Dysfunctional AMPA receptor trafficking in traumatic brain injury by. Joshua David Bell. A thesis submitted in conformity with the requirements for the degree of.
Dysfunctional AMPA receptor trafficking in traumatic brain injury by Joshua David Bell
A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy, Institute of Medical Science University of Toronto © Copyright by Joshua David Bell, 2010
Dysfunctional AMPA receptor trafficking in traumatic brain injury Joshua D Bell, Institute of Medical Science University of Toronto: Doctor of Philosophy; 2010
Abstract Traumatic brain injury (TBI) is a devastating public health problem for patients and their families. The neurodegeneration that follows TBI is complex, but can be broadly subdivided into primary and secondary damage. Though primary damage is irreversible and therefore unsalvageable, extensive literature aimed at understanding the tissue, cellular, inflammatory and subcellular processes following the injury have proven unequivocally that secondary pathophysiological events are delayed and progressive in nature. Understanding these secondary events at the cellular levels is critical in the eventual establishment of targeted therapeutics aimed at limiting progressive injury after an initial trauma. One such secondary event is referred to in the literature as excitotoxicity; a sustained and de-regulated activation of glutamate receptors that leads to rapid cytotoxic edema and calcium overload. Our understanding of excitotoxicity has evolved to include not only a role for elevated extracellular glutamate in mediating neuronal damage, but also post-synaptic receptor modifications that render glutamate profoundly more toxic to injured neurons than healthy tissue. In this thesis, we explored the hypothesis that glutamate excitotoxicity can be perpetuated by trauma-induced post-synaptic modification of the AMPA receptor. Specifically, we used a cortical culture model of TBI as well as the fluid percussion
ii
injury device to test the hypothesis that TBI confers a reduction of surface GluR2 protein, an AMPA receptor subunit that limits neuronal calcium permeability. We conjectured that this decrease in the expression of surface GluR2 would increase the expression of calcium-permeable AMPA receptors, thereby rendering neurons vulnerable to secondary excitotoxic injury. We further investigated the subcellular mechanisms responsible for the internalization of surface GluR2, and the phenotypic consequences of GluR2 endocytosis in both models. Our data revealed that both models of TBI resulted in a regulated signaling cascade leading to the phosphorylation and internalization of GluR2. By exogenously interrupting the trafficking of GluR2 protein with an inhibitory peptide, we further observed that GluR2 internalization was mediated by a protein interaction involving protein interacting with C kinase 1 (PICK1) and protein kinase C alpha (PKCα), two PDZ domain-containing proteins that mediate GluR2 trafficking during constitutive synaptic plasticity. We observed that GluR2 endocytosis was NMDA receptor dependent, and resulted in increased neuronal calcium permeability, augmented AMPA receptor mediated electrophysiological activity and increased susceptibility to delayed cell death. Finally, we demonstrated that the interruption of GluR2 trafficking is cytoprotective, suggesting that sustaining surface GluR2 protein protects neurons against secondary injury. Overall, our findings suggest that experimental TBI promotes the expression of injurious GluR2-lacking AMPA receptors, thereby enhancing cellular vulnerability to secondary excitotoxicity.
iii
Acknowledgements I want to thank a number of people who’ve helped me throughout my PhD program. Firstly, I want to thank my supervisor, Andrew Baker. Dr Baker has been the ideal supervisor, allowing me to pursue my own intellectual curiosities, while shaping my interests into scientifically testable hypotheses. He has supported my ideas, challenged my thinking, and helped me grow tremendously as a person and a scientist. He is a credit to the IMS graduate program and the university. Outside of the lab, Dr Baker has also fostered collaboration and connectivity with other labs, taking me to numerous conferences where I’ve been able to share my data with audiences well beyond U of T. I cannot possibly thank him enough for all that he has taught me about research and about life. I also want to acknowledge my fellow lab members. I want to thank Dr Eugene Park for his guidance and willingness to discuss ideas, but more importantly for his friendship. Dr Park’s experience in the lab made my program infinitely easier, as he shared with me all that he knew about completing a successful doctoral program in the IMS, invaluable knowledge that unquestionably contributed to my successes. I want to also acknowledge Dr Jinglu Ai, whose creativity and hard work inspired the preliminary data in the early days of this work back in 2006. Elaine Liu, our technician, is also worthy of significant thanks. Elaine has been a beacon of unwavering support in my pursuit of higher education, and made the downtime in the lab as well as lab get-togethers much more enjoyable. Finally, I want to thank Dr Carlo Santaguida. Though he arrived when my data collection was finished, his friendship and his exceedingly generous will in providing medical advice to me during my thesis writing days expedited the process of thesis completion significantly. He has been a constant alleviator of anxieties! I look forward to further collaboration with Carlo and many more tasty shawarmas. I want to thank Dr Beverley Orser, Dr Michael Fehlings, and Dr Peter Pennefather for their participation in my Program Advisory Committee. The collective advice that I have received has helped me appreciate the complexity of interpreting results, designing experiments, and thoughtfully considering alternatives to my hypotheses. It was significantly easier for me to mature as a scientist while working with iv
such stunning examples of academic success. I feel truly lucky to have had the opportunity to share my work with these scientists. Lastly, I want to thank my family. There is no way that I could have completed this work without Rachel, the love of my life, the maker of lunches on extremely busy days, and the most unconditionally devoted person I have ever met. Rachel was a huge part of this work, and as Dr Baker constantly says, this lab would fall apart without her. I also want to thank my Dad and my brother. Though they didn’t understand a word of what is written on these pages, they did flip through the thesis and humor me. Their support has meant the world to me. Finally, this thesis is dedicated to my mom, Judi, who never got to see or share in any of my successes. I hope I’ve made you proud. Formal Acknowledgements: Figures 7-8 are taken from the following manuscript: Bell JD, Ai J, Chen Y, and Baker AJ. (2007) Mild in vitro trauma induces rapid Glur2 endocytosis, robustly augments calcium permeability and enhances susceptibility to secondary excitotoxic insult in cultured Purkinje cells. Brain;130:2528-42. Figures 11-24 (excluding 19, 21, and 23) are taken from the following manuscript: Bell JD, Park E, Ai J, and Baker AJ (2009). PICK1-mediated GluR2 endocytosis contributes to cellular injury after neuronal trauma. Cell Death Differ 16:1665–1680.
v
Table of Contents Abstract.......................................................................................................... ii Acknowledgements ...................................................................................... iv Table of Contents ......................................................................................... vi List of Abbreviations ................................................................................... xi Chapter 1: Introduction ............................................................................... 1 1.1 Clinical Overview and Epidemiology of Traumatic Brain Injury ............................ 1 1.1.1 TBI Epidemiology ............................................................................................. 1 1.1.2 Cost of TBI ........................................................................................................ 3 1.1.3 Classification of TBI Severity ........................................................................... 4 1.2 Pathophysiology of a traumatic brain injury............................................................. 6 1.2.1 Primary Injury.................................................................................................... 6 1.2.2 Mechanical forces affecting cerebral tissue after TBI ....................................... 9 1.3 Mechanisms of secondary injury after TBI ............................................................ 12 1.3.1 Intracranial pressure and secondary ischemia.................................................. 12 1.3.2 Sub-cellular mechanisms of secondary injury ................................................. 14 1.4 Glutamate Excitotoxicity ........................................................................................ 15 1.4.1 Glutamate......................................................................................................... 15 1.4.2 Glutamate Release ........................................................................................... 19 1.4.3 Glutamate Receptors........................................................................................ 24 1.4.3.1 NMDARs .................................................................................................. 25 1.4.3.2 AMPARs – Discovery and function ......................................................... 28 1.4.3.3 Kainate receptors ...................................................................................... 35 1.4.3.4 Metabotropic Glutamate Receptors (mGluRs) ......................................... 36 1.4.4 The concept of excitotoxicity........................................................................... 37 1.4.4.1 De-regulation of glutamate release ........................................................... 39 1.4.4.2 An alternative look at excitotoxicity:........................................................ 43 Post-synaptic glutamate receptor dysfunction ...................................................... 43 1.4.4.3 Consequences of excitotoxicity: Ca2+-dependent neurodegeneration ...... 45 1.4.4.4 Oxidative stress and Mitochondrial Injury ............................................... 46 1.5 AMPA Receptor Trafficking: GluR2-lacking AMPA Receptors as sources of calcium influx ............................................................................................................... 50 1.5.1 Modification of the AMPA Receptor GluR2 content. ..................................... 52 1.5.1.1 Epigenetic silencing of GluR2 .................................................................. 52 1.5.1.2 Local trafficking of GluR2 protein ........................................................... 54 1.5.1.2.1 NSF/AP2 Site interactions in GluR2 trafficking ............................... 56 1.5.1.2.2 AMPA receptor c-terminal PDZ interactions .................................... 57 1.5.1.2.3 PDZ Interactions in GluR2 trafficking .............................................. 63
vi
1.5.1.2.4 GluR2 trafficking in synaptic plasticity............................................. 69 1.5.1.2.5 GluR2 trafficking in TBI ................................................................... 71 1.6 Rationale for proposed study .................................................................................. 73 1.7 Statement of Hypotheses......................................................................................... 76 1.7.1 General Hypotheses ......................................................................................... 76 1.7.2 Specific Hypotheses......................................................................................... 76 1.8 Statement of Objectives .......................................................................................... 78
Chapter 2 – Model Characterization and General.................................. 79 Methods........................................................................................................ 79 Chapter 2: General Methods ..................................................................... 80 2.1 Preface..................................................................................................................... 80 2.2 In vitro methods ...................................................................................................... 80 2.2.1 Isolation and dissociation of cortical cell cultures........................................... 80 2.2.2 In Vitro Model of TBI...................................................................................... 81 2.2.2.1 Use of stretch injury models in TBI literature .......................................... 82 2.2.2.2 The Stretch + NMDA model .................................................................... 83 2.2.2.3 Toxicity studies: Dose response characterization of stretch pressures ..... 86 2.2.2.4 Carboxyfluorescein assays of membrane permeability ............................ 90 2.2.3 Protein extraction and quantification ............................................................... 94 2.2.4 Co-Immunoprecipitation of GluR2 endocytotic proteins ................................ 95 2.2.5 SDS-PAGE ...................................................................................................... 95 2.2.6 Immmunoblotting ............................................................................................ 96 2.2.7 Acid Strip Immunofluorescence ...................................................................... 97 2.2.8 [Ca2+] Measurement......................................................................................... 99 2.2.9 Secondary AMPA Toxicity............................................................................ 100 2.2.10 Whole cell electrophysiology ...................................................................... 101 2.3 TAT peptides ........................................................................................................ 102 2.3.1 The HIV-1 TAT protein transduction domain ............................................... 103 2.3.2 Design of PICK1 inhibitory TAT peptides.................................................... 104 2.4 In vivo Methods .................................................................................................... 111 2.4.1 Fluid percussion trauma................................................................................. 111 2.4.2 Slice Electrophysiology ................................................................................. 111 2.4.3 TUNEL staining............................................................................................. 112 2.5 Contributions......................................................................................................... 113 2.6 Statistics ................................................................................................................ 114
Chapter 3: GluR2 trafficking in modeled brain trauma ...................... 115 3.1 Preface................................................................................................................... 116 3.2 Phosphorylation of GluR2 serine 880 following in vitro trauma correlates with susceptibility to AMPA toxicity ................................................................................. 116 3.2.1 NMDA receptor dependence of GluR2 phosphorylation .............................. 118 vii
3.3 In vitro trauma increases PICK1-PKCa binding................................................... 119 3.4 PKCa is embedded in the NMDAR complex: ...................................................... 122 PKCa co-precipitates with PSD-95......................................................................... 122 3.5 Traumatic injury increases GluR2 endocytosis .................................................... 126 3.6 PICK1-mediated endocytosis of GluR2 following fluid percussion trauma ........ 130 3.7 Summary of results ............................................................................................... 133
Chapter 4: Phenotypic AMPAR changes in modeled brain trauma ... 137 4.1 Preface................................................................................................................... 138 4.2 AMPAR-mediated mEPSC activity following in vitro traumatic injury.............. 138 4.3 AMPA receptor-mediated calcium influx following in vitro trauma: .................. 140 4.4 Interfering with GluR2 endocytosis is cytoprotective in vitro.............................. 144 4.5 Hippocampal CA1 is hyperexcitable following fluid percussion trauma: Excitability is lowered with TAT-QSAV application ................................................ 147 4.6 Hippocampal CA1 Naspm sensitivity increases after FPI.................................... 150 4.7 Occluding GluR2 endocytosis reduces apoptotic cell death:................................ 156 4.8 Summary ............................................................................................................... 157
Chapter 5: Discussion, Limitations and Future Directions .................. 161 5.1 Preface................................................................................................................... 162 5.2 Corroborating studies............................................................................................ 162 5.3 Co-operation of Stretch + NMDA ........................................................................ 167 5.4 Limitations of the current study............................................................................ 168 5.4.1 Non-specific Tat peptide interactions ............................................................ 168 5.4.2 Non-specific effects of Tat peptide transduction ........................................... 169 5.4.3 Co-precipitation: What does it mean?............................................................ 170 5.4.4 TNFα-induced AMPA receptor trafficking: An alternative mechanism of calcium-permeable AMPA receptor expression ..................................................... 172 5.5 Future Directions .................................................................................................. 173 5.5.1 Total GluR2 levels are reduced by 24 hours following trauma ..................... 174 5.5.2 GluR1 trafficking may increase following trauma: ....................................... 175 5.5.3 Tat-QSAV treatment does not occlude induced synaptic plasticity: ............. 178 5.5.4 – Does inhibition of the PICK1 PDZ domain represent a future anti-excitotoxic therapy?................................................................................................................... 181 5.6 Significance of Findings ....................................................................................... 185 5.7 Conclusions........................................................................................................... 186
References Cited........................................................................................ 195
viii
List of Figures Chapter 1: Introduction FIGURE 1. Mechanisms of cytotoxicity following traumatic brain injury. FIGURE 2. Schematic diagram of an AMPA receptor subunit. FIGURE 3. Processes leading to excitotoxicity after CNS injury. FIGURE 4. GluR2 subunit domain structure. FIGURE 5. Steps involved in the intracellular trafficking of the GluR2 subunit. Chapter 2: General Methods FIGURE 6. The cell injury controller and schematic of experimental paradigm. FIGURE 7. Dose-Response Characterization of stretch injury model FIGURE 8. Mild injury does not increase non-specific neuronal cell membrane permeability. FIGURE 9. Design of PICK1-inhibitory Tat peptides and mechanisms of Tat-peptide uptake. FIGURE 10. Transduction of dansyl-Tat-QSAV into cultured cortical neurons and brain slices in vivo. Chapter 3: GluR2 trafficking in modeled brain trauma FIGURE 11. Stretch + NMDA increases S880 phosphorylation of GluR2 and vulnerability to secondary AMPA toxicity FIGURE 12. Stretch + NMDA confers association of PKCa with PICK1. FIGURE 13. PKCa co-precipitates with PSD-95: A potential link from the NMDAR to GluR2 endocytosis. FIGURE 14. Stretch + NMDA increases GluR2 endocytosis FIGURE 15. In vivo traumatic brain injury (TBI) promotes GluR2 phosphorylation and association with PICK1.
ix
Chapter 4: Phenotypic AMPAR changes in modeled brain trauma FIGURE 16. Traumatic injury in vitro increases AMPAR mEPSC amplitude and sensitivity to intracellular polyamines. FIGURE 17. Stretch + NMDA promotes calcium influx through calcium-permeable AMPARs. FIGURE 18. Inhibiting GluR2 endocytosis is neuroprotective FIGURE 19. Post-injury CA1 hyperexcitability is attenuated by Tat-QSAV treatment. FIGURE 20. CA1 hippocampal physiology is sensitive to antagonists of calciumpermeable AMPA receptors after TBI. FIGURE 21. Perturbing GluR2 endocytosis affords cytoprotection from apoptosis at 24 hours following fluid percussion trauma. Chapter 5: Discussion and Limitations FIGURE 22. Summary of proposed signaling in TBI FIGURE 23. Total GluR2 protein levels are reduced at 24 hours following FPI. FIGURE 24. Stretch + NMDA increases GluR1 S845 phosphorylation FIGURE 25. Hippocampal LTP is preserved with PICK1 inhibition.
x
List of Abbreviations ABP – AMPA receptor binding protein AMPA - α-amino-3-hydroxyl-5-methyl-4-isoxazole-propionate AMPAR – AMPA receptor ANOVA – Analysis of variance AP2 – Adaptor protein 2 APAF-1 – Apoptosis peptidase activating factor 1 ATP – Adenosine Triphosphate BAR domain - Bin–Amphiphysin–Rvs domain CA1/3 – Cornus ammonis area 1/3 CNQX - 6-cyano-7-nitroquinoxaline-2,3-dione CNS – Central Nervous System CoIP – Co-immunoprecipitation CP-AMPARs – Calcium-permeable AMPA receptors CPP- Cerebral perfusion pressure DAI – Diffuse Axonal Injury D-MEM - Dulbecco’s modified eagle medium EAA – Excitatory amino acid FDU – (+)-5-fluor-2’-deoxyuridine fEPSP – Field excitatory post-synaptic potential GCS – Glasgow Coma Score GluR2 – Glutamate receptor subunit, 2 GRIP – Glutamate receptor interacting protein HBSS - Hank’s balanced salt solution HIV – Human immunodeficiency virus ICP – Intracranial pressure IgG – Immunoglobulin G L-NAME - γ-nitro-L-Arginine-Methyl Ester LTP/LTD – Long term potentiation/depression MAP – Mean arterial pressure
xi
mEPSC – Miniature excitatory post-synaptic current mRNA – Messenger Ribonucleic Acid Naspm – 1-naphthyl acetyl spermine NCX – Sodium calcium exchanger NMDA - N-methyl-D-aspartic acid NMDAR – NMDA receptor nNOS – Neuronal nitric oxide synthase NSF - N-Ethylmaleimide-Sensitive Fusion Protein OGD – Oxygen glucose deprivation PDZ - Post synaptic density protein (PSD95), Drosophila disc large tumor suppressor (DlgA), zonula occludens-1 protein (zo-1) PBS - Phosphate buffered saline PI – Propidium Iodide PICK1- Protein interacting with C Kinase 1 PKCα – Protein kinase C, alpha PMSF - Phenylmethylsulphonyl fluoride PSD-95 - Post-synaptic density protein, 95 kDa P.S.I – Pounds per square inch PTD – Protein transduction domain RIPA - Radio-Immunoprecipitation Assay ROS – Reactive Oxygen Species SDS-PAGE - Sodium dodecyl sulfate polyacrylamide gel electrophoresis SNARE - Soluble NSF Attachment Protein Receptors TARP – Transmembrane AMPA receptor regulatory protein TAT – Transacting activator of transcription TBI – Traumatic Brain Injury TUNEL - Terminal deoxynucleotidyl transferase dUTP nick end labeling
xii
Chapter 1: Introduction 1.1 Clinical Overview and Epidemiology of Traumatic Brain Injury 1.1.1 TBI Epidemiology Traumatic brain injury (TBI) continues to be a leading cause of death and disability in both developed and developing nations1-5. In North America, TBI is recognized as the leading cause of mortality and morbidity in young adults (15 to 44 years of age)6, while incidences of brain trauma continue to rise in the developing world as rates of vehicle use outpace the implementation of safety infrastructure and effective neurosurgical critical care initiatives7-9. Thus, the acute management and chronic treatment of head trauma is a global issue, with some estimates suggesting that by the year 2020 TBI will rank as the third most prevalent cause of worldwide mortality and disability10. Epidemiological studies are highly varied with respect to statistical estimates of TBI incidence, largely because of differing head injury inclusion criteria, and variability in classification of hospital admission. The United States Center for Disease Control (CDC) estimates that approximately 1.4 million Americans sustain a TBI each year11. Of those, approximately 1.1 million people will be treated and subsequently released from an emergency department, 235,000 will require long-term hospitalization, and 50,000 will die11-13. These figures represent a startling statistic; the number of people hospitalized each year for traumatic brain injuries exceed those diagnosed with multiple sclerosis, breast cancer, and spinal cord injury combined14. In Europe, epidemiological data for TBI is scarce, but most estimates indicate an annual aggregate incidence of hospitalized and fatal TBI of approximately 235 per 100, 00015. Globally, patients that succumb to their
1
injuries post TBI are thought to represent approximately one third of all injury-related deaths11,12,16. The major causes of head injury highlight the susceptibility of certain populations to TBI. For example, falls are thought to represent approximately 30% of all brain injuries, and occur most frequently in children aged 0-4 and in adults over the age of 7511,13. Events described as being “struck by or against” account for 20% of TBIs, and are thought to represent most of the 475,000 injuries sustained in the United States by children aged 0-14, largely from participation in youth sports and other recreational activities11. Motor vehicle accidents account for 20% of TBIs, with the highest rate of TBI-induced hospitalization from MVA occurring in adolescents aged 15-1911. Collectively, these statistics highlight the susceptibility of young children and young adults to TBI, and point to a need for new safety initiatives for young athletes and inexperienced drivers. Interestingly however, while TBI related emergency department (ED) visits are dominated by young children and adolescents (1696.1 per 100,000), the highest rate of TBI-related hospitalization and death actually occurs in adults aged 75 and older (322.7 per 100,000)11, indicating that health care costs are more significantly affected by injuries to the elderly, despite the high number of emergency department visits by children. Additional causes of TBI include assault (11%), non-motor vehicle transport (e.g., cycling, rollerblading, skateboarding) (3%), idiopathic injuries (9%), and other causes of brain trauma (e.g., blast injury and suicide) (7%). At all age groups and causes, TBI is three times more likely to effect males than females17, primarily due to participation in
2
violent athletic and recreational programs and the more aggressive risk-taking behaviours of males.
1.1.2 Cost of TBI Recent estimates suggest that approximately 5.3 million people in the United States are currently living with a disability as a result of an acquired TBI5, which highlights brain injury as a public heath issue with a significant economic burden. The most recent estimate of the cost of TBI to the United States health care system is $37.8 billion USD18, which, when averaged across all injury severities (i.e., mild, moderate and severe TBI), translates to a per person cost of $115,500 USD19,20. Approximately 65% of the cost of traumatic brain injury is accrued by TBI survivors, and is related to direct hospital costs, long-term care and rehabilitation, while 35% of the cost is associated with head injury deaths20. Overshadowed by the significant monetary drain on the health care system are the intangible challenges facing the families of survivors, and the longer term psychosocial, functional and neuro-cognitive disabilities suffered by survivors of TBI that impact society on the whole. Some epidemiologists have relied on standardized scales such as the functional capacity index (FCI) and life years lost to injury (LLI) to estimate these more abstract costs21, which measure inabilities to perform basic tasks including eating, hearing and speaking, and ambulation, and take into account both life expectancy and the number of years of professional productivity lost to injury. One study estimated that TBI results, on average, in 43 years of reduced functional capacity, a frightening statistic when coupled with the data reflecting the number of individuals affected by a head injury per year. These trends continue in Europe, where TBI accounts for the greatest number
3
of total years lived with disability resulting from trauma22. Moreover, when the psychosocial and emotional sequelae (including depression, anxiety, confusion and loneliness23) in individuals sustaining TBI are taken into account, it is easy to grasp the overall burden that brain trauma places on the individual and society as a collective, and to gain an appreciation for the data suggesting that indirect TBI costs more than triple those related to hospitalization and emergency treatment21.
1.1.3 Classification of TBI Severity
An individual is said to have sustained a traumatic brain injury if he/she has cranio-cerebral trauma caused by an external force, and associated with neurological or neuropsychological abnormalities, loss of consciousness, skull fracture, intracranial lesions or death. However, beyond identifying a patient as having suffered acquired brain trauma, the severity classification of TBI is of long-standing interest to both clinicians and researchers interested in predicting outcome and providing post-acute medical care. In most clinical settings, TBI is classified on the basis of single indicators including the Glasgow Coma Scale (GCS), duration of post-traumatic amnesia (PTA) and duration of loss of consciousness (LOC). Indeed all of these indices have demonstrated good predictive value in classifying TBI, with higher GCS scores and brief losses of consciousness associated with what is generally termed “mild” TBI24-31. However, recent studies indicate a number of confounds in classifying TBI according to these individual scales, citing issues in GCS predictability when patients are either intoxicated at the time of injury or given roadside sedation. Further, systemic and psychological shock sustained
4
from poly-trauma has been shown to contribute significantly to durations of PTA, skewing the validity of this measure as a predictor of TBI severity. Thus recent efforts have aimed at classifying TBI according to more reliable measures, most of which focus on a combination of GCS and neuro-anatomical pathologies. For example, Malec et al., at the Mayo Clinic classify TBI as a) Moderate-Severe (definite), b) Mild (probable) and c) Symptomatic (possible)32. Moderate-Severe TBI is said to have occurred if one or more of the following is present: death, LOC greater than 30 minutes, PTA of greater than 24 hours, GCS below 13 and not invalidated by confounding factors, and/or intra-parenchymal/subdural/epidural hematoma, subarachnoid hemorrhage, cerebral or hemorrhagic contusion or brain stem injury. If none of the above has occurred from injury, a mild TBI is diagnosed if there is LOC less than 30 minutes, PTA less than 24 hours, or a depressed, basilar or linear skull fracture with dura intact. If there is no skull fracture, LOC, or PTA at all, the patient is classified as having a symptomatic TBI, if they experience one of dizziness, confusion, blurred vision, nausea, or headache. It is likely that mild and symptomatic TBI are considerably under-diagnosed, with the reported values of 100-300 cases per 100,000 representing a highly conservative estimate of the prevalence33. Accordingly, mild TBI is thought to be somewhat of a “silent” epidemic, due to the under-diagnosis of the condition, coupled with the frequency of residual deficits resulting from it. Indeed one study by Thornhill et al., (2000) identified that one year after TBI, 1260 of 1397 (90%) disabled patients included in their analysis had sustained a mild injury34.
5
The principal complaints from patients sustaining a mild TBI include mood disturbances (including irritability and anxiety), loss of employment due to difficulty with concentration, and increased fatigue34. These impairments, though mild when compared with the physical disabilities sustained by those suffering from a more severe injury, can have a profound impact on an individual’s socialization and quality of life, pointing to the need for clinical hyper-vigilance when a case of mild TBI is suspected.
1.2 Pathophysiology of a traumatic brain injury 1.2.1 Primary Injury
A traumatic brain injury is not self-limiting, but rather is an evolving biological injury that stems from an initial trauma. Accordingly, the pathophysiological mechanisms that lead to neurological deterioration after a head injury can be classified into two separate categories of insults: primary and secondary. The primary injury that occurs following TBI consists of the physical perturbation of the cerebral tissue and vasculature, and is the major determinant of functional outcome35. The extent of primary injury depends almost exclusively on the type of physical load (i.e., force) placed on the brain at the time of injury. For instance, TBI can occur as a result of a blunt impact to the skull, rapid acceleration or deceleration, a penetrating object (e.g., gunshot), or blast waves from an explosion, each of which will produce a unique primary injury profile. Generally speaking, the type of primary injury is classified as focal or diffuse by radiological imaging of structural damage. Diffuse injuries, caused by inertial forces,
6
include macroscopic alterations such as white matter lesions and shearing (known as diffuse axonal injury, DAI), brain swelling, and tearing of blood vessels causing microhemmorhages. Focal injuries on the other hand primarily include contusions (microvasculature injury) and hematomas (both intra-cerebral and extradural). In many cases (e.g., MVAs), more than one of these pathologies is present, representing either a multi-pronged primary injury from one impact (e.g., closed head impact causing both contusion and DAI) or a manifestation of injury resulting from more than one external force (e.g., rotational injury followed by direct impact). The nature, intensity, direction, and duration of the external forces causing primary injury will dictate the pattern and extent of damage, and accordingly has an enormous impact on functional outcome after TBI. In static crush injuries and focal trauma (e.g., a blow to the head), a large proportion of the energy is absorbed by the skull, often limiting damage to superficial structures (e.g., a depressed skull fracture). Extra-dural bleeds, although problematic if left untreated because they will increase intracranial pressure, can often be removed through neurosurgical evacuation, and outcome in these situations is favorable. The poorest outcome is usually associated with diffuse axonal injury, resulting from rotational and inertial forces placed on the brain (discussed in detail in the next section). DAI is characterized radiologically by multiple lesions and disconnection of white matter tracts, appearing often throughout the deep and subcortical white matter and in midline structures including the splenium of the corpus callosum and brainstem. Usually, patients with DAI remain in a lengthy coma and, if they regain consciousness, have significant neuropsychological sequelae and physical disability. The acceleration-deceleration forces responsible for DAI can also have a
7
profound impact on cerebrovascular integrity, resulting in both vascular stenosis and shearing of vessels, leading to multiple intra-parenchymal hemorrhages. In the most severe cases, patients with DAI who survive rapidly lapse into coma, and remain unconscious, vegetative, or severely disabled until life support is withdrawn. Interestingly, the global demographics of the types of primary TBI are changing, as contusions become more frequent than diffuse injuries. Some authors explain this trend by citing an increase in the prevalence of falls in older patients, coupled with decreases in the frequency of high-velocity traffic accidents in young adults due to implementation of more effective safety measures and crackdown on “road-racing”. Another type of primary injury that is increasing in frequency is that sustained from a blast injury (i.e., shockwave-induced brain trauma following an explosion) as military conflict in Afghanistan and Iraq continues to escalate. Although less understood than penetrating injury sustained in combat, blast injuries result in early brain swelling, subarachnoid hemorrhage, and vasospasm36,37, sparking increased research efforts aimed at understanding the interplay between shockwave physics and the corresponding biological injury. Primary injury is irreversible, and is not amenable to therapeutic intervention. Accordingly, the efforts that make the biggest difference in preventing primary TBI are those of safety awareness, and changes to public policy. To this end, primary prevention includes changes to speed limits, enforcement of seat-belt use, and improved road engineering, primarily in underdeveloped countries. Further, socio-cultural attitudes play an important role in prevention of TBI, and should include increased awareness of the dangers of alcohol abuse when participating in certain activities, and increased helmet
8
use in both recreational activities and organized sports. Along this vein, a greater understanding by coaching staff of return-to-play guidelines following concussion will undoubtedly minimize incidents of TBI in both youth and professional athletics.
1.2.2 Mechanical forces affecting cerebral tissue after TBI An understanding of the biomechanics of traumatic brain injury is essential in the development of effective treatment strategies, and will provide the theoretical knowledge base necessary to understand the rationale and physics behind the in vitro and in vivo injury devices described later in this thesis. The compliant properties of cerebral tissue leave the brain susceptible to a variety of mechanical deformations during an impact. Exactly how physical perturbation of grey and white matter transfers to injury at the cellular level is unknown, but some of the basic mechanisms of mechanical injury have been mapped out for decades. The first of these mechanisms is pressure loading. The concussive effects of pressure on the brain were identified in the early TBI literature38, and are now known to reflect the dissipation of energy throughout the brain from pressure gradients generated in the intracranial space at the time of injury39. The direction of propagation of this intracranial pressure front effects the elastic deformation of cerebral tissue, and accordingly, impacts the type and level of strain experienced by the structures within the brain40. To calculate tissue strain (i.e., a deformation representing the relative displacement of tissue), some simple formulae can be applied. Strain (e) is calculated as:
e=λ–1
9
where λ represents the stretch ratio, a definition of tissue deformation expressed as:
λ = l/lo,
where l = the length of deformed tissue and lo = original tissue length
The calculation of tissue strain (as well as its regional distribution) induced by a uniform pressure load is important in understanding how a blow to the head translates to tissue injury. Clinically, it is impossible to measure the strain experienced by cerebral tissue during impact, but experimental approaches have generated both simulated intracranial pressure patterns produced during impact41 as well as calculated the corresponding regional strain of brain tissue in response to said pressure loading40 (identified through the use of finite element models, or FEMs). One such model identified that pressure loading of 3.5 atmospheres (atm) produces brainstem strain (e) in excess of 10%, a level of axonal strain higher than in any other regions, and similar to that produced during herniation of the brain stem through the foramen magnum40. Indeed this level of loading is reflective of a severe brain injury (calculated to occur when pressure loading exceeds 235 kPA, or 2.3 atm42) where brain stem herniation is a common response to rapidly elevated ICP. Brainstem injury arising from shear stress plays a prominent role in neurological dysfunction following pressure loading, and accounts for the vast majority of TBI-induced death by neurological criteria43.
10
The identification of pressure as a major contributor to TBI pathophysiology led to the development of animal models implementing a so-called “percussion-concussion”. One such model, the fluid percussion injury device (FPI), will be discussed in detail in the next chapter, and is used in this thesis as an in vivo TBI experimental paradigm. The second type of load placed on the brain during TBI is an inertial load that results from rapid head rotational motions, common in motor-vehicle accidents, and in some cases, falls and assaults44. The human brain has a moment of inertia (I) – that is, a resistance to change in its rotation rate. When the forces that resist the rotation of the head are overcome by sufficient changes to rotational acceleration (i.e., angular velocity over time), there is an instantaneous change to the angular momentum of the head, and unrestricted movement causing dynamic shear, tensile, and compressive strains on cerebral tissue44. Angular momentum is represented by the formula: L=Iω where L = angular momentum, I = moment of inertia, and ω = angular velocity. Thus one can see the direct relationship between changes to angular velocity and corresponding angular momentum. Rapid changes to angular velocity are responsible for diffuse axonal injury (DAI), the shearing and stretching of neuronal white matter discussed previously that result in a large number of swollen and disconnected axons. The duration of this axonal stretching plays an important role in the resultant injury. Under normal circumstances, human brain tissue is ductile to stretch, rapidly regaining its original geometry when deformed (e.g., during a concussive force). This is because the deformation in this scenario is generally quite slow. However, when axonal strain is applied rapidly (e.g., during an MVA), the tissue acts stiffer, exhibiting a more
11
brittle character. This can be easily recapitulated when one stretches and ordinary piece of plasticine (i.e., rapid stretching will break the material). This is a classic visco-elastic response to rapid deformation, and occurs in the human brain as it does in other materials, causing damage to the axonal cytoskeleton, and sometimes, physical disconnection. The mass effects of the brain thus result in the white matter literally pulling itself apart. The forces that result in this type of tensile elongation occur in 50 ms or less45. Because the mass effects of the human brain play such a large role in the impact of rotational acceleration on axonal integrity, the reproduction of this phenomenon in an animal model (where the brain is much smaller) has proven difficult. Indeed the Holbourn scaling relationship (which summarizes the acceleration needed to produce injury across varying brain sizes) predicts that the inertial forces necessary to produce DAI in a rat (with a brain weighing just 2 g) would need to approach an unachievable 8000% of those that produce DAI in a human46. Thus, to produce a clinically relevant level of axonal and neuronal stretching, investigators have relied on in vitro models of tissue strain which are also utilized in this thesis, and are discussed in the next chapter.
1.3 Mechanisms of secondary injury after TBI 1.3.1 Intracranial pressure and secondary ischemia Over the last few decades, we have learned much about factors associated with worse outcomes following traumatic brain injury. There is a substantial body of work that has analyzed the systemic and intracranial physiologically targeted interventions that might reduce secondary injury and make a difference in outcomes. The first such
12
intervention targets elevated intracranial pressure (ICP) after TBI. ICP is the pressure measured within the skull, and therefore, exerted on the brain (ICP can also be measured intra-parenchymally or intra-ventricularly, but all estimates suggest that these values should be identical). Normal ICP in healthy adults is generally below 20 mmHg; however, post-TBI edema of the brain or the development of an epidural hematoma or subdural hemorrhage can dramatically raise ICP, causing internal or external herniation of the brain, with distortion and pressure on cranial nerves and vital neurological centres. To treat elevated ICP, both neurosurgical and physiologic approaches are employed. Neurosurgically, a decompressive craniectomy can allow for the expansion of the brain as it swells without increasing ICP, while an intraventricular catheter can relieve pressure by removing cerebrospinal fluid. Evacuation of a hematoma will also relieve pressure. Physiologically, osmotherapy (increasing the osmolarity of the blood) serves to draw water out of tissues and reduce cerebral edema47,48 while simultaneously increasing blood pressure to counteract the effects of ICP on cerebral perfusion (discussed next). A second but related major physiological intervention targets hypotension and ischemic injury. Ischemic brain damage (reduced blood flow) after TBI is frequently superimposed on the primary injury, and can manifest as either widespread or perilesional. Maintenance of cerebral blood flow depends on a balance between ICP and the arterial pressure of the blood, mean arterial pressure (MAP). Indeed cerebral perfusion is defined as the difference between mean arterial pressure and intracranial pressure:
CPP = MAP - ICP
13
From this relationship, it is easy to see that when ICP is increased, the perfusion of the brain is decreased, resulting in inadequate tissue oxygenation and ischemic injury. Normal CPP falls around 80 mmHg, but when reduced to 50 mmHg or lower, there is metabolic evidence of impaired electrophysiology and tissue ischemia. Indeed clinical studies have demonstrated a correlation between poor neurological outcome and a reduction of CPP below 70 mmHg for a sustained period49,50. Notably, cerebral oxygenation can also be impaired after TBI following more focal micro-vascular destruction, coagulation and stenosis51-53, which results in smaller and more localized infarction. Basic science investigations have corroborated this evidence, demonstrating the sub-cellular expression of hypoxia-inducible factors after destruction of cerebral microvasculature52 following TBI.
1.3.2 Sub-cellular mechanisms of secondary injury In addition to complications of systemic and intracranial physiology, primary injury after TBI is exacerbated by discrete secondary sub-cellular processes that are more elusive to conventional imaging techniques and therapeutic intervention. An understanding of these more complex mechanisms of cell death is integral in the establishment of effective “neuroprotective” treatments for delayed cellular death and dysfunction after TBI, as interventions in systemic or intracranial physiology provide little protection against tissue injury at the cellular level. For example, even when ICP and CPP are restored to normal levels, there remain ongoing sequelae of damage to nervous tissue perpetuated by a number of cytotoxic processes. These include oxidative and nitrosative injury54-65 (free radical injury, lipid peroxidation, DNA fragmentation) glial proliferation and dysfunction66-68 (swelling of astrocytic foot processes, reversal of
14
neurotransmitter reuptake and reactive astrocytosis), inflammation69-80 (invasion of the injury site by microglia and release of proinflammatory cytokines), white matter and cytoskeletal deterioration81-94 (demyelination and proteolysis of the cytoskeleton), apoptotic cell death89,95-106 (both intrinsic and extrinsic) and finally, excitotoxicity and aberrant ionic homeostasis in neurons68,107-120. Each of these interrelated processes contributes to known mechanisms of grey and white matter injury after TBI and a number of comprehensive reviews exist for each topic. Accordingly, emphasis in this section will be placed on excitotoxicity (cell death mediated by hyper-activation of glutamate receptors) as it is a critical initiating factor in the progression of a number of these cascades and is the focus of the cell signaling studied in this particular thesis.
1.4 Glutamate Excitotoxicity 1.4.1 Glutamate Glutamate is the major excitatory neurotransmitter in the mammalian central nervous system, an observation that dates back to the 1950’s121-123. It is a ubiquitous amino acid (estimated to participate in signaling at over half of all brain synapses124) with two stereoisomeric configurations, L and D. In mammals, L-glutamate is the only physiologically relevant conformation of the molecule, and thus any further reference to glutamate refers to the L-glutamate stereoisomer. To gain a full appreciation for the process of excitotoxicity, it is necessary to review glutamatergic pharmacology and physiology, beginning with the synthesis of glutamate and ultimately concluding with
15
Figure 1. Mechanisms of subcellular injury following brain trauma. Microcirculatory derangements involve stenosis (1) and loss of microvasculature, and the blood–brain barrier may break down as a result of astrocyte foot processes swelling (2). Proliferation of astrocytes ("astrogliosis") (3) is a characteristic of injuries to the central nervous system, and their dysfunction results in a reversal of glutamate uptake (4) and neuronal depolarization through excitotoxic mechanisms. In injuries to white and grey matter, calcium influx (5) is a key initiating event in a molecular cascades resulting in delayed cell death or dysfunction as well as delayed axonal disconnection. In neurons, calcium and zinc influx though channels in the AMPA and NMDA receptors results in excitotoxicity (6), generation of free radicals, mitochondrial dysfunction and postsynaptic receptor modifications. These mechanisms are not ubiquitous in the traumatized brain but are dependent on the subcellular routes of calcium influx and the degree of injury. Calcium influx into axons (7) initiates a series of protein degradation cascades that result in axonal disconnection (8). Inflammatory cells also mediate secondary injury, through the release of proinflammatory cytokines (9) that contribute to the activation of cell-death cascades or postsynaptic receptor modifications.
16
Figure 1. Mechanisms of subcellular injury following brain trauma. Adapted with permission from Park, Bell and Baker, 2008, CMAJ, “Traumatic Brain Injury: Can the consequences be stopped”. 178 (9), 1163-1170. Copied under license from Access Copyright. Further reproduction prohibited.
17
mechanisms of glutamate-induced neuronal death. This will highlight both the physiological action of glutamate and its importance in regulating neuronal transmission, as well as the pathological nature of aberrant glutamatergic signaling. The structure of glutamate is that of any other amino acid found in the human body; that is, a central carbon atom bonded to 3 moieties: 1) a carboxyl group (COOH), 2) an amino group (NH3), and 3) a distinctive side-chain, termed an R group. In glutamate, this R group is CH2CH2COO-, an ionized form of CH2CH2COOH (pKa 4.1) that exists at physiological pH levels125. Notably, the ionized R group is what distinguishes the nomenclature of L-glutamic acid (unionized) and the more common term, glutamate (ionized). To use glutamate as an intercellular signal, neurons and glia have collectively developed a system which comprises an input, output, and termination of glutamate signaling. Glutamate does not cross the blood-brain barrier, and thus must be synthesized in neurons from local precursors124. Of these, the precursor with the highest prevalence is glutamine, the most abundant free amino acid in the body (500-900 µmol/l) and released primarily by astrocytes in the brain126. Peri-synaptic glutamine is taken up by neurons through pre-synaptic excitatory amino acid transporters (EAAT1-5, discussed later), and metabolized to glutamate by the mitochondrial enzyme glutaminase. An alternative form of glutamate synthesis involves phosphate-activated transamination (transfer of an amino group from an amino acid to an α-keto acid) of 2-oxoglutarate (also termed α-ketoglutaric acid), an intermediate of the tricarboxylic acid (Krebs) cycle127. Indirectly then, neuronal glucose metabolism also plays a key role in glutamate synthesis.
18
Glutamate, similar to all neurotransmitters, is stored pre-synaptically in cytosolic vesicles, a process which is dependent on the activity of another transmembrane glutamate transporter, VGLUT (vesicular glutamate transporter). VGLUTs (3 genes have been identified, VGLUT1-3) regulate the packaging of glutamate into vesicles using an electrochemical proton gradient, established by vacuolar-type proton ATPase127. Because of the remarkably strict substrate recognition ability of VGLUT (i.e., the protein only recognizes L-glutamate and a few cyclic glutamate analogues), it is frequently used as an immunocytochemical marker of glutamatergic nerve terminals. Glutamatergic vesicles (with a glutamate concentration of ~ 100 mmol/l) are transported along axonal microtubules to the presynaptic plasma membrane, where they fuse with exocytotic machinery and form the SNARE complex, a protein-protein interaction involving vesicular synaptobrevin and synaptotagmin, and membrane bound syntaxin and SNAP-25. This anchors the glutamatergic vesicle to the plasma membrane, allowing for subsequent exocytosis of the vesicle’s constituents.
