Downloaded by Mcgill University 10/04/2015 14:12:50

CELLULAR MECHANISMS OF BRAIN-DERIVED NEUROTROPHIC FACTOR MEDIATED SYNAPSE REORGANIZATION FOLLOWING HIPPOCAMPAL INJURY

By, Raminder Gill, B. Sc.

Department of Pharmacology & Therapeutics

McGill University

Montréal, Québec, Canada

April 2015

A thesis submitted to McGill University in partial fulfillment of the requirements of the degree of Doctor of Philosophy

TABLE OF CONTENTS ...... i ABSTRACT ...... iv RÉSUMÉ ...... vii ACKNOWLEDGMENTS ...... x ABBREVIATIONS ...... xii PREFACE ...... xiv CONTRIBUTIONS TO ORIGINAL SCIENCE ...... xvi CHAPTER 1. INTRODUCTION ...... 1 1.1. GENERAL BACKGROUND ...... 1 1.2. OVERVIEW OF THE HIPPOCAMPUS ...... 2 1.3. NEURONAL MORPHOLOGY AND FUNCTIONALITY ...... 7 1.3.1. Dendritic Spine and Axon Development ...... 8 1.3.2. Excitatory Neurotransmission in CA1 Pyramidal Neurons ...... 10 1.3.3. Synaptic Plasticity of Excitatory Synapses ...... 13 1.3.4. Inhibitory Neurotransmission in CA1 Pyramidal Neurons ...... 16 1.3.5. Synaptic Plasticity of Inhibitory Synapses ...... 19 1.4. HIPPOCAMPUS AND INJURY ...... 23 1.4.1. Epilepsy ...... 23 1.4.2. Ischemia and Stroke ...... 27 1.4.3. Defects in Plasticity Following Brain Injury ...... 30 1.5. NEUROTROPHIC FACTORS AND THEIR RECEPTORS ...... 33 1.5.1. Overview of Neurotrophins ...... 33 1.5.2. Neurotrophin Receptors and Signaling Cascades ...... 35 1.5.2.1. Trk Receptor Signaling ...... 37 1.5.2.2. p75NTR Signaling ...... 38 1.6. BRAIN-DERIVED NEUROTROPHIC FACTOR ...... 41 1.6.1. Transcription, Secretion and Processing of BDNF ...... 41 1.6.2. Activity-Dependent Regulation of BDNF ...... 43 1.6.3. BDNF in Synapse Development...... 45 1.6.3.1. BDNF and Development of Excitatory Synapses ...... 45 1.6.3.2. BDNF and Development of Inhibitory Synapses ...... 47 1.6.3. BDNF and Glutamatergic Plasticity ...... 49 1.6.4.1. Role of BDNF in Excitatory Presynaptic Plasticity ...... 50 1.6.4.2. Role of BDNF in Excitatory Postsynaptic Plasticity ...... 50 1.6.5. BDNF and GABAergic plasticity ...... 51 1.6.6. BDNF and Brain Injuries ...... 53 1.7. ORGANOTYPIC HIPPOCAMPAL SLICE CULTURES ...... 57 CHAPTER 2. BDNF AND PROBDNF DIFFERENTIALLY REGULATE ISCHEMIA- INDUCED PLASTICITY OF GABAERGIC AND GLUTAMATERGIC SYNAPSES OF CA1 PYRAMIDAL NEURONS ...... 60 FOREWORD ...... 60 2.1. ABSTRACT ...... 61

2.2. INTRODUCTION ...... 62 2.3. MATERIALS AND METHODS...... 64 2.3.1. Ethics Statement ...... 64 2.3.2. Hippocampal Slice Cultures and Oxygen-Glucose Deprivation ...... 64 2.3.3. Propidium Iodide Staining, Immunofluorescence and Confocal Microscopy ...... 65 2.3.4. Dendrite Reconstructions, Spine Quantification and Puncta Quantification ...... 66 2.3.5. Electrophysiological Recordings and Analysis ...... 67 2.3.6. RT-qPCR ...... 69 2.3.7. Pharmacological Treatments ...... 69 2.3.8. Biolistic Gene Transfection ...... 70 2.3.9. Statistical Analysis ...... 71 2.4. RESULTS ...... 72 2.4.1. Area CA1 is selectively affected following OGD in organotypic hippocampal slices 72 2.4.2. OGD disrupts excitatory synapses morphologically and functionally ...... 74 2.4.3. OGD disrupts GABAergic inhibitory synapses morphologically and functionally .... 77 2.4.4. Expression of bdnf mRNA increases after OGD and TrkB-Fc treatment rescues OGD- induced excitatory synapse deficit ...... 84 2.4.5. TrkB-Fc treatment rescues OGD-induced inhibitory synapse deficit ...... 88 2.4.6. Blocking mBDNF prevents gephyrin downregulation but not dendritic spine loss .... 92 2.4.7. Preventing ERK1/2 activation and blocking GSK3β prevents gephyrin degradation, but not dendritic spine loss ...... 98 2.4.8. Blocking proBDNF and p75NTR prevents dendritic spine loss but not gephyrin loss 108 2.5. CONCLUSION ...... 112 2.5.1. OGD leads to loss of glutamatergic and GABAergic synapses in CA1 neurons and increases bdnf mRNA expression ...... 112 2.5.2. mBDNF disrupts GABAergic synapses after OGD ...... 114 2.5.3. proBDNF and p75NTR disrupt glutamatergic synapses after OGD ...... 116 CHAPTER 3. BLOCKING BDNF INHIBITS INJURY INDUCED HYPEREXCITABILITY OF HIPPOCAMPAL CA3 NEURONS ...... 119 FOREWORD ...... 119 3.1. ABSTRACT ...... 120 3.2. INTRODUCTION ...... 121 3.3. MATERIALS AND METHODS...... 123 3.3.1. Ethics Statement ...... 123 3.3.2. Hippocampal Slice Cultures and Schaffer Collateral Lesions...... 123 3.3.3. Visualization of CA3 Pyramidal Neurons ...... 124 3.3.4. Electrophysiological Recordings and Analysis ...... 124 3.3.5. Immunohistochemistry ...... 126 3.3.6. RT-qPCR ...... 127 3.3.7. Western Blotting ...... 128 3.3.8. Immunoprecipitation ...... 129 3.3.9. Pharmacological Treatments ...... 130 3.3.10 Statistical analysis...... 130 3.4. RESULTS ...... 131 3.4.1. Schaffer collateral transection in mouse organotypic slices leads to axonal sprouting and hyperexcitability 14 days following injury ...... 131

3.4.2. Intrinsic electrical properties of CA3 neurons remains unchanged after Schaffer collateral transection ...... 136 3.4.3. Hyperexcitability of CA3 neurons after Schaffer collateral transection can be attributed to increased excitatory inputs ...... 137 3.4.4. BDNF transcript and protein levels increase after Schaffer collateral lesion ...... 140 3.4.5. BDNF inhibition downregulates axonal sprouting following injury ...... 143 3.4.6. BDNF inhibition blocks the development of hyperexcitability 14 days post-lesion . 146 3.4.7. Chronic treatment with TrkB-Fc for 14 days after lesion inhibits lesion induced hyperexcitable network activity in area CA3 ...... 150 3.5. DISCUSSION ...... 155 3.5.1. Time course of BDNF expression following injury ...... 155 3.5.2. BDNF initiates axonal sprouting of CA3 pyramidal neurons after lesion ...... 156 3.5.3. Chronic TrkB-Fc treatment attenuates injury-induced hyperexcitability and network activity ...... 158 3.5.4. Trigger and target for BDNF release ...... 159 CHAPTER 4. INHIBITION OF METHYL CPG BINDING PROTEIN 2 PHOSPHORYLATION DOES NOT PREVENT INJURY-INDUCED HYPEREXCITABILITY ...... 161 FOREWORD ...... 161 4.1. ABSTRACT ...... 163 4.2. INTRODUCTION ...... 164 4.3. MATERIALS AND METHODS...... 167 4.3.1. Ethics Statement ...... 167 4.3.2. Hippocampal Slice Cultures and Schaffer Collateral Lesions...... 167 4.3.3. Immunohistochemistry ...... 168 4.3.4. Electrophysiological Recordings and Analysis ...... 169 4.3.5. Pharmacological Treatment ...... 170 4.3.6. Statistical Analysis ...... 170 4.4. RESULTS ...... 171 4.4.1. MeCP2 is phosphorylated at S421 in neurons and a subset of astrocytes as early as 1 hour following Schaffer collateral transection in area CA3 ...... 171 4.4.2. Selective pharmacological targeting of MeCP2 Ca2+-dependent phosphorylation ... 173 4.4.3. Inhibition of CaMKII does not prevent Schaffer collateral lesion-induced increase in firing of CA3 neurons ...... 176 4.5. DISCUSSION ...... 180 4.5.1. Role of pMeCP2S421 in BDNF-induced axonal remodeling ...... 180 4.5.2. Potential sources of BDNF release ...... 182 CHAPTER 5. DISCUSSION AND CONCLUSION ...... 185 5.1. SUMMARY...... 185 5.2. DISCUSSION AND FUTURE DIRECTIONS ...... 185 5.2.1. Pathological Role of BDNF in Microcircuitry Remodeling ...... 186 5.2.2. Source of BDNF Release ...... 190 5.3. CONCLUSION ...... 194 REFERENCES ...... 196 APPENDIX ...... 249

iii

ABSTRACT

Brain injury and neurological disorders can adversely impact the way that we communicate with the environment and therefore detrimentally affect quality of life for patients. Synapses, which are important neuronal structures that mediate communication between neurons, can become dysfunctional after brain injury. It is generally thought that synaptic dysfunction underlies the cognitive deficits that patients experience following brain injury and disease.

As such, synapses represent an interesting target for therapeutic intervention in order to limit the damage that brain insults have on cognition. In the case of post-traumatic epilepsy and ischemia, both excitatory and inhibitory synapses are remodelled, which can have devastating effects to existing functional neuronal networks. Though there are some theories on how trauma can lead to long-term functional deficits through neurocircuitry reorganization, there is still a paucity of information on the cellular mechanisms underlying synapse remodeling. In this thesis,

I studied the role of the neurotrophin, brain-derived neurotrophic factor (BDNF) in synaptic reorganization following hippocampal injury, a brain region which is important for learning and memory. BDNF plays a crucial role in development of both excitatory glutamatergic and inhibitory

GABAergic synapses. Interestingly, BDNF is highly upregulated after many different types of brain injury, including stroke and epilepsy. Some neuroscientists believe that this increase in

BDNF is an attempt by the brain to ameliorate injury, but may actually revert the central nervous system to a more juvenile and aberrant state thereby provoking further injury.

In my thesis I hypothesized that (1) BDNF can downregulate excitatory and inhibitory neurotransmission following ischemia, (2) BDNF mediates axonal reorganization and network hyperexcitability in a model of post-traumatic epilepsy and (3) BDNF-mediated axonal reorganization is due to a misappropriation of activity-dependent transcription of the Bdnf gene.

In order to test my hypotheses, I used organotypic hippocampal slice cultures and subjected them to two well-established in vitro models of hippocampal injury for long-term studies on neuronal networks: (1) oxygen-glucose deprivation, focusing on area CA1, the hippocampal region most susceptible to ischemia and (2) Schaffer collateral lesion, focusing on area CA3, the region where axon sprouting and hyperexcitability occurs in response to Schaffer collateral injury. I then combined confocal microscopy, immunofluorescence, molecular biology and electrophysiology to study synapse function, morphology and signaling.

I found that after ischemia to organotypic hippocampal slices, BDNF can downregulate

GABAergic synapses structurally and functionally through the high-affinity TrkB receptor.

Moreover, I found that proBDNF, the precursor protein of BDNF, can downregulate glutamatergic synapses structurally and functionally through the low-affinity p75NTR receptor. Accordingly, my findings identify distinct signaling cascades that specifically provoke acute excitatory or inhibitory synapse loss after ischemia. Therefore, these signaling cascades represent putative therapeutic targets for prevention of cognitive deficits following ischemic stroke.

I next wanted to determine if BDNF played a role in another type of hippocampal injury such as post-traumatic epilepsy. Using the Schaffer collateral transection model, I found that bdnf mRNA expression is upregulated shortly following a lesion and that scavenging BDNF with TrkB-

Fc prevented lesion-induced axonal remodeling and inhibited the formation of a recurrent network.

Given that axonal remodelling is a classic hallmark of post-traumatic epilepsy, my data identifies a specific therapeutic pathway that may prevent epileptogenesis in patients following traumatic brain injury.

Lastly, in order to better understand the source of this BDNF and also identify other therapeutic targets to prevent injury-induced synaptic reorganization, I tested the involvement of

v methyl CpG binding protein 2 (MeCP2) regulation of activity-dependent transcription of Bdnf on

CA3 pyramidal neuron hyperexcitability. I found that MeCP2 became phosphorylated at serine

421, a molecular switch for activating bdnf transcription, shortly following Schaffer collateral lesion. In addition, I found that this injury-induced pMeCP2 upregulation could be prevented by inhibiting Ca2+/Calmodulin kinase II (CaMKII). Interestingly, I found that inhibiting CaMKII did not prevent CA3 pyramidal neuron hyperexcitability, suggesting that Ca2+-dependent regulation of pMeCP2 does not underlie synaptic reorganization induced by BDNF.

Taken together, my results enhance our understanding of how BDNF-mediated synaptic plasticity can be misappropriated after hippocampal injury and that this underlies synaptic reorganization and dysfunction. In conclusion, my work provides a mechanistic basis for further study of BDNF signaling after acquired brain injuries in rodents and higher mammals in vivo.

Consequently, findings from my work may lead to the development of specific therapeutic targets that enhance cognitive recovery following brain injury.

RÉSUMÉ

Les lésions cérébrales et les troubles neurologiques peuvent altérer nos interactions avec l'environnement de sorte à réduire la qualité de vie des patients. Les synapses sont des structures neuronales importantes pour la communication entre neurones qui peuvent devenir dysfonctionnelles à la suite d'une lésion cérébrale. Cette dysfonction synaptique est généralement reconnue comme étant la cause des déficits cognitifs chez ces mêmes patients.

Les synapses représentent une cible thérapeutique intéressante afin de réduire les déficits cognitifs résultant des lésions cérébrales et des troubles neurologiques. Dans le cas de l’épilepsie post-traumatique et de l’ischémie, les synapses excitatrices et inhibitrices sont réorganisées, ce qui peut avoir des effets dévastateurs sur les réseaux neuronaux fonctionnels. Même si quelques théories permettent d'expliquer la façon dont un traumatisme peut induire à long terme des déficits fonctionnels grâce à une réorganisation neuronale, il y a encore un manque d’informations sur les mécanismes cellulaires sous-tendant cette réorganisation. Dans le cadre de cette thèse, j’ai étudié le rôle de la neurotrophine brain-derived neurotrophic factor (BDNF) au niveau de la réorganisation synaptique suite à une lésion de l’hippocampe, une région du cerveau qui est importante pour l’apprentissage et la mémoire. Le BDNF est une molécule cruciale pour le développement des synapses glutamatergiques et GABAergiques. L’expression de cette protéine est fortement augmentée suivant différents types de lésions cérébrales, y compris les accidents cérébro-vasculaires et l’épilepsie. Cette augmentation du BDNF pourrait être une tentative pour améliorer une blessure mais peut aussi détériorer le système nerveux central et provoquer des blessures plus graves dans le cas d'un cerveau plus juvénile et anormal.

Dans cette thèse, j’ai émis l’hypothèse que (1) le BDNF peut réguler négativement la neurotransmission excitatrice et inhibitrice suite à une ischémie, (2) le BDNF facilite la

vii réorganisation axonale et l’hyperexcitabilité du réseau neuronal dans un modèle d’épilepsie post- traumatique et (3) la réorganisation axonale facilitée par le BDNF est causé à un dérèglement de la transcription dépendante de l'activité du gène codant pour le bdnf.

Pour vérifier ces hypothèses, j’ai utilisé des cultures organotypiques d’hippocampe dans deux différents modèles reconnus pour l'étude à long terme des réseaux neuronaux suivant une ischémie ou une lésion: (1) la privation oxygène-glucose en ciblant la zone CA1, soit la région de l’hippocampe la plus sensible à l’ischémie; (2) la lésion collatérale de Schaffer en ciblant la zone

CA3, soit la région où la régénération d’axones et l’hyperexcitabilité se produisent en réponse à une atteinte. J’ai ensuite combiné la microscopie confocale, l’immunofluorescence, la biologie moléculaire et l’électrophysiologie pour étudier la fonction des synapses, leurs morphologies ainsi que leurs signalisations intracellulaires.

J'ai trouvé qu'en condition ischémique, le BDNF peut structurellement et fonctionnellement réguler à la baisse les synapses GABAergiques via le récepteur à haute affinité

TrkB. Par ailleurs, j’ai trouvé que la protéine précurseur du BDNF, le proBDNF, peut aussi réguler de façon structurelle et fonctionnelle les synapses glutamatergiques via le récepteur à basse affinité p75NTR. Par conséquent, mes résultats permettent d’identifier certaines cascades de signalisations qui provoquent la perte à court terme des synapses excitatrices ou inhibitrices suite à l’ischémie.

Ces mêmes cascades de signalisation représentent donc des cibles thérapeutiques possibles pour la prévention des déficits cognitifs après un évènement ischémique.

Par la suite, j'ai cherché à déterminer si le BDNF joue un rôle dans l’épilepsie post- traumatique. En utilisant toujours des cultures organotypique d’hippocampe dans le modèle de lésion collatérale de Schaffer, j’ai mis en évidence que l’expression d’ARNm de bdnf est augmentée deux heures après la lésion et que le blocage du BDNF par le TrkB-Fc empêche la

viii réorganisation axonale et inhibe la formation d’un réseau récurrent. Étant donné que le remodelage axonal est une caractéristique typique de l’épilepsie post-traumatique, mes résultats identifient une approche thérapeutique qui peut empêcher l’épileptogenèse chez les patients ayant subi une lésion cérébrale traumatique.

Finalement, j'ai tenté de mieux comprendre la source du BDNF de sorte à éventuellement identifier d’autres cibles thérapeutiques empêchant la réorganisation synaptique causé par une lésion cérébrale. J’ai évalué la régulation par le méthyl CpG protéine 2 (MeCP2) de la transcription dépendante de l'activité du gène codant pour le bdnf au niveau de l’hyperexcitabilité des neurones pyramidaux de la région CA3. J'ai trouvé qu'un résidu essentiel de MeCP2 pour l'activation de la transcription du gène codant pour le bdnf, la sérine 421, est phosphorylé immédiatement après la lésion induite dans le modèle collatérale Schaffer. Cette phosphorylation peut être prévenue par l’inhibition de la Ca2+/calmodulin kinase II (CaMKII). De plus, l’inhibition de CaMKII n’empêche pas l’hyperexcitabilité des neurones pyramidaux de la région CA3, ce qui suggère que la régulation de la phosphorylation de MeCP2 par Ca2+ ne sous-tend pas la réorganisation synaptique causée par le BDNF.

Dans l'ensemble, mes résultats permettent de mieux comprendre comment la plasticité synaptique facilitée par le BDNF, et plus particulièrement la réorganisation et le dysfonctionnement synaptique, peuvent être altérés suite à une lésion à l’hippocampe. En conclusion, mon travail apporte une base mécanistique nécessaire à une étude in vivo plus approfondie de la signalisation du BDNF suite aux lésions cérébrales chez les rongeurs et mammifères d’ordres supérieurs. Les résultats de mon travail pourront éventuellement mener au développement de nouvelles cibles thérapeutiques pour améliorer la récupération des fonctions cognitives après une lésion cérébrale.

ACKNOWLEDGMENTS

First and foremost I would like to thank my supervisor Dr. Anne McKinney for supporting me intellectually and emotionally throughout my PhD. These last years have been very formative for me and I’m grateful for that.

Secondly, I would like to thank past and present members of the McKinney lab who have helped me along the way (professionally, personally and for all the laughs) and making this journey so memorable, in no particular order: Dr. Philip Chang, Dr. David Verbich, Micaela Das Gupta,

Dr. George Prenosil, Saman Sizdahkhani, Fiorella Guido, Melanie Chan, and Andy Gao. But most especially Emily Deane, who made this process more fun and debaucherous than I could have imagined. In addition, I would like to thank François Charron for putting up with my requests and making all the organotypic cultures used in this thesis. Likewise, I would like to thank the other student’s in the Bellini, IPN program and department of Pharmacology & Therapeutics for all their help and scientific discussion: Dr. Bryan Daniels, Dr. Varin Gosein, Mark Arrousseau, Patricia

Brown, Dr. Alina Ilie, Dr. Ziv Machnes and I apologize for leaving out anyone I may have forgotten. I would like to thank the administrative staff in the Pharmacology & Therapeutics office,

Hélène Duplessis, Tina Tremblay, Marianne Casey and Chantal Grignon for their support. Also a big thank you to Dr. Terry Hébert for helping to translate the abstract of my thesis into French.

I would also like to take this time to thank collaborators who have helped in the process of this PhD: Dr. Shiva Tyagarajan, Dr. Jean-Marc Fritschy and Zahra Thirouin. Additionally, many thanks are needed to the funding agencies who have given us the means to carry out this work:

Canadian Institute of Health Research (CIHR) and the Savoy Epilepsy Foundation.

I would like to thank my thesis committee members for their scientific guidance: Dr. Derek

Bowie, Dr. Graziella Di Cristo (Université de Montreal), Dr. Radan Capek and Dr. Philip Barker.

I would also like to thank Prof. Linda Cooper whose courses drastically improved my writing skills, which infinitely helped in writing this thesis.

I must thank my loving partner J.P., who has been my rock through this writing process and that is true for the entire Palerme family. I would also like to thank all my friends for their support throughout my PhD studies.

Last but most important, I would like to thank my family for putting up with me the past few years! It’s been a long journey and I could never have completed my PhD without them, in particular my brother and sister-in-law who have supported me throughout this endeavour. But most especially, I dedicate this thesis to my parents who have given me complete and unwavering support. In so many uncountable ways, they have been there for me throughout this whole process and it would not have been possible to complete this PhD without them. I love you so much and thank you for never giving up on me!

ABBREVIATIONS

α1-GABAAR subunit α1 of the GABAAR

α2-GABAAR subunit α2 of the GABAAR AMPA α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid BDNF brain-derived neurotrophic factor CA1 cornu ammonis area 1 CA3 cornu ammonis area 3 Ca2+ calcium ions CaMKII Ca2+/Calmodulin-dependent protein kinase II cAMP cyclic adenosine monophosphate Cl- chloride ions DNA deoxyribonucleic acid DPL days post-lesion EPSP excitatory post-synaptic potential ERK1/2 extracellular-signal-regulated kinases 1/2 fEPSP field EPSP GABA γ-aminobutyric acid

GABAAR GABA receptor type A; ionotropic

GABAB GABA receptors type B; metabotropic GAD65 glutamate decarboxylase 65 GAD67 glutamate decarboxylase 67 GAP43 growth-associated protein 43 GFAP glial fibrillary acidic protein Glu glutamate GFP green fluorescent protein GSK3β glycogen synthase kinase 3 β HIF1α hypoxia-inducible factor 1 α HPL hours post-lesion IEI inter-event interval KCC2 neuron-specific K+-Cl- cotransporter LTD long-term depression LTP long-term potentiation MAGE melanoma-associated antigen mBDNF mature BDNF Mg2+ magnesium ions

xii

MeCP2 methyl-CpG binding protein 2 mEPSC miniature excitatory post-synaptic current mIPSC miniature inhibitory post-synaptic current mGluR metabotropic glutamate receptor MMP matrix metalloproteinases mRNA messenger ribonucleic acid Na+ sodium ions NGF nerve growth factor NMDA N-methyl-D-aspartate NRAGE neurotrophin receptor p75 interacting MAGE homologue NT-3 neurotrophin-3 NT-4/5 neurotrophin-4/5 OGD oxygen-glucose deprivation p75NTR neurotrophin receptor p75

P2X4 purinergic type 2 ligand-gated ion channel subunit 4 PI propidium iodide PI3K phosphoinositol-3 kinase PKA protein kinase A PKC protein kinase C PLC phospholipase C PSD95 postsynaptic density protein 95 PTE post-traumatic epilepsy pMeCP2 phosphorylated MeCP2 pTrkB phosphorylated TrkB RT-qPCR reverse transcription quantitative polymerase chain reaction sEPSC spontaneous EPSC TrkA tropomyosin-related kinase A TrkB tropomyosin-related kinase B TrkC tropomyosin-related kinase C VGAT vesicular GABA transporter

xiii

PREFACE

This thesis is written in manuscript form in compliance with thesis preparation guidelines of

Graduate and Postdoctoral Studies of McGill University and contains three manuscripts. Chapters

2, 3, and 4 of my thesis are modified versions of the following three manuscripts:

1. Gill, R., Thirouin, Z., Chang, P.-K., Tyagarajan, S., and McKinney, R.A. (2015) BDNF and proBDNF differentially regulate ischemia-induced plasticity of GABAergic and glutamatergic synapses of CA1 pyramidal neurons. Nat Neurosci. (in submission).  R. G., S. T. and R. A. M. designed experiments; R. G. performed all experiments except electrophysiology which were carried out by Z. T. and P. K.-Y. C.; R. G. performed all data analysis, preparation of figures and manuscript.

2. Gill, R., Chang, P.-K., Prenosil, G.A., Deane, E.C., and McKinney, R.A. (2013) blocking BDNF inhibits injury induced hyperexcitability of hippocampal CA3 neurons. European Journal of Neuroscience. 38(11):3554  R. G. and R. A. M. designed experiments; R. G. performed all experiments except electrophysiology which was carried out by G. A. P. and P. K.-Y. C.; R. G. performed all data analysis, preparation of figures and manuscript.

3. Gill, R., Chang, P.-K., and McKinney, R.A. (2015) blocking MeCP2-mediated activity- dependent transcription of bdnf does not prevent hyperexcitability of hippocampal CA3 neurons. (in preparation).  R. G. and R. A. M. designed experiments; R. G. performed all experiments except electrophysiology which was carried out by P. K.-Y. C.; R. G. performed all data analysis, preparation of figures and manuscript.

xiv

I have not included the following manuscripts that I have co-authored in the body of my thesis because although it pertains to hippocampal injury, synapse remodeling and related methods, they do not directly relate to the main focus of my thesis. Copies of these manuscripts appear in the appendix:

4. Chang, P.-K., Prenosil, G.A., Verbich, D., Gill, R., and McKinney, R.A. (2014) Prolonged ampakine exposure prunes dendritic spines and increases presynaptic release probability for enhanced long-term potentiation in the hippocampus. European Journal of Neuroscience. 40(5):2766

5. Machnes, Z.M., Huang, T.C., Chang, P.-K., Gill, R., Reist, N., Dezsi, G., Ozturk, E., Charron, F., O’Brien, T.J., Jones, N.C., McKinney, R.A., and Szyf, M. (2013) DNA methylation mediates persistent epileptiform activity in vitro and in vivo. PLoS One. 8(10):e76299

6. Queval, A., Ghattamaneni, N.R., Perrault, C.M., Gill, R., Mirzaei, M., McKinney, R.A., and Juncker, D. (2010) Chamber and microfluidic probe for microperfusion of organotypic brain slices. Lab on a Chip. 10(3):326

CONTRIBUTIONS TO ORIGINAL SCIENCE

The findings that I have outlined in this thesis show that hippocampal injury can activate signalling cascades that eventually lead to BDNF-dependent remodeling of microcircuitry and this can be detrimental to neuronal function. In Chapter 2, I investigate the potential role of BDNF and proBDNF in ischemia-induced synapse loss. I found that after ischemia, BDNF can downregulate

GABAergic synapses structurally and functionally and that this occurred through the high-affinity

TrkB receptor. In contrast, I found that proBDNF can downregulate glutamatergic synapses structurally and functionally and that this occurred through the low-affinity p75NTR receptor. This bifurcation in BDNF signaling on specific types of synapses is very surprising as it demonstrates for the first time that in a pathophysiological context TrkB and p75NTR can act in concert to the detriment of CA1 pyramidal neurons, which are very important principal cells involved in hippocampal learning and memory. Therefore, this data adds an important piece to our understanding of:

(1) Synapse loss in ischemia, (2) The role of BDNF in pathophysiology of ischemia (3) Bifurcated signaling of BDNF and proBDNF in neuronal network reorganization

This work was carried out in collaboration with Dr. Shiva Tyagarajan from the University of Zurich. Dr. McKinney, Dr. Tyagarajan and I designed the experiments. I carried out all oxygen- glucose deprivation and drug treatments for: immunofluorescence, RT-qPCR and electrophysiology. I carried out all molecular biology, confocal imaging and quantification of synapses (dendritic spines and GABAergic synapses). Z. Thirouin and Dr. P. K-Y. Chang carried out electrophysiology experiments, though I analyzed all traces. I have written the manuscript and generated all figures. Dr. McKinney edited the manuscript.

xvi

In Chapter 3, I investigate the role of BDNF in injury-induced axon remodeling and epileptiform activity. I found that after Schaffer collateral injury, mimicking a penetrating head wound, the expression of bdnf mRNA increases as early as 2 hours after injury, similar to findings in patients with traumatic brain injury. Moreover, inhibiting BDNF, and other TrkB ligands, I could prevent axon sprouting and subsequent network hyperexcitability. This data adds to a growing body of literature demonstrating that BDNF-TrkB signaling is activated following axonal injury and contributes to the malformation of excitatory autosynapses, leading to network hyperexcitability. My findings add an important piece to our understanding of:

(1) Injury-induced increase in BDNF (2) That pharmacological inhibition of TrkB ligands can prevent axon sprouting and network reorganization (3) How BDNF-mediated axonal sprouting may contribute to the development of post- traumatic epilepsy

In contribution to this data, Dr. McKinney, E. Deane and I designed the experiments. I carried out all lesions and drug treatments for immunofluorescence, RT-qPCR and electrophysiology. I carried out all molecular biology, confocal imaging and quantification of axon sprouting. Dr. P. K.-Y. Chang and Dr. G. A. Prenosil performed electrophysiology experiments, though I analyzed all traces. I have written the manuscript and generated all figures. Dr. McKinney edited the manuscript.

In Chapter 4, I investigated the role of activity-dependent of transcription Bdnf as a means to develop a specific therapeutic target to prevent axon sprouting and hyperexcitability. The transcriptional repressor MeCP2 is known to regulate activity-dependent transcription of Bdnf.

Here I demonstrated that mediators of Ca2+ signaling (CaMKK, CaMKII, NMDA receptors and voltage-gated Ca2+ channels) can all deactivate MeCP2 through serine 421 phosphorylation. I

xvii further demonstrated that inhibiting this deactivation does not prevent CA3 pyramidal neuron hyperexcitability. My findings contribute to our understanding that:

(1) MeCP2-dependent transcription likely does not affect axon remodeling and hyperexcitability (2) Other mechanisms of BDNF transcription or release likely underlie axonal sprouting

In contribution to this data, Dr. McKinney and I designed the experiments. I carried out all lesions and drug treatments for immunofluorescence and electrophysiology. I carried out all molecular biology, confocal imaging and quantification of phosphorylation of MeCP2. Dr. P. K.-

Y. Chang performed electrophysiology experiments, though I analyzed all traces. I have written the manuscript and generated all figures. Dr. McKinney edited the manuscript.

In addition to the work I have presented in my thesis, I have included the following 3 manuscripts in which I am a co-author in the appendix:

(1) The first is titled, “DNA methylation mediates persistent epileptiform activity in vitro and

in vivo”, published in PLoS One, where we demonstrated that epileptiform activity results

in long-lasting changes in DNA methylation in cultured hippocampal slices and in an

animal model of epilepsy. Furthermore, we demonstrated that inhibiting changes in DNA

methylation can prevent epileptogenesis and thus identified DNA methylation as a novel

therapeutic target to prevent epilepsy.

(2) The second article to which I have contributed is entitled “Prolonged ampakine exposure

prunes dendritic spines and increases presynaptic release probability for enhanced long-

term potentiation in the hippocampus”, published in the European Journal of

Neuroscience. Here, we demonstrated that prolonged treatment with the ampakine CX546,

a potential treatment for Alzheimer’s, leads to loss of dendritic spines on CA1 pyramidal

xviii

neurons. We found that this dendritic spine loss was mediated by BDNF and that this

BDNF-mediated dendritic spine loss primed CA1 synapses for long-term potentiation, a

cellular model of learning and memory.

(3) Lastly, I have contributed to the development of a microfluidic probe for organotypic

hippocampal slice cultures, modified for live imaging on an upright confocal microscope,

this probe was described in the article entitled “Chamber and microfluidic probe for

microperfusion of organotypic brain slices” published in Lab on a Chip.

xix

CHAPTER 1. INTRODUCTION

1.1. GENERAL BACKGROUND

Despite its utmost importance in our ability to interact with our environment, the brain remains one of the least understood organs in the human body. Given this, even the smallest brain trauma can have a large effect on our daily lives. Many theories have been put forward as to how trauma can lead to long-term functional deficits through neurocircuitry reorganization. In spite of this, there is still a lot that we do not know about how this occurs. In my thesis, I studied the effect of certain proteins in disrupting neurocircuitry, as a way to potentially enhance or encourage recovery.

Neurons mainly use electrochemical neurotransmission as a way to communicate with each other through functional structures known as synapses. For many years it was thought that once synapses developed they became static. However, over the course of the last 60 years neuroscientists have established that functional circuits between neurons are dynamic. This

‘plasticity’ of neurons can be seen on two levels: (1) synaptic morphology and (2) on a functional level (Engert & Bonhoeffer, 1999; Maletic-Savatic et al., 1999; De Roo et al., 2008). With recent advances in technology, we have also vastly improved our capabilities for high-resolution imaging of ultrastructures and improved our capability to record electrical activity from neurons, allowing us to better understand their plasticity.

Given the importance of synapses in neuronal function, there has been intensive study on the development, stability and disruption of synaptic structures using animal models, so we can better understand how the brain functions and how it may be disrupted by injury. The

1 hippocampus, a brain region important for learning and memory, has played a crucial role in our understanding of synapses. It is one the best studied brain regions and there are many injuries and disorders that can affect the hippocampus, for example: epilepsy, stroke and Alzheimer’s disease.

Therefore, I have chosen to study hippocampal injury for my thesis.

In order to understand hippocampal injury, I will first discuss the gross and cellular morphology of the hippocampus, as well as functionality of principal neurons in the hippocampus under physiological conditions. Subsequently, I will discuss how hippocampal synapses and microcircuitry can remodel under pathological conditions.

1.2. OVERVIEW OF THE HIPPOCAMPUS

The hippocampal formation is found in the medial temporal lobe of the forebrain. It has a fascinating structure, such that its neurons are organized in a very unique network. Given its uniqueness and its importance in learning and memory it has attracted the attention of many neuroscientists throughout history.

For many years anatomists were intrigued by the hippocampus but did not know what function it served in cognition. The importance of the hippocampus to learning and memory was not characterized until the 1950s at McGill University, when neurosurgeon Dr. Wilder Penfield noticed that unexpectedly two of his patients developed severe anterograde amnesia after having undergone unilateral hippocampal resections to treat epilepsy. Dr. Brenda Milner, who had observed the deficits in Dr. Penfield’s patient, further studied a patient of Dr. William Scoville who had performed an experimental bilateral hippocampal resection on Henry Gustave Molaison

(Patient H. M.) to treat his intractable epilepsy. Dr. Scoville and Dr. Milner found that remarkably

Patient H. M. developed even more severe anterograde amnesia than the previous patients that Dr.

Milner had seen (Scoville & Milner, 1957), indicating that the hippocampus was an important structure for memory formation. Underlying this memory formation is a unique network of neurons.

Camillo Golgi, a forefather of contemporary neuroscience, first published detailed descriptions of the hippocampal formation and organization of neuronal cells (Golgi et al., 2001).

In his seminal paper, he developed and used a technique to stain cells in the brain which highlighted the unique gross morphology of the hippocampus (Fig. 1.1A) and nicely demonstrated the different regions of the hippocampal formation.

Building on this work, Ramon y Cajal, another forefather of contemporary neuroscience, discovered the existence and importance of pyramidal neurons (Elston, 2003; Garcia-Lopez et al.,

2006), a specific excitatory principal neuron that releases glutamate as its neurotransmitter.

Pyramidal neurons are found only in the cortex, hippocampus and the amygdala. Within the hippocampus, pyramidal neurons are found in cornu ammonis (CA) areas 1-4. Hippocampal pyramidal neuron cell bodies are neatly aligned within stratum pyramidale (Fig. 1.2). This is the site of the majority of a neuron’s cellular metabolism and protein synthesis, though these activities can also occur in mature dendrites (Ostroff et al., 2002) and in axons during development (Kleiman et al., 1990; Bassell et al., 1998). One axon and several primary dendritic branches shoot off from the CA1-4 pyramidal cell bodies, one primary dendrite projects apically towards stratum radiatum and two project basally towards stratum oriens (Fig. 1.2). These dendrites receive inputs from different locations from within or outside the hippocampus (Fig. 1.2) (Andersen, 2007).

The gross morphology of the hippocampus demonstrates the mostly unidirectional nature of the synaptic connections (Amaral & Lavanex, 2007). This is outlined in Fig 1.1B, which

Figure 1.1 – The hippocampal formation. (A) Hand drawing of the hippocampus from Camillo

Golgi in 1884 and (b) rendering of the hippocampus highlighting some of the important axonal pathways in the hippocampus. Mossy fibers emanating from dentate gyrus (DG; orange) granule cells form synapses on area CA3 (blue) pyramidal neurons, which when send Schaffer collateral axons onto area CA1 (green) pyramidal neurons. Figure modified from (Golgi et al., 2001; Neves et al., 2008)

Figure 1.2 – CA1 pyramidal neuron. Maximum intensity project obtained from a confocal image of CA1 pyramidal cell expressing membrane-tagged GFP from a mouse hippocampal slice culture.

CA1 cell bodies line up within stratum pyramidale, while dendrites extend into stratum radiatum and stratum oriens. Synapses found on dendrites stratum radiatum and stratum oriens are innervated by Schaffer collaterals from CA3 pyramidal neurons. Distal apical dendrites in stratum lacunosum moleculare are innervated by axons from the perforant pathway (P. K.-Y. Chang & R.

A. McKinney, unpublished data).

5 demonstrates that axons from neurons in the entorhinal cortex innervate dentate gyrus granule cells, this is known as the perforant pathway, which in turn will project axons known as mossy fibers onto CA3 pyramidal neurons. CA3 pyramidal neuronal axons will form postsynaptic connections with CA1 pyramidal neurons (Fig. 1.1B), this axonal bundle is known as the Schaffer collateral pathway, named after Karoly Schaffer, the scientist who first described them (Schaffer,

1892; Szirmai et al., 2012). Additionally, in vivo it has been demonstrated that some CA3-CA3 collaterals do form (MacVicar & Dudek, 1980a; b; Miles & Wong, 1986; Amaral & Witter, 1989), but they represent only a small proportion of CA3 axonal connections. Historically, the CA3-CA1 synapse formed by the Schaffer collateral pathway is one of the most well-studied synapses in the central nervous system (Yuste, 2011). CA1 neurons send almost all their axons out of the hippocampus into the subiculum and onto the entorhinal cortex (Fig 1.1B) (Spruston & McBain,

2007), though there are a very small minority of CA1-CA1 synapse forming collaterals (Radpour

& Thomson, 1991; Deuchars & Thomson, 1996).

The hippocampal network is maintained in balance with inhibitory neurotransmission coming from inhibitory interneurons, which are locally projecting neurons that release the neurotransmitter γ-aminobutyric acid (GABA). Inhibitory interneurons are important for regulating complex interactions among excitatory principal cells, including a crucial role in maintaining population oscillations. There are 16 different subtypes of interneurons in the hippocampus alone (Parra et al., 1998). Each interneuron subtype contains distinct neuropeptides, which may function as cotransmitters, and they often contain a variety of calcium (Ca2+) binding proteins (Parra et al., 1998). These neuropeptides and Ca2+-binding proteins are used as markers for the different subtypes. Inhibitory synapses can form on CA1 pyramidal neuron cell bodies and the axon initial segment (Bourne & Harris, 2011; Harris & Weinberg, 2012), however roughly

90% of GABAergic synapses on CA1 pyramidal neurons are found on dendritic shafts (Megias et al., 2001; Klausberger, 2009). Basket cells are the best studied interneurons; they are characterized by highly ramified axonal arbors and distinctively high expression of parvalbumin (Kawaguchi et al., 1987; Katsumaru et al., 1988), a Ca2+-binding protein. Typically basket cell dendrites are aspiny or spotted with few dendritic spines, which are cellular compartments that contain the majority of excitatory postsynaptic sites (Freund & Buzsaki, 1996). Basket cells can innervate anywhere from 1500 to 2000 pyramidal neurons (Sik et al., 1995) and can be innervated by up to

2000 excitatory inputs, therefore the degree of convergence at basket cells is enormous (Amaral

& Lavanex, 2007).

It is important to note that there are also many different non-neuronal cells in the central nervous system such as astrocytes and microglia, both of which can play an important role in synaptic remodeling in the hippocampus (Wake et al., 2013; Haydon & Nedergaard, 2014).

However, in this thesis I do not concentrate on these cell types and therefore I will not discuss them further. In the next section, I will explain the morphology and functionality of excitatory and inhibitory synapses in the hippocampus.

1.3. NEURONAL MORPHOLOGY AND FUNCTIONALITY

Dendrites and axons of neurons are distinct molecular and functional compartments.

Dendrites integrate synaptic inputs in order to trigger the generation of action potentials at the cell body level. These action potentials will then propagate along the axon of this neuron and will pass this information forward through presynaptic contacts on the target cell.

In 1959, during his morphological studies with electron microscopy, Dr. E. George Gray found that synapses on pyramidal neurons in the forebrain could be categorized into two types: (1)

Type I or asymmetric: which were more dense in the postsynaptic compartment compared to the presynaptic compartment and were almost always associated with a postsynaptic dendritic spine and (2) Type II or symmetric: which had equivalently dense pre- and postsynaptic compartments and were typically found on the cell soma or directly on the shaft of the dendrite (Gray, 1959a; b).

Work from Per Andersen and others demonstrated that Type II synapses were inhibitory (Andersen et al., 1964b; a). Inhibitory neurotransmission and plasticity within CA1 pyramidal neurons will be further discussed in sections 1.3.4. and 1.3.5. Andersen’s laboratory also demonstrated that

Type I synapses were excitatory (Andersen et al., 1966a; Andersen et al., 1966b). These synapses were formed on dendritic spines and were glutamatergic in nature. I will first discuss how these synapses develop and then discuss their function and plasticity in the following section.

1.3.1. Dendritic Spine and Axon Development

In the hippocampus, stem cells of pyramidal neurons originate from the ventricular germinal layers and migrate to stratum pyramidale (Altman & Bayer, 1990). During migration pyramidal neurons form a leading process and trailing process, each becoming the axon or dendrite

(Polleux & Snider, 2010). However, there is little known about which specific hippocampal factors guide axons and dendrites to their targets (Yu & Bargmann, 2001; Martinez & Soriano, 2005).

Though, this process requires the coordination of multiple cytoskeletal molecules such as actin and microtubules, some of which is coordinated via the Ras and Rho family of small GTPases

(Hall & Lalli, 2010). Another such protein is growth associated protein of 43 kDa (GAP43), which has been shown to be enriched in axon growth cones during development (Meiri et al., 1986; Van

Lookeren Campagne et al., 1989; Meiri & Gordon-Weeks, 1990; Van Lookeren Campagne et al.,

1990; Maier et al., 1999) and can be reactivated during high levels of activity in the mature hippocampus (Son et al., 1997; Ramakers et al., 1999).

Synapses begin to develop and mature in the first postnatal week in rodents (Harris, 1999).

Two different models have been put forward as to how axons and dendrites come to together to form synapses. One theory postulates that as axons cross dendrites in the neuropil, axons attract filopodial extensions from the dendrite, which are spine-like structures that do not have a distinct head, or that filopodia spontaneously protrude from the dendrite at an axon-dendrite crossing

(Jontes & Smith, 2000; Yuste & Bonhoeffer, 2004). However, there is no direct evidence demonstrating that axons can ‘pull’ or attract filopodia, which would then mature into dendritic spines. Alternatively, a second theory hypothesizes that during development dendrites randomly send out filopodia into the neuropil and some of which may come in contact with a presynaptic terminal (Papa et al., 1995; Dailey & Smith, 1996; Ziv & Smith, 1996; Okabe et al., 2001;

Lohmann & Bonhoeffer, 2008). It is thought that these filopodial extensions have the potential to form stable synapses, this likely occurs when the filopodia comes in contact with an active presynaptic zone (Ziv & Smith, 1996) that triggers local Ca2+ transients within the filopodia

(Lohmann et al., 2005; Lohmann & Bonhoeffer, 2008). There are many molecules that have been identified in synapse formation and development, including the neurotrophin brain-derived neurotrophic factor (BDNF), who’s role in synapse formation will be discussed in further detail in section 1.6.3.

1.3.2. Excitatory Neurotransmission in CA1 Pyramidal Neurons

In adults, once dendritic spines form they become stable structures, though their morphology can be quite plastic. Dendritic spines are heterogeneous in shape (Fig. 1.3B) and can be categorized into morphological subgroups based on the size of their head and neck. Within the spine, the active zone of the synapse and the dendrite is separated by 1-3 μm. This allows the spine to act as an autonomous biochemical and electrical compartment (Harris & Stevens, 1989; Harris et al., 1992; Svoboda et al., 1996; Araya et al., 2006; Tonnesen et al., 2014), having its own regulated membrane-trafficking events that can shuttle components into and out of the spine membrane (Richards et al., 2004; Calabrese et al., 2006; Hugel et al., 2009).

Dendritic spines are thought to be the morphological correlate of long-term memory in the brain and the strength of their synaptic connections are fairly plastic (Segal, 2005). This plasticity occurs through a highly dynamic mechanism, which is actin-dependent (McKinney, 2005).

Electron microscopy studies demonstrated that the ultrastructure of dendritic spines is characterized by a thick electron-dense band present at the head of the spine (Gray, 1959b), which directly apposes presynaptic boutons containing neurotransmitter vesicles. This electron-dense region consists of neurotransmitter receptors, scaffolding molecules and signaling molecules

(Sheng, 2001), all essential for the function of the synapse. For instance, post-synaptic density protein of 95 kDa (PSD95) (Cho et al., 1992; Kistner et al., 1993) is found in this electron dense region. PSD95 is an important scaffolding molecule in excitatory synapses and helps cluster ionotropic glutamate receptors at the postsynaptic density, including: (1) N-methyl-D-aspartate or

NMDA-type, (2) α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid or AMPA-type and (3) kainate glutamate receptors (Kim & Sheng, 2004). NMDA receptors are composed of a heteromeric tetramer consisting of subunits GluN1,2A-D,3A-B and are permeable to sodium ions

Figure 1.3 – Organotypic slice preparation from a mouse expressing membrane-tagged GFP under the Thy1 promoter. (A) (i) Mature organotypic slice (>21 days in vitro). Note the presence of a subset of GFP-positive CA1 cells; CA3 pyramidal neurons were filled with AlexaFluor 594.

(ii) CA1 pyramidal neuron overview, repeated from Fig. 1.2 for illustration purposes. (iii) Portion of tertiary basal dendrite from a CA1 GFP-expressing pyramidal neuron. (B) (i) 3D volume rendered portion of tertiary dendrite. Dendritic spines are classified into three main types: (ii) Short stubby (iii) mushroom type spines, consisting of a short neck and mushroom shaped head and (iv) thin spines, which have elongated necks with small, pin-shaped heads. Modified from (McKinney,

2005; Hugel et al., 2009).

(Na+) and Ca2+ (Paoletti et al., 2013), in addition it has a magnesium ion (Mg2+) block at resting membrane potentials (Mayer et al., 1984; Nowak et al., 1984). AMPA receptors are dimers of dimers consisting of subunits GluA1-4 and their ion pore allows the influx of Na+ and Ca2+ ions

(Hollmann & Heinemann, 1994), depending on the subunit composition. In the adult hippocampus, the vast majority of AMPA receptors are impermeable to Ca2+ due to the presence of GluA2 subunits that do not allow Ca2+ influx (Bowie, 2012). AMPA receptors are important in the maintenance of dendritic spines (McKinney et al., 1999a).

During neurotransmission, in response to incoming electrical activity, synaptic vesicles in the presynaptic neuron will release glutamate into the synaptic cleft. In turn, this glutamate binds to AMPA receptors and causes channel opening leading to Na+ influx which will depolarize the postsynaptic membrane. Once the membrane is sufficiently depolarized, the Mg2+ block of NMDA receptors is relieved allowing for the entry of Ca2+ into the postsynaptic cell (Nicoll et al., 1988;

Yuste et al., 1999). This increase in Ca2+ is important for signal transduction and synaptic plasticity

(Yuste et al., 1999). For example, Ca2+/calmodulin-dependent protein kinase II (CaMKII), a kinase which is linked to the postsynapse by PSD95 (Suzuki et al., 2008), becomes activated by an increase in intracellular Ca2+ and is important for the induction of a cellular model for learning and memory, known as long-term potentiation (LTP) (Wayman et al., 2008). It is important to note that a vast number of proteins have been implicated in LTP (Sanes & Lichtman, 1999) and a wide range of mechanisms have emerged at different synapses across the mammalian brain. However, for the sake of brevity and in the context of this thesis I will focus on mechanisms that are salient at the CA3-CA1 synapse; unless otherwise stated the mechanisms discussed below pertain to this synapse.

1.3.3. Synaptic Plasticity of Excitatory Synapses

Since Donald Hebb first described the theory of synaptic plasticity (Hebb, 1949), there has been considerable efforts in elucidating the cellular mechanisms behind it. The first steps toward understanding synaptic plasticity were made by Timothy Bliss and Terje Lømo, who demonstrated that high frequency stimulation of presynaptic fibers could elicit a robust increase in the efficacy of the postsynaptic cell (Bliss & Lomo, 1973; Lomo, 2003), termed LTP or also known as hebbian plasticity. Since then, scientists have elucidated that for LTP to occur at CA3-CA1 synapses the following steps are required: (1) NMDA receptor activation (Collingridge et al., 1983), (2) a rise in postsynaptic Ca2+ (Lynch et al., 1983), (3) persistent postsynaptic depolarization (Malinow &

Miller, 1986; Wigstrom et al., 1986) and (4) AMPA receptor insertion at the postsynapse (Shi et al., 1999; Bredt & Nicoll, 2003).

Many of these studies looked at the electrophysiological response of the postsynaptic cell, however, early electron microscopy studies also showed that high frequency stimulation could increase the size (Fifkova & Van Harreveld, 1977; Desmond & Levy, 1986) and number

(Trommald et al., 1996; Geinisman et al., 2000) of dendritic spines on the postsynaptic cell.

The link between LTP and spine structure was suggested by the finding that the size of the postsynaptic density is related to the size of the spine head (Harris et al., 1992) and the number of

AMPA receptors within it (Nusser et al., 1998; Kharazia & Weinberg, 1999; Takumi et al., 1999) and ultimately an increase in the number of synapses (Luscher et al., 2000). The use of two-photon uncaging of glutamate made the ultimate structure-function link, which demonstrated that there was a strong correlation between the number of AMPA receptors and the size of a spine and this could be modulated by LTP (Matsuzaki et al., 2001; Noguchi et al., 2005; Asrican et al., 2007).

All of which brings us to our current understanding of LTP on a structural level. Briefly and

13 simplified, in response to postsynaptic Ca2+ increases via NMDA receptors AMPA receptors are trafficked or inserted into the postsynaptic density in order to increase the postsynaptic cell’s efficacy to presynaptic glutamate release, which will then lead to spine head enlargement

(Matsuzaki et al., 2004). Notably, protein synthesis is required to maintain this functional and structural plasticity (Kelleher et al., 2004; Tanaka et al., 2008; Yang et al., 2008).

For a long time, it was thought that LTP was the sole mechanism that regulated learning and memory on a cellular level, however, it seemed likely that there should be a mechanism that could reverse the potentiation caused by learning. Interestingly, in 1982, long-term depression

(LTD) was discovered (Ito & Kano, 1982), which is characterized by a long-term decrease in synaptic strength in response to low frequency stimulation. It is thought that LTD occurs through the following mechanism: (1) low frequency stimulation activates NMDA receptors (Dudek &

Bear, 1992) or G-protein coupled metabotropic glutamate receptors (Bolshakov & Siegelbaum,

1994), which causes a (2) small influx of Ca2+ into the postsynaptic cell (Mulkey & Malenka,

1992), (3) leading to a depotentiation of the synapse through a mechanism reliant on the Ca2+- dependent protein phosphatase calcineurin (Mulkey et al., 1994) and (4) thus leading to the internalization of AMPA receptors (Carroll et al., 1999). The activation of this cellular cascade eventually leads to spine shrinkage in response to internalization of AMPA receptors (Zhou et al.,

2004).

In order for dendritic spines to remodel in response to LTP or LTD stimuli, the underlying actin cytoskeleton must also be remodeled (Okamoto et al., 2004; Sala & Segal, 2014). In the case of LTP, actin cytoskeleton remodeling is important for long-lasting maintenance of potentiation.

For example in the dentate gyrus, LTP leads to a long lasting increase in filamentous actin (f-actin) of up to 5 weeks (Fukazawa et al., 2003). Moreover, if actin polymerization is blocked with

14 cytochalasin D, a selective inhibitor for the end cap of actin filaments, it can impair long-term maintenance of LTP, but not short-term (Udo et al., 2005). The family of small Rho GTPases play an important role in this actin-dependent remodeling during synaptic plasticity, including: RhoA

(Tashiro et al., 2000; Tashiro & Yuste, 2004; Briz et al., 2015), Rac1 (Bongmba et al., 2011; Tan et al., 2011), and Cdc42 (Tashiro et al., 2000). When Rac1 is knocked down in vivo both LTP in area CA1 and hippocampal learning are impaired in rodents (Haditsch et al., 2009). Therefore, dendritic spines rely on actin remodeling for the type of long-term plasticity required for learning.

Changes in postsynaptic structure and function have been well described in both LTP and

LTD, but, it is important to note that some evidence suggests that presynaptic changes also aid in postsynaptic potentiation. It is well known that mossy fiber LTP in area CA3 is independent of

NMDA receptors and is mostly induced presynaptically (Nicoll & Malenka, 1995; Nicoll &

Schmitz, 2005). However, at the CA3-CA1 synapse, for many years, long-term synaptic plasticity was described only by postsynaptic mechanisms. Recent advances in high-resolution in vivo imaging have demonstrated that there may be changes in probability of presynaptic vesicle release and change in quantal content of presynaptic vesicles during LTP (Emptage et al., 1999; Emptage et al., 2003; Ward et al., 2006; Enoki et al., 2009). Though this idea is still controversial (Granger

& Nicoll, 2014), given that two-photon uncaging of glutamate, which completely bypasses the presynapse, can potently induce LTP (Matsuzaki et al., 2001).

The mechanisms described above pertain to classical hebbian plasticity, though now there are several identified mechanisms of non-hebbian plasticity. For example, emerging evidence suggests that activity can alter intrinsic neuronal excitability or principle neurons (Daoudal &

Debanne, 2003; Zhang & Linden, 2003), which can be mediated by regulating the expression or the biophysical properties of voltage- and Ca2+-gated ion channels (Guzman-Karlsson et al., 2014).

Another important non-hebbian form of homeostatic plasticity is synaptic scaling, which involves bidirectional changes in post-synaptic activity in response to chronically elevated or depressed activity levels (Turrigiano & Nelson, 2004). It has been hypothesized that synaptic scaling is a homeostatic adaptation within a neural circuit that can maintain proper directionality of information flow (Kim & Tsien, 2008).

Next, I would like to describe the function of inhibitory synapses and I will focus only on mechanisms of inhibitory neurotransmission and plasticity on hippocampal pyramidal neurons.

1.3.4. Inhibitory Neurotransmission in CA1 Pyramidal Neurons

GABA is the main inhibitory neurotransmitter in the hippocampus and it is synthesized by inhibitory interneurons via glutamic acid decarboxylase (GAD), an enzyme which has two isoforms: GAD65 and GAD67 (Erlander et al., 1991). Both GAD65 and -67 are often used as markers for GABAergic interneurons (Benson et al., 1994; Fukuda et al., 1997; Shi et al., 1999).

GABA is released presynaptically and goes on to bind postsynaptic GABAA receptors (GABAAR), the ionotropic type of GABA receptors that mediate fast inhibitory neurotransmission (Sieghart,

1995) or GABAB receptors, which are G-protein coupled receptors that mediate slow inhibitory neurotransmission and are found both pre- and post-synaptically (Craig & McBain, 2014).

GABAAR are heteromeric pentamers made up of 5 subunits from a total of 20, including:

α(1-6), β(1-3), γ(1-3), δ, ε, π, θ, and ρ(1-3) (Sieghart, 1995). Therefore, GABAAR are quite structurally diverse and theoretically can yield up to 23 subunit combinations, though the α1β2γ2 accounts for a large proportion of GABAAR in the mammalian central nervous system (McKernan

& Whiting, 1996). In the mature adult nervous system, binding of GABA to postsynaptic

- GABAAR causes channel opening and the influx of chloride ions (Cl ) leading to hyperpolarization of the postsynaptic membrane, thereby inhibiting action potential propagation in the postsynaptic neuron. However, during development GABA can be depolarizing at resting membrane potentials, i.e. excitatory (Ben-Ari et al., 1989; Thompson & Gahwiler, 1989). For GABA to be a hyperpolarizing transmitter, neurons must maintain a tight Cl- gradient that requires a low

- intracellular Cl concentration (Rivera et al., 1999; Blaesse et al., 2009), such that when GABAAR are activated Cl- flows in to the cell, making the membrane more negative. This Cl- gradient is maintained by K-Cl cotransporter (KCC2), a membrane-bound transporter that exchanges potassium ions (K+) for Cl- (Payne, 1997). The expression of KCC2 is developmentally regulated, in rodents its expression only rises in the second postnatal week (Rivera et al., 1999), meaning that

- prior to this when GABAAR are activated, Cl flows out of the cell, making GABA excitatory.

Increasing KCC2 expression shifts the chloride equilibrium potential more negative, thereby decreasing inhibitory post-synaptic current (IPSC) amplitudes at resting membrane potential.

Changes in KCC2 expression and subsequent breakdown in the Cl- gradient following injury, this is expanded on further in section 1.6.2.

Gray’s work on type II synapses demonstrated that inhibitory synapses were symmetrical in nature and their post-synaptic density was not nearly as dense as type I excitatory synapses

(Gray, 1959a). In contrast to excitatory synapses, the postsynaptic density of GABAergic synapses contains only a few scaffolding proteins. Gephyrin is the main scaffolding protein that forms the core of GABAergic postsynapses (Sassoe-Pognetto et al., 2000). Though we lack full structural information of gephyrin, it is believe that gephyrin self-assembles into a hexagonal lattice by auto- aggregation of gephyrin trimers (Fritschy et al., 2008). Gephyrin interacts directly with GABAAR subunits clustering them to the postsynaptic membrane (Tretter et al., 2008; Mukherjee et al.,

2011; Tretter et al., 2011; Kowalczyk et al., 2013). The most well characterized interaction of gephyrin is with the γ2-subunit, as gephyrin clusters are completely absent in γ2-subunit knockout mice (Essrich et al., 1998) and transfection of γ2-deficient neurons with the transmembrane spanning domain 4 of γ2 is sufficient for recruiting gephyrin and GABAAR to the plasma membrane (Alldred et al., 2005). Gephyrin is brought to postsynaptic sites by neuroligin-2

(Poulopoulos et al., 2009; Giannone et al., 2013), which is a cell adhesion molecule that is part of the neuroligin(1-4) family that have fundamental roles in synapse formation and regulation

(Scheiffele et al., 2000; Graf et al., 2004). It is thought that neuroligin-2 is first brought to the postsynaptic membrane, recruiting gephyrin, which then in turn brings GABAAR to the synapse.

Evidence suggests that GABAAR clustering is noticeably disrupted in gephyrin knockout mice

(Kneussel et al., 1999). Kneussel and colleagues found that α2- and γ2-GABAAR subunit immunostaining is no longer punctated and is more diffused in hippocampal neuronal cultures

-/- made from gephyrin mice. Surprisingly, the group was still able to record GABAA-mediated miniature inhibitory postsynaptic currents (mIPSC), a measure of inhibitory synapse strength, though they reported a significant reduction in the amplitude implying there are less receptors at the synapse. Moreover, immunofluorescence studies in the spinal cord suggest that some

GABAARs at the plasma membrane do not colocalize with gephyrin immunostaining (Kneussel et al., 2001; Lorenzo et al., 2014). This suggests that there are both gephyrin-dependent and gephyrin-independent GABAAR clustering mechanisms. Gephyrin is also a known modulator of

GABAergic synaptic plasticity. Though considerable attention has been devoted to plasticity of excitatory synapse much less is known about plasticity of inhibitory synapses.

1.3.5. Synaptic Plasticity of Inhibitory Synapses

In many cases of GABAergic synaptic plasticity, the induction of both postsynaptic and presynaptic forms requires non-GABAergic stimuli, often from nearby excitatory synapses

(Castillo et al., 2011). For instance, inhibitory synapses are known to be potentiated in response to

NMDA receptor stimulation (Marsden et al., 2007; Petrini et al., 2014), which recruits GABAAR to the postsynaptic membrane via recruitment of gephyrin (Petrini et al., 2014). Gephyrin is a highly dynamic molecule, time-lapse imaging from neurons expressing GFP-tagged gephyrin demonstrated that gephyrin clusters are constantly remodeled and can move within dendrites concurrently with its presynaptic terminal over several micrometers (Dobie & Craig, 2011).

Moreover, it is thought that gephyrin size can correlate to synaptic strength (Varley et al., 2011;

Vlachos et al., 2013).

Regulation of the size of gephyrin has been linked to many post-translational modifications including phosphorylation (Zita et al., 2007; Tyagarajan et al., 2011; Tyagarajan et al., 2013;

Flores et al., 2015), S-nitrosylation (Dejanovic & Schwarz, 2014) and palmitoylation (Dejanovic et al., 2014). It has been shown that altering gephyrin cluster size and number through phosphorylation can alter GABAergic activity (Tyagarajan et al., 2011; Flores et al., 2015).

Tyagarajan and colleagues demonstrated that serine 270 (ser270) of gephyrin can be phosphorylated by glycogen synthase kinase 3 β (GSK3β) and this negatively regulates the clustering of gephyrin and GABAAR in dissociated hippocampal neurons (Fig 1.4) (Tyagarajan et al., 2011). They demonstrated this by transfecting GFP-tagged gephyrin containing a serine to alanine point mutation (S268A), a dephospho-gephyrin mutant, into dissociated hippocampal neurons. They found that this point mutation resulted in a higher density of gephyrin clusters and in turn resulted in a higher frequency of GABAA-mediated mIPSC (Tyagarajan et al., 2011).

Figure 1.4 – ERK1/2 regulates gephyrin cluster size and GSK3β regulates number of gephyrin clusters. (a) Schematic representation of gephyrin domains and the identified phosphorylation sites. Ser270 and Ser268 are recognized targets of GSK3β and ERK1/2 kinase activities, respectively. (b) Stimulation of receptor tyrosine kinases (RTKs) by ligand binding can activate Ras and downstream signaling cascades Ras/MAPK and PI3K/Akt leading to gephyrin phosphorylation at Ser268 by ERK. Akt also regulates GSK3β activity, the kinase responsible for

Ser270 phosphorylation. Modified from (Zacchi et al., 2014).

Furthermore, using similar methods the same group demonstrated that extracellular-signal- regulated kinase 1/2 (ERK1/2) can phosphorylate gephyrin at serine 268 (ser268) and a global increase in network activity was sufficient to generate this effect (Tyagarajan et al., 2011).

Moreover, phosphorylation of ser268 negatively regulated the size of gephyrin clusters (Fig 1.4) and in turn affected the amplitude of GABAA-mediated mIPSCs (Tyagarajan et al., 2013).

Interestingly, if gephyrin is phosphorylated at both ser268 and ser270, it becomes a substrate for calpain-dependent cysteine cleavage and subsequently targeted for proteasomal degradation

(Tyagarajan et al., 2013). This tight regulation by GSK3β and ERK1/2 acts as a molecular switch to quickly enhance or dampen GABAergic neurotransmission. Therefore GSK3β (Peineau et al.,

2007) and ERK1/2 (Davis et al., 2000a; Yang et al., 2004; Zhai et al., 2013) are important for not only glutamatergic synaptic plasticity, but as well as regulation of inhibitory plasticity through gephyrin. Interestingly, as network activity increases GSK3B and ERK1/2 are thus able to orchestrate both excitation and inhibition in order to enhance excitatory postsynaptic potentiation.

Several other signaling cascades have emerged at the GABAergic postsynapse that are important for inhibitory plasticity, including: protein kinase A (PKA), protein kinase C (PKC), and CaMKII. Many of these kinases converge at important phosphorylation sites on GABAAR subunits themselves, thereby regulating channel function and activity (Vithlani et al., 2011;

Kullmann et al., 2012). CaMKII phosphorylation of GABAAR subunits is quite interesting as it could mediate crosstalk between excitatory and inhibitory synapses. In vitro studies demonstrate that CaMKII can phosphorylate GABAAR subunits β1-3 and γ2 on their intracellular domains

(McDonald & Moss, 1994; 1997) and can phosphorylate gephyrin (Flores et al., 2015). Moreover, these phosphorylation events can significantly potentiate amplitudes of GABAA-mediated currents in heterologous cells (Houston & Smart, 2006). Scientists are only beginning to fully understand

21 the cellular signals which mediate GABAAR plasticity, but it is very interesting that some of the players of inhibitory synaptic plasticity overlap with excitatory synaptic plasticity.

However, in contrast to excitatory synapses, some of the best characterized forms of inhibitory plasticity involve changes in presynaptic release of GABA. One important mechanism for LTD of inhibitory synapses (iLTD) on CA1 pyramidal neurons involves endocannabinoid retrograde signaling. Endocannabinoid signaling can differentially regulate short-term versus long-term inhibitory plasticity. For short-term plasticity, in response to incoming electrical activity postsynaptic metabotropic glutamate receptor 1 (mGluR1) at excitatory synapses activate phospholipase C (PLC). PLC activation leads to the production of the endocannabinoid 2- arachidonoylglycerol, which diffuses across the synaptic cleft and activates presynaptic cannabinoid receptor type 1 (CB1) receptors leading to temporary depression of evoked and spontaneous GABA release (Ohno-Shosaku et al., 2001; Wilson & Nicoll, 2001). Whereas for long-term inhibitory plasticity, 2-arachidonoylglycerol release is dependent on both AMPA and

NMDA receptors (Chevaleyre & Castillo, 2003), which then activates CB1 receptors.

Interestingly, BDNF can also be a retrograde signal for presynaptic plasticity of GABAergic synapses, but this will be discussed further in section 1.6.5.

Given that we have now covered hippocampal anatomy, cellular morphology and physiology in depth, I would now like to discuss hippocampal injuries and how this can affect cognition and neuronal function.

1.4. HIPPOCAMPUS AND INJURY

There are many brain injuries and disorders that are known to affect the hippocampus. For example in Alzheimer’s disease, a neurodegenerative disorder, the hippocampus is highly effected and patients tend to have varying degrees of memory deficits that worsen with age (Selkoe, 2002).

Two of the best studied disorders that affect the hippocampus are epilepsy and ischemic stroke.

1.4.1. Epilepsy

It is well known that the hippocampus is vulnerable to epilepsy (Sommer, 1880), a condition that is characterized by recurrent and unprovoked seizures. Approximately 1% of epilepsies can be of genetic or congenital origin and this is often manifested in children (Berkovic et al., 2006). Epilepsy in an adult is typically a disorder acquired through a process known as epileptogenesis, resulting from a brain injury (infection, tumour, head trauma etc.) to a previously normal brain (Goldberg & Coulter, 2013). Initially, these injuries can cause an imbalance in neurotransmitters, loss of ion gradients or disruption of the blood-brain barrier (Pitkanen &

Lukasiuk, 2011) all of which can lead to seizure development in an acute timeframe. However, it is well known that if a patient has one seizure their chances of developing more seizures and subsequently developing epilepsy goes up significantly (Hauser & Lee, 2002), i.e. seizures can beget more seizures. There are two main types of seizures that are exhibited by patients: (1) focal seizures, where the patient typically can experience an aura accompanied by confusion and unresponsiveness or (2) generalized seizures (including tonic, clonic, tonic-clonic and absence seizures), where typically the patient loses consciousness and this may or may not involve contraction of muscles and limbs (Berg et al., 2010). A seizure can last from a few seconds to more

23 than five minutes at which point it is known as status epilepticus (Trinka et al., 2012). It is important to note that many patients progress from focal seizures to generalized seizures and sometimes eventually to status epilepticus; this process is thought to be the functional consequence of a progressive process that begins with epileptogenesis.

Epileptogenesis activates many complex cellular cascades through which neuronal networks become progressively dysfunctional and aberrant. In this timeframe many functional, structural and biochemical changes are occurring. Evidence suggests that during epileptogenesis the expression of many proteins such as neurotransmitter receptor subunits, ion channels/transporters, neurotrophins and many cellular second messengers, are rapidly changing

(Rakhade & Jensen, 2009; Varvel et al., 2015). These changes in protein expression have been linked to DNA methylation and other epigenetic changes that can enhance or suppress DNA transcription (Qureshi & Mehler, 2010; Machnes et al., 2013). However, it is poorly understood how injury can generate transient alterations in protein expression and which then leads to permanent structural and functional changes within neuronal networks.

This is most apparent in post-traumatic epilepsy (PTE), which is a type of acquired temporal lobe epilepsy after traumatic brain injury. Acute seizures can occur in 25% of cases following immediately traumatic brain injury (Temkin, 2001) and chronic epilepsy develops in anywhere from 10-30% of patients who have severe injuries, particularly in patients with head wounds which penetrate the Dura (Frey, 2003). The onset of epilepsy can be delayed from weeks to months or even years after the injury has occurred (Frey, 2003), termed late-onset PTE. These patients respond very poorly to traditional anti-convulsant therapies. Moreover, anti-convulsants have no effect in preventing the development of late-onset PTE (Young et al., 1983b; a; Temkin et al., 1990; Temkin et al., 1999; Temkin, 2009). They are also poor candidates for resection

24 surgeries, given that they have already sustained a head trauma. Remarkably, despite the fact that a large proportion of patients who develop late-onset PTE sustain cortical injuries, many of them will eventually develop epileptic foci directly in the hippocampus, as monitored by intracranial electroencephalography (Marks et al., 1995). However, exactly what occurs between injury and the silent period before the presentation of epilepsy in late-onset variety of PTE is poorly understood.

Through post-mortem studies of epilepsy patients, anatomists have discovered the key hallmarks of temporal lobe epilepsy and PTE. Firstly, hippocampal sclerosis, a pathological condition with severe neuronal cell loss and gliosis within area CA1, is the most common finding in patients with epilepsy, including patients with PTE (Corsellis, 1957; Falconer et al., 1964;

Falconer, 1974). Often, this is accompanied by axonal sprouting in different regions of the hippocampus (Sutula et al., 1989; Lehmann et al., 2000), which can lead to defects in neuronal circuitry and propagate electrical activity that can eventually lead to seizures (Isokawa &

Levesque, 1991; Maru et al., 2002; Sutula & Dudek, 2007). This is further exacerbated by astrogliosis within the hippocampus of patients (Cendes et al., 2014). Consequently, neuronal cell death, gliosis and axon sprouting are considered the hallmarks of PTE.

These events have been modelled in rodents using cortical under-cutting (Hoffman et al.,

1994), fluid-percussion injury (Kharatishvili et al., 2006), stroke (Epsztein et al., 2008), excitotoxic agents such as picrotoxin (Kaplan & Williamson, 1978; Hamani & Mello, 2002) or kainic acid (Nadler et al., 1978; Ben-Ari, 1985) and kindling, in which repeat electrical stimuli is given to the hippocampus of rodents (Pinel & Rovner, 1978; Williams et al., 2007). These models recapitulate the classical hallmarks (circuit dysfunction, neuronal death, reactive gliosis and axonal sprouting). For example, rodents treated with kainic acid can develop reactive gliosis (Ding et al.,

2000) and cell death within area CA3 (Ben-Ari et al., 1979). It is thought that following kainic acid administration, as a compensatory mechanism to this cell death, mossy fibers from hippocampal dentate granule cells begin to sprout (Davenport et al., 1990; Routbort et al., 1999).

One theory suggests that sprouting of axons is caused by the reactivation of mechanisms that guide axons during development. Axonal sprouting is thought to lead to functional deficits within circuits and has been modeled with axon transection. For example, in vivo lesion of mossy fibers in animals that have not experienced seizures (Laurberg & Zimmer, 1981; Hannesson et al., 1997) or cortical- undercutting (Salin et al., 1995) can cause potent axon remodeling. Kainic acid treatment

(McNamara & Routtenberg, 1995; McNamara & Lenox, 2000) and lesion of Schaffer collaterals

(McKinney et al., 1997; Gill et al., 2013) can lead to axonal sprouting which is marked by re- expression of GAP43, a molecule important for axon growth and elongation (see section 1.3.1.).

Previously, our laboratory has modeled late-onset PTE in vitro, where we demonstrated that Schaffer collateral transection can lead to axonal sprouting in area CA3 (McKinney et al.,

1997). This resulted in more reciprocal excitatory synaptic connections between CA3 neurons and led to hyperexcitability of CA3 neurons 14 days following transection (McKinney et al., 1997). I used this model in Chapter 3 and 4 to elucidate some of the cellular mechanisms that underlie injury-induced axonal sprouting.

Interestingly, seizures following stroke accounts for up to 11% of new cases of adult epilepsy and the severity of epilepsy correlates with stroke size and severity (Hauser et al., 1993;

Silverman et al., 2002). In the next section, I will discuss further on how ischemic stroke can affect the hippocampus and neurocircuitry.

1.4.2. Ischemia and Stroke

Stroke is the second leading cause of death in developed countries (Strong et al., 2007).

Some patients recover from stroke but approximately 30% remain disabled and go on to develop psychiatry conditions and/or neurological deficits in speech, movement, learning and memory

(Strong et al., 2007; Hackett et al., 2014). The large majority of strokes in adults are ischemic strokes, which are caused by a reduction in blood flow to the brain. The brain is heavily dependent on blood flow for continuous supply of oxygen and glucose. During cerebral ischemic strokes blockade of blood flow is often focal. In the central core regions of the insult there is usually a complete arrest of blood flow and this area evolves rapidly toward death within minutes.

Surrounding the core is the penumbra region, in which blood flow is diminished but not arrested.

This neuronal tissue lies transiently above the threshold of cell death for a slightly longer period of time than the core region (Xing et al., 2012). Therefore, the penumbra would be the most likely site of pharmacological intervention in hopes of preventing further cell death. In particular, the hippocampus seems to be more susceptible to ischemia than other brain regions. CA1 pyramidal neuron cell death has been noted in patients (Brierley & Cooper, 1962; Siesjo, 1981; Zola-Morgan et al., 1986) and memory loss in a subset of patients is prevalent even years after the initial stroke

(Stewart et al., 1996).

Animal models of cerebral ischemia, which attempt to reproduce conditions of ischemic stroke, have also reported specific neuronal loss of CA1 pyramidal neurons (Pulsinelli & Brierley, 1979;

Pulsinelli, 1985; Crepel et al., 1989; Schmidt-Kastner & Freund, 1991) and an association with long-term hyperexcitability in the hippocampus (Congar et al., 2000; Crepel et al., 2003; Epsztein et al., 2006). For example, forebrain ischemia, where there is a bilateral occlusion of the common carotid artery, can lead to CA1 pyramidal neuron cell death as early as 2 minutes after ischemia

Figure 1.5 - Pathways of ischemic cell death in the brain. An overview of the processes that are

+ activated and involved in ischemic cell death. Note: pHi: intracellular pH; Nai: intracellular Na ;

2+ - Cai: intracellular Ca ; FFA: free fatty acids; PAF: platelet-activating factor; e Transport: electron transport. Modified from (Lipton, 1999).

28 onset and other brain regions follow within a few minutes (Smith et al., 1984).

Using animal models, scientists have determined that the initial consequence of ischemia is fundamentally an energy problem. In the affected region, adenosine triphosphate (ATP) consumption increases while production decreases, ultimately leading to lactate acidosis and a decoupling of important ion gradients in neurons (Mies et al., 1990). Subsequently, many multicellular downstream signaling cascades are activated to the detriment of the neuron (Fig 1.5).

A widely held theory suggests that the disruption of ATP stores can lead to neurotransmitter release and inhibition of reuptake, especially glutamate (Benveniste et al., 1984; Hagberg et al., 1985;

Nishizawa, 2001). This increase of glutamate can lead to a cellular influx of Ca2+ triggering the activation of proteases and promoting cell swelling, ultimately leading to cell death (Lipton, 1999).

Therefore, over the last 30 years there has been a push to develop neuroprotective therapies.

Currently, first-line therapy for embolic stroke is intravenous delivery of tissue plasminogen activator (tPA), which will break down clots (NINDS, 1995). Due the limited time window of use

(currently, <3 hours after first sign of stroke) and the number of contraindications for thrombolysis, only approximately 2-5% of patients receive tPA (Donnan et al., 2011). Therefore, other treatment strategies are needed. However, since the discovery of tPA 20 years ago there has been a successive failure of neuroprotective drugs in clinical trial, including therapies targeting: glutamate receptors

(Davis et al., 2000b; Albers et al., 2001; Saver et al., 2015), GABA receptors (Lyden et al., 2001),

Ca2+ channels (Horn et al., 2001), Na+ channels (Gibson et al., 2010) and free radicals (Diener et al., 2008) among others. So in order to develop new neuroprotective strategies, we need to better understand the cellular mechanisms underlying synaptic disruption after ischemic brain injury.

1.4.3. Defects in Plasticity Following Brain Injury

Synaptic plasticity is vital for learning and memory, however, the mechanisms underlying synaptic plasticity (as described in section 1.3.) may be misappropriated in response to brain injury.

It’s well known that dendritic spines become malformed in patients with epilepsy, stroke, dementia and depression (Fiala et al., 2002; Calabrese et al., 2006). Given that synapse function and plasticity is directly linked to dendritic spine morphology, the disruption of spines is likely to negatively influence cognition and function.

For example, loss of dendritic spine density is observed in patients with epilepsy (Scheibel et al., 1974; Vaquero et al., 1982; Swann et al., 2000) and in animal models of epilepsy (Jiang et al., 1998; Isokawa, 2000). In vivo two-photon imaging of mice expressing GFP in pyramidal neurons has demonstrated that 30 minutes after a kainate-induced seizure there is acute axonal blebbing, dendritic beading and loss of dendritic spines (Zeng et al., 2007). Whether this structural loss is persistent seems to depend on the severity and length of status epilepticus subjected to the mouse (Guo et al., 2012). On a functional level, this synaptic disruption can affect cognition. For example, after pilocarpine-induced seizures mice exhibit deficits in learning and memory tests, such as the Morris water maze or novel object exploration (Letty et al., 1995; Groticke et al., 2007;

Muller et al., 2009). In epilepsy, changes in excitation are also coupled with changes in inhibition.

For instance, mutations in GABAAR represents a large proportion of known genetic mutations that can lead to epilepsy. For instance, there are known mutations in α1- (Delgado-Escueta et al., 2013),

δ- (Dibbens et al., 2004) and γ2- (Lachance-Touchette et al., 2011) subunits that can cause epilepsy. Post mortem staining demonstrates a downregulation of GABAAR subunits in a subset of patients, even when there is no known cause of epilepsy (Loup et al., 2006; Loup et al., 2009).

The clinical implication of seizure-propagated synapse loss and alterations in the balance between

30 excitation-inhibition is that: (a) there will be reactive synaptogenesis (Sutula & Dudek, 2007) and

(b) patients may experience memory loss and learning deficits (Helmstaedter et al., 1991;

Helmstaedter, 2002).

Therefore, an important key to understanding how epileptogenesis occurs is to understand how mechanisms of synaptic plasticity can be misappropriated after seizures leading to reactive synaptogenesis that can further propagate seizure-activity. Beck and colleagues have shown that activity-dependent synaptic plasticity, i.e. induction of LTP, is impaired in the seizure focus in patients with temporal lobe epilepsy (Beck et al., 2000). Interestingly, neurotrophins, growth factors which are important modulators of synaptic plasticity, are known to be involved in the epileptogenesis process (discussed further in section 1.6.6.); however, we do not know how modulators of plasticity can encourage epileptogenesis.

One theory suggests that high levels glutamate release during seizures and subsequent increase in intracellular Ca2+ through NMDA receptors can cause dendritic spine retraction

(Delorenzo et al., 2005; McNamara et al., 2006; Wong & Guo, 2013). Large increases in intracellular Ca2+ can lead to the disruption of many signaling cascades. In particular, Ca2+ increases after seizures can activate calcineurin, a Ca2+-activated phosphatase, which can mediate dendritic spine loss in the hippocampus and neocortex following pilocarpine-induced (Kurz et al.,

2008) or kainate-induced (Zeng et al., 2007) status epilepticus. As mentioned in section 1.3.3., calcineurin is important for spine shrinkage in response LTD, therefore a compelling argument can be made to suggest that during status epilepticus this mechanism of synaptic plasticity is overtaken and misappropriated leading to dendritic spine loss.

Interestingly, synapse loss and cognitive impairment in patients with ischemic stroke can mirror what happens in patients with epilepsy. Following ischemia, it is well known that patients develop motor deficits and this can be coupled with memory loss and cognitive impairments

(Roding et al., 2009; Boussi-Gross et al., 2014; Lee & Pyun, 2014). There are very few studies that have looked at the ultrastructure of synapses in stroke patients, but there is an underlying theory that excitatory neurons, which have high metabolic needs, can be damaged due to diminishing availability of nutrients. In animal models, scientists have observed that during ischemia, dendrites of cortical pyramidal neurons can become beaded and there is an appearance of varicosities along dendritic shafts accompanied by a loss in dendritic spines (Park et al., 1996;

Hasbani et al., 2001; Enright et al., 2007). This loss of spines can be persistent depending on the severity and length of ischemia (Obeidat et al., 2000; Zhang et al., 2005). Ischemia to acute rat hippocampal slices can impair CA1 pyramidal neuron excitability (Zhang et al., 2008), which may impair synaptic plasticity. Along these lines, CA1 neurons have been shown to have increased

NMDA-mediated responses following ischemia to organotypic hippocampal slice cultures (Gee et al., 2006) and an overall decrease of glutamate receptors but a specific increase in Ca2+-permeable

AMPA receptors in ischemia to dissociated hippocampal neurons (Liu et al., 2006). Taken together, this would result in a drastic influx of Ca2+ into the cell, activating signaling cascades that may disrupt synapses. However, inhibiting glutamate receptors (Davis et al., 2000b; Albers et al., 2001; Saver et al., 2015) and Ca2+ channels (Horn et al., 2001) has been shown to not be effective in patients in clinical trials. Therefore, we must further determine the exact mechanisms of plasticity that are misappropriated after ischemia.

Neurotrophins, proteins which are important for synaptic plasticity, are highly upregulated after brain insults (Lindvall et al., 1994). However, what role they are playing in synaptic function

32 after injury is not known. First, I will discuss neurotrophin function and then I will go further in to the role of a particular member, BDNF, in plasticity and injury.

1.5. NEUROTROPHIC FACTORS AND THEIR RECEPTORS

Neurotrophins are a family of highly conserved proteins that are important for the survival, development, and function of neurons (Lewin & Barde, 1996). In the 1950s, nerve growth factor

(NGF) was discovered as protein with trophic properties, given that it promoted survival and growth of chick sensory neurons (Levi-Montalcini & Hamburger, 1951). Subsequently, BDNF was the second member of the neurotrophin family to be discovered (Barde et al., 1982; Leibrock et al., 1989). Shortly following the discovery of BDNF, neurotrophin-3 (NT-3) (Maisonpierre et al., 1990a; Maisonpierre et al., 1990b) and neurotrophin-4/5 (NT-4/5) (Hallbook et al., 1991; Ip et al., 1992) were described. Thus, NGF, BDNF, NT-3 and NT-4/5 became the canonical neurotrophin family. There are other trophic factors that have demonstrated homology to classical neurotrophins, such as: glial-derived neurotrophic factor (Lin et al., 1993) and ciliary neurotrophic factor (Manthorpe et al., 1980), however I will not be discussing non-canonical neurotrophins in this thesis.

