Synaptic Mechanisms Underlying Treatment of Depression and Bipolar Disorder Approved by Supervisory Committee

SYNAPTIC MECHANISMS UNDERLYING TREATMENT OF DEPRESSION AND BIPOLAR DISORDER

APPROVED BY SUPERVISORY COMMITTEE

______Lisa Monteggia, Ph.D.

______Mark Goldberg, M.D.

______Ege Kavalali, Ph.D.

______Adrian Rothenfluh, Ph.D.

Dedicated to my parents, Heinrich and Jo Ann, my brothers Chris and Adam and their families, my grandfather Linwood, my dog Lina, my boyfriend Jerry, and the rest of my family and friends for their unending love and support.

SYNAPTIC MECHANISMS UNDERLYING TREATMENT OF DEPRESSION AND BIPOLAR DISORDER

ERINN SOMMER GIDEONS

DISSERTATION

Presented to the Faculty of the Graduate School of Biomedical Sciences

The University of Texas Southwestern Medical Center at Dallas

In Partial Fulfillment of the Requirements

For the Degree of

DOCTOR OF PHILOSOPHY

The University of Texas Southwestern Medical Center

Dallas, Texas

August 2016

iii

Erinn Sommer Gideons, 2016

ACKNOWLEDGEMENTS

I would first like to thank my advisor Dr. Lisa Monteggia for her continual support, guidance, and patience during my time at UT-Southwestern. She has taught me how to be a successful scientist on many levels and is a true role model of a successful woman in science.

I am extremely grateful for her mentorship and guidance. I would also like to thank Dr. Ege

Kavalali not only for his experimental advice, but also for the many conversations about topics not dealing with science. I would not be an electrophysiologist without his influence. I want to think the current and past members of the Monteggia and Kavalali labs, especially

Dr. Megumi Adachi, Dr. Anita Autry, and Pei-Yi Lin for taking me under their wings and teaching me about science and their continued friendship. In addition, I want to thank Dr.

Melissa Mahgoub for being my rock in and out of lab. I would not have made it through the many steps and missteps of my Ph.D. without her unending support. I also want to thank my closest classmates Austin Reese and Aroon Karra for their friendship, conversations, and scientific insights as we navigated graduate school together.

I would like to thank the members of my thesis committee, Dr. Adrian Rothenfluh,

Dr. Mark Goldberg, and Dr. Ege Kavalali for their feedback and advice over the years.

Many special thanks to Dr. Stuart Ravnik for his insight and advice on many topics, and encouragement of my time with STARS. Finally, I want to thank the University of Texas

Southwestern Medical Center at Dallas, the Graduate School of Biomedical Sciences, and the entire Departments of Psychiatry and Neuroscience for providing top notch training, and the

National Institute of Neurological Diseases and Stroke for funding and monetary support.

Also thank you to UTSW WISMAC for travel support.

I must also thank my gym family, including the members and coaches of Tiger’s Den

Crossfit, Crossfit Bovine, and Heat Barbell Club, especially JD Thorne, Colin and Rebecca

Bilodeau, Meghan Courtney, Dr. Blake Wu, Melissa McIntyre, and Calren Moore. I would never have survived the past four years without the love and friendship I have received from being a part of this extended family.

In the end, I would not be where I am and who I am without the love and support of my family. Thank you to my parents Heinrich and Jo Ann for always believing in me and pushing me to succeed. Also, thank you to both of my brothers, Chris and Adam, I would not be the resilient person I am without both of your influences. Thank you also to Valli and

Kerri for being the sisters that I always wanted. Thank you to all of the Carter Girls, especially the original Carter Girl my grandma Louise Garland Carter, for showing me how to be a strong woman in all aspects of life. I must also thank my Opi, Linwood Gideons, for his love and gracious character. I also want to thank my dog Lina for always putting a smile on my face at the end of a long day. Finally, I want to thank my boyfriend, Jerry Rodriguez, for your unending love, patience, and support during this stressful time.

vii

SYNAPTIC MECHANISMS UNDERLYING TREATMENT OF DEPRESSION AND BIPOLAR DISORDER

Erinn S. Gideons, Ph.D.

The University of Texas Southwestern Medical Center at Dallas, 2016

Supervising Professor: Lisa Monteggia, Ph.D.

Ketamine is a N-methyl-D-aspartate receptor (NMDAR) antagonist that elicits rapid antidepressant responses in depressed patients. However, ketamine can also produce psychotomimetic effects, which limits its widespread use. The field has been exploring the mechanism of ketamine’s antidepressant action to assist in identifying drugs that may also produce the rapid effects without the potential side effects. Memantine is a NMDAR antagonist similar to ketamine in many regards but does not produce antidepressant effects in patients. Behavioral experiments in mice recapitulated clinical findings showing that ketamine but not memantine has antidepressant-like effects in two common antidepressant

viii efficacy tests. Ketamine and memantine effectively blocked NMDAR-mediated mEPSCs in the absence of Mg2+. However, in physiological levels of extracellular Mg2+ only ketamine was able to block the NMDAR at rest. This difference between ketamine and memantine extended to intracellular signaling coupled to NMDAR at rest, in that ketamine inhibits the phosphorylation of eukaryotic elongation factor 2 (eEF2) resulting in an augmentation of subsequent protein expression of brain-derived neurotrophic factor (BDNF), that is not triggered by memantine These results demonstrate significant differences between the efficacies of ketamine and memantine on NMDA receptor mediated neurotransmission that impacts downstream intracellular signaling which is hypothesized as the trigger for rapid antidepressant responses.

In a subsequent study, the therapeutic effect of lithium, a mood stabilizer and a common treatment for Bipolar Disorder (BD) is being investigated. Lithium has antidepressant and antimanic effects in patients with BD that can be recapitulated in animal models. While lithium is effective as a mood stabilizer, the mechanisms that underlie its therapeutic effect are unclear. Lithium has previously been shown to decrease the overall phosphorylation status of eEF2, which increases BDNF protein translation at the synapse. In this study, clinically effective doses of lithium were shown to result in antidepressant and antimanic-like effects in mice. We report that neither eEF2 kinase nor BDNF are necessary for the antidepressant effects of lithium. Additionally, eEF2 kinase is not required for the antimanic effects of lithium. However, BDNF appears to be necessary for lithium’s antimanic actions. To begin to understand whether this requirement of BDNF in lithium’s anti-manic action was due to synaptic changes we performed electrophysiological analysis on

ix primary hippocampal neurons treated with lithium. Chronic lithium treatment caused a significant decrease in AMPAR-mEPSC amplitude, which requires both BDNF and its high affinity receptor, TrkB. Additionally, chronic lithium treatment caused a significant decrease in surface expression of the GluA1 subunit of the AMPAR. Collectively, this data demonstrates that BDNF is required for the antimanic effects of lithium, and that lithium’s effects on the regulation of AMPARs are BDNF and TrkB dependent, which may underlie its behavioral effect.

TABLE OF CONTENTS

Dedication ...... ii Acknowledgements ...... v Abstract ...... viii Table of Contents ...... xi Prior Publications ...... xii List of Figures ...... xiii List of Abbreviations ...... xiv

Chapter 1: Introduction ...... 1 Treatments for MDD and BD ...... 4 Molecular mechanisms underlying Ketamine’s rapid antidepressant effects ...... 7 Molecular mechanisms of lithium still under debate ...... 11 Concluding Remarks ...... 16

Chapter 2: Mechanisms Underlying the Differential Effectiveness of Memantine and Ketamine in Rapid Antidepressant Responses ...... 19 Introduction ...... 19 Materials and Methods ...... 21 Results ...... 25 Discussion ...... 30

Chapter 3: Essential Role of BDNF/TrkB Dependent AMPAR Downscaling in Lithium’s Antimanic Effect ...... 40 Introduction ...... 40 Materials and Methods ...... 43 Results ...... 49 Discussion ...... 54

Chapter 4: Conclusions and Future Directions ...... 65

References ...... 74

Vitae ...... 89

PRIOR PUBLICATIONS

Gideons E.S., Kavalali E.T., Monteggia L.M. (2016) Essential Role of BDNF/TrkB dependent AMPAR downscaling in lithium’s antimanic effect (Manuscript submitted for publication).

Gideons E.S. and Monteggia L.M. (2016) Interplay between eEF2/eEF2-K and mTOR in regulating synaptic protein synthesis (Invited review for The Neuroscientist).

Gideons E.S., Kavalali E.T., and Monteggia L.M. (2014) Mechanisms underlying differential effectiveness of memantine and ketamine in rapid antidepressant responses. Proceedings of the National Academy of Sciences of the United States, 111(23): 8649-54.

Monteggia L.M., Gideons E.S., Kavalali E.T. (2013) The role of eukaryotic elongation factor 2 kinase in rapid antidepressant action of ketamine. Biological Psychiatry, 73(12): 1199-1203.

Brown, R.T and Gideons, E.S. (2006) Etiology of developmental disorders: good science, bad science, and pseudoscience. International Journal of Anthropology, 21(1), 75-82.

xii

LIST OF FIGURES

Figure 1-1. Synaptic mechanism underling ketamine’s rapid antidepressant action ...... 18

Figure 2-1. Memantine (Mem) treatment does not cause a fast-acting antidepressant effect ...... 35

Figure 2-2. AP5, ketamine and memantine block NMDAR-mEPSCs in the absence of Mg2 ...... 36

Figure 2-3. Memantine does not block the NMDAR component of mEPSCs when physiological concentrations of Mg2+ are present ...... 37

Figure 2-4. MK-801 blocks the NMDAR with and without Mg2+ present during recording ...... 38

Figure 2-5. Differential effects of ketamine and memantine on eEF2 phosphorylation and BDNF protein expression at three different time points following treatment ...... 39

Figure 3-1. Chronic lithium treatment causes antidepressant-like effect in mice and increases BDNF expression ...... 59

Figure 3-2. eEF2-K is not required for the antidepressant-like or antimanic-like effects of lithium ...... 60

Figure 3-3. BDNF expression is required for the antimanic-like effect of lithium ...... 61

Figure 3-4. Chronic lithium treatment causes a significant decrease in AMPAR mEPSC amplitude and surface GluA1 expression...... 62

Figure 3-5. BDNF expression is required for the lithium-mediated decrease in AMPAR mEPSC amplitude ...... 63

Figure 3-6. TrkB expression is required for the lithium-mediated decrease in AMPAR mEPSC amplitude ...... 64

xiii

LIST OF ABBREVIATIONS

ACC – anterior cingulate cortex

Akt – protein kinase B

AMPAR – alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor

AMPH - amphetamine

ANOVA – Analysis of variance

AP5 – (2R)-amino-5-phosphonovaleric acid

ARAC – cytosine arabinofuranoside

BD – Bipolar Disorder

BDNF – brain-derived neurotrophic factor

BSA – bovine serum albumin

2+ Ca – calcium

CaM – calmodulin

CamKII – calcium/calmodulin protein kinase II cAMP – cyclic adenosine monophosphate

CA1 – cornus ammonis 1, region I of hippocampus proper

CREB – cAMP response element-binding protein

DAT – dopamine transporter

DIV – days in vitro eEF2 – eukaryotic elongation factor 2 eEF2-K – eukaryotic elongation factor 2-kinase

xiv

FST – forced swim test

GAPDH – Glyceraldehyde 3-phosphate dehydrogenase

GDNF – glial cell-line derived neurotrophic factor

GFP – green fluorescent protein

GluA- AMPAR subunit

GluN – NMDAR subunit

GRIA2 – glutamate receptor, ionotropic, AMPA 2

GSK3β – glycogen synthase kinase 3 beta

ICV - intracerebroventricular

IP – intraperitoneal iPSC – induced pluripotent stem cells

Ket – ketamine

KCl – potassium chloride

KO – knock out

LiCl – lithium chloride

MAOI – monoamine oxidase inhibitor

MAPK – mitogen activated protein kinase

MAP1B – microtubule-associated protein 1B

MDD – Major Depressive Disorder

Mem – memantine mEPSC – miniature excitatory post synaptic current

Mg2+ – magnesium

xv mGluR – metabotropic glutamate receptor

MK-801 – Dizocilpine mRNA – messenger RNA mTOR- mammalian target of rapamycin mV – millivolt

NaCl – sodium chloride

NBQX – 2,3-dihydroxy-6-nitro-7-sulfamoyl-benzo[f]quinoxaline-2,3-dione

NET – norepinephrine transporter

NGF – nerve growth factor

NMDAR – N-methyl d-aspartate receptor

NSF – novelty suppressed feeding test

NT 3 – neurotrophin 3

NT4/5 – neurotrophin 4/5 pA – pico-amperes

PFC – prefrontal cortex

PI3K – phosphoinositide 3-kinase

PLC γ – phospholipase C gamma

PKA – cAMP-dependent protein kinase

PTX – picrotoxin

QPCR – quantitative polymerase chain reaction

RIPA – radio immunoprecipitation assay

RNA – ribonucleic acid

xvi

SEM – standard error of the mean

Ser – serine

SERT – serotonin transporter

SNRI – selective norepinephrine reuptake inhibitor

SSRI – selective serotonin reuptake inhibitors

TCA – tricyclic antidepressant

Thr – threonine

TrkB – tropomyosin receptor kinase B tRNA – transfer RNA

TTX- tetrodotoxin

VGlut1 – vesicular glutamate transporter 1 gene

WT – wild type

xvii

CHAPTER 1 INTRODUCTION

Psychiatric disease affects one in three people at some point in their lifetime

(Silberberg et al., 2015). Major depressive disorder (MDD) is the most commonly diagnosed psychiatric disease and number one form of disability in the United States (Greenberg et al.,

2015). MDD has a 12 month prevalence of 6.7% in the US adult population; women are almost twice as likely to be diagnosed with MDD as men (Kessler et al., 2005b).

Additionally more people are affected by milder forms of depression, such as persistent depressive disorder and seasonal affective disorder (Kessler et al., 2005a). MDD is characterized by thoughts of worthlessness, inappropriate guilt, anxiety, anhedonia, feelings of despair, and changes in sleep patterns and diet (American Psychiatric Association, 2013).

In addition, depression affects the person’s family, work and school performance, and their general health (Hays et al., 1995). The causes of depression are not well understood, although many hypotheses exist, including the monoamine hypothesis, which suggests that depression is caused by alternations in the monoamine neurotransmitters. The monoamine hypothesis was proposed following observations from the 1950s that imipramine and iproniazid, drugs which increase monoamine neurotransmitters, mediate an antidepressant effect while resperine, a drug that decreases monoamine neurotransmitters, is associated with depressive behavior (Schildkraut, 1965). Treatments for MDD range from mental health counseling to antidepressant medications, which will be covered in more detail below.

Bipolar Disorder (BD) is the sixth leading cause of disability worldwide (Schmitt and

Falkai, 2014) and has a life time prevalence from 1-2.4% in the United States (Merikangas et

al., 2007). BD has a similar incidence in men and women (Farren et al., 2012) and is usually diagnosed in late adolescence and early adulthood (Merikangas et al., 2007). The defining criterion of BD is presence of a manic episode that ranges from mild to extremely severe

(Barnett and Smoller, 2009). Manic episodes are characterized by a long period of feeling high, extreme irritability, racing thoughts, impulsivity, restlessness, distractibility, and decreased sleep (American Psychiatric Association, 2013). Generally, a depressive episode occurs subsequent to manic periods (Muzina et al., 2007). Manic episodes can last three to six months and depressive episodes can last for over six months if the patient is left untreated; however the majority of the time patients tend to be in a euthymic state (Muzina et al., 2007). Although the causes of BD are unclear, genetic influence is estimated to account for approximately 60-80% of the risk of developing BD (Kerner, 2014). BD patients are typically treated with mood stabilizers, anticonvulsants, antidepressants, and/or antipsychotic medications. Psychotherapy is also an effective treatment for some aspects of the disease.

Animal models of depression and behavioral tests that are predictive of antidepressant efficacy have been developed to investigate the mechanisms underlying the disorder and its treatment. To study the molecular mechanisms related to depression, a depressive-like state can be induced in rodents by subjecting them to chronic forms of stress that last days to weeks. The depressive-like behaviors can be ameliorated following sub-chronic and chronic dosing with monoaminergic based therapies and acute dosing of ketamine (Nestler and

Hyman, 2010; Krishnan and Nestler, 2011). However, it is not clear how well these depression-like models truly replicate the human condition or the underlying biology of the depression. Nevertheless, treatment of rodents with antidepressants is used to study

2 molecular and cellular mechanisms of antidepressants. Two common tests that are predictive of antidepressant effects are the Forced Swim Test (FST) and Novelty Suppressed Feeding

(NSF), which measure behavioral despair and anxiety, respectively. In the FST, a mouse is placed in a beaker of water and allowed to swim for 6 minutes. The amount of time in the last 4 minutes of the test the mouse spends not actively trying to escape is recorded as immobility time. Antidepressant drugs decrease immobility time in comparison to vehicle treatment (Porsolt et al., 1977; Monteggia et al., 2004). In the NSF, animals are food deprived for 24 hrs and then placed in an open field chamber containing a single pellet of chow. The experimenter measures the amount of it takes the mouse to eat the piece of food, with lower latency suggesting an antidepressant-like effect (Bodnoff et al., 1988).

The majority of preclinical models for BD focus on recreating either the depressive or manic aspects of the disorder as modeling the cyclical nature of BD has proven difficult in animals (Cosgrove et al., 2016). Mania is commonly modeled in rodents using pharmacological methods. Acute and chronic administration of psychostimulants increases locomotor activity of the animal that is blunted by concomitant treatment with lithium and other mood stabilizers (Antelman et al., 1998; Gould and Einat, 2007; Sharma et al., 2016).

Multiple studies have found that chronic lithium treatment blocks the hyperlocomotor effects of a single dose of amphetamine (Gould et al., 2007; Flaisher-Grinberg and Einat, 2010).

Another pharmacological method used to induce mania in rodents in the intracerebroventricular (i.c.v.) administration of ouabain, a sodium-potassium ATPase inhibitor. Similar to stimulants, ouabain causes increased locomotor activity and risk taking behavior in rodents that can be ameliorated by chronic administration of lithium (Li et al.,

1997). While these models allow for research on the mechanisms of antimanic drugs, there are very few rodent models that exhibit both manic and depressive behaviors highlighting the inherent difficulty in developing animal models of complex human diseases.

Treatments for MDD and BD

Classic antidepressant drugs target the monoaminergic neurotransmitter system.

Antidepressant drugs encompass tricyclic antidepressants (TCA), monoamine oxidase inhibitors (MAOI), selective serotonin reuptake inhibitors (SSRI), and norepinephrine reuptake inhibitors (SNRI). Monoamine based antidepressants increase the extracellular amount of monoamine neurotransmitters, serotonin and norepinephrine, by blocking either the serotonin transporter (SERT) or the norepinephrine transporter (NET) or both (Blier et al., 1987; Zemlan and Garver, 1990; Hamon and Blier, 2013) or inhibiting the breakdown of monoamines (Shulman et al., 2013). They can also have effects on the dopamine transporter

(DAT) (Hamon and Blier, 2013). Although monoamine based antidepressants produce significant antidepressant effects in some patients with depression (Trivedi et al., 2006), there are caveats associated with these drugs. Clinical trials have repeatedly shown a delayed onset of antidepressant efficacy, typically in the range of several weeks (Trivedi et al., 2006), which is especially concerning for patients exhibiting suicidal tendencies. Additionally, approximately two thirds of patients on monoaminergic based antidepressants never reach full remission of their depression and only approximately 50% of patients report significant relief of their symptoms (Trivedi et al., 2006). For these reasons, there has been great interest in developing faster acting antidepressant treatments that are effective for a larger percentage

4 of the patient population. In 2000, Berman et al. reported that the N-methyl-d-aspartate receptor (NMDAR) open channel antagonist, ketamine, caused an antidepressant effect in treatment resistant MDD patients within 30 min-1hr that was sustained for multiple days

(Berman et al., 2000). This discovery opened a completely new line of research into the molecular basis of rapid antidepressant efficacy. Having a better understanding of the molecular mechanisms underlying ketamine’s antidepressant effects may provide information that is key to advancing new treatments for depression.

