UNIVERSITY of CALIFORNIA Los Angeles Characterizing

UNIVERSITY OF CALIFORNIA

Los Angeles

Characterizing Neurotrauma and Astroglial Injury Biomarkers

by Proteomics and Mass Spectrometry

A dissertation submitted in partial satisfaction

of the requirements for the degree Doctor of Philosophy

in Biochemistry and Molecular Biology

Sean Shen

2016

Sean Shen

2016 ABSTRACT OF THE DISSERTATION

Characterizing Neurotrauma and Astroglial Injury Biomarkers

by Proteomics and Mass Spectrometry

Sean Shen

Doctor of Philosophy in Biochemistry and Molecular Biology

University of California, Los Angeles, 2016

Professor Joseph Ambrose Loo, Chair

Neurotraumatic injury has long been a leading cause of death and disability worldwide. Recently, the debilitating long term effects of chronic, mild traumatic brain injuries (TBIs) have gained increased public attention. In order to protect individuals most at risk (e.g. military personnel and athletes) from such injuries, improved diagnostics in the form of a biomarker panel capable of rapidly and sensitively detecting mild TBIs are needed. Despite the large number of TBI biomarker studies in the literature, the development of a clinically relevant protein signature remains elusive.

In contrast to diseases with singular mechanistic dysfunction, neurotrauma is characterized by the disruption to multiple cellular pathways that contribute to the sequelae of secondary pathophysiology that determines patient outcome and recovery.

This complexity has been a confounding factor in the identification of effective biomarkers.

In an effort to circumvent this hurdle, our group implemented a central nervous system ii

(CNS) specific cell injury model to examine preferentially released injury related proteins as candidate diagnostics.

Comparative analysis of a TBI CSF proteome and preferentially released proteins from our injury model revealed a panel of astroglial injury related candidate biomarkers including aldolase C (ALDOC), brain lipid binding protein (BLBP), glutamine synthetase

(GS), astrocytic phosphoprotein PEA15 (PEA15), and glial fibrillary acidic protein (GFAP) and its trauma-generated breakdown products (BDPs). Immediate and robust release of

ALDOC, BLBP, and PEA15 were associated more with acute cell wounding than cell death observed after biomechanical injury. In contrast, GFAP release correlated primarily with cell death. The sensitivity and selectivity of our biomarkers for neurotrauma were evaluated in human TBI and Yucatan swine spinal cord injury (SCI) CSF samples.

Verification studies demonstrated the ability of our astroglial biomarker panel to differentiate injury from non-injury with elevated signals detectable an hour after injury.

Additionally, differential CSF concentration kinetics were observed over a 1-week period post-injury indicative of a long diagnostic window. CSF concentration of biomarkers

GFAP, ALDOC, and BLBP correlated strongly with the extent of tissue loss after SCI at 7 days. Taken together, our data demonstrates the successful application of proteomics to the identification and verification of new neurotrauma biomarkers that exhibit potential for not only detecting but defining injury severity.

iii

This dissertation of Sean Shen is approved.

Gal Bitan

Jorge Torres

Joseph Ambrose Loo, Committee Chair

University of California, Los Angeles

2016

TABLE OF CONTENTS

CHAPTER 1: TRAUMATIC BRAIN INJURY – CLINICAL AND MOLECULAR PATHOLOGIES ...... 1 1.1 INTRODUCTION ...... 1 1.2 CLASSIFICATON OF TRAUMATIC BRAIN INJURY ...... 3 1.3 CLINICAL PATHOLOGIES OF NEUROTRAUMA ...... 6 1.4 THE MOLECULAR PATHOPHYSIOLOGY OF NEUROTRAUMA ...... 10 1.5 CONCLUSION ...... 18 1.6 REFERENCES ...... 20 CHAPTER 2: ADDRESSING THE NEEDS OF TRAUMATIC BRAIN INJURY WITH CLINICAL PROTEOMICS ...... 30 2.1 INTRODUCTION ...... 30 2.2 DISCUSSION ...... 32 2.3 CONCLUSIONS ...... 53 2.4 FIGURES ...... 54 2.5 TABLES ...... 55 2.6 REFERENCES ...... 56 CHAPTER 3: NEW ASTROGLIAL INJURY DEFINED BIOMARKERS FOR NEUROTRAUMA ASSESSMENT ...... 78 3.1 INTRODUCTION ...... 78 3.2 RESULTS ...... 81 2.3 DISCUSSION ...... 93 3.4 METHODS ...... 102 2.5 FIGURES ...... 112 3.6 TABLES ...... 129 3.7 SUPPLEMENTAL FIGURES ...... 158 3.8 REFERENCES ...... 174 CHAPTER 4: ASSESSMENT OF ASTROGLIAL INJURY DEFINED BIOMARKERS IN SPINAL CORD INJURY ...... 198 4.1 INTRODUCTION ...... 198 4.2 RESULTS ...... 201 4.3 DISCUSSION ...... 207

4.4 METHODS ...... 209 4.5 FIGURES ...... 213 4.6 TABLES ...... 233 4.7 REFERENCES ...... 239 CHAPTER 5: CHARACTERIZING THE PREFERENTIAL RELEASE OF PROTEIN SUBPOPULATIONS BY INJURED ASTROCYTES ...... 243 5.1 INTRODUCTION ...... 243 5.2 RESULTS ...... 245 5.3 DISCUSSION ...... 250 5.4 METHODS ...... 255 5.5 FIGURES ...... 259 5.6 TABLES ...... 269 5.7 REFERENCES ...... 312 CHAPTER 6: FUTURE DIRECTIONS FOR SPINAL CORD AND HEAD TRAUMA ... 320 6.1 INTRODUCTION ...... 320 6.2 RESULTS ...... 321 6.3 DISCUSSION ...... 325 6.4 CONCLUDING REMARKS ...... 328 6.5 METHODS ...... 333 6.6 FIGURES ...... 335 6.7 TABLES ...... 337 6.8 REFERENCES ...... 352

LIST OF FIGURES AND TABLES

Figure 2.1: ...... 54 Table 2.1: ...... 55 Figure 3.1: ...... 112 Figure 3.2: ...... 113 Figure 3.3: ...... 116 Figure 3.4: ...... 118 Figure 3.5: ...... 120 Figure 3.6: ...... 123 Figure 3.7: ...... 125 Figure 3.8: ...... 128 Table 3.1: ...... 130 Table 3.2: ...... 131 Table 3.3: ...... 133 Table 3.4: ...... 148 Table 3.5: ...... 149 Table 3.6: ...... 152 Table 3.7: ...... 154 Table 3.8: ...... 155 Table 3.9: ...... 157 S3.1: ...... 158 S3.2: ...... 160 S3.3: ...... 161 S3.4: ...... 162 S3.5: ...... 163 S3.6: ...... 164 S3.7: ...... 165 S3.8: ...... 167 S3.9: ...... 168 S3.10: ...... 170 S3.11: ...... 171 S3.12: ...... 173 Figure 4.1: ...... 213 Figure 4.2: ...... 214 Figure 4.3: ...... 215 Figure 4.4: ...... 216 Figure 4.5: ...... 218 Figure 4.6: ...... 219 Figure 4.7: ...... 220 Figure 4.8: ...... 222 Figure 4.9: ...... 224 Figure 4.10: ...... 225 Figure 4.11: ...... 226 Figure 4.12: ...... 228 Figure 4.13: ...... 230 Figure 4.14: ...... 232 vii

Table 4.1: ...... 233 Table 4.2: ...... 235 Table 4.3: ...... 236 Table 4.4: ...... 238 Figure 5.1: ...... 259 Figure 5.2: ...... 261 Figure 5.3: ...... 262 Figure 5.4: ...... 263 Figure 5.5: ...... 265 Figure 5.6: ...... 266 Figure 5.7: ...... 267 Figure 5.8: ...... 268 Table 5.1: ...... 269 Table 5.2: ...... 271 Table 5.3: ...... 273 Table 5.4: ...... 274 Table 5.5: ...... 276 Table 5.6: ...... 289 Table 5.7: ...... 299 Table 5.8: ...... 311 Figure 6.1: ...... 335 Figure 6.2: ...... 336 Table 6.1: ...... 338 Table 6.2: ...... 339 Table 6.3: ...... 341 Table 6.4: ...... 344 Table 6.5: ...... 351

viii

DEDICATION

I would like to dedicate this work to my mother (Xinfang), father (Zhongnan), and to all

my family and friends who have supported me throughout this long journey.

To my WindRose family, thanks for your constant encouragement.

To the memory of Daniel Cho, my brother from another mother.

ACKNOWLEDGEMENTS

I would like to thank my mentor, Professor Joseph A. Loo, for giving me this research opportunity and for his constant support and guidance throughout my graduate studies. I am forever thankful to Joe for taking a chance on me and trusting me to not only work with top of the line mass spectrometers but also to be so closely involved in the installation of new instruments. I will remember the encouraging, insightful, and kind demeanor that Joe conducts himself as something to aspire toward.

I would like to thank Ina-Beate Wanner for the opportunity to work on her traumatic brain injury project. Her words of encouragement, criticisms, and patience have pushed me throughout my graduate training. I have met few people in my life with as much passion and dedication to her work. I would like to also thank members of the Wanner lab

Julia Halford and Jacklynn Levine for all their help on the neurobiology side of the project and collaborative discussions.

Additionally, I would like to thank Rachel Loo for all the knowledge and expertise she has shared with me over the years on sample preparation and biochemical methods.

To past and present Loo lab members, I would like to extend my appreciation for your support, both professional and personal. I would especially like to thank Carly

Ferguson, Hong Nguyen, Pete Wongkongkathep, Huilin Li, Reid O’Brien Johnson, Dyna

Shirasaki, and my undergraduate student Eric Wang. I would also like to express my sincere gratitude to my good friends Keith Cheung and Subhajit Poddar for their advice and input throughout my graduate work.

Finally, I would like to thank my committee members, Professors Gal Bitan,

Catherine Clarke, Jorge Torres, James Wohlschlegel, and Joe Loo for their guidance and advice.

Chapter 2 of this dissertation is version of a version of S. Shen, R. R. O. Loo, I. B.

Wanner, J. A. Loo, Addressing the needs of traumatic brain injury with clinical proteomics.

Clinical proteomics 11, 1-13 (2014), reprinted with permissions. I would like to acknowledge my co-authors Rachel R. Ogorzalek Loo, Ina B. Wanner, and Joseph A.

Loo.

Chapter 3 of this dissertation is a manuscript in preparation entitled “New Astroglial

Injury Defined Biomarkers for Neurotrauma Assessment.” I would like to acknowledge Ina

B. Wanner, Julia Halford, Kyohei Itamura, and Jacklynn Levine their work on cellular characterization and immunoblot studies in addition to data analysis, assembly, and figure preparation; Gregg Czerwieniec for work on MRM-MS and mass spectrometry studies;

Dalton Dietrich, Paul Vespa, David Hovda for their input on traumatic brain injury (TBI) pathophysiology; and Ross Bullock, Paul Vespa, Thomas Glenn, and Stefania Mondella for providing TBI patient biofluid samples and input on clinical analysis.

I would like to acknowledge Ina B. Wanner, Julia Halford, Kyohei Itamura, and

Jacklynn Levine for their work on preparing stretch injured astrocytes, immunoblot analysis of swine SCI CSF samples, and immunohistology of spinal cords described in

Chapters 4 and 5 of this dissertation.

I would like to acknowledge our Department of Defense collaborators Rachel

Kinsler, Andrew Mayer, Jonathan DeShaw, and Salam Rahmatalla for their work in developing and executing the spinal cord injury model, cerebrospinal fluid (CSF) collection, post-injury transportation, and surgery which contributed to the elucidation of astroglial injury defined biomarkers in assessing swine spinal cord injury and characterization of swine SCI CSF proteomics in Chapters 4 and 6 respectively.

xii

I would also like to acknowledge my undergraduate student Eric Wang for his contributions to sample preparation for the work presented in Chapter 4.

xiii

VITA

2003 – 2007 Bachelors of Arts, University of California, Berkeley

Molecular Cell Biology

2007 – 2009 Research Associate I, WindRose Analytica

2009 – 2010 Research Associate II, Ajinomoto Althea

2011 – 2012 Teaching Assistant, UCLA

2014 - 2015 Technology Fellow, UCLA Office of Intellectual Property

2015 Excellence in Biochemical Research Fellowship

xiv

PUBLICATIONS

Shen, S., Loo, R. R. O., Wanner, I.-B. & Loo, J. A. Addressing the Needs of Traumatic Brain Injury with Clinical Proteomics. Clinical Proteomics 2014, 11:11.

Dzialo, M.C., Travaglini, K.J., Shen, S., Loo, J.A., Clarke, S.G. A New Type of Protein Lysine Methyltransferase Trimethylates Lys-79 of Elongation Factor 1A. Biochemical and Biophysical Research Communications 2014, 455:382-389

Dzialo, M.C., Travaglini, K.J., Shen, S., Roy, K., Chanfreau, G.F., Loo, J.A., Clarke, S.G. Translational Roles of Elongation Factor 2 Protein Lysine Methylation. Journal of Biological Chemistry 2014, 289:30511-30524

Buehler, D.*, Marsden, M.*, Shen, S., Toso, D. B., Wu, X., Loo, J. A., Zhou, Z. H., Kickhoefer, V. A., Wender, P. A., Zack, J. A., & Rome, L. H. Bioengineered Vaults: Self- Assembling Protein Shell-Lipophilic Core Nanoparticles for Drug Delivery. ACS Nano 2014, 8:7723-7732

PRESENTATIONS

Shen, S., Halford, J., Wanner, I.B., Loo, J.A. Characterizing Traumatic Brain Injury with New Astroglial Injury Biomarkers Measured by Targeted MS. American Society of Mass Spectrometry Annual Conference. San Antonio, TX. June 2016

Shen, S., Itamura, K., Halford, J., Wanner, I.B., Loo, J.A. Measuring Acute Traumatic Brain Injury Biomarkers by Targeted Mass Spectrometry. American Society of Mass Spectrometry Annual Conference. St. Louis, MO. June 2015

Shen, S., Wanner, I.B., Loo, J.A. Discovery and Verification of Neurotrauma Markers by High Mass Accuracy/High Resolution Mass Spectrometry. American Society of Mass Spectrometry Annual Conference. Baltimore, MA. June 2014

Shen, S., Wanner, I.B., Czerwieniec, G., Loo, J.A. Selection and Quantification of Neurotrauma Markers by Mass Spectrometry. American Society of Mass Spectrometry Annual Conference. Minneapolis, MN. June 2013

Shen, S., Ferguson, C., Loo, R.R.O., Loo, J.A. Highly Multiple Charging with 2- nitrophloronolgluncinol by MALDI Time-of-Flight Mass Spectrometry. American Society of Mass Spectrometry Annual Conference. Vancouver, CA. May 2012

CHAPTER 1: TRAUMATIC BRAIN INJURY – CLINICAL AND MOLECULAR PATHOLOGIES

1.1 INTRODUCTION

Impact and Healthcare Significance of Traumatic Brain Injury (TBI)

Neurotrauma to the central nervous system (CNS) is a serious public health worldwide. Most commonly, neurotrauma is experienced in the form a traumatic brain injury or TBI. Examining the US alone, TBI is most common in infants and toddlers, adolescents and the elderly (1). The US National Institute of Neurological Disorders and

Stroke estimates that 2.5-6.5 million Americans have had one or multiple TBIs. In the US military there were over 212,000 service men and women diagnosed with some form of

TBI between January 2000-May 2011, roughly accounting for one-third of all injured US soldiers, making TBI the signature injury of the wars in Iraq and Afghanistan compared to past wars (2). TBI contributes to over one third of all injury-related deaths, yet 75-90% of all brain trauma cases are considered to be mild TBI (mTBI), many without visible wounds that often are undiagnosed (3). The documented long term disability associated with repeated head trauma coupled with inadequate diagnostic measures highlight the immediate need for increased understanding of TBI pathology and how to treat it. Better diagnostic tools are needed to detect head injuries, especially mTBI, as well as to confirm and monitor the severity of TBI in order to determine the best course of action acutely and later post-injury. This is of special urgency for military personnel and athletes of all kinds who are most at risk for repeated head injury. This introduction presents TBI as a biomechanical injury and characterizes the clinical pathologies and their underlying molecular processes. Chapter 2 will discuss some of the challenges to discovering new 1

biomarkers. In Chapters 3-6, results of proteomic efforts in this field will be discussed as they relate to insights that can be gleamed for future study and development of neurotrauma diagnostics and therapeutics.

1.2 CLASSIFICATON OF TRAUMATIC BRAIN INJURY

TBI is a Biomechanical Injury

Traumatic brain injury is defined as focal or diffuse brain damage from external trauma. The principle mechanisms of focal brain damage include contusion, laceration, and intracranial hemorrhage. Diffuse brain injury occurs from abrupt acceleration and deceleration type injuries that result in diffuse axonal damage and brain edema. While a

TBI is initiated from one of the above two types of primary injury, assessment and treatment is further complicated by the onset of secondary non-mechanical damage (4-

6). No treatment exists for primary injuries outside of preventative measures. Secondary pathologies ranging from ischemic events to edema, however, are sensitive to therapeutic interventions. Proper management of secondary sequelae is essential to positive long term patient outcome and brain function.

Classification of TBI severity

Classification of the severity of traumatic brain injury is of clinical interest as it directly affects the type of acute and post-acute medical care administered. Typically, TBI severity is determined based on single indicators such as the Glasgow Coma Scale

(GCS), duration of post-traumatic amnesia, and loss of consciousness. The GCS is a neurocognitive examination of eye function, response to verbal commands, and motor function graded on a scale of 1-14 where scores of 13 or higher correlate with a mild brain injury, 9-12 to a moderate injury, and 8 or lower a severe injury (7). And while these measures correlate with severity and outcome, each may be influenced by indirect factors. Early sedation and patient intoxication have both been demonstrated to have

suppressive effects on GCS values (8, 9). Classification schemes that combine single indicators of GCS, loss of consciousness, amnesia as well as a myriad of other clinical criteria such as patient survival, presentation of hematoma/hemorrhage, and patient reported symptoms have demonstrated effectiveness ex post facto (10). However, classification after the fact offers little benefit to patient treatment and outcome. Coupled with the frequent lack of complete documentation of severity indicators, standardization of classification for clinical and research purposes necessitates a simple yet unencumbered diagnostic.

Concussive Neurotrauma

Perhaps the most impactful type of TBI is concussive injury. While concussive injuries are classified as a mild TBI (mTBI), they nonetheless possess the potential to affect a patient’s long term mental status. Studies estimate that anywhere between 1.5 and 4 million US athletes suffer a concussive mTBI annually (11, 12). Early symptoms of a concussion include but are not limited to changes in behavior, loss of emotional control, impairments to memory/attention, headache, and in rare cases catastrophic brain injury known as the second impact syndrome (13). Second impact syndrome (SIS) is defined as when a patient sustains a head injury, most commonly a mTBI, and subsequently endures a second injury before the symptoms of the first have fully cleared, resulting in catastrophic and typically fatal brain swelling (14, 15). However, it is scientifically unclear whether it is the repeated injury or delayed onset of cerebral swelling form the initial blow to the head that is responsible for these rare occurrences (16). Despite this controversy, the anecdotal reports of SIS have brought attention to the dangers of chronic head

trauma. Studies to this effect have identified a positive correlation of greater symptom severity, time to recover, and the earlier onset of age associated cognitive decline (also known as chronic traumatic encephalopathy (CTE)) in patients with a history of repeated concussions. These findings refuted the previous perception of concussive sports injuries as benign. Public awareness and outcry from this new information has been instrumental in changing how our sports medicine professionals treat and manage those individuals most at risk. Major sports organizations, most prominently the National Football League

(NFL), have been at the forefront of this backlash resulting in the institution of new protocols and procedures to ensure player safety.

1.3 CLINICAL PATHOLOGIES OF NEUROTRAUMA

Diffuse Axonal Injury

Biomechanical stretching of neuronal axons causes membrane disruption and depolarization. This increase in axolemmal permeability has been shown to persist for 6 hours post-injury (17, 18) with the influx of calcium. As a result, neurofilaments undergo compaction by calcium activated calpain proteolysis or neurofilament phosphorylation, leading to loss of stability and breakdown (19, 20). As axons begin to develop abnormalities and breaks, an accumulation of organelles occurs at the site of damage due to the continued transport along intact segments. Signs of axotomy or axonal severing can be observed as early as 4 hours post-injury and persist for days to weeks

(21).

Edema and Elevated Intracranial Pressure

Brain edema, or swelling, is a critical pathophysiology resulting from neurotrauma

(TBI, ischemia, etc.). Brain edema is defined as the abnormal accumulation of fluid within the parenchyma. In most organ systems the parenchyma refers to structural and connective tissues. In the brain, however, the parenchyma is comprised of the functional tissue consisting of neurons and glial cells. Edema is categorized as either vasogenic or cytotoxic (22). Vasogenic edema occurs when excess fluid accumulates in the brain around cells, usually originating from blood vessels. This is believed to occur following a traumatic compromise of the blood brain barrier (23). Swelling resulting from the accumulation of fluid within cells is classified as cytotoxic edema. Cytotoxic edema most commonly results from ischemic events, where inadequate oxygen and glucose content

impede cell survival. Both forms of edema are experienced in TBI, making treatment of resulting neuropathologies challenging.

Edema is well documented in TBI to raise intracranial pressure (ICP), a secondary pathology of the initial mechanical insult that is frequently associated with death and poor prognosis among TBI survivors (24). Assessing the extent of swelling by computed tomography (CT) scans (25) shortly after injury has demonstrated high correlation between patient outcome to severity of brain swelling. In severe cases, patient mortality may occur in as little at 36 hours despite aggressive clinical interventions (22). A study of the proteomic alterations may reveal trends between protein levels and patient ICP. Such a multidisciplinary study could identify signature proteins that may act as a less invasive surrogate measure for ICP, traditionally monitored by insertion of a catheter into different areas of the brain. Furthermore, potential surrogate protein markers could be monitored during treatment (both clinical and experimentally) as an indication of the modality’s effectiveness.

Neuroinflammation

Following the onset of cell death, a complex interplay of immunological and inflammatory responses is observed in neurotrauma. Both the primary insult and resulting secondary sequelae activate cellular mediators ranging from proinflammatory cytokines, prostaglandins, and components of the complement system. These mediators then induce chemokines and adhesion molecules, recruit immune cells, and activate glial cells

(26). While, many of these components of the inflammatory response contribute to acute

and chronic neurological detriment, the immune response is also responsible for long- term repair and recovery post-trauma.

Delayed CNS injury is another hallmark of neurotrauma with inflammatory response mediators implicated in the process. Upregulation of cellular adhesion molecules are responsible for tissue infiltration by leukocytes. These leukocytes are then responsible for the elimination of injured but also adjacent healthy tissues based on spreading depressions. This occurs on a time scale ranging from hours to weeks as astrocytes and microglial begin to synthesize the structural filament components of the neuroscar (26).

Acutely after injury proinflammatory enzymes tumor necrosis factor (TNFα), interleukin-1 (IL-1), and interleukin-6 (IL-6) are upregulated. IL-1, released immediately following CNS damage from activated glial cells, induces a variety of beneficial actions involved in restoring ionic balance through reduction in EAA glutamate release, enhancement of gamma-amino butyric acid (GABA), the primary neuroinhibitory signaling amino acid, and modulation of NMDA. IL-1 also upregulates the production of nitric oxide

(NO) which contributes both protective and neurotoxic effects (27). TNFα is a key mediator of tissue inflammation implicated in the development of an assortment of neurological conditions. TNF related signaling pathways function through two receptors, p55 and p75. While the functions of p75 in the brain are unknown, activation of p55 is responsible for the induction of apoptosis in the CNS. IL-6 is involved in various signaling pathways that lead to gene activation related to recovery processes (26, 28).

Modulation of these inflammatory mediators have shown positive experimental evidence for in the CNS damages at both acute and late (48h) time-points post-TBI.

Agents inhibiting TNFα improve short and long-term outcome in rats (29) while studies have shown IL-6 to have neuroprotective effects by increasing CNS healing (30).

Transgenic mice lacking complement C3 or C5 exhibit reduced secondary damage compared to control (26). While not necessarily specific to TBI neurotrauma, changes in neuroimmune responses are critical to treatment of neurotrauma sequelae.

Chemical/biochemical agents with the ability to modulate these immune responses may represent potential therapies in TBI management.

Cerebral Blood Flow

In the healthy brain, cerebral blood flow (CBF) is tightly coupled to cerebral glucose metabolism. However, post-trauma, cerebral blood flow is deregulated leading to a decoupling of blood and oxygen flow with cellular energy requirements. Experimental evidence in rat fluid procession models have shown a decrease in CBF by as much as

50% of normal levels in the post-traumatic state. This reduction in CBF limits the oxygen available to cerebral tissue necessary to meet the metabolic needs of injured and recovering cells in a damaging energy crisis (31, 32).

1.4 THE MOLECULAR PATHOPHYSIOLOGY OF NEUROTRAUMA

General Molecular Pathophysiology of TBI

Upon sustaining a TBI, mechanical shear and deformation forces initiate a complex cascade of neurochemical and metabolic events. These events begin with the impairment of cerebral blood flow (CBF) resulting in ischemic conditions (33, 34). The resulting anaerobic conditions and increased cellular metabolism results in an energy crisis, As energy dependent ion pumps fail, ionic balance is disrupted resulting in the indiscriminate release of excitatory amino acids (EEA) that only furthers ion imbalances and causes the activation of signaling pathways for cell death (4).

Ionic Imbalance and Neurotransmitter Release

Acutely following mechanical trauma to brain tissue, neuronal membranes become compromised, axons are stretched, and voltage-gated potassium channels are opened.

Increases in potassium extracellular potassium concentrations cause nonspecific axon depolarization and results in the indiscriminant release of excitatory amino acid (EEA) glutamate which further increases extracellular potassium concentrations through the activation of, N-methyl-D-aspartate (NMDA), α-amino-3-hydroxy-5-methyl-4- isoxazolpropionate (AMPA). This massive excitation of neurons is followed by neuronal suppression, a phenomenon known as spreading depression (35) where the sections of the brain undergo waves of electrophysiological hyperactivity followed by a waves of inactivity. Distinct from classical spreading depression, post-traumatic spreading depression affects both focal and diffuse areas of injury simultaneously. Early loss of

consciousness, memory loss, and confusion may be manifestations of a spreading depression state post-injury (32).

Under normal conditions, elevated extracellular potassium are absorbed by surrounding astroglial cells (36). Potassium released by neurons causes a passive influx of potassium into surrounding astrocytes. This causes an astrocyte depolarization and leads to current conduction along the cell and to cells coupled to them. As this potassium generated current is propagated to the endfeet of astrocytes which terminate on the surface of cerebral arterioles, potassium is siphoned from astrocyte feet onto their adjacent arteriole walls. The increase in potassium content in arterioles causes vasodilation and is important to the regulation of cerebral blood flow (36). While this process is sufficient in accommodating mild perturbations in extracellular potassium, it is unable to compensate for the larger levels of ionic imbalance generated from injury as post-traumatic astrocytes exhibit reduction/loss of inward potassium uptake and subsequent conduction (37). This loss of ionic homeostasis is likely involved in the impairments to learning and memory that patients experience after a TBI.

Neurons restore ionic balance by activating energy dependent sodium/potassium pumps. In post-TBI conditions however, energy stores are rapidly depleted resulting in rapid by inefficient anaerobic glycolysis acutely 30 minutes to 4 hours after injury in rats

(38). This increase in glycolysis known as hyper-metabolism results from diminished CBF and the disparity between cellular glucose supply and demand creates an energy crisis.

It is hypothesized that this energy deficient state is responsible for post-injury vulnerability where cerebral tissue is less equipped to respond adequately to subsequent injury leading to increased trauma severity and extended post-acute deficits (32). Additionally,

increased anaerobic production of energy results in an extracellular accumulation of lactate which contributes to acidosis, membrane compromise, and cerebral swelling (13).

Mitochondrial Dysfunction in TBI

In addition to increased potassium ion levels, calcium accumulation is also observed in the wake of neurotrauma (39). Elevated extracellular potassium ion concentrations in the post-traumatic state, triggers the unregulated release of excitatory amino acids that bind and activate NMDA receptors. Activated NMDA receptors create a pore that allows calcium ions to enter the cells. Calcium is key to the pathophysiology of trauma induced secondary sequela. When intracellular calcium increases above normal homeostatic levels, attack and digestion of cellular proteins, lipids and DNA occur as a result of the activation of proteases, lipases, and nucleases (40). As a result of increased calcium dependent enzymatic activation, cells are subject to an overproduction of neurotoxin free radicals, disruption of cytoskeletal organization, and or signaling cascades leading to cell death.

Neurons and glia cells respond to this increase in intracellular calcium by sequestering the excess within the mitochondria (41). Under normal conditions, calcium provides benefit to the mitochondrial by stimulating oxidation-phosphorylation and ATP synthesis. However, overloaded mitochondrial calcium has been shown to activate the production of reactive oxygen species. Increases in ROS can further modulate calcium dynamics by augmenting the calcium surge, thus generating a self-amplified loop of cellular damage through calcium dependent initiation of apoptosis or necrosis.

Additionally, calcium stimulates oxidative phosphorylation by allosteric activation of tricarboxylic acid (TCA) cycle enzymes leading to faster respiratory chain activity and increased oxygen consumption that is restricted under ischemic conditions experienced in the post-TBI state. Additionally, calcium concentrations in excess of physiological conditions disrupt the respiration process by increasing cytochrome c dislocation from the inner membrane through either competitive inhibition of negatively charged cadiolipin binding sites or activation of cyctochrome c release pathways (42).

Altered Glucose Metabolism in TBI

The human brain functions primary on glucose and the energy it generates through the glycolytic and tricarboxylic acid pathways. Alterations in cerebral glucose metabolism

(CMRglc) is a hallmark response to neurotrauma. Chemically labeled (18F-DG and 14C-

DG) glucose analogs have been used extensively in the study of glucose metabolism.

Once cleaved, these isotopes are trapped in cells allowing glucose accumulation to be monitored. Using autoradiographic visualization, rapid glucose uptake is observed acutely in the post-traumatic state followed by an extensive period of glucose metabolism depression (38). Immediately following head injury, large increases in cerebral glucose metabolism are observed. This increase has been shown to be attenuated by administration of kynurenic acid, an inhibitor of NMDA receptors involved in proliferation of cellular ionic imbalances (43). Consequently, this initial increase in glucose metabolism is believed to be in response to the higher cellular energy demands necessary to restore ionic homeostasis and neuronal membrane potential. This period of hyperglycolysis has been observed for up to 8 days in severe TBI patients (44). The described acute CMRglc

period is followed by a period of metabolic depression in both animal and human studies

(45) that correlate with the magnitude of injury severity. Consistent with experimental data, glucose metabolism rates in select regions of the brain – thalamus, cerebellum, and brain stem – showed significant positive correlation with levels of consciousness measured by the Glasgow Coma Scale (GCS) (46)

This delayed wave of glucose metabolism is believed to be the result of the contributions of changes in cerebral blood flow, defects in glucose transporter function, and or decreased metabolic demand for glucose. The rapid increase in glucose metabolism in the acute phase of injury correlates to increased consumption during a period of blood flow decline (41) generating an energy crisis. One explanation is that the increased energy burden quickly depletes glucose stores and in the presence of insufficient glucose replenishment from blood, glycolysis pathways are unable to keep up and CMRglc rates decline. However, experimental data proving this is so far unclear. In contrast, experimental studies rats showed no change to blood glucose after injury suggesting no substrate limitation (47).

A second reason is proposed to be related to decreases in neuronal glucose transporter GLUT1 (48) resulting in impaired glucose transport from blood to brain cells.

Hattori et al (46) demonstrated lower glucose accumulation in brain regions within contusion sites. It is possible that inhibition of glucose transport across the blood brain barrier is substantially affected in the post-TBI state.

Lastly, cells may experience a decreased metabolic demand as it prioritizes other repair related functions in response to trauma. As it relates to glycolytic processing, proton nuclear magnetic resonance (NMR) studies have uncovered increases in the amount of

glucose diverted to the synthesis of nucleic acid precursors in the acute time phase (3-24 hours) post-injury (49). This increase in DNA synthesis is likely a cellular response to both

DNA damage and upregulation of genes involved in repair and recovery pathways. In support of this, nicotinamide dinucleotide (NAD+), essential electron acceptors in the respiratory pathways, concentrations have been shown to decrease after injury (50). This can be explained by higher NAD+ consumption from DNA repair enzymes as Poly-ADP ribose polymerase (PARP) (51) in response to the elevated cellular levels of ROS. Thus, reductions in NAD+ levels may be responsible for glycolytic inhibition as the cell reorganizes its needs in the aftermath of traumatic injury.

Astrocytes and their Response to Injury

Astrocytes are an abundant class of glial cell in the central nervous system that provides both structural and functional support to neurons. In healthy tissue, astrocytes play crucial roles in functions related to energy provision, blood flow, regulation, maintenance of ionic balance, and neurotransmitter recycling (52, 53).

In healthy tissue, astrocytes regulate important and related functions between cerebral blood flow and the metabolic demand of neurons. Studies have demonstrated the ability of astrocytes to elicit bidirectional vasculature changes in adjacent blood vessels through activation of calcium sensitive signaling pathways (54). Astrocytes regulate changes in CBF in response to the metabolic needs of neurons. An increase in astrocytic anaerobic glycolysis is observed under reduced oxygen conditions, leading to increases in lactate release that result in increased vasodilation (55). Astrocytes also play crucial roles in maintaining ionic homeostasis through uptake and release of water in

response to neuropeptide signals and bidirectional aquaporin channels (56). Regulation of EAA glutamate is another central role of astrocytes in their neuronal interactions. EEAs are cleared from neuronal synapses by astrocytes via glutamate transporters, recycled back to glutamine, and then released and re-absorbed by neurons (57).

Perhaps most interesting, is the key role astrocytes play in response to injury.

Following CNS insult such as mechanical trauma, infection, ischemia, and neurodegenerative disease, astrocytes undergo a changes to molecular expression and morphology known collectively as reactive astrogliosis. Reactive astrocytes are characterized by increased expression of glial fibrillary acidic protein (GFAP) among other intermediate filament proteins involved in the hallmark star-like morphological change associated with cellular hypertrophy (52). New evidence has defined the mechanism of reactive astrogliosis to be a graded one. Changes to gene expression and intra/intercellular signaling are proportional to the severity of injury. Despite the appearance of hypertrophy, mild and moderate cases of astrogliosis are believed to be recoverable after injury resolution (53). In severe cases, higher activation of astrocytes is documented to result in the formation of a glial scar that creates a barrier that limits the spread of inflammation (58).

While reactive astrogliosis has traditionally been associated with the formation of a glial scar that inhibits axonal regeneration, new research has identified a myriad of beneficial and essential injury responses (59). Reactive astrocytes confer neuronal protection through the uptake of excitotoxic levels of glutamate that accompanies indiscriminant membrane depolarization (60-62). Ablation experiments have also implicated reactive astrocytes in limiting the infiltration of inflammatory cells, repair of the

blood brain barrier (63), protection against immune related demyelination (63), and reduction of hydrocephalus (60, 64). Molecular mediators of reactive astrogliosis are released by a number of CNS cells and while much is still unknown, considerable evidence suggest that different signaling mechanisms (STAT3, interleukin-6, leukemia inhibitory factor) may initiate functional changes proportionate to the extent of injury (62,

65, 66).

1.5 CONCLUSION

Traumatic brain injury is a major healthcare crisis for which there is currently inadequate diagnostic, let alone therapeutic, measures. In order to tackle this silent epidemic, we must first develop conclusive metrics for injury identification and classification for both clinical and research standardization. Currently, there is no clinically validated biofluid marker for TBI/neurotrauma despite the volume of TBI biomarker studies. And while the list of potential candidate biomarkers identified by proteomics is encouraging to the mission, it raises the question of which potential biomarkers should be prioritized for verification purposes. It is here that systems biology and an understanding of the underlying molecular mechanisms associated with observed clinical pathologies could provide an additional level of selection on top of biofluid abundance and tissue enrichment to aid researchers in discriminating the top candidates for validation studies.

While an effective biomarker does not necessarily require a direct relationship to the biology of injury, biomarkers with biological relevance to the resulting molecular sequelae have been shown to be the most promising. An example of such markers is

GFAP, an astroglial-specific intermediate filament, whose expression increases shortly after injury as astrocytes attempt to maintain homeostasis and promote recovery (67).

Mediators of this processes such as IL-1β may offer insights into potential therapeutic targets for future study (68). In our project, we have benefited from the understanding of calcium mediated calpain activation resulting from ionic imbalances in membrane compromised cells that lead to proteolytic breakdown products (69) that may offer additional nuance into the severity and progression of TBI. Other examples include

elevated post-trauma levels of proteins involved in free radical clearance (70), stress response (71), and immune response (26) associated with altered metabolic, signaling, and repair functions.

A more complete understanding of the mechanisms involved in injury response is also an essential tool for researchers to better design injury models to isolate specific responses within the complex web of neurotraumatic insult. In our research, this has helped us to select astrocytes as an in vitro model given their involvement in so many regulatory elements ranging from maintenance of ionic and fluid homeostasis to metabolism to blood flow. Alterations in signaling pathways, molecular pathologies of membrane permeability, and a graded reactivity lead us to believe that investigation of astroglial responses to injury may help to identify proteomic signatures of injury that are released proportionate to the severity of injury.

1.6 REFERENCES

1. M. Faul, L. Xu, M. M. Wald, V. G. Coronado, Traumatic Brain Injury in the United

States: Emergency Department Visits, Hospitalizations and Deaths 2002–2006.

(Centers for Disease Control and Prevention, National Center for Injury Prevention

and Control, Atlanta, GA, 2010).

2. J. E. Risdall, D. K. Menon, Traumatic brain injury. Phil Trans Royal Soc London,

Series B, Biol Sci 366, (2011).

3. W. A. Gordon, M. Brown, M. Sliwinski, M. R. Hibbard, N. Patti, M. J. Weiss, R.

Kalinsky, M. Sheerer, The enigma of "hidden" traumatic brain injury. J Head

Trauma Rehabil 13, (1998).

4. C. Werner, K. Engelhard, Pathophysiology of traumatic brain injury. British journal

of anaesthesia 99, 4-9 (2007); published online EpubJul (10.1093/bja/aem131).

5. L. F. Marshall, Head injury: recent past, present, and future. Neurosurgery 47, 546-

561 (2000).

6. J. Nortje, D. K. Menon, Traumatic brain injury: physiology, mechanisms, and

outcome. Current Opinion in Neurology 17, 711-718 (2004).

7. G. Teasdale, B. Jennett, Assessment of coma and impaired consciousness: a

practical scale. The Lancet 304, 81-84 (1974).

8. R. D. Zafonte, F. M. Hammond, N. R. Mann, D. L. Wood, K. L. Black, S. R. Millis,

Relationship between glasgow coma scale and functional outcome. American

journal of physical medicine & rehabilitation 75, 364-369 (1996).

9. M. P. Kelly, C. T. Johnson, N. Knoller, D. A. Drubach, M. M. Winslow, Substance

abuse, traumatic brain injury and neuropsychological outcome. Brain injury 11,

391-402 (1997).

10. J. F. Malec, A. W. Brown, C. L. Leibson, J. T. Flaada, J. N. Mandrekar, N. N. Diehl,

P. K. Perkins, The mayo classification system for traumatic brain injury severity. J

Neurotrauma 24, 1417-1424 (2007); published online EpubSep

(10.1089/neu.2006.0245).

11. J. A. Langlois, W. Rutland-Brown, M. M. Wald, The Epidemiology and Impact of

Traumatic Brain Injury. Journal of Head Trauma Rehabilitation 21, 375-378 (2006).

12. K. M. Guskiewicz, M. McCrea, S. W. Marshall, Cumulative Effects Associated With

Recurrent Concussion in Collegiate Football Players The NCAA Concussion

Study. JAMA 19, 2549-2555 (2003).

13. G. Barkhoudarian, D. A. Hovda, C. C. Giza, The molecular pathophysiology of

concussive brain injury. Clinics in sports medicine 30, 33-48, vii-iii (2011);

published online EpubJan (10.1016/j.csm.2010.09.001).

14. R. C. Cantu, Second-impact syndrome. Clinics in sports medicine 17, 37-44

(1998).

15. P. McCrory, Does second impact syndrome exist? Clinical Journal of Sport

Medicine 11, 144-149 (2001).

16. P. McCory, G. Davis, M. Makdissi, Second Impact Syndrome or Cerebral Swelling

after Sporting Head Injury. Current Sports Medicine Reports 11, 21-23 (2012).

17. J. T. Povlishock, E. Pettus, in Mechanisms of Secondary Brain Damage in

Cerebral Ischemia and Trauma. (Springer, 1996), pp. 81-86.

18. E. H. Pettus, C. W. Christman, M. L. Giebel, J. T. Povlishock, Traumatically

Induced Altered Membrane Permeability: Its Relationship to Traumatically Induced

Reactive Axonal Change. Journal of Neurotrauma 11, 507-522 (1994).

19. G. V. W. Johnson, J. A. Greenwood, A. C. Costello, J. C. Troncoso, The regulatory

role of calmodulin in the proteolysis of individual neurofilament proteins by calpain.

Neurochemical Research 16, 869-873 (1991)10.1007/bf00965535).

20. R. A. Nixon, The Regulation of Neurofilament Protein Dynamics by

Phosphorylation: Clues to Neurofibrillary Pathobiology Brain Pathology 3, 29-38

(1993).

21. J. T. Povlishock, C. W. Christman, The pathobiology of traumatically induced

axonal injury in animals and human: a review of current thoughts. Journal of

Neurotrauma 12, 555-564 (1995).

22. A. Marmarou, A review of progress in understanding the pathophysiology and

treatment of brain edema. Neurosurgery Focus 22, 1-10 (2007).

23. P. Barzó, A. Marmarou, P. Fatouros, F. Corwin, J. Dunbar, Magnetic resonance

imaging-monitored acute blood-brain barrier changes in experimental traumatic

brain injury. Journal of neurosurgery 85, 1113-1121 (1996).

24. H. Feldmann, G. Klages, F. Gärtner, J. Scharfenberg, in Proceedings of the 6th

European Congress of Neurosurgery. (Springer, 1979), pp. 74-77.

25. H. M. Eisenberg, H. E. Gary Jr, E. F. Aldrich, C. Saydjari, B. Turner, M. A. Foulkes,

J. A. Jane, A. Marmarou, L. F. Marshall, H. F. Young, Initial CT findings in 753

patients with severe head injury: a report from the NIH Traumatic Coma Data Bank.

Journal of neurosurgery 73, 688-698 (1990).

26. S. M. Lucas, N. J. Rothwell, R. M. Gibson, The role of inflammation in CNS injury

and disease. British journal of pharmacology 147 Suppl 1, S232-240 (2006);

published online EpubJan (10.1038/sj.bjp.0706400).

27. C. Bogdon, Nitric oxide and the immune response. Nature 2, 907-916 (2001).

28. N. J. Van Wagoner, E. N. Benveniste, Interleukin-6 expression and regulation in

astrocytes. Journal of neuroimmunology 100, 124-139 (1999).

29. E. Tobinick, N. M. Kim, G. Reyzin, H. Rodriguez-Romanacce, V. DePuy, Selective

TNF Inhibition for Chronic Stroke and Traumatic Brain Injury. CNS drugs 26, 1051-

1070 (2012).

30. S. A. Loddick, A. V. Turnbull, N. J. Rothwell, Cerebral interleukin-6 is

neuroprotective during permanent focal cerebral ischemia in the rat. Journal of

Cerebral Blood Flow & Metabolism 18, 176-179 (1998).

31. F. Velarde, D. Fisher, D. Hovda, P. Adelson, D. Becker, Fluid percussion injury

induces prolonged changes in cerebral blood flow. J Neurotrauma 9, 402 (1992).

32. C. C. Giza, D. A. Hovda, The Neurometabolic Cascade of Concussion. Journal of

Athletic Training 36, 228-235 (2001).

33. D. A. Bruce, A. Alavi, L. Bilaniuk, C. Dolinskas, W. Obrist, B. Uzzell, Diffuse

cerebral swelling following head injuries in children: the syndrome of “malignant

brain edema”. Journal of neurosurgery 54, 170-178 (1981).

34. H. K. Richards, S. Simac, S. Piechnik, J. D. Pickard, Uncoupling of cerebral blood

flow and metabolism after cerebral contusion in the rat. Journal of Cerebral Blood

Flow & Metabolism 21, 779-781 (2001).

35. A. J. Church, R. D. Andrew, Spreading Depression Expands Traumatic Injury in

Neocortical Brain Slices. Journal of Neurotrauma 22, 277-290 (2005).

36. O. B. Paulson, E. A. Newman, Does the Release of Potassium from Astrocyte

Endfeet Regulate Cerebral Blood Flow? Science 237, 896-898 (1987).

37. R. D'Ambrosio, D. O. Maris, M. S. Grady, R. H. Winn, D. Janigro, Impaired K+

Homeostasis and Altered Electrophysiological Properties of Post-Traumatic

Hippocampal Glia. Journal of Neuroscience 19, 8152-8162 (1999).

38. A. Yoshino, D. A. Hovda, T. Kawamata, Y. Katayama, D. P. Becker, Dynamic

changes in local cerebral glucose utilization following cerebral concussion in rats:

evidence of a hyper-and subsequent hypometabolic state. Brain research 561,

106-119 (1991).

39. C. L. Osteen, A. H. Moore, M. L. Prins, D. A. Hovda, Age-Dependency of

45Calcium Accumulation Following Lateral Fluid Percussion: Acute and Delayed

Patterns. Journal of Neurotrauma 18, 141-162 (2001).

40. W. Young, I. Koreh, Potassium and Calcium Changes in Injured Spinal Cords.

Brain research 365, 42-53 (1985).

41. M. Prins, T. Greco, D. Alexander, C. C. Giza, The pathophysiology of traumatic

brain injury at a glance. Disease models & mechanisms 6, 1307-1315 (2013);

published online EpubNov (10.1242/dmm.011585).

42. T. I. Peng, M. J. Jou, Oxidative stress caused by mitochondrial calcium overload.

Ann N Y Acad Sci 1201, 183-188 (2010); published online EpubJul

(10.1111/j.1749-6632.2010.05634.x).

43. T. Kawamata, Y. Katayama, D. A. Hovda, A. Yoshino, D. P. Becker, Administration

of excitatory amino acid antagonists via microdialysis attenuates the increase in

glucose utilization seen following concussive brain injury. Journal of Cerebral

Blood Flow & Metabolism 12, 12-24 (1992).

44. M. Bergsneider, D. A. Hovda, E. Shalmon, D. F. Kelly, P. M. Vespa, N. A. Martin,

M. E. Phelps, D. L. McArthur, M. J. Caron, J. F. Kraus, Cerebral hyperglycolysis

following severe traumatic brain injury in humans: a positron emission tomography

study. Journal of neurosurgery 86, 241-251 (1997).

45. M. O’Connell, A. Seal, J. Nortje, P. Al-Rawi, J. Coles, T. Fryer, D. Menon, J.

Pickard, P. Hutchinson, in Intracranial Pressure and Brain Monitoring XII.

(Springer, 2005), pp. 165-168.

46. N. Hattori, S.-C. Huang, H.-M. Wu, E. Yeh, T. C. Glenn, P. M. Vespa, D. McArthur,

M. E. Phelps, D. A. Hovda, M. Bergsneider, Correlation of regional metabolic rates

of glucose with Glasgow Coma Scale after traumatic brain injury. Journal of

Nuclear Medicine 44, 1709-1716 (2003).

47. M. L. Prins, D. A. Hovda, Mapping cerebral glucose metabolism during spatial

learning: interactions of development and traumatic brain injury. Journal of

neurotrauma 18, 31-46 (2001).

48. R. Balabanov, H. Goldman, S. Murphy, G. Pellizon, C. Owen, J. Rafols, P. Dore-

Duffy, Endothelial cell activation following moderate traumatic brain injury.

Neurological research, (2013).

49. B. L. Bartnik, R. L. Sutton, M. Fukushima, N. G. Harris, D. A. Hovda, S. M. Lee,

Upregulation of pentose phosphate pathway and preservation of tricarboxylic acid

cycle flux after experimental brain injury. Journal of neurotrauma 22, 1052-1065

(2005).

50. M. A. Satchell, X. Zhang, P. M. Kochanek, C. E. Dixon, L. W. Jenkins, J. Melick,

C. Szabó, R. S. Clark, A dual role for poly‐ADP‐ribosylation in spatial memory

acquisition after traumatic brain injury in mice involving NAD+ depletion and

ribosylation of 14‐3‐3γ. Journal of neurochemistry 85, 697-708 (2003).

51. C. C. Alano, P. Garnier, W. Ying, Y. Higashi, T. M. Kauppinen, R. A. Swanson,

NAD+ Depletion Is Necessary and Sufficient forPoly (ADP-Ribose) Polymerase-1-

Mediated Neuronal Death. The Journal of Neuroscience 30, 2967-2978 (2010).

52. Z. Yang, K. K. W. Wang, Glial fibrillary acidic protein: from intermediate filament

assembly and gliosis to neurobiomarker. Cell 38, 364-374 (2015).

53. M. V. Sofroniew, Molecular dissection of reactive astrogliosis and glial scar

formation. Trends in neurosciences 32, 638-647 (2009).

54. G. R. Gordon, S. J. Mulligan, B. A. MacVicar, Astrocyte control of the

cerebrovasculature. Glia 55, 1214-1221 (2007).

55. G. R. Gordon, H. B. Choi, R. L. Rungta, G. C. Ellis-Davies, B. A. MacVicar, Brain

metabolism dictates the polarity of astrocyte control over arterioles. Nature 456,

745-749 (2008).

56. M. Simard, M. Nedergaard, The neurobiology of glia in the context of water and

ion homeostasis. Neuroscience 129, 877-896 (2004).

57. K. Kam, R. Nicoll, Excitatory synaptic transmission persists independently of the

glutamate–glutamine cycle. The Journal of Neuroscience 27, 9192-9200 (2007).

58. R. R. Voskuhl, R. S. Peterson, B. Song, Y. Ao, L. B. J. Morales, S. Tiwari-Woodruff,

M. V. Sofroniew, Reactive astrocytes form scar-like perivascular barriers to

leukocytes during adaptive immune inflammation of the CNS. The Journal of

neuroscience 29, 11511-11522 (2009).

59. M. V. Sofroniew, H. V. Vinters, Astrocytes: biology and pathology. Acta

Neuropathol 119, (2010).

60. T. G. Bush, N. Puvanachandra, C. H. Horner, A. Polito, T. Ostenfeld, C. N.

Svendsen, L. Mucke, M. H. Johnson, M. V. Sofroniew, Leukocyte infiltration,

neuronal degeneration, and neurite outgrowth after ablation of scar-forming,

reactive astrocytes in adult transgenic mice. Neuron 23, 297-308 (1999).

61. J. D. Rothstein, M. Dykes-Hoberg, C. A. Pardo, L. A. Bristol, L. Jin, R. W. Kuncl,

Y. Kanai, M. A. Hediger, Y. Wang, J. P. Schielke, Knockout of glutamate

transporters reveals a major role for astroglial transport in excitotoxicity and

clearance of glutamate. Neuron 16, 675-686 (1996).

62. R. A. Swanson, W. Ying, T. M. Kauppinen, Astrocyte influences on ischemic

neuronal death. Current molecular medicine 4, 193-205 (2004).

63. J. R. Faulkner, J. E. Herrmann, M. J. Woo, K. E. Tansey, N. B. Doan, M. V.

Sofroniew, Reactive astrocytes protect tissue and preserve function after spinal

cord injury. The Journal of Neuroscience 24, 2143-2155 (2004).

64. Z. Zador, S. Stiver, V. Wang, G. T. Manley, in Aquaporins. (Springer, 2009), pp.

159-170.

65. S. Okada, M. Nakamura, H. Katoh, T. Miyao, T. Shimazaki, K. Ishii, J. Yamane, A.

Yoshimura, Y. Iwamoto, Y. Toyama, Conditional ablation of Stat3 or Socs3

discloses a dual role for reactive astrocytes after spinal cord injury. Nature

medicine 12, 829-834 (2006).

66. A. D. R. Garcia, N. B. Doan, T. Imura, T. G. Bush, M. V. Sofroniew, GFAP-

expressing progenitors are the principal source of constitutive neurogenesis in

adult mouse forebrain. Nature neuroscience 7, 1233-1241 (2004).

67. R. Hausmann, R. Riess, A. Fieguth, P. Betz, Immunohistochemical investigations

on the course of astroglial GFAP expression following human brain injury.

International journal of legal medicine 113, 70-75 (2000).

68. L. M. Herx, V. W. Yong, Interleukin-1β is required for the early evolution of reactive

astrogliosis following CNS lesion. Journal of Neuropathology & Experimental

Neurology 60, 961-971 (2001).

69. L. Papa, L. M. Lewis, J. L. Falk, Z. Zhang, S. Silvestri, P. Giordano, G. M. Brophy,

J. A. Demery, N. K. Dixit, I. Ferguson, M. C. Liu, J. Mo, L. Akinyi, K. Schmid, S.

Mondello, C. S. Robertson, F. C. Tortella, R. L. Hayes, K. K. Wang, Elevated levels

of serum glial fibrillary acidic protein breakdown products in mild and moderate

traumatic brain injury are associated with intracranial lesions and neurosurgical

intervention. Ann Emerg Med 59, 471-483 (2012); published online EpubJun

(10.1016/j.annemergmed.2011.08.021).

70. A. Lewen, P. Matz, P. H. CHAN, Free radical pathways in CNS injury. Journal of

neurotrauma 17, 871-890 (2000).

71. J. Truettner, R. Schmidt-Kastner, R. Busto, O. Alonso, J. Loor, W. D. Dietrich, M.

Ginsberg, Expression of brain-derived neurotrophic factor, nerve growth factor,

and heat shock protein HSP70 following fluid percussion brain injury in rats.

Journal of neurotrauma 16, 471-486 (1999).

CHAPTER 2: ADDRESSING THE NEEDS OF TRAUMATIC BRAIN INJURY

WITH CLINICAL PROTEOMICS

2.1 INTRODUCTION

A general goal of “proteomics” is to comprehend the relationship between the body’s proteins and how they change by disease to understand human pathophysiology, and ultimately to provide therapeutic and diagnostic tools. The completion of the human genome provided researchers with the blueprint for life; proteomics offers the potential means for analyzing the expressed genome.

Proteomics attempts to determine how genes function within the genome and how they communicate with each other to (hopefully) lead to important new insights into disease mechanisms. The potential of proteomics to advance biomedical research is high because the key functional components of biochemical systems and the cellular targets of therapeutic agents, namely proteins, are being studied. Mapping proteomes from injured tissues, cells and biofluids can potentially reveal new protein targets to explore mechanisms of insults and to provide candidate lists for new disease indicators or injury biomarkers as diagnostic or prognostic tools for the clinic.

A biomarker could be simply a molecule, such as a protein whose presence or abundance in a biological sample signals a disease or insult to an organ. Thus, they are quantifiable molecules that indicate a pathophysiological process. A biomarker in accessible body fluids or tissues could greatly enhance our ability to identify patients at risk, with invisible wounds or predict outcome of serious injury. A sensitive and specific disease or injury marker such as an early protein abnormality could provide

a warning sign prior to being symptomatic, and hence could result in more effective preventative care or treatment options to improve outcome.

2.2 DISCUSSION

The challenges of clinical proteomics and biomarkers

The goal of clinical proteomics to discover new disease or injury biomarkers is challenging. Beyond the number of human genes coding for proteins, proteins are processed and modified, comprising an important dimension of information to which present proteomic technologies have but limited access. The total mRNA population, accounting for alternate splicing, RNA editing, and use of alternate promoters could contain 250,000 transcripts, while various protein modifications could increase the size of the human proteome to over 500,000 members (1). Cellular proteins and their post-translational modifications (PTMs) change with the cell cycle, environmental conditions, developmental stage, and metabolic state. Independent of these variables, biomarkers should reliably detect changes in health status, a specific disease, or indicate whether an insult like a toxic exposure or trauma has occurred.

Clearly, we need proteomic approaches that advance beyond identifying proteins to elucidating their co- and post-translational modifications, to following the dynamics of those modifications, and to linking those modifications to specific diseases or cellular responses to an insult that inflicted an organ. Despite all of the significant advances in technologies in proteomics since its inception in the mid-1990s, with the development of more sensitive mass spectrometry detectors and more selective and specific strategies for sample processing and handling, no clinically validated disease biomarker has been discovered by proteomics to date (2).

Meeting the challenge with targeted screens, focused selection strategies, and clinical validation

What are the major factors that hindered finding robust disease and injury biomarkers and how can these be overcome? The complexity of clinical samples themselves is a significant limiting factor. Plasma and serum, i.e., blood, have been biofluids of choice for measuring levels of proteins and other biomolecules for clinical testing, as they can be sampled noninvasively. Plasma is a protein-rich information source containing what blood circulation has encountered on its journey throughout the body and tissue perfusion. The tremendous analytical challenge of the large number of plasma proteins lays in their unbalanced abundance: albumin constitutes over 50% of the plasma proteins (at 30–50 mg/mL) and the most abundant 22 proteins in plasma represent approximately 99% of the total protein content in plasma leaving the majority of proteins at very low abundance.

The estimated dynamic range of protein concentrations in human plasma may be up to 12 orders of magnitude (3).

Disease or insults trigger acute events, secondary and chronic sequelae, including inflammation, wound healing, and adaptive changes that the compromised body undergoes in response to the unhealthy state. In an effort to identify original disease causes or injury factors a simple experimental model can facilitate a targeted screen circumventing secondary, less disease-specific events. As such, scientific experimental model design follows controlled strategies for reproducibility and simplicity that can facilitate the initial discovery by limiting candidate markers to those proteins that are related to a disease origin or injury cause (4,5). One common

proteomics workflow involves a 2-dimensional separation prior to protein identification to reduce sample complexity (Figure 1). Proteins can be sorted by charge (isoelectric point) and size using two-dimensional polyacrylamide gel electrophoresis (2D-PAGE) and can be enzymatically digested within the gel matrix.

Despite being developed over 3 decades ago (6,7), 2D-PAGE remains one of the most powerful separation techniques for proteomic workflows and was instrumental in early protein biomarker research. Following separation, gels are stained and differentially expressed protein spots excised, enzymatically digested with trypsin, and identified by MS requiring only sufficiently accurate mass measurements (low part-per million range) performed on one or two tryptic peptides to identify silver stained protein spots (8).

A second strategy advocates first enzymatically (e.g., with trypsin) or chemically cleaving (“breaking”) a complex mixture of cellular proteins, and then “sorting” the peptides by one or more steps of chromatography. MS analyzes the recovered fragments as in the previous approach, and software matches the fragments to the proteins from which they are derived. Examples of this experimental approach include multidimensional protein identification technology (MudPIT) that couples two or more dimensions of chromatographic separations, e.g., strong cation exchange

(SCX) with reversed phase chromatography (9,10). While the outlined approaches have been instrumental in biomarker discovery research, the extensive sample preparation and time required in gel fractionation and long HPLC LC-MS/MS analyses make discovery proteomics feasible for only limited numbers of samples per project (11,12). A simplified disease or injury model using a controlled

experimental design may help to relieve a proteomic screen from confounding complexities of clinical samples (4,13-15).

