The Downstream Targets of Complex I Dysfunction in Bipolar Disorder

Helena Kyunghee Kim

A thesis submitted in conformity with the requirements for the degree of PhD

Graduate Department of Pharmacology and Toxicology

University of Toronto

© Copyright by Helena Kyunghee Kim (2015)

ii

Title: The Downstream Targets of Complex I Dysfunction in Bipolar Disorder

Name: Helena Kyunghee Kim, Doctor of Philosophy

Department of Pharmacology and Toxicology, University of Toronto 2015

ABSTRACT

Mitochondrial complex I dysfunction is consistently reported in bipolar disorder (BD),

implicating its role in increased oxidative stress. Therefore, the overall aim of my PhD was to

examine potential downstream targets of complex I dysfunction in BD. First, microarray studies

examining complex I subunits in patients with BD or schizophrenia were re-evaluated to

examine if complex I defect was specific to BD. Results revealed that complex I subunits that are

involved in the electron transfer process are decreased only in patients with BD, suggesting that

oxidative stress from complex I dysfunction may be specific to BD. Next, using post-mortem

brain, I confirmed lower levels of complex I and its subunit, NDUFS7, in the frontal cortex of

patients with BD, supporting the role of complex I dysfunction in the pathophysiology of BD. As

an extension of a post-mortem brain study from our group demonstrating the mitochondria and

the synapse as targets of oxidative stress, I first examined downstream targets of complex I

dysfunction in the post-mortem brain, focusing on the mitochondria and the synapse. To

demonstrate the mitochondria as a target of complex I dysfunction, I measured the activation of

the NLRP3-inflammasome, which is a marker of mitochondrial oxidative stress. Results revealed

that increased activation of the NLRP3-inflamamsome may be specific to BD, suggesting that

mitochondria may be a target of complex I dysfunction. To examine the synapse as a target of

complex I defect, I measured oxidative modifications to the dopaminergic synapse as increased

dopamine signaling is known to underlie mania. Oxidative modifications to dopaminergic

iii

proteins, the dopamine transporter and tyrosine hydroxylase, were altered in the post-mortem

prefrontal cortex of patients with BD, demonstrating the synapse as another potential target of

complex I dysfunction in BD. Downstream targets of complex I dysfunction were further

explored in cell models, allowing us to directly inhibit the electron transfer process of complex I

using rotenone. The effect of lithium was also tested as it is uniquely used for the treatment of

BD and therefore allows us to assess if the observed alterations may be relevant to BD

pathology. Inhibiting complex I with rotenone increased protein oxidation and nitration, and

caused an increase in methylation and hydroxymethylation of DNA. Moreover, lithium pre-

treatment was able to decrease these alterations, suggesting that they may be downstream targets

of complex I defect in BD. Together, these studies suggest that complex I dysfunction may be

specific to BD, and that it may play an important role in the pathophysiology of this disease.

iv

ACKNOWLEDGEMENTS

First and foremost, I would like to thank my two supervisors, Dr. L. Trevor Young and Dr. Ana

C. Andreazza for their continuous support. In addition to the technical knowledge and the

incredible training opportunities I received under their supervision, I also learned to appreciate

the process that goes into producing high quality data, telling a compelling story, and taking

ownership of my work. Their enthusiasm for research and passion that they poured into their

work were also contagious, and I am certain that I will not be as passionate about research today

if it was not for their guidance.

I would also like to thank my supervisory committee members, Dr. Stephen Kish and Dr.

Rebecca Laposa, for their impeccable advice and support throughout my PhD. Their instruction

helped shape my presentation and writing style, and truly allowed me to grow as a researcher. I

would also like to express my gratitude for their immense help with this thesis.

Furthermore, I would like to thank my collaborators from Brazil who I am very grateful to have

had a chance to work with, Camila Nascimento and Gustavo Scola. The learning experience I

shared with them are memories I will always treasure, and I want to thank them for all the new

things they introduced me to inside and outside the lab. I would also like to thank my amazing

students, Karina Mendonca, Cameron Isaacs-Trepanier, Nika Elmi, Wenjun Chen, and David

(Pok Yik) Yeung, who were some of the most intelligent, persistent, patient and diligent

individuals I have had the good fortune to meet.

I would also like to thank my mother (Eunja Yun), father (Jaegeun Kim), and my brother (Steven

Kim) for their encouragement and unyielding support. They had faith in me even in times I did

v

not, and I will not be here today without them. Lastly, I would like to thank my friend, Jackie

Liu, for her support and friendship.

vi

TABLE OF CONTENTS

CHAPTER 1. Introduction…………………………………………………………...…..………..1

1.1 The pathophysiology of bipolar disorder………………………………………….....………..1

1.1.1 Mitochondrial dysfunction and oxidative stress in bipolar disorder…………....………...4

1.1.2 Inflammation in bipolar disorder……………………………………………....…..……...7

1.1.3 Dopamine dysregulation in bipolar disorder………………………………..…....………11

1.2 Lithium…………………………………………………………………………………..…...14

1.3 Downstream effects of complex I dysfunction in bipolar disorder………………….....……17

1.3.1 Complex I dysfunction and the NLRP3-inflammasome in bipolar disorder………..…..18

1.3.2 Complex I dysfunction and the dopamine system in bipolar disorder…………………...20

1.4 Scope and Hypotheses…………………………………………………………………..…...22

1.4.1 Scope…...………………………………………………………………………………...22

1.4.2 Aims and Hypotheses…………………………………………………………………....24

CHAPTER 2. A Fresh Look at Complex I in Microarray Data: Clues to Understanding Disease-

Specific Mitochondrial Alterations in Bipolar Disorder…………………………………….…...27

- Statement of significance and impact …………………………………………………...34

CHAPTER 3. Nod-like Receptor Pyrin Containing 3 in the Post-mortem Frontal Cortex from

Patients with Bipolar Disorder: a Potential Mediator between Mitochondria and Inflammation 37

- Statement of significance and impact……………………………………………………58

CHAPTER 4. Oxidation and Nitration in Dopaminergic Areas Prefrontal Cortex from Patients

with Bipolar Disorder and Schizophrenia………………………………………………………..60

- Statement of significance and impact……………………………………………………86

vii

CHAPTER 5a. Glutathione-mediated Effects of Lithium in Decreasing Protein Oxidation

Induced by Mitochondrial Complex I Dysfunction……………………………………………..88

CHAPTER 5b. Lithium Reduces the Effects of Rotenone-induced Complex I Dysfunction on

DNA Methylation and Hydroxymethylation in Rat Cortical Primary Neurons………………..102

- Statement of significance and impact (Chapter 5a&b)…………………………………127

CHAPTER 6. Summary and General Discussion………………………………………………129

CHAPTER 7. Conclusions……………………………………………………………………..139

References………………………..……………………………………………………………..142

Copyright Releases……………………………………………………………………………..163

viii

LIST OF TABLES

CHAPTER 3

Table 1. Subject information……………………………………………………………………52

CHAPTER 4

Table 1. Subject information by group…………………………………………………………83

ix

LIST OF FIGURES

CHAPTER 2

Figure 1. Summary of mitochondrial complex I gene alterations in bipolar disorder and

schizophrenia……………………………………………………………………………………33

CHAPTER 3

Figure 1. Levels of electron transport chain components complex I, II, III, IV, V and

nicotinamide nucleotide transhydrogenase (NNT) in the post-mortem frontal cortex of patients

with BD, SCZ and non-psychiatric controls (CTL)……………………………………………53

Figure 2. Levels of NLRP3-inflammasome components NLRP3 (A), ASC (B) and caspase 1 (C)

in the brain homogenate (homogenate), cytoplasmic and crude mitochondrial fractions

(mitochondrial) in the post-mortem frontal cortex of patients with BD, SCZ and non-psychiatric

controls (CTL)…………………………………………………………………………………54

Figure 3. Levels of inflammatory factors downstream of NLRP3-inflammasome activation in the

IL-1β pathway, IL-1β, IL-6, TNF-α, and in the IL-18 pathway, IFNγ & IL-10 in the post-mortem

frontal cortex of patients with BD, schizophrenia (SCZ) and non-psychiatric controls

(CTL)………………………………………………………………………………………….55

CHAPTER 4

Figure 1. Immunohistochemistry……………………………………………………………...79

Figure 2. CPM intensity and 3NT intensity…………………………………………………...80

Figure 3. FRET between DAT and CPM, and TH and 3NT…………………………………..81

x

Figure 4.Correlations…………………………………...………………………………………82

CHAPTER 5a

Figure 1. Lithium prevents decrease in complex I activity and cell viability induced by

mitochondrial complex I dysfunction……………………………………………………………97

Figure 2. Glutathione is important for lithium’s ability to decrease protein carbonylation, but not

nitration…..………………………………………………………………………………………98

CHAPTER 5b

Figure 1. Complex I activity measurement in rat cortical primary neurons with lithium pre-

treatment (0.75mM) and rotenone treatment at different concentrations (5nM, 10nM, and

50nM)…………………………………………………………………………………………..116

Figure 2. The effect of lithium pre-treatment (0.75mM) and rotenone treatment at different

concentrations (5nM, 10nM, and 50nM) on ATP levels in primary cortical neurons…………117

Figure 3. The effect of lithium pre-treatment (0.75mM) and rotenone treatment at different

concentrations (5nM, 10nM, and 50nM) on cell mortality in primary cortical neurons measured

by the MTT assay………………………………………………………………………………118

Figure 4. The effect of lithium pre-treatment (0.75mM) and rotenone treatment at different

concentrations (5nM, 10nM, and 50nM) on apoptosis and necrosis in primary cortical

neurons………………………………………………………………………………………….119

Figure 5. Changes in methylation and hydroxymethylation levels with 7 days of lithium pre-

treatment (0.75mM) and 30 minutes of rotenone treatment (5nM, 10nM, and 50nM) in primary

cortical neurons…………………………………………………………………………………120

xi

Figure 6. Changes in methylation and hydroxymethylation levels of 30 minutes of H2O2

treatment (5µM, 10µM or 50µM) in primary cortical neurons………………………………121

Schema 1. Mitochondrial complex I dysfunction induced by rotenone and its effects on cell

viability and DNA methylation and hydroxymethylation: the role of lithium………………122

Online resource 1. Cross-reactivity for antibodies was assayed by adding different combinations

of the primary and secondary antibodies for assessing resultant fluorescence intensity……123

Online resource 2. Changes in methylation and hydroxymethylation levels with 7 days of lithium

pre-treatment followed by rotenone treatment in primary cortical neurons…………………124

Online resource 3. Changes in methylation and hydroxymethylation levels with H2O2 treatment

in primary cortical neurons…………………………………………………………………..124

1

1. INTRODUCTION

Bipolar disorder (BD) is a chronic mood disorder characterized by alternating episodes of mania

and depression. According to WHO, it is the sixth leading cause of disability worldwide, where

approximately 50% of the patients attempt suicide during their lifetime. Complex I dysfunction in

the electron transfer process and consequent generation of oxidative stress, which occurs when the

production of reactive oxygen or nitrogen species surpasses the cell’s capacity to maintain them at

an optimal level, is one of the most consistent findings in patients with BD. Therefore, the primary

aim of my PhD was to examine potential downstream targets of complex I dysfunction in BD.

1.1 The pathophysiology of bipolar disorder

Bipolar disorder is a mood disorder characterized by alternating episodes of mania/hypomania,

depression and/or mixed episodes. Mania, which is a defining feature of BD, is a sustained period in

which patients experience elevated mood and/or irritability, rapid speech, and impulsive behaviors

(Barnett & Smoller, 2009). Patients in the depressive phase exhibit persistent loss of interest in

pleasurable activities and/or depressed mood. Lastly, mixed episodes feature symptoms of both

depression and mania (Anderson, Haddad, & Scott, 2012). There are two subtypes of BD, BD I and

BD II. Patients with BD I exhibit one or more episodes of mania or mixed episodes with at least one

episode of depression, and patients with BD II have one or more episodes of depression with

hypomania (Muller-Oerlinghausen, Berghofer, & Bauer, 2002). In addition to the burden of the

symptoms of the disease, patients may require social welfare due to failure to maintain a job,

intervention from the justice system, and/or assistance from family members (Kilbourne et al.,

2004). BD is also associated with mild to moderate impairments in cognitive function, selective

attention and memory, which are present regardless of the state of the patient (Anderson et al., 2012;

Rajkowska, 2000). BD is also hypothesized to be a progressive disorder, where responsiveness to

therapy and cognitive ability declines with increasing number of recurrences (Kupfer, 2005). Indeed,

2

due to the chronic nature of the disorder, where greater than 90% of patients experience recurrent

episodes (Muller-Oerlinghausen et al., 2002), development of novel and more efficacious treatments

is of great interest.