1.5.1. Overview of Neurotrophins

Classically, neurotrophins were thought to be important for neurodevelopment and cell survival (Lewin & Barde, 1996). However, over the last 20 years neuroscientists have uncovered an important role for neurotrophins in synaptic plasticity in the adult nervous system (Schuman,

1999; Poo, 2001). Given that the hippocampus is a highly plastic region there is good reason to

33 believe that the expression of neurotrophins would be quite high. Through a series of thorough mRNA and protein expression studies, we now know that neurotrophins have both a distinct and overlapping distribution within the peripheral and central nervous systems (Lewin & Barde, 1996).

Specifically within the hippocampus, the expression of neurotrophins is overlapping. Table 1.1 summarizes the expression of NGF, BDNF, NT-3 and NT-4/5 within the hippocampus.

Table 1.1. Neurotrophin mRNA and protein expression in adult rodent hippocampus

Neurotrophin CA1 CA3 Dentate Gyrus References (Ernfors et al., 1990b; Hofer et al., 1990; Low NGF Low expression High expression Phillips et al., 1990; expression Conner et al., 1992; Das et al., 2001) (Ernfors et al., 1990b; Phillips et al., 1990; Moderate Moderate BDNF High expression Isackson et al., 1991; expression expression Lindholm et al., 1994; Yan et al., 1997) Low Low to Moderate (Ernfors et al., 1990b; NT-3 High expression expression expression Lindholm et al., 1994) Low Moderate (Lindholm et al., 1994; NT-4/5 Low expression expression expression Friedman et al., 1998)

Table 1.1 highlights that BDNF and NT-4/5 have moderately high expression within hippocampal area CA3 and CA1. Therefore, they seem to be ideal candidates to mediate synaptic plasticity at the Schaffer collateral-CA1 synapses. Interestingly, the expression of BDNF is modulated with activity (Zafra et al., 1990; Castren et al., 1992; Castren et al., 1993; Ichisaka et al., 2003; Rattiner et al., 2004), but in contrast the expression of NT-4/5 is not (Castren et al.,

1993; Ichisaka et al., 2003; Rattiner et al., 2004). Moreover, BDNF is required for the induction of LTP at the CA3-CA1 synapses, but neither NT-3 (Chen et al., 1999; Ma et al., 1999) nor NT-

4/5 (Chen et al., 1999) are required. Consequently, BDNF is an important player of synaptic plasticity in this region, which I will cover further in sections 1.6.4. and 1.6.5. In the next section

I will discuss neurotrophin receptors and downstream signaling.

1.5.2. Neurotrophin Receptors and Signaling Cascades

Neurotrophins mediate their effects through two types of receptors: (1) high-affinity tropomyosin-related kinases (Trk), which belong to the receptor tyrosine kinase superfamily and

(2) the low-affinity p75 neurotrophin receptor (p75NTR), which is part of the tumour necrosis factor receptor superfamily (Levitan & Kaczmarek, 2002). The high-affinity Trk receptors are further subdivided into: TrkA, TrkB and TrkC (Fig. 1.6). Through a series of elegant binding studies, we now know that TrkA binds preferentially to NGF (Kaplan et al., 1991a; Kaplan et al., 1991b; Klein et al., 1991), TrkB binds preferentially to BDNF (Klein et al., 1991; Soppet et al., 1991), NT-4/5

(Berkemeier et al., 1991) and NT-3 to a lesser extent (Klein et al., 1991; Soppet et al., 1991) and

NT-3 binds TrkC preferentially (Fig 1.6) (Lamballe et al., 1991).

The Trk receptors come in two main variants (Fig. 1.6): a full-length form which maintains the intrinsic tyrosine kinases activity and a truncated form which has the same binding affinity for their respective neurotrophins but do not have intracellular tyrosine kinases activity (Middlemas et al., 1991; Barbacid, 1994). Very little is known about the truncated Trk receptors, at first it was thought that they act as dominant-negative to full-length Trk receptors or as an extracellular buffer for neurotrophins (Jing et al., 1992; Eide et al., 1996; Fryer et al., 1997). Interestingly, over- expressing truncated TrkB in acute slices can increase dendritic length (Yacoubian & Lo, 2000) and deleting truncated TrkB in mice can decrease dendritic length (Carim-Todd et al., 2009).

Figure 1.6 – Neurotrophins and their receptors. NGF binds with high-affinity to TrkA, BDNF and NT-4/5 bind with high-affinity to TrkB, NT-3 binds with high-affinity to TrkC and to a lesser extent TrkC. TrkA-C come in two main isoforms, full-length (intracellular tyrosine kinase activity) or truncated (non-catalytic). All 4 group members bind with low-affinity to p75NTR. Figure modified from (Levitan & Kaczmarek, 2002).

Therefore, we might speculate that this is mediated through members of the Rho GTPases.

Recently it’s been shown in heterologous cells transfected with truncated TrkC that NT-3 binding to truncated TrkC receptors can activate Rac1 in order to enhance membrane protrusions and ruffling (Esteban et al., 2006). Moreover, binding of BDNF to truncated TrkB in astrocytes can inhibit Rho GTPases (Ohira et al., 2005). Therefore, there is some evidence that truncated receptors may have a physiological role beyond acting as a dominant-negative. However, given that beyond overexpression studies, which may enhance their dominant-negative effect, we do not know much about truncated Trk receptors. Therefore, moving forward all references made to Trk receptors refer to the full-length variant of the receptors, unless otherwise stated.

Neurotrophin mRNAs are typically transcribed as proproteins, following translation, proneurotrophins are cleaved into mature neurotrophins (Lessmann et al., 2003). All mature neurotrophins bind with low-affinity to p75NTR (Chao et al., 1986; Radeke et al., 1987; Ernfors et al., 1990a; Rodriguez-Tebar et al., 1990; Hallbook et al., 1991; Rodriguez-Tebar et al., 1992), whereas proneurotrophins have been shown to bind with much higher-affinity to p75NTR (Lee et al., 2001) especially when p75NTR is in complex with sortilin, a VPS10-domain containing protein

(Nykjaer et al., 2004; Chen et al., 2005). In the following two sections I will discuss in further depth Trk receptor and p75NTR signaling, respectively.

1.5.2.1. Trk Receptor Signaling

There is a high homology of the intracellular domain of the Trk receptors and thus it is likely that the signaling cascades they activate are similar (Barbacid, 1994; Friedman & Greene,

1999). Mature neurotrophins typically bind as homodimers to Trk receptors, inducing receptor dimerization and autophosphorylation of tyrosine residues in the activation loop of the intracellular

37 domain of the receptors (Poo, 2001; Chao, 2003). This triggers autophosphorylation of residues outside of the activation loop recruiting adaptor proteins, which contain domains and motifs that interact with the intracellular domain of Trk receptors. Trk receptors are known to activate PLCγ,

Ras-MAPK/Erk and phosphoinositol-3 kinase (PI3K) as seen in Fig 1.7 (Friedman & Greene,

1999; Huang & Reichardt, 2003). These pathways mediate the cell survival and gene transcription aspects of neurotrophins. Trk signaling cascades were identified via homology recognition to other receptor tyrosine kinases and subsequently point mutations were made to test the role of conserved phosphorylation sites in signal transduction. For example, pY490 (TrkA)/pY515 (TrkB) interacts with Shc, a family of adaptor proteins, which create a scaffold for the recruitment of Ras

(Kavanaugh & Williams, 1994; Stephens et al., 1994). This phosphorylation event and subsequent

Ras activation is important for neurite outgrowth (Inagaki et al., 1995). Moreover, pY785(TrkA)/pY816 (TrkB) are known to recruit PLCγ (Loeb et al., 1994; Stephens et al., 1994;

Qian et al., 1998), which can regulate gene transcription during LTP and learning and memory

(Gruart et al., 2007).

Interestingly, neurotrophins also activate Cdc42 and Rac in order to affect changes in cell motility, axonal growth and growth cone dynamics (Yuan et al., 2003). Given that Trks can induce changes to the cytoskeleton and that CaMKII can be activated through the PLCγ, Trk receptors are ideal candidates for modulation of synaptic plasticity.

1.5.2.2. p75NTR Signaling

p75NTR is the low-affinity neurotrophin receptor. Classically, it was thought that p75NTR was a cell death receptor given that it has a conserved intracellular death domain (Liepinsh et al.,

1997; Bamji et al., 1998). However, with further study, scientists have identified several signaling

Figure 1.7 – Trk receptors activate three main intracellular signalling pathways. Recruitment of adaptors to pY515 leads to activation of the Ras-MAPK/ERK and the PI3K cascade, which both promote neuronal differentiation and growth. Whereas recruitment and activation of PLCγ through phosphorylation of Y816 results in the stimulation of PKC and promotes the release of Ca2+ from internal stores and subsequent activation of CaMKII. Taken from (Minichiello, 2009).

39 cascades that are known to be activated following neurotrophin binding to p75NTR. Once proneurotrophins or mature neurotrophins bind p75NTR several adaptor proteins can be recruited to elicit downstream signaling, including: melanoma-associated antigen (MAGE) and neurotrophin receptor p75 interacting MAGE homologue (NRAGE) (Reichardt, 2006). NRAGE recruitment typically leads to cell cycle arrest and sometimes cell death (Salehi et al., 2000).

However, p75NTR can interact with different Trk receptors to modulate neurotrophin action

(Hempstead et al., 1991; Barker & Shooter, 1994; Friedman & Greene, 1999), allowing for the possibility for a non-cell death role of p75NTR.

In particular, p75NTR can modulate Rho small GTPases. For example, direct interaction of

RhoA and p75NTR can inhibit neurite outgrowth (Yamashita et al., 1999), which would imply that p75NTR may regulate axon or dendritic growth. Interestingly, p75NTR null mice have increased spine density (Zagrebelsky et al., 2005) and proBDNF binding to p75NTR can negatively regulate dendritic arborisation and spine density in vivo (Yang et al., 2014). RhoA in particular has been shown to be downstream of p75NTR-mediated dendritic spine remodeling (Gehler et al., 2004;

Domeniconi et al., 2005; Sun et al., 2012). Another member of the Rho small GTPase super family,

Rac1, which is known to regulate dendritic spine morphology (Luo et al., 1996; Nakayama et al.,

2000; Tashiro & Yuste, 2004; Tolias et al., 2005) is also known to be downstream of p75NTR

(Schweigreiter et al., 2004; Zeinieh et al., 2014). However, it remains to be determined whether

Rac1-mediated dendritic spine retraction occurs following p75NTR activation by proneurotrophins or neurotrophins.

As mentioned in section 1.5.1. BDNF is a very important neurotrophin in the hippocampus, especially given that it can be modulated by activity. In the next section I will discuss BDNF in depth in terms of its processing, its role in development and in synaptic plasticity.

1.6. BRAIN-DERIVED NEUROTROPHIC FACTOR

1.6.1. Transcription, Secretion and Processing of BDNF

In rodents, the BDNF gene locus consists of five exons (termed exons I-V) that encode for the BDNF protein, each containing their own distinctive promoters (Metsis et al., 1993; Timmusk et al., 1993). Eight discrete mRNAs arise from transcription of these exons, though the fully processed and cleaved mature BDNF is identical with all transcript forms (Metsis et al., 1993;

Timmusk et al., 1993). bdnf transcripts are produced with a signal peptide directing them to the endoplasmic reticulum for translation, processing and secretion (Mowla et al., 1999).

Neurotrophins are secreted via two mechanisms either: (1) constitutively or (2) through a regulated pathway in response to neuronal activity. In the constitutive pathway, mature BDNF and proBDNF leave the endoplasmic reticulum, transit through the trans Golgi network and accumulate into dense core vesicles (Dieni et al., 2012), which are constitutively released at the cell body.

However, BDNF seems to be preferentially sorted into the regulated pathway (Goodman et al.,

1996; Farhadi et al., 2000; Adachi et al., 2005). In this pathway, secretory vesicles containing

BDNF can be transported by motor protein complexes to subcellular secretion sites. It has been highly debated whether mature BDNF is release pre- or postsynaptically and there has been evidence demonstrating for both using fluorescent protein-tagged BDNF transfected into dissociated cortical and hippocampal neurons (Egan et al., 2003; Adachi et al., 2005; Lou et al.,

2005). In contrast, knock-in mice expressing Myc-tagged BDNF have shown that BDNF can be exclusively found in presynaptic terminals, but not postsynaptic terminals in adult mouse

41 hippocampus (Dieni et al., 2012). We will further discuss how activity can regulate BDNF secretion in section 1.6.2.

During transit through the trans Golgi network, proBDNF can be cleaved by pro-protein convertases to mature BDNF (Lessmann et al., 2003). However, recent studies demonstrate that proBDNF can also be released in significant quantities by neurons (Yang et al., 2009), though this idea remains controversial (Matsumoto et al., 2008). Moreover, cleavage of proBDNF can occur in the extracellular space by proteinases such as: matrix metalloproteinases (MMP1-9) (Lee et al.,

2001; Hwang et al., 2005), plasmin (Lee et al., 2001; Pang et al., 2004; Hwang et al., 2005) and tissue plasminogen activator (Pang et al., 2004). The expression of MMP9 in particular can be regulated by neuronal activity (Wang et al., 2008), generating a feedback loop by which cleavage is enhanced when needed, in a spatiotemporal manner, in response to external stimuli.

Interestingly, in humans a particular single nucleotide polymorphism in the BDNF gene has been identified (Neves-Pereira et al., 2002; Egan et al., 2003; Hong et al., 2005) and shown to decrease the efficiency of BDNF secretion resulting in learning deficits in patients (Egan et al.,

2003). This polymorphism at nucleotide 196 in the pro-region of the gene results in an amino acid substitution from valine in position 66 to methionine (Val66Met) (Egan et al., 2003). Patients with this polymorphism have higher susceptibilities for mood disorders (Rybakowski, 2008), schizophrenia (Notaras et al., 2015), depression (Hosang et al., 2014) and post-traumatic stress disorder (Frielingsdorf et al., 2010). Although, how reduced or misregulated BDNF can cause neurological deficits is poorly understood, but given that activity is important for BDNF secretion, this may underlie cognitive deficits seen patients with the Val66Met polymorphism.

1.6.2. Activity-Dependent Regulation of BDNF

Several lines of evidence suggest that external stimuli can activate BDNF gene expression.

For example in rodents, light stimulation increases bdnf mRNA in the visual cortex (Castren et al.,

1992), likewise whisker stimulation increases bdnf mRNA expression in the barrel cortex

(Rocamora et al., 1996). Given this, it is unsurprising that membrane depolarization or increases in intracellular Ca2+ alone can promote BDNF transcription (Zafra et al., 1990; Zafra et al., 1991;

Ghosh et al., 1994; West et al., 2001). In particular, transcription of exon III in rats and exon IV in mice is induced by activity; note: the nomenclature for rats and mice are slightly different and exon III in rats corresponds to exon IV in mice (Aid et al., 2007). Interestingly, these exons are under control of a Ca2+ response element (CRE) (Tao et al., 2002), making them ideal candidates for activity-dependent transcription of BDNF.

Recently, epigenetic mechanisms that regulate BDNF transcription at exon IV were identified in mice (Martinowich et al., 2003). They demonstrated in dissociated hippocampal neurons that neuronal depolarization triggers demethylation of the BDNF promoter, which led to the dissociation of the transcriptional repressor methyl CpG binding protein 2 (MeCP2) from the promoter region (Martinowich et al., 2003). Moreover, binding of MeCP2 to the promoter region of bdnf can be turned on and off by the Ca2+-dependent phosphorylation of serine 421 of MeCP2

(pMeCP2S421), by either CaMKII or CaMKIV (Chen et al., 2003). Once MeCP2 dissociates from the promoter, transcription machinery can be recruited to transcribe bdnf mRNA. pMeCP2S421 can regulate BDNF-dependent dendritic growth and spine maturation (Zhou et al., 2006), therefore this phosphorylation site is very important for carrying out BDNF-mediated structural remodeling.

MeCP2 is also expressed by astrocytes and other glial cells (Skene et al., 2010; Diaz de Leon-

Guerrero et al., 2011), but its role there is less understood. MeCP2 is mutated in more than 90%

43 of clinically diagnosed cases of Rett Syndrome (Amir et al., 1999), a pervasive developmental disorder which is characterized by neurological deficits, seizures, stereotypic hand wringing and failure to grow (Hagberg et al., 1983). Using mouse models of Rett syndrome, consisting of

MeCP2 deletion, we now know there are many targets of MeCP2 regulation of transcription, including BDNF (Li & Pozzo-Miller, 2014). Interestingly, induction of LTP is impaired in MeCP2 knock-out mice (Moretti & Zoghbi, 2006) and they have less glutamatergic synapses (Chao et al.,

2007). These deficits can be rescued by BDNF (Larimore et al., 2009).

Not only is Ca2+ important for activity-dependent transcription of BDNF, it can also trigger its vesicular release. Long-lasting dendritic depolarization can activate L-type voltage-gated Ca2+ channels, which then trigger Ca2+-induced Ca2+ release from intracellular stores, and this sustained

Ca2+ signal can trigger BDNF secretion by activating CaMKII (Balkowiec & Katz, 2002; Kolarow et al., 2007). Similarly, PKA, a kinase which is important in synaptic plasticity (Nguyen & Woo,

2003; Kandel, 2012), can also induce BDNF secretion (Patterson et al., 2001; Kolarow et al.,

2007). This suggests there is an orchestrated response to activity that coordinates three important regulators of synaptic plasticity: BDNF, CaMKII and PKA (Gottmann et al., 2009). I will further discuss the role of BDNF in plasticity in sections 1.6.4. and 1.6.5.

It is important to note that under pathological conditions BDNF can also be released by non-neuronal glial cells, such as astrocytes (Bergami et al., 2008; Parpura & Zorec, 2010) and microglia (Miwa et al., 1997; Trang et al., 2011). For example, after spinal cord injury in mice, microglia are known to release BDNF, which can then affect neurotransmission by inducing KCC2 downregulation (Coull et al., 2005), the importance of KCC2 is outlined in section 1.3.4.

There is a very well described role of BDNF in neuronal survival during development, in particular of vestibular, trigeminal, hippocampal and cerebellar neurons (Huang & Reichardt,

2001). However, I will only discuss the role of BDNF in hippocampal synapse development.

1.6.3. BDNF in Synapse Development

Beyond the well-established effects on neuronal survival, BDNF also profoundly influences dendritic arborisation (McAllister et al., 1996) and axon growth cone progression and branching (Cohen-Cory & Fraser, 1995). It is so potent, that applying exogenous BDNF to dissociated hippocampal neurons can induce the formation of both inhibitory and excitatory synapses (Vicario-Abejon et al., 1998). Similarly, inhibiting endogenous BDNF with a function blocking antibody leads to enhanced synapse elimination during development (Hu et al., 2005).

1.6.3.1. BDNF and Development of Excitatory Synapses

At glutamatergic synapses, BDNF can enhance synapse formation in two ways: (1) by inducing axon growth and axon-dendrite contact and (2) by enhancing pre- and post-synaptic functional differentiation. In conjunction with AMPA receptor-mediated postsynaptic currents,

BDNF signaling is important for formation of mature spines (Tyler & Pozzo-Miller, 2003). It is thought that this coordination between BDNF and activity may stabilize contact formation by triggering Ca2+ transients in the postsynaptic compartment, as mentioned in section 1.3.1. Studies suggest that these Ca2+ transients are in part mediated by BDNF signaling (Lang et al., 2007;

Lohmann & Bonhoeffer, 2008). The frequency of Ca2+ transients in dissociated hippocampal neurons is decreased by inhibiting BDNF, but not completely attenuated, and there is a global

45 increase in Ca2+ transients when BDNF is bath applied to hippocampal neurons (Lang et al., 2007).

Moreover, the same study demonstrated that BDNF-mediated local Ca2+ transients occur at directly at synaptic sites (Lang et al., 2007), implying that BDNF signalling is important for dendritic spine stabilization. However, it has also been shown that overexpression of endogenous

BDNF can lead to spine retraction (Horch et al., 1999), though this may be a concentration- dependent effect of BDNF. Interestingly, cyclic AMP (cAMP), a second messenger which activates PKA, regulates BDNF-induced dendritic spine formation (Ji et al., 2005). Therefore, many mediators of synaptic plasticity are themselves important for inducing glutamatergic synapse formation and stabilization.

Interestingly, different modes of application of exogenous BDNF can elicit differential signaling in downstream neurons and can affect the growth of neurites, the precursors to axons and dendrites. An elegant study from the laboratory of Dr. Bai Lu has demonstrated that a gradual increase in BDNF application can sustain activation of TrkB and ERK1/2 for much longer compared to fast acute application, independent of the concentration (Ji et al., 2010). In addition, acute application seemed to induce more primary dendrites that were longer in length, while gradual application seemed to increase neurite ramification (Ji et al., 2010). This is quite interesting given that secreted BDNF can also act as an autocrine factor for axon development

(Cheng et al., 2011), therefore the release and concentration of BDNF are very important for the wide array of functions attributed to BDNF.

The second method through which BDNF enhances synapse formation is by enhancing differentiation of pre- and postsynaptic compartments. BDNF is important for the coordination of presynaptic vesicle pools in terms of the number of presynaptic terminals and the size of the vesicle pool (Vicario-Abejon et al., 2002). More specifically, the formation of docked vesicles is directly

46 influenced by BDNF (Collin et al., 2001; Tyler & Pozzo-Miller, 2001; Marshak et al., 2007), a step which is required for presynapse maturation. This goes hand in hand with the fact that increased amplitudes of AMPA-mediated miniature excitatory postsynaptic currents (mEPSCs) are consistently observed when dissociated hippocampal neurons are chronically treated with

BDNF (Vicario-Abejon et al., 1998; Collin et al., 2001; Tyler & Pozzo-Miller, 2001; Tyler &

Pozzo-Miller, 2003; Copi et al., 2005), which may imply a change in presynaptic release probability. On the postsynaptic side, BDNF can enhance the expression of AMPA receptors thereby also affecting the maturation of dendritic spines.

However, it is important to note that BDNF knockout mice do not demonstrate a significant reduction in the number of glutamatergic synapses in the somatosensory cortex (Korte et al., 1995;

Patterson et al., 1996; Itami et al., 2003). Only when dissociated hippocampal neurons are made from these mice is there reduced synapse formation (Singh et al., 2006). Interestingly, in vivo,

TrkB conditional knockdown at the Schaffer collateral synapse can cause severe CA3-CA1 synapse loss (Luikart et al., 2005), indicating that there is likely compensation by NT-3 or NT4/5 when BDNF is reduced during development (Gottmann et al., 2009).

Many studies have looked at the effects of BDNF on developing excitatory circuits and synapses. Much less is known about the role of BDNF at inhibitory synapses, I will summarize what we do know from these studies in the following section.

1.6.3.2. BDNF and Development of Inhibitory Synapses

The first evidence that BDNF could enhance the maturation of GABAergic synapses came from a study demonstrating that exogenous BDNF could accelerate the maturation of dissociated striatal GABAergic neurons (Mizuno et al., 1994). Moreover, when BDNF was overexpressed in

47 vivo in the visual cortex, it accelerated the maturation of GABAergic synapses and enhanced the number of parvalbumin-positive interneurons (Huang et al., 1999). Giving way to the idea that

BDNF was also important for GABAergic synapse development, not only glutamatergic,

Interestingly, visual deprivation of rodents can slow down the maturation of GABAergic synapses in the visual cortex, which can be rescued with BDNF (Gianfranceschi et al., 2003; Lee et al.,

2006). Given that electrical activity induced by light and eye opening increases BDNF (Castren et al., 1992), this could be the mechanism through which GABAergic synapses mature, at least in the visual system.

Most of these effects are presynaptic in origin. Studies have shown that chronic exogenous application of BDNF can increase the number of GAD65/67-immunpositive GABAergic presynaptic terminals in developing hippocampal neurons (Bolton et al., 2000; Marty, 2000;

Yamada et al., 2002; Ohba et al., 2005), and enhances GABA release (Yamada et al., 2002;

Baldelli et al., 2005; Ohba et al., 2005). How BDNF can regulate GAD67 expression is not well understood, but, this mechanism is so salient that even humans with the Val66Met polymorphism have reduced prefrontal cortex GAD67 mRNA expression (Hashimoto & Lewis, 2006).

On the postsynaptic side, chronic application of BDNF during development can increase the expression of cell surface GABAAR, which increases the amplitude of mIPSC events (Yamada et al., 2002; Mizoguchi et al., 2003a; Mizoguchi et al., 2003b). In accordance with these findings, there is decreased function of GABAergic neurotransmission in heterozygous BDNF+/- mice

(Kohara et al., 2007). From this we could infer that BDNF is important for the appropriate development of GABAergic synapses. However, BDNF also plays a very important role in synaptic plasticity at mature synapses. I will first discuss the role of BDNF at glutamatergic synapses and then at GABAergic synapses.

1.6.3. BDNF and Glutamatergic Plasticity

Rapidly following the discovery of BDNF, many labs demonstrated that the high frequency neuronal stimulation required to induce LTP (tetanus burst stimulation) could increase the expression of BDNF mRNA (Patterson et al., 1992; Castren et al., 1993) and cell surface TrkB receptors (Du et al., 2000). These findings implicated BDNF in LTP, but more enticing was the demonstration by Lohof and Poo showing that the addition of exogenous BDNF can induce a rapid and reversible increase in the frequency of mEPSCs in xenopus nerve muscle preparations (Lohof et al., 1993). This discovery was shortly followed by a flurry of reports showing that neurotrophins can enhance excitatory synaptic transmission at central synapses (Kim et al., 1994; Lessmann et al., 1994; Kang & Schuman, 1995; Levine et al., 1995a) and suppress inhibitory transmission (Kim et al., 1994; Tanaka et al., 1997; Frerking et al., 1998), which will be discussed further in section

1.6.5. These findings are further supported by studies demonstrating that LTP is impaired in both

BDNF-/- and BDNF+/- mice (Korte et al., 1995), a defect that can be rescued by reintroducing

BDNF (Korte et al., 1996). Furthermore, pharmacologically scavenging BDNF with TrkB-Fc

(Figurov et al., 1996; Kang et al., 1997) or anti-BDNF (Chen et al., 1999) also blocks the induction of LTP. All of which unequivocally demonstrated that BDNF is important for LTP, which BDNF promotes through both presynaptic and postsynaptic mechanisms. Since then, many studies have demonstrated a differential role of BDNF depending on whether treatment was chronic or acute. I will attempt summarize the vast number of studies demonstrating the effect of BDNF on excitatory synaptic plasticity, first by discussing presynaptic mechanisms and then postsynaptic in the following subsections.

1.6.4.1. Role of BDNF in Excitatory Presynaptic Plasticity

In the 1990’s many labs provided direct reproducible evidence that acute exogenous application of BDNF could enhance presynaptic release of glutamate by studying evoked glutamatergic synaptic transmission and mEPSCs in dissociated hippocampal neurons (Lessmann et al., 1994; Lessmann & Heumann, 1998; Li et al., 1998; Berninger et al., 1999) or in hippocampal slices (Kang & Schuman, 1995; Tyler et al., 2006). In addition, increasing endogenous levels of

BDNF released from neurons in response to high-frequency stimulation can also enhance glutamate release (Magby et al., 2006; Shen et al., 2006). These effects of BDNF are thought to be mediated by TrkB-dependent phosphorylation of Rab3a and synapsin I, proteins which are associated with neurotransmitter vesicle release (Jovanovic et al., 2000; Thakker-Varia et al.,

2001). Accordingly, it seems that the presynaptic effects of BDNF on glutamate release rely on the enhancement of vesicle turnover at glutamatergic presynaptic terminals.

1.6.4.2. Role of BDNF in Excitatory Postsynaptic Plasticity

Postsynaptic enhancement of glutamatergic synapses by BDNF is well-reported in the literature. The first reported effects of postsynaptic plasticity came from dissociated hippocampal neurons where acute BDNF treatment led to an increase in glutamatergic synaptic transmission, namely an increase in spontaneous firing and increase in the frequency and amplitude of excitatory postsynaptic currents (Levine et al., 1995a; b). This enhancement arises from TrkB/MAPK- dependent phosphorylation of GluN2B, which prolonged channel opening (Levine et al., 1995a; b; Lin et al., 1998; Crozier et al., 1999). This is just one of many postsynaptic mechanisms through

BDNF acts. It can also enhance the formation of new synapses (Vicario-Abejon et al., 1998; Shen et al., 2006; Walz et al., 2006), modulate AMPA receptors (Lessmann & Heumann, 1998;

Rutherford et al., 1998), induce insertion of new glutamate receptors (Itami et al., 2003; Caldeira et al., 2007a; Caldeira et al., 2007b) and increase intracellular Ca2+ (Berninger et al., 1993;

Kovalchuk et al., 2002; Amaral & Pozzo-Miller, 2007a; b). All of which leads to an increase in mEPSC amplitude and frequency (Akaneya et al., 1997; Patterson et al., 2001).

Given the strong evidence that BDNF can modulate glutamatergic synapses pre- and postsynaptically, it is no surprise that LTP is impaired in area CA1 of BDNF-/- mice (Korte et al.,

1995). Using CA1-specific BDNF knockdown, most studies suggest BDNF induces a presynaptic effect on LTP (Xu et al., 2000; Zakharenko et al., 2003). However, one study demonstrated by transfecting a dominant-negative PLCγ construct into acute hippocampal slices that hippocampal

LTP is supported by both pre- and postsynaptic mechanism through BDNF/TrkB/ PLCγ signaling

(Gartner et al., 2006). Thus, BDNF is a very important modulator of glutamatergic synaptic plasticity.

1.6.5. BDNF and GABAergic plasticity

BDNF can have differential effects on GABAergic synaptic transmission depending on the time and nature of BDNF exposure. When BDNF is acutely applied (a few minutes), many scientists have reported a reduction of GABAergic neurotransmission in dissociated hippocampal neurons (Brunig et al., 2001) and in acute slices (Tanaka et al., 1997; Frerking et al., 1998) when recording from excitatory neurons, which was dependent on TrkB-mediated PKC activation

(Tanaka et al., 1997; Frerking et al., 1998) and subsequent internalization of GABAAR subunits

(Brunig et al., 2001). Conversely, some reports demonstrate that BDNF can actually cause an enhancement of GABAergic neurotransmission of inhibitory interneurons, mainly through a presynaptic mechanism (Mizoguchi et al., 2003a; Wardle & Poo, 2003; Jovanovic et al., 2004).

This data implies that BDNF may play a role in homeostatic mechanisms of maintaining a tight- balance between excitation and inhibition when released acutely.

Fewer studies have looked at chronic effects of exogenous BDNF exposure and most have been carried out in dissociated hippocampal neurons. They demonstrate that chronic exposure to

BDNF can actually increase the number of GABAergic synapses (Bolton et al., 2000; Aguado et al., 2003) and increase GABAergic neurotransmission (Bolton et al., 2000; Paul et al., 2001;

Baldelli et al., 2002; Palizvan et al., 2004). However, on the contrary reduced levels of BDNF in heterozygous BDNF+/- mice enhances GABAergic neurotransmission rather than reduces it

(Olofsdotter et al., 2000; Henneberger et al., 2002; Henneberger et al., 2005). Therefore, the chronic effect of BDNF on GABAergic synapses is conflicting, which can make it difficult to interpret the exact role of BDNF at inhibitory synapses. Both approaches have their limitations, dissociated cultures lack an intact network that is comparable to in vivo and therefore may over or under emphasize the effect of BDNF. On the other hand, a chronic reduction of BDNF in vivo may introduce compensatory mechanisms which may mask the role of BDNF at GABAergic synapses.

Therefore, more studies on the chronic effect of BDNF on GABAergic studies should be undertaken in organotypic hippocampal slices.

BDNF has been implicated in many neurological conditions in which both glutamatergic and GABAergic synaptic plasticity is disrupted, for example in epilepsy and stroke. It is well know that neurotrophins are highly upregulated after brain insults (Lindvall et al., 1994), though, what role they are playing in synaptic function after injury is still poorly understood. In the following section I will discuss the evidence linking BDNF to these conditions and highlight what we still do not know.

1.6.6. BDNF and Brain Injuries

Up to this point, I have introduced the physiology of the developing and mature hippocampus and I have highlighted how plasticity can be maladaptive after acquired brain injuries, such as epilepsy and stroke. It has been hypothesized that following injury, the central nervous system may revert to a developmental state in order to recover cognitive function following injury and in the process of doing so may activate these maladaptive processes. Many theories suggest that BDNF may be a potential therapeutic target to treat these neurological conditions, but prior to developing therapies directly targeting BDNF-TrkB signaling, we must understand the role of BDNF in injury-induced microcircuitry remodeling.

Studies have demonstrated an increase in BDNF mRNA in dentate gyrus granule cells and pyramidal neurons (Ernfors et al., 1991; Isackson et al., 1991; Timmusk et al., 1993; Nibuya et al., 1995) following induction of epilepsy in rodent. This is coupled with an increase in TrkB protein expression (Ampuero et al., 2007) and TrkB phosphorylation (Danzer et al., 2004).

Moreover, increased BDNF levels are also observed in the hippocampi of humans with epilepsy

(Mathern et al., 1997; Takahashi et al., 1999). Given how salient this BDNF upregulation is, it led to the idea that BDNF could contribute to the lasting structural and functional changes underlying epileptogenesis (Gall et al., 1991; Gall et al., 1997; Jankowsky & Patterson, 2001; Binder &

Scharfman, 2004). Since then, it has emerged that BDNF itself is a potent inducer of seizures and epileptogenesis, as transgenic mice that overexpress BDNF develop spontaneous seizures (Croll et al., 1999) and intra-hippocampal injection of BDNF is sufficient to induce seizure activity in vivo (Scharfman et al., 2002; Binder & Scharfman, 2004). Conversely, heterozygous BDNF+/- mice have decreased seizure susceptibility (Kokaia et al., 1995). However, how BDNF confers epileptogenesis is not well understood.

So far, two hypotheses have emerged: (1) elevated BDNF following initial precipitating injury, either first seizure or head trauma, can lead to structural changes that gives rise to epilepsy and (2) increased BDNF potentiates glutamatergic transmission, thereby increasing excitatory drive in hippocampal circuits and thus contributes to epileptogenesis (Scharfman, 2005). The first hypothesis relies largely on the fact that it is well known that BDNF facilitates dendritic spine growth, axon remodeling and neurogenesis in the hippocampus (see section 1.6.3.1.). A definitive association between BDNF and epileptogenesis has yet to be demonstrated (Scharfman, 2005).

Experimental traumatic brain injury inducing penetrating trauma to rodents demonstrates an increase in BDNF mRNA and protein for several hours in the hippocampus (Hicks et al., 1997;

Grundy et al., 2000) and cortex (Nieto-Sampedro et al., 1982; Oyesiku et al., 1999; Truettner et al., 1999; Griesbach et al., 2002). However, up until recently the role of TrkB receptors was less understood. Axonal sprouting following Schaffer collateral lesion, a model of post-traumatic epilepsy, is dependent on TrkB receptors (Dinocourt et al., 2006; Aungst et al., 2013). Moreover, this type of lesion leads to an increase in population activity in areas CA1 and CA3 following the relief of inhibitory drive with a low dose of bicuculline and this is also dependent on TrkB activation (Dinocourt et al., 2011; Aungst et al., 2013). Whether this plasticity is mediated through

BDNF and whether BDNF expression is increased by Schaffer collateral lesion mimicking a traumatic brain injury is not known. In addition, whether structural modifications of synapses can actually lead to CA3 neuronal hyperexcitability is still poorly characterized.

These key pieces of evidence lead us to believe that, at least in the hippocampus, TrkB negatively influences hyperexcitability and seizure-like activity. Though, it is also important to note that in the cortical undercutting model, GABAergic neuron survival depends on BDNF-TrkB and is important to confer protection from epileptogenesis (Prince et al., 2009; Scharfman, 2013).

This suggests a dual role of BDNF in epileptogenesis and thus we must better understand how

BDNF alters structural remodeling or potentiation of glutamatergic transmission following injury.

In addition, the signaling cascade that links injury to BDNF upregulation is unknown. The prevailing theory suggests that injury induces an increase in glutamate release and Ca2+ (Arundine

& Tymianski, 2004), which can potentially overproduce BDNF through the regulated pathway of activity-dependent transcription and release. However, there is little evidence to confirm this hypothesis. Lesion of mossy fibers at the dentate gyrus can lead to an increase in Ca2+ as demonstrated by the use of Ca2+-sensitive dyes (Muller et al., 2010). However, whether lesion- induced Ca2+ increases can activate MeCP2 de-repression of bdnf mRNA transcription is unknown.

Interestingly, BDNF expression has also been shown to be upregulated after ischemia

(Lindvall et al., 1994), especially within the hippocampus (Lindvall et al., 1992; Takeda et al.,

1993; Bejot et al., 2011b; Gottlieb et al., 2013; Neumann et al., 2015) and cortex (Comelli et al.,

1993) following transient forebrain ischemia. It has also been shown that hippocampal TrkB receptor expression can increase after forebrain ischemia (Merlio et al., 1993). However, the role of BDNF in ischemia remains controversial as there is conflicting evidence on whether this increase in BDNF is detrimental or protective. Intravenous administration of BDNF (Schabitz et al., 2007) or overexpression of BDNF with viral vectors (Yu et al., 2013) has been shown to aid in the recovery of motor function in rodents after stroke. This data suggests that elevated BDNF is beneficial in recovery after stroke. However, other studies using similar stroke models, with either heterozygous BDNF+/- transgenic mice (Nygren et al., 2006) or mice expressing the Val66Met polymorphism (Qin et al., 2014), which have reduced levels of BDNF release (Egan et al., 2003),

55 demonstrate that less BDNF has a positive outcome on motor recovery. Given this discrepancy, there is clearly a need to understand the role that BDNF plays at synapses following ischemia.

BDNF can significantly alter glutamatergic and GABAergic neurotransmission, both of which are affected in stroke and epilepsy. Moreover, patients with stroke can develop epilepsy, sometimes even years after initial insult (Camilo & Goldstein, 2004; Chang et al., 2014;

Wannamaker et al., 2015). Given that BDNF is upregulated in these conditions, one may speculate that BDNF may play a role in comorbidity of epilepsy and stroke.

Interestingly, bipolar disorder, which is linked to misregulation of BDNF (Neves-Pereira et al., 2002; Sklar et al., 2002; Green & Craddock, 2003; Binder & Scharfman, 2004; Tsai, 2004b), can be comorbid with both stroke (Santos et al., 2011; Carota & Bogousslavsky, 2012; Hackett et al., 2014) and epilepsy (Gaitatzis et al., 2004). It may be that these comorbidities occur in a subset of patients that have altered BDNF signaling. In addition, depression is also linked to BDNF misregulation (Tsai, 2004a; Hosang et al., 2014). Both stroke (Dwyer Hollender, 2014; Hackett et al., 2014) and epilepsy (Tellez-Zenteno et al., 2007) patients can develop depression. However, how BDNF dysregulation affects mood disorders is still a matter of speculation (Binder &

Scharfman, 2004) and even less understood is the role of BDNF in comorbidities between stroke, epilepsy and mood disorders. However, it may be that dysregulation of BDNF can induce long- term microcircuitry changes in the central nervous system thereby increasing a patient’s probability of developing one of these neurological conditions. Therefore, understanding the cellular mechanisms the role of BDNF in structural remodeling of synapses may lead to new therapies to treat many different neurological conditions. Throughout my thesis I have focused on the role of BDNF in synapse remodeling after injury to the hippocampus and specifically, I have studied the following objectives:

Chapter 2. The role of BDNF and its receptors in ischemia-induced morphological and

functional microcircuitry remodeling.

Chapter 3. The role of BDNF in Schaffer collateral lesion-induced morphological and

functional microcircuitry remodeling.

Chapter 4. The role of activity-dependent de-repression of MeCP2 on Schaffer collateral

lesion-induced microcircuitry remodeling.

In order to carry out this work, I used organotypic hippocampal slice preparations, which are a reduced model of the hippocampus. I will discuss the use of these cultures in the last section.

1.7. ORGANOTYPIC HIPPOCAMPAL SLICE CULTURES

As outlined in previous sections, there are many reasons why neuroscientists study the hippocampus. Firstly, it has a unique anatomy that is well structured and thoroughly characterized.

Secondly, it is a very important brain structure for learning and memory. Lastly, the connectivity of the neurons is mostly unidirectional (Amaral & Lavanex, 2007), which renders it a suitable model for the study of synaptic plasticity. Given that the hippocampus is in a deep-lying brain region, it is difficult to functionally and morphologically study these synapses in vivo.

In response to this, many neuroscientists have been removing the brain of rodents and creating tissue preparations in order to study neuronal function. There are several different methods to do this, the three most widely used are: (1) acute hippocampal slices, (2) organotypic hippocampal slice preparations or (3) dissociated primary neuronal cultures.

Dissociated neuronal cultures, involve mechanical disruption of isolated hippocampi and plating them in culture dishes. One advantage to this method is that this renders the neurons to a lower density population of cells which allows easy transfection of plasmid DNA to study the effect of one protein in a cellular signalling cascade (Goslin et al., 1998). However, on the other hand, given the lower density of neurons being plated in this method, it is more difficult to study the network-driven events and electrophysiological activity that underlie synaptic plasticity.

Moreover, these cultures lose the organized synaptic network seen in vivo (Papa et al., 1995).

Therefore, it can be difficult to study hippocampal injuries that are driven by whole network events, such as in epilepsy. On the other hand, neurons in acute slices are more accessible for electrophysiological or morphological study. However, it is very difficult to keep these slices alive for any length of time greater than 12 hours (Ankri et al., 2014; Buskila et al., 2014).

Therefore, since I wished to perform long-term studies in an intact network, I utilized organotypic slices throughout this thesis, which are much more feasible for this type of study.

When organotypic slices are prepared, initially there is reactive axonal sprouting that occurs, however after 3 week in vitro the synapses stabilize and the network closely resembles in vivo

(Frotscher & Gahwiler, 1988; Gahwiler et al., 1997). At this time point organotypic slices maintain an active network in which you can induce cellular models of learning and memory, such as long- term potentiation (LTP) and long-term depression (LTD) similar to acute hippocampal slices

(Bonhoeffer et al., 1989; Debanne et al., 1994). In addition, the density of excitatory synaptic connections between neurons, namely post-synaptic excitatory dendritic spines, is similar to postnatal day 15 in rodents (McKinney 1999). This suggests that ‘mature’ organotypic slices form stable synapses which can undergo synaptic plasticity (Gahwiler et al., 1997). Moreover, given that organotypic cultures can be kept in vitro for several weeks to months, I could pursue long-

58 term studies with pharmacological and genetic manipulations. Organotypic hippocampal slice cultures have been widely used for studies of brain injury and repair (Noraberg et al., 2005). The preparation of these cultures is described in detail in sections 2.3.2., 3.3.2, and 4.3.2.

Given that organotypic hippocampal slice cultures are maintained in a controlled environment, I am able to use different methods to induce brain injury while closely resembling an in vivo injury. Here I used two established methods: (1) a lesion of the Schaffer collaterals in order to mimic a penetrating head trauma which can cause the formation of recurrent hyperexcitable network within area CA3 (McKinney et al., 1997) and (2) oxygen-glucose deprivation (OGD), an in vitro model of ischemia (Gee et al., 2006; Gee et al., 2010). Interestingly, when exposed to OGD, CA1 pyramidal neurons are the most susceptible to damage as seen in vivo

(Newell et al., 1995; Bonde et al., 2005).

Using these methods throughout my thesis I asked how synapses and microcircuitry evolved after brain injury, taking in to consideration the aims outlined in section 1.6.6. and looking at areas CA3 (presynaptically) and CA1 (postsynaptically) of the hippocampus. In the next chapter,

I will use ischemia to organotypic hippocampal slice cultures to dissect the role of BDNF in hippocampal synaptic remodeling.

CHAPTER 2. BDNF AND PROBDNF DIFFERENTIALLY REGULATE ISCHEMIA-INDUCED PLASTICITY OF GABAERGIC AND GLUTAMATERGIC SYNAPSES OF CA1 PYRAMIDAL NEURONS

FOREWORD

It is well known that ischemia can downregulate both glutamatergic and GABAergic synapses and much research has been devoted to finding novel therapeutic interventions that will protect synapses after ischemia. In Chapter 2, I set out to understand some of the cellular mechanisms behind ischemia-induced synapse loss. Previously, it has been shown that BDNF is upregulated in patients and in rodent models of stroke and brain trauma. There is a discrepancy in the literature as to whether this increase in BDNF is neuroprotective or detrimental. Clearly, the exact role of BDNF in ischemia is still poorly understood. Under physiological conditions it has been shown that BDNF-TrkB negatively regulates the functionality of GABAergic synapses under physiological conditions. While, its precursor protein proBDNF through p75NTR can decrease dendritic spine density.

Given this, I hypothesized that following an in vitro model of ischemia mature BDNF binding through TrkB receptors can downregulate GABAergic synapses and proBDNF binding through p75NTR receptors can downregulate glutamatergic synapses.

To test my hypothesis, I subjected organotypic cultures to oxygen-glucose deprivation and used electrophysiology, confocal microscopy and molecular biology to better understand which mechanisms lead to the specific disruption of either glutamatergic or GABAergic synapses. The work presented in this chapter has been recently submitted to a peer reviewed journal.

2.1. ABSTRACT

Following ischemia, both glutamatergic and GABAergic synapses can be disrupted in the brain. However, the cellular mechanism underlying ischemia-induced synapse loss is poorly understood. The neurotrophin BDNF, important for synaptic plasticity, is upregulated in stroke patients and in rodent models of ischemia. Intriguingly, BDNF has been described as both neuroprotective and detrimental to recovery following ischemia in rodents. Given this discrepancy, there is a clear need to better understand if and how BDNF, either the proprotein form proBDNF or mature BDNF, disrupts synapses after ischemia. Using an established ischemia model of oxygen-glucose deprivation (OGD) in mouse organotypic hippocampal slices we report that proBDNF specifically disrupts glutamatergic synapses via p75NTR and mature BDNF downregulates GABAergic synapses via TrkB. In addition, we identified that second messengers

ERK1/2 and GSK3β, downstream of TrkB, can prevent GABAergic synapse loss following OGD.

Ischemia is a complex pathology and our data simplifies this complexity by identifying the molecular basis for observed plasticity changes at both inhibitory and excitatory synapses

2.2. INTRODUCTION

It is well established that following ischemic brain injury there is an interruption of oxygen and nutrient supply to neurons, which can disrupt synaptic integrity. Interestingly, both glutamatergic dendritic spines (Hasbani et al., 2001; Zhang et al., 2005) and GABAergic synapses

(Sette et al., 1993; Alicke & Schwartz-Bloom, 1995) are known to be affected by ischemia, leading to cognitive deficits in memory and function. After an ischemic event, the CNS activates certain plasticity mechanisms in an attempt to maintain a balance between excitation and inhibition, which can lead to comorbidities of epilepsy and depression in patients with ischemic stroke.

In order to protect synapses, many preclinical neuroprotective therapies have been developed, including therapies targeting: glutamate receptors (Davis et al., 2000b; Albers et al.,

2001; Saver et al., 2015), GABA receptors (Lyden et al., 2001), calcium channels (Horn et al.,

2001), sodium channels (Gibson et al., 2010) and free radicals (Diener et al., 2008) among others, but have failed to show efficacy in clinical trials. Therefore, there exists a need to better understand the cellular mechanisms underlying synaptic disruption after ischemia.

Following ischemia in rodent models, the expression of brain-derived neurotrophic factor

(BDNF), a neurotrophin that plays a very important role in synaptic plasticity and cell survival

(Park & Poo, 2013), has been shown to be upregulated (Lindvall et al., 1992; Takeda et al., 1993;

Bejot et al., 2011a; Gottlieb et al., 2013). However, there is conflicting evidence on whether this increase in BDNF is detrimental or protective. Many groups have administered BDNF, either intravenously (Schabitz et al., 2007) or with viral vectors (Yu et al., 2013), and have seen a recovery of motor function in rodents after stroke. These findings suggest that elevated BDNF is beneficial in recovery after stroke. However, other studies using similar stroke models, with either heterozygous bdnf-/+ transgenic mice (Nygren et al., 2006) or mice expressing the Val66Met

62 polymorphism (Qin et al., 2014), which has reduced levels of BDNF release (Egan et al., 2003), demonstrate that less BDNF has a positive outcome on motor recovery. Given this discrepancy, there is a need to better understand the role that BDNF might play at synapses following ischemia.

BDNF is expressed as a proprotein, proBDNF, and emerging evidence suggests that mature

BDNF (mBDNF) and proBDNF play opposing roles in synapse function (Deinhardt & Chao,

2014). It is widely thought that mBDNF, binding to TrkB receptors, can positively regulate excitatory synapses (Tanaka et al., 2008). mBDNF can also bind to p75NTR receptors with much lower affinity (Rodriguez-Tebar et al., 1990; Huang & Reichardt, 2001; Gentry et al., 2004).

Whereas proBDNF, binding with high affinity to p75NTR receptors, is thought to negatively regulate dendritic spines (Zagrebelsky et al., 2005; Yang et al., 2014). On the other hand, mBDNF is known to induce the internalization of GABAA receptor subunits (Brunig et al., 2001; Mou et al., 2013) and gephyrin, the main postsynaptic scaffolding protein at inhibitory synapses. Gephyrin degradation can occur by second messengers downstream of mBDNF (Tyagarajan et al., 2013).

Taken together, we hypothesize that mBDNF-TrkB signaling leads to the degradation of inhibitory synapses and proBDNF-p75NTR signaling degrades excitatory synapses after ischemia.

To test our hypothesis, we subjected mouse organotypic hippocampal slice cultures to oxygen-glucose deprivation (OGD), an established in vitro model of ischemia. To study morphological remodeling of synapses, we used confocal microscopy to image excitatory dendritic spines from CA1 pyramidal neurons expressing membrane-tagged GFP and used immunohistochemistry for gephyrin to image inhibitory synapses. Finally, we used electrophysiology to assess the functionality of excitatory and inhibitory synapses.

Here, we report that mBDNF via TrkB disrupts GABAergic synapses and proBDNF via p75NTR disrupts glutamatergic synapses after OGD.

2.3. MATERIALS AND METHODS

2.3.1. Ethics Statement

All animal handling procedures were carried out consistent with guidelines set by the

Canadian Council on Animal Care and the National Institutes of Health in the US. All procedures were approved by the Animal Resource Committee of the School of Medicine at McGill University and are outlined in McGill University Animal Handling Protocol #5057.

2.3.2. Hippocampal Slice Cultures and Oxygen-Glucose Deprivation

We have chosen to study the hippocampus as it possesses a unique unidirectional network that is preserved within the organotypic culture system (Gahwiler et al., 1997), making it an ideal candidate to study microcircuitry remodeling. Organotypic hippocampal slices were prepared using the roller-tube method, as previously described (Gahwiler, 1981; Gahwiler et al., 1997).

Briefly, 400 μm thick hippocampal slices were made from postnatal day 6-8 transgenic mice expressing membrane-targeted MARCKS-enhanced green fluorescent protein (GFP) under the

Thy-1 promoter in a subpopulation of CA1 cells (De Paola et al., 2003) and plated on glass coverslips. Slices were adhered in place with a clot of chicken plasma (Cocalico Biologicals;

Reamstown, PA, USA) coagulated with thrombin (Invitrogen). Coverslips were placed in flat- sided culture tubes with antibiotic-free serum-containing media and maintained in a dry-air roller drum incubator at 36C for three weeks prior to experimentation. An established model of oxygen- glucose deprivation (OGD) was used as a model of ischemia (Gee et al., 2006), briefly, mature organotypic hippocampal slices were placed in a glass dish containing glucose-free Tyrode’s

64 solution (in mM: NaCl, 137; KCl, 2.7; CaCl2, 2.5; MgCl2, 2; NaHCO3, 11.6; NaH2PO4, 0.4; pH

7.4) containing: 2 mM 2-Deoxyglucose, 3 mM sodium azide and 8 mM sucrose, for 4-5 minutes, and were then returned to normal culture media for 90 minutes, 24 hours or 1 week. Control slices were exposed to Tyrode’s solution (in mM: NaCl, 137; KCl, 2.7; CaCl2, 2.5; MgCl2, 2; NaHCO3,

11.6; NaH2PO4, 0.4; glucose, 5.6; pH 7.4) for 4-5 minutes and returned to culture media.

2.3.3. Propidium Iodide Staining, Immunofluorescence and Confocal Microscopy

90 minutes or 1 week following OGD induction, hippocampal slices were placed in 5 µM propidium iodide diluted in culture media for 15 minutes, after which slices were fixed. Slices were fixed for 1 hour at room temperature in 4% paraformaldehyde made up in 0.1 M phosphate buffer (PB), pH 7.4. Slices were then mounted using fluorescent mounting medium (Dako Canada;

Mississauga, Canada).

For immunofluorescence, 90 minutes or 24 hours following OGD, slices were fixed with

4% paraformaldehyde and washed in 0.1 M PB several times. Then, they were permeabilized in

0.4% Triton X-100 and blocked with 1.5% heat inactivated horse serum overnight at 4°C. Primary antibodies were incubated for five days at 4°C in permeabilizing buffer. The following antibodies were employed: DyLight 550 anti-HIF1α tagged (1:250, Pierce ThermoScientific), Alexa Fluor

488 tagged anti-NeuN (1:250, Millipore), anti-Gephyrin (1:500, Synaptic Systems), anti-α1

GABAAR and anti-α2 GABAAR (1:3000, both kind gifts from Dr. J.-M. Fritschy, University of

Zurich, Zurich, Switzerland; for more information see (Fritschy & Mohler, 1995) and anti-p75NTR

(kind gift from Dr. P, Barker, McGill University, Montreal, Quebec). Following several washes with 0.1 M PB containing 1.5% heat inactivated horse serum, when necessary slices were

65 incubated overnight at 4°C diluted in 0.1 M PB containing 1.5% heat inactivated horse serum. One of the following secondary antibodies: 1:250 anti-mouse DyLight 650 (Jackson ImmunoResearch,

Burlington, ON, Canada), 1:250 anti-guinea pig DyLight 550 (Jackson ImmunoResearch,

Burlington, ON, Canada) or 1:250 anti-rabbit Alexa Fluor 594 (Invitrogen Molecular Probes).

Following several washes with 0.1 M PB containing 1.5% heat inactivated horse serum, slices were mounted with fluorescent mounting medium onto microscope slides.

Following mounting, slices were imaged in area CA1 or CA3 using a Leica TCS SP2 scan head (Leica Microsystems) on a Leica DM6000 B upright microscope, equipped with a HCX PL

APO 63× NA 1.4 oil immersion objective using a 543 nm HeNe laser line. Image stacks were collected at Z = 0.3 μm and averaged 2-3 times to improve signal-to-noise ratio. For quantification, image stacks were obtained with identical parameters (laser intensity, filters, pinhole size, photomultiplier tube gain and offset). Representative images are maximum intensity projections of 5 sections from each stack.

2.3.4. Dendrite Reconstructions, Spine Quantification and Puncta Quantification

Three-dimensional confocal stacks were deconvolved with Huygens Essentials software

(Scientific Volume Imaging, Hilversum, The Netherlands) using a full maximum likelihood extrapolation algorithm. Stacks were then imported and rendered using the Surpass function in

Imaris software (Bitplane AG). Spine detection/classification program automatically detected the length of the spine head and neck. From the ratio of the diameter and the length of the head and the neck of the spines, it was possible to distinguish the stubby, mushroom, long, thin spines, dendrite branches, and filopodia. These classifications were based on previously established

66 criteria (McKinney et al., 1999a; McKinney, 2010). Lastly, n values for spine analysis represent

~75-100 µm of dendrite from 1-2 cells imaged from each slice. The number and volume of gephyrin puncta were quantified using the Spot function of Imaris software, which calculates differentiates puncta based on the fluorescence intensity.

2.3.5. Electrophysiological Recordings and Analysis

Slices were transferred into a temperature-controlled chamber (25°C) mounted on an upright microscope (DM LFSA, Leica Microsystems) and continuously perfused with external solution. Patch recording electrodes were pulled from borosilicate glass (GC150TC; Clark

Instruments, Old Sarum, Salisbury UK). All electrophysiological recordings were made using an

Axopatch 200A amplifier (Molecular Devices, Sunnyvale, CA, U.S.A.).

AMPA-mediated mEPSCs were gathered from whole-cell voltage-clamp recordings of

CA1 pyramidal neurons obtained at 25°C using electrodes with resistances of 4-5 MΩ and filled with intracellular solution containing (in mM): K-Gluconate, 120; EGTA, 1; HEPES, 10; Mg-

ATP, 5; Na-GTP 0.5; NaCl, 5; KCl, 5; phosphocreatine, 10; 295 mOsm; pH adjusted with KOH to 7.3. mEPSCs were recorded at -60 mV and in the presence of 1 µM tetrodotoxin (TTX), 15 µM

3-[(R)-2-carboxypiperazin-4-yl]-propyl-1-phosphonic acid (CPP), 100 µM picrotoxin, and 1 µM

CGP55845 in the external Tyrode solution. Access resistance was monitored with brief test pulses at regular intervals (2-3 minutes) throughout the experiment. Access resistance was usually 10-13

MΩ and data were discarded if the resistance deviated more than 10% through the course of the experiment. Series resistance of the access pulse and decay time was also used for the calculation of total membrane capacitance. After the holding current had stabilized, data were recorded at a

67 sampling frequency of 10 kHz and filtered at 2 kHz for 10 to 15 minutes. We found that the input resistance and resting membrane potential of OGD or control cells did not differ significantly.

GABAAR-mediated mIPSCs were gathered from whole-cell voltage-clamp recordings of

CA1 pyramidal neurons obtained at 25°C using electrodes with resistances of 4-5 MΩ and filled with intracellular solution containing (in mM): CsCl, 140; NaCl, 4; 0.5, CaCl2; HEPES, 10;

EGTA, 5; QX-314, 2; Mg-ATP, 2; Na-GTP 0.5; 290 mOsm; pH adjusted with CsOH to 7.36. mIPSCs were recorded at -60 mV and in the presence of 1 µM TTX, 25 µM CPP, 5 µM CGP55845,

5 µm 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX) and 0.3 µm strychnine in external Tyrode solution. Access resistance was monitored with brief test pulses at regular intervals (2-3 minutes) throughout the experiment. After the holding current had stabilized, data were recorded at a sampling frequency of 10 kHz and filtered at 2 kHz for 10 to 15 minutes.

All mEPSCs and mIPSCs were detected offline using the Mini Analysis Software

(Synaptosoft, Decatur, GA, USA). The amplitude threshold for mEPSC and mIPSCs detection was set at four times the root-mean-square value of a visually event-free recording period. From every experiment, 5 minutes of stable recording was randomly selected for blinded analysis of amplitude and inter-event interval. The data obtained was then used to plot cumulative histograms with an equal contribution from every cell. For statistical analysis, data were averaged for every single cell. It should be noted that the amplitude analysis was conducted only on single mEPSCs and mIPSCs that did not have subsequent events occurring during their rising and decaying phases.

For frequency analysis, all selected events were considered.

2.3.6. RT-qPCR

90 minutes following OGD, areas CA1 and CA3 were microdissected, snap frozen on dry ice and preserved in RNAlater (Ambion Inc., Austin, TX, USA) until experimentation. 5 slices were used for each experimental group. Each condition was tested with 4 biological samples in triplicate. Total RNA was extracted using Aurum™ Total RNA Mini Kit (BIO-RAD, Mississauga,

ON, Canada) according to the manufacturer’s protocol and 1 μg of RNA was reverse transcribed to cDNA using 20U of AMV reverse transcriptase (Roche Diagnostics), as per manufacturer's recommendation. Subsequently, qPCRs were performed using 30 ng of cDNA in a 20 μl reaction with EVA green mastermix (Solis BioDyne). Primers for bdnf mRNA were designed to target all

BDNF transcript variants (forward primer: 5'-TGC AGG GGC ATA GAC AAA AGG-3' and reverse primer: 5'-CTT ATG AAT CGC CAG CCA ATT CTC-3') and gapdh mRNA (forward primer: 5'-TGC CCC CAT GTT TGT GAT G-3' and reverse primer: 5'-TGT GGT CAT CAG CCC

TTC C-3') was used as an internal control. All qPCR reactions were performed under the following condition conditions: 40 cycles; denaturation at 95°C for 15s, annealing at 62°C for 25s and extension at 72°C. All qPCR data was quantified by ∆∆Ct method and normalized to gapdh.

2.3.7. Pharmacological Treatments

To scavenge BDNF, slices were treated with TrkB-Fc (R&D Systems; Minneapolis, MN,

USA), a fusion protein in which the BDNF binding site of the TrkB receptor replaces the Fc fragment of a human IgG1 antibody. We found that TrkB-Fc treatment to hippocampal cultures for 24 hours downregulated TrkB receptor phosphorylation (data not shown). TrkB-Fc was diluted in culture media at a final concentration of 10 μg/ml and treatment began immediately following

69 induction of OGD. ERK activation was inhibited using 30 µM of MEK inhibitor PD98059 (Tocris

Biosciences, ON, Burlington, Canada), GSK3β activity was inhibited using 25 µM GSK3β-IX

(Tocris Biosciences) and calpain activity was inhibited using 30 µM MDL28170. MDL28170,

PD98059 and GSK3β-IX were diluted in dimethyl sulfoxide (Invitrogen) and treatment began 16 hours prior to OGD induction and continued for 90 minutes after. Control sister cultures were treated with control culture media containing dimethyl sulfoxide only. The following function blocking antibodies were used: 1:200 anti-p75NTR (kind gift from Dr. P. Barker, McGill University,

Montreal, QC, Canada; Rex antibody for more information see (Weskamp & Reichardt, 1991),

1:200 anti-proBDNF (kind gift from Dr. Philip Barker, McGill University, Montreal, Canada) and

1:100 anti-BDNF (N-9, Developmental Studies Hybridoma Bank, University of Iowa, IA, USA; for more information see (Kolbeck et al., 1999). Function blocking antibody treatment began 2 hours prior to OGD induction and continued for 90 minutes after.

2.3.8. Biolistic Gene Transfection

Cartridges were prepared according to manufacturer’s protocol (Bio-Rad, Helios Gene

Gun). Briefly, 15 mg of gold particles (1 μm diameter) were first coated with 0.05 M spermidine.

15 μg of plasmid DNA expressing tdTomato and 45 μg of wildtype gephyrin-GFP (gephyrinWT-

GFP) or dephosphorylation mutant gephyrin-GFP S268A/S270A (gephyrinS268A/S270A-GFP).

Plasmids were then precipitated onto the particles by adding CaCl2. The coated particles were resuspended into 100% ethanol and infused into Tefzel tubing, which were then coated with the particles. Coated tubing was cut into 0.5 inch cartridges which were then transfected into mature organotypic slice cultures by shooting at a distance of 2 cm with a pressure of 200 psi through a

70 nylon mesh. Following 48 hours, slices which expressed target plasmids in CA1 pyramidal neurons were processed with OGD or control Tyrode solution.

2.3.9. Statistical Analysis

Comparison between two groups were made using two-tailed independent Student’s t-test.

Comparisons between multiple groups were made using one-way ANOVA with post hoc Tukey's test. Cumulative probability plots were compared using Kolmogorov-Smirnov (K-S) test for probability distributions. Results are expressed as mean ± S.E.M.

2.4. RESULTS

Here, we set out to study some of the mechanisms which underlie synaptic dysfunction following ischemia. Previously, it has been shown mBDNF protein expression is upregulated following stroke in rodents (Bejot et al., 2011a; Bejot et al., 2011b; Choi et al., 2015) and in patients (Yang et al., 2011). We wish to determine if this increase is detrimental or protective after hippocampal injury and CA1 synapse remodeling.

2.4.1. Area CA1 is selectively affected following OGD in organotypic hippocampal slices

First, we tested an established model of OGD in organotypic hippocampal slice cultures

(Gee et al., 2006; Gee et al., 2010) and we found an increase in the expression of a known marker of cell death, propidium iodide, within area CA1 in 40% of cultures 90 minutes following OGD

(Fig. 2.1A; n = 4 slices presenting cell death out of 10 slices). This propidium iodide-positive immunolabeling was comparable in area CA3 or dentate gyrus of control and 90 minutes post-

OGD cultures (Fig. 1A). Control cultures did not depict any cell death (n = 10 slices) and propidium iodide staining was comparable to control cultures 1 week following OGD (Fig. 2.1A; n = 5 slices). Given that 4 out of 10 cultures had propidium-iodide positive cell death we used a known marker for ischemia to confirm if we had indeed induced hypoxia, by immunolabeling for hypoxia-inducible factor 1 α (HIF1α), a known marker for ischemia (Bergeron et al., 1999). We found a significant 1.5-fold increase in HIF1α expression in area CA1 90 minutes after OGD (Fig.

2.1B, C; n = 5 slices) compared to CA1 in sister control cultures (n = 4 slices). Moreover, CA3

HIF1α expression in OGD-treated slices was comparable to control CA1 (*p < 0.05, two-tailed independent Student’s t-test). These findings are consistent with known models of ischemia and

Figure 2.1 – OGD selectively affects area CA1. (A) Propidium iodide (PI) control and OGD slices. As a positive control for PI staining, sister cultures were treated with saturated potassium chloride (KCl) for 5 minutes, which caused widespread cell death in the hippocampus. Scale = 300

µm, CA1: Cornu Ammonis area 1, CA3: Cornu Ammonis area 3, and DG: dentate gyrus. (B)

Example maximum intensity projections of cultures immunostained for HIF1α. Scale = 10 µm.

OGD (Kirino et al., 1984; Simon et al., 1984; Laake et al., 1999; Rytter et al., 2003; Calderone et al., 2004), which confirmed that this is a suitable model to address the mechanisms of synaptic disruption and remodeling following ischemia. Since we observed an increase in propidium iodide cell death and HIF1α expression in area CA1 as early as 90 minutes, we selected this time point to assess synapses on morphological level.

2.4.2. OGD disrupts excitatory synapses morphologically and functionally

To evaluate the morphological effect of OGD on excitatory synapses on CA1 pyramidal neurons, we used organotypic hippocampal slice cultures made from transgenic mice that express membrane-tagged GFP in a subset of CA1 neurons, allowing us to categorize dendritic spines by their shape using confocal microscopy (McKinney, 2010). We found a decrease in the total number of dendritic spines on tertiary dendrites 90 minutes following OGD (Fig. 2.2A, B; Table 2.1; **p

< 0.01, one-way ANOVA with post-hoc Tukey’s test) compared to control cells (Fig. 2.2A, B;

Table 2.1). This decrease was specific to mushroom and long-thin subtype of spines (Fig. 2.2B;

Table 2.1); no significant difference was observed in stubby spines (Fig. 2.2B; Table 2.1).

To determine whether this morphological change yielded in a functional deficit, we recorded AMPA-mediated miniature excitatory postsynaptic currents (mEPSCs) from CA1 pyramidal neurons within 24 hours after OGD, a time period at which dendritic spines remain downregulated (Fig. 2.2A, B). We found a significant increase in the IEI of mEPSCs of OGD cells

(Fig. 2C, E; 4,015.75 ± 1,644.66 ms; n = 11 cells from 6 slices; *p < 0.05, two-tailed independent

Student’s t-test) compared to control (Fig. 2.2C, E; 773.14 ± 84.25 ms; n = 14 cells from 8 slices).

In addition, we could see a rightward shift in the cumulative probability of the IEI (Fig. 2.2F; *p<

0.05, Kolmogorov-Smirnov test). Taken together, this indicates there are overall less events

Figure 2.2 – OGD induces morphological and functional deficits in excitatory synapses. (A)

Example tertiary dendrites from CA1 pyramidal neurons expressing membrane-tagged GFP in control and OGD cultures. Scale = 2 µm. (B) Dendritic spine quantification categorized into stubby, mushroom and long thin subtypes (*p < 0.05 and **p < 0.01, one-way ANOVA with post- hoc Tukey’s test). (C) Example traces of AMPA-mediated mEPSC recordings of CA1 pyramidal neurons from control and OGD slices after 24 hours. (D) mEPSCs average traces from CA1 pyramidal neurons of control and OGD slices. (E) Quantification of mEPSC IEIs (*p < 0.05, two- tailed independent Student’s t-test). (F) Cumulative probability of mEPSCs IEIs (p < 0.05, K-S test). (G) Quantification of mEPSC amplitudes (**p < 0.01 two-tailed independent Student’s t- test). (H) Cumulative probability of mEPSCs amplitudes (p < 0.05, K-S test).

Table 2.1: Normalized gephyrin expression and dendritic spine density in area CA1 from control or OGD slices at 90 minutes and 24 hours.

Gephyrin (A. U.) Number of Puncta Volume of Puncta n Control 3 subregions within CA1 averaged each 1.00 ± 0.030 1.00 ± 0.052 + 90 min from 10 slices Control 3 subregions within CA1 averaged each 1.00 ± 0.074 1.00 ± 0.070 + 24H from 5 slices OGD 3 subregions within CA1 averaged each 0.40 ± 0.052 0.53 ± 0.033 + 90 min from 10 slices OGD 3 subregions within CA1 averaged each 0.28 ± 0.035 1.06 ± 0.094 + 24H from 5 slices Dendritic Spines (spines/μm of dendrite) Mush- Total Stubby Long-thin n room Control 1.763 ± 0.506 ± 0.451 ± 0.796 ± 386 spines from 219.4 μm of dendrite + 90 min 0.083 0.051 0.047 0.062 from 8 total dendrites from 4 slices Control 1.695 ± 0.592 ± 0375 ± 0.783 ± 375 spines from 196.2 μm of dendrite + 24H 0.059 0.043 0.038 0.065 from 8 total dendrites from 4 slices OGD 1.254 ± 0.397 ± 0.233 ± 0.552 ± 295 spines from 224.6 μm of dendrite + 90 min 0.113 0.039 0.061 0.038 from 8 total dendrites from 4 slices OGD 1.316 ± 0.404 ± 0.170 ± 0.715 ± 327 spines from 249.5 μm of dendrite + 24H 0.095 0.05 0.025 0.105 from 8 total dendrites from 4 slices

76 occurring 24 hours after OGD. Moreover, we found a significant decrease in the amplitude of mEPSCs of cells in OGD treated slices (Fig. 2.2D, G; 9.51 ± 1.30 pA; n = 11 cells from 6 slices;

**p < 0.01, two-tailed independent Student’s t-test) compared to control cells (Fig. 2.2D, G; 14.18

± 0.93 pA; n = 14 cells from 8 slices). We also observed a leftward shift in the cumulative probability of the amplitudes of mEPSCs in OGD-treated slices compared to control (Fig. 2.2H;

*p < 0.05, Kolmogorov-Smirnov test), indicating there were smaller events in OGD-treated slices.

Our findings indicate that the OGD-induced loss in spines resulted in a functional loss of excitatory synapses. We next wished to determine the effect of OGD on inhibitory synapses.

2.4.3. OGD disrupts GABAergic inhibitory synapses morphologically and functionally

We evaluated the morphological effect of OGD on inhibitory GABAergic synapses in area

CA1 at 90 minutes following OGD by immunostaining for VGAT and gephyrin (pre- and postsynaptic markers of GABAergic synapses respectively). We found a significant downregulation in the number of gephyrin puncta at 90 minutes following OGD (Fig. 2.3A, B; Table 2.1; ***p <

0.0001, one-way ANOVA with post-hoc Tukey’s test) compared to control (Fig. 2.3A, B). In addition, we found that the total volume of the remaining puncta was significantly smaller 90 minutes following OGD (Fig. 2.3A, C; Table 2.1; ***p < 0.0001, two-tailed independent Student’s t-test) compared to control gephyrin (Fig. 2.3A, C; Table 2.1). The remaining gephyrin puncta tended to appose a VGAT-positive presynaptic terminal (Fig. 2.3A; Table 2.1), indicating that they have a presynaptic partner and may be putatively active. In addition, this downregulation of gephyrin was specific to area CA1, whereas gephyrin expression was comparable to control in dentate gyrus and area CA3 (Fig. 2.4A, B, C).

Figure 2.3 – OGD induces morphological and functional deficits in GABAergic inhibitory synapses on CA1 pyramidal neurons. (A) Example images of maximum intensity projections of area CA1 from organotypic hippocampal slices immunostained for gephyrin and VGAT in control and OGD cultures. Scale = 2 µm. (B) Quantification of number of gephyrin puncta per confocal stack (consisting of five 512x512 pixel z-planes each; ***p < 0.0001, two-tailed independent

Student’s t-test; gephyrin puncta values were normalized to control). (C) Quantification of the total volume of the remaining puncta (***p < 0.0001, two-tailed independent Student’s t-test). (D)

Example traces of GABAA-mediated mIPSC recordings of CA1 pyramidal neurons of control slices and slices 24 hours following OGD. (E) Average trace of mIPSCs. (F) Quantification of mIPSC IEIs (*p < 0.05, two-tailed independent Student’s t-test). (G) Cumulative probability of mIPSCs IEIs (p < 0.01, K-S test). (H) Quantification of mIPSC amplitudes. (I) Cumulative probability of mIPSC amplitudes.

Figure 2.4 – Gephyrin downregulation occurs specifically within area CA1. Example images from maximum intensity projections of different regions from organotypic hippocampal slice preparations immunostained for gephyrin and α1-GABAAR subunit in (A) control cultures and (B) cultures at 90 minutes following OGD. Scale = 20 μm. (C) Gephyrin immunofluorescence quantification within area CA1, CA3 and dentate gyrus (***p < 0.0001, one-way ANOVA post- hoc Tukey’s test).

Given this, we wished to determine whether this morphological change yielded in a functional deficit in GABAergic inhibitory neurotransmission of CA1 pyramidal neurons.

Therefore, we recorded GABAA-mediate miniature inhibitory postsynaptic currents (mIPSCs) within 24 hours of OGD induction and control cultures, a time point at which gephyrin puncta number are still downregulated (Fig. 2.3A, B; Table 2.1). We found a significant increase in the

IEI of mIPSCs of CA1 neurons in OGD-treated slices (Fig. 2.3D, F; 202.04 ± 15.51 ms; n = 13 cells from 9 slices; *p < 0.05, two-tailed independent Student’s t-test) compared to control (Fig.

2.3D, F; 152.79 ± 12.83 ms; n = 7 cells from 4 slices). We observed a rightward shift in the cumulative probability of the IEI in mIPSCs from cells of OGD cultures compared to control (Fig.

3G; **p < 0.01, Kolmogorov-Smirnov test), indicating there were overall few mIPSC events in

CA1 pyramidal neurons following OGD. However, we did not observe any change in the amplitude of mIPSCs in CA1 neurons from OGD slices (Fig. 2.3E, H; 33.09 ± 1.52 pA; n = 13 cells from 9 slices) compared to control slices (Fig. 2.3E, H; 31.49 ± 2.51 pA; n = 9 cells from 4 slices); additionally, the cumulative probability of the amplitudes of mIPSCs from CA1 neurons in control were similar to those from OGD-treated slices (Fig. 2.3I).

Given that we did not observed a GABAA-mediated mIPSC amplitudes, we immunostained for α1-GABAAR and α2-GABAAR, the main synaptic GABAAR subunits in the hippocampus

(Kasugai et al., 2010). Though α1-GABAAR can also be extrasynaptic. We found a significant decrease in both α1-GABAAR and α2-GABAAR within the tertiary dendrites of CA1 neurons 90 minutes following OGD (Fig. 2.5B, C and Fig. 6B, C; Table 2.2; *p < 0.05, two-tailed independent

Student’s t-test) compared to control (Fig. 2.5B, C and Fig. 2.6B, C). Interestingly, some α1-

GABAAR and α2-GABAAR immunopositive puncta were present that did not colocalized with gephyrin (Fig. 2.5A, 2.6A). This reveals that though there is a downregulation in these subunits,

Figure 2.5 – OGD downregulates α1-GABAAR subunits and some puncta do not colocalize with gephyrin. (A) Example tertiary dendrites from CA1 pyramidal neurons of control and OGD slices immunostained for α1-GABAAR and for gephyrin. Scale = 2 µm. White arrows highlight

α1-GABAAR puncta not colocalized with gephyrin. (B) Example dendrites repeated from panel

(A) demonstrating colocalization of α1-GABAAR and GFP. (C) Quantification of α1-GABAAR within CA1 pyramidal neuron dendrites (*p < 0.05, two-tailed independent Student’s t-test).

Figure 2.6 – OGD downregulates α2-GABAAR subunits and some puncta do not colocalize with gephyrin. (A) Example tertiary dendrites from CA1 pyramidal neurons of control and OGD slices immunostained for α2-GABAAR and for gephyrin. Scale = 2 µm. White arrows highlight

α2-GABAAR puncta not colocalized with gephyrin. (B) Example dendrites repeated from panel

(A) demonstrating colocalization of α2-GABAAR and GFP. (C) Quantification of α2-GABAAR within CA1 pyramidal neuron dendrites (*p < 0.05, two-tailed independent Student’s t-test).

83 this was not proportional to the downregulation in gephyrin. It may be possible that these synapses remain active without gephyrin.

Table 2.2: Normalized α1- and α2-GABAAR subunit expression within CA1 pyramidal neuron tertiary dendrites from control or 90 minutes post-OGD slices.

α1-GABAAR Expression puncta/μm dendrite n Control 265.5 μm of dendrite from 9 total 1.56 ± 0.18 + 90 min dendrites from 4 slices OGD 288.8 μm of dendrite from 9 total 0.951 ± 0.15 + 90 min dendrites from 4 slices α2-GABAAR puncta/μm dendrite puncta/μm dendrite n Control 374.41 μm of dendrite from 16 total 1.04 ± 0.10 + 90 min dendrites from 6 slices OGD 530.01 μm of dendrite from 12 total 0.542 ± 0.065 + 90 min dendrites from 5 slices

Taken together, our findings show that gephyrin downregulation resulted in a loss of

GABAergic synapses on CA1 pyramidal neurons. We next wished to determine what role, if any,

TrkB ligands, including proBDNF or mBDNF, could play in OGD-induced synapse dysfunction.

2.4.4. Expression of bdnf mRNA increases after OGD and TrkB-Fc treatment rescues OGD-induced excitatory synapse deficit

We tested for bdnf mRNA expression, which would account for both proBDNF and mBDNF, after OGD. We carried out RT-qPCR on micro-dissected CA1 and CA3 regions from either control or hippocampal slice cultures 90 minutes after they had been exposed to OGD. We found a significant 1.8-fold increase of bdnf mRNA expression in area CA1 compared to control

(Fig. 2.7A; n = 4 separate biological samples; *p < 0.05, two-tailed independent Student’s t-test).

Additionally, we observed a non-significant increase in area CA3 as well (Fig. 2.7A; n = 4 biological samples; p = 0.08). Elevated bdnf expression was attenuated one week after OGD (Fig.

2.7A; n = 4 biological samples).

Next, we wondered if either mBDNF or proBDNF were regulating excitatory synapse loss.

Therefore, we blocked TrkB ligands with the 10ug/ml of TrkB-Fc (a scavenger for endogenous

TrkB ligands) and quantified the dendritic spines from hippocampal slices cultures 90 minutes following OGD or from sister control cultures (with our without TrkB-Fc). We found that the observed decrease in total spine density caused by OGD (Fig. 2.7B, C; Table 2.3; *p < 0.05, one- way ANOVA with post-hoc Tukey’s test) was rescued when OGD cultures were treated with

TrkB-Fc after OGD (Fig. 2.7B, C), which were comparable to total spine densities of control (Fig.

2.7B, C) and control treated with TrkB-Fc (Fig. 2.7B, C). OGD specifically led to a decrease in mushroom spines and long-thin subtype of dendritic spines, which were all rescued with TrkB-Fc treatment (Fig. 2.7C; Table 2.3); no significant difference was observed in stubby spines (Fig.

2.7C; Table 2.3).

To determine whether this morphological rescue yielded in a functional rescue, we recorded AMPA-mediated mEPSCs from CA1 pyramidal neurons from all groups. We found that the OGD-induced increase in IEI (Fig. 2.7D, F; 1,865.28 ± 371.88 ms; n = 13 cells from 8 slices;

*p < 0.05, one-way ANOVA with post-hoc Tukey’s test) was rescued by TrkB-Fc treatment 24 hours after the induction of OGD (Fig. 2.7D, F; 868.43 ± 81.90 ms; n = 17 cells from 9 slices), which were comparable to sister control slice cultures (Fig. 2.7D, F; 934.50 ± 131.56 ms; n = 13 cells from 7 slices) and controls treated with TrkB-Fc (Fig. 2.7D, F; 991.95 ± 76.05 ms; n = 14 cells from 9 slices). Cumulative probability histogram demonstrated a rightward shift for the IEIs of mEPSCs recorded from neurons in slices 24 hours after OGD treatment compared to control,

Figure 2.7 – OGD increase expression of bdnf mRNA after 90 minutes and scavenging BDNF rescues OGD-induced excitatory synapse deficits. (A) Normalized mRNA expression of bdnf at 90 minutes and 1 week following OGD, within area CA1 and CA3, compared to control cultures

(n = 4 separate biological samples each containing 5 pooled micro-dissected regions; *p < 0.05, two-tailed independent Student’s t-test). (B) Example dendrites from CA1 neurons. Scale = 2 µm.

< 0.05, one-way ANOVA with post-hoc Tukey’s test). (D) AMPA-mediated mEPSC example traces. (E) mEPSCs average traces. (F) Quantification of mEPSC IEIs (*p < 0.05, one-way

ANOVA with post-hoc Tukey’s test). (G) Cumulative probability of mEPSC IEIs (p < 0.05, K-S test). (H) Quantification of mEPSC amplitudes (*p < 0.05, one-way ANOVA with post-hoc

Tukey’s test) (I) Cumulative probability of mEPSC amplitudes (p < 0.05, K-S test).

Table 2.3: Normalized gephyrin expression and dendritic spine density from control, control treated with TrkB-Fc, OGD and OGD slices treated with TrkB-Fc.

Gephyrin (A. U.) Number of Puncta Volume of Puncta n Control 3 subregions within CA1 averaged 1.00 ± 0.040 1.00 ± 0.032 + 90 min each from 15 slices Control 3 subregions within CA1 each from 9 0.95 ± 0.038 0.98 ± 0.033 + TrkB-Fc slices OGD 3 subregions within CA1 each from 0.27 ± 0.030 0.67 ± 0.060 + 90 min 13 slices OGD 3 subregions within CA1 each from 0.86 ± 0.047 1.02 ± 0.040 + TrkB-Fc 13 slices Dendritic Spines (spines/μm of dendrite) Mush- Total Stubby Long-thin n room Control 1.217 ± 0.410 ± 0.302 ± 0.505 ± 537 spines from 440.0 μm of dendrite + 90 min 0.050 0.029 0.021 0.042 from 16 total dendrites from 9 slices Control 1.389 ± 0.449 ± 0.368 ± 0.531 ± 601 spines from 445.4 μm of dendrite + TrkB-Fc 0.095 0.059 0.040 0.077 from 12 total dendrites from 7 slices OGD 0.951 ± 0.348 ± 0.226 ± 0.378 ± 315 spines from 333.1 μm of dendrite + 90 min 0.063 0.032 0.036 0.025 from 12 total dendrites from 8 slices OGD 1.193 ± 0.373 ± 0.297 ± 0.523 ± 467 spines from 394.2 μm of dendrite + TrkB-Fc 0.070 0.031 0.037 0.033 from 14 total dendrites from 8 slices

87 control treated with TrkB-Fc and OGD slices treated with TrkB-Fc (Fig. 2.7G; *p < 0.05,

Kolmogorov-Smirnov test). Indicating that TrkB-Fc treatment rescued the deficit in the number of mEPSC events. Furthermore, mEPSC amplitude that was reduced in CA1 neurons 24 hours following OGD (Fig. 7E, H; 9.85 ± 1.01 pA; n = 13 cells from 8 slices) were also rescued with

TrkB-Fc treatment (Fig. 7E, H; 13.18 ± 0.77 pA; n = 17 cells from 9 slices; *p < 0.05, one-way

ANOVA with post-hoc Tukey’s test), which were comparable to control (Fig. 2.7E, H; 12.25 ±

0.64 pA; n =13 cells from 7 slices) and control treated with TrkB-Fc (Fig. 2.7E, H; 11.78 ± 0.44 pA; n = 14 cells from 9 slices). Moreover, the leftward shift observed in the cumulative probability distribution of amplitudes in OGD slices was attenuated in OGD slices treated with TrkB-Fc, which were comparable to control and control treated with TrkB-Fc (Fig. 2.7I; *p < 0.05,

Kolmogorov-Smirnov test).

