1996. 16(23): p. 7599-609.

157. Yamamoto, S., et al., Proliferation of parenchymal neural progenitors in response to injury in the adult rat spinal cord. Exp Neurol, 2001. 172(1): p. 115-27.

158. McTigue, D.M., P. Wei, and B.T. Stokes, Proliferation of NG2-positive cells and altered oligodendrocyte numbers in the contused rat spinal cord. J Neurosci, 2001. 21(10): p. 3392-400.

159. Sellers, D.L., D.O. Maris, and P.J. Horner, Postinjury niches induce temporal shifts in progenitor fates to direct lesion repair after spinal cord injury. J Neurosci, 2009. 29(20): p. 6722-33.

221

160. Watt, S.M., et al., Human endothelial stem/progenitor cells, angiogenic factors and vascular repair. Journal of The Royal Society Interface, 2010. 7(Suppl 6): p. S731-S751.

161. Hill, C.E., M.S. Beattie, and J.C. Bresnahan, Degeneration and sprouting of identified descending supraspinal axons after contusive spinal cord injury in the rat. Exp Neurol, 2001. 171(1): p. 153-69.

162. Horner, P.J., Proliferation and differentiation of progenitor cells throughout the intact adult rat spinal cord. J. Neurosci., 2000. 20: p. 2218-2228.

163. Kakulas, B.A., A review of the neuropathology of human spinal cord injury with emphasis on special features. J. Spinal Cord Med., 1999. 22: p. 119-124.

164. Brodbelt, A.R. and M.A. Stoodley, Post-traumatic syringomyelia: a review. Journal of Clinical Neuroscience, 2003. 10(4): p. 401-408.

165. Williams, B., Pathogenesis of post-traumatic syringomyelia. Br J Neurosurg, 1992. 6(6): p. 517-20.

166. Perrouin-Verbe, B., et al., Post-traumatic syringomyelia and post-traumatic spinal canal stenosis: a direct relationship: review of 75 patients with a spinal cord injury. Spinal Cord, 1998. 36(2): p. 137-43.

167. Schwartz, E.D., et al., Posttraumatic syringomyelia: pathogenesis, imaging, and treatment. AJR Am J Roentgenol, 1999. 173(2): p. 487-92.

168. Stoodley, M.A., Pathophysiology of syringomyelia. J Neurosurg, 2000. 92(6): p. 1069-70; author reply 1071-3.

169. Klekamp, J., et al., Disturbances of cerebrospinal fluid flow attributable to arachnoid scarring cause interstitial edema of the cat spinal cord. Neurosurgery, 2001. 48(1): p. 174-85; discussion 185-6.

170. Klekamp, J., et al., Treatment of syringomyelia associated with arachnoid scarring caused by arachnoiditis or trauma. J Neurosurg, 1997. 86(2): p. 233-40.

171. Hulsebosch, C.E., et al., Mechanisms of chronic central neuropathic pain after spinal cord injury. Brain Res Rev, 2009. 60(1): p. 202-13.

172. Gwak, Y.S. and C.E. Hulsebosch, Remote astrocytic and microglial activation modulates neuronal hyperexcitability and below-level neuropathic pain after spinal injury in rat. Neuroscience, 2009. 161(3): p. 895-903.

173. Bruce, J.C., M.A. Oatway, and L.C. Weaver, Chronic pain after clip-compression injury of the rat spinal cord. Exp Neurol, 2002. 178(1): p. 33-48.

174. Werhagen, L., et al., Neuropathic pain after traumatic spinal cord injury--relations to gender, spinal level, completeness, and age at the time of injury. Spinal Cord, 2004. 42(12): p. 665-73.

222

175. Bao, F., et al., An integrin inhibiting molecule decreases oxidative damage and improves neurological function after spinal cord injury. Exp Neurol, 2008. 214(2): p. 160-7.

176. Gwak, Y.S., et al., Propentofylline attenuates allodynia, glial activation and modulates GABAergic tone after spinal cord injury in the rat. Pain, 2008. 138(2): p. 410-22.

177. McKinley, W.O., R.T. Seel, and J.T. Hardman, Nontraumatic spinal cord injury: Incidence, epidemiology, and functional outcome. Archives of Physical Medicine and Rehabilitation, 1999. 80(6): p. 619-623.

178. New, P.W. and V. Sundararajan, Incidence of non-traumatic spinal cord injury in Victoria, Australia: a population-based study and literature review. Spinal Cord, 2007. 46(6): p. 406-411.

179. Ones, K., et al., Comparison of functional results in non-traumatic and traumatic spinal cord injury. Disability and Rehabilitation, 2007. 29(15): p. 1185-1191.

180. BOHLMAN, H.H. and S.E. EMERY, The Pathophysiology of Cervical Spondylosis and Myelopathy. Spine, 1988. 13(7): p. 843-846.

181. Baptiste, D.C. and M.G. Fehlings, Pharmacological approaches to repair the injured spinal cord. J. Neurotrauma, 2006. 23: p. 318-334.

182. Richardson, P.M., U.M. McGuinness, and A.J. Aguayo, Axons from CNS grafts regenerate into PNS grafts. Nature, 1980. 284: p. 264-265.

183. Lee, Y.S., et al., Motor recovery and anatomical evidence of axonal regrowth in spinal cord-repaired adult rats. J. Neuropathol. Exp. Neurol., 2004. 63: p. 233-245.

184. Takami, T., Schwann cell but not olfactory ensheathing glia transplants improve hindlimb locomotor performance in the moderately contused adult rat thoracic spinal cord. J. Neurosci., 2002. 22: p. 6670-6681.

185. Cummings, B.J., Human neural stem cells differentiate and promote locomotor recovery in spinal cord-injured mice. Proc. Natl Acad. Sci. USA, 2005. 102: p. 14069-14074.

186. Karimi-Abdolrezaee, S., et al., Delayed transplantation of adult neural precursor cells promotes remyelination and functional neurological recovery after spinal cord injury. J. Neurosci., 2006. 26: p. 3377-3389.

187. Keirstead, H.S., et al., Human Embryonic Stem Cell-Derived Oligodendrocyte Progenitor Cell Transplants Remyelinate and Restore Locomotion after Spinal Cord Injury. J. Neurosci., 2005. 25(19): p. 4694-4705.

188. Pollack, A., Geron Is Shutting Down Its Stem Cell Clinical Trial, in The New York Times2011: New York.

189. StemCells Inc.: Clinical Trials. 2012 November 10, 2012; Available from: http://www.stemcellsinc.com/Therapeutic-Programs/Clinical-Trials.htm.

223

190. StemCells, I. StemCells, Inc. Completes Enrollment of First Cohort in Landmark Chronic Spinal Cord Injury Trial. 2011.

191. ReNeuron Update on stroke clinical trial. 2011.

192. Health, U.S.N.I.o., TotipotentRX Cell Therapy Pvt. Ltd.: NCT01490242, C. Trials, Editor 2013, U.S. National Institutes of Health and National Library of Medicine.

193. Thuret, S., L.D.F. Moon, and F.H. Gage, Therapeutic interventions after spinal cord injury. Nature Reviews Neuroscience, 2006. 7(8): p. 628-643.

194. Stirling, D.P., Minocycline treatment reduces delayed oligodendrocyte death, attenuates axonal dieback, and improves functional outcome after spinal cord injury. J. Neurosci., 2004. 24: p. 2182-2190.

195. Nógrádi, A., et al., Delayed riluzole treatment is able to rescue injured rat spinal motoneurons. Neuroscience, 2007. 144(2): p. 431-438.

196. Stutzmann, J.M., et al., The effect of riluzole on post-traumatic spinal cord injury in the rat. Neuroreport, 1996. 7(2): p. 387-392.

197. Schwartz, G. and M.G. Fehlings, Evaluation of the neuroprotective effects of sodium channel blockers after spinal cord injury: improved behavioral and neuroanatomical recovery with riluzole. Journal of Neurosurgery: Spine, 2001. 94(2): p. 245-256.

198. Kaptanoglu, E., et al., Erythropoietin exerts neuroprotection after acute spinal cord injury in rats: effect on lipid peroxidation and early ultrastructural findings. Neurosurgical Review, 2004. 27(2): p. 113-120.

199. Rabchevsky, A.G., et al., Cyclosporin A Treatment Following Spinal Cord Injury to the Rat: Behavioral Effects and Stereological Assessment of Tissue Sparing. Journal of Neurotrauma, 2001. 18(5): p. 513-522.

200. Koda, M., et al., Granulocyte colony-stimulating factor (G-CSF) mobilizes bone marrow-derived cells into injured spinal cord and promotes functional recovery after compression-induced spinal cord injury in mice. Brain Research, 2007. 1149(0): p. 223-231.

201. Park, H.C., et al., Treatment of complete spinal cord injury patients by autologous bone marrow cell transplantation and administration of granulocyte-macrophage colony stimulating factor. Tissue Eng, 2005. 11(5-6): p. 913-22.

202. Borgens, R.B., R. Shi, and D. Bohnert, Behavioral recovery from spinal cord injury following delayed application of polyethylene glycol. Journal of Experimental Biology, 2002. 205(1): p. 1-12.

203. Luo, J., R. Borgens, and R. Shi, Polyethylene glycol immediately repairs neuronal membranes and inhibits free radical production after acute spinal cord injury. Journal of Neurochemistry, 2002. 83(2): p. 471-480.

224

204. Baptiste, D.C., et al., Systemic Polyethylene Glycol Promotes Neurological Recovery and Tissue Sparing in Rats After Cervical Spinal Cord Injury. Journal of Neuropathology & Experimental Neurology, 2009. 68(6): p. 661-676 10.1097/NEN.0b013e3181a72605.

205. Kwon, B.K., et al., Hypothermia for spinal cord injury. The spine journal : official journal of the North American Spine Society, 2008. 8(6): p. 859-874.

206. Dietrich, W.D., 3rd, Therapeutic hypothermia for spinal cord injury. Crit Care Med, 2009. 37(7 Suppl).

207. Dietrich, W.D., et al., Hypothermic Treatment for Acute Spinal Cord Injury. Neurotherapeutics, 2011. 8(2): p. 229-239.

208. Levi, A.D., et al., Clinical application of modest hypothermia after spinal cord injury. J Neurotrauma, 2009. 26(3): p. 407-15.

209. Lo, T.P., et al., Systemic hypothermia improves histological and functional outcome after cervical spinal cord contusion in rats. The Journal of Comparative Neurology, 2009. 514(5): p. 433-448.

210. Yu, C.G., et al., Beneficial effects of modest systemic hypothermia on locomotor function and histopathological damage following contusion-induced spinal cord injury in rats. Journal of Neurosurgery: Spine, 2000. 93(1): p. 85-93.

211. Rosenstein, J.M. and J.M. Krum, New roles for VEGF in nervous tissue--beyond blood vessels. Experimental Neurology, 2004. 187(2): p. 246-253.

212. Liu, Y., et al., An engineered transcription factor which activates VEGF-A enhances recovery after spinal cord injury. Neurobiology of Disease, 2010. 37(2): p. 384-393.

213. Faulkner, J.R., et al., Reactive Astrocytes Protect Tissue and Preserve Function after Spinal Cord Injury. J. Neurosci., 2004. 24(9): p. 2143-2155.

214. Noble, L.J., et al., Matrix Metalloproteinases Limit Functional Recovery after Spinal Cord Injury by Modulation of Early Vascular Events. J. Neurosci., 2002. 22(17): p. 7526-7535.

215. Engesser-Cesar, C., et al., Voluntary wheel running improves recovery from a moderate spinal cord injury. J. Neurotrauma, 2005. 22: p. 157-171.

216. Bouyer, L.J., Animal models for studying potential training strategies in persons with spinal cord injury. J. Neurol. Phys. Ther., 2005. 29: p. 117-125.

217. Edgerton, V.R., et al., Rehabilitative therapies after spinal cord injury. J. Neurotrauma, 2006. 23: p. 560-570.

218. Raineteau, O. and M.E. Schwab, Plasticity of motor systems after incomplete spinal cord injury. Nature Rev. Neurosci., 2001. 2: p. 263-273.

225

219. Nash, M.S., Exercise as a health-promoting activity following spinal cord injury. J. Neurol. Phys. Ther., 2005. 29: p. 87-103.

220. Otrock, Z.K., et al., Understanding the biology of angiogenesis: Review of the most important molecular mechanisms. Blood Cells, Molecules, and Diseases, 2007. 39(2): p. 212-220.

221. Patan, S., Vasculogenesis and Angiogenesis

Angiogenesis in Brain Tumors, M. Kirsch and P.M. Black, Editors. 2004, Springer US. p. 3-32.

222. Carmeliet, P., Mechanisms of angiogenesis and arteriogenesis. Nature Medicine, 2000. 6(4): p. 389-395.

223. Fong, G.H., et al., Increased hemangioblast commitment, not vascular disorganization, is the primary defect in flt-1 knock-out mice. Development, 1999. 126(13): p. 3015-3025.

224. Ferrara, N., Role of vascular endothelial growth factor in the regulation of angiogenesis. Kidney International, 1999. 56(3): p. 794-814.

225. Schmidt, A., K. Brixius, and W. Bloch, Endothelial Precursor Cell Migration During Vasculogenesis. Circulation Research, 2007. 101(2): p. 125-136.

226. Hendrix, M.J.C., et al., Vasculogenic mimicry and tumour-cell plasticity: lessons from melanoma. Nature Reviews Cancer, 2003. 3(6): p. 411-421.

227. Ribatti, D., et al., Postnatal vasculogenesis. Mechanisms of Development, 2001. 100(2): p. 157-163.

228. Thurston, G., et al., Leakage-Resistant Blood Vessels in Mice Transgenically Overexpressing Angiopoietin-1. Science, 1999. 286(5449): p. 2511-2514.

229. Gale, N.W. and G.D. Yancopoulos, Growth factors acting via endothelial cell-specific receptor tyrosine kinases: VEGFs, Angiopoietins, and ephrins in vascular development. Genes & Development, 1999. 13(9): p. 1055-1066.

230. Maisonpierre, P.C., et al., Angiopoietin-2, a Natural Antagonist for Tie2 That Disrupts in vivo Angiogenesis. Science, 1997. 277(5322): p. 55-60.

231. Stetler-Stevenson, W.G., Matrix metalloproteinases in angiogenesis: a moving target for therapeutic intervention. The Journal of Clinical Investigation, 1999. 103(9): p. 1237-1241.

232. Suri, C., et al., Increased Vascularization in Mice Overexpressing Angiopoietin-1. Science, 1998. 282(5388): p. 468-471.

233. Peters, K.G., et al., Functional Significance of Tie2 Signaling in the Adult Vasculature. Recent Prog Horm Res, 2004. 59(1): p. 51-71.

226

234. Chavakis, E. and S. Dimmeler, Regulation of Endothelial Cell Survival and Apoptosis During Angiogenesis. Arteriosclerosis, Thrombosis, and Vascular Biology, 2002. 22(6): p. 887-893.

235. Byrne, A.M., D.J. Bouchier-Hayes, and J.H. Harmey, Angiogenic and cell survival functions of Vascular Endothelial Growth Factor (VEGF). Journal of Cellular and Molecular Medicine, 2005. 9(4): p. 777-794.

236. Nowak, D.G., et al., Expression of pro- and anti-angiogenic isoforms of VEGF is differentially regulated by splicing and growth factors. Journal of Cell Science, 2008. 121(20): p. 3487-3495.

237. Drake, C.J., D.A. Cheresh, and C.D. Little, An antagonist of integrin alpha v beta 3 prevents maturation of blood vessels during embryonic neovascularization. Journal of Cell Science, 1995. 108(7): p. 2655-2661.

238. Tolsma, S.S., M.S. Stack, and N. Bouck, Lumen Formation and Other Angiogenic Activities of Cultured Capillary Endothelial Cells Are Inhibited by Thrombospondin-1. Microvascular Research, 1997. 54(1): p. 13-26.

239. Ferrara, N., H.-P. Gerber, and J. LeCouter, The biology of VEGF and its receptors. Nature Medicine, 2003. 9(6): p. 669-676.

240. Greenberg, D.A. and K. Jin, From angiogenesis to neuropathology. Nature, 2005. 438(7070): p. 954-9.

241. Shweiki, D., et al., Vascular endothelial growth factor induced by hypoxia may mediate hypoxia-initiated angiogenesis. Nature, 1992. 359(6398): p. 843-5.

242. Zachary, I. and G. Gliki, Signaling transduction mechanisms mediating biological actions of the vascular endothelial growth factor family. Cardiovasc Res, 2001. 49(3): p. 568-81.

243. Harper, S.J. and D.O. Bates, VEGF-A splicing: the key to anti-angiogenic therapeutics? Nature Reviews Cancer, 2008. 8(11): p. 880-887.

244. Leung, D.W., et al., Vascular endothelial growth factor is a secreted angiogenic mitogen. Science, 1989. 246(4935): p. 1306-9.

245. Marti, H.H., Vascular endothelial growth factor. Adv Exp Med Biol, 2002. 513: p. 375-94.

246. Neufeld, G., et al., The Neuropilins: Multifunctional Semaphorin and VEGF Receptors that Modulate Axon Guidance and Angiogenesis. Trends in Cardiovascular Medicine, 2002. 12(1): p. 13-19.

247. Sondell, M., G. Lundborg, and M. Kanje, Vascular endothelial growth factor has neurotrophic activity and stimulates axonal outgrowth, enhancing cell survival and

227

Schwann cell proliferation in the peripheral nervous system. J Neurosci, 1999. 19(14): p. 5731-40.

248. Ortega, N., F.-E. L'Faqihi, and J. Plouet, Control of vascular endothelial growth factor angiogenic activity by the extracellular matrix. Biology of the Cell, 1998. 90: p. 381-390.

249. Sondell, M., F. Sundler, and M. Kanje, Vascular endothelial growth factor is a neurotrophic factor which stimulates axonal outgrowth through the flk-1 receptor. European Journal of Neuroscience, 2000. 12(12): p. 4243-4254.

250. Sondell, M., G. Lundborg, and M. Kanje, Vascular endothelial growth factor stimulates Schwann cell invasion and neovascularization of acellular nerve grafts. Brain Research, 1999. 846(2): p. 219-228.

251. Clauss, M., et al., Vascular permeability factor: a tumor-derived polypeptide that induces endothelial cell and monocyte procoagulant activity, and promotes monocyte migration. The Journal of Experimental Medicine, 1990. 172(6): p. 1535-1545.

252. Klagsbrun, M. and P. A. D'Amore, Vascular endothelial growth factor and its receptors. Cytokine &amp; Growth Factor Reviews, 1996. 7(3): p. 259-270.

253. Cross, M.J., et al., VEGF-receptor signal transduction. Trends in Biochemical Sciences, 2003. 28(9): p. 488-494.

254. Svensson, B., et al., Vascular endothelial growth factor protects cultured rat hippocampal neurons against hypoxic injury via an antiexcitotoxic, caspase-independent mechanism. J Cereb Blood Flow Metab, 2002. 22(10): p. 1170-5.

255. Hoeben, A., et al., Vascular Endothelial Growth Factor and Angiogenesis. Pharmacological Reviews, 2004. 56(4): p. 549-580.

256. Katoh, O., et al., Expression of the Vascular Endothelial Growth Factor (VEGF) Receptor Gene, KDR, in Hematopoietic Cells and Inhibitory Effect of VEGF on Apoptotic Cell Death Caused by Ionizing Radiation. Cancer Research, 1995. 55(23): p. 5687-5692.

257. Shalaby, F., et al., Failure of blood-island formation and vasculogenesis in Flk-1-deficient mice. Nature, 1995. 376(6535): p. 62-66.

258. Gerber, H.-P., et al., Vascular Endothelial Growth Factor Regulates Endothelial Cell Survival through the Phosphatidylinositol 3′-Kinase/Akt Signal Transduction Pathway: REQUIREMENT FOR Flk-1/KDR ACTIVATION. Journal of Biological Chemistry, 1998. 273(46): p. 30336-30343.

259. Gerber, H.-P., V. Dixit, and N. Ferrara, Vascular Endothelial Growth Factor Induces Expression of the Antiapoptotic Proteins Bcl-2 and A1 in Vascular Endothelial Cells. Journal of Biological Chemistry, 1998. 273(21): p. 13313-13316.

260. Fulton, D., et al., Regulation of endothelium-derived nitric oxide production by the protein kinase Akt. Nature, 1999. 399(6736): p. 597-601.

228

261. Dimmeler, S., et al., Activation of nitric oxide synthase in endothelial cells by Akt-dependent phosphorylation. Nature, 1999. 399(6736): p. 6001-605.

262. Rousseau, S., et al., p38 MAP kinase activation by vascular endothelial growth factor mediates actin reorganization and cell migration in human endothelial cells. Oncogene, 1997. 15(18): p. 2169-2177.

263. Qi, J.H. and L. Claesson-Welsh, VEGF-Induced Activation of Phosphoinositide 3-Kinase Is Dependent on Focal Adhesion Kinase. Experimental Cell Research, 2001. 263(1): p. 173-182.

264. Eliceiri, B.P., et al., Selective Requirement for Src Kinases during VEGF-Induced Angiogenesis and Vascular Permeability. Molecular Cell, 1999. 4(6): p. 915-924.

265. Soker, S., et al., Neuropilin-1 Is Expressed by Endothelial and Tumor Cells as an Isoform-Specific Receptor for Vascular Endothelial Growth Factor. Cell, 1998. 92(6): p. 735-745.