1.4.2 Glutamate Release An understanding of glutamatergic vesicle release is critical in the discussion of excitotoxic processes, as aberrant vesicle fusion is thought to be an important initiating factor in excitotoxic neuron death. Vesicle release is a calcium-dependent process, with vesicle-bound synaptotagmin serving as an intracellular calcium sensor. The calcium responsible for glutamatergic vesicle release is thought to originate from pre-synaptic N and P/Q-type (n, representing neural, p/q meaning purkinje) calcium channels126,128,129, voltage-gated ion channels found in excitable cells. These channels -- which are activated at depolarized membrane potentials and are responsible for the fidelity of synaptic
19
transmission from neuron to neuron -- were identified to be in close proximity to glutamatergic vesicle docking sites, creating a calcium micro-domain that serves as an immediate trigger for vesicle fusion and exocytosis. As such, the probability of release of a glutamatergic vesicle is dependent on the type and density of these pre-synaptic Ca2+ channels expressed and their individual proximity to and interaction with neighbouring transmitter release machinery. Of the calcium channel subtypes, it has been demonstrated that P/Q-type calcium channels contribute to approximately 50% of the presynaptic calcium influx responsible for glutamatergic vesicle fusion, evidenced by a marked inhibition of glutamate release by presynaptic blockade of these channels with Agatoxin IVA130. N-type calcium channels by contrast contribute to only 30% of the total pre-synaptic calcium entry, leading some authors to conclude that the P/Q-type channel interacts more tightly with the release machinery than does the N-type channel at glutamatergic synapses130-132. Mutations in these pre-synaptic calcium channels have a profound impact on neuronal functioning due to their influence on glutamatergic vesicle release. Mutations in the
1A
subunit of pre-synaptic voltage-gated P/Q-type channels have been identified in
two strains of mice, known as the tottering and leaner mice133,134. The mutations which occur at the S4-S5 linker region of the third transmembrane domain near the poreforming region of the channel, markedly reduce voltage-dependent inactivation of the calcium channels during prolonged depolarization, increase glutamate release, and produce a behavioural phenotype of motor seizures135. Accordingly, de-regulation of pre-synaptic calcium channel activation is a critical contributor to aberrations in glutamatergic vesicle fusion, plays a key role in de-regulation of cortical and
20
hippocampal circuitry, and –as will be discussed – contributes significantly to excitotoxic neuronal injury. These pre-synaptic, voltage-gated calcium channels, (and indirectly, glutamatergic vesicle release) are also profoundly regulated by another pre-synaptic protein, the G-protein coupled metabotropic glutamate auto-receptors. Metabotropic glutamate receptors (mGluRs, discussed briefly in the next section Glutamate Receptors) are seven transmembrane domain-containing proteins that bind synaptic glutamate both pre and post-synaptically (the latter of which is discussed later, along with description of the receptor itself). Pre-synaptic mGluRs serve the unique function of acting as a glutamatergic autoreceptor; that is, a glutamate receptor that, upon binding glutamate, provides negative feedback onto transmitter release machinery, thereby reducing glutamatergic vesicle fusion and synaptic transmission. Indeed throughout the CNS, mGluR agonists consistently reduce transmission at glutamatergic synapses (reviewed extensively by 136,137). Some of the precise mechanisms by which pre-synaptic mGluRs inhibit glutamate release are known, and a large body of evidence describes the effects of mGluR activation on pre-synaptic voltage-gated calcium channel activation. Agonists acting on mGluRs reduce current density and calcium influx originating from N, L, and P/Q-type calcium channels found in isolated neocortical, striatal, cerebellar, hippocampal, and retinal ganglion neurons, thereby preventing glutamatergic vesicle fusion in all of these cell types138-142. The mechanism of this inhibition involves translocation of the G-protein βγ moiety, an observation that was made through an elegant experiment that injected Gβγ cDNA into adult rat sympathetic neurons and observed tonic inhibition of N-type calcium
21
channel current143 (represented by a positive shift in the voltage dependence and a slowing of channel activation). At around the same time, a separate investigation found that transfection of neurons with Gβγ, but not Gα, induced a marked inhibition of P/Qtype voltage-gated calcium channels, corroborating the evidence for this moiety in channel inhibition144. However, other studies suggest that the mechanism by which mGluRs inhibit glutamate vesicle release is independent of voltage-gated calcium channel modulation. For instance, L-AP4, a phosphonic derivative of glutamate, potent mGluR agonist and synaptic depressant, induces a marked reduction in miniature excitatory postsynaptic current (mEPSC) frequency in hippocampal CA1 pyramidal cells, while the broad spectrum voltage-gated calcium channel antagonist cadmium completely abolishes mEPSC activity in this cell type145-147. These results have lead authors to suggest that the mechanism of mGluR-mediated inhibition of glutamate release is in fact quite different from voltage-gated channel blockade. To address this discrepancy, other studies have examined the influence of mGluR activation on pre-synaptic potassium channel activation, a modulatory effect that would also decrease glutamate release. Indeed it has been observed that mGluR activation activates pre-synaptic outward potassium conductances in visual cortex, raising the possibility that mGluR activation reduces glutamatergic signaling by a mechanism involving pre-synaptic potassium channels148. A third mechanism through which pre-synaptic calcium is kept in homeostatic balance is through the activity of the sodium/calcium (Na+/Ca2+) exchanger, another protein that dynamically modulates the release of glutamate. Na+/Ca2+ exchangers are 11 transmembrane domain ion transporters found in almost all tissues of the body including
22
the brain, where their mRNA is abundant in the cortex, hippocampus, dentate gyrus, thalamus, and cerebellum149. These exchangers play a critical role in the maintenance of cytosolic calcium by pumping calcium ions out of the cell, using an electrogenic sodium gradient as energy and thereby making this protein an anti-porter. The majority of Na+/Ca2+ exchangers have a transport stoichiometry of 3Na+:1Ca2+, pumping 3 sodium ions into the cell for every one calcium ion pumped out150. At the cellular level, Na+/Ca2+ exchangers play a role both pre and postsynaptically at glutamatergic synapses. Presynaptically – where the exchanger is most abundant relative to other sites151-153 – the protein plays a role in calcium-dependent neurotransmitter release by regulating [Ca2+] at nerve terminals152. When calcium enters the pre-synaptic terminal, it is only required for a brief period of time (pre-synaptic calcium transients last less than a millisecond154-156), and must be rapidly extruded157 to prevent the aberrant and pathological event of uncontrolled vesicle fusion. After depolarization-induced Ca2+ entry, Ca2+ efflux from isolated nerve terminals (synaptosomes) is markedly slowed by the removal of extracellular sodium158,159, suggesting that if the electrochemical gradient required for function of the Na+/Ca2+ exchanger is altered, so too are pre-synaptic calcium dynamics, and by association, transmitter release. Pre-synaptic calcium extrusion by the Na+/Ca2+ exchangers is therefore among the most critical regulators of neurotransmission, and a dysfunction of this protein has dire consequences on neuronal cell viability and function. As will be discussed in the section on excitotoxicity, dysfunction of the sodiumcalcium exchanger can lead to significantly augmented glutamate release at nerve terminals, manifest through both spontaneous vesicular exocytosis and synaptic
23
facilitation160. Usually this is observed by loading the presynaptic terminal with sodium, through ionophores or other compounds. The augmentation of pre-synaptic vesicle release by high sodium is a phenomenon that has been observed in synapses of invertebrates161-164, the frog neuromuscular junction165-168, and at both peripheral169,170 and central mammalian synapses153,171-173. In TBI, NCX dysfunction occurs through two processes that will be discussed: proteolysis of the exchanger by activated proteases, and reversal of the exchanger due to uncontrolled loading of presynaptic sodium, thereby reversing the electrogenic gradient required for exchanger function. When uncontrolled, this is turn leads to a number of rapidly fatal cellular processes and progression of secondary insult caused by hyperactivation of glutamate receptors.
1.4.3 Glutamate Receptors Fast synaptic communication between nerve cells involves the control of transmembrane electrostatic potential by a host of ion channels, including glutamate receptors. Glutamate receptors are located in the postsynaptic membrane, and activated by neurotransmitters (specifically, glutamate) that are released from the presynaptic cell. Generally, glutamate receptors are closed in the resting state, but open in response to the binding of agonist (i.e., they are ligand-gated), allowing selected ions to flow down their electrochemical gradients through an internal pore (i.e., they are also ionotropic). This ion flux mediates a local depolarization (positive change in membrane potential), representing an excitatory signal that can be further processed by the post-synaptic cell. The magnitude, duration, and type of signal depends on the subtype of glutamate receptor passing current, as each channel has distinct kinetics and permeability (i.e., ionic selectivity) that will characterize the depolarization.
24
Glutamate receptors are responsible for most excitatory signaling in the brain, and are thought to play an instrumental role in the synaptic plasticity that mediates learning and memory formation. Similarly, the physiological significance of glutamate receptor function is highlighted by the involvement of these receptors in a number of CNS disease states, including motor neuron disease, pain, epilepsy, stroke, and as discussed in this thesis, brain trauma. On the basis of their response to synthetic chemical agonists and sequence-homology criteria, three ionotropic glutamate receptor subtypes have been identified: the N-methyl-D-aspartate (NMDA) receptor, the kainate receptor, and the αamino-3-hydroxy-5-methyl-4-isoxazole propionic acid (AMPA) receptor. The general characteristics of each channel will be discussed in this section, as pathophysiological activity at all of these receptors has been implicated in excitotoxic injury following TBI. However, as this thesis examines the specific role of the AMPA receptor in mediating excitotoxic injury, a more thorough introduction to AMPA receptor function is necessary to accurately clarify and justify the specific aims and hypothesis of this thesis. This will follow this section. 1.4.3.1 NMDARs The NMDA receptor is a hetero-oligomeric assembly of integral membrane protein subunits. This modular construction has aided in the identification of receptor makeup and influence of subunit composition on the electrophysiological and pharmacological properties of the channel. The NMDA receptor is generally accepted as a hetero-tetrameric assembly of four subunits, two of which are known as obligatory NR1-type subunits, and the other two of the regionally localized NR2-type. In certain developmental periods and in restricted
25
brain regions, NR2 can be replaced with subunits of the NR3 subtype. Receptor isoforms result from incorporating more than eight alternatively spliced variants of NR1 (a-h), and peptides encoded by four separate NR2 genes (A-D)174-176. As a result, the receptor is termed a dimer of dimers, with one dimer homomeric for NR1, and the other for NR2. Each NMDA receptor subunit has two extracellular, globular domains: a ligand binding domain (LBD) for binding of agonist (e.g., glutamate on NR2, glycine on NR1 and NR3) and an n-terminal domain (NTD)177,178. All subunits also contain three transmembrane domains, and an intracellular c-terminal (CT) domain, which contains a number of serine and tyrosine kinase phosphorylation sites that regulate channel gating and receptor trafficking. Many of these phosphorylation sites are neighboured by PDZdomains (discussed later), which serve as protein:protein interaction motif’s necessary to keep intracellular scaffolds close to the receptor complex. One such PDZ interaction occurs through binding of the NR2B c-terminus to post-synaptic density protein, 95 kDa (PSD-95), an interaction that regulates the post-synaptic production of nitric oxide, and activation of Ras GTPases among other notable downstream effectors. The NMDA receptor plays in integral role in physiological excitatory CNS neurotransmission as well as pathological disease states. Two main signals generated simultaneously by the receptor complex are responsible for the information conveyed at these channels; the first is a depolarizing current, and the second is a biochemical signal of calcium influx. Upon glutamate binding to the ligand binding domain in the presence of glycine, the channel opens a cation-permeable pore causing a transient membrane depolarization. However, in addition to this dependence of channel opening on agonist binding, the NMDA receptor has a second dependence; membrane potential. At
26
hyperpolarized membrane potentials (including resting membrane potential), NMDA receptors are blocked by sub-millimolar extracellular concentrations of magnesium (Mg2+). Magnesium tightly binds the channel pore, and consequently reduces the NMDA receptor component of synaptic currents considerably. However, when neurons are depolarized (e.g., by activation of neighbouring glutamate receptors of the non-NMDA subtype – see below), the magnesium block is partially expelled, allowing both sodium and calcium influx through the receptor complex. This unique property renders the Ca2+ influx through NMDA receptors a type of neuronal coincidence detector for the simultaneous occurrence of both depolarization and synaptic release of glutamate. Calcium influx from the NMDA receptor triggers events crucial to neuronal survival and plasticity. For example, calcium micro-domains located near the NMDA receptor play an important role in synapse to nucleus signaling, triggering the transcription of many pro-survival neuronal proteins. This occurs through a cascade involving extracellular signal related kinase (ERK1/2), which undergoes nuclear translocation in response to NMDA receptor activation and phosphorylates the cyclicadenosine-monophosphate (cAMP)-response element binding protein (CREB)179,180. CREB is ubiquitously expressed transcription factor that initiates the transcription of a number of anti-apoptotic factors, including brain derived neurotrophic factor (BDNF)181, as well as anti-apoptotic bcl-2 proteins that inhibit the initiation of programmed cell death (apoptosis)182. NMDAR-derived calcium binding to cytosolic calmodulin also initiates the transcription of CREB-dependent proteins, since CREB phosphorylation also occurs via activation of calmodulin-dependent activation of calmodulin kinase IV (CaMKIV)183.
27
NMDA receptor activation is also responsible for remodeling the synapse during period of synaptic plasticity. The most thoroughly characterized examples of such synaptic plasticity in the mammalian nervous system are long-term potentiation (LTP) and long-term depression (LTD), which involve changes to the post-synaptic response of neurons following various patterns of electrophysiological or chemical stimulation. These events will be discussed in detail in the section on AMPA receptors, as although they are initiated by activation of NMDARs, they primarily involve the trafficking of AMPA receptors from cytosolic and extrasynaptic sites to the plasma membrane, and vice versa. Pathological activation of the NMDA receptor is implicated in numerous diseases of the central nervous system, though the clinical failure of NMDA receptor antagonists has brought this hypothesis under much scientific scrutiny in the last few years. The section that follows this will discuss the involvement of the NMDAR in traumatic brain injury-induced neuronal death and dysfunction. 1.4.3.2 AMPARs – Discovery and function
Similar to the NMDA receptor, AMPA receptors are hetero-oligomeric proteins made up of globular subunits, in this case termed GluR1-4 (also termed GluRA-D). The polypeptides encoding AMPAR subunit makeup were first identified through expression cloning in oocytes in 1994184, and the sequence predicted the functional domains that each subunit is now known to contain. The cloned GluR1 polypeptides contained a hydrophobic signal sequence, and four hydrophobic regions, which correspond to the four transmembrane domains that span the plasma membrane as α helices.
28
Epitope tagging185 and glycosylation analysis184 subsequently identified the rest of the AMPA receptor subunit topology. A large extracellular N-terminal region was identified, which is followed by the first transmembrane domain. Following this, the second transmembrane domain was identified as the channel pore region, which does not actually traverse the membrane, but rather dips into it from the cytosolic side. The poreforming domain is followed by a true transmembrane domain, an extracellular loop, the third transmembrane domain, and finally, the cytoplasmic tail, otherwise known as the cterminus region. Each subunit also contains an extracellular domain known as the S1-S2 site, which is the primary binding site for the endogenous agonist glutamate (Figure 2). GluR1 mRNA is expressed in most brain regions, but is absent from the thalamus and mesencephalon, anatomical locations which are known to express AMPA sensitive channels184. This led to the subsequent homology cloning of three additional AMPA receptor subunits, GluR2, GluR3 and GluR4, which were found to be highly related to the originally cloned GluR1184,186. AMPA receptor subunit mRNA is initially translated on the rough endoplasmic reticulum (ER), where subunit dimerization occurs, and a high mannose glycosylation attaches to specific asparagine residues in the first extracellular domain. Following ER synthesis, the receptors transit through the golgi network, where the high mannose sugars are modified to the complex carbohydrates seen in mature receptors. Receptors are further trafficked to dendrites or axons, where they are inserted either extrasynaptically (for GluR1187,188) or directly into the synapse (as is the case for GluR2188). Unlike the NMDA receptor, AMPA receptor subunits are also sorted and stored in cytoplasmic vesicles, which allows for the dynamic trafficking of receptors both to (exocytosis) and
29
from (endocytosis) the membrane during synaptic plasticity (discussed in the next section, AMPA receptor trafficking). Like the NMDA receptor, AMPA receptors are also a dimer of dimers. In the forebrain, including the hippocampus and cerebral neocortex, the predominantly expressed subunits are GluR1 and GluR2, with low levels of GluR3 and GluR4. Thus, the major neuronal population -- pyramidal cells -- expresses AMPARs primarily comprised of hetero-tetramers of GluR1 and GluR2. At one point it was hypothesized that GluR2/3 was the other major heteromer in cortical neurons, but the expression of GluR3 is low in this cell type (i.e., ~ 10% of GluR1 or GluR2 levels), suggesting that GluR2/3 is not a predominant subunit combination189. All AMPA receptors are glutamate-gated channels whose post-synaptic activation provides the primary sodium-dependent depolarization during excitatory neurotransmission in the brain. Indeed synaptic strength is almost entirely mediated by the ultimate density of AMPA receptors that accumulate at dendritic synapses190. However, of all of the AMPA receptor subunits, GluR2 is responsible for dictating the channel biophysics as well as ionic permeability. AMPA receptors that contain GluR2 (in contrast to those lacking GluR2, for example GluR1 homomeric channels) have a number of identifiable properties: 1) they are impermeable to divalent cations (including calcium and zinc); 2) they have a lower single channel conductance than receptors lacking GluR2; 3) they exhibit linear current-voltage relationships and 4) they are not subject to blockade by intracellular polyamines. GluR2 dictates these processes as a result of its amino acid makeup. Most mature GluR2 protein contains a positively charged arginine residue (R+) within the re-entrant
30
membrane loop (i.e., the channel pore region) at position 607 in place of the genomically encoded neutral glutamine (Q) residue. This change arises from hydrolytic RNA editing of a single adenosine base to inosine by the adenosine deaminase enzyme ADAR2. Notably, this Q/R editing is exclusive to GluR2, and therefore does not occur in any of the other AMPAR subunits. The addition of this positive charge into the pore of AMPA receptor channels containing GluR2 lowers the single-channel conductance, prevents the passage of divalent cations through the receptor, and also repulses the intracellular blockade of the channel by similarly charged polyamines (e.g., spermine) at positive voltages, thereby sustaining a linear relationship between membrane voltage and current amplitude (as opposed to the inwardly rectifying relationship observed when patching AMPA receptors lacking GluR2). In addition to RNA editing, AMPA receptor molecular diversity is further complicated by alternative RNA splicing of GluR1-4. Each AMPA receptor subunit exists as either of two distinct isoforms, termed “flip” and “flop”, both of which are generated by alternative splicing of a 114 base pair region immediately adjacent to another RNA editing site, the R/G site. This splicing process introduces a functionally critical cassette of 38 amino acids (either flip or flop) into the extracellular loop, which controls AMPAR desensitization and recovery following agonist binding. Differentially spliced subunits also exhibit varying sensitivity to allosteric modulators; for example, cyclothiazide, which reduces AMPA receptor desensitization, is only active in flip, but not flop variants of recombinant receptors. RNA splicing is also developmentally regulated, with only flip splice forms expressed in early postnatal mammalian life, followed by expression of GluR flop isoforms later in development.
31
AMPA receptors are also complemented on the plasma membrane by transmembrane AMPA receptor regulatory proteins, or TARPs. These proteins, including the most well characterized stargazin, co-assemble stoichiometrically with native receptors, acting as auxiliary subunits that are required for receptor maturation, trafficking, and other channel functions191,192. Further detail of TARP-mediated AMPA receptor modulation are reviewed elsewhere192. The ionic permeability of > 95% of native AMPA receptors is exclusively monovalent due to the presence of edited GluR2 in the receptor complex; however, there are a number of neuronal inputs that are capable of modifying AMPA receptor ionic permeability to include passage of divalent cations, via the removal of GluR2. This modification of AMPA receptor ionic permeability to include calcium influx is a critical mediator of both synaptic plasticity and excitotoxic neuron death in a number of CNS diseases, including ischemia, brain trauma, epilepsy, and motor neuron disease. The mechanisms through which GluR2 expression is altered under both physiological and pathological conditions are highly complex -- involving epigenetic changes to mRNA editing as well as intracellular protein:protein interactions between PDZ domains -- and are discussed at length in the sections that follow.
32
Figure 2. Schematic diagram of an AMPA receptor subunit. All receptor subunits have a similar structure and topology. The N-terminal domain (NTD) is followed by S1, which together with S2 forms the glutamate binding site (Glu). Of the four hydrophobic segments, three span the membrane, while one (domain 2) dips into the membrane from the cytoplasmic face and contributes to the channel pore. The alternatively spliced flip/ flop region and the C-terminal PDZ ligand, which interacts with intracellular PDZ domains, are shown.
33
Figure 2. Schematic diagram of an AMPA receptor subunit. Modified with permission from Bredt & Nicoll, 2003. “AMPA receptor trafficking at excitatory synapses”. Neuron. 40, 361-379.
34
1.4.3.3 Kainate receptors Kainate receptors are another class of ionotropic glutamate receptor, made up of subunits KA1-2, and GluR5-7. KA1 and KA2 on their own do not form functional ion channels, but when expressed in conjunction with GluR5-7 will form a channel that allows ion flux in response to glutamate stimulation184. Kainate receptor mRNA can be detected in a number of brain regions including the hippocampus, cerebellum, amygdala, and spinal cord. Generally speaking, kainate receptors are sodium channels, triggering a local depolarization of membrane potential upon agonist binding. However, these receptors have demonstrated calcium permeability in recombinant systems when certain subunit combinations are applied, alluding to the possibility that endogenous kainate receptor subtypes might also play an important role in post-synaptic calcium signaling193,194. Similar to other glutamate receptors, kainate receptor electrophysiology can also be modulated by intracellular effectors. For example, protein kinase Adependent phosphorylation of GluR6 increases kainate receptor single channel conductance, by increasing the coupling efficiency of glutamate binding and channel opening195. Kainate receptors also mediate glutamate release and contribute to excitotoxic neuronal damage, particularly the death of oligodendroglial cells196. Pre-synaptically, investigators report that kainate receptors reduce glutamate exocytosis197, while postsynaptically, kainate receptors couple to c-Jun N-terminal kinase (JNK) activation, initiating an apoptotic cascade that contributes to neuronal and glial cell death in both epilepsy and cerebral ischemia198,199. This has prompted emerging therapies aimed at
35
uncoupling the kainate receptor from its downstream apoptotic machinery with the use of inhibitory peptides198,199. 1.4.3.4 Metabotropic Glutamate Receptors (mGluRs)
Metabotropic glutamate receptors have been discussed previously in the context of regulating synaptic transmission, through their modulation of pre-synaptic voltage gated calcium channels. Post-synaptically however, these 7-transmembrane domain single peptide proteins couple to G-protein activation, resulting in slow, modulatory effects on neurotransmission. Therefore, unlike the ionotropic NMDA and AMPA receptors, mGluRs are not ion channels. However, their modulation of neuronal physiology is responsible for many types of synaptic plasticity, including long-term depression of post-synaptic glutamatergic EPSCs. Their effect on glutamate dependent ion flux is therefore indirect, but nonetheless critical to CNS excitatory signaling. There are 8 different mGluR subtypes that have been identified (mGluR1-8), which are subdivided into three groups based on their sequence homology and their associated signal transduction pathways. Group I mGluRs consist of mGluR1 and mGluR5, which couple intracellularly to phospholipase C and generation of inositol triphosphate (IP3)200,201. This cascade is responsible for liberation of intracellular calcium stores from the endoplasmic reticulum. Group I mGluRs have also demonstrated inhibitory activity on excitatory EPSCs in the hippocampus, through G-protein independent activation of tyrosine kinases202. Group II mGluRs consist of mGluR2 and mGluR3, and are largely responsible for the pre-syaptic effects on N and P/Q-type voltage gated calcium channels discussed previously203,204. These receptors also have an
36
inhibitory effect on adenylyl cyclase signaling, reducing intracellular levels of cyclic AMP (cAMP) and activation of its downstream effectors including voltage-gated calcium channels, protein kinase A (PKA) and cyclic nucleotide gated ion channels137. Group III mGluRs include mGluR4,6,7, and 8, and have modulatory properties similar to the group II mGluRs. They also act as glutamatergic autoreceptors, reducing pre-synaptic glutamate release through modulation of voltage-gated calcium channels137. Because of the inhibitory effect of mGluRs on pre-synaptic glutamate release, the activation of these receptors has gained much attention for the treatment of a number of CNS disorders involving excitotoxicity, including TBI205-208. In the following section, evidence will be presented that suggests that reduced activity of pre-synaptic mGluRs might contribute to excitotoxic glutamatergic signaling following CNS trauma.
1.4.4 The concept of excitotoxicity The neurotoxic potential of glutamate was first proposed by Lucas and Newhouse in 1957, when they discovered that injections of L-glutamate could destroy the inner layers of the mouse retina209. Twenty years later, Olney described the cerebral lesions associated with injection of kainate (structurally related to glutamate) to young animals lacking an intact blood-brain barrier. Olney’s initial findings were also critical to our modern understanding of how glutamate kills neurons, as his data described rapid cellular swelling near dendrosomal components, now known to be particularly enriched in the excitatory amino acid (EAA) receptors which were just discussed. It was in 1969 that he coined the term “excitotoxicity”, to refer to neuronal death induced by excitatory amino acids.
37
Our understanding of how glutamate receptor over-activation induces neuronal death is rooted in important ion substitution experiments performed in the late 80’s. It has long been understood that stimulation of glutamate receptors increases the post-synaptic concentrations of both intracellular sodium and calcium, and a separate role for these ions has been established in excitotoxic neuron death. First, investigators have demonstrated that neuronal cultures exposed to glutamate exhibit immediate and irreversible sodiummediated cell swelling, even in the absence of extracellular calcium210. However, a role for calcium was identified later when other groups described delayed (i..e., long term) glutamate-induced neuronal death when extracellular sodium was removed211. Indeed cell death in this model was attenuated only in the absence of both extracellular sodium and calcium. These observations provided for a simple model of excitotoxicity consisting of two components: an early sodium-mediated cell swelling, and a more delayed, calcium-dependent neuronal degeneration, which can be reproduced through the use of calcium ionophores115. Little debate exists that there is a strong correlation between intracellular calcium concentrations and neuronal injury induced by glutamate. It is well understood that elevated intracellular calcium is the initiating factor in many neurotoxic cascades, including the uncoupling of mitochondrial electron transport from ATP synthesis, the activation of proteolytic enzymes (e.g., calpains) that cleave the neuronal cytoskeleton, endonucleases that fragment nuclear DNA, production of reactive oxygen and reactive nitrogen species, and the initiation of programmed cell death (apoptosis). However, there are two schools of thought directed at understanding how glutamate shifts from an important mediator of neuronal excitatory physiology to an endogenous neurotoxin
38
following injury to the central nervous system. The first hypothesis is that following CNS trauma (including ischemia and TBI), there exists prolonged activation of glutamate receptors due to elevated levels of extracellular glutamate, resulting in increased calcium influx. The second hypothesis suggests that injury to the CNS induces changes to glutamate receptor function, allowing excessive entry of extracellular calcium. In TBI, there exists evidence for the involvement of both of these phenomena, which may in fact also occur at the same time. 1.4.4.1 De-regulation of glutamate release
Microdialysis studies have reported that following traumatic brain injury, extracellular glutamate is markedly increased113,212-217, in some cases up to 9 days following injury218. Indeed these clinical observations have been recapitulated by animal models of TBI219-221. Accordingly, there have been a number of hypotheses put forward to explain how glutamate release and/or reuptake are altered following TBI, resulting in excessive extracellular glutamate accumulation. The first relates to dysfunction of presynaptic calcium extrusion and subsequent glutamate vesicle fusion caused by reversal and failure of the sodium-calcium exchanger. TBI, as discussed, causes a reduction of cerebral perfusion pressure when intracranial pressure increases. This hypoperfusion deprives the cell of both oxygen and glucose. As is well understood from the basics of cellular respiration, glucose is the primary method of ATP production. When ATP levels are depleted, there is a dysfunction of the sodium-potassium exchanger (the membranebound ion pump responsible for maintaining neuronal resting membrane potential). This results in neuronal depolarization and accumulation of intracellular sodium. As was
39
discussed previously, the sodium-calcium exchanger operates on an electrogenic sodium gradient, such that when intracellular sodium is markedly increased, operation of the pump ceases (failing to extrude calcium) and in some cases will reverse, pumping calcium into the cell. Calcium will also enter the cell through the activation of voltagegated calcium channels. This elevation of pre-synaptic calcium in turn triggers the fusion of glutamatergic vesicles in an unregulated fashion, resulting in excessive glutamate exocytosis and toxic concentrations of the transmitter in the synaptic cleft. There is also some evidence that sodium-calcium exchanger (NCX) function ceases due to proteolytic cleavage. Proteolytic inactivation of NCX has been demonstrated in cellular models of excitotoxicity, as well as in whole animal CNS injury, where it was noted that calpain inhibition (preventing NCX cleavage) or expression of NCX lacking the calpain cleavage moiety protects against excitotoxicity222. Further, inhibition of the sodium-calcium exchanger activity has shown neuroprotective properties in both an animal model of TBI, as well as following cellular strain injury, suggesting perhaps that the reverse operation of the protein contributes to neuronal death223,224. A second mechanism by which extracellular glutamate is thought to increase is through dysfunction of astrocytic glutamate transporters, known as excitatory amino acid transporters (EAAT). Five EAAT subtypes have been cloned to date (EAAT1-5), two of which (EAAT1-2) exist primarily in astrocytes225. Astrocytic EAAT2 accounts for > 90% of total glutamate transport in the brain226-228, the majority of which is involved in clearance of synaptic glutamate following regulated excitatory neurotransmission68,226. A number of studies have identified both dysfunction and reversal of EAATs following CNS injury, leading to the hypothesis that impaired clearance of synaptic
40
glutamate, or reversal of astrocytic transporters leads to increases in extracellular glutamate levels following TBI. Firstly, transient down-regulation of EAAT1 and EAAT2 has been reported in the ipsilateral cerebral cortex following controlled cortical impact (CCI, an in vivo experimental model of TBI), concomitant with a reduction in [3H]-D-aspartate binding229. This study is corroborated by evidence that down-regulation of EAAT1&2 levels in the ipsilateral and contralateral cortex after CCI are associated with a rise in CSF glutamate levels, reaching a maximum at 48 h following the injury230. Other mechanisms aside from protein down-regulation are also thought to be involved in EAAT dysfunction. Notably, inhibition of astrocytic glycolysis, a key component of glucose metabolism, causes reversal of glutamate transporter activity. Indeed as discussed glucose delivery is impaired following TBI, and this mechanism may play a role in aberrant extracellular glutamate accumulation. EAAT transporter activity has also been shown to reverse under ischemic conditions231,232, which is frequently an insult superimposed on cerebral tissue following TBI.
Collectively, the data suggesting that
EAAT protein is lost following traumatic injury coupled with reversal of transporter activity during ischemia suggest that the activity of these proteins may play a critical role in excitotoxic neuron death following TBI. Accumulation of extracellular glutamate might also occur via cytoplasmic leakage through damaged cellular membranes. The intra to extracellular ration of glutamate is approximately 1000:1, suggesting that membrane shearing or cellular lysis from cytotoxic edema might contribute to the early rise of glutamate into the extracellular space. Indeed very high levels of dialysate glutamate are reported as a microdialysis probe is lowered into the brain’s parenchyma, producing a laceration injury. A similar
41
cellular leakage is thought to occur in the shear stress zone produced after tissue compression or contusion. Leakage of plasmatic glutamate through disrupted blood-brain barrier (BBB) dysfunction has also been proposed as a mechanism, albeit minor, of augmented extracellular glutamate. It is well established that TBI induces a disruption in blood-brain barrier integrity, and as the plasma concentration of glutamate is ~ 50 μM (i.e., 50x that of the extracellular space), plasmatic glutamate can leak into the interstitial space after injury233,234. Finally, as discussed, glutamate release is profoundly affected by pre-synaptic activation of group II metabotropic glutamate autoreceptors, which slow vesicular exocytosis through inhibition of voltage-gated calcium channel activity. A number of studies have identified a loss of this inhibitory activity following TBI, leading to the hypothesis that pre-synaptic mGluR dysfunction contributes to excitotoxic glutamate release. Indeed loss of group II mGluR mRNA and protein were reported following experimental diffuse brain injury and lateral fluid percussion, a phenomena reported up to 7 days following trauma206,235. Accordingly, authors have tested the efficacy of group II mGluR activation following TBI in attenuating neuronal injury. Indeed administration of both a group II and III mGluR agonist 30 min after lateral FPI has attenuated both neurotoxic extracellular glutamate accumulation and improved functional outcome following the injury236,237. This approach has also improved neuronal survival in cellular models of TBI, suggesting that augmentation of glutamatergic autoreceptor activity can attenuate excitotoxic neuronal death.
42
1.4.4.2 An alternative look at excitotoxicity: Post-synaptic glutamate receptor dysfunction The second hypothesis of how excitotoxicity occurs relates to the post-traumatic dysfunction of glutamate receptor activity, and has emerged out of some contradictions of the initial hypotheses related to augmented glutamate release as the cause of excitotoxicity per se. According to some authors, the concept that high extracellular glutamate is the key to excitotoxicity in TBI conflicts with important and convincing experimental data. A number of studies employing intracerebral microdialysis have indeed shown that cortical injury markedly increases the concentration of extracellular glutamate (discussed in the previous section). However, it has also been demonstrated through rapid sample collection at 2 minute intervals that this increase is often transient, peaking within five minutes of impact and rapidly declining to control levels238-240. Notably, much of the data reporting augmented extracellular glutamate levels following trauma are not specific to excitatory amino acids, with similar abnormalities reported for gamma-aminobutyric acid (GABA), taurine, ascorbate, and adenosine233,241,242. Importantly, these increases occurred on the same time scale and at the same magnitude as glutamate release. Thus, research efforts have also focused on identifying injuryinduced changes to glutamate receptor function, in an effort to understand how glutamate might prove neurotoxic in the absence of extracellular accumulation, or in the presence of only moderately elevated glutamate levels. Augmented glutamate receptor function has deleterious effects on neuronal survival and function in two ways; it imparts a vulnerability to secondary excitotoxicity, and it may interfere with constitutive glutamatergic physiology, such that ordinarily
43
innocuous stimulation of glutamate receptors is rendered deleterious to cellular survival. The concept that a traumatic injury to the CNS can change the ionic permeability, kinetics, or subunit composition of glutamate receptors is supported by a number of investigations. Nearly 15 years ago, it was demonstrated that mechanically injured neurons exhibit a reduced voltage-dependent magnesium blockade of the NMDA receptor243. Indeed this loss of magnesium blockade resulted in substantially larger NMDA-induced calcium influx, equivalent to stimulating control neurons in magnesiumfree extracellular solution. This result has been corroborated by other studies that have shown that mechanical stretch injury initiates large calcium transients that originate from the NMDA receptor, and are accordingly antagonized by AP-5244-246, and that neuronal stretch enhances NMDAR activity by increasing maximal NMDAR current, and steadystate current density247. Unsurprisingly, this enhanced NMDAR current translates to a marked vulnerability to otherwise innocuous levels of both glutamate and NMDA. Investigators have repeatedly shown that treatment of sub-lethally stretched neuronal cultures with L-glutamate augments cell death via the influx of NMDAR-derived calcium117,119,120,248. Thus, in the absence of abnormally high levels of extracellular glutamate, traumatic injury can impart a change to post-synaptic receptor function that translates to increased susceptibility to glutamate receptor stimulation. These findings are not limited to activity at the NMDA receptor. Much attention has been paid to trauma-induced changes to AMPA receptor function as well, as these receptors mediate the majority of ionotropic neurotransmission. Changes to AMPA receptor function and ionic permeability following traumatic injury have been reported. Agonist (i.e., AMPA)-activated currents recorded from traumatically injured neurons
44
exhibit marked potentiation, with increases in both AMPA and kainate mean steady-state current density. AMPA receptor kinetics are also affected by stretch injury, with trauma resulting in a significant increase in both the 20-80% activation rate and desensitization time constant (τ)249,250. Further, traumatic injury to cortical neurons has twice been shown to augment AMPAR-mediated calcium influx251,252, suggesting that traumatic injury is capable of increasing the divalent ion permeability of an ordinarily calcium-impermeable receptor. Notably, these changes to AMPA receptor function appeared to be mediated by regulated signaling pathways, as the effects of trauma on AMPAR current density were abolished by application of NMDA receptor antagonists and inhibitors of protein kinase C, suggesting a potential calcium-dependent modification of AMPA receptor function. However, though the phenomenon of trauma-induced increases in AMPA receptor function has been consistently reported, the mechanism through which trauma modifies AMPA receptor function has not been mapped out. 1.4.4.3 Consequences of excitotoxicity: Ca2+-dependent neurodegeneration The excessive stimulation of glutamate receptors following TBI from either augmented glutamate release or altered post-synaptic glutamate receptor function can only be tolerated for a short period of time due to the cytotoxic effects of elevated intracellular calcium. As discussed, neurons possess specialized homeostatic mechanisms to ensure the strict regulation of cytosolic free calcium, which include the activity of sodium-calcium exchangers and calcium buffering proteins, as well as calcium sequestration into organelles. However, excessive calcium influx can override these regulatory processes, leading to the inappropriate activation of Ca2+-dependent processes
45
that are normally dormant or operate at low levels, causing metabolic derangements and eventual cell death. 1.4.4.4 Oxidative stress and Mitochondrial Injury One mechanism through which stimulation of glutamate receptors causes calcumdependent cell death is through the production of reactive oxygen and nitrogen species (ROS/RNS) downstream of the NMDA receptor. The NMDAR is structurally connected to the intracellular, calcium-dependent synthase responsible for the generation of neuronal nitric oxide (nNOS), via a membrane associated guanylate kinase (MAGUK) protein scaffold. This protein, known as post-synaptic density, 95 kDa (PSD-95) binds to both nNOS as well as the c-terminus of the NR2B NMDA receptor subunit. As such, calcium influx from the NMDA receptor is placed in close proximity to nNOS, effectively coupling activity of the NMDAR to generation of post-synaptic nitric oxide (NO). While NO (by definition a free radical) is an important second messenger involved in a number of constitutive neuronal regulatory pathways and reacts slowly with most biological molecules, when combined with other free radicals it is remarkably reactive and has acutely cytotoxic effects. When excessive free calcium is sequestered by the mitochondria in an effort to restore intracellular calcium homeostasis, the elevated calcium level in the mitochondria increases the production of the superoxide anion. The reaction of this mitochondrialderived superoxide with NMDA-derived nitric oxide produces the highly reactive nitrating species peroxynitrite (an oxidant with activities similar to that of the hydroxyl radical and nitrogen dioxide radical). Peroxynitrite, which investigators have shown is produced in excess following experimental traumatic brain injury119,248,253 as well as
46
following the exposure of neuronal cultures to excessive L-glutamate, produces nitration of amino acid aromatic rings254,255, lipid peroxidation254,255 and DNA fragmentation248,254256
— all of which are rapidly fatal cellular processes responsible for excitotoxic cell
death. Calcium overload from glutamate receptor over-activation also plays a critical role in early mitochondrial swelling257-260. The excessive sequestration of calcium by mitochondria causes not only superoxide generation, but also mitochondrial membrane depolarization, the opening of membrane permeability transition pores and the release of initiating factors of programmed cell death (apoptosis), including for example cytochrome C. Once released into the cytosol, cytochrome C, through binding apoptosome activating factor 1 (APAF-1), activates and recruits caspase-9261-263, a cysteine protease responsible for the progression of apoptotic cascades to the point of cell death. The loss of mitochondrial function during excitotoxicity is cyclical in nature, as it not only eliminates calcium buffering capacity and initiates apoptosis, but it also contributes indirectly (via loss of ATP synthesis) to the influx of calcium resulting from bioenergetic failure of the previously discussed ATP-dependent ion pumps. Cytoprotective approaches to excitotoxic degeneration have therefore targeted mitochondrial function to attenuate the multi-pronged effects of mitochondrial damage. For example, cyclosporin A, an immunosuppressant and inhibitor of the mitochondrial membrane-permeability-transition pore, has been shown to significantly reduce neuronal cell loss following TBI, thus illustrating the importance of these processes.
47
Figure 3. Processes leading to excitotoxicity after CNS injury. A) Excitotoxicity can occur following pre-synaptic depolarization caused by failure of ATP-dependent sodium and calcium extrusion. Excessive calcium accumulating via voltage-gated calcium channels and reversal of sodium/calcium exchangers leads to a deregulation of glutamatergic vesicle fusion and massive glutamate exocytosis. Other sources of synaptic glutamate accumulation include plasmatic leakage through disrupted blood-brain barrier dysfunction, impaired glial-dependent glutamate uptake, as well as leakage through damaged cellular membranes following cytotoxic edema. B) An alternative hypothesis proposes that excitotoxicity can occur via a dysfunction of post-synaptic receptors. These aberrations can include intracellular modifications leading to increased receptor calcium permeability, decreased receptor desensitization, and increases in mean steady-state current densities. Together these processes also lead to cell swelling and calcium accumulation. Both theories of excitotoxicity involve a cytotoxic role for calcium, which leads to potent oxidative injury, DNA fragmentation, cytoskeletal proteolysis, and the initiation of apoptosis.
48
Figure 3. Processes leading to excitotoxicity after CNS injury
49
There is unquestionably a role for the NMDA receptor in the pathophysiology of excitotoxic injury after TBI. However, treatments aimed at reducing NMDA receptor functioning have proved to be impractical, due to interference with physiologic receptor function and suppression of the pro-survival NMDA receptor signaling discussed previously. Indeed all clinical trials for TBI employing NMDA receptor antagonists were stopped prematurely due to adverse side effects and a lack of efficacy in improving functional outcome. Rather, there has emerged a general consensus that antagonists of AMPA receptors, (e.g., the quinoxalinedione NBQX) are much more effective than NMDA receptor antagonists in attenuating neuronal cell death during periods of excitotoxicity, even when given as late as 24 hours following injury to the CNS264-266. However, it is not completely understood how native AMPA receptors mediate excitotoxic neuronal cell death, due to their generally poor permeability to calcium ions.