Our data indicates that blocking TrkB ligands after OGD can rescue excitatory synapse loss. Next, we wished to test if TrkB-Fc treatment would rescue inhibitory synapses after OGD.

2.4.5. TrkB-Fc treatment rescues OGD-induced inhibitory synapse deficit

In order to determine the effect of TrkB-Fc treatment on inhibitory synapses, we immunostained OGD and control cultures, with or without TrkB-Fc, for gephyrin. We found that the number of gephyrin puncta was downregulated 90 minutes after OGD (Fig. 2.8A, B; Table

2.3; ***p < 0.0001, one-way ANOVA with post-hoc Tukey’s test) and was rescued by TrkB-Fc

(Fig. 2.8A, B), comparable to control (Fig. 2.8A, B) and sister control cultures treated with TrkB-

Fc (Fig. 2.8A, B; Table 2.3). In addition, the volume of gephyrin puncta were significantly smaller

90 minutes following OGD compared to control, control treated with TrkB-Fc and OGD slices treated with TrkB-Fc (Fig. 2.8A, C; ***p < 0.0001, one-way ANOVA with post-hoc Tukey’s test).

Figure 2.8 – TrkB-Fc treatment rescues OGD-induced inhibitory synapse deficits. (a)

Example images of maximum intensity projections of organotypic hippocampal slices immunostained for gephyrin. Scale = 2 µm. (b) Quantification of the number of gephyrin puncta

(***p < 0.0001, one-way ANOVA with post-hoc Tukey’s test; all gephyrin values were normalized to control). (c) Quantification of gephyrin puncta volume (***p < 0.0001, one-way

ANOVA with post-hoc Tukey’s test). (d) Example traces of GABAA-mediated mIPSC recordings of CA1 pyramidal neurons. (e) Average trace of mIPSCs. (f) Quantification of mIPSC IEIs (*p <

0.05, one-way ANOVA with post-hoc Tukey’s test). (g) Cumulative probability histogram of mIPSC IEIs (p < 0.05, K-S test for untreated OGD compared to control). (h) Quantification of mIPSC amplitudes. (i) Cumulative probability histogram of mIPSC amplitudes.

In addition, we found that gephyrin puncta specifically within CA1 pyramidal neuron dendrites were rescued in OGD slices with TrkB-Fc treatment (Fig. 2.9A, B; 1.25 ± 0.158 gephyrin puncta/µm dendrite; n = 14 dendrites from 6 slices; ***p < 0.0001, one-way ANOVA with post- hoc Tukey’s test) compared to untreated OGD slices (Fig. 2.9A, B; 0.407 ± 0.088 gephyrin puncta/µm dendrite; n = 10 dendrites from 4 slices) and control dendrites (Fig. 2.9A, B; 1.28 ±

0.094 gephyrin puncta/µm dendrite; n = 10 dendrites from 4 slices). TrkB-Fc treatment did not affect gephyrin expression within control dendrites (Fig. 2.9A, B; 1.36 ± 0.10 gephyrin puncta/µm dendrite; n = 10 dendrites from 5 slices).

To test whether this morphological rescue was accompanied by a functional rescue, we recorded GABAA-mediated mIPSCs from CA1 pyramidal neurons from all groups. We found that the increase in IEI caused by OGD (Fig. 2.8D, F; 196.46 ± 47.59 ms; n = 10 cells from 7 slices;

*p < 0.05, one-way ANOVA with post-hoc Tukey’s test) was rescued with TrkB-Fc treatment

(Fig. 2.8D, F; 130.06 ± 6.53 ms; n = 11 cells from 6 slices) down to control levels (Fig. 2.8D, F;

136.72 ± 16.72 ms; n = 10 cells from 8 slices). Interestingly, we observed a non-significant trend towards an increase in the IEI of control cultures treated with TrkB-Fc (Fig. 2.8D, F; 188.41 ±

18.45 ms; n = 12 cells from 7 slices; p = 0.06, two-tailed independent Student’s t-test, comparison made to control). Cumulative probability demonstrated a rightward shift for the IEIs of mIPSCs recorded from neurons in slices 24 hours after OGD compared to control and OGD slices treated with TrkB-Fc (*p < 0.05, Kolmogorov-Smirnov test), indicating that the OGD-induced decrease in the number of events was corrected by TrkB-Fc treatment. Furthermore, mIPSC amplitudes were unchanged 24 hours following OGD (Fig. 2.8E, H; 31.33 ± 2.45 pA; n = 10 cells from 7 slices) compared to control (Fig. 2.8E, H; 27.99 ± 2.19 pA; n = 10 cells from 8 slices), control treated with TrkB-Fc (Fig. 2.8E, H; 29.86 ± 3.05 pA; n = 12 cells from 7 slices) and TrkB-Fc

Figure 2.9 – Gephyrin puncta are specifically downregulated within CA1 neurons and this deficit is rescued by TrkB-Fc. (A) Example tertiary dendrites from CA1 pyramidal neurons from control organotypic hippocampal slices, control slices treated with TrkB-Fc, slices 90 minutes after

OGD (with or without TrkB-Fc), immunostained for gephyrin. Scale = 2 µm. (B) Quantification of gephyrin puncta CA1 dendrites (***p < 0.0001, one-way ANOVA with post-hoc Tukey’s test).

91 treated OGD slices (Fig. 2.8E, H; 30.02 ± 1.13 pA; n = 11 cells from 6 slices). No difference was observed in the cumulative probability of the amplitudes of mIPSCs from CA1 neurons in any condition (Fig. 2.8I). Interestingly, we also observed that TrkB-Fc treatment for 90 minutes after

OGD could also rescue PI-positive cell death within area CA1, comparable to control levels (Fig.

2.10A, B, C, D).

Our data indicates that blocking TrkB ligands after OGD can rescue inhibitory synapse loss. However, since TrkB-Fc can inhibit not only mBDNF but other TrkB ligands such as NT4/5

(Shelton et al., 1995) and proBDNF (Fayard et al., 2005; Yang et al., 2009), we wished to test the role of mBDNF in OGD-induced synapse loss.

2.4.6. Blocking mBDNF prevents gephyrin downregulation but not dendritic spine loss

In order to test the role of mBDNF in OGD-led excitatory synapse loss, we pretreated cultures with anti-mBDNF (N-9, a function blocking antibody) for 2 hours prior to OGD and then quantified the dendritic spines; control sister cultures were processed simultaneously with and without anti-mBDNF treatment. Dendritic spine quantification revealed a significant downregulation of total dendritic spines in OGD slices (Fig. 2.11A, B; Table 2.4 for values) and anti-mBDNF pretreated OGD slices (Fig. 2.11A, B; **p < 0.001 one-way ANOVA with post-hoc

Tukey’s test) compared to control (Fig. 2.11A, B) and control cultures pretreated with anti- mBDNF (Fig. 2.11A, B). Specifically, mushroom and long-thin subtype of dendritic spines were downregulated 90 minutes following OGD with or without anti-mBDNF treatment (Fig. 2.11B;

Table 2.4), compared to control and control treated with anti-BDNF. Moreover, we carried out

AMPA-mediated mEPSC recordings on CA1 pyramidal neurons and we found that the increase in

Figure 2.10 – TrkB-Fc treatment rescues cell death in area CA1 90 minutes after OGD.

Propidium iodide staining from (A) control cultures, (B) untreated cultures 90 minutes following

OGD, (C) TrkB-Fc treated cultures 90 minutes following OGD. (D) Sister culture treated with saturated KCl for 5 minutes as a positive control for PI-stained cell death. Scale = 300 µm.

Figure 2.11 – Blocking mBDNF prevents gephyrin downregulation but not glutamatergic synapse loss. (A) Example tertiary dendrites from CA1 pyramidal neurons of organotypic hippocampal slices immunostained for gephyrin, with or without anti-mBDNF (N-9) treatment of

OGD or control cultures. Scale = 2 µm. (B) Quantification of dendritic spines categorized into stubby, mushroom and long thin subtypes (**p < 0.001 one-way ANOVA with post-hoc Tukey’s test). (C) Quantification of the number of gephyrin puncta (***p < 0.0001, one-way ANOVA with post-hoc Tukey’s test; all gephyrin values were normalized to control). (D) Quantification of gephyrin puncta volume (***p < 0.0001, one-way ANOVA with post-hoc Tukey’s test). (E)

AMPA-mediated mEPSC example traces. (F) mEPSCs average traces. (G) Quantification of mEPSC IEIs (**p < 0.01, two-tailed independent Student’s t-test). (H) Cumulative probability of mEPSC IEIs (p < 0.05, K-S test, compared to control). (I) Quantification of mEPSC amplitudes

(**p < 0.01, two-tailed independent Student’s t-test with Bonferroni correction, compared to control) (J) Cumulative probability of mEPSC amplitudes (p < 0.05, K-S test, compared to control).

Table 2.4: Normalized gephyrin expression and dendritic spine density from control, control pretreated with anti-mBDNF (N-9), OGD and OGD slices pretreated with anti-BDNF.

Gephyrin (A. U.) Number of Puncta Volume of Puncta n Control 3 subregions within CA1 averaged 1.00 ± 0.030 1.00 ± 0.058 + 90 min each from 11 slices Control 3 subregions within CA1 averaged 0.98 ± 0.038 1.07 ± 0.056 + anti-BDNF each from 6 slices OGD 3 subregions within CA1 averaged 0.44 ± 0.035 0.66 ± 0.050 + 90 min each from 12 slices OGD 3 subregions within CA1 averaged 0.85 ± 0.062 0.88 ± 0.054 + anti-BDNF each from 8 slices Dendritic Spines (spines/μm of dendrite) Mush- Total Stubby Long-thin n room Control 1.452 ± 0.497 ± 0.387 ± 0.567 ± 1434 spines from 987.1 μm of dendrite + 90 min 0.044 0.027 0.021 0.030 from 34 total dendrites from 12 slices Control 1.368 ± 0.450 ± 0.292 ± 0.624 ± 687 spines from 514.6 μm of dendrite + anti-BDNF 0.030 0.030 0.023 0.045 from 17 total dendrites from 6 slices OGD 0.950 ± 0.439 ± 0.216 ± 0.295 ± 761 spines from 804.6 μm of dendrite + 90 min 0.048 0.031 0.012 0.023 from 31 total dendrites from 9 slices OGD 1.049 ± 0.380 ± 0.258 ± 0.402 ± 546 spines from 523.1 μm of dendrite + anti-BDNF 0.062 0.021 0.026 0.039 from 19 total dendrites from 8 slices

IEI in OGD cultures (Fig. 2.11E, F, G; 777.76 + 113.10 ms; n = 13 cells from 7 slices) was comparable to OGD treated with anti-mBDNF (679.20 + 57.70 ms; n = 13 cells from 7 slices) and both were different from control (499.62 + 43.49 ms; n = 11 cells from 7 slices; **p < 0.01, two- tailed independent Student’s t-test). Cumulative probability distribution of IEIs was comparable between OGD and OGD treated with anti-mBDNF when compared to control (Fig. 2.11H; *p <

0.05, Kolmogorov-Smirnov test). Moreover, decrease in amplitudes of mEPSCs in OGD cells (Fig.

2.11E, F, I; 11.83 ± 0.79 pA) were also not rescued by anti-mBDNF treatment (11.63 + 0.57 pA), which were both different from control (16.16 + 1.29 pA; **p < 0.01, two-tailed independent

Student’s t-test); cumulative probability distribution of the amplitudes was similar in OGD and

OGD treated with anti-mBDNF (Fig. 2.11J; *p < 0.05, Kolmogorov-Smirnov test). This data indicates that inhibiting mBDNF does not rescue excitatory deficits caused by OGD.

Interestingly, immunostaining for gephyrin demonstrated that anti-mBDNF treatment was able to prevent the downregulation of gephyrin puncta number (Fig. 2.11A, C; Table 2.4) compared to OGD only slices (Fig. 2.11A, C; ***p < 0.0001, one-way ANOVA with post-hoc

Tukey’s test); anti-mBDNF treatment prevented OGD-induced downregulation of gephyrin comparable to control (Fig. 2.11A, C) and control pretreated with anti-mBDNF (Fig. 2.11A, C).

Additionally, the volume of remaining gephyrin puncta was significantly smaller 90 minutes following OGD (Fig. 2.11A, D; Table 2.4) compared to control (Fig. 2.11A, D), control pretreated with anti-mBDNF (Fig. 2.11A, D) and OGD slices pretreated with anti-mBDNF (Fig. 2.11A, D;

***p < 0.0001, one-way ANOVA with post-hoc Tukey’s test).

Taken together, our results strongly indicate excitatory synapse loss is not mediated by mBDNF, but inhibitory synapses loss is. Next, we wished to establish the downstream signaling cascade through which mBDNF leads to gephyrin downregulation.

2.4.7. Preventing ERK1/2 activation and blocking GSK3β prevents gephyrin degradation, but not dendritic spine loss

Previously, it has been demonstrated that second messengers ERK1/2 and GSK3β, which are downstream of TrkB, can regulate gephyrin cluster size and number (Tyagarajan et al., 2013); see figure 1.4 for signaling cascade. We now wished to determine whether these second messengers also play a role in synapse disruption after OGD. Therefore we used selective inhibitors against GSK3β (25 µM GSK3β-IX) and MEK (30 µM PD98059) which prevents the activation of ERK1/2. Given that both drugs had low cell permeability, we pretreated slices with these inhibitors and processed them with or without OGD, keeping treated and untreated sister cultures as controls. We found that inhibiting ERK1/2 activation and blocking GSK3β did not prevent spine loss in OGD treated cultures (Fig. 2.12A, B; Table 2.5), which were the same as untreated OGD (Fig. 2.12A, B; # p < 0.0001, one-way ANOVA with post-hoc Tukey’s test); both groups were significantly different than control (Fig. 2.12A, B; Table 2.5) and control drug treated cultures (Fig. 2.12A, B). Specifically, the mushroom and stubby type dendritic spines were most affected by OGD and were also not rescued with ERK1/2 and GSK3β treatment (Fig. 2.12B; **p

< 0.01, one-way ANOVA with post-hoc Tukey’s test).

In contrast, downregulation of gephyrin puncta was prevented in OGD slices which had been pretreated with GSK3β-IX and PD98059 (Fig. 2.12A, C; Table 2.5), compared to untreated

OGD slices (Fig. 2.12A, C; **p < 0.001; one-way ANOVA with post-hoc Tukey’s test). GSK3β-

IX and PD98059 pretreated OGD slices were similar to control (Fig. 2.12A, C) and pretreated sister control cultures (Fig. 2.12A, C). Moreover, the volume of remaining gephyrin puncta was significantly smaller 90 minutes following OGD (Fig. 2.12A, D; **p < 0.001; one-way ANOVA with post-hoc Tukey’s test) compared to control (Fig. 2.12A, D), control pretreated with GSK3β-

Figure 2.12 – Preventing ERK1/2 activation and blocking GSK3β prevents gephyrin degradation, but not dendritic spine loss after OGD. (A) Example tertiary dendrites from CA1 pyramidal neurons from control and 90 minute post-OGD slices, immunostained for gephyrin, with or without pharmacological treatment (GSK3β inhibitor: GSK3β-IX and MEK inhibitor:

PD98059). Scale = 2 µm. (B) Dendritic spine quantification categorized into stubby, mushroom and long thin subtypes (**p < 0.01 one-way ANOVA with post-hoc Tukey’s test). (C)

Quantification of the number of gephyrin puncta (**p < 0.001, one-way ANOVA with post-hoc

Tukey’s test; all gephyrin values were normalized to control). (D) Quantification of gephyrin puncta volume (**p < 0.001, one-way ANOVA with post-hoc Tukey’s test).

Table 2.5: Normalized gephyrin expression and dendritic spine density from control, control pretreated with PD98059 + GSK3β-IX, OGD and OGD slices pretreated with PD98059 + GSK3β-IX.

Gephyrin (A. U.) Number of Puncta Volume of Puncta n Control 3 subregions within CA1 averaged 1.00 ± 0.077 1.00 ± 0.038 + 90 min each from 11 slices Control 3 subregions within CA1 averaged + PD98059 1.12 ± 0.082 0.91 ± 0.050 each from 9 slices + GSK3β-IX OGD 3 subregions within CA1 averaged 0.44 ± 0.083 0.67 ± 0.042 + 90 min each from 9 slices OGD 3 subregions within CA1 averaged + PD98059 1.16 ± 0.092 0.96 ± 0.072 each from 9 slices + GSK3β-IX Dendritic Spines (spines/μm of dendrite) Mush- Total Stubby Long-thin n room Control 1.409 ± 0.558 ± 0.348 ± 0.502 ± 830 spines from 589.4 μm of dendrite + 90 min 0.055 0.026 0.023 0.028 from 21 total dendrites from 11 slices Control 1.460 ± 0.488 ± 0.288 ± 0.684 ± 800 spines from 553.2 μm of dendrite + PD98059 0.069 0.022 0.020 0.049 from 19 total dendrites from 8 slices + GSK3β-IX OGD 1.086 ± 0.362 ± 0.272 ± 0.452 ± 673 spines from 623.0 μm of dendrite + 90 min 0.063 0.027 0.031 0.033 from 22 total dendrites from 8 slices OGD 1.151 ± 0.393 ± 0.232 ± 0.526 ± 697 spines from 606.1 μm of dendrite + PD98059 0.052 0.023 0.019 0.041 from 21 total dendrites from 8 slices + GSK3β-IX

100

IX and PD98059 (Fig. 2.12A, D) and OGD slices pretreated GSK3β-IX and PD98059 (Fig. 2.12A,

D). Interestingly, we found that treating OGD cultures with PD98059 alone was sufficient to rescue the volume of gephyrin puncta (Fig. 2.13A, C; Table 2.6) but not the number (Fig. 2.13A,

B; † p < 0.01, one-way ANOVA with post-hoc Tukey’s test). However, GSK3β-IX treated OGD slices were able to partially recover the number of puncta (Fig. 2.13A, B).

Mutation studies have shown that GSK3β phosphorylates gephyrin on serine 270 (ser270) to negatively regulate the number of gephyrin clusters (Tyagarajan et al., 2011) and ERK1/2 phosphorylates gephyrin at serine 268 (ser268) to negatively regulate the size of gephyrin clusters

(Tyagarajan et al., 2013). In order to determine if phosphorylation of these serine residues were important for OGD-induced gephyrin downregulation, we used biolistic transfection of GFP- tagged gephyrin where serines 268 and 270 were mutated to alanines (gephyrinS268A/S270A) with serine residues into CA1 pyramidal neurons. We found that gephyrinS268A/S270A (Fig. 2.14; see Table

2.7 for values; p < 0.01 one-way ANOVA with post-hoc Tukey’s test) was resistant to OGD compared to wildtype gephyrin (gephyrinWT) (Fig. 2.14; Table 2.7), which were downregulated by

OGD compared to control transfections (Fig. 2.14; Table 2.7). Interestingly, as reported previously non-OGD mutant transfected gephyrin puncta were larger than control (Fig. 2.14; Table 2.7; **p

< 0.01, one-way ANOVA with post-hoc Tukey’s test).

Moreover, we showed that following ERK1/2 and GSK3β phosphorylation gephyrin becomes a substrate for calpain and it is then cleaved and degraded (Tyagarajan et al., 2013).

However, blocking calpain, with 30 μM MDL28170 treatment, rescues gephyrin puncta volume

(Fig. 2.15A, C; Table 2.8; ***p < 0.0001, one-way ANOVA with post-hoc Tukey’s test) but only partially rescues the number of gephyrin puncta (Fig. 2.15A, B).

101

Figure 2.13 – GSK3β regulates the number of gephyrin clusters and ERK1/2 regulates gephyrin cluster size. (A) Example images from maximum intensity projections of area CA1 from organotypic hippocampal slice preparations immunostained for gephyrin in control and OGD cultures with or without treatment with GSK3β-IX and/or PD98059. Scale = 2 µm. (B)

Quantification of gephyrin puncta number (*p < 0.05, † p < 0.01, # p < 0.0001, one-way ANOVA with post-hoc Tukey’s test; all gephyrin values were normalized to control). (C) Quantification of gephyrin volume (*p < 0.05, one-way ANOVA with post-hoc Tukey’s test).

102

Table 2.6: Normalized gephyrin expression in control and OGD slices with or without PD98059 and/or GSK3β-IX.

Gephyrin (A. U.) Number of Puncta Volume of Puncta n 3 subregions within CA1 Control 1.00 ± 0.071 1.00 ± 0.042 averaged each from 10 slices Control 3 subregions within CA1 1.59 ± 0.147 1.39 ± 0.148 + PD98059 averaged each from 4 slices Control 3 subregions within CA1 1.15 ± 0.070 1.84 ± 0.262 + GSK3β-IX averaged each from 5 slices Control 3 subregions within CA1 + PD98059 1.31 ± 0.90 1.18 ± 0.173 averaged each from 5 slices + GSK3β-IX 3 subregions within CA1 OGD 0.18 ± 0.030 0.56 ± 0.058 averaged each from 9 slices OGD 3 subregions within CA1 0.32 ± 0.054 1.03 ± 0.138 + PD98059 averaged each from 5 slices OGD 3 subregions within CA1 0.42 ± 0.086 0.79 ± 0.081 + GSK3β-IX averaged each from 5 slices OGD 3 subregions within CA1 + PD98059 0.90 ± 0.121 1.07 ± 0.096 averaged each from 5 slices + GSK3β-IX

103

Figure 2.14 – Mutant gephyrin that is insensitive to TrkB signaling is resistant to OGD. (a)

Maximum intensity projection of example CA1 pyramidal neurons which had been biolistitally transfected with a plasmid expressing tdTomato. Scale = 10 µm. Neurons were cotransfected with either wildtype or mutant gephyrin. (b) Example tertiary dendrites from CA1 pyramidal neurons from control and 90 minute post-OGD slices which were transfected with either gephyrinWT-GFP or gephyrinS268A/S270A-GFP. (c) Dendritic spine quantification (**p < 0.01 one-way ANOVA with post-hoc Tukey’s test). (d) Quantification of number of gephyrin per μm of tertiary dendrite (**p

< 0.01, one-way ANOVA with post-hoc Tukey’s test). (e) Quantification of gephyrin puncta volume (**p < 0.01, one-way ANOVA with post-hoc Tukey’s test).

104

Table 2.7: Gephyrin density in tertiary dendrites of CA1 pyramidal neurons and dendritic spine density from cells transfected with wildtype or mutant gephyrin (with or without OGD)

Gephyrin Density (puncta/µm) Volume (µm3) n Control + 11 total dendrites from 4 0.383 ± 0.058 0.085 ± 0.012 GephyrinWT slices Control + 12 total dendrites from 5 Gephyrin 0.470 ± 0.092 0.220 ± 0.043 slices S268A/S270A OGD + 14 total dendrites from 5 0.108 ± 0.015 0.046 ± 0.01 GephyrinWT slices OGD + 8 total dendrites from 5 Gephyrin 0.435 ± 0.12 0.210 ± 0.049 slices S268A/S270A Dendritic Spines (spines/μm of dendrite) Mush- Total Stubby Long-thin n room 436 spines from 451.6 μm Control + 1.050 ± 0.312 ± 0.288 ± 0.450 ± of dendrite from 11 total GephyrinWT 0.052 0.027 0.019 0.035 dendrites from 4 slices Control + 500 spines from 547.3 μm 1.050 ± 0.388 ± 0.250 ± 0.411 ± Gephyrin of dendrite from 12 total 0.090 0.039 0.031 0.060 S268A/S270A dendrites from 5 slices 305 spines from 491.1 μm OGD + 0.744 ± 0.315 ± 0.218 ± 0.211 ± of dendrite from 14 total GephyrinWT 0.056 0.027 0.028 0.028 dendrites from 5 slices OGD + 157 spines from 314.5 μm 0.686 ± 0.231 ± 0.160 ± 0.295 ± Gephyrin of dendrite from 8 total 0.064 0.035 0.028 0.038 S268A/S270A dendrites from 5 slices

105

Figure 2.15 – Blocking calpain rescues gephyrin puncta volume but not number. (A) Example maximum intensity projections of control and OGD slices with our without MDL28170 treatment, immunostained for gephyrin. Scale = 2 µm. (B) Quantification of gephyrin puncta number (***p

< 0.0001, one-way ANOVA with post-hoc Tukey’s test; all gephyrin values were normalized to control). (C) Quantification of gephyrin puncta volume (*p < 0.05, one-way ANOVA with post- hoc Tukey’s test).

106

Table 2.8: Normalized gephyrin expression in control and OGD slices with or without MDL28170

Gephyrin (A. U.) Number of Puncta Volume of Puncta n 3 subregions within CA1 Control 1.00 ± 0.102 1.00 ± 0.081 averaged each from 5 slices Control 3 subregions within CA1 1.81 ± 0.105 1.61 ± 0.193 + MDL28170 averaged each from 4 slices 3 subregions within CA1 OGD 0.21 ± 0.052 0.45 ± 0.032 averaged each from 5 slices OGD 3 subregions within CA1 0.42 ± 0.050 1.15 ± 0.118 + MDL28170 averaged each from 5 slices

107

Our results demonstrate that signaling downstream of TrkB is important in OGD-induced gephyrin downregulation but not dendritic spine disruption. Therefore, we next asked whether proBDNF signaling via p75NTR may underlie dendritic spine disruption.

2.4.8. Blocking proBDNF and p75NTR prevents dendritic spine loss but not gephyrin loss

We next used function blocking antibodies to inhibit proBDNF or p75NTR to determine if they played any role in dendritic spine loss after OGD. As seen previously, we found a significant downregulation in the total number of spines 90 minutes following OGD (Fig. 2.16A, B; Table

2.9; † p < 0.0001 one-way ANOVA with post-hoc Tukey’s test) compared to sister control cultures

(Fig. 2.16A, B; Table 2.9), control cultures treated with anti-proBDNF (Fig. 2.16A, B) and control cultures treated with anti-p75NTR (Fig. 2.16A, B; n = 15 dendrites from 6 slices). This downregulation was specific to mushroom and long-thin subtype of dendritic spines (Fig. 2.16A,

B; Table 2.8; *p < 0.05 and † p < 0.0001 one-way ANOVA with post-hoc Tukey’s test). OGD- induced downregulation of dendritic spines was prevented by anti-proBDNF (Fig. 2.16A, B; Table

2.8) and with anti-p75NTR (Fig. 2.16A, B).

We found that the number of gephyrin puncta were not rescued with either anti-proBDNF

(Fig. 2.16A, D; Table 2.9) or anti-p75NTR (Fig. 2.16A, C; ***p < 0.0001; one-way ANOVA with post-hoc Tukey’s test) and were comparable to untreated OGD (Fig. 2.16A, C). Neither anti- proBDNF (Fig. 2.16A, C, D) nor anti-p75 (Fig. 2.16A, C, D) treatment of control sister cultures affected gephyrin puncta number or volume, which were comparable to untreated control (Fig.

2.16A, C, D). Additionally, volume of the remaining puncta was lower in treated and untreated

OGD slices compared to treated and untreated control (Fig. 2.16A, D; ***p < 0.0001; one-way

ANOVA with post-hoc Tukey’s test).

108

Figure 2.16 – Inhibiting proBDNF and p75NTR prevents dendritic spine retraction but not gephyrin loss. (A) Example tertiary dendrites from CA1 pyramidal neurons from control organotypic hippocampal slices, control slices pretreated with function blocking anti-proBDNF, control slices pretreated with function blocking anti-p75NTR (Rex), slices 90 minutes following

OGD, OGD slices pretreated function blocking anti-proBDNF, and OGD slices pretreated function blocking anti-p75NTR (Rex), immunostained for gephyrin. Scale = 2 µm. (B) Dendritic spine quantification categorized into stubby, mushroom and long thin subtypes († p < 0.0001 one-way

ANOVA with post-hoc Tukey’s test). (C) Quantification of the number of gephyrin puncta (***p

< 0.0001, one-way ANOVA with post-hoc Tukey’s test, compared to control and treated controls; all gephyrin values were normalized to control). (D) Quantification of gephyrin puncta volume

(***p < 0.0001, one-way ANOVA with post-hoc Tukey’s test, compared to control and treated controls).

109

Table 2.9: Normalized gephyrin expression and dendritic spine density from control and OGD slices with or without anti-proBDNF or anti-p75NTR (Rex) pre-treatment.

Gephyrin (A. U.) Number of Puncta Volume of Puncta n 3 subregions within CA1 averaged each Control 1.00 ± 0.025 1.00 ± 0.042 from 18 slices Control 3 subregions within CA1 averaged each 0.97 ± 0.071 0.95 ± 0.068 + anti-proBDNF from 6 slices Control 3 subregions within CA1 averaged each 0.97 ± 0.057 1.01 ± 0.087 + anti-p75NTR from 6 slices 3 subregions within CA1 averaged each OGD 0.43 ± 0.029 0.63 ± 0.034 from 20 slices OGD 3 subregions within CA1 averaged each 0.47 ± 0.034 0.65 ± 0.023 + anti-proBDNF from 16 slices OGD 3 subregions within CA1 averaged each 0.47 ± 0.038 0.62 ± 0.033 + anti-p75NTR from 9 slices Dendritic Spines (spines/μm of dendrite) Mush- Long- Total Stubby n room thin 1.427 ± 0.558 ± 0.370 ± 0.562 ± 1840 spines from 1291.5 μm of dendrite Control 0.036 0.026 0.019 0.027 from 43 total dendrites from 15 slices Control 1.370 ± 0.495 ± 0.406 ± 0.470 ± 608 spines from 449.8 μm of dendrite + anti-proBDNF 0.066 0.025 0.033 0.042 from 15 total dendrites from 6 slices Control 1.416 ± 0.517 ± 0.343 ± 0.551 ± 578 spines from 405.1 μm of dendrite + anti-p75NTR 0.050 0.047 0.026 0.027 from 15 total dendrites from 6 slices 0.954 ± 0.418 ± 0.200 ± 0.327 ± 1091 spines from 1144.0 μm of dendrite OGD 0.039 0.025 0.011 0.023 from 39 total dendrites from 14 slices OGD 1.400 ± 0.462 ± 0.316 ± 0.618 ± 1261 spines from 907.3 μm of dendrite + anti-proBDNF 0.034 0.018 0.021 0.024 from 30 total dendrites from 13 slices OGD 1.351 ± 0.488 ± 0.300 ± 0.563 ± 1261 spines from 899.2 μm of dendrite + anti-p75NTR 0.040 0.024 0.020 0.023 from 32 total dendrites from 11 slices

110

Taken together, these findings indicate that preventing the loss of excitatory synapses does not prevent the loss of inhibitory synapses, demonstrating an independence of inhibitory and excitatory synapse loss following OGD. Moreover, we demonstrate a bifurcation of signaling where proBDNF activates p75NTR to disrupt excitatory synapses and mBDNF via TrkB disrupts inhibitory synapses in area CA1 following hippocampal OGD.

111

2.5. CONCLUSION

Here, we demonstrate that following OGD to hippocampal slice cultures there is loss of both glutamatergic and GABAergic synapses in area CA1, which are dependent on the activation of p75NTR and TrkB receptors respectively.

2.5.1. OGD leads to loss of glutamatergic and GABAergic synapses in CA1 neurons and increases bdnf mRNA expression

Our results reveal a downregulation of both dendritic spines and AMPA-mediated mEPSCs in CA1 pyramidal neurons ≤24 hours post-OGD, demonstrating that there is structural and functional loss of excitatory synapses. These findings are consistent with imaging of dendritic spines on pyramidal neurons in the cortex (Zhang et al., 2005; Brown et al., 2008; Wu et al., 2014) and hippocampal area CA1 (Pokorny & Trojan, 1983; Gonzalez-Burgos et al., 2007; Kocsis et al.,

2014) in rodents following in vivo stroke models and also after OGD in dissociated hippocampal neurons (Blanco-Suarez et al., 2014; Blanco-Suarez & Hanley, 2014). Interestingly, overall we did not observe a downregulation of stubby spines, however, the size of this spine subtype is close to the limit of light resolution (Harris et al., 1992) and therefore we may not have been able to distinguish subtle differences in stubby spine expression.

Not only did we see a decrease in excitation, we also observed a decrease in GABAergic inhibitory synapses. We observed a decrease in the number of gephyrin puncta and GABAA- mediate mIPSC events ≤24 hours post-OGD, indicating there was both structural and functional loss of inhibitory synapses. Interestingly, it has been shown that gephyrin (Mele et al., 2014) and

GABAAR (Mileson et al., 1992; Li et al., 1993; Alicke & Schwartz-Bloom, 1995; Montori et al.,

2012; Mele et al., 2014) expression decrease within area CA1 after in vivo transient forebrain

112 ischemia and this is coupled with a reduction in frequency and amplitude of GABAA-mediated mIPSCs in CA1 pyramidal neurons (Xu & Pulsinelli, 1994; Zhan et al., 2006), suggesting that inhibitory neurotransmission is compromised. Gephyrin is the main scaffolding protein at

GABAergic synapses and it is generally thought that gephyrin is required for GABAAR clustering

(Kneussel et al., 1999). However, it has been shown in the spinal cord (Kneussel et al., 2001;

Lorenzo et al., 2014) and the retina (Fischer et al., 2000) that some GABAAR can cluster without the presence of gephyrin, though it’s not known whether these GABAAR are functional. Given that there are some cases where GABAAR can cluster without gephyrin, there may be other clustering mechanisms that can compensate when gephyrin is knocked down. Since we found that the proportion of gephyrin downregulation was not necessarily 1:1 with the increase in IEI of

GABAA-mediated mIPSCs, we tested for GABAAR subunit expression, specifically the synaptic varieties: α1 and α2. Interestingly, we found that some GABAAR subunits remained after OGD and a small proportion of those did not colocalize with gephyrin. Additionally, we did not observe a chance in mIPSCs amplitude at ≤24 hours following OGD, though at this timepoint we found that gephyrin puncta volume had rebounded (see Supplemental Fig. 2c, e). Having established the effect of OGD on synapses, we wished to determine if mBDNF was related to the disruption of synapses.

We first tested the expression of bdnf and we found an increase in its expression 90 minutes following OGD, as observed by other groups (Lindvall et al., 1992; Takeda et al., 1993;

Bejot et al., 2011a; Bejot et al., 2011b; Gottlieb et al., 2013). In order to see if this increase in mBDNF was playing a role in synapse loss, we inhibited BDNF with TrkB-Fc. Remarkably, we were able to rescue the downregulation of both glutamatergic and GABAergic synapses. TrkB-Fc is known to inhibit all TrkB ligands, including: mBDNF, proBDNF, proNT4/5 and NT4/5,

113

However, given that it has been shown that NT4/5 expression does not increase in the hippocampus following (Royo et al., 2007), we believe the effect of TrkB-Fc on synapse neuroprotection was specific to proBDNF and mBDNF.

2.5.2. mBDNF disrupts GABAergic synapses after OGD

We next wanted to specifically test the role of mBDNF on synapse disruption. Using a specific function blocking antibody for mBDNF (N-9 antibody), we demonstrated that inhibiting mBDNF protected gephyrin puncta to OGD. Interestingly, acute mBDNF application causes the internalization of gephyrin and GABAAR subunits in primary amygdala neuronal cultures (Mou et al., 2013) and in dissociated hippocampal neurons (Brunig et al., 2001), which was coupled with a reduction in mIPSC frequency (Brunig et al., 2001). Given that gephyrin is responsible for

GABAAR clustering at synapses (Fritschy et al., 2008; Tyagarajan & Fritschy, 2014) and that it can regulate inhibitory GABAergic synaptic plasticity (Tyagarajan et al., 2011; Flores et al.,

2015), we believe that blocking mBDNF rescued GABAergic synapses from OGD-induced downregulation.

Most mBDNF binds through TrkB, which is its high-affinity receptor, though mBDNF can also bind with low-affinity to p75NTR (Rodriguez-Tebar et al., 1990), we next wished to tease apart which receptor was mediating the effects of mBDNF on GABAergic synapses. Previously, we have shown that GSK3β phosphorylates gephyrin on serine 270 (ser270) to negatively regulate the number of gephyrin clusters in dissociated hippocampal neurons (Tyagarajan et al., 2011).

Moreover, we demonstrated that ERK1/2 phosphorylates gephyrin at serine 268 (ser268) to negatively regulate the size of gephyrin clusters (Tyagarajan et al., 2013). Interestingly, ERK1/2 and GSK3β are second messengers that are known to be downstream of BDNF (Reichardt, 2006).

114

Therefore, we wanted to test if this signaling cascade was being activated downstream of TrkB- mBDNF. Here, we found that gephyrin clustering could be prevented by pharmacologically blocking both GSK3β and ERK1/2. Moreover, we similarly found that the OGD-induced downregulation of gephyrin cluster number and size were dependent on GSK3β and ERK1/2 respectively. Additionally, gephyrin phosphorylation at ser268 renders gephyrin as a substrate for calpain-dependent cysteine cleavage and subsequently it is targeted for proteasomal degradation

(Tyagarajan et al., 2013). We were also able to rescue gephyrin volume following OGD by pharmacologically inhibiting calpain, in accordance with our previous findings.

Recently, it has been shown that hippocampal long-term potentiation (LTP), a cellular model of learning and memory, is impaired one month after ischemia in rodents and this can be rescued by GABAAR inhibition (Li et al., 2013). They also reported a decrease in ERK1/2 expression and activation in the hippocampus 30 days after stroke (Li et al., 2013). It could be that during the post-stroke recovery phase, ERK1/2 is downregulated in order to prevent hyperexcitability within area CA1 by increasing GABAergic inhibition and decreasing glutamatergic excitation, leading to long-term cognitive impairments. This is in concordance with our findings, which suggest that ERK1/2 are initially over-activated to compensate for lack of oxygen and nutrients, but over time may be under-activated to compensate for this initial increase.

Interestingly, inhibiting mBDNF did not prevent dendritic spine loss. It is not surprising that inhibiting mBDNF did not rescue dendritic spines given the well-established role of mBDNF on glutamatergic synapse formation, stability and plasticity (Gottmann et al., 2009). Additionally,

ERK1/2 and GSK3β inhibition also did not prevent dendritic spine loss, this is likely due to the fact that ERK1/2 (Alonso et al., 2004) and GSK3β (Peineau et al., 2007) are important for excitatory synapse maintenance and plasticity; as is calpain (Amini et al., 2013).

115

Taken together, this demonstrated that mBDNF-TrkB signaling was not important for dendritic spine loss.

2.5.3. proBDNF and p75NTR disrupt glutamatergic synapses after OGD

We next asked whether p75NTR was playing a role in dendritic spine loss. To test this we used specific function blocking antibodies against p75NTR (Rex antibody) and proBDNF, which thought to bind p75NTR with higher affinity (Lee et al., 2001; Kotlyanskaya et al., 2013). Our data revealed that dendritic spine disruption could be inhibited by function blocking antibodies again proBDNF and p75NTR. Interestingly, it is highly debated whether proBDNF is released in large quantities from neurons into the extracellular space and whether it has an important physiological role (Pang et al., 2004; Matsumoto et al., 2008; Barker, 2009; Yang et al., 2009). However, there is evidence demonstrating that proBDNF is cleaved by matrix metalloproteinases (Lee et al., 2001;

Hwang et al., 2005), plasmin (Lee et al., 2001; Pang et al., 2004; Hwang et al., 2005) and tissue plasminogen activator (Pang et al., 2004), which are all located in the extracellular space.

Therefore, it may be possible that in pathological conditions there is an excess of proBDNF, causing the cleavage machinery to be overwhelmed. Though the expression of proBDNF after ischemia has not been tested by us or others, our mRNA analysis of bdnf assesses BDNF prior to release and cleavage, thereby including both pools. Moreover, ultrastructure studies have demonstrated that p75NTR puncta are expressed by CA1 pyramidal neurons at the synaptic cleft

(Woo et al., 2005). It has long been known that p75NTR activation (Woo et al., 2005) and proBDNF

(Woo et al., 2005; Yang et al., 2014) can facilitate hippocampal long-term depression (LTD). This would indicate that proBDNF and p75NTR may negatively regulate dendritic spines. Recently, it was demonstrated that proBDNF overexpression, through a cleavage-resistant version of the

116 protein, leads to reduced dendritic arborisation and spine density in vivo and these effects were dependent on p75NTR (Yang et al., 2014). Moreover, there is an increase in spine density in p75NTR knockout mice (Zagrebelsky et al., 2005). Taken together, there is considerable evidence that proBDNF-p75NTR signaling negatively regulate glutamatergic synapses on CA1 pyramidal neurons. Given that proBDNF has a higher affinity for p75NTR than for TrkB, and BDNF has a higher affinity for TrkB than for p75NTR (Lee et al., 2001; Teng et al., 2005; Barker, 2009), it seems that following proBDNF translation, secretion, release and extracellular cleavage there very well could be bifurcation in signaling proBDNF/mBDNF signaling.

Here we demonstrated that proBDNF and p75NTR are responsible for OGD-induced downregulation of dendritic spines, however, we do not know through which downstream pathway this could be. Likely, the family of small Rho GTPases that are critical for actin cytoskeletal remodeling at dendritic spines, including RhoA, Rac1 and Cdc42, would play an important role.

RhoA in particular has been shown to be downstream of p75NTR-mediated dendritic spine remodeling (Gehler et al., 2004; Domeniconi et al., 2005; Sun et al., 2012). Whether RhoA is mediating dendritic spine retraction is unknown. Another member of the RhoA small GTPases super family, Rac1, is also known to be downstream of p75NTR (Schweigreiter et al., 2004; Zeinieh et al., 2014) and can regulate dendritic spine morphology (Luo et al., 1996; Nakayama et al., 2000;

Tashiro & Yuste, 2004; Tolias et al., 2005). Recently, it was demonstrated that Rac1 mediates spine disruption after OGD in dissociated hippocampal neurons (Blanco-Suarez et al., 2014), leading us to believe that RhoA and Rac1 are likely downstream of p75NTR-induced dendritic spine disruption.

In conclusion, here we demonstrate that proBDNF/mBDNF bidirectional signaling plays an important role in synapse disruption after ischemia. Given the conflicting evidence on whether

117 mBDNF is detrimental (Nygren et al., 2006; Qin et al., 2014) or protective (Schabitz et al., 2007;

Yu et al., 2013) following stroke, our data may better explain these discrepancies. In particular, the motor recovery seen in mice expressing the Val66Met polymorphism (Qin et al., 2014), which has reduced levels of both mBDNF and proBDNF (Egan et al., 2003). Furthermore, findings from our work may lead to better understand of neuroprotective strategies for cognitive enhancement following ischemic stroke.

118

CHAPTER 3. BLOCKING BDNF INHIBITS INJURY INDUCED HYPEREXCITABILITY OF HIPPOCAMPAL CA3 NEURONS

FOREWORD

In Chapter 2, I demonstrated that BDNF and its precursor protein can disrupt excitatory and inhibitory synapses in a model of ischemia. Interestingly, we observed an increase in BDNF following ischemia and patients of stroke often develop seizures or even epilepsy years after initial insult. In Chapter 3, I wanted to understand whether BDNF played a role in hyperexcitability after brain injuries. Therefore, I used a model in which we have previously demonstrated that lesion of the Schaffer collaterals, mimicking a traumatic brain injury, to hippocampal slices can induce hyperexcitability of CA3 neurons due to potent axon sprouting and formation of excitatory synapses. This model mimics post-traumatic epilepsy, with a delayed onset of axon sprouting as seen in patients. The mechanism of this sprouting is still unclear, thought it has been previously shown that there is a reduction in axon sprouting in conditional TrkB knockdown mice. However, the role of BDNF in lesion-induced axon sprouting is poorly understood.

Here, I hypothesized that BDNF initiates injury-induced axonal remodeling and hyperexcitability, leading to changes in network activity.

To test my hypothesis, I made lesions to the Schaffer collateral pathway in organotypic hippocampal slice cultures and probed for GAP43, a marker for regenerating axons, with immunofluorescence. Furthermore, I tested for the formation of a recurrent network and CA3 pyramidal neuron hyperexcitability with electrophysiology. BDNF expression was tested with RT- qPCR and this increase was blocked with a scavenger for BDNF, TrkB-Fc. The following chapter was published in the European Journal of Neuroscience in 2013.

119

3.1. ABSTRACT

Brain trauma can disrupt synaptic connections, which in turn can prompt axons to sprout and form new connections. If these new axonal connections are aberrant, hyperexcitability can result. It has been shown that ablating TrkB, a receptor for brain-derived neurotrophic factor

(BDNF), can reduce axonal sprouting after hippocampal injury. However, it is unknown if inhibiting BDNF-mediated axonal sprouting will reduce hyperexcitability. Given this, here our purpose is to determine if pharmacologically blocking BDNF inhibits hyperexcitability after injury-induced axonal sprouting in the hippocampus. To induce injury, we made Schaffer collateral lesions to organotypic hippocampal slice cultures. As published by others, we observed at 50% reduction in axonal sprouting in cultures treated with a BDNF blocker (TrkB-Fc) 14 days after injury. Furthermore, lesioned cultures treated with TrkB-Fc were less hyperexcitable compared to lesioned untreated cultures. Using electrophysiology we observed a 2-fold decrease in the number of CA3 neurons that exhibited bursting responses after lesion with TrkB-Fc treatment, while we found no change in intrinsic neuronal firing properties. Lastly, evoked field

EPSP recordings indicated an increase in network activity within area CA3 after lesion, which was prevented with chronic TrkB-Fc treatment. Taken together, our results demonstrate that blocking

BDNF attenuates injury-induced hyperexcitability of hippocampal CA3 neurons. Axonal sprouting has been found in patients with post-traumatic epilepsy. Therefore, our data suggests that blocking the BDNF-TrkB signaling cascade shortly after injury may be a potential therapeutic target for the treatment of post-traumatic epilepsy.

120

3.2. INTRODUCTION

Brain injury can physically disrupt axons and synapses within neuronal networks. When these connections are disrupted, axons may respond by sprouting their collaterals. Often, this process is aberrant and can lead to the functional disturbance of an existing neuronal network. For example, there is considerable evidence for mossy fiber sprouting from hippocampal dentate granule cells following excitotoxic injury with kainic acid (Davenport et al., 1990; Routbort et al.,

1999), lesion of existing mossy fibers (Laurberg & Zimmer, 1981; Hannesson et al., 1997) or injured cortex (Salin et al., 1995). However, the underlying cellular events that cause axonal growth following injury are poorly understood.

Neurotrophic factors, in particular brain-derived neurotrophic factor (BDNF) and its receptor TrkB, are important for axonal growth during development (Huang & Reichardt, 2001).

Davies and colleagues were the first to show that BDNF application to dissociated chick sensory neurons lead to profuse neurite outgrowth (Davies et al., 1986). More recent studies have demonstrated that BDNF can lead to neurite elongation or branching through distinct signaling cascades which are dependent on the concentration of BDNF in the extracellular space (Ji et al.,

2010). Moreover, the application of extracellular BDNF promotes neurite differentiation to axons

(Shelly et al., 2007; Shelly et al., 2011) and BDNF is required and sufficient for the formation of axons in cultured hippocampal neurons (Cheng et al., 2011). We have also reported that exogenous

TrkB ligands can induce axon sprouting in unlesioned organotypic hippocampal cultures, which increased the activity of CA3 neurons (Schwyzer et al., 2002).

Previously, we have shown that transecting the Schaffer collateral pathway, which connects CA3 pyramidal neurons to ones in CA1, can lead to axonal sprouting in area CA3

121

(McKinney et al., 1997). This resulted in more reciprocal excitatory synaptic connections between

CA3 cells and led to hyperexcitability of CA3 neurons 14 days following transection (McKinney et al., 1997). Recent evidence has demonstrated that BDNF and TrkB expression is upregulated

72 hours after Schaffer collateral injury (Dinocourt et al., 2006; Aungst et al., 2013). Moreover, conditional partial knock-down of the BDNF receptor, TrkB, can reduce CA3 axonal sprouting after Schaffer collateral lesion (Dinocourt et al., 2006). However, it remains to be determined whether inhibiting this BDNF-mediated axonal sprouting can attenuate the hyperexcitability of

CA3 neurons. It is also unknown, how this axonal sprouting functionally affects the hippocampal network within area CA3. Moreover, the hippocampal localization of BDNF and the time course of its expression have not been established after Schaffer collateral lesion.

Therefore, we hypothesize that BDNF initiates injury-induced axonal remodeling and hyperexcitability, leading to changes in network activity. To test our hypothesis, we transected the

Schaffer collateral pathway of organotypic hippocampal slice preparations. Following which, we immunostained for growth-associated protein of 43 kDa (GAP43), a known marker for axonal growth, to determine the extent of axonal sprouting. Expression of BDNF and TrkB were measured through quantitative PCR and Western blot analysis. Lastly, we used electrophysiology to determine the functional effect of the aberrant sprouting in area CA3.

Here, we demonstrate that transection of the Schaffer collateral pathway can upregulate

BDNF to induce axonal sprouting and hyperexcitability of hippocampal CA3 neurons, which leads to an increase in network activity within area CA3.

122

3.3. MATERIALS AND METHODS

3.3.1. Ethics Statement

All animal handling procedures were carried out consistent with guidelines set by the

3.3.2. Hippocampal Slice Cultures and Schaffer Collateral Lesions

Briefly, 7-day-old C57B6 mice were decapitated and a bilateral hippocampal dissection was performed, from which 400 µm thick transverse hippocampal slices were made and plated on glass coverslips. Slices were adhered in place with a clot of chicken plasma (Cocalico Biologicals;

Reamstown, PA, USA) coagulated with thrombin (Invitrogen GIBCO). Coverslips were placed in flat-sided culture tubes with antibiotic-free serum-containing media and maintained in a dry-air roller drum incubator at 36C for three weeks prior to experimentation. To prevent excitotoxic injury during lesioning, slices (> 21 days in vitro) were placed into cutting solution containing 0.5

μM tetrodotoxin (Alomone Labs, Jerusalem, Israel) and 10 mM MgCl2.

123

Lesions were made to the Schaffer collaterals between area CA3 and CA1 with a razor blade shard; cuts extended from the alveus layer to stratum radiatum. Control sister cultures were also exposed to the cutting media for the same period of time, but not lesioned. Cultures were then returned to normal culture media and were processed simultaneously in all experiments.

3.3.3. Visualization of CA3 Pyramidal Neurons

Farnesylated mCherry protein was expressed specifically in CA3 neurons by infecting control and 14 days post-lesion organotypic hippocampal slices derived with Semliki Forest Virus strain PD, which selectively targets neurons (Lundstrom et al., 2003; Lundstrom & Ehrengruber,

2003). The infection was achieved by microinjection using picospritzer (picospritzer III, Parker;

Cleveland, OH, USA) of 1-2 μl of virus via a glass pipette into area CA3 with 5-10 PSI with 25-

50 ms pulses at 0.1 Hz for 15-20 minutes. Approximately 16-20 hours after infection, the cultures were fixed in 4% paraformaldehyde for 1 hour at room temperature and mounted onto slides for confocal imaging.

3.3.4. Electrophysiological Recordings and Analysis

Slices were placed in a temperature-controlled chamber (30°C) mounted on an upright microscope (DM LFSA, Leica Microsystems) and continuously perfused with Tyrode solution containing: 137 mM NaCl, 2.7 mM KCl, 2.5mM CaCl2, 2 mM MgCl2, 11.6 mM NaHCO3, 0.4 mM NaH2PO4, and 5.6 mM glucose (pH 7.4). To assess excitability, traces were recorded from

CA3 neurons (2 cells/slice) in current-clamp mode. Whole-cell recording electrodes were pulled from borosilicate glass (resistance of 4-6 MΩ; GC150TC; Clark Instruments, UK) using a P-97

124 electrode puller (Stutter Instrument Co., Novato, CA, USA). All electrophysiological recordings were made using an Axopatch 200A amplifier (Molecular Devices, Sunnyvale, CA, USA). Signals were recorded at 20 kHz and then filtered through a Bessel low-pass filter at 2 kHz using the

Clampex 9.2 acquisition program (Axon Instruments Inc.). Recording pipettes were filled with intracellular solution containing: 120 mM K-Gluconate, 1 mM EGTA, 10 mM HEPES, 5 mM Mg-

ATP, 0.5 mM Na-GTP, 5 mM NaCl, 5 mM KCl, 10 mM phosphocreatine, and pH was adjusted to 7.3 with KOH, and 285-295 mOsm. CA3 neurons were chosen at approximately 300 μm from the site of lesion in the case of transected slices. Once whole-cell configuration was established, cells were monitored for 5 minutes for equilibration of internal solution and to ensure that the seal and opening were maintained; cells were then recorded for 20 minutes. Input resistance and membrane potential were monitored at regular intervals throughout the experiment and no observable differences were noted between conditions. Data were discarded if recordings in cells showed a seal resistance <5 GΩ, a holding current >-100 pA, or resistance deviated more than

±20% through the course of recording. Also, cells that failed to last the entire 20 minutes of recording were excluded from further analysis. The digitized data was coded prior to analysis so that analysis could be carried out blinded. Upward deflecting EPSPs greater than 40 mV in amplitude were considered to be somatic action potentials and were detected offline. Action potential threshold was identified as the membrane potential at the start point of the first action potential. Three or more action potentials were grouped into bursts if the inter-spike interval was smaller than 600 ms. For spontaneous EPSP frequency analysis, traces displaying typical EPSP upward deflections are considered and analyzed as EPSPs if they are greater than 3 mV in size.

Action potentials, were not taken into account in EPSP analysis.

125

To assess network activity, evoked field EPSPs (fEPSPs) were recorded in CA3 with recording electrode placed in stratum oriens at 30C. Field recording pipettes were pulled to 0.5

MΩ and filled with external Tyrode solution. The extracellular solution consisted of regular

Tyrode solution without the addition of any pharmacological agents. CA3 axonal stimulation was achieved using an insulated platinum-iridium bipolar electrode (50 μm diameter, 25 μm layer of

Teflon insulation (A-M Systems Inc., Carlsborg, WA, USA). Stimulation strength ranged from 20 to 200 μA to elicit a minimum response and a single stimulus was delivered every 20 seconds.

Stimulus strength was increased every five stimuli by 10% to a maximum of 300% of minimum response. Signal was low-pass filtered at 2 kHz and recorded at 20 kHz using the PCLAMP 9.2 software (Axon Instrument Inc.). The digitized data was coded prior to analysis so that analysis could be carried out blinded. fEPSPs were detected offline and analyzed for compound responses.

Compounded responses were considered to be any fEPSPs responses that included two or more downward deflections following the delivery of a stimulus before reaching maximum amplitude.

The number of peaks during the rise phase were calculated and compared to corresponding stimulation strengths.

3.3.5. Immunohistochemistry

Slices were fixed at 4°C overnight in 4% paraformaldehyde made up in 0.1 M PB, pH 7.4.

Following fixation, slices were washed in 0.1 M PB, permeabilized in 0.4% Triton X-100 and blocked with 1.5% heat inactivated horse serum overnight at 4°C. Primary antibodies were incubated for five days at 4°C in permeabilizing buffer with a 1:20 dilution of GAP43 mouse antisera (clone 10E8/E7, provided by Dr. K. Meiri, Tufts University, Boston, MA, USA) (Meiri et al., 1986), 1:250 dilution of anti-pTrkB (Y515; Abcam) or 1:250 of anti-PSD95 (Millipore).

126

Following several 0.1 M PB washes, slices were incubated overnight at 4°C in 1:250 secondary

Alexa Fluor 594 anti-mouse-IgG1 (Invitrogen Molecular Probes) antibody diluted in 0.1 M PB containing 1.5% heat inactivated horse serum. Finally, slices were mounted with DAKO

Fluorescent Mounting medium (Dako Canada; Mississauga, Canada) onto microscope slides.

Following mounting, slices were imaged in area CA3 using a Leica TCS SP2 scanhead (Leica

Microsystems) on a Leica DM6000 B upright microscope, equipped with a HCX PL APO 40× NA

1.4 oil immersion objective using a 543 nm HeNe laser line. Image stacks were collected at Z =

0.4 μm and averaged 4 times to improve signal-to-noise ratio. For quantification of the presence of GAP43 fibers, image stacks were obtained with identical parameters (laser intensity, filters, pinhole size, photomultiplier tube gain and offset). Representative images are maximum intensity projections of 5 sections from each stack.

3.3.6. RT-qPCR

Slices were lesioned in cutting solution and areas CA1 and CA3 were microdissected, snap frozen on dry ice and preserved in RNAlater (Ambion Inc., Austin, TX, USA) until experimentation. Control cultures were exposed to cutting media and were maintained for 6 hours prior to microdissection. 5-6 slices were used for each experimental group. Each condition was tested with 3 biological samples in triplicate. Total RNA was extracted using TRIzol reagent

(Invitrogen) and 1 μg of RNA was reverse transcribed to cDNA using 20U of AMV reverse transcriptase (Roche Diagnostics), as per manufacturer’s recommendation. Subsequently, qPCRs were performed using 2 μl of cDNA in a 20 μl reaction containing SYBR green mastermix (Roche

Diagnostics), and 0.5 μM of forward and reverse primers (all primers were obtained from

Invitrogen). Reactions were performed using a Roche LightCycler LC480 under the following

127 conditions: denaturation step 10 minutes at 95°C, 45× [denaturing at 95°C for 10 seconds, annealing for 10 seconds and extension at 72°C for 10 seconds], and lastly a final extension step at 72°C for 10 minutes. This was followed by melting curves and cooling cycles. The following primers and annealing temperatures were used for experimentation: BDNF targeting all mRNA transcript variants [sense 5’-GAA GAG CTG GAT GAG GAC-3’ and antisense 5’-TTC AGT

TGG CCT TTT GAT ACC-3’] annealing temperature 59°C, TrkB targeting transcript for the full- length receptor [sense 5’-TGA TGG CAG AGG GTA ACC-3’ and antisense 5’-CTA CTA TCG

GGT CGG TGG C-3’] annealing temperature 57°C, GAPDH [sense 5’-AAA TGG TGA AGG

TCG GTG TG-3’ and antisense 5’-TGA AGG GGT CGT TGA TGG-3’] annealing temperature

59°C, and β-actin [sense 5’-TGG TGG GTA TGG GTC AGA AGG ACT C-3’ and antisense 5’-

CAT GGC TGG GGT GTT GAA GGT CTC A-3’] annealing temperature 59°C. Reaction efficiencies for each assay were calculated from serial dilutions of cDNA. All qPCR data was quantified and normalized by Roche LightCycler 480 software.

3.3.7. Western Blotting

10 control or 10 lesioned cultures were grouped together for each paradigm (defined as one sample group). Slices were freed from the plasma clot, pooled, and lysed in ice-cold RIPA buffer

(1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate, 50 mM Tris-HCl

(pH 8.0), 150 mM NaCl) and completeMini protease inhibitors (Roche). Protein concentration was determined using a BCA dye-binding assay (Thompson Scientific), according to the manufacturer’s protocol, with bovine serum albumin (BSA) as a standard. 25 µg of total protein was separated by SDS-PAGE (10% or 15% resolving) under reducing conditions and wet- transferred onto 0.22 µm pore size polyvinelidenedifluoride membranes (Millipore). Membranes

128 were blocked for 1 hour at room temperature with 5% BSA in tris-buffered saline with 0.05% tween-20 (TBST). Membranes were then incubated at 4°C overnight with anti-BDNF (Santa Cruz

Biotechnology 1:500), anti-TrkB (BD Falcon 1:1000), or anti-β-tubulin (Sigma-Aldrich, 1:3000; internal loading control) diluted in 5% BSA in TBST. Primary antibodies were revealed by horseradish peroxidase-conjugated secondary antibodies (1:50000, Bio-Rad Laboratories) diluted in 5% BSA in TBST for one hour at room temperature. Immunoreactive bands were detected with enhanced chemiluminescence (GE Healthcare) according to the manufacturer’s protocol and revealed using Kodak BioMax Light Film (Sigma-Aldrich). Blots were analyzed with Adobe

Photoshop software for mean pixel intensity.

3.3.8. Immunoprecipitation

To detect the phosphorylation of TrkB, immunoprecipitation was carried out on 10 pooled control or 3 hours post-lesion slices with 500 μl of ice-cold RIPA buffer containing completeMini protease and phosphatase inhibitors (Roche). Lysates were incubated overnight at 4°C with 1 μg of primary anti-pTrkB/TrkA antibody (pY515; Cell Signaling, Beverly, MA, USA). 50 μl of

Protein G-Agarose beads (Thermo Scientific) were added to the lysates and incubated for 90 minutes at 4°C. After three washes with 0.1 M PB, the beads were spun down and boiled at 95°C for 5 minutes in 2x Laemmle buffer (Bio-Rad Laboratories) containing β-mercaptoethanol.

Proteins were separated by SDS-PAGE and analyzed by Western blot. TrkB expression was assessed with anti-TrkB (BD Falcon 1:1000).

129

3.3.9. Pharmacological Treatments

To scavenge BDNF, slices were treated with TrkB-Fc (R&D Systems; Minneapolis, MN,

USA), a fusion protein in which the BDNF binding site of the TrkB receptor replaces the Fc fragment of a human IgG1 antibody. As a control, we treated sister lesioned cultures with IgG-Fc.

We found that TrkB-Fc treatment to hippocampal cultures for 24 hours downregulated TrkB receptor phosphorylation (data not shown). TrkB-Fc or IgG-Fc was diluted in culture media at a final concentration of 10 μg/ml and chronic treatment began 12 hours following transection of the

Schaffer collateral of the slices maintained in vitro >21 days. Control culture media or drug- containing media was replaced 3 times a week for 1-3 weeks following transection.

3.3.10 Statistical analysis

Student’s t-test (Bonferroni correction was applied when necessary), Fisher’s exact test, or one-way ANOVA with post hoc Tukey’s test were used when appropriate. In cases where data were not normally distributed (tested with Lilliefors) a Kruskal-Wallis Test with post hoc Mann-

Whitney U-Test was utilized, unless otherwise specified. Results are expressed as mean ± S.E.M and n values represent number of slices, unless otherwise stated. For Fig. 2B and 6B, linear regression was performed for each slice and then slopes were compared using one-way ANOVA with post hoc Tukey’s test.

130

3.4. RESULTS

3.4.1. Schaffer collateral transection in mouse organotypic slices leads to axonal sprouting and hyperexcitability 14 days following injury

We wished to determine the role of BDNF and TrkB in injury-induced hyperexcitability.

Previously, we studied axonal sprouting by transecting the Schaffer collateral pathway in mature rat hippocampal slice cultures (McKinney et al., 1997). Here, we sought to consolidate our findings in a mouse organotypic hippocampal system and re-establish the timeframe for axonal sprouting.

Initially, we tested if a similar outcome and timeframe could be observed following lesion to Schaffer collaterals in mouse organotypic hippocampal slices. Three days after transection of the Schaffer collateral pathway, we observed a small gap between area CA3 and area CA1 (Fig.

3.1A), isolating the two areas. We saw the presence of a few propidium iodide positive cells close to the lesion site in area CA3 and CA1 (data not shown), resulting in selective cell death close to the lesion site, similar to our previous observations in rat slices (McKinney et al., 1997; McKinney et al., 1999b).

However, what effect did this lesion have axonal morphology and reorganization of axonal fibers? 14 days after transection, we microinjected a virus expressing an mCherry plasmid into area CA3, which transfected the mCherry plasmid into a pyramidal cell situated close to the lesion site (>500 μm). We found that the typical Schaffer collateral morphology, a parallel bundle of axonal fibers, was disrupted and the axonal processes were undulating and twisting, seeming to lack a primary direction of growth (Fig. 3.2A, B). To discern if this disrupted morphology was due to newly sprouted axons, we performed immunofluorescence staining for GAP43, an established

131

Figure 3.1 – GAP43 expression is downregulated in 21 day in vitro cultures. (A) Bright field micrograph of a Nissl-stained (>21 DIV) hippocampal slice 3 days post-lesion (DPL); scale = 500

μm. Note the presence of a gap between CA3 and CA1 at the lesion site. (B) Many GAP43 immunopositive fibers were found in young mouse hippocampal preparations (3 DIV), though this is downregulated in mature slices (C). Scale bar = 50 μm.

132

Figure 3.2 – Schaffer collateral lesion induced

axonal sprouting and hyperexcitability at 14

days post-lesion (DPL) in area CA3. (A) A CA3

neuron, visualized with virus transfection of an

mCherry plasmid in a 14 DPL preparation. (B)

Expanded inset of (A), showing axons regrowing

towards the lesion site. (C) GAP43

immunopositive fibers were downregulated in

mature slices (>21 DIV). (D) When mature

hippocampal cultures were lesioned, GAP43

immunopositive fibers were observed in area

CA3 adjacent to the lesion site after 14 DPL. (C)-

(D) are representative images of maximum

intensity projections rendered from confocal

stacks of slices immunohistochemically stained

for GAP43 and all images were taken in area CA3

(n = 10 slices for each paradigm). Note: dotted

line indicates site of lesion; scale of (A), (C), and

(D) is 50 μm and scale of (B) is 20 μm. (E)

Representative traces of whole-cell current-

clamp electrophysiological recordings of CA3

pyramidal neurons. (F) Quantification of action

potential frequency. We found a significant

133 increase in action potential firing at 14 DPL versus control. (G) Analysis of burst frequency. We detected a significant increase in burst frequency in 14 DPL slices compared to control. Note: for

(F) and (G) **p < 0.05, two-tailed students’ t-test; data are depicted as mean ± S.E.M. (H)

Percentage of CA3 neurons exhibiting spontaneous action potentials. (I) Percentage of cells exhibiting spontaneous bursting. Note: for (H) and (I) **p < 0.01, two-tailed Fisher’s exact test.

134 marker for axonal growth (Meiri et al., 1986). GAP43 is highly upregulated during development when there is a high level of axonal growth, but it is downregulated when an animal matures to adulthood, we find that in hippocampal slice cultures the expression of GAP43 is downregulated at 21 days in vitro (Fig. 3.1B, C), the time point at which we made Schaffer collateral lesions. We found that 14 days following lesion of Schaffer collaterals, at a time point where GAP43 is normally downregulated (Fig. 3.2C), a number of fibers were GAP43 immunoreactive immediately adjacent the lesion site in area CA3 but not CA1 (Fig. 3.2D). The GAP43 immunopositive fibers are visible for 327.2±12.8 µm (n = 24 slices) away from the lesion site into area CA3. Hence, lesion of the Schaffer collateral pathway leads to the upregulation of GAP43 and axonal sprouting of CA3 pyramidal neurons. Does the observed axonal sprouting in mouse hippocampal slices result in hyperexcitability of CA3 neurons, as seen after rat hippocampal culture transections?

To determine if a functional change accompanied the morphological changes, we measured spontaneous activity with whole-cell current clamp recordings from CA3 pyramidal neurons in either control or 14 days post-lesion slices. Compared to control unlesioned slices, we found that lesioning markedly increased excitation, with increased action potential firing (Fig. 3.2E, F;

3.95±1.17 action potentials per minute; n = 20 cells) compared to control sister cultures (0.40±0.16 action potentials per minute; n = 14 cells). Moreover, there was a significant increase in the average number of bursts (defined as 3 or more consecutive spikes with inter-spike intervals <600 ms) in slices 14 days following lesion (Fig. 3.2G; 0.56 ± 0.16 bursts per minute; n = 21 cells) compared to control (0.094 ± 0.034 bursts per minute; n = 14 cells). In addition, we found that 90% of 14 days post-lesion cultures (n = 18 out of 20 cells) exhibited spontaneous action potentials, whereas

42.9% of control cultures fired spontaneously (Fig. 3.2H; n = 6 out of 14 cells). Furthermore, 75%

135 of CA3 neurons in 14 days post-lesion slices (n = 21 out of 28 cells) bursted, compared to only

31.6% of control CA3 neurons (Fig. 3.2I; n = 6 out of 19 cells). Therefore, 14 days after lesion

CA3 neurons are more likely to have spontaneous hyperexcitability compared to CA3 neurons from control cultures.

These results show 14 days after Schaffer collateral transection in mouse organotypic hippocampal cultures, we consistently observe GAP43 labeling of de novo axons and an increase in excitation, comparable to our previously published data in rats. Thus, confirming that this is a suitable model to address the mechanisms of reactive axonal sprouting-induced hyperexcitability.

Since we observed an increase in hyperexcitability, we then ventured to determine why the CA3 neurons were more excitable.

3.4.2. Intrinsic electrical properties of CA3 neurons remains unchanged after Schaffer collateral transection

Next, we evaluated the intrinsic electrical properties of CA3 pyramidal cells as we had observed an increase in the firing and bursting of these neurons. We found that the input resistance of CA3 pyramidal cells was unchanged in 14 days post-lesion slices (124.9 ± 20.9 MΩ; n = 16) compared to control cultures (126 ± 28.8 MΩ; n = 12). Next, we looked at the resting membrane potential of these neurons, and we found they were also unchanged in 14 days post-lesion slices (-

71.8 ± 0.48 mV; n = 16) compared to control cultures (-69.7 ± 1.4 mV; n = 12), also within the normal range of CA3 pyramidal neurons (Andersen, 2007). Lastly, we tested the input/output response of CA3 pyramidal cells by performing current step experiments. We found no change in the number of action potentials fired in response to various injected currents, some examples are

136 shown in Table 1. Therefore, we concluded that Schaffer collateral lesion did not affect the intrinsic electrical properties of CA3 pyramidal neurons 14 days post-injury.

Our next question was to determine if there were any changes to GABA production, as an increase in hyperexcitability could be related to changes in inhibitory GABAergic transmission.

Therefore, we immunostained for glutamate decarboxylase 67 (GAD67) to determine if GABA production was increased 14 days following injury, the time when we had seen the most robust axonal sprouting. We found that GAD67 staining was comparable to control (Fig. 3.3A, B, C), indicating that there was no change in the production of GABA after injury.

Table 3.1: Input/output response of CA3 pyramidal neurons remains unchanged 14 DPL

# of APs Fired at Various Injected Currents -15 pA 0 pA 25 pA 50 pA n Control 0.0 ± 0.0 0.0 ± 0.0 39.91 ± 12.21 71.1 ± 10.33 11 Lesion 0.0 ± 0.0 0.0 ± 0.0 38.08 ± 9.82 54.38 ± 9.18 13 p - - 0.906 0.674

3.4.3. Hyperexcitability of CA3 neurons after Schaffer collateral transection can be attributed to increased excitatory inputs

Since we did not observe any changes in the intrinsic electrical properties of CA3 pyramidal neurons 14 days after lesion or in GAD67 staining, we performed evoked field potential

(fEPSPs) recordings to further characterize network activity in area CA3. We stimulated axons in area CA3 closer to dentate gyrus and recorded potentials in area CA3 near the lesion site. We observed that there was an increase in compounded fEPSPs (defined as responses that included

137

Figure 3.3 – GAD67 expression is unchanged by Schaffer collateral lesion. (A) Control slices widely expressed GAD67, a general marker for inhibitory neurons, in area CA3. Scale bar = 50

μm; dotted line indicates site of lesion. (B) We found this to be unchanged in 14 DPL slices. (C)

The mean fluorescence intensity was quantified and we found no significant difference between both groups (measurements in arbitrary units). (A) and (B) are representative images of maximum intensity projections rendered from confocal stacks of slices immunohistochemically stained for

GAD67 and all images were taken in area CA3 (n = 5 slices for each paradigm).

138

Figure 3.4 – Axonal sprouting leads

to an increase in area CA3 network

activity. (A) Representative traces of

evoked fEPSP recordings within area

CA3 for 100%, 150% and 200% of

minimal stimulation. Note: the arrows

depict additional peaks in compounded

fEPSP responses. (B) Quantification of

the number of peaks as a function of

stimulation strength. There is a

significant increase in the number of

compounded fEPSPs with an

incremental increase in stimulation

strength in 14 DPL cultures compared

to control (*p < 0.05; linear regression

was performed for each slice and then

the two slopes were compared using

independent students’ t-test, we then

made pairwise comparisons for

individual data points with unprotected independent students’ t-test). (C) Quantification of percentage of cultures depicting compounded fEPSPs (*p < 0.05; Fisher’s Exact Probability test, two-tailed).

139 two or more downward deflections following stimulus before reaching maximum amplitude) in relation to the stimulation strength in area CA3 of 14 days post-lesion slices, when compared to control (Fig. 3.4A). We found that with the same stimulation intensity used in control, we were able to recruit more fibers in 14 days post-lesion slices (Fig. 3.4A, B). Moreover, overall, we found an increase in the percentage of cultures that depicted compounded fEPSPs at any stimulation intensity in 14 days post-lesion slices (100%; n = 10; Fig. 3.4C) compared to control (50%; n = 8;

Fig. 3.4C). This indicates that there are more connections, likely due to the formation of a recurrent network.

Our next question was to determine what cellular signal could be causing these axons to sprouting and form new connections. Since recent evidence has suggested that BDNF was important for injury-induced axonal sprouting following Schaffer collateral transection (Dinocourt et al., 2006; Aungst et al., 2013), we next wished to determine when BDNF was upregulated following injury.

3.4.4. BDNF transcript and protein levels increase after Schaffer collateral lesion

In order to see area-specific changes in BDNF expression, we used qPCR to quantify bdnf transcripts shortly after injury. We observed a 2-fold increase in bdnf mRNA expression, which was solely observed in area CA3 and not in area CA1 (Fig. 3.5A). Furthermore, this 2-fold increase was sustained until 4 hours post-lesion and then downregulated to control levels by 6 hours (Fig.

3.5A). Conversely, full-length trkb transcript expression did not vary in either CA3 or CA1 up to

6 hours after transection (Fig. 3.5B). As we detected a transient change in expression of bdnf mRNA, we also wanted to know if this resulted in changes of BDNF protein levels post-injury.

140

Figure 3.5 – BDNF expression was upregulated 2 hours following injury. (A) Region specific mRNA expression of BDNF 2, 4 and 6 hours post-lesion (HPL). (B) Full-length TrkB transcript expression 2, 4 and 6 hours following injury. Note: for (A) and (B) n = 3 separate experiments of

5-6 pooled slices for each group, β-actin was used as a reference gene, *p < 0.05 (two-tailed students’ t-test with Bonferroni correction) and error bars are ± S.E.M. (C) Western blot of TrkB and BDNF protein expression at 24 HPL. (D) Western blot of TrkB after immunoprecipitation using a phospho-Trk antibody (Y515; an autophosphorylation site for dimerization) at 3 HPL,

Note: β-tubulin was used as a loading control for (C)-(D) and all Western blots were carried on as n = 3 in separate experiments of 10 pooled slices for each group. (E) The increase in pTrkB (Y515) was especially prominent in area CA3 at 3 HPL and many pTrkB puncta co-localized with PSD95, a marker for excitatory post-synaptic terminals. Scale bar = 5 µm. Note: representative images are maximum intensity projections rendered from confocal stacks immunostained for pTrkB and

141

PSD95. In the merge panel pTrkB is represented in red, PSD95 in green and colocalization is yellow. (F) Quantification of fluorescence intensity normalized to control (**p < 0.01; two-tailed students’ t-test and A.U. stands for arbitrary units).

142

Using Western blot analysis, we found an increase in BDNF protein expression 24 hours post- lesion, but no change in TrkB expression (Fig. 3.5C). Since we did not observe a change in TrkB expression, we next determined if BDNF was activating TrkB receptors by probing for autophosphorylation of TrkB at tyrosine 515 (Y515), a known site of TrkB autophosphorylation following BDNF binding TrkB. By performing a Western blot, we found an increase in pTrkB

(Y515) compared to controls at 3 hours following injury (Fig. 3.5D). Immunofluorescence revealed a 2-fold increase of pTrkB (Y515; Fig. 3.5E, F), which was mostly located adjacent to the lesion site (Fig. 3.5E), similar to the distribution of GAP43 immunoreactivity at 14 days post- lesion.

Taken together, our data show that lesioning Schaffer collaterals results in elevated BDNF and hyperexcitability. Next, we asked if BDNF is necessary for injury-induced sprouting.

3.4.5. BDNF inhibition downregulates axonal sprouting following injury

To determine whether BDNF is necessary for sprouting after Schaffer collateral lesion, we blocked BDNF signaling by chronically treating lesioned slices with TrkB-Fc, a BDNF scavenger, for 14 days post-lesion and assessed axonal sprouting with GAP43 immunofluorescence. We found that 31.3% (n = 5 of 16) of TrkB-Fc treated 14 days post-lesion cultures had GAP43 immunoreactive fibers in area CA3 (Fig. 3.6Dii, E), while 88.9% (n = 24 of 27) of untreated 14 days post-lesion cultures exhibited immunoreactive GAP43 axons in area CA3 (Fig. 3.6Cii, E).

Moreover, TrkB-Fc treated lesioned cultures were comparable to control sister cultures (Fig.

3.6Aii, E). Our results show that blocking BDNF reduced axonal sprouting following transection of Schaffer collaterals.

143

144

Figure 3.6 – Chronic BDNF blockade downregulates GAP43 expression 7, 14 and 21 days following lesion to Schaffer collaterals. (A) We detected little GAP43 immunoreactivity in control slices at 7 days (i) 14 days (ii) and 21 days (iii) after exposure to cutting solution. Scale bar = 50 μm; dotted line indicates site of lesion. (B) GAP43-immunopositive fibers were also not seen in TrkB-Fc treated control slices at 7 days (i) 14 days (ii) and 21 days (iii) following experiment onset. (C) At 7 DPL (i) a few GAP43-immunopositive fibers were observed, which disappeared by 21 DPL (iii); as seen earlier, GAP43 expression reached its peak at 14 DPL (ii).

(D) TrkB-Fc significantly decreased the expression of GAP43 immunoreactivity at 14 DPL (ii) and completely attenuated GAP43-labelled axonal sprouting at 7 DPL (i) and 21 DPL (ii). (B-D) are representative images of maximum intensity projections rendered from confocal stacks of slices immunohistochemically stained for GAP43 and all images were taken in area CA3. (E)

Percentage of cultures expressing GAP43. We found a significant increase in the percentage of slices expressing GAP43-immunopositive fibers 14 DPL and this was significantly decreased with

TrkB-Fc treatment (***p < 0.001; two-tailed Fisher’s exact test; note: control data were pooled from different time points).

145

However, it is possible that the TrkB-Fc treatment either accelerated or simply delayed the axonal sprouting. Therefore, we looked at GAP43 expression at 7 and 21 days post- injury. We found that when we treated lesioned slices for 7 days with TrkB-Fc, no cultures depicted GAP43 immunoreactivity (n = 5), in comparison to 37.5% of untreated 7 days post-lesion slices (n = 8;

Fig. 3.6Ci, Di, E). TrkB-Fc treated 7 days post-lesion cultures were comparable to control sister cultures (Fig. 3.6Ai, Di, E). Moreover, lesioned slices that were treated with TrkB-Fc for 21 days also did not depict any GAP43 immunoreactivity (0%; n = 6; Fig. 3.6Diii, E) and were comparable to control cultures (0%; n = 5; Fig. 3.6Aiii, E). However, 50.3% of 21 days post-lesion cultures depicted GAP43-immunopositive fibers (n = 8; Fig. 3.6Ciii, E), when GAP43 expression is downregulated indicating that the connections are maturing. Lastly, when we returned 14 days post-lesion cultures treated with TrkB-Fc containing media to control culture media for an additional week, we did not detect GAP43-immunopositive sprouting (data not shown). This indicates that there is an early time window during which BDNF can induce axonal sprouting and after this window is closed, axonal sprouting does not occur.

From these findings, we conclude that BDNF is necessary for lesion-induced sprouting, but is BDNF-induced axonal sprouting required for injury-induced hyperexcitability of CA3 pyramidal neurons?

3.4.6. BDNF inhibition blocks the development of hyperexcitability 14 days post- lesion

To test if BDNF-induced axonal sprouting is necessary for hyperexcitability, we recorded spontaneous action potential firing from CA3 cells. We found that 75% of lesioned cultures (Fig.

3.7A, B; n = 21 out of 28 cells) exhibited spontaneous action potentials 14 days post-transection,

146

Figure 7 – Chronic treatment with

TrkB-Fc for 14 DPL prevents the

development of hyperexcitability

following transection of Schaffer

collaterals. (A) Representative whole-

cell current clamp recordings. (B)

Percentage of cells depicting

spontaneous activity. The number of

slices with spontaneous action

potentials was significantly increased

in 14 DPL slices but was attenuated

with TrkB-Fc treatment, comparable to

control cultures. We found no change

in the 14 DPL IgG-Fc treated group

(**p < 0.01 between 14 DPL + IgG-Fc

and all groups excluding 14 DPL; *p <

0.05 between 14 DPL and all groups

excluding 14 DPL + IgG-Fc; Fisher’s

Exact Probability test, two-tailed). (C)

Quantification of action potential

frequency. The observed increase in

firing rate of CA3 neurons in 14 DPL

cultures was decreased in TrkB-Fc

147 treated lesioned slices, comparable to control sister cultures; no change was seen in 14 DPL treated with IgG-Fc (***p < 0.005 between 14 DPL and all groups excluding 14 DPL + IgG-Fc; *p < 0.05 between 14 DPL + IgG-Fc and all groups excluding 14 DPL; Kruskal-Wallis Test with post hoc

Mann-Whitney U-Test). (D) Percentage of cells with spontaneous bursting. The number of cultures that depicted spontaneous bursting after 14 DPL was increased, which was not seen in 14

DPL that were treated with TrkB-Fc. No change in the 14 DPL IgG-Fc treated group (*p < 0.05 between 14 DPL and all groups excluding 14 DPL + IgG-Fc; Fisher’s Exact Probability test, two- tailed). (E) Quantification of burst frequency. The observed increase in burst frequency of 14 DPL slices was attenuated with TrkB-Fc treatment 14 DPL cultures, which behaved like control CA3 neurons; no change was seen in 14 DPL slices treated with IgG-Fc (**p < 0.01 between 14 DPL and all groups excluding 14 DPL + IgG-Fc; *p < 0.05 between 14 DPL + IgG-Fc and all groups excluding 14 DPL; Kruskal-Wallis Test with post hoc Mann-Whitney U-Test). (C) and (E) are mean ± S.E.M. (F) Cumulative probability distribution of inter-spike-intervals demonstrated a leftward shift in the 14 DPL and 14 DPL IgG-Fc treated cultures (***p < 0.001 Kolmogorov-

Smirnov Test).

148 which led to an increase in firing rate (Fig. 3.7C; 4.11±1.07 action potentials per minute; n = 28 cells). This was attenuated in TrkB-Fc treated lesioned slices (Fig. 3.7A, B; 44%; n = 11 out of 25 cells), which depicted significantly fewer action potentials (Fig. 3.7A, C; 1.22±0.48 action potentials per minute; n = 25 cells). Treated lesioned slices were similar to control cultures (Fig.

3.7A, B; 40%; n = 8 out of 20 cells) with a spontaneous firing frequency of 0.46±0.16 action potentials per minute (Fig. 3.7C; n = 20 cells). TrkB-Fc treatment did not affect control cultures, of which 40.9% (Fig.3.7A, B; n = 9 out of 22 cells) of TrkB-Fc treated control cultures exhibited spontaneous firing (Fig. 3.7C; 0.68±0.28 action potentials per minute; n = 22 cells). To ensure that the Fc fragment was not itself affecting the functionality of the neurons, we treated lesioned sister cultures with IgG-Fc and we found that those cultures were comparable to untreated lesioned slices, as 87.5% (Fig. 3.7A, B; n = 14 out of 16 cells) were firing spontaneously (Fig. 3.7C;

5.44±2.64 action potentials per minute; n = 16 cells).

When we quantified the burst frequency (3 or more spikes were grouped into bursts if the inter-spike interval was ≤600 ms), we found that 67.9% (Fig. 3.7A, D; n = 19 out of 28 cells) of

14 days post-lesion cultures exhibited spontaneous bursting (Fig. 3.7E; 0.51±0.13 bursts per minute; n = 28 cells). This increase in excitability was significantly attenuated with TrkB-Fc treatment, where 28% of treated lesioned slices (Fig. 3.7A, D; n = 7 out of 25 cells) were bursting at a frequency of 0.10±0.04 bursts per minute (Fig. 3.7E; n = 25 cells). These slices were similar to controls, where 20% (Fig. 3.7A, D; n = 4 out of 20 cells) of sister control slices exhibited spontaneous bursts (Fig. 3.7E; 0.06±0.02 bursts per minute; n = 20 cells). We observed spontaneous bursts in 31.8% (Fig. 3.7A, D; n = 7 out of 22 cells) of TrkB-Fc treated controls (Fig.

