1 THE FUNCTIONAL INTEGRATION OF ADULT NEURAL PROGENITOR CELLS INTO INTACT AND COMPROMISED PRIMARY NEURAL NETWORKS IN A MICROELECTRODE ARRAY CO CULTURE SYSTEM By CRYSTAL LYNN STEPHENS A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2012
2 2012 Crystal Lynn Stephens
3 To my family, William, Patricia, Cheryl, Chanel and Benjamin Stephens
4 ACKNOWLEDGMENTS I acknowledge the J. Crayton Pruitt Family Department of Biomedical Engineering for giving me the opportunity, and providing an enriching, didactic environment that encouraged me to work and learn in an exciting field of study I thank the members of the Stem Cell Research Laboratory, Rachel, Lan, Aditya and Vik for being wonderful lab mates. I am grateful for their help and support with my research, and for the fun times we had in and out of the lab. I know I have made many l asting friendships, and have many fond memories throughout the years. I also thank my undergraduate researchers, Ellen and Crystal, for their excellent work and assistance. I thank my committee members, Dr. Mingzhou Ding and Dr. Charles Jason Frazier for their essential recommendations for my projects, and for driving me to aim higher with my goals and to be confident in my work. I thank Dr. Brandi Ormerod, my advisor and committee chair, for her investment of time and thought into my research. Her guidanc e and encouragement was invaluable to my research and my growth as a scientist, and as a person. I thank Dr. Tom DeMarse, my co chair, for his guidance and support. I truly appreciate the time and work he devoted to my education and research projects. Las tly, I thank my family for their love and encouragement throughout my years in school. I thank my parents for lending me the confidence to pursue a doctorate in a field that I am passionate about even though I may find it challenging at times. I also thank my brother and sisters for supporting me even as they pursue their own careers. I hope that I can be as supportive and inspiring for them, as they were for me.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF FIGURES ................................ ................................ ................................ .......... 8 LIST OF ABBREVIATIONS ................................ ................................ ............................. 9 ABSTRACT ................................ ................................ ................................ ................... 10 CHAPTER 1 GENERAL INTRODUCTION ................................ ................................ .................. 12 The Concept of Neuroregeneration ................................ ................................ ........ 12 Neural Progenitor Cell Origins and Other Stem Cell Sources ................................ 13 The Neurogenic Niche Provides Clues for Therapeutic Strategies ......................... 17 Adult Neurogenesis: A Benchmark for Functional Integration ................................ 19 Transplantation Studies Provide Evidence of Functional Integration ...................... 21 Microelectrode Array (MEA) Technology ................................ ................................ 24 Investigating NPC Addition to Damaged Neural Netwo rks Using a Hypoxic Ischemic Stroke Model ................................ ................................ ........................ 26 Specific Aims ................................ ................................ ................................ .......... 29 2 ADULT NEURAL PROGENITOR CELLS REACTIVATE SUPERBURSTING IN MATURE NEURAL NETWORKS ................................ ................................ ........... 32 Introduction ................................ ................................ ................................ ............. 32 Material and Methods ................................ ................................ ............................. 33 Preparation of Microelectrode Arrays Dishes and 8 well LabTek Chamber Slides ................................ ................................ ................................ ............ 33 Neural Progenitor Cell Harvest and Expansion ................................ ................ 34 Preparation of NPC Cultures ................................ ................................ ............ 35 Preparation of Primary Cultures ................................ ................................ ....... 36 Preparation of NPCs for Co Culture ................................ ................................ 37 Immunocytochemistry ................................ ................................ ...................... 37 Confocal and Epifluorescent Microscopy ................................ ......................... 38 Electrophysiological Data Collection ................................ ................................ 39 Statistical Analyses ................................ ................................ .......................... 41 Results ................................ ................................ ................................ .................... 42 Adult NPC Generated Neurons that Retain Immature Phenotypes are Incapable of Ensemble Activity ................................ ................................ ..... 42 Primary Neural Cultures Develop Neural Ensemble Behavior Coincident with Neuronal Maturation ................................ ................................ .............. 47 NPCs Evoke Re Emergent Superbursting In Pre Established Mature Neural Networks ................................ ................................ ................................ ....... 53
6 Co Culture with Electrophysiologically Active Primary Neural Cultures Promotes Differentiation and Maturation of NPC derived Neurons ............... 54 Discussion ................................ ................................ ................................ .............. 57 3 MICROELECTRODE A RRAY TECHNOLOGY REVEALS THAT VARIABLE DURATION OXYGEN GLUCOSE DEPRIVATION MAY MODEL DIFFERENT ASPECTS OF STROKE ................................ ................................ ......................... 62 Introduction ................................ ................................ ................................ ............. 62 Material and Methods ................................ ................................ ............................. 64 Generation of Primary Cultures ................................ ................................ ........ 64 MEA and Multiwell Chamber LabTek Preparation ................................ ............ 64 Oxygen Glucose Deprivation ................................ ................................ ............ 65 Immunocytochemistry and Microscopy ................................ ............................. 66 Electrophysiology ................................ ................................ ............................. 67 Statistical Analysis ................................ ................................ ............................ 69 Results ................................ ................................ ................................ .................... 70 Mature Neurons are Particularly Vulnerable to the Effects of Long Duration OGD ................................ ................................ ................................ .............. 70 Neuronal Network Activity in Primary Cultures Matured Along an Expected Timeline ................................ ................................ ................................ ......... 75 OGD Exposure Affects Network Activity in a Time Dependent Fashion ........... 77 Neuronal Activity Recovers in an OGD Exposure Duration Dependent Manner ................................ ................................ ................................ .......... 79 Discussion ................................ ................................ ................................ .............. 83 4 ADULT NEURAL PROGENITOR CELLS PARTIALLY PROTECT MICROELECTRODE ARRAY PLATED NEURAL NETWORKS FROM O XYGEN GLUCOSE DEPRIVATION INDUCED DEATH AND SILENCING ......... 89 Introduction ................................ ................................ ................................ ............. 89 Material and Met hods ................................ ................................ ............................. 92 Primary Culture Preparation ................................ ................................ ............. 92 Neural Progenitor Cells ................................ ................................ .................... 94 Human Dermal Fibroblasts ................................ ................................ ............... 95 Oxygen Glucose Deprivation and Co Culture ................................ .................. 96 Immunocytochemistry and Microscopy ................................ ............................. 97 Electrophysiology ................................ ................................ ............................. 98 Statistical Analysis ................................ ................................ ............................ 99 Results ................................ ................................ ................................ .................. 100 NPCs Transiently Pro tect Neural Cultures from Death and Oxidative Stress 100 OGD Exposed Primary Cells Did Not Affect NPC Fate ................................ .. 105 Neuronal Death Following OGD is Less Severe Following NPC Addition ...... 108 OGD Increases Cu/Zn SOD in Neurons Immediately, and in Astrocytes After 14 Days ................................ ................................ .............................. 111 NPC Addition After OGD Protects Neural Activity and May Lead to New Activity ................................ ................................ ................................ ......... 113
7 NPC Addition Leads to Partial Recovery of Network Activity Following OGD 113 Discussion ................................ ................................ ................................ ............ 117 5 GENERAL DISCUSSION ................................ ................................ ..................... 123 Summary ................................ ................................ ................................ .............. 123 Future Work ................................ ................................ ................................ .......... 128 Conclusions ................................ ................................ ................................ .......... 132 APPENDIX: CHAPTER 2 SUPPLEMENTARY DATA ................................ ................. 134 LIST OF REFERENCES ................................ ................................ ............................. 137 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 154
8 LIST OF FIGURES Figure page 2 1 Neurons in adult hippocampal NPC derived cultures fail to develop network activity. ................................ ................................ ................................ ............... 45 2 2 Adult hippocampal NPCs produced neurons that retain immature phenotypes and few astrocytes. ................................ ................................ ............................. 46 2 3 Primary cultures develop stable network bursting within weeks of plating. ......... 49 2 4 NPCs and/or their progeny revert neuronal ensemble activity to an immature status. ................................ ................................ ................................ ................. 50 2 5 Dissociated primary cultures produce robust numbers of neurons, astrocytes and oligodendrocytes ................................ ................................ ......................... 52 2 6 Adult NPC progeny mature more quickly and robustly when co cultured with electrophysiologically mature oligodendrocyte and astrocyte rich primary cultures. ................................ ................................ ................................ .............. 56 3 1 Significant cell death occurs after 120 min of OGD exposure, primarily in mature neurons ................................ ................................ ................................ .. 72 3 2 OGD increases the presence of vesicular glutamate and GABA transporters, and Cu/Zn SOD ................................ ................................ ................................ .. 74 3 3 Primary cultures provide a stable baseline of activity 35 40d in vitro ................. 76 3 4 MEA recordings during OGD reveal an increase in asynchronous activity before complete silencing ................................ ................................ ................... 78 3 5 OGD causes decreased spike amplitudes, silencing of network activity by ~15 min, but full recovery if reperfused within 20 min ................................ ......... 82 4 1 NPC ad dition ameliorates initial cell death, but OGD increases the presence of pyknotic nuclei, Cu/Zn SOD and immature neurons ................................ ..... 104 4 2 Co culture with OGD exposed cells does not affect NPC fate .......................... 107 4 3 Cell death following OGD is less severe with NPC addition ............................. 110 4 4 Oxidative stress increases in neurons following OGD, but is not changed by NPC addition ................................ ................................ ................................ .... 112 4 5 NPC addition may lead to new activity. ................................ ............................ 114 4 6 NPC addition a fter OGD leads to partial recovery of neural activity ................. 116
9 LIST OF ABBREVIATION S BDNF Bone derived neurotrophic factor BrdU B romodeoxyuridine CNS Central nervous system DAPI diamidino 2 phenylindole DMEM EGF Epidermal growth factor ES Equine serum ESC Embryonic stem cell FBS Fetal bovine serum FEP Fl u o rinated ethylene propylene FGF Fibroblast growth factor GABA gamma Am inobutyric acid GFAP Glial fibrillary acidic protein GFP Green fluorescent protein LTP Long term potentiation MEA Microelectrode array NMDA N Methyl D aspartic acid NPC Neural progenitor cell NT 3 Neurotrophin 3 OGD Oxygen glucose deprivation PBS Phosphate buffered saline PEI P olyethyleneimine vGAT Vesicular GABA transporter vGlut Vesicular glutamate transporter
10 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy THE FUNCTIONAL INTEGRATION OF ADULT NEURAL PROGENITOR CELLS INTO INTACT AND COMPROMISED PRIMARY NEURAL NETWORKS IN A MICROELECTRODE ARRAY CO CULTURE SYSTEM By Crystal Lynn Stephens May 2012 Chair: Brandi Ormerod C ochair: Thomas DeMarse Major: Biomedical Engineering Neural stem/progenitor cell research is on the cusp of achieving real strategies for central nervous system repair. Recent studies provide evidence that grafts can survive, functionally integrate, and ultimately augment damaged neural networks. However, understanding the impact of functional integration on existing brain circuitry is an understudi ed area of research. This dissertation employs m icroelectrode array s (MEAs), which facilitate real time optical and ele ctrophysiological data collection across entire neural network s, to first determine the faculty for adult hippocampal neural progenitor c ell (NPC) derivatives to produce spontaneous action potentials and network bursts when cultured alone, and then to investigate the effects of adult NPC addition to intact or damaged primary neural networks MEA plated primary cultures transition through h i ultimately refined into mature, regular burst patterns. When plated alone, adult NPC progeny fail to establish synchronized bursting that primary and embryonic stem cell derived cultures readily form. However, NPCs evok e re emergent superbursting in primary neural cultures. Developmental superbursting is thought to accompany heightened plasticity both in culture
11 preparations, and across brain regions. This work also uses online MEA recordings to show activity degradation during hypoxia ischemia via oxygen glucose deprivation (OGD), and investigate the ability of subsequent NPC addition to protect or restore network activity Spike rates initially rise during the first 5min of OGD due to increased asynchronous activity bet ween bursts. Action potentials decrease in amplitude, and all activity ultimately ceases by ~16min. Spontaneous activity recover s following reperfusion after 20min OGD but little recovery is observed after longer durations Shorter duration OGD may effect ively model the silenced circuits found in the penumbra of a stroke while longer OGD might effectively model the dying circuits found in the core. NPC addition 24hrs following 3hr OGD increase s low level proliferation, and decreases cell death, preferentia lly protecting mature neurons from oxidative stress and death for at least a week after OGD exposure. Although co culture in OGD compromised primary cultures did not affect NPC behavior, NPCs partially rescued network activity. The MEA co culture model est ablished in this dissertation is an effective bioassay for examining the effects of potential neuroregenerative therapies on existing network activity, before these strategies are translated in vivo
12 CHAPTER 1 GENERAL INTRODUCTION The Concept of Neuroregeneration Before the discovery of adult neurogenesis by Altman and Das in the 1960s (Altman and Das, 1965) the scientific community generally believed that the central nervous system (CNS) had little or no ability to regenerate. The complexity of the structure and function of the brain was thought to be too delicate to be disrupted by the introduction of new cells. This dogma, established by Ramn y Cajal persisted across decades of neuroscience research (for review, Colucci D'Amato et al., 2006 ). Decades later, the in vitro isolation and expansion of dividing neural cells in the 1990s (Reynolds and Weiss, 1992; Lois and Alvarez Buylla, 1993; Luskin, 1993) erupted a new could be utilized for neural repair (Gage, 2000) If new neurons were added to the intact adult brain on a daily basis, perhaps they could be introduced into an injured or diseased brain to replace or augment damaged tissue. Although animal models an d even clinical trials have supported the notion that neural stem cell transplantation, or their derivatives, is a feasible treatment strategy, studies have provided mixed results in their demonstrations of functional and behavioral recovery (for example, Freed et al., 2001; Olanow et al., 2003 ). This is partly due to confounding factors, such as inconsistent number and/or type of cells grafted, and varying methods of transplantation. Means to increase harvest yield, quickly proliferate, and differentiate s tem cells into specific target cell types are currently under investigation. However, functionally integrating cell grafts into mature neural tissue, promoting cell survival and controlling interactions with existing neural tissue to replace
13 damaged or dea d cells remains elusive. Perhaps a highly sensitive, noninvasive, in vitro system for testing the effects of new treatments could be used to answer some of these questions in an inexpensive, yet remarkably informative, manner. Neural Progenitor Cell Origin s and Other Stem Cell Sources Stem cells are unique populations of cells that exhibit clonality, the ability to self renew, and potentiality, the ability to generate multiple cell types. Recent convention refers to less clonal and less potential cells as p rogenitor cells (Fabel et al., 2003) Therefore, neural stem cells harvested from adult neural tissue that divide, but generate only cell types of the CNS will be described in this dissertation as neural progenitor cells (NPCs). NPCs are committed to diffe rentiating into the neuron, astrocyte and oligodendrocyte cell subtypes found in the brain and spinal cord (Gage, 2000) Neurons are the functional cells of the nervous system that communicate with one another typically through chemical messages transmitte d across synapses that generate propagated spikes of cell membrane voltages, called action potentials. Astrocytes and oligodendrocytes generally serve as supporting cells for the neural network (Kandel et al., 2000) There are several different sources of NPCs, each with unique advantages for neural repair (Ormerod et al., 2008) NPCs can be produced from embryonic stem cells (ESCs) or harvested from fetal or adult brain tissue (Palmer et al., 1997; Svendsen et al., 1999; Palmer et al., 2001) ESCs were first generated from mouse blastocysts by culturing the inner cell mass from pre implantation embryos on fibroblast feeder layers. The resulting cultures contain populations of ESCs that grow in colonies (for review, Smith, 2001 ). Thomson et al. later isol ated human ESCs that do not require feeder layers (Thomson et al., 1998) ESC derived neurons develop spontaneous action
14 potentials after 21 days and limited synchronous bursting after 5 6 weeks (Ban et al., 2007; Illes et al., 2007) However, they can dev elop into any cell type, including teratomas, and of course have obvious ethical issues. Although the production of teratomas, or a tumor containing all three germ layers, in immunodeficient mice is actually a test of clonality and multipotency used to sho w the successful generation of ESCs (Smith, 2001) their proficiency in forming teratomas is a serious health risk and barrier to their clinical use. The risk of tumorogenesis, however, can be overcome by pre differentiating all cells before engraftment (B rederlau et al., 2006) attempted to overcome some of the ethical and medical issues of ESCs by introducing ESC specific genes into somatic cells using retroviral vectors to generate induced pluripotent stem cells (iPSCs) (Takahashi and Yamanaka, 2006; Yu et al., 2006; Takahashi et al., 2007) Given the correct conditions iPSCs can generate cell types from any germ layer, including mature neurons, but these cells also require pre differentiation before engraftment (Takahashi et al., 2007). NPCs can also be derived from immortalized lineages of transgenic mice that can be readily propagated, and can differentiate and integrate in an engraftment site specific manner, but these cells lack the ability to survive long t erm when implanted into adult mouse brains (Snyder et al., 1992) Furthermore, immortalized neural stem cell lines differ significantly from nonimmortalized lines, and even subclones of the same line, because they secrete additional growth factors and cyto kines, express different genetic markers, and do not maintain karyotypic stability (Mi et al., 2005) These cell lines
15 cannot be generalized and are used with caution depending on the goal of the experiment. Adult NPC s exist throughout the fetal and adult mammalian brain but only actively divide to produce neuroblasts in the subventricular zone (SVZ) of the lateral ventricles, and the subgranular zone (SGZ) of the dentate gyrus, supporting neurogenesis in the olfactory bulbs and hippocampal granule cell lay er, respectively, throughout life. In the rodent brain, and presumably the human brain, the rostral migratory stream contains progenitor cells that migrate from the SVZ, differentiating and dividing as the cells travel the long distance towards the olfacto ry bulb where they continually replenish the mature neuronal population (Doetsch et al., 1997) Evidence of a rostral migratory stream in the adult human brain is still under debate (Sanai et al., 2004; Curtis et al., 2007) New cells of the SGZ of the den tate gyrus migrate to the granular cell layer (GCL) of the hippocampus, where they mature into granule cell neurons and exhibit electrophysiology that may contribute to long term memory formation and influence behavior (Aimone et al., 2006) Adult NPCs can be harvested from adult brains and grown continuously in vitro as floating neurospheres or monolayers (Gage et al., 1998) In addition, media conditions have now been formulated that promote differentiation of neural or glial populations, however the shif t in fate is not that robust (Takahashi et al., 1999) Nonetheless, the addition of neurons to existing hippocampal and olfactory bulb neural circuits suggests that the brain holds potential to integrate new neurons following injury or disease. The origin of NPCs, or whether more than one type of NPC exists, remains under debate. M ature quiescent astrocytes start to proliferate after injury and have the
16 potential to form neurospheres in vitro However, these cells do not stray from astrocyte lineage in vivo (Buffo et al., 2008) Due to the different neurogenic niches in the adult brain, the types of cells that are produced may be influenced by the interactions between NPCs and their local environment that predict their mature fate as well as their li neage (Goldman, 2003) For example, NPCs harvested from the SVZ differentiate into hippocampal neurons when transplanted into the SGZ (Suhonen et al., 1996) Recent experiments suggest that there may be two distinct populations of neural progenitors in th e adult CNS (Zhao et al., 2008) In the adult SVZ, a GFAP positive stem cell gives rise to proliferative NPCs that produce migrating neuroblasts, as well as oligodendrocytes (Merkle et al., 2004) The work of Steiner et al supported Goldman by demonstratin g that proliferating cells in the SGZ could either be glial fibrillary acidic protein (GFAP, known to be an astrocyte marker) positive or double cortin (DCX, an immature neuronal marker) positive but never both, indicating two independent populations of dividing cells in the glial and neuronal lineages, respectively. Because proliferating cells had the propensity to express GFAP but no other astrocyte markers or astrocytic characteristics, GFAP expression alone may not be enough to characterize a cell as an astrocyte (Steiner et al., 2004) Current techniques used to harvest adult NPCs are not specific to any one cell type. I f more than one progenitor cell exists, adult NPC lines would have a heterogeneous population. The cells used in for the experiments in this dissertation do not express GFAP when fixed immediately after a 12 h our exposure to BrdU, a cell division marker, suggesting that they are not a GFAP+ve population of cells. Regardless of the true origin of endogenous progenitor
17 cells, they may be great sources of regenerative potential if researchers can engineer them to produce the desired phenotypes for any particular treatment. Adult NPCs were chosen for the experiments in this dissertation because they can be harvested from multiple adult, livi ng and cadaveric CNS regions, they exhibit restricted neural lineage potential, and are relatively safe and karyotypically stable at moderate passages (Reynolds and Weiss, 1992; Palmer et al., 1995; Arsenijevic et al., 2001; Palmer et al., 2001; Seaberg et al., 2005) These properties are advantageous in that adult NPCs could potentially be harvested and transplanted autologously, which reduces the need for donors, and diminishes the risk of immune response to the transplant. Recipients would also be safe f rom the risk of tumor development, and the outcome of cell lines, given the right conditions, could be generalized. However, unlike ESC derived neurons, the ability of adult NPC derivatives to form neural networks is unknown. Therefore, experiments include d in Chapter 2 of this dissertation will plate NPCs alone, and look for development of synchronous bursting evidence of network formation The Neurogenic Niche Provides Clues for Therapeutic Strategies Neurogenesis is a dynamic process that is affected b y both exogenous factors and cell intrinsic factors. The anatomy and physiology of regions promoting neurogenesis has been examined extensively (Doetsch et al., 1997; Palmer et al., 1997) however, t he key to maintaining neurogenesis in these unique niches of the brain remains elusive. The observation that some areas of the brain encourage neurogenesis while others appear to inhibit it despite widespread presence of NPCs is interesting because if the mechanisms that cause this phenomenon were discovered, n eurogenesis could be directed where needed for neural repair. Modifying the environment through external
18 behavior (i.e. exercise) modifies behavior via NPC mediated neurogenesis, which is an exciting revelation indicative of an extremely plastic and dynamic process that could be manipulated with clinical therapies. Defining these neurogenic niches is the first step in ultimately engineering therapeutic strategies that either recruit endogenous NPCs, or direct transplanted NPCs. New lineage informa tion is expected in the future as more is being delineated about CNS cell biology with new technologies. Researchers are gathering lots of information about NPC biology that should provide a toolbox for transplant or recruitment of endogenous NPCs for neural repair. There is some evidence that links neurogenesis to the anatomical organization of vasculature (Mercier et al., 2002; Mirzadeh et al., 2008; Tavazoie et al., 2008) Key factors known to be necessary to grow NPCs alone in culture in order to maintain proliferation and prevent differentiation have been identified whereas other conditions encourage neuronal cell fate and survival. Fibroblast growth factor 2 (FGF) and epidermal growth factor (EGF), for example, activate downstream pathways responsible f or proliferation. Laminin, a CNS extracellular matrix molecule, will encourage neural growth when used as a substrate in culture. Extracellular signaling molecules will also lead to progenitor differentiation. Retinoic acid, for example, will cause a small shift neuronal cell fate and neurotrophins will aid in neural sur vival (Takahashi et al., 1999) Full understanding of the mechanisms driving NPC division and differentiation of their progeny must be obtained before NPCs can have clinical applications. T he recipient brain region may need to be prepared, by perhaps adding these factors during transplant, if the target region is not a reliable source for generating them alone.
19 Adult Neurogenesis: A Benchmark for Functional I ntegration In order to engineer t ailored strategies for a particular disease or injury with a predictable and favorable outcome, researchers need to decode how to control functional integration or the formation of new networks, of transplanted or endogenous NPC derived neurons. Adult neur ogenesis is an excellent example to turn to for clues about how to introduce transplanted cells, because new endogenous neurons integrate into the hippocampus and olfactory bulb in the adult mammalian brain on a daily basis. The hippocampus is involved in the encoding and storage of episodic memory in humans (Scoville and Milner, 2000; Leuner et al., 2006) Researchers have assumed for years that the new GCL neurons in the adult brain are involved with this type of learning, b ut due to the indirect nature of the experiments, relationships between new neuron number and measures of hippocampal function remain correlational. Rodent models focus on using navigational memory related tasks such as the Morris water maze, in which rats learn to swim to a hidden platform in a pool of water to ask how new neurons contribute to the integrity of the hippocampus. Changes in neurogenesis are measured by quantifying both the number of new cells (birthdated by a synthesis phase marker, such as the thymidine analogue bromodeoxyuridine; BrdU) produced in the dentate gyrus and their phenotype as well as by the number of new neurons that survive over time (Ormerod and Galea, 2001) The contribution of new granule neurons to hippocampal function ha s not yet been revealed, but studies have shown that strains of mice with high level neurogenesis outperform strains with low level neurogenesis on hippocampus dependent t asks (Kempermann and Gage, 2002) and neurogenesis has been linked to learning and me mory by several groups (Gould et al., 1999; Kempermann and Gage, 2002; Ormerod
20 et al., 2004; Jaako Movits and Zharkovsky, 2005; Winocur et al., 2006) Manipulations that chronically deplete hippocampal neurogenesis also produce concomitant impairments in c ognition, particularly hippocampus dependent memory (Raber et al., 2004; Rola et al., 2004; Snyder et al., 2005) These data further support the notion that established neuronal circuits in the adult brain are amenable to the functional integration of new neurons In addition, introducing animals to a running wheels or exposing them to enriched environments increases the rate at which new neurons are produced and the proportion of neurons that survive over time, respectively (Ehninger and Kempermann, 2003; Tashiro et al., 2007) Although evidence is mounting that neurons produced in the adult hippocampus do integrate into existing circuitry to improve function, variability in the results of many studies, likely due to variability in experiment (BrdU injectio n) and task (trials per day, room setup, etc), have made the establishment of a direct connection difficult (Piatti et al., 2011; Marrone et al., 2012; Song et al., 2012) The maturation of endogenous new adult hippocampal neurons closely follows the devel opment of the embryonic brain, taking about 4 weeks to exhibit both GABAergic and glutamatergic synapses, although the exact function of these new cells remains unknown (Esposito et al., 2005) When injected into the dentate gyrus, whole cell recordings of these cells do show some evidence for spontaneous a ctivity (van Praag et al., 2002; Piatti et al., 2011) An approach in which the integration of new cells can be monitored over time may facilitate our understanding of whether transplanted cells or mobili zed endogenous cells could be enticed to integrate functionally into a diseased or injured circuit.
