ACUTE MICROGLIAL RESPONSES TO SINGLE NON-EPILEPTOGENIC VS. EPILEPTOGENIC SEIZURES IN MOUSE HIPPOCAMPUS

A Dissertation submitted to the Faculty of the Graduate School of Arts and Sciences of Georgetown University in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Neuroscience

By

Alberto Sepulveda-Rodriguez, B.S.

Washington, D.C.

June 5, 2019

Copyright 2019 by Alberto Sepulveda-Rodriguez

All Rights Reserved

ii

ACUTE MICROGLIAL RESPONSES TO SINGLE NON-EPILEPTOGENIC VS. EPILEPTOGENIC SEIZURES IN MOUSE HIPPOCAMPUS

Alberto Sepulveda-Rodriguez, B.S.

Thesis Advisor: Stefano Vicini, Ph.D.

ABSTRACT

As the resident macrophage of the central nervous system, microglia are in a uniquely privileged position to both affect and be affected by neuroinflammation, neuronal activity and injury, which are all hallmarks of seizures and the epilepsies. In one of the prototypical rodent models of seizure-induced epilepsy, hippocampal microglia become activated after prolonged, damaging seizures known as Status

Epilepticus (SE). However, since SE comprises both neuronal hyperactivity and injury, the specific mechanisms triggering this microglial activation remain unclear, as does its relevance to the ensuing epileptogenic processes.

The present studies employed another well-established seizure model, electroconvulsive shock (ECS), to study the effect of paroxysmal/ictal neuronal hyperactivity on mouse hippocampal microglia, in the absence of concomitant neuronal degeneration. Unlike SE, ECS did not cause neuronal injury and did not alter hippocampal CA1 microglial and astrocytic density, morphology nor baseline process motility. In contrast, both ECS and SE produced a similar increase in ATP-directed microglial process motility in acute slices and similarly upregulated expression of the

iii

chemokine CCL2. As opposed to the general pro-inflammatory environment produced after SE (where induction of CCL2, TNF, IL-1 and IL-6 signaling among others have been reported), the acute microglial response to ECS was more selective: I found upregulated expression of Ccl2 message but unchanged transcript levels of the pro- inflammatory cytokine Tnf. Whole-cell patch-clamp recordings of hippocampal CA1sr microglia showed that ECS enhanced purinergic currents mediated by P2X7 receptors, in the absence of changes in passive properties or voltage-gated currents or changes in receptor expression.

This differs from previously described alterations in intrinsic characteristics and purinoreceptor upregulation which coincided with enhanced purinergic currents following SE. These ECS-induced effects point to a common “seizure signature” in hippocampal microglia, characterized by functional alterations to purinergic signaling.

The data herein demonstrate that ictal activity per se can drive changes in microglial physiology. These described physiological changes (which up until now have been exclusively associated with prolonged and damaging seizures) are of added interest as they may be relevant to electroconvulsive therapy which remains a gold-standard treatment for depression.

iv

ACKNOWLEDGEMENTS

“O Lord, that lends me life, lend me a heart replete with thankfulness!”

- William Shakespeare, Henry VI

The research and writing of this thesis are dedicated to:

 Karen Gale, who picked my application and called me from the hospital to

propose that we submit the CONACyT fellowship that supported my work

these past 5 years. Nothing would make me feel prouder than when others

would comment on how she would be proud of my progress and project.

 My mother Cecilia and my father Alberto, their spouses and all my siblings,

but especially Gua for the constant inspiration. I honestly feel that she

deserved a CONACyT award more than me, and am certain she would have

received one if they still existed.

 The rest of my family, who have set me up for success no matter where or

what I choose to dedicate myself to.

 My talented, intelligent and beautiful partner Meghan, and her entire

family that has welcome me with open arms from the very start.

 My caring mentors Stefano Vicini and Patrick Forcelli, who have taught me

PLENTY of physiology, pharmacology and other biosciences, but even more

v

about life. You continue to inspire me with your paradoxical youth and

wisdom respectively.

 My amazingly diligent thesis committee (chaired by John Partridge and

featuring Lindsay DeBiase, Kathleen Maguire-Zeiss and Patrick Forcelli)

who have been instrumental for this project. I could not have asked for

anything more than what they gave me. Their time, effort and energy

collectively made this a more rigorous and impactful study, as well as a

better story.

 Other faculty that helped me and cared about me and my project without

any obligation to do so, namely Bob Yasuda, Kathy Conant, Robin Tucker,

Ken Kellar, Tom Coate, Brent Harris, Ludise Malkova, Barry Wolfe, Bill

Rebeck, JY Wu and Mark Burns among many others.

 My close friends and labmates Nour, Adam, Selena, Lorenza, Irene and

David, and obviously Steph too because ephys is hard. Also, the Forcelli-

Malkova lab for being so welcoming and making me feel like I had a 2nd

home there.

 Megan Huizenga, Pinggan Li, Tahiyana Khan, James Ma, Colby Carlone,

Lorenzo Bozzelli, Carissa Winland, Yan Li and the rest of my coauthors.

 The students and fellows that allowed me to learn even more by training

them, namely Tahi, Serena, Colby, Rebecca and Pinggan.

vi

 My happy hour/venting/dreaming support group, who helped me cope:

Erin, Megan, Kaela, Gabby, Lorenzo, Jeremiah, Steph D, Edith, Steph S, Evan,

Thien, Saf, and many others.

 The animals that were used in my experiments, and the DCM and GUACUC

staff devoted to protecting them while enabling our research.

 All the other faculty, staff and students that supported me and kept me

sane through this endeavor. It really does take a village, and there are

entirely too many people to list.

 My best friends David and Martin. I would do anything for you guys, and I

know you would and indeed have done the same for me.

 The brothers of my fraternity Delta Upsilon, particularly my gigantic and

raunchy Alpha Upsilon pledge class, the classes of 2011 and 2014 seniors

who like me are mostly overachievers while always having a good time,

Trevor and Kyle, the presidents that preceded me and were always

available to guide me, and the officers that served on my executive board

during those trying times for DU Rochester. Dikaia Upotheke indeed.

 The faculty and staff at my alma mater, particularly Ben Hayden who

introduced me to neuroscience research, Renee Miller who awoke in me a

latent fascination with neurochemistry and pharmacology in me, and Anne

Shields for her continued mentorship.

 The IPN and BGE for funding my last 5 years here together with CONACyT.

vii

 GU for providing me a place to not only learn and research my own science,

but also become exposed to a never-ending flux of interesting people, from

amazing scientists and clinicians to policymakers, world leaders and even

Nobel Prize laureates. It will truly be an honor and a privilege to call myself

a Georgetown alum soon.

Eternally grateful,

ALBERTO

viii

TABLE OF CONTENTS

1. INTRODUCTION ...... 1

A. Seizures and epilepsy ...... 3

i. Treatment approaches against the epilepsies ...... 4

ii. Status epilepticus ...... 7

B. Sequelae of damaging seizures have potential links to epileptogenesis...... 9

i. Neural plasticity, neurogenesis and structural changes ...... 10

ii. Neuronal and blood brain brain barrier damage, metabolic or oxidative

stress and inflammation ...... 11

C. Microglia: A labile cell type enabled by unique physiology ...... 14

i. Roles in development, physiology and pathology ...... 15

ii. Microglia become “activated” in response to a wide array of insults...... 16

iii. Morphology and motility: “Scouting” microglia from box scores vs. game

film ...... 18

iv. Physiology of microglia ...... 22

v. P2 purinoreceptors underlie many basic functions of microglia ...... 25

vi. Focusing on P2X7R as a key mediator of microglial function in health and

disease ...... 28

D. Microglial responses to seizures and potential roles in epilepsy ...... 32

i. A survey of studies on acute microglial responses to status epilepticus ...... 34

ii. Experimental status epilepticus induces a particular activation state in

murine hippocampal microglia ...... 35

ix

E. A gap in our knowledge: “Benign” seizures and microglia ...... 39

i. Non-epileptogenic seizures are common and can even be beneficial ...... 40

ii. Microglial response to “benign” seizures: A field clouded by a dearth of

relevant physiology data ...... 43

F. Emerging hypothesis ...... 44

2. MATERIALS, METHODS AND EXPERIMENTAL DESIGN ...... 46

A. Experimental animals and seizure models ...... 46

i. Electroconvulsive shock (ECS)-induced seizures ...... 47

ii. Pilocarpine-induced status epilepticus (SE) ...... 48

B. Fixed tissue sectioning ...... 49

C. FluoroJade C staining ...... 50

D. Confocal imaging of hippocampal area CA1 in brain slices ...... 51

E. Astrocytic cell density and area ...... 51

F. Microglial cell density and area ...... 52

G. Immunostaining and microglial morphometry ...... 52

H. Preparation of acute brain slices ...... 53

I. Live imaging of microglial motility in situ in acute hippocampal slices ...... 54

i. Basal motility or “surveillance” ...... 54

ii. Responsive motility assay ...... 55

J. Patch-clamp electrophysiology ...... 55

i. Passive or intrinsic membrane properties and current-voltage relation ...... 56

ii. Quantification and biophysical characterization of P2X currents ...... 57

iii. Pharmacological characterization of P2X currents with Brilliant Blue G ..... 57

x

K. Microglial gene expression analysis ...... 58

i. Immunomagnetic isolation of microglia ...... 58

ii. Lysis of isolated cells and extraction and purification of total RNA ...... 60

iii. RT-qPCR ...... 60

L. Quantification of CCL2 and total protein in hippocampal lysates ...... 63

M. Experimental design and statistical analysis ...... 63

3. ELECTROCONVULSIVE SHOCK: A NON-DAMAGING AND PUTATIVELY NON-

EPILEPTOGENIC SEIZURE MODEL...... 65

A. Introduction ...... 65

B. Experimental approach ...... 65

C. Results ...... 66

i. Electroconvulsive shock does not cause hippocampal astrogliosis 24 hours

after seizures ...... 66

ii. Unlike status epilepticus, single electroconvulsive shock seizures do not

result in acute neuronal injury ...... 69

D. Summary and conclusion ...... 70

4. ELECTROCONVULSIVE SHOCK HAS NO EFFECT ON MICROGLIAL DENSITY OR

MORPHOLOGY IN CA1 ...... 71

A. Introduction ...... 71

B. Experimental approach ...... 71

C. Results ...... 72

i. Electroconvulsive shock does not cause acute proliferation of microglia

in CA1 ...... 72

xi

ii. Unlike status epilepticus, electroconvulsive shock has no effect on CA1sr

microglial morphology ...... 76

D. Summary and conclusion ...... 80

5. EFFECTS OF ELECTROCONVULSIVE SHOCK ON MICROGLIAL PROCESS

MOTILITY IN CA1 ...... 81

A. Introduction ...... 81

B. Experimental approach ...... 82

C. Results ...... 83

i. Electroconvulsive shock has no effect on baseline microglial motility...... 83

ii. Like status epilepticus, electroconvulsive shock enhances the response

velocity of microglial processes towards ATP ...... 85

iii. Electroconvulsive shock-induced enhancement of microglial response to

ATP is concentration-dependent ...... 88

D. Summary and conclusion ...... 90

6. EFFECTS OF ELECTROCONVULSIVE SHOCK ON CA1SR MICROGLIAL

MEMBRANE PHYSIOLOGY ...... 91

A. Introduction ...... 91

B. Experimental approach ...... 91

C. Results ...... 94

i. Electroconvulsive shock has no effect on intrinsic properties of CA1sr

microglia ...... 94

ii. ATP-evoked currents are modulated by divalent cations and enhanced by

electroconvulsive shock ...... 98

xii

iii. Electroconvulsive shock-enhanced, ATP-evoked microglial currents are

Brilliant Blue G-sensitive ...... 101

D. Summary and conclusion ...... 103

7. EFFECTS OF ELECTROCONVULSIVE SHOCK ON MICROGLIAL PURINERGIC

GENE EXPRESSION ...... 104

A. Introduction ...... 104

B. Experimental approach ...... 104

C. Results ...... 105

i. Electroconvulsive shock does not affect microglial expression of P2Y

metabotropic receptors ...... 105

ii. Electroconvulsive shock does not affect microglial expression of P2X

ionotropic receptors ...... 107

D. Summary and conclusion ...... 108

8. EFFECTS OF ELECTROCONVULSIVE SHOCK ON NEUROINFLAMMATION IN

THE HIPPOCAMPUS ...... 109

A. Introduction ...... 109

B. Experimental approach ...... 110

C. Results ...... 110

i. Electroconvulsive shock induces increased microglial expression of Ccl2,

but not Tmem119 or Tnf ...... 110

ii. Electroconvulsive shock, like status epilepticus, induces expression of CCL2

protein in the hippocampus ...... 113

D. Summary and conclusion ...... 115

xiii

9. DISCUSSION, FUTURE DIRECTIONS AND CONCLUSION ...... 116

A. Summary of findings ...... 116

B. Discussion and caveats ...... 117

C. Future directions: Microglial response to electroconvulsive shock and… ...... 131

i. The role of fractalkine signaling ...... 132

ii. The role of microglial [Ca2+]i ...... 133

iii. The role of astrocytes ...... 134

iv. Its effects on seizure propensity and sequelae ...... 136

v. Its potential role in the beneficial or deleterious effects of ECT ...... 138

D. Conclusion ...... 138

APPENDIX I: EXTENDED DATA ...... 140

A. 2D-projected and binarized 3D tracings of microglia in mouse CA1sr ...... 141

i. Under control conditions ...... 142

ii. 24 hours after electroconvulsive shock ...... 149

iii. 24 hours after status epilepticus ...... 156

REFERENCES ...... 161

xiv

LIST OF FIGURES

Figure 1: Electroconvulsive shock does not cause acute hippocampal astrogliosis or neurodegeneration ...... 68

Figure 2: Electroconvulsive shock does not trigger acute microglial proliferation in hippocampal area CA1 ...... 74

Figure 3: Electroconvulsive shock does not cause microglial somatic changes in hippocampal area CA1 ...... 75

Figure 4: Electroconvulsive shock does not result in simplification of microglial arborization in CA1sr ...... 78

Figure 5: Electroconvulsive shock does not trigger changes in microglial morphometry in CA1sr ...... 79

Figure 6: Electroconvulsive shock does not alter hippocampal microglial baseline motility ...... 84

Figure 7: Seizure-induced potentiation of responsive microglial process extension behavior ...... 86

Figure 8: Seizures enhance the velocity of microglia responding to 3 mM [ATP] ..... 87

Figure 9: Electroconvulsive shock potentiates microglial motility in response to 1 and 3 but not 10 mM [ATP] ...... 89

Figure 10: Microglial patch-clamp electrophysiology experiments ...... 93

Figure 11: Electroconvulsive shock does not affect intrinsic membrane properties of CA1sr microglia ...... 95

Figure 12: Electroconvulsive shock does not affect current-voltage relation in CA1sr microglia ...... 97

Figure 13: Purinergic currents in CA1sr microglia potentiated by 0CaMg ...... 99

Figure 14: Electroconvulsive shock-enhanced, ATP-evoked microglial cationic currents potentiated by 0CaMg ...... 100

Figure 15: Electroconvulsive shock-enhanced, ATP-evoked microglial currents are mostly mediated by P2X7R ...... 102

xv

Figure 16: Effects of electroconvulsive shock on microglial expression of P2Y purinoreceptor genes ...... 106

Figure 17: Effects of electroconvulsive shock on microglial expression of P2X purinoreceptor genes ...... 108

Figure 18: Electroconvulsive shock increases expression of Ccl2 in microglia ...... 112

Figure 19: Electroconvulsive shock increases expression of CCL2 protein in hippocampus ...... 114

xvi

TABLE OF ABBREVIATIONS

Abbreviation Term aCSF artificial cerebrospinal fluid AEDs antiepileptic drugs ANOVA analysis of variance ATP/ADP adenosine tri-/diphosphate BBB blood-brain barrier BBG brilliant blue G, a food dye and P2X7 antagonsit BSA bovine serum albumin CA1sr stratum radiatum of hippocampal area Cornu Ammonis 1 C-C motif chemokine ligand 2, a.k.a. CCL2 monocyte chemoattractant protein-1 (MCP-1) cDNA complementary DNA CNS central nervous system CSF1R colony stimulating factor 1 receptor C-X3-C motif chemokine receptor 1, a.k.a CX3CR1 fractalkine receptor ddH2O distilled deionized water DIC differential interference contrast imaging ECS (maximal) electroconvulsive shock ECT electroconvulsive therapy ELISA enzyme-linked immunosorbent assay FJC FluoroJade C GABA γ-aminobutyric acid GFP/eGFP green fluorescent protein/enhanced GFP GPCR(s) G protein-coupled receptor(s) I/V current/voltage ICV intracerebroventricular IFN interferon IH intrahippocampal IL interleukin IN intranasal IP intraperitoneal KA kainic acid LJP liquid junction potential LPS lipopolysaccharides LTP/LTD long-term potentiation/depression

xvii

Abbreviation Term MACS magnetic-activated cell sorting MIP(s) maximal intensity projection(s) MPC(s) microglial process convergence event(s) mTLE mesial TLE NeuN neuronal nuclear antigen Rbfox3 NO nitric oxide P# postnatal day # P2 purinergic ATP-sensitive receptor P2X ionotropic P2 purinoreceptor P2Y metabotropic P2 purinoreceptor P2Z P2X7 receptor a.k.a. cytolytic receptor PBS phosphate-buffered saline Pilo pilocarpine qPCR quantitative polymerase chain reaction RMP resting membrane potential ROI(s) region(s) of interest ROS reactive oxygen species RT room temperature, 18-23 C RT (reaction) reverse transcription of RNA into cDNA s.c. subcutaneous SE Status Epilepticus SNP(s) single nucleotide polymorphism(s) SR101 sulforhodamine 101, an astrocyte-selective fluorescent dye THE tonic hindlimb extension, the endpoint that defines ECS TLE temporal lobe epilepsy Tmem119 transmembrane protein 119, a microglial marker TNF tumor necrosis factor Tt tetanus toxin UTP/UDP uridine tri-/diphosphate Vm membrane/holding voltage

Other commonly used abbreviations: units of time (d: day; h: hour; min: minute; s: second), SI prefixes (p: pico; n: nano; µ: micro; m: milli; k: kilo; M: mega; G: giga) and units (mol: mole; g: gram; L: liter; m: meter; M: molar or mol/L; A: ampere; F: farad; Ω: ohm; V: volt; Hz: hertz or s-1), symbols for dimensions (t: time; x and y: vertical and horizontal; z: depth or height) or for statistics (SD: standard deviation; SEM: standard error about the mean) and other constants and units (g: force of gravity at the surface of the Earth or 9.81m/s-1; Ct: threshold cycle of qPCR; FC: fold change; px: pixel).

xviii

1. INTRODUCTION

Microglia, as the resident immune cells of the central nervous system, are in a privileged position to both affect and be affected by the neuroinflammation and neural injury that accompany many of the epilepsies. Microglia are canonically seen as the first line of response to any sign of trouble in the brain. In various rodent models of prolonged seizures that are associated with extensive neuronal damage (Status

Epilepticus or SE), these specialized glial cells acutely mount robust and complex responses that are characterized by the upregulation of purinergic signaling (Avignone et al., 2008). It remains unclear what, if any, is the role of these microglial changes in the epileptogenic process that ensues after SE. Microglia can mount responses independently to different aspects of the SE “insult” (namely, neuronal hyperactivity and neuronal damage), demanding that responses related to or involved in the epileptogenesis be dissociated from those that are merely coincident. Without this information, the field cannot responsibly move forward in the search for microglial- related treatment targets against (or biomarkers for) epileptogenic mechanisms.

To address this research question, I have employed the mouse maximal electroconvulsive shock (ECS) model, in which short but severe seizures can be reproduced reliably, without neuronal damage or eventual epileptogenesis. Using approaches informed by the extensive body of work onto the SE-induced microglial activation pattern, I have discovered a mild activation pattern in hippocampal microglia following ECS. This response is partly reminiscent of the upregulation in purinergic signaling seen after SE, suggesting that the complex SE-induced activation indeed

- 1 -

comprised a component that was purely a response to the seizure activity. These commonalities could be a “signature” common to self-sustaining seizure activity on hippocampal microglia.

To start this introduction to my dissertation, I will briefly define, review and compare seizures and the epilepsies and discuss potentially epileptogenic sequelae of severe seizure activity. After I review the literature on the role of microglia in epilepsy, I will review what is known about their responses to seizures, with a focus on physiological changes that can acutely impact their interactions with neurons. Lastly, I will explain the importance of filling the gap in our collective knowledge that this project aims to address: how do microglia respond to “benign” seizures? How specific features of microglial activation contribute to subsequent epileptogenesis and how seizure activity, per se, triggers changes in microglial responses is understudied. Through my thesis work, I demonstrate that murine hippocampal microglia can react acutely to single non-epileptogenic seizures, following activation patterns that are partially reminiscent of the SE-induced activation extensively described in the literature. Thus, since they are also associated with ECS-induced seizures, some key features of the microglial activation pattern observed after SE may not be related to the epileptogenic process and further work is needed to fully characterize the interplay between microglia, seizures and epilepsy.

- 2 -

A. Seizures and epilepsy

Seizures are defined as sudden periods of uncontrolled paroxysmal electrical activity in the central nervous system/CNS (Devinsky et al., 2018). To quote directly from the groundbreaking “A study of convulsions” by J. Hughlings Jackson: “A convulsion is but a symptom and implies only that there is an occasional, an excessive and a disorderly discharge of nerve tissue…“ (Jackson, 1870). Others have understood the use of this last adjective as implying incoordination, anatomical disorganization or aperiodicity of the underlying activity (Forcelli, 2011, p. 1). I question this interpretation since the main topic of Jackson’s seminal piece was the orderly progression of motor symptoms in simple partial seizures that we now call “Jacksonian march” in his honor. The fact that he continuously refined his views on the epilepsies (Jackson, 1873, 1874, 1890) drawing from what was cutting edge neurophysiology at the time (like the motor effects of cortical stimulation experiments in dogs by Fritsch and Hitzig, 1870) further leads me to give the father of modern epileptology the benefit of the doubt on this issue. If one assumes that he meant “disorderly” as in the conduct1, or referring to something characteristic of a disorder, this basic description of seizures rings remarkably true to this day. Depending on the brain structures affected, the length and severity of the ictal activity and many genetic and environmental factors, seizures may or may not be associated with acute changes in level of alertness, movement (convulsions), perception or behavior. They also may or may not lead to brain damage and eventual epileptogenesis and cognitive and neurological decline.

1 That is to say, referring to something “insurgent, tumultuous, rough and ungoverned”.

- 3 -

With an incidence of around 50/100,000 persons per year and a prevalence of around 7/1000 persons per year, epilepsy (defined simply as more than one unprovoked seizure, or the predisposition to them) is the 4th most prevalent neurological disease, affecting 1 in 26 people in the US over their lifetime (Hauser et al.,

1991; Banerjee et al., 2009; England et al., 2012). Slightly over 1% of the US population suffers from epilepsy any given year (Strine et al., 2005; England et al., 2012), following a bimodal age distribution with peaks in pediatric (Cui et al., 2015) and geriatric populations (Ip et al., 2018). Seizures themselves are (by definition) even more prevalent: at least one seizure can be expected in up to 1 in 10 people worldwide over their lifetime (Hauser et al., 1991; Benbadis and Hauser, 2000; Lawn et al., 2015; Rizvi et al., 2017).

Similar epidemiologies have been reported historically worldwide for both seizures (Das et al., 2000; Rizvi et al., 2017) and the epilepsies (Olafsson and Hauser, 1999; Fiest et al.,

2017).

i. Treatment approaches against the epilepsies

There are critical and unmet needs in understanding the basic mechanisms underlying the development of epilepsy (epileptogenesis), given that no disease- modifying treatments are currently available in the clinic and that a large proportion of cases do not respond to pharmacotherapy (Löscher et al., 2013), particularly patients suffering from (mesial) temporal lobe epilepsy or (m)TLE.

- 4 -

Following W.R. Gowers’ famous aphorism that “seizures beget seizures” (first quoted after his “Gulstonian Lectures on Epilepsy” in 1880, but more commonly cited from his 1881 book), patients with even relatively mild and infrequent seizures (which are unlikely to be dangerous per se) have historically been treated in an aggressive manner (Reynolds et al., 1983). This is still commonplace despite the mass of tangible evidence against the original claim in the aphorism: “unprovoked1” first and even second seizures do not always associate with increased risk of seizure recurrence (Hauser et al.,

1990; Berg and Shinnar, 1991, 1997; Hauser et al., 1998; Hauser and Lee, 2002). Even more worryingly, clinical trial data do not support the standard approach that results from it: indeed, early and aggressive treatment of new-onset seizures does not decrease seizure burden (Marson et al., 2005) or improve quality of life (Jacoby et al., 2007) in the medium or long term when compared to a “wait and see” approach where treatment for seizures that are not life-threatening (a.k.a. treatment-uncertain unprovoked first or second seizures) is deferred. Treatment options are limited in number and efficacy: primarily pharmacologically with antiepileptic drugs (AEDs) or surgically for resection of the epileptic foci or implantation of electrodes for chronic neurostimulation.

However, up to a third of patients suffering from epilepsy fail to respond to 3 or more lines of AED treatment and are ineligible for surgical treatment (Devinsky et al., 2018).

This zeal to eliminate or minimize seizure burden is at once a cause and a consequence of the significant societal stigma that is placed on patients, which only serves to

1 In most cases, this is defined as those seizures which are not symptomatic of birth complications or developmental problems, or of infection, trauma or stroke; even the first of these “provoked” seizures does associate positively with increased risk of seizure recurrence and mortality (Hesdorffer et al., 1998, 2009).

- 5 -

aggravate the numerous psychiatric problems that are often co-morbid with the epilepsies (Selassie et al., 2014). Making things worse, AEDs are remarkably dirty drugs with high teratogenic potential and many acute and chronic off-target effects on top of the myriad “on-target side effects” that are expected from the unspecific modulation of neurotransmission and excitability that underlies the efficacy of many AEDs (Rogawski and Löscher, 2004; Toman and Goodman, 1948). Particularly during development, exposure to many AEDs can increase apoptotic neurodegeneration (Bittigau et al., 2002) and disrupt synaptic refinement and maturation (Forcelli et al., 2012; Al-Muhtasib et al.,

2018), leading to several psychiatric and behavioral effects (Chen et al., 2017b). This is especially troublesome for women of child-bearing age that suffer from epilepsy, as AED use is not recommended during pregnancy and breastfeeding. This is both because of the effects on neuronal and synaptic maturation mentioned above, but also because of the increased likelihood of significant congenital malformations (Weston et al., 2016) and the potential for hormonal or metabolic changes that undermine the treatment efficacy (Reisinger et al., 2013), for example by affecting drug pharmacokinetics (Leppik and Rask, 1988). These drugs can also have enduring deleterious effects in adult populations with epilepsy, ranging from cosmetic effects (Chen et al., 2015) to psychiatric and behavioral abnormalities (Chen et al., 2017a). These neurological symptoms of drug toxicity are particularly complicated to evaluate in individuals with cognitive disability, but there have been observations of more frequent and severe adverse events associated with AED use in those populations (Sipes et al., 2011;

Copeland et al., 2017). There is clearly a trade-off between these side effects of treatment and the adverse effects of treatment-uncertain seizures and this likely

- 6 -

underlies the lack of long-term benefits from early and aggressive treatment. Moreover, all pharmacotherapies currently approved by the FDA for the treatment of epilepsy were evaluated and even designed on the basis of their efficacy against seizures (Kehne et al., 2017). Only recently has the field focused on specifically targeting treatment- refractory cases (Barker-Haliski et al., 2017; Kienzler‐Norwood et al., 2017) and epilepsies in special populations or cases (like women in Scharfman et al., 2009; or in

D’Amour et al., 2015; viral infection in Stewart et al., 2010; and SUDEP in Kalume et al.,

2013; or in Tupal and Faingold, 2019). Similarly, a large effort is underway to find treatments that target epileptogenesis itself (West et al., 2018). Concurrently, much research has been done on potential therapies for the comorbidities of epilepsy (Slomko et al., 2014; Cho et al., 2015). The epilepsies are the network disorder par excellence, meaning that insights into its pathogenesis could generalize to other psychiatric or cognitive disorders that are also likely arising from network deficiencies. There is an urgent need for better understanding of the mechanisms behind the development of epilepsy to inform and direct these research efforts.

ii. Status epilepticus

Status Epilepticus (or SE) is defined in clinical parlance as a prolonged continuously self-sustaining seizure: historically requiring that the crisis last longer than

30 minutes, since 2015 this has been amended by the International League Against

Epilepsy to 5 minutes (Trinka et al., 2015). SE in humans is a medical emergency that accounts for around 1% of ER visits yearly in the US, but it is also used in the lab to

- 7 -

model seizure-induced epileptogenesis in rodents and other animals. In the prototypical chemoconvulsant-induced SE models (Turski et al., 1983, 1984; Ben-Ari, 1985;

Goodman, 1998; reviewed in Lévesque et al., 2016) of Temporal Lobe Epilepsy (TLE), the experimental subjects are acutely challenged with systemic treatment of pilocarpine or kainic acid/kainate (agonists at muscarinic or ionotropic glutamatergic receptors, respectively) which causes severe and prolonged seizures. After a mostly asymptomatic latent period lasting a couple of weeks, the animals enter a chronic phase where they display interictal spikes and spontaneous seizures. In these well-established models of temporal lobe epilepsy, acute neuronal injury is observed widely through the hippocampus and other brains structures to a greater degree than is seen in clinical cases (Löscher, 2011), which reduces their face validity. The prototypical SE-induced models of epileptogenesis have been further refined by direct delivery of chemoconvulsant intracerebrally to the hippocampal or amygdaloid formations, resulting in more limited neuronal damage but paradoxically also more reliable and treatment-resilient epileptogenesis. These SE models, together with kindling-based models (where repetitive non-convulsive stimuli “kindle” themselves to eventually trigger severe seizures and even epileptogenesis) are of particular interest given their predictive validity for TLE (Löscher, 2002). Interestingly, the initial seizure intensity and the ensuing epileptogenesis are correlated with each other, but not with the degree of injury: while most or all animals treated have neuronal damage, only animals that display extended self-sustained seizure activity will reliably develop chronic epilepsy and related comorbidities (Gröticke et al., 2007, 2008). After seizures, epileptogenesis mechanisms have highly complex bi-directional interactions with changes in brain

- 8 -

structure and function, particularly in the limbic system. On the balance, the available evidence indicates that after SE, the seizure crisis is the critical factor, given that it correlates with epileptogenesis better than the injury does. Likewise, in certain genetic models (such as channelopathies, any of the mutations in ion channels that directly cause enhanced neuronal excitability and trigger seizures among other deficits), neuronal injury and structural changes are apparently secondary to the seizure activity

(Salgueiro-Pereira et al., 2019). This is not always true, as mice harboring disease- causing Scn1b mutations show defects in neuronal patterning during ontogeny before they develop seizures and other symptoms (Brackenbury et al., 2013). These complex causal links between function and structure and between these changes and epilepsy/epileptogenesis are further clouded by the fact that neuronal activity drives differentiation, migration, wiring, network role and fate of neurons, especially during development (Hebb, 1949). Similarly, in other genetic models of inherited epilepsies such as tuberous sclerosis or cortical/sub-cortical heterotopias, basic dysfunctions in neuronal physiology or organization cause epileptiform network activity and imbalanced excitation/inhibition, which in turn causes injury and cognitive and behavioral comorbidities (Martineau et al., 2019). There are thus complex bidirectional relations between structural and functional change during epileptogenesis.

B. Sequelae of damaging seizures have potential links to epileptogenesis

Although in some cases seizures can be beneficial, severe, prolonged and/or frequent seizures can trigger epileptogenesis, as demonstrated by the wide use of SE and

- 9 -

kindling models to study the different processes that may underlie the progression to a chronic “epileptic” phase. Through these processes, seizures can “hijack” mechanisms necessary for normal brain functions for epileptogenic purposes (Avanzini et al., 2014;

Klein et al., 2018).

i. Neural plasticity, neurogenesis and structural changes

Among all the changes observed after SE, atypical cellular excitability and circuit remodeling or potentiation are of primordial importance, as they are the ultimate cause of the paroxysmal activity in eventual recurrent seizures. SE triggers aberrant neurogenesis and results in miswired newly-born neurons (most importantly ectopic hilar cells) that integrate into the network and promote pathogenesis (Cho et al., 2015;

Iyengar et al., 2015; Jain et al., 2019). The excessive mossy fiber sprouting observed after SE (Shibley and Smith, 2002) is similarly believed to enable epileptogenesis by providing input into these new neurons (Pierce et al., 2005). These diverse plasticity mechanisms which normally enable higher-level functions in healthy physiology can be pathologically recruited by the seizure activity itself, but many seizure sequelae have been associated with epileptogenesis as well, yielding new potential targets or biomarkers against epileptogenesis.

- 10 -

ii. Neuronal and blood brain brain barrier damage, metabolic or oxidative stress and inflammation

In adult subjects, experimental SE is associated with widespread neuronal damage, particularly in structures of the limbic system (Turski et al., 1984; Pollard et al.,

1994; Avignone et al., 2008), although the precise relation between the seizure crisis, the neuronal injury and the epileptogenesis remains unclear. For instance, SE can trigger different mechanisms of neuronal cell death to different degrees depending on the precise model used and brain area studied among other factors (Bengzon et al., 2002;

Limbrick et al., 2003; Mikati et al., 2003; Schauwecker, 2003; Kondratyev and Gale,

2004; Fabene et al., 2007; Tokuhara et al., 2007; Schauwecker, 2012).

Through both direct toxicity/off-target effects of the chemoconvulsant and inflammatory signaling secondary to the epileptic crisis and ensuing injury, severe seizure models such as SE cause extensive blood-brain barrier (BBB) damage acutely and milder and more focal increases in BBB permeability chronically (Marchi et al.,

2007; van Vliet et al., 2007; Marchi et al., 2012; Gorter et al., 2015). Sustained, unresolved brain inflammation after seizures leads to persistent BBB damage and promotes seizure recurrence (Librizzi et al., 2012) through leakage of damaging molecules from the peripheral circulation. For example, among others, albumin can increase excitability after neuronal uptake (Noé et al., 2016) and thrombin can activate various signaling mechanisms through cleavage of PAR-1 receptors and components of the extracellular matrix (Isaev et al., 2015).

By their very nature, extended periods of organized neuronal hyperactivity disturb metabolic processes (Van Landingham and Lothman, 1991) and trigger

- 11 -

oxidative stress throughout the brain by producing reactive oxygen species and downregulating antioxidant defenses (Ingvar, 1986; Milatovic et al., 2002; Rettenbeck et al., 2015). These SE-induced changes also contribute to and to a certain degree are affected by, the neuronal injury, the induction of neurogenesis, the changes in BBB integrity and the inflammatory response induced by SE (Bengzon et al., 2002; Limbrick et al., 2003; Mikati et al., 2003; Schauwecker, 2003; Kondratyev and Gale, 2004; Fabene et al., 2007; Tokuhara et al., 2007; Schauwecker, 2012).

Like other insults to the brain (Klein et al., 2018), SE triggers a specific inflammatory response that, while acutely beneficial after isolated damage, oftentimes contributes to the pathogenesis chronically through the recruitment of peripheral immune mechanisms and the ensuing amplification of the primary injury (Avignone et al., 2008; Marchi et al., 2009; Janigro et al., 2013; Kim et al., 2013; Varvel et al., 2016;

Tian et al., 2017; Feng et al., 2019). This response is characterized by a strong transient increase in the expression of the pro-inflammatory mediators TNFα, IL-1β, IL-6 and

CCL2 and the pro-inflammatory prostaglandin-producing enzyme cyclooxygenase-

2/COX-2, particularly in the hippocampus. After SE, these changes peak between 24 and

48 hours and resolve within 5-7 days.

In a similar fashion, endogenous constraining mechanisms act to control and terminate seizure activity as part of a homeostatic response at the synaptic (Postnikova et al., 2017), cellular (Raimondo et al., 2015) and micro- and macro-circuit (Forcelli and

Gale, 2014) level. While seizure onset, propagation and sustenance is correlated with increases in intracellular chloride in principal neurons (Lillis et al., 2012), seizure control and termination is associated with neuronal sodium accumulation (Krishnan

- 12 -

and Bazhenov, 2011). Perhaps from temporal “spillover” of these mechanisms, synapses, neurons and networks stay relatively inactive and hypoexcitable immediately after most seizures. Plasticity mechanisms are also affected: for instance, long-term potentiation

(LTP) is transiently impaired in rat CA1 slices postictally (Barr et al., 1997). On one hand, these seizure-inhibiting processes serve to limit the extent and damage of the epileptic crisis and to lower the vulnerability to another similar insult in the short term.

On the other, their acute effects on the networks that permit normal brain function likely underlie the transient cognitive and behavioral symptoms that characterize the postictal state.

Being neuronal network pathologies, the epilepsies are highly comorbid with various developmental, psychiatric and neurologic disorders (Strine et al., 2005; Selassie et al., 2014). Indeed, all the same mechanisms that can bring about recurrent aberrant paroxysmal activity to cause epilepsy can also reorganize activity in subtler ways and contribute to or cause these comorbidities (Gröticke et al., 2007, 2008; Slomko et al.,

2014; Bozzi et al., 2018; Aguilar et al., 2018; Martineau et al., 2019). Unsurprisingly given the central importance of the hippocampus and other limbic structure to learning, rodent models of TLE were recently linked with specific disruptions to ”what-where- when”/episodic-like memory (Inostroza et al., 2013).