266. Gitay-Goren, H., et al., The binding of vascular endothelial growth factor to its receptors is dependent on cell surface-associated heparin-like molecules. Journal of Biological Chemistry, 1992. 267(9): p. 6093-8.

267. Ferrara, N. and T. Davis-Smyth, The Biology of Vascular Endothelial Growth Factor. Endocrine Reviews, 1997. 18(1): p. 4-25.

268. Ikeda, E., et al., Hypoxia-induced Transcriptional Activation and Increased mRNA Stability of Vascular Endothelial Growth Factor in C6 Glioma Cells. Journal of Biological Chemistry, 1995. 270(34): p. 19761-19766.

269. Levy, A.P., N.S. Levy, and M.A. Goldberg, Post-transcriptional Regulation of Vascular Endothelial Growth Factor by Hypoxia. Journal of Biological Chemistry, 1996. 271(5): p. 2746-2753.

270. Levy, A.P., et al., Transcriptional Regulation of the Rat Vascular Endothelial Growth Factor Gene by Hypoxia. Journal of Biological Chemistry, 1995. 270(22): p. 13333-13340.

271. Wang, G.L., et al., Hypoxia-inducible factor 1 is a basic-helix-loop-helix-PAS heterodimer regulated by cellular O2 tension. Proceedings of the National Academy of Sciences, 1995. 92(12): p. 5510-5514.

272. Salceda, S. and J. Caro, Hypoxia-inducible Factor 1α (HIF-1α) Protein Is Rapidly Degraded by the Ubiquitin-Proteasome System under Normoxic Conditions: ITS STABILIZATION BY HYPOXIA DEPENDS ON REDOX-INDUCED CHANGES. Journal of Biological Chemistry, 1997. 272(36): p. 22642-22647.

273. Pugh, C.W. and P.J. Ratcliffe, The von Hippel–Lindau tumor suppressor, hypoxia-inducible factor-1 (HIF-1) degradation, and cancer pathogenesis. Seminars in Cancer Biology, 2003. 13(1): p. 83-89.

229

274. Maxwell, P.H., et al., The tumour suppressor protein VHL targets hypoxia-inducible factors for oxygen-dependent proteolysis. Nature, 1999. 399(6733): p. 271-275.

275. Ruohola, J.K., et al., Vascular endothelial growth factors are differentially regulated by steroid hormones and antiestrogens in breast cancer cells. Molecular and Cellular Endocrinology, 1999. 149(1–2): p. 29-40.

276. Hyder, S.M. and G.M. Stancel, Regulation of Angiogenic Growth Factors in the Female Reproductive Tract by Estrogens and Progestins. Molecular Endocrinology, 1999. 13(6): p. 806-811.

277. Hyder, S.M., et al., Uterine Expression of Vascular Endothelial Growth Factor Is Increased by Estradiol and Tamoxifen. Cancer Research, 1996. 56(17): p. 3954-3960.

278. Lee, M.-Y., et al., Expression of vascular endothelial growth factor mRNA following transient forebrain ischemia in rats. Neuroscience Letters, 1999. 265(2): p. 107-110.

279. Bartholdi, D., B.P. Rubin, and M.E. Schwab, VEGF mRNA Induction Correlates With Changes in the Vascular Architecture Upon Spinal Cord Damage in the Rat. European Journal of Neuroscience, 1997. 9(12): p. 2549-2560.

280. Ma, Y., et al., Effects of vascular endothelial growth factor in ischemic stroke. Journal of Neuroscience Research, 2012. 90(10): p. 1873-1882.

281. Skold, M.K., et al., VEGF and VEGF Receptor Expression after Experimental Brain Contusion in Rat. Journal of Neurotrauma, 2005. 22(3): p. 353-367.

282. Sun, Y., et al., VEGF-induced neuroprotection, neurogenesis, and angiogenesis after focal cerebral ischemia. The Journal of Clinical Investigation, 2003. 111(12): p. 1843-1851.

283. Fitch, M.T. and J. Silver, Activated Macrophages and the Blood–Brain Barrier: Inflammation after CNS Injury Leads to Increases in Putative Inhibitory Molecules. Experimental Neurology, 1997. 148(2): p. 587-603.

284. McGraw, J., G.W. Hiebert, and J.D. Steeves, Modulating astrogliosis after neurotrauma. Journal of Neuroscience Research, 2001. 63(2): p. 109-115.

285. Jin, K.L., X.O. Mao, and D.A. Greenberg, Vascular endothelial growth factor: direct neuroprotective effect in in vitro ischemia. Proc Natl Acad Sci U S A, 2000. 97(18): p. 10242-7.

286. Jin, K.L., et al., Induction of vascular endothelial growth factor and hypoxia-inducible factor-1α by global ischemia in rat brain. Neuroscience, 2000. 99(3): p. 577-585.

287. Issa, R., et al., Vascular endothelial growth factor and its receptor, KDR, in human brain tissue after ischemic stroke. Laboratory investigation; a journal of technical methods and pathology, 1999. 79(4): p. 417-425.

230

288. Tsao, M.N., et al., Upregulation of Vascular Endothelial Growth Factor Is Associated with Radiation-Induced Blood-Spinal Cord Barrier Breakdown. Journal of Neuropathology & Experimental Neurology, 1999. 58(10): p. 1051-1060.

289. Vaquero, J., et al., Vascular endothelial growth/permeability factor in spinal cord injury. Journal of Neurosurgery: Spine, 1999. 90(2): p. 220-223.

290. Monacci, W.T., M.J. Merrill, and E.H. Oldfield, Expression of vascular permeability factor/vascular endothelial growth factor in normal rat tissues. American Journal of Physiology - Cell Physiology, 1993. 264(4): p. C995-C1002.

291. Peters, K.G., C. De Vries, and L.T. Williams, Vascular endothelial growth factor receptor expression during embryogenesis and tissue repair suggests a role in endothelial differentiation and blood vessel growth. Proceedings of the National Academy of Sciences, 1993. 90(19): p. 8915-8919.

292. Krum, J.M. and J.M. Rosenstein, VEGF mRNA and Its Receptor flt-1 Are Expressed in Reactive Astrocytes Following Neural Grafting and Tumor Cell Implantation in the Adult CNS. Experimental Neurology, 1998. 154(1): p. 57-65.

293. Krum, J.M. and J.M. Rosenstein, Transient Coexpression of Nestin, GFAP, and Vascular Endothelial Growth Factor in Mature Reactive Astroglia Following Neural Grafting or Brain Wounds. Experimental Neurology, 1999. 160(2): p. 348-360.

294. Papavassiliou, E., et al., Vascular endothelial growth factor (vascular permeability factor) expression in injured rat brain. Journal of Neuroscience Research, 1997. 49(4): p. 451-460.

295. Cobbs, C.S., et al., Vascular endothelial growth factor expression in transient focal cerebral ischemia in the rat. Neuroscience Letters, 1998. 249(2–3): p. 79-82.

296. Lennmyr, F., et al., Expression of vascular endothelial growth factor (VEGF) and its receptors (Flt-1 and Flk-1) following permanent and transient occlusion of the middle cerebral artery in the rat. Journal of Neuropathology and Experimental Neurology, 1998. 57(9): p. 874-882.

297. Wang, Y., et al., VEGF-overexpressing transgenic mice show enhanced post-ischemic neurogenesis and neuromigration. J Neurosci Res, 2007. 85(4): p. 740-7.

298. Widenfalk, J., et al., Vascular endothelial growth factor improves functional outcome and decreases secondary degeneration in experimental spinal cord contusion injury. Neuroscience, 2003. 120(4): p. 951-60.

299. Kaneko, S., et al., A selective Sema3A inhibitor enhances regenerative responses and functional recovery of the injured spinal cord. Nat Med, 2006. 12(12): p. 1380-1389.

300. Ohab, J.J., et al., A Neurovascular Niche for Neurogenesis after Stroke. J. Neurosci., 2006. 26(50): p. 13007-13016.

231

301. Peters, M.C., P.J. Polverini, and D.J. Mooney, Engineering vascular networks in porous polymer matrices. Journal of Biomedical Materials Research, 2002. 60(4): p. 668-678.

302. Raab, S. and K. Plate, Different networks, common growth factors: shared growth factors and receptors of the vascular and the nervous system. Acta Neuropathologica, 2007. 113(6): p. 607-626.

303. Bearden, S.E. and S.S. Segal, Neurovascular Alignment in Adult Mouse Skeletal Muscles. Microcirculation, 2005. 12(2): p. 161-167.

304. Gene Therapy, 2011, Human Genome Project.

305. Volpers, C. and S. Kochanek, Chapter 8: Viral Gene Transfer into Endothelial Cells, in Methods In Endothelial Cell Biology, H.G. Augustin, Editor 2004, Springer Verlag. p. 73-82.

306. Stewart, P.L., Chapter 1: Adenovirus Structure, in Adenoviral Vectors for Gene Therapy, D.T. Curiel and J.T. Douglas, Editors. 2002, Elsevier Science. p. 1-18.

307. Laupacis, A., et al., Cyclosporin A: a powerful immunosuppressant. Canadian Medical Association Journal, 1982. 126(9): p. 1041-1046.

308. Chirmule, N., et al., Immune responses to adenovirus and adeno-associated virus in humans. Gene Therapy, 1999. 6(9): p. 1574-1583.

309. Davidson, B.L. and X.O. Breakefield, Viral vectors for gene delivery to the nervous system. Nature Reviews Neuroscience, 2003. 4(5): p. 353-364.

310. Grieger, J. and R. Samulski, Adeno-associated Virus as a Gene Therapy Vector: Vector Development, Production and Clinical Applications, in Gene Therapy and Gene Delivery Systems, D. Schaffer and W. Zhou, Editors. 2005, Springer Berlin / Heidelberg. p. 119-145.

311. Surosky, R.T., et al., Adeno-associated virus Rep proteins target DNA sequences to a unique locus in the human genome. Journal of Virology, 1997. 71(10): p. 7951-9.

312. Lu, Y., Recombinant Adeno-Associated Virus As Delivery Vector for Gene Therapy—A Review. Stem Cells and Development, 2004. 13(1): p. 133-145.

313. Fischer, A.C., et al., Successful transgene expression with serial doses of aerosolized rAAV2 vectors in rhesus macaques. Molecular Therapy, 2003. 8(6): p. 918-926.

314. Nicklin, S.A., et al., Efficient and Selective AAV2-Mediated Gene Transfer Directed to Human Vascular Endothelial Cells. Molecular Therapy, 2001. 4: p. 174–181.

315. Bartlett, J.S., R.J. Samulski, and T.J. McCown, Selective and Rapid Uptake of Adeno-Associated Virus Type 2 in Brain. Human Gene Therapy, 1998. 9(8): p. 1181-1186.

232

316. Rebar, E.J., et al., Induction of angiogenesis in a mouse model using engineered transcription factors. Nat Med, 2002. 8(12): p. 1427-32.

317. Siddiq, I., et al., Treatment of Traumatic Brain Injury Using Zinc-Finger Protein Gene Therapy Targeting VEGF-A. Journal of Neurotrauma, 2012. Online: ahead of print.

318. Dai, Q., et al., Engineered zinc finger-activating vascular endothelial growth factor transcription factor plasmid DNA induces therapeutic angiogenesis in rabbits with hindlimb ischemia. Circulation, 2004. 110(16): p. 2467-75.

319. Yu, J., et al., An engineered VEGF-activating zinc finger protein transcription factor improves blood flow and limb salvage in advanced-age mice. Faseb J, 2006. 20(3): p. 479-81.

320. Li, Y., et al., In mice with type 2 diabetes, a vascular endothelial growth factor (VEGF)-activating transcription factor modulates VEGF signaling and induces therapeutic angiogenesis after hindlimb ischemia. Diabetes, 2007. 56(3): p. 656-65.

321. Price, S.A., et al., Gene transfer of an engineered transcription factor promoting expression of VEGF-A protects against experimental diabetic neuropathy. Diabetes, 2006. 55(6): p. 1847-54.

322. Siddiq, I., Upregulation of VEGF-A using Engineered Zinc Finger Protein Gene Therapy Increases Cell Survival After Lateral Fluid Percussion Injury in Rats, in Institute of Medical Science2011, University of Toronto: Toronto.

323. Tator, C.H. and M.G. Fehlings, Review of the secondary injury theory of acute spinal cord trauma with emphasis on vascular mechanisms. Journal of Neurosurgery, 1991. 75(1): p. 15-26.

324. Glaser, J., et al., Neutralization of the chemokine CXCL10 reduces apoptosis and increases axon sprouting after spinal cord injury. Journal of Neuroscience Research, 2006. 84(4): p. 724-734.

325. Kaneko, S., et al., A selective Sema3A inhibitor enhances regenerative responses and functional recovery of the injured spinal cord. Nature Medicine, 2006. 12(12): p. 1380-1389.

326. Yoshihara, T., et al., Neuroprotective Effect of Bone Marrow–Derived Mononuclear Cells Promoting Functional Recovery from Spinal Cord Injury. Journal of Neurotrauma, 2007. 24(6): p. 1026-1036.

327. Fawcett, J.W. and R.A. Asher, The glial scar and central nervous system repair. Brain Research Bulletin, 1999. 49(6): p. 377-391.

328. Bonkowski, D., et al., The CNS microvascular pericyte: pericyte-astrocyte crosstalk in the regulation of tissue survival. Fluids and Barriers of the CNS, 2011. 8(1): p. 8.

233

329. Azzouz, M., et al., VEGF delivery with retrogradely transported lentivector prolongs survival in a mouse ALS model. Nature, 2004. 429(6990): p. 413-417.

330. Lambrechts, D., et al., VEGF is a modifier of amyotrophic lateral sclerosis in mice and humans and protects motoneurons against ischemic death. Nat Genet, 2003. 34(4): p. 383-394.

331. Siddiq, I., et al., Treatment of Traumatic Brain Injury Using Zinc-Finger Protein Gene Therapy Targeting VEGF-A. J Neurotrauma, 2012.

332. Hobson, M.I., C.J. Green, and G. Terenghi, VEGF enhances intraneural angiogenesis and improves nerve regeneration after axotomy. Journal of Anatomy, 2000. 197(4): p. 591-605.

333. Kim, H.M., et al., VEGF Delivery by Neural Stem Cells Enhances Proliferation of Glial Progenitors, Angiogenesis, and Tissue Sparing after Spinal Cord Injury. PLoS ONE, 2009. 4(3): p. e4987.

334. Liu, P.Q., et al., Regulation of an endogenous locus using a panel of designed zinc finger proteins targeted to accessible chromatin regions. Activation of vascular endothelial growth factor A. J Biol Chem, 2001. 276(14): p. 11323-34.

335. Ikeda, Y., M. Wang, and S. Nakazawa, Simple quantitative evaluation of blood-brain barrier disruption in vasogenic brain edema. Acta neurochirurgica. Supplementum, 1994. 60: p. 119-120.

336. Saria, A. and J.M. Lundberg, Evans blue fluorescence: quantitative and morphological evaluation of vascular permeability in animal tissues. Journal of Neuroscience Methods, 1983. 8(1): p. 41-49.

337. Basso, D.M., M.S. Beattie, and J.C. Bresnahan, A sensitive and reliable locomotor rating scale for open field testing in rats. J Neurotrauma, 1995. 12(1): p. 1-21.

338. Koopmans, G.C., et al., The assessment of locomotor function in spinal cord injured rats: the importance of objective analysis of coordination. J Neurotrauma, 2005. 22(2): p. 214-25.

339. Hamers, F.P., G.C. Koopmans, and E.A. Joosten, CatWalk-assisted gait analysis in the assessment of spinal cord injury. J Neurotrauma, 2006. 23(3-4): p. 537-48.

340. Gensel, J.C., et al., Behavioral and histological characterization of unilateral cervical spinal cord contusion injury in rats. J Neurotrauma, 2006. 23(1): p. 36-54.

341. Fehlings, M.G., et al., Motor evoked potentials recorded from normal and spinal cord-injured rats. Neurosurgery, 1987. 20(1): p. 125-130.

342. Nashmi, R., et al., Serial recording of somatosensory and myoelectric motor evoked potentials: role in assessing functional recovery after graded spinal cord injury in the rat. J Neurotrauma, 1997. 14(3): p. 151-9.

234

343. Holm, S., A simple sequentially rejective multiple test procedure. Scandinavian Journal of Statistics, 1979. 6(2): p. 65-70.

344. Glantz, S.A., Primer of Biostatistics. Sixth ed2005: McGraw-Hill Companies,Inc. 520.

345. Holtz, A., B. Nystrom, and B. Gerdin, Relation between spinal cord blood flow and functional recovery after blocking weight-induced spinal cord injury in rats. Neurosurgery, 1990. 26(6): p. 952-7.

346. Ohab, J.J., et al., A Neurovascular Niche for Neurogenesis after Stroke. The Journal of Neuroscience, 2006. 26(50): p. 13007-13016.

347. Bearden, S.E. and S.S. Segal, Microvessels Promote Motor Nerve Survival and Regeneration Through Local VEGF Release Following Ectopic Reattachment. Microcirculation, 2004. 11(8): p. 633-644.

348. Casella, G.T.B., et al., New Vascular Tissue Rapidly Replaces Neural Parenchyma and Vessels Destroyed by a Contusion Injury to the Rat Spinal Cord. Experimental Neurology, 2002. 173(1): p. 63-76.

349. Schumacher, P.A., J.H. Eubanks, and M.G. Fehlings, Increased calpain I-mediated proteolysis, and preferential loss of dephosphorylated NF200, following traumatic spinal cord injury. Neuroscience, 1999. 91(2): p. 733-744.

350. von Euler, M., Å. Seiger, and E. Sundström, Clip Compression Injury in the Spinal Cord: A Correlative Study of Neurological and Morphological Alterations. Experimental Neurology, 1997. 145(2): p. 502-510.

351. Grossman, S.D., L.J. Rosenberg, and J.R. Wrathall, Temporal-Spatial Pattern of Acute Neuronal and Glial Loss after Spinal Cord Contusion. Experimental Neurology, 2001. 168(2): p. 273-282.

352. Noble, L.J. and J.R. Wrathall, Blood-spinal cord barrier disruption proximal to a spinal cord transection in the rat: Time course and pathways associated with protein leakage. Experimental Neurology, 1988. 99(3): p. 567-578.

353. Noble, L.J. and J.R. Wrathall, The blood-spinal cord barrier after injury: pattern of vascular events proximal and distal to a transection in the rat. Brain Research, 1987. 424(1): p. 177-188.

354. Benton, R.L. and S.R. Whittemore, VEGF165 therapy exacerbates secondary damage following spinal cord injury. Neurochem Res, 2003. 28(11): p. 1693-703.

355. Gruner, J.A., A monitored contusion model of spinal cord injury in the rat. Journal of Neurotrauma, 1992. 9(2): p. 123-126.

356. Rivlin, A.S. and C.H. Tator, Regional spinal cord blood flow in rats after severe cord trauma. Journal of Neurosurgery, 1978. 49(6): p. 844-853.

235

357. Sandler, A.N. and C.H. Tator, Effect of acute spinal cord compression injury on regional spinal cord blood flow in primates. Journal of Neurosurgery, 1976. 45(6): p. 660-676.

358. Fehlings, M.G. and C.H. Tator, The relationships among the severity of spinal cord injury, residual neurological function, axon counts, and counts of retrogradely labeled neurons after experimental spinal cord injury. Exp Neurol, 1995. 132(2): p. 220-8.

359. Choi, U.H., et al., Hypoxia-inducible expression of vascular endothelial growth factor for the treatment of spinal cord injury in a rat model. J Neurosurg Spine, 2007. 7(1): p. 54-60.

360. Facchiano, F., et al., Promotion of regeneration of corticospinal tract axons in rats with recombinant vascular endothelial growth factor alone and combined with adenovirus coding for this factor. J Neurosurg, 2002. 97(1): p. 161-8.

361. Dijkers, M., T. Bryce, and J. Zanca, Prevalence of chronic pain after traumatic spinal cord injury: A systematic review. Journal of Rehabilitation Research & Development 2009. 46(1): p. 13-30.

362. Nesic, O., et al., Vascular Endothelial Growth Factor and Spinal Cord Injury Pain. Journal of Neurotrauma, 2010. 27(10): p. 1793-1803.

363. Fehlings, M.G. and R.G. Perrin, The Timing of Surgical Intervention in the Treatment of Spinal Cord Injury: A Systematic Review of Recent Clinical Evidence. Spine, 2006. 31(11S (Supplement)): p. S28-S35.

364. Bracken, M.B., et al., A Randomized, Controlled Trial of Methylprednisolone or Naloxone in the Treatment of Acute Spinal-Cord Injury. New England Journal of Medicine, 1990. 322(20): p. 1405-1411.

365. Flynn, J.R., et al., The role of propriospinal interneurons in recovery from spinal cord injury. Neuropharmacology, 2011. 60(5): p. 809-822.

366. Pearse, D.D., et al., cAMP and Schwann cells promote axonal growth and functional recovery after spinal cord injury. Nature Medicine, 2004. 10(6): p. 610-616.

367. Ballermann, M. and K. Fouad, Spontaneous locomotor recovery in spinal cord injured rats is accompanied by anatomical plasticity of reticulospinal fibers. European Journal of Neuroscience, 2006. 23(8): p. 1988-1996.

368. Defrin, R., et al., Characterization of chronic pain and somatosensory function in spinal cord injury subjects. Pain, 2001. 89(2-3): p. 253-263.

369. Österberg, A. and J. Boivie, Central pain in multiple sclerosis – Sensory abnormalities. European Journal of Pain, 2010. 14(1): p. 104-110.