There are many different types of pharmacological treatments for BD, which can be effective in treating the symptoms of mania or depression or both. Lithium was first reported as a treatment for mania in the late 1940s (Cade, 1949) and was subsequently shown to have antidepressant effects in bipolar patients (Gershon, 1972). Additionally, it is commonly used as adjunctive therapy with traditional antidepressants in MDD patients and has well documented anti-suicidal effects in MDD and BD patients (Coppen et al., 1991; Muller-

Oerlinghausen et al., 1992; Linden et al., 1994; Muller-Oerlinghausen et al., 1994). Lithium is the most common mood stabilizer, but other classes of drugs have gained popularity due to the side effect profile associated with lithium use. Anticonvulsants, such as valproic acid, lamotrigine, and gabapentin, were originally developed to treat seizures but are now also used as mood stabilizers (Selle et al., 2014). Patients experiencing extreme manic episodes may sometimes be treated with atypical antipsychotics, but these drugs are usually prescribed in combination with antidepressant medications (Selle et al., 2014). Alternatively, antidepressants are given to BD patients while in depressive episodes, though antidepressant use alone can increase the risk of rebound mania (Koszewska and Rybakowski, 2009). The

5 molecular mechanisms underlying the effectiveness of these different drugs as mood stabilizers in BD are still under debate.

One common pathway that is linked to both depression and BD and their treatments is brain-derived neurotrophic factor (BDNF) and its high affinity receptor, tropomyosin receptor kinase B (TrkB). BDNF release, activation of TrkB, and signaling pathways downstream of TrkB are critical for synaptic plasticity (Leal et al., 2014), which is thought to be necessary for antidepressant responses (Martinowich et al., 2007). Activation of the TrkB receptor, as measured by autophosphorylation of tyrosine 674/5, can then activate three downstream signaling pathways, phosphoinositol 3-kinase (PI3K), phospholipase C γ (PLC

γ), and mitogen-activated protein kinase [MAPK or extracellular signal related kinase

(ERK)]. Antidepressant and antimanic medications have been shown to increase BDNF protein expression in both patients in blood serum and brain tissue from pre-clinical rodent models (Autry and Monteggia, 2012). BDNF mRNA is also increased following chronic treatment with antidepressants and the mood stabilizer lithium (Autry and Monteggia, 2012;

Tunca et al., 2014). Moreover, direct infusion of BDNF into the midbrain or the hippocampus is able to induce an antidepressant-like within days in rodents (Siuciak et al.,

1997; Shirayama et al., 2002). Previous work has shown that BDNF expression is required for the antidepressant-like effects of both traditional antidepressants and the fast-acting antidepressant ketamine (Monteggia et al., 2004; Adachi et al., 2008; Autry et al., 2011).

TrkB expression and activation are also required for the antidepressant-like effects of monoaminergic antidepressants and ketamine in preclinical rodent models (Saarelainen et al.,

2003; Autry et al., 2011). Clinical research has shown that BDNF protein expression is

6 significantly decreased in blood serum and cerebrospinal fluid (CSF) of MDD and BD patients (Fernandes et al., 2011; de Azevedo Cardoso et al., 2014). Additionally, pre-clinical rodent models of both diseases have decreased BDNF expression in the hippocampus and prefrontal cortex (PFC) (Duman and Monteggia, 2006; Fries et al., 2015). However, the requirement for BDNF or TrkB in the behavioral effects of lithium and other mood stabilizers is at this point unknown.

Molecular mechanisms underlying Ketamine’s rapid antidepressant effects

The discovery of the fast-acting and long-lasting antidepressant effects of ketamine opened up many new areas of research on elucidating the mechanisms underlying antidepressant efficacy. Trullas and Skolnick first reported that multiple NMDAR antagonists produced antidepressant-like effects in mice and rats (Trullas and Skolnick,

1990). A decade later, Berman et al. found that a single low-dose injection of the NMDAR antagonist ketamine caused rapid antidepressant effects in treatment-resistant MDD patients

(Berman et al., 2000). This initial report has since been replicated in multiple clinical trials in MDD patients (Zarate et al., 2006a; Liebrenz et al., 2007; DeWilde et al., 2015).

Ketamine is also effective in treating depressive episodes in BD patients (Zarate et al., 2012).

However, ketamine can cause psychotomimetic and anesthetic effects at higher doses and also potential abuse liability, which could limit widespread use as an antidepressant. As ketamine does not directly target the monoaminergic neurotransmitter system, there has been considerable research investigating the mechanisms underlying ketamine’s rapid

7 antidepressant effects (Li et al., 2010; Autry et al., 2011; Monteggia et al., 2013; Kavalali and Monteggia, 2015).

In molecular studies, ketamine increased protein translation of BDNF and other mRNA transcripts. The increase in protein translation was shown to occur through desuppression of eukaryotic elongation factor 2 (eEF2) as evidenced by a decrease in the ratio of phosphorylated eEF2 to total eEF2 (Autry et al., 2011). eEF2 is tonically phosphorylated at threonine 56 (Thr56) by eEF2 kinase (eEF2K), which represses the activity of eEF2 (Ryazanov et al., 1988; Ryazanov et al., 1991). When eEF2K is inhibited, unphosphorylated eEF2 can bind to the ribosome and catalyze the elongation step of protein translation (Ryazanov et al., 1988). In a series of studies, Sutton and Schuman found that eEF2K activity and dendritic protein translation are linked to spontaneous release of glutamatergic vesicles and subsequent calcium influx through the NMDAR (Sutton et al.,

2004; Sutton et al., 2006; Sutton et al., 2007). Under baseline conditions, calcium influx through the NMDAR activates eEF2K, which suppresses protein translation. However, antagonizing the NMDAR when neuronal activity is inhibited by the sodium channel blocker tetrodotoxin (TTX) mediates a decrease in eEF2 phosphorylation and an increase in protein translation specifically at the synapse within minutes (Sutton et al., 2007). Importantly, decreasing the phosphorylation of eEF2 with pharmacological inhibition of eEF2K by rottlerin or NH125 also increased BDNF protein expression and caused an antidepressant- like effect in the FST and NSF (Autry et al., 2011). More recent studies using constitutive eEF2K knock out (KO) mice demonstrated that eEF2K is required for the rapid antidepressant effects of ketamine (Nosyreva et al., 2013). Taken together, these results

8 indicate that protein translation mediated by the eEF2K pathway is necessary for the antidepressant effects of ketamine.

The rapid antidepressant effect of ketamine relies on blockade of the NMDAR at rest and subsequent increase of protein translation and potentiation of evoked alpha-amino-3- hydroxy-5-methyl-4-isoxazolepropionic acid receptor (AMPAR) responses. In behavioral studies, ketamine’s rapid antidepressant-like effects were blocked by pretreatment with either the protein translation inhibitor anisomycin or the AMPAR antagonist 2,3-dihydroxy-6-nitro-

7-sulfamoyl-benzo[f]quinoxaline-2,3-dione (NBQX) (Li et al., 2010; Autry et al., 2011).

Using acute hippocampal slices from wildtype mice, it was found that a 30 min ketamine incubation caused potentiation of AMPAR mediated evoked neurotransmission that lasted for at least an hour (Autry et al., 2011; Nosyreva et al., 2013). This potentiation was blocked by both anisomycin and NBQX, again indicating that protein translation and AMPAR activity are required for ketamine’s effects. In addition, potentiation of AMPAR responses following ketamine treatment was not seen in hippocampal slices from conditional BDNF KO mice or constitutive eEF2K KO mice (Nosyreva et al., 2013). Acute ketamine treatment of hippocampal slices was also shown to increase the surface expression of AMPAR subunits 1 and 2 (GluA1, GluA2), which was lost in hippocampal slices from BDNF KO and eEF2K

KO mice (Nosyreva et al., 2013). These results show that eEF2K mediated protein translation of BDNF is required for AMPAR potentiation following ketamine treatment.

Based on these findings, I would like to present the following model for the molecular mechanisms underlying the rapid antidepressant effects of ketamine. Blockade of the NMDAR by ketamine at rest decreases Ca2+ flow through the channel pore, which

9 reduces the activity of eEF2K. As eEF2K activity decreases, the amount of unphosphorylated eEF2 at the synapse increases, allowing for eEF2 to bind to the ribosome catalyzing protein translation elongation of multiple mRNA transcripts, such as BDNF

(Figure 1-1). Concomitantly, homeostatic mechanisms mediate increased surface expression

GluA1 and GluA2 mediates a potentiation of evoked AMPAR responses. Taken together, the increase in dendritic protein translation and AMPAR potentiation following NMDAR blockade point to homeostatic plasticity as the cellular mechanism for the rapid antidepressant effect of ketamine.

The discovery of ketamine’s fast-acting antidepressant effects caused much excitement in clinical and pre-clinical research, however, ketamine has side effects that limit its widespread use. At high doses, ketamine is routinely used as an effective anesthetic in humans, especially children and the elderly, as well as in veterinary medicine (Carter and

Story, 2013; Peltoniemi et al., 2016). At subanesthetic doses, ketamine causes psychotomimetic effects, such as hallucinations and difficulty concentrating, and short-term memory impairment and is often abused as “Special K” (Krystal et al., 1994). Therefore, many clinicians and pre-clinical researchers have initiated research into other compounds that affect NMDAR functioning without causing anesthesia and psychotomimetic effects associated with ketamine. One NMDAR antagonist that was initially tested in MDD patients was memantine, as it is a drug approved by the Food and Drug Administration (FDA) for the treatment of mild to moderate Alzheimer’s disease (Fleischhacker et al., 1986; Wilkinson,

2001). However, memantine does not produce an antidepressant effect after acute or chronic treatment as demonstrated in three separate placebo controlled clinical trials (Zarate et al.,

2006b; Ferguson and Shingleton, 2007; Lenze et al., 2012). These results were especially surprising since memantine binds in the pore of the NMDAR channel similar to ketamine

(Kotermanski et al., 2009). Therefore, an open question remains as to why two drugs that target the NMDAR in similar manners do not both produce antidepressant effects.

Molecular Mechanisms of Lithium Still Under Debate

Historically, lithium was recognized for its ability to treat manic and depressive episodes as early as the 1870s, (Lenox and Watson, 1994; Mitchell and Hadzi-Pavlovic,

2000; Shorter, 2009), with some anecdotal reports of lithium springs as “crazy waters” dating even further back. However, prescribing lithium to treat manic or depressed patients was not widespread until the mid-1900s and was not approved by the FDA for use in treating acute mania in the U.S. until 1970 (Fieve, 1970). Lithium is effective in treating both acute manic episodes and as a prophylactic therapy for long-term management of both depressive and manic episodes (Schou et al., 1954; Baastrup and Schou, 1967; Cundall et al., 1972).

Patients are typically given an initial low dose and then the dose is increased with monitoring of lithium concentration in the blood serum (Grandjean and Aubry, 2009b). Lithium efficacy is known to be dose-dependent and shows strong correlation with blood serum concentration

(Grandjean and Aubry, 2009b). The clinically effective lithium dosage range for maintenance treatment in BD patients is 0.5-2.5 mmol/L (Amdisen, 1980). Although lithium is the first-line choice for maintenance of BD, there are multiple documented side effects associated with lithium treatment including: polydipsia, polyuria, myoclonus,

11 hypothyroidism, and renal toxicity (Grandjean and Aubry, 2009a), therefore the lowest effective dose of lithium is optimal for long-term maintenance treatment.

The mood stabilizing effects of lithium are well documented in both clinical and pre- clinical literature (Malhi et al., 2013), however the molecular and cellular mechanisms underlying its antimanic and antidepressant effects are still under debate. Many mechanisms have been proposed, yet there are only a few molecular effects that have a strong consensus

(Du et al., 2004a; Wu et al., 2014; Beurel et al., 2015). Lithium’s inhibition of glycogen synthase kinase 3 (GSK3) was first reported from in vitro studies and since has been replicated in vivo many times over (Klein and Melton, 1996; Stambolic et al., 1996; Jope,

2011). There are two isoforms of GSK3, alpha (α) and beta (β), that are both inhibited by lithium, although the β isoform has been shown to be more important for lithium’s mood stabilizing effects. Lithium inhibits GSK3 directly by competing for Mg2+ binding, which is required for kinase activity (Klein and Melton, 1996). Lithium inhibits GSK3 indirectly by increasing activity of upstream kinases, such as Akt, that phosphorylate serine 9 (Ser9) or

Ser21 on GSK3β or GSK3α, respectively, which decreases the kinase activity of GSK3

(Stambolic et al., 1996). Many studies have shown that specific GSK3 inhibitors also mediate antidepressant and antimanic behaviors (Jope, 1999; Manji and Lenox, 2000; Phiel and Klein, 2001), suggesting that inhibition of GSK3 is an important mediator of lithium’s therapeutic effects. Furthermore, genetic manipulation to increase GSK3 activity induces manic-like and depressive-like behaviors (Prickaerts et al., 2006; Polter et al., 2010), while decreasing the expression of GSK3 causes the reverse (O'Brien et al., 2004; Beaulieu et al.,

2008; Kaidanovich-Beilin et al., 2009). However, GSK3 is known to phosphorylate over 100

12 substrates (Sutherland, 2011) and is involved in various cellular functions, including apoptosis, glucose homeostasis, and insulin signaling among others (Cohen and Frame, 2001;

Kaidanovich-Beilin and Woodgett, 2011). At this time, it is still unknown which of these various substrates and functions are involved in GSK3’s mood regulation. Additionally, increased GSK3 phosphorylation is linked to many other psychotropic drugs, including ketamine (Liu et al., 2013) and amphetamine (Mines and Jope, 2012), and is also seen in the ouabain-induced mania model (Yu et al., 2010), indicating that inhibition of GSK3 is not specific to lithium-mediated mood stabilization.

Another mechanism proposed to underlie lithium’s behavioral effects involves neurotrophin signaling. Lithium treatment in patients, pre-clinical rodent models, and neuronal cultures increases BDNF expression and activation of its high affinity receptor

TrkB. Lithium treatment increases BDNF protein expression in the blood serum of patients

(Tramontina et al., 2009; de Sousa et al., 2011), in the hippocampus and frontal cortex of both mice and rats (Fukumoto et al., 2001; Jornada et al., 2010), and in hippocampal and cortical neuronal cultures (Dwivedi and Zhang, 2014; Park et al., 2015). Additionally, chronic lithium treatment elevates BDNF mRNA expression in patients, pre-clinical models, and cultured neurons (Fukumoto et al., 2001; Omata et al., 2008; Dwivedi and Zhang, 2014;

Tunca et al., 2014). Importantly, lithium treatment also increases autophosphorylation of

TrkB in the anterior cingulate cortex (ACC) and in cortical culture, suggesting that lithium activates the TrkB receptor and downstream signaling cascades (Hashimoto et al., 2002;

Rantamaki et al., 2006). Intriguingly, Hashimoto et al. found that lithium protected cortical neurons from glutamate induced excitotoxicity in a BDNF and TrkB activity dependent

13 manner, implying that BDNF could be required for other actions of lithium (Hashimoto et al.,

2002). TrkB transduces the BDNF signal by activation of ERK1/2 and PI3-K/Akt signaling pathways, both of which are known to regulate cell survival (Wada, 2009). There is evidence that lithium increases expression of neurotrophin-3 (NT-3) (Walz et al., 2008), NT-4/5 (Walz et al., 2009), nerve growth factor (NGF) (Angelucci et al., 2003), and glial cell line-derived neurotrophic factor (GDNF) (Angelucci et al., 2003) suggesting that lithium activates multiple neuroprotective pathways in addition to BDNF/TrkB, although the role of these growth factors in lithium action have not been well studied.

Lithium has been shown to activate multiple pathways that result in increased BDNF mRNA and protein expression. One possible pathway is through inhibition of GSK3β, which has been shown to increase BDNF expression through cyclic adenosine monophosphate

(cAMP) response element-binding protein (CREB)-mediated transcription (Boer et al.,

2007). However, increased BDNF expression and activation of TrkB stimulates Akt, which phosphorylates and inhibits GSK3 (Johnson-Farley et al., 2006; Smillie et al., 2013).

Therefore it is currently unknown which of these actions of lithium is primary and which is secondary. Additionally, lithium treatment has been shown to increase cAMP leading to activation of cAMP-dependent protein kinase (PKA), which phosphorylates and activates

CREB (Lorenzi et al., 2013). Lithium also activates the ERK/MAP kinase and ERK1/2 cascades, both of which lead to CREB activation and increased BDNF transcription (Yan et al., 2007). Recently it was shown that lithium decreases the ratio of phosphorylated eEF2 to total eEF2 in a human neuronal cell line and in the hippocampus of mice treated with lithium,

14 similar to ketamine’s effect on eEF2 and suggestive of another pathway leading to increased

BDNF protein expression (Karyo et al., 2010).

The glutamatergic system has been implicated in both the pathophysiology of BD and lithium treatment. Studies have found elevated glutamine/glutamate ratios in the brains of

BD patients in manic episodes suggestive of increased glutamatergic activity (Michael et al.,

2003; Ongur et al., 2008). Post-mortem samples from the ACC of BD patients had a significant increase in the expression of the vesicular glutamate transporter 1 (VGluT1) which is associated with increased glutamate release (Eastwood and Harrison, 2010).

Additionally, chronic lithium treatment is associated with significant downregulation of the

GRIA2, the gene encoding the GluA2 subunit of the AMPAR, in a human neuronal cell line

(Seelan et al., 2008). Pre-clinical studies have also shown direct links between mania, lithium, and the AMPAR. Chronic treatment of rats and mice with lithium at therapeutic concentrations results in decreased membrane expression of GluA1 and GluA2 in the hippocampus (Du et al., 2003; Gray et al., 2003; Du et al., 2004c; Du et al., 2004b; Du et al.,

2008; Du et al., 2010). Conversely, a single injection of amphetamine, which is routinely used to model mania, increases GluA1 surface expression in the hippocampus of rats (Mao et al., 2015). Intriguingly, increased GluA2 surface expression was found in the hippocampus of rats treated with the antidepressant imipramine, which is known to cause rebound mania in

BD patients (Gray et al., 2003; Du et al., 2008). In electrophysiological experiments, the reduction in surface expression of AMPAR subunits results in decreased AMPAR-mediated signaling. In cultured neurons, acute and chronic lithium incubation causes a significant decrease in AMPAR miniature excitatory postsynaptic current (mEPSC) amplitude while not

15 affecting AMPAR mEPSC frequency, indicative of a postsynaptic effect of lithium (Wei et al., 2010; Ankolekar and Sikdar, 2015). Acute hippocampal slices taken from rats following chronic lithium treatment show a reduced AMPA/NMDA ratio in CA1 pyramidal neurons suggesting decreased AMPAR signaling (Du et al., 2008). Additionally, it was recently shown that chronically increased BDNF expression is associated with decreased surface expression of the AMPAR, suggesting a possible link between BDNF/TrkB signaling and

AMPAR endocytosis (Reimers et al., 2014). Taken together, these studies point to a disease relevant mechanism of lithium treatment on AMPARs in that increased glutamatergic signaling during manic episodes is ameliorated with lithium through down regulation of

AMPAR receptors at the synapse possibly involving BDNF.

Concluding Remarks

The molecular mechanisms underlying the efficacy of pharmacological treatments for

MDD and BD are not well understood. However, there is a growing body of evidence that

BDNF/TrkB signaling is intimately involved in the effects of the rapid antidepressant ketamine and the mood stabilizer lithium. Additionally, activation of the eEF2/eEF2K pathway that is known to increase BDNF expression, is required for ketamine’s antidepressant effects and has recently been implicated in the molecular effects of lithium.

Ketamine’s rapid antidepressant effects in treatment-resistant patients and in pre- clinical rodent models caused much excitement in the field of psychiatry. However, the known side effect profile of ketamine prevents its widespread use for the treatment of depression. It was widely hypothesized that memantine, another open channel NMDAR

16 blocker, would be effective as an antidepressant, yet multiple clinical trials have shown no antidepressant effects of memantine. Therefore, I seek to elucidate potential differences between ketamine and memantine.

Lithium’s mood stabilizing effects are well documented, yet the molecular and cellular mechanisms mediating the antidepressant and antimanic actions are not well understood. BDNF is intricately involved in both BD and lithium treatment, but it is unknown whether expression of BDNF is required for lithium’s behavioral actions.

Additionally, as outlined above, AMPAR signaling is also affected in opposing manners by

BD and lithium. However, it is unknown whether lithium’s effects on AMPAR signaling are linked to its therapeutic profile. Thus, I want to determine the requirement for BDNF/TrkB in lithium’s antidepressant and antimanic effects and lithium-mediated decrease in AMPAR signaling.

Collectively, these studies aim to provide a better understanding of the molecular actions of psychiatric medications and a framework for development of new pharmacological treatments.

Figure 1-1. Synaptic mechanism underling ketamine’s rapid antidepressant action. (Top) Under resting conditions, spontaneous glutamate release and NMDAR activation leads to stimulation of eEF2-kinase triggering eEF2 phosphorylation and silencing of BDNF translation at the synapse. (Bottom) Ketamine-mediated use-dependent blockade of tonic NMDAR activity at rest decreases activation of eEF2-kinase resulting in a gradual decrease of eEF2 phosphorylation and desuppression of BDNF translation. (Figure taken from Monteggia, Gideons, Kavalali 2013).