3.7E; 0.10±0.04 bursts per minute; n = 22 cells). Lastly, IgG-Fc treated lesioned cultures were comparable to untreated lesioned slices, where 50% (Fig. 3.7A, D; n = 8 out of 16 cells) of cells

149 showed spontaneous bursting at a frequency of 0.24±0.07 bursts per minute (Fig. 3.7E; n = 16 cells).

When the inter-spike intervals were plotted as a cumulative probability distribution, we found a leftward shift in the distribution of 14 days post-lesion and IgG-Fc treated lesioned slices

(Fig. 3.7F), whereas TrkB-Fc treated lesioned slices were comparable to control and TrkB-Fc treated control cultures (Fig. 3.7F). This indicated the inter-spike intervals were much shorter in

14 days post-lesion and IgG-Fc treated lesioned slices compared to other conditions.

This evidence strongly indicates that BDNF blockade attenuates Schaffer collateral injury-induced hyperexcitability of individual CA3 neurons. Next, we wished to determine if BDNF blockade could inhibit the recurrent network formation.

3.4.7. Chronic treatment with TrkB-Fc for 14 days after lesion inhibits lesion induced hyperexcitable network activity in area CA3

To test if BDNF was important for sprouting-induced network activity, we carried out evoked field potential recordings. We observed that there was an increase in the number of compounded fEPSPs in relation to the stimulation strength in area CA3 of 14 days post-lesion slices versus control and this was attenuated with TrkB-Fc treatment (Fig. 3.8A, B). To control for non-specific effects of the Fc fragment, we treated lesioned sister cultures with IgG-Fc and we found no difference versus untreated lesioned slices (Fig. 3.8A, B). Moreover, unlesioned TrkB-

Fc treated slices behaved like control (Fig. 3.8A, B). Control cultures typically depicted non- compounded fEPSPs at any given moment (Fig. 3.8A), with only 50% of slices showing compounded fEPSPs (Fig. 3.8C; n = 8); this was comparable in TrkB-Fc treated control slices

150

Figure 3.8 – Chronic treatment with TrkB-Fc for 14 DPL prevents lesioned-induced increase in network activity. (A) Representative traces of evoked fEPSP recordings. Note: arrows depict additional peaks in compounded fEPSP responses and scale bars correspond to values of control scale bar. (B) Quantification of the number of peaks as a function of stimulation strength. We found that 14 DPL slices required less stimulation than control slices to induce compound fEPSPs.

This was not observed in 14 DPL slices treated with TrkB-Fc, which were comparable to control unlesioned TrkB-Fc treated sister cultures (*p < 0.05; linear regression was performed for each slice and then slopes were compared using one-way ANOVA with post hoc Tukey’s test, following which we performed two-tailed independent Student’s t-test for pairwise comparisons between data points). Moreover, IgG-Fc treated lesioned slices were similar to 14 DPL slices. (C)

Quantification of percentage of cultures depicting compounded fEPSPs. We found that all 14 DPL

151 slices depicted compounded fEPSPs. Though, TrkB-Fc treated slices were comparable to control and TrkB-Fc treated control slices. IgG-Fc treated 14 DPL slices were similar to 14 DPL slices

(*p < 0.05 between 14 DPL and all groups excluding 14 DPL + IgG-Fc; vice versa, between 14

DPL + IgG-Fc and all groups excluding 14 DPL; Fisher’s Exact Probability test, two-tailed). Note: control and 14 DPL data are repeated from Fig. 4B, C for the purpose of comparison.

152

(Fig. 3.8B, C; 62.5%; n = 8). However, we detected compounded fEPSPs in 100% 14 days post- lesion cultures (Fig. 3.8B, C; n = 11) and 100% of IgG-Fc treated lesioned cultures (Fig. 3.8B, C; n = 9), indicating that there were increased excitatory inputs within CA3, after lesioning.

Interestingly, TrkB-Fc treatment of lesioned slices were comparable to control (Fig. 3.8B, C; 30%; n = 10). In addition, we quantified individual spontaneous excitatory postsynaptic potentials

(sEPSPs) and we found an increase in the number of sEPSPs in 14 days post-lesion slices (Fig.

3.9A, B; 33.84 ± 4.02; n = 26) compared with control (Fig. 3.9A, B; 22.98 ± 3.56; n = 20); this increase was blocked by TrkB-Fc treatment (Fig. 3.9A, B; 22.33 ± 3.23; n = 24). We did not see a change in sEPSPs of CA3 pyramidal cells of control slices treated with TrkB-Fc (Fig. 3.9A, B;

18.0 ± 2.53; n = 22). Therefore, our data demonstrates that lesion-induced hyperexcitability and increased network activity are attributable to BDNF-induced axonal sprouting.

153

Figure 3.9 – Spontaneous EPSPs of CA3 Pyramidal Neurons Increase After Schaffer

Collateral Transection. (A) Representative traces of whole-cell patch electrophysiological recordings of CA3 pyramidal neurons. (B) Quantification of sEPSPs. We found an increase in sEPSPs 14 DPL, compared to control cultures (*p < 0.05, two-tailed students’-test). Furthermore, we found this increase in sEPSPs was attenuated with TrkB-Fc treatment for 14 DPL; we did not see a change in sEPSPs of CA3 pyramidal cells of control slices treated with TrkB-Fc. Note: data are depicted as mean ± S.E.M.

154

3.5. DISCUSSION

Here, we demonstrate that following Schaffer collateral transection in mouse organotypic hippocampal slices, there is an increase in GAP43-immunopositive axon collaterals specific to area CA3 close to the lesion site, which coincides with an increase in extrinsic excitability of CA3 pyramidal neurons coupled with an increase in network activity. Pharmacological inhibition of

BDNF attenuated the observed morphological and functional changes. Furthermore, we show that

BDNF is increased in a tight therapeutic time window and activates TrkB receptors, which we measured via phosphorylation. Therefore, we show that BDNF is necessary for lesion-mediated axonal sprouting and the subsequent increase in excitation.

3.5.1. Time course of BDNF expression following injury

The present results also reveal the timeframe for bdnf mRNA upregulation following

Schaffer collateral transection in area CA3. We found that transection of Schaffer collaterals lead to a 2-fold change in bdnf mRNA transcripts at 2 and 4 hours following injury, which was specific to area CA3. Interestingly, no changes were observed in area CA1, consistent with the observation that GAP43-immunoreactivity was not observed in area CA1 at any of the time point tested following lesion. Moreover, the post-lesion change in mRNA expression in area CA3 led to an increase in BDNF protein expression 24 hours after transection, consistent with the previous finding that BDNF expression was increased at 24 and 48 hours after Schaffer collateral transection (Dinocourt et al., 2006). Interestingly, the BDNF increase immediately precedes

GAP43 upregulation and axonal sprouting. Moreover, upregulation of BDNF in the hippocampus over a similar time frame has been reported with in vivo animal models of brain injury (Mudo et

155 al., 1993; Hicks et al., 1997; Hicks et al., 1998; Grundy et al., 2000; Griesbach et al., 2002), in children following severe head trauma (Chiaretti et al., 2003) and in vivo kainic acid injury (Zafra et al., 1990; Dugich-Djordjevic et al., 1992a; Dugich-Djordjevic et al., 1992b). Since BDNF mRNA and protein expression was increased after injury, we tested for TrkB expression at 24 hours post-injury and did not observe any change in expression. However, a previous study reported increased TrkB expression between 24 and 48 hours after Schaffer collateral lesion

(Dinocourt et al., 2006). Since, intracellular phosphorylation of TrkB receptors occurs after brain injury in rodents (Binder et al., 1999; Hu et al., 2004), we tested for TrkB receptor activation by phosphorylation of Y515, a residue that mediates Shc binding following activation of the TrkB receptor (Huang & Reichardt, 2001). Trk receptor activation of the Shc pathway is responsible for the local axon outgrowth effects of neurotrophins via Ras and ERK second messengers (Huang &

Reichardt, 2003). We found a 2-fold increase in pTrkB (Y515), 3 hours post-lesion, which was specifically localized in area CA3. Therefore, our finding that there is an increase of pTrkB (Y515)

3 hours following Schaffer transection indicates that the Shc pathway has been activated and may underlie the observed axon outgrowth. Taken together with the temporally restricted increase in

BDNF, our results suggest that BDNF induces axonal sprouting following Schaffer collateral lesion.

3.5.2. BDNF initiates axonal sprouting of CA3 pyramidal neurons after lesion

To determine if the observed increase in BDNF is necessary for axonal sprouting we interfered with the increase in BDNF by TrkB-Fc treatment to lesioned cultures. We found a marked downregulation of GAP43 expressing axon fibers at 7, 14, 21 days post-lesion compared to untreated lesion cultures. This finding builds on Dinocourt and colleagues, who showed that

156 partial knock-down of TrkB receptors downregulated axonal sprouting 5 days after Schaffer collateral injury (Dinocourt et al., 2006). Since we chronically blocked BDNF for up to 21 days after transection, our data indicates that at least for 3 weeks, no compensatory mechanism was initiated to overcome the BDNF blockade to induce axonal sprouting. Exactly how BDNF regulates this aberrant sprouting is currently unknown. However, on the dendritic side, emerging evidence shows that TrkB may regulate synapse number in early development by stimulating filopodial motility (Luikart et al., 2008) and increasing the probability that a dendritic filopodium will encounter a nearby axon. This suggests that BDNF facilitates excitatory synaptogenesis, which could disrupt the subtle balance between excitation and inhibition in favor of excitation to yield hyperexcitability. It has also been shown that BDNF can enhance glutamatergic synaptic transmission (Lu, 2003; Gottmann et al., 2009) and therefore BDNF may potentially exacerbate the hyperexcitability we found following Schaffer collateral injury.

Interestingly, BDNF levels can also enhance the formation of GABAergic synapses

(Vicario-Abejon et al., 1998; Marty et al., 2000; Rico et al., 2002) and increase the expression of

GAD67 in cultured hippocampal neurons (Bolton et al., 2000; Yamada et al., 2002) and GAD65 in organotypic hippocampal slices (Marty et al., 2000). Here, we observed no change in the expression of GAD67, though we did not test for the expression of GAD65. However, whether there is an alteration of inhibitory transmission after Schaffer collateral injury to organotypic slices remains to be determined. BDNF can affect GABAergic transmission by regulating cell surface expression of GABAA receptor subunits (Brunig et al., 2001; Jovanovic et al., 2004; Kanematsu et al., 2006); it is unknown whether there is a change in GABAA receptor subunit expression after

Schaffer collateral injury. In addition, exogenous application of BDNF to primary hippocampal neurons can increase the frequency (Bolton et al., 2000) and amplitude (Brunig et al., 2001) of

157 miniature inhibitory post-synaptic currents. Previously we have shown that there is no change in miniature inhibitory events 14 days following Schaffer collateral lesion in rat organotypic hippocampal slices, indicating that inhibitory drive has not changed at least in a rat culture system

(McKinney et al., 1997).

3.5.3. Chronic TrkB-Fc treatment attenuates injury-induced hyperexcitability and network activity

As axonal sprouting caused by Schaffer collateral transection increases the connectivity between CA3 pyramidal neurons, it may facilitate the synchronous discharge of the CA3 cell population (McKinney et al., 1997). Therefore, we tested the hypothesis that BDNF-mediated axonal sprouting can lead to hyperexcitability of excitatory CA3 pyramidal neurons 14 days after transection. We found that chronically treated TrkB-Fc slices had less spontaneous action potentials compared to untreated lesioned slices. Moreover, Schaffer collateral lesion caused an increase in compound fEPSPs in area CA3, consistent with epileptiform activity and hyperexcitability, comparable to in vivo Schaffer collateral lesion (Aungst et al., 2013). This change in network activity was also attenuated with TrkB-Fc treatment. Interestingly, epileptogenesis in a kindling model is much more difficult to induce in BDNF heterozygous knockout mice (Kokaia et al., 1995) or with pre-treatment of TrkB-Fc (Binder et al., 1999).

Moreover, BDNF over expressing transgenic mice can have spontaneous seizures (Croll et al.,

1999). Taken together, this indicates BDNF-mediated local microcircuitry remodeling may be important for epileptogenesis.

This may have implications for post-traumatic epilepsy, as recurrent epileptic seizures are a common clinical consequence of traumatic brain injury (Annegers et al., 1998). Others have

158 shown that Schaffer collateral lesion can result in increased network activity after the addition of a mildly proconvulsive amount of bicuculline (0.1 μM), a GABAA receptor antagonist (Aungst et al., 2013). Moreover, evidence from in vivo models of epilepsy has shown that axonal sprouting can occur in the hippocampus after induction of status epilepticus using convulsants. This is well characterized in hippocampal granule cells, where an insult leads to aberrant mossy fiber sprouting and a recurrent network with increased excitatory drive (Tauck & Nadler, 1985; Cronin et al.,

1992; Okazaki et al., 1999). This occurrence has also been described in patients with temporal lobe epilepsy (de Lanerolle et al., 1989; Sutula et al., 1989; Houser, 1990) and in rodent models of head trauma (Golarai et al., 2001; Santhakumar et al., 2001; Dinocourt et al., 2011). As there is considerable evidence correlating axonal sprouting and hyperexcitability in epilepsy models,

BDNF maybe a potential therapeutic target.

3.5.4. Trigger and target for BDNF release

Our evidence implicates BDNF in injury-related axonal sprouting and hyperexcitability.

However, little is known about the cell types responsible for the observed increase in BDNF and the location of the essential TrkB receptors remains unknown. Both BDNF and TrkB expression are widespread in the hippocampus and TrkB receptors are found on both neurons and astrocytes

(Ernfors et al., 1990b; Wetmore et al., 1990; Armanini et al., 1995). Interestingly, there is an induction of non-catalytic truncated trkb mRNA in glial cells 6-14 days following lesion of the perforant pathway (Beck et al., 1993). As truncated TrkB receptors sequester and internalize

BDNF (Biffo et al., 1995; Alderson et al., 2000), it could be possible that unique intracellular pathways are activated through this truncated TrkB-mediated internalization, leading to axonal sprouting.

159

Emerging evidence shows that activity-dependent transcription of BDNF is under epigenetic regulation (Martinowich et al., 2003; Zhou et al., 2006; Bredy et al., 2007). A common epigenetic modification is methylation of cytosine residues that are adjacent to guanine residues

(CpG) (Klose & Bird, 2006). Methyl-CpG binding protein 2, MeCP2, can bind to methylated CpG sites and repress transcription (Jones et al., 1998; Nan et al., 1998), such as in the case of the bdnf gene (Martinowich et al., 2003). It has been shown that Ca2+-dependent phosphorylation of MeCP2 increases transcription of bdnf (Chen et al., 2003). Since hippocampal lesion can lead to excess glutamate release and increased Ca2+ intake by neurons (Muller et al., 2010), it would be interesting to probe the involvement of MeCP2 in injury-induced BDNF release.

In summary, our study provides direct evidence that BDNF is critical for the development of new axons and increased excitation after injury to the pyramidal cell axon. Furthermore, we observed that BDNF inhibition attenuated lesion-induced hippocampal hyperexcitability and changes in network activity. The evidence shown in the present study suggests that TrkB-BDNF signaling could be a novel therapeutic target. However, as BDNF plays an essential role for learning and memory it would be important to study signaling cascades downstream of BDNF to develop viable treatments targeting post-traumatic epilepsy.

160

CHAPTER 4. INHIBITION OF METHYL CPG BINDING PROTEIN 2 PHOSPHORYLATION DOES NOT PREVENT INJURY-INDUCED HYPEREXCITABILITY

FOREWORD

In Chapter 3, we identified that bdnf mRNA expression increases 2 hours post-lesion in area CA3 after Schaffer collateral lesion. Moreover, inhibiting BDNF and other TrkB ligands with

TrkB-Fc prevents axonal sprouting and epileptiform activity. However, given that BDNF is important for cell survival and synaptic plasticity, it would be difficult to develop therapies targeting BDNF-TrkB signaling to prevent axonal sprouting in the hippocampus after traumatic brain injury in humans. Therefore, in Chapter 4 I set out to identify other more specific therapeutic targets. As BDNF is known to be modulated by physiological and seizure-like activity in vivo, I wanted to determine if there was any role for activity-dependent transcription of BDNF in hyperexcitability. This transcription is dependent on phosphorylation of MeCP2, a transcriptional repressor, on serine 421, which acts as a molecular switch to induce bdnf transcription.

Given this, I hypothesized that Schaffer collateral lesion causes the phosphorylation of MeCP2 at S421, relieving transcriptional repression of bdnf, causing axon sprouting- induced hyperexcitability of CA3 pyramidal neurons.

In order to test this hypothesis, I made Schaffer collateral lesions in organotypic cultures, in which we have previously shown can induce axon sprouting and CA3 neuron hyperexcitability.

I tested for MeCP2 phosphorylation by using a specific antibody targeting pMeCPS421 and this phosphorylation event was blocked using several specific inhibitors of Ca2+ mediators (CaMKII,

161

CaMKK, NMDA receptors and voltage gated Ca2+ channels). Lastly, hyperexcitability of CA3 neurons was tested using whole-cell current clamp recordings of spontaneous action potentials.

162

4.1. ABSTRACT

Axonal sprouting induced by brain injury, can often lead to recurrent synapse formation and hyperexcitability of excitatory neurons. Previously, we demonstrated that bdnf mRNA expression increases in area CA3 after Schaffer collateral lesion. Moreover, scavenging BDNF prevents lesion-induced axon sprouting and CA3 pyramidal neuron hyperexcitability. Since BDNF is important for synaptic plasticity, it remains an inappropriate target to prevent axon sprouting.

Therefore, we must identify other more specific therapeutic targets to prevent BDNF-induced axonal sprouting. Interestingly, BDNF transcription is regulated by methyl CpG binding protein 2

(MeCP2) in an activity-dependent manner, through the phosphorylation of serine 421 of MeCP2

(pMeCP2S421). Given this, we hypothesized that pMeCP2S421-dependent bdnf transcription is misregulated following Schaffer collateral transection leading to hyperexcitability caused by axonal sprouting. In order to address this hypothesis, we made Schaffer collateral lesions to organotypic hippocampal slice cultures and probed for pMeCP2S421. We then prevented phosphorylation of MeCP2 by targeting CaMKII and tested for CA3 pyramidal neuron hyperexcitability. We found that inhibiting CaMKII did not prevent lesion-induced CA3 hyperexcitability, suggesting that pMeCP2S421 does not regulate BDNF-mediated axonal sprouting.

163

4.2. INTRODUCTION

Head trauma, particularly penetrating brain injury, often results in axonal sprouting.

Synapses formed by sprouting axons can be aberrant and induces the formation of a recurrent network leading to hyperexcitability of target neurons (Nadler, 2003; Sutula & Dudek, 2007).

Interestingly, this mechanism can be activated following traumatic brain injury in patients, a subset of whom develop post-traumatic epilepsy (PTE) (Frey, 2003). Using an established in vitro model of PTE (McKinney et al., 1997), we have shown that Schaffer collateral lesion of organotypic hippocampal slices can increase bdnf mRNA expression, a neurotrophin important for synaptic plasticity and neuronal development (Huang & Reichardt, 2001), as early as 2 hours post-lesion in area CA3 (Gill et al., 2013). Moreover, blocking the TrkB receptor of BDNF can abolish axon sprouting and hyperexcitability of CA3 pyramidal neurons (Gill et al., 2013).

However, as mentioned above, BDNF is important for cell survival and synaptic plasticity, making BDNF and TrkB inappropriate therapeutic targets to prevent axonal sprouting in the hippocampus after traumatic brain injury in patients. Thus, we must identify other, more specific therapeutic targets that can prevent BDNF upregulation following hippocampal injury.

BDNF transcription can occur in an activity-dependent manner. In fact, it has been demonstrated that in many instances activity-dependent transcription of BDNF is under epigenetic regulation (Martinowich et al., 2003; Zhou et al., 2006; Bredy et al., 2007). One such epigenetic mechanism is regulated by methylation of cytosine residues that are adjacent to guanine residues

(methylated CpGs) (Klose & Bird, 2006). Methyl-CpG binding protein 2 (MeCP2) can bind to methylated CpG sites and repress transcription (Jones et al., 1998; Nan et al., 1998), such as in the case of the bdnf gene (Martinowich et al., 2003). It has been shown that Ca2+-dependent

164 phosphorylation of MeCP2 at serine 421 (pMeCP2S421) increases transcription of bdnf (Chen et al., 2003). S421 phosphorylation of MeCP2 causes MeCP2 to dissociate from the promoter region and allow bdnf transcription (Zhou et al., 2006). Furthermore, pMeCP2S421 is regulated by activity- induced Ca2+ influx, occurring from many sources including NMDA receptors or L-type voltage gated Ca2+ channels (Fig. 4.1) (Chen et al., 2003; Tao et al., 2009). Ca2+ entry can activate CaMKII and CaMKIV via Ca2+/Calmodulin kinase (CaMKK) (Zhou et al., 2006; Tao et al., 2009), kinases important for synaptic plasticity. CaMKII or CaMKIV can then directly phosphorylate MeCP2 in the nucleus (Buchthal et al., 2012), changing its affinity for the promoter region, dissociating itself from the bdnf exon IV promoter and allowing its transcription.

Given that hippocampal lesions can lead to excess glutamate release and increased Ca2+ intake in neurons (Muller et al., 2010), we hypothesized that this increase could induce the Ca2+- dependent phosphorylation of MeCP2 and may underlie an early signaling cascade that increases bdnf expression to induce hyperexcitability of CA3 pyramidal neurons.

In order to test our hypothesis, we mimicked a penetrating head trauma by lesioning the

Schaffer collateral pathway, which induces axon sprouting and subsequent CA3 network reorganization (McKinney et al., 1997; Gill et al., 2013). We then probed for pMeCP2S421 using immunofluorescence and imaged in area CA3 using confocal microscopy. CA3 pyramidal neuron hyperexcitability was assessed by recording spontaneous firing of action potentials. We used several pharmacological inhibitors to prevent pMeCP2S421, as outlined in Fig. 4.1.

Here, we demonstrate that blocking NMDA receptors, CaMKII, CaMKK and L-type

2+ voltage gated Ca channels can all decrease pMeCP2S421 in area CA3. However, after choosing

CaMKII inhibition as an ideal target to block pMeCP2S421, we found that inhibiting CaMKII did not prevent Schaffer collateral-lesion induced CA3 hyperexcitability.

165

Figure 4.1 – Activity-dependent regulation of Bdnf exon IV. (A) Schematic of Ca2+ influx by neuronal activity leading to phosphorylation of MeCP2 at S421. (B) Bdnf gene locus contains many exons with at least 12 alternatively spliced transcripts, all encoding for one protein. Note: text in red indicates small molecule inhibitors that target specific steps of cascade. Schematic modified from (Hong et al., 2005)

166

4.3. MATERIALS AND METHODS

4.3.1. Ethics Statement

All animal handling procedures were carried out consistent with guidelines set by the

4.3.2. Hippocampal Slice Cultures and Schaffer Collateral Lesions

Organotypic hippocampal slices were prepared using the roller-tube method, as previously described (Gahwiler, 1981; Gahwiler et al., 1997). Briefly, 7-day-old C57B6 mice were decapitated and a bilateral hippocampal dissection was performed, from which 400 µm thick transverse hippocampal slices were made and plated on glass coverslips. Slices were adhered in place with a clot of chicken plasma (Cocalico Biologicals; Reamstown, PA, USA) coagulated with thrombin (Invitrogen GIBCO). Coverslips were placed in flat-sided culture tubes with antibiotic- free serum-containing media and maintained in a dry-air roller drum incubator at 36°C for three weeks prior to experimentation. To prevent excitotoxic injury during lesioning, slices (> 21 days in vitro) were placed into cutting solution containing 0.5 μM TTX (Alomone Labs, Jerusalem,

Israel) and 10 mM MgCl2. Lesions were made to the Schaffer collaterals between area CA3 and

CA1 with a razor blade shard; cuts extended from the alveus layer to stratum radiatum. Control sister cultures were also exposed to the cutting media for the same period of time, but not lesioned.

Cultures were returned to normal culture media and processed simultaneously in all experiments.

167

4.3.3. Immunohistochemistry

Slices were fixed at 4°C overnight in 4% paraformaldehyde made up in 0.1 M PB, pH 7.4.

Following fixation, slices were washed in 0.1 M PB, permeabilized in 0.4% Triton X-100 and blocked with 1.5% heat inactivated horse serum overnight at 4°C. Primary antibodies were incubated for five days at 4°C in permeabilizing buffer with a 1:250 dilution of: rabbit anti- pMeCP2S421 (Abgent, San Diego, CA, USA), rabbit anti-MeCP2 (Millipore), mouse anti-GFAP

(Sigma- Aldrich) and anti-NeuN tagged to Alexa Fluor 488 (Millipore). Following several 0.1 M

PB washes, slices were incubated overnight at 4°C in 1:250 of either secondary Alexa Fluor 594 anti-rabbit (Invitrogen Molecular Probes) or Alexa Fluor 488 anti-mouse (Invitrogen Molecular

Probes) antibody diluted in 0.1 M PB containing 1.5% heat inactivated horse serum. Finally, slices were mounted with DAKO Fluorescent Mounting medium (Dako Canada; Mississauga, Canada) onto microscope slides. Following mounting, slices were imaged in area CA3 using a Leica TCS

SP2 scanhead (Leica Microsystems) on a Leica DM6000 B upright microscope, equipped with a

HCX PL APO 40× NA 1.4 oil immersion objective using a 543 nm HeNe laser line. Image stacks were collected at Z = 0.4 μm and averaged 4× to improve signal-to-noise ratio.

Immunofluorescence was quantified using ImageJ software from confocal stacks that were obtained with identical parameters (laser intensity, filters, pinhole size, photomultiplier tube gain and offset). Representative images are maximum intensity projections of 5 sections from each stack.

168

4.3.4. Electrophysiological Recordings and Analysis

CA3 neurons (up to 2 cells/slice) in current-clamp mode. Whole-cell recording electrodes were pulled from borosilicate glass (resistance of 4-6 MΩ; GC150TC; Clark Instruments, UK) using a

P-97 electrode puller (Stutter Instrument Co., Novato, CA, USA). All electrophysiological recordings were made using an Axopatch 200A amplifier (Molecular Devices, Sunnyvale, CA,

USA). Signals were recorded at 20 kHz and then filtered through a Bessel low-pass filter at 2 kHz using the Clampex 9.2 acquisition program (Axon Instruments Inc.). Recording pipettes were filled with intracellular solution containing: 120 mM K-Gluconate, 1 mM EGTA, 10 mM HEPES, 5 mM

Mg-ATP, 0.5 mM Na-GTP, 5 mM NaCl, 5 mM KCl, 10 mM phosphocreatine, and pH was adjusted to 7.3 with KOH, and 285-295 mOsm. CA3 neurons were chosen at approximately 250

μm from the site of lesion in the case of transected slices. Once whole-cell configuration was established, cells were monitored for 5 minutes for equilibration of internal solution and to ensure that the seal and opening were maintained; cells were then recorded for 20 minutes. Input resistance and membrane potential were monitored at regular intervals throughout the experiment and no observable differences were noted between conditions. Data were discarded if recordings in cells showed a seal resistance <5 GΩ, a holding current >-100 pA, or resistance deviated more than ±20% through the course of recording. Also, cells that failed to last the entire 20 minutes of recording were excluded from further analysis. The digitized data was coded prior to analysis so that analysis could be carried out blinded. Upward deflecting EPSPs greater than 40 mV in

169 amplitude were considered to be somatic action potentials and were detected offline. Action potential threshold was identified as the membrane potential at the start point of the first action potential. Three or more action potentials were grouped into bursts if the inter-spike interval was smaller than 600 ms.

4.3.5. Pharmacological Treatment

The following drugs were used: 10 μM of KN-62 (inhibits CaMKII), 15 µM 3-[(R)-2- carboxypiperazin-4-yl]-propyl-1-phosphonic acid (CPP; inhibits NMDA receptors), 10 µM STO-

609 (inhibitor of CaMKK) or 20 µM nimodepine (inhibits L-type voltage gated Ca2+ channels).

All drugs were obtained from Tocris Biosciences and were diluted in dimethyl sulfoxide. Cultures were treated for 1 hour or 14 days following Schaffer collateral transection, after which slices were processed for immunohistochemistry or electrophysiology. Sister cultures were treated with normal culture media containing an equivalent amount of dimethyl sulfoxide.

4.3.6. Statistical Analysis

Two-way independent student’s t-test was used to compare <2 groups to control. When additional groups were present all groups were compared to control using a one-way ANOVA with post-hoc Dunnett’s test. Results are expressed as mean ± S.E.M and n values represent number of slices, unless otherwise stated.

170

4.4. RESULTS

Here, we set out to determine the cellular mechanism of BDNF upregulation following

Schaffer collateral lesion, mimicking a penetrating head injury. Using an established model of post-traumatic epilepsy (McKinney et al., 1997), we have shown that Schaffer collateral lesion leads to an increase in bdnf mRNA expression as early as 2 hours post-lesion (HPL) and blocking its TrkB receptor ligands, can abolish the development of axonal sprouting and hyperexcitability within area CA3 (Gill et al., 2013). Given that BDNF is an important modulator of cell survival and synaptic plasticity, it would be difficult to target BDNF-TrkB signaling outright as method of pharmacological intervention to inhibit axonal sprouting and recurrent network formation after brain injury. Therefore, we wished to identify other targets that could be used to inhibit a specific pool of BDNF, such as targeting the activity-dependent transcription of bdnf.

4.4.1. MeCP2 is phosphorylated at S421 in neurons and a subset of astrocytes as early as 1 hour following Schaffer collateral transection in area CA3

Given the role of MeCP2 in activity-dependent transcription of bdnf, we first probed pMeCP2S421 at 1 and 3 hours following Schaffer collateral transection to organotypic hippocampal slice cultures by immunofluorescence. We found a significant increase in the mean fluorescence intensity of pMeCP2S421 at 1 HPL (Fig. 4.2A, B; 1.62 ± 0.28 A. U.; n = 6 slices; *p < 0.05, two- way independent student’s t-test) in area CA3, which was sustained for at least 3 HPL (Fig. 4.2A,

B; 1.61 ± 0.19; n = 6 slices), compared to control (Fig. 4.2A, B; 1.00 ± 0.13 A.U.; n = 6 slices; note: all values were normalized to control). Additionally, close to the lesion site we observed some pMeCP2S421 expression also in area CA1.

171

Figure 4.2 – Schaffer collateral lesion leads to increased phosphorylation of MeCP2 at S421.

(A) Example maximum intensity projections of organotypic hippocampal slice culture which have been immunostained for pMeCP2S421 and NeuN, a marker for neuronal nuclei. Slices were imaged in area CA3. Scale bar = 20 μm and dotted line demarcates lesion site. Notes: green fibers are

GFP-expressing mossy fibers. (B) Quantification of pMeCP2S421 mean fluorescence intensity in area CA3 at 1 HPL (n = 6 slices; *p < 0.05, two-way independent student’s t-test), 3 HPL (n = 6 slices) and control (n = 6 slices; all values were normalized to control). Note: arrowheads denote pMeCP2S421-immunopositive cell bodies that do not colocalized with NeuN.

172

In addition, we observed that the majority of pMeCP2S421 colocalized with NeuN, a marker for neuronal nuclei, indicating that MeCP2 primarily resides within neuronal nuclei. However, a subset of pMeCP2S421-immunpositive nuclei were not colocalized with NeuN, we hypothesized they may be within astrocytes. Therefore, we immunostained lesioned slices with glial fibrillary acidic protein (GFAP), a marker for astrocytes, and found that a subset of GFAP-immunopositive cells colocalized with pMeCP2S421 (Fig. 4.3A, B), particularly in astrocytes adjacent to the lesion.

Next, we set out to determine whether the total protein expression of MeCP2 was altered by Schaffer collateral lesion, which could underlie the increase in phosphorylation of MeCP2 that we had observed. Therefore, at 1 and 3 HPL we immunostained control and lesioned slices with a

MeCP2 antibody that recognizes both phosphorylated and unphosphorylated forms of MeCP2. We found that overall protein expression for MeCP2 did not change at 1HPL in area CA3 (Fig. 4.4A,

B; 0.94 ± 0.059 A. U.; n = 5 slices) or 3HPL (Fig. 4.4A, B; 1.04 ± 0.11 A. U.; n = 5 slices), compared to control (Fig. 4.4A, B; 1.00 ± 0.11 A. U.; n = 5 slices; all data were normalized to control). Indicating that overall pMeCP2S421 increased relative to total MeCP2 in area CA3 after lesion. Subsequently, we wished to block the phosphorylation of MeCP2 to test if this blocked

BDNF-induced hyperexcitability within area CA3 after Schaffer collateral lesion.

4.4.2. Selective pharmacological targeting of MeCP2 Ca2+-dependent phosphorylation

We selectively targeted different sources of Ca2+ within neurons that have been shown to activate Ca2+-dependent phosphorylation of MeCP2. As seen in Fig. 4.1, we targeted 3 key steps of Ca2+ entry into neurons with 4 selective small molecule inhibitors: (1) NMDA receptors were inhibited with 15 µM of CPP, (2) L-type voltage-gated Ca2+ channels were inhibited with 20 µM

173

Figure 4.3 – pMeCP2S421 also colocalizes to a subset of GFAP-labelled astrocytes adjacent to the lesion site. (A) Example maximum intensity projection obtained from a hippocampal slice culture 6 HPL (n = 4 slices) that was immunostained for pMeCP2S421 and GFAP, a marker for astrocytes. Scale bar = 30 μm. (B) Inset of (A) demonstrating a GFAP-immunopositive astrocyte colocalizing with pMeCP2S421. Arrow demarcates astrocyte cell body. Scale bar = 5 μm.

174

Figure 4.4 – Total MeCP2 expression is unchanged in area CA3 following Schaffer collateral lesion. (A) Example maximum intensity projections of organotypic hippocampal slice cultures which have been immunostained for MeCP2 and NeuN. Slices were imaged in area CA3. Scale bar = 20 μm and dotted line demarcates lesion site. (B) Quantification of immunofluorescence of total MeCP2 at 1 HPL (n = 5 slices), 3 HPL (n = 5 slices) and control (n = 5 slices; all values were normalized to control).

175 of nimodepine, (3) CaMKK were inhibited with 10 µM of STO-609, which is known to block both

CaMKIV and CaMKII and lastly, (4) CaMKII was selectively inhibited with 10 µM KN-.

Treatment began immediately following lesion and 1 hour after treatment slices were immunostained for pMeCP2S421. We found an approximately 3-fold increase in pMeCP2S421 immunofluorescence in area CA3 in 1 HPL slices (Fig 4.5A, B; n = 8 slices; ***p < 0.001, one- way ANOVA with post-hoc Dunnett’s test) compared to control (Fig 4.5A, B; n = 10 slices). In addition, the increase in pMeCP2S421 was attenuated by KN-62 (Fig 4.5A, B; n = 8 slices) and CPP

(n = 8 slices) treatment. Both nimodepine (n = 8 slices; *p < 0.05, one-way ANOVA with post- hoc Dunnett’s test) and STO-609 (n = 8 slices) significantly downregulated pMeCP2S421 when compared to untreated lesioned cultures, but did not completely downregulate to control levels (*p

< 0.05, one-way ANOVA with post-hoc Dunnett’s test). This may potentially have been due to a lower effective concentration of STO-609 given that it has a low cell permeability with DMSO and potentially for both nimodepine and STO-609 1 hour of treatment may have been insufficient for these inhibitors to penetrate the dense neuropil. Given that we have previously shown that chronic NMDA receptor inhibition itself induces axonal sprouting and CA3 hyperexcitability in organotypic hippocampal slices (McKinney et al., 1999b), we did not further consider NMDA receptor inhibition as a candidate. We then proceeded to test whether KN-62 treatment would prevent the increase in firing of CA3 neurons, as previously seen (Gill et al., 2013).

4.4.3. Inhibition of CaMKII does not prevent Schaffer collateral lesion-induced increase in firing of CA3 neurons

We next blocked CaMKII with 10 μM KN-62, for 14 days following Schaffer collateral transection, a time point at which we have previously observed an increase CA3 neuron firing,

176

2+ Figure 4.5 – Selective inhibition of Ca sources can downregulate pMeCP2S421 in area CA3.

(A) Example maximum intensity projections of organotypic hippocampal slice cultures which have been immunostained for pMeCP2 (S421). Slices were then imaged in area CA3. Scale bar =

20 μm and dotted line demarcates lesion site. (B) Quantification of mean fluorescence intensity of slices: 1 HPL (n = 8 slices; ***p < 0.001, one-way ANOVA with post-hoc Dunnett’s test, comparisons made to control), 1 HPL treated with STO-609 (n = 8 slices; *p < 0.05, one-way

ANOVA with post-hoc Dunnett’s test), with nimodepine (n = 8 slices), with CPP (n = 8 slices) or with KN-62 (n = 8 slices) and control (n = 10 slices; all values were normalized to control).

177 that was caused by an increase in recurrent collaterals and induced an increase in the intrinsic hyperexcitability of CA3 neurons and increase in network activity (Gill et al., 2013). We recorded spontaneous action potential firing from CA3 neurons (<250 µm from lesion site) and we found that the input resistance and resting membrane potential of OGD or control cells did not differ significantly. Moreover, at 14DPL there was a significant increase in the frequency of action potentials (Fig. 4.6A, B; 2.88 ± 0.96 APs/min recording; n = 10 cells from 5 slices; *p < 0.05, two- way independent student’s t-test) compared to control (Fig. 4.6A, B; 0.43 ± 0.18 APs/min recording; n = 8 cells from 4 slices) and control treated with KN-62 (Fig. 4.6A, B; 0.38 ± 0.12

APs/min recording; n = 7 cells from 4 slices). Interestingly, KN-62 did not prevent the increase in firing rate in lesioned slices (Fig. 4.6A, B; 2.21 ± 0.90 APs/min recording; n = 9 cells from 5 slices; p = 0.069, two-way independent student’s t-test). Likewise, lesion caused an increase in bursting activity (Fig. 4.6A, C; 0.34 ± 0.14 bursts /min recording; p = 0.062; bursts were defined as three or more action potentials occurring at an inter-event interval <600 ms), compared to control (Fig.

4.6A, C; 0.039 ± 0.018 bursts/min recording) and control treated with KN-62 (Fig. 4.6A, C; 0.056

± 0.028 bursts/min recording). Bursting activity was also observed in slices treated with KN-62

(Fig. 4.6A, C; 0.34 ± 0.11 bursts/min recording; *p < 0.05, two-way independent student’s t-test).

Taken together, our data indicates that preventing CaMKII phosphorylation of MeCP2 at

S421 does prevent increase in intrinsic CA3 pyramidal neuron excitability and this suggests, that

MeCP2 phosphorylation may not regulate lesion-induced BDNF increase and axon sprouting.

178

Figure 4.6 – CaMKII does mediate Schaffer collateral lesion-induced increase of CA3 pyramidal neuron firing. (A) Representative traces of whole-cell current-clamp electrophysiological recordings of CA3 pyramidal neurons from control, control treated with KN-

62, 14 DPL and 14 DPL slices treated with KN-62. (B) Quantification of spontaneous action potential frequency at 14 DPL (n = 10 cells from 5 slices; *p < 0.05, † p = 0.069, two-way independent student’s t-test), control (n = 8 cells from 4 slices), control treated with KN-62 (n = 7 cells from 4 slices) and 14 DPL slices treated with KN-62 (n = 9 cells from 5 slices). (C) Analysis of burst frequency († p = 0.062, *p < 0.05, two-way independent student’s t-test; bursts were defined as 3 or more action potentials occurring at an inter-event interval <600 ms).

179

4.5. DISCUSSION

Here, we demonstrate that following Schaffer collateral transection in mouse organotypic hippocampal slices there is an increase in pMeCP2S421 in area CA3 close to the lesion site, coinciding with an increase in firing rate of CA3 pyramidal neurons. The increased action potential firing and bursting activity is likely due to an increase in new excitatory synapses formed on CA3 neurons by axon sprouting as we have previously shown (McKinney et al., 1997; Gill et al., 2013).

Pharmacological inhibition of CaMKII, CaMKK, NMDA receptors and voltage-gated Ca2+ channels all attenuated pMeCP2S421. Given that CaMKII elicited the most significant decrease in pMeCP2S421, we inhibited CaMKII for 14 DPL and found that it did not prevent lesion-induced

CA3 pyramidal neuron hyperexcitability.

4.5.1. Role of pMeCP2S421 in BDNF-induced axonal remodeling

Given that BDNF is important for axon growth during development (Cohen-Cory & Fraser,

1995), it is not surprising that BDNF can remodel of axons in response to injury (Mathern et al.,

1997; Cohen-Cory, 1999). However, due to the importance of BDNF for synaptic plasticity and neuronal survival, BDNF is an unlikely target for prevention of axonal sprouting in particular after penetrating head injury. Here, we found that pharmacological inhibition of different mediators of

2+ Ca signaling all attenuated pMeCP2S421, suggesting that inhibiting these would then inhibiting activity-dependent transcription of BDNF in order to prevent axonal sprouting and subsequent hyperexcitability of CA3 pyramidal neurons. However, we did not block lesion-induced hyperexcitability of CA3 pyramidal neurons by preventing pMeCP2S421. Therefore, our data suggests that activity-dependent transcription of Bdnf may not be essential for axonal sprouting

180 and network reorganization within area CA3. However, it is important to note that without measuring mRNA levels from treated slices or measuring GAP43-immunopositive axonal spouting it is difficult to know whether the blockade of a single Ca2+ mediator is sufficient to prevent MeCP2-dependent transcription. Hence, it would be interesting to test whether individual or combination of these inhibitors (nimodepine – L-type voltage-gated Ca2+ channels; CPP –

NMDA receptors; KN-62 – CaMKII; STO-609 – CaMKK) can prevent bdnf mRNA upregulation specifically from exon IV. Unfortunately, current methodologies of RT-qPCR, though sensitive, cannot detect bdnf mRNA arising from exon IV in organotypic hippocampal slice cultures.

Studies have shown that MeCP2-/- mice can depict distinctive changes in gene expression in particular of genes important for neuronal function (Tudor et al., 2002; Chahrour et al., 2008;

Skene et al., 2010), including transcription of bdnf (Chen et al., 2003). Despite being an important regulator of neuronal genes, MeCP2-/- mice develop normally and only begin to show neurological deficits at 6 weeks after birth and typically die after 16-20 weeks due to suppressed appetite (Guy et al., 2001; Weng et al., 2011). This fits with literature demonstrating that MeCP2 expression is very low in new born mice but increases with development reaching a peak after 5 weeks (Skene et al., 2010), coinciding with the postnatal week where neurological deficits begin. Given that organotypic cultures are made from P6-8 mice and that typically the morphology of pyramidal neurons in mature slice culture are comparable to P15 mice (Harris et al., 1992; McKinney et al.,

1999a), it may be possible that there are other mechanisms in place during development that regulate activity-dependent transcription of BDNF. For instance, bdnf at exon I is also known to be regulated by Ca2+, as its promoter region contains a Ca2+ response element (Pruunsild et al.,

2011). Moreover, mRNA transcripts produced by the Bdnf gene can also be regulated through non- coding regulatory regions, such as the 3’ untranslated region (3’ UTR) (Timmusk et al., 1993).

181

Long or short 3’ UTR mRNA can have different subcellular localization and different stability

(Timmusk et al., 1993; Greenberg et al., 2009). For instance, bdnf mRNA with longer 3’ UTR are more likely to be targeted to the dendrite (Gorski et al., 2003; An et al., 2008) and short 3’ UTR to the soma (An et al., 2008). Thus, alternating the ratio and stability of these two distinct pools may promote synaptogenesis or axonal remodeling after brain trauma or injury.

Interestingly, altered regulation of 3’ UTR in MeCP2-/- mice can also lead to reduction of bdnf mRNA (Wu et al., 2010). However, it is important to note that animals with a complete

MeCP2 knockout from birth may activate other compensatory mechanisms during development.

Additionally, a total MeCP2 knockout mouse would have complete loss of gene function and thus it would be difficult tease apart the specific activity of the S421-specific transcription activation.

In order to tease this apart, recently, the Greenberg group has developed a S421A (serine 421 to alanine) knock-in mouse, thereby decoupling the activity-dependent switch of MeCP2 (Cohen et al., 2011). They demonstrated that MeCP2 is more likely a master regulator of neuronal chromatin remodeling in response to activity, rather than a specific regulator of certain genes (Cohen et al.,

2011; Rutlin & Nelson, 2011). Thus, MeCP2 seems to regulate the expression of many genes simultaneously and may not be a viable therapeutic target. In addition, our data also suggests that

MeCP2 is not an active participant in activity-induced neuronal regulation or synaptic remodeling.

4.5.2. Potential sources of BDNF release

It has been shown that MeCP2 expression is significantly higher in neurons than in other cell types (Skene et al., 2010). Correspondingly, our data demonstrated a specific increase in pMeCP2S421 in a large proportion of CA3 neurons adjacent to the lesion and also a few GFAP- positive astrocytes. Initially, this indicated to us that BDNF was likely being released from neurons

182 in response to excess glutamate release and Ca2+ uptake. However, given that inhibiting pMeCP2S421 prevent the development of CA3 pyramidal neuron hyperexcitability, it suggests that activity-dependent release through this mechanism is not required for BDNF-induced axonal sprouting. It may be possible that this is occurring through a different activity-dependent mechanisms or that BDNF release is not occurring from neurons. BDNF is known to be released from glial cells, including astrocytes (Bergami et al., 2008; Parpura & Zorec, 2010) and microglia

(Miwa et al., 1997; Trang et al., 2011).

Interestingly, there are many pathophysiological instances where BDNF can be released in high quantities from microglia. For example, lipopolysaccharide-induced inflammation can cause microglial cultures to release BDNF (Miwa et al., 1997). After this initial report, many other studies were published suggesting that microglia release BDNF in models of ischemia (Lee et al.,

2002), multiple sclerosis (Stadelmann et al., 2002), neuropathic pain (Coull et al., 2005), and traumatic brain injury (Batchelor et al., 1999; Dougherty et al., 2000). Moreover, the notion that microglia actively participate in synaptic plasticity is growing (Wake et al., 2013). Given that Ca2+ and glutamate can cause microglia to become activated (Noda et al., 2000; Kettenmann et al.,

2011), a hippocampal lesion may cause BDNF release from microglia, thereby beginning the axonal remodeling process.

In summary, our study suggests that pMeCP2S421-mediated activity-dependent bdnf transcription through CaMKII is not necessary for the hyperexcitability of CA3 neurons induced by Schaffer collateral lesion. The evidence shown in this study implies that there may be other mechanisms of release that may play a role in BDNF-mediated effects on axon sprouting. In addition, we provide further insight on the role of MeCP2 in regulating BDNF and therefore our work has implications for not only traumatic brain injury but also epilepsy and Rett syndrome,

183 where MeCP2 is known to be mutated in 90% of cases (Amir et al., 1999). Findings from our work improves our understanding of the pathological role of BDNF in neurological conditions and will help identify novel therapeutic targets to prevent cognitive deficits after traumatic brain injuries.

184

CHAPTER 5. DISCUSSION AND CONCLUSION

5.1. SUMMARY

In this thesis, I have outlined my findings demonstrating that BDNF and its receptors can contribute to injury-induced microcircuitry reorganization within the hippocampus. We have demonstrated this in two ways: (1) BDNF-TrkB signaling can downregulate GABAergic synapses and proBDNF-p75NTR signaling can downregulate glutamatergic synapses, both structurally and functionally after ischemia and (2) BDNF-TrkB induces axonal remodeling and functional network reorganization following a lesion mimicking a traumatic brain injury. Moreover, we have shown that activity-dependent transcription of bdnf mediated by MeCP2 is not required for CA3 neuron hyperexcitability caused by Schaffer collateral lesion.

5.2. DISCUSSION AND FUTURE DIRECTIONS

Our data supports the theory that BDNF upregulation occurs after injury in an attempt by the brain to ameliorate injury, but may actually revert the central nervous system to a more juvenile state further exacerbating injury. Thus, BDNF signaling is an important therapeutic target for intervention in many different neurological conditions, including epilepsy and ischemia. Our data also has broader implications on microcircuitry remodeling as a whole in other disorders such as bipolar disorder or depression, where BDNF is also known to be misregulated (Neves-Pereira et al., 2002; Green & Craddock, 2003; Tsai, 2004b; Rybakowski, 2008).

185

5.2.1. Pathological Role of BDNF in Microcircuitry Remodeling

In this thesis, I demonstrate that after hippocampal injury bdnf mRNA expression can increase shortly after injury onset, 1.5-2 hours after ischemia or Schaffer collateral lesion. These results mirror what has been seen in ischemia, epilepsy, traumatic brain injury or spinal cord injury where BDNF expression increases as early as 2 hours post-injury, including in: rodents (Ernfors et al., 1991; Lindvall et al., 1992; Takeda et al., 1993; Hicks et al., 1997; Santhanam et al., 2010;

Vanelderen et al., 2010; Bejot et al., 2011a; Bejot et al., 2011b; Gottlieb et al., 2013), in primates

(Tonchev et al., 2008; Sato et al., 2009) and in human patients (Chiaretti et al., 2003; Rodrigues et al., 2008). However, there has been considerable debate as to whether this increase in BDNF is protective or detrimental (Lindvall et al., 1994; Xie & Longo, 2000) for recovery, suggesting that

BDNF signaling is complex. This differential signaling can be explained by a number of factors:

(1) concentration of BDNF release, (2) availability of cleavage proteins for transformation of proBDNF to mature BDNF, (3) timeframe of release, (4) source of BDNF secretion (neuronal vs glial) and (5) availability and expression of TrkB and p75NTR receptors. Therefore, therapies which directly modulate BDNF must be specific in targeting one of these five factors, rather than modulating BDNF or TrkB receptors as a whole. In this thesis, I examined the cellular mechanism and signaling of BDNF in order to identify more specific targets for treatment of brain injuries.

Firstly, I found that following ischemia to organotypic hippocampal slices bidirectional signaling at GABAergic and glutamatergic synapses was activated by proBDNF-p75NTR and

BDNF-TrkB, respectively. The data I outlined in this thesis describes this bifurcated signaling on pyramidal neurons and as of yet we are still unaware how this acute increase in proBDNF and mature BDNF can affect GABAergic or glutamatergic postsynapses on interneurons. Interestingly, downstream signaling of mature BDNF and TrkB, downregulation of GABAergic synapses on

186

CA1 pyramidal neurons was regulated by ERK1/2 and GSK3β, two molecules that are important for synaptic plasticity. To the best of our knowledge, this is the first report demonstrating simultaneous bifurcated signaling of proBDNF and mature BDNF, at both glutamatergic and

GABAergic synapses, in a pathological context. Intriguingly, this can occur in axon guidance during development of the visual system (Marler et al., 2010; Xu & Henkemeyer, 2012). This supports the theory that brain injury may indeed reset the adult central nervous system back to a juvenile or developmental state in an erroneous attempt to rescue synaptic deficits after injury.

Interestingly, intravenous delivery of BDNF (Schabitz et al., 2007) or overexpression of

BDNF with viral vectors (Yu et al., 2013) has been shown to aid in the recovery of motor function in rodents after stroke. These studies suggest that elevated BDNF is beneficial in recovery after stroke. However, other studies using similar stroke models, with either heterozygous BDNF+/- transgenic mice (Nygren et al., 2006) or mice expressing the Val66Met polymorphism of the

BDNF gene (Qin et al., 2014), which have reduced levels of BDNF release (Egan et al., 2003), demonstrate that less BDNF has a positive outcome on motor recovery. Likewise, patients with the Val66Met polymorphism also demonstrate enhanced cognitive recovery after traumatic brain injuries (Barbey et al., 2014) and stroke (Cramer et al., 2012; Di Lazzaro et al., 2015). This indicates that having life-long decreased levels of BDNF in patients is beneficial for recovery, though we cannot exclude compensation from other neurotrophins. It’s also important to note that a lot of literature demonstrates that enhancement of BDNF-TrkB signaling can increase post-stroke recovery (Galvin & Oorschot, 2003; Schabitz et al., 2007; Yu et al., 2013), in particular when enhancement of BDNF is delayed (Han et al., 2012). It may be possible that immediately after injury, in response to excess glutamate and Ca2+, there is a large surge in proBDNF release which may override extracellular cleavage thus increase the concentration of proBDNF available at

187 synapses. This would then be detrimental to excitatory synapses, especially since p75NTR expression increases in the penumbra shortly following ischemia (Angelo et al., 2009). This would also increase the amount of mature BDNF available at GABAergic synapses, effectively leading to their disruption. Over time this expression might recover to baseline and then BDNF-TrkB signaling could be neuroprotective by augmenting neurogenesis and neuronal survival.

Interestingly, patients with the Val66Met polymorphism have reduced proBDNF as well as mature

BDNF levels (Egan et al., 2003), therefore they are less likely to incur maladaptive plasticity after injury, corresponding to post-injury functional outcome in this cohort of patients (Cramer et al.,

2012; Barbey et al., 2014; Di Lazzaro et al., 2015).

I have also observed maladaptive plasticity after Schaffer collateral lesion, where I ascertained that BDNF binding to TrkB receptors was necessary for axonal sprouting and hyperexcitability of neuronal networks in area CA3. Interestingly, it has been demonstrated in in vitro (Dinocourt et al., 2006) and in vivo (Aungst et al., 2013) that TrkB can play an important role in axonal sprouting after injury. TrkB receptors, are known to be upregulated after Schaffer collateral lesion in organotypic hippocampal slices and conditional knockdown of TrkB receptors can downregulate this sprouting (Dinocourt et al., 2006). Moreover, in vivo, TrkB receptor activation is critical for axonal sprouting and subsequent network hyperexcitability (Aungst et al.,

2013). I demonstrate that pharmacological inhibition of BDNF can prevent this hyperexcitability.

It is also important to note that in different in vivo model of post-traumatic epilepsy, cortical undercutting of rodents, BDNF may be important for the survival of GABAergic interneurons in the cortex (Prince et al., 2009; Prince, 2012). On the other hand, recent evidence also suggests that

Schaffer collateral lesion can downregulate the K+/Cl- cotransporter KCC2 (Shulga et al., 2008), making GABA excitatory. Given that increased BDNF can decrease the expression of KCC2

188

(Rivera et al., 2005), it seems likely that this mechanism may shift the balance in favour of excitation, leading to epileptogenesis. Though, we have previously shown in rat organotypic cultures that Schaffer collateral lesion did not induce any changes to GABAergic transmission

(McKinney et al., 1997). Therefore, at this time it is unclear whether the net effect of BDNF-TrkB signaling is anti- or pro-epileptogenic. Though one thing is clear, if pharmacological inhibition of

BDNF-TrkB signaling was used to prevent the development of epilepsy in patients following traumatic brain injury, it would be important to target a specific pool of BDNF, given that BDNF-

TrkB signaling is important for many functions (Huang & Reichardt, 2001).

Thus, I next wanted to determine if activity-dependent release of BDNF via MeCP2 could be a viable target to prevent Schaffer collateral lesion-induced axonal sprouting and subsequent hyperexcitability. Phosphorylation of MeCP2 at S421 (pMeCP2S421) causes MeCP2 to dissociate from the promoter region of the Bdnf gene and allows transcription of bdnf mRNA (Zhou et al.,

2+ 2006). Furthermore, pMeCP2S421 is regulated by activity-induced Ca influx, from many sources including NMDA receptors or L-type voltage gated Ca2+ channels (Chen et al., 2003; Tao et al.,

2009). Ca2+ entry can activate CaMKK and then CaMKII and CaMKIV (Zhou et al., 2006; Tao et al., 2009), which directly phosphorylate MeCP2 in the nucleus (Buchthal et al., 2012), changing its affinity for the promoter region, dissociating itself from the bdnf exon IV promoter and allowing its transcription. When we prevented pMeCP2S421 by inhibiting CaMKII, we did not prevent CA3 pyramidal neuron hyperexcitability. It may be possible that this mechanism is not necessary for

BDNF-mediated axonal sprouting or that we did not use sufficient blockade to prevent bdnf exon

IV mRNA upregulation. However, it is difficult to measure bdnf mRNA expression from a specific exon via RT-qPCR due to the very low expression levels in organotypic hippocampal slices.

189

In addition, it has been shown that another phosphorylation site also regulates MeCP2 association with DNA, serine 80 (S80). Phosphorylation of MeCP2 at S80 (pMeCP2S80) is constitutive in resting state neurons, until neuronal activity leads to the dephosphorylation by an unknown phosphatase (Tao et al., 2009; Damen & Heumann, 2013). Therefore, in order to induce activity-dependent transcription of Bdnf, MeCP2 must be phosphorylated at S421 and dephosphorylated at S80 (Fig. 5.1A). Given this information, we have recently probed for pMeCP2S80 after Schaffer collateral lesion (Fig. 5.1B, C). We observed that pMeCP2S80 expression remained unchanged up to 3 hours following transection (Fig. 5.1B, C), suggesting that MeCP2 remains bound to DNA after Schaffer collateral lesion. Therefore, we can speculate that MeCP2- mediated transcription of bdnf at exon IV is not necessary for lesion-induced hyperexcitability caused by axonal sprouting in area CA3.

Taken together, we must identify more specific signaling cascades or distinct mechanisms of BDNF release in order to develop precise therapy to prevent the negative effects of BDNF, but not the positive. An alternate approach may be to target specific pools of BDNF-release, in particular the pool of BDNF that resides in non-neuronal cells may be a viable therapeutic target.

5.2.2. Source of BDNF Release

For a long time, it was thought that only neurons could release BDNF, given their importance in neuronal development and survival (Huang & Reichardt, 2001). However, with advancing knowledge on the role of glial cells in synaptic plasticity, it has become abundantly clear that astrocytes (Bergami et al., 2008; Parpura & Zorec, 2010) and microglia (Miwa et al.,

1997; Dougherty et al., 2000; Coull et al., 2005) can also release BDNF. Moreover, given the

190

Figure 5.1 – pMeCP2S80 expression remains unchanged at 1 and 3 hours following Schaffer collateral lesion. (A) Schematic of Ca2+ influx by neuronal activity leading to phosphorylation of MeCP2 at S421 and dephosphorylation at S80. (i-iii) Different activation states of MeCP2.

MeCP2 only dissociates from promoter if S421 is phosphorylated and S80 is dephosphorylated.

(B) Maximum intensity projections of organotypic hippocampal slice cultures which have been immunostained for pMeCP2 (S80), imaged in area CA3. Antibody obtained from Millipore and immunofluorescence protocol found in section 4.3.3. Scale bar = 20 μm and dotted line demarcates lesion site. (C) Quantification of mean fluorescence intensity of slices: 1 HPL (n = 5 slices), 3 HPL (n = 5 slices) and control (n = 5 slices). Modified from (Hong et al., 2005; Chao

& Zoghbi, 2009).

191 bifurcation of proBDNF/mature BDNF signaling, it may also be possible that the sources of proBDNF and mature BDNF are different.

proBDNF released by neurons in response to theta burst stimulation, a protocol typically used to induce cellular models of learning and memory, can be up taken by perineural astrocytes by binding astrocytic p75NTR, causing the endocytosis of the ligand-receptor complex (Bergami et al., 2008). This restricts the availability of proBDNF, which can then be re-secreted at a later time point by astrocytes, potentially in response to glutamate (Bergami et al., 2008). This leads to a very interesting idea that proBDNF-p75NTR signaling in astrocytes may affect dendritic spine remodeling, as recent evidence suggests that astrocytes play a crucial role in dendritic spine remodeling (Haber & Murai, 2006; Haber et al., 2006; Verbich et al., 2012; Perez-Alvarez et al.,

2014). It would be interesting to see if p75NTR-BDNF signaling from astrocytes underlies ischemia- induced synapse remodeling. This could be done by ablating p75NTR receptors for neurons through:

(1) transfection of plasmid DNA encoding siRNA against p75NTR using biolistic methods in organotypic slice cultures or (2) developing a transgenic mouse in which p75NTR is specifically ablated in astrocytes, potentially through Cre-lox recombination with a GFAP-driven promoter.

Developing a transgenic mouse requires considerable effort and therefore, initially a biolistic approach would be ideal to determine if this mechanism is important for BDNF-mediated pathological synapse remodeling.

Moreover, BDNF can also be released in high quantities from microglia in a pathophysiological context. This was first observed in dissociated microglial cultures stimulated with lipopolysaccharide (Miwa et al., 1997), an endotoxin found on the outer membrane of Gram- negative bacteria that can elicit a strong immune response (Sawada et al., 1989; Suzumura et al.,

1991). We now know that microglia release BDNF in models of ischemia (Lee et al., 2002),

192 multiple sclerosis (Stadelmann et al., 2002), neuropathic pain (Coull et al., 2005), and traumatic brain injury (Batchelor et al., 1999; Dougherty et al., 2000). Given that the primary function of microglia is to sense and react to alterations of the extracellular milieu in response to injury or infection (Kettenmann et al., 2011), the fact that microglia release BDNF is an important point to solidify the role of BDNF in neurological disorders. Microglial release of BDNF is best characterized in the spinal cord of neuropathic pain models (Trang et al., 2011), which has been demonstrated through a series of elegant studies from the De Koninck and Salter laboratories. They have shown that altered Cl- homeostasis in the spinal dorsal horn is a key mechanism underlying neuropathic pain (Coull et al., 2003) and this alteration is caused by microglial release of BDNF

(Coull et al., 2005). In turn, BDNF binds to TrkB receptors inducing the downregulation of KCC2

(Coull et al., 2005), which maintains the neuronal Cl- gradient. Interestingly, this is dependent on the purine receptor P2X4, which binds to ATP (Ulmann et al., 2008). In a subsequent study, the same groups used a transgenic mouse in which BDNF expression was deleted specifically from microglia (Ferrini et al., 2013), in order to demonstrate the microglial specific effect on neuropathic pain development through the P2X4-BDNF-TrkB-KCC2 axis.

For many years it was thought that microglia were only important for immune response in the brain, but now we know that microglia can actively participate in synaptic plasticity

(Kettenmann et al., 2011; Wake et al., 2013). Many brain traumas can cause excess Ca2+ and glutamate release, including traumatic brain injury, epilepsy and stroke. Furthermore, Ca2+ and glutamate can cause microglia to become activated (Noda et al., 2000; Kettenmann et al., 2011).

Therefore, it may be possible that injury-induced increases in Ca2+ and glutamate would activate microglia, releasing BDNF and thereby leading to network reorganization in response to injury.

Additionally, there is an initial surge of ATP in ischemia, as discussed in section 1.4.2. on page

193

28, which could also induce the release of BDNF from microglia, especially since P2X4 receptors are found not only in the spinal cord but also in highly plasticity regions in the brain, such as the hippocampus and cortex (Bo et al., 2003). This represents an interesting new therapeutic avenue, however, more research must be done to identify if this cascade is being activated following brain trauma. In this thesis, we did not test for microglial activation or for BDNF release from microglia.

The following two experiments would be necessary to test for this after OGD or Schaffer collateral lesion of organotypic hippocampal slices: (1) determine if microglia are activated after injury by immunostaining for Iba1, a Ca2+-binding protein found only in activated microglia (Ito et al., 1998;

Sasaki et al., 2001) and (2) making organotypic slices from a transgenic mouse in which BDNF expression is specifically deleted from microglia, as in the CD11b-cre BdnfloxP/loxP mouse (Ferrini et al., 2013) and then subjecting these slices to our injury models.

Microglial or astrocytic release of BDNF may be an ideal approach to target a specific pool of BDNF in order to prevent pathological levels of BDNF release, which can lead to synapse reorganization, rather than targeting all BDNF as a whole.

5.3. CONCLUSION

In conclusion, the models I have outlined in this thesis attempt to mimic events that occur in brain trauma in order to dissect the cellular and molecular mechanisms that underlie synaptic reorganization. My data enhances our understanding of how BDNF-mediated synaptic plasticity can be misappropriated after hippocampal injury. Consequently, findings from my work enhances our understanding of injury-induced maladaptive processes and may lead to the development of specific therapeutic targets that enhance cognitive recovery. My data provides a mechanistic basis

194 for further study of BDNF signaling after acquired brain injuries in rodents and higher mammals in vivo.

Maladaptive plasticity and synaptic reorganization has been described in patients of multiple different cognitive disorders, including but not limited to: traumatic brain injury, post- traumatic epilepsy and ischemic stroke. Beyond this, my studies on maladaptive plasticity has implications for other synaptic disorders such as Alzheimer’s disease, depression, schizophrenia and bipolar disease, where BDNF is known to be misregulated. Finally, a better understanding of pathological plasticity is critical to produce novel therapies for patients of neurological disorders in order to improve their quality of life.

195

REFERENCES

Adachi, N., Kohara, K. & Tsumoto, T. (2005) Difference in trafficking of brain-derived neurotrophic factor between axons and dendrites of cortical neurons, revealed by live-cell imaging. BMC neuroscience, 6, 42. Aguado, F., Carmona, M.A., Pozas, E., Aguilo, A., Martinez-Guijarro, F.J., Alcantara, S., Borrell, V., Yuste, R., Ibanez, C.F. & Soriano, E. (2003) BDNF regulates spontaneous correlated activity at early developmental stages by increasing synaptogenesis and expression of the K+/Cl- co-transporter KCC2. Development, 130, 1267-1280. Aid, T., Kazantseva, A., Piirsoo, M., Palm, K. & Timmusk, T. (2007) Mouse and rat BDNF gene structure and expression revisited. Journal of neuroscience research, 85, 525-535. Akaneya, Y., Tsumoto, T., Kinoshita, S. & Hatanaka, H. (1997) Brain-derived neurotrophic factor enhances long-term potentiation in rat visual cortex. The Journal of neuroscience : the official journal of the Society for Neuroscience, 17, 6707-6716. Albers, G.W., Goldstein, L.B., Hall, D., Lesko, L.M. & Aptiganel Acute Stroke, I. (2001) Aptiganel hydrochloride in acute ischemic stroke: a randomized controlled trial. Jama, 286, 2673-2682. Alderson, R.F., Curtis, R., Alterman, A.L., Lindsay, R.M. & DiStefano, P.S. (2000) Truncated TrkB mediates the endocytosis and release of BDNF and neurotrophin-4/5 by rat astrocytes and schwann cells in vitro. Brain research, 871, 210-222. Alicke, B. & Schwartz-Bloom, R.D. (1995) Rapid down-regulation of GABAA receptors in the gerbil hippocampus following transient cerebral ischemia. Journal of neurochemistry, 65, 2808-2811. Alldred, M.J., Mulder-Rosi, J., Lingenfelter, S.E., Chen, G. & Luscher, B. (2005) Distinct gamma2 subunit domains mediate clustering and synaptic function of postsynaptic GABAA receptors and gephyrin. The Journal of neuroscience : the official journal of the Society for Neuroscience, 25, 594-603. Alonso, M., Medina, J.H. & Pozzo-Miller, L. (2004) ERK1/2 activation is necessary for BDNF to increase dendritic spine density in hippocampal CA1 pyramidal neurons. Learning & memory, 11, 172-178. Altman, J. & Bayer, S.A. (1990) Prolonged sojourn of developing pyramidal cells in the intermediate zone of the hippocampus and their settling in the stratum pyramidale. The Journal of comparative neurology, 301, 343-364. Amaral, D.G. & Lavanex, P. (2007) Hippocampal Neuroanatomy. In Andersen, P., Morris, R., Amaral, D.G., Bliss, T., O'Keefe, J. (eds) The Hippocampus Book. Oxford Press, New York City, NY, pp. 37-114. Amaral, D.G. & Witter, M.P. (1989) The three-dimensional organization of the hippocampal formation: a review of anatomical data. Neuroscience, 31, 571-591.

196

Amaral, M.D. & Pozzo-Miller, L. (2007a) BDNF induces calcium elevations associated with IBDNF, a nonselective cationic current mediated by TRPC channels. Journal of neurophysiology, 98, 2476-2482. Amaral, M.D. & Pozzo-Miller, L. (2007b) TRPC3 channels are necessary for brain-derived neurotrophic factor to activate a nonselective cationic current and to induce dendritic spine formation. The Journal of neuroscience : the official journal of the Society for Neuroscience, 27, 5179-5189. Amini, M., Ma, C.L., Farazifard, R., Zhu, G., Zhang, Y., Vanderluit, J., Zoltewicz, J.S., Hage, F., Savitt, J.M., Lagace, D.C., Slack, R.S., Beique, J.C., Baudry, M., Greer, P.A., Bergeron, R. & Park, D.S. (2013) Conditional disruption of calpain in the CNS alters dendrite morphology, impairs LTP, and promotes neuronal survival following injury. The Journal of neuroscience : the official journal of the Society for Neuroscience, 33, 5773-5784. Amir, R.E., Van den Veyver, I.B., Wan, M., Tran, C.Q., Francke, U. & Zoghbi, H.Y. (1999) Rett syndrome is caused by mutations in X-linked MECP2, encoding methyl-CpG-binding protein 2. Nature genetics, 23, 185-188. Ampuero, E., Dagnino-Subiabre, A., Sandoval, R., Zepeda-Carreno, R., Sandoval, S., Viedma, A., Aboitiz, F., Orrego, F. & Wyneken, U. (2007) Status epilepticus induces region-specific changes in dendritic spines, dendritic length and TrkB protein content of rat brain cortex. Brain research, 1150, 225-238. An, J.J., Gharami, K., Liao, G.Y., Woo, N.H., Lau, A.G., Vanevski, F., Torre, E.R., Jones, K.R., Feng, Y., Lu, B. & Xu, B. (2008) Distinct role of long 3' UTR BDNF mRNA in spine morphology and synaptic plasticity in hippocampal neurons. Cell, 134, 175-187. Andersen, P., Blackstad, T.W. & Lomo, T. (1966a) Location and identification of excitatory synapses on hippocampal pyramidal cells. Experimental brain research, 1, 236-248. Andersen, P., Eccles, J.C. & Loyning, Y. (1964a) Location of Postsynaptic Inhibitory Synapses on Hippocampal Pyramids. Journal of neurophysiology, 27, 592-607. Andersen, P., Eccles, J.C. & Loyning, Y. (1964b) Pathway of Postsynaptic Inhibition in the Hippocampus. Journal of neurophysiology, 27, 608-619. Andersen, P., Holmqvist, B. & Voorhoeve, P.E. (1966b) Excitatory synapses on hippocampal apical dendrites activated by entorhinal stimulation. Acta physiologica Scandinavica, 66, 461-472. Andersen, P., Morris, R., Amaral, D., Bliss, T. O’Keefe, J. (2007) The Hippocampus Book. Oxford University Press, New York City, New York. Angelo, M.F., Aviles-Reyes, R.X., Villarreal, A., Barker, P., Reines, A.G. & Ramos, A.J. (2009) p75 NTR expression is induced in isolated neurons of the penumbra after ischemia by cortical devascularization. Journal of neuroscience research, 87, 1892-1903. Ankri, L., Yarom, Y. & Uusisaari, M.Y. (2014) Slice it hot: acute adult brain slicing in physiological temperature. Journal of visualized experiments : JoVE, e52068. Annegers, J.F., Hauser, W.A., Coan, S.P. & Rocca, W.A. (1998) A population-based study of seizures after traumatic brain injuries. The New England journal of medicine, 338, 20-24.

197

Araya, R., Jiang, J., Eisenthal, K.B. & Yuste, R. (2006) The spine neck filters membrane potentials. Proceedings of the National Academy of Sciences of the United States of America, 103, 17961-17966. Armanini, M.P., McMahon, S.B., Sutherland, J., Shelton, D.L. & Phillips, H.S. (1995) Truncated and catalytic isoforms of trkB are co-expressed in neurons of rat and mouse CNS. The European journal of neuroscience, 7, 1403-1409. Arundine, M. & Tymianski, M. (2004) Molecular mechanisms of glutamate-dependent neurodegeneration in ischemia and traumatic brain injury. Cellular and molecular life sciences : CMLS, 61, 657-668. Asrican, B., Lisman, J. & Otmakhov, N. (2007) Synaptic strength of individual spines correlates with bound Ca2+-calmodulin-dependent kinase II. The Journal of neuroscience : the official journal of the Society for Neuroscience, 27, 14007-14011. Aungst, S., England, P.M. & Thompson, S.M. (2013) Critical role of trkB receptors in reactive axonal sprouting and hyperexcitability after axonal injury. Journal of neurophysiology, 109, 813-824. Baldelli, P., Hernandez-Guijo, J.M., Carabelli, V. & Carbone, E. (2005) Brain-derived neurotrophic factor enhances GABA release probability and nonuniform distribution of N- and P/Q-type channels on release sites of hippocampal inhibitory synapses. The Journal of neuroscience : the official journal of the Society for Neuroscience, 25, 3358-3368. Baldelli, P., Novara, M., Carabelli, V., Hernandez-Guijo, J.M. & Carbone, E. (2002) BDNF up- regulates evoked GABAergic transmission in developing hippocampus by potentiating presynaptic N- and P/Q-type Ca2+ channels signalling. The European journal of neuroscience, 16, 2297-2310. Balkowiec, A. & Katz, D.M. (2002) Cellular mechanisms regulating activity-dependent release of native brain-derived neurotrophic factor from hippocampal neurons. The Journal of neuroscience : the official journal of the Society for Neuroscience, 22, 10399-10407. Bamji, S.X., Majdan, M., Pozniak, C.D., Belliveau, D.J., Aloyz, R., Kohn, J., Causing, C.G. & Miller, F.D. (1998) The p75 neurotrophin receptor mediates neuronal apoptosis and is essential for naturally occurring sympathetic neuron death. The Journal of cell biology, 140, 911-923. Barbacid, M. (1994) The Trk family of neurotrophin receptors. Journal of neurobiology, 25, 1386- 1403. Barbey, A.K., Colom, R., Paul, E., Forbes, C., Krueger, F., Goldman, D. & Grafman, J. (2014) Preservation of general intelligence following traumatic brain injury: contributions of the Met66 brain-derived neurotrophic factor. PloS one, 9, e88733. Barde, Y.A., Edgar, D. & Thoenen, H. (1982) Purification of a new neurotrophic factor from mammalian brain. The EMBO journal, 1, 549-553. Barker, P.A. (2009) Whither proBDNF? Nature neuroscience, 12, 105-106. Barker, P.A. & Shooter, E.M. (1994) Disruption of NGF binding to the low affinity neurotrophin receptor p75LNTR reduces NGF binding to TrkA on PC12 cells. Neuron, 13, 203-215.

198

Bassell, G.J., Zhang, H., Byrd, A.L., Femino, A.M., Singer, R.H., Taneja, K.L., Lifshitz, L.M., Herman, I.M. & Kosik, K.S. (1998) Sorting of beta-actin mRNA and protein to neurites and growth cones in culture. The Journal of neuroscience : the official journal of the Society for Neuroscience, 18, 251-265. Batchelor, P.E., Liberatore, G.T., Wong, J.Y., Porritt, M.J., Frerichs, F., Donnan, G.A. & Howells, D.W. (1999) Activated macrophages and microglia induce dopaminergic sprouting in the injured striatum and express brain-derived neurotrophic factor and glial cell line-derived neurotrophic factor. The Journal of neuroscience : the official journal of the Society for Neuroscience, 19, 1708-1716. Beck, H., Goussakov, I.V., Lie, A., Helmstaedter, C. & Elger, C.E. (2000) Synaptic plasticity in the human dentate gyrus. The Journal of neuroscience : the official journal of the Society for Neuroscience, 20, 7080-7086. Beck, K.D., Lamballe, F., Klein, R., Barbacid, M., Schauwecker, P.E., McNeill, T.H., Finch, C.E., Hefti, F. & Day, J.R. (1993) Induction of noncatalytic TrkB neurotrophin receptors during axonal sprouting in the adult hippocampus. The Journal of neuroscience : the official journal of the Society for Neuroscience, 13, 4001-4014. Bejot, Y., Mossiat, C., Giroud, M., Prigent-Tessier, A. & Marie, C. (2011a) Circulating and brain BDNF levels in stroke rats. Relevance to clinical studies. PloS one, 6, e29405. Bejot, Y., Prigent-Tessier, A., Cachia, C., Giroud, M., Mossiat, C., Bertrand, N., Garnier, P. & Marie, C. (2011b) Time-dependent contribution of non neuronal cells to BDNF production after ischemic stroke in rats. Neurochemistry international, 58, 102-111. Ben-Ari, Y. (1985) Limbic seizure and brain damage produced by kainic acid: mechanisms and relevance to human temporal lobe epilepsy. Neuroscience, 14, 375-403. Ben-Ari, Y., Cherubini, E., Corradetti, R. & Gaiarsa, J.L. (1989) Giant synaptic potentials in immature rat CA3 hippocampal neurones. The Journal of physiology, 416, 303-325. Ben-Ari, Y., Lagowska, J., Tremblay, E. & Le Gal La Salle, G. (1979) A new model of focal status epilepticus: intra-amygdaloid application of kainic acid elicits repetitive secondarily generalized convulsive seizures. Brain research, 163, 176-179. Benson, D.L., Watkins, F.H., Steward, O. & Banker, G. (1994) Characterization of GABAergic neurons in hippocampal cell cultures. Journal of neurocytology, 23, 279-295. Benveniste, H., Drejer, J., Schousboe, A. & Diemer, N.H. (1984) Elevation of the extracellular concentrations of glutamate and aspartate in rat hippocampus during transient cerebral ischemia monitored by intracerebral microdialysis. Journal of neurochemistry, 43, 1369- 1374. Berg, A.T., Berkovic, S.F., Brodie, M.J., Buchhalter, J., Cross, J.H., van Emde Boas, W., Engel, J., French, J., Glauser, T.A., Mathern, G.W., Moshe, S.L., Nordli, D., Plouin, P. & Scheffer, I.E. (2010) Revised terminology and concepts for organization of seizures and epilepsies: report of the ILAE Commission on Classification and Terminology, 2005-2009. Epilepsia, 51, 676-685.

199

Bergami, M., Santi, S., Formaggio, E., Cagnoli, C., Verderio, C., Blum, R., Berninger, B., Matteoli, M. & Canossa, M. (2008) Uptake and recycling of pro-BDNF for transmitter- induced secretion by cortical astrocytes. The Journal of cell biology, 183, 213-221. Bergeron, M., Yu, A.Y., Solway, K.E., Semenza, G.L. & Sharp, F.R. (1999) Induction of hypoxia- inducible factor-1 (HIF-1) and its target genes following focal ischaemia in rat brain. The European journal of neuroscience, 11, 4159-4170. Berkemeier, L.R., Winslow, J.W., Kaplan, D.R., Nikolics, K., Goeddel, D.V. & Rosenthal, A. (1991) Neurotrophin-5: a novel neurotrophic factor that activates trk and trkB. Neuron, 7, 857-866. Berkovic, S.F., Mulley, J.C., Scheffer, I.E. & Petrou, S. (2006) Human epilepsies: interaction of genetic and acquired factors. Trends in neurosciences, 29, 391-397. Berninger, B., Garcia, D.E., Inagaki, N., Hahnel, C. & Lindholm, D. (1993) BDNF and NT-3 induce intracellular Ca2+ elevation in hippocampal neurones. Neuroreport, 4, 1303-1306. Berninger, B., Schinder, A.F. & Poo, M.M. (1999) Synaptic reliability correlates with reduced susceptibility to synaptic potentiation by brain-derived neurotrophic factor. Learning & memory, 6, 232-242. Biffo, S., Offenhauser, N., Carter, B.D. & Barde, Y.A. (1995) Selective binding and internalisation by truncated receptors restrict the availability of BDNF during development. Development, 121, 2461-2470. Binder, D.K., Routbort, M.J., Ryan, T.E., Yancopoulos, G.D. & McNamara, J.O. (1999) Selective inhibition of kindling development by intraventricular administration of TrkB receptor body. The Journal of neuroscience : the official journal of the Society for Neuroscience, 19, 1424-1436. Binder, D.K. & Scharfman, H.E. (2004) Brain-derived neurotrophic factor. Growth factors, 22, 123-131. Blaesse, P., Airaksinen, M.S., Rivera, C. & Kaila, K. (2009) Cation-chloride cotransporters and neuronal function. Neuron, 61, 820-838. Blanco-Suarez, E., Fiuza, M., Liu, X., Chakkarapani, E. & Hanley, J.G. (2014) Differential Tiam1/Rac1 activation in hippocampal and cortical neurons mediates differential spine shrinkage in response to oxygen/glucose deprivation. Journal of cerebral blood flow and metabolism : official journal of the International Society of Cerebral Blood Flow and Metabolism, 34, 1898-1906. Blanco-Suarez, E. & Hanley, J.G. (2014) Distinct subunit-specific alpha-amino-3-hydroxy-5- methyl-4-isoxazolepropionic acid (AMPA) receptor trafficking mechanisms in cultured cortical and hippocampal neurons in response to oxygen and glucose deprivation. The Journal of biological chemistry, 289, 4644-4651. Bliss, T.V. & Lomo, T. (1973) Long-lasting potentiation of synaptic transmission in the dentate area of the anaesthetized rabbit following stimulation of the perforant path. The Journal of physiology, 232, 331-356.

200

Bo, X., Kim, M., Nori, S.L., Schoepfer, R., Burnstock, G. & North, R.A. (2003) Tissue distribution of P2X4 receptors studied with an ectodomain antibody. Cell and tissue research, 313, 159-165. Bolshakov, V.Y. & Siegelbaum, S.A. (1994) Postsynaptic induction and presynaptic expression of hippocampal long-term depression. Science, 264, 1148-1152. Bolton, M.M., Pittman, A.J. & Lo, D.C. (2000) Brain-derived neurotrophic factor differentially regulates excitatory and inhibitory synaptic transmission in hippocampal cultures. The Journal of neuroscience : the official journal of the Society for Neuroscience, 20, 3221- 3232. Bonde, C., Noraberg, J., Noer, H. & Zimmer, J. (2005) Ionotropic glutamate receptors and glutamate transporters are involved in necrotic neuronal cell death induced by oxygen- glucose deprivation of hippocampal slice cultures. Neuroscience, 136, 779-794. Bongmba, O.Y., Martinez, L.A., Elhardt, M.E., Butler, K. & Tejada-Simon, M.V. (2011) Modulation of dendritic spines and synaptic function by Rac1: a possible link to Fragile X syndrome pathology. Brain research, 1399, 79-95. Bonhoeffer, T., Staiger, V. & Aertsen, A. (1989) Synaptic plasticity in rat hippocampal slice cultures: local "Hebbian" conjunction of pre- and postsynaptic stimulation leads to distributed synaptic enhancement. Proceedings of the National Academy of Sciences of the United States of America, 86, 8113-8117. Bourne, J.N. & Harris, K.M. (2011) Coordination of size and number of excitatory and inhibitory synapses results in a balanced structural plasticity along mature hippocampal CA1 dendrites during LTP. Hippocampus, 21, 354-373. Boussi-Gross, R., Golan, H., Volkov, O., Bechor, Y., Hoofien, D., Schnaider Beeri, M., Ben- Jacob, E. & Efrati, S. (2014) Improvement of Memory Impairments in Poststroke Patients by Hyperbaric Oxygen Therapy. Neuropsychology. Bowie, D. (2012) Redefining the classification of AMPA-selective ionotropic glutamate receptors. The Journal of physiology, 590, 49-61. Bredt, D.S. & Nicoll, R.A. (2003) AMPA receptor trafficking at excitatory synapses. Neuron, 40, 361-379. Bredy, T.W., Wu, H., Crego, C., Zellhoefer, J., Sun, Y.E. & Barad, M. (2007) Histone modifications around individual BDNF gene promoters in prefrontal cortex are associated with extinction of conditioned fear. Learning & memory, 14, 268-276. Brierley, J.B. & Cooper, J.E. (1962) Cerebral complications of hypotensive anaesthesia in a healthy adult. Journal of neurology, neurosurgery, and psychiatry, 25, 24-30. Briz, V., Zhu, G., Wang, Y., Liu, Y., Avetisyan, M., Bi, X. & Baudry, M. (2015) Activity- Dependent Rapid Local RhoA Synthesis Is Required for Hippocampal Synaptic Plasticity. The Journal of neuroscience : the official journal of the Society for Neuroscience, 35, 2269- 2282. Brown, C.E., Wong, C. & Murphy, T.H. (2008) Rapid morphologic plasticity of peri-infarct dendritic spines after focal ischemic stroke. Stroke; a journal of cerebral circulation, 39, 1286-1291.