1.5 AMPA Receptor Trafficking: GluR2-lacking AMPA Receptors as sources of calcium influx Indeed the vast majority of AMPA receptors in the CNS exhibit a low permeability to divalent cations, due to the presence of the GluR2 subunit in the receptor heteromer. Accordingly, a reduction of the AMPA receptor GluR2 content would be expected to have a dramatic impact on neuronal physiology and resistance to excitotoxic injury. AMPA receptors lacking GluR2, as discussed earlier, exhibit higher single channel conductances as well as permeability to both calcium and zinc. These attributes make the GluR2-lacking AMPA receptor a powerful mediator of neuronal signaling. Many physiological and pathophysiological processes involve the dynamic regulation of
50
the AMPA receptor GluR2 content, which itself is not static but subject to remodeling from a variety of neuronal inputs. Changes to AMPA receptor GluR2 levels have been observed during synaptic plasticity (including the induction of LTP267-269 where GluR2lacking AMPARs play a role in increasing basal synaptic strength), but also in disease states, including drug abuse270-275, epilepsy276-278 and ischemia279-286. In the latter scenarios, it is clear that the expression of calcium-permeable AMPA receptors during periods of excitotoxicity imparts neuronal vulnerability to cellular injury due to augmented cytosolic Ca2+ loads. While some investigators have suggested that the NMDA receptor mediates the majority of glutamate-dependent excitotoxicity, it has been shown that in fact over-activation of calcium-permeable AMPA receptors – when they are expressed – results in levels of neuronal cell death similar to loading cells with Ca2+ via NMDA receptor activation287. Investigators have also repeatedly demonstrated that calcium entry through calcium-permeable AMPA receptors triggers marked intracellular production of reactive oxygen species as well as severe mitochondrial depolarization and injury comparable to that produced by excessive stimulation of the NMDA receptor288-290. Accordingly, oxygen radical scavengers and inhibitors of oxygen radical production have demonstrated a marked cytoprotective efficacy against cell death mediated by AMPA receptors291,292. Thus, while GluR2-containing AMPARs require excessive amounts of stimulation to induce neuronal death (likely due to neuronal depolarization and secondary Ca2+ influx through voltage-sensitive Ca2+ channels293-296), the subset of AMPA receptors lacking GluR2 appear to be particularly lethal sources of calcium. Indeed this can also be observed with the widespread neuronal damage that follows the glutamatergic stimulation
51
of certain types of neurons that express endogenous calcium-permeable AMPARs287-290, and the marked neuroprotection afforded by AMPA receptor antagonists in experimental disease states involving excitotoxicity (reviewed by 297). Accordingly, this has promoted a tremendous amount of interest in identifying the mechanisms responsible for the aberrant expression of calcium-permeable AMPA receptors in neuronal populations ordinarily expressing GluR2-containing channels, as this may shed important insight into how AMPA receptors mediate excitotoxic neuronal death.
1.5.1 Modification of the AMPA Receptor GluR2 content. It is clear that the cytotoxic potential of AMPA receptor stimulation is almost entirely dependent on the presence or absence of GluR2 in the receptor complex. Recent studies have demonstrated that the remodeling of the AMPA receptor GluR2 content is a consequence of the redistribution and trafficking of AMPA receptor subunits, as well as epigenetic reprogramming of RNA editing (reviewed by 298). The molecular mechanisms underlying activity-dependent remodeling of the subunit composition and permeability of synaptic AMPA receptors are being further examined as potential targets of antiexcitotoxic therapy, as it appears many of them contribute in critical ways to glutamatedependent neuronal death following CNS injury. 1.5.1.1 Epigenetic silencing of GluR2 Insights from studies of global or transient forebrain ischemia have shed important light on one mechanism of GluR2 mRNA and protein loss in certain neuronal populations during excitotoxicity. Ischemia -- which shares with brain trauma the
52
involvement of excitotoxicity in expanding the primary lesion into widespread neuronal damage – involves an early rise in intracellular calcium and a delayed rise in free zinc, similar to observations made in dying neurons following TBI. The neurons exhibiting these rises in free Ca2+ and Zn2+ (primarily CA1 hippocampal cells) also demonstrate a concomitant reduction in GluR2 mRNA and protein abundance281,285,299,300, inducing a long-lasting switch in AMPA receptor phenotype, from GluR2-containing, to GluR2lacking280,283. Indeed following the excitotoxic input delivered during ischemia, AMPA receptors physiology exhibits marked inward rectification of EPSCs, calcium permeability, as well as sensitivity to polyamine antagonism280,283, three physiological hallmarks of GluR2-lacking receptors. Following ischemia, antagonism of GluR2-lacking AMPA receptors (but not GluR2-containing or NMDA receptors) affords significant neuroprotection, suggesting that excessive activity at these receptors can initiate substantial neuronal loss. This, along with the evidence that acute knockdown of GluR2 protein by anti-sense oligonucleotides causes death of hippocampal neurons even in the absence of CNS injury301, has perpetuated the hypothesis that a reduction of GluR2 protein is a causal mechanism of cell death during excitotoxic injury. The loss of GluR2 in ischemic injury is thought to involve transcriptional silencing of its mRNA production. GluR2 mRNA transcription is under strict control by repressor element-1 silencing transcription factor (REST), a transcriptional repressor that actively represses neural specific genes important to synaptic plasticity and development302-304. For example, REST functions, using epigenetic modifications, to silence target genes in neural progenitor cells during development to maintain particular receptor phenotypes305,306. However, under pathological scenarios where a loss of a
53
particular gene is undesirable, REST can mediate neuronal death. In CNS injury, REST binds the GluR2 promoter, and functions through chromatin remodeling to suppress GluR2 protein in neurons destined to die from excitotoxic insult281,285. Indeed it has been shown that acute knockdown of REST affords significant cytoprotection in ischemia, and preserves GluR2 protein levels281,285. GluR2 editing is also affected by neuronal injury. As discussed, the RNA-editing enzyme adenosine deaminase acting on RNA 2 (ADAR2) is responsible for editing GluR2 RNA to contain the positively charged arginine residue in place of the genomic glutamine, thereby governing the ionic permeability of the AMPA receptor channel pore. Accordingly, investigators have shown that ischemic injury inhibits the activity of ADAR2, rendering a substantial portion of GluR2 RNA in its unedited form, and thereby increasing the proportion of calcium-permeable AMPA receptors307. Indeed this increased population of GluR2-lacking AMPARs imparts neuronal vulnerability to delayed cell death. Direct delivery of ADAR2 or constitutively active cAMP response element binding protein (CREB), which induces ADAR2 expression, restores Q/R editing and protects vulnerable neurons from cell death307. Thus, reduced GluR2 Q/R editing further contributes to neuronal vulnerability in excitotoxic injury. 1.5.1.2 Local trafficking of GluR2 protein
The mechanisms discussed above are intriguing examples of how total GluR2 protein expression can be altered by injury to the brain, and how this can manifest as a susceptibility to delayed excitotoxic injury. However, there are other important mechanisms of GluR2 regulation that do not involve suppression of the protein’s
54
expression, but rather incorporate endo and exocytotic trafficking of the protein both to and from the plasma membrane. Much of our understanding of GluR2 trafficking comes from studies of synaptic plasticity and the induction of long-term potentiation (LTP). LTP involves a long-lasting increase in the efficacy of synaptic transmission, and is generally induced by highfrequency stimulation of afferent fibers. One mechanism through which LTP has been proposed to involve is an activity-dependent switch in AMPA receptor subtype. For example, high frequency stimulation of the Schaffer collateral-CA1 synapse, which ordinarily expresses GluR2-containing, calcium impermeable receptors, causes a transient incorporation of GluR2-lacking receptors, thereby increasing both local calcium influx as well as single channel conductance. Together, these two properties of calciumpermeable AMPA receptors increase basal synaptic strength, initiating a larger postsynaptic depolarization per quanta of pre-synaptic vesicle released. The time scales on which these changes occur (i.e., within minutes) do not favor the hypothesis that these phenomena are observed due to transcriptional silencing of GluR2 mRNA. It is also unlikely that de novo GluR1 protein translation can occur on this time scale. Indeed even with exogenous expression of AMPA receptor mRNA on isolated dendrites, it takes hours for protein to be translated, folded, assembled, exported, and trafficked to the post-synaptic membrane via secretory machinery308. Thus, recent efforts have focused on identifying a more rapid mechanism of modifying the AMPA receptor GluR2 content. It is now known that mechanisms exist in neurons for the subunit specific trafficking of AMPA receptors to and from synapses309. This was first illustrated by
55
recombinant expression of GluR2 homomeric AMPA receptors in hippocampal CA1 neurons, where the channels were found to be constitutively incorporated into synaptic sites310,311. Further work identified that the proteins responsible for GluR2 trafficking regulated the movement of the subunit via protein-protein interactions with the cterminus of the subunit. The best characterized of these interactions are at the proximal N-ethylamide-sensitive fusion protein (NSF)/Adaptor protein 2 (AP2) site, and at the distal post synaptic density protein (PSD95), Drosophila disc large tumor suppressor (DlgA), and zonula occludens-1 protein (zo-1) (PDZ) site (Figure 2).
1.5.1.2.1 NSF/AP2 Site interactions in GluR2 trafficking NSF -- an ATPase that is involved in a number of membrane fusion events312 – interacts directly with the GluR2 carboxy (c)-terminus313-315, and is thought to play a critical role in the stabilization of the subunit’s surface expression. The interaction with NSF on the GluR2 c-terminus is located at a membrane proximal site313. Though the binding site involves a motif that is completely novel to this protein interaction, other proteins may coassemble with the GluR2-NSF complex, including α and β SNAPs314,316 (soluble NSF attachment proteins). At the same site, AP2, an adaptor protein critical for clathrin-mediated endocytosis317, also associates with GluR2318,319. The AP2 binding motif overlaps, but is not identical to the NSF site318,319. The role of these protein interactions was identified using targeted disruption with dominant negative peptide decoys mimicking the binding sites. These experiments revealed that these binding partners are involved in the constitutive and activity-
56
dependent regulation of AMPA receptor surface expression313,315,318,319. Indeed the GluR2-NSF interaction is required to maintain receptor surface expression, with virallyexpressed or intracellularly delivered inhibitory peptides (the most well recognized being pep2m) resulting in either a complete loss of AMPA receptor surface expression320, or a ~40% reduction of AMPA receptor EPSC amplitude313, depending on the experimental preparation. It has also been shown that the loss of this protein interaction is involved in certain types of endogenous synaptic plasticity, including NMDA-receptor dependent long-term depression (LTD). This was noted by a marked occlusion of NMDARmediated LTD by prior treatment of neurons with pep2m321,322. Similar effects have been reported for the AP2 site, which is thought to mediate the recruitment and subsequent formation of clathrin-coated pits during AMPA receptor endocytosis323,324, which also occurs during NMDA receptor-dependent LTD. Some work has hypothesized that during this process cytosolic hippocalcin acts as a calcium sensor, linking NMDAR-derived calcium to AP2-dependent internalization of AMPA receptors325.
1.5.1.2.2 AMPA receptor c-terminal PDZ interactions Three proteins, glutamate receptor interacting protein (GRIP)326, AMPA receptor binding protein (ABP, also known as GRIP2)327, and protein interacting with C kinase 1 (PICK1)328,329 interact with the AMPA receptor at the extreme c-terminal PDZ binding site. Prior to describing the nature of these interactions and their role in AMPA receptor trafficking, a brief discussion of PDZ domains is necessary.
57
Figure 4. GluR2 subunit domain structure. The C-terminal amino acids are shown, to help identify the binding sites for NSF/AP-2 at the membrane proximal site (blue) and the distal PICK1 binding site (at the PDZ ligand, green). Transmembrane domains are shown in yellow and the flip/flop alternative splicing region is shown in grey shade. Palmitoylation sites are noted by arrowheads, as is the Q/R editing site in transmembrane domain 2, where the critical RNA editing occurs in the GluR2 pore, controlling ionic permeability. Other alternative splicing regions are underlined, and in vivo phosphorylation sites are denoted by bolded residues.
58
Figure 4. GluR2 subunit domain structure Modified with permission from Isaac et al., 2007. “The role of the GluR2 subunit in AMPA receptor function and synaptic plasticity”. Neuron, 54, 859-871.
59
PDZ domains are modular protein-interaction domains of approximately 90 amino acids that function in specialized binding to the extreme c-terminal sequences of other proteins. PDZ domains (of which ~ 440 have been identified in 259 different proteins in humans330,331), are named after the first three proteins identified as carrying them; the postsynaptic density protein PSD-95/SAP90, the Drosophila septate junction protein Discs-large, and the tight junction protein zonula occludens-1 (ZO-1). Since their initial identification, PDZ and PDZ-like domains have been recognized in numerous proteins from organisms as diverse as bacteria, plants, yeast, metazoans, and Drosophila and are among the most common protein domains represented in sequenced genomes331. PDZ domains generally function as scaffolds as part of an assembly of large multimeric protein complexes, involved in signal transduction and protein trafficking. Any one protein may contain more than one PDZ domain, and may also contain PDZ domains of differing specificity. Based on their general ligand specificity, PDZ domains can be broadly divided into several categories. Type I PDZ domains, including those found on PSD-95 and its homologous family members, bind carboxy termini with the following consensus amino acid sequence332,333: Type I PDZ domain: (S/T)-2X-1(V/I/L)0
where S represents serine, T represents threonine, V represents valine, I represents isoleucine, and L represents leucine. X in this scenario can represent any amino acid. The superscript numbers above the amino acid symbols represent the relative position of
60
the amino acid relative to the c-terminal end (i.e., 0 represents the most c-terminal amino acid in the entire protein sequence). An example of a –COOH terminus containing this sequence would be that of the NMDA NR2B subunit, which binds the PDZ domain of PSD-95 via its KLSSIESDV sequence334, thus classifying this interaction as a type I PDZ interaction. Conversely, type II PDZ domains are found in proteins such as the fourth and fifth PDZ domain of GRIP and the PDZ domain of calmodulin-sensitive kinase (CASK), and bind a consensus sequence of333:
Type II PDZ domain: X-3-Φ-2- X-1-Φ0
where X represents any amino acid, and Φ represents a hydrophobic residue, (preferably tyrosine or phenylalanine at P−2)335,336. A third type of PDZ domain present in neuronal nitric-oxide synthase shows a preference for aspartate at P−2 (i.e., a DXV c-terminal motif)337,338, although it also accepts other residues (e.g. isoleucine)339. Additionally, another kind of binding has been described and is exemplified by an internal (non c-terminal) sequence in neuronal nitricoxide synthase that binds to syntrophin's PDZ domain and the second PDZ domain of PSD-95340,341. It should be noted that possession of these consensus sequences does not guarantee that the protein is involved in a PDZ interaction. Similarly, it is clear that other residues must contribute to specificity for a given PDZ domain. Notably, several proteins and ion channels that have an (S/T)XV motif do not bind PDZ domains under conditions
61
in which certain other ligands do. For example, the neuronal inwardly rectifying K+ channels Kir3.2 and Kir3.3 – which possess COOH-terminal sequences in both cases of ESKV – the Na+ channel Nav1.5 (c-terminus of ESIV), and diacylglycerol kinase ζ (cterminus of ETAV) do not bind to PSD-95342-344, which is well recognized for accepting type I PDZ ligands as binding partners. Additionally, the β1 adrenergic receptor does not interact with either of the first two PDZ domains of PSD-95, despite conforming to the (S/T)XV motif345. Thus, beyond these consensus sequences, there are other auxiliary factors involved in stabilization of a PDZ interaction. These factors are reviewed elsewhere330. The common structure of PDZ domains contains six β strands, (βA–βF), and two α helices (αA and αB), which fold in an overall six-stranded β sandwich. The c-terminal peptides discussed above bind the PDZ domain as an anti-parallel β strand, in a groove between the βB strand and the αB helix. Within the βA–βB connecting loop, there is a conserved sequence of Gly-Leu-Gly-Phe (GLGF), which participates in hydrogen bond co-ordination of the c-terminal carboxylate group. For example, in a type I interaction, the serine or threonine residue on the ligand c-terminus occupies the −2 position, where the side chain hydroxyl group forms a hydrogen bond with the N-3 nitrogen of a histidine residue at position αB1 in the PDZ domain itself. In type II interactions, the hydrophobic residue at the −2 position of the peptide ligand interacts with a similar hydrophobic amino acid in the αB1 position of the PDZ domain. Finally, PDZ protein-protein interactions can be modulated through phosphorylation of certain residues on the c-terminal ligand. For example, serine phosphorylation at position −2 in the inward rectifier K+ channel Kir2.3 by protein kinase
62
A (PKA) disrupts binding to the PDZ domains of PSD-95. The association of β2adrenergic receptor with Na+/H+ exchanger regulatory factor (NHERF) is abolished in a similar fashion by phosphorylation at position −2 by G-protein-coupled receptor kinase 5 (GRK5). Not only can phosphorylation disrupt PDZ binding, but it can also regulate specificity of PDZ protein interactions, with certain PDZ substrates preferring phosphorylated moieties. A well characterized example of this is in the phosphorylation of the GluR2 c-terminal serine residue at P-3, which enhances the binding of its c-terminal PDZ ligand (SVKI) to PICK1, while disrupting the interaction with GRIP (discussed in detail below).
1.5.1.2.3 PDZ Interactions in GluR2 trafficking Because AMPA receptors themselves lack motor domains, the receptors must associate with protein partners that assist in their trafficking. The extreme c-terminus of GluR2 contains a type II PDZ binding motif (SVKI) that is involved in the trafficking of the subunit both to and from the plasma membrane. The interaction of this PDZ ligand with its various PDZ binding partners is regulated by the phosphorylation state of the P-3 serine residue. Constitutively, this serine is not phosphorylated, stabilizing the interaction of the SVKI motif with the 5th PDZ domain of membrane-bound GRIP, a 7 PDZ domaincontaining AMPA receptor anchoring protein lacking a catalytic domain. The 4th PDZ domain of GRIP plays a role, through intramolecular interactions, in stabilization of SVKI-GRIP binding. It was through mutagenesis analysis that the role for the GRIPGluR2 interaction was first revealed. In transfected hippocampal neurons, GluR2 mutants lacking the PDZ binding motif did not accumulate at synapses in the manner seen with
63
wild-type subunits. Moreover, mutating a single residue, preventing GluR2-GRIP binding, reduced synaptic accumulation of GluR2, suggesting that the role of the GluR2GRIP PDZ interaction is in preventing endocytosis of the subunit. Therefore, following NSF facilitated fusion of GluR2-containing vesicles with the post-synaptic membrane (discussed previously), GRIP acts as an anchor, to stabilize surface subunit expression346349
. To complicate matters, the synaptic expression of GRIP itself is regulated through
N-terminal cysteine palmitoylation350,351 (i.e., the covalent attachment of a 16 carbon saturated fatty acid352). This modification of the GRIP N-terminus is required for trafficking of the protein to synapses, with unpalmitoylated isoforms of GRIP residing exclusively in the cytosol353. The GluR2 c-terminal PDZ ligand also interacts with another PDZ-domain containing protein known as protein interacting with C kinase alpha 1 (PICK1). PICK1 is a peripheral membrane protein initially cloned as one of the proteins interacting with protein kinase C α (PKCα) from a yeast two-hybrid screen354. It contains two structurally known domains, the PDZ domain, as well as a crescent shaped dimeric Bin– Amphiphysin–Rvs (BAR) domain355, which interacts with negatively charged curved membranes during membrane fusion events356,357, and intramolecularly binds the PDZ domain in an auto-inhibitory fashion358. In addition, there are three regions that border these two domains: a short N-terminal region of 18 residues before the PDZ domain enriched with acidic residues, a linker region of 40 residues between the PDZ and BAR domains, and a c-terminal region characterized with a stretch of acidic residues355.
64
The structure and function of PICK1 are quite unique among the human genome; PICK1 is the only known protein to contain both a BAR domain and a PDZ domain, and it is also the only known protein to contain a PDZ domain that accepts both Type I and Type II PDZ ligands as binding partners (due to the presence of a lysine, K83, in the critical α B1 position)355. To date, PICK1 has been shown to interact with over 40 proteins, most of which are membrane proteins, including receptors, transporters, and ion channels. Of these interactions, the interaction of PICK1 with the c-terminus of GluR2 is the best characterized, both in terms of its biochemical regulation and impact on neuronal physiology. The interaction of PICK1 with GluR2 was first identified through yeast two hybrid screening328,329, and has subsequently been verified through coimmunoprecipitation (CoIP) from heterologous cells359 and later through in vivo CoIP in rat brain homogenates358,360. Indeed when expressed in heterologous cells, PICK1 and GluR2 form many co-clusters that are abolished upon mutation of the PICK1 PDZ domain (K27D28 to AA) or deletion of the GluR2 c-terminal PDZ ligand329,357. It is now known that the function of the PICK1-GluR2 interaction is in modification of GluR2 surface expression, with the vast majority of studies in this area pointing to a role for this interaction in GluR2 endocytosis. That is, binding of PICK1 to the GluR2 c-terminus has been identified as a critical event in the internalization of this subunit from the post-synaptic plasma membrane. The evidence supporting this hypothesis is extensive. Through immunohistochemistry, surface biotinylation, and subcellular fractionation of membrane components, many investigators have shown that PICK1 transfection into hippocampal neurons and heterologous cell lines reduces surface
65
GluR2 protein expression267,268,360,357. Further, known mechanisms of inducing the endocytosis of GluR2 in cultured neurons (including bath application of NMDA and PKC-activating phorbol esters) are inhibited through mutation of the PICK1 PDZ domain or peptide-mediated interference of GluR2-PICK1 binding361-365. The interaction of GluR2 with PICK1 is preceded by a number of important biochemical events that regulate this protein complex. The most critical of these events is the phosphorylation of the GluR2 P-3 serine residue by protein kinase C alpha (PKCα). As discussed, the GluR2 c-terminal SVKI PDZ ligand is constitutively bound to GRIP, to anchor the subunit to the membrane. However, the phosphorylation of the serine residue within this moiety (serine 880, or S880), by PKCα, is capable of disrupting the GluR2GRIP interaction, whilst favouring an intermolecular interaction between GluR2 and PICK1349,350,362,366-373. Moreover, this phosphorylation of GluR2 is further dependent upon the trafficking of PKCα to the plasma membrane by PICK1, in a second PDZ interaction involving the PKCα c-terminal type I QSAV PDZ ligand with the PICK1 PDZ domain349,360,369,370,372. The binding of PKCα’s PDZ ligand to PICK1 is dependent on activation of the kinase. This occurs endogenously through the influx of intracellular calcium, and exogenously with PKC activators such as phorbol esters. Activation of PKCα exposes the PDZ ligand for binding the PICK1 PDZ domain, by altering its conformation from folded to linear360. Finally, localization of the PKCα-PICK1 complex to the plasma membrane is dependent on an interaction between the PICK1 BAR domain and GRIP, which brings the complex in close structural proximity to GluR2 itself360,367,374. Indeed the steps involved in this GluR2 endocytotic cascade are quite
66
Figure 5. Steps involved in the intracellular trafficking of the GluR2 subunit. 1) Influx of intracellular calcium (usually through the NMDA receptor) activates cytosolic PKCα, freeing up its PDZ ligand (QSAV) for binding available PDZ domains. 2) Binding of PKCα’s PDZ ligand to PICK1’s PDZ domain disrupts the auto-inhibitory interaction between the PICK1 PDZ and BAR domains, exposing the PICK1 BAR domain for binding other proteins. 3) PICK1 traffics activated PKCα to the GRIP/GluR2 complex, through the interaction of the PICK1 BAR domain with a 55 amino acid binding region (Br) sequence in GRIP. The PICK1 PDZ domain is now in close proximity to the GluR2 PDZ ligand SVKI. 4) PICK1 competes with ABP/GRIP for binding the GluR2 SVKI ligand. 5) PKCα phosphorylates serine 880 in SVKI to SPO4VKI. 6) GluR2 phosphorylated at serine 880 is no longer able to bind to GRIP, its synaptic anchor. GluR2 binds PICK1 through the SVKI-PDZ domain interaction. 7) GluR2 is internalized from the cell surface.
67
Figure 5. Steps involved in the intracellular trafficking of the GluR2 subunit. Modified with permission from Lu and Ziff, 2005, Neuron. “PICK1 interacts with ABP/GRIP to regulate AMPA receptor trafficking”. 47, 407-421.
68
complicated, and as a result are outlined for clarification in Figure 5 and its accompanying figure caption. It is now known that this cascade plays a critical role in the activity-dependent trafficking of AMPA receptors, as well as in modulation of the GluR2 content of synaptic AMPARs, thereby controlling critical properties of AMPA receptor biophysics and initiating synaptic plasticity. Forms of synaptic plasticity including LTD and LTP are thought of as cellular analogues of learning and memory, with PICK1-mediated trafficking of GluR2 playing an integral role in these synaptic modifications. Indeed there is good evidence that PICK1-mediated internalization of GluR2 is one of the mechanisms through which neurons increase their basal excitability and calcium permeability.
1.5.1.2.4 GluR2 trafficking in synaptic plasticity Several comprehensive reviews exist on the subject of AMPA receptor trafficking as a cellular mechanism of synaptic plasticity. Accordingly, the balance of this section will focus specifically on the evidence supporting an involvement of PICK1-mediated trafficking of GluR2 in modulating AMPA receptor phenotype. To understand the role of these proteins in initiating changes to AMPA receptor biophysics, an effective experimental strategy is transfection, that is, viral-mediated upregulation of a protein in a native neuronal population or cell line. Indeed studies of PICK1 transfection into hippocampal slices have yielded important information on the role of the PICK1-GluR2 interaction in modulating the properties of surface AMPA receptors. When expressed exogenously, PICK1 increases AMPA receptor EPSC amplitude, induces inward rectification of the current-voltage relationship, as well as
69
confers a sensitivity of the AMPA receptor population to polyamine antagonism, all defining characteristics of GluR2-lacking, calcium-permeable AMPA receptors267,268. Concomitant immunocytochemstry and subcellular fractionation revealed that in fact surface GluR2 expression (but not total protein) had markedly decreased after PICK1 transfection, without an appreciable change in GluR1 surface levels, arguing against a role for PICK1 in AMPA receptor exocytosis in the reported increase in EPSC amplitude267,268. Perhaps more compelling is the requirement for endogenous kinase activity in regulating these effects, as PICK1-mediated GluR2 removal after PICK1 transfection is abolished through PKC inhibition and NMDA receptor blockade, clarifying an integral role for a regulated signaling cascade in modifying the AMPA receptor GluR2 content. Other experiments have corroborated these results. When PICK1 is transfected into the hippocampus in the presence of GluR2 c-terminal peptides (acting as dominant negative decoys for PICK1 binding, thereby disrupting endogenous PDZ interactions), the AMPA receptor phenotype is also unchanged, suggesting a requirement for PICK1-SVKI binding in GluR2 protein removal267,268. Indeed this interaction between GluR2 and PICK1 is also associated with endogenous phenotypic changes to AMPA receptor physiology, specifically during a switch from GluR2containing to GluR2-lacking receptors276,373,375. Collectively, these experiments provide a compelling role for PICK1 in decreasing the GluR2 content of synaptic AMPA receptors, resulting in an increase in both synaptic strength and calcium permeability through the increased expression of GluR2-lacking receptors. The PICK1-mediated switch in AMPA receptor subunit composition is reminiscent of the previously discussed observations that the AMPA receptor GluR2
70
content is reduced following the induction of LTP269,376, resulting in a higher average single channel conductance and strengthening of synaptic inputs. Therefore, investigators have examined the role of the PICK1-GluR2 protein interaction in initiating LTP. Firstly, experiments that replicated the PKC and NMDA receptor-dependent effects of PICK1 transfection on AMPA receptor function also showed that LTP is occluded after PICK1 upregulation268. Secondly, acute knockdown of PICK1 expression with the use of shRNA interferes with the initiation and maintenance of hippocampal LTP268. Thirdly, expression of PICK1 binding, PDZ-ligand peptides mimicking the GluR2 c-terminus interfere with the development of LTP268. Finally, LTP is absent in hippocampal slices taken from PICK1 knockout mice268. These experiments examining the physiological role of PICK1 demonstrate a clear role for the PICK1-mediated decrease in surface GluR2 in synaptic strengthening as well as increasing neuronal calcium permeability.
1.5.1.2.5 GluR2 trafficking in TBI A loss of surface GluR2 protein following injury to the CNS is, by all logical assumptions, an undesirable situation. The most problematic consequence of reduced GluR2 surface expression is probably heightened neuronal calcium permeability, which as discussed, would predispose neurons to excitotoxic cellular injury. Indeed a reduction in the population of AMPA receptors containing GluR2 would not only impart susceptibility to elevated extracellular glutamate, but might also impart lethality upon synaptic concentrations of glutamate that under other circumstances remain innocuous. Certainly the neuroprotective effects of sustaining surface GluR2 were demonstrated in
71
the studies highlighting REST-dependent decreases in GluR2 mRNA following ischemia280,283, which are likely due to resilience to excitotoxic injury. There is evidence to suggest that modification of the AMPA receptor GluR2 content might occur following TBI. As highlighted previously, traumatic injury to neuronal cultures dramatically increases AMPA receptor mediated depolarizations249,250, supporting a role for receptors with a higher single-channel conductance in neuronal signaling after TBI. This is further supported by the findings that calcium-permeable AMPA receptors appear in cortical neurons following in vitro stretch injury and in vivo spinal cord trauma, studies which indeed demonstrate a GluR2-lacking AMPA receptor phenotype in traumatically injured neuronal populations252,377. Also, excessive stimulation of NMDA receptors occurs following trauma378, providing the necessary NMDA receptor stimulation and calcium influx required for PKC activation during GluR2 endocytosis. Indeed studies have shown that following TBI, PKC activity markedly increases, and moreover undergoes a translocation from a constitutively cytosolic residence to membrane-bound, suggesting PKC-dependent modification of membrane-embedded substrates379. Plenty of evidence also exists for the neuroprotective effects of AMPA receptor antagonism after TBI380-382. Although this does not necessarily suggest that GluR2lacking receptors contribute to neuronal physiology, it does highlight the possibility that pathological events are initiated at AMPAergic synapses, which are known to initiate substantially more cell death when GluR2-lacking receptors are present (discussed previously). The cytoprotective effects of AMPA receptor blockade might be due not only to a reduction of intracellular calcium, but also zinc. It is well recognized that
72
AMPA receptors lacking GluR2 are highly zinc permeable189,383, and there exists indeed a marked cytotoxic elevation in free ionic zinc in neurons following experimental brain trauma384-387. Free zinc is taken up by mitochondria in an effort to restore zinc homeostasis but, similar to the effects of mitochondrial calcium uptake, this leads to potent mitochondrial dysfunction, prolonged loss of mitochondrial membrane potential and free radical generation.
A reduction of GluR2 protein in the traumatic brain might also predispose neuronal populations to situations of relative ischemia, by contributing to heightened metabolic demand at a time where glucose delivery is impaired. Neuronal hyperexcitability (i.e., an enhancement of constitutive depolarizations induced at AMPAergic synapses) is a possible consequence of GluR2 loss, reflected by the previously discussed increases in average AMPA-mediated EPSC amplitude. In the traumatic brain, this might translate clinically to epileptic discharges and increases in cerebral metabolic rate of glucose metabolism (CMRG) 388,389. Coupled with posttraumatic injury to microvasculature (as described, a situation of decreased cerebral perfusion), hyperexcitable neuronal populations create a situation of relative ischemia, which is clinically the biggest contributor to secondary injury after TBI 390 .
1.6 Rationale for proposed study Our understanding of traumatic brain injury (TBI) has evolved considerably from a simple self-limiting physical trauma, to an evolving and progressive biological injury amenable to meaningful intervention. In order to design a therapeutic approach to the treatment of secondary injury and neuronal dysfunction following brain trauma, an
73
understanding of the aberrant molecular events occurring at the cellular level is necessary. Excitotoxicity, a major contributor to secondary injury events after TBI, can occur through two mechanisms as discussed; either through substantial elevation of extracellular glutamate (e.g., triggered by uncontrolled vesicle fusion or a dysfunction of astrocytic glutamate transporters) or alternatively, by changes to the post-synaptic response to physiologic glutamate that render stimulation of glutamate receptors cytotoxic. The present study was undertaken, in a very broad sense, to examine this alternative hypothesis of excitotoxic cell death following traumatic brain injury; that is, an appreciable change to glutamate receptor function that confers neuronal injury during excitotoxicity. Specifically however, since the GluR2 subunit has dramatic control over AMPA receptor ionic permeability and conductance, the study was performed to investigate the possibility that a reduction of surface GluR2 protein contributes to secondary injury following TBI. The study aimed to investigate the involvement of GluR2 trafficking in TBI through biochemical assays (i.e., western blotting, coimmunoprecipitation, and immunofluorescence) as well as any associated changes to AMPA receptor phenotype that result from aberrant GluR2 trafficking with the use of whole cell and field electrophysiology, calcium imaging, and vulnerability to excitotoxic injury. A wealth of information exists regarding the involvement of specific proteinprotein interactions in the regulated endocytosis of GluR2 from the plasma membrane, and this work aimed to examine the activity of these proteins following neuronal trauma, as well as the cytoprotective efficacy of inhibiting the protein-protein interactions responsible for GluR2 internalization.
74
To examine these pathways in mechanistic detail, and yet with sufficient whole animal (and therefore clinical) relevance, it was decided to use a multi-system approach, from a cell-free system in some assays, to a cell culture model, and finally, an in vivo model of TBI. The cell culture model of TBI used in this thesis involved a mechanical stretch injury -- which was first established as exhibiting sublethal properties and therefore suitable for examining the susceptibility of neuronal populations to secondary injury without the added confound of cell death initiated by mechanical trauma – coupled with mild excitotoxicity, to include a model that contains the heterogeneous sequelae of insults (both mechanical and biochemical) faced by injured neurons following TBI. The excitotoxic injury was a low concentration of NMDA, which was applied to the cultures immediately following stretch to mimic glutamatergic excitotoxicity, and more specifically, activation of extrasynaptic NMDA receptors. Excitotoxicity and the activation of these extrasynaptic receptors are documented pathophysiological phenomena noted following both fluid percussion injury in rats, and following TBI in humans. A further elaboration of this model, including the physics of the injury, data on its initial characterization (including dose response curves for cell death and severity of the stretch injury) and further explanation of its rationale is found in the next section. The whole animal model used in this thesis was the well established fluid percussion injury device (FPI), which involves the extradural injection of a column of fluid to the rat brain, a model which reproduces many of the clinical consequences of TBI in humans, including diffuse axonal shearing, contusion, and widespread neuronal injury.
75
1.7 Statement of Hypotheses 1.7.1 General Hypotheses In this thesis, the following main hypothesis was tested: A reduction of surface GluR2 protein contributes to secondary injury after TBI. This hypothesis was fashioned based on the data supporting a role for GluR2 endocytosis in increasing both neuronal calcium permeability as well as single channel conductance. Given that both early sodium-mediated cell swelling and calciumdependent apoptotic cell death contribute to neuronal injury following TBI, we hypothesized that a reduction of surface GluR2 protein might contribute to these processes. Previous findings further report that AMPA receptor conductances markedly increase following trauma and that AMPA receptor antagonism is cytoprotective, observations which might involve a mechanism of surface GluR2 downregulation. Collectively, we conjectured that this post-synaptic modification of AMPAergic synapses might underlie excitotoxic neuronal death after TBI.
1.7.2 Specific Hypotheses In addition to our main hypothesis which proposes that GluR2 endocytosis contributes to cellular injury after TBI, we sought to investigate the mechanisms responsible for its internalization. This lead to the following sub-hypotheses:
i)
Post-TBI GluR2 endocytosis is mediated by the intracellular machinery responsible for constitutive and activity-dependent trafficking
This sub-hypothesis is based on the extensive literature highlighting the involvement of specific intracellular cascades in modulating the GluR2 content of surface AMPA
76
receptors. Specifically, this study examined the involvement of the NMDA receptor in mediating GluR2 internalization after trauma, as well as the role of PICK1 PDZ interactions in injury-induced AMPA receptor modification. The purpose of examining these cascades was to differentiate between non-specific effects of GluR2 internalization – for example, those mediated by applying a mechanical force – and the involvement of regulated intracellular signaling in GluR2 trafficking. ii) Targeted inhibition of GluR2 endocytosis increases cellular survival after TBI The purpose of this hypothesis was to examine the relevance of GluR2 trafficking in neuronal survival after TBI. Although it is possible that GluR2 trafficking occurs following injury to the CNS, the cytotoxic relevance of this phenomena is unknown. Accordingly, it is of utmost importance when trying to parse out mechanisms of neuronal death and dysfunction after injury to identify if various phenomena actually contribute to cell death or if their effects are tangential (or even endogenously cytoprotective). To inhibit GluR2 internalization, we designed a custom, cell-permeable peptide inhibitor of PICK1-PKCα protein binding, which was validated as an inhibitor of this protein-protein interaction prior to its introduction as a putative cytoprotective compound. Further detail on the design and testing of the compound is presented in the following chapter. The primary use of this PICK1-PKCα inhibitor was not in an effort to examine a novel treatment for TBI. Rather, this compound was used to validate the involvement of a specific mechanism in GluR2 trafficking. Some of the final experiments did employ this inhibitor in cell survival assays, but this was more with the purpose of examining the role of GluR2 endocytosis in conferring vulnerability to neuronal damage.
77
1.8 Statement of Objectives In order to address our hypotheses, the following specific aims were defined: 1) To investigate if calcium-permeable AMPA receptors contribute to neuronal physiology after TBI 2) To examine the intracellular mechanisms responsible for the expression of CPAMPARs 3) To investigate the physiological significance and cytoprotective efficacy of interrupting the expression of CP-AMPARs after TBI.
78
Chapter 2 – Model Characterization and General Methods
79
Chapter 2: General Methods 2.1 Preface The following section contains detailed methodologies for the entire repertoire of methods employed in this thesis. They are subdivided into methods used for the in vitro model of TBI, as well as the whole animal preparation. As the former paradigm involved some characterization (e.g., assays of membrane integrity and a dose-response relationship for cellular injury vs. injury severity), the data on the initial use of the model is presented as well. Each methodology is accompanied by a supporting rationale for its use in examining GluR2 trafficking following traumatic brain injury. This section also contains a list of contributions to the data collected in this thesis.
2.2 In vitro methods All procedures described here were approved by the Animal Care Committee at St. Michael’s Hospital and complied with regulations of the Canadian Council on Animal Care.
2.2.1 Isolation and dissociation of cortical cell cultures For the stretch injury model described below, cortical cultures containing both neurons and glia were prepared from E16-17 Wistar rats (Charles River Laboratories, Wilmington, MA). Primary cultures were grown on 6-well BioFlex culture plates
80
(FlexCell, Hillsborough, NC). Pregnant animals were anesthetized with isofluorane and sacrificed via decapitation. Embryos were surgically removed, isolated from the amniotic sac, and decapitated. Embryo heads were placed in 20 ml 1 x Hank’s Balanced Salt Solution (HBSS, Invitrogen Corp. Carlsbad, CA). Brains were removed and placed in a separate dish containing 20 ml supplemented HBSS. Cerebral cortices dissected from whole brains using microdissection forceps, were incubated in 2 ml of 0.1% trypsin (Sigma-Aldrich, St.Louis, MO) at 37 ºC for 10 mins, and placed in 2 ml HBSS. Tissue was triturated by glass pipette 10-20 times, and seeding medium (DMEM/F-12 containing 10% Horse serum, Invitrogen) was added. Cortical cells were centrifuged for 5 mins at 1200 rpm, triturated again, re-centrifuged at 700 rpm for 1 minute, and seeded in plating medium (Neurobasal medium containing 2% B-27 supplement, 1% Fetal Bovine Serum, 0.5 mM L-glutamine, 25 μM glutamic acid, Invitrogen) onto poly-Llysine (5 μg/ml; Sigma) coated plates at a density of 1 x 106 cells/well. Cell counts were done by loading PBS, Typan Blue (Sigma-Aldrich) and 50 μl of cell suspension into a hemocytometer. Ninety-six hours after isolation, cells were fed with fresh maintenance medium (Neurobasal medium containing 2% B-27 supplement, 0.5 mM L-Glutamine, Invitrogen) containing 10 μM FDU (5 mM Uridine, 5 mM (+)-5-Fluor-2’-Deoxyuridine, Invitrogen) and left to incubate for 48 hours to halt the growth of non-neuronal cells. Cells were fed with maintenance medium every 3-4 days until stretch assays. We used the cells for experiments 11-14 days after isolation consistent with previous in vitro stretch assays 243,248,249.
2.2.2 In Vitro Model of TBI
81
2.2.2.1 Use of stretch injury models in TBI literature As was previously discussed in the section on the biophysics of traumatic brain injury, the rotational strain produced by rapid changes to angular velocity that occurs during many types of TBI is nearly impossible to reproduce in an animal model of TBI, due to the mass effects of the human brain in axonal and somatic injury. As a result, a number of models have been developed to study TBI at the cellular level, by reproducing cellular trauma in an in vitro system. One of these models, which will be discussed here and is used in this thesis, is an electronically controlled pneumatic device that allows the study of morphologic, physiologic and biochemical responses of cultured neurons to trauma. The device used in this thesis to injure cortical monolayers was the Cell Injury Controller II (Custom Design and Fabrication, Virginia Commonwealth University, Richmond, VA, USA, Figure 5). An inlet on this injury controller is connected to a tank of compressed nitrogen, and the controller regulates the pressure and duration of a pulse of air that is delivered through a closed tube system to an adapter that fits with an airtight seal into the top of each tissue culture well. The exact millisecond duration of the valve opening and the air pressure pulse is tightly controlled by a valve and timer (1-1000 msec) on the unit’s controls. Moreover, an external output on the system allows recording of the exact time and duration of the electrical pulse on an oscilloscope or polygraph. The air pressure pulse is delivered by pressing a trigger on the controller unit. The air between the unit and the culture plate is immediately vented into the atmosphere once the air pulse is delivered, allowing a rapid deformation and subsequent rebound of the membrane in the individual wells. The injury
82
process is repeated six times per plate (since each plate contains six wells), and an entire plate (~ 4-6 million cells) is used as an n of 1, per injury condition evaluated. To allow for the deformation of the cells, they were cultured on BioFlex’s SilasticTM tissue culture plates, which consist of a flexible silicone elastomer membrane, with a total growth area of 57.75cm2 (9.62 cm2 per well). When used in conjunction with the cell injury controller, the plates allow for uniform radial and circumferential strain. The model of injury was initially characterized by Ellis et al., in 1995 when they established a dose response-relationship between pressure intensity and both membrane deformation and lactate dehydrogenase release as a marker of cell death. Subsequently, the model has been used by numerous laboratories to characterize mechanisms of secondary injury in TBI. Notably, this is the same model used in the experiments highlighting trauma-induced augmentation of AMPA receptor current density.