370. Xu, X.J., et al., Chronic pain-related syndrome in rats after ischemic spinal cord lesion: a possible animal model for pain in patients with spinal cord injury. Pain, 1992. 48(2): p. 279-290.

236

371. Vierck Jr, C.J., P. Siddall, and R.P. Yezierski, Pain following spinal cord injury: animal models and mechanistic studies. Pain, 2000. 89(1): p. 1-5.

372. Hoheisel, U., et al., Pathophysiological activity in rat dorsal horn neurones in segments rostral to a chronic spinal cord injury. Brain Research, 2003. 974(1–2): p. 134-145.

373. Choi, J.-S., et al., Upregulation of Vascular Endothelial Growth Factor Receptors Flt-1 and Flk-1 Following Acute Spinal Cord Contusion in Rats. Journal of Histochemistry & Cytochemistry, 2007. 55(8): p. 821-830.

374. Nör, J.E., et al., Vascular endothelial growth factor (VEGF)-mediated angiogenesis is associated with enhanced endothelial cell survival and induction of Bcl-2 expression. The American journal of pathology, 1999. 154(2): p. 375-384.

375. Keyt, B.A., et al., The carboxyl-terminal domain (111-165) of vascular endothelial growth factor is critical for its mitogenic potency. J Biol Chem, 1996. 271(13): p. 7788-95.

376. Fassbender, J., S. Whittemore, and T. Hagg, Targeting Microvasculature for Neuroprotection after SCI. Neurotherapeutics, 2011. 8(2): p. 240-251.

377. Ohab, J.J., et al., A neurovascular niche for neurogenesis after stroke. Journal of Neuroscience, 2006. 26(50): p. 13007-13016.

378. Neufeld, G., O. Kessler, and Y. Herzog, The Interaction of Neuropilin-1 and Neuropilin-2 with Tyrosine-Kinase Receptors for VEGF, D. Bagnard, Editor 2003, Springer US. p. 81-90.

379. Zhang, Z.G., et al., VEGF enhances angiogenesis and promotes blood-brain barrier leakage in the ischemic brain. J Clin Invest, 2000. 106(7): p. 829-38.

380. Sun, Y., et al., VEGF-induced neuroprotection, neurogenesis, and angiogenesis after focal cerebral ischemia. J Clin Invest, 2003. 111(12): p. 1843-51.

381. Ng, Y.S., et al., Differential expression of VEGF isoforms in mouse during development and in the adult. Developmental Dynamics, 2001. 220(2): p. 112-121.

382. Barleon, B., et al., Migration of human monocytes in response to vascular endothelial growth factor (VEGF) is mediated via the VEGF receptor flt-1. Blood, 1996. 87(8): p. 3336-3343.

383. Schratzberger, P., et al., Reversal of experimental diabetic neuropathy by VEGF gene transfer. The Journal of Clinical Investigation, 2001. 107(9): p. 1083-1092.

384. Bakshi, A., et al., Mechanically engineered hydrogel scaffolds for axonal growth and angiogenesis after transplantation in spinal cord injury. Journal of Neurosurgery: Spine, 2004. 1(3): p. 322-329.

237

385. Kim, H.M., et al., Ex Vivo VEGF Delivery by Neural Stem Cells Enhances Proliferation of Glial Progenitors, Angiogenesis, and Tissue Sparing after Spinal Cord Injury. PLoS ONE, 2009. 4(3): p. e4987.

386. Stirling, D.P., et al., Minocycline Treatment Reduces Delayed Oligodendrocyte Death, Attenuates Axonal Dieback, and Improves Functional Outcome after Spinal Cord Injury. The Journal of Neuroscience, 2004. 24(9): p. 2182-2190.

387. Weaver, L.C., et al., Methylprednisolone Causes Minimal Improvement after Spinal Cord Injury in Rats, Contrasting with Benefits of an Anti-Integrin Treatment. Journal of Neurotrauma, 2005. 22(12): p. 1375-1387.

388. Pan, W. and A.J. Kastin, Increase in TNFα Transport after SCI Is Specific for Time, Region, and Type of Lesion. Experimental Neurology, 2001. 170(2): p. 357-363.

389. Klusman, I. and M.E. Schwab, Effects of pro-inflammatory cytokines in experimental spinal cord injury. Brain Research, 1997. 762(1): p. 173-184.

390. Yezierski, R.P., et al., Neuronal Damage Following Intraspinal Injection of a Nitric Oxide Synthase Inhibitor in the Rat. J Cereb Blood Flow Metab, 1996. 16(5): p. 996-1004.

391. Azzouz, M., et al., Increased motoneuron survival and improved neuromuscular function in transgenic ALS mice after intraspinal injection of an adeno-associated virus encoding Bcl-2. Human Molecular Genetics, 2000. 9(5): p. 803-811.

392. Burger, C., et al., Recombinant AAV viral vectors pseudotyped with viral capsids from serotypes 1, 2, and 5 display differential efficiency and cell tropism after delivery to different regions of the central nervous system. Mol Ther, 2004. 10(2): p. 302-17.

393. Madsen, D., et al., Adeno-associated virus serotype 2 induces cell-mediated immune responses directed against multiple epitopes of the capsid protein VP1. Journal of General Virology, 2009. 90(11): p. 2622-2633.

394. Hicklin, D.J. and L.M. Ellis, Role of the Vascular Endothelial Growth Factor Pathway in Tumor Growth and Angiogenesis. Journal of Clinical Oncology, 2005. 23(5): p. 1011-1027.

395. Austin, J.W., et al., The effects of intrathecal injection of a hyaluronan-based hydrogel on inflammation, scarring and neurobehavioural outcomes in a rat model of severe spinal cord injury associated with arachnoiditis. Biomaterials, 2012. 33(18): p. 4555-4564.

396. Kang, C.E., et al., A New Paradigm for Local and Sustained Release of Therapeutic Molecules to the Injured Spinal Cord for Neuroprotection and Tissue Repair. Tissue Engineering, 2008. 15(3): p. 595-604.

397. Naldini, A. and F. Carraro, Role of inflammatory mediators in angiogenesis. Curr Drug Targets Inflamm Allergy, 2005. 4(1): p. 3-8.

238

398. Imhof, B.A. and M. Aurrand-Lions, Angiogenesis and inflammation face off. Nature Medicine, 2006. 12(2): p. 171-172.

399. Folkman, J., W. Li, and R. Casey, Inflammation and Angiogenesis, in Progress in Immunology, F. Melchers, et al., Editors. 1989, Springer Berlin Heidelberg. p. 761-764.

400. McMahon, S.S., et al., Effect of cyclosporin A on functional recovery in the spinal cord following contusion injury. J Anat, 2009. 215(3): p. 267-79.

401. Diaz-Ruiz, A., et al., Cyclosporin-A inhibits lipid peroxidation after spinal cord injury in rats. Neuroscience Letters, 1999. 266(1): p. 61-64.

402. Sullivan, P.G., et al., Dose-response curve and optimal dosing regimen of cyclosporin A after traumatic brain injury in rats. Neuroscience, 2000. 101(2): p. 289-295.

403. Sullivan, P.G., et al., Mitochondrial permeability transition in CNS trauma: Cause or effect of neuronal cell death? Journal of Neuroscience Research, 2005. 79(1-2): p. 231-239.

404. Okonkwo, D.O., et al., Cyclosporin A limits calcium-induced axonal damage following traumatic brain injury. Neuroreport, 1999. 10(2): p. 353-358.

405. Scheff, S.W. and P.G. Sullivan, Cyclosporin A significantly ameliorates cortical damage following experimental traumatic brain injury in rodents. J Neurotrauma, 1999. 16(9): p. 783-92.

406. Patel, C.B., et al., Effect of VEGF Treatment on the Blood-Spinal Cord Barrier Permeability in Experimental Spinal Cord Injury: Dynamic Contrast-Enhanced Magnetic Resonance Imaging. Journal of Neurotrauma, 2009. 26(7): p. 1005-1016.

407. Bunge, M.B., Book Review: Bridging Areas of Injury in the Spinal Cord. The Neuroscientist, 2001. 7(4): p. 325-339.

408. Hari Shanker, S., Neurotrophic Factors in Combination: A Possible new Therapeutic Strategy to Influence Pathophysiology of Spinal Cord Injury and Repair Mechanisms. Current Pharmaceutical Design, 2007. 13(18): p. 1841-1874.

409. Lutton, C., et al., Combined VEGF and PDGF Treatment Reduces Secondary Degeneration after Spinal Cord Injury. Journal of Neurotrauma, 2012. 29(5): p. 957-970.

239

Neurobiology of Disease 37 (2010) 384–393

Contents lists available at ScienceDirect

Neurobiology of Disease

j ourna l homepage: www.e lsev ie r.com/ locate /ynbd i

An engineered transcription factor which activates VEGF-A enhances recovery afterspinal cord injury

Yang Liu a, Sarah Figley a,d, S. Kaye Spratt b, Gary Lee b, Dale Ando b,Richard Surosky b, Michael G. Fehlings a,c,d,⁎a Department of Genetics and Development, Toronto Western Research Institute, and Spinal Program, Krembil Neuroscience Centre, University Health Network, Toronto, Ontario, Canadab Department of Therapeutic Development, Sangamo BioSciences, Pt. Richmond, CA, USAc Department of Surgery, University of Toronto, Ontario, Canadad Institute of Medical Sciences, University of Toronto, Ontario, Canada

⁎ Corresponding author. Division of Neurosurgery,Development, University Health Network, University4WW-449, Toronto, ON, Canada M5T 2S8. Fax: +1 416

E-mail address: Michael.Fehlings@uhn.on.ca (M.G. FAvailable online on ScienceDirect (www.scienced

0969-9961/$ – see front matter © 2009 Elsevier Inc. Adoi:10.1016/j.nbd.2009.10.018

a b s t r a c t

a r t i c l e i n f o

Article history:Received 8 May 2009Revised 9 October 2009Accepted 22 October 2009Available online 29 October 2009

Keywords:SCIZFP-VEGFAngiogenesisNeuroprotectionMolecular therapy

Spinal cord injury (SCI) leads to local vascular disruption and progressive ischemia, which contribute tosecondary degeneration. Enhancing angiogenesis through the induction of vascular endothelial growthfactor (VEGF)-A expression therefore constitutes an attractive therapeutic approach. Moreover, emergingevidence suggests that VEGF-A may also exhibit neurotrophic, neuroprotective, and neuroproliferativeeffects. Building on this previous work, we seek to examine the potential therapeutic benefits of anengineered zinc finger protein (ZFP) transcription factor designed to activate expression of all isoforms ofendogenous VEGF-A (ZFP-VEGF). Administration of ZFP-VEGF resulted in increased VEGF-A mRNA andprotein levels, an attenuation of axonal degradation, a significant increase in vascularity and decreased levelsof apoptosis. Furthermore, ZFP-VEGF treated animals showed significant improvements in tissuepreservation and neurobehavioural outcomes. These data suggest that activation of VEGF-A via theadministration of an engineered ZFP transcription factor holds promise as a therapy for SCI and potentiallyother forms of neurotrauma.

© 2009 Elsevier Inc. All rights reserved.

Introduction

Spinal cord injury (SCI) is a leading cause of death and neurologicaldisability. The pathophysiology of SCI involves a primary mechanicalinjury followed by a series of secondary molecular and cellular events(Fehlings et al., 1989). Compromised blood flow, hemorrhage, cordcompression, intravascular thrombosis, and vasospasm contribute toischemia, which initiates events that impair angiogenesis, amongother reparative processes (Tator and Fehlings, 1991). Angiogenicfactors, such as vascular endothelial growth factor (VEGF)-A, promotethe proliferation of vascular endothelial cells and angiogenesis(Shweiki et al., 1992). Emerging evidence suggests that VEGF-A alsohas neurotrophic, neuroprotective, and neuroproliferative effects(Greenberg and Jin, 2005). VEGF-A is a homodimeric glycoproteinthat is expressed as multiple splice variants encoded by a single gene(Leung et al., 1989). The most common and best-studied isoforms inthe central nervous system are VEGF121, VEGF165 and VEGF189.Increased expression of VEGF-A and its receptors during hypoxic/

and Division of Genetics andof Toronto, 399 Bathurst St.603 5745.ehlings).irect.com).

ll rights reserved.

24

ischemic injury to the brain and spinal cord suggests that VEGF-Acould play a neuroprotective role in these pathophysiologicalprocesses. Previous approaches using VEGF-A have relied on theintroduction of a single splice isoform of VEGF-A, which result indisappointing outcomes in human clinical trials because of increasedperipheral edema (Baumgartner et al., 2000; Rajagopalan et al., 2003).These suboptimal results may stem from the fact that the centralnervous system expresses several isoforms of VEGF-A. A morecomprehensive approach, in which several isoforms of this gene areexpressed, may havemore striking results. To this end, a panel of zinc-finger protein transcription factors (ZFPs) have been successfullydesigned that bind with high affinity to diverse DNA sequencespresent within the VEGF-A locus and that are capable of activating theexpression of multiple splice variants of the endogenous chromo-somal VEGF-A gene. It has been demonstrated previously that theexpression of ZFPs in vivo can induce expression of the VEGF-Aprotein, stimulate angiogenesis, and accelerate wound healing (Rebaret al., 2002). Evidence for a potentially therapeutic biophysiologicaleffect of ZFPs has been reported in animal models of hindlimbischemia (Dai et al., 2004; Xie et al., 2006; Yu et al., 2006) and diabetes(Li et al., 2007; Price et al., 2006).

To determine the effects of VEGF-A expression in a rodentmodel ofSCI, we employed a virally delivered ZFP transcription factor whichactivates VEGF-A expression (ZFP-VEGF). Recombinant adenoviral

0

385Y. Liu et al. / Neurobiology of Disease 37 (2010) 384–393

(Ad) and adeno-associated viral (AAV) vectors are both promisingcandidates for gene therapy. Ad vectors result in rapid, high yieldtransgene expression; however, gene expression is transient (Her-mens et al., 1997). In contrast to Ad vectors, transgene expression byan AAV vector increases at a relatively slower rate and is sustained fora longer period (Burger et al., 2004). We have, therefore, performedtwo separate sets of experiments to determine both the acute (usingan Ad vector) and long-term (using an AAV vector) effects of ZFP-VEGF. We report novel findings indicating that ZFP-VEGF deliverypromotes neuroprotection and angiogenesis after SCI, resulting insignificant tissue sparing and neurobehavioural recovery at morechronic time points.

Materials and methods

All animal experiments were conducted with approval from theAnimal Care Committee, University Health Network (Toronto,Canada).

Viral vector constructs

The VEGF-A-activating ZFP and controls were provided in viralvectors by Sangamo BioSciences (Pt. Richmond, CA) and have beenpreviously described (Liu et al., 2001; Price et al., 2006). A diagram ofthe ZFP-VEGF expression cassette is illustrated in Fig. 1. 32E-p65 is a378 amino acid multidomain protein that is composed of threefunctional regions: (a) the nuclear localization signal (NLS) of thelarge T-antigen of SV40, (b) a designed 3-finger zinc-fingered protein(32E) that binds to a 9 base-pair target DNA sequence (GGGGGTGAC)present in the human VEGF-A promoter region and (c) thetransactivation domain from the p65 subunit of human NF-κB,which is identical to VZ+434, subcloned into pVAX1 (Invitrogen,San Diego, CA) with expression driven by the human cytomegalovirus(CMV) promoter. Adenoviral (Ad5-32Ep65 or Ad5-DsRed) and AAV(AAV2-32Ep65 or AAV2-GFP) vectors were packaged by transfectingT-REx-293 cells (Invitrogen, San Diego, CA). T-REx-293 cells in ten-stack cell factories were inoculated with Ad vectors at a multiplicity ofinfection (MOI) of 50 to 100 particles per cell. When adenoviralmediated cytopathy effect (CPE) was observed, cells were harvestedand lysed by three cycles of freezing and thawing. Crude lysates wereclarified by centrifugation. AAV2-32Ep65 is a first generation secondsingle-stranded AAV-2 vector. AAV vectors were produced in five-stack cell factories. 293 cells were seeded at 4×107 and grown 3 daysprior to transfection. The calcium phosphate method was used fortransfection. After three days, the cells were harvested and AAV waspurified by two rounds of cesium chloride density gradient centrifu-gation. The cesium chloride was removed by dialysis againstphosphate buffer saline (PBS) with additional sodium chloride to200 mM and Pluronic F 68 (Sigma Aldrich, St. Lois, MO) to 0.01%.Infectious titers of the Ad vectors were quantified using the Adeno-XRapid Titer kit (Clontech, Mountain View, CA). The genome copynumber of AAV was determined using Taqman polymerase chainreaction (PCR) (Applied Biosystems, Foster City, CA).

Fig. 1. ZFP-VEGF Expression Cassette. CMV pro—cytomegalovirus promoter/enhancer;NLS—nuclear localization sequence; ZFP-VEGF—engineered VEGF transcriptionalactivator; NF-κB p65 TAD—transactivation domain from the p65 subunit of humanNF-κB; bGH pA—bovine growth hormone polyadenylation sequence. Arrow indicatestranscription initiation site. Control viruses Ad-DsRed and AAV-GFP have been designedwith both VEGF-ZFP and NF-κB p65 domains deleted, and either DsRed or GFP domainsinserted, respectively.

241

SCI and intraspinal microinjection

The aneurysm clip compression model of SCI used in ourlaboratory has been characterized extensively and described previ-ously (Fehlings and Tator, 1995). Briefly, adult female Wistar rats(250–300 g; Charles River, Montreal, Canada) received laminectomiesof midthoracic vertebral segments T6-T7. A modified clip calibrated toa closing force of 35 g was applied extradurally to the cord forduration of 1 min. Ad vectors were used for acute phase (3–10 days)and AAV vectors were used for a more chronic evaluation period(6 weeks). The animals were divided into two groups in a randomizedand “blinded” manner, Ad/AAV-ZFP-VEGF treatment group and Ad-DsRed/AAV-GFP control group. Using a stereotaxic frame and glasscapillary needle (tip diameter 60 μm) connected to a Hamiltonmicrosyringe, a total of 5×108 viral particles (Ad) or 7.5×106 (AAV)plaque forming units (PFU) were injected into the dorsal spinal cordimmediately after SCI. Four 2.5 μl (10 μl total) intraspinal injectionswere made bilaterally at 2 mm rostral and caudal of the injury site.The injection rate is 0.5 μl/min and at the end of injection, thecapillary was left in the cord for at least 1 min to allow diffusion fromthe injection site. A subcutaneous injection of 10mg/kg of cyclosporinA was administered daily starting 24 h prior to the SCI until the end ofthe experiments for immunosuppression in Ad vector injected animalgroups, which have been shown to illicit a non-specific immunereaction.

Measurement of VEGF mRNA by real-time PCR

Three days after SCI and viral vector injection, spinal cord tissues(5-mm piece of spinal cord, centered on the injury site) were takenand homogenized in Trizol (Invitrogen, Burlington, Canada) underRNAse-free conditions. RNeasy Mini Reagent Set (Qiagen Inc.,Mississauga, Canada) was used to isolate RNA. Two micrograms oftotal RNA were reverse transcribed using SuperScript™ II RNase H-reverse transcriptase. The mRNA level of VEGF isoforms werequantified by real-time PCR on the ABI 7900 HT Fast Real-Time PCRSystem using SYBR green PCR master mix reagent kit (AppliedBiosystems, Foster City, CA). Primers for VEGF-A (Sigma, Oakville,Canada) isoforms were as follow (Adris et al., 2005; Zhang et al.,2002): VEGF-A isoform common forward primer, 5′-GCC AGC ACATAG GAG AGA TGA GC-3′; VEGF120 reverse primer, 5′-GGC TTG TCACAT TTT TCT GG-3′; VEGF164 reverse primer, 5′-CAA GGC TCA CAGTGA TTT TCT GG-3′; VEGF188 reverse primer, 5′-AAC AAG GCT CACAGT GAA CGC T-3′; hypoxanthine phosphoribosyl transferase 1(HPRT1) forward primer, 5′-GCC CCA AAA TGG TTA AGG TT-3′;HPRT1 reverse primer, 5′-CCG CTG TCT TTT AGG CTT TG-3′. The real-time quantitative PCR assay was performed twice, and each samplewas tested in triplicate. “No-template” and “no-amplification” con-trols were included for each gene, and melt curves showed a singlepeak, confirming specific amplification (Bustin and Nolan, 2004). Thethreshold cycle (CT) for each gene was determined and normalizedagainst the housekeeping gene HPRT1.

Western blotting

A 5-mm length of the spinal cord centered at the injury site wastaken and 20 μg of proteinwas loaded into 7.5% or 12% polyacrylamidegels (Bio-Rad, Mississauga, Canada). Membranes were probed witheither monoclonal anti-NF200 antibody (1:2000; Sigma, Oakville,Canada), rabbit IgG anti-VEGF-A antibody (1:100; Santa CruzBiotechnology, Santa Cruz, CA), or rabbit IgG anti-NFκBp65 (1:1000;Santa Cruz Biotechnology, Santa Cruz, CA). NFκBp65 rabbit polyclonalantibody recognizes the p65 activation domain in the ZFP-VEGFtreated animals. Primary antibodies were labeled with horseradishperoxidase-conjugated secondary antibodies (goat anti-mouse/rabbitIgG, 1:3000; Jackson Immuno Research Laboratories, West Grove, PA),

386 Y. Liu et al. / Neurobiology of Disease 37 (2010) 384–393

and bands were imaged using an enhanced chemiluminescence (ECL)detection system (Perkin Elmer, Woodbridge, Canada). Mousemonoclonal, beta-actin (Chemicon International, Inc., Temecula, CA)was immunoblotted as a loading control. Gel-Pro Plus Analyzersoftware (Media Cybernetics Inc., MD) was used for integrated opticaldensity (OD) analysis.