201

Brunig, I., Penschuck, S., Berninger, B., Benson, J. & Fritschy, J.M. (2001) BDNF reduces miniature inhibitory postsynaptic currents by rapid downregulation of GABA(A) receptor surface expression. The European journal of neuroscience, 13, 1320-1328. Buchthal, B., Lau, D., Weiss, U., Weislogel, J.M. & Bading, H. (2012) Nuclear calcium signaling controls methyl-CpG-binding protein 2 (MeCP2) phosphorylation on serine 421 following synaptic activity. The Journal of biological chemistry, 287, 30967-30974. Buskila, Y., Breen, P.P., Tapson, J., van Schaik, A., Barton, M. & Morley, J.W. (2014) Extending the viability of acute brain slices. Scientific reports, 4, 5309. Calabrese, B., Wilson, M.S. & Halpain, S. (2006) Development and regulation of dendritic spine synapses. Physiology, 21, 38-47. Caldeira, M.V., Melo, C.V., Pereira, D.B., Carvalho, R., Correia, S.S., Backos, D.S., Carvalho, A.L., Esteban, J.A. & Duarte, C.B. (2007a) Brain-derived neurotrophic factor regulates the expression and synaptic delivery of alpha-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid receptor subunits in hippocampal neurons. The Journal of biological chemistry, 282, 12619-12628. Caldeira, M.V., Melo, C.V., Pereira, D.B., Carvalho, R.F., Carvalho, A.L. & Duarte, C.B. (2007b) BDNF regulates the expression and traffic of NMDA receptors in cultured hippocampal neurons. Molecular and cellular neurosciences, 35, 208-219. Calderone, A., Jover, T., Mashiko, T., Noh, K.M., Tanaka, H., Bennett, M.V. & Zukin, R.S. (2004) Late calcium EDTA rescues hippocampal CA1 neurons from global ischemia-induced death. The Journal of neuroscience : the official journal of the Society for Neuroscience, 24, 9903-9913. Camilo, O. & Goldstein, L.B. (2004) Seizures and epilepsy after ischemic stroke. Stroke; a journal of cerebral circulation, 35, 1769-1775. Carim-Todd, L., Bath, K.G., Fulgenzi, G., Yanpallewar, S., Jing, D., Barrick, C.A., Becker, J., Buckley, H., Dorsey, S.G., Lee, F.S. & Tessarollo, L. (2009) Endogenous truncated TrkB.T1 receptor regulates neuronal complexity and TrkB kinase receptor function in vivo. The Journal of neuroscience : the official journal of the Society for Neuroscience, 29, 678- 685. Carota, A. & Bogousslavsky, J. (2012) Mood disorders after stroke. Frontiers of neurology and neuroscience, 30, 70-74. Carroll, R.C., Lissin, D.V., von Zastrow, M., Nicoll, R.A. & Malenka, R.C. (1999) Rapid redistribution of glutamate receptors contributes to long-term depression in hippocampal cultures. Nature neuroscience, 2, 454-460. Castillo, P.E., Chiu, C.Q. & Carroll, R.C. (2011) Long-term plasticity at inhibitory synapses. Current opinion in neurobiology, 21, 328-338. Castren, E., Pitkanen, M., Sirvio, J., Parsadanian, A., Lindholm, D., Thoenen, H. & Riekkinen, P.J. (1993) The induction of LTP increases BDNF and NGF mRNA but decreases NT-3 mRNA in the dentate gyrus. Neuroreport, 4, 895-898.

202

Castren, E., Zafra, F., Thoenen, H. & Lindholm, D. (1992) Light regulates expression of brain- derived neurotrophic factor mRNA in rat visual cortex. Proceedings of the National Academy of Sciences of the United States of America, 89, 9444-9448. Cendes, F., Sakamoto, A.C., Spreafico, R., Bingaman, W. & Becker, A.J. (2014) Epilepsies associated with hippocampal sclerosis. Acta neuropathologica, 128, 21-37. Chahrour, M., Jung, S.Y., Shaw, C., Zhou, X., Wong, S.T., Qin, J. & Zoghbi, H.Y. (2008) MeCP2, a key contributor to neurological disease, activates and represses transcription. Science, 320, 1224-1229. Chang, C.S., Liao, C.H., Lin, C.C., Lane, H.Y., Sung, F.C. & Kao, C.H. (2014) Patients with epilepsy are at an increased risk of subsequent stroke: a population-based cohort study. Seizure, 23, 377-381. Chao, H.T. & Zoghbi, H.Y. (2009) The yin and yang of MeCP2 phosphorylation. Proceedings of the National Academy of Sciences of the United States of America, 106, 4577-4578. Chao, H.T., Zoghbi, H.Y. & Rosenmund, C. (2007) MeCP2 controls excitatory synaptic strength by regulating glutamatergic synapse number. Neuron, 56, 58-65. Chao, M.V. (2003) Neurotrophins and their receptors: a convergence point for many signalling pathways. Nature reviews. Neuroscience, 4, 299-309. Chao, M.V., Bothwell, M.A., Ross, A.H., Koprowski, H., Lanahan, A.A., Buck, C.R. & Sehgal, A. (1986) Gene transfer and molecular cloning of the human NGF receptor. Science, 232, 518-521. Chen, G., Kolbeck, R., Barde, Y.A., Bonhoeffer, T. & Kossel, A. (1999) Relative contribution of endogenous neurotrophins in hippocampal long-term potentiation. The Journal of neuroscience : the official journal of the Society for Neuroscience, 19, 7983-7990. Chen, W.G., Chang, Q., Lin, Y., Meissner, A., West, A.E., Griffith, E.C., Jaenisch, R. & Greenberg, M.E. (2003) Derepression of BDNF transcription involves calcium-dependent phosphorylation of MeCP2. Science, 302, 885-889. Chen, Z.Y., Ieraci, A., Teng, H., Dall, H., Meng, C.X., Herrera, D.G., Nykjaer, A., Hempstead, B.L. & Lee, F.S. (2005) Sortilin controls intracellular sorting of brain-derived neurotrophic factor to the regulated secretory pathway. The Journal of neuroscience : the official journal of the Society for Neuroscience, 25, 6156-6166. Cheng, P.L., Song, A.H., Wong, Y.H., Wang, S., Zhang, X. & Poo, M.M. (2011) Self-amplifying autocrine actions of BDNF in axon development. Proceedings of the National Academy of Sciences of the United States of America, 108, 18430-18435. Chevaleyre, V. & Castillo, P.E. (2003) Heterosynaptic LTD of hippocampal GABAergic synapses: a novel role of endocannabinoids in regulating excitability. Neuron, 38, 461-472. Chiaretti, A., Piastra, M., Polidori, G., Di Rocco, C., Caresta, E., Antonelli, A., Amendola, T. & Aloe, L. (2003) Correlation between neurotrophic factor expression and outcome of children with severe traumatic brain injury. Intensive care medicine, 29, 1329-1338.

203

Cho, K.O., Hunt, C.A. & Kennedy, M.B. (1992) The rat brain postsynaptic density fraction contains a homolog of the Drosophila discs-large tumor suppressor protein. Neuron, 9, 929-942. Choi, Y., Goo Kang, S. & Kam, K.Y. (2015) Changes in the BDNF-immunopositive cell population of neocortical layers I and II/III after focal cerebral ischemia in rats. Brain research. Cohen-Cory, S. (1999) BDNF modulates, but does not mediate, activity-dependent branching and remodeling of optic axon arbors in vivo. The Journal of neuroscience : the official journal of the Society for Neuroscience, 19, 9996-10003. Cohen-Cory, S. & Fraser, S.E. (1995) Effects of brain-derived neurotrophic factor on optic axon branching and remodelling in vivo. Nature, 378, 192-196. Cohen, S., Gabel, H.W., Hemberg, M., Hutchinson, A.N., Sadacca, L.A., Ebert, D.H., Harmin, D.A., Greenberg, R.S., Verdine, V.K., Zhou, Z., Wetsel, W.C., West, A.E. & Greenberg, M.E. (2011) Genome-wide activity-dependent MeCP2 phosphorylation regulates nervous system development and function. Neuron, 72, 72-85. Collin, C., Vicario-Abejon, C., Rubio, M.E., Wenthold, R.J., McKay, R.D. & Segal, M. (2001) Neurotrophins act at presynaptic terminals to activate synapses among cultured hippocampal neurons. The European journal of neuroscience, 13, 1273-1282. Collingridge, G.L., Kehl, S.J. & McLennan, H. (1983) Excitatory amino acids in synaptic transmission in the Schaffer collateral-commissural pathway of the rat hippocampus. The Journal of physiology, 334, 33-46. Comelli, M.C., Guidolin, D., Seren, M.S., Zanoni, R., Canella, R., Rubini, R. & Manev, H. (1993) Time course, localization and pharmacological modulation of immediate early inducible genes, brain-derived neurotrophic factor and trkB messenger RNAs in the rat brain following photochemical stroke. Neuroscience, 55, 473-490. Congar, P., Gaiarsa, J.L., Popovici, T., Ben-Ari, Y. & Crepel, V. (2000) Permanent reduction of seizure threshold in post-ischemic CA3 pyramidal neurons. Journal of neurophysiology, 83, 2040-2046. Conner, J.M., Muir, D., Varon, S., Hagg, T. & Manthorpe, M. (1992) The localization of nerve growth factor-like immunoreactivity in the adult rat basal forebrain and hippocampal formation. The Journal of comparative neurology, 319, 454-462. Copi, A., Jungling, K. & Gottmann, K. (2005) Activity- and BDNF-induced plasticity of miniature synaptic currents in ES cell-derived neurons integrated in a neocortical network. Journal of neurophysiology, 94, 4538-4543. Corsellis, J.A. (1957) The incidence of Ammon's horn sclerosis. Brain : a journal of neurology, 80, 193-208. Coull, J.A., Beggs, S., Boudreau, D., Boivin, D., Tsuda, M., Inoue, K., Gravel, C., Salter, M.W. & De Koninck, Y. (2005) BDNF from microglia causes the shift in neuronal anion gradient underlying neuropathic pain. Nature, 438, 1017-1021.

204

Coull, J.A., Boudreau, D., Bachand, K., Prescott, S.A., Nault, F., Sik, A., De Koninck, P. & De Koninck, Y. (2003) Trans-synaptic shift in anion gradient in spinal lamina I neurons as a mechanism of neuropathic pain. Nature, 424, 938-942. Craig, M.T. & McBain, C.J. (2014) The emerging role of GABAB receptors as regulators of network dynamics: fast actions from a 'slow' receptor? Current opinion in neurobiology, 26, 15-21. Cramer, S.C., Procaccio, V., Americas, G. & Investigators, G.I.S. (2012) Correlation between genetic polymorphisms and stroke recovery: analysis of the GAIN Americas and GAIN International Studies. European journal of neurology : the official journal of the European Federation of Neurological Societies, 19, 718-724. Crepel, V., Epsztein, J. & Ben-Ari, Y. (2003) Ischemia induces short- and long-term remodeling of synaptic activity in the hippocampus. Journal of cellular and molecular medicine, 7, 401-407. Crepel, V., Represa, A., Beaudoin, M. & Ben-Ari, Y. (1989) Hippocampal damage induced by ischemia and intra-amygdaloid kainate injection: effect on N-methyl-D-aspartate, N-(1-[2- thienyl]cyclohexyl)piperidine and glycine binding sites. Neuroscience, 31, 605-612. Croll, S.D., Suri, C., Compton, D.L., Simmons, M.V., Yancopoulos, G.D., Lindsay, R.M., Wiegand, S.J., Rudge, J.S. & Scharfman, H.E. (1999) Brain-derived neurotrophic factor transgenic mice exhibit passive avoidance deficits, increased seizure severity and in vitro hyperexcitability in the hippocampus and entorhinal cortex. Neuroscience, 93, 1491-1506. Cronin, J., Obenaus, A., Houser, C.R. & Dudek, F.E. (1992) Electrophysiology of dentate granule cells after kainate-induced synaptic reorganization of the mossy fibers. Brain research, 573, 305-310. Crozier, R.A., Black, I.B. & Plummer, M.R. (1999) Blockade of NR2B-containing NMDA receptors prevents BDNF enhancement of glutamatergic transmission in hippocampal neurons. Learning & memory, 6, 257-266. Dailey, M.E. & Smith, S.J. (1996) The dynamics of dendritic structure in developing hippocampal slices. The Journal of neuroscience : the official journal of the Society for Neuroscience, 16, 2983-2994. Damen, D. & Heumann, R. (2013) MeCP2 phosphorylation in the brain: from transcription to behavior. Biological chemistry, 394, 1595-1605. Danzer, S.C., He, X. & McNamara, J.O. (2004) Ontogeny of seizure-induced increases in BDNF immunoreactivity and TrkB receptor activation in rat hippocampus. Hippocampus, 14, 345-355. Daoudal, G. & Debanne, D. (2003) Long-term plasticity of intrinsic excitability: learning rules and mechanisms. Learning & memory, 10, 456-465. Das, K.P., Chao, S.L., White, L.D., Haines, W.T., Harry, G.J., Tilson, H.A. & Barone, S., Jr. (2001) Differential patterns of nerve growth factor, brain-derived neurotrophic factor and neurotrophin-3 mRNA and protein levels in developing regions of rat brain. Neuroscience, 103, 739-761.

205

Davenport, C.J., Brown, W.J. & Babb, T.L. (1990) Sprouting of GABAergic and mossy fiber axons in dentate gyrus following intrahippocampal kainate in the rat. Experimental neurology, 109, 180-190. Davies, A.M., Thoenen, H. & Barde, Y.A. (1986) The response of chick sensory neurons to brain- derived neurotrophic factor. The Journal of neuroscience : the official journal of the Society for Neuroscience, 6, 1897-1904. Davis, S., Vanhoutte, P., Pages, C., Caboche, J. & Laroche, S. (2000a) The MAPK/ERK cascade targets both Elk-1 and cAMP response element-binding protein to control long-term potentiation-dependent gene expression in the dentate gyrus in vivo. The Journal of neuroscience : the official journal of the Society for Neuroscience, 20, 4563-4572. Davis, S.M., Lees, K.R., Albers, G.W., Diener, H.C., Markabi, S., Karlsson, G. & Norris, J. (2000b) Selfotel in acute ischemic stroke : possible neurotoxic effects of an NMDA antagonist. Stroke; a journal of cerebral circulation, 31, 347-354. de Lanerolle, N.C., Kim, J.H., Robbins, R.J. & Spencer, D.D. (1989) Hippocampal interneuron loss and plasticity in human temporal lobe epilepsy. Brain research, 495, 387-395. De Paola, V., Arber, S. & Caroni, P. (2003) AMPA receptors regulate dynamic equilibrium of presynaptic terminals in mature hippocampal networks. Nature neuroscience, 6, 491-500. De Roo, M., Klauser, P., Garcia, P.M., Poglia, L. & Muller, D. (2008) Spine dynamics and synapse remodeling during LTP and memory processes. Progress in brain research, 169, 199-207. Debanne, D., Gahwiler, B.H. & Thompson, S.M. (1994) Asynchronous pre- and postsynaptic activity induces associative long-term depression in area CA1 of the rat hippocampus in vitro. Proceedings of the National Academy of Sciences of the United States of America, 91, 1148-1152. Deinhardt, K. & Chao, M.V. (2014) Shaping neurons: Long and short range effects of mature and proBDNF signalling upon neuronal structure. Neuropharmacology, 76 Pt C, 603-609. Dejanovic, B. & Schwarz, G. (2014) Neuronal nitric oxide synthase-dependent S-nitrosylation of gephyrin regulates gephyrin clustering at GABAergic synapses. The Journal of neuroscience : the official journal of the Society for Neuroscience, 34, 7763-7768. Dejanovic, B., Semtner, M., Ebert, S., Lamkemeyer, T., Neuser, F., Luscher, B., Meier, J.C. & Schwarz, G. (2014) Palmitoylation of gephyrin controls receptor clustering and plasticity of GABAergic synapses. PLoS biology, 12, e1001908. Delgado-Escueta, A.V., Koeleman, B.P., Bailey, J.N., Medina, M.T. & Duron, R.M. (2013) The quest for juvenile myoclonic epilepsy genes. Epilepsy & behavior : E&B, 28 Suppl 1, S52- 57. Delorenzo, R.J., Sun, D.A. & Deshpande, L.S. (2005) Cellular mechanisms underlying acquired epilepsy: the calcium hypothesis of the induction and maintainance of epilepsy. Pharmacology & therapeutics, 105, 229-266. Desmond, N.L. & Levy, W.B. (1986) Changes in the postsynaptic density with long-term potentiation in the dentate gyrus. The Journal of comparative neurology, 253, 476-482.

206

Deuchars, J. & Thomson, A.M. (1996) CA1 pyramid-pyramid connections in rat hippocampus in vitro: dual intracellular recordings with biocytin filling. Neuroscience, 74, 1009-1018. Di Lazzaro, V., Pellegrino, G., Di Pino, G., Corbetto, M., Ranieri, F., Brunelli, N., Paolucci, M., Bucossi, S., Ventriglia, M.C., Brown, P. & Capone, F. (2015) Val66Met BDNF gene polymorphism influences human motor cortex plasticity in acute stroke. Brain stimulation, 8, 92-96. Diaz de Leon-Guerrero, S., Pedraza-Alva, G. & Perez-Martinez, L. (2011) In sickness and in health: the role of methyl-CpG binding protein 2 in the central nervous system. The European journal of neuroscience, 33, 1563-1574. Dibbens, L.M., Feng, H.J., Richards, M.C., Harkin, L.A., Hodgson, B.L., Scott, D., Jenkins, M., Petrou, S., Sutherland, G.R., Scheffer, I.E., Berkovic, S.F., Macdonald, R.L. & Mulley, J.C. (2004) GABRD encoding a protein for extra- or peri-synaptic GABAA receptors is a susceptibility locus for generalized epilepsies. Human molecular genetics, 13, 1315-1319. Diener, H.C., Lees, K.R., Lyden, P., Grotta, J., Davalos, A., Davis, S.M., Shuaib, A., Ashwood, T., Wasiewski, W., Alderfer, V., Hardemark, H.G., Rodichok, L., Saint, I. & Investigators, I.I. (2008) NXY-059 for the treatment of acute stroke: pooled analysis of the SAINT I and II Trials. Stroke; a journal of cerebral circulation, 39, 1751-1758. Dieni, S., Matsumoto, T., Dekkers, M., Rauskolb, S., Ionescu, M.S., Deogracias, R., Gundelfinger, E.D., Kojima, M., Nestel, S., Frotscher, M. & Barde, Y.A. (2012) BDNF and its pro- peptide are stored in presynaptic dense core vesicles in brain neurons. The Journal of cell biology, 196, 775-788. Ding, M., Haglid, K.G. & Hamberger, A. (2000) Quantitative immunochemistry on neuronal loss, reactive gliosis and BBB damage in cortex/striatum and hippocampus/amygdala after systemic kainic acid administration. Neurochemistry international, 36, 313-318. Dinocourt, C., Aungst, S., Yang, K. & Thompson, S.M. (2011) Homeostatic increase in excitability in area CA1 after Schaffer collateral transection in vivo. Epilepsia, 52, 1656- 1665. Dinocourt, C., Gallagher, S.E. & Thompson, S.M. (2006) Injury-induced axonal sprouting in the hippocampus is initiated by activation of trkB receptors. The European journal of neuroscience, 24, 1857-1866. Dobie, F.A. & Craig, A.M. (2011) Inhibitory synapse dynamics: coordinated presynaptic and postsynaptic mobility and the major contribution of recycled vesicles to new synapse formation. The Journal of neuroscience : the official journal of the Society for Neuroscience, 31, 10481-10493. Domeniconi, M., Zampieri, N., Spencer, T., Hilaire, M., Mellado, W., Chao, M.V. & Filbin, M.T. (2005) MAG induces regulated intramembrane proteolysis of the p75 neurotrophin receptor to inhibit neurite outgrowth. Neuron, 46, 849-855. Donnan, G.A., Davis, S.M., Parsons, M.W., Ma, H., Dewey, H.M. & Howells, D.W. (2011) How to make better use of thrombolytic therapy in acute ischemic stroke. Nature reviews. Neurology, 7, 400-409.

207

Dougherty, K.D., Dreyfus, C.F. & Black, I.B. (2000) Brain-derived neurotrophic factor in astrocytes, oligodendrocytes, and microglia/macrophages after spinal cord injury. Neurobiology of disease, 7, 574-585. Du, J., Feng, L., Yang, F. & Lu, B. (2000) Activity- and Ca(2+)-dependent modulation of surface expression of brain-derived neurotrophic factor receptors in hippocampal neurons. The Journal of cell biology, 150, 1423-1434. Dudek, S.M. & Bear, M.F. (1992) Homosynaptic long-term depression in area CA1 of hippocampus and effects of N-methyl-D-aspartate receptor blockade. Proceedings of the National Academy of Sciences of the United States of America, 89, 4363-4367. Dugich-Djordjevic, M.M., Tocco, G., Lapchak, P.A., Pasinetti, G.M., Najm, I., Baudry, M. & Hefti, F. (1992a) Regionally specific and rapid increases in brain-derived neurotrophic factor messenger RNA in the adult rat brain following seizures induced by systemic administration of kainic acid. Neuroscience, 47, 303-315. Dugich-Djordjevic, M.M., Tocco, G., Willoughby, D.A., Najm, I., Pasinetti, G., Thompson, R.F., Baudry, M., Lapchak, P.A. & Hefti, F. (1992b) BDNF mRNA expression in the developing rat brain following kainic acid-induced seizure activity. Neuron, 8, 1127-1138. Dwyer Hollender, K. (2014) Screening, diagnosis, and treatment of post-stroke depression. The Journal of neuroscience nursing : journal of the American Association of Neuroscience Nurses, 46, 135-141. Egan, M.F., Kojima, M., Callicott, J.H., Goldberg, T.E., Kolachana, B.S., Bertolino, A., Zaitsev, E., Gold, B., Goldman, D., Dean, M., Lu, B. & Weinberger, D.R. (2003) The BDNF val66met polymorphism affects activity-dependent secretion of BDNF and human memory and hippocampal function. Cell, 112, 257-269. Eide, F.F., Vining, E.R., Eide, B.L., Zang, K., Wang, X.Y. & Reichardt, L.F. (1996) Naturally occurring truncated trkB receptors have dominant inhibitory effects on brain-derived neurotrophic factor signaling. The Journal of neuroscience : the official journal of the Society for Neuroscience, 16, 3123-3129. Elston, G.N. (2003) Cortex, cognition and the cell: new insights into the pyramidal neuron and prefrontal function. Cerebral cortex, 13, 1124-1138. Emptage, N., Bliss, T.V. & Fine, A. (1999) Single synaptic events evoke NMDA receptor- mediated release of calcium from internal stores in hippocampal dendritic spines. Neuron, 22, 115-124. Emptage, N.J., Reid, C.A., Fine, A. & Bliss, T.V. (2003) Optical quantal analysis reveals a presynaptic component of LTP at hippocampal Schaffer-associational synapses. Neuron, 38, 797-804. Engert, F. & Bonhoeffer, T. (1999) Dendritic spine changes associated with hippocampal long- term synaptic plasticity. Nature, 399, 66-70. Enoki, R., Hu, Y.L., Hamilton, D. & Fine, A. (2009) Expression of long-term plasticity at individual synapses in hippocampus is graded, bidirectional, and mainly presynaptic: optical quantal analysis. Neuron, 62, 242-253.

208

Enright, L.E., Zhang, S. & Murphy, T.H. (2007) Fine mapping of the spatial relationship between acute ischemia and dendritic structure indicates selective vulnerability of layer V neuron dendritic tufts within single neurons in vivo. Journal of cerebral blood flow and metabolism : official journal of the International Society of Cerebral Blood Flow and Metabolism, 27, 1185-1200. Epsztein, J., Ben-Ari, Y., Represa, A. & Crepel, V. (2008) Late-onset epileptogenesis and seizure genesis: lessons from models of cerebral ischemia. The Neuroscientist : a review journal bringing neurobiology, neurology and psychiatry, 14, 78-90. Epsztein, J., Milh, M., Bihi, R.I., Jorquera, I., Ben-Ari, Y., Represa, A. & Crepel, V. (2006) Ongoing epileptiform activity in the post-ischemic hippocampus is associated with a permanent shift of the excitatory-inhibitory synaptic balance in CA3 pyramidal neurons. The Journal of neuroscience : the official journal of the Society for Neuroscience, 26, 7082- 7092. Erlander, M.G., Tillakaratne, N.J., Feldblum, S., Patel, N. & Tobin, A.J. (1991) Two genes encode distinct glutamate decarboxylases. Neuron, 7, 91-100. Ernfors, P., Bengzon, J., Kokaia, Z., Persson, H. & Lindvall, O. (1991) Increased levels of messenger RNAs for neurotrophic factors in the brain during kindling epileptogenesis. Neuron, 7, 165-176. Ernfors, P., Ibanez, C.F., Ebendal, T., Olson, L. & Persson, H. (1990a) Molecular cloning and neurotrophic activities of a protein with structural similarities to nerve growth factor: developmental and topographical expression in the brain. Proceedings of the National Academy of Sciences of the United States of America, 87, 5454-5458. Ernfors, P., Wetmore, C., Olson, L. & Persson, H. (1990b) Identification of cells in rat brain and peripheral tissues expressing mRNA for members of the nerve growth factor family. Neuron, 5, 511-526. Essrich, C., Lorez, M., Benson, J.A., Fritschy, J.M. & Luscher, B. (1998) Postsynaptic clustering of major GABAA receptor subtypes requires the gamma 2 subunit and gephyrin. Nature neuroscience, 1, 563-571. Esteban, P.F., Yoon, H.Y., Becker, J., Dorsey, S.G., Caprari, P., Palko, M.E., Coppola, V., Saragovi, H.U., Randazzo, P.A. & Tessarollo, L. (2006) A kinase-deficient TrkC receptor isoform activates Arf6-Rac1 signaling through the scaffold protein tamalin. The Journal of cell biology, 173, 291-299. Falconer, M.A. (1974) Mesial temporal (Ammon's horn) sclerosis as a common cause of epilepsy. Aetiology, treatment, and prevention. Lancet, 2, 767-770. Falconer, M.A., Serafetinides, E.A. & Corsellis, J.A. (1964) Etiology and Pathogenesis of Temporal Lobe Epilepsy. Archives of neurology, 10, 233-248. Farhadi, H.F., Mowla, S.J., Petrecca, K., Morris, S.J., Seidah, N.G. & Murphy, R.A. (2000) Neurotrophin-3 sorts to the constitutive secretory pathway of hippocampal neurons and is diverted to the regulated secretory pathway by coexpression with brain-derived neurotrophic factor. The Journal of neuroscience : the official journal of the Society for Neuroscience, 20, 4059-4068.

209

Fayard, B., Loeffler, S., Weis, J., Vogelin, E. & Kruttgen, A. (2005) The secreted brain-derived neurotrophic factor precursor pro-BDNF binds to TrkB and p75NTR but not to TrkA or TrkC. Journal of neuroscience research, 80, 18-28. Ferrini, F., Trang, T., Mattioli, T.A., Laffray, S., Del'Guidice, T., Lorenzo, L.E., Castonguay, A., Doyon, N., Zhang, W., Godin, A.G., Mohr, D., Beggs, S., Vandal, K., Beaulieu, J.M., Cahill, C.M., Salter, M.W. & De Koninck, Y. (2013) Morphine hyperalgesia gated through microglia-mediated disruption of neuronal Cl(-) homeostasis. Nature neuroscience, 16, 183-192. Fiala, J.C., Spacek, J. & Harris, K.M. (2002) Dendritic spine pathology: cause or consequence of neurological disorders? Brain research. Brain research reviews, 39, 29-54. Fifkova, E. & Van Harreveld, A. (1977) Long-lasting morphological changes in dendritic spines of dentate granular cells following stimulation of the entorhinal area. Journal of neurocytology, 6, 211-230. Figurov, A., Pozzo-Miller, L.D., Olafsson, P., Wang, T. & Lu, B. (1996) Regulation of synaptic responses to high-frequency stimulation and LTP by neurotrophins in the hippocampus. Nature, 381, 706-709. Fischer, F., Kneussel, M., Tintrup, H., Haverkamp, S., Rauen, T., Betz, H. & Wassle, H. (2000) Reduced synaptic clustering of GABA and glycine receptors in the retina of the gephyrin null mutant mouse. The Journal of comparative neurology, 427, 634-648. Flores, C.E., Nikonenko, I., Mendez, P., Fritschy, J.M., Tyagarajan, S.K. & Muller, D. (2015) Activity-dependent inhibitory synapse remodeling through gephyrin phosphorylation. Proceedings of the National Academy of Sciences of the United States of America, 112, E65-72. Frerking, M., Malenka, R.C. & Nicoll, R.A. (1998) Brain-derived neurotrophic factor (BDNF) modulates inhibitory, but not excitatory, transmission in the CA1 region of the hippocampus. Journal of neurophysiology, 80, 3383-3386. Freund, T.F. & Buzsaki, G. (1996) Interneurons of the hippocampus. Hippocampus, 6, 347-470. Frey, L.C. (2003) Epidemiology of posttraumatic epilepsy: a critical review. Epilepsia, 44 Suppl 10, 11-17. Friedman, W.J., Black, I.B. & Kaplan, D.R. (1998) Distribution of the neurotrophins brain-derived neurotrophic factor, neurotrophin-3, and neurotrophin-4/5 in the postnatal rat brain: an immunocytochemical study. Neuroscience, 84, 101-114. Friedman, W.J. & Greene, L.A. (1999) Neurotrophin signaling via Trks and p75. Experimental cell research, 253, 131-142. Frielingsdorf, H., Bath, K.G., Soliman, F., Difede, J., Casey, B.J. & Lee, F.S. (2010) Variant brain- derived neurotrophic factor Val66Met endophenotypes: implications for posttraumatic stress disorder. Annals of the New York Academy of Sciences, 1208, 150-157. Fritschy, J.M., Harvey, R.J. & Schwarz, G. (2008) Gephyrin: where do we stand, where do we go? Trends in neurosciences, 31, 257-264.

210

Fritschy, J.M. & Mohler, H. (1995) GABAA-receptor heterogeneity in the adult rat brain: differential regional and cellular distribution of seven major subunits. The Journal of comparative neurology, 359, 154-194. Frotscher, M. & Gahwiler, B.H. (1988) Synaptic organization of intracellularly stained CA3 pyramidal neurons in slice cultures of rat hippocampus. Neuroscience, 24, 541-551. Fryer, R.H., Kaplan, D.R. & Kromer, L.F. (1997) Truncated trkB receptors on nonneuronal cells inhibit BDNF-induced neurite outgrowth in vitro. Experimental neurology, 148, 616-627. Fukazawa, Y., Saitoh, Y., Ozawa, F., Ohta, Y., Mizuno, K. & Inokuchi, K. (2003) Hippocampal LTP is accompanied by enhanced F-actin content within the dendritic spine that is essential for late LTP maintenance in vivo. Neuron, 38, 447-460. Fukuda, T., Heizmann, C.W. & Kosaka, T. (1997) Quantitative analysis of GAD65 and GAD67 immunoreactivities in somata of GABAergic neurons in the mouse hippocampus proper (CA1 and CA3 regions), with special reference to parvalbumin-containing neurons. Brain research, 764, 237-243. Gahwiler, B.H. (1981) Organotypic monolayer cultures of nervous tissue. Journal of neuroscience methods, 4, 329-342. Gahwiler, B.H., Capogna, M., Debanne, D., McKinney, R.A. & Thompson, S.M. (1997) Organotypic slice cultures: a technique has come of age. Trends in neurosciences, 20, 471- 477. Gaitatzis, A., Carroll, K., Majeed, A. & J, W.S. (2004) The epidemiology of the comorbidity of epilepsy in the general population. Epilepsia, 45, 1613-1622. Gall, C., Lauterborn, J., Bundman, M., Murray, K. & Isackson, P. (1991) Seizures and the regulation of neurotrophic factor and neuropeptide gene expression in brain. Epilepsy research. Supplement, 4, 225-245. Gall, C.M., Lauterborn, J.C., Guthrie, K.M. & Stinis, C.T. (1997) Seizures and the regulation of neurotrophic factor expression: associations with structural plasticity in epilepsy. Advances in neurology, 72, 9-24. Galvin, K.A. & Oorschot, D.E. (2003) Continuous low-dose treatment with brain-derived neurotrophic factor or neurotrophin-3 protects striatal medium spiny neurons from mild neonatal hypoxia/ischemia: a stereological study. Neuroscience, 118, 1023-1032. Garcia-Lopez, P., Garcia-Marin, V. & Freire, M. (2006) Three-dimensional reconstruction and quantitative study of a pyramidal cell of a Cajal histological preparation. The Journal of neuroscience : the official journal of the Society for Neuroscience, 26, 11249-11252. Gartner, A., Polnau, D.G., Staiger, V., Sciarretta, C., Minichiello, L., Thoenen, H., Bonhoeffer, T. & Korte, M. (2006) Hippocampal long-term potentiation is supported by presynaptic and postsynaptic tyrosine receptor kinase B-mediated phospholipase Cgamma signaling. The Journal of neuroscience : the official journal of the Society for Neuroscience, 26, 3496- 3504. Gee, C.E., Benquet, P., Demont-Guignard, S., Wendling, F. & Gerber, U. (2010) Energy deprivation transiently enhances rhythmic inhibitory events in the CA3 hippocampal network in vitro. Neuroscience, 168, 605-612.

211

Gee, C.E., Benquet, P., Raineteau, O., Rietschin, L., Kirbach, S.W. & Gerber, U. (2006) NMDA receptors and the differential ischemic vulnerability of hippocampal neurons. The European journal of neuroscience, 23, 2595-2603. Gehler, S., Gallo, G., Veien, E. & Letourneau, P.C. (2004) p75 neurotrophin receptor signaling regulates growth cone filopodial dynamics through modulating RhoA activity. The Journal of neuroscience : the official journal of the Society for Neuroscience, 24, 4363-4372. Geinisman, Y., Disterhoft, J.F., Gundersen, H.J., McEchron, M.D., Persina, I.S., Power, J.M., van der Zee, E.A. & West, M.J. (2000) Remodeling of hippocampal synapses after hippocampus-dependent associative learning. The Journal of comparative neurology, 417, 49-59. Gentry, J.J., Barker, P.A. & Carter, B.D. (2004) The p75 neurotrophin receptor: multiple interactors and numerous functions. Progress in brain research, 146, 25-39. Ghosh, A., Carnahan, J. & Greenberg, M.E. (1994) Requirement for BDNF in activity-dependent survival of cortical neurons. Science, 263, 1618-1623. Gianfranceschi, L., Siciliano, R., Walls, J., Morales, B., Kirkwood, A., Huang, Z.J., Tonegawa, S. & Maffei, L. (2003) Visual cortex is rescued from the effects of dark rearing by overexpression of BDNF. Proceedings of the National Academy of Sciences of the United States of America, 100, 12486-12491. Giannone, G., Mondin, M., Grillo-Bosch, D., Tessier, B., Saint-Michel, E., Czondor, K., Sainlos, M., Choquet, D. & Thoumine, O. (2013) Neurexin-1beta binding to neuroligin-1 triggers the preferential recruitment of PSD-95 versus gephyrin through tyrosine phosphorylation of neuroligin-1. Cell reports, 3, 1996-2007. Gibson, L.M., Brazzelli, M., Thomas, B.M. & Sandercock, P.A. (2010) A systematic review of clinical trials of pharmacological interventions for acute ischaemic stroke (1955-2008) that were completed, but not published in full. Trials, 11, 43. Gill, R., Chang, P.K., Prenosil, G.A., Deane, E.C. & McKinney, R.A. (2013) Blocking brain- derived neurotrophic factor inhibits injury-induced hyperexcitability of hippocampal CA3 neurons. The European journal of neuroscience, 38, 3554-3566. Golarai, G., Greenwood, A.C., Feeney, D.M. & Connor, J.A. (2001) Physiological and structural evidence for hippocampal involvement in persistent seizure susceptibility after traumatic brain injury. The Journal of neuroscience : the official journal of the Society for Neuroscience, 21, 8523-8537. Goldberg, E.M. & Coulter, D.A. (2013) Mechanisms of epileptogenesis: a convergence on neural circuit dysfunction. Nature reviews. Neuroscience, 14, 337-349. Golgi, C., Bentivoglio, M. & Swanson, L. (2001) On the fine structure of the pes Hippocampi major (with plates XIII-XXIII). 1886. Brain research bulletin, 54, 461-483. Gonzalez-Burgos, I., Letechipia-Vallejo, G., Lopez-Loeza, E., Morali, G. & Cervantes, M. (2007) Long-term study of dendritic spines from hippocampal CA1 pyramidal cells, after neuroprotective melatonin treatment following global cerebral ischemia in rats. Neuroscience letters, 423, 162-166.

212

Goodman, L.J., Valverde, J., Lim, F., Geschwind, M.D., Federoff, H.J., Geller, A.I. & Hefti, F. (1996) Regulated release and polarized localization of brain-derived neurotrophic factor in hippocampal neurons. Molecular and cellular neurosciences, 7, 222-238. Gorski, J.A., Zeiler, S.R., Tamowski, S. & Jones, K.R. (2003) Brain-derived neurotrophic factor is required for the maintenance of cortical dendrites. The Journal of neuroscience : the official journal of the Society for Neuroscience, 23, 6856-6865. Goslin, K., Asmussen, H. & Banker, G. (1998) Rat Hippocampal Neurons in Low-Density Culture. In Banker, G., Goslin, K. (eds) Culturing Nerve Cells. MIT Press, Cambridge, MA, pp. 339-370. Gottlieb, M., Bonova, P., Danielisova, V., Nemethova, M., Burda, J. & Cizkova, D. (2013) Brain- derived neurotrophic factor blood levels in two models of transient brain ischemia in rats. General physiology and biophysics, 32, 139-142. Gottmann, K., Mittmann, T. & Lessmann, V. (2009) BDNF signaling in the formation, maturation and plasticity of glutamatergic and GABAergic synapses. Experimental brain research, 199, 203-234. Graf, E.R., Zhang, X., Jin, S.X., Linhoff, M.W. & Craig, A.M. (2004) Neurexins induce differentiation of GABA and glutamate postsynaptic specializations via neuroligins. Cell, 119, 1013-1026. Granger, A.J. & Nicoll, R.A. (2014) Expression mechanisms underlying long-term potentiation: a postsynaptic view, 10 years on. Philosophical transactions of the Royal Society of London. Series B, Biological sciences, 369, 20130136. Gray, E.G. (1959a) Axo-somatic and axo-dendritic synapses of the cerebral cortex: an electron microscope study. Journal of anatomy, 93, 420-433. Gray, E.G. (1959b) Electron microscopy of synaptic contacts on dendrite spines of the cerebral cortex. Nature, 183, 1592-1593. Green, E. & Craddock, N. (2003) Brain-derived neurotrophic factor as a potential risk locus for bipolar disorder: evidence, limitations, and implications. Current psychiatry reports, 5, 469-476. Greenberg, M.E., Xu, B., Lu, B. & Hempstead, B.L. (2009) New insights in the biology of BDNF synthesis and release: implications in CNS function. The Journal of neuroscience : the official journal of the Society for Neuroscience, 29, 12764-12767. Griesbach, G.S., Hovda, D.A., Molteni, R. & Gomez-Pinilla, F. (2002) Alterations in BDNF and synapsin I within the occipital cortex and hippocampus after mild traumatic brain injury in the developing rat: reflections of injury-induced neuroplasticity. Journal of neurotrauma, 19, 803-814. Groticke, I., Hoffmann, K. & Loscher, W. (2007) Behavioral alterations in the pilocarpine model of temporal lobe epilepsy in mice. Experimental neurology, 207, 329-349. Gruart, A., Sciarretta, C., Valenzuela-Harrington, M., Delgado-Garcia, J.M. & Minichiello, L. (2007) Mutation at the TrkB PLC{gamma}-docking site affects hippocampal LTP and associative learning in conscious mice. Learning & memory, 14, 54-62.

213

Grundy, P.L., Patel, N., Harbuz, M.S., Lightman, S.L. & Sharples, P.M. (2000) Glucocorticoids modulate BDNF mRNA expression in the rat hippocampus after traumatic brain injury. Neuroreport, 11, 3381-3384. Guo, D., Arnspiger, S., Rensing, N.R. & Wong, M. (2012) Brief seizures cause dendritic injury. Neurobiology of disease, 45, 348-355. Guy, J., Hendrich, B., Holmes, M., Martin, J.E. & Bird, A. (2001) A mouse Mecp2-null mutation causes neurological symptoms that mimic Rett syndrome. Nature genetics, 27, 322-326. Guzman-Karlsson, M.C., Meadows, J.P., Gavin, C.F., Hablitz, J.J. & Sweatt, J.D. (2014) Transcriptional and epigenetic regulation of Hebbian and non-Hebbian plasticity. Neuropharmacology, 80, 3-17. Haber, M. & Murai, K.K. (2006) Reshaping neuron-glial communication at hippocampal synapses. Neuron glia biology, 2, 59-66. Haber, M., Zhou, L. & Murai, K.K. (2006) Cooperative astrocyte and dendritic spine dynamics at hippocampal excitatory synapses. The Journal of neuroscience : the official journal of the Society for Neuroscience, 26, 8881-8891. Hackett, M.L., Kohler, S., O'Brien, J.T. & Mead, G.E. (2014) Neuropsychiatric outcomes of stroke. The Lancet. Neurology, 13, 525-534. Haditsch, U., Leone, D.P., Farinelli, M., Chrostek-Grashoff, A., Brakebusch, C., Mansuy, I.M., McConnell, S.K. & Palmer, T.D. (2009) A central role for the small GTPase Rac1 in hippocampal plasticity and spatial learning and memory. Molecular and cellular neurosciences, 41, 409-419. Hagberg, B., Aicardi, J., Dias, K. & Ramos, O. (1983) A progressive syndrome of autism, dementia, ataxia, and loss of purposeful hand use in girls: Rett's syndrome: report of 35 cases. Annals of neurology, 14, 471-479. Hagberg, H., Lehmann, A., Sandberg, M., Nystrom, B., Jacobson, I. & Hamberger, A. (1985) Ischemia-induced shift of inhibitory and excitatory amino acids from intra- to extracellular compartments. Journal of cerebral blood flow and metabolism : official journal of the International Society of Cerebral Blood Flow and Metabolism, 5, 413-419. Hall, A. & Lalli, G. (2010) Rho and Ras GTPases in axon growth, guidance, and branching. Cold Spring Harbor perspectives in biology, 2, a001818. Hallbook, F., Ibanez, C.F. & Persson, H. (1991) Evolutionary studies of the nerve growth factor family reveal a novel member abundantly expressed in Xenopus ovary. Neuron, 6, 845- 858. Hamani, C. & Mello, L.E. (2002) Spontaneous recurrent seizures and neuropathology in the chronic phase of the pilocarpine and picrotoxin model epilepsy. Neurological research, 24, 199-209. Han, J., Pollak, J., Yang, T., Siddiqui, M.R., Doyle, K.P., Taravosh-Lahn, K., Cekanaviciute, E., Han, A., Goodman, J.Z., Jones, B., Jing, D., Massa, S.M., Longo, F.M. & Buckwalter, M.S. (2012) Delayed administration of a small molecule tropomyosin-related kinase B ligand promotes recovery after hypoxic-ischemic stroke. Stroke; a journal of cerebral circulation, 43, 1918-1924.

214

Hannesson, D.K., Armitage, L.L., Mohapel, P. & Corcoran, M.E. (1997) Time course of mossy fiber sprouting following bilateral transection of the fimbria/fornix. Neuroreport, 8, 2299- 2303. Harris, K.M. (1999) Structure, development, and plasticity of dendritic spines. Current opinion in neurobiology, 9, 343-348. Harris, K.M., Jensen, F.E. & Tsao, B. (1992) Three-dimensional structure of dendritic spines and synapses in rat hippocampus (CA1) at postnatal day 15 and adult ages: implications for the maturation of synaptic physiology and long-term potentiation. The Journal of neuroscience : the official journal of the Society for Neuroscience, 12, 2685-2705. Harris, K.M. & Stevens, J.K. (1989) Dendritic spines of CA 1 pyramidal cells in the rat hippocampus: serial electron microscopy with reference to their biophysical characteristics. The Journal of neuroscience : the official journal of the Society for Neuroscience, 9, 2982-2997. Harris, K.M. & Weinberg, R.J. (2012) Ultrastructure of synapses in the mammalian brain. Cold Spring Harbor perspectives in biology, 4. Hasbani, M.J., Schlief, M.L., Fisher, D.A. & Goldberg, M.P. (2001) Dendritic spines lost during glutamate receptor activation reemerge at original sites of synaptic contact. The Journal of neuroscience : the official journal of the Society for Neuroscience, 21, 2393-2403. Hashimoto, T. & Lewis, D.A. (2006) BDNF Val66Met polymorphism and GAD67 mRNA expression in the prefrontal cortex of subjects with schizophrenia. The American journal of psychiatry, 163, 534-537. Hauser, W.A., Annegers, J.F. & Kurland, L.T. (1993) Incidence of epilepsy and unprovoked seizures in Rochester, Minnesota: 1935-1984. Epilepsia, 34, 453-468. Hauser, W.A. & Lee, J.R. (2002) Do seizures beget seizures? Progress in brain research, 135, 215-219. Haydon, P.G. & Nedergaard, M. (2014) How Do Astrocytes Participate in Neural Plasticity? Cold Spring Harbor perspectives in biology, 7. Hebb, D.O. (1949) The Organization of Behavior: A Neuropsychological Theory. Wiley & Sons, New York, NY. Helmstaedter, C. (2002) Effects of chronic epilepsy on declarative memory systems. Progress in brain research, 135, 439-453. Helmstaedter, C., Pohl, C., Hufnagel, A. & Elger, C.E. (1991) Visual learning deficits in nonresected patients with right temporal lobe epilepsy. Cortex; a journal devoted to the study of the nervous system and behavior, 27, 547-555. Hempstead, B.L., Martin-Zanca, D., Kaplan, D.R., Parada, L.F. & Chao, M.V. (1991) High- affinity NGF binding requires coexpression of the trk proto-oncogene and the low-affinity NGF receptor. Nature, 350, 678-683. Henneberger, C., Juttner, R., Rothe, T. & Grantyn, R. (2002) Postsynaptic action of BDNF on GABAergic synaptic transmission in the superficial layers of the mouse superior colliculus. Journal of neurophysiology, 88, 595-603.

215

Henneberger, C., Kirischuk, S. & Grantyn, R. (2005) Brain-derived neurotrophic factor modulates GABAergic synaptic transmission by enhancing presynaptic glutamic acid decarboxylase 65 levels, promoting asynchronous release and reducing the number of activated postsynaptic receptors. Neuroscience, 135, 749-763. Hicks, R.R., Numan, S., Dhillon, H.S., Prasad, M.R. & Seroogy, K.B. (1997) Alterations in BDNF and NT-3 mRNAs in rat hippocampus after experimental brain trauma. Brain research. Molecular brain research, 48, 401-406. Hicks, R.R., Zhang, L., Dhillon, H.S., Prasad, M.R. & Seroogy, K.B. (1998) Expression of trkB mRNA is altered in rat hippocampus after experimental brain trauma. Brain research. Molecular brain research, 59, 264-268. Hofer, M., Pagliusi, S.R., Hohn, A., Leibrock, J. & Barde, Y.A. (1990) Regional distribution of brain-derived neurotrophic factor mRNA in the adult mouse brain. The EMBO journal, 9, 2459-2464. Hoffman, S.N., Salin, P.A. & Prince, D.A. (1994) Chronic neocortical epileptogenesis in vitro. Journal of neurophysiology, 71, 1762-1773. Hollmann, M. & Heinemann, S. (1994) Cloned glutamate receptors. Annual review of neuroscience, 17, 31-108. Hong, E.J., West, A.E. & Greenberg, M.E. (2005) Transcriptional control of cognitive development. Current opinion in neurobiology, 15, 21-28. Horch, H.W., Kruttgen, A., Portbury, S.D. & Katz, L.C. (1999) Destabilization of cortical dendrites and spines by BDNF. Neuron, 23, 353-364. Horn, J., de Haan, R.J., Vermeulen, M. & Limburg, M. (2001) Very Early Nimodipine Use in Stroke (VENUS): a randomized, double-blind, placebo-controlled trial. Stroke; a journal of cerebral circulation, 32, 461-465. Hosang, G.M., Shiles, C., Tansey, K.E., McGuffin, P. & Uher, R. (2014) Interaction between stress and the BDNF Val66Met polymorphism in depression: a systematic review and meta- analysis. BMC medicine, 12, 7. Houser, C.R. (1990) Granule cell dispersion in the dentate gyrus of humans with temporal lobe epilepsy. Brain research, 535, 195-204. Houston, C.M. & Smart, T.G. (2006) CaMK-II modulation of GABA(A) receptors expressed in HEK293, NG108-15 and rat cerebellar granule neurons. The European journal of neuroscience, 24, 2504-2514. Hu, B., Liu, C., Bramlett, H., Sick, T.J., Alonso, O.F., Chen, S. & Dietrich, W.D. (2004) Changes in trkB-ERK1/2-CREB/Elk-1 pathways in hippocampal mossy fiber organization after traumatic brain injury. Journal of cerebral blood flow and metabolism : official journal of the International Society of Cerebral Blood Flow and Metabolism, 24, 934-943. Hu, B., Nikolakopoulou, A.M. & Cohen-Cory, S. (2005) BDNF stabilizes synapses and maintains the structural complexity of optic axons in vivo. Development, 132, 4285-4298. Huang, E.J. & Reichardt, L.F. (2001) Neurotrophins: roles in neuronal development and function. Annual review of neuroscience, 24, 677-736.

216

Huang, E.J. & Reichardt, L.F. (2003) Trk receptors: roles in neuronal signal transduction. Annual review of biochemistry, 72, 609-642. Huang, Z.J., Kirkwood, A., Pizzorusso, T., Porciatti, V., Morales, B., Bear, M.F., Maffei, L. & Tonegawa, S. (1999) BDNF regulates the maturation of inhibition and the critical period of plasticity in mouse visual cortex. Cell, 98, 739-755. Hugel, S., Abegg, M., de Paola, V., Caroni, P., Gahwiler, B.H. & McKinney, R.A. (2009) Dendritic spine morphology determines membrane-associated protein exchange between dendritic shafts and spine heads. Cerebral cortex, 19, 697-702. Hwang, J.J., Park, M.H., Choi, S.Y. & Koh, J.Y. (2005) Activation of the Trk signaling pathway by extracellular zinc. Role of metalloproteinases. The Journal of biological chemistry, 280, 11995-12001. Ichisaka, S., Katoh-Semba, R., Hata, Y., Ohshima, M., Kameyama, K. & Tsumoto, T. (2003) Activity-dependent change in the protein level of brain-derived neurotrophic factor but no change in other neurotrophins in the visual cortex of young and adult ferrets. Neuroscience, 117, 361-371. Inagaki, N., Thoenen, H. & Lindholm, D. (1995) TrkA tyrosine residues involved in NGF-induced neurite outgrowth of PC12 cells. The European journal of neuroscience, 7, 1125-1133. Ip, N.Y., Ibanez, C.F., Nye, S.H., McClain, J., Jones, P.F., Gies, D.R., Belluscio, L., Le Beau, M.M., Espinosa, R., 3rd, Squinto, S.P. & et al. (1992) Mammalian neurotrophin-4: structure, chromosomal localization, tissue distribution, and receptor specificity. Proceedings of the National Academy of Sciences of the United States of America, 89, 3060-3064. Isackson, P.J., Huntsman, M.M., Murray, K.D. & Gall, C.M. (1991) BDNF mRNA expression is increased in adult rat forebrain after limbic seizures: temporal patterns of induction distinct from NGF. Neuron, 6, 937-948. Isokawa, M. (2000) Remodeling dendritic spines of dentate granule cells in temporal lobe epilepsy patients and the rat pilocarpine model. Epilepsia, 41 Suppl 6, S14-17. Isokawa, M. & Levesque, M.F. (1991) Increased NMDA responses and dendritic degeneration in human epileptic hippocampal neurons in slices. Neuroscience letters, 132, 212-216. Itami, C., Kimura, F., Kohno, T., Matsuoka, M., Ichikawa, M., Tsumoto, T. & Nakamura, S. (2003) Brain-derived neurotrophic factor-dependent unmasking of "silent" synapses in the developing mouse barrel cortex. Proceedings of the National Academy of Sciences of the United States of America, 100, 13069-13074. Ito, D., Imai, Y., Ohsawa, K., Nakajima, K., Fukuuchi, Y. & Kohsaka, S. (1998) Microglia-specific localisation of a novel calcium binding protein, Iba1. Brain research. Molecular brain research, 57, 1-9. Ito, M. & Kano, M. (1982) Long-lasting depression of parallel fiber-Purkinje cell transmission induced by conjunctive stimulation of parallel fibers and climbing fibers in the cerebellar cortex. Neuroscience letters, 33, 253-258. Jankowsky, J.L. & Patterson, P.H. (2001) The role of cytokines and growth factors in seizures and their sequelae. Progress in neurobiology, 63, 125-149.

217

Ji, Y., Lu, Y., Yang, F., Shen, W., Tang, T.T., Feng, L., Duan, S. & Lu, B. (2010) Acute and gradual increases in BDNF concentration elicit distinct signaling and functions in neurons. Nature neuroscience, 13, 302-309. Ji, Y., Pang, P.T., Feng, L. & Lu, B. (2005) Cyclic AMP controls BDNF-induced TrkB phosphorylation and dendritic spine formation in mature hippocampal neurons. Nature neuroscience, 8, 164-172. Jiang, M., Lee, C.L., Smith, K.L. & Swann, J.W. (1998) Spine loss and other persistent alterations of hippocampal pyramidal cell dendrites in a model of early-onset epilepsy. The Journal of neuroscience : the official journal of the Society for Neuroscience, 18, 8356-8368. Jing, S., Tapley, P. & Barbacid, M. (1992) Nerve growth factor mediates signal transduction through trk homodimer receptors. Neuron, 9, 1067-1079. Jones, P.L., Veenstra, G.J., Wade, P.A., Vermaak, D., Kass, S.U., Landsberger, N., Strouboulis, J. & Wolffe, A.P. (1998) Methylated DNA and MeCP2 recruit histone deacetylase to repress transcription. Nature genetics, 19, 187-191. Jontes, J.D. & Smith, S.J. (2000) Filopodia, spines, and the generation of synaptic diversity. Neuron, 27, 11-14. Jovanovic, J.N., Czernik, A.J., Fienberg, A.A., Greengard, P. & Sihra, T.S. (2000) Synapsins as mediators of BDNF-enhanced neurotransmitter release. Nature neuroscience, 3, 323-329. Jovanovic, J.N., Thomas, P., Kittler, J.T., Smart, T.G. & Moss, S.J. (2004) Brain-derived neurotrophic factor modulates fast synaptic inhibition by regulating GABA(A) receptor phosphorylation, activity, and cell-surface stability. The Journal of neuroscience : the official journal of the Society for Neuroscience, 24, 522-530. Kandel, E.R. (2012) The molecular biology of memory: cAMP, PKA, CRE, CREB-1, CREB-2, and CPEB. Molecular brain, 5, 14. Kanematsu, T., Yasunaga, A., Mizoguchi, Y., Kuratani, A., Kittler, J.T., Jovanovic, J.N., Takenaka, K., Nakayama, K.I., Fukami, K., Takenawa, T., Moss, S.J., Nabekura, J. & Hirata, M. (2006) Modulation of GABA(A) receptor phosphorylation and membrane trafficking by phospholipase C-related inactive protein/protein phosphatase 1 and 2A signaling complex underlying brain-derived neurotrophic factor-dependent regulation of GABAergic inhibition. The Journal of biological chemistry, 281, 22180-22189. Kang, H. & Schuman, E.M. (1995) Long-lasting neurotrophin-induced enhancement of synaptic transmission in the adult hippocampus. Science, 267, 1658-1662. Kang, H., Welcher, A.A., Shelton, D. & Schuman, E.M. (1997) Neurotrophins and time: different roles for TrkB signaling in hippocampal long-term potentiation. Neuron, 19, 653-664. Kaplan, B.J. & Williamson, P.D. (1978) Electroencephalogram and somatosensory evoked potential changes after administration of six convulsant drugs. Experimental neurology, 59, 124-136. Kaplan, D.R., Hempstead, B.L., Martin-Zanca, D., Chao, M.V. & Parada, L.F. (1991a) The trk proto-oncogene product: a signal transducing receptor for nerve growth factor. Science, 252, 554-558.

218

Kaplan, D.R., Martin-Zanca, D. & Parada, L.F. (1991b) Tyrosine phosphorylation and tyrosine kinase activity of the trk proto-oncogene product induced by NGF. Nature, 350, 158-160. Kasugai, Y., Swinny, J.D., Roberts, J.D., Dalezios, Y., Fukazawa, Y., Sieghart, W., Shigemoto, R. & Somogyi, P. (2010) Quantitative localisation of synaptic and extrasynaptic GABAA receptor subunits on hippocampal pyramidal cells by freeze-fracture replica immunolabelling. The European journal of neuroscience, 32, 1868-1888. Katsumaru, H., Kosaka, T., Heizmann, C.W. & Hama, K. (1988) Immunocytochemical study of GABAergic neurons containing the calcium-binding protein parvalbumin in the rat hippocampus. Experimental brain research, 72, 347-362. Kavanaugh, W.M. & Williams, L.T. (1994) An alternative to SH2 domains for binding tyrosine- phosphorylated proteins. Science, 266, 1862-1865. Kawaguchi, Y., Katsumaru, H., Kosaka, T., Heizmann, C.W. & Hama, K. (1987) Fast spiking cells in rat hippocampus (CA1 region) contain the calcium-binding protein parvalbumin. Brain research, 416, 369-374. Kelleher, R.J., Govindarajan, A. & Tonegawa, S. (2004) Translational regulatory mechanisms in persistent forms of synaptic plasticity. Neuron, 44, 59-73. Kettenmann, H., Hanisch, U.K., Noda, M. & Verkhratsky, A. (2011) Physiology of microglia. Physiological reviews, 91, 461-553. Kharatishvili, I., Nissinen, J.P., McIntosh, T.K. & Pitkanen, A. (2006) A model of posttraumatic epilepsy induced by lateral fluid-percussion brain injury in rats. Neuroscience, 140, 685- 697. Kharazia, V.N. & Weinberg, R.J. (1999) Immunogold localization of AMPA and NMDA receptors in somatic sensory cortex of albino rat. The Journal of comparative neurology, 412, 292- 302. Kim, E. & Sheng, M. (2004) PDZ domain proteins of synapses. Nature reviews. Neuroscience, 5, 771-781. Kim, H.G., Wang, T., Olafsson, P. & Lu, B. (1994) Neurotrophin 3 potentiates neuronal activity and inhibits gamma-aminobutyratergic synaptic transmission in cortical neurons. Proceedings of the National Academy of Sciences of the United States of America, 91, 12341-12345. Kim, J. & Tsien, R.W. (2008) Synapse-specific adaptations to inactivity in hippocampal circuits achieve homeostatic gain control while dampening network reverberation. Neuron, 58, 925-937. Kirino, T., Tamura, A. & Sano, K. (1984) Delayed neuronal death in the rat hippocampus following transient forebrain ischemia. Acta neuropathologica, 64, 139-147. Kistner, U., Wenzel, B.M., Veh, R.W., Cases-Langhoff, C., Garner, A.M., Appeltauer, U., Voss, B., Gundelfinger, E.D. & Garner, C.C. (1993) SAP90, a rat presynaptic protein related to the product of the Drosophila tumor suppressor gene dlg-A. The Journal of biological chemistry, 268, 4580-4583.

219

Klausberger, T. (2009) GABAergic interneurons targeting dendrites of pyramidal cells in the CA1 area of the hippocampus. The European journal of neuroscience, 30, 947-957. Kleiman, R., Banker, G. & Steward, O. (1990) Differential subcellular localization of particular mRNAs in hippocampal neurons in culture. Neuron, 5, 821-830. Klein, R., Jing, S.Q., Nanduri, V., O'Rourke, E. & Barbacid, M. (1991) The trk proto-oncogene encodes a receptor for nerve growth factor. Cell, 65, 189-197. Klose, R.J. & Bird, A.P. (2006) Genomic DNA methylation: the mark and its mediators. Trends in biochemical sciences, 31, 89-97. Kneussel, M., Brandstatter, J.H., Gasnier, B., Feng, G., Sanes, J.R. & Betz, H. (2001) Gephyrin- independent clustering of postsynaptic GABA(A) receptor subtypes. Molecular and cellular neurosciences, 17, 973-982. Kneussel, M., Brandstatter, J.H., Laube, B., Stahl, S., Muller, U. & Betz, H. (1999) Loss of postsynaptic GABA(A) receptor clustering in gephyrin-deficient mice. The Journal of neuroscience : the official journal of the Society for Neuroscience, 19, 9289-9297. Kocsis, K., Knapp, L., Gellert, L., Olah, G., Kis, Z., Takakuwa, H., Iwamori, N., Ono, E., Toldi, J. & Farkas, T. (2014) Acetyl-L-carnitine normalizes the impaired long-term potentiation and spine density in a rat model of global ischemia. Neuroscience, 269, 265-272. Kohara, K., Yasuda, H., Huang, Y., Adachi, N., Sohya, K. & Tsumoto, T. (2007) A local reduction in cortical GABAergic synapses after a loss of endogenous brain-derived neurotrophic factor, as revealed by single-cell gene knock-out method. The Journal of neuroscience : the official journal of the Society for Neuroscience, 27, 7234-7244. Kokaia, M., Ernfors, P., Kokaia, Z., Elmer, E., Jaenisch, R. & Lindvall, O. (1995) Suppressed epileptogenesis in BDNF mutant mice. Experimental neurology, 133, 215-224. Kolarow, R., Brigadski, T. & Lessmann, V. (2007) Postsynaptic secretion of BDNF and NT-3 from hippocampal neurons depends on calcium calmodulin kinase II signaling and proceeds via delayed fusion pore opening. The Journal of neuroscience : the official journal of the Society for Neuroscience, 27, 10350-10364. Kolbeck, R., Bartke, I., Eberle, W. & Barde, Y.A. (1999) Brain-derived neurotrophic factor levels in the nervous system of wild-type and neurotrophin gene mutant mice. Journal of neurochemistry, 72, 1930-1938. Korte, M., Carroll, P., Wolf, E., Brem, G., Thoenen, H. & Bonhoeffer, T. (1995) Hippocampal long-term potentiation is impaired in mice lacking brain-derived neurotrophic factor. Proceedings of the National Academy of Sciences of the United States of America, 92, 8856-8860. Korte, M., Griesbeck, O., Gravel, C., Carroll, P., Staiger, V., Thoenen, H. & Bonhoeffer, T. (1996) Virus-mediated gene transfer into hippocampal CA1 region restores long-term potentiation in brain-derived neurotrophic factor mutant mice. Proceedings of the National Academy of Sciences of the United States of America, 93, 12547-12552. Kotlyanskaya, L., McLinden, K.A. & Giniger, E. (2013) Of proneurotrophins and their antineurotrophic effects. Science signaling, 6, pe6.

220

Kovalchuk, Y., Hanse, E., Kafitz, K.W. & Konnerth, A. (2002) Postsynaptic Induction of BDNF- Mediated Long-Term Potentiation. Science, 295, 1729-1734. Kowalczyk, S., Winkelmann, A., Smolinsky, B., Forstera, B., Neundorf, I., Schwarz, G. & Meier, J.C. (2013) Direct binding of GABAA receptor beta2 and beta3 subunits to gephyrin. The European journal of neuroscience, 37, 544-554. Kullmann, D.M., Moreau, A.W., Bakiri, Y. & Nicholson, E. (2012) Plasticity of inhibition. Neuron, 75, 951-962. Kurz, J.E., Moore, B.J., Henderson, S.C., Campbell, J.N. & Churn, S.B. (2008) A cellular mechanism for dendritic spine loss in the pilocarpine model of status epilepticus. Epilepsia, 49, 1696-1710. Laake, J.H., Haug, F.M., Wieloch, T. & Ottersen, O.P. (1999) A simple in vitro model of ischemia based on hippocampal slice cultures and propidium iodide fluorescence. Brain research. Brain research protocols, 4, 173-184. Lachance-Touchette, P., Brown, P., Meloche, C., Kinirons, P., Lapointe, L., Lacasse, H., Lortie, A., Carmant, L., Bedford, F., Bowie, D. & Cossette, P. (2011) Novel alpha1 and gamma2 GABAA receptor subunit mutations in families with idiopathic generalized epilepsy. The European journal of neuroscience, 34, 237-249. Lamballe, F., Klein, R. & Barbacid, M. (1991) trkC, a new member of the trk family of tyrosine protein kinases, is a receptor for neurotrophin-3. Cell, 66, 967-979. Lang, S.B., Stein, V., Bonhoeffer, T. & Lohmann, C. (2007) Endogenous brain-derived neurotrophic factor triggers fast calcium transients at synapses in developing dendrites. The Journal of neuroscience : the official journal of the Society for Neuroscience, 27, 1097- 1105. Larimore, J.L., Chapleau, C.A., Kudo, S., Theibert, A., Percy, A.K. & Pozzo-Miller, L. (2009) Bdnf overexpression in hippocampal neurons prevents dendritic atrophy caused by Rett- associated MECP2 mutations. Neurobiology of disease, 34, 199-211. Laurberg, S. & Zimmer, J. (1981) Lesion-induced sprouting of hippocampal mossy fiber collaterals to the fascia dentata in developing and adult rats. The Journal of comparative neurology, 200, 433-459. Lee, B. & Pyun, S.B. (2014) Characteristics of Cognitive Impairment in Patients With Post-stroke Aphasia. Annals of rehabilitation medicine, 38, 759-765. Lee, E.J., Gibo, T.L. & Grzywacz, N.M. (2006) Dark-rearing-induced reduction of GABA and GAD and prevention of the effect by BDNF in the mouse retina. The European journal of neuroscience, 24, 2118-2134. Lee, R., Kermani, P., Teng, K.K. & Hempstead, B.L. (2001) Regulation of cell survival by secreted proneurotrophins. Science, 294, 1945-1948. Lee, T.H., Kato, H., Chen, S.T., Kogure, K. & Itoyama, Y. (2002) Expression disparity of brain- derived neurotrophic factor immunoreactivity and mRNA in ischemic hippocampal neurons. Neuroreport, 13, 2271-2275.

221

Lehmann, T.N., Gabriel, S., Kovacs, R., Eilers, A., Kivi, A., Schulze, K., Lanksch, W.R., Meencke, H.J. & Heinemann, U. (2000) Alterations of neuronal connectivity in area CA1 of hippocampal slices from temporal lobe epilepsy patients and from pilocarpine-treated epileptic rats. Epilepsia, 41 Suppl 6, S190-194. Leibrock, J., Lottspeich, F., Hohn, A., Hofer, M., Hengerer, B., Masiakowski, P., Thoenen, H. & Barde, Y.A. (1989) Molecular cloning and expression of brain-derived neurotrophic factor. Nature, 341, 149-152. Lessmann, V., Gottmann, K. & Heumann, R. (1994) BDNF and NT-4/5 enhance glutamatergic synaptic transmission in cultured hippocampal neurones. Neuroreport, 6, 21-25. Lessmann, V., Gottmann, K. & Malcangio, M. (2003) Neurotrophin secretion: current facts and future prospects. Progress in neurobiology, 69, 341-374. Lessmann, V. & Heumann, R. (1998) Modulation of unitary glutamatergic synapses by neurotrophin-4/5 or brain-derived neurotrophic factor in hippocampal microcultures: presynaptic enhancement depends on pre-established paired-pulse facilitation. Neuroscience, 86, 399-413. Letty, S., Lerner-Natoli, M. & Rondouin, G. (1995) Differential impairments of spatial memory and social behavior in two models of limbic epilepsy. Epilepsia, 36, 973-982. Levi-Montalcini, R. & Hamburger, V. (1951) Selective growth stimulating effects of mouse sarcoma on the sensory and sympathetic nervous system of the chick embryo. The Journal of experimental zoology, 116, 321-361. Levine, E.S., Dreyfus, C.F., Black, I.B. & Plummer, M.R. (1995a) Brain-derived neurotrophic factor rapidly enhances synaptic transmission in hippocampal neurons via postsynaptic tyrosine kinase receptors. Proceedings of the National Academy of Sciences of the United States of America, 92, 8074-8077. Levine, E.S., Dreyfus, C.F., Black, I.B. & Plummer, M.R. (1995b) Differential effects of NGF and BDNF on voltage-gated calcium currents in embryonic basal forebrain neurons. The Journal of neuroscience : the official journal of the Society for Neuroscience, 15, 3084- 3091. Levitan, I.B. & Kaczmarek, L.K. (2002) The Neuron: Cell and Molecular Biology. Oxford University Press. Lewin, G.R. & Barde, Y.A. (1996) Physiology of the neurotrophins. Annual review of neuroscience, 19, 289-317. Li, H., Siegel, R.E. & Schwartz, R.D. (1993) Rapid decline of GABAA receptor subunit mRNA expression in hippocampus following transient cerebral ischemia in the gerbil. Hippocampus, 3, 527-537. Li, W., Huang, R., Shetty, R.A., Thangthaeng, N., Liu, R., Chen, Z., Sumien, N., Rutledge, M., Dillon, G.H., Yuan, F., Forster, M.J., Simpkins, J.W. & Yang, S.H. (2013) Transient focal cerebral ischemia induces long-term cognitive function deficit in an experimental ischemic stroke model. Neurobiology of disease, 59, 18-25. Li, W. & Pozzo-Miller, L. (2014) BDNF deregulation in Rett syndrome. Neuropharmacology, 76 Pt C, 737-746.

222

Li, Y.X., Zhang, Y., Lester, H.A., Schuman, E.M. & Davidson, N. (1998) Enhancement of neurotransmitter release induced by brain-derived neurotrophic factor in cultured hippocampal neurons. The Journal of neuroscience : the official journal of the Society for Neuroscience, 18, 10231-10240. Liepinsh, E., Ilag, L.L., Otting, G. & Ibanez, C.F. (1997) NMR structure of the death domain of the p75 neurotrophin receptor. The EMBO journal, 16, 4999-5005. Lin, L.F., Doherty, D.H., Lile, J.D., Bektesh, S. & Collins, F. (1993) GDNF: a glial cell line- derived neurotrophic factor for midbrain dopaminergic neurons. Science, 260, 1130-1132. Lin, S.Y., Wu, K., Levine, E.S., Mount, H.T., Suen, P.C. & Black, I.B. (1998) BDNF acutely increases tyrosine phosphorylation of the NMDA receptor subunit 2B in cortical and hippocampal postsynaptic densities. Brain research. Molecular brain research, 55, 20-27. Lindholm, D., da Penha Berzaghi, M., Cooper, J., Thoenen, H. & Castren, E. (1994) Brain-derived neurotrophic factor and neurotrophin-4 increase neurotrophin-3 expression in the rat hippocampus. International journal of developmental neuroscience : the official journal of the International Society for Developmental Neuroscience, 12, 745-751. Lindvall, O., Ernfors, P., Bengzon, J., Kokaia, Z., Smith, M.L., Siesjo, B.K. & Persson, H. (1992) Differential regulation of mRNAs for nerve growth factor, brain-derived neurotrophic factor, and neurotrophin 3 in the adult rat brain following cerebral ischemia and hypoglycemic coma. Proceedings of the National Academy of Sciences of the United States of America, 89, 648-652. Lindvall, O., Kokaia, Z., Bengzon, J., Elmer, E. & Kokaia, M. (1994) Neurotrophins and brain insults. Trends in neurosciences, 17, 490-496. Lipton, P. (1999) Ischemic cell death in brain neurons. Physiological reviews, 79, 1431-1568. Liu, B., Liao, M., Mielke, J.G., Ning, K., Chen, Y., Li, L., El-Hayek, Y.H., Gomez, E., Zukin, R.S., Fehlings, M.G. & Wan, Q. (2006) Ischemic insults direct glutamate receptor subunit 2-lacking AMPA receptors to synaptic sites. The Journal of neuroscience : the official journal of the Society for Neuroscience, 26, 5309-5319. Loeb, D.M., Stephens, R.M., Copeland, T., Kaplan, D.R. & Greene, L.A. (1994) A Trk nerve growth factor (NGF) receptor point mutation affecting interaction with phospholipase C- gamma 1 abolishes NGF-promoted peripherin induction but not neurite outgrowth. The Journal of biological chemistry, 269, 8901-8910. Lohmann, C. & Bonhoeffer, T. (2008) A role for local calcium signaling in rapid synaptic partner selection by dendritic filopodia. Neuron, 59, 253-260. Lohmann, C., Finski, A. & Bonhoeffer, T. (2005) Local calcium transients regulate the spontaneous motility of dendritic filopodia. Nature neuroscience, 8, 305-312. Lohof, A.M., Ip, N.Y. & Poo, M.M. (1993) Potentiation of developing neuromuscular synapses by the neurotrophins NT-3 and BDNF. Nature, 363, 350-353. Lomo, T. (2003) The discovery of long-term potentiation. Philosophical transactions of the Royal Society of London. Series B, Biological sciences, 358, 617-620.

223

Lorenzo, L.E., Godin, A.G., Wang, F., St-Louis, M., Carbonetto, S., Wiseman, P.W., Ribeiro-da- Silva, A. & De Koninck, Y. (2014) Gephyrin clusters are absent from small diameter primary afferent terminals despite the presence of GABA(A) receptors. The Journal of neuroscience : the official journal of the Society for Neuroscience, 34, 8300-8317. Lou, H., Kim, S.K., Zaitsev, E., Snell, C.R., Lu, B. & Loh, Y.P. (2005) Sorting and activity- dependent secretion of BDNF require interaction of a specific motif with the sorting receptor carboxypeptidase e. Neuron, 45, 245-255. Loup, F., Picard, F., Andre, V.M., Kehrli, P., Yonekawa, Y., Wieser, H.G. & Fritschy, J.M. (2006) Altered expression of alpha3-containing GABAA receptors in the neocortex of patients with focal epilepsy. Brain : a journal of neurology, 129, 3277-3289. Loup, F., Picard, F., Yonekawa, Y., Wieser, H.G. & Fritschy, J.M. (2009) Selective changes in GABAA receptor subtypes in white matter neurons of patients with focal epilepsy. Brain : a journal of neurology, 132, 2449-2463. Lu, B. (2003) BDNF and activity-dependent synaptic modulation. Learning & memory, 10, 86-98. Luikart, B.W., Nef, S., Virmani, T., Lush, M.E., Liu, Y., Kavalali, E.T. & Parada, L.F. (2005) TrkB has a cell-autonomous role in the establishment of hippocampal Schaffer collateral synapses. The Journal of neuroscience : the official journal of the Society for Neuroscience, 25, 3774-3786. Luikart, B.W., Zhang, W., Wayman, G.A., Kwon, C.H., Westbrook, G.L. & Parada, L.F. (2008) Neurotrophin-dependent dendritic filopodial motility: a convergence on PI3K signaling. The Journal of neuroscience : the official journal of the Society for Neuroscience, 28, 7006- 7012. Lundstrom, K., Abenavoli, A., Malgaroli, A. & Ehrengruber, M.U. (2003) Novel Semliki Forest virus vectors with reduced cytotoxicity and temperature sensitivity for long-term enhancement of transgene expression. Molecular therapy : the journal of the American Society of Gene Therapy, 7, 202-209. Lundstrom, K. & Ehrengruber, M.U. (2003) Semliki Forest virus (SFV) vectors in neurobiology and gene therapy. Methods in molecular medicine, 76, 503-523. Luo, L., Hensch, T.K., Ackerman, L., Barbel, S., Jan, L.Y. & Jan, Y.N. (1996) Differential effects of the Rac GTPase on Purkinje cell axons and dendritic trunks and spines. Nature, 379, 837-840. Luscher, C., Nicoll, R.A., Malenka, R.C. & Muller, D. (2000) Synaptic plasticity and dynamic modulation of the postsynaptic membrane. Nature neuroscience, 3, 545-550. Lyden, P., Jacoby, M., Schim, J., Albers, G., Mazzeo, P., Ashwood, T., Nordlund, A. & Odergren, T. (2001) The Clomethiazole Acute Stroke Study in tissue-type plasminogen activator- treated stroke (CLASS-T): final results. Neurology, 57, 1199-1205. Lynch, G., Larson, J., Kelso, S., Barrionuevo, G. & Schottler, F. (1983) Intracellular injections of EGTA block induction of hippocampal long-term potentiation. Nature, 305, 719-721. Ma, L., Reis, G., Parada, L.F. & Schuman, E.M. (1999) Neuronal NT-3 is not required for synaptic transmission or long-term potentiation in area CA1 of the adult rat hippocampus. Learning & memory, 6, 267-275.

224

Machnes, Z.M., Huang, T.C., Chang, P.K., Gill, R., Reist, N., Dezsi, G., Ozturk, E., Charron, F., O'Brien, T.J., Jones, N.C., McKinney, R.A. & Szyf, M. (2013) DNA methylation mediates persistent epileptiform activity in vitro and in vivo. PloS one, 8, e76299. MacVicar, B.A. & Dudek, F.E. (1980a) Dye-coupling betwen CA3 pyramidal cells in slices of rat hippocampus. Brain research, 196, 494-497. MacVicar, B.A. & Dudek, F.E. (1980b) Local synaptic circuits in rat hippocampus: interactions between pyramidal cells. Brain research, 184, 220-223. Magby, J.P., Bi, C., Chen, Z.Y., Lee, F.S. & Plummer, M.R. (2006) Single-cell characterization of retrograde signaling by brain-derived neurotrophic factor. The Journal of neuroscience : the official journal of the Society for Neuroscience, 26, 13531-13536. Maier, D.L., Mani, S., Donovan, S.L., Soppet, D., Tessarollo, L., McCasland, J.S. & Meiri, K.F. (1999) Disrupted cortical map and absence of cortical barrels in growth-associated protein (GAP)-43 knockout mice. Proceedings of the National Academy of Sciences of the United States of America, 96, 9397-9402. Maisonpierre, P.C., Belluscio, L., Friedman, B., Alderson, R.F., Wiegand, S.J., Furth, M.E., Lindsay, R.M. & Yancopoulos, G.D. (1990a) NT-3, BDNF, and NGF in the developing rat nervous system: parallel as well as reciprocal patterns of expression. Neuron, 5, 501-509. Maisonpierre, P.C., Belluscio, L., Squinto, S., Ip, N.Y., Furth, M.E., Lindsay, R.M. & Yancopoulos, G.D. (1990b) Neurotrophin-3: a neurotrophic factor related to NGF and BDNF. Science, 247, 1446-1451. Maletic-Savatic, M., Malinow, R. & Svoboda, K. (1999) Rapid dendritic morphogenesis in CA1 hippocampal dendrites induced by synaptic activity. Science, 283, 1923-1927. Malinow, R. & Miller, J.P. (1986) Postsynaptic hyperpolarization during conditioning reversibly blocks induction of long-term potentiation. Nature, 320, 529-530. Manthorpe, M., Skaper, S., Adler, R., Landa, K. & Varon, S. (1980) Cholinergic neuronotrophic factors: fractionation properties of an extract from selected chick embryonic eye tissues. Journal of neurochemistry, 34, 69-75. Marks, D.A., Kim, J., Spencer, D.D. & Spencer, S.S. (1995) Seizure localization and pathology following head injury in patients with uncontrolled epilepsy. Neurology, 45, 2051-2057. Marler, K.J., Poopalasundaram, S., Broom, E.R., Wentzel, C. & Drescher, U. (2010) Pro- neurotrophins secreted from retinal ganglion cell axons are necessary for ephrinA- p75NTR-mediated axon guidance. Neural development, 5, 30. Marsden, K.C., Beattie, J.B., Friedenthal, J. & Carroll, R.C. (2007) NMDA receptor activation potentiates inhibitory transmission through GABA receptor-associated protein-dependent exocytosis of GABA(A) receptors. The Journal of neuroscience : the official journal of the Society for Neuroscience, 27, 14326-14337. Marshak, S., Nikolakopoulou, A.M., Dirks, R., Martens, G.J. & Cohen-Cory, S. (2007) Cell- autonomous TrkB signaling in presynaptic retinal ganglion cells mediates axon arbor growth and synapse maturation during the establishment of retinotectal synaptic connectivity. The Journal of neuroscience : the official journal of the Society for Neuroscience, 27, 2444-2456.

225

Martinez, A. & Soriano, E. (2005) Functions of ephrin/Eph interactions in the development of the nervous system: emphasis on the hippocampal system. Brain research. Brain research reviews, 49, 211-226. Martinowich, K., Hattori, D., Wu, H., Fouse, S., He, F., Hu, Y., Fan, G. & Sun, Y.E. (2003) DNA methylation-related chromatin remodeling in activity-dependent BDNF gene regulation. Science, 302, 890-893. Marty, S. (2000) Differences in the regulation of neuropeptide Y, somatostatin and parvalbumin levels in hippocampal interneurons by neuronal activity and BDNF. Progress in brain research, 128, 193-202. Marty, S., Wehrle, R. & Sotelo, C. (2000) Neuronal activity and brain-derived neurotrophic factor regulate the density of inhibitory synapses in organotypic slice cultures of postnatal hippocampus. The Journal of neuroscience : the official journal of the Society for Neuroscience, 20, 8087-8095. Maru, E., Kanda, M. & Ashida, H. (2002) Functional and morphological changes in the hippocampal neuronal circuits associated with epileptic seizures. Epilepsia, 43 Suppl 9, 44-49. Mathern, G.W., Babb, T.L., Micevych, P.E., Blanco, C.E. & Pretorius, J.K. (1997) Granule cell mRNA levels for BDNF, NGF, and NT-3 correlate with neuron losses or supragranular mossy fiber sprouting in the chronically damaged and epileptic human hippocampus. Molecular and chemical neuropathology / sponsored by the International Society for Neurochemistry and the World Federation of Neurology and research groups on neurochemistry and cerebrospinal fluid, 30, 53-76. Matsumoto, T., Rauskolb, S., Polack, M., Klose, J., Kolbeck, R., Korte, M. & Barde, Y.A. (2008) Biosynthesis and processing of endogenous BDNF: CNS neurons store and secrete BDNF, not pro-BDNF. Nature neuroscience, 11, 131-133. Matsuzaki, M., Ellis-Davies, G.C., Nemoto, T., Miyashita, Y., Iino, M. & Kasai, H. (2001) Dendritic spine geometry is critical for AMPA receptor expression in hippocampal CA1 pyramidal neurons. Nature neuroscience, 4, 1086-1092. Matsuzaki, M., Honkura, N., Ellis-Davies, G.C. & Kasai, H. (2004) Structural basis of long-term potentiation in single dendritic spines. Nature, 429, 761-766. Mayer, M.L., Westbrook, G.L. & Guthrie, P.B. (1984) Voltage-dependent block by Mg2+ of NMDA responses in spinal cord neurones. Nature, 309, 261-263. McAllister, A.K., Katz, L.C. & Lo, D.C. (1996) Neurotrophin regulation of cortical dendritic growth requires activity. Neuron, 17, 1057-1064. McDonald, B.J. & Moss, S.J. (1994) Differential phosphorylation of intracellular domains of gamma-aminobutyric acid type A receptor subunits by calcium/calmodulin type 2- dependent protein kinase and cGMP-dependent protein kinase. The Journal of biological chemistry, 269, 18111-18117. McDonald, B.J. & Moss, S.J. (1997) Conserved phosphorylation of the intracellular domains of GABA(A) receptor beta2 and beta3 subunits by cAMP-dependent protein kinase, cGMP-

226

dependent protein kinase protein kinase C and Ca2+/calmodulin type II-dependent protein kinase. Neuropharmacology, 36, 1377-1385. McKernan, R.M. & Whiting, P.J. (1996) Which GABAA-receptor subtypes really occur in the brain? Trends in neurosciences, 19, 139-143. McKinney, R.A. (2005) Physiological roles of spine motility: development, plasticity and disorders. Biochemical Society transactions, 33, 1299-1302. McKinney, R.A. (2010) Excitatory amino acid involvement in dendritic spine formation, maintenance and remodelling. The Journal of physiology, 588, 107-116. McKinney, R.A., Capogna, M., Durr, R., Gahwiler, B.H. & Thompson, S.M. (1999a) Miniature synaptic events maintain dendritic spines via AMPA receptor activation. Nature neuroscience, 2, 44-49. McKinney, R.A., Debanne, D., Gahwiler, B.H. & Thompson, S.M. (1997) Lesion-induced axonal sprouting and hyperexcitability in the hippocampus in vitro: implications for the genesis of posttraumatic epilepsy. Nature medicine, 3, 990-996. McKinney, R.A., Luthi, A., Bandtlow, C.E., Gahwiler, B.H. & Thompson, S.M. (1999b) Selective glutamate receptor antagonists can induce or prevent axonal sprouting in rat hippocampal slice cultures. Proceedings of the National Academy of Sciences of the United States of America, 96, 11631-11636. McNamara, J.O., Huang, Y.Z. & Leonard, A.S. (2006) Molecular signaling mechanisms underlying epileptogenesis. Science's STKE : signal transduction knowledge environment, 2006, re12. McNamara, R.K. & Lenox, R.H. (2000) Differential regulation of primary protein kinase C substrate (MARCKS, MLP, GAP-43, RC3) mRNAs in the hippocampus during kainic acid-induced seizures and synaptic reorganization. Journal of neuroscience research, 62, 416-426. McNamara, R.K. & Routtenberg, A. (1995) NMDA receptor blockade prevents kainate induction of protein F1/GAP-43 mRNA in hippocampal granule cells and subsequent mossy fiber sprouting in the rat. Brain research. Molecular brain research, 33, 22-28. Megias, M., Emri, Z., Freund, T.F. & Gulyas, A.I. (2001) Total number and distribution of inhibitory and excitatory synapses on hippocampal CA1 pyramidal cells. Neuroscience, 102, 527-540. Meiri, K.F. & Gordon-Weeks, P.R. (1990) GAP-43 in growth cones is associated with areas of membrane that are tightly bound to substrate and is a component of a membrane skeleton subcellular fraction. The Journal of neuroscience : the official journal of the Society for Neuroscience, 10, 256-266. Meiri, K.F., Pfenninger, K.H. & Willard, M.B. (1986) Growth-associated protein, GAP-43, a polypeptide that is induced when neurons extend axons, is a component of growth cones and corresponds to pp46, a major polypeptide of a subcellular fraction enriched in growth cones. Proceedings of the National Academy of Sciences of the United States of America, 83, 3537-3541.