21 Transplantation S tudies Provide Evidence of Functional Integration NPCs hold the potential for the clinical treatment of neurodegenerative disease such as disease, or stroke, to name a few (Levy et al., 2004; Roybon et al., 2004) disease essentially impairs motor skills and speech as a result of decreased stimulation of the motor cortex from the basal ganglia due to insufficient dopamine pro duction (Hoglinger et al., 2004) NPC trans plantation successful replace the dopaminergic cells. In 1989, Lindvall and colleagues reported the resu lts of the first human clinical trial, which transplanted ventral mesencephalon tissue from 8 10 week old aborted human fetuses into the putamen of modest motor improvement in 2 late stage patients. Of the approximately 300 P patients who have since undergone fetal tissue transplants, only a few have enjoyed long term positive clinical benefits (Clarkson, 2001; Le Belle and Svendsen, 2002; Winkler et al., 2005) Because of the efficacy of this appr oach in some patients the National Institutes of Health initiated two double blind placebo controlled studies in 1993 In the Denver Col umbia trial (Freed et al., 2001) no persistent improvement in the motor part of the UPDRS was observed among transplan t patients (n = 19; at 12 months). However, in a subgroup of younger patients (< 60 years old) a statistically significant 30 35% reduction was observed across the post graft evaluation period. In the Tampa Mount S inai trial (Olanow et al., 2003) a progressive 6 month post surgical improvement in UPDRS scores was observed but gradually declined to no significant difference at 24 months. There are likely compounding factors that led to the failure of the placebo controlled trials that include lack of immunosuppression, quality and
22 consistency of transplanted tissue, patient selection, transplant parameters and unexpected side effects such as dyskinesia. on transplanted cells is not fully understood, the eff ects of hormones, stress, neuroinflammation and lif e experience on endogenous NPCs has been well documented (Gould et al., 1992; Cameron et al., 1993; Nilsson et al., 1999; van Praag et al., 1999; Mizumatsu et al., 2003; Monje et al., 2003; Ormerod et al., 2003) Clearly these crucial issues remain to be resolved and are discussed in a recent review by Winkler et al., 2005 The clinical improvement shown in the subgroup of younger patients provides hope that stem cell strategies would produce functional imp rovement if more parameters for success were understood. Monitoring t he extent, to which implanted NPCs become functional in vivo, is very limited using current engraftment techniques. There have been more experiments examining neural networks generated fr om ESCs than from adult NPC derived cells perhaps because adult NPC derived cells are more difficult to grow in culture than ESCs. ESCs can give rise to inhibitory and excitatory neurons (Bain et al., 1995; Strubing et al., 1995) hESC derived neurons rece ive postsynaptic currents both in vitro and in vivo and begin to burst via synaptic integration (Weick et al., 2011) When injected into fetal mouse brains, mouse ESC derived NPCs demonstrate widespread migration and differentiation into multi ple neuronal subtypes (Wernig et al., 2004) Looking at single cell development in vitro they appear to extend outgrowths contacting other ESCs and have been shown to become electrically excitable with patch clamp after 11 to 14 days (Copi et al., 2005) When plated onto adult hippocampal slices, they
23 migrate into the slice and receive input from primary cells, expressing excitatory (AMPA and NMDA) and inhibitory (GABA) receptors (Benninger et al., 2003) When pre differentiated human ESCs were transplanted into a 6 O HDA (6 hydroxydopamine) had excellent survival rates and expressed neuronal markers. However, few displayed a dopaminergic neuronal phenotype. Reversal of lesion induced motor deficits was not observed. In addition, ESCs pre differentiated for less than 20 days developed into severe teratoma, indicating that prolonged pre differentiation is essential when using ESC deriv ed NPCs (Brederlau et al., 2006) In a similar experiment, adult derived NPCs pre differentiated into dopaminergic neurons were transplanted into the striatum of the same 6 OHDA lesion rat model with the addition of glial cell line deri ved neurotrophic factor (GDNF), known to enhance the survival of dopaminergic n eurons. However, l ong term survival of the implanted NPCs was not established and behavior after transplantation did not improve (Ostenfeld et al., 2002) Plating NPCs on an established long term biologically relevant neural circuit will have many advantag es. In vitro analysis will allow for an easily accessible model to examine the interactions between NPCs, and between NPCs and primary neural cells. The cellular activity can be tightly controlled and monitored for long periods of time, which allows for an in depth analysis of how the surrounding e nvironment influences cell fate, and in turn, how transplanted cells influence existing brain circuitry. Grafted cells could potentially cause serious problems, and it is important that we understand these possibi lities. Generating NPCs for transplant into the adult brain for the purpose of neural repair is a growing focus of research but it remains unknown how and if these
24 cells will functionally integrate into the existing neuronal circuits. The idea of using cel l transplants to treat neurodegenerative disease is not new, and early work has shown some promise and some pitfalls. Although transplant studies can detect effects of functional integration through changes in behavior, and then posthumously look for histo logical markers as evidence of differentiation, synaptogenesis and synaptic activity, MEAs can serve as a missing link, an assay of electrophysiological changes in real time. A system stripped of systems regulators, known to influence neurogenesis, where p rogenitor cell progeny could be directly observed over time, may contribute to a clearer picture about added cell fate and functionality Microelectrode Array (MEA) Technology For this dissertation, co culture methods and MEA technology were combined to d evelop a novel in vitro cell transplantation model in order to provide simultaneous access to optical and electrophysiological data across an entire neural network from 59 electrodes, while easily manipulated by adding cells, factors and/or simulated injur y. Primary, embryonic day 18 dissociated cortical cells have been well characterized on MEAs, and consistently develop a mature, stable network of monolayer cultured primary cells that can be grown for long term experiments lasting months or even years (Ka mioka et al., 1996; Potter and DeMarse, 2001) MEAs are an effective platform for investigating the internal dynamics of neuronal networks, pattern generation and evolution of network organization as the dissociated fetal cells organize themselves into spo ntaneously active dynamic systems. After these cultured networks have matured they become functionally stable in terms of axon density, inter circuit connectivity, and plasticity (Gross and Kowalski, 1999) permitting chronic recording and manipulation. Therefore, primary cultures grown on MEAs appear to establish a biological neural
25 circuit that could be used to examine how NPC transplantation would influence a diseased or damaged circuit and in turn, how a n established circuit would influence the behavior of transplanted NPCs. Van Pelt, et al characterized the development of cultured neural networks on MEAs. Spontaneous action potentials are observed within one week with periods of elevated firing rates. I n the second and third week s, network bursts, short episodes of sustained synchronous firing at many sites followed by a variable quiescent phase with few action potentials, are present. The bursts evolve gradually, broadening to about two seconds in the t hird week followed by drastic shortening after one month. In earlier weeks (before 30 days in vitro ), network bursts are preceded by a ramp like phase of low firing, followed by almost or complete silences. At later stages such bursts have a sudden onset and are often followed by an after discharge of lower intensity firing that is later described as superbursts At the end of three weeks in vitro bursts reach their max imum duration. This is attributed to maximum dendritic spine and shaft synapse density numbers. GABAergic inhibitory feedback and dendritic spine pruning cause these long, less intense bursts, which shift the balance of synaptic excitation and inhibition towards th e latter (van Pelt et al., 2004) In immature neurons, gamma aminobutyric acid (GABA), the inhibitory neurotransmitter, exerts dual properties (both excitatory and inhibitory actions). This phenomenon is caused by the reversed potential of the GABA A receptor activated chloride conductance (E GABA) that reverses at 35mV. On the one hand, GABA A receptor mediated depolarization excites immature neurons. Also, it activates voltage gated Ca 2+ channels and potentiates the activity of the NMDA receptors (excitatory) via
26 attenuation of their voltage dependent Mg 2+ block. GABAergic synapses i n the fetal rat hippocampus are established before glutamatergic ones. Glutamate synapses will not be established as long as the target neuron has not reached a certain degree of maturity, which occ urs in 3 4 weeks (Ben Ari, 2001) ESCs cul tured on MEAs g ive a good indication of how adult NPCs would behave. ESC derived neural culture s developed spontaneous action potentials aft er 21 days (Mistry et al., 2002) and synchronous burst patterns after 5 6 weeks (Illes et al., 2007) A ctivity development of ESC d erivatives is slightly delayed compared to primary cortical cultures, which exhibit spontaneous action potentials as early as three days in vitro and synchronous network bursting by two weeks in vitro (Maeda et al., 1995) Thus, adult NPCs should be expect ed to also show delayed maturation. Electronic stimulation or delivery of differentiation factors may also be needed to control progenitor cell fate, synapse formation and electrical response patterns. For therapeutic applications NPC derived neurons must integrate into the brain by forming the correct syna pses as to be beneficial to brain function A dding new neurons that develop undesired de novo networks c ould cause more deleterious effects than the disease that is being treated by causing problems suc h as epilepsy from over excitation of the surrounding tissue, or loss of brain function due to inhibition (Buzski et al., 1991) The MEA co culture model described in this dissertation could help to understand the effects of added NPCs to existing CNS circuitry, a vital part of engineering transplantation strategies. Investigating NPC Addition to Damaged Neural Networks Using a Hypoxic Ischemic Stroke Model Applying the MEA co culture assay system to test the effect of transplanted cells on damaged neur al circuitry could uncover valuable information about the mechanisms
27 of functional recovery in the absence of confounding systems regulators. Previous experiments used MEAs to monitor the effects of hypoxia/ischemia on cardiomyocytes (Yeung et al., 2009) ; however, stroke models in cortical networks have only previously been recorded using 1 7 electrodes in vivo (Kass and Lipton, 1982; Rosner et al., 1986) Ischemic stroke is an excellent prototype of acquired brain injury, because in vitro and in vivo model s have been established, are well characterized, and affect many areas of the brain Cell death and loss of function in the central nervous system (CNS) due to ischemia is a complicated cascade of cellular mechanisms that in recent years has gradually been decipher ed. Briefly, lack of blood flow leads to a rapid depletion of ATP that halts active Na + /K + pumps (Madl and Burgesser, 1993) and in turn causes cells to swell, and reverse glutamate transporters. Located mainly on astrocytes, glutamate transporter s use the sodium gradient to clean glutamate from the extracellular space to prevent neuronal excitotoxicity (Zerangue and Kavanaugh, 1996) However, ischemic conditions reverse the gradient, and glutamate is released out into the extracellular space (Szat kowski et al., 1990) During severe ischemia, neurons develop a large membrane depolarization due to the influx of Na + (Calabresi et al., 1999) u ltimately leading to mitochondrial dysfunction, generation of reactive oxygen species (including the over expression of Cu/Zn superoxide dismutase (SOD; Fukui et al., 2002) and cell death (Rossi et al., 2000) White matter is not excluded from injury as immature oligodendrocytes and oligodendrocyte progenitors are particularly susceptible to ischemia related death (Fern and Moller, 2000; Tekkok and Goldberg, 2001; Back et al., 2002) Previous studies have shown that NPC grafts increase neuron survival and
28 promote limited recovery following ischemia in vivo however, how they preserve funct ion and whether they increase survival of other cell types remains unclear (Hodges et al., 1996; Sinden et al., 1997; Modo et al., 2002; Harms et al.) Hypotheses regarding the mechanisms of functional improvement following the transplantation of stem cell s or their progeny suggest that they may provide trophic support that thwarts further cell death, stimulate an endogenous repair process or the production/secretion of neurotransmitter, or perhaps integrate into existing circuitry to amplify compromised si gnaling (Ben Hur et al., 2004; Ormerod et al., 2008) Neural stem cell grafts into ischemic CA1 lesions in the rat hippocampus exhibit pre and post synaptic markers and develop properties of mature, normal hippocampal neurons within the infarct (Aoki et a l., 1993) NPCs also induce plasticity through secretion of trophic factors (Andres et al. 2011 ; Modo et al., 2003; Lee et al., 2007) Additionally, s ystems regulators might have an influence or interaction with any protective effects or functional integr ation of added cells A post hypoxic/ischemic environment may affect the behavior of NPCs because, for one, NPCs have metabolic glutamate receptors that are implicated in the regulation of proliferation, differentiation, and survival (Melchiorri et al., 20 07) Proliferation of endogenous NPCs does in fact increase within minutes of ischemia in the dentate gyrus (Liu et al., 1998; Takagi et al., 1999; Kee et al., 2001; Yagita et al., 2001) Neurogenesis is also induced in nonneurogenic regions, specifically the striatum, where neuroblasts migrate from the SVZ after stroke and differentiate into functional neurons resembling those that died in the lesion (Arvidsson et al., 2002; Thored et al., 2006) Activated neural stem cells also contribute to stroke induce d neurogenesis and neuroblast migration toward the infarct boundary (Zhang et al., 2004)
29 How and if stem cells or their derivatives are influenced by these mechanisms remains unclear. Applying MEA technology to stroke models can facilitate online monitori ng of network activity during a hypoxic ischemic event, and may shed light upon how to improve the longevity of new neurons. Specific Aims The goal of this dissertation is to investigate the ability of adult NPC generated neurons to form neuronal networks, and influence existing neural activity in mature neuronal networks that are intact, or compromised by OGD. Concurrently, exhaustive phenotyping of the composition of NPC progeny, and the primary cultures, is performed to better understand the electrophysi ological data, and in turn determine how the surrounding primary cells influence NPC maturation The experiments described in Chapter 2 aim to study the effects of adding adult NPCs on electrophysiological activity in an in vitro cell culture model of a mature neural network. The hypothesis for this study is that g iven the correct environment, NPC derived neuron s c ould functionally integrate into an existing neural network to significantly impact its activity as the new cell population differentiates and matures NPCs will first be plated alone to determine whether they are capable of forming neural networks and spontaneous bursting. Then, primary neural cells will be plated and allowed to reach a mature, stable state of bursting before GFP expressing NPC s are added to the cultures. The addition of NPCs is expected to initially lower network activity as they generate the appropriate proteins for transmission but as the cells mature, they will function similarly to neurons surrounding them, thus strengthen ing the network. C hanges in spike rate per channel, and overall spike rate burst rate, and superburst
30 rate per array are measured. Concurrently, changes to NPC fate will be examined immunocytochemically. The studies described in Chapter 3 aim to establish an ischemic stroke model by applying OGD to primary cultures on MEAs. Recovery of network activity along with cell death rates will be measured after different lengths of OGD exposure times in order to determine a benchmark for modeling the effects of NPC addition on a compromised network. For the first time, MEA plated cultures are recorded during OGD to delineate how network activity breaks down during ischemic conditions. Particularly vulnerable cell phenotypes will be detected immunocytochemically, in addition to cell death markers and Cu/Zn SOD. Finally, Chapter 4 uses the established in vitro brain injury model for stroke to observe changes in NPC derived neurons due to neural activity in injured neural networks. Twenty four hours following OGD expos ure, NPCs, NPC conditioned media or fibroblasts are added to the post ischemic primary neural cultures in order to determine if any neuroprotective effects are due to NPC addition, trophic factors secreted by NPC derived cells, or by the physical addition of cells (Boehler et al., 2007) Changes in cell death, oxidative stress, and the fate and survival in GFP positive NPC progeny will be examined immunocytochemically. The hypothesis for this study is that NPC derived cells will functionally integrate into a mature network in an attempt to repair damaged primary neural cultures Neural activity after the cell addition will be monitored to 1) look for any rescued activity due to the addition of NPC s as evidence that the NPCs can r epair the damaged network and 2) observe NPC derived cell survival and differentiation on any remaining active sites versus inactive sites. The work
31 in this dissertation will demonstrate the convenience and applicability of the MEA co culture system to tes t the effects of NPC addition and/or other potential therapies on a biologically relevant neural network.
32 CHAPTER 2 ADULT NEURAL PROGENI TOR CELLS REACTIVATE SUPERBURSTING IN MATURE NEURAL NETWOR KS Introduction Reports of potentiated behavioral recovery in some animal models of human CNS syndromes suggest that transplanted stem cell derivatives may replace or augment damaged neural networks (Nikkhah et al., 1994b) However, the mechanism(s) behind potentiated rec overy in animal studies remain elusive (Winkler et al., 2005; Ormerod et al., 2008). Monitoring the effects of stem cell derivatives on neural activity as they differentiate and mature could provide insight about the optimal parameters (i.e. cell numbers a nd endpoints) to employ in preclinical studies (Winkler et al., 2005) Here we use microelectrode array (MEA) technology, which allows the longitudinal collection of both optical and electrophysiological data across neural cultures (Potter and DeMarse, 200 1; Maccione et al., 2010), to test whether long term adult rat neural progenitor cell (NPC) generated cultures generate de novo networks and/or integrate functionally into pre established rat primary neural networks. Employing stem cell sources that generate neurons with spontaneous networking capability may facilitate stem cell strategies seeking to replace dead neural circuits. MEA technology has revealed that human and mouse embryonic stem cell (ESC) derived neurons generate intensity dependent evoked potentials and spontaneous synchronized network bursts within ~5 6 weeks of plating that are remarkably similar to those generated by networked primary neurons (Ban et al., 2007; Illes et al., 2007; Heikkila et al., 200 9) Because NPCs can be harvested from multiple adult living and cadaveric CNS regions, exhibit restricted neural lineage potential, and are relatively safe and karyotypically stable at moderate passages (Reynolds and Weiss, 1992;
33 Palmer et al., 1995; Arse nijevic et al., 2001; Palmer et al., 2001; Seaberg et al., 2005), they are attractive candidates for CNS replacement strategies. However, their ability to generate neurons with de novo networking capability is unknown. Here we examine whether adult rat hip pocampal NPC derived neurons spontaneously generate novel networks, like their human and mouse ESC counterparts. Decades of MEA research have revealed that primary neurons reliably generate activity patterns that stabilize with time (Gross and Kowalski, 1 999; van Pelt et al., 2004; Wagenaar et al., 2006) For example, fetal primary hippocampal and cortical neurons initially generate isolated action potentials (within 3 5 days) but eventually generate synchronized action potentials (within ~ 2 weeks) that e volve into short high frequency bursts of activity (within ~ 4 weeks) that persist for the lifetime of the culture. Often before stabilizing, these ex vivo neural networks transiently generate clusters of bursts called superbursts (Wagenaar et al., 2006) reminiscent of GABA receptor mediated recurrent burst s generated by neu ral networks in diverse regions of the developing mammalian brain (de Lima and Voigt, 1997; Voigt et al., 2001; Opitz et al., 2002) Here we also test whether the activity of pre existi ng electrophysiogically mature rat neuronal networks are affected by the addition of adult rat hippocampal NPCs Material and Methods Preparation of Microelectrode Arrays Dishes and 8 well LabTek Chamber Slides Electrophysiological data was collected from cells plated on microelectrode array (MEA) dishes with a square grid of 60 planar Ti/TiN microelectrodes 30 m in diameter, spaced 200 m apart (Multichannel Systems, Reutlingen, Germany) and immuno cyt ochemical data was collected from cells plated on 8 well LabTek chambers
34 that were paraformaldehyde fixed at 14 day intervals throughout the experiment. MEA and LabTek plated cultures were generated from the same pool of dissociated cells for each experiment. Just before use, a 100 l drop of 0.1% (w/v) polyethy leneimine (PEI) in borate buffer (Sigma Aldrich ) solution was placed on the center of each autoclaved MEA dish or sterile LabTek well After one hour the dishes and chambers were rinsed and dried before a 20 l droplet of mouse laminin (Sigma; 1mg/ml in PB S) was placed on the center of each MEA (covering the electrode region) or center of each LabTek well which was then incubated at 37 C for one hour Cell suspension was then placed directly onto the laminin droplet, which ensured identical plating densities in the dishes and wells. Once the cells adhered to the laminin covered substrate (~20min), media was topped off to 1ml in MEAs and 400 l per well in the LabTek chambers. Neural Progenitor Cell Harvest and Expansion The rat hippocampal NPCs used in this study were descended from a line generated by Dr. Hiroki Toda (Toda et al., 2003) Briefly, the minced hippocampi of adult female Fisher 344 rats (Harlan, Sprague Dawley) were digested in papain (2.5 units/m l; Worthington Biochemical, Freehold, NJ), dispase II (1 unit/ml; Roche Applied Science) and DNase I (250 units/ml; Worthington Biochemical) solution at 37 C for 45 min and then triturated gently. The dissociated tissue was fractionated in a 50% and then 2 0% Percoll solution by centrifugation for 10min at 20,000 x g to capture NPCs. Fractionated cells were plated onto poly L ornithine/laminin coated dishes in DMEM/F 12 (1:1) medium containing 10% fetal bovine serum (FBS) for 24 h ours and then the medium was replaced with DMEM/F12 (1:1) supplemented with N2 (Invitrogen) and 20ng/ml recombinant human fibroblast growth factor 2 (FGF 2; Genzyme, Cambridge,
35 MA). At approximately passage 20 the NPCs were infected with the replication deficient GFP expressing reco mbinant retrovirus, LZRS CAMut4GFP to express green fluorescent protein (GFP) and passaged weekly until abundant GFP expression was observed All GFP + NPCs used in the current study had been passaged ~32 35 times Previous work has shown that adult NPCs re main genetically stable after this moderate number of passages (Palmer et al., 2001) Preparation of NPC Cultures To test whether long term NPC cultures establish dynamic neural networks and to determine their cellular composition after long term culture, NPCs were plated on MEA dishes and in 8 well LabTek chambers, respectively. To test whether they modulate the activity of mature neural networks, they were also co cultured with mature bursting neural cultures. In both cases, NPCs were thawed and grown fo r 1 week on polyornithine/laminin coated 100mm tissue culture composed of 20ng/ml fibroblast growth factor ( FGF ; Peprotech, Inc Rocky Hill, NJ) 20 ng/ml epidermal growth factor ( EGF ; Sigma ) and N2 ( Gibco /Invitrogen) in DMEM/F12 with L glutamine (Mediatech, Inc) before being passaged in 0.1% trypsin at 37 C. The 2ng/ml FGF, 100nM all trans retinoic acid (Acros/Thermo Fisher) 10ng/ml brain deriv ed neurotrophic factor (BDNF ; Peprotech, Inc), 10ng/ml neurotrophin 3 (NT 3 ; Peprotech, Inc.), and 1% (v/v) fetal bovine serum in D MEM/F12 w ith L glutamine to drive progeny more toward a neuronal versus glial fate (Takahashi et al., 1999) A 100l droplet of media containing 10,000 NPCs was placed in the center of empty PEI /laminin coated MEAs (n=5) or PEI/laminin coated 8 well LabTek chambers ( n=3) Electrophysiological
36 activity was recorded weekly for 8wks from MEA plated cultures and phenotypes were qua ntified using immunocytochemistry and confocal microscopy in LabTek plated NPC cultures that were paraformaldehyde fixed at 2, 4 and 6 weeks after plating. In both growth and differentiation conditions ~70% of cells in the NPC generated cultures expressed green fluorescent protein (data not shown). Preparation of Primary Cultures The neocortices (without hippocampi) dissected from embryonic Day 18 (E18) Sprague Dawley rats (n = 2 per experiment) were shipped to the University of Florida from Brain Bits (Sp ringfield, Illinois) in Hibernate B solution and were stored at 4 C until use. Within 48h of arrival, the tissue was dissociated enzymatically in a solution of p apain ( 20 units /ml ; W orthington Biochemical; Lakewood, NJ) and deoxyribonuclease 1 (2000 units/ ml, Worthington Biochemical) in GlutaMax supplemented D MEM (1x ; I nvitrogen) at 37 C for 30min before being triturated 10x gently and pelleted for 3min at 1500rpm. The pelleted cells were re growth media consisting of 10% equine serum (Hyclone) + 1% Cellgro Antibiotic Antimycotic solution (Mediatech, Inc.; Herndon, VA) in DMEM/GlutaMax (Invitrogen) and plated by placing a 100l droplet of primary culture growth media containing 25,000 cells in the center of PEI laminin coat ed MEA dishes for electrophysiological characterization ( n = 12 ) or PEI/laminin coated LabTeks for immunocytochemical characterization ( n = 17; > 2 wells per staining set). Thus, 25K cells were plated per array or LabTek well (~ 4,000/mm 2 ), which is within plating densities known to generate robust and stable network wide electrophysiological activity (for example Wagenaar et al., 2006). Every 3 days, 50% of the media was replaced with fresh growth media. Media changes were conducted at
37 least 24 hours prior to a recording session to avoid disrupting electrophysiological activity. Preparation of NPCs for Co C ulture At 40 days after plating, dishes were randomly assigned to either the 10,000 NPC, 50,000 NPC, NPC conditioned media or fresh media control group and the co culture experiment was initiated. Media was removed from the primary culture and a NPCs was added to the center of each LabTek chamber well plated primary culture and a 100l drop let containing either 10,000 or 50,000 NPCs was added to each MEA plated primary culture ( Appendix A). Once the NPCs adhered to the primary cultures (5 10 minutes), primary culture growth media wa s added to fill the well or dish ( ~ 400l per LabTeK cha mber well and 1ml per MEA dish) and a teflon lid with a FEP membrane was added (Potter and DeMarse, 2001). Control cultures were fed with 25% NPC conditioned media or fresh media containing no cells. Every 24 hours (to compensate for increased cell densiti es), 50% of the media was replaced with fresh media. Immuno cyto chemistry Lab Tek plated cells were fixed with 4% paraformaldehyde for 10 minutes and rinsed repeatedly between immunostaining steps with TBS (pH 7.4) The cells were blocked in 3% normal donkey serum (NDS) and then incubated overnight in a cocktail of three primary antibodies that included rabbit or mouse anti Tubulin (1:500; Covance; to visualize immature neurons) mouse anti neuronal nuclei ( NeuN ; 1:500; Chemicon I nternational; to visualize mature neurons), guinea pig anti glial fibrillary acidic protein ( GFAP ; 1:750; Advanced Immunochemical; to visualize astrocy tes) or mouse anti S100 (1:500; Sigma Aldrich; to visualize astrocytes), rabbit anti NG2 (1:1000;
38 Chemic on International; to visualize oligodendrocyte precursors ) rabbit anti rat vesicular GABA transporter (vGAT ; 1:500; Chemicon International; t o visualize inhibitory neurons), guinea pig anti vesicular glutamate transporter 1 (vGlut1 ; 1:5000; Chemicon Inter national; t o visualize excitatory neurons) and/or rabbit anti ki67 (1:500; Novocastro distributed by Leica Microsystems; to visualize dividing cells) at 4C. The following day, the sect ions were incubated in the appro priate FITC Cy3 or Cy5 conjugated s econdary antibody (anti goat IGg, anti mouse IGg, anti guinea pig IGg, and anti rabbit IGg, respectively 1:500; Jackson ImmunoResearch) for 4 hours Then cells were DAPI stained (1:20,000 in TBS for 20min; Calbiochem) and then cover slipped under PVA DABCO A subset of MEA plated cells was stained identically to confirm concordance among their phenotypes with LabTek plated cells. In addition, because mean ( S.E.M.) percent of cells co expressing S100 and GFAP was high (96.83 3.16% at 2 weeks, 84.46 8.41% at 4 weeks and 93.53 4.13% at 6 weeks), we employed anti GFAP to phenotype astroglia. Confocal and Epifluorescent Microscopy Confocal lmages (512x512 pixels) of LabTek plated cultures were taken using a Zeiss LSM 710 fully spectral Laser Scanning Confocal Microscope (with 405, 440, 488, 532, 635 laser lines) with a 20x objective (3x optical zoom) through the z plane (5 stacks) of DAPI positive cells and GFP/DAPI + cells in primary cultures and in co cultures. Visible laser line intensities were maintained b elow 15%. DAPI positive cells with appropriate neuron, astrocyte and oligodendrocyte morphologies and intense staining with the appropriate antibodies in primary cultures served as positive controls to which exposure levels and gain were set to exclude no nspecific staining among GFP +
39 cells (NPCs). Cell densities and phenotypes were determined from images taken in 8 10 non overlapping fields of view taken diagonally from each of the 4 corners of > 2 wells (> 500 cells per well). In primary cultures and NPC cultures, the density of DAPI + or DAPI/GFP + cells (number primary cells or NPC generated cells per mm 2 respectively ) the proportion of DAPI + or DAPI/GFP + cells expressing neuronal ( tubulin and/ or NeuN) or glial (GFAP or NG2) markers and the fraction of neurons ( tubulin + or NeuN + ) co labeled with vGlut or vGAT was calculated. In co cultures, the fraction of DAPI/GFP + cells relative to DAPI + cells, the proportion of DAPI/GFP + cells co labeled with neuron al or glial markers and the fraction of new neurons (GFP/ tubulin + ) co labeled with vGlut or vGAT was calculated. Fluorescent images of immunostained MEAs were taken using a Zeiss reflected light fluorescent microscope using a 20x or 40x objective and cel l densities and phenotypes determined as described above Electrophysiological Data Collection Changes in extracellular potential were recorded at each of the 59 recording electrodes over 10 minute sessions. In the NPC culture experiment, activity was reco rded every 2 3 days for 56 days after plating to monitor isolated and synchronous activity. In the co culture experiment, activity was recorded every 3 days from 3 39 days after plating to establish the development of mature electrophysiological activity a mong primary cultures and then every 2 days for 90 days after NPC addition to monitor changes in activity among the mature neural networks The MEA signals sampled at 25 kHz, were amplified and digitized using Multichannel Systems (Reutlingen, Germany) ME A 10 60 amplifier and A/D hardware. The cultures were allowed to equilibrate for ~2min after being placed into the C0 2 and temperature controlled recording chamber
40 before recording was initiated. Spike events were extracted and stored online for later offline processing using a software package custom written by Dr. Thomas B. DeMarse. Action potentials ( or spikes) were detected online as upward or downward excursions beyond 5x the estimated median threshold of quiescent noise estimated during a series of brief 1s periods over 15s before each recording session E l ectrodes reports of minimum steady firing rates (Latham et al., 2000; Boehler et al., 2007) Spike w aveforms occurring 1ms before and 2ms after peak spike height were saved and used to remove duplicate detections of multiphasic spikes and visually inspect for artifacts All data and relevant contextual information aforementioned were recorded to a file from which spike, burst and superburst frequencies were quantified offline We calculated and report t he mean ( S.E.M.) time after plating that action potentials emerged (in weeks after plating), mean spike rate per session adjusted for active electrodes (Hz) and mean percent active electrodes per session in NPC generated cultures, primary cultures and NPC/primary cell co cultures. Network bursts are sudden increases followed by sudden decreases in firing frequencies among neurons in bra in regions, organotypic cultures or cell cultures (Gross and Kowalski, 1999) We defined network bursts based upon inter spike intervals similarly to previous reports that defined network bursts as activity with 0.1 to 2s inter spike intervals recorded at > 3 or > 14 electrodes (Gross and Kowalski, 1999; Tabak and Latham, 2003) Activity that was reliably categorized as a network burst by visual inspection typically showed action potentials across > 10 electrodes with < 0.2s inter spike intervals Because t hese parameters fell within the conservative end of
41 previously reported definitions, we chose to employ them to define network bursts in our culture and co culture experiments. We calculated and report t he mean ( S.E.M.) time after plating that network bur sts emerged (in days and weeks) and burst rates (per minute and per session) in NPC generated cultures, primary cultures and NPC/primary cell co cultures. No evidence of contamination was noted in any of the MEA dishes during the extended culture period. Superbursts were defined as a cluster of >10 consecutive network bursts (> 30 bursts/min) separated from other clusters by a quiescent period >10x longer than the mean inter burst interval within the cluster. These parameters were chosen in reference to pr evious work that defined a superburst as at least 50% of all large and medium bursts occurring inside tight clusters with inter superburst intervals > than 10x longer than inter burst intervals (Wagenaar et al., 2006) We confirmed these parameters by visual inspection of raster plots in which the smallest distinguishable superburst met these criteria. We calculated and report the mean ( S.E.M.) number of superbursts detected per recording session, the percent network bursts occurring within a superburst as well as burst rates detected between and within superbursts in NPC generated cultures, primary cultures and co cultures. Statistical Analy se s All group statistical tests were conducted using Statplus Software (Ana lysoft Incorporated; Vancouver, BC, Canada). To determine whether adult NPC generated cultures develop network wide activity, we compared the dependent variables (percent active electrodes, spike rates (Hz), burst rates, superburst rates, percent activity in superbursts number of superbursts and spike rate s) across weeks 1
42 7 after plating using one way ANOVAs Changes in viability (cell densities) and c ell phenotypes ( percent immature neuron, mature neuron, astrocyte, or oligodendrocyte precursor and perce nt neurons expressing glutamatergic and GABAergic phenotypes) among NPC progeny were examined across time after plating ( 2, 4, and 6 w ee ks ) using one way ANOVAs To confirm that primary culture s exhibited mature bursting activity before NPC addition (and r andom group assignment) and to characterize changes in their bursting activity after co culture, we examined the dependent variables ( spike rate, burst rate, percent change in spike rate from baseline, percent change in burst rate from baseline, percent ac tive electrodes, percent activity in superbursts and number of superbursts ) between groups (fresh media control, NPC conditioned media control, 10 K NPCs added and 50 K NPCs added) across days after plating (3 39) or w ee ks after NPC addition (weeks 1 3) using two way ANOVAs To phenotype primary culture d cells and co cultured NPC progeny we examined cell density and cell phenotype ( percent dividing cells and percent immature neuron s transitioning neurons, mature neuron s astrocyte s and oligodendrocyte precursor s or percent neurons expressin g glutamatergic or GABAergic phenotypes) across time after the initiation of co culture (2, 4, and 6 weeks) using one way ANOVAs. Newman Keuls p ost hoc tests were employed to explore main and interaction effects and a ll levels were set at 0.05. Results Adult NPC Generated Neurons that Retain Immature Phenotypes are Incapable of Ensemble Activity Adult NPCs were plated on MEAs and observed for several weeks. Because activity across NPC generated cultures did not evolv e robustly across recording sessions within each week we analyzed the mean ( S.E.M.) percent of plate wide
43 active electrodes, electrode wide spike rates and plate wide synchronized bursts collected during the last se ssion of each week (Fig ure 2 1). Action p otentials emerged in the 2 nd week after plating ( Figures 2 1A B), presumably coincident with the emergence of active neurons. Activity at some electrodes persisted, disappeared or emerged latently ( Figures 2 1 C K ) suggesting asynchronous neuronal differentiati on and turnover among adult NPC generated neurons. The percent active electrodes (> 0.1 Hz) across MEA plates did increase with time ( F (6,28 ) = 5.42, p < 0.001; Figure 2 1A). Relative to the 1 st (zero activity) and 2 nd week after plating the percent active electrodes significantly increased in the 5 th week ( p 0.05) and peaked in the 6 th week ( p 0.05). However, the peak percent of active electrodes (~9%) was much lower than the peak percentage we had expected from primary neurons or pluripotent stem cell derived neurons ( Figure 2 1A vs. 2 3C, for example), suggesting that few functional NPC generat ed neurons were present. Spike rates on active electrodes remained consistently low across weeks (~1Hz; F (6,28 ) = 1.03; n.s. ; Figure 2 1B) suggesting that NPC generated neurons failed to integrate into a functional network. In fact, raster plots of action potentials recorded at each electrode over time confirmed that although adult NPC derived neurons fired action potentials, their activity never synchronized into bursts characteristic of networked neurons ( Figures 2 1C, F and I) MEA plated NPC derived cultures were imaged a fter each weekly recording session using phase contrast microscopy to monitor their location and density relative to each electrode Figure 2 1 E, H and K shows adult NPC generated cells at the same electrode locations 2, 4 and 6wks a fter plating D issociated NPCs were small round
44 and phase bright upon plating Many of these cells proliferated over the first 2 w eek s producing dense colonies of cells that had extended a process ( Figure 2 1 E ) Shriveled phase dark cells floating above the adherent cell layer, indicative of cell death, were observed at this time point During the 3 rd and 4 th w ee k, proliferation and death stabilized at a low level and cells migrating out of the colonies established during the proliferative phase appeared to increase the culture confluency ( Figure 2 1 H ). At this point, m ost cells possessed extensive branche s and either round or oval cell bodies typical of neurons or flattened irregular shaped cell bodies typical of astrocytes. Although low level proliferation may have persisted through later weeks, the increased culture confluency appeared to be mediated by more elaborate branching and further migration ( Figure 2 1K vs. H ). Cultures were fixed and stained at 2, 4 and 6 wks after plating and the proportion NPC derived cells that expressed immature neuroblast ( tubulin + ), differentiated neuron (NeuN + ), astrocyte (GFAP + ), and oligodendrocyte (NG2 + ) phenotypes was scored ( Figure 2 2). Oligodendrocytes were not detected at any time poin t. An ANOVA on phenotypes revealed that preferred phenotypes changed significantly across weeks ( F (4,18) = 6.00; p < 0.001; Figure 2 2). While the percent of NPC derived immature neurons increased across weeks (week 6 > weeks 2 and 4; p < 0.001), the prop ortion of astrocytes and mature neurons remained consistently low (week 2 = week 4 = week 6; n.s. ). By the 6 th week after plating, a significantly higher percent of adult NPC generated cells expressed immature neuronal phenotypes versus either astrocyte ( p < 0.001) or mature neuronal ( p < 0.001) phenotypes.