BD is a highly complex disorder and multiple pathological pathways have been reported in brain

areas associated with emotional control, cognition and memory, such as the prefrontal cortex,

anterior cingulate cortex, hippocampus and the amygdala (Chen, Suckling, Lennox, Ooi, &

Bullmore, 2011). For example, reduced brain volume has been reported in the prefrontal cortex

(Almeida et al., 2009), hippocampus, amygdala, anterior cingulated gyrus and medial and inferior

frontal gyri (Lyoo et al., 2004; Malhi, Tanious, Das, Coulston, & Berk, 2013). In agreement with

these findings, decreased neuronal density was also reported in the prefrontal cortex (Rajkowska,

2000). Some other findings include increased protein and mRNA levels of pro-apoptotic factors

(Rao, Harry, Rapoport, & Kim, 2010), lower levels of brain derived neurotrophic factor (BDNF)

(Cunha et al., 2006), and increased glutamatergic and dopaminergic signaling in patients with BD

(Berk et al., 2007; Messiha, Agallianos, & Clower, 1970; Rao et al., 2010).

Genetic alterations, such as Val66Met polymorphism of the BDNF gene (Lohoff et al., 2005),

and genetic variations at the calcium channel CACNA1C have also been reported in BD (Ferreira et

al., 2008). In addition to genetic polymorphisms, recent studies have focused on epigenetic

modifications in BD, with a strong focus on methylation (Abdolmaleky et al., 2006; Connor &

Akbarian, 2008; D'Addario et al., 2012; Huzayyin et al., 2013). DNA methylation occurs when DNA

methyltransferase transfers a methyl group to the cytosine residue at the 5’ carbon position (Cheng,

Cahill, Kasai, Nishimura, & Loeb, 1992), and methylation of regulatory regions prevents the binding

of transcription factors to the DNA, inhibiting transcription (Bestor, 2000; Bird, 2002;

Lertratanangkoon, Wu, Savaraj, & Thomas, 1997; Reik, Dean, & Walter, 2001). Methylation is also

the most common epigenetic alteration in mammals, and is known to be affected by a large number

3

of different factors, such as nicotine, environmental toxins, and psychostimulants (Bird, 2002;

Connor & Akbarian, 2008; Reik et al., 2001). Alterations in methylation levels were reported in

patients with BD, where BDNF promoter region was found to be hypermethylated (D'Addario et al.,

2012), and promoter of the catechol-o-methyltransferase gene was found to be hypomethylated

(Abdolmaleky et al., 2006). A related alteration is hydroxylation of methylcytosine by the ten eleven

translocation enzyme (Tahiliani et al., 2009). Formation of 5-hydroxymethylcytosine is thought to

promote the demethylation of DNA, and also to inhibit the binding of 5-methylctyosine recognizing

proteins, limiting their ability to repress transcription (Jin, Kadam, & Pfeifer, 2010). The ten eleven

translocation enzyme requires molecular oxygen (Jin et al., 2010). Since the mitochondria are

primary consumers of oxygen in the cell, mitochondrial dysfunction may also contribute to different

hydroxymethylation patterns. Moreover, studies have shown that reactive oxygen species can cause

various DNA damage including both hypermethylation and hypomethylation using cancer cells

(Chestnut et al., 2011; H. J. Kang et al., 2012; Kang, Zhang, Kim, Bae, & Hyun, 2012), implicating

methylation and hydroxymethylation as possible downstream targets of mitochondrial dysfunction

and oxidative stress.

Indeed, mitochondrial complex I dysfunction has been strongly implicated in BD (Iwamoto,

Bundo, & Kato, 2005; Kato, 2005; Kato & Kato, 2000; Konradi, Sillivan, & Clay, 2012). Complex I

is involved in many different pathways in the brain, including the production of reactive oxygen and

nitrogen species (Halliwell, 2007). Production of these molecules can have multiple downstream

effects, as they can cause direct modifications to transporters and enzymes, changing their function,

and alter levels of genetic expression as mentioned above (Campos et al., 2007; K. A. Kang et al.,

2012), or act as signaling molecules to affect downstream pathways (Halliwell & Gutteridge, 2007;

H. K. Kim, Andreazza, Yeung, Isaacs-Trepanier, & Young, 2014). The following sections will

begin with a discussion of the role of mitochondrial complex I dysfunction in BD. I will then discuss

4

two potential downstream targets of complex I dysfunction in BD: dopaminergic dysfunction and

inflammation. Lithium, which is uniquely used for the treatment of BD, is also discussed as it

improves complex I activity and decrease oxidative stress, making it a good proxy for understanding

complex I dysfunction in BD.

1.1.1 Mitochondrial dysfunction and oxidative stress in bipolar disorder

Mitochondria are primarily responsible for oxidative phosphorylation, which produces adenosine

triphosphate (ATP) by relaying electrons extracted from nicotinamide adenine dinucleotide (NADH)

and flavin adenine dinucleotide (FADH2) through the electron transport chain (ETC) consisting of 5

multi-protein complexes located at the inner mitochondrial membrane. The process of electron

transfer is coupled to proton pumping into the intermembrane space of the mitochondria. This

process creates a proton gradient, which is used by the ATP-synthase (complex V) to phosphorylate

ATP. Molecular oxygen is the final acceptor of electrons, becoming water (Fariss, Chan, Patel, Van

Houten, & Orrenius, 2005). Studies have reported lower levels of high energy phosphates in the

brain of patients with BD since the 1990s using magnetic resonance spectroscopy (Deicken, Weiner,

& Fein, 1995; Kato et al., 1998; Kato, Takahashi, Shioiri, & Inubushi, 1993), suggesting decreased

efficiency of the mitochondria. Patients with BD were also found to have lower mRNA levels of

genes encoding mitochondrial proteins related to oxidative phosphorylation, suggesting simpairment

in this process (Konradi et al., 2004; Sun, Wang, Tseng, & Young, 2006). Decreased efficiency of

the ETC can compromise the process of electron transfer, resulting in increased production of

superoxide anions from the reaction between leaked electrons and oxygen (Fariss et al., 2005).

Complex I of the ETC consists of 45 or 46 subunits arranged into 2 functional arms, which are

the hydrophilic arm and the hydrophobic arm. The hydrophilic arm is largely responsible for the

transfer of electrons and contains subunits containing iron-sulfur clusters, while the hydrophobic arm

is responsible for proton pumping (Green & Kroemer, 2004; Lenaz, 2001). Interestingly, patients

5

with BD were found to have lower mRNA and protein levels of NADH dehydrogenase Fe-S protein

7 (NDUFS7), which is a subunit of complex I that contains the last iron-sulfur cluster N2 responsible

for the reduction of ubiquinone to ubiquinol (Andreazza, Shao, Wang, & Young, 2010), and levels

of this subunit were found to correlate with complex I activity (Andreazza et al., 2010). Because

complex I is the main site of electron leakage (Halliwell & Gutteridge, 2007; Jeong & Seol, 2008;

Sherer et al., 2003), these studies suggest that patients with BD may be more susceptible to the

production of superoxide anions. This is important because while products of oxidative stress are

found in both schizophrenia and BD (Ng, Berk, Dean, & Bush, 2008), the contributing sources may

be different for the two disorders. More specifically, while complex I dysfunction may be a major

contributor to oxidative stress production BD, other pathways may be involved for schizophrenia

(Ng et al., 2008).

Superoxide anions produced during this process is converted to hydrogen peroxide by the

superoxide dismutase. Hydrogen peroxide, which is relatively unreactive, can go through the Fenton

reaction in the presence of iron ions to produce hydroxyl radicals, which are highly reactive

(Halliwell, 1992). Indeed, hydroxyl radicals can cause lipid peroxidation, converting membrane

lipids to lipid hydroperoxides and releasing reactive unsaturated aldehydes (Halliwell, 1992). It can

also oxidize DNA, producing 8-hydroxydeoxyguanosine (8-OHdG), and oxidize proteins to cause

disulfide bridge formation or carbonylation (Halliwell, 2001). Superoxide anion can also react with

nitric oxide to produce peroxynitrite, a reactive nitrogen species (RNS) (Halliwell & Gutteridge,

1984; Jeong & Seol, 2008). Peroxynitrite can act similarly to the hydroxyl radical, oxidizing protein

sulfhydryls, lipids, and DNA (LaVoie & Hastings, 1999). On the other hand, peroxynitrite can also

produce a different type of oxidative damage called nitration, where it reacts with tyrosine residues

to produce 3-nitrotyrosine (Halliwell & Gutteridge, 2007). The central nervous system may be

particularly vulnerable to oxidative damage due to the richness of polyunsaturated fatty acid side

6

chains that are more sensitive to attack by free radicals, lower levels of antioxidant enzymes such as

superoxide dismutase and glutathione peroxidase, high oxygen consumption and high concentrations

of iron in certain brain areas (Halliwell, 1992).

Production of ROS and RNS is neutralized by antioxidant enzymes, including the

aforementioned superoxide dismutase, which converts the superoxide anion to hydrogen peroxide.

Hydrogen peroxide is then converted to water by catalase (Halliwell & Gutteridge, 2007). The main

antioxidant system in the brain is glutathione, which can neutralize both ROS and RNS by using its

sulfhydryl group (Gawryluk, Wang, Andreazza, Shao, & Young, 2011). Glutathione peroxidase

catalyzes the reaction between hydrogen peroxide and glutathione to produce water, and glutathione-

S-transferase conjugates glutathione to oxidized products, making them non-toxic (Dringen, 2000).

Oxidized glutathione groups are returned to their reduced form by glutathione reductase (Yao &

Keshavan, 2011). Oxidative stress can also occur due to the reduction in antioxidant defenses. Lower

levels of antioxidant enzymes including glutathione peroxidase, superoxide dismutase, catalase,

glutathione S-transferase and glutathione (Benes, Matzilevich, Burke, & Walsh, 2006) were found in

the post-mortem brain of patients with BD. Indeed, N-acetylcysteine, a cysteine derivative that

increases glutathione production, was found to decrease depression and improve global functioning

in patients with BD compared to placebo in a double-blind randomized study (Berk et al., 2008;

Magalhães et al., 2011).

While complex I dysfunction and increased oxidative stress are well-established in BD

(Andreazza et al., 2010; Andreazza, Wang, Salmasi, Shao, & Young, 2013; Scola, Kim, Young, &

Andreazza, 2013; Sun et al., 2006), the specific downstream targets have yet to be elucidated. To

answer this question, a recent study published from our group identified the mitochondria and the

synapse as two targets of oxidative stress in the frontal cortex of patients with BD (Andreazza et al.,

2013). To extend on these findings, the first part of my PhD aimed to examine the downstream

7

targets of complex I dysfunction in BD with a focus on the mitochondria and the synapse in the post-

mortem brain of patients with BD. In choosing a marker of mitochondrial oxidative damage, we

decided to examine the nod-like receptor, pyrin containing 3 (NLRP3)-inflammasome, a redox

sensor in the inflammatory system that was found to be downstream of mitochondrial ROS

production from complex I dysfunction in vitro (Zhou, Yazdi, Menu, & Tschopp, 2011). For a

marker of synaptic oxidative damage, we chose to examine oxidative modifications to the dopamine

system, which exists largely in the synapse and is known to underlie mania, a defining feature of BD

(Berk et al., 2007; Gerner, Post, & Bunney, 1976). The following sections will provide a review of

findings supporting inflammatory and dopaminergic abnormalities in BD.