The presentation of these acute sequelae differs greatly between models with similar outcome profiles and thus the exact interplay between them and epileptogenesis has not been fully elucidated. Since the acute disruption of plasticity mechanisms induced by SE after systemic kainate treatment is dependent on the epileptogenesis potential of the SE crisis and not on the seizure intensity or degree of brain damage

- 13 -

(Suárez et al., 2012), it seems that epileptogenesis is a critical component of the puzzle and not just the result of uncontrolled plasticity or responses to injury. As such, putatively epileptogenic processes make for excellent treatment targets against epilepsies and related comorbidities. In an example that perfectly demonstrates this, a recent study was published by the Represa group that shows how in a rat model of

Dcx1-mutant subcortical heterotopia, cognitive defects come after seizures and stopping the epileptogenesis also ameliorated the behavioral comorbidities observed (Martineau et al., 2019). The primordial importance of epileptogenesis is also evidenced by the shift of strategies in the field, as it has become one of the main targets going forward for many therapy research efforts (Löscher, 2011; Kehne et al., 2017; Klein et al., 2018).

C. Microglia: A labile cell type enabled by unique physiology

Since being first described as a distinct non-neuronal brain cell type by Pio del Rio

Hortega (1919; reviewed in Rezaie and Male, 2002; reviewed and translated from

Spanish in Sierra et al., 2016), microglia have long been thought by the field to simply be passive surveyors until their mounted response was needed after infection or injury.

More recently, however, complex and crucial roles have been found for microglia in many developmental (Paolicelli et al., 2011; Schafer et al., 2013), physiological

(Tremblay et al., 2011) and pathological (Salter and Stevens, 2017) brain processes.

Similarly, recent advances in the field have also shown morphological, functional and gene expression differences between microglia from different brain areas. In particular,

De Biase et al. (2017) have shown that microglia from various basal ganglia nuclei are

- 14 -

quite different from their cortical counterparts and among themselves. Furthermore, these regional differences return upon repopulation after pharmacological or pharmacogenetic depletion of microglia, suggesting that they are driven by local cues, perhaps from astrocytes.

i. Roles in development, physiology and pathology

Indeed, in neural development, microglia (which, like hematopoietic cells, derive from the mesoderm and infiltrate CNS tissues early in embryogenesis) contribute to both synaptic pruning and neuronal fate determination during each critical period. By exploiting their macrophage-like cell biology, microglia can use phagocytosis and phagoptosis to engulf and eliminate synapses or whole neurons (Paolicelli et al., 2011;

Sokolowski et al., 2014; Zhan et al., 2014). Moreover, microglia employ the fractalkine system (Mizuno et al., 2003; Cardona et al., 2006; Ragozzino et al., 2006; Corona et al.,

2010; Fuhrmann et al., 2010) and have even repurposed the complement system

(Stevens et al., 2007; Fonseca et al., 2017) to communicate with neurons and to tag elements for sparing or destroying. Additionally, microglia assist normal healthy function through mechanisms such as trogocytosis-induced synaptic remodeling

(Weinhard et al., 2018), synaptic displacement (Chen et al., 2014) and through their complex role interacting with the BBB and with peripheral inflammation. All of these microglial functions can be equally deleterious and contribute to neuronal health and long-term outcomes in disease (Colonna and Butovsky, 2017; Wolf et al., 2017), with recent research highlighting the important complement-mediated role of microglia in

- 15 -

the pathogenesis of Alzheimer’s disease (Hong et al., 2016), viral encephalitis (Vasek et al., 2016) and schizophrenia (Schizophrenia Working Group of the Psychiatric Genomics

Consortium et al., 2016; Wang et al., 2019).

ii. Microglia become “activated” in response to a wide array of insults

Although microglia clearly have roles in normal physiology, they are still best known for their potential to respond to injury or changing conditions. Although outdated (Mittelbronn, 2014), the classic immunological view of two poles of microglial

”activation” is still useful to introduce the possible consequences of activation and to classify other response patterns. Today, it is widely accepted that different insults or combinations of insults under different conditions each cause distinct microglial activation patterns (Sousa et al., 2018; Merienne et al., 2019).

These “microglial signatures” in response to environmental changes, specific diseases or disruptions to homeostasis have recently come into vogue in the search for neuroinflammation-related/microglial treatment targets or biomarkers for epilepsy

(Butovsky and Weiner, 2018).

Canonically, like for other macrophages (Shapouri‐Moghaddam et al., 2018), microglial activation was described along a linear spectrum with two opposing poles termed M1 and M2 (Orihuela et al., 2016; Tang and Le, 2016). In immunology, the M1 pole is associated with “classical” activation (prototypically by IFN-γ or TNF-α, or by

LPS, a component of Gram-negative bacterial cell walls) and characterized by the expression of iNOS and/or COX-2 and the production of oxidative ROS and of

- 16 -

traditionally pro-inflammatory signaling molecules like IL-1β, IL-6, TNF-α, CCL2, IFN-γ and CD16/32. On the opposite end of the response spectrum, “alternative” activation

(prototypically by IL-4 or IL-13, by immunoglobulins or by glucocorticoids) polarizes microglia to the M2 state which is characterized by the expression of Arg1 and/or PPAR-

γ and the release of neurotrophic factors and anti-inflammatory molecules such as IL-4,

IL-10, IL-13, TGF-β, CD206, Fizz1, Ym1 and IL-1ra. These microglial activation states

(reviewed in Lynch, 2009) are also self-sustaining to a certain degree, as microglia can act as autocrine targets of their own pro- or anti-inflammatory signaling.

In a manner reminiscent of the well-studied “respiratory burst” (Iles and Forman,

2002; Thomas, 2017) conserved in many immune-like responses across clades (Keppler et al., 1989; Bradley et al., 1992; Jabs et al., 1997; Dahlgren and Karlsson, 1999), microglia exploit their own cellular metabolism to shape their responses through bioenergetic regulatory mechanisms. Like most other mammalian cell types, microglia can utilize both oxidative and glycolytic pathways, but they seemingly must simultaneously increase aerobic glycolysis and decrease respiration when undergoing activation by several stimuli (Ghosh et al., 2018). As such, bioavailability of substrate

(mostly glucose), rate of glycolysis and function of mitochondrial processes can all influence the ensuing inflammatory signaling through a wide spectrum of both transcriptional and post-translational regulatory mechanisms. While these metabolic signals certainly exert control on physiological functions of microglia, they are especially remarkable for their roles in pathology. Particularly, glucose metabolism directly impacts superoxide formation by NADPH oxidase since glucose is the required substrate to reform NADPH in the hexose monophosphate shunt/pentose phosphate pathway

- 17 -

(Kruger and von Schaewen, 2003). Interestingly, microglia can also metabolize ketones bodies like β-hydroxybutyrate. They also respond to these ketones as metabolic signals:

β-hydroxybutyrate has been shown to inhibit histone de-acetylases and activate microglial receptors like GRP109A receptors. The net effect of this is acute suppression of microglial activation and shifting of the response phenotype to a more neuroprotective one. This is particularly enticing since it is widely accepted that ketogenic diets and fasting greatly ameliorate many epilepsies dependent on production of ketones, which could suggest a potential microglial role in this clinically effective

“treatment”.

As an isolated population of neuroimmune non-ectodermal cells, microglia rely on tightly regulated self-renewal through mitotic and apoptotic processes to maintain their numbers in check during healthy function (Wirenfeldt et al., 2007; Réu et al., 2017).

Another important aspect of “activation” in general is that it is often associated with microglial proliferation, which is thought to happen through dysfunction of their normal self-limiting renewal cycle and excessive clonal expansion of activated populations (Tay et al., 2017).

iii. Morphology and motility: “Scouting” microglia from box scores vs. game film

Together with the induction of marker expression and cytokine release, changes in microglial morphology have consistently been associated with activation. In particular, increases in phagocytic and even more so in secretory activity require simplification of microglia, with activated cells losing their complex ramified profile to

- 18 -

eventually adopt an amoeboid, macrophage-like morphology (Lynch, 2009). Given that application of apyrase (an ATP-hydrolyzing enzyme) results in altered microglial morphology and motility, it was long thought that these behaviors were regulated by ambient ATP among other signals (Madry and Attwell, 2015). More recently, however, it has become evident that apyrase application resulted in depolarization (due to increased [K+]o from the apyrase buffer). The ensuing microglial changes reflect the effect of this depolarization and not of ATP depletion, since dialyzed apyrase did not elicit a similar response (Madry et al., 2018). Together with the proliferation/accumulation, this morphological change is a classic hallmark of

“activation” and indeed counting and classifying stained microglia in brain sections has long been the first line of detection for microglial activation. Historically, microglia were identified visually after non-selective staining by traditional silver methods. Today, it is more common to stain in a more selective fashion: histochemically with the NDPase reaction or immuno-histochemically against marker molecules such as CD11b (which marks all macrophages) and Iba1 (which is problematic for quantitative morphometry given its modulation by many insults). Increasingly, more specific markers like the colony-stimulating factor 1 receptor CSF1R (Elmore et al., 2014), the fractalkine receptor CX3CR1 (Nishiyori et al., 1998; Boddeke et al., 1999; Hughes et al., 2002), the metabotropic purinergic receptor P2Y12 (Sasaki et al., 2003) and the transmembrane protein Tmem119 (Satoh et al., 2016) are used to stain, image and reconstruct microglia and measure their morphologies quantitatively. More recently, the advent of transgenic reporter mouse lines expressing fluorescent proteins under the control of markers like

CX3CR1, Iba1 and CSF1R has kickstarted functional studies into microglial dynamics and

- 19 -

motility in vivo and in situ through acute slices (Hirasawa et al., 2005; Jung et al., 2000;

Sasmono, 2003). The constantly changing motility/dynamic morphology of microglia is enabled by their complex physiology, with various receptors, ion channels and signaling mechanisms playing a role (Madry and Attwell, 2015).

Despite regional, temporal and individual variability even in the healthy CNS

(Lawson et al., 1990), microglial distribution at rest is remarkably uniform and most

“resting” cells occupy an exclusive territory that they constantly "patrol" by coordinated extending and retracting of branching processes (Nimmerjahn et al., 2005). This rapid process motility behavior underlies the microglia-neuron contact that has many established roles in neural plasticity (Tremblay et al., 2010; Weinhard et al., 2018).

Similarly, in a manner reminiscent of their physiological role in development, microglial motility also has roles in determination of neuronal cell fate (Hines et al., 2009; Wake et al., 2009). Microglial process motility is primarily guided by chemoattraction towards

ATP/ADP (Haynes et al., 2006; Wu et al., 2007; Irino et al., 2008; Ohsawa et al., 2010,

2012) or chemorepellence away from adenosine (Orr et al., 2009). The extension elicited by ATP is modulated by NO (Duan et al., 2009), particularly in white matter and the spinal cord (Dibaj et al., 2010). This irregular and seemingly stochastic motility under physiological conditions becomes organized in response to injury (Ohsawa and Kohsaka,

2011; Sokolowski et al., 2014), with ATP from damaged neurons recruiting processes from neighboring microglia (Davalos et al., 2005). In the last decade, these coordinated process motility behaviors have been described and measured in in vivo and in slice assays of microglial physiology (Avignone et al., 2008; Eyo et al., 2014, 2015, 2016). Of particular interest are the velocity and extent of the directional microglial process

- 20 -

motility in response to a laser injury or local ATP source, as well as the frequency and dynamics of spontaneous Microglial Process Convergence events (MPCs) that rely on the same signaling pathways. The accepted mechanism for this response is that local ATP gradients activate P2Y12 receptors on microglial processes, triggering the changes in voltage-dependent currents, integrin activation and actin polymerization required for directional process extension (Wu et al., 2007; Ohsawa et al., 2010). Interestingly, this same receptor and similar downstream signaling also mediate slower microglial cell translocation over longer distances (Eyo et al., 2018b). Other related channels, stores and pores mediate membrane rearrangement by altering the osmotic pressure and thus the volume of different sections. While MPCs in slices (whose frequency is increased by excess glutamate and reductions in extracellular [Ca2+]) are also dependent on ATP, it seems that these complex coordinated responses are following other chemical gradients as well, as the process velocity in these extensions is greater than that achieved by ATP by itself, at any concentration (Eyo et al., 2015). As microglial “activation” has normally been associated with morphological simplification, enhanced process motility behaviors have long been considered exclusively in the context of development, physiology or plasticity, not pathology. As I review later in this introduction, SE and related insults have been shown to increase the velocity of hippocampal microglial processes responding to simulated injury (Avignone et al., 2008), the area “monitored” by a single cell (Avignone et al., 2015) and the frequency of MPCs (Eyo et al., 2016). In stark comparison to most activation states, the brain-wide acute microglial response to SE couples some traditional facets of “activation” with preserved surveillance motility behavior (Avignone et al., 2015).

- 21 -

iv. Physiology of microglia

Underlying their wide array of actions and roles in development, health and disease, microglia have a unique physiology and pharmacology among brain cells, with the potential to express a variety of receptors for neurotransmitters, neuromodulators, biomolecules and “danger” signals. Their complex physiology has been authoritatively reviewed by Kettenmann, Hanisch, Noda and Verkhratsky (Kettenmann et al., 2011), but it remains unclear the degree to which findings from in vitro cell culture experiments can be extended to microglia in vivo, since they are particularly severely affected by immortalization into stable cell lines, or by dissection, dissociation and expansion before primary cell culture experiments. For these and other reasons (including phagocytosis of other cellular elements), a bewildering proportion of the channelome and the sensome has been detected in microglial cell lines, primary microglia and even acutely isolated or labeled microglia.

Below, I will introduce the relevant receptors and ion channels that are most widely accepted to be expressed in microglia, as well as their membrane properties. As reviewed in exhaustive detail by Kettenman et al. (2011), microglia can possess functional receptors for:

 purines like adenosine (A1, A2A, A2B, A3) and ATP (P2 receptors, reviewed

in more detail in the following paragraph);

 small molecule neurotransmitters such as glutamate (α-amino-3-hydroxy-

5-methyl-4-isoxazolepropionic acid/AMPA receptors: GluA1, GluA2, GluA3

and GluA4 subunits; and metabotropic glutamate receptors from group I:

- 22 -

mGluR5a, group II: mGluR2 and mGluR3; group III: mGluR4, mGluR6 and

mGluR8) and γ-aminobutyric acid/GABA (like the metabotropic receptors

GABAB1a, GABAB1b, GABAB1c);

 classic neuromodulators like acetylcholine (ionotropic nicotinic

receptors/nAChRs, particularly α7 but also α3, α5, α6 and β4),

norepinephrine (α1A, α2A, β1 and β2) and dopamine (D1 and D2);

 neurohormones and other neuromodulatory molecules like platelet-

activating factor/PAF, bradykinin, histamine, endothelin, cannabinoids,

angiotensin II, somatostatin, gluco- and mineralocorticoids, opioids,

neurokinin/substance P, vasoactive intestinal peptide/VIP and

neurotrophins;

 cytokines and chemokines (such as CCR1, CCR2, CCR3, CCR5, CXCR1,

CXCR3, CX3CR1, TNFR1 and TNFR2);

 interferons and interleukins (like IFNγR, IFNAR, IL-R1/IL-1R2, IL-2R/IL-

15R, IL-4R, IL-10R, IL-13R and IL-18R);

 PAMPs or DAMPs, pathogen- or danger-associated molecular patterns like

LPS and unmethylated CpG in DNA (such as the co-receptor CD14 and the

toll-like receptors TLR1/2, TLR6/2, TLR3, TLR4, TLR5, TLR7/8 or TLR9);

 and, cellular components of the complement system (C3aR, C5aR/CD88).

Microglia also have a biogenic electrical membrane potential (further discussed in the next paragraph) and they express a wide variety of other passive or gated ion channels, including: sodium channels (Nav1.1, Nav1.5, Nav1.6), calcium channels

- 23 -

(ORAI1/CRAC), potassium channels (inward rectifier Kir2.1; delayed rectifiers Kv1.1,

Kv1.2, Kv1.3 and Kv1.5; Ca2+-dependent BK/KCa1.1; small-conductance SK1, SK2, SK3 and

SK4; and, Gi protein-activated GIRK), anion channels (volume-regulated Cl-, Best, CLIC-

1), proton channels (Hv1), water channels/aquaporins (like the prototypically astrocytic

AQP4) and connexins (Cx32, Cx36, Cx43 and Cx45). Of all these ion channels, Kv1.3 are perhaps the most powerful regulators of microglial function, as they contribute to killing neurons through a mechanism that involves the post-activation respiratory burst but not necessarily the resulting increased NO production (Fordyce et al., 2005). Critically, these channels interact with a series of cotransporters, pumps and calcium adaptors and stores to maintain ionic and osmotic homeostasis during basal and responsive motility and other morphological rearrangements. Also notable is the role they play during periods of intense metabolic and physiological stress from increased secretory, phagocytic or proliferative activity, when these same mechanisms must respond to changes in osmotic pressure and ionic concentrations (particularly Ca2+) to avoid irreversible cellular damage.

At “rest” in situ or in vivo, healthy ramified hippocampal microglia display a bimodally distributed resting membrane potential (with peaks around 52 and 23 mV) and their membrane has a very high input resistance and a shallow and mostly linear current/voltage (I/V) relation, with relatively small currents and having no rectification

(Boucsein et al., 2003). Acute depolarization of microglia results in simplification of their morphology and reduced extent of surveillance motility (Madry et al., 2018). On the other hand, cultured microglia at rest possess slightly more negative resting membrane

- 24 -

potentials (bimodal distribution with peaks at 61 and 29 mV) and much larger currents with strong inward rectification (steep, non-linear I/V). These culture-induced changes in membrane physiology are reminiscent of microglia acutely activated by injury, or of amoeboid microglia in situ. Thus, these are likely artifacts from the dissociation, isolation or expansion of cells required for in vitro studies.

v. P2 purinoreceptors underlie many basic functions of microglia

P2 purinergic receptors underlie the response of microglia to ATP and other bioactive purines like ADP, UTP and UDP (Boucsein et al., 2003). This complex system of metabotropic and ionotropic signaling regulates many vital microglial functions at rest

(Inoue, 2008) and also has complex and well-established roles in the transition to different modes of functional activation (Koizumi et al., 2013).

Metabotropic purinoreceptors belong to the P2Y family which includes eight identified members (P2Y1, P2Y2, P2Y4, P2Y6, P2Y11, P2Y12, P2Y13 and P2Y14). Like other archetypal class A G protein-coupled receptors (GPCRs), P2Ys comprise seven hydrophobic transmembrane domains with an intracellular C-terminus and an extracellular N-terminus. Similar to other GPCRs, P2Ys can form homo- or hetero-dimers among themselves and with other receptor families, with consequences in pharmacology and downstream signaling bias. However, only the simple or homodimerized receptors have been extensively studied to date in homologous expression systems or in vivo. Microglia are thought to express P2Y2, P2Y6, P2Y12 and

P2Y13. Colloquially known as the “eat-me” receptor, P2Y6 is preferentially activated by

- 25 -

UDP and holds particular importance for its role in mediating organized microglial phagocytosis (Koizumi et al., 2007; Neher et al., 2014). The microglial-endemic P2Y12 receptor (sometimes termed “find-me”) is preferentially activated by ATP/ADP and famously serves to guide microglial processes towards ATP released from an injury

(Davalos et al., 2005). This extension is triggered through changes in actin dynamics regulated by activation of the Akt (Irino et al., 2008) and integrin beta-1/CD29 (Ohsawa et al., 2010) signaling pathways and induction of an outward potassium-selective current (Wu et al., 2007). Changes in P2Y receptor expression and function have been observed in animal models of TLE as well as in surgical or autopsy brain samples from patients with drug-refractory epilepsies (Alves et al., 2017).

Ionotropic purinoreceptors are homo- or heterotrimeric and belong to the P2X family (part of the two-transmembrane domain containing ENaC superfamily), of which seven separate members have been identified: P2X1, P2X2, P2X3, P2X4, P2X5, P2X6 and

P2X7 (Khakh et al., 2001; North, 2002). This family of receptors has been tied into the interplay between excitability and neuroinflammation after epileptogenic insults

(Henshall and Engel, 2015). P2X1, P2X4 and P2X7 have been consistently associated with microglia. P2X1 is expressed preferentially by embryonic microglia with only low levels detected in cells from adult brains (Xiang and Burnstock, 2005). The P2X4 receptor is known for being fast and sensitive to ATP relative to other members of the family

(Stokes et al., 2017; Suurväli et al., 2017). Increased P2X4 transcript and protein expression has been detected after SE and under systemic inflammation conditions

(Avignone et al., 2008; Ulmann et al., 2013; Xu et al., 2016). P2X4 in spinal microglia is thought to be specially relevant in the pathogenesis on neuropathic pain after nerve

- 26 -

injury (reviewed in Inoue, 2019). This is interesting since there are many common mechanisms underlying both epilepsy and neuropathic pain, which is why new AED candidates are often co-opted for testing of potential effects against animal models of neuropathic pain as part of screening at the NINDS-funded, University of Utah-based

Epilepsy Therapy Screening Program (or ETSP, reviewed in Kehne et al., 2017).

P2X receptors assemble into trimeric channels and heterotrimers may possess emergent pharmacological or biophysical properties (Saul et al., 2013). P2X7 was for a long time thought to be the only member of the family to not form heterotrimers

(Dubyak, 2007), but recent evidence suggests that functional receptors containing both

X4 and X7 can form after heterologous expression and exist endogenously in homologous systems like macrophages (Guo et al., 2007).

P2X7 and P2Y12 receptors are growing in popularity for imaging microglia and their “activation” in humans through radioactive ligands and PET (Janssen et al., 2014;

Beaino et al., 2017; Janssen et al., 2018), quickly rising to challenge the previous consensus marker: translocator protein 18 kDA (TSPO or peripheral benzodiazepine receptor), the use of which is troublesome given that it fails to bind to the protein in many humans due to a single polymorphism. Similarly, positive preliminary results have been recently reported that support the use of radioligands selective to CSF1R to image microglia in a clinical setting (Horti et al., 2019). Going forward, it is clear that the field will move away from attempting to image “microglial activation” and towards more specific, relevant and mechanistic questions, like attempting to image microglial phagocytic, secretory or proliferative activity, microglial purinergic binding or microglial sensitivity to other neurotransmitters and cytokines (Tronel et al., 2017).

- 27 -

vi. Focusing on P2X7R as a key mediator of microglial function in health and disease

The mammalian P2X7 receptor is widely expressed in microglia and macrophages

(Nuttle et al., 1993; Falzoni et al., 1995; Ferrari et al., 1996; Donnelly-Roberts et al.,

2009) and possesses unique and fascinating biophysical properties among the members of the P2X family (Khakh et al., 2001; Jarvis and Khakh, 2009; Bhattacharya and Biber,

2016), to the degree that they were first identified functionally as a separate P2 subfamily termed P2Z (Buisman et al., 1988; Surprenant et al., 1996).

These receptors normally conduct cations non-selectively and although they are somewhat calcium-permeable under physiological conditions, extracellular divalent cations greatly inhibit the amplitude of response current by allosterically “clogging” the channel pore (Michel et al., 1999; Yan et al., 2011). Thus, currents are greatly amplified in divalent cation-free media. Together with selective pharmacological agents like the potent agonist BzATP (Young et al., 2007), the reversible allosteric antagonist textile dye

Brilliant Blue G (BBG) in low micromolar concentrations (Jiang et al., 2000) or the irreversible competitive antagonist oxidized ATP at high micromolar concentrations

(Murgia et al., 1993), this biophysical signature is routinely exploited for unmasking and identification of P2X7 receptors in patch-clamp experiments (Avignone et al., 2008;

Sepulveda-Rodriguez et al., 2019).

These receptors also possess multiple extra- and intracellular protein- and lipid- binding sites (including one with affinity for bacterial LPS and roles in inflammation) which could mediate metabotropic functions (Denlinger et al., 2001). Unlike most other known receptors, continuous ligand binding results in reversible sensitization instead of

- 28 -

a dulled response (Rassendren et al., 1997). Interestingly, this potentiation of activity occurs at the single-receptor level and does not require recruitment of additional

P2X7Rs: individual P2X7 increase their permeability to eventually form a large macropore (Virgilio et al., 2018). This pore becomes large enough to readily conduct not only anions like Cl- and divalent cations like Ca2+ which regulates intracellular signaling cascades (Schachter et al., 2008; Liang et al., 2015), but also larger biomolecules like large organic ions including NMDG or nucleotides including ATP (Brandao-Burch et al.,

2012) and even nanometer-scale macromolecules like the dye YOPRO-1 (Browne et al.,

2013). Multiple non-exclusive mechanisms have been proposed to underlie this transient non-selectivity and increase in permeability, including pore dilation dependent on the composition or signaling of the 2nd transmembrane domain (Sun et al., 2013) or of the intracellular domain/carboxyl tail (Becker et al., 2008; Wickert et al., 2013) or on the cysteine palmitoylation of plasma membrane lipids (Karasawa et al., 2017). Other potentially receptor-autonomous mechanisms have been suggested for macropore formation: through recruitment of accessory proteins that modify the channel structure and/or Pannexin1 hemichannels (Bao et al., 2004; Locovei et al., 2007; Schachter et al.,

2008; Shoji et al., 2014). Recently, the gate and selectivity filter of the P2X7R has been positively identified as the S342 residue in the 2nd transmembrane helix (Pippel et al.,

2017). This same study provided some evidence that suggests that the loss of selectivity upon extended agonist binding was in fact not due to pore dilation but rather a “striking susceptibility to remain in the open state for longer when the channel pore contains slowly or nonpermeating cations”. Given the importance of the C-terminal intracellular tail for the sensitization behavior and for other quasi-metabotropic intracellular

- 29 -

signaling, it is important to note that P2X7 receptors from different species have different C-terminal domain sequences which likely explain their different pharmacology and biophysics (Donnelly-Roberts et al., 2009; Wickert et al., 2013). Most critically, a spontaneous P451L mutation in the cytoplasmic tail of the receptor has been found in many inbred mouse strains and it impairs the formation of the macropore structure to some degree (Adriouch et al., 2002).

The increased permeability of P2X7R upon prolonged activation has well- established roles in cytotoxicity (Di Virgilio et al., 1990; Ferrari et al., 1997) and has been termed a “cytolytic channel” by itself (Surprenant et al., 1996; Virgilio et al., 1998). or “death complex” when cooperating with other molecules like Pannexin1 (Pelegrin and Surprenant, 2006; Locovei et al., 2007). P2X7 has been shown to initiate processing of the NLRP3 inflammasome (Franceschini et al., 2015; Karmakar et al., 2016) to coordinate the production and release of pro-inflammatory mediators like IL-1β (Ferrari et al., 2006). In a similar fashion, activation of the receptor in several immune cells including microglia also leads to increased ROS production and release (Harada et al.,

2003; Mead et al., 2012; Bartlett et al., 2013; Wang and Sluyter, 2013).

Interestingly, in systems with either heterologous or endogenous expression of the receptors, P2X7R and particularly its roles in inflammation are negatively regulated by P2X4R activity (Kawano et al., 2012). Possible mechanisms accounting for this effect could be agonist competition due to the exquisite sensitivity of P2X4R (Suurväli et al.,

2017) or perhaps direct inhibition of the necessary ROS production from their increased calcium permeability (Scherz‐Shouval et al., 2007). Another key regulator of P2X7R

- 30 -

expression and function is the transcription factor Specificity protein 1 or Sp1 (García-

Huerta et al., 2012).

Since they can release ATP through the pore upon sensitization, P2X7Rs are paradoxically also critical regulators of the extracellular levels of ATP, their own endogenous cognate agonist (Brandao-Burch et al., 2012; Amores-Iniesta et al., 2017).

Indeed, it is thought to be an important contributor to the ATP release into the extracellular medium that is observed after ictal activity in general (Wieraszko et al.,

1989; Abiega et al., 2016; Beamer et al., 2019). Besides inflammation, other traditional aspects of “activation” responses like proliferation and morphological changes are dependent on P2X7-induced macropore formation, revealing a contradictory trophic role of the pore conductance at endogenous levels of ligand binding (Monif et al., 2009).

As is to be expected given all the above, P2X7 has been associated with many neurological diseases and particularly with inflammatory processes underlying neurodegeneration (Sperlágh and Illes, 2014; Bhattacharya and Biber, 2016). Recently, the potential role of this channel and its many single nucleotide polymorphisms (SNPs) in diseases as diverse as depression (Deussing and Arzt, 2018) and cancer (Burnstock and Knight, 2018) have rekindled interest in it as a treatment target. The labs of David

Henshall and Tobias Engel at RCSI among others have found increased expression of this receptor in TLE patients and animal models (Jimenez‐Pacheco et al., 2013).

Transient pharmacological antagonism of P2X7 during the acute or latent phases of SE results in reduction of adverse outcomes in the chronic phase, so this approach has been suggested as a novel therapy against epileptogenesis (Engel et al., 2012; Jimenez‐

Pacheco et al., 2013; Fischer et al., 2016; Jimenez-Pacheco et al., 2016; Beamer et al.,

- 31 -

2017). Besides this long-lasting relief from epileptogenesis, P2X7 antagonists also potentiate the acute anticonvulsant efficacy of AEDs like carbamazepine, despite having no effect by themselves on the seizure models tested (Fischer et al., 2016; Nieoczym et al., 2017). Because they are BBB permeable and safe, the potent and selective P2X7 antagonists AFC-5128 and JNJ-47965567 make particularly good drug candidates. BBG is also non-toxic and efficiently enters the CNS, but it has significant cosmetic side effects

(blue coloring of membranes, skin and sclera) which dampen enthusiasm for its use.

Another promising approach is through the use of specific nanobodies (high-affinity single-domain camelid antibodies) which can be systemically injected and combine the specificity of immunologics with the potency of small drug inhibitors (Danquah et al.,

2016). Before these pore-blocking nanobodies were developed, antibodies tested consistently failed to prevent channel opening and pore formation despite binding to the receptor.

D. Microglial responses to seizures and potential roles in epilepsy

As the first line of immune defense in the CNS, microglia can react acutely to a wide spectrum of external stimuli, including but not limited to the hyperactivity, injury and other sequelae of SE. As briefly summarized in Table 1 below, multiple studies by independent laboratories have proven that microglia in hippocampus mount a mostly pro-inflammatory complex response to epileptogenic insults in acquired models of epilepsy (Shaw et al., 1990; Akiyama et al., 1994; Chen et al., 2005; Penkowa et al., 2005;

Avignone et al., 2008; Shapiro et al., 2008; Foresti et al., 2009; Menteyne et al., 2009;

- 32 -

Ulmann et al., 2013; Eyo et al., 2014; Arisi et al., 2015; Avignone et al., 2015; Patterson et al., 2015; Rettenbeck et al., 2015; Eyo et al., 2016; Sabilallah et al., 2016; Tian et al.,

2017; Wyatt-Johnson et al., 2017; Bosco et al., 2018; Eyo et al., 2018a; Kalozoumi et al.,

2018; Klement et al., 2018; Feng et al., 2019).

The particular activation state in hippocampal microglia after SE (characterized by enhanced purinergic signaling) was first described in a seminal study from Etienne

Audinat’s laboratory (Avignone et al., 2008). Since then, the Audinat laboratory and other groups (most notably Long-jun Wu’s teams at Rutgers U then the Mayo Clinic) have delved into other features of the response (Menteyne et al., 2009; Ulmann et al.,

2013; Avignone et al., 2015; Rettenbeck et al., 2015; Schartz et al., 2016; Wyatt-Johnson et al., 2017) and also the mechanisms that likely underlie this “particular” activation

(Foresti et al., 2009; Eyo et al., 2014, 2016; Tian et al., 2017; Bosco et al., 2018; Eyo et al.,

2018a; Feng et al., 2019).

Furthermore, a recent study where Tsc1 (a gene in which mutation is associated with tuberous sclerosis complex) was selectively knocked out in microglia suggests that noninflammatory microglial changes could be the main driver of epileptogenesis in this novel mouse model of tuberous sclerosis (Zhao et al., 2018). A different group, however, had previously described microglial activation (Zhang et al., 2016) and epileptogenesis

(Zou et al., 2017) after neuronal- and astrocytic-selective mutations and has raised concerns regarding the specificity of the genetic tools used to target microglia (Zhang et al., 2018).

- 33 -

i. A survey of studies on acute microglial responses to status epilepticus

Table 1: Previous studies on status epilepticus-induced microglial activation Reference SE model Relevant post-SE findings morphological activation of microglia, Shaw et al. (1990) rat/Tt, IH neuronal cell loss microglial activation, complement Akiyama et al. (1994) rat/KA, ICV system recruitment, inflammation neurodegeneration, microglial activation Chen et al. (2005) mouse/KA, IN and astrogliosis enhanced purinergic signaling in Avignone et al. (2008) mouse/KA, IP microglia neurodegeneration, microglial activation Shapiro et al. (2008) rat/Pilo, IP and astrogliosis increased expression of CCL2 and Foresti et al. (2009) rat/Pilo, IP CCR2 proteins Menteyne et al. (2009) mouse/KA, IP enhanced Kv1.3 expression in microglia

Ulmann et al. (2013) mouse/KA, IP P2X4 receptors in activated microglia NMDA hyperactivity recruits microglial Eyo et al. (2014) mouse/KA, IP processes through P2Y12 receptor Avignone et al. (2015) mouse/KA, IP altered morphological dynamics Arisi et al. (2015) Rat/Pilo, IP Increased CCL2, CCL3, CCL5 and IL-1β inflammatory response correlates with Patterson et al. (2015) juvenile rat/febrile epileptogenesis Rettenbeck et al. (2015) rat/electrical increased microglial ROS production mouse/Pilo+KA, fractalkine signaling regulates microglial- Eyo et al. (2016) IP neuronal contact pro-inflammatory changes and their Sabilallah et al. (2016) mouse/KA, IN correlation with epileptogenesis CCL2 and IL-1β signaling promote Tian et al. (2017) mouse/KA, IP neuronal cell death Wyatt-Johnson et al. microglial proliferation and rat/Pilo, IP (2017) morphological changes Bosco et al. (2018) mouse/KA, IP RNAseq of hippocampal microglia glial and inflammatory changes during Kalozoumi et al. (2018) mouse/KA, IH epileptogenesis pericytosis and pericyte-microglia Klement et al. (2018) mouse/KA, IH clustering in cerebrovasculature microglial and peripheral immune Feng et al. (2019) mouse/KA, IP system contribution to gliosis Table Abbreviations: Tt: tetanus toxin; KA: kainic acid; Pilo: pilocarpine; febrile: induced hyperthermia; electrical: self-sustaining SE after amygdala stimulation; IH: intrahippocampal; ICV: intracerebroventricular; IP: intraperitoneal; IN: intranasal.

- 34 -

ii. Experimental status epilepticus induces a particular activation state in murine hippocampal microglia

Much like other epileptogenic insults to the brain (Klein et al., 2018), status epilepticus (SE) induces a wide spectrum of changes in pro- and anti-inflammatory cytokine expression and is accompanied by microglial proliferation and morphological/physiological activation (Avignone et al., 2008, 2015).

Notably, SE rapidly triggers enhanced purinergic signaling in microglia which correlates positively with increased velocity of microglial process motility, following a temporal and anatomical pattern that suggests relevance to the induced neuroinflammation and the eventual epileptogenesis. Previously described increases in expression of purinergic receptor or potassium channel transcripts were correlated with increased density of ATP- and voltage-gated currents respectively. The purinergic receptors P2Y12 (Eyo et al., 2014), P2X4 (Ulmann et al., 2013) and P2X7 (Engel et al.,

2012; Jimenez‐Pacheco et al., 2013; Jimenez-Pacheco et al., 2016) and the voltage-gated delayed rectifier potassium channel Kv1.3 (Menteyne et al., 2009) are critically involved in the particular microglial activation pattern after SE. Strikingly, despite unchanged process velocity during basal motility after SE, hippocampal microglia cover more area during surveillance and extend their processes faster towards a point source of ATP or a focal laser injury (Avignone et al., 2008, 2015).

Concomitantly, microglia populate the hippocampal pyramidal cell layer with higher density and increase the frequency of MPCs and of individual contact with neurons (Eyo et al., 2014, 2016, 2018a). Infiltrating monocytes have been shown to

- 35 -

engage in a feed-forward loop with neuroinflammation that contributes to the microgliosis (Feng et al., 2019) and exacerbates the neuronal injury (Varvel et al., 2016).

This purinergic-centered response originally described by Avignone and colleagues (2008) is also accompanied by other related changes that draw from more traditional elements of activation: simplification in microglial morphology, increased proliferation (detectable through greater cell density but also through immunoreactivity against the proliferation markers Ki-67 and MAC2), upregulated pro-inflammatory signaling (particularly TNFα, IL-1β, IL-6 and CCL2) and increased expression of markers of classic activation (Iba1 and F4/80).