A straightforward selection of suitable marker candidates from the ‘long list’ of identified injury or disease specifically changed proteins should arrive at a manageable ‘short list’ of possible disease marker candidates. A tailored selection strategy will consider injury cause, marker candidates with the necessary reporting power for the cause as well as organ specificity and exclusion of proteins normally present in healthy plasma and tissues. The subsequent validation of selected disease or injury markers from a group of candidates may occur stepwise starting with a preclinical smaller cohort of patients and controls, allowing to test for normality (16).

Following initial confirmation, a larger subject cohort can be enrolled in clinical trials allowing for receiver operating characteristic curve analyses that will establish the basis for biomarker suitability in the clinic (17). Currently, the majority of biomarker validation studies have been performed by enzyme-linked immunosorbent assay

(ELISA). This highly sensitive method is limited for use early in the verification process, as antibody pairs have to be optimized for specificity and sensitivity for each marker separately. As mass spectrometry measurements improve in sensitivity to match immunoassay detection limits (pg/ mL), a targeted and quantitative mass spectrometry application can provide multiplex capacity and absolute specificity by gas-phase sequence determination, making it an ideal alternative for assessing validity of selected marker candidates.

The need for markers of Traumatic Brain Injury (TBI)

Neurotrauma to the central nervous system (CNS) is a serious public health problem in the US; among US civilians, TBI is most common in infants and toddlers, adolescents and the elderly (18). The US National Institute of Neurological Disorders and Stroke estimates that 2.5-6.5 million Americans have had one or multiple TBIs

(19). In the US military there were over 212,000 service men and women diagnosed with some form of TBI between January 2000-May 2011, roughly accounting for one- third of all injured US soldiers, making TBI the signature injury of the wars in Iraq and

Afghanistan compared to past wars (20). TBI contributes to over one third of all injury- related deaths, yet 75-90% of all brain trauma cases are considered to be mild TBI

(mTBI), many without visible wounds that often are undiagnosed (21). Better diagnostic tools are needed to detect head injuries, especially mTBI, to confirm and to monitor the severity of TBI in order to determine the best course of action acutely and later post-injury. The neurotrauma field has currently still no chemical diagnostic marker in clinical use. Here we will outline briefly the spectrum of TBI and give examples where a surrogate chemical marker assay for TBI would be of great benefit to patients, high risk populations, their families and doctors.

Head injuries can be classified into penetrating and non-penetrating TBI.

Penetrating TBI involves physical compromise of the skull by an external object resulting in specific, focused injury most commonly characterized by hemorrhages and lesions. Non-penetrating TBI, is much more difficult to assess, as injuries may not be visible or located precisely. Closed head injuries are caused by rapid acceleration and deceleration of the brain within the skull and inflict shear and deformation forces on gray matter tissue and white matter tracts (22). Each trauma

patient is a unique injury case with individual complexity, thus the field distinguishes mainly between severe and mild TBI (mTBI) as opposite ends of a clinical spectrum of manifestations. Evaluating and predicting outcome in severe TBI is often problematic, especially for patients without visible wounds such as infants.

Diagnostic neurotrauma tools include imaging techniques, neurocognitive examinations, and for severe TBI patients, the determination of post-traumatic amnesia, but they provide only estimates of the dynamically evolving injury process.

Functional MRI (fMRI) and the detection of regional blood flow changes (e.g., PET scans) are not always available, cannot be obtained in critically ill patients, and are not definitive. Radiological brain scans on infants and toddlers are widely considered problematic because the radiation dose endangers the developing brain.

Absence of imaging in the pediatric clinical praxis prevents distinguishing brain injury from frequent intestinal flu or even infant irritability (23). Non-accidental head injury, or “Shaken Baby” syndrome, caused by rotation-acceleration strains on the brain in the still loosely connected infant skull causes bleeding and swelling that can lead to catastrophic intracranial damage and can severely impair normal brain development (and can even lead to death) (23). Undiagnosed victims may be sent back to continued abuse. On the other hand, imaging does not distinguish inflicted head injury from non-traumatic bleeding, originating from a trauma independent condition – a situation in which legal authorities, parents and care-givers would greatly benefit from an assay for brain trauma-specific chemicals (24,25).

Mechanical impacts traumatizing the brain obviously need to be clinically differentiated from trauma in other organs or from other non-traumatic brain injuries

like stroke, ischemia, bleeding diseases, poisoning, epilepsy or chronic degenerative diseases for proper treatment and activities in the operating room and the courtrooms (26). Monitoring daily progression of a severe TBI patient by repeated imaging can be quite impractical, considering life supporting intensive care instrumentation. A fluid derived chemical marker for compromised brain cell viability will be a useful added measure of the patients evolving status and could aid in outcome prognosis.

For the vast majority of mTBI/concussion patients, there are no objective diagnostic or prognostic tools (27). A ready diagnostic tool at point of care acutely after TBI is needed especially for high-risk individuals (e.g., athletes, military personnel). An objective and unambiguous trauma biochemical assay would be valuable for legal authorities in forensic cases that currently rely on neuropsychological testing that lacks premorbid base rates and is subject to malingering and subjective interpretation. Thus, for high-risk groups, for mild and severe TBI cases as well as for all pediatric neurotrauma patients, there is an urgent need for an accurate, unambiguous chemical measure indicating that a significant impact to the brain had occurred.

Moreover, a second hit to a concussed, vulnerable brain can, in rare cases, have a catastrophic outcome with permanent brain damage or even death (known as the second impact syndrome) (28). Several repeated concussions over time can in later years cumulate in irreversible brain damage with devastating psychological and cognitive decline, a pathological condition now defined as chronic traumatic encephalopathy (CTE) (29,30). Military personnel and veterans with mild TBI often

suffer from post-traumatic stress disorder (PTSD) after being exposed to blast waves from explosive devices (31,32).

Certain areas of the brain may be more susceptible to concussive trauma. A recent study investigated longitudinal changes in global and regional brain volume in patients one year after mTBI and correlated such changes with clinical and neurocognitive metrics. Magnetic resonance imaging data showed measureable global brain atrophy, larger than that in control subjects one year after mTBI.

Atrophy was found in specific regions of the cingulate cortex independent of the site of initial trauma. The cingulate cortex’s role in rational cognitive functions such as empathy, impulse control, and emotion correlate strongly with the patient’s observed clinical symptoms of increased depression and anxiety (33). These finding are supported by an independent study of National Football League players and referees using positron emission tomography (PET) with a tau specific tracer that showed higher densities of tau tangles in regions of the brain involved in a nearby region (caudate nucleus) that is also associated with learning, memory, emotion, and language comprehension. The deposition of tau tangles is consistent with those observed in CTE autopsy patients (34).

Current evaluation of concussion is basically an assessment of neurocognitive deficits, often not immediate and requires extensive neuropsychological testing that is subject to motivational confounds, while critical care treatment decisions have to be made immediately by emergency clinical personnel and surgeons. Severity classification of TBI patients relies on assessing the level of consciousness, commonly with the Glasgow Coma Scale (GCS), which is an insensitive measure.

Testing relies on verbal communication, and proper motor control and eye function, which are often impaired after TBI. Brain function-altering substances such as drugs, alcohol, pain medication, sedatives, or even induced coma as part of emergency and intensive care routine obviously compromise the use of memory recall and the GCS.

Although predictors of TBI exist, such as the Standardized Assessment of

Concussion test, these tools offer little insight into the pathology of the disease beyond determining whether a concussion has occurred or not. Because of this lack of insight into TBI, licensed health care providers of concussive sports injury are conservative in their approach to player safety after injury with the hope that coordination between sideline and clinical practitioners will aid in improving our understanding of the extent of impairment for various types of sports related concussions (35).

Current potential TBI biomarker candidates

An ideal biomarker should be both specific to head trauma as well as sufficiently sensitive to be measured and quantified reproducibly in patient blood or other peripheral or proximal fluid samples (such as CSF) by an assay of choice.

These markers should be acutely released into the fluids following injury and show a distinct temporal signal pattern. The identification of a unique TBI biomarker(s) or surrogate brain cell injury markers that meet these criteria would provide physicians with an objective method for early diagnosis of brain injury and enable early assessment of severity, intervention, and monitoring disease progression (36).

Multiple neurotrauma signature markers would allow for correlation analyses with

improved statistical power using multivariant logistic regression or similar analyses

(37). Finding candidate TBI markers is pursued typically by these strategies: (1)

Classical deduction chooses proteins with literature reported association to brain injury or its secondary events like inflammation, axonal degeneration or reactive astrogliosis. (2) Hypothesis driven animal trauma model studies report changes in specific proteins using available antibodies or pathway tailored kits (38). (3) Discovery of trauma associated proteins using a proteomic screen of samples derived from animal injury models or small patient cohorts (39-43). Surprisingly, few screens address the impact of mechanical trauma on brain cells, i.e., cell death (13,14,44). After briefly summarizing currently investigated candidate TBI markers, we will evaluate challenges and alternatives in identifying TBI markers.

Inflammatory markers

Part of the pathology of CNS injury is characterized by secondary effects, including the inflammatory response to TBI. Cytokines are key mediators in the process of (neuro)inflammation (45) and increased concentrations of these compounds have been associated with severe CNS injury as well as post-traumatic hypoxia (46,47). For example, elevated levels of interleukin-10 (IL-10), an anti- inflammatory cytokine, was measured in low pg/mL levels in CSF and low-mid pg/mL levels in serum (Table 1) and correlated with severe TBI determined by the GCS

(46,48-50). Higher Il-10 serum levels illustrate the systemic nature of an inflammatory response. Such responses are systemic in nature and not specific to

TBI, but occur with any insult, hence inflammatory markers are not ‘pointing to brain

injury’.

Neuronal markers

With their elongated axonal and dendritic processes, neurons are exposed to shear forces associated with the whiplash trauma of a concussion. Acute plasma membrane permeablility, or mechanoporation, compromises cell integrity and is linked to diffuse axonal injury in response to a mechanical impact (71-73). Tau protein is a member of microtubule-associated proteins involved in maintaining cytoskeletal structure and axonal transport. It is expressed by CNS neurons and oligodendrocytes and found primarily in axons (74). Traditionally used in the diagnosis of Alzheimer’s disease, elevated levels of Tau in CSF and serum have been linked to CNS insults like TBI and stroke (62,75). CSF and serum studies of

TBI patients have measured elevated Tau protein concentrations in the 1000 ng/mL range in young adult TBI patients, whereas it is three orders of magnitude lower in neonates with brain insults (59,65,66). Because of Tau’s chronic accumulation after various CNS insults, it seems less useful as an acute head trauma marker.

Mylein basic protein (MBP) is released with myelin debris that accumulates with axonal damage in the injured brain or spinal cord. MBP is one of three proteins comprising the myelin sheath essential for axonal impulse conduction (76). MBP markers have shown promise in the appraisal of TBI with serum levels in the low- mid ng/mL range (77,78). Similar to GFAP (vide infra), studies have demonstrated degradation of MBP isoforms as a result of TBI (79,80).

Neuron Specific Enolase (NSE), Microtubule-associated protein 2 (MAP-2)

and ubiquitin C-term hydrolase L1 (UCH-L1) all display differential expression patterns in TBI patients. NSE, a glycolytic enzyme isoform of neurons, has been documented to increase following head trauma (77), but has a slow elimination process, making it difficult to distinguish between primary and secondary injuries

(81). Additionally, NSE is released during the process of hemolysis, making it difficult to pin down the source of injury (82).

Microtubule-associated protein 2 (MAP-2) is a cytoskeletal-associated protein localized to dendrites of neurons that is believed to function in the growth and maturation of dendrites as well as cytoskeletal organization (83). Previous studies have demonstrated that MAP-2 is absent from damaged regions of the brain and that serum levels increase early after injury (84). Mondello et al. assessed the long-term release of MAP-2 in blood 6 months post trauma by ELISA immunoassay and found that severe TBI patients had significantly higher serums levels of MAP-2 compared to normal non-TBI patients. TBI patients in a vegetative state, as assessed by the GCS, however, showed no increase in serum MAP-2 versus controls. This suggests that

MAP-2 could provide insight into the mechanism of neuronal remodeling as well as discriminate between patients with deficits in consciousness and increased risk of unfavorable outcomes (70).

Ubiquitin C-terminal hydrolase-L1 (UCH-L1) has been identified in a cell death culture assay and is verified by ELISA to be significantly increased in TBI patients

(85). Neurodegenerative marker UCH-L1 fluid levels are also elevated in ischemia, vasospasm, infarction, and carbon monoxide poisoning (86-88). UCH-L1 is a proteolytically stable, abundant neuronal protein (69,70,85,89). Future studies will

show whether these proteins would be present in mTBI subjects without significant brain cell death.

Trauma specific breakdown products of neuronal and glial cytoskeletal proteins

Spectrin breakdown products (SBDPs) have been identified as potential TBI biomarkers in rat CSF fluid (90). αII-Spectrin is the submembraneous cortical cytoskeleton of neurons and astroglia, sharing 50-59% homology with the abundant erythroid α-spectrin (44,91). Cell-death associated spectrin fragments of molecular weight 150 kDa (SBDP150) and two N-terminal fragments at 145 kDa (SBDP145) and 120 kDa (SBDP120) cleaved by calpain and caspase-3 have been identified in a cell death culture model (87,92,93). Using a sandwich ELISA methodology,

Mondello et al. showed both SBDP145 and SBDP120 increased in patients post-

TBI, with SBDP145 present immediately post-trauma and SBDP120 most accurately measured 24 hr post-injury. SBDP CSF levels greater than specific thresholds were shown to correlate with poor outcome and mortality and the temporal expression of SBDP for non-surviving patients differed from that of surviving patients. Thus, if cross-reactivity and breakdown specificity is controlled, SBDPs in

CSF may aid to predict the severity of injury and mortality (69).

Astroglial markers

Astroglia are the most abundant cells in the human cerebral cortex (94) and respond to insult by becoming reactive, a process that involves gene expression, morphological changes, proliferation, and the formation of a glial scar around lesions

(95-99). However, astrocytes are also trauma victims as they are especially vulnerable to acidosis, pressure elevation, and hypoxic/ischemic damage, known comorbidities of TBI (100-103). Human astrocytes display very long thin processes that cross through several laminae from the pia to the ventricular walls, so called interlaminar processes and are hence vulnerable to shear and deformation forces similar to those that cause diffuse axonal injury in white matter tracks (104,105). Two of the most well studied TBI marker candidates are S100β and glial fibrillary acidic protein (GFAP), both glial proteins. S100β is a calcium binding protein that is predominantly produced by astrocytes within the CNS. Because S100β is also produced in a variety of non-CNS cells (e.g., lymphocytes, bone marrow, adipocytes, and glia of peripheral nerves), brain specificity is its problem (106). It has, however, been reported that the few extracranial sources of S100β are short lived compared to S100β from cerebral lesions (107,108). Elevated S100β concentrations have been measured in the ng/mL and pg/mL range in TBI patient CSF and serum, respectively

(51,53). Despite the immediate spike in S100β levels, it has been found that S100β measurements taken 24 hours post TBI offer the most prognostic value for patient outcome due to initial interference from external S100β (52). S100β is released into the perivascular space immediately following blood brain barrier (BBB) compromise and may serve as a BBB-permeability marker (109). Additionally, higher levels of

S100β have been correlated to patients suffering from post-traumatic hypoxia, demonstrating the interrelation between secondary effects and amplified biomarker expulsion (46).

GFAP is an intermediate filament that is highly enriched in CNS astroglia, but

is also expressed in Schwann cells and olfactory ensheathing glia of peripheral nerves (110-112). GFAP levels are persistently elevated after severe TBI in CSF and serum, relate to poor outcome, and are predictive for mortality (54,55,113). Serum levels of GFAP show high variability or no elevation after mTBI, yet reports are confounded by varying delimitation of ‘mild’ as to include more moderate cases with lesions and positive imaging signals or not. Thus the discriminative power of GFAP as a mild neurotrauma biomarker is conflicted (56,114). Measured CSF levels of several biomarkers in boxers acutely after one or repeated blows to the head as well as after 14 days, revealed elevated levels of GFAP with large variations among the boxers suffering a concussion (56). GFAP breakdown products are found after

TBI and are being explored as insult-specific markers (114-116).

Strategies for addressing the challenges in identifying and validating new TBI biomarkers

For a brain cell specific protein to be a trauma marker, either it should be selectively expressed in response to the trauma and then discharged into fluids, or cytosolic proteins released solely from injured neurons and glia with compromised membrane integrity or dying brain cells (73,117). A suitable study design to identify fluid-derived trauma specific proteins would employ a targeted proteomic screen on a defined trauma model. Experimental animal injury models were developed with the effort to mimic human TBI as closely as possible while underlying cellular and molecular mechanisms of acute trauma are still scantly investigated. The predominant criterion is to recapitulate the clinical manifestation of TBI over studies using

simplified reproducible trauma models with the goal to determine primary cellular injury consequences (4,5). Most commonly used injury models include focal injuries with the animal’s head in a fixed position, like fluid percussion and controlled cortical impact, which produce a focal contusion with hematoma and hemorrhage while the dura remains intact (118,119). Also used is Marmarou’s weight drop model where distributed forces cause diffuse injury with the animal’s head unrestrained in a helmet and the brain is therefore subjected to rotational forces as well (120,121). Blast injury models historically use shock tubes and larger animals, but have been adapted recently to rodents as well as investigated for milder blast effects from explosion exposures in the field (122).

Developing biochemical markers of TBI by proteomics and mass spectrometry

Proteomic studies of injured brain or spinal cord tissue are being done in these injury models and are providing lists of protein changes that are difficult to interpret due to the complexity of events at and around a dynamically changing lesion site and variations between models (39,40,42). Injury zones are not reproducibly defined from lab to lab as histopathological analyses have for long not followed standardized analysis and reporting criteria (5). Tissue derived protein signals are products of a changing composition of viable, injured, and dead cells as well as infiltrating non-neural cells, that complicate the interpretation of proteomic studies (97,99). An effort has been made in recent years to standardize and compare severities of commonly used TBI animal models across centers (123).

Defining common data elements for collection, analysis protocols, and reporting of

fluid samples and histopathological defining features of injury models will help this field in interpreting proteomic and biomarker preclinical studies as well as clinical data collection and interpretation (124-126).

Proteomic TBI marker projects on biofluids using rodent injury models have been few due to naturally limited available fluid amounts (42,127), but biofluid neurotrauma marker candidates have been studied in pig blast injury models (128-

132). Human proteomic analyses have started from severe trauma patient’s CSF and plasma from individual patients (41,133). Bioinformatics analysis tools are expected to facilitate systems level understanding of neurotrauma protein changes (134,135).

While bioinformatics tools are indispensable for classification, consensus-based data collection, and data mining, they will not make the bottleneck of biomarker candidate selection much easier.

Hanrieder et al. describes a workflow using matrix-assisted laser desorption ionization time-of-flight (MALDI- TOF) MS/MS in conjunction with off-line nano-LC sample fractionation (136). In their study, ventricular CSF samples from 3 severe TBI patients displaying different symptoms were taken at various time points post-trauma and analyzed by nano-LC MALDI-TOF MS/MS to determine temporal protein expression changes. CSF samples were digested with trypsin and labeled with isobaric tags for relative and absolute quantitation using the iTRAQ method

(137,138). Labeled tryptic digests were then separated on a nano-flow LC system equipped with online fraction collection capable of depositing fractions directly onto

MALDI sample plates for MALDI-TOF MS/MS-based identification and quantification.

Several proteins were increased after injury. Additionally, relative quantification using

iTRAQ labeling revealed temporal changes in protein expression for several inflammation-related proteins (e.g., serum amyloid, fibroinogen alpha chain, ceruloplasmin) as well as known neurotrauma-related proteins (GFAP, NSE).

Due to the confounding complexity of clinical TBI and clinic-resembling animal injury models, we propose a targeted proteomic screen using a well-characterize in vitro cell-based trauma model as a starting point for TBI marker candidate identification (139-144). This will limit protein changes to those directly related to an acute mechanical trauma by applying an abrupt pressure pulse inflicting shear forces and deformation onto cortical brain cells in a reproducible fashion at various severities

(142). We are finding robust cellular release patterns that correlate with cell injury and cell death of rodent and human astrocytes matured and stretched in a prototype of this injury model (139). A suitable selection strategy needs to be applied to any trauma-release protein list to eliminate proteins found in healthy human plasma and to focus on brain-specific proteins (145,146).

Verifying biochemical markers for TBI by proteomics and mass spectrometry

One analytical challenge that is unique to TBI for measuring candidate biomarkers lays in the unpredictably fluctuating protein concentrations among CSF samples from different TBI patients (low microgram/ml to several mg/ml range). This may be due to as variables such as the patient’s varying blood–brain barrier integrity, hemorrhage, brain cell protein leakage, as well as waves of brain cell death. This is unlike healthy CSF or plasma with constant and physiologically controlled protein levels allowing for sample preparation with reproducible protein amounts (147).

These injury specific variables can be addressed only by relating all measurements to raw, unprocessed sample volume regardless of depletion and other processing steps including optimizing protein amounts for trypsin digestion or immunoassay applications. There are also injury-related but not-trauma specific secondary changes in protein composition in trauma CSF that could be caused by secondary infection due to hospitalization that could reduce protein amounts or bacterial proteins present in the samples. Such samples should be omitted from an initial biomarker validation study.

The accepted “gold standard” of single-protein measurements is the ELISA immunoassay, which takes advantage of the specificity and diversity of IgG antigen recognition. Yet, while ELISA is well touted for its high sensitivity (~1 pg/mL), it is not without limitations (148). ELISA methods rely on antibodies for protein detection and assay development ideally uses two antibodies against different epitopes of the candidate TBI marker. Non-specific binding of immunoglobulins to abundant plasma proteins may contribute to a background problem, limiting the availability of suitable highly specific antibodies ideally from different host animals to cancel out non- specific binding. The availability of such antibody pairs often requires de novo generation, lengthening the assay development time. Thus, the lack of multiplex capacity may exclude using the ELISA platform as initial validation tool of candidate

TBI markers in patient samples (149).

By not relying on antibody-antigen binding, quantitative mass spectrometry is well suited to meet the challenge of overcoming the verification bottleneck where immunoassays cannot be applied. MRM-MS is quickly becoming the preferred

method of candidate biomarker verification because of the discriminating power of mass analyzers to accurately measure and quantify multiple specific proteins within a single sample set. Specific peptide fragments (via trypsin digestion) corresponding to the candidate proteins are selected to act as stoichiometric representatives (or surrogates) within a complex patient CSF or blood sample. The mass spectrometer

(usually a triple quadrupole analyzer) is then set to scan for the precursor peptide ion, fragment the precursor in the collision cell, and then select for a specific precursor fragment (known as a transition). Because the mass spectrometer is not expending resources scanning through all the ions within a complex patient sample, the signal from less abundant peptides are no longer being masked by highly abundant ions, partially addressing the problems with high dynamic range limitations. Additionally, MRM provides a more cost-effective alternative for quantification compared to traditional ELISA methods by using stable isotope- labeled internal standards of the selective candidate peptides. Using the method of isotope dilution (150), isotopically labeled peptides are spiked into digested CSF or blood samples and the relative peak heights between the endogenous peptides and isotope-labeled peptides are used to quantify selected candidate biomarkers. This approach has been greatly aided by the increased availability of stable isotope- labeled standard (SIS) peptides manufactured and sold by life science companies

(151). MRM-MS has long been a method of choice for detecting marker metabolites for amino acids, organic acids, and fatty acid disorders in newborns (152). The success of these quantitative methods in candidate biomarker discovery/verification has been well documented in a variety of samples such as synovial fluid (153), CSF

(154), and plasma (155).

The MRM-MS platform is ideally suited to address the challenge of validating several marker candidates at once (multiplexing) and measuring their levels together with candidate TBI markers reported in the literature. This is in large part due to advances in in pre-analysis enrichment methods (156) as well as improvements in both sensitivity and speed of modern mass spectrometers that allow for detection and quantitation in the low-mid ng/mL concentration range. Hybrid Orbitrap mass spectrometers such as the Q-Exactive have demonstrated the ability to detect up to

10 amol of heavy SIS peptides in the presence of 10 ng- 1 ug of yeast tryptic digest background with up to 10 ppm mass accuracy (157). Coupled with the high re- solving power of Orbitrap detectors (up to 140 K for the Q-Exactive) and fast duty cycles to collect full MS/MS spectra, these instruments should be able to confidently identify surrogate peptides. When comparing the low cost of SIS peptide generation from commercial sources to the cost of antibody generation and capacity to multiplex more than ten within a single analytical sample, the mass spectrometry platform is a feasible choice for TBI candidate marker verification for the early preclinical validation stage. Following this initial verification, antibodies will be generated only for the most robustly detected TBI marker candidates for ELISA assay development for future clinical trials and diagnostic use.

2.3 CONCLUSIONS

Combining a targeted screen, a focused selection strategy, and a stepwise approach from preclinical validation towards clinical translation offers a feasible pipeline for candidate TBI marker identification and preparation for its diagnostic use. Validation through a stepwise increasing sample cohort and moving from severe TBI CSF to matching plasma samples and then to mTBI plasma samples will provide verification where experimental analyses and patient samples are matched with appropriate positive controls along the way.

Moreover, the emergence of targeted MS-technologies brings promise to the development of an efficient biomarker discovery to verification pipeline for TBI. This pipeline could consist of the initial application of proteomics technologies in the form of comparative 2D-PAGE and shotgun LC-MS/MS to identify and discover candidate biomarkers from trauma and healthy subject samples. This is followed by the development of quantitative MRM-MS to assess the biological significance of these markers followed by clinical validation in a larger scale. With the possibility of multiplexing using proteomic methods such as MRM-MS, the time required for preclinical verification can be reduced as tens of marker candidate proteins can be monitored concurrently in clinical samples. This process will help narrow the pool of potential surrogates from which the most specific and easily measured candidates can be chosen for clinical validation and assay development.

2.4 FIGURES

Figure 2.1: Candidate Biomarker Discovery and Verification Workflow.

Bottom-up proteomics strategies, such as shotgun proteomics (multidimensional LC-

MS/MS) and 2D-PAGE/MS, can be applied to identify putative candidate markers (left).

Candidate protein markers can be subsequently verified and confirmed by targeted proteomics using standard ELISA methods or multiple reaction monitoring (MRM)-MS

(right). MRM-MS offers the advantages of an antibody-independent platform with capabilities for multiplexing.

2.5 TABLES

Surrogate Marker Process or Source, Concentration in TBI Biofluids (ng/mL) Cell Type CSF Serum IL-10 Inflammation 0.001-0.060 (children [36]) 0.050-0150 [38] 0.002-0.005 (adult [,37]) S100B Astroglia 1.0-15.0 [39] 0.01-0.70 [37,40,41] GFAP Astroglia 9.0-22.0 [42] 0.14-15.0 [43] NFL/NFH/ P-NFH Axon 0.13-2.5 and 49–562 [44,45] NA MBP Axon/oligodendrocytes NA 0.50-100.0 [37,46] Tau / amyloid β Axon Tau: 0.035-5.72 (neonate, [47]) 0.91- 5.1 [54,55] 1,519.6 – 2,308 (adult, [48-52]) Aβ 1–42: 1.17 (adult, [53]) NSE Neuron 10-30 [37] 10-20 [46] UCH-L1 Neuron 20-300 [56] 1.0-15 [56] α-spectrin-II BDP Neuron + astroglia 0.0-100 [57] NA MAP-2 Neuron NA 0.04-0.06 [58]

Table 2.1: Candidate Marker Biofluid Concentrations

Selected examples of reported TBI markers for which some concentrations were found;

NA – not available.

2.6 REFERENCES

1. Alam SL, Atkins JF, Gesteland RF: Programmed ribosomal frameshifting: much

ado about knotting! Proc Natl Acad Sci U S A 1999, 96:14177–14179.

2. Poste G: Bring on the biomarkers. Nature 2011, 469:156–157.

3. Anderson L: Candidate-based proteomics in the search for biomarkers of

cardiovascular disease. J Physiol 2005, 563:23–60.

4. Spaethling JM, Geddes-Klein DM, Miller WJ, von Reyn CR, Singh P, Mesfin M,

Bernstein SJ, Meaney DF: Linking impact to cellular and molecular sequelae of

CNS injury: modeling in vivo complexity with in vitro simplicity. Prog Brain Res

2007, 161:27–39.

5. Kazanis I: CNS injury research; reviewing the last decade: methodological errors

and a proposal for a new strategy. Brain Res Brain Res Rev 2005, 50:377–386.

6. Klose J: Protein mapping by combined isoelectric focusing and electrophoresis of

mouse tissues. Humangenetik 1975, 26:231–243.

7. O'Farrell PH: High resolution two-dimensional electrophoresis of proteins.

J Biol Chem 1975, 250:4007–4021.

8. Nielsen ML, Bennett KL, Larsen B, Moniatte M, Mann M: Peptide end sequencing

by orthogonal MALDI tandem mass spectrometry. J Proteome Res 2002, 1:63–

71.

9. Washburn MP, Wolters D, Yates JR: Large-scale analysis of the yeast proteome

by multidimensional protein identification technology. Nature Biotechnol 2001,

19:242–247.

10. Wolters DA, Washburn MP, Yates JR III: An automated multidimensional protein

identification technology for shotgun proteomics. Anal Chem 2001, 73:5683–

5690.

11. Matt P, Fu Z, Fu Q, Van Eyk JE: Biomarker discovery: proteome fractionation and

separation in biological samples. Physiol Genomics 2007, 33:12–17.

12. Parker CE, Pearson TW, Anderson NL, Borchers CH: Mass-spectrometry-based

clinical proteomics – a review and prospective. Analyst 1830, 2010:135.

13. Siman R, McIntosh TK, Soltesz KM, Chen Z, Neumar RW, Roberts VL: Proteins

released from degenerating neurons are surrogate markers for acute brain

damage. Neurobiol Dis 2004, 16:311–320.

14. Siman R, Toraskar N, Dang A, McNeil E, McGarvey M, Plaum J, Maloney E,

Grady MS: A panel of neuron-enriched proteins as markers for traumatic brain

injury in humans. J Neurotrauma 2009, 26:1867–1877.

15. Rifai N, Gillette MA, Carr SA: Protein biomarker discovery and validation: the long

and uncertain path to clinical utility. Nat Biotechnol 2006, 24:971–983.

16. Shapiro SS, Wilk MB: An analysis of variance test for normality (complete

samples). Biometrika 1965, 52:591–611.

17. Borchers CH, Parker CE: Improving the biomarker pipeline. Clin Chem 2010,

56:1786–1788.

18. Faul M, Xu L, Wald MM, Coronado VG: Traumatic Brain Injury in the United

States: Emergency Department Visits, Hospitalizations and Deaths 2002–2006.

Atlanta, GA: Centers for Disease Control and Prevention, National Center for

Injury Prevention and Control; 2010. http://www.cdc.gov/traumaticbraininjury/

pdf/blue_book.pdf.

19. National Institute of Neurological Disorders and Stroke: Traumatic Brain Injury:

Hope Through Research. 2002. http://www.ninds.nih.gov/disorders/

tbi/tbi_htr.pdf.

20. Risdall JE, Menon DK: Traumatic brain injury. Phil Trans Royal Soc London,

Series B, Biol Sci 2011, 366:241–250.

21. Gordon WA, Brown M, Sliwinski M, Hibbard MR, Patti N, Weiss MJ, Kalinsky R,

Sheerer M: The enigma of "hidden" traumatic brain injury. J Head Trauma Rehabil

1998, 13:39–56.

22. North SH, Shriver-Lake LC, Taitt CR, Ligler FS: Rapid Analytical Methods for On-

Site Triage for Traumatic Brain Injury. Ann Rev Anal Chem 2012, 5:35–56.

23. Squier W: The "Shaken Baby" syndrome: pathology and mechanisms.

Acta Neuropathol 2011, 122:519–542.

24. Geddes JF, Tasker RC, Hackshaw AK, Nickols CD, Adams GGW, Whitwell HL,

Scheimberg I: Dural haemorrhage in non-traumatic infant deaths: does it explain

the bleeding in 'shaken baby syndrome'? Neuropath Appl Neuro 2003, 29:14–22.

25. Laposata ME, Laposata M: Children with signs of abuse: when is it not child

abuse? Am J Clin Pathol 2005, 123 (Suppl): S119–S124.

26. Yokobori S, Hosein K, Burks S, Sharma I, Gajavelli S, Bullock R: Biomarkers for

the clinical differential diagnosis in traumatic brain injury–a systematic review.

CNS Neurosci Ther 2013, 19:556–565.

27. Bettermann K, Slocomb JE: Clinical Relevance of Biomarkers for Traumatic Brain

Injury. In Biomarkers for Traumatic Brain Injury. Edited by Thurston D.

Cambridge: Royal Society of Chemistry; 2012:1–18.

28. Cobb S, Battin B: Second-impact syndrome. J Sch Nurs 2004, 20:262–267.

29. Baugh CM, Stamm JM, Riley DO, Gavett BE, Shenton ME, Lin A, Nowinski CJ,

Cantu RC, McKee AC, Stern RA: Chronic traumatic encephalopathy:

neurodegeneration following repetitive concussive and subconcussive brain

trauma. Brain Imaging Behav 2012, 6:244–254.

30. Stein TD, Alvarez VE, McKee AC: Chronic traumatic encephalopathy: a spectrum

of neuropathological changes following repetitive brain trauma in athletes and

military personnel. Alzheimers Res Ther 2014, 6:4.

31. Kennedy JE, Leal FO, Lewis JD, Cullen MA, Amador RR: Posttraumatic stress

symptoms in OIF/OEF service members with blast-related and non-blast-related

mild TBI. Neurorehabil 2010, 26:223–231.

32. Cifu DX, Taylor BC, Carne WF, Bidelspach D, Sayer NA, Scholten J, Campbell

EH: Traumatic brain injury, posttraumatic stress disorder, and pain diagnoses in

OIF/OEF/OND Veterans. J Rehabil Res Dev 2014, 50:1169–1176.

33. Zhou Y, Kierans A, Kenul D, Ge Y, Rath J, Reaume J, Grossman RI, Lui YW: Mild

traumatic brain injury: longitudinal regional brain volume changes. Radiology

2013, 267:880–890.

34. Small GW, Kepe V, Siddarth P, Ercoli LM, Merrill DA, Donoghue N, Bookheimer

SY, Martinez J, Omalu B, Bailes J, Carrio J: PET scanning of brain tau in retired

national football league players: preliminary findings. Am J Geriat Psych 2013,

12:138–144.

35. Giza CC, Kutcher JS, Ashwal S, Barth J, Getchius TSD, Gioia GA, Gronseth GS,

Guskiewicz K, Mandel S, Manley G, McKeag DB, Thurman DJ, Zafonte R:

Summary of evidence-based guideline update: evaluation and management of

concussion in sports: Report of the Guideline Development Subcommittee of the

American Academy of Neurology. Neurology 2013, 80:2250–2257.

36. Bakay RAE, Ward AA Jr: Enzymatic changes in serum and cerebrospinal fluid in

neurological I njury. J Neurosurg 1983, 58:27–37.

37. Diaz-Arrastia R, Wang KK, Papa L, Sorani MD, Yue JK, Puccio AM, McMahon

PJ, Inoue T, Yuh EL, Lingsma HF, Maas AI, Valadka AB, Okonkwo DO, Manley,

The Track-Tbi Investigators GT, Casey IS, Cheong M, Cooper SR, Dams-

O'Connor K, Gordon WA, Hricik AJ, Menon DK, Mukherjee P, Schnyer DM, Sinha

TK, Vassar MJ: Acute biomarkers of traumatic brain injury: relationship between

plasma levels of ubiquitin C-Terminal Hydrolase-L1 and Glial Fibrillary Acidic

Protein. J Neurotrauma 2014, 31:19–25.

38. Light M, Minor KH, DeWitt P, Jasper KH, Davies SJ: Multiplex array proteomics

detects increased MMP-8 in CSF after spinal cord injury. J Neuroinflammation

2012, 9:122.

39. Yan X, Liu J, Luo Z, Ding Q, Mao X, Yan M, Yang S, Hu X, Huang J, Luo Z:

Proteomic profiling of proteins in rat spinal cord induced by contusion injury.

Neurochem Int 2010, 56:971–983.

40. Boutte AM, Yao C, Kobeissy F, May Lu XC, Zhang Z, Wang KK, Schmid K,

Tortella FC, Dave JR: Proteomic analysis and brain-specific systems biology in a

rodent model of penetrating ballistic-like brain injury. Electrophoresis 2012,

33:3693–3704.

41. Sjodin MO, Bergquist J, Wetterhall M: Mining ventricular cerebrospinal fluid from

patients with traumatic brain injury using hexapeptide ligand libraries to search

for trauma biomarkers. J Chromatogr B Analyt Technol Biomed Life Sci 2010,

878:2003. 2012.

42. Crawford F, Crynen G, Reed J, Mouzon B, Bishop A, Katz B, Ferguson S, Phillips

J, Ganapathi V, Mathura V, Roses A, Mullan M: Identification of plasma

biomarkers of TBI outcome using proteomic approaches in an APOE mouse

model. J Neurotrauma 2012, 29:246–260.

43. Ottens AK, Bustamante L, Golden EC, Yao C, Hayes RL, Wang KK, Tortella FC,

Dave JR: Neuroproteomics: a biochemical means to discriminate the extent and

modality of brain injury. J Neurotrauma 2010, 27:1837–1852.

44. Guingab-Cagmat JD, Newsom K, Vakulenko A, Cagmat EB, Kobeissy FH,

Zoltewicz S, Wang KK, Anagli J: In vitro MS-based proteomic analysis and

absolute quantification of neuronal-glial injury biomarkers in cell culture system.

Electrophoresis 2012, 33:3786–3797.

45. Woodcock T, Morganti-Kossmann MC: The Role of Markers of Inflammation in

Traumatic Brain Injury. Front Neurol 2013, 4:18.

46. Yan EB, Satgunaseelan L, Paul E, Bye N, Nguyen P, Agyapomaa D, Kossmann

T, Rosenfeld JV, Morganti-Kossmann MC: Post-Traumatic Hypoxia Is Associated

with Prolonged Cerebral Cytokine Production, Higher Serum Biomarker Levels,

and Poor Outcome in Patients with Severe Traumatic Brain Injury. J Neurotrauma

2014. doi:10.1089/neu.2013.3087.

47. Kadhim HJ, Duchateau J, Sebire G: Cytokines and brain injury: invited review. J

Intensive Care Med 2008, 23:236–249.

48. Bell MJ, Kochanek PM, Doughty LA, Carcillo JA, Adelson PD, Clark RS,

Wisniewski SR, Whalen MJ, DeKosky ST: Interleukin-6 and interleukin-10 in

cerebropinal fluid after severe traumatic brain injury in children. J Neurotrauma

1997, 14:451–457.

49. Schneider Soares FM, Menezes de Souza N, Libório Schwarzbold M, Paim Diaz

A, Costa Nunes J, Hohl A, Nunes Abreu da Silva P, Vieira J, Lisboa de Souza R,

Moré Bertotti M, Schoder Prediger RD, Neves Linhares M, Bafica A, Walz R:

Interleukin-10 Is an independent biomarker of severe traumatic brain injury

prognosis. Neuroimmunomodulation 2012, 19:377–385.

50. Kamm K, VanderKolk W, Lawrence C, Jonker M, Davis AT: The effect of

traumatic brain injury upon the concentration and expression of interleukin-1β and

Interleukin-10 in the Rat. J Trauma: Injury, Infection, Critical Care 2006, 60:152–

157.

51. Goyal A, Failla MD, Niyonkuru C, Amin K, Fabio A, Berger RP, Wagner AK:

S100b as a prognostic biomarker in outcome prediction for patients with severe

traumatic brain injury. J Neurotraum 2013, 30:946–957.

52. Egea-Guerrero JJ, Murillo-Cabezas F, Gordillo-Escobar E, Rodríguez-Rodríguez

A, Enamorado-Enamorado J, Revuelto-Rey J, Pacheco-Sánchez M, León-Justel

A, Domínguez-Roldán JM, Vilches-Arenas A: S100B protein may detect brain

death development after severe traumatic brain injury. J Neurotraum 2013,

30:1762–1769.

53. Petzold A, Keir G, Lim D, Smith M, Thompson EJ: Cerebrospinal fluid (CSF) and

serum S100B: release and wash-out pattern. Brain Research Bulletin 2003,

61:281–285.

54. Fraser DD, Close TE, Rose KL, Ward R, Mehl M, Farrell C, Lacroix J, Creery D,

Kesselman M, Stanimirovic D, Hutchison JS: Severe traumatic brain injury in

children elevates glial fibrillary acidic protein in cerebrospinal fluid and serum*.

Pediatric Critical Care Medicine 2011, 12:319–324.

55. Nylén K, Öst M, Csajbok LZ, Nilsson I, Blennow K, Nellgård B, Rosengren L:

Increased serum-GFAP in patients with severe traumatic brain injury is related to

outcome. J Neurol Sci 2006, 240:85–91.

56. Neselius S, Brisby H, Theodorsson A, Blennow K, Zetterberg H, Marcusson J:

CSF-biomarkers in Olympic boxing: diagnosis and effects of repetitive head

trauma. PLoS One 2012, 7: e33606.

57. Neselius S, Zetterberg H, Blennow K, Marcusson J, Brisby H: Increased CSF

levels of phosphorylated neurofilament heavy protein following bout in amateur

boxers. PLoS One 2013, 8: e81249.

58. Berger RP, Adelson PD, Pierce MC, Dulani T, Cassidy LD, Kochanek PM: Serum

neuron-specific enolase, S100B, and myelin basic protein concentrations after

inflicted and noninflicted traumatic brain injury in children. J Neurosurg (Pediatrics

1) 2005, 103:61–68.

59. Takahashi K, Hasegawa S, Maeba S, Fukunaga S, Motoyama M, Hamano H,

Ichiyama T: Serum tau protein level serves as a predictive factor for neurological

prognosis in neonatal asphyxia. Brain and Development 2013.

10.1016/j.braindev.2013.10.007.

60. Liliang P-C, Liang C-L, Weng H-C, Lu K, Wang K-W, Chen H-J, Chuang J-H: Tau

proteins in serum predict outcome after severe traumatic brain injury. J Surg Res

2010, 160:302–307.

61. Neselius S, Zetterberg H, Blennow K, Randall J, Wilson D, Marcusson J, Brisby

H: Olympic boxing is associated with elevated levels of the neuronal protein tau

in plasma. Brain Inj 2013, 27:425–433.

62. Franz G, Beer R, Kampfl A, Engelhardt K, Schmutzhard E, Ulmer H,

Deisenhammer F: Amyloid beta 1–42 and tau in cerebrospinal fluid after severe

traumatic brain injury. Neurology 2003, 60:1457–1461.

63. Shiiya N, Kunihara T, Miyatake T, Matsuzaki K, Yasuda K: Tau protein in the

cerebrospinal fluid is a marker of brain injury after aortic surgery. Ann Thorac Surg

2004, 77:2034–2038.

64. Tsitsopoulos PP, Marklund N: Amyloid-beta peptides and tau protein as

biomarkers in cerebrospinal and interstitial fluid following traumatic brain injury: a

review of experimental and clinical studies. Front Neurol 2013, 4:79.

65. Zemlan FP, Rosenberg WS, Luebbe PA, Campbell TA, Dean GE, Weiner NE,

Cohen JA, Rudick RA, Woo D: Quantification of axonal damage in traumatic brain

injury: affinity purification and characterization of cerebrospinal fluid tau proteins.

J Neurochem 1999, 72:741–750.

66. Emmerling MR, Morganti-Kossmann MC, Kossmann T, Stahel PF, Watson MD,

Evans LM, Mehta PD, Spiegel K, Kuo YM, Roher AE, Raby CA: Traumatic brain

injury elevates the Alzheimer's amyloid peptide A beta 42 in human CSF. A

possible role for nerve cell injury. Ann N Y Acad Sci 2000, 903:118–122.

67. Raby CA, Morganti-Kossmann MC, Kossmann T, Stahel PF, Watson MD, Evans

LM, Mehta PD, Spiegel K, Kuo YM, Roher AE, Emmerling MR: Traumatic brain

injury increases beta-amyloid peptide 1–42 in cerebrospinal fluid. J Neurochem

1998, 71:2505–2509.

68. Mondello S, Akinyi L, Buki A, Robicsek SA, Gabrielli A, Tepas J, Papa L, Brophy

GM, Tortella F, Hayes RL, Wang KKW: Clinical Utility of Serum Levels of

Ubiquitin-C Terminal Hydrolase as a Biomarker for Severe Traumatic Brain Injury.

Neurosurgery 2012, 70:666–675.

69. Mondello S, Robicsek SA, Gabrielli A, Brophy GM, Papa L, Tepas J, Robertson

C, Buki A, Scharf D, Jixiang M, Akinyi L, Muller U, Wang KKW, Hayes RL: αII-

Spectrin Breakdown Products (SBDPs): Diagnosis and Outcome in Severe

Traumatic Brain Injury Patients. J Neurotrauma 2010, 27:1203–1213.

70. Mondello S, Gabrielli A, Catani S, D’Ippolito M, Jeromin A, Ciaramella A, Bossù

P, Schmid K, Tortella F, Wang KKW, Hayes RL, Formisano R: Increased levels

of serum MAP-2 at 6-months correlate with improved outcome in survivors of

severe traumatic brain injury. Brain Injury 2012, 26:1629–1635.

71. Whalen MJ, Dalkara T, You Z, Qiu J, Bermpohl D, Mehta N, Suter B, Bhide PG,

Lo EH, Ericsson M, Moskowitz MA: Acute plasmalemma permeability and

protracted clearance of injured cells after controlled cortical impact in mice. J

Cereb Blood Flow Metab 2008, 28:490–505.

72. Farkas O, Lifshitz J, Povlishock JT: Mechanoporation induced by diffuse

traumatic brain injury: an irreversible or reversible response to injury? J Neurosci

2006, 26:3130–3140.

73. Barbee KA: Mechanical cell injury. Ann N Y Acad Sci 2005, 1066:67–84.

74. Binder LI, Frankfurter A, Rebhun LI: The Distribution of Tau in the Mammalian

Central Nervous System. J Cell Biol 1985, 101:1371–1378.

75. Bitsch A, Horn C, Kemmling Y, Seipelt M, Hellenbrand U, Stiefel M, Ciesielczyk

B, Cepek L, Bahn E, Ratzka P, Prange H, Otto M: Serum Tau protein level as a

marker of axonal damage in acute ischemic stroke. European Neurology 2001,

47:45–51.

76. Arroyo EJ, Schere SS: On the molecular architecture of myelinated fibers.

Histochem Cell Biol 2000, 133:1–18.

77. Woertgen C, Rothoerl RD, Holzschuh M, Metz C, Brawanski A: Comparison of

serial S-100 and NSE serum measurements after severe head injury. Acta

Neurochirurgica 1997, 139:1161–1165.

78. Yamazaki Y, Yada K, Morii S, Kitahara T, Ohwada T: Diagnostic significance of

serum neuron-specific enolase and myelin basic protein assay in patients with

acute head injury. Surgical Neurology 1995, 43:267–270.

79. Liu MC, Akle V, Zheng W, Kitlen J, O'Steen B, Larner SF, Dave JR, Tortella FC,

Hayes RL, Wang KKW: Extensive degradation of myelin basic protein isoforms

by calpain following traumatic brain injury. J Neurochem 2006, 98:700–712.

80. Ottens AK, Golden EC, Bustamante L, Hayes RL, Denslow ND, Wang KKW:

Proteolysis of multiple myelin basic protein isoforms after neurotrauma:

characterization by mass spectrometry. J Neurochem 2008, 104:1404–1414.

81. Ross SA, Cunningham RT, Johnston CF, Rowlands BJ: Neuron-specific enolase

as an aid to outcome predicition in head injury. Brit J Neurosurg 1996, 10:471–

476.

82. Johnsson P: Markers of cerebral ischemia after cardiac surgery.

J Cardiothorac Vasc Anesth 1996, 10:120–126.

83. Caceres A, Payne MR, Binder LI, Steward O: Immunocytochemical Localization

of Actin and Microtubule-Associated Protein MAP2 in Dendritic Spines. Proc Natl

Acad Sci USA 1983, 80:1738–1742.

84. Kitagawa K, Matsumoto M, Niinobe M, Mikoshiba K, Hata R, Ueda H, Handa N,

Fukunaga R, Isaka Y, Kimura K, Kamada T: Microtubule-Associated Protein 2 as

a sensitive marker for cerebral ischemic damage-immunohistochemical

invetigation of dendritic damage. Neuroscience 1989, 31:401–411.

85. Berger RP, Hayes RL, Richichi R, Beers SR, Wang KKW: Serum Concentrations

of Ubiquitin C-Terminal Hydrolase-L1 and αII-Spectrin Breakdown Product 145

kDa Correlate with Outcome after Pediatric TBI. J Neurotrauma 2012, 29:162–

167.

86. Liu MC, Akinyi L, Scharf D, Mo JX, Larner SF, Muller U, Oli MW, Zheng WR,

Kobeissy F, Papa L, Lu XC, Dave JR, Tortella FC, Hayes RL, Wang KKW:

Ubiquitin C-terminal hydrolase-L1 as a biomarker for ischemic and traumatic brain

injury in rats. Eur J Neurosci 2010, 31:722–732.

87. Siman R, Giovannone N, Toraskar N, Frangos S, Stein SC, Levine JM, Kumar

MA: Evidence that a panel of neurodegeneration biomarkers predicts vasospasm,

infarction, and outcome in aneurysmal subarachnoid hemorrhage. PLoS One

2011, 6:e28938.

88. Pang L, Wu Y, Dong N, Xu DH, Wang DW, Wang ZH, Li XL, Bian M, Zhao HJ,

Liu XL, Zhang N: Elevated serum ubiquitin C-terminal hydrolase-L1 levels in

patients with carbon monoxide poisoning. Clin Biochem 2014, 47:72–76.

89. Papa L, Akinyi L, Liu MC, Pineda JA, Tepas JJ, Oli MW, Zheng W, Robinson G,

Robicsek SA, Gabrielli A, Heaton SC, Hannay HJ, Demery JA, Brophy GM, Layon

J, Robertson CS, Hayes RL, Wang KKW: Ubiquitin C-terminal hydrolase is a

novel biomarker in humans for severe traumatic brain injury. Crit Care Med 2010,

38:138–144.

90. Pike BR, Flint J, Dutta S, Johnson E, Wang KKW, Hayes RL: Accumulation of

non-erythroid αII-spectrin and calpain-cleaved αII-spectrin breakdown products in

cerebrospinal fluid after traumatic brain injury in rats. J Neurochem 2001,

78:1297–1306.

91. Tomas M, Marin MP, Portoles M, Megias L, Gomez-Lechon MJ, Renau-Piqueras

J: Ethanol affects calmodulin and the calmodulin-binding proteins neuronal nitric

oxide synthase and alphaII-spectrin (alpha-fodrin) in the nucleus of growing and

differentiated rat astrocytes in primary culture. Toxicol In Vitro 2007, 21:1039–

1049.

92. Siman R, Giovannone N, Hanten G, Wilde EA, McCauley SR, Hunter JV, Li X,

Levin HS, Smith DH: Evidence That the Blood Biomarker SNTF Predicts Brain

Imaging Changes and Persistent Cognitive Dysfunction in Mild TBI Patients.

Front Neurol 2013, 4:190.

93. Riederer BM, Zagon IS, Goodman SR: Brain spectrin (240/235) and brain

spectrin(240/235E): two distinct spectrin subtypes with different locations within

Mammalian neural cells. J Cell Biol 1986, 102:2088–2097.

94. Azevedo FA, Carvalho LR, Grinberg LT, Farfel JM, Ferretti RE, Leite RE, Jacob

Filho W, Lent R, Herculano-Houzel S: Equal numbers of neuronal and

nonneuronal cells make the human brain an isometrically scaled-up primate

brain. J Comp Neurol 2009, 513:532–541.

95. Reier PJ, Houle JD: The glial scar: its bearing on axonal elongation and

transplantation approaches to CNS repair. Adv Neurol 1988, 47:87–138.

96. McGraw J, Hiebert GW, Steeves JD: Modulating astrogliosis after neurotrauma.

J Neurosci Res 2001, 63:109–115.

97. Wanner IB, Anderson MA, Song B, Levine J, Fernandez A, Gray-Thompson Z,

Ao Y, Sofroniew MV: Glial scar borders are formed by newly proliferated,

elongated astrocytes that interact to corral inflammatory and fibrotic cells via

STAT3-dependent mechanisms after spinal cord injury. J Neurosci 2013,

33:12870–12886.

98. Sofroniew MV, Vinters HV: Astrocytes: biology and pathology. Acta Neuropathol

2010, 119:7–35.

99. Burda JE, Sofroniew MV: Reactive gliosis and the multicellular response to CNS

damage and disease. Neuron 2014, 81:229–248.

100. Li DR, Zhang F, Wang Y, Tan XH, Qiao DF, Wang HJ, Michiue T, Maeda H:

Quantitative analysis of GFAP- and S100 protein-immunopositive astrocytes to

investigate the severity of traumatic brain injury. Legal Med 2012, 14:84–92.

101. Gelot A, Villapol S, Billette de Villemeur T, Renolleau S, Charriaut-Marlangue C:

Astrocytic demise in the developing rat and human brain after hypoxic-ischemic

damage. Dev Neurosci 2009, 31:459–470.

102. Chen Y, Swanson RA: Astrocytes and brain injury. J Cereb Blood Flow Metab

2003, 23:137–149.

103. Liu D, Smith CL, Barone FC, Ellison JA, Lysko PG, Li K, Simpson IA: Astrocytic

demise precedes delayed neuronal death in focal ischemic rat brain. Brain Res

Mol Brain Res 1999, 68:29–41.

104. Colombo JA, Yanez A, Lipina SJ: Interlaminar astroglial processes in the cerebral

cortex of non human primates: response to injury. J Hirnforsch 1997, 38:503–

512.

105. Colombo JA, Yanez A, Puissant V, Lipina S: Long, interlaminar astroglial cell

processes in the cortex of adult monkeys. J Neurosci Res 1995, 40:551–556.

106. Shakeri M, Mahdkhah A, Panahi F: S100B protein as a post-traumatic biomarker

for prediction of brain death in association with patient outcomes. Arch Trauma

Res 2013, 2:76–80.

107. Korfias S, Stranjalis G, Psachoulia C, Vasiliadis C, Pitaridis M, Boviatsis E, Sakas

DE: Slight and short-lasting increase of serum S-100B protein in extra-cranial

trauma. Brain Injury 2006, 20:867–872.

108. Thelin EP, Johannesson L, Nelson D, Bellander B-M: S100B is an important

outcome predictor in traumatic brain injury. J Neurotraum 2013, 30:519–528.

109. Marchi N, Cavaglia M, Fazio V, Bhudia S, Hallene K, Janigro D: Peripheral

markers of blood–brain barrier damage. Clinica Chimica Acta 2004, 342:1–12.

110. Mancardi GL, Cadoni A, Tabaton M, Schenone A, Zicca A, De Martini I, Bianchini

D, Damiani G, Zaccheo D: Schwann cell GFAP expression increases in axonal

neuropathies. J Neurol Sci 1991, 102:177–183.

111. Pellitteri R, Spatuzza M, Stanzani S, Zaccheo D: Biomarkers expression in rat

olfactory ensheathing cells. Front Biosci 2010, 2:289–298.

112. Eng LF, Ghirnikar RS: GFAP and astrogliosis. Brain Pathol 1994, 4:229–237.

113. Lumpkins KM, Bochicchio GV, Keledjian K, Simard JM, McCunn M, Scalea T:

Glial fibrillary acidic protein is highly correlated with brain injury. J Trauma 2008,

65:778–782. discussion 782–774.

114. Papa L, Lewis LM, Falk JL, Zhang Z, Silvestri S, Giordano P, Brophy GM, Demery

JA, Dixit NK, Ferguson I, Liu MC, Mo J, Akinyi L, Schmid K, Mondello S,

Robertson CS, Tortella FC, Hayes RL, Wang KKW: Elevated levels of serum glial

fibrillary acidic protein breakdown products in mild and moderate traumatic brain

injury are associated with intracranial lesions and neurosurgical intervention. Ann

Emerg Med 2012, 59:471–483.

115. Okonkwo DO, Yue JK, Puccio AM, Panczykowski DM, Inoue T, McMahon PJ,

Sorani MD, Yuh EL, Lingsma HF, Maas AI, Valadka AB, Manley, Transforming

R, Clinical Knowledge In Traumatic Brain Injury Investigators Including GT, Casey

SS, Cheong M, Cooper SR, Dams-O'Connor K, Gordon WA, Hricik AJ,

Hochberger K, Menon DK, Mukherjee P, Sinha TK, Schnyer DM, Vassar MJ:

GFAP-BDP as an acute diagnostic marker in traumatic brain injury: results from

the prospective transforming research and clinical knowledge in traumatic brain

injury study. J Neurotrauma 2013, 30:1490–1497.

116. Zoltewicz JS, Scharf D, Yang B, Chawla A, Newsom KJ, Fang L: Characterization

of Antibodies that Detect Human GFAP after Traumatic Brain Injury. Biomarker

Insights 2012, 7:71–79.

117. Lafrenaye AD, McGinn MJ, Povlishock JT: Increased intracranial pressure after

diffuse traumatic brain injury exacerbates neuronal somatic membrane poration

but not axonal injury: evidence for primary intracranial pressure-induced neuronal

perturbation. J Cereb Blood Flow Metab 2012, 32:1919–1932.

118. Thompson HJ, Lifshitz J, Marklund N, Grady MS, Graham DI, Hovda DA,

McIntosh TK: Lateral fluid percussion brain injury: a 15-year review and

evaluation. J Neurotrauma 2005, 22:42–75.

119. Smith DH, Soares HD, Pierce JS, Perlman KG, Saatman KE, Meaney DF, Dixon

CE, McIntosh TK: A model of parasagittal controlled cortical impact in the mouse:

cognitive and histopathologic effects. J Neurotrauma 1995, 12:169–178.

120. Foda MA, Marmarou A: A new model of diffuse brain injury in rats. Part II:

Morphological characterization. J Neurosurg 1994, 80:301–313.

121. Marmarou A, Foda MA, van den Brink W, Campbell J, Kita H, Demetriadou K: A

new model of diffuse brain injury in rats. Part I: Pathophysiology and

biomechanics. J Neurosurg 1994, 80:291–300.

122. Risling M, Davidsson J: Experimental animal models for studies on the

mechanisms of blast-induced neurotrauma. Front Neurol 2012, 3:30.

123. Kochanek PM, Bramlett H, Dietrich WD, Dixon CE, Hayes RL, Povlishock J,

Tortella FC, Wang KK: A novel multicenter preclinical drug screening and

biomarker consortium for experimental traumatic brain injury: operation brain

trauma therapy. J Trauma 2011, 71: S15–S24.