2.2.2.2 The Stretch + NMDA model Prior to stretch injury, the culture medium in our studies was replaced with 2 ml HEPES buffered saline (concentrations in mM: 121 NaCl, 5 KCl, 20 glucose, 10 HEPES acid, 7 HEPES-Na salt, 3 NaHCO3, 1 Na-pyruvate, 1.8 CaCl2, and 0.01 glycine, adjusted to pH 7.4 with NaOH). On the basis of data suggesting that forces resulting in tensile elongation following TBI occur in 50 ms or less, the duration of the stretch injury was set to 50 ms. To establish an initial dose response relationship between cellular injury and pressure in our culture system, the applied pressure levels ranged from 2.5 (mild) to 7.5 (severe) pounds per square inch (psi), representative of the rotational acceleration/deceleration injury resulting from rapid changes to angular velocity (ω), and
83
subsequently, momentum (L). According to the formula: Impulse (J) = ΔP = FΔt, (where P = momentum, F = force in newtons, and t = time) and F = pressure (2.9 psi) x area (1.49 in2), we calculated that at 2.9 psi (the pressure used in the majority of the experiments) J(on cells) ≈ 9.6 N·s. In the intact mammalian brain, tissue peripheral to the necrotic core of trauma undergoes not only mechanical strain, but is also subject to excitotoxic glutamatergic spillage 241,391 from dead or dying neurons that are injured during the primary injury event. This post-trauma excitotoxic environment is largely lost in vitro, but may play an important role in progressive neuronal injury. Our intention with this model was also to replicate in culture a similar or equivalent biomechanical loading and biochemical environment as found in in vivo TBI. Thus, immediately following the mild stretch, 10 μM NMDA was added to the wells for 1 hour to mimic this excitotoxic stimulation, a combinatorial method that has been used by a number of laboratories to mimic both mechanical injury and glutamatergic receptor stimulation. Previous studies of this dose of NMDA in cortical cultures have demonstrated no lethality, and in fact promotion of neuroprotection against subsequent challenges 392. To block NR2b and NR2a containing NMDA receptors respectively, 5 µM -[2-(4-Hydroxyphenoxy)ethyl]-4-[(4methylphenyl)methyl ]-4-piperidinol (Co101244) hydrochloride and (2R*,4S*)-4-(3Phosphonopropyl)-2-piperidinecarboxylic acid (PPPA, 100 nm) (Tocris Biosciences, Ellisville, MO) was bath applied with NMDA. Ki values of PPPA are 0.13 and 0.47 μm for NR2A and NR2B respectively, which ensured specificity of our approach.
84
Figure 6. The cell injury controller and schematic of experimental paradigm.
85
2.2.2.3 Toxicity studies: Dose response characterization of stretch pressures
In order to establish the in vitro model, dose–response (injury–cell death) experiments were performed on the culture cells. Neurons underwent varying levels of stretch (2.5–7.5 psi) in 2 ml HEPES buffer as described earlier. Wells were subsequently loaded with 10 µg/ml PI (warmed in 37°C water bath). The quantitative measurements of PI fluorescence were used as a determination of the prevalence of cell death using a Victor3V multiwell plate fluorescence scanner (PerkinElmer, Wellesley, MA, USA) controlled by Workout software (Dazdaq, Finland). All parameters including the size and number of scanning area, the duration of scanning, etc. were kept constant by using the same protocol for all groups. A second dye, fluorescein diacetate (FDA) was used as a marker of healthy, viable cells, as observed by us and others102,393-395. It has been reported that damaged membranes lose their capacity to retain FDA, and thus will not fluoresce396. In brief, immediately following stretch, baseline PI and FDA fluorescence readings were taken, cells were incubated at 37°C in the absence of CO2 and a subsequent reading was taken 20 h later. Cell death along the continuum of mechanical deformation was normalized to unstretched wells exposed to 1 mM glutamate for 1 h (Glu). This exposure routinely produced nearly 100% cell death in our observations, and that of others119,397-399, and thus PI fluorescence for each condition was normalized to these wells. Cell death was calculated according to the formula: Fraction dead = F20 – F0/F20GLU – F0GLU, where F20 = PI fluorescence 20 h post-stretch, F0 = initial PI fluorescence, F20GLU = PI fluorescence of cells 20 h post-exposure of 1 mM Glu for 1 h, F0GLU = initial PI fluorescence of 1 mM Glu exposed wells. Cells exposed to 1 mM Glu were identical cultures from the same
86
To validate our stretch paradigm as an experimental model of delayed neuron death, we first examined cell viability at 20 h post-injury along a continuum of stretch amplitudes as assessed by uptake of propidium iodide. A robust gradient was observed in PI uptake from magnitudes ranging between 3.5 and 7.5 psi (e.g. cell death averaged 23.2 ± 2.45%, n = 3 for 3.5 psi versus 70.1 ± 6.5%, n = 3 for 7.5 psi, Fig. 7). Each magnitude tested resulted in significantly greater PI uptake than the lower pressure tested (P < 0.05– 0.001, Fig. 7). However, stretch at 2.5 psi did not alter PI uptake relative to controls (P > 0.05). Both conditions averaged approximately 11.5 (± 0.85% for control, ± 1.73% for 2.5 psi, n = 3 for both conditions) of the PI uptake relative to wells treated with 1 mM Glu (Fig. 7). Mildly stretched neurons also stained brightly with FDA, whereas severely injured neurons did not (data not quantified, see Fig. 7). This initial data suggested that insult at 2.5 psi does not confer delayed cell death on its own. It was thus termed, “sublethal”, allowing us to examine the impact of this injury on vulnerability to secondary insults.
87
Figure 7. Dose-Response Characterization of stretch injury model A) Injury magnitude-propidium iodide (PI) uptake dose-response curve for stretch pressures ranging from 2.5 to 7.5 psi. The percentage of cell death was normalized to 1mM glutamate exposed cells at 20 h post-stretch. Note the absence of increased cell death at 20 h in mildly stretched neurons (2.5 psi) as compared to control wells. (B) Representative FDA (green fluorescence, marker of viability) and PI (red fluorescence, marker of cell death) micrographs of mildly injured (B1, B2) and severely injured (B3, B4) neurons. Scale Bars = 200 mm. *P < 0.05, ** P < 0.01, *** P < 0.001. Error bars represent SEM, and each condition represents an experiment repeated in triplicate (i.e., n = ~ 3 x 106 cells total).
88
Figure 7. Dose-Response Characterization of stretch injury model
89
2.2.2.4 Carboxyfluorescein assays of membrane permeability As discussed in the introduction, it is possible that stretched cells may exhibit enhanced calcium permeability as a result of changes to membrane permeability. As this thesis intended to employ assays of calcium imaging as well as whole cell electrophysiology, it was important to examine the impact of the stretch injury on membrane integrity prior to proceeding with these experiments to avoid potentially problematic confounds. Plasma membrane permeability following mechanical stretch was assessed by evaluating uptake of the ordinarily impermeant fluorescent molecule, carboxyfluorescein, (CBF, MW = 380 Da, radius = 0.5 nm; Sigma). We adopted this technique (established by Geddes-Klein et al.,245,400) for use in stretch-induced alterations to cell permeability. The technique however, has also previously been implemented to detect permeability changes in electroporated cells401,402. Immediately prior to injury, cells were treated with 100 µM CBF, and nuclei were stained with Hoechst 33 342 (20 µg/ml; Molecular Probes, Eugene, OR, USA). Neurons were stretched in the presence of CBF and incubated at 37°C, 5% CO2 for 10 min to maximize diffusion of the dye into cells245. Cells were then rinsed with buffer to ensure the removal of extracellular CBF. Sections of membranes were detached (0.75 in.), placed in HEPES buffer, and fluorescent images were taken from five different areas per section of membrane. Cells positively stained with CBF were later counted and normalized to the total number of Hoeschtpositive nuclei. This procedure was repeated 3–4 times in each condition, across separate cell isolations.
90
Applications of CBF enabled us to image uptake of an ordinarily impermeable fluorescent molecule and as a result determine immediate, post-stretch alterations to plasma membrane permeability in injured neurons. Representative CBF micrographs are shown in Fig. 8 for control cultures, mildly stretched cultures, and severely stretched cultures. We observed almost no CBF uptake (denoted by bright green staining) in both control and mildly stretched cultures (quantified at 6.5 ± 1.31% and 5.6 ± 1.91% CBF positive neurons, respectively, n = 3 for both conditions, Fig. 8). CBF uptake was significantly higher in severely stretched cultures ( 33.6 ± 4.03%, n = 3, P < 0.001, Fig. 8A). Thus, CBF uptake was a function of the pressure exerted on cultures, and did not increase in mildly stretched neurons relative to controls. It should be noted, however, that changes to permeability that would have occurred more than 10 min post-stretch would not have been accounted for. However, recent work suggests that plasma membrane permeability changes are transient and repaired rapidly following stretch, if they occur at all245.
91
Figure 8. Mild injury does not increase non-specific neuronal cell membrane permeability. A) Representative micrographs of CBF uptake and Hoescht staining in control, mildly injured and severely injured neurons. Based on this data, mild stretch does not confer the development of non-specific membrane holes or tears from mechanical deformation, suggesting the preservation of membrane integrity. As a positive control however, severe injury does significantly increase CBF uptake (P < 0.001). B) Quantification of the percentage of carboxyfluorescein (CBF) positive cells normalized to Hoescht positive nuclei in each condition.
92
Figure 8. Mild injury does not increase non-specific neuronal cell membrane permeability.
93
Our biochemical investigation into GluR2 trafficking following TBI involved two primary assay methods; co-immunoprecipitation and western blotting. Coimmunoprecipitation is an assay method that examines protein-protein interactions, which we used to measure the progression of the GluR2 endocytotic cascade. The rationale for this approach was based on the assumption that heightened activity of the GluR2 endocytotic cascade would yield a more robust interaction between the previously described PDZ proteins upstream of GluR2 internalization. Along this vein, we examined the intracellular interaction between PICK1 and PKCα, the phosphorylation of GluR2 at serine 880 (critical for subunit internalization), the interaction between GluR2 and PICK1, as well as a novel protein interaction between PKCα and PSD-95, which we hypothesized might underlie an NMDA receptor dependence of GluR2 trafficking. The following methods were used during these assays:
2.2.3 Protein extraction and quantification Following in vitro treatment, cells (an entire six well plate, repeated 3 times, for a total of 18 wells per condition) were washed twice with ice-cold HEPES solution. Lysis buffer (250 µL per well) containing protease inhibitors (50 mM Tris-HCl, 1% NP-40, 150 mM NaCl, 1 mM EDTA, 1 mM PMSF, 1 µg/ml aprotinin, 1 µg/ml leupeptin, 1 µg/ml pepstatin, 1mM NaF) was added and cell suspension was agitated at room temperature for 20 minutes. Cell lysates were collected and centrifuged at 4°C (10,000 rpm), and the pellet was discarded. Protein quantification was determined using the modified Lowry method 403. Following quantification, aliquots of 500 µg protein per condition were
94
collected and frozen at -80°C for subsequent immunoprecipitation for proteins of interest. Similar procedures were used in vivo, using homogenates of cortical tissue.
2.2.4 Co-Immunoprecipitation of GluR2 endocytotic proteins All immunoprecipitation procedures were performed at 4°C or on ice. 5 µg of polyclonal rabbit anti-PICK1 (Abcam Inc, Cambridge, MA), or 5 µg of polyclonal rabbit anti-PSD-95 (Chemicon, Billerica, MA) was incubated with 500 µg of cell lysate and mixed by inversion overnight. Before being added to the antibody-lysate mixture, 50 µl of Protein A agarose beads were washed 3 times with 500 µl of PBS (each time spun for 30 seconds at 10,000 rpm). After washing, protein A agarose beads were added to the antibody-lysis complex and incubated overnight to capture the antibody-antigen complex. As a negative control, we also incubated samples in the absence of primary anti-sera, with only protein A agarose. The antigen-antibody-bead complex was collected by pulse centrifugation (centrifuged at 14,000 rpm for 5 seconds). The supernatant was discarded, and the beads were washed 3 times in ice-cold PBS. Bead complexes were then resuspended in 60 µl 2x sample buffer (0.5 M Tris-HCl pH 6.8, 20% glycerol, 10% SDS, 1% bromophenol blue, 5% β-mercaptoethanol), and boiled for 5 minutes. The beads were pelleted by centrifugation, and SDS-PAGE was performed with the supernatant.
2.2.5 SDS-PAGE For western blotting of whole cell lysates, 20 µg of boiled sample was loaded into each lane in 2x sample buffer. For electrophoresis of immunoprecipitated samples, 20 µl supernatant was loaded per lane. For probing of phospho-serine880ct GluR2 and phospho-serine845 GluR1, 7% Tris/glycine gels were used, whereas a 12% gel was used
95
to probe for PKCα following the immunoprecipitation of PICK1 and PSD-95. Protein samples were transferred onto nitrocellulose membranes for immunoblotting.
2.2.6 Immmunoblotting After transfer, membranes were blocked in 5% blotting grade non-fat dry-milk (BioRad) in TBS-T (0.01 M Tris, 0.1 M NaCl, 0.05% Tween 20) for 1 hour at RT. To probe for phosho-serine880ct- GluR2, the immunogen (Chemicon, polyclonal, rabbit, 1:1000, diluted in 5% milk block) was a thyroglobulin-conjugated synthetic peptide corresponding to amino acids 873-883 of rat GluR2, with a phosphorylated serine at position 880 (LVYGIE(PO4S)VKIA). Immunoblotting for total GluR2 was performed with a polyclonal rabbit anti-GluR2 (1:1000, Chemicon) antibody. Phosphorylated GluR1 at serine 845 was detected using a polyclonal antibody to PS845 (Abcam, 1:400). Following immunoprecipitation with PICK1, we probed for PKCα using a mouse, monoclonal anti- PKCα antibody (1:350, Upstate Biotechnology, Lake Placid, NY). We sought to verify this interaction using both the aforementioned monoclonal antibody (1:350) and a separate rabbit anti- PKCα (1:350, Abcam) antibody. Hence, the immunoblots presented using the latter antibody contain a heavy chain IgG band at 55 kDa (because both the immunoprecipitating and immunoblotting antibody were polyclonals hosted in rabbit), whereas the blots using the monoclonal PKCα antibody do not contain the heavy chain IgG band. All primary antibodies were incubated overnight at 4°C. After washing in TBS-T, HRP-conjugated goat anti-rabbit IgG (1:3000, Chemicon) or HRP-conjugated goat anti-mouse (1:3000, Chemicon) was added for 1 hour at RT. We visualized immunoreactivty using an ECL western blotting detection kit (Perkin-Elmer).
96
To verify equal protein loading in whole cell blotting, membranes were re-probed with mouse anti-beta actin (1:2000, Sigma), mouse anti-ERK (1:40,000, Sigma), and HRPconjugated goat anti-mouse (1:3000). For immunoprecipitation, membranes were reprobed for the immunoprecipitating protein (PICK1, 1:500 or PSD-95, 1;1000), and results were normalized to the amount of IP protein per lane. All immunoblotting and immunoprecipitation experiments were repeated in triplicate, with densitometry performed within the linear range of analysis. Densitometry analysis was performed using Gel-Pro Analyzer software (Media Cybernetics, San Diego, CA). Integrated optical density of PKCα in both immunoprecipitation conditions was expressed as a ratio of PKCα:PICK1 or PKCα:PSD-95. All results are normalized to control cultures, which are assumed to represent 100% expression.
2.2.7 Acid Strip Immunofluorescence In addition to examining the protein-protein interactions that underlie GluR2 trafficking, we also wanted to visualize GluR2 protein inside our cultured neurons. To visualize the internalization of GluR2, 2 µg/mL monoclonal anti-GluR2 (Chemicon) recognizing the extracellular N-terminus was bath applied to live cultures in medium. Cells were incubated at 37°C for 10 minutes, and washed with warm HEPES containing buffer to remove unbound antibody. Where appropriate, cells where then subjected to our model of injury. To examine the effect of blocking NR2b-containing NMDA receptors on AMPAR internalization, 5 µM Co101244 hydrocholoride was bath applied with NMDA. Following injury, cells were incubated for 1 hour at 37°C, and washed with ice-cold HEPES buffered saline to stop endocytosis. After the wash, cells underwent a 4 minute
97
acid strip using ice cold solution of 0.2 M acetic acid, and 0.5 M NaCl, pH 2.8. Cells were thoroughly washed in buffer again, and fixed for 15 minutes in 4% paraformaldehyde. After fixation, cells were permeabilized with 0.1% Triton-X (or not permeabilized as a negative control), and anti-rabbit Alexa 488 secondary antibody was applied (1:1000, diluted in 4% NGS in PBS) for 1 hour at RT. As a second negative control, live cells were fixed, permeabilized, and incubated with anti-rabbit Alexa 488 secondary antibody alone (i.e., no primary antisera). To visualize fluorescence, images were acquired on a Leica DMIRE2 confocal microscope using a 20X objective, digitally magnified 16X on dendritic spines. Image capture settings were standardized for all images. A Z-series projection of 3–4 images at 0.5 μm step intervals was used for each image capture and settings were always in the linear range of signal intensity. To quantify dendritic immunofluorescent staining, we examined 1-2 distal dendrites per cell which contained distinct protruding spines and did not exceed 50 μm in length or 3 μm in width of dendritic shaft. Using ImageProPlus software (Silver Spring, MD) we calculated the area occupied by fluorescent puncta for each process, and divided this by the total area of the process. We collected data for 10 cells per condition per trial, and repeated this in triplicate across separate cultures. In each condition, cells were selected under bright field optics, and the investigator was blind to the treatment condition. Spine sizes were quantified by measuring the diameter of the spine head (after fixation) using Image J. A line was manually drawn in image J across the head of the spine, which was then converted from pixels to micrometers using the scale bars of the
98
image. For the various types of dendritic spines, we measured the maximum diameter (i.e., the head of a mushroom and thin spine, the base of a stubby spine).
2.2.8 [Ca2+] Measurement Our hypothesis was that GluR2 internalization would deregulate intracellular calcium homeostasis, and so it was necessary to visualize intracellular calcium dynamics. Cortical cells were incubated with 5 µM Fura-PE3 AM (a calcium-binding dye, Teflabs, Texas) for 40 min at 37ºC, and then washed three times with HEPES buffered saline and left to incubate further for another 40 minutes to maximize dye hydrolysis. In our model of mild injury, neurons were incubated with the dye for 40 minutes along with 20 μM Naspm (1-Napthylacetyl Spermine, Sigma, to selectively block GluR2-lacking AMPARs). Cells were washed, injured, and allowed to incubate for 60 minutes, to remain temporally consistent with previous assays. Circular selections of membranes (0.75’’ diameter) were then removed from the well using a membrane sectioner, and perfused in HEPES buffer at room temperature. After collecting 150 seconds (30 epochs of 5 seconds each) of stable baseline data, cells were perfused with HEPES containing 100 µM AMPA and 50 µM cyclothiazide (CTZ, for allosteric regulation of desensitization) for 45 epochs, and then returned to control HEPES. Cells were excited for 500 ms alternately at 340 and 380 nm at 5 second intervals, and an image from each excitation wavelength was captured using a High Performance cooled CCD camera (Sensicam, Cooke, Eugene OR). Volumetric flow rate of both HEPES buffer and AMPA + CTZ containing buffer was 1.2 mL/minute. The emission intensity at 340 nm was
99
divided by the intensity at 380 nm, to calculate increases in cytosolic calcium. Figure 5a provides a temporal schematic of our calcium imaging experiments. To analyze regions of interest, cells were selected using SlidebookTM software (Intelligent Imaging Innovations Inc, Denver, CO), with three parameters monitored by the experimenter: emission at 340 nm, emission at 380 nm, and the ratio of the two values. Calcium imaging was done in triplicate for control cells, and quadruplicate in the injury condition, across separate cell culture isolations. Calcium imaging of injury + Naspm treated cells was also repeated in triplicate. We quantified three data parameters: i) the amount of time between peak emission and return to baseline, ii) integration of the area under each calcium curve as an estimate of the relative quantity of excess cytosolic calcium, and iii) compared values of peak emission during AMPA + CTZ perfusion (normalized to baseline ratios).
2.2.9 Secondary AMPA Toxicity We further sought to examine the vulnerability to excitotoxicity of neurons that had exhibited an internalization of GluR2 protein. Cortical cells underwent stretch alone, stretch + 10 uM NMDA, or 10 uM NMDA alone in 2 ml HEPES buffer as described above, and left to incubate for 1 hour at 37ºC. Wells were subsequently loaded with 10 μg/ml propidium iodide (PI) warmed in 37 ºC water bath. The quantitative measurements of PI fluorescence were used as a determination of the prevalence of cell death using a Victor3V multi-well plate scanner (PerkinElmer, Wellesley, MA) controlled by Workout software (Dazdaq, Finland). All parameters including the size and number of scanning area and the duration of scanning were kept constant by using the same protocol for all
100
groups. One hour following stretch, 10 μM NMDA, or the dual insult, baseline PI readings were taken. Cells were subsequently exposed to 30 μM AMPA and further incubated at 37 ºC in the absence of CO2 for 1 hour. Cells were washed with buffer, and subsequent readings were taken 20 hours later. Cell death in each condition was compared to unstretched wells exposed to 1 mM glutamate (Glu) for 1 hour, which routinely produced a nearly 100% increase in cell death. Cell death was calculated according to the formula: % increase in cell death = F20/F0, where F20 = PI fluorescence 20 hours post insult and F0 = Initial PI fluorescence. Cell death was thus normalized to baseline readings. Cells exposed to 1 mM Glu were identical cultures from the same dissection, in the same plate.
2.2.10 Whole cell electrophysiology Whole-cell patch-clamp recording was performed at room temperature in cultured control neurons, as well as at one hour following injury, to examine any phenotypic changes to AMPA receptor physiology (e.g., sensitivity to polyamines or changes to whole cell current amplitude or frequency). The extracellular solution during recording was comprised of (concentrations in mM): 128 NaCl, 5 KCl, 1.8 CaCl2, 1 Na-Pyruvate, 17 HEPES acid, 20 D-Glucose, 3 NaHCO3, 1 MgSO4, 0.001 tetrodotoxin, 0.01 AP-5. Intracellular solution was comprised of (concentration in mM): 128 CsOH, 111 gluconic acid, 4 NaOH, 10 CsCl, 2 MgCl, 10 HEPES acid, 4 Na2ATP, 0.4 Na3GTP, 30 Sucrose, 0.1 1-napthylacetyl spermine (Naspm), pH 7.3, 299 mOsm. Extracellular solution during sodium-free recordings consisted of: 128 Choline Chloride, 5 KCl, 1.8 CaCl2, 17 HEPES acid, 20 D-Glucose, 1 MgSO4. Holding potential was maintained at -70 mV, and
101
AMPAR-mediated mEPSCs were recorded and filtered at 2 kHz using Clampex software (Axon Instruments, Union City, CA). mEPSC amplitude was assayed using MiniAnalysis software (Synaptosoft, Decatur, GA). Event threshold was set to 5 pA, and each mEPSC was analyzed individually. In examining the effects of 10 μM Tat-QSAV and AAAA, neurons were treated post-injury, and the peptide remained in the wells until recording.
2.3 TAT peptides Molecular cloning of the proteins involved in GluR2 trafficking has yielded a tremendous amount of insight into potential methods of perturbing subunit endocytosis. A widely used method of interfering with protein-protein binding is the exogenous introduction of a primary amino acid sequence mimicking the binding moiety of one of the proteins involved in the interaction. By introducing one of these dominant-negative decoy peptides, the experimenter can effectively bind their protein of interest, thereby perturbing the endogenous protein interaction. For example, if one is interested in interfering with a PDZ interaction, a well-accepted experimental strategy involves occupying the endogenous PDZ domain involved in the interaction with the use of a peptide mimicking the PDZ ligand of the associated protein. Although in theory this is an effective strategy, peptides in general are not cellpermeable, as the plasma membrane of cells presents a tight barrier to the passage of foreign hydrophilic extracellular cargoes. To apply peptides intracellularly, they can be introduced via electroporation, single cell microinjection, or via fusion to a virus, although these methods have a number of shortcomings, including lack of clinical applicability and a massive immune response in the case of viral injection. An alternative approach is the introduction of a molecular chaperone protein that is capable of crossing
102
the plasma membrane, carrying with it the amino acid cargo that will be used in perturbing the protein interaction of interest. One of these chaperones is a protein transduction domain (PTD) produced by the human immunodeficiency virus type 1 (HIV-1) transacting activator of transcription (TAT) protein.
2.3.1 The HIV-1 TAT protein transduction domain The first example of this type of protein transduction was observed when the full length HIV Tat protein was found to be capable of entering mammalian cells and activating transcription from an HIV long terminal repeat promoter construct. Subsequently, studies defined the specific region of the protein that mediated cellular uptake, which was identified as an 11 amino acid arginine rich and therefore highly cationic sequence. This sequence (YGRKKRRQRRR), when fused to other peptides or oligonucleotides, demonstrated membrane transduction properties on its own and allowed fused cargoes to carry out intracellular functions ranging from cytoskeletal reorganization to recombination of genomic DNA. Indeed Tat peptides can also transfer much larger molecules, including 45 nm iron beads, lambda phage, adenovirus, lipsosomes complexed with plasmid DNA, and nanoparticles. The mechanisms of protein transduction have been largely mapped out. It is thought that the cationic charged amino acids present on the Tat PTD allow the peptide to form tight and rapid interactions with ubiquitous extracellular glycosaminoglycans, including heparin sulfate, heparin, and chondroitin sulfate B located on lipid rafts. This hypothesis originally grew out of the observation that externally added heparin or heparinase III inhibits cellular uptake of Tat PTD, as does the interference with lipid-raft
103
dependent macropinocytosis, a specialized form of fluid phase endocytosis. The binding of the Tat PTD to anionic extracellular glycoproteins and phospholipids is therefore thought to be the primary step of protein translocation because it is so electrostatically favourable. Following the stimulation of macropinocytosis and trafficking of the macropinosome, a drop in pH is thought to mediate the release of cargoes into the cytosol or nucleus from enclosed vesicles. This escape from macropinosomes is widely accepted as the rate limiting step of Tat-mediated protein transduction, with a number of experimental strategies seeking to enhance cargo release through photoacceleration strategies and the development of molecules capable of destabilizing macropinosome lipid bilayers, such as the influenza HA2 pH sensitive fusion domain, which enhances Tat peptide release.
2.3.2 Design of PICK1 inhibitory TAT peptides The technique of coupling a transduction domain to a small PDZ-ligand has shown to be extremely effective both in transducing into cortical neurons, and in perturbing protein-protein interactions both in vivo and in vitro119,334. We chose to adopt this technique to perturb the interaction between PKCα and PICK1, thereby interfering with the protein interactions responsible for GluR2 endocytosis. The structural PDZ interaction between PICK1 and PKCα is well established, and as discussed it is known that the extreme C-terminus of PKCα, upon activation by calcium, binds the PICK1 PDZdomain via its unique –QSAV sequence 360,404-407. Thus, we chose to create a 15 amino acid peptide made up of the transduction domain of the HIV-1 transacting activator of
104
transcription (TAT) protein and the unique PKCα PDZ-ligand QSAV, for a final sequence (synthesized by CPC Scientific, San Jose CA) of: Tat:QSAV: Tyr-gly-arg-lys-lys-arg-arg-glu-arg-arg-arg-glu-ser-ala-val
By introducing this exogenous PKC sequence into our cultured neurons (and later into the whole animal brain), it was our intention to bind the PDZ domain of PICK1, thereby inhibiting the PDZ interactions necessary to carry out the internalization of GluR2. Peptides were tagged with a dansyl chloride moiety for visualization of transduction. Control peptides (Tat-AAAA) had an alanine quadruplet in place of the QSAV sequence, which served as a negative control for non-specific effects of peptide transduction on GluR2 trafficking. Indeed this AAAA sequence does not represent a functional binding domain for any known proteins. Peptides were made using solid phase Fmoc chemistry, where the first amino acid was covalently linked to a solid support with the alpha amino group protected by an Fmoc (9-fluorenylmethyloxycabonyl) moiety, as described by 408. Using piperidine (a deprotection agent) the alpha amino group was freed in preparation for coupling the next amino acid in the sequence. Stepwise addition continued until the desired peptide length was obtained. After the last amino acid was added, one additional deprotection step was performed to remove the last moiety on the N terminal amino acid. Peptides were removed from the solid support by adding trifluoroacetic acid (TFA).
To first identify if our peptides could transduce cultured cortical neurons, peptides were bath-applied for 30 minutes at a concentration of 10 µM in HEPES buffer. Cultures
105
were washed thoroughly to remove un-transduced peptide. Subsequently, sections of SilasticTM membranes were cut, removed from the plate, and placed in HEPES buffer. Live-cell fluorescence was visualized by fluorescence microscopy, and fixed cells were imaged using confocal microscopy. All parameters for image capture were kept constant among images (aperture, gain, black level, number of passes for Kallman integration). For confocal imaging, cells were fixed 30 minutes after peptide application, mounted on slides, and imaged (see below for confocal imaging details). We observed that Tat peptides rapidly transduced our cortical cultures, represented by marked dansyl fluorescence in both the soma and dendrites (Figure 10). We further observed that Tat peptides accumulated in coronal brain slices, confirming that our compound was able to transduce the membrane of neurons in vivo (Figure 10). However, unlabeled peptides were used for all experiments not involving visualization to eliminate the possible effects of the conjugate on the actions of the drug.
106
Figure 9. Design of PICK1-inhibitory Tat peptides and mechanisms of Tat-peptide uptake. A) Dansyl chloride was conjugated to both the Tat protein transduction domain as well as the PKCα QSAV PDZ ligand. This QSAV sequence binds the PICK1 PDZ domain. Negative control peptides contained an AAAA moiety instead of the active PDZ ligand. B) Mechanism of Tat peptide internalization. Tat-mediated transduction occurs by macropinocytosis. Cationic peptides bind to cell surface proteoglycans on lipid rafts, stimulating the formation of a macropinosome. Macropinosomes decrease their pH, and the membrane of the macropinosome vesicle destabilizes, releasing intracellular cargoes. Peptide release can be enhanced with the addition of membrane-destabilizing agents.
107
Figure 9. Design of PICK1-inhibitory Tat peptides and mechanisms of Tatpeptide uptake.
108
Figure 10. Transduction of dansyl-Tat-QSAV into cultured cortical neurons and brain slices in vivo. A) Dansyl chloride-conjugated TAT peptides (10 μM) successfully transduce live cortical cultures (fixed images taken at 20 min after peptide application). Scale bars: 10 mm low magnification, 2 mm high magnification. B) Dansyl-Tat-QSAV also accumulates in native brain slices (live images taken at 40x), indicating successful transduction of the peptide in vivo. Note that some neurons contain visible cytosolic accumulation, while others demonstrate marked accumulation of the peptide around the plasma membrane, presumably from different stages of peptide pinocytosis.
109
A)
B)
Figure 10. Transduction of dansyl-Tat-QSAV into cultured cortical neurons and brain slices (cortex) in vivo.
110
2.4 In vivo Methods 2.4.1 Fluid percussion trauma The fluid percussion injury (FPI) model has been extensively characterized in the rat model of TBI409. In brief, male Wistar rats (280-350 g) were anesthetized with 2.0– 2.5% halothane delivered in compressed air. Temperature was maintained by a thermal heating blanket at 37°C. A craniotomy (~ 2–3 mm diameter) was performed in the right lateral hemisphere, such that the medial edge of the craniotomy was approximately 2 mm from the midline suture, midway between bregma and lambda. A polyethylene tube was fixed to the opening with cyanoacrylate adhesive and dental acrylic, filled with 0.9% isotonic saline and attached to the FPI device. Rats were subject to a 2.0 atmosphere extradural fluid percussion impact. Bone wax was used to close the hole in the skull, and scalps were sutured prior to recovery in a temperature-controlled chamber. Tat peptides (dissolved in ddH2O) were administered via intraperitoneal or intravenous injection at 1-3 mg/kg after closure of the head incision (i.e., approximately 10 minutes after the impact).
2.4.2 Slice Electrophysiology Similar to our in vitro recordings, we sought to examine any changes to the primarily AMPA receptor-mediated hippocampal field responses following fluid percussion trauma. All rat slice recordings were performed between 3 and 6 hours after fluid percussion trauma. Stimulation (0.1 ms in duration) was delivered using a bipolar tungsten electrode over a range of 40-90 μA generated by a Grass S88 stimulator (Grass Instrument, West Warwick, RI) and delivered through a PSIU6 isolation unit. Recording electrodes were pulled from filamented borosilicate glass capillary tubes with a P-97 Flaming/Brown micropipette (Sutter Instruments Co.). Pipettes with resistances of 2– 111
3 MΩ were backfilled with 150 mM NaCl. Signals were digitally recorded using an Axopatch 200B amplifier (Axon Instruments, Foster City, CA). All recordings were performed at room temperature and analyzed by pCLAMP software (Axon Instruments). Extracellular solution (perfused at a rate of 7 ml/min) during all recordings consisted of (concentration in mM): 126 NaCl, 3 KCl, 1.4 KH2PO4, 2.4 CaCl2, 1.3 MgSO4, 26 NaHCO3, 20 glucose, and, when necessary, 0.02 Naspm, bubbled with carbogen (95% O2, 5% CO2), 285 ± 5 mOsm. One hour after FPI, animals were decapitated, and the brains were extracted in ice-cold aCSF. Recordings were performed on 450 µm transverse hippocampal slices. Slices were acclimated to room temperature for a minimum of 60 minutes prior to recording. Recording electrodes were placed in the CA1 stratum pyramidale, with stimulation electrodes placed in the schaffer collateral tract. For sensitivity to the synthetic polyamine 1-naphthylacetyl spermine, 5 minutes of perfusion with control aCSF was followed by a 7.5 minute perfusion with aCSF + 20 µM Naspm. Slices were returned to control aCSF after Naspm treatment. For electrophysiological recordings, Tat peptides were administered to animals at a concentration of 3 mg/kg I.V (in a 1 mL volume of saline as vehicle) following FPI but prior to decapitation and brain slicing.
2.4.3 TUNEL staining To examine the prevalence of apoptotic cell death, we performed terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick end labeling (TUNEL), which labels DNA strand breaks initiated by cleavage of genomic DNA during programmed cell
112
death. The strand breaks, or “nicks”, can be identified by labeling the free 3’-OH terminals with modified nucleotides in an enzymatic reaction. Anesthetized rats were transcardially perfused with 0.9% isotonic saline followed by 4% paraformaldehyde. The brain was postfixed overnight in 4% paraformaldehyde 0.5 M acetate solution before paraffin embedding. Coronal brain slices were sectioned at 10µm thickness. After deparaffinization of sections, slides were treated with proteinase K, and the TUNEL labeling procedure was carried out according to the manufacturer’s protocol. For quantification of TUNEL labeling, three areas of the cortex were taken, each with a field of view 820 x 650 μm (533 mm2). Areas corresponded to medial, core, and lateral to the fluid percussion impact site. Sections were taken from Bregma – 4.3 mm, according to Paxinos and Watson (1998). Data was normalized to the total number of cells (labeled with a Hoescht counterstain) identified in the field of view; this translated to a sampling of approximately 2500-3000 cells per animal.
2.5 Contributions The experimental data presented in this thesis was collected and analyzed entirely by the author. This included isolation and dissociation of cell cultures, protein lysis and quantification, western blotting, co-immunoprecipitation, single cell and slice electrophysiology, calcium imaging, toxicity assays, peptide design and testing, and immunofluorescence. The one exception is the TUNEL staining, which was performed by Ms. Elaine Liu. Quantification for this assay was performed by the author. Technical assistance for molecular biology and electrophysiology, as well as manuscript editing for publication of the findings was provided by Dr Eugene Park. Dr Jinglu Ai, Dr Loren
113
Martin, and Mr. Zikai Zhou, also generously provided their knowledge in electrophysiology to help set up the experiments.
2.6 Statistics All in vitro data are representative of trials repeated at least three times across separate cell culture isolations unless otherwise indicated. In vivo data was collected with an n of 4-6 animals, unless otherwise indicated. Data are presented as mean ± SEM. One-way ANOVAs with post-hoc Tukey tests, or Dunn tests (in cases where tests of normality failed) were used to identify significant differences between treatment conditions in all assays.
114
Chapter 3: GluR2 trafficking in modeled brain trauma
115
3.1 Preface This section of data comprises the molecular biology and biochemistry employed in this thesis. Here, we examined the cell signaling involved in GluR2 trafficking in two models of brain trauma: in vitro cortical injury, as well as in the lateral fluid percussion injury preparation. The assays examined the various critical endpoints in GluR2 endocytosis, and accordingly, the involvement of certain proteins in mediating posttraumatic internalization of GluR2. The first component of this section involves presentation of the in vitro findings; western blotting of GluR2 phosphorylation, coprecipitation of GluR2 trafficking proteins, and acid strip immunofluorescence on dendritic spines of cortical cultures. The second component examines these phenomena at an acute time point following whole animal trauma.
3.2 Phosphorylation of GluR2 serine 880 following in vitro trauma correlates with susceptibility to AMPA toxicity In our first assay, we sought to examine the prevalence of phosphorylated GluR2 in traumatized cortical cultures. Indeed the plethora of experimental evidence in the literature that suggests GluR2 endocytosis is preceded by PKCα-dependent serine 880 phosphorylation. 349,369 led us to investigate this post-transcriptional modification, as we hypothesized this might influence the synaptic composition of GluR2. We examined this protein modification and its role in delayed cell death in our previously established model of sublethal mechanical trauma followed by mild excitotoxicity. Relative to control cultures, our model of TBI produced a rapid increase in detectable levels of serine 880 phosphorylated GluR2 [GluR2 phosphorylation = 164 ± 10.3% of control, p < 0.01, 116
Figure 11A and 11B, but see also quantification in Figure 12]. However, because our injury employed two distinct insults to the cultures to mimic TBI (mechanical strain as well as application of NMDA), we further examined the impact of these individual injuries on GluR2 phosphorylation in an attempt to parse out which component might be responsible for the reported effect. Notably, neither stretch injury alone nor application of 10 μM NMDA for 1 hour had a significant effect on GluR2 phosphorylation [GluR2 phosphorylation = 109 ± 2.3% of control, p = 0.12 for 10 μM NMDA vs. control; GluR2 phosphorylation = 91 ± 8.3% of control, p = 0.16 for stretch vs. control, Figure 11A and 11B), suggesting a synergistic co-operation between mechanical trauma and stimulation of the NMDA receptor in increasing the phosphorylation of GluR2. Co-operation between stretch injury and stimulation of the NMDA receptor is indeed a phenomenon reported throughout the literature by a number of different investigators, although discussion of the mechanism responsible for this synergy will occur later in this chapter. With the development of a model that increased GluR2 phosphorylation, we sought to investigate the possibility of an increased vulnerability of the injured cultures to AMPA receptor-mediated excitotoxicity. Indeed our hypothesis was that phosphorylation of GluR2 would lead to a reduction of surface protein expression, thereby increasing the inherent cytotoxicity of AMPA receptor stimulation. In line with this hypothesis, our in vitro TBI model of stretch + NMDA resulted in an increased vulnerability of cortical cells to a one hour challenge of 30 μM AMPA (24.98 ± 4.8% increase in cell death, p < 0.05, Figure 1D and 1E), evidenced through markedly greater cellular uptake of propidium iodide at 24 hours following early post-injury AMPA receptor stimulation. Accordingly, conditions that did not enhance the expression of phosphorylated GluR2 did
117
not result in delayed cell death following exposure to 30 µM AMPA challenge (stretch + AMPA = 0.97 ± 2.8% increase in cell death; NMDA + AMPA = 5.7 ± 2.5%, p > 0.05, n = 3 cultures, Figure 1D). Stretch + NMDA without a secondary AMPA treatment also did not result in an increase in delayed cell death (4.12 ± 1.4% increase, p > 0.05). Representative micrographs of PI uptake are presented in Figure 11C. Collectively, these initial results suggested that stretch injury coupled with NMDAR stimulation resulting in GluR2 phosphorylation conferred heightened sensitivity to excitotoxic challenge of a dose of AMPA that remained innocuous in normal conditions. This has immediate implications in the pathophysiology of secondary excitotoxic injury after TBI which will be discussed later.
3.2.1 NMDA receptor dependence of GluR2 phosphorylation The synergistic effect of NMDA and stretch injury on GluR2 phosphorylation suggests perhaps a trauma-induced modification of the NMDA receptor that increases receptor calcium influx. However, to further understand the downstream effectors responsible for GluR2 phosphorylation after NMDA application to traumatized cultures, we sought to understand which subpopulation of NMDA receptors might be mediating this effect. As discussed in the introduction, the two predominant subtypes of the NMDA receptor are NR1/NR2A, as well as NR1/NR2B, the former which is primarily synaptic, the latter extrasynaptic. We treated cultures with antagonists of both NMDA receptor subtypes. Significant differences in GluR2 phosphorylation were indeed detected among treatment groups (p < 0.001, F = 26.197). Post-hoc analysis revealed attenuation of GluR2 phosphorylation by selective antagonism of NR2B-containing NMDA receptors (32.7 ± 6.1% of control, p < 0.001, Figure 1F), while antagonism of NR2A-containing
118
NMDARs was ineffective (Figure 11E). Thus, this suggests the likely possibility that extrasynaptic NMDA receptor stimulation is primarily responsible for trauma-induced GluR2 phosphorylation. Data for GluR2 phosphorylation is quantified in Figure 12.
3.3 In vitro trauma increases PICK1-PKCa binding The data suggesting that GluR2 phosphorylation increases in this model of trauma raises the possibility that the endogenous cellular machinery responsible for GluR2 trafficking is activated post-injury. To examine the involvement of GluR2 endocytotic proteins in mediating this post-traumatic phosphorylation, we next incubated cultures with our 15 amino acid TAT peptide (Tat-QSAV) that mimics the extreme c-terminus PDZ-binding motif of PKCα, thereby designed to inhibit the PICK1-PKCα protein interaction (discussed previously as a critical interaction during GluR2 internalization from the cell surface). After confirming successful transduction of the peptide (Figure 10), we examined the effects of Tat-QSAV on the protein interaction between PKCα and PICK1 in the in vitro injury paradigm. First, we observed that Stretch + NMDA significantly augmented PKCα-PICK1 binding (169 ± 5.6% of control; p < 0.01, Figure 12A), an effect similarly attenuated by NR2B-containing NMDA receptor antagonism (65.6 ± 9.6% of control levels, p < 0.01, Figure 12A). Because this interaction is dependent on activation of PKCα, this result suggested an NMDA receptor dependent activation of the kinase, leading to increased PICK1 binding. We further examined the
119
Figure 11. Stretch + NMDA increases S880 phosphorylation of GluR2 and vulnerability to secondary AMPA toxicity. A) Western blot of GluR2 phosphorylation at c-terminus serine residue 880. Stretch + NMDA (but neither condition alone) markedly increased phosphorylation. Membranes were stripped and re-probed for β-actin as a loading control. (B) Data represented in (A), quantified as integrated optical density (IOD) normalized to control values, which are assumed to represent 100% expression. (C) Representative micrographs of propidium iodide (PI) fluorimetry after stretch + AMPA, NMDA + AMPA, stretch + NMDA or stretch + NMDA + AMPA. Scale bars = 75 μm. (D) Plate scanner quantification of PI uptake in all toxicity studies performed. Treatments that do not enhance the expression of GluR2S880ct (stretch alone, or 10 μM NMDA alone) do not increase the vulnerability of cortical cells to subsequent challenge of 30 μM AMPA. (E) Antagonizing NR2b-contaning NMDA receptors attenuates the injuryinduced increase in GluR2 phosphorylation. (F) NR2A antagonism does not reduce GluR2 phosphorylation. ERK1,2 was used as a loading control. **P < 0.01.