1.1.2 Inflammation and bipolar disorder

Medical conditions associated with the activation of the inflammatory system are more

commonly diagnosed in patients with BD than in the normal population, particularly diabetes

mellitus, cardiovascular disease and autoimmune diseases (Barbosa et al., 2013). Chronic pain

conditions such as arthritis and headache are also more prevalent in BD (Goldstein, Fagiolini,

Houck, & Kupfer, 2009; Goldstein, Kemp, Soczynska, & McIntyre, 2009; Kilbourne et al., 2004),

which is in agreement with studies reporting a strong association between pain and inflammatory

cytokines IL-1β, IL-6, and TNF-α (J. Zhang et al., 2015). Patients with BD also show higher

monocyte phagocytic activity (Padmos et al., 2008). The first study demonstrating increased levels

of inflammatory cytokines in patients with BD was published in 1995 (Maes, Bosmans, Calabrese,

Smith, & Meltzer, 1995). Since then, an accumulating number of studies have demonstrated mild

and chronic activation of the inflammatory system in the brain and periphery of patients with BD

(Goldstein, Kemp, et al., 2009; Hope et al., 2011; Kauer-Sant'Anna et al., 2009; Leboyer et al., 2012;

Liu et al., 2004; Rao et al., 2010). Indeed, even after controlling for various demographic factors

such as age, sex, and even clinical variables, cytokine levels were found to be strongly associated

8

with BD (Goldstein, Kemp, et al., 2009). While the role of the inflammatory system in BD is

unclear, Barbosa et al (2014)’s hypotheses for increased monocyte activity in BD can also be applied

to the general inflammatory system (Barbosa et al., 2014). More specifically, presence of

inflammatory activation in BD can be explained by 1. The illness itself induces activation of the

inflammatory system, 2. Inflammatory activation may contribute to the development of BD, and/or

3. A factor may be responsible for both the development of BD and inflammatory conditions.

Studies examining genetic alterations in the inflammatory system reported polymorphisms in the

IL1B gene in patients with BD, and greater levels of haplotypic combinations in the IL1RA gene,

particularly in BD patients with a family history of BD, schizophrenia, or major depression (Papiol

et al., 2008). Moreover, expression levels of 19 inflammatory mRNAs were found to be altered, with

IL-6 being one of the strongest markers separating patients with BD from healthy controls (Padmos

et al., 2008). Interestingly, some studies reported increased activation of the inflammatory system

during mood episodes, where patients with mania were found to have greater levels of IL-6, TNF-α,

and soluble IL-2 receptor compared to patients with euthymia as well as controls (Barbosa et al.,

2013; Modabbernia, Taslimi, Brietzke, & Ashrafi, 2013; Munkholm, Braüner, Kessing, & Vinberg,

2013). However, findings are inconsistent and studies examining associations between treatment and

cytokine levels, as well as inflammation and symptom severity have largely reported negative

findings, suggesting that inflammatory activation may be a trait marker rather than a state marker in

BD (Goldstein, Kemp, et al., 2009). Another topic that has been garnering much attention is the

interaction between inflammatory cytokines in the periphery and in the central nervous system, as

inflammatory cytokines are larger and cannot easily pass through the blood-brain barrier (Barbosa et

al., 2013). Some methods of transport under physiological conditions include cytokine receptors,

endothelial activation and active transport (Barbosa et al., 2013). Cytokines can also pass through

the blood brain barrier when the blood brain barrier becomes leaky, which occurs with increased

9

inflammation in the brain, suggesting that neuroinflammation can exacerbate cytokine exchange

between the central nervous system and the periphery (Brietzke, Stabellini, Grassis-Oliveira, &

Lafer, 2011). Interestingly, cytokine activation profile differs between the central nervous system

and the periphery in BD, with increased levels of cytokines in the IL-2, TNF-α and IL-6 pathways in

the periphery, suggesting the activation of T cells (Kauer-Sant'Anna et al., 2009; Muller-

Oerlinghausen et al., 2002; Munkholm et al., 2013; O'Brien, Scully, Scott, & Dinan, 2006; Ortiz-

Domínguez et al., 2007; Rao et al., 2010), and greater levels of cytokines and receptors in the IL-1

pathway, including IL-1β, IL-1 receptor, NF-κB, glial fibrillary acidic protein, and cFOS in the

prefrontal cortex and cerebrospinal fluid of patients with BD (Rao et al., 2010; Söderlund et al.,

2011). Moreover, treatment with medications that are used in BD such as lithium and antipsychotics

was found to decrease levels of inflammatory cytokines, including IL-10, IL-6, IL-2 and IFN-γ,

suggesting that amelioration of inflammatory activation may partly contribute to their effect in BD

(Boufidou, Nikolaou, Alevizos, Liappas, & Christodoulou, 2004; Padmos et al., 2008).

Activation of the inflammatory system may have important implications for BD as inflammatory

cytokines can influence long-term potentiation, neuronal survival, astrocyte development, and

neurogenesis (Barbosa et al., 2013). Furthermore, inflammatory cytokines can contribute to changes

in neurotransmitter levels such as dopamine and cause cognitive decline, possibly contributing to the

behavioral symptoms of BD (Berk et al., 2013; Felger et al., 2013; Felger & Miller, 2012). As such,

anti-inflammatory agents have received attention as potential treatments for BD (Barbosa et al.,

2013). Celecoxib, a cyclooxygenase-2 inhibitor, was tested on patients with BD with mixed or

depressive episodes in a double-blind, randomized and placebo-controlled study as an adjunct (Nery

et al., 2008). The findings of this study revealed that patients treated with celecoxib experienced

faster reduction in their depressive symptoms compared to the control group.

10

The cause for inflammatory activation in BD appears to be multifactorial, with a few studies

demonstrating genetic alterations, and others suggesting that it may be occurring downstream of

other changes that occur in BD, such as excitatoxicity, apoptotic signaling, or activation of the

hypothalamic-pituitary-adrenal axis (Goldstein, Kemp, et al., 2009; Leboyer et al., 2012). Recent

studies have shown that mitochondrial complex I dysfunction and oxidative stress may contribute to

activation of the inflammatory system (López-Armada, Riveiro-Naveira, Vaamonde-García, &

Valcárcel-Ares, 2013; Zhou et al., 2011), suggesting that complex I dysfunction in BD may be a

contributing factor to inflammation as well. More specifically, signs of mitochondrial dysfunction,

such as release of mitochondrial DNA or ATP through the opening of permeability transition pores,

were found to correlate with increased levels of IL-1β, and inhibition of complex I using chemicals

such as rotenone was found to have the same effect (López-Armada et al., 2013; Zhou et al., 2011).

Indeed, drugs that decrease oxidative stress, such as N-acetyl cysteine were found to decrease the

release of inflammatory cytokines (Zhou et al., 2011). Lithium, which is the gold-standard of

treatment for the management of BD, was also shown to improve complex I activity, decrease

oxidative stress, and decrease pro-inflammatory cytokines (Burgess et al., 2001; Chuang, 2005;

Goldstein, Kemp, et al., 2009; Quiroz, Gould, & Manji, 2004), further suggesting the importance of

these pathways in BD. However, whether complex I dysfunction and subsequent release of

mitochondrial ROS underlies increased inflammation in BD has yet to be explored. Specific

mechanisms that may be involved in the link between complex I defect and inflammation in BD will

be further discussed in section 1.3.1, which examines potential downstream targets of complex I

dysfunction in BD.

In addition to activating the inflammatory system through generating mitochondrial oxidative

stress, complex I dysfunction can have other downstream effects (Andreazza et al., 2010; Naoi et al.,

2005; Watabe & Nakaki, 2007). As a continuation of a previous study by our group identifying the

11

synapse as a target of oxidative stress in BD (Andreazza et al., 2013), we decided to examine the

dopaminergic synapse as a potential downstream target of complex I dysfunction, since increased

dopamine signaling is strongly implicated in mania (Berk et al., 2007; Cousins, Butts, & Young,

2009). The next section will provide an overview of dopaminergic dysregulation in BD.

1.1.3 Dopamine dysregulation in bipolar disorder

Much attention has been given to the dopaminergic synapse as a site of damage in BD as

increased levels of dopamine is widely accepted to be one of the main mechanisms underlying mania

(Berk et al., 2007; Manji & Lenox, 2000; Salvadore et al., 2010). Dopamine is a neurotransmitter

that is involved in locomotion, reward, emotion and cognition (Giros et al., 1992) through four

pathways: the nigrostiratal pathway, which is involved in the motor system and includes substantia

nigra pars compacta and the striatum, the tuberinfundibular pathway, which is in the hypothalamus,

the mesolimbic pathway, which begins at the ventral tegmental area and project to the ventral

striatum, hippocampus, and septum, and finally the mescortical pathway, which also originates from

the ventral tegmental area and project to the frontal and temporal cortices (Ciliax et al., 1999;

Cousins et al., 2009). The mesocortical pathway has received much attention in BD as it contributes

to working memory, reward, impulsivity and emotional processing, characteristics often reported to

be altered in BD (Seamans & Yang, 2004).

Dopamine is regulated by many different enzymes, including tyrosine hydroxylase, which is the

rate-limiting enzyme for dopamine synthesis (Daubner, Le, & Wang, 2011). As the concentration of

its precursor, tyrosine, does not affect the rate in which dopamine is synthesized, the proper

functioning of tyrosine hydroxylase is critical for the dopamine system (Cousins et al., 2009). L-

DOPA produced by tyrosine hydroxylase is converted to dopamine by aromatic amino acid

decarboxylase. Dopamine’s actions are primarily terminated by the dopamine transporter, which

uptakes dopamine from the synapse (Cousins et al., 2009; Jaber, Jones, Giros, & Caron, 1997),

12

although it can also be oxidized by catechol-o-methyl transferase into homovanilic acid, or auto-

oxidized (Cousins et al., 2009). Once released into the synapse, dopamine exerts its effects through

two different families of dopamine receptors, D1 and D2. Activation of receptors of the D1 family,

consisting of D1 and D5, result in activation of the adenylyl cyclase pathway through its coupling

with Gα. Activation of D2 receptors, which consist of D2, D3 and D4, inhibits adenylyl cyclase

through Gαi/o (Elsworth & Roth, 1997).

First report of greater dopamine levels in patients with BD was in 1970, when Messiha et al.

reported higher levels of conjugated and free dopamine in the urine of patients in the manic phase

(Messiha et al., 1970). Administration of dopaminergic agonists such as amphetamine, which

reverses the direction of dopamine transport such that synaptic dopamine levels increase, and

pramipexole and bromocriptine (Cousins et al., 2009; Silverstone, 1984) leads to manic behavior as

measured by the Young Mania Scale in healthy individuals and in patients with BD (Anand et al.,

2000; Berk et al., 2007). The same is observed when L-DOPA, a dopamine precursor, is

administered (Cousins et al., 2009). Behavioral effects of dopaminergic agonists can be reversed by

medications used in BD, such as lithium, adding support to the dopamine theory of BD (Cousins et

al., 2009; Van Kammen & Murphy, 1975). Moreover, antagonizing the dopamine system with

antipsychotics and alpha-methyl-para-tyrosine, an inhibitor of dopamine synthesis, reduces manic

behavior (Berk et al., 2007; Cousins et al., 2009). Some studies also suggest that bipolar depression

may represent hypodopaminergic states, as bupropion, which inhibits dopamine removal from the

synapse, is effective in treating depression (Berk et al., 2007), and levels of homovanillic acid, a

metabolite of dopamine, in the cerebrospinal fluid of patients with BD in the depressive phase is

lower after drug washout (Cousins et al., 2009). Moreover, lower dopamine levels may be

underlying depression in Parkinson’s disease, which can be ameliorated by administering

dopaminergic agonists (Goldberg, Burdick, & Endick, 2004). The presence of cognitive dysfunction

13

and impaired working memory also suggests dopamine dysfunction in BD, as optimal levels of

dopamine is necessary for working memory, where too much or too little dopamine is thought to

impair performance (Berk et al., 2011). Indeed, both patients with mania or depression exhibit

lowered sustained attention, and lower scores on tests for executive function and working memory

(Clay, Sillivan, & Konradi, 2010; Cousins et al., 2009; H. W. Kim, Rapoport, & Rao, 2010;

Martinez-Aran et al., 2004). These findings lead to the dopamine dysregulation model of BD, which

states that increased dopamine transmission that underlies mania is followed by a subsequent

downregulation of elements of the dopamine system, resulting in the depressive phase (Berk et al.,

2007).

First-line treatment for acute mania is antipsychotics, which act by blocking dopamine D2

receptors (Anderson et al., 2012), prompting scientists to examine dopamine receptor alterations in

BD. Studies examining dopamine receptors by positron emission tomography have reported

inconsistent findings, with one study reporting alterations in D1-receptor binding (Suhara et al.,

1992), increased D2 receptor density (Pearlson et al., 1995; Wong et al., 1997), and other studies

reporting no alterations in D2 receptor density between BD and controls (Pearlson et al., 1995).

Differences in the results may be due to the different areas examined or the radioligand used

(Cousins et al., 2009). Genetic studies have also examined dopamine receptors, in which

predominantly negative results were reported for the D1 receptor (Mitchell et al., 1992; Nöthen et

al., 1992), D2 receptor (Cousins et al., 2009; Furlong et al., 1998; Nöthen et al., 1992), the D3

receptor (Elvidge et al., 2001; Savoye et al., 1998), D4 (Serretti et al., 1999) and the D5 receptor

genes (Muir et al., 2001). Coincidentally, genetic studies examining catechol-o-methyl transferase

also reported inconsistent findings (Craddock & Sklar, 2013), and mRNA levels of this enzyme was

not found to differ in the prefrontal cortex of patients with BD compared to controls (Tunbridge,

Burnet, Sodhi, & Harrison, 2004). As genetic studies suggest that dopaminergic abnormality in BD

14

may not occur due to alterations in gene expression, downstream pathways may be involved, such as

post-translational alterations. In fact, dopaminergic proteins are susceptible to oxidative

modifications due to their abundance in cysteine residues (Hastings, Lewis, & Zigmond, 1996;

Hastings & Zigmond, 1994), suggesting that complex I dysfunction and oxidative stress may

contribute to dopamine dysregulation in BD.