45 Figure 2 1. Neurons in adult hippocampal NPC derived cultures fail to develop network activity. The electrophysiological activity of adult hippocampal NPC generated cultures was monitored for 7wks (n = 5). (A) The mean ( S.E.M.) % active electrodes significantly increased after 5wks ( s < 0.05), but only ~8% of electrodes detected action potentials (compare Figures 2 3A C) (B) M ean ( S.E.M.) spike rates (Hz) on active electrodes did not increase across weeks. Raster plots from recordings at 2 (C), 4 (F), and 6 (I) weeks confirm action potentials do not synchronize Waveform maps captured at 2 (D), 4 (G), and 6 (J) weeks after plating s how persistent, transient an d latent activity at electrodes Phase contrast micrographs (20x objective) show extensive migration and branching among cells in the same MEA location following 2 (E), 4 (H) and 6 (K) weeks after plating. Arrows (C, F, I) and b oxes (D K) show activity and cells at the same electrode 2, 4, and 6wks after plating. p < 0.05 relative to plating.
46 Figure 2 2 Adult hippocampal NPCs produced neurons that retain immature phenotypes and few astrocytes. The proportion of GFP expressing a dult hippocampal NPC derived cells expressing immature ( tubulin) and mature (NeuN) neuronal markers or astrocyte (GFAP) and oligodendrocyte precursor (NG2) markers at 2, 4 and 6wks after plating was quantified on confocal laser scanning microscope digita l images (20x objective with 3x digital zoom; n = > 500 cells in 4 wells per staining condition). No NG2 + cells were detected. (A) While the % of adult NPC generated immature neurons increased across weeks (weeks 2 and 4 < week 6; s < 0.001), the % of as trocytes and mature neurons did not increase from week 2 (weeks 4 and 6 n.s. ). By the 6 th week after plating, more immature neurons were detected than astrocytes ( p < 0.001) or mature neurons ( p < 0.001), suggesting that adult NPCs generate primarily neuro ns that retain immature phenotypes. (B E) show confocal images of GFP expressing NPC progeny (in green) that co express the immature neuronal marker tubulin (in red) at 2wks (B) and 6wks after plating (D) or the astroglial marker GFAP (in red) at 2wks (C ) and 6wks (E) after plating. Scale bar 25 m. *** p <0.001 between phenotypes; p < 0.001 within wks.
47 P rimary Neural Cultures Develop Neural Ensemble Behavior Coincident with Neuronal Maturation Cells were isolated from E18 rat cortices and plated on MEAs. Activity within each primary culture was recorded for 10min every 3d after plating to confirm that neurons generated stable and mature synchronized network wide bursting prior to co culture with NPCs. As expected, the mean (S.E.M.) spike rates and burst rates generated by neurons in primary cultures increased across days after plating (spike rate: F (12,104 ) = 16.02; p < 0.001 and burst rate: F (12,104 ) = 41.04; p < 0.001 ; below ) In addition, the mean ( S.E.M.) percent of active electrodes withi n MEA dishes increased across days after plating until stabilizing on day 14 ( F (12,104 ) = 40.91; p < 0. 001; Figure 2 3C), consistent with the timelines reported by other groups (Gross and Kowalski, 1999; van Pelt et al., 2004; Wagenaar et al., 2006) Action potentials emerged within 3 5 days (mean S.E.M: 0.14 0.03Hz; Figure 2 3A) and network bursts emerged by the 2 nd w ee k ( Day s 8 14) after plating ( 0.98 0.07bursts/min; Figure 2 3B) when increases in percent active electrodes within each MEA dish had alread y plateaued ( Figure 2 3C). During the 3 rd w ee k ( D ays 15 21) after plating, spike frequency increased to 1. 03 0. 18 Hz; (Day 20 vs. Day s 3 and 5; p values < 0.05 ). Burst rates also increased to 3.13 0.63/min (Day 20 vs. Day s 3 17 p values < 0.04). At the beginning of the 4 th w ee k after plating (~ Day 26), the cultures began generating prolonged clusters of bursts separated by several minutes of quiescence Figure 2 4 B). During this transient superburst period (~ Day s 26 33) we observ ed peak spike rates (2. 65 0. 15 Hz; day s 26 36 > day s 20 and 23; p values < 0.05), network burst rates (13.79 0.80/min; days 26 33 > day 23 > day 20; p
48 values < 0.004) across the majority of active electrodes (mean S.E.M: 80 1 5 94%; Figure 2 3 and 2 4 ). N etwork bursting plateaued during the 5 th w ee k after plating ( day s 29 35) with mean spike rates of 2.80 0.19 Hz and burst rates of 14.14 0.73/min ( Figure 2 3) We added NPCs, NPC conditioned media or fresh media to MEAs on day 39 and resumed recordings. To better understand how cell composition promotes the evolution of bursting, we fixed and stained primary cells in parallel with the development of ensemble activity (2 weeks), superburst patterns (4 weeks) and network stabilization (6 weeks) in their MEA plated counterparts ( Figure 2 5). The percent of differentiated cells detected across weeks was consistently high (~81 85%; Figure 2 5A; F (2, 30 ) = 0.19 ; n.s .). While the percent of glial phenotypes detected remained consistently high across weeks, decreased percentages of immature (week 2 > weeks 4 and 6; p values < 0.05) and transitioning (weeks 2 and 4 > week 6; p values < 0.001) neurons but increased percentages of mature neurons were detected (weeks 2 and 4 < week 6; p = 0.08 and 0.02, respecti vely; effect of phenotype: F (4,30) = 138.6 4 ; p < 0.001 ; phenotype by week interaction effect: F ( 8 30 ) = 4.44 ; p < 0.001; Figur e 2 5A D). The majority of 4wk old neurons expressed vesicular glutamate and a small proportion expressed vesicular GABA ( Figure 2 5E F ) suggesting the presence of abundant functional synapses among these networked primary neurons
49 Figure 2 3. Primary cultures develop stable network bursting within weeks of plating. (A) Action potentials appeared within 3d of plating. Mean ( S.E.M) spike rates (Hz) increased in the 3 rd wk (20d vs. 3 5 d ; s < 0.05) and then peaked in the 4 th wk when superbursts were ob served (26 36d > 20 23 d ; s < 0.05) before plateauing ( 26 39 d ; n.s .). (B) Network bursts were first detected in the 2 nd wk after plat ing and increased in the 3 rd wk ( 20 d vs. 3 17 d ; s < 0.05) peaking in the 4 th wk when superbu rsts were detected ( 26 33 d > 23 d > 20 d ; s < 0.005) before plateauing ( 26 39 d n.s. ). Inset raster plot showing action potentials detected simultaneously on most electrodes (a b urst), twice within 5s. (C) M ean ( S.E.M) % active electrodes incr eased rapidly across days ( 14 39 d > 11, 8 d > 5 d > 3 d ; s < 0.05) before plateauing on Day 14.
50 Figure 2 4. NPCs and/or their progeny revert neuronal ensemble activity to an immature status. (A) A fluorescent image showing a GFP + NPC derived (in green) Tubulin + (in red) neuron among MEA plated primary cells. The co culture was fixed and immunostained to show primary (DAPI + ) and NPC derived ( DAPI/GFP + ) astrocytes (GFAP + ; in blue) and neurons ( Tubulin + ; in red) 50 days after NPC addition. Scale bar 100 m. (B) Representative raster plots
51 showing action potentials detected in primary cultures by electrodes 1 59 (y axis) during a 600s session (x axis). Superbursts were detected in all primary cultures during the 4 th week after plating (Days 23 33; First Row) before permanent stable bursting activity emerged in the 6 th week (Days 35 39; Second Row). Superbursting re emerges in primary cultures 1wk after 50K NPCs (and 2wks after 10K NPCs) are added and persists for at least 3wks (Days 41 61), whereas fresh media control cultures continue exhibiting stable burst patterns (Third Row and C). Inset a 10s zoom showing that superbursts are composed of high frequency bursts. The addition of 50K NPCs also induced persistent superbursting activity in 64d old fresh media controls (Bottom Row). (C) The mean (+S.E.M.) number of superbursts detected in primary cultures increases the 4 th week and then subsides in the 5 th week after plating as stable bursting activity emerges in the 6 th week. Superburst frequency in mature primary cultures increased significantly in the 2 nd and 3 rd week after 50K NPCs were added ( p < 0.0001) and in the 3 rd week after 10K NPCs were added ( p < 0.05), but not at any time point after NPC conditioned media was added, relative to fresh media controls (n = 3 cultures per group). Superburst frequencies were similar in immature (4wk old) cultures and following NPC addition. (D) Shows th e mean (S.E.M.) % of activity in superbursts recorded after NPC or media addition. Across sessions, most activity consisted of superbursts beginning in the 2 nd week after the addition of 50K NPCs ( p < 0.05) and the 3 rd week after the addition of 10K NPC s ( p < 0.05), relative to fresh media. After emergence, superbursting activity persisted for the duration of the experiment in the 50K NPC added group ( p < 0.01) and tended to persist in the 10K NPC added group ( p = 0.07). We found that inter superbu rst and intra superburst burst rates (B inset) were similar to those detected in primary cultures as they transitioned to a mature state (~4wks after plating), suggesting that NPCs induce the re emergence of superbursting activity, normally produced by neurons as they mature. We examined inter superburst (E) and intra superburst burst rates (B inset and F) to determine whether the strength of bursting and superbursting was NPC density dependent. Inter superburst rates sign ificantly decreased in the 2 nd and 3 rd weeks after 50K NPCs ( p values < 0.05), but not in the 10K NPC supplemented or NPC conditioned media groups, relative to fresh media controls. This effect is likely because superburst frequency was also significantly elevated in the 50K NPC group (C). (F) Within superbursts, mean (S.E.M.) burst rate frequencies were similar in primary cultures as they transitioned to a mature state (~4wks after plating) and following NPC addition. These data suggest that NPCs induce the re emergence of superbursting activity, typically produced by neurons as they mature. p < 0.05, ** p < 0.01, *** p < 0.001, relative to fresh media controls.
52 Figure 2 5. Dissociated primary cultures produce robust numbers of neurons, astrocytes and oligodendrocytes. (A) The % DAPI + primary cells expressing markers of immature neurons ( Tubulin), transitioning neurons ( Tubulin and NeuN), mature neurons (NeuN), astrocytes (GFAP) and oligodendrocytes (NG2) were calculated at 2, 4 and 6wks after plati ng. While the % of astrocytes and oligodendrocyte precursors remained consistent across wks ( p n.s. ), the percentage of immature (wk 2 > wks 4 and 6; p < 0.05) and transitioning (wks 2 and 4 > wk 6; p < 0.001) neurons decreased while the % of mature neurons increased (wks 2 and 4 < wk 6; p = 0.08 and 0.02, respectively). (B F) Representative confocal images of DAPI + (in gray) neurons, astrocytes and oligodendrocyte precursors detected in 6 week old primary cultures (B D). B shows immature ( Tubuli n; red) and mature (NeuN; green) neurons. Note that NeuN and Tubulin is co expressed in some cells. Many neurons in 4 wk old cultures expressed (E) vesicular GABA (vGAT) or (F) vesicular glutamate (vGlut), suggesting the presence of functional excitato ry and inhibitory neurons at the same time point that superbursting activity emerges. Scale bar 50 m. p < 0.05, ** p < 0.01 and p = 0.08 relative to week 2.
53 NPCs Evoke Re E mergent Superbursting In Pre Established Mature Neural Networks We recorded electrophysiological activity for 10 minutes every 48 hours for an additional 50 days to determine the effects of adding undifferentiated adult NPCs to these mature neural networks. There was no statistically significant effect of NPC addition on spike rate (percent change from Day 39 baseline; effect of group: F (3,96 ) = 0.54, n.s. ) at any point across days after the initiation of co culture (effect of day: F ( 11 ,96) = 1.41 n.s. and interaction effect: F ( 33,143 ) = 0.64 n.s. ). There was, howeve r, an effect of NPC addition on burst rate (percent change from Day 39 baseline; effect of group: F (3,96) = 9.67, p < 0.001) that persisted across days after the initiation of co culture ( effect of day: F ( 11 ,96) = 0.6 5 n.s. ; interaction effect: F ( 33 143 ) = 0.24 n.s. ) Relative to control MEAs maintained in fresh media (+1.9 5.0%), burst rate was unaffected by the addition of NPC conditioned media ( 7.8 2.3%; n.s. ), but significantly increased following the addition of 10K NPCs (70.2 1.79%; p < 0 .001) and tended to increase following the addition of 50K NPCs (37.8 8.1%; p = 0.07). Raster plots showing recorded action potentials across electrodes revealed that neurons in adult NPC supplemented primary cultures, but not media supplemented primary cultures, were generating superbursts reminiscent of those generated transiently by developing primary cultures ( Figure 2 4B). An ANOVA revealed that superburst frequency was elevated by NPC addition ( F (3, 96 ) = 18.91; p < 0.001), but differentially across w eeks after addition (effect of week: F (2,96 ) = 10.97; p < 0.001; week by group interaction effect: F (6,107 ) = 2.90; p < 0.05). Specifically, the frequency of network wide superbursts increased in cultures supplemented with 10K NPCs ( p < 0.05) or 50K NPCs ( p < 0.001), but did not change among cultures supplemented with NPC conditioned media ( n .s. ), relative to fresh
54 media. The onset of re emergent superbursts among electrophysiologically mature primary cultures was NPC number dependent. Relative to neurons i n the fresh media control group, neurons in the 50K NPC supplemented group generated significantly more superbursts in week 2 ( p < 0.001) and week 3 ( p < 0.001), whereas neurons in the 10K NPC supplemented group only generated increased numbers of superbursts in week 3 ( p < 0.05; Figure 2 4C). Once superbursting activity was induced by NPC addition (effect of day: F (6,56) = 8.80; p < 0.001), the majority of electrophysiological events recorded during a session consisted of superbursts (effect of grou p: F (3,56) = 20.80; p < 0.001; Figure 2 4D). While the majority of electophysiological events recorded among media control cultures consisted of bursts (with few if any superbursts detected), about 50% of the events were superbursts ~ 1 week after 50K NPCs were added ( p values < 0.05 ) and ~2 weeks after 10K NPCs were added < 0.05) The majority of events were superbursts by the 2 nd week after 50K NPCs were added ( > 75%) and by the 3 rd week after 10K NPCs were added ( > 50%). Superburst rates, as well as inter and intra superburst burst rates among NPC supplemented cultures did not differ from those measured in developing primary cultures. Co Culture with Electrophysiologically Active Primary Neural Cultures Promotes Differentiation and Maturation of NPC derived Neurons The changes in neuronal ensemble behavior paralleled changes in the composition of GFP expressing NPC derived cells in parallel co cultures that were stained immunocytochemically every 2 weeks after co culture ( Figure 2 6). As with the cultur es generated by adult NPCs plated alone, we did not detect NPC derived oligodendrocyte precursors in co cultures and therefore excluded NG2 + phenotypes from our analysis. Relative to NPCs cultured in isolation, co culture with primary
55 neurons promoted an i ncrease in NPC generated cells that expressed the immature neuronal marker tubulin versus either the astrocyte marker GFAP ( p < 0.05) or the mature neuronal marker NeuN ( p < 0.01, with overall significance of co culture on phenotype by ANOVA, F (2,63) = 6.60; p < 0.01). The phenotypes among co cultured NPC progeny changed significantly across weeks after co culture (interaction between week and phenotype: F (4,71) = 9.02; p < 0.001). At 2 weeks, immature neuronal phenotypes were observed more frequently than either astrocyte phenotypes ( p < 0.001) or mature neuronal phenotypes ( p < 0.001) and at 4 weeks, similar proportions of immature neurons and astrocytes were detected ( n.s .) but more astrocytes were observed than mature neurons ( p < 0.05). The percen t of NPC generated cells expressing astrocyte phenotypes peaked at 4 weeks (week 2 < week 4 and week 6; p < 0.05). By 6 weeks, all phenotypes were observed among NPC generated cells with the same frequency. Unlike the NPCs plated alone, many co cultured NPCs generated neurons that transitioned from an immature (week 2 > weeks 4 and 6; p values < 0.001) to a mature state (week 2 < week 6, p < 0.05 and week 2 < week 4, p = 0.09). We also note that some NPC derived neurons died or lost their phenotype betwe en 2 and 4 weeks after co culture because the percent of total neurons declined significantly (~79% at week 2 versus ~40% at weeks 4 and 6). The majority of immature neurons expressed vesicular glutamate and this decreased with time coincident with an incr ease in the expression of vesicular GABA ( Figure 2 6).
56 Figure 2 6. Adult NPC progeny mature more quickly and robustly when co culture d with electrophysiologically mature oligodendrocyte and astrocyte rich primary cultures. The proportion of adult hippo campal NPC derived (GFP + ) cells expressing tubulin, NeuN, GFAP or NG2 at 2, 4 and 6wks in co culture with mature primary neural networks was quantified (A) Most NPC derived cells were immature neurons after 2wks in co culture ( p < 0.001 versus astrocyt es and mature neurons), astrocytes after 4 weeks in co culture ( p < 0.05 versus mature neurons) and all phenotypes were observed with the same frequency by 6wks in co culture ( n .s. ). Co cultured adult NPCs generated astrocytes latently (week 2 < week 4 an d week 6; p produced neurons capable of transitioning from an immature ( tubulin; week 2 > weeks 4 and 6; p week 2; p 2 (~79%) to 4 and 6(~40%), suggesting an early vulnerability to cell death. The majority of immature neurons that expressed vesicular glutamate decreased with time coincident with an increase in the expression of vesicular GABA. Confocal laser scanning digit al images show adult NPC derived (DAPI/GFP + ) cells co labeled with the immature neuronal marker tubulin (B) and the astrocyte marker GFAP (C) S cale bar 50 m. p < 0.05, *** p < 0.00 1 versus 2wks in co culture.
57 Discussion Neurons in primary neocortic al cultures synchronize their activity into network wide bursts (Pine, 1980; Gross and Kowalski, 1999; van Pelt et al., 2004; Wagenaar et al., 2006) This activity transitions through a superbursting phase before stabilizing into a mature state that has be en described as analogous to the superbursting activity that is associated with network refinement in the developing mammalian brain (de Lima and Voigt, 1997; Voigt et al., 2001; Opitz et al., 2002) Several prominent differences were observed between prim ary neural networks and NPC generated neural networks. Primary neural cultures contained large numbers of astrocytes and oligodendrocytes whereas NPC generated cultures only contained a few astrocytes and no oligodendrocytes ( Figure 2 2 vs. Figure 2 6). In add ition, many neurons in primary cultures transitioned from an immature to a mature state, whereas NPC alone generated cultures contained primarily immature neurons at equivalent times. Our finding that adult NPC derived neurons, at least grown in standard c ulture conditions, contain few astrocytes and fail to develop synchronized bursting activity complements the body of work showing that mature glia are required for efficient synaptogenesis and the maturation of network attributes (Banker, 1980; Eroglu and Barres; Kucukdereli et al.) The addition of NPCs to a pre established mature network evoked a surprising re emergence of superbursting activity. Superbursts (~0.2s; Figure 2 5) detected in primary cultures as they mature electrophysiologically are shorter than the continuous ~4s long bursts ( ~3 4Hz ) recorded in seizure models (Jefferys and Haas, 1982) and more accurately resemble network bursts detected across brain regions developmentally (Galli and Maffei, 1988; Ben Ari et al., 1989; O'Donovan, 1989;
58 Pein ado, 2000) Superbursting activity is mediated by depolarizing currents at GABA A receptor s on the maturing neurons that activate voltage gated Ca 2+ channels which potentiate the activity of the NMDA receptors (Ben Ari, 2001) Slow Ca 2+ wave (0.1 0.5Hz) network formation (Khazipov et al., 2001; Dupont et al., 2006) and long term potentiation (Lledo et al., 1995; Berridge, 1998). We postulate that because transi ent superbursts among primary cultures are Ca 2+ wave mediated (Opitz et al., 2002; Pasquale et al., 2008) and indistinguishable from re emergent superbursts in adult NPC supplemented cultures, that adult NPCs may induce plasticity throughout mature neurona l networks. The ability to establish synchronous bursting suggests that NPCs trigger a remodeling of connectivity and/or synaptic attributes within established networks, potentially by secreting factors that act on networked neurons or by generating de nov o connections with neurons within the pre formed network. Of course, dual patch clamp experiments would be required to definitively show that NPC generated neurons have integrated functionally into pre established networks but the inability of conditioned medium to evoke bursting activity in the current study is more consistent with the latter hypothesis and we speculate young NPC derived neurons may mediate these effects. Identifying whether similar mechanisms underlie superbursting activity and NPC induce d re emergent superbursting activity may clarify the functional and clinical importance of this phenomenon. Astrocytes present in the primary electrophysiologically mature cultures are likely to promote NPC derived neuron connectivity in the co culture context. Our primary cultures (~46% astrocytes/~8% oligodendrocyte precursors; 2 weeks) increased adult
59 NPC generated astrocyte and mature neuron yields ( Figure 2 2 versus 2 6). These data, and reports that adult NPCs generate functional neurons more efficie ntly when co cultured either with astrocytes and brain derived growth factor (BDNF) (Song et al., 2002a; Song et al., 2002b) or in co culture with BDNF and LTP producing neurons (Babu et al., 2009) suggest that they could generate networking neurons in t he appropriate niche. In vivo multicellular recordings have shown that functional processes, such as motor learning and working memory, are mediated by neurons that fire with spatiotemporal coherence (Buzski, 2010) These neuronal networks increase in s ize and interconnectivity with the complexity of the functional process (Pessoa, 2008) C ell replacement strategies may benefit by employing cell s that spontaneously generate neurons with network modifying attributes. ESCs are already attractive candidates because their neuronal progeny quickly network in culture, within grafts and with graft host neurons (Ban et al., 2007; Illes et al., 2007; Heikkila et al., 2009; Tonnesen et al., 2011) In cont rast, adult NPCs possess a more restricted application in replacement or augmentation strategies because their neuronal progeny do not establish networks in isolation ( Figure 2 1) but may beneficially induce renewed plasticity in pre existing networks. Ulti mately, the use of cells in therapy will require consideration of both intrinsic properties of the graft and properties of the target niche. We have shown here that cells and niche reciprocally influence each other. Embryonic stem cells are able to generat e many neuronal subtypes while tissue derived NPC progeny appear restricted to more limited subsets of GABAergic or glutamatergic neurons (Peljto and Wichterle; Svendsen
60 et al., 1998; Dziewczapolski et al., 2003). These differences may be in part, due to a loss of potential as neural progenitors become more specialized over time. For example, ESC derived NPCs that are propagated in monolayer cultures gradually loose networking capability (Illes et al., 2009) similar to fetal tissue harvested NPCs that are propagated in vitro (Mistry et al., 2002; O'Shaughnessy et al., 2009; Lepski et al., 2011). I n vivo adult NPCs spontaneously generate neurons that integrate into existing neural circuits (van Praag et al., 1999; Nissant and Pallotto, 2011), but whether the y can or should generate de novo networks in the absence of a pre existing network has never been explored. Transplanted ESC and tissue derived NPCs only generate neurons in neurogenic regions of the adult brain, but abortive neurogenesis can be stimulate d in extra neurogenic regions by injury or focal cell death (Hoehn et al., 2005) ESC derived neuronal grafts may bypass niche effects because they contain neurons that are intrinsically capable of establishing networks with one other and host neurons eve n in extra neurogenic regions (Tonnesen et al., 2011) In the present study, our use of MEAs highlight their utility in defining both intrinsic properties of stem cell derived networks and in assessing how a stem cell derived population may alter characte ristics of an existing network. The persistence of bursting activity produced by mature MEA plated primary neurons illustrates the stability of network properties established on MEA and the ability to quantitatively measure dynamic effects of maturing stem cell derivatives on neural activity. The MEA system such as hormones and neuroinflammatory molecules (Tanapat et al., 1999; Monje et al., 2003; Ormerod and Galea, 2003; Ekdahl et al., 2009) and parameters hypothesized to
61 affect transplant efficacy, such as cell number (Winkler et al., 2005) For example, we detected superbursting activity in primary cultures 1 week after 50,000 NPCs and 2 weeks after 10,000 NPCs were add ed ( Figure 2 5). Cellular injury can also be modeled using MEAs. For example, we are currently exploring how adult NPCs modulate the effects of oxygen glucose deprivation (OGD) on network activity and have early indications that NPC addition restores superb ursting activity in mature networks following OGD. Our present data in combination with earlier work suggests that stem cell strategies could be tailored and even combined to improve the outcome of specific CNS injuries or syndromes. For example, large in juries from stroke may benefit from replacing neuronal networks lost in the injury core and awakening silenced networks in the injury penumbra (Fisher, 1997; Hicks et al., 2009; Shimada and Spees, 2011) One might speculate that ESC derived neurons are att ractive candidates for generating de novo core circuitry because they spontaneously generate neurons with networking capability (Ban et al., 2007; Illes et al., 2007; Heikkila et al., 2009) while adult NPCs may be attractive candidates for stimulating pers istent plasticity among silenced networks, such as those found in the stroke penumbra.
62 CHAPTER 3 MICROELECTRODE ARRAY TECHNOLOGY REVEALS T HAT VARIABLE DURATIO N OXYGEN GLUCOSE DEPRIVATION MAY MODEL DIFFERENT ASPECTS OF STROKE Introduction Ischemic stroke is a leading cause of death worldwide and stroke survivors face compromised quality of life due to loss of function and its associated decrease in occupational utility (for review Dhamoon et al. 2010 ) Both in vitro and in vivo models of hypoxia ischemia have revealed mechanisms associated with stroke induced cell death (f or example Szatkowski et al., 1990; Madl and Burgesser, 1993; Calabresi et al., 1999) Fewer studies have identified mechanisms that affect the quiescent neu ral circuit s that survive ischemic stroke but likewise lose their function (Zhu et al., 2011) Although in vitro models cannot fully recapitulate the exquisite cytoarchitecture of the brain, simplified biologically relevant circuits exposed to oxygen gluc ose deprivation (OGD) may provide insight into the mechanisms that damage neurons in vivo because they arguably produce similar stroke pathology to those observed in situ (Harms et al., 2001; Meloni et al., 2001) Multi electrode array (MEA) technology per mits the simultaneous collection of optical and electrophysiological data within a culture for months (Potter and DeMarse, 2001) and t he development of activity in cultures established from a number of brain regions has been well characterized. For exampl e, fetal rat dissociated cortical cultures develop spontaneous action potentials and network bursts that evolve predictably over several weeks to reach a mature, functionally stable state by 35 40 days (Kamioka et al., 1996; van Pelt et al., 2004; Stephens et al., 2012) Since the neural activity can be tightly controlled, monitored for long periods, and is highly sensitive to pharmacological manipulations, MEAs offer a stable platform for
63 performing in depth analyses of how hypoxia/ischemia and subsequent therapies influence neural networks. These cultures are attractive options for inexpensive high throughput assays for screening new pharmacological or cellular engraftment strategies because they are highly sensitive, non invasive, reduce the need for anim al experiments, and provide less restricted access for electrophysiological and immuno cy tochemical data collection. To our knowledge, MEAs have only been employed to monitor the effects of hypoxia/ischemia on cardiomyocytes (Krinke et al., 2009; Yeung et a l., 2009) The behavior of neuronal networks during OGD has not been examined using MEAs, but can give an unprecedented view of the network degradation of spontaneous activity from 59 sites simultaneously that has only previously been recorded from a handf ul of electrodes placed in vivo (Kass and Lipton, 1982; Rosner et al., 1986) Here we combine OGD with MEA technology and immuno cy tochemisty to measure the breakdown of the activity in an established neural network during hypoxic ischemic conditions. We e mploy different durations of OGD to identify exposure levels corresponding with moderate to significant levels of cell death, and observe the immediate and delayed effects of variable durations of OGD on neuronal activity. Experiments are mirrored on multi w e ll chamber LabTeks to for specific cell phenotypes in combination with cell death and distress markers in order to investigate how and by influencing what cell populations could treatment options ameliorate some of the damage in the aftermath stroke.
