The Role of Potassium Ion and Water Channels in an Animal Model Of

Home , Astrogliosis, KCNC1, KCNC3, KCNK9

The Role of Potassium Ion and Water Channels in an Animal Model of

Multiple Sclerosis

DISSERTATION

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University

Peter Isaac Jukkola, B.S.

Graduate Program in Biomedical Science

The Ohio State University

2014

Dissertation Committee:

Chen Gu, Advisor

Brian Kaspar

Amy Lovett-Racke

John Oberdick

Peter Isaac Jukkola

2014

ABSTRACT

Neuroinflammation and demyelination in multiple sclerosis (MS) lead first to neuronal dysfunction, then to the potential for neurodegeneration. Voltage-gated potassium ion channels have key roles in maintaining the resting membrane potential of a neuron in readiness for neurotransmission. This includes repolarization of the membrane following an action potential, which involves a depolarization event mediated by voltage-gated sodium channels. Myelin, the many-layered lipid sheath surrounding axons, is known to participate in molecular interactions with the axonal membrane to target ion channels to important, specialized locations along the axon. The water channel aquaporin-4 (AQP4), expressed in astrocyte endfeet, also plays an important role in maintaining ionic and fluid homeostasis in the neuronal environment. Because progression of disease and permanent disability in MS appears related to the degree of neuronal loss, neuroprotective treatments are needed in MS. Although a nonspecific Kv channel blocker is currently approved for symptomatic treatment of MS, it is not known to provide any neuroprotective effect, it has significant side effects, and its mechanism of action is not clear. In my dissertation research, I used two animal models of MS, chronic or relapsing-remitting experimental autoimmune encephalomyelitis (chEAE or rrEAE) to mimic progressive or relapsing- remitting MS, respectively, and to characterize the effect of inflammatory, demyelinating

ii lesions upon the expression and localization patterns of AQP4 and key Kv channels. I found that Kv 1.2, expressed in myelinated axons in spinal cord (SC) white matter (WM), was redistributed in lesioned areas from its normal location at the juxtaparanode (JXP).

The JXP localization could be recovered in remitting rrEAE, but not late chEAE. Kv 2.1, normally clustered on the soma and proximal dendrites of alpha motor neurons located in

SC ventral gray matter (GM), was declustered and reduced in expression levels in EAE.

Kv 2.1 expression recovered to a more normal pattern in remitting rrEAE, but not late chEAE. Kv 1.4 expression was markedly increased in EAE SCWM surrounding lesions in reactive astrocytes, and this effect was enhanced in remitting EAE but not late chonic

EAE. AQP4 and other astrocyte proteins were upregulated in SCWM lesioned areas in

EAE, but reactive astrocytes in other CNS regions showed varying patterns of protein expression. These changes may reveal compensatory or neuroprotective effects of ion and water channel expression and regulation in the CNS during neuroinflammatory disease. My studies pave the way for more detailed studies in this field to determine the regulatory mechanisms and functional consequences of channel alteration in response to disease. As these elements of channel biology are better understood, better-targeted therapies with fewer side effects can be developed for the treatment of neurobiological disorders such as MS.

iii

DEDICATION

This document is dedicated to my wife Kristie and my children, Timothy, Adelyn,

Johannah, and Samuel. Their love and support gave me the courage to pursue this course

to completion.

ACKNOWLEDGMENTS

I am thankful to: Dr. Kevin M. Kelly for giving me my start in biomedical research and for encouraging me to pursue graduate studies; to Dr. Haiyan Fu and Dr. Amy Lovett-

Racke for allowing me to rotate in their labs; to Dr. Ichiko Nishijima for her mentorship through my PhD candidacy exam; to my current mentor Dr. Chen Gu for challenging me to be my best; to my coworkers Dr. Mingxuan Xu, Dr. Yuanzheng Gu, and Dr. Joshua

Barry for their support, companionship and the many procedures they taught me; to

Paula Monsma and the OSU Neurobiology Imaging Core for maintaining the microscope facilities that played a large part in my studies; to my dissertation committee members

Dr. Brian Kaspar, Dr. Amy Lovett-Racke, and Dr. John Oberdick; to the numerous faculty members who have contributed to my education; and to Nationwide Children’s

Hospital and The Ohio State University for institutional support.

VITA

May 1996 ...... Aliquippa Baptist Temple Academy

1996-1998 ...... Biological Science, Community College of

Beaver County

1998-2000 ...... B.S., Biology, Geneva College

2001-2007…………………………………..Research Assistant/Sr. Research Assistant,

Allegheny-Singer Research Institute

2007 to present ...... Graduate Research Associate, Department

of Medicine, The Ohio State University

Publications

Barry J, Gu Y, Jukkola P, O’Neill B, Gu H, Mohler PJ, Thirtamara Rajamani K, and Gu

C (2014). Ankyrin-G Directly Binds to Kinesin-1 to Transport Voltage-Gated Na+

Channels into Axons. Developmental Cell. Jan 27;28(2):117-31

Jukkola P, Guerrero T, Gray V, and Gu C. Astrocytes Differentially Respond to

Inflammatory Autoimmune Insults and Imbalances of Neural Activity. Acta

Neuropathologica Communications. 1:70 (23 October 2013)

Barry J, Xu M, Gu Y, Dangel A, Jukkola P, Shrestha C, and Gu C (2013). Activation of

Conventional Kinesin Motors in Clusters by Shaw Voltage-Gated Potassium

Channels. Journal of Cell Science May 1;126(Pt 9):2027-41.

Gardner A, Jukkola P, Gu C (2012). Myelination of rodent hippocampal neurons in

culture. Nature Protocols. Sep 6;7(10):1774-82.

Jukkola, PI, Lovett-Racke AE, Zamvil SS, Gu C (2012). K+ channel alterations in the

progression of experimental autoimmune encephalomyelitis. Neurobiology of

Disease. Aug;47(2):280-93.

Kelly KM, Shiau DS, Jukkola PI, Miller ER, Mercadante AL, Quigley MM, Nair SP,

Sackellares JC (2011). Effects of age and cortical infarction on EEG dynamic

changes associated with spike wave discharges in F344 rats. Experimental

Neurology Nov;232(1):15-21.

Jukkola PI, Rogers JT, Kaspar BK, Weeber EJ, Nishijima I (2011). Secretin deficiency

causes impairment in survival of neural progenitor cells in mice. Human

Molecular Genetics. Mar 1;20(5):1000-7.

DiRosario J, Divers E, Wang C, Etter J, Charrier A, Jukkola P, Auer H, Best V,

Newsom DL, McCarty DM and Fu H (2009) Innate and adaptive immune

activation in the brain of MPS IIIB mouse model. J Neurosci Res 87:978-990.

Nair SP, Jukkola PI, Quigley M, Wilberger A, Shiau DS, Sackellares JC, Pardalos PM,

Kelly KM (2008). Absence seizures as resetting mechanisms of brain dynamics.

Cybernetics and Systems Analysis 44(5):664-72.

vii

Kelly KM, Kharlamov EA, Downey KL, Jukkola PI, Grayson DR (2008). Expression

of GABAA receptor α1 subunit mRNA and protein in rat neocortex following

photothrombotic infarction. Brain Research 1210: 29-38.

Kelly KM, Jukkola PI, Kharlamov EA, Downey KL, McBride JW, Strong R, Aronowski

J (2006). Long-term video-EEG recordings following transient unilateral middle

cerebral and common carotid artery occlusion in Long–Evans rats. Experimental

Neurology 201:495-506.

Kharlamova EA, Jukkola PI, Schmitt KL, Kelly KM (2003). Electrobehavioral

characteristics of epileptic rats following photothrombotic brain infarction.

Epilepsy Research 56:185-203.

Fields of Study

Major Field: Biomedical Sciences

Area of Research Emphasis: Biology of Neurological Disorders

viii

TABLE OF CONTENTS

ABSTRACT ...... II DEDICATION...... IV ACKNOWLEDGMENTS ...... V VITA...... VI TABLE OF CONTENTS ...... IX LIST OF TABLES ...... XII LIST OF FIGURES ...... XIII CHAPTER 1: INTRODUCTION ...... 1 NEUROINFLAMMATION AND NEURODEGENERATION...... 1 DEMYELINATION ...... 2 MULTIPLE SCLEROSIS ...... 3 NEUROPROTECTION ...... 4 ION CHANNELS IN NEUROTRANSMISSION ...... 5 ION AND WATER CHANNELS IN THE NEUROVASCULAR UNIT ...... 7 ASTROCYTE CHANNEL PROTEINS ...... 10 Water channel AQP4 ...... 10 Inwardly rectifying potassium channel Kir4.1 ...... 13 Ca2+ and Voltage-activated BK channels ...... 17 Two-pore-domain K+ channels ...... 19 Voltage-gated calcium channels ...... 21 PERSPECTIVES FOR ASTROCYTIC ION CHANNEL RESEARCH ...... 24 Recently identified channel proteins expressed in astrocytes ...... 24 Regulation of astrocytic channel functions by intracellular signaling pathways ..... 27 SUMMARY ...... 29 OVERALL GOAL ...... 30 CHAPTER 2: K+ CHANNEL ALTERATIONS IN EAE ...... 31 INTRODUCTION ...... 31 ix

METHODS ...... 33 RESULTS ...... 42 Demyelinated lesions in chEAE and rrEAE ...... 42

Disrupted Kv1.2/Kvβ2 JXP clustering in demyelinated lesion areas in EAE ...... 46

Altered patterns of Kv1.2 targeting along axons in various lesion areas ...... 55

Somatodendritic Kv2.1 in spinal cord motor neurons during EAE progression ...... 62

Upregulation of Kv1.4 channels around spinal cord lesions in EAE progression .... 68 DISCUSSION ...... 81

Localized regulation of JXP clusters of Kv 1.2/ Kv β2 channels ...... 81

Downregulation of Kv2.1 and upregulation of Kv1.4 during EAE progression ...... 83

Kv1.4- and GFAP-positive astrocytes around EAE lesions and their potential function ...... 84 CONCLUSION ...... 86 CHAPTER 3: ASTROCYTE REACTIVITY TO INFLAMMATION AND NEURONAL DYSFUNCTION ...... 87 INTRODUCTION ...... 87 METHODS ...... 90 RESULTS ...... 97

AQP4, Kv1.4, Vimentin and GFAP are upregulated in SCWM at different stages of EAE ...... 98 Differential alterations of myelin and astrocytes in the late and remitting EAE in SCWM ...... 102 Cerebellar astrocytes are differentially activated at different stages of EAE ...... 102 Hippocampal astrocytes are activated at the peak and late stages of EAE ...... 110 Cortical astrocytes are activated at the peak and late stages of EAE ...... 113

Astrocyte alterations in AnkG and Kv3.1 KO mice ...... 116 DISCUSSION ...... 122 Both fibrous and protoplasmic astrocytes are activated during EAE, but only fibrous astrocytes are selectively damaged around lesion sites ...... 122 Neuronal activity regulates astrocytic protein expression ...... 127 Differential localization of AQP4 in fibrous and protoplasmic astrocytes ...... 128 Functional consequences of astrocyte activation ...... 129 CONCLUSIONS ...... 130 CHAPTER 4: IN VIVO ION CHANNEL TRANSPORT MECHANISMS ...... 131 x

KV3.1 CLUSTERS AND ACTIVATES KINESIN MOTORS ...... 131 INTRODUCTION ...... 131 METHODS ...... 133 RESULTS ...... 134 DISCUSSION ...... 138 ANKYRIN-G SERVES AS AN ADAPTOR FOR KINESIN-MEDIATED TRANSPORT OF VOLTAGE-GATED SODIUM CHANNELS ...... 140 INTRODUCTION ...... 140 METHODS ...... 141 RESULTS ...... 145 DISCUSSION ...... 149 CHAPTER 5: CONCLUSIONS ...... 150 OVERVIEW ...... 150 CONTRIBUTION TO ION AND WATER CHANNEL RESEARCH ...... 150 CONTRIBUTION TO NEUROPROTECTION FIELD ...... 153 CONTRIBUTION TO MS RESEARCH ...... 155 CONCLUSION ...... 158 REFERENCES ...... 150

LIST OF TABLES

Table 1. Channel proteins expressed in astrocytes ...... 26

xii

LIST OF FIGURES

Figure 1. Sectioning strategy for brain and spinal cord...... 37

Figure 2. Characterization of the Chronic EAE model...... 44

Figure 3. Characterization of the Relapsing-Remitting EAE model...... 45

Figure 4. Altered Kv1.2 JXP targeting in lesioned areas in chEAE spinal cord...... 47

Figure 5. Kv1.2 JXP clustering was altered in rrEAE lesions...... 49

Figure 6. Kvβ2 JXP clustering was disrupted in lesions of chEAE and rrEAE spinal cord...... 50

Figure 7. Expression of Kv1.2 in the areas with or without a lesion in spinal cord white matter at different stages of chEAE and rrEAE...... 51

Figure 8. Kv1.2 expression patterns in brain regions during EAE progression...... 53

Figure 9. The global expression of Kv1.2/ Kvβ2 in cerebral cortex and spinal cord at different stages of chEAE and rrEAE revealed by Western blotting...... 54

Figure 10. Alterations of Kv1.2 targeting patterns along axons within EAE lesions.

...... 57

Figure 11. Different patterns of altered Kv1.2 JXP targeting in demyelinated lesions...... 59

xiii

Figure 12. Different types of lesions in spinal cord white matter of chEAE and rrEAE...... 61

Figure 13. Alterations of somatodendritic Kv2.1 channels in spinal cord motor neurons in chEAE...... 63

Figure 14. Kv2.1 levels in motor neurons of lower spinal cord in rrEAE correlated with clinical signs...... 65

Figure 15. Quantification of fluorescence intensity of Kv2.1 in spinal cord alpha motor neurons...... 67

Figure 16. Upregulation of Kv1.4 channels in GFAP-positive cells around EAE lesions ...... 70

Figure 17. Specificity of the rabbit polyclonal anti- Kv1.4 antibody...... 72

Figure 18. High-magnification confocal imaging of Kv1.4 upregulation in chEAE and rrEAE...... 74

Figure 19. Up-regulated Kv1.4 channels partially colocalized with NG2 but not with

YFP-positive axons in Thy1-YFP transgenic mice...... 76

Figure 20. Global expression of Kv1.4 channels revealed by Western blotting of cerebral cortex and spinal cord at different stages of chEAE and rrEAE ...... 77

Figure 21. Kv1.4 levels decreased in lesions of spinal cord white matter during the late relapse phase of rrEAE...... 80

Figure 22. Clinical scores and body weight of mice with chEAE...... 97

Figure 23. Activation of astrocytes in chEAE spinal cord white matter...... 99

xiv

Figure 24. High magnification confocal image stacks were obtained from control and EAE Thy1-YFP transgenic mice...... 101

Figure 25. Astrocytic proteins are altered in lesion sites of at the late stage of chEAE and at the remitting stage of rrEAE...... 104

Figure 26. High magnification view of a lesion site at the late stage of chEAE and at the remitting stage of rrEAE...... 105

Figure 27. Summary of the alteration of astrocytic proteins in chEAE and rrEAE.

...... 106

Figure 28. Differentially altered expression of astrocytic proteins in the cerebellum of EAE mice...... 108

Figure 29. The confocal images of cerebellar WM from control and EAE Thy1-YFP transgenic mice ...... 109

Figure 30. Activation of astrocytes in the hippocampus of EAE mice...... 111

Figure 31. High magnification confocal image stacks were obtained from control and EAE Thy1-YFP transgenic mice...... 112

Figure 32. Alteration of GFAP and AQP4 in the cortex of EAE mice...... 114

Figure 33. High magnification confocal image stacks were obtained from control and EAE Thy1-YFP transgenic mice...... 115

Figure 34. Upregulation of GFAP and AQP4 in the cerebellum in AnkG and Kv3.1

KO mice...... 119

Figure 35. Confocal imaging of the granule cell layer in a WT, AnkG KO , and

Kv3.1 KO mice...... 120

Figure 36. Altered astrocytes in the hippocampus and cortex in AnkG KO and

Kv3.1 KO mice...... 121

Figure 37. Diagram of regulation of fibrous and protoplasmic astrocytes by inflammation and neuronal activities...... 125

Figure 38. Decreased KIF5 clustering in cerebellar neurons of Kv3.1 KO mice. . 137

Figure 39. AnkG Deletion or Disruption of AnkG-KIF5 Binding by Expressing a

KIF5B Tail Fragment Reduces the Axonal Level of Nav Channels In Vivo ...... 148

xvi

CHAPTER 1: INTRODUCTION

NEUROINFLAMMATION AND NEURODEGENERATION

Neuroinflammation can lead to reduced neuronal function and to neurological disability. The CNS is immune-privileged, protected by the blood-brain barrier (BBB) so that immune cells have limited access to neurons. When the BBB breaks down or when resident immune cells (microglia) within the CNS recognize foreign material, inflammatory cells such as T lymphocytes and macrophages gain increased access and may mount an immune response to the foreign particles or, potentially, an autoimmune response against the body’s own tissues. Immune activity within the brain can alter the chemical environment of the neurons by the release of cytokines, reactive oxygen species, and other metabolites. This can cause neuronal dysfunction and swelling of axons, potentially leading to neurodegeneration.

DEMYELINATION

Inflammation can also damage other cells within the brain. Demyelination is the loss of the myelin sheath, composed of many layers of membrane produced by oligodendrocytes that wrap around axons. Myelin provides a protective, insulating layer for axons, and also alters the mode of neurotransmission from a slower impulse traveling along the axon to high-speed saltatory conduction in which the nerve impulse seems to jump from one node of Ranvier to the next. The node of Ranvier is a highly specialized area of the axon located between two segments of myelin. Adjacent this is the paranode, where the many loops of the myelin membrane layers are tethered to the axonal membrane by molecular interactions. Flanking these paranodes on either side of the node are the juxtaparanodes (JXP), just underneath the myelin membranes. Ion channels expressed in the nodal area have highly specific localization patterns, with Nav channels concentrated in the node and Kv1 channels clustered in the juxtaparanodes. Molecular interactions between the myelin and axonal membranes, including cell adhesion molecules, neurofascins, and scaffolding molecules, hold these channels in their proper location and exclude them from areas of the nodal region where they do not belong.

Thus, when the myelin is stripped away from the axons in demyelinating disease conditions, we would expect significant alteration of the localization patterns of ion channels in the axonal membrane.

MULTIPLE SCLEROSIS

Multiple Sclerosis is an inflammatory, demyelinating disease. Although the initial causative factor for MS is not known, it is clear that disease pathology is characterized by immune-mediated destruction of myelin within the CNS. The effects of demyelination in the heavily myelinated white matter areas of the brain are clearly recognized clinically by magnetic resonance imaging (MRI) in MS patients, and histologically by lipid-directed stains such as Luxol fast blue, and have thus received the most attention from the research community. However, the effects of neuroinflammation and demyelination in the gray matter areas of the brain, which have a much lesser degree of myelination in normal physiology, are now being recognized more and more. In both white matter and gray matter areas of the brain, these effects can include activation of microglia and astrocytes as well as neuronal dysfunction and loss. Depending on the location of the lesions or affected areas in the brain, MS patients can experience sensory, motor, or cognitive symptoms.

Four clinical courses of disease are now recognized in MS. First, relapsing- remitting MS (RR-MS) is the most common form of disease, and is the most treatable using currently available medications, most of which are focused on reduction of inflammation and inhibition of (auto)immune responses. RR-MS is characterized by acute episodes of inflammation and increased disability, after which the patient can recover nearly to the level of function they experienced prior to the inflammatory episode. Over time, however, the effects of these episodes take their toll, and disease

3 gradually worsens. The second type, primary progressive MS (PP-MS) is characterized by continual worsening of disease, but does not include the acute inflammatory episodes that transiently increase disability as in RR-MS. Secondary progressive MS (SP-MS) begins with a clinical course similar to RR-MS, but at some point converts to a disease course more similar to PP-MS, with continually worsening disease. Finally, progressive relapsing MS (PR-MS) is characterized by continually worsening disability, but does include acute episodes of inflammation that significantly increase disability and only partially recover following the episode. In all of these courses, the primary driver of disease progression and permanent disability is thought to be the degree of neuronal loss.

Even in the case of demyelination of axons and neuronal dysfunction, it seems possible that recovery and remyelination can occur as long as the axons are still intact. Thus, a clear need exists in the field of MS for neuroprotective treatments that will preserve axonal integrity and allow for remyelination and restoration of neuronal function.

NEUROPROTECTION

I would define neuroprotection as the maintenance of neuronal health and function despite threats from neuroinflammation or disease processes. In other words, it is important to provide overall support to the neuron to preserve its health, in addition to merely restoring function. A treatment could potentially drive a neuron to continue functioning and generating action potentials in spite of threatening disease conditions, but

4 this theoretically could lead to an overwhelming of the metabolic processes necessary for neuronal health. Indeed, the converse may also be true. For example, a treatment that reduces the excitability of a neuron may allow important cell processes such as the production of ATP to keep up with the demand; however, if the reduction of neuronal activity is continued for too long a time, activity-dependent support of the neuron in the form of growth factors could theoretically be reduced to levels detrimental to the neuron.

Thus, some treatments may be beneficial in the short term, but cannot be continued for an extended period of time without consequence.

ION CHANNELS IN NEUROTRANSMISSION

Neurons are responsible for communication in the nervous system. Sensory neurons receive input from peripheral stimuli and send nerve impulses through afferent pathways to the brain, where the axon terminals will provide input to numerous neurons.

The signal is then integrated and interpreted using the amazing computational ability of neural circuits in the brain, and a decision is made to act on this information. The brain then sends nerve impulses through efferent pathways back to the periphery, where the neurons are connected to effector tissues such as muscle, which responds to the neuronal input by contracting so that movement can occur.

Ion channels are essential for the generation and transmission of these nerve impulses. Neurons receive input along their dendrites in the form of neurotransmitters

5 released from the axon terminal of an upstream neuron. These neurotransmitters bind receptors on the postsynaptic membrane and (ionotropic) receptors open channels to allow ions to flow through the cell membrane. Sodium ions (Na+) flowing into the cell cause a local depolarization of the membrane, opening nearby voltage-gated sodium

(Nav) channels to allow further influx of Na+ into the cell. The membrane depolarization also opens voltage-gated potassium ion (K+) channels (Kv channels) to allow K+ to flow out of the cell and restore the membrane potential to its resting level. If enough depolarizing currents are generated on the neuronal dendritic and somatic cell membrane, they sum at the axon hillock. When a threshold level of depolarization is reached, the densely packed Nav channels at the axon initial segment (AIS) open to permit a large influx of Na+, generating an action potential (AP) which is then propagated along the axon by activation of adjacent Nav channels. The propagated action potential reaches the

2+ axon terminal, voltage-gated calcium ion (Ca )channels (Cav channels) open to allow influx of Ca2+ into the cell. This promotes the fusion of synaptic vessels containing neurotransmitters with the presynaptic membrane. The neurotransmitters contained in the vesicles are then released into the synaptic cleft for activation of receptors on the postsynaptic membrane of the next neuron. Axonal Kv channels also open to return the axon to its resting membrane potential so that another AP can be generated. Thus, action potential generation and propagation requires a balance of Nav and Kv channels in the axonal membrane for continued function. If many Nav channels are located along an axon with few Kv channels, the Kv channels cannot efficiently restore the membrane potential, and Nav channels remain in their inactivated state and cannot generate a new

6 action potential. If too many Kv channels are located on a portion of the axonal membrane, they hyperpolarize the membrane and Nav channels do not reach their threshold for activation, resulting in the blocking of conduction in the axonal membrane.

This is the major current hypothesis for the role of potassium channels in MS. 4- aminopyridine (4-AP or Ampyra®), a nonspecific Kv channel blocker currently approved for symptomatic treatment of MS, is thought to act by blocking potassium channels on the axonal membrane so that their hyperpolarizing influence is reduced and conduction of the action potential is no longer blocked.

ION AND WATER CHANNELS IN THE NEUROVASCULAR UNIT

Neurons do not function in isolation, but within the cellular context of the neurovascular unit, the interactive and supportive environment provided by glia

(astrocytes, microglia, and oligodendrocytes) and cells of the vasculature (endothelial cells, pericytes, and smooth muscle cells). Although ion channels likely play significant roles in the regulation of all of these cell types, I would like to focus here on the role of astrocytes, which are the most abundant glial cell type and which have a central role in neurovascular coupling – the bidirectional communication between neurons and the vasculature. Astrocytes have important roles in modulating neuronal function, recycling glutamate, maintaining ionic homeostasis, and buffering potassium. Ion and water channels are functionally active in these processes. Because the roles of channel proteins

7 in astrocyte function are less well-known than their roles in neuronal function, I will describe here in some detail the functions of several channel proteins that are expressed in astrocytes. I will characterize their roles in normal physiological function and in many diseases, including MS.

Astrocytes are a diverse cell population distributed throughout the CNS and are divided into several major groups. Protoplasmic astrocytes are found in gray matter areas of the CNS, and their processes contact synapses and blood vessel capillaries. Fibrous astrocytes are located in white matter, and their endfeet contact nodes of Ranvier and capillaries (Barres, 2008). Radial glial cells are a more specialized population of astroglia that have a radial morphology and are usually vimentin-positive (Sild and Ruthazer,

2011). These cells have important roles in development, but some cell types in this category are also present in adulthood, such as Bergmann glia in the cerebellum and

Muller cells in the retina (Sild and Ruthazer, 2011). In addition to serving as neuronal and glial precursor cells, they help to direct migration and synaptic development of newly generated neurons (Sild and Ruthazer, 2011). Continued studies in the astrocyte field will likely lead to the identification of new subpopulations of astrocytes with distinct properties and functions.

Astrocytic functions are regulated by a number of channel proteins, which are embedded in the cell membrane and allow the passage of water or ions into or out of the cells. Ion and water channels are very important to the normal physiology of many cell types, and defects in channel function can lead to diseases called channelopathies. Ion channels are especially recognized for their functions in neurons, including synaptic

8 transmission and the generation and propagation of action potentials, but they have many important roles in non-excitable cells as well, including astrocytes. The water channel

AQP4 regulates fluid homeostasis in the CNS, and is the target of a disease-causing autoantibody in neuromyelitis optica (NMO) (Lennon et al., 2004, Hinson et al., 2012,

Waters and Vincent, 2008), a disease with some similarities to MS (Barnett and Sutton,

2012). In the CNS, AQP4 is predominantly localized to astrocytic endfeet contacting

+ blood capillaries. The inwardly rectifying K channel, Kir4.1, is also present in astrocytic endfeet and is involved in K+ buffering (Higashimori and Sontheimer, 2007, Stephan et al., 2012, Bay and Butt, 2012). Kir4.1 is likely a target of the autoimmune response in a subgroup of MS patients (Srivastava et al., 2012). Furthermore, other K+ channels, including Ca2+-sensitive K+ channels and two-pore K+ channels, are also found in astrocytes (Armstrong et al., 2005, Benesova et al., 2009, Bouhy et al., 2011, Filosa et al.,

2006, Girouard et al., 2010). Voltage-gated Kv1.4 channels are highly upregulated in activated astrocytes around EAE (experimental autoimmune encephalomyelitis; an animal model of MS) lesions (Jukkola et al., 2012), although their functional role in this context remains unclear. Astrocytes can participate in cell-to-cell communication by propagation of Ca2+ waves (Charles et al., 1991, Cornell-Bell et al., 1990), which require channel proteins in the cell membrane and in the endoplasmic reticulum. Astrocytes can influence other cells through gap junctions and by the release of gliotransmitters to directly modulate synaptic function (Shigetomi et al., 2008, Halassa et al., 2007,

Nedergaard et al., 2003, Perea et al., 2009, Volterra and Meldolesi, 2005). Thus, several

9 of these channels play critical roles in neurovascular coupling and are potential targets for the treatment of neurological disorders.

ASTROCYTE CHANNEL PROTEINS

Water channel AQP4

Aquaporins are channel proteins allowing water to be transported through the plasma membrane. They are divided into three subgroups: (1) classical aquaporins

(AQP0, 1,2,4,5 and 8), which transport only water; (2) aquaglyceroporins (AQP3, 7, 9, and 10), which transport both water and various small molecules; and (3) unorthodox aquaporins (AQP6, 11, and 12), which are identified as aquaporins by sequence homology, but remain to be fully characterized (Rojek et al., 2008). Two isoforms (M1 and M23) have been discovered for AQP4, which differ only in the N-terminus by 22 amino acid residues (Umenishi and Verkman, 1998). The aquaporin α-subunit contains 6 transmembrane (TM) domains, along with two additional membrane-embedded domains that do not span the entire bilayer (Papadopoulos and Verkman, 2013). Although the monomeric α-subunit form contains the water pore, aquaporins most often form stable tetramers (Papadopoulos and Verkman, 2013). AQP4 function is regulated by channel gating, subcellular distribution, phosphorylation, protein-protein interactions and orthogonal array formation (Yukutake and Yasui, 2010). These orthogonal arrays are composed of aggregates of many AQP4 channels in one area of the cell surface to

10 increase membrane permeability. Interestingly, the center of the array contains the AQP4

M23 isoform, while the outside edge contains the M1 isoform (Rossi et al., 2012).

Aquaporins are expressed widely throughout the body, likely in every type of tissue. Some channels seem restricted to certain tissues, such as AQP2 which is primarily expressed in kidney, while AQP3 has been reported in at least 20 tissues and confirmed to have functional roles in about half of these (Rojek et al., 2008). Some tissues with high fluid-handling needs express multiple aquaporins, such as the kidneys and the eyes with eight aquaporins each (Hamann et al., 1998, Radin et al., 2012). AQP1, 4, and 9 have known functions within the CNS (Papadopoulos and Verkman, 2013). AQP4 is highly expressed in astrocytic endfeet contacting both blood vessels and neuronal synapses, and is involved in regulation of the flow of water and solutes into and within the CNS. It is involved in bulk flow of cerebrospinal fluid into the brain parenchyma and helps to regulate the size of particles that can pass the BBB (Iliff et al., 2012). AQP4 is expressed at higher levels in more excitable brain regions than less excitable regions (Aoyama et al.,

2012). This is likely because more excitable areas have higher metabolic demand. AQP4 is a major player in neurovascular coupling with regulatory effects both at the blood vessels and at neuronal synapses.

Polarized expression of AQP4 in astrocytic endfeet allows for efficient regulation of water in the extracellular spaces adjacent to vascular and synaptic structures. AQP4 control of extracellular water content helps to regulate ionic concentrations, influencing

K+ buffering and calcium signaling. AQP4 KO mice may have improved function in cytotoxic models of edema, but perform worse in models of edema focused on blood

11 vessel fluid leak (Verkman et al., 2006). AQP4 dysfunction leads to impaired water and solute handling abilities, and may contribute to the abnormal buildup of water (edema) or proteins that occurs in many neurological disorders (Iliff et al., 2012). Furthermore,

AQP4 KO mice have impaired memory formation and reduced hippocampal synaptic plasticity (Fan et al., 2013, Scharfman and Binder, 2013, Skucas et al., 2011).