227

Mele, M., Ribeiro, L., Inacio, A.R., Wieloch, T. & Duarte, C.B. (2014) GABA(A) receptor dephosphorylation followed by internalization is coupled to neuronal death in in vitro ischemia. Neurobiology of disease, 65, 220-232. Merlio, J.P., Ernfors, P., Kokaia, Z., Middlemas, D.S., Bengzon, J., Kokaia, M., Smith, M.L., Siesjo, B.K., Hunter, T., Lindvall, O. & et al. (1993) Increased production of the TrkB protein tyrosine kinase receptor after brain insults. Neuron, 10, 151-164. Metsis, M., Timmusk, T., Arenas, E. & Persson, H. (1993) Differential usage of multiple brain- derived neurotrophic factor promoters in the rat brain following neuronal activation. Proceedings of the National Academy of Sciences of the United States of America, 90, 8802-8806. Middlemas, D.S., Lindberg, R.A. & Hunter, T. (1991) Trkb, a Neural Receptor Protein-Tyrosine Kinase - Evidence for a Full-Length and 2 Truncated Receptors. Mol Cell Biol, 11, 143- 153. Mies, G., Paschen, W. & Hossmann, K.A. (1990) Cerebral blood flow, glucose utilization, regional glucose, and ATP content during the maturation period of delayed ischemic injury in gerbil brain. Journal of cerebral blood flow and metabolism : official journal of the International Society of Cerebral Blood Flow and Metabolism, 10, 638-645. Miles, R. & Wong, R.K. (1986) Excitatory synaptic interactions between CA3 neurones in the guinea-pig hippocampus. The Journal of physiology, 373, 397-418. Mileson, B.E., Ehrmann, M.L. & Schwartz, R.D. (1992) Alterations in the gamma-aminobutyric acid-gated chloride channel following transient forebrain ischemia in the gerbil. Journal of neurochemistry, 58, 600-607. Minichiello, L. (2009) TrkB signalling pathways in LTP and learning. Nature reviews. Neuroscience, 10, 850-860. Miwa, T., Furukawa, S., Nakajima, K., Furukawa, Y. & Kohsaka, S. (1997) Lipopolysaccharide enhances synthesis of brain-derived neurotrophic factor in cultured rat microglia. Journal of neuroscience research, 50, 1023-1029. Mizoguchi, Y., Ishibashi, H. & Nabekura, J. (2003a) The action of BDNF on GABA(A) currents changes from potentiating to suppressing during maturation of rat hippocampal CA1 pyramidal neurons. The Journal of physiology, 548, 703-709. Mizoguchi, Y., Kanematsu, T., Hirata, M. & Nabekura, J. (2003b) A rapid increase in the total number of cell surface functional GABAA receptors induced by brain-derived neurotrophic factor in rat visual cortex. The Journal of biological chemistry, 278, 44097- 44102. Mizuno, K., Carnahan, J. & Nawa, H. (1994) Brain-derived neurotrophic factor promotes differentiation of striatal GABAergic neurons. Developmental biology, 165, 243-256. Montori, S., Dos Anjos, S., Poole, A., Regueiro-Purrinos, M.M., Llorente, I.L., Darlison, M.G., Fernandez-Lopez, A. & Martinez-Villayandre, B. (2012) Differential effect of transient global ischaemia on the levels of gamma-aminobutyric acid type A (GABA(A)) receptor subunit mRNAs in young and older rats. Neuropathology and applied neurobiology, 38, 710-722.

228

Moretti, P. & Zoghbi, H.Y. (2006) MeCP2 dysfunction in Rett syndrome and related disorders. Current opinion in genetics & development, 16, 276-281. Mou, L., Dias, B.G., Gosnell, H. & Ressler, K.J. (2013) Gephyrin plays a key role in BDNF- dependent regulation of amygdala surface GABAARs. Neuroscience, 255, 33-44. Mowla, S.J., Pareek, S., Farhadi, H.F., Petrecca, K., Fawcett, J.P., Seidah, N.G., Morris, S.J., Sossin, W.S. & Murphy, R.A. (1999) Differential sorting of nerve growth factor and brain- derived neurotrophic factor in hippocampal neurons. The Journal of neuroscience : the official journal of the Society for Neuroscience, 19, 2069-2080. Mudo, G., Persson, H., Timmusk, T., Funakoshi, H., Bindoni, M. & Belluardo, N. (1993) Increased expression of trkB and trkC messenger RNAs in the rat forebrain after focal mechanical injury. Neuroscience, 57, 901-912. Mukherjee, J., Kretschmannova, K., Gouzer, G., Maric, H.M., Ramsden, S., Tretter, V., Harvey, K., Davies, P.A., Triller, A., Schindelin, H. & Moss, S.J. (2011) The residence time of GABA(A)Rs at inhibitory synapses is determined by direct binding of the receptor alpha1 subunit to gephyrin. The Journal of neuroscience : the official journal of the Society for Neuroscience, 31, 14677-14687. Mulkey, R.M., Endo, S., Shenolikar, S. & Malenka, R.C. (1994) Involvement of a calcineurin/inhibitor-1 phosphatase cascade in hippocampal long-term depression. Nature, 369, 486-488. Mulkey, R.M. & Malenka, R.C. (1992) Mechanisms underlying induction of homosynaptic long- term depression in area CA1 of the hippocampus. Neuron, 9, 967-975. Muller, C.J., Groticke, I., Bankstahl, M. & Loscher, W. (2009) Behavioral and cognitive alterations, spontaneous seizures, and neuropathology developing after a pilocarpine- induced status epilepticus in C57BL/6 mice. Experimental neurology, 219, 284-297. Muller, C.M., Vlachos, A. & Deller, T. (2010) Calcium homeostasis of acutely denervated and lesioned dentate gyrus in organotypic entorhino-hippocampal co-cultures. Cell calcium, 47, 242-252. Nadler, J.V. (2003) The recurrent mossy fiber pathway of the epileptic brain. Neurochemical research, 28, 1649-1658. Nadler, J.V., Perry, B.W. & Cotman, C.W. (1978) Intraventricular kainic acid preferentially destroys hippocampal pyramidal cells. Nature, 271, 676-677. Nakayama, A.Y., Harms, M.B. & Luo, L. (2000) Small GTPases Rac and Rho in the maintenance of dendritic spines and branches in hippocampal pyramidal neurons. The Journal of neuroscience : the official journal of the Society for Neuroscience, 20, 5329-5338. Nan, X., Ng, H.H., Johnson, C.A., Laherty, C.D., Turner, B.M., Eisenman, R.N. & Bird, A. (1998) Transcriptional repression by the methyl-CpG-binding protein MeCP2 involves a histone deacetylase complex. Nature, 393, 386-389. Neumann, J.T., Thompson, J.W., Raval, A.P., Cohan, C.H., Koronowski, K.B. & Perez-Pinzon, M.A. (2015) Increased BDNF protein expression after ischemic or PKC epsilon preconditioning promotes electrophysiologic changes that lead to neuroprotection. Journal

229

of cerebral blood flow and metabolism : official journal of the International Society of Cerebral Blood Flow and Metabolism, 35, 121-130. Neves-Pereira, M., Mundo, E., Muglia, P., King, N., Macciardi, F. & Kennedy, J.L. (2002) The brain-derived neurotrophic factor gene confers susceptibility to bipolar disorder: evidence from a family-based association study. American journal of human genetics, 71, 651-655. Neves, G., Cooke, S.F. & Bliss, T.V. (2008) Synaptic plasticity, memory and the hippocampus: a neural network approach to causality. Nature reviews. Neuroscience, 9, 65-75. Newell, D.W., Barth, A., Papermaster, V. & Malouf, A.T. (1995) Glutamate and non-glutamate receptor mediated toxicity caused by oxygen and glucose deprivation in organotypic hippocampal cultures. The Journal of neuroscience : the official journal of the Society for Neuroscience, 15, 7702-7711. Nguyen, P.V. & Woo, N.H. (2003) Regulation of hippocampal synaptic plasticity by cyclic AMP- dependent protein kinases. Progress in neurobiology, 71, 401-437. Nibuya, M., Morinobu, S. & Duman, R.S. (1995) Regulation of BDNF and trkB mRNA in rat brain by chronic electroconvulsive seizure and antidepressant drug treatments. The Journal of neuroscience : the official journal of the Society for Neuroscience, 15, 7539-7547. Nicoll, R.A., Kauer, J.A. & Malenka, R.C. (1988) The current excitement in long-term potentiation. Neuron, 1, 97-103. Nicoll, R.A. & Malenka, R.C. (1995) Contrasting properties of two forms of long-term potentiation in the hippocampus. Nature, 377, 115-118. Nicoll, R.A. & Schmitz, D. (2005) Synaptic plasticity at hippocampal mossy fibre synapses. Nature reviews. Neuroscience, 6, 863-876. Nieto-Sampedro, M., Lewis, E.R., Cotman, C.W., Manthorpe, M., Skaper, S.D., Barbin, G., Longo, F.M. & Varon, S. (1982) Brain injury causes a time-dependent increase in neuronotrophic activity at the lesion site. Science, 217, 860-861. NINDS, r.-P.S.S.G. (1995) Tissue Plasminogen Activator for Acute Ischemic Stroke. New England Journal of Medicine, 333, 1581-1588. Nishizawa, Y. (2001) Glutamate release and neuronal damage in ischemia. Life sciences, 69, 369- 381. Noda, M., Nakanishi, H., Nabekura, J. & Akaike, N. (2000) AMPA-kainate subtypes of glutamate receptor in rat cerebral microglia. The Journal of neuroscience : the official journal of the Society for Neuroscience, 20, 251-258. Noguchi, J., Matsuzaki, M., Ellis-Davies, G.C. & Kasai, H. (2005) Spine-neck geometry determines NMDA receptor-dependent Ca2+ signaling in dendrites. Neuron, 46, 609-622. Noraberg, J., Poulsen, F.R., Blaabjerg, M., Kristensen, B.W., Bonde, C., Montero, M., Meyer, M., Gramsbergen, J.B. & Zimmer, J. (2005) Organotypic hippocampal slice cultures for studies of brain damage, neuroprotection and neurorepair. Current drug targets. CNS and neurological disorders, 4, 435-452.

230

Notaras, M., Hill, R. & van den Buuse, M. (2015) A role for the BDNF gene Val66Met polymorphism in schizophrenia? A comprehensive review. Neuroscience and biobehavioral reviews, 51, 15-30. Nowak, L., Bregestovski, P., Ascher, P., Herbet, A. & Prochiantz, A. (1984) Magnesium gates glutamate-activated channels in mouse central neurones. Nature, 307, 462-465. Nusser, Z., Lujan, R., Laube, G., Roberts, J.D., Molnar, E. & Somogyi, P. (1998) Cell type and pathway dependence of synaptic AMPA receptor number and variability in the hippocampus. Neuron, 21, 545-559. Nygren, J., Kokaia, M. & Wieloch, T. (2006) Decreased expression of brain-derived neurotrophic factor in BDNF(+/-) mice is associated with enhanced recovery of motor performance and increased neuroblast number following experimental stroke. Journal of neuroscience research, 84, 626-631. Nykjaer, A., Lee, R., Teng, K.K., Jansen, P., Madsen, P., Nielsen, M.S., Jacobsen, C., Kliemannel, M., Schwarz, E., Willnow, T.E., Hempstead, B.L. & Petersen, C.M. (2004) Sortilin is essential for proNGF-induced neuronal cell death. Nature, 427, 843-848. Obeidat, A.S., Jarvis, C.R. & Andrew, R.D. (2000) Glutamate does not mediate acute neuronal damage after spreading depression induced by O2/glucose deprivation in the hippocampal slice. Journal of cerebral blood flow and metabolism : official journal of the International Society of Cerebral Blood Flow and Metabolism, 20, 412-422. Ohba, S., Ikeda, T., Ikegaya, Y., Nishiyama, N., Matsuki, N. & Yamada, M.K. (2005) BDNF locally potentiates GABAergic presynaptic machineries: target-selective circuit inhibition. Cerebral cortex, 15, 291-298. Ohira, K., Kumanogoh, H., Sahara, Y., Homma, K.J., Hirai, H., Nakamura, S. & Hayashi, M. (2005) A truncated tropomyosin-related kinase B receptor, T1, regulates glial cell morphology via Rho GDP dissociation inhibitor 1. The Journal of neuroscience : the official journal of the Society for Neuroscience, 25, 1343-1353. Ohno-Shosaku, T., Maejima, T. & Kano, M. (2001) Endogenous cannabinoids mediate retrograde signals from depolarized postsynaptic neurons to presynaptic terminals. Neuron, 29, 729- 738. Okabe, S., Miwa, A. & Okado, H. (2001) Spine formation and correlated assembly of presynaptic and postsynaptic molecules. The Journal of neuroscience : the official journal of the Society for Neuroscience, 21, 6105-6114. Okamoto, K., Nagai, T., Miyawaki, A. & Hayashi, Y. (2004) Rapid and persistent modulation of actin dynamics regulates postsynaptic reorganization underlying bidirectional plasticity. Nature neuroscience, 7, 1104-1112. Okazaki, M.M., Molnar, P. & Nadler, J.V. (1999) Recurrent mossy fiber pathway in rat dentate gyrus: synaptic currents evoked in presence and absence of seizure-induced growth. Journal of neurophysiology, 81, 1645-1660. Olofsdotter, K., Lindvall, O. & Asztely, F. (2000) Increased synaptic inhibition in dentate gyrus of mice with reduced levels of endogenous brain-derived neurotrophic factor. Neuroscience, 101, 531-539.

231

Ostroff, L.E., Fiala, J.C., Allwardt, B. & Harris, K.M. (2002) Polyribosomes redistribute from dendritic shafts into spines with enlarged synapses during LTP in developing rat hippocampal slices. Neuron, 35, 535-545. Oyesiku, N.M., Evans, C.O., Houston, S., Darrell, R.S., Smith, J.S., Fulop, Z.L., Dixon, C.E. & Stein, D.G. (1999) Regional changes in the expression of neurotrophic factors and their receptors following acute traumatic brain injury in the adult rat brain. Brain research, 833, 161-172. Palizvan, M.R., Sohya, K., Kohara, K., Maruyama, A., Yasuda, H., Kimura, F. & Tsumoto, T. (2004) Brain-derived neurotrophic factor increases inhibitory synapses, revealed in solitary neurons cultured from rat visual cortex. Neuroscience, 126, 955-966. Pang, P.T., Teng, H.K., Zaitsev, E., Woo, N.T., Sakata, K., Zhen, S., Teng, K.K., Yung, W.H., Hempstead, B.L. & Lu, B. (2004) Cleavage of proBDNF by tPA/plasmin is essential for long-term hippocampal plasticity. Science, 306, 487-491. Paoletti, P., Bellone, C. & Zhou, Q. (2013) NMDA receptor subunit diversity: impact on receptor properties, synaptic plasticity and disease. Nature reviews. Neuroscience, 14, 383-400. Papa, M., Bundman, M.C., Greenberger, V. & Segal, M. (1995) Morphological analysis of dendritic spine development in primary cultures of hippocampal neurons. The Journal of neuroscience : the official journal of the Society for Neuroscience, 15, 1-11. Park, H. & Poo, M.M. (2013) Neurotrophin regulation of neural circuit development and function. Nature reviews. Neuroscience, 14, 7-23. Park, J.S., Bateman, M.C. & Goldberg, M.P. (1996) Rapid alterations in dendrite morphology during sublethal hypoxia or glutamate receptor activation. Neurobiology of disease, 3, 215- 227. Parpura, V. & Zorec, R. (2010) Gliotransmission: Exocytotic release from astrocytes. Brain research reviews, 63, 83-92. Parra, P., Gulyas, A.I. & Miles, R. (1998) How many subtypes of inhibitory cells in the hippocampus? Neuron, 20, 983-993. Patterson, S.L., Abel, T., Deuel, T.A., Martin, K.C., Rose, J.C. & Kandel, E.R. (1996) Recombinant BDNF rescues deficits in basal synaptic transmission and hippocampal LTP in BDNF knockout mice. Neuron, 16, 1137-1145. Patterson, S.L., Grover, L.M., Schwartzkroin, P.A. & Bothwell, M. (1992) Neurotrophin expression in rat hippocampal slices: a stimulus paradigm inducing LTP in CA1 evokes increases in BDNF and NT-3 mRNAs. Neuron, 9, 1081-1088. Patterson, S.L., Pittenger, C., Morozov, A., Martin, K.C., Scanlin, H., Drake, C. & Kandel, E.R. (2001) Some forms of cAMP-mediated long-lasting potentiation are associated with release of BDNF and nuclear translocation of phospho-MAP kinase. Neuron, 32, 123-140. Paul, J., Gottmann, K. & Lessmann, V. (2001) NT-3 regulates BDNF-induced modulation of synaptic transmission in cultured hippocampal neurons. Neuroreport, 12, 2635-2639. Payne, J.A. (1997) Functional characterization of the neuronal-specific K-Cl cotransporter: implications for [K+]o regulation. The American journal of physiology, 273, C1516-1525.

232

Peineau, S., Taghibiglou, C., Bradley, C., Wong, T.P., Liu, L., Lu, J., Lo, E., Wu, D., Saule, E., Bouschet, T., Matthews, P., Isaac, J.T., Bortolotto, Z.A., Wang, Y.T. & Collingridge, G.L. (2007) LTP inhibits LTD in the hippocampus via regulation of GSK3beta. Neuron, 53, 703-717. Perez-Alvarez, A., Navarrete, M., Covelo, A., Martin, E.D. & Araque, A. (2014) Structural and functional plasticity of astrocyte processes and dendritic spine interactions. The Journal of neuroscience : the official journal of the Society for Neuroscience, 34, 12738-12744. Petrini, E.M., Ravasenga, T., Hausrat, T.J., Iurilli, G., Olcese, U., Racine, V., Sibarita, J.B., Jacob, T.C., Moss, S.J., Benfenati, F., Medini, P., Kneussel, M. & Barberis, A. (2014) Synaptic recruitment of gephyrin regulates surface GABAA receptor dynamics for the expression of inhibitory LTP. Nature communications, 5, 3921. Phillips, H.S., Hains, J.M., Laramee, G.R., Rosenthal, A. & Winslow, J.W. (1990) Widespread expression of BDNF but not NT3 by target areas of basal forebrain cholinergic neurons. Science, 250, 290-294. Pinel, J.P. & Rovner, L.I. (1978) Experimental epileptogenesis: kindling-induced epilepsy in rats. Experimental neurology, 58, 190-202. Pitkanen, A. & Lukasiuk, K. (2011) Mechanisms of epileptogenesis and potential treatment targets. The Lancet. Neurology, 10, 173-186. Pokorny, J. & Trojan, S. (1983) Chronic changes in the receptive field of the pyramidal cells of the rat hippocampus after intermittent postnatal hypoxia. Physiologia Bohemoslovaca, 32, 393-402. Polleux, F. & Snider, W. (2010) Initiating and growing an axon. Cold Spring Harbor perspectives in biology, 2, a001925. Poo, M.M. (2001) Neurotrophins as synaptic modulators. Nature reviews. Neuroscience, 2, 24-32. Poulopoulos, A., Aramuni, G., Meyer, G., Soykan, T., Hoon, M., Papadopoulos, T., Zhang, M., Paarmann, I., Fuchs, C., Harvey, K., Jedlicka, P., Schwarzacher, S.W., Betz, H., Harvey, R.J., Brose, N., Zhang, W. & Varoqueaux, F. (2009) Neuroligin 2 drives postsynaptic assembly at perisomatic inhibitory synapses through gephyrin and collybistin. Neuron, 63, 628-642. Prince, D.A., Parada, I., Scalise, K., Graber, K., Jin, X. & Shen, F. (2009) Epilepsy following cortical injury: cellular and molecular mechanisms as targets for potential prophylaxis. Epilepsia, 50 Suppl 2, 30-40. Prince, D.A.P., I.; Graber, K. (2012) Traumatic Brain Injury and Posttraumatic Epilepsy Jasper's Basic Mechanisms of the Epilepsies [Internet]. National Center for Biotechnology Information (US), Bethesda, MD. Pruunsild, P., Sepp, M., Orav, E., Koppel, I. & Timmusk, T. (2011) Identification of cis-elements and transcription factors regulating neuronal activity-dependent transcription of human BDNF gene. The Journal of neuroscience : the official journal of the Society for Neuroscience, 31, 3295-3308. Pulsinelli, W.A. (1985) Selective neuronal vulnerability: morphological and molecular characteristics. Progress in brain research, 63, 29-37.

233

Pulsinelli, W.A. & Brierley, J.B. (1979) A new model of bilateral hemispheric ischemia in the unanesthetized rat. Stroke; a journal of cerebral circulation, 10, 267-272. Qian, X., Riccio, A., Zhang, Y. & Ginty, D.D. (1998) Identification and characterization of novel substrates of Trk receptors in developing neurons. Neuron, 21, 1017-1029. Qin, L., Jing, D., Parauda, S., Carmel, J., Ratan, R.R., Lee, F.S. & Cho, S. (2014) An adaptive role for BDNF Val66Met polymorphism in motor recovery in chronic stroke. The Journal of neuroscience : the official journal of the Society for Neuroscience, 34, 2493-2502. Qureshi, I.A. & Mehler, M.F. (2010) Epigenetic mechanisms underlying human epileptic disorders and the process of epileptogenesis. Neurobiology of disease, 39, 53-60. Radeke, M.J., Misko, T.P., Hsu, C., Herzenberg, L.A. & Shooter, E.M. (1987) Gene transfer and molecular cloning of the rat nerve growth factor receptor. Nature, 325, 593-597. Radpour, S. & Thomson, A.M. (1991) Coactivation of Local Circuit NMDA Receptor Mediated epsps Induces Lasting Enhancement of Minimal Schaffer Collateral epsps in Slices of Rat Hippocampus. The European journal of neuroscience, 3, 602-613. Rakhade, S.N. & Jensen, F.E. (2009) Epileptogenesis in the immature brain: emerging mechanisms. Nature reviews. Neurology, 5, 380-391. Ramakers, G.M., McNamara, R.K., Lenox, R.H. & De Graan, P.N. (1999) Differential changes in the phosphorylation of the protein kinase C substrates myristoylated alanine-rich C kinase substrate and growth-associated protein-43/B-50 following Schaffer collateral long-term potentiation and long-term depression. Journal of neurochemistry, 73, 2175-2183. Rattiner, L.M., Davis, M., French, C.T. & Ressler, K.J. (2004) Brain-derived neurotrophic factor and tyrosine kinase receptor B involvement in amygdala-dependent fear conditioning. The Journal of neuroscience : the official journal of the Society for Neuroscience, 24, 4796- 4806. Reichardt, L.F. (2006) Neurotrophin-regulated signalling pathways. Philosophical transactions of the Royal Society of London. Series B, Biological sciences, 361, 1545-1564. Richards, D.A., De Paola, V., Caroni, P., Gahwiler, B.H. & McKinney, R.A. (2004) AMPA- receptor activation regulates the diffusion of a membrane marker in parallel with dendritic spine motility in the mouse hippocampus. The Journal of physiology, 558, 503-512. Rico, B., Xu, B. & Reichardt, L.F. (2002) TrkB receptor signaling is required for establishment of GABAergic synapses in the cerebellum. Nature neuroscience, 5, 225-233. Rivera, C., Voipio, J. & Kaila, K. (2005) Two developmental switches in GABAergic signalling: the K+-Cl- cotransporter KCC2 and carbonic anhydrase CAVII. The Journal of physiology, 562, 27-36. Rivera, C., Voipio, J., Payne, J.A., Ruusuvuori, E., Lahtinen, H., Lamsa, K., Pirvola, U., Saarma, M. & Kaila, K. (1999) The K+/Cl- co-transporter KCC2 renders GABA hyperpolarizing during neuronal maturation. Nature, 397, 251-255. Rocamora, N., Welker, E., Pascual, M. & Soriano, E. (1996) Upregulation of BDNF mRNA expression in the barrel cortex of adult mice after sensory stimulation. The Journal of neuroscience : the official journal of the Society for Neuroscience, 16, 4411-4419.

234

Roding, J., Glader, E.L., Malm, J., Eriksson, M. & Lindstrom, B. (2009) Perceived impaired physical and cognitive functions after stroke in men and women between 18 and 55 years of age--a national survey. Disability and rehabilitation, 31, 1092-1099. Rodrigues, E., Oliveira, C., Cambrussi, A., Gomes, G., Godoy, D., Brondani da Rocha, A., Grivicich, I. & Regner, A. (2008) Increased serum brain derived neurotrophic factor (BDNF) following isolated severe traumatic brain injury in humans. Brain Injury, 22, 165. Rodriguez-Tebar, A., Dechant, G. & Barde, Y.A. (1990) Binding of brain-derived neurotrophic factor to the nerve growth factor receptor. Neuron, 4, 487-492. Rodriguez-Tebar, A., Dechant, G., Gotz, R. & Barde, Y.A. (1992) Binding of neurotrophin-3 to its neuronal receptors and interactions with nerve growth factor and brain-derived neurotrophic factor. The EMBO journal, 11, 917-922. Routbort, M.J., Bausch, S.B. & McNamara, J.O. (1999) Seizures, cell death, and mossy fiber sprouting in kainic acid-treated organotypic hippocampal cultures. Neuroscience, 94, 755- 765. Royo, N.C., LeBold, D., Magge, S.N., Chen, I., Hauspurg, A., Cohen, A.S. & Watson, D.J. (2007) Neurotrophin-mediated neuroprotection of hippocampal neurons following traumatic brain injury is not associated with acute recovery of hippocampal function. Neuroscience, 148, 359-370. Rutherford, L.C., Nelson, S.B. & Turrigiano, G.G. (1998) BDNF has opposite effects on the quantal amplitude of pyramidal neuron and interneuron excitatory synapses. Neuron, 21, 521-530. Rutlin, M. & Nelson, S.B. (2011) MeCP2: phosphorylated locally, acting globally. Neuron, 72, 3- 5. Rybakowski, J.K. (2008) BDNF gene: functional Val66Met polymorphism in mood disorders and schizophrenia. Pharmacogenomics, 9, 1589-1593. Rytter, A., Cronberg, T., Asztely, F., Nemali, S. & Wieloch, T. (2003) Mouse hippocampal organotypic tissue cultures exposed to in vitro "ischemia" show selective and delayed CA1 damage that is aggravated by glucose. Journal of cerebral blood flow and metabolism : official journal of the International Society of Cerebral Blood Flow and Metabolism, 23, 23-33. Sala, C. & Segal, M. (2014) Dendritic spines: the locus of structural and functional plasticity. Physiological reviews, 94, 141-188. Salehi, A.H., Roux, P.P., Kubu, C.J., Zeindler, C., Bhakar, A., Tannis, L.L., Verdi, J.M. & Barker, P.A. (2000) NRAGE, a novel MAGE protein, interacts with the p75 neurotrophin receptor and facilitates nerve growth factor-dependent apoptosis. Neuron, 27, 279-288. Salin, P., Tseng, G.F., Hoffman, S., Parada, I. & Prince, D.A. (1995) Axonal sprouting in layer V pyramidal neurons of chronically injured cerebral cortex. The Journal of neuroscience : the official journal of the Society for Neuroscience, 15, 8234-8245. Sanes, J.R. & Lichtman, J.W. (1999) Can molecules explain long-term potentiation? Nature neuroscience, 2, 597-604.

235

Santhakumar, V., Ratzliff, A.D., Jeng, J., Toth, Z. & Soltesz, I. (2001) Long-term hyperexcitability in the hippocampus after experimental head trauma. Annals of neurology, 50, 708-717. Santhanam, A.V., Smith, L.A. & Katusic, Z.S. (2010) Brain-derived neurotrophic factor stimulates production of prostacyclin in cerebral arteries. Stroke; a journal of cerebral circulation, 41, 350-356. Santos, C.O., Caeiro, L., Ferro, J.M. & Figueira, M.L. (2011) Mania and stroke: a systematic review. Cerebrovascular diseases, 32, 11-21. Sasaki, Y., Ohsawa, K., Kanazawa, H., Kohsaka, S. & Imai, Y. (2001) Iba1 is an actin-cross- linking protein in macrophages/microglia. Biochemical and biophysical research communications, 286, 292-297. Sassoe-Pognetto, M., Panzanelli, P., Sieghart, W. & Fritschy, J.M. (2000) Colocalization of multiple GABA(A) receptor subtypes with gephyrin at postsynaptic sites. The Journal of comparative neurology, 420, 481-498. Sato, Y., Chin, Y., Kato, T., Tanaka, Y., Tozuka, Y., Mase, M., Ageyama, N., Ono, F., Terao, K., Yoshikawa, Y. & Hisatsune, T. (2009) White matter activated glial cells produce BDNF in a stroke model of monkeys. Neuroscience research, 65, 71-78. Saver, J.L., Starkman, S., Eckstein, M., Stratton, S.J., Pratt, F.D., Hamilton, S., Conwit, R., Liebeskind, D.S., Sung, G., Kramer, I., Moreau, G., Goldweber, R., Sanossian, N., Investigators, F.-M. & Coordinators (2015) Prehospital use of magnesium sulfate as neuroprotection in acute stroke. The New England journal of medicine, 372, 528-536. Sawada, M., Kondo, N., Suzumura, A. & Marunouchi, T. (1989) Production of tumor necrosis factor-alpha by microglia and astrocytes in culture. Brain research, 491, 394-397. Schabitz, W.R., Steigleder, T., Cooper-Kuhn, C.M., Schwab, S., Sommer, C., Schneider, A. & Kuhn, H.G. (2007) Intravenous brain-derived neurotrophic factor enhances poststroke sensorimotor recovery and stimulates neurogenesis. Stroke; a journal of cerebral circulation, 38, 2165-2172. Schaffer, K. (1892) Beitrag zur Histologie der Ammonshornformation. Archiv f. mikrosk. Anatomie, 39, 611-632. Scharfman, H.E. (2005) Brain-derived neurotrophic factor and epilepsy--a missing link? Epilepsy currents / American Epilepsy Society, 5, 83-88. Scharfman, H.E. (2013) Cutting through the complexity: the role of brain-derived neurotrophic factor in post-traumatic epilepsy (Commentary on Gill et al.). The European journal of neuroscience, 38, 3552-3553. Scharfman, H.E., Goodman, J.H., Sollas, A.L. & Croll, S.D. (2002) Spontaneous limbic seizures after intrahippocampal infusion of brain-derived neurotrophic factor. Experimental neurology, 174, 201-214. Scheibel, M.E., Crandall, P.H. & Scheibel, A.B. (1974) The hippocampal-dentate complex in temporal lobe epilepsy. A Golgi study. Epilepsia, 15, 55-80.

236

Scheiffele, P., Fan, J., Choih, J., Fetter, R. & Serafini, T. (2000) Neuroligin expressed in nonneuronal cells triggers presynaptic development in contacting axons. Cell, 101, 657- 669. Schmidt-Kastner, R. & Freund, T.F. (1991) Selective vulnerability of the hippocampus in brain ischemia. Neuroscience, 40, 599-636. Schuman, E.M. (1999) Neurotrophin regulation of synaptic transmission. Current opinion in neurobiology, 9, 105-109. Schweigreiter, R., Walmsley, A.R., Niederost, B., Zimmermann, D.R., Oertle, T., Casademunt, E., Frentzel, S., Dechant, G., Mir, A. & Bandtlow, C.E. (2004) Versican V2 and the central inhibitory domain of Nogo-A inhibit neurite growth via p75NTR/NgR-independent pathways that converge at RhoA. Molecular and cellular neurosciences, 27, 163-174. Schwyzer, L., Mateos, J.M., Abegg, M., Rietschin, L., Heeb, L., Thompson, S.M., Luthi, A., Gahwiler, B.H. & McKinney, R.A. (2002) Physiological and morphological plasticity induced by chronic treatment with NT-3 or NT-4/5 in hippocampal slice cultures. The European journal of neuroscience, 16, 1939-1948. Scoville, W.B. & Milner, B. (1957) Loss of recent memory after bilateral hippocampal lesions. Journal of neurology, neurosurgery, and psychiatry, 20, 11-21. Segal, M. (2005) Dendritic spines and long-term plasticity. Nature reviews. Neuroscience, 6, 277- 284. Selkoe, D.J. (2002) Alzheimer's disease is a synaptic failure. Science, 298, 789-791. Sette, G., Baron, J.C., Young, A.R., Miyazawa, H., Tillet, I., Barre, L., Travere, J.M., Derlon, J.M. & MacKenzie, E.T. (1993) In vivo mapping of brain benzodiazepine receptor changes by positron emission tomography after focal ischemia in the anesthetized baboon. Stroke; a journal of cerebral circulation, 24, 2046-2057; discussion 2057-2048. Shelly, M., Cancedda, L., Heilshorn, S., Sumbre, G. & Poo, M.M. (2007) LKB1/STRAD promotes axon initiation during neuronal polarization. Cell, 129, 565-577. Shelly, M., Cancedda, L., Lim, B.K., Popescu, A.T., Cheng, P.L., Gao, H. & Poo, M.M. (2011) Semaphorin3A regulates neuronal polarization by suppressing axon formation and promoting dendrite growth. Neuron, 71, 433-446. Shelton, D.L., Sutherland, J., Gripp, J., Camerato, T., Armanini, M.P., Phillips, H.S., Carroll, K., Spencer, S.D. & Levinson, A.D. (1995) Human trks: molecular cloning, tissue distribution, and expression of extracellular domain immunoadhesins. The Journal of neuroscience : the official journal of the Society for Neuroscience, 15, 477-491. Shen, W., Wu, B., Zhang, Z., Dou, Y., Rao, Z.R., Chen, Y.R. & Duan, S. (2006) Activity-induced rapid synaptic maturation mediated by presynaptic cdc42 signaling. Neuron, 50, 401-414. Sheng, M. (2001) Molecular organization of the postsynaptic specialization. Proceedings of the National Academy of Sciences of the United States of America, 98, 7058-7061. Shi, S.H., Hayashi, Y., Petralia, R.S., Zaman, S.H., Wenthold, R.J., Svoboda, K. & Malinow, R. (1999) Rapid spine delivery and redistribution of AMPA receptors after synaptic NMDA receptor activation. Science, 284, 1811-1816.

237

Shulga, A., Thomas-Crusells, J., Sigl, T., Blaesse, A., Mestres, P., Meyer, M., Yan, Q., Kaila, K., Saarma, M., Rivera, C. & Giehl, K.M. (2008) Posttraumatic GABA(A)-mediated [Ca2+]i increase is essential for the induction of brain-derived neurotrophic factor-dependent survival of mature central neurons. The Journal of neuroscience : the official journal of the Society for Neuroscience, 28, 6996-7005. Sieghart, W. (1995) Structure and pharmacology of gamma-aminobutyric acidA receptor subtypes. Pharmacological reviews, 47, 181-234. Siesjo, B.K. (1981) Cell damage in the brain: a speculative synthesis. Journal of cerebral blood flow and metabolism : official journal of the International Society of Cerebral Blood Flow and Metabolism, 1, 155-185. Sik, A., Penttonen, M., Ylinen, A. & Buzsaki, G. (1995) Hippocampal CA1 interneurons: an in vivo intracellular labeling study. The Journal of neuroscience : the official journal of the Society for Neuroscience, 15, 6651-6665. Silverman, I.E., Restrepo, L. & Mathews, G.C. (2002) Poststroke seizures. Archives of neurology, 59, 195-201. Simon, R.P., Swan, J.H., Griffiths, T. & Meldrum, B.S. (1984) Blockade of N-methyl-D-aspartate receptors may protect against ischemic damage in the brain. Science, 226, 850-852. Singh, B., Henneberger, C., Betances, D., Arevalo, M.A., Rodriguez-Tebar, A., Meier, J.C. & Grantyn, R. (2006) Altered balance of glutamatergic/GABAergic synaptic input and associated changes in dendrite morphology after BDNF expression in BDNF-deficient hippocampal neurons. The Journal of neuroscience : the official journal of the Society for Neuroscience, 26, 7189-7200. Skene, P.J., Illingworth, R.S., Webb, S., Kerr, A.R., James, K.D., Turner, D.J., Andrews, R. & Bird, A.P. (2010) Neuronal MeCP2 is expressed at near histone-octamer levels and globally alters the chromatin state. Molecular cell, 37, 457-468. Sklar, P., Gabriel, S.B., McInnis, M.G., Bennett, P., Lim, Y., Tsan, G., Schaffner, S., Kirov, G., Jones, I., Owen, M., Craddock, N., DePaulo, J.R. & Lander, E.S. (2002) Family-based association study of 76 candidate genes in bipolar disorder: BDNF is a potential risk locus. Brain-derived neutrophic factor. Molecular psychiatry, 7, 579-593. Smith, M.L., Auer, R.N. & Siesjo, B.K. (1984) The density and distribution of ischemic brain injury in the rat following 2-10 min of forebrain ischemia. Acta neuropathologica, 64, 319- 332. Sommer, W. (1880) Erkrankung des Ammonshorns als aetiologisches Moment der Epilepsie. Archiv für Psychiatrie und Nervenkrankheiten, 10, 631-675. Son, H., Davis, P.J. & Carpenter, D.O. (1997) Time course and involvement of protein kinase C- mediated phosphorylation of F1/GAP-43 in area CA3 after mossy fiber stimulation. Cellular and molecular neurobiology, 17, 171-194. Soppet, D., Escandon, E., Maragos, J., Middlemas, D.S., Reid, S.W., Blair, J., Burton, L.E., Stanton, B.R., Kaplan, D.R., Hunter, T., Nikolics, K. & Parada, L.F. (1991) The neurotrophic factors brain-derived neurotrophic factor and neurotrophin-3 are ligands for the trkB tyrosine kinase receptor. Cell, 65, 895-903.

238

Spruston, N. & McBain, C.J. (2007) Structural and Functional Properties of Hippocampal Neurons. In Andersen, P., Morris, R., Amaral, D.G., Bliss, T., O'Keefe, J. (eds) The Hippocampus Book. Oxford Press, New York City, NY, pp. 133-201. Stadelmann, C., Kerschensteiner, M., Misgeld, T., Bruck, W., Hohlfeld, R. & Lassmann, H. (2002) BDNF and gp145trkB in multiple sclerosis brain lesions: neuroprotective interactions between immune and neuronal cells? Brain : a journal of neurology, 125, 75-85. Stephens, R.M., Loeb, D.M., Copeland, T.D., Pawson, T., Greene, L.A. & Kaplan, D.R. (1994) Trk receptors use redundant signal transduction pathways involving SHC and PLC-gamma 1 to mediate NGF responses. Neuron, 12, 691-705. Stewart, F.M., Sunderland, A. & Sluman, S.M. (1996) The nature and prevalence of memory disorder late after stroke. The British journal of clinical psychology / the British Psychological Society, 35 ( Pt 3), 369-379. Strong, K., Mathers, C. & Bonita, R. (2007) Preventing stroke: saving lives around the world. The Lancet. Neurology, 6, 182-187. Sun, Y., Lim, Y., Li, F., Liu, S., Lu, J.J., Haberberger, R., Zhong, J.H. & Zhou, X.F. (2012) ProBDNF collapses neurite outgrowth of primary neurons by activating RhoA. PloS one, 7, e35883. Sutula, T., Cascino, G., Cavazos, J., Parada, I. & Ramirez, L. (1989) Mossy fiber synaptic reorganization in the epileptic human temporal lobe. Annals of neurology, 26, 321-330. Sutula, T.P. & Dudek, F.E. (2007) Unmasking recurrent excitation generated by mossy fiber sprouting in the epileptic dentate gyrus: an emergent property of a complex system. Progress in brain research, 163, 541-563. Suzuki, T., Du, F., Tian, Q.B., Zhang, J. & Endo, S. (2008) Ca2+/calmodulin-dependent protein kinase IIalpha clusters are associated with stable lipid rafts and their formation traps PSD- 95. Journal of neurochemistry, 104, 596-610. Suzumura, A., Marunouchi, T. & Yamamoto, H. (1991) Morphological transformation of microglia in vitro. Brain research, 545, 301-306. Svoboda, K., Tank, D.W. & Denk, W. (1996) Direct measurement of coupling between dendritic spines and shafts. Science, 272, 716-719. Swann, J.W., Al-Noori, S., Jiang, M. & Lee, C.L. (2000) Spine loss and other dendritic abnormalities in epilepsy. Hippocampus, 10, 617-625. Szirmai, I., Buzsaki, G. & Kamondi, A. (2012) 120 years of hippocampal Schaffer collaterals. Hippocampus, 22, 1508-1516. Takahashi, M., Hayashi, S., Kakita, A., Wakabayashi, K., Fukuda, M., Kameyama, S., Tanaka, R., Takahashi, H. & Nawa, H. (1999) Patients with temporal lobe epilepsy show an increase in brain-derived neurotrophic factor protein and its correlation with neuropeptide Y. Brain research, 818, 579-582. Takeda, A., Onodera, H., Sugimoto, A., Kogure, K., Obinata, M. & Shibahara, S. (1993) Coordinated expression of messenger RNAs for nerve growth factor, brain-derived

239

neurotrophic factor and neurotrophin-3 in the rat hippocampus following transient forebrain ischemia. Neuroscience, 55, 23-31. Takumi, Y., Ramirez-Leon, V., Laake, P., Rinvik, E. & Ottersen, O.P. (1999) Different modes of expression of AMPA and NMDA receptors in hippocampal synapses. Nature neuroscience, 2, 618-624. Tan, A.M., Chang, Y.W., Zhao, P., Hains, B.C. & Waxman, S.G. (2011) Rac1-regulated dendritic spine remodeling contributes to neuropathic pain after peripheral nerve injury. Experimental neurology, 232, 222-233. Tanaka, J., Horiike, Y., Matsuzaki, M., Miyazaki, T., Ellis-Davies, G.C. & Kasai, H. (2008) Protein synthesis and neurotrophin-dependent structural plasticity of single dendritic spines. Science, 319, 1683-1687. Tanaka, T., Saito, H. & Matsuki, N. (1997) Inhibition of GABAA synaptic responses by brain- derived neurotrophic factor (BDNF) in rat hippocampus. The Journal of neuroscience : the official journal of the Society for Neuroscience, 17, 2959-2966. Tao, J., Hu, K., Chang, Q., Wu, H., Sherman, N.E., Martinowich, K., Klose, R.J., Schanen, C., Jaenisch, R., Wang, W. & Sun, Y.E. (2009) Phosphorylation of MeCP2 at Serine 80 regulates its chromatin association and neurological function. Proceedings of the National Academy of Sciences of the United States of America, 106, 4882-4887. Tao, X., West, A.E., Chen, W.G., Corfas, G. & Greenberg, M.E. (2002) A calcium-responsive transcription factor, CaRF, that regulates neuronal activity-dependent expression of BDNF. Neuron, 33, 383-395. Tashiro, A., Minden, A. & Yuste, R. (2000) Regulation of dendritic spine morphology by the rho family of small GTPases: antagonistic roles of Rac and Rho. Cerebral cortex, 10, 927-938. Tashiro, A. & Yuste, R. (2004) Regulation of dendritic spine motility and stability by Rac1 and Rho kinase: evidence for two forms of spine motility. Molecular and cellular neurosciences, 26, 429-440. Tauck, D.L. & Nadler, J.V. (1985) Evidence of functional mossy fiber sprouting in hippocampal formation of kainic acid-treated rats. The Journal of neuroscience : the official journal of the Society for Neuroscience, 5, 1016-1022. Tellez-Zenteno, J.F., Patten, S.B., Jette, N., Williams, J. & Wiebe, S. (2007) Psychiatric comorbidity in epilepsy: a population-based analysis. Epilepsia, 48, 2336-2344. Temkin, N.R. (2001) Antiepileptogenesis and seizure prevention trials with antiepileptic drugs: meta-analysis of controlled trials. Epilepsia, 42, 515-524. Temkin, N.R. (2009) Preventing and treating posttraumatic seizures: the human experience. Epilepsia, 50 Suppl 2, 10-13. Temkin, N.R., Dikmen, S.S., Anderson, G.D., Wilensky, A.J., Holmes, M.D., Cohen, W., Newell, D.W., Nelson, P., Awan, A. & Winn, H.R. (1999) Valproate therapy for prevention of posttraumatic seizures: a randomized trial. Journal of neurosurgery, 91, 593-600.

240

Temkin, N.R., Dikmen, S.S., Wilensky, A.J., Keihm, J., Chabal, S. & Winn, H.R. (1990) A randomized, double-blind study of phenytoin for the prevention of post-traumatic seizures. The New England journal of medicine, 323, 497-502. Teng, H.K., Teng, K.K., Lee, R., Wright, S., Tevar, S., Almeida, R.D., Kermani, P., Torkin, R., Chen, Z.Y., Lee, F.S., Kraemer, R.T., Nykjaer, A. & Hempstead, B.L. (2005) ProBDNF induces neuronal apoptosis via activation of a receptor complex of p75NTR and sortilin. The Journal of neuroscience : the official journal of the Society for Neuroscience, 25, 5455- 5463. Thakker-Varia, S., Alder, J., Crozier, R.A., Plummer, M.R. & Black, I.B. (2001) Rab3A is required for brain-derived neurotrophic factor-induced synaptic plasticity: transcriptional analysis at the population and single-cell levels. The Journal of neuroscience : the official journal of the Society for Neuroscience, 21, 6782-6790. Thompson, S.M. & Gahwiler, B.H. (1989) Activity-dependent disinhibition. II. Effects of extracellular potassium, furosemide, and membrane potential on ECl- in hippocampal CA3 neurons. Journal of neurophysiology, 61, 512-523. Timmusk, T., Palm, K., Metsis, M., Reintam, T., Paalme, V., Saarma, M. & Persson, H. (1993) Multiple promoters direct tissue-specific expression of the rat BDNF gene. Neuron, 10, 475-489. Tolias, K.F., Bikoff, J.B., Burette, A., Paradis, S., Harrar, D., Tavazoie, S., Weinberg, R.J. & Greenberg, M.E. (2005) The Rac1-GEF Tiam1 couples the NMDA receptor to the activity- dependent development of dendritic arbors and spines. Neuron, 45, 525-538. Tonchev, A.B., Boneva, N.B., Kaplamadzhiev, D.B., Kikuchi, M., Mori, Y., Sahara, S. & Yamashima, T. (2008) Expression of neurotrophin receptors by proliferating glia in postischemic hippocampal CA1 sector of adult monkeys. Journal of neuroimmunology, 205, 20-24. Tonnesen, J., Katona, G., Rozsa, B. & Nagerl, U.V. (2014) Spine neck plasticity regulates compartmentalization of synapses. Nature neuroscience, 17, 678-685. Trang, T., Beggs, S. & Salter, M.W. (2011) Brain-derived neurotrophic factor from microglia: a molecular substrate for neuropathic pain. Neuron glia biology, 7, 99-108. Tretter, V., Jacob, T.C., Mukherjee, J., Fritschy, J.M., Pangalos, M.N. & Moss, S.J. (2008) The clustering of GABA(A) receptor subtypes at inhibitory synapses is facilitated via the direct binding of receptor alpha 2 subunits to gephyrin. The Journal of neuroscience : the official journal of the Society for Neuroscience, 28, 1356-1365. Tretter, V., Kerschner, B., Milenkovic, I., Ramsden, S.L., Ramerstorfer, J., Saiepour, L., Maric, H.M., Moss, S.J., Schindelin, H., Harvey, R.J., Sieghart, W. & Harvey, K. (2011) Molecular basis of the gamma-aminobutyric acid A receptor alpha3 subunit interaction with the clustering protein gephyrin. The Journal of biological chemistry, 286, 37702- 37711. Trinka, E., Hofler, J. & Zerbs, A. (2012) Causes of status epilepticus. Epilepsia, 53 Suppl 4, 127- 138.

241

Trommald, M., Hulleberg, G. & Andersen, P. (1996) Long-term potentiation is associated with new excitatory spine synapses on rat dentate granule cells. Learning & memory, 3, 218- 228. Truettner, J., Schmidt-Kastner, R., Busto, R., Alonso, O.F., Loor, J.Y., Dietrich, W.D. & Ginsberg, M.D. (1999) Expression of brain-derived neurotrophic factor, nerve growth factor, and heat shock protein HSP70 following fluid percussion brain injury in rats. Journal of neurotrauma, 16, 471-486. Tsai, S.J. (2004a) Down-regulation of the Trk-B signal pathway: the possible pathogenesis of major depression. Medical hypotheses, 62, 215-218. Tsai, S.J. (2004b) Is mania caused by overactivity of central brain-derived neurotrophic factor? Medical hypotheses, 62, 19-22. Tudor, M., Akbarian, S., Chen, R.Z. & Jaenisch, R. (2002) Transcriptional profiling of a mouse model for Rett syndrome reveals subtle transcriptional changes in the brain. Proceedings of the National Academy of Sciences of the United States of America, 99, 15536-15541. Turrigiano, G.G. & Nelson, S.B. (2004) Homeostatic plasticity in the developing nervous system. Nature reviews. Neuroscience, 5, 97-107. Tyagarajan, S.K. & Fritschy, J.M. (2014) Gephyrin: a master regulator of neuronal function? Nature reviews. Neuroscience, 15, 141-156. Tyagarajan, S.K., Ghosh, H., Yevenes, G.E., Imanishi, S.Y., Zeilhofer, H.U., Gerrits, B. & Fritschy, J.M. (2013) Extracellular signal-regulated kinase and glycogen synthase kinase 3beta regulate gephyrin postsynaptic aggregation and GABAergic synaptic function in a calpain-dependent mechanism. The Journal of biological chemistry, 288, 9634-9647. Tyagarajan, S.K., Ghosh, H., Yevenes, G.E., Nikonenko, I., Ebeling, C., Schwerdel, C., Sidler, C., Zeilhofer, H.U., Gerrits, B., Muller, D. & Fritschy, J.M. (2011) Regulation of GABAergic synapse formation and plasticity by GSK3beta-dependent phosphorylation of gephyrin. Proceedings of the National Academy of Sciences of the United States of America, 108, 379-384. Tyler, W.J. & Pozzo-Miller, L. (2003) Miniature synaptic transmission and BDNF modulate dendritic spine growth and form in rat CA1 neurones. The Journal of physiology, 553, 497- 509. Tyler, W.J. & Pozzo-Miller, L.D. (2001) BDNF enhances quantal neurotransmitter release and increases the number of docked vesicles at the active zones of hippocampal excitatory synapses. The Journal of neuroscience : the official journal of the Society for Neuroscience, 21, 4249-4258. Tyler, W.J., Zhang, X.L., Hartman, K., Winterer, J., Muller, W., Stanton, P.K. & Pozzo-Miller, L. (2006) BDNF increases release probability and the size of a rapidly recycling vesicle pool within rat hippocampal excitatory synapses. The Journal of physiology, 574, 787-803. Udo, H., Jin, I., Kim, J.H., Li, H.L., Youn, T., Hawkins, R.D., Kandel, E.R. & Bailey, C.H. (2005) Serotonin-induced regulation of the actin network for learning-related synaptic growth requires Cdc42, N-WASP, and PAK in Aplysia sensory neurons. Neuron, 45, 887-901.

242

Ulmann, L., Hatcher, J.P., Hughes, J.P., Chaumont, S., Green, P.J., Conquet, F., Buell, G.N., Reeve, A.J., Chessell, I.P. & Rassendren, F. (2008) Up-regulation of P2X4 receptors in spinal microglia after peripheral nerve injury mediates BDNF release and neuropathic pain. The Journal of neuroscience : the official journal of the Society for Neuroscience, 28, 11263-11268. Van Lookeren Campagne, M., Oestreicher, A.B., Van Bergen en Henegouwen, P.M. & Gispen, W.H. (1990) Ultrastructural double localization of B-50/GAP43 and synaptophysin (p38) in the neonatal and adult rat hippocampus. Journal of neurocytology, 19, 948-961. Van Lookeren Campagne, M., Oestreicher, A.B., Van Bergen en Henegowen, P.M. & Gispen, W.H. (1989) Ultrastructural immunocytochemical localization of B-50/GAP43, a protein kinase C substrate, in isolated presynaptic nerve terminals and neuronal growth cones. Journal of neurocytology, 18, 479-489. Vanelderen, P., Rouwette, T., Kozicz, T., Roubos, E., Van Zundert, J., Heylen, R. & Vissers, K. (2010) The role of brain-derived neurotrophic factor in different animal models of neuropathic pain. European journal of pain, 14, 473 e471-479. Vaquero, J., Oya, S., Cabezudo, J.M. & Bravo, G. (1982) Morphological study of human epileptic dendrites. Neurosurgery, 10, 720-724. Varley, Z.K., Pizzarelli, R., Antonelli, R., Stancheva, S.H., Kneussel, M., Cherubini, E. & Zacchi, P. (2011) Gephyrin regulates GABAergic and glutamatergic synaptic transmission in hippocampal cell cultures. The Journal of biological chemistry, 286, 20942-20951. Varvel, N.H., Jiang, J. & Dingledine, R. (2015) Candidate drug targets for prevention or modification of epilepsy. Annual review of pharmacology and toxicology, 55, 229-247. Verbich, D., Prenosil, G.A., Chang, P.K., Murai, K.K. & McKinney, R.A. (2012) Glial glutamate transport modulates dendritic spine head protrusions in the hippocampus. Glia, 60, 1067- 1077. Vicario-Abejon, C., Collin, C., McKay, R.D. & Segal, M. (1998) Neurotrophins induce formation of functional excitatory and inhibitory synapses between cultured hippocampal neurons. The Journal of neuroscience : the official journal of the Society for Neuroscience, 18, 7256- 7271. Vicario-Abejon, C., Owens, D., McKay, R. & Segal, M. (2002) Role of neurotrophins in central synapse formation and stabilization. Nature reviews. Neuroscience, 3, 965-974. Vithlani, M., Terunuma, M. & Moss, S.J. (2011) The dynamic modulation of GABA(A) receptor trafficking and its role in regulating the plasticity of inhibitory synapses. Physiological reviews, 91, 1009-1022. Vlachos, A., Reddy-Alla, S., Papadopoulos, T., Deller, T. & Betz, H. (2013) Homeostatic regulation of gephyrin scaffolds and synaptic strength at mature hippocampal GABAergic postsynapses. Cerebral cortex, 23, 2700-2711. Wake, H., Moorhouse, A.J., Miyamoto, A. & Nabekura, J. (2013) Microglia: actively surveying and shaping neuronal circuit structure and function. Trends in neurosciences, 36, 209-217.

243

Walz, C., Jungling, K., Lessmann, V. & Gottmann, K. (2006) Presynaptic plasticity in an immature neocortical network requires NMDA receptor activation and BDNF release. Journal of neurophysiology, 96, 3512-3516. Wang, X.B., Bozdagi, O., Nikitczuk, J.S., Zhai, Z.W., Zhou, Q. & Huntley, G.W. (2008) Extracellular proteolysis by matrix metalloproteinase-9 drives dendritic spine enlargement and long-term potentiation coordinately. Proceedings of the National Academy of Sciences of the United States of America, 105, 19520-19525. Wannamaker, B.B., Wilson, D.A., Malek, A.M. & Selassie, A.W. (2015) Stroke after adult-onset epilepsy: A population-based retrospective cohort study. Epilepsy & behavior : E&B, 43, 93-99. Ward, B., McGuinness, L., Akerman, C.J., Fine, A., Bliss, T.V. & Emptage, N.J. (2006) State- dependent mechanisms of LTP expression revealed by optical quantal analysis. Neuron, 52, 649-661. Wardle, R.A. & Poo, M.M. (2003) Brain-derived neurotrophic factor modulation of GABAergic synapses by postsynaptic regulation of chloride transport. The Journal of neuroscience : the official journal of the Society for Neuroscience, 23, 8722-8732. Wayman, G.A., Lee, Y.S., Tokumitsu, H., Silva, A.J. & Soderling, T.R. (2008) Calmodulin- kinases: modulators of neuronal development and plasticity. Neuron, 59, 914-931. Weng, S.M., Bailey, M.E. & Cobb, S.R. (2011) Rett syndrome: from bed to bench. Pediatrics and neonatology, 52, 309-316. Weskamp, G. & Reichardt, L.F. (1991) Evidence that biological activity of NGF is mediated through a novel subclass of high affinity receptors. Neuron, 6, 649-663. West, A.E., Chen, W.G., Dalva, M.B., Dolmetsch, R.E., Kornhauser, J.M., Shaywitz, A.J., Takasu, M.A., Tao, X. & Greenberg, M.E. (2001) Calcium regulation of neuronal gene expression. Proceedings of the National Academy of Sciences of the United States of America, 98, 11024-11031. Wetmore, C., Ernfors, P., Persson, H. & Olson, L. (1990) Localization of brain-derived neurotrophic factor mRNA to neurons in the brain by in situ hybridization. Experimental neurology, 109, 141-152. Wigstrom, H., Gustafsson, B., Huang, Y.Y. & Abraham, W.C. (1986) Hippocampal long-term potentiation is induced by pairing single afferent volleys with intracellularly injected depolarizing current pulses. Acta physiologica Scandinavica, 126, 317-319. Williams, P.A., Hellier, J.L., White, A.M., Staley, K.J. & Dudek, F.E. (2007) Development of spontaneous seizures after experimental status epilepticus: implications for understanding epileptogenesis. Epilepsia, 48 Suppl 5, 157-163. Wilson, R.I. & Nicoll, R.A. (2001) Endogenous cannabinoids mediate retrograde signalling at hippocampal synapses. Nature, 410, 588-592. Wong, M. & Guo, D. (2013) Dendritic spine pathology in epilepsy: cause or consequence? Neuroscience, 251, 141-150.

244

Woo, N.H., Teng, H.K., Siao, C.J., Chiaruttini, C., Pang, P.T., Milner, T.A., Hempstead, B.L. & Lu, B. (2005) Activation of p75NTR by proBDNF facilitates hippocampal long-term depression. Nature neuroscience, 8, 1069-1077. Wu, F., Catano, M., Echeverry, R., Torre, E., Haile, W.B., An, J., Chen, C., Cheng, L., Nicholson, A., Tong, F.C., Park, J. & Yepes, M. (2014) Urokinase-type plasminogen activator promotes dendritic spine recovery and improves neurological outcome following ischemic stroke. The Journal of neuroscience : the official journal of the Society for Neuroscience, 34, 14219-14232. Wu, H., Tao, J., Chen, P.J., Shahab, A., Ge, W., Hart, R.P., Ruan, X., Ruan, Y. & Sun, Y.E. (2010) Genome-wide analysis reveals methyl-CpG-binding protein 2-dependent regulation of microRNAs in a mouse model of Rett syndrome. Proceedings of the National Academy of Sciences of the United States of America, 107, 18161-18166. Xie, Y. & Longo, F.M. (2000) Neurotrophin small-molecule mimetics. Progress in brain research, 128, 333-347. Xing, C., Arai, K., Lo, E.H. & Hommel, M. (2012) Pathophysiologic cascades in ischemic stroke. International journal of stroke : official journal of the International Stroke Society, 7, 378- 385. Xu, B., Gottschalk, W., Chow, A., Wilson, R.I., Schnell, E., Zang, K., Wang, D., Nicoll, R.A., Lu, B. & Reichardt, L.F. (2000) The role of brain-derived neurotrophic factor receptors in the mature hippocampus: modulation of long-term potentiation through a presynaptic mechanism involving TrkB. The Journal of neuroscience : the official journal of the Society for Neuroscience, 20, 6888-6897. Xu, N.J. & Henkemeyer, M. (2012) Ephrin reverse signaling in axon guidance and synaptogenesis. Seminars in cell & developmental biology, 23, 58-64. Xu, Z.C. & Pulsinelli, W.A. (1994) Responses of CA1 pyramidal neurons in rat hippocampus to transient forebrain ischemia: an in vivo intracellular recording study. Neuroscience letters, 171, 187-191. Yacoubian, T.A. & Lo, D.C. (2000) Truncated and full-length TrkB receptors regulate distinct modes of dendritic growth. Nature neuroscience, 3, 342-349. Yamada, M.K., Nakanishi, K., Ohba, S., Nakamura, T., Ikegaya, Y., Nishiyama, N. & Matsuki, N. (2002) Brain-derived neurotrophic factor promotes the maturation of GABAergic mechanisms in cultured hippocampal neurons. The Journal of neuroscience : the official journal of the Society for Neuroscience, 22, 7580-7585. Yamashita, T., Tucker, K.L. & Barde, Y.A. (1999) Neurotrophin binding to the p75 receptor modulates Rho activity and axonal outgrowth. Neuron, 24, 585-593. Yan, Q., Rosenfeld, R.D., Matheson, C.R., Hawkins, N., Lopez, O.T., Bennett, L. & Welcher, A.A. (1997) Expression of brain-derived neurotrophic factor protein in the adult rat central nervous system. Neuroscience, 78, 431-448. Yang, C.H., Huang, C.C. & Hsu, K.S. (2004) Behavioral stress modifies hippocampal synaptic plasticity through corticosterone-induced sustained extracellular signal-regulated

245

kinase/mitogen-activated protein kinase activation. The Journal of neuroscience : the official journal of the Society for Neuroscience, 24, 11029-11034. Yang, J., Harte-Hargrove, L.C., Siao, C.J., Marinic, T., Clarke, R., Ma, Q., Jing, D., Lafrancois, J.J., Bath, K.G., Mark, W., Ballon, D., Lee, F.S., Scharfman, H.E. & Hempstead, B.L. (2014) proBDNF negatively regulates neuronal remodeling, synaptic transmission, and synaptic plasticity in hippocampus. Cell reports, 7, 796-806. Yang, J., Siao, C.J., Nagappan, G., Marinic, T., Jing, D., McGrath, K., Chen, Z.Y., Mark, W., Tessarollo, L., Lee, F.S., Lu, B. & Hempstead, B.L. (2009) Neuronal release of proBDNF. Nature neuroscience, 12, 113-115. Yang, L., Zhang, Z., Sun, D., Xu, Z., Yuan, Y., Zhang, X. & Li, L. (2011) Low serum BDNF may indicate the development of PSD in patients with acute ischemic stroke. International journal of geriatric psychiatry, 26, 495-502. Yang, Y., Wang, X.B., Frerking, M. & Zhou, Q. (2008) Spine expansion and stabilization associated with long-term potentiation. The Journal of neuroscience : the official journal of the Society for Neuroscience, 28, 5740-5751. Young, B., Rapp, R.P., Norton, J.A., Haack, D., Tibbs, P.A. & Bean, J.R. (1983a) Failure of prophylactically administered phenytoin to prevent early posttraumatic seizures. Journal of neurosurgery, 58, 231-235. Young, B., Rapp, R.P., Norton, J.A., Haack, D., Tibbs, P.A. & Bean, J.R. (1983b) Failure of prophylactically administered phenytoin to prevent late posttraumatic seizures. Journal of neurosurgery, 58, 236-241. Yu, S.J., Tseng, K.Y., Shen, H., Harvey, B.K., Airavaara, M. & Wang, Y. (2013) Local administration of AAV-BDNF to subventricular zone induces functional recovery in stroke rats. PloS one, 8, e81750. Yu, T.W. & Bargmann, C.I. (2001) Dynamic regulation of axon guidance. Nature neuroscience, 4 Suppl, 1169-1176. Yuan, X.B., Jin, M., Xu, X., Song, Y.Q., Wu, C.P., Poo, M.M. & Duan, S. (2003) Signalling and crosstalk of Rho GTPases in mediating axon guidance. Nature cell biology, 5, 38-45. Yuste, R. (2011) Dendritic spines and distributed circuits. Neuron, 71, 772-781. Yuste, R. & Bonhoeffer, T. (2004) Genesis of dendritic spines: insights from ultrastructural and imaging studies. Nature reviews. Neuroscience, 5, 24-34. Yuste, R., Majewska, A., Cash, S.S. & Denk, W. (1999) Mechanisms of calcium influx into hippocampal spines: heterogeneity among spines, coincidence detection by NMDA receptors, and optical quantal analysis. The Journal of neuroscience : the official journal of the Society for Neuroscience, 19, 1976-1987. Zacchi, P., Antonelli, R. & Cherubini, E. (2014) Gephyrin phosphorylation in the functional organization and plasticity of GABAergic synapses. Frontiers in cellular neuroscience, 8, 103. Zafra, F., Castren, E., Thoenen, H. & Lindholm, D. (1991) Interplay between glutamate and gamma-aminobutyric acid transmitter systems in the physiological regulation of brain-

246

derived neurotrophic factor and nerve growth factor synthesis in hippocampal neurons. Proceedings of the National Academy of Sciences of the United States of America, 88, 10037-10041. Zafra, F., Hengerer, B., Leibrock, J., Thoenen, H. & Lindholm, D. (1990) Activity dependent regulation of BDNF and NGF mRNAs in the rat hippocampus is mediated by non-NMDA glutamate receptors. The EMBO journal, 9, 3545-3550. Zagrebelsky, M., Holz, A., Dechant, G., Barde, Y.A., Bonhoeffer, T. & Korte, M. (2005) The p75 neurotrophin receptor negatively modulates dendrite complexity and spine density in hippocampal neurons. The Journal of neuroscience : the official journal of the Society for Neuroscience, 25, 9989-9999. Zakharenko, S.S., Patterson, S.L., Dragatsis, I., Zeitlin, S.O., Siegelbaum, S.A., Kandel, E.R. & Morozov, A. (2003) Presynaptic BDNF required for a presynaptic but not postsynaptic component of LTP at hippocampal CA1-CA3 synapses. Neuron, 39, 975-990. Zeinieh, M., Salehi, A., Rajkumar, V. & Barker, P.A. (2014) p75NTR-dependent Rac1 activation requires receptor cleavage and activation of an NRAGE and NEDD9 signaling cascade. Journal of cell science. Zeng, L.H., Xu, L., Rensing, N.R., Sinatra, P.M., Rothman, S.M. & Wong, M. (2007) Kainate seizures cause acute dendritic injury and actin depolymerization in vivo. The Journal of neuroscience : the official journal of the Society for Neuroscience, 27, 11604-11613. Zhai, S., Ark, E.D., Parra-Bueno, P. & Yasuda, R. (2013) Long-distance integration of nuclear ERK signaling triggered by activation of a few dendritic spines. Science, 342, 1107-1111. Zhan, R.Z., Nadler, J.V. & Schwartz-Bloom, R.D. (2006) Depressed responses to applied and synaptically-released GABA in CA1 pyramidal cells, but not in CA1 interneurons, after transient forebrain ischemia. Journal of cerebral blood flow and metabolism : official journal of the International Society of Cerebral Blood Flow and Metabolism, 26, 112-124. Zhang, H., Schools, G.P., Lei, T., Wang, W., Kimelberg, H.K. & Zhou, M. (2008) Resveratrol attenuates early pyramidal neuron excitability impairment and death in acute rat hippocampal slices caused by oxygen-glucose deprivation. Experimental neurology, 212, 44-52. Zhang, S., Boyd, J., Delaney, K. & Murphy, T.H. (2005) Rapid reversible changes in dendritic spine structure in vivo gated by the degree of ischemia. The Journal of neuroscience : the official journal of the Society for Neuroscience, 25, 5333-5338. Zhang, W. & Linden, D.J. (2003) The other side of the engram: experience-driven changes in neuronal intrinsic excitability. Nature reviews. Neuroscience, 4, 885-900. Zhou, Q., Homma, K.J. & Poo, M.M. (2004) Shrinkage of dendritic spines associated with long- term depression of hippocampal synapses. Neuron, 44, 749-757. Zhou, Z., Hong, E.J., Cohen, S., Zhao, W.N., Ho, H.Y., Schmidt, L., Chen, W.G., Lin, Y., Savner, E., Griffith, E.C., Hu, L., Steen, J.A., Weitz, C.J. & Greenberg, M.E. (2006) Brain-specific phosphorylation of MeCP2 regulates activity-dependent Bdnf transcription, dendritic growth, and spine maturation. Neuron, 52, 255-269.

247

Zita, M.M., Marchionni, I., Bottos, E., Righi, M., Del Sal, G., Cherubini, E. & Zacchi, P. (2007) Post-phosphorylation prolyl isomerisation of gephyrin represents a mechanism to modulate glycine receptors function. The EMBO journal, 26, 1761-1771. Ziv, N.E. & Smith, S.J. (1996) Evidence for a role of dendritic filopodia in synaptogenesis and spine formation. Neuron, 17, 91-102. Zola-Morgan, S., Squire, L.R. & Amaral, D.G. (1986) Human amnesia and the medial temporal region: enduring memory impairment following a bilateral lesion limited to field CA1 of the hippocampus. The Journal of neuroscience : the official journal of the Society for Neuroscience, 6, 2950-2967.

248

APPENDIX

The appendix contains the following three peer-reviewed articles:

1. Chang, P.-K., Prenosil, G.A., Verbich, D., Gill, R., and McKinney, R.A. (2014) Prolonged ampakine exposure prunes dendritic spines and increases presynaptic release probability for enhanced long-term potentiation in the hippocampus. European Journal of Neuroscience. 40(5):2766

2. Machnes, Z.M., Huang, T.C., Chang, P.-K., Gill, R., Reist, N., Dezsi, G., Ozturk, E., Charron, F., O’Brien, T.J., Jones, N.C., McKinney, R.A., and Szyf, M. (2013) DNA methylation mediates persistent epileptiform activity in vitro and in vivo. PLoS One. 8(10):e76299

3. Queval, A., Ghattamaneni, N.R., Perrault, C.M., Gill, R., Mirzaei, M., McKinney, R.A., and Juncker, D. (2010) Chamber and microfluidic probe for microperfusion of organotypic brain slices. Lab on a Chip. 10(3):326

249

European Journal of Neuroscience, Vol. 40, pp. 2766–2776, 2014 doi:10.1111/ejn.12638

MOLECULAR AND SYNAPTIC MECHANISMS

Prolonged ampakine exposure prunes dendritic spines and increases presynaptic release probability for enhanced long-term potentiation in the hippocampus

Philip K.-Y. Chang,1 George A. Prenosil,1 David Verbich,2 Raminder Gill1 and R. Anne McKinney1,2 1Department of Pharmacology & Therapeutics, McGill University, Bellini Life Science Complex, Room 167, 3649 Promenade Sir-William-Osler, Montreal, QC, H3G 0B1, Canada 2Department of Neurology & Neurosurgery, McGill University, Montreal, QC, Canada

Keywords: CX546, LTP, organotypic hippocampal slices, synaptic plasticity, synaptic transmission

Abstract CX546, an allosteric positive modulator of a-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid-type ionotropic glutamate receptors (AMPARs), belongs to a drug class called ampakines. These compounds have been shown to enhance long-term potentiation (LTP), a cellular model of learning and memory, and improve animal learning task performance, and have augmented cognition in neurodegenerative patients. However, the chronic effect of CX546 on synaptic structures has not been examined. The structure and integrity of dendritic spines are thought to play a role in learning and memory, and their abnormalities have been implicated in cognitive disorders. In addition, their structural plasticity has been shown to be important for cognitive function, such that dendritic spine remodeling has been proposed as the morphological correlate for LTP. Here, we tested the effect of CX546 on dendritic spine remodeling following long-term treatment. We found that, with prolonged CX546 treatment, organotypic hippocampal slice cultures showed a signiﬁcant reduction in CA3–CA1 excitatory synapse and spine density. Electrophysiological approaches revealed that the CA3–CA1 circuitry compensates for this synapse loss by increasing synaptic efﬁcacy through enhancement of presynaptic release probability. CX546-treated slices showed prolonged and enhanced potentiation upon LTP induction. Furthermore, structural plasticity, namely spine head enlargement, was also more pronounced after CX546 treatment. Our results suggest a concordance of functional and structural changes that is enhanced with prolonged CX546 exposure. Thus, the improved cognitive ability of patients receiving ampakine treatment may result from the priming of synapses through increases in the structural plasticity and functional reliability of hippocampal synapses.

Introduction In the brain, neurons communicate with each other through special- 1999; Maletic-Savatic et al., 1999; Stamatakou et al., 2013), spine ised structures called synapses. The integrity of synaptic structures branching (Toni et al., 1999), actin-dependent remodeling (Matus, dictates the fidelity of neurotransmission, and dysgenesis of dendritic 2000; Kramar et al., 2006; Hotulainen & Hoogenraad, 2010), and spines is often observed in neurological disorders (Levenga et al., an increase in spine volume accompanied by insertion of a-amino- 2012). Dendritic spines are small, morphologically and functionally 3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA)-type iono- diverse, protrusions that serve as excitatory postsynaptic sites on den- tropic glutamate receptors (AMPARs) (Matsuzaki et al., 2004; Kopec drites of principal neurons in the brain (McKinney & Thompson, et al., 2006; Bosch & Hayashi, 2012). Therefore, a structural plastic- 2009). In the mature hippocampus, spines are classified into three dis- ity enhancement of spines may, in theory, promote the propensity for tinct subtypes: stubby, thin, and mushroom (McKinney, 2005; Hill & learning and memory acquisition. Although series of experiments Zito, 2013). This diversity in dendritic spine morphologies has been have shown that a class of pharmaceutical cognitive enhancers called proposed to represent differences in the functional characteristics of ampakines can enhance spine actin polymerisation following LTP synapses (Bourne & Harris, 2007). Learning models, such as long-term induction (Baudry et al., 2012; Kramar et al., 2012), less is known potentiation (LTP), have been shown to induce dendritic spine struc- about their long-term effects on dendritic spines. tural modifications (De Roo et al., 2008; Hill & Zito, 2013; Jaafari Ampakines are novel benzamide compounds that allosterically et al., 2013), including new spine formation (Engert & Bonhoeffer, modulate AMPARs by blocking their desensitisation and deactivation (Arai & Kessler, 2007). At the cellular level, ampakines facilitate LTP and upregulate the neurotrophin brain-derived neurotrophic Correspondence: R. Anne McKinney, 1Department of Pharmacology & Therapeutics, factor (BDNF) (Lauterborn et al., 2003; Rex et al., 2006), favoring as above. their use for the treatment of cognitive deficits (Lynch et al., 2008). E-mail: [email protected] On the basis of these findings, ampakines have been subjected to Received 8 November 2013, revised 24 April 2014, accepted 28 April 2014 clinical trials for various neurological disorders (Ingvar et al., 1997;

Arai & Kessler, 2007; Kanju et al., 2008). Although the behavioral treated daily with 250 lM (2,3-dihydro-1,4-benzodioxin-6-yl)-1-pipe- and acute neurophysiological effects of ampakines are well docu- ridinylmethanone (CX546) for 2 weeks. At this concentration, mented (Arai & Kessler, 2007), less is known about synaptic CX546 has been previously reported to show the positive effects of changes with prolonged use. Therefore, we investigated whether ampakine treatment (St€aubli et al., 1994a,b; Lauterborn et al., ampakines’ positive effects on synaptic plasticity are associated with 2003). All drugs were obtained from Tocris Biosciences (Ellisville, any postsynaptic structural alterations following chronic treatment. MO, USA), and were prepared in serum-containing medium. Sister In this study, we chronically treated organotypic hippocampal cultures were treated daily with control medium to eliminate possi- slice cultures for 2 weeks with the ampakine CX546, to reflect the ble handling artefacts. Secondary and tertiary dendritic branches time course of animal experiments with clinically relevant concen- from either apical or basal dendrites of CA1 pyramidal neurons were trations (St€aubli et al., 1994a,b; Ingvar et al., 1997; Hampson et al., imaged with a Leica TCS SP2 scanhead (Leica Microsystems, Hei- 1998a,b; Yamada, 1998; Lauterborn et al., 2003; Porrino et al., delberg, Germany) on a Leica DM6000 B upright microscope, 2005; Simmons et al., 2009). Surprisingly, after chronic ampakine equipped with an HCX PL APO 9 63 NA 1.4 oil immersion objec- treatment, we observed a profound loss of dendritic spines on CA1 tive, or a DM LFSA upright microscope equipped with an HCX pyramidal neurons that was not accompanied by cell death. We also APO 9 63 NA 0.9 water immersion objective, with a 488-nm Ar or found a reduction in spontaneous AMPA-mediated miniature excit- a 543-nm HeNe laser line. Image stacks were collected at atory postsynaptic current (mEPSC) frequency, but an increase in Z = 0.25 lm, and averaged six times to improve the signal-to-noise presynaptic neurotransmitter release probability. When challenged ratio. For the recovery experiments, after 2 weeks of CX546 treat- with LTP induction protocols, the treated cultures showed more ment, the culture medium containing the ampakine was exchanged robust potentiation and spine enlargement. Taken together, our data for ampakine-free control medium. The cultures were returned to the provide evidence that the learning enhancement associated with incubator for 1 week before confocal images were collected as ampakine treatment is attributable, in part, to synapse priming. We mentioned above and quantified as described below. propose that this priming mechanism works by a combination of morphological (increased structural plasticity) and functional Electrophysiological recordings (increased presynaptic neurotransmitter release probability) changes that allow synapses to better respond to patterned stimulation and Slices were transferred into a temperature-controlled chamber learning events. (30 °C) mounted on an upright microscope (DM LFSA; Leica Microsystems), and continuously perfused with Tyrode solution containing: 137 mM NaCl, 2.7 mM KCl, 2.5 mM CaCl2,2mM MgCl2, Materials and methods 11.6 mM NaHCO3, 0.4 mM NaH2PO4, and 5.6 mM glucose (pH 7.4). Patch and field recording electrodes were pulled from borosili- Organotypic hippocampal slice cultures cate glass (GC150TC; Clark Instruments, Old Sarum, Salisbury, All procedures for animal handling were carried out according to UK). All electrophysiological recordings were performed with an the guidelines of the Canadian Council on Animal Care, and were Axopatch 200A amplifier (Molecular Devices, Sunnyvale, CA, approved by the Animal Resource Committee of the School of USA). For comparison of the results between CX546-treated and Medicine at McGill University (McGill University Animal Han- untreated cultures, data were always obtained from sister cultures. dling Protocol no. 5057). Organotypic hippocampal slice cultures were chosen because of their ability to be treated chronically in a AMPA-mediated mEPSCs controlled environment in an organised tissue preparation. Further- more, in these cultures, the neural architecture is preserved, and Whole-cell voltage-clamp recordings from CA1 pyramidal neurons can be readily imaged with confocal microscopy. In brief, hippo- were obtained at 30 °C by the use of electrodes with resistances of campal slices (400 lm) from postnatal day 6–8 transgenic mice 4–5MΩ and filled with an intracellular solution containing 120 mM expressing membrane-targeted MARCKS–enhanced green fluores- potassium gluconate, 1 mM EGTA, 10 mM Hepes, 5 mM Mg-ATP, cent protein (eGFP) under the Thy-1 promoter in a subpopulation 0.5 mM Na-GTP, 5 mM NaCl, 5 mM KCl, 10 mM phosphocreatine of CA1 cells (De Paola et al., 2003; Verbich et al., 2012) were (295 mOsm; pH adjusted with KOH to 7.3). mEPSCs were recorded prepared as previously described with the roller-tube method at À60 mV and in the presence of 1 lM tetrodotoxin, 15 lM 3-[(R)- (G€ahwiler et al., 1997; Machnes et al., 2013). Briefly, hippocampal 2-carboxypiperazin-4-yl]-propyl-1-phosphonic acid (CPP), 100 lM slices were prepared with a tissue-chopper, embedded in a plasma picrotoxin and 1 lM CGP55845 in the external Tyrode solution. clot on glass coverslips, placed and sealed in flat-sided tubes with Access resistance was monitored with brief test pulses at regular serum-containing medium, and incubated in a dry-air roller-drum intervals (2–3 min) throughout the experiment. The access resistance incubator at 36 °C. The culture medium consisted of 25% heat- was usually 10–13 MΩ, and data were discarded if the resistance inactivated horse serum, 25% Hank’s balanced salt solution, and deviated by > 20% through the course of the experiment. Series 50% Basal medium Eagle. Slice cultures were maintained in the resistance of the access pulse and decay time were also used for the incubator with a roller-drum for 3 weeks before experimentation to calculation of total membrane capacitance. After the holding current allow for maturation (Gill et al., 2013). The spine distribution is had stabilised, data were recorded at a sampling frequency of not different in cultures maintained for 3–6 weeks (Chang et al., 10 kHz, and filtered at 2 kHz for 10–15 min. All mEPSCs were 2013). detected offline with MINI ANALYSIS software (Synaptosoft, Decatur, GA, USA). The amplitude threshold for mEPSC detection was set at four times the root-mean-square value of a visually event-free Chronic pharmacological treatment with CX546 recording period. From every experiment, 300 events were randomly All pharmacological treatments were started in cultures after selected for blinded analysis of amplitude and inter-event interval. ≥ 3 weeks in vitro. Sister cultures which were prepared from the The data obtained were then used to plot cumulative histograms same mouse litter and received similar handling were chosen and with an equal contribution from every cell. For statistical analysis,

© 2014 Federation of European Neuroscience Societies and John Wiley & Sons Ltd European Journal of Neuroscience, 40, 2766–2776 2768 P. K.-Y. Chang et al. data were averaged for every cell. It should be noted that the ampli- USA), was prepared at a 1 : 250 dilution in 0.1 M PB containing tude analysis was conducted only on single mEPSCs that did not 1.5% horse serum, and incubated for 2 h at room temperature. have subsequent events occurring during their rising and decaying Slices were then mounted with Dako Fluorescent Mounting medium phases. For frequency analysis, all selected events were considered. (Dako Canada, Mississauga, ON, Canada) onto microscope slides prior to imaging and subsequent blinded analysis. Localisation and quantification of immunopositive puncta were performed with the Evoked (paired-pulse) excitatory postsynaptic potential ‘Spots’ reconstruction function in IMARIS (Bitplane AG, Zurich, Swit- (EPSP) recordings zerland). Whole-cell voltage-clamp recordings were conducted as above, but without tetrodotoxin in the external bath solution. A surgical cut Chemically induced LTP (chemLTP) and time-lapse confocal was made between CA3 and CA1. Paired pulses (0.1 ls, 10– microscopy 40 lA) were provided to Schaffer collateral axons every 15 s with a constant current stimulator (IS4 Stimulator; HIFO, Zurich, Switzer- Slices were transferred to a chamber mounted on an upright micro- land). Stimulation was achieved with a monopolar glass electrode scope (DM LFSA; Leica Microsystems) equipped with the Leica (made from a patch electrode) filled with Tyrode solution. The use TCS SP2 scanhead, and perfused continuously with Tyrode solution of a monopolar glass electrode allowed us to probe for responses at 30 °C (as above). Image stacks were collected at 1-min time with a minimal amount of polysynaptic content while recording intervals, and, after 10–15 min of baseline imaging, the perfusing from a patch-clamped CA1 pyramidal cell. Access resistance was solution was changed to the chemLTP solution (Kopec et al., 2006, monitored with a test pulse (À5 mV, 50 ms) at the beginning of 2007; Makino & Malinow, 2009), comprising MgCl2-lacking each recorded trace, and was found to be similar to that in the Tyrode’s solution with 100 nM rolipram, 50 lM forskolin, and experiments described above. The traces were analysed offline by 100 lM picrotoxin, for exactly 16 min, as previously reported the use of IGOR PRO 6 software with a custom-designed program, by (Kopec et al., 2006, 2007; Makino & Malinow, 2009). After measuring the peak amplitude of every EPSC (WaveMetrics, Lake 16 min, control Tyrode solution was perfused for the remainder of Oswego, OR, USA). the experiment (~45 min). The time-lapse images were taken with an HCX APO 9 63 0.9 NA (Leica) water immersion long working distance lens at Z = 0.25 lm. Secondary and tertiary dendritic Field EPSP (fEPSP) recordings and LTP branches from either apical or basal dendrites were acquired. As no fEPSPs were recorded in the CA1 stratum oriens at 30 °C by the difference between basal and apical dendrites was observed, data use of recording electrodes with tip resistances of 0.5–1MΩ and were subsequently pooled for blinded analysis. fEPSP measurements filled with Tyrode solution containing 100 lM picrotoxin to suppress were also obtained from the chemLTP solution-treated cultures. inhibitory responses at the recording site, in order to isolate excit- Electrical stimulation (40–200 lA) was delivered every 20 s, but no atory responses following the stimulation. The extracellular solution stimulation was provided during the perfusion of chemLTP solution consisted of regular Tyrode solution without the addition of any and 40 min thereafter, to prevent excessive synaptic activation. To pharmacological agents. A surgical cut was made between CA3 and test for the dependency of N-methyl-D-aspartate receptor (NMDAR) CA1. Schaffer collateral axons were then stimulated with an insu- activation mediated by chemLTP, we recorded fEPSPs in the lated platinum–iridium bipolar electrode [diameter, 50 lm; 25-lm presence of the NMDAR blocker CPP during chemLTP induction. layer of Teflon insulation (A-M Systems, Carlsborg, WA, USA)]. The stimulus strength was between 40 lA and 200 lA, to achieve Cell viability assay with propidium iodide (PI) one-half to one-third of the maximum response, and stimulation was provided every 20 s. LTP was induced with a theta-burst stimulation Hippocampal slice cultures were tested for cell death after chronic (TBS) paradigm consisting of bursts of five pulses at 100 Hz; the ampakine treatment. Briefly, 5 lg/mL PI (Sigma-Aldrich, St Louis, bursts were applied five times at intervals of 200 ms, and delivered MO, USA) was added to the culture, which was returned to the five times every 10 s. Data analysis was performed blinded by the incubator for 10 min and then imaged with epifluorescence micros- use of IGOR PRO 6 software with a custom-designed program by mea- copy (Axioplane Imaging Fluorescence Microscope; Carl Zeiss suring the initial slope of every fEPSP. In the case of polysynaptic MicroImaging, Hamburg, Germany). As a positive control for cell + fEPSP responses that were not rising smoothly, only the first slope death and PI staining, we added excess K solution (KCl, 50 mM) was measured. Every three data points were pooled into one, and to the culture for 20 min. normalised against 10 min of baseline recordings. Data were discarded if the normalised fEPSP slope during the baseline recording Three-dimensional reconstructions and spine quantification period changed by > 15%. Three-dimensional and four-dimensional image stacks were first deconvolved with HUYGENS ESSENTIALS software (Scientific Volume Immunohistochemical staining of synaptic proteins Imaging, Hilversum, The Netherlands) by use of a full maximum Following chronic treatment, hippocampal slice cultures were likelihood extrapolation algorithm. Stacks were then imported and removed from the coverglass and fixed in 4% paraformaldehyde dis- rendered with SURPASS function in IMARIS (Bitplane AG). IMARIS 6.3.1 solved in 0.1 M phosphate buffer (PB) overnight at 4 °C. Following (Bitplane AG) was used for semi-automated dendritic spine identifi- fixation, the cultures were washed in 0.1 M PB, permeabilised in cation and quantification of all dendritic spines. The spine detection/ 0.4% Triton X-100, and blocked with 1.5% horse serum overnight classification program automatically detected the length of the spine at 4 °C. Primary anti-synaptophysin antibody (Zymed, South San head and neck. From the ratio of the diameter and the length of Francisco, CA, USA) was incubated for 5 days at 4 °C in the per- the head and the neck of the spines, it was possible to distinguish meabilising solution at a 1 : 400 dilution. The secondary antibody, the stubby, mushroom, long, thin spines, dendrite branches, and filo- anti-Cy-5 (Jackson ImmunoResearch Laboratories, West Grove, PA, podia. These classifications were based on previously established

© 2014 Federation of European Neuroscience Societies and John Wiley & Sons Ltd European Journal of Neuroscience, 40, 2766–2776 Ampakines alter synaptic and dendritic spine plasticity 2769 criteria (McKinney et al., 1999; McKinney, 2010). Following spine chronically treated cultures (Fig. 1C; control, n = 16 cells, one classification, spine motility was measured by the changes in vol- dendritic segment of 20–50 lm per cell; CX546-treated, n = 19 ume of the spines over time. Spine volumes were measured with cells; P < 0.01). IMARIS 6.3.1 (Bitplane AG) as previously described (Chang et al., 2013). Briefly, the time-lapse confocal image stacks were four- CX546 treatment does not cause cell death, but decreases dimensionally rendered under the SURPASS function, whereby an the number of synaptic contacts formed on dendritic spines isosurface was built on the basis of the fluorescence intensity of the four-dimensional image. The spines of interest were isolated at As spine loss can be indicative of neuronal degeneration, and excess each time point, and the volume was measured on the basis of the glutamate-mediated neurotransmission can be toxic to neurons, we structure constructed by the isosurface. The same parameters for examined culture viability after prolonged ampakine exposure. We image reconstruction were used for all time points of an experi- used PI to assay necrotic cell death, and found comparably low mental series, for both treated and control cultures. We then mea- numbers of PI-positive cells in the control and treated sister cultures, sured the volume changes in all thin and mushroom spines, and indicating that a vast majority of cells in the cultures were still via- excluded stubby spines, because the size of stubby spines is close ble following 2 weeks of ampakine treatment (Fig. 2A; n = 5 to the resolution limit of light microscopy, preventing proper volume analysis. A Statistical analysis Student’s t-tests were performed, and Bonferroni’s correction was applied when appropriate; the Mann–Whitney U-test was applied to non-parametric data sets. Results are expressed as mean Æ standard error of the mean. All data analyses were performed blind; however, the striking phenotypes in some experiments prevented proper blinding of the experimenters.