64 Material and Methods Generation of Primary Cultures Embryonic day 18 (E18) sprague dawley rat neocortices (n=2) were purchased and shipped from Brain Bits (Springfield, Illinois) in Hibernate B solution and stored overnight at 4C upon arrival. The follow ing day, they were dissociated in a solution of papain and DNase I (20 units/ml and 2000 units/ml, respectively; Worthington Biochemical; Lakewood, NJ) l) in GlutaMax supplemented DMEM (Invitrogen) at 37C for 30min. The cells were then triturated ~10x gen tly, pelleted for 3min (1500rpm) and resuspended in DMEM/ GlutaMax (Gibco; Grand Island, NY) containi ng 10% equine serum (Hyclone ; Logan, UT ) and 1% Cellgro Antibiotic Antimycotic (Mediatech, Inc.; Herndon, VA). A 100l droplet of media containing ~ 25K ce lls directly onto a laminin droplet placed o nto the center of each polyethyleneimine (PEI) laminin coated MEA (n = 1 8 ) and LabTek (n = 1 6 ; > 2 wells per staining set), producing a cell density of ~4,000 cells/mm 2 known to generate robust network wide electr ophysiological activity (for example, Wagenaar et al., 1999). Every 3d (> 24h prior to a recording session), 50% of the media was replaced with fresh media. MEA and Multiwell Chamber LabTek Preparation Electrophysiological data was recorded from cells plat ed on MEAs consisting of a Multichannel Systems, Reutlingen, Germany) and immunocytochemical data was collected from cells plated on 8 well LabTek chambers. The cultures plated on MEAs and in LabTeks wells were generated from the same pool of dissociated cells for each Aldrich ; St. Louis, MO ) was placed on the center of each autoclaved MEA dish or sterile La bTek
65 of mouse laminin (Sigma Aldrich; 1mg/ml in PBS) was placed onto the center of their surfaces for 1h at 37 C. Once cells adhered to the laminin substrate (~10min) med ia Oxygen G lucose Deprivation We first examined the effects of variable duration OGD on the death of subtypes of cells in a subset of LabTek plated cultures. Once neurons in MEA plated primary cul tures exhibited the synchronized bursting activity typical of electrophysiologically mature neural networks (day 40) LabTek and MEA plated cultures were assigned to either the OGD control group (100% fresh media change ) or the 60 min 120 min 180 min of O GD (n = 3 MEAs and n = 6 LabTeks per group) Activity was recorded from 3 OGD exposed MEA plated cultures during the exposure. First, we simultaneously measured the dissolved oxygen (DO) content in cell media and the oxygen content in the air inside of th e incubator for 180min in order to measure the rate at which oxygen leaves the media ( Figure 1E; n=3). An ANOVA showed a change in DO overtime ( F (19,62) = 47.0; p < 0.001), and post hoc analyses showed DO began to drop from baseline after 30min ( p = 0.062), and significantly dropped after 60min (all p < 0.001). Therefore, we decided to keep the cell media in the hypoxic incubator overnight before use in the OGD experiments. Pipetting the media into the cultures in preparation for OGD would indeed re oxygenate the media, but studies have shown that oxygen uptake by cells exceeds the oxygen diffusion coefficient in a cell density dependent manner, and therefore we believe that the cellular environment rapidly becomes hypoxic (Bambrick et al.)
66 All media was rep laced on MEA and LabTeK plated cultures assigned randomly to OGD groups with DMEM devoid of glucose and sodium pyruvate and containing low L glutamine (GIBCO cat. no. 11966) that had been incubated in hypoxic conditions (37C, 0.1% O 2 5.0% CO 2 94.1% N 2 ) overnight. These hypoxic DMEM fed cultures were placed immediately into a hypoxic incubator (37C, 0.1% O 2 5.0% CO 2 94.1% N 2 ) for 60, 120 or 180min. Immediately after OGD exposure, the hypoxic DMEM was replaced with fresh media and the cultures were ret urned to a normoxic incubator (37C, 20.95% O 2 5.0% CO 2 ). All media on MEA and LabTek plated cultures assigned randomly to the control group was completely replaced with fresh media just before and after the experimental groups were exposed to OGD. Immunocytochemistry and M icroscopy LabTek plated cells were stained with PI (250mg/ml TBS; Sigma Aldrich) for 15 min and then fixed with 4% paraformaldehyde for 10min and rinsed repeatedly between immunostaining steps with tris buffered saline (TBS; pH 7. 4). The cells were blocked in a solution of 3% normal donkey serum (NDS) and 0.1% triton x in TBS and then incubated overnight in a cocktail of 3 primary antibodies that could include: rabbit or mouse anti Tubulin (1:500; Covance), mouse anti neuronal nu clei (NeuN; 1:500; Chemicon International), guinea pig anti glial fibrillary acidic protein (GFAP; 1:750; Advanced Immunochemical), rabbit anti NG2 (1:1000; Chemicon International), and Cu/Zn SOD (1:500; Millipore) at 4C. The following day, the sections w ere incubated in the appropriate FITC Cy3 or Cy5 conjugated maximally adsorbed secondary antibody (1:500; Jackson ImmunoResearch) for 4hrs. Subsets of cells immunostained with mouse anti NeuN or rabbit anti tubulin were also stained with rabbit anti rat
67 vesicular GABA A transporter (vGAT; 1:500; Chemicon International), and/or guinea pig anti vesicular glutamate transporter 1 (vGlut1; 1:5000; Chemicon International) and then the appropriate secondaries. All immunostained cells were DAPI stained (1:20,0 00 in TBS for 20min; Calbiochem) before being coverslipped under PVA DABCO. lmages of immunostained LabTek plated cultures were taken using a Zeiss LSM 710 fully spectral Laser Scanning Confocal Microscope with a 63x oil objective (5 z stacks). Visible la ser line intensities were maintained below 15%. DAPI + cells with neuron, astrocyte and oligodendrocyte morphologies and intense staining with the appropriate antibodies in primary cultures served as positive controls to which exposure levels and gain were set to exclude nonspecific staining. The percent of PI + DAPI labeled nuclei were quantified as a measure of cell death. Densities of mature and immature neurons, astrocytes, oligodendrocyte precursors, as well as percent PI + and percent Cu/Zn SOD + cells of each phenotype was quantified to detect any cell type particularly vulnerable to OGD. Immature and mature neurons surround by vesicular glutamate and GABA were measured to detect neuronal responses to OGD. The sum of five non overlapping images taken diag onally from each corne r of >2 wells (>500 cells/well) was averaged to find the mean cell density, or % of each cell type, or total DAPI labeled cells. Electrophysiology As described previously (Stephens et al., 201 2 ), changes in extracellular potential were recorded over 10min sessions every 3d from 3 43 d after plating. Cultures were equilibrated for ~2min after being placed into the amplifier before recording was initiated. Their activity was sampled at 25 kHz and signal was amplified and digitized usin g Multichannel Systems (Reutlingen, Germany) MEA1060 amplifier and A/D
68 hardware. Spike events were extracted and stored online for further offline processing using a software package custom written by Dr. Thomas B. DeMarse. In the subset of MEA plated cel ls that we re recorded during OGD exposure, recordings began immediately after the hypoxic incubator door was closed and continued for the 180min exposure to OGD. At the starting point of each recording, the acquisition software estimated the quiescent nois e signal Action potentials, or spikes, were detected online as signal deviation s beyond five times the median threshold from the quiescent signal. Electrodes exhibiting spike frequencies based on previous reports of minimum firing rates (Latham et al., 2000; Boehler et al., 2007) Waveforms between 1ms before and 2ms after detected spikes were saved and used to remove duplicate detections of multiphasic spikes and visually inspect for artifacts. N etwork bursts were defined as spike detected at > 10 electrodes within 0.2sec inter spike intervals. All data and relevant contextual information aforementioned were recorded to a file from which spike and burst frequency, spike amplitude, percent active el ectrodes and burst duration were quantified offline. The percent of spikes in interburst intervals during minutes 1 4 of OGD was compared to baseline by taking samples 5 interburst intervals from each of the 3 MEAs and comparing using a paired t est. The mean and standard deviation of the spike rates, burst rates and burst durations for each minute of the baseline recordings were calculated, and used to obtain the normal score of spike rates, burst rates and burst durations of each minute during O GD.
69 Statistical Analysis Statistical tests were conducted using Statistica Software (StatSoft, Inc; Tulsa, OK). We examined c hanges in viability (PI labeled cells) as percent of total DAPI labeled nuclei across OGD exposure times (0, 60, 120, or 180 min) u sing a one way ANOVA. Percent dead cells (PI) and oxidative stress ( percent SOD labeled cells) that were co labeled with cell phenotypes (immature or mature neuron, astrocyte and oligodendrocyte) were examined between control and OGD with two way ANOVAs. A lso, cell density between each phenotype and control or OGD, and % vGlut or vGAT between tubulin + or NeuN + neurons and control or OGD were examined with two way ANOVAs. For the MEA analysis, the dependent variables (spike rate, burst rate, spike amplitude, percent active electrodes, and burst duration) across days after plating were compared using repeated measures ANOVAs. Changes in spike rate, spike amplitude, burst rate, percent active ele ctrodes, and burst duration during OGD were determined using repeated measures ANOVAs across minutes of OGD exposure. The % spikes during interburst intervals during minutes 1 4 of OGD was compared to baseline using a test. The baseline spike rate at each electrode was correlated with spike rate during minutes 1 the percent of baseline spike rate, and percent active electrodes, between OGD exposure times (20, 60 or 180 min) and t ime of recording (1h or 3d after exposure) were compared using two way ANOVAs. Newman Keuls post hoc tests were employed unless otherwise noted,
70 Results Mature N eurons are P articularly V ulnerable to the E ff ects of L ong D uration OGD We first examined total cell death in fresh media control cultures versus cultures exposed to OGD for 60, 120 or 180 min ( Figure 1A B). An ANOVA on the mean ( S.E.M.) percent of DAPI + nuclei co labeled with the cell death marker PI revealed significant effects of OGD exposure duration on cell death ( F (3,23) = 31.51; p < 0.001). Significantly more dead cells were found in cultures exposed to OGD for either 120 minutes ( p < 0.001) or 180 minutes ( p < 0.001; Figure 3 1A), relative to cont rol cultures. We further characterized the effects of a 180 min OGD exposure on cell death and oxidative stress because this duration of OGD exposure killed ~50% of the culture, which often serves as a benchmark for exploring the mechanisms of stroke induc ed cell death (Meloni et al., 2001; Hirt et al., 2004) We next quantified the percent of NeuN + mature neurons, tubulin + immature neurons, GFAP + astrocytes and NG2 + oligodendrocyte precursors co labeled with the cell death marker PI in OGD exposed and co ntrol cultures ( Figure 3 1C). An ANOVA revealed significant effects of OGD exposure ( F (1,24) = 191.12; p < 0.001), dying cell subtype ( F (3,24) = 5.12; p < 0.01) and a significant OGD exposure by dying cell subtype interaction ( F (3,31) = 5.88; p < 0.01) on PI + cells. Cell death among all cell subtypes was similarly low in control cultures. OGD exposure increased the death of all cell subtypes ( p < 0.001), but more OGD induced death was detected among NeuN + mature neurons than all other cell types ( all p values < 0.05) and more dead NG2 + oligodendrocyte precursors were detected than GFAP + astrocytes ( p < 0.05).
71 PI binds to the DNA of cells with compromised nuclear membranes, indicative of necrosis and apoptosis. Cells that undergo rapid necrosis may not remain intact, and therefore would not be stained by PI. Because PI does not label all stages of apoptotic and necrotic cell death, we calculated densities of immature neurons, mature neurons, astrocytes and oligodendrocyte precursors as a secondary measure of vulnerability to OGD induced cell death ( Figure 3 1D L). An ANOVA revealed that although cell subtype densities differed ( F (3,40) = 44.7; p < 0.001), there was neither an effect of OGD on cell subtype densities ( F (1,40) = 1.47; n.s. ) nor a statistically sign ificant OGD exposure by cell subtype density interaction ( F (3,47) = 0.600; n.s. ). Specifically, astrocytes were detected in greater densities than all other cell types ( all p values < 0.001). Because our proportion data suggested that neurons were more vul nerable to the effects of OGD than glia, we ran planned comparisons on densities of mature and immature neurons in control versus OGD exposed cultures. Planned comparisons revealed lower densities of both mature and immature neurons in OGD exposed cultures versus control cultures ( p < 0.01 and p < 0.05, respectively). PI labels only cells whose nuclei are compromise but still intact. Although PI staining showed cell death across all phenotypes, the loss of cell density of neurons s uggests that rapid necrosi s may occur in neurons, whereas glia may succumb to apoptosis.
72 Figure 3 1. Significant cell death occurs after 120 min of OGD exposure, primarily in mature neurons. Primary cultures were exposed to 60, 120 or 180 min of OGD or fresh media change (control ) at 40d in vitro (A) Immediately following OGD exposure, cultures were exposed to propidium iodide (PI) fixed, and DAPI stained. The mean (S.E.M.) perc ent PI labeled dead/dying cells of total DAPI labeled nuclei was calculated. 120 min and 180 min OGD exposure s resulted in significant levels of cell death compared to control ( < 0.001). (B) A confocal image of cells after 180 min of OGD exposure displays the DAPI labeled nuclei of all cells (grey) and the PI labeled nuclei of dead/dying cells (red). (C) The percent PI + cells expressing different phenotype markers was quantified. (D) Cell densities of different phenotypes were compared between control (no OGD) cultures and cultures fixed immediately following 180 min OGD (no reperfusion). The expressi on of NeuN + mature neurons had decreased, but the densities of other phenotypes including tubulin + immature neurons, GFAP + astrocytes and NG2 + oligodendrocytes remained unchanged. (E L) Confocal images of neurons, astrocytes and oligodendrocyte precursor s 1d after OGD exposure. Note the DAPI labeled nucleus in all cells (grey). ( E) Normal NeuN + mature neurons appear healthy relative to ( F) NeuN + mature neurons after OGD that have
73 shriveled nuclei. ( G) tubulin + immature neurons in normal cultures versus ( H ) OGD exposed cultures that have less extensive branching. GFAP + astrocytes in normal cultures ( I ) versus 1d post OGD cultures (J) that show little change in morphology. NG2 + oligodendrocyte precursors in normal cultures ( K ) versus 1d post OGD cultures (L) that have noticeably condensed and fragmented processes. Scale bar 50 m; p <0.05, ** p <0.01, *** p <0.001 for OGD relative to controls; p <0.05 between phenotypes To examine whether glutamatergic or GABAergic signaling was altered by OGD exposure t he p ercent of mature and immature neurons expressing vesicular glutamate (vGlut) or vesicular GABA (vGAT) was quantified in OGD exposed and control culture s ( Figure 3 2A). An ANOVA revealed significant effects of OGD exposure ( F (1,24) = 6.70; p < 0.05), and neur on age ( F (3,24) = 25.4; p < 0.001) as well as a significant OGD exposure and neuron age interaction effect ( F (3,31) = 3.06; p < 0.05) on the expression of vGlut and vGAT. We found a higher percent of NeuN + mature neurons expressing vGlut or vGAT than the p ercent tubulin + neurons expressing either vesicle ( < 0.001). Among mature neurons, more were consistent with a glutamatergic (vGlut + ) versus GABAergic (vGAT + ) phenotype ( p < 0.05) whereas similar percentage s of immature neurons expressed glutamatergic and GABAergic phenotypes ( p = 0.22). Relative to a complete media change, OGD exposure decreased the number of mature neurons expressing either vGAT ( p < 0.01) or vGlut ( p < 0.05), whereas the number of immature neurons expressing either vesicle was unaffected by exposure to OGD ( Figure 3 2 A E). Of the surviving cells, we quantified the percent of neurons and glia that expressed the oxidative stress marker Cu/Zn SOD ( Figure 3 2 F G) An ANOVA found significant effects of cell subtype ( F (3, 18 ) = 10.4 ; p < 0.0 01 ) and exposure to OGD ( F (1,18 ) = 8.48 ; p < 0.0 5 ) and an OGD exposure by cell subtype interaction ( F (3, 18 ) = 3.41 ; p < 0. 05 ).
74 Specifically, astrocytes were more likely to exp ress Cu/Zn SOD than oligodendrocytes ( p < 0.05 ) and tended to express Cu/Zn SOD more than neurons ( all p values < 0.1 ) However, OGD increased neuronal Cu/Zn SOD ( all p values < 0.01). Figure 3 2. OGD increases the presence of vesicular glutamate and GA BA transporters, and Cu/Zn SOD. (A) The co expression of mature or immature neuronal markers (NeuN or tubulin, respectively) and vesicular glutamate (vGlut) or GABA A (vGAT ) transporter relative to total neurons in culture was quantified, and found to have increased in mature neurons following OGD exposure. (B C) Confocal images show immature neurons ( tubulin; red) co expressing (B) vGlut, and (C) vGAT (green). (D E) Confo cal images show mature neurons (NeuN; red) co expressing (D) vGlut, and (E) vGAT (green). (F) Cells fixed and stained for Cu/Zn SOD, which is expressed in cells as a response to oxidative stress, revealed an increase in SOD + neurons following OGD (G) Confo cal images display a Cu/Zn SOD + (red) cell 1d following OGD and DAPI labeled (grey) cell nuclei. Scale bars 50 m, p <0. 1 p <0.05, ** p <0.01
75 Neuronal Network Activity in Primary Cultures Matured Along an Expected Timeline The activity generated by neuron s within each MEA plated primary culture was recorded for 10 min every 2 3 days to confirm that it had matured into the synchronized bursting expected of mature neuronal networks prior to OGD exposure (Kamioka et al., 1996; van Pelt et al., 2004; Stephens et al., 2012) The mean (S.E.M.) spike rate, indicative of unit activity increased across days after plating ( F (14,210) = 24.1; p < 0.001; Figure 3 3A). Higher spike rates were detected during all recording sessions (Days 6 38) relative to the first recordi ng session ( < 0.001 versus Day 3), and generally increased across weeks (Days 9 and 12 < Days 26 38; Days 15 21 < Days 28 36; < 0.05) until plateauing in approximately the 4 th week after plating. Similarly, the mean (S.E.M.) percent of active electrodes within MEA dishes also increased across days after plating ( F (14,210) = 173; p < 0.001; Figure 3 3 B), with increases detected over each recording session (Days 15 38 > 12, 9 > 6 > 3; < 0.001) until stabilizing on Day 15. Burst rates indicative of network activity, increased across days after plating ( F (14,210) = 30.1; p < 0.001; Figure 3 3 C). Burst duration ( Figure 3 3 D) increased across days after plating ( F (14,210) = 18.8; p < 0.001), peaking when superburst patterns were observed ( D a ys 12 18 > D ays 21 38; all p < 0.05) and then plateaued by D ay 24 ( D ays 24 38, n.s. ) with mean (S.E.M.) burst durations of 0.761 0.097sec at 1.75 0.25 bursts/min. Because network bursting detected on ~75% of active electrodes had plateaued with mean ( S.E.M) spike rates of 2.27 0.25Hz on day 38, we began the OGD experiment on day 40.
76 Figure 3 3. Primary cultures provide a stable baseline of activity 35 40d in vitro The electrophysiological activity of primary cultures plated on MEAs was recorded for 10min every 2 prior to the OGD experiment. (A) Action potentials were detected within 3 6d in vitro Mean (S.E.M.) spike rates (Hz) increased after three weeks in vitro (day 24 vs 3 and 6; s < 0.05) and then plateaued by 28d (28 38d; n.s. ). (B) The mean (S.E.M.) percent active electrodes increased rapidly (days 15 38 > 12, 9 > 6 > 3; < 0.001) until plateauing by 18d. (C) Network bursts were first detected in the second week in vitro. Mean (S.E.M.) burst rate s significantly increased after three weeks in vitro (day 24 vs. days 9 21; < 0.05) before stabilizing by 34d (days 34 38; n.s. ). (D) Mean (S.E.M.) burst duration peaked during immature bursting activity (days 12 18 > days 21 38; < 0.05) and then plateaued by day 24 (days 24 38, n.s. ).
77 OGD Exposure Affects Network Activity in a T ime D ependent Fashion We next recorded from a subset of MEA plated cultures (n = 3) during OGD exposure to better understand how variable OGD e xposure durations may affect neuronal activity. Figure 3 4A C shows a representative raster plot of action potentials detected on each electrode (1 59; y axis) during the first 30 min of OGD (x axis). Almost immediately after OGD exposure, asynchronous act ion potentials emerged during the normally relatively quiescent inter burst intervals (between network bursts). Burst durations decreased across time and burst activity appeared to degrade by ~ 12 min after the onset of OGD exposure. A moderate amount of as ynchronous activity persisted until ~16 min after the onset of OGD exposure, when the cultures became relati vely silent, with the exception of a few occasional isolated action potentials. In support of th e hypothesis that OGD exposure initially desynchroni zes activity in neuronal networks, we found that the mean ( S.E.M.) percent of spikes generated between bursts increased during the first 4 min of OGD exposure, relative to baseline (t( 14 ) = 4.39 ; p < 0.0 01 ; Figure 3 4 D ). Consistent with the qualitative dat a we observed on raster plots of neuronal activity, an ANOVA confirmed that the mean (S.E.M.) percent of active electrodes decreased across time after onset of exposure to OGD ( F (30,6 2 ) = 13.2 ; p < 0.001; Figure 3 4 E ). Relative to baseline (~77% active electrodes), the percent active electrodes decrea sed significantly by ~ 12 min after the onset of OGD exposure when only ~37% of the electrodes were active ( p < 0.05), continually declining to the point that on average, ac tivity was only detected on ~2 electrodes per min at ~ 27 min.
78 Figure 3 4. MEA recordings during OGD reveal an increase in asynchronous activity before complete silencing. (A) A raster plot of action potentials detected during the first 30min of OGD (x axis) by electrodes 1 59 (y axis) reveals network bursts were entangled by brief periods of rapid asynchronous firing of individual electrodes that was followed by total silence by ~16min. (B) A zoomed in rasterplot of the boxed 10 sec region during OGD shows more asynchronous firing compared to 10 sec of the same MEA before OGD (C) where the majority of spikes occur during bursts. (D) The percent of spikes during interburst inte rvals increased during minutes 2 4 of OGD, indicating a loss of synchronicity. ( E ) Mean (S.E.M.) percent active electrodes during OGD exposure significantly decreased by 13min. ( F ) Mean (S.E.M.)
79 normalized spike rate of only active electrodes per MEA rev eals that OGD initially results in an increase in spike rates per electrode that is silenced by 16 min. Inset reveals a significant correlation between baseline spike rate and increase in spike rate in the first 4 min of OGD (n = 140) p < 0.10* p < 0.05, ** p < 0.01, *** p < 0.001 We next examined mean (S.E.M.) normalized spike rates on active electrodes ( Figure 3 4 F ) during OGD exposure. An ANOVA revealed that spike rates changed significantly with the duration of OGD exposure ( F (30, 1821 ) = 7.2 ; p < 0.001). During the first 5 min, according to a Fisher LSD post hoc, spike rates either significantly increased (2 and 4 min; p values < 0.01) or tended to increase (3 and 5 min p values < 0.10) relative to baseline. In fact, baseline spike rate detected o n an electrode correlated positively with the spike rate recorded on the same electrode during the first 4 min of OGD exposure when elevated rates were detected ( Pearson correlation r values = 0.82, 0.77, 0.72 and 0.64, respectively; all p values < 0.001; Figure 3 4C Inset), suggesting that spike rate may be related to the number of neurons in contact with each electrode. By 8 min of OGD exposure, spike rates were significantly lower than baseline (8 30 min; all p values < 0.05), with little to no activity de tected by ~17 min after the onset of OGD exposure. Neuronal Activity Recovers in an OGD Exposure Duration D ependent Manner We first examined the effects of OGD on network activity by quantifying changes in burst rate relative to baseline over the first 16 min of OGD exposure that bursts were detected. An ANOVA revealed that burst rates changed with exposure to OGD ( F (15,30) = 6.15; p < 0.001; Figure 3 5A ), tending to decrease at 15 min ( p = 0.06), and significantly decreasing by 16 min ( p < 0.05). Although bu rst duration decreased with time ( F (15,30) = 2.48; p < 0.02; Figure 3 5B ), there was no significant change from baseline, suggesting that exposure to OGD produced highly variable burst durations.
80 Because we observed robust activity for the first 10 min of OG D exposure, we examined the amplitudes of spikes detected on 10 randomly selected active electrodes per MEA during baseline and during the first few minutes of OGD exposure ( Figure 3 5C ). An ANOVA revealed significant effects of electrode ( F (2,54) = 6.43; p < 0.01), a significant effect of time ( F (9,486) = 4.27; p < 0.001), and the effect of OGD approached significance ( F (1,54) = 13.2; p = 0.09). However, there was a significant interaction effect of OGD and time ( F (9,486) = 3.43; p = 0.0004). The overall mean (S.E.M.) baseline spike 9 were significantly lower compared to baseline (all p < 0.05 ). The inset is an example of spike amplitude at one electrode, showing all spike forms recorded at the same electrode before OGD at baseline, and during OGD. MEA plated cultures that exhibited the synchronous bursting activity indicative of mature neuron al networks (> 40 day old cultures) were exposed to 20 min, 60 min or 180 min OGD (n = 4). Baseline data were recorded 24h prior to OGD exposure (baseline), and recoveries of neural activity data were recorded 1h and 72h after OGD exposure. The spike rates as percent of baseline spike rate ( Figure 3 5D ), and active electrodes, as percent of electrodes active at baseline ( Figure 3 5E ), were calculated. An ANOVA revealed a significant effects of OGD exposure duration ( F (2,27) = 17.4; p < 0.001) and recovery ti me ( F (2,27) = 25.5; p < 0.001) and a significant OGD exposure duration by recovery time interaction effect ( F (4,35) = 4.74; p < 0.01) on spike rates. An ANOVA also revealed significant effects of OGD exposure duration ( F (2,27) = 762; p < 0.001) and recover y time ( F (2,27) = 661; p values < 0.001) as well as a significant OGD
81 exposure duration by recovery time interaction effect ( F (4,35) = 191 ; p < 0.01) on the percent of active electrodes. Following a 20 min exposure to OGD, spike rates had returned to base line by 1h ( p = n.s .) and remained at baseline when measured 72h after OGD exposure ( p = n.s .). Interestingly, the percent of electrodes recording activity did not differ from baseline 1h after ( n.s. ) but slightly and significantly increased 3h after ( p < 0.05) a 20 min exposure to OGD. Following either a 60 min or 180 min OGD exposure, both spike rates and the percent of active electrodes were significantly lower than baseline when measured either 1 or 72h after exposure ( p values < 0.001). Note that altho ugh neuronal activity appeared to recover within 72h following a 60min OGD exposure, there was a large degree of variability. Taken together these data reveal that neural activity is silenced before significant cell death occurs and that the degree of recovery in neuronal activity is related to the duration of OGD exposure.