CHAPTER 2

MECHANISMS UNDERLYING THE DIFFERENTIAL EFFECTIVENESS OF

MEMANTINE AND KETAMINE IN RAPID ANTIDEPRESSANT RESPONSES

Introduction

Ketamine is a noncompetitive glutamate N-methyl-D-aspartate receptor (NMDAR; also called GluN) antagonist that has been shown to mediate rapid antidepressant efficacy in patients with treatment resistant major depression (Berman et al., 2000; Zarate et al., 2006a;

Price et al., 2009). The antidepressant effects of ketamine are fast acting with some patients reporting effects as soon as 30 min to within a few hours following a single i.v. low-dose injection of ketamine. However, ketamine can produce adverse psychotomimetic effects, which may limit its use as an antidepressant. Traditional antidepressant drugs target the monoamine system and typically require several weeks of treatment in order to mediate a therapeutic effect. There is an urgent need for rapid antidepressant drugs and the clinical data with ketamine suggests that blocking the NMDAR may be a viable therapeutic target.

Memantine is a noncompetitive NMDAR antagonist that has been approved by the

U.S. Food and Drug Administration for the treatment of Alzheimer’s disease. Memantine is a well-tolerated drug that at therapeutic doses lacks the psychotomimetic effects (Parsons et al., 1999) observed with ketamine. However, attempts to test memantine as an antidepressant in individuals with major depression have yielded mixed results following long-term drug treatment with no evidence of rapid antidepressant effects (Zarate et al.,

2006b; Ferguson and Shingleton, 2007; Lenze et al., 2012) A better understanding of why

19 ketamine, but not memantine, produces a fast-acting antidepressant response has clinical implications and may provide novel information critical for the development of rapid antidepressant therapeutics based on NMDA receptor antagonism with fewer side effects.

There is much interest in identifying the molecular mechanism that underlies the rapid antidepressant response of ketamine. In recent work we demonstrated that the fast- acting antidepressant effect of ketamine requires deactivation of eukaryotic elongation factor

2 kinase (eEF2K) and subsequent desuppression of brain derived neurotrophic factor

(BDNF) protein translation in the hippocampus (Autry et al., 2011; Nosyreva et al., 2013).

We hypothesize that low-dose ketamine mediates its rapid antidepressant response by blockade of spontaneous glutamate release mediated NMDAR activity. This in turn decreases calcium (Ca2+) flow through the receptor, inhibiting eEF2K activity and resulting in decreased levels of phosphorylated eukaryotic elongation factor 2 (eEF2) (Sutton et al.,

2004; Sutton et al., 2006; Sutton et al., 2007) and desuppression of BDNF protein synthesis

(Autry et al., 2011; Kavalali and Monteggia, 2012; Nosyreva et al., 2013). In this study, we compared ketamine with memantine in their effectiveness to block NMDAR activation during spontaneous neurotransmission and subsequently inhibiting eEF2K and increasing

BDNF protein translation. Our results reveal key differences between the effects of ketamine and memantine on resting NMDAR mediated neurotransmission and subsequent intracellular signaling pathways that may explain the mechanistic differences between these two drugs in eliciting rapid antidepressant effects.

Materials and Methods

Mice and Drug Treatments. Male C57BL/6 mice aged 6-8 weeks old were habituated to the animal facility for one week prior to testing. Mice were kept on a 12/12 light-dark cycle and allowed ad libitum access to food and water, except where indicated. Mice were injected intraperitoneal (i.p.) to more closely mimic the route of administration in humans. Mice were injected with drug 30 min, 8 hrs, or 24 hrs prior to testing or sacrifice to assess behavior and molecular events at the time of initial antidepressant responses with the exception of the studies examining locomotor activity in which mice were injected and immediately placed in the boxes to assess drug effects with time. Memantine hydrochloride (3.0, 10, or 20 mg/kg)

(Sigma) and ketamine (2-3 mg/kg) (Fort Dodge Animal Health) were dissolved in saline.

Experiments were conducted by an observer blind to drug treatment. All procedures were approved by the Institutional Animal Care and Use Committee at UT Southwestern Medical

Center.

Locomotor Activity. Following a one-hour habituation period in the testing room, mice were injected with saline, ketamine, or memantine then placed under red light into standard cages and locomotor activity was measured for 60 minutes by photocell beams linked to computer acquisition software (San Diego Instruments).

Forced Swim Test (FST). The FST was performed according to published protocols (Autry et al., 2011). Briefly, mice were video-recorded in a 4 L glass beaker containing 3 L of 24 ±

1°C water for 6 minutes. The last 5 minutes of each 6-minute trial were scored by a blinded observer to determine the time spent immobile.

Novelty Suppressed Feeding (NSF). The NSF was performed according to published protocols (Autry et al., 2011). Mice were food deprived for 24 hours prior to the test and then habituated to the behavioral room for 1 hour prior to testing. A mouse was placed into a 42 x

42 cm open field with a food pellet placed in the center, and allowed to explore for up to 5 minutes with the time to initiate eating the food pellet determined. To assess differences in appetite, the amount of food consumed in a 5-minute period for each mouse in their home cage was measured.

Cell Culture. Dissociated hippocampal cultures were prepared as previously described

(Autry et al., 2011). Whole hippocampi were dissected from postnatal day 0-3 (P0-P3)

C57BL/6 mice, then trypsinized (~10 mg/ml trypsin, Invitrogen), mechanically dissociated, and plated on Matrigel coated coverslips. At 1 day in vitro (DIV), 4 M cytosine arabinoside

(4 ARAC, Sigma) was added; at 4 DIV the ARAC concentration was reduced to 2 M. All experiments were done on cultures that were 14-21 DIV.

Electrophysiology. Whole-cell patch-clamp recordings were performed on hippocampal pyramidal neurons. Data were acquired using a MultiClamp 700B amplifier and Clampex

10.0 software (Molecular Devices). Recordings were sampled at 100 s and filtered at 2 kHz with a gain of 5. The external Tyrode solution contained (in mM): 150 NaCl, 4 KCl, 2 CaCl2,

10 glucose, 10 HEPES, pH 7.4, ~300 mOsm. External MgCl2 concentration was either 0 mM or 1.25 mM depending on the experiment. The pipette internal solution contained (in mM):

110 K-gluconate, 20 KCl, 10 NaCl, 10 HEPES, 0.6 EGTA, 4 Mg-ATP, 0.3 Na-GTP, 10

22 lidocaine N-ethyl bromide (QX-314), pH 7.3, ~300 mOsm. Pipettes had resistances between

3-6 MΩ. The junction potential between the external solution and internal solution was ~12 mV and was subtracted from all recordings. NMDAR mediated miniature excitatory postsynaptic potentials (mEPSCs) were recorded in the presence 1 M tetrodotoxin (TTX,

EMD Millipore), 50 M picrotoxin (PTX, Sigma), and 10 M 2,3-dihydroxy-6-nitro-7- sulfamoyl-benzo[f]quinoxaline-2,3-dione (NBQX, Sigma) for 4 min before and 4 min after the application of 50 M AP5, 50 M ketamine, 50 M memantine or 50 M MK-801.

Dual-component mEPSCS were recorded at -67 mV in the presence 1 M TTX and 50 M

PTX for 4 min before and 4 min after the application of 50 M AP5, 50 M ketamine, 50

M memantine, or 50 M MK-801.

Data Analysis. To isolate mEPSCs recorded in the absence or presence of the NMDAR antagonists, events were selected using a template search in pClamp 10.0. The experimenter was blind to drug condition for time shift analysis and averaging of mEPSCs. For comparison of mEPSCs in each cell before and after drug application, 100 mEPSCs under steady state conditions (~ 2 min into the recording or after solution exchange) were averaged.

Charge transfer calculations were performed for before and after comparisons for NMDAR- mEPSCs on the entire 4 minute recording.

Protein Quantification. Anterior hippocampal slices (2-3/mouse, ~1 mm thick) were dissected and flash frozen from mice 30 min following drug injection. Hippocampal tissue was lysed in a buffer containing phosphatase and protease inhibitors (Roche). Protein

23 concentration was quantified with Bradford analysis. Approximately 30 μg of protein were run on SDS-PAGE gels then transferred to polyvinylidene fluoride (PVDF) membranes activated in methanol. Primary antibodies were used at the following dilutions: BDNF

(Abcam, ab108319) 1:1000, GAPDH (Cell Signaling, 2118s) 1:50000, phosphorylated eEF2

(Thr56) and total eEF2 (Cell Signaling, 2331s and 2332, respectively) 1:2000. Total and phospho-eEF2 primary antibody dilutions included 5% BSA. After washing, the membranes were incubated with anti-rabbit secondary antibodies: BDNF 1:5000, GAPDH 1:10000, phospho-eEF2 and total eEF2 1:5000. Protein bands were detected using ECL and exposed to film. Following development of phospho-eEF2 bands, membranes were stripped in buffer

(25 mM glycine, 1% SDS, pH 2) before blocking and reprobing with total eEF2 primary antibody. The films were analyzed using ImageJ (NIH). Phospho-eEF2 and total eEF2 intensity were measured as a ratio normalized to GAPDH. BDNF protein was normalized to

GAPDH.

Statistical Analysis. Data are reported as mean ± SEM. Statistical differences were assessed using unpaired or paired 2-tailed Student’s t-test, one-way ANOVA, or two-way ANOVA when appropriate. Tukey and Bonferroni post hoc tests were used when appropriate.

Statistical significance was defined as p < 0.05.

Results

Acute memantine treatment does not trigger a fast-acting antidepressant response

We assessed whether memantine affects locomotor activity immediately following drug treatment. In all experiments, we included a ketamine group as a direct comparison, which has previously been shown to elicit an antidepressant response in mice 30 minutes after administration without effects on locomotor activity at this time point (Mantovani et al.,

2003; Maeng et al., 2008; Autry et al., 2011; Nosyreva et al., 2013). To examine a range of doses for the effects of this drug in vivo, we injected memantine at 20 mg/kg, a dose reported to have neuroprotective effects in rodents (Emnett et al., 2013), 10 mg/kg a dose that blocks morphine dependence (Popik and Skolnick, 1996), and 3 mg/kg, a dose that prevents estrogen-dependent tolerance to morphine. We found that a 3-, 10-, or 20- mg/kg dose of memantine did not have any effect on total locomotor activity in comparison to the saline treated mice during the 60 minute testing period (Fig. 1A, insert). We examined the data in

5-minute epochs and also did not find any significant differences between mice treated with memantine compared to saline (Fig. 1A). Ketamine caused an initial increase in locomotor activity during the first 10 minutes following drug administration, however there were no significant differences in activity during the remaining 50 min of the test, specifically at the

30 min time point when antidepressant responses are measured (Fig. 1A).

We next examined whether memantine produces rapid antidepressant effects in the forced swim test (FST). In agreement with previous data, ketamine significantly reduced immobility time 30 min following injection, suggestive of an antidepressant response (Fig.

1B) (Autry et al., 2011; Nosyreva et al., 2013). In contrast, a single injection of memantine

25 at 3 or 10 mg/kg did not significantly alter immobility time in the FST at 30 minutes (Fig.

1B). Using separate cohorts of mice, we found that ketamine caused a significant decrease in immobility time at 8 or 24 h following acute administration in the FST that was not observed with either 3 or 10 mg/kg of memantine (Fig. 1B). Rather unexpectedly, mice treated with

20 mg/kg memantine had severe adverse effects in the FST, including general motor instability and drowning; thus, to avoid potential complications and undue stress to the animals, we did not use this dose in vivo in subsequent experiments.

A different cohort of C57BL/6 mice was used to assess whether memantine triggers a rapid antidepressant effect in the novelty suppressed feeding (NSF) test. In agreement with previous data, a single low dose injection of ketamine administered 30 min prior to testing significantly reduced the time to initiate eating in the mice, suggestive of an antidepressant response (Fig. 1C) (Autry et al., 2011; Nosyreva et al., 2013). However, a single injection of memantine at either 3 or 10 mg/kg did not reduce the time for the mice to initiate feeding compared to saline. We also found that neither ketamine nor memantine treatment had any significant effect on the mice’s appetite ruling out a possible confound to the NSF test (Fig.

1D).

Memantine exhibits reduced NMDAR block in physiological magnesium

We examined ketamine, memantine, and the commonly used NMDAR antagonist R-2- amino-5-phosphonopentanoate (AP5) in their ability to block NMDA-mediated miniature excitatory post-synaptic currents (mEPSCs) in cultured hippocampal neurons. To record

NMDA-mEPSCs, the extracellular recording solution did not include the endogenous

NMDAR pore blocker, magnesium (Mg2+); however the α-amino-3-hydroxy-5-methyl-4- isoxazolepropionic acid receptor (AMPAR) antagonist 2,3-dihydroxy-6-nitro-7-sulfamoyl- benzo[f]quinoxaline-2,3-dione (NBQX) was added. To measure the total decrease in charge transfer conferred by the NMDAR antagonists, baseline NMDAR-mEPSCs were recorded for 4 min. Each of the individual NMDAR antagonists, AP5 (Fig, 2A), ketamine (Fig, 2C), or memantine (Fig. 2E), was then perfused into the bath, and recordings were continued for an additional 4 min. Analysis before and after drug application by an observer blind to the treatment revealed that perfusion of AP5, ketamine, and memantine resulted in a significant and similar reduction in charge transfer of NMDAR-mEPSCs (Fig. 2 B, D, F). This finding is in agreement with recent results demonstrating equal efficacy of ketamine and memantine in blockade of NMDAR mediated responses (Emnett et al., 2013).

Next, we evaluated whether ketamine and memantine show similar efficacy in blockade of spontaneous glutamate release mediated NMDAR responses under physiological levels of Mg2+ blockade. Previous work from our group has shown that acute treatment with

AP5 under physiological conditions, 1.25 mM Mg2+ and -67 mV holding potential, causes a significant decrease in decay time and charge transfer of mEPSCs (Espinosa and Kavalali,

2009) suggesting a significant contribution of NMDARs to glutamatergic neurotransmission at rest. Using whole-cell patch-clamp methods in the presence of 1.25 mM Mg2+, dual- component mEPSCs were recorded from dissociated hippocampal neurons before and subsequent to treatment with AP5 (Fig. 3A), ketamine (Fig. 3F), or memantine (Fig. 3K). As previously reported, AP5 perfusion caused a significant reduction in the area and decay time of mEPSCs (Fig, 3 C and D) (Espinosa and Kavalali, 2009). We found that ketamine

27 treatment also significantly decreased mEPSC area and decay time, indicative of blocking the

NMDAR component of the mEPSC (Fig. 3 H and I). In contrast, memantine treatment caused no significant change in mEPSC area or decay time under physiological conditions

(Fig. 3 M and N). We also found that none of the NMDAR antagonists examined caused a change in the average amplitude of mEPSCs (Fig. 3 E, J, and O).

To test our model that blockade of the NMDAR at rest mediates a fast-acting antidepressant response, we assessed the ability of another noncompetitive NMDAR antagonist, Dizocilpine (MK-801), to block NMDAR-mEPSCs in the absence of Mg2+ and the NMDAR component of mEPSCs in the presence of physiological concentrations of

Mg2+. We chose MK-801 because we had previously demonstrated that acute treatment with

MK-801 causes a significant decrease in immobility time in the FST 30 min after drug administration (Autry et al., 2011). Similar to AP5, ketamine, and memantine, perfusion of

MK-801 caused a significant decrease in charge transfer of NMDAR-mEPSCs (Fig. 4A and

B), indicating that it is able to block the NMDAR at rest when no Mg2+ is included in the recording solution. MK-801 treatment also caused a significant decrease in decay time of mEPSCs, a strong trend towards a decrease in the area of mEPSCs in the presence of 1.25 mM Mg2+, and no change in mEPSC amplitude which closely mimics the effects of AP5 and ketamine (Fig. 4 C- F).

Ketamine and memantine have differing intracellular signaling effects

The fast-acting antidepressant effect of ketamine is dependent on protein translation (Autry et al., 2011). The increase in protein translation following ketamine is hypothesized to be

28 mediated through blockade of NMDARs at rest, which inhibits eEF2K, resulting in decreased phosphorylation of eEF2 followed by desuppression of BDNF protein translation. We examined whether memantine treatment affects eEF2 phosphorylation and BDNF expression in the hippocampus by Western blot analysis. In agreement with previous data, ketamine treatment triggered a significant decrease in phosphorylation of eEF2 (Fig. 5 A and C) and a significant increase in BDNF protein levels compared to vehicle 30 min following injection

(Fig. 5 B and D) (Autry et al., 2011; Nosyreva et al., 2013). In contrast, memantine did not alter the phosphorylation level of eEF2 or total eEF2 (Fig. 5 A and C) and did not have any significant effect on BDNF protein levels (Fig. 5 B and D).

We previously demonstrated that ketamine-mediated effects on eEF2 phosphorylation and BDNF protein abundance are transient and disappear by 24 h post-injection (Autry

2011). However, to determine whether memantine may mediate effects on eEF2 phosphorylation and BDNF levels at later time points, we examined these protein levels 8 or

24 h after acute injection. As with previous data, ketamine treatment did not cause any significant changes in eEF2 phosphorylation at 8 h (Fig. 5 E and G) or 24 h (Fig. 5 I and K).

Additionally, there was no change in BDNF protein 8 h (Fig. 5 F and H) or 24 h (Fig. 5 J and

L) following ketamine injection. Similarly, memantine treatment did not cause any changes in eEF2 phosphorylation or BDNF protein levels 8 h (Fig. 5 E- H) or 24 h (Fig. 5 I- L) after drug administration.

Discussion

In this study, we used behavioral, electrophysiological, and biochemical approaches to compare the actions of ketamine and memantine on antidepressant-like effects in behavioral models, spontaneous NMDAR-mEPSCs and downstream signaling in the hippocampus to work out a mechanistic explanation for why ketamine, but not memantine, is able to exert rapid antidepressant actions. In this way, we recapitulated the clinical findings of ketamine and memantine in mice, showing that ketamine, but not memantine, has antidepressant-like effects in behavioral models. Electrophysiological analysis revealed that ketamine and memantine antagonize the NMDAR at rest when Mg2+ is absent. However, only ketamine blocks the NMDAR at rest when physiological concentrations of Mg2+ are included in the external solution, providing a key functional difference between ketamine and memantine in their ability to block NMDAR function at rest. The differential functional effects of ketamine and memantine on NMDAR-mEPSCs extend to intracellular signaling coupled to

NMDARs at rest. We found that memantine does not inhibit the phosphorylation of eEF2 or augment subsequent BDNF protein expression, which are critical determinants of ketamine- mediated antidepressant efficacy. Collectively, these results identify key functional differences between ketamine and memantine in their ability to suppress NMDAR function at rest, and thus inhibit the eEF2 kinase signaling pathway, providing insight into the mechanistic basis for NMDA receptor antagonism and rapid antidepressant action.

Recent clinical findings demonstrating that ketamine shows rapid antidepressant effects in patients with major depression (Berman et al., 2000; Zarate et al., 2006a; Price et

30 al., 2009) has triggered a great deal of interest in the field of depression research. However, even low-dose ketamine used in the depression studies causes psychotomimetic effects in some patients, with the potential for abuse (Salvadore and Singh, 2013). To circumvent these potential liabilities associated with ketamine, there has been interest in investigating whether memantine possesses the antidepressant properties of ketamine. However, in two recent clinical trials chronic memantine did not elicit an antidepressant response in depressed patients compared to patients given placebo (Zarate et al., 2006b; Lenze et al., 2012). In a separate open-label trial, depressed patients treated with chronic memantine did show some clinical improvement in ratings of their depression symptoms 1 week after starting daily dosing and continuing for at least 12 wk; however this study did not contain a placebo control group (Ferguson and Shingleton, 2007). Ketamine has faster pharmacokinetics following in vivo administration than memantine, and is likely to reach peak concentration in brain much faster than memantine. In addition, in vitro studies suggest that ketamine has slightly higher potency than memantine. However, given that the clinical studies utilized chronic memantine treatment yet memantine lacked antidepressant properties, it seems unlikely that the clinical differences between these two drugs are due to differences in pharmacokinetics and/or affinity. The clinical findings demonstrating differences between ketamine and memantine in triggering rapid antidepressant responses are rather surprising as both drugs are noncompetitive NMDAR antagonists that block the receptor when it is in an open configuration (Kotermanski et al., 2009; Emnett et al., 2013). Importantly, the two compounds do not show significant differences in their ability to block NMDAR-mediated synaptic or extrasynaptic currents in the absence of physiological Mg2+ (Emnett et al., 2013).

Although our findings agree with this earlier work, we could detect a biologically significant difference in their differential ability to block the NMDAR component of mEPSCs in the presence of physiological levels of Mg2+. Therefore, we propose that the disparity in rapid antidepressant responses between these compounds is likely due to differences in how they affect NMDAR function at rest under physiological conditions.