While antipsychotics are best recognized as decreasing dopamine signaling in mania, lithium

was also shown to decrease dopamine levels and increase dopamine uptake (Dunigan & Shamoo,

1995; Messiha et al., 1970). In addition to decreasing dopamine signaling, lithium also has a wide

range of effects, including decreasing glutamate levels, stabilizing calcium levels, decreasing

apoptosis, pro-inflammatory cytokines, and importantly, increasing complex I activity (Bearden et

al., 2007; Cordeiro, Gundersen, & Umbach, 2002; Cui, Shao, Young, & Wang, 2007; J. S. Lai,

Zhao, Warsh, & Li, 2006; Nascimento et al., 2014; Quiroz et al., 2004; Schäfer, Goodenough,

Moosmann, & Behl, 2004; Tan et al., 2011). Complex I dysfunction and generation of oxidative

stress can impact many systems due to the ability of ROS and RNS to directly modify proteins and

impact downstream pathways, suggesting that at least a part of lithium’s actions may be by

improving complex I function (Nascimento et al., 2014; Sun et al., 2006). Additionally, lithium is

uniquely used for the treatment of BD, making lithium a good proxy for studying complex I

dysfunction in BD. The next section further examines the use of lithium and its mechanisms of

action in BD.

1.2 Lithium

Lithium has been the first-line treatment for long-term management of BD since being

introduced 60 years ago, and is used uniquely for the treatment of BD (Malhi et al., 2013; Muller-

Oerlinghausen et al., 2002). While it is most recognized for its ability to stabilize mood and prevent

suicidal ideation during euthymia (Bauer, 2004; Burgess et al., 2001; Malhi et al., 2013), lithium is

15

also prescribed for acute mania, although its efficacy is limited as it takes 6-10 days to exert its

effects. Hence, lithium is often prescribed with antipsychotics (Malhi et al., 2013). Evidence

supporting lithium’s efficacy for bipolar depression is less consistent (Bhagwagar & Goodwin, 2002;

Fountoulakis, Gonda, Siamouli, & Rihmer, 2009), and recent clinical trials did not show differences

between lithium and placebo (Malhi et al., 2013). At a global level, patients who were prescribed

lithium were reported to have greater total gray matter volume compared to patients who were not

prescribed lithium and even healthy controls in areas implicated in BD, such as the anterior

cingulate, frontal cortex, amygdala, and hippocampus (Bearden et al., 2007; Hallahan et al., 2011;

Kempton, Geddes, Ettinger, Williams, & Grasby, 2008). Studies have also shown that lithium

attenuates dopamine release (Ferrie, Young, & McQuade, 2006) and decrease glutamate

transmission while increasing GABA release (Malhi et al., 2013), suggesting that it acts by

preventing alterations in neurotransmission. More specifically, lithium treated rats had lower levels

of dopamine (Gambarana et al., 1999), and lithium was shown to cause downregulation of the

NMDA receptor and increase glutamate reuptake (Dixon & Hokin, 1998), which could together be

contributing to decreasing manic symptoms. Furthermore, lithium was found to increase GABA

levels in the cerebrospinal fluid of patients (Ahluwalia, Grewaal, & Singhal, 1981), decreasing

glutamate-induced excitatoxicity.

Lithium was also shown to impact many signaling pathways through its ability to interfere with

magnesium binding (Malhi et al., 2013). For example, lithium inhibits inositol monophosphatase (Q.

Li et al., 2010; Malhi et al., 2013) and adenylyl cyclase (Quiroz et al., 2004), and minimize cAMP

fluctuations from stimulation, which may contribute to its ability to stabilize neurotransmission

(Bachmann, Schloesser, Gould, & Manji, 2005; El Khoury, Petterson, Kallner, Aberg-Wistedt, &

Stain-Malmgren, 2002; Quiroz et al., 2004). Lithium’s ability to regulate calcium levels are also well

documented, where it was shown to block calcium uptake, decrease calcium stores and its

16

intracellular levels (Dubovsky, Thomas, Hijazi, & Murphy, 1994; El Khoury et al., 2002; Perova,

Kwan, Li, & Warsh, 2010; Shalbuyeva, Brustovetsky, & Brustovetsky, 2007). Since calcium is an

important signaling molecule for many cellular processes, lithium’s ability to stabilize calcium levels

most likely plays an important role (Malhi et al., 2013). Lithium’s inhibition of glycogen synthase

kinase 3β is of interest, as it is a target of monoaminergic systems and produce hyperactivity in mice

(Prickaerts et al., 2006).

In addition, lithium is known to have neuroprotective effects, as it was shown to decrease pro-

apoptotic proteins, including Bax and p53, while increasing anti-apoptotic proteins, such bcl-2 (Q. Li

et al., 2010; Malhi et al., 2013). Lithium was also found to increase BDNF levels after 5 days of

treatment, which is very similar to how long lithium takes to have an effect in patients (6-10 days),

strongly implicating BDNF in lithium’s neuroprotective effects (Hashimoto et al., 2002).

Importantly, lithium was shown to protect against oxidative stress (Cui et al., 2007; Machado-Vieira,

Andreazza, et al., 2007; Shao, Cui, Young, & Wang, 2008; Sun et al., 2006). For example, lithium

was able to inhibit lipid peroxidation and protein oxidation in primary rat cortical cells (Frey et al.,

2006). Moreover, lithium was able to prevent other downstream effects of oxidative stress produced

by complex I inhibition or hydrogen peroxide, such as release of cytochrome c in hippocampal and

neuroblastoma cells (Cui et al., 2007; J. S. Lai et al., 2006). In addition, lithium was shown to

increase the activity of superoxide dismutase and glutathione peroxidase in the brain of rats, and

normalize superoxide dismutase to catalase ratios (Machado-Vieira, Andreazza, et al., 2007;

Machado-Vieira, Dietrich, et al., 2007), further demonstrating its ability to increase antioxidant

defenses. Lithium was also found to inhibit amphetamine-induced lipid peroxidation in rats,

suggesting that lithium can counteract oxidative stress produced by high levels of dopamine as well

(Frey et al., 2006). While the exact mechanism by which lithium exerts its antioxidant effects is

unknown, studies have demonstrated that lithium increases glutathione content (Shao et al., 2008),

17

expression levels of glutamate-cysteine ligase, and increase expression levels of glutathione S-

transferase (Shao et al., 2008). A number of studies have also shown that lithium improves complex

I function, which would decrease mitochondrial production of ROS (Bachmann et al., 2009; Quiroz,

Gray, Kato, & Manji, 2008). Also, patients with BD who were taking lithium were found to have

higher mRNA levels of complex I components compared to patients who were not (Sun et al., 2006).

These studies suggest that lithium may be able to prevent downstream effects of complex I

dysfunction and oxidative stress in BD.

1.3 Downstream effects complex I dysfunction in bipolar disorder

Complex I dysfunction in the electron transfer process can lead to increased leakage of electrons,

which can react with molecular oxygen to produce the superoxide anion (Halliwell & Gutteridge,

2007; Sherer et al., 2003). The superoxide anion can undergo a cascade of reactions to produce ROS

and RNS that can accumulate and lead to oxidative stress (Halliwell, 2007). Oxidative stress can

have many downstream effects as ROS and RNS can directly modify biomolecules such as proteins,

lipids and DNA, and also act as signaling molecules (Halliwell, 1992, 2006; Sies, 1991). Protein

oxidation and nitration products, such as 3-nitrotyrosine (Andreazza et al., 2009; Andreazza et al.,

2010), protein carbonyl groups (Andreazza et al., 2010) and disulfide bridges (H. K. Kim et al.,

2014) have also been found to be increased in the post-mortem brain of patients with BD,

particularly in the frontal cortex. Protein oxidation and nitration can produce aggregates and direct

alteration of protein function (Beal, 2002). Moreover, oxidative modifications can accumulate in

sufficient amounts to disrupt essential cellular processes when the production of oxidative stress

surpasses the cell’s antioxidant defense system (Halliwell, 2001), resulting in apoptosis or necrosis

depending on the degree of severity (Ng et al., 2008). Evidence of increased lipid peroxidation is one

of the most consistent findings in BD (Brown, Andreazza, & Young, 2014), where lipid peroxidation

products such as 4-hydroxynonenal, malondialdehyde, and thiobarbituric acid reactive substances

18

were reported in the brain and periphery of patients with BD (Andreazza et al., 2008; Brown et al.,

2014). Because products of lipid peroxidation can disturb cellular function by forming adducts,

increased levels of these products have functional implications (Kuloglu et al., 2002). RNA and

DNA are also targets of oxidative stress, as guanine residues are sensitive to oxidative attack, and

form 8-hydroxydeoxyguanosine in DNA (Barzilai & Yamamoto, 2004; Cheng et al., 1992).

Oxidative damage to DNA was also found to be increased in patients with BD, as higher 8-OHdG

levels were found in lymphocytes and the post-mortem hippocampus (Che, Wang, Shao, & Young,

2010; Soeiro-de-Souza et al., 2013).

Interestingly, a recent study by our group showed that complex I dysfunction and oxidative stress

can specifically target different subcellular fractions by examining the post-mortem frontal cortex of

patients with BD (Andreazza et al., 2013), suggesting that ROS and RNS may have specific

downstream targets rather than affecting all systems in the same direction. More specifically, the

myelin, synapse and mitochondrial fractions were found to be targets of oxidative modifications. In

the myelin fraction, 4-hydroxynonenal and 8-isoprostane levels were increased in patients with BD,

in agreement with previous studies supporting increased lipid peroxidation in patients with BD

(Andreazza et al., 2008; Brown et al., 2014; Wang, Shao, Sun, & Young, 2009). Lipid peroxidation

in patients with BD has been extensively studied by other members from our group, demonstrating

lipid peroxidation as one of the most consistent findings in BD and that it may contribute to

decreased white matter integrity (Andreazza et al., 2008; Brown et al., 2014; Versace et al., 2013).

These findings suggest that lipid peroxidation in the myelin fraction may be a downstream target of

complex I dysfunction in BD. Therefore, my PhD project aimed to extent on these findings and

examine specific downstream targets of complex I dysfunction in BD in the mitochondria and the

synapse. We first examined the mitochondria by using a marker of complex I dysfunction, NLRP3.

Not only does NLRP3 allow us to examine mitochondrial release of ROS through complex I defect,

19

its activation also causes the release of pro-inflammatory cytokines (Zhou et al., 2011). Increased

levels of inflammatory cytokines are consistently reported in the brain and peripheral samples of

patients with BD (Goldstein, Kemp, et al., 2009; Rao et al., 2010), making NLRP3 of particular

interest for the examination of complex I dysfunction in BD.

1.3.1 Complex I dysfunction and the NLRP3-inflammasome in bipolar disorder

A recent study demonstrated that complex I dysfunction and subsequent production of

mitochondrial ROS can cause the activation of a redox sensor, NLRP3 (Sorbara & Girardin, 2011;

Tschopp & Schroder, 2010; Zhou et al., 2011), suggesting that it can be used as a marker of

mitochondrial oxidative stress. More specifically, NLRP3 is a redox sensor in the inflammatory

system (Zhou et al., 2011) and a member of the NOD-like receptor family that are responsible for

activating IL-1β and IL-18 upon sensing a variety of different molecules including

lipopolysaccharide, ATP, and monosodium urate (Agostini et al., 2004; Bryant & Fitzgerald, 2009).

As it is highly unlikely that the ligand binding site of NLRP3 is flexible enough to allow it to

recognize such a large variety of triggers, it has been hypothesized that a common pathway must

exist for these molecules to activate NLRP3 (Schroder, Zhou, & Tschopp, 2010; Tschopp &

Schroder, 2010). NLRP3 is also activated by markers of mitochondrial oxidative damage, such as

opening of mitochondrial permeability transition pores and release of mitochondrial DNA and ATP,

demonstrating NLRP3 is a marker for mitochondrial oxidative stress (Schroder et al., 2010; Shimada

et al., 2012; Tschopp & Schroder, 2010; Zhong et al., 2013; Zhou et al., 2011). Indeed, superoxide

anions produced by complex I dysfunction can act as signaling molecules through redox sensors in

the cell that trigger downstream pathways by undergoing structural modifications upon reacting with

ROS or RNS (Jones, 2010). Studies have shown that treatment with rotenone, a complex I inhibitor,

causes activation of NLRP3 and that molecular triggers fail to activate NLRP3 when mitochondrial

ROS release is inhibited (N. Li et al., 2003; Zhou et al., 2011), suggesting that complex I

20

dysfunction and consequent release of superoxide anions may be the common pathway (Zhou et al.,

2011).

NLRP3 consists of three domains, a pyrin domain, a C-terminal leucine-rich domain, and a

central nucleotide binding domain (Agostini et al., 2004; Bryant & Fitzgerald, 2009). In its resting

state, NLRP3 is localized in the cytoplasm with its central nucleotide binding domain bound to the

leucine rich domain, which inhibits it from oligomerizing (Tschopp & Schroder, 2010). Upon its

activation, NLRP3 undergoes a conformational change exposing its oligomerization and pyrin

domain, allowing it to recruit another cytoplasmic protein, apoptosis-associated speck-like protein

containing a CARD (ASC) through pyrin-pyrin interaction (Zhou et al., 2011). The CARD domain

of ASC allows for the recruitment of procaspase-1, which completes the assembly of the NLRP3-

inflammasome. It was also found that inflammasomes co-localize with the mitochondrial membrane

using co-localization analysis and subcellular fractionation (Zhou et al., 2011). Assembly of the

inflammasome allows for the release of active caspase 1, which then cleaves IL-1β and IL-18 to their

active forms, triggering the inflammatory cascade. These pathways lead to the activation of MYD88

and NF-κB, resulting in increased expression of IL-6, TNFα, and prostaglandin E2, which have all

been shown to be increased in patients with BD (Eder, 2009; Sollberger, Strittmatter, Garstkiewicz,

Sand, & Beer, 2014).