AQP4 has been identified as the target of autoantibodies produced in NMO, formerly considered a variant of MS but now recognized as a distinct disease (Barnett and Sutton, 2012). AQP4-specific autoantibodies are produced concurrently with disease onset and have direct involvement in the disease pathology of NMO (Bennett et al.,

2009). These antibodies target one of multiple binding sites located in the three extracellular loops, the intracellular loops, or the N- and C-termini. The binding can occur with monomeric or tetrameric AQP4 (Iorio et al., 2013). This produces an immune response that results in damage to astrocytes and breakdown of the BBB, followed by myelin loss and oligodendrocyte apoptosis (Brosnan, 2013, Hinson et al., 2010, Jarius and Wildemann, 2010). AQP4 may also be a mediator of MS disease pathology, since

EAE susceptibility is almost eliminated in AQP4 KO mice (Li et al., 2011).

Interestingly, secretion of proinflammatory cytokines was also reduced from AQP4 KO astrocytes in vitro (Amiry-Moghaddam et al., 2005). In contrast, AQP4 deficiency significantly increased the extent of neuron loss and demyelination, and caused increased motor dysfunction in a spinal cord contusion injury model (Kimura et al., 2010). As might be expected from the changes in synaptic plasticity, AQP4 deficiency also affects seizure activity. AQP4 KO mice have a higher threshold for seizure activity induced by

12 either pentylenetetrazole injection or by electrical stimulation (Binder et al., 2004, Binder and Steinhauser, 2006), but have a longer duration of generalized tonic-clonic seizures, likely due to altered K+ handling (Binder et al., 2006). Perivascular loss of AQP4 in the hippocampus has been associated with temporal lobe epilepsy (Eid et al., 2005). Despite some progress, the physiological and pathological functions of aquaporins remain to be fully understood.

Inwardly rectifying potassium channel Kir4.1

The K+ channel superfamily is encoded by about 80 genes, and can be categorized into three major families based on structure and function (Gu and Barry, 2011). The

+ voltage-gated K (Kv) channel family consists of 12 different subfamilies denoted Kv1-

12, encoded by about 40 separate genes (Gu and Barry, 2011, Gu et al., 2012); (Gu and

Gu, 2010, Gu and Gu, 2011, Barry, 2013, Gu et al., 2006, Xu et al., 2007b, Xu et al.,

2010). Kv channel α-subunits have 6 TM segments, with one pore-forming loop located between segments 5 and 6. Four α-subunits tetramerize to form a functional potassium- selective channel. As a general rule, K+ channel α-subunits can form channel complexes with other members of the same subfamily. An additional specialized subfamily, the Slo family, can be grouped with the Kv channels, although these channels can be activated by intracellular signaling as well as the membrane potential (Gu and Barry, 2011). The Slo family includes the Maxi-K (or BK) channel, or Slo1, which is activated by both voltage

2+ + and Ca (Gu and Barry, 2011). The second family, inwardly rectifying K (Kir) channels, is composed of 7 subfamilies (Gu and Barry, 2011, Hibino et al., 2010). The α-subunits

13 of these channels have only two TM segments, which are connected by a pore-forming loop (Gu and Barry, 2011). Kir channel α-subunits also form tetramers containing a single potassium-selective pore, and conduct current more efficiently in the inward direction

+ (Gu and Barry, 2011, Hibino et al., 2010). The third family, two-pore domain K (K2p) channels, has 15 members which can be divided into six subfamilies (Enyedi and Czirjak,

2010, Gu and Barry, 2011). The K2p α-subunit has four TM segments and two pore- forming loops, and two subunits dimerize to form a channel complex with dual potassium-selective pores (Enyedi and Czirjak, 2010, Gu and Barry, 2011). Most K2p channels are functionally characterized by currents having little voltage or time dependence, also known as “leak” K+ currents (Gu and Barry, 2011).

Kir channels are inwardly rectifying channels that preferentially allow K+ ions to flow into the cell. The Kir α-subunit contains two transmembrane domains and one pore- forming loop, and four α-subunits homo- or hetero-tetramerize to form a functional channel. Fifteen Kir α-subunits have been identified in humans and rodents, and these are divided into seven subfamilies (Kir1-Kir7). Channels of the Kir2 subfamily are strong inward rectifiers and are considered the “classical” Kir channels (Hibino et al., 2010). Kir3 channels are also strong rectifiers, but are distinguished by their gating by G-protein- coupled receptors (Hibino et al., 2010). Channels in the Kir4, 5, and 7 families are intermediate inward rectifiers, and are known as “transport” channels along with Kir1.1 which is a weak inward rectifier (Hibino et al., 2010). Kir6 channels are also weak rectifiers, and are sensitive to ATP (Hibino et al., 2010). Here we focus on the role of

Kir4.1, an intermediate inward rectifier expressed in astrocytes.

Kir2 channels are localized in skeletal and cardiac muscle cells and smooth muscle cells of the vasculature, and in neurons throughout the brain at low levels. Kir3 channels are also expressed throughout the brain. Kir6 channels are expressed in the pancreas, heart, vascular smooth muscle, and certain areas of the brain such as hypothalamus. Kir1.1 can be found in kidney, and Kir7.1 is expressed in the retina and choroid plexus (Hibino et al., 2010). Kir4.1 channels are expressed in the CNS and localized in astrocytic endfeet, and so are well-situated to remove K+ ions from extracellular spaces adjacent to blood vessels and synapses, in what is often referred to as

+ a “K buffering” role. A loss of Kir4.1 function leads to neuronal dysfunction and seizure activity, as shown by several different studies. In animal models, mice with a Kir4.1 genetic polymorphism causing increased seizure susceptibility were shown to have impaired astrocytic K+ and glutamate uptake (Inyushin et al., 2010). A conditional KO under control of a GFAP promoter confirmed that this effect was mediated by astrocytic

+ Kir4.1, also showing impaired K and glutamate uptake, ataxia, seizures, and premature death (Djukic et al., 2007). Synaptic function was also affected, with the Kir4.1 cKO causing enhanced short-term potentiation (Djukic et al., 2007). Thus, the function of astrocytic Kir4.1 is clearly coupled to neuronal activity. Although Kir4.1 and AQP4 are both expressed in astrocytic endfeet, they appear to have no direct functional interaction.

Kir4.1 channel expression and function were no different in AQP4 KO mice (Ruiz-Ederra et al., 2007, Zhang and Verkman, 2008), and inhibition or knockdown of Kir4.1 caused no change in the water permeability of AQP4 (Zhang and Verkman, 2008). Also, no

15 change in AQP4 expression was observed in Kir4.1 KO mice (Hsu et al., 2011). However,

AQP4 may support K+ buffering by altering the concentration in the extracellular space.

Several diseases are known to be related to Kir4.1. Hippocampal sclerosis in temporal lobe epilepsy has been shown to be associated with loss of Kir4.1 in astrocytic endfeet (Heuser et al., 2012). In animal models of epilepsy, disruption of the BBB led to serum albumin leakage into the CNS and uptake in astrocytes, mediated by the TGFβ receptor. This led to downregulation of Kir4.1 and the development of neuronal hyperexcitability and epileptiform activity (Cacheaux et al., 2009, Ivens et al., 2007).

+ Reduced Kir4.1 activity likely leads to increased extracellular K and glutamate, contributing to brain excitability and epileptogenesis (de Lanerolle et al., 2010). Kir4.1 has also been shown to have effects in demyelinating conditions. A spinal cord injury model showed up to 80% reduction in Kir4.1 expression in the lesion area, and demonstrated that a previously discovered neuroprotective agent, 17-β-oestradiol, could partially restore Kir4.1 expression and function (Olsen et al., 2006). Also, a recent study of serum IgG identified Kir4.1 as the target of an autoantibody present in about half of the

MS patients studied (Srivastava et al., 2012). When Kir4.1 serum antibodies were injected into mice, it caused complement activation in Kir4.1-expressing areas and led to altered expression patterns of GFAP in the cerebellar molecular layer (Srivastava et al., 2012), showing that the autoantibody was indeed pathogenic. Kir4.1 KO mice have impaired myelination and oligodendrocyte maturation in development, leading to neuronal degeneration and a poor survivability to only about 2-3 weeks of age (Neusch et al.,

2001). Kir4.1 has not yet been studied in EAE models, but would be an interesting target

16 for future study in wildtype and Kir4.1 inducible-KO or astrocyte-specific (GFAP- conditional) KO mice.

Ca2+ and Voltage-activated BK channels

Maxi-K (BK or Slo1) channels are large-conductance K+ channels that are activated by voltage and intracellular calcium to allow efflux of K+ ions from the cell.

The α-subunit has a similar structure to Kv channels, but contains a 7th TM domain located at the N-terminal end of the protein. The BK channel also has a large carboxyl terminus which mediates its gating by intracellular Ca2+ (Salkoff et al., 2006). Four subunits tetramerize to form a functional ion channel. Four β-subunit genes exist for the

BK channel, which are important in regulating channel trafficking and function. Other members grouped within this subfamily are Slo2.1, Slo2.2, and Slo3, which are also large-conductance channels, but are regulated by Na+ (Slo2.1, 2.2), ATP (Slo2.1), or pH

(Slo3) rather than Ca2+ (Salkoff et al., 2006). Slo3 channel α-subunits contain 7 TM domains and one pore-forming loop similar to BK channels, while Slo2 channels contain only 6 TM domains and one pore-forming loop (Salkoff et al., 2006).

BK channels are expressed in neurons, often in axons and presynaptic terminals.

In the cerebellum, they are expressed in Purkinje cell soma, in cerebellar basket cell terminals, and in the molecular layer (Misonou et al., 2006). In astrocytes, they are primarily located in the endfeet. In addition to α-subunits, BKβ4 (strong expression) and

BKβ1 (weak expression) subunits are expressed in astrocytes in some regions (Seidel et

17 al., 2011). A recent study suggests that spontaneous Ca2+ oscillations keep BK channels activated and increase basal extracellular (perivascular) levels of K+ (Koide et al., 2012).

The roles of BK channels in disease processes are only beginning to be elucidated. A genetic study associating a BK channel mutation with paroxysmal dyskinesia and generalized epilepsy showed increased channel function (Du et al., 2005), and the corresponding mutation in mice produced similar results (Wang et al., 2009). It is not known whether these mutations affect the function of BK channels expressed in astrocytes. Nonetheless, BK channel expression is known to be altered in several diseases. A chemoconvulsant-induced seizure model was found to increase BK activity, resulting in elevated excitability of cortical pyramidal neurons (Shruti et al., 2008).

Interestingly, in the pilocarpine model of epilepsy, the BK channel is downregulated significantly, perhaps as a compensatory mechanism (Ermolinsky et al., 2008, Pacheco

Otalora et al., 2008). The BKβ4 auxiliary subunit regulates both the voltage- and Ca2+- gating sensitivities and the membrane trafficking of BK channels (Shruti et al., 2008)

(Wang and Brenner, 2006), and BKβ4 KO mice develop spontaneous seizures (Brenner et al., 2005). In addition, a recent study found that BK channels are reduced along with

AQP4 and Kir4.1 levels in astrocytic endfeet both in AD patient tissues and in amyloid precursor protein transgenic mice (Wilcock et al., 2009). Astrocytic BK channels do appear to have a significant role in neurovascular coupling, likely through regulation of

+ the extracellular K level in conjunction with Kir4.1 in order to control vasodilation or vasoconstriction. Ca2+ concentration in the astrocyte endfeet is known to affect vasodilation and vasoconstriction (Girouard et al., 2010), and this is mediated at least in

18 part by BK channel activation by Ca2+ (Girouard et al., 2010, Koide et al., 2012).

Interestingly, in a model of subarachnoid hemorrhage, BK channel activation instead produced vasoconstriction, suggesting a biphasic vascular response to K+ levels (Koide et al., 2012). Low extracellular concentrations of K+ produce vasodilation, while high concentrations cause vasoconstriction. This is due to the activity of another inward

+ rectifier K channel, Kir2.1, expressed in vascular smooth muscle cells (Filosa et al.,

2006). These inward-rectifier K+ currents hyperpolarize the smooth muscle cell and close voltage-dependent calcium channels to reduce the SMC Ca2+ concentration and cause vasodilation (Girouard et al., 2010), since calcium binding is necessary for smooth muscle contraction. More study is needed to determine whether BK channels within the neurovascular unit have a causative role in neurological disorders.

Two-pore-domain K+ channels

Two-pore domain potassium channel α-subunits are composed of four TM segments and two pore-forming loops. They dimerize to form two K+ permeable pores.

They contribute to “leak currents” and help to set the resting membrane potential of the cell. The 15 known members of this family are named for their defining characteristic:

TWIK (Tandem-pore-domain Weak Inwardly-rectifying K+ channel), THIK (Halothane-

Inhibited), TREK (TWIK-RElated), TRAAK (Arachidonic Acid-stimulated), TASK

(Acid-Sensitive), TALK (ALkaline pH activated), TRESK (TWIK-RElated Spinal). As the names imply, two-pore domain K+ channels can be modulated by a number of

19 different factors, including pH, neurotransmitters, anesthetics, osmolarity, and oxygen tension.

Two-pore domain K+ channels are localized in both neurons and glia in the CNS.

They are expressed throughout astrocytes and are not polarized to endfeet. TASK-1, -2, and -3 are all expressed in cerebellar astrocytes (Rusznak et al., 2004). TWIK-1 and

TREK-1 are highly expressed in hippocampal astrocytes, while TASK-1 is more weakly expressed (Zhou et al., 2009). These channels, with TREK-1 being the greatest contributor, mediate a large portion of the passive membrane K+ conductance that maintains the resting membrane potential of mature hippocampal astrocytes (Zhou et al.,

2009). TREK-1 channels can be modulated by GPCRs, which apparently render them permeable to glutamate and thus promote glutamate release from astrocytes (Woo et al.,

2012). In neurons, TWIK-1 can also be activated by a GPCR through the 5-HT1A receptor and the Gαi/PKA mediated signaling pathway, allowing serotonin to reduce neuronal excitability (Deng et al., 2007). All members of the TREK family may be involved in neuroprotection, modulated by neuroprotective factors including riluzole, polyunsaturated fatty acids and lysophospholipids (Mathie and Veale, 2007).

Two-pore domain K+ channels may be involved in several diseases, including

MS. Both TASK-1 deficiency and pharmacological inhibition of TASK-1 were shown to reduce the severity of EAE and impair T cell proliferation and cytokine production

(Bittner et al., 2009). TASK-1 KO mice perform worse in ischemia, resulting in larger lesion volumes, but it is unclear whether this effect is mediated within the brain or by altered blood pressure due to loss of TASK-1 in the heart, adrenal gland, or vasculature

(Muhammad et al., 2010). A TASK-3 mutation was identified as the cause of a mental retardation dysmorphism syndrome and completely abolished channel currents both in the homodimeric form and when heterodimerized with TASK-1 (Barel et al., 2008). A sheep model of fetal alcohol syndrome caused significant cerebellar Purkinje cell loss due to ethanol induced changes in pH, and this effect was abolished when TASK-1 and

TASK-3 were blocked concurrently with ethanol administration (Ramadoss et al., 2008).

In the pilocarpine model of epilepsy, TASK-1 and TASK-2 are both upregulated in hippocampal astrocytes (Kim et al., 2009, Kim et al., 2008). Interestingly, TASK-2 expression increases in the astrocytic endfeet as well as the soma, which may indicate dysregulation of the BBB (Kim et al., 2009). Further study may bring to light significant roles of other two-pore channels in health and disease pathology.

Voltage-gated calcium channels

Voltage-gated calcium (Cav) channels are synthesized as a single α-subunit protein containing four homologous domains that contain six TM domains each. There are ten known α-subunit genes in the Cav family, which have been divided into three subfamilies, Cav 1-3. They also have classical labels based on properties and originally discovered location. The Cav 1 family contains four L-type (“Long-lasting” conductance) channels that are important for gene transcription involving calcium-mediated signal pathways (Lehmann-Horn and Jurkat-Rott, 1999) (Zamponi et al., 2009). The Cav 2 family contains P-type (“Purkinje cell”) channels (Cav 2.1) and N-type (“Neuronal”) channels (Cav 2.2) that are important for neurotransmitter release from synaptic

21 terminals, as well as R-type (“toxin Resistant”) channels (Cav 2.3) (Lehmann-Horn and

Jurkat-Rott, 1999) (Zamponi et al., 2009). The Cav 3 family contains three T-type

(“Transient” current) channels that are important for modulation of neuronal excitability and firing patterns (Lehmann-Horn and Jurkat-Rott, 1999, Zamponi et al., 2009). L-, P- and N-type channels are all activated by high voltages, while R-type channels are activated by intermediate voltages and T-type channels by low voltage (Lehmann-Horn and Jurkat-Rott, 1999). In addition, most Cav channel α-subunits exist in complex with other subunits that regulate their activity: α2δ, β, and γ subunits (Buraei and Yang, 2013).

The four α2δ subunit genes encode two subunits each, α2 and δ (Buraei and Yang, 2013).

Four different β- subunit genes are known to exist, and eight γ-subunit genes (Buraei and

Yang, 2013). In addition to voltage, calcium channels can be regulated by G-protein coupled receptors, either by direct interaction or through cAMP-mediated signaling

(Zamponi et al., 2009).

Cav channels are located primarily in neurons in the CNS. The Cav 1 and Cav 3 families are primarily expressed somatodendritically, while Cav 2 family members are expressed in synaptic terminals. Cav channels have no or low-level expression in astrocytes under normal conditions, but they may be highly induced or upregulated in reactive astrocytes. Interestingly, Cav channels in neurons can assemble closely with BK channels in macromolecular complexes where they become the rate-limiting step in BK channel activity (Berkefeld and Fakler, 2013). It is not known whether this coexpression and channel interaction occurs in the astrocyte endfeet.

Calcium signaling is the major form of excitability or activity for astrocytes, and can directly affect neurotransmission by release of glutamate from astrocyte endfeet contacting synapses. Besides the influx of calcium through voltage-gated calcium channels at the cell membrane, calcium signaling in astrocytes also involves intracellular mechanisms of calcium release, termed store-operated calcium release. This mechanism is calcium-dependent, so that a small current passing through the plasma membrane can be amplified by intracellular calcium release from the endoplasmic reticulum mediated by inositol triphosphate receptors (IP3R) (Leybaert and Sanderson, 2012). Furthermore, this calcium release can self-propagate by further stimulation of adjacent IP3Rs both within the cell and intercellularly through gap junctions in a travelling “calcium wave” that can modulate activity at endfoot processes over a wider area (Leybaert and

Sanderson, 2012). Interestingly, calcium signaling can be differentially regulated at the level of the astrocyte process (Arizono et al., 2012); Calcium signaling can therefore be matched to needs of surrounding neurovascular unit at the process level. As shown by two-photon imaging, this local Ca2+ signaling in astrocyte processes modulates neurotransmission in nearby synapses (Di Castro et al., 2011), and is dependent upon inositol triphosphate receptor 2 (IP3R2)-mediated intracellular calcium release. Astrocyte

Ca2+ signaling is required for cholinergic-induced synaptic plasticity (LTP), also mediated by IP3R signaling, and results in glutamate release to act upon synaptic mGluR’s (Navarrete et al., 2012). Calcium signaling can also be affected by AQP4- mediated changes in extracellular osmolarity (Thrane et al., 2011).

Neuronal Cav channels are involved in many neurological diseases, and disease- causing mutations have been mapped throughout the α1-subunit in various channels

(Cain and Snutch, 2011). Cav1 (L-type ) channel mutations have been linked to retinal dysfunction and to disorders of cardiac and skeletal muscle (Brinkmeier et al., 2000).

Cav3 (T-type) channel mutations have been linked to autism and epilepsy, while Cav2.1

(P-type) channel mutations are associated with ataxia and migraine (Cain and Snutch,

2011). Although these are likely mediated by Cav channel dysfunction in neurons, astrocytes may also play a role in disease processes. Cav1.3 (L-type) and Cav2.1 (P-type) channel expression is induced in hippocampal reactive astrocytes following pilocarpine- induced status epilepticus (Xu et al., 2007a). In APP transgenic mice, Cav1.2 channels are induced in reactive astrocytes surrounding amyloid plaques, and their expression increases with age of the mice and the number of plaques (Willis et al., 2010). The potential beneficial or detrimental aspects of upregulation of Cav channels in astrocytes have not yet been fully explored, but could alter neuronal responses over a wide area through intracellular and intercellular Ca2+ wave activity.

PERSPECTIVES FOR ASTROCYTIC ION CHANNEL RESEARCH

Recently identified channel proteins expressed in astrocytes

In this review, we have discussed the astrocytic channels known to have important effects in neurovascular coupling, but many other channels exist that also could

24 play important roles in the neurovascular unit (Table 1). Kv1.6 expression has been reported in astrocyte culture (Smart et al., 1997) but does not appear to be altered in reactive astrocytes (Wilcock et al., 2009). Voltage-gated sodium (Nav) channels Nav1.1,

1.2, 1.3, and 1.6 are expressed at low levels in astrocytes (Black et al., 2010). Of these, only Nav1.2 is slightly upregulated in reactive astrocytes in MS lesions, but Nav1.5 is induced and highly upregulated in reactive astrocytes in MS lesions, cerebrovascular accidents, and brain tumors (Black et al., 2010). Transient receptor potential (TRP) channels are permeable to more than one cation, and TRPV1 (Mannari et al., 2013),

TRPV4 (Butenko et al., 2012, Dunn et al., 2013), and TRPC1 (Reyes et al., 2013) channels have recently been reported to be expressed in astrocytes. TRPV4 is involved in astrocytic calcium signaling (Butenko et al., 2012, Dunn et al., 2013) and appears to be upregulated in reactive astrocytes in ischemic conditions (Butenko et al., 2012). TRPV4 may also interact with AQP4 function (Benfenati et al., 2011). Finally, we discuss gap junctions expressed in astrocytes. Connexins are proteins which hexamerize to form a

“hemichannel” in the cell membrane which allows the passage of ions and molecules smaller than about 1 kD into or out of the cell (Laird, 2006, Giaume et al., 2013) . When coupled to a hemichannel on an adjacent cell, a gap junction is formed, which allows molecules to pass between the cytoplasm of the two cells. The major connexins expressed in astrocytes are Cx43 and Cx30, and these are thought to have a role in several diseases including multiple sclerosis and Alzheimer’s disease (Koulakoff et al., 2012, Masaki et al., 2013) (Table 1). As these and other channels are more fully studied, they may be found to have important roles in neurovascular coupling.

Resting Astrocytes Activated Astrocytes Water channel AQP4(Iliff et al., 2012, Aoyama et al., 2012, Fan AQP4 (Jukkola et al., 2013) et al., Papadopoulos and Verkman, 2013) Inward Rectifier K+ channels

Kir4.1(Djukic et al., 2007)Heuser et al., 2012) Kir2.3(Perillan et al., 2002, Seifert et al., 2009) Kv channels

BK channel (Wilcock et al., 2009). Kv1.4(Jukkola et al., 2012)

Kv1.6(Smart et al., 1997) Two-pore domain K+ channels TWIK-1(Zhou et al., 2009) TREK-1(Zhou et al., 2009) TASK-1(Rusznak et al., 2004) TASK-1(Rusznak et al., 2004) TASK-2(Rusznak et al., 2004) TASK-2(Rusznak et al., 2004) TASK-3(Rusznak et al., 2004) Cav channels

Cav1.2(Willis et al., 2010)

Cav1.3(Xu et al., 2007a)

Cav2.1(Xu et al., 2007a) Nav channels

Nav1.1(Black et al., 2010) Nav1.2(Black et al., 2010)

Nav1.2(Black et al., 2010) Nav1.5(Black et al., 2010)

Nav1.3(Black et al., 2010)

Nav1.6(Black et al., 2010) TRP channels TRPV1(Mannari et al., 2013) TRPV4(Benesova et al., 2009, Amici et al., 2007) TRPV4(Amici et al., 2007) TRPC1(Reyes et al., 2013) Gap junctions Connexin 30(Koulakoff et al., 2012) Connexin 43(Masaki et al., 2013) Table 1. Channel proteins expressed in astrocytes

Regulation of astrocytic channel functions by intracellular signaling pathways

In the past, reactive astrocytes have been viewed in a negative light, as proinflammatory cells that form glial scars which hinder remyelination and axon growth.

We now know from both in vivo and in vitro studies that activated astrocytes also mediate CNS myelination by promoting the migration, proliferation, and differentiation of oligodendrocyte progenitor cells (Ishibashi et al., 2006, Nash et al., 2011a, Sorensen et al., 2008, Watkins et al., 2008). Astrocytes promote myelination in response to electrical activity by releasing the cytokine leukemia inhibitory factor (LIF) (Ishibashi et al., 2006).

Other factors produced from astrocytes can also impact myelination (Moore et al., 2011), including platelet-derived growth factor (PDGF), fibroblast growth factor (FGF), ciliary neurotrophic factor (CNTF), and insulin-like growth factor (IGF), as well as gliotransmitters such as glutamate and ATP. Because ion channels are clearly involved in myelination-promoting functions of astrocytes, especially gliotransmission and the modulation of electrical activity, the role of astrocytic ion channels in myelination will be an important topic for future research.

Astrocytes are now recognized as an important player in MS, ALS, AD, epilepsy, stroke, spinal cord injury, and other neurological disorders (Sofroniew and Vinters,

2010). Astrocytes modify their morphology and function in response to neuroinflammation or injury by altering expression of many genes, including ion channel genes. These changes are regulated by specific signaling pathways that produce a measured, context-dependent response to the insult (Sofroniew and Vinters, 2010).

Although they normally do not fire action potentials, astrocytes express Na+ channels and

K+ channels (Black et al., 1998, Black et al., 1994a, Black et al., 1994b, Djukic et al.,

2007, Neusch et al., 2006, Olsen et al., 2006, Schaller et al., 1995, Zhou et al., 2009).

Many ion channels are also upregulated in activated astrocytes (Table 1).Further study of the mechanisms and functional importance of ion channel alterations in reactive astrocytosis will be vital for the understanding and treatment of many neurological diseases.

Relatively little is known about the factors that regulate the targeting of astrocytic channels and the induction of channel expression in reactive astrocytes. In vitro studies of protein kinase C epsilon (PKCε) show that this kinase can upregulate Cav channels in astrocytes (Burgos et al., 2007b), and can alter cytoskeletal protein expression to induce a stellate morphology which may be analogous to reactive gliosis (Burgos et al., 2007a,

Burgos et al., 2011). It would be interesting to determine what other transcription factors and signal pathways are activated in reactive astrocytes, which might in turn induce or upregulate channel expression or trafficking. The localization of channel proteins to specific subcellular areas such as the endfeet may have important roles in normal astrocyte function and in reactive astrocytosis. Some of the details of trafficking and anchoring of AQP4 and Kir4.1 have been shown (Noell et al., 2011, Potokar et al., 2013,

Steiner et al., 2012), but the targeting mechanisms of other channels are unexplored in astrocytes and may be an important topic for future research. It would also be interesting to determine whether astrocytic channel activity itself can induce or regulate signaling pathways, perhaps promoting the upregulation of other channels during reactive

28 astrocytosis. The research field of astrocyte channel regulatory mechanisms is wide-open with exciting potential for future investigation.

SUMMARY

The function or dysfunction of astrocytes within the neurovascular unit may have broad implications for health and disease pathology. The neurovascular unit is known to be perturbed in many devastating diseases, and many questions remain about the role of reactive astrocytes in the progression of, or the response to, disease processes. Astrocytic ion and water channels are situated at key interfaces between astrocytes and either neurons or blood vessels. Understanding the mechanisms of astrocytic channel activity will promote the research and discovery of potential targets for drug therapy in diseases involving neurovascular abnormalities, and will inform the strategies necessary to restore health within the neurovascular unit.

OVERALL GOAL

The previous discussion highlighted the importance of ion and water channels to both neuronal and astrocytic function. I have discussed the relationship of myelin to ion channel localization, and have shown that ion and water channels have a relationship to many neurobiological diseases, including MS. I have also introduced potential roles of reactive astrocytes in restoring neuronal health, and potential involvement of ion channels in this process. Therefore, I sought to explore the role of potassium ion and water channel targeting and expression patterns in the CNS in response to inflammatory demyelination in EAE.

CHAPTER 2: K+ CHANNEL ALTERATIONS IN EAE

INTRODUCTION

MS is an inflammatory, demyelinating disease of the CNS that is thought to have an autoimmune pathogenesis. Myelin damage can impair action potential (AP) propagation along axons, leading to various neurophysiological abnormalities. Ion channels are implicated in neuronal dysfunction in MS, such as axon conduction failure and axonal degeneration (Waxman, 1982, Kornek et al., 2001, Waxman, 2002, Judge and

Bever, 2006, Waxman, 2006). In particular, axonal Kv channels, to hyperpolarize membrane potentials towards the resting level, play critical roles in regulating the initiation, waveform, frequency and uni-directional propagation of APs (Hodgkin and

Huxley, 1952, Hille, 2001, Debanne, 2004, Gu and Barry, 2011). Blocking axonal Kv channel activity can enhance axon conduction, which may underlie the beneficial effects of 4-AP in symptomatic treatment of MS, such as improving balance, vision, walking speed and leg strength, etc (Hayes, 2004, Judge and Bever, 2006, Goodman et al., 2009,

Espejo and Montalban, 2012, Goodman et al., 2010). However, the treatment does not

31 cure MS and has side effects. This appears consistent with complex expression and targeting patterns of Kv channels.

Kv1 channels are a major subfamily of axonal Kv channels present in both the

CNS and peripheral nervous system (PNS). Kv1.1, Kv1.2, and Kvβ2 are clustered in the

JXP regions under the myelin sheath along myelinated axons in brain and spinal cord

(Wang et al., 1993, Rhodes et al., 1995, Wang et al., 1995, Rhodes et al., 1997, Rasband et al., 1999, Vabnick et al., 1999, Rasband and Shrager, 2000, Trimmer and Rhodes,

2004). These channels constrain repetitive firing of APs in normal myelinated axons

(Zhou et al., 1998), and reduce axonal excitability in pre-myelinated axons during early development and demyelinated axons in diseases (Rasband et al., 1998, Sinha et al.,

2006). Furthermore, they are also present in unmyelinated axons, such as cerebellar basket cell terminals (Wang et al., 1994, Rhodes et al., 1997). Kv1.4, carrying transient currents, resides in unmyelinated hippocampal mossy fibers and in small diameter and unmyelinated axons of dorsal root ganglion neurons, but not in the JXP regions of large diameter and myelinated axons (Sheng et al., 1992, Cooper et al., 1998, Rasband et al.,

2001). In contrast to axonal Kv1 channels, Kv2 channels are mainly localized in the somatodendritic regions of projection neurons, such as cortical pyramidal neurons and spinal cord motor neurons (Lim et al., 2000, Muennich and Fyffe, 2004), and therefore their targeting should not be regulated by myelin.

EAE is widely used as an animal model of human CNS demyelinating diseases, including MS. Although EAE and MS are different diseases, they share many aspects in disease pathogenic processes and EAE has been used as a model to successfully develop

32 several effective treatments for MS (Lublin et al., 1987, Zamvil and Steinman, 1990,

Steinman, 2005, Slavin et al., 2010). The chEAE and rrEAE models are induced with different myelin peptides in different mouse strains to mimic the progressive and remitting-relapsing types of MS, respectively (Youssef et al., 2002). Therefore, EAE is a good model to determine the role of Kv channels in the progression of inflammatory demyelination. In particular, the changes at the remitting stage may provide important clues for neuronal survival and remyelination.