Increased TNFα signaling is nearly ubiquitous in inflammatory states and effectively gives glia and the immune system control over synaptic strength (Beattie et al., 2002). It is known to regulate neurotransmission and initiate excitotoxicity through increased trafficking of AMPA receptors to the synapse (Leonoudakis et al., 2004) and even selective recruitment of Ca2+-permeable receptors lacking the GluA2 subunit

(Stellwagen et al., 2005). Additionally, a subpopulation of cells from slices treated with

LPS displayed increased Ca2+ currents after stimulation of AMPA receptors likely resulting from effects of proinflammatory signaling, canonically through TNFα (Winland et al., 2017). This is thought to happen both at the level of the cell (through changes to voltage activated Ca2+ channels) and of the individual receptor (by regulating GluA2 content/S845 phosphorylation). The finding of increased CCL2 and IL-1β levels post-SE are particularly interesting given the association of these cytokines with the interplay between inflammation and the epilepsies (Vezzani and Baram, 2007; Vezzani et al.,

2011). Both of these molecules have established roles in proconvulsive (Chiavegato et

- 36 -

al., 2014; Cerri et al., 2016) and epileptogenic mechanisms (Plata-Salamán et al., 2000;

Tian et al., 2017). Increased brain expression of the canonical receptors for these molecules has been reported after SE and is also thought to contribute to the ensuing pathology (Ravizza and Vezzani, 2006; Foresti et al., 2009). CCR2 is putatively found in patrolling monocytes (giving the cognate ligand CCL2 its other name, monocyte chemoattractant protein 1 or MCP-1) and signaling through this pathway contributes to monocyte infiltration, neuroinflammation, microgliosis and neuronal cell death after SE

(Foresti et al., 2009; Tian et al., 2017; Varvel et al., 2016; Feng et al., 2019). Interestingly, one of the consequences of SE-induced CCL2/CCR2 signaling is the increased production and secretion of IL-1β from microglia and eventually also monocytes (Tian et al., 2017).

Its receptor IL-1R is located mostly in neurons in the naïve CNS, but neuronal expression is increased and expression is also widely found in astrocytes after SE (Ravizza and

Vezzani, 2006). Over activation of this pathway in neurons leads to neuronal hyperexcitability and subsequent excitotoxic damage-induced cell death (Vezzani and

Viviani, 2015; Eyo et al., 2016; Tian et al., 2017), potentially by indirectly inhibiting

GABAA (Wang et al., 2000) and enhancing NMDAR neurotransmission by Src kinase activation (Viviani et al., 2003), through a novel signaling mechanism that is not transcription-dependent (Balosso et al., 2008). In astrocytes this signaling may also contribute to the pathology, for example by downregulating glutamate uptake through transporters (Ye and Sontheimer, 1996; Hu et al., 2000) or by upregulating secretion of inflammatory factors like NO which potentiate neurotransmitter release (Casamenti et al., 1999). In combination with the concomitant increase in TNFα, the increase in IL-1β can also trigger calcium-dependent glutamate release from astrocytes (Bezzi et al., 2001;

- 37 -

Vezzani et al., 2008). This pathologic functional reversal of a basic astrocytic role has been suggested as an underappreciated source of excitation in epilepsy (Tian et al.,

2005). Not coincidentally, a spontaneous mutation (a SNP with currently unknown effects) in the gene encoding human IL-1 was identified as a risk factor for epilepsy through a meta-analytical approach of genetics studies (Kauffman et al., 2008).

In human patients with temporal lobe epilepsy, similar morphological and functional microglial activation has been found in the proximity of lesions and in neurodegenerating or sclerotic tissues (Morin-Brureau et al., 2018). This study also observed effects dependent on “time since last seizure”: the activated microglial state was switching from a transient response to the seizure activity to a chronic response to the injury. In addition, this group described a transient component of activation after seizures that is only present in areas without damaged neurons. While the causality between the lack of neuronal damage and the resolution of microglial activation is unclear, this finding is consistent with my general hypothesis that hippocampal microglia mount separate additive responses to the neuronal damage and neuronal hyperactivity seen after SE.

As previously mentioned, perhaps the most striking features of this activation pattern are the maintenance and even enhancement of microglial process motility in a markedly pro-inflammatory environment. This is because the majority of microglial

“activation” states normally coincide with stark decreases in morphological complexity

(Lynch, 2009; Karperien et al., 2013; Hoogland et al., 2015) and even greater reductions in baseline microglial process motility underlying “surveillance” (Stence et al., 2001;

Parkhurst and Gan, 2010; Winland et al., 2017).

- 38 -

Given its timing and anatomical profile, it is tempting to speculate that this microglial response to SE is an active contributor to the epileptogenic process. However, since little is known about the effect of short acute seizures/non-epileptogenic hyperactivity on these cells, it is imprudent to assume that these microglial changes may represent putative antiepileptogenic targets. Moreover, because experimental SE in rodents also acutely causes extensive hippocampal degeneration (Turski et al., 1984;

Pollard et al., 1994) and BBB breakdown (van Vliet et al., 2007; Marchi et al., 2012;

Gorter et al., 2015), it still remains unclear how much of the post-SE microglial activation is a result of the seizure activity per se, as opposed to the brain damage downstream of the epileptic crisis: it is plausible that the SE-induced injury (and not the epileptic crisis itself) is the chief contributor to the particular SE-induced microglial activation.

E. A gap in our knowledge: “Benign” seizures and microglia

As I have outlined above, findings from the most widely used models for seizure- induced epileptogenesis are complicated to interpret, given the presence of many confounding factors. Particularly, since SE models cause significantly more BBB and neuronal damage than SE in humans (and much more so than the average non-SE seizure) and moreover since those processes along with seizure activity itself can activate microglia, it remains unclear what if any part of the microglial response to SE is truly related to the pathology. Thus, it is imprudent to continue the search for microglial- or neuroinflammation-related biomarkers and treatment targets against

- 39 -

epileptogenesis without knowing the responses to injury/damage and to seizure activity. While the responses of microglia and other cells to brain damage has been extensively characterized (for example, in the realm of hypoxia/ischemia, toxicity, trauma or neurodegenerative disease), their reactions to paroxysmal activity seen in seizures is much less well known.

i. Non-epileptogenic seizures are common and can even be beneficial

Electroconvulsive shock (ECS) in rodents is a nearly perfect model of electroconvulsive therapy (ECT) in humans. As a gold standard treatment for depression and other mood disorders, ECT has an extensive history of safe and effective clinical use

(Wright, 1990; UK ECT Review Group, 2003). Unlike those in the SE model, individual

ECT/ECS-induced seizures are not thought to be acutely damaging nor epileptogenic

(Devanand et al., 1994; Anderson, 2018), nor is their continued use associated with increased risk of dementia or other adverse outcomes (Devanand et al., 1991; Osler et al., 2018). In this way, they may provide a better model of the average unprovoked seizure in humans, since most of these are neither damaging nor epileptogenic (Hauser et al., 1990; Berg and Shinnar, 1991).

The observation that seizure activity can result in modified behavior predates modern medicine and indeed, it was incidental observations of transient post-ictal improvement of psychiatric symptoms in patients that suffered from both mood disorders and epilepsy that spurred the rise of shock therapy in the early 20th century

(Fink, 1984; Wright, 1990; Berrios, 1997). Although initially induced by stimulant drugs

- 40 -

like pentylenetetrazol/Metrazol or even by hormones like insulin, therapeutic seizures were only widely used around the world upon the introduction of ECT as the most prevalent shock treatment approach (Leiknes et al., 2012). In a sad turn of events, the vast therapeutic potential of ECT against schizophrenia and major depressive disorder/MDD among other psychiatric conditions has been undermined throughout its history by the societal stigma against it and the related admittedly misguided and sometimes outright cruel overuse on the part of physicians, institutions and mental health authorities (Sadowsky, 2017).

Today, the safety and efficacy of ECT combined with the decreased latency to effect onset has resulted in a resurgence of its use in the clinic (Fink, 2009; Leiknes et al.,

2012; Weiner and Reti, 2017). At last, the historical stigma associated with ECT is being reversed, in part thanks to wider and better informed public coverage and conversations in the popular media, perhaps sparked by celebrity endorsers of this therapy (Fisher,

2011). The field is poised to make an even greater comeback as it leaves behind the unscientific practices of yore. Recent reports show that ECT effectively halves the 30-day returns to ER of patients with various psychiatric disorders (Slade et al., 2017). This gives ECT a much better prognosis than any pharmacotherapy and confirms it as the only treatment besides ketamine with a sufficiently fast onset for use in acute cases

(Sackeim, 2017).

Novel approaches to ECT stimulation like shorter or unilateral pulses have reduced the deleterious effects of this treatment like acute transient bouts of retrograde amnesia (Sackeim et al., 2008; Verwijk et al., 2012; Magid et al., 2013; Spaans et al.,

2013; Loo et al., 2014; Bansod et al., 2018). The overall safety of this procedure is

- 41 -

excellent: when properly administered, right unilateral ECT with ultrabrief pulses is associated with only infrequent and transient amnesia and it is seemingly no more dangerous than sham procedures under anesthesia and far safer than childbirth (Watts et al., 2011; Tørring et al., 2017). Although not much is known about the potential mechanism underlying the therapeutic mood stabilizing effect (or the deleterious amnesiac side effect), it has been observed that the expression and self-sustenance of the seizure is required (Perera et al., 2004). This backs up the parsimonious interpretation that since insulin shock therapy, Metrazol-induced seizures and ECT all have similar effects, it is the seizure itself and not the induction method that is the main driver behind therapeutic hyperactivity. ECT has been used successfully to treat behavioral and affective disturbances in patients with neurodegenerative and/or psychiatric disorders like Parkinson’s disease (Borisovskaya et al., 2016) or schizophrenia and unipolar or bipolar MDDs among others (UK ECT Review Group,

2003).

Besides these mood-stabilizing effects and perhaps through related mechanisms,

ECS in animals reliably confers protection from a variety of insults to neurons like SE

(Kondratyev et al., 2001) or adrenalectomy (Masco et al., 1999). Many other concomitant changes have been described acutely in the whole brain (and particularly the limbic system) after ECT including increases in neurotrophic signaling, in neurogenesis and in metabolism. These could explain both the antidepressant and the neuroprotective effects of ECT and ECS. Interestingly, neuroprotection has actually been implicated as a mechanism underlying the chronic mood-stabilizing effects of ECT

(Schwieler et al., 2016; Gan et al., 2017).

- 42 -

Given that depression is the most prevalent psychiatric disorder and perhaps the largest drain on society and that ECT has proven to be a supremely cheap, fast, safe and effective treatment alternative to psycho- or pharmacotherapy (Ross et al., 2018), it is of utmost importance to better understand the mechanisms recruited both beneficially and pathologically by these evoked seizures.

ii. Microglial response to “benign” seizures: A field clouded by a dearth of relevant physiology data

Other published work has studied the microglial response after ECS (for example

Jinno and Kosaka, 2008; or Jansson et al., 2009), but only found significant changes in microglial density and morphological/functional activation after chronic ECS stimulation (10-30 seizures). Other studies had already found a concomitant astrogliosis and changes in hippocampal interneurons after repeated ECS (Orzi et al., 1990). All these previous studies found that single ECS seizures failed to elicit any observable changes, including no effect on number or morphology of hippocampal microglia and astrocytes.

As I have reviewed previously, hippocampal microglia mount a particular activation pattern after experimental SE, characterized by changes in morphology and induction of pro-inflammatory signaling (which are expected to result from and exacerbate the severe neuronal injury), but also by maintenance and even enhancement of process motility together with changes in physiology (which I hypothesize are a specific response to the hyperactivity during the seizure crisis). Since individual ECS

- 43 -

seizures do not cause damage, one would not expect the ensuing acute microglial response to include the proliferation and morphological simplification seen after SE or other injurious events. This explains why previous efforts which mostly relied on count, density and shape of Iba1 staining had failed to see any effect on microglia after single

ECS seizures. Before our recent study (Sepulveda-Rodriguez et al., 2019), there had been no physiology studies employing live-cell imaging in functional assays and patch-clamp experiments (necessary to detect many aspects of post-SE activation and particularly those that are apparently more symbolic of seizure activity like purinergic signaling potentiation), leading to the premature conclusion that microglia do not respond to individual benign maximal seizures. The immune system and neuroinflammation are recognized as likely targets of ECT (van Buel et al., 2015, 2017; Kranaster et al., 2018;

Yrondi et al., 2018), as is the purinergic signaling (Sadek et al., 2011), which lends further credence to the possibility that the response I described after ECS could have a role in either the deleterious or beneficial aspects of ECT.

F. Emerging hypothesis

The overarching hypothesis behind my work was that the particular SE-induced hippocampal microglial activation extensively described in the literature was in fact a combined response to different facets of the SE “insult”. By comparing and contrasting the microglial effects of non-damaging, non-epileptogenic seizures (from the maximal electroconvulsive shock/ECS model) to those of injurious, epileptogenic seizures (from

Status Epilepticus/SE models), I aimed to disentangle the different facets of the

- 44 -

complicated activation pattern. I hypothesized that parts of that activation pattern were likely a response to the extensive damage and also that a distinct, dissociable subset of the changes were specifically caused by the seizure itself (that is to say, the paroxysmal hyperactivity per se).

- 45 -

2. MATERIALS, METHODS AND EXPERIMENTAL DESIGN1

A. Experimental animals and seizure models

We used postnatal day (P)30-45 male and female CX3CR1GFP/+ mice, which express green fluorescent protein in place of one allele of the fractalkine receptor gene, resulting in fluorescently labeled microglia (Jung et al., 2000). CX3CR1GFP/GFP mice

(#005582) were obtained from the Jackson Laboratory (Bar Harbor, ME, USA) and backcrossed onto a C57BL/6J (#000664, Jackson Laboratory) congenic background before this study. All animals used in this study were housed under specific-pathogen free conditions on a standard 12:12h light:dark cycle (lights on from 6am to 6pm) with water and standard chow available ad libitum. Lab staff or veterinary and husbandry technicians from the Department of Comparative Medicine performed a general check on the health of each litter and all breeding adults at least every 3 days. Cages were changed at least once every week. Sentinel animals living in cages with bedding pooled from the entire colony were routinely tested for infestation, infection and other diseases.

1 Most of the data in this dissertation have been previously published (Sepulveda- Rodriguez et al., 2019, see full citation in References). As such, certain media herein are reused or adapted from that publication including Figures 1E; 2B; 3A; 4; 5A, B, C; 6; 7; 8; 9; 10A; 11; 12; 13; 14; 15; 16; 17, 18 and 19.

Sepulveda-Rodriguez et al. (2019) is an open-access article distributed under the terms of the Creative Commons Attribution 4.0 International license (CCBY4.0), permitting unrestricted use, distribution and reproduction in any medium provided the original material and license are properly attributed.

- 46 -

Experimental and control mice were heterozygous littermates resulting from homozygote  wild type mating. Different individuals in a litter were identified by toe clipping at P7. To establish the breeding program and occasionally renew breeders, tail tissue was collected concurrently and sent out for genotyping (through Transnetyx

Cordoba, TN, USA). Littermates were weaned after P21 and randomized to treatment and control groups. Matched control animals were handled and treated in the same manner except that they received sham (0 mA) shocks or saline/vehicle instead of pilocarpine injections (see below). Vehicle and sham control animals were pooled in a single control group for all analyses. All procedures described herein were performed with the approval of the Georgetown University Animal Care and Use Committee under protocol 2016-1299.

We chose to employ two seizure models that recruit essentially the entire brain

(and particularly the hippocampus) to a similar degree as measured by 2-DG imaging or by induction of immediate early gene expression (Ingvar, 1986; Morinobu et al., 1997;

Hsieh et al., 1998; Ji et al., 1998; Scorza et al., 2002; Dyrvig et al., 2012; Sinel’nikova et al.,

2013) , but have little else in common. Unlike the prolonged crisis that defines Status

Epilepticus models, electroconvulsive shock-induced seizures are brief, non-injurious, non-epileptogenic and are not associated with behavioral deficits neither acutely nor later in life.

- 47 -

i. Electroconvulsive shock (ECS)-induced seizures

Maximal (tonic-clonic) seizures were elicited by transcorneal stimulation (as in

Peterson, 1998) using an Ugo Basile Electro-Convulsive Device (#57800, Stoelting Co.,

Wood Dale, IL, USA) and custom-built transcorneal electrodes. Animals received 0.5% tetracaine HCl eyedrops (Alcon, Ft. Worth, TX, USA) 15-30 minutes before stimulation.

Shocks consisting of 0.3 s long trains of 0.9 ms-wide square pulses (17 mA for females,

19 mA for males) at 200 Hz reliably evoked a Tonic Hindlimb Extension (THE, 180  to the torso) response lasting between 12 and 20 s with negligible mortality. Maximal ECS seizures manifest as a tonic-clonic flexion/extension sequence, which starts with a tonic forelimb extension, followed by hindlimb flexion and terminates in full tonic hindlimb extension.

Stimulation protocol and intensities were adapted from the literature (Frankel et al., 2001; Ferraro et al., 2002) and the associated dataset “Frankel1” in the Mouse

Phenome Database (The Jackson Laboratory), publicly accessible at https://phenome.jax.org/projects/Frankel1/protocol.

ii. Pilocarpine-induced status epilepticus (SE)

For the Status Epilepticus (SE) model, I employed systemic administration of pilocarpine, a standard protocol that results in a chronic phase with spontaneous seizures after a latent period (Curia et al., 2008). On the day of the SE induction, after animals have been acclimated to housing conditions and handling (all injections were i.p. and all drugs were dissolved in sterile 0.9% NaCl unless noted otherwise): 30

- 48 -

minutes after pretreating with 1 mg/kg each of scopolamine methylbromide and terbutaline, I injected 260-320 mg/kg pilocarpine HCl; and observed the animals as they progressed through modified Racine stages (1 = mouth and face automatisms, 2 = head bobbing, 3 = unilateral forelimb clonus, 4 = bilateral forelimb clonus and 5 = rearing and falling (Racine, 1972) and into SE (defined as continuous seizures over stage 3 for longer than 5 minutes). SE was terminated after 2 h by injecting diazepam (10 mg/kg).

Concurrently, mice received 0.25 mL (s.c.) of sterile warmed acetated Ringer’s + 5% dextrose. Ethosuximide (150 mg/kg, s.c., in phosphate-buffered saline/PBS) was administered 6 hours after the start of seizures, together with 0.3 mL of sterile 0.9%

NaCl. During and immediately after SE, mice were kept huddled/touching in a bare warmed (30-31 C) cage. This unique SE induction protocol (to date used exactly in this manner only by us in Sepulveda-Rodriguez et al., 2019) was adapted from personal communications with Dr. Helen Scharfman (Nathan Kline Insitute/NYU) and based on previously published studies (Cho et al., 2015; Iyengar et al., 2015; Jain et al., 2019). It reliably resulted in SE in all tested mice with relatively limited acute mortality (<15%).

B. Fixed tissue sectioning

24 hours after the induction of the seizures, animals were anesthetized with unmetered isoflurane (Patterson Veterinary, Greeley, CO, USA) or pentobarbital (>100 mg/kg) and intracardially perfused with cold PBS. Brains were quickly excised and drop-fixed overnight in 4% paraformaldehyde (#18505, Ted Pella Inc., Redding, CA,

USA) + 4% sucrose in PBS at 4 C. 50-100 μm hippocampal slices were cut horizontally

- 49 -

using a Vibratome Series 1000 (Vibratome) for immunostaining and morphometry. For

FluoroJade C studies, fixed brains were cryoprotected overnight in 30% sucrose in PBS before freezing in isopentane in dry ice (78 °C) and sectioned at 25 μm on a cryostat

(CM1850, Leica Biosystems, Nussloch, Germany) before immediately mounting on a series of 10 to 12 gelatin-subbed slides per brain.

C. FluoroJade C staining

To visualize neuronal damage, I used FluoroJade C, a polyanionic fluorescein derivative that can selectively mark degenerating neurons (Schmued et al., 2005). I employed an FJC Ready-to-Dilute kit (TR-100-FJC, Biosensis, Temecula, CA, USA) and followed the manufacturer’s instructions, except for halving the time in potassium permanganate. Briefly, after drying, slides were treated with basic ethanol solution for 5 minutes before transfer into 70% ethanol for 2 minutes, then rinsed in distilled deionized water (ddH2O) for 2 minutes. After incubating in a 0.06% potassium permanganate solution for 5 minutes, followed by a 2 min rinse in ddH2O, samples were stained in an acidified 0.001% FJC working solution for 10 minutes in the dark. After staining, slides were washed three times for 1 min in ddH2O, then placed on a slide warmer at 40 °C until dry before being cleared in xylene for 2 minutes and coverslipped with D.P.X. mounting medium (#13510, Electron Microscopy Sciences, Hatfield, PA,

USA).

Fluorescence photomicrographs from 3 sections per slide were captured on an upright microscope (i80, Nikon Instruments, Melville, NY, USA) with a QIClick camera

- 50 -

(QImaging, Surrey, B.C., Canada), using a standard FITC filter set and a 0.65NA 40X objective (Nikon Instruments). Images were captured by a blinded investigator using the same imaging conditions throughout. FJC positive cells in each image were manually counted by 2 blinded investigators and cell counts were averaged from at least 3 sections per animal.

D. Confocal imaging of hippocampal area CA1 in brain slices

Confocal Z-or ZT-stacks were taken using a laser scanning microscope system

(Thor Imaging Systems Division, Sterling, VA, USA) equipped with 488/561/642 nm lasers and Green/Red/Far-red filters and dichroic mirror and mounted on an upright

Eclipse FN1 microscope (Nikon Instruments, Melville, NY, USA). 284  284  20 μm (xyz) volumes of horizontal hippocampal slices containing CA1sr were imaged with a long working distance 60X water-dipping objective (CFI Fluor 60XW, NA = 1.0, WD = 2 mm,

Nikon). Differential interference contrast (DIC) images (on acute and fixed slices) or fluorescent images of NeuN (for neuronal nuclei) staining (on fixed slices only, see section on immunostaining for details) were used to identify and confirm the region of interest centered on CA1sr (with guidance from the atlas by Franklin and Paxinos,

2008).

E. Astrocytic cell density and area

We exploited sulforhodamine 101 (SR101) staining, which is used in vivo and in acute slices from adult rodents, resulting in specific red fluorescent labelling of live

- 51 -

astrocytes (Nimmerjahn et al., 2004) which preferentially uptake the anionic dye

(Schnell et al., 2012, 2015). For astrocytic density and area quantification I took volumetric Z-stacks of acute slices stained with SR101 (see section on preparation of acute slices for details) by imaging 21 planes of 2048  2048 pixel (px), 1 μm apart. I analyzed the maximal intensity projections (MIPs) across the Z axis, referring to the 3D

Z-stack for the manual cell counting analysis if necessary. SR101+ cells from a 284  284

 20 μm field containing CA1sr per hemi section were manually traced using the ROI tools in FIJI and the number and area of ROIs were measured.

F. Microglial cell density and area

For microglial density quantification I captured volumetric 2048  2048 px Z- stacks of fixed slices from perfused brains by imaging 21 planes 1 μm apart. I analyzed the MIPs across the Z axis, referring to the 3D Z-stack for the manual cell counting analysis if necessary. Cells were counted and the area of their soma measured in a single

284  284  20 μm field containing CA1sr per hemi section from manually drawn ROIs in

FIJI.

G. Immunostaining and microglial morphometry

Following perfusion and sectioning, slices were processed free-floating for immunofluorescence against GFP to better visualize microglia and their fine processes and against NeuN to identify stratum pyramidale. Sections were blocked and permeabilized for 2 hours in 0.5% TritonX-100 and 10% normal goat serum in PBS.

- 52 -

Next, slices were incubated overnight at 4 °C with mouse anti-GFP (1:1000, Chemicon

MAB3850, MilliporeSigma, Burlington, MA, USA) to stain microglia and rabbit anti-NeuN

(1:500, ABN78, MilliporeSigma) to stain neuronal nuclei. Slices were washed and then incubated at RT for 1 hour with secondary antibodies (1:1000 each; goat anti-mouse

AlexaFluor647, #A-21235, Thermo Fisher Scientific, Waltham, MA, USA; goat anti-rabbit

Cy3, #111-165-144, Jackson ImmunoResearch Laboratories, West Grove, PA, USA).

Sections were mounted and coverslipped using VectaShield fluorescent mounting media

(H-1200, Vector Laboratories, Burlingame, CA, USA).

Individual microglia were traced using the “FilamentTracer” tool in Imaris 7.4.2

(Bitplane, Concord, MA, USA) from Z-stacks of fixed anti-GFP stained slices with 41 planes of 4096  4096 px taken at 0.5 μm apart. I compared microglia morphometrically by extracting patterns of 3D Sholl crossings, numbers of branching points and primary branches and total filament tree lengths for each traced cell.

H. Preparation of acute brain slices

24 hours after seizure induction, mice were killed by decapitation and brains were rapidly dissected into an ice-cold sucrose aCSF slicing solution as previously described

(Al-Muhtasib et al., 2018; Sepulveda-Rodriguez et al., 2019) and in a manner consistent with the AVMA guidelines on euthanasia of laboratory animals (AVMA, 2013). The slicing solution contained (in mmol/L): 88 NaCl, 2.7 KCl, 0.5 CaCl2, 6.6 MgSO4 anhydrate,

1.4 NaH2PO4, 26 NaHCO3, 25 dextrose and 75 sucrose (all chemicals from Sigma, St.

Louis, MO, USA unless otherwise noted). 300 μm horizontal hippocampal slices were cut

- 53 -

on a Vibratome 3000 Plus Sectioning System (Vibratome, St. Louis, MO, USA), in ice-cold sucrose aCSF as above. Sections were recovered for 30 minutes at 32 °C in normal/recording aCSF solution, containing in mmol/L: 124 NaCl, 4.5 KCl, 1.2 Na2HPO4,

26 NaHCO3, 2 CaCl2, 1 MgCl2 and 10 dextrose. Slices were then transferred to room temperature (RT = 22-24 °C) and equilibrated for for greater than 10 minutes before use. In some experiments, slices were incubated with 1.8 μM sulforhodamine 101 in aCSF for 10 minutes to mark astrocytes (Nimmerjahn et al., 2004), see relevant section for details on cell counting. To limit artifactual microglial activation from dissection/sectioning, all slices were used within 5 hours of euthanasia and all microglia studied had somata at least 10 μm away from the cut surfaces. aCSF solutions were maintained at pH = 7.4 by bubbling with carbogen gas (95% O2/5% CO2, Roberts Oxygen,

Rockville, MD, USA). All experiments were conducted at RT.

I. Live imaging of microglial motility in situ in acute hippocampal slices

For the baseline and responsive motility experiments, I took 1024  1024 px ZT- stacks of acute slices by imaging 11 planes 2 μm apart every 30 s. If necessary, MIP

(maximal intensity projections over the z-axis) timelapses were registered using the

StackReg plugin (Thevenaz et al., 1998) in FIJI (Wayne Rasband/NIH, Bethesda, MD,

USA) before the motility analyses.

- 54 -

i. Basal motility or “surveillance”

Spontaneous motility of microglial processes directly mediates the physical microglia-neuron contact that is a prerequisite for many microglial functions like phagocytosis of synaptic terminals. I imaged this baseline behavior over 20 minutes in

CA1sr of either naïve slices or in the presence of a 0 mM [ATP] containing (aCSF-only) pipette (controls for responsive motility, see below). Motility analysis was performed in

FIJI by adapting the method described in Eyo et al. (2018a). I first manually cropped, then automatically thresholded and binarized the ROIs. The area above threshold at the end of the time-lapse movie (t = 20 minutes) was then measured and normalized to the area above threshold of the first frame of the movie (t = 0, extension index E.I. = 1.0).

The extension index through time of each time-lapse movie was then determined.

ii. Responsive motility assay

Responsive motility of microglial processes following ATP gradients is a vital endogenous response to injury (Davalos et al., 2005), and is a sensitive and reproducible in-slice assay of microglial purinergic signaling. In an assay adapted from the work of

Avignone et al. (2008), I lowered a patch pipette containing 1, 3, or 10 mM [Na-ATP] in aCSF into CA1sr, 10-20 μm deep and imaged the volume surrounding (284  284  20

μm) it for 20 minutes. I quantified process velocity with the Manual Tracking plugin in

FIJI. Between 3 and 8 responding processes per slice were manually tracked as they moved towards the pipette. Control experiments with 0 mM ATP (aCSF only in the

- 55 -

pipette) did not elicit any appreciable directional motility in the processes of nearby microglia.

J. Patch-clamp electrophysiology

Whole-cell patch-clamp recordings were performed under DIC illumination on fluorescently identified ramified cells, with 4.56.5 MΩ pipettes pulled from Wiretrol II borosilicate glass capillaries (Drummond, Broomall, PA, USA) filled with an internal solution as described in Avignone et al. (2008), containing (in mmol/L): 132 K- gluconate, 11 HEPES, 0.1 EGTA and 4 MgCl2, pH = 7.35 adjusted with KOH.

Liquid junction potential (LJP = 14 mV) was calculated in Clampex 11 (pClamp,

Molecular Devices, San Jose, CA, USA). The LJP correction was only applied to the reported RMP values. Patch-clamp experiments were performed with a MultiClamp

700B amplifier (Molecular Devices). Recordings were digitized at 20 kHz and low-pass

Bessel-filtered at 2 kHz with a DigiData 1440 (Molecular Devices) and a personal computer running Clampex 11. Recordings were analyzed offline with Clampfit 11.3

(pClamp). Results are shown as Current Density (C.D., in pA/pF) to take varying cell sizes/capacitances into account. Access resistance was monitored periodically during the experiment (with 5 mV pulses as above) and recordings with change greater than

20% were discarded.

- 56 -

i. Passive or intrinsic membrane properties and current-voltage relation

Resting membrane potential (RMP) was measured in current clamp (I = 0) mode, immediately after break-in to minimize effect of dialysis. All other data were recorded in voltage clamp configuration. Input resistance, access resistance and cell capacitance were calculated from 5 averaged current responses to brief (10 ms) 5 mV hyperpolarizing voltage steps from Vm = 60 mV. Internal and access resistance were derived by using Ohm’s law on the steady state and maximal amplitude of the response, respectively. Membrane capacitance was calculated from the time constant derived from a simple exponential fit of the decay phase. I/V relations were calculated using 500 ms square voltage steps from Vm = 60 mV to a series of voltages from Vm = 120 mV to

Vm = +50 mV every 10 mV.

ii. Quantification and biophysical characterization of P2X currents

For the ATP-evoked current studies, cells were held at Vm = 60 mV, with 500 ms continuous voltage ramps from Vm = 120 mV to +50 mV delivered every 10 s.

1 mM Na-ATP in normal or divalent cation-free/0CaMg aCSF (aCSF as described above, minus CaCl2 and MgCl2) was locally perfused via a custom-made Y-tube apparatus

(Hevers and Lüddens, 2002; Murase et al., 1989).

iii. Pharmacological characterization of P2X currents with Brilliant Blue G

For the P2X7 receptor-dependent current studies, the specific P2X7R antagonist

Brilliant Blue G (BBG, #B0770, Sigma) was dissolved in aCSF to create a 1 mM stock

- 57 -

solution (Jiang et al., 2000). Randomized slices were preincubated in 5 μM BBG in aCSF for 30 minutes (Avignone et al., 2008) before recording. Electrophysiological experiments were then conducted as outlined above but in the constant presence of 5

μM BBG in the bath aCSF.

K. Microglial gene expression analysis i. Immunomagnetic isolation of microglia

Microglial isolation was performed 24 hours after ECS seizures, exploiting the specificity of the Magnetic Activated Cell Sorting (MACS) approach with anti-Cd11b

MicroBeads (#130-049-601, all MACS supplies in this section are from Miltenyi Biotec,

Gaithersburg, MD, USA), tightly adhering to the manufacturer’s standard protocol for single cell dissociation and microglial isolation from adult brains, except for complete omission of the erythrocyte/red blood cell lysis step (since it can activate microglia and it is redundant given the perfusion).

Mice were anesthetized and perfused transcardially with cold PBS. Brains were rapidly dissected on ice and stored in MACS Tissue Storage Solution (#130-100-008) until dissociation could proceed (normally within 5 minutes and always less than 15 minutes). Brain tissues were sliced along eight parallel sagittal planes with a sterile scalpel, further homogenized by slowly pipetting with a P1000 tip 2-3 times, then placed together with the enzyme mix from the Adult Brain Dissociation Kit (#130-107-677) in a

C Tube (#130-093-237) for processing in the gentleMACS Octo Dissociator with Heaters

(#130-096-427) using the recommended program (37C_ABDK_01). The dissociated

- 58 -

tissue was resuspended in PBS and applied to a 30 µm SmartStrainer (#130-098-458).

All subsequent steps were performed at 4 °C except for the magnetic column separation.

The resulting single-cell suspension was centrifuged at 300  g for 10 minutes at 4 °C

(all centrifugation steps were performed at this temperature unless indicated, spinning on an Allegra X-30R machine with a JS-4.2 rotor, Beckman Coulter, Brea, CA, USA) and the supernatant aspirated. The cell pellet was resuspended in 3100 µL of PBS and mixed with 900 µL of ice-cold Debris Removal Solution. 4 mL of cold PBS was then overlaid on top and the tubes were centrifuged for 10 min at 3000  g (a swinging-bucket rotor is particularly critical for this centrifugation step). Three phases formed, the top two were discarded and cold PBS was added to the tube to bring the final volume to 15 mL, before being centrifuged at 1000  g for 10 minutes and aspirated completely. The pellet was resuspended in 10 mL cold 0.5% BSA in PBS and centrifuged at 300  g for 10 minutes.

The supernatant was again aspirated completely and the cell pellet was carefully resuspended in 90 µL cold 0.5% BSA to which 10 µL of anti-Cd11b MicroBeads were added, before incubation in the dark for 15 minutes at 4 °C. The cells were washed in cold 0.5% BSA and centrifuged at 300  g for 5 minutes, supernatant was aspirated completely and the pellet was resuspended in 500 µL 0.5% BSA. Microglia were magnetically isolated through MACS MS columns (#130-042-201) placed into MiniMACS

Separator magnets (#130-042-102). Unlabeled cells were collected in the original flow- through and after 3  500 µL washes. Labeled microglia were collected by flushing the column after removal from the magnetic field.

- 59 -

After MACS isolation and before cell lysis as indicated below, the positive fraction cell suspension was loosely pelleted by centrifuging briefly at 300  g. Most of the supernatant was aspirated and discarded, leaving  250 µL of fluid.

ii. Lysis of isolated cells and extraction and purification of total RNA

Cell pellets were immediately lysed with 0.8 mL cold Trizol LS (Invitrogen

#10296028, ThermoFisher Scientific) and homogenized by thorough pipette and vortex mixing. If necessary, samples were stored at 80 °C before further processing. After a 5 min initial incubation, 0.2 mL chloroform (C7559, Sigma) was added and samples were incubated for 2-3 minutes before being centrifuged for 15 minutes at 12,000  g and 4

°C. The aqueous phase was transferred to a new tube with 0.5 mL isopropanol (I9516,

Sigma) and RNA was precipitated by incubating for 10 minutes at RT. After being centrifuged for 10 minutes at 12,000  g and 4 °C, the supernatant was discarded and the RNA washed twice by resuspending the pellet in 75% ethanol (E7023, Sigma) in

UltraPure nuclease-free ddH2O (Invitrogen #10977023, ThermoFisher Scientific), then vortexing briefly before spinning down for 5 minutes at 7,500  g and 4 °C. After the 2nd ethanol wash was removed, RNA was air-dried for 10 minutes before resuspending by thorough pipetting in 25 μL RNase-free water. Samples were then incubated at 60 °C for

10-15 minutes before RNA quantity and purity were determined by standard spectrophotometry methods (NanoDrop 1000, ThermoFisher Scientific). If necessary,

RNA samples were stored at 80 °C before cDNA synthesis and qPCR analysis downstream.

- 60 -

iii. RT-qPCR

cDNA was synthesized from RNA samples by reverse transcription using random hexamer primers in the SuperScript IV First Strand Synthesis system (Invitrogen

#18091050, ThermoFisher Scientific) according to the manufacturer’s directions. If necessary, cDNA samples were stored at 20 °C before qPCR analysis. 10% of the 20 μL

RT reaction output was used in a 20 μL qPCR reaction for each technical replicate.

Hydrolysis probe-based qPCR reactions were run in multiplex on a 4-channel

(FAM/Green, HEX/Yellow, TEX615/Orange, Cy5/Red) Mic cycler instrument (Bio

Molecular Systems, Upper Coomera, Australia) using mouse gene expression assays from the PrimePCR line (Bio-Rad, Hercules, CA, USA) as follows:

Table 2: qPCR probes BioRad PrimePCR Assay Design Gene Target Fluorophore Unique Assay ID (see below) Actb qMmuCEP0039589 FAM exonic Tmem119 qMmuCEP0042925 Cy5 exonic Tnf qMmuCEP0028054 HEX exonic Ccl2 qMmuCEP0056726 Cy5 exonic P2rx1 qMmuCIP0031612 HEX intron-spanning P2rx4 qMmuCIP0028782 TEX615 intron-spanning P2rx7 qMmuCIP0042331 Cy5 intron-spanning P2ry6 qMmuCIP0029813 HEX intron-spanning P2ry12 qMmuCEP0057087 TEX615 exonic

Assay Design: For my purposes (studying mRNA abundance), intron-spanning designs are preferred over exonic ones whenever possible. Since amplicons that span introns can only be amplified from cDNA reverse transcribed from spliced mRNA transcripts, the confounding effect of gDNA contamination can be mitigated. On the other hand, exonic amplicons can give false positives if there is more contaminating

- 61 -

gDNA present than target mRNA, which highlights the importance of running no RT controls for these assays. In my hands, non-microglial brain RNA (from healthy tissue or from original or negative MACS fractions from naive brains) probed with Tnf and Ccl2 routinely amplified at around the same Ct as no RT controls from the same samples, meaning those transcripts were below the limit of detection. To account for this, I only quantified inflammatory markers (Tnf and Ccl2) in microglial RNA samples, where these transcripts are readily detectable. Regardless, microglia or other Cd11b+ macrophages are the most relevant and likely source of these pro-inflammatory factors.