120
Figure 11. Stretch + NMDA increases S880 phosphorylation of GluR2 and vulnerability to secondary AMPA toxicity
121
effects of Tat-QSAV on perturbing this increase in PKCα-PICK1 binding, and found that the compound, when administered pre-stretch, successfully disrupted the interaction between PKCα and PICK1 following injury (68.3 ± 16.7% of control, p < 0.01, Figure 12A). Tat-AAAA (our other peptide lacking an intact PDZ binding motif) however, was ineffective (162.9 ± 6.5% of control, Figure 12A) suggesting that the interference with PKCα-PICK1 binding did result from non-specific effects of Tat peptide transduction. Data for these co-precipitation experiments are quantified in Figure 12B. Tat-QSAV, but not Tat-AAAA, also interfered with trauma-induced phosphorylation of GluR2 (respectively, 90.5 ± 18.3% of control, p < 0.05, vs. 161.8 ± 11.1% of control, p = 0.43 vs. stretch + NMDA, Figure 12C), suggesting further that this PKCα-PICK1 increase was potentially involved in the post-traumatic phosphorylation of GluR2. We also assayed for total GluR2 protein expression, which we found did not differ in any of the treatment conditions (p = 0.67, Figure 12C). Thus, the increased phosphorylation of GluR2 did not translate to a reduction of total cellular protein, in contrast to the epigenetic silencing of GluR2 described in cerebral ischemia. Data of total and phosphorylated GluR2 is quantified in Figure 12D.
3.4 PKCa is embedded in the NMDAR complex: PKCa co-precipitates with PSD-95 Our previous data demonstrated that in vitro trauma confers the association of PICK1 with PKCα. It is known that PKCα activation increases its binding with PICK110 and that PKCα activation can occur endogenously via binding of intracellular calcium18.
122
Figure 12. Stretch + NMDA confers association of PKCa with PICK1. A) As assayed through co-immunoprecipitation at one hour following trauma, in vitro injury promotes an NR2B-dependent association of PKCa with PICK1. TaT-QSAV, relative to TATAAAA and injured (untreated) cultures, markedly diminishes bound levels of PKCa to PICK1, suggesting this compound can successfully compete for the endogenous PICK1 PDZ domain. Membranes were stripped and re-probed for PICK1. (B) Quantification of data presented in (A). Data are expressed as the ratio of PKCa/PICK1, and each condition is normalized to control levels. (C) TaT-QSAV, but not TaT-AAAA, also reduces posttraumatic S880 phosphorylation of GluR2, highlighting the potential involvement of the PICK1-PKCa increase in downstream phosphorylation of GluR2. (D) Total GluR2 does not change in any of the treatment conditions. (E) Quantification of total and phosphorylated GluR2 across conditions.
123
Figure 12. Stretch + NMDA confers association of PKCa with PICK1.
124
Given the NR2b-dependence of the PKCα:PICK1 increase, we sought to understand further the mechanism behind increased activation of PKCα in this preparation. PKCα’s type I PDZ ligand (QSAV) has the potential to form a stable PDZ interaction with another protein, PSD-95, which contains three PDZ domains, and is also structurally connected to the NMDA receptor. PSD-95 is a membrane-associated guanylate kinase (MAGUK) scaffolding protein that plays an important role in linking calcium derived from the NMDA receptor (particularly NR2B-containing receptors) to activation of downstream substrates, including for example neuronal nitric oxide synthase (nNOS), a protein activated by calcium influx following ligand-mediated opening of the NMDAR channel. We hypothesized that since the increased association between PKCα and PICK1 was also NMDA receptor dependent (i.e., the binding between these proteins was perturbed by an NR2B antagonist) that a similar scaffold was provided by PSD-95 to PKCα activation. Thus, we attempted to co-precipitate PKCα with PSD-95. We first observed co-immunoprecipitation of PSD-95 with PKCα (Figure 13A and 13B), with PICK1 pull-down, whole cell lysates and bound nNOS (a known binding partner of PSD95) as positive controls (Figure 2H). This association was a novel finding, the first to demonstrate that PKCα might be physically embedded within the NMDA receptor complex. Subsequently, we observed that the PKCα-PSD-95 interaction was also markedly increased after Stretch + NMDA (168 ± 30.3% of control levels, p < 0.05, Figures 13C and D). Both NR2b-antagonism (74 ± 28.1% of control, p < 0.01 compared to stretch + NMDA) as well as Tat-QSAV (123 ± 9.2% of control, p < 0.05 compared to stretch + NMDA) attenuated the increase in PKCα-PSD-95 binding (Figure 2I, J and K). Tat-
125
AAAA had no observable effect (171 ± 22.4% of control, p < 0.01 relative to control, Figure 2J and K). Indeed this data confirmed that the association between PKCα and PSD-95 was likely occurring via the PKC PDZ-ligand, as mimicking this sequence attenuated the interaction. This structural association between PKCa and the NMDAR complex sheds further mechanistic light on how signaling at NR2b-containing NMDARs might lead to post-traumatic GluR2 phosphorylation, and ultimately, endocytosis. 3.5 Traumatic injury increases GluR2 endocytosis
Though we had biochemical data suggesting that GluR2 endocytosis might be occurring (and had identified a possible mechanism of the NMDA receptor dependence of the effect), we sought to have direct evidence for GluR2 internalization from the cell surface. To this end, we employed a protocol known as acid-strip immunofluorescence, a technique which labels surface receptors, allows for endocytosis to proceed, and subsequently strips away any remaining staining on the surface of the neuron with an acidic solution that destabilizes the antibody-antigen complex. When the cells are permeabilized, the assay detects internalized protein that was initially present on the surface of the cell. One hour after our traumatic injury, this acid-strip immunofluorescence revealed significant internalization of GluR2 (ratio of internal GluR2:dendrite area = 0.038 ± .003 relative to 0.012 ± .001 in control cultures, p < 0.001, Figure 14A and 14B). This provided us with evidence that surface GluR2 protein was being internalized into the cytosol. However, since bath application of NMDA can cause GluR2 endocytosis, controls were run with both 10 and 50 μM NMDA alone (the latter as a positive control).
126
Figure 13. PKCa co-precipitates with PSD-95: A potential link from the NMDAR to GluR2 endocytosis. A-B) PKCa co-immunoprecipitates with PSD-95 in cortical cell lysates. PICK1 I.P was used as a positive control for the PKCa immunoblot in (A), and blotting for nNOS was used as a positive control for the PSD-95 I.P in (B). Note the lack of nNOS in the PICK1 I.P. (C) PKCa and PSD-95 exhibit a stronger interaction after Stretch + NMDA. Antagonism of NR2b-containing NMDA receptors with Co101244 attenuates this increase. Membranes were stripped and re-probed for PSD-95. (D) Identical co-immunoprecipitation experiments as outlined in (C), using a polyclonal antibody to PKCa. This antibody also recognized higher levels of bound PKCa to PSD-95 in conditions in which GluR2 phosphorylation was increased. TAT-QSAV attenuates the injury-induced increase in PKCa–PSD-95 co-immunoprecipitation (far right lane), but TAT-AAAA is ineffective. (E) Quantification of co-precipitated PKCa with PSD-95, expressed as the ratio of PKCa/PSD-95, and normalized to control values. # P < 0.05 versus control; ## P < 0.01 versus control **P < 0.01; ***P < 0.001; *P < 0.05.
127
Figure 13. PKCa co-precipitates with PSD-95: A potential link from the NMDAR to GluR2 endocytosis
128
10 μM NMDA on its own did not increase internalized GluR2 (ratio = 0.009 ± .002, 10 μM NMDA vs. 0.012 ± .001, control, p > 0.05, Figure 3A and B), a stark contrast to the effect of this dose of NMDA when combined with stretch injury. As expected, 50 μM NMDA did cause a significant increase in internalized GluR2 (ratio = 0.029 ± .005, p < 0.01 vs control, p > 0.05 vs stretch + 10 μM NMDA). Thus, we observed that stretch injury significantly augmented the GluR2 endocytotic response of a low dose of NMDA, with a synergistic effect of the mechanical injury and the excitotoxin similar to what was observed in the assays of GluR2 phosphorylation. We employed similar antagonistic approaches to what was used in our assays of phosphorylated GluR2. NR2b-antagonism significantly reduced GluR2 internalization (ratio = 0.023 ± .0004, p < 0.05 relative to stretch + NMDA), as did Tat-QSAV (ratio = 0.022 ± .004, p < 0.05). The compounds did not differ significantly in their levels of attenuation (p = 0.47). Importantly, both of our negative controls (non-permeabilized cells and cells incubated only with secondary antibody, Figure 14i) exhibited only diffuse background staining, indicating the efficacy of our acid-strip protocol in eliminating the binding of our primary antibody to surface receptors, as well as the specificity of our staining for GluR2. Thus, the data was in line with our hypothesis that NMDA and PICK1-mediated GluR2 phosphorylation leads to subunit endocytosis. In addition to examining the impact of the stretch injury on GluR2 endocytosis, we also examined cytoarchitechtural changes to the injured neurons. Relative to control cultures (0.37 ± .01 spines per μm, mean spine head diameter = 0.76 ± .04 μm, n = 21 cells) stretch + NMDA also had the incidental effect of decreasing dendritic spine density and increasing the mean diameter of remaining spine heads (0.26 ± .02 spines per μm,
129
mean spine head diameter = 1.14 ± .06 μm, n = 28 cells, p < 0.001 relative to control for spine density and diameter, Figure 3D and 3E). We hypothesized that GluR2 endocytosis was contributing to this morphological damage since surface GluR2 stabilizes dendritic spines through an extracellular interaction between the GluR2 N-terminus and presynaptic N-cadherin 410. Indeed Tat-QSAV preserved dendritic spine density and reduced average spine size in injured neurons (0.38 ± .01 spines per μm, mean spine head diameter = 0.76 ± .02 μm, n = 19 cells, p < 0.001 relative to injured (untreated) for spine density and mean diameter, Figure 3C, 3D and 3E). There was no statistical difference between injured cultures treated with Tat-QSAV and uninjured cultures (p = 0.40 for spine density, p = 0.49 for spine diameter). NR2b antagonism resulted in a mean spine diameter similar to controls (0.83 ± .01 μm, n = 24 cells, p < 0.001 vs injured (untreated), p = 0.11 vs control, Figure 3E) but did not rescue dendritic spine density (0.29 ± .01 spines per μm, p = 0.45, Figure 3D] suggesting that NMDAR blockade was less effective in restoring normal dendrite morphology relative to the Tat-QSAV peptide. These results suggest that preventing GluR2 endocytosis also helps preserve neuronal morphology after traumatic injury, and corroborates the evidence that GluR2 protein was in fact internalized.
3.6 PICK1-mediated endocytosis of GluR2 following fluid percussion trauma Our in vitro findings raised the possibility that traumatic injury to a population of neurons is capable of inducing the trafficking and internalization of GluR2 protein, an AMPA receptor modification that might impart vulnerability to secondary excitotoxicity. To validate this hypothesis, we next assayed cortical and hippocampal GluR2
130
Figure 14. Stretch + NMDA increases GluR2 endocytosis. (a) Inverted confocal phase contrast images of cortical dendritic spines were overlayed with staining of internalized GluR2 after acid stripping. Stretch +10 μM NMDA conferred distinct GluR2-positive puncta in dendritic spines, whereas control neurons did not (far left panel). NR2b antagonism (Co101244) and TAT-QSAV significantly decrease GluR2 internalization after Stretch + NMDA, but GluR2 endocytosis was still higher than controls. In all conditions, arrows indicate spines that stained positively for internalized GluR2; 50 μM NMDA was used as a positive control, and resulted in intense staining along the dendrite of internalized GluR2. (ai) Negative controls of non-permeabilized cells, and cultures treated only with secondary antibody. (b) Quantification of immunofluorescent data expressed as the ratio of internalized GluR2/area of dendrite. *P < 0.05 versus control, # P 0.05 vs FPI).
3.7 Summary of results Our biochemical data employing two models of experimental TBI revealed that neuronal trauma promotes the endocytosis of GluR2 surface protein. We observed in our cortical injury model that GluR2 is phosphorylated at serine 880, internalized from dendritic spines, and that subunit trafficking can be interrupted by perturbing the binding between PICK1 and PKCα. We further identified a likely mechanism of the NMDA receptor dependence of GluR2 phosphorylation, highlighting a novel protein interaction between PKC and PSD-95, the NMDAR-bound scaffolding protein that links
133
Figure 15. In vivo traumatic brain injury (TBI) promotes GluR2 phosphorylation and association with PICK1. (a) Representative immunoblot of PS880 GluR2 in injured cortex 1 h after 2 atmosphere fluid percussion injury. ERK 1/2 is used as a loading control. (b) Representative coimmunoprecipitation of PICK1 with GluR2 and PKCa after forebrain trauma, showing GluR2 endocytosis 1 h after the injury (c) Quantification of all GluR2/PICK1 coprecipitation experiments. (d) TAT-QSAV, but not a control peptide, can perturb PICK1–PKCa protein interactions in vivo. (e) Animals treated after trauma with 1 mg/kg TAT-QSAV show significantly less co-precipitation of GluR2 with PICK1 1 h after injury, suggesting this peptide can effectively prevent GluR2 endcocytosis in injured animals. TAT-AAAA has no effect on the injury-induced increase in GluR2/PICK1 (f) Quantification of GluR2/PICK1 co-precipitation with or without injection of TAT peptides.
134
Figure 15. In vivo traumatic brain injury (TBI) promotes GluR2 phosphorylation and association with PICK1
135
glutamatergic calcium influx to downstream effector proteins. Upon investigation of these effects in vivo, we observed a similar post-traumatic GluR2 phosphorylation. We further reported an upregulation in the association between GluR2 and PICK1, a biochemical indication that the subunit was being internalized from the cell surface. Finally, exogenous interference with the PICK1-PKC interaction following intravenous peptide injection prevented the association of GluR2 with PICK1, suggesting that posttraumatic GluR2 endocytosis in vivo is also dependent on the trafficking of PKC to the plasma membrane by PICK1.
136
Chapter 4: Phenotypic AMPAR changes in modeled brain trauma
137
4.1 Preface The previous chapter described, at the molecular level, changes to the trafficking of the GluR2 subunit that were observed in our two models of traumatic brain injury. The biochemical and immunocytochemical data provided evidence that traumatic injury imparted the endocytosis of GluR2. However, there were no observations made with respect to any phenotypic changes to AMPA receptor behaviour that occurred following this reduction of surface GluR2 protein. As discussed in the introduction, there is compelling basic science evidence that PICK1-mediated endocytosis of GluR2 confers the increased expression of calcium-permeable, GluR2-lacking AMPA receptors, and that these receptors impart neuronal vulnerability to cell death and damage. In this section of the thesis, we examined the effects of GluR2 endocytosis on AMPA receptor-mediated electrophysiology, calcium influx, and neuronal death. This section employed an analysis of post-injury AMPA receptor whole-cell miniature excitatory post-synaptic events, free calcium concentrations, AMPA-receptor mediated field potentials, and finally, the influence of interfering with GluR2 endocytosis on delayed cellular death and apoptosis in both our in vitro injury paradigm and our whole animal TBI preparation.
4.2 AMPAR-mediated mEPSC activity following in vitro traumatic injury Occluding GluR2 endocytosis reduces AMPAR mEPSC amplitude To examine the contribution of GluR2-lacking AMPA receptors to neuronal physiology following traumatic injury, we took advantage of a number of the characteristics of calcium-permeable AMPA receptors. As was discussed in detail in the 138
introduction, it is known that GluR2-lacking AMPA receptors have a higher single channel conductance than receptors containing GluR2189 and are sensitive to polyamine antagonism. To investigate if these changes occurred to the AMPA receptors native to our neuronal population, we performed whole cell patch clamp of neurons at one hour following the traumatic injury, and isolated AMPA receptor mediated responses by antagonizing voltage-gated sodium channels and NMDA receptors. After stretch + NMDA, AMPAR-mediated mEPSCs indeed exhibited significantly larger amplitudes than control neurons (26.76 ± 1.62 pA vs.18.33 ± 0.69 pA, p < 0.01, Figure 4B), as well as a 36.4 ± 5.4 % reduction in amplitude following application of 1-naphthylacetyl spermine (Naspm), a polyamine antagonist of GluR2lacking but not GluR2-containing AMPARs (Figure 16C and D). Control mEPSCs did not demonstrate Naspm sensitivity (control + Naspm = 18.38 ± 0.81 pA), consistent with the presence of predominantly GluR2-containing AMPARs in control cortical neurons. Naspm treatment did not significantly alter the frequency of mEPSCs, which were also unchanged between control and injured cultures [control + Naspm = 0.36 ± .04 Hz; control alone = 0.43 ± .01 Hz; injury = 0.31 ± .06 Hz; injury + Naspm = 0.45 ± .01 Hz, Figure 16E]. There are a number of mechanisms through which AMPA receptor whole cell currents might increase, including phosphorylation of the channel, increased agonist potency, or a reduction of desensitization. To directly measure whether GluR2 trafficking was contributing to the increased whole cell currents we incubated injured cultures with Tat-QSAV prior to patch. Notably, Tat-QSAV reduced mEPSC amplitudes in injured cultures to 14.72 ± 0.95 pA. Tat-AAAA treatment reduced amplitudes to 22.13 ± 0.58
139
pA. Both treatment amplitudes were significantly lower than injury levels (Figure 4E, and 4F). However, mEPSCs were significantly reduced in QSAV treated cultures relative to AAAA treated cultures (p < 0.05), suggesting a significant effect of PICK1 inhibition independent of any effects that peptide transduction alone may have on excitability (Figure 4F). The mechanisms through which Tat peptide transduction might have nonspecific effects on glutamatergic receptor physiology are discussed in the next section.
4.3 AMPA receptor-mediated calcium influx following in vitro trauma: Polyamine antagonism of GluR2-lacking AMPARs lowers cytosolic Ca2+ load The cytotoxicity of AMPA receptor stimulation that occurs following a reduction of surface GluR2 protein is largely dependent on the excessive influx of calcium through calcium-permeable AMPA receptors. Indeed the expression of these receptors is innocuous when extracellular calcium is chelated or removed from the bath. Thus, we sought to visualize post-injury intracellular calcium dynamics following perfusion with AMPA (schematic in Figure 17A), to examine if GluR2 endocytosis augmented cytosolic Ca2+ loads. Prior to stimulation of the cells with AMPA, we recorded baseline calcium following the in vitro injury, to first measure the impact of our model on intracellular calcium. Baseline calcium of control neurons was significantly lower than in neurons exposed to stretch + NMDA (0.11 ± .01 vs. 0.19 ± .01, respectively, p < 0.01, Figure 5B and 5C), indicating the insult affected cytosolic Ca2+ levels prior to AMPAR stimulation. Indeed this observation was critical to our hypothesis, as our work suggested calcium-
140
Figure 16. Traumatic injury in vitro increases AMPAR mEPSC amplitude and sensitivity to intracellular polyamines. A) Representative AMPAR-mediated mEPSC traces of injured and control neurons (one hour after insult) showing an increase in average mEPSC amplitude. Ai) Average mEPSC traces overlaid. Black trace = Control neurons, Red trace = Injured neurons. B) Representative traces showing that the amplitude of AMPAR-mediated mEPSCs is not influenced by inclusion of polyamines (Naspm) in the patch pipette. C) Following trauma however, AMPAR mEPSCs demonstrate sensitivity to Naspm, an antagonist of GluR2-lacking receptors. D-E) Quantification of mEPSC amplitude and frequency in the two treatment conditions. F) Post-injury co-precipitation of PICK1 and PKCα in the presence of Tat-QSAV and Tat-AAAA and resultant mEPSC activity. QSAV-treated neurons exhibited a significant reduction from AAAA treated cells in mEPSC amplitude and bound PKCα:PICK1. G) Quantification of mEPSC amplitudes in all conditions. Neurons were held at -70 mV. ** p < 0.01 vs control; * p < 0.05 vs control; ## p < 0.01 vs injured; # p < 0.05; † p < 0.01 vs Tat-AAAA.
141
Figure 16. Traumatic injury in vitro increases AMPAR mEPSC amplitude and sensitivity to intracellular polyamines.
142
dependent activation of PKCα initially after the injury. We further observed the effects of AMPA receptor activation on intracellular calcium at one hour after trauma. After applying AMPA, peak emissions normalized to baseline ratios did not differ between control and injured neurons (2.07 ± 0.45 x baseline vs. 1.84 ± 0.12 x baseline, respectively, p = 0.19) (Figure 5B and 5E). However, injured neurons exhibited significantly longer calcium extrusion times (5.93 ± 1.59 minutes vs. 1.65 ± 0.51 minutes respectively, Figure 5B and 5D p < 0.01). Integration for the area under the curve as a surrogate indicator of intracellular calcium levels indicated a 2.11 fold larger area relative to control neurons (62.37 ratio·epochs vs. 29.62 ratio·epochs, Figure 5F). We tested the efficacy of Naspm (an antagonist of GluR2-lacking receptors) in reducing peak Ca2+ in injured neurons and in improving calcium extrusion. Baseline calcium of Naspm-treated injured cells was comparable to injured (untreated) cells (0.21 ± 0.01 vs. 0.19 ± .01, respectively, p = 0.13, Figure 5B, and 5C), suggesting that GluR2lacking AMPARs were not responsible for the initial trauma-induced elevation of baseline emission ratios. However, after AMPA application, peak calcium was significantly lower in Naspm-treated injured cells relative to injury (untreated) (1.52 ± .04 x baseline, p < 0.01 vs. values for injury, Figure 5E, and 5G), suggesting a contribution of calcium-permeable AMPARs in the initial rise of Ca2+ in injured neurons during perfusion of AMPA. We have previously shown that Naspm does not impact Ca2+ influx in control neurons 251. Further, the time from peak to extrusion in Naspm-treated injured neurons was 0.88 ± 0.21 minutes, a significant reduction from that of injured (untreated) cells (p < 0.05) but not of control cells (p = 0.12, Figure 5B, and 5D). Integration of the Naspm-treated calcium curve yielded a value of 28.29 ratio · epochs, a
143
value similar to that obtained from control cultures (29.62 ratio · epochs). Thus, our calcium imaging data suggested not only that GluR2-lacking AMPA receptors mediate calcium influx following in vitro injury, but also that their expression protracts calcium extrusion.
4.4 Interfering with GluR2 endocytosis is cytoprotective in vitro Tat-QSAV and Naspm reduce excitotoxicity There is clear evidence for the involvement of elevated cytosolic calcium in mediating neuronal death during excitotoxicity. However, we had no direct evidence at this point that suggested the expression of calcium-permeable AMPA receptors were necessarily involved in the cytotoxicity of AMPA in this preparation. To test this, it was necessary to examine the cytoprotective efficacy of both calcium-permeable AMPA receptor antagonism as well as interfering with GluR2 trafficking. We repeated the previous toxicity assays of stretch + NMDA followed one hour later by a 30 μM AMPA challenge. Post-injury treatments included 20 μM Tat-QSAV, 20 μM Tat-AAAA, and 100 μM Naspm. Stretch + NMDA again resulted in a marked susceptibility to secondary AMPA toxicity (23.3 ± 5.9% increase in cell death, n = 3 cultures, Figure 6B). However, Tat-QSAV applied with stretch + NMDA afforded significant cytoprotection against AMPA excitotoxicity 20 hours after injury [9.58 ± 2.9% increase in cell death, n = 4 cultures, p < 0.05, Figure 6B]. Naspm also demonstrated a trend towards cytoprotection against cell death conferred by AMPA [1.78 ± 5.7% increase in cell death, n = 3 cultures, p = 0.055, Figure 6B].
144
Figure 17. Stretch + NMDA promotes calcium influx through calcium-permeable AMPARs. A) Temporal schematic of calcium imaging experiments B) Fura PE3 data over the entire recording period. Baseline ratios of neurons exposed to Stretch + NMDA are significantly higher than those of control neurons. As well, after perfusion of 100 μM AMPA and 50 μM CTZ the duration of excess cytosolic Ca2+ is prolonged. Selective antagonism of GluR2-lacking AMPARs (100 μM Naspm) lowers peak AMPA-induced Ca2+ and mitigates the prolonged elevation in intracellular calcium. C) Quantification of baseline Fura ratios (340/380 nm). D) Quantification of Δt of peak calcium levels to return to baseline E) Quantification of peak ratio normalized to baseline. 100 μM Naspm reduces peak calcium. F) Integration of the calcium curves shown in (B) reveals a 2.11fold increase in the area under Stretch + NMDA curve relative to control neurons. There are no error bars in this graph because these are the integrals of the mean calcium curves. G) Representative Fura-PE3 micrographs of baseline (left column) and peak (right column) emission in control neurons (top row), Stretch + NMDA (middle row) and Stretch + NMDA + 100 μM Naspm (bottom row). Scale bars = 40 μm.
145
Figure 17. Stretch + NMDA promotes calcium influx through calcium-permeable AMPARs.
146
Tat-AAAA demonstrated no attenuation of AMPA-induced cell death (29.2 ± 3.9% increase in cell death, n = 3 cultures, Figure 6B). Importantly, there was no significant difference in cell death between groups at 1 hour after the insult. These results suggest that a portion of the delayed (i.e., secondary) cell death that occurs in this model of trauma could be prevented through preservation of surface GluR2 or antagonizing GluR2-lacking AMPARs.
4.5 Hippocampal CA1 is hyperexcitable following fluid percussion trauma: Excitability is lowered with TAT-QSAV application To ascertain a measure of AMPA receptor phenotype in the injured whole animal, we performed CA1 field recordings following Schaffer collateral stimulation, a wellestablished glutamatergic synapse which we demonstrated to be an almost entirely AMPA receptor mediated response after complete rundown following 6-cyano-7nitroquinoxaline-2,3-dione (CNQX) application (Figure 19B). We first measured the gross amplitude of the CA1 AMPA-receptor mediated evoked population spike over a range of 12 stimulation amplitudes (ranging from 10-120 μA). Statistically, we ran a two-way repeated measure ANOVA, with independent variables of stimulation amplitude and treatment (i.e., Ctrl, FPI, and FPI + 3 mg/kg Tat-QSAV I.V). Significant differences were detected among our treatment groups (P = 0.002). To follow up the two-way ANOVA and parse out where the differences lay, one way ANOVA followed by StudentNeuman Keuls tests were performed to identify differences between groups at each individual stimulation amplitude. Over a range of 10-80 μA, we observed that fluid percussion trauma markedly increased the CA1 evoked response, (P < 0.05, Figure 19C).
147
Figure 18. Inhibiting GluR2 endocytosis is neuroprotective. A) Top row: representative propidium iodide fluorimetry 20 hours after exposure of cortical neurons to Stretch + NMDA + AMPA, with or without the presence of polyamines or inhibitory peptides. AMPA was applied for 1 hour, with or without peptide/polyamine treatment, at 1 hour following Stretch + NMDA. Bottom row: brightfield images of the corresponding field represented in top row. B) Quantification of normalized PI fluorimetry by plate scanning at 1 hour and 20 hours after AMPA treatment. Scale bar 200 μm. # p < 0.05 vs control; * p < 0.05 vs injured.
148
Figure 18. Inhibiting GluR2 endocytosis is neuroprotective
149
However, when FPI animals were treated intravenously with 3 mg/kg Tat-QSAV, the evoked response was significantly lower across all stimulation amplitudes, suggesting potentially the involvement of PICK1-dependent processes in mediating this synaptic potentiation. At higher stimulation amplitudes (80-120 μA), a marked depression of the population spike was maintained in Tat-QSAV treated animals (P < 0.05 for all stimulation amplitudes), despite a lack of a significant difference between control and injured animals over these treatment points (P > 0.05 for all). Notably, we were able to achieve a similar potentiation of the CA1 population spike with exogenous PKC activation, which was performed via application of 1 μM phorbol 12-myristate 13-acetate (phorbol ester, PMA, Figure 19D). Collectively, these results suggest a partial enhancement of CA1 population spike amplitude by FPI that could be attenuated by interfering with PICK1-dependent protein interactions or mimicked by activation of protein kinase C, two proteins which play a key role in the removal of surface GluR2 protein. Given that GluR2-lacking AMPA receptors have a higher single channel conductance per receptor complex, and that Tat-QSAV reduced CA1 evoked responses, we next hypothesized that perhaps AMPA receptors devoid of the GluR2 subunit were contributing to the elevation of CA1 population spike amplitude.
4.6 Hippocampal CA1 Naspm sensitivity increases after FPI Occlusion of CA1 Naspm sensitivity is achieved through interference with GluR2 endocytosis. Having observed CA1 hyperexcitability that was attenuated by PICK1 inhibition, we next investigated whether a traumatically injured hippocampus demonstrated an
150
Figure 19. Post-injury CA1 hyperexcitability is attenuated by Tat-QSAV treatment. A) Schematic of recording procedure in sagittal hippocampal slices. Recording electrodes were placed in the stratum pyramidale of area CA1, while stimulation occurred at the axons of the schaffer collateral tracts originating in area CA3. DG = dentate gyrus. B) CA1 population spike amplitude is nearly completely abolished during perfusion of the slice with 20μM CNQX, an indication that this synapse is an appropriate measure of AMPA receptor-mediated evoked responses. C) CA1 excitability is markedly increased 3-6 hours following fluid percussion injury, an effect that is attenuated by intravenous treatment of animals with 3 mg/kg Tat-QSAV. The effect is particularly noticeable at lower stimulation amplitudes (10-80 μA). D) Potentiation of the CA1 evoked response can be achieved via perfusion with phorbol esters (PMA), exogenous activators of PKC, which stimulate the endocytosis of GluR2.
151
Figure 19. Post-injury CA1 hyperexcitability is attenuated by Tat-QSAV treatment
152
increased expression of calcium permeable AMPARs, as these receptors have a higher single channel conductance, and as has been discussed, contribute to progressive excitotoxic cell death and dysfunction251,252,276,279,280,287,411. We found during recordings from FPI rats that CA1 population spikes exhibited a Naspm-induced rundown to 58.9 ± 1.7% of baseline, a significantly greater inhibition than sham animals (78.9 ± 0.79%, p < 0.05, Figure 20A). This increased sensitivity of CA1 physiology to antagonists of calcium-permeable AMPA receptors suggests that these receptors contribute more significantly to synaptic transmission in injured animals relative to controls. However, injecting animals intravenously with Tat-QSAV (3 mg/kg) following the traumatic injury occluded Naspm-induced rundown of CA1 population spike amplitude (88.2 ± 5.6 %, Figure 20B), providing evidence that GluR2 trafficking is integral in the expression of calcium-permeable AMPARs. Notably, Naspm sensitivity was preserved in Tat-AAAA injected animals (59.3 ± 8.3% of baseline (p < 0.01 vs sham and QSAV, p > 0.05 vs FPI, Figure 20B). As a final positive control, we also treated animals (3 mg/kg) with a GluR2 c-terminal PICK1 binding peptide that has been used throughout the literature to interfere with AMPA receptor trafficking, Tat-SVKI. This peptide similarly mimics a PICK1 PDZ binding motif by replicating the GluR2 c-terminal PDZ ligand. Naspm-induced population spike rundown was occluded (97.2 ± 14.1% of baseline, p < 0.05 vs FPI) in Tat-SVKI treated animals in a similar fashion to those animals treated with Tat-QSAV (Figure 20B), providing further evidence that PICK1-mediated GluR2 endocytosis was involved in the post-traumatic expression of calcium-permeable AMPARs. Collectively, our biochemical and electrophysiological data suggested that calcium-permeable AMPA receptors were expressed via GluR2 endocytosis following whole animal trauma.
153
Figure 20. CA1 hippocampal physiology is sensitive to antagonists of calciumpermeable AMPA receptors after TBI. A) Naspm-induced rundown of CA1 population spike amplitude was significantly greater in injured animals, supporting the in vitro findings that these molecular modifications lead to incorporation of phenotypically different channels. Representative traces illustrating rundown of population spike amplitude during the recording period appear above the graph. B) Prevention of GluR2 endocytosis with Tat-QSAV or Tat-SVKI, both PICK1 binding peptides, significantly reduces CA1 naspm sensitivity. Tat-AAAA was ineffective in occluding naspm sensitivity. * p < 0.05. ** p < 0.05.
154
Figure 20. CA1 hippocampal physiology is sensitive to antagonists of calciumpermeable AMPA receptors after TBI
155
4.7 Occluding GluR2 endocytosis reduces apoptotic cell death: Post-traumatic DNA fragmentation is reduced by interfering with GluR2 trafficking Biochemically, our data suggested that the GluR2 subunit was internalized following fluid percussion trauma. Electrophysiologically, we identified a contribution for GluR2-lacking receptors to hippocampal physiology. However, our hypothesis was that the expression of these receptors played a significant role in the susceptibility of neurons to secondary injury following brain trauma. Thus, it was necessary to ultimately examine the cytoprotective efficacy of Tat-QSAV, and thereby delineate whether the aberrant trafficking of GluR2 has any cytotoxic implications. We performed TUNEL staining of coronal brain slices at 24 hours following fluid percussion injury, quantifying the prevalence of DNA fragmentation with a sampling of approximately 3000 cortical cells per animal. TUNEL staining was accompanied by a nuclear counter-stain for Hoescht 33342, allowing us to quantify data as the percentage of cells identified as TUNEL positive (thereby normalizing the data to cell density). Quantification of slices was performed completely blind. Following fluid percussion trauma, 6.08 ± 1.49% of cells were identified as TUNEL positive (n = 6). However, there was a marked reduction of TUNEL positive cells in slices obtained from animals treated intravenously with Tat-QSAV (1.47 ± 0.6% TUNEL positive, n = 6 animals, P < 0.05), as well as a reduction of chromatin condensation (Figure 21D, arrowheads), suggesting that interference with PICK1-dependent protein interactions reduces cortical DNA fragmentation following experimental TBI (quantification in Figure 21E). One problem with the use of TUNEL staining is the tendency of the method to identify DNA strand breaks that occur from cytotoxic processes other than apoptosis (e.g., fragmentation produced by reactive oxygen species such as peroxynitrite). To
156
confirm that the TUNEL staining was corroborated by other assays of apoptotic cell death, we performed co-precipitation experiments with cytochrome c and apoptotic peptidase activating factor 1 (APAF-1). The cyt-c-APAF-1 complex is recognized as an important initiator of apoptotic cell death, which binds and cleaves procaspase-9, releasing the mature and activated form of the cysteine protease. In turn, caspase-9 cleaves and activates the effector caspases 3 and 7, which carry out the execution phase of programmed cell death. As evidenced in Figure 21F, TUNEL staining was accompanied by an observable interaction between cytochrome-c and APAF-1, an interaction that was only present in injured tissue and was confirmed by both positive and negative controls (Figure 21F). Thus, this biochemical data provided further evidence that apoptotic cell death indeed was occurring following FPI.
4.8 Summary These data provide evidence for an increased contribution of GluR2-lacking AMPA receptors to neuronal physiology following TBI. We observed in both our whole cell patch clamp and CA1 field electrophysiological assays that the sensitivity of the AMPA receptor response to a selective antagonist of GluR2-lacking receptors (Naspm) was significantly increased. We also observed that AMPAR-mediated calcium influx was augmented following trauma, and could be similarly attenuated by Naspm application. Further evidence for the involvement of higher-conductance GluR2-lacking receptors in neuronal signaling post-TBI comes from our observation that AMPAR mEPSC amplitudes are increased, as is the basal excitability of hippocampal CA1, an
157
electrophysiological response which we showed to be almost entirely AMPA receptor mediated. To ascertain the potential involvement of GluR2 trafficking in the subsequent expression of GluR2-lacking receptors, we treated both neurons and cultures with our peptide inhibitor of the PICK1-PKC protein interaction. We found not only that this peptide inhibitor occluded the expression of calcium-permeable receptors and dampened AMPA receptor-mediated electrophysiological responses, but also that the compound, when administered post-injury, provided cytoprotection against apoptotic cell death in vivo. Collectively, our results provide evidence for a cascade of GluR2 endocytosis which promotes the cytotoxic expression of calcium-permeable AMPA receptors.
158
Figure 21. Perturbing GluR2 endocytosis affords cytoprotection from apoptosis at 24 hours following fluid percussion trauma. A) Sham tissue is non-reactive for TUNEL staining, a marker of endonuclease-mediated DNA overhang initiated during programmed cell death. Top panel, Hoescht 33342 nuclear stain, a non-specific marker of cellular nuclei. B) Top row: TUNEL staining is markedly increased following fluid percussion trauma. Bottom row: intravenously administered Tat-QSAV reduces the prevalence of TUNEL positive cells. C) Contralateral tissue is non-reactive for TUNEL staining in both treatment conditions. D) 40 x magnification of cortical cells revealing co-localization of TUNEL positive neurons with condensed chromatin (arrowheads), two hallmarks of the terminal stages of apoptosis. Nuclei of Tat-QSAV treated animals are less condensed, and do not co-localize with TUNEL staining to the same extent. E) Quantification of TUNEL positive neurons normalized to the total number of cells in the sampling area. F) Co-precipitation experiments with pull-down of cytochrome c and blotting for bound APAF-1 reveal a cyt-c-APAF1 complex only in injured tissue. Far left lane: positive control of APAF-1 (whole cell lysate). Negative controls and sham tissue do not demonstrate binding between cytochrome c and APAF-1. Cyt-c IgG appears only in lanes where co-precipitation was performed.
159
Figure 21. Perturbing GluR2 endocytosis affords cytoprotection from apoptosis at 24 hours following fluid percussion trauma.
160
Chapter 5: Discussion, Limitations and Future Directions
161
5.1 Preface The specific aims of this thesis were directed at investigating the hypothesis that a reduction of surface GluR2 protein contributes to neuronal vulnerability to secondary injury following TBI, by increasing the population of calcium-permeable AMPA receptors. In two experimental models, in vitro and in vivo TBI, we described molecular and phenotypic alterations to AMPA receptor trafficking and physiology that had profound effects on neuronal viability. At the cellular level, we found that the endocytosis of surface GluR2 protein after trauma contributes to the expression of GluR2-lacking AMPARs, and the susceptibility of neurons to excitotoxicity (see figure 22). Accordingly, our data employing the use of Tat peptides intended to disrupt GluR2 trafficking suggest that GluR2 internalization is an aberrant event occurring in traumatized neurons that contributes to delayed neuronal death and calcium overload.
5.2 Corroborating studies Consistent with the observations presented in this work, several other studies have proposed that a reduction of surface GluR2 contributes to secondary injury and neuronal death after CNS insult. Firstly, ischemic incorporation of GluR2-lacking AMPARs and association of GluR2 with PICK1 was reported in cultured hippocampal neurons279. In that study, internalization of GluR2 was associated with a similar polyamine-sensitive increase in mEPSC amplitude. Over the course of our investigation (i.e., simultaneous to our study), an independent lab investigated the mechanism of GluR2 internalization in this ischemic model and found an identical cascade to what is reported in this thesis. Indeed it was shown in mid 2009 that activation of NMDA receptors following ischemia 162
leads to a PICK1-dependent switch in AMPA receptor subunit composition from GluR2containing to GluR2-lacking. Moreover, the investigation showed that peptides that interfere with the GluR2 c-terminal PDZ interaction with PICK1 occlude the expression of GluR2-lacking AMPARs and provide cytoprotection in hippocampal neurons exposed to OGD375. This finding further supports our hypothesis that NMDA receptor activation following TBI might lead to an identical reduction of surface GluR2 via the PDZ interactions responsible for subunit trafficking. Concomitant to the undertaking of our study, other experimental paradigms of CNS injury reported aberrant GluR2 trafficking in conditions involving neuronal hyperexcitability and calcium overload. GluR2 S880 phosphorylation, GluR1 S845 phosphorylation (discussed next in Future directions), and enhanced AMPAR mEPSCs were reported during neonatal epilepsy, a condition also marked by neuronal hyperexcitability276. Indeed that study suggested that the simultaneous removal of GluR2 coupled with the delivery of GluR1 was capable of remodeling the AMPAergic synapse to become profoundly more calcium-permeable. Neuronal hyperexcitability in the spinal dorsal horn also contributes to the pathophysiology of chronic pain, another condition that has shown to involve aberrant NMDAR and PKC-dependent GluR2 trafficking. In a study that employed an animal model of peripheral inflammation, it was demonstrated that nociceptive hypersensitivity induces synaptic GluR2 internalization in dorsal horn neurons, an effect mediated by serine 880 phosphorylation, and activation of PKC downstream of NMDA receptors.
163
Figure 22. Schematic demonstrating proposed signaling involved in post-traumatic internalization of GluR2 and subsequent expression of GluR2-lacking AMPARs. A) After TBI, intracellular calcium coming through the NMDA receptor activates PKCα via its association with PSD-95 in the NMDAR complex. Activated PKCα binds PICK1, and is trafficked to the membrane where it phosphorylates GluR2 at serine 880. GluR2 associates with PICK1 and is internalized from the cell surface, enhancing the expression of GluR2-lacking AMPARs. B) Proposed mechanism of cytoprotection. Antagonism of GluR2-lacking AMPA receptors with Naspm, or occluding the binding of PKCα with PICK1 and/or PSD-95 via Tat-QSAV reduces GluR2 internalization and expression of calcium-permeable AMPARs after TBI.
164
Figure 22. Schematic demonstrating proposed signaling involved in post-traumatic internalization of GluR2 and subsequent expression of GluR2-lacking AMPARs.
165
Perhaps most compelling however was the observed switch from GluR2containing AMPARs to GluR2-lacking AMPARs reported after a more severe mechanical stretch injury - characterized by marked inward-rectification of the AMPA receptor current-voltage relationship - 252, in a study also demonstrating neuroprotective effects of Naspm antagonism. Collectively, this work has built a growing body of evidence suggesting that the loss of surface GluR2 protein is an important contributing factor to neuronal dysfunction and cell death in excitotoxic CNS disease. In order to prevent the loss of surface GluR2, the intracellular mechanisms responsible for its aberrant endocytosis need to be mapped out, a problem which this study begins to address. The finding that prevention of GluR2 endocytosis reduces secondary injury after TBI is supported by many investigations that have recapitulated the result that injured neurons are dramatically more susceptible to glutamatergic stimulation than healthy cells. A number of cell culture models using mechanical injury devices have shown increased excitotoxin lethality following trauma119,120,251,412. However, the more compelling evidence comes from whole animal models. One such study used the fluid percussion device followed by microdialysis of glutamate to investigate a possible co-operation between trauma and excitatory amino acids in mediating neuronal damage after TBI. This study was undertaken on the basis of the observations that cerebral glutamate levels measured in patients by microdialysis (16-350 μM) are sufficient to kill neurons in culture, but not in the intact brain of the normal rat413. The authors therefore sought to identify a synergistic effect between excitatory amino acid–mediated damage and other
166
posttrauma mechanisms. Following central FPI, the authors reported that glutamate perfusion produced a lesion significantly larger than both FPI + mock CSF perfusion, and glutamate perfusion alone. Furthermore, the lesion volume of the FPI + glutamate group exceeded the summed mean volumes from the FPI + mock CSF, and glutamate alone groups. This highlights a clear susceptibility of injured tissue to glutamate receptor stimulation. Coupled with the observations reported by us and others that CA1 hippocampal glutamatergic transmission is significantly augmented following trauma, this data supports the theory that trauma induces a post-synaptic modification of glutamate receptor functioning, which might include the type of AMPA receptor remodeling reported in this thesis.