82 Figure 3 5. OGD causes decreased spike amplitudes, silencing of network activity by ~15 min, but full recovery if reperfused within 20 min. ( A ) Normalized burst rate during the first 15 min of OGD reveals a gradual decline in synchronous activity. ( B ) However, burst duration generally remains unchanged, although OGD initially results in higher variability. ( C ) The amplitude of action potentials was significa ntly smaller during OGD compared to baseline (D) Mean (S.E.M.) percent change in spike rate approximately 60 min and 3d following 20 min, 60 min or 180 min of OGD exposure reveals that 20 min of OGD results in recovery of activity, however, 60 180min of O GD results in a significant loss of action potentials. (E) Mean (S.E.M.) percent active electrodes following 20 min, 60 min or 180 min OGD from electrodes active at baseline reveal that 3d after 20 min OGD there was an increase in active electrodes, but 6 0 180min of OGD exposure resulted in significant loss of active electrodes. p < 0.10* p < 0.05, ** p < 0.01, *** p < 0.001
83 Discussion In this study, we found that MEA plated primary neurons exposed to OGD could provide an excellent platform for study ing the mechanisms that silence neurons following shorter duration OGD exposures and those that produce cell death, particularly among neurons at longer duration OGD Specifically we found little cell death following a 20 60 min exposure to OGD and significant ce ll death following a 120 180 min exposure to OGD. Acute cell death was particularly detected in the mature neuron population by PI labeling, as well as cell density, immediately following OGD. There was also a loss of integrity of the morphology of neuron s and oligodendrocytes, with notable swollen and fragmented processes. Activity recorded during OGD detected an initial increase in asynchronous spiking, and a decrease in spike amplitude. The extent to which spike rate increased per electrode correlated w ith the baseline spike rate of the individual electrode, and an increase in interburst activity. There was no initial change to burst rate or duration, but network bursts subsided by approximately 15 minutes. Activity recovered after reperfusion from short er durations of OGD (20 min), but remained silent after longer durations (60 min). As expected, cell death among electrophysiologically mature neuronal cultures was dependant upon the duration of exposure to OGD. Consistent with other work showing that 30 min of exposure produced approximately 20% cell death in hippocampal slice culture s (Hirt et al., 2004) we found that 20 60 min of OGD also produced ~20% cell death. This was not significantly elevated relative to the rate of cell death observed in control cultures. Our 120 180 min duration s of exposure to OGD killed 35 45% of cell in culture. Interestingly, 120 160 min of O GD in purely neuronal cultures produce s ~100% death (Meloni et al., 2002) Indeed, mature neurons
84 appeared to be significantly more vulnerable to cell death that either immature neurons or glial. Therefore, these data suggest that glia protect neurons even in culture systems of dissociated cells. Prior to OGD exp osure the activity of neurons in primary cultures evolved along an expected timeline (Kamioka et al., 1996; Wagenaar et al., 2006; Stephens et al., 2012) They produce d isolated action potentials within a few days after plating, network wide bursts after about one week, and after evolving through a transient superburstin g period, generated a slow The number of active electrodes during OGD exposure significantly decreased by 12 min compared to baseline (37.3 7.8%; p < 0.03), and gradually decreased to ~1 electrode by 27 min. Interestin gly, normalized spike rate initially increased during the first 5 min of OGD, similar to intracellular recordings on slice preps that reported a rapid initial membrane depolarization at 6 min of OGD (Tanaka et al., 1997; Pisani et al., 1998) Stroke induce d reductions in blood flow rapidly deplete intracellular ATP which inactivates ATP dependent Na + /K + transporters on neuron and astrocyte membranes and ATP dependent glutamate transporters on astrocyte membranes (Madl and Burgesser, 1993) Inactiv ated Na + /K + transporter s lead to prolonged elevations in intracellular Na + concentrations that produce cell swelling and prolonged depolarizations that contribute to increased extracellular glutamate levels (Calabresi et al., 1999) Inactivated glutamate transporters on astrocytes now not only fail to mop up extracellular glutamate, but spill additional glutamate into the extracellular space (Szatkowski et al., 1990; Zerangue and Kavanaugh, 1996) Th is could explain the initial spik e rate increase and asynchronous activity observed during the first few minutes of
85 OGD. Failure to maintain the Na + /K + gradient and consequential chronic depolarization may explain the decrease in spike amplitude, and how spike rates declined after 8 minut es of OGD exposure followed by complete silencing by 17 minutes. P revious work report ed reduced spike amplitudes after 30 min ute OGD in hippocampal slices (Mielke et al., 2007) and an initial increase in beat frequency but reduction in beat amplitude in c ardiomyocytes plated on MEAs (Yeung et al., 2009) In the current study, stable normalized burst rate s interspersed with asynchronous spik es during the first 4 minutes of OGD also suggest a gradual degradation of synchronous activity. Indeed, burst duratio n decreased during with time of OGD exposure until no bursts were detected at 16 min utes Interestingly, there was a significant positive correlation between baseline spike rate per electrode and spike rate during minutes 1 4 of OGD. If an increase in over all activity at individual electrodes is correlated with baseline activity, then elevated spike rate s at the onset of OGD likely relates to a higher number of overexcited neurons in contact with th at particular electrode. During severe ischemia, prolonged depolarizations are also associated with the rise of intracellular N a + whereas neuronal death is associated with increased intracellular Ca 2+ (Calabresi et al., 1999) Ultimately, elevated intracellular Ca 2+ levels render mitochondria dysfunction al. Coupled with generation of re active oxygen species, including the overexpression of Cu/Zn SOD (Fukui et al., 2002) and the activation of proteases, phospholipases, and endonucleases, neurons ultimately succumb to cell death (Rossi et al., 2000) Cu/Zn SOD blocks cytosolic release of cytochrome C a critical step in apoptosis, and thus reduc es apoptosis after transient focal cerebral ischemia (Fujimura et al., 2000) The percent of Cu/Zn SOD cells was significantly
86 higher in neurons one day following OG D, but consistently high in astrocytes (~40%) This coincides with reports that free radical production peaks after reperfusion (Perez Velazquez et al., 1997) Although SOD over expression is a sign of oxidative stress, it can also be neuroprotective (Zaghl oul et al., 2012) Therefore, perhaps astrocytes are less vulnerable to OGD because they primarily produce more Cu/Zn SOD. Previous work has shown that transient focal ischemia in the brain increases v glut1 (Kim et al., 2005) until 3d reperfusion, that the n decreases after 7d. The proportion of v Glut and v GAT positive, mature neurons increased following OGD, but there was no change in the percent of v Glut or v GAT positive immature neurons. Ischemic conditions lead to cleavage of v GAT via calpains inducing a loss of synaptic delivery, leading to a homogeneous distribution along neurites for ~24h following MCA o The v GAT antibody recognizes both v GAT and the cleaved v GAT, which may explain the increase (Gomes et al., 2011) However, activation of glutamate t ransporters results in release of GABA through reversal of glial GABA transporters, which may function as a negative feedback to combat excitation in pathological conditions such as ischemia (Heja et al., 2009) However, there are conflicting reports that v GAT may actually decrease after 180min reperfusion (Vemuganti, 2005) 10 minute OGD produced irreversible loss of field potential from striatal slices but AMPA anatagonists (not NMDA antagonists) allowed for recovery (Calabresi et al., 1999) Thus, neurop rotection may be dependent on reducing glutamate release. Interestingly, oligodendrocyte precursors exhibited increased vulnerability to OGD induced cell death. Death of oligodendrocytes and their progenitors is mediated by Ca 2+ influx via nonNMDA receptor s by the release of glutamate from autologous reversed
87 glutamate transport (Fern and Moller, 2000; Tekkok and Goldberg, 2001; Back et al., 2002) NG2 + oligodendrocyte precursors in normal cultures versus 1d post OGD cultures had noticeably condensed and fr agmented processes ( Figure 3 1K versus 1L), which coincides with previous observations of fragmented and swollen myelinating processes following 60 min MCAo in rats (McIver et al.) NG2 + cells may be susceptible to ischemia because they receive functional glutamatergic synapses from neurons (Mangin and Gallo, 2000) and like neurons, display an inward current and increase in internal Ca2 + during ischemia (Domercq et al.) However, they also actively proliferate in all areas in the brain, which may hide any l oss in cell density (Back et al., 2002) Notably, some of the mitotic cells that we detected expressed oligodendrocyte precursor phenotypes. Reports of recovery following shorter durations of OGD show conflicting results. A 30 min exposure in hippocampal slice cultures led to reduced activity, evident in reduced evoked potential amplitude localized by 6 locations in the CA1, induced by stimulation in the CA3 (Hirt et al., 2004) Using one electrode on the dentate gyrus following OGD, however, there was alm ost no recovery of evoked po pulation bursts after 10 min OGD (Kass and Lipton, 1982) Previous work using 4 7 electrodes placed in the cortex in vivo during transient middle cerebral artery occlusion (MCAo) saw recovery of neuronal activity fro m 10 180 min OGD (Rosner et al., 1986) With the utility of recording from 59 sites, MEAs provide more data simultaneously, thus less variation in results can be achieved relative to what is found by just a few electrodes. Neural activity recovered following reperfusi on after a 20 minute OGD exposure in our experiment, but significantly diminished following a 60 minute OGD exposure. However,
88 a 60 minute OGD exposure did not lead to significant cell death, indicating that neuron silencing occurs while cells remain viabl e. In the brain, an area called the penumbra, encircles the necrotic tissue of the stroke core (Astrup et al., 1981) The penumbra is composed functionally silent, yet structurally intact tissue in which blood supply is constrained but energy metabolism is preserved. If s horter durations of OGD cultures are functionally silent but structurally intact, and longer OGD exposed cultures are completely dead, perhaps this model would be most useful by customizing the level of damage to fit the type of injury for the treatment strategy being tested. The MEA OGD model provides a global view into the disintegration of neuronal network s similar to ischemic stroke pathology described in situ Furthermore, the extent of cell death in stroke models may not be the ideal benchmark for establishing OGD models, because activity is silenced long before significant cell death is observed. Future research could benefit by utilizing this model by applying treatments or cells following OGD, and monitoring MEA recordings for funct ional recovery of individual neurons, as well as network bursts. Perhaps shorter durations of exposure provide a more representative model that could be customized for the type of treatment strategy (treating penumbra or core) OGD exposed MEA plated neuro ns nonetheless, appear to be an effective platform for investigating the effects of hypoxic ischemic stroke on neuronal networks, and could potentially be utilized as an effective bioassay for cell transplantation or pharmacological treatment on mature neu ral networks.
89 CHAPTER 4 ADULT NEURAL PROGENI TOR CELLS PARTIALLY PROTECT MICROELECTRO DE ARRAY PLATED NEURAL NETWOR KS FROM OXYGEN GLUCOSE DEPRIVATION INDUCED DEATH AND SI LENCING Introduction Neural stem and progenitor cell s trategies to repair the central nervous syste m (CNS) have improved function in some animal models of human disease and injury (Nikkhah et al., 1994a; Dziewczapolski et al., 2003; Ben Hur et al., 2004; Gao et al., 2006) Understanding whether transplantable cells provide trophic s upport or integrate functionally into surviving neural circuits would be beneficial for optimizing these strategies and subsequently translating them into the clinical setting I n vitro assay system s that collect optical and electrophysiological data simul taneously online may provide insight into how neural stem cell progeny mature over several weeks and affect the activity of these damaged neural circuits, which could be applied to animal models. Neurons in MEA plated primary cultures harvested from embry onic day 18 rat neocortices have been well characterized to generate activity that e volve s predictably over several weeks to a mat ure, stable bursting state by 35 40 days (Kamioka et al., 1996; van Pelt et al., 2004 ; Stephens et al., 2012 ) In these cultur es, asynchronous spike activity emerges coincident with immature neuronal phenotypes (within days) and permanently robust numbers of astrocytes and oligodendrocyte precursors. Superbursting activity emerges coincident with neuronal phenotypes that appear t ransitional between immature and mature (~2 weeks), and permanent bursting activity emerges coincident with mature neuronal phenotypes (~4 weeks). Previously, we foun d that NPCs re evoke developmental superbursting activity in electrophysiologically
90 mature neural cultures, potentially because some mature neuronal NPC progeny integrate functionally into neural networks ( Chapter 2 ). MEA plated cultures are amenable to the development of injury models because their activity can be evoked or monitored for mont hs or years after injury, and/or following the addition of potentially therapeutic factors or cells to their media after injury (Potter and DeMarse, 2001; Scarlatos et al., 2008) Their optical transparency permits the concomitant examination of cell death and the behavior of transplanted cells. The effects of oxygen glucose deprivation (OGD) on cardiomyocytes behavior has been characterized using MEA technology (Yeung et al., 2009) but the effects of OGD on network activity in stroke models has only been investigated previously using 1 7 electrodes in vivo (Kass and Lipton, 1982; Rosner et al., 1986) MEA plated neural cultures may be particularly valuable as in vitro models of stroke because the well known effects of stroke on both cell death and neuronal network activity can be monitored simultaneously using MEA technology, and culture models of stroke are thought to resemble the in vivo post stroke milieu (Meloni et al., 2001; Harms et al., 2010) For example, stroke induced blood hypofusion rapidly de pletes ATP stores, which inactivates ATP dependent Na + /K + pumps on neurons (Madl and Burgesser, 1993) and reverses glutamate transporters primarily expressed on astrocytes (Szatkowski et al., 1990; Zerangue and Kavanaugh, 1996; Calabresi et al., 1999) Th e combination of prolonged neuronal depolarizations and inefficient clearance of now excess glutamate from the extracellular space by astrocytes leads to excitotoxic cell death and increased oxidative stress among surviving cells primarily because of mitoc hondrial dysfunction
91 (Rossi et al., 2000; Fukui et al., 2002) We have previously shown that MEA plated cultures contain all of the cell players that mediate the effects of stroke in vivo (Chapter 3), and that while all cell subtypes express markers of cell death following exposure to long duration OGD, mature neurons and oligodendrocytes may be particularly vulnerable. Here we investigate the effects of NPCs on neural activity and cell de ath in MEA plated cultures exposed to OGD. NPCs transplanted into the cortex, hippocampus, or striatum of rodents following middle cerebral artery occlusion can improve neuronal survival and promote behavioral recovery (Hodges et al., 1996; Sinden et al ., 1997; Modo et al., 2002; Harms et al. 2010 ) Some evidence suggests that neuronal NPC progeny integrate functionally into the post ischemic hippocampal CA1 region because they eventually express pre and post synaptic markers and resemble normal hippoc ampal neurons found within the infarct (Aoki et al., 1993) Other evidence suggests that NPCs stimulate plasticity or reduce cell death among compromised circuits by secreting trophic factors or by perhaps physically buffering the compromised circuitry fr om damage (Andres et al. 2010 ; Modo et al., 2003; Boehler et al., 2007; Lee et al., 2007) Identifying these mechanisms may facilitate their potential capacity for neuroregeneration. Of course, a hypoxic/ischemic environment may also affect NPC behavior beneficially or detrimentally. The hypothesis that transplantable NPCs are as exquisitely sensitive to sy as endogenous NPCs (for example, hormones, stress, neuroinflammation glutamaterigic transmission and lif e experience (Gould et al., 1992; Cameron et al., 1993; Nilsson et al., 1999; van Praag et al., 1999; Mizumatsu et al., 2003; Monje et al., 2003; Ormerod et al., 2003; Nacher and McEwen, 2006) is highly
92 conceivable. NPCs express metabolic glutamate receptors that would likely be high ly activated in an excitotoxic environment to modulate the proliferative, differentiative and survival promoting activities they have been implicated in (Melchiorri et al., 2007) Indeed, NPCs resident to the neurogenic hippocampal dentate gyrus proliferat e in response to ischemia (Liu et al., 1998; Takagi et al., 1999; Kee et al., 2001; Yagita et al., 2001) Furthermore, neurons are produced in non neurogenic regions, such as the striatum, following ischemia but their survival is transient (Arvidsson et al ., 2002; Zhang et al., 2004; Thored et al., 2006) Therefore, we also investigate the effects of the post OGD environment on the survival and phenotypes of NPCs and their progeny. Material and Methods Primary Culture Preparation MEA and LabTek plated pri mary cultures were generated from the same pool of dissociated cells for each experiment. Before plating, polyethylene imine (PEI) in borate buffer (Sigma Aldrich; St. Louis, MO) was placed on the center of each autoclaved MEA dish or s terile LabTek well After 1h the dishes and chambers was placed on the center of each MEA or LabTek well and incubated at 37 C for 1h. E mbryonic d ay 18 (E18) Sprague Daw ley rat neocortices were purchased from Brain Bits (Springfield, Illinois) and dissociated in p apain ( 20 units /ml ; W orthington Biochemical; Lakewood, NJ) and DNase 1 (2000 units/ml, Worthington Biochemical) in GlutaMax supplemented DMEM (Gibco; Grand Island, NY) at 37 C for 30min before gentle triturat ion, and pelleted by centrifugation for 3min at 1500rpm. The cells were re Cellgro Antibiotic Antimycotic (Mediatech, Inc.; Herndon, VA) in DMEM/Glu taMax and plated by placing a
93 100l droplet of cell suspension containing 25K cells directly onto a laminin droplet placed in the center of the MEAs ( n = 21) or LabTeks ( n = 24; > 2 wells per staining set) to ensure equal plating densities of ~4,000 cells/m m 2 Once the cells adhered to the substrate (~ LabTek well and 1ml/MEA. One half of the media was replaced with fresh media 24h prior to recording sessions, or every 72h to avoid disrupting electrophysiological activity. Each MEA and LabTek was assigned randomly to one condition group (control, OGD) and treatment group (10K NPCs, 10 K HDF s, NPC conditioned media) after the recording session conducted on day 3 8 confirmed that the electrophysiological activity among MEA plated pr imary cultures had matured and the OGD experiment initiated on D ay 40. As expected, neurons in primary cultures generated mean spike rates and burst rates that increased across days after plating (spike rate: F (14,168 ) = 28.8; p < 0.00 1 and burst rate: F ( 14,168 ) = 29.1; p < 0.00 1 ) but that did not differ by eventual condition (spike rate: F (1,12 ) = 0.68; n.s. and burst rate: F (1,12 ) = 0.07; n.s. ) or treatment (spike rate: F (2,12 ) = 0.01; n.s. and burst rate: F (2,12 ) = 1.57; n.s. ) groups and no interaction between day and eventual condition or treatment group on either spike rate ( F (14,168 ) = 1.34; n.s. and F (28,168 ) = 0.89; n.s., respectively) or burst rate ( F (14,168 ) = 0.71; n.s and F (28,168 ) = 0.61; n.s ., respectively) was observed. The mean percent acti ve electrodes within MEA dishes increased across days after plating ( F (14,168 ) = 198.8; p < 0. 001), with increases detected over each recording session (days 15 38 > 12, 9 > 6 > 3; all p < 0.001) until stabilizing on day 15. Eventual condition and treatmen t groups were not different ( F (1,12 ) = 1.12; n.s and F (2,12 ) = 1.89; n.s ., respectively), and the interaction between days after plating and eventual condition or treatment group on percent active electrodes did not
94 approach significance ( F (14,168 ) = 1.11; n.s and F (28,168 ) = 1.41; n.s ., respectively). Burst duration increased across days after plating ( F (12,144 ) = 13.5; p < 0. 001), peaking when immature bursting was observed (days 12 18 > days 21 38; all p < 0.05) and then plateaued by day 24 (day s 24 38, n.s. ) with mean ( S.E.M.) burst duration of 0.41 0.235sec at 5.97 0.75bursts/min. Again, eventual condition and treatment groups were not different ( F (1,12 ) = 0.02; n.s and F (2,12 ) = 1.28; n.s ., respectively), and the interaction between days after plating and eventual group did not approach statistical significance ( F (12,144 ) = 1.41; n.s .). However, there was an interaction effect between eventual treatment group and day ( F (24,144 ) = 2.0; p < 0 .01), probably due to high variability in burst d uration during immature activity (days 12 18). Because planned comparisons confirmed that mean spike rates, burst rates, percent active electrodes and burst durations recorded before day 40 were consistent between groups, in line with previous reports (Ste phens et al., 2012) we felt confident that the random distribution of primary cultures did not confound our interpretation of the effects of cell addition on neuronal activity following OGD exposure. Neural Progenitor Cells The adult rat hippocampal NPC line used in the present study was generated from the hippocampus of a male GFP expressing Wistar TgN(CAG GFP)184Ys rat obtained from our breeding colony maintained by University of Florida Animal Care Services and descended from breeding pairs obtained from t he Rat Resource and Research Center (University of Missouri, MO). All animals produced in this specific pathoge n free facility housed colony ( maintained on a 12:12h light:dark cycle at 23 1 C ) were pair housed after weaning in shoebox cages and had free access to water and food (Teklad LM 485;
95 Harlan Laboratories). Briefly, the rat was anaesthetized with CO 2 and de capitated. The hippocampi were rapidly dissected, minced and digested enzymatically in papain (2.5 un its/ml ), dispase II (1 unit/ml; Roche Applied Science ; Indianapolis, IN ) and D Nase I (250 units/ml ) at 37 C for 45 min. The dissociated tissue was fractiona ted in a 50% and then 20% Percoll solution by cen trifugation (20,000xg) for 10min. The f ractionated cells were plated onto poly L ornithine/laminin coated dishes in DMEM/F 12 (1:1) medium (Mediatech; Manassas, VA) containing 10% fetal bovine serum (FBS) fo r 24h and then the medium was replaced with DMEM/F12 (1:1) w/ L glutamine supplemented with N2 (Invitrogen) and 20ng/ml recombinant human fibroblast growth factor 2 (FGF 2; Peprotech, Inc. Rocky Hill, NJ ). NPCs were passaged 10x before use Before co cultu re, NPCs were resuspended in primary growth media to wash out any remnant growth factors that may influence primary cells. Conditioned media was collected fresh from NPCs grown on 10cm tissue culture plates in 8ml of primary cell growth media, and centrifu ged to filter any aggregate NPCs that could have been collected. Human D ermal F ibroblasts Adul t HDF s (Cascade B iologics; Portland, OR) were plated at ~150,000 cells were plated into T 25 culture flasks with 5ml media ( Media 106 (G ibco ), 2% FBS 1% a ntibioti c a ntimycotic hydrocortisone (1g/ml), heparine (10g/ml), epithelial growth factor ( EGF; 10ng/ml) and FGF 2 (3ng/ml) ), grown to confluency (~7d) and passaged via trypsinization. After three passages, 1x10 6 cel ls were suspended in 1ml media, 1ml lentitangerine (produced by D D ; titer=1.8x10 7 TU/mL), and 10 l polybrene ( Sigma Aldrich) The cells were incubated in virus for 30min at room temperature and then overnight at 37 C. The following da y the media was replaced. At 5d post infection
96 58.03+/ .03 % cells expressed m tangerine. HDF s were passaged ~6x after infection before being used in co culture. Oxygen G lucose D eprivation and Co Culture Before the MEA experiment, we first examined total cell death in 40d LabTek plated primary cultures exposed to O GD for 60, 120 or 180 min, or 100% fresh media change (control), in order to find an optimum exposure time to conduct the experiments. Immediately following OGD, the cultures were PI stained and fixed. An ANOVA on the mean ( S.E.M.) % of DAPI + nuclei co la beled with the cell death marker PI revealed significant effects of OGD exposure duration on cell death ( F (3,23 ) = 31.51; p < 0.001). Significantly more dead cells were found in cultures exposed to OGD for either 120 minutes ( p < 0.001) or 180 minutes ( p < 0.001; Figure 1A), relative to control cultures. We further characterized the effects of a 180 min OGD exposure on cell death and oxidative stress because this duration of OGD exposure killed ~50% of the culture, which often serves as a benchmark for explor ing the mechanisms of stroke induced cell death (Meloni et al., 2001; Hirt et al., 2004) On day 40, media was removed and replaced with DMEM containing no glucose, no sodium pyruvate, low L glutamine (G ibco ) pre exposed to hypoxia overnight, f or MEAs / LabTeKs in OGD groups,. T he dishes were immediately placed in the hypoxic incubator ( 37 C, 0.1% O 2 5.0% CO 2 94.1% N 2 ) for 3h Media was completely replaced with fresh media in non OGD groups in order to rule out any effect due to media change. After OGD the DMEM was replaced with fresh media, and the cultures were returned to an incubator at normal conditions ( 37 C, 20.95 % O 2 5.0% CO 2 ) 24h after OGD, the media was removed and 100l of primary growth media containing 10K
97 NPCs, 10K HDFs or no cells was a dded to the center of each LabTek chamber and MEA. Once the NPCs adhered to the primary cultures (5 10min), media was added to fill the dish. Every 24h 48h, ~50% of the media was replaced with fresh media. The conditioned media control group was given 75% fresh media, 25% NPC conditioned media, for the remainder of the experiment. Immuno cy tochemistry and Microscopy Lab Tek plated cells were exposed to PI (250 g/ml TBS; Sigma Aldrich) to visualize dead/dying cells, fixed with 4% paraformaldehyde for 10min and rinsed repeatedly between immunostaining steps with TBS (pH 7.4) The cells were blocked in 3% normal donkey serum (NDS) and then incubated overnight in a cocktail of 3 primary antibodies that included rabbit or mouse anti Tubulin (1:500; Covance ) mouse anti neuronal nuclei ( NeuN ; 1:500; Chemicon International), guinea pig anti glial fibrillary acidic protein ( GFAP ; 1:750; Advanced Immunochemical), rabbit anti NG2 (1:1000; Chemicon International ), and Cu Zn superoxide dismutase (SOD; 1:500; Millapor e) at 4C. The following day, the sect ions were incubated in the appro priate FITC Cy3 or Cy5 conjugated secondary antibody (1:500; Jackson ImmunoResearch) for 4hrs Then cells were DAPI stained (1:20,000 in TBS for 20min; Calbiochem) and cover slipped under PVA DABCO. lmages of LabTek plated cultures were taken using a Zeiss LSM 710 fully spectral Laser Scanning Confocal Microscope with a 63 x oil objective (5 z stacks) of DAPI + and GFP/DAPI + cells in primary cultures and in co cultures. Visible laser line intensities w ere maintained below 15%. DAPI + cells with appropriate neuron, astrocyte and oligodendrocyte morphologies and intense staining with the appropriate antibodies in
98 primary cultures served as positive controls to which exposure levels and g ain were set to exclude nonspecific staining among GFP + cells (NPCs). 8 10 non overlapping images were taken diagonally from each corner of > 2 wells (>500 cells/ well). We measured the % PI + and %SOD + DAPI labeled nuclei, and primary or GFP+ (NPC generated) cells co labeled with the mature neuronal marker NeuN, the immature neuronal marker tubulin, the astrocyte marker GFAP, and the oligodendrocyte precursor marker NG2. Mitotic and pyknotic nuclei were noted in each image. Fluorescent images of immunostain ed MEAs were taken using a Zeiss reflected light fluorescent microsco pe using a 20x or 40x objective. Electrophysiology Electrophysiological data was collected from cells plated on MEAs with a square grid of 60 planar Ti/TiN (Multichannel Systems, Reutlingen, Germany) As described previously (Stephens et al., 2012) changes in extracellular potential were recorded over 10min sessions every 3d from 3 39d after plating to es tablish a basline, and then every 4d from 21 68d after OGD to record recovery Briefly, the signal from 59 recording electrodes was sampled at 25 kHz, and amplified and digitized using Multichannel Systems (Reutlingen, Germany) MEA 10 60 amplifier and A/D ha rdware. Cultures were allowed to equilibrate for ~2min after being placed into the amplifier before recording was initiated. Action potentials, or spike events, were detected online as upward or downward excursions beyond 5x the median threshold from the q uiescent signal that was estimated during a Waveforms between 1ms before and 2ms after detected spike s were saved and used to remove duplicate detections of
99 multiphasic spikes and visually inspect for a rtifacts These data was saved for further offline processing using a software package custom written by Dr. Thomas B. DeMarse. E l previous reports of minimum firing rates (Latham et al., 2000; Boehler et al., 2007) Network bursts were defined in previous reports as >3 or >14 electrodes spiking within 0.1 2sec inter spike intervals (Gross and Kowalski, 1999; Tabak and Latham, 2003; van Pelt et al., 2004) We categorized a network burst as a cluster of spikes across >10 electrodes within 0.2sec inter spike intervals From these data, spike and burst frequency, number of active electrodes and burst duration were quantified offline Statistical Analysis Statistical tests were conduct ed using St atistica Software ( StatSoft, Inc ; Tulsa, OK) Viability (% total PI+ and pyknotic cells), oxidative stress (% total SOD labeled cells), division (% mitotic), densities and proportions of neurons (NeuN + tubulin + or tubulin/NeuN + ) or glia (G FAP + or NG2 + ) primary (DAPI + ) or NPC (DAPI/GFP + ) cell origin was compared between OGD versus fresh media exposed cultures supplemented with NPC conditioned media versus cells (NPCs or HDFs) at 0, 1, 7, 14 or 28 days after reperfusion using mixed ANOVAs. U nit activity (% change in spike rate ) and network activity ( burst rate, percent active electrodes, and burst duration) were compared between OGD versus fresh media exposed cultures supplemented with NPC conditioned media versus cells (NPCs versus HDF s) across days using repeated measures ANOVAs Newman Keuls p ost hoc tests were employed to explore main and levels = 0.05.
100 Results NPCs Transiently Protect Neural Cultures from Death and Oxidative Stress Primary cultures were exposed to 3h OGD or a complete media change at 40d in vitro before being PI stained and fixed 1d, 7d, 14d, or 28d after reperfusion with and without cell addition ( Figure 4 1A G). An ANOVA revealed significant effects of OGD ( F (1 30 ) = 153 ; p < 0. 001), cell addition ( F (2 30 ) = 15.4 ; p < 0. 001), and survival time after reperfusion ( F (3 90 ) = 83.8 ; p < 0. 001) and significant interactions between OGD and survival time after reperfusion ( F (3 90 ) = 74.5 ; p < 0. 001), cell addition and surviv al time after reperfusion ( F (6 90 ) = 7.59 ; p < 0.001) as well as an overall interaction ( F (6 90 ) = 4.86 ; p = 0. 001) on dead cells (Fig 4 1A) The mean ( S.E.M.) % of PI + (dead) cells was significantly increased by OGD ( p < 0.001) and more cell death was observed on Days 1 ( p values < 0.001) and 7 ( p values < 0.001) versus 14 and 28. Regardless of OGD exposure or survival time after reperfusion, cell death was significantly lower in NPC versus NPC conditioned media ( p < 0.05 ) or fibroblast ( p < 0.001) su pplemented cultures. In control cultures, fibroblast addition slightly increased cell death on Day 1 ( p values < 0.1) relative to NPC + cells were detected in OGD exposed cultures that had been supplemented with NPCs relative to either NPC conditioned media or fibroblasts on Day 1 ( all p values < 0.05), the percentage was not decreased to those observed in control cultures ( all p values < 0.05). In addition, the beneficial effect of NPC versus NPC conditione d media or fibroblast addition among OGD exposed cultures was not apparent at 7d, 14 or 28d after reperfusion ( p values = n.s .). Our data include several important findings. First, we found similar effects of OGD on cell death among cultures fed every 72h with fresh media ( data not shown ) or with NPC conditioned media ( Figure 1A), suggesting that
101 any protective effects of NPCs on OGD exposed cultures are cell contact dependent. Second, we found that fibroblasts slightly increase cell death. Third, we show th at NPCs may partially mitigate the immediate, but perhaps not later, effects of OGD exposure on cell death. To further determine the effect of OGD with and without cell addition on overall cell death, we calculated the proportion of pyknotic DAPI + nuclei indicative of dying cells ( Figure 4 1B). An ANOVA revealed significant effects of OGD on the percent of pyknotic cells ( F (1,78) = 103.0; p < 0.001) and days after reperfusion ( F (3,234) = 25.0; p < 0.001), but no effect of cell addition ( F (2,78) = 2.05; n.s. ). OGD interacted with days after reperfusion ( F (3,234) = 14.4; p < 0.001) to affect the percent of dying cells observed. Overall, relatively few pyknotic cells were detected. However, OGD increased the percent of pyknotic nuclei ( < 0.001) detected at Days 1 ( p values < 0.001 ) and 7 ( p values < 0.001 ) relative to Days 14 and 28. To quantify the effects of OGD on oxidative stress, we quantified the proportion of cells that expressed Cu/Zn SOD in cultures exposed to a full media change or OGD, 1d, 7d, 1 4d, or 28d after reperfusion with and without cell addition ( Figure 4 1C). An ANOVA revealed significant effects of OGD ( F (1 42 ) = 37.3 ; p < 0. 001), cell addition ( F (2 42 ) = 22.0 ; p < 0. 001), and survival time after reperfusion ( F (3 126 ) = 6.69 ; p < 0. 001), and interaction effects between OGD exposure and survival time after reperfusion ( F (3 126 ) = 4.20 ; p < 0.01), cell addition and survival time after reperfusion ( F (6 126 ) = 4.76 ; p < 0.001), and OGD exposure and cell addition ( F (2,42 ) = 4.07 ; p < 0.05), but no overall interaction ( F (6 126 ) = 1.36 ; n.s. ). The percent of cells expressing Cu/Zn SOD was significantly higher after OGD exposure (p < 0.001) and at 1 versus Days 14 and 28
102 after OGD exposure ( p values < 0.01). In control cultures, the perc ent of cells expressing Cu/Zn SOD was similarly low regardless of cell addition but in OGD exposed cultures, the percent of cells expressing Cu/Zn SOD was significantly higher in the NPC conditioned media supplemented cultures on Day 1 versus either the fi broblast (p < 0.001) or NPC supplemented cultures (p values < 0.001) and in NPC conditioned media supplemented control cultures (p values < 0.001). These data suggest that either fibroblasts or NPCs can minimize oxidative stress in cultures exposed to OG D. As neurons transition from immature to mature they alter their expression of neuron specific proteins. For example, young neurons that exhibit immature morphological and electrophysiological phenotypes express proteins such as tubulin while those exp ressing more mature phenotypes express proteins such as NeuN. We also find that there is a transitional neuronal phenotype that expresses both tubulin and NeuN ( Figure 4 1D; and Chapter 2 ). To ascertain whether immature, transitioning or mature neurons in our primary cultures may be selectively vulnerable to the effects of OGD, we quantified densities of tubulin + immature neurons, tubulin/ NeuN + transitioning neurons and NeuN + mature neurons. An ANOVA revealed no significant effect of OGD ( F (1 42 ) = 0.12; n.s. ), but did reveal significant effects of neuron age ( F (2 42 ) = 28.0 ; p < 0.001) and survival time after reperfusion ( F (3 126 ) = 3.38 ; p < 0. 05) on cell densities. Furthermore OGD by neuron age ( F (2 42 ) = 26.4 ; p < 0. 001), neuron age by surviva l time after reperfusion ( F (6 126 ) = 7.27 ; p < 0. 001) and an overall ( F (6 126 ) = 2.83 ; p < 0. 05) interaction effects were observed. Overall, higher densities of immature neurons were observed than either transitioning ( p < 0.001) or mature neurons ( p < 0.001).