PCR reactions were prepared in duplicate or triplicate with Bioline SensiFast

Probe No-ROX mastermix (Thomas Scientific, Swedesboro, NJ, USA) following the manufacturer’s directions for multiplex assays (initial denaturation/polymerase activation for 3 min @ 95 °C; then 40 cycles of 10 s denaturation @ 95 °C, 50 s annealing/extension @ 60 °C, signal acquisition). The following multiplex qPCR assay combinations were validated by comparing their results to those from parallel singleplex reactions (data not shown):

Table 3: qPCR multiplex assays. Validated Multiplex Target 1 Target 2 Target 3 Target 4 qPCR Assays FAM HEX TEX615 Cy5 P2X Receptors Actb P2rx1 P2rx4 P2rx7 Actb P2yr6 P2yr12 Tmem119 P2YRs/µglial Marker Inflammation Markers Actb Tnf Ccl2

Threshold cycle was automatically determined and averaged across replicates by the Mic cycler manager software (Bio Molecular Systems). Amplification traces for were manually checked for quality control: for instance, I checked for low SDs across technical replicates and weeded out potential false positive signals using no RT controls in each

- 62 -

run. Fold changes were determined using the 2∆∆Ct method, with expression of all transcripts normalized to mean Actb levels in the control group.

L. Quantification of CCL2 and total protein in hippocampal lysates

After decapitation, hippocampi were dissected in ice-cold PBS, flash frozen in isopentane cooled in dry ice (78 °C) and lysed in 200 μL of RIPA buffer + HALT inhibitor cocktail/hemi-hippocampus (both from ThermoFisher Scientific). Lysates were probed for CCL2 using Quantikine ELISAs (MJE00, R&D Systems, Minneapolis, MN, USA) and for total protein using a Pierce BCA assay (#23227, ThermoFisher Scientific), both according to manufacturers’ directions.

M. Experimental design and statistical analysis

Our primary question of interest was the degree to which ECS would alter microglial responses. The impact of SE on the same endpoints I measure for ECS have been previously reported and replicated by others (Avignone et al., 2008; Foresti et al.,

2009; Menteyne et al., 2009; do Nascimento et al., 2012; Eyo et al., 2014; Arisi et al.,

2015; Avignone et al., 2015; Eyo et al., 2016; Schartz et al., 2016; Tian et al., 2017; Wyatt-

Johnson et al., 2017; Bosco et al., 2018; Feng et al., 2019).

For comparison purposes, I included a Status Epilepticus group as a positive control in subset of experiments. I focused the analysis on the hippocampal subfield CA1 and particularly CA1sr, as the microglial response to status epilepticus has been best characterized in this region. All data were analyzed by investigators blinded to

- 63 -

treatment status using FIJI, Clampfit 10.7 and 11, Excel (Microsoft, Redmond, WA, USA) and Prism (GraphPad, LaJolla, CA, USA). Results are presented in text as mean 

Standard Error of the Mean (SEM) and graphically as violin plots with all datapoints shown and the mean, median and quartiles indicated. N is number of animals used and n is number of slices/fields or number of cells for the patch-clamp electrophysiology studies. I did not detect any sex differences in microglial density, morphology and motility, or in microglial gene expression and thus combined data across sexes for further statistical analysis. Similarly, wherever applicable, control groups for both seizure models (saline vehicle for SE, sham shock for ECS) were combined.

Data were statistically analyzed by two-way ANOVA followed by Sidak’s multiple comparisons test (microglial 3D Sholl analysis data; qPCR data), by one-way ANOVA followed by post-hoc multiple comparisons with Tukey’s test (other microglial morphometry data; CCL2 ELISA analysis data), by Kruskal-Wallis non-parametric test followed by multiple comparisons with Dunn’s correction (for the non-normally distributed BBG preincubation electrophysiology data), or by Fisher’s exact, Mann-

Whitney, or t- tests for all other comparisons. I established statistical significance at p <

0.05, applied recommended multiple comparison corrections where appropriate and computed all p-values from two-tailed distributions.

- 64 -

3. ELECTROCONVULSIVE SHOCK: A NON-DAMAGING AND PUTATIVELY NON-

EPILEPTOGENIC SEIZURE MODEL

A. Introduction

The basic premise behind this project was that ECS could induce maximal seizures that are neither injurious nor epileptogenic. As reviewed in the introduction in

Chapter 1, spontaneous seizures like these are common in humans even in the absence of epilepsy (Hauser et al., 1990) and when evoked by ECT they can even be beneficial to the affective symptoms of several psychiatric and neurological diseases.

Although there exists much evidence that ECT is not associated with brain damage or accelerated cognitive or neurological decline in humans, the injurious potential of the ECS seizure was directly tested in my model of choice. To accomplish this, the effect of ECS on hippocampal CA1 astrocytes and neurons was studied 24 h after seizures. Others have previously reported that SE acutely triggers proliferation and morphological and functional changes in astrocytes in hippocampus (Shapiro et al.,

2008; Wilcox et al., 2015; Plata et al., 2018) and that it causes extensive neuronal damage and neurodegeneration brain-wide (Turski et al., 1984; Ben-Ari, 1985; Covolan and Mello, 2000). In contrast the hypothesis here was that that ECS would result in neither changed astrocyte density and size nor in increased neuronal cell death in hippocampal area CA1.

- 65 -

B. Experimental approach

See Chapter 2 for more details on experimental procedures.

To directly evaluate the injurious potential of the chosen seizure models, the state of hippocampal astrocytes and neurons was examined 24 h after ECS or control treatment.

To study astrocytes, sulforhodamine 101 staining was employed in acute slices resulting in selective labeling of live astrocytes (Nimmerjahn et al., 2004). Astrocytic somata were counted and their area as well as the total area stained were measured in

FIJI.

FluoroJade C staining (Schmued et al., 2005) was performed on thin cryosections from perfused and fixed brains to quantify neurodegeneration 24 h after seizures. An SE group was included in each processed batch as a positive control.

C. Results i. Electroconvulsive shock does not cause hippocampal astrogliosis 24 hours after seizures

As shown in the representative MIPs in Figure 1A, SR101 treatment of acute slices resulted in selective staining of astrocytes in hippocampus. As quantified in Figure

1B, SR101+ astrocytes did not increase in density 24 h after ECS: per 106 m3 imaged, there were an average of 29.2  4.1 cells in n = 25 slices from N = 7 control mice and 21.4

 2.4 in n = 24 slices from N = 6 ECS-treated mice and non-parametric statistical analysis revealed no significant difference (U(668,556.5) = 256.5; p = 0.3901, Mann-Whitney test).

- 66 -

- 67 -

Figure 1: Electroconvulsive shock does not cause acute hippocampal astrogliosis or neurodegeneration. A. Representative Maximal Intensity Projections (MIPs) taken from CA1 of acute horizontal slices stained in sulforhodamine 101 (SR101) to mark astrocytes 24 hours after sham shock (left, Control) or after a single maximal electroconvulsive shock seizure (right, ECS). B. Density of SR101+ astrocytes, in cells per 106 m3 of CA1 volume imaged. ECS had no effect as compared to control 24 hours after seizures. C. Cumulative projected area of SR101+ astrocytes as a percentage of each 284  284  20 μm field. There was no significant difference between slices from control and ECS animals. D. Average area of individual astrocytic somata in each field. The size of the soma of SR101+ cells did not change as a function of treatment group. E. Number of FluoroJade C positive cells per field imaged (320m x 240m). Control and ECS-exposed animals have significantly lower degenerating cell densities in CA1 than SE-exposed animals. **** : p < 0.0001. F. Representative fluorescent microscopy images of FJC and DAPI staining in area CA1 of horizontal slices of mouse hippocampus, 24h after sham treatment (left), ECS (middle) or SE (right).

- 68 -

Similarly, ECS failed to affect the area of fluorescent signal per cell (Figure 1C,

58.9  1.8 for 1196 control cells and 59.7  2.2 for 826 ECS cells, slice/field and animal numbers as above, t(47) = 0.2766, p = 0.7833, t-test) and the total area of fluorescence per field (Figure 1D, average of 3.33  0.07% for control slices and 3.24  0.06% for slices from animals in the ECS group, n and N as above, t(47) = 0.9354, p = 0.3544, t-test).

In conclusion, SR101 staining of live astrocytes in CA1 from acute hippocampal slices failed to detect any ECS effect 24 h after the seizure.

ii. Unlike status epilepticus, single electroconvulsive shock seizures do not result in acute neuronal injury

To verify that ECS treatment was not associated with neuronal damage in the hippocampus, I counted FluoroJade C (FJC) positive cells. FJC is a polyanionic dye that selectively marks degenerating neurons (Schmued et al., 2005).

As quantified in Figure 1E (and exemplified in Figure 1F), FJC+ cells were commonly found in hippocampal area CA1 of mice from SE but not ECS or control groups: per 320 m x 240 m field imaged, control animals had 4.2  0.36 FJC-positive cells on average (N = 9), comparable to the 3.6  0.54 found in ECS-treated animals (N =

5). Conversely, animals in the SE group had an average of 16.3  1.5 cells (N = 7) per field.

Analysis by one-way ANOVA revealed a statistically significant effect of treatment

(F(2,18) = 53.1, p = 0.00000003), that was driven by the SE group, which differed from both control (q(18) = 13.13, p = 0.00000008, Tukey’s test) and ECS groups (q(18) = 11.89,

- 69 -

p = 0.0000003, Tukey’s test). Control and ECS groups did not differ from each other

(q(18) = 0.612, p = 0.902, Tukey’s test). Thus, as hypothesized, single ECS did not cause acute neurodegeneration, a profile different from that following SE, which is associated with high levels of degeneration brain-wide.

D. Summary and conclusion

First, the astrocytic response to ECS was studied by counting and measuring

SR101-labelled live astrocytes in acute slices 24 h after the seizure. ECS did not cause changes in astrocyte density, nor in total or individual cell area. Second, FJC+

(degenerating) neurons were counted in CA1 of perfused and fixed brains from mice 24 h control, ECS or SE treatment.

Unlike SE, ECS did not induce increased neuronal cell death. In conclusion, as initially hypothesized, single ECS in mice is not associated with astrogliosis nor with neuronal cell death 24 h after the seizures in the hippocampal area CA1.

- 70 -

4. ELECTROCONVULSIVE SHOCK HAS NO EFFECT ON MICROGLIAL DENSITY OR

MORPHOLOGY IN CA1

A. Introduction

Given that ECS did not cause astrocytic changes (proliferation or atrophy) or neuronal cell death in hippocampus 24 h after the seizure and is not associated with epileptogenesis, I can use this model to study the effect of neuronal hyperactivity on hippocampal microglia, in the absence of overt neurodegeneration.

One of the most readily apparent changes to microglia after activation by a variety of insults including SE are the markedly increased density and decreased morphological complexity (Avignone et al., 2008). Cellular proliferation is likely the product of clonal expansion of activated microglia to increase the strength of the response. Morphological simplification is usually associated with increases in phagocytic and secretory functions at the expense of baseline surveillance and homeostatic maintenance. As such, these characteristics were the first to be studied in my ECS model.

B. Experimental approach

See Chapter 2 for more details on experimental procedures.

To begin to characterize the effects of single ECS on microglia in hippocampal area CA1, their density and morphology were analyzed by counting and reconstructing

- 71 -

anti-GFP immunoboosted1 microglia from CX3CR1GFP/+ brains (Jung et al., 2000) perfused and fixed 24 h after sham or shock treatment. Neuronal nuclei were visualized by anti-NeuN immunofluorescence to assist in anatomical location.

Microglial somata were manually counted in FIJI to determine density. Processes were traced and reconstructed using Filament Tracer in Imaris, before performing 3D

Sholl analysis and extracting relevant morphometry measures of each cell such as the number of branch points and primary branches, the total length of the process tree and the average length per process.

C. Results i. Electroconvulsive shock does not cause acute proliferation of microglia in CA1

A single ECS-induced tonic-clonic seizure did not result in observable differences in microglial density 24 hour after the seizures in the CA1 hippocampal subfield. As shown in the examples in Figure 2A (hippocampus at low magnification) and Figure 3A

(CA1 at higher magnification) and the summary data in Figure 2B (by slice/field left, by animal right), controls (N = 13 mice, n = 56 fields) had a mean cell density of 27  2.9 microglia/106 μm3 of volume imaged in CA1, compared to 282.7 in the ECS group (N =

12, n = 53). Statistical analysis with t-test revealed no significant difference at the level of the slices (t(107) = 0.949, p = 0.34) or the individual animals (t(23) = 0.367, p = 0.717).

1 Immunoboosting is amplifying an existing GFP signal by immunofluorescent staining against GFP using a dye with similar spectra (i.e. Alexa 488).

- 72 -

Microglial somata become enlarged before and during proliferative responses. As quantified in Figure 3B, the average area of the somata of microglia in the CA1 hippocampal region did not differ as a function of treatment: 55.9  3.7 μm2 for controls vs. 51.7  3.1 μm2 for ECS-treated (t(107) = 0.8645, p = 0.39, t-test).

As was expected given that neither cell density was affected by ECS, the ratio of microglial to astrocytic counts was similarly unchanged in hippocampal area CA1 24 h after the seizure: 1.63  0.18 for controls vs. 1.56  0.18 for ECS (t(47) = 0.3138, p = 0.76, t-test, Figure 3C).

- 73 -

Figure 2: Electroconvulsive shock does not trigger acute microglial proliferation in hippocampal area CA1. A. Representative MIPs from immunofluorescence images showing CX3CR1-GFP+ microglia in green and NeuN+ neurons in red in horizontal hippocampal slices of mice exposed to control (left) or ECS (right), 24 hours after seizures. B. Density of microglia in CA1sr microglia, in cells per 106 m3 of CA1sr volume. Neither the count per slice (left) nor per animal (right) was affected 24 hours after ECS as compared to controls.

- 74 -

Figure 3: Electroconvulsive shock does not cause microglial somatic changes in hippocampal area CA1. A. Representative MIPs from immunofluorescence images showing CX3CR1-GFP+ microglia in green and NeuN+ neurons in red in CA1 (str. pyramidale left of dashed line and str. radiatum right of dashed line) of control (left) and ECS (right) mice 24 h after seizures. B. Average area of microglial somata in each field. The size of the soma of CX3CR1-GFP+ microglial cells did not change as a function of treatment group. C. Ratio of CX3CR1-GFP+ microglia to SR101+ astrocyte counts per field is not altered 24 h after ECS as compared to sham shocked controls.

- 75 -

ii. Unlike status epilepticus, electroconvulsive shock has no effect on CA1sr microglial morphology

To investigate the effect of the experimental treatment on microglial morphology, confocal z-stacks of individual cells in CA1sr were reconstructed after perfusion, fixation and immunofluorescent amplification of GFP. Representative tracings from control

(left), ECS (center) or SE (right) are shown in Figure 4A. A total of 38 individual cells were reconstructed: 14 cells from 7 control animals, 14 cells from 7 ECS animals and 10 cells from 5 SE animals. 2D-projected tracings from each cell used in this analysis are included in Appendix I.

Microglia from control animals were highly ramified and had long and complex processes with regular branching, as is expected under physiological conditions.

Compared to controls and unlike ECS animals, SE animals displayed clear morphological activation of hippocampal microglia, as has been previously described by others

(Avignone et al., 2008, 2015; Wyatt-Johnson et al., 2017).

The 3D Sholl profile was more compact in slices from SE-exposed animals as compared to slices from controls (F(1,1320) = 1761, p < 10-15), with significantly fewer crossings from 2m to 33m from the cell body (all p < 0.0001, Tukey’s tests, Figure 4B).

Cells from SE-exposed animals had significantly fewer branching points (68  4.2 in control vs. 9.4  2.0 in SE, q(35) = 12.4, p = 0.0000000007, Tukey’s test, Figure 5A) and primary branches (9.5  0.53 in control vs. 3.7  0.42 in SE, q(35) = 12.73, p =

0.0000000004, Tukey’s test, Figure 5B), a significantly shorter total process length (748

 44 m in control vs. 178  26 m in SE, q(35) = 13.4, p = 0.0000000001, Tukey’s test,

- 76 -

Figure 5C) and significantly longer average branch length (11.07  0.23 m in control vs.

20.16  1.68 m in SE, q(34) = 11.12, p = 0.00000001, Tukey’s test, Figure 5D).

On the other hand, microglia from the hippocampus of ECS-exposed animals had no detectable differences in morphology when compared to controls. Their 3D Sholl profile was similar to control cells (p > 0.9 at all radii, Tukey’s multiple comparisons tests, Figure 4B) and they also had no significant changes in total number of branching points (76  5.5, q(35) = 1.79, p = 0.43, Tukey’s test, Figure 5A), number of primary branches (8.36  0.31, q(35) = 2.747, p = 0.142, Tukey’s test, Figure 5B), cumulative process length (843  43 m, q(35) = 2.422, p = 0.22, Tukey’s test, Figure 5C) or average length per branch (11.39  0.46 m, q(34) = 0.4366, p = 0.95, Tukey’s test, Figure 5D).

- 77 -

Figure 4: Electroconvulsive shock does not result in simplification of microglial arborization in CA1sr. A. Representative binarized images of individual traced and reconstructed immunofluorescence-boosted CX3CR1-GFP+ microglia from CA1sr of control (left), ECS (middle) or SE (right) treated animals, 24 h after seizures (see other cells in APPENDIX I). Microglial morphology in CA1sr was drastically changed after SE. B. SE, but not ECS, significantly decreased the number of 3D Sholl crossings in CA1sr microglia for sphere radii between 2m and 33m as compared to controls. * : p < 0.05

- 78 -

Figure 5: Electroconvulsive shock does not trigger changes in microglial morphometry in CA1sr. SE had significant effects on a variety of morphometric measures: it decreased the number of branch points (A) and primary branches (B) per cell, decreased the cumulative length of the average filament tree length (C) and increased the average length of their branches (D); while ECS had no acute effect on any of the measured parameters. **** : p < 0.001.

- 79 -

D. Summary and conclusion

To search for “visible” proliferative and morphological activation of microglia, their density and the size of their soma were quantified and individual cells were traced and reconstructed for 3D Sholl analysis and to extract morphometry measures such as number of total branching points and primary branches and length of the total filament tree and of the average branch.

ECS did not result in changes to microglial density or microglial morphology, in stark contrast to the obvious effects SE has on these measurements.

- 80 -

5. EFFECTS OF ELECTROCONVULSIVE SHOCK ON MICROGLIAL PROCESS

MOTILITY in CA1

A. Introduction

To further probe for potential seizure-induced changes in microglial function, the effect of ECS on the two basic process motility phenotypes (stochastic “surveillance” extensions and retractions at baseline and coordinated responsive motility to extracellular ATP) was evaluated.

The constant background motility of processes from microglia in a “resting state” is not well understood, with explanations for its role ranging from pure stochastic surveillance/sampling of the environment in their exclusive domain (Nimmerjahn et al.,

2005) to dynamic responses to the activity of individual synapses closeby (Wake et al.,

2009; Abiega et al., 2016). Microglial contact can enable phagocytosis (Brown and

Neher, 2014; Abiega et al., 2016), displace synapses (Chen et al., 2014), promote plasticity through location-specific neurotrophic signaling (Parkhurst et al., 2013;

Cserép et al., 2019) or remodel synaptic structures (Tremblay et al., 2010, 2012), for example by trogocytosis/nibbling (Weinhard et al., 2018). Microglial process ramification and motility, as well as the release of pro-inflammatory factors like IL-1β, are regulated by membrane potential, which in turn depends greatly on the expression and function of the tandem-pore domain halothane-inhibited potassium channel subunit

THIK1 (Madry et al., 2017).

- 81 -

On the other hand, responsive motility is a well-preserved response to injury through which microglia react to damage to even single neuron through ATP signaling

(Davalos et al., 2005). This endogenous response has been coopted as a sensitive in slice assay of microglial purinergic signaling (Avignone et al., 2008). The importance of microglial motility to brain development and homeostasis is readily apparent through the study of P2Y12 knockout mice. These mice display perturbations in neuronal plasticity during critical periods of development (Sipe et al., 2016), deficient blood vessel repair after injury (Lou et al., 2016) and even worsened outcomes in experimental models of epilepsy (Eyo et al., 2014).

B. Experimental approach

See Chapter 2 for more details on experimental procedures.

Baseline and responsive motility were imaged in the presence of no pipette/aCSF-only pipette, or of pipette with 3 mM ATP in aCSF respectively, 24 h after the seizures in acute horizontal slices of hippocampus from CX3CR1GFP/+ mice. To quantify surveillance, an Extension Index (defined as the area of extensions divided by the area of retractions, normalized to 1 at t = 0) was measured over 20 minutes of baseline imaging (Eyo et al., 2018a). Processes extending towards the ATP-containing pipette were tracked using the Manual Tracking plugin in FIJI to quantify the velocity of responsive motility (assay adapted from Avignone et al., 2008). An SE group was included as a positive control for this assay that was new to our lab. To better

- 82 -

understand the effect observed after ECS, the motility evoked by pipettes containing 1 mM and 10 mM ATP was also studied to form a dose-response relation.

C. Results i. Electroconvulsive shock has no effect on baseline microglial motility

To evaluate baseline motility, an extension index was calculated by dividing the mean area of process extensions by the mean area of retracted processes (as in Eyo et al., 2018a) in an imaged field over time, as illustrated by the examples in Figure 6A

(fields) and Figure 6B (individual cells).

Over 20 minutes of imaging under baseline conditions, I measured similar mean extension indices of 1.090.04 for control slices and 1.070.05 for ECS slices (N = 7 and

6 animals; n = 17 and 12 slices; t(27) = 0.269, p = 0.79, t-test, Figure 6C).

- 83 -

Figure 6: Electroconvulsive shock does not alter hippocampal microglial baseline motility. A,B. Representative time-coded images (t=0 in red, t=20 minutes in green, overlap in yellow) of CA1sr fields in A or single cells in B from slices from control (right/top) and ECS-treated (left/bottom) animals. C. ECS had no effect on mean Extension Index (area of extensions/area of retractions) per field after 20 minutes of imaging under baseline conditions.

- 84 -

ii. Like status epilepticus, electroconvulsive shock enhances the response velocity of microglial processes towards ATP

I next tested for one of the more distinctive consequences of the potentiated purinergic signaling changes seen in SE-induced microglial activation: enhanced responsive motility of microglial processes (Avignone et al., 2008, 2015). This motility across increasing ATP gradients (and towards point sources of extracellular ATP) is an important endogenous response to injury (Davalos et al., 2005), as well as a sensitive in- slice assay of microglial purinergic signaling.

To study the effect of ECS on this, microglia were imaged as their processes extended in response to a pipette containing ATP in aCSF. As exemplified in the top panel of Figure 7, microglia from control animals slowly mounted a response in the form of a narrowing circle, formed by the leading edge of the processes as they advanced towards the pipette. Importantly, this directional motility was not evoked by aCSF-only pipettes, confirming that the response was to ATP and not the pipette.

As shown in the middle and bottom rows of Figure 7 and in stark contrast to the responses in control slices, slices from ECS and SE animals displayed similarly enhanced microglial process motility towards 3 mM ATP in a patch pipette 24 hours after the seizures. Statistical analysis by one-way ANOVA revealed a significant effect of treatment group (F(2,91) = 17.9, p < 10-15).

As quantified in Figure 8, controls (n = 39 slices from N = 10 mice) displayed a mean process velocity of 2.7  0.04 m/min, compared to 4.3  0.12m/min for 38 slices from 10 ECS mice and 4.4  0.23m/min for 17 slices from 5 SE mice (by Tukey’s

- 85 -

multiple comparison test; C vs. ECS: q(91) = 15.7, p = 0.0000000004; C vs. SE: q(91) = 13.1, p = 0.0000000004; ECS vs. SE: q(91) = 0.793, p = 0.84).

Figure 7: Seizure-induced potentiation of responsive microglial process extension behavior. Representative MIPs of confocal ZT-stacks showing the time-course of the microglial response (in green) to 3 mM ATP in a patch pipette in acute hippocampal slices from animals in control (top), ECS (middle) and SE (bottom) groups. Right-most panel in each row shows examples of analyzed process tracks.

- 86 -

Figure 8: Seizures enhance the velocity of microglia responding to 3 mM [ATP]. ECS and SE similarly increased the average process velocity during the CA1 hippocampal microglial response to a patch pipette containing 3 mM [ATP] in aCSF. **** : p < 0.001.

- 87 -

iii. Electroconvulsive shock-induced enhancement of microglial response to ATP is concentration-dependent

As shown in Figure 9, ECS enhancement of ATP-directed microglial process motility was relatively much smaller but still statistically significant when the concentration of ATP in the patch pipette solution was 1 mM (1.5  0.05 m/min for control vs. 1.8  0.05 m/min for ECS, t(18) = 3.79, p = 0.0013, t-test, n = 10 slices per group, from N = 3 animals per group). The positive effect of ECS on microglial responsive process velocity was not detectable when the pipette contained an assay-saturating concentration of ATP (10 mM; 5.3  0.14 m/min for control vs. 5.4  0.17 m/min for

ECS, t(17) = 0.233, p = 0.82, t-test, n = 10 and 9 slices per group, from N = 3 animals per group).

- 88 -

Figure 9: Electroconvulsive shock potentiates microglial motility in response to 1 and 3 but not 10 mM [ATP]. ECS-induced enhancement of microglial responsive motility is concentration-dependent. Besides the potentiated response to 3 mM ATP (control and ECS data from Figure 8 are repeated here for comparison), there was a small but significant difference in the microglial responses to 1 mM ATP (see inset), while the responses to 10 mM ATP were not significantly different. **** : p < 0.001.

- 89 -

D. Summary and conclusion

To study the effect of ECS on microglial motility in CA1 in acute hippocampal slices taken 24 h after the seizure, two assays were used: an Extension Index was calculated to evaluate baseline surveillance motility, whereas the velocity of the responsive motility to an ATP source was quantified from manual tracking.

ECS had no measurable effect on the “random” microglial process extension and retraction that mediates baseline surveillance at “resting” states.

Surprisingly, ECS potentiated ATP-responsive microglial process motility, to a similar degree than SE did. This enhancement of responsive microglial motility was dose-dependent, as a smaller ECS effect was measured in assays with 1 mM [ATP] than those with 3 mM. The responses to 10 mM [ATP] did not differ between slices from control and ECS-exposed animals.

Thus, the microglial response to ECS does indeed include some facets that have been described after SE, which suggests that this subset of changes could comprise a novel conserved signature of general ictal activity in microglia in mouse hippocampus.

- 90 -

6. EFFECTS OF ELECTROCONVULSIVE SHOCK ON CA1sr MICROGLIAL

MEMBRANE PHYSIOLOGY

A. Introduction

For further characterization of hippocampal microglia after ECS, the effects of

ECS on passive and intrinsic membrane characteristics of hippocampal area CA1sr microglia were studied through whole-cell patch-clamp electrophysiology recordings.

To directly test the hypothesis that ECS enhanced ionotropic purinergic signaling mechanisms, purinergic currents were measured after being induced by local perfusion of 1 mM [ATP] aCSF. Ionic and pharmacological manipulations were also used to identify the likely receptor subunits mediating the observed currents.

B. Experimental approach

See Chapter 2 for more details on experimental procedures.

After patching fluorescently identified cells in acute slices from CX3CR1GFP/+ brains 24 h after seizures (see Figure 10A), resting membrane potential was measured in current clamp I = 0 mode, before switching to voltage-clamp mode and giving brief hyperpolarizing pulses to measure input resistance and membrane capacitance. To account for differences in cell area and thus exposure to extracellular medium and agonist, currents were expressed as densities (pA/pF) rather than amplitudes (pA).

Current-voltage relation was then determined from a series of hyperpolarizing and depolarizing voltage steps as illustrated by Figure 10B.

- 91 -

For the agonist-evoked current studies, voltage ramp protocols (as exemplified in

Figure 10C) were repeated before and during the time that 1 mM [ATP] in aCSF was locally applied through a y-tube (Hevers and Lüddens, 2002), in the presence (normal) or absence (0CaMg/aCSF minus calcium chloride and magnesium chloride) of divalent cations in the extracellular media to potentiate P2X-mediated currents (Michel et al.,

1999; Yan et al., 2011).

For the P2X7R-mediated current studies, experiments were performed as above, except that slices were preincubated in the specific antagonist Brilliant Blue G which was also present in the media during recording (Jiang et al., 2000).

- 92 -

Figure 10: Microglial patch-clamp electrophysiology experiments. A. Representative photomicrograph showing CX3CR1-GFP+ microglia in CA1sr superimposed with a 60X DIC image of the tissue. B. Voltage step protocol used to study passive properties and voltage-activated currents and example currents evoked in response. C. Voltage ramp protocol used for agonist-evoked current studies and representative current responses.

- 93 -

C. Results

I used patch-clamp electrophysiology in the whole-cell configuration targeting fluorescently tagged microglia, as had previously been done after SE (Avignone et al.,

2008). I first studied intrinsic and passive properties of microglia before measuring and characterizing purinoreceptor-mediated currents. i. Electroconvulsive shock has no effect on intrinsic properties of CA1sr microglia

The membrane physiological properties of a total of 21 cells from 4 control animals and 14 cells from 3 ECS animals were recorded. As measured by whole-cell patch-clamp in acute hippocampal slices, control microglia displayed the expected bi- modal distribution of negatively polarized resting membrane potentials (with a mean of

47.1  4.7 mV, Figure 11A), high input resistance (1.77  0.32 G, Figure 11B) and relatively low membrane capacitance (38.4  1.5 pF, Figure 11C).

ECS did not affect the intrinsic and passive electrophysiological properties of

CA1sr microglia: the distribution of resting membrane potential (ECS mean RMP = 42.8

 4.9 mV, t(29) = 0.635, p = 0.53, t-test, n = 17 and 14, Figure 11A) and the mean input resistance (ECS: 1.94  0.4 G, t(32) = 0.348, p = 0.73, t-test, n=20 and 14, Figure 11B) and membrane capacitance (ECS: 36.2  1.3 pF, t(33) = 1.02, p = 0.316, t-test, n = 21 and

14, Figure 11C) were all similar across both groups (N = 3 and 4 mice for the control and

ECS groups respectively) and in accordance with previously published ranges (Boucsein et al., 2000, 2003; Avignone et al., 2008; Menteyne et al., 2009; Kettenmann et al., 2011;

Verkhratsky and Noda, 2014; de Biase et al., 2017; Madry et al., 2018).

- 94 -

Figure 11: Electroconvulsive shock does not affect intrinsic membrane properties of CA1sr microglia. ECS did not affect resting membrane potential (A), input resistance (B) or membrane capacitance (C) of ramified microglial cells patched in CA1sr in acute horizontal slices taken 24 h after treatment.

- 95 -

Patched microglial cells from control slices displayed negligible voltage-activated currents, yielding mostly linear I/V relations. 24 h after ECS, I did not detect any induction of Kv voltage-activated potassium currents or changes in I/V curves in the voltage range tested (effect of treatment group: F(1,468) = 0.165, p = 0.69, two-way

ANOVA, n = 16 and 12 cells, see summary data in Figure 12A and individual I/Vs in

Figure 12B). Analysis of linearity of the responses using a rectification index (absolute amplitude current evoked at 70 mV / at +40 mV) equally failed to yield significant differences between groups: microglia from control animals had a mean rectification index of 0.39  0.05, compared to 0.44  0.09 in cells from ECS-treated animals (t(24) =

0.829, p = 0.415, t-test, n = 15 and 11).

- 96 -

Figure 12: Electroconvulsive shock does not affect current-voltage relation in CA1sr microglia. Current density to voltage relation (I/V) in each treatment group (A) or in individual patched cells (B). Microglial I/V was unchanged after ECS. Current amplitudes were measured at steady state during 500 ms voltage steps (from Vm = 60 mV) and divided by cell capacitance to obtain current density values

- 97 -

ii. ATP-evoked currents are modulated by divalent cations and enhanced by electroconvulsive shock

Local perfusion of 1mM Na-ATP through a Y-tube application device rapidly and reproducibly evoked inward currents in cells from control slices which reversed near

0mV and were potentiated by divalent cation-free aCSF, putatively identifying them as cationic purinergic currents mediated by P2X receptors (Michel et al., 1999; Khakh et al.,

2001; Jarvis and Khakh, 2009; Yan et al., 2011).

ECS significantly increased the current density of the response under both conditions, as illustrated in the example traces in Figure 13 and as summarized in Figure

14. In normal aCSF, control cells had a mean current density of 0.152  0.052 pA/pF in response to 1 mM ATP when held at Vm = −60 mV, which was significantly smaller than the 0.596  0.073 pA/pF evoked in ECS cells (t(18) = 5.07, p = 0.000079, t-test, n = 12 and

8, Figure 14A). This increased conductance was not associated with or due to selectivity or voltage-gating differences since the evoked currents had similar reversal potential and linear I/V, as shown in the summarized ATP-evoked current data in Figure 14B.

In divalent cation-free aCSF, control cells had a mean evoked current density of

5.02  1.28 pA/pF, which was significantly smaller than the 11.6  1.62 pA/pF measured in ECS cells (t(21) = 3.23, p = 0.0041, t-test, n = 12 and 11, Figure 14C). In the same way as under normal aCSF conditions, the evoked currents had similar reversal potential and linear I/V as shown in the summarized ATP-evoked current data in Figure 14D, suggesting unaltered channel pore selectivity and gating.

- 98 -

Figure 13: Purinergic currents in CA1sr microglia potentiated by 0CaMg. Representative voltage-clamp traces showing the currents induced by local application of 1 mM [ATP] (black bar) in normal or 0CaMg aCSF (orange bar), in microglia in slices from control (left) vs. ECS (right) animals. Cells were held at Vm = −60 mV, with 500 ms ramps from −120 mV to +50 mV every 10 s.

- 99 -

Figure 14 Electroconvulsive shock-enhanced, ATP-evoked microglial cationic currents potentiated by 0CaMg. Peak current density (current amplitude/cell capacitance, pA/pF) of 1 mM ATP-evoked currents in normal (A) and divalent cation- free/0CaMg aCSF (C) at Vm = −60 mV. ECS resulted in significantly increased current densities under both recording conditions. Panels on the right show average ATP- induced current/voltage relation obtained by subtracting the I/V curve obtained before from that obtained during the ATP application under normal (B) or 0CaMg (D) conditions. As is expected for P2X-mediated nonselective cationic currents, the I/V relations are linear and reverse around 0mV. ** : p < 0.01; **** : p < 0.001.

- 100 -

iii. Electroconvulsive shock-enhanced, ATP-evoked microglial currents are

Brilliant Blue G-sensitive

Next, to identify the receptors underlying these enhanced currents, slices were preincubated with and used in the presence of, the purinergic inhibitor Brilliant Blue

G/BBG (5 μM) which preferentially blocks P2X7 over P2X4 and P2X1 (Jiang et al., 2000).

Microglia in BBG-treated slices from both control and ECS-treated animals failed to display a current response to 1 mM ATP in normal aCSF (as shown in Figure 15A, n =

6 and 11 cells from N = 2 animals/group, Kruskal-Wallis test, H(3) = 27.8, p = 0.000004).

This inhibitory effect was particularly striking in the ECS group, where I detected measurable currents in all cells incubated in normal aCSF and failed to detect any response in any cells from slices incubated with BBG. Multiple comparisons tests revealed a significant decrease in current density between cells incubated in aCSF vs. those incubated in BBG for the ECS group (Dunn’s test, p = 0.000002). Unsurprisingly given that the currents evoked in control slices incubated with aCSF were small (and only detectable in half the cells I recorded from), the responses in this group did not differ following incubation in BBG (Dunn’s test, p = 0.411). However, the proportion of cells that displayed detectable currents was clearly different after BBG treatment in both groups: 8/15 in vehicle vs. 0/10 in BBG for the controls (p = 0.0077 by Fisher’s exact test), 12/12 vs. 0/10 for the ECS-treated cells (p = 0.000002 by Fisher’s exact test).

Similarly, as summarized in Figure 15B, BBG preincubation drastically reduced the current density elicited by 1 mM ATP in 0CaMg/divalent cation-free aCSF: I recorded mean responses of only 0.156  0.073 pA/pF for 6 cells from 2 control animals and 0.92

- 101 -

 0.38 pA/pF for 10 cells from 2 ECS animals (Kruskal-Wallis test, H(3) = 29.9, p =

0.000001). Multiple comparisons tests revealed significant differences in ATP-evoked current densities from cells from control slices exposed to BBG versus those that were not BBG-treated (Dunn’s test, p = 0.0052) and cells from ECS slices exposed to BBG versus those that incubated in control aCSF (Dunn’s test, p = 0.00002).

These findings are consistent with higher affinity, larger conductance, or increased number of P2X7-containing receptors in hippocampal microglia post-ECS.

Figure 15: Electroconvulsive shock-enhanced, ATP-evoked microglial currents are mostly mediated by P2X7R. A. Peak current density (current amplitude/cell capacitance, pA/pF) of 1 mM [ATP]-evoked currents in normal aCSF at Vm = −60 mV, with or without preincubation in the specific P2X7 antagonist Brilliant Blue G (BBG). BBG significantly reduced the current density evoked by 1 mM [ATP] in normal aCSF in ECS microglia only. **** : p < 0.001. B. Peak current density (pA/pF) of 1 mM [ATP]- evoked currents in divalent cation-free/0CaMg aCSF at Vm = −60 mV, with or without BBG preincubation. BBG significantly reduced the current density evoked by 1 mM [ATP] in 0CaMg aCSF in both control and ECS microglial cells. ** : p < 0.01; **** : p < 0.001.