The above studies suggest that complex I dysfunction may lead to the activation of the NLRP3-

inflammasome through the generation of mitochondrial ROS, making this system a marker of

mitochondrial oxidative damage. The aforementioned study has found the mitochondrial and the

synapse as targets of oxidative damage. While the synapse is a highly complex system consisting of

numerous transporters, enzymes, and receptors, we decided to focus on the dopamine system, as

increased activity of the dopamine system is known to underlie mania, a defining feature of BD.

1.3.2 Complex I dysfunction and the dopamine system in bipolar disorder

21

In vitro and animal studies have demonstrated that dopamine synthesis and transport are

influenced by oxidative stress and complex I inhibition, suggesting that the dopamine system may

also be a target of complex I dysfunction in BD (H. K. Kim & Andreazza, 2012; Maragos, Zhu,

Chesnut, & Dwoskin, 2002; Offen, Ziv, Sternin, Melamed, & Hochman, 1996; S. U. Park, Ferrer,

Javitch, & Kuhn, 2002; Watabe & Nakaki, 2008). In fact, bursts of dopamine synthesis and release

that occur during mania could make dopaminergic synaptic terminals particularly vulnerable to

oxidative damage, as dopamine and L-DOPA are highly susceptible to oxidation by ROS and RNS,

and produce more ROS upon being oxidized (H. K. Kim & Andreazza, 2012; Miyazaki et al., 2006;

S. U. Park et al., 2002). Moreover, oxidative metabolism of dopamine is increased in the presence of

ROS and RNS (Graham, 1978; Graham, Tiffany, Bell, & Gutknecht, 1978; LaVoie & Hastings,

1999; Tse, McCreery, & Adams, 1976), suggesting that complex I dysfunction could also be

contributing to oxidative damage in the dopaminergic system. Oxidation of dopamine results in the

formation of additional free radicals that are cytotoxic (Asanuma, Miyazaki, Diaz-Corrales, &

Ogawa, 2004), such as the superoxide anion, hydrogen peroxide and hydroxyl radicals, and

cytotoxic dopamine/DOPA-quinones which are electron poor and act similarly to ROS (Hastings,

2009; Hastings et al., 1996; Hastings & Zigmond, 1994). Quinones can produce functional

modifications to cysteine containing proteins by forming 5-cysteinyl dopamine (Maker, Weiss,

Silides, & Cohen, 1981; Tse et al., 1976). Indeed, protein-bound dopamine caused by intrastriatal

dopamine injections can be prevented by adding antioxidants (Hastings et al., 1996), and DA-

induced cytotoxicity can be prevented by adding glutathione and N-acetyl cysteine, which increases

glutathione levels (C. T. Lai & Yu, 1997), suggesting that dopamine-induced toxicity is through the

production of oxidative stress. Moreover, inducing mania-like states with d-amphetamine and

increasing synaptic levels of dopamine were shown to increase lipid peroxidation and protein

oxidation (Frey et al., 2006; Valvassori et al., 2010).

22

Two markers of dopamine-rich areas are tyrosine hydroxylase (Elsworth & Roth, 1997) and the

dopamine transporter (Akil et al., 1999). Areas immunoreactive for the dopamine transporter and

tyrosine hydroxylase overlap extensively in the human brain, suggesting that both are good markers

for the dopamine synapse (Akil et al., 1999; Lewis et al., 2001), These two proteins are particularly

interesting because they are also vulnerable to oxidative/nitrosative modifications (Ara et al., 1998;

Blanchard-Fillion et al., 2001; Fleckenstein, Metzger, Beyeler, Gibb, & Hanson, 1997; S. U. Park et

al., 2002). Tyrosine hydroxylase is a target of peroxynitrite-induced tyrosine nitration in vitro, and

nitration of a single tyrosine residue was found to be able to reduce its activity (Ara et al., 1998).

While tyrosine hydroxylase is particularly susceptible to nitration (Blanchard-Fillion et al., 2001),

dopamine transporter is susceptible to disulfide formation (S. U. Park et al., 2002). Indeed, the

dopamine transporter is rich in cysteine residues and is a target of products of dopamine oxidation,

resulting in decreased uptake of dopamine from the synapse and further increasing dopamine

signaling (S. B. Berman & Hastings, 1999; S. B. Berman, Zigmond, & Hastings, 1996; LaVoie &

Hastings, 1999; Miyazaki et al., 2006). Hence, since oxidation of dopamine makes dopaminergic

synapses more vulnerable to oxidative stress, and mitochondrial production of ROS/RNS from

complex I dysfunction would increase dopamine oxidation, subsequent oxidative damage to

dopaminergic proteins at the synapse may be contributing to dopamine dysregulation observed in

BD.

While complex I dysfunction may have many downstream effects such as the aforementioned

oxidation of dopaminergic proteins at the synapse and activation of the NLRP3-inflammasome in

animal and in vitro studies (Maragos et al., 2002; Sherer et al., 2003; Zhou et al., 2011), it remains to

be seen whether complex I dysfunction is underlying the alterations that are observed in BD.

Moreover, lithium, which was shown to improve complex I function, was also found to decrease

dopamine signaling and pro-inflammatory cytokines, suggesting that lithium may be exerting its

23

effects at least in part by ameliorating alterations that occur downstream of complex I defect and

oxidative stress.

1.4 Scope and Hypotheses

1.4.1 Scope

Bipolar disorder is a serious mood disorder with recurrent episodes, where symptom severity,

response to medications and cognitive ability decrease with recurrent episodes, bringing attention to

the urgent need for the development of novel and more efficacious treatments to prevent relapse in

patients with BD (Berk et al., 2011). Although studies have reported numerous alterations in the

brain and peripheral cells of patients with BD, one of the most consistent findings is alterations in

proteins of the mitochondrial electron transport chain, especially in complex I (Kato & Kato, 2000;

Konradi et al., 2004; Konradi et al., 2012; Sun et al., 2006). Complex I dysfunction can result in

increased generation of oxidative stress, which is also consistently reported in patients with BD

(Andreazza et al., 2008; Andreazza et al., 2013; Halliwell & Gutteridge, 2007). Indeed, increased

levels of carbonyl groups, 3-nitrotyrosine, lipid peroxidation products, and oxidized DNA and RNA

have been found in patients with BD (Andreazza et al., 2009; Andreazza et al., 2008; Andreazza et

al., 2010; Brown et al., 2014). This has important implications, as studies have demonstrated a wide

array of targets of oxidative stress in vitro and in animal models by structurally and functionally

modifying proteins that are susceptible to oxidative modifications (Halliwell, 1992, 2001) or by

acting as signaling molecules through redox sensors. What is unknown is whether complex I

dysfunction and oxidative stress underlie the changes that are observed in BD.

Complex I dysfunction leads to increased generation of ROS when subunits that are involved in

the electron transfer process are dysfunctional, allowing for electrons to escape to react with

molecular oxygen, producing the superoxide anion (Halliwell & Gutteridge, 2007). To examine if

24

complex I dysfunction and susceptibility to increased generation of ROS is specific to BD, we re-

examined microarray studies that measured complex I subunits in patients with BD or schizophrenia.

I also aimed to confirm complex I dysfunction in the post-mortem brain of patients with BD, which

was reported by two independent studies from our group (Andreazza et al., 2010; Andreazza et al.,

2013).

A recent study from our group identified the mitochondria and the synapse as two targets of

oxidative stress in the post-mortem brain of patients with BD (Andreazza et al., 2013). Therefore, I

aimed to extend on these findings by studying downstream targets of complex I dysfunction in the

post-mortem brain of patients with BD, focusing on the mitochondria and the synapse. To study the

mitochondria as a downstream target of complex I defect, I examined activation of the NLRP3-

inflammasome, which is a marker of mitochondrial oxidative stress (Zhou et al., 2011), in the post-

mortem frontal cortex. Activation of the NLRP3-inflammasome also results in the release of

inflammatory cytokines (Schroder et al., 2010), further suggesting that this marker may be relevant

for BD as increased inflammation is consistently found in patients with BD (Berk et al., 2011; Rao et

al., 2010). To examine the synapse as a target of complex I dysfunction, I examined oxidative

modifications to the dopamine system, as dopaminergic proteins largely reside in the synapse and

increased levels of synaptic dopamine is known to underlie manic symptoms (Berk et al., 2007;

Cousins et al., 2009).

To further understand the downstream targets of complex I in BD, we used cell-models, allowing

for direct inhibition of the electron transfer process in complex I using rotenone (Maragos et al.,

2002; Naoi et al., 2005; Sherer et al., 2003). Also, lithium was used to examine if the observed

alterations may be relevant to BD pathology, as lithium is uniquely used for the treatment of BD (J.

S. Lai et al., 2006; Maurer, Schippel, & Volz, 2009; Sun et al., 2006; Washizuka, Iwamoto,

Kakiuchi, Bundo, & Kato, 2009).

25

Examining post-mortem brains allows us to confirm complex I dysfunction in BD, and to

directly observe alterations that may be occurring downstream of complex I defect and oxidative

stress in patients. On the other hand, experiments using cell models provide us with the opportunity

to directly manipulate complex I functioning and observe its consequences. In examining the

downstream targets of complex I dysfunction using post-mortem brains and cell models, we hoped

to benefit from the advantages provided by both approaches in elucidating the role of complex I

dysfunction in the pathophysiology of BD.

1.4.2 Aims and hypotheses

The overall aim of my PhD was to examine downstream targets of complex I dysfunction in BD,

since complex I dysfunction may have specific targets through the production of oxidative stress as

demonstrated by a recent study from our group (Andreazza et al., 2013).

1. Re-examination of microarray studies:

a. Aim 1: To review microarray studies examining the post-mortem brain of patients

with BD or schizophrenia to determine if complex I dysfunction in the electron

transfer process is specific to patients with BD, making patients with BD more

susceptible to having increased generation of oxidative stress.

Hypothesis: Decreased levels of complex I subunits involved in the electron transfer

process will be unique to patients with BD, suggesting increased susceptibility to the

production of ROS in these patients.

2. Post-mortem brain studies: Studies in post-mortem brain aimed to confirm complex I

dysfunction in patients with BD and extend on a previous study from our group identifying

the mitochondria and the synapse as two targets of oxidative stress in the post-mortem frontal

cortex of patients with BD.

26

a. Aim 2: To confirm complex I dysfunction in the post-mortem brain of patients with

BD.

Hypothesis: Patients with BD will have decreased levels of complex I and its subunit,

NDUFS7, compared to non-psychiatric controls in the post-mortem frontal cortex.

b. Aim 3: To examine the mitochondria as a target of complex I dysfunction in BD by

measuring the activation of the NLRP3-inflammasome, a marker of mitochondrial

oxidative stress.

Hypothesis: Mitochondria will be found to be a potential downstream target of

complex I dysfunction in patients with BD, as shown by increased activation of the

NLRP3-inflamamsome in the post-mortem frontal cortex.

c. Aim 4: To examine the synapse as a target of complex I dysfunction in BD by

measuring oxidative modifications to the dopaminergic synapse, since dopaminergic

dysregulation underlies mania.

Hypothesis: The synapse will be found to be a potential downstream target of

complex I dysfunction in patients with BD, as shown by increased oxidative

modifications to the dopaminergic synapse in the post-mortem prefrontal cortex.

3. Cell model studies: Studies in cell models aimed to further examine downstream targets of

complex I dysfunction in a more controlled environment, allowing for the direct inhibition of

the electron transfer process in complex I using rotenone. The effect of lithium was also

examined, allowing us to see if resulting alterations from complex I inhibition may be

relevant to BD pathology.

a. Aim 5: To examine potential downstream targets of complex I dysfunction by using

rotenone to block the process of electron transfer in complex I, and to examine the

relevance of alterations occurring from rotenone treatment to BD by testing if lithium

27

is able to decrease these changes. Protein oxidation, nitration, and changes in

methylation and hydroxymethylation of DNA were measured as potential

downstream targets of complex I defect.

Hypothesis: Inhibition of the electron transfer process of complex I using rotenone

will result in increased generation of oxidative damage, as measured by increased

protein oxidation, nitration, and methylation and hydroxymethylation of DNA.

Lithium pre-treatment will be able to decrease these alterations produced by rotenone

treatment.

By testing these hypotheses in post-mortem brain and cell models, I aim to elucidate the

potential downstream targets of complex I dysfunction in BD, which will lead to a better

understanding of the pathways that are involved in the pathophysiology of this highly complex

disease.