Results

CX546 treatment decreases dendritic spine densities B First, the effect of chronic ampakine treatment on dendritic spine morphology was examined, as these structures have been shown to be important for cognitive functions. We treated sister hippocampal cultures with either control or CX546-containing culture medium for 2 weeks, which is a similar duration as used in animal studies. Pre- C vious studies have found that, at a concentration of 250 lM, the acute presence of the ampakine CX546 can enhance LTP and upregulate BDNF. By imaging membrane-targeted green fluorescent protein (mGFP)-positive cells from transgenic hippocampal cultures, we found that overall morphology of CA1 pyramidal cells remained unchanged with chronic treatment (Fig. 1A). Interestingly, once we focused on dendritic spines, we observed a significant 70% decrease in total spine density in cultures chronically treated with CX546 (Fig. 1B). Under control conditions, the total spine density was 1.04 Æ 0.08 spines/lm, whereas the chronically CX546-treated cultures had a total spine density of 0.33 Æ 0.06 spines/lm (Fig. 1C; control, n = 16 cells, one dendritic segment of 20–50 lm per cell; CX546-treated, n = 19 cells; P < 0.001). This change in dendritic Fig. 1. Two weeks of CX546 treatment decreases dendritic spine densities spine density did not result from further developmental changes fol- without affecting overall hippocampal CA1 pyramidal neuron morphology. lowing the initial 3-week maturation period, as the spine distribution (A) Examples of three-dimensional reconstructions of a hippocampal CA1 pyramidal neuron from a control culture (left) and a CX546-treated culture remained similar in cultures maintained for up to 6 weeks thereafter (right). CA1 pyramidal neurons have similar dendritic arborisations in both (McKinney et al., 1999). When we classified spines by their the control and treated cultures. Scale bar: 40 lm. (B) Left: example of ter- respective subtype in each condition, we found that stubby spines tiary portions of dendrites in the control cultures. Right: example of portions decreased in density from 0.26 Æ 0.03 spines/lm to 0.10 Æ of treated dendrites of hippocampal CA1 pyramidal neurons. Note that, in 0.01 spines/lm, and mushroom spines decreased in density from the control, numerous spines of various subtypes are present. In the chroni- Æ l Æ l cally CX546-treated hippocampal culture, very few spines appear to be visi- 0.32 0.04 spines/ m to 0.07 0.02 spines/ m, in CX546-treated ble. In particular, the mushroom spines have disappeared, with remaining cultures as compared with control sister cultures (Fig. 1C; control, spines being mainly stubby and thin. Scale bar: 2 lm. (C) Graph showing n = 16 cells, one dendritic segment of 20–50 lm per cell; CX546- spine densities of dendritic spines with (black bar) or without (white bar) l treated, n = 19 cells; P < 0.005). Interestingly, of the few remaining chronic treatment with 250 M CX546, including the densities of various spine subtypes. The densities of all spine types decreased significantly along spines on dendrites from treated cultures, the majority were with the total numbers of spines. Control, n = 16 cells, one dendritic seg- thin spines. The density of thin spines decreased from 0.47 Æ ment of 20–50 lm per cell. CX546-treated, n = 19 cells. **P < 0.01, 0.06 spines/lm in control cultures to 0.16 Æ 0.04 spines/lmin ***P < 0.005 and ****P < 0.001.

= = control cultures, n 5 CX546-treated cultures, n 5 negative-con- A trol cultures, and n = 5 positive-control cultures). We also tested other time points to eliminate the possibility that the lack of PI staining following 2 weeks of treatment was attributable to the clearance of dead cells at earlier time points. Hence, we examined PI staining following 1, 2 and 7 days of CX546 treatment. We found no significant differences in PI staining at any time point that we examined (Fig. 2A; n = 5 control cultures and n = 5 CX546-treated cultures at each time point). Additionally, we determined whether the removal of CX546 would rescue the spine loss, another indication of cell viability. We found that dendritic spine density in chronically CX546-treated cultures returned to control levels after 1 week of incubation in control medium (Fig. S1A and B: total, control, B 1.04 Æ 0.08spines/lm; total, CX546 + recovery, 1.31 Æ 0.05 spines/lm; stubby, control, 0.26 Æ 0.03 spines/lm; stubby, CX546 + recovery, 0.32 Æ 0.03 spines/lm; mushroom, control, 0.32 Æ 0.04 spines/lm; mushroom, CX546 + recovery, 0.38 Æ 0.02 spines/lm; thin, control, 0.47 Æ 0.06 spines/lm; thin, CX546 + recovery, 0.60 Æ 0.06 spines/lm; control, n = 14 cells, one dendritic segment of 20–50 lm per cell; CX546, n = 16 cells; P > 0.05). The facts that cell death was negligible and that the spines recovered after ampakine removal strongly suggested that the changes observed did not result from toxicity of the ampakine treatment. The ensuing question was whether loss of dendritic spines is accompanied by the pruning of synaptic contacts. Previous work from our group has shown that spine loss is not a perfect correlate of synapse loss, as the synaptic machinery can retract back onto the dendritic shaft when C spines are lost (Mateos et al., 2007). Therefore, we immunostained control and treated sister cultures for the presynaptic terminal marker synaptophysin, and co-localisation of these puncta with mGFP-positive dendrites was used to indicate putative excitatory synapses. Similarly to what we have shown, our experiments revealed that the majority of mature dendritic spines are in contact with putative presynaptic boutons (Richards et al., 2005). However, in CX546-treated cultures, we detected a loss of synapses that were localised on dendritic spines (Fig. 2B and C: total synapse density, control, 13.84 Æ 0.67 contacts/10 lm; total synapse density, CX546-treated, 9.08 Æ 0.74 contacts/10 lm; spine synapse density, control, 11.31 Æ 0.0.50 contacts/10 lm; spine synapse density, CX546-treated, 6.08 Æ 0.68 contacts/10 lm; control, n = 15 cells, one dendritic segment of 20–40 lm per cell; CX546-treated, n = 17 cells; Fig. 2. Two weeks of CX546 treatment does not affect the viability of hip- P < 0.05), but not of those located on the dendritic shaft (Fig. 2B pocampal slice cultures but decreases putative synapse density. (A) PI stain- and C: shaft synapse density, control, 2.52 Æ 0.35 contacts/10 lm; ing: fluorescent positive staining (white signal) indicates cell death, whereas shaft synapse density, CX546-treated, 2.99 Æ 0.24 contacts/10 lm; viable cells do not show staining (black background). The negative control > showed no staining, whereas maximum staining was observed in the positive P 0.05). This result showed that CX546 decreases the numbers of control. Similar numbers of PI-labeled cells are seen in both control and trea- dendritic spines and synapses. The fact that cell death was similar in ted cultures, indicating that chronic CX546 treatment does not cause cell control and CX546-treated cultures strongly suggested that the death. Slices are shown in the same orientation. DG, dentate gyrus. = = = observed decreases in spine density and putative synapse density did Scale bar: 1 mm. n 5 control slices, n 5 CX546-treated slices, n 5 negative-control slices, n = 5 positive-control slices. (B) Representative not result from excitotoxicity. As synaptic structure can dictate the three-dimensional reconstruction of dendritic segments (white) from control integrity of neurotransmission and even cognitive function, the loss and CX546-treated cultures immunostained with anti-synaptophysin antibody of presynaptic and postsynaptic structures prompted us to assess the (red) to determine the location of the putative synapses. Puncta immunostain- functionality of the slice cultures after CX546 treatment. ing positively for synaptophysin were reconstructed in IMARIS (Bitplane, AG), and are shown as red spheres in contact with dendritic spines. Scale bar: 2 lm. (C) Quantification of the densities of putative synapses located on the CX546 treatment decreases the frequency of mEPSCs, but dendritic shaft or spine between the control and treated cultures. There was a significant decrease in the density of spine synapses between the control and increases the reliability of synaptic transmission by treated cultures. Spine synapses, control, 11.31 Æ 0.50 contacts/10 lm; spine enhancement of synaptic vesicle release probability synapses, CX546-treated, 6.08 Æ 0.68 contacts/10 lm; shaft synapses, control, 2.52 Æ 0.35 contacts/10 lm; shaft synapses, CX546-treated, To assess the functional impact of spine loss caused by CX546 2.99 Æ 0.24 contacts/10 lm; total synapses, control, 13.84 Æ 0.67 contacts/ treatment, we recorded AMPAR-mediated mEPSCs to determine the 10 lm; total synapses, CX546-treated, 9.08 Æ 0.74 contacts/10 lm; n = 10 integrity of synapses. Spontaneous mEPSCs were recorded from dendrites from 10 cells from six slices (three control; three treated); CA1 pyramidal neurons from either control or CX546-treated *P < 0.05.

© 2014 Federation of European Neuroscience Societies and John Wiley & Sons Ltd European Journal of Neuroscience, 40, 2766–2776 Ampakines alter synaptic and dendritic spine plasticity 2771 cultures (Fig. 3A). We found an increase in the IEI (Fig. 3B: A control, 350.44 Æ 50.71 ms; CX546-treated, 625.17 Æ 100.14 ms; control, n = 11 slices, one or two cells per slice, 300 random events selected per cell; CX546-treated, n = 11 slices; P < 0.05), with no significant effect on the average amplitude of mEPSCs, in chronically treated cultures (Fig. 3C: control, 20.39 Æ 0.99 pA; CX546- treated, 20.36 Æ 0.88 pA; control, n = 11 slices, CX546-treated, n = 11 slices; P > 0.05). To further explore the functional consequences of such morphological alteration, we next determined the capacitance of the cell, as this is a correlate of total membrane sur- B face area. In neurons from control cultures, a cell capacitance of 13.8 Æ 1.25 pF was measured (n = 23 cells, one or two cells per slice). Neurons from treated sister cultures had a significantly reduced cell capacitance of 9.4 Æ 0.9 pF (n = 20 cells; P < 0.01), indicating a change in total membrane surface area. Concurrent with the loss of synaptophysin-positive contacts, the decrease in mEPSC frequency further confirms the loss of functional synapses following CX546 treatment. We also examined whether functional recovery was possible following CX546 withdrawal. When we recorded spontaneous mEPSCs in those cultures that were chronically treated for 2 weeks and left to recover in control medium thereafter, we found C that there were no discernible differences from control cultures in either IEI (Fig. S2: control, 359.61 Æ 42.92 ms; CX546-treated, 297.01 Æ 39.84 ms; control, n = 7 slices; CX546-treated, n = 7 slices; P > 0.05) or amplitude (Fig. S2: control, 20.14 Æ 2.00 pA; CX546-treated, 12.88 Æ 1.70 pA; control, n = 7 slices; CX546-treated, n = 7 slices; P > 0.05). The normalisation of IEI to control level corresponded to the restoration of dendritic spine density 1 week following the removal of CX546-containing media. As the reduction in mEPSC frequency did not appear to result from cell death, but rather from a decrease in the number of syn- DE apses, we were interested in determining whether the remaining synapses have greater efficacy. To determine whether the presynaptic neurotransmitter release probability was altered by the chronic CX546 treatment, we performed paired-pulse facilitation (PPF) experiments on both control and chronically CX546-treated sister cultures. It is widely recognised that, whereas a high presynaptic neurotransmitter release probability is usually reflected in a weaker PPF, a low neurotransmitter release probability will generally lead to an increased PPF (Katz & Miledi, 1968; Manabe et al., 1993; Mitra et al., 2012). When the degree of PPF was compared between Fig. 3. Two weeks of CX546 treatment decreases AMPA-mediated mEPSC sister control and CX546-treated cultures, we observed smaller PPF frequency but not amplitude, and the PPF ratio decreases. (A) Example traces in the CX546-treated cultures [Fig. 3D and E: 25-ms inter-stimulus obtained from CA1 pyramidal neurons of sister control cultures (top) and interval (ISI), control, 1.07 Æ 0.15; 25-ms ISI, CX546-treated, cultures treated chronically with 250 lM CX546 (bottom) for 2 weeks. (B) 0.40 Æ 0.13; 50-ms ISI, control, 0.73 Æ 0.078; 50-ms ISI, CX546- Graphical comparison of the difference between control and CX546-treated Æ Æ slices and their averaged mEPSC IEIs (left) and cumulative distribution treated, 0.42 0.11; 75-ms ISI, control, 0.57 0.12; 75-ms ISI, (right) obtained from 300 randomly selected events from each cell, showing CX546-treated, 0.19 Æ 0.065; control, n = 11; CX546-treated, a significant increase in the IEIs in neurons from CX546-treated cultures. n = 11; P < 0.05], indicating a higher release probability at CA3– Gray: control cultures. Black: CX546-treated cultures. *P < 0.05. (C) Graph- CA1 synapses. ical comparison of averaged mEPSC amplitude (left) and cumulative distribution (right) of control and CX546-treated cultures, showing that there is no This result shows that CX546 led to an overall synapse loss while change in amplitude; P > 0.05. Control, n = 11 slices, 300 random events increasing the neurotransmission reliability of the remaining synapses selected per slice. CX546-treated, n = 11 slices. (D) Example of PPF. Indi- through increases in release probability, which is similar to the pre- vidual traces are in gray, and averaged traces are in black: Left: control cul- synaptic phenomenon observed following LTP induction (Enoki tures. Right: CX546-treated cultures. (E) PPF ratios at different ISIs for fi et al., 2009). control and CX546-treated cultures. There are signi cant decreases in PPF ratios at 250 ms, 50- and 75-ms ISIs. Open circles: control. Black circles: CX546-treated. n = 22 slices (11 control; 11 CX546-treated). *P < 0.05. CX546 treatment leads to prolonged LTP Previous studies have shown that ampakines can induce more sus- still show functional plasticity. We used a TBS protocol at CA3– tained LTP in the hippocampus (St€aubli et al., 1994a; Rex et al., CA1 synapses to study potentiation after chronic treatment. We 2006) and that an increase in presynaptic release probability can detected LTP in both control and ampakine-treated sister cultures result from patterned stimulation (Enoki et al., 2009). We wanted to (Fig. 4A and B). Interestingly, with CX546 treatment, fEPSPs determine whether, after chronic ampakine treatment, synapses can remained more potentiated for longer (Fig. 4C: first 10 min,

154% Æ 14.1% baseline; last 10 min, 152.6% Æ 15.8% baseline; shown to enhance actin polymerisation following LTP (Baudry n = 13 slices, one slice per experiment; P > 0.05) than in control et al., 2012; Kramar et al., 2012). cultures (Fig. 4C: first 10 min, 168% Æ 19.5% baseline; last 10 min, 120% Æ 10.3% baseline; n = 20 slices, one slice per one CX546 treatment enhances activity-dependent structural experiment; P < 0.05). Next, we investigated whether this robust modifications of dendritic spines during LTP LTP found after ampakine treatment may be linked to possible structural plasticity of dendritic spines, as ampakines have been Postsynaptic remodeling has shown a strong association with LTP stimulation protocols, and is widely believed to be the structural substrate for learning and memory. To determine whether morphological modification can be enhanced with chronic CX546 treatment, A we monitored postsynaptic structural changes induced by LTP through time-lapse imaging of mGFP-positive CA1 dendrites. We used chemLTP as a form of LTP induction for global potentiation, as previously reported (Kopec et al., 2006, 2007; Makino & Mali- now, 2009). TBS-induced LTP is not suitable for optical studies, because of the challenge of reliably stimulating a sufficient number of synapses in the imaging field (Kopec et al., 2006, 2007; Makino & Malinow, 2009). The chemLTP method that we chose induces an NMDAR-dependent form of LTP that can be successfully blocked by the NMDAR blocker CPP (Fig. S3A), as shown previously by others (Kopec et al., 2006, 2007; Makino & Malinow, 2009). In B order to dissect the changes in dendritic spine morphology after chemLTP induction, we measured the volume changes in all thin and mushroom spines. Analysis of stubby spines was excluded, because the size of stubby spines is close to the resolution limit of light microscopy, preventing proper volume measurements. Spine volumes were determined to provide an indication of structural modification during LTP, because it has been shown that spine volumes can increase as a result of AMPAR insertion (Matsuzaki et al., 2004). As different spine subtypes have been reported to have different properties (Bourne & Harris, 2007), the spines were quantified according to their subclasses. Interestingly, when we determined the volume fluctuation of thin spines, we found a large elevation in the averaged volumes of thin spines in CX546-treated cultures (Fig. 5A and C, left: control, À4.17% Æ 7.22% increase in spine volume; CX546-treated, 29.2% Æ 16.3% increase in spine volume; C n = 9 cells, one dendritic segment of 20–30 lm per cell; CX546- treated, n = 11 cells; P < 0.05). Moreover, following the initial volume enlargement in the treated cultures, this increase in size was sustained until the end of the time-lapse experiment (~45 min after chemLTP induction). This was not seen under control conditions. Although there were observable volume increases in some spines, the averaged volume change of all spines was not significantly changed, because not all spines increased in size. On average, 77.00% Æ 11.40% of thin spines had a volume increase in CX546- treated cultures, as compared with only 37.92% Æ 10.69% in control cultures (P < 0.05; Mann–Whitney U-test). Furthermore, when we determined whether spine volumes were also changed in mushroom spines, we found a transient increase after chemLTP induction in control cultures, as the increase in volume quickly returned to the Fig. 4. Two weeks of CX546 treatment increases the sustainability and duration of TBS-induced LTP. (A) Representative averaged fEPSPs from a basal level (Fig. 5B and C). In contrast, the volumes of mushroom control culture and a chronically ampakine-treated culture at baseline (solid spines in CX546-treated cultures increased after induction, and line), 0–10 min after TBS (dotted line), and 50–60 min after TBS (gray line). showed, although not significantly, a trend towards an increase (B) LTP at Schaffer collateral synapses. The mean percentage change from (Fig. 5C: control, À1.60% Æ 9.55% increase in spine volume; baseline of the fEPSP slope is plotted against time. Potentiation of fEPSP CX546-treated, 22.6% Æ 11.3% increase in spine volume; n = 9 following TBS protocol (arrow) reached a higher level in CX546-treated cul- – l tures (black) than in control cultures (gray). Control: n = 20 slices, one slice cells, one dendritic segment of 20 30 m per cell; CX546-treated, per experiment. CX546-treated: n = 13 slices. P > 0.05. Also, we observed a n = 11 cells; P > 0.05). Of all of the mushroom spines, 55.00% Æ decrease in fEPSP slope following TBS in the control but not treated cultures 14.14% in CX546-treated cultures and 38.33% Æ 11.37% in control < (white triangle indicates P 0.05). (C) Graphical comparison of different cultures showed signs of a volume increase (P > 0.05; Mann–Whit- time periods and corresponding potentiation between control cultures (white) and treated cultures (black). A significant drop in fEPSP slope was observed ney U-test). Additionally, the observed dendritic spine volume in the control but not the treated cultures. Control: n = 18 slices, one slice enlargement following chemLTP induction was NMDAR-dependent, per one experiment; *P < 0.05. CX546-treated: n = 13 slices; P > 0.05. as the spine volume increases were blocked by CPP, an NMDAR

A U-test). We concluded that CX546 treatment signiﬁcantly increased dendritic spine structural plasticity in response to LTP.

Discussion Emerging lines of evidence suggest that LTP is accompanied by structural modifications of dendritic spines. However, this view has been challenged, because, using similar experimental conditions, others have not observed consistent postsynaptic modification of spine structures. The present study supports the idea that conforma- tional modifications occur concordantly at CA3–CA1 synapses during LTP. More interestingly, the pharmacological enhancement of B LTP is closely linked with more pronounced dendritic spine enlargement following CX546 treatment through destabilisation of previously established synapses. The first striking consequence of long-term CX546 application to hippocampal slice cultures was an unexpected decrease in dendritic spine density, with a propensity for the long, thin subclass of spines. These findings are unexpected, as the presence of dendritic spines has been shown to be important for learning and memory, and their dysgenesis has been observed in various forms of mental impairment (Pur- pura, 1974; Irwin et al., 2000). Also, previous studies exposing rats to enriched environments or training rodents excessively as paradigms to increase learning and memory reported increases in spine density C (Comery et al., 1996; Jones et al., 1997). Thus, as a ‘learning drug’, CX546 was expected to increase dendritic spine density. This discrepancy in spine density could be explained by the fact that the slice cultures were not exposed to any type of learning-like stimulus during the initial incubation stage. The effect of CX546 was evident when the slice cultures were challenged with LTP, where the CX546-treated slices showed more prolonged and enhanced LTP than control sister cultures. It has been shown by pharmacological manipulations that glutamate receptor activation leads to changes in dendritic spine distribution. For example, excess glutamate has excitotoxic effects on neurons that can result in spine loss, and chronic blockade of AMPARs decreases spine density and converts spine synapses to asymmetric Fig. 5. Two weeks of CX546 treatment enhances and sustains dendritic shaft synapses (McKinney et al., 1999; Mateos et al., 2007). Thus, spine enlargement associated with chemLTP. (A) Three-dimensional recon- CX546 treatment enhances AMPAR activation, and the subsequently structed image from a time-lapse experiment showing thin spines from either control or CX546-treated cultures. Zero indicates the induction of increased neuronal activity may lead to decreases in spine density as an chemLTP. Arrowheads point to the enlargement of spines. Increases in vol- adaptive structural modification. Furthermore, chronic application of ume can be observed following initial chemLTP; however, enlargement of ampakines can lead to the downregulation of AMPAR subunit protein l spine heads was not uniform in all spines. Scale bar: 1 m. (B) Three- levels (Lauterborn et al., 2009), a potential precursor of spine loss. dimensional reconstructed image from a time-lapse experiment showing mushroom spines from either control or CX546-treated scultures. Zero indi- Using electrophysiological recordings, we found that the decrease cates the start of chemLTP, Arrowheads point to the enlargement of spines. in spine density was accompanied by a decrease in the frequency of More spines from treated cultures showed an increase in volume that was mEPSCs but no change in amplitudes, whereas CX546-treated cul- sustained until the end of the experiment. (C) Left: normalised spine vol- tures had an increased release probability. The decrease in cell ume of thin spines during chemLTP time-lapse experiments. The thin spine capacitance indicates a decrease in cell membrane surface area stem- volume in the treated cultures (white markers) increased dramatically following chemLTP, whereas the thin spine volume in the control cultures did ming from dendritic spine loss. Moreover, the observed decrease in not (black markers). The black bar above the graph indicates the start of spine density suggests fewer excitatory synapses, although caution is chemLTP. n = 9 cells, one dendritic segment of 20–30 lm per cell. warranted, because spine loss is not a concomitant of synapse loss = < CX546-treated, n 11 cells; *P 0.05. Right: normalised spine volume of (McKinney et al., 1999; Mateos et al., 2007). However, through mushroom spines before, during and after chemLTP. There appears to be a trend for increasing mushroom spine volume in chronically treated cultures immunostaining of synaptophysin, we revealed that the total number following the start of chemLTP; the control mushroom spines do not show of putative contacts was decreased mostly through the loss of con- this increase. n = 9 cells, one dendritic segment of 20–30 lm per cell. tacts formed on dendritic spines. Together, these findings show that CX546-treated, n = 11 cells; P > 0.05. chronic CX546 exposure reduces excitatory synapse density while increasing the release probability of these remaining synapses. Post- synaptic effects can indeed regulate presynaptic properties as a blocker (Fig. S3B). On average, only 18.79% Æ 4.60% of all spines homeostatic response (Branco et al., 2008; Yu & Goda, 2009; Mitra showed a volume increase following chemLTP induction in the et al., 2012). It is therefore plausible that chronic enhancement of presence of CPP, as compared with 46.38% Æ 3.44% in control cul- excitatory neurotransmission provokes a homeostatically driven tures (control, n = 5 cells, one dendritic segment of 20–30 lm per decrease in synapse number, while increasing the reliability of the cell; chemLTP + CPP, n = 4 cells; P < 0.05; Mann–Whitney remaining synapses.

This disparity between learning enhancement and the loss of den- tion of BDNF induced by ampakine treatment. It has been shown dritic spines may be attributable to the downstream effect of CX546 previously that BDNF is required for persistent enlargement of den- treatment. A key feature of CX546 and other ampakines is their dritic spines in response to LTP induction (Tanaka et al., 2008). ability to cause sustained increases in BDNF levels (Lauterborn Through the use of CX546, we have shown spine volume enlarge- et al., 2003, 2009; Rex et al., 2006; Simmons et al., 2009). BDNF, ment after LTP induction in chronically CX546-treated slices. These as a pro-survival neurotrophin in the developing central nervous sys- results support previous findings that LTP is expressed structurally tem, is involved in LTP and synaptic plasticity in the mature brain as increases in spine density (Engert & Bonhoeffer, 1999; Maletic- (Huang & Reichardt, 2001; Lu, 2003). Spine morphologies are gov- Savatic et al., 1999) and spine volume (Ostroff et al., 2002; Matsu- erned by actin dynamics (Lippman & Dunaevsky, 2005; McKinney, zaki et al., 2004; Tanaka et al., 2008). Nonetheless, a study found 2005; Tada & Sheng, 2006), and numerous studies have shown that no postsynaptic alterations after LTP, but only an increase in the BDNF can also regulate actin dynamics through downstream signal- probability of transmitter release accounting for functional plasticity ing pathways involving small GTPases and the Ras family of pro- at CA3–CA1 synapses (Enoki et al., 2009). These different views teins (Smart & Halpain, 2000; Ethell & Pasquale, 2005; Tada & regarding LTP may, in fact, both be valid, as we have shown that Sheng, 2006; Sekino et al., 2007; Biou et al., 2008). Although, CX546 enhances both postsynaptic structural plasticity and presy- morphologically, BDNF has been shown to increase spine density naptic transmitter release probability. The facts that our observed and promote dendritic growth (Horch & Katz, 2002; Alonso et al., spine phenotype in ampakine-treated slices is typically long and 2004; Ji et al., 2005), in one study BDNF was overexpressed, and a thin, and that, upon chemLTP induction, we see a trend towards decrease in the number of spines in organotypic cortical slices was more mushroom spines and thin spine enlargement, suggest that found (Horch et al., 1999). It was reported that BDNF rapidly ampakines may increase the propensity for learning by altering spine altered the stability of dendritic spines, allowing greater turnover of structure. spines, and thus, presumably, synapses (Horch et al., 1999). The In conclusion, our data provide evidence for structural modifica- authors hypothesised that, upon receiving novel synaptic inputs, tions occurring concordantly at CA3–CA1 synapses during LTP that neurons can quickly deconstruct previously established connections are coupled to changes in presynaptic release probability and that in a BDNF-dependent manner to better accommodate the newly this process can be enhanced with CX546. We offer a plausible established pattern of inputs. The actions of BDNF may offer plausi- mechanism for how synapse pruning and increased synaptic reliabil- ble mechanisms for the decrease and change in dendritic spine dis- ity lead to learning enhancement in studies using ampakines. These tribution following CX546 treatment. First, the differential binding modified parameters allow for the increased potentiation of syn- of BDNF to either TrkB or p75NTR receptors can have opposing apses. Although CX546 has been proven to be efficacious in effects on neuron and dendritic development (Lu et al., 2005; Teng enhancing learning and memory, further investigations are warranted et al., 2005). Whereas the activation of TrkB receptors has been regarding their prolonged use in neuropathological conditions such shown to promote neuronal survival, differentiation, outgrowth, and as Alzheimer’s disease, where there is already a dramatic loss of LTP, signaling through p75NTR receptors can mediate apoptosis, dendritic spines (Baloyannis, 2009). More studies are required to pruning, neurite retraction, and long-term depression in neurons (Lu confirm whether ampakine treatment can lead to enhanced dendritic et al., 2005; Deinhardt & Chao, 2013). However, the exclusive spine reduction and/or increased structural remodeling in disease antagonism of these two receptors has been debated, as the blockade states. of either receptor can lead to either an increase or a decrease in dendritic spine density (Chapleau & Pozzo-Miller, 2012). Another pos- Supporting Information sible explanation for BDNF-mediated changes in dendritic spines is the speed at which BDNF is released (Ji et al., 2010). A slow rise Additional supporting information can be found in the online in BDNF level in cultured hippocampal neurons led to sustained version of this article: TrkB activation and elongation of spine necks, whereas acute BDNF Fig. S1. Removal of CX546 following a 2-week chronic treatment treatment resulted in transient TrkB activation, an increase in spine restored dendritic spine densities to control levels. density, and enlargement of spine heads (Ji et al., 2010). Therefore, Fig. S2. Removal of CX546 following a 2-week chronic treatment our observed dendritic spine loss and shift towards thin spines may restored AMPA-mediated mEPSC frequency to control levels. be the result of sustained elevation of BDNF caused by ampakine Fig. S3. ChemLTP induces NMDAR-dependent LTP. treatment and possibly the activation of p75NTR. Consequently, ampakines may improve information processing at synapses, in a Acknowledgements BDNF-dependent manner, by pruning, refining circuits, and increasing vesicular release, as an overall mechanism for priming neurons We thank Francßois Charron for technical assistance and Dr Pico Caroni for for increased synaptic plasticity, as proposed by previous studies generously providing the transgenic animals. P. K.-Y. Chang was supported by the Foundation of Ataxia Charlevoix-Saguenay. R. Gill was supported by (Horch et al., 1999; Horch & Katz, 2002). the Savoy Foundation. D. Verbich received a Banting and Best scholarship Once the effects of chronic CX546 treatment were established in from the Canadian Institutes of Health Research (CIHR). R. A. McKinney naive mature slices, we examined any possible structural changes was supported by CHIR (MOP133611). The funders had no role in study during LTP. Previous findings have demonstrated that, when applied design, data collection and analysis, decision to publish, or preparation of the acutely, ampakines decrease the threshold for LTP induction (Rex manuscript. The authors declare that no competing interests exist. et al., 2006). We now report that chronically treating hippocampal slice cultures with CX546 leads to greater LTP than in control slices, similarly to the decreased threshold for LTP induction with Abbreviations acute treatment. This decreased threshold and robustness may result AMPA, a-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid; AMPAR, from the enhanced synaptic transmission reliability combined with a-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor; BDNF, the longer-lasting spine enlargement following LTP induction in brain-derived neurotrophic factor; chemLTP, chemically induced long-term CX546-treated slices. Again, this can be explained by the upregula- potentiation; CPP, 3-[(R)-2-carboxypiperazin-4-yl]-propyl-1-phosphonic acid-

EPSC, excitatory postsynaptic current; EPSP, excitatory postsynaptic poten- Horch, W.H., Kruttgen,€ A., Portbury, S.D. & Katz, L.C. (1999) Destabiliza- tial; fEPSP, field excitatory postsynaptic potential; IEI, inter-event interval; tion of cortical dendrites and spines by BDNF. Neuron, 23, 353–364. ISI, inter-stimulus interval; LTP, long-term potentiation; mEPSC, miniature Hotulainen, P. & Hoogenraad, C.C. (2010) Actin in dendritic spines: con- excitatory postsynaptic current; mGFP, membrane-targeted green fluorescent necting dynamics to function. J. Cell Biol., 189, 619–629. protein; NMDAR, N-methyl-D-aspartate receptor; PB, phosphate buffer; PI, Huang, E.J. & Reichardt, L.F. (2001) Neurotrophins: roles in neuronal devel- propidium iodide; PPF, paired-pulse facilitation; TBS, theta-burst stimulation. opment and function. Annu. Rev. Neurosci., 24, 677–736. Ingvar, M., Ambros-Ingerson, J., Davis, M., Granger, R., Kessler, M., Rog- ers, G.A., Schehr, R.S. & Lynch, G. (1997) Enhancement by an ampakine References of memory encoding in humans. Exp. Neurol., 146, 553–559. Irwin, S.A., Galvez, R. & Greenough, W.T. (2000) Dendritic spine structural Alonso, M., Medina, J.H. & Pozzo-Miller, L. (2004) ERK1/2 activation is anomalies in fragile-X mental retardation syndrome. Cereb. Cortex, 10, necessary for BDNF to increase dendritic spine density in hippocampal 1038–1044. 11 – CA1 pyramidal neurons. Learn. Memory, , 172 178. Jaafari, N., Konopacki, F.A., Owen, T.F., Kantamneni, S., Rubin, P., Craig, Arai, A.C. & Kessler, M. (2007) Pharmacology of ampakine modulators: T.J., Wilkinson, K.A. & Henley, J.M. (2013) SUMOylation is required for 8 from AMPA receptors to synapses and behavior. Curr. Drug Targets, , glycine-induced increases in AMPA receptor surface expression (Chem- – 583 602. LTP) in hippocampal neurons. PLoS One, 8, e52345. ’ Baloyannis, S.J. (2009) Dendritic pathology in Alzheimer s disease. Ji, Y., Pang, P.T., Feng, L. & Lu, B. (2005) Cyclic AMP controls BDNF- 283 – J. Neurol. Sci., , 153 157. induced TrkB phosphorylation and dendritic spine formation in mature Baudry, M., Kramar, E., Xu, X., Zadran, H., Moreno, S., Lynch, G., Gall, C. hippocampal neurons. Nat. Neurosci., 8, 164–172. & Bi, X. (2012) Ampakines promote spine actin polymerization, long-term Ji, Y., Lu, Y., Yang, F., Shen, W., Tang, T.T.-T., Feng, L., Duan, S. & Lu, potentiation, and learning in a mouse model of Angelman syndrome. B. (2010) Acute and gradual increases in BDNF concentration elicit dis- 47 – Neurobiol. Dis., , 210 215. tinct signaling and functions in neurons. Nat. Neurosci., 13, 302–309. Biou, V., Brinkhaus, H., Malenka, R.C. & Matus, A. (2008) Interactions Jones, T.A., Klintsova, A.Y., Kilman, V.L., Sirevaag, A.M. & Greenough, between drebrin and Ras regulate dendritic spine plasticity. Eur. J. Neuro- W.T. (1997) Induction of multiple synapses by experience in the visual 27 – sci., , 2847 2859. cortex of adult rats. Neurobiol. Learn. Mem., 68,13–20. Bosch, M. & Hayashi, Y. (2012) Structural plasticity of dendritic spines. Kanju, P.M., Parameshwaran, K., Sims, C., Bahr, B.A., Shonesy, B.C. & 22 – Curr. Opin. Neurobiol., , 383 388. Suppiramaniam, V. (2008) Ampakine CX516 ameliorates functional defi- Bourne, J. & Harris, K.M. (2007) Do thin spines learn to be mushroom cits in AMPA receptors in a hippocampal slice model of protein accumula- 17 – spines that remember? Curr. Opin. Neurobiol., , 381 386. tion. Exp. Neurol., 214,55–61. Branco, T., Staras, K., Darcy, K.J. & Goda, Y. (2008) Local dendritic activ- Katz, B. & Miledi, R. (1968) The role of calcium in neuromuscular facilita- 59 – ity sets release probability at hippocampal synapses. Neuron, , 475 485. tion. J. Physiol., 195, 481–492. Chang, P.K.Y., Boridy, S., McKinney, R.A. & Maysinger, D. (2013) Letroz- Kopec, C.D., Li, B., Wei, W., Boehm, J. & Malinow, R. (2006) Glutamate ole potentiates mitochondrial and dendritic spine impairments induced by receptor exocytosis and spine enlargement during chemically induced b 2013 – -amyloid. J. Aging Res., ,1 11. long-term potentiation. J. Neurosci., 26, 2000–2009. Chapleau, C.A. & Pozzo-Miller, L. (2012) Divergent roles of p75NTR and Kopec, C.D., Real, E., Kessels, H.W. & Malinow, R. (2007) GluR1 links ’ Trk receptors in BDNF s effects on dendritic spine density and morphol- structural and functional plasticity at excitatory synapses. J. Neurosci., 27, 2012 – ogy. Neural Plast., ,1 9. 13706–13718. Comery, T.A., Stamoudis, C.X., Irwin, S.A. & Greenough, W.T. (1996) Kramar, E., Lin, B., Rex, C.S., Gall, C.M. & Lynch, G. (2006) Integrin-dri- Increased density of multiple-dead dendritic spines on medium-sized spiny ven actin polymerization consolidates long-term potentiation. Proc. Natl. neurons of the striatum in rats reared in a complex environment. Neurobiol. Acad. Sci. USA, 103, 5579–5584. 66 – Learn. Memory, ,93 96. Kramar, E.A., Babayan, A.H., Gavin, C.F., Cox, C.D., Jafari, M., Gall, De Paola, V., Arber, S. & Caroni, P. (2003) AMPA receptors regulate C.M., Rumbaugh, G. & Lynch, G. (2012) Synaptic evidence for the dynamic equilibrium of presynaptic terminals in mature hippocampal net- efficacy of spaced learning. Proc. Natl. Acad. Sci. USA, 109, 5121– 6 – works. Nat. Neurosci., , 491 500. 5126. De Roo, M., Klauser, P., Garcia, P.M., Poglia, L. & Muller, D. (2008) Spine Lauterborn, J.C., Truong, G.S., Baudry, M., Bi, X., Lynch, G. & Gall, C.M. dynamics and synapse remodeling during LTP and memory processes. (2003) Chronic elevation of brain-derived neurotrophic factor by ampakines. 169 – Prog. Brain Res., , 199 207. J. Pharmacol. Exp. Ther., 307, 297–305. Deinhardt, K. & Chao, M.V. (2013) Shaping neurons: long and short range Lauterborn, J.C., Pineda, E., Chen, L.Y., Ramirez, E.A., Lynch, G. & Gall, effects of mature and proBDNF signalling upon neuronal structure. Neuro- C.M. (2009) Ampakines cause sustained increases in brain-derived neuro- 76 – pharmacology, , 603 609. trophic factor signaling at excitatory synapses without changes in AMPA Engert, F. & Bonhoeffer, T. (1999) Dendritic spine changes associated with receptor subunit expression. Neuroscience, 159, 283–295. 399 – hippocampal long-term synaptic plasticity. Nature, ,66 70. Levenga, J., Willemsen, R., Mara, D. & De La Rafael, T. (2012) Perturba- Enoki, R., Hu, Y.-L., Hamilton, D. & Fine, A. (2009) Expression of long-term tion of dendritic protrusions in intellectual disability. Prog. Brain Res., plasticity at individual synapses in hippocampus is graded, bidirectional, 197, 153–168. 62 – and mainly presynaptic: optical quantal analysis. Neuron, , 242 253. Lippman, J. & Dunaevsky, A. (2005) Dendritic spine morphogenesis and Ethell, I.M. & Pasquale, E.B. (2005) Molecular mechanisms of dendritic plasticity. J. Neurobiol., 64,47–57. 75 – spine development and remodeling. Prog. Neurobiol., , 161 205. Lu, B. (2003) BDNF and activity-dependent synaptic modulation. Learn. € Gahwiler, B.H., Capogna, M., Debanne, D., McKinney, R.A. & Thompson, Memory, 10,86–98. S.M. (1997) Organotypic slice cultures: a technique has come of age. Lu, B., Pang, P.T. & Woo, N.H. (2005) The yin and yang of neurotrophin 20 – Trends Neurosci., , 471 477. action. Nat. Rev. Neurosci., 6, 603–614. Gill, R., Chang, P.K.Y., Prenosil, G.A., Deane, E.C. & McKinney, R.A. Lynch, G., Rex, C.S., Chen, L.Y. & Gall, C.M. (2008) The substrates of (2013) Blocking brain-derived neurotrophic factor inhibits injury-induced memory: defects, treatments, and enhancement. Eur. J. Pharmacol., 585, 38 hyperexcitability of hippocampal CA3 neurons. Eur. J. Neurosci., , 2–13. – 3554 3566. Machnes, Z.M., Huang, T.C.T., Chang, P.K.Y., Gill, R., Reist, N., Dezsi, G., Hampson, R.E., Rogers, G., Lynch, G. & Deadwyler, S.A. (1998a) Facilita- Ozturk, E., Charron, F., O’Brien, T.J., Jones, N.C., McKinney, R.A. & tive effects of the ampakine CX516 on short-term memory in rats: correla- Szyf, M. (2013) DNA methylation mediates persistent epileptiform activity 18 – tions with hippocampal neuronal activity. J. Neurosci., , 2748 2763. in vitro and in vivo. PLoS One, 8, e76299. Hampson, R.E., Rogers, G., Lynch, G. & Deadwyler, S.A. (1998b) Facilita- Makino, H. & Malinow, R. (2009) AMPA receptor incorporation into syn- tive effects of the ampakine CX516 on short-term memory in rats: apses during LTP: the role of lateral movement and exocytosis. Neuron, enhancement of delayed-nonmatch-to-sample performance. J. Neurosci., 64, 381–390. 18 – , 2740 2747. Maletic-Savatic, M., Malinow, R. & Svoboda, K. (1999) Rapid dendritic Hill, T.C. & Zito, K. (2013) LTP-induced long-term stabilization of individ- morphogenesis in CA1 hippocampal dendrites induced by synaptic activ- 33 – ual nascent dendritic spines. J. Neurosci., , 678 686. ity. Science, 283, 1923–1927. Horch, H.W. & Katz, L.C. (2002) BDNF release from single cells elicits Manabe, T., Wyllie, D.J., Perkel, D.J. & Nicoll, R.A. (1993) Modulation of 5 – local dendritic growth in nearby neurons. Nat. Neurosci., , 1177 1184. synaptic transmission and long-term potentiation: effects on paired pulse

facilitation and EPSC variance in the CA1 region of the hippocampus. spine head protrusions in hippocampal slice cultures. Proc. Natl. Acad. J. Neurophysiol., 70, 1451–1459. Sci. USA, 102, 6166–6171. Mateos, J.M., Luthi,€ A., Savic, N., Stierli, B., Streit, P., G€ahwiler, B.H. & Sekino, Y., Kojima, N. & Shirao, T. (2007) Role of actin cytoskeleton in McKinney, R.A. (2007) Synaptic modiﬁcations at the CA3–CA1 synapse dendritic spine morphogenesis. Neurochem. Int., 51,92–104. after chronic AMPA receptor blockade in rat hippocampal slices. J. Phys- Simmons, D.A., Rex, C.S., Palmer, L., Pandyarajan, V., Fedulov, V., Gall, iol., 581, 129–138. C.M. & Lynch, G. (2009) Up-regulating BDNF with an ampakine rescues Matsuzaki, M., Honkura, N., Ellis-Davies, G.C.R. & Kasai, H. (2004) Struc- synaptic plasticity and memory in Huntington’s disease knockin mice. tural basis of long-term potentiation in single dendritic spines. Nature, Proc. Natl. Acad. Sci. USA, 106, 4906–4911. 429, 761–766. Smart, F.M. & Halpain, S. (2000) Regulation of dendritic spine stability. Matus, A. (2000) Actin-based plasticity in dendritic spines. Science, 290, Hippocampus, 10, 542–554. 754–758. Stamatakou, E., Marzo, A., Gibb, A. & Salinas, P.C. (2013) Activity-depen- McKinney, R.A. (2005) Physiological roles of spine motility: development, dent spine morphogenesis: a role for the actin-capping protein Eps8. plasticity and disorders. Biochem. Soc. T., 33, 1299–1302. J. Neurosci., 33, 2661–2670. McKinney, R.A. (2010) Excitatory amino acid involvement in dendritic spine St€aubli, U., Perez, Y., Xu, F.B., Rogers, G., Ingvar, M., Stone-Elander, S. & formation, maintenance and remodelling. J. Physiol., 588, 107–116. Lynch, G. (1994a) Centrally active modulators of glutamate receptors McKinney, R.A. & Thompson, S.M. (2009) Glutamate regulation of facilitate the induction of long-term potentiation in vivo. Proc. Natl. Acad. dendritic spine form and function. In Squire, L.R. (Ed.), Encyclopedia of Sci. USA, 91, 11158–11162. Beuroscience. Academic Press, Oxford, pp. 905–911. St€aubli, U., Rogers, G. & Lynch, G. (1994b) Facilitation of glutamate recep- McKinney, R.A., Capogna, M., Durr, R., G€ahwiler, B.H. & Thompson, S.M. tors enhances memory. Proc. Natl. Acad. Sci. USA, 91, 777–781. (1999) Miniature synaptic events maintain dendritic spines via AMPA Tada, T. & Sheng, M. (2006) Molecular mechanisms of dendritic spine mor- receptor activation. Nat. Neurosci., 2,44–49. phogenesis. Curr. Opin. Neurobiol., 16,95–101. Mitra, A., Mitra, S.S. & Tsien, R.W. (2012) Heterogeneous reallocation of Tanaka, J.-I., Horiike, Y., Matsuzaki, M., Miyazaki, T., Ellis-Davies, G.C.R. presynaptic efﬁcacy in recurrent excitatory circuits adapting to inactivity. & Kasai, H. (2008) Protein synthesis and neurotrophin-dependent struc- Nat. Neurosci., 15, 250–257. tural plasticity of single dendritic spines. Science, 319, 1683–1687. Ostroff, L.E., Fiala, J.C., Allwardt, B. & Harris, K.M. (2002) Teng, H.K., Teng, K.K., Lee, R., Wright, S., Tevar, S., Almeida, R.D., Ker- Polyribosomes redistribute from dendritic shafts into spines with mani, P., Torkin, R., Chen, Z.-Y., Lee, F.S., Kraemer, R.T., Nykjaer, A. enlarged synapses during LTP in developing rat hippocampal slices. & Hempstead, B.L. (2005) ProBDNF induces neuronal apoptosis via acti- Neuron, 35, 535–545. vation of a receptor complex of p75NTR and sortilin. J. Neurosci., 25, Porrino, L.J., Daunais, J.B., Rogers, G.A., Hampson, R.E. & Deadwyler, 5455–5463. S.A. (2005) Facilitation of task performance and removal of the effects of Toni, N., Buchs, P.A., Nikonenko, I., Bron, C.R. & Muller, D. (1999) LTP sleep deprivation by an ampakine (CX717) in nonhuman primates. PLoS promotes formation of multiple spine synapses between a single axon ter- Biol., 3, 1639–1652. minal and a dendrite. Nature, 402, 421–425. Purpura, D.P. (1974) Dendritic spine ‘dysgenesis’ and mental retardation. Verbich, D., Prenosil, G.A., Chang, P.K.Y., Murai, K.K. & McKinney, R.A. Science, 186, 1126–1128. (2012) Glial glutamate transport modulates dendritic spine head protru- Rex, C.S., Lauterborn, J.C., Lin, C.-Y., Kramar, E.A., Rogers, G.A., Gall, sions in the hippocampus. Glia, 60, 1067–1077. C.M. & Lynch, G. (2006) Restoration of long-term potentiation in middle- Yamada, K.A. (1998) AMPA receptor activation potentiated by the AMPA aged hippocampus after induction of brain-derived neurotrophic factor. modulator 1-BCP is toxic to cultured rat hippocampal neurons. Neurosci. J. Neurophysiol., 96, 677–685. Lett., 249, 119–122. Richards, D.A., Mateos, J.M., Hugel, S., de Paola, V., Caroni, P., G€ahwiler, Yu, L.M.Y. & Goda, Y. (2009) Dendritic signalling and homeostatic adapta- B.H. & McKinney, R.A. (2005) Glutamate induces the rapid formation of tion. Curr. Opin. Neurobiol., 19, 327–335.

© 2014 Federation of European Neuroscience Societies and John Wiley & Sons Ltd European Journal of Neuroscience, 40, 2766–2776 DNA Methylation Mediates Persistent Epileptiform Activity In Vitro and In Vivo

Ziv M. Machnes1, Tony C. T. Huang1, Philip K. Y. Chang1, Raminder Gill1, Nicholas Reist1, Gabriella Dezsi2, Ezgi Ozturk2, Francois Charron1, Terence J. O’Brien2, Nigel C. Jones2, R. Anne McKinney1*☯, Moshe Szyf1*☯

1 Department of Pharmacology and Therapeutics McGill University, McGill University, Montreal, Quebec, Canada, 2 Department of Medicine (Royal Melbourne Hospital), Melbourne Brain Centre, University of Melbourne, Parkville, Victoria, Australia

Abstract

Epilepsy is a chronic brain disorder involving recurring seizures often precipitated by an earlier neuronal insult. The mechanisms that link the transient neuronal insult to the lasting state of epilepsy are unknown. Here we tested the possible role of DNA methylation in mediating long-term induction of epileptiform activity by transient kainic acid exposure using in vitro and in vivo rodent models. We analyzed changes in the gria2 gene, which encodes for the GluA2 subunit of the ionotropic glutamate, alpha-amino-3-hydroxy-5-methyl-4-isoxazole proprionic acid receptor and is well documented to play a role in epilepsy. We show that kainic acid exposure for two hours to mouse hippocampal slices triggers methylation of a 5’ regulatory region of the gria2 gene. Increase in methylation persists one week after removal of the drug, with concurrent suppression of gria2 mRNA expression levels. The degree of kainic acid- induced hypermethylation of gria2 5’ region varies between individual slices and correlates with the changes in excitability induced by kainic acid. In a rat in vivo model of post kainic acid-induced epilepsy, we show similar hypermethylation of the 5’ region of gria2. Inter-individual variations in gria2 methylation, correlate with the frequency and intensity of seizures among epileptic rats. Luciferase reporter assays support a regulatory role for methylation of gria2 5’ region. Inhibition of DNA methylation by RG108 blocked kainic acid-induced hypermethylation of gria2 5’ region in hippocampal slice cultures and bursting activity. Our results suggest that DNA methylation of such genes as gria2 mediates persistent epileptiform activity and inter-individual differences in the epileptic response to neuronal insult and that pharmacological agents that block DNA methylation inhibit epileptiform activity raising the prospect of DNA methylation inhibitors in epilepsy therapeutics.

Citation: Machnes ZM, Huang TCT, Chang PKY, Gill R, Reist N, et al. (2013) DNA Methylation Mediates Persistent Epileptiform Activity In Vitro and In Vivo. PLoS ONE 8(10): e76299. doi:10.1371/journal.pone.0076299 Editor: Jorg Tost, CEA - Institut de Genomique, France Received January 8, 2013; Accepted August 25, 2013; Published October 2, 2013 Copyright: © 2013 Machnes et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: This work was supported by Canadian Institute Health Research grants MS (MOP-42411), RAM (MOP 86724) and by NHMRC project grants to NJ and TO (566544 & 1006077). This work was also supported by the Sackler program in epigenetics and psychobiology at McGill University (MS). MS is a fellow of the Canadian Institute for Advanced research and is supported by a GSK/CIHR professorship in pharmacology. NR held an NSERC USRA. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing interests: The authors have declared that no competing interests exist. * E-mail: [email protected] (MS); [email protected] (RAM) ☯ These authors contributed equally to this work.

Introduction birth [5], can be long lasting, and can affect brain-related phenotypes in both rodents and humans [6]. Furthermore, it Epigenetic mechanisms are known to maintain long-lasting has been previously shown, that inhibition of DNA methyl gene expression programs. These mechanisms involve several transferases (DNMTs) could affect excitatory levels of regulation, including chemical modification of the DNA neurotransmission in the hippocampus [7,8]. These molecule by adding methyl groups at specific positions, often mechanisms may explain the persistence of acquired epilepsy involving the dinucleotide sequence CpG [1]. Such modification long after the original trigger has receded and account for inter- regulates the binding of the different transcription regulators, individual variations in development of epilepsy, in addition to both enhancers and repressors, and the transcription or in the absence of genetic heterogeneity. machinery to control the expression of specific genes [2-4]. Several lines of evidence are consistent with the hypothesis Recent data supports the hypothesis that differential DNA that epilepsy might be mediated by epigenetic processes methylation patterns can form in response to experiences after [9-12]. A popular antiepileptic drug, valproic acid, is a histone-

PLOS ONE | www.plosone.org 1 October 2013 | Volume 8 | Issue 10 | e76299 DNA Methylation and Epilepsy

deacetylase inhibitor [13] that induces DNA demethylation in The proximal promoter was generally unmethylated in all cultured cells [14,15] and in the brain [16]. Furthermore, recent CpG sites (<5%) in both KA treated slices and controls with analysis of hippocampi from mice acutely treated with the minute changes in DNA methylation between KA and controls chemo-convulsant, kainic acid (KA) demonstrated widespread (Figure 1B). However, a 5’ region positioned at -595 to -804 changes in DNA methylation [17]. To test whether DNA upstream of the transcription start site showed measurable methylation plays a causal role in epileptogenesis, however, it levels of methylation and small but nevertheless significant is important to determine whether genes critical for differences in methylation between KA and control (CpG sites epileptogenesis are regulated by DNA methylation in response 35-39). to a transient initial insult and whether these DNA methylation The 5’ region (-766-804) that exhibited consistent changes in changes are essential for epileptogenesis. DNA methylation with KA treatment in slices from individual In this study we tested this hypothesis by examining the mice (boxed in Figure 1A and 1B, CpG sites 37-39) was changes in DNA methylation in the upstream regulatory regions analyzed in silico using TRANSFAC [26]. The analysis of gria2, the gene coding the GluA2 subunit of the AMPA identified several transcription factor binding sites, including receptor. Evidence indicates that the presence of GluA2 CCAAT/enhancer-binding protein beta (C/EBP beta) and the subunit (encoded by gria2) in the heteromeric AMPA receptors Glucocorticoid Receptor (GR), which was previously shown to impermeabilizes it to calcium [18,19], preventing possible localize in Gria2 positive cells in the hippocampus and this calcium-mediated toxicity. Early and lasting downregulation of colocalization was affected by epilepsy [27,28]. gria2 expression observed in epilepsy models suggest that it These results represent DNA methylation levels in a pool of plays a critical role in initiating the epileptogenic cascade, 5 slices from different mice at the end of two hours KA maintaining neuronal hyperexcitability [19,20] and is critical for treatment. However, it is well known that there are inter- the pathophysiology of mesial temporal lobe epilepsy (MTLE), individual differences in both animals and humans in the the most common form of epilepsy acquired in adulthood [21]. liability to developing epilepsy in response to a single brain Furthermore, knockdown of gria2 in young rats resulted in insult [29]. A plausible hypothesis is that variations in DNA seizure-like behavior and neurodegeneration [22]. The methylation of critical genes in the brain are associated with molecular mechanism mediating this phenomenon remains these differences. Genetically homogenous inbred strains are unclear, but epigenetic changes such as REST targeted an ideal system to test this hypothesis. We first tested whether regulation of gria2 expression by histone hypoacetylation in there are differences in the DNA methylation state of the KA response to KA treatment [23], have been implicated. responsive 5’ region in the gria2 promoter between five individual mice. The results depicted in Figure 1C (see mice I, In this study we used a gene-targeted approach by II, III) demonstrate variability in the basal state of methylation of monitoring methylation at specific CpG sites in gria2, a gene the three CpG sites in this region between slices derived from implicated by several lines of data in epileptogenesis, in an in littermate mice (the technical variability in measurement of the vitro mouse and an in vivo rat model of epileptogenesis same mouse is indicated by the standard error), in addition to triggered by KA. We hypothesize that this change in the differences observed between slices derived from mice of methylation is persistent and that inter-individual variation in different litters (Figure 1C mice IV and V). In certain cases gria2 methylation is associated with differences in epileptic where an individual CpG site showed high methylation levels in bursting activity in vitro and the severity of epilepsy developed the basal state (e.g CpG 39 in mice I and IV), we observed in an in vivo model. We then tested whether these methylation decreased methylation after the 2 hour KA treatment. This events functionally down-regulated the gria2 promoter activity observation further highlights the intriguing finding that the DNA and whether the non-nucleoside DNA methyltransferase methylation response 2 hours after exposure to KA is different inhibitor N-Phthalyl-L-tryptophan (RG108) [24] blocked between slices from littermates. methylation of gria2 and epileptogenic bursting. We then determined whether the gria2 gene would remain hypermethylated after KA was removed, serving as a “memory” Results of a transient exposure to the epileptiform inducer. We examined the DNA methylation state of slices that were treated Epileptiform bursts triggered by KA in the with KA for 2 hours and were then maintained in standard hippocampus are associated with inter-individual culture media for one week in absence of the agonist and variability of immediate and persistent changes in a compared it to untreated controls. The results presented in DNA methylation of a 5’ regulatory region of the gria2 Figure 1D show that transiently treated slices exhibited gene. enduring increases in methylation of all three CpGs in the 5’ KA treatment of mature organotypic cultured hippocampus region (Ctrl 6.8±2.7, 15.1±5.36, 20.7±.8 KA 35.5±7.5, 36.6±2.2, slices is a well-established in vitro model for inducing 36.2±5.6 respectively) and a 2.4 fold in the average DNA epileptiform activity [25]. Using this model we examined the methylation in the whole 5’ region of the gria2 promoter (Figure state of DNA methylation of the proximal promoter of the gria2 1A boxed) over control cultures (n=5, p=0.002). The difference gene (Figure 1A) as well as a second region upstream to the in DNA methylation between treated and control slices after 1 proximal promoter (boxed; -766--804) in hippocampal slices week of incubation in drug free medium was higher than after 2 hours of treatment with KA compared to drug-free immediately after exposure to KA for 2 hours (Figure 1C). cultured slices using pyrosequencing (Figure 1B). Importantly, each of the 3 CpGs sites in this region was more

PLOS ONE | www.plosone.org 2 October 2013 | Volume 8 | Issue 10 | e76299 DNA Methylation and Epilepsy

Figure 1. Methylation changes in gria2 5’ region in response to KA induced epileptiform activity. (A) Physical map of the gria2 5’ regulatory region and promoter region. CpG sites are marked by balloons and transcription factor predictions in the analyzed 5’ region are indicated above the physical map. (B) State of methylation of CpG sites in the proximal promoter of the gria2 and the 5’ region in control and KA treated hippocampal slices (n=3 technical replicates); CpG 32 and 34 were not analyzed due to sequence restrictions. (C) Methylation differences between control and KA treated slices derived from the same mouse (n=4 technical replicates for each individual mouse) immediately after KA treatment for two hours, and (D) 1 week after removal of the drug (n=4 technical replicates, SD indicate technical errors). Inter-individual differences are apparent between littermate mice (Mouse I, II, III) and between mice from different litters (mouse IV, V) (E) Gria2 mRNA expression levels in control and KA treated slices as measured by qPCR immediately after KA treatment (2 hours, n=5) and 1 week after removal of the drug (1 week, n=9). Gria2 mRNA levels were normalized to TBP/GAPDH expression based on NormFinder. * p<0.05 ** p<0.01 *** p<0.005 as determined by Mann-Whitney U-test. doi: 10.1371/journal.pone.0076299.g001

PLOS ONE | www.plosone.org 3 October 2013 | Volume 8 | Issue 10 | e76299 DNA Methylation and Epilepsy

methylated in the treatment cultures than in control cultures significant increase in PI positive cells after the 2 hour KA (Figure 1D). Inter-individual differences between slices derived treatment. This difference correlates to less than 2% of the from the same littermate mice in the basal DNA methylation overall number of cells (9266.7±1430.4 as evaluated by state and their response to 2 hours transient KA exposure were manual counting of dissociated neurons in a Neubauer- amplified following 1 week incubation in absence of the drug Improved cell counting chamber). Thirdly, we used BrdU (Figure 1D). incorporation to evaluate cell proliferation in slices of both We therefore tested whether the increase in DNA treated and untreated samples (Figure 3A). There was no methylation in response to KA is associated with reduced significant difference in the total number of proliferating cells expression of gria2 mRNA. In order to accurately evaluate the between the treatment and control samples either immediately levels of gria2 mRNA levels we used NormFinder analysis to after the KA treatment (Ctrl 672.5±11.79, KA 663.7±30.78, test possible reference genes which have previously been p=0.799) or after 1 week of recovery (Ctrl: 230.5±2.23, KA reported to be stable under seizure conditions 260.0±12.39, p=0.290; Figure 3B). These minor differences (Glyceraldehyde-3-phosphate-dehydrogenase - gapdh, TATA between the number of control and KA treated dying and binding protein -tbp, Hypoxanthine phosphoribosyl-transferase proliferating cells suggest that the persistent DNA methylation - hprt1, Neuron specific enolase nse1) [30,31]. We found that changes in response to KA treatment must represent a much all four genes displayed a variability level well below the larger percentage of cells that changed DNA methylation than standard cutoff value previously established of 0.15 (Figure the small number of dying or proliferating cells. S1), indicating the value of any one of these genes as an In summary, inter-individual DNA methylation differences in accurate single reference gene. In order to increase the this region of the gria2 promoter exist in otherwise genetically accuracy of our measurement, we used the combination of identical mice. Most importantly, there are differences in the gapdh and tbp which was indicated by NormFinder to be the responsivity of the DNA methylation state of these CpG sites to most stable. We found high variability in the gria2 mRNA KA insult and these differences are enhanced during the drug- expression levels (n=5 cultures per condition) immediately after free incubation period creating large differences in the long- the 2 hour treatment in the KA-treated slices compared to term DNA methylation state between different individuals with a controls, similar to the methylation response. The average “history” of transient exposure to KA insult. The response of the expression of gria2 mRNA in the transiently treated slices that DNA methylation state does not seem to reflect a variation in were incubated in drug free medium for 1 week was 21.5% cell death or proliferation. lower than in control slices (p=0.006) (Figure 1E), and showed variable levels of reduction. Variations in DNA Methylation associate with For each allele, the methylation profile of a specific site is differences in bursting activity either methylated or not methylated, with each cell contributing The inter-individual differences in DNA methylation as an extreme of either zero or one hundred percent methylation observed in the gria2 promoter region beg the question of to the cumulative measurement (i.e. percent of methylation whether variations in DNA methylation in response to KA are measured gives an indication of the percentage of cells that are associated with differences in long-term electrophysiological methylated in the slice). We considered an important set of activity of the hippocampal slices. We therefore conducted potential confounders namely that the DNA methylation membrane potential recording in single-cell current-clamp changes observed could have been caused by either DNA mode of control and KA-treated slices after one-week recovery synthesis or cell death. These are especially important (Figure 4A,B). Spikes were detected offline and 3 or more considering that the hippocampus is a heterogenic tissue and spikes were grouped into bursts if the inter-spike interval was that changes in gria2 levels have been previously associated smaller than 600 ms. We found that 10 out of 19 slices with neuronal cell death associated with epileptic injuries [32]. exhibited spontaneous bursting activity of the measured Here we used several methods to evaluate the physiological pyramidal neurons after the KA treatment, with an average of condition of our slices. First, we applied Nissl staining (Figure 6.21±3.39 bursts per slice (64.79±30.26 spikes per slice). Two 2A) to examine the structural integrity of the hippocampal control slices out of 16 did have some bursting activity, with an slices. No swelling or vacuolization of cells, typically seen average of 0.25±0.14 bursts (3.56±2.11 spikes per slice, during cell death or apoptosis, was observed suggesting that p=0.029). We then compared the DNA methylation level at the the integrity of the hippocampus was not affected by 2 hours gria2 5’ promoter region in slices that exhibited bursting (n=4) treatment with KA. Secondly, propidium iodine (PI) staining versus non-bursting slices (n=4) (Figure 4C). DNA methylation was performed to test for cell death by necrosis. We evaluated levels were determined in four technical replicates for each cell death induced by exposure to KA for 12h and found severe slice from the same mouse. cell death throughout the slice similar to previous reports Significant hypermethylation was observed in all CpGs in the [33,34]. A positive control for the PI staining was also examined region of gria2 in the bursting slices relative to the conducted by incubating a slice in high KCl solution prior to the non-bursting slices (p<0.001, p<0.001, p=0.017 respectively). staining (Figure 2B). A small amount of PI positive cells was The difference in average DNA methylation of the 3 CpGs in observed in both the control and treated slices (Ctrl 52.0±12.78 this region was highly significant as well (p<0.001). KA 147.5±23.40 p=0.025) immediately after the treatment and We then correlated bursting and DNA methylation in this an even smaller number 1 week post-treatment (Ctrl gria2 5’ region across all individual hippocampal slices. This 43.8±13.72 KA 35.25±6.02 p=0.619; Figure 2C). We found a analysis provided compelling evidence for a correlation

PLOS ONE | www.plosone.org 4 October 2013 | Volume 8 | Issue 10 | e76299 DNA Methylation and Epilepsy

Figure 2. Cell death in hippocampus organotypic culture slices after KA treatment. (A) Nissl staining of mature hippocampus slices displaying normal neuronal organization in both control and KA treated slices 1 week after removal of the drug. (B) Propidium Iodine (PI) staining of the cultures at 2 hours treatment with KA (KA 2h), control (Ctrl 2h) and 1 week after removal of KA (Ctrl 1 week, KA 1 week) to evaluate cell death. 12h KA (KA 12 hours) was used as a positive control for levels of cell death reported in other KA models and 15.7M KCl was used as PI staining positive control (KCl control). PI positive cells are labeled in red. White arrows point at sample PI positive cells (C) Quantification of total number of PI positive cells in the different conditions show a small but significant increase in the number of positive cells immediately after 2h KA treatment compared to control, and no significant change after 1 week recovery (n=4). * p<0.05. doi: 10.1371/journal.pone.0076299.g002

PLOS ONE | www.plosone.org 5 October 2013 | Volume 8 | Issue 10 | e76299 DNA Methylation and Epilepsy

Figure 4. Correlation between epileptiform burst activity and gria2 5’ region CpG methylation levels 1 week after removal of drug. (A) Sample neurons’ membrane potentials recording in current-clamp mode of Control hippocampal slices and (B) 2 hours KA treated slices 1 week after removal of the drug and incubation in drug free medium. Three or more spikes were grouped into bursts if the inter-spike interval was smaller than 600 ms. (C) Average methylation levels of gria2 5’ region CpGs in either slices that exhibit spontaneous bursting and those that don’t exhibit bursting 1 week after 2 hours exposure to KA (n=4). (D) Correlation between bursting frequency and average probe methylation levels (p=0.017, Pearson’s Figure 3. BrdU staining for visualization of proliferating r=0.7995). *p<0.05 ***p<0.005. cells after KA treatment. (A) BrdU incorporation [30] in the doi: 10.1371/journal.pone.0076299.g004 different regions of hippocampus slice cultures (CA1, CA3 and dentate gyrus – DG) after 2 hour treatment (Ctrl 2h, KA 2h) and DNA Methylation of the gria2 promoter in a rat in vivo 1 week recovery (Ctrl -1 week, KA -1 week). Neuronal cells model of chronic epilepsy (NeuN positive) are stained green (B) Quantification of the total We then examined whether gria2 DNA methylation behaves number of proliferating cells (BrdU positive) show no significant in-vivo similarly to what was observed ex-vivo. In order to difference between KA treated slices and control immediately assess this point, we induced Status Epilepticus (SE) in Wistar after 2 hour treatment, or after 1 week of recovery (n=5). rats by intraperitoneal injection of KA. SE was terminated after doi: 10.1371/journal.pone.0076299.g003 four hours with diazepam. The animals were left to recover for ten weeks, at which point we performed a two-week period of between burst frequency and DNA methylation levels at the video-EEG recording (Figure 5A) to record the frequency and examined gria2 promoter region (Pearson correlation p=0.017 severity of spontaneous seizures. Our results revealed high r=0.7995; Figure 4D). variability in the physiological response of the individual rats to the KA treatment (Table 1) ranging from a total of 2 class 0 seizures to 111 seizures (out of which 33 were class IV and V). Collectively, these animals were obtained from 3 litters, but no association existed between severity of the epilepsy and the

PLOS ONE | www.plosone.org 6 October 2013 | Volume 8 | Issue 10 | e76299 DNA Methylation and Epilepsy

Table 1. Seizure number and severity in rat 10 weeks after KA induced SE.