- 102 -

D. Summary and conclusion

The physiology of microglia in hippocampal area CA1sr was examined by whole- cell patch-clamp recordings. Activation of microglia after SE has been reported to alter cell capacitance and input resistance and to induce first outward then inward voltage- dependent currents. In the absence of any changes to these measures, ECS (like SE) potentiated the amplitude and density of the currents produced by purinergic agonists.

Biophysical and pharmacological interrogation (with divalent cation-free media and the specific antagonist BBG respectively) revealed that these enhanced currents were mediated by P2X7-containing receptors.

Again, a select subset of the SE-induced activation effects was also observed after

SE, suggesting that these particular changes in purinergic function could form part of a

“seizure signature” in murine hippocampal microglia.

- 103 -

7. EFFECTS OF ELECTROCONVULSIVE SHOCK ON MICROGLIAL PURINERGIC

GENE EXPRESSION

A. Introduction

To evaluate the simplest explanation for the enhanced microglial ATP-induced responsive motility and currents, the expression of several relevant purinergic receptor genes was studied in microglial RNA samples.

In SE models, these effects on purinergic system function were accompanied by increased expression of transcript levels for ATP receptors like the GPCRs P2Y6 and

P2Y12 and the ligand-gated ion channels P2X4 and P2X7.

B. Experimental approach

See Chapter 2 for more details on experimental procedures.

24 h after seizures, microglia were immunomagnetically isolated before being processed for RNA extraction and purification, followed by reverse transcription and quantitative multiplex polymerase chain reaction using hydrolysis probe-based assays.

The transcripts levels for P2rx1, P2rx4, P2rx7, P2ry6 and P2ry12 were quantified relative to Actb levels using the standard 2-Ct approach to determine fold changes. Expression levels of the microglial marker Tmem119 (Bennett et al., 2016) were also used for relative quantification of Actb and the relevant microglial purinoreceptors, which yielded the same results (data not shown).

- 104 -

C. Results

As summarized in Figure 16 for P2Y metabotropic receptor transcripts and in

Figure 17 for P2X ionotropic receptor transcripts, the previously described changes in microglial purinergic function were not accompanied by increases in expression of receptor mRNA, as measured by qPCR: two-way repeated measures mixed model

ANOVA analysis showed no effect of treatment group (F(1,25) = 2.213, p = 0.149), of interaction between treatment group and gene (F(4,25) = 1.377, p = 0.270) or of gene assayed (F(4,25) = 0.8903, p = 0.484). Multiple comparison analysis of each gene assayed also failed to reveal any differences, as explained below. Average threshold cycle numbers for the Actb normalizer transcript were as follows for control and ECS-exposed microglial RNA samples respectively: 16.0  0.15, 15.9  0.15.

i. Electroconvulsive shock does not affect microglial expression of P2Y metabotropic receptors

There were no statistically significant differences or trends in relative expression between control and ECS microglial samples (all evaluated by Sidak’s test for multiple comparisons) for P2ry6 (control FC = 1.05  0.15, ECS FC = 1.06  0.13, t(25) = 0.5179, p >

0.99, N = 6/group) or P2yr12 (control FC = 1.00  0.01, ECS FC = 0.83  0.16, t(25) = 1.026, p > 0.99, N = 6/group). See Figure 16 for graphically presented data. Average threshold cycle numbers were as follows for control and ECS-exposed microglial RNA samples:

P2ry6: 22.4  0.77, 21.8  0.79; P2ry12: 21.0  0.40, 20.6  0.43.

- 105 -

The lack of changes in transcript levels suggest that translational or posttranslational alterations, or changes in intracellular signaling downstream likely account for the enhanced P2Y12 receptor function that is observed in the responsive motility assays in hippocampal microglia 24h after ECS.

Figure 16: Effects of electroconvulsive shock on microglial expression of P2Y purinoreceptor genes. Relative fold change for each microglial purinergic receptor transcript assayed (the metabotropic receptors P2ry6 and P2ry12 was determined by the 2–∆∆Ct method, normalizing to Actb levels. ECS did not have any statistically significant effects on the microglial expression of the studied purinoreceptor genes.

- 106 -

ii. Electroconvulsive shock does not affect microglial expression of P2X ionotropic receptors

There were no statistically significant differences in relative expression between control and ECS microglial samples (all evaluated by Sidak’s test for multiple comparisons) for P2rx1 (control FC = 1.02  0.08, ECS FC = 1.05  0.15, t(25) = 0.2205, p >

0.99, N = 6/group), P2rx4 (control FC = 1.02  0.08, ECS FC = 1.09  0.16, t(25) = 0.4724, p

> 0.99, N = 6/group) or P2rx7 (control FC = 1.05  0.14, ECS FC = 0.66  0.15, t(25) =

2.476, p = 0.102, N = 6/group). See graphic representation of data in Figure 17. Average threshold cycle numbers were as follows for control and ECS-exposed microglial RNA samples: P2rx1: 26.2  0.72, 25.5  0.61; P2rx4: 23.1  0.24, 22.8  0.42, P2rx7: 22.2 

0.53, 22.0  0.55.

The lack of changes in transcript levels suggest that translational or posttranslational alterations, changes in intracellular signaling or changes in receptor subunit composition must account for the enhanced P2X7R-dependent currents observed in mouse hippocampal microglia 24h after ECS.

- 107 -

Figure 17: Effects of electroconvulsive shock on microglial expression of P2X purinoreceptor genes. Relative fold change for each microglial purinergic receptor transcript assayed (the ionotropic receptors P2rx1, P2rx4 and P2rx7) was determined by the 2–∆∆Ct method, normalizing to Actb levels. ECS did not have any statistically significant effects on the microglial expression of the studied purinoreceptor genes.

D. Summary and conclusion

The relative expression levels of various microglial-relevant purinergic receptor genes were quantified from RNA purified from isolated microglia 24 h sham or ECS.

Surprisingly and unlike after SE, the ECS-enhanced purinergic function described in

Chapters 5 and 6 could not be explained by increased expression of receptor mRNA as measured by qPCR.

This indicates that different mechanisms are recruited by microglia in response to ECS and SE, but with similar resulting outcomes as related to the physiology of the purinergic signaling system. The lack of changes in transcript levels suggest that translational or posttranslational alterations account for the enhanced P2Y12-dependent motility and P2X7-dependent currents post-ECS.

- 108 -

8. EFFECTS OF ELECTROCONVULSIVE SHOCK ON NEUROINFLAMMATION IN

THE HIPPOCAMPUS

A. Introduction

Since “activated” microglia are at once a result, target and source of proinflammatory molecules, the relative expression levels of two important cytokines

(TNF and CCL2) were quantified in the ECS model.

TNF is one the central pro-inflammatory mediators and it was found to be significantly upregulated after SE (Avignone et al., 2008). Interestingly, aberrantly upregulated TNF signaling is thought to underlie the pathogenic losses in blood-brain barrier integrity described after SE (Marchi et al., 2007; Librizzi et al., 2012; Kim et al.,

2013; Klement et al., 2018).

CCL2 (also known as MCP-1) is a canonically pro-inflammatory chemokine whose expression was found to be significantly upregulated after SE (Avignone et al., 2008;

Foresti et al., 2009; Arisi et al., 2015). CCL2 has additionally been mechanistically implicated in the extensive neuronal cell death that follows Status Epilepticus (Tian et al., 2017). CCL2 signaling is extensively intertwined with epilepsy, simultaneously serving as a cause and consequence of both neuronal hyperexcitability and injury (Bozzi and Caleo, 2016). Indeed, besides the connection to post-SE injury noted above, CCL2 has been implicated in the seizure-promoting effects of systemic inflammation (Cerri et al., 2016). Critically, it has also been shown to directly upregulate microglial purinergic signaling by trafficking more receptors to the plasmalemma (Toyomitsu et al., 2012).

- 109 -

B. Experimental approach

See Chapter 2 for more details on experimental procedures.

To start to characterize the neuroinflammatory milieu associated with the microglial response to ECS seizures, the expression of Tnf and Ccl2 message was quantified in microglial RNA samples and also of CCL2 protein in whole hippocampal lysate. Additionally the transcript level of Tmem119, a relatively new microglial marker

(Bennett et al., 2016) which is upregulated after SE (Bosco et al., 2018) was also determined. 24 h after seizures, microglia were immunomagnetically isolated before being processed for RNA extraction and purification, reverse transcription and quantitative multiplex polymerase chain reaction using hydrolysis probe-based assays.

The transcripts levels for the targeted genes were quantified relative to Actb levels using the standard 2-Ct approach to determine fold changes. The expression of the microglial marker Tmem119 was found to be unchanged after ECS. Since there was little variability both between and within experimental groups, Tmem119 levels were also used for relative quantification of Actb, Tnf and Ccl2, yielding similar results (data not shown).

C. Results i. Electroconvulsive shock induces increased microglial expression of Ccl2, but not

Tmem119 or Tnf

When microglial RNA samples were probed by qPCR for expression of the pro- inflammatory cytokines Ccl2 and Tnf and the microglial marker Tmem119 relative to Actb, two-way repeated measures mixed model ANOVA analysis showed significant effects of

- 110 -

treatment group (F(1,28) = 5.311, p = 0.03), interaction between treatment group and gene

(F(2,28) = 5.815, p = 0.008) as well as of gene assayed (F(2,28) = 5.156, p = 0.01), as summarized in Figure 18. Average threshold cycle values were as follows for control and

ECS-exposed microglial RNA samples respectively: Ccl2: 26.2  0.31, 25.2  0.38;

Tmem119: 21.4  0.62, 20.6  0.65, Tnf: 23.1  0.60, 22.4 0.37.

Ccl2 expression was found to be altered after ECS: Ccl2 mRNA levels were statistically significantly upregulated in ECS microglia compared to controls (control FC

= 1.07  0.18, ECS FC = 3.56  0.97, t(28) = 3.959, p = 0.001, N = 5/group).

No statistically significant differences were found in relative expression of the microglial marker Tmem119 (control FC = 1.01  0.07, ECS FC = 0.91  0.07, t(28) = 0.175, p > 0.99, N = 6/group) or the pro-inflammatory cytokine Tnf (control FC = 1.43  0.45,

ECS FC = 1.41  0.25, t(28) = 0.039, p > 0.99, N = 6/group).

- 111 -

Figure 18: Electroconvulsive shock increases expression of Ccl2 in microglia. Relative fold changes for each transcript assayed (the microglial marker antigen Tmem119 and the pro-inflammatory cytokine Tnf and chemokine Ccl2) were determined by the 2–∆∆Ct method, normalizing to Actb levels. Compared to sham controls, ECS resulted in significantly higher relative expression of Ccl2 mRNA in ECS microglial samples and no significant changes for the other two measured transcripts. **** : p < 0.001.

- 112 -

ii. Electroconvulsive shock, like status epilepticus, induces expression of CCL2 protein in the hippocampus

As shown in Figure 19, ECS and SE induced a similar increase in CCL2 protein levels as measured by ELISA in hippocampal lysates taken 24 hours after the seizures.

As expected, lysates from control animals displayed low levels of CCL2 (0.9  0.12 pg

CCL2/mg total protein; N = 16 hippocampi from 16 animals). One-way ANOVA revealed a statistically significant treatment group effect (F(2,32) = 6.13, p = 0.0056). This effect was driven by a significant increase in CCL2 in the ECS (1.6  0.18 pg CCL2/mg total protein, N = 11, q(32) = 4.48, p = 0.0091) and SE (1.5  0.27 pg CCL2/mg total protein, N =

8, q(32) = 3.61, p = 0.041) groups, as compared to the control group by Tukey’s test.

There was no significant difference in CCL2 protein expression between the ECS and the

SE groups (q(32) = 0.416, p = 0.95).

- 113 -

Figure 19: Electroconvulsive shock increases expression of CCL2 protein in hippocampus. Hippocampi were lysed 24 h after seizures and the ratios of CCL2 to total protein were determined by ELISA and BCA assays respectively. ECS and SE induced a similar statistically significant upregulation of relative CCL2 expression. * : p < 0.05.

- 114 -

D. Summary and conclusion

The relative expression levels of the transcripts for the proinflammatory genes

Ccl2 and Tnf were quantified in purified microglial RNA 24 h after sham shock or after

ECS. Levels of Tmem119, a novel microglial surface antigen with unknown function were also measured. Finally, expression of CCL2 protein was quantified relative to total protein levels in hippocampal lysates taken 24 h after control, ECS or SE treatment.

In contrast to the SE-induced neuroinflammation which comprises significantly increased expression of all three of these genes among others (Bosco et al., 2018), exposure to ECS only resulted in increased levels of Ccl2 transcript in microglia. Relative

CCL2 protein levels were elevated to a similar degree in hippocampus after ECS and SE.

- 115 -

9. DISCUSSION, FUTURE DIRECTIONS AND CONCLUSION

A. Summary of findings

In brief, I report here a complex spectrum of functional changes to hippocampal microglia in response to maximal electroconvulsive shock seizures in mice. This ECS- induced state of mild “activation” features changes that partially overlap with those described after Status Epilepticus.

ECS (unlike SE) did not cause observable differences in astrocytic morphology or density, neuronal degeneration, microglial proliferation/density, morphology, spontaneous motility, or intrinsic electrophysiological properties in the mouse hippocampus 24 h after the seizure.

On the other hand, similarly as after SE, ECS resulted in enhanced ATP-responsive motility and ATP-evoked currents in microglia, however unlike after SE, this happened in the absence of measurable changes in receptor expression.

Moreover, also like SE, ECS resulted in increased gene and protein expression of

Ccl2/CCL2, a chemokine with an established role in neuron-glia-inflammation crosstalk in healthy states and in seizures/epilepsy. However, no ECS-induced changes were detected in Tnf or Tmem119 levels (which are increased after SE).

Thus, single brief non-injurious seizures (ECS) recapitulated to a certain degree some of the facets of the microglial activation state described after prolonged, injurious seizures (SE). This partial overlap of activation patterns suggests that for a subset of features, seizure activity per se drives microglial responses.

- 116 -

B. Discussion and caveats

The commonly used chemoconvulsant models of SE-induced epileptogenesis are associated with severe damage throughout the brain and particularly in the hippocampus, above and beyond the sclerosis typically associated with human patients with TLE (Löscher, 2011; Becker, 2018). As the first line of defense in the CNS, microglia immediately respond to this and any other cellular damage (Davalos et al., 2005).

However, microglia also acutely react to the neuronal hyperactivity that underlies seizures (Eyo et al., 2014; Abiega et al., 2016). It is thus possible that the “particular” microglial activation pattern seen in mouse hippocampus after SE (Avignone et al.,

2008) is in fact a composite response to the seizure itself and the ensuing BBB and neuronal damage. My data seem to support this idea: ECS, which is neither epileptogenic nor damaging, induced only a subset of features seen after SE.

Thus, the greater pro-inflammatory activation, marked proliferation, morphological simplification, enhancement of baseline motility and changes in intrinsic electrophysiological properties observed in hippocampal microglia after SE but not after

ECS may be a response to the SE-induced neuronal damage. This is in accordance with the established pro-inflammatory role of microglia after CNS injury (Davalos et al., 2005;

Nimmerjahn et al., 2005). The reported time course and anatomical profile of the

“visible” morphological microglial response to SE lend additional credence to this correlation, since the changes in proliferation and morphology after SE are localized to areas and time periods where neuronal cell death is expected (Covolan and Mello, 2000;

Wyatt-Johnson et al., 2017).

- 117 -

While SE and ECS seizures clearly manifest differently in EEG and behavioral measures, both types of seizures have been shown to recruit the hippocampal formation as measured by induction of immediate early genes or by increased 2-deoxyglucose binding (Ingvar, 1986; Morinobu et al., 1997; Hsieh et al., 1998; Ji et al., 1998; Scorza et al., 2002; Dyrvig et al., 2012; Sinel’nikova et al., 2013). Besides having divergent outcomes both in the short and long terms, the two seizure models chosen for these experiments are obviously also of very different durations. It would be interesting to learn whether repeated ECS with a similar seizure burden as SE would result in a more pro-inflammatory microglial activation pattern. Indeed, ECT in the clinic is only useful if given at least semi-chronically, so these additional data would also inform how my current study fits into the beneficial sequelae of seizures. In my opinion, it is likely that the effect of ECS repetition on microglial activation will depend entirely on the duration of the intertrial period. If the ECS-induced microglial response subsides after 5-7 days

(as is the case after SE), I would expect no cumulative effects from less than weekly repeated stimulation. ECT is given on average once or twice a month (Fink, 2009), meaning that my data could be a good model for the microglial response to ECT in human patients. It is important to recall, however, that the wide-reaching neuroprotective effect of ECS as described in rodents requires seizures to be clustered more closely in time, with frequencies ranging from daily to three times a day (Masco et al., 1999; Kondratyev et al., 2001).

SE models relying on systemic or local chemoconvulsant application could additionally be having direct drug effects on the microglia, although single microglia do not express kainate or muscarinic receptors (Hammond et al., 2019). This potential

- 118 -

confound is mostly mitigated by the finding that different models of SE induction are associated with similar ensuing microglial activation (Chen et al., 2005; Avignone et al.,

2008; Menteyne et al., 2009; Eyo et al., 2014; Arisi et al., 2015; Avignone et al., 2015; Eyo et al., 2016; Schartz et al., 2016; Tian et al., 2017; Wyatt-Johnson et al., 2017; Feng et al.,

2019), suggesting that the seizure and damage likely drive the response as opposed to the drug.

My data also show that the hippocampal microglial responses to ECS and SE also have characteristics in common. Both models similarly enhanced the velocity motility of microglial processes in response to a local source of ATP in slices, a sensitive functional assay of P2Y12 receptor signaling in microglia, since this is the receptor underlying the process extensions (Davalos et al., 2005; Haynes et al., 2006; Ohsawa et al., 2010). Since the ECS-induced increase in velocity was not detected in response to saturating ATP concentrations, I concluded that changes in number or function of purinergic receptors in microglia underlie these changes in physiology, as opposed to differences in ambient

ATP levels or ATP-induced ATP release from astrocytes. This is notable given that multiple groups have reported increased ATP levels after brain stimulation and seizures

(Wu and Phillis, 1978; Wieraszko et al., 1989; Beamer et al., 2019). Paradoxically, adenosine, which mediates microglial process retraction (Orr et al., 2009), is also increased in the brain after seizures (Ilie et al., 2012; Lovatt et al., 2012). Interestingly, my whole-cell patch-clamp experiments on microglia in hippocampal slices from ECS- exposed animals also revealed a greater P2X current amplitude and density, chiefly mediated by P2X7-containing receptors. These ATP-gated ion channels possess biophysical characteristics like calcium permeability and conductance

- 119 -

sensitization/pore size increase (Young et al., 2007; Liang et al., 2015) that likely mediate their well-established roles in epilepsy and neuroinflammation (Henshall and

Engel, 2015; Amhaoul et al., 2016; Beamer et al., 2017). Although microglia display variable and complex rectification with unknown pharmacology and biophysics at strongly hyperpolarized or depolarized potentials, at rest they normally have linear responses at physiological potentials. Thus, even if whole current traces may show a trend towards changes in rectification, there may be little physiological relevance since microglia tend to adopt a limited range of membrane potentials. More work is needed to elucidate the mechanisms, significance and potential role of this rectification at extreme voltages. Since P2Y12-directed motility in response to ATP is the first phase of the microglial response to neuronal injury and P2X7 receptors have a role in the ensuing neuroinflammation, it seems that the seizures in either of the studied models are priming microglia to mount both a faster and stronger response to future insults.

I failed to detect any differences in passive properties (resting membrane potential/RMP, input resistance/IR or membrane capacitance/Cm) or in voltage- activated currents following ECS. These data are consistent with the lack of an effect on morphology and baseline motility, since IR and Cm directly reflect cell membrane properties and changes in voltage-activated potassium currents underlie pathogenic microglial-neuron contacts resulting in cell death after SE (Fordyce et al., 2005;

Menteyne et al., 2009). In this study, some cells expressed greater current densities than the tightly packed average population, both under hyper- and depolarizing potentials. It would be interesting to test whether these cells are expressing delayed and/or inward rectifying potassium channels through pharmacology and single-cell qPCR.

- 120 -

Thus, my results point to parallel enhancements in metabotropic and ionotropic purinergic signaling within hippocampal microglia after ECS, in the absence of changes in morphology, baseline motility or intrinsic electrophysiological characteristics. In this way, one can reimagine the complex SE-induced activation state as comprising a particular response to seizures, as well as a parallel response to neuronal injury.

CCL2 (C-C motif chemokine ligand 2, also known as monocyte chemoattractant protein 1 or MCP-1) is a canonically pro-inflammatory signaling molecule that has been strongly implicated in the post-SE neuronal injury (Foresti et al., 2009; Arisi et al., 2015;

Bozzi and Caleo, 2016; Tian et al., 2017) as well as in the seizure-enhancing effects of systemic inflammation (Cerri et al., 2016). Unexpectedly, given the lack of neuronal damage (and epileptogenesis) after ECS as compared to SE, significantly upregulated expression of Ccl2 mRNA in microglia was found after ECS. This resulted in similarly increased CCL2 protein expression in hippocampal lysates from ECS-exposed and SE- exposed animals. This change was not accompanied by increased levels of Tnf mRNA, another pro-inflammatory cytokine whose expression is increased post-SE (Avignone et al., 2008). Interestingly, upregulated TNF signaling is thought to underlie the pathogenic losses in blood-brain barrier integrity described after SE (Marchi et al., 2007; Kim et al.,

2013; Noé et al., 2016; Klement et al., 2018). Although the protein experiments were performed on hippocampal lysates obtained from rapid decapitation without transcardial perfusion, where contamination from blood could cloud my analysis, the qPCR data is derived from MACS-purified microglial samples from perfused animals, strengthening my conclusions.

- 121 -

SE, inflammation and subsequent injury/epileptogenesis are intricately linked through at least two related signaling pathways: fractalkine and interleukin 1-beta.

Neuronal-microglial fractalkine signaling and subsequent astrocytic and neuronal IL-1R activation have been extensively implicated in the pathogenesis mechanism observed after SE (Ravizza and Vezzani, 2006; Ali et al., 2015; Arisi et al., 2015; Eyo et al., 2016;

Tian et al., 2017), as well as in other models of acquired epilepsies in rodents like kindling (Plata-Salamán et al., 2000).

Besides complicating the picture as far as the role of CCL2 in neuronal injury post-SE, my novel finding of increased hippocampal CCL2/microglial Ccl2 after ECS is of particular interest given the fact that this chemotactic cytokine has been found to directly enhance purinergic signaling in microglia by promoting trafficking of receptors like P2X4 to the microglial plasmalemma (Toyomitsu et al., 2012); whether CCL2 influences trafficking of other receptors such as P2X7 is unknown. However, by some reports, microglial as well as neuronal P2X7 levels are increased in TLE patients and rodent SE models (Jimenez‐ Pacheco et al., 2013), while transient inhibition of P2X7 resulted in lasting decreases of post-SE neurodegeneration, gliosis and epileptogenesis

(Engel et al., 2012; Jimenez-Mateos et al., 2015; Jimenez-Pacheco et al., 2016).

Unlike the observed increase in CCL2 protein levels which is correlated with an increase in Ccl2 mRNA abundance, the increases in purinergic receptor function could not be explained by changes in microglial gene expression. Indeed, I observed no significant gross differences in microglial expression of P2rx1, P2rx4, P2rx7, P2ry6 and

P2ry12 mRNAs post-ECS. Interestingly, the distribution (but not average) of P2yr12 message expression was clearly affected: whereas naïve and control mice all clustered

- 122 -

around a certain relative expression level, microglia from ECS-exposed mice showed a much wider distribution. Under normal conditions, P2Y12 expression changes linearly with surface area, which explains the tight clustering seen in control mice. Thus, these data can be interpreted as a dysregulation of expression of a transcript that is normally under strict regulation. On the other hand, the relative expression levels of Tmem119, another microglial cell surface marker, were similarly tightly clustered in both experimental groups. To rule out a potential confound from normalizing to actin, the analysis was repeated by comparing expression of every purinergic/inflammatory target gene to levels of Tmem119 (this yielded no differences for any transcript). Single cell gene expression data would be needed to unequivocally rule out the possibility that only certain individual microglial cells or subpopulations of microglia possess more transcripts for this receptor. While RT-qPCR was the right approach for a first pass, this type of analysis relies solely on relative quantification of message, and thus cannot detect other causes of increased receptor activity. For example, there could be an existing readily available pool of mRNA from which to transcribe new protein. Similarly, alternative splicing or of transcripts could be involved, as could the machinery that degrades existing mRNA. Any of those possibilities could explain the observed physiological effect even in the absence of measurable differences in gene expression.

Notably, while SE models seem to robustly change the number, morphology and purinergic physiology of hippocampal microglia, the effect of SE on purinergic gene expression has incited controversy in the field and seems to depend not only on timing, strain, age and model, but also on laboratory. For instance, different groups have reported unchanged (Bosco et al., 2018), decreased (Alves et al., 2017) and increased

- 123 -

(Avignone et al., 2008) expression of purinergic receptor gene P2ry12 in the latent phase of similar models of SE. Likewise, the transcript for P2rx7 has been reported as both being increased (Avignone et al., 2008; Jimenez-Pacheco et al., 2016) and unaffected

(Bosco et al., 2018) post-SE. Thus, the direction and even existence of an SE-induced effect on expression of purinergic receptor genes remains controversial. While a change in microglial expression of P2rx7 would have been the most parsimonious explanation for the increased BBG-sensitive/P2X7R-dependent current density, several possibilities, including biophysical and/or pharmacological changes in the P2X7 channel properties could explain these findings and will have to be further investigated.

SE models have almost universally been employed in the search of putatively pro- epileptogenic changes that may trigger chronic and spontaneously recurrent seizures.

Considering my results, the findings from previous studies on SE-induced microglial activation must be adjusted to account for the “baseline” response pattern to neuronal hyperactivity in the absence of overt neuronal damage and of eventual epileptogenesis.

In this way, the purinergic changes I describe in hippocampal microglia likely do no

Other published work has studied the microglial response after ECS (Jinno and

Kosaka, 2008; Jansson et al., 2009), but only found significant changes in microglial density and morphological/functional activation after chronic ECS stimulation (10 to 30 seizures). Concordant with this present study, one of these previous studies found that single ECS seizures failed to elicit changes in number or morphology of microglia (Jinno and Kosaka, 2008).

Since people suffering from epilepsy normally present with single unprovoked seizures as opposed to status epilepticus (Hauser et al., 1991; Hesdorffer et al., 1998;

- 124 -

Olafsson and Hauser, 1999; Banerjee et al., 2009), my studies with ECS may model human seizures with higher validity than SE or chronic stimulation protocols

(unsurprising since by design, these model epileptogenesis rather than acute seizures).

It remains unclear for now what the role of microglial changes is in deciding the differential outcomes after the seizures from each model. Since ECS in rodents is a near- perfectly valid model for ECT in humans, my data may additionally point to a potential role in ECT’s effects, although ECT in the clinic is administered chronically (Perera et al.,

2004; Fink, 2009) and to date no mood-stabilizing effects have been shown after a single

ECT seizure. Further research (including chronic treatments and experiments on mouse models of depression) are needed to elucidate whether the observed microglial response extends to ECT as used in the clinic.

The mice used in this study were all of age corresponding to a “young adult” life stage: P30-P45 in mice corresponds approximately to the adolescent/teenage years and early 20s in humans (Flurkey et al., 2007; Dutta and Sengupta, 2016). At these ages the

BBB (Sweeney et al., 2019), brain circuits (Agoston, 2017) and microglia (Ginhoux et al.,

2010; Thion and Garel, 2017; Thion et al., 2018) are thought to be mature, effectively negating any potential interaction effects of developmental processes with seizures. In the clinic, however, the epilepsies have much greater prevalence in pediatric and geriatric populations (Hauser et al., 1991; Fiest et al., 2017; Ip et al., 2018). While it is certainly true that these models may lack face validity insofar as their relative age during the seizures, I chose this age range based on their widely recognized predictive validity against human convulsive seizures (for ECS) and human TLE (for SE) (Löscher,

2011). Moreover, it is important to establish a “baseline” response to neuronal

- 125 -

hyperactivity in the absence of changes due to development or aging. Mice of these ages also make for an ECT model with almost perfect face, construct and predictive validity

(Jonckheere et al., 2018; Alemu et al., 2019).

In contrast to my studies using ECS and SE on these “young adult” mice, preliminary data from our lab suggests that exposure to similar seizure burdens during development (P7-P10) does not result in similar observable activation of microglial morphology or physiology. Interestingly, rodent SE models oftentimes cause less cellular damage and are less epileptogenic during those time frames (Oakley et al., 2009;

Buckmaster and Haney, 2012), perhaps from the increased potential for plasticity inherent to developmental periods. Studying the effects of age on these microglial

“behaviors” together with the interaction of developmental and/or aging processes with the inflammatory and epileptogenic responses elicited by SE will be necessary to fully elucidate their interplay. These data would surely reveal even more meaningful insights into how my current findings in mouse hippocampal microglia relate to seizures, epilepsy and ECT in human populations.

After backcrossing the CX3CR1-GFP allele, I elected to conduct all my studies on mice with a congenic C57Bl/6 background, since this is the most popular inbred mouse strain used in neurobiology in general and in the niche of SE-induced microglial responses specifically. Relative to other commonly used lines like DBA, 129S or BALB/c, and particularly compared to the vulnerable FVB/N line, this strain is more resistant to seizures, SE and excitotoxic damage (Schauwecker and Steward, 1997; Schauwecker,

2003, 2012). As such, it would be interesting to test whether the microglial changes I describe after ECS (partially overlapping with what others have described after SE)

- 126 -

translate to other genetic backgrounds. Since these other mouse lines generally have lower seizure thresholds (Frankel et al., 2001; Ferraro et al., 2002) and are more prone to excitotoxicity (Schauwecker and Steward, 1997; Schauwecker, 2003), I would suspect that the SE-induced hippocampal microglial response is even more tilted towards a classical/pro-inflammatory/M1 activation. Interestingly, vulnerability to seizure- induced damage only partially correlates with epileptogenesis (Avanzini et al., 2014;

Dingledine et al., 2014), suggesting that more complex causal interactions are at play. It is also entirely plausible that the ECS model I used would result in more stress and damage in these more vulnerable mouse lines, unlike what I have shown for mice on a

C57Bl/6 background.

A potential confound that is often raised regarding the study of microglia in acute brain slices is that the slicing process itself may alter microglial state or responses.

Following others (Avignone et al., 2008; Menteyne et al., 2009), I employed the protective cold sucrose slicing technique and used slices within five hours or less after slicing. Slices from control and ECS animals deteriorated in a similar fashion, as approximated by the ease of obtaining patch-clamp recordings from ramified GFP+ cells deep in the slice and the quality of those recordings, or visually through confocal imaging. Moreover, my findings with SE mirror those previously reported by others using multiple slice preparation approaches. It does remain possible that ECS or SE makes neurons more vulnerable to the slicing protocol, meaning that some of my observations in slice could arise from a seizure/slicing interaction. Imaging microglial motility in vivo would be the ideal solution, but this is very difficult to accomplish given the fine resolution needed to see the processes and the anatomical depth of the

- 127 -

hippocampus. Additionally, the genetic line that I used to identify microglia requires at least a partial knockdown of the native CX3CR1 allele, which can have independent and interactive effects on microglial biology (Jung et al., 2000; Wolf et al., 2013; Zhan et al.,

2014). Imaging microglial motility in the recently created and soon to be available

Tmem119-GFP mice could address this concern (Kaiser and Feng, 2019).

Instead of using FluoroJade C/FJC to quantify neuronal cell death, the number of remaining neurons could have been counted using Nissl stains or NeuN immunolabeling.

Neuronal viability could also be studied directly by exploiting the live-cell dye calcein-

AM (the -acetoxymethyl ester easily enters every cell, but only live cells will possess the esterase activity that enables calcein fluorescence). Comparing DAPI staining to

Propidium Iodide/PI staining without permeabilization would serve to directly probe for structural damage of cell and nuclear membranes (under these conditions, DAPI enters and marks all cell nuclei, whereas PI only marks cells with compromised membrane integrity). Other facets of neurodegeneration could also be detected through

TUNEL staining (to quantify DNA damage) or by measuring levels of components of the apoptotic signaling cascade (like activated Caspase 3) through western blot or ELISA.

Although FJC is a well-established marker of neurodegeneration in the mouse brain

(Schmued et al., 2005), the possibility nevertheless remains that ECS caused reversible and thus undetectable stress on hippocampal neurons. To better address this question, I could have looked at signs of healthy neuronal metabolism, and even confirm a lack of oxidative or other cellular stressors by evaluating their effects at the molecular level.

From another approach, I could have harnessed the patch-clamp method to measure membrane properties of individual hippocampal neurons and then dye-fill them. Images

- 128 -

of these cells could be examined for subtler signs of damage such as dendrite beading indicative of mitochondrial dysfunction (Greenwood et al., 2007), or the axonal swelling that is characteristic of traumatic brain injuries (Foda and Marmarou, 1994).

This study could also be improved upon by using other methods besides staining with sulforhodamine 101/SR101 to study astrocytic density and morphology, such as by exploiting existing transgenic lines that drive fluorescent reporter proteins from astrocyte-specific promoters such as Aldh1l1 or GFAP (Srinivasan et al., 2016). Although most published uses of SR101 have focused on other brain areas (namely cortex), the mechanisms underlying the selective uptake of the dye by astrocytes have been mostly studied in hippocampus (Schnell et al., 2012). Indeed, it is now well-accepted that multispecific thyroid hormone transporter (OATP1C1) activity mediates the astrocyte- specific staining of SR101, at least in the hippocampus of naïve adult mice (Schnell et al.,

2015).

Additionally, I could have used biolistic labeling (in vitro) or viral vectors (in vivo) encoding fluorescent reporters under the control of the synthetic promoter GfaABC1D, which has even greater selectivity for astrocytes than the prototypical GFAP promoter

(Shigetomi et al., 2013). Sparse labeling through this approach also enables imaging of fine astrocytic processes and leaflets, which is impossible in a fully labeled sample given the overlapping fluorescent signals and the minute (sub diffraction-threshold) size of their processes. Similarly, individual astrocytes can be studied after patch-clamp and cell filling, for example with neurobiotin. By combining small dyes (like the AlexaFluor series) and large marker molecules (like biotins) in the intracellular solution, one could even study the gap-junction coupling between astrocytes which has been reported as

- 129 -

impaired in the epilepsies (Oberheim et al., 2008; Tian et al., 2005) and also after SE

(Plata et al., 2018).

It has also recently been reported that SR101 can induce ictal activity

(Rasmussen et al., 2016). This would certainly be a confound during the microglial imaging studies, as seizure-like events, neuronal hyperactivity and excessive glutamatergic transmission can all influence microglial motility (Eyo et al., 2014).

However, these proconvulsive effects of SR101 were only detectable in vivo at concentrations roughly two orders of magnitude greater than what was used for staining slices in my experiments (100-250 μM vs. 1.8 μM).

For the P2X7R inhibition studies, it would have been more rigorous to use the pharmacologically inert dye Coomassie Brilliant Blue R(ed) as a control, as it shares many other characteristics with the BBG/Coomassie Brilliant Blue G(reen) dye that was used to selectively block P2X7-containing receptors (Jiang et al., 2000). I also could have used the prototypical agonist BzATP to evoke currents. This modified purine is selective for P2X7 over other P2X subunits and is 5-10 fold more potent than ATP (Young et al.,

2007). Since BBG has also been shown to inhibit Pannexin1, a hemichannel with well- established roles in ATP release, it would have been advantageous to confirm the role of

P2X7 by using the Pannexin1-selective antagonist FD&C Blue No. 1 (Wang et al., 2013).

Should it have no effect on measured currents, Pannexin1 can be ruled out from the list of potential mechanisms underlying the observed effect. Conversely, if FD&C Blue No. 1 inhibits the evoked currents (as was seen for BBG), it is likely that Pannexin1 activity partially underlies the 0CaMg-enhanced purinergic currents seen after ECS. The causality here remains very unclear, given that P2X7 activity can regulate its own

- 130 -

expression and localization, and increase [ATP] either directly or through Pannexin1 recruitment (Engel et al., 2017). The irreversible competitive P2X7-selective inhibitor oxidized ATP also comes with its own set of challenges, as it is a fragile and “sticky” molecule. ATP and oxidized ATP have also been shown to sterically inhibit Pannexin1 activity by “clogging” the pore (Qiu and Dahl, 2009; Qiu et al., 2012), which clouds the picture even further.

Along a similar vein, I could also have used clopidogrel, an FDA-approved antiplatelet medication, to block P2Y12 receptors (Herbert and Savi, 2003) during the responsive motility imaging studies and confirm their canonically assumed role

(Davalos et al., 2005; Haynes et al., 2006; Irino et al., 2008; Ohsawa et al., 2010). I could also have used the prototypical agonist 2MeSADP to further identify the receptors underlying the enhanced response (Dorsam and Kunapuli, 2004; Avignone et al., 2008;

Zhang et al., 2014).