124. Yue JK, Vassar MJ, Lingsma HF, Cooper SR, Okonkwo DO, Valadka AB, Gordon

WA, Maas AI, Mukherjee P, Yuh EL, Puccio AM, Schnyer DM, Manley GT, Track-

Tbi I, Casey SS, Cheong M, Dams-O'Connor K, Hricik AJ, Knight EE, Kulubya

ES, Menon DK, Morabito DJ, Pacheco JL, Sinha TK: Transforming research and

clinical knowledge in traumatic brain injury pilot: multicenter implementation of the

common data elements for traumatic brain injury. J Neurotrauma 2013, 30:1831–

1844.

125. Hicks R, Giacino J, Harrison-Felix C, Manley G, Valadka A, Wilde EA: Progress

in developing common data elements for traumatic brain injury research: version

two–the end of the beginning. J Neurotrauma 2013, 30:1852–1861.

126. Manley GT, Diaz-Arrastia R, Brophy M, Engel D, Goodman C, Gwinn K, Veenstra

TD, Ling G, Ottens AK, Tortella F, Hayes RL: Common data elements for

traumatic brain injury: recommendations from the biospecimens and biomarkers

working group. Arch Phys Med Rehabil 2010, 91:1667–1672.

127. Lubieniecka JM, Streijger F, Lee JH, Stoynov N, Liu J, Mottus R, Pfeifer T, Kwon

BK, Coorssen JR, Foster LJ, Grigliatti TA, Tetzlaff W: Biomarkers for severity of

spinal cord injury in the cerebrospinal fluid of rats. PLoS One 2011, 6: e19247.

128. Svetlov SI, Prima V, Kirk DR, Gutierrez H, Curley KC, Hayes RL, Wang KKW:

Morphologic and biochemical characterization of brain injury in a model of

controlled blast overpressure exposure. J Trauma: Injury, Infection, Critical Care

2010, 69:795–804.

129. Svetlov SI, Prima V, Glushakova O, Svetlov A, Kirk DR, Gutierrez H, Serebruany

VL, Curley KC, Wang KK, Hayes RL: Neuro-glial and systemic mechanisms of

pathological responses in rat models of primary blast overpressure compared to

"composite" blast. Front Neurol 2012, 3:15.

130. Ahmed FA, Kamnaksh A, Kovesdi E, Long JB, Agoston DV: Long-term

consequences of single and multiple mild blast exposure on select physiological

parameters and blood-based biomarkers. Electrophoresis 2013, 34:2229–2233.

131. Ahmed F, Gyorgy A, Kamnaksh A, Ling G, Tong L, Parks S, Agoston D:

Time-dependent changes of protein biomarker levels in the cerebrospinal fluid

after blast traumatic brain injury. Electrophoresis 2012, 33:3705–3711.

132. Gyorgy A, Ling G, Wingo D, Walker J, Tong L, Parks S, Januszkiewicz A, Baumann

R, Agoston DV: Time-dependent changes in serum biomarker levels after blast

traumatic brain injury. J Neurotrauma 2011, 28:1121–1126.

133. Zetterberg H, Smith DH, Blennow K: Biomarkers of mild traumatic brain injury in

cerebrospinal fluid and blood. Nature Rev Neurol 2013, 9:201–210.

134. Guingab-Cagmat JD, Cagmat EB, Hayes RL, Anagli J: Integration of proteomics,

bioinformatics, and systems biology in traumatic brain injury biomarker discovery.

Front Neurol 2013, 4:61.

135. Agoston DV, Risling M, Bellander BM: Bench-to-bedside and bedside back to the

bench; coordinating clinical and experimental traumatic brain injury studies. Front

Neurol 2012, 3:3.

136. Hanrieder J, Wetterhall M, Enblad P, Hillered L, Bergquist J: Temporally resolved

differential proteomic analysis of human ventricular CSF for monitoring traumatic

brain injury biomarker candidates. J Neurosci Meth 2009, 177:469–478.

137. Ross PL: Multiplexed Protein Quantitation in Saccharomyces cerevisiae Using

Amine-reactive Isobaric Tagging Reagents. Mol Cell Proteomics 2004, 3:1154–

1169.

138. Wiese S, Reidegeld KA, Meyer HE, Warscheid B: Protein labeling by iTRAQ: A

new tool for quantitative mass spectrometry in proteome research. Proteomics

2007, 7:340–350.

139. Wanner IB: An in vitro trauma model to study rodent and human astrocyte

reactivity. Methods Mol Biol 2012, 814:189–219.

140. Wanner IB, Deik M, Torres M, Rosendahl AR, Neary JT, Lemmon VP, Bixby JL: A

new in vitro model of the glial scar inhibits axon growth. Glia 2008, 56:1691–1709.

141. Sondej M, Doran P, Loo JA, Wanner I: Sample preparation of primary astrocyte

cellular and released proteins for 2-D gel electrophoresis and protein identification

by mass spectrometry. In Sample preparation in biological mass spectrometry.

Edited by Ivanov A, Lazarev A. Dordrecht: Springer; 2011:829–849.

142. Ellis EF, McKinney JS, Willoughby KA, Liang S, Povlishock JT: A new model for

rapid stretch-induced injury of cells in culture: characterization of the model using

astrocytes. J Neurotrauma 1995, 12:325–339.

143. Ellis EF, Willoughby KA, Sparks SA, Chen T: S100B protein is released from rat

neonatal neurons, astrocytes, and microglia by in vitro trauma and anti-S100

increases trauma-induced delayed neuronal injury and negates the protective

effect of exogenous S100B on neurons. J Neurochem 2007, 101:1463–1470.

144. Rzigalinski BA, Weber JT, Willoughby KA, Ellis EF: Intracellular free calcium

dynamics in stretch-injured astrocytes. J Neurochem 1998, 70:2377–2385.

145. Schenk S, Schoenhals GJ, de Souza G, Mann M: A high confidence, manually

validated human blood plasma protein reference set. BMC Med Genomics 2008,

1:41.

146. Omenn GS, States DJ, Adamski M, Blackwell TW, Menon R, Hermjakob H,

Apweiler R, Haab BB, Simpson RJ, Eddes JS, Kapp EA, Moritz RL, Chan DW, Rai

AJ, Admon A, Aebersold R, Eng J, Hancock WS, Hefta SA, Meyer H, Paik YK, Yoo

JS, Ping P, Pounds J, Adkins J, Qian X, Wang R, Wasinger V, Wu CY, Zhao X,

Zeng R, Archakov A, Tsugita A, Beer I, Pandey A, Pisano M, Andrews P, Tammen

H, Speicher DW, Hanash SM: Overview of the HUPO Plasma Proteome Project:

results from the pilot phase with 35 collaborating laboratories and multiple

analytical groups, generating a core dataset of 3020 proteins and a publicly-

available database. Proteomics 2005, 5:3226–3245.

147. Waybright TJ: Preparation of human cerebrospinal fluid for proteomics biomarker

analysis. Methods Mol Biol 2013, 1002:61–70.

148. Theilacker N, Roller EE, Barbee KD, Franzreb M, Huang X: Multiplexed protein

analysis using encoded antibody-conjugated microbeads. J Royal Soc Interf 2011,

8:1104–1113.

149. Kingsmore SF: Multiplexed protein measurement: technologies and applications

of protein and antibody arrays. Nat Rev Drug Discov 2006, 5:310–320.

150. Barr JR, Maggio VL, Patterson DG Jr, Cooper GR, Henderson LO, Turner WE,

Smith SJ, Hannon H, Needham LL, Sampson EJ: Isotope dilution-mass

spectrometric quantification of specific proteins: model application with

apoliprotein A-1. Clin Chem 1996, 42:1672–1682.

151. Gerber SA, Rush J, Stemman O, Kirschner MW, Gygi SP: Absolute quantification

of proteins and phosphoproteins from cell lysates by tandem MS. Proc Natl Acad

Sci U S A 2003, 100:6940–6945.

152. Roschinger W, Olgemoller B, Fingerhut R, Liebl B, Roscher AA: Advances in

analytical mass spectrometry to improve screening for inherited metabolic

diseases. Eur J Pediat 2003, 162: S67–S76.

153. Liao H, Wu J, Kuhn E, Chin W, Chang B, Jones MD, O'Neil S, Clauser KR, Karl J,

Hasler F, Roubenoff R, Zolg W, Guild BC: Use of mass spectrometry to identify

protein biomarkers of disease severity in the synovial fluid and serum of patients

with rheumatoid arthritis. Arth Rheum 2004, 50:3792–3803.

154. Struys EA, Jansen EEW, De Meer K, Jakobs C: Determination of S-

Adenosylmethionine and S-Adenosulhomocysteine in Plasma and Cerebrospinal

Fluid by Stable-Isotope DIlution Tandem Mass Spectrometry. Clin Chem 2000,

46:1650–1656.

155. Anderson L, Anderson NG, Haines LR, Hardie DB, Olafson RW, Pearson TW:

Mass Spectrometric Quantification of Peptides and Proteins Using Stable Isotope

Standards and Capture by Anti-peptide Antibodies (SISCAPA). J Proteome Res

2004, 3:235–244.

156. Ahn YH, Lee YJ, Lee YJ, Kim Y-S, Ko JH, Yoo JS: Quantitative Analysis of an

Aberrant Glycoform of TIMP1 from Colon Cancer Serum by L-PHA-Enrichment

and SISCAPA with MRM Mass Spectrometry. J Proteome Res 2009, 8:4216–

4224.

157. Zhang Y, Hao Z, Kellmann M, Huhmer A: HR/AM Targeted Peptide Quantitation

on a Q Exactive MS: a Unique Combination of High Selectivity, Sensitivity, and

Throughput. 2012. http://planetorbitrap.com/data/uploads/ZFS1334248563484_A

N554_63517_HRAM-Q_Exactive_0412S.pdf

CHAPTER 3: NEW ASTROGLIAL INJURY DEFINED BIOMARKERS FOR

NEUROTRAUMA ASSESSMENT

3.1 INTRODUCTION

Traumatic brain injury (TBI) is “the most complex disease known to man” (1). It is a global public health concern affecting over 2.5 million cases per year in the United

States alone and is the leading cause of death and disability among the youth (2). The spectrum of TBI covers a wide range of severities with multiple adverse outcomes (3).

Severe TBI, characterized by extended periods of coma, results in variable degrees of brain dysfunction that are difficult to predict (4). The most common TBIs are mild, and occur frequently particularly during sports practice and routine military operations. Some mild TBI patients develop persistent or even permanent neurological deficits, which would be desirable to predict (5, 6). Repeated sub-concussive impacts and cumulative mild TBIs can increase the risk for neurological deficits, so real-time diagnosis is essential for safe return-to play/duty to prevent repeated exposure of at-risk individuals (7).

Injury evolution and eventual outcome are difficult to prognosticate, and current approaches to assess TBI patients, based on Glasgow Coma Scale (GCS) scores or computed tomography (CT) scans, are often insufficient to adequately capture the TBI complexity (8-10). While CT scans are common, and readily report macroscopic brain lesions, the detection of diffuse microstructural injuries and metabolic dysfunction after

TBI require more refined structural and functional imaging techniques that are less commonly available (11, 12).

Information about microstructural injuries such as fiber damage and cell membrane permeability (mechanoporation) can provide insight into very early cellular and structural pathophysiology and are accompanied by protein alterations and degradation at the molecular level (13-19). We believe that further understanding of cellular injury mechanisms, and their protein signatures are a resource for sensitive and acute diagnostic tools, needed for accurate assessment of neurotrauma injury magnitude (20).

Multiple studies have been conducted to identify biomarker candidates that can offer superior diagnostic and prognostic information. GFAP, neuron specific enolase, neurofilaments, tau, ubiquitin C-terminal hydrolase (UCHL1) and S100β have been chosen based on neuropathological presence or known expression in neurons or glia without elucidating their trauma-induced release mechanisms. The utility of these biomarkers is limited by short-lived biofluid presence, extra-cranial sources, delayed circulatory appearance and age-dependent liabilities, all of which hamper their individual clinical translation (21-23). Most protein biomarker mining studies have drawbacks, including failure to address fluid changes, differences in time-scales between rodent and human pathophysiology and proteomic analytical challenges that obstruct the identification of new, low abundance biomarkers (24-28). Furthermore, selection of suitable biomarker candidates from untargeted (i.e., global) proteomic discovery experiments are difficult and are typically unsuccessful (29-32). These examples emphasize the need for a new class of biomarkers associated directly and immediately with a traumatic impact to the brain.

In this study, astrocytes were chosen, given their central roles in the neurovascular unit, brain metabolism, blood-flow and blood-brain barrier (BBB) (33-37).

Astrocytes outnumber neurons in the human neocortical white matter, are vulnerable and responsive to injury and thereby participate to white matter injury (16, 38-41). Significant amounts of cellular proteins were found to be released into fluids minutes after abrupt astrocyte stretch-injury, suggesting astrocytes as the ideal carriers for neurotrauma biomarkers (32).

In the present work, we determined a TBI CSF proteome and a strategy to overcome the typical ‘proteomics bottlenecks’. This selection strategy capitalized on astrocyte-enrichment and our previous work on trauma-released proteins in a simple culture injury model to identify a panel of new biomarker candidates (32, 37, 42). The study identified astroglial injury-defined (AID) biomarker release from traumatized human astrocytes, documents their elevation in TBI patients during the first post-injury days, and explores their presence in serum of mild TBI patients. Importantly, establishing biomarkers of cell wounding and cell death may provide future biosignatures of brain cell compromise and demise that could facilitate our understanding of TBI pathophysiology.

3.2 RESULTS

Cerebrospinal fluid of TBI patients carries a signature of trauma-released astroglial proteins.

Analytical liquid chromatography-tandem mass spectrometry (LC-MS/MS) identified proteins in CSF of 19 TBI patients and 9 control subjects. Among the 484 proteins identified in TBI CSF, 232 were unique to TBI, while 252 proteins overlapped with the control CSF proteome (Figure 3.1). To select neurotrauma biomarker candidates, we determined the overlap with previously published trauma-released proteins determined from fluids of stretch-injured mouse astrocytes using a 2D gel analysis approach (Supplement Figure S3.1) (32). To improve specificity, we then determined astrocyte-enriched proteins using gene array data, and excluded proteins present in healthy plasma or abundant in extracranial tissues (42-45). Four candidate astroglial injury-associated proteins were identified: aldolase C (ALDOC) which is one of the most astrocyte-enriched proteins and also one of the highest expressed proteins in the brain

(32, 46), glutamine synthetase (GS), brain lipid binding protein (BLBP), and astrocytic phosphoprotein 15 (PEA15).

Traumatized human astrocytes show membrane wounding, reactivity and cell death at different times post-injury.

Differentiated, serum-free human astrocytes grown on deformable membranes received pressure-pulses that produced diffuse shear and stretch injuries reminiscent of an abrupt traumatic force (32, 47, 48). Subpopulations of traumatized human astrocytes displayed membrane wounding, died or underwent reactivity by acquiring star-shaped

morphology and GFAP up-regulation (Figure 3.2) as previously shown (48-50). Wounded cells displayed beaded, fragmented and amputated processes with elevated GFAP signals as soon as 30min post-stretching, earlier than reported gene expression changes

(Figure 3.2C) (50, 51). By analyzing nuclear shape and membrane integrity, intact, wounded and demised astrocytes could be distinguished (see Methods, Figure 2D-F). In both mild and severely stretched cultures the population of wounded cells, increased 16- fold over controls 30min after injury, constituting a fraction of 20% of stretched astrocytes

(Figure 3.2E, S3.2A). By two days after stretching, this population decreased substantially. By contrast, the rate of cell death was low at 30min post- injury and continued to rise over time until two days post-injury, when cell death differed significantly between mild and severe pressure pulsed cultures (Figure 3.2F, S3.2B). Mechanical trauma caused severity-independent acute membrane wounding and protracted severity- dependent cell death.

Different biomarkers and release kinetics correlate with astrocyte wounding and cell death.

Fluid immunoblotting from control and stretched cultures provided release kinetics of AID biomarkers over time (Figure 3.3A). GFAP, ALDOC, BLBP and PEA15 levels showed logarithmic release levels up to three orders of magnitude (Figure 3.3B-E). GFAP fluid levels show clear temporal and severity differences, while ALDOC, BLBP and PEA15 levels remained similarly elevated at each timepoint and across the two severities (Figure

3.3B-E). GFAP fluid levels were 5-7 fold elevated between 30min to 1-2 days post-injury

(p<0.013, Figure 3.3B). After mild stretching, GFAP levels at 30min rose only three-fold,

which was not significantly different from control levels. In contrast, fluid levels of ALDOC,

BLBP and PEA15 at 30min after mild stretching rose much higher and significantly over those of controls (60-460x; p<0.0001). Their release after mild stretching was not significantly different from severe stretching at any timepoint post-injury (2-3x). Yet, GFAP release levels were significantly higher 5h after severe versus those after mild stretching

(5x, p=0.042). The data suggest the release of cytosolic biomarkers ALDOC, BLBP and

PEA15 relates to both early cell wounding and later cell death, while the release of cytoskeletal GFAP, particularly its small proteolytic fragments, follows the slow accumulation of dead cells. We tested for an association between astroglial biomarker release and cell fates by plotting each biomarker’s levels against cell wounding and cell death rates from cultures of all conditions, and determining their Spearman correlations

(Figure 3.3F-I). ALDOC, BLBP, and PEA15 associated with cell wounding. ALDOC and

BLBP also correlated with cell death. GFAP had the strongest correlation with cell death, while it did not correlate with cell wounding. ALDOC and PEA15 release also correlated well with each other while all other marker pairs had moderate correlation (Table 3.5).

This is the first report of the early, robust release of ALDOC, BLBP and PEA15 from human astrocytes after mild and severe injury.

Characterization of glial fibrillary acidic protein break-down products

We also identified and measured new trauma-generated small GFAP proteolytic breakdown products (BDPs, 18, 20 and 25kD) that only appeared 1-2 days after injury

(S3.3). Intact GFAP and its BDPs were further characterized through immunoprecipitation from stretched astrocyte conditioned media and whole cell lysates. Peptide mapping of

immunoprecipitated products revealed a core sequence of amino acids common to all observed breakdown products (including our new lower MW species) starting at alanine residue 71 (S3.4). Generation of smaller BDPs appears to result from additional C- terminal cleavage. These N and C-terminal cleavages are believed to be the result of calpain and caspase activation following injury and are consistent with the reported cleavage sites reported in the literature (52). However, we were unable to identify BDP unique peptides by PRM-MS based on the reported non-tryptic N-terminal cleavage site.

This leaves some room for doubt for the exact sequence of these BDPs.

Trauma caused astroglial biomarker depletion and disassembly in wounded and dying cell populations.

Cell analyses using dye uptake and biomarker immunofluorescence further substantiated the correlation between biomarker release and cell fate. Viable GFAP- expressing control astrocytes displayed cytoskeletal filaments (Figure 3.4A1). Acutely after stretching, a population of GFAP-positive cells lost cytoskeletal fiber definition

(Figure 3.4A2). While the number of GFAP-positive cells did not decrease significantly after injury, stretching did significantly increase the fraction of astrocytes with non-fibrous

GFAP (Figure 3.4I). GFAP filament disruption was associated with cell integrity compromise (Figure 3.4E).

Control images show robust BLBP expression in GFAP-positive astrocytes. By

30min after stretching, bright BLBP signal decreased, while the remaining GFAP signal distribution were altered (S3.5). The population of cells with bright BLBP signals nearly disappeared 30min after stretching (Figure 3.4C2, G). ALDOC and PEA15 were

ubiquitously expressed in control astrocytes (Figures 3.4B, D). Membrane wounding, with occasional blebbing 30min after stretching, was associated with ALDOC and PEA15 signal decreases (Figures 3.4B2, and D2, arrowheads and arrows) and occurred in 29-

39% of leaky cells, and in 11-14% of cells with intact membranes (Figure 3.4F, H). Rapid post-injury resealing of mechanoporated cells could explain the depletion of markers from cells with intact membranes. Altogether, ~17% of astrocytes were depleted of PEA15 and

ALDOC acutely after stretching. Cell fluorescence measurements 30min after injury confirmed acute protein loss by demonstrating significant signal intensity reduction from control levels for ALDOC (by 34%), BLBP (by 29%), and PEA15 (by 43%) in subpopulations of stretched astrocytes (Figure 3.4J, p<0.001).

This human trauma model documents hyper-acute release of cytosolic markers

ALDOC, BLBP and PEA15 was associated with cytosolic protein loss in a subpopulation of wounded astrocytes, likely through plasmalemmal irregularities as was also suggested by previous time-lapse studies (32). In contrast, GFAP was temporarily retained, undergoing cytoskeletal filament loss and redistribution followed by delayed release with further fragmentation. The two trauma-inflicted release kinetics highlight different, cell- fate associated astroglial biomarker classes.

Clinical study documents distinct AID biomarker CSF and blood profiles in TBI patients across all severities.

We measured AID biomarkers, apolipoprotein B (APOB), a serum protein, prostaglandin synthase (PTGDS), a CSF standard, and known biomarkers GFAP and

S100β in CSF of 26 severe TBI patients and 13 control subjects (Table 3.6, Figure 3.5).

ALDOC, BLBP, PEA15 and small BDPs of GFAP presence in clinical blood samples was established in 22 severe and 15 mild TBI patients along with 12 control subjects (Figure

3.8). The average age was 42 among severe and mild TBI patients, and 39 among control subjects. Severe and mild TBI patients included 71% male, and 59% of control subjects were male. Severe TBI patients had on average a GCS score of 5.5, ranging between 3 to 11 and the survival rate was 85%. Most severe TBI patients (92%) had intracerebral hemorrhage with one or more neuroimaging findings of contusions, subdural hematoma, subarachnoid or intraventricular hemorrhage. Diffuse axonal injury was reported in 42%, epidural hematomas and edema were each found in 21%, ischemia in 8% and midline shift in 4% of the severe TBI patients. Injury causes among all TBI patients involved motor vehicles in 43%, falls in 41% and other causes including violence and football in 16% of the cases (Table 3.6). Multiple samples per patient from different post-injury times were analyzed from 8 TBI patients (6 severe and 3 mild TBI patients), and data were separated by day post-injury except post-injury days 4 and 5, which were averaged because individual immunoblot biomarker optical density (OD) measures did not significantly differ

(not shown). Variances in biomarker levels were larger between TBI patients than within subjects.

Boxplots of normalized immunoblot densities show significantly elevated TBI CSF levels for GFAP, ALDOC, BLBP, GS, PEA15 and S100B, which were 2-4 orders of magnitude greater than control levels, or controls had no measurable signal as for PEA15, small (18-25 kD) GFAP BDPs and ALDOC BDP (Figures 3.5 D-K, Table 3.7, S3.6 and not shown). Longitudinal trajectories differed between new and known astroglial biomarkers in TBI CSF, as GFAP and S100β signals decreased significantly after the first

day post-injury, with mean GFAP levels over 11-fold lower on later post-injury days

(Figure 3.5A, D). In contrast, ALDOC and GS means remained elevated across all days post-injury (Figure 3.5A, F, G). BLBP and PEA15 signals fluctuated over time and across patients without significant mean decreases across days post-injury (Figure 3.5A, H, I).

About 50% of all TBI CSF samples displayed a new trauma-generated 38kD ALDOC fragment, found predominantly on later post-injury days (Figure 3.5C). The temporal differences in enzymatic cleavage pattern between GFAP and ALDOC as well as the short half-lives of BLBP and PEA15 document highly variable proteolysis and clearance kinetics. Overall, AID biomarkers, including their BDPs, were robustly elevated after TBI and had an extended detection window when compared to GFAP and S100. Time after injury responses were also measured by MRM-MS which demonstrated similar trends compared to immunoblot densities for GFAP, ALDOC, BLBP, and GS (S3.7)

Analyses included also evaluation of fluid standards. The CSF standard PTGDS was secreted at high levels in healthy controls and was decreased over ten-fold in TBI patients; it was often undetectable on injury day and early post-injury days suggesting dysregulated CSF balance (53, 54) (Figure 3.5A, K). Serum protein apolipoprotein B

(APOB), absent in control CSF, was significantly elevated acutely after TBI with concentrations decreasing during subsequent post-injury days (Figure 3.5A, J). A reported ischemic episode on the third post-injury day in one severe TBI patient was accompanied by a secondary CSF peak of APOB as well as BLBP and PEA15 levels, the latter two documenting secondary astroglial damage (Figure 3.5A). These are the first quantitative APOB measurements used as intraventricular bleeding marker in human TBI,

although APOB has been documented previously in a rodent spinal cord injury CSF proteomic study (25).

Preliminary correlation of AID biomarkers with TBI patient survival were explored for three markers. Acutely depleted PTGDS levels were later restored to higher, near control, levels in TBI survivors but recovered less or not at all in TBI patients who had died (S3.8A). Mean PEA15 levels for surviving TBI patients were over one-thousand fold higher than those of non-surviving TBI patients (S3.8B). Small GFAP fragments (25 kD doublet, 20 kD, 18 kD BDP) were measured in TBI patients for the first time (Figure 3.5A,

B, S3.8C). Small GFAP BDP amounts differed between survivors and non-survivors by an order of magnitude more than total GFAP levels (S3.8C). Each TBI patient had different large and small GFAP fragment profiles, which is a first indication for individual degradation kinetics reflecting patient heterogeneity. Together with the unique association to cell death, small GFAP fragments add specificity to TBI severity assessment. While we were unable to measure GFAP BDPs by MRM-MS, intact GFAP,

ALDOC, BLBP, and GS concentrations were assessed in relation to TBI patient survival as well (S3.9). Similarly, elevated biomarker concentrations were observed in deceased patients compared to survivors. However, given the low number of mortalities compared to survivors, additional clinical studies are needed to determine outcome correlations of this panel.

AID biomarker concentrations were quantified using antibodies and quantitative mass spectrometry.

Our work represents the first use of multiple reaction monitoring (MRM)-MS to systematically quantify levels of neurotrauma biomarkers in TBI patients (Figure 3.6).

MRM-MS of GFAP, ALDOC and BLBP show higher levels in TBI than control CSF (Figure

3.6A). Biomarker concentrations were determined using the ratio of the amount of CSF- derived endogenous peptides to the known amount of isotope-labeled standard peptides

(Figure 3.6 B-D). GFAP immunoblot densities and MRM measurements showed high correlations between the two independent methods (Figure 3.6B). A comparison of immunoblot and MRM longitudinal CSF measurements shows matching profiles for astroglial biomarkers in one severe TBI survivor from as early as 3h to 5 days post-injury

(S3.10).

MRM-MS provides antibody-independent concentration comparisons among AID biomarkers in TBI patients’ CSF (Figures 3.6C, D). Highest levels were measured for

ALDOC, followed by GFAP (2.5-fold lower) on injury day, and both were significantly higher than levels for BLBP (29-70-fold lower) and GS (3 orders of magnitude less, Figure

3.6C, Table 3.7). By 3 days post-TBI, ALDOC concentrations were significantly higher

GFAP concentrations, differing by an order of magnitude, documenting prolonged

ALDOC stability in CSF over that of GFAP (Figure 3.6D). MRM and immunoblot pure protein measurements resulted in similar detection limits for ALDOC, BLBP and GFAP

(Table 3.8). Matching MRM and immunoblot interquartile concentration ranges show a wide dynamic range in AID biomarker levels after TBI (Tables 3.7, 3.8).

Multivariate Factor analysis of AID biomarkers documents TBI patient diversity.

Spearman correlation coefficients between biomarker pairs are listed (Table 3.9).

Strongest correlations were between cell death-associated, lower GFAP fragments, serum protein APOB (0.9), and S100 (0.87-0.85) that may reflect an association between intraventricular bleeding and astroglial demise. BLBP and PEA15 correlated robustly (0.8), suggesting similar CSF profiles. Markers with different temporal profiles and stability tended to correlate poorly like ALDOC and GS levels with those of GFAP, its small BDPs, and S100β (Figure 3.5A, Table 3.9). Astroglial biomarker levels related negatively with PTGDS indicating diverging profiles (53, 54). Proteolytic fragments for

ALDOC and GFAP did not co-vary, suggesting different proteolytic degradation patterns

(see Discussion). The correlations support the diversity of this panel and reflect differences in biomarker appearance and clearance after TBI.

We used an exploratory machine-learning Factor analysis for unsupervised grouping of AID biomarker profiles into ‘Factors’ based on Spearman correlations and known for revealing common underlying trends (55). The algorithm sorted AID biomarkers into two factors. These two factors and PTGDS accounted for 84% of the cohort’s biomarker variance. Factor A was comprised of GFAP, its lower BDPs, S100β and APOB.

‘Factor B’ contained ALDOC, its 38kD fragment, BLBP, GS and PEA15 (Figure 3.7A).

Thus, Factor A reflected markers of astroglial demise and bleeding corresponding with tissue loss, whereas Factor B represented markers of astroglial wounding, associating with tissue compromise. The resulting categories were reliable, as they had high communality expressed by factor correlation coefficients (Figure 3.7A, Cronbach’s ).

This TBI patient based, unbiased factor classification independently confirmed a grouping of astroglial biomarkers based on release from wounded or dying astrocytes in vitro

(Figures 3.2-3.4). Factors A and B separated TBI patients between controls and non- survivors, (Figure 3.7B). Exploratory classification tree analysis partitioned controls and

TBI patients using a Factor B boundary and a Factor A boundary separated TBI survivors from non-survivors with one outlier (Figure 3.7B, C) (56). Both factors were robustly elevated in TBI patients versus controls (Figure 3.7D, E). Factor A temporal CSF profiles decreased over post-injury days and differed between TBI survivors and non-survivors, while Factor B profiles were more stable over time and indifferent to survival status

(Figures 3.7D, E).

AID biomarkers are elevated in the blood after severe TBI and in a subgroup of mild TBI patients.

ALDOC, BLBP, PEA15 and GFAP BDPs were detected in blood samples depleted of abundant proteins. All four markers were robustly elevated in 50 plasma and serum samples of 22 severe TBI patients compared to 12 control subjects (Figure 3.8A-E, Table

3.6). Their concentrations reached up to 20 ng/ml (Table 3.8). Cell wounding markers

ALDOC, BLBP and PEA15 were significantly elevated over controls as early as 3h on injury day in blood of severe TBI patients (Figure 3.8A-E, S8). Blood ALDOC levels rose significantly between injury day (88-fold over controls) and subsequent two post-injury days (over 300-fold above controls, Figure 3.8A, C). Mean injury day blood levels for

BLBP and PEA15 were elevated over control levels (122-fold and 40-fold, respectively).

(Figure 3.8A, D, E). For the same TBI patient, BLBP and PEA15 were elevated in serum prior to their presence in CSF (S3.11). In contrast, a 25 kD GFAP fragment was consistently absent on injury day and appeared robustly on the first post-injury day in

blood of severe TBI patients, beginning as early as 22h post-TBI, GFAP was first elevated in CSF followed by overnight decrease and appearance in the circulation, with levels up to 4 orders of magnitude higher than those of controls in CSF (Figure 3.8A, B, S3.11).

ALDOC levels showed at 3 and 34h in CSF and in blood (S3.11A). These same patient observations illustrate that different fluid kinetics for these four astroglial biomarkers can exist.

The presence of AID biomarkers was explored in 15 mild TBI patients within the first injury day, a relevant time window for mild TBI diagnosis (57). Preliminary data show elevation of ALDOC, BLBP and PEA15 in serum as early as one hour after mild TBI compared to control serum and at similar levels as found after severe TBI (Figure 3.8F,

S3.12). ALDOC was present in 80%, PEA15 in 60% and BLBP in 47% of this cohort of mild TBI patients irrespective of CT status. In contrast, GFAP BDPs (37, 25 or 20 kD) were only found in 27% of the samples (S3.12). In some mild TBI patients, sera signals appeared in the same range as those in severe TBI sera, consistent with similarly observed in vitro release profiles from wounded human astrocytes after mild and severe stretching.

2.3 DISCUSSION

A new panel of astroglial, injury-defined (AID) biomarkers from the TBI CSF proteome and a list of trauma-released, astrocyte-enriched proteins is presented. AID biomarkers are qualified by release from wounded and dying human traumatized astrocytes. Clinical confirmation in TBI patients shows robust AID biomarker elevation in

CSF and blood of TBI patients with broad post-injury profiles and provides a new concept for a biosignature of brain cell compromise and demise.

There are unmet requirements for assessing TBI patients.

Presently, the initial assessment of TBI patients relies on the GCS scores and on

CT scans, which both correlate poorly with outcome and functional compromise after TBI

(58, 59). Mild TBI victims are assessed using behavioral testing and cognitive questionnaires (60). These tests rely on subjective self-reporting and require baseline assessment. An unbiased brain injury signature is needed that can assess TBI and identify complicated injuries among mild TBI patients (5, 57). On-site diagnosis of TBI patients, particularly of athletes, military personnel and urgent care situations, would provide early information for advising on initial treatment decisions or transportation choices. Instant release with the primary trauma event and concomitant presence in the circulation are prerequisites for future real-time neurotrauma biomarkers. Based on this rationale, connecting acute cellular injury processes and biomarker presence can facilitate this goal. Few animal studies address primary cellular injury events or early biomarker release (61-64). After mild TBI, GFAP’s passage into the circulation is delayed, making its use as urgent care tool less helpful (65). In contrast, UCHL1 declines on injury

day limiting interpretation of samples collected at various times after injury (65, 66). Thus, limitations exist among current neurotrauma biomarker candidates (67, 68). This study provides biomarkers of brain cell wounding that address current limitations, as they were released instantly and robustly after severe and mild TBI.

Overcoming proteomic bottlenecks of biomarker identification and a multiplex standardized TBI assay.

Proteomics provides a comprehensive view of protein changes after neurotrauma, yet clinically useful neurotrauma biomarkers remain elusive (29, 31). A major hurdle has been the selection of clinically relevant biomarkers from extensive lists of identified proteins, which our controlled human trauma model and selection strategy have cleared

(31, 69) Typically, one or two biomarkers are investigated, resulting in diverse profiles and sensitivities due to non-standardized assays (23, 70). Further, efficient throughput quantifying multiple biomarker candidates requires a standardized assay to enable biomarker comparisons (71, 72) (23, 70). MRM-MS is favored as a multiplex, antibody- independent assay for standardized simultaneous measurement of multiple candidate biomarkers, but until the present study, had not yet been applied systematically in the neurotrauma field (57, 70, 73).

Astrocyte trauma responses are heterogeneous.

We previously documented molecular heterogeneity among human and mouse astrocytes based on variable expression of astroglial markers and their different trauma responses (32, 48). Morphological signs of astrocyte wounding were found early and

scattered in stretch-injured cultures, which could indicate both, selective astrocyte vulnerability and ‘hot spots’ of focal tensile forces. Distinct subpopulations of wounded human astrocytes underwent depletion of cytosolic proteins and GFAP filament disassembly, while adjacent cells seemed unchanged, illustrating the diffuse injury distribution in this trauma model. Rapid post-injury GFAP filament disassembly and brighter immunofluorescence associated with cell membrane wounding were found prior to reported gene expression changes (32, 51). Such GFAP changes are similar to reported alteration in GFAP antigenicity after acid treatment, mediated by rapidly elevated calcium, and calpain activation, but are new in mechanically wounded astrocytes (74, 75).

The depletion of key metabolic proteins, together with cytoskeletal filament disassembly in mechanoporated astrocytes, likely exacerbates their compromise, making a subpopulation of traumatized astrocytes vulnerable to a second mechanical blow or other stressors that can lead to cell death.

Mechanical trauma-induced reactivity with characteristic shape changes and upregulation of astroglial markers occurred within hours and evolved over days post-injury in human astrocytes (48). Trauma-induced activation of signal transducer and activator of transcription 3 increases oxidative metabolism and upregulates expression of GFAP,

ALDOC, BLBP and PEA15 during reactive gliosis, boosting astroglial resilience and could amplify delayed release in case of a secondary insult, as was observed in one TBI patient

(32, 39, 76-78). Thus, acute membrane wounding, neuroprotective astrogliosis as well as delayed astroglial demise document cell fate heterogeneity that is reflected in diverse astroglial markers, their release, and expression after neurotrauma.

AID biomarkers monitor acute trauma pathology of astroglial membrane wounding and fiber damage.

Plasmalemmal permeability (mechanoporation) is an early and enduring pathology in acutely traumatized brain and spinal cord tissues, and mechanoporation is a hallmark of diffuse axonal injury, which is characterized by process beading and fragmentation (13,

14, 79, 80). Diffuse axonal damage is also a morbidity in mild TBI patients with post- concussive symptoms (81, 82). How long mechanoporated cells endure in a compromised state, and whether they recover or undergo protracted cell death are open questions (83). The present study documents profiles of astroglial biomarker release after pressure-pulse stretching that correlated with cellular features of human astroglial wounding, mechanoporation, and delayed cell death. We show human astroglial fiber damage, including beading and process disintegration, shortly after mechanical trauma.

Astroglial fiber damage is also seen in vivo early after mouse spinal cord crush injury, in the traumatically injured primate cortex, reported as clasmatodendrosis, and in human cerebral cortex after traumatic intracranial injury where it is associated with protein degradation markers (16, 32, 84). This histopathology is particularly relevant for human white matter injury, because astrocytes outnumber neurons in the human neocortical white matter, and human astrocytes carry over-sized processes (38, 41, 85, 86). Thus, astroglial wounding-released biomarkers may provide biofluid-accessible tools to investigate diffuse glial fiber damage acutely after neurotrauma.

AID biomarkers may have possible roles as biosignatures and for manifesting metabolic depression.

Astrocytes maintain a high rate of active oxidative glucose metabolism, express overall high levels of glycolytic and tricarboxyl acid (TCA) cycle enzymes and carry a large number of mitochondria in their perisynaptic processes (37, 42). Astroglial GS and BLBP have important roles for synaptic plasticity at the tripartite synapse complex in glutamate recycling and regulation of fatty acid uptake (87, 88). ALDOC is a central glycolytic enzyme providing the substrate for lactate and ATP production and its product glyceraldehyde-3-phosphate controls cell fate and astrocyte-neuron crosstalk (36).

PEA15 is a main regulator of glucose metabolism, and high PEA15 levels make cells resistant to glucose deprivation by adapting to different metabolic states (89, 90).

Astrocytes couple synaptic metabolic demand as they adjust local blood flow by ensheathing both compartments with their endfeet (88, 91). Vital astroglial metabolism is essential for maintaining neuro-metabolic coupling and brain energy homeostasis.

The majority of TBI patients undergo metabolic depression indicated by reduced cerebral oxidative metabolism and associated with an imbalance of lactate, pyruvate, glutamate and glucose (11, 92, 93). Decreased cerebral glucose metabolism is measured in mild and severe TBI patients using positron emission tomography scanning (11, 58).

Reduced oxidative glucose metabolism is demonstrated after lateral fluid percussion, a rat cerebral concussion model with astrocyte metabolism being initially reduced and sub- acutely supportive for restoring metabolic homeostasis (94-96). One reason for impaired energetic needs of injured cells may be selective depletion of several glycolytic and TCA cycle enzymes as well as GS, PEA15 and BLBP as they are acutely released in vitro.

ALDOC and PEA15 are reduced in vivo and in perilesional wounded astrocytes acutely after mouse crush spinal cord injury (32). The present study links fluid release of BLBP,

ALDOC and PEA15 with concomitant cellular protein loss in traumatized human astrocytes, supports a concept of cellular compromise and identifies biofluid AID markers signals as possible candidate indicators of metabolic depression after TBI.

GFAP release and degradation associate with astrocyte death after severe TBI

It has been assumed that biomarkers are released due to cell death after neurotrauma (23). Severe TBI with lesions and contusions are associated with tissue demise, vascular damage, and perilesional, irreversible astrocyte swelling leading to cytotoxic edema (97-99). Astrocyte demise is reported after lateral fluid percussion in the rat cortex and in the peri-lesional mouse spinal cord one day after contusion (100, 101).

Human post-mortem cerebral and hippocampal cell counts document progressive astroglial death by means of different mortality times after TBI, that relate to injury severity

(102). Our human trauma model shows delayed and severity-dependent astroglial cell death preceded by GFAP release and associated with the appearance of small GFAP

BDPs. Similar findings, albeit only considering total GFAP signal, were obtained by stretching rat hippocampal slice cultures (103). Thus, trauma-inflicted astroglial cell death can be monitored using GFAP, particularly by its small, more selectively cell death- associated fragments. GFAP degradation is related to caspase activation in models of

Alzheimer’s and Alexander disease (104-106). The present study is the first to associate small GFAP fragment generation with cell death in traumatized human astrocytes. Our explorative clinical data indicates small GFAP BDP levels differ substantially between survivors and non-survivors compared to those of total GFAP amounts. Thus, small

GFAP fragments in biofluids may help to specifically monitor astroglial demise after TBI.

Kinetic diversity among astroglial neurotrauma biomarkers

Biomarker profiles fluctuate given injury complexity and irregular secondary events in severe TBI patients (107). Astroglial biomarkers also displayed marker-specific kinetics. ALDOC had remarkable CSF stability over time after TBI and is reported to last up to three weeks in sheep and cow blood (108, 109). A 38kD ALDOC BDP of later post- injury days could be a product of proteolysis by calpain or cathepsin (110-112). In contrast, GFAP displayed massive degradation into large fragments that had been previously detected after TBI, amyotrophic lateral sclerosis and oxidative frontotemporal lobe degeneration (113-115), and also into small fragments associated with caspase activation, which have not been measured after TBI (104-106). We quantified large and small GFAP fragments in TBI patients in this study, and observed that overall GFAP CSF levels decreased drastically from the second post-injury day onwards, concurrent with fragments appearing transiently in blood. Delayed biomarker passage into the circulation may occur via glial-mediated overnight CSF clearance (116). In contrast, BLBP and

PEA15 signals were more variable, which could reflect short biofluid stability, (117).

Overall, these observations reveal a new diversity in biofluid kinetics among different astroglial biomarkers.

Hyper-acute presence of AID markers indicates astroglial release immediate to the traumatic impact and documents direct passage into the circulation. This may be possible, since perivascular astrocyte sheets cover endothelial tubes nearly completely in brain microvessels and are part of the blood-brain diffusion barrier (118, 119). Astrocytes function as gatekeepers in the neuro-vascular unit and their damage results in BBB

permeability (33, 120). TBI causes microvasculature disruption and astroglial fiber damage (16, 80, 121, 122). ALDOC is present in astroglial process endings and could be directly released with their rupture (32). Animal and clinical studies show BBB disruption in the early hours post-injury after mild TBI, blast shock waves, and mild fluid percussion injury (123-125). Blood elevation of cytosolic astroglial marker, S100, indicates BBB permeability and is elevated after mild TBI including repeated sub-concussive events

(126-128). Hence, cytosolic astroglial proteins are situated for immediate release into the circulation upon BBB disruption after TBI.

AID biomarker panel limitation, uniqueness and significance for future TBI patient assessment and monitoring

Confounding variables of age and gender were matched in this study, while medications and comorbidities were not controlled for. Methodological rigor is provided by using two technically independent assays to validate biomarker measurements. Data analysis was separated by day, yet considering the short-lived nature of some biomarkers, future finer resolved kinetic studies are advised.

Selection of candidate biomarkers was achieved using astrocyte enrichment and trauma-release, not warranting brain exclusiveness. All markers are highly enriched in the CNS; ALDOC is one of the most abundant brain proteins and is highly brain enriched

(46, 129). BLBP, GS and PEA15 are also highest expressed in the CNS, with selective presence in other tissues (S1). To our knowledge, the combined biofluid elevation of any two or more astrocyte-enriched biomarkers presented here points exclusively and sensitively to brain injury.

100

Biomarker panels are anticipated to improve TBI patient assessment compared to a single biomarker (130). Several neurotrauma biomarkers have been previously combined to evaluate patients with brain or spinal cord injuries (131, 132). Unsupervised factor analysis has been used for psychological and cognitive self-rated scores to assess

TBI patients (133, 134). To our knowledge, this is the first study applying Factor analysis to a small neurotrauma biomarker panel. AID biomarker factors derived from TBI patients coincided with cell fate assignment, thereby clinically validating the trauma model’s classification. Aldolases (ALDOC) and fatty acid binding proteins (BLBP/FABP7) have been previously considered as biomarkers of brain injury and cerebrovascular disease, but isoforms were not always distinguished (117, 135). Linking cell fate and biomarkers is unique to this study and delivers novel fluid biosignatures for traumatized brain tissue.

Correlating biomarkers and cell fates and documenting differences in individual biomarker’s kinetic are new observations with significance for future neurotrauma biomarker studies. Overall, our translational and exploratory clinical studies document elevation of AID biomarkers in CSF and blood of severe and in serum of mild TBI patients, supporting their further study for TBI patient assessment.

101

3.4 METHODS

Donors, patients and samples

All CSF and plasma samples were collected prospectively under protocols approved by the local ethics committee of all the sites involved and stored at UCLA.

Written informed consent was obtained from patients or legal authorized representatives before enrollment. The CSF samples from TBI subjects were collected directly from ventriculostomy catheters, every 6h up to a maximum of 5d following injury. Blood samples were collected by venipuncture. CSF and blood samples were aliquoted, and stored at−80°C until the time of analysis.

Adult patients with severe head injury and requiring a ventricular catheter for intracranial pressure monitoring were included in the discovery set. Inclusion criteria were a Glasgow Coma Scale (GCS) score of eight or less post-resuscitation or on presentation.

Exclusion criteria were no informed consent, patients younger than 18 years of age, female patients that were or may have been pregnant, known history of neurological disease, and Injury Severity Score greater than 15. Treatment of patients, according to international guidelines, was targeted at a normal ICP and maintaining cerebral perfusion pressure.

The CSF control samples were obtained by lumbar drain from patients with an unruptured aneurysm or was donated from healthy subjects (Precision Med). This study group included also adult patients presenting to the hospital emergency department within

4 hours after sustaining blunt trauma to the head resulting in mild TBI (GCS 13-15) or moderate TBI (GCS 9-12). Individuals under the age of 18 years, pregnant women,

102

prisoners, subjects who did not require a CT scan as part of their clinical evaluation or with previous history of psychotic illness or neurological disease were excluded.

CSF Proteomics

CSF volumes from TBI patients and healthy subjects corresponding to 50-300µg of total protein content measured by BCA assay was dried down by vacuum centrifugation and reconstituted in 100 µL of 50 mM ammonium bicarbonate pH 8.3 solution (Sigma-

Aldrich),0.1% deoxycholatic acid. Cysteine disulfides were reduced by addition of tris(2- carboxyethyl)-phosphine (10 mM, Thermo Scientific) and incubated at 50°C for 1h then adjusted to room temperature. Free cysteines were alkylated with iodoacetamide (20mM,

Sigma-Aldrich). for 30 min at 37°C in the dark. Trypsin (500ng/µL 50 mM ammonium bicarbonate, sequencing grade, Promega) was added to CSF samples at a 1:25 enzyme to protein ratio and digested for 16-18h overnight at 37°C. Samples were acidified with

5% formic acid (v/v) and centrifuged at 13K rpm to pellet deoxycholic acid precipitate. The supernatant was then transferred to a separate microcentrifuge tube and dried by vacuum centrifugation.

CSF tryptic digests were reconstituted in 100µL of 0.1% formic acid, 3% acetonitrile for LC-MS/MS. Samples were desalted using a C18 trap column connected to C18 PepMap reversed phase HPLC column for peptide separation. CSF samples were analyzed using a LTQ-Orbitrap or Q-Exactive Orbitrap mass spectrometer (Thermo).

Peptide separation was done in a 60 or 120min gradient from 5-35% of mobile phase

(100% acetonitrile, 0.1% formic acid). Analysis on the Q-Exactive was performed in the positive ion mode with settings: resolution –70,000; m/z range–300-2000; maximum MS1

103

injection time–50ms; MS automatic gain control (AGC) target–1x106. Acquisition was set to record up to 10 confirmatory product ion spectra (MS2) per full scan spectrum by selecting precursor ions of decreasing signal intensity with 30sec dynamic and charge state exclusions to exclude signals with unassigned charge, charge +1, and charges >+5.

MS2 instrument settings were: resolution – 35,000; maximum MS2 injection time –

100ms; MS2 AGC target – 2x105; fixed first mass m/z–100.

The data was searched using Mascot (Matrix Science) against the human subset of the SwissProt database. Oxidation of methionine was set as a variable modification with carbamidomethylation of cysteine was set as a fixed modification. Enzyme specificity was set to C-terminal cleavage at arginine and lysine with up to 2 mixed cleavages allowed. Strict m/z error tolerances were set to 15 ppm in MS mode and 0.01Da in MS2 mode. Peptide spectral matches were validated against a decoy database using the percolator algorithm at a 5% false discovery rate.

Human astroglial injury model, cell permeability and viability assay, immunocytochemistry

Primary human astrocytes were prepared from donated, de-identified human fetal cerebral neocortex at 16-19 gestational weeks as described (48). Briefly, in calcium and magnesium-free Hank’s buffered saline solution (HBSS) mechanically dissociated tissue was filtered through 70 µm and 10 µm nylon meshes (Nitex) into culture medium (DMEM-

F12) with 10% fetal bovine serum (FBS, Atlanta Biol.). Neural progenitor cells were removed by 30mincentrifugation at 30,000 X g in a HBSS-buffered 33% Percoll gradient

(Sigma). The top fraction was washed and diluted in DMEM/F12, 10% FBS and astrocytes cultured in T150 cell culture-treated plastic flasks (Corning). Confluent cultures

104

were shaken for 4 days at 200rpm on a shaker in an incubator. Astrocytes were treated in 0.25% trypsin/EDTA followed by gently mechanical dissociation, and washed cells were seeded onto collagen I-coated silastic membrane culture plates (6 well Bioflex,

Flexcell Intl.) at a density of ~ 135,000 human cells / 962mm2. Upon confluence, medium was replaced by DMEM/F12 with 10% heat-inactivated horse serum (Atlanta Biol.) that was then stepwise reduced. Serum-free astrocytes in 2ml DMEM/F12 were stretch- injured using one mild (2.6-4.0psi) or severe (4.4-5.3psi) 50ms nitrogen pressure pulse with the CIC II pressure controller (Custom Design and Fabrication Inc.). Cell death rates significantly differed between mild and severe pulses, providing 2 outcome defined distinct severities in this human trauma model (see Results, Figure 3.2).

Cells were incubated with 0.025 µg/ml propidium iodide (PI) in Leibowitz’ L15

(Gibco) for 10min at 37°C followed by four rinses in L15. Dye was crosslinked to DNA of leaky cells by 5min exposure to UV light. Cells were fixed in freshly depolymerized 4% paraformaldehyde in Tris-buffered saline for 30min at 4°C. Rinsed cells were permeabilized with 0.3% Triton in buffer and blocked in 5% normal donkey serum in buffer, followed by overnight, incubation with primary antibody at 4°C diluted in blocking solution (Table 3.1). After rinsing, secondary antibodies (Table 3.2) were applied in blocking solution for 1h at room temperature. Cultures were rinsed and stained in bisbenzimide nuclear dye (Hoechst, 1:75 in distilled water) for 5min, rinsed, dried and coverslipped (Fluorogel, Biomedia). Hardened cultures were mounted on slides (32, 48).

Immunoblotting, sub-saturated densitometry and technical variance

105

Culture trauma fluids, conditioned medium, was collected from 6 cultures (12mL) and samples prepared as described (32). Briefly, protease inhibitors (Roche) and dithiothreitol (5mM, Calbiochem) were added and fluids concentrated by ultrafiltration to one twentieth of the original volume (Vivaspin, VWR). Clinical biofluid samples were treated following common data element recommendations (136). CSF and blood samples were thawed and supplied with EDTA (pH7.4 to 1mM) and protease inhibitors bestatin

(40µM), pepstatin A (10µM) and phosphoramidon (10µM). Samples were centrifuged for

10min at 16,060g at 4ºC to remove lipids. Plasma and serum samples were depleted of albumin and immunoglobulins (IgGs) using immunoaffinity columns (ProteoPrep,

PROTIA Sigma) and concentrated by ultrafiltration (Vivaspin, VWR). Depletion removed

~ 85% of original protein concentration, including ~10-15% non-albumin or non-IgG protein.

Samples were denatured (5min at 100ºC followed by ice), biofluids reduced, by adding 1% β mercaptoethanol and all samples adjusted to Laemmli Sample Buffer

(2%SDS, 125mM TrisHCL pH6.8, 10% glycerol, 0.6% bromphenol blue) followed by immunoblotting as previously described (32). Fluid analyses were normalized by volume

(30μL/lane). Proteins were separated in 200mM glycine 25mM Tris 0.1% SDS for 30min at 100V followed by 1h at 120V in and transferred onto nitrocellulose (Hybond-ECL,

Amersham) with 20% methanol in same buffer. Proteins were reversibly stained using

0.1% Ponceau S in 5% acetic acid. Protein concentrations were determined by assay

(Pierce 660) against bovine serum albumin dilutions. Proteins > 30kD and 10-25kD were separated on 10% and 15% polyacrylamide, 2% SDS Tris-based gels. Respectively. A molecular weight standard (Precision Plus Kaleidoscope, Bio-Rad) and His-tagged pure

106

proteins ALDOC, BLBP, PEA15 (EnCor Biotech. Inc) and GFAP full size, 37 and 20kD fragments (Abbott Diagnostics) at various concentrations diluted in 0.5% bovine serum albumin were analyzed in parallel.

Blots were blocked for 30min with 10% non-fat milk in Tris-buffered saline with

0.05% Tween-20 (TBST) before overnight incubation at 4ºC with primary antibodies diluted in 5% BSA in TBST (Table 3.1) (32). Isoform specific antibodies were used for

ALDOC and BLBP, because organs outside the CNS express other isoforms of these proteins that are released after injury (137, 138). Washed blots were incubated for 1hr at room temperature with peroxidase-conjugated secondary antibody (Thermo, Table 3.2).

Washed blots were incubated for 5min in enhanced chemiluminescence substrate (West

Pico ECL, Thermo Scientific). Film (Denville) captured signal, using same sequence of exposure lengths consistently.

Signal levels were measured from scanned films using a bio-imaging and analysis system with background correction (Autochemie Systems, UVP). Post-hoc normalization of sub-saturated signals across multiple exposures covered 2-4 orders of magnitude. The relative standard deviation for scaled exposure readings ranged from 11-33%. Wet experimental replicates produced signals that varied from the sample mean by 20±14%.

Overall, analysis variance fell one order of magnitude below significant cross-condition differences. In the trauma model, variation in release of each biomarker due to base astroglial expression varied no more than 1-2 z-scores (not shown).

We controlled for the combined use of serum and plasma samples by comparing serum and plasma marker signals in the same subject, which resulted in similar results for all blood-compatible biomarkers, except for GFAP, which showed additional non-

107

specific bands in plasma compared to serum samples (not shown). Hence, only the new and specific 25 kDa GFAP BDP was quantified in blood.

Quantitation of Biomarkers in CSF using multiple reaction monitoring mass-spectrometry

Peptides specific to each biomarker were designed and same, synthetic standard peptides with heavy isotope labeled arginine (6C144N15) and lysine (6C132N15) were purchased (Thermo Scientific). Peptide standards were prepared in 5% acetonitrile

(5pmol/µL) and spiked into CSF samples to concentrations between 25-75pmol per mL of CSF. CSF samples are then reduced, alkylated, and digested as described above.

Digested CSF peptides were dried by vacuum centrifugation and reconstituted in

0.1% formic acid, 3% acetonitrile in water. Samples were desalted using an on-line C18 trap column prior to LC-MS/MS analysis. Peptides were separated on a 5%-35% gradient of mobile phase B (0.1% formic acid in acetonitrile) over 40 min on a C18 PepMap reversed phase HPLC column. Samples were analyzed using either a Q-Exactive

Orbitrap MS or a 4000 QTRAP triple quadrupole MS (AB Sciex). MRM-MS analysis was performed with the Q-Exactive (by parallel reaction monitoring) targeting an inclusion list of precursor peptide ions (Table 3.3) for MS2 analysis with the following parameters: resolution 17500, AGC target 2x105, maximum ion injection time 50ms, isolation window

3.0Da, fixed first mass 100, normalized collision energy 27.

Biomarker specific precursor peptide ions are listed in Table 3.3. These precursor ions were fragmented by higher energy collisional dissociation or collision activated dissociation depending on the MS instrument, into their component product ions.

Biomarker abundance was calculated based on the area under the curve (AUC) of the

108

precursor to product ion transitions for each biomarker specific peptide using Pinpoint

(Thermo) and Skyline (MacCoss Lab). 3 transitions were summed per biomarker specific peptide and the ratio of the endogenous peptide signal to its heavy labeled counterpart was determined. Biomarker concentrations were calculated based on each peptide’s endogenous to heavy standard signal ratio, heavy standard concentration, protein molecular weight (MW), and a dimensional conversion factor according to the formula:

푒푛푑표푔푒푛표푢푠 25푓푚표푙 Endogenous protein concentration (ng/mL) = 푟푎푡푖표 × × 푝푟표푡푒푖푛 푀푊 × 푠푡푎푛푑푎푟푑 µ퐿

1 . 1000

Statistical Analyses

Optical density (OD) measurements were log-transformed for normal distribution, standardized and replicates were averaged. Multiple replicates of each MRM sample were measured (analytical replicates). Same day patient samples were independently prepared 3 times assuring experimental MRM-MS and immunoblotting consistencies

(experimental replicates). Same day patient replicates with different draw times were analyzed in parallel and averaged for graphs and statistical analyses and are shown.

Signal specificity was validated using multiple specific antibodies for ALDOC, GFAP and

BLBP (Table 3.1).

Cell fate (death, wounding) mean differences between early and late post-injury and between different pressure pulse severities were determined using mixed model analysis of variance (ANOVA) allowing for non-constant variance and random donors effects. Percent leaky/wounded and dead cells were normally distributed but had dissimilar standard deviations. Donor-paired log-transformed biomarker culture fluid 109

levels were compared over time and at two different stretch severities using repeated measures ANOVA, mixed model with homogeneous variances over time (139). The associations between biomarker levels and astroglial fate were quantified using the

Spearman correlation coefficient. (rs and p, Sigmaplot). Single cell immunofluorescence densities were compared between stretched and control astrocytes using Student’s or

Welch’s t-tests or Mann-Whitney U test depending on data distribution. TBI patient CSF and blood biomarker densitometry levels were compared to controls and over time by

ANOVA with independence of each timepoint (for CSF) and repeated measures ANOVA, mixed model with non-constant intra-class variances over time.

Spearman correlation of ranks was determined between biomarker immunoblot densities and MRM-MS biomarker specific peptide measures. A quantile-quantile plot was generated assessing the strength of the Spearman correlation. MRM-MS concentration differences across biomarkers were compared using repeated measures

ANOVA mixed model with non-constant variance.

A multivariate factor analysis based on Spearman correlations, was conducted for

CSF samples across all injury days. Signals for GFAP (50-37kDa) and total GFAP

(Spearman correlation 0.988) differed only slightly, as did those of 40 kDa ALDOC and total ALDOC (0.983), so one entry ‘ALDOC’ and ‘GFAP’ was used (Table 3.9). GFAP small BDPs (25-18 kD) and ALDOC 38 kD BDP signals varied from their main bands and were treated as additional biomarkers. Factor extraction was made for maximal differences between factors (varimax criterion). Factor values were computed by adding each marker after multiplying each biomarkers’ signal with its weight (loading). Biomarker

110

weight cutoff for loading was 0.51. Factor values were only computed for CSF samples with available readings for all factor components (biomarkers).