Seizure number and severity Animal no. Class 0 Class 1 Class 2 Class 3 Class 4 1 23 0 4 0 9 2 2 0 0 0 0 3 8 0 4 0 3 4 11 0 3 0 4 5 15 0 5 2 2 6 1 0 0 0 1 7 10 0 2 1 7 8 7 0 4 0 2 9 16 0 9 2 5 10 37 0 4 1 26 11 48 0 4 0 35 12 17 0 25 3 18 13 51 2 24 1 32 doi: 10.1371/journal.pone.0076299.t001

of the video-EEG recording. Similar to the mouse in vitro model, we found very low methylation levels (<10%) and no significant differences throughout the promoter, in both the control and high seizing animals (data not shown), and high inter-individual differences in the methylation state of the gria2 5’ promoter region in different rats. The high convulsive seizure rats exhibited higher DNA methylation in all CpG sites in the Figure 5. DNA Methylation changes in rat gria2 gene gria2 5’ promoter region (Figure 5C). We determined whether promoter 5. ‘ region in epileptic and control rats and their inter-individual differences in seizure behaviour (recorded correlation with seizures. number of convulsive seizures class IV and V) correlated with (A) Representative electrographic recording of a seizure with these inter-individual differences in DNA methylation of the synchronous video using Compumedics software. (B) Physical tested gria2 region (n=13). Our results plotted in Figure 5C map of the rat gria2 promoter (+320 - -664). CpG sites are show that, similar to the in vitro results, there is a highly marked by balloons and predicted transcription factors significant correlation between the state of DNA methylation of common to the mouse and rat 5’ region (-562--664) are gria2 in the hippocampus of individual epileptic rats and their indicated above the physical map. (C) Correlation between bursting frequency and average methylation levels of the gria2 seizure behaviour (p=0.006, Pearson’s r=0.72). Furthermore, 5’ region (p=0.006, Pearson’s r=0.7190, n=13). (D) mRNA we found a significant reduction of 26.3% (p<0.001; n=13) in gria2 expression levels in control and KA treated rats mRNA expression levels of gria2 in the same treated rats measured by qPCR 10 weeks after initial SE. ***p<0.005. relative to the controls, 10 weeks after induction of SE with KA doi: 10.1371/journal.pone.0076299.g005 (Figure 5D), as have been reported previously in the literature, and as we found in the in vitro model. The low seizure rats expressed on average 11.5% higher levels of mRNA than the litter. This observation demonstrates that the severity of the high seizure rats but the difference between the high and low epilepsy which develops from this insult (all animals become seizure groups did not reach statistical significance due to high epileptic following KA-induced SE) varies between different variability in expression in the group and reduced numbers individuals from the same litter similarly to what was observed once the treated group was subdivided to low and high with the mouse hippocampal slices in culture (Figures 1C and seizures. It is important to note that, on review of the video- D and 4D). We used this model to address the following questions: First, would the gria2 promoter in rat hippocampus EEG, no animals experienced Class IV or V in the 24 hours become hypermethylated in response to KA-induced SE and immediately preceding cull, which suggests that the changes in second, whether there are inter-individual differences in DNA methylation are not due to the acute effects of convulsive methylation that correlate with the frequency of seizures? seizures. We focused on a 5’ upstream region (Figure 5B) that was previously shown to regulate gria2 in the rat [35] and contains Gria2 promoter activity is modulated by the state of binding sites for transcription factors found in the differentially methylation of the 5’ region (-528--719) methylated gria2 5’ region in the mouse (Figure 1A). DNA was The gria2 promoter was previously found to contain several isolated from whole hippocampi from the rats after completion regulatory regions that interact with different transcription

PLOS ONE | www.plosone.org 7 October 2013 | Volume 8 | Issue 10 | e76299 DNA Methylation and Epilepsy

factors such as REST and NRF-1 [36]. We utilized transient transfection luciferase reporter assays to determine whether the gria2 5’ region (-528--719) that exhibited differential methylation in rats functionally regulated transcriptional activity and whether that activity was silenced by DNA methylation. We first generated a plasmid that contained the gria2 5’ region (-528--719) (Figure 6A) upstream to a Luciferase reporter gene in a pCpGL-Basic plasmid in either sense or antisense (reverse) direction (scheme in Figure 6A). The vector sequences were previously engineered to have no CpGs and are therefore not methylated by CpG methyltransferases [37], thus any effect of DNA methylation on expression of the reporter gene would be caused by the methylated CpG sites in the inserted test regions. We then subjected the plasmid to either in vitro methylation with the bacterial CpG methyltransferase M.SssI or to mock methylation. Following in vitro methylation of the gria2 5’ region we inserted by ligation the proximal unmethylated gria2 promoter (+320 -528) in the sense orientation (Figure 6A) downstream to the methylated 5’ region. The state of methylation of this construct recapitulates the state of methylation of the gria2 promoter in vivo in high convulsive rats. The ligated patch methylated or unmethylated constructs were directly transfected into SH-Sy5y (human neuronal cell line) cells without passing the plasmid through bacterial cloning to keep the state of methylation of the gria2 5’ region and unmethylated promoter region. Our results show that DNA methylation of the gria2 5’ region silenced luciferase activity when it was inserted in the sense orientation to the promoter but not in the antisense orientation (Figure 6B). Enhancer regions could direct activity in both orientations but promoters act only in the 5‘orientation to transcription start site. We therefore tested whether this 5’ region has independent promoter activity in absence of the proximal gria2 promoter. Our results show that the 5‘ region of gria2 could direct luciferase transcriptional activity independently and that this activity is silenced by methylation of its CpG sites. We further validated the presence of a transcript upstream to the known TSS and downstream to the gria2 5' region by RT- PCR using forward 5' primers residing on the 3’ edge of gria2 5'-region and reverse 3' primers around the known TSS. The amplified Figure 6. DNA methylation of the 5’ region silences the fragment was sequenced to verify that the fragment indeed activity of the gria2 promoter as determined by a transient represents mRNA upstream to the known TSS (see Figure S2 transfection Luciferase reporter assay. (A) Physical map of for physical map and sequence). Together, these results the gria2-Luciferase reporter construct. The 5’ region CpGs in suggest that the gria2 5’ region (-528--719) that we found to be the probe were in-vitro methylated (black lollypop), or mock differentially methylated in response to KA treatment in vivo methylated (empty lollypop). The CpGs in the promoter region has promoter activity in vitro that is silenced by DNA were left unmethylated (empty lollypop). Additional constructs methylation and that methylation of this region silences also the were designed as controls; One containing the full promoter but downstream unmethylated proximal gria2 promoter. with the 5’ region in a reverse direction (Antisense); second, containing only the 5’ region (5’ region); third containing only Inhibition of DNA methylation blocks epileptogenesis the unmethylated promoter (Promoter); forth containing Our results established that new DNA methylation events plasmid without any promoter sequence (Empty vector). (B) occur in response to KA treatment in vitro, that they could The indicated constructs were transfected into SH-Sy5y persist and amplify following removal of KA and they correlate (human neuronal cell line). 48h after transfection the cells were with bursting electrical activity. The remaining crucial question harvested, extracts were prepared and assayed for Luciferase is that of cause and effect; are the DNA methylation events a activity and the values were normalized to total protein result of epileptogenesis and the changing cellular landscape in concentration. Results are average of (n=3) transfections +/- the brain [17], or do the DNA methylation changes triggered by SEM. *p<0.05 **p<0.01 ***p<0.005 determined by a Student t the initial insult play a causal role in persistent bursting in vitro test with Holm-Bonferroni correction. doi: 10.1371/journal.pone.0076299.g006

PLOS ONE | www.plosone.org 8 October 2013 | Volume 8 | Issue 10 | e76299 DNA Methylation and Epilepsy

and seizures in vivo? We used a catalytic inhibitor of DNA methyltransferase RG108 to test whether DNA methylation activity is required for the development of bursting activity in hippocampal slices in response to KA. Mature organotypic hippocampal slices were treated for 2 hours with either the DNA methylation inhibitor RG108 (100µM), KA (6µM), a combination of KA and RG108 or left untreated as controls. The slices were left to recover for one week without the drugs. The results presented in Figure 7A show that while KA treatment induced methylation changes as displayed by the hypermethylation of all the CpG sites in the gria2 5’ region, the DNA methylation inhibitor RG108 blocked this persistent increase in DNA methylation in response to KA. RG108 by itself had no effect on the state of DNA methylation. This suggests that there is no active demethylation or new DNA synthesis of this region in untreated slices that will necessitate the presence of DNA methylation activity to maintain the DNA methylation state. However, DNA methylation activity is required for KA induced DNA hypermethylation. We then determined whether blocking KA induced DNA methylation would also block bursting activity. As seen by the bursting frequency in Figure 7B,C, while RG108 had no significant effect on cultures that were not exposed to KA with an average of 0.58±0.34 bursts (6.25±3.69 spikes, n=12), it completely blocked bursting activity that is normally induced by KA in all of the slices (n=14) showing bursting activity (3.86±3.70 spikes per slice). The percentage of bursting slices were Ctrl 12.5%, KA 53%, RG108 25%, RG108 + KA 0%. When bursting activities and DNA methylation of the gria2 5’ region were correlated across samples of all conditions, there was a significant correlation between gria2 DNA methylation and bursting activity (Figure 7D, p=0.003 Figure 7. Effect of DNA methylation inhibitor RG108 on KA Pearson’s r=0.7763). This data supports the conclusion that induced DNA methylation of the gria2 5’ region and DNA methylation activity is required for the development of epileptiform bursting 1 week following a transient 2 hour epileptogenesis in response to KA and is consistent with the exposure. (A) Methylation differences between control (Ctrl, hypothesis that increased DNA methylation of gria2 5’ region n=4 technical replicates on 3 different mice of origin), KA as well as other putative genes that were not measured in this treated (KA, n=4 technical replicates on 3 slices from different study is involved in epileptogenesis. individual mice), RG108 (RG108 n=4 technical replicates on slices from 2 different mice) and RG108 and KA treated slices Discussion (RG108+KA n=4 technical replicates on 2 different mice of origin) 1 week after removal of the drug and incubation in drug Epilepsy can be triggered by either physical or free medium. (B) Sample neurons’ membrane potentials neurochemical insults to the brain in humans and animals. The recording in current-clamp mode of Ctrl, KA, RG108 and critical question is what are the mechanisms that mediate long- RG108 + KA treated samples. (C) Average spontaneous term consequences of these transient insults, such as chronic bursting activity of hippocampus slices after treatment with the seizures? Another unresolved question is what are the different drugs. Significant increase in bursting activity can be mechanisms responsible for the inter-individual variation in the observed in KA treated slices (n=19) vs. control (n=16). RG108 chronic response to a transient epileptoform insult? Human and combined with KA treatment (RG108+KA, n=14) blocks the animal epilepsy studies show changes in gene expression spontaneous bursting induces by KA treatment (n=19). profiles [38-40], suggesting that mechanisms involved in Treatment with RG108 alone (n=12) does not have any epileptogenesis are registered in the genome and that there significant effect on bursting compared to control slices (n=16). must be genomic mechanisms that mediate the long-lasting Significance between the different conditions in multiple- changes in gene expression in response to transient neural comparisons was calculated using Student t-test with Holm- insult. Bonferroni correction for multiple comparisons. (D) Correlation We used in vitro and in vivo rodent models of between bursting frequency and average probe methylation epileptogenesis triggered by transient KA exposure to test the levels in all the treatments samples (n=12) (Ctrl, KA, RG108 plausibility that DNA methylation mechanisms are involved in and RG108 + KA) (p=0.003, Pearson’s r=0.7763). * p<0.05 ** long-term genomic memory of earlier neural insults. We p<0.01 *** p<0.005. doi: 10.1371/journal.pone.0076299.g007

PLOS ONE | www.plosone.org 9 October 2013 | Volume 8 | Issue 10 | e76299 DNA Methylation and Epilepsy

focused on a single, well-established candidate gene, gria2 as [18,22,41] are consistent with a critical role of DNA methylation a plausible example, although we recognize that it is not in genes such as gria2 in the long-lasting electrophysiological exclusively targeted by KA and that several functional gene response to KA insult. pathways are potentially involved in mediating the insult’s long- The fact that a DNA methylation inhibitor completely blocks lasting effects [17]. Nevertheless, the study of DNA methylation long-term bursting in response to a transient KA treatment alterations in this gene enabled us to establish the first strongly supports DNA methylation involvement in principles of DNA methylation involvement in epileptogenesis. epileptogenesis. Although we used a general DNA methylation A causal role for gria2 in epileptogenesis in young rats has inhibitor, which blocked increased DNA methylation of other been previously established. GluA2 encoded by gria2 is the genes as well, the results nevertheless provided evidence that glutamate receptor subunit that inhibits Ca2+ permeability of the methylation of new sites in the genome is required for AMPA/kainate receptors. GluA2 expression is deregulated in epileptogenesis. The possibility that DNA methylation mediates hippocampal regions after KA treatment [18,41]. Knockdown of the long-term maintenance of an epileptogenic status points to gria2 in hippocampi of young rats induced seizure-like behavior new therapeutic directions in epilepsy. Further studies are [22], supporting a causal role of gria2 down-regulation in required to delineate the genome-wide changes in DNA epileptogenesis. methylation affected by RG108 and their role in Our study of gria2 demonstrates that DNA methylation epileptogenesis, and the possible side effects of RG108 changes are triggered by KA using in vivo (Figure 5) and in treatment on normal tissue. Interestingly, RG108 doesn’t alter vitro models of epileptogenesis (Figure 1) and that they last the basal state of gria2 methylation, suggesting that DNA beyond the initial exposure in vivo (Figure 5) and in vitro methyltransferase activity is required for the KA-mediated (Figure 1) in two rodent species. Moreover, the long-term changes in methylation, but its methylation state is not differences in DNA methylation between treatment and control dynamically maintained in the normal state, since no cell are enhanced following a period of incubation in absence of the division occurs in most of the hippocampus cells. drug, suggesting a consolidation of the long-term response in An excellent correlation was observed between the gria2 the post-insult period. These results conform with the methylation and both the bursting in the in vitro model and the previously suggested hypothesis of the involvement of late- seizures in the in vivo model, suggesting a quantitative onset methylation-mediated gene silencing and relationship between DNA methylation of gria2 and its epileptogenesis [42]. electrophysiological function (Figure 4D, Figure 5C). A well- Our data is consistent with the hypothesis that DNA known enigma is why some individuals are more prone to methylation of the gria2 5’ promoter region is involved in its developing epilepsy in response to an earlier brain trauma. The regulation during epileptogenesis and that DNA methylation is early identification of higher-risk subjects is of paramount required for epileptogenesis. First, the increase in methylation importance when developing protective and preventive observed following one week of incubation in the absence of strategies. One plausible explanation for variances in the the drug (Figure 1D) is associated with a significant reduction susceptibility to epilepsy is genetic differences. However, as in levels of gria2 mRNA expression in the in vitro mouse model differences in epileptogenesis appear among inbred, (Figure 1E) and the methylation levels of the gria2 5’ region in genetically homogenous mice and rats, an alternative epileptic rats in vivo (Figure 5C) is associated with reduced hypothesis of a significant “epigenetic” contribution to this inter- expression of the mRNA (Figure 5D). Second, changes in DNA individual variation should be considered. Remarkably, inter- methylation and mRNA expression, which were measured in individual differences in bursting in response to KA are seen the heterogeneous cell population of the hippocampus were not only in rats in vivo (where one could attribute variations in not due to proliferation or death of either glial or neuronal cells response to slight differences in experimental manipulation of (Figures 2, 3). Third, the promoter activity of gria2 is silenced in animals), but also when brain slices from genetically transient transfection reporter assays by the same methylation homogeneous mice are placed under identical events that were shown to occur in the rat in vivo model (Figure pharmacological, physical and chemical conditions in vitro and 6B). Fourth, the level of gria2 methylation after long-term drug- are provided an identical concentration of drug in solution in free incubation, following a single transient insult correlates random Brownian motion. What we observed is that the initial with the amount of seizures measured in vivo by video-EEG relatively small DNA methylation response and the alterations (Figure 5), and with the levels of bursting in vitro measured by that are consolidated following a long-term incubation period in single-cell patch clamp recording of pyramidal neurons (Figure absence of the drug, exhibited large inter-individual differences 4). Fifth, inhibition of DNA methylation, with a catalytic DNA (Figure 1C, 1D and 4C). Importantly, these differences in methyltransferase inhibitor RG108 (Figure 7B,C) which causes methylation correlated with differences in bursting in vitro inhibition of methylation of gria2 results in parallel inhibition of (Figure 4D) and seizures in vivo (Figure 5C). While we have to bursting (Figure 7A). Moreover, the extent of change in gria2 acknowledge that the averaged methylation differences for methylation in response to RG108 correlates with the reduction CpG -37 to -39 between KA treated and control slices are in bursting (Figure 7D). Although RG108 is a general DNA initially small, it becomes clear that these relatively small methylation inhibitor and most probably impacts the state of changes with high inter-individual variability are further methylation of several genes in addition to gria2, considered amplified with time (Figure 1D). Even though we did not find altogether, these data along with previous publications any correlation between basal methylation level and the level of demonstrating a causal role for gria2 in epileptogenesis change immediately after treatment, or between the basal

PLOS ONE | www.plosone.org 10 October 2013 | Volume 8 | Issue 10 | e76299 DNA Methylation and Epilepsy

methylation levels and susceptibility for long-term bursting cultivation in a dry-air roller incubator at 36 degrees. Each activity, these data provide strong support for the hypothesis organotypic slice was cultured for 3 weeks to allow that there is an inter-individual difference in the DNA differentiation and maturation of neural circuitries. The methylation response to KA. This difference is at least partly morphology of the resulting slices is similar to those observed responsible for the inter-individual differences in in vivo [25]. To induce epileptiform activity, cultures were electrophysiological response and the chronicity of the epileptic incubated with KA (6 µM) for two hours. Hippocampal slices state. were incubated with either 6 µM KA alone, 100µM RG108 There is evidence that DNA methylation differences driven alone (Sigma-Aldrich, St. Louis, MO), or RG108 + KA. Control by experience and not exclusively by genetics also appear in cultures were incubated in fresh medium for the same duration humans. Notably, differences in DNA methylation were shown of time. After 2 hours, the slices were either removed for to emerge between monozygotic twins later in life, as well as immediate processing, or incubated in fresh drug-free medium during gestation [43-49]. These differences were associated for 1 week in the dry-air roller incubator. All culturing methods with psychological disorders such as schizophrenia and bipolar were carried out according to standards of the Canadian disorder [50]. It is therefore plausible that as in rodents, Council on Animal Care (CCAC)-approved protocols. variations in DNA methylation may play an important role in creating inter-individual differences in the emergence of Rat Post-KA Induced Status Epilepticus (SE) epilepsy following brain trauma. If indeed DNA methylation The post-kainic acid induced SE rat model closely resembles mechanisms are involved, there are clear implications on the clinical features, ontogenesis, imaging and diagnosis, prediction, prevention and therapeutics of brain histopathological brain changes of human MTLE [51,52]. At injury and consequent epilepsy in humans. A critical question is 9-10 weeks of age, rats were exposed to 5 mg/kg KA [40] whether some of the DNA methylation differences between which induced SE, a period of sustained seizure activity. Four individuals are present before the exposure to the insult and hours after the initiation of SE (defined as being unresponsive whether there are DNA methylation marks in peripheral tissue to external stimuli, and experiencing convulsive seizures), that could predict susceptibility to developing epilepsy either animals were given an injection of diazepam (5mg/kg ip) to prior to or following a brain insult. Identification of such marks cease the SE. Control animals received saline (0.9%, 2ml/kg will have a profound impact on the treatment and prevention of ip) coupled with diazepam. Animals were then left to recover, epilepsy. and over the ensuing weeks, spontaneous recurrent seizures develop, leading to a diagnosis of epilepsy. Materials and Methods At 8 weeks post-SE, all rats were surgically implanted with extradural recording electrodes, as described in previous Animals publications [53]. Briefly, rats were anaesthetized with All animal handling procedures and experiments were isoflurane (5% induction, 2-3% maintenance) and a midline approved by The Mcgill University Facility Animal Care incision made on the scalp. The connective tissue was Committee (FACC) in accordance with The Canadian Council removed, and 6 burr holes were drilled into the skull. Brass on Animal Care (permit no. 5057) and the University of recording electrodes were then gently screwed into the holes, Melbourne Animal Ethics Committee in accordance with the and held in place by dental cement. Australian Code of Practice for the Care and Use of Animals for At 10 weeks post-SE, all animals underwent 2 weeks of Scientific Purposes (permit no. 0911543). C57BL/6 mice were continuous video-EEG recording (Compumedics, Australia) to bred at the Faculty of Medicine animal facility at McGill record the frequency and severity of spontaneous seizures. At University, Montreal. Male Wistar rats were inbred in the the completion of the 2 weeks recording, animals were rapidly Department of Zoology, University of Melbourne Biological decapitated, and the brains excised and immersed in ice-cold

Research Facility (BRF), and housed from 5 weeks of age in artificial CSF containing 125mM NaCl, 3mM KCl, 6mM MgCl2, the Department of Medicine (RMH) University of Melbourne 1mM CaCl2, 1.25mM NaH2PO4, 25mM NaHCO3, and 10.6mM BRF until the completion of the study. All facilities were under glucose. The hippocampus structure was then microdissected controlled temperature (20 °C) and lighting conditions (12 h using a dissecting microscope. One hippocampus from each light/dark cycle – lights on at 0600 h) and animals had ad animal was taken for RNA extraction, and the other one was libitum access to food and water. snap-frozen for DNA extraction and pyrosequencing analysis of DNA methylation of gria2 gene 5’ regulatory region. Mouse hippocampal Organotypic Hippocampal cultures Mice were sacrificed at postnatal day 6 by cervical DNA Extraction dislocation. The brains were rapidly excised, and the Whole hippocampus tissue was digested overnight with hippocampi microdissected out and sliced to 400µm in 400µg/ml Proteinase K (Roche, Germany) in Lysis Buffer thickness by McIlvain tissue chopper. Each slice was mounted consisting of 10mM Tris (pH 8.0), 400mM NaCl, 2mM EDTA, onto a poly-D-lysine-coated glass cover slide with chicken and 1% SDS (Sigma, St. Louis). The solutions were then plasma (Cocalico, New Jersey) and thrombin (Sigma-Aldrich, saturated with NaCl and the homogenates were centrifuged. St. Louis, MO), and inserted into a plastic tube containing The supernatant was transferred to a new tube and was 500µL media (25% heat-inactivated horse serum, 50% basal treated with 17 µg/ml RNAse A (Fermentas International, medium Eagle, 25% balanced salt solution; pH 7.4) for Canada) and was then subjected to phenol-chloroform

PLOS ONE | www.plosone.org 11 October 2013 | Volume 8 | Issue 10 | e76299 DNA Methylation and Epilepsy

extraction. The DNA was then precipitated out of the aqueous followed by 45 cycles of 10 sec at 95 °C, 10 sec at 56 °C and phase by adding 3 volumes of -80°C 95% ethanol and 20µg 10sec at 72 °C. Amplification was followed by an extension for glycogen (Roche, Germany) as a carrier followed by high 10min at 72 °C and a melting curve cycle. Stability of the speed centrifugation. The resultant DNA pellet was washed in reference genes was determined using NormFinder (http:// 700µl 75% EtOH, centrifuged again and the EtOH was air www.mdl.dk/publicationsnormfinder.htm). For rat gria2 mRNA dried. The dry DNA pellet was then suspended in Tris-EDTA expression, Taqman Multiplex real-time PCR was carried out buffer (10mM Tris-HCl (pH 7.5) and 1mM EDTA (pH 8.0)). using an ABI prism 7000 sequence detector (Applied Biosystems, Warrington, UK), with 2 µL cDNA, 18 µM each RNA Extraction primer, 5 µM probe, and Universal TaqMan 2 × PCR Mastermix RNA was extracted from whole hippocampi using a Trizol (Applied Biosystems) to a final volume of 25 µL. All samples reagent protocol (Invitrogen, Life Technologies, Carlsbad, CA) were run in triplicate. Primers and minor groove binder (MGB) according to the manufacturer’s protocol. Extracted RNA was TaqMan probes for gria2 (Rn00568514_m1) and gapdh treated with DNAse (New England Biolabs) according to the (Rn017775763) were designed by Applied Biosystems (Assay- manufacturer’s protocol. 500 ng of total RNA was used as on-Demand) to avoid genomic amplification. The thermal template for cDNA synthesis using AMV reverse transcriptase cycling conditions used during the PCR were: 2 min at 50 °C, 10 min at 95 °C, followed by 40 cycles at 95 °C for 15 s and (Roche Diagnostics, Laval, QC, Canada) and poly[12]16VN (IDT Technologies) as recommended by the manufacturer. 60 °C for 1 min. gria2 mRNA levels were normalized to gapdh and the values were calculated relative to untreated controls. Bisulfite Pyro Sequencing Nissl Staining Epigentek Bisulfite Kits (Qiagen) were used for bisulfite conversion of DNA as described in the manufacturer’s manual. Organotypic hippocampal slice cultures were fixed in 0.1 M 1µg of genomic DNA was used for a single-conversion phosphate buffer (PB; pH 7.4) with 4% paraformaldehyde reaction. Samples were prepared by performing nested PCR overnight at 4 °C. Slices were then washed several times and with one of the nested primers carrying a 5’ biotin modification. dehydrated for 10 minutes in 30%, 50%, 70% ethanol and then Primers used are listed in Table S1. stained with 0.5% cresyl violet. Further dehydration was All primers designed against bisulfite-converted DNA using performed in 90% and 100% ethanol and 45 minutes in 100% either MethPrimer online software [54] or pyromark assay xylene. Nissl-stained slices were mounted using Permount design software (Qiagen) were synthesized by IDT mounting medium (Fisher Scientific) and imaged with an Technologies. PCR reactions were conducted using Taq upright Zeiss Axioplan 2 microscope equipped with a 2.5x Polymerase (Fermentas International, Canada). Reaction objective and a Zeiss Axiocam high-resolution color digital conditions consisted of initial denaturation/enzyme activation at camera (Carl Zeiss AG). 95°C for 3 min, then 40 cycles of 95°C for 30 sec, annealing for 30 seconds at the temperatures listed below for the different In –Vitro Electrophysiological Recordings and Analysis primers, 72 °C extension for 30 seconds, and completed with a Slice cultures were inserted into a temperature-controlled final extension step at 72 °C for 4 minutes. Annealing chamber (30°C) mounted on an upright microscope (BX53, temperatures were: mouse outside PCR 54.8°C; mouse nested Olympus Corporation, Canada) and continuously perfused with PCR 49.6°C; rat outside PCR 50.8°C; rat nested PCR 55.1°C. Tyrode solution containing: 137 mM NaCl, 2.7 mM KCl, 2.5mM

The nested 5’ primer contained a 5’ biotinylated nucleotide. CaCl2, 2 mM MgCl2, 11.6 mM NaHCO3, 0.4 mM NaH2PO4, and Pyro Sequencing was then performed using a PyroMark Q24 5.6 mM Glucose (pH 7.4). To determine excitability, whole-cell machine (Qiagen) using the protocol described in the patch clamp recordings were performed in current-clamp mode manufacturer’s manual. In brief, biotinylated PCR products from CA1 neurons using borosilicate glass pipettes (3-5 MΩ; were incubated with streptavidin sepharose beads (GE GC150TC; Clark Instruments, UK). All recordings were made healthcare, Canada) for 15min in room temperature. using an Axopatch 200A amplifier (Molecular Devices, Unbiotinylated strand was removed by denaturing with 0.2M Sunnyvale, CA, U.S.A.). Patch pipettes were filled with NaOH, and the beads were washed with 10mM Tris (pH=7.5). intracellular solution containing: 120 mM K-Gluconate, 1 mM Beads were released into 24 well plate containing 25µl EGTA, 10 mM HEPES, 5 mM Mg-ATP, 0.5 mM Na-GTP, 5mM annealing solution and 0.3mM sequencing primer per well. The NaCl, 5mM KCl, 10mM Phosphocreatine, and pH was adjusted plate was loaded onto the PyroMark Q24 machine using to 7.3 with KOH. Cells that did not survive the entire recording specific sequencing assay runs. The results were analyzed were excluded from the final analysis. Spikes were detected with PyroMark® Q24 Software (Qiagen). offline and 3 or more spikes were grouped into bursts if the inter-spike interval was smaller than 600 ms. Real-time Quantitative PCR Mouse mRNA expression levels were quantified by real-time Cell Viability Assay with Propidium Iodide quantitative PCR on a Roche LightCycler® 480 Real-Time Hippocampal slice cultures were tested for cell death. Briefly, PCR System using LightCycler480® DNA SYBRGreen I master slices were incubated with 5 µg/ml of propidium iodide (PI) mix (Roche, Mannheim, Germany). Primers were purchased (Sigma-Aldrich, St. Louis, MO) and imaged (Axioplane Imaging from IDT Technologies and are listed in Table S1. Amplification Fluorescence Microscope, Carl Zeiss MicroImaging GmbH, was performed using the following cycles: 10min at 95 °C Hamburg, Germany). Incubation of mouse hippocampus

PLOS ONE | www.plosone.org 12 October 2013 | Volume 8 | Issue 10 | e76299 DNA Methylation and Epilepsy

culture-slices in 15.7M KCl for 5 minutes was used to induce conducted over-night at 22°C. Plasmids were transformed into cell death in the KCl positive control slices. PIR-1 competent bacteria (Invitrogen, Life Technologies, Carlsbad, CA) and plated on agar plates with Zeocin BrdU Assessment of Cell Proliferation (Invitrogen, Life Technologies, Carlsbad, CA) selective Hippocampal slice cultures were incubated with BrdU 0.5 µM antibiotics (25µg/ml). Positive colonies were picked and grown for 3 days. Following the indicated treatments, slices were fixed overnight in LB media containing 25µg/ml. Plasmid was at in 0.1 M phosphate buffer (PB; pH 7.4) with 4% extracted using Quickclean 5M Miniprep kit (Genscript, USA) paraformaldehyde overnight at 4 °C and washed extensively. and tested for the insert by digestion with the same restriction The slices were then removed from their coverslips for BrdU enzymes and visualization on a 1.2% agarose gel (1.2 gr staining. DNA was denatured using 50% formamide in 2 × SSC agarose in TBE buffer -89mM Tris-Base, 89mM Boric Acid, for 1.5 hour for 65 °C. After washing in 2 × SSC buffer, they 2mM EDTA (pH=8.0)). A positive clone was grown in 500ml LB were incubated at 37 °C in 2N HCL. The slices were then with selective antibiotics (25µg/ml Zeocin) and purified with washed in 0.1 M PB and permeablized for 2 days at 4 °C in 0.1 Qiagen Maxiprep kit, according to the manufacturer manual. M PB, 0.4% Triton X100 and 10% heat-inactivated horse serum. Primary antibodies (sheep a-BrdU, Fitzerald Industries, The plasmid containing the 5’ region was in vitro methylated MA, USA) were applied overnight at 4 °C in 0.1 M PB + 0.4% using M.SssI CpG methyltransferase (New England Biolabs) in Triton X-100 and 5% NHS. Secondary antibodies (donkey a- a 400µl reaction with the following ingredients: 10µg vector, 2µl sheep Alexa-568 secondary antibodies; Molecular Probes, OR, SAM, 40µl Buffer, 3µl Enzyme. After 4 hours of incubation, 2µl USA) were then added overnight at concentration of 1:300. of SAM and 1µl enzyme were added and the reaction mixture Dentate gyrus region of the hippocampal slices were imaged was incubated for additional 4 hours. Mock methylation on an upright confocal microscope (Leica DM6000 B upright followed the same protocol with exclusion of the M.SssI microscope; HCX PL APO 63× NA 1.4 oil immersion objective). enzyme. Methylated plasmid was cleaned using Quickclean 5M PCR purification kit (Genscript, USA). Nuclei Separation from Slice Culture We then inserted an unmethylated PCR amplified fragment Nuclei were separated from mouse hippocampus slice using containing the gria2 promoter: Insert and vector were digested the protocol described by Matevossian et al., 2008 [55] with the with BamHI and NcoI restriction enzymes (Fermentas following two modification: First, each sample consisted of 5 International, Canada) in 2xTango buffer, by incubation for 4 slices and lyzed in 1ml of lysis buffer; Second, the nuclei were hours at 37c. Digestion products were cleaned using suspended in a final volume of 500 µL. Nuclei were mixed 1:1 Quickclean 5M PCR purification kit (Genscript, USA). Vector with Trypan Blue Stain (Gibco, USA) and quantified on a and the gria2 promoter insert were ligated in a 1:2 molecule Neubauer-Improved cell counting chamber (LaborOptic, UK). number ratio with T4 Ligase (Frementas International, Canada) based on the manufacturer recommendation. Incubation was Luciferase cCnstruct Assembly conducted over-night at 22 °C. For the promoter plasmid, the Rat gria2 5’ region and promoter (Figure 6 for physical map) promoter region PCR fragment was ligated under the same were amplified from 5ng of rat genomic DNA by a PCR reaction conditions to an empty pCpGL-Basic plasmid. After inserting using the following primers: Promoter region Forward - TTT the 'unmethylated' DNA, the ligated plasmids were purified GGA TCC GAA GCT AAA GTT CAc agt ttt ggg ag; Reverse - from 1.0% agarose gel using Qiaquick Gel Extraction kit TTT CCA TGG AAT TAG ATC CTC TGC ATT GTG AG; Probe region (sense) Forward - TTT CTG CAG TTC AAG AGC AAT resulting in removal of excess unligated unmethylated DNA. CCA CAG G; Reverse - TTT GGA TCC CTA TGA TGC AAG The “patch” methylated constructs were transfected directly CAT AAT TCC; Probe region (antisense) Forward - TTT GGA into SH-Sy5y neuroblastoma cells without cloning in bacterial TCC TTC AAG AGC AAT CCA CAG G; Reverse - TTT CTG cells to maintain the state of methylation of the patch. During CAG CTA TGA TGC AAG CAT AAT TCC according to the cloning in bacterial cells the CpG methylation pattern is lost following cycle: 5min at 95°C followed by 40 cycles of 35 sec at since E. coli does not harbor the DNA methyltransferase 95°C, 35 sec at 56°C and 70sec/40sec (respectively) at 72°C. required to copy CpG methylation during replication. Amplification was followed by 10 min extension at 72°C. The PCR product was purified using Quickclean 5M PCR Cell Line and Transfection purification kit (Genscript, USA). PCR product containing the 5’ SH-Sy5y human neuroblastoma cells (Sigma, USA) were region and pCpGL-Basic plasmid (promoterless Luciferase grown in F12:DMEM 1:1 media (Gibco, Invitrogen, Life reporter plasmid lacking any CpGs in its sequence), generously Technologies, Carlsbad, CA) in controlled environment (5% donated by Dr. Rehi’s lab, University hospital, Regensburg, CO , 37 °C). Cells were transfected using X-tremeGene HD Germany [37]; were digested with PstI and BamHI restriction 2 enzymes (Fermentas International, Canada) in 2xTango buffer, (Roche, Germany) according to the manufacturer by incubation for 4 hours at 37°C. Digestion products were recommendations. In brief, 100,000 cells were platted on 6 well cleaned using Quickclean 5M PCR purification kit (Genscript, plates and grown to 80% confluency at the day of transfection. USA). Vector and insert were ligated in a 1:2 molecule number 1µg of vector was mixed with the transfection reagent (1:3 ratio with T4 Ligase (Frementas International, Canada) based ratio) and added to the cell media. Cells were harvested 48 on the manufacturer recommendation. Incubation was hour post transfection.

PLOS ONE | www.plosone.org 13 October 2013 | Volume 8 | Issue 10 | e76299 DNA Methylation and Epilepsy

Luciferase Activity Assay correlated with the number of convulsive seizures experienced, Cells were washed with PBS (Invitrogen, Life Technologies, or number of recorded bursts in the slices, using Pearson’s Carlsbad, CA) and harvested using a cell scraper (Sarstedt, correlation analysis. Significance between the different Germany). Cells were spun down (5min at 1000rpm) and lysed conditions in multiple-comparisons was calculated using the in 30µl lysis buffer (25mM Tris-Phosphate (pH=7.8), 10% Student t-test with Holm-Bonferroni correction for multiple Glycerol, 1% Triton-X, 1mg/ml BSA, 2.5mM EDTA (pH=8.0) comparisons [57]. All analyses were performed using Graph and 1x Complete-mini EDTA free protein inhibitors (Roch, pad Prism or Microsoft Excel software, and results were Germany)) for 5 min on ice. The lysate was spun for 5 min at considered statistically significant when p<0.05. 13,000 rpm and the supernatant was removed for activity evaluation. 10µl of the lysate was added to each well with Supporting Information 100µl Luciferase assay substrate (Promega, USA). The reactions were read using FluoStar Optime (BMG labtech, Figure S1. NormFinder analysis of expression stability of Offenburg, Germany) and the results were normalized to the several reference genes. Variability values calculated by total protein concentration per sample. NormFinder for 4 reference genes and a recommended The protein concentration was measured using Bradford combined reference. Note that all reference genes are below reaction assay (Biorad, USA) according to the manufacturer the acceptable variability cutoff previously defined by recommendations, and read using DU730 UV/Vis Wierschke et al. 2010 [31] as v<0.15. spectrophotometer (Beckman Coulter, USA). (TIF)

Validation of Upstream Initiation Site Figure S2. Upstream gria2 initiation site. An RT-PCR To test whether a 5' transcript is initiated upstream to the reaction was used to validate the existence of an alternate known gria2 TSS we designed 5' forward primers from the 3’ upstream gria2 TSS in vivo. (A) Physical map of the rat gria2 region of the proposed regulatory region (“5’ region”) where a promoter and 5’ region. CpG sites are marked by balloons and novel transcript would initiate if indeed this promoter is numbers indicate distance from previously known TSS. Primers functional in vivo 5'-GGAATTATGCTTGCATCATAG and 3' spanning from the 3’ end of the gria2 5’ region (red arrow reverse primer corresponding to the region of the known TSS marked F) to the 5’ end of the known TSS (red arrow marked 5'- AATATCAGCACCCTCCCAT (see physical map in Figure R) were used to amplify DNAse-treated rat hippocampus RNA. S2), RT-PCR reaction was conducted on DNAse treated (1 µg (B) PCR product was produced only after reverse-transcription per 20 µl for 1 h at 37°C) rat hippocampus RNA to exclude the (cDNA) and not when conducting the PCR reaction directly on possibility that we amplified genomic DNA rather than RNA. A the DNAse-treated RNA (RT -- ). (C) The PCR product was PCR product was detected using the following cycle: 5 min at subjected to Sanger sequencing and aligned to the gria2 98 °C followed by 46 cycles of 35 sec at 98 °C, 35 sec at 56 °C promoter, from the 5’ region to the previously reported TSS (C). and 30 sec at 72 °C. Amplification was followed by 10min (TIF) extension at 72 °C. The PCR was run, visualized andeluted from an E-Gel Clonewell 2% SYBR safe gel (Invitrogen, Israel) Table S1. PCR primers. ® using the E-Gel Agarose electrophoresis system (Invitrogen, (DOCX) Israel). The PCR product was subjected to Sanger sequencing for confirmation of the sequence and its alignment to the DNA Acknowledgements upstream of the known TSS using Multialin software [56]. We would like to thank Kim Powell for her excellent technical Statistical Analysis assistance. For the DNA methylation results, we used Mann-Whitney U- test. For mRNA expression analysis, a Student’s t-test Author Contributions compared between treatment groups. To measure correlation between DNA methylation and either bursting or seizures, we Conceived and designed the experiments: ZMM RAM NCJ MS. first averaged the methylation levels at all 5 CpG sites in the rat Performed the experiments: ZMM TCTH PKYC RG NR FC TO gria2 5’ region and the 3 CpG sites in the mouse gria2 5’ EO GD. Analyzed the data: ZMM TCTH PKYC RG NR TO EO region. The average DNA methylation value per animal was GD. Wrote the manuscript: ZMM MS NJ RAM.

References

1. Razin A (1998) CpG methylation, chromatin structure and gene concepts in MeCP2 research. Neural Plast, 2012: 415825. PubMed: silencing-a three-way connection. EMBO J 17: 4905-4908. doi:10.1093/ 22474603 emboj/17.17.4905. PubMed: 9724627. 4. Portela A, Liz J, Nogales V, Setién F, Villanueva A et al. (2013) DNA 2. Szyf M (2009) Epigenetics, DNA Methylation, and Chromatin Modifying methylation determines nucleosome occupancy in the 5'-CpG islands Drugs. Annu Rev Pharmacol Toxicol 49: 243-263. doi:10.1146/ of tumor suppressor genes. Oncogene. PubMed: 23686312 annurev-pharmtox-061008-103102. PubMed: 18851683. 5. Weaver IC, Cervoni N, Champagne FA, D’Alessio AC, Sharma S et al. 3. Zachariah RM, Rastegar M (2012) Linking epigenetics to human (2004) Epigenetic programming by maternal behavior. Nat Neurosci 7: disease and Rett syndrome: the emerging novel and challenging 847-854. doi:10.1038/nn1276. PubMed: 15220929.

PLOS ONE | www.plosone.org 14 October 2013 | Volume 8 | Issue 10 | e76299 DNA Methylation and Epilepsy

6. Szyf M (2011) The early life social environment and DNA methylation: server: PROMO and MALGEN. Nucleic Acids Res 31: 3651-3653. doi: DNA methylation mediating the long-term impact of social 10.1093/nar/gkg605. PubMed: 12824386. environments early in life. Epigenetics 6: 971-978. doi:10.4161/epi. 27. Patel A, Bulloch K (2003) Type II glucocorticoid receptor 6.8.16793. PubMed: 21772123. immunoreactivity in the mossy cells of the rat and the mouse 7. Nelson ED, Kavalali ET, Monteggia LM (2008) Activity-Dependent hippocampus. Hippocampus 13: 59-66. doi:10.1002/hipo.10045. Suppression of Miniature Neurotransmission through the Regulation of PubMed: 12625458. DNA Methylation. J Neurosci 28: 395-406. doi:10.1523/JNEUROSCI. 28. Bao GS, Cheng XQ, Hua Y, Wang ZD, Liu ZG (2011) Changes of 3796-07.2008. PubMed: 18184782. glucocorticoid receptor mRNA expression in basolateral amygdale- 8. Levenson JM, Roth TL, Lubin FD, Miller CA, Huang IC et al. (2006) kindled rats. Chin Med J (Engl) 124: 2622-2627. PubMed: 22040414. Evidence That DNA (Cytosine-5) Methyltransferase Regulates Synaptic 29. French JA, Williamson PD, Thadani VM, Darcey TM, Mattson RH et al. Plasticity in the Hippocampus. J Biol Chem 281: 15763-15773. doi: (1993) Characteristics of medial temporal lobe epilepsy: I. Results of 10.1074/jbc.M511767200. PubMed: 16606618. history and physical examination. Ann Neurol 34: 774-780. doi:10.1002/ 9. Qureshi IA, Mehler MF (2010) Epigenetic mechanisms underlying ana.410340604. PubMed: 8250525. human epileptic disorders and the process of epileptogenesis. 30. Pernot F, Dorandeu F, Beaup C, Peinnequin A (2010) Selection of Neurobiol Dis 39: 53-60. doi:10.1016/j.nbd.2010.02.005. PubMed: reference genes for real-time quantitative reverse transcription- 20188170. polymerase chain reaction in hippocampal structure in a murine model 10. Kobow K, Blümcke I (2011) The methylation hypothesis: Do epigenetic of temporal lobe epilepsy with focal seizures. J Neurosci Res 88: chromatin modifications play a role in epileptogenesis? Epilepsia 52: 1000-1008. PubMed: 19937810. 15-19. doi:10.1111/j.1528-1167.2011.03145.x. PubMed: 21732935. 31. Wierschke S, Gigout S, Horn P, Lehmann TN, Dehnicke C et al. (2010) 11. Kobow K, Auvin S, Jensen F, Löscher W, Mody I et al. (2012) Finding a Evaluating reference genes to normalize gene expression in human better drug for epilepsy: Antiepileptogenesis targets. Epilepsia 53: epileptogenic brain tissues. Biochem Biophys Res Commun 403: 1868-1876. doi:10.1111/j.1528-1167.2012.03716.x. PubMed: 385-390. doi:10.1016/j.bbrc.2010.10.138. PubMed: 21081112. 23061663. 32. Grooms SY, Opitz T, Bennett MVL, Zukin RS (2000) Status epilepticus 12. Kobow K, Jeske I, Hildebrandt M, Hauke J, Hahnen E et al. (2009) decreases glutamate receptor 2 mRNA and protein expression in Increased Reelin Promoter Methylation Is Associated With Granule Cell hippocampal pyramidal cells before neuronal death. Proc Natl Acad Sci Dispersion in Human Temporal Lobe Epilepsy. J Neuropathol Exp U S A 97: 3631-3636. doi:10.1073/pnas.97.7.3631. PubMed: Neurol 68: 356-364. doi:10.1097/NEN.0b013e31819ba737. PubMed: 10725374. 19287316. 33. Routbort MJ, Bausch SB, McNamara JO (1999) Seizures, cell death, 13. Phiel CJ, Zhang F, Huang EY, Guenther MG, Lazar MA et al. (2001) and mossy fiber sprouting in kainic acid-treated organotypic Histone deacetylase is a direct target of valproic acid, a potent hippocampal cultures. Neuroscience 94: 755-765. doi:10.1016/ anticonvulsant, mood stabilizer, and teratogen. J Biol Chem 276: S0306-4522(99)00358-9. PubMed: 10579566. 36734-36741. doi:10.1074/jbc.M101287200. PubMed: 11473107. 34. Holopainen IE, Järvelä J, Lopez-Picon FR, Pelliniemi LJ, Kukko- 14. Detich N, Bovenzi V, Szyf M (2003) Valproate induces replication- Lukjanov TK (2004) Mechanisms of kainate-induced region-specific independent active DNA demethylation. J Biol Chem 278: neuronal death in immature organotypic hippocampal slice cultures. 27586-27592. doi:10.1074/jbc.M303740200. PubMed: 12748177. Neurochem Int 45: 1-10. doi:10.1016/j.neuint.2004.01.005. PubMed: 15. Milutinovic S, D’Alessio AC, Detich N, Szyf M (2007) Valproate induces 15082216. widespread epigenetic reprogramming which involves demethylation of 35. Myers SJ, Peters J, Huang Y, Comer MB, Barthel F et al. (1998) specific genes. Carcinogenesis 28: 560-571. PubMed: 17012225. Transcriptional Regulation of the GluR2 Gene: Neural-Specific 16. Dong E, Guidotti A, Grayson DR, Costa E (2007) Histone Expression, Multiple Promoters, and Regulatory Elements. J Neurosci hyperacetylation induces demethylation of reelin and 67-kDa glutamic 18: 6723-6739. PubMed: 9712644. acid decarboxylase promoters. Proc Natl Acad Sci U S A 104: 36. Traynelis SF, Wollmuth LP, McBain CJ, Menniti FS, Vance KM et al. 4676-4681. doi:10.1073/pnas.0700529104. PubMed: 17360583. (2010) Glutamate Receptor Ion Channels: Structure, Regulation, and 17. Miller-Delaney SF, Das S, Sano T, Jimenez-Mateos EM, Bryan K et al. Function. Pharmacol Rev 62: 405-496. doi:10.1124/pr.109.002451. (2012) Differential DNA methylation patterns define status epilepticus PubMed: 20716669. and epileptic tolerance. J Neurosci 32: 1577-1588. doi:10.1523/ 37. Klug M, Rehli M (2006) Functional analysis of promoter CpG JNEUROSCI.5180-11.2012. PubMed: 22302800. methylation using a CpG-free luciferase reporter vector. Epigenetics 1: 18. Friedman LK, Pellegrini-Giampietro DE, Sperber EF, Bennett MV, 127-130. doi:10.4161/epi.1.3.3327. PubMed: 17965610. Moshé SL et al. (1994) Kainate-induced status epilepticus alters 38. Xi ZQ, Xiao F, Yuan J, Wang XF, Wang L et al. (2009) Gene glutamate and GABAA receptor gene expression in adult rat expression analysis on anterior temporal neocortex of patients with hippocampus: an in situ hybridization study. J Neurosci 14: 2697-2707. intractable epilepsy. Synapse 63: 1017-1028. doi:10.1002/syn.20681. PubMed: 8182436. PubMed: 19623530. 19. Pellegrini-Giampietro DE, Gorter JA, Bennett MV, Zukin RS (1997) The 39. Jamali S, Bartolomei F, Robaglia-Schlupp A, Massacrier A, Peragut JC GluR2 (GluR-B) hypothesis: Ca(2+)-permeable AMPA receptors in et al. (2006) Large-scale expression study of human mesial temporal neurological disorders. Trends Neurosci 20: 464-470. doi:10.1016/ lobe epilepsy: evidence for dysregulation of the neurotransmission and S0166-2236(97)01100-4. PubMed: 9347614. complement systems in the entorhinal cortex. Brain 129: 625-641. doi: 20. Tanaka H, Grooms SY, Bennett MV, Zukin RS (2000) The AMPAR 10.1093/brain/awl001. PubMed: 16399808. subunit GluR2: still front and center-stage. Brain Res 886: 190-207. doi: 40. Arion D, Sabatini M, Unger T, Pastor J, Alonso-Nanclares L et al. 10.1016/S0006-8993(00)02951-6. PubMed: 11119696. (2006) Correlation of transcriptome profile with electrical activity in 21. Scharfman HE, Gray WP (2007) Relevance of seizure-induced temporal lobe epilepsy. Neurobiol Dis 22: 374-387. doi:10.1016/j.nbd. neurogenesis in animal models of epilepsy to the etiology of temporal 2005.12.012. PubMed: 16480884. lobe epilepsy. Epilepsia 48 Suppl 2: 33-41. doi:10.1111/j. 41. Suzuki F, Hirai H, Onteniente B, Riban V, Matsuda M et al. (2000) 1528-1167.2007.01065.x. PubMed: 17571351. Long-term increase of GluR2 α-amino-3-hydroxy-5-methylisoxazole-4- 22. Friedman LK, Koudinov AR (1999) Unilateral GluR2(B) hippocampal propionate receptor subunit in the dispersed dentate gyrus after knockdown: a novel partial seizure model in the developing rat. J intrahippocampal kainate injection in the mouse. Neuroscience 101: Neurosci 19: 9412-9425. PubMed: 10531445. 41-50. doi:10.1016/S0306-4522(00)00359-6. PubMed: 11068135. 23. Huang Y, Doherty JJ, Dingledine R (2002) Altered histone acetylation 42. Kobow K, Blümcke I (2012) The emerging role of DNA methylation in at glutamate receptor 2 and brain-derived neurotrophic factor genes is epileptogenesis. Epilepsia 53: 11-20. doi:10.1111/epi.12031. PubMed: an early event triggered by status epilepticus. J Neurosci 22: 23216575. 8422-8428. PubMed: 12351716. 43. Bell JT, Saffery R (2012) The value of twins in epigenetic epidemiology. 24. Brueckner B, Boy RG, Siedlecki P, Musch T, Kliem HC et al. (2005) Int J Epidemiol, 41: 140–50. PubMed: 22253312. Epigenetic reactivation of tumor suppressor genes by a novel small- 44. Bell JT, Spector TD (2011) A twin approach to unraveling epigenetics. molecule inhibitor of human DNA methyltransferases. Cancer Res 65: Trends Genet 27: 116-125. doi:10.1016/j.tig.2010.12.005. PubMed: 6305-6311. doi:10.1158/0008-5472.CAN-04-2957. PubMed: 16024632. 21257220. 25. Gähwiler BH, Capogna M, Debanne D, McKinney RA, Thompson SM 45. Fraga MF, Ballestar E, Paz MF, Ropero S, Setien F et al. (2005) (1997) Organotypic slice cultures: a technique has come of age. Trends Epigenetic differences arise during the lifetime of monozygotic twins. Neurosci 20: 471-477. doi:10.1016/S0166-2236(97)01122-3. PubMed: Proc Natl Acad Sci U S A 102: 10604-10609. doi:10.1073/pnas. 9347615. 0500398102. PubMed: 16009939. 26. Farré D, Roset R, Huerta M, Adsuara JE, Roselló L et al. (2003) 46. Nguyen A, Rauch TA, Pfeifer GP, Hu VW (2010) Global methylation Identification of patterns in biological sequences at the ALGGEN profiling of lymphoblastoid cell lines reveals epigenetic contributions to

PLOS ONE | www.plosone.org 15 October 2013 | Volume 8 | Issue 10 | e76299 DNA Methylation and Epilepsy

autism spectrum disorders and a novel autism candidate gene, RORA, 52. Jupp B, Williams J, Binns D, Hicks RJ, Cardamone L et al. (2012) whose protein product is reduced in autistic brain. FASEB J 24: Hypometabolism precedes limbic atrophy and spontaneous recurrent 3036-3051. doi:10.1096/fj.10-154484. PubMed: 20375269. seizures in a rat model of TLE. Epilepsia 53: 1233-1244. doi:10.1111/j. 47. Kaminsky ZA, Tang T, Wang SC, Ptak C, Oh GH et al. (2009) DNA 1528-1167.2012.03525.x. PubMed: 22686573. methylation profiles in monozygotic and dizygotic twins. Nat Genet 41: 53. Jones NC, Salzberg MR, Kumar G, Couper A, Morris MJ et al. (2008) 240-245. doi:10.1038/ng.286. PubMed: 19151718. Elevated anxiety and depressive-like behavior in a rat model of genetic 48. Petronis A (2006) Epigenetics and twins: three variations on the theme. generalized epilepsy suggesting common causation. Exp Neurol 209: Trends Genet 22: 347-350. doi:10.1016/j.tig.2006.04.010. PubMed: 254-260. doi:10.1016/j.expneurol.2007.09.026. PubMed: 18022621. 16697072. 54. Li LC, Dahiya R (2002) MethPrimer: designing primers for methylation 49. Singh SM, Murphy B, O’Reilly R (2002) Epigenetic contributors to the PCRs. Bioinformatics 18: 1427-1431. doi:10.1093/bioinformatics/ discordance of monozygotic twins. Clin Genet 62: 97-103. doi: 18.11.1427. PubMed: 12424112. 10.1034/j.1399-0004.2002.620201.x. PubMed: 12220446. 55. Matevossian A, Akbarian S (2008) Neuronal nuclei isolation from 50. Dempster EL, Pidsley R, Schalkwyk LC, Owens S, Georgiades A et al. human postmortem brain tissue. J Vis Exp. PubMed: 19078943 (2011) Disease-associated epigenetic changes in monozygotic twins 56. Corpet F (1988) Multiple sequence alignment with hierarchical discordant for schizophrenia and bipolar disorder. Hum Mol Genet 20: clustering. Nucleic Acids Res 16: 10881-10890. doi:10.1093/nar/ 4786-4796. doi:10.1093/hmg/ddr416. PubMed: 21908516. 16.22.10881. PubMed: 2849754. 51. Morimoto K, Fahnestock M, Racine RJ (2004) Kindling and status 57. Ludbrook J (1998) MULTIPLE COMPARISON PROCEDURES epilepticus models of epilepsy: rewiring the brain. Prog Neurobiol 73: UPDATED. Clin Exp Pharmacol Physiol 25: 1032-1037. doi:10.1111/j. 1-60. doi:10.1016/j.pneurobio.2004.03.009. PubMed: 15193778. 1440-1681.1998.tb02179.x. PubMed: 9888002.

PLOS ONE | www.plosone.org 16 October 2013 | Volume 8 | Issue 10 | e76299 View Article Online / Journal Homepage / Table of Contents for this issue

PAPER www.rsc.org/loc | Lab on a Chip Chamber and microﬂuidic probe for microperfusion of organotypic brain slices†

Arthur Queval,ab Nageswara R. Ghattamaneni,ab Cecile M. Perrault,ab Raminder Gill,c Maryam Mirzaei,ab R. Anne McKinney*c and David Juncker*abd

Received 12th August 2009, Accepted 22nd October 2009 First published as an Advance Article on the web 19th November 2009 DOI: 10.1039/b916669f

Microfluidic systems are increasingly being used for the culture and study of dissociated cells because they require only minute amounts of materials while enabling drug screening and chemotaxis studies down to the single cell level. However, the culture of organized tissue, such as brain slices, has been more difficult to adapt to microfluidic devices. Here, wepresent a microfluidic system, comprising (i) a perfusion chamber for the culture of organotypic slices that is compatible with high resolution imaging on inverted microscopes, and (ii) a novel transparent microfluidic probe (MFP) for the localized microperfusion of the brain tissue. The MFP is made in poly(dimethylsiloxane), features six micrometre-scale apertures and can be assembled within a few hours in a standard laboratory. Each aperture can indiscriminately be used either for the injection or aspiration of solutions, giving rise to many possible combinations. The MFP was successfully used for the perfusion of a small number of cells in a brain slice with concurrent confocal fluorescence imaging of the perfused dye and sub-cellular structures within the tissue.

Introduction preferentially runs along the sagittal plan, to which the cutting plane is aligned, and allows obtaining tissue sections with a stereotypic Tissue slices derived from many different brain areas are important neuronal microcircuit architecture that reflects the connectivity of tools to study the morphological, pharmacological and physiolog- the hippocampus, and which is maintained during the thinning of ical properties of neuronal circuits. Acute slices are used immedi- the slices. Briefly, after cutting the slices on a tissue chopper, they are ately after slicing, and are thus easy to prepare, retain the fixed onto a specially made 12 24 mm2 large coverslip using cytoarchitecture of theirbrain region of origin and areused by many chicken plasma and thrombin, and then cultivated and rotated in 1,2 electrophysiologists and biochemists. Unfortunately these slices a dedicated rollertube system. The roller-tubes are half-filled with can only survive for a short period of time, yet many biological media, and inclined so that the slice is periodically immersed and processes only occur within the course of days to weeks. Organo- exposed to air allowing an oxygen nutrient interchange. After 2 typic slice culture was developed to allow pieces of brain tissue to be weeks of culture the slices have reconnected and thinned to two to Published on 19 November 2009. Downloaded by McGill University 10/04/2015 14:12:50. maintained in vitro for periods of many weeks and up to months three cells thick and are ready for experimentation.1 while still maintaining the cytotarchitecture of brain region (hence The major advantage, and rationale, for using such cultures is the name organotypic cultures), and which are particularly useful that the tissue slices uphold some of the functions of the intact 3,4 for long-term studies. During preparation of the cultures certain brain while being amenable to high resolution imaging using afferant neuronal fibers are cut and therefore some degenerate in confocal microscopy. They thus permit the visualization of slice cultures, yet surprisingly few changes in synaptic rearrange- processes and events with subcellular resolution of structures, ment have been observed. The cultured brain slices, for the most such as the hippocampus, that are buried inside the CNS and part, will develop structural and functional activities that are similar which cannot be visualized with high accuracy in live animals. 5,6 to those they would develop in the brain. Preserving the viability of organotypic slices for long term analysis A specific organotypic slice cultivation method called the roller- over hours, or sometimes even days requires a physiological drum technique, which was established by Gahwiler€ in 1981, allows environment which is provided by a perfusion chamber used to thinning the brain slices of the hippocampus from 400 mm down to continuously circulate oxygenated and heated culture media.7–9 3 70 mm. The connectivity of the neurons in the hippocampus Many environmental chambers have been developed for studying different aspects of organotypic slices including development or aBiomedical Engineering Department, McGill University, 740, Dr. Penfield their response to pharmacological, electrophysiological and Ave, Montreal, Quebec, H3A1A4, Canada biochemical stimuli.10–13 These perfusion setups have mostly been bMcGill University and Genome Quebec Innovation Centre, 740, Dr. Penfield Ave, Montreal, Quebec, H3A1A4, Canada developed for upright microscopes because conventional orga- cDepartment of Pharmacology and Therapeutics, McGill University, notypic slices are typically 400 mm thick, which exceeds the Bellini Life Science Complex, 3649 Sir William Osler, Montreal, working distance of high magnification objectives, and therefore Quebec, H3W0B1. E-mail: [email protected] only permit imaging from above using an upright microscope. The d Department of Neurology and Neurosurgery, 740, Dr. Penfield Ave, use of closed chamber leads to rapid degeneration of organotypic Montreal, Quebec, H3A1A4, Canada. E-mail: [email protected] † Electronic supplementary information (ESI) available: Technical slices, but a recent study using PDMS microfluidic chamber drawings; design files; Fig. S1, and Video S1. See DOI: 10.1039/b916669f allowed for measurements of approximately one hour and showed

326 | Lab Chip, 2010, 10, 326–334 This journal is ª The Royal Society of Chemistry 2010 View Article Online

local perfusion by taking advantage of the laminar flow (which however does not allow single point application), but without provision for high resolution imaging.14 An add-on to electrophysiological chambers was also proposed featuring a substrate with microfluidic conduits and openings distributed at predefined points.15 This chamber allows for local perfusion, but without control of the exact location, nor confinement, and is not suited for an inverted microscope. Open chambers for inverted confocal imaging such as the commercial RC-30 series from Warner Instruments are designed for round coverslips and are not compatible with the standard 12 24 mm2 coverslips imposed by the roller-tube cultures.16 In the upright configuration, a water immersion lens is used and placed as close as 200 mm above the organotypic slices for high resolution imaging. Despite the small gap, capillaries are commonly inserted from the side into the slice for patch clamping and intercellular recording with the help of micromanipulators. However, placement and alignment of multiple capillaries is cumbersome, and repositioning cannot be done easily. It is also possible to use two capillaries for carrying out microperfusion, but because of the open configuration it is difficult to control the perfusion area and the percolation radius leading to uncontrolled diffusion of the chemicals beyond the intended processing volume. The microfluidic probe (MFP) allows for localized delivery of Fig. 1 Schematic of the perfusion chamber and the MFP used for local a microfluidic stream to a surface while avoiding the need for perfusion of organotypic slices while concurrently imaging with an closed microfluidic systems.17,18 The basic principle of the MFP is inverted confocal microscope. (a) Cross section of the perfusion chamber to inject a solution through one aperture into a small gap between and MFP along with syringe pumps, and (b) close-up view showing the the MFP and a substrate surface, and immediately reaspirate it MFP, the perfusion flow (in green) which penetrates into the slice below, the coverslip, and the microscope objective. back through a second aperture, forming a hydrodynamically- shaped microjet. The MFP is characterized by a blunt, flat tip a few 2 hundred micrometres in diameter with originally two apertures at MFP, (ii) compatibility with the 12 24 mm coverslips used the center of the tip, which are each connected to a different pump. for the roller-drum organotypic slices, (iii) perfusion of slices for A MFP is operated by positioning it a few micrometres above several hours, (iv) compatibility with inverted microscope and a substrate surface immersed in a solution. The microjet locally high N.A. lenses, (v) failure-free operation to prevent spilling of perfuses the substrate surface and the exact shape is determined by solutions onto the microscope, and (vi) finally, to ensure the the ratio of injection and aspiration flow rate and the gap between viability of the organotypic slices, the perfusion chamber must Published on 19 November 2009. Downloaded by McGill University 10/04/2015 14:12:50. the MFP tip and the substrate. The hydrodynamic confinement allow for rapid insertion and perfusion of the coverslips prevents diffusion of the injected chemicals beyond the intended carrying the organotypic slices following their extraction from zone, and the MFP can be scanned across the substrate and the culture tubes to prevent drying and degradation of the positioned at different locations. For live visualization, the MFP tissue. needs to be mounted atop of an inverted microscope. The chamber that was designed according to these criteria is In this article, we present a microfluidic perfusion setup for the assembled out of three main modules, (a) a main plate (b) an local perfusion of roller-drum organotypic slices using a MFP adapter plate (both made of aluminium), and (c) a bottom with concurrent confocal imaging on an inverted microscope, support plate made of thin steel sheets. This configuration is Fig. 1. The setup comprises (a) a novel perfusion chamber that is modular so that only the metal adapter plate and the coverslip compatible with high numerical aperture lenses and that allows holder need to be replaced (while the main plate can be reused) rapid mounting of the roller-drum slices and their cultivation for to accommodate a coverslip of a different dimensions. The several hours, and (b) a novel, transparent MFP made entirely in chamber was designed using 3D design software (Inventor 3D, PDMS with 6 injection/aspiration apertures used for local Autodesk) and optimized iteratively. The final design is shown perfusion of the slices. As an illustration of the potential of this in an exploded view in Fig. 2, and additional details can be system for localized processing, we show confocal images of the found in the technical drawings in the ESI and in the original local perfusion of an organotypic slice with a fluorescent dye. design files.† The main plate fits into the motorized microscopy stages from both Prior Scientific and Applied Scientific Instruments. Design Perfusion chamber Microfluidic probe

The design of the perfusion chamber has to satisfy multiple The opening in the perfusion chamber after assembly is requirements, including (i) accessibility from the top by the 10 22 mm2 in size and imposes constraints on the size of

This journal is ª The Royal Society of Chemistry 2010 Lab Chip, 2010, 10, 326–334 | 327 View Article Online

Solutions, USA) and glued together. The lower steel sheet features a 10 22 mm2 hole to support the 12 24 mm2 coverslip (Kaeslin, Switzerland) that serves as the bottom of the perfusion chamber. The upper sheet has a 12.5 24.5 mm2 cut- out for inserting the coverslip and has the same thickness as the coverslips (100 mm) so that they are level. An interchangeable adapter plate made of stainless steel with a 10 22 mm2 cut-out (a second adapter plate with a 20 20 mm2 opening for accommodating 22 22 mm2 coverslips was also made) receives a rectangular, soft rubber sealing ring (Shore A10) with 1 mm width and 500 mm thickness (McMaster, Elmhurst, USA). The ring serves as an inner gasket and produces a liquid-tight seal when the coverslip is pressed against it. An outer gasket is formed by a second rubber sheet (see Fig. 2a) and sandwiched between the steel sheets and the plate. A vacuum is applied through a hole drilled through the adapter plate and presses the steel sheet against the outer gasket and against the cover slip which is pressed against the inner gasket, and produces a liquid tight seal around the perfusion chamber opening (See Fig. 2a). The perfusion channels, the vacuum line and the perfusion bath are machined entirely in the metal adapter plate. In order to reduce leaks and improve the overall hermeticity, silicone O-rings have been added to each access (media inlet, liquid outlet, vacuum). A hydrophobic and chemically resistant 500-micrometre-thick PTFE sheet (McMaster Elmhurst, USA) was inserted between the metal cover and the main plate and serves as the top cover of the embedded channels and prevents leakage. A 1-cm-thick main plate machined in Al with a cut-out for the adapter plate completes the setup. The adapter plate and main plate are screwed together using 6 screws. The main plate also has 3 threaded holes for fixing nylon plastic quick-turn Luer lock couplings (McMaster) feeding the inlet and outlet of the flow chamber and the vacuum line, and two large holes for handling (see Fig. 2b). As a safety measure in the event of vacuum failure, two magnets were also glued to the steel sheets and can be held in place by two counter-magnets positioned on the main plate. Published on 19 November 2009. Downloaded by McGill University 10/04/2015 14:12:50. Fig. 2 Design of the perfusion chamber. (a) Exploded view of the The perfusion chamber is connected to a vacuum pump perfusion chamber for inverted microscope. The primary characteristic of (vacuum pressure: 75 torr, Maxidry, Fisher Scientific) for the the chamber is that it can be assembled in seconds following the insertion coverslip sealing while a peristaltic pump (flow rates: 0.002–35 ml of the coverslip with an organotypic slice. (b) Schematic of the chamber min1, REGLO Digital ISM 834, Ismatec) ensures the media setup connected to a vacuum pump for coverslip sealing and a peristaltic recirculation inside the perfusion bath. The peristaltic tubing pump for perfusion. (c) Photograph of the chamber. vacuum line is linked to the chamber by nylon quick connectors, allowing a quick and tight connection. the MFP which needs to be inserted inside the chamber with the Finally, the plate comprises 3 set screws with micrometre pitch freedom to move without touching the sidewalls of the chamber. that can be used to adjust co-planarity of the coverslip and the Up to now, MFPs were made out of brittle 3 7mm2 Si chips imaging plane of the microscope, which is important for opera- and manipulated with a 8 mm wide rod,17 which would allow for tions with the MFP that needs to operate with a constant gap.18 only 2 mm of lateral movement inside the chamber opening. We therefore designed a new MFP entirely made out of PDMS and Microfabrication of the MFP shaped as a rod to help increase the area accessible to the probe, as schematically shown in Fig. 1. A new fabrication process was The MFP is made by replica molding of 2 identical PDMS pieces developed to make the MFP and is described in detail below. from raised photoresist structures patterned on a wafer, followed by bonding them together with a thin membrane in between as described below and shown in Fig. 3. Materials & methods Microfabrication of the molds for the MFP. Fig. 3, steps 1–3 Fabrication of the perfusion chamber delineate the mold fabrication. SU8-25 (MicroChem) was spin- The perfusion chamber comprises two steel sheets each 100 mmin coated on a 4-inch wafer (Silicon Wafers) at 500 rpm for 10 s and thickness that were cut by laser machining (Sefar Printing 1000 rpm for 30 s resulting in a 40-mm-thick layer. The wafer was

328 | Lab Chip, 2010, 10, 326–334 This journal is ª The Royal Society of Chemistry 2010 View Article Online

50/E, Proxxon) and a 280 mm drill bit. Debris were removed using Scotch tape to prevent subsequent clogging of the microchannels. A thin PDMS layer (Fig. 3, step 7) was spin-coated on a clean Si wafer (precoated with a trichlorofluorosilane) at 1800 rpm for 30 s yielding a 40 mm thick layer and partially cured at 60 C for 10–15 min. One of the molded PDMS blocks was then slightly pressed and bonded to the thin PDMS layer for 5 min at 60 C. Next, the two PDMS pieces are detached, and a second molded PDMS block aligned under a stereomicroscope, sealed against the back-side of the membrane, and the entire setup bonded for 24 h at 65 C. The probe is finalized by cutting it using a razor blade, first perpendicularly to the channels, and then at an angle to form a ‘‘blunt’’ tip a few hundred micrometres in width. Each probe was then trimmed to a rod approximately 15 5 5mm3 in size.

Microfluidic probe imaging station and syringe pumps An inverted confocal microscope (Eclipse C1si confocal, Nikon) equipped with a precision motorized stage (H117 ProScan II, Prior Scientific) was used for real-time imaging of organotypic slices. Alternatively, an inverted microscope (TE2000, Nikon) also with a high precision stage (MS-2000, Applied Scientific Instrumentation) and an electron multiplied CCD camera (QuantEM:512SC, Photometrics) was used. Both microscopes were equipped with a custom designed environmental chamber (Precision Plastics) with temperature and humidity control, and a fixed platform. The PDMS MFP was clamped to the set-up using a home-made handling rod that was described previously.17 A photograph of the MFP positioned inside the perfusion chamber is shown in Fig. 4. The MFP was attached to a XYZ- PhiTheta micromanipulator (MX7600, Siskyou, with computer-controlled XYZ) itself screwed onto the platform. The micromanipulator is used adjust the position of the probe in the Fig. 3 Microfluidic probe and its fabrication process. (a) Schematic of center of the imaging field (XY), the gap between the MFP, the microfabrication process of the PDMS-MFP using multilayer soft- the substrate (Z) and the parallelism between the MFP and the Published on 19 November 2009. Downloaded by McGill University 10/04/2015 14:12:50. lithography. (b) Overview of the PDMS-MFP. (c)–(d) Enlarged views of substrate (PhiTheta, using two manual goniometers). Each the 2-dimensional array of holes for two different design width. aperture of the MFP was connected using a polymer capillary soft baked for 5 min at 65 C then 15 min at 95 C on a hot plate. with 360 mm outer diameter and 50 mm inner diameter (Upchurch Next, it was exposed using an mask aligner (EVG 620, Electronic Scientific) to the injection and aspiration syringes (Hamilton) Vision) with 250 mJ cm2. The wafer was post-baked on a hot with volumes varying between 1 ml and 100 ml depending on the plate for 1 min at 65 C, 4 min at 95, and then developed for experiment, and which were controlled using precision pumps 6 min in SU-8 developer solution (MicroChem). The structured mold was activated in an air plasma (Plasmaline 415, Tegal) for 30 s, coated with a Trichlorofluorosilane (1H,1H,2H,2H-per- fluorooctyl silane, Aldrich) using a desiccator, and then placed in an oven at 115 C for 1 h.

Replication of the mold into PDMS. Fig. 3, steps 4–6 outline the replication process. PDMS (Sylgard 184, Dow Corning Corporation) was mixed in a 1 : 10 ratio and poured as a 2–3 mm thick layer on the mold within a Petri dish, and degassed in a desiccator connected to a vacuum source for 15 min to remove bubbles. The wafer was placed in an oven at 60 C for 30–35 min to partially cure the PDMS, which was peeled of the mold and placed in a clean Petri dish and cut into blocks. Fig. 4 Photograph of the microperfusion setup mounted on an inverted Drilling, assembly, and trimming of the MFP. Access holes microscope comprising the home made perfusion chamber and the were drilled into the PDMS blocks with a microdrill (Micromot PDMS-MFP clamped with the probe holder.

This journal is ª The Royal Society of Chemistry 2010 Lab Chip, 2010, 10, 326–334 | 329 View Article Online

(Nemesys, Cetoni GmbH, Germany). The injection rate was set 320 320 900 nm3. Multiple optical sections were taken at 100 pl s1 and the aspiration at 1 nl s1 unless indicated across the slice in the microperfused area. otherwise. The MFP was positioned over an area of interest by moving the automated XY stage of the microscope; the MFP Results and discussion thus always remains at the centre of the imaging field while different areas of the sample are moved under the probe. All Perfusion chamber operations were automated and computer controlled using the The short working distance and the large diameter of high 19 Labview software, or with a joystick. For flow confinement magnification objectives used for confocal microscopy require 1 studies, 0.1 mgml of fluorescein or rhodamine in PBS were that the area around the substrate provide sufficient clearance to used. For the perfusion of the organotypics slice, 10 kDa dextran prevent crashing the objective into the mechanical support of the conjugated to Alexa 647 dye (Invitrogen) at a concentration of coverslip when imaging excentrically. Here, the coverslip is 1 75 mgml was used. supported by a 1 mm edge on the lower steel sheet which protrudes only 100 mm beyond the coverslip and thus can Numerical simulation of the shear stress below the probe accommodate a 60 oil immersion objective with a working distance of 0.13 mm (the working distance is defined as the gap The fluid flow shear stress were calculated using water as a fluid between the objective and a standard 0.17 mm thick coverslip and no slip boundary conditions using a numerical simulations when focusing on the sample immediately behind the coverslip, software (Comsol Multiphysics 3.4, Comsol Inc.) run on an and is larger here because the coverslip is only 100 mm thick). The eight-core, 64 bit computer (Xeon, Dell) with 10 Gb of RAM. sealing of the coverslip is ensured by the 1 mm wide sealing ring The simulations were run under steady state conditions and the that is fixed on the adaptor plate. The ring presses on the top side boundary condition were set as zero pressure at the edge of the of the coverslip, which is wet when mounting it, and which blunt tip of the probe. proved to be a source of leakage in the initial tests. The use of a very soft ring allowed the overcoming of this problem. The Transgenic mice space between the inner and outer sealing ring forms a 5 mm wide channel that allows the distribution of the vacuum around the Variegated mice were generated by using standard techniques. A coverslip. The steel sheets are thus pressurized against the back- construct was generated where the cDNA for enhanced GFP was side of the coverslip which in turn presses it on the sealing ring fused to the membrane-anchoring domain (first 41 amino acids) and ensures the tightness of the perfusion chamber. of a palmitoylated mutant of MARCKS29 under the Thy1 Neurons in organotypic slices are very sensitive to environ- promoter. Twenty-five distinct lines were produced, each with mental changes and need to be perfused as quickly as possible 20 subtly different patterns of expression. Of these, L15 mice were after extraction from the culture tube. With the current chamber, chosen because they had a low, but consistent, number of organotypic slices are mounted into the chamber by placing the mGFP-labeled cells within the CA1 area of the hippocampus. supporting coverslip into the opening in the steel sheets resting on a work bench, and then positioning the rest of the perfusion Hippocampal organotypic slices chamber above it. Two holes machined in the main plate serve as

Published on 19 November 2009. Downloaded by McGill University 10/04/2015 14:12:50. guide to two magnets glued on the steel plate which ensures Organotypic slice cultures were used for these experiments alignment of the coverslip relative to the opening that will form because they provide the major advantage of exhibiting the perfusion chamber. Then, the vacuum can be initiated to preserved tissue-specific organization of synaptic connections in firmly clamp the coverslip to the ring on the adapter plate, and an in vitro preparation suitable for imaging studies. Slices the perfusion media added to the chamber. As a safety measure (400 mm thick) were prepared from the hippocampi of 6-day-old to avoid spilling of media onto the microscope below in the event L15 mice and maintained in roller tubes for 2–4 weeks before use, of vacuum failure, two counter magnets are placed in the main 6 as previously described, and fixed in 4% paraformaldehyde plate to hold the steel sheets and the coverslip firmly in place. before imaging. In summary, the perfusion chamber introduces four major innovations comprising (i) an overall novel chamber concept Confocal imaging permitting ultra-fast mounting by vertical insertion. (ii) A dual force application by magnetic forces and vacuum, which had not Slice cultures were transferred to a recording chamber mounted been used in this context. (iii) A sealing mechanism with indirect on the microscope that was heated (32 C). The slices were pressurization along a one-millimetre-wide sealing ring which continuously perfused in Tyrode solution comprising 137 mM helps maximizing the useful imaging and working area. NaCl, 2.7 mM KCl, 2.5 mM CaCl2, 2 mM MgCl2, 11.6 mM (iv) Finally, an adapter plate with embedded fluidic conduits NaHCO3, 0.4 mM NaH2PO4, and 5.6 mM glucose. mGFP was which allows the use of the same setup for multiple cover slip sizes. excited by using the 488 nm argon laser line, and the dextran with a HeNe 639 nm laser. The mGFP-labeled pyramidal cells were Fabrication and assembly of the MFP imaged using either a 60 oil immersion objective (Nikon CFI Plan Apochromat VC (violet corrected) 60 oil, N.A. 1.40, with We successfully made up to 10 PDMS-MFP a day using the 0.13 mm working distance and designed for 170 mm substrate fabrication method presented here without need for clean-room thickness) or a 40 Nikon CFI Plan Apochromat, (N.A. 0.95, access, except for the initial fabrication of the master mold. The W.D. 0.14 mm; C.C.0.11–0.23) with voxel dimensions of successive bonding of two partially cured, molded PDMS blocks

330 | Lab Chip, 2010, 10, 326–334 This journal is ª The Royal Society of Chemistry 2010 View Article Online

to the membrane is an innovation that allowed for rapid assembly. The parameters were optimized so that the PDMS was mechanically stable while still having sufficiently uncured polymer chains at the interface of the two layers for cross-linking between them upon additional curing. A practical advantage of this process is that misalignments discovered after bringing the pieces into contact can readily be corrected by simply removing the block and repeating the alignment procedure, which is not possible when using oxygen plasma bonding for example.21,22 A single mold was used to make multiple probes, and a single probe was reused multiple times for different experiments without noticeable deterioration. The mechanical rigidity of the MFP is dependent on the thickness of the top and bottom layers and can be varied easily during replication of the mold by adjusting the thickness of the prepolymer layer. The rigidity is important because the probe is only fixed at the rear end, and because capillaries need to be plugged into the holes drilled on each side. Since the outer diameter of the capillaries is 360 mm and is larger than the diameter of the holes of the probe which are 280 mm, they can be simply plugged in, but the apertures have to be spaced sufficiently for easy plugging. The spacing between the microchannels Fig. 5 Fluorescent micrographs of a PDMS-MFP operated using decreases as it gets closer to the aperture. The design includes different injection/aspiration configurations that are possible using 6 different microchannel widths, from 20 mmto50mm. apertures. (a–g) Combinations with between 1–3 injection and 1–3 Whereas the width of the microchannels is defined by the aspiration ports in operation. The aspiration side can be identified by the mask, the depth of the channel can be chosen when making the narrow streams. (h) Diagonal alignment of aspiration and injection with mold by changing the thickness of the spin-coated SU-8 photo- two fluorescence dyes produces a black line where the two dyes diffuse and mutually quench one another. resist layer. The design can thus be easily modified to match to desired microfluidic applications for customized flow patterns. 17 Compared to the original MFP made of Si, the PDMS MFP aspiration and 3 for injection, and by varying the flow rates, features a reduced footprint as described above, a fabrication different shapes can be generated. Finally, in Fig. 5h, 2 different method that can be carried out using photolithograhy only, and type of solutions, one containing fluorescein and one containing allows adjusting the height of the apertures during fabrication, 6 rhodamine, are injected through two apertures in diagonal (top instead of 2 apertures, and is transparent. left and bottom right) and aspirated back by two other aspira- tions (bottom left and top right). Between the two apertures 23

Published on 19 November 2009. Downloaded by McGill University 10/04/2015 14:12:50. Operation and characterization of the MFP a stagnation point is formed, and limited mixing takes place at the interface between the green fluorescein and red rhodamine. The MFP was mounted on the inverted microscope and lowered Fluorescein and rhodamine constitute a fluorescence resonance into the perfusion chamber on the inverted microscope, all of energy transfer pair and we ascribe the loss in fluorescence to which form the microperfusion setup, Fig. 4. The flow properties quenching effects, which gives rise to the black line. We further of the PDMS-MFP with 6 apertures were characterized by observed that the hydrophobic rhodamine dye diffused into the injecting fluorescein in water and visualizing the flow with an PDMS.24 The configuration of the PDMS-MFP resembles 17,18 inverted fluorescence microscope. The hydrodynamic the one of multibarrel pipettes,25 but the resolution that can be confinement and shape of the injected microfluidic stream were achieved is lower when using large apertures as shown here. adjusted by changing the aspiration (Qasp) and injection (Qinj) Conversely, the MFP with a blunt tip is robust, its coplanar ¼ flow rate ratio. Experiments showed that a ratio of Qasp/Qinj 10 geometry facilitates the confinement of the injected fluid over or higher produced an effective confinement of the microjet large area,17 and it does not break upon contact with the regardless of the size of the gap. The number of combinations substrate while yielding and preventing damage to the tissue that can be realized with 6 apertures, each being used for injec- because it is soft. tion, aspiration or quiescent is 3n 3 ¼ 726 for n ¼ 6, whereas ‘‘3’’ accounts for the fact that at least one opening must be used Calculation of the shear stress at the substrate surface for aspiration or injection (i.e. the three combinations with all apertures used either for injection, aspiration or quiescent does Numerical simulations in 3D were carried out to simulate the not allow for hydrodynamic confinement). A subset of injection– flow of media and to calculate the shear stress at the substrate aspiration configurations were tested using 10 nl s1 for injection surface or at the top of an organotypic slice following exposure and 100 nl s1 for aspiration, Fig. 5. Fig. 5a–e show different to the flow of a microfluidic probe. The coplanar geometry of the configurations with a single aspiration only, and between 1–3 probe and the small gap between probe and slice leads to a high injections ports. The aspiration openings are the ones where the resistance, and therefore the recirculated bath fluid flows around stream is narrowly focused. In Fig. 5f–g, 3 apertures are used for the probe and does not add significantly to the shear stress. The

This journal is ª The Royal Society of Chemistry 2010 Lab Chip, 2010, 10, 326–334 | 331 View Article Online

Microperfusion and imaging of hippocampal brain slices

Brain slices can be used to study learning and memory, they exhibit similar morphology and synaptic transmission to in vivo preparations. They can also be used as models for diseases such as epilepsy, for studying regeneration following lesions inflicted on the slices using a knife or chemicals, and following the growth of neurons across the lesion. We first tested imaging of organotypic slices immobilized in our perfusion chamber with the inverted confocal microscope using 10,20,40, and 60 objective. A representative XY image plane for each objective is shown in Fig. S1 in the ESI.† To illustrate local microperfusion a fixed organotypic brain slice of the hippocampi was perfused with a fluorescent dye using the MFP. The MFP was first lowered onto the substrate holding the slice on a bare area for parallelization, then raised, positioned above the area of interest, and then lowered down to 90 mm above the surface of the glass, which corresponds to a gap of 20 mm between the probe and the slice. Fluorescently labeled dextran that does not react with the tissue, but can diffuse inside of the slice, was injected through a single aperture and aspirated through an opposing aperture. We perfused the slices for up to 30 min and tested flow rates of up to 10 nl s1 for injection and 100 nl s1 for aspiration without observing detachment of cells. The MFP used for this experiment had 40 70 mm2 apertures that were 55 mm apart, and a flow rate of 5 nl s1 and 25 nl s1 respectively was used resulting in a hydrodynamic confinement zone that is 114 mm long Fig. 6 Result of a finite element modelization of the shear stress for two and 83 mm wide. The MFP was positioned in the CA1 region of the 2 1 40 40 mm sized apertures 60 mm apart, a 40 mm gap and 100 pl s and organotypic slice (see Fig. 7a). The distribution of the dye above 1nls1 injection and aspiration flow rates, respectively. (a) Cross-section and in the organotypic slice was imaged with the confocal through the middle of the apertures. The maximal shear stress on the substrate is at the projected edge of the aspiration aperture. (b) Variation microscope using a 40 objective and triple laser exposure at of the maximal shear stress at the substrate surface as function of the gap. 403 nm for imaging of the stained cell nuclei (blue in the picture), 488 nm for mGFP (green) and 639 nm for exciting the fluorescent dye (red) (see Fig. 7 and video S1 in the ESI†). The characteristic teardrop shape of the hydrodynamic flow is apparent in the XY simulation was based on a probe with one injection and one section, and the cross-sections show that the dextran penetrates aspiration aperture with a flow rate of 100 pl s 1 and 1 nl s 1, Published on 19 November 2009. Downloaded by McGill University 10/04/2015 14:12:50. 32 mm deep into the 70 mm thick slice after 12 min of perfusion. respectively. The size of the apertures is 40 40 mm2 and they are The three XZ sections at different positions under the MFP reveal separated by 60 mm. The shear stress was computed for gaps that the dye penetrates deeper in the centre than at the edges, varying from 10 to 40 mm, Fig. 6. The result show that the shear replicating the distribution profile seen in a liquid.17 The confined stress at the surface decreases in a quadratic manner with the dye perfuses about 25 cells at once, with a relatively sharp inverse of the gap when keeping the flow rate constant. The shear boundary. Whereas these parameters replicate flow in open space, stress is highest below the aspiration aperture where the local the transport of the dye is likely to be dominated by diffusion. flow velocity is the highest, and varies linearly with the flow rate Indeed, there is a continuous increase in dye intensity in the first as expected (data not shown). The injection flow rate is much few minutes. In addition, the penetration is sometimes inhomo- smaller, and does not contribute significantly to the shear stress. geneous, depending on the microstructure of the tissue. Following Thus, the shear stress produced by the probe can be reduced by successful perfusion of 10-kDa-dextran inside a fixed tissue, it will either increasing gap or reducing the aspiration (and injection) be important to study the diffusion of different chemicals and flow rate, or increasing the aperture size, to remain below a crit- proteins inside live tissues, and to quantify the mass transport ical value depending on the particular cell and condition. There using for example fluorescence recovery after photobleaching have been no studies on the effect of shear stress on organotypic experiments.27 It will be interesting to see how the transport in the slices. In the case of dissociated neurons, one study showed that slice along the vertical and horizontal planes will vary as function cells can withstand stresses of 1.5 Pa when they are applied of the composition and mass of the chemical being perfused, and intermittently for periods of 2 h, although they detach for stresses of the flow parameters. of 0.5 Pa when applied continuously.26 Here, only the outermost cells are exposed to a high flow, which mostly comprises glial cells. The maximal shear using the conditions described above is Conclusions 0.33 Pa, which is below the threshold for detachment of dissociated cells, and which should thus permit long term micro- We have developed a microfluidic setup comprising of a perfu- perfusion of live cells with the MFP. sion chamber for the culture of organotypic slices over several

332 | Lab Chip, 2010, 10, 326–334 This journal is ª The Royal Society of Chemistry 2010 View Article Online

Fig. 7 Bright field image and confocal cross-sections of an organotypic slice from L15 transgenic mice. (a) Overview of a mature mouse organotypic hippocampal slice culture, which has been Nissl-stained to show the general morphology of preparation. (b–j) XY cross sections from 76 mmto33mm above the cover slip with a 5.4 mm separation between each image plane. Cell nuclei are labelled with DAPI (blue), and the cell membranes labelled with mGFP are green. The slice was locally perfused with a fluorescent dextran (red) using the MFP with a ratio of 5 between the aspiration and injection flow rates. The cells appear in the second image plane, whereas the dye gradually vanishes and disappears in the last section corresponding to a penetration depth of 32 mm. All the images of the confocal stack are shown as video S1 in the ESI.† (j–l) XZ cross-sections along the dashed lines shown in (f) illustrating the confinement of the stream along the vertical plane. The two dashed lines in (k) indicate the top of the slice.

hours and a novel MFP made of PDMS used for localized Acknowledgements perfusion of a small number of cells within a slice while permitting concurrent confocal imaging on an inverted microscope. The The authors wish to thank the CIHR Regeneration and Nano- Published on 19 November 2009. Downloaded by McGill University 10/04/2015 14:12:50. PDMS-MFPs presented in this report are straightforward to medicine team grant, the McGill Nanotools Microfab Labora- assemble, and feature 6 apertures allowing different combina- tory (funded by CFI, NSERC and VRQ), Tadayuki Shimada, tions of chemicals to be injected and re-aspirated into the MFP. Matthieu Nannini and Nikolas Pekas for their help and advice. The dimensions of the MFP can be changed, additional layers Franc¸ois Charron for his excellent technical support in preparing added, or entirely redesigned to meet a particular requirement if the organotypic cultures. R. A. M. holds an FRSQ Senior Salary desired. The MFP may also be used in combination with award and Hugh and Helene Memorial Award. D. J. holds a conventional perfusion chamber.16 a CRC chair. Optical techniques can also be used for locally delivering calcium or neurotransmitters by photo-uncaging with high speed References and high spatial control.28 However, these methods are limited to 1B.H.Gahwiler,€ M. Capogna, D. Debanne, R. A. McKinney and chemicals that can be photo-uncaged, and unlike when using the S. M. Thompson, Trends Neurosci., 1997, 20, 471–477. MFP, they cannot be applied continuously over long periods of 2 S. Cho, A. Wood and M. R. Bowlby, Curr. Neuropharmacol., 2007, 5, time. 19–33. € The potential of the MFP for biological studies will be 3B.Gahwiler, J. Neurosci. Methods, 1981, 4, 329. 4 L. Stoppini, P. A. Buchs and D. Muller, J. Neurosci. Methods, 1991, explored in the future using organotypic slices and by locally 37, 173–182. delivering multiple chemicals that can be used to promote the 5 D. Debanne, B. H. Gahwiler and S. M. Thompson, J. Physiol., 1998, regeneration of neurons, or for delivery of viruses for localized 507, 237–247. 6 R. A. McKinney, M. Capogna, R. Durr,€ B. H. Gahwiler€ and transfection. This PDMS MFP, combined with the inverted S. M. Thompson, Nat. Neurosci., 1999, 2, 44–49. perfusion chamber presented here, offer great opportunities for 7 W. Thiemann, R. Malisch and K. G. Reymann, Brain Res. Bull., 1986, neuroscientists to study neuron interactions where high spatial 17, 1–4. resolution is required together with long-term, steady-state 8 I. D. Forsythe and R. T. Coates, J. Neurosci. Methods, 1988, 25, 19– 27. perfusion of drugs. e.g. the study of synapse remodelling within 9 E. J. Anderson, K. Tate and L. Melissa, BioMed. Eng. Online, 2007, 6, organotypic slices following injury. 46.

This journal is ª The Royal Society of Chemistry 2010 Lab Chip, 2010, 10, 326–334 | 333 View Article Online

10 C. L. Ho, T. Y. Mou, P. S. Chiang, C. L. Weng and N. H. Chow, 20 V. De Paola, S. Arber and P. Caroni, Nat Neurosci, 2003, 6, 491–500. BioTechniques, 2005, 38, 267–273. 21 S. W. Lee and S. S. Lee, Microsyst. Technol., 2008, 14, 205–208. 11 A. Y. Budantsev, Biotech. Histochem., 2007, 82, 263–266. 22 D. A. Chang-Yen, D. G. Myszka and B. K. Gale, 12 J. F. Koerner and C. W. Cotman, J. Neurosci. Methods, 1983, 7, 243. J. Microelectromech. Syst., 2006, 15, 1145–1151. 13 H. L. Haas, B. Schaerer and M. Vosmansky, J. Neurosci. Methods, 23 M. A. Qasaimeh, P. Sanyal, R. Safavieh, C. M. Perrault, A. Queval 1979, 1, 323. and D. Juncker, Proceedings of the The Twelfth International 14 A. J. Blake, T. M. Pearce, N. S. Rao, S. M. Johnson and Conference on Miniaturized Systems for Chemistry and Life J. C. Williams, Lab Chip, 2007, 7, 842–849. Sciences, pp. 841–843, San Diego, California, 2009. 15 J. S. Mohammed, H. H. Caicedo, C. P. Fall and D. T. Eddington, Lab 24 T. L. ter Laak, F. J. M. Busser and J. L. M. Hermens, Anal. Chem., Chip, 2008, 8, 1048–1055. 2008, 80, 3859–3866. 16 http://www.warneronline.com/index.cfm. 25 O. Feinerman, A. Rotem and E. Moses, Nat. Phys., 2008, 4, 967–973. 17 D. Juncker, H. Schmid and E. Delamarche, Nat. Mater., 2005, 4, 622– 26 I. A. Kim, S. A. Park, Y. J. Kim, S.-H. Kim, H. J. Shin, Y. J. Lee, 628. S. G. Kang and J.-W. Shin, J. Biosci. Bioeng., 2006, 101, 120–126. 18 C. M. Perrault, M. A. Qasaimeh and D. Juncker, J. Vis. Exp., 2009, 27 N. B. Cole, C. L. Smith, N. Sciaky, M. Terasaki, M. Edidin and 28, 1418. J. Lippincott-Schwartz, Science, 1996, 273, 797–801. 19 C. M. Perrault, T. Brastaviceanu, K. Anderson, Y. Kabakibo and 28 V. Nikolenko, K. E. Poskanzer and R. Yuste, Nat. Methods, 2007, 4, D. Juncker, unpublished work. 943–950. Published on 19 November 2009. Downloaded by McGill University 10/04/2015 14:12:50.

334 | Lab Chip, 2010, 10, 326–334 This journal is ª The Royal Society of Chemistry 2010