C. Future directions: Microglial response to electroconvulsive shock and…

Going forward, this research could be taken a variety of different directions. I briefly outline some of the most pressing follow-up questions and how I would approach them below. Beyond potentially uncovering new biomarkers and treatment targets for seizure-induced epileptogenesis, these experiments could also reveal novel microglial roles underlying effects of ECT as used in the clinic. Besides the topics featured below, this study would also benefit from more comprehensive morphometry of microglial processes, for instance to quantify and compare the thickness/diameter of processes

- 131 -

after ECS. Moreover, since gene expression as measured by qPCR did not explain the observed effects, it is of utmost interest to complement my results by dissecting each possible stage of regulation for the relevant purinergic receptors by studying things like: lifetime of the transcripts, potential alternative RNA splicing or protein maturation/folding, post-translational modifications that would impact receptor trafficking and/or function, or cellular localization. By combining techniques such as

RNAseq, masspec, surface biotinylation assays and imaging/physiology, the causal mechanisms underlying the changes I have described herein.

i. The role of fractalkine signaling

The fractalkine pathway (CX3CL1-CX3CR1) is well recognized as an avenue for neuron-to-microglia communication, with critical roles in development, health and pathology. Homozygous carriers of the CX3CR1-GFP reporter allele are in fact knockouts

(KOs) for the receptor (Jung et al., 2000) and have been extensively used to study the role of fractalkine signaling in various conditions. These constitutive KO mice have deficient synaptic pruning during ontogeny (Paolicelli et al., 2011) which results in long- lasting anatomical, physiological and behavioral differences (Corona et al., 2010; Zhan et al., 2014). The most notable deficits in these mice are in social behavior, which is specially intriguing given the fact that Cx3cr1 polymorphisms are associated with increased risk of schizophrenia and autism spectrum disorders in humans (Ishizuka et al., 2017). Mice lacking this receptor are also protected from neuron loss in models of

Alzheimer’s (Fuhrmann et al., 2010), although further experiments are needed to tease

- 132 -

apart whether this really reflects a role for CX3CR1 in neurodegeneration or if this protection is a result of the developmental effects or of life-long “disability” from the differences in social behavior.

CX3CR1 is also intricately intertwined with epilepsy and the microglial response to SE. When compared to wild-type (WT) controls, KOs have decreased seizure thresholds and display more severe and damaging SE after similar chemoconvulsant dose (Eyo et al., 2016). Importantly, this same study showed that the heterozygous

CX3CR1GFP/+ animals I used for my studies did not differ from WT controls in their reaction to SE.

As such, I can envision relatively simple experiments that test whether the effects

I have previously described are dependent on fractalkine signaling. The main methodology issue is that different ECS stimuli will probably be necessary to safely and reliably elicit maximal seizures from the KOs. Like the AD study above, there would also be potential confounds from the constitutive and unselective nature of the knockout, which could be partially addressed by parallel experiments with RNAi in heterozygous mice, genetic rescue in KO mice or creating inducible KOs.

ii. The role of microglial [Ca2+]i

Like most mammalian cells, microglia employ calcium as a 2nd messenger and thus need to keep fluctuations in intracellular calcium concentrations under strict control. Calcium imaging of microglia in vivo or in slice has proven to be very technically demanding, and adequate tools have only become publicly available recently (Gee et al.,

- 133 -

2014). Given their unique physiology and pharmacology, relatively little is known about the causes and effects of “calcium spikes” that occur in resting or responding microglia.

However, it has been shown that microglia rapidly respond to individual damaged neurons by transiently increasing [Ca2+]i (Eichhoff et al., 2011). More recently healthy aging has also been associated with dysfunctions in microglial calcium signaling

(Olmedillas del Moral et al., 2019). Of particular interest in the context of the present study is the finding that treatment with LPS (pro-inflammatory) and bicuculline (pro- excitatory) synergistically increase the frequency of microglial calcium spikes, and in fact co-treatment with both even introduces novel phenomenon of coordinated spiking or “calcium waves” (Pozner et al., 2015). It is of primordial importance going forward to study the effect of SE and ECS on microglial calcium dynamics (spikes and waves), and particularly to investigate how they may relate to the changes that others have described after SE and that I have described here after ECS.

iii. The role of astrocytes

Astrocytes interact extensively with microglia and neurons and have well established roles in seizures and epilepsy, for example through potassium buffering and glutamatergic signaling (reviewed in Tian et al., 2005). Microglial activation increases modulation of synaptic transmission by astrocytes (Pascual et al., 2012), Astrocytes can contribute to epileptogenesis in models of TLE by releasing neurotransmitters and regulating neuronal NMDA receptors (Clasadonte et al., 2013). After SE, astrocytes become atrophic and dysfunctional (Wilcox et al., 2015) and seemingly inhibit synaptic

- 134 -

plasticity chronically through reduced Ca2+ activity (Plata et al., 2018). Astrocytic Ca2+ waves have been posited as master regulators of brain physiology (Scemes and Giaume,

2006) and they can have impactful roles in regulating other CNS cells. In neurons, they can increase the frequency of spontaneous and mini glutamatergic synaptic currents

(Fiacco and McCarthy, 2004). Their occurrence has also been found to incite complex purinergic responses in microglia (Schipke et al., 2002).

I can imagine at least three different non mutually exclusive mechanisms through which astrocytes could have a role in the microglial response to SE and ECS. First and foremost, it is still unclear if the astrocytic changes observed after SE are due to hyperactivity, neuronal injury, or their interaction. Thus, astrocytes could have a role in the unique microglial activation induced by SE simply through their response to neuron or BBB damage. I could address this by studying the astrocytic response to non- damaging seizures in more detail than I have here. It would be particularly interesting to learn if astrocytic physiology (for instance, neurotransmitter uptake, K+ buffering, Ca2+ activity or purinergic signaling) is changed by the paroxysmal/ictal activity itself.

Similarly, since astrocytic number and morphology do not correlate with astrocytic

“activation states” (A1/A2), it remains possible that hippocampal astrocytes are activated after ECS, albeit along a different activation “path” as after SE. A technically simple way to approach this question would be to immunomagnetically separate astrocytes (for example by using microbeads with antibodies against GLAST1) and then probe for markers of activation at the gene expression level by qPCR or ideally RNAseq, or at the protein level through western blot or ELISA. In my ECS model, in the absence of overt neurodegeneration (as shown in Figure 1) or measurable changes in microglial

- 135 -

morphology (as shown in Figure 4 and Figure 5), I would not expect astrocytes to be significantly activated, although my own present data prove that glia are capable of important physiological responses happening without any overt morphological changes.

Astrocytes could also be involved in the acute response that sensitizes microglial purinergic signaling during and immediately after the seizure, perhaps by amplifying excessive ATP release from the injury as is normal, or by amplifying “signaling levels” of ambient ATP aberrantly. This could be tested by inhibiting this signaling in astrocytes in vivo during that epoch and measuring the effect on the microglial response to different seizure models, or by infusing ATP or dialyzed apyrase intracerebrally.

Similarly, ATP-induced ATP release mechanisms in astrocytes could be potentiated at the 24 h timepoint such that I observe functionally increased purinergic signaling in microglia. Following the same logic as above, these mechanisms could be inhibited during the slice experiments to probe for astrocytic-dependent microglial effects of seizure activity.

iv. Its effects on seizure propensity and sequelae

Given that CX3CR1 and P2Y12 KOs have reduced seizure thresholds, more severe

SE and reduced SE-induced potentiation of microglial purinergic signaling (Eyo et al.,

2014, 2016), it has been hypothesized that the microglial phenotype observed after SE could be anticonvulsive, neuroprotective or antiepileptogenic (Hiragi et al., 2018). On the other hand, since a postnatal peak in microglial development correlates with increased sensitivity to seizure triggers, it has also been postulated that some actions of

- 136 -

these cells can be proconvulsive (Kim et al., 2015). As is usually the case, the answer probably lies somewhere in between, and microglia likely have divergent context- dependent roles in both seizures and epilepsy.

Although the effects of seizures on acute excitability are well characterized in the long and short term for different models, it is unknown if single electroconvulsive shocks like the ones I applied have an effect on seizure thresholds at the 24 h timepoint at which microglia were affected. It would be interesting to directly test this by measuring the CC50 for maximal seizures at this stage. By manipulating the microglial activation after ECS1 and seeing if it modulates the effect of ECS on the threshold for a

2nd ECS 24 h after, one could finally assign a putatively pro-or anti-convulsive role to these microglial changes. This could also be tangentially addressed by depleting microglia through pharmacogenetic (using the Cd11b-TK mice introduced in Heppner et al., 2005) or pharmacological (exploiting CSF1R inhibition as first used in Elmore et al.,

2014) approaches.

The effects of multiple seizures could also be studied to determine whether these microglial changes are maintained constant or summative after multiple ECS, or whether they are specific to the first seizure. Lastly, it would be interesting to test for microglial roles in any changes induced by multiple ECS such as neuroprotection from SE or reversal of behavioral alterations.

1 For example: by inhibiting purinoreceptors, by using fractalkine KOs if they do have less microglial activation after ECS, or by using minocycline which inhibits pro-inflammatory polarization (Kobayashi et al., 2013).

- 137 -

v. Its potential role in the beneficial or deleterious effects of ECT

Finally, along a similar line of thinking as the above experiments, one could evaluate the potential for microglial roles in the well-known effects of ECT. For instance, after inducing rodent models of anxiety and “depression”-like chronic unpredictable stress (Monleon et al., 1995; Nollet et al., 2013), ECS could be chronically administered with and without microglial manipulations as previously mentioned and the effect measured on these models with recognized face, construct and predictive validity.

D. Conclusion

I have described herein a state of microglial “activation” in the mouse hippocampal area CA1sr one day after a single electroconvulsive shock-evoked seizure.

Surprisingly, the observed microglial changes partially overlapped in function with those described after epileptogenic Status Epilepticus seizures. I posit that the changes present in the response to both models could represent a “signature” of maximal seizures in hippocampal microglia, with little sensitivity to the degree of damage or ensuing epileptogenesis. These microglial responses are characterized by potentiation of purinergic signaling and result in increased responsive process motility.

Although the exact causes, mechanisms and consequences of this motility behavior are still not completely understood, a trio of non-peer-reviewed publications that surfaced on the biorxiv preprint server in the last months point to its importance. In the first such study (submitted to the server in January 2019), Martin Fuhrmann’s group at DZNE posits that “Microglial motility depends on neuronal activity and promotes

- 138 -

structural plasticity in the hippocampus” (Nebeling et al., 2019). This observation is in accordance with my findings of potentiated motility after even brief periods of neuronal hyperactivity. Conversely, in their paper uploaded February 21st 2019, Ania Majewska’s team at the University of Rochester show that “Noradrenergic signaling in wakeful states inhibits microglial surveillance and synaptic plasticity in the mouse visual cortex”

(Stowell et al., 2019). One day later, on February 22nd 2019, the laboratory of Long-Jun

Wu at the Mayo Clinic submitted their preprint entitled “Neuronal network activity controls microglial process surveillance in awake mice via norepinephrine signaling”

(Liu et al., 2019). Their combined results raise additional caveats that must be applied to slice and culture studies of microglial motility, since noradrenergic neurotransmission is either impaired or absent in these preparations.

Repeated ECS in mice is a near-perfect model of electroconvulsive therapy (ECT) in humans. Thus, my results also set the stage for a potential role of microglial changes in mediating either ECT’s established benefits (anti-depressant/mood-stabilizing) or its equally well-known deleterious effects (confusion or amnesia). This new knowledge further calls for renewed consideration of the importance of using non-damaging and non-epileptogenic seizure models as a “baseline” to tease out the relative contribution of epileptogenic seizures (and thus epileptogenic processes) vs. that of paroxysmal activity per se. This is important in every case but particularly for studies into microglia, neuroinflammation or neuroimmunology, as activity, injury and their interaction differentially recruit inflammatory responses acutely. These data also prompt numerous avenues of future research to continue resolving the crux of causes and effects of seizure-induced microglial activation

- 139 -

APPENDIX I: EXTENDED DATA

Injection of AAV5 encoding GfaABC1D-tdTomato into CX3CR1GFP/+ hippocampus. GFP+ microglia are shown in green and infected astrocytes are shown in red.

- 140 -

A. 2D-projected and binarized 3D tracings of microglia in mouse CA1sr

The following cells were manually reconstructed using the Filament Tracer module in Imaris (Bitplane) for analysis published in Figure 3 of Sepulveda-Rodriguez et al. (2019), which was modified into Figure 4 and Figure 5 of this dissertation.

Please refer to Chapter 2, section G “Immunostaining and microglial morphometry” for further methodology details. The 3D reconstruction files (.ims) for the filament tree of each cell shown here will be made available online through

NeuroMorpho.org.

- 141 -

i. Under control conditions

Scalebars: 10 m

- 142 -

Scalebars: 10 m

- 143 -

Scalebars: 10 m

- 144 -

Scalebars: 10 m

- 145 -

Scalebars: 10 m

- 146 -

Scalebars: 10 m

- 147 -

Scalebars: 10 m

- 148 -

ii. 24 hours after electroconvulsive shock

Scalebars: 10 m

- 149 -

Scalebars: 10 m

- 150 -

Scalebars: 10 m

- 151 -

Scalebars: 10 m

- 152 -

Scalebars: 10 m

- 153 -

Scalebars: 10 m

- 154 -

Scalebars: 10 m

- 155 -

iii. 24 hours after status epilepticus

Scalebars: 10 m

- 156 -

Scalebars: 10 m

- 157 -

Scalebars: 10 m

- 158 -

Scalebars: 10 m

- 159 -

Scalebars: 10 m

- 160 -

REFERENCES

Abiega O et al. (2016) Neuronal Hyperactivity Disturbs ATP Microgradients, Impairs Microglial Motility, and Reduces Phagocytic Receptor Expression Triggering Apoptosis/Microglial Phagocytosis Uncoupling. PLoS Biol 14:e1002466. Adriouch S, Dox C, Welge V, Seman M, Koch-Nolte F, Haag F (2002) Cutting Edge: A Natural P451L Mutation in the Cytoplasmic Domain Impairs the Function of the Mouse P2X7 Receptor. J Immunol 169:4108–4112. Agoston DV (2017) How to Translate Time? The Temporal Aspect of Human and Rodent Biology. Front Neurol 8. Aguilar BL, Malkova L, N’Gouemo P, Forcelli PA (2018) Genetically Epilepsy-Prone Rats Display Anxiety-Like Behaviors and Neuropsychiatric Comorbidities of Epilepsy. Front Neurol 9:476. Akiyama H, Tooyama I, Kondo H, Ikeda K, Kimura H, McGeer EG, McGeer PL (1994) Early response of brain resident microglia to kainic acid-induced hippocampal lesions. Brain Res 635:257–268. Alemu JL, Elberling F, Azam B, Pakkenberg B, Olesen M (2019) Electroconvulsive treatment prevents chronic restraint stress-induced atrophy of the hippocampal formation-A stereological study. Brain Behav 9:e01195. Ali I, Chugh D, Ekdahl CT (2015) Role of fractalkine-CX3CR1 pathway in seizure-induced microglial activation, neurodegeneration, and neuroblast production in the adult rat brain. Neurobiol Dis 74:194–203. Al-Muhtasib N, Sepulveda-Rodriguez A, Vicini S, Forcelli PA (2018) Neonatal phenobarbital exposure disrupts GABAergic synaptic maturation in rat CA1 neurons. Epilepsia 59:333–344. Alves M, Gomez-Villafuertes R, Delanty N, Farrell MA, O’Brien DF, Miras-Portugal MT, Hernandez MD, Henshall DC, Engel T (2017) Expression and function of the metabotropic purinergic P2Y receptor family in experimental seizure models and patients with drug-refractory epilepsy. Epilepsia 58:1603–1614. Amhaoul H, Ali I, Mola M, Van Eetveldt A, Szewczyk K, Missault S, Bielen K, Kumar- Singh S, Rech J, Lord B, Ceusters M, Bhattacharya A, Dedeurwaerdere S (2016) P2X7 receptor antagonism reduces the severity of spontaneous seizures in a chronic model of temporal lobe epilepsy. Neuropharmacology 105:175–185. Amores-Iniesta J, Barberà-Cremades M, Martínez CM, Pons JA, Revilla-Nuin B, Martínez- Alarcón L, Di Virgilio F, Parrilla P, Baroja-Mazo A, Pelegrín P (2017) Extracellular ATP Activates the NLRP3 Inflammasome and Is an Early Danger Signal of Skin Allograft Rejection. Cell Rep 21:3414–3426. Anderson IM (2018) Does electroconvulsive therapy damage the brain? Lancet Psychiatry 5:294–295. Arisi GM, Foresti ML, Katki K, Shapiro LA (2015) Increased CCL2, CCL3, CCL5, and IL- 1β cytokine concentration in piriform cortex, hippocampus, and neocortex after pilocarpine-induced seizures. J Neuroinflammation 12:1–7.

- 161 -

Avanzini G, Forcelli PA, Gale K (2014) Are There Really “Epileptogenic” Mechanisms or Only Corruptions of “Normal” Plasticity? In: Issues in Clinical Epileptology: A View from the Bench, Advances in Experimental Medicine and Biology (Scharfman HE, Buckmaster PS eds), pp95–107. Dordrecht, Netherlands: Springer. Avignone E, Lepleux M, Angibaud J, Nägerl UV (2015) Altered morphological dynamics of activated microglia after induction of status epilepticus. J Neuroinflammation 12. Avignone E, Ulmann L, Levavasseur F, Rassendren F, Audinat E (2008) Status Epilepticus Induces a Particular Microglial Activation State Characterized by Enhanced Purinergic Signaling. J Neurosci 28:9133–9144. Balosso S, Maroso M, Sanchez-Alavez M, Ravizza T, Frasca A, Bartfai T, Vezzani A (2008) A novel non-transcriptional pathway mediates the proconvulsive effects of interleukin- 1β. Brain 131:3256–3265. Banerjee PN, Filippi D, Hauser WA (2009) The descriptive epidemiology of epilepsy—A review. Epilepsy Res 85:31–45. Bansod A, Sonavane SS, Shah NB, De Sousa AA, Andrade C (2018) A Randomized, Nonblind, Naturalistic Comparison of Efficacy and Cognitive Outcomes With Right Unilateral, Bifrontal, and Bitemporal Electroconvulsive Therapy in Schizophrenia. J ECT 34:26–30. Bao L, Locovei S, Dahl G (2004) Pannexin membrane channels are mechanosensitive conduits for ATP. FEBS Lett 572:65–68. Barker-Haliski ML, Johnson K, Billingsley P, Huff J, Handy LJ, Khaleel R, Lu Z, Mau MJ, Pruess TH, Rueda C, Saunders G, Underwood TK, Vanegas F, Smith MD, West PJ, Wilcox KS (2017) Validation of a Preclinical Drug Screening Platform for Pharmacoresistant Epilepsy. Neurochem Res 42:1904–1918. Barr DS, Hoyt KL, Moore SD, Wilson WA (1997) Post-ictal depression transiently inhibits induction of LTP in area CA1 of the rat hippocampal slice. Epilepsy Res 27:111–118. Bartlett R, Yerbury JJ, Sluyter R (2013) P2X7 receptor activation induces reactive oxygen species formation and cell death in murine EOC13 microglia. Mediators Inflamm 2013:271813. Beaino W, Janssen B, Kooij G, van der Pol SMA, van Het Hof B, van Horssen J, Windhorst AD, de Vries HE (2017) Purinergic receptors P2Y12R and P2X7R: potential targets for PET imaging of microglia phenotypes in multiple sclerosis. J Neuroinflammation 14:259. Beamer E, Conte G, Engel T (2019) ATP release during seizures – A critical evaluation of the evidence. Brain Res Bull In Press:S0361923018307391. Beamer E, Fischer W, Engel T (2017) The ATP-Gated P2X7 Receptor As a Target for the Treatment of Drug-Resistant Epilepsy. Front Neurosci 11:21. Beattie EC, Stellwagen D, Morishita W, Bresnahan JC, Ha BK, Zastrow MV, Beattie MS, Malenka RC (2002) Control of Synaptic Strength by Glial TNFα. Science 295:2282– 2285. Becker AJ (2018) Review: Animal models of acquired epilepsy: insights into mechanisms of human epileptogenesis. Neuropathol Appl Neurobiol 44:112–129. Becker D, Woltersdorf R, Boldt W, Schmitz S, Braam U, Schmalzing G, Markwardt F (2008) The P2X7 Carboxyl Tail Is a Regulatory Module of P2X7 Receptor Channel Activity. J Biol Chem 283:25725–25734.

- 162 -

Ben-Ari Y (1985) Limbic seizure and brain damage produced by kainic acid: mechanisms and relevance to human temporal lobe epilepsy. Neuroscience 14:375–403. Benbadis SR, Hauser WA (2000) An estimate of the prevalence of psychogenic non-epileptic seizures. Seizure 9:280–281. Bengzon J, Mohapel P, Ekdahl CT, Lindvall O (2002) Neuronal apoptosis after brief and prolonged seizures In: Progress in Brain Research 135, Do Seizures Damage the Brain? , pp111–119. Bennett ML, Bennett FC, Liddelow SA, Ajami B, Zamanian JL, Fernhoff NB, Mulinyawe SB, Bohlen CJ, Adil A, Tucker A, Weissman IL, Chang EF, Li G, Grant GA, Hayden Gephart MG, Barres BA (2016) New tools for studying microglia in the mouse and human CNS. Proc Natl Acad Sci 113:E1738–E1746. Berg AT, Shinnar S (1997) Do seizures beget seizures? An assessment of the clinical evidence in humans. J Clin Neurophysiol 14:102–110. Berg AT, Shinnar S (1991) The risk of seizure recurrence following a first unprovoked seizure: a quantitative review. Neurology 41:965–972. Berrios GE (1997) The scientific origins of electroconvulsive therapy: a conceptual history. Hist Psychiatry 8:105–119. Bezzi P, Domercq M, Brambilla L, Galli R, Schols D, De Clercq E, Vescovi A, Bagetta G, Kollias G, Meldolesi J, Volterra A (2001) CXCR4-activated astrocyte glutamate release via TNFalpha: amplification by microglia triggers neurotoxicity. Nat Neurosci 4:702–710. Bhattacharya A, Biber K (2016) The microglial ATP-gated ion channel P2X7 as a CNS drug target. Glia 64:1772–1787. Bittigau P, Sifringer M, Genz K, Reith E, Pospischil D, Govindarajalu S, Dzietko M, Pesditschek S, Mai I, Dikranian K, Olney JW, Ikonomidou C (2002) Antiepileptic drugs and apoptotic neurodegeneration in the developing brain. Proc Natl Acad Sci 99:15089–15094. Boddeke EW, Meigel I, Frentzel S, Biber K, Renn LQ, Gebicke-Härter P (1999) Functional expression of the fractalkine (CX3C) receptor and its regulation by lipopolysaccharide in rat microglia. Eur J Pharmacol 374:309–313. Borisovskaya A, Bryson WC, Buchholz J, Samii A, Borson S (2016) Electroconvulsive therapy for depression in Parkinson’s disease: systematic review of evidence and recommendations. Neurodegener Dis Manag 6:161–176. Bosco DB, Zheng J, Xu Z, Peng J, Eyo UB, Tang K, Yan C, Huang J, Feng L, Wu G, Richardson JR, Wang H, Wu L-J (2018) RNAseq analysis of hippocampal microglia after kainic acid-induced seizures. Mol Brain 11:34. Boucsein C, Kettenmann H, Nolte C (2000) Electrophysiological properties of microglial cells in normal and pathologic rat brain slices. Eur J Neurosci 12:2049–2058. Boucsein C, Zacharias R, Färber K, Pavlovic S, Hanisch U-K, Kettenmann H (2003) Purinergic receptors on microglial cells: functional expression in acute brain slices and modulation of microglial activation in vitro. Eur J Neurosci 17:2267–2276. Bozzi Y, Caleo M (2016) Epilepsy, Seizures, and Inflammation: Role of the C-C Motif Ligand 2 Chemokine. DNA Cell Biol 35:257–260. Bozzi Y, Provenzano G, Casarosa S (2018) Neurobiological bases of autism-epilepsy comorbidity: a focus on excitation/inhibition imbalance. Eur J Neurosci 47:534–548.

- 163 -

Brackenbury WJ, Yuan Y, O’Malley HA, Parent JM, Isom LL (2013) Abnormal neuronal patterning occurs during early postnatal brain development of Scn1b-null mice and precedes hyperexcitability. Proc Natl Acad Sci 110:1089–1094. Bradley DJ, Kjellbom P, Lamb CJ (1992) Elicitor- and wound-induced oxidative cross- linking of a proline-rich plant cell wall protein: A novel, rapid defense response. Cell 70:21–30. Brandao-Burch A, Key ML, Patel JJ, Arnett TR, Orriss IR (2012) The P2X7 Receptor is an Important Regulator of Extracellular ATP Levels. Front Endocrinol 3:41. Brown GC, Neher JJ (2014) Microglial phagocytosis of live neurons. Nat Rev Neurosci 15:209–216. Browne LE, Compan V, Bragg L, North RA (2013) P2X7 Receptor Channels Allow Direct Permeation of Nanometer-Sized Dyes. J Neurosci 33:3557–3566. Buckmaster PS, Haney MM (2012) Factors affecting outcomes of pilocarpine treatment in a mouse model of temporal lobe epilepsy. Epilepsy Res 102:153–159. Buisman HP, Steinberg TH, Fischbarg J, Silverstein SC, Vogelzang SA, Ince C, Ypey DL, Leijh PC (1988) Extracellular ATP induces a large nonselective conductance in macrophage plasma membranes. Proc Natl Acad Sci 85:7988–7992. Burnstock G, Knight GE (2018) The potential of P2X7 receptors as a therapeutic target, including inflammation and tumour progression. Purinergic Signal 14:1–18. Butovsky O, Weiner HL (2018) Microglial signatures and their role in health and disease. Nat Rev Neurosci 19:622–635. Cardona AE, Pioro EP, Sasse ME, Kostenko V, Cardona SM, Dijkstra IM, Huang D, Kidd G, Dombrowski S, Dutta R, Lee J-C, Cook DN, Jung S, Lira SA, Littman DR, Ransohoff RM (2006) Control of microglial neurotoxicity by the fractalkine receptor. Nat Neurosci 9:917–924. Casamenti F, Prosperi C, Scali C, Giovannelli L, Colivicchi MA, Faussone-Pellegrini MS, Pepeu G (1999) Interleukin-1β activates forebrain glial cells and increases nitric oxide production and cortical glutamate and GABA release in vivo: implications for Alzheimer’s disease. Neuroscience 91:831–842. Cerri C, Genovesi S, Allegra M, Pistillo F, Püntener U, Guglielmotti A, Perry VH, Bozzi Y, Caleo M (2016) The Chemokine CCL2 Mediates the Seizure-enhancing Effects of Systemic Inflammation. J Neurosci 36:3777–3788. Chen B, Choi H, Hirsch LJ, Katz A, Legge A, Buchsbaum R, Detyniecki K (2017a) Psychiatric and behavioral side effects of antiepileptic drugs in adults with epilepsy. Epilepsy Behav EB 76:24–31. Chen B, Choi H, Hirsch LJ, Moeller J, Javed A, Kato K, Legge A, Buchsbaum R, Detyniecki K (2015) Cosmetic side effects of antiepileptic drugs in adults with epilepsy. Epilepsy Behav EB 42:129–137. Chen B, Detyniecki K, Choi H, Hirsch L, Katz A, Legge A, Wong R, Jiang A, Buchsbaum R, Farooque P (2017b) Psychiatric and behavioral side effects of anti-epileptic drugs in adolescents and children with epilepsy. Eur J Paediatr Neurol 21:441–449. Chen Z, Duan R, Quezada H, Mix E, Nennesmo I, Adem A, Winblad B, Zhu J (2005) Increased microglial activation and astrogliosis after intranasal administration of kainic acid in C57BL/6 mice. J Neurobiol 62:207–218.

- 164 -

Chen Z, Jalabi W, Hu W, Park H-J, Gale JT, Kidd GJ, Bernatowicz R, Gossman ZC, Chen JT, Dutta R, Trapp BD (2014) Microglial displacement of inhibitory synapses provides neuroprotection in the adult brain. Nat Commun 5:4486. Chiavegato A, Zurolo E, Losi G, Aronica E, Carmignoto G (2014) The inflammatory molecules IL-1β and HMGB1 can rapidly enhance focal seizure generation in a brain slice model of temporal lobe epilepsy. Front Cell Neurosci 8. Cho K-O, Lybrand ZR, Ito N, Brulet R, Tafacory F, Zhang L, Good L, Ure K, Kernie SG, Birnbaum SG, Scharfman HE, Eisch AJ, Hsieh J (2015) Aberrant hippocampal neurogenesis contributes to epilepsy and associated cognitive decline. Nat Commun 6:6606. Clasadonte J, Dong J, Hines DJ, Haydon PG (2013) Astrocyte control of synaptic NMDA receptors contributes to the progressive development of temporal lobe epilepsy. Proc Natl Acad Sci 110:17540–17545. Colonna M, Butovsky O (2017) Microglia Function in the Central Nervous System During Health and Neurodegeneration. Annu Rev Immunol 35:441–468. Copeland L, Meek A, Kerr M, Robling M, Hood K, McNamara R (2017) Measurement of side effects of anti-epileptic drugs (AEDs) in adults with intellectual disability: A systematic review. Seizure 51:61–73. Corona AW, Huang Y, O’Connor JC, Dantzer R, Kelley KW, Popovich PG, Godbout JP (2010) Fractalkine receptor (CX3CR1) deficiency sensitizes mice to the behavioral changes induced by lipopolysaccharide. J Neuroinflammation 7:93. Covolan L, Mello LE (2000) Temporal profile of neuronal injury following pilocarpine or kainic acid-induced status epilepticus. Epilepsy Res 39:133–152. Cserép C et al. (2019) Microglia monitor and protect neuronal function via specialized somatic purinergic junctions. bioRxiv 606079. Cui W, Kobau R, Zack MM, Helmers S, Yeargin-Allsopp M (2015) Seizures in Children and Adolescents Aged 6-17 Years - United States, 2010-2014. MMWR Morb Mortal Wkly Rep 64:1209–1214. Curia G, Longo D, Biagini G, Jones RSG, Avoli M (2008) The pilocarpine model of temporal lobe epilepsy. J Neurosci Methods 172:143–157. Dahlgren C, Karlsson A (1999) Respiratory burst in human neutrophils. J Immunol Methods 232:3–14. D’Amour J, Magagna-Poveda A, Moretto J, Friedman D, LaFrancois JJ, Pearce P, Fenton AA, MacLusky NJ, Scharfman HE (2015) Interictal spike frequency varies with ovarian cycle stage in a rat model of epilepsy. Exp Neurol 269:102–119. Danquah W et al. (2016) Nanobodies that block gating of the P2X7 ion channel ameliorate inflammation. Sci Transl Med 8:366ra162. Das CP, Sawhney IM, Lal V, Prabhakar S (2000) Risk of recurrence of seizures following single unprovoked idiopathic seizure. Neurol India 48:357–360. Davalos D, Grutzendler J, Yang G, Kim JV, Zuo Y, Jung S, Littman DR, Dustin ML, Gan W- BB (2005) ATP mediates rapid microglial response to local brain injury in vivo. Nat Neurosci 8:752–758. de Biase LM, Schuebel KE, Fusfeld ZH, Jair K, Hawes IA, Cimbro R, Zhang H-Y, Liu Q-R, Shen H, Xi Z-X, Goldman D, Bonci A (2017) Local Cues Establish and Maintain Region-Specific Phenotypes of Basal Ganglia Microglia. Neuron 95.

- 165 -

del Rio-Hortega P (1919) El “tercer elemento de los centros nerviosos”. IV. Poder fagocitario y movilidad de la microglía. Bol Soc Esp Biol VIII:154–166. Denlinger LC, Fisette PL, Sommer JA, Watters JJ, Prabhu U, Dubyak GR, Proctor RA, Bertics PJ (2001) Cutting Edge: The Nucleotide Receptor P2X7 Contains Multiple Protein- and Lipid-Interaction Motifs Including a Potential Binding Site for Bacterial Lipopolysaccharide. J Immunol 167:1871–1876. Deussing JM, Arzt E (2018) P2X7 Receptor: A Potential Therapeutic Target for Depression? Trends Mol Med 24:736–747. Devanand DP, Dwork AJ, Hutchinson ER, Bolwig TG, Sackeim HA (1994) Does ECT alter brain structure? Am J Psychiatry 151:957–970. Devanand DP, Verma AK, Tirumalasetti F, Sackeim HA (1991) Absence of cognitive impairment after more than 100 lifetime ECT treatments. Am J Psychiatry 148:929– 932. Devinsky O, Vezzani A, O’Brien TJ, Jette N, Scheffer IE, de Curtis M, Perucca P (2018) Epilepsy. Nat Rev Dis Primer 4:18024. Di Virgilio F, Pizzo P, Zanovello P, Bronte V, Collavo D (1990) Extracellular ATP as a possible mediator of cell-mediated cytotoxicity. Immunol Today 11:274–277. Dibaj P, Nadrigny F, Steffens H, Scheller A, Hirrlinger J, Schomburg ED, Neusch C, Kirchhoff F (2010) NO mediates microglial response to acute spinal cord injury under ATP control in vivo. Glia 58:1133–1144. Dingledine R, Varvel NH, Dudek FE (2014) When and How Do Seizures Kill Neurons, and Is Cell Death Relevant to Epileptogenesis? In: Issues in Clinical Epileptology: A View from the Bench, Advances in Experimental Medicine and Biology (Scharfman HE, Buckmaster PS eds), pp109–122. Dordrecht, Netherlands: Springer. do Nascimento AL, Dos Santos NF, Campos Pelágio F, Aparecida Teixeira S, de Moraes Ferrari EA, Langone F (2012) Neuronal degeneration and gliosis time-course in the mouse hippocampal formation after pilocarpine-induced status epilepticus. Brain Res 1470:98–110. Donnelly-Roberts DL, Namovic MT, Han P, Jarvis MF (2009) Mammalian P2X7 receptor pharmacology: comparison of recombinant mouse, rat and human P2X7 receptors. Br J Pharmacol 157:1203–1214. Dorsam RT, Kunapuli SP (2004) Central role of the P2Y12 receptor in platelet activation. J Clin Invest 113:340–345. Duan Y, Sahley CL, Muller KJ (2009) ATP and NO dually control migration of microglia to nerve lesions. Dev Neurobiol 69:60–72. Dubyak GR (2007) Go It Alone No More—P2X7 Joins the Society of Heteromeric ATP- Gated Receptor Channels. Mol Pharmacol 72:1402–1405. Dutta S, Sengupta P (2016) Men and mice: Relating their ages. Life Sci 152:244–248. Dyrvig M, Hansen HH, Christiansen SH, Woldbye DPD, Mikkelsen JD, Lichota J (2012) Epigenetic regulation of Arc and c-Fos in the hippocampus after acute electroconvulsive stimulation in the rat. Brain Res Bull, The Impact of Missense Mutations on Human Behavior 88:507–513. Eichhoff G, Brawek B, Garaschuk O (2011) Microglial calcium signal acts as a rapid sensor of single neuron damage in vivo. Biochim Biophys Acta 1813:1014–1024.

- 166 -

Elmore MRP, Najafi AR, Koike MA, Dagher NN, Spangenberg EE, Rice RA, Kitazawa M, Matusow B, Nguyen H, West BL, Green KN (2014) Colony-Stimulating Factor 1 Receptor Signaling Is Necessary for Microglia Viability, Unmasking a Microglia Progenitor Cell in the Adult Brain. Neuron 82:380–397. Engel T, Brennan GP, Sanz-Rodriguez A, Alves M, Beamer E, Watters O, Henshall DC, Jimenez-Mateos EM (2017) A calcium-sensitive feed-forward loop regulating the expression of the ATP-gated purinergic P2X7 receptor via specificity protein 1 and microRNA-22. Biochim Biophys Acta Mol Cell Res 1864:255–266. Engel T, Gomez-Villafuertes R, Tanaka K, Mesuret G, Sanz-Rodriguez A, Garcia-Huerta P, Miras-Portugal TM, Henshall DC, Diaz-Hernandez M (2012) Seizure suppression and neuroprotection by targeting the purinergic P2X7 receptor during status epilepticus in mice. FASEB J 1616–1628. England MJ, Liverman CT, Schultz AM, Strawbridge LM (2012) Epilepsy across the spectrum: Promoting health and understanding. Epilepsy Behav EB 25:266–276. Eyo UB, Bispo A, Liu J, Sabu S, Wu R, DiBona VL, Zheng J, Murugan M, Zhang H, Tang Y, Wu L-J (2018a) The GluN2A Subunit Regulates Neuronal NMDA receptor-Induced Microglia-Neuron Physical Interactions. Sci Rep 8:828. Eyo UB, Gu N, De S, Dong H, Richardson JR, Wu L-J (2015) Modulation of Microglial Process Convergence Toward Neuronal Dendrites by Extracellular Calcium. J Neurosci 35:2417–2422. Eyo UB, Mo M, Yi M-H, Murugan M, Liu J, Yarlagadda R, Margolis DJ, Xu P, Wu L-J (2018b) P2Y12R-Dependent Translocation Mechanisms Gate the Changing Microglial Landscape. Cell Rep 23:959–966. Eyo UB, Peng J, Murugan M, Mo M, Lalani A, Xie P, Xu P, Margolis DJ, Wu L-JJ (2016) Regulation of Physical Microglia-Neuron Interactions by Fractalkine Signaling after Status Epilepticus. eNeuro 3. Eyo UB, Peng J, Swiatkowski P, Mukherjee A, Bispo A, Wu L-J (2014) Neuronal Hyperactivity Recruits Microglial Processes via Neuronal NMDA Receptors and Microglial P2Y12 Receptors after Status Epilepticus. J Neurosci 34:10528–10540. Fabene PF, Merigo F, Galiè M, Benati D, Bernardi P, Farace P, Nicolato E, Marzola P, Sbarbati A (2007) Pilocarpine-induced status epilepticus in rats involves ischemic and excitotoxic mechanisms. PloS One 2:e1105. Falzoni S, Munerati M, Ferrari D, Spisani S, Moretti S, Di Virgilio F (1995) The purinergic P2Z receptor of human macrophage cells. Characterization and possible physiological role. J Clin Invest 95:1207–1216. Feng L, Murugan M, Bosco DB, Liu Y, Peng J, Worrell GA, Wang H, Ta LE, Richardson JR, Shen Y, Wu L (2019) Microglial proliferation and monocyte infiltration contribute to microgliosis following status epilepticus. Glia glia.23616. Ferrari D, Chiozzi P, Falzoni S, Dal Susino M, Collo G, Buell G, Di Virgilio F (1997) ATP- mediated cytotoxicity in microglial cells. Neuropharmacology 36:1295–1301. Ferrari D, Pizzirani C, Adinolfi E, Lemoli RM, Curti A, Idzko M, Panther E, Virgilio FD (2006) The P2X7 Receptor: A Key Player in IL-1 Processing and Release. J Immunol 176:3877–3883.