103 Regardless of neuron age, neuron densities were higher at 14 and 28 days after reperfusion relative to Days 7 ( p values < 0.05). These effects largely appear due to the fact that densities of immature neurons increased at 14 and 28 days after OGD exposure ( p values < 0.001). These data suggest that young neurons may either be produced or driven to express tubulin in the weeks following OGD. Because the densities of immature neurons increased after OGD exposure, we quantified mitotic phase nuclei to determine whether OGD potentially stimulated the production of new neurons. Mitotic phase nuclei were detected at extremely low levels (< 1%; data not shown). However, an ANOVA revealed significant effects of OGD ( F (1,78 ) = 10.5 ; p < 0.01) and cell addi tion ( F (2,312 ) = 8.94 ; p = 0.001), but not survival time after reperfusion ( F (3,234 ) = 0.20 ; n.s. ) on the % of mitotic cells. Overall OGD exposure increased the % percent of mitotic nuclei ( p < 0.01) a higher % of mitotic nuclei were detected in NPC versu s NPC conditioned media or fibroblast supplemented cultures regardless of OGD exposure ( p values < 0.01). While OGD exposure and cell addition could have stimulated the proliferation of latent NPCs in our primary cultures, the low density of mitotic figure s that we detected after OGD did not warrant further investigation.
104 Figure 4 1. NPC addition ameliorates initial cell death, but OGD increases the presence of pyknotic nuclei Cu/Zn SOD and immature neurons. (A) The percent of PI + dead/dying cells incr eased following OGD, but by 14d was no different from controls. When comparing the different treatment groups (n = 4), there was a significant increase in cell death following OGD in all groups; however, fewer dead cells were found in the NPCs added group 1d post OGD. (B) Pyknotic nuclei were also quantified to measure dying cells (n = 4), and increased in all groups following OGD that declined with time, and did not differ from controls 28d post OGD. (C) % Cu/Zn SOD cells increased latently, increasing 7 14d following OGD. There were fewer Cu/Zn SOD + cells in the NPCs added group and no difference in the HDFs added group (n = 4). (D) The transition of neurons from immature ( tubulin + ), to transition ( tub ulin + and NeuN + ), mature (NeuN + ) states overtime was monitored (n = 8). OGD actually increased the number of immature neurons, 14 and 28 days after OGD exposu re Confocal images of (E) transition state neurons, ( F ) middle cell in pyknosis, with a fragmente d nucleus, and (G) bottom left nucleus in mitotic phase, with condensed, splitting chromatin, compared to normal top right nucleus. Mitotic phase nuclei, although present at low levels (<1%), did increase after OGD. Scale bar 50 m. p <0.1, p <0.05, ** p <0 .01, *** p <0.001
105 OGD Exposed Primary Cells Did Not Affect NPC Fate The obvious goal of stem/progenitor cell strategies is to promote the integrity of compromised circuitry, but testing the effects of variable post injury milieus on the behavior of neural s tem/progenitor cells may also be critical to generating clinically relevant strategies. We first examined the fate of GFP + adult hippocampal NPCs 1d, 7d, 14d or 28d after their addition to OGD exposed or control cultures ( Figure 4 2A). An ANOVA revealed simi lar fates among GFP + co cultured with OGD exposed versus control cultures ( F (1,24) = 0.01; n.s. ), but preferred phenotypes among co cultured GFP + cells ( F (3 24 ) = 60.9 ; p < 0. 001) that changed significantly with time after co culture ( F (3 72 ) = 17.0 ; p < 0.001), and interacted significantly with time after co culture ( F (9,72) = 6.03; p < 0.001). A higher percent of NPC derived (GFP + ) cells expressed tubulin versus NG2, GFAP or NeuN ( all p values < 0.001), more NeuN than GFAP or NG2 ( all p values < 0.05) and more GFAP than NG2 (p = 0.05). Relative to Day 1, the percent of GFP/ tubulin + neurons increased at Day 7 ( p < 0.001), Day 14 ( p < 0.001) and Day 28 ( p < 0.001). More GFP/NeuN+ progeny were detected at Days 28 and 14 versus Days 1 and 7 when none wer e detected ( p values < 0.01). The percent of GFP/GFAP + astrocytes remained consistently low across days in co culture and few GFAP/NG2 + oligodendrocyte precursors were detected. Interestingly, NPCs grown alone for 28d differentiated into 7.7 1.3% tubulin + immature neurons and 7 9 3.2% GFAP + astrocytes (data not shown). The proportion of GFP + NPCs to total DAPI labeled cells in co culture approached significance among conditions ( F (1 120 ) = 3.67 ; p = 0. 060) and significantly changed with time ( F (3 120 ) = 28.0 ; p < 0. 001), but there was no interaction effect ( F (3 120 ) = 1.06 ; n.s. ).
106 Overall, there were fewer NPC derived cells in the OGD group ( p = 0.058). There were ~3 5% GFP + NPCs of all cells in co culture 1d after plating that increased over t ime (all p < 0.001), suggesting that NPCs continued to divide ( Figure 4 2B). By 28d, there was a significantly smaller percent of GFP + cells in the OGD exposed co cultures than the controls ( p < 0.05), which could be contributed to faster differentiation int o mature cells and/or an increase in primary cell division. The proportion of m tangerine + HDFs ( Figure 4 2G) to total DAPI labeled cells in co culture did not differ with condition ( F (1 120 ) = 0.07 ; n.s. ), but changed over time ( F (3 120 ) = 11.2 ; p < 0. 001) After one day in co culture, HDFs composed ~3 5% of total DAPI labeled cells and decreased slowly over time (14d and 28d versus 1d, all p < 0.05), suggesting that either the HDFs may not survive long term in co culture.
107 Figure 4 2. Co culture with OGD exposed cells does not affect NPC fate. The proportion of adult hippocampal NPC derived (GFP + ) cells expressing immature neuronal ( tubulin), matu re neuronal (NeuN), astroglial (GFAP) or oligodendroctye precursor (NG2) at 1d, 7d, 14d and 28d after addition to OGD exposed or non OGD cultures was quantified on confocal laser scanning microscope digital images (n = 4). (A) NPC progeny began to express immature neuronal and glial markers 1 7d in co culture. NeuN + mature neurons derived from NPCs were observed 14d in co culture, and few (~2%) NG2 + oligodendrocyte precursors derived from NPCs were observed in co culture with OGD exposed primary cells. (B) The proportion of GFP + NPCs to total DAPI labeled cells in co culture was ~3 5% at plating and increased over time, suggesting that NPCs continued to divide (n = 16). By 28d, there was a siginificantly smaller percent GFP + cells in the OGD exposed co cultu res than the controls (* p < 0.05), which could be contributed to faster differentiation into mature cells (evident in greater % phenotypes at 14d) and/or and increase in primary cell division. (C F) Confocal images of GFP + NPC progeny (green) expressing tubulin ( C ; red), NeuN ( D ; red GFAP ( E ; red) and NG2 ( F ; red). Note that all cell nuclei are DAPI labeled (grey). (G) The proportion of mtangerine + HDFs to total DAPI labeled cells in co culture was ~3 5% at plating and decreased slowly over time (n = 16) Inset is a confocal image showing mtangerine + HDFs (orange) and DAPI labeled nuclei (grey). Scale bar 50 m.
108 Neuronal Death Following OGD is Less Severe Following NPC Addition The percent of PI + cells co labeled with NeuN, tubulin, GFAP and NG2 wa s quantified in order to detect a cell type that may be more susceptible to OGD ( Figure 4 3A). There was, in fact, an effect of phenotype ( F (3 72 ) = 12.5 ; p < 0. 001), an interaction of OGD and phenotype ( F (3 72 ) = 5.49 ; p < 0. 01), treatment group and phenotype ( F (6 72 ) = 3.31 ; p < 0. 01), and day and phenotype ( F (9 216 ) = 6.39 ; p < 0. 001), as well as an overall interaction effect ( F (18 216 ) = 2.44 ; p < 0. 01). The mean ( S.E.M.) percent of PI + cells significantly increased following exposure to OGD ( p < 0.001) and on Days 0 14 after reperfusion ( all p values < 0.001). The percent of PI + NeuN labeled mature neurons was higher (~19%) than the percent PI + tubulin labeled immature neurons (~13%; p < 0.01), GFAP + astrocytes (~6 %; p < 0.01) and NG2 + oligo dend rocyte precursors (~11 %; p < 0.01). In fact, astrocytes had the lowest proportion of dead/dying cells (all p < 0.05). The conditioned media and HDFs added treatment groups had significantly higher PI + mature neurons one day after OGD (all p < 0.001), however, the percent of PI + mature neurons in the NPCs added group was not statistically higher than the conditioned media non OGD control group. However, by one week, the conditioned media OGD group was not different from non OGD controls, the HDFs added OGD group had more dead/dying mature neurons, and the percent PI + NeuN labeled cells approached significance in the NPCs added group ( p = 0.072). One day following OGD, the conditioned media OGD group and HDFs added had significantly higher proportions of PI + immature neurons and astrocytes ( < 0.001), but by 7d was not significantly different from controls. On the other hand, the percent PI + immature neurons in the NPCs added OGD did not differ from the conditioned media control at 1d 14d and 28d,
109 but approached significance at 7d following OGD ( p = 0.072). In addition, the percent of PI + astrocytes in the NPCs added group did not differ from conditioned media controls. The conditioned media OGD group and HDFs added OGD group had significantly higher pr oportions of PI + oligodendrocyte precursors (all p < 0.001), but by 14d were not significantly different from controls. The percent PI + NG2 labeled cells in NPCs added OGD group did not differ from controls at 1, 14 and 28, but was significantly different at 7d ( p < 0.001). These data suggest NPC addition ameliorated some cell death, at least temporarily.
110 Figure 4 3. Cell death following OGD is less severe with NPC addition. Cells were stained with propidium iodide (PI) and subsequently fixed immediatel y following 3h of OGD or after total fresh media change (control), or 1d, 7d, 14d, or 28d later. (A) The percent PI + cells expressing different phenotype markers was quantified (n = 4), and revealed that NPC addition may protect mature neurons, and glia, a t least after 1d. (B E) Confocal images display PI + nuclei (red) of dead/dying cells 1d following OGD co expressing tubulin (cyan; B ), NeuN (cyan; C ), GFAP (cyan; D ) and NG2 (cyan; E ). Note that all cell nuclei are DAPI labeled (grey). Scale bar 50 m. p <0. 1, p <0.05, and *** p <0.001 versus conditioned media control; BOLD p <0.05 versus conditioned media OGD
111 OGD Increases Cu/Zn SOD in Neurons Immediately, and in Astrocytes After 14 Days When the percent of Cu/Zn SOD + NeuN tubulin GFAP and NG2 labeled cells was quantified to detect a cell type that may react to OGD with Cu/Zn SOD more than others ( Figure 4 4A), there was an effect of phenotype ( F (3 72 ) = 87.6 ; p < 0. 001), an interaction of condition and phenotype ( F (3 72 ) = 2.76 ; p < 0. 05), treatment and phenotype ( F (6 72 ) = 6.15 ; p < 0. 001), and day and phenotype approached significance ( F (9 216 ) = 1.72 ; p = 0. 086), as well as an overall interaction effect of all four independent variables ( F (18 216 ) = 1.82 ; p < 0. 05). After 1d, the conditioned media OGD group had a greater number of SOD + mature and immature neurons (all p < 0.001) that subsided by 7 28d ( n.s. ). On the other hand, there was a greater percent of SOD + astrocytes at 7d and 14d (all p < 0.01). The percent SO D + cells in the HDFs added and NPCs added OGD groups generally did not differ from the conditioned media control group, except for more Cu/Zn SOD + astrocytes at 7d (all p < 0.001).
112 Figure 4 4. Oxidative stress increases in neurons following OGD, but is not changed by NPC addition. (A) Cells were fixed and stained for Cu/Zn SOD, which is expressed in cells as a response to oxidative stress, immediately following (0d) 3h of OGD or after total fresh media change (control), or 1d, 7d, 14d, or 28d later (n = 4). The %Cu/Zn SOD + cells expressing different phenotype did not change in response to NPC addition. (B E) Confocal images display Cu/Zn SOD + (red) cells 1d following OGD co expressing immature neuronal marker tubulin (cyan; B ), mature neuronal marker NeuN (cyan; C ), astrocyte marker GFAP (cyan; D ) and oligodendrocyte precursor marker NG2 (cyan; E ). Note that all cell nuclei are DAPI labeled (grey). Scale bar 50 m. p <0. 1, ** p <0.01, and *** p <0.001 versus conditioned media control; BOLD p <0.05 versus conditioned media OGD
113 NPC Addition After OGD Protects Neural Activity and May Lead to New Activity In order to characterize recovered activity following OGD, the type of activity detected in groups following OGD was quantified as Recovered (active elect rodes at baseline that remained active after OGD), Transiently New (inactive electrodes at baseline that became active for 1 day after OGD), and Lasting New (inactive electrodes at baseline that became active for >1 day after OGD; Figure 4 5A B). The proport ions of the different types of recovery changed with treatment group ( F (2 18 ) = 4.25 ; p < 0. 05), and there was a preferred type of recovery ( F (2 18 ) = 258 ; p < 0. 001), but there was no significant interaction effect ( F (4 26 ) = 1.19 ; n.s. ). All groups had s imilar numbers of transiently new electrodes ( n.s. ), which was the preferred type of recovery compared to Recovered and Lasting New ( all p < 0.001). Although Lasting New activity appeared in the NPCs added group, it appeared on only a few electrodes (n.s.) Interestingly, the NPCs added group had significantly more recovered activity (all p < 0.05). NPC Addition Leads to Partial Recovery of Network Activity Following OGD The electrophysiological activity of primary cultures plated on MEAs was recorded for 10 min every 3 4d following OGD and subsequent cell addition, to record any recovery of activity following OGD ( Figure 4 6) Mean (S.E.M.) spike rates (Hz) changed with condition ( F (1 11 ) = 18.8 ; p < 0. 01), and there was an overall interaction effect of con dition, treatment and day ( F (14 84 ) = 1.85 ; p < 0. 05). Spike rates in the OGD groups were significantly lower than the control groups ( p < 0.01). Compared to the conditioned media control group, spike rates recovered in the NPCs added OGD group (days 1; p = 0.062; days7 14, n.s. ), but recovery did not last (day 28; p < 0.05).
114 Figure 4 5. NPC addition may lead to new activity (A) MEA context plots show the added spikeforms of detected spikes on the electrodes respective to the arrangement on the array. Th e example shown is 20s from an NPCs added OGD MEA before OGD at baseline day 38, 14d post OGD/NPC addition and 24d post OGD/NPC addition. Below are the respective raster plots of the spikes detected at all 59 electrodes (y axis) across time (x axis). The c olored boxes highlight examples of electrodes and corresponding spikes on the rasterplots that was categorized as Recovered (red; active electrodes at baseline that remained active after OGD), Transient New (green; inactive electrodes at baseline that beca me active for 1 day after OGD), and Lasting
115 New (blue; inactive electrodes at baseline that became active for >1 day after OGD). (B) All groups had similar numbers of transiently new electrodes, but few electrodes recovered exclusively in the NPCs added gr oup (n = 3). However, the NPCs added group had significantly more lasting new activity. Reflected light fluorescent images merge with bright field images show (C) GFP expressing NPCs, and (D) m tangerine expressing HDFs after 1d in co culture with primary neural cells plated on MEAs. The conditioned media or HDFs added OGD groups did not show recovery (all p < 0.05; Figure 4 6A). The mean (S.E.M.) percent active electrodes had a significant effect of condi tion ( F (1 11 ) = 83.1 ; p < 0. 001), and day ( F (7 84 ) = 4.14 ; p = 0. 001), but not with treatment ( F (2 11 ) = 0.508 ; n.s. ). However, there was an overall interaction effect ( F (14 84 ) = 2.21 ; p < 0. 05). There were significantly fewer active electrodes in the OG D groups (all p < 0.001). However, the NPCs added OGD group had significantly more active electrodes compared to the conditioned media OGD group (days 1 14, all p < 0.05; day 28, p = 0.099; Figure 4 6B). Network bursts, on the other hand, had a significant e ffect of OGD ( F (1 11 ) = 88.1 ; p < 0. 001) but did not have an overall interaction effect of condition, treatment and day ( F (14 84 ) = 0.49 ; n.s. ). Although bursts were observed in the NPCs added group after OGD, the rates were small and variable and did not differ from the conditioned media or HDFs added groups ( Figure 4 6C). Mean (S.E.M.) burst duration significantly changed with condition ( F (1 11 ) = 6.92 ; p < 0. 05), but not with day ( F (7,84 ) = 0.81 ; n.s. ), or with treatment ( F (2 11 ) = 2.13 ; n.s. ). However, t here was an overall interaction effect ( F (14 84 ) = 2.29 ; p < 0. 01). Burst duration peaked in the NPC added OGD group after 4 8d (all p < 0.10) in co culture and then plateaued by day 16 ( Figure 4 6D).
116 Figure 4 6. NPC addition after OGD leads to partial re covery of neural activity The electrophysiological activity of primary cultures plated on MEAs was recorded for 10min every 4d following OGD and subsequent cell addition 1d later, to record any recovery of activity following OGD (n = 3). ( A ) Mean (S.E.M.) spike rates (Hz) recovered in the NPCs added OGD group, but not in the conditioned media or HDFs added OGD groups after 8d in co culture. (B) The mean (S.E.M.) percent active electrodes show that activity recovered in the NPCs added OGD gro up, but not in the conditioned media or HDFs added groups. The number of active electrodes was not different in the NPCs added OGD group compared to the conditioned media OGD group after 24d. (C) Network bursts also recovered in the NPCs added group, howev er, the rates were small and variable and did not differ from the conditioned media or HDFs added groups. (D) Mean (S.E.M.) burst duration peaked in the NPC added OGD group after 4 8d in co culture and then plateaued by day 16. p <0.05 NPC OGD versus CM c ontrol; p <0.05 NPC OGD versus CM OGD
117 Discussion In this study, we demonstrated that MEA plated primary neurons exposed to OGD could be an effective platform for investigating the effec ts of OGD and potentially protective/regenerative effects of adult NPCs on mature neural netwo rks. As expected, OGD produced significant cell death and oxidative stress that persisted for a week after OGD exposure. Surprisingly, we noted a latent (2 4 weeks post injury) increase in the proportion of primary cells that exp ressed the immature neuronal marker tubulin, which could reflect injury induced neurogenesis or perhaps even de differentiation. Primary cells in OGD exposed cultures that were supplemented with NPCs exhibited low level proliferation as well as decreased death among mature neurons and oxidative stress among all cells, at least in the first week after injury. Importantly the post OGD milieu of these primary cultures did not overtly affect the fate of NPCs or their progeny. Network activity in electrophysi ologically mature primary cultures was silenced immediately by exposure to OGD. In fresh media control cultures, NPCs but not NPC conditioned media or HDFs stimulated a developmental pattern of superbursting activity ( Chapter 2 ). In OGD exposed culture s NPCs, but not NPC conditioned media or HDFs partially protected single unit and network activity for ~ 2 weeks, which may be significant in a clinical setting where combined approaches are employed. Our observations that NPC conditioned media neither affe cts the activity of control or OGD exposed cultures and that novel activity emerges in both control and OGD exposed primary cultures after NPC supplementation suggests that NPC progeny are capable of integrating functionally into established neural circuit s potentially after injury. In vivo and in vitro ischemic stroke can kill cells in a number of ways. Signs of necrotic cell death include aberrant Ca 2+ and Na + homeostasis, mitochondrial and
118 organelle swelling, ruptured plasma membranes and randomly degra ded DNA (Walker et al., 1988) Signs of apoptotic cell death include condensed cytoplasm, fusing of the endoplasmic reticulum with the plasma membrane and nuclear material that becomes ( Wyllie et al., 1981) Because PI and other cell death markers cannot differentiate between necrotic and apoptotic cell death (Banasiak et al., 2000) we quantified pyknotic (condensed and/or fragmented) nuclei as a secondary measure of apoptosis following O GD. Overall, OGD exposure increased the percent of PI + dead cells and pyknotic dying cells in the 2 weeks after exposure. NPC addition decreased the percent PI + cells but did not impact the percent of pyknotic cells. While these data tempt speculation that NPCs may protect neurons from the primary necrotic cell death typically observed immediately after stroke but not the secondary apoptotic cell death typically observed later, more experiments are required to reveal the mechanisms by which NPCs promote neu ronal survival after stroke. Our result do concur however, with the work of others showing that NPCs increase cell survival following ischemia in vivo (Harms et al., 2010) Among cells that survived OGD exposure, a large proportion expressed Cu/Zn SOD, which is a marker of cell stress that blocks the cytosolic release of cytochrome C ultimately reducing apoptosis after ischemia (Fujimura et al., 2000) The percent of Cu/Zn S OD cells was significantly higher 1 14d following OGD, peaking 1d after treatment. This coincides with reports that free radical production peaks after reperfusion (Perez Velazquez et al., 1997) Interestingly, the percent of neurons in OGD exposed culture s supplemented with either NPCs or fibroblasts (but not NPC conditioned media) remained at control levels. Although SOD up regulation is a sign of
119 oxidative stress, it is also neuroprotective (Fujimura et al., 2000) and interestingly remains upregulated in astrocytes following OGD exposure regardless of cell supplementation. For example, transgenic mice that overexpression Cu/Zn SOD had smaller infarcts following ischemic stroke (Kinouchi et al., 1991) Therefore, perhaps astrocytes are less vulnerable to O GD because they primarily produce Cu/Zn SOD. Although an increase in Cu/Zn SOD after OGD further demonstrates that OGD in cultured cells reflects what is seen in situ, NPCs do not appear to be protective by affecting Cu/Zn SOD directly. Lower Cu/Zn SOD in NPC added groups may indicate that primary cells are facing less oxidative stress, perhaps either by NPC progeny blocking excitoxicity, or offering trophic support. Mitotic phase nuclei among primary cells were also quantified in order to estimate prolife ration. Although we detected mitosis in less than 1% of all cells, OGD exposure increased the percent of mitotic nuclei particularly after NPCs (but not NPC conditioned media or fibroblasts) were added to the cultures. Interestingly the impact of transplan table cells on the behavior of endogenous NPCs may be more robust than in vitro NPCs are resident to the neurogenic hippocampal dentate gyrus and SVZ (Liu et al., 1998; Takagi et al., 1999; Kee et al., 2001; Yagita et al., 2001) New n eurons are even generated in non neurogenic regions following ischemic injury but their survival is transient (Arvidsson et al., 2002; Thored et al., 2006) Transplanted NPCs may in fact stimulate the local production of factors that promote the proliferat ion, differentiation and survival of endogenous NPCs, such as EGF, FGF, GDNF and PSA NCAM (Craig et al., 1996; Kuhn et al., 1997; Kitagawa et al., 1998; Iwai et al., 2001)
120 Understanding whether the post injury or disease milieu that cells will be transpl anted into affects their behavior may be critical for the success of stem/progenitor cell strategies for human CNS syndromes. In both OGD and fresh media exposed cultures we found that ~3 5% of all cells were GFP + (and therefore of NPC origin) 1d after pl ating. The proportion of NPCs increased similarly over time, regardless of condition, suggesting that NPCs proliferated. We did however, detect a slight but significant decrease in the percent of GFP + cells in OGD versus fresh media control exposed co cul tures at 28d, which could represent a post stroke milieu effect on the speed with which NPC progeny divided and/or differentiated, or even an increase in primary cell division. Although the surrounding environment can influence NPC survival and fate (Ormer od et al., 2008) the percent of NPC derived (GFP + ) cells acquiring neuronal or glial phenotypes quantified at 28 was unaffected by OGD, suggesting that there was no significant effect on their fates. In fact, the proportions of 4 week old NPC progeny dete cted in OGD and fresh media exposed co cultures in the current study were remarkably similar to the proportions of phenotypes of 4 week old co cultured NPC progeny (derived from a different cell line) in our previous work ( Chapter 2 ). Together these d ata suggest that transplantable adult NPCs may be fairly resilient to the effects of post injury milieus. Both single unit and network activity was rapidly silenced by exposure to OGD. In addition to protecting primary cells and particularly neurons from OGD induced cell death and oxidative stress, we found that NPCs also partially but transiently protected neurons from the effects of OGD induced silencing. In control cultures, we found that NPCs, but not fibroblasts or NPC conditioned media, stimulated a developmental
121 pattern of superbursting activity consistent with our previous work ( Chapter 2 ). In NPC supplemented cultures, we also detected novel activity over time suggesting that new NPC generated neurons are capable of integrating functionally in to existing neuronal circuitry. This evidence is further supported by our finding that NPC conditioned media neither stimulates superbursting activity nor the emergence of novel activity in electrophysiologically mature cultures. The most exciting data sho wed that although activity never returned to control levels in OGD exposed cultures following NPC supplementation they did exhibit activity for several weeks after the insult. Spike rates on active electrodes almost retained control levels in OGD exposed cultures supplemented with NPCs but not with fibroblasts or NPC conditioned media until the 4 th week after OGD and we did detect more active electrodes among OGD exposed cultures supplemented with NPCs versus NPC conditioned media. The ability of NPCs to protect network activity was less pronounced than their ability to protect single unit activity. For example, more network bursts were observed in OGD exposed cultures supplemented with NPCs versus either fibroblasts or NPC conditioned media but the rates were small and variable and did not recover to those observed in control cultures. Although OGD exposed cultures supplemented with NPCs (but not fibroblasts or NPC conditioned media) exhibited burst durations that peaked 4 8d and plateaued ~16d after OGD (resembling the evolution of burst durations in developing primary cultures) they did not maintain bursting activity much longer than 2 3 weeks after the injury. Discovering how to prolong this effect may be critical for increasing the efficacy of transpla nt, or delaying cell death until other treatments can be applied.
122 All groups had similar numbers of transiently new electrodes, but few electrodes had lasting new activity and were not significantly different from zero. However, the NPCs added group had si gnificantly more recovered activity. Neural stem cell grafts into ischemic CA1 lesions in the rat hippocampus exhibit pre and post synaptic markers and have been shown to develop properties of mature, normal hippocampal neurons within the infarct (Aoki et al., 1993) Indeed, some f unctional improvement has been reported in a nimal models of stroke following NPC transplantation despite variable rates of survival among the transplanted cells and their progeny (Nikkhah et al., 1994a; Hodges et al., 1996; Sinde n et al., 1997; Modo et al., 2002; Dziewczapolski et al., 2003; Ben Hur et al., 2004; Gao et al., 2006) Hypotheses regarding the mechanisms of functional improvement in these studies have included that transplanted cells and/or their progeny provide trophic support that thwarts further cell death, stimulate an endogenous repair process or the production/secretion of neurotransmitter, or perhaps integrate into existing circuitry to amplify compromised signaling (Ben Hur et al., 2004; Ormerod et al., 20 08) Neural stem cells could also induce plasticity by secreting factors such as VEGF and Apolipoprotein E (Andres et al.; Modo et al., 2003; Lee et al., 2007) The se results demonstrate that the MEA OGD co culture model provides a n effective benchmark for investigating the efficacy of potential therapies on the damage caused by ischemic stroke because electrophysiological and immunocytochemical analyses resemble ischemic stroke pathology described in situ Future research could benefit by utilizing thi s model by applying treatments and cells following OGD, and monitoring MEA recordings for functional recovery of individual neurons, as well as network bursts.