Previous work has shown that ketamine and memantine block the NMDAR by binding inside of the ion channel in an area overlapping the binding site for Mg2+, which blocks the channel in a voltage-dependent manner (Kashiwagi et al., 2002). In the absence of

Mg2+, ketamine is more potent than memantine in the blockade of the two most highly expressed GluN2 subunits in the hippocampus, GluN2A and GluN2B, as evidenced by lower

IC50 values (GluN1/GluN2A: memantine ~0.80 μM vs ketamine ~0.33 μM; GluN1/GluN2B: memantine ~0.57 μM vs ketamine ~0.31 μM) (Kotermanski and Johnson, 2009), which may indicate a higher affinity for the receptor. Under more physiological conditions in the

2+ presence of Mg , ketamine (IC50: GluN1/GluN2A ~5.35 μM; GluN1/GluN2B ~5.08 μM) is still more potent than memantine (IC50: GluN1/Glu2A ~13.4 μM; GluN1/GluN2B ~ 10.4

μM)(Kotermanski and Johnson, 2009). However, the difference seen between the ability of memantine and ketamine to block GluN2A and GluN2B containing NMDA receptors is not extended to GluN2C and GluN2D containing NMDAR without Mg2+ present

(GluN1/GluN2C: memantine ~0.52 μM vs ketamine ~0.51 μM; GluN1/GluN2D: memantine

~0.54 μM vs ketamine ~0.83 μM) (Kotermanski and Johnson, 2009). There also appears to be no significant difference in blockade of GluN2C and GluN2D containing NMDARs by memantine (GluN1/GluN2C: ~1.61 μM; GluN1/GluN2D: ~1.76 μM) or ketamine

(GluN1/GluN2C: ~1.18 μM; GluN1/GluN2D: ~2.95 μM) when Mg2+ is included in the external solution (Kotermanski and Johnson, 2009). Additionally, it seems unlikely that any differences found between ketamine and memantine would be mediated by GluN3 containing

NMDARs as they are not sensitive to block by Mg2+ or use dependent blockers (Chatterton et al., 2002) (Smothers and Woodward, 2007). Because our experiments specifically assess the ability of memantine to block NMDAR-mediated transmission in the presence of Mg2+ in the hippocampus, the two-fold difference between the abilities of ketamine and memantine to block NMDA receptor mediated transmission in the presence of Mg2+ could partly explain our observations. Recent electrophysiological studies have shown that blockade of the

NMDAR by Mg2+ is incomplete, even at physiological resting potentials, allowing for a significant amount of current to flow through the NMDAR (Espinosa and Kavalali, 2009;

Povysheva and Johnson, 2012; Nosyreva et al., 2013). These results are in agreement with classical studies that assessed the degree of NMDAR blockade by Mg2+ at near-resting membrane potentials (Collingridge et al., 1988b, a; Jahr and Stevens, 1990b, a) . Therefore, our current findings demonstrating that ketamine, but not memantine, blocks the NMDAR- mediated component of mEPSCs when Mg2+ is present may provide a functional explanation for how these two compounds have differing effects as fast acting antidepressants.

The importance of blockade of NMDAR-mEPSCs as a key determinant in the rapid antidepressant action of ketamine extends to intracellular signaling coupled to NMDAR at rest. Blockade of NMDAR-mEPSCs has been linked to the control of dendritic protein translation by decreasing calcium influx through the NMDAR into the synapse, leading to decreased phosphorylation of eEF2 by eEF2 kinase and desuppression of protein translation

(Sutton et al., 2004; Sutton et al., 2006), thus providing a potential mechanism for ketamine’s rapid antidepressant effects. The rapid antidepressant effects of ketamine have also been suggested to be mediated by mammalian target of rapamycin (mTOR)-dependnt synapse formation, although it remains unclear how blockade of the NMDAR activates mTOR (Li

Science 2010). BDNF is a potent activator of mTOR; thus, the blockade of NMDAR- mEPSCs and inactivation of eEF2K, leading to decreased phosphorylation of eEF2, and ultimately, up-regulation of BDNF protein expression may well explain the mTOR findings

(Monteggia, Gideons, Kavalali 2013). In this study, we found that memantine does not inhibit the phosphorylation of eEF2 or augment subsequent expression of BDNF, necessary requirements for ketamine mediated antidepressant efficacy (Autry et al., 2011; Kavalali and

Monteggia, 2012; Nosyreva et al., 2013). The differential actions of ketamine and memantine on NMDAR function at rest, coupled with differential effects on downstream intracellular signaling pathways coupled to these receptors, further corroborate the functional requirements of NMDAR mediated neurotransmission at rest as a necessary determinant of rapid antidepressant responses.

In the present study, our data strengthen and extend our previous findings that decreased eEF2 phosphorylation triggered by ketamine-mediated blockade of NMDAR- mEPSCs is critical for the rapid antidepressant effect (Autry et al., 2011; Kavalali and

Monteggia, 2012; Nosyreva et al., 2013). These findings provide a mechanistic explanation for why ketamine, but not memantine, is able to exert rapid antidepressant actions, which provides important information for the development of more effective antidepressants based on NMDAR antagonism with fewer side effects.

Figure 2-1. Memantine (Mem) treatment does not cause a fast-acting antidepressant effect. (A) Ketamine causes a significant increase in locomotor activity at the 5 and 10 min intervals [(Two-way ANOVA interaction F44,420 = 1.904, p=0.0007, time F11,420 = 25.87, p < 0.0001 and treatment effects F4,420 = 2.784, p = 0.0264; Tukey’s post hoc analysis for ket: 5 min vs saline *P =.01; 10 min vs saline **P =.001, n=8/group)]. (Inset) There were no significant differences in total number of beam breaks over 1 h. (B) Ket treatment caused a significant decrease in immobility in the FST compared with the saline control group at 30 min, 8 h, and 24 h following injection, whereas Mem (3 mg/kg, 10 mg/kg) did not cause a significant decrease in immobility time [Two-way ANOVA F3,107=28.96, P <0.0001, Tukey’s post hoc analysis 30 min: saline vs Ket ***P=0.0002, saline vs Mem 3 mg/kg P=0.501, saline vs Mem 10 mg/kg p=0.972 n=9-10/group; 8 h: saline vs Ket****P<0.0001, saline vs Mem 3 mg/kg p=0.991, saline vs Mem 10 mg/kg P=0.904; 24 hrs: saline vs Ket **P=0.0017, saline vs Mem 3 mg/kg P=0.584, saline vs Mem 10 mg/kg P=0.996). (C) Single dose of Mem treatment (30 min, 3mg/kg, 10 mg/kg) did not impact latency to feed in the NSF test compared to the saline treated group. However, acute Ket treatment (3 mg/kg) resulted in a significant decrease in latency to feed compared to the saline-treated group [ANOVA F3,35=4.97,* P =0.0056, Bonferroni’s post hoc comparison saline vs Mem 3mg/kg P >0.9999, saline vs Mem10 mg/kg P >0.999 saline vs Ket P=.0465, n=9-10/group]. (D) Appetite posttest following the NSF test indicates that the total amount of food consumed is indistinguishable across groups (ANOVA F3,35=1.999, P=0.1321).

Figure 2-2. AP5, ketamine and memantine block NMDAR-mEPSCs in the absence of Mg2+. (Left) Example traces recorded before and after incubation of the neurons with AP5 (A), ketamine (C), or memantine (E). (Right) Quantification of the charge transfer in a 4-min period before and after applying AP5 (B), ketamine (D), or memantine (F). The application of AP5, ketamine and memantine caused a significant decrease in charge transfer (Student’s paired t test, AP5: *P =0.019, n=6 coverslips, ketamine: *P=0.025, n=6 coverslips, memantine: *P=0.043, n=7 coverslips).

Figure 2-3. Memantine does not block the NMDAR component of mEPSCs when physiological concentrations of Mg2+ are present. (Left) Representative traces before and after applying AP5 (A), ketamine (F), or memantine (K); AMPAR mEPSCs are still present following NMDAR antagonist incubation. (Center) Average traces of 100 mEPSCs before (black trace) and after (red trace) perfusion of AP5 (B), ketamine (G), or memantine (L), with the calculated difference trace (gray). (C, H, and M) Both AP5 and ketamine caused a significant decrease in mEPSC area, whereas no significant differences were measured with memantine (Student’s paired t test, AP5: *P=0.031, n=7 coverslips, ketamine: *P=0.039, n=8 coverslips, memantine p=0.357, n=15 coverslips). (D, I, and N) Application of AP5 (D) and ketamine (I) caused a significant decrease in mEPSC decay time compared to control groups, with no changes detected with memantine (N) (Student’s paired t test, AP5: *P=0.013, n=7 coverslips, ketamine: **P=0.003, n=8 coverslips, memantine: P=0.695, n=15 coverslips). (E, J, and O) mEPSC amplitude was not affected by AP5 (E), ketamine (J), or memantine (O) application (Student’s paired t test, AP5: P=0.244, n=7 coverslips, ketamine: P=0.926, n=8 coverslips, memantine: P=0.153, n=15 coverslips).

Figure 2-4. MK-801 blocks the NMDAR with and without Mg2+ present during recording. (A) Example traces recorded before and after perfusion with MK-801. (B) Application of MK-801 caused a significant decrease in charge transfer (Student’s paired t test **P=0.004, n=6 coverslips). (C) Example traces of mEPSCs recorded in 1.25 mM Mg2+ before and after incubation with MK-801. (D) Average traces of 100 mEPSCs recorded before (black trace) and after (red trace) perfusion with MK-801 with the calculated difference trace (gray trace). (E- G). Perfusion of MK-801 causes an almost significant decrease in mEPSC area (Student’s paired t test p=0.07), a significant decrease in mEPSC decay time (Student’s paired t test **P=0.003) and no change in mEPSC amplitude (Student’s paired t test p=0.427) (n=14 coverslips).

Figure 2-5. Differential effects of ketamine and memantine on eEF2 phosphorylation and BDNF protein expression at three different time points following treatment. (A, C) Densitometric analysis of phosphorylated eEF2 (P-eEF2) levels revealed ketamine caused a significant decrease in the ratio of phosphorylated eEF2/total eEF2 (T-eEF2) at 30 min, whereas memantine (3 mg/kg and 10mg/kg) did not impact phosphorylation of eEF2. [ANOVA, F3,44= 3.579, *P=0.021, Tukey’s post hoc analysis saline vs ketamine *P=0.018, saline vs memantine 3 mg/kg P=0.569, saline vs memantine 10 mg/kg P=0.930 n= 10-13 per group)]. (B, D) Densitometric analysis of BDNF levels revealed memantine did not have a significant effect on BDNF protein expression 30 min after injection; however ketamine treatment caused a significant increase in BDNF protein [ANOVA, F3,49=5.893 , P=0.0016, Tukey’s post hoc analysis saline vs ketamine *P=0.02, saline vs memantine 3 mg/kg P=0.999, saline vs memantine 10 mg/kg P=0.745 (n= 12-15 per group)]. (E and G) Neither ketamine nor memantine caused a significant change in the levels of P-eEF2 8 h after injection as shown by densiometric analysis (ANOVA, F3,21=0.0828, P=0.969). (F and H) There was no change in the amount of BDNF protein in the hippocampus 8 h following injection with ketamine or memantine (ANOVA, F3,24=1.006, P=0.407). (I and K) Densiometric analysis of p-eEF2 levels showed no difference in P-eEF2/T-eEF2 between saline and ketamine or memantine treatment 24 h after injection (ANOVA, F3,16=0.731, P=0.552). (J and L) Densiometric analysis of BDNF protein showed no difference between saline, ketamine, or memantine treatment 24 h following injection (ANOVA, F3,25=0.206, P=0.891).

CHAPTER 3

ESSENTIAL ROLE OF BDNF-TRKB DEPENDENT AMPAR DOWNSCALING IN

LITHIUM’S ANTIMANIC EFFECT

Introduction

Lithium was initially described as a mood stabilizer over 60 years ago (Cade, 1949) and was subsequently shown to have antidepressant effects (Gershon, 1972). Lithium is widely prescribed for the treatment of Bipolar Disorder (BD) (Poolsup et al., 2000; Mitchell, 2013) and is also used as adjunctive therapy for unipolar depression; however the cellular and molecular mechanisms that underlie its effectiveness as both an antidepressant and antimanic agent are still under debate. Many molecular effects of lithium have been described , including increased brain-derived neurotrophic factor (BDNF) expression (Wu et al., 2014) and inhibition of glycogen synthase kinase-3β (GSK3β) (Klein and Melton, 1996). Additionally lithium has been shown to decrease α-amino-3-hydroxyl-5-methyl-4-isoxazolepropionic acid receptor (AMPAR) miniature excitatory post-synaptic current (mEPSC) amplitude and surface expression of

AMPAR subunits 1 and 2 (GluA1, GluA2) (Gray et al., 2003; Seelan et al., 2008; Du et al.,

2010; Wei et al., 2010; Ankolekar and Sikdar, 2015). However, which of these effects are the primary actions of lithium is currently unknown.

BDNF and its high affinity receptor tropomyosin receptor kinase B (TrkB) are closely associated with both the pathophysiology and treatment of many psychiatric conditions, including BD (Autry and Monteggia, 2012). Both manic and depressive states have been associated with significantly decreased BDNF blood serum levels in BD patients in comparison to patients in euthymic states and healthy controls (Tunca et al., 2014; Fernandes et al., 2015). In rodents, decreased BDNF mRNA and protein are found in the hippocampus following

40 experimental interventions that cause manic-like (Frey et al., 2006a; Jornada et al., 2010; Fries et al., 2015) and depressive-like behaviors (Smith et al., 1995; Tsankova et al., 2006).

Alternatively, lithium usage has been associated with increased BDNF protein levels in the serum of BD patients (Cunha et al., 2006; Tramontina et al., 2009; de Sousa et al., 2011). In rodent models, chronic lithium treatment also increases BDNF mRNA and protein expression in the hippocampus and cortex (Fukumoto et al., 2001; Yasuda et al., 2009; Jornada et al., 2010).

BDNF expression and TrkB activity are increased following treatment with lithium in neuronal culture (Hashimoto et al., 2002). Increased BDNF expression and TrkB activation have been shown to increase the phosphorylation of GSK3β at serine 9 (Xia et al., 2010; Smillie et al.,

2013), which inhibits the activity of GSK3β and is thought to be one of the main molecular mechanisms of lithium (Jope, 2011).

Lithium is also known to decrease glutamatergic signaling, which has been proposed to counteract the increase in glutamate and neuronal hyperexcitability found in BD patients and pre-clinical rodent models (Chuang et al., 2002; Mertens et al., 2015). Lithium treatment has previously been shown to decrease AMPAR signaling. Chronic lithium treatment was associated with decreased expression of GRIA2, the gene encoding GluA2, in a human neuronal cell line

(Seelan et al., 2008). Chronic lithium dosing in rats and chronic lithium treatment of cultured neurons caused a significant decrease in surface expression of two AMPAR subunits, GluA1 and

GluA2, in the hippocampus (Gray et al., 2003; Du et al., 2004b; Du et al., 2008). Other studies have shown both acute and chronic treatment with lithium caused a significant decrease in

AMPAR miniature excitatory post-synaptic current (mEPSC) amplitudes but did not affect mEPSC frequency, indicating a post-synaptic effect of lithium (Wei et al., 2010; Ankolekar and

Sikdar, 2015). Taken together, these studies point to a direct effect of lithium on AMPAR functioning through changes in surface expression.

There is an ongoing debate concerning the primary targets underlying lithium’s effectiveness. Previous work has shown that BDNF/TrkB signaling is necessary for the antidepressant effect of traditional antidepressants (Saarelainen et al., 2003; Monteggia et al.,

2004; Adachi et al., 2008; Ibarguen-Vargas et al., 2009) and the fast-acting antidepressant ketamine (Autry et al., 2011; Lepack et al., 2015). However, it is currently unknown if BDNF or its receptor TrkB are required for lithium’s behavioral effects or lithium’s effects on AMPAR mEPSC. In this study, we test for the necessity of BDNF in lithium’s antidepressant and antimanic effects. We then investigate the requirement of either BDNF or TrkB for the lithium mediated decrease in AMPAR mEPSC amplitude. Our results reveal a requirement for neurotrophic signaling in lithium’s behavioral and cellular effects that start to shed light on the mechanism of action of lithium.

Materials & Methods

Animals

Male C57BL/6 mice aged 6-8 weeks were habituated to the animal facilities for at least 7 days prior to behavior testing. The mice were maintained on a 12/12 light dark cycle with access to food and water ad libitum, unless otherwise noted for lithium treatment groups. The inducible

BDNF knockout (KO) mice were generated as previously described (Monteggia et al., 2004). All behavioral testing was done with age and weight matched mice and groups were balanced by genotype. All analysis was done with an experimenter blind to genotype and treatment condition.

Animal protocols were approved by the Institutional Care and Use Committee at UT

Southwestern Medical Center.

Lithium Treatment in vivo

Mice were given a 0.2% lithium chloride (LiCl) diet (Harlan Teklad) for 4 days and then switched to chow containing 0.4% LiCl (Harlan Teklad) for the remainder of the study, which lasted from 11-17 days total. All mice received water as well as a bottle of 0.9% sodium chloride (NaCl) to control for ion imbalances known to occur with lithium administration.

Control mice were kept on the same chow except it lacked lithium.

Measurement of lithium in serum

Trunk blood was collected from all animals after the completion of the lithium treatment and behavioral testing. Whole blood was kept on ice until it was spun at 3,000 RPM at 4˚C for 10 min to separate red blood cells and serum. Lithium ion counts were made using a flame photometer (Jenway PFP7) and concentration was calculated following the construction of a

43 standard curve. Mice with serum lithium concentrations below 0.5 mM and above 3 mM were excluded from behavioral and biochemical analyses.

Behavior

Forced Swim Test. The forced swim test (FST) was performed in accordance with published protocols (Gideons et al., 2014). In brief, mice were video-recorded in a glass 4-L beaker with 3

L of 23 ± 2 ̊C water for 6 min. The last 5 min of each trial were scored by an observer blinded to drug condition and genotype to determine immobility time.

Locomotor Activity. Mice were habituated to the testing room for 1 hr and then given an intraperitoneal (i.p.) injection of saline or amphetamine hydrochloride (Sigma) at 2 mg/kg dissolved in saline. Mice were then immediately placed in standard home cages under red light and locomotor activity was measured for 2 h by photocell beams linked to computer acquisition software (San Diego Instruments). The total beam counts for the 2 h period were collected as a measurement of amphetamine induced hyperactivity.

Cell culture

Dissociated hippocampal cultures were prepared as previously described (Reese and Kavalali,

2015). Whole hippocampi were dissected from postnatal day 0-3 (P0-P3) C57BL/6, floxed

BDNF (fl.BDNF), or floxed TrkB (fl.TrkB) mice as indicated. The hippocampi were trypsinized

(~10 mg/mL trypsin; Invitrogen), dissociated mechanically, and plated on Matrigel (Corning

Biosciences)-coated coverslips for electrophysiology or directly onto tissue culture treated plates for protein collection for western blot analysis. At 1 d in vitro (DIV), 4 M cytosine arabinoside

(ARAC; Sigma) was added. At 4 DIV, the ARAC concentration was decreased to 2 M with a

44 media change. Treatment with LiCl (1 mM) or NaCl (1 mM ) was initiated at 4 DIV and lasted for 11-17 days. The fl.BDNF and fl.TrkB cultures were infected with lentivirus expressing Cre-

GFP or eGFP control at 4 DIV. All electrophysiology experiments and protein collected for western blot analysis were done on 15-21 DIV cultures.

Electrophysiology

As described previously, whole-cell patch clamp recordings were performed on hippocampal pyramidal neurons (Autry et al., 2011). The external Tyrode’s solution contained (in mM) 150

NaCl, 4 KCl, 2 CaCl2, 1.25 MgCl2, 10 glucose, and 10 Hepes (ph 7.4) at ~300 mOsm. The pipette internal solution contained (in mM): 110 K-gluconate, 20 KCl, 10 NaCl, 10 Hepes, 0.6

EGTA, 4 Mg-ATP, 0.3 Na-GTP, and 10 lidocaine N-ethyl bromide (pH 7.3) at ~300 mOsm.

Pipettes had a resistance between 3-7 M. The junction potential between the internal and external solutions was ~12 mV and was subtracted from all recordings. AMPAR-mediated mEPSCs were recorded in the presence of 50 M (2R)-amino-5-phosphonovaleric acid (AP5;

Tocris), 1 M tetrodotoxin (TTX; EMD Millipore), and 50 M picrotoxin (PTX; Sigma). Data were acquired using a MultiClamp 700B amplifier and Clampex 10.0 software (Molecular

Devices). Recordings were sampled at 100 s, filtered at 2kHz with a gain of 5. No more than 3 recordings were obtained per coverslip. AMPAR-mEPSCs were analyzed from a 3-5 min recording using MiniAnalysis software by an experimenter blind to drug condition and genotype.