In this study, we systematically examined the expression and targeting patterns of several key Kv channels in the CNS during the progression of chEAE and rrEAE. Our results demonstrate that they are differentially altered in various CNS cell types during

EAE progression, and hence provide important insights regarding effects of Kv channel blockers on immune-mediated demyelination.

METHODS

Reagents and Antibodies

The nuclear dye (Hoechst 33342) and the lipophilic dyes (Fluoromyelin-green and

Fluoromyelin-red) were purchased from Invitrogen (Carlsbad, CA, USA). The following antibodies were used in our study, rabbit polyclonal anti- Kv1.2, anti- Kvβ2, anti- Kv1.4, and anti- Kv2.1 antibodies (Alomone Labs, Jerusalem, Israel), anti- Kv1.2 and anti- contactin associated protein 2 (Caspr2) antibodies (Millipore, Billerica, MA, USA), and

33 anti-neurofilament 200 antibody (Sigma, , St Louis, MO, USA), mouse monoclonal anti-

Kv1.2, anti- Kvβ2, and anti- Kv1.4 antibodies (clone #: K14/16, K17/70, and K13/31, respectively; UC Davis/NIH Neuromab Facility, Davis, CA, USA), anti-β-tubulin

(Millipore), rat monoclonal anti-myelin basic protein (MBP) antibody (Chemicon,

Temecula, CA, USA), goat polyclonal anti-GFAP and anti-NG2 antibodies (AbCAM,

Cambridge, MA, USA), and Dylight 488-, Dylight 649-, Cy3-, and Cy5-conjugated secondary antibodies (Jackson ImmunoResearch Laboratories, West Grove, PA, USA).

Myelin oligodendrocyte glycoprotein (MOG) peptide 35-55

(MEVGWYRSPFSRVVHLYRNGK) was purchased from Pro-Spec (Rehovot, Israel), and proteolipoprotein (PLP) peptide 139-151 (HCLGKWLGHPDKF) from Anaspec

(Fremont, CA, USA).

Induction of chronic EAE

Complete Freund's Adjuvant (CFA) was prepared by adding ground inactivated mycobacteria tuberculosis H37Ra (Difco Laboratories, Detroit, MI, USA) to Incomplete

Freund's Adjuvant (Difco Laboratories) at 4 mg/mL. CFA was passed through a glass syringe with an equal volume of sterile-filtered PBS (control) or PBS containing myelin oligodendrocyte glycoprotein (MOG) peptide 35-55 (1 mg/mL final concentration). 12- week-old female C57BL/6J mice (Jackson Laboratories, Bar Harbor, ME, USA) were immunized with 100 μL of MOG/CFA or CFA only (control) by subcutaneous injection at four sites in the belly and in each hind footpad. Pertussis toxin was injected at 0 and 2 days post immunization (DPI). Thy1-YFP transgenic mice (Jackson Laboratories) were

34 maintained on a C57BL/6 background, and thus were immunized with the same protocol.

Control mice received the same injection of CFA/PBS without the peptide. The chEAE experiment was performed 5 times, 10-17 mice per experiment. Clinical scores were assigned on a scale of 0-6 [0 = no symptoms, 1 = loss of tail tone, 2 = hindlimb paresis, 3

= moderate paralysis, 4 = paraplegia (complete hindlimb paralysis), 5 = quadriplegia, 6 = death or moribund state]. Grade 6 animals were removed from the study.

Induction of remitting-relapsing EAE

Solution preparation and immunization of 12-week-old female SJL/J mice

(Jackson Laboratories) for rrEAE were performed identically to chEAE except for the use of PLP peptide 139-151. Control mice received the same injection of CFA/PBS without the peptide. The rrEAE experiment was performed 3 times, 12-15 mice per experiment.

Cardiac perfusion and tissue fixation

Animals were deeply anesthetized with 250 mg/kg avertin (12.5 mg/mL 2,2,2- tribromoethanol dissolved in ddH20/0.025% 2-methyl-2butanol) (Sigma). The thoracic cavity was opened to expose the heart, and the mice were perfused with 20-30 mL ice- cold PBS followed by 20 mL 4% formaldehyde in PBS. The brain and spinal cord were carefully removed and post-fixed overnight in 4% formaldehyde in PBS (Figure 1A). The brain was cut into 3-mm blocks using an acrylic brain matrix (Braintree Scientific,

Braintree, MA, USA) and placed in 30% sucrose for at least 24 hr before sectioning.

Sectioning of brain and spinal cord

Brain and spinal cord tissues were arranged in the same block, embedded in optimal cutting temperature (OCT) media (Sakura Finetek USA, Inc., Torrance, CA,

USA), and stored at -80 °C until sectioning. The spinal cord was cut into segments and arranged for both transverse and longitudinal sectioning (Figure 1B). The tissue blocks were cut with a Microm HM550 cryostat (Thermo Scientific, Waltham, MA, USA) and the 40-μm sections were collected on Superfrost Plus microscope slides (FisherScientific,

Pittsburgh, PA, USA) for storage at -20°C.

Figure 1. Sectioning strategy for brain and spinal cord. A, The brain and spinal cord were dissected from the perfused mouse, cut into tissue blocks, and arranged for embedding in OCT media. B, The end result of sectioning was a double array of 40µm tissue sections for each slide. Both longitudinal and transverse sections were collected at varying intervals from the same spinal cord.

Immunofluorescent staining

Sections were incubated in PBS/0.2% Triton X-100 for 1 hr at room temperature to permeabilize the tissue, then blocked with 2.5% normal goat or donkey serum

(matched with the host species of the secondary antibody) for 1 hr at room temperature.

The primary antibodies were then added in blocking solution, and the sections were incubated for 3 hr at room temperature, then overnight at 4°C. The next day, the sections were rinsed 10 × 5 min at RT, the appropriate secondary antibody was added in blocking solution, and the sections were incubated for 3 hr at RT. Then, the sections were incubated in Hoechst 33342 and/or Fluoromyelin for 10 min at RT. Sections were rinsed

10 × 5 min at RT and coverslipped using tris-buffered Fluoro-Gel mounting media

(Electron Microscopy Sciences, Hatfield, PA, USA).

Fluorescence light microscope and image capture

Fluorescence images were captured with a Spot CCD camera RT slider

(Diagnostic Instruments, Sterling Heights, MI, USA) in a Zeiss upright microscope,

Axiophot, using Plan Apo objectives, 2.5×/0.075, 10×/0.30, and 20×/0.50, saved as 16-bit

TIFF files, and analyzed with MetaMorph (Molecular Devices, Downingtown, PA, USA) and Sigmaplot 11.0 (Systat Software, Inc., Chicago, IL, USA) for fluorescence intensity quantification. Exposure times were controlled so that the pixel intensities in the tissues of interest were below saturation, and the same exposure time was used within each group of an experiment. Image brightness and contrast were adjusted using Adobe

Photoshop 7.0 (Adobe Systems Incorporated, San Jose, CA, USA). All fluorescence intensity measurements were taken from the original captured images.

Areas of interest were outlined with the region measurement tool in MetaMorph by tracing the borders of the lesions as shown by Hoechst staining. The traced region was then overlaid onto the Kv channel image, and the average intensity of the region was recorded. For control images, the average intensity of the spinal cord white matter was measured. The background intensity of the slide was measured and subtracted from region measurements for each image. For measurement of cell processes, the line measurement tool was used to trace each Kv channel-stained process directly, and the average fluorescence intensity along this line was recorded. Measurements of non- lesioned white matter regions of the same image were obtained, and were subtracted from the measurements of cell processes. For motor neurons, the somatodendritic region revealed by Kv2.1 staining was outlined and the average fluorescence intensity was recorded. Gray matter background intensity was measured for each image and subtracted from the intensity of the cells. For each experimental group, the same immunostaining and imaging procedures were used so that the fluorescence intensities can be compared within the group.

Confocal microscopy

High-magnification confocal images were captured with a Leica TCS SL confocal imaging system (Leica Microsystems, Mannheim, Germany), using a 100× HCX Plan

Apo CS oil immersion objective (numerical aperture = 1.40). Multiple channels were

39 acquired simultaneously, and the signal was averaged over six scans. Channel crosstalk was largely eliminated through optimization of the laser line intensity by acousto-optical tunable excitation filters, and by spectral detectors allowing precisely-defined bandwidth adjustment. Images were saved as 8-bit TIFF files and adjusted for brightness and contrast using Adobe Photoshop 7.0.

Western blotting

Mice were killed with carbon dioxide, and brain and spinal cord tissues were quickly removed, flash-frozen in liquid nitrogen, and stored at -80°C until use. The tissue samples were then weighed and homogenized in 1:4 (w/v) Laemmli Sample Buffer (Bio-

Rad Laboratories, Hercules, CA, USA) with 5% β-mercaptoethanol (Sigma). Protein samples were resolved by SDS-PAGE and transferred to a PVDF membrane (GE

Healthcare, Piscataway, N, USA) for western blotting. The membrane was rinsed in

TBST (50 mM Tris-Cl (pH 7.5), 150 mM NaCl, and 0.1% Tween20), incubated in 5% dry milk/TBST blocking solution for 1 hr at RT, and then incubated with the primary antibody in blocking solution overnight at 4°C. Membranes were then rinsed in TBST 4 ×

10 min, incubated with a horseradish peroxidase-conjugated secondary antibody in blocking solution for 1 hr, and rinsed in TBST 5 × 10 min. Membranes were then incubated in ECL Plus (GE Healthcare) chemiluminescent solution for 5 min, wrapped in plastic, and exposed to X-ray film (Kodak, Rochester, NY, USA). Developed film was digitized with an Epson 3590 scanner (Epson America, Inc., Long Beach, CA, USA). The total intensities of protein bands were measured and quantified with NIH ImageJ.

Background was subtracted for each band. The β-tubulin loading ratios were obtained by normalizing against the control band. The Kv channel alterations were obtained by first normalizing against their controls and further normalizing with the β-tubulin loading ratio. All Western blotting experiments were performed three times.

Hippocampal neuron culture and transfection

The hippocampal neuron culture was prepared as previously described from rat hippocampi at the embryonic day 18 (E18) (Xu et al., 2007b, Barry et al., 2010, Xu et al.,

2010). The neurons were transfected with cDNA constructs at 5 DIV (day in vitro), fixed at 7 DIV, and stained with a rabbit polyclonal anti- Kv 1.4 antibody. YFP- Kv1.2 was previously described (Gu and Gu, 2011, Gu et al., 2012). YFP- Kv1.4 was made by inserting the coding sequence of yellow fluorescence protein into the N-terminal region of rat Kv1.4 between G58 and G59 with an engineered NheI site. The construct was confirmed with sequencing.

Statistical analysis

For both immunofluorescence imaging and Western blotting data, we performed

One-Way ANOVA when comparing 2 or 3 experimental groups to one control group, and unpaired t-test when comparing two groups, using Sigmaplot 11.0. Results are provided as mean ± SEM.

RESULTS

Demyelinated lesions in chEAE and rrEAE

To determine potential alterations of Kv1 channels in the pathogenic process of demyelinating diseases, chEAE was induced by immunizing C57BL/6 mice with

MOG35-55, which exhibited a rapid onset and sustained clinical EAE (Figure 2A). Since acute EAE peaked around day 20, this time point (17-20 DPI) was used to determine the initial effects of inflammation and demyelination on Kv channel expression (Figure 2A).

In addition, day 34 was selected to determine the distribution of Kv channels in the chronic phase. SJL/J mice, immunized with PLP139-151, were used as a model of remitting-relapsing MS, the most common MS subtype. Analysis was performed on day

14 (clinical peak), 34 (remission) and 66 (relapse) to determine the changes associated with clinical signs (Figure 3A). For immunohistochemistry studies, brains and spinal cords of EAE and control mice were dissected as shown in Figure 1. Coronal sections were cut from brain, and both longitudinal and cross sections were cut from spinal cord

(Figure 1). In longitudinal sections of spinal cord, Fluoromyelin staining readily detected the areas of demyelination in the dorsal (left) and ventral (right) white matter (Figure

2A). Enhanced nuclear dye staining illustrated the inflammatory cells present in the regions of myelin loss. In the remitting phase of rrEAE (25-34 DPI; showing a reduction in clinical scores by at least 2 points), both the number and the size of lesions decreased, correlating well with the EAE signs (Figure 3B), indicating that remyelination may occur.

In contrast, at the late stage of chEAE, the lesions were persistent, even somewhat

42 increased (data not shown). Taken together, these models provide a foundation for analyzing Kv channels in the CNS at different clinical stages observed in MS patients.

We asked the following questions. Do the expression and localization of Kv1 channels change during the progression of demyelinating disease, and if so, how? Do changes of channel proteins correlate with demyelination? Does Kv1 channel expression in distinct brain regions change differently during disease progression? Is there any difference in Kv channel expression and localization between normal mice and those in the remitting phase of rrEAE? Are dendritic as well as axonal Kv channels altered during EAE?

Figure 2. Characterization of the Chronic EAE model. (A) Mean clinical scoring in female C57BL/6J mice from immunization (day 0) with MOG peptide 35–55 until 41 days postimmunization (DPI) showed a typical disease course for chEAE. (B) Inflammatory, demyelinating lesions of chEAE spinal cord (SC) white matter (WM) were demonstrated by Fluoromyelin green (FMG) and Hoechst nuclear staining in longitudinal SC sections. Left, FMG labeling (green in the merged image on the right) revealed dorsal and ventral WM flanking gray matter (GM). Middle, cell nuclei were labeled with the DNA-binding dye Hoechst (blue in merged image). Multiple lesioned areas (red arrows) were observed in chEAE SCWM (bottom) at peak disease severity [clinical score (CS) = 4], characterized by dense clusters of cell nuclei (indicating inflammatory cells) colocalized with areas where FMG was absent (indicating demyelination). 44

Figure 3. Characterization of the Relapsing-Remitting EAE model. (A) Mean clinical scores (CS) in female SJL/J mice from immunization (day 0) with PLP peptide 139–151 until 65 DPI revealed a typical rrEAE disease course featuring an early peak, a gradual return to near normal, and a relapse phase. (B) Lesions in rrEAE SCWM showed myelin loss and clustered cell nuclei similar to chEAE (Figure 2) at peak disease severity. At the remitting stage, cell nuclei clusters were less dense and myelin loss was less severe. Control (top), peak stage (middle), and remitting stage (bottom; note that this mouse previously scored CS = 4 at the peak stage) longitudinal SC sections are shown. Left, FMG staining. Right, merged FMG (green) and Hoechst (blue) staining. Red arrows, lesions with reduced FMG and clustered nuclear staining. Scale bars, 1500 μm. 45

Disrupted Kv1.2/Kvβ2 JXP clustering in demyelinated lesion areas in EAE

We recently determined that myelin, via a cell adhesion molecule, plays a critical role in regulating the biophysical properties as well as the JXP clustering of axonal Kv1.2 channels (Gu and Gu, 2011). Thus, demyelination observed in MS may affect axonal conduction through altering both activity and localization of these channels. We examined Kv1.2 expression and distribution in EAE. In control mice, the highly clustered pattern of Kv1.2 staining was observed in the white matter of spinal cord (Figure 4), consistent with previous reports that Kv1.2 is clustered in the JXP regions of myelinated axons (Wang et al., 1993, Rhodes et al., 1995, Vabnick et al., 1999, Rasband and

Shrager, 2000). The clusters decreased or disappeared in the lesion sites for both chEAE

(Figure 4) and rrEAE (Figure 5). In some lesion sites, Kv1.2 staining appeared slightly increased, but not in clusters (Figure 4 bottom and Figure 5 middle). The smooth staining signals were likely contributed by Kv 1.2 expression in proliferating immune cells and demyelinated axons. In some lesion sites in the rrEAE remitting phase, Kv1.2 clusters were still present, most likely resulting from partial demyelination and/or remyelination, as indicated by the reduction but not complete absence of the Fluoromyelin staining

(Figure 5 bottom). Overall, disruption of Kv1.2 clustering was highly restricted to the lesion site. In non-lesion areas, Kv1.2 clustering remained normal (Figure 4,Figure 5).

Figure 4. Altered Kv1.2 JXP targeting in lesioned areas in chEAE spinal cord. (A) JXP clustering of Kv1.2 in myelinated axons of SCWM was altered in chEAE lesions. Left, Kv1.2 staining with signals inverted. Right, merged images with Kv1.2 staining shown in red, FMG in green, and nuclear staining in blue. Control (top), peak stage (middle), and late stage (bottom) SC longitudinal sections are shown. DPI and CS of each mouse are indicated. Scale bars, 200 μm.

Similar results were obtained in both chEAE and rrEAE for Kvβ2, an auxiliary subunit of Kv1 channels (Figure 6). The average fluorescence intensity of Kv1.2 remained unchanged in lesion versus non-lesion areas at different stages of both chEAE and rrEAE

(Figure 7), indicating that there were changes in the distribution but not total amount of

Kv1.2 channels. Therefore, demyelination-induced disruption of Kv1/ Kvβ2 clustering is a local effect.

Figure 5. Kv1.2 JXP clustering was altered in rrEAE lesions. Shown are staining images from spinal cord sections of the mice of control (top), at the peak stage (middle), and at the remitting stage (bottom).

Figure 6. Kvβ2 JXP clustering was disrupted in lesions of chEAE and rrEAE spinal cord. Top, chEAE; bottom, rrEAE. Lesion areas are enclosed with dashed green lines and indicated with white arrows in merged images.

Figure 7. Expression of Kv1.2 in the areas with or without a lesion in spinal cord white matter at different stages of chEAE (left) and rrEAE (right). In chEAE, the average Kv1.2 staining intensities of spinal cord white matter are 667 ± 94 (AU, n = 18) in control areas, 703 ± 57 (AU, n = 13) in lesion areas from peak chEAE, and 571 ± 76 (AU, n = 5) in lesion areas from late chEAE. In rrEAE, the average Kv1.2 staining intensities are 653 ± 67 (AU, n = 16) in control areas, 571 ± 33 (AU, n = 15) in lesion areas from the peak stage, and 596 ± 43 (AU, n = 20) in lesion areas from the remitting stage. No significant difference was detected among the groups (One-Way ANOVA followed by Dunn's test, p > 0.05). Average fluorescence intensities of different areas of interest were measured, quantified, and shown as mean ± SEM. Background fluorescence intensity was subtracted.

To determine whether Kv1.2 channel expression and distribution are altered in myelinated and unmyelinated axons, we examined other brain regions during EAE progression. In both chEAE and rrEAE, the clustering of Kv1.2 channels in unmyelinated cerebellar basket cell terminals remained unchanged (Figure 8A), which is consistent with previous studies showing that different molecular mechanisms determine Kv1.2 clustering in different CNS regions (Ogawa et al., 2010). However, lesions in the cerebellar white matter also disrupted Kv1.2 clustering similar to those in spinal cord white matter (data not shown). In the time frame studied, we did not observe any lesions in the corpus callosum, or any change of Kv1.2 clustering there (Figure 8B). The overall expression levels of Kv1.2 and Kvβ2 in cerebral cortex and spinal cord remained constant at different stages of chEAE and rrEAE revealed by Western blotting, indicating there is no global alteration of these channel proteins during the progression of EAE (Figure

9A,B), consistent with the immunostaining results (Figure 7).

Figure 8. Kv1.2 expression patterns in brain regions during EAE progression. (A) Kv1.2 clustering in cerebellar basket cell terminals was not affected during chEAE progression. Coronal sections of cerebellum from control (top) and chEAE (bottom) mice are shown. DPI and CS are provided on the left. Kv1.2 staining is in red, FMG in green, and Hoechst staining in blue. Images with 10-fold higher magnification showing Kv1.2 staining (red) in basket cell terminals are on the far right. (B) No lesion was observed in corpus callosum and no change in Kv1.2 JXP targeting. The image was from corpus callosum linking two hemispheres in a coronal section of mouse brain. Images with 5- fold higher magnification showing Kv1.2 staining (red) in JXP clusters of myelinated axons in corpus callosum are on the far right. Scale bars, 200 µm.

Figure 9. The global expression of Kv1.2/ Kvβ2 in cerebral cortex and spinal cord at different stages of chEAE (A) and rrEAE (B) revealed by Western blotting. Molecular weights were labeled on the left in kDa. Each set of experiment contains four different time points, control, and at early peak, peak, and late stages. Equal amounts of tissue samples were loaded on the gel within each set. Mouse anti- Kv1.2 and rabbit anti- Kvβ2 antibodies were used in the Western blotting. Mouse monoclonal anti-β-tubulin was used as a control for protein loading. The intensities of protein bands were measured, normalized with the controls, further normalized with the ratio of β-tubulin staining, and provided as mean ± SEM (bottom). Background intensity was subtracted for each measurement.

Altered patterns of Kv1.2 targeting along axons in various lesion areas

To determine whether the reduction of Kv1.2 JXP clustering is due to the disruption of the integrity of myelinated axons in lesions, we induced chEAE in Thy1-

YFP transgenic mice, which contain a subset of neurons that express YFP. Using confocal microscopy, we visualized the YFP-positive axons and Kv1.2 localization pattern. In control, Kv1.2 staining was clustered in the JXP regions of both YFP-positive and YFP-negative axons (Figure 10C). Kv1.2 highly colocalized with a JXP marker,

Caspr2 (Figure 10C bottom). Within EAE lesions, Kv1.2 clusters markedly decreased and

Kv1.2 staining was no longer restricted to the JXP regions (Figure 10A, D top). In the borderline region of a lesion, varicosities along axons were often observed, which may represent retraction bulbs of broken axons or abnormal axonal enlargements, where Kv1.2 channels were diffusely distributed (Figure 10 B, D bottom). As shown by a recent study, some varicosities in the axons within lesions can recover (Nikic et al., 2011). Some axons remained myelinated in borderline regions, reflected by the presence of Kv1.2 JXP clusters (Figure 10 B, D bottom).

Figure 10. Alterations of Kv1.2 targeting patterns along axons within EAE lesions. (A, B) Lesions in a Thy1-YFP transgenic mouse at peak disease severity. YFP+ axons (green) traverse the lesion area revealed by Hoechst nuclear staining (blue). (C) Kv1.2 JXP clustering along myelinated axons in spinal cord white matter of a CFA-injected control Thy1-YFP transgenic mouse. In upper panels, YFP-positive axons are shown in green and anti- Kv1.2 staining in red. The lower panels show Kv1.2 (red in merged) co- clustered with Caspr2 (blue in merged), a JXP marker, along a YFP-expressing axon. White arrowheads, putative JXP regions flanking nodes of Ranvier. (D) In upper panels, Kv1.2 JXP clusters within a chEAE lesion (see low-mag view in A) were eliminated, while several YFP-positive axons were still present. In lower panels, this area was from the borderline of a chEAE lesion (see low-mag view in B). A potential retraction bulb was present, whereas some Kv1.2 clusters were present in surrounding axons. DPI and CS are provided on the left. Scale bars, 20 µm.

Figure 10. Alterations of Kv1.2 targeting patterns along axons within EAE lesions.

In various lesion areas, we observed three different types of Kv1.2 alterations in the JXP clustering patterns at EAE lesions of spinal cord white matter (Figure 11A).

First, the Kv1.2 JXP cluster was elongated only on one side of the node of Ranvier, which may represent the partial or asymmetric demyelination of the axon (Figure 11A,

Asymmetric). Second, the distribution of Kv1.2 became uniform over a segment of completely-demyelinated axon (Figure 11A, Uniform). Third, Kv1.2 staining became uniform but more concentrated in an axonal varicosity (Figure 11A, Varicosity). In addition to altered distribution patterns, significant reduction of Kv1.2 staining intensity along axonal segments around nodes of Ranvier was also observed during the peak stage of both chEAE (Control: 1033 ± 38 AU, n = 83; Peak: 857 ± 21 AU, n = 174; One Way

ANOVA followed by Dunn's test, p < 0.01) and rrEAE (Control: 949 ± 30 AU, n = 167;

Peak: 727 ± 20 AU, n = 149; One Way ANOVA followed by Dunn's test, p < 0.01)

(Figure 11B). Further reduction was observed for the chEAE late phase (Late: 603 ± 30

AU, n = 55; One Way ANOVA followed by Dunn's test, p < 0.01) (Figure 11B). In sharp contrast, the staining intensity significantly recovered in the remitting phase of rrEAE

(Late: 877 ± 23 AU, n = 264; One Way ANOVA followed by Dunn's test, p > 0.05)

(Figure 11B).

Figure 11. (A) Different patterns of altered Kv1.2 JXP targeting in demyelinated lesions. Kv1.2 staining in red and FMG staining in green. Control, normal Kv1.2 JXP targeting in control mouse; Asymmetric, only one of the Kv 1.2 JXP clusters elongated; Uniform, Kv1.2 distributed uniformly along demyelinated axons; Varicosity, Kv1.2 distribution pattern in enlarged axonal segments or retraction bulbs. Scale bars, 10 μm. (B) Average Kv1.2 staining intensities along axonal segments near nodes of Ranvier of axons within and outside the lesions. One-Way ANOVA followed by Dunn's test, ** p < 0.01.

Different types of lesions were observed in this study. Based on the staining of

Fluoromyelin and nuclear dye, we divided them into three major groups. The first group, type α (approximately 28% in peak chEAE; 28% in late chEAE; 17% in peak rrEAE;

10% in remitting rrEAE)(Figure 12B), includes a cyst of cells pushing axons aside without breaking them (Figure 12A). The second group, type β (approximately 62% in peak chEAE; 59% in late chEAE; 57% in peak rrEAE; 31% in remitting rrEAE)(Figure

12B), includes a dense core of cells with many broken axons failing to cross the lesion

(Figure 12A). This is a major portion of lesions observed at the peak stage of both chEAE and rrEAE, which likely represent local permanent lesions in spinal cord white matter.

The third group, type γ (approximately 10% in peak chEAE; 14% in late chEAE; 26% in peak rrEAE; 59% in remitting rrEAE)(Figure 12B), includes lesions with less dense cell bodies and partially demyelinated areas (Figure 12A), which were more frequent in the remitting phase of rrEAE and may represent the recovering and remyelinating phase of lesions.

Figure 12. Different types of lesions in spinal cord white matter of chEAE and rrEAE. (A) Illustration of three major types (α, β and γ) of lesions observed in spinal cord white matter of chEAE and rrEAE mice. Black lines, axons revealed by Kv1.2/ Kvβ2 staining and YFP fluorescence of the Thy1-YFP transgenic mice; Blue circles, proliferating cells revealed by nuclear dye (Hoechst) staining (potentially the proliferated inflammatory cells); Green areas, myelinated axons labeled with FMG; White areas, demyelination lesions with no FMG staining; Light green areas, partially demyelinated areas. (B) Percentage of different lesion types at the peak and late stages of chEAE (Peak: n = 61; Late: n = 58) and rrEAE (Peak: n = 70; Late (remitting): n = 70).

Somatodendritic Kv2.1 in spinal cord motor neurons during EAE progression

Alterations of axonal Kv1.2/Kvβ2 in EAE progression are largely consistent with the notion that their JXP clustering is dictated by myelin. Since both motor and sensory systems are affected in MS, we examined two other Kv channels in EAE progression,

Kv2.1 expressed in the alpha motor neurons and Kv1.4 expressed in sensory neurons. In contrast to the axonal Kv1 channels, Kv2.1 channels are mainly localized in clusters in soma and proximal dendrites of spinal cord motor neurons (Lim et al., 2000, Muennich and Fyffe, 2004), and were initially used as a negative control for Kv channels in spinal cord gray matter. Surprisingly, we found a marked reduction of Kv2.1 channels in motor neurons in EAE especially in the lower spinal cord (Figure 13A).

Figure 13. Alterations of somatodendritic Kv2.1 channels in spinal cord motor neurons in chEAE. (A) Kv2.1 levels in motor neurons in lower spinal cord of chEAE were significantly reduced. Signals are inverted in single channel images. In merged images, Kv2.1 staining is shown in red and nuclear dye (Hoechst) staining in blue. DPI and CS are provided on the left. Dashed green lines indicate the border between gray matter (GM) and white matter (WM). (B) Reduction of somatodendritic Kv2.1 channels in motor neurons in EAE of Thy1-YFP transgenic mice. In merged images, YFP is shown in green and Kv2.1 staining in red.

To determine whether the reduction of Kv2.1 levels is due to neuronal cell death or channel reduction within neurons, we induced chEAE in Thy1-YFP transgenic mice.

Although some reduction of YFP-positive neurons in spinal cord gray matter was observed, consistent with previous reports (Aharoni et al., 2011, Vogt et al., 2009), Kv2.1 channel levels were clearly reduced in many YFP-positive motor neurons in the lower spinal cord in chEAE (Figure 13B). Next, we examined whether Kv2.1 levels in these motor neurons were reduced in rrEAE. Clear reduction of Kv2.1 was observed at the peak stage, similar to chEAE; yet, Kv2.1 channel levels became largely normal during the remitting phase (Figure 14).

Figure 14. Kv2.1 levels in motor neurons of lower spinal cord in rrEAE correlated with clinical signs. The rrEAE mice of control (top), at the peak phase (middle), and at the remitting phase (bottom), are shown.

In upper spinal cord, the level of Kv2.1 remained unchanged in chEAE (Figure 15, top). In lower spinal cord, the level of Kv2.1 was markedly reduced in both the early and late stages of chEAE (Control: 154 ± 11.4 (AU), n = 50; Peak: 111 ± 8.6 (AU), n = 56;

Late: 87 ± 4.0 (AU), n = 54) (Figure 15, bottom). In contrast, Kv2.1 levels were significantly reduced at the peak stage of rrEAE, but increased to normal levels in the remitting phase (Control: 138 ± 5.2 (AU), n = 67; Peak: 94 ± 4.7 (AU), n = 56; Late

(remitting): 131 ± 7.0 (AU), n = 64) (Figure 15), correlating with clinical signs. The

Kv2.1 reduction was in gray matter, and did not correlate with nearby lesion sites in white matter. Therefore, the alteration of Kv2.1 channel levels is not a consequence of demyelination.

Figure 15. Quantification of fluorescence intensity of Kv2.1 in spinal cord alpha motor neurons. (top) Average Kv2.1 levels in motor neurons of upper spinal cord in chEAE [(Left) Control: 98 ± 6.4 (AU), n = 51; Peak: 97 ± 6.8 (AU), n = 55; Late: 96 ± 5.6 (AU), n = 54)] and rrEAE [(Right) Control: 113 ± 6.9 (AU), n = 61; Peak: 81 ± 3.3 (AU), n = 63; Late (remitting): 122 ± 5.9 (AU), n = 63)]. (bottom) Average Kv2.1 levels in the motor neurons of lower spinal cord in chEAE and rrEAE (see text for statistics). One-Way ANOVA followed by Dunn's test. ** p < 0.01. Scale bars, 150 μm in (A) and (C); 50 μm in (B).