- 167 -

Ferrari D, Villalba M, Chiozzi P, Falzoni S, Ricciardi-Castagnoli P, Virgilio FD (1996) Mouse microglial cells express a plasma membrane pore gated by extracellular ATP. J Immunol 156:1531–1539. Ferraro TN, Golden GT, Smith GG, DeMuth D, Buono RJ, Berrettini WH (2002) Mouse strain variation in maximal electroshock seizure threshold. Brain Res 936:82–86. Fiacco TA, McCarthy KD (2004) Intracellular astrocyte calcium waves in situ increase the frequency of spontaneous AMPA receptor currents in CA1 pyramidal neurons. J Neurosci Off J Soc Neurosci 24:722–732. Fiest KM, Sauro KM, Wiebe S, Patten SB, Kwon C-S, Dykeman J, Pringsheim T, Lorenzetti DL, Jetté N (2017) Prevalence and incidence of epilepsy: A systematic review and meta-analysis of international studies. Neurology 88:296–303. Fink M (2009) Electroconvulsive therapy: a guide for professionals and their patients. New York, NY: Oxford University Press. Fink M (1984) Meduna and the origins of convulsive therapy. Am J Psychiatry 141:1034– 1041. Fischer W, Franke H, Krügel U, Müller H, Dinkel K, Lord B, Letavic MA, Henshall DC, Engel T (2016) Critical Evaluation of P2X7 Receptor Antagonists in Selected Seizure Models. PLOS ONE 11:e0156468. Fisher C (2011) Shockaholic. New York: Simon & Schuster. Flurkey K, M. Currer J, Harrison DE (2007) Chapter 20 - Mouse Models in Aging Research In: The Mouse in Biomedical Research (Fox JG, Davisson MT, Quimby FW, Barthold SW, Newcomer CE, Smith AL eds), pp637–672. Burlington: Academic Press. Foda MA, Marmarou A (1994) A new model of diffuse brain injury in rats. Part II: Morphological characterization. J Neurosurg 80:301–313. Fonseca MI, Chu S-H, Hernandez MX, Fang MJ, Modarresi L, Selvan P, MacGregor GR, Tenner AJ (2017) Cell-specific deletion of C1qa identifies microglia as the dominant source of C1q in mouse brain. J Neuroinflammation 14:48. Forcelli PA (2011) Sequelae of Neonatal Antiepileptic Drug Exposure in Rats (Doctoral Dissertation). Washington, D.C., USA: Georgetown University. Forcelli PA, Gale K (2014) Brain Circuits Responsible for Seizure Generation, Propagation, and Control: Insights from Preclinical Research. In: Epilepsy Topics (Holmes M ed). Forcelli PA, Janssen MJ, Vicini S, Gale K (2012) Neonatal exposure to antiepileptic drugs disrupts striatal synaptic development. Ann Neurol 72:363–372. Fordyce CB, Jagasia R, Zhu X, Schlichter LC (2005) Microglia Kv1.3 Channels Contribute to Their Ability to Kill Neurons. J Neurosci 25:7139–7149. Foresti ML, Arisi GM, Katki K, Montañez A, Sanchez RM, Shapiro LA (2009) Chemokine CCL2 and its receptor CCR2 are increased in the hippocampus following pilocarpine- induced status epilepticus. J Neuroinflammation 6:1–11. Franceschini A, Capece M, Chiozzi P, Falzoni S, Sanz JM, Sarti AC, Bonora M, Pinton P, Di Virgilio F (2015) The P2X7 receptor directly interacts with the NLRP3 inflammasome scaffold protein. FASEB J 29:2450–2461. Frankel WN, Taylor L, Beyer B, Tempel BL, White HS (2001) Electroconvulsive Thresholds of Inbred Mouse Strains. Genomics 74:306–312. Franklin KBJ, Paxinos G (2008) The mouse brain in stereotaxic coordinates, 3rd ed. ed. Amsterdam, Netherlands: Elsevier AP.

- 168 -

Fritsch G von, Hitzig E (1870) Über die elektrische Erregbarkeit des Grosshirns. In: Archiv für Anatomie, Physiologie und Wissenschaftliche Medicin , pp300–332. Leipzig, Germany: Veit. Fuhrmann M, Bittner T, Jung CKE, Burgold S, Page RM, Mitteregger G, Haass C, LaFerla FM, Kretzschmar H, Herms J (2010) Microglial Cx3cr1 knockout prevents neuron loss in a mouse model of Alzheimer’s disease. Nat Neurosci 13:411–413. Gan J-L, Duan H-F, Cheng Z-X, Yang J-M, Zhu X-Q, Gao C-Y, Zhao L-M, Liang X-J (2017) Neuroprotective Effect of Modified Electroconvulsive Therapy for Schizophrenia: A Proton Magnetic Resonance Spectroscopy Study. J Nerv Ment Dis 205:480–486. García-Huerta P, Díaz-Hernandez M, Delicado EG, Pimentel-Santillana M, Miras-Portugal MT, Gómez-Villafuertes R (2012) The specificity protein factor Sp1 mediates transcriptional regulation of P2X7 receptors in the nervous system. J Biol Chem 287:44628–44644. Gee JM, Smith NA, Fernandez FR, Economo MN, Brunert D, Rothermel M, Morris SC, Talbot A, Palumbos S, Ichida JM, Shepherd JD, West PJ, Wachowiak M, Capecchi MR, Wilcox KS, White JA, Tvrdik P (2014) Imaging Activity in Neurons and Glia with a Polr2a-based and Cre-dependent GCaMP5G-IRES-tdTomato Reporter Mouse. Neuron 83:1058–1072. Ghosh S, Castillo E, Frias ES, Swanson RA (2018) Bioenergetic regulation of microglia. Glia 66:1200–1212. Ginhoux F, Greter M, Leboeuf M, Nandi S, See P, Gokhan S, Mehler MF, Conway SJ, Ng LG, Stanley ER, Samokhvalov IM, Merad M (2010) Fate Mapping Analysis Reveals That Adult Microglia Derive from Primitive Macrophages. Science 330:841–845. Goodman JH (1998) Experimental Models of Status Epilepticus. In: Neuropharmacology Methods in Epilepsy Research, Methods in the Life Sciences: Cellular and Molecular Neuropharmacology (Peterson SL, Albertson TE eds), pp95–126. Boca Raton: CRC Press. Gorter JA, van Vliet EA, Aronica E (2015) Status epilepticus, blood-brain barrier disruption, inflammation, and epileptogenesis. Epilepsy Behav EB 49:13–16. Gowers WR (1881) Epilepsy and other chronic convulsive disorders: their causes, symptoms, and treatment. London, England: J&A Churchill. Gowers WR (1880) The Gulstonian Lectures on Epilepsy. Br Med J 1:353–357. Greenwood SM, Mizielinska SM, Frenguelli BG, Harvey J, Connolly CN (2007) Mitochondrial dysfunction and dendritic beading during neuronal toxicity. J Biol Chem 282:26235–26244. Gröticke I, Hoffmann K, Löscher W (2008) Behavioral alterations in a mouse model of temporal lobe epilepsy induced by intrahippocampal injection of kainate. Exp Neurol 213:71–83. Gröticke I, Hoffmann K, Löscher W (2007) Behavioral alterations in the pilocarpine model of temporal lobe epilepsy in mice. Exp Neurol 207:329–349. Guo C, Masin M, Qureshi OS, Murrell-Lagnado RD (2007) Evidence for functional P2X4/P2X7 heteromeric receptors. Mol Pharmacol 72:1447–1456. Hammond TR et al. (2019) Single-Cell RNA Sequencing of Microglia throughout the Mouse Lifespan and in the Injured Brain Reveals Complex Cell-State Changes. Immunity 50:253-271.e6.

- 169 -

Harada H, Tsukimoto M, Ikari A, Takagi K, Suketa Y (2003) P2X7 Receptor-induced generation of reactive oxygen species in rat mesangial cells. Drug Dev Res 59:112– 117. Hauser WA, Annegers JF, Kurland LT (1991) Prevalence of Epilepsy in Rochester, Minnesota: 1940–1980. Epilepsia 32:429–445. Hauser WA, Lee JR (2002) Do seizures beget seizures? In: Progress in Brain Research 135, Do Seizures Damage the Brain? , pp215–219. New York, NY: Elsevier. Hauser WA, Rich SS, Annegers JF, Anderson VE (1990) Seizure recurrence after a 1st unprovoked seizure: an extended follow-up. Neurology 40:1163–1170. Hauser WA, Rich SS, Lee JR, Annegers JF, Anderson VE (1998) Risk of recurrent seizures after two unprovoked seizures. N Engl J Med 338:429–434. Haynes SE, Hollopeter G, Yang G, Kurpius D, Dailey ME, Gan W-B, Julius D (2006) The P2Y12 receptor regulates microglial activation by extracellular nucleotides. Nat Neurosci 9:1512–1519. Hebb DO (1949) The Organization of Behavior: A Neuropsychological Theory, 1st ed. Oxford, England: Wiley. Henshall DC, Engel T (2015) P2X purinoceptors as a link between hyperexcitability and neuroinflammation in status epilepticus. Epilepsy Behav EB, SI:Status Epilepticus 49:8–12. Heppner FL, Greter M, Marino D, Falsig J, Raivich G, Hövelmeyer N, Waisman A, Rülicke T, Prinz M, Priller J, Becher B, Aguzzi A (2005) Experimental autoimmune encephalomyelitis repressed by microglial paralysis. Nat Med 11:146–152. Herbert J-M, Savi P (2003) P2Y12, a new platelet ADP receptor, target of clopidogrel. Semin Vasc Med 3:113–122. Hesdorffer DC, Benn EKT, Cascino GD, Hauser WA (2009) Is a first acute symptomatic seizure epilepsy? Mortality and risk for recurrent seizure. Epilepsia 50:1102–1108. Hesdorffer DC, Logroscino G, Cascino G, Annegers JF, Hauser WA (1998) Risk of unprovoked seizure after acute symptomatic seizure: effect of status epilepticus. Ann Neurol 44:908–912. Hevers W, Lüddens H (2002) Pharmacological heterogeneity of γ-aminobutyric acid receptors during development suggests distinct classes of rat cerebellar granule cells in situ. Neuropharmacology 42:34–47. Hines DJ, Hines RM, Mulligan SJ, Macvicar BA (2009) Microglia processes block the spread of damage in the brain and require functional chloride channels. Glia 57:1610–1618. Hiragi T, Ikegaya Y, Koyama R (2018) Microglia after Seizures and in Epilepsy. Cells 7. Hirasawa T, Ohsawa K, Imai Y, Ondo Y, Akazawa C, Uchino S, Kohsaka S (2005) Visualization of microglia in living tissues using Iba1-EGFP transgenic mice. J Neurosci Res 81:357–362. Hong S, Beja-Glasser VF, Nfonoyim BM, Frouin A, Li S, Ramakrishnan S, Merry KM, Shi Q, Rosenthal A, Barres BA, Lemere A, Selkoe DJ, Stevens B (2016) Complement and microglia mediate early synapse loss in Alzheimer mouse models. Science 352. Hoogland ICM, Houbolt C, van Westerloo DJ, van Gool WA, van de Beek D (2015) Systemic inflammation and microglial activation: systematic review of animal experiments. J Neuroinflammation 12:114.

- 170 -

Horti AG et al. (2019) PET imaging of microglia by targeting macrophage colony-stimulating factor 1 receptor (CSF1R). Proc Natl Acad Sci 116:1686–1691. Hsieh TF, Simler S, Vergnes M, Gass P, Marescaux C, Wiegand SJ, Zimmermann M, Herdegen T (1998) BDNF Restores the Expression of Jun and Fos Inducible Transcription Factors in the Rat Brain Following Repetitive Electroconvulsive Seizures. Exp Neurol 149:161–174. Hu S, Sheng WS, Ehrlich LC, Peterson PK, Chao CC (2000) Cytokine Effects on Glutamate Uptake by Human Astrocytes. Neuroimmunomodulation 7:153–159. Hughes PM, Botham MS, Frentzel S, Mir A, Perry VH (2002) Expression of fractalkine (CX3CL1) and its receptor, CX3CR1, during acute and chronic inflammation in the rodent CNS. Glia 37:314–327. Iles KE, Forman HJ (2002) Macrophage signaling and respiratory burst. Immunol Res 26:95– 105. Ilie A, Raimondo JV, Akerman CJ (2012) Adenosine release during seizures attenuates GABAA receptor-mediated depolarization. J Neurosci 32:5321–5332. Ingvar M (1986) Cerebral Blood Flow and Metabolic Rate during Seizuresa. Ann N Y Acad Sci 462:194–206. Inostroza M, Brotons-Mas JR, Laurent F, Cid E, Prida LM de la (2013) Specific Impairment of “What-Where-When” Episodic-Like Memory in Experimental Models of Temporal Lobe Epilepsy. J Neurosci 33:17749–17762. Inoue K (2019) Role of the P2X4 receptor in neuropathic pain. Curr Opin Pharmacol 47:33– 39. Inoue K (2008) Purinergic systems in microglia. Cell Mol Life Sci 65:3074–3080. Ip Q, Malone DC, Chong J, Harris RB, Labiner DM (2018) An update on the prevalence and incidence of epilepsy among older adults. Epilepsy Res 139:107–112. Irino Y, Nakamura Y, Inoue K, Kohsaka S, Ohsawa K (2008) Akt activation is involved in P2Y12 receptor-mediated chemotaxis of microglia. J Neurosci Res 86:1511–1519. Isaev D, Lushnikova I, Lunko O, Zapukhliak O, Maximyuk O, Romanov A, Skibo GG, Tian C, Holmes GL, Isaeva E (2015) Contribution of protease-activated receptor 1 in status epilepticus-induced epileptogenesis. Neurobiol Dis 78:68–76. Ishizuka K et al. (2017) Rare genetic variants in CX3CR1 and their contribution to the increased risk of schizophrenia and autism spectrum disorders. Transl Psychiatry 7:e1184. Iyengar SS, LaFrancois JJ, Friedman D, Drew LJ, Denny CA, Burghardt NS, Wu MV, Hsieh J, Hen R, Scharfman HE (2015) Suppression of Adult Neurogenesis Increases the Acute Effects of Kainic Acid. Exp Neurol 264:135–149. Jabs T, Tschöpe M, Colling C, Hahlbrock K, Scheel D (1997) Elicitor-stimulated ion fluxes and O2− from the oxidative burst are essential components in triggering defense gene activation and phytoalexin synthesis in parsley. Proc Natl Acad Sci 94:4800–4805. Jackson JH (1890) The Lumleian Lectures on convulsive seizures. The Lancet 135:685–688. Jackson JH (1874) On the scientific and empirical investigations of epilepsies. Med Press Circ 3:316–360. Jackson JH (1873) On the anatomical, physiological, and pathological investigation of epilepsies. West Rid Lunatic Asylum Med Rep 3:315–349. Jackson JH (1870) A study of convulsions. Trans St Andrews Med Grad Assoc 3:162–204.

- 171 -

Jacoby A, Gamble C, Doughty J, Marson A, Chadwick D, Medical Research Council MESS Study Group (2007) Quality of life outcomes of immediate or delayed treatment of early epilepsy and single seizures. Neurology 68:1188–1196. Jain S, LaFrancois JJ, Botterill JJ, Alcantara-Gonzalez D, Scharfman HE (2019) Adult neurogenesis in the mouse dentate gyrus protects the hippocampus from neuronal injury following severe seizures. Hippocampus. Janigro D, Iffland PH, Marchi N, Granata T (2013) A role for inflammation in status epilepticus is revealed by a review of current therapeutic approaches. Epilepsia 54 Suppl 6:30–32. Janssen B, Vugts DJ, Funke U, Spaans A, Schuit RC, Kooijman E, Rongen M, Perk LR, Lammertsma AA, Windhorst AD (2014) Synthesis and initial preclinical evaluation of the P2X7 receptor antagonist [11C]A-740003 as a novel tracer of neuroinflammation. J Label Compd Radiopharm 57:509–516. Janssen B, Vugts DJ, Windhorst AD, Mach RH (2018) PET Imaging of Microglial Activation-Beyond Targeting TSPO. Mol Basel Switz 23. Jansson L, Wennström M, Johanson A, Tingström A (2009) Glial cell activation in response to electroconvulsive seizures. Prog Neuropsychopharmacol Biol Psychiatry 33:1119– 1128. Jarvis MF, Khakh BS (2009) ATP-gated P2X cation-channels. Neuropharmacology, Ligand- Gated Ion Channels 56:208–215. Ji R-R, Schlaepfer TE, Aizenman CD, Epstein CM, Qiu D, Huang JC, Rupp F (1998) Repetitive transcranial magnetic stimulation activates specific regions in rat brain. Proc Natl Acad Sci 95:15635–15640. Jiang L, Mackenzie A, North R, Surprenant A (2000) Brilliant blue G selectively blocks ATP- gated rat P2X(7) receptors. Mol Pharmacol 58:82–88. Jimenez-Mateos EM et al. (2015) microRNA targeting of the P2X7 purinoceptor opposes a contralateral epileptogenic focus in the hippocampus. Sci Rep 5:17486. Jimenez-Pacheco A, Diaz-Hernandez M, Arribas-Blázquez M, Sanz-Rodriguez A, Olivos-Oré LA, Artalejo AR, Alves M, Letavic M, Miras-Portugal M, Conroy RM, Delanty N, Farrell MA, O’Brien DF, Bhattacharya A, Engel T, Henshall DC (2016) Transient P2X7 Receptor Antagonism Produces Lasting Reductions in Spontaneous Seizures and Gliosis in Experimental Temporal Lobe Epilepsy. J Neurosci 36:5920–5932. Jimenez‐Pacheco A, Mesuret G, Sanz‐Rodriguez A, Tanaka K, Mooney C, Conroy R, Miras‐ Portugal M, Diaz‐Hernandez M, Henshall DC, Engel T (2013) Increased neocortical expression of the P2X7 receptor after status epilepticus and anticonvulsant effect of P2X7 receptor antagonist A‐438079. Epilepsia 54:1551–1561. Jinno S, Kosaka T (2008) Reduction of Iba1-expressing microglial process density in the hippocampus following electroconvulsive shock. Exp Neurol 212:440–447. Jonckheere J, Deloulme JC, Dall’Igna G, Chauliac N, Pelluet A, Nguon A-S, Lentini C, Brocard J, Denarier E, Brugière S, Coute Y, Heinrich C, Porcher C, Holtzmann J, Andrieux A, Suaud-Chagny M-F, Gory-Fauré S (2018) Short- and long-term efficacy of electroconvulsive stimulation in animal models of depression: The essential role of neuronal survival. Brain Stimulat 11.

- 172 -

Jung S, Aliberti J, Graemmel P, Sunshine M, Kreutzberg GW, Sher A, Littman DR (2000) Analysis of Fractalkine Receptor CX3CR1 Function by Targeted Deletion and Green Fluorescent Protein Reporter Gene Insertion. Mol Cell Biol 20:4106–4114. Kaiser T, Feng G (2019) Tmem119-EGFP and Tmem119-CreERT2 transgenic mice for labeling and manipulating microglia. bioRxiv 624825. Kalozoumi G, Kel-Margoulis O, Vafiadaki E, Greenberg D, Bernard H, Soreq H, Depaulis A, Sanoudou D (2018) Glial responses during epileptogenesis in Mus musculus point to potential therapeutic targets. PLOS ONE 13. Kalume F, Westenbroek RE, Cheah CS, Yu FH, Oakley JC, Scheuer T, Catterall WA (2013) Sudden unexpected death in a mouse model of Dravet syndrome. J Clin Invest 123:1798–1808. Karasawa A, Michalski K, Mikhelzon P, Kawate T (2017) The P2X7 receptor forms a dye- permeable pore independent of its intracellular domain but dependent on membrane lipid composition. eLife 6. Karmakar M, Katsnelson MA, Dubyak GR, Pearlman E (2016) Neutrophil P2X7 receptors mediate NLRP3 inflammasome-dependent IL-1β secretion in response to ATP. Nat Commun 7:10555. Karperien A, Ahammer H, Jelinek HF (2013) Quantitating the subtleties of microglial morphology with fractal analysis. Front Cell Neurosci 7:3. Kauffman MA, Moron DG, Consalvo D, Bello R, Kochen S (2008) Association study between interleukin 1 beta gene and epileptic disorders: a HuGe review and meta- analysis. Genet Med 10:83–88. Kawano A, Tsukimoto M, Mori D, Noguchi T, Harada H, Takenouchi T, Kitani H, Kojima S (2012) Regulation of P2X7-dependent inflammatory functions by P2X4 receptor in mouse macrophages. Biochem Biophys Res Commun 420:102–107. Kehne JH, Klein BD, Raeissi S, Sharma S (2017) The National Institute of Neurological Disorders and Stroke (NINDS) Epilepsy Therapy Screening Program (ETSP). Neurochem Res 42:1894–1903. Keppler LD, Baker CJ, Atkinson MM (1989) Active Oxygen Production During a Bacteria- Induced Hypersensitive Reaction in Tobacco Suspension Cells. Phytopathology 79:974–978. Kettenmann H, Hanisch U-K, Noda M, Verkhratsky A (2011) Physiology of microglia. Physiol Rev 91:461–553. Khakh BS, Burnstock G, Kennedy C, King BF, North RA, Séguéla P, Voigt M, Humphrey PPA (2001) International Union of Pharmacology. XXIV. Current Status of the Nomenclature and Properties of P2X Receptors and Their Subunits. Pharmacol Rev 53:107–118. Kienzler‐Norwood F, Costard L, Sadangi C, Müller P, Neubert V, Bauer S, Rosenow F, Norwood BA (2017) A novel animal model of acquired human temporal lobe epilepsy based on the simultaneous administration of kainic acid and lorazepam. Epilepsia 58:222–230. Kim I, Mlsna LM, Yoon S, Le B, Yu S, Xu D, Koh S (2015) A postnatal peak in microglial development in the mouse hippocampus is correlated with heightened sensitivity to seizure triggers. Brain Behav 5:e00403.

- 173 -

Kim J-E, Ryu HJ, Kang T-C (2013) Status epilepticus induces vasogenic edema via tumor necrosis factor-α/ endothelin-1-mediated two different pathways. PloS One 8:e74458. Klein P et al. (2018) Commonalities in epileptogenic processes from different acute brain insults: Do they translate? Epilepsia 59:37–66. Klement W, Garbelli R, Zub E, Rossini L, Tassi L, Girard B, Blaquiere M, Bertaso F, Perroy J, de Bock F, Marchi N (2018) Seizure progression and inflammatory mediators promote pericytosis and pericyte-microglia clustering at the cerebrovasculature. Neurobiol Dis 113:70–81. Kobayashi K, Imagama S, Ohgomori T, Hirano K, Uchimura K, Sakamoto K, Hirakawa A, Takeuchi H, Suzumura A, Ishiguro N, Kadomatsu K (2013) Minocycline selectively inhibits M1 polarization of microglia. Cell Death Dis 4:e525. Koizumi S, Ohsawa K, Inoue K, Kohsaka S (2013) Purinergic receptors in microglia: Functional modal shifts of microglia mediated by P2 and P1 receptors. Glia 61:47–54. Koizumi S, Shigemoto-Mogami Y, Nasu-Tada K, Shinozaki Y, Ohsawa K, Tsuda M, Joshi BV, Jacobson KA, Kohsaka S, Inoue K (2007) UDP acting at P2Y6 receptors is a mediator of microglial phagocytosis. Nature 446:1091–1095. Kondratyev A, Gale K (2004) Latency to onset of status epilepticus determines molecular mechanisms of seizure-induced cell death. Mol Brain Res 121:86–94. Kondratyev A, Sahibzada N, Gale K (2001) Electroconvulsive shock exposure prevents neuronal apoptosis after kainic acid-evoked status epilepticus. Brain Res Mol Brain Res 91:1–13. Kranaster L, Hoyer C, Aksay SS, Bumb JM, Müller N, Zill P, Schwarz MJ, Sartorius A (2018) Antidepressant efficacy of electroconvulsive therapy is associated with a reduction of the innate cellular immune activity in the cerebrospinal fluid in patients with depression. World J Biol Psychiatry 19:379–389. Krishnan GP, Bazhenov M (2011) Ionic dynamics mediate spontaneous termination of seizures and postictal depression state. J Neurosci 31:8870–8882. Kruger NJ, von Schaewen A (2003) The oxidative pentose phosphate pathway: structure and organisation. Curr Opin Plant Biol 6:236–246. Lawn N, Chan J, Lee J, Dunne J (2015) Is the first seizure epilepsy—and when? Epilepsia 56:1425–1431. Lawson LJ, Perry VH, Dri P, Gordon S (1990) Heterogeneity in the distribution and morphology of microglia in the normal adult mouse brain. Neuroscience 39:151–170. Leiknes KA, Jarosh-von Schweder L, Høie B (2012) Contemporary use and practice of electroconvulsive therapy worldwide. Brain Behav 2:283–344. Leonoudakis D, BRAITHWAITE SP, BEATTIE MS, BEATTIE EC (2004) TNFα-induced AMPA-receptor trafficking in CNS neurons; relevance to excitotoxicity? Neuron Glia Biol 1:263–273. Leppik IE, Rask CA (1988) Pharmacokinetics of antiepileptic drugs during pregnancy. Semin Neurol 8:240–246. Lévesque M, Avoli M, Bernard C (2016) Animal models of temporal lobe epilepsy following systemic chemoconvulsant administration. J Neurosci Methods, Methods and Models in Epilepsy Research 260:45–52.

- 174 -

Liang X, Samways DS, Wolf K, Bowles EA, Richards JP, Bruno J, Dutertre S, DiPaolo RJ, Egan TM (2015) Quantifying Ca2+ Current and Permeability in ATP-gated P2X7 Receptors. J Biol Chem 290:7930–7942. Librizzi L, Noè F, Vezzani A, Curtis M de, Ravizza T (2012) Seizure-induced brain-borne inflammation sustains seizure recurrence and blood–brain barrier damage. Ann Neurol 72:82–90. Lillis KP, Kramer MA, Mertz J, Staley KJ, White JA (2012) Pyramidal cells accumulate chloride at seizure onset. Neurobiol Dis 47:358–366. Limbrick DD, Sombati S, DeLorenzo RJ (2003) Calcium influx constitutes the ionic basis for the maintenance of glutamate-induced extended neuronal depolarization associated with hippocampal neuronal death. Cell Calcium 33:69–81. Liu Y, Li Y, Eyo UB, Chen T, Umpierre A, Zhu J, Bosco DB, Dong H, Wu L-J (2019) Neuronal network activity controls microglial process surveillance in awake mice via norepinephrine signaling. bioRxiv 557686. Locovei S, Scemes E, Qiu F, Spray DC, Dahl G (2007) Pannexin1 is part of the pore forming unit of the P2X(7) receptor death complex. FEBS Lett 581:483–488. Loo CK, Katalinic N, Smith DJ, Ingram A, Dowling N, Martin D, Addison K, Hadzi-Pavlovic D, Simpson B, Schweitzer I (2014) A randomized controlled trial of brief and ultrabrief pulse right unilateral electroconvulsive therapy. Int J Neuropsychopharmacol 18. Löscher W (2011) Critical review of current animal models of seizures and epilepsy used in the discovery and development of new antiepileptic drugs. Seizure 20:359–368. Löscher W (2002) Animal models of epilepsy for the development of antiepileptogenic and disease-modifying drugs. A comparison of the pharmacology of kindling and post- status epilepticus models of temporal lobe epilepsy. Epilepsy Res 50:105–123. Löscher W, Klitgaard H, Twyman RE, Schmidt D (2013) New avenues for anti-epileptic drug discovery and development. Nat Rev Drug Discov 12:757–776. Lou N, Takano T, Pei Y, Xavier AL, Goldman SA, Nedergaard M (2016) Purinergic receptor P2RY12-dependent microglial closure of the injured blood-brain barrier. Proc Natl Acad Sci U S A 113:1074–1079. Lovatt D, Xu Q, Liu W, Takano T, Smith NA, Schnermann J, Tieu K, Nedergaard M (2012) Neuronal adenosine release, and not astrocytic ATP release, mediates feedback inhibition of excitatory activity. Proc Natl Acad Sci 109:6265–6270. Lynch MA (2009) The multifaceted profile of activated microglia. Mol Neurobiol 40:139– 156. Madry C, Arancibia-Cárcamo LI, Kyrargyri V, Chan VT, Hamilton NB, Attwell D (2018) Effects of the ecto-ATPase apyrase on microglial ramification and surveillance reflect cell depolarization, not ATP depletion. Proc Natl Acad Sci 115:E1608–E1617. Madry C, Attwell D (2015) Receptors, ion channels, and signaling mechanisms underlying microglial dynamics. J Biol Chem 290:12443–12450. Madry C, Kyrargyri V, Lorena Arancibia-Cárcamo I, Jolivet R, Kohsaka S, M. Bryan R, Attwell D (2017) Microglial Ramification, Surveillance, and Interleukin-1β Release Are Regulated by the Two-Pore Domain K + Channel THIK-1. Neuron 97. Magid M, Truong L, Trevino K, Husain M (2013) Efficacy of right unilateral ultrabrief pulse width ECT: a preliminary report. J ECT 29:258–264.

- 175 -

Marchi N, Angelov L, Masaryk T, Fazio V, Granata T, Hernandez N, Hallene K, Diglaw T, Franic L, Najm I, Janigro D (2007) Seizure‐Promoting Effect of Blood–Brain Barrier Disruption. Epilepsia 48:732–742. Marchi N, Fan Q, Ghosh C, Fazio V, Bertolini F, Betto G, Batra A, Carlton E, Najm I, Granata T, Janigro D (2009) Antagonism of peripheral inflammation reduces the severity of status epilepticus. Neurobiol Dis 33:171–181. Marchi N, Granata T, Ghosh C, Janigro D (2012) Blood–brain barrier dysfunction and epilepsy: Pathophysiologic role and therapeutic approaches. Epilepsia 53:1877–1886. Marson A, Jacoby A, Johnson A, Kim L, Gamble C, Chadwick D, Medical Research Council MESS Study Group (2005) Immediate versus deferred antiepileptic drug treatment for early epilepsy and single seizures: a randomised controlled trial. The Lancet 365:2007–2013. Martineau FS, Fournier L, Buhler E, Watrin F, Sargolini F, Manent J-B, Poucet B, Represa A (2019) Spared cognitive and behavioral functions prior to epilepsy onset in a rat model of subcortical band heterotopia. Brain Res 1711:146–155. Masco D, Sahibzada N, Switzer R, Gale K (1999) Electroshock seizures protect against apoptotic hippocampal cell death induced by adrenalectomy. Neuroscience 91:1315– 1319. Mead EL, Mosley A, Eaton S, Dobson L, Heales SJ, Pocock JM (2012) Microglial neurotransmitter receptors trigger superoxide production in microglia; consequences for microglial–neuronal interactions. J Neurochem 121:287–301. Menteyne A, Levavasseur F, Audinat E, Avignone E (2009) Predominant Functional Expression of Kv1.3 by Activated Microglia of the Hippocampus after Status epilepticus. PLOS ONE 4:e6770. Merienne N, Meunier C, Schneider A, Seguin J, Nair SS, Rocher AB, Le Gras S, Keime C, Faull R, Pellerin L, Chatton J-Y, Neri C, Merienne K, Déglon N (2019) Cell-Type- Specific Gene Expression Profiling in Adult Mouse Brain Reveals Normal and Disease-State Signatures. Cell Rep 26:2477-2493.e9. Michel AD, Chessell IP, Humphrey PPA (1999) Ionic effects on human recombinant P2X7 receptor function. Naunyn Schmiedebergs Arch Pharmacol 359:102–109. Mikati MA, Abi‐Habib RJ, El Sabban ME, Dbaibo GS, Kurdi RM, Kobeissi M, Farhat F, Asaad W (2003) Hippocampal Programmed Cell Death after Status Epilepticus: Evidence for NMDA-Receptor and Ceramide-Mediated Mechanisms. Epilepsia 44:282–291. Milatovic D, Gupta RC, Dettbarn W-D (2002) Involvement of nitric oxide in kainic acid- induced excitotoxicity in rat brain. Brain Res 957:330–337. Mittelbronn M (2014) The M1/M2 immune polarization concept in microglia: a fair transfer? Neuroimmunol Neuroinflammation 1:6–7. Mizuno T, Kawanokuchi J, Numata K, Suzumura A (2003) Production and neuroprotective functions of fractalkine in the central nervous system. Brain Res 979:65–70. Monif M, Reid CA, Powell KL, Smart ML, Williams DA (2009) The P2X7 Receptor Drives Microglial Activation and Proliferation: A Trophic Role for P2X7R Pore. J Neurosci 29:3781–3791.

- 176 -

Monleon S, Parra A, Simon VM, Brain PF, D’Aquila P, Willner P (1995) Attenuation of sucrose consumption in mice by chronic mild stress and its restoration by imipramine. Psychopharmacology (Berl) 117:453–457. Morin-Brureau M, Milior G, Royer J, Chali F, LeDuigou C, Savary E, Blugeon C, Jourdren L, Akbar D, Dupont S, Navarro V, Baulac M, Bielle F, Mathon B, Clemenceau S, Miles R (2018) Microglial phenotypes in the human epileptic temporal lobe. Brain 141:3343–3360. Morinobu S, Strausbaugh H, Terwilliger R, Duman RS (1997) Regulation of c-Fos and NGF1-A by antidepressant treatments. Synapse 25:313–320. Murase K, Ryu PD, Randic M (1989) Excitatory and inhibitory amino acids and peptide- induced responses in acutely isolated rat spinal dorsal horn neurons. Neurosci Lett 103:56–63. Murgia M, Hanau S, Pizzo P, Rippa M, Virgilio FD (1993) Oxidized ATP. An irreversible inhibitor of the macrophage purinergic P2Z receptor. J Biol Chem 268:8199–8203. Nebeling FC, Poll S, Schmid LC, Mittag M, Steffen J, Keppler K, Fuhrmann M (2019) Microglia motility depends on neuronal activity and promotes structural plasticity in the hippocampus. bioRxiv 515759. Neher JJ, Neniskyte U, Hornik T, Brown GC (2014) Inhibition of UDP/P2Y6 purinergic signaling prevents phagocytosis of viable neurons by activated microglia in vitro and in vivo. Glia 62:1463–1475. Nieoczym D, Socała K, Wlaź P (2017) Evaluation of the Anticonvulsant Effect of Brilliant Blue G, a Selective P2X7 Receptor Antagonist, in the iv PTZ-, Maximal Electroshock- , and 6 Hz-Induced Seizure Tests in Mice. Neurochem Res 42:3114–3124. Nimmerjahn A, Kirchhoff F, Helmchen F (2005) Resting Microglial Cells Are Highly Dynamic Surveillants of Brain Parenchyma in Vivo. Science 308:1314–1318. Nimmerjahn A, Kirchhoff F, Kerr JN, Helmchen F (2004) Sulforhodamine 101 as a specific marker of astroglia in the neocortex in vivo. Nat Methods 1. Nishiyori A, Minami M, Ohtani Y, Takami S, Yamamoto J, Kawaguchi N, Kume T, Akaike A, Satoh M (1998) Localization of fractalkine and CX3CR1 mRNAs in rat brain: does fractalkine play a role in signaling from neuron to microglia? FEBS Lett 429:167–172. Noé FM, Bellistri E, Colciaghi F, Cipelletti B, Battaglia G, de Curtis M, Librizzi L (2016) Kainic acid-induced albumin leak across the blood-brain barrier facilitates epileptiform hyperexcitability in limbic regions. Epilepsia 57:967–976. Nollet M, Guisquet A-ML, Belzung C (2013) Models of Depression: Unpredictable Chronic Mild Stress in Mice. Curr Protoc Pharmacol 61:5.65.1-5.65.17. North RA (2002) Molecular physiology of P2X receptors. Physiol Rev 82:1013–1067. Nuttle LC, el-Moatassim C, Dubyak GR (1993) Expression of the pore-forming P2Z purinoreceptor in Xenopus oocytes injected with poly(A)+ RNA from murine macrophages. Mol Pharmacol 44:93–101. Oakley JC, Kalume F, Yu FH, Scheuer T, Catterall WA (2009) Temperature- and age- dependent seizures in a mouse model of severe myoclonic epilepsy in infancy. Proc Natl Acad Sci 106:3994–3999. Oberheim NA, Tian G-F, Han X, Peng W, Takano T, Ransom B, Nedergaard M (2008) Loss of astrocytic domain organization in the epileptic brain. J Neurosci Off J Soc Neurosci 28:3264–3276.