123 CHAPTER 5 GENERAL DISCUSSION Summary Treatments or even cures for CNS injury or neurodegenerati ve disease could be based on NPC transplantation to replace dead or damaged tissue. Recent breakthroughs in stem cell research suggest that NPC derived cells can produce CNS cell subtypes, including mature neurons which may synapse with target areas to af fect their function (Aoki et al., 1993; Stephens et al., 2012) Experiments have shown that ESC derived NPC progeny readily form de novo neural networks, and f unctionally integrate into existing host circuitry (Copi et al., 2005; Lepski et al., 2011; Tonne sen et al., 2011) More recent data, including the data contained in this dissertation, suggest that adult tissue derived NPC progeny may not form networks independently, but may also integrate functionally into existing neural networks (Chapters 2 and 4) demonstrated by stimulating plasticity among intact neural circuits (Chapter 2), and protecting cells and preserving activity among networks compromised by injury (Chapter Cha pter 4 shows that transplanted NPCs are surprisingly unaffected by the post stroke milieu. Nevertheless any effects th at the addition of NPCs may have on existing neural activity and in turn, any effects existing tissue may have on progenitor cell fate are questions that must be addressed before effective treatments can be developed The work in this dissertation began by asking whether adult NPCs form de novo neural networks. Based on current knowledge of neural stem cell biology adult NPCs in the right conditions will hypothetically form glia and mature neurons with functional synapses that are capable of integrating into the surrounding network as they do in the
124 daily turnover of new neurons in the adult hippocampus and olfactory bulb (Zhao et al., 2008) Although the exact function of these new cells remains unknown, endogenous NPC derived hippocampal neurons exhibit both GABAergic and glutamatergic synapses within approximately four weeks (Esposito et al., 2005) and show some evidence of spontaneous activity (van Praag et al., 2002) Before investigating how adult derived NPCs affect existing activity of mature primary neurons, it was important to first examine whether adult NPC derived neurons were capable of integrating functionally with one another by testing their ability to develop spontaneous activity and network bursting independently. The capability of NPC derived neurons to autonomously form networks would be highly desirable for prospective treatments seeking to replace dead circu itry. However, extraneous novel circuits or heavily modified existing circuits may even compromise residual function. For example, cognitive deficits are often reported in animal models of epilepsy that are associated with drastically elevated rates of ect opic neurogenesis (Parent et al., 1997) If adult derived NPCs possess a unique ability to augment existing neural networks without being able to form new networks of their own, adult NPCs may be perfectly adept for reparative therapies seeking to amplify, strengthen or reinforce damaged neural circuitry. MEAs were previously used to test spontaneous networking ability in ESC derived cultures (Illes et al., 2007) ; however, the networking capability of purely NPC derived cells was put into question and teste d on MEAs for the first time in Chapter 2 of this dissertation. Similar to ESC derived neural cultures, NPC development wa s hypothesized to be slightly delayed and develop on a much smaller scale compared to primary neural populations. As expected, NPCs p lated alone on MEAs in normal culture
125 conditions displayed delayed development of spontaneous action potentials, as well as delayed expression of mature neural and glial markers. Interestingly, NPC progeny grown alone in culture did not produce synchronous network wide bursts, even after eight weeks in normal culture conditions. This lack of network capability of adult NPCs may or may not be useful in treatment strategies, depending on its purpose and target. Adult NPCs could be utilized when the goal of tr ansplant is to augment existing neural activity, rather than replace the entire circuit. However, electrical stimulation and/or manipulation of the microenviro nment with external factors may be necessary to promote network formation (Piatti et al., 2011) For example, NPC progeny matured faster in co culture than when grown alone (Chapter 2). Perhaps optimizing culture conditions and techniques to direct differentiation could change the future of NPC applications. Although several groups have linked the con tribution of new neurons in the hippocampus to learning and memory (Gould et al., 1999; Kempermann and Gage, 2002; Ormerod et al., 2004; Winocur et al., 2006) a direct connection between functional integration of new neurons and its affect on existing cir cuitry remains unclear. Chapter 2 of this dissertation took advantage of well characterized, stable activity in primary rat cortical cultures, and used it as a benchmark to directly observe how NPCs influence established activity. After two weeks in co cul ture, as NPCs began to express mature neural and glial markers, network burst patterns reverted to superbursts, to reflect those observed during network development in vitro and in various areas in the developing cortex The appearance of superbursts, indi cative of induced plasticity, occurred approximately two weeks after addition at Day 40, and again after an addition
126 at Day 64. Superbursting was also observed in NPC supplemented, non OGD control cultures during the experiments described in Chapter 4 usin g a completely different NPC line. Highly excitable, nave young neurons in primary cultures hypothetically drive transient superbursts when networks are developing at 4 5 weeks in primary neural cultures, and again by added NPC generated young neurons 1 2 weeks after NPC addition. This is an exciting discovery that demonstrates NPC addition does influence neural activity that may be favorable in some therapeutic strategies where the existing neural network is damaged and needs remodeling or enhancement, bu t may not be favorable in other strategies where disrupting existing circuitry could impede normal brain functions. The stroke penumbra, consisting of surviving but silenced neural circuits, is an excellent example of post injury environment where improvem ent to function could be made by stimulating plasticity, perhaps without generating novel networks. Next, the co culture MEA model of NPC transplant was taken a step further, and applied for the first time to a compromised neural network, by modeling hypo xic ischemic stroke with oxygen glucose deprivation (Chapter 3). Recording primary neural cultures during OGD is a novel application of MEA technology, and gives an unprecedented view of the massive depolarization and network degradation of spontaneous act ivity that has only previously been recorded from a handful of electrodes placed in vivo (Kass and Lipton, 1982; Rosner et al., 1986) The behavior of network activity in primary cultures closely mirrors what is seen in situ, but data is collected from 59 sites simultaneously. This model adds a new, comprehensive view of activity during ischemia that encompasses neuron behavior on a network scale, adding
127 to the picture of stroke dynamics in culture and in vivo. Spike rates initially increased during the fir st five minutes of OGD due to the emergence of asynchronous activity between bursts. Action potentials decreased in amplitude, and all activity ultimately ceased by approximately sixteen minutes. Neural activity recovered following reperfusion after 20 min ute OGD, but little recovery was observed after longer durations. Shorter duration OGD may effectively model the silenced circuits found in the penumbra of a stroke, while longer OGD might effectively model the dying circuits found with the stroke core. By defining the parameters needed to silence neuronal activity but maintain viability, this work has pioneered a culture model of a stroke penumbra versus stroke core. Each substrate can be utilized in the future to develop approaches to treat both areas. Ov erall, OGD exposure increased the percent of PI + dead cells and pyknotic dying cells two weeks after exposure. NPC addition decreased the percent PI + cells but did not impact the percent of pyknotic cells. These results may point to NPCs as protecting neurons from the primary necrotic cell death typically observed immediately after stroke, but not the secondary apoptotic cell death typically observed later. However, more experiments are required to r eveal the mechanisms by which NPCs promote neuronal survival after stroke. Although we detected mitosis in less than 1% of all cells, OGD exposure increased the percent of mitotic nuclei, particularly after NPCs were added to the cultures. Interestingly, t he impact of transplantable cells on the behavior of in vitro In addition, co culture in OGD compromised primary cultures did not affect the rate of NPC differentiation or phenotype relative to t hose in co culture with intact primary cells. We
128 also detected novel activity over time in NPC supplemented cultures, suggesting that new NPC generated neurons are capable of becoming functional in existing neuronal circuitry. This evidence is further supp orted by our finding that NPC conditioned media neither stimulates superbursting activity nor the emergence of novel activity in electrophysiologically mature cultures. The most exciting data showed that although activity never returned to control levels in OGD exposed cultures following NPC supplementation, they did exhibit activity for several weeks after the insult. These results demonstrated that this MEA co culture model could be an effective bioassay for potential neuroregenerative therapies on exist ing network activity. An in vitro assay system may help the translation of NPC transplant strategies to clinic by investigating these effects in a highly sensitive, biologically relevant neuronal network. Future Work Additional studies that build on the M EA based co culture system may reveal that NPCs have the capability to form de novo networks on their own, given the correct conditions. In Chapter 2, NPC derived cultures did not produce many glia, however, a certain number of glial cells are necessary fo r synapse formation and network development (Johnson et al., 2007; Kucukdereli et al., 2011) Future experiments to optimize different media conditions that promote heterogenous cell fate, through chemical and/or electrical stimulation may potentiate the r egenerative capabilities of adult derived NPCs. Results in Chapter 2 and 4 suggest that NPCs functionally integrate, because they altered existing neural activity in primary cultures. Unfortunately, MEA technology lacks the ability to isolate single unit a ctivity and conclusively identify which neuron is the source of the particular activity in question.
129 Future experiments should look directly at if and how NPC derived neurons become functional, and by what mechanism(s) do they influence existing activity. Researchers have and are currently finding many factors that influence NPC fate and functional it y in the brain. When added to primary neural cultures, NPCs differentiate into mature neurons and glia at a faster rate; exhibiting neural and glial markers wit hin two weeks, versus six weeks when cultured alone (Chapter 2). NPC derived mature cells also display more mature morphology, in terms of more extensive branching, than when grown alone. Although NPCs did not exhibit de novo networking ability, perhaps th e co cultured NPC derived neurons were able to form synapses with primary cells and other NPC progeny, and participate in network bursting. Co culture with other stem or progenitor cells, such as oligodendrocyte progenitor cell lines, could yield the right number of astrocytes and/or oligodentrocytes for spontaneous network bursting activity to develop. Moreover changing the culture conditions by growth factor supplemented media and/or chronic electrical stimulation, which is known to affect NPCs and the d evelopment of neurons, may also direct NPC derived neuron s to form network bursts (Tateno et al., 2005; Wagenaar et al., 2005) MEAs have the capability to stimulate, as well as record, activity. An experiment that applies chronic stimulation of NPC derive d cultures for days or weeks after plating may induce network formation. Whole cell patch clamp is often paired with MEAs (Pancrazio et al., 2001; O'Shaughnessy et al., 2003; Meyer et al., 2008; Chiappalone et al., 2009) because it is low throughput, but gives refined, cell specific information, while MEAs give more global data of the impact on neuronal networks. I would have liked to pair whole cell patch clamp methods with my MEA co culture experiment in order to determine if and when
130 NPC derived neuron s became functional, and if participated in network bursts. Whole cell patch clamp under current clamp mode can record spontaneous activity of an NPC derived neuron and would determine if NPC derived neurons become spontaneously active and participate in n etwork bursts. Furthermore, voltage clamp recordings could reveal the fundamental mechanisms underlying NPC functional integration. This analysis should be performed to investigate whether the changes observed in network activity occurred directly because of functional integration of NPC progeny, or indirectly, by secreted factors to induce neurogenesis and/or plasticity within the primary neural cultures. The latter seems unlikely because the conditioned media groups yielded no change in activity. However, patch clamp techniques paired with MEA analysis could provide direct evidence of NPC derived neuron and primary cell functional interactions. Live cell imaging is another method of providing evidence of the functional integration of NPC derived neurons th at could be paired with MEA recordings. Synaptic stimuli produce Ca 2+ elevations that are largely restricted to dendritic spines, where activated synapses terminate. The intracellular Ca 2+ increase is due to, for the most part, NMDA receptor activation by glutamate. Live cell imaging can report transient changes in neuronal cytosolic Ca 2+ signaling to show the release of glutamate from individual synaptic terminals I n the MEA plated cultures, live Ca 2+ imaging could s how network bursting, and the participation of individual neurons in network bursts. Specifically, a red shifted Ca 2+ indicator, such as rhod 3 (Molecular Probes; Eugene, OR), would be ideal for imaging in GFP + NPC derived neurons. A non fluorescent Rhod 3 is passively loaded into cell s in a wash, where intracellular esterases cleave the dye to a cell impermeant, active form that fluoresces upon Ca 2+ binding (Ieda et al.,
131 2010) MEAs can be imaged on a reflected light fluorescent microscope, and bursts can be video recorded with the Rho d 3. T he intensity of Rhod 3 in specific cells can be plotted as spikes for comparison to the MEA recordings. Improve ments to the translational potential of the variable duration OGD model of stroke could also be made with future experiments For example, a time series experiment of OGD exposure with 5 10 minute intervals could refine the 20 60 minute, silenced penumbra model versus the 2 3 hour, stroke core model. In addition, a time series on NPC supplementation could investigate immediate versus delayed neuroprotective effects of adding NPCs after hours or days or reperfusion. In addition, pairing patch clamp experiments with future applications of the ischemic stroke model could also provide valuable information about whether NPCs become active, are the cause of some or all of the activity recovered after OGD, and whether the activity is due to rescued primary neurons, or even new neurons generated by primary cells. Endogenous NPCs in the primary cultures could be activated by factors secreted by the add ed NPCs. The mere presence of endogenous NPCs could be ruled out by the application of an anti mitotic, or irradiation after plating. I would also like to delve further into how NPCs rescue mature neurons from OGD induced death and silencing. Future experi ments pairing patch clamp with the MEA and/or stimulation of the cultures, may detect the mechanisms behind the survival of primary neurons, and define possible treatment s trategies for silencing and regeneration. Lastly, combining the MEA co culture tran splantation model with in vivo studies designed to measure effects of NPC transplant on intact brain circuitry is essential to moving any effective treatments to clinic. The work for my dissertation developed a
132 simplified in vitro model of NPC transplantat ion that is easily accessible, relatively inexpensive, does not require the use of animals, is highly sensitive to manipulation, and can be paired with extensive immunocytochemical analysis to monitor changes in a biologically relevant neural network. Howe ver, it is merely an assay for testing the efficacy of NPC transplant or other strategies for neural repair without confounding factors found in vivo Future researchers should consider combining the MEA based co culture model when developing a translation al approach for scaling up therapies found to be beneficial through this assay. Conclusions Although some NPC transplantation experiments have yielded successful graft survival, and even limited behavioral recovery (Jgamadze et al., 2012) researchers are not yet capable of manipulating NPC derived neurons and glia to control function al integration and effectively treat the injured or diseased CNS. Some may stand to argue that studying treatment strategies in vitro is impractical, because any solution that is found by this assay would probably not work when translated in vivo Of course, many variables are excluded in the in vitro model that could drastically alter the fate and functionality of NPCs and their progen y in vivo (i.e. horm ones, inflammation, vasculature; Dua n et al., 2008; Mu et al., 2012 ). However, such a highly controllable environment is crucial to the scientific method, and it would be a disadvantage to stem cell research to underestimate or disregar d its utility. As quickly as investigative technology is advancing, researchers are persistently discovering new ways to regulate NPC behavior (Kim et al., 2012) As strategies are conceived in vitro additional variables can be introduced to gradually eng ineer an in vivo strategy that preserves the
133 In conclusion, the MEA co culture model of NPC transplantation established through this dissertation research can help understand part of the basic mechanisms that accompany the functional i ntegration of transplanted neural cells. My work has contributed to the need for further knowledge of how to engineer connections to target neural networks, or induce plasticity in existing neural networks, to prevent unintentional consequences, such as up setting normal brain function or causing epilepsy. Therefore my findings, together with future extensions of this work will hopefully help build a foundation for establishing transplant strategies for neural repair. This work developed an in vitro model of NPC transplantation that is easily accessible in terms o f monitoring cells optically and electrophysiologically on a network wide scale across thousands of cells that can be paired with extensive immunocytochemical analysis to provide an assay for testing the efficacy of NPC transplant or other strategies for neural repair without confounding factors found in vivo This model gives insight into the networking ability of cells by plating NPCs alone, and the impact of cell engraftment on existing neural circ uitry by way of co culture. Furthermore, this model demonstrated its utility by the adding the application of hypoxic ischemic conditions to simulate a post stroke environment. This work, along with future extensions of this research, will establish a foun dation for testing customized strategies for neural repair before introducing them into much more complex in vivo environments, and ultimately to the clinic.
134 APPENDIX: CHAPTER 2 SUPPLEMENT ARY DATA NPC media optimization for co culture First, we confirmed that NPC cultures would survive long term in media formulations typical of MEA experiments with primary cultures that predictably develop mature bursting activity. A 100l droplet of each of eight media formulations conta ining 5,000 NPCs was placed in the center of PEI/laminin coated wells in a 12 well culture dish (to model the surface area of an MEA dish), and the cells were grown for 21 days (n = 3 wells per media formulation). The eight media formulations chosen were: (1) NPC differentiation media prepared with GlutaMax supplemented DMEM, (2) NPC differentiation media prepared with high glucose DMEM/F12, (3) N2 supplemented NPC differentiation media prepared with GlutaMax supplemented DMEM, (4) N2 supplemented NPC diffe rentiation media prepared with high glucose DMEM/F12, (5) primary culture growth media (10% ES) prepared in GlutaMax DMEM, and (6) primary culture growth media (10%) prepared in high glucose DMEM/F12, (7) 10% fetal bovine serum in Glutamax supplemented DM EM and (8) 10% fetal bovine serum in high glucose DMEM/F12. On Day 21, propidium iodide (250 g/ml) was added to each well for 15min to label the DNA of non viable cells before the cells were paraformaldehyde fixed and DAPI stained. Densities of live (GFP + ) and dead (PI/GFP + ) cells were counted on images taken using a Zeiss AxioObserver microscope (20x objective). We found that media formulation ( F ( 3 16 ) = 10 10 p = 0.00 05 ) and GlutaMax supplementation ( F ( 1 16 ) = 8 96 p = 0.00 8) significantly impacted cell death. Cell death was significantly lower among NPC cultures grown in primary culture growth media (10% equine serum; 17.33 3.26%), relative to 10% fetal bovine serum in DMEM (29.17 5.16%; p = 0.022),
135 N2 supplemented NPC differentiation media (32.50 3.63% ; p = 0.013) or NPC differentiation media (42.67 3.60%; p = 0.0004). Cell death was highest in media without N2 supplementation relative to all other formulations ( p Overall GlutaMax supplemented DMEM slightly, but significantly increased cell death relative to glutamine supplementation (25.5 3.87% versus 35.33 3.15%; p < 0.010). However, we chose to use GlutaMax supplemented primary culture growth media for our initial co culture experiments because it only slightly increased cell death (23.00 3.78%) and promotes the reliable development of electrophysiological activity among neurons in primary cultures. NPC plating density optimization Next we tested plating densities that would permit the long term culture of proliferative adult NPCs at cell densities that produced consistent monolayer cultures. A placed into the center of a PEI/laminin coated well in a 12 well tissue culture plate (n = 3 wells per condition) a nd the cells grown for 21 days. Every three days, ~50% of the media was replaced to model the feeding regime used in MEA experiments. Cell densities were quantified at Days 2, 7, 14 and 21 after plating, using digital phase contrast images (20x objective) taken in random well locations (n = 7 locations in 3 wells per plating density). As expected, cell density was directly proportional to plating density two days after plating ( F (3,2 4 ) = 16.35, p < 0.0001; Newman Keuls p < 0.0 10; Mean S.E.M cells/mm 2 : 1K = 37.3 7.4; 5K = 313.4 29.3; 10K = 644.1 85.6; 15K = 706.7 124.2). Note that although some death occurs after plating, cells rapidly migrate out of the droplet zone reducing the cell densities observed at plating. At later time
136 points, cell densities were not affected by plating densities ( F (3,2 4 ) = 2 14 p = 0.12) but did change across day after plating ( F (2,48 ) = 31 55 p < 0.0001; interaction effect : F (6,48 ) = 0 456 p = 0.837). Specifically, cell fold increases diminished from their mean peak of ~9.7 fold at Day 7 (1K = 12 fold; 5K = 12 fold; 10K = 8 fold and 15K = 8 fold) to ~ 1.3 fold on Day 14 ( p fold on Day 21 ( p employ the 10K NPCs/droplet concentration for further experiments because this concentration sufficiently covered the well surface without overwhelming the culture dish, produced cells with a healthy appearance, and would comprise ~5% of the co culture. Because we hypothesized that NPC derived cells co cultured with established neural networks may undergo less extensive proliferation, we also included a 50K NPC/droplet condition in our co culture experiment.
137 LIST OF REFERENCES Aimone JB, Wiles J, Gage FH (2006) Potential role for adult neurogenesis in the encoding of time in new memories. Nat Neurosci 9:723 727. Altman J, Das GD (1965) Post natal origin of microneurones in the rat brain. Nature 207:953 956. Andres RH Horie N, Slikker W, Keren Gill H, Zhan K, Sun G, Manley NC, Pereira MP, Sheikh LA, McMillan EL, Schaar BT, Svendsen CN, Bliss TM, Steinberg GK Human neural stem cells enhance structural plasticity and axonal transport in the ischaemic brain. Brain 134:17 77 1789. Aoki H, Onodera H, Yae T, Jian Z, Kogure K (1993) Neural grafting to ischemic CA1 lesions in the rat hippocampus: an autoradiographic study. Neuroscience 56:345 354. Arsenijevic Y, Villemure JG, Brunet JF, Bloch JJ, Deglon N, Kostic C, Zurn A, Aeb ischer P (2001) Isolation of multipotent neural precursors residing in the cortex of the adult human brain. Exp Neurol 170:48 62. Arvidsson A, Collin T, Kirik D, Kokaia Z, Lindvall O (2002) Neuronal replacement from endogenous precursors in the adult brain after stroke. Nat Med 8:963 970. Astrup J, Siesjo BK, Symon L (1981) Thresholds in cerebral ischemia the ischemic penumbra. Stroke 12:723 725. Babu H, Ramirez Rodriguez G, Fabel K, Bischofberger J, Kempermann G (2009) Synaptic Network Activity Induces N euronal Differentiation of Adult Hippocampal Precursor Cells through BDNF Signaling. Front Neurosci 3:49. Back SA, Han BH, Luo NL, Chricton CA, Xanthoudakis S, Tam J, Arvin KL, Holtzman DM (2002) Selective vulnerability of late oligodendrocyte progenitors to hypoxia ischemia. J Neurosci 22:455 463. Bain G, Kitchens D, Yao M, Huettner JE, Gottlieb DI (1995) Embryonic stem cells express neuronal properties in vitro. Dev Biol 168:342 357. Bambrick LL, Kostov Y, Rao G In vitro cell culture pO(2) is significantl y different from incubator pO(2). Biotechnol Prog. Ban J, Bonifazi P, Pinato G, Broccard FD, Studer L, Torre V, Ruaro ME (2007) Embryonic stem cell derived neurons form functional networks in vitro. Stem Cells 25:738 749. Banasiak KJ, Xia Y, Haddad GG (200 0) Mechanisms underlying hypoxia induced neuronal apoptosis. Prog Neurobiol 62:215 249.
138 Banker GA (1980) Trophic interactions between astroglial cells and hippocampal neurons in culture. Science 209:809 810. Ben Ari Y (2001) Developing networks play a simi lar melody. Trends Neurosci 24:353 360. Ben Ari Y, Cherubini E, Corradetti R, Gaiarsa JL (1989) Giant synaptic potentials in immature rat CA3 hippocampal neurones. J Physiol 416:303 325. Ben Hur T, Idelson M, Khaner H, Pera M, Reinhartz E, Itzik A, Reubino ff BE (2004) Transplantation of human embryonic stem cell derived neural progenitors improves behavioral deficit in Parkinsonian rats. Stem Cells 22:1246 1255. Benninger F, Beck H, Wernig M, Tucker KL, Brustle O, Scheffler B (2003) Functional integration of embryonic stem cell derived neurons in hippocampal slice cultures. J Neurosci 23:7075 7083. Berridge MJ (1998) Neuronal calcium signaling. Neuron 21:13 26. Boehler MD, Wheeler BC, Brewer GJ (2007) Added astroglia promote greater synapse density and high er activity in neuronal networks. Neuron Glia Biol 3:127 140. Brederlau A, Correia AS, Anisimov SV, Elmi M, Paul G, Roybon L, Morizane A, Bergquist F, Riebe I, Nannmark U, Carta M, Hanse E, Takahashi J, Sasai Y, Funa K, Brundin P, Eriksson PS, Li JY (2006) Transplantation of human embryonic stem cell derived cells to a rat model of Parkinson's disease: effect of in vitro differentiation on graft survival and teratoma formation. Stem Cells 24:1433 1440. Buffo A, Rite I, Tripathi P, Lepier A, Colak D, Horn AP Mori T, Gotz M (2008) Origin and progeny of reactive gliosis: A source of multipotent cells in the injured brain. Proc Natl Acad Sci U S A 105:3581 3586. Buzski G (2010) Neural syntax: cell assemblies, synapsembles, and readers. Neuron 68:362 385. Buzs ki G, Masliah E, Chen LS, Horvth Z, Terry R, Gage FH (1991) Hippocampal grafts into the intact brain induce epileptic patterns. Brain Res 554:30 37. Calabresi P, Marfia GA, Centonze D, Pisani A, Bernardi G (1999) Sodium influx plays a major role in the me mbrane depolarization induced by oxygen and glucose deprivation in rat striatal spiny neurons. Stroke 30:171 179. Cameron HA, Woolley CS, McEwen BS, Gould E (1993) Differentiation of newly born neurons and glia in the dentate gyrus of the adult rat. Neuros cience 56:337 344.
139 Chiappalone M, Casagrande S, Tedesco M, Valtorta F, Baldelli P, Martinoia S, Benfenati F (2009) Opposite changes in glutamatergic and GABAergic transmission underlie the diffuse hyperexcitability of synapsin I deficient cortical networks Cereb Cortex 19:1422 1439. Clarkson ED (2001) Fetal tissue transplantation for patients with Parkinson's disease: a database of published clinical results. Drugs Aging 18:773 785. Colucci D'Amato L, Bonavita V, di Porzio U (2006) The end of the central d ogma of neurobiology: stem cells and neurogenesis in adult CNS. Neurol Sci 27:266 270. 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 J Neurophysiol 94:4538 4543. Craig CG, Tropepe V, Morshead CM, Reynolds BA, Weiss S, van der Kooy D (1996) In vivo growth factor expansion of endogenous subependymal neural precursor cell populations in the adult mouse brain. J Neurosci 16:2649 2658. Cur tis MA, Kam M, Nannmark U, Anderson MF, Axell MZ, Wikkelso C, Holtas S, van Roon Mom WM, Bjork Eriksson T, Nordborg C, Frisen J, Dragunow M, Faull RL, Eriksson PS (2007) Human neuroblasts migrate to the olfactory bulb via a lateral ventricular extension. S cience 315:1243 1249. de Lima AD, Voigt T (1997) Identification of two distinct populations of gamma aminobutyric acidergic neurons in cultures of the rat cerebral cortex. J Comp Neurol 388:526 540. Dhamoon MS, Moon YP, Paik MC, Boden Albala B, Rundek T, S acco RL, Elkind MS (2010) Quality of life declines after first ischemic stroke. The Northern Manhattan Study. Neurology 75:328 334. Doetsch F, Garcia Verdugo JM, Alvarez Buylla A (1997) Cellular composition and three dimensional organization of the subvent ricular germinal zone in the adult mammalian brain. J Neurosci 17:5046 5061. Domercq M, Perez Samartin A, Aparicio D, Alberdi E, Pampliega O, Matute C P2X7 receptors mediate ischemic damage to oligodendrocytes. Glia 58:730 740. Duan X, Kang E, Liu CY, Ming GL, Song H (2008) Development of neural stem cell in the adult brain. Curr Opin Neurobiol 18:108 115. Dupont E, Hanganu IL, Kilb W, Hirsch S, Luhmann HJ (2006) Rapid developmental switch in the mechanisms driving early cortical columnar networks. Nature 439:79 83.
140 Dziewczapolski G, Lie DC, Ray J, Gage FH, Shults CW (2003) Survival and differentiation of adult rat derived neural progenitor cells transplanted to the striatum of hemiparkinsonian rats. Exp Neurol 183:653 664. Ehninger D, Kempermann G (2003) R egional effects of wheel running and environmental enrichment on cell genesis and microglia proliferation in the adult murine neocortex. Cereb Cortex 13:845 851. Ekdahl CT, Kokaia Z, Lindvall O (2009) Brain inflammation and adult neurogenesis: the dual rol e of microglia. Neuroscience 158:1021 1029. Eroglu C, Barres BA (2011) Regulation of synaptic connectivity by glia. Nature 468:223 231. Esposito MS, Piatti VC, Laplagne DA, Morgenstern NA, Ferrari CC, Pitossi FJ, Schinder AF (2005) Neuronal differentiation in the adult hippocampus recapitulates embryonic development. J Neurosci 25:10074 10086. Fabel K, Toda H, Palmer T (2003) Copernican stem cells: regulatory constellations in adult hippocampal neurogenesis. J Cell Biochem 88:41 50. Fern R, Moller T (2000) Rapid ischemic cell death in immature oligodendrocytes: a fatal glutamate release feedback loop. J Neurosci 20:34 42. Fisher M (1997) Characterizing the target of acute stroke therapy. Stroke 28:866 872. Freed CR, Greene PE, Breeze RE, Tsai WY, DuMouchel W Kao R, Dillon S, Winfield H, Culver S, Trojanowski JQ, Eidelberg D, Fahn S (2001) Transplantation of embryonic dopamine neurons for severe Parkinson's disease. N Engl J Med 344:710 719. Fujimura M, Morita Fujimura Y, Noshita N, Sugawara T, Kawase M, Chan PH (2000) The cytosolic antioxidant copper/zinc superoxide dismutase prevents the early release of mitochondrial cytochrome c in ischemic brain after transient focal cerebral ischemia in mice. J Neurosci 20:2817 2824. Fukui S, Ookawara T, Nawashiro H, Suz uki K, Shima K (2002) Post ischemic transcriptional and translational responses of EC SOD in mouse brain and serum. Free Radic Biol Med 32:289 298. Gage FH (2000) Mammalian neural stem cells. Science 287:1433 1438. Gage FH, Kempermann G, Palmer TD, Peterso n DA, Ray J (1998) Multipotent progenitor cells in the adult dentate gyrus. J Neurobiol 36:249 266. Galli L, Maffei L (1988) Spontaneous impulse activity of rat retinal ganglion cells in prenatal life. Science 242:90 91.
141 Gao J, Prough DS, McAdoo DJ, Grady JJ, Parsley MO, Ma L, Tarensenko YI, Wu P (2006) Transplantation of primed human fetal neural stem cells improves cognitive function in rats after traumatic brain injury. Exp Neurol 201:281 292. Goldman S (2003) Glia as neural progenitor cells. Trends Neur osci 26:590 596. Gomes JR, Lobo AC, Melo CV, Inacio AR, Takano J, Iwata N, Saido TC, de Almeida LP, Wieloch T, Duarte CB (2011) Cleavage of the vesicular GABA transporter under excitotoxic conditions is followed by accumulation of the truncated transporter in nonsynaptic sites. J Neurosci 31:4622 4635. Gould E, Cameron HA, Daniels DC, Woolley CS, McEwen BS (1992) Adrenal hormones suppress cell division in the adult rat dentate gyrus. J Neurosci 12:3642 3650. Gould E, Beylin A, Tanapat P, Reeves A, Shors TJ (1999) Learning enhances adult neurogenesis in the hippocampal formation. Nat Neurosci 2:260 265. Gross GW, Kowalski JM (1999) Origins of Activity Patterns in Self Organizing Neuronal Networks in Vitro. Journal of Intelligent Material Systems and Structure s 10:558 564. Harms C, Lautenschlager M, Bergk A, Katchanov J, Freyer D, Kapinya K, Herwig U, Megow D, Dirnagl U, Weber JR, Hortnagl H (2001) Differential mechanisms of neuroprotection by 17 beta estradiol in apoptotic versus necrotic neurodegeneration. J Neurosci 21:2600 2609. Harms KM, Li L, Cunningham LA (2010) Murine neural stem/progenitor cells protect neurons against ischemia by HIF 1alpha regulated VEGF signaling. PLoS One 5:e9767. Heikkila TJ, Yla Outinen L, Tanskanen JM, Lappalainen RS, Skottman H, Suuronen R, Mikkonen JE, Hyttinen JA, Narkilahti S (2009) Human embryonic stem cell derived neuronal cells form spontaneously active neuronal networks in vitro. Exp Neurol 218:109 116. Heja L, Barabas P, Nyitrai G, Kekesi KA, Lasztoczi B, Toke O, Tarkanyi G, Madsen K, Schousboe A, Dobolyi A, Palkovits M, Kardos J (2009) Glutamate uptake triggers transporter mediated GABA release from astrocytes. PLoS One 4:e7153. Hicks AU, Lappalainen RS, Narkilahti S, Suuronen R, Corbett D, Sivenius J, Hovatta O, Jolkkone n J (2009) Transplantation of human embryonic stem cell derived neural precursor cells and enriched environment after cortical stroke in rats: cell survival and functional recovery. Eur J Neurosci 29:562 574. Hirt L, Badaut J, Thevenet J, Granziera C, Regl i L, Maurer F, Bonny C, Bogousslavsky J (2004) D JNKI1, a cell penetrating c Jun N terminal kinase inhibitor, protects against cell death in severe cerebral ischemia. Stroke 35:1738 1743.