Upregulation of Kv1.4 channels around spinal cord lesions in EAE progression

Kv1.1, Kv1.2 and Kv1.4 are the three most abundant Kv1 subunits expressed in mammalian brain. They often form heteromeric complexes predominantly in axons of some neurons, but their composition varies dramatically. The Kv1.1 subunit appears to be segregated into two major subpopulations: one associated with Kv1.2 and one associated with Kv1.4 (Vacher et al., 2008). In particular, Kv1.4 colocalizes with Kv1.1 in some unmyelinated axons of dorsal root ganglion neurons (Rasband et al., 2001). Kv1.4 expression in spinal cord white matter in adult mice is close to background (Figure 16 A,

B upper panels). Surprisingly, Kv1.4 was markedly up-regulated around lesion sites in both chEAE and rrEAE at the peak stage (Figure 16 A,B middle panels). Its level declined at the late stage of chEAE, but continued to increase at the remitting stage of rrEAE (Figure 16 A,B lower panels). Interestingly, Kv1.4 was present in both round- and fiber-like structures, most likely corresponding to the cell bodies and processes of one or multiple types of cells (Figure 16 A,B).

Figure 16. Upregulation of Kv1.4 channels in GFAP-positive cells around EAE lesions. (A) Kv1.4 channel expression markedly increased in spinal cord WM around lesions in chEAE. In SCWM, Kv1.4 staining is shown (signals inverted on the left and in red in the merged images on the right) in a control mouse (top), and at the peak (middle) and late (bottom) stages of chEAE. GFAP staining is shown in green and Hoechst staining in blue. DPI and CS are provided on the left. Lesion areas are enclosed with green dashed lines and indicated with white arrows in merged images. (B) Kv1.4 channels markedly increased in SCWM around lesions in rrEAE. The late stage in rrEAE (bottom) is the remitting stage and this mouse had a CS at 4 at the peak stage. (C) Average fluorescence intensities of Kv1.4 in normal and lesion areas. In chEAE: Control, 615 ± 44 (AU), n = 27; Peak, 564 ± 28 (AU), n = 41; Late, 543 ± 58 (AU), n = 18. In rrEAE: Control, 469 ± 34 (AU), n = 28; Peak 683 ± 28 (AU), n = 47; Late (remitting), 752 ± 52 (AU), n = 28.

Figure 16. Upregulation of Kv1.4 channels in GFAP-positive cells around EAE lesions

By performing co-staining with the astrocyte marker, GFAP, we found that the processes containing up-regulated Kv1.4 were exclusively GFAP positive (Figure 16A,B;

Figure 18A,B), indicating that Kv1.4 expression was specifically enhanced in astrocytes around lesions. The specificity of the rabbit polyclonal anti- Kv1.4 antibody was confirmed by robust and specific staining of cultured cells transfected with YFP-tagged

Kv1.4, but not those transfected with YFP-tagged Kv1.2 (Figure 17).

Figure 17. Specificity of the rabbit polyclonal anti- Kv1.4 antibody. Cultured hippocampal neurons were transfected with either YFP- Kv1.4 (A) or YFP-Kv1.2 (B) at 5 DIV, fixed at 7 DIV, and stained with a rabbit polyclonal anti-Kv1.4 antibody (red). YFP is in green. Scale bars, 100 m.

GFAP staining was clear in spinal cord white matter of control mice and increased around the lesion sites in EAE. The average fluorescence intensity of Kv1.4 in lesion and non-lesion areas in spinal cord white matter was rather consistent during the three time points in chEAE, but significantly increased in the peak and remitting phases of rrEAE. (Figure 16C). In contrast, the Kv1.4 staining intensity along each cellular process significantly increased in both chEAE (Control: 150 ± 13 (AU), n = 162; Peak:

242 ± 10 (AU), n = 394; Late: 192 ± 8 (AU), n = 201) and rrEAE (Control: 68 ± 12 (AU), n = 45; Peak: 229 ± 12 (AU), n = 235; Late: 292 ± 11 (AU), n = 336) (Figure 18C). This result also indicates the expression level of Kv1.4 in normal white matter is low, but not absolute zero. This is why the average level of Kv1.4 remained constant in chEAE

(Figure 16C left), even when its level clearly increased in astrocytic processes (Figure

18C left). In contrast, the increase of Kv1.4 levels in rrEAE appeared to be more global

(Figure 16C right). Nonetheless, there was a clear increase of Kv1.4 levels in GFAP- positive astrocytic processes around lesion sites in both chEAE and rrEAE. Actually the number of Kv1.4-positive processes decreased in late chEAE but increased in remitting rrEAE. Importantly, we only observed Kv1.4 upregulation in fiber-like GFAP-positive astrocytes in spinal cord white matter, but never in GFAP-positive astrocytes in spinal cord gray matter (Figure 18D,E).

Figure 18. High-magnification confocal imaging of Kv1.4 upregulation in chEAE and rrEAE. (A) Confocal images of Kv1.4 colocalizing with GFAP at the peak (top) and late (bottom) phases of chEAE. (B) Confocal images of Kv1.4 colocalizing with GFAP at the peak (top) and remitting (bottom) phases of rrEAE. (C) Average fluorescence intensities of Kv1.4 along individual processes around the lesions in chEAE and rrEAE mice. Background fluorescence intensity was subtracted for each measurement. One-Way ANOVA followed by Dunn's test. ** p < 0.01, * p < 0.05. (D) High-mag. image of a GFAP-positive (green) astrocyte expressing Kv1.4 (red) in spinal cord white matter during chEAE. (E) High-mag. image of a GFAP-positive (green) astrocyte in spinal cord gray matter during chEAE. No Kv1.4 (red) was expressed. Scale bars, 25 μm in (A) and (B); 50 μm in (D) and (E).

Up-regulated Kv1.4 channels only partially colocalized with NG2, a marker for oligodendrocyte precursor cells, in both chEAE and rrEAE (Figure 19). NG2 staining was up-regulated around the lesion sites (Figure 19). The Kv1.4-positive processes were not NG2 positive (Figure 19). These processes were not likely axons either, as revealed in

Thy1-YFP transgenic mice (Figure 19). No YFP-positive axon expressing Kv1.4 was ever observed (Figure 19C). To determine the global changes of Kv1.4 in brain and spinal cord during EAE progression, we performed the Western blotting. In chEAE, there was no significant change detected at the total protein level of Kv1.4 (Figure 20), consistent with the immunostaining result (Figure 16C left). In rrEAE, a significant increase of Kv1.4 protein level was observed in spinal cord, but not in cortex (Figure 20), consistent with the immunostaining result in spinal cord (Figure 16C). Therefore, the expression of Kv1.4 indeed appeared to increase in the spinal cord of rrEAE mice in the peak and remitting phases.

Figure 19. Up-regulated Kv1.4 channels partially colocalized with NG2 but not with YFP-positive axons in Thy1-YFP transgenic mice. (A) NG2 staining increased around the lesions in spinal cord white matter of chEAE. NG2 staining signals are inverted on the left image and shown in green in merged images on the right. Kv1.4 staining is shown in red and nuclear dye in blue. Lesion areas are enclosed with green dashed lines and indicated with white arrows in merged images. (B) NG2 staining increased around the lesions in spinal cord white matter of rrEAE. (C) Up-regulated Kv1.4 around the lesions did not colocalize with YFP-positive axons around chEAE lesions. Kv1.4 staining is in red and YFP in green in the merged image (right). Scale bars, 200 μm.

Figure 20. Global expression of Kv1.4 channels revealed by Western blotting of cerebral cortex and spinal cord at different stages of chEAE (D) and rrEAE (E). Molecular weight was labeled on the left in kDa. Equal amounts of tissue samples were loaded on the gel. Normalized intensity of protein bands in the Western blotting is shown at the bottom.

Since changes in Kv1.2, Kv2.1 and Kv1.4 correlated with severity of EAE signs in the late phase of chEAE and remitting phase of rrEAE, we wondered how these channels changed in the late relapse phase of rrEAE (Figure 21A). At day 66, lesions were bigger and more abundant in the white matter of spinal cord and Kv1.2 JXP clustering was affected more extensively (Figure 21B). Kv2.1 further reduced (data not shown).

Importantly, the Kv1.4 levels were significantly reduced compared to the remitting phase

(Kv1.4 average intensity: remitting, 292 ± 11 (AU), n = 336; relapsing, 130 ± 5 (AU), n =

279; p < 0.01) (Figure 21C-E), supporting the observation that Kv channel expression is related to lesion formation and clinical signs of rrEAE. Thus, Kv channels play critical, yet diverse, roles in the functional consequences of immune-mediated demyelination.

Figure 21. Kv1.4 levels decreased in lesions of spinal cord white matter during the late relapse phase of rrEAE. Kv1.4 levels decreased in lesions of spinal cord white matter during the late relapse phase of rrEAE. (A) Clinical scores of two rrEAE mice. Both were sacrificed at 66 DPI during the late relapse stage. (B) Kv1.2 staining in spinal cord white matter of rrEAE at the relapse stage. Lesion areas are enclosed with green dashed lines and indicated with white arrows. Kv1.2 staining is inverted on the top and in red in the merged image at the bottom. FMG staining is in green and nuclear dye staining in blue. (C) Kv1.4 staining decreased in the GFAP positive processes around the lesions. Kv1.4 staining is inverted in the left and in red in the merged image on the right. GFAP staining is inverted in the middle and in green in the merged image on the right. The nuclear dye staining is in blue in the merged image. (D) Decreased Kv1.4 staining only partially colocalized with NG2 staining. NG2 staining is inverted in the middle and in green in the merged image on the right. (E) Statistical results of Kv1.4 staining intensity in the processes around the lesions. Unpaired t-test; **, p < 0.01. Scale bars, 250 μm.

Figure 21. Kv1.4 levels decreased in lesions of spinal cord white matter during the late relapse phase of rrEAE.

DISCUSSION

This is the first report of a systematical study of the expression and distribution patterns of three key Kv channels in the CNS during the progression of two EAE models.

This study contributes to better understanding of the relationship of Kv channels to MS pathogenesis and of the therapeutic mechanism of Kv channel blockers such as 4-AP.

Localized regulation of JXP clusters of Kv 1.2/ Kv β2 channels

Demyelinated lesions were present in spinal cord white matter in both chEAE and rrEAE (Figures 2, 3). The JXP clustering of Kv1.2/ Kvβ2 was disrupted apparently only in the demyelinated lesion site (Figures 4, 5, 6). In surrounding areas, they were still correctly clustered to the JXP regions. In some lesions, especially those from the remitting phase of rrEAE, Kv1.2 clusters were also present, which might be due to early or partial demyelination (Figure 5). However, the overall effect of demyelination on

Kv1.2/ Kvβ2 clustering is clear. This is consistent with the notion that myelin clusters Kv1 channels into the JXP regions. Blocking the activity of these exposed Kv channels along demyelinated axons should enhance axonal conduction and hence potentially lead to improvement of disease symptoms of MS, which may underlie the beneficial effects of the 4-AP treatment.

In this study, we observed variations in the patterns of lesion and alteration of Kv

1.2 clusters (Figures 4, 7). Lesions differed at different stages and in different EAE models. We divide them into three groups, type α, β, and γ (Figure 12). Although their

81 cellular and molecular identities are not the focus of the present study, they may be of interest for future investigation. Especially, in the rrEAE, how type β lesions apparently containing many broken axons in the initial exacerbation may be repaired in the remitting phase is of high interest for future studies. A recent study surprisingly shows that the varicosities forming along axons are reversible (Nikic et al., 2011), which appears to support what we have observed in this study. Moreover, completely or partially disrupted clustering patterns of Kv1.2/ Kvβ2 were observed, correlating with the degree of demyelination and axonal damage (Figure 5C,D). In the remitting phase of rrEAE, the number and size of the lesions decreased and Kv1.2 clustering recovered close to levels observed in control animals, suggesting remyelination.

We also observed weak signals from Kv1.2 channel staining in round cells within some lesions, which suggests the expression in other cell types, immune cells or glial cells, at lower levels. Kv1 channels are indeed expressed in immune/inflammatory cells

(Kettenmann et al., 1993, Wulff et al., 2003a, Wulff et al., 2003b, Chandy et al., 2004,

Beeton and Chandy, 2005, Farber and Kettenmann, 2005, Fordyce et al., 2005, Pannasch et al., 2006, Nutile-McMenemy et al., 2007, Wu et al., 2009). Moreover, Kv1 channel activity is required for the proliferation of oligodendrocyte precursor cells (Attali et al.,

1997, Chittajallu et al., 2002), while oligodendrocyte proliferation and subsequent remyelination may be important in the remitting phase of MS. Nonetheless, neuronal dysfunction is the direct manifestation of MS.

Downregulation of Kv2.1 and upregulation of Kv1.4 during EAE progression

In this study, we surprisingly observed that Kv2.1 expression in motor neurons of lower spinal cord was markedly down-regulated, correlating with EAE severity (Figures

13, 15). At the late stage of chEAE induced in Thy1-YFP transgenic mice, YFP-positive motor neurons in spinal cord decreased, suggesting neuronal death, consistent with previous reports in MOG-induced EAE (Aharoni et al., 2011, Vogt et al., 2009). This process is most likely irreversible, consistent with sustained clinical signs in the late chEAE. In the rrEAE remitting phase, Kv2.1 staining was restored back to the normal level, suggesting that motor neuron death may not be very abundant in rrEAE, otherwise it would be difficult to understand how these motor neurons were regenerated in less than a month. Previous reports differ regarding motor neuron loss in PLP-induced EAE, showing the specific genetic background and EAE induction conditions may be important. Significant motor neuron loss was noted in adoptive transfer EAE in SJL/J mice (Vogt et al., 2009), while no loss was observed in active EAE in (SJL×Balb/c)F1 mice (Aharoni et al., 2011). Interestingly, another study shows reversible effects of EAE induced in rats on motor neuron dendritic arbors (Zhu et al., 2003), consistent with our results in rrEAE. Nonetheless, we observed many surviving motor neurons in Thy1-YFP mice. In these neurons, Kv2.1 levels were markedly reduced and its clusters were eliminated (Figure 13B).

Kv2.1 channels form large clusters in somatodendritic regions of cortical/hippocampal pyramidal neurons and spinal cord motor neurons. Regulation of

Kv2.1 clustering, extensively studied, may involve neuronal activity, protein

83 phosphorylation, and actin cytoskeleton (O'Connell et al., 2010, Misonou et al., 2006,

Misonou et al., 2004). However, the functional consequence of Kv2.1 clustering currently remains controversy (O'Connell et al., 2010, Misonou et al., 2004). In EAE, it remains unclear whether the downregulation of Kv2.1 levels and clusters is part of the pathogenic process resulting in enhanced neuronal excitability or a compensatory mechanism to increase suppressed excitability. It will be interesting to determine whether Kv2.1 downregulation occurs in MS patients in future investigation.

In sharp contrast to the restricted alteration of Kv1.2/ Kvβ2 JXP clustering

(Figures 4, 6), the Kv2.1 downregulation did not correlate with a nearby lesion (Figures

13, 14). The Kv1.4 upregulation occurred not only within lesion areas, but also in the surrounding areas (Figure 16). Therefore, diffusible factor(s) may be involved in regulating Kv2.1 and Kv1.4.

Kv1.4- and GFAP-positive astrocytes around EAE lesions and their potential function

The Kv1.4-positive processes that were markedly increased surrounding EAE lesions were exclusively colocalized with GFAP (Figures 16, 18), a marker usually used to identify differentiated astrocytes in vitro and in vivo. This result suggests Kv1.4 is upregulated in activated astrocytes in spinal cord white matter. Neural stem cells giving rise to neurons and oligodendrocytes can also transiently express GFAP (Doetsch et al., 1999,

Seri et al., 2001, Garcia et al., 2004). A recent study shows that reactive astrocytes in

EAE were derived by proliferation and phenotypic transformation of fibrous astroglia in spinal cord white matter (Guo et al., 2012). Therefore, it is most likely that Kv1.4 is up-

84 regulated in reactive astrocytes around EAE lesions. NG2, a marker for oligodendrocyte precursor cells, was only partially colocalized with Kv1.4-positive cells (Figure 19A,B).

Our results are consistent with the upregulation of Kv1.4 channels observed in damaged white matter in the rat model of spinal cord injury (Edwards et al., 2002) and in demyelinated lesions of mouse spinal cord of MOG-induced EAE (Herrero-Herranz et al., 2007). However, these studies show that the up-regulated Kv1.4 mainly resides in

OPCs and oligodendrocytes, different from our results. Importantly, in our immunofluorescence staining of EAE, mouse antibody was not used, eliminating nonspecific signals from endogenous antibodies produced by mice during EAE. In the present study, the Kv1.4-positive processes were not colocalized with myelin marker

FMG, nor axonal marker YFP in Thy1-YFP transgenic mice (Figure 19C).

Kv1.4 is expressed by a subset of astrocytes surrounding EAE lesions. GFAP- positive astrocytes are abundant in spinal cord white matter of control mice, when Kv1.4 was at a level close to the background (Figure 16A,B). Given that Kv1.4 is up-regulated continuously in the remitting phase of rrEAE, but down-regulated in the late phase of chEAE and the relapse phase of rrEAE (Figures 16, 18, 19), these Kv1.4-positive cells may play a neuroprotective role. For example, Nav1.5 channels are present in astrocytes and important in astrocyte function (Sontheimer et al., 1994), but are inactivated in a depolarized resting membrane potential (Sontheimer and Waxman, 1992). Upregulation

+ of Kv1.4 could provide a more negative resting potential to make these Na channels operative, and thus to enhance astrocyte electrical signaling. If up-regulated Kv1.4 plays

85 an essential role in either the proliferation or the function of these cells, blocking Kv1.4 activity may have long-term side effects, preventing the potential remitting process.

CONCLUSION

This study not only has demonstrated demyelination-induced disruption of Kv1.2/

Kvβ2 JXP clustering, but also surprisingly has revealed down-regulated Kv2.1 in motor neuron soma and up-regulated Kv1.4 in spinal cord astrocytes correlating with EAE severity. More importantly, by comparing their changes at different stages of chEAE and rrEAE, especially the remitting stage, we provide novel insights into motor neuron damage/recovery and the potential role of Kv1.4-positive astrocytes in remyelination.

Therefore, this study has laid a solid foundation for developing an effective treatment for

MS symptoms via modulation of ion channel activity.

CHAPTER 3: ASTROCYTE REACTIVITY TO INFLAMMATION AND

NEURONAL DYSFUNCTION

INTRODUCTION

Normal functions of the CNS rely critically on the proper structure and function of the vascular system. Blood vessels provide neurons with oxygen and nutrients and protect them from toxins and pathogens. Neurons, in turn, control blood vessel dilation and contraction. The neurovascular unit consists of neurons, glia (astrocytes, microglia and oligodendrocytes), and vascular cells (endothelia, pericytes and smooth muscle cells). Vascular cells form the BBB and maintain the chemical and cellular composition of the neuronal microenvironment, which is required for proper functioning of neuronal synapses and circuits. Accumulating evidence shows some neurodegenerative diseases, such as Alzheimer’s disease, amyotrophic lateral sclerosis, and multiple sclerosis (MS), are initiated and perpetuated by vascular abnormalities (Zlokovic, 2008, Quaegebeur et al., 2011, Ransohoff and Brown, 2012). Understanding neurovascular coupling will advance the diagnosis, therapy, and prevention of these diseases.

Astrocytes play a key role in neurovascular coupling. Astrocytes outnumber neurons in human brain and spinal cord (Nedergaard et al., 2003) and fulfill important

87 roles in regulating synapse formation and function (Eroglu and Barres, 2010) and axon myelination (Sorensen et al., 2008, Nash et al., 2011b), as well as neurovascular coupling

(Petzold and Murthy, 2011). They are a diverse group of cells in the CNS and can be divided into two major groups. Protoplasmic astrocytes are found in gray matter and their endfoot processes ensheath synapses. Fibrous astrocytes reside in white matter and their endfeet contact myelin membranes and nodes of Ranvier (Barres, 2008). Endfeet of both types of astrocytes cover more than 90% of brain capillaries (Barres, 2008, Dunn and

Nelson, 2010).

MS is an inflammatory demyelinating disease with unknown origin. In brain biopsies of MS patients, lesions are characterized by extensive loss of myelin, axonal damage and inflammation (Lucchinetti et al., 2011, Brosnan, 2013, King et al., 2009,

Young et al., 2010). Activation of immune cells and disruption of the BBB are early events in MS pathology and have been extensively studied, but the role of astrocytes in

MS progression is relatively unexplored (Brosnan, 2013). GFAP was isolated from old

MS plaques in 1971 (Eng et al., 1971). However, astrocytes only began to gain attention in MS recently due to two channel proteins that are expressed at astrocytic endfeet. The water channel AQP4 regulates water balance, and its auto-antibody can lead to neuromyelitis optica (NMO) (Lennon et al., 2004, Hinson et al., 2012, Waters and

Vincent, 2008), arguably considered as a variant of MS (Barnett and Sutton, 2012). In the

CNS, AQP4 is predominantly localized to astrocytic endfeet contacting blood capillaries.

In NMO, astrocytic AQP4 is the primary target of an immune response that results in profound damage to astrocytes, breakdown of the BBB, secondary loss of myelin, and

88 apoptosis of oligodendrocytes (Brosnan, 2013, Hinson et al., 2012, Jarius and

+ Wildemann, 2010). An inwardly rectifying K channel, Kir4.1, is also present in astrocytic endfeet, and is involved in K+ buffering (Higashimori and Sontheimer, 2007,

Stephan et al., 2012, Bay and Butt, 2012). Kir4.1 is likely a target of the autoimmune response in a subgroup of MS patients (Srivastava et al., 2012). Our previous studies show that astrocytes in spinal cord white matter are activated during EAE, an animal model for MS (Jukkola et al., 2012). Kv1.4 is upregulated in astrocytes during EAE

(Jukkola et al., 2012). How astrocytes are altered in different brain regions is not known.

In the past, the astrocyte has been viewed as a cell that promotes inflammation and forms glial scars that hinder remyelination and axon growth. We now know that astrocytes can enhance CNS myelination by promoting the migration, proliferation, and differentiation of oligodendrocyte progenitor cells. Promotion of CNS myelination by astrocytes has been demonstrated in many culture models (Sorensen et al., 2008, Nash et al., 2011b,

Ishibashi et al., 2006, Watkins et al., 2008).

Astrocytes express neurotransmitter receptors that allow them to respond to neuronal activity (Schipke and Kettenmann, 2004). Astrocytes are activated in epilepsy and ataxia, diseases often caused by imbalances of excitatory and inhibitory neural networks. Glutamate receptors and transporters expressed on astrocytes are altered in epilepsy (Binder and Steinhauser, 2006), while a key enzyme in the glutamate cycle, glutamine synthetase, is down-regulated (Eid et al., 2008, Eid et al., 2004). AQP4 polarization to astrocytic endfeet was disrupted in the latent phase of the kainate model of epilepsy, before the onset of spontaneous seizure activity (Alvestad et al., 2013).

Astrocyte dysfunction and alterations are involved in other neurodegenerative diseases, including the spinocerebellar ataxia by Ataxin-7 mutant (Custer et al., 2006). Therefore, astrocytes play a key role in mediating neurovascular coupling. A better understanding of alterations to neurovascular units in different regions and in response to different stimuli will contribute to our understanding of (1) why certain brain regions are vulnerable to specific insults in certain neurological diseases, and (2) how CNS lesions initiate, extend, and repair in terms of neurons and glia.

In this study, we systematically examined alterations of astrocytic proteins on animal models for different diseases, including MS, neurodegenerative ataxia and epilepsy. Astrocytes are broadly activated in the CNS including both spinal cord and brain by autoimmune inflammation in EAE. On the other hand, altered neuronal activity, particularly synaptic activity, can also activate astrocytes. Thus, astrocytes are a good indicator for alterations in neurovascular coupling and a mediator of the vicious cycles in a number of neurodegenerative diseases.

METHODS

Reagents and antibodies

The nuclear dye (Hoechst 33342) and the lipophilic dyes (Fluoromyelin-green and

Fluoromyelin-red) were purchased from Invitrogen (Carlsbad, CA, USA). The following antibodies were used in our study: rabbit polyclonal anti- Kv1.4, anti-AQP4 (Alomone

Labs, Jerusalem, Israel), and anti-PKCγ (Santa Cruz Biotechnology, Dallas, TX); goat

90 polyclonal anti-GFAP (AbCAM, Cambridge, MA, USA); chicken polyclonal anti-

Vimentin (Millipore, Temecula, CA); and Dylight 488-, Dylight 649-, Cy3-, and Cy5- conjugated secondary antibodies (Jackson Immuno Research Laboratories, West Grove,

PA, USA). Myelin oligodendrocyte glycoprotein (MOG) peptide 35–55

(MEVGWYRSPFSRVVHLYRNGK) was purchased from Pro-Spec (Rehovot, Israel), and proteolipoprotein (PLP) peptide 139–151 (HCLGKWLGHPDKF) from Anaspec

(Fremont, CA, USA). Ground inactivated mycobacteria tuberculosis H37Ra and

Incomplete Freund’s Adjuvant were from Difco Laboratories (Detroit, MI, USA).

Induction of chronic EAE

Chronic EAE was induced in 12-week-old female C57BL/6 and Thy1-YFP transgenic mice (Jackson Laboratories, Bar Harbor, ME, USA) according to previously published methods (Jukkola et al., 2012). Briefly, myelin oligodendrocyte glycoprotein

(MOG) peptide 35–55 (1 mg/mL final concentration) was emulsified in sterile-filtered

PBS and Complete Freund’s Adjuvant (CFA) containing 2 mg/mL (final concentration) ground inactivated mycobacteria tuberculosis H37Ra. Mice were immunized with 100 μL of MOG/CFA or CFA only (control) by subcutaneous injection at four sites in the belly and in each hind footpad. Pertussis toxin was administered by tail-vein injection at 0 and

2 days post immunization (DPI). Disease progression was monitored by daily clinical scoring on a scale of 0–6 [0 = no symptoms, 1 = loss of tail tone, 2 = hindlimb paresis, 3 = moderate paralysis, 4 = paraplegia (complete hindlimb paralysis), 5 = quadriplegia, 6 = death or moribund state]. Grade 6 animals were removed from the study.

Induction of remitting-relapsing EAE

As previously published (Jukkola et al., 2012), procedures for the induction and monitoring of rrEAE in 12-week-old female SJL/J mice (Jackson Laboratories) were identical to chEAE procedures except for the use of a different myelin peptide, PLP 139–

151.

Cardiac perfusion, tissue fixation and sectioning

Animal tissues were collected and processed according to previously published methods, with minor modification (Jukkola et al., 2012). Briefly, mice were deeply anesthetized with avertin and perfused through the heart with 20–30 mL ice-cold PBS followed by 20 mL 4% formaldehyde in PBS (FA/PBS). The brain and spinal cord of

EAE mice were carefully removed and post-fixed overnight in 4% FA/PBS. Wildtype,

AnkG−/−, and Kv 3.1−/− mice were post-fixed in 4% FA/PBS for just 1 h, which allowed improved staining resolution of fine astrocyte processes. Tissues were cryoprotected in

30% sucrose for at least 24 hr, cut into 3-mm blocks using an acrylic brain matrix

(Braintree Scientific, Braintree, MA, USA), embedded in optimal cutting temperature

(OCT) media (Sakura Finetek USA, Inc., Torrance, CA, USA), and stored at −80°C until sectioning. The tissue blocks were cut with a Microm HM550 cryostat (Thermo

Scientific, Waltham, MA, USA) and the 40-μm sections were collected on Superfrost

Plus microscope slides (FisherScientific, Pittsburgh, PA, USA) for storage at −20°C.

Immunofluorescent staining

Tissues were stained according to previously published methods (Jukkola et al.,

2012), with minor modification. Briefly, sections were permeabilized in PBS/0.4% Triton

X-100 for 1 hr at room temperature (RT), blocked with 2.5% normal donkey serum for

1 hr at RT, and incubated in primary antibodies in blocking solution for 3 hr at RT and overnight at 4°C. The next day, the sections were rinsed 10 × 5 min, incubated for 3 hr with secondary antibodies in blocking solution, counterstained in Hoechst 33342 and/or

Fluoromyelin for 10 min, and again rinsed 10 × 5 min (all at RT). Slides were coverslipped using tris-buffered Fluoro-Gel mounting media (Electron Microscopy

Sciences, Hatfield, PA, USA). Staining for each primary antibody was performed at least twice on tissue from 3–6 mice per experimental cohort, except for the late rrEAE cohort in which only 2 mice were available.

Fluorescence light microscope and image capture

Fluorescence microscopy and image analysis procedures were adapted from previously published methods (Jukkola et al., 2012, Gu and Gu, 2011, Gu and Gu, 2010).

Images were captured with a Spot CCD camera RT slider (Diagnostic Instruments,

Sterling Heights, MI, USA) in a Zeiss Axiophot upright microscope using10×/0.30 and

20×/0.50 Plan Apo objectives and saved as 16-bit TIFF files. Exposure times were controlled so that the pixel intensities in the tissues of interest were below saturation, and the same exposure time was used for each group within an experiment. Image brightness

93 and contrast were adjusted using Adobe Photoshop 7.0 (Adobe Systems Incorporated,

San Jose, CA, USA).

Images to be quantified were chosen from 2 representative experiments for each antibody, and compared between experimental cohorts (3 mice quantified per cohort).

The number of images quantified is noted on the bar graphs in each figure. Images were analyzed with MetaMorph (Molecular Devices, Downingtown, PA, USA) and Sigmaplot

11.0 (Systat Software, Inc., Chicago, IL, USA) for fluorescence intensity quantification and statistical testing. Staining intensities for GFAP, Vimentin, Kv1.4, and AQP4 were measured by using the MetaMorph region measurement tool to sample and automatically calculate the average pixel fluorescence intensities of small circles (~ 50-pixel area) drawn on astrocyte soma and processes (10–20 sampled per image), and these intensities were averaged for each image. The background intensity of the slide was then measured and subtracted from the astrocyte staining intensity for each image. To account for variations in staining intensity from images obtained in different staining experiments, the data were normalized to controls stained side-by-side. The image fluorescence intensity values were averaged to obtain the mean fluorescence intensity for each experimental group, expressed as mean ± SEM. These fluorescence intensity measurements reflect the level of protein present within the cells stained by a particular antibody (higher intensity correlates with more protein), and do not give any indication of lesion size or cell number. To quantify cell numbers in hippocampus or cortex, the area of the region of interest in each image was first measured using the MetaMorph region measurement tool. Then, the GFAP + cells present in these regions were counted

94 manually and expressed as the number of cells per mm2, mean ± SEM. Statistical significance was determined between two groups using an unpaired Student’s t test, and among three or more groups using One-Way ANOVA followed by Fisher’s test.

Statistically significant differences from the Control group (for EAE experiments) or the

Wildtype group (for KO mice) are shown by an asterisk (*) in figures. All fluorescence intensity measurements were taken from the originally captured images.

Confocal microscopy and 3D reconstruction

High-magnification confocal images were captured with a Leica TCS SL confocal imaging system (Leica Microsystems, Mannheim, Germany), using a 100× HCX Plan

Apo CS oil immersion objective (numerical aperture = 1.40). Multiple channels were acquired simultaneously, or sequentially only if needed to eliminate channel bleed- through. Channel bleed-through was largely eliminated through optimization of the laser line intensity by acousto-optical tunable excitation filters, and by spectral detectors allowing precisely-defined bandwidth adjustment. The signal was averaged over eight scans in linescan mode. Images were saved as 8-bit TIFF files and adjusted for brightness and contrast using Adobe Photoshop 7.0.