5.3 Co-operation of Stretch + NMDA In our in vitro model, we observed that 10 μM NMDA did not result in the internalization or phosphorylation of GluR2 unless it was combined with stretch injury. Notably, we also observed that while stretch + NMDA on its own was not immediately cytotoxic, it imposed a marked vulnerability to secondary AMPA insult. There are a number of possibilities to explain the cooperative effects of stretch injury and NMDA on both AMPA receptor trafficking and neuronal susceptibility to secondary injury. Mechanical trauma reduces the magnesium block of the NMDA receptor243 potentially allowing a previously innocuous dose of NMDA to initiate substantially more calcium influx in injured neurons vs controls. Indeed NMDA is markedly more lethal to stretched neurons than uninjured cultures119 and initiates larger calcium transients245 after sublethal stretch. These findings help in the understanding of how the two insults might cooperate
167
in calcium-dependent PKC-activation and GluR2 phosphorylation. Also, mechanical trauma elevates intracellular superoxide levels in cortical neurons119,120. Superoxide plays an important role in PKC activation via thiol oxidation414, including the regulation of kinase activity during LTP415,416 when PKC is active in the post-synaptic density and plays a role in GluR2 removal268. It is possible that oxidative modification causes preferential binding of PKC to various substrates, and it would be worthwhile to investigate the hypothesis that superoxide is responsible for the post-injury PKCa-PSD95 association. In this scenario, PKCa – structurally connected to PSD-95 and embedded in the NMDAR protein complex after stretch – would be primed for activation from subsequent NMDAR stimulation.
5.4 Limitations of the current study 5.4.1 Non-specific Tat peptide interactions It is important to recognize the possibility that occupying the PDZ-domains of PICK1 and/or PSD-95 with Tat-QSAV might be cytoprotective in a more non-specific fashion than inhibiting PKCα binding. The PDZ-domain of PICK1 interacts with at least 45 other known PDZ-ligands. Occupying this domain with a -QSAV peptide could conceivably interfere with other PICK1 protein interactions. Further, there might conceivably exist other intracellular PDZ targets of the –QSAV sequence present on our peptide. Thus, while we demonstrate the successful perturbation of the PICK1-PKCa and PSD-95-PKCa association, we cannot definitively exclude the possibility that the cytoprotective effect of the compound is mediated elsewhere. Future work will include knocking down the expression of PKCα and/or PICK1 and investigating if the cytoprotective effects of the peptide are occluded.
168
Prior to the knockdown experiments however, proteomics can be used to identify the PDZ domains with which Tat-QSAV interacts. Studies have been carried out which have employed cloning of the ~ 470 human PDZ proteins in the SMART database, followed by fusion of the proteins to GST. By coating plates with anti-GST antibody, these cloned PDZ proteins can be immobilized and probed with potential binding partners. Incubation of individual wells with labeled Tat-QSAV would be a highthroughput method of identifying the interacting partners of our peptide. Moreover, immobilization of GST-PICK1 followed by probing with purified PKCα in the presence of varying concentrations (e.g., 0.001-100 μM) of our tat peptide inhibitor could be done to identify the IC50 of the peptide. This would provide valuable information on the intracellular concentration of the peptide necessary to achieve sufficient inhibition of the protein-protein interaction.
5.4.2 Non-specific effects of Tat peptide transduction In our whole cell patch clamp electrophysiological assays, we observed a significant non-specific effect of tat peptide transduction on miniature AMPA receptormediated EPSCs. Indeed incubation of our injured cultures with Tat-AAAA, a nonfunctional negative control peptide significantly reduced the amplitude of injured events. Though our active PDZ-ligand (QSAV) induced a further and significant reduction from our inactive control (providing a role for GluR2 trafficking in the increased event amplitude), the mechanism by which our Tat-AAAA peptide decreased AMPA-mediated mEPSCs is at present unknown. The simplest explanation is that the peptide may have
169
non-specifically blocked AMPA channels. However, based on current insight into tat peptide transduction, there are other likely explanations. Tat peptides, as discussed in the introduction, enter cells through lipid-raft dependent fluid phase macropinocytosis. The interaction of the cationic Tat PTD with lipid rafts – enriched in cholesterol and anionic sphingolipids – is an electrostatic interaction followed by endocytosis of the raft along with the extracellular peptide cargoes417. The link between this mechanism and its effect on the functioning of AMPA receptors can be made via the studies describing the effects of lipid raft depletion on AMPA receptor surface expression and electrophysiology. It is now known that AMPA receptors are associated in detergent-resistant membranes in dendritic spines with the cholesterol and sphingolipids present on lipid rafts, and that raft depletion reduces the density of AMPA receptors found on dendritic spines418,419. It is therefore conceivable that the macropinocytosis following Tat peptide transduction might be accompanied by a loss of AMPA receptor surface expression, translating unsurprisingly to a decrease in the amplitude of AMPA receptor mediated events. To parse this out, an appropriate further experiment would include bath application of tat-peptides to our cortical cultures, followed by immunocytochemical GluR1 N-terminal surface labeling to examine the density of AMPA receptors following Tat-mediated protein transduction.
5.4.3 Co-precipitation: What does it mean? Interestingly, in the present study, the reduction of surface GluR2 and subsequent AMPA receptor potentiation was NMDA receptor dependent. The simplest explanation is that the NMDAR dependence arises because of the structural link between PKCα and
170
PSD-95. The -QSAV sequence on PKCα’s extreme c-terminus is a type I PDZ ligand with the potential to form a stable interaction with two of PSD-95’s PDZ domains333. However, our co-precipitation data does not rule out the possibility that PKCα is indirectly associated with PSD-95, via a binding partner that is able to bind both the kinase’s PDZ-ligand and one of PSD-95’s PDZ domains. Indeed this problem is the hallmark shortcoming of using co-immunoprecipitation as an assay method. While co-immunoprecipitation can demonstrate that two proteins are found in the same cellular complex, the assay does not prove that the two proteins are physically touching one another, that is, directly associated. Since all of our coimmunoprecipitation experiments were performed using cell lysates, it is possible that two co-immunoprecipitating proteins in our experiments were linked together by a third protein that acts as a scaffold. This possibility is more likely in certain scenarios than others. The PKC-PICK1 as well as GluR2-PICK1 protein interactions have been extensively defined through yeast-two hybrid screening and direct recombinant protein pull-down assays. However, an appropriate further experiment to fully validate the interaction between PSD-95 and PKCα that we report would involve purification of both proteins, followed by dot blotting. Indeed immobilization of purified GST-labeled PSD95 onto a nitrocellulose membrane, followed by incubation with myc-tagged purified PKCα would definitively identify that the proteins are capable of a direct interaction. It was our intention to perform this experiment, however, a lack of expertise in protein cloning (and therefore an inability to secure purified PSD-95 protein) in our laboratory prevented us from doing so. Nonetheless, this assay is necessary to verify that the kinase is a direct PDZ binding partner of PSD-95.
171
5.4.4 TNFα-induced AMPA receptor trafficking: An alternative mechanism of calcium-permeable AMPA receptor expression An alternative mechanism for the neuronal phenotype observed in this work (i.e., an increased expression of calcium-permeable AMPA receptors) involves a proinflammatory cytokine that is central to the inflammatory response that occurs after cerebral trauma. Tumor necrosis factor alpha (TNFα), released from neighbouring
glial cells during CNS inflammation, was recently shown to increase AMPA receptor surface expression ex vivo, and specifically, to increase the expression of calcium-permeable AMPA receptors420. Indeed brain slices incubated with TNFα exhibit a marked increase in naspm sensitivity during whole-cell AMPAR patch clamp, and also exhibit marked elevations in AMPAR-derived free calcium421,422. The evidence that this process might contribute to excitotoxicity following CNS trauma is compelling. For example, co-injection of TNFα with low doses of kainic acid produces marked neuronal death in vivo that far exceeds injection of the glutamate agonist alone423. Secondly, it was recently shown during spinal cord injury that TNFα induces the surface trafficking of GluR2-lacking AMPA receptors, thereby imparting a vulnerability to secondary excitotoxic injury mediated at AMPAergic synapses. Interestingly, in this paradigm, a soluble TNFα receptor mitigated these effects, providing convincing evidence that this cytokine can remodel the glutamatergic synapse during CNS injury to include calciumpermeable AMPA receptors377.
172
In our experimental paradigm, there is likely to be a significant increase in parenchymal TNFα levels. It has been shown that fluid percussion trauma induces marked elevations in TNFα levels424-427, and it is also known that injured neurons will release TNFα themselves428, something which may have occurred during our in vitro stretch injury protocol. Thus it is possible that this mechanism contributed to the increased naspm sensitivity that we observed after trauma. However, our observations that peptide-mediated PICK1 inhibition attenuated the expression of calcium-permeable receptors suggest that if TNFα-mediated AMPA receptor trafficking occurs after experimental TBI, it is likely in parallel to the effects that we observed.
5.5 Future Directions The data presented in this thesis raise some compelling questions that remain to be answered. For example, we have yet to identify the intracellular events that follow the internalization of GluR2 protein (e.g., protein degradation or recycling). We have also not investigated the trafficking of GluR1, another AMPA receptor subunit whose surface delivery, rather than internalization, might mimic the phenotype we observed in many of our experiments. Moreover, it is unknown at present what impact the application of GluR2 endocytotic inhibitory peptides has on physiologic synaptic plasticity, given that the PICK1-dependent expression of GluR2-lacking AMPARs is a mechanism critical to the development of LTP. In the interest of pursuing these questions, we have collected preliminary data that begins to address these issues.
173
5.5.1 Total GluR2 levels are reduced by 24 hours following trauma One intriguing difference between our work and other work that has studied the influence of GluR2 expression on cell survival during CNS injury is that our work was focused on local protein trafficking, as opposed to global regulation of protein transcription. Indeed the work studying REST-dependent epigenetic silencing of GluR2 expression in cerebral ischemia identified that by silencing RNA transcription, total GluR2 protein levels were reduced by 24-48 hours following global ischemia. Whether or not our acute experiments highlighting protein endocytosis (performed within hours of the injury) translated to a down-regulation of total protein at a more delayed time point is largely unknown. Accordingly, we investigated total GluR2 expression at 24 hours following FPI in a small sample of four animals. Notably, we observed that total GluR2 expression was down-regulated by 24 hours after FPI (67.3 ± 11.1% of control in ipsilateral cortex, 54.4 ± 7.2% in contralateral cortex, n = 4, P < 0.01, figure 23), a finding that has been reported by other investigators in both experimental TBI and spinal cord injury429-431, particularly in apoptotic neurons101. It remains to be seen if the early endocytosis that we report is the mechanism responsible for the delayed down-regulation of GluR2, although there is evidence that GluR2 endocytosis leads to lysosomal degradation, and subsequently, the expression of calcium permeable AMPA receptors. Indeed one particularly relevant investigation reported that the endosomal protein NEEP21 associates with the PDZ scaffolding molecule GRIP1 and GluR2; and that when the NEEP21-GRIP interaction is lost, GluR2 surface expression decreases, causing GluR2 accumulation in early endosomes and lysosomes, and inward rectification of AMPAR EPSCs (a property of GluR2-lacking
174
AMPARs)432. Thus, it is conceivable that PICK1 targets internalized GluR2 to acid hydrolase-filled lysosomes, where a reduction of total protein would occur. Appropriate experiments to parse out whether post-traumatic GluR2 endocytosis leads to protein loss might include a) examination of total GluR2 levels with an without Tat-QSAV, to examine if interfering with GluR2 trafficking influences total protein expression at 24 hours, and b) to co-precipitate GluR2 with known late endosomal or lysosomal proteins (e.g., RAB7 and RAB9 and mannose 6-phosphate receptors) after TBI, to examine if this is the mechanism of protein loss.
5.5.2 GluR1 trafficking may increase following trauma: NO-mediated GluR1 serine 845 phosphorylation occurs following traumatic injury An alternative mechanism of an increased population of GluR2-lacking AMPA receptors might be through increased exocytotic delivery of GluR1, thereby allowing for the incorporation of GluR1 homomeric channels. One way through which GluR1 delivery occurs is through nitric oxide-mediated phosphorylation of a critical GluR1 serine reside (845), that allows for binding of the subunit with cyclic GMP-dependent kinase II and delivery to the plasma membrane433. Indeed, mild mechanical trauma coupled with NMDA receptor activation produces high levels of nitric oxide (NO) through the NR2bPSD-95-nNOS cascade in cortical neurons119. Accordingly, we performed western blots for GluR1 phosphorylation at S845. Following stretch + NMDA, phosphorylated GluR1 increased to 302 ± 47.6% of control levels (p < 0.05, Figure 24A). Notably, cells treated with an NR2b antagonist did not exhibit a significant increase in phosphorylated GluR1 relative to control cultures (p = 0.15, Figure 23A), suggesting the mechanism of this phosphorylation was NR2b-dependent, likely because of the structural scaffold between
175
Figure 23. Total GluR2 protein levels are reduced at 24 hours following FPI. A) Western blot of total GluR2 protein in ipsilateral and contralateral cortex. ERK 1,2 was used as a loading control. B) Quantification of total GluR2 protein levels normalized to sham animals.
176
Figure 23. Total GluR2 protein levels are reduced at 24 hours following FPI.
177
NR2B and nNOS activation. We also probed simultaneously for total GluR1 levels, of which there was no significant difference between treatments (p = 0.71, Figure 23A). We also identified another nitric oxide dependent modification of the AMPA receptor through whole-cell electrophysiology. We observed that injured cortical neurons displayed markedly increased activity (qualitatively) relative to control neurons when sodium-free extracellular solution was perfused (Figure 24B). Given that the dominant cation in these experiments was choline, the most likely-explanation for this activity would be calcium-mediated currents. Notably, when we inhibited nitric oxide synthase activity with L-NG-Nitroarginine methyl ester (L-Name), we were unable to replicate the sodium-free firing of the neurons (Figure 24B). This preliminary western blotting and electrophysiological data suggest that perhaps nitric oxide dependent delivery of GluR1 accompanies GluR2 trafficking in the expression of calcium-permeable AMPA receptors.
5.5.3 Tat-QSAV treatment does not occlude induced synaptic plasticity: Hippocampal LTP is preserved with PICK1 inhibition Our initial hypotheses were crafted based on the physiological role of PICK1 in remodeling the AMPAergic synapse during synaptic plasticity. Indeed a major cellular mechanism underlying activity-dependent plasticity of glutamatergic transmission is the regulated trafficking of AMPARs, particularly the trafficking of GluR2. As discussed in the introduction, PICK1-mediated control of GluR2 surface levels is a key mechanism in the induction of LTP in hippocampal CA1. The removal of GluR2, and therefore the incorporation of GluR2-lacking, higher conductance channels, is thought to underlie a lasting increase in synaptic efficacy during the induction of learning and memory. Thus
178
Figure 24. Stretch + NMDA increases NO-dependent GluR1 S845 phosphorylation. A) GluR1 S845 increases after Stretch + NMDA and is mitigated by NR2b antagonism (quantification on right). B) Removal of extracellular sodium abolishes AMPARmediated mEPSCs in cortical neurons held at -70 mV. C) AMPAR mEPSCs persist in the absence of extracellular sodium in cortical neurons 1 hour following Stretch + NMDA. D) Addition of 100 μM L-NAME to inhibit nNOS and GluR1 phosphorylation largely attenuates the sodium-free firing in injured cortical neurons.
179
Figure 24. Stretch + NMDA increases GluR1 S845 phosphorylation
180
we had a final interest in examining the impact of Tat-QSAV on the induction of LTP in the hippocampus, to investigate the physiological significance of inhibiting GluR2 trafficking. We further sought to identify the impact of FPI, an injury known to induce GluR2 internalization, on the induction of LTP. Given that we observed GluR2 internalization and associated hyperexcitability in the hippocampus, we conjectured that perhaps LTP in this area would be occluded. Indeed it is known that LTP is impaired in the hippocampus after FPI, and we thought that an increase in basal excitability due to remodeling of the AMPAergic response might underlie this impairment. However, contrary to this hypothesis, we observed that there was no impairment in LTP induction in injured animals. In slices from control animals (n = 4), population spike amplitude was maintained at 155.3 ± 13.2% of baseline (30th epoch used for analysis). Following FPI (n = 7, 3-6 hours after the injury), population spike amplitude increased to 151.1 ± 13.1% of baseline, an insignificant difference from uninjured animals (figure 25, P > 0.05). Similarly, intravenous injection of Tat-QSAV (3 mg/kg, n = 4) was without effect on hippocampal population spike LTP, with baseline levels increasing to 153.1 ± 18.2%. Thus, these results suggest one of two possibilities: 1) that GluR2 trafficking is not involved in the induction of LTP in hippocampal CA1, or 2) that following injury, mechanisms outside of AMPA receptor trafficking are responsible for LTP induction.
5.5.4 – Does inhibition of the PICK1 PDZ domain represent a future antiexcitotoxic therapy? The concept that PICK1-mediated protein interactions might underlie neurological dysfunction in a number of disorders is beginning to gain considerable
181
Figure 25. Hippocampal LTP is unaffected following FPI, and uninfluenced by TatQSAV treatment. 100 Hz theta burst stimulation of the schaffer collateral tract was applied to induce LTP of the population spike in stratum pyramidale of the CA1 cell layer. Top panel: LTP of the population spike amplitude is successfully induced in both control and injured animals. Bottom panel: Treatment of injured animals with Tat-QSAV is without effect on CA1 LTP. Slices were stimulated 3-6 hours following FPI.
182
Figure 25. Hippocampal LTP is unaffected following FPI, and uninfluenced by TatQSAV treatment.
183
attention in neurobiology literature. In general, PICK1 serves as an attractive molecular target because of its nature as a PDZ domain, which has been identified as a putative drug target across a variety of different diseases. For example, blocking the PDZ interaction between the NMDA receptor and PSD-95 with membrane-permeable peptides results in selective inhibition of neuronal nitric oxide synthase (nNOS) activation and a dramatic reduction of ischemic injury following experimental stroke334. In cancer, recent evidence suggests that blocking the PDZ domains of Na+/H+ exchanger regulatory factor 1 (NHERF-1), dishevelled, or AF-6 might have tumor suppressing potential434. Finally, our data here is the first to show that inhibition of the PICK1 PDZ domain reduces cell death following CNS injury involving excitotoxicity. Inhibition of the PICK1 PDZ domain has evolved from a conceptual idea to a reality in recent months based largely on the emerging data reporting that PICK1medaited protein interactions contribute to neuropathic pain, excitotoxicity, and drug addiction272,435. At the forefront of this effort is an investigation that screened approximately 44,000 compounds as small-molecule inhibitors of the PICK1 PDZ domain436. Remarkably, a non-peptide small molecular inhibitor of the PICK1 PDZ domain was identified (known as FSC231) which has an affinity for the domain similar to that of the endogenous peptide ligand (Ki ~10 μM). Physiologically, FRET and coimmunoprecipitation experiments demonstrated that FSC231 crossed the plasma membrane and inhibited the interaction between GluR2 and PICK1 in cultured neurons. Moreover, FSC231 interfered with GluR2 trafficking, which is consistent with inhibiting PICK1’s involvement in GluR2 endocytosis. Finally, FSC231 inhibited both LTD and LTP expression in CA1 hippocampal neurons, consistent with inhibition of PICK1’s
184
bidirectional effect on synaptic plasticity. The work in this thesis has laid the foundation for the evaluation of FSC231 as a putative therapeutic against secondary cell death after TBI.
5.6 Significance of Findings The data presented in this thesis contribute to a novel understanding of the neuronal mechanisms responsible for excitotoxic cell death following traumatic injury to the brain. Much of the current literature on glutamate receptor-mediated cellular injury following trauma highlights increased extracellular glutamate as the initiating event in excitotoxicity; however, our data introduces the possibility that trauma-induced postsynaptic receptor modification can impart lethality upon otherwise innocuous glutamate levels. As previously discussed, these data are corroborated by a number of investigations describing similar trafficking of the GluR2 subunit under pathological conditions, and provide a plausible mechanism responsible for the previous observations detailing the cytotoxicity of physiological glutamate in traumatized neurons. In addition to presenting a different conceptual understanding of excitotoxicity, these data also provide a mechanism through which these changes can occur. Our peptide-mediated interventional approach has highlighted PICK1 as the major contributor to post-traumatic GluR2 endocytosis, and our data has further elucidated a potential NMDA receptor-dependent mechanism through which GluR2 internalization is initiated. Hopefully, this will translate in the future to a more targeted therapeutic approach to excitotoxicity that circumvents the shortcomings and potential non-specific effects of global glutamate receptor antagonism after TBI. The development of FSC231 as a small molecule inhibitor of the PICK1 PDZ domain is an example of such efforts, and as stated
185
by the authors of that study, was inspired in part by the data presented in this thesis and its associated publications.
5.7 Conclusions
1) Neuronal trauma confers the endocytosis of the AMPA receptor GluR2 subunit, evidenced by subunit phosphorylation, internalization, and a physical association with its major trafficking proteins at early time-points after traumatic injury in vivo and in vitro.
2) The trafficking of GluR2 protein increases the expression of calcium-permeable AMPA receptors, evidenced through whole cell and field electrophysiology and imaging of cellular calcium dynamics. Importantly, perturbation of GluR2 endocytosis reduces the expression of calcium-permeable AMPA receptors.
3) Interruption of GluR2 trafficking confers cytoprotection against excitotoxic injury in two experimental paradigms -- in vitro stretch injury, where PICK1-binding peptides protected against AMPA toxicity, and fluid percussion trauma, where these same peptides reduced apoptotic cell death 24 hours after trauma.
186
Reflective Appendix This brief appendix was written in the days following the final oral defense of the work, and reflects some of the more compelling and clinically relevant issues that were raised during that meeting.
Relevance of GluR2 endocytosis to white matter injury: Over the course of the discussion, it was postulated that the mechanisms described in this thesis might also contribute to white matter injury, a prominent pathophysiological feature underlying severe functional impairment post TBI. This connection was made based on certain work demonstrating a marked susceptibility of oligodendrocyte cell cultures to AMPA receptor-mediated neuronal injury. Indeed if a post-traumatic modification of the GluR2 content occurred in oligos in a fashion similar to what was seen in our neuronal population, this might lead to eventual oligo cell injury, axonal demylenation, and a withdrawn trophic support for neurons in our whole animal preparation. Notably, McDonald et al., (1998) showed in their Nature Medicine paper that rodent oligodendrocytes are highly susceptible to AMPAR-mediated excitotoxicity, both in culture and following stereotaxic injection of AMPAR agonists. This is a compelling finding demonstrating the sensitivity of a non-neuronal cell type to AMPAmediated cell death with immediate relevance to axonal viability. However, a follow up study examining the clinical applicability of this phenomenon described the salient observation that rodent and human oligodendrocytes differ vastly in their levels of AMPA receptor expression, leading to a cautionary interpretation of McDonald et al.’s findings. In 2004, Wosik et al demonstrated that in
187
fact human oligos express low levels of ionotropic glutamate receptors in vitro and are resistant to high and sustained doses of AMPA/Kainate, even in the presence of reduced receptor desensitization, which would exacerbate stimulation of the receptor. In the same investigation, the group performed identical experiments in rodent oligos, demonstrating a marked vulnerability to identical doses of the AMPAR activators. This observation raises important questions about the clinical relevance of the pre-clinical work supporting AMPAR-mediated cell death in oligos as a critical mechanism of white matter injury in CNS disease.
Model descriptors: Why the stretch injury is described as an impulse: Another relevant concept mentioned during the final defense was an inquiry into our decision to describe the stretch model in terms of the impulse (J) inflicted on the neurons. We chose to offer an approximation of the impulse experienced by the neurons to help in the replication of the model by others. Because impulse is equal to FΔt, and most models have strict control over the duration of injury (ie. it is relatively easy to keep this variable constant from lab to lab because most systems have control over the duration of valve opening), the largest variable in this equation would be the force, which we thought easier to standardize than other variables, such as pressure (P) strain (e) or the stretch ratio (λ). Force, being equal to pressure·area, can be calculated by labs lacking the necessary equipment to measure strain/stretch. For example, our lab lacks a high-speed camera necessary to make accurate strain measurements experienced by the cultures (this would require measurement of axon length before and during the stretch injury). Thus we have found it difficult to recapitulate the models which describe neurons as having
188
undergone a certain level of strain (eg., 130% of initial axon length). Though this appears to be the standard in the literature, we imagine that other labs have similar difficulties in designing a system to replicate a specific level of strain. In the thesis and its publications, we also described the pressure exerted on the cells for all experiments (i.e., between 2.5 and 2.9 psi). However, in the literature this can be confusing. The initial papers characterizing the stretch injury model described a dose response of pressure-cell death that ranged over pressures from 10-70 psi (Ellis et al., 1995). However, these were the regulator pressure readings, not the pressure measured by the transducer following rebound of the silastic membrane. Also, because not all labs have tissue culture wells of the same size (ie., 35 mm as in our study), the force exerted on the neurons will change if the pressure is the only variable that is standardized. Indeed we have found this to be the case with our fluid percussion injury device. By standardizing force (and time) labs can adjust the pressure exerted on the cells given a certain well size, such that the force (pressure x area) is constant (and therefore impulse is too, if time is also a constant). It has been reported that an internal chamber pressure during stretch injury from between 5-7 psi correlates to a strain on axons of 0.58-0.77, or 58-77% beyond its initial length (Smith et al, JNeurosci, 1999). Given the linear relationship between pressure and tissue deformation described in this model (Ellis et al, 1995), we can approximate that our tissue strain would measure close to half of the lower end of this approximation, or approximately 30%. Notably, this level of strain for a mild injury model is highly comparable to other labs measurements of strain during mild injury (Arundine et al., 2004, Lau et al., 2006, JNeurosci). The next best approximation we can make with
189
respect to comparing our injury with other labs because of this shortcoming in measuring tissue strain is with biological/biochemical outcome. Our model is sub-lethal, and does not appear to alter membrane integrity. These are endpoints that we used to compare to other models in the literature and establish the differences between mild/moderate/severe. We believe that impulse offers some approximation of the in vivo situation. As mentioned in the thesis on page 24, the forces that result in this tensile elongation during TBI are thought to occur in 50 ms, the duration of stretch applied in our model (and accounted for in our calculation of impulse). The situation becomes more complicated when looking at the force applied to the neurons and how this compares to the intact brain. One shortcoming of the model is that it is two dimensional, and thus difficult to approximate how this type of stretching corresponds to a three dimensional environment found in the intact brain. In terms of biological outcome, we believe our model is analogous to an in vivo mild trauma based on the following: 1) there is no cell death after the injury, 2) there are no observable changes to membrane permeability, 3) there is no evidence of axotomy, and 4) there is no accumulation of cytoskeletal swellings. Indeed in a moderate injury, these are salient pathophysiological features of more severe axonal injury models and accordingly, more severe in vivo models and clinical TBI. Further we believe that what does occur in this model is analogous to a mild injury in vivo, as we observed GluR2 phosphorylation in both a mild FPI and our stretch model, as well as electrophysiological hyperexcitability.
190
Can we be sure that NR2B containing NMDARs are extrasynaptic? In this thesis, we found that GluR2 internalization was an NR2B-dependent phenomenon. There is a plethora of literature suggesting that NR2B is primarily an extrasynaptic protein. Indeed early papers described a reduction of ifenprodil sensitivity (an indication of NR2B subunit expression) of synaptic activity as neurons developed in culture, promoting the hypothesis that the proportion of synaptic NR2B was lessened in favor of NR2A expression, and that NR2B containing receptors are extrasynaptic. However, experiments in the last 5 years have shown that in the presence of MK-801, an antagonist of synaptic NMDARs, the eletrophysiological response of NMDARs is not abolished completely by ifenprodil, suggesting the presence of NR2A subunits at extrasynaptic sites (see for example Thomas et al., 2006). We did not perform any experiments examining the distribution of these subunits, and thus it cannot be definitively concluded that the NR2B-containing receptors mediating GluR2 phosphorylation are exclusively extrasynaptic. Future experiments will include parsing this out in greater detail. How else might we examine the impact of non-specific Tat peptide transduction on AMPAR mEPSCs? One of the findings that were slightly problematic in this thesis was that TatAAAA dampened AMPAR-mediated events in a non-specific fashion. We thought it pertinent to address this point in slightly greater detail, and to discuss experimental approaches that might help us further understand this observation. One way to identify if the peptide was non-specifically blocking channels or if AMPARs were removed from the surface following peptide transduction would be to mutate the Tat sequence to a non-
191
functional moiety that contains an equal number of positive charges. In this scenario, one would need to confirm that the peptide was not being taken up into cells by similarly tagging it with dansyl chloride. If the neurons stained negatively for the dansyl but exhibited blockade of AMPAR mediated currents, this would suggest that the nonspecific antagonistic action of the peptide (and therefore any peptide) was extracellular (ie fitting into the channel pore, blocking the binding site, allosteric modulation etc). However, if the currents were unaffected, this would suggest the peptide needs to be intracellular to exert its effects and would support a role for endocytosis in the antagonism of AMPAR mEPSCs. One might also stimulate macropinocytosis of other molecules (e.g., eosinophil cationic proteins) to see if there is a similar depression of AMPAR-mediated events. If so, one could suggest that endocytosis of any cargo perturbs the surface expression of AMPARs. One might also apply the tat peptide in the presence of inhibitors of endocytosis (cyclodextran or chlorpromazine) to ensure that this process is critical to the blockade.
Final Thoughts: One last consideration in this thesis is that of the relevance of GluR2 endocytosis in mediating secondary neuronal cell injury after TBI in the context of all other sequelae present in traumatized brain tissue. In TBI research, it is frequently difficult to contribute more than a minor “piece of the puzzle” to the vast array of knowledge surrounding mechanisms of cell death after trauma. Thus any one observation can seem diminished, as it will likely only address a minor contributor to cell injury. Whether the effects that
192
we observed are a dominant contributor to post-TBI neuronal injury or dysfunction is unknown, particularly in the clinical context. However, the data supporting a cytoprotective role for PICK1 inhibition combined with the corroborating studies that were undertaken at the same time as ours, suggest that this is unlikely to be an epiphenomenon without relevance to neuronal survival. Thus, we think further pursuit of this mechanism is warranted.
193
194
References Cited 1.
Guerrero, JL, Thurman, DJ and Sniezek, JE, (2000) Emergency department visits associated with traumatic brain injury: United States, 1995-1996. Brain Inj 14: 181-6.
2.
Thurman, DJ, Branche, CM and Sniezek, JE, (1998) The epidemiology of sportsrelated traumatic brain injuries in the United States: recent developments. J Head Trauma Rehabil 13: 1-8.
3.
Sosin, DM, Sniezek, JE and Thurman, DJ, (1996) Incidence of mild and moderate brain injury in the United States, 1991. Brain Inj 10: 47-54.
4.
Waxweiler, RJ, Thurman, D, Sniezek, J, Sosin, D and O'Neil, J, (1995) Monitoring the impact of traumatic brain injury: a review and update. J Neurotrauma 12: 509-16.
5.
Thurman, DJ, Alverson, C, Dunn, KA, Guerrero, J and Sniezek, JE, (1999) Traumatic brain injury in the United States: A public health perspective. J Head Trauma Rehabil 14: 602-15.
6.
Sauaia, A, Moore, FA, Moore, EE, Moser, KS, Brennan, R, Read, RA et al., (1995) Epidemiology of trauma deaths: a reassessment. J Trauma 38: 185-93.
7.
Steyerberg, EW, Mushkudiani, N, Perel, P, Butcher, I, Lu, J, McHugh, GS et al., (2008) Predicting outcome after traumatic brain injury: development and international validation of prognostic scores based on admission characteristics. PLoS Med 5: e165; discussion e165.
8.
Mock, C, Quansah, R, Krishnan, R, Arreola-Risa, C and Rivara, F, (2004) Strengthening the prevention and care of injuries worldwide. Lancet 363: 2172-9.
9.
Finfer, SR and Cohen, J, (2001) Severe traumatic brain injury. Resuscitation 48: 77-90.
10.
Murray, CJ and Lopez, AD, (1997) Alternative projections of mortality and disability by cause 1990-2020: Global Burden of Disease Study. Lancet 349: 1498-504.
11.
Langlois JA, R-BW, Thomas KE, (2006) Traumatic Brain Injury in the United States: Emergency Department Visits, Hospitalizations, and Deaths. Atlanta
195
(GA): Centers for Disease Control and Prevention, National Center for Injury Prevention and Control. pp. 1-70. 12.
Rutland-Brown, W, Langlois, JA, Thomas, KE and Xi, YL, (2006) Incidence of traumatic brain injury in the United States, 2003. J Head Trauma Rehabil 21: 5448.
13.
Bruns, J, Jr. and Hauser, WA, (2003) The epidemiology of traumatic brain injury: a review. Epilepsia 44 Suppl 10: 2-10.
14.
Langlois, JA and Sattin, RW, (2005) Traumatic brain injury in the United States: research and programs of the Centers for Disease Control and Prevention (CDC). J Head Trauma Rehabil 20: 187-8.
15.
Tagliaferri, F, Compagnone, C, Korsic, M, Servadei, F and Kraus, J, (2006) A systematic review of brain injury epidemiology in Europe. Acta Neurochir (Wien) 148: 255-68; discussion 268.
16.
Harris, OA, Bruce, CA, Reid, M, Cheeks, R, Easley, K, Surles, MC et al., (2008) Examination of the management of traumatic brain injury in the developing and developed world: focus on resource utilization, protocols, and practices that alter outcome. J Neurosurg 109: 433-8.
17.
Adekoya N, TD, White D, Webb K., (2002) Surveillance for traumatic brain injury deaths--United States, 1989-1998. . MMWR. Surveillance summaries : Morbidity and mortality weekly report. Surveillance summaries 51: 1-14.
18.
Marshall, LF, (2000) Epidemiology and cost of central nervous system injury. Clin Neurosurg 46: 105-12.
19.
Max, W, Rice, DP and MacKenzie, EJ, (1990) The lifetime cost of injury. Inquiry 27: 332-43.
20.
Max W, ME, Rice D., (1991) Head Injury: Cost and Consequences. Journal of Head Trauma Rehabilitation 6: 76-91.
21.
Luchter, S and Walz, MC, (1995) Long-term consequences of head injury. J Neurotrauma 12: 517-26.
22.
Maas, AI, Stocchetti, N and Bullock, R, (2008) Moderate and severe traumatic brain injury in adults. Lancet Neurol 7: 728-41.
196
23.
Morton, MV and Wehman, P, (1995) Psychosocial and emotional sequelae of individuals with traumatic brain injury: a literature review and recommendations. Brain Inj 9: 81-92.
24.
Klonoff, PS, Snow, WG and Costa, LD, (1986) Quality of life in patients 2 to 4 years after closed head injury. Neurosurgery 19: 735-43.
25.
Klonoff, PS, Costa, LD and Snow, WG, (1986) Predictors and indicators of quality of life in patients with closed-head injury. J Clin Exp Neuropsychol 8: 469-85.
26.
Levin, HS, Grossman, RG, Rose, JE and Teasdale, G, (1979) Long-term neuropsychological outcome of closed head injury. J Neurosurg 50: 412-22.
27.
Levin, HS, Gary, HE, Jr., Eisenberg, HM, Ruff, RM, Barth, JT, Kreutzer, J et al., (1990) Neurobehavioral outcome 1 year after severe head injury. Experience of the Traumatic Coma Data Bank. J Neurosurg 73: 699-709.
28.
Levin, HS, (1995) Prediction of recovery from traumatic brain injury. J Neurotrauma 12: 913-22.
29.
Whyte, J, Cifu, D, Dikmen, S and Temkin, N, (2001) Prediction of functional outcomes after traumatic brain injury: a comparison of 2 measures of duration of unconsciousness. Arch Phys Med Rehabil 82: 1355-9.
30.
Brown, AW, Malec, JF, McClelland, RL, Diehl, NN, Englander, J and Cifu, DX, (2005) Clinical elements that predict outcome after traumatic brain injury: a prospective multicenter recursive partitioning (decision-tree) analysis. J Neurotrauma 22: 1040-51.
31.
Dikmen, S, Machamer, J, Temkin, N and McLean, A, (1990) Neuropsychological recovery in patients with moderate to severe head injury: 2 year follow-up. J Clin Exp Neuropsychol 12: 507-19.
32.
Malec, JF, Brown, AW, Leibson, CL, Flaada, JT, Mandrekar, JN, Diehl, NN et al., (2007) The mayo classification system for traumatic brain injury severity. J Neurotrauma 24: 1417-24.
33.
Cassidy, JD, Carroll, LJ, Peloso, PM, Borg, J, von Holst, H, Holm, L et al., (2004) Incidence, risk factors and prevention of mild traumatic brain injury: results of the WHO Collaborating Centre Task Force on Mild Traumatic Brain Injury. J Rehabil Med: 28-60.
197
34.
Thornhill, S, Teasdale, GM, Murray, GD, McEwen, J, Roy, CW and Penny, KI, (2000) Disability in young people and adults one year after head injury: prospective cohort study. Bmj 320: 1631-5.
35.
Helmy, A, Vizcaychipi, M and Gupta, AK, (2007) Traumatic brain injury: intensive care management. Br J Anaesth 99: 32-42.
36.
Taber, KH, Warden, DL and Hurley, RA, (2006) Blast-related traumatic brain injury: what is known? J Neuropsychiatry Clin Neurosci 18: 141-5.
37.
Armonda, RA, Bell, RS, Vo, AH, Ling, G, DeGraba, TJ, Crandall, B et al., (2006) Wartime traumatic cerebral vasospasm: recent review of combat casualties. Neurosurgery 59: 1215-25; discussion 1225.
38.
Denny-Brown D, aR, W.R., (1941) Experimental cerebral concussion. Brain 64: 92-164.
39.
Sullivan, HG, Martinez, J, Becker, DP, Miller, JD, Griffith, R and Wist, AO, (1976) Fluid-percussion model of mechanical brain injury in the cat. J Neurosurg 45: 521-34.
40.
Thibault, LE, Meaney, DF, Anderson, BJ and Marmarou, A, (1992) Biomechanical aspects of a fluid percussion model of brain injury. J Neurotrauma 9: 311-22.
41.
Zhang, L, Yang, KH and King, AI, (2004) A proposed injury threshold for mild traumatic brain injury. J Biomech Eng 126: 226-36.
42.
Ward CC, CM, Nahum AM., (1980) Intracranial pressure - A brain injury criterion. Proceedings, 24th Stapp Car Crash Conf.,: SAE Paper No. 801304.
43.
Skoglund, TS and Nellgard, B, (2005) Long-time outcome after transient transtentorial herniation in patients with traumatic brain injury. Acta Anaesthesiol Scand 49: 337-40.
44.
Smith, DH, Meaney, DF and Shull, WH, (2003) Diffuse axonal injury in head trauma. J Head Trauma Rehabil 18: 307-16.
45.
Metz, H, McElhaney, J and Ommaya, AK, (1970) A comparison of the elasticity of live, dead, and fixed brain tissue. J Biomech 3: 453-8.
46.
Meaney, DF, Margulies, SS and Smith, DH, (2001) Diffuse axonal injury. J Neurosurg 95: 1108-10. 198
47.
Noppens, RR, Christ, M, Brambrink, AM, Koerner, IP, Heimann, A and Kempski, O, (2006) An early bolus of hypertonic saline hydroxyethyl starch improves long-term outcome after global cerebral ischemia. Crit Care Med 34: 2194-200.
48.
Ziai, WC, Toung, TJ and Bhardwaj, A, (2007) Hypertonic saline: first-line therapy for cerebral edema? J Neurol Sci 261: 157-66.
49.
McGraw, CP, (1989) A cerebral perfusion pressure greater than 80 mmHg is more beneficial. Hoff, J. T., Betz, A.L. ed. (Springer-Verlag, Berlin).
50.
Rosner M.J., R, S.D., Johnson, A.H., (1985) Cerebral perfusion: management protocol and clinical results. J.Neurosurgery 83: 949-962.
51.
Stein, SC, Chen, XH, Sinson, GP and Smith, DH, (2002) Intravascular coagulation: a major secondary insult in nonfatal traumatic brain injury. J Neurosurg 97: 1373-7.
52.
Park, E, Bell, JD, Siddiq, IP and Baker, AJ, (2009) An analysis of regional microvascular loss and recovery following two grades of fluid percussion trauma: a role for hypoxia-inducible factors in traumatic brain injury. J Cereb Blood Flow Metab 29: 575-84.
53.
Harhangi, BS, Kompanje, EJ, Leebeek, FW and Maas, AI, (2008) Coagulation disorders after traumatic brain injury. Acta Neurochir (Wien) 150: 165-75; discussion 175.
54.
Ansari, MA, Roberts, KN and Scheff, SW, (2008) Oxidative stress and modification of synaptic proteins in hippocampus after traumatic brain injury. Free Radic Biol Med 45: 443-52.
55.
Deng-Bryant, Y, Singh, IN, Carrico, KM and Hall, ED, (2008) Neuroprotective effects of tempol, a catalytic scavenger of peroxynitrite-derived free radicals, in a mouse traumatic brain injury model. J Cereb Blood Flow Metab 28: 1114-26.
56.
Opii, WO, Nukala, VN, Sultana, R, Pandya, JD, Day, KM, Merchant, ML et al., (2007) Proteomic identification of oxidized mitochondrial proteins following experimental traumatic brain injury. J Neurotrauma 24: 772-89.
57.
Singh, IN, Sullivan, PG and Hall, ED, (2007) Peroxynitrite-mediated oxidative damage to brain mitochondria: Protective effects of peroxynitrite scavengers. J Neurosci Res 85: 2216-23.
199
58.
Singh, IN, Sullivan, PG, Deng, Y, Mbye, LH and Hall, ED, (2006) Time course of post-traumatic mitochondrial oxidative damage and dysfunction in a mouse model of focal traumatic brain injury: implications for neuroprotective therapy. J Cereb Blood Flow Metab 26: 1407-18.
59.
Tavazzi, B, Signoretti, S, Lazzarino, G, Amorini, AM, Delfini, R, Cimatti, M et al., (2005) Cerebral oxidative stress and depression of energy metabolism correlate with severity of diffuse brain injury in rats. Neurosurgery 56: 582-9; discussion 582-9.
60.
Bayir, H, Kagan, VE, Borisenko, GG, Tyurina, YY, Janesko, KL, Vagni, VA et al., (2005) Enhanced oxidative stress in iNOS-deficient mice after traumatic brain injury: support for a neuroprotective role of iNOS. J Cereb Blood Flow Metab 25: 673-84.
61.