Histochemistry

Histological processingAnimals were perfused transcardially with 4% paraformaldehyde

(PFA) in 0.1 M PBS. Then, the tissues were cryoprotected in 20%sucrose in PBS. A 1 cm length of the spinal cord centered at the injurysite was embedded in tissue-embedding medium. The injuredsegment was snap frozen and sectioned on a cryostat at a thicknessof 12 μm. Serial spinal cord sections at 500 μm intervals were stainedwith myelin-selective pigment luxol fast blue (LFB) and the cellularstain hematoxylin–eosin (HE) to identify the injury epicenter. Tissuesections displaying the largest proportion of cystic cavity comparedwith total cross-sectional area were taken to represent the focal pointof the injury epicenters.

ImmunohistochemistryThe following primary antibodies were used: mouse anti-NeuN

(1:500; Chemicon International, Inc., Temecula, CA) for neurons,mouse anti-GFAP (1:500; Chemicon International, Inc., Temecula, CA)

Fig. 2. Transduction efficiency of Ad-DsRed in spinal cord. (A) Photomicrographs showing acompression injury and Ad-DsRed injection. Ds-Red signal was detected in the gray matteneurons (NeuN), astrocytes (GFAP) and oligodendrocytes (CC1). Scale bar 300 μm for A; 20

24

for astrocytes, mouse anti-APC (CC1, 1:100; Calbiochem, San Diego,CA) for oligodendrocytes, mouse anti-RECA-1 (1:25; Serotec Inc.,Raleigh, NC) for endothelial cells. The sections were rinsed three timesin PBS after primary antibody incubation and incubated with eitherfluorescent Alexa 568, 647 or 488 goat anti-mouse/rabbit secondaryantibody (1:400; Invitrogen, Burlington, Canada) for 1 h. The sectionswere rinsed three times with PBS and coverslipped with Mowiolmounting medium containing DAPI (Vector Laboratories, Inc.,Burlingame, CA) to counterstain the nuclei. The images were takenusing a Zeiss 510 laser confocal microscope.

Quantification of blood vesselsSections were used for immunofluorescence studies with a

monoclonal antibody specific for a rat endothelial cell antibody,RECA-1. As shown in Fig. 4A, the counting of vessels was performed on4 selected fields (ventral horn, dorsal horn, left and right lateralcolumns) each section under 25× magnification (0.14 mm2). Thenumber of vessels was calculated on two sections (one rostral and onecaudal) at 2 mm and two sections at 4 mm away from the epicenterfor each animal. Labeling and quantification of apoptotic cells. An insitu terminal-deoxy-transferase mediated dUTP nick end-labeling(TUNEL) apoptosis kit (Chemicon International, Inc., Temecula, CA)was used to label apoptotic cells as described in the manufacturer'sinstructions. The numbers of TUNEL-positive nuclei were counted atthe epicenter, as well as at 1, 2 and 3mm (rostral and caudal) from theinjury epicenter. Apoptotic nuclei were found randomly distributed

transverse section of rat spinal cord obtained adjacent to the injury site 10 days afterr and white matter. (B) Confocal images show that the Ad-DsRed vectors transfectedμm for B.

2

387Y. Liu et al. / Neurobiology of Disease 37 (2010) 384–393

throughout the cord, and the count was performed on 4 randomlyselected fields (0.14 mm2) from each section. In order to ensure thatthe selected field included only spinal cord tissue and not the injurycavity, the field was chosen in the anteriolateral white matter andincluded some anterior gray matter. This position was maintainedeven in sections where the cavitation was less pronounced in order tosample consistently.

Assessment of tissue sparing and cavity formation at the injury siteTissue sparing and cavity formation was analyzed 6 weeks after

SCI, at the center of the lesion, 2 mm above and 2 mm below theepicenter. Sections were stained with LFB-HE. The measurementswere carried out on coded slides using ImageJ software (MediaCybernetics Inc., MD). Cross-sectional residual tissue and cavity areaswere normalized with respect to total cross-sectional area and theareas were calculated every 500 μm within the rostrocaudalboundaries of the injury site.

Behavioural testing

Locomotor recovery of the animals was assessed by twoindependent observers using the 21 point Basso, Beattie, andBresnahan (BBB) open field locomotor score (Basso et al., 1995)from 1 to 6 weeks after SCI. The BBB scale was used to assess hindlimb

Fig. 3. Transduction efficiency of AAV-GFP and ZFP-VEGF gene transfer evaluation. (A) AAV2-after SCI and vector injection. Scale bar is 350 μm. (B) AAV2-GFP vectors were predominantlrabbit polyclonal antibody recognizes the p65 activation domain in the ZFP-VEGF treated anipresent in both the control and treatment groups. The lower band (arrow) corresponds to tpositive control for NFκB p65 in HEK293 cells. Lane 1 shows ZFP-VEGF transduced HEK293marker.

243

locomotor recovery including joint movements, stepping ability,coordination, and trunk stability. Testing was done every week on ablinded basis and the duration of each session was 4 min per rat.Scores were averaged across both the right and left hindlimbs toarrive at a final motor recovery score for each week of testing.

Statistical analysis

Data were analyzed with Sigma Stat software. For comparison ofgroups over time (BBB behavioural testing) or distance (tissuesparing), we used two-way analysis of variance (ANOVA) withrepeated measures, followed by the post-hoc Bonferroni test. Forcomparison of simple effects, Student's t test was used. In all figures,the mean value±SEM are used to describe the results. Statisticalsignificance was accepted for p values ofb0.05.

Results

Delivery of the ZFP-VEGF to the spinal cord

To evaluate the transduction efficiency of the Ad or AAV vectors invivo, Ad-DsRed or AAV-GFP were injected into animals with SCI. TheAd-DsRedfluorescent signalwas detected in the gray andwhitematterof the injured spinal cord 10 days after SCI (Fig. 2A). Furthermore, as

GFP-positive cells were still observed andmostly detected in the gray matter at 6 weeksy expressed in neurons (NeuN). Scale bar=20 μm. (C) Western blot showing NFκB p65mals. The higher molecular weight band (upper band) is from endogenous p65 and washe ZFP-VEGF and was only present in the treated animals. (D) Western blot showing acells, lane 2 displays non-transduced HEK293 cells and lane 3 is a molecular weight

Fig. 4. Ad vector mediated ZFP-VEGF treatment increased VEGF mRNA and proteinexpression at 3 days after vector injection. (A) VEGF mRNA levels encoding forVEGF120, 164 and 188 isoforms were measured by real-time PCR. The bar graphillustrates that administration of ZFP-VEGF resulted in an increase of VEGF mRNAcompared with DsRed control group in non-SCI and SCI groups. Relative mRNA levelsare expressed as the mean±SEM, n=12 in non-SCI groups and n=11 in SCI groups;⁎⁎pb0.01, ⁎pb0.05. (B) Western blot showing administration of ZFP-VEGF resulted inincreased VEGF-A protein levels, and (C) VEGF 42 kDa protein significantly increased intreated animals compared with control group. Optical density (OD) of VEGF-A wasnormalized to actin. Data are presented as mean±SEM, n=6/group.

388 Y. Liu et al. / Neurobiology of Disease 37 (2010) 384–393

seen in Fig. 2B, confocal images show that the Ad-DsRed vectortransfected neurons, astrocytes and oligodendrocytes. The biodistri-bution of Ad-DsRed was found in 21.6±3.3 % (mean±SEM) ofneurons, 36.3±3.4% of astrocytes and 20.2±2.3% of oligodendrocytesin the regions of the cord where DsRed expression was observed(n=3). Examination of the immune response elicited by the Advectors was achieved by the immunocytochemical visualization ofmacrophages/microglia and T cells. There was no difference in theextent of inflammation seen in the plain injured and Ad-ZFP-VEGF/Ad-DsRed group with cyclosporin A administration (data notshown).

Following AAV (serotype 2) delivery, GFP-positive cells were stillseen at 6 weeks after SCI and injection. GFP signal was detectedmainly in the gray matter (Fig. 3A), and AAV vectors appeared topreferentially transfect neurons (Fig. 3B). Quantification with anti-NeuN labeling revealed that 37.1±4.9 % (mean±SEM) neurons weretransfected in the regions of the cord where GFP expression wasobserved (n=3). Because the ZFP-VEGF viral construct contains thep65 subunit of the human NFκb transcription factor as the activationdomain (Price et al., 2006), delivery of the ZFP-VEGF was confirmedby immunoblotting using an NFκb p65 antibody for the presence ofthe transcription factor (Fig. 3C). As a positive control, HEK293 cellswere transduced with ZFP-VEGF and cell lysates were processed forimmunoblotting using the same NFκb p65 antibody. The results areshown in Fig. 3D. These results demonstrate localized gene transfer tothe injured spinal cord.

ZFP-VEGF treatment increases VEGF mRNA and protein expression

Animals were sacrificed 3 days after injection, and mRNAexpression of three abundant isoforms, VEGF120, VEGF164 andVEGF188, were measured by quantitative real-time PCR. As shownin Fig. 4A, administration of ZFP-VEGF resulted in a significantincrease in mRNA of VEGF-A splice variants, VEGF120 and VEGF164,both in non-SCI and SCI groups as compared with Ad-DsRed controlanimals. VEGF-A protein expression was assessed by Western blotusing a polyclonal anti-VEGF antibody, which detects a double bandat 42 kDa and 21 kDa and is recommended for the detection of the189, 165 and 121 amino acid splice variants of VEGF. The VEGFprotein (42 kDa) was significantly increased by 2.2-fold in ZFP-VEGFtreated animals versus DsRed treated controls (Fig. 4B and C). Theseresults suggest that, in agreement with earlier studies in skeletalmuscle (Dai et al., 2004; Rebar et al., 2002; Yu et al., 2006), ZFP-VEGF increases VEGF mRNA and protein expression in the injuredspinal cord of rats.

ZFP-VEGF promotes an angiogenic response after traumatic SCI

To quantify the angiogenic response to ZFP-VEGF, we performedimmunostaining with RECA-1, a monoclonal antibody specific forendothelial cells, at 10 days (Ad vector) and 6 weeks (AAV vector)following SCI, in animals that received immediate vector inocula-tion. At the injury epicenter, spinal cord tissue was significantlydisrupted and angiogenesis was unable to be properly quantified.Therefore, capillary density in spinal cord tissue sections wasquantified 2 mm and 4 mm, both caudal and rostral, from thelesion epicenter (Fig. 5A). ZFP-VEGF treated rats displayed a clearincrease in vascular density in the lesion penumbra when comparedwith control rats both in Ad vector (Fig. 5B) and AAV vector(Fig. 5C) application. A statistically significant increase in the numberof blood vessels was observed in the Ad-ZFP-VEGF treatment groupcompared with the Ad-DsRed control group in non-SCI groups, andSCI animals at 2 mm (38.88±2.59 versus 24.58±1.53, pb0.01) and4 mm (44.50±3.15 versus 27.13±2.11, pb0.01) distal to the injurysite (Fig. 5D). As shown in Fig. 5E, analysis of capillary density at6 weeks after injury by AAV injection yielded similar results in non-

24

SCI group and SCI rats at 4 mm away from the epicenter (AAV-ZFP-VEGF=51.58±2.16, AAV-GFP control=40.25±2.07, pb0.01).

ZFP-VEGF treatment attenuates axonal degradation and post-traumaticapoptosis

To assess the neuroprotective effects of ZFP-VEGF after SCI, thelevel of degradation of neurofilament protein (NF200), a hallmark ofneurodegeneration in the SCI model, was assessed in the injuredregion of the cord. Previous publications from our laboratory havedemonstrated a progressive loss of NF200 after SCI (Karimi-Abdolrezaee et al., 2004; Schumacher et al., 2000). NF200 degradationwas reduced at 3 and 7 days post-injury with Ad-ZFP-VEGF treatment(Fig. 6A), and axonal preservation was observed at 6 weeks afterinjury with AAV-ZFP-VEGF treatment (Fig. 6B). As shown in Fig. 6C,the NF200 content was significantly (pb0.05) increased by 2-fold at7 days and by 2.5-fold at 6 weeks post-injury in ZFP-VEGF treatedanimals versus control animals.

4

Fig. 5. Angiogenic response of VEGF-A with an increase in capillary density. (A) Diagram of experimental parameter illustrates serial injured spinal cord transverse sections. Rightpanel illustrates the area of the cord used for RECA-1 counting. Panels (B) and (C) are representative sections taken 4mm rostral to the epicenter from a ZFP-VEGF treated and controlanimal respectively immunostained with RECA-1 at (B) 10 days (Ad vector mediated) and (C) 6 weeks (AAV vector mediated) after SCI. Top panels: low-power representation ofcord stained with RECA-1; scale 200 μm. The boxed areas in top panels correspond to the enlarged areas in lower panels (bar=50 μm). An increased angiogenic response wasobserved in the ZFP-VEGF treated group. (D) Bar graph depicting the RECA-1-positive cell counts 10 days after SCI by Ad vector mediated treatment. The ZFP-VEGF administration ledto a significant increase in capillary density in non-SCI and SCI animals (2 mm and 4 mm away from the epicenter) as compared with the control group. (E) Bar graph showing theRECA-1-positive cell counts 6 weeks after SCI by AAV vector mediated treatment. The ZFP-VEGF administration led to a significant increase in capillary density in non-SCI group, andSCI animals at 4 mm away from the injury site as compared with the control group. Data are presented as mean±SEM, n=5/group; ⁎⁎pb0.01, ⁎pb0.05.

389Y. Liu et al. / Neurobiology of Disease 37 (2010) 384–393

Previous studies from our group have shown that apoptotic celldeath occurs as early as 6 h following SCI, peaks at seven days and isstill evident at 14 days post injury (Casha et al., 2001). To determinewith the effects of ZFP-VEGF treatment on apoptotic cell death, in situ

245

terminal-deoxy-transferase mediated dUTP nick end-labeling(TUNEL) staining was performed 10 days after injury. TUNEL-positivecells were found throughout the gray and white matter in the injuredspinal cord, with the greatest concentration close to the lesion site.

Fig. 6. ZFP-VEGF administration attenuated axonal degradation. (A) Western blotdemonstrating that administration of Ad-ZFP-VEGF resulted in the attenuation ofNF200 degradation at 3 and 7 days after injury. (B) Administration of AAV-ZFP-VEGFresulted in the attenuation of NF200 degradation at 6 weeks after injury. (C) RelativeOD values of control versus ZFP-VEGF treated animals. Inhibition of NF200 degradationwas significantly different in control versus the ZFP-VEGF treatment groups at 7 daysand 6 weeks after injury. Optical density of NF200 was normalized to actin. Error barsare expressed as standard error of means, n=6/group at 3 and 7 days, and n=6 inAAV-ZFP-VEGF treatment group, n=5 in AAV-GFP control group at 6 weeks; ⁎pb0.05.

Fig. 7. Ad-ZFP-VEGF administration reduced post-traumatic apoptosis after SCI. Panels(A) and (B) are representative sections taken 2 mm rostral to the epicenter from a ZFP-VEGF treated and Ad-DsRed control animal respectively immunostained with TUNEL at10 days after SCI. Top panels: low-power representation of cord stained with TUNEL ingreen and nuclear marker DAPI in blue; scale 200 μm. The boxed areas in the top panelscorrespond to the enlarged areas in the lower panels (bar=50 μm). A reduction ofTUNEL-positive cells was observed in the ZFP-VEGF treated group (C) Bar graphshowing the TUNEL cell counts 10 days after SCI. There was a significant decrease inTUNEL-positive cell death in the ZFP-VEGF treatment group versus control group at1 mm, 2 mm and 3 mm away (rostral and caudal) from the lesion epicenter. Values aremean±SEM, n=5/group; ⁎⁎pb0.01.

390 Y. Liu et al. / Neurobiology of Disease 37 (2010) 384–393

TUNEL-stained nuclei were counted at the injury epicenter, and at 1, 2,and 3 mm from the injury epicenter. As shown in Fig. 7, ZFP-VEGFtreatment was associated with a significant reduction in counts ofTUNEL-positive cells at 1 mm (pb0.01), 2 mm (pb0.01) and 3 mm(pb0.01) from the lesion epicenter. The reduction in apoptotic celldeath at the injury epicenter approached (p=0.09) but did not attainsignificance.

ZFP-VEGF enhances tissue preservation at the lesion site

Six weeks after SCI, spinal cord cross-sections were stainedserially with LFB-HE. Spinal cords from AAV-ZFP-VEGF treated ratsexhibited a greater extent of spared tissue and decreased cavityformation in all sections, up to 2 mm rostral and caudal to theinjury epicenter when compared to tissue sections from AAV-GFPcontrol rats (Fig. 8A). Measurements of residual tissue or cavity sizetaken from cross-sectional areas were expressed as a percent of thetotal cross-section area of the section. A comparison of percentnormalized residual tissue, which is represented in Fig. 8B, wasperformed by two-way ANOVA with the two factors representingtreatment and distance from the injury epicenter. This experimentrevealed an overall significant improvement in tissue preservationin the ZFP-VEGF treated group (p=0.005). The percentage ofremaining tissue was 72.55±4.13% in ZFP-VEGF treated animalsand 53.73±4.41% in control rats at injury epicenter (p=0.002).There was also significantly increased preserved neural tissue inZFP-VEGF treated animals at 1.0 (p=0.007) and 0.5 mm (p=0.036)

24

rostral and at 0.5 mm (p=0.011) caudal to the epicenter. As shownin Fig. 8C, the cavity area was significantly decreased in ZFP-VEGF-treated rats at 1.5 and 1.0 mm rostral and epicenter.

ZFP-VEGF promotes functional neurobehavioural recovery after SCI

The hind limbs of experimental animals were completelyparalyzed after SCI. Hind-limb performance gradually improved inboth experimental groups. However, at 3 weeks and thereafter, theAAV-ZFP-VEGF treated rats displayed significant behavioural im-provement compared with control group (Fig. 9). At 6 weeks, theAAV-GFP control rats reached an average score of 6, indicatingnon-functional movement of the three hind-limb joints withoutcoordinated sweeping of their hind legs. In contrast, the ZFP-VEGFtreated rats reached an average score of 8, indicating the ability to

6

Fig. 8. AAV-ZFP-VEGF improves spinal cord tissue preservation and decreases cavityformation. Panel (A) shows representative sections taken 0.5 mm rostral to theepicenter from a ZFP-VEGF treated and AAV-GFP control animal, respectively, stainedwith LFB and HE (scale 200 μm) 6 weeks after SCI. ZFP-VEGF treated spinal cordexhibited a larger extent of spared tissue and decreased cavity formation than controlanimal tissue. (B) Percent normalized residual tissue and (C) cavity area in ZFP-VEGFtreatment (■) versus control (◊) group. There was a significant different between ZFP-VEGF treated animals versus controls by two-way ANOVA with post-hoc (Bonferroni)test. Single (pb0.05) and double (pb0.01) asterisks indicate significantly increasedtissue area or decreased cavity formation in ZFP-VEGF treated animals at different sitesfrom injured epicenter. Data are mean±SEM (bars) values (n=12 in AAV-ZFP-VEGFtreatment group, and n=11 in AAV-GFP control group).

Fig. 9. The graph depicts functional hindlimb recovery over time after SCI. Treatmentwith ZFP-VEGF resulted in improved BBB scores versus controls. The two groups differfrom each other at pb0.02 by two-way ANOVA with repeated measures. ⁎Bonferronipost-hoc tests showed that scores differed at pb0.04 at every time point 3 weeksafter SCI and thereafter. Data are mean±SEM (bars) values (n=12 in AAV-ZFP-VEGFtreatment group, and n=11 in AAV-GFP control group).

391Y. Liu et al. / Neurobiology of Disease 37 (2010) 384–393

make sweeping movements with their hind legs or coordinatedplantar placement of the hind limbs without weight support.

Discussion

The results of this study indicate that ZFP-VEGF treatment leadsto an angiogenic response and has neuroprotective effects in theinjured spinal cord of rodents. These effects include an attenuationof axonal degradation, reduced post-traumatic apoptosis, enhancedtissue sparing at the lesion site, and improved neurobehaviouraloutcomes. Hence, the ZFP-VEGF strategy as described here holdspromise as a potential therapy for SCI and other forms of CNSinjury.

247

Disruption of the vasculature and ischemia are key elements of SCI

SCI results in the disruption of spinal cord blood flow and theonset of spinal cord ischemia. Approaches that address the onsetand downstream consequence of ischemic injury are thereforeattractive treatment options for SCI. VEGF-A is angiogenic, and hasbeen recognized as an important signaling molecule in the nervoussystem. We examined the effects of a ZFP transcription factordesigned to increase the expression of all major VEGF-A isoforms ina well-characterized clip compression model of SCI. Previous datahave demonstrated that delivery of ZFP-VEGF can increase endog-enous VEGF-A expression in striated skeletal muscle and promoteangiogenesis in both a mouse ear and rabbit hind-limb ischemiamodels (Dai et al., 2004; Price et al., 2006; Rebar et al., 2002; Yu etal., 2006). More recent findings have demonstrated that intramus-cular delivery of ZFP-VEGF was able to improve perfusion, limittissue apoptosis, and promote angiogenesis after hind-limb ischemiain ApoE knockout mice fed a high-cholesterol diet (Xie et al., 2006).Similarly, gene transfer with a transcription factor designed toincrease VEGF-A expression improved recovery in an ischemicmouse limb (Li et al., 2007). The VEGF-A isoforms examined in ourreport were upregulated at the mRNA and protein levels in SCI afterZFP-VEGF treatment. The significant increase in VEGF120 andVEGF164 is particularly noteworthy. In the human CNS, theVEGF121 and VEGF165 isoforms constitute the majority of VEGF-Aexpression and appear to be the major players in the process ofangiogenesis in the spinal cord (Keyt et al., 1996).