Classification tree analysis partitioned the CSF sample cohort by determining factor thresholds using Factors A and B amounts calculated of TBI survivors, non- survivors and Controls (56). Statistical analyses were conducted using Sigmaplot, Excel,

Instat (Graphpad), JMP and SAS version 9.4.

111

2.5 FIGURES

Figure 3.1: Venn diagram shows astrocyte proteomic signature of neurotrauma in cerebrospinal fluid

Venn diagram documents LC-MS/MS proteomes of CSF from 19 severe TBI patients (484 proteins) and 9 healthy subjects (402 proteins, Crl, Table 3.4). A published astrocyte trauma-release proteome of 59 proteins showed 38 proteins overlapped (64%, purple outline) with the clinical CSF proteomes (32). A subset of 14 proteins of the CSF and trauma model proteomes were 2-fold astrocyte enriched (black outline, (42). Proteins also present in healthy plasma or abundant in tissues outside the CNS were excluded. Thus,

Aldolase C (ALDOC) was identified among 5 candidates present in TBI and control CSF.

Among 4 proteins exclusive in TBI CSF was glutamine synthetase (GS). Among the additional 5 trauma-released, astrocyte enriched proteins were astrocytic phosphoprotein

15 (PEA15) and brain-lipid binding protein (BLBP), which were considered despite absence in CSF proteomes, due to limited LC-MS/MS sensitivity.

112

Figure 3.2: Mechanical trauma causes acute membrane wounding, reactivity and delayed cell death in human astrocytes

(A) GFAP (green) is weakly expressed in uninjured, differentiated neocortical astrocytes.

(B) Reactive astrocytes 1d post-injury were star-shaped with enlarged processes and

113

upregulated GFAP. (C) Acutely wounded, mechanoporated astrocytes 30min after stretching had taken up PI (red), had bright GFAP signals and beaded (arrows), disintegrated or amputated (asterisks) processes. (D) Nuclear morphologies differentiate

(D1) viable astrocytes with large, oval-shaped pale Höchst-positive nuclei (blue); (D2) membrane-wounded astrocytes show large pale Höchst stained, non-pyknotic nuclei with

PI-positive nucleoli (pink); and (D3) dead astrocytes with condensed chromatin, bright

Höchst and PI-positive small, pyknotic nuclei (pink). E) After mild (2.6 – 4.0 PSI, small red dot) and severe (4.4 - 5.3 PSI, large red dot p<0.001) pressure-pulse stretching, median fraction of leaky cells was elevated at 30min (p<0.0001) and 2d post-injury (P<0.01, mild) and was decreased between 30min and 2d post-injury (triangles, P< 0.01). (F) Cell death fractions were elevated at 30min (p<0.05, mild; p<0.01 severe; *) and 2d post-injury

(p<0.01, mild; p<0.001 severe; *) in stretched cultures. Severe stretched cultures had higher cell death rates than mild ones at 2d post-injury (black dot, p<0.001). The increase between early and late cell death rates was significant for both severities (triangle, p<0.0001).

114

115

Figure 3.3: Human astroglial biomarker release is defined by membrane wounding and cell death after mechanical trauma

(A) Immunoblots for GFAP, ALDOC, BLBP and PEA15 from conditioned medium (fluid) samples of unstretched (control, Crl), mild (2.6-4 psi, small red dot) and severe (4.4-5.3 psi, large red dot) stretched astrocyte cultures at 30min (30’) and 2d post- injury. Small

GFAP BDPs (25-18 kD) are absent at 30min and appear 2d post- injury whereas ALDOC,

BLBP and PEA15 fluid signals are present 30min post-injury and at 2d. Ponceau S shows protein amounts of same volumes per lane. B-E: Fluid sample geometric means of optical densities (OD) for total GFAP (B), ALDOC (C), BLBP (D) and PEA15 (E) of unstretched and at 30min, 5h, 1d and 2d after mild and severe stretching. Asterisks indicate significant differences between stretch-injured and control (GFAP 5h mild stretch: p = 0.005, all others p < 0.001; number of donors on x-axis). GFAP levels increased between early and later time-points (triangles, among mild stretched between 30min and 1d: p=0.018, between 30min and 2d p = 0.012; among severe stretched between 30min and 1d: p =

0.013, between 30min and 2d: p = 0.01). GFAP release differed at 5h between mild and severe stretching (black dot, p=0.042). F-J: Biplots correlate biomarker levels for GFAP

(F), ALDOC (G) BLBP (H) and PEA15 (I) on the y-axes with percent membrane wounded astrocytes (red) and percent cell death (black) on x-axes. Spearman correlations (rs) are given with p-values and best fit lines for significant ones.

116

117

Figure 3.4: Acute cell wounding is associated with depletion of astroglial markers and GFAP filament disruption

Human astrocytes show GFAP (white, A), ALDOC (B), BLBP (C) and PEA15 (D) signals

(green) in unstretched (control) and 30min post-stretched cultures. (A1) PI-negative, intact astrocytes display filament assembled, fibrous GFAP. (A2) Membrane-wounded,

PI-positive (pink) astrocytes had homogeneous, non-fibrous GFAP. (B1) Intact astrocytes express ALDOC. (B2) Leaky astrocytes (PI-positive, red) show plasmalemma blebbling

(arrowheads) and dim ALDOC. (C1) Group of intact astrocytes with bright BLBP expression. (C2) Membrane-wounded astrocytes had dim BLBP signal. (D1) Intact astrocytes expressed PEA15 homogeneously. (D2) PEA15 was depleted from PI-positive astrocytes (red nuclei, arrows). (E) Proportions of fibrous (striped) and non-fibrous (gray)

GFAP in intact (blue) and leaky (pink) astrocytes with percentage of each population listed. Stretching increased non-fibrous GFAP signals in intact (p=0.006) and leaky populations (p<0.001, n=5). (F) Intact astrocytes had strong ALDOC expression (bright, green). Stretching increased the fraction of ALDOC depleted cells (dim), in intact (p=0.02) and leaky (p=0.03) astrocytes, and signal loss was greater in leaky than intact stretched cells (p<0.001, n=6). (G) Brightly BLBP-stained GFAP-positive intact astrocyte population decreased 30min after stretching (p=0.007) and almost disappeared from leaky cells

(n=5). (H) The majority of intact astrocytes were PEA15-positive and their percentage diminished 30min after stretching in intact and leaky astrocytes (p<0.01), with greater signal loss in leaky than intact stretched cells (p<0.0001, n=6). (I) Percent GFAP expressing astrocytes with fibrous and non-fibrous GFAP cell populations. The shift from fibrous to non-fibrous GFAP fraction changed acutely post-injury (p<0.01, asterisk),

118

without significant reduction in GFAP-expressing cells. (J) Percentages of bright and dim cells differed between control and 30min post-stretch cultures for ALDOC (p=0.001),

BLBP (p=0.007) and PEA15 (p=0.003). (K) Stretching reduced cellular ALDOC, BLBP, and PEA15 fluorescence intensities 30 min post-injury (p<0.001, n=3-5).

119

Figure 3.5: CSF profiles of marker panel in TBI patients on injury day and consecutive 5 days are diverse

120

(A) Immunoblots of GFAP (50kD with BDPs 37, 25, 20 and 18kD), S100β (10kD), ALDOC

(40kD), GS (45kD), BLBP (15kD) and PEA15 (15kD) of 30 µl CSF samples from injury day (i) and subsequent 5 post-injury days (i+1 to i+5) of a severe 54 year old male TBI patient (1a.-1f.) alongside 30µl control CSF of a 24 year old male (I.). Bleeding indicator

APOB (130 and 250kD) had variable intensity over time post-injury and was absent from healthy CSF; CSF marker PTGDS (22kD) had robust signal in Crl CSF but was absent acutely after TBI and 1d post-injury, and signals recovered stepwise on subsequent post- injury days. (B) Six CSF samples (30µl/lane) from four TBI patients (2.-5.) show variable signals of GFAP and large BDPs (50-37kD), and new small GFAP BDPs (25/23kD doublet, 20kD, 18kD) on injury day (patients 2., 3., 4a.) and 1d post-injury (4b., 5.) and control CSF of a 22 year old male (II.). (C) CSF immunoblots (30µl/lane) show full size

ALDOC (40kD) in five TBI patients (6.-10.) and variable intensity of 38kD ALDOC BDP on four days post-injury in three TBI patients (8.-10.) while a Crl subject showed no

ALDOC (III.). (D-K) Jitterplots (replicates averaged) and box-and-whisker plots, median

(line) and geometric mean (dashed) show logarithmic scaled immunoblot optical densities

(OD) of GFAP, S100β and AID biomarker CSF signals of 20-25 TBI patients on injury day and subsequent 5 post-injury days and 8-11 Controls (n: subjects numbers per day). (D)

Total GFAP (separate BDPs, S5) was elevated on all TBI days (black *, p<0.05) and declined over time (red *, p<0.001, i+4/5 p=0.002). (E) S100β was increased on each TBI day versus controls. (F) ALDOC (p< 0.004) and (G) GS (p < 0.001) were elevated on each day in TBI CSF versus Crls without significant decline. (H) BLBP (p< 0.03) and (I)

PEA15 (p< 0.004) had elevated mean levels in TBI on indicated days. Serum protein

APOB (J) was elevated in TBI versus Crl CSF (p<0.005). CSF standard PTGDS was

121

decreased in TBI versus Crl (p< 0.004) with levels depleted to various extents followed by recovery.

122

Figure 3.6: MRM mass spectrometry provides concentration comparison of AID biomarkers

(A) MRM-MS traces of specific peptides for GFAP, ALDOC and BLBP of three product ion traces (y #) with given mass over charge (m/z) values and their retention time (min, x-axis) of biomarker specific precursor ions of m/z 549.816 (for GFAP-specific peptide), m/z 526.970 (ALDOC) and m/z 446.256 (BLBP). Traces are of CSF samples from a 21 year old male severe TBI patient (left) and a 24 year old male Control (right). (B) Biplot

123

shows log MRM values of endogenous/standard GFAP peptide ion ratios (x-axis) over log GFAP immunoblot densities in TBI patients’ CSF samples with regression line and

Spearman correlation (r.s.=0.874, p<0.0001). Insert scatterplot of normal distributed residuals (y-axis) over normal quantiles (x-axis, Pearson coefficient R2=0.991, p<0.0001) validates accuracy of the two independent methods. (C) Mean MRM concentrations for

GFAP, BLBP, GS and ALDOC on injury day in TBI patients. ALDOC had 2.5-fold higher concentrations than GFAP. GFAP and ALDOC levels were over two orders larger than those of BLBP (p<0.001) and over three orders higher than GS (p<0.002). (D) Mean CSF concentrations on the third post-injury day of ALDOC were 10-fold higher than those of

GFAP (p=0.008). BLBP levels were lower than ALDOC and GFAP levels (p<0.001) and mean ALDOC levels were three orders of magnitude higher than those of GS (p=0.02).

124

Figure 3.7: Unsupervised multivariate biomarker analysis stratifies TBI patients using factors

(A) Unsupervised factor analysis grouped biomarkers S100, GFAP, small GFAP BDPs and APOB into Factor A (gray), ALDOC, 38 kD ALDOC BDP, BLBP, GS and PEA15 into

Factor B (green), given with their respective loading and Cronbach’s coefficients  for each factors reliability. (B) Scatterplot shows Factors A (x-axis) and Factor B (y-axis) CSF biomarker levels (z-units) from 12 subjects with signals for all biomarkers and they were partitioned between control and TBI (green dashed line) and between survivors and non- survivors (gray line). N=number of observations. (C) Classification tree boundaries that

125

partitioned controls, survivors and non-survivors of TBI using Factor thresholds. (D)

Standardized means of Factor A plotted over time post-injury show difference between

TBI survivor and non-survivors on several post-injury days (p<0.03) and decreases over days post-injury in both TBI patient groups (P<0.005). (E) Factor B means differed between TBI and Controls (P<0.001) but means did not differ significantly in survival or temporal profiles. n=number of subjects (D, E).

126

127

Figure 3.8: AID biomarkers are elevated in severe and mild TBI patient’s blood with unique kinetics

(A) Immunoblot signals for GFAP (25 kD BDP), aldolase A+C, ALDO (mab E9), PEA15 and BLBP in depleted 30µl plasma samples of a control subject (VI.) and 3 severe TBI patients on injury day and following 2-4 post-injury days. B-E) Scatterplots show levels in plasma with temporal profiles for GFAP 25 kD BDP (B), ALDOC (mab 5C9) (C), BLBP

(D) and PEA15 (E). Same patient data shown in A are connected by gray lines. (B) GFAP

25 kD BDP was absent on injury day and elevated on post-injury days 1-5 (p<0.0001).

(C) ALDOC levels were elevated in TBI (p<0.027) on injury day, first and second post- injury days (p<0.009) followed by decrease thereafter (between i+1 and i+4/5 p=0.041).

(D) Mean BLBP levels were increased on injury day in TBI (p=0.0067), stayed elevated on the first post-injury day and decreased subsequently (p< 0.017). (E) PEA15 levels were increased on injury day in TBI (p=0.024) and decreased thereafter (p<0.036). (F)

Pilot data show acute post-injury serum presence (see post-injury h) of ALDO (mab E9),

BLBP and PEA15 in CT-positive and CT-negative mild TBI patients while GFAP BDPs were absent or weak.

128

3.6 TABLES

Name Antibodies (Company) Epitope

Rabbit polyclonal anti-GFAP (DAKO, Whole cow GFAP; recognizes full size

Z0334) GFAP and large and small breakdown

products (BDPs). GFAP Chicken polyclonal anti GFAP Whole bovine GFAP; Recognizes full

(ThermoFisher Scientific, PA1-10004) size GFAP, large and small BDPs

Rabbit affinity purified polyclonal anti- Recombinant ALDOC fragment amino

ALDOC (Genetex, GTX102284) acids 10-163 (P09972)

Rabbit Serum 88 (Encor, gift) Recombinant whole ALDOC and BDP

Several monoclonal ALDOC Mab 1A1: C-terminal peptide ALDOC antibodies (Encor): IgG1 mab 1A1 Mab E9: Recombinant whole protein

(MCA-1A1), IgG1 mab E9 (MCA-E9), Mab 4A9: N-terminal peptide

IgG1 mab 4A9 (MCA-4A9), IgG1 mab (MPHSYPALSAEQKKELS)

5C9 Mab 5C9: N-terminus

Rabbit IgG fraction polyclonal anti GS GS peptide amino acids 357-373,

(Sigma, G2781) GS Mouse mab IgG2A to GS clone 6 (BD Full size GS

Transduction, 610517)

Rabbit polyclonal affinity purified anti Human PEA15 peptide surrounding PEA15 PEA15 (Cell Signaling) Leu60

Affinity purified rabbit polyclonal anti – GST-tagged recombinant full size BLBP FABP7 (Millipore) human FABP7, specific to BLBP

129

= FABP7 Affinity purified rabbit polyclonal anti- C-terminal human FABP7 peptide

FABP7 clone RB22973(Abgent) amino acids 104-132, specific to BLBP

Rabbit affinity purified polyclonal IgG Unspecified APOB peptide

APOB anti-APOB (PTGlab, 20578-1-AP) APOB 120-130 kD observed band, full

size 516 kD

Rabbit affinity purified IgG anti-PTGDS Synthetic human PTGDS peptide amino PTGDS (USBiological, P9053-24D) acids 120-190

Table 3.1: Primary antibodies

Listed are primary antibodies, commercial source, and epitopes used for Western blotting and immunocytochemistry.

130

Application Host, target, conjugate Dilution Company Catalog #

Goat anti-rabbit IgG, HRP 1:10,000 Thermo Fisher 31460

Goat anti-mouse IgG, HRP 1:10,000 Thermo Fisher 31430

blotting

Immuno Goat anti-chicken IgY, HRP 1:10,000 Thermo Fisher SA1-72012

711-545- Donkey anti-rabbit IgG, AlexaFluor 488 1:150 JacksonImmuno 152

711-605- Donkey anti-rabbit IgG, AlexaFluor 647 1:150 JacksonImmuno 152

711-165- Donkey anti-rabbit IgG, Cy 3 1:250 JacksonImmuno 152

Donkey anti-mouse IgG, AlexaFluor 715-545- 1:200 JacksonImmuno 488 151

715-165- Donkey anti-mouse IgG, Cy 3 1:150 JacksonImmuno 151

Immunocytochemistry Donkey anti-chicken IgY, AlexaFluor 703-605- 1:80 JacksonImmuno 647 155

705-545- Donkey anti-goat IgG, AlexaFluor 488 1:100 JacksonImmuno 003

712-585- Donkey anti-rat IgG, AlexaFluor 594 1:250 JacksonImmuno 150

Table 3.2: Secondary antibodies

Listed are secondary detection antibodies used for Western blots and immunocytochemistry with dilution and commercial sources.

131

Name Peptide Sequence Measured MRM Transition

GFAP ALAAELNQLR(Heavy) 554.821 (2+) --> 924.514 (1+, y8)

554.821 (2+) --> 853.477 (1+, y7)

554.821 (2+) --> 782.439 (1+, y6)

ALAAELNQLR(Light) 549.816 (2+) --> 914.505 (1+, y8)

549.816 (2+) --> 843.468 (1+, y8)

549.816 (2+) --> 722.431 (1+, y8)

LADVYQAELR (Heavy) 594.758 (2+) --> 1003.508 (1+, y8)

594.758 (2+) --> 789.413 (1+, y6)

594.758 (2+) --> 626.350 (1+, y5)

LADVYQAELR (Light) 589.314 (2+) --> 993.500 (1+, y8)

589.314 (2+) --> 779.405 (1+, y6)

589.314 (2+) --> 616.341 (1+, y5)

ALDOC TPSALAILENANVLAR (Heavy) 831.974 (2+) --> 1193.688 (1+ y11)

831.974 (2+) --> 1122.651 (1+ y10)

831.974 (2+) --> 1009.566 (1+ y9)

TPSALAILENANVLAR (Light) 826.970 (2+) --> 1183.679 (1+, y11)

826.970 (2+) --> 1112.642 (1+, y10)

826.970 (2+) --> 999.558 (1+, y9)

GS DIVEAHYR (Heavy) 506.758 (2+) --> 784.398 (1+, y6)

506.758 (2+) --> 685.329 (1+, y5)

506.758 (2+) --> 556.287 (1+, y4)

DIVEAHYR (Light) 501.753 (2+) --> 774.389 (1+, y6)

501.753 (2+) --> 675.321 (1+, y5)

501.753 (2+) --> 546.278 (1+, y4)

132

BLBP ALGVGFATR (Heavy) 451.260 (2+) --> 717.392 (1+, y7)

= FABP7 451.260 (2+) --> 660.370 (1+, y6)

451.260 (2+) --> 561.302 (1+, y5)

ALGVGFATR (Light) 446.256 (2+) --> 707.384 (1+, y7)

446.256 (2+) --> 650.362 (1+, y6)

446.256 (2+) --> 551.294 (1+, y5)

APOB SPAFTDLHLR (Heavy) 389.545 (3+) --> 764.429 (1+, y6)

389.545 (3+) --> 663.381 (1+, y5)

389.545 (3+) --> 491.771 (2+, y8)

SPAFTDLHLR (Light) 386.208 (3+) --> 754.421 (1+ y6)

386.208 (3+) --> 653.373 (1+ y5)

386.208 (3+) --> 486.767 (2+ y8)

PTGDS APEAQVSVQPNFQQDK (Heavy) 897.449 (2+) --> 1297.663 (1+, y11)

897.449 (2+) --> 1198.594 (1+, y10)

897.449 (2+) --> 884.435 (1+, y7)

APEAQVSVQPNFQQDK (Light) 893.442 (2+) --> 1289.648 (1+, y11)

893.442 (2+) --> 1190.580 (1+, y10)

893.442 (2+) --> 876.421 (1+, y7)

Table 3.3: MRM peptides and ion transitions

Human CSF biomarker-specific peptide precursor ions were selected for MRM-MS based on the above peptide and ion transition list. MRM-MS was operated in positive ion mode. m/z and charge state (CS [z]) values for each peptide listed.

133

Accession Protein Name

TBI Only CSF Proteins O94760 N(G),N(G)-dimethylarginine dimethylaminohydrolase 1 P11142 Heat shock cognate 71 kDa protein P18206 Vinculin P15104 Glutamine synthetase (GS) (EC 6.3.1.2) P12277 Creatine kinase B-type Q06830 Peroxiredoxin-1 P31946 14-3-3 protein beta/alpha P62258 14-3-3 protein epsilon P61981 14-3-3 protein gamma P63104 14-3-3 protein zeta/delta P23528 Cofilin-1 O75874 Isocitrate dehydrogenase cytoplasmic P00558 Phosphoglycerate kinase 1 P13796 Plastin-2 P67936 Tropomyosin alpha-4 chain Q13885 Tubulin beta-2A chain P27348 14-3-3 protein theta Q16555 Dihydropyrimidinase-related protein 2 P21333 Filamin-A Q12765 Secernin-1 P06753 Tropomyosin alpha-3 chain P14136 Glial fibrillary acidic protein P30041 Peroxiredoxin-6 P08670 Vimentin P80108 Phosphatidylinositol-glycan-specific phospholipase D P00491 Purine nucleoside phosphorylase P25713 Metallothionein-3 P00918 Carbonic anhydrase 2 Q01469 Fatty acid-binding protein, epidermal P30043 Flavin reductase Q06033 Inter-alpha-trypsin inhibitor heavy chain H3 P02545 Prelamin-A/C P26447 Protein S100-A4 P09382 Galectin-1 P09429 High mobility group protein B1 P26583 High mobility group protein B2 P18669 Phosphoglycerate mutase 1 Q71U36 Tubulin alpha-1A chain P04040 Catalase P21291 Cysteine and glycine-rich protein 1 P26038 Moesin P06703 Protein S100-A6 134

P28799 Granulins Heat shock-related 70 kDa protein 2 P35998 26S protease regulatory subunit 7 P68032 Actin, alpha cardiac muscle 1 P63261 Actin, cytoplasmic 2 P00568 Adenylate kinase isoenzyme 1 Q01518 Adenylyl cyclase-associated protein 1 P12814 Alpha-actinin-1 P04083 Annexin A1 P04114 Apolipoprotein B-100 P02655 Apolipoprotein C-II Q13790 Apolipoprotein F O14791 Apolipoprotein L1 O95445 Apolipoprotein M P08519 Apolipoprotein(a) P07738 Bisphosphoglycerate mutase P04003 C4b-binding protein alpha chain P20851 C4b-binding protein beta chain P05937 Calbindin P00915 Carbonic anhydrase 1 P16152 Carbonyl reductase P15169 Carboxypeptidase N catalytic chain P22792 Carboxypeptidase N subunit 2 P49913 Cathelicidin antimicrobial peptide O43866 CD5 antigen-like P06276 Cholinesterase P00740 Coagulation factor IX P05160 Coagulation factor XIII B chain P02745 Complement C1q subcomponent subunit A P31146 Coronin-1A P02741 C-reactive protein P06732 Creatine kinase M-type P13716 Delta-aminolevulinic acid dehydratase P81605 Dermcidin P15090 Fatty acid-binding protein, adipocyte P02792 Ferritin light chain Q9UGM5 Fetuin-B O75636 Ficolin-3 P05062 Fructose-bisphosphate aldolase B P06744 Glucose-6-phosphate isomerase P35754 Glutaredoxin-1 P78417 Glutathione S-transferase omega-1 P09211 Glutathione S-transferase P P69891 Hemoglobin subunit gamma-1 P26927 Hepatocyte growth factor-like protein 135

P10412 Histone H1.4 P16401 Histone H1.5 P62805 Histone H4 Q86YZ3 Hornerin Q14520 Hyaluronan-binding protein 2 P01591 Immunoglobulin J chain P02533 Keratin, type I cytoskeletal 14 P35908 Keratin, type II cytoskeletal 2 epidermal P13647 Keratin, type II cytoskeletal 5 P02788 Lactotransferrin P30740 Leukocyte elastase inhibitor P18428 Lipopolysaccharide-binding protein P00338 L-lactate dehydrogenase A chain P14151 L-selectin Q9Y5Y7 Lymphatic vessel endothelial hyaluronic acid receptor 1 P14174 Macrophage migration inhibitory factor P14780 Matrix metalloproteinase-9 P11137 Microtubule-associated protein 2 P19105 Myosin regulatory light chain 12A P12882 Myosin-1 P12883 Myosin-7 P35579 Myosin-9 P59665 Neutrophil defensin 1 P80188 Neutrophil gelatinase-associated lipocalin P30044 Peroxiredoxin-5, mitochondrial P02775 Platelet basic protein P02776 Platelet factor 4 P20742 Pregnancy zone protein P07737 Profilin-1 P27918 Properdin P25786 Proteasome subunit alpha type-1 P28072 Proteasome subunit beta type-6 P05109 Protein S100-A8 Q9UK55 Protein Z-dependent protease inhibitor Q92954 Proteoglycan 4 P31150 Rab GDP dissociation inhibitor alpha P52565 Rho GDP-dissociation inhibitor 1 P52566 Rho GDP-dissociation inhibitor 2 P0DJI8 Serum amyloid A-1 protein P0DJI9 Serum amyloid A-2 protein P02743 Serum amyloid P-component P04278 Sex hormone-binding globulin Q9H299 SH3 domain-binding glutamic acid-rich-like protein 3 P10599 Thioredoxin P07996 Thrombospondin-1 136

P62328 Thymosin beta-4 P37837 Transaldolase P29401 Transketolase P68363 Tubulin alpha-1B chain P68366 Tubulin alpha-4A chain P07437 Tubulin beta chain P68371 Tubulin beta-4B chain P09936 Ubiquitin carboxyl-terminal hydrolase isozyme L1 P04275 von Willebrand factor P61604 10 kDa heat shock protein, mitochondrial P62191 26S protease regulatory subunit 4 P17980 26S protease regulatory subunit 6A P43686 26S protease regulatory subunit 6B P62195 26S protease regulatory subunit 8 Q13200 26S proteasome non-ATPase regulatory subunit 2 P51665 26S proteasome non-ATPase regulatory subunit 7 P52209 6-phosphogluconate dehydrogenase, decarboxylating P00325 Alcohol dehydrogenase 1B Q9NZD4 Alpha-hemoglobin-stabilizing protein P20160 Azurocidin P02730 Band 3 anion transport protein Q562R1 Beta-actin-like protein 2 Q13938 Calcyphosin P62158 Calmodulin P08311 Cathepsin G P29762 Cellular retinoic acid-binding protein 1 Q15782 Chitinase-3-like protein 2 O43405 Cochlin P32320 Cytidine deaminase P19957 Elafin P12724 Eosinophil cationic protein P02794 Ferritin heavy chain Q05315 Galectin-10 P00739 Haptoglobin-related protein P0DMV8 Heat shock 70 kDa protein 1A P08238 Heat shock protein HSP 90-beta P69892 Hemoglobin subunit gamma-2 Q14103 Heterogeneous nuclear ribonucleoprotein D0 P22492 Histone H1t P20671 Histone H2A type 1-D O60814 Histone H2B type 1-K P68431 Histone H3.1 P01877 Ig alpha-2 chain C region P01880 Ig delta chain C region P01743 Ig heavy chain V-I region HG3 137

P23083 Ig heavy chain V-I region V35 P06331 Ig heavy chain V-II region ARH-77 P01824 Ig heavy chain V-II region WAH P01769 Ig heavy chain V-III region GA P01762 Ig heavy chain V-III region TRO P01779 Ig heavy chain V-III region TUR P01594 Ig kappa chain V-I region AU P01604 Ig kappa chain V-I region Kue P01605 Ig kappa chain V-I region Lay P01608 Ig kappa chain V-I region Roy P01610 Ig kappa chain V-I region WEA P01611 Ig kappa chain V-I region Wes P01616 Ig kappa chain V-II region MIL P04206 Ig kappa chain V-III region GOL P06311 Ig kappa chain V-III region IARC/BL41 P01624 Ig kappa chain V-III region POM P01623 Ig kappa chain V-III region WOL P04211 Ig lambda chain V region 4A P01701 Ig lambda chain V-I region NEW P01702 Ig lambda chain V-I region NIG-64 P04208 Ig lambda chain V-I region WAH P06889 Ig lambda chain V-IV region MOL A0M8Q6 Ig lambda-7 chain C region P04220 Ig mu heavy chain disease protein P09960 Leukotriene A-4 hydrolase P08637 Low affinity immunoglobulin gamma Fc region receptor III-A Q9BZG9 Ly-6/neurotoxin-like protein 1 P40121 Macrophage-capping protein P08493 Matrix Gla protein P02686 Myelin basic protein P20916 Myelin-associated glycoprotein P24158 Myeloblastin P05164 Myeloperoxidase P60660 Myosin light polypeptide 6 P29966 Myristoylated alanine-rich C-kinase substrate P22894 Neutrophil collagenase P59666 Neutrophil defensin 3 P08246 Neutrophil elastase P10153 Non-secretory ribonuclease P20472 Parvalbumin alpha O75594 Peptidoglycan recognition protein 1 P15259 Phosphoglycerate mutase 2 P0CG48 Polyubiquitin-C P31949 Protein S100-A11 P80511 Protein S100-A12 138

P04271 Protein S100-B P48539 Purkinje cell protein 4 P48741 Putative heat shock 70 kDa protein 7 Q9HD89 Resistin P63313 Thymosin beta-10 P07951 Tropomyosin beta chain Q13509 Tubulin beta-3 chain P04350 Tubulin beta-4A chain Q9BW30 Tubulin polymerization-promoting protein family member 3 Q9Y279 V-set and immunoglobulin domain-containing protein 4

TBI and Control CSF Proteins P02649 Apolipoprotein E P10909 Clusterin P09972 Fructose-bisphosphate aldolase C P60709 Actin, cytoplasmic 1 P07195 L-lactate dehydrogenase B chain P13645 Keratin, type I cytoskeletal 10 P00441 Superoxide dismutase P06733 Alpha-enolase P06396 Gelsolin P19823 Inter-alpha-trypsin inhibitor heavy chain H2 P30086 Phosphatidylethanolamine-binding protein 1 P04075 Fructose-bisphosphate aldolase A P40925 Malate dehydrogenase, cytoplasmic P32119 Peroxiredoxin-2 P04406 Glyceraldehyde-3-phosphate dehydrogenase P00751 Complement factor B P62937 Peptidyl-prolyl cis-trans isomerase A P01023 Alpha-2-macroglobulin P01019 Angiotensinogen P16070 CD44 antigen P0C0L5 Complement C4-B P01034 Cystatin-C Q12805 EGF-containing fibulin-like extracellular matrix protein 1 P18065 Insulin-like growth factor-binding protein 2 P36955 Pigment epithelium-derived factor Q13228 Selenium-binding protein 1 Q14515 SPARC-like protein 1 P19320 Vascular cell adhesion protein 1 P36222 Chitinase-3-like protein 1 Q9UBP4 Dickkopf-related protein 3 O14594 Neurocan core protein O00584 Ribonuclease T2 O14498 Immunoglobulin superfamily containing leucine-rich repeat protein 139

P01871 Ig mu chain C region P06681 Complement C2 P07108 Acyl-CoA-binding protein P07225 Vitamin K-dependent protein S P08294 Extracellular superoxide dismutase P09486 SPARC P12259 Coagulation factor V P13473 Lysosome-associated membrane glycoprotein 2 P17900 Ganglioside GM2 activator P19022 Cadherin-2 P43251 Biotinidase P49908 Selenoprotein P P78324 Tyrosine-protein phosphatase non-receptor type substrate 1 Q08380 Galectin-3-binding protein Q12841 Follistatin-related protein 1 Q13449 Limbic system-associated membrane protein Q13740 CD166 antigen Q14118 Dystroglycan Q8WXD2 Secretogranin-3 Q96GW7 Brevican core protein Q9P121 Neurotrimin P02768 Serum albumin P02787 Serotransferrin P01024 Complement C3 P00450 Ceruloplasmin P01008 Antithrombin-III P00738 Haptoglobin P02656 Apolipoprotein C-III P02790 Hemopexin P07339 Cathepsin D P07602 Prosaposin P07858 Cathepsin B P08571 Monocyte differentiation antigen CD14 P08697 Alpha-2-antiplasmin P09871 Complement C1s subcomponent P13591 Neural cell adhesion molecule 1 P16870 Carboxypeptidase E P23142 Fibulin-1 P43652 Afamin P61769 Beta-2-microglobulin Q12907 Vesicular integral-membrane protein VIP36 P02763 Alpha-1-acid glycoprotein 1 P19652 Alpha-1-acid glycoprotein 2 P01011 Alpha-1-antichymotrypsin P01009 Alpha-1-antitrypsin 140

P04217 Alpha-1B-glycoprotein P02765 Alpha-2-HS-glycoprotein P02647 Apolipoprotein A-I P02652 Apolipoprotein A-II P06727 Apolipoprotein A-IV P02654 Apolipoprotein C-I P05090 Apolipoprotein D P17174 Aspartate aminotransferase, cytoplasmic O75882 Attractin P98160 Basement membrane-specific heparan sulfate proteoglycan core protein P02749 Beta-2-glycoprotein 1 Q96KN2 Beta-Ala-His dipeptidase P55290 Cadherin-13 O94985 Calsyntenin-1 Q96IY4 Carboxypeptidase B2 Q9NQ79 Cartilage acidic protein 1 P43121 Cell surface glycoprotein MUC18 P00742 Coagulation factor X P00748 Coagulation factor XII P02452 Collagen alpha-1 P08123 Collagen alpha-2 P02746 Complement C1q subcomponent subunit B P02747 Complement C1q subcomponent subunit C P00736 Complement C1r subcomponent Q9NZP8 Complement C1r subcomponent-like protein P0C0L4 Complement C4-A P01031 Complement C5 P13671 Complement component C6 P10643 Complement component C7 P07357 Complement component C8 alpha chain P07358 Complement component C8 beta chain P07360 Complement component C8 gamma chain P02748 Complement component C9 P00746 Complement factor D P08603 Complement factor H Q03591 Complement factor H-related protein 1 P36980 Complement factor H-related protein 2 P05156 Complement factor I P08185 Corticosteroid-binding globulin Q16610 Extracellular matrix protein 1 P02671 Fibrinogen alpha chain P02675 Fibrinogen beta chain P02679 Fibrinogen gamma chain P02751 Fibronectin P09104 Gamma-enolase 141

P22352 Glutathione peroxidase 3 P69905 Hemoglobin subunit alpha P68871 Hemoglobin subunit beta P02042 Hemoglobin subunit delta P05546 Heparin cofactor 2 P04196 Histidine-rich glycoprotein P16403 Histone H1.2 P22692 Insulin-like growth factor-binding protein 4 P24592 Insulin-like growth factor-binding protein 6 Q16270 Insulin-like growth factor-binding protein 7 P35858 Insulin-like growth factor-binding protein complex acid labile subunit P19827 Inter-alpha-trypsin inhibitor heavy chain H1 Q14624 Inter-alpha-trypsin inhibitor heavy chain H4 P29622 Kallistatin P35527 Keratin, type I cytoskeletal 9 P04264 Keratin, type II cytoskeletal 1 P01042 Kininogen-1 P02750 Leucine-rich alpha-2-glycoprotein P51884 Lumican P61626 Lysozyme C P07333 Macrophage colony-stimulating factor 1 receptor P01033 Metalloproteinase inhibitor 1 P20774 Mimecan P02144 Myoglobin Q96PD5 N-acetylmuramoyl-L-alanine amidase O15394 Neural cell adhesion molecule 2 O00533 Neural cell adhesion molecule L1-like protein P55058 Phospholipid transfer protein P03952 Plasma kallikrein P05155 Plasma protease C1 inhibitor P05154 Plasma serine protease inhibitor P00747 Plasminogen Q15113 Procollagen C-endopeptidase enhancer 1 P41222 Prostaglandin-H2 D-isomerase P02760 Protein AMBP Q99497 Protein deglycase DJ-1 P06702 Protein S100-A9 P00734 Prothrombin P14618 Pyruvate kinase PKM P02753 Retinol-binding protein 4 P07998 Ribonuclease pancreatic Q86VB7 Scavenger receptor cysteine-rich type 1 protein M130 P35542 Serum amyloid A-4 protein P27169 Serum paraoxonase/arylesterase 1 O00391 Sulfhydryl oxidase 1 142

P22105 Tenascin-X P05452 Tetranectin P05543 Thyroxine-binding globulin Q15582 Transforming growth factor-beta-induced protein ig-h3 P02766 Transthyretin P60174 Triosephosphate isomerase P02774 Vitamin D-binding protein P04004 Vitronectin P54289 Voltage-dependent calcium channel subunit alpha-2/delta-1 P25311 Zinc-alpha-2-glycoprotein P63267 Actin, gamma-enteric smooth muscle P05067 Amyloid beta A4 protein P51693 Amyloid-like protein 1 O43505 Beta-1,4-glucuronyltransferase 1 P80723 Brain acid soluble protein 1 P07711 Cathepsin L1 P13987 CD59 glycoprotein Q8TCZ2 CD99 antigen-like protein 2 Q8N126 Cell adhesion molecule 3 Q8NFZ8 Cell adhesion molecule 4 P10645 Chromogranin-A P12109 Collagen alpha-1 chain Q12860 Contactin-1 Q9P0K1 Disintegrin and metalloproteinase domain-containing protein 22 Q13822 Ectonucleotide pyrophosphatase/phosphodiesterase family member 2 O94919 Endonuclease domain-containing 1 protein P54764 Ephrin type-A receptor 4 P61916 Epididymal secretory protein E1 Q96KK5 Histone H2A type 1-H P01876 Ig alpha-1 chain C region P01857 Ig gamma-1 chain C region P01859 Ig gamma-2 chain C region P01860 Ig gamma-3 chain C region P01861 Ig gamma-4 chain C region P01764 Ig heavy chain V-III region 23 P01766 Ig heavy chain V-III region BRO P01767 Ig heavy chain V-III region BUT P01768 Ig heavy chain V-III region CAM P01781 Ig heavy chain V-III region GAL P01765 Ig heavy chain V-III region TIL P01834 Ig kappa chain C region P01593 Ig kappa chain V-I region AG P01597 Ig kappa chain V-I region DEE P01598 Ig kappa chain V-I region EU P01613 Ig kappa chain V-I region Ni 143

P01606 Ig kappa chain V-I region OU P01617 Ig kappa chain V-II region TEW P01619 Ig kappa chain V-III region B6 P04207 Ig kappa chain V-III region CLL P01620 Ig kappa chain V-III region SIE P04433 Ig kappa chain V-III region VG P04434 Ig kappa chain V-III region VH P01625 Ig kappa chain V-IV region Len P80748 Ig lambda chain V-III region LOI P01714 Ig lambda chain V-III region SH P0CG04 Ig lambda-1 chain C regions P0CG05 Ig lambda-2 chain C regions P0CG06 Ig lambda-3 chain C regions Q9Y6R7 IgGFc-binding protein B9A064 Immunoglobulin lambda-like polypeptide 5 P01344 Insulin-like growth factor II Q92876 Kallikrein-6 O94772 Lymphocyte antigen 6H P04156 Major prion protein P41271 Neuroblastoma suppressor of tumorigenicity 1 P05408 Neuroendocrine protein 7B2 Q92823 Neuronal cell adhesion molecule O95502 Neuronal pentraxin receptor O15240 Neurosecretory protein VGF Q02818 Nucleobindin-1 P10451 Osteopontin Q96FE7 Phosphoinositide-3-kinase-interacting protein 1 Q9H3G5 Probable serine carboxypeptidase CPVL Q9UHG2 ProSAAS Q92520 Protein FAM3C Q99435 Protein kinase C-binding protein NELL2 P05060 Secretogranin-1 P13521 Secretogranin-2 O75326 Semaphorin-7A Q5TFQ8 Signal-regulatory protein beta-1 isoform 3 P04216 Thy-1 membrane glycoprotein P13611 Versican core protein Q8TAG5 V-set and transmembrane domain-containing protein 2A Q8TEU8 WAP, Kazal, immunoglobulin, Kunitz and NTR domain-containing protein 2

Healthy CSF Proteins Only P11021 78 kDa glucose-regulated protein P24593 Insulin-like growth factor-binding protein 5 Q08431 Lactadherin P04180 Phosphatidylcholine-sterol acyltransferase 144

Q06481 Amyloid-like protein 2 Q16620 BDNF/NT-3 growth factors receptor P11362 Fibroblast growth factor receptor 1 P22304 Iduronate 2-sulfatase Q86UX2 Inter-alpha-trypsin inhibitor heavy chain H5 Q96KG7 Multiple epidermal growth factor-like domains protein 10 P23471 Receptor-type tyrosine-protein phosphatase zeta Q9NPR2 Semaphorin-4B O14773 Tripeptidyl-peptidase 1 P30530 Tyrosine-protein kinase receptor UFO Q99969 Retinoic acid receptor responder protein 2 P26992 Ciliary neurotrophic factor receptor subunit alpha O95967 EGF-containing fibulin-like extracellular matrix protein 2 P21802 Fibroblast growth factor receptor 2 P98095 Fibulin-2 Q8NBJ4 Golgi membrane protein 1 P21246 Pleiotrophin O15031 Plexin-B2 P51888 Prolargin O60883 Prosaposin receptor GPR37L1 O75711 Scrapie-responsive protein 1 Q9Y646 Carboxypeptidase Q Q01459 Di-N-acetylchitobiase P40189 Interleukin-6 receptor subunit beta Q92859 Neogenin Q6UX71 Plexin domain-containing protein 2 P23470 Receptor-type tyrosine-protein phosphatase gamma O60241 Adhesion G protein-coupled receptor B2 P49641 Alpha-mannosidase 2x P55283 Cadherin-4 Q9BY67 Cell adhesion molecule 1 Q9Y287 Integral membrane protein 2B Q9HCB6 Spondin-1 P08253 72 kDa type IV collagenase P27797 Calreticulin P02461 Collagen alpha-1 chain P12111 Collagen alpha-3 chain P14625 Endoplasmin Q9UBQ6 Exostosin-like 2 P14314 Glucosidase 2 subunit beta O75144 ICOS ligand Mannosyl-oligosaccharide 1,2-alpha-mannosidase IA P16035 Metalloproteinase inhibitor 2 Q7Z7M0 Multiple epidermal growth factor-like domains protein 8 P32004 Neural cell adhesion molecule L1 145

P14543 Nidogen-1 P23515 Oligodendrocyte-myelin glycoprotein Q6UXB8 Peptidase inhibitor 16 P23284 Peptidyl-prolyl cis-trans isomerase B Q96S96 Phosphatidylethanolamine-binding protein 4 Q96NZ9 Proline-rich acidic protein 1 Q9NYQ8 Protocadherin Fat 2 P23468 Receptor-type tyrosine-protein phosphatase delta Q13332 Receptor-type tyrosine-protein phosphatase S P34096 Ribonuclease 4 Q9Y6N7 Roundabout homolog 1 Q8WZ42 Titin Q24JP5 Transmembrane protein 132A Q9BRK5 45 kDa calcium-binding protein O94910 Adhesion G protein-coupled receptor L1 O00468 Agrin P07686 Beta-hexosaminidase subunit beta Q9BQT9 Calsyntenin-3 Q8N3J6 Cell adhesion molecule 2 Q99674 Cell growth regulator with EF hand domain protein 1 Q6UW01 Cerebellin-3 Q16568 Cocaine- and amphetamine-regulated transcript protein P39060 Collagen alpha-1 chain P08174 Complement decay-accelerating factor Q02246 Contactin-2 Q9C0A0 Contactin-associated protein-like 4 P07585 Decorin P09417 Dihydropteridine reductase P52799 Ephrin-B2 O94769 Extracellular matrix protein 2 Q8IWU5 Extracellular sulfatase Sulf-2 Q9UBX5 Fibulin-5 P14207 Folate receptor beta Q6MZW2 Follistatin-related protein 4 O00451 GDNF family receptor alpha-2 P48058 Glutamate receptor 4 Q16769 Glutaminyl-peptide cyclotransferase Q9Y2T3 Guanine deaminase Q8IZP7 Heparan-sulfate 6-O-sulfotransferase 3 P18136 Ig kappa chain V-III region HIC P01622 Ig kappa chain V-III region Ti Q969P0 Immunoglobulin superfamily member 8 Q9NX62 Inositol monophosphatase 3 Q9UMF0 Intercellular adhesion molecule 5 O43291 Kunitz-type protease inhibitor 2 146

Q8N2S1 Latent-transforming growth factor beta-binding protein 4 Q9NT99 Leucine-rich repeat-containing protein 4B P42785 Lysosomal Pro-X carboxypeptidase P09603 Macrophage colony-stimulating factor 1 P22897 Macrophage mannose receptor 1 P55083 Microfibril-associated glycoprotein 4 Q16653 Myelin-oligodendrocyte glycoprotein Q9NY97 N-acetyllactosaminide beta-1,3-N-acetylglucosaminyltransferase 2 Q9ULB1 Neurexin-1 Q9P2S2 Neurexin-2 Q9Y4C0 Neurexin-3 Q9NPD7 Neuritin O94856 Neurofascin Q7Z3B1 Neuronal growth regulator 1 Q15818 Neuronal pentraxin-1 P47972 Neuronal pentraxin-2 Q5BLP8 Neuropeptide-like protein C4orf48 Q99574 Neuroserpin Q14112 Nidogen-2 Q14982 Opioid-binding protein/cell adhesion molecule Q99983 Osteomodulin P19021 Peptidyl-glycine alpha-amidating monooxygenase P01127 Platelet-derived growth factor subunit B Q9NZ53 Podocalyxin-like protein 2 P01210 Proenkephalin-A Q5FWE3 Proline-rich transmembrane protein 3 P01303 Pro-neuropeptide Y O15354 Prosaposin receptor GPR37 O60888 Protein CutA P48745 Protein NOV homolog Q8WZA1 Protein O-linked-mannose beta-1,2-N-acetylglucosaminyltransferase 1 Q9Y5F6 Protocadherin gamma-C5 C9JVW0 Putative transmembrane protein INAFM1 Q92932 Receptor-type tyrosine-protein phosphatase N2 Q16849 Receptor-type tyrosine-protein phosphatase-like N P78509 Reelin O75787 Renin receptor Q9BZR6 Reticulon-4 receptor Q6NW40 RGM domain family member B Q93091 Ribonuclease K6 Q9BYH1 Seizure 6-like protein Q6UXD5 Seizure 6-like protein 2 Q53EL9 Seizure protein 6 homolog Q96PX8 SLIT and NTRK-like protein 1 Q8WVQ1 Soluble calcium-activated nucleotidase 1 147

Q9H4F8 SPARC-related modular calcium-binding protein 1 O60279 Sushi domain-containing protein 5 Q08629 Testican-1 Q92563 Testican-2 O43493 Trans-Golgi network integral membrane protein 2 O75509 Tumor necrosis factor receptor superfamily member 21 Q6UX73 UPF0764 protein C16orf89 Q9UPU3 VPS10 domain-containing receptor SorCS3 A6NLU5 V-set and transmembrane domain-containing protein 2B Q15904 V-type proton ATPase subunit S1 Q9ULF5 Zinc transporter ZIP10

Table 3.4: TBI and Control CSF proteomes

Analytical mass spectrometry, by LC-MS/MS on LTQ-Orbitrap and Q-Exactive Orbitrap mass spectrometers was conducted on CSF from 19 severe and moderate TBI patients and 9 Control donors. Proteins were identified by MASCOT database searching

(SwissProt, Homo sapiens, ≥2 unique peptides, 95% peptide confidence) to arrive at cumulative conservative TBI and healthy CSF protein lists using same conditions. These

CSF proteomes served as clinical correlates to determine the signature of a previously identified astrocytic trauma-release proteome. Proteins from the mouse trauma release proteome found in TBI CSF are highlighted in yellow (32). Astrocyte enriched proteins are highlighted in blue (42). Proteins from both categories of trauma release and astrocyte enrichment are highlighted in green. Italics indicate proteins present in plasma proteomes

(43, 44).

148

Total % Dead ALDOC BLBP PEA15 GFAP

-0.0402 0.337 0.759 0.499 0.785 r s (Spearman) % 0.809 0.155 < 0.001 0.0338 < 0.001 p value Leaky 38 19 20 18 20 Number of observations

0.719 0.452 0.543 0.405

% Dead < 0.001 0.044 0.0194 0.0754

19 20 18 20

0.595 0.493 0.456 Total < 0.001 < 0.001 < 0.001 GFAP 54 42 54

0.398 0.681

ALDOC 0.0084 < 0.001

43 55

0.499

BLBP < 0.001

Table 3.5: Spearman correlations between culture trauma fluid biomarkers and astrocyte cell fates

The p-values for Spearman correlations of ranks are shown for culture-released total

GFAP (full size band and all fragments), ALDOC, BLBP and PEA15 as well as their correlation to rates of cell wounding (leaky membranes) and cell death. Numbers of culture observations for control and all post-stretching data are included.

149

150

151

Table 3.6: TBI patient data, clinical samples and conducted experimental analyses

(A) Figure IDs are roman italic numbers that refer to signals of control subject CSF and blood samples shown in Figures 3.5 and 3.8. Gender and age are listed for 17 healthy subjects who donated CSF, obtained by lumbar drain, blood or both. Sample measurements and replicates are listed for immunoblotting (IB), MRM-MS and proteome analyses (LC-MS/MS). (B) Italicized figure IDs list severe TBI patients whose CSF, obtained by ventriculostomy or/and blood samples are shown in Figures 3.5 and 3.8 with gender and age. Injury cause including motor vehicle accidents (MVA) and gunshot wound (GSW) is listed. Post-resuscitation GCS scores, survival and computed tomography (CT) findings are given. CT scan reports include presence of intracerebral hemorrhage (ICH) including one or more findings of contusion (Cnts), subdural hematoma

(SDH), subarachnoid hemorrhage (SAH) or intraventricular hemorrhage (IVH). Further, 152

presence of epidural hematoma (EDH), diffuse Axonal Injury (DAI), ischemia (Is) and edema or midline shift (Edm, mdls) are reported. Post-injury days are given for CSF, plasma and serum samples averaging multiple same-day samples for analyses. Italicized figure IDs indicate mild TBI patients whose serum IB data are shown in Figure 3.8F.

Gender, age, injury cause, GCS score and CT normal (CT-) or abnormal CT (CT+) is listed with scan findings as well as post-injury hrs are given for serum samples.

153

GFAP ALDOC BLBP GS PEA15 APOB PTGDS

Minimum OD 0.001 0.017 0.005 0.000 0.002 0.001 0.00039

Maximum OD 6.629 1.391 3.660 0.982 4.173 0.845 1.116

Orders of magnitude 4.1 1.9 2.9 3.4 3.3 3.0 3.5

# of observations 54 102 48 51 44 51 40

Table 3.7: Dynamic ranges of biomarker levels in CSF

Listed are minimum and maximum optical density readings for astrocyte injury biomarkers

GFAP, ALDOC, BLBP, GS and PEA15, as well as bleeding indicator APOB and CSF standard PTGDS. Resulting signal ranges are given in orders of magnitude and number of observation are listed (italic).

154

Interquartile Concentration Range Detection Limit Marker Immunoblot MRM

Immunoblot MRM Blood CSF CSF

ALDOC 0.2-1ng 58pg 1ng/mL - 13.3 ng/mL 600ng/mL - 1.3μg/mL 361ng/mL-1.5μg/mL

BLBP ~50pg 1pg 0ng/mL - 20ng/mL 2.3ng/mL - 20ng/mL 3.1ng/mL-14.4ng/mL

GFAP 8-40pg 1pg 267pg/mL - 20ng/mL 2.7ng/mL - 253ng/mL 86ng/mL-544ng/mL

Table 3.8: Concentrations and detection limits of ALDOC, BLBP and GFAP in TBI

CSF and blood

Pure protein dilution series in 0.5 % serum albumin were used to calibrate immunoblot densitometry signals and determine their approximate detection limits in TBI CSF and blood. Known amounts of heavy isotope labeled biomarker-specific peptides were used as standards to determine MRM-MS interquartile ranges in TBI CSF.

155

Spearman p n = Variable by Variable corr. value observations GFAP large < Very GFAP total 0.9877 64 BDPs 0.0001 strong < ALDOC total ALDOC 40 kD 0.9833 121 0.0001 GFAP small < APOB 0.898 BDPs 0.001 42 GFAP small < S100β 0.87 BDPs 0.001 54 APOB S100β 0.847 0 44 <.0001 PEA15 BLBP 0.8054 * 46 GFAP small < GFAP total 0.757 BDPs 0.001 64 Strong <.0001 S100β GFAP total 0.7391 * 54 APOB GS 0.726 0 44 <.0001 BLBP ALDOC total 0.6816 * 56 <.0001 PEA15 S100β 0.6772 * 43 <.0001 GS AldoC total 0.6724 * 53 APOB BLBP 0.638 0 44 <.0001 GS BLBP 0.603 * 49 APOB ALDOC total 0.602 0 46 GFAP small < BLBP 0.59 BDPs 0.001 54 Moderate <.0001 BLBP S100β 0.5833 * 51 <.0001 GS S100β 0.5826 * 46 <.0001 PEA15 GFAP total 0.5755 * 47 GFAP small < GS 0.573 BDPs 0.001 51 GFAP small < PEA15 0.572 BDPs 0.001 47 <.0001 PEA15 ALDOC total 0.5589 * 49 ALDOC 38kD GS 0.549 0.0009 BDP 33 APOB GFAP total 0.541 0.0002 42 <.0001 PEA15 GS 0.5334 * 49 ALDOC 38kD BLBP 0.532 0.0003 BDP 41

156

<.0001 BLBP GFAP total 0.5149 * 54 APOB PEA15 0.506 0.0002 48 ALDOC 38kD < ALDOC total 0.506 BPD 0.001 59 GFAP small < ALDOC total 0.477 BDPs 0.001 61 0.0005 ALDOC total S100β 0.4503 * 56 0.0017 ALDOC total GFAP total 0.3927 * 61 Weak 0.0190 GS GFAP total 0.3275 * 51 P<0.05 ALDOC 38kD PEA15 0.309 0.0749 BDP 34 Very weak ALDOC 38kD APOB 0.261 0.1353 BDP 34 PTGDS ALDOC total 0.017 0.893 61 ALDOC 38kD PTGDS -0.024 0.8779 BDP 42 ALDOC 38kD GFAP small -0.029 0.8628 BDP BDPs 39 ALDOC 38kD S100β -0.04 0.808 BDP 39 None PTGDS GS -0.13 0.3641 59 PTGDS APOB -0.183 0.2344 44 PTGDS BLBP -0.198 0.1473 57 ALDOC 38kD GFAP total -0.201 0.221 BDP 39 GFAP small PTGDS -0.251 0.055 BDPs 59 PTGDS PEA15 -0.307 0.0356 47 P<0.05 PTGDS S100β -0.314 0.0248 54 PTGDS GFAP total -0.446 0.0004 58 Moderate

Table 3.9: Biomarker panel correlations from CSF of severe TBI patients

Spearman rank correlation coefficients are given for all pairs of new and known astroglial neurotrauma biomarkers and CSF standards with their p-values and number of CSF samples analyzed. Coefficients >0.8-0.99 are considered very strong, 0.6-0.8 strong, moderate >0.4-0.6, weak <0.4 and divergent if coefficients were <-0.3.

157

3.7 SUPPLEMENTAL FIGURES

S3.1: The flow chart illustrates the AID biomarker selection strategy

Flow chart shows steps used to arrive at this astroglial biomarker panel. First, TBI and control CSF proteomes both generated by same LC-MS/MS settings were compiled and compared against each other (Table 3.4). 59 significantly trauma-changed proteins were previously identified by 2D gel analysis of culture medium (CM) from stretched astrocytes, the majority of which was present in the CSF proteomes (32). Then, astrocyte enriched proteins, by larger than 2-fold, were selected among the CSF and the trauma model proteome lists (42). From the resulting 14 candidates, those proteins present in healthy donor plasma were removed, including coactosin-like protein 1, heat shock cognate

71kDa protein, vinculin, apolipoprotein E, clusterin and lactate dehydrogenase B (43, 44).

Proteins with dominant expression outside the central nervous system (CNS) were also excluded: transgelin, F-box only protein 2 and N, N-dimethyl arginine dimethyl 158

aminohydrolase 1 (45). Resulting astroglial neurotrauma biomarker candidates were

ALDOC, GS, BLBP and PEA15, all with predominant expression in the CNS, with some presence outside of the CNS for GS and PEA15 (45).

159

S3.2: Human astrocyte cell leak and cell death populations show trends over time with combined severities

Shown are arithmetic means of percent (A) leaky (membrane wounded) and (B) dead astrocytes at various timepoints post-stretching, combining different pressure pulses.

Percent leaky cells were elevated at 30min against control and percent dead cells were elevated by 2 days against control and 30min post-injury levels using Kruskal-Wallis

ANOVA on ranks (p<0.001). One-day post-injury cell death rates nearly reached significance versus controls (p=0.052) in n=8-13 donors.

160

S3.3: Culture trauma temporal fluid profiles of full size GFAP plus large BDPs and small GFAP BDPs are different

(A) Mean fluid levels of full size GFAP (50 kD) and larger proteolytic breakdown products

(large BDPs: 47, 42, 37 kD) were significantly elevated over controls at all times post- stretching (asterisks), showed severity difference at 5h (black dot) and their release increased significantly between 30min and 1-2 days post-injury (triangles). (B) Small

GFAP fragments (25, 20 and 18kD BDPs) were only released significantly more than controls by 1 and 2 days post-injury (asterisks). Levels also differed significantly between timepoints (triangles).

161

S3.4: Peptide mapping of glial fibrillary acidic protein break-down products

Glial fibrillary acidic protein (GFAP) break-down products (BDPs) were immunoprecipitated from the conditioned medium (CM) and whole cell lysates (WCL) of stretch injured astrocytes. Immunoprecipitated proteins were separated by SDS-PAGE and stained with Sypro Ruby. Bands corresponding to the intact 49 kDa GFAP and 42,

38, 25, and 20 kDa BDPs were excised and in-gel digested with trypsin. LC-MS/MS peptide mapping identified a common core region starting from alanine residue 71 for all

BDPs. Lower molecular weight BDPs are generated from additional C-terminal cleavages.

162

S3.5: BLBP and GFAP are co-expressed in human astrocytes

A population of control human astrocytes show robust GFAP (white) and BLBP (green) expression (A, B). 30min post-stretching, BLBP signals were depleted. The same cells retained GFAP, but with altered distribution showing filament-disassembled or focal presence in process endings (C, D). Membrane-wounded cells have PI-positive nuclei

(red).