28

CHAPTER 2

Title: A Fresh Look at Complex I in Microarray Data: Clues to Understanding Disease-Specific

Mitochondrial Alterations in Bipolar Disorder

Authors: Gustavo Scola*, Helena Kyunghee Kim*, L. Trevor Young, Ana Cristina Andreazza

*Authors GS and HKK contributed equally to this letter.

Journal: Biological Psychiatry

Volume: 2

Pages: e4-5

Work performed by the student:

The student performed all work included in this chapter, including literature review, and writing of

the manuscript.

29

A fresh look at complex I in microarray data: clues to understanding disease-specific

involvement of mitochondrial alterations in bipolar disorder

Scola G1*, Kim HK2*, Young LT1,2, Andreazza AC1,2

1. Departments of Psychiatry and Pharmacology, University of Toronto, ON, Canada

2. Centre for Addiction and Mental Health, Toronto, ON, Canada

30

To the Editor:

Mitochondrial dysfunction and the consequent generation of oxidative stress damage have

been consistently reported in bipolar disorder (BD) and schizophrenia (SCZ). Microarray studies are

a major source of evidence for this mechanism, where the expression of many electron transport

chain (ETC) genes from complexes I-V was found to be decreased in patients with BD in the frontal

cortex (Choi et al., 2011; Jurata et al., 2004; Sun et al., 2006; Washizuka et al., 2009; Washizuka et

al., 2003) and hippocampus (Konradi et al., 2004). Mitochondrial ETC complex I is one of the

major sites for the generation of reactive oxygen species (ROS) (Fariss et al., 2005). Recently,

volume 71, issue 11 of the Biological Psychiatry journal addressed the implications of using

antioxidants for the maintenance of redox balance in psychiatric disorders (O. M. Dean, Bush, &

Berk, 2012; Hardan et al., 2012; Shungu, 2012). Considering the complexity of mitochondrial

complex I and its significance for psychiatric disorders, we re-examined the reported microarray

findings and grouped the altered complex I subunits by their relevance for ROS generation.

Complex I is a large complex consisting of 45-46 subunits, each of which plays a unique role

for the structure or activity of this complex, or ROS generation (Green & Kroemer, 2004; Lenaz,

2001). Subunits of complex I are arranged in four main subcomplexes, which are λ, α-λ, γ, and β.

The hydrophilic arm, which is responsible for the transfer of electrons, contains subcomplexes λ and

α-λ, and the hydrophobic arm, which is largely responsible for the pumping of protons, consists of γ

and β. Upon reviewing the literature, we identified 34 microarray studies examining BD and/or SCZ.

Of these studies, we have selected the microarray studies that included probes for complex I genes.

Finally, 10 studies were selected to be included in this report, as they reported alterations in complex

I genes in BD or SCZ. Microarrays enable the examination of the expression pattern of a significant

portion of the human genome encompassing multiple systems, making it ideal for the exploration of

the pathophysiology of complex disorders such as BD and SCZ (Mirnics, Levitt, & Lewis, 2006). In

31

total, 18 genes were found to be reduced or increased in their levels of expression in BD or SCZ,

where 8 were found to be altered in BD, 6 in SCZ, and surprisingly, only 4 genes were found to be

altered in both BD and SCZ (Figure 1). Combined findings of these studies revealed that in BD,

expression of iron-sulfur cluster-containing subunits within the hydrophilic arm were reduced,

suggesting that patients with BD may be more prone to having a dysfunction in the electron transfer

process. For instance, NDUFV1, which contains FMN and N3, and initiates the electron transfer

process from NADH to the iron-sulfur clusters, was found to be downregulated in BD. Also,

NDUFS1, which contains N1b, N4, N5, and N7, and NDUFS8, which contains N6a and N6b, were

also found to be downregulated. Finally, NDUFS7, which contains N2, and is responsible for the

reduction of ubiquinone to ubiquinol, was also found to be reduced in BD (Janssen, Nijtmans, van

den Heuvel, & Smeitink, 2006) (Figure 1). Although 4 subunits were found to be down-regulated in

the α-ʎ subcomplex in BD, these subunits are non-catalytic and we therefore will focus on the ʎ

subunit. In contrast, gene expression alterations in SCZ were found to be scattered throughout

complex I, with up and downregulation, and did not include alterations in subunits directly involved

in electron transfer. These data may suggest that while both patients with BD and SCZ may have a

reduction in complex I functionality, patients with BD may be more prone to deficiencies in the

electron transfer process, which could increase the probability of electrons escaping the electron

transport chain to react with molecular oxygen, causing a cascade of reactions to increase the

production of ROS such as the hydroxyl radical (Fariss et al., 2005). In fact, gene expression levels

of NDUSF7 was found to be reduced in two microarray studies in BD (Iwamoto et al., 2005; Sun et

al., 2006), and this finding was supported by a separate study, where protein levels of NDUFS7 was

found to be reduced in patients with BD (Andreazza et al., 2010). Importantly, this reduction in

NDUFS7 levels was found to be correlated with decreased complex I activity (Andreazza et al.,

32

2010). This indicates that altered gene expression identified in microarray studies could have

functional implications for complex I activity and ROS generation.

There are several caveats to these data. For example, different platforms were used for data

acquisition and analysis (Konradi et al., 2004; Mirnics et al., 2006), which could generate variability

between studies. Although post-mortem brain allows for direct investigation of human pathology,

there are inherent limitations involved in the use of samples acquired post-mortem, such as pH, post-

mortem interval, and small sample size (Andreazza et al., 2010; Iwamoto et al., 2005).

In summary, the data from the microarray studies discussed here suggest important

differences in the expression of complex I genes between BD and SCZ. In patients with BD, there is

down-regulation specifically in genes involved in electron transfer in complex I. On the other hand,

altered genes in SCZ were found to be scattered through complex I and include increased as well as

decreased expression levels. The findings reported here suggest that in bipolar disorder, emphasis

must be placed on the electron transfer chain, which may contribute to the elucidation of the

pathogenesis of this condition.

Acknowledgements

The authors would like to acknowledge CNPq/Brazil (Gustavo Scola) and CIHR as sources

of funding in support of this report.

33

Figures

Figure 1. Summary of mitochondrial complex I gene alterations in bipolar disorder and

schizophrenia. Gene alterations and direction of alteration (arrows) in bipolar disorder (BD;

highlighted in grey) and schizophrenia (SCZ) are divided into different brain areas and

subcomplexes where the alterations were found. Downregulations are shown in blue, and

upregulations are shown in red. Mitochondrial complex I subunits NDUFV1, NDUFS1, NDUFS8,

and NDUFS7 within the λ subcomplex were found to be downregulated in BD. These subunits

contain iron-sulfur clusters that are directly involved in the electron transfer process from NADH to

ubiquinone.

34

Statement of significance and impact:

BD and SCZ contain many pathological similarities, including calcium alterations, decreased

numbers of neurons, neurotransmitter imbalance, and inflammation (Clay et al., 2010). Many studies

have reported alterations in energy metabolism and mitochondrial proteins in BD and SCZ, in

particular for those belonging to oxidative phosphorylation (Clay et al., 2010; Iwamoto, Kakiuchi,

Bundo, Ikeda, & Kato, 2004). However, a study by our group reported that only patients with BD

have lower protein levels of NDUFS7, a complex I subunit that is critical for the proper transfer of

electrons to ubiquinone, and may contribute to electron leakage when it is dysfunctional (Andreazza

et al., 2010). Hence, this review addressed Aim 1, which was to identify if complex I dysfunction in

the electron transfer process is specific to patients with BD. This review of microarray studies

examining complex I subunits in BD and SCZ revealed that patients with BD have lower levels of

complex I subunits that are involved specifically in the process of electron transfer, while patients

with SCZ have increases or decreases in subunits that are responsible for different functions, such as

proton pumping or structural support. It should be noted, however, that decreased ability to pump

protons may also contribute to production of mitochondrial ROS indirectly, such as by decreasing

ATP production and producing stress in the cell (Schütt, Aretz, Auffarth, & Kopitz, 2012). On the

other hand, inhibition in the electron transfer process leads to the direct production of superoxide

anions, and rotenone, which interferes with electron transferring subunits in complex I, was

consistently shown to increase mitochondrial ROS (Bashkatova, Alam, Vanin, & Schmidt, 2004;

Boveris & Chance, 1973; N. Li et al., 2003; Sherer et al., 2003; Sorbara & Girardin, 2011). These

findings suggest that patients with BD may be particularly vulnerable to having increased generation

of mitochondrial ROS.

Mitochondrial ROS can directly modify enzymes and transporters, altering their function,

and activate downstream pathways by acting as signaling molecules (Haddad, 2002; Lenaz, 2001),

35

suggesting that it may be contributing to other alterations observed in BD. A recent study from our

group demonstrated that mitochondria and the synapse are targets of oxidative stress in the post-

mortem frontal cortex of patients with BD (Andreazza et al., 2013). To extend on these findings, I

decided to first examine the post-mortem brain to confirm complex I dysfunction, and then to

examine a marker of mitochondrial oxidative stress, the NLRP3-inflammasome (Zhou et al., 2011).

Once the NLRP3-inflammasome is activated following complex I dysfunction and release of

mitochondrial ROS, it releases pro-inflammatory cytokines, making this pathway particularly

interesting as inflammation is consistently reported in patients with BD (Goldstein, Kemp, et al.,

2009; Leboyer et al., 2012; Liu et al., 2004).

36

CHAPTER 3

Title: Nod-like receptor pyrin containing 3 (NLRP3) in the postmortem frontal cortex from patients

with bipolar disorder: a potential mediator between mitochondria and inflammation

Authors: Helena Kyunghee Kim, Ana Cristina Andreazza, Wenjun Chen, L. Trevor Young

Work performed by the student:

The student performed all work included in this chapter, including planning and performing the

experiments, data analysis, and preparation of the manuscript.

37

Nod-like receptor pyrin containing 3 (NLRP3) in the postmortem frontal cortex from patients

with bipolar disorder: a potential mediator between mitochondria and inflammation

Helena Kyunghee Kim1, Ana Cristina Andreazza1,2, Wenjun Chen1, L. Trevor Young1,2

1Departments of Psychiatry and Pharmacology, University of Toronto, Canada

2Campbell Family Mental Health Research Institute, Centre for Addiction and Mental Health

38

Abstract

Mitochondrial complex I dysfunction and inflammation are consistently reported in bipolar disorder

(BD). Mitochondrial production of reactive oxygen species (ROS), which occurs when complex I is

working ineffectively, was recently linked to activation of a redox sensor in the inflammatory

system, the nod-like receptor family pyrin domain-containing 3 (NLRP3). Upon its activation,

NLRP3 recruits apoptosis-associated speck-like protein (ASC) and caspase-1 to form the NLRP3-

inflamamsome, activating IL-1β and IL-18. This study aimed to examine if inflammation may be a

downstream target of complex I dysfunction through the NLRP3-inflamamsome in BD. Post-mortem

frontal cortex from patients with BD (N=9), schizophrenia (N=10), and non-psychiatric controls

(N=9) were donated from the Harvard Brain Tissue Resource Center. Levels of NLRP3, ASC and

caspase-1 were measured by western blotting and ELISA, and levels of downstream cytokines were

measured using Luminex. While we found no effects of age, sex or post-mortem delay, lower levels

of complex I (F2,25=3.46, p<0.05) and NDUFS7, a subunit of complex I (F2,25= 4.13, p<0.05), were

found in patients with BD compared with healthy controls. Mitochondrial NLRP3 (F2,25=3.86,

p<0.05) and ASC (F2,25=4.61, p<0.05) levels were higher in patients with BD. However, levels of

caspase 1 (F2,25=4.13, p<0.05 for both), IL-1β (F2,25=7.05, p<0.01), IL-6 (F2,25=5.48, p<0.05), TNFα

(F2,25=7.14, p<0.01) and IL-10 (F2,25=5.02, p<0.05) were found to be increased in both BD and

schizophrenia. These findings suggest that inflammation in the frontal cortex may occur both in

patients with BD and schizophrenia, while complex I dysfunction and NLRP3-inflammasome

activation may be more specific to BD.

39

Evidence supporting mitochondrial dysfunction, such as decreased mitochondrial respiration,

metabolism (Kato et al., 1993) and mitochondrial DNA polymorphisms (Kato, 2008) are

consistently reported in BD (Clay et al., 2010), a debilitating disorder affecting 1-2% of the

population worldwide (Merikangas et al., 2011). In particular, decreased mRNA and protein levels

of subunits that are involved in the transfer of electrons in complex I, which is a member of the

electron transport chain, may be a finding that is specific to BD (Iwamoto et al., 2005; Konradi et al.,

2012; Scola, Kim, Young, & Andreazza, 2012). In microarray studies, for example, NDUFS7, the

final iron-sulfur cluster-containing subunit involved in electron transfer within complex I, was found

to be lower in patients with BD (Iwamoto et al., 2005; Scola et al., 2013; Sun et al., 2006). Protein

levels of NDUFS7 were also found to be lower in two separate studies examining post-mortem

brains from patients with BD (Andreazza et al., 2010; Andreazza et al., 2013). Decreased efficiency

of the electron transfer process within complex I has important implications, as it can result in

increased leakage of electrons that react with molecular oxygen to form the superoxide anion

(Halliwell & Gutteridge, 2007). The superoxide anion can undergo further reactions to form

powerful ROS (Halliwell & Gutteridge, 2007). In support of these findings, increased oxidative

damage to proteins, lipids and DNA are consistently reported in the brain and periphery of patients

with BD (Andreazza et al., 2010; Andreazza et al., 2013; Clay et al., 2010).