For 3D reconstruction, z-stacks of images were collected with a 0.5-μm slice interval. Collapsed images were created using a maximum intensity projection of the z- stack. For each 3D image, the z-stack was visualized in three-dimensional cross section, and the cross-bars were centered on pertinent features of the image. The three one dimensional images were then exported together as an 8-bit TIFF file. Supplemental

95 movies were created using the Maximum Projection with Animation tool within the Leica

Confocal Software (v2.61). A maximum intensity projection of the z-stack was rotated 90 degrees around the Y-axis in a series of 18 or 36 steps and exported as an AVI file, and converted to an MPEG-1 file (MPG).

AnkG and Kv3.1 KO mice

AnkG KO mice were previously published (Jenkins and Bennett, 2001). The

Kv3.1 knockout (KO) mouse line was kindly provided by Dr. R. Joho at UT

Southwestern Medical Center and has been maintained using a PCR-based genotyping procedure as previously described (Ho et al., 1997, Hurlock et al., 2009, Sanchez et al.,

2000). The Kv3.1 KO mice were backcrossed with C57BL/6 for ten generations.

Although very rare, seizure convulsions (last only 10–15 sec; mouse returns to normal within a few minutes) can be observed occasionally from some Kv3.1 homozygote (−/−) but not heterozygote (+/−) mice. The following primers were used: forward primer

31 F775 (for both WT and knockout, 5′- GCG CTT CAA CCC CAT CGT GAA CAA

GA -3′), reverse primer 31R991 (for WT, 5′- GGC CAC AAA GTC AAT GAT ATT

GAG GG -3′), and reverse primer PNR278 (for knockout, 5′- CTA CTT CCA TTT GTC

ACG TCC TGC AC -3′). Three Kv 3.1 KO (−/−) and three control C57BL/6 mice of either sex at the age of 2–4 months were used in this study.

All animal experiments have been conducted in accordance with the NIH Animal

Use Guidelines and were approved by the Ohio State University Institutional Animal

Care and Use Committee (IACUC).

RESULTS

Figure 22. Clinical scores (A) and body weight (B) of mice with chEAE.

AQP4, Kv1.4, Vimentin and GFAP are upregulated in SCWM at different stages of EAE

To understand how inflammatory autoimmune insults affect astrocytes in the

CNS, we performed the MOG35-55-induced chronic EAE (chEAE) (Figure 22), as previously described (Jukkola et al., 2012). First, we examined spinal cord white matter

(SCWM) and gray matter (SCGM). In SCWM of control mice, AQP4 and Vimentin

(Vim) were colocalized in GFAP-positive astrocytes (Figure 23A, C). At the peak stage of chEAE, GFAP was significantly upregulated around demyelinated lesion sites, where

AQP4 was also upregulated (Figure 23B). The Kv1.4 level was upregulated together with

Vim around the lesion sites (Figure 23C,D). Both GFAP and Vim are upregulated in reactive astrocytes. Although both are intermediate filaments, they differ in temporal and spatial expression patterns in astrocytes. Vim is expressed in radial glia and immature astrocytes, as well as other cell types, whereas GFAP replaces Vim in many mature astrocytes (Bignami et al., 1982, Lazarides, 1982, Eriksson et al., 2009). In adult CNS,

Vim expression may indicate astrocyte proliferation. In the present study, we use it as a marker for distinct subpopulations of astrocytes. In our recent study, we found that Kv1.4 was upregulated in astrocytes of SCWM near the EAE lesion sites (Jukkola et al., 2012), but not in astrocytes in other regions of the CNS. Interestingly, both Kv1.4 and Vim expression were upregulated in SCWM (Figure 23C, D). Therefore, we examined Vim expression in other regions of the CNS during EAE in following studies, which may suggest potential astrocytic proliferation.

Figure 23. Activation of astrocytes in chEAE spinal cord white matter. White matter (WM) and gray matter (GM) in spinal cord longitudinal sections were stained with FMG (green) and nuclear dye (blue). (A) and (B) contain both GM and WM tissue, while (C) and (D) show only WM. Spinal cord sections, control (A) and EAE peak (B), were costained for GFAP (green, top), AQP4 (red), Hoechst (blue), and FMG (green, bottom). Co-staining of Kv1.4 (green, top), Vim (red), Hoechst (blue) and FMG (green, bottom) were also performed on control (C) and EAE (D) spinal cord sections.

Using high magnification confocal imaging, we examined alterations of the neurovascular unit (axons, astrocytes, and capillaries) of Thy1-YFP transgenic mice.

Axons were revealed by YFP fluorescence, astrocytes were stained for GFAP, and capillaries were revealed with AQP4 staining. In control, fibrous astrocytes in SCWM extended processes parallel with axons, while AQP4 in astrocyte endfeet colocalized with putative nodes of Ranvier along YFP-positive axons (Figure 24A). At the peak stage of chEAE, GFAP and AQP4 staining disappeared in the center of the lesion, but was often upregulated surrounding lesion sites (Figure 24B). YFP-positive axons were often fragmented near the lesion, indicating axonal degeneration (Figure 24B). EAE lesions were only observed in WM but not in GM.

100

Figure 24. High magnification confocal image stacks were obtained from control (A) and EAE (B) Thy1-YFP transgenic mice. Images contain YFP (green), GFAP (blue) and AQP4 (red). The collapsed 2D image is on the left, and single-plane images in three dimensions (3D) are on the right. In (A), the crossbars are centered on a putative node of Ranvier. In (B), the crossbars are centered on the AQP4+/GFAP + lesion edge. Scale bars, 50 μm.

101

Differential alterations of myelin and astrocytes in the late and remitting EAE in SCWM

We further compared altered protein expression in astrocytes at the late stage of chEAE and at the remitting stage of remitting-relapsing EAE (rrEAE). In late chEAE, demyelinated lesion areas were very extensive. Both AQP4 and GFAP colocalized very well within and around the lesion sites (Figure 25A, Figure 27A). Kv1.4 and Vim levels also markedly increased in these dense astrocytic fibers (Figure 25B, Figure 27A). In contrast, lesions in remitting rrEAE were fewer and much smaller, and enclosed by

GFAP- and AQP4-positive processes (Figure 25C, Figure 27B). The increase of Kv1.4,

Vim and GFAP but not AQP4 levels were significant at the remitting stage (Figure 25C,

D; Figure 27B). High magnification confocal images revealed GFAP staining in extended processes of astrocytes, colocalizing with AQP4 in lesion sites at the late stage of chEAE

(Figure 26A). In contrast, at the remitting stage of rrEAE, a small lesion was fully enclosed by GFAP- and AQP4-positive processes (Figure 26B). This result indicates that activated astrocytes may deter the infiltration, proliferation and migration of immune cells to limit the lesion at the initial and remitting stages of EAE.

Cerebellar astrocytes are differentially activated at different stages of EAE

EAE lesions were observed in cerebellar white matter, especially in EAE mice with high clinical scores (Jukkola et al., 2012). We wondered whether cerebellar gray matter including various astrocytes is affected during EAE. In the cerebellum of control mice, GFAP was expressed at relatively low levels in WM (highest), granule cell layer 102

(medium), and molecular layer (lowest) (Bergmann glia). In the molecular layer,

Bergmann glia expressed both GFAP and Vim at low levels, whereas protoplasmic astrocytes surrounding Purkinje neuron soma expressed higher levels of GFAP but with no expression of Vim (Figure 28A). AQP4 was expressed rather diffusedly but concentrated around blood capillaries (Figure 28A). At the peak stage of rrEAE, the

GFAP level significantly increased in both Bergmann glia and protoplasmic astrocytes

(Figure 28B, E). At the remitting stage, GFAP levels reduced (Figure 28C, E). At the relapsing stage, GFAP staining increased again (Figure 28D, E). The Vim level in the

Bergmann glia did not change much, and it was not expressed in the protoplasmic astrocytes (Figure 28). The AQP4 level only moderately increased at the peak stage of rrEAE (Figure 28).

103

Figure 25. Astrocytic proteins are altered in lesion sites of at the late stage of chEAE and at the remitting stage of rrEAE. A and B, staining patterns of astrocytic proteins at the late stage of chEAE. C and D, staining pattern of astrocytic proteins at the remitting stage of rrEAE.

104

Figure 26. A, High magnification view of a lesion site at the late stage of chEAE. B, Confocal images of a lesion site at the remitting stage of rrEAE. Collapsed 2D image is on the top, and single-plane images in three dimensions are on the bottom. In (A), the crossbars are centered on an astrocyte with colocalizing AQP4 and GFAP. In (B), the crossbars emphasize the absence of AQP4 and GFAP in the lesion core. Scale bars, 50 μm.

105

Figure 27. A, Summary of the alteration of astrocytic proteins in chEAE. B, Summary of the alteration of astrocytic proteins in rrEAE. The “n” number of quantified images is provided for each bar. One-way ANOVA followed by Fisher’s test. Asterisk (*) shows significant difference from Control staining intensity for each antibody, p < 0.05.

106

Therefore, astrocytic GFAP levels in cerebellar gray matter are altered during

EAE, in addition to the changes in cerebellar WM, where EAE lesions were sometimes observed. The alterations of neurovascular units in cerebellar WM were similar to those of SCWM, except that all lesion sites in cerebellar white matter were enclosed, overall smaller than those in SCWM (Figure 29A, B). AQP4 was more concentrated in the granule cell layer. Its expression increased around lesions in cerebellar WM. In the molecular layer, its expression was around the blood vessels.

Fibrous GFAP + and AQP4+ processes were observed in SCWM and white matter of cerebellum and corpus callosum. Their organization is quite similar in all of these areas. Although upregulation of GFAP was also observed in corpus callosum, no lesion was observed there at any stage of EAE (data not shown).

107

Figure 28. Differentially altered expression of astrocytic proteins in the cerebellum of EAE mice. A, The confocal image stack of cerebellar molecular layer that was stained for GFAP (green), AQP4 (red) and Vim (blue) from a control mouse. Collapsed 2D image is on the left and single-plane images in 3D are on the right. B, The images at the peak stage of an rrEAE mouse. C, The images at the remitting stage of an rrEAE mouse. D, The images at the relapsing stage of an rrEAE mouse. In (A-D), the crossbars are centered on astrocytic endfeet with colocalizing AQP4 and GFAP. E, Summary of changes of protein levels during rrEAE in cerebellar molecular layer. One-Way ANOVA followed by Fisher’s test. *, significant difference from Control for each antibody, p < 0.05. The “n” number of quantified images is provided for each bar. Scale bars, 50 μm.

108

Figure 29. A, The confocal images of cerebellar WM from a Thy1-YFP transgenic mouse. The crossbars are centered on astrocytic endfeet with colocalizing AQP4 and GFAP. B, The confocal images of cerebellar WM at the peak stage of chEAE. The crossbars are centered on the lesion border with colocalized GFAP and AQP4. YFP (green), AQP4 (red) and GFAP (blue). Collapsed 2D image is on the left and single-plane images in 3D are on the right.

109

Hippocampal astrocytes are activated at the peak and late stages of EAE

Although no lesion was clearly observed in the hippocampus during the course of

EAE, we surprisingly observed the upregulation of GFAP staining in the hippocampus.

GFAP-positive protoplasmic astrocytes were observed in the hippocampi of control mice, providing non-overlapping but complete coverage of the entire hippocampus (Figure

30A). During EAE, GFAP levels markedly increased, while AQP4 levels did not change

(Figure 30A). In contrast to SCWM astrocytes and Bergmann glia, hippocampal astrocytes normally do not express Vim, but during EAE its expression level significantly increased (Figure 30B, C, and F). We further examined the changes of the neurovascular unit within the hippocampus using the Thy1-YFP transgenic mice. In the control, the hippocampal neuron soma, dendrites and axons were revealed by YFP fluorescence.

Astrocytes were stained with GFAP antibody, and AQP4 staining revealed ensheathed blood capillaries (Figure 31, top). During EAE, neuronal morphology including axons, dendrites and dendritic spines, remained largely unchanged, whereas GFAP staining markedly increased. The level of AQP4 did not change, but perivascular GFAP astrocyte staining significantly increased (Figure 31, bottom).

110

Figure 30. Activation of astrocytes in the hippocampus of EAE mice. A, Increased GFAP but not AQP4 staining in the hippocampus of EAE mice. B, Increased Vim staining in the hippocampus during EAE progression. C, Enlarged images of individual astrocytes in the hippocampus clearly show the increase of Vim staining. (D) Summary of the levels of astrocytic proteins at different stages during EAE progression. One-Way ANOVA followed by Fisher’s test, *, significant difference from Control for each antibody, p < 0.01. (E) GFAP + cell number was not changed during EAE. The “n” number of quantified images is provided for each bar. Scale bars, 200 μm in A and B.

111

Figure 31. High magnification confocal image stacks were obtained from control (top) and EAE (bottom) Thy1-YFP transgenic mice. Images contain YFP (green), GFAP (blue) and AQP4 (red). The collapsed 2D image is on the left, and single-plane images in 3D are on the right. The crossbars reveal astrocytic endfeet with colocalizing AQP4 and GFAP. Scale bars, 50 μm.

112

Cortical astrocytes are activated at the peak and late stages of EAE

In the cortex of control mice, GFAP staining was relatively weak, almost invisible in some areas. Like in the hippocampus, AQP4 staining was mainly concentrated around small-diameter blood capillaries. At the peak stage of rrEAE, GFAP markedly increased, whereas AQP4 did not appear to change much. At the remitting stage of rrEAE, GFAP staining was significantly increased. At the relapsing stage of rrEAE, GFAP staining intensity further increased (Figure 32A,B). Perivascular GFAP appeared increased significantly (Figure 32A). There was no Vim staining in the cortex, even in EAE progression, indicating that the properties of cortical astrocytes are different from those of hippocampal astrocytes, even though they are all protoplasmic astrocytes in GM.

Next, we performed high-resolution confocal imaging of the cortex using Thy1-

YFP mice. The dendrites and spines were clearly seen in cortical pyramidal neurons

(Figure 33). Astrocytes and blood vessels intertwined with neurons. Dendritic spines partially colocalized with background AQP4, which possibly represented the synapse- ensheathing endfeet of astrocytes. Astrocytic endfeet are known to colocalize with synapses. Dendritic spines are the postsynaptic structures of excitatory synapses. In the cortex of EAE mice, dendritic branches and spines were not changed, but GFAP intensity and the number of GFAP + astrocytes increased (Figure 32B, C; Figure 33). In particular, the GFAP + processes surrounding blood capillaries positive for AQP4 appeared to increase significantly and capillaries appeared dilated (Figure 33). Thus, cortical astrocytes in EAE are primarily altered at the perivascular endfeet. The dendritic spines and surrounding AQP4 staining puncta do not appear to be altered (Figure 33). 113

Figure 32. Alteration of GFAP and AQP4 in the cortex of EAE mice. A, Alterations of astrocytic proteins in the cortex at different stages of rrEAE. B, Summary of the alterations of intensities. C, Summary of increased GFAP + cells. (B, C) One-Way ANOVA followed by Fisher’s test. *, significant difference from Control for each antibody, p < 0.05. **, significant difference from Control for each antibody, p < 0.01. The “n” number of quantified images is provided for each bar. Scale bars, 200 μm.

114

Figure 33. High magnification confocal image stacks were obtained from control (top) and EAE (bottom) Thy1-YFP transgenic mice. Images contain YFP (green), GFAP (blue) and AQP4 (red). The collapsed 2D image is on the left, and single-plane images in 3D are on the right. Scale bars, 50 μm.

115

Astrocyte alterations in AnkG and Kv3.1 KO mice

To understand how neuronal activity regulates astrocyte and BBB function, we focused our studies on AnkG and Kv3.1 KO mice. Crucial for efficient initiation and saltatory propagation of action potentials along myelinated axons of vertebrates, Nav channels are clustered at the axon initial segment and nodes of Ranvier (Stuart et al.,

+ 1997, Clark et al., 2009). The clustering of Nav channels, as well as some K channels and cell adhesion molecules at axon initial segments and nodes, is mediated by AnkG

(Jenkins and Bennett, 2001, Zhou et al., 1998, Salzer, 2003, Bennett and Baines, 2001,

Pan et al., 2006). AnkG links these key membrane proteins to the actin cytoskeleton via spectrins and functions as a gate to maintain the axon-dendrite polarity (Song et al., 2009,

Sobotzik et al., 2009). Since most nodes in cerebellar WM have no adaptor protein AnkG

(Jenkins and Bennett, 2001), the clustering of Nav channels at these nodes is eliminated and thereby nodal excitability is altered, which may conversely impact the astrocytic endfeet around the nodes. It has never been shown before how astrocytes are altered in

AnkG KO mice.

The Kv3 (Shaw) channel subfamily contains 4 members (Kv3.1 to Kv3.4). Kv3.1 and Kv3.2 carry sustained currents, while Kv3.3 and Kv3.4 carry transient currents. Kv3 channels play a critical role in rapid spiking of action potentials in some neurons (Rudy and McBain, 2001, Jonas et al., 2004, Gu and Barry, 2011). Axonal targeting and unique channel biophysical properties of Kv3 channels—high activation threshold, fast activation/deactivation kinetics—allow neurons to fire action potentials at high frequency

(Rudy and McBain, 2001, Coetzee et al., 1999, Gu, 2013, Gu et al., 2012, Xu et al.,

116

2007b, Xu et al., 2010). KO mice were generated to study the functions of different Kv3 channels. Kv3.1 KO mice are viable and fertile (Ho et al., 1997). Their spontaneous locomotor and exploratory activities remain unchanged, although their coordinated motor skill and muscle contraction are worsened. Kv3.1 is expressed in cerebellar granule cells and deep nuclei neurons, which are critical for timing or gait patterning in motor functions (Hurlock et al., 2009, Zagha et al., 2008, Hurlock et al., 2008). Kv3.1b has also been shown to localize in some nodes of Ranvier in spinal cord white matter (Devaux et al., 2003). We were interested to determine how alteration of neuronal activity in the absence of AnkG or Kv3.1 would affect astrocytes and BBB function.

Since the AnkG KO is missing the AnkG isoform primarily expressed in the cerebellum and Kv3.1 is highly expressed in the granule cells in the cerebellum, we focused our studies on the cerebellum. In AnkG KO mice, GFAP was increased in the cerebellar molecular layers, from high to low in the following order, molecular layer,

Purkinje cell layer, and granule cell layer. PKCγ in WT mice labeled the Purkinje neurons including soma and dendrites (Figure 34A). In the AnkG KO mice, dendritic branches of some Purkinje neurons reduced, and some Purkinje neurons were degenerated, where GFAP staining signals markedly increased (Figure 34B, D). These signals likely result from both increased GFAP expression and increased number of astrocytes. The increased GFAP-positive astrocyte processes were not as oriented as the regular Bergmann glia cells (Figure 34B). In Kv 3.1 KO mice, GFAP signal intensity also significantly increased in the molecular layer of the cerebellum (Figure 34C, D). In summary, in the molecular layer both GFAP and AQP4 staining increased in both AnkG

117 and Kv3.1 KO mice (Figure 34D). We next examined the granule cell layer. In AnkG KO mice, GFAP and AQP4 staining in the granule cell layer were higher (Figure 35B,D). In

Kv3.1 KO mice, GFAP and AQP4 staining were both higher (Figure 35C, D).

Finally, we examined astrocytic proteins in the hippocampus and cerebral cortex.

In the hippocampi of AnkG and Kv3.1 KO mice, GFAP staining intensity was increased

(Figure 36A-C). Interestingly, empty pockets of GFAP and AQP4 staining were found in

Kv3.1 KO mice (Figure 36C), potentially indicating the death of some astrocytes. In the cortex of AnkG KO and Kv3.1 KO mice, GFAP staining intensity was also increased, but in a non-uniform way (Figure 36D-F). The density of the brain capillary network appeared increased in both the hippocampus and cortex in Kv3.1 KO mice (Figure

36C,F).

118

Figure 34. Upregulation of GFAP and AQP4 in the cerebellum in AnkG and Kv3.1 KO mice. A, High magnification image stacks of cerebellar molecular layer stained with anti-PKCγ (green) and anti-GFAP (red) antibodies. The collapsed 2D images are on the top, and single-plane images in 3D are on the bottom. B, Confocal image stacks from the AnkG KO mice. C, Confocal image stacks from the Kv3.1 KO mice. In (A,C), the crossbars reveal radially oriented GFAP + Bergmann glial processes in (A) WT and (C) Kv3.1 KO mice. In (B), the crossbars are centered on highly upregulated GFAP + astrocyte processes in the absence of Purkinje neurons in an AnkG KO mouse. D, Normalized fluorescence intensity in the molecular layer. One-Way ANOVA followed by Fisher’s test. *, significant difference from Wildtype for each antibody, p < 0.05. **, significant difference from Wildtype for each antibody, p < 0.01. The “n” number of quantified images is provided for each bar. Scale bars, 50 μm.

119

Figure 35. (A), A single confocal image of the granule cell layer in a WT mouse. (B), An image from the AnkG KO mice. (C), An image from the Kv3.1 KO mice. D, Normalized fluorescence intensity in the granule cell layer. One-Way ANOVA followed by Fisher’s test. *, significant difference from Wildtype for each antibody, p < 0.05. **, significant difference from Wildtype for each antibody, p < 0.01. The “n” number of quantified images is provided for each bar. Scale bars, 50 μm.

120

Figure 36. Altered astrocytes in the hippocampus and cortex in AnkG KO and Kv3.1 KO mice. The confocal image stacks of hippocampus (A-C) and cortex (D-F) were costained for GFAP (green) and AQP4 (red) from WT (A,D), AnkG KO (B,E) and Kv3.1 KO (C,F) mice. The collapsed 2D image is on the top and single-plane images in 3D are on the bottom. The crossbars are centered on astrocyte endfeet with colocalizing AQP4 and GFAP. Scale bars, 100 μm.

121

DISCUSSION

In this study, we show alterations of neurovascular units in different regions of the brain and spinal cord in response to inflammation and altered neuronal activity. The analysis of neurovascular units has been carried out with high-resolution confocal imaging and 3D reconstruction. This study mainly focuses on astrocytic proteins and provides subcellular structural details regarding disease progression induced by either inflammatory insults or neuronal defects. Our studies indicate striking intrinsic heterogeneity of astrocytes in the CNS, which likely underlies their differential responses to various stimuli (Figure 37).

Both fibrous and protoplasmic astrocytes are activated during EAE, but only fibrous

astrocytes are selectively damaged around lesion sites

The organization of fibrous astrocyte processes, AQP4+ endfeet and axons is quite similar in the WM of spinal cord, cerebellum, and corpus callosum. GFAP and AQP4 were upregulated during EAE. During EAE progression, lesions were observed in WM of spinal cord and cerebellum (Figures 23, 24, 25,26, 29), but not in corpus callosum (data not shown). Dramatic reorganization of neurovascular units was observed only in WM at or surrounding the lesion sites, but not in gray matter . Thus fibrous astrocytes appeared to be specifically damaged in inflammatory demyelination. Both axonal demyelination and degeneration were observed in the EAE spinal cord. At the remitting stage, the lesion

122 size and number were reduced in spinal cord, and the levels of GFAP reduced in SCWM and cerebellar GM, but not cortical GM (Figures 25, 26, 27, 28, 32).

Vim is an intermediate filament and one of the markers for astrocytes. In control animals, Vim is expressed in SCWM and cerebellar Bergmann glia, but not in SCGM, cerebellar Purkinje and granule cell layers, hippocampus, and cortex (Figures 23, 28, 30,

32). During EAE, the level of Vim markedly increased in SCWM and hippocampus

(Figures 23, 30). Its level increased moderately in cerebellar Bergmann glia (Figure 28), and did not increase at all in other layers of cerebellum, or in cortex (Figures 28, 32).

This result shows that astrocytes that are intrinsically different respond to inflammation differently. In particular, our result that GFAP is upregulated in the hippocampus during

EAE is consistent with a recent study (Wu et al., 2012). Our finding that Vim is upregulated in hippocampal astrocytes during EAE indicates a distinct region-specific alteration in EAE. Taken together, our results add to the growing body of evidence that the hippocampus is affected in EAE, despite the lack of clear immune-cell mediated lesions characteristic of spinal cord. Recent evidence shows that in addition to astrocytes, microglia are activated in EAE (Kurkowska-Jastrzebska et al., 2013). The hippocampal

CA1 region decreases in volume in EAE, with specific loss of GABAergic interneurons

(Ziehn et al., 2010). Synaptic function and immunostaining are also reduced in EAE, but can be recovered with testosterone (Ziehn et al., 2012b) or estriol treatment (Ziehn et al.,

2012a), which are interesting results due to the gender disparity in MS incidence.

Because astrocyte activation and changes in neuronal activity occur in EAE, this could

123 correlate with the role of astrocytes in response to neuronal defects that we observed in

AnkG−/− and Kv3.1−/− mice.

124

Figure 37. Diagram of regulation of fibrous and protoplasmic astrocytes by inflammation and neuronal activities. A, Fibrous astrocytes in the WM with endfeet contacting blood capillaries and nodes of Ranvier. B, Protoplasmic astrocytes in the GM with endfeet contacting blood capillaries and synapses. (This figure was created by Dr. Chen Gu.)

125

Increased GFAP staining around blood vessels is observed in EAE mice, which may reflect increased proliferation of perivascular astrocytes in gray matter of the brain.

In a recent study using live, two-photon imaging to investigate the astrocyte response to a cortical stab-wound injury, a limited number of astrocytes (~20%) were found to proliferate, but ~70-80% of these were located immediately adjacent to blood vessels

(Bardehle et al., 2013). No astrocytes were found to migrate toward the lesion, but some became polarized and extended GFAP + processes toward the injury site (Bardehle et al.,

2013). It is unclear whether these polarized processes contributed to increased staining of perivascular astrocyte endfeet. Further study is needed to elucidate the cytokines or other secreted factors that may induce perivascular astrocyte proliferation, and how these results obtained in the local stab-wound response might correlate with the more general autoimmune insult in EAE. An interesting candidate is interferon-γ, a cytokine with major effects in MS and EAE (Popko et al., 1997) that has been shown to increase astrocyte proliferation in vitro and in vivo (Yong et al., 1991). Moreover, a recent study indicates that in normal physiological conditions astrocytic endfeet play a major role in regulating the flow of cerebrospinal fluid and solutes into the brain parenchyma, and that

AQP4 is important to this process (Iliff et al., 2012). It is possible that perivascular astrocyte proliferation may augment the astrocytic roles in fluid and solute handling and help to prevent the leakage of the BBB during EAE progression. The analysis of neurovascular units can provide more detail regarding damaged nerve tissues during the progression of various neurological diseases. Although lesions were not observed in EAE in gray matter, astrocyte morphology did change, GFAP staining was increased, and

126 astrocytes proliferated, all of which may be indicators of neuroprotective function of astrocytes.

Neuronal activity regulates astrocytic protein expression

Purkinje neuron degeneration was observed in AnkG KO mice (Jenkins and

Bennett, 2001). Highly ordered Bergmann glial processes became irregular, and GFAP staining markedly increased at the same time. The highest staining of GFAP was often associated with the loss of Purkinje cells (Figure 34B). Kv3.1 is mainly expressed in cerebellar molecular and granule cell layers in WT mice. The highest upregulation of

GFAP was found in the molecular cell layers in Kv3.1 KO mice (Figure 34C,D).

The patchy loss of AQP4 staining in the hippocampus of Kv3.1 KO mice may indicate the loss of astrocytes. Interestingly, this was not observed in AnkG KO mice or

EAE mice at any stage (Figure 36C). This phenomenon of patchy AQP4 loss has been reported previously in the spinal cord gray matter of mice injected with AQP4 autoantibody-positive serum from NMO patients (Chan et al., 2012), but has not previously been reported in any KO mice. Kv3.1 channels are expressed in parvalbumin- positive fast spiking interneurons in the hippocampus, but it is unclear how interneuronal dysfunction in the absence of Kv3.1 could cause loss of astrocytes. A loss of astrocytes has been observed in the pilocarpine-induced epilepsy model (Kim et al., 2010), which is interesting because although very rare, we did see occasional seizures in our Kv3.1 KO mice. Indeed, synaptic transmission does appear to play a major role in regulating astrocytic protein expression.

127

Differential localization of AQP4 in fibrous and protoplasmic astrocytes

AQP4 were highly colocalized with GFAP + processes in the WM, different from the pattern in the GM where it is highly concentrated in astrocytic endfeet contacting blood vessels. AQP4 expression increased in SCWM in EAE, which is consistent with a recent study in EAE and correlates with increased AQP4 levels previously shown in MS patient tissues (Miyamoto et al., 2009, Sinclair et al., 2007). This contrasts with loss of

AQP4 expression in NMO, a demyelinating disease in which AQP4 autoantibodies are not only a highly specific biomarker for the disease, but are involved in disease pathology as well (Bennett et al., 2009), despite some controversy for the latter (Kinoshita and

Nakatsuji, 2012). In the early stage of NMO but not MS, AQP4 and GFAP are lost before the loss of MBP staining in NMO lesions, which can occur in WM or GM in mainly perivascular areas (Misu et al., 2007). Patient MRIs also show that NMO lesions occur in both WM and GM (Matthews et al., 2013). AQP4 autoantibody-induced demyelination is thought to occur by oligodendrocyte apoptosis secondary to loss of astrocytes and astrocytic trophic support (Bukhari et al., 2012, Marignier et al., 2010). The BBB is more severely disrupted in NMO than in MS (Tomizawa et al., 2012), but T cells are not required for lesion formation, as shown in mice administered IgG from NMO patients

(Saadoun et al., 2011). EAE susceptibility is almost eliminated in AQP4−/− mice (Li et al., 2011, Li et al., 2009), and secretion of proinflammatory cytokines from AQP4 −/− astrocytes was reduced in vitro (Li et al., 2011). However, AQP4 deficiency significantly increased the extent of neuron loss, demyelination, and motor dysfunction in a spinal cord contusion injury model (Kimura et al., 2010). Taken together, these results suggest

128 that AQP4 facilitates disease progression after breakdown of the BBB, but also has a neuroprotective effect, both of which could be mediated by the central role of astrocytes in neurovascular coupling.

Functional consequences of astrocyte activation

Astrocytes are increasingly recognized as an important player in MS, amyotrophic lateral sclerosis, Alzheimer’s disease, epilepsy, stroke, spinal cord injury, and other neurological disorders (Sofroniew and Vinters, 2010). In response to a variety of CNS disorders and pathologies, astrocytes undergo tremendous changes in morphology and function, becoming activated or reactive astrocytes. These changes, which may be either reversible or irreversible, are not an all-or-none phenomenon, but are finely graded in a context-dependent manner regulated by specific signaling pathways (Sofroniew and

Vinters, 2010). Astrocytes exhibit excitability mainly by increasing intracellular Ca2+ concentration (Charles et al., 1991, Cornell-Bell et al., 1990). Astrocytes relay information to other cells through gap junctions and by releasing gliotransmitters

(Nedergaard et al., 2003, Shigetomi et al., 2008, Halassa et al., 2007, Perea et al., 2009,

Volterra and Meldolesi, 2005). Astrocytes promote myelination in response to electrical activity by releasing the cytokine leukemia inhibitor factor (Ishibashi et al., 2006), or through other means, such as by coating axons with myelin-promoting extracellular matrix molecules, providing lipids for myelin synthesis, or modulating electrical activity.