163

S3.6: TBI patient CSF trajectories are shown separately for full size plus large fragments and for small GFAP BDPs

Plots show CSF levels in Control and TBI patients of normalized GFAP optical densities

(A) for “upper” signal range (50kD and large BDPs <50-37kD) and (B) for “lower” GFAP bands (BDPs sizes between 18-25kD). All severe TBI patients had robust elevation in upper GFAP levels upper bands on injury day followed by decreases on later post-injury days (red asterisks). Smaller sized GFAP BDPs had variable signal levels and means decreased between one and 4-5 post-injury days. (A) Large GFAP signals (50-37kD) were elevated on all TBI days versus Crl (p<0.06) and declined over time as indicated in red (p <0.05, repeated measures ANOVA, see Methods). (B) Small GFAP BDPs (25-

18kD) were elevated in TBI versus Crl (p < 0.03) and declined between first post-injury day later post-injury days (p< 0.05).

164

S3.7: TBI patient CSF trajectories for astroglial injury-defined biomarkers measured by MRM-MS

Astroglial injury-defined biomarkers GFAP, ALDOC, BLBP, and GS were measured in

TBI patient CSF by MRM-MS. (A) GFAP was elevated on all TBI days compared to control

(grey *, p <0.01). (B) ALDOC was elevated on all TBI days compared to control (grey *, p

<0.01). (C) BLBP was elevated on all TBI days compared to control (grey *, p <0.05). (D)

165

GS was elevated on all TBI days compared to control (grey *, p <0.05).

166

S3.8: Temporal CSF trends for PTGDS, PEA15 and small GFAP BDPs differ between

TBI survivors and non-survivors

Explorative trend lines plot geometric mean levels of (A) PTGDS, (B) PEA15 and (C) small GFAP fragments in Controls (black), TBI survivors (red) and non-survivors (blue) with lower and upper bound error bars (95% confidence interval). CSF post-injury trajectories of GFAP lower BDPs are significantly elevated in non-survivors versus survivors by 28-fold on the first and 388-fold on the third post-injury day. PEA15 means are elevated over up to three orders of magnitude in non-survivors versus survivors. Mean

PTGDS levels decrease more in survivors than non-survivors, and levels recover gradually for survivors, resulting in significantly higher means on the third post-injury day compared to non-survivors of TBI. Statistical test was multiple measures ANOVA, mixed model, with non-constant intraclass variance (139).

167

S3.9: Temporal CSF trends for GFAP, ALDOC, BLBP, and GS differ between TBI survivors and non-survivors

Explorative trend lines plot geometric mean levels of CSF biomarker concentrations between TBI survivors and non-survivors. Survivor concentration profiles over time was shown in grey, blue, pink, and orange for (A) GFAP, (B) ALDOC, (C) BLBP, and (D) GS respectively. Non-survivor traces are displayed in red dashed lines. Patient n is listed by

168

color under each injury day post-TBI. SEM around the geometric mean are displayed by the error bars.

169

S3.10: Immunological and mass spectrometry measurements show comparable

CSF profiles of a severe TBI patient

Shown are temporal profiles of (A) GFAP, (B) ALDOC, (C) GS and (D) BLBP in longitudinal CSF samples of a severe TBI patient ‘1’ every 6h post-TBI. Biomarker levels were measured using immunoblot densitometry (continued lines, optical densities, left y- axes) and MRM-MS (dashed lines, ng/ml, right y-axes).

170

S3.11: Appearance of AID biomarkers in CSF and serum are different

Presence of GFAP, ALDOC, BLBP and PEA15 at 3 and 34h (first day post-injury) in CSF and serum of a severe TBI patient (patient 1, Table 3.6). Temporal profile of same

171

markers in concurrent CSF and serum samples of same patient. AID markers were elevated in serum acutely post-TBI, prior to GFAP elevation.

172

S3.12: Immunoblots shows distinct presence of AID biomarkers on injury day among 15 mild TBI patients

Composite shows immunoblots of pure proteins, depleted sera from control (Crl) and 15 mild TBI patients (clinical data see Table 3.6) between 1-31h post-concussion for GFAP

25 kD BDP, ALDOC, BLBP and PEA15. Absence (-) or presence (+) of CT findings in those patients are given. Exposure times match those shown for severe TBI patients

(Figure 3.8A).

173

3.8 REFERENCES

1. J. L. Wheble, D. K. Menon, TBI-the most complex disease in the most complex

organ: the CENTER-TBI trial-a commentary. Journal of the Royal Army Medical

Corps 162, 87-89 (2016); published online EpubApr (10.1136/jramc-2015-

000472).

2. B. Roozenbeek, A. I. Maas, D. K. Menon, Changing patterns in the epidemiology

of traumatic brain injury. Nature reviews. Neurology 9, 231-236 (2013); published

online EpubApr (10.1038/nrneurol.2013.22).

3. S. B. Rosenbaum, M. L. Lipton, Embracing chaos: the scope and importance of

clinical and pathological heterogeneity in mTBI. Brain imaging and behavior 6, 255-

282 (2012); published online EpubJun (10.1007/s11682-012-9162-7).

4. M. E. Brogan, J. J. Provencio, Spectrum of catastrophic brain injury: coma and

related disorders of consciousness. J Crit Care 29, 679-682 (2014); published

online EpubAug (10.1016/j.jcrc.2014.04.014).

5. A. Buki, N. Kovacs, E. Czeiter, K. Schmid, R. P. Berger, F. Kobeissy, D. Italiano,

R. L. Hayes, F. C. Tortella, E. Mezosi, A. Schwarcz, A. Toth, O. Nemes, S.

Mondello, Minor and repetitive head injury. Advances and technical standards in

neurosurgery 42, 147-192 (2015)10.1007/978-3-319-09066-5_8).

6. K. M. Helmick, C. A. Spells, S. Z. Malik, C. A. Davies, D. W. Marion, S. R. Hinds,

Traumatic brain injury in the US military: epidemiology and key clinical and

research programs. Brain imaging and behavior 9, 358-366 (2015); published

online EpubSep (10.1007/s11682-015-9399-z).

174

7. R. A. Stern, D. O. Riley, D. H. Daneshvar, C. J. Nowinski, R. C. Cantu, A. C.

McKee, Long-term consequences of repetitive brain trauma: chronic traumatic

encephalopathy. PM R 3, S460-467 (2011); published online EpubOct

(10.1016/j.pmrj.2011.08.008).

8. G. L. Sternbach, The Glasgow coma scale. J Emerg Med 19, 67-71 (2000).

9. E. W. Steyerberg, N. Mushkudiani, P. Perel, I. Butcher, J. Lu, G. S. McHugh, G.

D. Murray, A. Marmarou, I. Roberts, J. D. Habbema, A. I. Maas, Predicting

outcome after traumatic brain injury: development and international validation of

prognostic scores based on admission characteristics. PLoS medicine 5, e165;

discussion e165 (2008); published online EpubAug 5

(10.1371/journal.pmed.0050165).

10. N. A. Mushkudiani, C. W. Hukkelhoven, A. V. Hernandez, G. D. Murray, S. C. Choi,

A. I. Maas, E. W. Steyerberg, A systematic review finds methodological

improvements necessary for prognostic models in determining traumatic brain

injury outcomes. Journal of clinical epidemiology 61, 331-343 (2008); published

online EpubApr (10.1016/j.jclinepi.2007.06.011).

11. M. Bergsneider, D. A. Hovda, S. M. Lee, D. F. Kelly, D. L. McArthur, P. M. Vespa,

J. H. Lee, S. C. Huang, N. A. Martin, M. E. Phelps, D. P. Becker, Dissociation of

cerebral glucose metabolism and level of consciousness during the period of

metabolic depression following human traumatic brain injury. Journal of

neurotrauma 17, 389-401 (2000); published online EpubMay

(10.1089/neu.2000.17.389).

175

12. T. B. Meier, P. S. Bellgowan, R. Singh, R. Kuplicki, D. W. Polanski, A. R. Mayer,

Recovery of cerebral blood flow following sports-related concussion. JAMA

neurology 72, 530-538 (2015); published online EpubMay

(10.1001/jamaneurol.2014.4778).

13. O. Farkas, J. Lifshitz, J. T. Povlishock, Mechanoporation induced by diffuse

traumatic brain injury: an irreversible or reversible response to injury? The Journal

of neuroscience : the official journal of the Society for Neuroscience 26, 3130-3140

(2006).

14. K. A. Barbee, Mechanical cell injury. Ann N Y Acad Sci 1066, 67-84 (2005);

published online EpubDec (10.1196/annals.1363.006).

15. J. T. Povlishock, E. H. Pettus, Traumatically induced axonal damage: evidence for

enduring changes in axolemmal permeability with associated cytoskeletal change.

Acta neurochirurgica. Supplement 66, 81-86 (1996).

16. J. A. Colombo, A. Yanez, S. J. Lipina, Interlaminar astroglial processes in the

cerebral cortex of non human primates: response to injury. J Hirnforsch 38, 503-

512 (1997).

17. D. H. Smith, K. Uryu, K. E. Saatman, J. Q. Trojanowski, T. K. McIntosh, Protein

accumulation in traumatic brain injury. Neuromolecular medicine 4, 59-72

(2003)10.1385/NMM:4:1-2:59).

18. A. Buki, R. Siman, J. Q. Trojanowski, J. T. Povlishock, The role of calpain-mediated

spectrin proteolysis in traumatically induced axonal injury. Journal of

neuropathology and experimental neurology 58, 365-375 (1999).

176

19. M. A. Hemphill, S. Dauth, C. J. Yu, B. E. Dabiri, K. K. Parker, Traumatic brain injury

and the neuronal microenvironment: a potential role for neuropathological

mechanotransduction. Neuron 85, 1177-1192 (2015); published online EpubMar

18 (10.1016/j.neuron.2015.02.041).

20. S. Mondello, U. Muller, A. Jeromin, J. Streeter, R. L. Hayes, K. K. Wang, Blood-

based diagnostics of traumatic brain injuries. Expert Review of Molecular

Diagnostics 11, 65-78 (2011).

21. A. Petzold, Glial fibrillary acidic protein is a body fluid biomarker for glial pathology

in human disease. Brain research 1600, 17-31 (2015); published online EpubMar

10 (10.1016/j.brainres.2014.12.027).

22. P. K. Dash, J. Zhao, G. Hergenroeder, A. N. Moore, Biomarkers for the diagnosis,

prognosis, and evaluation of treatment efficacy for traumatic brain injury.

Neurotherapeutics : the journal of the American Society for Experimental

NeuroTherapeutics 7, 100-114 (2010).

23. S. H. Chou, C. S. Robertson, M. Participants in the International Multi-disciplinary

Consensus Conference on the Multimodality, Monitoring biomarkers of cellular

injury and death in acute brain injury. Neurocritical care 21 Suppl 2, S187-214

(2014); published online EpubDec (10.1007/s12028-014-0039-z).

24. A. M. Boutte, C. Yao, F. Kobeissy, X. C. May Lu, Z. Zhang, K. K. Wang, K. Schmid,

F. C. Tortella, J. R. Dave, Proteomic analysis and brain-specific systems biology

in a rodent model of penetrating ballistic-like brain injury. Electrophoresis 33, 3693-

3704 (2012); published online EpubDec (10.1002/elps.201200196).

177

25. J. M. Lubieniecka, F. Streijger, J. H. Lee, N. Stoynov, J. Liu, R. Mottus, T. Pfeifer,

B. K. Kwon, J. R. Coorssen, L. J. Foster, T. A. Grigliatti, W. Tetzlaff, Biomarkers

for severity of spinal cord injury in the cerebrospinal fluid of rats. PloS one 6,

e19247 (2011)10.1371/journal.pone.0019247).

26. S. Roche, A. Gabelle, S. Lehmann, Clinical proteomics of the cerebrospinal fluid:

Towards the discovery of new biomarkers. Proteomics Clin Appl 2, 428-436

(2008); published online EpubMar (10.1002/prca.200780040).

27. M. O. Sjodin, J. Bergquist, M. Wetterhall, Mining ventricular cerebrospinal fluid

from patients with traumatic brain injury using hexapeptide ligand libraries to

search for trauma biomarkers. Journal of chromatography. B, Analytical

technologies in the biomedical and life sciences 878, 2003-2012 (2010).

28. J. Hanrieder, M. Wetterhall, P. Enblad, L. Hillered, J. Bergquist, Temporally

resolved differential proteomic analysis of human ventricular CSF for monitoring

traumatic brain injury biomarker candidates. Journal of neuroscience methods

177, 469-478 (2009).

29. P. N. Lizhnyak, A. K. Ottens, Proteomics: in pursuit of effective traumatic brain

injury therapeutics. Expert review of proteomics 12, 75-82 (2015); published online

EpubFeb (10.1586/14789450.2015.1000869).

30. G. Poste, Bring on the biomarkers. Nature 469, 156-157 (2011).

31. G. Poste, Biospecimens, biomarkers, and burgeoning data: the imperative for

more rigorous research standards. Trends in molecular medicine 18, 717-722

(2012); published online EpubDec (10.1016/j.molmed.2012.09.003).

178

32. J. Levine, E. Kwon, M. Sondej, P. Paez, G. Czerwieniec, Y. Ao, M. V. Sofroniew,

J. A. Loo, I. B. Wanner, Traumatically injured astrocytes release a proteomic

signature modulated by STAT3 dependent cell survival. Glia 64, 668-694 (2016).

33. V. Muoio, P. B. Persson, M. M. Sendeski, The neurovascular unit - concept review.

Acta Physiol (Oxf) 210, 790-798 (2014); published online EpubApr

(10.1111/apha.12250).

34. P. J. Magistretti, Neuron-glia metabolic coupling and plasticity. Exp Physiol 96,

407-410 (2011).

35. L. Hertz, Bioenergetics of cerebral ischemia: a cellular perspective.

Neuropharmacology 55, 289-309 (2008); published online EpubSep

(10.1016/j.neuropharm.2008.05.023).

36. P. Mamczur, B. Borsuk, J. Paszko, Z. Sas, J. Mozrzymas, J. R. Wisniewski, A.

Gizak, D. Rakus, Astrocyte-neuron crosstalk regulates the expression and

subcellular localization of carbohydrate metabolism enzymes. Glia 63, 328-340

(2015); published online EpubFeb (10.1002/glia.22753).

37. D. Lovatt, U. Sonnewald, H. S. Waagepetersen, A. Schousboe, W. He, J. H. Lin,

X. Han, T. Takano, S. Wang, F. J. Sim, S. A. Goldman, M. Nedergaard, The

transcriptome and metabolic gene signature of protoplasmic astrocytes in the adult

murine cortex. The Journal of neuroscience : the official journal of the Society for

Neuroscience 27, 12255-12266 (2007); published online EpubNov 7

(10.1523/JNEUROSCI.3404-07.2007).

38. D. P. Pelvig, H. Pakkenberg, A. K. Stark, B. Pakkenberg, Neocortical glial cell

numbers in human brains. Neurobiology of aging 29, 1754-1762 (2008).

179

39. I. B. Wanner, M. A. Anderson, B. Song, J. Levine, A. Fernandez, Z. Gray-

Thompson, Y. Ao, M. V. Sofroniew, Glial scar borders are formed by newly

proliferated, elongated astrocytes that interact to corral inflammatory and fibrotic

cells via STAT3-dependent mechanisms after spinal cord injury. The Journal of

neuroscience : the official journal of the Society for Neuroscience 33, 12870-12886

(2013); published online EpubJul 31 (10.1523/JNEUROSCI.2121-13.2013).

40. D. R. Li, T. Ishikawa, L. Quan, D. Zhao, T. Michiue, B. L. Zhu, H. J. Wang, H.

Maeda, Morphological analysis of astrocytes in the hippocampus in mechanical

asphyxiation. Leg Med (Tokyo) 12, 63-67 (2010); published online EpubMar

(10.1016/j.legalmed.2009.11.005).

41. F. A. Azevedo, L. R. Carvalho, L. T. Grinberg, J. M. Farfel, R. E. Ferretti, R. E.

Leite, W. Jacob Filho, R. Lent, S. Herculano-Houzel, Equal numbers of neuronal

and nonneuronal cells make the human brain an isometrically scaled-up primate

brain. The Journal of comparative neurology 513, 532-541 (2009).

42. J. D. Cahoy, B. Emery, A. Kaushal, L. C. Foo, J. L. Zamanian, K. S.

Christopherson, Y. Xing, J. L. Lubischer, P. A. Krieg, S. A. Krupenko, W. J.

Thompson, B. A. Barres, A transcriptome database for astrocytes, neurons, and

oligodendrocytes: a new resource for understanding brain development and

function. The Journal of neuroscience : the official journal of the Society for

Neuroscience 28, 264-278 (2008).

43. G. S. Omenn, D. J. States, M. Adamski, T. W. Blackwell, R. Menon, H. Hermjakob,

R. Apweiler, B. B. Haab, R. J. Simpson, J. S. Eddes, E. A. Kapp, R. L. Moritz, D.

W. Chan, A. J. Rai, A. Admon, R. Aebersold, J. Eng, W. S. Hancock, S. A. Hefta,

180

H. Meyer, Y. K. Paik, J. S. Yoo, P. Ping, J. Pounds, J. Adkins, X. Qian, R. Wang,

V. Wasinger, C. Y. Wu, X. Zhao, R. Zeng, A. Archakov, A. Tsugita, I. Beer, A.

Pandey, M. Pisano, P. Andrews, H. Tammen, D. W. Speicher, S. M. Hanash,

Overview of the HUPO Plasma Proteome Project: results from the pilot phase with

35 collaborating laboratories and multiple analytical groups, generating a core

dataset of 3020 proteins and a publicly-available database. Proteomics 5, 3226-

3245 (2005).

44. S. Schenk, G. J. Schoenhals, G. de Souza, M. Mann, A high confidence, manually

validated human blood plasma protein reference set. BMC Med Genomics 1, 41

(2008).

45. M. Kapushesky, T. Adamusiak, T. Burdett, A. Culhane, A. Farne, A. Filippov, E.

Holloway, A. Klebanov, N. Kryvych, N. Kurbatova, P. Kurnosov, J. Malone, O.

Melnichuk, R. Petryszak, N. Pultsin, G. Rustici, A. Tikhonov, R. S. Travillian, E.

Williams, A. Zorin, H. Parkinson, A. Brazma, Gene Expression Atlas update--a

value-added database of microarray and sequencing-based functional genomics

experiments. Nucleic acids research 40, D1077-1081 (2012); published online

EpubJan (10.1093/nar/gkr913).

46. H. Haimoto, K. Kato, Highly Sensitive Enzyme-Immunoassay for Human-Brain

Aldolase-C. Clin Chim Acta 154, 203-212 (1986); published online EpubFeb 15

(Doi 10.1016/0009-8981(86)90032-X).

47. E. F. Ellis, J. S. McKinney, K. A. Willoughby, S. Liang, J. T. Povlishock, A new

model for rapid stretch-induced injury of cells in culture: characterization of the

model using astrocytes. Journal of neurotrauma 12, 325-339 (1995).

181

48. I. B. Wanner, An in vitro trauma model to study rodent and human astrocyte

reactivity. Methods in molecular biology 814, 189-219 (2012)10.1007/978-1-

61779-452-0_14).

49. J. Strotmann, I. Wanner, J. Krieger, K. Raming, H. Breer, Expression of odorant

receptors in spatially restricted subsets of chemosensory neurones. Neuroreport

3, 1053-1056 (1992).

50. I. B. Wanner, M. Deik, M. Torres, A. R. Rosendahl, J. T. Neary, V. P. Lemmon, J.

L. Bixby, A new in vitro model of the glial scar inhibits axon growth. Glia 56, 1691-

1709 (2008).

51. D. A. Hinkle, S. A. Baldwin, S. W. Scheff, P. M. Wise, GFAP and S100beta

expression in the cortex and hippocampus in response to mild cortical contusion.

Journal of neurotrauma 14, 729-738 (1997).

52. Z. Zhang, J. S. Zoltewicz, S. Mondello, K. J. Newsom, Z. Yang, B. Yang, F.

Kobeissy, J. Guingab, O. Glushakova, S. Robicsek, S. Heaton, A. Buki, J. Hannay,

M. S. Gold, R. Rubenstein, X. C. Lu, J. R. Dave, K. Schmid, F. Tortella, C. S.

Robertson, K. K. Wang, Human traumatic brain injury induces autoantibody

response against glial fibrillary acidic protein and its breakdown products. PLoS

One 9, e92698 (2014)10.1371/journal.pone.0092698).

53. K. Watanabe, Y. Urade, M. Mader, C. Murphy, O. Hayaishi, Identification of beta-

trace as prostaglandin D synthase. Biochemical and biophysical research

communications 203, 1110-1116 (1994); published online EpubSep 15

(10.1006/bbrc.1994.2297).

182

54. J. Clausen, Proteins in normal cerebrospinal fluid not found in serum. Proc Soc

Exp Biol Med 107, 170-172 (1961).

55. L. R. Fabrigar, D. T. Wegener, Exploratory factor analysis. Understanding statistics

(Oxford University Press, Oxford ; New York, 2012), pp. viii, 159 p.

56. L. Breiman, Classification and regression trees. The Wadsworth

statistics/probability series (Wadsworth International Group, Belmont, Calif.,

1984), pp. x, 358 p.

57. H. Zetterberg, K. Blennow, Fluid markers of traumatic brain injury. Molecular and

cellular neurosciences 66, 99-102 (2015); published online EpubMay

(10.1016/j.mcn.2015.02.003).

58. M. Bergsneider, D. A. Hovda, D. L. McArthur, M. Etchepare, S. C. Huang, N.

Sehati, P. Satz, M. E. Phelps, D. P. Becker, Metabolic recovery following human

traumatic brain injury based on FDG-PET: time course and relationship to

neurological disability. The Journal of head trauma rehabilitation 16, 135-148

(2001).

59. A. R. Mayer, P. S. Bellgowan, F. M. Hanlon, Functional magnetic resonance

imaging of mild traumatic brain injury. Neurosci Biobehav Rev 49, 8-18 (2015);

published online EpubFeb (10.1016/j.neubiorev.2014.11.016).

60. P. McCrory, W. H. Meeuwisse, M. Aubry, B. Cantu, J. Dvorak, R. J. Echemendia,

L. Engebretsen, K. Johnston, J. S. Kutcher, M. Raftery, A. Sills, B. W. Benson, G.

A. Davis, R. G. Ellenbogen, K. Guskiewicz, S. A. Herring, G. L. Iverson, B. D.

Jordan, J. Kissick, M. McCrea, A. S. McIntosh, D. Maddocks, M. Makdissi, L.

Purcell, M. Putukian, K. Schneider, C. H. Tator, M. Turner, Consensus statement

183

on concussion in sport: the 4th International Conference on Concussion in Sport

held in Zurich, November 2012. British journal of sports medicine 47, 250-258

(2013); published online EpubApr (10.1136/bjsports-2013-092313).

61. A. M. Boutte, Y. Deng-Bryant, D. Johnson, F. C. Tortella, J. R. Dave, D. A. Shear,

K. E. Schmid, Serum Glial Fibrillary Acidic Protein Predicts Tissue Glial Fibrillary

Acidic Protein Break-Down Products and Therapeutic Efficacy after Penetrating

Ballistic-Like Brain Injury. Journal of neurotrauma 33, 147-156 (2016); published

online EpubJan 1 (10.1089/neu.2014.3672).

62. S. Mondello, D. A. Shear, H. M. Bramlett, C. E. Dixon, K. E. Schmid, W. D. Dietrich,

K. K. Wang, R. L. Hayes, O. Glushakova, M. Catania, S. P. Richieri, J. T.

Povlishock, F. C. Tortella, P. M. Kochanek, Insight into Pre-Clinical Models of

Traumatic Brain Injury Using Circulating Brain Damage Biomarkers: Operation

Brain Trauma Therapy. Journal of neurotrauma 33, 595-605 (2016); published

online EpubMar 15 (10.1089/neu.2015.4132).

63. J. S. Zoltewicz, S. Mondello, B. Yang, K. J. Newsom, F. Kobeissy, C. Yao, X. C.

Lu, J. R. Dave, D. A. Shear, K. Schmid, V. Rivera, T. Cram, J. Seaney, Z. Zhang,

K. K. Wang, R. L. Hayes, F. C. Tortella, Biomarkers track damage after graded

injury severity in a rat model of penetrating brain injury. Journal of neurotrauma 30,

1161-1169 (2013); published online EpubJul 1 (10.1089/neu.2012.2762).

64. M. J. Whalen, T. Dalkara, Z. You, J. Qiu, D. Bermpohl, N. Mehta, B. Suter, P. G.

Bhide, E. H. Lo, M. Ericsson, M. A. Moskowitz, Acute plasmalemma permeability

and protracted clearance of injured cells after controlled cortical impact in mice.

Journal of cerebral blood flow and metabolism : official journal of the International

184

Society of Cerebral Blood Flow and Metabolism 28, 490-505 (2008); published

online EpubMar (10.1038/sj.jcbfm.9600544).

65. L. Papa, G. M. Brophy, R. D. Welch, L. M. Lewis, C. F. Braga, C. N. Tan, N. J.

Ameli, M. A. Lopez, C. A. Haeussler, D. I. Mendez Giordano, S. Silvestri, P.

Giordano, K. D. Weber, C. Hill-Pryor, D. C. Hack, Time Course and Diagnostic

Accuracy of Glial and Neuronal Blood Biomarkers GFAP and UCH-L1 in a Large

Cohort of Trauma Patients With and Without Mild Traumatic Brain Injury. JAMA

neurology 73, 551-560 (2016); published online EpubMay 1

(10.1001/jamaneurol.2016.0039).

66. W. Carr, A. M. Yarnell, R. Ong, T. Walilko, G. H. Kamimori, U. da Silva, R. M.

McCarron, M. L. LoPresti, Ubiquitin carboxy-terminal hydrolase-l1 as a serum

neurotrauma biomarker for exposure to occupational low-level blast. Frontiers in

neurology 6, 49 (2015)10.3389/fneur.2015.00049).

67. P. Shahim, T. Linemann, D. Inekci, M. A. Karsdal, K. Blennow, Y. Tegner, H.

Zetterberg, K. Henriksen, Serum Tau Fragments Predict Return to Play in

Concussed Professional Ice Hockey Players. Journal of neurotrauma, (2016);

published online EpubMay 2 (10.1089/neu.2014.3741).

68. B. A. Plog, M. Nedergaard, Why have we not yet developed a simple blood test for

TBI? Expert review of neurotherapeutics 15, 465-468 (2015); published online

EpubMay (10.1586/14737175.2015.1031112).

69. M. Latterich, J. E. Schnitzer, Streamlining biomarker discovery. Nature

biotechnology 29, 600-602 (2011); published online EpubJul (10.1038/nbt.1917).

185

70. F. G. Strathmann, S. Schulte, K. Goerl, D. J. Petron, Blood-based biomarkers for

traumatic brain injury: Evaluation of research approaches, available methods and

potential utility from the clinician and clinical laboratory perspectives. Clinical

biochemistry, (2014); published online EpubJan 31

(10.1016/j.clinbiochem.2014.01.028).

71. T. A. Addona, X. Shi, H. Keshishian, D. R. Mani, M. Burgess, M. A. Gillette, K. R.

Clauser, D. Shen, G. D. Lewis, L. A. Farrell, M. A. Fifer, M. S. Sabatine, R. E.

Gerszten, S. A. Carr, A pipeline that integrates the discovery and verification of

plasma protein biomarkers reveals candidate markers for cardiovascular disease.

Nature biotechnology 29, 635-643 (2011).

72. M. A. Gillette, S. A. Carr, Quantitative analysis of peptides and proteins in

biomedicine by targeted mass spectrometry. Nature methods 10, 28-34 (2013);

published online EpubJan (10.1038/nmeth.2309).

73. J. Lemoine, T. Fortin, A. Salvador, A. Jaffuel, J. P. Charrier, G. Choquet-

Kastylevsky, The current status of clinical proteomics and the use of MRM and

MRM(3) for biomarker validation. Expert review of molecular diagnostics 12, 333-

342 (2012); published online EpubMay (10.1586/erm.12.32).

74. Y. B. Lee, S. Du, H. Rhim, E. B. Lee, G. J. Markelonis, T. H. Oh, Rapid increase

in immunoreactivity to GFAP in astrocytes in vitro induced by acidic pH is mediated

by calcium influx and calpain I. Brain research 864, 220-229 (2000).

75. S. Du, A. Rubin, S. Klepper, C. Barrett, Y. C. Kim, H. W. Rhim, E. B. Lee, C. W.

Park, G. J. Markelonis, T. H. Oh, Calcium influx and activation of calpain I mediate

186

acute reactive gliosis in injured spinal cord. Experimental neurology 157, 96-105

(1999); published online EpubMay (10.1006/exnr.1999.7041).

76. R. E. White, D. M. McTigue, L. B. Jakeman, Regional heterogeneity in astrocyte

responses following contusive spinal cord injury in mice. The Journal of

comparative neurology 518, 1370-1390 (2010); published online EpubApr 15

(10.1002/cne.22282).

77. N. C. Reich, STAT3 revs up the powerhouse. Science signaling 2, pe61

(2009)10.1126/scisignal.290pe61).

78. J. Wegrzyn, R. Potla, Y. J. Chwae, N. B. Sepuri, Q. Zhang, T. Koeck, M. Derecka,

K. Szczepanek, M. Szelag, A. Gornicka, A. Moh, S. Moghaddas, Q. Chen, S.

Bobbili, J. Cichy, J. Dulak, D. P. Baker, A. Wolfman, D. Stuehr, M. O. Hassan, X.

Y. Fu, N. Avadhani, J. I. Drake, P. Fawcett, E. J. Lesnefsky, A. C. Larner, Function

of mitochondrial Stat3 in cellular respiration. Science 323, 793-797 (2009);

published online EpubFeb 6 (10.1126/science.1164551).

79. D. Kilinc, G. Gallo, K. A. Barbee, Mechanical membrane injury induces axonal

beading through localized activation of calpain. Experimental neurology 219, 553-

561 (2009); published online EpubOct (10.1016/j.expneurol.2009.07.014).

80. J. Sword, T. Masuda, D. Croom, S. A. Kirov, Evolution of neuronal and astroglial

disruption in the peri-contusional cortex of mice revealed by in vivo two-photon

imaging. Brain : a journal of neurology 136, 1446-1461 (2013); published online

EpubMay (10.1093/brain/awt026).

81. J. J. Bazarian, J. Zhong, B. Blyth, T. Zhu, V. Kavcic, D. Peterson, Diffusion tensor

imaging detects clinically important axonal damage after mild traumatic brain

187

injury: a pilot study. Journal of neurotrauma 24, 1447-1459 (2007); published

online EpubSep (10.1089/neu.2007.0241).

82. Kirov, II, A. Tal, J. S. Babb, J. Reaume, T. Bushnik, T. A. Ashman, S. Flanagan,

R. I. Grossman, O. Gonen, Proton MR spectroscopy correlates diffuse axonal

abnormalities with post-concussive symptoms in mild traumatic brain injury.

Journal of neurotrauma 30, 1200-1204 (2013); published online EpubJul 1

(10.1089/neu.2012.2696).

83. A. D. Lafrenaye, M. J. McGinn, J. T. Povlishock, Increased intracranial pressure

after diffuse traumatic brain injury exacerbates neuronal somatic membrane

poration but not axonal injury: evidence for primary intracranial pressure-induced

neuronal perturbation. Journal of cerebral blood flow and metabolism : official

journal of the International Society of Cerebral Blood Flow and Metabolism 32,

1919-1932 (2012); published online EpubOct (10.1038/jcbfm.2012.95).

84. K. Sakai, T. Fukuda, K. Iwadate, Beading of the astrocytic processes

(clasmatodendrosis) following head trauma is associated with protein degradation

pathways. Brain injury : [BI] 27, 1692-1697

(2013)10.3109/02699052.2013.837198).

85. J. A. Colombo, S. Gayol, A. Yanez, P. Marco, Immunocytochemical and electron

microscope observations on astroglial interlaminar processes in the primate

neocortex. Journal of neuroscience research 48, 352-357 (1997); published online

EpubMay 15 (

86. N. A. Oberheim, X. Wang, S. Goldman, M. Nedergaard, Astrocytic complexity

distinguishes the human brain. Trends in neurosciences 29, 547-553 (2006).

188

87. J. R. Gerstner, W. M. Vanderheyden, T. LaVaute, C. J. Westmark, L. Rouhana, A.

I. Pack, M. Wickens, C. F. Landry, Time of day regulates subcellular trafficking,

tripartite synaptic localization, and polyadenylation of the astrocytic Fabp7 mRNA.

The Journal of neuroscience : the official journal of the Society for Neuroscience

32, 1383-1394 (2012); published online EpubJan 25 (10.1523/JNEUROSCI.3228-

11.2012).

88. A. Schousboe, S. Scafidi, L. K. Bak, H. S. Waagepetersen, M. C. McKenna,

Glutamate metabolism in the brain focusing on astrocytes. Advances in

neurobiology 11, 13-30 (2014)10.1007/978-3-319-08894-5_2).

89. A. Eckert, B. C. Bock, K. E. Tagscherer, T. L. Haas, K. Grund, J. Sykora, C. Herold-

Mende, V. Ehemann, M. Hollstein, H. Chneiweiss, O. D. Wiestler, H. Walczak, W.

Roth, The PEA-15/PED protein protects glioblastoma cells from glucose

deprivation-induced apoptosis via the ERK/MAP kinase pathway. Oncogene 27,

1155-1166 (2008); published online EpubFeb 14 (10.1038/sj.onc.1210732).

90. P. Mergenthaler, A. Kahl, A. Kamitz, V. van Laak, K. Stohlmann, S. Thomsen, H.

Klawitter, I. Przesdzing, L. Neeb, D. Freyer, J. Priller, T. J. Collins, D. Megow, U.

Dirnagl, D. W. Andrews, A. Meisel, Mitochondrial hexokinase II (HKII) and

phosphoprotein enriched in astrocytes (PEA15) form a molecular switch governing

cellular fate depending on the metabolic state. Proceedings of the National

Academy of Sciences of the United States of America 109, 1518-1523 (2012);

published online EpubJan 31 (10.1073/pnas.1108225109).

91. P. J. Magistretti, Neuron-glia metabolic coupling and plasticity. J Exp Biol 209,

2304-2311 (2006); published online EpubJun (10.1242/jeb.02208).

189

92. N. R. Stein, D. L. McArthur, M. Etchepare, P. M. Vespa, Early cerebral metabolic

crisis after TBI influences outcome despite adequate hemodynamic resuscitation.

Neurocritical care 17, 49-57 (2012); published online EpubAug (10.1007/s12028-

012-9708-y).

93. P. Vespa, M. Bergsneider, N. Hattori, H. M. Wu, S. C. Huang, N. A. Martin, T. C.

Glenn, D. L. McArthur, D. A. Hovda, Metabolic crisis without brain ischemia is

common after traumatic brain injury: a combined microdialysis and positron

emission tomography study. Journal of cerebral blood flow and metabolism :

official journal of the International Society of Cerebral Blood Flow and Metabolism

25, 763-774 (2005); published online EpubJun (10.1038/sj.jcbfm.9600073).

94. B. L. Bartnik, S. M. Lee, D. A. Hovda, R. L. Sutton, The fate of glucose during the

period of decreased metabolism after fluid percussion injury: a 13C NMR study.

Journal of neurotrauma 24, 1079-1092 (2007)

95. D. A. Hovda, A. Yoshino, T. Kawamata, Y. Katayama, D. P. Becker, Diffuse

prolonged depression of cerebral oxidative metabolism following concussive brain

injury in the rat: a cytochrome oxidase histochemistry study. Brain research 567,

1-10 (1991).

96. B. L. Bartnik-Olson, U. Oyoyo, D. A. Hovda, R. L. Sutton, Astrocyte oxidative

metabolism and metabolite trafficking after fluid percussion brain injury in adult

rats. Journal of neurotrauma 27, 2191-2202 (2010); published online EpubDec

(10.1089/neu.2010.1508).

190

97. R. Bullock, W. L. Maxwell, D. I. Graham, G. M. Teasdale, J. H. Adams, Glial

swelling following human cerebral contusion: an ultrastructural study. Journal of

neurology, neurosurgery, and psychiatry 54, 427-434 (1991).

98. N. Hattori, S. C. Huang, H. M. Wu, W. Liao, T. C. Glenn, P. M. Vespa, M. E. Phelps,

D. A. Hovda, M. Bergsneider, PET investigation of post-traumatic cerebral blood

volume and blood flow. Acta neurochirurgica. Supplement 86, 49-52 (2003).

99. R. L. Sutton, D. A. Hovda, P. D. Adelson, E. C. Benzel, D. P. Becker, Metabolic

changes following cortical contusion: relationships to edema and morphological

changes. Acta neurochirurgica. Supplementum 60, 446-448 (1994).

100. X. Zhao, A. Ahram, R. F. Berman, J. P. Muizelaar, B. G. Lyeth, Early loss of

astrocytes after experimental traumatic brain injury. Glia 44, 140-152 (2003);

published online EpubNov (10.1002/glia.10283).

101. J. M. Lytle, J. R. Wrathall, Glial cell loss, proliferation and replacement in the

contused murine spinal cord. The European journal of neuroscience 25, 1711-

1724 (2007); published online EpubMar (10.1111/j.1460-9568.2007.05390.x).

102. D. R. Li, F. Zhang, Y. Wang, X. H. Tan, D. F. Qiao, H. J. Wang, T. Michiue, H.

Maeda, Quantitative analysis of GFAP- and S100 protein-immunopositive

astrocytes to investigate the severity of traumatic brain injury. Leg Med (Tokyo) 14,

84-92 (2012); published online EpubMar (10.1016/j.legalmed.2011.12.007).

103. V. Di Pietro, A. M. Amorini, G. Lazzarino, K. M. Yakoub, S. D'Urso, G. Lazzarino,

A. Belli, S100B and Glial Fibrillary Acidic Protein as Indexes to Monitor Damage

Severity in an In Vitro Model of Traumatic Brain Injury. Neurochemical research

40, 991-999 (2015); published online EpubMay (10.1007/s11064-015-1554-9).

191

104. P. E. Mouser, E. Head, K. H. Ha, T. T. Rohn, Caspase-mediated cleavage of glial

fibrillary acidic protein within degenerating astrocytes of the Alzheimer's disease

brain. The American journal of pathology 168, 936-946 (2006); published online

EpubMar (10.2353/ajpath.2006.050798).

105. T. T. Rohn, L. W. Catlin, W. W. Poon, Caspase-cleaved glial fibrillary acidic protein

within cerebellar white matter of the Alzheimer's disease brain. International

journal of clinical and experimental pathology 6, 41-48 (2013).

106. M. H. Chen, T. L. Hagemann, R. A. Quinlan, A. Messing, M. D. Perng, Caspase

cleavage of GFAP produces an assembly-compromised proteolytic fragment that

promotes filament aggregation. ASN neuro 5, e00125

(2013)10.1042/AN20130032).

107. R. B. Borgens, P. Liu-Snyder, Understanding secondary injury. The Quarterly

review of biology 87, 89-127 (2012).

108. D. G. Jones, Stability and storage characteristics of enzymes in sheep blood.

Research in veterinary science 38, 307-311 (1985).

109. D. G. Jones, Stability and storage characteristics of enzymes in cattle blood.

Research in veterinary science 38, 301-306 (1985).

110. S. Pontremoli, E. Melloni, M. Michetti, F. Salamino, B. Sparatore, B. L. Horecker,

Limited proteolysis of liver aldolase and fructose 1,6-bisphosphatase by lysosomal

proteinases: effect on complex formation. Proceedings of the National Academy

of Sciences of the United States of America 79, 2451-2454 (1982).

111. M. K. Offermann, J. F. Chlebowski, J. S. Bond, Action of cathepsin D on fructose-

1,6-bisphosphate aldolase. The Biochemical journal 211, 529-534 (1983).

192

112. M. F. Hopgood, S. E. Knowles, F. J. Ballard, Proteolysis of N-ethylmaleimide-

modified aldolase loaded into erythrocyte ghosts: prevention by inhibitors of

calpain. The Biochemical journal 259, 237-242 (1989).

113. J. S. Zoltewicz, D. Scharf, B. Yang, A. Chawla, K. J. Newsom, L. Fang,

Characterization of Antibodies that Detect Human GFAP after Traumatic Brain

Injury. Biomark Insights 7, 71-79 (2012)10.4137/BMI.S9873).

114. K. Fujita, T. Kato, M. Yamauchi, M. Ando, M. Honda, Y. Nagata, Increases in

fragmented glial fibrillary acidic protein levels in the spinal cords of patients with

amyotrophic lateral sclerosis. Neurochemical research 23, 169-174 (1998).

115. A. Martinez, M. Carmona, M. Portero-Otin, A. Naudi, R. Pamplona, I. Ferrer, Type-

dependent oxidative damage in frontotemporal lobar degeneration: cortical

astrocytes are targets of oxidative damage. Journal of neuropathology and

experimental neurology 67, 1122-1136 (2008); published online EpubDec

(10.1097/NEN.0b013e31818e06f3).

116. B. A. Plog, M. L. Dashnaw, E. Hitomi, W. Peng, Y. Liao, N. Lou, R. Deane, M.

Nedergaard, Biomarkers of traumatic injury are transported from brain to blood via

the glymphatic system. The Journal of neuroscience : the official journal of the

Society for Neuroscience 35, 518-526 (2015); published online EpubJan 14

(10.1523/JNEUROSCI.3742-14.2015).

117. M. M. Pelsers, T. Hanhoff, D. Van der Voort, B. Arts, M. Peters, R. Ponds, A. Honig,

W. Rudzinski, F. Spener, J. R. de Kruijk, A. Twijnstra, W. T. Hermens, P. P.

Menheere, J. F. Glatz, Brain- and heart-type fatty acid-binding proteins in the brain:

193

tissue distribution and clinical utility. Clinical chemistry 50, 1568-1575 (2004);

published online EpubSep (10.1373/clinchem.2003.030361).

118. T. M. Mathiisen, K. P. Lehre, N. C. Danbolt, O. P. Ottersen, The perivascular

astroglial sheath provides a complete covering of the brain microvessels: an

electron microscopic 3D reconstruction. Glia 58, 1094-1103 (2010).

119. M. Nuriya, T. Shinotsuka, M. Yasui, Diffusion properties of molecules at the blood-

brain interface: potential contributions of astrocyte endfeet to diffusion barrier

functions. Cerebral cortex 23, 2118-2126 (2013); published online EpubSep

(10.1093/cercor/bhs198).

120. C. L. Willis, L. Leach, G. J. Clarke, C. C. Nolan, D. E. Ray, Reversible disruption

of tight junction complexes in the rat blood-brain barrier, following transitory focal

astrocyte loss. Glia 48, 1-13 (2004); published online EpubOct

(10.1002/glia.20049).

121. M. M. Saw, J. Chamberlain, M. Barr, M. P. Morgan, J. R. Burnett, K. M. Ho,

Differential disruption of blood-brain barrier in severe traumatic brain injury.

Neurocritical care 20, 209-216 (2014); published online EpubApr

(10.1007/s12028-013-9933-z).

122. A. F. Logsdon, B. P. Lucke-Wold, R. C. Turner, J. D. Huber, C. L. Rosen, J. W.

Simpkins, Role of Microvascular Disruption in Brain Damage from Traumatic Brain

Injury. Compr Physiol 5, 1147-1160 (2015); published online EpubJul 1

(10.1002/cphy.c140057).

123. R. R. Hicks, D. H. Smith, D. H. Lowenstein, R. Saint Marie, T. K. McIntosh, Mild

experimental brain injury in the rat induces cognitive deficits associated with

194

regional neuronal loss in the hippocampus. Journal of neurotrauma 10, 405-414

(1993).

124. A. Korn, H. Golan, I. Melamed, R. Pascual-Marqui, A. Friedman, Focal cortical

dysfunction and blood-brain barrier disruption in patients with Postconcussion

syndrome. J Clin Neurophysiol 22, 1-9 (2005).

125. A. K. Shetty, V. Mishra, M. Kodali, B. Hattiangady, Blood brain barrier dysfunction

and delayed neurological deficits in mild traumatic brain injury induced by blast

shock waves. Front Cell Neurosci 8, 232 (2014)10.3389/fncel.2014.00232).

126. D. Vajtr, O. Benada, J. Kukacka, R. Prusa, L. Houstava, P. Toupalik, R. Kizek,

Correlation of ultrastructural changes of endothelial cells and astrocytes occurring

during blood brain barrier damage after traumatic brain injury with biochemical

markers of BBB leakage and inflammatory response. Physiological research /

Academia Scientiarum Bohemoslovaca 58, 263-268 (2009).

127. N. Marchi, J. J. Bazarian, V. Puvenna, M. Janigro, C. Ghosh, J. Zhong, T. Zhu, E.

Blackman, D. Stewart, J. Ellis, R. Butler, D. Janigro, Consequences of repeated

blood-brain barrier disruption in football players. PloS one 8, e56805

(2013)10.1371/journal.pone.0056805).

128. N. Marchi, M. Cavaglia, V. Fazio, S. Bhudia, K. Hallene, D. Janigro, Peripheral

markers of blood-brain barrier damage. Clin Chim Acta 342, 1-12 (2004); published

online EpubApr (10.1016/j.cccn.2003.12.008).

129. V. J. Willson, R. J. Thompson, Human brain aldolase C4 isoenzyme: purification,

radioimmunoassay, and distribution in human tissues. Annals of clinical

biochemistry 17, 114-121 (1980).

195

130. N. D. Silverberg, A. J. Gardner, J. R. Brubacher, W. J. Panenka, J. J. Li, G. L.

Iverson, Systematic review of multivariable prognostic models for mild traumatic

brain injury. Journal of neurotrauma 32, 517-526 (2015); published online EpubApr

15 (10.1089/neu.2014.3600).

131. B. K. Kwon, F. Streijger, N. Fallah, V. K. Noonan, L. M. Belanger, L. Ritchie, S. J.

Paquette, T. Ailon, M. C. Boyd, J. Street, C. G. Fisher, M. F. Dvorak, Cerebrospinal

Fluid Biomarkers To Stratify Injury Severity and Predict Outcome in Human

Traumatic Spinal Cord Injury. Journal of neurotrauma, (2016); published online

EpubAug 15 (10.1089/neu.2016.4435).

132. J. E. Buonora, A. M. Yarnell, R. C. Lazarus, M. Mousseau, L. L. Latour, S. B. Rizoli,

A. J. Baker, S. G. Rhind, R. Diaz-Arrastia, G. P. Mueller, Multivariate analysis of

traumatic brain injury: development of an assessment score. Front Neurol 6, 68

(2015)10.3389/fneur.2015.00068).

133. D. L. Bohac, J. F. Malec, A. M. Moessner, Factor analysis of the Mayo-Portland

Adaptability Inventory: Structure and validity. Brain Injury 11, 469-482 (1997);

published online EpubJul (Doi 10.1080/713802185).

134. J. DeJong, J. Donders, A confirmatory factor analysis of the California Verbal

Learning Test--Second Edition (CVLT-II) in a traumatic brain injury sample.

Assessment 16, 328-336 (2009); published online EpubDec

(10.1177/1073191109336989).

135. M. Asaka, T. Kimura, S. Nishikawa, M. Saitoh, T. Miyazaki, T. Takatori, E. Alpert,

Serum aldolase isozyme levels in patients with cerebrovascular diseases. The

American journal of the medical sciences 300, 291-295 (1990).

196

136. G. T. Manley, R. Diaz-Arrastia, M. Brophy, D. Engel, C. Goodman, K. Gwinn, T. D.

Veenstra, G. Ling, A. K. Ottens, F. Tortella, R. L. Hayes, Common data elements

for traumatic brain injury: recommendations from the biospecimens and

biomarkers working group. Archives of physical medicine and rehabilitation 91,

1667-1672 (2010); published online EpubNov (10.1016/j.apmr.2010.05.018).

137. G. Haralambie, Serum Aldolase Isoenzymes in Athletes at Rest and after Long-

Lasting Exercise. Int J Sports Med 2, 31-36 (1981)DOI 10.1055/s-2008-1034581).

138. O. Ruzgar, A. K. Bilge, Z. Bugra, S. Umman, E. Yilmaz, B. Ozben, B. Umman, M.

Meric, The use of human heart-type fatty acid-binding protein as an early

diagnostic biochemical marker of myocardial necrosis in patients with acute

coronary syndrome, and its comparison with troponin-T and creatine kinase-

myocardial band. Heart and vessels 21, 309-314 (2006); published online

EpubSep (10.1007/s00380-006-0908-2).

139. M. J. Crowder, D. J. Hand, in Monographs on statistics and applied probability.

(Chapman and Hall, London ; New York, 1990), chap. 1-3, pp. 1-59.

197

CHAPTER 4: ASSESSMENT OF ASTROGLIAL INJURY DEFINED BIOMARKERS

IN SPINAL CORD INJURY

4.1 INTRODUCTION

The long term debilitating effects of CNS injury in the form of head or spinal cord trauma is a major occupational concern for military personnel. High quality care and symptom mitigation for cases of severe neurotrauma starts with rapid, safe, and effective field transportation to emergency medical care centers (1, 2). Presently, there are reported clinical cases of spinal cord injury (SCI) patients’ conditions worsening as a result of transport to hospitals (3-5). Field care providers have reported on the severe pain experienced by casualties of SCI and traumatic brain injury (TBI) patients during bumpy and high vibrational ground and air transport. The present standard of care dictates the immobilization of SCI patients prior to transport, a procedure that may delay time to treatment. Currently, the interactions of patient immobilization, dynamic transport environment, and recovery and outcome are not clearly understood. It is believed that

SCI and TBI patients may be especially sensitive to repeated vibrational shock resulting from vehicle transport (2, 6) Because the extent of patient recovery is heavily dependent on early treatment (hours post-injury), a better understanding regarding transport effects is needed to better manage SCI casualties (7, 8).

In order to accurately assess whether the effects of vehicle transport exacerbate patient outcome, accurate diagnostics tools are needed to quantitatively assess and monitor injury state from the initial point of care to the medical treatment center to post- treatment patient outcome. This requires further preclinical research into the dynamics of

198

vehicular shock on pathology to better optimize transport protocols for improved patient outcome. Biomarkers capable of detecting the presence of injury that also relate to the severity of injury represent a minimally invasive, quantitative assessment of disease progression. To this end, we applied our astroglial injury derived (AID) protein biomarkers, initially identified for head injury, to the study of the effects of ground medical vehicle evacuation in a recoverable spinal cord injury animal model. While exact disease pathologies between head and spine injury differ, the high concentration of astroglia in the spinal cord make the application of our AID biomarkers to SCI assessment relevant as all astrocytes are believed to react to injury through a process of reactive astrogliosis with characteristic upregulation of GFAP expression and development of star like morphology (9). Consequently, biomechanical trauma induced plasmalemmal permeability and cell death should result in similar populations of astroglial protein release in a SCI model. What this study will uncover is whether these markers are (1) specific to

SCI, (2) whether AID marker concentrations are capable of stratifying injury severity, and

(3) whether transport post-SCI affects biofluid concentrations of our AID biomarkers.

Additionally, the occurrence of traumatic head trauma has been documented to occur concomitantly with spinal cord injury as a result of both classes of injuries resulting from high kinetic accidents such as falls and traffic accidents (10). SCIs and TBIs are the naturally occurring secondary injuries from the indirect forces of the initial trauma.

Cervical SCI, for example, occurs from the indirect forces to the spine from initial head injury (11). Because TBI patients suffer from cognitive and emotional deficits (e.g. attention deficits, inability to concentrate, memory loss, irritability, and impulse control), rehabilitation of patients suffering from both TBI concurrent with SCI is typically results in

199

poor patient outcome (12-14). The co-occurrence of TBI with SCI is estimated at approximately 40-60% (15, 16), making the need for better diagnostics tools essential as patients suffering from both conditions will require modified rehabilitative approaches.

The application of astroglial injury markers to SCI may set the foundation for better diagnostics of concomitant injury. Should AID biomarkers demonstrate SCI utility, future studies in dual TBI/SCI injury models may uncover distinct concentration profiles that indiscriminately identifies the presence of both injuries, allowing physicians to adjust treatment modalities to optimize patient care and recovery. In this chapter, we examine the application of 3 previously described (Chapter 3) biomarkers, glial fibrillary acidic protein (GFAP), aldolase C (ALDOC), and brain lipid binding protein (BLBP), for astroglial injury.

200

4.2 RESULTS

A total of 21 yucatan swine specimens were injured using a modified weight drop injury mechanism onto an exposed spinal cord (17). These animals were divided into 3 experimental groups – uninjured (sham) animals (n=7), SCI injured animals without vehicular transport (7), and SCI injured animals subjected to vehicular transport.

Cerebrospinal fluid (CSF) was extracted by lumbar puncture for each at animal within all three experimental groups at the following 5 time-points – baseline (pre-SCI), 15-30m post-SCI, 2-3h post-SCI, 2d post-SCI, and 7d post-SCI (Figure 4.1). Animals in our vehicle transport cohort were subjected to ground transport under experimentally controlled vibrational forces along the Aberdeen Test Center track (Figure 4.2) in New

Mexico. Transported animals were exposed to vehicle speeds ranging from 5 mph to 25 mph with respect to the extent of vibrational forces with a maximum exposure time not exceeding 1 hour.

AID biomarkers are sensitively detected in CSF and specific to SCI

Significantly elevated (p ≤ 0.02) AID biomarker concentrations for GFAP, ALDOC, and BLBP were observed acutely after SCI injury compared to pre-injury baseline levels

(Figure 4.3). Geometric means of GFAP CSF concentrations were measured at 25.9 ng/mL (± 23.9) and 15.1 ng/mL (± 31.3) 20 minutes and 2.7 hours after SCI respectively before dropping off to baseline levels by 2 days post SCI. Similar observations were made for ALDOC, displaying acute concentrations of 14.5 ng/mL (± 38.3) and 1.59 ng/mL (±

38.7) before dropping out in the post-acute period. BLBP levels also exhibited an early concentration drop off with concentrations of 13.9 ng/mL (± 34.0) at 20 minutes and 6.3

201

ng/mL (± 27.9) at 2.7 hours. Natural log transformed geometric means were calculated due to the large concentration differences observed between injury and non-injury conditions. Unlike arithmetic mean, geometric means tend to dampen the effects of outliers. The large standard errors (SE) measured from aggregative concentration values at each time-point can be explained by the large spread in concentration for GFAP (0-

3484.9 ng/mL), ALDOC (0-9397.0 ng/mL), and BLBP (0-2132.8 ng/mL). This variance is further assessed in later sections through evaluation of single animal biomarker profiles and differences in pathophysiological injury responses. GS was also evaluated as part of our AID SCI panel but was too weakly detected by PRM-MS for meaningful interpretation

(Figure 4.14).

High CSF biomarker concentration variance observed within SCI injured animals

Examination of individual animal CSF concentrations post-injury revealed distinct temporal profiles within our SCI swine cohort (Figure 4.4). As a whole, GFAP, ALDOC, and BLBP concentrations were observed to rise hyper-acutely (within 20 minutes) after surgery and weight drop before returning to base line levels by 2 days. 10 and 9 swine specimens maintained elevated GFAP and BLBP concentrations through the post- transport time-point (2.7h) while 1 animal displayed elimination of GFAP and BLBP respectively within the same time frame. 3 and 4 animals displayed no observable GFAP and BLBP readings (Figure 4.4A, C). ALDOC levels were similarly maintained in 7 of 14 specimens through the two acute post-SCI time-points with concentrations dropping out for 4 animals by 2.7h. ALDOC was not observed in 3 of 14 animals (Figure 4.4A). These

202

animal specific differences in biomarker CSF concentration highlight inter-specimen heterogeneity.

Weight-drop SCI varies in injury severity and outcome

Severity of pathophysiological damage was assessed at the impact site of surgically excised spinal cords from specimens sacrificed at 7 days. Considerable physical injury and molecular pathophysiology was observed at the impact site between animals in the SCI cohort. This was displayed in the varying degrees of bleeding and hemorrhage present at the injury cavity (Figure 4.5A). Various amounts of astroglial white matter damage, visualized by immunohistological staining, were observed around the injury site as shown by GFAP beading representing glial fiber fragmentation (Figure 4.5B).

Consistent with visual observations of differential lesion severities, total injury, measured by the combined expansion of tissue loss and white matter fragmentation, confirmed a range of injury severities between animals (Figure 4.5C, Table 4.1). Injury severity was further assessed by each individual animal’s ability to recovery ambulation after surgery rated by the Porcine Thoracic Injury Behavioral Scale (PTIBS) (17). Compared to baseline mobility scores, all animals exhibited diminished ambulation following SCI (Table 4.1).

Exceptions to this trend were observed for animals exhibiting minor tissue loss cavitation and white matter beading. Overall, a negative correlation (Spearman r.s. = -0.885, p-value

<0.001) was observed between SCI cavity length and recoverable mobility (Figure 4.6).

Biomarker levels and trajectories document trauma severity by the presence and size of an injury lesion

203

Immunohistological assessment revealed differential responses to SCI injury, resulting in a range of ambulatory recoveries. This physiological response difference is believed to explain the large standard deviations in acute CSF biomarker readings presented in Figure 4.3. To further investigate the relationship between trauma physiology and proximal fluid concentration, CSF concentrations of GFAP, ALDOC, and BLBP were distinguished between lesion positive (n=9) and negative (n=3) animals. Figure 4.7 demonstrates a distinct concentration difference between animals with measurable injury cavities at 7 days compared to those without. Biomarker concentration values were natural log transformed to calculate the geometric mean. This minimizes the skewing effects of non-normally distributed data that results from both injury response and individual specimen heterogeneity. GFAP, ALDOC, BLBP levels remained elevated in lesion positive animals for up to 2.7 hours after injury in biofluid compared to animals exhibiting minimal tissue loss with correspondingly high mobility 7 days after injury. T- tests (one-tailed) between cavity +/- animal showed highly significant (p-values < 0.01) differences in proximal fluid concentrations for GFAP, ALDOC, and BLBP 0.3h after injury. Additionally, 2.7h hour differences were also significant for biomarkers GFAP and

ALDOC but not BLBP (Figure 4.8).

Next, we examined the correlation between GFAP, ALDOC, and BLBP concentrations with cavity length. Strong (r.s. = 0.773, p-value < 0.001) to very strong (r.s.

= 0.991 and 0.829, p-values < 0.001) spearman correlations were observed for BLBP,

GFAP, and ALDOC respectively (Figure 4.9). Having demonstrated a relationship between physiological injury and biomarker concentrations, further stratification of injury severity was evaluated based on a combination of tissue loss diameter and radiation of

204

astroglial damage visualized by GFAP immunohistochemical staining around the injury site. Total injury was classified based on the total diameter of tissue loss and astroglial disintegration around the impact site (Table 4.1). However, lower correlation was obtained between this classification metric and biofluid concentration of our markers (Table 4.2).

Strong negative correlation was observed between 7 day PTIBS scores and AID biomarker CSF concentrations for GFAP and ALDOC with weaker association for BLBP

(Figure 4.10). Based on these findings, further SCI segregation was performed based on cavity size and 7 day ambulation scores. Using these two criteria, 14 SCI were further stratified into mild (n=5), moderate (n=6), and severe SCI (n=3) injury groups. This more in-depth injury classification uncovered additional differences between animals based on

AID biomarker concentrations (Figure 4.11). Statistical analysis highlighted the strongest distinction between both severe and mild (p-value < 0.001) and severe and moderate (p- value < 0.01) SCI for ALDOC at 0.3h post-injury (Figure 4.12). Significant distinctions were also observed for GFAP and BLBP at 0.3h between the three different injury groups.