Mendez, DR, Cherian, L, Moore, N, Arora, T, Liu, PK and Robertson, CS, (2004) Oxidative DNA lesions in a rodent model of traumatic brain injury. J Trauma 56: 1235-40.
62.
Chang, EF, Wong, RJ, Vreman, HJ, Igarashi, T, Galo, E, Sharp, FR et al., (2003) Heme oxygenase-2 protects against lipid peroxidation-mediated cell loss and impaired motor recovery after traumatic brain injury. J Neurosci 23: 3689-96.
63.
Uryu, K, Laurer, H, McIntosh, T, Pratico, D, Martinez, D, Leight, S et al., (2002) Repetitive mild brain trauma accelerates Abeta deposition, lipid peroxidation, and cognitive impairment in a transgenic mouse model of Alzheimer amyloidosis. J Neurosci 22: 446-54.
64.
Lewen, A, Fujimura, M, Sugawara, T, Matz, P, Copin, JC and Chan, PH, (2001) Oxidative stress-dependent release of mitochondrial cytochrome c after traumatic brain injury. J Cereb Blood Flow Metab 21: 914-20.
65.
Awasthi, D, Church, DF, Torbati, D, Carey, ME and Pryor, WA, (1997) Oxidative stress following traumatic brain injury in rats. Surg Neurol 47: 575-81; discussion 581-2.
66.
Laird, MD, Vender, JR and Dhandapani, KM, (2008) Opposing roles for reactive astrocytes following traumatic brain injury. Neurosignals 16: 154-64.
67.
Floyd, CL and Lyeth, BG, (2007) Astroglia: important mediators of traumatic brain injury. Prog Brain Res 161: 61-79.
200
68.
Yi, JH and Hazell, AS, (2006) Excitotoxic mechanisms and the role of astrocytic glutamate transporters in traumatic brain injury. Neurochem Int 48: 394-403.
69.
Pedersen, MO, Larsen, A, Stoltenberg, M and Penkowa, M, (2009) Cell death in the injured brain: roles of metallothioneins. Prog Histochem Cytochem 44: 1-27.
70.
Harting, MT, Jimenez, F, Adams, SD, Mercer, DW and Cox, CS, Jr., (2008) Acute, regional inflammatory response after traumatic brain injury: Implications for cellular therapy. Surgery 144: 803-13.
71.
Israelsson, C, Bengtsson, H, Kylberg, A, Kullander, K, Lewen, A, Hillered, L et al., (2008) Distinct cellular patterns of upregulated chemokine expression supporting a prominent inflammatory role in traumatic brain injury. J Neurotrauma 25: 959-74.
72.
Lloyd, E, Somera-Molina, K, Van Eldik, LJ, Watterson, DM and Wainwright, MS, (2008) Suppression of acute proinflammatory cytokine and chemokine upregulation by post-injury administration of a novel small molecule improves long-term neurologic outcome in a mouse model of traumatic brain injury. J Neuroinflammation 5: 28.
73.
Goodman, JC, Van, M, Gopinath, SP and Robertson, CS, (2008) Proinflammatory and pro-apoptotic elements of the neuroinflammatory response are activated in traumatic brain injury. Acta Neurochir Suppl 102: 437-9.
74.
Morganti-Kossmann, MC, Satgunaseelan, L, Bye, N and Kossmann, T, (2007) Modulation of immune response by head injury. Injury 38: 1392-400.
75.
Hutchinson, PJ, O'Connell, MT, Rothwell, NJ, Hopkins, SJ, Nortje, J, Carpenter, KL et al., (2007) Inflammation in human brain injury: intracerebral concentrations of IL-1alpha, IL-1beta, and their endogenous inhibitor IL-1ra. J Neurotrauma 24: 1545-57.
76.
Chen, G, Shi, JX, Hang, CH, Xie, W, Liu, J and Liu, X, (2007) Inhibitory effect on cerebral inflammatory agents that accompany traumatic brain injury in a rat model: a potential neuroprotective mechanism of recombinant human erythropoietin (rhEPO). Neurosci Lett 425: 177-82.
77.
Lenzlinger, PM, Morganti-Kossmann, MC, Laurer, HL and McIntosh, TK, (2001) The duality of the inflammatory response to traumatic brain injury. Mol Neurobiol 24: 169-81.
201
78.
Morganti-Kossmann, MC, Rancan, M, Otto, VI, Stahel, PF and Kossmann, T, (2001) Role of cerebral inflammation after traumatic brain injury: a revisited concept. Shock 16: 165-77.
79.
Arand, M, Melzner, H, Kinzl, L, Bruckner, UB and Gebhard, F, (2001) Early inflammatory mediator response following isolated traumatic brain injury and other major trauma in humans. Langenbecks Arch Surg 386: 241-8.
80.
Morganti-Kossman, MC, Lenzlinger, PM, Hans, V, Stahel, P, Csuka, E, Ammann, E et al., (1997) Production of cytokines following brain injury: beneficial and deleterious for the damaged tissue. Mol Psychiatry 2: 133-6.
81.
Zhang, Z, Ottens, AK, Sadasivan, S, Kobeissy, FH, Fang, T, Hayes, RL et al., (2007) Calpain-mediated collapsin response mediator protein-1, -2, and -4 proteolysis after neurotoxic and traumatic brain injury. J Neurotrauma 24: 460-72.
82.
Park, E, Liu, E, Shek, M, Park, A and Baker, AJ, (2007) Heavy neurofilament accumulation and alpha-spectrin degradation accompany cerebellar white matter functional deficits following forebrain fluid percussion injury. Exp Neurol 204: 49-57.
83.
Ai, J, Liu, E, Park, E and Baker, AJ, (2007) Structural and functional alterations of cerebellum following fluid percussion injury in rats. Exp Brain Res 177: 95112.
84.
Buki, A and Povlishock, JT, (2006) All roads lead to disconnection?--Traumatic axonal injury revisited. Acta Neurochir (Wien) 148: 181-93; discussion 193-4.
85.
Chung, RS, Staal, JA, McCormack, GH, Dickson, TC, Cozens, MA, Chuckowree, JA et al., (2005) Mild axonal stretch injury in vitro induces a progressive series of neurofilament alterations ultimately leading to delayed axotomy. J Neurotrauma 22: 1081-91.
86.
Stone, JR, Okonkwo, DO, Dialo, AO, Rubin, DG, Mutlu, LK, Povlishock, JT et al., (2004) Impaired axonal transport and altered axolemmal permeability occur in distinct populations of damaged axons following traumatic brain injury. Exp Neurol 190: 59-69.
87.
Buki, A, Farkas, O, Doczi, T and Povlishock, JT, (2003) Preinjury administration of the calpain inhibitor MDL-28170 attenuates traumatically induced axonal injury. J Neurotrauma 20: 261-8.
202
88.
Baker, AJ, Phan, N, Moulton, RJ, Fehlings, MG, Yucel, Y, Zhao, M et al., (2002) Attenuation of the electrophysiological function of the corpus callosum after fluid percussion injury in the rat. J Neurotrauma 19: 587-99.
89.
Buki, A, Okonkwo, DO, Wang, KK and Povlishock, JT, (2000) Cytochrome c release and caspase activation in traumatic axonal injury. J Neurosci 20: 2825-34.
90.
Buki, A, Siman, R, Trojanowski, JQ and Povlishock, JT, (1999) The role of calpain-mediated spectrin proteolysis in traumatically induced axonal injury. J Neuropathol Exp Neurol 58: 365-75.
91.
Bramlett, HM, Kraydieh, S, Green, EJ and Dietrich, WD, (1997) Temporal and regional patterns of axonal damage following traumatic brain injury: a betaamyloid precursor protein immunocytochemical study in rats. J Neuropathol Exp Neurol 56: 1132-41.
92.
Posmantur, RM, Kampfl, A, Liu, SJ, Heck, K, Taft, WC, Clifton, GL et al., (1996) Cytoskeletal derangements of cortical neuronal processes three hours after traumatic brain injury in rats: an immunofluorescence study. J Neuropathol Exp Neurol 55: 68-80.
93.
Pettus, EH, Christman, CW, Giebel, ML and Povlishock, JT, (1994) Traumatically induced altered membrane permeability: its relationship to traumatically induced reactive axonal change. J Neurotrauma 11: 507-22.
94.
Christman, CW, Grady, MS, Walker, SA, Holloway, KL and Povlishock, JT, (1994) Ultrastructural studies of diffuse axonal injury in humans. J Neurotrauma 11: 173-86.
95.
Zhang, X, Graham, SH, Kochanek, PM, Marion, DW, Nathaniel, PD, Watkins, SC et al., (2003) Caspase-8 expression and proteolysis in human brain after severe head injury. Faseb J 17: 1367-9.
96.
Sullivan, PG, Keller, JN, Bussen, WL and Scheff, SW, (2002) Cytochrome c release and caspase activation after traumatic brain injury. Brain Res 949: 88-96.
97.
Harter, L, Keel, M, Hentze, H, Leist, M and Ertel, W, (2001) Caspase-3 activity is present in cerebrospinal fluid from patients with traumatic brain injury. J Neuroimmunol 121: 76-8.
98.
Keane, RW, Kraydieh, S, Lotocki, G, Alonso, OF, Aldana, P and Dietrich, WD, (2001) Apoptotic and antiapoptotic mechanisms after traumatic brain injury. J Cereb Blood Flow Metab 21: 1189-98.
203
99.
Raghupathi, R, Graham, DI and McIntosh, TK, (2000) Apoptosis after traumatic brain injury. J Neurotrauma 17: 927-38.
100.
Clark, RS, Kochanek, PM, Adelson, PD, Bell, MJ, Carcillo, JA, Chen, M et al., (2000) Increases in bcl-2 protein in cerebrospinal fluid and evidence for programmed cell death in infants and children after severe traumatic brain injury. J Pediatr 137: 197-204.
101.
O'Dell, DM, Raghupathi, R, Crino, PB, Eberwine, JH and McIntosh, TK, (2000) Traumatic brain injury alters the molecular fingerprint of TUNEL-positive cortical neurons In vivo: A single-cell analysis. J Neurosci 20: 4821-8.
102.
Pike, BR, Zhao, X, Newcomb, JK, Glenn, CC, Anderson, DK and Hayes, RL, (2000) Stretch injury causes calpain and caspase-3 activation and necrotic and apoptotic cell death in septo-hippocampal cell cultures. J Neurotrauma 17: 28398.
103.
Beer, R, Franz, G, Schopf, M, Reindl, M, Zelger, B, Schmutzhard, E et al., (2000) Expression of Fas and Fas ligand after experimental traumatic brain injury in the rat. J Cereb Blood Flow Metab 20: 669-77.
104.
Newcomb, JK, Zhao, X, Pike, BR and Hayes, RL, (1999) Temporal profile of apoptotic-like changes in neurons and astrocytes following controlled cortical impact injury in the rat. Exp Neurol 158: 76-88.
105.
Clark, RS, Chen, J, Watkins, SC, Kochanek, PM, Chen, M, Stetler, RA et al., (1997) Apoptosis-suppressor gene bcl-2 expression after traumatic brain injury in rats. J Neurosci 17: 9172-82.
106.
Colicos, MA and Dash, PK, (1996) Apoptotic morphology of dentate gyrus granule cells following experimental cortical impact injury in rats: possible role in spatial memory deficits. Brain Res 739: 120-31.
107.
Bell, JD, Park, E, Ai, J and Baker, AJ, (2009) PICK1-mediated GluR2 endocytosis contributes to cellular injury after neuronal trauma. Cell Death Differ.
108.
Bigford, GE, Alonso, OF, Dietrich, D and Keane, RW, (2009) A novel protein complex in membrane rafts linking the NR2B glutamate receptor and autophagy is disrupted following traumatic brain injury. J Neurotrauma 26: 703-20.
109.
Beschorner, R, Dietz, K, Schauer, N, Mittelbronn, M, Schluesener, HJ, Trautmann, K et al., (2007) Expression of EAAT1 reflects a possible
204
neuroprotective function of reactive astrocytes and activated microglia following human traumatic brain injury. Histol Histopathol 22: 515-26. 110.
Yi, JH, Pow, DV and Hazell, AS, (2005) Early loss of the glutamate transporter splice-variant GLT-1v in rat cerebral cortex following lateral fluid-percussion injury. Glia 49: 121-33.
111.
Wagner, AK, Bayir, H, Ren, D, Puccio, A, Zafonte, RD and Kochanek, PM, (2004) Relationships between cerebrospinal fluid markers of excitotoxicity, ischemia, and oxidative damage after severe TBI: the impact of gender, age, and hypothermia. J Neurotrauma 21: 125-36.
112.
Ruppel, RA, Kochanek, PM, Adelson, PD, Rose, ME, Wisniewski, SR, Bell, MJ et al., (2001) Excitatory amino acid concentrations in ventricular cerebrospinal fluid after severe traumatic brain injury in infants and children: the role of child abuse. J Pediatr 138: 18-25.
113.
Brown, JI, Baker, AJ, Konasiewicz, SJ and Moulton, RJ, (1998) Clinical significance of CSF glutamate concentrations following severe traumatic brain injury in humans. J Neurotrauma 15: 253-63.
114.
Palmer, AM, Marion, DW, Botscheller, ML, Swedlow, PE, Styren, SD and DeKosky, ST, (1993) Traumatic brain injury-induced excitotoxicity assessed in a controlled cortical impact model. J Neurochem 61: 2015-24.
115.
Arundine, M and Tymianski, M, (2004) Molecular mechanisms of glutamatedependent neurodegeneration in ischemia and traumatic brain injury. Cell Mol Life Sci 61: 657-68.
116.
Aarts, MM and Tymianski, M, (2004) Molecular mechanisms underlying specificity of excitotoxic signaling in neurons. Curr Mol Med 4: 137-47.
117.
Arundine, M and Tymianski, M, (2003) Molecular mechanisms of calciumdependent neurodegeneration in excitotoxicity. Cell Calcium 34: 325-37.
118.
Sattler, R and Tymianski, M, (2001) Molecular mechanisms of glutamate receptor-mediated excitotoxic neuronal cell death. Mol Neurobiol 24: 107-29.
119.
Arundine, M, Aarts, M, Lau, A and Tymianski, M, (2004) Vulnerability of central neurons to secondary insults after in vitro mechanical stretch. J Neurosci 24: 8106-23.
205
120.
Arundine, M, Chopra, GK, Wrong, A, Lei, S, Aarts, MM, MacDonald, JF et al., (2003) Enhanced vulnerability to NMDA toxicity in sublethal traumatic neuronal injury in vitro. J Neurotrauma 20: 1377-95.
121.
Curtis, DR and Watkins, JC, (1960) The excitation and depression of spinal neurones by structurally related amino acids. J Neurochem 6: 117-41.
122.
Curtis, DR, Phillis, JW and Watkins, JC, (1960) The chemical excitation of spinal neurones by certain acidic amino acids. J Physiol 150: 656-82.
123.
Hayashi, T, (1952) A physiological study of epileptic seizures following cortical stimulation in animals and its application to human clinics. Jpn J Physiol 3: 46-64.
124.
Purves, D, Augustine, G.J., Fitzpatrick, D., Hall, W.C., LaMantia, A.S., McNamara, J.O., Williams, S.M. , (2004) Neuroscience: Chapter 6; Neurotransmitters and their receptors. (Sinauer Associates., Sunderland, MA).
125.
Mark, LP, Prost, RW, Ulmer, JL, Smith, MM, Daniels, DL, Strottmann, JM et al., (2001) Pictorial review of glutamate excitotoxicity: fundamental concepts for neuroimaging. AJNR Am J Neuroradiol 22: 1813-24.
126.
Meldrum, BS, (2000) Glutamate as a neurotransmitter in the brain: review of physiology and pathology. J Nutr 130: 1007S-15S.
127.
Moriyama, Y and Yamamoto, A, (2004) Glutamatergic chemical transmission: look! Here, there, and anywhere. J Biochem 135: 155-63.
128.
Millan, C and Sanchez-Prieto, J, (2002) Differential coupling of N- and P/Q-type calcium channels to glutamate exocytosis in the rat cerebral cortex. Neurosci Lett 330: 29-32.
129.
Dunlap, K, Luebke, JI and Turner, TJ, (1995) Exocytotic Ca2+ channels in mammalian central neurons. Trends Neurosci 18: 89-98.
130.
Qian, J and Noebels, JL, (2001) Presynaptic Ca2+ channels and neurotransmitter release at the terminal of a mouse cortical neuron. J Neurosci 21: 3721-8.
131.
Qian, J and Noebels, JL, (2000) Presynaptic Ca(2+) influx at a mouse central synapse with Ca(2+) channel subunit mutations. J Neurosci 20: 163-70.
132.
Mintz, IM, Sabatini, BL and Regehr, WG, (1995) Calcium control of transmitter release at a cerebellar synapse. Neuron 15: 675-88.
206
133.
Fletcher, CF, Lutz, CM, O'Sullivan, TN, Shaughnessy, JD, Jr., Hawkes, R, Frankel, WN et al., (1996) Absence epilepsy in tottering mutant mice is associated with calcium channel defects. Cell 87: 607-17.
134.
Doyle, J, Ren, X, Lennon, G and Stubbs, L, (1997) Mutations in the Cacnl1a4 calcium channel gene are associated with seizures, cerebellar degeneration, and ataxia in tottering and leaner mutant mice. Mamm Genome 8: 113-20.
135.
Wakamori, M, Yamazaki, K, Matsunodaira, H, Teramoto, T, Tanaka, I, Niidome, T et al., (1998) Single tottering mutations responsible for the neuropathic phenotype of the P-type calcium channel. J Biol Chem 273: 34857-67.
136.
Glaum, SR, Miller, R.J., (1994) Acute regulation of synaptic transmission by metabotropic glutamate receptors. In "The Metabotropic Glutamate Receptors" Conn, P. J., Patel, J. ed. (Humana, Totowa, NJ). 137.
Conn, PJ and Pin, JP, (1997) Pharmacology and functions of metabotropic glutamate receptors. Annu Rev Pharmacol Toxicol 37: 205-37.
138.
Sayer, RJ, Schwindt, PC and Crill, WE, (1992) Metabotropic glutamate receptormediated suppression of L-type calcium current in acutely isolated neocortical neurons. J Neurophysiol 68: 833-42.
139.
Swartz, KJ and Bean, BP, (1992) Inhibition of calcium channels in rat CA3 pyramidal neurons by a metabotropic glutamate receptor. J Neurosci 12: 4358-71.
140.
Stefani, A, Pisani, A, Mercuri, NB, Bernardi, G and Calabresi, P, (1994) Activation of metabotropic glutamate receptors inhibits calcium currents and GABA-mediated synaptic potentials in striatal neurons. J Neurosci 14: 6734-43.
141.
Rothe, T, Bigl, V and Grantyn, R, (1994) Potentiating and depressant effects of metabotropic glutamate receptor agonists on high-voltage-activated calcium currents in cultured retinal ganglion neurons from postnatal mice. Pflugers Arch 426: 161-70.
142.
Chavis, P, Fagni, L, Bockaert, J and Lansman, JB, (1995) Modulation of calcium channels by metabotropic glutamate receptors in cerebellar granule cells. Neuropharmacology 34: 929-37.
143.
Ikeda, SR, (1996) Voltage-dependent modulation of N-type calcium channels by G-protein beta gamma subunits. Nature 380: 255-8.
207
144.
Herlitze, S, Garcia, DE, Mackie, K, Hille, B, Scheuer, T and Catterall, WA, (1996) Modulation of Ca2+ channels by G-protein beta gamma subunits. Nature 380: 258-62.
145.
Gereau, RWt and Conn, PJ, (1995) Multiple presynaptic metabotropic glutamate receptors modulate excitatory and inhibitory synaptic transmission in hippocampal area CA1. J Neurosci 15: 6879-89.
146.
Scanziani, M, Gahwiler, BH and Thompson, SM, (1995) Presynaptic inhibition of excitatory synaptic transmission by muscarinic and metabotropic glutamate receptor activation in the hippocampus: are Ca2+ channels involved? Neuropharmacology 34: 1549-57.
147.
Tyler, EC and Lovinger, DM, (1995) Metabotropic glutamate receptor modulation of synaptic transmission in corticostriatal co-cultures: role of calcium influx. Neuropharmacology 34: 939-52.
148.
Sladeczek, F, Momiyama, A., Takahashi, T., (1993) Presynaptic inhibitory action of a metabotropic glutamate receptor agonist on excitatory transmission in visual cortical neurons. Proc. R. Soc. London B 253: 297–303.
149.
Wroblewski, JT, (1992) Na+/Ca*' exchange and glutamate neurotoxicity. Neurosci. Facts 3: 76.
150.
Yu, SP and Choi, DW, (1997) Na(+)-Ca2+ exchange currents in cortical neurons: concomitant forward and reverse operation and effect of glutamate. Eur J Neurosci 9: 1273-81.
151.
Juhaszova, M, Shimizu, H, Borin, ML, Yip, RK, Santiago, EM, Lindenmayer, GE et al., (1996) Localization of the Na(+)-Ca2+ exchanger in vascular smooth muscle, and in neurons and astrocytes. Ann N Y Acad Sci 779: 318-35.
152.
Luther, PW, Yip, RK, Bloch, RJ, Ambesi, A, Lindenmayer, GE and Blaustein, MP, (1992) Presynaptic localization of sodium/calcium exchangers in neuromuscular preparations. J Neurosci 12: 4898-904.
153.
Reuter, H and Porzig, H, (1995) Localization and functional significance of the Na+/Ca2+ exchanger in presynaptic boutons of hippocampal cells in culture. Neuron 15: 1077-84.
154.
Borst, JG and Sakmann, B, (1996) Calcium influx and transmitter release in a fast CNS synapse. Nature 383: 431-4.
208
155.
Llinas, R, Sugimori, M and Silver, RB, (1992) Microdomains of high calcium concentration in a presynaptic terminal. Science 256: 677-9.
156.
Llinas, R, Sugimori, M and Silver, RB, (1992) Presynaptic calcium concentration microdomains and transmitter release. J Physiol Paris 86: 135-8.
157.
Aarhus, LL, Shepherd, JT, Tyce, GM, Verbeuren, TJ and Vanhoutte, PM, (1983) Contractions of canine vascular smooth muscle cells caused by ouabain are due to release of norepinephrine from adrenergic nerve endings. Circ Res 52: 501-7.
158.
Blaustein, MP, Goldman, WF, Fontana, G, Krueger, BK, Santiago, EM, Steele, TD et al., (1991) Physiological roles of the sodium-calcium exchanger in nerve and muscle. Ann N Y Acad Sci 639: 254-74.
159.
Sanchez-Armass, S and Blaustein, MP, (1987) Role of sodium-calcium exchange in regulation of intracellular calcium in nerve terminals. Am J Physiol 252: C595603.
160.
Blaustein, MP and Lederer, WJ, (1999) Sodium/calcium exchange: its physiological implications. Physiol Rev 79: 763-854.
161.
Atwood, HL, Charlton, MP and Thompson, CS, (1983) Neuromuscular transmission in crustaceans is enhanced by a sodium ionophore, monensin, and by prolonged stimulation. J Physiol 335: 179-95.
162.
Atwood, HL, Swenarchuk, LE and Gruenwald, CR, (1975) Long-term synaptic facilitation during sodium accumulation in nerve terminals. Brain Res 100: 198202.
163.
Charlton, MP and Atwood, HL, (1977) Modulation of transmitter release by intracellular sodium in squid giant synapse. Brain Res 134: 367-71.
164.
Charlton, MP, Thompson, CS, Atwood, HL and Farnell, B, (1980) Synaptic transmission and intracellular sodium: ionophore induced sodium loading of nerve terminals. Neurosci Lett 16: 193-6.
165.
Birks, RI and Cohen, MW, (1968) The influence of internal sodium on the behaviour of motor nerve endings. Proc R Soc Lond B Biol Sci 170: 401-21.
166.
Birks, RI and Cohen, MW, (1968) The action of sodium pump inhibitors on neuromuscular transmission. Proc R Soc Lond B Biol Sci 170: 381-99.
209
167.
Kelly, JS, (1998) Antagonism between Na+ and Ca2+ at the neuromuscular junction. Nature 205: 296.
168.
Meiri, H, Erulkar, SD, Lerman, T and Rahamimoff, R, (1981) The action of the sodium ionophore, monensin, or transmitter release at the frog neuromuscular junction. Brain Res 204: 204-8.
169.
Blaszkowski, TP and Bogdanski, DF, (1971) Possible role of sodium and calcium ions in retention and physiological release of norepinephrine by adrenergic nerve endings. Biochem Pharmacol 20: 3281-94.
170.
Kirpekar, SM and Wakade, AR, (1968) Release of noradrenaline from the cat spleen by potassium. J Physiol 194: 595-608.
171.
Bouron, A and Reuter, H, (1996) A role of intracellular Na+ in the regulation of synaptic transmission and turnover of the vesicular pool in cultured hippocampal cells. Neuron 17: 969-78.
172.
Collard, KJ, (1989) On the mechanism by which extracellular sodium depletion causes 5-hydroxytryptamine release from rat brain synaptosomes. Biochim Biophys Acta 984: 319-25.
173.
Taglialatela, M, Amoroso, S, Canzoniero, LM, Di Renzo, GF and Annunziato, L, (1988) Membrane events and ionic processes involved in dopamine release from tuberoinfundibular neurons. II. Effect of the inhibition of the Na+-Ca++ exchange by amiloride. J Pharmacol Exp Ther 246: 689-94.
174.
Moriyoshi, K, Masu, M, Ishii, T, Shigemoto, R, Mizuno, N and Nakanishi, S, (1991) Molecular cloning and characterization of the rat NMDA receptor. Nature 354: 31-7.
175.
Monyer, H, Sprengel, R, Schoepfer, R, Herb, A, Higuchi, M, Lomeli, H et al., (1992) Heteromeric NMDA receptors: molecular and functional distinction of subtypes. Science 256: 1217-21.
176.
Sommer, B, Monyer, H, Wisden, W, Verdoorn, TA, Burnashev, N, Sprengel, R et al., (1992) Glutamate-gated ion channels in the brain. Genetic mechanism for generating molecular and functional diversity. Arzneimittelforschung 42: 209-10.
177.
Wo, ZG and Oswald, RE, (1995) Unraveling the modular design of glutamategated ion channels. Trends Neurosci 18: 161-8.
210
178.
Madden, DR, (2002) The structure and function of glutamate receptor ion channels. Nat Rev Neurosci 3: 91-101.
179.
Hardingham, GE, Arnold, FJ and Bading, H, (2001) A calcium microdomain near NMDA receptors: on switch for ERK-dependent synapse-to-nucleus communication. Nat Neurosci 4: 565-6.
180.
Hardingham, GE, Arnold, FJ and Bading, H, (2001) Nuclear calcium signaling controls CREB-mediated gene expression triggered by synaptic activity. Nat Neurosci 4: 261-7.
181.
Tao, X, Finkbeiner, S, Arnold, DB, Shaywitz, AJ and Greenberg, ME, (1998) Ca2+ influx regulates BDNF transcription by a CREB family transcription factordependent mechanism. Neuron 20: 709-26.
182.
Wang, JM, Chao, JR, Chen, W, Kuo, ML, Yen, JJ and Yang-Yen, HF, (1999) The antiapoptotic gene mcl-1 is up-regulated by the phosphatidylinositol 3-kinase/Akt signaling pathway through a transcription factor complex containing CREB. Mol Cell Biol 19: 6195-206.
183.
Bok, J, Wang, Q, Huang, J and Green, SH, (2007) CaMKII and CaMKIV mediate distinct prosurvival signaling pathways in response to depolarization in neurons. Mol Cell Neurosci 36: 13-26.
184.
Hollmann, M and Heinemann, S, (1994) Cloned glutamate receptors. Annu Rev Neurosci 17: 31-108.
185.
Bennett, JA and Dingledine, R, (1995) Topology profile for a glutamate receptor: three transmembrane domains and a channel-lining reentrant membrane loop. Neuron 14: 373-84.
186.
Wisden, W and Seeburg, PH, (1993) Mammalian ionotropic glutamate receptors. Curr Opin Neurobiol 3: 291-8.
187.
Zamanillo, D, Sprengel, R, Hvalby, O, Jensen, V, Burnashev, N, Rozov, A et al., (1999) Importance of AMPA receptors for hippocampal synaptic plasticity but not for spatial learning. Science 284: 1805-11.
188.
Passafaro, M, Piech, V and Sheng, M, (2001) Subunit-specific temporal and spatial patterns of AMPA receptor exocytosis in hippocampal neurons. Nat Neurosci 4: 917-26.
211
189.
Isaac, JT, Ashby, M and McBain, CJ, (2007) The role of the GluR2 subunit in AMPA receptor function and synaptic plasticity. Neuron 54: 859-71.
190.
Borges, K and Dingledine, R, (1998) AMPA receptors: molecular and functional diversity. Prog Brain Res 116: 153-70.
191.
Chen, L, Chetkovich, DM, Petralia, RS, Sweeney, NT, Kawasaki, Y, Wenthold, RJ et al., (2000) Stargazin regulates synaptic targeting of AMPA receptors by two distinct mechanisms. Nature 408: 936-43.
192.
Ziff, EB, (2007) TARPs and the AMPA receptor trafficking paradox. Neuron 53: 627-33.
193.
Bleakman, D, (1999) Kainate receptor pharmacology and physiology. Cell Mol Life Sci 56: 558-66.
194.
Bortolotto, ZA, Clarke, VR, Delany, CM, Parry, MC, Smolders, I, Vignes, M et al., (1999) Kainate receptors are involved in synaptic plasticity. Nature 402: 297301.
195.
Traynelis, SF and Wahl, P, (1997) Control of rat GluR6 glutamate receptor open probability by protein kinase A and calcineurin. J Physiol 503 ( Pt 3): 513-31.
196.
Sanchez-Gomez, MV and Matute, C, (1999) AMPA and kainate receptors each mediate excitotoxicity in oligodendroglial cultures. Neurobiol Dis 6: 475-85.
197.
Chittajallu, R, Vignes, M, Dev, KK, Barnes, JM, Collingridge, GL and Henley, JM, (1996) Regulation of glutamate release by presynaptic kainate receptors in the hippocampus. Nature 379: 78-81.
198.
Yu, CZ, Li, C, Pei, DS, Zong, YY, Shi, Q, Wen, XR et al., (2009) Neuroprotection against transient focal cerebral ischemia and oxygen-glucose deprivation by interference with GluR6-PSD95 protein interaction. Neurochem Res 34: 2008-21.
199.
Pei, DS, Wang, XT, Liu, Y, Sun, YF, Guan, QH, Wang, W et al., (2006) Neuroprotection against ischaemic brain injury by a GluR6-9c peptide containing the TAT protein transduction sequence. Brain 129: 465-79.
200.
Aramori, I and Nakanishi, S, (1992) Signal transduction and pharmacological characteristics of a metabotropic glutamate receptor, mGluR1, in transfected CHO cells. Neuron 8: 757-65.
212
201.
Abe, T, Sugihara, H, Nawa, H, Shigemoto, R, Mizuno, N and Nakanishi, S, (1992) Molecular characterization of a novel metabotropic glutamate receptor mGluR5 coupled to inositol phosphate/Ca2+ signal transduction. J Biol Chem 267: 13361-8.
202.
Heuss, C, Scanziani, M, Gahwiler, BH and Gerber, U, (1999) G-proteinindependent signaling mediated by metabotropic glutamate receptors. Nat Neurosci 2: 1070-7.
203.
Tanabe, Y, Nomura, A, Masu, M, Shigemoto, R, Mizuno, N and Nakanishi, S, (1993) Signal transduction, pharmacological properties, and expression patterns of two rat metabotropic glutamate receptors, mGluR3 and mGluR4. J Neurosci 13: 1372-8.
204.
Chavis, P, Shinozaki, H, Bockaert, J and Fagni, L, (1994) The metabotropic glutamate receptor types 2/3 inhibit L-type calcium channels via a pertussis toxinsensitive G-protein in cultured cerebellar granule cells. J Neurosci 14: 7067-76.
205.
Fei, Z, Zhang, X, Bai, HM, Jiang, XF and Wang, XL, (2006) Metabotropic glutamate receptor antagonists and agonists: potential neuroprotectors in diffuse brain injury. J Clin Neurosci 13: 1023-7.
206.
Fei, Z, Zhang, X, Jiang, XF, Huang, WD and Bai, HM, (2005) Altered expression patterns of metabotropic glutamate receptors in diffuse brain injury. Neurosci Lett 380: 280-3.
207.
Bai, HM, Wang, WM, Li, TD and Fei, Z, (2004) Changes in metabotropic glutamate receptor 4 expression and the effects of L-2-amino-4phosphonobutyrate in a rodent model of diffuse brain injury. Chin J Traumatol 7: 233-8.
208.
Fei, Z, Zhang, X and Liu, EY, (2003) Changes of metabotropic glutamate receptor subtype 1a in diffuse brain injury with secondary brain insults and the effects of 2-methyl-4-carboxyphenylglycine. Chin J Traumatol 6: 270-4.
209.
Lucas, DR and Newhouse, JP, (1957) The toxic effect of sodium L-glutamate on the inner layers of the retina. AMA Arch Ophthalmol 58: 193-201.
210.
Rothman, SM, (1985) The neurotoxicity of excitatory amino acids is produced by passive chloride influx. J Neurosci 5: 1483-9.
211.
Choi, DW, (1987) Ionic dependence of glutamate neurotoxicity. J Neurosci 7: 369-79.
213
212.
Sakowitz, OW, Stover, JF, Sarrafzadeh, AS, Unterberg, AW and Kiening, KL, (2007) Effects of mannitol bolus administration on intracranial pressure, cerebral extracellular metabolites, and tissue oxygenation in severely head-injured patients. J Trauma 62: 292-8.
213.
Reinert, M, Khaldi, A, Zauner, A, Doppenberg, E, Choi, S and Bullock, R, (2000) High level of extracellular potassium and its correlates after severe head injury: relationship to high intracranial pressure. J Neurosurg 93: 800-7.
214.
Alessandri, B, Doppenberg, E, Zauner, A, Woodward, J, Choi, S and Bullock, R, (1999) Evidence for time-dependent glutamate-mediated glycolysis in headinjured patients: a microdialysis study. Acta Neurochir Suppl 75: 25-8.
215.
Bullock, R, Zauner, A, Woodward, JJ, Myseros, J, Choi, SC, Ward, JD et al., (1998) Factors affecting excitatory amino acid release following severe human head injury. J Neurosurg 89: 507-18.
216.
Alessandri, B, Reinert, M, Young, HF and Bullock, R, (2000) Low extracellular (ECF) glucose affects the neurochemical profile in severe head-injured patients. Acta Neurochir Suppl 76: 425-30.
217.
Yamamoto, T, Rossi, S, Stiefel, M, Doppenberg, E, Zauner, A, Bullock, R et al., (1999) CSF and ECF glutamate concentrations in head injured patients. Acta Neurochir Suppl 75: 17-9.
218.
Vespa, P, Prins, M, Ronne-Engstrom, E, Caron, M, Shalmon, E, Hovda, DA et al., (1998) Increase in extracellular glutamate caused by reduced cerebral perfusion pressure and seizures after human traumatic brain injury: a microdialysis study. J Neurosurg 89: 971-82.
219.
Matsushita, Y, Shima, K, Nawashiro, H, Wada, K, Tsuzuki, N and Miyazawa, T, (2000) Real time monitoring of glutamate following fluid percussion brain injury with hypoxia in the rat. Acta Neurochir Suppl 76: 207-12.
220.
Bullock, R, Butcher, SP, Chen, MH, Kendall, L and McCulloch, J, (1991) Correlation of the extracellular glutamate concentration with extent of blood flow reduction after subdural hematoma in the rat. J Neurosurg 74: 794-802.
221.
Globus, MY, Alonso, O, Dietrich, WD, Busto, R and Ginsberg, MD, (1995) Glutamate release and free radical production following brain injury: effects of posttraumatic hypothermia. J Neurochem 65: 1704-11.
214
222.
Bano, D, Young, KW, Guerin, CJ, Lefeuvre, R, Rothwell, NJ, Naldini, L et al., (2005) Cleavage of the plasma membrane Na+/Ca2+ exchanger in excitotoxicity. Cell 120: 275-85.
223.
Floyd, CL, Gorin, FA and Lyeth, BG, (2005) Mechanical strain injury increases intracellular sodium and reverses Na+/Ca2+ exchange in cortical astrocytes. Glia 51: 35-46.
224.
Zhao, X, Gorin, FA, Berman, RF and Lyeth, BG, (2008) Differential hippocampal protection when blocking intracellular sodium and calcium entry during traumatic brain injury in rats. J Neurotrauma 25: 1195-205.
225.
Danbolt, NC, (2001) Glutamate uptake. Prog Neurobiol 65: 1-105.
226.
Tanaka, K, Watase, K, Manabe, T, Yamada, K, Watanabe, M, Takahashi, K et al., (1997) Epilepsy and exacerbation of brain injury in mice lacking the glutamate transporter GLT-1. Science 276: 1699-702.
227.
Danbolt, NC, Storm-Mathisen, J and Kanner, BI, (1992) An [Na+ + K+]coupled L-glutamate transporter purified from rat brain is located in glial cell processes. Neuroscience 51: 295-310.
228.
Haugeto, O, Ullensvang, K, Levy, LM, Chaudhry, FA, Honore, T, Nielsen, M et al., (1996) Brain glutamate transporter proteins form homomultimers. J Biol Chem 271: 27715-22.
229.
Rao, VL, Baskaya, MK, Dogan, A, Rothstein, JD and Dempsey, RJ, (1998) Traumatic brain injury down-regulates glial glutamate transporter (GLT-1 and GLAST) proteins in rat brain. J Neurochem 70: 2020-7.
230.
van Landeghem, FK, Stover, JF, Bechmann, I, Bruck, W, Unterberg, A, Buhrer, C et al., (2001) Early expression of glutamate transporter proteins in ramified microglia after controlled cortical impact injury in the rat. Glia 35: 167-79.
231.
Phillis, JW, Ren, J and O'Regan, MH, (2000) Transporter reversal as a mechanism of glutamate release from the ischemic rat cerebral cortex: studies with DL-threobeta-benzyloxyaspartate. Brain Res 880: 224.
232.
Hamann, M, Rossi, DJ, Marie, H and Attwell, D, (2002) Knocking out the glial glutamate transporter GLT-1 reduces glutamate uptake but does not affect hippocampal glutamate dynamics in early simulated ischaemia. Eur J Neurosci 15: 308-14.
215
233.
Koizumi, H, Fujisawa, H, Ito, H, Maekawa, T, Di, X and Bullock, R, (1997) Effects of mild hypothermia on cerebral blood flow-independent changes in cortical extracellular levels of amino acids following contusion trauma in the rat. Brain Res 747: 304-12.
234.
Povlishock, JT, Lyeth, B., (1989) Traumatically induced blood-brain barrier disruption: a conduit for the passage of circulating excitatory neurotransmitters. Society for Neuroscience Abstracts 15: 1113.
235.
Gong, QZ, Phillips, LL and Lyeth, BG, (1999) Metabotropic glutamate receptor protein alterations after traumatic brain injury in rats. J Neurotrauma 16: 893-902.
236.
Allen, JW, Ivanova, SA, Fan, L, Espey, MG, Basile, AS and Faden, AI, (1999) Group II metabotropic glutamate receptor activation attenuates traumatic neuronal injury and improves neurological recovery after traumatic brain injury. J Pharmacol Exp Ther 290: 112-20.
237.
Faden, AI, Ivanova, SA, Yakovlev, AG and Mukhin, AG, (1997) Neuroprotective effects of group III mGluR in traumatic neuronal injury. J Neurotrauma 14: 88595.
238.
Nilsson, P, Ronne-Engstrom, E, Flink, R, Ungerstedt, U, Carlson, H and Hillered, L, (1994) Epileptic seizure activity in the acute phase following cortical impact trauma in rat. Brain Res 637: 227-32.
239.
Obrenovitch, TP and Urenjak, J, (1997) Is high extracellular glutamate the key to excitotoxicity in traumatic brain injury? J Neurotrauma 14: 677-98.
240.
Obrenovitch, TP and Urenjak, J, (1997) Altered glutamatergic transmission in neurological disorders: from high extracellular glutamate to excessive synaptic efficacy. Prog Neurobiol 51: 39-87.
241.
Nilsson, P, Hillered, L, Ponten, U and Ungerstedt, U, (1990) Changes in cortical extracellular levels of energy-related metabolites and amino acids following concussive brain injury in rats. J Cereb Blood Flow Metab 10: 631-7.
242.
Hillered, L, Nilsson, P, Ungerstedt, U and Ponten, U, (1990) Trauma-induced increase of extracellular ascorbate in rat cerebral cortex. Neurosci Lett 113: 32832.
243.
Zhang, L, Rzigalinski, BA, Ellis, EF and Satin, LS, (1996) Reduction of voltagedependent Mg2+ blockade of NMDA current in mechanically injured neurons. Science 274: 1921-3.
216
244.
Geddes-Klein, DM, Serbest, G, Mesfin, MN, Cohen, AS and Meaney, DF, (2006) Pharmacologically induced calcium oscillations protect neurons from increases in cytosolic calcium after trauma. J Neurochem 97: 462-74.
245.
Geddes-Klein, DM, Schiffman, KB and Meaney, DF, (2006) Mechanisms and consequences of neuronal stretch injury in vitro differ with the model of trauma. J Neurotrauma 23: 193-204.
246.
Geddes, DM, LaPlaca, MC and Cargill, RS, 2nd, (2003) Susceptibility of hippocampal neurons to mechanically induced injury. Exp Neurol 184: 420-7.
247.
Lea, PMt, Custer, SJ, Stoica, BA and Faden, AI, (2003) Modulation of stretchinduced enhancement of neuronal NMDA receptor current by mGluR1 depends upon presence of glia. J Neurotrauma 20: 1233-49.
248.
Lau, A, Arundine, M, Sun, HS, Jones, M and Tymianski, M, (2006) Inhibition of caspase-mediated apoptosis by peroxynitrite in traumatic brain injury. J Neurosci 26: 11540-53.
249.
Goforth, PB, Ellis, EF and Satin, LS, (1999) Enhancement of AMPA-mediated current after traumatic injury in cortical neurons. J Neurosci 19: 7367-74.
250.
Goforth, PB, Ellis, EF and Satin, LS, (2004) Mechanical injury modulates AMPA receptor kinetics via an NMDA receptor-dependent pathway. J Neurotrauma 21: 719-32.
251.
Bell, JD, Ai, J, Chen, Y and Baker, AJ, (2007) Mild in vitro trauma induces rapid Glur2 endocytosis, robustly augments calcium permeability and enhances susceptibility to secondary excitotoxic insult in cultured Purkinje cells. Brain 130: 2528-42.
252.
Spaethling, JM, Klein, DM, Singh, P and Meaney, DF, (2008) Calciumpermeable AMPA receptors appear in cortical neurons after traumatic mechanical injury and contribute to neuronal fate. J Neurotrauma 25: 1207-16.
253.
Bayir, H, Kagan, VE, Clark, RS, Janesko-Feldman, K, Rafikov, R, Huang, Z et al., (2007) Neuronal NOS-mediated nitration and inactivation of manganese superoxide dismutase in brain after experimental and human brain injury. J Neurochem 101: 168-81.