Quantitative RT-PCR

Briefly, fresh frozen hippocampi were dissected and total RNA was extracted using Trizol

(Invitrogen) following the manufacturer’s instruction. Conditions for cDNA synthesis,

45 amplification, and primer sequences were described previously (Adachi et al., 2008). The fold change in BDNF expression was normalized to the housekeeping gene, GAPDH.

Protein Quantification

Anterior hippocampal slices (~1 mm thick, 2-3 per mouse) were dissected and flash-frozen following 11 days of lithium treatment or immediately following the last behavioral test depending on the experiment. Hippocampal tissue was lysed in a radio immunoprecipitation assay (RIPA) buffer containing: 50 mM Tris pH 7.4, 1% Igepal, 0.1% SDS, 0.5% Na deoxycholate, 4 mM EDTA, 150 mM NaCl, phosphatase inhibitors (10 mM Na pyrophosphate,

50 mM NaF, 2 mM Na orthovanadate), and protease inhibitors (complete Mini tablets, Roche).

Protein concentration was quantified with the Quick-Start Bradford assay, (Bio-Rad).

Approximately 30 g of protein was electrophoresed on SDS-PAGE gels and then transferred to nitrocellulose membranes. The membranes were incubated with primary antibodies overnight at the following dilutions: BDNF (Abcam), 1:1000, GAPDH (Cell Signaling) 1:50,000, phosphorylated eEF2 (Thr56) and total eEF2 (Cell Signaling) 1:2000. Primary antibody for phospho-eEF2 included 5% BSA. After washing, the membranes were incubated in anti-rabbit secondary antibodies: BDNF, 1:5000, GAPDH, 1:10,000, and phospho-eEF2 and total eEF2,

1:5,000. Protein bands were detected using ECL then exposed to film. Following development of phospho-eEF2 bands, membranes were stripped using Restore PLUS Western Blot Stripping

Buffer (ThermoScientific), put in block, and then in primary antibody for total eEF2 overnight.

BDNF expression was normalized to GAPDH. Phospho-eEF2 and total eEF2 expression were normalized to GAPDH and then plotted as a ratio.

Cell surface AMPAR expression measurement

Membrane biotinylation experiments were performed as previously described (Nosyreva et al.,

2013). Dissociated hippocampal cultures from C57BL/6 mice were incubated in Tyrode’s solution containing 1 mg/ml sulfo-NHS-LC-biotin (Pierce) for 20 min on ice. The biotin reactions were quenched by incubating the cultures in Tris-buffered saline (TBS) with 15 mM ammonium chloride for 5 min on ice, and then washed twice with TBS for 5 min on ice.

Following the second TBS wash, the cultures were lysed in RIPA buffer (as described above) for

10 min on ice and spun at 12000 rpm for 5 min to remove non-solubilized material. Total protein concentration was quantified by Quik-Start Bradford assay (Bio-rad). 100 g of protein from each sample was incubated with 100 L of washed UltraLink NeutrAvidin (Pierce) immobilized beads and rotated overnight at 4 ̊C. Beads were washed with three times with RIPA buffer, followed by three washes with TBS at 4 ̊C. Protein was eluted from the beads with SDS-

PAGE sample buffer supplemented with -mercaptaethanol (BME) for 10 min at 95 ̊C. Eluted surface protein and 20 g of total protein in SDS-PAGE-BME sample buffer were resolved by

10% SDS-PAGE gel, transferred to nitrocellulose, and probed with anti-GluA1 antibody

(1:1000, Chemicon) and anti-GAPDH antibody (1:50,000) overnight. Secondary anti-rabbit antibodies were at 1:2000 and 1:10,000 for GluA1 and GAPDH respectively. Surface GluA1 over total GluA1 ratio is reported.

Statistics

Data are reported as mean ± SEM. Statistical differences in the FST, locomotor tests, western blot, and QPCR, and mEPSC amplitude and frequency were assessed using unpaired two-tailed

Student’s t test or one-way ANOVA when appropriate. Tukey, Bonferroni, and Dunnett post

47 hoc tests were used when appropriate. Differences in the cumulative probability histograms were assessed with Kolmogorov-Smirnov test. Statistical significance was defined at p ≤ 0.05.

Results

Chronic lithium treatment causes antidepressant-like effect and increases BDNF.

The lithium treatment protocol consisted of 0.2% lithium chloride (LiCl) chow for four days, followed by an increase to 0.4% LiCl chow for the rest of treatment and testing period, which lasted 11-17 days depending on the experiment (Fig 3-1A). We determined the lithium serum concentration in treated mice to ensure it was within the therapeutic range of 0.5-2 mM lithium

(Amdisen, 1980). In our experiments, we found an average lithium concentration of 1 mM in the blood serum of lithium treated mice (Fig 3-1B). The forced swim test (FST) is commonly used in preclinical psychiatric research as antidepressant drugs, and mood stabilizers such as lithium, reduce the immobility time (Nestler and Hyman, 2010) allowing for the examination of the possible mechanism of action of these drugs. To show that the lithium treatment was causing the expected behavioral effects, we tested mice in the FST. Mice that received lithium treatment showed a significant reduction in immobility time compared to untreated mice (Fig 3-1C). We then analyzed whether the lithium treatment regulated the levels of phosphorylated eEF2 compared to total eEF2 or the expression of BDNF as previously reported. Mice were sacrificed following 11 days of lithium treatment and the hippocampus was rapidly removed. Using Q-PCR to the coding exon of BDNF, lithium treatment significantly increased BDNF mRNA expression

(Fig 3-1D). Western blot analysis revealed that lithium treatment significantly increased BDNF protein levels in the hippocampus (Fig 3-1E) and decreased the ratio of phosphorylated eEF2 relative to total eEF2 (Fig 3-1F) although this was due to an increase in total eEF2 rather than a decrease in the level of phosphorylated eEF2.

BDNF is required for the antimanic effect of lithium

Since lithium regulated the ratio of phosphorylated eEF2 to total eEF2 and BDNF expression, we investigated whether these molecules were necessary for lithium’s behavioral action. We used constitutive eEF2 kinase null knockout (KO) mice to examine the antidepressant and antimanic effects of lithium. In the FST, we found that lithium treatment significantly reduced the time of immobility in the littermate control (CTL) mice compared to vehicle injected mice as expected

(Fig 3-2A). Similarly, lithium treatment also significantly reduced the immobility time in the eEF2 kinase KO mice compared to vehicle treated eEF2 kinase KO mice suggesting that eEF2 kinase was not required for the antidepressant effects of lithium (Fig 3-2A). Inducible BDNF forebrain specific KO mice were also examined in the FST. Littermate control (CTL) mice that received lithium showed a significant reduction in immobility compared to vehicle treated CTL mice (Fig 3-3A). Likewise, BDNF KO mice that received lithium also showed a significant reduction in immobility compared to vehicle treated BDNF KO mice that did not receive lithium

(Fig 3-3A) suggesting that BDNF is not necessary for the antidepressant-like effect of lithium in the FST.

To investigate whether eEF2 kinase or BDNF were required for lithium’s antimanic effects, we used the amphetamine (AMPH) hyperlocomotor test in which lithium blunts the increased locomotor activity that occurs following acute AMPH injection (Gould et al., 2007;

Flaisher-Grinberg and Einat, 2010). In the first set of experiments we tested eEF2 kinase KO and littermate CTL mice. The CTL mice treated with acute AMPH show the expected significant increase in the total number of horizontal beam breaks (Fig 3-2B) compared to vehicle treated mice. Lithium treatment of CTL mice resulted in a slight trend towards an increase in the total number of beam breaks that was not significantly different from CTL mice treated with chronic

50 lithium and then given acute AMPH. Similarly, the eEF2 kinase KO mice showed similar behavioral effects compared to the CTL mice in all treatment groups, in two different cohorts of mice, suggesting that the loss of eEF2 kinase was not impacting the behavioral responses compared to littermate CTL mice (Fig 3-2B). In the next set of experiments, the inducible

BDNF KO mice were tested and the littermate CTL mice were found to have a statistically significant increase in the number of beam breaks following acute AMPH, with chronic lithium treatment producing a slight increase in beam breaks that was not significantly different compared to mice receiving chronic lithium and AMPH (Fig 3-3B). In contract, while inducible

BDNF KO mice receiving acute AMPH showed a significant increase in the number of beam breaks compared to vehicle treated BDNF KOs, and a slight increase to chronic lithium treatment, the BDNF KO mice receiving chronic lithium and acute AMPH still had a significant increase in the number of beam breaks demonstrating that lithium treatment does not block the hyperlocomotor effects of AMPH in mice with a forebrain deletion of BDNF and thus suggesting that BDNF is required for the antimanic effect of lithium (Fig 3-3B). Next, we assessed the combined effect of lithium and amphetamine on BDNF protein expression in the hippocampus.

Chronic lithium treatment caused a significant increase in BDNF protein expression in comparison to control and acute amphetamine injection alone (Fig 3-3C). Interestingly, a single injection of amphetamine reversed the increase in BDNF expression caused by lithium alone

(Fig 3-3C).

BDNF and TrkB required for lithium-mediated decreases in AMPAR mEPSC amplitude

Since BDNF was required for lithium’s antimanic effects in rodent models, we were intrigued whether BDNF was involved in lithium’s effects on synaptic function. Previous work has shown

51 that lithium treatment of dissociated neurons results in a significant decrease in AMPAR mEPSC amplitude that has been suggested to underlie the antimanic effects of lithium (Du et al., 2008;

Wei et al., 2010; Ankolekar and Sikdar, 2015). To explore whether BDNF is required for lithium’s effects on AMPAR mEPSC amplitude, we first incubated dissociated wild-type hippocampal neurons with 1 mM LiCl or 1 mM NaCl, to control for changes in osmolarity, for

11-15 days and then recorded AMPAR mEPSCs (Fig 3-4A). Lithium treatment produced a significant decrease in AMPAR mEPSC amplitude in comparison to control and NaCl treated wild-type neurons (Fig 3-4B). To determine if lithium’s effect on AMPAR mEPSC amplitude was due to synaptic scaling, the amplitudes from each condition were plotted in rank order and then analyzed with a linear fit equation. We found that lithium treatment mediated a 41% and

26% decrease in slope in comparison to control and NaCl treatment, respectively (Fig 3-4D) indicating the lithium treatment caused a downward scaling of all AMPAR mEPSC amplitudes.

The lithium effects were not generalized effects on synaptic measures as there were no changes in mEPSC frequency compared to control or NaCl treated neurons (Fig 3-4B, C, D, E). To explore whether the lithium mediated decrease in mEPSC amplitude was due to changes in surface expression of GluA1, we performed surface biotinylation experiments of our chronic lithium treated dissociated neurons. We found that lithium treatment, in comparison to vehicle or

NaCl treated neurons, resulted in a significant decrease in GluA1 surface expression relative to total GluA1 (Fig 3-4F).

To explore whether BDNF was required for lithium’s effects on AMPAR mEPSC amplitude, we used dissociated hippocampal cultures from fl.BDNF mice that were infected with lentivirus expressing Cre recombinase tagged with GFP (GFPCre) or GFP alone as a control.

Previous studies from our laboratory have shown that lenti-GFPCre can efficiently knock down

52 endogenous gene expression without triggering cell death (Nelson et al., 2006; Akhtar et al.,

2009). As expected, lithium treatment resulted in a significant decrease in AMPAR mEPSC amplitude in comparison to control or NaCl treated GFP infected neurons (Fig. 3-5A, C, D).

Similarly, flBDNF neurons infected with lenti-GFPCre and treated with NaCl were indistinguishable in AMPAR mEPSC amplitude compared to vehicle treatment (Fig. 3-5B, C,

D). In contrast, lithium treatment of fl.BDNF neurons infected with lenti-GFPCre did not result in a significant decrease in AMPAR mEPSC amplitude suggesting that BDNF was required for lithium’s effect on AMPAR mEPSC amplitude (Figure 3-5B, C, D). We also examined mEPSC frequency in dissociated neurons from the six treatment groups and did not observe any significant changes with any of the conditions (data not shown).

To further confirm that BDNF signaling is required for lithium mediated decreases in

AMPAR mediated synaptic transmission, we explored the requirement of TrkB. We utilized fl.TrkB dissociated neurons that were infected with either lenti-GFPCre or GFP. Consistent with our previous results, lithium treatment of GFP infected fl.TrkB neurons resulted in a significant decrease in AMPAR mEPSC amplitude compared to control or NaCl treated neurons (Fig 3-6A,

C, D). The fl.TrkB neurons infected with lenti-GFPCre had indistinguishable AMPA mEPSC amplitudes compared to GFP infected neurons demonstrating that loss of TrkB does not affect this synaptic measure (Fig. 3-6A, C, D). The NaCl treatment of fl.TrkB neurons infected with lenti-GFPCre did not alter AMPAR mEPSC amplitude (Fig 3-6B, C, D). However in agreement with the BDNF data, lithium treatment of fl.TrkB neurons infected with lenti-GFPCre did not affect AMPAR mEPSC amplitude demonstrating a requirement for TrkB in lithium’s effects on

AMPAR mEPSC amplitude (Fig. 3-6B, C, D.) The six treatment groups did not affect mEPSC frequency demonstrating a specific effect of lithium on mEPSC properties (data not shown).

Discussion

In the current study, we report a novel role for BDNF and TrkB in both the behavioral and electrophysiological effects of chronic lithium treatment. Our results confirm previous reports that chronic lithium treatment in rodents causes an increase in BDNF mRNA and protein in the hippocampus (Fukumoto et al., 2001; Yasuda et al., 2009; Jornada et al., 2010) and an antidepressant-like effect in the FST. We then tested whether BDNF is required for the antidepressant or antimanic effects of lithium. Intriguingly, we found that BDNF is not required for the antidepressant effects of lithium; however BDNF expression is required for lithium’s antimanic effect in the commonly used amphetamine hyperactivity test. We then investigated whether lithium’s known effects on AMPAR functioning also required BDNF/TrkB signaling.

We replicated prior studies showing lithium causes a significant decrease in AMPAR mEPSC amplitude, likely through decreased membrane AMPAR expression (Du et al., 2003; Gould et al., 2007; Du et al., 2008; Wei et al., 2010; Ankolekar and Sikdar, 2015). Finally, we found that both BDNF and its receptor TrkB are required for the lithium-mediated decrease in AMPAR mEPSC amplitude.

We have previously found that ketamine treatment caused a decrease in the phosphorylation of eEF2 and inhibition of eEF2-K caused an antidepressant-like in mice (Autry et al., 2011) Moreover, eEF2-K is required for ketamine’s rapid antidepressant-like effect

(Nosyreva et al., 2013). Lithium treatment has also been shown to decrease the overall phosphorylation of eEF2 in cultured neuronal-like cells and hippocampal lysates (Karyo et al.,

2010). Therefore we wanted to determine the necessity of eEF2-K in lithium’s behavioral effects. We found that chronic lithium treatment mediated both an antidepressant-like and

54 antimanic-like effect in constitutive eEF2-K KO mice, suggesting that eEF2-K is not required for lithium’s mood stabilizing effects. This is in contrast to our results with ketamine, which shows dissociation between the molecular requirements underlying the behavioral effects of lithium and rapid antidepressants.

BDNF expression is increased in the hippocampus following treatment with traditional antidepressants, acute low dose ketamine, and lithium (Autry and Monteggia, 2012). We found that lithium treatment produces a robust increase in BDNF expression in the hippocampus.

Previous work has demonstrated that BDNF is necessary for the antidepressant effects of traditional antidepressants and the fast-acting antidepressant ketamine (Monteggia et al., 2004;

Adachi et al., 2008; Autry et al., 2011). Therefore, we investigated whether BDNF would be required for the antidepressant effects of lithium. Rather surprisingly, we found that lithium produced an antidepressant response in the inducible BDNF KO mice, a line that has previously been shown to be unresponsive to the antidepressant-like effects of traditional antidepressants and ketamine in the FST. While lithium is more regarded as a mood stabilizer, it does produce antidepressant effects and is commonly used in conjunction with traditional antidepressants.

These data suggest that BDNF is not necessarily required for all drugs with antidepressant effects to mediate antidepressant-related responses.

To examine the antimanic effects of lithium we tested mice in the amphetamine hyperlocomotor paradigm (Frey et al., 2006a; Frey et al., 2006b; Gould et al., 2007). We found that our lithium treatment reduced the amphetamine induced hyperactivity in CTL mice. In contrast, lithium was unable counter the hyperactivity induced by amphetamine demonstrating a requirement for BDNF in this antimanic assessed effect of lithium. While the amphetamine hyperlocomotor paradigm is a commonly used behavioral paradigm to assess lithium’s

55 behavioral effects, few genes have been shown to be required for its antimanic effects. The demonstration that BDNF is required for the antimanic effect suggests a starting point to mechanistically link the behavioral effects of lithium.

As our data revealed a novel and unexpected role for BDNF in the antimanic effects of lithium, we examined whether it was involved in lithium’s effects on synaptic transmission.

Previous work has shown that lithium mediates a significant decrease in AMPAR mEPSC amplitude, a mechanism that has been hypothesized to underlie its antimanic effects (Du et al.,

2008). We therefore examined whether BDNF was required for lithium’s effects on AMPAR mEPSC amplitude. Primary hippocampal neurons from fl.BDNF mice were infected with lentivirus expressing Cre recombinase to knock-down the gene of interest. We found that lithium triggered a significant decrease on AMPAR mEPSC amplitude that was occluded in neurons lacking BDNF. To further confirm the requirement of BDNF in lithium’s effects on

AMPAR mEPSC amplitude, we examined lithium’s effect in primary hippocampal neurons in neurons with knockdown of TrkB. In agreement with the BDNF data, we found that TrkB was required for lithium to decrease AMPAR mEPSC amplitude. While our data revealed a significant requirement for BDNF in lithium’s effects on AMPAR mEPSC amplitude, the knockdown of TrkB fully blocked lithium’s effect on synaptic transmission. The more robust effect of TrkB on lithium mediated AMPAR mEPSC amplitude may be attributable to some low level expression of BDNF acting on TrkB receptors in the BDNF knockdown experiments or alternatively some minor effects of neurotrophin 3 (NT3) or neurotrophin 4 (NT4) activating the

TrkB receptor to impact lithium’s effects on synaptic transmission. We also found that lithium, as well as knockdown of either BDNF or TrkB, did not impact AMPAR mEPSC frequency suggesting that lithium’s effect on synaptic transmission is due to postsynaptic effects.

Previous in vivo and in vitro studies have shown that lithium decreases AMPAR signaling through reducing GluA1 and GluA2 surface expression (Du et al., 2004b; Du et al.,

2008; Wei et al., 2010). Lithium treatment has also been shown to decrease phosphorylation of

Thr840 on GluA1, which is associated with decreased AMPAR signaling (Szabo et al., 2009).

The electrophysiological findings in the current study suggest that lithium decreased AMPAR surface expression in hippocampal neurons in a BDNF-TrkB dependent manner. These findings are rather unexpected as increased BDNF expression is typically associated with increased

AMPAR surface expression (Autry et al., 2011; Nosyreva et al., 2013). However, there is precedent for BDNF to induce scaling in AMPAR surface expression. Specifically, acute BDNF has been shown to increase GluA1 and GluA2 surface expression, whereas chronic BDNF can decrease surface expression of GluA1 and GluA2 (Reimers et al., 2014). Our findings suggest that chronic lithium treatment increases BDNF expression in a chronic manner leading to synaptic downscaling. Collectively, our findings demonstrate that the lithium is decreasing

AMPAR mEPSCs through a BDNF dependent process, such as endocytosis, that reduces surface expression of GluA1.

Altered glutamatergic signaling and hyperactivity have been routinely implicated in the pathology of BD. Mania in bipolar patients has been associated with elevated glutamate signaling (Ongur et al., 2008; Lan et al., 2009). Additionally, increased expression of vesicular glutamate transporter (VGlut1) mRNA and glutamate have been observed in post-mortem BD brain tissue (Eastwood and Harrison, 2010). A recent study examining IPSC derived neurons from bipolar patients found lower action potential thresholds, increased evoked and spontaneous action potentials, and larger action potential amplitudes in comparison to neurons derived from heathy controls, indicating that neurons derived from BD patients are hyperactive (Mertens et al.,

2015). The hyperactivity in BD patient derived neurons was reversed with lithium treatment

(Mertens et al., 2015). Therefore, it is possible that lithium decreases AMPAR surface expression to counteract the increased glutamatergic signaling and hyperactivity associated with mania.