Additionally, BLBP was observed to be significantly different at 2.7h post injury as well

(Figure 4.12).

Effect of transportation inconclusive

The assessment of transportation related effects on biofluid biomarker concentrations was inconclusive. Natural log transformed means for each marker over time did not show apparent increases in concentration or immediate alternations in kinetic profiles (Figure 4.13) when segregated (n=7 for each group) by whether ground transportation effects were experienced. It was noted that the profile of GFAP showed a

205

slower reduction in slope from 20m to 2.7h after injury. The effects of transportation on the rate of CSF concentration changes was subsequently evaluated but no significance

(p-value < 0.05) was established between animals (Table 4.3). The lowest p-values (0.14-

0.15) were observed for GFAP between 20m and 2.7h and ALDOC between 20m and 7 days.

206

4.3 DISCUSSION

The contusion SCI model used in this study produced less severe injuries to those reported in the literature using the same weight drop contusion apparatus (17, 18). In comparison to these studies where injured animals never recovered ambulation, half of the animals investigated in this study experienced recoverable injury along with higher mean PTIBS scores. Despite precise optimization of injury settings, histopathology revealed a range of injury severities at the site of impact. While our measured injury response was less consistent compared to previous groups, this injury response heterogeneity allowed a graded scale of severity, based on both lesion presence, size, diffuse astroglial damage expansion around the site of injury, and functional recovery, to be established and assessed in relation to the sensitivity and prognostic utility of our astroglial injury derived (AID) biomarker panel of GFAP, ALDOC, and BLBP.

Rapid identification of GFAP, ALDOC, and BLBP in SCI but not healthy animal

CSF demonstrated the successful application of AID biomarkers, developed originally for traumatic head injury, to the study of spinal cord contusion-like injuries. Our results highlighted the utility of AID biomarkers to not only identify the presence of SCI but also provide quantitative metrics that associate with pathophysiological and functional observations. In the acute post-injury phase (20m – 2.7h), GFAP, ALDOC, and BLBP displayed a graded CSF concentration response when animals were segregated based on presence of a 7 day lesion cavity as well as the combined expansion of tissue loss and astroglial fiber damage. Analysis of the relationship between tissue damage and behavioral locomotion at 7 days also revealed a strong association. Taken together, these findings suggest that robust and immediate elevation of GFAP, ALDOC, and BLBP can

207

successfully identify the severity of SCI and also predict the degree of functional recovery a week after injury.

Acute phase detection of surrogate signals of injury offers the promise of improved trauma management under hostile environments where decisions with long-lasting health implications need to be made quickly. Graded concentration signals may equip medical personnel with better diagnostic and prognostic tools to triage wounded soldiers in combat. In addition to severe injury, our AID biomarkers demonstrated the capacity to evaluate mild forms of injury where minor or no tissue loss was experienced. A sensitive measure for mild SCI has some applications for rapid on-field diagnostics that can be used to evaluate athletes immediately after injury to allow proper resting of individuals who have undergone a neurotraumatic event.

208

4.4 METHODS

Spinal Cord Injury

Spinal cord injury (SCI) was generated using an existing contusion injury model as described by Lee et al. (17). In brief, this procedure involves a surgical laminectomy of the thoracolumbar spine, followed by a spinal cord contusion that is caused by a weight

(50g) drop onto the exposed cord from 10 cm. The incision is closed, without fixation of the modified spinal vertebrae, leaving a possibly unstable spinal column. A consistent weight strike is critical to obtaining consistent injury severities across subjects. This procedure was performed by our collaborators at the Department of Defense.

CSF sample preparation

Protease inhibitors bestatin (40 micromoles per liter [µM]), pepstatin (1 microgram per milliliter [µg/ml]) and phosphoramidon (10 micromolar [µM]) and EDTA (1 millimolar

[mM]) were added to swine CSF samples followed by delipidation (centrifugation 10 min at 16,060 x g). Protein concentration was determined using Pierce 660 assay and a Tris- bovine serum albumin (BSA) dilution series.

Quantitative histopathology

Histopathological analyses were done 1 wk post-injury. Coronal (longitudinal) free floating 60 micrometer (µm) vibratome sections were prepared for each spinal segment containing the injury site, stained with Sudan Black to quench tissue autofluorescence, and permeabilized using a 0.5% Triton X100 solution. Nonspecific protein binding was blocked and a rabbit (rb) anti-GFAP (Dako) antibody was incubated overnight, followed

209

by detection using an AlexaFluor 488 donkey anti-rb secondary antibody. Bleeding was detected using a goat Cy3-conjugated anti-swine immunoglobulin (SwIgG) stain on repeated sections of each injured and sham animal.

Spinal tissue atrophy: rostro-caudal lesion expansion and white matter fiber damage

(clasmatodendrosis)

Longitudinal sections were collected ~0.5 mm from the dorsal surface until past the lesion, approximately over a depth of 3.5 to 4 mm (~48 sections/cord). As a proxy for spinal tissue atrophy, the average rostro-caudal length of the cavity was determined from measurements in 2 to 4 sections, covering a depth of >1 mm. All available sections of animals without lesion were examined to confirm lesion absence. Adjacent spinal segments were sectioned and stained if the injury expanded beyond the edge of the initial trimmed injury segment.

The abundance of astroglial process injury was used as proxy for white matter fiber damage. Glial fibrillary acidic protein (GFAP) staining was used to identify white matter and to quantify clasmatodendrosis (19-21). White matter was identified by uniquely organized, highly aligned, and brightly stained GFAP fibers.

Quantitation of Biomarkers in CSF using parrallel reaction monitoring- mass spectrometry

Synthetic standard peptides designed with stable isotope labeled arginine (6C13

14H 4N15 2O) and lysine (6C13 14H 2N15 2O) were purchased (Thermo Scientific) corresponding to our surrogate biomarker peptides. Peptide standards were prepared in

5% acetonitrile in water at a concentration of 5pmol/µL. Heavy peptide standards were

210

spiked into CSF samples to concentrations of 25 fmol/µL. CSF samples are then reduced, alkylated, and digested as described previously (Chapter 3).

Digested CSF peptides are dried by vacuum centrifugation and reconstituted in

0.1% formic acid, 3% acetonitrile in water. Samples are desalted using an on-line C18 trap column prior to LC-MS/MS analysis. Peptides were separated on a 5%-35% gradient of mobile phase B (0.1% formic acid in acetonitrile) over 40 minutes on a C18 PepMap

(Thermo) reversed phase HPLC column. Samples were analyzed by a parallel-reaction- monitoring (PRM-MS) workflow on a Q-Exactive Orbitrap MS operating in targeted-MS2 mode with an inclusion list of precursor peptide ions (Table 4.4) for MS2 analysis with the following parameters: resolution 17500, AGC target 2x105, maximum ion injection time

50ms, isolation window 3.0 Da, fixed first mass 100, and normalized collision energy

(NCE) 27.

Multiple-reaction-monitoring-mass spectrometry (MRM-MS) measured biomarker peptide specific precursor-product ion transitions isolated for monitoring. These precursor ions were fragmented by higher-energy collisional dissociation (HCD) into their component ions. Biomarker abundance was calculated based on the area under the curve

(AUC) of precursor to product ion transitions of each biomarker specific peptide using

Skyline (MacCoss Lab). The 3 transitions were summed and ratios of endogenous peptide to their heavy labeled counterparts were determined. Biomarker concentrations were calculated based on each peptide’s ratio of endogenous peptide AUC over added standard, heavy labeled peptide AUC, concentration of the labeled standard peptide, protein molecular weight (MW), and a dimensional conversion factor according to the following formula:

211

Endogenous protein concentration (ng/mL) = [endogenous/standard ratio × 50fmol heavy standard × protein MW × 1/1000]/[0.02 x raw CSF volume].

Statistical analysis

Biomarker concentration associations with functional recovery and immunohistology were performed by Spearman analysis using SIgmaplot. Determination of significance between injury severity stratifications, transportation effects, and development of tissue loss cavitation was performed by one-tailed t-test analysis using

Excel (Microsoft) and Graphpad (Prism).

212

4.5 FIGURES

Figure 4.1: Yucatan swine spinal cord injury (SCI) experimental design

21 Yucatan swine specimens were exposed to spinal cord contusion via a previously developed injury model (17). Animals were divided into 3 groups of 7 representing uninjured, SCI injured with vehicle transport, and SCI injured without transport cohorts.

SCI injured animals were first subjected to surgical laminectomy of the thoracolumbar spine to allow for a 10 cm weight drop onto the exposed spinal cord to induce contusion.

Following injury, the incision is closed without fixation of the spinal vertebrae. All 7 uninjured animals were vehicle transported. CSF samples were collected for biofluid analysis at a pre-SCI baseline time-point, a 15-30m post-SCI acute time-point, a 2-3h post-SCI acute time-point representing the post-transport condition, and 2 post-acute time-points at 2d and 7d post-SCI.

213

Figure 4.2: Test track layout and pass sequence for transportation effect assessment

The figure above displays the vehicle route used to evaluate effects of field transport.

Routes labeled 1-5 represent areas of the test track with different levels of surface unevenness that SCI animals were driven over to assess the impact of vibrational forces.

214

Figure 4.3: AID biomarkers GFAP, ALDOC, and BLBP acutely elevated after SCI

Comparison of cerebrospinal fluid (CSF) levels of glial fibrillary acidic protein (GFAP, A), aldolase C (ALDOC, B), and brain lipid binding protein (BLBP, C) in Yucatan swine at pre- spinal cord injury (SCI) baseline (Bl) versus acute (20m, 2.7h) and post-acute (2d,

7d) time-points after SCI. For GFAP, ALDOC, and BLBP, CSF concentrations were predominantly elevated within the first 24h of injury compared to baseline levels. Average

CSF collection times are displayed on the x-axis with geometric means of concentration values (ng/mL ± SD) on the y-axis. 215

Figure 4.4: Individual temporal concentration profiles demonstrate animal specific biomarker responses to SCI

AID biomarker concentrations for glial fibrillary acidic protein (GFAP, A), aldolase C

(ALDOC, B), and brain lipid binding protein (BLBP, C) were plotted with respect to time pre- and post-spinal cord injury (SCI) from individual Yucatan swine cerebrospinal fluid.

While GFAP, ALDOC, and BLBP concentrations are all elevated post-SCI, distinct differences (log10 y-axis) in protein concentrations (ng/mL) are observed between

216

animals. Concentration values of 0 were adjusted to 0.1 to accommodate the log10 y-axis scaling. Concentration values from 14 different animals are represented at each time- point.

217

Figure 4.5: Heterogeneous injury response observed in spinal cord injury (SCI) cohort

10 cm weight drop induced SCI contusion resulted in heterogeneous injury response at the site of impact (A). Tissue loss at the injury site was quantified by the diameter of lesion with bruises and hemorrhage. Astroglial damage was measured by visualization of white matter glial fibrillary acidic protein (GFAP) staining around the site of injury. Glial fiber disintegration (B) represents reversible, diffuse white matter injury. Total injury (C) was measured by the combination of both lesion cavity diameter (red) and expansion of white matter fragmentation (grey).

218

100 rs = -0.885 p < 0.001

Cavity Length (mm) Cavity

0.1 2 4 6 8 10 Crawl Recovery (PTIBS) Walk

Figure 4.6: SCI-related cavity length correlates negatively to animal recovery

A spearman correlation of r.s. = -0.885 (p <0.001) was observed for spinal cord injury site cavitation, measured by tissue loss, and ambulatory recovery at 7 days, measured by the

Porcine Thoracic Injury Behavioral Scale (PTIBS). PTIBS is graded from 1-10, with higher values representing higher recovered mobility. Data from this biplot indicates that high tissue (7 days) loss associates with poor recovery of walking at 7 days.

219

Figure 4.7: AID biomarkers display different concentration and temporal dynamics between animals with varying degrees of injury in cerebrospinal fluid

14 SCI injured animals were separated into cavity negative (-, yellow) and positive (+, red) injury groups based on extent of tissue damage. 7 uninjured (sham) pigs were also analyzed as the control group (grey). Geometric means of PRM-MS CSF concentrations for glial fibrillary acidic protein (GFAP, A), aldolase C (ALDOC, B), and brain lipid binding protein (BLBP, C) were plotted for 5 experimental time points (baseline (Bl), 20m, 2.7h,

220

2d, and 7d post-SCI) with 69% confidence intervals on a log10 y-scale. Concentration values of 0 were changed to 0.01 to accommodate log scaling for graphic visualization.

221

Figure 4.8: Animals that develop an injury lesion exhibit significantly higher CSF biomarker concentrations acutely after injury

2d, and 7d post-SCI) with 69% confidence intervals on a log10 y-scale. Concentration

222

values of 0 were changed to 0.01 to accommodate log scaling for graphic visualization.

A one-tailed, t-test was used to evaluate mean differences in biomarker concentrations between cavity + and – animals. GFAP and ALDOC levels were determined to be statistically higher up to 2.7h after injury while BLBP displayed quicker clearance kinetics and was only significantly different from the lesion negative group at 20m post-SCI.

223

Figure 4.9: CSF concentrations of GFAP, ALDOC, and BLBP associate positively with extent of tissue loss measured at 7 days

Biplots show strong negative Spearman correlation between CSF concentrations of

GFAP (A), ALDOC (B), and BLBP (C) and 7 day injury site cavitation measured by tissue loss.

224

Figure 4.10: GFAP, ALDOC, and BLBP may be predictive of functional recovery in spinal cord injury

Biplots show strong negative Spearman correlation between CSF concentrations of

GFAP (A), ALDOC (B), and BLBP (C) and ambulatory recovery graded by the Porcine

Thoracic Injury Behavioral Scale (PTIBS).

225

Figure 4.11: AID biomarkers demonstrate differences in CSF concentration profiles over time between injury severity assessed by cavity formation and 7 day ambulatory recovery

14 SCI injured animals were separated into mild (yellow, n=5), moderate (orange, n=6), and severe (red, n=3) injury groups based on extent of tissue damage and recovery of ambulation at 7 days (Table 4.1). 7 uninjured (sham) pigs were also analyzed as the control group (grey). Geometric means of PRM-MS CSF concentrations for glial fibrillary acidic protein (GFAP, A), aldolase C (ALDOC, B), and brain lipid binding protein (BLBP,

C) were plotted for 5 experimental time points (baseline (Bl), 20m, 2.7h, 2d, and 7d post-

226

SCI) with 69% confidence intervals on a log10 y-scale. Concentration values of 0 were changed to 0.01 to accommodate log scaling for graphic visualization.

227

Figure 4.12: AID biomarkers may be capable of distinguishing between varying levels of injury severity in the hyper acute post-injury period.

14 SCI injured animals were separated into mild (yellow ,n=5), moderate (orange, n=6), and severe (red, n=3) injury groups based on extent of tissue damage and recovery of ambulation at 7 days (Table 4.1). 7 uninjured (sham) pigs were also analyzed as the control group (grey). Geometric means of PRM-MS CSF concentrations for glial fibrillary acidic protein (GFAP, A), aldolase C (ALDOC, B), and brain lipid binding protein (BLBP,

C) were plotted for 5 experimental time points (baseline (Bl), 20m, 2.7h, 2d, and 7d post-

228

SCI) with 69% confidence intervals on a log10 y-scale. Concentration values of 0 were changed to 0.01 to accommodate log scaling for graphic visualization.

229

Figure 4.13: Vibrational forces from ground transport did not significantly impact biofluid concentrations of AID biomarkers

21 animals were separated into sham (grey, n=7), SCI only (orange, n=7), and SCI + transportation (blue, n=7) experimental groups. Geometric means of PRM-MS CSF concentrations for glial fibrillary acidic protein (GFAP, A), aldolase C (ALDOC, B), and brain lipid binding protein (BLBP, C) were plotted for 5 experimental time points (baseline

(Bl), 20m, 2.7h, 2d, and 7d post-SCI) with 69% confidence intervals on a log10 y-scale.

230

Concentration values of 0 were changed to 0.01 to accommodate log scaling for graphic visualization.

231

Figure 4.14: GS was weakly detected in pig SCI CSF samples

(A) Line and (B) bar graph representations of GS CSF concentration in sham injured and

SCI (separated into cavity negative (yellow) and positive (red)) swine. Natural log transformed means are plotted with corresponding 69% confidence intervals. However,

GS was poorly detected in CSF by MRM-MS, resulting in too few values for adequate analysis.

232

4.6 TABLES

0.3h CSF 0.3h CSF 0.3h CSF

PTIBS Total-Injury Cavity GFAP ALDOC BLBP

Animal (7day) (mm) (mm) (ng/mL) (ng/mL) (ng/mL)

43-090 9.8 47.6 0.0 0.0 0.0 1.1

46-091 8.4 8.8 6.8 24.1 33.7 0.0

42-131 9.9 0.0 0.0 0.0 0.0 0.0

47-094 9.9 5.6 0.0 0.0 0.0 0.0

46-030 4.5 8.8 7.2 64.4 9.0 67.2

42-115 9.8 6.2 0.9 17.7 35.8 26.2

42-101 3 13.4 9.9 75.6 1893.5 0.0

43-031 9.3 0.0 0.0 0.0 2.3 37.4

43-082 4.2 11.2 7.9 57.0 1.4 202.7

42-017 9.3 9.4 8.3 85.7 1.7 95.0

42-127 3.8 11.7 10.2 355.0 40.6 289.2

42-068 1.9 27.9 18.0 401.7 717.7 323.8

45-157 3.4 32.0 30.8 1729.4 668.5 389.6

42-132 3.1 46.0 36.0 2584.9 9397.0 2132.8

Table 4.1: Spinal cord impact site injury severity measured by astroglial beading and tissue loss cavity size

Extent of pathophysiological damage following weight drop spinal cord injury (SCI) was quantified by physical and histological examination of excised spinal cord from animals sacrificed after the 7 day time-point. Tissue loss at 7 days was quantified by the diameter

233

of the lesion cavity and extent of clasmatodendrosis was quantified by the distance astroglial beading measured from the site of injury visualized by immunostaining. Total injury was measured by combination of both these parameters. Ambulatory recovery at 7 days is presented as a Porcine Thoracic Injury Behavioral Scale (PTIBS) score. PTIBS is graded from 1-10, with higher values representing higher recovered mobility. Animals highlighted in green, yellow, and red represent our classification of mild, moderate, and severe SCI respectively.

234

Spearman Correlations Cavity Length Total Injury Expansion Biomaker (mm) PTIBS (mm) GFAP r.s. 0.991 -0.847 0.647 p *** *** * ALDOC r.s. 0.824 -0.848 0.484 p *** *** ns BLBP r.s. 0.773 -0.622 0.560 p *** * *

Table 4.2: Spearman correlations for biofluid biomarker concentrations to immunohistology and functional recovery

The table lists Spearman correlations (r.s.) between biomarkers GFAP, ALDOC, BLBP and corresponding histopathological (tissue loss diameter and total injury expansion including tissue cavity and diffusion of astroglial fiber damage) and functional observations (mobility at 7 days post-injury was assessed using the Porcine Thoracic

Injury Behavioral Scale (PTIBS))

235

SCI SCI+T Biomarker Slope [ng/(mL*h)] Mean (ng/(mL*h)) Mean (ng/(mL*h)) p-value GFAP m (20m-2.7h) -74.9 55.1 0.14 m (20m-7d) -2.7 -2.0 0.40 ALDOC m (20m-2.7h) -42.0 193.4 0.36 m (20m-7d) -1.5 -8.5 0.15 BLBP m (20m-2.7h) -51.1 -34.7 0.34 m (20m-7d) -1.4 -2.1 0.48

Table 4.3: t-test of AID biomarker slope changes between transported and un- transported animals

Effects of transport on CSF concentration of astroglial injury-defined biomarkers were measured by assessing the rate of changes within the hyper-active post-injury time-points

(20m, 2.7h) and the hyper-acute and post-acute post-injury time-points (20m, 2d). Rate of change was defined by change in biomarker concentration over time in hours.

236

Name Peptide Sequence Measured MRM Transition

GFAP ALAAELNQLR(Heavy) 554.821 (2+) --> 924.514 (1+, y8) 554.821 (2+) --> 853.477 (1+, y7) 554.821 (2+) --> 782.439 (1+, y6) ALAAELNQLR(Light) 549.816 (2+) --> 914.505 (1+, y8) 549.816 (2+) --> 843.468 (1+, y8) 549.816 (2+) --> 722.431 (1+, y8) LADVYQAELR (Heavy) 594.758 (2+) --> 1003.508 (1+, y8) 594.758 (2+) --> 789.413 (1+, y6) 594.758 (2+) --> 626.350 (1+, y5) LADVYQAELR (Light) 589.314 (2+) --> 993.500 (1+, y8) 589.314 (2+) --> 779.405 (1+, y6) 589.314 (2+) --> 616.341 (1+, y5)

ALDOC TPSALAILENANVLAR (Heavy) 831.974 (2+) --> 1193.688 (1+ y11) 831.974 (2+) --> 1122.651 (1+ y10) 831.974 (2+) --> 1009.566 (1+ y9) TPSALAILENANVLAR (Light) 826.970 (2+) --> 1183.679 (1+, y11) 826.970 (2+) --> 1112.642 (1+, y10) 826.970 (2+) --> 999.558 (1+, y9)

GS DIVEAHYR (Heavy) 506.758 (2+) --> 784.398 (1+, y6) 506.758 (2+) --> 685.329 (1+, y5) 506.758 (2+) --> 556.287 (1+, y4) DIVEAHYR (Light) 501.753 (2+) --> 774.389 (1+, y6) 501.753 (2+) --> 675.321 (1+, y5) 501.753 (2+) --> 546.278 (1+, y4)

BLBP ALGVGFATR (Heavy) 451.260 (2+) --> 717.392 (1+, y7) = FABP7 451.260 (2+) --> 660.370 (1+, y6) 451.260 (2+) --> 561.302 (1+, y5) ALGVGFATR (Light) 446.256 (2+) --> 707.384 (1+, y7)

237

446.256 (2+) --> 650.362 (1+, y6) 446.256 (2+) --> 551.294 (1+, y5)

Table 4.4: PRM-MS precursor ion inclusion list and measured transitions

The above table lists PRM-MS transitions monitored for quantitative mass spectrometry analysis of biomarkers GFAP, ALDOC, BLBP, and GS in Yucatan swine SCI CSF.

238

4.7 REFERENCES

1. M. Hauswald, A re-conceptualisation of acute spinal care. Emergency medicine

journal : EMJ 30, 720-723 (2013); published online EpubSep (10.1136/emermed-

2012-201847).

2. N. Theodore, B. Aarabi, S. S. Dhall, D. E. Gelb, R. J. Hurlbert, C. J. Rozzelle, T.

C. Ryken, B. C. Walters, M. N. Hadley, Transportation of patients with acute

traumatic cervical spine injuries. Neurosurgery 72 Suppl 2, 35-39 (2013); published

online EpubMar (10.1227/NEU.0b013e318276edc5).

3. E. R. Haut, B. T. Kalish, D. T. Efron, A. H. Haider, K. A. Stevens, A. N. Kieninger,

E. E. Cornwell, 3rd, D. C. Chang, Spine immobilization in penetrating trauma: more

harm than good? The Journal of trauma 68, 115-120; discussion 120-111 (2010);

published online EpubJan (10.1097/TA.0b013e3181c9ee58).

4. J. S. Harrop, G. M. Ghobrial, R. Chitale, K. Krespan, L. Odorizzi, T. Fried, M.

Maltenfort, M. Cohen, A. Vaccaro, Evaluating initial spine trauma response: injury

time to trauma center in PA, USA. Journal of clinical neuroscience : official journal

of the Neurosurgical Society of Australasia 21, 1725-1729 (2014); published online

EpubOct (10.1016/j.jocn.2014.03.011).

5. A. E. Ropper, M. T. Neal, N. Theodore, Acute management of traumatic cervical

spinal cord injury. Practical neurology 15, 266-272 (2015); published online

EpubAug (10.1136/practneurol-2015-001094).

6. M. N. Hadley, Cervical Spine Immobilization before Admission to the Hospital.

Neurosurgery 50, S7-S17 (2002).

239

7. H. J. Hachen, Emergency transportation in the event of acute spinal cord lesion.

Paraplegia 12, 33-37 (1974); published online EpubMay (10.1038/sc.1974.6).

8. G. A. Zach, W. Seiler, P. Dollfus, Treatment results of spinal cord injuries in the

Swiss Parplegic Centre of Basle. Paraplegia 14, 58-65 (1976); published online

EpubMay (10.1038/sc.1976.9).

9. A. E. Kerstetter, R. H. Miller, Isolation and culture of spinal cord astrocytes.

Methods in molecular biology 814, 93-104 (2012)10.1007/978-1-61779-452-0_7).

10. A. Tolonen, J. Turkka, O. Salonen, E. Ahoniemi, H. Alaranta, Traumatic brain injury

is under-diagnosed in patients with spinal cord injury. Journal of rehabilitation

medicine 39, 622-626 (2007); published online EpubOct (10.2340/16501977-

0101).

11. A. Mahmoud, S. Rengachary, R. Zafonte, Biomechanics of associated spine

injuries in head-injured patients. Topics in Spinal Cord Injury Rehabilitation 5, 41-

46 (1999).

12. E. Roth, G. Davidoff, P. Thomas, R. Doljanac, M. Dijkers, S. Berent, J. Morris, G.

Yarkony, A controlled study of neuropsychological deficits in acute spinal cord

injury patients. Paraplegia 27, 480-489 (1989); published online EpubDec

(10.1038/sc.1989.75).

13. J. N. K. Hsiang, T. Yueng, A. L. M. Yu, W. S. Poon, High-risk mild head injury.

Journal of Neurosurger 87, 234-238 (1997).

14. M. R. Garnett, A. M. Blamire, B. Rajagopalan, P. Styles, T. A. D. Cadoux-Hudson,

Evidence for cellular damage in normal-appearing white matter correlates with

240

injury severity in patients following traumatic brain injury: A magnetic resonance

spectroscopy study. Brain 123, 1403-1409 (2000).

15. G. Davidoff, E. Roth, J. Morris, J. Bleiberg, P. R. Meyer, Assessment of closed

head injury in trauma-related spinal cord injury. Paraplegia 24, 97-104 (1986).

16. G. Davidoff, P. Thomas, M. Johnson, S. Berent, M. Dijkers, R. Doljanac, Closed

head injury in acute traumatic spinal cord injury: incidence and risk factors.

Archives of physical medicine and rehabilitation 69, 869-872 (1988).

17. J. H. Lee, C. F. Jones, E. B. Okon, L. Anderson, S. Tigchelaar, P. Kooner, T.

Godbey, B. Chua, G. Gray, R. Hildebrandt, P. Cripton, W. Tetzlaff, B. K. Kwon, A

novel porcine model of traumatic thoracic spinal cord injury. Journal of

neurotrauma 30, 142-159 (2013); published online EpubFeb 1

(10.1089/neu.2012.2386).

18. F. Streijger, J. H. Lee, N. Manouchehri, A. D. Melnyk, J. Chak, S. Tigchelaar, K.

So, E. B. Okon, S. Jiang, R. Kinsler, K. Barazanji, P. A. Cripton, B. K. Kwon,

Responses of the Acutely Injured Spinal Cord to Vibration that Simulates Transport

in Helicopters or Mine-Resistant Ambush-Protected Vehicles. Journal of

neurotrauma, (2016); published online EpubJul 5 (10.1089/neu.2016.4456).

19. J. A. Colombo, A. Yanez, S. J. Lipina, Interlaminar astroglial processes in the

cerebral cortex of non human primates: response to injury. J Hirnforsch 38, (1997).

20. U. Ito, Y. Hakamata, E. Kawakami, K. Oyanagi, Degeneration of astrocytic

processes and their mitochondria in cerebral cortical regions peripheral to the

cortical infarction heterogeneity of their disintegration is closely associated with

241

disseminated selective neuronal necrosis and maturation of injury. Stroke 40,

2173-2181 (2009).

21. M. G. Salter, R. Fern, The mechanisms of acute ischemic injury in the cell

processes of developing white matter astrocytes. Journal of Cerebral Blood Flow

& Metabolism 28, 588-601 (2008).

242

CHAPTER 5: CHARACTERIZING THE PREFERENTIAL RELEASE OF PROTEIN

SUBPOPULATIONS BY INJURED ASTROCYTES

5.1 INTRODUCTION

Traumatic central nervous system (CNS) injury is caused by direct physical damage to the brain or spinal cord resulting diffuse axonal damage in addition to immediate hemorrhage and contusion at the site of injury. Traumatic brain injury (TBI) is perhaps the most common form of CNS damage that affects more than 57 million hospitalizations globally. Over 5 million people are estimated to live with TBI-related disabilities and is the most common cause of disability in individuals under 30 (1, 2).

Traumatic brain injury is of increasing concern to military personnel, emergency responders, and athletes with the leading causes from blast injury, violence, and falls.

Despite the immense health and financial costs associated with TBI, current evaluation of injury is limited by the insensitivity of standard neurocognitive assessments such as the Glasgow Coma Scale (GCS). These tests measure a patient’s level of consciousness and cognitive function and rely solely on verbal communication, motor function, and memory related responses. At best, improper assessment results from deficiencies in clinical expertise and at worst, these test may be subject to motivational confounds from business pressures. Better diagnostic measures, in the form of an easy to administer and decipher biofluid biomarker assay offer the potential of unambiguous identification of injury with the promise of quantitative and standardized severity categorization.

243

Despite advances in proteomic screening methods, no established biomarker for

TBI is currently used clinically in the US. Proteomic screening of injured CNS tissue and patient biofluids have identified a wealth of potential biomarkers for neurotrauma. The challenge, however, lies in the interpretation of these long candidate lists due to the complexity of events at and around a dynamically changing injury site and variations between trauma models (3-5). Tissue derived protein signals are products of a changing composition of viable, injured, and dead cells as well as infiltrating non-neural cells, all of which complicate the interpretation of proteomic studies (6, 7).

Due to the confounding complexity of clinical TBI and clinic-resembling animal injury models, we propose a targeted proteomic screen using a well-characterized in vitro cell-based trauma model as a starting point for TBI marker candidate identification (8-12).

This will limit protein changes to those directly related to an acute mechanical trauma by applying an abrupt pressure pulse inflicting shear forces and deformation onto cortical brain cells in a reproducible fashion at various severities. We identified robust cellular release patterns that correlate with cell injury and cell death and apply a suitable selection strategy that builds on our previous work in clinical samples toward the ultimate goal of a blood based diagnostic.

244

5.2 RESULTS

Astrocyte cultures derived from 3 different human fetal cerebral neocortex donations were subjected to biomechanical trauma from nitrogen pressure pulses to evaluate the impact of injury on released and intracellular proteomes. Additionally, the effect of injury severity was assessed using either mild or severe stretch parameters defined by the application of 50 ms of nitrogen gas flow at 2.6-4.0 or 4.4-5.3 psi respectively. The effect of injury on astrocyte release proteomes and intracellular proteomes were assessed by collection and trypsin digestion of conditioned medium (CM) and whole cell lysates (WCL). Relative quantitation between injured and un-stretched astrocytes was performed using isobaric labeling with TMT 6-plex mass tags (Figure 5.1) and LC-MS/MS. This approach allows for the simultaneous comparison of proteomes before and after treatment based on the ratios of the relative intensities of the differential reporter ions cleaved from the isobaric mass tags during higher-energy collisional dissociation (HCD). Because precursor co-isolation negatively impacts reporter ion quantification, trypsin digested samples were subjected to strong cation exchange (SCX) pre-fractionation. Pre-fractionation reduces the occurrence of precursor co-elution through improved chromatographic separation to improve the accuracy of quantitative results (13-15). Additionally, peptides with co-isolation percentages higher than 50% were also excluded from TMT ratio calculations (Methods).

Stretched astrocytes release different populations of injury-related proteins

TMT ratios for CM proteins released after exposure to mild or severe pressure stretch conditions were averaged to determine relative fold changes to control conditions

245

(un-stretched). Intensities from replicate analyses were evaluated by paired, 1-tailed t- test to determine statistically significant (p < 0.05) quantitative observations. A fold change (FC) of greater than 2 compared to control was selected for biological significance. These statistical and biological boundaries were selected to identify released proteins of interest from our injury model and visualized in the volcano plots presented in

Figure 5.2. For each time-point evaluated, severe stretching resulted in more abundant

(FC>3 and above) release of injury-related proteins compared to mild stretch conditions.

Difference in differentially released proteins were most apparent between severe and mild stretching at 5 and 24 hours (Figure 5.3). At 48 hours, minimal change in CM protein abundances were observed between mild and severe injury groups.

Distinct size profiles observed for injury released proteins in conditioned medium

Cellular membrane permeability is a hallmark of diffuse axonal injury in traumatic brain injury (TBI) (16, 17). This phenomenon has also been demonstrated by our group following biomechanical injury in vitro (Chapter 3). Building upon this, we examined the relationship between released protein molecular weights (MW) and mechanoporated astrocytes using our trauma model. Using the released protein selection criteria described above, the MW distributions of CM proteins with at least a 2-fold increase relative to control were compared. Figure 5.4A displays median MWs of differentially released

(FC>2) proteins for all experimental conditions along with 1.5 interquartile ranges.

Comparison of stretch severities at 5 and 24 hours denote lower median MWs and tighter size distributions in elevated CM proteins for astrocytes stretched with 4.0 psi or lower.

No real difference is observed at 48 hours between mild and severe injury groups,

246

consistent with minimal changes in release abundances examined previously. Looking deeper at the effect on stretching on the various sub-populations of differentially released proteins, an inverse relationship between CM abundance (measured by FC) and mean

MWs is observed (Figure 5.4B, Table 5.1) in the mildly stretched cohort. At both 5 and 24 hours, the mean MW of proteins with 3-fold CM increase are measured at 36.6 ± 26.67 kDa and 15.7 ± 7.95 kDa compared to 57.96 ± 74.5 kDa and 45.87 ± 58.4 kDa respectively. These observations are suggestive of the preferential release of lower MW proteins from injured astrocytes. Overall, released protein MWs rise with time after injury for all stretch conditions that likely relate to temporal changes in cell death dynamics.

Intracellular protein expression dynamics relatively unchanged following injury

Astrocyte whole cell lysates (WCL) corresponding to post-injury CM fluid samples were also harvested at 5, 24, and 48 hours post-stretch injury. Given prior evidence suggestive of protein leakage from injured cells, a corresponding change in intracellular protein concentrations was expected from cell lysates. However, in contrast to the significant changes to CM protein concentrations, little deviation from baseline protein levels was measured from TMT labeled WCL peptides. Even at more modest expression deviations of ± 1.25-fold relative to control, very few WCL proteins were statistically or biologically altered in our stretch trauma model (Figure 5.5).

Select proteins display corresponding expression and release trends

Looking further at the WCL proteins displaying differential expression (Table 5.2), thymosin beta-4, thymosin beta-10, and 14 kDa phosphohistidine phosphatase were

247

observed to exhibit a 25% reduction in intracellular protein concentration at 5 hours post stretch injury. All 3 proteins also exhibited a correspondingly high increase (2

CM protein concentration that is consistent with subpopulations of cells with membrane irregularities as demonstrated in Chapter 3. Peptidyl-prolyl cis-trans isomerase FKBP10, myosin regulatory light polypeptide 9, and reticulon-4 all showed a modest, but statistically significant (~1.25-fold) increase from baseline levels. These proteins exhibited around a 2-fold increase in CM concentration after injury with the exception of reticulon-

4 (FC ~ 1.25). The high MW of reticulon-4 is consistent with our hypothesis. At 24 hours, thymosin beta proteins and transgelin continue to be decreased in WCL but highly increased (FC>3) in CM. Adenylyl cyclase-associated protein 1 is decreased in WCL at

48h and very highly increased in CM. Ubiquitin-conjugating enzyme E2 N, calponin-3, actin-related protein-2, and plasminogen activator inhibitor 1 are 1.25-fold elevated in

WCL with very high CM elevations. This population of proteins exhibiting matching intracellular and released protein dynamics in response to injury may represent good biomarkers for biomechanical neurotrauma.

Identification of potential astrocyte injury protein signatures with respect to time and severity

A major goal of our astrocyte injury model is to identify highly abundant, ideally central nervous system (CNS) enriched protein biomarkers sensitive and specific to neurotraumatic injury. As part of this characterization, several stepwise comparisons were performed in an attempt to identify both differentiating (with respect to time and injury severity) and common trauma signatures. First, differentially released (FC>2) CM

248

proteins were compared at all time-points (5h, 24h, and 48h) with respect to our trauma severity (Figure 5.6). This uncovered a common core of proteins between mild and severe stretch released proteins as well as subgroups of proteins potentially specific to both severity of injury and time after injury. Because a strict fold change cut off of 2 was selected, these injury-time specific subpopulations were further curated manually to eliminate proteins that were also observed with approximate (FC slightly less than 2) abundance changes at other time-points (Table 5.3). A last round of selection was performed to identify proteins observed from both severities of stretch injury and at time points post-trauma (Figure 5.7).

Top neurotraumatic injury biomarker identification

Signature injury markers identified from our quantitative astrocyte trauma model were further narrowed using the filtering strategy presented in Figure 5.8. As described previously (Chapter 3), we established a human TBI cerebrospinal fluid (CSF) traumatome that included proteins observed in TBI CSF only as well as proteins measured in both 19 TBI and 9 healthy patients. Next, to reduce the contributions from non-CNS specific organ systems and toward the development of a blood based assay, proteins derived from blood (18, 19) were filtered. The final candidate list (Table 5.4) was graded on the following criteria: (1) Observed in human TBI CSF traumatome

(preferentially TBI CSF only), (2) observed at multiple time-points in in vitro trauma model,

(3) low MW, (4) demonstrates relationship between WCL and CM concentrations in response to injury, (5) not observed in blood proteome and (6) relative elevation across all stretch trauma conditions.

249

5.3 DISCUSSION

Application of quantitative proteomics to a well-defined in vitro injury mode was used to further characterize cellular changes with regard to both release and intracellular proteomes following traumatic injury in astrocytes. Our findings contribute to the continued elucidation of astrocyte response to injury with an emphasis on temporal dynamics and reported plasma membrane compromise as they relate to the identification of diagnostic protein signatures for traumatic CNS injury.

Mechanical injury may result in preferential release of lower molecular weight proteins

Pressure pulse (2.6 – 5.3 psi, 50 ms) induced stretch injury resulted in significant changes to the release proteomes of astrocytes in culture. Differential release of proteins was inferred from increase in protein abundances measured in conditioned media (CM) compared to healthy un-stretched cells at three time-points from 5 to 48 hours after injury.

Plasmalemma damage is a documented cellular response that occurs early after injury before the onset of other injury related sequelae (17, 20-22). Membrane irregularities coupled with the wide spread occurrence of apoptotic and neurotic cell death (23-26) after

TBI are responsible for the changes to protein abundance in astrocyte CM. Both our studies (Chapter 3) and published literature support the findings that the early sequelae after neurotraumatic insult is dominated by changes to the cellular integrity. The non- discriminant release of proteins from mechanoporated cells is most likely a function of cellular protein molecular weight. This is supported by our in vitro study which identified subpopulations of more differentially released proteins (FC>3) that occupy a much lower distribution of molecular weights (MWs) compared to the entire population of differentially

250

released proteins (defined as FC>2). This preferential release of lower MW protein species is most apparent under what we define as mild pressure stretching due to the lower extent of early cell death as previously described (Chapter 3).

Further evidence for preferential release of lower MW species is presented in the comparison of intracellular expression changes measured in WCL to corresponding increases in CM concentrations. Reticulon-4, also known as Nog-66 receptor 4, was observed to be 1.25-fold (p<0.05) elevated in whole cell lysates (WCLs) at 5 hours after injury but only mildly elevated in CM (FC ~1.3). In contrast to smaller proteins observed elevated in both WCL and CM at 5 hours, the modest increase in reticulon-4 CM abundance despite elevated cellular expression is likely a consequence of its larger (102 kDa) size. In stark contrast, 3 proteins, thymosin beta-4, thymosin beta-10, and transgelin

(MW range 5-23 kDa) were observed to be highly elevated (average FC>3) in CM in the acute time-point despite a decrease in intracellular concentrations. While it is unclear whether increased injury induces increased expression of these proteins, this evidence supports the notion that their low MW facilitates their cellular departure after trauma.

Mild trauma may generate comparable levels of cell death as severe trauma

Severe stretch trauma resulted in higher CM concentrations of released proteins.

This is believed to result from the complete release of cytosolic contents accompanying increased cell death (27, 28). While differences in highly differentially released proteins

(FC>3) were most apparent between 5 and 24 hours, differences between mild and severe trauma were minimized by 48 hours with near equal percentages of matching proteins between injury conditions exhibiting 3-fold or higher concentration increases in

251

CM. This indirect evidence suggests that the extent of secondary sequelae related to cell death processes may be comparable between mild and severe TBI, highlighting the health risk of these invisible wounds. Additionally, from a diagnostic standpoint, this finding points to the importance of time with respect to detection of injury as well as biofluid marker concentrations as they relate to patient prognoses. Time-related increases in cell death will greatly alter the effective concentrations of disease-associated proteins, necessitating differential acceptance ranges depending on when diagnostics were administered. It is important to note that what we have defined as mild and severe in our in vitro trauma system may not correlate with observed pathophysiological manifestations of clinical defined injury severity in patients.

New low molecular weight acute neurotraumatic injury signatures

Comparative analysis of release profiles after injury resulted in the identification of subsets of protein signatures that may potentially aid in the elucidation of the underlying molecular cascade of traumatic astrocyte injury. Stretching of astrocytes in vitro yielded increased release of ribosomal protein subunits, heat shock proteins, components of the ubiquitin-proteasome complex, and caspases associated with increases in protein synthesis, stress response (29-31), injury-related abnormal protein degradation (32, 33), and apoptosis (34). It is still unclear whether biomechanical trauma is responsible for the systematic alteration of pathways associated with disease despite the presentation of a host of molecular pathologies (35, 36). Consequently, it is possible that identified proteins with increased release profiles may be purely associated with increased membrane permeability. Because of this, preferential value was assigned to low MW proteins in our

252

selection of new neurotraumatic biomarkers. In addition to MW considerations, priority was given based on representation within our previously established human TBI CSF traumatome (Chapter 3) and proteins not normally present in blood circulation. However, representation in the blood protein did not necessarily exclude candidate proteins given the fact that well established TBI marker ubiquitin carboxyl-terminal hydrolase isozyme

L1 (UCHL1) (37) is naturally present in blood but with diagnostically relevant concentrations after injury (38). Aldolase C (39) and protein 14-3-3 (40) were also identified by our filtering strategy, further strengthening the validity of our approach as our group has characterized these two candidates in clinical samples. Of particular interest are new candidates from the thymosin family of proteins (5 kDa), 14 kDa phosphohistidine phosphatase (14 kDa), and transgelin (23 kDa) that exhibited corresponding decreases in intracellular protein concentrations after injury. This is strongly suggestive of their preferential leak into surrounding fluid. Ezrin, an actin-related protein, was another interesting hit from our screen that is reported to be present in high concentration extracellularly after TBI. Ezrin localization has been observed on astrocyte lamellipodia and around cellular debris suggesting dual roles in motility and immune recruitment after injury (41, 42). Superoxide dismutase (SOD1) presents another biological relevant candidate given the increases in brain reactive oxygen species (ROS) generation that accompanies mitochondrial dysfunction after trauma. Increased SOD1 expression is consistent with the neuroprotective process with previous studies demonstrating partial ablation of ischemia related symptoms with SOD1 overexpression or supplementation

(43, 44).

253

Conclusions

The proteomic response to both injury-induced astrocyte plasmalemma compromise and cell death have been evaluated in this work with strong implications for preferential release of low MW species after injury. This was observed to be especially apparent with regard to injury severity at acute time points within 24 hours of injury in vitro that is believed to be related to the early onset of membrane compromise but prior to the onset of secondary sequelae leading to widespread cell death. This early occurrence of membrane compromise is of special interest due to its diagnostic implications. Low MW biomarkers that are preferentially released as a result of cellular mechanoporation may capture the development of early traumatic sequelae specific for TBI in the hyper-acute post-injury period. Applications of this are of particular importance to healthcare for athletes where a rapid diagnostic tool may prevent incorrect medical clearance of players with diffuse axonal damage from a concussive injury. We have presented a manually curated list of low MW candidates that are robustly detected not only in vitro but also in clinical TBI patient CSF. Future verification in biological relevant injury systems will determine the utility of these candidates and whether protein size should be a consideration for neurotraumatic injury biomarkers.

254

5.4 METHODS

Fetal astrocyte culture and mechanical stretch injury

Primary human astrocytes were prepared as previously described (45). Donated human fetal cerebral neocortex at 16-19 gestational weeks was de-identified, cleaned and mechanically dissociated in calcium and magnesium-free Hank’s buffered saline solution (HBSS) before filtering through 70 µm and 10 µm nylon meshes (Nitex) in culture medium (DMEM-F12) with 10% fetal bovine serum (FBS). Astrocytes were separated from neural progenitor cells by 30min centrifugation at 30,000xg (J6B Sorvall centrifuge, rotor SA600) in a HBSS-buffered 33% Percoll gradient (Sigma). The top fraction was washed and diluted in DMEM/F12, 10% FBS and cultured in T150 cell culture-treated plastic flasks (Corning). Confluent cultures were shaken for 4 days at 200rpm on a shaker in an incubator. Astrocytes were mechanically dissociated following brief treatment in

0.25% trypsin/EDTA, washed, collected by centrifugation at 400 x g in a clinical centrifuge

(IEC) and seeded onto collagen I-coated silastic membrane culture plates (6 well Bioflex) at a density of ~ 135,000 human cells / 962mm2. Upon confluence, medium was replaced by DMEM/F12 with 10% heat-inactivated horse serum (Atlanta Biol.) that was subsequently stepwise reduced. Differentiated serum-free astrocytes in 2ml DMEM/F12 were stretch-injured using one mild (2.6-4.0psi) or one severe (4.4-5.3psi) 50ms nitrogen pressure pulse with the CIC II pressure controller (Custom Design and Fabrication Inc.).

TMT isobaric labeling

Stretched and un-stretched fetal human astrocytes lysates (WCL) and their conditioned media (CM) were labeled with isobaric TMT sixplex mass tags (Thermo).

255

Cultured astrocytes were stretched at various severities, described above and harvested along with their surrounding culture media for analysis at various time-points. Astrocyte

WCLs were precipitated using cold acetone and then re-suspended in 100 µL of 100 mM triethyl ammonium bicarbonate (TEAB), 0.1% deoxycholate. Samples were then reduced and alkylated with TCEP and IAM for 1 hour (55°C) and 30 minutes (37°C) respectively.

2.5 µg of trypsin per 100 µg of protein added for overnight digestion at 37°C. CM samples were measured by BCA and 100 µg of sample treated directly to reduction, alkylation, and trypsin digestion.

TMT isobaric label reagent sets were used per manufacturer’s instructions. Prior to use TMT label reagents were equilibrated to room temperature. 0.8 mg vials of each label (m/z 126-131) were reconstituted in 41 µL of anhydrous acetonitrile (ACN) and allowed to dissolve for 5 minutes with occasional vortexing. Entire aliquots of WCL or CM were added directly to TMT reagent vials. Labeling reaction was carried out at room temperature for 1 hour and then quenched with 8 µL of 5% hydroxylamine with 15 minute incubation. Equal volumes of sample were then combined for fractionation and analysis.

Replicates from three separate experiments were labeled with TMT mass tags as described in the table below (both WCL and CM). Samples were labeled according to

Table 5.5.

Offline SCX fractionation

TMT labeled samples were fractionated by C18/SCX spin-tips prepared in-house.

200 µL Eppendorf tips were packed (in order) with equal amounts of SCX and C18 packing (Empore). Spin-tips were conditioned sequentially with 100 µL of methanol, C18

256

elution buffer (80% ACN, 5% acetic acid (HOAc)), and loading buffer (3% ACN, 0.5%

HOAc), spinning down at 2K x g between solvents changes. 100 µL of SCX buffer (30%

ACN, 500 mM ammonium acetate, 0.5% HOAc) was added to condition SCX packing following by 100 µL of loading buffer to re-equilibrate. Up to 50 µg of sample was then added to the conditioned spin-tip with 100 µL loading buffer and spun down at 2K x g. 20

µL of C18 elution buffer was added to elute peptides onto SCX filter. Stepwise elution with increasing ammonium acetate (25, 50, 100, 200, and 500 mM) in 30% ACN, 0.5%

HOAc was performed to fractionate samples. Fractionated samples were dried by vacuum centrifugation and reconstituted with 3% ACN, 0.1% formic acid (FA) in water to concentration of 0.5 mg/mL.

Protein identification and TMT quantification by nano-LC-MS/MS

Fractionated peptides were injected onto an Acclaim PepMap 100, 75 µm X 2cm

C18 (Thermo) trap column and EASY-Spray PepMap RSLC, C18, 2µm, 75 µM X 25 cm analytical column (Thermo) attached to an EASY nLC 1000 (Proxeon). The flow rate of the mobile phase was set to 300 nL/min. Peptides were separated with a 0.1% FA in water (A) and 0.1% FA in ACN (B) mobile phase system as follows: 5-35% B over 90 minutes, 35-60% B over 30 minutes. Peptides were introduced from the nano-HPLC to

Q-Exactive (Thermo), an Orbitrap mass spectrometer, operating with a Top10 duty cycle consisting of 1 full scan (70,000 resolution, AGC 1e6, 100 ms max IT, 200-2000 m/z) followed by 10 consecutive data-dependent MS2 (HCD) acquisitions (17,500 resolution,

AGC 1e5, 100 ms max IT, 4 m/z isolation window, 100 m/z fixed first mass, and NCE 30).

257

A dynamic exclusion of 30 secs was applied and peptides with unassigned charge or charge state 1 were excluded from MS2 analysis.

Raw data files were searched in Proteome Discoverer v1.4 (Thermo) configured with MASCOT (Matrix Science) to identify and quantify proteins based on mass, peptide spectral matches (PSM), and reporter ion intensities. Peptide mass data was matched against the human SwissProt database with the following search parameters: Enzyme – trypsin, 2 missed cleavages allowed, 10 ppm and 0.05 Da MS1 and MS2 mass tolerances, static modifications – carbamidomethyl (C), and dynamic modifications –

TMT6plex (K, N-term), oxidation (M). Protein identifications were validated by searching against a reverse sequence decoy database with a FDR of 0.05 and a minimum of 2 unique peptides. Common contaminant proteins were manually excised from protein ID lists (46). Relative quantitation of TMT reported ions were performed off the most confident centroid peak with 20 ppm mass tolerance. Precursor ions with high co-isolation interference were excluded from ratio calculations in Proteome Discoverer based on the following formula:

푝푟푒푐푢푟푠표푟 푖푛푡푒푛푠푖푡푦 푖푛 푖푠표푙푎푡푖표푛 푤푖푛푑표푤 % Isolation Interference = 100 x [1-( )]. 푡표푡푎푙 푖푛푡푒푛푠푖푡푦 푖푛 푖푠표푙푎푡푖표푛 푤푖푛푑표푤

Statistical analysis was performed using a combination of Excel (Microsoft) and Graph

Pad 7 (Prism).

258

5.5 FIGURES

Figure 5.1: Relative quantitation by TMT isobaric mass tags workflow

259

260

Figure 5.2: Astrocyte-released protein fold changes measured in conditioned media at after injury

Conditioned media (CM) was collected pre-injury and at 5 (A), 24 (B), and 48 (C) hours

(h) post mild (left) or severe (right) stretch injury. Astrocyte released proteins from different conditions were cleaved with trypsin and labeled with isobaric TMT mass tags. Dotted vertical and horizontal lines designate boundaries for measured 2-fold CM concentrations compared to control and statistical significance (p-value <0.05) of change calculated by paired, 1-tailed t-test. Proteins with fold change (FC) >2, >3, and >4 are represented in yellow, orange, and red respectively. Higher severity mechanical injury induces higher subpopulations of proteins released with respect to FC.

261

CM Fold Change 100 FC>2 80 FC>3 FC>4 60

% Proteins 20

5h Mild 24h Mild 48h Mild 5h Severe 24h Severe 48h Severe

Figure 5.3: Higher intensity pressure pulse stretching induces greater release of injury-related proteins from astrocytes

The percentage of proteins with relative fold change (FC) increases in condition medium

(CM) relative to un-stretched cells are represented in bar graph format. FCs of >2, >3, and >4 measured by TMT labeling are represented in yellow, orange, and red respectively for CM collected at 5, 24, and 48 hours (h) following mild or severe pressure stretching.

262

Figure 5.4: Stretch injury may induce preferential release of lower molecular weight protein species

Molecular weight (MW) analysis of astrocyte-released proteins in conditioned media (CM) at various 5, 24, and 48 hours (h) after mild (M) or severe (S) stretch injury. (A) Box whisker plots displaying median MWs with Tukey (1.5 interquartile ranges) whiskers of 2- fold differentially released astrocyte proteins in CM for each experimental group. From 5h to 24h, higher stretch severity induces release proteins exhibiting a higher MW range with little difference in protein sizes by 48 hours with regard to extent of injury. (B) Bar graphs displaying mean MWs of proteins released with fold changes (FCs) of >2 and >3 at 5h,

24h, and 48h collection time-points post mild mechanical injury. Error bars display full range of released protein MWs. A general trend is observed that possibly highlights the preferential release of lower MW proteins in response to injury.

263

264

Figure 5.5: Little change observed in intracellular protein concentrations in response to injury from astrocyte whole cell lysates

Whole cell lysate (WCL) was collected pre-injury and after mild (left) and severe (right) stretch injury at 5 hours (h) (A), 24h (B), and 48h (C). Intracellular proteins were cleaved with trypsin and labeled with isobaric TMT mass tags. Dotted vertical and horizontal lines designate boundaries for measured fold changes (FC) of <0.5, 0<0.75, >1.25, and >2 (left to right). WCL concentrations were compared to control and statistical significance (p- value <0.05) of change calculated by paired, 1-tailed t-test. Proteins with FC <0.75 and

>1.25 are represented in light blue and yellow respectively. Minimal protein concentration changes within cells were measured in response to injury.

265

Figure 5.6: Comparison of proteomic release profiles resulting from astrocytes injured by mild and severe stretching

Comparative analysis of conditioned medium (CM) proteins with 2-fold elevation relative to control after mild or severe stretch injury at 5, 24, and 48 hour (h) collection time-points after injury. Distinct injury severity related protein profiles are observed at 5h (A), 24h (B), and 48h (C).

266

Figure 5.7: A common subpopulation of injury released proteins may represent strong candidates of neurotraumatic injury

Proteins observed in both mild and severe injury groups at each time-point (5h, 24h, and

48h, Figure 5.6) were compared against each other. The 32 proteins overlapping between all three groups represent the most robustly observed injury released proteins in astrocyte conditioned media. This list of 32 proteins were further filter to arrive at final candidate lists.

267

Figure 5.8: Astrocyte injury related release biomarker filtering strategy

Differentially (FC>2) released injury proteins identified by relative quantitation with TMT isobaric mass labels were filtered using the above scheme. Severity/Time specific proteins were manually curated based on exclusivity to time or injury conditions. Proteins were then compared to the UCLA human traumatic brain injury (TBI) cerebrospinal fluid

(CSF) proteome (Chapter 3) in an attempt to identify potentially robust in vivo candidates.

Finally a blood protein filter was applied based on published blood proteomes (18, 19).

Manual addition of select proteins was also added to our candidate lists (Table 5.3) as described in the Results and Discussion sections.

268

5.6 TABLES

5h Mild 24h Mild 48h Mild MW (kDa) FC>2 FC>3 FC>2 FC>3 FC>2 FC>3 Minimum 5.05 5.05 5.05 5.05 5.05 5.05 25% Percentile 22.6 17.93 18 9.086 27.97 26.79 Median 36.67 28.75 27.73 14.87 45.97 42.59 75% Percentile 59.81 47.5 47.14 23.06 72.72 70.39 Maximum 628.7 123.5 284.4 27.37 628.7 628.7 Mean 57.96 36.6 45.87 15.7 70.29 66.54 Std. Deviation 74.5 26.67 58.35 7.95 81.11 78.75 n 235 53 51 6 365 325

5h Severe 24h Severe 48h Severe MW (kDa) FC>2 FC>3 FC>2 FC>3 FC>2 FC>3 Minimum 5.05 5.05 5.05 5.05 5.05 5.05 25% Percentile 22.81 20.32 23.73 22.81 27.91 27.42 Median 38.57 31.33 40.4 39.4 45.97 44.72 75% Percentile 61.58 51.61 65.29 68.6 73.58 73.21 Maximum 628.7 628.7 468.5 468.5 628.7 628.7 Mean 57.37 52.11 61.79 65.27 70.72 68.53 Std. Deviation 73.34 78.23 66.98 75.62 81.52 78.53 n 236 125 199 129 361 340

Table 5.1: Differentially released CM protein molecular weight statistics

269

Differentially Expressed WCL Proteins (5h Mild) Accession Description MW [kDa] P63313 Thymosin beta-10 5.0 P62328 Thymosin beta-4 5.0 Q96AY3 Peptidyl-prolyl cis-trans isomerase FKBP10 64.2 Q9NQC3 Reticulon-4 129.9 P05121 Plasminogen activator inhibitor 1 45.0 Q9UDY4 DnaJ homolog subfamily B member 4 37.8 P24844 Myosin regulatory light polypeptide 9 19.8

Differentially Expressed WCL Proteins (5h Severe) Accession Description MW [kDa] Q9NRX4 14 kDa phosphohistidine phosphatase 13.8

Differentially Expressed WCL Proteins (24h Mild) Accession Description MW [kDa] P63313 Thymosin beta-10 5.0 P62328 Thymosin beta-4 5.0 Q01995 Transgelin 22.6 Q8NBS9 Thioredoxin domain-containing protein 5 47.6 P62266 40S ribosomal protein S23 15.8 Q7KZF4 Staphylococcal nuclease domain-containing protein 1 101.9 P50454 Serpin H1 46.4

Differentially Expressed WCL Proteins (24h Severe) Accession Description MW [kDa] P62328 Thymosin beta-4 5.0 Q01995 Transgelin 22.6 Q8NBS9 Thioredoxin domain-containing protein 5 47.6

Differentially Expressed WCL Proteins (48h Severe) Accession Description MW (kDa) P61088 Ubiquitin-conjugating enzyme E2 N 17.1 P05121 Plasminogen activator inhibitor 1 45.0 P35268 60S ribosomal protein L22 14.8 Q9Y2B0 Protein canopy homolog 2 20.6 P62424 60S ribosomal protein L7a 30.0

Differentially Expressed WCL Proteins (48h Severe) Accession Description MW (kDa) Q01518 Adenylyl cyclase-associated protein 1 51.9 P61088 Ubiquitin-conjugating enzyme E2 N 17.1 Q99439 Calponin-2 33.7 P61160 Actin-related protein 2 44.7 P05121 Plasminogen activator inhibitor 1 45.0 P35268 60S ribosomal protein L22 14.8 Q9Y2B0 Protein canopy homolog 2 20.6

270

Table 5.2: Differentially expressed whole cell lysate proteins at post-injury conditions

Whole cell lysate proteins with expression differences at 5, 24, and 48 hours (h) after mild or severe stretch injury. Proteins with a 25% reduction in intracellular concentrations are displayed in italics. Remaining proteins represent intracellular increases of 25%. Bolded proteins exhibited corresponding increases in time-point and stretch severity matched conditioned medium.