Various factors produced from astrocytes can also impact myelination (Moore et al.,

2011), such as platelet-derived growth factor, fibroblast growth factor, ciliary

129 neurotrophic factor, leukemia inhibitor factor, and insulin-like growth factor, as well as gliotransmitters including glutamate and ATP. Interestingly, some of these factors can further stimulate astrocytes via a positive feedback loop. Functional consequences and signaling pathways of astrocyte activation will be interesting topics for future studies in various diseases.

CONCLUSIONS

Taken together, the analysis of neurovascular units can provide more detail regarding the progression of diseases induced by either inflammatory insults or neuronal dysfunction. High-resolution confocal imaging with 3D reconstruction is a powerful tool to illustrate the changes under abnormal conditions. Better understanding their roles contributes to the development of novel strategies of neuroprotection and repair for various diseases, through activity-dependent regulation of neurovascular coupling.

130

CHAPTER 4: IN VIVO ION CHANNEL TRANSPORT MECHANISMS

Many ion channels may be affected by demyelinating disease, and as shown in chapter 2, may be altered in their distribution pattern in neurons. Therefore, it is necessary to consider the mechanisms of targeting and transport of these ion channels throughout the neuron. This chapter describes collaborative work in which I have taken part in the lab, working out the mechanism of transport for key channels in the neuron.

The major biochemical and cell biological experiments for these studies were performed by Dr. Joshua Barry and Dr. Yuanzheng Gu. My role was in relating the exciting in vitro findings to relevant physiological situations in vivo, to determine whether the mechanisms worked out in the primary neuron cultures were in fact occurring in neurons within their physiological context.

KV3.1 CLUSTERS AND ACTIVATES KINESIN MOTORS

INTRODUCTION

Previous studies in the Gu lab had identified Kv3.1 as the first ion channel known to directly bind KIF5 motor proteins (Xu et al., 2010). This binding requires

131 tetramerization, so that only fully assembled Kv3.1 channels will bind KIF5 (Xu et al.,

2010). Dr. Joshua Barry performed systematic mutagenesis studies and identified a 70- residue region of the KIF5 tail that formed the Kv3.1 binding site, including three residues that were crucial for binding (Barry, 2013). He then used surface plasmon resonance to measure the binding affinity between Kv3.1 and KIF5, which turned out to be a high-affinity interaction (Kd = 6.0±1.4×10−8 M)(Barry, 2013). The KIF5 tail contains binding sites for the motor domain and for microtubules (Wong and Rice, 2010), both of which can inactivate the KIF5 motor. The high-affinity interaction of Kv 3.1with

KIF5 competes for these inhibitory binding sites, and activates the kinesin motor when binding occurs. Pulldown assays performed by Dr. Barry suggested that up to four KIF5 motors could be bound by a single tetramerized Kv 3.1 channel (Barry, 2013). Thus,

Kv3.1 can cluster KIF5 motors by binding several at the same time. With likely many

Kv3.1 channels loaded on a carrier vesicle, this effect would be compounded and many, perhaps hundreds, of KIF5 motors could be bound to a single vesicle (Barry, 2013).

Interestingly, other cargo molecules with binding sites overlapping the Kv3.1 binding site,

SNAP25 and VAMP2, did not cluster or activate the KIF5 motor (Barry, 2013), probably because they are not tetramerized. To follow up on these studies, Dr. Yuanzheng Gu performed many live cell imaging experiments to confirm that KIF5 and Kv3.1 do functionally interact during transport within the axon, and that Kv3.1 has an activating effect upon KIF5 (Barry, 2013).

132

My role in this project was to relate these in vitro findings to physiologically relevant situations in animal models, to determine the importance of the interaction between KIF5 and Kv3.1 in vivo.

METHODS

Kv 3.1 knockout mouse backcrossing and genotyping

The Kv3.1 knockout (KO) mouse line was kindly provided by Dr R. Joho at

University of Texas Southwestern Medical Center. Because the Kv 3.1 KO mice were obtained on a mixed genetic background, I backcrossed them with a C57BL/6 mouse line for ten generations. Since then, the mice have been maintained using a Kv3.1 KO heterozygote X heterozygote breeding scheme, because Kv3.1 homozygotes do not breed well. In routine maintenance of this mouse line, I have occasionally observed seizures in

Kv3.1 homozygous (−/−) but not heterozygous (+/−) mice. A PCR-based procedure has been used to genotype the mice, as described previously (Ho et al., 1997, Sanchez et al.,

2000, Hurlock et al., 2009). The following primers were used: forward primer 31F775

(for both WT and knockout, 5′-GCG CTT CAA CCC CAT CGT GAA CAA GA-3′), reverse primer 31R991 (for WT, 5′-GGC CAC AAA GTC AAT GAT ATT GAG GG-3′), and reverse primer PNR278 (for knockout, 5′-CTA CTT CCA TTT GTC ACG TCC

TGC AC-3′). Three Kv3.1 KO (−/−) and three control BL6 mice of either sex at the age of 2–4 months were used in the immunofluorescence study.

133

Immunofluorescence staining on mouse brain sections and cluster quantification

After cardiac perfusion and tissue fixation, mouse cerebellum was removed, post- fixed, and then embedded in optimal cutting temperature media (Sakura Finetek USA,

Inc., Torrance, CA) and stored at −80°C until sectioning. Coronal sections (40-µm thickness) of mouse cerebellum were cut with a Microm HM550 cryostat (Thermo

Scientific, Waltham, MA) and collected on Superfrost Plus microscope slides

(FisherScientific, Pittsburgh, PA). Immunostaining and imaging were performed as previously described (Jukkola et al., 2012). The whole images captured with a 40× objective lens were processed for quantification using Metamorph by flattening the background, thresholding the images to include only the 5% highest intensity pixels, and binarizing the image. Kif5B puncta were quantified in ImageJ using the ‘Analyze

Particles’ tool to count all pixel clusters ranging in size from 9–300 pixels.

RESULTS

We sought to determine the effects of Kv3.1 deficiency upon KIF5 clustering in vivo. First, I confirmed the genotype of the Kv3.1 KO mice using PCR (Figure 38B).

Western blot studies performed by Dr. Barry also revealed the absence of Kv3.1b protein in the Kv3.1 KO mice, while expression of KIF5 and other KIF5-binding proteins was not altered (Figure 38A). I performed immunohistochemical studies in wildtype and Kv3.1

134

KO mice, and studied KIF5 staining patterns in the cerebellar molecular layer. Kv3.1b is normally expressed throughout the cerebellar molecular and granule cell layers (Weiser et al., 1994, Grigg et al., 2000, Rudy and McBain, 2001), but is absent from Purkinje neurons (Puente et al., 2010). This staining pattern suggests that Kv3.1 is present in parallel fiber axons of the molecular layer, having their origin in granule cell layer neurons. Immunostaining revealed the loss of Kv3.1b in the cerebellum of Kv3.1 KO mice (Figure 38C,D). The KIF5B staining pattern showed numerous puncta throughout the molecular layer, with an apparent reduction in Kv3.1 KO mice compared to the wildtype (Figure 38C,D). Because thousands of these puncta were present within a single image field collected at 40X, manually counting the puncta was likely neither feasible nor reliable, and I set out to find a method to quantify the puncta automatically.

To minimize the fluctuations in image fluorescence intensity based on uneven illumination by the microscope’s mercury light source, I used the Metamorph software to

“flatten” the background and allow the individual puncta to stand out. It was then necessary to threshold and binarize the image so that only the isolated puncta would remain in the image field. I realized that the thresholding was very important to this process, and would determine how many puncta would show up, and the size of the puncta in the binarized image. Therefore, I needed an objective criterion for thresholding the image, and settled on excluding all data except for the brightest 5% of the data. This yielded a binarized image with clearly isolated puncta. I then used the “Analyze

Particles” function in the NIH ImageJ software to quantify the numbers of puncta. I set the size limit for the quantified puncta between 9-300 pixels, thinking that anything larger

135 or smaller than this might be artifactual. To confirm the accuracy of this paradigm, I overlaid several processed images containing only puncta between 9-300 pixels in size with their original images. This provided me with confirmation that the puncta isolated by software manipulation were in fact those that could be distinguished visually in the original image as standing out against the background.

This method revealed a significant reduction in the numbers of puncta (KIF5B clusters) in the cerebellar molecular layer of the Kv3.1 KO mice compared to the wildtype (Figure 38E,F) (WT: 3017±137; Kv 3.1−/−: 2545±121; P = 0.016). These data confirmed that Kv3.1 does have a significant interaction with KIF5 in normal physiology.

Further studies performed by Dr. Barry demonstrated a similar effect in primary cultures of cerebellar granular neurons established from either Kv3.1 KO or wildtype mice. The

Kv3.1 KO neurons were shown to contain significantly fewer KIF5B clusters along their axons by the computer-based quantification method described above.

136

Figure 38. Decreased KIF5 clustering in cerebellar neurons of Kv 3.1 KO mice. (A) Western blot studies performed by Dr. Joshua Barry show expression of Kv3.1b, KIF5 and KIF5-binding proteins in brains of Kv3.1 KO (−/−) and wild-type (+/+) mice. Soluble (left) and crude membrane (right) fractions are compared. Molecular weights are indicated in kDa. (B) PCR genotyping results of Kv3.1 KO heterozygotes (+/−), homozygotes (−/−), and wild-type (+/+) mice. The DNA ladder is indicated in bp. (C,D) KIF5B and Kv3.1b expression in the cerebellum, shown here in coronal sections from wild-type (C) and Kv3.1 KO (D) mice. P, Purkinje cell layer; GL, granule cell layer; ML, molecular layer. (E) KIF5B clusters were quantified in wild-type (left) and Kv3.1−/− (right) sections, after images were thresholded and binarized to clarify pixel clusters. (F) The number of KIF5 clusters per image (40× objective lens) was reduced in Kv3.1 KO mice. The number of images quantified is indicated within the bars;*P<0.05 (t-test). Scale bars: 300 µm (C,D); 100 µm (E). All experiments for panels B-F were performed by Peter Jukkola.

137

DISCUSSION

Kv3.1 KO mice were observed to have reduced clustering of KIF5 motor proteins in the cerebellar molecular layer (Figure 38F), and these in vivo findings correlated well with reductions of KIF5 clustering observed in Kv3.1-deficient primary cerebellar granule cell neuron cultures. Together, these data are a provide compelling physiological confirmation of the detailed biochemical studies which revealed the mechanism by which

Kv3.1 clusters and activates KIF5 motor proteins to promote their own trafficking to the axon.

KIF5 clustering was reduced in Kv3.1 KO, but was not eliminated (Figure 38E,F).

This reveals that Kv3.1 is not the only mechanism for clustering of KIF5. Other members of the same Kv channel family, Kv3.3 and Kv3.4, are also expressed in the cerebellum.

These channels also feature a T1 domain with high sequence homology to Kv3.1, and likely can cluster KIF5 in a similar manner. Other cargo molecules for KIF5 may also be able to induce clustering.

The targeting and trafficking of ion channels to their proper location within the neuron are very important to neuronal function. These studies have revealed novel roles for Kv3.1 channels in regulating their own transport by regulation of KIF5 motor proteins. In addition to the important function of this relationship in normal physiology, trafficking of ion channels may also be important in the course of disease, as implied by the redistribution of Kv1.2 channels shown in chapter 2. Kv3.1 and Kv3.3 are both expressed in the spinal cord, and potentially could be altered in their localization under

138 the neuroinflammatory and demyelinating conditions present in EAE or MS. Kv3.1 is an especially interesting target for future study because it is the channel most sensitive to 4-

AP, the nonspecific channel blocker currently approved for symptomatic treatment of

MS. However, many other channels may be involved in the pathology of diseases such as MS, and it will be very important to understand the ion channel targeting and trafficking mechanisms of each under normal and pathophysiological conditions.

139

ANKYRIN-G SERVES AS AN ADAPTOR FOR KINESIN-MEDIATED

TRANSPORT OF VOLTAGE-GATED SODIUM CHANNELS

INTRODUCTION

Voltage-gated sodium channels are essential to neuronal function and the firing of an action potential, and are targeted to very specific locations in the proximal axon (the axon initial segment or AIS) and in the distal axon (nodes of Ranvier). This requires an active targeting and transport mechanism to efficiently move these channels to their proper location within the cell.

My colleague Dr. Joshua Barry elucidated the direct binding relationships of Ank-

G to KIF5 and to Nav channels. Multiple binding sites were shown on the Ank-G for both KIF5 and Nav channels, suggesting that Ank-G can bind both proteins at the same time and serve as an adaptor protein for transport of Nav channels during axonal transport

(Barry et al., 2014). Dr. Barry then validated these relationships in vitro with small interfering RNA (siRNA) knockdown of Ank-G or with the use of Ank-G dominant- negative constructs, which demonstrated that disrupting Ank-G function markedly reduced Nav channel levels at the axon initial segment (AIS). Indeed, Nav channels were reduced along the whole axon, showing that the major transport mechanism for Nav channels was disrupted, not just the docking mechanism to the axonal membrane (Barry et al., 2014). Dr. Yuanzheng Gu used these same siRNA or dominant-negative Ank-G

140 constructs in electrophysiological studies to confirm that the neurons were functionally altered. Action potential firing was decreased, and inward and outward currents were markedly reduced (Barry et al., 2014).

Dr. Yuanzheng Gu also performed important live cell imaging studies using fluorescently tagged KIF5, Ank-G and Nav channel proteins, or protein fragments containing the binding sites of these proteins. These studies demonstrated the functional relationship of these proteins and their close association in transport throughout the axon.

My role in these studies was again in helping to validate these findings in vivo, working with Dr. Joshua Barry and our collaborators in Dr. Howard Gu’s lab at The Ohio

State University, Dr. Keerthi Thirtamara Rajamani and Dr. Brian O’Neill. In short, we found significant evidence that Ank-G mediated transport of Nav channels does occur in physiological situations in animal models. Here I will describe the methods we used in these studies and the results that were obtained.

METHODS

AnkG KO Mice, Cardiac Perfusion, Tissue Fixation, and Sectioning

AnkG KO mice have been previously described (Jenkins and Bennett, 2001), and were obtained from our collaborator Dr. Peter Mohler at The Ohio State University.

Briefly, the AnkG isoform that is knocked out in these mice is predominantly expressed

141 in the cerebellum, so that the major deficits in these mice are related to the cerebellum, although some effects were also seen in spinal cord white matter axons in our studies.

I performed nearly all of the procedures of cardiac perfusion, tissue fixation, and sectioning as previously described (Jukkola et al., 2012, Jukkola et al., 2013). Briefly, mice were deeply anesthetized with avertin and perfused through the heart with 20–

30 mL ice-cold PBS followed by 20 mL 4% formaldehyde in PBS (FA/PBS). The brain and spinal cord of wildtype and AnkG−/− mice, and the brains of AAV-injected mice, were post-fixed in 4% FA/PBS for just 1 h because our AnkG antibody is sensitive to the degree of fixation of tissue. Tissues were cryoprotected in 30% sucrose for at least 24 hr, cut into 3-mm blocks using an acrylic brain matrix (Braintree Scientific, Braintree, MA,

USA), embedded in optimal cutting temperature (OCT) media (Sakura Finetek USA,

Inc., Torrance, CA, USA), and stored at −80°C until sectioning. The tissue blocks were cut with a Microm HM550 cryostat (Thermo Scientific, Waltham, MA, USA) and the 40-

μm sections were collected on Superfrost Plus microscope slides (FisherScientific,

Pittsburgh, PA, USA) for storage at −20°C.

Immunofluorescent Staining of Sections of Brain and Spinal Cord

Sections were incubated in PBS/0.3% Triton X-100 for 1 hr at room temperature

(RT) to permeabilize the tissue, and then blocked with 2.5% normal goat or donkey serum (matched with the host species of the secondary antibody) for 1 hr at RT. The primary antibodies were then added in blocking solution and the sections were incubated for 3 hr at RT, and then overnight at 4°C. The next day, the sections were rinsed 10 ×

142

5 min at RT, the appropriate secondary antibody was added in blocking solution, and the sections were incubated for 3 hr at RT. In this case, class-specific goat anti-mouse IgG -

2a or -2b secondary antibodies were used so that two mouse monoclonal primary antibodies could be distinguished. Sections were rinsed 10 × 5 min at RT and coverslipped using tris-buffered Fluoro-Gel mounting media (Electron Microscopy

Sciences). Both Dr. Joshua Barry and I performed these immunstaining experiments.

Confocal Microscopy

High-magnification confocal images were captured with a Leica TCS SL confocal imaging system (Leica Microsystems) using a 100× HCX Plan Apo CS oil immersion objective (numerical aperture = 1.40). Multiple channels were acquired simultaneously and the signal was averaged over six scans. Channel crosstalk was eliminated through optimization of the laser line intensity by acousto-optical tunable excitation filters, and by spectral detectors allowing precisely defined bandwidth adjustment. Images were saved as 8-bit TIFF files and adjusted for brightness and contrast using Adobe Photoshop 7.0.

These studies were performed by both Dr. Joshua Barry and I.

AAV2 Construction and Production

Dr. Joshua Barry prepared YFP, YFP-T70, and YFP-T70RKR constructs and inserted them into an adeno-associated virus (AAV) vector (kindly provided by the Viral

Core facility of The Research Institute at Nationwide Children’s Hospital) by subcloning.

The Viral Core facility of The Research Institute at Nationwide Children’s Hospital then

143 used these vectors to produce AAV2 viruses expressing YFP (viral titer: 2.1 × 1013),

13 13 YFP-T70 (viral titer: 1.9 × 10 ), or YFP-T70RKR (viral titer: 1.0 × 10 ).

Surgeries and Microinjection of Viral Vectors

YFP, YFP-T70, or YFP-T70RKR was overexpressed in mouse cerebellum by stereotaxic injection of recombinant AAV2 virus performed by Dr. Keerthi Thirtamara

Rajamani and Dr. Brian O’Neill. Each mouse was unilaterally injected on the right side, leaving the uninjected left side as a within-subject control. Cohorts of mice injected with

AAV-YFP (n = 6), AAV-YFP-T70 (n = 6), or AAV-YFP-T70RKR (n = 6) were produced for between-subject comparisons. After injection, the mice were sutured and administered postoperative care for 1 week. Mouse cerebellum was fixed and sectioned 2–3 weeks after injection for immunohistochemistry studies.

Statistical Analysis

Dr. Joshua Barry performed the statistical analysis with Sigmaplot 12.0, using one-way ANOVA when comparing two or three experimental groups with one control group, and unpaired t tests when comparing two groups. Results are provided as mean ±

SEM.

144

RESULTS

In light of the clear AnkG-mediated transport mechanism for Nav channels shown in vitro, we sought to determine whether Nav channels would also be altered in physiological contexts in the axons of mice deficient for Ank-G. Dr. Joshua Barry and I studied AnkG KO mice by immunohistochemistry in the cerebellum and spinal cord. We confirmed that AnkG was indeed missing from the AIS and nodes of Ranvier in cerebellar white matter axons (Figure 39A,C), and showed that this was associated with loss of Nav channels in these locations (Figure 39E). Dr. Barry quantified the percentage of nodes expressing AnkG and Nav channels, and found that cerebellar AnkG staining was eliminated and Nav channel staining was drastically reduced (Figure 39B).

A smaller, yet significant, reduction of both proteins was also observed in the spinal cord

(Figure 39B). Dr. Barry quantified the ratio of AnkG or Nav channels to Kv1.2 channels located at the juxtaparanodes, and found a significant reduction in the AnkG KO mice compared to wildtype (Figure 39D,F). Thus, Kv1.2 localization was not affected by

AnkG deficiency and likely has a mechanism of transport involving different motor and adaptor proteins (Gu et al., 2006).

To further corroborate these findings, we wondered if overexpression of a fragment of the KIF5 tail containing the AnkG binding site could disrupt Nav channel transport in vivo. Dr. Barry inserted DNA constructs for YFP, YFP-T70 (a fusion protein containing the AnkG-binding domain of KIF5), and YFP-T70RKR (containing the AnkG- binding domain of KIF5 with key residues mutated) into a viral vector and adeno-

145 associated virus type 2 (AAV2) viruses containing the constructs were produced by the

Viral Core facilities at The Research Institute at Nationwide Children’s Hospital.

Dr.Keerthi Thirtamara Rajamani and Dr. Brian O’Neill injected the viruses into the cerebellar deep nuclei of wildtype mice. At three weeks postinjection, I perfused the mice and sectioned their brains for immunohistochemistry and confocal microscopy studies. Dr. Barry and I located YFP-labeled sections infected with virus, immunostained them for AnkG and Kv1.2, and imaged them by confocal microscopy. Dr. Barry had the larger role in these imaging studies, and the data presented in Figure 39 were generated by him.

In the mice infected with YFP alone, clear expression of juxtaparanodal Kv1.2 and nodal Nav channels or AnkG was present in YFP –labeled axons (Figure 39H,I). When the T70 fragment of KIF5 was expressed, nodal Nav channels and AnkG were reduced in the nodes, as shown in Figure 39 (J,K) and quantified as the ratio of Nav or AnkG to

Kv1.2 (FAnkG/FKv1.2: GFP, 0.24 ± 0.03; YFP-T70, 0.11 ± 0.03; YFP-T70RKR, 0.25 ± 0.04;

FNav/FKv1.2: GFP, 0.24 ± 0.03; YFP-T70, 0.06 ± 0.02; YFP-T70RKR, 0.24 ± 0.06)(Figure

39L). This implies that the binding of AnkG to endogenous KIF5 was disrupted by the

T70 fragment so that AnkG-mediated transport of Nav channels could not occur. In the mice infected with the mutated KIF5 fragment T70RKR, the ratio of AnkG or Nav channels to Kv1.2 was at levels comparable to the mice infected with YFP alone. This shows that the mutated binding site could not disrupt binding of AnkG, and AnkG- mediated Nav channel transport was not impeded.

146

Figure 39. AnkG Deletion or Disruption of AnkG-KIF5 Binding by Expressing a KIF5B Tail Fragment Reduces the Axonal Level of Nav Channels In Vivo. Mouse cerebellar sections from wildtype, AnkG-/-, or AAV-injected wildtype mice were costained for Kv1.2 and AnkG (or Nav channels). (A) Confocal images of the Purkinje neuron AIS (revealed by AnkG staining in red) and cerebellar basket cell terminal (shown by Kv1.2 in green) in WT (top) and AnkG−/− (bottom) mice. White dots outline the Purkinje neuron soma. (B) Summary of the percentage of nodes of Ranvier with AnkG and Nav channel staining in WT and AnkG−/− mice in the cerebellum (Cereb) and spinal cord (SC). The nodal regions were defined by a pair of JXP Kv1.2 clusters. (C) High-magnification confocal images of AnkG (red) and Kv1.2 (green) staining in cerebellar WM myelinated axons of WT (top) and AnkG−/− (bottom) mice. (D) Summary of the ratio of AnkG and Kv1.2 staining fluorescence intensities along the axonal segment. (E) High-magnification confocal images of pan- Nav (red) and Kv1.2 (green) staining in cerebellar WM myelinated axons of WT (top) and AnkG−/− (bottom) mice. (F) Summary of the ratio of pan- Nav and Kv1.2 staining fluorescence intensities. (G) Diagram showing location of AAV virus injection into the mouse cerebellum. (H) Neurons in cerebellar nuclei expressing AAV-GFP. (I) Axons expressing AAV-GFP in cerebellar WM. (J) Axons expressing either GFP (top, blue) or YFP-T70 (bottom, blue) were costained for endogenous AnkG (red) and Kv 1.2 (green). (K) Axons expressing either GFP (top, blue) or YFP-T70 (bottom, blue) were costained for endogenous pan- Nav (red) and Kv1.2 (green) channels. (L) Summary of the effect of AAV-mediated expression of dominant-negative construct of KIF5B tail domain. White arrowheads: JXP region; white arrows: nodes of Ranvier. In (B), (D), (F), and (L), results are provided as mean ± SEM. Scale bars, 200 μm (H and I), 50 μm (A), and 10 μm (C, E, J, and K). An unpaired t test was used in (B), (D), and (F); one-way ANOVA followed by Dunnett’s test was used in (L). ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001. (Images in panel A were contributed by Peter Jukkola. All other image data for this figure was generated by Dr. Joshua Barry from tissue perfused and sectioned by Peter Jukkola. Please refer to text for other specific contributions to this collaborative project).

147

Figure 39. AnkG Deletion or Disruption of AnkG-KIF5 Binding by Expressing a KIF5B Tail Fragment Reduces the Axonal Level of Nav Channels In Vivo

148

DISCUSSION

These in vivo studies confirm the in vitro mechanistic studies and show that

AnkG is a critical adaptor protein for KIF5B-mediated transport of Nav channels to their precise location within the axon. Given the essential role of Nav channels in action potential generation at the AIS and in saltatory conduction at the nodes of Ranvier, this transport system clearly has a vital role in normal physiology. Proper Nav channel localization is essential for neurotransmission.

The described transport mechanism may also have a significant relevance to diseases such as MS. Nav channels have been shown to be redistributed along the axon following demyelination (Craner et al., 2004a, Craner et al., 2004b, Moll et al., 1991), similar to the redistribution of Kv1 channels as shown in chapter 2. It is unknown whether an active transport mechanism is involved in transporting the channels to the demyelinated axonal membranes, or if the redistribution occurs by simple diffusion of the channels in the absence of the anchoring molecules regulated by myelination. Certainly, in the case of remyelination of these axons, I expect that there would be a role for kinesins in repopulating nodes of Ranvier with Nav channels as they are re-established.

149

CHAPTER 5: CONCLUSIONS

OVERVIEW

These studies provided clear evidence that ion and water channels are indeed altered by inflammatory, demyelinating disease processes. Surprisingly, this occurred to an even greater extent than we initially expected, with significant change occurring in gray matter in addition to the white matter changes that were our primary focus at the beginning of our studies. Based on related literature, these channel alterations may have neuroprotective effects and could potentially be targeted in the development of disease- modifying therapies for MS and other diseases.

CONTRIBUTION TO ION AND WATER CHANNEL RESEARCH

This research opened up new avenues for studies of ion channel regulation in disease states. Previously, the field has been focused on the activity of a single potassium channel, Kv1.2, and its potential role in blocking conduction along an axon after redistribution to a demyelinated area. My research highlights the changes that occur in other potassium channels that may have varied effects upon the disease process. How

150 many other channels could potentially be involved? With forty different Kv channels and numerous other ion channels that have important roles in the CNS, these studies likely have only scratched the surface of the subject. Therefore, it is essential that more research is done in this area to promote a better understanding of Kv channel roles in disease processes. This would certainly promote better informed treatment, but could also lead to new treatment options since many of these Kv channels can be targeted by specific blockers or modulators. It is my hope that the foundation laid by these studies will help to generate more interest in the study of ion channels in MS and other diseases.

Although Kv1.2 should no longer be regarded as the only channel that is important in demyelinating disease, it is still a highly interesting channel that should be studied further. Perhaps most intriguing is the localization of Kv1.2 at the juxtaparanode. A recent study showed that axonal degeneration begins with focal axonal swellings, and that these preferentially occur near nodes of Ranvier (Nikic et al., 2011) . These swellings were related to mitochondrial dysfunction, and were also shown to be potentially reversible (Nikic et al., 2011). My Kv1.2 studies likewise showed that a significant recovery of Kv1.2 channel expression patterns can occur in rrEAE (Figure 10), suggesting that remyelination and restoration of neuronal function is possible while the axons are still intact. It would be very interesting to discover whether Kv1.2 function or dysfunction has any role in producing the focal axonal swellings, and whether there is any connection between Kv1.2 and mitochondrial function at the nodes. Regardless, these studies highlight the need for further development of new strategies for neuroprotection and remyelination.

151

The results of my Kv2.1 studies showed that Kv channels in gray matter areas could also be affected in EAE, with no lesion nearby. Thus, Kv channel expression patterns can be regulated over long distances, which may be due to either a circulating factor or to signaling within a neuron whose axon may pass through a lesioned area distal to the location of the neuronal soma. It would be very interesting to perform proteomics studies to find the potential cytokines or other factors that could cause declustering of

Kv2.1 in EAE or other disease conditions. Recently, the chemokine receptor CXCR4 was shown to be able to induce declustering, leading to a short-term neuroprotective effect in hippocampal neurons (Shepherd et al., 2012). More study is needed to show whether this pathway is involved in producing the declustering effect in spinal cord alpha motor neurons in EAE. Considering the potential for neuroprotection, this is a very interesting research direction that could open up new treatment options if Kv2.1 can be modulated in a beneficial way.

My Kv1.4 studies revealed a potential role of this Kv channel in reactive astrocytes in EAE. Since Kv1.4 expression in astrocytes is not yet well-characterized, the initial steps in the follow-up for these studies should be to determine whether the Kv1.4 channels are functionally active in these reactive astrocytes, and how this affects the activity of the cell. Does it promote proliferation of the cell, as has been suggested in a spinal cord injury model (Edwards et al., 2002)? Does Kv1.4 interact with Nav currents from channels expressed in reactive astrocytes (Black et al., 2010)? In short, more studies are needed to determine the effect of the observed changes in the expression of

Kv1.4 during the course of disease, and the mechanisms by which the changes occur.

152

Finally, my studies have highlighted the alteration of AQP4 expression and localization in EAE lesioned areas. Given the vital roles of AQP4 in maintaining the ionic and fluid environment of the brain, the reorganization of AQP4+ processes at the border of lesions within the glial scar may indicate important roles in the restoration of ionic homeostasis in the lesioned area and the prevention or resolution of lesion- associated edema.

CONTRIBUTION TO NEUROPROTECTION FIELD

My research points to potential neuroprotective mechanisms mediated by Kv channels. Current thought in the field of neurodegeneration in MS centers around the

+ 2+ roles of Na and Ca -mediated toxicity, leading to neuronal loss. It may be that Kv channel activity is an important counterbalance to these degenerative mechanisms.

Kv1.2 redistribution to demyelinated axons is likely necessary for the neuron, because Nav channels are also known to be redistributed to these areas (Waxman et al.,

2004, Waxman, 2006, Craner et al., 2004b). Although Kv1.2 channel redistribution to demyelinated areas is thought to lead to blocking of axonal conduction, this is not necessarily always detrimental. Reduced activity of the neuron would likely allow the cell to maintain the proper ionic concentrations of sodium and potassium on either side of the membrane by the Na+,K+, ATPase pump, without depletion of the ATP required to drive this process. This, in turn, would prevent excess accumulation of Na+ within the

153 axon, which is thought to cause reversal of the Na+, Ca2+ exchanger, influx of Ca2+ into the axon, and activation of Ca2+-dependent proteases detrimental to axonal stability . A neuroprotective strategy for Kv1.2 modulation should take into account both the need for neuronal function and the need to avoid an unsustainable level of activity that could lead to neuronal loss.