ROS also serve as important signaling molecules through redox sensors. Recent studies have

suggested that ROS produced by the mitochondria may be a potent activator of the inflammatory

system (López-Armada et al., 2013). Increasing levels of mitochondrial ROS production by

inhibiting complex I, for instance, was found to result in greater levels of inflammatory factors such

as NF-κB and IL-1β (N. Li et al., 2003). These findings have important implications for BD, as

increased levels of inflammatory cytokines have been found in the brain and periphery of patients

with BD (B. Dean et al., 2013; Kauer-Sant'Anna et al., 2009; Y. K. Kim, Jung, Myint, Kim, & Park,

40

2007; Leboyer et al., 2012; Munkholm et al., 2013; O'Brien et al., 2006; Rao, Kellom, Reese,

Rapoport, & Kim, 2012). More specifically, IL-6 and TNF-α were reported to be increased in studies

examining peripheral samples from patients with BD (Y. K. Kim et al., 2007; Munkholm et al.,

2013), and studies examining the central nervous system have reported increases in cytokines and

receptors involved in the IL-1 pathway (B. Dean et al., 2013; Rao et al., 2010; Söderlund et al.,

2011).

Nod-like receptor pyrin containing 3 (NLRP3) is a redox sensor in the inflammatory system

that has been implicated as a potential link between mitochondrial ROS production and

inflammation (Tschopp & Schroder, 2010; Zhou et al., 2011). NLRP3, which normally resides in the

cytoplasm, undergoes a structural change upon the release of mitochondrial ROS and migrates to the

mitochondria, allowing it to be close to the site of damage (Zhou et al., 2011). This structural change

allows NLRP3 to recruit two other cytoplasmic proteins - apoptosis-associated speck-like protein

containing a CARD (ASC) and caspase-1 which form the NLRP3-inflammasome (Schroder et al.,

2010; Shimada et al., 2012; Tschopp & Schroder, 2010; Zhou et al., 2011). The assembly of the

inflammasome allows for the cleavage of caspase-1 into its active form, releasing it into the

cytoplasm to cleave pro-IL-1β and pro-IL-18 to their biologically active forms (Perregaux & Gabel,

1994). IL-1β and IL-18 are then released from the cells to activate downstream pathways involving

MYD88 and NF-κB (Eder, 2009), leading to increased expression of other inflammatory molecules

such as IL-6 and TNF-α (Bryan et al., 2010; A. Weber, Wasiliew, & Kracht, 2010). What remains

unknown, however, is whether inflammation is a downstream target of mitochondrial production of

ROS in BD through the NLRP3-inflamamsome. Therefore, in this study, we examined

mitochondrial dysfunction, localization and levels of NLRP3, ASC, and caspase-1, and levels of

downstream cytokines of the inflammasome pathway in frontal cortex of patients with BD compared

to healthy controls.

41

Materials and Methods

Subjects

Frozen post-mortem frontal cortex (BA9,10,24) tissues were generously donated from the Harvard

Brain Tissue Resource Centre (Table 1). These well-characterized samples have been in previous

publications by our lab and others (Andreazza et al., 2013; Rao et al., 2012). Briefly, subject groups

consisted of patients with BD (N=9), schizophrenia (SCZ; N=10) and non-psychiatric controls

(N=9). There were no between-group differences for sex (F2,25 = 0.023, p = 0.98), age (F2,25 = 0.18, p

= 0.83) and post-mortem interval (F2,25 = 0.14, p = 0.87). Diagnoses were established according to

DSM-IV criteria. Samples were coded numerically, and the experimenters were blind to the codes

until all the experiments and data analyses were completed.

Measurement of mitochondrial ETC proteins in the post-mortem frontal cortex

Levels of mitochondrial ETC components complex I (NADH-ubiquinone oxidoreductase), II

(succinate ubiquinone oxidoreductase), III (ubiquinone cytochrome c oxidoreductase), IV

(cytochrome c oxidase) and V (ATP synthase) and nicotinamide nucleotide transhydrogenase (NNT)

were measured using the Human oxidative phosphorylation magnetic bead panel kit (EMD

Millipore, #H0XPSMAG-16K) according to manufacturer’s instructions. Briefly, tissue extracts

were prepared with the lysis buffer provided in the kit followed by centrifugation at 14,000xg for 20

minutes. Samples (5ug total protein) were incubated with antibody-coated magnetic beads followed

by detection antibodies and streptavidin-phycoerythrin. Luminex Magpix (EMD Millipore) was used

to read the plates. Analysis was performed with the xPONENT software, with results expressed as

median fluorescence intensity (MFI). NDUFS7 levels were measured by immunoblotting using a

previously published protocol (Andreazza et al., 2013). First, brain homogenate was prepared by

42

sonication in a mannitol-sucrose buffer (225mM mannitol, 75mM sucrose, 30mM Tris-HCL at

pH7.4) (Wieckowski, Giorgi, Lebiedzinska, Duszynski, & Pinton, 2009). 15ug of proteins were

loaded onto 12% polyacrylamide gels and transferred to polyvinylidene difluoride (PVDF)

membranes. Blotting was performed using a primary antibody against NDUFS7 (Santa Cruz, Dallas,

TX, USA, sc-98644; 2h; 1:1000) followed by a secondary goat anti-rabbit IgG horseradish

peroxidase antibody (Abcam Inc, Cambridge, MA, USA, #ab97051; 1h; 1:2000). For loading

control, mouse monoclonal anti-beta actin (Abcam Inc, #ab8226, 1hr; 1:1000; secondary antibody –

1:3000, 30min) was used. Membranes were visualized by incubating with ECL reagents and

analyzed using Versa Doc from Bio-Rad (Bio-Rad Laboratories Ltd., Mississauga, Canada).

Measurement of NLRP3 in the post-mortem frontal cortex

First, post-mortem brain was homogenized by sonication and applied to a sugar gradient (mannitol +

sucrose) to isolate the crude mitochondrial fraction and the cytosolic fraction using a previously

reported method that preserves the association between mitochondrial membranes and NLRP3-

inflammasome (Wieckowski et al., 2009; Zhou et al., 2011). Samples were first homogenized as

described above, and centrifuged twice at 740g for 5min to collect the supernatant. The supernatant

was then centrifuged at 9000g for 10 minutes. The resulting supernatant was collected as the

cytosolic fraction, and the pellet containing the crude mitochondrial fraction was resuspended and

centrifuged twice at 10,000g for 10 minutes. The resulting pellet was then collected and resuspended

in mitochondrial resuspending buffer (250mM mannitol, 5mM HEPES, 0.5mM EGTA at pH7.4).

The following procedures were performed on the brain homogenate as well as the cytoplasmic and

crude mitochondrial fractions. To measure NLRP3 levels, we performed western blotting according

to a previously published protocol (Andreazza et al., 2013). Samples (10ug for brain homogenate,

20ug for cytoplasmic fraction, and 35ug for mitochondrial fraction) were loaded on to 10%

43

polyacrylamide gels, and transferred to PVDF membranes. The membranes were stained with

Memcode reversible stain (Pierce BCA protein Assay kit, 23227) according to manufacturer’s

instructions and blocked (1h, 5% BSA). Incubation with primary anti-NLRP3 polyclonal antibody

(Millipore Co, Billerica, MA, USA, #ABF23; 2h; 1:500 for mitochondria, 1:2000 for cytoplasmic,

1:3000 for homogenate) was followed by a secondary goat anti-rabbit IgG horseradish peroxidase

antibody (Abcam Inc; 1h; 1:2000 for mitochondrial, 1:3000 for cytoplasmic, 1:5000 for whole

brain). Membranes were incubated with ECL reagents and analyzed using Versa Doc from Bio-Rad

(Bio-Rad Laboratories Ltd., Mississauga, Canada). For loading control, Memcode was used for the

crude mitochondrial fraction, and β-actin was used for cytoplasmic fraction and brain homogenate.

Measurement of ASC and caspase-1 in the post-mortem frontal cortex

ASC and caspase-1 levels were measured in the brain homogenate, cytoplasmic and mitochondrial

fractions using the ELISA technique. For ASC, samples were diluted to 20μg/ml for the brain

homogenate, cytoplasmic and mitochondrial fractions. For caspase 1, samples were diluted to

10ug/mL for the brain homogenate and cytoplasmic fraction, and 20ug/mL for the mitochondrial

fraction. 100μl/well of the samples were loaded onto 96w microlon ELISA plates (Gibco Inc,

Sainte-Marie, Quebec, Canada). ASC peptide (Abcam Inc, #ab22875) was used to create a curve

ranging from 10ng - 0pg to allow for quantification of ASC in the samples. THP-1 cell lysate was

used as a positive control for caspase-1(Chanput, Mes, & Wichers, 2014). Samples were incubated

overnight at 4°C on a shaker followed by blocking using 1% BSA (1hr, 37°C). Then, samples were

incubated with either primary anti-ASC antibody (Abcam, #ab111852; 2h; 1:32000 for homogenate,

1:20000 for cytoplasmic and mitochondrial fractions) or the primary anti-caspase-1 antibody

(Abcam, #ab62698; 2h; 1:5000 for all). Samples were incubated with horse radish peroxidase-

conjugated secondary antibody (Abcam; 1h; #ab6885; 1:4000 RT for ASC, #ab97051; 1:2000 RT

44

for caspase-1), and visualized by adding 100uL/well of TMB solution followed by 100uL/well of

stopping solution (1.2M HCl). Synergy H1 hybrid reader (BioTekR, Winooski, VT, USA) was used

to read the optical density at 450nm using the Gen 5 1.11 software.

Measurement of downstream cytokines of the NLRP3-inflammasome pathway in the post-mortem

frontal cortex

Levels of IL-1β, IFN-γ, IL-10, IL-6, and TNF-α in the brain homogenate were measured using the

human cytokine magnetic bead panel Milliplex Map kit (EMD Millipore, #HCYTOMAG-60K)

according to manufacturer’s instructions. Briefly, samples (20ug total protein) were incubated with

antibody-coated magnetic beads followed by detection antibodies and streptavidin-phycoerythrin.

Luminex Magpix (EMD Millipore) was used to read the plates. Analysis was performed with the

xPONENT software, where the amount of cytokines in the samples was quantified using a standard

curve provided by the kit.

Statistical analysis

Statistical analysis was performed using SPSS ver 21. Kolmogorov-Smirnov test was used to

determine normal distribution of data. For parametrically distributed data, one-way ANOVA was

used to examine between-group differences followed by Least Significant Difference post-hoc test to

examine the effect of diagnosis. For non-parametrically distributed data, Kruskal-Wallis test was

used to examine between-group differences followed by Mann-Whitney U test to examine the effect

of diagnosis. Pearson’s correlation test was used to measure any associations between age, sex,

postmortem interval (PMI) and measured variables. If the correlation was significant, ANCOVA was

performed with age/sex/PMI as a covariate. Pearson’s correlation test was also used to examine

correlations. Data are presented as mean±standard error of the mean (SEM).

45

Results

Complex I and NDUSF7

Levels of complex I, complex II, complex III, complex IV, complex V and nicotinamide nucleotide

transhydrogenase were measured using the Luminex system. Age, sex and PMI did not correlate

with any of the variables. We found a significant group difference for complex I (F2,25 = 3.46, p <

0.05), where lower levels of complex I was found in patients with BD (Fig. 1) compared to non-

psychiatric controls (p < 0.05). Complex II, III, IV, V and nicotinamide nucleotide transhydrogenase

did not differ between groups. Since the decrease in BD was specific to complex I, and because

previous studies reported decreased levels of NDUFS7, a complex I subunit, in patients with

BD(Andreazza et al., 2010; Andreazza et al., 2013), levels of NDUFS7 were measured using

western blotting. We found a significant group difference, (F2,25=4.13, p < 0.05), where patients with

BD were found to have lower levels of NDUFS7 (Fig. 1G) than patients with SCZ (p < 0.05) and

non-psychiatric controls (p < 0.05).