271

Unique 5h Protein Signatures Accession Description 5hr 5h Severe MW(kDa) P68402 Platelet-activating factor Mild2.7 25.6 P61247 40Sacetylhydrolase ribosomal protein IB subunit S3a 2.5 29.9 Q13643 Four and a halfbeta LIM domains 2.2 31.2 P20930 proteFilaggrinin 3 2.2 434.9 P15121 Aldose reductase 2.2 35.8 P62841 40S ribosomal protein S15 2.2 17.0 Q14315 Filamin-C 2.1 290.8 P29692 Elongation factor 1-delta 2.1 31.1 P35237 Serpin B6 2.1 42.6 P17174 Aspartate aminotransferase, 2.0 46.2 P05387 60S acidiccytoplasmic ribosomal protein 2.0 11.7 Q9BWD1 Acetyl-CoA acetyltransferase,P2 4.8 41.3 O95782 AP-2 complexcytosolic subunit alpha-1 3.4 107.5 P28300 Protein-lysine 6-oxidase 2.3 46.9

Unique 24h Protein Signatures Accession Description 24h 24h Severe MW(kDa) P12109 Collagen alpha-1(VI) chain Mild 6.3 108.5 P24821 Tenascin 6.1 240.7 Q4ZHG4 Fibronectin type III domain- 5.7 205.4 Q15121 Astrocyticcontaining phosp proteinhoprotein 1 5.6 15.0 P30101 Protein disulfidePEA-15isomerase A3 4.5 56.7 Q9H4D0 Calsyntenin-2 4.5 106.9 P05997 Collagen alpha-2(V) chain 4.4 144.8 Q8NBS9 Thioredoxin domain-containing 4.4 47.6 P39687 Acidic leucineprotein-rich 5 nuclear 4.3 28.6 P01008 phosphoproteinAntithrombin 32-III family 4.3 52.6 Q12805 EGF-containingmember fibulinA -like 4.3 54.6 P07237 extracellularProtein disulfide matrix-isomerase protein 1 4.2 57.1 Q92626 Peroxidasin homolog 4.2 165.2 P49327 Fatty acid synthase 4.0 273.3 P28838 Cytosol aminopeptidase 3.8 56.1 P35555 Fibrillin-1 3.8 312.0 P01033 Metalloproteinase inhibitor 1 3.8 23.2 Q969H8 UPF0556 protein C19orf10 3.8 18.8 P09972 Fructose-bisphosphate 3.8 39.4 Q02818 Nucleobindinaldolase C - 1 3.8 53.8 P08572 Collagen alpha-2(IV) chain 3.5 167.4 P18065 Insulin-like growth factor- 3.4 34.8 P16035 Metalloproteinasebinding protein inhibitor 2 2 3.4 24.4 P05204 Non-histone chromosomal 3.3 9.4 P07996 Thrombospondinprotein HMG-17- 1 3.3 129.3 P98160 Basement membrane-specific 3.3 468.5 O94985 heparanCalsyntenin sulfate proteoglycan-1 3.3 109.7 Q9UI42 Carboxypeptidasecore protein A4 3.3 47.3 Q15063 Periostin 3.3 93.3 Q15582 Transforming growth factor- 3.3 74.6 Q9UBP4 betaDickkopf-induced-related protein protein ig-h3 3 3.2 38.4 P05121 Plasminogen activator inhibitor 3.2 45.0 P19827 Inter-alpha-trypsin1 inhibitor 3.1 101.3 P01709 Ig lambdaheavy chain chain V -H1II region 3.0 11.6 P21810 BiglycanMGC 3.0 41.6 Q14767 Latent-transforming growth 2.9 194.9 P12107 factorCollagen beta alpha-binding-1(XI) protein chain 2 2.8 181.0 272

P23284 Peptidyl-prolyl cis-trans 2.8 23.7 O14818 Proteasomeisomerase subunit B alpha 2.8 27.9 P24593 Insulin-liketype growth-7 factor- 2.7 30.6 P02461 Collagenbinding alpha protein-1(III) 5 chain 2.5 138.5 Q08380 Galectin-3-binding protein 2.5 65.3 Q16270 Insulin-like growth factor- 2.4 29.1 P62906 60S ribosomalbinding protein protein 7 L10a 2.4 24.8 Q14766 Latent-transforming growth 2.3 186.7 Q96HC4 PDZfactor and beta LIM-binding domain protein protein 1 5 2.0 63.9

Unique 48h Protein Signatures Accession Description 48hr 48h Severe MW(kDa) Q14697 Neutral alpha-glucosidase AB Mild3.8 106.8 O00410 Importin-5 7.9 123.5 Q01105 Protein SET 5.1 33.5 P26639 Threonine--tRNA ligase, 3.3 83.4 cytoplasmic

Table 5.3: Potential time and injury severity specific signatures

Unique released protein signatures identified from workflows presented in Figure 5.6 and

5.7. Results of comparative analysis were manually filtered to include for not only injury related release proteins observed only under specified time and trauma levels but also proteins that were measured in only low abundances (~2-fold less) at other time-points.

For 24 hour unique protein signatures, the presence of high 48 hour concentrations in CM was ignored given increase in associated cell death resulting in high overall released protein abundances.

273

Top Neurotraumatic Injury Biomarker Candidates 5hr 5h 24h 24h 48h 48h MW Accession Description M S M S M S (kDa) P09493 Tropomyosin alpha-1 chain 2.5 3.1 2.1 6.8 5.3 7.6 32.7 P62328/ Thymosin beta-4 / P63313 Thymosin beta-10 4.4 5.0 3.5 5.4 8.4 11.9 5.0 P62937 Peptidyl-prolyl cis-trans isomerase A 3.5 3.8 2.0 2.5 5.1 7.1 18.0 P00441 Superoxide dismutase [Cu-Zn] 2.6 3.0 3.8 4.8 4.9 6.3 15.9 Ubiquitin carboxyl-terminal hydrolase P09936 isozyme L1 3.7 3.9 2.6 3.8 4.8 6.7 24.8 P63104 14-3-3 protein zeta/delta 2.8 2.7 2.2 3.5 4.4 6.1 27.7 P09972 Fructose-bisphosphate aldolase C 2.4 2.4 1.9 3.8 3.6 4.9 39.4 Q01995 Transgelin 3.5 3.9 2.8 3.4 6.3 9.1 22.6 14 kDa phosphohistidine Q9NRX4 phosphatase 2.4 2.3 3.8 3.1 3.6 5.1 13.8 P62158 Calmodulin 2.2 3.5 2.4 4.8 5.1 8.2 16.8 P07951 Tropomyosin beta chain 2.1 2.6 2.3 3.6 5.1 7.1 32.8 Q15121 Astrocytic phosphoprotein PEA-15 2.8 3.6 1.8 5.6 5.8 8.5 15.0 P15311 Ezrin 2.6 3.7 2.9 3.1 4.8 8.3 69.4

Table 5.4: Top protein biomarker candidates for neurotraumatic injury

List of top neurotraumatic injury markers displaying proteomic fold changes relative to control in conditioned medium (CM) at 5, 24, and 48 hours (h) after mild (M) or severe (S) stretch injury. This list was manually curated based on whole cell lysate (WCL) and CM protein dynamics in a stretch injury astrocyte culture system and filtering strategy against human traumatic brain injury (TBI) cerebrospinal fluid (CSF) proteomes and blood proteomes (Figure 5.8). Proteins highlighted in red and orange represent proteins observed in TBI CSF and both TBI and healthy CSF respectively. Italicized entries represent proteins observed in the blood proteome (18, 19, 47). Bolded proteins represent proteins with corresponding WCL and CM protein abundance changes in response to stretch injury in vitro. Because many promising TBI biomarkers are also present normally in blood (ubiquitin carboxyl-terminal hydrolase isozyme L1, fructose-bisphosphate

274

aldolase C), overlap with blood proteome lists did not exclude inclusion in our candidate list.

275

Experiment 1 (Internal ID TMT3) Replicate Injury Severity Time (h) TMT Label A Un-stretched 0 126 B Un-stretched 0 127 A Mild 5 128 B Mild 5 129 A Severe 5 130 B Severe 5 131

Experiment 2 (Internal ID TMT4) Replicate Injury Severity Time (h) TMT Label A Un-stretched 0 126 B Un-stretched 0 127 A Mild 24 128 B Mild 24 129 A Severe 24 130 B Severe 24 131

Experiment 3 (Internal ID TMT5) Replicate Injury Severity Time (h) TMT Label A Un-stretched 0 126 A Mild 5 128 B Severe 5 129 A Mild 48 130 B Severe 48 131

Table 5.5: In vitro astrocyte WCL and CM TMT labeling table

276

# MW Accession Description PSMs 5h M 5h S Peptides [kDa] P07437 Tubulin beta chain 21 285 6.5 9.9 49.6 Q71U36 Tubulin alpha-1A chain 22 231 6.0 7.8 50.1 O60664 Perilipin-3 2 2 4.9 6.5 47.0 Nicotinamide N- P40261 8 41 4.2 6.1 29.6 methyltransferase Q9NTK5 Obg-like ATPase 1 3 6 3.5 5.8 44.7 P06703 Protein S100-A6 3 142 4.5 5.5 10.2 O00410 Importin-5 8 14 4.1 5.4 123.5 Glyceraldehyde-3-phosphate P04406 20 1264 4.2 5.4 36.0 dehydrogenase P30044 Peroxiredoxin-5, mitochondrial 4 10 3.7 5.1 22.1 P62328 Thymosin beta-4 4 14 4.4 5.0 5.0 P62857 40S ribosomal protein S28 2 2 3.9 5.0 7.8 P61970 Nuclear transport factor 2 4 19 3.7 5.0 14.5 Glutathione S-transferase P78417 5 9 3.6 5.0 27.5 omega-1 P04792 Heat shock protein beta-1 14 46 3.1 4.9 22.8 P58546 Myotrophin 3 6 3.5 4.9 12.9 Isocitrate dehydrogenase O75874 10 35 3.4 4.8 46.6 [NADP] cytoplasmic P31946 14-3-3 protein beta/alpha 17 288 3.6 4.8 28.1 Branched-chain-amino-acid P54687 3 3 3.4 4.8 42.9 aminotransferase, cytosolic Acetyl-CoA acetyltransferase, Q9BWD1 2 15 4.8 41.3 cytosolic Glyoxalase domain-containing Q9HC38 2 3 3.4 4.8 34.8 protein 4 40S ribosomal protein S4, X P62701 3 6 3.3 4.6 29.6 isoform Ubiquitin-conjugating enzyme E2 P68036 6 10 3.2 4.6 17.9 L3 P20962 Parathymosin 2 5 2.7 4.6 11.5 Q04917 14-3-3 protein eta 9 180 3.2 4.6 28.2 Protein phosphatase 1 regulatory Q96C90 2 2 4.5 15.9 subunit 14B Q9UGI8 Testin 3 4 3.5 4.5 48.0 Puromycin-sensitive P55786 16 28 3.1 4.4 103.2 aminopeptidase P30085 UMP-CMP kinase 6 31 3.4 4.4 22.2 P12277 Creatine kinase B-type 4 15 2.9 4.4 42.6 P30041 Peroxiredoxin-6 6 31 2.8 4.3 25.0 Serine/threonine-protein phosphatase 2A 65 kDa P30153 8 19 3.1 4.3 65.3 regulatory subunit A alpha isoform P08729 Keratin, type II cytoskeletal 7 14 48 3.0 4.2 51.4 277

Q14847 LIM and SH3 domain protein 1 9 16 2.9 4.1 29.7 Ubiquitin-conjugating enzyme E2 Q15819 2 8 3.2 4.1 16.4 variant 2 P05783 Keratin, type I cytoskeletal 18 9 26 2.9 4.1 48.0 F-actin-capping protein subunit P47756 6 20 3.2 4.1 31.3 beta Four and a half LIM domains Q13642 11 45 3.1 4.1 36.2 protein 1 P17655 Calpain-2 catalytic subunit 6 13 3.3 4.0 79.9 P27348 14-3-3 protein theta 17 216 3.2 4.0 27.7 P68363 Tubulin alpha-1B chain 22 234 3.2 4.0 50.1 Microtubule-associated protein P46821 12 19 2.9 3.9 270.5 1B Q15843 NEDD8 2 2 2.9 3.9 9.1 P39019 40S ribosomal protein S19 4 6 2.8 3.9 16.1 P02511 Alpha-crystallin B chain 5 29 2.5 3.9 20.1 Q01995 Transgelin 21 1730 3.5 3.9 22.6 Ubiquitin carboxyl-terminal P09936 15 437 3.7 3.9 24.8 hydrolase isozyme L1 Myosin light chain kinase, Q15746 2 3 2.9 3.9 210.6 smooth muscle Q53FA7 Quinone oxidoreductase PIG3 2 2 2.8 3.9 35.5 P20618 Proteasome subunit beta type-1 4 10 2.9 3.8 26.5 P18085 ADP-ribosylation factor 4 3 8 2.7 3.8 20.5 Peptidyl-prolyl cis-trans P62937 19 333 3.5 3.8 18.0 isomerase A P61204 ADP-ribosylation factor 3 4 10 3.0 3.8 20.6 Alcohol dehydrogenase P14550 9 27 3.0 3.8 36.5 [NADP(+)] Staphylococcal nuclease Q7KZF4 6 15 2.8 3.8 101.9 domain-containing protein 1 Q14019 Coactosin-like protein 7 20 3.0 3.8 15.9 Protein phosphatase Q9Y570 2 6 2.9 3.7 42.3 methylesterase 1 Phosphoacetylglucosamine O95394 5 14 2.9 3.7 59.8 mutase Farnesyl pyrophosphate P14324 3 3 3.0 3.7 48.2 synthase Q9H4A4 Aminopeptidase B 4 9 2.9 3.7 72.5 6-phosphogluconate P52209 15 46 2.8 3.7 53.1 dehydrogenase, decarboxylating Translationally-controlled tumor P13693 6 54 3.3 3.7 19.6 protein P15311 Ezrin 22 134 2.6 3.7 69.4 Astrocytic phosphoprotein PEA- Q15121 5 102 2.8 3.6 15.0 15 P34932 Heat shock 70 kDa protein 4 9 16 2.9 3.6 94.3 P23528 Cofilin-1 15 439 3.0 3.6 18.5

278

Ubiquitin-like modifier-activating P22314 15 29 3.7 3.6 117.8 enzyme 1 P61160 Actin-related protein 2 8 37 2.6 3.6 44.7 Q15404 Ras suppressor protein 1 7 17 2.9 3.6 31.5 P28066 Proteasome subunit alpha type-5 2 2 2.9 3.5 26.4 P61960 Ubiquitin-fold modifier 1 2 2 2.8 3.5 9.1 Q15942 Zyxin 6 10 2.7 3.5 61.2 Q12765 Secernin-1 4 7 2.7 3.5 46.4 Ubiquitin-40S ribosomal protein P62979 11 140 2.6 3.5 18.0 S27a P22392 Nucleoside diphosphate kinase B 8 90 3.0 3.5 17.3 P60981 Destrin 10 207 2.9 3.5 18.5 P60900 Proteasome subunit alpha type-6 3 3 2.6 3.5 27.4 P62158 Calmodulin 3 10 2.2 3.5 16.8 Heterogeneous nuclear P61978 2 2 2.4 3.5 50.9 ribonucleoprotein K Fatty acid-binding protein, Q01469 3 8 2.2 3.5 15.2 epidermal P52565 Rho GDP-dissociation inhibitor 1 3 10 3.3 3.4 23.2 Q14974 Importin subunit beta-1 12 23 2.6 3.4 97.1 Q99584 Protein S100-A13 3 5 2.4 3.4 11.5 S-phase kinase-associated P63208 3 3 2.7 3.4 18.6 protein 1 P10599 Thioredoxin 7 28 2.7 3.4 11.7 Q96FW1 Ubiquitin thioesterase OTUB1 3 5 2.7 3.4 31.3 O95782 AP-2 complex subunit alpha-1 2 2 3.4 107.5 Q06830 Peroxiredoxin-1 10 38 2.7 3.4 22.1 Ras GTPase-activating-like P46940 21 54 2.7 3.4 189.1 protein IQGAP1 P12955 Xaa-Pro dipeptidase 2 2 2.5 3.4 54.5 Q9NVA2 Septin-11 4 8 2.5 3.4 49.4 P53396 ATP-citrate synthase 10 20 2.8 3.4 120.8 Programmed cell death 6- Q8WUM4 3 3 2.2 3.3 96.0 interacting protein P11766 Alcohol dehydrogenase class-3 3 3 2.7 3.3 39.7 P36871 Phosphoglucomutase-1 11 34 2.7 3.3 61.4 P37837 Transaldolase 5 16 2.4 3.3 37.5 SH3 domain-binding glutamic O75368 5 13 2.8 3.3 12.8 acid-rich-like protein Serine/arginine-rich splicing Q07955 2 32 2.2 3.3 27.7 factor 1 P06733 Alpha-enolase 29 1290 3.0 3.3 47.1 Q99497 Protein DJ-1 9 75 2.7 3.3 19.9 P68371 Tubulin beta-4B chain 19 255 2.5 3.3 49.8 UTP--glucose-1-phosphate Q16851 5 11 2.5 3.3 56.9 uridylyltransferase 279

Macrophage migration inhibitory P14174 4 49 2.9 3.3 12.5 factor Myristoylated alanine-rich C- P29966 4 4 2.6 3.2 31.5 kinase substrate P32119 Peroxiredoxin-2 5 15 2.3 3.2 21.9 P08107 Heat shock 70 kDa protein 1A/1B 9 20 2.4 3.2 70.0 ATP-dependent 6- Q01813 phosphofructokinase, platelet 3 4 2.4 3.2 85.5 type Neuroblast differentiation- Q09666 9 31 2.4 3.2 628.7 associated protein AHNAK P00558 Phosphoglycerate kinase 1 28 267 2.7 3.2 44.6 P21266 Glutathione S-transferase Mu 3 9 31 2.9 3.2 26.5 P62241 40S ribosomal protein S8 3 8 2.5 3.2 24.2 Adenylyl cyclase-associated Q01518 17 74 2.5 3.2 51.9 protein 1 Q9ULV4 Coronin-1C 8 26 2.5 3.1 53.2 P67936 Tropomyosin alpha-4 chain 16 130 2.7 3.1 28.5 P37802 Transgelin-2 15 255 2.7 3.1 22.4 P14618 Pyruvate kinase PKM 37 1061 2.8 3.1 57.9 P09382 Galectin-1 9 217 3.1 3.1 14.7 P25786 Proteasome subunit alpha type-1 5 8 2.2 3.1 29.5 P09493 Tropomyosin alpha-1 chain 15 126 2.5 3.1 32.7 Cysteine and glycine-rich protein P21291 11 94 2.6 3.1 20.6 1 P09211 Glutathione S-transferase P 11 135 2.9 3.1 23.3 Cytoplasmic dynein 1 heavy Q14204 9 12 2.8 3.1 532.1 chain 1 SH3 domain-binding glutamic Q9H299 5 29 3.3 3.0 10.4 acid-rich-like protein 3 Cysteine and glycine-rich protein Q16527 6 19 2.5 3.0 20.9 2 P13639 Elongation factor 2 21 92 2.6 3.0 95.3 P09960 Leukotriene A-4 hydrolase 14 35 2.5 3.0 69.2 Chloride intracellular channel Q9Y696 16 251 3.6 3.0 28.8 protein 4 Dihydropyrimidinase-related Q16555 8 13 2.4 3.0 62.3 protein 2 BTB/POZ domain-containing Q96CX2 3 5 2.3 3.0 35.7 protein KCTD12 P26641 Elongation factor 1-gamma 8 17 2.6 3.0 50.1 Q92820 Gamma-glutamyl hydrolase 4 11 2.6 3.0 35.9 P00441 Superoxide dismutase [Cu-Zn] 8 65 2.6 3.0 15.9 P06744 Glucose-6-phosphate isomerase 17 127 2.6 3.0 63.1 Four and a half LIM domains Q14192 7 50 3.8 3.0 32.2 protein 2 P07195 L-lactate dehydrogenase B chain 16 157 2.5 3.0 36.6

280

1,4-alpha-glucan-branching Q04446 5 6 2.5 2.9 80.4 enzyme Q9Y617 Phosphoserine aminotransferase 14 138 2.6 2.9 40.4 O75083 WD repeat-containing protein 1 27 126 2.6 2.9 66.2 P07737 Profilin-1 10 284 2.7 2.9 15.0 P80723 Brain acid soluble protein 1 5 7 2.4 2.9 22.7 P26038 Moesin 31 292 2.6 2.9 67.8 Q05682 Caldesmon 16 95 2.8 2.9 93.2 P68032 Actin, alpha cardiac muscle 1 20 1041 2.7 2.9 42.0 Glycogen phosphorylase, brain P11216 4 5 2.4 2.8 96.6 form Chloride intracellular channel O00299 10 34 2.5 2.8 26.9 protein 1 P26022 Pentraxin-related protein PTX3 9 34 2.2 2.8 41.9 P00338 L-lactate dehydrogenase A chain 29 452 2.7 2.8 36.7 P62258 14-3-3 protein epsilon 15 226 2.2 2.8 29.2 P12814 Alpha-actinin-1 57 1326 2.5 2.8 103.0 P07602 Prosaposin 8 71 2.4 2.8 58.1 P00966 Argininosuccinate synthase 13 64 2.6 2.8 46.5 Rab GDP dissociation inhibitor P50395 21 235 2.4 2.8 50.6 beta Threonine--tRNA ligase, P26639 6 8 2.2 2.8 83.4 cytoplasmic P62826 GTP-binding nuclear protein Ran 7 25 2.3 2.8 24.4 Q16658 Fascin 16 99 3.5 2.8 54.5 P63104 14-3-3 protein zeta/delta 20 358 2.8 2.7 27.7 P60842 Eukaryotic initiation factor 4A-I 11 30 3.1 2.7 46.1 P18669 Phosphoglycerate mutase 1 16 273 2.5 2.7 28.8 Cyclin-dependent kinase inhibitor P42771 2 2 2.1 2.7 16.5 2A, isoforms 1/2/3 P13645 Keratin, type I cytoskeletal 10 9 12 2.1 2.7 58.8 Protein-glutamine gamma- P21980 6 14 3.0 2.7 77.3 glutamyltransferase 2 P68104 Elongation factor 1-alpha 1 15 144 2.4 2.7 50.1 P48163 NADP-dependent malic enzyme 4 5 2.7 64.1 Heat shock cognate 71 kDa P11142 30 304 2.6 2.7 70.9 protein Q9UBG0 C-type mannose receptor 2 7 11 2.4 2.7 166.6 P07858 Cathepsin B 3 5 2.2 2.7 37.8 P18206 Vinculin 62 559 2.5 2.7 123.7 Fructose-bisphosphate aldolase P04075 21 412 2.4 2.7 39.4 A Actin-related protein 2/3 complex O15144 11 31 2.1 2.7 34.3 subunit 2 P60709 Actin, cytoplasmic 1 22 1650 2.4 2.6 41.7 Q9Y3B8 Oligoribonuclease, mitochondrial 6 11 2.2 2.6 26.8 281

P13797 Plastin-3 20 95 2.3 2.6 70.8 Heterogeneous nuclear P09651 6 43 2.9 2.6 38.7 ribonucleoprotein A1 P51911 Calponin-1 7 12 2.2 2.6 33.1 P31949 Protein S100-A11 5 68 2.4 2.6 11.7 P60174 Triosephosphate isomerase 16 525 2.4 2.6 30.8 P07951 Tropomyosin beta chain 16 133 2.1 2.6 32.8 P00568 Adenylate kinase isoenzyme 1 6 17 2.6 2.6 21.6 Dihydropyrimidinase-related Q14195 17 64 2.3 2.6 61.9 protein 3 LIM and cysteine-rich domains Q9NZU5 9 16 2.0 2.6 40.8 protein 1 Phosphatidylethanolamine- P30086 11 66 2.6 2.5 21.0 binding protein 1 Peptidyl-prolyl cis-trans Q96AY3 6 16 2.0 2.5 64.2 isomerase FKBP10 Ubiquitin-conjugating enzyme E2 P61088 5 12 2.1 2.5 17.1 N ATP-dependent RNA helicase Q92499 2 2 3.7 2.5 82.4 DDX1 P61158 Actin-related protein 3 9 31 2.3 2.5 47.3 P21333 Filamin-A 107 1119 2.5 2.5 280.6 P49720 Proteasome subunit beta type-3 3 3 2.2 2.5 22.9 Serine/threonine-protein Q14738 phosphatase 2A 56 kDa 4 7 2.9 2.5 69.9 regulatory subunit delta isoform Peptidyl-prolyl cis-trans P62942 4 16 2.3 2.5 11.9 isomerase FKBP1A P61981 14-3-3 protein gamma 18 297 2.3 2.5 28.3 O43707 Alpha-actinin-4 55 782 2.2 2.5 104.8 Eukaryotic translation initiation P63241 3 229 2.3 2.5 16.8 factor 5A-1 Heterogeneous nuclear Q14103 6 10 2.0 2.4 38.4 ribonucleoprotein D0 Malate dehydrogenase, P40925 12 61 2.5 2.4 36.4 cytoplasmic Thioredoxin domain-containing Q8NBS9 7 8 2.0 2.4 47.6 protein 5 P19022 Cadherin-2 6 10 1.8 2.4 99.7 Fructose-bisphosphate aldolase P09972 8 90 2.4 2.4 39.4 C Immunoglobulin superfamily O14498 containing leucine-rich repeat 6 20 2.1 2.4 46.0 protein Q9Y490 Talin-1 37 90 2.3 2.3 269.6 P04080 Cystatin-B 4 73 2.9 2.3 11.1 Thioredoxin domain-containing Q9BRA2 5 13 2.4 2.3 13.9 protein 17 O43852 Calumenin 13 38 2.0 2.3 37.1

282

P28300 Protein-lysine 6-oxidase 2 4 2.3 46.9 P05388 60S acidic ribosomal protein P0 4 11 2.4 2.3 34.3 14 kDa phosphohistidine Q9NRX4 7 23 2.4 2.3 13.8 phosphatase P24534 Elongation factor 1-beta 5 10 2.5 2.3 24.7 P35052 Glypican-1 7 10 1.9 2.3 61.6 P08670 Vimentin 36 735 2.0 2.3 53.6 Cullin-associated NEDD8- Q86VP6 19 46 2.3 2.3 136.3 dissociated protein 1 P41250 Glycine--tRNA ligase 16 97 2.4 2.3 83.1 Guanine nucleotide-binding P62873 protein G(I)/G(S)/G(T) subunit 2 2 1.9 2.2 37.4 beta-1 Beta-hexosaminidase subunit P07686 5 8 2.0 2.2 63.1 beta Q9NY33 Dipeptidyl peptidase 3 7 15 2.0 2.2 82.5 P98095 Fibulin-2 4 8 2.0 2.2 126.5 P08758 Annexin A5 13 81 2.2 2.2 35.9 Q15417 Calponin-3 7 13 1.9 2.2 36.4 P14314 Glucosidase 2 subunit beta 3 5 1.7 2.2 59.4 P27797 Calreticulin 13 53 1.9 2.2 48.1 Chondroitin sulfate proteoglycan Q6UVK1 10 20 2.1 2.2 250.4 4 Q15149 Plectin 22 44 1.9 2.2 531.5 EGF-like repeat and discoidin I- O43854 5 9 2.0 2.2 53.7 like domain-containing protein 3 Q13219 Pappalysin-1 21 49 2.1 2.2 180.9 P02462 Collagen alpha-1(IV) chain 18 86 2.0 2.1 160.5 P29401 Transketolase 16 94 2.1 2.1 67.8 O95084 Serine protease 23 3 4 1.8 2.1 43.0 P51397 Death-associated protein 1 3 5 3.6 2.1 11.2 High mobility group protein P52926 3 3 2.0 2.1 11.8 HMGI-C Adipocyte enhancer-binding Q8IUX7 8 13 1.7 2.1 130.8 protein 1 Q00610 Clathrin heavy chain 1 34 106 2.1 2.1 191.5 P25787 Proteasome subunit alpha type-2 7 24 2.0 2.1 25.9 Tryptophan--tRNA ligase, P23381 11 38 2.0 2.1 53.1 cytoplasmic P60660 Myosin light polypeptide 6 6 24 2.0 2.1 16.9 P48637 Glutathione synthetase 6 19 2.3 2.1 52.4 P05452 Tetranectin 5 15 2.0 2.1 22.5 P07339 Cathepsin D 13 68 1.9 2.0 44.5 Inter-alpha-trypsin inhibitor heavy Q86UX2 2 18 2.0 104.5 chain H5 P28074 Proteasome subunit beta type-5 4 11 1.9 2.0 28.5

283

Q15084 Protein disulfide-isomerase A6 4 6 1.8 2.0 48.1 Q9BY76 Angiopoietin-related protein 4 3 7 1.7 2.0 45.2 P08238 Heat shock protein HSP 90-beta 25 201 2.0 2.0 83.2 Q92743 Serine protease HTRA1 11 84 1.8 2.0 51.3 P27658 Collagen alpha-1(VIII) chain 13 45 1.8 1.9 73.3 P27816 Microtubule-associated protein 4 10 33 2.2 1.9 120.9 Rab GDP dissociation inhibitor P31150 14 167 2.0 1.9 50.6 alpha P15502 Elastin 6 18 1.9 1.9 68.4 P35579 Myosin-9 55 232 2.4 1.9 226.4 Heterogeneous nuclear P22626 9 27 2.0 1.9 37.4 ribonucleoproteins A2/B1 P29279 Connective tissue growth factor 4 8 1.9 1.9 38.1 Q16610 Extracellular matrix protein 1 10 27 1.6 1.8 60.6 P28799 Granulins 5 12 1.8 1.8 63.5 Q92626 Peroxidasin homolog 23 65 1.8 1.8 165.2 O60565 Gremlin-1 2 2 2.3 1.8 20.7 Heat shock protein HSP 90- P07900 24 225 2.0 1.8 84.6 alpha Beta-hexosaminidase subunit P06865 5 11 1.7 1.8 60.7 alpha O75369 Filamin-B 43 218 2.0 1.8 278.0 P35442 Thrombospondin-2 26 116 1.8 1.8 129.9 P07585 Decorin 14 96 1.7 1.8 39.7 Hyaluronan and proteoglycan P10915 15 127 1.9 1.8 40.1 link protein 1 P16870 Carboxypeptidase E 9 19 1.8 1.8 53.1 Nucleosome assembly protein 1- P55209 3 5 1.8 45.3 like 1 P06396 Gelsolin 29 188 1.9 1.7 85.6 P07355 Annexin A2 17 61 1.7 1.7 38.6 P14625 Endoplasmin 14 69 1.8 1.7 92.4 Acidic leucine-rich nuclear Q92688 phosphoprotein 32 family 4 13 1.6 1.7 28.8 member B P35556 Fibrillin-2 10 25 1.6 1.7 314.6 P30101 Protein disulfide-isomerase A3 15 59 1.9 1.7 56.7 Insulin-like growth factor-binding P24593 9 48 1.5 1.7 30.6 protein 5 Q9NRN5 Olfactomedin-like protein 3 6 12 1.8 1.7 46.0 P12110 Collagen alpha-2(VI) chain 13 34 1.6 1.7 108.5 P08572 Collagen alpha-2(IV) chain 43 144 1.6 1.6 167.4 P12107 Collagen alpha-1(XI) chain 15 49 1.6 1.6 181.0 Secreted frizzled-related protein Q96HF1 9 22 1.6 1.6 33.5 2 Q15063 Periostin 38 427 1.7 1.6 93.3 284

P36578 60S ribosomal protein L4 4 8 2.0 1.6 47.7 P12111 Collagen alpha-3(VI) chain 45 124 1.7 1.6 343.5 P08603 Complement factor H 11 20 1.6 1.6 139.0 O76061 Stanniocalcin-2 6 78 1.5 1.5 33.2 Coiled-coil domain-containing Q76M96 16 69 1.7 1.5 108.1 protein 80 P02545 Prelamin-A/C 8 42 1.8 1.5 74.1 P61769 Beta-2-microglobulin 4 41 1.6 1.5 13.7 O00391 Sulfhydryl oxidase 1 20 123 1.7 1.5 82.5 P08476 Inhibin beta A chain 16 35 1.6 1.5 47.4 Q99715 Collagen alpha-1(XII) chain 59 175 1.5 1.5 332.9 Q9UBP4 Dickkopf-related protein 3 8 34 1.5 1.4 38.4 P01033 Metalloproteinase inhibitor 1 8 236 1.6 1.4 23.2 Q9UI42 Carboxypeptidase A4 10 29 1.4 1.4 47.3 Q9UBX5 Fibulin-5 5 12 1.5 1.4 50.1 P07237 Protein disulfide-isomerase 16 95 1.6 1.4 57.1 P07942 Laminin subunit beta-1 17 42 1.5 1.4 197.9 Q16777 Histone H2A type 2-C 3 25 1.4 1.4 14.0 P10909 Clusterin 10 91 1.3 52.5 Insulin-like growth factor-binding P18065 15 97 1.6 1.3 34.8 protein 2 P29692 Elongation factor 1-delta 4 9 2.1 1.3 31.1 P13611 Versican core protein 16 87 1.5 1.3 372.6 Transforming growth factor-beta- Q15582 28 676 1.5 1.3 74.6 induced protein ig-h3 P12109 Collagen alpha-1(VI) chain 34 285 1.5 1.3 108.5 Q9NQC3 Reticulon-4 2 2 1.5 1.2 129.9 Cellular nucleic acid-binding P62633 3 10 1.9 1.2 19.4 protein O15540 Fatty acid-binding protein, brain 3 34 1.5 1.2 14.9 Urokinase-type plasminogen P00749 4 20 1.3 0.8 48.5 activator P01011 Alpha-1-antichymotrypsin 2 3 0.8 47.6 P02790 Hemopexin 2 3 1.2 0.8 51.6 Q6NXT2 Histone H3.3C 2 5 0.7 0.7 15.2 P01024 Complement C3 5 40 0.7 187.0 P01709 Ig lambda chain V-II region MGC 2 15 0.7 11.6 P02787 Serotransferrin 3 30 0.6 77.0 P00738 Haptoglobin 6 227 0.6 45.2 Isopentenyl-diphosphate Delta- Q13907 2 4 26.3 isomerase 1 Bifunctional glutamate/proline-- P07814 2 4 170.5 tRNA ligase O14818 Proteasome subunit alpha type-7 4 5 27.9

285

Glutamine--fructose-6-phosphate Q06210 3 5 78.8 aminotransferase [isomerizing] 1 Q01105 Protein SET 2 2 33.5 P42574 Caspase-3 2 3 31.6 O75711 Scrapie-responsive protein 1 2 2 11.1 P62269 40S ribosomal protein S18 2 3 17.7 Cold-inducible RNA-binding Q14011 2 2 18.6 protein Q96FQ6 Protein S100-A16 3 7 11.8 O95678 Keratin, type II cytoskeletal 75 4 8 59.5 N(G),N(G)-dimethylarginine O95865 3 8 29.6 dimethylaminohydrolase 2 Cytosolic non-specific Q96KP4 3 4 52.8 dipeptidase Q9BQE3 Tubulin alpha-1C chain 19 216 49.9 P09104 Gamma-enolase 7 558 47.2 Q15181 Inorganic pyrophosphatase 4 6 32.6 Procollagen-lysine,2- O00469 6 6 84.6 oxoglutarate 5-dioxygenase 2 P62081 40S ribosomal protein S7 3 4 22.1 Q13509 Tubulin beta-3 chain 15 180 50.4 Spectrin beta chain, non- Q01082 5 5 274.4 erythrocytic 1 P10768 S-formylglutathione hydrolase 3 4 31.4 P04264 Keratin, type II cytoskeletal 1 14 88 66.0 Asparagine--tRNA ligase, O43776 4 7 62.9 cytoplasmic Q13308 Inactive tyrosine-protein kinase 7 7 12 118.3 P62273 40S ribosomal protein S29 2 2 6.7 Q05707 Collagen alpha-1(XIV) chain 2 2 193.4 Myosin regulatory light P24844 6 20 19.8 polypeptide 9 O95965 Integrin beta-like protein 1 4 7 53.9 Actin-related protein 2/3 complex P59998 3 3 19.7 subunit 4 P13667 Protein disulfide-isomerase A4 2 3 72.9 Q6IBS0 Twinfilin-2 2 2 39.5 P46781 40S ribosomal protein S9 2 2 22.6 P35527 Keratin, type I cytoskeletal 9 7 130 62.0 P00491 Purine nucleoside phosphorylase 2 4 32.1 T-complex protein 1 subunit P49368 2 2 60.5 gamma Glycogen phosphorylase, liver P06737 4 4 97.1 form Q16363 Laminin subunit alpha-4 3 7 202.4 26S proteasome non-ATPase Q13200 4 7 100.1 regulatory subunit 2 286

Nuclease-sensitive element- P67809 4 8 35.9 binding protein 1 P55060 Exportin-2 2 2 110.3 Q9P1F3 Costars family protein ABRACL 2 2 9.1 Guanine nucleotide-binding P63244 2 2 35.1 protein subunit beta-2-like 1 P01034 Cystatin-C 8 51 15.8 Mitogen-activated protein kinase P28482 2 6 41.4 1 P46783 40S ribosomal protein S10 2 4 18.9 P36222 Chitinase-3-like protein 1 4 4 42.6 Keratin, type II cytoskeletal 2 P35908 11 64 65.4 epidermal Q86YZ3 Hornerin 12 20 282.2 Q08629 Testican-1 4 4 49.1 Microfibril-associated P55083 5 105 28.6 glycoprotein 4 O60814 Histone H2B type 1-K 2 3 13.9 P11021 78 kDa glucose-regulated protein 14 77 1.8 72.3 P30043 Flavin reductase (NADPH) 2 2 2.3 22.1 Spectrin alpha chain, non- Q13813 8 19 1.8 284.4 erythrocytic 1 Q14315 Filamin-C 26 205 2.1 290.8 Aspartate aminotransferase, P17174 6 10 2.0 46.2 cytoplasmic P35237 Serpin B6 6 12 2.1 42.6 P04083 Annexin A1 7 51 1.9 38.7 P05387 60S acidic ribosomal protein P2 3 4 2.0 11.7 P13489 Ribonuclease inhibitor 7 12 2.0 49.9 P11047 Laminin subunit gamma-1 13 35 1.6 177.5 N(G),N(G)-dimethylarginine O94760 5 12 1.5 31.1 dimethylaminohydrolase 1 P61247 40S ribosomal protein S3a 2 3 2.5 29.9 P28070 Proteasome subunit beta type-4 4 11 1.7 29.2 Q96D15 Reticulocalbin-3 2 3 1.6 37.5 Q08431 Lactadherin 4 18 1.5 43.1 MAM domain-containing protein Q7Z304 16 47 1.7 77.5 2 P50454 Serpin H1 17 94 1.5 46.4 P02795 Metallothionein-2 3 21 3.0 6.0 Peptidyl-prolyl cis-trans P23284 8 123 1.5 23.7 isomerase B Platelet-activating factor P68402 3 23 2.7 25.6 acetylhydrolase IB subunit beta Transitional endoplasmic P55072 4 4 2.5 89.3 reticulum ATPase P15018 Leukemia inhibitory factor 2 24 1.8 22.0 287

Serine/threonine-protein P36873 phosphatase PP1-gamma 2 4 1.6 37.0 catalytic subunit P51858 Hepatoma-derived growth factor 2 4 1.9 26.8 P24043 Laminin subunit alpha-2 11 20 1.7 343.7 P20908 Collagen alpha-1(V) chain 30 127 1.5 183.4 Four and a half LIM domains Q13643 4 8 2.2 31.2 protein 3 P07093 Glia-derived nexin 14 121 1.6 44.0 P62841 40S ribosomal protein S15 2 2 2.2 17.0 Q02818 Nucleobindin-1 13 52 1.5 53.8 P21810 Biglycan 21 674 1.4 41.6 P02751 Fibronectin 103 5607 1.5 262.5 Q9Y4K0 Lysyl oxidase homolog 2 10 46 1.4 86.7 Q14697 Neutral alpha-glucosidase AB 4 6 2.2 106.8 P24821 Tenascin 48 370 1.4 240.7 Procollagen-lysine,2- Q02809 23 122 1.4 83.5 oxoglutarate 5-dioxygenase 1 Latent-transforming growth factor Q14767 27 127 1.5 194.9 beta-binding protein 2 Q9P2E9 Ribosome-binding protein 1 3 6 1.5 152.4 Chitinase domain-containing Q9BWS9 3 13 1.4 44.9 protein 1 C-type lectin domain family 11 Q9Y240 2 2 1.8 35.7 member A P15121 Aldose reductase 4 11 2.2 35.8 Latent-transforming growth factor Q14766 14 34 1.5 186.7 beta-binding protein 1 Q14118 Dystroglycan 4 8 1.7 97.4 P16035 Metalloproteinase inhibitor 2 11 39 1.5 24.4 Basement membrane-specific P98160 heparan sulfate proteoglycan 30 146 1.4 468.5 core protein P02461 Collagen alpha-1(III) chain 44 179 1.5 138.5 P02765 Alpha-2-HS-glycoprotein 2 3 0.5 39.3 P35555 Fibrillin-1 76 522 1.5 312.0 P20930 Filaggrin 5 6 2.2 434.9 P01023 Alpha-2-macroglobulin 22 200 1.1 163.2 P62906 60S ribosomal protein L10a 2 4 1.6 24.8 Fibronectin type III domain- Q4ZHG4 11 22 1.5 205.4 containing protein 1 P07996 Thrombospondin-1 46 387 1.4 129.3 O94985 Calsyntenin-1 8 26 1.3 109.7 O00468 Agrin 15 44 1.6 217.1 P41222 Prostaglandin-H2 D-isomerase 2 2 1.4 21.0 P14136 Glial fibrillary acidic protein 8 33 1.3 49.8 288

P06753 Tropomyosin alpha-3 chain 10 79 2.6 32.9 Q9UJ70 N-acetyl-D-glucosamine kinase 5 9 1.8 37.4 P16070 CD44 antigen 2 15 1.5 81.5 P46777 60S ribosomal protein L5 3 6 1.8 34.3 Q12841 Follistatin-related protein 1 22 718 1.4 35.0 P46108 Adapter molecule crk 2 2 1.7 33.8 Q9H4D0 Calsyntenin-2 10 22 1.3 106.9 Insulin-like growth factor-binding P17936 11 42 1.6 31.7 protein 3 EGF-containing fibulin-like Q12805 17 89 1.3 54.6 extracellular matrix protein 1 Insulin-like growth factor-binding Q16270 21 1187 1.3 29.1 protein 7 P51884 Lumican 12 92 1.2 38.4 Q08380 Galectin-3-binding protein 21 137 1.3 65.3 P05121 Plasminogen activator inhibitor 1 24 1641 1.3 45.0 Polypeptide N- Q10471 acetylgalactosaminyltransferase 9 46 1.3 64.7 2 P02452 Collagen alpha-1(I) chain 86 1945 1.3 138.9 P05997 Collagen alpha-2(V) chain 37 117 1.4 144.8 P61916 Epididymal secretory protein E1 5 14 1.3 16.6 P08123 Collagen alpha-2(I) chain 74 1563 1.4 129.2 P02768 Serum albumin 16 381 1.3 69.3 P08253 72 kDa type IV collagenase 25 131 1.2 73.8 P09486 SPARC 20 3614 1.3 34.6 EGF-containing fibulin-like O95967 4 18 1.3 49.4 extracellular matrix protein 2

Table 5.6: Differentially released conditioned medium proteins at 5 hours

Astrocyte-released proteins measured in condition media from TMT experiments comparing protein abundance between mild and severe stretched cells in conditioned media collected 5 hours post-injury.

289

# MW Accession Description PSMs 24h M 24h S Peptides (kDa) P61158 Actin-related protein 3 9 12 1.4 7.5 47.3 Heterogeneous nuclear P22626 10 72 2.2 7.2 37.4 ribonucleoproteins A2/B1 P09493 Tropomyosin alpha-1 chain 15 198 2.1 6.8 32.7 P12109 Collagen alpha-1(VI) chain 13 50 1.3 6.3 108.5 Malate dehydrogenase, P40925 11 87 2.3 6.2 36.4 cytoplasmic P24821 Tenascin 10 13 6.1 240.7 P31946 14-3-3 protein beta/alpha 16 199 1.8 5.9 28.1 Fibronectin type III domain- Q4ZHG4 5 11 5.7 205.4 containing protein 1 Astrocytic phosphoprotein PEA- Q15121 5 10 1.8 5.6 15.0 15 P62328 Thymosin beta-4 4 22 3.5 5.4 5.0 O75369 Filamin-B 22 63 5.3 278.0 P29401 Transketolase 14 74 2.1 5.2 67.8 P61163 Alpha-centractin 4 34 2.1 4.9 42.6 P62158 Calmodulin 4 31 2.4 4.8 16.8 P00441 Superoxide dismutase [Cu-Zn] 3 40 3.8 4.8 15.9 P62258 14-3-3 protein epsilon 13 215 1.7 4.7 29.2 P00558 Phosphoglycerate kinase 1 19 189 1.9 4.7 44.6 Microtubule-associated protein P46821 5 8 4.6 270.5 1B P30101 Protein disulfide-isomerase A3 15 46 1.6 4.5 56.7 Q9H4D0 Calsyntenin-2 4 5 4.5 106.9 P05997 Collagen alpha-2(V) chain 12 22 4.4 144.8 P61981 14-3-3 protein gamma 15 225 1.9 4.4 28.3 Thioredoxin domain-containing Q8NBS9 5 6 1.3 4.4 47.6 protein 5 Acidic leucine-rich nuclear P39687 phosphoprotein 32 family 2 2 1.5 4.3 28.6 member A P01008 Antithrombin-III 7 19 0.8 4.3 52.6 P12814 Alpha-actinin-1 56 1056 1.9 4.3 103.0 EGF-containing fibulin-like Q12805 9 43 4.3 54.6 extracellular matrix protein 1 P13797 Plastin-3 16 29 1.7 4.3 70.8 Cysteine and glycine-rich protein P21291 9 142 2.1 4.3 20.6 1 P26038 Moesin 29 297 2.1 4.3 67.8 P07237 Protein disulfide-isomerase 15 50 1.6 4.2 57.1 Q92626 Peroxidasin homolog 3 3 4.2 165.2 P01024 Complement C3 5 20 4.1 187.0 Eukaryotic translation initiation Q15056 2 2 3.1 4.0 27.4 factor 4H 290

Aspartate aminotransferase, P17174 4 4 1.8 4.0 46.2 cytoplasmic P37802 Transgelin-2 14 170 1.7 4.0 22.4 P49327 Fatty acid synthase 2 2 4.0 273.3 O75083 WD repeat-containing protein 1 10 42 1.8 3.9 66.2 Q9BVA1 Tubulin beta-2B chain 15 116 2.1 3.9 49.9 Q14847 LIM and SH3 domain protein 1 5 25 2.2 3.9 29.7 P09382 Galectin-1 6 119 2.0 3.9 14.7 Protein phosphatase 1 regulatory Q96C90 2 3 3.9 15.9 subunit 14B P51397 Death-associated protein 1 2 3 2.9 3.9 11.2 Ubiquitin-conjugating enzyme E2 P68036 5 15 1.7 3.8 17.9 L3 P28838 Cytosol aminopeptidase 2 2 3.8 56.1 P52565 Rho GDP-dissociation inhibitor 1 3 18 2.3 3.8 23.2 P35555 Fibrillin-1 23 86 1.3 3.8 312.0 Peptidyl-prolyl cis-trans P62942 3 12 2.2 3.8 11.9 isomerase FKBP1A P0CG48 Polyubiquitin-C 3 23 2.4 3.8 77.0 P01033 Metalloproteinase inhibitor 1 7 158 1.5 3.8 23.2 Q969H8 UPF0556 protein C19orf10 2 2 3.8 18.8 Ubiquitin carboxyl-terminal P09936 14 241 2.6 3.8 24.8 hydrolase isozyme L1 Thioredoxin domain-containing Q9BRA2 2 4 3.8 13.9 protein 17 Fructose-bisphosphate aldolase P09972 7 81 1.9 3.8 39.4 C P23396 40S ribosomal protein S3 4 5 3.8 26.7 P04083 Annexin A1 5 15 1.3 3.8 38.7 Q02818 Nucleobindin-1 4 5 3.8 53.8 Heat shock protein HSP 90- P07900 21 146 1.8 3.7 84.6 alpha P06396 Gelsolin 19 113 1.3 3.7 85.6 O43707 Alpha-actinin-4 43 681 1.6 3.7 104.8 Four and a half LIM domains Q14192 6 24 1.9 3.7 32.2 protein 2 P26022 Pentraxin-related protein PTX3 3 6 3.7 41.9 P07951 Tropomyosin beta chain 17 204 2.3 3.6 32.8 P10909 Clusterin 8 85 1.2 3.6 52.5 P08238 Heat shock protein HSP 90-beta 17 109 1.6 3.5 83.2 P63104 14-3-3 protein zeta/delta 20 328 2.2 3.5 27.7 P08572 Collagen alpha-2(IV) chain 14 57 3.5 167.4 P27816 Microtubule-associated protein 4 7 12 1.7 3.5 120.9 Rab GDP dissociation inhibitor P31150 14 51 1.5 3.5 50.6 alpha

291

SH3 domain-binding glutamic Q9H299 5 41 3.3 3.5 10.4 acid-rich-like protein 3 P35579 Myosin-9 21 27 2.3 3.4 226.4 Q53FA7 Quinone oxidoreductase PIG3 2 2 3.4 35.5 P62826 GTP-binding nuclear protein Ran 4 6 1.8 3.4 24.4 P07737 Profilin-1 9 172 1.9 3.4 15.0 P35237 Serpin B6 4 5 1.4 3.4 42.6 Insulin-like growth factor-binding P18065 11 103 1.3 3.4 34.8 protein 2 P06703 Protein S100-A6 3 26 2.3 3.4 10.2 Q01995 Transgelin 19 989 2.8 3.4 22.6 P60174 Triosephosphate isomerase 15 305 2.1 3.4 30.8 P16035 Metalloproteinase inhibitor 2 4 11 3.4 24.4 Q71U36 Tubulin alpha-1A chain 15 163 3.4 50.1 Glyceraldehyde-3-phosphate P04406 14 369 1.8 3.4 36.0 dehydrogenase P60709 Actin, cytoplasmic 1 19 1242 1.8 3.4 41.7 Fructose-bisphosphate aldolase P04075 18 368 2.1 3.4 39.4 A P30085 UMP-CMP kinase 5 26 1.7 3.4 22.2 P10599 Thioredoxin 3 20 2.0 3.4 11.7 Phosphatidylethanolamine- P30086 10 118 2.2 3.3 21.0 binding protein 1 Non-histone chromosomal P05204 2 2 3.3 9.4 protein HMG-17 SH3 domain-binding glutamic O75368 4 7 2.1 3.3 12.8 acid-rich-like protein P07996 Thrombospondin-1 33 247 3.3 129.3 Basement membrane-specific P98160 heparan sulfate proteoglycan 11 23 3.3 468.5 core protein P19105 Myosin regulatory light chain 12A 3 6 1.4 3.3 19.8 O94985 Calsyntenin-1 2 2 3.3 109.7 Q9UI42 Carboxypeptidase A4 7 46 3.3 47.3 P21333 Filamin-A 84 791 1.8 3.3 280.6 Serine/threonine-protein phosphatase 2A 65 kDa P30153 2 2 3.3 65.3 regulatory subunit A alpha isoform Translationally-controlled tumor P13693 3 42 1.9 3.3 19.6 protein Q15063 Periostin 15 80 1.2 3.3 93.3 Q14315 Filamin-C 20 70 1.5 3.3 290.8 Transforming growth factor-beta- Q15582 13 49 1.2 3.3 74.6 induced protein ig-h3 Q9NY33 Dipeptidyl peptidase 3 4 4 3.2 82.5 Q15417 Calponin-3 4 7 1.9 3.2 36.4

292

P21266 Glutathione S-transferase Mu 3 4 12 2.2 3.2 26.5 Q9UBP4 Dickkopf-related protein 3 6 14 3.2 38.4 P05121 Plasminogen activator inhibitor 1 15 430 3.2 45.0 P32119 Peroxiredoxin-2 6 51 1.8 3.2 21.9 Q92820 Gamma-glutamyl hydrolase 3 3 3.2 35.9 P07195 L-lactate dehydrogenase B chain 13 193 1.8 3.2 36.6 P49720 Proteasome subunit beta type-3 3 5 3.2 22.9 P09211 Glutathione S-transferase P 6 64 2.3 3.2 23.3 P68371 Tubulin beta-4B chain 14 86 1.6 3.1 49.8 P68032 Actin, alpha cardiac muscle 1 17 856 2.0 3.1 42.0 P06733 Alpha-enolase 22 526 2.4 3.1 47.1 Chloride intracellular channel Q9Y696 11 89 1.8 3.1 28.8 protein 4 Q9Y617 Phosphoserine aminotransferase 12 60 1.5 3.1 40.4 P18669 Phosphoglycerate mutase 1 15 149 2.0 3.1 28.8 Inter-alpha-trypsin inhibitor heavy P19827 2 3 3.1 101.3 chain H1 P80723 Brain acid soluble protein 1 2 2 3.1 22.7 P15311 Ezrin 23 179 2.9 3.1 69.4 P30046 D-dopachrome decarboxylase 4 9 1.7 3.1 12.7 14 kDa phosphohistidine Q9NRX4 6 34 3.8 3.1 13.8 phosphatase P01709 Ig lambda chain V-II region MGC 2 6 3.0 11.6 Q13219 Pappalysin-1 4 4 3.0 180.9 P23528 Cofilin-1 13 253 2.5 3.0 18.5 P02795 Metallothionein-2 3 5 2.0 3.0 6.0 P31949 Protein S100-A11 5 65 1.9 3.0 11.7 6-phosphogluconate P52209 2 3 3.0 53.1 dehydrogenase, decarboxylating P21810 Biglycan 15 174 1.2 3.0 41.6 P14618 Pyruvate kinase PKM 34 636 1.9 3.0 57.9 P18206 Vinculin 64 551 1.9 3.0 123.7 P28066 Proteasome subunit alpha type-5 3 9 1.6 3.0 26.4 Heterogeneous nuclear Q14103 4 7 3.0 38.4 ribonucleoprotein D0 P61970 Nuclear transport factor 2 2 2 1.5 3.0 14.5 P04792 Heat shock protein beta-1 13 60 1.9 2.9 22.8 Q562R1 Beta-actin-like protein 2 8 435 2.9 42.0 P15531 Nucleoside diphosphate kinase A 5 51 2.9 17.1 Latent-transforming growth factor Q14767 11 36 2.9 194.9 beta-binding protein 2 Chloride intracellular channel O00299 6 12 1.8 2.9 26.9 protein 1 Q16658 Fascin 11 29 1.6 2.9 54.5

293

Protein phosphatase 1 regulatory Q96T49 2 2 2.9 63.5 inhibitor subunit 16B P12107 Collagen alpha-1(XI) chain 3 12 2.8 181.0 Rab GDP dissociation inhibitor P50395 18 100 1.5 2.8 50.6 beta Four and a half LIM domains Q13642 9 35 1.7 2.8 36.2 protein 1 P08107 Heat shock 70 kDa protein 1A/1B 7 14 1.6 2.8 70.0 Q06830 Peroxiredoxin-1 9 135 2.3 2.8 22.1 P04080 Cystatin-B 3 83 1.6 2.8 11.1 Peptidyl-prolyl cis-trans P23284 7 27 1.3 2.8 23.7 isomerase B O14818 Proteasome subunit alpha type-7 4 4 1.8 2.8 27.9 P37837 Transaldolase 5 6 1.8 2.8 37.5 Mitogen-activated protein kinase P28482 2 2 2.8 41.4 1 P48637 Glutathione synthetase 3 3 2.7 52.4 Thioredoxin reductase 1, Q16881 3 3 2.7 70.9 cytoplasmic P07355 Annexin A2 12 53 1.8 2.7 38.6 P60660 Myosin light polypeptide 6 4 9 1.6 2.7 16.9 Nicotinamide N- P40261 4 13 1.9 2.7 29.6 methyltransferase Ubiquitin-conjugating enzyme E2 P61088 4 12 1.8 2.7 17.1 N Ubiquitin-like modifier-activating P22314 5 5 2.7 117.8 enzyme 1 Cysteine and glycine-rich protein Q16527 5 14 1.6 2.7 20.9 2 Insulin-like growth factor-binding P24593 8 70 2.7 30.6 protein 5 P48163 NADP-dependent malic enzyme 3 3 2.6 64.1 P07339 Cathepsin D 13 77 2.6 44.5 Q9UBR2 Cathepsin Z 2 2 2.6 33.8 Isocitrate dehydrogenase O75874 2 2 2.6 46.6 [NADP] cytoplasmic P16152 Carbonyl reductase [NADPH] 1 3 6 2.0 2.6 30.4 P68363 Tubulin alpha-1B chain 15 163 2.6 50.1 P25786 Proteasome subunit alpha type-1 6 16 2.0 2.6 29.5 Nucleosome assembly protein 1- Q99733 2 2 2.5 42.8 like 4 P02461 Collagen alpha-1(III) chain 22 47 2.5 138.5 Heterogeneous nuclear Q32P51 3 7 2.3 2.5 34.2 ribonucleoprotein A1-like 2 P27348 14-3-3 protein theta 15 158 1.7 2.5 27.7 P41250 Glycine--tRNA ligase 2 3 2.5 83.1 Q00610 Clathrin heavy chain 1 2 2 2.5 191.5