VEGF-A induces angiogenesis and promotes neuroprotection

Recent publications have also examined the role of VEGF-A inmodels of SCI, with varying results. Choi et al. (2007) used a hypoxia-inducible VEGF-A expression system to treat rats with SCI andobserved neuroprotective effects and enhanced VEGF-A expression.Another group used an adenovirus coding for VEGF165, delivered viamatrigel, in a partial spinal cord transection model. They observed asignificant increase in vessel volume and a reduction in the retrogradedegeneration of corticospinal tract axons (Facchiano et al., 2002).

However, Benton and Whittemore (2003) reported an exacerba-tion of lesion size and increased inflammation after the delivery of2 μg of recombinant VEGF165 directly into the contused spinal cord3 days post-SCI. This study highlights several factors which are likelyto be critical in the successful application of VEGF-A as a therapy forSCI. Themethod of VEGF delivery is likely a critical factor. In our study,we injected the ZFP-VEGF adjacent to the injury epicenter as the

392 Y. Liu et al. / Neurobiology of Disease 37 (2010) 384–393

perilesional ischemic penumbra is likely the zone, which wouldbenefit themost from approaches to enhance angiogenesis. Moreover,Ad and AAV viral vectors were used in the acute and chronic studiesaccording to their different expression characteristics. The use ofcyclosporine A as an immunosuppressive agent following Ad vector isalso an important factor which merits commentary. Of note, it hasbeen reported that cyclosporine A has modest neuroprotective effectsin models of contusive SCI (McMahon et al., 2009). Importantly, theeffects of cyclosporine A were controlled for in our experiments.Control rats who received Ad-DsRed also received cyclosporine A.Thus any potential beneficial effects of cyclosporine A were accountedfor in the design of our experimental plan and would not havecontributed to the differential neuroprotective effects seen with ZFP-VEGF constructs.

The administration of ZFP-VEGF has the ability to upregulateseveral isoforms of VEGF-A, mimicking endogenous expression.WhenVEGF therapy is applied, the potent vascular permeability effect of thistreatment must also be considered. The study from Rebar et al. (2002)have shown that the neovasculature resulting from activation of theendogenous VEGF-A by engineered ZFPs is not hyperpermeable, incomparison with the vasculature induced by VEGF-A164 alone. In ourexperiments, ZFP-VEGF treatment resulted in increased vessel densityin both SCI and non-SCI animals. However, we observed no adverseeffects of ZFP-VEGF in uninjured animals, as assessed by histomor-phological and neurobehavioural outcomes (see SupplementaryFigure). It is possible that the ZFP-induced vasculature is morephysiologically mature and that this is due to the induced expressionof the natural VEGF-A splice variants. There was a slight, transientreduction in locomotor performance in non-SCI animals due to theintraspinal microinjection. In future studies we aim to investigatealternative (and potentially less invasive) methods for administrationof ZFP-VEGF following SCI, which may result in a more clinicallyrelevant delivery system. The timing of VEGF-A therapy could play animportant role in mediating the optimal effects, although this issuerequires further study. Based on our current data and other positiveresults of VEGF-A treatment in SCI and brain ischemic injury/stroke, itwould appear that VEGF-A has in the potential to promote recoveryfrom spinal cord trauma (Kaya et al., 2005; Wang et al., 2006).

Possible neuroprotective mechanisms of VEGF-A treatment

Recent studies indicate that VEGF-A can no longer be characterizedsolely as an endothelial mitogen. It is increasingly apparent that VEGF-A may exert multiple roles in the CNS that contribute to neurotrophicand neuroprotective effects. By stimulating angiogenesis with VEGF,we hypothesized that ischemia, and its contribution to secondarydamage of the spinal cord, could be counteracted. Indeed, weobserved higher blood vessel densities and increased axonal preser-vation. The differences between treatments correlated well tobehavioural outcomes and the amount of spared tissue. However,the mechanisms underlying the neuroprotective effects of VEGF-A inthe setting of SCI are not well understood. A number of in vitro and invivo studies have demonstrated that VEGF-A is a potent neurotrophicfactor which also confers protection to injured neurons (Jin et al.,2000; Rosenstein and Krum, 2004). Some studies have suggested thatthis neuroprotective effect is mediated by activation of the intracel-lular tyrosine kinase domains that influence several downstreamsignaling pathways, including mitogen-activated protein kinases(MAPK) and phosphoinositide 3-kinase (PI3K)/Akt (Kaya et al.,2005; Sondell et al., 1999; Svensson et al., 2002). It has also beenproposed that the binding of VEGF-A to the VEGFR-2 receptor onneurons stimulates dimerization of VEGFR-2 forming a complex withneuropilin-1, thus activating the PI3-K/Akt and the extracellular-regulated kinase (ERK-1/-2) signaling pathways. The PI3-K/Aktpathway is an important regulator of cell proliferation and survivaland has been shown to mediate the anti-apoptotic effects of VEGF in

24

endothelial cells (Larrivee and Karsan, 2000; Thakker et al., 1999), aswell as to transduce neuroprotective effects in an immortalizedneuronal cell line (Jin et al., 2000). Thus, VEGF-A could play asignificant role in spinal cord repair and is emerging as a central playerin neurodegenerative disorders (Storkebaum and Carmeliet, 2004;Storkebaum et al., 2004) with potent direct neuroprotective (Marti,2002; Sun et al., 2003) and neurotrophic (Rosenstein and Krum, 2004;Sondell et al., 1999) functions.

A review secondary injury theory by Tator and Fehlings (1991)suggested that petechial hemorrhages and edema occur within15 min of injury. During the first 4 h, myelin sheaths are disrupted,axonal degeneration can be observed, and ischemic endothelial injuryoccurs. We injected ZFP-VEGF immediately following SCI in order tostem the initial cascade of secondary injury, but ZFP-VEGF adminis-tration immediately following injury is not clinically feasible. Thetiming of gene production is an important issue and the perifocalregion may have substantially wider timeframes for rescue. It hasbeen reported that delayed (48 h after ischemia) administration ofrecombinant human VEGF165 to ischemic rats enhanced angiogenesisin the penumbra and significantly improved neurological recovery(Sun et al., 2003; Zhang et al., 2000). While a future study willdetermine the effective time window for ZFP-VEGF administrationfollowing SCI, the present study demonstrates that the administrationof ZFP-VEGF immediately following SCI exerts potent neuroprotectiveeffects. To the best of our knowledge, this report is the first toinvestigate the therapeutic efficiency of ZFP-VEGF in a SCI model. Themajor findings of our study suggest this approach warrants investi-gation as a novel approach to treat SCI.

Acknowledgments

The authors would like to thank Wendy Zhang for assistance withthe histochemistry experiments, Jian Wang and Behzad Azad for theirhelp with behavioural testing, Julio Furlan for assistance with dataanalysis and Allyson Tighe for her editorial reviews. This study wassupported by Sangamo BioSciences and the Krembil Chair in NeuralRepair and Regeneration (held by Dr. Michael G. Fehlings). Theauthors thank Philip Gregory and Edward Rebar of SangamoBioSciences for scientific review.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.nbd.2009.10.018.

References

Adris, S., et al., 2005. Quantification of vascular endothelial growth factor andneuropilins mRNAs during rat brain maturation by real-time PCR. Cell. Mol.Neurobiol. 25, 1035–1041.

Basso, D.M., et al., 1995. A sensitive and reliable locomotor rating scale for open fieldtesting in rats. J. Neurotrauma 12, 1–21.

Baumgartner, I., et al., 2000. Lower-extremity edema associated with gene transfer ofnaked DNA encoding vascular endothelial growth factor. Ann. Intern. Med. 132,880–884.

Benton, R.L., Whittemore, S.R., 2003. VEGF165 therapy exacerbates secondary damagefollowing spinal cord injury. Neurochem. Res. 28, 1693–1703.

Burger, C., et al., 2004. Recombinant AAV viral vectors pseudotyped with viral capsidsfrom serotypes 1, 2, and 5 display differential efficiency and cell tropism afterdelivery to different regions of the central nervous system. Molec. Ther. 10,302–317.

Bustin, S.A., Nolan, T., 2004. Pitfalls of quantitative real-time reverse-transcriptionpolymerase chain reaction. J. Biomol. Tech. 15, 155–166.

Casha, S., et al., 2001. Oligodendroglial apoptosis occurs along degenerating axons andis associated with FAS and p75 expression following spinal cord injury in the rat.Neuroscience 103, 203–218.

Choi, U.H., et al., 2007. Hypoxia-inducible expression of vascular endothelial growthfactor for the treatment of spinal cord injury in a rat model. J. Neurosurg. Spine 7,54–60.

Dai, Q., et al., 2004. Engineered zinc finger-activating vascular endothelial growth factortranscription factor plasmid DNA induces therapeutic angiogenesis in rabbits withhindlimb ischemia. Circulation 110, 2467–2475.

8

393Y. Liu et al. / Neurobiology of Disease 37 (2010) 384–393

Facchiano, F., et al., 2002. Promotion of regeneration of corticospinal tract axons in ratswith recombinant vascular endothelial growth factor alone and combined withadenovirus coding for this factor. J. Neurosurg. 97, 161–168.

Fehlings, M.G., Tator, C.H., 1995. The relationships among the severity of spinal cordinjury, residual neurological function, axon counts, and counts of retrogradelylabeled neurons after experimental spinal cord injury. Exp. Neurol. 132, 220–228.

Fehlings, M.G., et al., 1989. The relationships among the severity of spinal cord injury,motor and somatosensory evoked potentials and spinal cord blood flow.Electroencephalogr. Clin. Neurophysiol. 74, 241–259.

Greenberg, D.A., Jin, K., 2005. From angiogenesis to neuropathology. Nature 438,954–959.

Hermens, W.T., et al., 1997. Transient gene transfer to neurons and glia: analysis ofadenoviral vector performance in the CNS and PNS. J. Neurosci. Methods 71, 85–98.

Jin, K.L., et al., 2000. Vascular endothelial growth factor: direct neuroprotective effect inin vitro ischemia. Proc. Natl. Acad. Sci. U. S. A. 97, 10242–10247.

Karimi-Abdolrezaee, S., et al., 2004. Temporal and spatial patterns of Kv1.1 and Kv1.2protein and gene expression in spinal cord white matter after acute and chronicspinal cord injury in rats: implications for axonal pathophysiology afterneurotrauma. Eur. J. Neurosci. 19, 577–589.

Kaya, D., et al., 2005. VEGF protects brain against focal ischemia without increasingblood-brain permeability when administered intracerebroventricularly. J. Cereb.Blood Flow Metab. 25, 1111–1118.

Keyt, B.A., et al., 1996. The carboxyl-terminal domain (111–165) of vascular endothelialgrowth factor is critical for its mitogenic potency. J. Biol. Chem. 271, 7788–7795.

Larrivee, B., Karsan, A., 2000. Signaling pathways induced by vascular endothelialgrowth factor (review). Int. J. Mol. Med. 5, 447–456.

Leung, D.W., et al., 1989. Vascular endothelial growth factor is a secreted angiogenicmitogen. Science 246, 1306–1309.

Li, Y., et al., 2007. In mice with type 2 diabetes, a vascular endothelial growth factor(VEGF)-activating transcription factor modulates VEGF signaling and inducestherapeutic angiogenesis after hindlimb ischemia. Diabetes 56, 656–665.

Liu, P.Q., et al., 2001. Regulation of an endogenous locus using a panel of designed zincfinger proteins targeted to accessible chromatin regions. Activation of vascularendothelial growth factor A. J. Biol. Chem. 276, 11323–11334.

Marti, H.H., 2002. Vascular endothelial growth factor. Adv. Exp. Med. Biol. 513,375–394.

McMahon, S.S., et al., 2009. Effect of cyclosporin A on functional recovery in the spinalcord following contusion injury. J. Anat. 215, 267–279.

Price, S.A., et al., 2006. Gene transfer of an engineered transcription factor promotingexpression of VEGF-A protects against experimental diabetic neuropathy. Diabetes55, 1847–1854.

Rajagopalan, S., et al., 2003. Regional angiogenesis with vascular endothelial growthfactor in peripheral arterial disease: a phase II randomized, double-blind,

249

controlled study of adenoviral delivery of vascular endothelial growth factor121 in patients with disabling intermittent claudication. Circulation 108,1933–1938.

Rebar, E.J., et al., 2002. Induction of angiogenesis in a mouse model using engineeredtranscription factors. Nat. Med. 8, 1427–1432.

Rosenstein, J.M., Krum, J.M., 2004. New roles for VEGF in nervous tissue-beyond bloodvessels. Exp. Neurol. 187, 246–253.

Schumacher, P.A., et al., 2000. Pretreatment with calpain inhibitor CEP-4143 inhibitscalpain I activation and cytoskeletal degradation, improves neurological function,and enhances axonal survival after traumatic spinal cord injury. J. Neurochem. 74,1646–1655.

Shweiki, D., et al., 1992. Vascular endothelial growth factor induced by hypoxia maymediate hypoxia-initiated angiogenesis. Nature 359, 843–845.

Sondell, M., et al., 1999. Vascular endothelial growth factor has neurotrophic activityand stimulates axonal outgrowth, enhancing cell survival and Schwann cellproliferation in the peripheral nervous system. J. Neurosci. 19, 5731–5740.

Storkebaum, E., Carmeliet, P., 2004. VEGF: a critical player in neurodegeneration. J. Clin.Invest. 113, 14–18.

Storkebaum, E., et al., 2004. VEGF: once regarded as a specific angiogenic factor, nowimplicated in neuroprotection. BioEssays 26, 943–954.

Sun, Y., et al., 2003. VEGF-induced neuroprotection, neurogenesis, and angiogenesisafter focal cerebral ischemia. J. Clin. Invest. 111, 1843–1851.

Svensson, B., et al., 2002. Vascular endothelial growth factor protects cultured rathippocampal neurons against hypoxic injury via an antiexcitotoxic, caspase-independent mechanism. J. Cereb. Blood Flow Metab. 22, 1170–1175.

Tator, C.H., Fehlings, M.G., 1991. Review of the secondary injury theory of acute spinalcord trauma with emphasis on vascular mechanisms. J. Neurosurg. 75, 15–26.

Thakker, G.D., et al., 1999. The role of phosphatidylinositol 3-kinase in vascularendothelial growth factor signaling. J. Biol. Chem. 274, 10002–10007.

Wang, Y., et al., 2006. Vascular endothelial growth factor improves recovery ofsensorimotor and cognitive deficits after focal cerebral ischemia in the rat. BrainRes. 1115, 186–193.

Xie, D., et al., 2006. An engineered vascular endothelial growth factor-activatingtranscription factor induces therapeutic angiogenesis in ApoE knockout mice withhindlimb ischemia. J. Vasc. Surg. 44, 166–175.

Yu, J., et al., 2006. An engineered VEGF-activating zinc finger protein transcriptionfactor improves blood flow and limb salvage in advanced-age mice. FASEB J. 20,479–481.

Zhang, Z.G., et al., 2000. VEGF enhances angiogenesis and promotes blood-brain barrierleakage in the ischemic brain. J. Clin. Invest. 106, 829–838.

Zhang, L., et al., 2002. Different effects of glucose starvation on expression and stabilityof VEGF mRNA isoforms in murine ovarian cancer cells. Biochem. Biophys. Res.Commun. 292, 860–868.

A Spinal Cord Window Chamber Model for In VivoLongitudinal Multimodal Optical and Acoustic Imagingin a Murine ModelSarah A. Figley1,2., Yonghong Chen3., AzusaMaeda4, Leigh Conroy4, Jesse D. McMullen3, Jason I. Silver3,

Shawn Stapleton4, Alex Vitkin3,4,5, Patricia Lindsay5, Kelly Burrell2, Gelareh Zadeh2,

Michael G. Fehlings1,2, Ralph S. DaCosta3,4,5*

1 Institute of Medical Science, University of Toronto, Toronto, Ontario, Canada, 2 Toronto Western Research Institute, Krembil Neuroscience Program, University Health

Network, Toronto, Ontario, Canada, 3Ontario Cancer Institute, University Health Network, Princess Margaret Hospital, Toronto, Ontario, Canada, 4Department of Medical

Biophysics, University of Toronto, Toronto, Ontario, Canada, 5Department of Radiation Physics, University Health Network, Princess Margaret Hospital, Toronto, Ontario,

Canada

Abstract

In vivo and direct imaging of the murine spinal cord and its vasculature using multimodal (optical and acoustic) imagingtechniques could significantly advance preclinical studies of the spinal cord. Such intrinsically high resolution andcomplementary imaging technologies could provide a powerful means of quantitatively monitoring changes in anatomy,structure, physiology and function of the living cord over time after traumatic injury, onset of disease, or therapeuticintervention. However, longitudinal in vivo imaging of the intact spinal cord in rodent models has been challenging,requiring repeated surgeries to expose the cord for imaging or sacrifice of animals at various time points for ex vivo tissueanalysis. To address these limitations, we have developed an implantable spinal cord window chamber (SCWC) device andprocedures in mice for repeated multimodal intravital microscopic imaging of the cord and its vasculature in situ. Wepresent methodology for using our SCWC to achieve spatially co-registered optical-acoustic imaging performed serially forup to four weeks, without damaging the cord or induction of locomotor deficits in implanted animals. To demonstrate thefeasibility, we used the SCWC model to study the response of the normal spinal cord vasculature to ionizing radiation overtime using white light and fluorescence microscopy combined with optical coherence tomography (OCT) in vivo. In vivopower Doppler ultrasound and photoacoustics were used to directly visualize the cord and vascular structures and tomeasure hemoglobin oxygen saturation through the complete spinal cord, respectively. The model was also used forintravital imaging of spinal micrometastases resulting from primary brain tumor using fluorescence and bioluminescenceimaging. Our SCWC model overcomes previous in vivo imaging challenges, and our data provide evidence of the broaderutility of hybridized optical-acoustic imaging methods for obtaining multiparametric and rich imaging data sets, includingover extended periods, for preclinical in vivo spinal cord research.

Citation: Figley SA, Chen Y, Maeda A, Conroy L, McMullen JD, et al. (2013) A Spinal Cord Window Chamber Model for In Vivo Longitudinal Multimodal Optical andAcoustic Imaging in a Murine Model. PLoS ONE 8(3): e58081. doi:10.1371/journal.pone.0058081

Editor: Chin-Tu Chen, The University of Chicago, United States of America

Received June 19, 2012; Accepted January 30, 2013; Published March 14, 2013

Copyright: � 2013 Figley et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permitsunrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Funding: This research was funded in part by Cancer Care Ontario (DaCosta holds a CCO Research Chair in Cancer Imaging; https://www.cancercare.on.ca/), theCanadian Institutes of Health Research (awarded to Dr. DaCosta, reference number 93578; http://www.cihr-irsc.gc.ca/e/193.html) and the Ontario Ministry ofHealth and Long Term Care(http://www.health.gov.on.ca/en/). The funders had no role in study design, data collection and analysis, decision to publish, orpreparation of the manuscript.

Competing Interests: The authors have declared that no competing interests exist.

* E-mail: rdacosta@uhnres.utoronto.ca

. These authors contributed equally to this work.

Introduction

Most in vivo imaging of the spinal cord in animals (and humans)

has been conducted using computed tomography (CT), magnetic

resonance imaging (MRI), diffusion tensor imaging (DTI) or

ultrasound imaging [1,2,3,4]. While these non-invasive imaging

techniques allow in vivo serial imaging of the cord in preclinical

models, image resolution is suboptimal for visualizing vital

microscopic anatomical structures, such as the vasculature and

neural tracts. Furthermore, such imaging techniques suffer from

poor tissue specificity, and typically require an exogenous contrast

agent to differentiate vasculature from solid tissue structures.

Alternatively, optical imaging could provide a unique and

powerful method of studying the intact spinal cord and its

vasculature in situ at structural and functional levels longitudinally

and at sub-micrometer resolutions (e.g. at the cellular level).

However, the anatomy and location of the intact spinal cord is

close to the heart and lungs, and therefore results in cord motion

during imaging. Thus, in vivo spinal cord imaging contains

inherent challenges for optical imaging compared to other central

nervous system (CNS) targets, such as the retina or cerebral cortex,

which can be readily accessed using in vivo optically-based imaging

techniques, either directly or via intracranial transparent window

chamber implants, respectively [1,5,6,7]. Moreover, the vascular

structures of the spinal cord are predominantly located in the grey

PLOS ONE | www.plosone.org 1 March 2013 | Volume 8 | Issue 3 | e58081

250

matter, making it difficult to image using traditional microscopy

techniques, such as confocal fluorescence microscopy as they are

unable to penetrate deep enough into the spinal cord tissue to

image the microvasculature of the grey matter [8,9].