In conclusion, our data strengthen the previously described links between BDNF/TrkB, glutamatergic signaling, and the mood stabilizer lithium. We show that BDNF expression is necessary for the antimanic effects of lithium. At the synaptic level, lithium’s decrease of

AMPAR mEPSC amplitude requires both BDNF and TrkB expression. These findings provide novel insight into the molecular and cellular mechanisms underlying lithium’s effects.

Figure 3-1. Chronic lithium treatment causes antidepressant- like effect in mice and increases BDNF expression. (A). Timeline of chronic lithium treatment and behavioral testing. (B). (Left) Example standard curve of lithium counts used to calculate lithium concentration in blood serum in control and lithium-treated mice. (Right). Lithium concentration in blood serum following chronic lithium treatment. (C). LiCl chow caused a significant decrease in immobility in the FST compared to the control group (Student’s unpaired t test ***P<0.0001, n=10 mice per group). (D). Chronic lithium treatment caused a significant increase in BDNF mRNA expression in the hippocampus (Student’s unpaired t test *P=0.04, n=8-9 mice per group). (E). Chronic lithium treatment caused a significant increase in BDNF protein expression in the hippocampus (Student’s unpaired t test **P=0.002, n=10 mice per group). (F). Chronic lithium treatment caused a significant decrease in the phosphorylated eEF2/total eEF2 ratio in the hippocampus (Student’s unpaired t test *P=0.02, n=10 mice per group).

Figure 3-2. eEF2-K is not required for the antidepressant-like or antimanic-like effects of lithium. (A). Chronic lithium treatment caused a significant decrease in immobility time in the FST in WT littermate controls and eEF2-K KO mice (ANOVA F3,38=5.116, P=0.004, Bonferonni’s post hoc comparison Veh WT vs LiCl WT *P=0.02, Veh eEF2-K KO vs LiCl eEF2-K KO *P=0.01, n=10-12 per group). (B). Acute amphetamine injection caused an increase in locomotor activity in WT and eEF2-K KO mice in comparison to the saline injected animals. Chronic lithium treatment blunted the increased locomotor activity following acute amphetamine injection in both WT and eEF2-K KO mice. (ANOVA F7,74=7.448, P<0.0001, Tukey’s multiple comparison WT Veh-Sal vs WT Veh-Amph ***P=0.0001, WT LiCl-Sal vs WT LiCl-Amph P=0.46, eEF2-K KO Veh-Sal vs eEF2-K KO Veh-Amph **P=0.004, eEF2-K KO LiCl-Sal vs eEF2-K KO LiCl-Amph P=0.28, WT Veh-Sal vs WT LiCl-Sal P=0.82, eEF2-K KO Veh-Sal vs eEF2-K KO LiCl-Sal P=0.81, n=9-11 mice per group).

Figure 3-3. BDNF expression is required for the antimanic-like effect of lithium. (A) Chronic lithium treatment caused a significant decrease immobility time in FST in both control littermates and BDNF KO mice in comparison to the control chow (ANOVA F3,90=18.1 P<0.0001, Dunnett’s multiple comparisons Veh Ctl vs LiCl Ctl ****P<0.0001, Veh BDNF KO vs LiCl BDNF KO ****P<0.0001, n=19-28 mice per group). (B). Acute amphetamine injection caused an increase in locomotor activity in control and BDNF KO mice in comparison to the saline injected animals. Chronic lithium treatment blunted the increased locomotor activity following acute amphetamine injection in the control littermate mice, but was unable to decrease the increased locomotor activity following amphetamine in the BDNF KO mice (ANOVA F7,98=12.75 P<0.0001, Tukey’s multiple comparisons Ctl Veh-Sal vs Ctl Veh- Amph ****P<0.0001, Ctl LiCl-Sal vs Ctl LiCl-Amph P=0.97, BDNF KO Veh-Sal vs BDNF KO Veh-Amph ****P<0.0001, BDNF KO LiCl-Sal vs BDNF KO LiCl-Amph *P=0.01, Ctl Veh-Sal vs Ctl LiCl-Sal P=0.23, BDNF KO Veh-Sal vs BDNF KO LiCl-Sal P= 0.44, n=11-15 mice per group). (C) Chronic lithium treatment caused an increase in BDNF protein expression in the hippocampus in control littermate animals. Acute amphetamine injection brought BDNF expression back to control levels. There is some residual BDNF expression in the inducible BDNF KO mice. (ANOVA F3,54=4.881, P=0.0045, Tukey’s multiple comparisons Ctl Veh-Sal vs Ctl LiCl-Sal **P=0.006, Ctl Veh-Amph vs Ctl LiCl-Amph P=0.99, n=13-16 mice per group).

Figure 3-4. Chronic lithium treatment causes a significant decrease in AMPAR mEPSC amplitude and surface GluA1 expression. (A) Timeline for treatment of dissociated hippocampal neurons with LiCl and NaCl. (B) Cumulative probability histogram showing a significant leftward shift (decrease) in the amplitudes of AMPAR mESPCs of cells treated chronically treated with 1 mM LiCl in comparison to control and 1mM NaCl treated neurons (Kolmogorov-Smirnov test: Control vs 1 mM LiCl P=1.75x10- 43, D=0.234, 1mM NaCl vs 1mM LiCl P=8.14x10-27, D=0.167). (C) Example traces of AMPAR mEPSCs from Control (top), 1 mM NaCl treated (middle) and 1 mM LiCl treated (bottom) dissociated hippocampal neurons. (D) Rank order plot of control AMPAR mEPSC amplitudes vs 1 mM LiCl AMPAR mEPSC amplitudes found a 41% decrease. Rank order plot of 1mM NaCl AMPAR mEPSC amplitudes vs 1 mM LiCl AMPAR mEPSC amplitudes found a 26% decrease. (Control vs LiCl line of best fit y=.59x, Control vs NaCl line of best fit y=.85x, Difference between NaCl and LiCl .26) (E) AMPAR mEPSC frequency is indistinguishable between control, NaCl treated, and LiCl treated neurons (ANOVA F4,57=0.129 P=0.97, n=10-17 recordings). (F) Surface biotinylation experiments found that chronic lithium incubation in dissociated hippocampal neurons caused a significant decrease in the surface/total GluA1 ratio (ANOVA F2,21=2.911 P=0.08, Dunnett’s multiple comparisons Control vs NaCl P=.58, Control vs LiCl P=0.04).

Figure 3-5. BDNF expression is required for the lithium-mediated decrease in AMPAR mEPSC amplitude. (A) Example traces from GFP infected fl.BDNF neurons with no treatment (control, top) or 11-17 day incubation with 1 mM NaCl (middle) or 1 mM LiCl (bottom). (B) Example traces from Cre infected fl.BDNF neurons with no treatment (control, top) or 11-17 day incubation with 1 mM NaCl (middle) or 1 mM LiCl (bottom). (C) Cumulative probability histogram showing a significant leftward shift (decrease) in AMPAR mEPSC amplitudes between fl.BDNF-GFP Control and fl.BDNF-GFP LiCl conditions. There is also a significant leftward shift (decrease) between fl.BDNF-GFP LiCl and fl.BDNF-Cre LiCl conditions (Kolmogorov-Smirnov test: GFP Control vs GFP LiCl P<0.0001 D=0.309, GFP LiCl vs Cre LiCl P<0.0001 D=0.19). (D) Lithium caused a significant decrease in AMPAR mEPSC amplitudes in fl.BDNF- GFP neurons compared to control fl.BDNF-GFP neurons. Lithium did not cause a significant change in AMPAR mEPSC amplitudes between fl.BDNF-Cre neurons in comparison to control fl.BDNF- Cre neurons (ANOVA F5,59=5.694 P=0.0002, Tukey’s multiple comparisons GFP Control vs GFP LiCl **P=0.003, Cre Control vs Cre LiCl P=0.197).

Figure 3-6. TrkB expression is required for the lithium-mediated decrease in AMPAR mEPSC amplitude. (A) Example traces from GFP infected fl.TrkB neurons, Control (top), 1 mM NaCl treatment (middle), and 1 mM LiCl treatment (bottom). (B) Examples from Cre infected fl.TrkB neurons, Control (top), 1 mM NaCl treatment (middle) and 1 mM LiCl treatment (bottom). (C) Cumulative probability histogram showing a significant leftward shift (decrease) in AMPAR mEPSC amplitudes from fl.TrkB-GFP neurons treated with 1 mM LiCl. (Kolmogorov-Smirnov test: GFP control vs GFP LiCl P<0.0001, D=0.181, GFP LiCl vs Cre LiCl P<0.0001, D=0.153). (D). Lithium caused a significant decrease in AMPAR mEPSC amplitude in fl.TrkB-GFP neurons compared to fl.Trk-GFP control. Lithium was unable to cause any significant changes in AMPAR mEPSC in fl.TrkB-Cre neurons compared to fl.TrkB-Cre control (ANOVA F5,76=5.107 P= 0.0004, Bonferonni multiple comparison test: GFP control vs GFP LiCl **P=0.002, Cre Control vs Cre LiCl P>0.999).

CHAPTER 4 CONCLUSIONS AND FUTURE DIRECTIONS

In the preceding chapters, I describe my recent discoveries that define mechanistic criteria related to NMDAR blockade mediating a rapid antidepressant effect. I also provide the first evidence regarding the requirement for BDNF/TrkB signaling in lithium’s antimanic effects.

Additionally, I show that BDNF and TrkB are required for the lithium-mediated AMPAR mEPSC amplitude downscaling in hippocampal neurons. However, I found that unlike BDNF, eEF2 kinase expression is not required for the behavioral effects of lithium. These data add to the growing knowledge of the molecular and cellular mechanisms required for treatment of

MDD and BD. While my present findings extend our understanding of the requirements for mediating a fast-acting antidepressant effect, there are many open questions relating to synaptic vesicle pools upstream and signaling downstream of the NMDAR that is important for rapid antidepressant effects. Additionally, although my results showing the necessity for BDNF/TrkB in both the behavioral and electrophysiological effects of lithium are a promising start, they reveal many new avenues of research into the role of BDNF in mania and AMPAR downscaling.

The rapid antidepressant effect of ketamine generated much excitement in the field of psychiatry, however the side effect profile of ketamine has limited its widespread use for the treatment of MDD and depressive episodes in BD patients. Subsequently, pre-clinical research began to elucidate the molecular requirements for the rapid antidepressant effects and determined that antagonizing the NMDAR at rest was the primary mechanism. Therefore, memantine, a

NMDAR antagonist FDA-approved to treat Alzheimer’s disease, was tested in placebo controlled clinical trials for its ability to mediate an antidepressant response in depressed patients. Rather surprisingly, treatment with memantine did not cause an antidepressant effect,

65 even when given for more than eight weeks (Zarate et al., 2006b). Thus, I wanted to directly compare ketamine and memantine in their ability to block the NMDAR at rest using electrophysiological recording methods. Previous work suggested that the key in uncovering differences in their NMDAR blockade would be physiological Mg2+ in the extracellular recording solution as ketamine has a higher binding affinity to NMDAR subunits, GluN2A and

GluN2B, than memantine which is exacerbated in the presence of Mg2+. As previously reported, both ketamine and memantine mediated a significant and equivalent blockade of the NMDAR when the AMPAR was blocked and no Mg2+ was included in the extracellular solution.

However, when Mg2+ was present, only ketamine was able to mediate a significant blockade of the NMDAR at rest as measured by decreased charge transfer and decreased decay time of the mEPSC. These results allowed me to conclude that this is the primary effect underlying the difference in ketamine and memantine’s ability to cause a rapid antidepressant response.

The differential effects of ketamine and memantine on NMDAR blockade extended to downstream signaling known to be required for ketamine’s rapid antidepressant effects. Our group had previously shown that BDNF protein expression was increased in the hippocampus 30 min after ketamine injection (Autry et al., 2011). Additionally, the increase in BDNF protein was found to be due to decreased activity of eEF2 kinase leading to decreased phosphorylation of eEF2 and ultimately desupression of mRNA translation. Interestingly, the changes in eEF2 phosphorylation and BDNF expression were back to control levels by 3 hours post injection, indicating that mechanisms downstream of BDNF/TrkB may mediate the longer lasting antidepressant effect of ketamine (Autry et al., 2011). My data extends these observations as I also found that ketamine caused a significant antidepressant-like effect in the FST at 8 hrs and 24 hrs following i.p. injection but no difference in BDNF expression or eEF2 phosphorylation in the

66 hippocampus at those time points. This data reinforces the notion that the eEF2/eEF2 kinase pathway which increased BDNF translation in necessary for the rapid antidepressant effects of ketamine while activation of TrkB and its downstream signaling cascade may be important for the antidepressant effects seen days after a single ketamine injection. Previous work has shown that specifically inhibiting eEF2 kinase causes an antidepressant effect in preclinical rodent models. Additionally, eEF2 kinase inhibitors increase local protein translation of multiple transcripts, including BDNF (Sutton et al., 2007; Autry et al., 2011). Therefore it has been hypothesized that inhibition eEF2 kinase might be a viable target for development of new rapidly acting antidepressants. However it should be noted that eEF2 kinase activity is known to regulate cell cycle progression and induction of autophagy and apoptosis in addition to protein translation elongation (Xie et al., 2014), indicating that extreme caution should be used in development of eEF2 kinase inhibitors as antidepressants.

The ability of ketamine to block the NMDAR at rest and the resulting increase in protein translation and synaptic potentiation highlights the importance of spontaneous neurotransmission and homeostatic plasticity. Spontaneous neurotransmission occurs independent of action potentials as a result of a single synaptic vesicle fusing with the presynaptic membrane and releasing neurotransmitter. Previous work has shown that spontaneous transmission is a key regulator of protein translation at the synapse and homeostatic plasticity. Activation of

NMDARs by spontaneously released vesicles has been shown to increase phosphorylation of eEF2 while blockade of the NMDAR causes a decrease in eEF2 phosphorylation in a synapse- specific manner (Sutton et al., 2006). Unphosphorylated eEF2 is able to bind the ribosome and mediate increases in local protein translation at the synapse. Additionally, blockade of the

NMDAR at rest induces homeostatic plasticity mechanisms that increase membrane expression

67 of AMPAR subunits GluA1 and GluA2 and causes increased mEPSC amplitude and potentiation of evoked AMPAR responses (Autry et al., 2011; Nosyreva et al., 2013; Izumi and Zorumski,

2014). Work from the Kavalali lab supports the idea that there are different pools of glutamatergic vesicles for spontaneous and activity-dependent release (Atasoy et al., 2008;

Ramirez et al., 2012; Bal et al., 2013), which suggests that regulation of the spontaneous vesicles and associated post-synaptic receptors is possible. Therefore specifically inhibiting release of glutamate from spontaneous vesicles might be another way to mediate a rapid antidepressant response while forgoing the known problems associated with antagonizing the NMDAR.

Lithium was one of the first recognized treatments for mania. However, there is an ongoing debate over the molecular and cellular mechanisms mediating lithium’s therapeutic effects. There are strong links between BD, lithium treatment of BD, and BDNF/TrkB signaling, yet the role of BDNF/TrkB in lithium’s mood stabilizing effects are poorly understood. Based on previous work demonstrating that BDNF and TrkB expression are required for the behavioral effects of traditional antidepressants and the rapid antidepressant effect of ketamine, I initially hypothesized that BDNF would also be necessary for lithium’s antidepressant effect. Rather surprisingly, I found that chronic lithium treatment mediated an antidepressant-like effect in inducible BDNF KO mice and their control littermates in the FST. However, lithium is known to have stronger antimanic effects than antidepressant effects suggesting that the underlying mechanisms could be separate. Therefore, I tested the lithium treated inducible BDNF KO mice in amphetamine induced hyperlocomotor antimanic efficacy paradigm and found that lithium did not cause an antimanic response in the BDNF KO mice indicating that BDNF is required for the antimanic effects of lithium. While this result is very exciting, it leads to many new questions.

One question is the role of BDNF’s high affinity receptor TrkB in lithium’s antimanic effects.

Additionally, BDNF expression is decreased in the broad forebrain in our inducible BDNF KO mice, therefore it would be interesting to know if there is any regional specificity for the requirement of BDNF in lithium’s antimanic effect. Also, although the amphetamine induced hyperlocomotor test is commonly used to test the efficacy of mood stabilizing treatments, it is only one test. Thus, it would be important to examine whether BDNF is required for lithium’s antimanic effects in other behavioral models of mania, such as the ouabain-induced mania model.

Lithium has also been found to decrease the overall phosphorylation of eEF2 in both cultured human neuronal-like cells and hippocampal lysates from mice, but it appeared to be due to an increase in the expression of total eEF2 not a decrease in phosphorylation, per se (Karyo et al., 2010). Previous work from the Monteggia laboratory had shown that eEF2 kinase inhibition caused an increase in BDNF expression and a rapid antidepressant effect in mice (Autry et al.,

2011). Additionally, experiments with eEF2 kinase KO mice showed that eEF2 kinase is required for the antidepressant effects of ketamine (Nosyreva et al., 2013), therefore I wanted to determine if eEF2 kinase is necessary for the behavioral effects of lithium. I treated constitutive eEF2 kinase KO mice with lithium then tested the mice in the FST to assess antidepressant effects of lithium and in the amphetamine hyperlocomotor test to investigate the antimanic effects of lithium. Unlike the inducible BDNF KO mice, the eEF2 kinase KO mice exhibited both antidepressant-like and antimanic-like behaviors with lithium treatment, indicating that eEF2 kinase is not required for the therapeutic effects of lithium.

The glutamatergic neurotransmitter system has been shown to be regulated in opposite ways by BD and lithium treatment. Previous work has found increased glutamate and expression of VGlut1 in BD patients (Eastwood and Harrison, 2010). In rodents, a single injection of

69 amphetamine, which is used to mimic mania, caused increased surface expression of GluA1 in the hippocampus (Mao et al., 2015). Conversely, lithium treatment is associated with significant downregulation of the gene encoding GluA2 in BD patients. Similarly, decreased AMPAR mEPSC amplitude and decreased surface expression of GluA1 and GluA2 is seen following lithium treatment in cortical cultures and hippocampal lysates, respectively (Du et al., 2003; Du et al., 2008; Wei et al., 2010; Ankolekar and Sikdar, 2015). Therefore I wanted to determine if chronic lithium treatment mediated a decrease in AMPAR mEPSC amplitude in hippocampal culture and if BDNF or TrkB were required for this decrease. As had been reported in cortical culture, I confirmed that chronic lithium treatment produces a significant decrease in AMPAR mESPC amplitude but does not affect frequency in dissociated hippocampal neurons.

Additionally, using rank order plotting, I found that lithium treatment caused an overall decrease of all mEPSC amplitudes in comparison to both control and NaCl treatment, suggestive of a homeostatic downscaling effect of lithium. I also confirmed that the lithium mediated decrease in AMPAR mEPSC amplitude was due to decreased AMPAR surface expression using surface biotinylation following chronic lithium treatment. Then using fl.BDNF and fl.TrkB neurons, I demonstrated that deletion of BDNF or TrkB caused AMPAR mEPSC amplitudes to return to control levels, indicating that both BDNF and TrkB are required for lithium’s effect on AMPAR signaling. While these results highlight a previously unknown role of BDNF/TrkB in regulating lithium’s effect on AMPAR signaling, the connection to behavior and underlying molecular mechanisms are unclear.

There are many experiments that can be done to gain a better understanding of the connection between BDNF/TrkB and lithium-mediated AMPAR mEPSC amplitude downscaling and its relevance for behavior. First, it will be important to identify the mechanistic link between

BDNF/TrkB and decreased AMPAR mEPSC amplitude. Previous work has shown that GSK3β inhibitors also cause decreased AMPAR mEPSC amplitude and AMPAR subunit surface expression (Wei et al., 2010). Additionally, increased BDNF expression and TrkB activation has been found to cause increased phosphorylation and inhibition of GSK3β (Mai et al., 2002;

Ortega et al., 2010; Smillie et al., 2013). Alternatively, increased GSK3β phosphorylation and inhibition has been shown to increase transcription of BDNF mRNA (Yasuda et al., 2009).

However, it is unknown which of these events occurs first in relation to lithium, increased BDNF expression or GSK3β inhibition. Therefore, I propose to directly test the necessity of

BDNF/TrkB signaling for GSK3β inhibition mediated decreased AMPAR mEPSC amplitude by treating fl.BDNF and fl.TrkB hippocampal cultures with a specific GSK3 inhibitor and recording

AMPAR mEPSCs. Additionally, the link between GSK3β inhibition, BDNF/TrkB signaling, and mood can be assessed by treating BDNF KO animals with a specific GSK3 inhibitor and using the amphetamine induced hyperlocomotor antimanic test to determine the requirement for

BDNF in the antimanic effects of GSK3 inhibition. These experiments would begin to answer whether the electrophysiological and behavioral effects of lithium are mediated through BDNF’s inhibition of GSK3 or inhibition of GSK3 causing an increase in BDNF.