294

Adenylyl cyclase-associated Q01518 12 38 1.9 2.5 51.9 protein 1 P61160 Actin-related protein 2 6 36 1.8 2.5 44.7 P51911 Calponin-1 4 5 2.5 33.1 Q08380 Galectin-3-binding protein 12 61 0.8 2.5 65.3 Peptidyl-prolyl cis-trans P62937 14 239 2.0 2.5 18.0 isomerase A P15289 Arylsulfatase A 2 2 2.4 53.6 P63167 Dynein light chain 1, cytoplasmic 2 2 2.4 10.4 Insulin-like growth factor-binding Q16270 17 904 2.4 29.1 protein 7 Macrophage migration inhibitory P14174 3 8 2.1 2.4 12.5 factor P62906 60S ribosomal protein L10a 4 5 1.7 2.4 24.8 P30044 Peroxiredoxin-5, mitochondrial 2 3 2.4 22.1 Q9NTK5 Obg-like ATPase 1 2 2 2.4 44.7 Actin-related protein 2/3 complex O15145 2 2 2.4 20.5 subunit 3 Ras GTPase-activating-like P46940 16 19 2.3 189.1 protein IQGAP1 Latent-transforming growth factor Q14766 2 5 2.3 186.7 beta-binding protein 1 P02462 Collagen alpha-1(IV) chain 6 27 2.3 160.5 Q9Y3B8 Oligoribonuclease, mitochondrial 2 3 2.3 26.8 P24534 Elongation factor 1-beta 5 11 2.0 2.3 24.7 P13667 Protein disulfide-isomerase A4 2 2 2.3 72.9 P00966 Argininosuccinate synthase 3 4 1.7 2.2 46.5 P14625 Endoplasmin 9 13 2.2 92.4 P01009 Alpha-1-antitrypsin 3 75 2.2 46.7 P27797 Calreticulin 10 106 1.5 2.1 48.1 LIM and cysteine-rich domains Q9NZU5 4 9 1.8 2.0 40.8 protein 1 Farnesyl pyrophosphate P14324 3 11 2.2 2.0 48.2 synthase Q96HC4 PDZ and LIM domain protein 5 2 2 2.0 63.9 P14314 Glucosidase 2 subunit beta 3 5 1.5 2.0 59.4 Coiled-coil domain-containing Q76M96 5 6 1.9 108.1 protein 80 P06753 Tropomyosin alpha-3 chain 8 87 1.9 32.9 P20742 Pregnancy zone protein 4 26 0.8 1.9 163.8 P00568 Adenylate kinase isoenzyme 1 5 9 3.2 1.9 21.6 P0C0L4 Complement C4-A 4 9 1.9 192.7 P02751 Fibronectin 72 1126 1.9 262.5 O00151 PDZ and LIM domain protein 1 2 2 1.9 36.0 Transforming growth factor beta- P61812 3 6 1.8 47.7 2

295

Q9NR12 PDZ and LIM domain protein 7 2 2 1.8 49.8 P08253 72 kDa type IV collagenase 14 42 1.8 73.8 Q14118 Dystroglycan 3 3 1.8 97.4 P30043 Flavin reductase (NADPH) 2 2 1.8 22.1 Q15084 Protein disulfide-isomerase A6 5 7 1.8 48.1 Puromycin-sensitive P55786 5 8 1.8 103.2 aminopeptidase Spectrin alpha chain, non- Q13813 9 11 2.1 1.8 284.4 erythrocytic 1 Actin-related protein 2/3 complex O15144 4 9 2.0 1.8 34.3 subunit 2 Q9Y281 Cofilin-2 6 139 1.6 1.7 18.7 P07093 Glia-derived nexin 14 62 1.7 44.0 Procollagen-lysine,2- Q02809 4 4 1.7 83.5 oxoglutarate 5-dioxygenase 1 Interleukin enhancer-binding Q12905 3 3 1.7 43.0 factor 2 P35241 Radixin 9 50 1.7 68.5 P23142 Fibulin-1 2 2 1.7 77.2 Q99538 Legumain 3 4 1.7 49.4 P00338 L-lactate dehydrogenase A chain 18 397 1.9 1.7 36.7 Hyaluronan and proteoglycan P10915 7 36 1.6 40.1 link protein 1 P28070 Proteasome subunit beta type-4 4 8 1.6 1.6 29.2 Q99439 Calponin-2 2 5 1.6 33.7 P07437 Tubulin beta chain 14 133 1.8 1.6 49.6 Keratin, type II cytoskeletal 2 P35908 5 16 1.6 65.4 epidermal Cytoplasmic dynein 1 heavy Q14204 3 3 1.6 532.1 chain 1 P05388 60S acidic ribosomal protein P0 4 6 1.6 34.3 P11021 78 kDa glucose-regulated protein 12 72 1.4 1.6 72.3 P02545 Prelamin-A/C 8 26 1.7 1.5 74.1 P02788 Lactotransferrin 2 13 1.5 78.1 O95336 6-phosphogluconolactonase 4 6 1.5 27.5 Actin-related protein 2/3 complex O15511 2 2 1.5 16.3 subunit 5 Heat shock cognate 71 kDa P11142 18 76 1.8 1.5 70.9 protein O76061 Stanniocalcin-2 3 13 1.5 33.2 P02511 Alpha-crystallin B chain 5 76 1.9 1.5 20.1 O43852 Calumenin 4 5 1.5 37.1 Q16181 Septin-7 2 4 1.4 50.6 P50454 Serpin H1 10 32 1.5 1.4 46.4 P60900 Proteasome subunit alpha type-6 2 3 1.4 27.4

296

High mobility group protein P52926 2 2 1.4 11.8 HMGI-C P68104 Elongation factor 1-alpha 1 10 47 1.9 1.4 50.1 P22392 Nucleoside diphosphate kinase B 6 83 2.2 1.4 17.3 Membrane primary amine Q16853 2 2 1.4 84.6 oxidase P29279 Connective tissue growth factor 4 6 1.4 38.1 P36871 Phosphoglucomutase-1 6 19 1.6 1.4 61.4 P25787 Proteasome subunit alpha type-2 2 2 1.4 25.9 COP9 signalosome complex Q99627 2 2 1.4 23.2 subunit 8 Q15181 Inorganic pyrophosphatase 3 3 1.4 32.6 P01023 Alpha-2-macroglobulin 10 134 1.4 163.2 P13639 Elongation factor 2 12 25 2.3 1.3 95.3 Eukaryotic translation initiation P63241 4 23 2.0 1.3 16.8 factor 5A-1 P08670 Vimentin 34 520 1.8 1.3 53.6 Nuclease-sensitive element- P67809 5 11 1.7 1.3 35.9 binding protein 1 P09960 Leukotriene A-4 hydrolase 7 11 2.1 1.3 69.2 Q05682 Caldesmon 23 189 2.3 1.3 93.2 P15121 Aldose reductase 3 5 1.3 35.8 Q99497 Protein DJ-1 7 53 1.9 1.3 19.9 Insulin-like growth factor-binding P17936 9 33 1.3 1.3 31.7 protein 3 P01034 Cystatin-C 4 11 1.2 15.8 Q14019 Coactosin-like protein 6 42 2.4 1.2 15.9 P60981 Destrin 4 17 2.1 1.2 18.5 P61769 Beta-2-microglobulin 2 14 1.2 13.7 P67936 Tropomyosin alpha-4 chain 16 152 2.5 1.2 28.5 Q14974 Importin subunit beta-1 3 3 1.1 97.1 Q8NCW5 NAD(P)H-hydrate epimerase 2 3 1.1 31.7 Q99584 Protein S100-A13 4 6 1.8 1.1 11.5 Serine/arginine-rich splicing Q07955 3 3 27.7 factor 1 MAM domain-containing protein Q7Z304 10 26 77.5 2 P02774 Vitamin D-binding protein 2 20 52.9 P13645 Keratin, type I cytoskeletal 10 5 11 58.8 Q01105 Protein SET 2 2 33.5 P20618 Proteasome subunit beta type-1 2 3 26.5 Q13185 Chromobox protein homolog 3 2 2 20.8 P11766 Alcohol dehydrogenase class-3 2 2 39.7 P08123 Collagen alpha-2(I) chain 54 715 129.2 P35052 Glypican-1 5 5 61.6 297

P07942 Laminin subunit beta-1 3 3 197.9 Q15019 Septin-2 3 3 41.5 Dihydropyrimidinase-related Q14195 4 5 61.9 protein 3 P35442 Thrombospondin-2 8 18 129.9 P08729 Keratin, type II cytoskeletal 7 8 17 51.4 Myosin light chain kinase, Q15746 3 4 210.6 smooth muscle P30041 Peroxiredoxin-6 3 3 25.0 P60983 Glia maturation factor beta 2 4 16.7 P35527 Keratin, type I cytoskeletal 9 10 97 62.0 P34932 Heat shock 70 kDa protein 4 3 3 94.3 P09104 Gamma-enolase 6 115 47.2 Q9ULV4 Coronin-1C 6 6 53.2 Probable histidine--tRNA ligase, P49590 2 2 56.9 mitochondrial P27658 Collagen alpha-1(VIII) chain 4 6 73.3 P00738 Haptoglobin 7 160 45.2 Serine/threonine-protein Q9BRF8 2 2 35.5 phosphatase CPPED1 Q15149 Plectin 5 5 531.5 Dihydropyrimidinase-related Q16555 3 3 62.3 protein 2 Q16610 Extracellular matrix protein 1 8 17 60.6 P09486 SPARC 16 837 34.6 P28074 Proteasome subunit beta type-5 2 3 28.5 Q99715 Collagen alpha-1(XII) chain 8 10 332.9 P07585 Decorin 4 5 39.7 P25789 Proteasome subunit alpha type-4 3 6 29.5 Polypeptide N- Q10471 acetylgalactosaminyltransferase 2 2 64.7 2 Q14393 Growth arrest-specific protein 6 2 2 79.6 Cold-inducible RNA-binding Q14011 2 2 18.6 protein P04264 Keratin, type II cytoskeletal 1 16 91 66.0 P02452 Collagen alpha-1(I) chain 69 1061 138.9 Cytosolic non-specific Q96KP4 2 2 52.8 dipeptidase Secreted frizzled-related protein Q96HF1 2 2 33.5 2 O00391 Sulfhydryl oxidase 1 9 26 82.5 Q12841 Follistatin-related protein 1 21 442 35.0 P08476 Inhibin beta A chain 2 2 47.4 Q9Y4K0 Lysyl oxidase homolog 2 3 3 86.7 P61586 Transforming protein RhoA 2 2 21.8 298

P51884 Lumican 11 88 38.4 Glyoxalase domain-containing Q9HC38 3 3 34.8 protein 4 P01860 Ig gamma-3 chain C region 2 2 41.3 P20908 Collagen alpha-1(V) chain 8 20 183.4 Q9H4A4 Aminopeptidase B 3 4 72.5 P46777 60S ribosomal protein L5 3 3 34.3 Cullin-associated NEDD8- Q86VP6 2 2 136.3 dissociated protein 1 O00468 Agrin 5 5 217.1 P63010 AP-2 complex subunit beta 2 2 104.5 P08758 Annexin A5 8 12 1.8 35.9 F-actin-capping protein subunit P47756 2 3 31.3 beta P07602 Prosaposin 9 19 1.3 58.1 Threonine--tRNA ligase, P26639 2 2 1.9 83.4 cytoplasmic Interleukin enhancer-binding Q12906 3 3 95.3 factor 3 P62241 40S ribosomal protein S8 2 6 24.2 P60842 Eukaryotic initiation factor 4A-I 6 14 1.8 46.1 N(G),N(G)-dimethylarginine O95865 5 5 1.9 29.6 dimethylaminohydrolase 2 P81605 Dermcidin 2 13 11.3 P06744 Glucose-6-phosphate isomerase 12 60 1.6 63.1 UTP--glucose-1-phosphate Q16851 6 9 1.6 56.9 uridylyltransferase Q9BUF5 Tubulin beta-6 chain 9 56 49.8 Q9Y490 Talin-1 18 21 2.2 269.6

Table 5.7: Differentially released conditioned medium proteins at 24 hours

Astrocyte-released proteins measured in condition media from TMT experiments comparing protein abundance between mild and severe stretched cells in conditioned media collected 24 hours post-injury.

299

# MW Accession Description PSMs 48h M 48h S Peptides (kDa) P62328 Thymosin beta-4 4 14 8.4 11.9 5.0 P07437 Tubulin beta chain 20 225 6.5 11.6 49.6 O60664 Perilipin-3 2 2 7.4 11.5 47.0 P20962 Parathymosin 2 5 5.4 11.5 11.5 P68032 Actin, alpha cardiac muscle 1 20 619 5.9 10.4 42.0 Q01995 Transgelin 20 1047 6.3 9.1 22.6 Astrocytic phosphoprotein PEA- Q15121 5 102 5.8 8.5 15.0 15 Q71U36 Tubulin alpha-1A chain 22 173 5.7 8.4 50.1 P15311 Ezrin 20 95 4.8 8.3 69.4 P37802 Transgelin-2 15 206 6.0 8.3 22.4 P22392 Nucleoside diphosphate kinase B 7 62 5.3 8.2 17.3 Ubiquitin-40S ribosomal protein P62979 11 140 5.1 8.2 18.0 S27a P62158 Calmodulin 3 10 5.1 8.2 16.8 P05388 60S acidic ribosomal protein P0 4 9 5.6 8.2 34.3 P60709 Actin, cytoplasmic 1 22 1101 5.3 8.1 41.7 SH3 domain-binding glutamic Q9H299 5 15 5.3 8.1 10.4 acid-rich-like protein 3 O00410 Importin-5 8 14 7.9 123.5 P62857 40S ribosomal protein S28 2 2 4.5 7.9 7.8 Translationally-controlled tumor P13693 6 37 5.7 7.8 19.6 protein P23528 Cofilin-1 15 332 5.4 7.7 18.5 P09493 Tropomyosin alpha-1 chain 15 109 5.3 7.6 32.7 P05783 Keratin, type I cytoskeletal 18 9 26 5.2 7.5 48.0 P30044 Peroxiredoxin-5, mitochondrial 4 10 5.1 7.4 22.1 P08670 Vimentin 34 567 5.2 7.3 53.6 Glyceraldehyde-3-phosphate P04406 20 907 4.2 7.3 36.0 dehydrogenase F-actin-capping protein subunit P47756 6 20 4.9 7.3 31.3 beta P67936 Tropomyosin alpha-4 chain 15 108 5.2 7.2 28.5 P13797 Plastin-3 20 73 4.6 7.2 70.8 Rab GDP dissociation inhibitor P50395 21 173 4.5 7.2 50.6 beta Q06830 Peroxiredoxin-1 9 36 4.8 7.2 22.1 P07951 Tropomyosin beta chain 16 112 5.1 7.1 32.8 Peptidyl-prolyl cis-trans P62937 19 265 5.1 7.1 18.0 isomerase A P42574 Caspase-3 2 3 4.6 7.0 31.6 P07195 L-lactate dehydrogenase B chain 16 115 4.7 6.9 36.6 P31946 14-3-3 protein beta/alpha 16 170 5.1 6.9 28.1

300

Q14847 LIM and SH3 domain protein 1 9 16 4.8 6.9 29.7 P00568 Adenylate kinase isoenzyme 1 6 14 5.5 6.9 21.6 P60981 Destrin 10 178 4.9 6.9 18.5 P08729 Keratin, type II cytoskeletal 7 14 48 5.0 6.8 51.4 P04792 Heat shock protein beta-1 14 46 4.8 6.8 22.8 Macrophage migration inhibitory P14174 4 33 4.4 6.8 12.5 factor P26641 Elongation factor 1-gamma 8 17 4.7 6.8 50.1 P06703 Protein S100-A6 3 142 5.1 6.7 10.2 Q05682 Caldesmon 15 88 4.9 6.7 93.2 Ubiquitin carboxyl-terminal P09936 15 263 4.8 6.7 24.8 hydrolase isozyme L1 Isocitrate dehydrogenase O75874 10 35 4.6 6.6 46.6 [NADP] cytoplasmic P10768 S-formylglutathione hydrolase 3 4 5.0 6.6 31.4 Glyoxalase domain-containing Q9HC38 2 3 4.4 6.6 34.8 protein 4 P26038 Moesin 31 201 4.6 6.6 67.8 Q14019 Coactosin-like protein 7 20 4.8 6.6 15.9 Protein phosphatase 1 regulatory Q96C90 2 2 6.6 15.9 subunit 14B P30041 Peroxiredoxin-6 6 31 4.4 6.5 25.0 P13639 Elongation factor 2 21 79 4.7 6.5 95.3 P09211 Glutathione S-transferase P 11 90 4.5 6.5 23.3 P00338 L-lactate dehydrogenase A chain 29 412 4.4 6.5 36.7 P30085 UMP-CMP kinase 6 31 4.7 6.5 22.2 P13489 Ribonuclease inhibitor 7 12 4.8 6.5 49.9 Q9UGI8 Testin 3 4 5.1 6.4 48.0 P07737 Profilin-1 9 148 4.6 6.4 15.0 P20618 Proteasome subunit beta type-1 4 10 4.7 6.4 26.5 P51911 Calponin-1 7 12 4.4 6.4 33.1 Ubiquitin-conjugating enzyme E2 Q15819 2 8 4.6 6.4 16.4 variant 2 P00558 Phosphoglycerate kinase 1 28 229 5.0 6.4 44.6 Q15843 NEDD8 2 2 4.5 6.4 9.1 P61204 ADP-ribosylation factor 3 4 10 3.9 6.4 20.6 Q99497 Protein DJ-1 9 55 4.4 6.4 19.9 P00441 Superoxide dismutase [Cu-Zn] 8 58 4.9 6.3 15.9 40S ribosomal protein S4, X P62701 3 6 4.5 6.3 29.6 isoform Q9Y617 Phosphoserine aminotransferase 14 74 4.5 6.3 40.4 Branched-chain-amino-acid P54687 3 3 4.2 6.3 42.9 aminotransferase, cytosolic Ubiquitin-conjugating enzyme E2 P68036 6 10 4.6 6.2 17.9 L3

301

Neuroblast differentiation- Q09666 9 31 4.7 6.2 628.7 associated protein AHNAK P58546 Myotrophin 3 6 4.9 6.2 12.9 P18206 Vinculin 59 371 4.5 6.2 123.7 Microtubule-associated protein P46821 12 19 4.5 6.2 270.5 1B Glutathione S-transferase P78417 5 9 4.6 6.2 27.5 omega-1 P27348 14-3-3 protein theta 17 107 4.6 6.1 27.7 P68104 Elongation factor 1-alpha 1 15 136 4.0 6.1 50.1 P21266 Glutathione S-transferase Mu 3 8 24 4.8 6.1 26.5 P80723 Brain acid soluble protein 1 5 7 5.1 6.1 22.7 O75711 Scrapie-responsive protein 1 2 2 5.5 6.1 11.1 P63104 14-3-3 protein zeta/delta 19 209 4.4 6.1 27.7 P51858 Hepatoma-derived growth factor 2 4 4.1 6.0 26.8 Q04917 14-3-3 protein eta 9 73 4.5 6.0 28.2 Serine/threonine-protein phosphatase 2A 65 kDa P30153 8 19 4.1 6.0 65.3 regulatory subunit A alpha isoform P18085 ADP-ribosylation factor 4 3 8 4.0 6.0 20.5 Q9BVA1 Tubulin beta-2B chain 17 204 3.9 6.0 49.9 P61981 14-3-3 protein gamma 18 187 4.3 5.9 28.3 Nicotinamide N- P40261 8 41 4.3 5.9 29.6 methyltransferase Farnesyl pyrophosphate P14324 3 3 4.5 5.9 48.2 synthase P60842 Eukaryotic initiation factor 4A-I 9 25 4.5 5.9 46.1 P09382 Galectin-1 9 167 4.5 5.9 14.7 P01033 Metalloproteinase inhibitor 1 8 193 5.1 5.9 23.2 P06733 Alpha-enolase 29 1110 4.5 5.9 47.1 P18669 Phosphoglycerate mutase 1 16 178 4.3 5.9 28.8 P68363 Tubulin alpha-1B chain 22 176 4.3 5.9 50.1 Q53FA7 Quinone oxidoreductase PIG3 2 2 4.1 5.9 35.5 Puromycin-sensitive P55786 16 28 4.0 5.9 103.2 aminopeptidase Malate dehydrogenase, P40925 11 36 4.1 5.9 36.4 cytoplasmic P61970 Nuclear transport factor 2 4 19 4.5 5.8 14.5 Q16658 Fascin 15 57 4.7 5.8 54.5 Protein phosphatase Q9Y570 2 6 4.3 5.8 42.3 methylesterase 1 P62258 14-3-3 protein epsilon 15 127 4.3 5.8 29.2 Dihydropyrimidinase-related Q14195 16 38 4.5 5.8 61.9 protein 3 Cullin-associated NEDD8- Q86VP6 18 38 4.3 5.8 136.3 dissociated protein 1 302

P61158 Actin-related protein 3 9 31 4.1 5.8 47.3 P12814 Alpha-actinin-1 56 921 4.0 5.7 103.0 P06744 Glucose-6-phosphate isomerase 16 102 4.5 5.7 63.1 Fructose-bisphosphate aldolase P04075 20 357 4.1 5.7 39.4 A Q9BQE3 Tubulin alpha-1C chain 19 158 4.5 5.7 49.9 S-phase kinase-associated P63208 3 3 4.4 5.7 18.6 protein 1 Q15404 Ras suppressor protein 1 7 17 4.2 5.7 31.5 P21333 Filamin-A 104 619 4.3 5.7 280.6 P30101 Protein disulfide-isomerase A3 15 56 3.8 5.7 56.7 P17655 Calpain-2 catalytic subunit 6 13 4.9 5.7 79.9 SH3 domain-binding glutamic O75368 5 13 4.4 5.6 12.8 acid-rich-like protein Q15942 Zyxin 6 10 3.8 5.6 61.2 Phosphoacetylglucosamine O95394 5 14 4.0 5.6 59.8 mutase P28066 Proteasome subunit alpha type-5 2 2 4.5 5.6 26.4 O95782 AP-2 complex subunit alpha-1 2 2 4.4 5.6 107.5 Thioredoxin domain-containing Q9BRA2 5 13 4.6 5.6 13.9 protein 17 P25786 Proteasome subunit alpha type-1 5 8 3.5 5.6 29.5 Q9ULV4 Coronin-1C 8 26 4.4 5.6 53.2 Myristoylated alanine-rich C- P29966 4 4 4.1 5.6 31.5 kinase substrate O43707 Alpha-actinin-4 54 525 4.0 5.5 104.8 Q9NTK5 Obg-like ATPase 1 3 6 3.7 5.5 44.7 Peptidyl-prolyl cis-trans P62942 2 11 4.3 5.5 11.9 isomerase FKBP1A Ubiquitin-conjugating enzyme E2 P61088 5 12 3.7 5.5 17.1 N P36871 Phosphoglucomutase-1 11 34 3.9 5.5 61.4 Heterogeneous nuclear P61978 2 2 4.0 5.5 50.9 ribonucleoprotein K High mobility group protein P52926 3 3 4.5 5.5 11.8 HMGI-C P32119 Peroxiredoxin-2 5 15 4.0 5.5 21.9 Q9UBG0 C-type mannose receptor 2 7 11 4.4 5.4 166.6 P14625 Endoplasmin 14 69 3.7 5.4 92.4 Peptidyl-prolyl cis-trans P23284 8 97 3.8 5.4 23.7 isomerase B Four and a half LIM domains Q13642 11 45 4.6 5.4 36.2 protein 1 P07602 Prosaposin 8 71 4.0 5.4 58.1 P60174 Triosephosphate isomerase 15 390 3.9 5.4 30.8 P26022 Pentraxin-related protein PTX3 9 34 4.3 5.4 41.9 P60660 Myosin light polypeptide 6 6 24 4.3 5.4 16.9 303

Myosin light chain kinase, Q15746 2 3 4.5 5.4 210.6 smooth muscle P35237 Serpin B6 4 7 4.0 5.3 42.6 P10599 Thioredoxin 7 28 4.0 5.3 11.7 Tryptophan--tRNA ligase, P23381 11 38 3.8 5.3 53.1 cytoplasmic P61769 Beta-2-microglobulin 4 35 5.1 5.3 13.7 Heterogeneous nuclear P22626 7 21 3.6 5.3 37.4 ribonucleoproteins A2/B1 P16035 Metalloproteinase inhibitor 2 7 20 4.4 5.3 24.4 P15121 Aldose reductase 3 5 4.3 5.3 35.8 Cysteine and glycine-rich protein P21291 11 56 4.3 5.3 20.6 1 Chloride intracellular channel O00299 10 34 4.3 5.2 26.9 protein 1 Alcohol dehydrogenase P14550 9 27 4.0 5.2 36.5 [NADP(+)] O75083 WD repeat-containing protein 1 26 76 4.1 5.2 66.2 Fatty acid-binding protein, Q01469 3 8 3.5 5.2 15.2 epidermal Ubiquitin-like modifier-activating P22314 15 29 3.8 5.2 117.8 enzyme 1 P61960 Ubiquitin-fold modifier 1 2 2 4.3 5.2 9.1 P53396 ATP-citrate synthase 10 20 3.8 5.2 120.8 Heat shock cognate 71 kDa P11142 30 230 3.7 5.1 70.9 protein UTP--glucose-1-phosphate Q16851 5 11 3.6 5.1 56.9 uridylyltransferase Q9Y490 Talin-1 36 84 3.9 5.1 269.6 14 kDa phosphohistidine Q9NRX4 5 10 3.6 5.1 13.8 phosphatase P52565 Rho GDP-dissociation inhibitor 1 3 10 4.7 5.1 23.2 Q01105 Protein SET 2 2 5.1 33.5 P68371 Tubulin beta-4B chain 18 194 3.6 5.1 49.8 Q96FW1 Ubiquitin thioesterase OTUB1 3 5 3.7 5.1 31.3 P27797 Calreticulin 13 53 3.6 5.1 48.1 P27816 Microtubule-associated protein 4 8 17 3.6 5.1 120.9 P31949 Protein S100-A11 5 66 4.3 5.1 11.7 Q92743 Serine protease HTRA1 11 84 4.6 5.1 51.3 Phosphatidylethanolamine- P30086 11 43 3.8 5.1 21.0 binding protein 1 Actin-related protein 2/3 complex O15144 10 25 3.7 5.1 34.3 subunit 2 P14618 Pyruvate kinase PKM 37 818 3.9 5.1 57.9 P24534 Elongation factor 1-beta 4 8 3.5 5.0 24.7 Q15417 Calponin-3 7 13 3.8 5.0 36.4 P05387 60S acidic ribosomal protein P2 3 4 3.6 5.0 11.7

304

Q14974 Importin subunit beta-1 12 23 4.1 5.0 97.1 1,4-alpha-glucan-branching Q04446 3 4 4.4 5.0 80.4 enzyme Q99584 Protein S100-A13 3 5 3.8 5.0 11.5 Adenylyl cyclase-associated Q01518 14 50 3.7 5.0 51.9 protein 1 Acetyl-CoA acetyltransferase, Q9BWD1 2 15 3.7 5.0 41.3 cytosolic Heterogeneous nuclear P09651 5 14 3.6 5.0 38.7 ribonucleoprotein A1 P14314 Glucosidase 2 subunit beta 3 5 3.6 5.0 59.4 Q92820 Gamma-glutamyl hydrolase 4 11 4.1 5.0 35.9 Q12765 Secernin-1 4 7 3.8 5.0 46.4 Fructose-bisphosphate aldolase P09972 8 54 3.6 4.9 39.4 C 6-phosphogluconate P52209 15 46 3.9 4.9 53.1 dehydrogenase, decarboxylating P98095 Fibulin-2 4 8 4.0 4.9 126.5 O43852 Calumenin 13 38 4.0 4.9 37.1 Staphylococcal nuclease Q7KZF4 6 15 3.7 4.9 101.9 domain-containing protein 1 P12109 Collagen alpha-1(VI) chain 34 285 4.5 4.9 108.5 Q9NVA2 Septin-11 4 8 4.0 4.9 49.4 EGF-like repeat and discoidin I- O43854 5 9 4.4 4.9 53.7 like domain-containing protein 3 P07858 Cathepsin B 3 5 3.8 4.9 37.8 P04080 Cystatin-B 4 54 4.0 4.9 11.1 P25787 Proteasome subunit alpha type-2 6 14 3.7 4.9 25.9 Serine/threonine-protein Q14738 phosphatase 2A 56 kDa 4 5 3.5 4.9 69.9 regulatory subunit delta isoform P62906 60S ribosomal protein L10a 2 4 3.5 4.9 24.8 O76061 Stanniocalcin-2 6 78 5.1 4.9 33.2 P28300 Protein-lysine 6-oxidase 2 4 4.0 4.9 46.9 Q9H4A4 Aminopeptidase B 4 9 3.9 4.8 72.5 Eukaryotic translation initiation P63241 3 216 3.6 4.8 16.8 factor 5A-1 P29401 Transketolase 15 51 3.6 4.8 67.8 Spectrin alpha chain, non- Q13813 6 9 3.7 4.8 284.4 erythrocytic 1 Glycogen phosphorylase, brain P11216 4 5 3.9 4.8 96.6 form P41250 Glycine--tRNA ligase 16 95 3.6 4.8 83.1 P12277 Creatine kinase B-type 4 15 3.0 4.8 42.6 P37837 Transaldolase 5 16 3.4 4.8 37.5 Q14315 Filamin-C 23 119 3.9 4.8 290.8 P60900 Proteasome subunit alpha type-6 3 3 3.5 4.8 27.4

305

Q00610 Clathrin heavy chain 1 30 73 3.8 4.8 191.5 P12955 Xaa-Pro dipeptidase 2 2 3.4 4.8 54.5 O60565 Gremlin-1 2 2 4.4 4.8 20.7 P02511 Alpha-crystallin B chain 5 29 3.5 4.7 20.1 P62241 40S ribosomal protein S8 3 8 3.6 4.7 24.2 P34932 Heat shock 70 kDa protein 4 9 16 3.5 4.7 94.3 P02545 Prelamin-A/C 7 11 3.7 4.7 74.1 P11021 78 kDa glucose-regulated protein 14 70 3.4 4.7 72.3 P08238 Heat shock protein HSP 90-beta 23 124 3.7 4.7 83.2 Ras GTPase-activating-like P46940 21 54 3.6 4.7 189.1 protein IQGAP1 Thioredoxin domain-containing Q8NBS9 7 8 3.5 4.6 47.6 protein 5 Cytoplasmic dynein 1 heavy Q14204 9 12 3.8 4.6 532.1 chain 1 Dihydropyrimidinase-related Q16555 8 13 3.5 4.6 62.3 protein 2 P28074 Proteasome subunit beta type-5 3 9 3.5 4.6 28.5 P35579 Myosin-9 54 223 3.8 4.6 226.4 Chloride intracellular channel Q9Y696 15 184 4.0 4.6 28.8 protein 4 BTB/POZ domain-containing Q96CX2 3 5 4.0 4.6 35.7 protein KCTD12 Q02818 Nucleobindin-1 13 46 4.6 4.6 53.8 Q15084 Protein disulfide-isomerase A6 4 6 3.2 4.6 48.1 ATP-dependent 6- Q01813 phosphofructokinase, platelet 3 4 3.1 4.5 85.5 type O00391 Sulfhydryl oxidase 1 19 116 4.5 4.5 82.5 P00966 Argininosuccinate synthase 11 26 3.4 4.5 46.5 Q15149 Plectin 22 44 3.6 4.5 531.5 P61160 Actin-related protein 2 8 37 3.6 4.5 44.7 Rab GDP dissociation inhibitor P31150 14 129 3.6 4.5 50.6 alpha Transitional endoplasmic P55072 2 2 3.1 4.4 89.3 reticulum ATPase P12110 Collagen alpha-2(VI) chain 13 34 4.2 4.4 108.5 P20908 Collagen alpha-1(V) chain 30 116 4.0 4.4 183.4 P07237 Protein disulfide-isomerase 16 77 3.5 4.4 57.1 LIM and cysteine-rich domains Q9NZU5 7 14 3.3 4.4 40.8 protein 1 P39019 40S ribosomal protein S19 4 6 3.2 4.4 16.1 Programmed cell death 6- Q8WUM4 3 3 3.7 4.4 96.0 interacting protein P02462 Collagen alpha-1(IV) chain 18 86 3.7 4.3 160.5

306

Immunoglobulin superfamily O14498 containing leucine-rich repeat 6 20 3.7 4.3 46.0 protein Protein-glutamine gamma- P21980 6 14 3.7 4.3 77.3 glutamyltransferase 2 P09960 Leukotriene A-4 hydrolase 13 30 3.3 4.3 69.2 Heterogeneous nuclear Q14103 4 8 3.4 4.3 38.4 ribonucleoprotein D0 P12111 Collagen alpha-3(VI) chain 45 124 3.9 4.3 343.5 Peptidyl-prolyl cis-trans Q96AY3 6 16 3.4 4.3 64.2 isomerase FKBP10 Serine/arginine-rich splicing Q07955 2 32 3.4 4.3 27.7 factor 1 Latent-transforming growth factor Q14767 26 82 4.0 4.3 194.9 beta-binding protein 2 P07355 Annexin A2 16 54 3.5 4.3 38.6 P08107 Heat shock 70 kDa protein 1A/1B 9 20 3.3 4.2 70.0 O75369 Filamin-B 36 125 3.5 4.2 278.0 P11766 Alcohol dehydrogenase class-3 3 3 3.3 4.2 39.7 Cyclin-dependent kinase inhibitor P42771 2 2 4.1 4.2 16.5 2A, isoforms 1/2/3 P06396 Gelsolin 29 160 3.4 4.2 85.6 O95965 Integrin beta-like protein 1 4 7 3.7 4.2 53.9 Acidic leucine-rich nuclear P39687 phosphoprotein 32 family 3 7 3.4 4.1 28.6 member A Heat shock protein HSP 90- P07900 24 149 3.4 4.1 84.6 alpha Q13219 Pappalysin-1 21 49 3.8 4.1 180.9 P04083 Annexin A1 7 35 3.4 4.1 38.7 P13611 Versican core protein 16 54 3.2 4.1 372.6 P50454 Serpin H1 17 67 2.9 4.1 46.4 P19022 Cadherin-2 6 10 3.2 4.0 99.7 Q9BY76 Angiopoietin-related protein 4 3 7 2.9 4.0 45.2 Hyaluronan and proteoglycan P10915 15 127 3.9 4.0 40.1 link protein 1 Four and a half LIM domains Q14192 6 17 3.5 4.0 32.2 protein 2 P07093 Glia-derived nexin 13 100 3.7 4.0 44.0 P05452 Tetranectin 5 15 3.8 4.0 22.5 Q96D15 Reticulocalbin-3 2 3 3.3 4.0 37.5 P49720 Proteasome subunit beta type-3 3 3 3.3 4.0 22.9 P07585 Decorin 14 96 3.9 4.0 39.7 P35442 Thrombospondin-2 26 113 3.8 4.0 129.9 Q9NY33 Dipeptidyl peptidase 3 5 11 3.3 4.0 82.5 Insulin-like growth factor-binding P18065 15 97 3.9 3.9 34.8 protein 2

307

N(G),N(G)-dimethylarginine O94760 4 8 3.1 3.9 31.1 dimethylaminohydrolase 1 Coiled-coil domain-containing Q76M96 16 69 3.8 3.9 108.1 protein 80 Aspartate aminotransferase, P17174 6 10 3.3 3.9 46.2 cytoplasmic P24043 Laminin subunit alpha-2 11 20 3.7 3.9 343.7 Q92626 Peroxidasin homolog 23 65 3.5 3.9 165.2 P35555 Fibrillin-1 73 446 3.2 3.9 312.0 Cysteine and glycine-rich protein Q16527 5 17 3.2 3.9 20.9 2 Q9Y3B8 Oligoribonuclease, mitochondrial 6 11 3.3 3.9 26.8 Adipocyte enhancer-binding Q8IUX7 8 13 3.5 3.9 130.8 protein 1 Q9NRN5 Olfactomedin-like protein 3 6 12 4.1 3.9 46.0 P01034 Cystatin-C 8 51 3.3 3.8 15.8 Beta-hexosaminidase subunit P07686 5 8 3.1 3.8 63.1 beta Q9UBP4 Dickkopf-related protein 3 8 34 3.6 3.8 38.4 O95084 Serine protease 23 3 4 3.5 3.8 43.0 Q9UI42 Carboxypeptidase A4 9 27 4.0 3.8 47.3 P11047 Laminin subunit gamma-1 13 35 3.0 3.8 177.5 P30043 Flavin reductase (NADPH) 2 2 3.5 3.8 22.1 Latent-transforming growth factor Q14766 14 34 3.5 3.7 186.7 beta-binding protein 1 P48163 NADP-dependent malic enzyme 4 5 3.2 3.7 64.1 Fibronectin type III domain- Q4ZHG4 11 17 2.9 3.7 205.4 containing protein 1 P08758 Annexin A5 9 24 3.1 3.7 35.9 P29279 Connective tissue growth factor 4 8 3.3 3.6 38.1 Chondroitin sulfate proteoglycan Q6UVK1 10 20 3.7 3.6 250.4 4 Q15063 Periostin 38 425 3.5 3.6 93.3 P05997 Collagen alpha-2(V) chain 36 110 3.6 3.5 144.8 P35556 Fibrillin-2 10 25 3.2 3.5 314.6 P16870 Carboxypeptidase E 9 19 3.1 3.5 53.1 P08123 Collagen alpha-2(I) chain 74 1436 3.7 3.5 129.2 P35052 Glypican-1 7 10 2.8 3.5 61.6 P07339 Cathepsin D 13 68 3.4 3.5 44.5 P41222 Prostaglandin-H2 D-isomerase 2 2 3.2 3.5 21.0 P21810 Biglycan 21 599 3.2 3.4 41.6 P02751 Fibronectin 102 4800 3.6 3.4 262.5 P08603 Complement factor H 11 20 3.1 3.4 139.0 Q9H4D0 Calsyntenin-2 10 20 3.1 3.4 106.9 P62826 GTP-binding nuclear protein Ran 7 25 2.7 3.4 24.4

308

P27658 Collagen alpha-1(VIII) chain 13 45 3.2 3.4 73.3 Threonine--tRNA ligase, P26639 6 8 3.3 83.4 cytoplasmic Transforming growth factor-beta- Q15582 28 647 3.2 3.3 74.6 induced protein ig-h3 MAM domain-containing protein Q7Z304 16 47 3.2 3.3 77.5 2 Procollagen-lysine,2- Q02809 22 117 3.8 3.3 83.5 oxoglutarate 5-dioxygenase 1 O00468 Agrin 14 29 2.9 3.3 217.1 P15502 Elastin 6 18 3.5 3.3 68.4 P02461 Collagen alpha-1(III) chain 44 177 3.3 3.3 138.5 Insulin-like growth factor-binding P24593 9 48 3.3 3.3 30.6 protein 5 P04264 Keratin, type II cytoskeletal 1 13 42 2.6 3.3 66.0 P13667 Protein disulfide-isomerase A4 2 3 2.8 3.3 72.9 P08572 Collagen alpha-2(IV) chain 43 144 3.0 3.3 167.4 Insulin-like growth factor-binding P17936 11 38 3.2 3.3 31.7 protein 3 Q12841 Follistatin-related protein 1 22 590 3.5 3.2 35.0 P13645 Keratin, type I cytoskeletal 10 9 12 2.6 3.2 58.8 P28799 Granulins 5 12 3.0 3.2 63.5 Q16610 Extracellular matrix protein 1 10 27 3.0 3.2 60.6 O94985 Calsyntenin-1 7 24 2.8 3.2 109.7 Q08431 Lactadherin 4 18 2.7 3.1 43.1 Microfibril-associated P55083 4 52 3.5 3.1 28.6 glycoprotein 4 Q9Y4K0 Lysyl oxidase homolog 2 9 38 2.7 3.1 86.7 Insulin-like growth factor-binding Q16270 21 879 2.9 3.1 29.1 protein 7 P02452 Collagen alpha-1(I) chain 86 1758 3.2 3.1 138.9 Beta-hexosaminidase subunit P06865 5 11 2.5 3.1 60.7 alpha P12107 Collagen alpha-1(XI) chain 15 47 3.0 3.1 181.0 P05121 Plasminogen activator inhibitor 1 24 1059 3.1 3.1 45.0 Q9UBX5 Fibulin-5 5 12 2.6 3.0 50.1 Q14118 Dystroglycan 4 8 3.1 3.0 97.4 P10909 Clusterin 9 86 2.9 3.0 52.5 P07996 Thrombospondin-1 46 364 3.8 3.0 129.3 Secreted frizzled-related protein Q96HF1 9 22 3.0 3.0 33.5 2 P09486 SPARC 20 3097 3.2 3.0 34.6 Q16363 Laminin subunit alpha-4 3 7 2.4 3.0 202.4 Q16777 Histone H2A type 2-C 2 2 2.6 2.9 14.0 P08476 Inhibin beta A chain 16 35 3.0 2.9 47.4

309

Chitinase domain-containing Q9BWS9 3 13 2.2 2.9 44.9 protein 1 Polypeptide N- Q10471 acetylgalactosaminyltransferase 6 16 2.7 2.9 64.7 2 Q99715 Collagen alpha-1(XII) chain 59 175 2.8 2.9 332.9 Serine/threonine-protein P36873 phosphatase PP1-gamma 2 4 2.4 2.9 37.0 catalytic subunit P07942 Laminin subunit beta-1 17 42 2.8 2.8 197.9 Nucleosome assembly protein 1- P55209 3 5 2.3 2.8 45.3 like 1 Keratin, type II cytoskeletal 2 P35908 9 15 2.3 2.8 65.4 epidermal O95678 Keratin, type II cytoskeletal 75 4 8 2.1 2.7 59.5 P08253 72 kDa type IV collagenase 23 74 2.5 2.7 73.8 P35527 Keratin, type I cytoskeletal 9 7 38 2.1 2.6 62.0 Q08380 Galectin-3-binding protein 21 128 2.8 2.6 65.3 EGF-containing fibulin-like Q12805 17 75 2.7 2.6 54.6 extracellular matrix protein 1 P24821 Tenascin 35 150 2.9 2.6 240.7 Basement membrane-specific P98160 heparan sulfate proteoglycan 23 99 2.3 2.5 468.5 core protein EGF-containing fibulin-like O95967 3 6 2.5 2.5 49.4 extracellular matrix protein 2 P51884 Lumican 12 75 2.3 2.3 38.4 P01023 Alpha-2-macroglobulin 11 13 1.6 2.0 163.2 P02768 Serum albumin 15 81 1.7 1.9 69.3 P02765 Alpha-2-HS-glycoprotein 2 3 1.9 39.3 Glutamine--fructose-6-phosphate Q06210 3 5 78.8 aminotransferase [isomerizing] 1 Inter-alpha-trypsin inhibitor heavy Q86UX2 2 18 104.5 chain H5 P48637 Glutathione synthetase 5 13 52.4 Q9P2E9 Ribosome-binding protein 1 3 6 152.4 Procollagen-lysine,2- O00469 6 6 84.6 oxoglutarate 5-dioxygenase 2 C-type lectin domain family 11 Q9Y240 2 2 35.7 member A Q96FQ6 Protein S100-A16 3 7 11.8 P28070 Proteasome subunit beta type-4 4 11 29.2 Bifunctional glutamate/proline-- P07814 2 4 170.5 tRNA ligase P62081 40S ribosomal protein S7 3 4 22.1 P62269 40S ribosomal protein S18 2 3 17.7 Platelet-activating factor P68402 3 11 25.6 acetylhydrolase IB subunit beta Q13308 Inactive tyrosine-protein kinase 7 7 12 118.3 310

Isopentenyl-diphosphate Delta- Q13907 2 4 26.3 isomerase 1 P46781 40S ribosomal protein S9 2 2 22.6 Q13509 Tubulin beta-3 chain 14 169 50.4 Q05707 Collagen alpha-1(XIV) chain 2 2 193.4 Glycogen phosphorylase, liver P06737 4 4 97.1 form Nuclease-sensitive element- P67809 4 8 35.9 binding protein 1 Guanine nucleotide-binding P62873 protein G(I)/G(S)/G(T) subunit 2 2 37.4 beta-1 26S proteasome non-ATPase Q13200 4 7 100.1 regulatory subunit 2 P55060 Exportin-2 2 2 110.3 P00491 Purine nucleoside phosphorylase 2 4 32.1 Q6IBS0 Twinfilin-2 2 2 39.5 Q9P1F3 Costars family protein ABRACL 2 2 9.1 Spectrin beta chain, non- Q01082 3 3 274.4 erythrocytic 1 Guanine nucleotide-binding P63244 2 2 35.1 protein subunit beta-2-like 1 Myosin regulatory light P24844 6 20 19.8 polypeptide 9 P02795 Metallothionein-2 3 10 6.0 Q15181 Inorganic pyrophosphatase 4 6 3.2 32.6 N(G),N(G)-dimethylarginine O95865 3 8 3.0 29.6 dimethylaminohydrolase 2 Q14697 Neutral alpha-glucosidase AB 4 6 3.8 106.8 Cytosolic non-specific Q96KP4 3 4 3.6 52.8 dipeptidase Actin-related protein 2/3 complex P59998 3 3 4.1 19.7 subunit 4 O14818 Proteasome subunit alpha type-7 4 5 3.4 27.9 Asparagine--tRNA ligase, O43776 4 7 4.0 62.9 cytoplasmic P09104 Gamma-enolase 7 539 3.4 47.2

Table 5.8: Differentially released conditioned medium proteins at 48 hours

Astrocyte-released proteins measured in condition media from TMT experiments comparing protein abundance between mild and severe stretched cells in conditioned media collected 48 hours post-injury.

311

5.7 REFERENCES

1. S. C. Choi, R. Bullock, Design and statistical issues in multicenter trials of severe

head injury. Neurological research 23, 190-192 (2001); published online EpubMar-

Apr (10.1179/016164101101198325).

2. J. A. Langlois, W. Rutland-Brown, M. M. Wald, The Epidemiology and Impact of

Traumatic Brain Injury. Journal of Head Trauma Rehabilitation 21, 375-378 (2006).

3. G. Franz, R. Beer, A. Kampfl, K. Engelhardt, E. Schmutzhard, H. Ulmer, F.

Deisenhammer, Amyloid beta 1–42 and tau in cerebrospinal fluid after severe

traumatic brain injury. Neurology 60, (2003).

4. N. Shiiya, T. Kunihara, T. Miyatake, K. Matsuzaki, K. Yasuda, Tau protein in the

cerebrospinal fluid is a marker of brain injury after aortic surgery. Ann Thorac Surg

77, (2004).

5. P. P. Tsitsopoulos, N. Marklund, Amyloid-beta peptides and tau protein as

biomarkers in cerebrospinal and interstitial fluid following traumatic brain injury: a

review of experimental and clinical studies. Front Neurol 4, (2013).

6. I. B. Wanner, M. A. Anderson, B. Song, J. Levine, A. Fernandez, Z. Gray-

Thompson, Y. Ao, M. V. Sofroniew, Glial scar borders are formed by newly

proliferated, elongated astrocytes that interact to corral inflammatory and fibrotic

cells via STAT3-dependent mechanisms after spinal cord injury. J Neurosci 33,

(2013).

7. J. E. Burda, M. V. Sofroniew, Reactive gliosis and the multicellular response to

CNS damage and disease. Neuron 81, (2014).

312

8. I. B. Wanner, An in vitro trauma model to study rodent and human astrocyte

reactivity. Methods in molecular biology 814, (2012).

9. I. B. Wanner, M. Deik, M. Torres, A. R. Rosendahl, J. T. Neary, V. P. Lemmon, J.

L. Bixby, A new in vitro model of the glial scar inhibits axon growth. Glia 56, (2008).

10. M. Sondej, P. Doran, J. A. Loo, I. Wanner, in Sample preparation in biological mass

spectrometry, A. Ivanov, A. Lazarev, Eds. (Springer, Dordrecht, 2011).

11. E. F. Ellis, J. S. McKinney, K. A. Willoughby, S. Liang, J. T. Povlishock, A new

model for rapid stretch-induced injury of cells in culture: characterization of the

model using astrocytes. Journal of neurotrauma 12, (1995).

12. E. F. Ellis, K. A. Willoughby, S. A. Sparks, T. Chen, S100B protein is released from

rat neonatal neurons, astrocytes, and microglia by in vitro trauma and anti-S100

increases trauma-induced delayed neuronal injury and negates the protective

effect of exogenous S100B on neurons. J Neurochem 101, (2007).

13. A. Sandberg, R. M. Branca, J. Lehtio, J. Forshed, Quantitative accuracy in mass

spectrometry based proteomics of complex samples: the impact of labeling and

precursor interference. Journal of proteomics 96, 133-144 (2014); published online

EpubJan 16 (10.1016/j.jprot.2013.10.035).

14. H. Wang, S. Alvarez, L. M. Hicks, Comprehensive comparison of iTRAQ and label-

free LC-based quantitative proteomics approaches using two Chlamydomonas

reinhardtii strains of interest for biofuels engineering. Journal of proteome research

11, 487-501 (2012); published online EpubJan 1 (10.1021/pr2008225).

313

15. M. Bantscheff, M. Boesche, D. Eberhard, T. Matthieson, G. Sweetman, B. Kuster,

Robust and Sensitive iTRAQ Quantification on an LTQ Orbitrap Mass

Spectrometer. Molecular & Cellular Proteomics 7, 1702-1713 (2008)10.1074/).

16. E. H. Pettus, C. W. Christman, M. L. Giebel, J. T. Povlishock, Traumatically

Induced Altered Membrane Permeability: Its Relationship to Traumatically Induced

Reactive Axonal Change. Journal of Neurotrauma 11, 507-522 (1994).

17. O. Farkas, J. Lifshitz, J. T. Povlishock, Mechanoporation Induced by Diffuse

Traumatic Brain Injury: An Irreversible or Reversible Response to Injury? The

Journal of Neuroscience 25, 3130-3140 (2006).

18. S. Schenk, G. J. Schoenhals, G. de Souza, M. Mann, A high confidence, manually

validated human blood plasma protein reference set. BMC medical genomics 1,

41 (2008); published online EpubSep 15 (10.1186/1755-8794-1-41).

19. V. Nanjappa, J. K. Thomas, A. Marimuthu, B. Muthusamy, A. Radhakrishnan, R.

Sharma, A. Ahmad Khan, L. Balakrishnan, N. A. Sahasrabuddhe, S. Kumar, B. N.

Jhaveri, K. V. Sheth, R. Kumar Khatana, P. G. Shaw, S. M. Srikanth, P. P. Mathur,

S. Shankar, D. Nagaraja, R. Christopher, S. Mathivanan, R. Raju, R. Sirdeshmukh,

A. Chatterjee, R. J. Simpson, H. C. Harsha, A. Pandey, T. S. Prasad, Plasma

Proteome Database as a resource for proteomics research: 2014 update. Nucleic

acids research 42, D959-965 (2014); published online EpubJan

(10.1093/nar/gkt1251).

20. M. J. Whalen, T. Dalkara, Z. You, J. Qiu, D. Bermpohl, N. Mehta, B. Suter, P. G.

Bhide, E. H. Lo, M. Ericsson, M. A. Moskowitz, Acute plasmalemma permeability

and protracted clearance of injured cells after controlled cortical impact in mice.

314

Journal of cerebral blood flow and metabolism : official journal of the International

Society of Cerebral Blood Flow and Metabolism 28, 490-505 (2008); published

online EpubMar (10.1038/sj.jcbfm.9600544).

21. R. H. Singleton, J. T. Povlishock, Identification and characterization of

heterogeneous neuronal injury and death in regions of diffuse brain injury:

evidence for multiple independent injury phenotypes. The Journal of neuroscience

: the official journal of the Society for Neuroscience 24, 3543-3553 (2004);

published online EpubApr 7 (10.1523/JNEUROSCI.5048-03.2004).

22. M. C. LaPlaca, G. R. Prado, Neural mechanobiology and neuronal vulnerability to

traumatic loading. Journal of Biomechanics 43, 71-78 (2010).

23. R. S. B. Clark, P. M. Kochanek, S. C. Watkins, M. Chen, C. E. Dixon, N. A.

Seidberg, J. Melick, J. E. Loeffert, P. D. Nathaniel, K. L. Jin, S. H. Graham,

Caspase-3 Mediated Neuronal Death AfterTraumatic Brain Injury in Rats. Journal

of Neurochemistry 74, 740-753 (2000).

24. A. Degterev, Z. Huang, M. Boyce, Y. Li, P. Jagtap, N. Mizushima, G. D. Cuny, T.

J. Mitchison, M. A. Moskowitz, J. Yuan, Chemical inhibitor of nonapoptotic cell

death with therapeutic potential for ischemic brain injury. Nature chemical biology

1, 112-119 (2005); published online EpubJul (10.1038/nchembio711).

25. R. S. Clark, P. M. Kochanek, P. D. Adelson, M. J. Bell, J. A. Carcillo, M. Chen, S.

R. Wisniewski, K. Janesko, M. J. Whalen, S. H. Graham, Increases in bcl-2 protein

in cerebrospinal fluid and evidence for programmed cell death in infants and

children after severe traumatic brain injury. The Journal of pediatrics 137, 197-204

(2000); published online EpubAug (10.1067/mpd.2000.106903).

315

26. A. K. Ottens, L. Bustamante, E. C. Golden, C. Yao, R. L. Hayes, K. K. Wang, F. C.

Tortella, J. R. Dave, Neuroproteomics: a biochemical means to discriminate the

extent and modality of brain injury. J Neurotrauma 27, 1837-1852 (2010);

published online EpubOct (10.1089/neu.2010.1374).

27. K. E. Saatman, K. J. F. Feeko, R. L. Pape, R. Raghupathi, Differential Behavioral

and Histopathological Responses to Graded Cortical Impact Injury in Mice. Journal

of Neurotrauma 23, 1241-1253 (2006).

28. J. M. Spaethling, D. M. Geddes-Klein, W. J. Miller, C. R. von Reyn, P. Singh, M.

Mesfin, S. J. Bernstein, D. F. Meaney, Linking impact to cellular and molecular

sequelae of CNS injury: Modeling in vivo complexity with in vitro simplicity. 161,

27-39 (2007)10.1016/s0079-6123(06)61003-0).

29. J. Y. Kim, N. Kim, Z. Zheng, J. E. Lee, M. A. Yenari, The 70 kDa heat shock protein

protects against experimental traumatic brain injury. Neurobiology of disease 58,

289-295 (2013); published online EpubOct (10.1016/j.nbd.2013.06.012).

30. J. Y. Kim, M. A. Yenari, The immune modulating properties of the heat shock

proteins after brain injury. Anatomy & cell biology 46, 1-7 (2013); published online

EpubMar (10.5115/acb.2013.46.1.1).

31. D. L. Feinstein, E. Galea, D. A. Aquino, G. C. Li, H. Xu, D. J. Reis, Heat Shock

Protein 70 Suppresses Astroglial-inducible Nitric-oxide Synthase Expression by

Decreasing NFkB Activation. The Journal of Biological Chemistry 271, 17724-

17732 (1996).

32. X. Yao, J. Liu, J. T. McCabe, Ubiquitin and ubiquitin-conjugated protein expression

in the rat cerebral cortex and hippocampus following traumatic brain injury (TBI).

316

Brain research 1182, 116-122 (2007); published online EpubNov 28

(10.1016/j.brainres.2007.08.076).

33. M. Majetschak, D. R. King, U. Krehmeier, L. T. Busby, C. Thome, S. Vajkoczy, K.

G. Proctor, Ubiquitin immunoreactivity in cerebrospinal fluid after traumatic brain

injury: Clinical and experimental findings. Critical Care Medicine 33, 1589-1594

(2005)10.1097/01.ccm.0000169883.41245.23).

34. K. Dikranian, R. Cohen, C. Mac Donald, Y. Pan, D. Brakefield, P. Bayly, A.

Parsadanian, Mild traumatic brain injury to the infant mouse causes robust white

matter axonal degeneration which precedes apoptotic death of cortical and

thalamic neurons. Experimental neurology 211, 551-560 (2008); published online

EpubJun (10.1016/j.expneurol.2008.03.012).

35. P. N. Lizhnyak, A. K. Ottens, Proteomics: in pursuit of effective traumatic brain

injury therapeutics. Expert review of proteomics 12, 75-82 (2015); published online

EpubFeb (10.1586/14789450.2015.1000869).

36. G. Barkhoudarian, D. A. Hovda, C. C. Giza, The molecular pathophysiology of

concussive brain injury. Clinics in sports medicine 30, 33-48, vii-iii (2011);

published online EpubJan (10.1016/j.csm.2010.09.001).

37. L. Papa, L. Akinyi, M. C. Liu, J. A. Pineda, J. J. Tepas, 3rd, M. W. Oli, W. Zheng,

G. Robinson, S. A. Robicsek, A. Gabrielli, S. C. Heaton, H. J. Hannay, J. A.

Demery, G. M. Brophy, J. Layon, C. S. Robertson, R. L. Hayes, K. K. Wang,

Ubiquitin C-terminal hydrolase is a novel biomarker in humans for severe traumatic

brain injury. Crit Care Med 38, 138-144 (2010); published online EpubJan

(10.1097/CCM.0b013e3181b788ab).

317

38. S. Mondello, A. Linnet, A. Buki, S. Robiscek, A. Gabrielli, J. Tepas, L. Papa, G. M.

Brophy, F. Tortella, R. L. Hayes, K. K. Wang, Clinical Utility of Serum Levels of

Ubiquitin C-Terminal Hydrolase as a Biomarker for Severe Traumatic Brain Injury.

Neurosurgery 70, 666-675 (2011)10.1227/NEU.0b013e318236a809).

39. E. P. Thelin, D. Just, A. Frostell, A. Haggmark-Manberg, M. Risling, M. Svensson,

P. Nilsson, B. M. Bellander, Protein profiling in serum after traumatic brain injury

in rats reveals potential injury markers. Behavioural brain research, (2016);

published online EpubAug 31 (10.1016/j.bbr.2016.08.058).

40. R. Siman, T. K. McIntosh, K. M. Soltesz, Z. Chen, R. W. Neumar, V. L. Roberts,

Proteins released from degenerating neurons are surrogate markers for acute

brain damage. Neurobiology of disease 16, 311-320 (2004); published online

EpubJul (10.1016/j.nbd.2004.03.016).

41. C. Loov, A. G. Nadadhur, L. Hillered, F. Clausen, A. Erlandsson, Extracellular

ezrin: a novel biomarker for traumatic brain injury. J Neurotrauma 32, 244-251

(2015); published online EpubFeb 15 (10.1089/neu.2014.3517).

42. C. Loov, G. Shevchenko, A. Geeyarpuram Nadadhur, F. Clausen, L. Hillered, M.

Wetterhall, A. Erlandsson, Identification of injury specific proteins in a cell culture

model of traumatic brain injury. PLoS One 8, e55983

(2013)10.1371/journal.pone.0055983).

43. M. Fujimura, T. Tominaga, P. H. Chan, Neuroprotective effect of an antioxidant in

ischemic brain injury. Neurocritical care 2, 59-66 (2005).

44. H. Endo, C. Nito, H. Kamada, F. Yu, P. H. Chan, Reduction in oxidative stress by

superoxide dismutase overexpression attenuates acute brain injury after

318

subarachnoid hemorrhage via activation of Akt/glycogen synthase kinase-3β

survival signaling. Journal of Cerebral Blood Flow & Metabolism 27, 975-982

(2007).

45. I. B. Wanner, An in vitro trauma model to study rodent and human astrocyte

reactivity. Methods in molecular biology 814, 189-219 (2012)10.1007/978-1-

61779-452-0_14).

46. K. Hodge, S. T. Have, L. Hutton, A. I. Lamond, Cleaning up the masses: exclusion

lists to reduce contamination with HPLC-MS/MS. Journal of proteomics 88, 92-103

(2013); published online EpubAug 2 (10.1016/j.jprot.2013.02.023).

47. G. S. Omenn, D. J. States, M. Adamski, T. W. Blackwell, R. Menon, H. Hermjakob,