To date, a few published reports have emerged on the use of

optical microscopy to visualize the mouse spinal cord in vivo. For

example, Kerschensteiner et al. used in vivo fluorescence imaging to

monitor individual fluorescent axons in the spinal cords of living

transgenic mice over several days after spinal injury [10]. Davalos

et al. used two-photon fluorescence imaging to study multiple

axons, microglia and blood vessels in the mouse spinal cord in vivo

[11]. Johannssen et al. labeled the superficial dorsal horn

populations with a Ca(2+) indicator, and were able to stabilize

the spinal cord sufficiently to permit functional imaging in

anaesthetized mice using two-photon fluorescence Ca(2+) micros-

copy [12]. Again, using two-photon fluorescence microscopy, Kim

et al. studied the migration of GFP(+) immune cells in the spinal

cord of CXCR6(gfp/+) mice during active experimental autoim-

mune encephalomyelitis using an intervertebral window approach

[13]. Dray et al. have successfully followed the dynamics of

degeneration-regeneration of injured spinal cord axons while

simultaneously monitoring the fate of the vascular network in the

same animal up to 4 months post-injury using multiphoton

fluorescence microscopy [14]. Finally, Codotte et al. recently

demonstrated the use of optical coherence tomography (OCT)

for structural and vascular imaging of the mouse spinal cord

without the use of a contrast agent; however, their studies did not

include repeated in vivo imaging [15]. These examples reflect

a major recent trend in spinal cord research to apply established

optical microscopy techniques to study the cord and its vascular

network in situ and over time at high resolution and in vivo.

However, a major drawback of all these approaches has been the

need for repeated surgeries to the vertebral column of the same

animal to expose the cord or, alternatively, animal sacrifice for ex

vivo tissue analysis.

Recently, Farrar et al. reported that they had overcome the

limitation of repeated surgical procedures by using a metal spinal

cord window chamber implanted between T11–T12 of the mouse

vertebral column for repeated optical imaging [16]. Briefly, the

spinal chamber held a glass coverslip in place and provided

continuous optical access to the cord for over five weeks, allowing

quantitative imaging of microglia and afferent axon dynamics after

laser-induced damage to the cord. Fenrich et al. also recently

developed a SCWC model to examine axonal regeneration

following a ‘pin-prick’ model of spinal cord injury [17]. While

these studies provide elegant designs for longitudinal in vivo spinal

imaging, both models utilize metallic components and conduct

multiphoton microscopy for high-resolution image acquisition.

However, metal devices are incompatible with other emerging

optically-enabled imaging techniques which could provide addi-

tional complementary biological information about the cord and,

in particular, its vasculature. For example, photoacoustic imaging

[18], which combines optical excitation and ultrasound detection,

can provide quantitative information about the vasculature

throughout the the full thickness of the cord at imaging depths

unacheivable with mutliphoton fluorescence microscopy. In

addition, multispectral photoacoustics can provide quantitative

information about the oxygenation status of the cord vasculature

[19,20]. Power Doppler ultrasound can also be used to determine

vascular density in vivo. Thus, while the window chamber

approach of Farrar et al. is a significant step forward for in vivo

optical imaging of the mouse spinal cord, it is limited to

mutliphoton fluorescence microscopy. Here, we report the de-

velopment and testing of an alternate design of a transparent

spinal cord window chamber (SCWC) implantable device

composed of either metal or polycarbonate materials for mice

(Figure 1; and rats, See File S1 and Figure S1) that overcomes the

need for repeated spinal surgeries. We demonstrate the feasibility

and utility of our approach to obtain multiparametic (morpho-

logical, structural, functional, and cellular) high-resolution imaging

data of the mouse spinal cord and its vasculature using multiple

complementary imaging techniques (including fluoresence micros-

copy, OCT, power Doppler ultrasound and photoacoustic

imaging) longitudinally and in vivo.

Methods

All animal procedures were conducted with approval from the

University Health Network Animal Care Committee (Animal Use

Protocol #2263 and #2609).

Mouse Window Chamber DesignsThe spinal cord window chamber (SCWC) devices were

modeled and designed using SolidWorksH software (SolidWorks

Corporation, Waltham, MA, USA) (Figure 1A). Window cham-

bers were 3D printed using a Fortus 3D Production printer

systems (Stratasys, Eden Prairie, MN, USA) using ABS-poly-

carbonate (Figure 1F) or machined to the same specifications out

of surgical grade stainless steel (Figure 1G). The metal SCWC

devices were used for X-ray irradiation experiments and sub-

sequent fluorescence and speckle-variance OCT (svOCT) imag-

ing. Since the metal device was thinner, it allowed for closer

contact between the tissue and imaging objective lenses. However,

since metal is incompatible with photoacoustic imaging, poly-

carbonate SCWC devices were installed in animals for the

photoacoustic and power Doppler ultrasound imaging. Both metal

and polycarbonate SCWC designs had eight circumferentially-

located holes for surgical sutures to secure the device to the dorsal

skin. The total weight of SCWC devices were 0.35 g (plastic) and

1.0 g (metal), which were well tolerated by the mice. The SCWC

coverslip had a diameter of 8 mm, permitting use of water-coupled

high-magnification microscope objective lenses for high-resolution

imaging in vivo. To restrain the mouse to the microscope stage

during imaging, four perpendicular extension arms were added to

the device to mechanically screw the animal to the microscope

stage, thus ensuring stability and minimizing movement during

intravital imaging (Figure 1). Standard glass coverslips of 8 mm

diameter (Cat. No. 5DE89, Grainger, Lake Forest, IL, USA) was

used for the mouse SCWC. Coverslips were held in place by

a metal ring clamp (Cat. No. 5DE89, Grainger, Lake Forest, IL,

USA) once the device was implanted in the animal.

Mouse Spinal Cord Window Chamber InstallationFemale athymic nude mice (NCRNU-F, Taconic, Hudson, NY,

USA) or C57BL6 (Jackson Laboratories, Bar Harbor, Maine,

USA) at 15–20 weeks, were anesthetized using a mixture of

ketamine (80 mg/kg) and xylazine (5 mg/kg) prior to surgical

installation of the SCWCs. Briefly, mice were placed in a sterile

surgical preparation area and the dorsal skin was disinfected with

70% isopropyl ethanol and 10% povidone-iodine. A 3–4 cm

incision, using a #15 blade scalpel, along the dorsal midline was

made in the lumbar region to expose the spine (Figure 1B), and

a two-level laminectomy at L2–L3 was performed using fine

scissors (Figure 1C and 1D). After exposing the spinal cord, India

ink (Pelikan, #221143, Hannover, Germany) was carefully applied

to the middle of the spinal cord using the tip of a sterile piece of

tissue paper, approximately the size of a 30 gauge needle. This

served as a landmark for longitudinal imaging, enabling us to

Spinal Cord Window Chamber for Multimodal Imaging

PLOS ONE | www.plosone.org 2 March 2013 | Volume 8 | Issue 3 | e58081

251

locate and track the same landmark between multiple imaging

sessions. Then, a small piece of customized artificial dura, made of

thin pliable and biocompatible silicon rubber (Eagerpolymers,

#0812, Chicago, IL, USA), was placed over the spinal cord to

prevent scar tissue formation (Figure 1E). The artificial dura was

prepared by polymerizing the optically transparent silicone rubber

with a curing agent (Cat. #0812, Eagerpolymers, Chicago, IL,

USA) to form a thin membrane which was spread in a Petri dish to

a thickness of 0.25 mm. A similar application of silicon rubber was

used by Shtoyerman et al. [21]. When the biocompatible artificial

dura was fully polymerized after 12 h, it was custom cut to size in

order to cover the exposed area of the spinal cord.

A sterile, light-weight SCWC device (Figure 1A) was implanted

and fixed to the superficial dorsal muscles and skin using standard

nylon sutures (Covidien, Syneture Monosof 5-0 sutures, Norwalk,

CT, USA) (Figure 1B). The SCWC was sutured tightly inside the

incision, with any additional skin being sutured together to create

a seal around the device. An 8 mm diameter glass coverslip was

placed inside the inner ring of the mounting device and then held

in place using a thin 8 mm diameter metal retaining ring. Animals

were administered anesthetics and underwent surgical procedures

for approximately 30 minutes each. Following surgical implanta-

tion of the device, mice were transferred to a temperature-

controlled recovery pad until awake and then returned to their

cages to fully recover. Animals were given oral antibiotics

(ClavamoxH) in water for 3 days following SCWC implantation

to prevent infection. At all times, with the exception of imaging

sessions, animals were allowed free access to food and water.

Figure 1. Mouse spinal cord window chamber (SCWC) device and surgical implantation procedures. (A) The SCWC device design anddimensions are shown. An 8 mm diameter glass coverslip was inserted into the SCWC device and held in place by a metal ring clamp once the deviceis surgically implanted into the mouse (shown in panel ‘‘F’’ and ‘‘G’’). Four radial extension arms have been built into the device in order to immobilizethe animal during imaging sessions. (B-E) Photographs showing step-by-step surgical procedures for implanting the SCWC in the mouse exposing thespinal cord at the L2–L3 vertebrae. (E) Artificial dura was placed on the dorsal surface of exposed spinal cord below the coverslip to prevent scartissue formation. Two separate devices were manufactured from either durable (F) polycarbonate or (G) light-weight surgical steel. Polycarbonatewas used to allow photoacoustic imaging in vivo. (H) X-ray images were taken following SCWC implantation to confirm the device had been placedover L2–L3 and demonstrate that the spinal cord and vertebrae remain structurally sound after implantation of the device. Scale bars = 1 cm.doi:10.1371/journal.pone.0058081.g001

Spinal Cord Window Chamber for Multimodal Imaging

PLOS ONE | www.plosone.org 3 March 2013 | Volume 8 | Issue 3 | e58081

252

Animals were monitored daily by veterinary staff for adverse

effects of the SCWC implants and any signs of decreased mobility,

infection, necrosis, or window chamber dehiscence.

Histology and ImmunostainingTo investigate the presence of an inflammatory or immune

response in the spinal cord caused by the surgical implantation of

the metallic and plastic SCWC devices, mice bearing chambers

were deeply anesthetized and transcardially perfused with 10%

formalin at 24, 48 and 72 h after implantation (n = 3 per group;

total n = 9) (Figure 2). Sham animals (naıve, no surgery or SCWC

implantation) were used as a control (n = 3). A 5-mm long section

of the spinal cord from directly below the SCWC was extracted

and fixed in 10% formalin for 24 hours and then embedded in

paraffin for tissue sectioning and histological staining. Tissues were

cut in longitudinal and axial sections, serially at a thickness of

4 mm, and fixed onto glass microscope slides (VWR, #48311-703,

Canada). Sections were stained with hematoxylin and eosin to

determine anatomical and cellular microstructures, and with Iba-1

(1:300, Cat. # 019-19741, Wako Chemicals USA, Richmond,

VA, USA) to assess inflammation following SCWC device

implantation [23,24]. Positive controls were used to confirm Iba-

1 reactivity of the antibody (mouse spinal cord tissue from 7 days

post-injury was used from animals receiving an 8 g clip-

compression spinal cord injury, as described previously by Yu

and Fehlings [25]).

White light micrographs were obtained of the H&E stained

tissue sections using a stereoscopic epifluorescence microscope

(Leica MZ FLIII, Leica Microsystems, Richmond Hill, ON,

Canada). For Iba-1, tiled images were taken at 20X magnification

using StereoInvestigatorH software (MicroBrightField Inc., Will-

iston, VT, USA), and images were quantified for fluorescent

intensity using ImageJ software (National Institutes of Health).

Data from each group was subject to a square-root transformation

(to adjust for any uneven distribution of normality and/or variance

within groups), and then a one-way ANOVA, with a Tukey post-

hoc analysis, was used to analyze the Iba-1 fluorescent intensity

data between control animals and SCWC animals at various time

points following implantation.

Western Blot AnalysisFollowing deep sedation, animals were sacrificed by decapita-

tion at 24, 48 or 72 hours following SCWC implantation (n = 3 per

group; total n = 9). Sham animals (naıve, no surgery or SCWC

implantation) were used as a control (n = 3). A 10 mm length of

the spinal cord centered under the SCWC was surgically removed.

Samples were mechanically homogenized in 100 ml of homoge-

nization buffer (0.1 M Tris, 0.5 M EDTA, 0.1% SDS, 1 M DTT

Figure 2. Visual and histological confirmation that SCWCimplantation does not damage the spinal cord structure orcause significant inflammation or infection. (A) White lightimages following SCWC implantation at 0, 1, 3 and 7 days showed nosigns of local infection, excessive bleeding around the installation site,or device rejection. SCWC remained optically clear for 29 days,permitting long-term high-resolution imaging of cord and vascularstructures. Yellow arrows indicate the location of the spinal cord. (B–E)Histological analysis and quantification of spinal cord tissue cross-sections cut directly below the caudal edge of the implanted SCWC. (B)H&E staining confirmed tissue morphology was intact after WCimplantation. (C) Representative Iba-1 immunohistochemistry imagesfrom spinal cords 24, 48 and 72 hours following SCWC implantation.Sham and spinal cord injury (SCI) (Iba-1 positive control) animals arealso shown for comparison. SCI positive control animals showeda significant increase in Iba-1 expression, * p,0.001. No notablechanges in Iba-1 expression in ex vivo spinal cord were observedbetween the SCWC implanted groups). (D) Western blot for Iba-1 priorto SCWC implant (sham), and at 24, 48, and 72 hours post-implantation

in athymic nude mice. (E) Bar graph representing the fluorescentintensity quantification of Iba-1, and no significant increase in Iba-1 wasobserved; n = 3 per group (p-values .0.212 for comparisons betweenall groups). (F) Bar graph showing quantification of Western blot data.Increases in Iba-1 protein were observed in animals receiving SCWCimplantation, although these increases were insignificant (p = 0.15);n = 3 per group. (G) Western blot for Iba-1 prior to SCWC implant (sham)(n = 3), and at 24 hours (n = 4), 3 days (n = 3), 10 days (n = 3) and 28 days(n = 3) post-implantation in C57BL6 mice. (H) Bar graph showingquantification of Western blot Iba-1 data (C57BL6 mice). No significantincreases in Iba-1 protein were observed (p = 0.405). It is anticipatedthat the slight increases in Iba-1 in the athymic nude mice may be dueto the laminectomy and surgical procedures performed. b-Actin wasused as a protein loading control. Scale bars = 400 mm. Data wastransformed (square-root transformation) and analyzed using a one-way ANOVA; Tukey post-hoc analysis. SCWC= spinal cord windowchamber.doi:10.1371/journal.pone.0058081.g002

Spinal Cord Window Chamber for Multimodal Imaging

PLOS ONE | www.plosone.org 4 March 2013 | Volume 8 | Issue 3 | e58081

253

solution, 100 mM PMSF, 1.7 mg/mL aprotinin, 1 mM pepstatin,

and 10 mM leupeptin) and centrifuged at 15,000 rpm for 10

minutes at 4uC. Supernatants were extracted and used for Western

blot analysis, where 10 mg of protein was loaded into 12%

polyacrylamide gels (Bio-Rad, Mississauga, Canada). Membranes

were probed with primary anti-Iba-1 antibody (1:500, Cat. # 016-

20001, Wako Chemicals USA, Richmond, VA, USA). Primary

antibodies were labeled with horseradish peroxidase-conjugated

secondary antibodies (goat anti-rabbit IgG, 1:2000; Jackson

ImmunoResearch Laboratories, West Grove, PA, USA), and

bands were imaged using an enhanced chemiluminescence (ECL)

detection system (Perkin Elmer, Woodbridge, Canada). Mouse

monoclonal beta-actin (Chemicon International, Inc., Temecula,

CA, USA) was immunoblotted as a loading control as per standard

protocol. Gel-Pro AnalyzerH software (Media Cybernetics,

Bethesda, MD, USA) was used for integrated optical density

(OD) analysis and quantification of Iba-1 protein expression

(Figure 2). Data from each group was subject to a square-root

transformation (to adjust for any uneven distribution of normality

and/or variance within groups), and then a one-way ANOVA was

used to statistically analyze the data between naıve and SCWC

implanted groups. A Tukey post-hoc was applied.

X-ray Micro-irradiation and Vascular InjuryTo demonstrate the feasibility of using the mouse SCWC for

imaging radiation response of the spinal cord and its vasculature

in vivo, we delivered ionizing radiation to the spinal cord using

a custom-designed small animal X-ray microirradiator system

(XRad225Cx, Precision X-Ray Inc., North Branford, CT, USA)

(Figure 3A). The fully automated microirradiator system was

controlled using a computer system that integrates cone beam

computer tomography (CT) imaging with focused X-ray delivery

technology, and was able to deliver a single focal radiation beam at

a dose of 30 Gy with a diameter of 3 mm directly to the spinal

cord at 2.5 Gy/min. The X-ray tube was mounted on a rotating

gantry with a flat panel detector located opposite to the isocenter,

which facilitated imaging and irradiation of the target at any given

angle. The irradiator was calibrated to ensure accurate dose

delivery with tissue phantoms using methods previously described

[22].

Figure 3. SCWC model permits X-ray microirradiation of the spinal cord in situ. (A) Anaesthetized mice were placed directly under themicro-irradiation collimator of the small animal irradiator for delivery of X-rays to the spinal cord through the coverglass of the window chamber. (B,C) In situ fluoroscopic imaging was used for image-guided delivery of the X-ray beam (centered on the crosshairs) to the spinal cord. (D) Custom fitradiochromic film was used to confirm the location of the irradiation beam which had a 3 mm diameter (as seen by the dark blue circle). Scalebar = 1 cm. SCWC= spinal cord window chamber.doi:10.1371/journal.pone.0058081.g003

Spinal Cord Window Chamber for Multimodal Imaging

PLOS ONE | www.plosone.org 5 March 2013 | Volume 8 | Issue 3 | e58081

254

Prior to irradiation, mice were anaesthetized by intraperitoneal

injection of ketamine (80 mg/kg) and xylazine (5 mg/kg) and were

secured on the stage at the radiation isocenter (n = 5). Fluoroscopy

images of anatomical features of the animal and the integrated

targeting software were used to align the center of the target to the

isocenter of the radiation beam in the three axes (X,Y,Z) by

automatic movement of the stage for an anterior-posterior (AP/

PA) radiation treatment. Radiation dosimetry was performed

using radiochromic EBT film (ISP Inc., Wayne, NJ, USA)

consisting of a radiosensitive monomer that polymerizes upon

irradiation. A white light image of the mouse was taken using

a stereoscopic epifluorescence microscope (Leica MZ FLIII, Leica

Microsystems, Richmond Hill, ON, Canada) immediately after X-

ray irradiation in order to visualize and spatially define the

radiation field to permit accurate spatial localization of the

treatment dose for subsequent intravital fluorescence imaging.

Intravital White Light and Fluorescence ImagingIn vivo white light and fluorescence imaging were performed on

mouse spinal cords at 1, 24, 48 hours, and up to 5 days after

irradiation (n = 5). Prior to each imaging session, mice were

anaesthetized by intraperitoneal injection of ketamine (80 mg/kg)

and xylazine (5 mg/kg) and placed within the custom-made

animal restraint and secured to the microscope stage with an

embedded heating pad to maintain the animal’s body temperature

during imaging.

White light and fluorescence macroscopic images of the spinal

cord were acquired through the transparent glass coverslip of the

window chamber using a stereoscopic epifluorescence microscope

(Leica MZ FLIII, Leica Microsystems, Richmond Hill, ON,

Canada). To visualize the spinal cord vasculature, FITC-

conjugated dextran (0.65 mg/mouse, MW = 20 kDa; Sigma–

Aldrich Corporation Ltd, Oakville, Canada) was administered

intravenously by tail vein prior to fluorescence imaging, and then

imaged using a 470 nm excitation filter set. Using this method,

macroscopic imaging allowed for the determination of radiobio-

logical changes to the spinal cord tissue and vasculature at the sub-

millimeter scale.

Intravital Bioluminescence and Fluorescence Imaging ofTumor Metastases In Vivo

To demonstrate the feasibility of using the mouse SCWC for

imaging tumor micrometastases in vivo, we used an inverted

confocal fluorescence imaging microscope (LSM510 Laser Scan-

ning Confocal Microscope, Carl Zeiss, Jena, Germany) to visualize

microvasculature and the micrometastases of WW426 medullo-

blastoma cells (both fluorescent and bioluminescent, expressing a c-

myc-GFP tag and Luc-RFP reporter construct) which were

injected intracranially 28 days prior to in vivo imaging. Cell culture:

WW426 cells were grown as adherent culture using

DMEM:FBS(10%) heat inactivated in standard non-treated tissue

culture. The cell line is predisposed for myc-C(GFP) cells to

become unattached, but the GFP (myc) signal has a positive

feedback loop whereby high myc-C(GFP) signal is required to

sustain high levels of myc-C(GFP). Thus, it was essential that the

floating cells are retained in the culture both during feeding and

splitting. When the cells reached 60–80% confluency, they were

split 1:3 (approximately every 5 days). Old media was removed

Figure 4. Longitudinal optical imaging of radiation response of the normal spinal cord and its vasculature through the SCWC.Whitelight, wide-field fluorescence, and svOCT images were taken at (A) 1 day before, (B) 1 hour, (C) 24 hours, and (D) 48 hours following a single 30 Gyradiation dose to the cord. White light images revealed significant radiation-induced hematoma in the spinal cord two days after irradiation (arrows).FITC-dextran was injected intravenously prior to acquiring the fluorescence images at each time point. Fiducial markers consisting of India ink (blackdots shown by arrows) on the white light images were used to as spatial landmarks to allow identification and long-term imaging of vascularstructures. Compared with vascular function, corresponding en-face projected svOCT images revealed that the posterior spinal cord vein and othervasculature did not suffer significant short-term radiation-induced structural damage. Scale bar = 1 mm. SCWC= spinal cord window chamber.svOCT= speckle variance optical coherence tomography.doi:10.1371/journal.pone.0058081.g004

Spinal Cord Window Chamber for Multimodal Imaging

PLOS ONE | www.plosone.org 6 March 2013 | Volume 8 | Issue 3 | e58081

255

from the flask, with floating cells, and kept in a 15 mL tube. 1 mL

trypsin was added to the flask and then placed at 37uC for 5

minutes, or until the cells dissociated. The trypsin was neutralized

with existing culture media, and the suspension was centrifuged to

retrieve all cells.