The link between increased BDNF expression and decreased AMPAR signaling is another open question. Previous work in our lab and others has found that an increase in BDNF expression is associated with increased AMPAR surface expression, increased AMPAR mEPSC amplitude and potentiation of evoked AMPAR responses (Sutton et al., 2006; Sutton et al., 2007;

Autry et al., 2011; Nosyreva et al., 2013; Nosyreva et al., 2014). However, it should be noted that the increased AMPAR signaling was seen within minutes to hours of the increase in BDNF expression. Additionally, a recent study found distinct differences in AMPAR surface

71 expression in cultured neurons following either 30 min or 24 hr incubation with exogenous

BDNF (Reimers et al., 2014). After 30 min incubation with BDNF, surface expression of both

GluA1 and GluA2 were significantly increased, however 24 hr BDNF treatment caused a significant decrease in GluA1 and GluA2 expression at the membrane, suggesting differential regulation of AMPAR membrane expression by acute or chronic BDNF exposure. These experiments were done with incubation of exogenous BDNF therefore it is unclear if similar mechanisms exist when endogenous BDNF expression is increased.

A recent study compared hippocampal-like neurons derived from BD patient iPSCs in a manic state to hippocampal-like neurons from healthy controls to assess potential differences due to BD (Mertens et al., 2015). The authors found that neurons from the BD patients had increased evoked and spontaneous action potentials and action potential amplitudes that were ameliorated by lithium treatment, suggestive of a direct link between neuronal hyperactivity and mania that is treated by lithium. While these results are exciting, the authors did not assess AMPAR signaling or BDNF expression in their studies. Therefore, it would be interesting to determine whether there are any differences in AMPAR neurotransmission or surface expression of AMPAR subunits between the BD patient neurons and healthy controls. Also, what effect does lithium treatment have on AMPARs in BD patient neurons. Additionally, BD is characterized by depressive episodes indicating that there could be different electrophysiological signatures associated with depression.

In conclusion, the current research has furthered our understanding the molecular mechanisms regulating the antidepressant effects of ketamine and the antimanic effects of lithium. These studies provide novel insight into the synaptic regulation of mood disorder treatments and pose many intriguing questions critical to the field of psychiatry. I hope that this

72 research will provide the field with new questions and avenues to answer them to expand our knowledge regarding the role of BDNF/TrkB and eEF2/eEF2K in antidepressant and antimanic efficacy.

REFERENCES

Adachi M, Barrot M, Autry AE, Theobald D, Monteggia LM (2008) Selective loss of brain- derived neurotrophic factor in the dentate gyrus attenuates antidepressant efficacy. Biological psychiatry 63:642-649.

Akhtar MW, Raingo J, Nelson ED, Montgomery RL, Olson EN, Kavalali ET, Monteggia LM (2009) Histone deacetylases 1 and 2 form a developmental switch that controls excitatory synapse maturation and function. The Journal of neuroscience : the official journal of the Society for Neuroscience 29:8288-8297.

Amdisen A (1980) Serum concentration and clinical supervision in monitoring of lithium treatment. Therapeutic drug monitoring 2:73-83.

American Psychiatric Association (2013) Diagnostic and Statistical Manual of Mental Disorders. In, 5th Edition. Washington, D.C.

Angelucci F, Aloe L, Jimenez-Vasquez P, Mathe AA (2003) Lithium treatment alters brain concentrations of nerve growth factor, brain-derived neurotrophic factor and glial cell line- derived neurotrophic factor in a rat model of depression. The international journal of neuropsychopharmacology / official scientific journal of the Collegium Internationale Neuropsychopharmacologicum 6:225-231.

Ankolekar SM, Sikdar SK (2015) Early postnatal exposure to lithium in vitro induces changes in AMPAR mEPSCs and vesicular recycling at hippocampal glutamatergic synapses. Journal of biosciences 40:339-354.

Antelman SM et al. (1998) The effects of lithium on a potential cycling model of bipolar disorder. Progress in neuro-psychopharmacology & biological psychiatry 22:495-510.

Atasoy D, Ertunc M, Moulder KL, Blackwell J, Chung C, Su J, Kavalali ET (2008) Spontaneous and evoked glutamate release activates two populations of NMDA receptors with limited overlap. The Journal of neuroscience : the official journal of the Society for Neuroscience 28:10151-10166.

Autry AE, Monteggia LM (2012) Brain-derived neurotrophic factor and neuropsychiatric disorders. Pharmacological reviews 64:238-258.

Autry AE, Adachi M, Nosyreva E, Na ES, Los MF, Cheng PF, Kavalali ET, Monteggia LM (2011) NMDA receptor blockade at rest triggers rapid behavioural antidepressant responses. Nature 475:91-95.

Baastrup PC, Schou M (1967) Lithium as a prophylactic agents. Its effect against recurrent depressions and manic-depressive psychosis. Archives of general psychiatry 16:162-172.

Bal M, Leitz J, Reese AL, Ramirez DM, Durakoglugil M, Herz J, Monteggia LM, Kavalali ET (2013) Reelin mobilizes a VAMP7-dependent synaptic vesicle pool and selectively augments spontaneous neurotransmission. Neuron 80:934-946.

Barnett JH, Smoller JW (2009) The genetics of bipolar disorder. Neuroscience 164:331-343.

Beaulieu JM et al. (2008) A beta-arrestin 2 signaling complex mediates lithium action on behavior. Cell 132:125-136.

Berman RM, Cappiello A, Anand A, Oren DA, Heninger GR, Charney DS, Krystal JH (2000) Antidepressant effects of ketamine in depressed patients. Biological psychiatry 47:351-354.

Beurel E, Grieco SF, Jope RS (2015) Glycogen synthase kinase-3 (GSK3): regulation, actions, and diseases. Pharmacology & therapeutics 148:114-131.

Blier P, de Montigny C, Chaput Y (1987) Modifications of the serotonin system by antidepressant treatments: implications for the therapeutic response in major depression. Journal of clinical psychopharmacology 7:24S-35S.

Bodnoff SR, Suranyi-Cadotte B, Aitken DH, Quirion R, Meaney MJ (1988) The effects of chronic antidepressant treatment in an animal model of anxiety. Psychopharmacology 95:298- 302.

Boer U, Eglins J, Krause D, Schnell S, Schofl C, Knepel W (2007) Enhancement by lithium of cAMP-induced CRE/CREB-directed gene transcription conferred by TORC on the CREB basic leucine zipper domain. The Biochemical journal 408:69-77.

Cade JF (1949) Lithium salts in the treatment of psychotic excitement. The Medical journal of Australia 2:349-352.

Carter J, Story DA (2013) Veterinary and human anaesthesia: an overview of some parallels and contrasts. Anaesthesia and intensive care 41:710-718.

Chatterton JE et al. (2002) Excitatory glycine receptors containing the NR3 family of NMDA receptor subunits. Nature 415:793-798.

Chuang DM et al. (2002) Neuroprotective effects of lithium in cultured cells and animal models of diseases. Bipolar disorders 4:129-136.

Cohen P, Frame S (2001) The renaissance of GSK3. Nature reviews Molecular cell biology 2:769-776.

Collingridge GL, Herron CE, Lester RA (1988a) Synaptic activation of N-methyl-D-aspartate receptors in the Schaffer collateral-commissural pathway of rat hippocampus. The Journal of physiology 399:283-300.

Collingridge GL, Herron CE, Lester RA (1988b) Frequency-dependent N-methyl-D-aspartate receptor-mediated synaptic transmission in rat hippocampus. The Journal of physiology 399:301- 312.

Coppen A, Standish-Barry H, Bailey J, Houston G, Silcocks P, Hermon C (1991) Does lithium reduce the mortality of recurrent mood disorders? Journal of affective disorders 23:1-7.

Cosgrove VE, Kelsoe JR, Suppes T (2016) Toward a Valid Animal Model of Bipolar Disorder: How the Research Domain Criteria Help Bridge the Clinical-Basic Science Divide. Biological psychiatry 79:62-70.

Cundall RL, Brooks PW, Murray LG (1972) A controlled evaluation of lithium prophylaxis in affective disorders. Psychological medicine 2:308-311.

Cunha AB, Frey BN, Andreazza AC, Goi JD, Rosa AR, Goncalves CA, Santin A, Kapczinski F (2006) Serum brain-derived neurotrophic factor is decreased in bipolar disorder during depressive and manic episodes. Neuroscience letters 398:215-219. de Azevedo Cardoso T et al. (2014) Neurotrophic factors, clinical features and gender differences in depression. Neurochemical research 39:1571-1578.

de Sousa RT, van de Bilt MT, Diniz BS, Ladeira RB, Portela LV, Souza DO, Forlenza OV, Gattaz WF, Machado-Vieira R (2011) Lithium increases plasma brain-derived neurotrophic factor in acute bipolar mania: a preliminary 4-week study. Neuroscience letters 494:54-56.

DeWilde KE, Levitch CF, Murrough JW, Mathew SJ, Iosifescu DV (2015) The promise of ketamine for treatment-resistant depression: current evidence and future directions. Annals of the New York Academy of Sciences 1345:47-58.

Du J, Quiroz J, Yuan P, Zarate C, Manji HK (2004a) Bipolar disorder: involvement of signaling cascades and AMPA receptor trafficking at synapses. Neuron glia biology 1:231-243.

Du J, Gray NA, Falke C, Yuan P, Szabo S, Manji HK (2003) Structurally dissimilar antimanic agents modulate synaptic plasticity by regulating AMPA glutamate receptor subunit GluR1 synaptic expression. Annals of the New York Academy of Sciences 1003:378-380.

Du J, Quiroz JA, Gray NA, Szabo ST, Zarate CA, Jr., Manji HK (2004b) Regulation of cellular plasticity and resilience by mood stabilizers: the role of AMPA receptor trafficking. Dialogues in clinical neuroscience 6:143-155.

Du J, Gray NA, Falke CA, Chen W, Yuan P, Szabo ST, Einat H, Manji HK (2004c) Modulation of synaptic plasticity by antimanic agents: the role of AMPA glutamate receptor subunit 1 synaptic expression. The Journal of neuroscience : the official journal of the Society for Neuroscience 24:6578-6589.

Du J et al. (2010) A kinesin signaling complex mediates the ability of GSK-3beta to affect mood-associated behaviors. Proceedings of the National Academy of Sciences of the United States of America 107:11573-11578.

Du J et al. (2008) The role of hippocampal GluR1 and GluR2 receptors in manic-like behavior. The Journal of neuroscience : the official journal of the Society for Neuroscience 28:68-79.

Duman RS, Monteggia LM (2006) A neurotrophic model for stress-related mood disorders. Biological psychiatry 59:1116-1127.

Dwivedi T, Zhang H (2014) Lithium-induced neuroprotection is associated with epigenetic modification of specific BDNF gene promoter and altered expression of apoptotic-regulatory proteins. Frontiers in neuroscience 8:457.

Eastwood SL, Harrison PJ (2010) Markers of glutamate synaptic transmission and plasticity are increased in the anterior cingulate cortex in bipolar disorder. Biological psychiatry 67:1010- 1016.

Emnett CM, Eisenman LN, Taylor AM, Izumi Y, Zorumski CF, Mennerick S (2013) Indistinguishable synaptic pharmacodynamics of the N-methyl-D-aspartate receptor channel blockers memantine and ketamine. Molecular pharmacology 84:935-947.

Espinosa F, Kavalali ET (2009) NMDA receptor activation by spontaneous glutamatergic neurotransmission. Journal of neurophysiology 101:2290-2296.

Farren CK, Hill KP, Weiss RD (2012) Bipolar disorder and alcohol use disorder: a review. Current psychiatry reports 14:659-666.

Ferguson JM, Shingleton RN (2007) An open-label, flexible-dose study of memantine in major depressive disorder. Clinical neuropharmacology 30:136-144.

Fernandes BS et al. (2011) Brain-derived neurotrophic factor as a state-marker of mood episodes in bipolar disorders: a systematic review and meta-regression analysis. Journal of psychiatric research 45:995-1004.

Fernandes BS et al. (2015) Peripheral brain-derived neurotrophic factor (BDNF) as a biomarker in bipolar disorder: a meta-analysis of 52 studies. BMC medicine 13:289.

Fieve RR (1970) Lithium in psychiatry. International journal of psychiatry 9:375-412.

Flaisher-Grinberg S, Einat H (2010) Strain-specific battery of tests for domains of mania: effects of valproate, lithium and imipramine. Frontiers in psychiatry 1:10.

Fleischhacker WW, Buchgeher A, Schubert H (1986) Memantine in the treatment of senile dementia of the Alzheimer type. Progress in neuro-psychopharmacology & biological psychiatry 10:87-93.

Frey BN, Andreazza AC, Cereser KM, Martins MR, Valvassori SS, Reus GZ, Quevedo J, Kapczinski F (2006a) Effects of mood stabilizers on hippocampus BDNF levels in an animal model of mania. Life sciences 79:281-286.

Frey BN, Valvassori SS, Reus GZ, Martins MR, Petronilho FC, Bardini K, Dal-Pizzol F, Kapczinski F, Quevedo J (2006b) Effects of lithium and valproate on amphetamine-induced oxidative stress generation in an animal model of mania. Journal of psychiatry & neuroscience : JPN 31:326-332.

Fries GR et al. (2015) Memory and brain-derived neurotrophic factor after subchronic or chronic amphetamine treatment in an animal model of mania. Journal of psychiatric research 68:329-336.

Fukumoto T, Morinobu S, Okamoto Y, Kagaya A, Yamawaki S (2001) Chronic lithium treatment increases the expression of brain-derived neurotrophic factor in the rat brain. Psychopharmacology 158:100-106.

Gershon S (1972) Lithium salts in the management of the manic-depressive syndrome. Annual review of medicine 23:439-452.

Gideons ES, Kavalali ET, Monteggia LM (2014) Mechanisms underlying differential effectiveness of memantine and ketamine in rapid antidepressant responses. Proceedings of the National Academy of Sciences of the United States of America 111:8649-8654.

Gould TD, Einat H (2007) Animal models of bipolar disorder and mood stabilizer efficacy: a critical need for improvement. Neuroscience and biobehavioral reviews 31:825-831.

Gould TD, O'Donnell KC, Picchini AM, Manji HK (2007) Strain differences in lithium attenuation of d-amphetamine-induced hyperlocomotion: a mouse model for the genetics of clinical response to lithium. Neuropsychopharmacology : official publication of the American College of Neuropsychopharmacology 32:1321-1333.

Grandjean EM, Aubry JM (2009a) Lithium: updated human knowledge using an evidence-based approach: part III: clinical safety. CNS drugs 23:397-418.

Grandjean EM, Aubry JM (2009b) Lithium: updated human knowledge using an evidence-based approach. Part II: Clinical pharmacology and therapeutic monitoring. CNS drugs 23:331-349.

Gray NA, Du J, Falke CS, Yuan P, Manji HK (2003) Lithium regulates total and synaptic expression of the AMPA glutamate receptor GluR2 in vitro and in vivo. Annals of the New York Academy of Sciences 1003:402-404.

Greenberg PE, Fournier AA, Sisitsky T, Pike CT, Kessler RC (2015) The economic burden of adults with major depressive disorder in the United States (2005 and 2010). The Journal of clinical psychiatry 76:155-162.

Hamon M, Blier P (2013) Monoamine neurocircuitry in depression and strategies for new treatments. Progress in neuro-psychopharmacology & biological psychiatry 45:54-63.

Hashimoto R, Takei N, Shimazu K, Christ L, Lu B, Chuang DM (2002) Lithium induces brain- derived neurotrophic factor and activates TrkB in rodent cortical neurons: an essential step for neuroprotection against glutamate excitotoxicity. Neuropharmacology 43:1173-1179.

Hays RD, Wells KB, Sherbourne CD, Rogers W, Spritzer K (1995) Functioning and well-being outcomes of patients with depression compared with chronic general medical illnesses. Archives of general psychiatry 52:11-19.

Ibarguen-Vargas Y, Surget A, Vourc'h P, Leman S, Andres CR, Gardier AM, Belzung C (2009) Deficit in BDNF does not increase vulnerability to stress but dampens antidepressant-like effects in the unpredictable chronic mild stress. Behavioural brain research 202:245-251.

Izumi Y, Zorumski CF (2014) Metaplastic effects of subanesthetic ketamine on CA1 hippocampal function. Neuropharmacology 86:273-281.

Jahr CE, Stevens CF (1990a) A quantitative description of NMDA receptor-channel kinetic behavior. The Journal of neuroscience : the official journal of the Society for Neuroscience 10:1830-1837.

Jahr CE, Stevens CF (1990b) Voltage dependence of NMDA-activated macroscopic conductances predicted by single-channel kinetics. The Journal of neuroscience : the official journal of the Society for Neuroscience 10:3178-3182.

Johnson-Farley NN, Travkina T, Cowen DS (2006) Cumulative activation of akt and consequent inhibition of glycogen synthase kinase-3 by brain-derived neurotrophic factor and insulin-like growth factor-1 in cultured hippocampal neurons. The Journal of pharmacology and experimental therapeutics 316:1062-1069.

Jope RS (1999) A bimodal model of the mechanism of action of lithium. Molecular psychiatry 4:21-25.

Jope RS (2011) Glycogen synthase kinase-3 in the etiology and treatment of mood disorders. Frontiers in molecular neuroscience 4:16.

Jornada LK, Moretti M, Valvassori SS, Ferreira CL, Padilha PT, Arent CO, Fries GR, Kapczinski F, Quevedo J (2010) Effects of mood stabilizers on hippocampus and amygdala BDNF levels in an animal model of mania induced by ouabain. Journal of psychiatric research 44:506-510.

Kaidanovich-Beilin O, Woodgett JR (2011) GSK-3: Functional Insights from Cell Biology and Animal Models. Frontiers in molecular neuroscience 4:40.

Kaidanovich-Beilin O et al. (2009) Abnormalities in brain structure and behavior in GSK-3alpha mutant mice. Molecular brain 2:35.

Karyo R, Eskira Y, Pinhasov A, Belmaker R, Agam G, Eldar-Finkelman H (2010) Identification of eukaryotic elongation factor-2 as a novel cellular target of lithium and glycogen synthase kinase-3. Molecular and cellular neurosciences 45:449-455.

Kashiwagi K, Masuko T, Nguyen CD, Kuno T, Tanaka I, Igarashi K, Williams K (2002) Channel blockers acting at N-methyl-D-aspartate receptors: differential effects of mutations in the vestibule and ion channel pore. Molecular pharmacology 61:533-545.

Kavalali ET, Monteggia LM (2012) Synaptic mechanisms underlying rapid antidepressant action of ketamine. The American journal of psychiatry 169:1150-1156.

Kavalali ET, Monteggia LM (2015) How does ketamine elicit a rapid antidepressant response? Current opinion in pharmacology 20:35-39.

Kerner B (2014) Genetics of bipolar disorder. The application of clinical genetics 7:33-42.

Kessler RC, Chiu WT, Demler O, Merikangas KR, Walters EE (2005a) Prevalence, severity, and comorbidity of 12-month DSM-IV disorders in the National Comorbidity Survey Replication. Archives of general psychiatry 62:617-627.

Kessler RC, Berglund P, Demler O, Jin R, Merikangas KR, Walters EE (2005b) Lifetime prevalence and age-of-onset distributions of DSM-IV disorders in the National Comorbidity Survey Replication. Archives of general psychiatry 62:593-602.

Klein PS, Melton DA (1996) A molecular mechanism for the effect of lithium on development. Proceedings of the National Academy of Sciences of the United States of America 93:8455- 8459.

Koszewska I, Rybakowski JK (2009) Antidepressant-induced mood conversions in bipolar disorder: a retrospective study of tricyclic versus non-tricyclic antidepressant drugs. Neuropsychobiology 59:12-16.

Kotermanski SE, Johnson JW (2009) Mg2+ imparts NMDA receptor subtype selectivity to the Alzheimer's drug memantine. The Journal of neuroscience : the official journal of the Society for Neuroscience 29:2774-2779.

Kotermanski SE, Wood JT, Johnson JW (2009) Memantine binding to a superficial site on NMDA receptors contributes to partial trapping. The Journal of physiology 587:4589-4604.

Krishnan V, Nestler EJ (2011) Animal models of depression: molecular perspectives. Current topics in behavioral neurosciences 7:121-147.

Krystal JH, Karper LP, Seibyl JP, Freeman GK, Delaney R, Bremner JD, Heninger GR, Bowers MB, Jr., Charney DS (1994) Subanesthetic effects of the noncompetitive NMDA antagonist, ketamine, in humans. Psychotomimetic, perceptual, cognitive, and neuroendocrine responses. Archives of general psychiatry 51:199-214.