254.
Halliwell, B, Zhao, K and Whiteman, M, (1999) Nitric oxide and peroxynitrite. The ugly, the uglier and the not so good: a personal view of recent controversies. Free Radic Res 31: 651-69.
217
255.
Radi, R, Beckman, JS, Bush, KM and Freeman, BA, (1991) Peroxynitrite-induced membrane lipid peroxidation: the cytotoxic potential of superoxide and nitric oxide. Arch Biochem Biophys 288: 481-7.
256.
Dawson, VL and Dawson, TM, (2004) Deadly conversations: nuclearmitochondrial cross-talk. J Bioenerg Biomembr 36: 287-94.
257.
Kim, JS, He, L, Qian, T and Lemasters, JJ, (2003) Role of the mitochondrial permeability transition in apoptotic and necrotic death after ischemia/reperfusion injury to hepatocytes. Curr Mol Med 3: 527-35.
258.
Kim, JS, He, L and Lemasters, JJ, (2003) Mitochondrial permeability transition: a common pathway to necrosis and apoptosis. Biochem Biophys Res Commun 304: 463-70.
259.
Kim, JS, Qian, T and Lemasters, JJ, (2003) Mitochondrial permeability transition in the switch from necrotic to apoptotic cell death in ischemic rat hepatocytes. Gastroenterology 124: 494-503.
260.
Lemasters, JJ, Qian, T, He, L, Kim, JS, Elmore, SP, Cascio, WE et al., (2002) Role of mitochondrial inner membrane permeabilization in necrotic cell death, apoptosis, and autophagy. Antioxid Redox Signal 4: 769-81.
261.
Zou, H, Yang, R, Hao, J, Wang, J, Sun, C, Fesik, SW et al., (2003) Regulation of the Apaf-1/caspase-9 apoptosome by caspase-3 and XIAP. J Biol Chem 278: 8091-8.
262.
Zou, H, Li, Y, Liu, X and Wang, X, (1999) An APAF-1.cytochrome c multimeric complex is a functional apoptosome that activates procaspase-9. J Biol Chem 274: 11549-56.
263.
Zou, H, Henzel, WJ, Liu, X, Lutschg, A and Wang, X, (1997) Apaf-1, a human protein homologous to C. elegans CED-4, participates in cytochrome c-dependent activation of caspase-3. Cell 90: 405-13.
264.
Buchan, AM, Li, H, Cho, S and Pulsinelli, WA, (1991) Blockade of the AMPA receptor prevents CA1 hippocampal injury following severe but transient forebrain ischemia in adult rats. Neurosci Lett 132: 255-8.
265.
Sheardown, MJ, (1993) The pharmacology of AMPA receptors and their antagonists. Stroke 24: I146-7; discussion I153-4.
218
266.
Sheardown, MJ, Suzdak, PD and Nordholm, L, (1993) AMPA, but not NMDA, receptor antagonism is neuroprotective in gerbil global ischaemia, even when delayed 24 h. Eur J Pharmacol 236: 347-53.
267.
Terashima, A, Cotton, L, Dev, KK, Meyer, G, Zaman, S, Duprat, F et al., (2004) Regulation of synaptic strength and AMPA receptor subunit composition by PICK1. J Neurosci 24: 5381-90.
268.
Terashima, A, Pelkey, KA, Rah, JC, Suh, YH, Roche, KW, Collingridge, GL et al., (2008) An essential role for PICK1 in NMDA receptor-dependent bidirectional synaptic plasticity. Neuron 57: 872-82.
269.
Plant, K, Pelkey, KA, Bortolotto, ZA, Morita, D, Terashima, A, McBain, CJ et al., (2006) Transient incorporation of native GluR2-lacking AMPA receptors during hippocampal long-term potentiation. Nat Neurosci 9: 602-4.
270.
Conrad, KL, Tseng, KY, Uejima, JL, Reimers, JM, Heng, LJ, Shaham, Y et al., (2008) Formation of accumbens GluR2-lacking AMPA receptors mediates incubation of cocaine craving. Nature 454: 118-21.
271.
Mameli, M, Balland, B, Lujan, R and Luscher, C, (2007) Rapid synthesis and synaptic insertion of GluR2 for mGluR-LTD in the ventral tegmental area. Science 317: 530-3.
272.
Bellone, C and Luscher, C, (2006) Cocaine triggered AMPA receptor redistribution is reversed in vivo by mGluR-dependent long-term depression. Nat Neurosci 9: 636-41.
273.
Hemby, SE, Tang, W, Muly, EC, Kuhar, MJ, Howell, L and Mash, DC, (2005) Cocaine-induced alterations in nucleus accumbens ionotropic glutamate receptor subunits in human and non-human primates. J Neurochem 95: 1785-93.
274.
Churchill, L, Swanson, CJ, Urbina, M and Kalivas, PW, (1999) Repeated cocaine alters glutamate receptor subunit levels in the nucleus accumbens and ventral tegmental area of rats that develop behavioral sensitization. J Neurochem 72: 2397-403.
275.
Fitzgerald, LW, Ortiz, J, Hamedani, AG and Nestler, EJ, (1996) Drugs of abuse and stress increase the expression of GluR1 and NMDAR1 glutamate receptor subunits in the rat ventral tegmental area: common adaptations among crosssensitizing agents. J Neurosci 16: 274-82.
219
276.
Rakhade, SN, Zhou, C, Aujla, PK, Fishman, R, Sucher, NJ and Jensen, FE, (2008) Early alterations of AMPA receptors mediate synaptic potentiation induced by neonatal seizures. J Neurosci 28: 7979-90.
277.
Prince, HK, Conn, PJ, Blackstone, CD, Huganir, RL and Levey, AI, (1995) Down-regulation of AMPA receptor subunit GluR2 in amygdaloid kindling. J Neurochem 64: 462-5.
278.
Pollard, H, Heron, A, Moreau, J, Ben-Ari, Y and Khrestchatisky, M, (1993) Alterations of the GluR-B AMPA receptor subunit flip/flop expression in kainateinduced epilepsy and ischemia. Neuroscience 57: 545-54.
279.
Liu, B, Liao, M, Mielke, JG, Ning, K, Chen, Y, Li, L et al., (2006) Ischemic insults direct glutamate receptor subunit 2-lacking AMPA receptors to synaptic sites. J Neurosci 26: 5309-19.
280.
Noh, KM, Yokota, H, Mashiko, T, Castillo, PE, Zukin, RS and Bennett, MV, (2005) Blockade of calcium-permeable AMPA receptors protects hippocampal neurons against global ischemia-induced death. Proc Natl Acad Sci U S A 102: 12230-5.
281.
Calderone, A, Jover, T, Noh, KM, Tanaka, H, Yokota, H, Lin, Y et al., (2003) Ischemic insults derepress the gene silencer REST in neurons destined to die. J Neurosci 23: 2112-21.
282.
Tanaka, H, Calderone, A, Jover, T, Grooms, SY, Yokota, H, Zukin, RS et al., (2002) Ischemic preconditioning acts upstream of GluR2 down-regulation to afford neuroprotection in the hippocampal CA1. Proc Natl Acad Sci U S A 99: 2362-7.
283.
Opitz, T, Grooms, SY, Bennett, MV and Zukin, RS, (2000) Remodeling of alphaamino-3-hydroxy-5-methyl-4-isoxazole-propionic acid receptor subunit composition in hippocampal neurons after global ischemia. Proc Natl Acad Sci U S A 97: 13360-5.
284.
Aronica, EM, Gorter, JA, Grooms, S, Kessler, JA, Bennett, MV, Zukin, RS et al., (1998) Aurintricarboxylic acid prevents GLUR2 mRNA down-regulation and delayed neurodegeneration in hippocampal CA1 neurons of gerbil after global ischemia. Proc Natl Acad Sci U S A 95: 7115-20.
285.
Gorter, JA, Petrozzino, JJ, Aronica, EM, Rosenbaum, DM, Opitz, T, Bennett, MV et al., (1997) Global ischemia induces downregulation of Glur2 mRNA and increases AMPA receptor-mediated Ca2+ influx in hippocampal CA1 neurons of gerbil. J Neurosci 17: 6179-88. 220
286.
Pellegrini-Giampietro, DE, Zukin, RS, Bennett, MV, Cho, S and Pulsinelli, WA, (1992) Switch in glutamate receptor subunit gene expression in CA1 subfield of hippocampus following global ischemia in rats. Proc Natl Acad Sci U S A 89: 10499-503.
287.
Lu, YM, Yin, HZ, Chiang, J and Weiss, JH, (1996) Ca(2+)-permeable AMPA/kainate and NMDA channels: high rate of Ca2+ influx underlies potent induction of injury. J Neurosci 16: 5457-65.
288.
Carriedo, SG, Sensi, SL, Yin, HZ and Weiss, JH, (2000) AMPA exposures induce mitochondrial Ca(2+) overload and ROS generation in spinal motor neurons in vitro. J Neurosci 20: 240-50.
289.
Sensi, SL, Yin, HZ, Carriedo, SG, Rao, SS and Weiss, JH, (1999) Preferential Zn2+ influx through Ca2+-permeable AMPA/kainate channels triggers prolonged mitochondrial superoxide production. Proc Natl Acad Sci U S A 96: 2414-9.
290.
Carriedo, SG, Yin, HZ, Sensi, SL and Weiss, JH, (1998) Rapid Ca2+ entry through Ca2+-permeable AMPA/Kainate channels triggers marked intracellular Ca2+ rises and consequent oxygen radical production. J Neurosci 18: 7727-38.
291.
Dykens, JA, Stern, A and Trenkner, E, (1987) Mechanism of kainate toxicity to cerebellar neurons in vitro is analogous to reperfusion tissue injury. J Neurochem 49: 1222-8.
292.
Patel, M, Day, BJ, Crapo, JD, Fridovich, I and McNamara, JO, (1996) Requirement for superoxide in excitotoxic cell death. Neuron 16: 345-55.
293.
Murphy, SN and Miller, RJ, (1989) Regulation of Ca++ influx into striatal neurons by kainic acid. J Pharmacol Exp Ther 249: 184-93.
294.
Miller, RJ, Murphy, SN and Glaum, SR, (1989) Neuronal Ca2+ channels and their regulation by excitatory amino acids. Ann N Y Acad Sci 568: 149-58.
295.
Weiss, JH, Hartley, DM, Koh, J and Choi, DW, (1990) The calcium channel blocker nifedipine attenuates slow excitatory amino acid neurotoxicity. Science 247: 1474-7.
296.
Choi, DW, Weiss, JH, Koh, JY, Christine, CW and Kurth, MC, (1989) Glutamate neurotoxicity, calcium, and zinc. Ann N Y Acad Sci 568: 219-24.
297.
Gill, R and Lodge, D, (1997) Pharmacology of AMPA antagonists and their role in neuroprotection. Int Rev Neurobiol 40: 197-232.
221
298.
Liu, SJ and Zukin, RS, (2007) Ca2+-permeable AMPA receptors in synaptic plasticity and neuronal death. Trends Neurosci 30: 126-34.
299.
Pellegrini-Giampietro, DE, Gorter, JA, Bennett, MV and Zukin, RS, (1997) The GluR2 (GluR-B) hypothesis: Ca(2+)-permeable AMPA receptors in neurological disorders. Trends Neurosci 20: 464-70.
300.
Bennett, MV, Pellegrini-Giampietro, DE, Gorter, JA, Aronica, E, Connor, JA and Zukin, RS, (1996) The GluR2 hypothesis: Ca(++)-permeable AMPA receptors in delayed neurodegeneration. Cold Spring Harb Symp Quant Biol 61: 373-84.
301.
Liu, S, Lau, L, Wei, J, Zhu, D, Zou, S, Sun, HS et al., (2004) Expression of Ca(2+)-permeable AMPA receptor channels primes cell death in transient forebrain ischemia. Neuron 43: 43-55.
302.
Ballas, N and Mandel, G, (2005) The many faces of REST oversee epigenetic programming of neuronal genes. Curr Opin Neurobiol 15: 500-6.
303.
Ballas, N, Grunseich, C, Lu, DD, Speh, JC and Mandel, G, (2005) REST and its corepressors mediate plasticity of neuronal gene chromatin throughout neurogenesis. Cell 121: 645-57.
304.
Ballas, N, Battaglioli, E, Atouf, F, Andres, ME, Chenoweth, J, Anderson, ME et al., (2001) Regulation of neuronal traits by a novel transcriptional complex. Neuron 31: 353-65.
305.
Paquette, AJ, Perez, SE and Anderson, DJ, (2000) Constitutive expression of the neuron-restrictive silencer factor (NRSF)/REST in differentiating neurons disrupts neuronal gene expression and causes axon pathfinding errors in vivo. Proc Natl Acad Sci U S A 97: 12318-23.
306.
Roopra, A, Sharling, L, Wood, IC, Briggs, T, Bachfischer, U, Paquette, AJ et al., (2000) Transcriptional repression by neuron-restrictive silencer factor is mediated via the Sin3-histone deacetylase complex. Mol Cell Biol 20: 2147-57.
307.
Peng, PL, Zhong, X, Tu, W, Soundarapandian, MM, Molner, P, Zhu, D et al., (2006) ADAR2-dependent RNA editing of AMPA receptor subunit GluR2 determines vulnerability of neurons in forebrain ischemia. Neuron 49: 719-33.
308.
Kacharmina, JE, Job, C, Crino, P and Eberwine, J, (2000) Stimulation of glutamate receptor protein synthesis and membrane insertion within isolated neuronal dendrites. Proc Natl Acad Sci U S A 97: 11545-50.
222
309.
Collingridge, GL, Isaac, JT and Wang, YT, (2004) Receptor trafficking and synaptic plasticity. Nat Rev Neurosci 5: 952-62.
310.
Shi, S, Hayashi, Y, Esteban, JA and Malinow, R, (2001) Subunit-specific rules governing AMPA receptor trafficking to synapses in hippocampal pyramidal neurons. Cell 105: 331-43.
311.
Hayashi, Y, Shi, SH, Esteban, JA, Piccini, A, Poncer, JC and Malinow, R, (2000) Driving AMPA receptors into synapses by LTP and CaMKII: requirement for GluR1 and PDZ domain interaction. Science 287: 2262-7.
312.
Rothman, JE, (1994) Intracellular membrane fusion. Adv Second Messenger Phosphoprotein Res 29: 81-96.
313.
Nishimune, A, Isaac, JT, Molnar, E, Noel, J, Nash, SR, Tagaya, M et al., (1998) NSF binding to GluR2 regulates synaptic transmission. Neuron 21: 87-97.
314.
Osten, P, Srivastava, S, Inman, GJ, Vilim, FS, Khatri, L, Lee, LM et al., (1998) The AMPA receptor GluR2 C terminus can mediate a reversible, ATP-dependent interaction with NSF and alpha- and beta-SNAPs. Neuron 21: 99-110.
315.
Song, I, Kamboj, S, Xia, J, Dong, H, Liao, D and Huganir, RL, (1998) Interaction of the N-ethylmaleimide-sensitive factor with AMPA receptors. Neuron 21: 393400.
316.
Hanley, JG, Khatri, L, Hanson, PI and Ziff, EB, (2002) NSF ATPase and alpha/beta-SNAPs disassemble the AMPA receptor-PICK1 complex. Neuron 34: 5367.
317.
Sorkin, A, (2004) Cargo recognition during clathrin-mediated endocytosis: a team effort. Curr Opin Cell Biol 16: 392-9.
318.
Lee, SH, Liu, L, Wang, YT and Sheng, M, (2002) Clathrin adaptor AP2 and NSF interact with overlapping sites of GluR2 and play distinct roles in AMPA receptor trafficking and hippocampal LTD. Neuron 36: 661-74.
319.
Man, HY, Lin, JW, Ju, WH, Ahmadian, G, Liu, L, Becker, LE et al., (2000) Regulation of AMPA receptor-mediated synaptic transmission by clathrindependent receptor internalization. Neuron 25: 649-62.
320.
Noel, J, Ralph, GS, Pickard, L, Williams, J, Molnar, E, Uney, JB et al., (1999) Surface expression of AMPA receptors in hippocampal neurons is regulated by an NSF-dependent mechanism. Neuron 23: 365-76.
223
321.
Luscher, C, Xia, H, Beattie, EC, Carroll, RC, von Zastrow, M, Malenka, RC et al., (1999) Role of AMPA receptor cycling in synaptic transmission and plasticity. Neuron 24: 649-58.
322.
Luthi, A, Chittajallu, R, Duprat, F, Palmer, MJ, Benke, TA, Kidd, FL et al., (1999) Hippocampal LTD expression involves a pool of AMPARs regulated by the NSF-GluR2 interaction. Neuron 24: 389-99.
323.
Carroll, RC, Beattie, EC, von Zastrow, M and Malenka, RC, (2001) Role of AMPA receptor endocytosis in synaptic plasticity. Nat Rev Neurosci 2: 315-24.
324.
Carroll, RC, Beattie, EC, Xia, H, Luscher, C, Altschuler, Y, Nicoll, RA et al., (1999) Dynamin-dependent endocytosis of ionotropic glutamate receptors. Proc Natl Acad Sci U S A 96: 14112-7.
325.
Palmer, CL, Lim, W, Hastie, PG, Toward, M, Korolchuk, VI, Burbidge, SA et al., (2005) Hippocalcin functions as a calcium sensor in hippocampal LTD. Neuron 47: 487-94.
326.
Dong, H, O'Brien, RJ, Fung, ET, Lanahan, AA, Worley, PF and Huganir, RL, (1997) GRIP: a synaptic PDZ domain-containing protein that interacts with AMPA receptors. Nature 386: 279-84.
327.
Srivastava, S, Osten, P, Vilim, FS, Khatri, L, Inman, G, States, B et al., (1998) Novel anchorage of GluR2/3 to the postsynaptic density by the AMPA receptorbinding protein ABP. Neuron 21: 581-91.
328.
Dev, KK, Nishimune, A, Henley, JM and Nakanishi, S, (1999) The protein kinase C alpha binding protein PICK1 interacts with short but not long form alternative splice variants of AMPA receptor subunits. Neuropharmacology 38: 635-44.
329.
Xia, J, Zhang, X, Staudinger, J and Huganir, RL, (1999) Clustering of AMPA receptors by the synaptic PDZ domain-containing protein PICK1. Neuron 22: 179-87.
330.
Kim, E and Sheng, M, (2004) PDZ domain proteins of synapses. Nat Rev Neurosci 5: 771-81.
331.
Hung, AY and Sheng, M, (2002) PDZ domains: structural modules for protein complex assembly. J Biol Chem 277: 5699-702.
332.
Kurakin, A, Swistowski, A, Wu, SC and Bredesen, DE, (2007) The PDZ domain as a complex adaptive system. PLoS One 2: e953.
224
333.
Lim, IA, Hall, DD and Hell, JW, (2002) Selectivity and promiscuity of the first and second PDZ domains of PSD-95 and synapse-associated protein 102. J Biol Chem 277: 21697-711.
334.
Aarts, M, Liu, Y, Liu, L, Besshoh, S, Arundine, M, Gurd, JW et al., (2002) Treatment of ischemic brain damage by perturbing NMDA receptor- PSD-95 protein interactions. Science 298: 846-50.
335.
Daniels, DL, Cohen, AR, Anderson, JM and Brunger, AT, (1998) Crystal structure of the hCASK PDZ domain reveals the structural basis of class II PDZ domain target recognition. Nat Struct Biol 5: 317-25.
336.
Songyang, Z, Fanning, AS, Fu, C, Xu, J, Marfatia, SM, Chishti, AH et al., (1997) Recognition of unique carboxyl-terminal motifs by distinct PDZ domains. Science 275: 73-7.
337.
Stricker, NL, Christopherson, KS, Yi, BA, Schatz, PJ, Raab, RW, Dawes, G et al., (1997) PDZ domain of neuronal nitric oxide synthase recognizes novel C-terminal peptide sequences. Nat Biotechnol 15: 336-42.
338.
Tochio, H, Zhang, Q, Mandal, P, Li, M and Zhang, M, (1999) Solution structure of the extended neuronal nitric oxide synthase PDZ domain complexed with an associated peptide. Nat Struct Biol 6: 417-21.
339.
Jaffrey, SR, Snowman, AM, Eliasson, MJ, Cohen, NA and Snyder, SH, (1998) CAPON: a protein associated with neuronal nitric oxide synthase that regulates its interactions with PSD95. Neuron 20: 115-24.
340.
Brenman, JE, Chao, DS, Gee, SH, McGee, AW, Craven, SE, Santillano, DR et al., (1996) Interaction of nitric oxide synthase with the postsynaptic density protein PSD-95 and alpha1-syntrophin mediated by PDZ domains. Cell 84: 757-67.
341.
Hillier, BJ, Christopherson, KS, Prehoda, KE, Bredt, DS and Lim, WA, (1999) Unexpected modes of PDZ domain scaffolding revealed by structure of nNOSsyntrophin complex. Science 284: 812-5.
342.
Nehring, RB, Wischmeyer, E, Doring, F, Veh, RW, Sheng, M and Karschin, A, (2000) Neuronal inwardly rectifying K(+) channels differentially couple to PDZ proteins of the PSD-95/SAP90 family. J Neurosci 20: 156-62.
343.
Gee, SH, Quenneville, S, Lombardo, CR and Chabot, J, (2000) Single-amino acid substitutions alter the specificity and affinity of PDZ domains for their ligands. Biochemistry 39: 14638-46.
225
344.
Hogan, A, Shepherd, L, Chabot, J, Quenneville, S, Prescott, SM, Topham, MK et al., (2001) Interaction of gamma 1-syntrophin with diacylglycerol kinase-zeta. Regulation of nuclear localization by PDZ interactions. J Biol Chem 276: 2652633.
345.
Hu, LA, Tang, Y, Miller, WE, Cong, M, Lau, AG, Lefkowitz, RJ et al., (2000) beta 1-adrenergic receptor association with PSD-95. Inhibition of receptor internalization and facilitation of beta 1-adrenergic receptor interaction with Nmethyl-D-aspartate receptors. J Biol Chem 275: 38659-66.
346.
Braithwaite, SP, Xia, H and Malenka, RC, (2002) Differential roles for NSF and GRIP/ABP in AMPA receptor cycling. Proc Natl Acad Sci U S A 99: 7096-101.
347.
Daw, MI, Chittajallu, R, Bortolotto, ZA, Dev, KK, Duprat, F, Henley, JM et al., (2000) PDZ proteins interacting with C-terminal GluR2/3 are involved in a PKCdependent regulation of AMPA receptors at hippocampal synapses. Neuron 28: 873-86.
348.
Hirbec, H, Francis, JC, Lauri, SE, Braithwaite, SP, Coussen, F, Mulle, C et al., (2003) Rapid and differential regulation of AMPA and kainate receptors at hippocampal mossy fibre synapses by PICK1 and GRIP. Neuron 37: 625-38.
349.
Seidenman, KJ, Steinberg, JP, Huganir, R and Malinow, R, (2003) Glutamate receptor subunit 2 Serine 880 phosphorylation modulates synaptic transmission and mediates plasticity in CA1 pyramidal cells. J Neurosci 23: 9220-8.
350.
Matsuda, S, Mikawa, S and Hirai, H, (1999) Phosphorylation of serine-880 in GluR2 by protein kinase C prevents its C terminus from binding with glutamate receptor-interacting protein. J Neurochem 73: 1765-8.
351.
Yamazaki, M, Fukaya, M, Abe, M, Ikeno, K, Kakizaki, T, Watanabe, M et al., (2001) Differential palmitoylation of two mouse glutamate receptor interacting protein 1 forms with different N-terminal sequences. Neurosci Lett 304: 81-4.
352.
el-Husseini Ael, D and Bredt, DS, (2002) Protein palmitoylation: a regulator of neuronal development and function. Nat Rev Neurosci 3: 791-802.
353.
Bredt, DS and Nicoll, RA, (2003) AMPA receptor trafficking at excitatory synapses. Neuron 40: 361-79.
354.
Staudinger, J, Zhou, J, Burgess, R, Elledge, SJ and Olson, EN, (1995) PICK1: a perinuclear binding protein and substrate for protein kinase C isolated by the yeast two-hybrid system. J Cell Biol 128: 263-71.
226
355.
Xu, J and Xia, J, (2006) Structure and function of PICK1. Neurosignals 15: 190201.
356.
Cao, M, Xu, J, Shen, C, Kam, C, Huganir, RL and Xia, J, (2007) PICK1-ICA69 heteromeric BAR domain complex regulates synaptic targeting and surface expression of AMPA receptors. J Neurosci 27: 12945-56.
357.
Jin, W, Ge, WP, Xu, J, Cao, M, Peng, L, Yung, W et al., (2006) Lipid binding regulates synaptic targeting of PICK1, AMPA receptor trafficking, and synaptic plasticity. J Neurosci 26: 2380-90.
358.
Engel, DC, Slemmer, JE, Vlug, AS, Maas, AI and Weber, JT, (2005) Combined effects of mechanical and ischemic injury to cortical cells: secondary ischemia increases damage and decreases effects of neuroprotective agents. Neuropharmacology 49: 985-95.
359.
Dev, KK, Nakanishi, S and Henley, JM, (2004) The PDZ domain of PICK1 differentially accepts protein kinase C-alpha and GluR2 as interacting ligands. J Biol Chem 279: 41393-7.
360.
Perez, JL, Khatri, L, Chang, C, Srivastava, S, Osten, P and Ziff, EB, (2001) PICK1 targets activated protein kinase Calpha to AMPA receptor clusters in spines of hippocampal neurons and reduces surface levels of the AMPA-type glutamate receptor subunit 2. J Neurosci 21: 5417-28.
361.
Lin, DT and Huganir, RL, (2007) PICK1 and phosphorylation of the glutamate receptor 2 (GluR2) AMPA receptor subunit regulates GluR2 recycling after NMDA receptor-induced internalization. J Neurosci 27: 13903-8.
362.
Steinberg, JP, Takamiya, K, Shen, Y, Xia, J, Rubio, ME, Yu, S et al., (2006) Targeted in vivo mutations of the AMPA receptor subunit GluR2 and its interacting protein PICK1 eliminate cerebellar long-term depression. Neuron 49: 845-60.
363.
Hanley, JG and Henley, JM, (2005) PICK1 is a calcium-sensor for NMDAinduced AMPA receptor trafficking. Embo J 24: 3266-78.
364.
Kakegawa, W and Yuzaki, M, (2005) A mechanism underlying AMPA receptor trafficking during cerebellar long-term potentiation. Proc Natl Acad Sci U S A 102: 17846-51.
227
365.
Xia, J, Chung, HJ, Wihler, C, Huganir, RL and Linden, DJ, (2000) Cerebellar long-term depression requires PKC-regulated interactions between GluR2/3 and PDZ domain-containing proteins. Neuron 28: 499-510.
366.
Kim, CH, Chung, HJ, Lee, HK and Huganir, RL, (2001) Interaction of the AMPA receptor subunit GluR2/3 with PDZ domains regulates hippocampal long-term depression. Proc Natl Acad Sci U S A 98: 11725-30.
367.
Lu, W and Ziff, EB, (2005) PICK1 interacts with ABP/GRIP to regulate AMPA receptor trafficking. Neuron 47: 407-21.
368.
Wang, JQ, Arora, A, Yang, L, Parelkar, NK, Zhang, G, Liu, X et al., (2005) Phosphorylation of AMPA receptors: mechanisms and synaptic plasticity. Mol Neurobiol 32: 237-49.
369.
Chung, HJ, Steinberg, JP, Huganir, RL and Linden, DJ, (2003) Requirement of AMPA receptor GluR2 phosphorylation for cerebellar long-term depression. Science 300: 1751-5.
370.
McDonald, BJ, Chung, HJ and Huganir, RL, (2001) Identification of protein kinase C phosphorylation sites within the AMPA receptor GluR2 subunit. Neuropharmacology 41: 672-9.
371.
Chung, HJ, Xia, J, Scannevin, RH, Zhang, X and Huganir, RL, (2000) Phosphorylation of the AMPA receptor subunit GluR2 differentially regulates its interaction with PDZ domain-containing proteins. J Neurosci 20: 7258-67.
372.
Shin, J, Shen, F and Huguenard, J, (2007) PKC and polyamine modulation of GluR2-deficient AMPA receptors in immature neocortical pyramidal neurons of the rat. J Physiol 581: 679-91.
373.
Bell, JD, Park, E, Ai, J and Baker, AJ, (2009) PICK1-mediated GluR2 endocytosis contributes to cellular injury after neuronal trauma. Cell Death Differ. Advance online publication, July 31, 2009; doi:10.1038/cdd.2009.106.
374.
Osten, P, Khatri, L, Perez, JL, Kohr, G, Giese, G, Daly, C et al., (2000) Mutagenesis reveals a role for ABP/GRIP binding to GluR2 in synaptic surface accumulation of the AMPA receptor. Neuron 27: 313-25.
375.
Dixon, RM, Mellor, JR and Hanley, JG, (2009) PICK1-mediated glutamate receptor subunit 2 (GluR2) trafficking contributes to cell death in oxygen/glucosedeprived hippocampal neurons. J Biol Chem 284: 14230-5.
228
376.
Lu, Y, Allen, M, Halt, AR, Weisenhaus, M, Dallapiazza, RF, Hall, DD et al., (2007) Age-dependent requirement of AKAP150-anchored PKA and GluR2lacking AMPA receptors in LTP. Embo J 26: 4879-90.
377.
Ferguson, AR, Christensen, RN, Gensel, JC, Miller, BA, Sun, F, Beattie, EC et al., (2008) Cell death after spinal cord injury is exacerbated by rapid TNF alphainduced trafficking of GluR2-lacking AMPARs to the plasma membrane. J Neurosci 28: 11391-400.
378.
Faden, AI, Demediuk, P, Panter, SS and Vink, R, (1989) The role of excitatory amino acids and NMDA receptors in traumatic brain injury. Science 244: 798800.
379.
Yang, K, Taft, WC, Dixon, CE, Todaro, CA, Yu, RK and Hayes, RL, (1993) Alterations of protein kinase C in rat hippocampus following traumatic brain injury. J Neurotrauma 10: 287-95.
380.
Belayev, L, Alonso, OF, Liu, Y, Chappell, AS, Zhao, W, Ginsberg, MD et al., (2001) Talampanel, a novel noncompetitive AMPA antagonist, is neuroprotective after traumatic brain injury in rats. J Neurotrauma 18: 1031-8.
381.
Wang, KK, Larner, SF, Robinson, G and Hayes, RL, (2006) Neuroprotection targets after traumatic brain injury. Curr Opin Neurol 19: 514-9.
382.
Maas, AI, (2001) Neuroprotective agents in traumatic brain injury. Expert Opin Investig Drugs 10: 753-67.
383.
Kwak, S and Weiss, JH, (2006) Calcium-permeable AMPA channels in neurodegenerative disease and ischemia. Curr Opin Neurobiol 16: 281-7.
384.
Frederickson, CJ, Maret, W and Cuajungco, MP, (2004) Zinc and excitotoxic brain injury: a new model. Neuroscientist 10: 18-25.
385.
Hellmich, HL, Eidson, KA, Capra, BA, Garcia, JM, Boone, DR, Hawkins, BE et al., (2007) Injured Fluoro-Jade-positive hippocampal neurons contain high levels of zinc after traumatic brain injury. Brain Res 1127: 119-26.
386.
Sensi, SL and Jeng, JM, (2004) Rethinking the excitotoxic ionic milieu: the emerging role of Zn(2+) in ischemic neuronal injury. Curr Mol Med 4: 87-111.
387.
Suh, SW, Frederickson, CJ and Danscher, G, (2006) Neurotoxic zinc translocation into hippocampal neurons is inhibited by hypothermia and is
229
aggravated by hyperthermia after traumatic brain injury in rats. J Cereb Blood Flow Metab 26: 161-9. 388.
Agrawal, A, Timothy, J, Pandit, L and Manju, M, (2006) Post-traumatic epilepsy: an overview. Clin Neurol Neurosurg 108: 433-9.
389.
Bergsneider, M, Hovda, DA, Shalmon, E, Kelly, DF, Vespa, PM, Martin, NA et al., (1997) Cerebral hyperglycolysis following severe traumatic brain injury in humans: a positron emission tomography study. J Neurosurg 86: 241-51.
390.
Nortje, J and Menon, DK, (2004) Traumatic brain injury: physiology, mechanisms, and outcome. Curr Opin Neurol 17: 711-8.
391.
Rose, ME, Huerbin, MB, Melick, J, Marion, DW, Palmer, AM, Schiding, JK et al., (2002) Regulation of interstitial excitatory amino acid concentrations after cortical contusion injury. Brain Res 943: 15-22.
392.
Soriano, FX, Papadia, S, Hofmann, F, Hardingham, NR, Bading, H and Hardingham, GE, (2006) Preconditioning doses of NMDA promote neuroprotection by enhancing neuronal excitability. J Neurosci 26: 4509-18.
393.
McKinney, JS, Willoughby, KA, Liang, S and Ellis, EF, (1996) Stretch-induced injury of cultured neuronal, glial, and endothelial cells. Effect of polyethylene glycol-conjugated superoxide dismutase. Stroke 27: 934-40.
394.
Weber, JT, Rzigalinski, BA, Willoughby, KA, Moore, SF and Ellis, EF, (1999) Alterations in calcium-mediated signal transduction after traumatic injury of cortical neurons. Cell Calcium 26: 289-99.
395.
Zhao, X, Newcomb, JK, Pike, BR, Wang, KK, d'Avella, D and Hayes, RL, (2000) Novel characteristics of glutamate-induced cell death in primary septohippocampal cultures: relationship to calpain and caspase-3 protease activation. J Cereb Blood Flow Metab 20: 550-62.
396.
Slemmer, JE, Matser, EJ, De Zeeuw, CI and Weber, JT, (2002) Repeated mild injury causes cumulative damage to hippocampal cells. Brain 125: 2699-709.
397.
Bruno, VM, Goldberg, MP, Dugan, LL, Giffard, RG and Choi, DW, (1994) Neuroprotective effect of hypothermia in cortical cultures exposed to oxygenglucose deprivation or excitatory amino acids. J Neurochem 63: 1398-406.
230
398.
David, JC, Yamada, KA, Bagwe, MR and Goldberg, MP, (1996) AMPA receptor activation is rapidly toxic to cortical astrocytes when desensitization is blocked. J Neurosci 16: 200-9.
399.
Sattler, R, Charlton, MP, Hafner, M and Tymianski, M, (1997) Determination of the time course and extent of neurotoxicity at defined temperatures in cultured neurons using a modified multiwell plate fluorescence scanner. J Cereb Blood Flow Metab 17: 455-63.
400.
Geddes, DM, Cargill, RS, 2nd and LaPlaca, MC, (2003) Mechanical stretch to neurons results in a strain rate and magnitude-dependent increase in plasma membrane permeability. J Neurotrauma 20: 1039-49.
401.
Bartoletti, DC, Harrison, GI and Weaver, JC, (1989) The number of molecules taken up by electroporated cells: quantitative determination. FEBS Lett 256: 4-10.
402.
Gift, EA and Weaver, JC, (2000) Simultaneous quantitative determination of electroporative molecular uptake and subsequent cell survival using gel microdrops and flow cytometry. Cytometry 39: 243-9.
403.
Peterson, GL, (1977) A simplification of the protein assay method of Lowry et al. which is more generally applicable. Anal Biochem 83: 346-56.
404.
Hanley, JG, (2006) Molecular mechanisms for regulation of AMPAR trafficking by PICK1. Biochem Soc Trans 34: 931-5.
405.
Dev, KK, (2007) PDZ domain protein-protein interactions: a case study with PICK1. Curr Top Med Chem 7: 3-20.
406.
Elkins, JM, Papagrigoriou, E, Berridge, G, Yang, X, Phillips, C, Gileadi, C et al., (2007) Structure of PICK1 and other PDZ domains obtained with the help of selfbinding C-terminal extensions. Protein Sci 16: 683-94.
407.
Leitges, M, Kovac, J, Plomann, M and Linden, DJ, (2004) A unique PDZ ligand in PKCalpha confers induction of cerebellar long-term synaptic depression. Neuron 44: 585-94.
408.
Chang, CD and Meienhofer, J, (1978) Solid-phase peptide synthesis using mild base cleavage of N alpha-fluorenylmethyloxycarbonylamino acids, exemplified by a synthesis of dihydrosomatostatin. Int J Pept Protein Res 11: 246-9.
231
409.
Dixon, CE, Lyeth, BG, Povlishock, JT, Findling, RL, Hamm, RJ, Marmarou, A et al., (1987) A fluid percussion model of experimental brain injury in the rat. J Neurosurg 67: 110-9.
410.
Saglietti, L, Dequidt, C, Kamieniarz, K, Rousset, MC, Valnegri, P, Thoumine, O et al., (2007) Extracellular interactions between GluR2 and N-cadherin in spine regulation. Neuron 54: 461-77.
411.
Santos, AE, Duarte, CB, Iizuka, M, Barsoumian, EL, Ham, J, Lopes, MC et al., (2006) Excitotoxicity mediated by Ca2+-permeable GluR4-containing AMPA receptors involves the AP-1 transcription factor. Cell Death Differ 13: 652-60.
412.
Ahmed, SM, Weber, JT, Liang, S, Willoughby, KA, Sitterding, HA, Rzigalinski, BA et al., (2002) NMDA receptor activation contributes to a portion of the decreased mitochondrial membrane potential and elevated intracellular free calcium in strain-injured neurons. J Neurotrauma 19: 1619-29.
413.
BULLOCK, R, FUJISAWA, H, and LANDOLT, H., (1993) Glutamate-induced neurotoxicity and acute brain damage. Kewpski, O. ed. (W Zuchschwerdt Verlag, München).
414.
Knapp, LT and Klann, E, (2000) Superoxide-induced stimulation of protein kinase C via thiol modification and modulation of zinc content. J Biol Chem 275: 24136-45.
415.
Knapp, LT and Klann, E, (2002) Potentiation of hippocampal synaptic transmission by superoxide requires the oxidative activation of protein kinase C. J Neurosci 22: 674-83.
416.
Knapp, LT and Klann, E, (2002) Role of reactive oxygen species in hippocampal long-term potentiation: contributory or inhibitory? J Neurosci Res 70: 1-7.
417.
Nichols, B, (2003) Caveosomes and endocytosis of lipid rafts. J Cell Sci 116: 4707-14.
418.
Hou, Q, Huang, Y, Amato, S, Snyder, SH, Huganir, RL and Man, HY, (2008) Regulation of AMPA receptor localization in lipid rafts. Mol Cell Neurosci 38: 213-23.
419.
Hering, H, Lin, CC and Sheng, M, (2003) Lipid rafts in the maintenance of synapses, dendritic spines, and surface AMPA receptor stability. J Neurosci 23: 3262-71.
232
420.
Stellwagen, D and Malenka, RC, (2006) Synaptic scaling mediated by glial TNFalpha. Nature 440: 1054-9.
421.
Ogoshi, F, Yin, HZ, Kuppumbatti, Y, Song, B, Amindari, S and Weiss, JH, (2005) Tumor necrosis-factor-alpha (TNF-alpha) induces rapid insertion of Ca2+permeable alpha-amino-3-hydroxyl-5-methyl-4-isoxazole-propionate (AMPA)/kainate (Ca-A/K) channels in a subset of hippocampal pyramidal neurons. Exp Neurol 193: 384-93.
422.
De, A, Krueger, JM and Simasko, SM, (2003) Tumor necrosis factor alpha increases cytosolic calcium responses to AMPA and KCl in primary cultures of rat hippocampal neurons. Brain Res 981: 133-42.
423.
Hermann, GE, Rogers, RC, Bresnahan, JC and Beattie, MS, (2001) Tumor necrosis factor-alpha induces cFOS and strongly potentiates glutamate-mediated cell death in the rat spinal cord. Neurobiol Dis 8: 590-9.
424.
Kita, T, Tanaka, T, Tanaka, N and Kinoshita, Y, (2000) The role of tumor necrosis factor-alpha in diffuse axonal injury following fluid-percussive brain injury in rats. Int J Legal Med 113: 221-8.
425.
Knoblach, SM, Fan, L and Faden, AI, (1999) Early neuronal expression of tumor necrosis factor-alpha after experimental brain injury contributes to neurological impairment. J Neuroimmunol 95: 115-25.
426.
Kita, T, Liu, L, Tanaka, N and Kinoshita, Y, (1997) The expression of tumor necrosis factor-alpha in the rat brain after fluid percussive injury. Int J Legal Med 110: 305-11.
427.
Fan, L, Young, PR, Barone, FC, Feuerstein, GZ, Smith, DH and McIntosh, TK, (1996) Experimental brain injury induces differential expression of tumor necrosis factor-alpha mRNA in the CNS. Brain Res Mol Brain Res 36: 287-91.
428.
Barone, FC, Arvin, B, White, RF, Miller, A, Webb, CL, Willette, RN et al., (1997) Tumor necrosis factor-alpha. A mediator of focal ischemic brain injury. Stroke 28: 1233-44.
429.
Grossman, SD, Rosenberg, LJ and Wrathall, JR, (2001) Relationship of altered glutamate receptor subunit mRNA expression to acute cell loss after spinal cord contusion. Exp Neurol 168: 283-9.
233
430.
Grossman, SD, Wolfe, BB, Yasuda, RP and Wrathall, JR, (1999) Alterations in AMPA receptor subunit expression after experimental spinal cord contusion injury. J Neurosci 19: 5711-20.
431.
Brown, KM, Wrathall, JR, Yasuda, RP and Wolfe, BB, (2004) Glutamate receptor subunit expression after spinal cord injury in young rats. Brain Res Dev Brain Res 152: 61-8.
432.
Steiner, P, Alberi, S, Kulangara, K, Yersin, A, Sarria, JC, Regulier, E et al., (2005) Interactions between NEEP21, GRIP1 and GluR2 regulate sorting and recycling of the glutamate receptor subunit GluR2. Embo J 24: 2873-84.
433.
Serulle, Y, Zhang, S, Ninan, I, Puzzo, D, McCarthy, M, Khatri, L et al., (2007) A GluR1-cGKII Interaction Regulates AMPA Receptor Trafficking. Neuron 56: 670-88.
434.
Dai, JL, Wang, L, Sahin, AA, Broemeling, LD, Schutte, M and Pan, Y, (2004) NHERF (Na+/H+ exchanger regulatory factor) gene mutations in human breast cancer. Oncogene 23: 8681-7.
435.
Park, JS, Voitenko, N, Petralia, RS, Guan, X, Xu, JT, Steinberg, JP et al., (2009) Persistent inflammation induces GluR2 internalization via NMDA receptortriggered PKC activation in dorsal horn neurons. J Neurosci 29: 3206-19.
436.
Thorsen, TS, Madsen, KL, Rebola, N, Rathje, M, Anggono, V, Bach, A et al., (2009) Identification of a small-molecule inhibitor of the PICK1 PDZ domain that inhibits hippocampal LTP and LTD. Proc Natl Acad Sci U S A.
234
235