Cell transplantation. The skulls of athymic nude mice

(n = 10) were surgically exposed and bur-holes were carefully

drilled to allow access to the posterior cerebellum. 46105 cells

(total volume 10 ml) were injected 2–3 mm deep into the

cerebellum at a rate of 2 ml/minute (n = 3). The needle remained

in the brain for 1 minute after the injection to prevent fluid

backflow.

The primary intracranial WW426 medulloblastoma tumors

took approx. 3 weeks to grow and then metastasize to the spinal

cord, and some animals did not develop spinal metastases in these

experiments. Our intention was to determine whether our WC

model was capable of imaging tumor micrometastases occurring in

the spinal cord in vivo, as a proof-of-concept. To test this, we used

in vivo bioluminescence imaging to non-invasively track the tumor

growth in the brain in the whole animal from the time of the initial

tumor cell implant prior to surgical implantation of the WC

device, since our previous experience with this tumor line showed

it was slow growing. This enabled us to implant the WC only once

the primary brain tumor was sufficiently grown and metastases to

the spine were likely. Bioluminescence imaging was performed

using an IVIS Spectrum imaging system (Caliper, MA, USA) by

injecting luciferin substrate intraperitoneally (150 mg/kg) prior to

each bioluminescence imaging session. We then implanted the

SCWC devices 27 days after intracranial tumor seeding and used

bioluminescence to confirm the presence of medulloblastoma

micrometastases within the spinal cord through the transparent

coverslip window.

Figure 5. SCWC permits structural, functional and oxygenation imaging of the intact spinal cord vasculature in situ. (A) PowerDoppler ultrasound (color) overlaid on a B-mode structural ultrasound (gray-scale) image obtained through the polycarbonate SCWC alonga longitudinal section of the normal spinal cord in vivo (device is shown in Figure 1F). The power Doppler depicts vascular architecture in severalvessels of the spinal cord. The color bar represents the signal intensity. (B) Corresponding multispectral photoacoustic imaging of the same crosssection of normal spinal cord permitted in situ measurement of hemoglobin oxygen saturation in the anterior spinal artery and posterior spinal vein.It demonstrated that the cord is well oxygenated. The color bar represents the relative hemoglobin oxygen saturation level. (C) Cross-sectionalDoppler OCT image demonstrated significant blood flow in the posterior spinal vein. The color bar represents the phase-shift of the backscatteredlight in radians which is proportional to the velocity of the red blood cells in the axial direction. (D) Corresponding structural OCT image of the spinalcord permitted visualization of key spinal cord features, including the glass coverslip (1), anterior spinal vein (2), white matter (3), and grey matter (4)of the intact cord. Scale Bars = 500 mm (A–D). SCWC= spinal cord window chamber. OCT=optical coherence tomography.doi:10.1371/journal.pone.0058081.g005

Spinal Cord Window Chamber for Multimodal Imaging

PLOS ONE | www.plosone.org 7 March 2013 | Volume 8 | Issue 3 | e58081

256

Since the tumor cells were GFP-positive, TRITC-conjugated

dextran was used to image the vasculature, and was administered

intravenously via the tail vein prior to fluorescence microscopy

(100 mg/kg body weight, MW = 155,000 Da; Sigma–Aldrich

Corporation Ltd, #T1287, Oakville, Canada). A Zeiss LSM510

confocal fluorescence microscope (Carl Zeiss, Jena, Germany) was

used to observe the location of the micrometastases in relation to

the vasculature. A 56Fluar objective (Carl Zeiss, Jena, Germany)

was used for intravital confocal fluorescence microscopy of the

cord as it had a 12.5 mm (NA 0.25) working distance and allowed

a wide area of the cord to be imaged without the need for tiling of

multiple images. Animals were imaged 1 day following SCWC

implantation (28 days post-cell transplantation).

Optical Coherence Tomography (OCT) ImagingOptical coherence tomography (OCT) was used for depth

resolved three dimensional structural and functional imaging of

the spinal cord and its vasculature in vivo. Imaging was performed

on anesthetized mice with a swept-source OCT system described

previously [22]. Briefly, a 36-kHz swept laser source with

a sweeping range of 110 nm centered at 1310 nm was used to

acquire depth resolved structural images of the intact in vivo spinal

cord up to ,2 mm in depth with an axial resolution of ,8 mm

and a lateral resolution of ,13 mm. Three dimensional structural

OCT images were acquired over 2.5 mm63 mm regions of the

cord within the window chamber.

Speckle variance OCT (svOCT) is a functional extension of

OCT that enabled depth resolved three-dimensional imaging of

in vivo spinal cord vasculature as small as ,20 mm in diameter

without the use of exogenous contrast agents [22]. The difference

in the temporal speckle statistics of blood and solid tissues provides

the contrast in svOCT. Three dimensional vascular images were

acquired over the same 2.5 mm63 mm region as the structural

images and vascular contrast was obtained by computing the

interframe speckle variance over four consecutive B-mode images.

svOCT is highly sensitive to motion such as breathing and

heartbeat; therefore mice were secured in a custom holding frame

to minimize motion artifacts.

Unavoidable motion artifacts caused by breathing created

bright streaks through the images in the scanning direction and

were minimized by applying a 363 median filter in the depth

direction, followed by a clamp to remove low-intensity pixel

values.

Doppler OCT imaging enabled real-time visualization of blood

flow in the posterior spinal vein. Two-dimensional B-mode

Doppler images were formed using the Kasai estimator to

determine the phase shift of scattering red blood cells over

consecutive A-scans [23]. Doppler imaging was performed with

2000 A-scans over a 1 mm region centered on the posterior spinal

vein with an ensemble length of eight. The imaging head was

angled ,15u relative to the surface of the chamber, providing

a Doppler angle of ,75u. (n = 4, for svOCT and Doppler OCT

imaging. Imaging parameters were optimized during OCT

sessions, using 3 mice).

Ultrasound and Photoacoustic ImagingIn vivo ultrasound, power Doppler, and photoacoustic imaging

of the polycarbonate (plastic – Figure 1F) SCWC-bearing animals

were performed using the Vevo2100 and Vevo LAZR systems

(VisualSonics Inc., Toronto, ON, Canada) with a 40 MHz centre

frequency transducer (LZ-550, VisualSonics Inc., Toronto, ON

Canada) at 24 hours following SCWC implantation (n = 2). These

experiments were conducted as terminal, end-point procedures.

The vascular hemoglobin oxygen saturation (sO2) was determined

by irradiating the window chamber with light of two different

wavelengths (750 nm and 850 nm). The built-in software on the

Vevo LAZR system automatically calculated sO2 based on the

received photoacoustic signals. The energy density of the laser

beam at the surface of the window chamber was approximately

3 mJ/cm2. The glass coverslip was not acoustically compatible;

therefore it was removed for these experiments. Sterile coupling

gel (LithoClear, Sonotech, Washington, USA) was applied to the

artificial dura above the spinal cord to facilitate the transmission of

acoustic waves between the tissue and ultrasound transducer,

therefore reducing air-tissue interface-based imaging artifacts.

Three-dimensional co-registered power Doppler and sO2 mea-

surements of the spinal cord were performed while the mouse was

breathing 100% oxygen mixed with 2% isoflurane for approx-

imately 20 minutes. In addition, during photoacoustic imaging, the

hemoglobin oxygen saturation recovery dynamics of the spinal

cord were measured by shifting the animal’s anesthetic mixture

from 100% to 7% oxygen for 1 minute. Quantification of the sO2

recovery measurement was performed in a region of interest

around the spinal cord in a single imaging plane. Post-processing

of sO2 and power Doppler images was performed using Amira

(Visage Imaging, San Diego, CA, USA). A median filter was

applied to the sO2 data set to reduce the effects of clutter. The data

sets were overlaid with an anatomical B-Mode image.

Results

Spinal Cord Window Chamber Design and ImplantationIn the present study, we designed and developed two types of

spinal cord window chamber (SCWC) devices (metal and plastic)

and the procedures to surgically implant them in mice to permit

longitudinal high-resolution multimodal optical and acoustic

imaging of the spinal cord and its vasculature (Figure 1A). Our

SCWC device was easily implanted following a two-level

laminectomy at L2–L3 (Figures 1B–1E, 1H) with both poly-

carbonate (Figure 1F) and metal (Figure 1G) compositions to

permit in situ imaging of the cord and its vasculature using a several

complementary intravital optical imaging modalities. The devices

Figure 6. Hemoglobin oxygen saturation (sO2) measurementmade using in vivo photoacoustic imaging. Baseline vascular sO2

was measured in situ for 1 minute while the animal breathed 100%oxygen mixed with 2% isoflurane. The animal was shifted to breathing7% oxygen mixed with 2% isoflurane for an additional 1 minute. Theoxygen concentration was then returned to 100%.doi:10.1371/journal.pone.0058081.g006

Spinal Cord Window Chamber for Multimodal Imaging

PLOS ONE | www.plosone.org 8 March 2013 | Volume 8 | Issue 3 | e58081

257

Figure 7. Intravital multispectral fluorescence microscopic imaging of medulloblastoma tumor metastasis to the spinal cord. (A) Invivo bioluminescence images of mice 7 days following intracranial tumor implantation of human WW426 medulloblastoma tumor cells,demonstrating local tumor growth. (B) SCWC was implanted 27 days after tumor implantation, when metastatic GFP+ tumor cells to the spinal cordcould be seen using both BLI and intravital two-photon images (color bar indicates bioluminescence signal intensity; BLI units are photons/s/cm2/Sr).The head of the mouse was covered in ‘‘B’’ to reduce the bioluminescence signal from the brain in order to detect lower bioluminescence from thetumor micrometastases. (C) Wide-field fluorescence imaging and (D) confocal fluorescence microscopy of the SCWC-bearing mouse 28 days afterinitial tumor implantation (1 day post-SCWC installation). The outline of the spinal cord is highlighted with the orange dotted line in ‘‘C’’. TRITC-dextran shows the posterior spinal cord vein. The arrows in ‘‘C’’ and ‘‘D’’ indicate the location of multiple tumor micrometastases in close proximity tothe spinal cord vasculature. Scale bars = 1 mm. SCWC= spinal cord window chamber. BLI = bioluminescence imaging.doi:10.1371/journal.pone.0058081.g007

Spinal Cord Window Chamber for Multimodal Imaging

PLOS ONE | www.plosone.org 9 March 2013 | Volume 8 | Issue 3 | e58081

258

were light weight and the animals tolerated them well for up to 1

month (See Video S1).

In a cohort of (non-irradiated) animals, we investigated the

possibility of the implanted SCWC devices causing local swelling

and/or infection at the surgical site, and examined the spinal tissue

directly below the window chamber for tissue damage and

inflammation. White light images following SCWC implantation

showed no visible hallmark signs of local infection or device

rejection at day 0, 1, 3 or 7 after SCWC implantation (Figure 2A).

There was no microstructural damage to the cord as determined

by ex vivo histological assessment using hematoxylin and eosin

staining in animals 24, 48 and 72 hours post-SCWC implantation

(Figure 2B). Furthermore, using ex vivo immunostaining of Iba-1,

an indicator of macrophage/microglia activation and inflamma-

tion, we confirmed that there was negligible inflammation in spinal

cord tissues from the time of the device implantation and up to

72 h after implantation, thereby indicating that the SCWC device

did not cause injury to the cord (Figure 2C, E). In contrast, spinal

cord injured (SCI) mice at 7 days post-injury, used as a positive

control for Iba-1 staining, showed an increase in Iba-1 expression,

which is consistent with previous reports [26,27]. In addition,

Western blot analysis and quantification of Iba-1 protein further

indicated a lack of an inflammatory response in the area below the

SCWC (Figure 2D, F). Although we observed a slight increase in

Iba-1 protein in animals with SCWC implanted, this increase is

likely due to the surgery and two-level laminectomy performed in

these animals, rather than the installation of the window chamber

mount itself. Sham animals, which did not receive a laminectomy,

were used as the control group for Iba-1 protein quantification.

Sham animals showed reduced Iba-1 expression; however, when

compared to C57BL6 or athymic nude animals with SCWCs

installed, no considerable changes in Iba-1 expression were

observed (p = 0.15). Overall, the data suggest that implanting

our SCWC design over the spinal cord is feasible in vivo, and does

not result in any discernible damage to the spinal cord tissue.

The SCWC remained optically clear for up to 29 days of

imaging, after which point the devices detached (suture failure)

and tissue growth into the window chamber area prevented

further imaging. On average, SWCWs remained optically clear for

21 days without need for intervention; however, if required, mild

tissue growth into the window chamber area was easily removed

prior to imaging, allowing imaging to be conducted out to 29 days.

The replacement of coverslips between imaging sessions was

simple and rapid (e.g. a few minutes). The devices that we

developed were easily sterilized by autoclave (for metal device) or

surgical disinfectant (for plastic device) and were compatible with

commercially-available glass coverslips of standard diameter.

Based on our qualitative observations and quantitative (ex vivo)

assessments following SCWC implantation, we observed that the

SCWC devices did not cause physical or biological damage to the

spinal cord or its vasculature. Thus, the surgical implantation

procedure or the prolonged use of an in vivo SCWC did not

compromise the integrity or the interpretation of imaging data

obtained in order to study the effect of a given treatment by

differentiating it from background biological response (e.g. that

might have occurred due to inflammation after surgical implan-

tation) (Figure 2).

Intravital Imaging of Radiation-induced Changes toSpinal Cord Vasculature

To demonstrate the utility of the animal model, we used our

SCWC model to study the biological response of the spinal cord

and the vasculature to X-ray irradiation. We specifically selected

a microirradiation approach to induce vascular damage, because it

could be delivered in a controlled, spatially-localized, and

reproducible manner using the small animal X-ray microirradiator

(Figure 3A, B, C). A benefit of using an implanted SCWC device

was that imaging of the cord could be performed in vivo before and

serially after irradiation in the same animal. Thus, each animal

could act as its own experimental pre-treatment control. This

reduced the number of animals required for experiments, as well

as controlled for individual differences in vascular organization

and branching within each mouse spinal cord.

Consistent with previous studies of spinal cord irradiation

[28,29,30], we observed significant radiation-induced hematoma

in the spinal cord white matter two days after a single 30 Gy

irradiation with a 3 mm beam diameter (Figure 3D, 4). FITC-

dextran was injected intravenously prior to acquiring the

fluorescence images at each time point to visualize spinal cord

vasculature, and revealed significant decrease in vascular function

in the posterior spinal cord vein and vasculature, as a result of

radiation-induced damage. These vascular changes occurred as

early as 24 h after treatment and worsened at day 2 (Figure 4).

Moreover, we observed significant radiation-induced hematoma in

the spinal cord white matter 2 days after a single 30 Gy

irradiation, which is consistent with the literature [28]. Increase

in vascular permeability occurred following irradiation, as seen by

the leakage of FITC-dextran from intact vasculature. Edema and

extravasation of red blood cells due to an increase in vascular

permeability following irradiation has been observed previously

[29,30]. Compared with this radiation-induced vascular dysfunc-

tion, corresponding svOCT images revealed that the posterior

spinal cord vein and vasculature did not suffer from significant

radiation-induced structural damage over the same 2 day period.

We used India ink markers placed directly on the cord surface as

spatial landmarks to allow identification and serial imaging of the

same vascular structures without the need for image alignment

post-acquisition. These data illustrated that the SCWC could be

used to follow the radiobiological response of the cord and its

vasculature at morphological, microstructural, and functional

levels. Fluorescence and svOCT imaging enabled clear in vivo

longitudinal imaging of the posterior vein as well as the

microscopic radial-branching vessels of approximately 25 mm

diameter. svOCT was able to resolve vessels and spinal cord

structure up to 500 mm in depth. Images were of high quality and

had sufficient signal-to-noise ratios as determined by comparison

between background fluorescence and svOCT intensities.

Intravital Power Doppler Ultrasound, Photoacoustic andDoppler OCT Imaging of the Spinal Cord and itsVasculature

To further demonstrate the use of the SCWC model for other

complementary imaging techniques we used power Doppler

ultrasound to highlight the vascular network of the spinal cord

(Figure 5A) [31]. Using the same animal, we measured sO2 in the

intact spinal cord using multispectral photoacoustic imaging

(Figure 5B). However, since power Doppler is more sensitive to

the detection of small vessels compared to photoacoustics, the data

shown in Figure 5A (power Doppler; see Video S2) displayed an

increased number of vascular structures in comparison to

Figure 5B (photoacoustics; see Video S2). Our SCWC method

permitted image-based sO2 measurements in spinal vessels that

would not have been possible without a laminectomy, since the

vertebrae would have prevented effective photoacoustic imaging.

Our method overcomes the impractical limitations involving the

use of traditional oxygen electrodes which must be placed within

the spine to measure vascular/tissue oxygenation and which only

measure sO2 in one small tissue volume (,1 mm3) at a time,

Spinal Cord Window Chamber for Multimodal Imaging

PLOS ONE | www.plosone.org 10 March 2013 | Volume 8 | Issue 3 | e58081

259

requiring the needle to be moved many times for multiple

measurements and possibly causing traumatic tissue damage to the

cord [32]. We also demonstrated the ability to measure sO2

recovery in real time (Figure 6; see Video S3). We found that

transitioning the mouse from breathing 100% to 7% oxygen for 1

minute decreased sO2 by approximately 23% and took approx-

imately 30 sec to return to baseline values (Figure 6; see Video S3).

Combining power Doppler ultrasound and spatially co-registered

photoacoustic imaging of the same mouse spinal cord enabled

tracking of vascular structure and sO2 dynamics in the same

mouse over time.

We also demonstrated the feasibility of using in vivo Doppler

OCT to image blood flow, while simultaneously capturing the

cross sectional structure of the spinal cord (Figure 5C, D). Doppler

OCT was able to image the posterior spinal vein only, compared

with photoacoustic or power Doppler ultrasound which provided

deeper tissue penetration to the anterior side of the cord. However,

a major advantage of OCT imaging was the ability to spatially

resolve anatomical microstructures of the spinal cord itself

(Figure 5D; see Video S4 and Video S5), which was not possible

using ultrasound imaging alone.

SCWC Allows for Visualization of Micrometastases in theSpinal Cord

To further highlight an additional preclinical research use of the

SCWC model, we demonstrated the intravital visualization of

tumor micrometastases within the spinal cord originating from

WW426 medulloblastoma cells transplanted intracranially

(Figure 7A). Using bioluminescence imaging (BLI), we were able

to track the migration of tumor cells down the spinal cord until

they were directly under the SCWC (28 days post-transplant)

(Figure 7B). Using epifluorescence microscopy, we were able to

identify localized tumor micrometastases at L2–L4, immediately

under the SCWC (Figure 7C). Furthermore, using TRITC-

dextran to mark the spinal vessels, we observed that the metastases

were in close proximity (up to 600 mm away) to the posterior spinal

vein where they could access oxygen and nutrients (Figure 7D).

Discussion

We have developed a new transparent window chamber device

and surgical implantation protocol for mice that overcomes the

inherent limitations of many previous in vivo spinal cord imaging

studies. In comparison to the SCWCs developed by Farrar et al.

and Fenrich et al., we have designed an alternative device, which is

compatible with additional optically-enabled imaging techniques,

e.g. photoacoustics, to obtain important complementary structur-

al, functional and oxygenation information about vasculature

in vivo. Our device design and surgical implantation methods are

less complex and easily implemented for in vivo spinal cord

imaging. Overall, this model enables direct in vivo, intravital

multimodal imaging of healthy and diseased spinal cord and its

vasculature over time. White light imaging provided high-

resolution information about cord anatomy and vasculature,

including hemorrhage that may occur as a result of damage cause

to the cord by irradiation [28]. When combined with injectable

fluorescent blood contrast agents, such as FITC- or TRITC-

dextran, intravital confocal fluorescence microscopy imaging

provided high intensity contrast-based images of the spinal cord

vascular network for vessels as small as ,25 mm in diameter.

svOCT provided a contrast agent-free method of imaging the

structure of blood vessels of the spinal cord. A limitation of both

fluorescence and svOCT in imaging the spinal cord of mice is the

lack of tissue-penetration to image through the full thickness of the

cords, which is approximately 1.5 mm in mice [33]. Using

photoacoustic imaging, the oxygenation level of the cord

vasculature was quantified while power Doppler ultrasound

provided visualization of vascular architecture through the

complete thickness of the spinal cord. Thus, our SCWC model

could be a useful tool for future imaging studies of vascular events

following spinal cord injury and their contribution in pathogenesis

[34].

The SCWC also permitted the use of a small animal micro-

irradiator to focally treat the spinal cord directly with X–rays,

followed by the imaging of the radiobiological response of the cord

and the vasculature in situ over time. To our knowledge, this is the

first attempt to study the spinal cord vascular response to radiation

in mice over time using a transparent spinal cord window chamber

and multimodal intravital optical imaging approach. Our results

demonstrated the feasibility of this new method for studying the

spinal cord vasculature and the sensitivity of this approach to

radiobiological changes associated with morphology, structure and

function induced by a single dose of 30 Gy. The small animal

microirradiator is capable of delivering a variety of clinically-

relevant radiation therapy doses and treatment regimens (e.g.

single and multiple fractions) in a variety of treatment beam

geometries [22]. When combined with our SCWC murine models,

the microirradiator system could offer an important new pre-

clinical experimental platform for studying the radiobiological

response of the spinal cord in murine models (including in the

presence of primary and metastatic tumors) longitudinally and at

cellular resolution [35]. This has not been possible to date despite

significant work on spinal cord radiobiology and ischemia [36,37].