- 177 -

Ohsawa K, Irino Y, Sanagi T, Nakamura Y, Suzuki E, Inoue K, Kohsaka S (2010) P2Y12 receptor-mediated integrin-beta1 activation regulates microglial process extension induced by ATP. Glia 58:790–801. Ohsawa K, Kohsaka S (2011) Dynamic motility of microglia: Purinergic modulation of microglial movement in the normal and pathological brain. Glia 59:1793–1799. Ohsawa K, Sanagi T, Nakamura Y, Suzuki E, Inoue K, Kohsaka S (2012) Adenosine A3 receptor is involved in ADP‐induced microglial process extension and migration. J Neurochem 121:217–227. Olafsson E, Hauser WA (1999) Prevalence of Epilepsy in Rural Iceland: A Population‐Based Study. Epilepsia 40:1529–1534. Olmedillas del Moral M, Asavapanumas N, Uzcategui N, Garaschuk O (2019) Healthy Brain Aging Modifies Microglial Calcium Signaling In Vivo. Int J Mol Sci 20:589. Orihuela R, McPherson CA, Harry GJ (2016) Microglial M1/M2 polarization and metabolic states. Br J Pharmacol 173:649–665. Orr AG, Orr AL, Li X-J, Gross RE, Traynelis SF (2009) Adenosine A2A receptor mediates microglial process retraction. Nat Neurosci 12. Orzi F, Zoli M, Passarelli F, Ferraguti F, Fieschi C, Agnati LF (1990) Repeated electroconvulsive shock increases glial fibrillary acidic protein, ornithine decarboxylase, somatostatin and cholecystokinin immunoreactivities in the hippocampal formation of the rat. Brain Res 533:223–231. Osler M, Rozing MP, Christensen GT, Andersen PK, Jørgensen MB (2018) Electroconvulsive therapy and risk of dementia in patients with affective disorders: a cohort study. Lancet Psychiatry 5:348–356. Paolicelli RC, Bolasco G, Pagani F, Maggi L, Scianni M, Panzanelli P, Giustetto M, Ferreira TA, Guiducci E, Dumas L, Ragozzino D, Gross CT (2011) Synaptic Pruning by Microglia Is Necessary for Normal Brain Development. Science 333:1456–1458. Parkhurst CN, Gan W-B (2010) Microglia dynamics and function in the CNS. Curr Opin Neurobiol 20:595–600. Parkhurst CN, Yang G, Ninan I, Savas JN, Yates JR, Lafaille JJ, Hempstead BL, Littman DR, Gan W-B (2013) Microglia Promote Learning-Dependent Synapse Formation through Brain-Derived Neurotrophic Factor. Cell 155:1596–1609. Pascual O, Achour SB, Rostaing P, Triller A, Bessis A (2012) Microglia activation triggers astrocyte-mediated modulation of excitatory neurotransmission. Proc Natl Acad Sci 109:E197–E205. Patterson KP, Brennan GP, Curran M, Kinney-Lang E, Dubé C, Rashid F, Ly C, Obenaus A, Baram TZ (2015) Rapid, Coordinate Inflammatory Responses after Experimental Febrile Status Epilepticus: Implications for Epileptogenesis. eNeuro 2:15. Pelegrin P, Surprenant A (2006) Pannexin-1 mediates large pore formation and interleukin- 1beta release by the ATP-gated P2X7 receptor. EMBO J 25:5071–5082. Penkowa M, Florit S, Giralt M, Quintana A, Molinero A, Carrasco J, Hidalgo J (2005) Metallothionein reduces central nervous system inflammation, neurodegeneration, and cell death following kainic acid-induced epileptic seizures. J Neurosci Res 79:522– 534.

- 178 -

Perera TD, Luber B, Nobler MS, Prudic J, Anderson C, Sackeim HA (2004) Seizure Expression During Electroconvulsive Therapy: Relationships with Clinical Outcome and Cognitive Side Effects. Neuropsychopharmacology 29:813–825. Peterson SL (1998) Electroshock. In: Neuropharmacology Methods in Epilepsy Research, Methods in the Life Sciences: Cellular and Molecular Neuropharmacology (Peterson SL, Albertson TE eds), pp1–26. Boca Raton: CRC Press. Pierce JP, Melton J, Punsoni M, McCloskey DP, Scharfman HE (2005) Mossy fibers are the primary source of afferent input to ectopic granule cells that are born after pilocarpine- induced seizures. Exp Neurol 196:316–331. Pippel A, Stolz M, Woltersdorf R, Kless A, Schmalzing G, Markwardt F (2017) Localization of the gate and selectivity filter of the full-length P2X7 receptor. Proc Natl Acad Sci 114:E2156–E2165. Plata A, Lebedeva A, Denisov P, Nosova O, Postnikova TY, Pimashkin A, Brazhe A, Zaitsev AV, Rusakov DA, Semyanov A (2018) Astrocytic Atrophy Following Status Epilepticus Parallels Reduced Ca2+ Activity and Impaired Synaptic Plasticity in the Rat Hippocampus. Front Mol Neurosci 11:215. Plata-Salamán CR, Ilyin SE, Turrin NP, Gayle D, Flynn MC, Romanovitch AE, Kelly ME, Bureau Y, Anisman H, McIntyre DC (2000) Kindling modulates the IL-1β system, TNF-α, TGF-β1, and neuropeptide mRNAs in specific brain regions. Mol Brain Res 75:248–258. Pollard H, Charriaut-Marlangue C, Cantagrel S, Represa A, Robain O, Moreau J, Ben-Ari Y (1994) Kainate-induced apoptotic cell death in hippocampal neurons. Neuroscience 63:7–18. Postnikova TY, Zubareva OE, Kovalenko AA, Kim KK, Magazanik LG, Zaitsev AV (2017) Status epilepticus impairs synaptic plasticity in rat hippocampus and is followed by changes in expression of NMDA receptors. Biochem Mosc 82:282–290. Pozner A, Xu B, Palumbos S, Gee JM, Tvrdik P, Capecchi MR (2015) Intracellular calcium dynamics in cortical microglia responding to focal laser injury in the PC::G5-tdT reporter mouse. Front Mol Neurosci 8. Qiu F, Dahl G (2009) A permeant regulating its permeation pore: inhibition of pannexin 1 channels by ATP. Am J Physiol Cell Physiol 296:C250-255. Qiu F, Wang J, Dahl G (2012) Alanine substitution scanning of pannexin1 reveals amino acid residues mediating ATP sensitivity. Purinergic Signal 8:81–90. Racine RJ (1972) Modification of seizure activity by electrical stimulation. II. Motor seizure. Electroencephalogr Clin Neurophysiol 32:281–294. Ragozzino D, Angelantonio SD, Trettel F, Bertollini C, Maggi L, Gross C, Charo IF, Limatola C, Eusebi F (2006) Chemokine Fractalkine/CX3CL1 Negatively Modulates Active Glutamatergic Synapses in Rat Hippocampal Neurons. J Neurosci 26:10488– 10498. Raimondo JV, Burman RJ, Katz AA, Akerman CJ (2015) Ion dynamics during seizures. Front Cell Neurosci 9:419. Rasmussen R, Nedergaard M, Petersen NC (2016) Sulforhodamine 101, a widely used astrocyte marker, can induce cortical seizure-like activity at concentrations commonly used. Sci Rep 6:30433.

- 179 -

Rassendren F, Buell GN, Virginio C, Collo G, North RA, Surprenant A (1997) The Permeabilizing ATP Receptor, P2X7. J Biol Chem 272:5482–5486. Ravizza T, Vezzani A (2006) Status epilepticus induces time-dependent neuronal and astrocytic expression of interleukin-1 receptor type I in the rat limbic system. Neuroscience 137:301–308. Reisinger TL, Newman M, Loring DW, Pennell PB, Meador KJ (2013) Antiepileptic drug clearance and seizure frequency during pregnancy in women with epilepsy. Epilepsy Behav EB 29:13–18. Rettenbeck ML, Rüden E-L von, Bienas S, Carlson R, Stein VM, Tipold A, Potschka H (2015) Microglial ROS production in an electrical rat post-status epilepticus model of epileptogenesis. Neurosci Lett 599:146–151. Réu P, Khosravi A, Bernard S, Mold JE, Salehpour M, Alkass K, Perl S, Tisdale J, Possnert G, Druid H, Frisén J (2017) The Lifespan and Turnover of Microglia in the Human Brain. Cell Rep 20:779–784. Reynolds EH, Elwes RD, Shorvon SD (1983) Why does epilepsy become intractable? Prevention of chronic epilepsy. The Lancet 2:952–954. Rezaie P, Male D (2002) Mesoglia & microglia--a historical review of the concept of mononuclear phagocytes within the central nervous system. J Hist Neurosci 11:325– 374. Rizvi S, Ladino LD, Hernandez-Ronquillo L, Téllez-Zenteno JF (2017) Epidemiology of early stages of epilepsy: Risk of seizure recurrence after a first seizure. Seizure 49:46– 53. Rogawski MA, Löscher W (2004) The neurobiology of antiepileptic drugs. Nat Rev Neurosci 5:553–564. Ross EL, Zivin K, Maixner DF (2018) Cost-effectiveness of Electroconvulsive Therapy vs Pharmacotherapy/Psychotherapy for Treatment-Resistant Depression in the United States. JAMA Psychiatry 75:713–722. Sabilallah M, Fontanaud P, Linck N, Boussadia B, Peyroutou R, Lasgouzes T, Rassendren FA, Marchi N, Hirbec HE (2016) Evidence for Status Epilepticus and Pro- Inflammatory Changes after Intranasal Kainic Acid Administration in Mice. PloS One 11. Sackeim HA (2017) Modern Electroconvulsive Therapy: Vastly Improved yet Greatly Underused. JAMA Psychiatry 74:779–780. Sackeim HA, Prudic J, Nobler MS, Fitzsimons L, Lisanby SH, Payne N, Berman RM, Brakemeier E-L, Perera T, Devanand DP (2008) Effects of pulse width and electrode placement on the efficacy and cognitive effects of electroconvulsive therapy. Brain Stimulat 1:71–83. Sadek A-R, Knight GE, Burnstock G (2011) Electroconvulsive therapy: a novel hypothesis for the involvement of purinergic signalling. Purinergic Signal 7:447–452. Sadowsky JH (2017) Electroconvulsive therapy in America: the anatomy of a medical controversy, Routledge studies in cultural history. New York: Routledge/Taylor & Francis Group. Salgueiro-Pereira AR, Duprat F, Pousinha PA, Loucif A, Douchamps V, Regondi C, Ayrault M, Eugie M, Stunault MI, Escayg A, Goutagny R, Gnatkovsky V, Frassoni C, Marie

- 180 -

H, Bethus I, Mantegazza M (2019) A two-hit story: Seizures and genetic mutation interaction sets phenotype severity in SCN1A epilepsies. Neurobiol Dis 125:31–44. Salter MW, Stevens B (2017) Microglia emerge as central players in brain disease. Nat Med 23:1018–1027. Sasaki Y, Hoshi M, Akazawa C, Nakamura Y, Tsuzuki H, Inoue K, Kohsaka S (2003) Selective expression of Gi/o-coupled ATP receptor P2Y12 in microglia in rat brain. Glia 44:242–250. Sasmono RT (2003) A macrophage colony-stimulating factor receptor-green fluorescent protein transgene is expressed throughout the mononuclear phagocyte system of the mouse. Blood 101:1155–1163. Satoh J, Kino Y, Asahina N, Takitani M, Miyoshi J, Ishida T, Saito Y (2016) TMEM119 marks a subset of microglia in the human brain. Neuropathology 36:39–49. Saul A, Hausmann R, Kless A, Nicke A (2013) Heteromeric assembly of P2X subunits. Front Cell Neurosci 7:250. Scemes E, Giaume C (2006) Astrocyte calcium waves: what they are and what they do. Glia 54:716–725. Schachter J, Motta AP, Zamorano A de S, Silva-Souza HA da, Guimarães MZP, Persechini PM (2008) ATP-induced P2X7-associated uptake of large molecules involves distinct mechanisms for cations and anions in macrophages. J Cell Sci 121:3261–3270. Schafer DP, Lehrman EK, Stevens B (2013) The “quad‐partite” synapse: Microglia‐synapse interactions in the developing and mature CNS. Glia 61:24–36. Scharfman HE, Malthankar-Phatak GH, Friedman D, Pearce P, McCloskey DP, Harden CL, Maclusky NJ (2009) A rat model of epilepsy in women: a tool to study physiological interactions between endocrine systems and seizures. Endocrinology 150:4437–4442. Schartz ND, Herr SA, Madsen L, Butts SJ, Torres C, Mendez LB, Brewster AL (2016) Spatiotemporal profile of Map2 and microglial changes in the hippocampal CA1 region following pilocarpine-induced status epilepticus. Sci Rep 6:24988. Schauwecker PE (2012) Strain differences in seizure-induced cell death following pilocarpine-induced status epilepticus. Neurobiol Dis 45:297–304. Schauwecker PE (2003) Genetic basis of kainate-induced excitotoxicity in mice: phenotypic modulation of seizure-induced cell death. Epilepsy Res 55:201–210. Schauwecker PE, Steward O (1997) Genetic determinants of susceptibility to excitotoxic cell death: implications for gene targeting approaches. Proc Natl Acad Sci U S A 94:4103– 4108. Scherz‐Shouval R, Shvets E, Fass E, Shorer H, Gil L, Elazar Z (2007) Reactive oxygen species are essential for autophagy and specifically regulate the activity of Atg4. EMBO J 26:1749–1760. Schipke CG, Boucsein C, Ohlemeyer C, Kirchhoff F, Kettenmann H (2002) Astrocyte Ca2+ waves trigger responses in microglial cells in brain slices. FASEB J Off Publ Fed Am Soc Exp Biol 16:255–257. Schizophrenia Working Group of the Psychiatric Genomics Consortium, Sekar A, Bialas AR, de Rivera H, Davis A, Hammond TR, Kamitaki N, Tooley K, Presumey J, Baum M, Van Doren V, Genovese G, Rose SA, Handsaker RE, Daly MJ, Carroll MC, Stevens B, McCarroll SA (2016) Schizophrenia risk from complex variation of complement component 4. Nature 530:177–183.

- 181 -

Schmued LC, Stowers CC, Scallet AC, Xu L (2005) Fluoro-Jade C results in ultra high resolution and contrast labeling of degenerating neurons. Brain Res 1035:24–31. Schnell C, Hagos Y, Hülsmann S (2012) Active sulforhodamine 101 uptake into hippocampal astrocytes. PloS One 7:e49398. Schnell C, Shahmoradi A, Wichert SP, Mayerl S, Hagos Y, Heuer H, Rossner MJ, Hülsmann S (2015) The multispecific thyroid hormone transporter OATP1C1 mediates cell- specific sulforhodamine 101-labeling of hippocampal astrocytes. Brain Struct Funct 220:193–203. Schwieler L, Samuelsson M, Frye MA, Bhat M, Schuppe-Koistinen I, Jungholm O, Johansson AG, Landén M, Sellgren CM, Erhardt S (2016) Electroconvulsive therapy suppresses the neurotoxic branch of the kynurenine pathway in treatment-resistant depressed patients. J Neuroinflammation 13:51. Scorza FA, Arida RM, Priel MR, Calderazzo L, Cavalheiro EA (2002) Glucose utilisation during status epilepticus in an epilepsy model induced by pilocarpine: a qualitative study. Arq Neuropsiquiatr 60:198–203. Selassie AW, Wilson DA, Martz GU, Smith GG, Wagner JL, Wannamaker BB (2014) Epilepsy beyond seizure: A population-based study of comorbidities. Epilepsy Res 108:305–315. Sepulveda-Rodriguez A, Li P, Khan T, Ma JD, Carlone CA, Bozzelli PL, Conant KE, Forcelli PA, Vicini S (2019) Electroconvulsive Shock Enhances Responsive Motility and Purinergic Currents in Microglia in the Mouse Hippocampus. eNeuro 6:ENEURO.0056-19.2019. Shapiro LA, Wang L, Ribak CE (2008) Rapid astrocyte and microglial activation following pilocarpine‐induced seizures in rats. Epilepsia 49:33–41. Shapouri‐Moghaddam A, Mohammadian S, Vazini H, Taghadosi M, Esmaeili S, Mardani F, Seifi B, Mohammadi A, Afshari JT, Sahebkar A (2018) Macrophage plasticity, polarization, and function in health and disease. J Cell Physiol 233:6425–6440. Shaw JAG, Perry VH, Mellanby J (1990) Tetanus toxin-induced seizures cause microglial activation in rat hippocampus. Neurosci Lett 120:66–69. Shibley H, Smith BN (2002) Pilocarpine-induced status epilepticus results in mossy fiber sprouting and spontaneous seizures in C57BL/6 and CD-1 mice. Epilepsy Res 49:109– 120. Shigetomi E, Bushong EA, Haustein MD, Tong X, Jackson-Weaver O, Kracun S, Xu J, Sofroniew MV, Ellisman MH, Khakh BS (2013) Imaging calcium microdomains within entire astrocyte territories and endfeet with GCaMPs expressed using adeno- associated viruses. J Gen Physiol 141:633–647. Shoji KF, Sáez PJ, Harcha PA, Aguila HL, Sáez JC (2014) Pannexin1 channels act downstream of P2X7 receptors in ATP-induced murine T-cell death. Channels 8:142– 156. Sierra A, de Castro F, Del Río-Hortega J, Rafael Iglesias-Rozas J, Garrosa M, Kettenmann H (2016) The “Big-Bang” for modern glial biology: Translation and comments on Pío del Río-Hortega 1919 series of papers on microglia. Glia 64:1801–1840. Sinel’nikova VV, Shubina LV, Gol’tyaev MV, Loseva EV, Kichigina VF (2013) Detection of c-Fos Expression in the Brains of Animals with a Pilocarpine Model of Temporal Lobe Epilepsy. Neurosci Behav Physiol 43:1084–1091.

- 182 -

Sipe G, Lowery R, Tremblay M-È, Kelly E, Lamantia C, Majewska A (2016) Microglial P2Y12 is necessary for synaptic plasticity in mouse visual cortex. Nat Commun 7. Sipes M, Matson JL, Belva B, Turygin N, Kozlowski AM, Horovitz M (2011) The relationship among side effects associated with anti-epileptic medications in those with intellectual disability. Res Dev Disabil 32:1646–1651. Slade EP, Jahn DR, Regenold WT, Case BG (2017) Association of Electroconvulsive Therapy With Psychiatric Readmissions in US Hospitals. JAMA Psychiatry 74:798– 804. Slomko AM, Naseer Z, Ali SS, Wongvravit JP, Friedman LK (2014) Retigabine calms seizure-induced behavior following status epilepticus. Epilepsy Behav EB 37:123– 132. Sokolowski JD, Chabanon-Hicks CN, Han CZ, Heffron DS, Mandell JW (2014) Fractalkine is a “find-me” signal released by neurons undergoing ethanol-induced apoptosis. Front Cell Neurosci 8:360. Sousa C, Golebiewska A, Poovathingal SK, Kaoma T, Pires‐Afonso Y, Martina S, Coowar D, Azuaje F, Skupin A, Balling R, Biber K, Niclou SP, Michelucci A (2018) Single‐cell transcriptomics reveals distinct inflammation‐induced microglia signatures. EMBO Rep 19:e46171. Spaans H-P, Verwijk E, Comijs HC, Kok RM, Sienaert P, Bouckaert F, Fannes K, Vandepoel K, Scherder EJA, Stek ML, Kho KH (2013) Efficacy and cognitive side effects after brief pulse and ultrabrief pulse right unilateral electroconvulsive therapy for major depression: a randomized, double-blind, controlled study. J Clin Psychiatry 74:e1029- 1036. Sperlágh B, Illes P (2014) P2X7 receptor: an emerging target in central nervous system diseases. Trends Pharmacol Sci 35:537–547. Srinivasan R, Lu T-Y, Chai H, Xu J, Huang BS, Golshani P, Coppola G, Khakh BS (2016) New Transgenic Mouse Lines for Selectively Targeting Astrocytes and Studying Calcium Signals in Astrocyte Processes In Situ and In Vivo. Neuron 92:1181–1195. Stellwagen D, Beattie EC, Seo JY, Malenka RC (2005) Differential Regulation of AMPA Receptor and GABA Receptor Trafficking by Tumor Necrosis Factor-α. J Neurosci 25:3219–3228. Stence N, Waite M, Dailey ME (2001) Dynamics of microglial activation: a confocal time- lapse analysis in hippocampal slices. Glia 33:256–266. Stevens B, Allen NJ, Vazquez LE, Howell GR, Christopherson KS, Nouri N, Micheva KD, Mehalow AK, Huberman AD, Stafford B, Sher A, Litke AM, Lambris JD, Smith SJ, John SWM, Barres BA (2007) The classical complement cascade mediates CNS synapse elimination. Cell 131:1164–1178. Stewart K-AA, Wilcox KS, Fujinami RS, White HS (2010) Theiler’s virus infection chronically alters seizure susceptibility. Epilepsia 51:1418–1428. Stokes L, Layhadi JA, Bibic L, Dhuna K, Fountain SJ (2017) P2X4 Receptor Function in the Nervous System and Current Breakthroughs in Pharmacology. Front Pharmacol 8. Stowell RD, Sipe GO, Dawes RP, Batchelor HN, Lordy KA, Bidlack JM, Brown E, Sur M, Majewska AK (2019) Noradrenergic signaling in wakeful states inhibits microglial surveillance and synaptic plasticity in the mouse visual cortex. bioRxiv 556480.

- 183 -

Strine TW, Kobau R, Chapman DP, Thurman DJ, Price P, Balluz LS (2005) Psychological Distress, Comorbidities, and Health Behaviors among U.S. Adults with Seizures: Results from the 2002 National Health Interview Survey. Epilepsia 46:1133–1139. Suárez LM, Cid E, Gal B, Inostroza M, Brotons-Mas JR, Gómez-Domínguez D, de la Prida LM, Solís JM (2012) Systemic injection of kainic acid differently affects LTP magnitude depending on its epileptogenic efficiency. PloS One 7:e48128. Sun C, Heid ME, Keyel PA, Salter RD (2013) The Second Transmembrane Domain of P2X7 Contributes to Dilated Pore Formation. PLOS ONE 8:e61886. Surprenant A, Rassendren F, Kawashima E, North RA, Buell G (1996) The cytolytic P2Z receptor for extracellular ATP identified as a P2X receptor (P2X7). Science 272:735– 738. Suurväli J, Boudinot P, Kanellopoulos J, Rüütel Boudinot S (2017) P2X4: A fast and sensitive purinergic receptor. Biomed J 40:245–256. Sweeney MD, Zhao Z, Montagne A, Nelson AR, Zlokovic BV (2019) Blood-Brain Barrier: From Physiology to Disease and Back. Physiol Rev 99:21–78. Tang Y, Le W (2016) Differential Roles of M1 and M2 Microglia in Neurodegenerative Diseases. Mol Neurobiol 53:1181–1194. Tay TL et al. (2017) A new fate mapping system reveals context-dependent random or clonal expansion of microglia. Nat Neurosci 20:793–803. Thion MS, Garel S (2017) On place and time: microglia in embryonic and perinatal brain development. Curr Opin Neurobiol, Glial Biology 47:121–130. Thion MS, Ginhoux F, Garel S (2018) Microglia and early brain development: An intimate journey. Science 362:185–189. Thomas DC (2017) The phagocyte respiratory burst: Historical perspectives and recent advances. Immunol Lett 192:88–96. Tian D-SS, Peng J, Murugan M, Feng L-JJ, Liu J-LL, Eyo UB, Zhou L-JJ, Mogilevsky R, Wang W, Wu L-JJ (2017) Chemokine CCL2-CCR2 Signaling Induces Neuronal Cell Death via STAT3 Activation and IL-1β Production after Status Epilepticus. J Neurosci 37:7878–7892. Tian G-F, Azmi H, Takano T, Xu Q, Peng W, Lin J, Oberheim N, Lou N, Wang X, Zielke HR, Kang J, Nedergaard M (2005) An astrocytic basis of epilepsy. Nat Med 11:973– 981. Tokuhara D, Sakuma S, Hattori H, Matsuoka O, Yamano T (2007) Kainic acid dose affects delayed cell death mechanism after status epilepticus. Brain Dev 29:2–8. Toman JEP, Goodman LS (1948) Anticonvulsants. Physiol Rev 28:409–432. Tørring N, Sanghani SN, Petrides G, Kellner CH, Østergaard SD (2017) The mortality rate of electroconvulsive therapy: a systematic review and pooled analysis. Acta Psychiatr Scand 135:388–397. Toyomitsu E, Tsuda M, Yamashita T, Tozaki-Saitoh H, Tanaka Y, Inoue K (2012) CCL2 promotes P2X4 receptor trafficking to the cell surface of microglia. Purinergic Signal 8:301–310. Tremblay M-È, Lowery RL, Majewska AK (2010) Microglial Interactions with Synapses Are Modulated by Visual Experience. PLoS Biol 8:e1000527. Tremblay M-È, Stevens B, Sierra A, Wake H, Bessis A, Nimmerjahn A (2011) The Role of Microglia in the Healthy Brain. J Neurosci 31:16064–16069.

- 184 -

Tremblay M-È, Zettel ML, Ison JR, Allen PD, Majewska AK (2012) Effects of aging and sensory loss on glial cells in mouse visual and auditory cortices. Glia 60:541–558. Trinka E, Cock H, Hesdorffer D, Rossetti AO, Scheffer IE, Shinnar S, Shorvon S, Lowenstein DH (2015) A definition and classification of status epilepticus--Report of the ILAE Task Force on Classification of Status Epilepticus. Epilepsia 56:1515–1523. Tronel C, Largeau B, Santiago Ribeiro MJ, Guilloteau D, Dupont A-C, Arlicot N (2017) Molecular Targets for PET Imaging of Activated Microglia: The Current Situation and Future Expectations. Int J Mol Sci 18. Tupal S, Faingold CL (2019) Fenfluramine, a serotonin-releasing drug, prevents seizure- induced respiratory arrest and is anticonvulsant in the DBA/1 mouse model of SUDEP. Epilepsia 60:485–494. Turski WA, Cavalheiro EA, Bortolotto ZA, Mello LM, Schwarz M, Turski L (1984) Seizures produced by pilocarpine in mice: A behavioral, electroencephalographic and morphological analysis. Brain Res 321:237–253. Turski WA, Cavalheiro EA, Schwarz M, Czuczwar SJ, Kleinrok Z, Turski L (1983) Limbic seizures produced by pilocarpine in rats: behavioural, electroencephalographic and neuropathological study. Behav Brain Res 9:315–335. UK ECT Review Group (2003) Efficacy and safety of electroconvulsive therapy in depressive disorders: a systematic review and meta-analysis. The Lancet 361:799–808. Ulmann L, Levavasseur F, Avignone E, Peyroutou R, Hirbec H, Audinat E, Rassendren F (2013) Involvement of P2X4 receptors in hippocampal microglial activation after status epilepticus. Glia 61:1306–19. van Buel EM, Patas K, Peters M, Bosker FJ, Eisel ULM, Klein HC (2015) Immune and neurotrophin stimulation by electroconvulsive therapy: is some inflammation needed after all? Transl Psychiatry 5:e609. van Buel EM, Sigrist H, Seifritz E, Fikse L, Bosker FJ, Schoevers RA, Klein HC, Pryce CR, Eisel UL (2017) Mouse repeated electroconvulsive seizure (ECS) does not reverse social stress effects but does induce behavioral and hippocampal changes relevant to electroconvulsive therapy (ECT) side-effects in the treatment of depression. PLOS ONE 12. Van Landingham KE, Lothman EW (1991) Self-sustaining limbic status epilepticus. I. Acute and chronic cerebral metabolic studies: limbic hypermetabolism and neocortical hypometabolism. Neurology 41:1942–1949. van Vliet EA, Araújo CS da, Redeker S, Schaik R van, Aronica E, Gorter J (2007) Blood– brain barrier leakage may lead to progression of temporal lobe epilepsy. Brain 130:521–534. Varvel NH, Neher JJ, Bosch A, Wang W, Ransohoff RM, Miller RJ, Dingledine R (2016) Infiltrating monocytes promote brain inflammation and exacerbate neuronal damage after status epilepticus. Proc Natl Acad Sci 113. Vasek MJ et al. (2016) A complement–microglial axis drives synapse loss during virus- induced memory impairment. Nature 534:538–543. Verkhratsky A, Noda M (2014) General Physiology and Pathophysiology of Microglia. In: Neuroinflammation and Neurodegeneration (Peterson PK, Toborek M eds), pp47–60. New York, NY: Springer New York.

- 185 -

Verwijk E, Comijs HC, Kok RM, Spaans H-P, Stek ML, Scherder EJA (2012) Neurocognitive effects after brief pulse and ultrabrief pulse unilateral electroconvulsive therapy for major depression: a review. J Affect Disord 140:233– 243. Vezzani A, Balosso S, Ravizza T (2008) The role of cytokines in the pathophysiology of epilepsy. Brain Behav Immun 22:797–803. Vezzani A, Baram TZ (2007) New Roles for Interleukin-1 Beta in the Mechanisms of Epilepsy. Epilepsy Curr 7:45–50. Vezzani A, French J, Bartfai T, Baram TZ (2011) The role of inflammation in epilepsy. Nat Rev Neurol 7:31–40. Vezzani A, Viviani B (2015) Neuromodulatory properties of inflammatory cytokines and their impact on neuronal excitability. Neuropharmacology 96:70–82. Virgilio FD, Chiozzi P, Falzoni S, Ferrari D, Sanz JM, Venketaraman V, Baricordi OR (1998) Cytolytic P2X purinoceptors. Cell Death Differ 5:191–199. Virgilio FD, Schmalzing G, Markwardt F (2018) The Elusive P2X7 Macropore. Trends Cell Biol 28:392–404. Viviani B, Bartesaghi S, Gardoni F, Vezzani A, Behrens MM, Bartfai T, Binaglia M, Corsini E, Luca MD, Galli CL, Marinovich M (2003) Interleukin-1β Enhances NMDA Receptor-Mediated Intracellular Calcium Increase through Activation of the Src Family of Kinases. J Neurosci 23:8692–8700. Wake H, Moorhouse AJ, Jinno S, Kohsaka S, Nabekura J (2009) Resting Microglia Directly Monitor the Functional State of Synapses In Vivo and Determine the Fate of Ischemic Terminals. J Neurosci 29:3974–3980. Wang B, Sluyter R (2013) P2X7 receptor activation induces reactive oxygen species formation in erythroid cells. Purinergic Signal 9:101–112. Wang J, Jackson DG, Dahl G (2013) The food dye FD&C Blue No. 1 is a selective inhibitor of the ATP release channel Panx1. J Gen Physiol 141:649–656. Wang M, Zhang L, Gage FH (2019) Microglia, complement and schizophrenia. Nat Neurosci 22:333–334. Wang S, Cheng Q, Malik S, Yang J (2000) Interleukin-1beta inhibits gamma-aminobutyric acid type A (GABA(A)) receptor current in cultured hippocampal neurons. J Pharmacol Exp Ther 292:497–504. Watts BV, Groft A, Bagian JP, Mills PD (2011) An examination of mortality and other adverse events related to electroconvulsive therapy using a national adverse event report system. J ECT 27:105–108. Weiner RD, Reti IM (2017) Key updates in the clinical application of electroconvulsive therapy. Int Rev Psychiatry 29:54–62. Weinhard L, Bartolomei G di, Bolasco G, Machado P, Schieber NL, Neniskyte U, Exiga M, Vadisiute A, Raggioli A, Schertel A, Schwab Y, Gross CT (2018) Microglia remodel synapses by presynaptic trogocytosis and spine head filopodia induction. Nat Commun 9:1228. West PJ, Saunders GW, Billingsley P, Smith MD, White HS, Metcalf CS, Wilcox KS (2018) Recurrent epileptiform discharges in the medial entorhinal cortex of kainate‐treated rats are differentially sensitive to antiseizure drugs. Epilepsia 59:2035–2048.

- 186 -

Weston J, Bromley R, Jackson CF, Adab N, Clayton-Smith J, Greenhalgh J, Hounsome J, McKay AJ, Tudur Smith C, Marson AG (2016) Monotherapy treatment of epilepsy in pregnancy: congenital malformation outcomes in the child. Cochrane Database Syst Rev 11:CD010224. Wickert LE, Blanchette JB, Waldschmidt NV, Bertics PJ, Denu JM, Denlinger LC, Lenertz LY (2013) The C-Terminus of Human Nucleotide Receptor P2X7 Is Critical for Receptor Oligomerization and N-Linked Glycosylation. PLOS ONE 8:e63789. Wieraszko A, Goldsmith G, Seyfried TN (1989) Stimulation-dependent release of adenosine triphosphate from hippocampal slices. Brain Res 485:244–250. Wilcox KS, Gee JM, Gibbons MB, Tvrdik P, White JA (2015) Altered structure and function of astrocytes following status epilepticus. Epilepsy Behav EB 49:17–19. Winland CD, Welsh N, Sepulveda-Rodriguez A, Vicini S, Maguire-Zeiss KA (2017) Inflammation alters AMPA-stimulated calcium responses in dorsal striatal D2 but not D1 spiny projection neurons. Eur J Neurosci 46:2519–2533. Wirenfeldt M, Dissing-Olesen L, Babcock AA, Nielsen M, Meldgaard M, Zimmer J, Azcoitia I, Leslie RGQ, Dagnaes-Hansen F, Finsen B (2007) Population Control of Resident and Immigrant Microglia by Mitosis and Apoptosis. Am J Pathol 171:617–631. Wolf SA, Boddeke HWGM, Kettenmann H (2017) Microglia in Physiology and Disease. Annu Rev Physiol 79:619–643. Wolf Y, Yona S, Kim K-W, Jung S (2013) Microglia, seen from the CX3CR1 angle. Front Cell Neurosci 7. Wright B (1990) An Historical Review of Electroconvulsive Therapy. Jefferson J Psychiatry 8. Wu L-J, Vadakkan KI, Zhuo M (2007) ATP-induced chemotaxis of microglial processes requires P2Y receptor-activated initiation of outward potassium currents. Glia 55:810– 821. Wu P, Phillis J (1978) Distribution and release of adenosine triphosphate in rat brain. Neurochem Res 3:563–571. Wyatt-Johnson SK, Herr SA, Brewster AL (2017) Status Epilepticus Triggers Time- Dependent Alterations in Microglia Abundance and Morphological Phenotypes in the Hippocampus. Front Neurol 8:700. Xiang Z, Burnstock G (2005) Expression of P2X receptors on rat microglial cells during early development. Glia 52:119–126. Xu J, Bernstein AM, Wong A, Lu X-H, Khoja S, Yang XW, Davies DL, Micevych P, Sofroniew MV, Khakh BS (2016) P2X4 Receptor Reporter Mice: Sparse Brain Expression and Feeding-Related Presynaptic Facilitation in the Arcuate Nucleus. J Neurosci 36:8902–8920. Yan Z, Khadra A, Sherman A, Stojilkovic SS (2011) Calcium-dependent block of P2X7 receptor channel function is allosteric. J Gen Physiol 138:437–452. Ye ZC, Sontheimer H (1996) Cytokine modulation of glial glutamate uptake: a possible involvement of nitric oxide. Neuroreport 7:2181–2185. Young MT, Pelegrin P, Surprenant A (2007) Amino acid residues in the P2X7 receptor that mediate differential sensitivity to ATP and BzATP. Mol Pharmacol 71:92–100.

- 187 -

Yrondi A, Sporer M, Péran P, Schmitt L, Arbus C, Sauvaget A (2018) Electroconvulsive therapy, depression, the immune system and inflammation: A systematic review. Brain Stimulat 11:29–51. Zhan Y, Paolicelli RC, Sforazzini F, Weinhard L, Bolasco G, Pagani F, Vyssotski AL, Bifone A, Gozzi A, Ragozzino D, Gross CT (2014) Deficient neuron-microglia signaling results in impaired functional brain connectivity and social behavior. Nat Neurosci 17:400–406. Zhang B, Zou J, Han L, Beeler B, Friedman JL, Griffin E, Piao Y-S, Rensing NR, Wong M (2018) The specificity and role of microglia in epileptogenesis in mouse models of tuberous sclerosis complex. Epilepsia 59:1796–1806. Zhang B, Zou J, Han L, Rensing N, Wong M (2016) Microglial activation during epileptogenesis in a mouse model of tuberous sclerosis complex. Epilepsia 57:1317– 1325. Zhang J, Zhang K, Gao Z-G, Paoletta S, Zhang D, Han GW, Li T, Ma L, Zhang W, Müller CE, Yang H, Jiang H, Cherezov V, Katritch V, Jacobson KA, Stevens RC, Wu B, Zhao Q (2014) Agonist-bound structure of the human P2Y12 receptor. Nature 509:119–122. Zhao X, Liao Y, Morgan S, Mathur R, Feustel P, Mazurkiewicz J, Qian J, Chang J, Mathern GW, Adamo MA, Ritaccio AL, Gruenthal M, Zhu X, Huang Y (2018) Noninflammatory Changes of Microglia Are Sufficient to Cause Epilepsy. Cell Rep 22:2080–2093. Zou J, Zhang B, Gutmann DH, Wong M (2017) Postnatal reduction of tuberous sclerosis complex 1 expression in astrocytes and neurons causes seizures in an age-dependent manner. Epilepsia 58:2053–2063.

- 188 -