My results showed declustering of Kv2.1 in motor neurons that was reversible in rrEAE, but sustained in chEAE (Figures 13, 14, 15). This seems to agree with recent work showing a short-term neuroprotective or long-term degenerative effect of Kv2.1 declustering (Shepherd et al., 2012). The short-term neuroprotective effect of Kv2.1 activity logically could occur by hyperpolarizing the cell and reducing neuronal activity, allowing the cell to keep up with the metabolic demand. The longer term neurodegenerative effect is likely due to induction of neuronal apoptosis. Kv2.1 has been shown to mediate a large efflux of K+ from cells undergoing apoptosis (Yu et al., 1997,

Pal et al., 2006). The reduction of K+ concentration in the cell releases inhibition of nucleases and caspases involved in dismantling the cell (Leung, 2010). If factors modulating Kv2.1 in EAE can be discovered, as previously discussed, efforts to develop treatments based on these factors will have to take into account both the beneficial short- term and detrimental long-term effects of Kv2.1 activity.

Since Nav channels are known to be upregulated in reactive astrocytes in MS and other disease conditions (Black et al., 2010), Kv1.4 may be upregulated in response to maintain the resting membrane potential. My results showed that Kv1.4 is upregulated during peak EAE in lesioned areas in both chEAE and rrEAE, and persists in remitting

154 rrEAE (Figures 16, 18). In light of this, we have postulated a neuroprotective effect for

Kv1.4 upregulation, since the remitting phase of rrEAE showed a high level of expression compared to the peak of disease, while the animals’ clinical scores returned to normal.

This Kv1.4 upregulation potentially could provide a measure of control for Nav- and Cav- mediated currents, the latter of which could have wide-ranging effects on neuronal activity through calcium wave and gliotransmission activities. Thus, we can hypothesize that Kv1.4 promotes a neuroprotective effect of reactive astrocytes under neuroinflammatory conditions by reducing the likelihood of Ca2+-mediated excitotoxicity. More study is needed to determine whether this is true.

CONTRIBUTION TO MS RESEARCH

Before my studies, the role of potassium ion channels in MS neuropathology was thought to be related primarily to Kv1.2 redistribution in demyelinated axons. My research opens up new avenues for research, suggesting that more channels may be involved in the pathology of MS, and that the major channels involved in the therapeutic effect of 4-AP may not be restricted to Kv1. Indeed, the nonspecific Kv-channel blocking activity of 4-AP could have significant effects on many Kv channels with varied roles, which may be either beneficial or detrimental to the restoration of neuronal health.

Further study is needed into the involvement of various ion channels in MS and other

155 diseases so that their roles can be elucidated, and treatment strategies can be developed to minimize the bad effects while maximizing the good effects.

My work also furthers research progress in the area of gray matter involvement in the pathology of MS/EAE. Although MS has most often been considered a white matter disease, gray matter pathology has increasingly been shown to be important in the course of disease. Gray matter atrophy as shown by MRI correlates well with disease progression (Geurts et al., 2009). In EAE models, gray matter areas have been shown to have increased microglial and astrocyte activation, loss of neurons, and altered synaptic function (Kurkowska-Jastrzebska et al., 2013, Ziehn et al., 2012a, Ziehn et al., 2012b,

Ziehn et al., 2010). In chapter 2, I have shown alterations in Kv channels localized in gray matter that were well correlated with disease severity as shown by clinical scoring

(Jukkola et al., 2012). In chapter 3, I have also shown the development of reactive astrocytosis in several different gray matter areas of the brain in response to EAE

(Jukkola et al., 2013). As the neuroinflammatory effects of MS/EAE in gray matter areas of the CNS become better understood, the research community will gain new insights into threats to neuronal health and cognition, and will be better equipped to develop new therapies.

The role of kinesin motors in MS is a relatively unexplored area of research.

Since ion channels are altered in their localization in EAE, as shown previously, it stands to reason that motor proteins are involved in either this redistribution or in the restoration of normal targeting that occurs in remyelination. MS has been associated with mutations in at least three different kinesins, including KIF5A (Alcina et al., 2010, Goris et al.,

156

2010, Hares et al., 2013, Koutsis et al., 2011). Defects in axonal transport have also been recognized in the course of EAE (Guy et al., 1989, O'Neill et al., 1998). Because kinesin-mediated transport is an ATP-dependent process, it may be that the ATP-depleted conditions present in demyelinated axons lead to a lack of the energy required for axonal transport. Accumulation of cellular components in areas of disturbed axonal transport may promote the development of axonal swellings (Kreutzer et al., 2011, Nikic et al.,

2011), which often lead to neurodegeneration. To my knowledge, kinesins themselves have not yet been studied in EAE. However, I believe this could be a highly interesting direction for future research. It could be quite informative to determine which kinesins mediate altered axonal transport in EAE, and to explore the potential cargoes for each affected kinesin. This could yield new insights into the mechanisms of axonal dysfunction and degeneration in EAE/MS, and might even reveal new targets for development of new drug therapies.

Considering the ion channel changes that we have discussed, is there potential for an ion channel-mediated cure for MS? It seems unlikely, yet this thought has intrigued me throughout my research. A complete cure for the disease in question indeed ought to be the ultimate goal of biomedical research, and the possibility deserves careful consideration. However, until the disease-initiating mechanisms of MS are clearly defined, this remains an open question. Changes in ion channel activity do not imply a causative role in disease, and indeed, the nature of the observed changes seems to indicate that they are more likely a consequence of disease pathology. Although autoantibodies to the inwardly rectifying potassium channel Kir4.1 have recently been

157 associated with MS (Srivastava et al., 2012), it is unclear what role this plays in the course of disease. Although a full cure presently seems out of reach, ion channel- mediated neuroprotective treatment for MS appears to be a promising research direction, because disease-related disability in MS is highly correlated to the degree of neuronal loss. With preservation of neuronal health and function, there is also potential for reparative processes to progress, which may lead to substantial recovery. Thus, the best strategy for treatment in MS would likely be a combination of the currently available anti-inflammatory medications with novel neuroprotective treatments as they are discovered. I look forward to the day when these neuroprotective treatments become available, and hope that my research may aid in their development.

CONCLUSION

Ion and water channels play essential roles in neurons and other cells of the CNS.

Animal modeling and systematic neuropathological studies are vital to the understanding of these roles in health and in neurobiological disorders such as Multiple Sclerosis. My work has opened up new avenues of research, which may lead to much more detailed functional and mechanistic analyses. As ion and water channel roles in disease are better understood, this will promote the development of better targeted treatment options with fewer side effects.

158

REFERENCES

Aharoni R, Vainshtein A, Stock A, et al. (2011) Distinct pathological patterns in

relapsing-remitting and chronic models of experimental autoimmune

encephalomyelitis and the neuroprotective effect of glatiramer acetate. J

Autoimmun. Nov;37(3):228-41.

Alcina A, Vandenbroeck K, Otaegui D, et al. (2010) The autoimmune disease-associated

KIF5A, CD226 and SH2B3 gene variants confer susceptibility for multiple

sclerosis. Genes Immun, 11, 439-45.

Alvestad S, Hammer J, Hoddevik EH, et al. (2013) Mislocalization of AQP4 precedes

chronic seizures in the kainate model of temporal lobe epilepsy. Epilepsy Res,

105, 30-41.

Amici SA, Dunn WA, Jr., Notterpek L (2007) Developmental abnormalities in the nerves

of peripheral myelin protein 22-deficient mice. J Neurosci Res, 85, 238-49.

Amiry-Moghaddam M, Lindland H, Zelenin S, et al. (2005) Brain mitochondria contain

aquaporin water channels: evidence for the expression of a short AQP9 isoform in

the inner mitochondrial membrane. Faseb J, 19, 1459-67.

159

Aoyama M, Kakita H, Kato S, Tomita M, Asai K (2012) Region-specific expression of a

water channel protein, aquaporin 4, on brain astrocytes. J Neurosci Res, 90, 2272-

80.

Arizono M, Bannai H, Nakamura K, et al. (2012) Receptor-selective diffusion barrier

enhances sensitivity of astrocytic processes to metabotropic glutamate receptor

stimulation. Sci Signal, 5, ra27.

Armstrong WE, Rubrum A, Teruyama R, Bond CT, Adelman JP (2005)

Immunocytochemical localization of small-conductance, calcium-dependent

potassium channels in astrocytes of the rat supraoptic nucleus. Journal of

Comparative Neurology, 491, 175-85.

Attali B, Wang N, Kolot A, Sobko A, Cherepanov V, Soliven B (1997) Characterization

of delayed rectifier Kv channels in oligodendrocytes and progenitor cells. J

Neurosci, 17, 8234-45.

Bardehle S, Kruger M, Buggenthin F, et al. (2013) Live imaging of astrocyte responses to

acute injury reveals selective juxtavascular proliferation. Nat Neurosci, 16, 580-6.

Barel O, Shalev SA, Ofir R, et al. (2008) Maternally inherited Birk Barel mental

retardation dysmorphism syndrome caused by a mutation in the genomically

imprinted potassium channel KCNK9. Am J Hum Genet, 83, 193-9.

Barnett MH, Sutton I (2012) Neuromyelitis optica: not a multiple sclerosis variant.

Current Opinion in Neurology, 25, 215-20.

Barres BA (2008) The mystery and magic of glia: a perspective on their roles in health

and disease. Neuron, 60, 430-40.

160

Barry J, Gu Y, Gu C (2010) Polarized targeting of L1-CAM regulates axonal and

dendritic bundling in vitro. Eur J Neurosci, 32, 1618-31.

Barry J, Gu Y, Jukkola P, et al. (2014) Ankyrin-g directly binds to Kinesin-1 to transport

voltage-gated na(+) channels into axons. Dev Cell, 28, 117-31.

Barry J, Xu, M., Gu, Y., Dangel, A., Jukkola, P., Shrestha, C. & Gu, C. (2013) Activation

of conventional kinesin motors in clusters by Shaw voltage-gated K+ channels.

Journal of Cell Science, 126, 2027-2041.

Bay V, Butt AM (2012) Relationship between glial potassium regulation and axon

excitability: a role for glial Kir4.1 channels. Glia, 60, 651-60.

Beeton C, Chandy KG (2005) Potassium channels, memory T cells, and multiple

sclerosis. Neuroscientist, 11, 550-62.

Benesova J, Hock M, Butenko O, Prajerova I, Anderova M, Chvatal A (2009)

Quantification of astrocyte volume changes during ischemia in situ reveals two

populations of astrocytes in the cortex of GFAP/EGFP mice. J Neurosci Res, 87,

96-111.

Benfenati V, Caprini M, Dovizio M, et al. (2011) An aquaporin-4/transient receptor

potential vanilloid 4 (AQP4/TRPV4) complex is essential for cell-volume control

in astrocytes. Proc Natl Acad Sci U S A, 108, 2563-8.

Bennett JL, Lam C, Kalluri SR, et al. (2009) Intrathecal pathogenic anti-aquaporin-4

antibodies in early neuromyelitis optica. Ann Neurol, 66, 617-29.

Bennett V, Baines AJ (2001) Spectrin and ankyrin-based pathways: metazoan inventions

for integrating cells into tissues. Physiol Rev, 81, 1353-92.

161

Berkefeld H, Fakler B (2013) Ligand-gating by Ca2+ is rate limiting for physiological

operation of BK(Ca) channels. J Neurosci, 33, 7358-67.

Bignami A, Raju T, Dahl D (1982) Localization of vimentin, the nonspecific intermediate

filament protein, in embryonal glia and in early differentiating neurons. In vivo

and in vitro immunofluorescence study of the rat embryo with vimentin and

neurofilament antisera. Dev Biol, 91, 286-95.

Binder DK, Oshio K, Ma T, Verkman AS, Manley GT (2004) Increased seizure threshold

in mice lacking aquaporin-4 water channels. Neuroreport, 15, 259-62.

Binder DK, Steinhauser C (2006) Functional changes in astroglial cells in epilepsy. Glia,

54, 358-68.

Binder DK, Yao X, Zador Z, Sick TJ, Verkman AS, Manley GT (2006) Increased seizure

duration and slowed potassium kinetics in mice lacking aquaporin-4 water

channels. Glia, 53, 631-6.

Bittner S, Meuth SG, Gobel K, et al. (2009) TASK1 modulates inflammation and

neurodegeneration in autoimmune inflammation of the central nervous system.

Brain, 132, 2501-16.

Black JA, Dib-Hajj S, Cohen S, Hinson AW, Waxman SG (1998) Glial cells have heart:

rH1 Na+ channel mRNA and protein in spinal cord astrocytes. Glia, 23, 200-8.

Black JA, Newcombe J, Waxman SG (2010) Astrocytes within multiple sclerosis lesions

upregulate sodium channel Nav1.5. Brain, 133, 835-46.

162

Black JA, Westenbroek R, Ransom BR, Catterall WA, Waxman SG (1994a) Type II

sodium channels in spinal cord astrocytes in situ: immunocytochemical

observations. Glia, 12, 219-27.

Black JA, Yokoyama S, Higashida H, Ransom BR, Waxman SG (1994b) Sodium

channel mRNAs I, II and III in the CNS: cell-specific expression. Brain Res Mol

Brain Res, 22, 275-89.

Bouhy D, Ghasemlou N, Lively S, et al. (2011) Inhibition of the Ca2-dependent K

channel, KCNN4/KCa3.1, improves tissue protection and locomotor recovery

after spinal cord injury. Journal of Neuroscience, 31, 16298-308.

Brenner R, Chen QH, Vilaythong A, Toney GM, Noebels JL, Aldrich RW (2005) BK

channel beta4 subunit reduces dentate gyrus excitability and protects against

temporal lobe seizures. Nat Neurosci, 8, 1752-9.

Brinkmeier H, Aulkemeyer P, Wollinsky KH, Rudel R (2000) An endogenous

pentapeptide acting as a sodium channel blocker in inflammatory autoimmune

disorders of the central nervous system. Nat Med, 6, 808-11.

Brosnan CR, CS (2013) The Astrocyte in Multiple Sclerosis Revisited. Glia, 61, 453-465.

Bukhari W, Barnett MH, Prain K, Broadley SA (2012) Molecular pathogenesis of

neuromyelitis optica. Int J Mol Sci, 13, 12970-93.

Buraei Z, Yang J (2013) Structure and function of the beta subunit of voltage-gated

Ca(2)(+) channels. Biochim Biophys Acta, 1828, 1530-40.

163

Burgos M, Calvo S, Molina F, et al. (2007a) PKCepsilon induces astrocyte stellation by

modulating multiple cytoskeletal proteins and interacting with Rho A signalling

pathways: implications for neuroinflammation. Eur J Neurosci, 25, 1069-78.

Burgos M, Fradejas N, Calvo S, Kang SU, Tranque P, Lubec G (2011) A proteomic

analysis of PKCepsilon targets in astrocytes: implications for astrogliosis. Amino

Acids, 40, 641-51.

Burgos M, Pastor MD, Gonzalez JC, et al. (2007b) PKCepsilon upregulates voltage-

dependent calcium channels in cultured astrocytes. Glia, 55, 1437-48.

Butenko O, Dzamba D, Benesova J, et al. (2012) The increased activity of TRPV4

channel in the astrocytes of the adult rat hippocampus after cerebral

hypoxia/ischemia. PLoS One, 7, e39959.

Cacheaux LP, Ivens S, David Y, et al. (2009) Transcriptome profiling reveals TGF-beta

signaling involvement in epileptogenesis. J Neurosci, 29, 8927-35.

Cain SM, Snutch TP (2011) Voltage-gated calcium channels and disease. Biofactors, 37,

197-205.

Chan KH, Zhang R, Kwan JS, et al. (2012) Aquaporin-4 autoantibodies cause

asymptomatic aquaporin-4 loss and activate astrocytes in mouse. J

Neuroimmunol, 245, 32-8.

Chandy KG, Wulff H, Beeton C, Pennington M, Gutman GA, Cahalan MD (2004) K+

channels as targets for specific immunomodulation. Trends Pharmacol Sci, 25,

280-9.

164

Charles AC, Merrill JE, Dirksen ER, Sanderson MJ (1991) Intercellular signaling in glial

cells: calcium waves and oscillations in response to mechanical stimulation and

glutamate. Neuron, 6, 983-92.

Chittajallu R, Chen Y, Wang H, et al. (2002) Regulation of Kv1 subunit expression in

oligodendrocyte progenitor cells and their role in G1/S phase progression of the

cell cycle. Proc Natl Acad Sci U S A, 99, 2350-5.

Clark BD, Goldberg EM, Rudy B (2009) Electrogenic tuning of the axon initial segment.

Neuroscientist, 15, 651-68.

Coetzee WA, Amarillo Y, Chiu J, et al. (1999) Molecular diversity of K+ channels. Ann

N Y Acad Sci, 868, 233-85.

Cooper EC, Milroy A, Jan YN, Jan LY, Lowenstein DH (1998) Presynaptic localization

of Kv1.4-containing A-type potassium channels near excitatory synapses in the

hippocampus. Journal of Neuroscience, 18, 965-974.

Cornell-Bell AH, Finkbeiner SM, Cooper MS, Smith SJ (1990) Glutamate induces

calcium waves in cultured astrocytes: long-range glial signaling. Science, 247,

470-3.

Craner MJ, Hains BC, Lo AC, Black JA, Waxman SG (2004a) Co-localization of sodium

channel Nav1.6 and the sodium-calcium exchanger at sites of axonal injury in the

spinal cord in EAE. Brain, 127, 294-303.

Craner MJ, Newcombe J, Black JA, Hartle C, Cuzner ML, Waxman SG (2004b)

Molecular changes in neurons in multiple sclerosis: altered axonal expression of

165

Nav1.2 and Nav1.6 sodium channels and Na+/Ca2+ exchanger. Proc Natl Acad

Sci U S A, 101, 8168-73.

Custer SK, Garden GA, Gill N, et al. (2006) Bergmann glia expression of polyglutamine-

expanded ataxin-7 produces neurodegeneration by impairing glutamate transport.

Nat Neurosci, 9, 1302-11. de Lanerolle NC, Lee TS, Spencer DD (2010) Astrocytes and epilepsy.

Neurotherapeutics, 7, 424-38.

Debanne D (2004) Information processing in the axon. Nat Rev Neurosci, 5, 304-16.

Deng PY, Poudel SK, Rojanathammanee L, Porter JE, Lei S (2007) Serotonin inhibits

neuronal excitability by activating two-pore domain k+ channels in the entorhinal

cortex. Mol Pharmacol, 72, 208-18.

Devaux J, Alcaraz G, Grinspan J, et al. (2003) Kv3.1b is a novel component of CNS

nodes. J Neurosci, 23, 4509-18.

Di Castro MA, Chuquet J, Liaudet N, et al. (2011) Local Ca2+ detection and modulation

of synaptic release by astrocytes. Nat Neurosci, 14, 1276-84.

Djukic B, Casper KB, Philpot BD, Chin LS, McCarthy KD (2007) Conditional knock-out

of Kir4.1 leads to glial membrane depolarization, inhibition of potassium and

glutamate uptake, and enhanced short-term synaptic potentiation. J Neurosci, 27,

11354-65.

Doetsch F, Caille I, Lim DA, Garcia-Verdugo JM, Alvarez-Buylla A (1999)

Subventricular zone astrocytes are neural stem cells in the adult mammalian brain.

Cell, 97, 703-16.

166

Du W, Bautista JF, Yang H, et al. (2005) Calcium-sensitive potassium channelopathy in

human epilepsy and paroxysmal movement disorder. Nat Genet, 37, 733-8.

Dunn KM, Hill-Eubanks DC, Liedtke WB, Nelson MT (2013) TRPV4 channels stimulate

Ca2+-induced Ca2+ release in astrocytic endfeet and amplify neurovascular

coupling responses. Proc Natl Acad Sci U S A, 110, 6157-62.

Dunn KM, Nelson MT (2010) Potassium channels and neurovascular coupling.

Circulation Journal, 74, 608-16.

Edwards L, Nashmi R, Jones O, et al. (2002) Upregulation of Kv 1.4 protein and gene

expression after chronic spinal cord injury. J Comp Neurol, 443, 154-67.

Eid T, Ghosh A, Wang Y, et al. (2008) Recurrent seizures and brain pathology after

inhibition of glutamine synthetase in the hippocampus in rats. Brain, 131, 2061-

70.

Eid T, Lee TS, Thomas MJ, et al. (2005) Loss of perivascular aquaporin 4 may underlie

deficient water and K+ homeostasis in the human epileptogenic hippocampus.

Proc Natl Acad Sci U S A, 102, 1193-8.

Eid T, Thomas MJ, Spencer DD, et al. (2004) Loss of glutamine synthetase in the human

epileptogenic hippocampus: possible mechanism for raised extracellular

glutamate in mesial temporal lobe epilepsy. Lancet, 363, 28-37.

Eng LF, Vanderhaeghen JJ, Bignami A, Gerstl B (1971) An acidic protein isolated from

fibrous astrocytes. Brain Res, 28, 351-4.

Enyedi P, Czirjak G (2010) Molecular background of leak K+ currents: two-pore domain

potassium channels. Physiol Rev, 90, 559-605.

167

Eriksson JE, Dechat T, Grin B, et al. (2009) Introducing intermediate filaments: from

discovery to disease. J Clin Invest, 119, 1763-71.

Ermolinsky B, Arshadmansab MF, Pacheco Otalora LF, Zarei MM, Garrido-Sanabria ER

(2008) Deficit of Kcnma1 mRNA expression in the dentate gyrus of epileptic rats.

Neuroreport, 19, 1291-4.

Eroglu C, Barres BA (2010) Regulation of synaptic connectivity by glia. Nature, 468,

223-31.

Espejo C, Montalban X (2012) Dalfampridine in multiple sclerosis: From symptomatic

treatment to immunomodulation. Clin Immunol.

Fan Y, Liu M, Wu X, et al. Aquaporin-4 promotes memory consolidation in Morris water

maze. Brain Struct Funct, 218, 39-50.

Fan Y, Liu M, Wu X, et al. (2013) Aquaporin-4 promotes memory consolidation in

Morris water maze. Brain Struct Funct, 218, 39-50.

Farber K, Kettenmann H (2005) Physiology of microglial cells. Brain Res Brain Res Rev,

48, 133-43.

Filosa JA, Bonev AD, Straub SV, et al. (2006) Local potassium signaling couples

neuronal activity to vasodilation in the brain. Nat Neurosci, 9, 1397-1403.

Fordyce CB, Jagasia R, Zhu X, Schlichter LC (2005) Microglia Kv1.3 channels

contribute to their ability to kill neurons. J Neurosci, 25, 7139-49.

Garcia AD, Doan NB, Imura T, Bush TG, Sofroniew MV (2004) GFAP-expressing

progenitors are the principal source of constitutive neurogenesis in adult mouse

forebrain. Nat Neurosci, 7, 1233-41.

168

Geurts JJ, Stys PK, Minagar A, Amor S, Zivadinov R (2009) Gray matter pathology in

(chronic) MS: modern views on an early observation. J Neurol Sci, 282, 12-20.

Giaume C, Leybaert L, C CN, J CS (2013) Connexin and pannexin hemichannels in brain

glial cells: properties, pharmacology, and roles. Front Pharmacol, 4, 88.

Girouard H, Bonev AD, Hannah RM, Meredith A, Aldrich RW, Nelson MT (2010)

Astrocytic endfoot Ca2+ and BK channels determine both arteriolar dilation and

constriction. Proc Natl Acad Sci U S A, 107, 3811-6.

Goodman AD, Brown TR, Edwards KR, et al. (2010) A phase 3 trial of extended release

oral dalfampridine in multiple sclerosis. Ann Neurol, 68, 494-502.

Goodman AD, Brown TR, Krupp LB, et al. (2009) Sustained-release oral fampridine in

multiple sclerosis: a randomised, double-blind, controlled trial. Lancet, 373, 732-

Goris A, Boonen S, D'Hooghe M B, Dubois B (2010) Replication of KIF21B as a

susceptibility locus for multiple sclerosis. J Med Genet, 47, 775-6.

Grigg JJ, Brew HM, Tempel BL (2000) Differential expression of voltage-gated

potassium channel genes in auditory nuclei of the mouse brainstem. Hear Res,

140, 77-90.

Gu C, Barry J (2011) Function and mechanism of axonal targeting of voltage-sensitive

potassium channels. Prog Neurobiol, 94, 115-32.

Gu C, Gu Y (2011) Clustering and activity tuning of kv1 channels in myelinated

hippocampal axons. J Biol Chem, 286, 25835-47.

169

Gu C, Zhou W, Puthenveedu MA, Xu M, Jan YN, Jan LY (2006) The microtubule plus-

end tracking protein EB1 is required for Kv1 voltage-gated K+ channel axonal

targeting. Neuron, 52, 803-16.

Gu Y, Barry J, McDougel R, Terman D, Gu C (2012) Alternative splicing regulates

Kv3.1 polarized targeting to adjust the maximal spiking frequency. J Biol Chem.

Gu Y, Barry, J., & Gu, C. (2013) Kv3 channel assembly, trafficking and activity are

regulated by zinc through different binding sites. The Journal of Physiology, 591,

2475-2490.

Gu Y, Gu C (2010) Dynamics of Kv1 channel transport in axons. PLoS One, 5, e11931.

Guo F, Maeda Y, Ma J, et al. (2012) Macroglial plasticity and the origins of reactive

astroglia in experimental autoimmune encephalomyelitis. J Neurosci, 31, 11914-

28.

Guy J, Ellis EA, Tark EF, 3rd, Hope GM, Rao NA (1989) Axonal transport reductions in

acute experimental allergic encephalomyelitis: qualitative analysis of the optic

nerve. Curr Eye Res, 8, 261-9.

Halassa MM, Fellin T, Haydon PG (2007) The tripartite synapse: roles for

gliotransmission in health and disease. Trends Mol Med, 13, 54-63.

Hamann S, Zeuthen T, La Cour M, et al. (1998) Aquaporins in complex tissues:

distribution of aquaporins 1-5 in human and rat eye. Am J Physiol, 274, C1332-

45.

170

Hares K, Kemp K, Rice C, Gray E, Scolding N, Wilkins A (2013) Reduced axonal motor

protein expression in non-lesional grey matter in multiple sclerosis. Mult Scler.

October 21; doi:10.1177/1352458513508836

Hayes KC (2004) The use of 4-aminopyridine (fampridine) in demyelinating disorders.

CNS Drug Rev, 10, 295-316.

Herrero-Herranz E, Pardo LA, Bunt G, Gold R, Stuhmer W, Linker RA (2007) Re-

expression of a developmentally restricted potassium channel in autoimmune

demyelination: Kv1.4 is implicated in oligodendroglial proliferation. Am J Pathol,

171, 589-98.

Heuser K, Eid T, Lauritzen F, et al. (2012) Loss of perivascular Kir4.1 potassium

channels in the sclerotic hippocampus of patients with mesial temporal lobe

epilepsy. J Neuropathol Exp Neurol, 71, 814-25.

Hibino H, Inanobe A, Furutani K, Murakami S, Findlay I, Kurachi Y (2010) Inwardly

rectifying potassium channels: their structure, function, and physiological roles.

Physiol Rev, 90, 291-366.

Higashimori H, Sontheimer H (2007) Role of Kir4.1 channels in growth control of glia.

Glia, 55, 1668-79.

Hille B (2001) Ion channels of excitable membranes. (Sinauer, Sunderland,

Massachusetts, 2001).

Hinson SR, McKeon A, Lennon VA (2010) Neurological autoimmunity targeting

aquaporin-4. Neuroscience, 168, 1009-18.

171

Hinson SR, Romero MF, Popescu BF, et al. (2012) Molecular outcomes of neuromyelitis

optica (NMO)-IgG binding to aquaporin-4 in astrocytes. Proc Natl Acad Sci U S

A, 109, 1245-50.

Ho CS, Grange RW, Joho RH (1997) Pleiotropic effects of a disrupted K+ channel gene:

reduced body weight, impaired motor skill and muscle contraction, but no

seizures. Proc Natl Acad Sci U S A, 94, 1533-8.

Hodgkin AL, Huxley AF (1952) Currents carried by sodium and potassium ions through

the membrane of the giant axon of Loligo. J Physiol, 116, 449-72.

Hsu MS, Seldin M, Lee DJ, Seifert G, Steinhauser C, Binder DK (2011) Laminar-specific

and developmental expression of aquaporin-4 in the mouse hippocampus.

Neuroscience, 178, 21-32.

Hurlock EC, Bose M, Pierce G, Joho RH (2009) Rescue of motor coordination by

Purkinje cell-targeted restoration of Kv3.3 channels in Kcnc3-null mice requires

Kcnc1. Journal of Neuroscience, 29, 15735-44.

Hurlock EC, McMahon A, Joho RH (2008) Purkinje-cell-restricted restoration of Kv3.3

function restores complex spikes and rescues motor coordination in Kcnc3

mutants. Journal of Neuroscience, 28, 4640-8.

Iliff JJ, Wang M, Liao Y, et al. (2012) A paravascular pathway facilitates CSF flow

through the brain parenchyma and the clearance of interstitial solutes, including

amyloid beta. Sci Transl Med, 4, 147ra111.

172

Inyushin M, Kucheryavykh LY, Kucheryavykh YV, et al. (2010) Potassium channel

activity and glutamate uptake are impaired in astrocytes of seizure-susceptible

DBA/2 mice. Epilepsia, 51, 1707-13.

Iorio R, Fryer JP, Hinson SR, et al. (2013) Astrocytic autoantibody of neuromyelitis

optica (NMO-IgG) binds to aquaporin-4 extracellular loops, monomers, tetramers

and high order arrays. J Autoimmun, 40, 21-7.

Ishibashi T, Dakin KA, Stevens B, et al. (2006) Astrocytes promote myelination in

response to electrical impulses. Neuron, 49, 823-32.

Ivens S, Kaufer D, Flores LP, et al. (2007) TGF-beta receptor-mediated albumin uptake

into astrocytes is involved in neocortical epileptogenesis. Brain, 130, 535-47.

Jarius S, Wildemann B (2010) AQP4 antibodies in neuromyelitis optica: diagnostic and

pathogenetic relevance. Nature Reviews Neurology, 6, 383-92.

Jenkins SM, Bennett V (2001) Ankyrin-G coordinates assembly of the spectrin-based

membrane skeleton, voltage-gated sodium channels, and L1 CAMs at Purkinje

neuron initial segments. J Cell Biol, 155, 739-46.

Jonas P, Bischofberger J, Fricker D, Miles R (2004) Interneuron Diversity series: Fast in,

fast out--temporal and spatial signal processing in hippocampal interneurons.

Trends in Neurosciences, 27, 30-40.

Judge SI, Bever CT, Jr. (2006) Potassium channel blockers in multiple sclerosis: neuronal

Kv channels and effects of symptomatic treatment. Pharmacol Ther, 111, 224-59.

173

Jukkola P, Guerrero T, Gray V, Gu C (2013) Astrocytes differentially respond to

inflammatory autoimmune insults and imbalances of neural activity. Acta

Neuropathol Commun, 1, 70.

Jukkola PI, Lovett-Racke AE, Zamvil SS, Gu C (2012) K+ channel alterations in the

progression of experimental autoimmune encephalomyelitis. Neurobiol Dis, 47,

280-93.

Kettenmann H, Banati R, Walz W (1993) Electrophysiological behavior of microglia.

Glia, 7, 93-101.

Kim D, Cavanaugh EJ, Kim I, Carroll JL (2009) Heteromeric TASK-1/TASK-3 is the

major oxygen-sensitive background K+ channel in rat carotid body glomus cells.

J Physiol, 587, 2963-75.