Another application of the SCWC models could be for preclinical

studies of photodynamic therapy of spinal cord tumors and

metastases [38], such that tumors could be irradiated by light and

then imaged with optical and/or other imaging techniques (e.g.

ultrasound) over time to measure the response of various tissue

components. We have demonstrated that fluorescence microscopy

can be used with our SCWC model to visualize tumor

micrometastatic colonies in relation to the spinal cord and its

vasculature. This approach could allow the study of spinal cord

pathophysiology of metastatic spread as well as tumor angiogenesis

at cellular-level resolution in vivo in future studies; however, the

system will require optimization to reduce motion (breathing)

artifacts during in vivo imaging. In addition to demonstrating the

capacity to identify and track tumor cells in vivo, our model, when

combined with high-resolution microscopy, may have the poten-

tial to observe cell-cell interactions (i.e. oligodendrocyte-neuron)

and cell motility (i.e. leukocyte trafficking through diapedesis),

which may be beneficial for visualizing CNS regeneration or

monitoring a localized inflammatory responses, respectively.

There is also a growing body of evidence pointing to the existence

of a subset of tumor cells with high tumorigenic potential in many

spine cancers that exhibit characteristics similar to stem cells [39].

Our intravital SCWC experimental model could be useful for such

emerging biological studies.

While there have been previous studies reported on in vivo

optical imaging of the spinal cord, they have required repeated

surgeries to remove the skin to access the cord for longitudinal

imaging [10,11,12,13,14,15,16,40,41]. Recently, two studies have

demonstrated the implementation of a transparent window

chamber approach to facilitate serial optical imaging of the spinal

cord in vivo; however, their SCWC have differed in materials and

design in comparison to our model [16,17]. In contradistinction to

our design, their spinal chamber incorporated metal components

situated deep in the vertebral column, and were located adjacent

to the vertebrae (although not in direct contact with the spinal

Spinal Cord Window Chamber for Multimodal Imaging

PLOS ONE | www.plosone.org 11 March 2013 | Volume 8 | Issue 3 | e58081

260

cord). These metal components stabilized the cordduring imaging

and the authors state that this technique produced a moderate

inflammatory response in the cord (microglial density was

increased at both 1d and 7d post-implantation). In our model,

we developed a chamber that was not fixed to the vertebrae, yet

still permitted optical imaging with few motion artifacts. Our

design involved components that are located superficially to the

spinal cord (implanted and stablizied in the dorsal skin/muscle),

preventing any contact between the spinal cord and the SCWC

device. In addition, unlike the Farrar and Fenrich designs, which

include metallic components, our plastic spinal chamber enabled

the use of real-time multispectral photoacoustic and ultrasound

imaging in addition to the fluoerscence and svOCT imaging

in vivo. This is the first time that OCT imaging has been performed

longitudinally in the in vivo spinal cord, and the first time

photoacoustic imaging has ever been used to image the spinal

cord in vivo. Thus, using a plastic chamber increased the number

of additional compatible imaging techniques that could be used to

study the spinal cord in greater detail. The transparent SCWC

model could offer a method of precise delivery of either ionizing or

ablative optical energy to the cord with a potential role in studies

of cellular-based regenerative enhancement of spinal cord injury

repair [42].

Our SCWC experimental model is practical, easy to use, and

involves a single surgical procedure to implant the window

chamber, which minimizes the possibility of focal traumatic

damage caused by repeated surgical exposure of the spinal cord. In

contrast to Farrar et al., we did not observe considerable local

tissue inflammation in the cord, as demonstrated by negligible

changes in Iba-1 expression in cord tissues after implantation

(Figure 2). This suggests that a SCWC implanted more

superficial/dorsal, rather than directly adjacent to the vertebral

column, may be a more effective model that does not induce

significant inflammation. Further, no motor-function deficits or

neuropathology were observed in the chamber-bearing mice (See

Video S1). Importantly, the metal and plastic spinal chambers were

designed to be lightweight and to minimize discomfort to the

animal as well as interference to the normal activities (e.g. feeding,

drinking, grooming, locomotion, etc.) of the animal. Therefore, the

SCWC we have developed for long term chronic imaging of the

normal and diseased spinal cord, including after radiation

treatment, is a robust, reproducible, and a useful murine

experimental model for imaging the spinal cord and its vasculature

with optically-enabled intravital imaging techniques.

Additionally, we have recently extended the SCWC model from

mice to rats (See File S1 and Figure S1). By developing a rat SCWC

model, we open the possibility of experimental studies in a larger

murine model that more accurately represents the human spinal

cord. This is important for studying the pathophysiology and

development of the cystic cavity following spinal cord injury (See

File S1) [43,44].

While our SCWC model overcomes previous challenges in

optical imaging of the living spinal cord, we have recognized a few

limitations. Firstly, our chambers were designed with an inner

diameter of 8 mm, which is wide enough to fit standard

microscope objective lens (from 1X –60X), yet had an overall

size and profile that were small enough to avoid restricting animal

movement after implantation. However, an 8 mm diameter

transparent window to the spinal cord allows microscopic

observation of the spinal cord for only 2 to 3 spinal segments at

the most. Increasing the size of the chamber would require more

vertebral segments to be removed and we found this risked

fracture of the vertebral column. For the purposes of longitudinal

and localized optical imaging of tissue, cellular and vascular

changes to the cord following traumatic and/or localized damage

(i.e. spinal cord injury, stroke, spinal tumors), an 8 mm diameter

window is sufficient. However, for the study of neurological

diseases, such as multiple sclerosis or amyotrophic lateral sclerosis,

which have widespread effects along the spinal cord, in vivo optical

imaging using our SCWC setup may not be appropriate to assess

the global deterioration of the whole cord [45,46]. Secondly, while

intravital confocal fluorescence microscopy (with exogenous blood

contrast agents) and svOCT (without contrast agents) provide

high-resolution imaging of the microvasculature which cannot be

achieved by other imaging techniques such as microCT or MRI,

these optical imaging methods cannot image through the full

thickness of the spinal cord. Therefore, only pial and white matter

vessels are accessible using fluorescence and svOCT, while the

majority of vascular structures in the grey matter remain

a challenge for optical imaging. One possible alternative is the

use of photoacoustic imaging which uses high-frequency ultra-

sound to image through the entire spinal cord following pulsed

laser excitation, including vessels of the grey matter. While yielding

useful structural information about tissue and vasculature, in-

travital multispectral photoacoustic imaging also provides impor-

tant functional information of spinal cord vascular oxygenation

status non-invasively over time.

Nevertheless, our research – in conjunction with previous

reports by Farrar et al. and Fenrich et al. – substantiate the need

and confirm the technical feasibility for such unique murine

window chamber models for in vivo longitudinal multimodal

imaging of the spinal cord. Future preclinical studies of the

healthy, diseased or injured spinal cord will benefit from the

availability of such robust experimental animal models and their

ability to exploit powerful multimodal and intravital imaging

techniques.

Supporting Information

Figure S1 Spinal cord window chamber designed for imaging of

the rat cord and vasculature. (A) The spinal cord window chamber

device was designed and printed in polycarbonate with a metal

ring to secure the 12-mm coverglass slip. (B) The SCWC device,

shown from a different angle, has lateral arms which retract the

dorsal muscles of the vertebrae and keep the spinal cord exposed

over days-to-weeks- for longitudinal optical imaging. (C) The

SCWC device is surgically implanted over the exposed spinal cord

providing a window for direct spinal cord imaging, following a two-

level laminectomy at T6–T7. (D) Images of the rat spinal cord

were taken 3 days after SCWC implantation. Intravital white light

at 2X magnification (left panel), and 6X magnification inset of

intravenous FITC-dextran (right panel) images are shown. Scale

bars = 1 cm. SCWC = spinal cord window chamber.

(TIF)

File S1

(DOCX)

Video S1 Behavioural and functional observation of mice 28 day

post-SCWC implantation. Athymic nude mice had spinal cord

window chambers implanted and were followed for 1 month to

examine their behaviour, motor function, grooming, and eating

habits, as well as to document any necrosis, inflammation or

infection surrounding the implantation site. No motor/beha-

vioural deficits were observed in the 28 day period. Similarly, no

observable inflammation, necrosis or infection resulted from spinal

cord window chamber (SCWC) implantation.

(MP4)

Spinal Cord Window Chamber for Multimodal Imaging

PLOS ONE | www.plosone.org 12 March 2013 | Volume 8 | Issue 3 | e58081

261

Video S2 Photoacoustic and Power Doppler imaging. Co-

registered power Doppler and oxygen saturation (sO2) measure-

ments of the spinal cord. Three-dimensional power Doppler image

of the spinal cord vasculature shown in orange demonstrates the

ability to image multiple vascular structures within the spinal cord.

Longitudinal section of the cord is illustrated by the structural

ultrasound image and overlaid photoacoustic image. Color bar

indicates the relative sO2 level of the vasculature. Imaging was

performed while the mouse was breathing 100% oxygen mixed

with 2% isoflurane.

(WMV)

Video S3 Spinal cord O2 saturation monitoring by photoacous-

tic imaging. Two-dimensional cross-section of the spinal cord

within the window chamber. Ultrasound structural image (left)

shows the outline of the window chamber as well as the artificial

dura that cover the spinal cord. The rectangle indicates the region

where photoacoustic image was acquired, and the circular region

of interest indicates the area that photoacoustic signal intensity was

measured. Photoacoustic image (right) displays the spinal cord

vasculature. Color bar indicates the relative oxygenation level of

the vasculature, and the scale bar illustrates the depth of imaging

from the transducer head. The animal’s anaesthetic mixture was

shifted from 100% to 7% oxygen for 1 minute, which corresponds

to the frame 58 to 103 (out of total 248 frames acquired) in this

video.

(AVI)

Video S4 3D OCT. Reconstructed three-dimensional structural

OCT image acquired over 2.5 mm63 mm regions of the cord

within the window chamber. Arterial spinal vein and the spinal

cord structure can be seen throughout the region of interest.

OCT = optical coherence tomography.

(MPG)

Video S5 3D OCT. The same reconstructed three-dimensional

structural OCT image as in Video S4 was made transparent for

better visualization to highlight the structure of the spinal cord.

OCT = optical coherence tomography.

(MPG)

Acknowledgments

The authors would like to thank Jiachuan Bu for his assistance with animal

care. We would also like to thank Dr. Annie Huang’s lab for providing the

WW426 cell line. The views expressed in this manuscript do not necessarily

reflect those of the Ontario Ministry of Health and Long Term Care

(OMHLTC). Dr. DaCosta holds a Cancer Care Ontario Research Chair

in Cancer Imaging.

Author Contributions

Designed the window chambers: JIS JDM. Conceived and designed the

experiments: RSD. Performed the experiments: SAF YC AM LC JDM JIS

SS PL KB GZ. Analyzed the data: SAF YC AM LC SS. Contributed

reagents/materials/analysis tools: AV GZ MGF RSD. Wrote the paper:

SAF YC AM LC RSD.

References

1. Stroman PW (2005) Magnetic Resonance Imaging of Neuronal Function in the

Spinal Cord: Spinal fMRI. Clinical Medicine & Research 3: 146–156.

2. Braun IF, Raghavendra BN, Kricheff II (1983) Spinal cord imaging using real-

time high-resolution ultrasound. Radiology 147: 459–465.

3. Moseley ME, Cohen Y, Kucharczyk J, Mintorovitch J, Asgari HS, et al. (1990)

Diffusion-weighted MR imaging of anisotropic water diffusion in cat central

nervous system. Radiology 176: 439–445.

4. McAfee PC, Bohlman HH, Han JS, Salvagno RT (1986) Comparison of

Nuclear Magnetic Resonance Imaging and Computed Tomography in the

Diagnosis of Upper Cervical Spinal Cord Compression. Spine 11: 295–304.

5. Tench CR, Morgan PS, Jaspan T, Auer DP, Constantinescu CS (2005) Spinal

Cord Imaging in Multiple Sclerosis. Journal of Neuroimaging 15: 94S–102S.

6. Madi S, Flanders AE, Vinitski S, Herbison GJ, Nissanov J (2001) Functional MR

Imaging of the Human Cervical Spinal Cord. American Journal of

Neuroradiology 22: 1768–1774.

7. Misgeld T, Kerschensteiner M (2006) In vivo imaging of the diseased nervous

system. Nat Rev Neurosci 7: 449–463.

8. Dommisse GF (1974) The Blood Supply of the Spinal Cord: A Critical Vascular

Zone in Spinal Surgery. J Bone Joint Surg Br 56-B: 225–235.

9. Marcus M, Heistad D, Ehrhardt J, Abboud F (1977) Regulation of total and

regional spinal cord blood flow. Circulation Research 41: 128–134.

10. Kerschensteiner M, Schwab ME, Lichtman JW, Misgeld T (2005) In vivo

imaging of axonal degeneration and regeneration in the injured spinal cord. Nat

Med 11: 572–577.

11. Davalos D, Lee JK, Smith WB, Brinkman B, Ellisman MH, et al. (2008) Stable

in vivo imaging of densely populated glia, axons and blood vessels in the mouse

spinal cord using two-photon microscopy. Journal of Neuroscience Methods

169: 1–7.

12. Johannssen HC, Helmchen F (2010) In vivo Ca2+ imaging of dorsal horn

neuronal populations in mouse spinal cord. The Journal of Physiology 588:

3397–3402.

13. Kim JV, Jiang N, Tadokoro CE, Liu L, Ransohoff RM, et al. (2010) Two-

photon laser scanning microscopy imaging of intact spinal cord and cerebral

cortex reveals requirement for CXCR6 and neuroinflammation in immune cell

infiltration of cortical injury sites. Journal of Immunological Methods 352: 89–

100.

14. Dray C, Rougon Gv, Debarbieux F (2009) Quantitative analysis by in vivo

imaging of the dynamics of vascular and axonal networks in injured mouse

spinal cord. Proceedings of the National Academy of Sciences 106: 9459–9464.

15. Cadotte DW, Mariampillai A, Cadotte A, Lee KKC, Kiehl T-R, et al. (2012)

Speckle variance optical coherence tomography of the rodent spinal cord:

in vivo feasibility. Biomed Opt Express 3: 911–919.

16. Farrar MJ, Bernstein IM, Schlafer DH, Cleland TA, Fetcho JR, et al. (2012)

Chronic in vivo imaging in the mouse spinal cord using an implanted chamber.

Nature Methods advance online publication.

17. Fenrich KK, Weber P, Hocine M, Zalc M, Rougon G, et al. (2012) Long-term

in vivo imaging of normal and pathological mouse spinal cord with sub-cellular

resolution using implanted glass windows. The Journal of Physiology.

18. Xu M, Wang LV (2006) Photoacoustic imaging in biomedicine. Review of

Scientific Instruments 77: 041101–041122.

19. Wang X, Xie X, Ku G, Wang LV, Stoica G (2006) Noninvasive imaging of

hemoglobin concentration and oxygenation in the rat brain using high-

resolution photoacoustic tomography. Journal of Biomedical Optics 11:

024015–024019.

20. Ntziachristos V (2010) Going deeper than microscopy: the optical imaging

frontier in biology. Nat Meth 7: 603–614.

21. Shtoyerman E, Arieli A, Slovin H, Vanzetta I, Grinvald A (2000) Long-Term

Optical Imaging and Spectroscopy Reveal Mechanisms Underlying the Intrinsic

Signal and Stability of Cortical Maps in V1 of Behaving Monkeys. The Journal

of Neuroscience 20: 8111–8121.

22. Clarkson R, Lindsay PE, Ansell S, Wilson G, Jelveh S, et al. (2011)

Characterization of image quality and image-guidance performance of a pre-

clinical microirradiator. Medical Physics 38: 845–856.

23. Imai Y, Ibata I, Ito D, Ohsawa K, Kohsaka S (1996) A Novel Geneiba1in the

Major Histocompatibility Complex Class III Region Encoding an EF Hand

Protein Expressed in a Monocytic Lineage. Biochemical and Biophysical

Research Communications 224: 855–862.

24. Sasaki Y, Ohsawa K, Kanazawa H, Kohsaka S, Imai Y (2001) Iba1 Is an Actin-

Cross-Linking Protein in Macrophages/Microglia. Biochemical and Biophysical

Research Communications 286: 292–297.

25. Yu W, Fehlings M (2011) Fas/FasL-mediated apoptosis and inflammation are

key features of acute human spinal cord injury: implications for translational,

clinical application. Acta Neuropathologica 122: 747–761.

26. Popovich PG, Wei P, Stokes BT (1997) Cellular inflammatory response after

spinal cord injury in sprague-dawley and lewis rats. The Journal of Comparative

Neurology 377: 443–464.

27. Donnelly DJ, Popovich PG (2008) Inflammation and its role in neuroprotection,

axonal regeneration and functional recovery after spinal cord injury.

Experimental Neurology 209: 378–388.

28. Powers BE, Beck ER, Gillette EL, Gould DH, LeCouter RA (1992) Pathology of

radiation injury to the canine spinal cord. International Journal of Radiation

Oncology*Biology*Physics 23: 539–549.

29. Hornsey S, Myers R, Jenkinson T (1990) The reduction of radiation damage to

the spinal cord by post-irradiation administration of vasoactive drugs. In-

ternational Journal of Radiation Oncology*Biology*Physics 18: 1437–1442.

30. Siegal T, Pfeffer MR (1995) Radiation-induced changes in the profile of spinal

cord serotonin, prostaglandin synthesis, and vascular permeability. International

Journal of Radiation Oncology*Biology*Physics 31: 57–64.

31. Bilgen M, Al-Hafez B (2006) Comparison of spinal vasculature in mouse and rat:

investigations using MR angiography. Neuroanatomy 5: 12–16.

Spinal Cord Window Chamber for Multimodal Imaging

PLOS ONE | www.plosone.org 13 March 2013 | Volume 8 | Issue 3 | e58081

262

32. Nix W, Capra N, Erdmann W, Halsey J (1976) Comparison of vascular

reactivity in spinal cord and brain. Stroke 7: 560–563.

33. Anderson CR, Ashwell KWS, Collewijn H, Conta A, Harvey A, et al. (2009)

The Spinal Cord: A Christopher and Dana Reeve Foundation Text and Atlas;

Charles W, George P, Gulgun KayaliogluA2 - Charles Watson GP, Gulgun K,

editors. San Diego: Academic Press. v p.

34. Sinescu C, Popa F, Grigorean V, Onose G, Sandu A, et al. (2010 ) Molecular

basis of vascular events following spinal cord injury. Journal of Medicine and

Life 3: 254–261.

35. Sahgal A, Bilsky M, Chang EL, Ma L, Yamada Y, et al. (2011) Stereotactic body

radiotherapy for spinal metastases: current status, with a focus on its application

in the postoperative patient. Journal of Neurosurgery: Spine 14: 151–166.

36. Kirkpatrick JP, van der Kogel AJ, Schultheiss TE (2010) Radiation Dose-

Volume Effects in the Spinal Cord. International Journal of Radiation

Oncology*Biology*Physics 76: S42–S49.

37. van der Kogel AJ (1993) Dose-volume effects in the spinal cord. Radiotherapy

and Oncology 29: 105–109.

38. Liu TW, Akens MK, Chen J, Wise-Milestone L, Wilson BC, et al. (2011)

Imaging of Specific Activation of Photodynamic Molecular Beacons in Breast

Cancer Vertebral Metastases. Bioconjugate Chemistry 22: 1021–1030.

39. Hsu W, Mohyeldin A, Shah SR, Gokaslan ZL, Quinones-Hinojosa A (2012)

Role of Cancer Stem Cells in Spine Tumors: Review of Current Literature.Neurosurgery. E-pub ahead of pr int . doi :10.1227/NEU.1220-

b1013e3182532e3182571.

40. Bhavna Y, Erturk A, Hellal F, Nadrigny F, Hurtado A, et al. (2009) ChronicallyCNS-Injured Adult Sensory Neurons Gain Regenerative Competence upon

a Lesion of Their Peripheral Axon. Current Biology 19: 930–936.41. Dibaj P, Nadrigny F, Steffens H, Scheller A, Hirrlinger J, et al. (2010) NO

mediates microglial response to acute spinal cord injury under ATP control

in vivo. Glia 58: 1133–1144.42. Fehlings M, Vawda R (2011) Cellular Treatments for Spinal Cord Injury: The

Time is Right for Clinical Trials. Neurotherapeutics 8: 704–720.43. Thuret S, Moon LDF, Gage FH (2006) Therapeutic interventions after spinal

cord injury. Nat Rev Neurosci 7: 628–643.44. Byrnes KR, Fricke ST, Faden AI (2010) Neuropathological differences between

rats and mice after spinal cord injury. Journal of Magnetic Resonance Imaging

32: 836–846.45. Lassmann H (2005) Multiple Sclerosis Pathology: Evolution of Pathogenetic

Concepts. Brain Pathology 15: 217–222.46. Rowland LP, Shneider NA (2001) Amyotrophic Lateral Sclerosis. New England

Journal of Medicine 344: 1688–1700.

Spinal Cord Window Chamber for Multimodal Imaging

PLOS ONE | www.plosone.org 14 March 2013 | Volume 8 | Issue 3 | e58081

263