142 Hodges H, Sowinski P, Fleming P, Kershaw TR, Sinden JD, Meldrum BS, Gray JA (1996) Contrasting effects of fetal CA1 and CA3 hippocampal grafts on deficits in spatial learning and working memory induced by global cerebral ischaemia in rats. Neuroscience 72:959 988. Hoehn BD, Palmer TD, Steinberg GK (2005) Neurogenesis in rats after focal cerebral ischemia is enhanced by indomethacin. Stroke 36:2718 2724. Hoglinger GU, Rizk P, Muriel MP, Duyckaerts C, Oertel WH, Caille I, Hirsch EC (2004) Dopamine depletion impairs precursor cell proliferation in Parkinson disease. Nat Neurosci 7:726 735. Ieda M, Fu JD, Delgado Olguin P, Vedantham V, Hayashi Y, Bruneau BG, Srivastava D (2010) Direct reprogramming of fibroblasts into functional cardiomyocytes by defined factors. Cell 142:375 386. Illes S, Fleischer W, Siebler M, Hartung H P, Dihne M (2007) Development and pharmacological modulation of embryonic stem cell derived neuronal network activity. Exp Neurol 207:171 176. Illes S, Theiss S, Hartung HP, Siebler M, Dihne M (2009) Niche dependent development of functional neuronal netwo rks from embryonic stem cell derived neural populations. BMC Neurosci 10:93. Iwai M, Hayashi T, Zhang WR, Sato K, Manabe Y, Abe K (2001) Induction of highly polysialylated neural cell adhesion molecule (PSA NCAM) in postischemic gerbil hippocampus mainly d issociated with neural stem cell proliferation. Brain Res 902:288 293. Jaako Movits K, Zharkovsky A (2005) Impaired fear memory and decreased hippocampal neurogenesis following olfactory bulbectomy in rats. Eur J Neurosci 22:2871 2878. Jefferys JG, Haas HL (1982) Synchronized bursting of CA1 hippocampal pyramidal cells in the absence of synaptic transmission. Nature 300:448 450. Jgamadze D, Bergen J, Stone D, Jang JH, Schaffer DV, Isacoff EY, Pautot S (2012) Colloids as mobile substrates for the implantatio n and integration of differentiated neurons into the mammalian brain. PLoS One 7:e30293. Johnson MA, Weick JP, Pearce RA, Zhang SC (2007) Functional neural development from human embryonic stem cells: accelerated synaptic activity via astrocyte coculture. J Neurosci 27:3069 3077. Kamioka H, Maeda E, Jimbo Y, Robinson HP, Kawana A (1996) Spontaneous periodic synchronized bursting during formation of mature patterns of connections in cortical cultures. Neurosci Lett 206:109 112.
143 Kandel ER, Schwartz JH, Jessel l TM (2000) Principles of neural science, 4th Edition. New York: McGraw Hill, Health Professions Division. Kass IS, Lipton P (1982) Mechanisms involved in irreversible anoxic damage to the in vitro rat hippocampal slice. J Physiol 332:459 472. Kee NJ, Pres ton E, Wojtowicz JM (2001) Enhanced neurogenesis after transient global ischemia in the dentate gyrus of the rat. Exp Brain Res 136:313 320. Kempermann G, Gage FH (2002) Genetic determinants of adult hippocampal neurogenesis correlate with acquisition, but not probe trial performance, in the water maze task. Eur J Neurosci 16:129 136. Khazipov R, Esclapez M, Caillard O, Bernard C, Khalilov I, Tyzio R, Hirsch J, Dzhala V, Berger B, Ben Ari Y (2001) Early development of neuronal activity in the primate hippoc ampus in utero. J Neurosci 21:9770 9781. Kim DS, Kwak SE, Kim JE, Won MH, Choi HC, Song HK, Kwon OS, Kim YI, Choi SY, Kang TC (2005) Bilateral enhancement of excitation via up regulation of vesicular glutamate transporter subtype 1, not subtype 2, immunore activity in the unilateral hypoxic epilepsy model. Brain Res 1055:122 130. Kim H, Cooke MJ, Shoichet MS (2012) Creating permissive microenvironments for stem cell transplantation into the central nervous system. Trends Biotechnol 30:55 63. Kinouchi H, Epst ein CJ, Mizui T, Carlson E, Chen SF, Chan PH (1991) Attenuation of focal cerebral ischemic injury in transgenic mice overexpressing CuZn superoxide dismutase. Proc Natl Acad Sci U S A 88:11158 11162. Kitagawa H, Abe K, Hayashi T, Mitsumoto Y, Koga N, Itoya ma Y (1998) Ameliorative effect of glial cell line derived neurotrophic factor on brain edema formation after permanent middle cerebral artery occlusion in rats. Neurol Res 20:333 336. Krinke D, Jahnke HG, Panke O, Robitzki AA (2009) A microelectrode based sensor for label free in vitro detection of ischemic effects on cardiomyocytes. Biosens Bioelectron 24:2798 2803. Kucukdereli H, Allen NJ, Lee AT, Feng A, Ozlu MI, Conatser LM, Chakraborty C, Workman G, Weaver M, Sage EH, Barres BA, Eroglu C (2011) Contro l of excitatory CNS synaptogenesis by astrocyte secreted proteins Hevin and SPARC. Proc Natl Acad Sci U S A 108:E440 449. Kuhn HG, Winkler J, Kempermann G, Thal LJ, Gage FH (1997) Epidermal growth factor and fibroblast growth factor 2 have different effect s on neural progenitors in the adult rat brain. J Neurosci 17:5820 5829. Latham PE, Richmond BJ, Nirenberg S, Nelson PG (2000) Intrinsic dynamics in neuronal networks. II. experiment. J Neurophysiol 83:828 835.
144 Le Belle JE, Svendsen CN (2002) Stem cells fo r neurodegenerative disorders: where can we go from here? BioDrugs 16:389 401. Lee HJ, Kim KS, Park IH, Kim SU (2007) Human neural stem cells over expressing VEGF provide neuroprotection, angiogenesis and functional recovery in mouse stroke model. PLoS One 2:e156. Lepski G, Maciaczyk J, Jannes CE, Maciaczyk D, Bischofberger J, Nikkhah G (2011) Delayed functional maturation of human neuronal progenitor cells in vitro. Mol Cell Neurosci 47:36 44. Leuner B, Gould E, Shors TJ (2006) Is there a link between adul t neurogenesis and learning? Hippocampus 16:216 224. Levy YS, Stroomza M, Melamed E, Offen D (2004) Embryonic and adult stem cells as a source for cell therapy in Parkinson's disease. J Mol Neurosci 24:353 386. Liu J, Solway K, Messing RO, Sharp FR (1998) Increased neurogenesis in the dentate gyrus after transient global ischemia in gerbils. J Neurosci 18:7768 7778. Lledo PM, Hjelmstad GO, Mukherji S, Soderling TR, Malenka RC, Nicoll RA (1995) Calcium/calmodulin dependent kinase II and long term potentiation enhance synaptic transmission by the same mechanism. Proc Natl Acad Sci U S A 92:11175 11179. Lois C, Alvarez Buylla A (1993) Proliferating subventricular zone cells in the adult mammalian forebrain can differentiate into neurons and glia. Pro c Natl Acad Sci U S A 90:2074 2077. Luskin MB (1993) Restricted proliferation and migration of postnatally generated neurons derived from the forebrain subventricular zone. Neuron 11:173 189. Maccione A, Gandolfo M, Tedesco M, Nieus T, Imfeld K, Martinoia S, Berdondini L (2010) Experimental Investigation on Spontaneously Active Hippocampal Cultures Recorded by Means of High Density MEAs: Analysis of the Spatial Resolution Effects. Front Neuroengineering 3:4. Madl JE, Burgesser K (1993) Adenosine triphosphat e depletion reverses sodium dependent, neuronal uptake of glutamate in rat hippocampal slices. J Neurosci 13:4429 4444. Maeda E, Robinson HP, Kawana A (1995) The mechanisms of generation and propagation of synchronized bursting in developing networks of co rtical neurons. J Neurosci 15:6834 6845. Mangin JM, Gallo V (2000) The curious case of NG2 cells: transient trend or game changer? ASN Neuro 3:e00052.
145 Marrone DF, Ramirez Amaya V, Barnes CA (2012) Neurons generated in senescence maintain capacity for funct ional integration. Hippocampus. McIver SR, Muccigrosso M, Gonzales ER, Lee JM, Roberts MS, Sands MS, Goldberg MP Oligodendrocyte degeneration and recovery after focal cerebral ischemia. Neuroscience 169:1364 1375. Melchiorri D, Cappuccio I, Ciceroni C, Spi nsanti P, Mosillo P, Sarichelou I, Sale P, Nicoletti F (2007) Metabotropic glutamate receptors in stem/progenitor cells. Neuropharmacology 53:473 480. Meloni BP, Majda BT, Knuckey NW (2001) Establishment of neuronal in vitro models of ischemia in 96 well m icrotiter strip plates that result in acute, progressive and delayed neuronal death. Neuroscience 108:17 26. Meloni BP, Majda BT, Knuckey NW (2002) Evaluation of preconditioning treatments to protect near pure cortical neuronal cultures from in vitro ische mia induced acute and delayed neuronal death. Brain Res 928:69 75. Mercier F, Kitasako JT, Hatton GI (2002) Anatomy of the brain neurogenic zones revisited: fractones and the fibroblast/macrophage network. J Comp Neurol 451:170 188. Merkle FT, Tramontin AD Garcia Verdugo JM, Alvarez Buylla A (2004) Radial glia give rise to adult neural stem cells in the subventricular zone. Proc Natl Acad Sci U S A 101:17528 17532. Meyer DA, Carter JM, Johnstone AF, Shafer TJ (2008) Pyrethroid modulation of spontaneous neu ronal excitability and neurotransmission in hippocampal neurons in culture. Neurotoxicology 29:213 225. Mi R, Luo Y, Cai J, Limke TL, Rao MS, Hoke A (2005) Immortalized neural stem cells differ from nonimmortalized cortical neurospheres and cerebellar gran ule cell progenitors. Exp Neurol 194:301 319. Mielke JG, Comas T, Ahuja T, Preston E, Mealing GA (2007) Synaptic activity and triphenyltetrazolium chloride metabolism are correlated after oxygen glucose deprivation in acute, but not cultured, hippocampal s lices. Brain Res 1176:113 123. Mirzadeh Z, Merkle FT, Soriano Navarro M, Garcia Verdugo JM, Alvarez Buylla A (2008) Neural stem cells confer unique pinwheel architecture to the ventricular surface in neurogenic regions of the adult brain. Cell Stem Cell 3:265 278. Mistry SK, Keefer EW, Cunningham BA, Edelman GM, Crossin KL (2002) Cultured rat hippocampal neural progenitors generate spontaneously active neural networks. Proc Natl Acad Sci U S A 99:1621 1626.
146 Mizumatsu S, Monje ML, Morhardt DR, Rola R, Palm er TD, Fike JR (2003) Extreme sensitivity of adult neurogenesis to low doses of X irradiation. Cancer Res 63:4021 4027. Modo M, Hopkins K, Virley D, Hodges H (2003) Transplantation of neural stem cells modulates apolipoprotein E expression in a rat model o f stroke. Exp Neurol 183:320 329. Modo M, Stroemer RP, Tang E, Patel S, Hodges H (2002) Effects of implantation site of stem cell grafts on behavioral recovery from stroke damage. Stroke 33:2270 2278. Monje ML, Toda H, Palmer TD (2003) Inflammatory blockad e restores adult hippocampal neurogenesis. Science 302:1760 1765. Mu Y, Lee SW, Gage FH (2012) Signaling in adult neurogenesis. Curr Opin Neurobiol 20:416 423. Nacher J, McEwen BS (2006) The role of N methyl D asparate receptors in neurogenesis. Hippocampu s 16:267 270. Nikkhah G, Bentlage C, Cunningham MG, Bjorklund A (1994a) Intranigral fetal dopamine grafts induce behavioral compensation in the rat Parkinson model. J Neurosci 14:3449 3461. Nikkhah G, Cunningham MG, Jdicke A, Knappe U, Bjrklund A (1994b) Improved graft survival and striatal reinnervation by microtransplantation of fetal nigral cell suspensions in the rat Parkinson model. Brain Res 633:133 143. Nilsson M, Perfilieva E, Johansson U, Orwar O, Eriksson PS (1999) Enriched environment increases neurogenesis in the adult rat dentate gyrus and improves spatial memory. J Neurobiol 39:569 578. Nissant A, Pallotto M (2011) Integration and maturation of newborn neurons in the adult olfactory bulb -from synapses to function. Eur J Neurosci 33:1069 1077 O'Donovan MJ (1989) Motor activity in the isolated spinal cord of the chick embryo: synaptic drive and firing pattern of single motoneurons. J Neurosci 9:943 958. O'Shaughnessy TJ, Liu JL, Ma W (2009) Passaged neural stem cell derived neuronal networks for a portable biosensor. Biosens Bioelectron 24:2365 2370. O'Shaughnessy TJ, Zim B, Ma W, Shaffer KM, Stenger DA, Zamani K, Gross GW, Pancrazio JJ (2003) Acute neuropharmacologic action of chloroquine on cortical neurons in vitro. Brain Res 959:280 286.
147 O lanow CW, Goetz CG, Kordower JH, Stoessl AJ, Sossi V, Brin MF, Shannon KM, Nauert GM, Perl DP, Godbold J, Freeman TB (2003) A double blind controlled trial of bilateral fetal nigral transplantation in Parkinson's disease. Ann Neurol 54:403 414. Opitz T, De Lima AD, Voigt T (2002) Spontaneous development of synchronous oscillatory activity during maturation of cortical networks in vitro. J Neurophysiol 88:2196 2206. Ormerod BK, Galea LA (2001) Reproductive status influences cell proliferation and cell surviv al in the dentate gyrus of adult female meadow voles: a possible regulatory role for estradiol. Neuroscience 102:369 379. Ormerod BK, Galea LA (2003) Reproductive status influences the survival of new cells in the dentate gyrus of adult male meadow voles. Neurosci Lett 346:25 28. Ormerod BK, Falconer EM, Galea LA (2003) N methyl D aspartate receptor activity and estradiol: separate regulation of cell proliferation in the dentate gyrus of adult female meadow vole. J Endocrinol 179:155 163. Ormerod BK, Lee TT Galea LA (2004) Estradiol enhances neurogenesis in the dentate gyri of adult male meadow voles by increasing the survival of young granule neurons. Neuroscience 128:645 654. Ormerod BK, Palmer TD, Caldwell MA (2008) Neurodegeneration and cell replacement Philos Trans R Soc Lond B Biol Sci 363:153 170. Ostenfeld T, Tai YT, Martin P, Deglon N, Aebischer P, Svendsen CN (2002) Neurospheres modified to produce glial cell line derived neurotrophic factor increase the survival of transplanted dopamine neurons. J Neurosci Res 69:955 965. Palmer TD, Ray J, Gage FH (1995) FGF 2 responsive neuronal progenitors reside in proliferative and quiescent regions of the adult rodent brain. Mol Cell Neurosci 6:474 486. Palmer TD, Takahashi J, Gage FH (1997) The adult rat hip pocampus contains primordial neural stem cells. Mol Cell Neurosci 8:389 404. Palmer TD, Schwartz PH, Taupin P, Kaspar B, Stein SA, Gage FH (2001) Cell culture. Progenitor cells from human brain after death. Nature 411:42 43. Pancrazio JJ, Keefer EW, Ma W, Stenger DA, Gross GW (2001) Neurophysiologic effects of chemical agent hydrolysis products on cortical neurons in vitro. Neurotoxicology 22:393 400.
148 Parent JM, Yu TW, Leibowitz RT, Geschwind DH, Sloviter RS, Lowenstein DH (1997) Dentate granule cell neurog enesis is increased by seizures and contributes to aberrant network reorganization in the adult rat hippocampus. J Neurosci 17:3727 3738. Pasquale V, Massobrio P, Bologna LL, Chiappalone M, Martinoia S (2008) Self organization and neuronal avalanches in ne tworks of dissociated cortical neurons. Neuroscience 153:1354 1369. Peinado A (2000) Traveling slow waves of neural activity: a novel form of network activity in developing neocortex. J Neurosci 20:RC54. Peljto M, Wichterle H Programming embryonic stem cel ls to neuronal subtypes. Curr Opin Neurobiol 21:43 51. Perez Velazquez JL, Frantseva MV, Carlen PL (1997) In vitro ischemia promotes glutamate mediated free radical generation and intracellular calcium accumulation in hippocampal pyramidal neurons. J Neuro sci 17:9085 9094. Pessoa L (2008) On the relationship between emotion and cognition. Nat Rev Neurosci 9:148 158. Piatti VC, Davies Sala MG, Esposito MS, Mongiat LA, Trinchero MF, Schinder AF (2011) The timing for neuronal maturation in the adult hippocampu s is modulated by local network activity. J Neurosci 31:7715 7728. Pine J (1980) Recording action potentials from cultured neurons with extracellular microcircuit electrodes. J Neurosci Methods 2:19 31. Pisani A, Calabresi P, Tozzi A, D'Angelo V, Bernardi G (1998) L type Ca2+ channel blockers attenuate electrical changes and Ca2+ rise induced by oxygen/glucose deprivation in cortical neurons. Stroke 29:196 201; discussion 202. Potter SM, DeMarse TB (2001) A new approach to neural cell culture for long term studies. J Neurosci Methods 110:17 24. Raber J, Rola R, LeFevour A, Morhardt D, Curley J, Mizumatsu S, VandenBerg SR, Fike JR (2004) Radiation induced cognitive impairments are associated with changes in indicators of hippocampal neurogenesis. Radiat Res 1 62:39 47. Reynolds BA, Weiss S (1992) Generation of neurons and astrocytes from isolated cells of the adult mammalian central nervous system. Science 255:1707 1710. Rola R, Raber J, Rizk A, Otsuka S, VandenBerg SR, Morhardt DR, Fike JR (2004) Radiation ind uced impairment of hippocampal neurogenesis is associated with cognitive deficits in young mice. Exp Neurol 188:316 330.
149 Rosner G, Graf R, Kataoka K, Heiss WD (1986) Selective functional vulnerability of cortical neurons following transient MCA occlusion i n the cat. Stroke 17:76 82. Rossi DJ, Oshima T, Attwell D (2000) Glutamate release in severe brain ischaemia is mainly by reversed uptake. Nature 403:316 321. Roybon L, Christophersen NS, Brundin P, Li JY (2004) Stem cell therapy for Parkinson's disease: where do we stand? Cell Tissue Res 318:261 273. Sanai N, Tramontin AD, Quinones Hinojosa A, Barbaro NM, Gupta N, Kunwar S, Lawton MT, McDermott MW, Parsa AT, Manuel Garcia Verdugo J, Berger MS, Alvarez Buylla A (2004) Unique astrocyte ribbon in adult human brain contains neural stem cells but lacks chain migration. Nature 427:740 744. Scarlatos A, Cadotte AJ, DeMarse TB, Welt BA (2008) Cortical networks grown on microelectrode arrays as a biosensor for botulinum toxin. J Food Sci 73:E129 136. Scoville WB, M ilner B (2000) Loss of recent memory after bilateral hippocampal lesions. 1957. J Neuropsychiatry Clin Neurosci 12:103 113. Seaberg RM, Smukler SR, van der Kooy D (2005) Intrinsic differences distinguish transiently neurogenic progenitors from neural stem cells in the early postnatal brain. Dev Biol 278:71 85. Shimada IS, Spees JL (2011) Stem and progenitor cells for neurological repair: minor issues, major hurdles, and exciting opportunities for paracrine based therapeutics. J Cell Biochem 112:374 380. Sinden JD, Rashid Doubell F, Kershaw TR, Nelson A, Chadwick A, Jat PS, Noble MD, Hodges H, Gray JA (1997) Recovery of spatial learning by grafts of a conditionally immortalized hippocampal neuroepithelial cell line into the ischaemia lesioned hippocampus. Neuroscience 81:599 608. Smith AG (2001) Embryo derived stem cells: of mice and men. Annu Rev Cell Dev Biol 17:435 462. Snyder EY, Deitcher DL, Walsh C, Arnold Aldea S, Hartwieg EA, Cepko CL (1992) Multipotent neural cell lines can engraft and participate in development of mouse cerebellum. Cell 68:33 51. Snyder JS, Hong NS, McDonald RJ, Wojtowicz JM (2005) A role for adult neurogenesis in spatial long term memory. Neuroscience 130:843 852. Song H, Stevens CF, Gage FH (2002a) Astroglia induce neurogenesis f rom adult neural stem cells. Nature 417:39 44. Song HJ, Stevens CF, Gage FH (2002b) Neural stem cells from adult hippocampus develop essential properties of functional CNS neurons. Nat Neurosci 5:438 445.
150 Song J, Christian K, Ming GL, Song H (2012) Modific ation of hippocampal circuitry by adult neurogenesis. Dev Neurobiol. Steiner B, Kronenberg G, Jessberger S, Brandt MD, Reuter K, Kempermann G (2004) Differential regulation of gliogenesis in the context of adult hippocampal neurogenesis in mice. Glia 46:41 52. Stephens CL, Toda H, Palmer TD, Demarse TB, Ormerod BK (2012) Adult neural progenitor cells reactivate superbursting in mature neural networks. Exp Neurol 234:20 30. Strubing C, Ahnert Hilger G, Shan J, Wiedenmann B, Hescheler J, Wobus AM (1995) Diffe rentiation of pluripotent embryonic stem cells into the neuronal lineage in vitro gives rise to mature inhibitory and excitatory neurons. Mech Dev 53:275 287. Suhonen JO, Peterson DA, Ray J, Gage FH (1996) Differentiation of adult hippocampus derived proge nitors into olfactory neurons in vivo. Nature 383:624 627. Svendsen CN, Caldwell MA, Ostenfeld T (1999) Human neural stem cells: isolation, expansion and transplantation. Brain Pathol 9:499 513. Svendsen CN, ter Borg MG, Armstrong RJ, Rosser AE, Chandran S Ostenfeld T, Caldwell MA (1998) A new method for the rapid and long term growth of human neural precursor cells. J Neurosci Methods 85:141 152. Szatkowski M, Barbour B, Attwell D (1990) Non vesicular release of glutamate from glial cells by reversed elec trogenic glutamate uptake. Nature 348:443 446. Tabak J, Latham PE (2003) Analysis of spontaneous bursting activity in random neural networks. Neuroreport 14:1445 1449. Takagi Y, Nozaki K, Takahashi J, Yodoi J, Ishikawa M, Hashimoto N (1999) Proliferation o f neuronal precursor cells in the dentate gyrus is accelerated after transient forebrain ischemia in mice. Brain Res 831:283 287. Takahashi J, Palmer TD, Gage FH (1999) Retinoic acid and neurotrophins collaborate to regulate neurogenesis in adult derived n eural stem cell cultures. J Neurobiol 38:65 81. Takahashi K, Yamanaka S (2006) Induction of pluripotent stem cells from mouse embryonic and adult fibroblast cultures by defined factors. Cell 126:663 676. Takahashi K, Tanabe K, Ohnuki M, Narita M, Ichisaka T, Tomoda K, Yamanaka S (2007) Induction of pluripotent stem cells from adult human fibroblasts by defined factors. Cell 131:861 872.
151 Tanaka E, Yamamoto S, Kudo Y, Mihara S, Higashi H (1997) Mechanisms underlying the rapid depolarization produced by depriv ation of oxygen and glucose in rat hippocampal CA1 neurons in vitro. J Neurophysiol 78:891 902. Tanapat P, Hastings NB, Reeves AJ, Gould E (1999) Estrogen stimulates a transient increase in the number of new neurons in the dentate gyrus of the adult female rat. J Neurosci 19:5792 5801. Tashiro A, Makino H, Gage FH (2007) Experience specific functional modification of the dentate gyrus through adult neurogenesis: a critical period during an immature stage. J Neurosci 27:3252 3259. Tateno T, Jimbo Y, Robinson HP (2005) Spatio temporal cholinergic modulation in cultured networks of rat cortical neurons: evoked activity. Neuroscience 134:439 448. Tavazoie M, Van der Veken L, Silva Vargas V, Louissaint M, Colonna L, Zaidi B, Garcia Verdugo JM, Doetsch F (2008) A specialized vascular niche for adult neural stem cells. Cell Stem Cell 3:279 288. Tekkok SB, Goldberg MP (2001) Ampa/kainate receptor activation mediates hypoxic oligodendrocyte death and axonal injury in cerebral white matter. J Neurosci 21:4237 4248. Tho mson JA, Itskovitz Eldor J, Shapiro SS, Waknitz MA, Swiergiel JJ, Marshall VS, Jones JM (1998) Embryonic stem cell lines derived from human blastocysts. Science 282:1145 1147. Thored P, Arvidsson A, Cacci E, Ahlenius H, Kallur T, Darsalia V, Ekdahl CT, Kok aia Z, Lindvall O (2006) Persistent production of neurons from adult brain stem cells during recovery after stroke. Stem Cells 24:739 747. Toda H, Tsuji M, Nakano I, Kobuke K, Hayashi T, Kasahara H, Takahashi J, Mizoguchi A, Houtani T, Sugimoto T, Hashimot o N, Palmer TD, Honjo T, Tashiro K (2003) Stem cell derived neural stem/progenitor cell supporting factor is an autocrine/paracrine survival factor for adult neural stem/progenitor cells. J Biol Chem 278:35491 35500. Tonnesen J, Parish CL, Sorensen AT, And ersson A, Lundberg C, Deisseroth K, Arenas E, Lindvall O, Kokaia M (2011) Functional integration of grafted neural stem cell derived dopaminergic neurons monitored by optogenetics in an in vitro Parkinson model. PLoS One 6:e17560. van Pelt J, Wolters PS, C orner MA, Rutten WL, Ramakers GJ (2004) Long term characterization of firing dynamics of spontaneous bursts in cultured neural networks. IEEE Trans Biomed Eng 51:2051 2062.
152 van Praag H, Christie BR, Sejnowski TJ, Gage FH (1999) Running enhances neurogenesi s, learning, and long term potentiation in mice. Proc Natl Acad Sci U S A 96:13427 13431. van Praag H, Schinder AF, Christie BR, Toni N, Palmer TD, Gage FH (2002) Functional neurogenesis in the adult hippocampus. Nature 415:1030 1034. Vemuganti R (2005) De creased expression of vesicular GABA transporter, but not vesicular glutamate, acetylcholine and monoamine transporters in rat brain following focal ischemia. Neurochem Int 47:136 142. Voigt T, Opitz T, de Lima AD (2001) Synchronous oscillatory activity in immature cortical network is driven by GABAergic preplate neurons. J Neurosci 21:8895 8905. Wagenaar DA, Pine J, Potter SM (2006) An extremely rich repertoire of bursting patterns during the development of cortical cultures. BMC Neurosci 7:11. Wagenaar DA Madhavan R, Pine J, Potter SM (2005) Controlling bursting in cortical cultures with closed loop multi electrode stimulation. J Neurosci 25:680 688. Walker NI, Harmon BV, Gobe GC, Kerr JF (1988) Patterns of cell death. Methods Achiev Exp Pathol 13:18 54. Weick JP, Liu Y, Zhang SC (2011) Human embryonic stem cell derived neurons adopt and regulate the activity of an established neural network. Proc Natl Acad Sci U S A 108:20189 20194. Wernig M, Benninger F, Schmandt T, Rade M, Tucker KL, Bussow H, Beck H, B rustle O (2004) Functional integration of embryonic stem cell derived neurons in vivo. J Neurosci 24:5258 5268. Winkler C, Kirik D, Bjorklund A (2005) Cell transplantation in Parkinson's disease: how can we make it work? Trends Neurosci 28:86 92. Winocur G Wojtowicz JM, Sekeres M, Snyder JS, Wang S (2006) Inhibition of neurogenesis interferes with hippocampus dependent memory function. Hippocampus 16:296 304. Wyllie AH, Beattie GJ, Hargreaves AD (1981) Chromatin changes in apoptosis. Histochem J 13:681 692 Yagita Y, Kitagawa K, Ohtsuki T, Takasawa K, Miyata T, Okano H, Hori M, Matsumoto M (2001) Neurogenesis by progenitor cells in the ischemic adult rat hippocampus. Stroke 32:1890 1896.
153 Yeung CK, Sommerhage F, Wrobel G, Law JK, Offenhausser A, Rudd JA, Ing ebrandt S, Chan M (2009) To establish a pharmacological experimental platform for the study of cardiac hypoxia using the microelectrode array. J Pharmacol Toxicol Methods 59:146 152. Yu J, Vodyanik MA, He P, Slukvin, II, Thomson JA (2006) Human embryonic s tem cells reprogram myeloid precursors following cell cell fusion. Stem Cells 24:168 176. Zaghloul N, Nasim M, Patel H, Codipilly C, Marambaud P, Dewey S, Schiffer WK, Ahmed M (2012) Overexpression of extracellular superoxide dismutase has a protective rol e against hyperoxia induced brain injury in neonatal mice. FEBS J 279:871 881. Zerangue N, Kavanaugh MP (1996) Flux coupling in a neuronal glutamate transporter. Nature 383:634 637. Zhang R, Zhang Z, Wang L, Wang Y, Gousev A, Zhang L, Ho KL, Morshead C, Chopp M (2004) Activated neural stem cells contribute to stroke induced neurogenesis and neuroblast migration toward the infarct boundary in adult rats. J Cereb Blood Flow Metab 24:441 448. Zhao C, Deng W, Gage FH (2008) Mechanisms and functional implicati ons of adult neurogenesis. Cell 132:645 660. Zhu JM, Zhao YY, Chen SD, Zhang WH, Lou L, Jin X (2011) Functional recovery after transplantation of neural stem cells modified by brain derived neurotrophic factor in rats with cerebral ischaemia. J Int Med Res 39:488 498.
154 BIOGRAPHICAL SKETCH Crystal Lynn Stephens was born in St. Petersburg, Florida in 1984. The oldest of four children, she grew up primarily in Clearwater, Florida, and graduated with honors from Countryside High School in 2002. Crystal obtained her B.S. in a gricultural and b iological e ngineering cum laude with a minor in b iomechanics (concentrating on an area in biological engineering) from the University of Florida in 2006, supported by a Bright Futures Florida Academic Schola rship. She continued her postgraduate work at the University of Florida in the J. Crayton Pruitt Family Department of Biomedical Engineering as a recipient of the University Alumni Fellowship. Upon completion of the Ph.D. program, Crystal plans to pursue a career as a research scientist in the private sector for a pharmaceutical or biotechnological company.