Kim JE, Kwak SE, Choi SY, Kang TC (2008) Region-specific alterations in astroglial

TWIK-related acid-sensitive K+-1 channel immunoreactivity in the rat

hippocampal complex following pilocarpine-induced status epilepticus. J Comp

Neurol, 510, 463-74.

Kim JE, Yeo SI, Ryu HJ, et al. (2010) Astroglial loss and edema formation in the rat

piriform cortex and hippocampus following pilocarpine-induced status

epilepticus. J Comp Neurol, 518, 4612-28.

Kimura A, Hsu M, Seldin M, Verkman AS, Scharfman HE, Binder DK (2010) Protective

role of aquaporin-4 water channels after contusion spinal cord injury. Ann Neurol,

67, 794-801.

174

King IL, Dickendesher TL, Segal BM (2009) Circulating Ly-6C+ myeloid precursors

migrate to the CNS and play a pathogenic role during autoimmune demyelinating

disease. Blood, 113, 3190-7.

Kinoshita M, Nakatsuji Y (2012) Where Do AQP4 Antibodies Fit in the Pathogenesis of

NMO? Mult Scler Int, 2012, 862169.

Koide M, Bonev AD, Nelson MT, Wellman GC (2012) Inversion of neurovascular

coupling by subarachnoid blood depends on large-conductance Ca2+-activated

K+ (BK) channels. Proc Natl Acad Sci U S A, 109, E1387-95.

Kornek B, Storch MK, Bauer J, et al. (2001) Distribution of a calcium channel subunit in

dystrophic axons in multiple sclerosis and experimental autoimmune

encephalomyelitis. Brain, 124, 1114-24.

Koulakoff A, Mei X, Orellana JA, Saez JC, Giaume C (2012) Glial connexin expression

and function in the context of Alzheimer's disease. Biochim Biophys Acta, 1818,

2048-57.

Koutsis G, Karadima G, Floroskufi P, Sfagos C, Vassilopoulos D, Panas M (2011) The

rs10492972 KIF1B polymorphism and disease progression in Greek patients with

multiple sclerosis. J Neurol, 258, 1726-8.

Kreutzer M, Seehusen F, Kreutzer R, et al. (2011) Axonopathy is associated with

complex axonal transport defects in a model of multiple sclerosis. Brain Pathol,

22, 454-71.

Kurkowska-Jastrzebska I, Swiatkiewicz M, Zaremba M, et al. (2013) Neurodegeneration

and inflammation in hippocampus in experimental autoimmune encephalomyelitis

175

induced in rats by one - Time administration of encephalitogenic T cells.

Neuroscience. Sep 17;248:690-8.

Laird DW (2006) Life cycle of connexins in health and disease. Biochem J, 394, 527-43.

Lazarides E (1982) Intermediate filaments: a chemically heterogeneous, developmentally

regulated class of proteins. Annu Rev Biochem, 51, 219-50.

Lee W, Reyes RC, Gottipati MK, et al. (2013) Enhanced Ca-dependent glutamate release

from astrocytes of the BACHD Huntington's disease mouse model. Neurobiol

Dis, 58C, 192-199.

Lehmann-Horn F, Jurkat-Rott K (1999) Voltage-gated ion channels and hereditary

disease. Physiol Rev, 79, 1317-72.

Lennon VA, Wingerchuk DM, Kryzer TJ, et al. (2004) A serum autoantibody marker of

neuromyelitis optica: distinction from multiple sclerosis. Lancet, 364, 2106-12.

Leung YM (2010) Voltage-gated K+ channel modulators as neuroprotective agents. Life

Sci, 86, 775-80.

Leybaert L, Sanderson MJ (2012) Intercellular Ca(2+) waves: mechanisms and function.

Physiol Rev, 92, 1359-92.

Li L, Zhang H, Varrin-Doyer M, Zamvil SS, Verkman AS (2011) Proinflammatory role

of aquaporin-4 in autoimmune neuroinflammation. Faseb J, 25, 1556-66.

Li L, Zhang H, Verkman AS (2009) Greatly attenuated experimental autoimmune

encephalomyelitis in aquaporin-4 knockout mice. BMC Neurosci, 10, 94.

176

Lim ST, Antonucci DE, Scannevin RH, Trimmer JS (2000) A novel targeting signal for

proximal clustering of the Kv2.1 K+ channel in hippocampal neurons. Neuron,

25, 385-97.

Lublin FD, Lavasa M, Viti C, Knobler RL (1987) Suppression of acute and relapsing

experimental allergic encephalomyelitis with mitoxantrone. Clin Immunol

Immunopathol, 45, 122-8.

Lucchinetti CF, Popescu BF, Bunyan RF, et al. (2011) Inflammatory cortical

demyelination in early multiple sclerosis. New England Journal of Medicine, 365,

2188-97.

Mannari T, Morita S, Furube E, Tominaga M, Miyata S (2013) Astrocytic TRPV1 ion

channels detect blood-borne signals in the sensory circumventricular organs of

adult mouse brains. Glia, 61, 957-71.

Marignier R, Nicolle A, Watrin C, et al. (2010) Oligodendrocytes are damaged by

neuromyelitis optica immunoglobulin G via astrocyte injury. Brain, 133, 2578-91.

Masaki K, Suzuki SO, Matsushita T, et al. (2013) Connexin 43 astrocytopathy linked to

rapidly progressive multiple sclerosis and neuromyelitis optica. PLoS One, 8,

e72919.

Mathie A, Veale EL (2007) Therapeutic potential of neuronal two-pore domain

potassium-channel modulators. Curr Opin Investig Drugs, 8, 555-62.

Matthews L, Marasco R, Jenkinson M, et al. (2013) Distinction of seropositive NMO

spectrum disorder and MS brain lesion distribution. Neurology, 80, 1330-7.

177

Misonou H, Menegola M, Mohapatra DP, Guy LK, Park KS, Trimmer JS (2006)

Bidirectional activity-dependent regulation of neuronal ion channel

phosphorylation. J Neurosci, 26, 13505-14.

Misonou H, Mohapatra DP, Park EW, et al. (2004) Regulation of ion channel localization

and phosphorylation by neuronal activity. Nat Neurosci, 7, 711-8.

Misu T, Fujihara K, Kakita A, et al. (2007) Loss of aquaporin 4 in lesions of

neuromyelitis optica: distinction from multiple sclerosis. Brain, 130, 1224-34.

Miyamoto K, Nagaosa N, Motoyama M, Kataoka K, Kusunoki S (2009) Upregulation of

water channel aquaporin-4 in experimental autoimmune encephalomyeritis. J

Neurol Sci, 276, 103-7.

Moll C, Mourre C, Lazdunski M, Ulrich J (1991) Increase of sodium channels in

demyelinated lesions of multiple sclerosis. Brain Res, 556, 311-6.

Moore CS, Abdullah SL, Brown A, Arulpragasam A, Crocker SJ (2011) How factors

secreted from astrocytes impact myelin repair. J Neurosci Res, 89, 13-21.

Muennich EA, Fyffe RE (2004) Focal aggregation of voltage-gated, Kv2.1 subunit-

containing, potassium channels at synaptic sites in rat spinal motoneurones. J

Physiol, 554, 673-85.

Muhammad S, Aller MI, Maser-Gluth C, Schwaninger M, Wisden W (2010) Expression

of the kcnk3 potassium channel gene lessens the injury from cerebral ischemia,

most likely by a general influence on blood pressure. Neuroscience, 167, 758-64.

Nash B, Ioannidou K, Barnett SC (2011a) Astrocyte phenotypes and their relationship to

myelination. J Anat, 219, 44-52.

178

Nash B, Thomson CE, Linington C, et al. (2011b) Functional duality of astrocytes in

myelination. J Neurosci, 31, 13028-38.

Navarrete M, Perea G, Fernandez de Sevilla D, et al. (2012) Astrocytes mediate in vivo

cholinergic-induced synaptic plasticity. PLoS Biol, 10, e1001259.

Nedergaard M, Ransom B, Goldman SA (2003) New roles for astrocytes: redefining the

functional architecture of the brain. Trends Neurosci, 26, 523-30.

Neusch C, Papadopoulos N, Muller M, et al. (2006) Lack of the Kir4.1 channel subunit

abolishes K+ buffering properties of astrocytes in the ventral respiratory group:

impact on extracellular K+ regulation. J Neurophysiol, 95, 1843-52.

Neusch C, Rozengurt N, Jacobs RE, Lester HA, Kofuji P (2001) Kir4.1 potassium

channel subunit is crucial for oligodendrocyte development and in vivo

myelination. J Neurosci, 21, 5429-38.

Nikic I, Merkler D, Sorbara C, et al. (2011) A reversible form of axon damage in

experimental autoimmune encephalomyelitis and multiple sclerosis. Nat Med, 17,

495-9.

Noell S, Wolburg-Buchholz K, Mack AF, et al. (2011) Evidence for a role of

dystroglycan regulating the membrane architecture of astroglial endfeet. Eur J

Neurosci, 33, 2179-86.

Nutile-McMenemy N, Elfenbein A, Deleo JA (2007) Minocycline decreases in vitro

microglial motility, beta1-integrin, and Kv1.3 channel expression. J Neurochem,

103, 2035-46.

179

O'Connell KM, Loftus R, Tamkun MM (2010) Localization-dependent activity of the

Kv2.1 delayed-rectifier K+ channel. Proc Natl Acad Sci U S A, 107, 12351-6.

O'Neill JK, Baker D, Morris MM, et al. (1998) Optic neuritis in chronic relapsing

experimental allergic encephalomyelitis in Biozzi ABH mice: demyelination and

fast axonal transport changes in disease. J Neuroimmunol, 82, 210-8.

Ogawa Y, Oses-Prieto J, Kim MY, et al. (2010) ADAM22, a Kv1 channel-interacting

protein, recruits membrane-associated guanylate kinases to juxtaparanodes of

myelinated axons. J Neurosci, 30, 1038-48.

Olsen ML, Higashimori H, Campbell SL, Hablitz JJ, Sontheimer H (2006) Functional

expression of Kir4.1 channels in spinal cord astrocytes. Glia, 53, 516-28.

Pacheco Otalora LF, Hernandez EF, Arshadmansab MF, et al. (2008) Down-regulation of

BK channel expression in the pilocarpine model of temporal lobe epilepsy. Brain

Res, 1200, 116-31.

Pal SK, Takimoto K, Aizenman E, Levitan ES (2006) Apoptotic surface delivery of K+

channels. Cell Death Differ, 13, 661-7.

Pan Z, Kao T, Horvath Z, et al. (2006) A common ankyrin-G-based mechanism retains

KCNQ and NaV channels at electrically active domains of the axon. J Neurosci,

26, 2599-613.

Pannasch U, Farber K, Nolte C, et al. (2006) The potassium channels Kv1.5 and Kv1.3

modulate distinct functions of microglia. Mol Cell Neurosci, 33, 401-11.

Papadopoulos MC, Verkman AS (2013) Aquaporin water channels in the nervous

system. Nat Rev Neurosci, 14, 265-77.

180

Perea G, Navarrete M, Araque A (2009) Tripartite synapses: astrocytes process and

control synaptic information. Trends Neurosci, 32, 421-31.

Perillan PR, Chen M, Potts EA, Simard JM (2002) Transforming growth factor-beta 1

regulates Kir2.3 inward rectifier K+ channels via phospholipase C and protein

kinase C-delta in reactive astrocytes from adult rat brain. J Biol Chem, 277, 1974-

80.

Petzold GC, Murthy VN (2011) Role of astrocytes in neurovascular coupling. Neuron,

71, 782-97.

Popko B, Corbin JG, Baerwald KD, Dupree J, Garcia AM (1997) The effects of

interferon-gamma on the central nervous system. Mol Neurobiol, 14, 19-35.

Potokar M, Stenovec M, Jorgacevski J, et al. (2013) Regulation of AQP4 surface

expression via vesicle mobility in astrocytes. Glia, 61, 917-28.

Puente N, Mendizabal-Zubiaga J, Elezgarai I, Reguero L, Buceta I, Grandes P (2010)

Precise localization of the voltage-gated potassium channel subunits Kv3.1b and

Kv3.3 revealed in the molecular layer of the rat cerebellar cortex by a pre-

embedding immunogold method. Histochem Cell Biol, 134, 403-9.

Quaegebeur A, Lange C, Carmeliet P (2011) The neurovascular link in health and

disease: molecular mechanisms and therapeutic implications. Neuron, 71, 406-24.

Radin MJ, Yu MJ, Stoedkilde L, et al. (2012) Aquaporin-2 regulation in health and

disease. Vet Clin Pathol, 41, 455-70.

181

Ramadoss J, Lunde ER, Ouyang N, Chen WJ, Cudd TA (2008) Acid-sensitive channel

inhibition prevents fetal alcohol spectrum disorders cerebellar Purkinje cell loss.

Am J Physiol Regul Integr Comp Physiol, 295, R596-603.

Ransohoff RM, Brown MA (2012) Innate immunity in the central nervous system.

Journal of Clinical Investigation, 122, 1164-71.

Rasband MN, Park EW, Vanderah TW, Lai J, Porreca F, Trimmer JS (2001) Distinct

potassium channels on pain-sensing neurons. Proc Natl Acad Sci U S A, 98,

13373-8.

Rasband MN, Shrager P (2000) Ion channel sequestration in central nervous system

axons. J Physiol, 525 Pt 1, 63-73.

Rasband MN, Trimmer JS, Peles E, Levinson SR, Shrager P (1999) K+ channel

distribution and clustering in developing and hypomyelinated axons of the optic

nerve. J Neurocytol, 28, 319-31.

Rasband MN, Trimmer JS, Schwarz TL, et al. (1998) Potassium channel distribution,

clustering, and function in remyelinating rat axons. J Neurosci, 18, 36-47.

Reyes RC, Verkhratsky A, Parpura V (2013) TRPC1-mediated Ca(2+) and Na(+)

signalling in astroglia: Differential filtering of extracellular cations. Cell Calcium,

54, 120-5.

Rhodes KJ, Keilbaugh SA, Barrezueta NX, Lopez KL, Trimmer JS (1995) Association

and colocalization of K+ channel alpha- and beta-subunit polypeptides in rat

brain. J Neurosci, 15, 5360-71.

182

Rhodes KJ, Strassle BW, Monaghan MM, BekeleArcuri Z, Matos MF, Trimmer JS

(1997) Association and colocalization of the Kv beta 1 and Kv beta 2 beta-

subunits with Kv1 alpha-subunits in mammalian brain K+ channel complexes.

Journal of Neuroscience, 17, 8246-8258.

Rojek A, Praetorius J, Frokiaer J, Nielsen S, Fenton RA (2008) A current view of the

mammalian aquaglyceroporins. Annu Rev Physiol, 70, 301-27.

Rossi A, Moritz TJ, Ratelade J, Verkman AS (2012) Super-resolution imaging of

aquaporin-4 orthogonal arrays of particles in cell membranes. J Cell Sci, 125,

4405-12.

Rudy B, McBain CJ (2001) Kv3 channels: voltage-gated K+ channels designed for high-

frequency repetitive firing. Trends Neurosci, 24, 517-26.

Ruiz-Ederra J, Zhang H, Verkman AS (2007) Evidence against functional interaction

between aquaporin-4 water channels and Kir4.1 potassium channels in retinal

Muller cells. J Biol Chem, 282, 21866-72.

Rusznak Z, Pocsai K, Kovacs I, et al. (2004) Differential distribution of TASK-1, TASK-

2 and TASK-3 immunoreactivities in the rat and human cerebellum. Cell Mol Life

Sci, 61, 1532-42.

Saadoun S, Waters P, Macdonald C, et al. (2011) T cell deficiency does not reduce

lesions in mice produced by intracerebral injection of NMO-IgG and complement.

J Neuroimmunol, 235, 27-32.

Salkoff L, Butler A, Ferreira G, Santi C, Wei A (2006) High-conductance potassium

channels of the SLO family. Nat Rev Neurosci, 7, 921-31.

183

Salzer JL (2003) Polarized domains of myelinated axons. Neuron, 40, 297-318.

Sanchez JA, Ho CS, Vaughan DM, Garcia MC, Grange RW, Joho RH (2000) Muscle and

motor-skill dysfunction in a K+ channel-deficient mouse are not due to altered

muscle excitability or fiber type but depend on the genetic background. Pflugers

Arch, 440, 34-41.

Schaller KL, Krzemien DM, Yarowsky PJ, Krueger BK, Caldwell JH (1995) A novel,

abundant sodium channel expressed in neurons and glia. J Neurosci, 15, 3231-42.

Scharfman HE, Binder DK (2013) Aquaporin-4 water channels and synaptic plasticity in

the hippocampus. Neurochem Int.

Schipke CG, Kettenmann H (2004) Astrocyte responses to neuronal activity. Glia, 47,

226-32.

Seidel KN, Derst C, Salzmann M, et al. (2011) Expression of the voltage- and Ca2+-

dependent BK potassium channel subunits BKbeta1 and BKbeta4 in rodent

astrocytes. Glia, 59, 893-902.

Seifert G, Huttmann K, Binder DK, et al. (2009) Analysis of astroglial K+ channel

expression in the developing hippocampus reveals a predominant role of the

Kir4.1 subunit. J Neurosci, 29, 7474-88.

Seri B, Garcia-Verdugo JM, McEwen BS, Alvarez-Buylla A (2001) Astrocytes give rise

to new neurons in the adult mammalian hippocampus. J Neurosci, 21, 7153-60.

Sheng M, Tsaur ML, Jan YN, Jan LY (1992) Subcellular Segregation of 2 a-Type K+

Channel Proteins in Rat Central Neurons. Neuron, 9, 271-284.

184

Shepherd AJ, Loo L, Gupte RP, Mickle AD, Mohapatra DP (2012) Distinct modifications

in Kv2.1 channel via chemokine receptor CXCR4 regulate neuronal survival-

death dynamics. J Neurosci, 32, 17725-39.

Shigetomi E, Bowser DN, Sofroniew MV, Khakh BS (2008) Two forms of astrocyte

calcium excitability have distinct effects on NMDA receptor-mediated slow

inward currents in pyramidal neurons. J Neurosci, 28, 6659-63.

Shruti S, Clem RL, Barth AL (2008) A seizure-induced gain-of-function in BK channels

is associated with elevated firing activity in neocortical pyramidal neurons.

Neurobiol Dis, 30, 323-30.

Sild M, Ruthazer ES (2011) Radial glia: progenitor, pathway, and partner. Neuroscientist,

17, 288-302.

Sinclair C, Kirk J, Herron B, Fitzgerald U, McQuaid S (2007) Absence of aquaporin-4

expression in lesions of neuromyelitis optica but increased expression in multiple

sclerosis lesions and normal-appearing white matter. Acta Neuropathol, 113, 187-

94.

Sinha K, Karimi-Abdolrezaee S, Velumian AA, Fehlings MG (2006) Functional changes

in genetically dysmyelinated spinal cord axons of shiverer mice: role of

juxtaparanodal Kv1 family K+ channels. J Neurophysiol, 95, 1683-95.

Skucas VA, Mathews IB, Yang J, et al. (2011) Impairment of select forms of spatial

memory and neurotrophin-dependent synaptic plasticity by deletion of glial

aquaporin-4. J Neurosci, 31, 6392-7.

185

Slavin A, Kelly-Modis L, Labadia M, Ryan K, Brown ML (2010) Pathogenic

mechanisms and experimental models of multiple sclerosis. Autoimmunity, 43,

504-13.

Smart SL, Bosma MM, Tempel BL (1997) Identification of the delayed rectifier

potassium channel, Kv1.6, in cultured astrocytes. Glia, 20, 127-34.

Sobotzik JM, Sie JM, Politi C, et al. (2009) AnkyrinG is required to maintain axo-

dendritic polarity in vivo. Proc Natl Acad Sci U S A, 106, 17564-9.

Sofroniew MV, Vinters HV (2010) Astrocytes: biology and pathology. Acta

Neuropathol, 119, 7-35.

Song AH, Wang D, Chen G, et al. (2009) A selective filter for cytoplasmic transport at

the axon initial segment. Cell, 136, 1148-60.

Sontheimer H, Fernandez-Marques E, Ullrich N, Pappas CA, Waxman SG (1994)

Astrocyte Na+ channels are required for maintenance of Na+/K(+)-ATPase

activity. J Neurosci, 14, 2464-75.

Sontheimer H, Waxman SG (1992) Ion channels in spinal cord astrocytes in vitro. II.

Biophysical and pharmacological analysis of two Na+ current types. J

Neurophysiol, 68, 1001-11.

Sorensen A, Moffat K, Thomson C, Barnett SC (2008) Astrocytes, but not olfactory

ensheathing cells or Schwann cells, promote myelination of CNS axons in vitro.

Glia, 56, 750-63.

Srivastava R, Aslam M, Kalluri SR, et al. (2012) Potassium channel KIR4.1 as an

immune target in multiple sclerosis. N Engl J Med, 367, 115-23.

186

Steiner E, Enzmann GU, Lin S, et al. (2012) Loss of astrocyte polarization upon transient

focal brain ischemia as a possible mechanism to counteract early edema

formation. Glia, 60, 1646-59.

Steinman L (2005) Blocking adhesion molecules as therapy for multiple sclerosis:

natalizumab. Nat Rev Drug Discov, 4, 510-8.

Stephan J, Haack N, Kafitz KW, et al. (2012) Kir4.1 channels mediate a depolarization of

hippocampal astrocytes under hyperammonemic conditions in situ. Glia, 60, 965-

78.

Stuart G, Schiller J, Sakmann B (1997) Action potential initiation and propagation in rat

neocortical pyramidal neurons. J Physiol, 505 ( Pt 3), 617-32.

Thrane AS, Rappold PM, Fujita T, et al. (2011) Critical role of aquaporin-4 (AQP4) in

astrocytic Ca2+ signaling events elicited by cerebral edema. Proc Natl Acad Sci U

S A, 108, 846-51.

Tomizawa Y, Yokoyama K, Saiki S, Takahashi T, Matsuoka J, Hattori N (2012) Blood-

brain barrier disruption is more severe in neuromyelitis optica than in multiple

sclerosis and correlates with clinical disability. J Int Med Res, 40, 1483-91.

Trimmer JS, Rhodes KJ (2004) Localization of voltage-gated ion channels in mammalian

brain. Annu Rev Physiol, 66, 477-519.

Umenishi F, Verkman AS (1998) Isolation and functional analysis of alternative

promoters in the human aquaporin-4 water channel gene. Genomics, 50, 373-7.

187

Vabnick I, Trimmer JS, Schwarz TL, Levinson SR, Risal D, Shrager P (1999) Dynamic

potassium channel distributions during axonal development prevent aberrant

firing patterns. J Neurosci, 19, 747-58.

Vacher H, Mohapatra DP, Trimmer JS (2008) Localization and targeting of voltage-

dependent ion channels in mammalian central neurons. Physiol Rev, 88, 1407-47.

Verkman AS, Binder DK, Bloch O, Auguste K, Papadopoulos MC (2006) Three distinct

roles of aquaporin-4 in brain function revealed by knockout mice. Biochim

Biophys Acta, 1758, 1085-93.

Vogt J, Paul F, Aktas O, et al. (2009) Lower motor neuron loss in multiple sclerosis and

experimental autoimmune encephalomyelitis. Ann Neurol, 66, 310-22.

Volterra A, Meldolesi J (2005) Astrocytes, from brain glue to communication elements:

the revolution continues. Nat Rev Neurosci, 6, 626-40.

Wang B, Brenner R (2006) An S6 mutation in BK channels reveals beta1 subunit effects

on intrinsic and voltage-dependent gating. J Gen Physiol, 128, 731-44.

Wang B, Rothberg BS, Brenner R (2009) Mechanism of increased BK channel activation

from a channel mutation that causes epilepsy. J Gen Physiol, 133, 283-94.

Wang H, Allen ML, Grigg JJ, Noebels JL, Tempel BL (1995) Hypomyelination alters K+

channel expression in mouse mutants shiverer and Trembler. Neuron, 15, 1337-

47.

Wang H, Kunkel DD, Martin TM, Schwartzkroin PA, Tempel BL (1993)

Heteromultimeric K+ channels in terminal and juxtaparanodal regions of neurons.

Nature, 365, 75-9.

188

Wang H, Kunkel DD, Schwartzkroin PA, Tempel BL (1994) Localization of Kv1.1 and

Kv1.2, two K channel proteins, to synaptic terminals, somata, and dendrites in the

mouse brain. J Neurosci, 14, 4588-99.

Waters P, Vincent A (2008) Detection of anti-aquaporin-4 antibodies in neuromyelitis

optica: current status of the assays. International Ms Journal, 15, 99-105.

Watkins TA, Emery B, Mulinyawe S, Barres BA (2008) Distinct stages of myelination

regulated by gamma-secretase and astrocytes in a rapidly myelinating CNS

coculture system. Neuron, 60, 555-69.

Waxman SG (1982) Membranes, myelin, and the pathophysiology of multiple sclerosis.

N Engl J Med, 306, 1529-33.

Waxman SG (2002) Ion channels and neuronal dysfunction in multiple sclerosis. Arch

Neurol, 59, 1377-80.

Waxman SG (2006) Axonal conduction and injury in multiple sclerosis: the role of

sodium channels. Nat Rev Neurosci, 7, 932-41.

Waxman SG, Craner MJ, Black JA (2004) Na+ channel expression along axons in

multiple sclerosis and its models. Trends Pharmacol Sci, 25, 584-91.

Weiser M, Vega-Saenz de Miera E, Kentros C, et al. (1994) Differential expression of

Shaw-related K+ channels in the rat central nervous system. J Neurosci, 14, 949-

72.

Wilcock DM, Vitek MP, Colton CA (2009) Vascular amyloid alters astrocytic water and

potassium channels in mouse models and humans with Alzheimer's disease.

Neuroscience, 159, 1055-69.

189

Willis M, Kaufmann WA, Wietzorrek G, et al. (2010) L-type calcium channel CaV 1.2 in

transgenic mice overexpressing human AbetaPP751 with the London (V717I) and

Swedish (K670M/N671L) mutations. J Alzheimers Dis, 20, 1167-80.

Wong YL, Rice SE (2010) Kinesin's light chains inhibit the head- and microtubule-

binding activity of its tail. Proc Natl Acad Sci U S A, 107, 11781-6.

Woo DH, Han KS, Shim JW, et al. (2012) TREK-1 and Best1 channels mediate fast and

slow glutamate release in astrocytes upon GPCR activation. Cell, 151, 25-40.

Wu CY, Kaur C, Sivakumar V, Lu J, Ling EA (2009) Kv1.1 expression in microglia

regulates production and release of proinflammatory cytokines, endothelins and

nitric oxide. Neuroscience, 158, 1500-8.

Wu X, Hsuchou H, Kastin AJ, Mishra PK, Pan W (2012) Upregulation of astrocytic

leptin receptor in mice with experimental autoimmune encephalomyelitis. J Mol

Neurosci, 49, 446-56.

Wulff H, Beeton C, Chandy KG (2003a) Potassium channels as therapeutic targets for

autoimmune disorders. Curr Opin Drug Discov Devel, 6, 640-7.

Wulff H, Calabresi PA, Allie R, et al. (2003b) The voltage-gated Kv1.3 K(+) channel in

effector memory T cells as new target for MS. J Clin Invest, 111, 1703-13.

Xu JH, Long L, Tang YC, Hu HT, Tang FR (2007a) Ca(v)1.2, Ca(v)1.3, and Ca(v)2.1 in

the mouse hippocampus during and after pilocarpine-induced status epilepticus.

Hippocampus, 17, 235-51.

190

Xu M, Cao R, Xiao R, Zhu MX, Gu C (2007b) The axon-dendrite targeting of Kv3

(Shaw) channels is determined by a targeting motif that associates with the T1

domain and ankyrin G. J Neurosci, 27, 14158-70.

Xu M, Gu Y, Barry J, Gu C (2010) Kinesin I transports tetramerized Kv3 channels

through the axon initial segment via direct binding. J Neurosci, 30, 15987-6001.

Yong VW, Moumdjian R, Yong FP, et al. (1991) Gamma-interferon promotes

proliferation of adult human astrocytes in vitro and reactive gliosis in the adult

mouse brain in vivo. Proc Natl Acad Sci U S A, 88, 7016-20.

Young NP, Weinshenker BG, Parisi JE, et al. (2010) Perivenous demyelination:

association with clinically defined acute disseminated encephalomyelitis and

comparison with pathologically confirmed multiple sclerosis. Brain, 133, 333-48.

Youssef S, Stuve O, Patarroyo JC, et al. (2002) The HMG-CoA reductase inhibitor,

atorvastatin, promotes a Th2 bias and reverses paralysis in central nervous system

autoimmune disease. Nature, 420, 78-84.

Yu SP, Yeh CH, Sensi SL, et al. (1997) Mediation of neuronal apoptosis by enhancement

of outward potassium current. Science, 278, 114-7.

Yukutake Y, Yasui M (2010) Regulation of water permeability through aquaporin-4.

Neuroscience, 168, 885-91.

Zagha E, Lang EJ, Rudy B (2008) Kv3.3 channels at the Purkinje cell soma are necessary

for generation of the classical complex spike waveform. Journal of Neuroscience,

28, 1291-300.

191

Zamponi GW, Lewis RJ, Todorovic SM, Arneric SP, Snutch TP (2009) Role of voltage-

gated calcium channels in ascending pain pathways. Brain Res Rev, 60, 84-9.

Zamvil SS, Steinman L (1990) The T lymphocyte in experimental allergic

encephalomyelitis. Annu Rev Immunol, 8, 579-621.

Zhang H, Verkman AS (2008) Aquaporin-4 independent Kir4.1 K+ channel function in

brain glial cells. Mol Cell Neurosci, 37, 1-10.

Zhou L, Zhang CL, Messing A, Chiu SY (1998) Temperature-sensitive neuromuscular

transmission in Kv1.1 null mice: role of potassium channels under the myelin

sheath in young nerves. J Neurosci, 18, 7200-15.

Zhou M, Xu G, Xie M, et al. (2009) TWIK-1 and TREK-1 are potassium channels

contributing significantly to astrocyte passive conductance in rat hippocampal

slices. J Neurosci, 29, 8551-64.

Zhu B, Luo L, Moore GR, Paty DW, Cynader MS (2003) Dendritic and synaptic

pathology in experimental autoimmune encephalomyelitis. Am J Pathol, 162,

1639-50.

Ziehn MO, Avedisian AA, Dervin SM, O'Dell TJ, Voskuhl RR (2012a) Estriol preserves

synaptic transmission in the hippocampus during autoimmune demyelinating

disease. Lab Invest, 92, 1234-45.

Ziehn MO, Avedisian AA, Dervin SM, Umeda EA, O'Dell TJ, Voskuhl RR (2012b)

Therapeutic testosterone administration preserves excitatory synaptic transmission

in the hippocampus during autoimmune demyelinating disease. J Neurosci, 32,

12312-24.

192

Ziehn MO, Avedisian AA, Tiwari-Woodruff S, Voskuhl RR (2010) Hippocampal CA1

atrophy and synaptic loss during experimental autoimmune encephalomyelitis,

EAE. Lab Invest, 90, 774-86.

Zlokovic BV (2008) The blood-brain barrier in health and chronic neurodegenerative

disorders. Neuron, 57, 178-201.

193