Lan MJ, McLoughlin GA, Griffin JL, Tsang TM, Huang JT, Yuan P, Manji H, Holmes E, Bahn S (2009) Metabonomic analysis identifies molecular changes associated with the pathophysiology and drug treatment of bipolar disorder. Molecular psychiatry 14:269-279.

Leal G, Comprido D, Duarte CB (2014) BDNF-induced local protein synthesis and synaptic plasticity. Neuropharmacology 76 Pt C:639-656.

Lenox RH, Watson DG (1994) Lithium and the brain: a psychopharmacological strategy to a molecular basis for manic depressive illness. Clinical chemistry 40:309-314.

Lenze EJ, Skidmore ER, Begley AE, Newcomer JW, Butters MA, Whyte EM (2012) Memantine for late-life depression and apathy after a disabling medical event: a 12-week, double-blind placebo-controlled pilot study. International journal of geriatric psychiatry 27:974-980.

Lepack AE, Fuchikami M, Dwyer JM, Banasr M, Duman RS (2015) BDNF release is required for the behavioral actions of ketamine. The international journal of neuropsychopharmacology / official scientific journal of the Collegium Internationale Neuropsychopharmacologicum 18.

Li N, Lee B, Liu RJ, Banasr M, Dwyer JM, Iwata M, Li XY, Aghajanian G, Duman RS (2010) mTOR-dependent synapse formation underlies the rapid antidepressant effects of NMDA antagonists. Science 329:959-964.

Li R, el-Mallakh RS, Harrison L, Changaris DG, Levy RS (1997) Lithium prevents ouabain- induced behavioral changes. Toward an animal model for manic depression. Molecular and chemical neuropathology / sponsored by the International Society for Neurochemistry and the World Federation of Neurology and research groups on neurochemistry and cerebrospinal fluid 31:65-72.

Liebrenz M, Borgeat A, Leisinger R, Stohler R (2007) Intravenous ketamine therapy in a patient with a treatment-resistant major depression. Swiss medical weekly 137:234-236.

Linden M, Helmchen H, Mackert A, Muller-Oerlinghausen B (1994) Structure and feasibility of a standardized stepwise drug treatment regimen (SSTR) for depressed inpatients. Pharmacopsychiatry 27 Suppl 1:51-53.

Liu RJ, Fuchikami M, Dwyer JM, Lepack AE, Duman RS, Aghajanian GK (2013) GSK-3 inhibition potentiates the synaptogenic and antidepressant-like effects of subthreshold doses of ketamine. Neuropsychopharmacology : official publication of the American College of Neuropsychopharmacology 38:2268-2277.

Lorenzi C, Pirovano A, Bosia M (2013) Molecular mechanisms of lithium salts in enhancement of cAMP-stimulated CREB-dependent gene transcription. Pharmacogenomics 14:127-128.

Maeng S, Zarate CA, Jr., Du J, Schloesser RJ, McCammon J, Chen G, Manji HK (2008) Cellular mechanisms underlying the antidepressant effects of ketamine: role of alpha-amino-3-hydroxy-5- methylisoxazole-4-propionic acid receptors. Biological psychiatry 63:349-352.

Mai L, Jope RS, Li X (2002) BDNF-mediated signal transduction is modulated by GSK3beta and mood stabilizing agents. Journal of neurochemistry 82:75-83.

Malhi GS, Tanious M, Das P, Coulston CM, Berk M (2013) Potential mechanisms of action of lithium in bipolar disorder. Current understanding. CNS drugs 27:135-153.

Manji HK, Lenox RH (2000) Signaling: cellular insights into the pathophysiology of bipolar disorder. Biological psychiatry 48:518-530.

Mantovani M, Pertile R, Calixto JB, Santos AR, Rodrigues AL (2003) Melatonin exerts an antidepressant-like effect in the tail suspension test in mice: evidence for involvement of N- methyl-D-aspartate receptors and the L-arginine-nitric oxide pathway. Neuroscience letters 343:1-4.

Mao LM, Xue B, Jin DZ, Wang JQ (2015) Dynamic increases in AMPA receptor phosphorylation in the rat hippocampus in response to amphetamine. Journal of neurochemistry 133:795-805.

Martinowich K, Manji H, Lu B (2007) New insights into BDNF function in depression and anxiety. Nature neuroscience 10:1089-1093.

Merikangas KR, Akiskal HS, Angst J, Greenberg PE, Hirschfeld RM, Petukhova M, Kessler RC (2007) Lifetime and 12-month prevalence of bipolar spectrum disorder in the National Comorbidity Survey replication. Archives of general psychiatry 64:543-552.

Mertens J et al. (2015) Differential responses to lithium in hyperexcitable neurons from patients with bipolar disorder. Nature 527:95-99.

Michael N, Erfurth A, Ohrmann P, Gossling M, Arolt V, Heindel W, Pfleiderer B (2003) Acute mania is accompanied by elevated glutamate/glutamine levels within the left dorsolateral prefrontal cortex. Psychopharmacology 168:344-346.

Mines MA, Jope RS (2012) Brain region differences in regulation of Akt and GSK3 by chronic stimulant administration in mice. Cellular signalling 24:1398-1405.

Mitchell PB (2013) Bipolar disorder. Australian family physician 42:616-619.

Mitchell PB, Hadzi-Pavlovic D (2000) Lithium treatment for bipolar disorder. Bulletin of the World Health Organization 78:515-517.

Monteggia LM, Gideons E, Kavalali ET (2013) The role of eukaryotic elongation factor 2 kinase in rapid antidepressant action of ketamine. Biological psychiatry 73:1199-1203.

Monteggia LM et al. (2004) Essential role of brain-derived neurotrophic factor in adult hippocampal function. Proceedings of the National Academy of Sciences of the United States of America 101:10827-10832.

Muller-Oerlinghausen B, Muser-Causemann B, Volk J (1992) Suicides and parasuicides in a high-risk patient group on and off lithium long-term medication. Journal of affective disorders 25:261-269.

Muller-Oerlinghausen B et al. (1994) Mortality during initial and during later lithium treatment. A collaborative study by the International Group for the Study of Lithium-treated Patients. Acta psychiatrica Scandinavica 90:295-297.

Muzina DJ, Kemp DE, McIntyre RS (2007) Differentiating bipolar disorders from major depressive disorders: treatment implications. Annals of clinical psychiatry : official journal of the American Academy of Clinical Psychiatrists 19:305-312.

Nelson ED, Kavalali ET, Monteggia LM (2006) MeCP2-dependent transcriptional repression regulates excitatory neurotransmission. Current biology : CB 16:710-716.

Nestler EJ, Hyman SE (2010) Animal models of neuropsychiatric disorders. Nature neuroscience 13:1161-1169.

Nosyreva E, Autry AE, Kavalali ET, Monteggia LM (2014) Age dependence of the rapid antidepressant and synaptic effects of acute NMDA receptor blockade. Frontiers in molecular neuroscience 7:94.

Nosyreva E, Szabla K, Autry AE, Ryazanov AG, Monteggia LM, Kavalali ET (2013) Acute suppression of spontaneous neurotransmission drives synaptic potentiation. The Journal of neuroscience : the official journal of the Society for Neuroscience 33:6990-7002.

O'Brien WT, Harper AD, Jove F, Woodgett JR, Maretto S, Piccolo S, Klein PS (2004) Glycogen synthase kinase-3beta haploinsufficiency mimics the behavioral and molecular effects of lithium. The Journal of neuroscience : the official journal of the Society for Neuroscience 24:6791-6798.

Omata N, Murata T, Takamatsu S, Maruoka N, Mitsuya H, Yonekura Y, Fujibayashi Y, Wada Y (2008) Neuroprotective effect of chronic lithium treatment against hypoxia in specific brain regions with upregulation of cAMP response element binding protein and brain-derived neurotrophic factor but not nerve growth factor: comparison with acute lithium treatment. Bipolar disorders 10:360-368.

Ongur D, Jensen JE, Prescot AP, Stork C, Lundy M, Cohen BM, Renshaw PF (2008) Abnormal glutamatergic neurotransmission and neuronal-glial interactions in acute mania. Biological psychiatry 64:718-726.

Ortega F, Perez-Sen R, Morente V, Delicado EG, Miras-Portugal MT (2010) P2X7, NMDA and BDNF receptors converge on GSK3 phosphorylation and cooperate to promote survival in cerebellar granule neurons. Cellular and molecular life sciences : CMLS 67:1723-1733.

Park SW et al. (2015) Effects of mood-stabilizing drugs on dendritic outgrowth and synaptic protein levels in primary hippocampal neurons. Bipolar disorders 17:278-290.

Parsons CG, Danysz W, Quack G (1999) Memantine is a clinically well tolerated N-methyl-D- aspartate (NMDA) receptor antagonist--a review of preclinical data. Neuropharmacology 38:735-767.

Peltoniemi MA, Hagelberg NM, Olkkola KT, Saari TI (2016) Ketamine: A Review of Clinical Pharmacokinetics and Pharmacodynamics in Anesthesia and Pain Therapy. Clinical pharmacokinetics.

Phiel CJ, Klein PS (2001) Molecular targets of lithium action. Annual review of pharmacology and toxicology 41:789-813.

Polter A et al. (2010) Deficiency in the inhibitory serine-phosphorylation of glycogen synthase kinase-3 increases sensitivity to mood disturbances. Neuropsychopharmacology : official publication of the American College of Neuropsychopharmacology 35:1761-1774.

Poolsup N, Li Wan Po A, de Oliveira IR (2000) Systematic overview of lithium treatment in acute mania. Journal of clinical pharmacy and therapeutics 25:139-156.

Popik P, Skolnick P (1996) The NMDA antagonist memantine blocks the expression and maintenance of morphine dependence. Pharmacology, biochemistry, and behavior 53:791-797.

Porsolt RD, Bertin A, Jalfre M (1977) Behavioral despair in mice: a primary screening test for antidepressants. Archives internationales de pharmacodynamie et de therapie 229:327-336.

Povysheva NV, Johnson JW (2012) Tonic NMDA receptor-mediated current in prefrontal cortical pyramidal cells and fast-spiking interneurons. Journal of neurophysiology 107:2232- 2243.

Price RB, Nock MK, Charney DS, Mathew SJ (2009) Effects of intravenous ketamine on explicit and implicit measures of suicidality in treatment-resistant depression. Biological psychiatry 66:522-526.

Prickaerts J, Moechars D, Cryns K, Lenaerts I, van Craenendonck H, Goris I, Daneels G, Bouwknecht JA, Steckler T (2006) Transgenic mice overexpressing glycogen synthase kinase 3beta: a putative model of hyperactivity and mania. The Journal of neuroscience : the official journal of the Society for Neuroscience 26:9022-9029.

Ramirez DM, Khvotchev M, Trauterman B, Kavalali ET (2012) Vti1a identifies a vesicle pool that preferentially recycles at rest and maintains spontaneous neurotransmission. Neuron 73:121- 134.

Rantamaki T, Knuuttila JE, Hokkanen ME, Castren E (2006) The effects of acute and long-term lithium treatments on trkB neurotrophin receptor activation in the mouse hippocampus and anterior cingulate cortex. Neuropharmacology 50:421-427.

Reese AL, Kavalali ET (2015) Spontaneous neurotransmission signals through store-driven Ca(2+) transients to maintain synaptic homeostasis. eLife 4.

Reimers JM, Loweth JA, Wolf ME (2014) BDNF contributes to both rapid and homeostatic alterations in AMPA receptor surface expression in nucleus accumbens medium spiny neurons. The European journal of neuroscience 39:1159-1169.

Ryazanov AG, Shestakova EA, Natapov PG (1988) Phosphorylation of elongation factor 2 by EF-2 kinase affects rate of translation. Nature 334:170-173.

Ryazanov AG, Rudkin BB, Spirin AS (1991) Regulation of protein synthesis at the elongation stage. New insights into the control of gene expression in eukaryotes. FEBS letters 285:170-175.

Saarelainen T et al. (2003) Activation of the TrkB neurotrophin receptor is induced by antidepressant drugs and is required for antidepressant-induced behavioral effects. The Journal of neuroscience : the official journal of the Society for Neuroscience 23:349-357.

Salvadore G, Singh JB (2013) Ketamine as a fast acting antidepressant: current knowledge and open questions. CNS neuroscience & therapeutics 19:428-436.

Schildkraut JJ (1965) The catecholamine hypothesis of affective disorders: a review of supporting evidence. The American journal of psychiatry 122:509-522.

Schmitt A, Falkai P (2014) Comorbidity, stigma and emotional perception in psychiatric disorders. European archives of psychiatry and clinical neuroscience 264:1-2.

Schou M, Juel-Nielsen N, Stromgren E, Voldby H (1954) The treatment of manic psychoses by the administration of lithium salts. Journal of neurology, neurosurgery, and psychiatry 17:250- 260.

Seelan RS, Khalyfa A, Lakshmanan J, Casanova MF, Parthasarathy RN (2008) Deciphering the lithium transcriptome: microarray profiling of lithium-modulated gene expression in human neuronal cells. Neuroscience 151:1184-1197.

Selle V, Schalkwijk S, Vazquez GH, Baldessarini RJ (2014) Treatments for acute bipolar depression: meta-analyses of placebo-controlled, monotherapy trials of anticonvulsants, lithium and antipsychotics. Pharmacopsychiatry 47:43-52.

Sharma AN, Fries GR, Galvez JF, Valvassori SS, Soares JC, Carvalho AF, Quevedo J (2016) Modeling mania in preclinical settings: A comprehensive review. Progress in neuropsychopharmacology & biological psychiatry 66:22-34.

Shirayama Y, Chen AC, Nakagawa S, Russell DS, Duman RS (2002) Brain-derived neurotrophic factor produces antidepressant effects in behavioral models of depression. The Journal of neuroscience : the official journal of the Society for Neuroscience 22:3251-3261.

Shorter E (2009) The history of lithium therapy. Bipolar disorders 11 Suppl 2:4-9.

Shulman KI, Herrmann N, Walker SE (2013) Current place of monoamine oxidase inhibitors in the treatment of depression. CNS drugs 27:789-797.

Silberberg D, Anand NP, Michels K, Kalaria RN (2015) Brain and other nervous system disorders across the lifespan - global challenges and opportunities. Nature 527:S151-154.

Siuciak JA, Lewis DR, Wiegand SJ, Lindsay RM (1997) Antidepressant-like effect of brain- derived neurotrophic factor (BDNF). Pharmacology, biochemistry, and behavior 56:131-137.

Smillie KJ, Pawson J, Perkins EM, Jackson M, Cousin MA (2013) Control of synaptic vesicle endocytosis by an extracellular signalling molecule. Nature communications 4:2394.

Smith MA, Makino S, Kvetnansky R, Post RM (1995) Effects of stress on neurotrophic factor expression in the rat brain. Annals of the New York Academy of Sciences 771:234-239.

Smothers CT, Woodward JJ (2007) Pharmacological characterization of glycine-activated currents in HEK 293 cells expressing N-methyl-D-aspartate NR1 and NR3 subunits. The Journal of pharmacology and experimental therapeutics 322:739-748.

Stambolic V, Ruel L, Woodgett JR (1996) Lithium inhibits glycogen synthase kinase-3 activity and mimics wingless signalling in intact cells. Current biology : CB 6:1664-1668.

Sutherland C (2011) What Are the bona fide GSK3 Substrates? International journal of Alzheimer's disease 2011:505607.

Sutton MA, Wall NR, Aakalu GN, Schuman EM (2004) Regulation of dendritic protein synthesis by miniature synaptic events. Science 304:1979-1983.

Sutton MA, Taylor AM, Ito HT, Pham A, Schuman EM (2007) Postsynaptic decoding of neural activity: eEF2 as a biochemical sensor coupling miniature synaptic transmission to local protein synthesis. Neuron 55:648-661.

Sutton MA, Ito HT, Cressy P, Kempf C, Woo JC, Schuman EM (2006) Miniature neurotransmission stabilizes synaptic function via tonic suppression of local dendritic protein synthesis. Cell 125:785-799.

Szabo ST et al. (2009) Glutamate receptors as targets of protein kinase C in the pathophysiology and treatment of animal models of mania. Neuropharmacology 56:47-55.

Tramontina JF, Andreazza AC, Kauer-Sant'anna M, Stertz L, Goi J, Chiarani F, Kapczinski F (2009) Brain-derived neurotrophic factor serum levels before and after treatment for acute mania. Neuroscience letters 452:111-113.

Trivedi MH et al. (2006) Evaluation of outcomes with citalopram for depression using measurement-based care in STAR*D: implications for clinical practice. The American journal of psychiatry 163:28-40.

Trullas R, Skolnick P (1990) Functional antagonists at the NMDA receptor complex exhibit antidepressant actions. European journal of pharmacology 185:1-10.

Tsankova NM, Berton O, Renthal W, Kumar A, Neve RL, Nestler EJ (2006) Sustained hippocampal chromatin regulation in a mouse model of depression and antidepressant action. Nature neuroscience 9:519-525.

Tunca Z et al. (2014) Alterations in BDNF (brain derived neurotrophic factor) and GDNF (glial cell line-derived neurotrophic factor) serum levels in bipolar disorder: The role of lithium. Journal of affective disorders 166:193-200.

Wada A (2009) Lithium and neuropsychiatric therapeutics: neuroplasticity via glycogen synthase kinase-3beta, beta-catenin, and neurotrophin cascades. Journal of pharmacological sciences 110:14-28.

Walz JC, Frey BN, Andreazza AC, Cereser KM, Cacilhas AA, Valvassori SS, Quevedo J, Kapczinski F (2008) Effects of lithium and valproate on serum and hippocampal neurotrophin-3 levels in an animal model of mania. Journal of psychiatric research 42:416-421.

Walz JC, Magalhaes PV, Giglio LM, Cunha AB, Stertz L, Fries GR, Andreazza AC, Kapczinski F (2009) Increased serum neurotrophin-4/5 levels in bipolar disorder. Journal of psychiatric research 43:721-723.

Wei J, Liu W, Yan Z (2010) Regulation of AMPA receptor trafficking and function by glycogen synthase kinase 3. The Journal of biological chemistry 285:26369-26376.

Wilkinson D (2001) Drugs for treatment of Alzheimer's disease. International journal of clinical practice 55:129-134.

Wu R, Fan J, Zhao J, Calabrese JR, Gao K (2014) The relationship between neurotrophins and bipolar disorder. Expert review of neurotherapeutics 14:51-65.

Xia Y, Wang CZ, Liu J, Anastasio NC, Johnson KM (2010) Brain-derived neurotrophic factor prevents phencyclidine-induced apoptosis in developing brain by parallel activation of both the ERK and PI-3K/Akt pathways. Neuropharmacology 58:330-336.

Xie CM, Liu XY, Sham KW, Lai JM, Cheng CH (2014) Silencing of EEF2K (eukaryotic elongation factor-2 kinase) reveals AMPK-ULK1-dependent autophagy in colon cancer cells. Autophagy 10:1495-1508.

Yan XB, Hou HL, Wu LM, Liu J, Zhou JN (2007) Lithium regulates hippocampal neurogenesis by ERK pathway and facilitates recovery of spatial learning and memory in rats after transient global cerebral ischemia. Neuropharmacology 53:487-495.

Yasuda S, Liang MH, Marinova Z, Yahyavi A, Chuang DM (2009) The mood stabilizers lithium and valproate selectively activate the promoter IV of brain-derived neurotrophic factor in neurons. Molecular psychiatry 14:51-59.

Yu HS, Kim SH, Park HG, Kim YS, Ahn YM (2010) Activation of Akt signaling in rat brain by intracerebroventricular injection of ouabain: a rat model for mania. Progress in neuropsychopharmacology & biological psychiatry 34:888-894.

Zarate CA, Jr., Singh JB, Carlson PJ, Brutsche NE, Ameli R, Luckenbaugh DA, Charney DS, Manji HK (2006a) A randomized trial of an N-methyl-D-aspartate antagonist in treatment- resistant major depression. Archives of general psychiatry 63:856-864.

Zarate CA, Jr., Singh JB, Quiroz JA, De Jesus G, Denicoff KK, Luckenbaugh DA, Manji HK, Charney DS (2006b) A double-blind, placebo-controlled study of memantine in the treatment of major depression. The American journal of psychiatry 163:153-155.

Zarate CA, Jr. et al. (2012) Replication of ketamine's antidepressant efficacy in bipolar depression: a randomized controlled add-on trial. Biological psychiatry 71:939-946.

Zemlan FP, Garver DL (1990) Depression and antidepressant therapy: receptor dynamics. Progress in neuro-psychopharmacology & biological psychiatry 14:503-523.