Components of the NLRP3-inflammasome in different subcellular fractions

Levels of NLRP3, ASC and caspase-1 were measured in the brain homogenate, cytosolic and crude

mitochondrial fractions. Western blotting was used to examine NLRP3, and the ELISA technique

was used to measure levels of ASC and caspase-1. Age, sex, and PMI did not correlate with any of

the measured variables. Between-group difference for NLRP3 was found in the crude mitochondrial

fraction (F2,25 = 3.83, p < 0.05; Figure 2C), with higher levels in patients with BD compared to

patients with SCZ (p < 0.05) and non-psychiatric controls (p < 0.05). There were no differences in

either the brain homogenate or the cytoplasmic fraction. Levels of ASC were also found to differ

between groups in the crude mitochondrial fraction (F2,25 = 4.61, p < 0.05; Figure 2F), with higher

46

levels in patients with BD compared to patients with SCZ (p < 0.05) and non-psychiatric controls (p

< 0.01). Caspase-1 was found to differ between groups in the crude mitochondrial fraction (F2,25

=4.13, p < 0.05; Figure 2I), with higher levels in patients with BD (p < 0.05) or SCZ (p < 0.05)

compared to non-psychiatric controls.

Cytokines downstream of the NLRP3-inflammasome

Levels of cytokines involved in the IL-1β pathway, including IL-1β, IL-6, TNF-α, and cytokines

involved in the IL-18 pathway, including IFN-γ and IL-10, were measured using the Luminex

system. Age, sex and PMI did not correlate with any of the variables. For the IL-1β pathway, a

between-group difference was found for IL-1β (F2,25 = 7.05 p < 0.01; Figure 3A), with higher levels

in patients with BD or SCZ (p < 0.01 for both) compared to non-psychiatric controls. IL-6 also

significantly differed between groups (F2,25 = 5.48, p < 0.05; Figure 3B), such that patients with BD

and SCZ (p < 0.01 for both) had higher levels than non-psychiatric controls. Similar results were

found for TNF-α (F2,25 = 7.14, p < 0.01, p < 0.01 for both BD and SCZ compared to non-psychiatric

controls; Figure 3C). For the IL-18 pathway, group difference was found in IL-10 (F2,25 = 5.02, p <

0.05; Figure 3F), with higher levels in patients with BD or SCZ compared to non-psychiatric

controls (p < 0.05 for both). IFN-γ did not differ between groups.

Discussion

In this study, we examined activation of the NLRP3-inflammasome in BD using post-

mortem frontal cortex from patients with BD, SCZ and non-psychiatric controls. Our results showed

decreased levels of complex I and NDUFS7 in patients with BD, suggesting complex I dysfunction,

which could in turn lead to greater production of mitochondrial ROS. Consistent with these findings,

we also found increased levels of NLRP3 and ASC in the crude mitochondrial fraction of patients

47

with BD, suggesting increased activation of the inflammasome as a result of complex I dysfunction

in these patients. On the other hand, we found that levels of caspase-1 and downstream cytokines are

increased in both patients with BD or SCZ, demonstrating activation of the inflammatory system in

both disorders. Together, these findings suggest that complex I dysfunction and subsequent

activation of the NLRP3-inflammasome in the frontal cortex may be specific to BD. Moreover,

while the NLRP3-inflammasome may contribute to increased levels of inflammatory cytokines in

patients with BD, inflammation in SCZ may occur due to other causes.

When we examined components of the electron transport chain, we only found a group

difference in complex I, where lower levels were found in patients with BD than non-psychiatric

controls. Complex I extracts electrons from NADH to transfer it to ubiquinone, reducing it to

ubiquinol. This is done by relaying electrons through the hydrophilic arm of complex I, which

contains iron-sulfur cluster containing subunits (Halliwell & Gutteridge, 2007). When this process is

disturbed, electrons can escape the system to react with molecular oxygen, producing the superoxide

anion (Halliwell, 2007). Because of this, complex I has been strongly implicated in mitochondrial

production of ROS, and disruption of complex I using chemicals such as rotenone, has been linked

to increased oxidative damage in cells and in animals (Bashkatova et al., 2004; Nascimento et al.,

2014; Sorbara & Girardin, 2011). Interestingly, a review of microarray studies revealed that patients

with BD have lower mRNA levels of complex I subunits that are directly involved in the electron

transfer process, while patients with SCZ show changes in subunits that are involved in other

functions of complex I, such as structural support or proton pumping (Scola et al., 2012). This

suggests that patients with BD may be particularly vulnerable to increased production of

mitochondrial ROS from complex I dysfunction. As complex I is a multi-protein structure, we

wanted to explore this further and examine NDUFS7, which contains the last iron-sulfur cluster

responsible for transferring electrons to ubiquinone(Janssen et al., 2006). Importantly, levels of

48

NDUFS7 were found to correlate with complex I activity (Andreazza et al., 2010), suggesting that it

is most likely important for its functioning. Our findings showed lower levels of NDUFS7 in patients

with BD compared to patients with SCZ and non-psychiatric controls. These findings are in

agreement with two independent studies that examined this subunit in post-mortem brain (Andreazza

et al., 2010; Andreazza et al., 2013), suggesting that it is a very consistent finding for BD.

Our findings having suggested complex I dysfunction in BD, we examined evidence of

NLRP3-inflammsome activation as mitochondrial ROS production is an important precursor for

inflammasome assembly (Zhou et al., 2011). We examined the three components of the NLRP3-

inflammasome, NLRP3, ASC and caspase-1, in the brain homogenate, cytoplasmic and crude

mitochondrial fractions of the post-mortem frontal cortex. Our findings showed that levels of

NLRP3 and ASC are increased only in the mitochondrial fraction from patients with BD, while no

group differences were found in the brain homogenate and the cytoplasmic fraction. A previous

study reported that activating the NLRP3-inflammasome caused an increase in co-localization

between the mitochondria and inflammasome components using microscopy (Zhou et al., 2011).

This finding was confirmed using subcellular fractionation studies in which NLRP3 and ASC were

found to be increased in the mitochondrial fraction and mitochondria-associated membranes, both of

which are part of the crude mitochondrial fraction (Wieckowski et al., 2009; Zhou et al., 2011).

Moreover, adding a complex I inhibitor, rotenone, caused an increase in IL-1β secretion (N. Li et al.,

2003), while the addition of a mitochondrial ETC inhibitor did not result in IL-1β secretion in Nlrp3

KO mice (Zhou et al., 2011). Importantly, activation of the inflammasome was completely blocked

by adding an ROS inhibitor, ammonium pyrrolidine dithiocarbamate, suggesting that NLRP3-

inflammsome assembly at the mitochondria is triggered by mitochondrial release of ROS (Zhou et

al., 2011). Thus, our finding of increased NLRP3 and ASC only in the mitochondrial fraction of

patients with BD suggests increased activation of the inflammsome in these patients. This is in

49

agreement with our results showing decreased levels of complex I and NDUFS7, as decreased levels

of complex I and resulting increase in mitochondrial ROS production may lead to increased

assembly of the NLRP3-inflammasome complex at the mitochondria (Zhou et al., 2011). As our

results suggested complex I dysfunction and increased activation of the NLRP3-inflammasome in

patients with BD, we correlated levels of complex I with NLRP3, ASC and caspase 1 levels in the

brain homogenate, cytoplasmic fraction and crude mitochondrial fraction (Figure 4). While we did

not find significant correlations, it is difficult to conclude that this suggests a lack of association

between complex I levels and the NLRP3-inflammasome due to the small number of subjects (N=9

in BD group).

Inflammation has been hypothesized to play a role in psychiatric disorders since the 1970s.

Since that time, evidence demonstrating cytokine alterations in BD and SCZ has accumulated

(Libíková et al., 1979; Maes et al., 1995; Munkholm et al., 2013; Müller, Myint, & Schwarz, 2012).

In order to examine if the NLRP3-inflammsome contributes to inflammation in BD, we measured

levels of cytokines that are activated downstream of the NLRP3-inflammasome in our samples.

Caspase-1 cleaves and activates two cytokines, IL-1β and IL-18, both of which are released outside

of the cells to activate different signaling pathways (Sollberger et al., 2014). We examined IL-1β,

IL-6, and TNF-α for the IL-1β pathway. For cytokines activated downstream of IL-18, also known

as IFN-γ inducing factor, we examined IFN-γ and IL-10 (Gracie, Robertson, & McInnes, 2003; A.

Weber et al., 2010). Our results indicated that IL-1β, IL-6, TNF-α and IL-10 are increased in patients

with BD and SCZ, suggesting an activation of the IL-1β pathway in these individuals, but not the IL-

18 pathway, since IFN-γ is a direct product of IL-18 activation. A previous study demonstrated that

the IL-18 pathway is unaffected by the NLRP3-inflammasome in microglial cells (Hanamsagar,

Torres, & Kielian, 2011), which are the primary immune cells of the brain (Dheen, Kaur, & Ling,

2007). This may be why we did not see group differences for IFN-γ despite evidence supporting

50

activation of the NLRP3-inflammasome in the BD group. It is also worth noting that while evidence

supporting activation of the NLRP3-inflammasome was stronger for patients with BD than SCZ,

higher levels of cytokines were found in both BD and SCZ. This is in agreement with previous

studies demonstrating inflammation in the brain and periphery of patients with BD or SCZ (B. Dean

et al., 2013; Kauer-Sant'Anna et al., 2009; Leboyer et al., 2012; Müller et al., 2012; Rao et al., 2010).

Indeed, altered levels of cytokines in the IL-2, TNF-α, IL-6 and IL-1 pathways are consistently

reported in the central nervous system and periphery of patients with BD or SCZ (Barbosa et al.,

2011; Bresee & Rapaport, 2009; Fineberg & Ellman, 2013; B. J. Miller, Buckley, Seabolt, Mellor, &

Kirkpatrick, 2011; Munkholm et al., 2013; Müller et al., 2012; Rao et al., 2010; Söderlund et al.,

2011). Similar findings in cytokine alterations in BD and SCZ may be due to symptoms that exist in

both of these disorders, such as psychosis (B. J. Miller et al., 2011). Moreover, it is important to note

that expression of cytokines is influenced by many different factors in addition to the inflammasome,

such as apoptosis, neurotransmitter imbalance (Berk et al., 2013; Dheen et al., 2007; Felger et al.,

2013; Felger & Miller, 2012) and other inflammatory pathways such as those involving

phospholipase A2 (Weerasinghe, Rapoport, & Bosetti, 2004), all of which have been implicated in

both BD and SCZ (Berk et al., 2007; Jarskog, 2006; H. W. Kim, Rapoport, & Rao, 2011; Rao et al.,

2012). Hence, while activation of the NLRP3-inflammasome may be contributing to the activation

of theses cytokines in patients with BD, the aforementioned factors could be playing a much larger

role in the activation of cytokines in patients with SCZ. These studies, in addition to our findings,

demonstrate the complex involvement of the inflammatory system in both of these conditions,

indicating that extensive future studies are required to further elucidate this interaction.

Limitations of this study include small sample size, which limits generalizability. Repeating

these experiments in a larger sample will be useful in validating the findings reported in this study.

Also, pre and postmortem factors such as agonal states and storage conditions may have influenced

51

the results (Chandana, Mythri, Mahadevan, Shankar, & Srinivas Bharath, 2009; Vawter et al., 2006).

To verify the effect of such factors, we correlated our findings with PMI, sex and age. Also, we did

not directly measure the activity of the NLRP3-inflammasome. While it has been demonstrated in

vitro that components of the NLRP3-inflammasome associate with the mitochondria and can be

found in the mitochondrial fraction when the inflammasome is activated (Zhou et al., 2011), it

should be noted that immunocontent does not always represent activation. In addition, as patients

may have been taking medications that could have influenced the results, it would be important to

verify these findings in a large sample with medication information available to examine their effect

on complex I and the NLRP3-inflamamsome. Moreover, we only examined the frontal cortex.

Examination of other areas implicated in psychiatric disorders may yield interesting results.

While chronic inflammation of the brain has been consistently demonstrated in BD, its

causes remain largely unknown. (Leboyer et al., 2012). This study is the first to demonstrate

activation of the NLRP3-inflammasome in the frontal cortex of patients with BD, implicating

inflammation as a downstream target of complex I dysfunction in BD and linking two of the most

commonly reported alterations in this disorder - mitochondrial abnormalities and inflammation

(Andreazza et al., 2010; Berk et al., 2011; Clay et al., 2010). Future studies are required to further

elucidate this pathway, and the extent of its contribution to inflammation in brain tissue in patients

with BD. For example, since BD is conceptualized as progressive (Berk et al., 2011), repeating

these experiments in a larger sample that allows for examination of the extent of NLRP3-

inflammasome activation in relation to age or number of episodes may yield interesting findings.

The effects of past or current treatment will also be of great interest. Furthermore, given that

identification of biomarkers of BD is of great interest, measuring the NLRP3-inflammasome in

peripheral samples to identify its potential to be used as a biomarker may be beneficial. Lastly,

52

examination of the specific molecules that are involved in the complex I-NLRP3-inflammasome

pathway may contribute to a better understanding of the pathophysiology of BD.

53

Table 1. Subject information

Control (N=9) BD (N=9) SCZ (N=10)

Age: mean (range) 79.0 (65-91) 76.1 (58-92) 76.8 (55-91)

PMI: mean (range); hr 21.4 (13.0-30.1) 20.9 (10.0-32.9) 19.4 (8.0-43.5)

Male 4 4 4

Female 5 5 6

54

Figures

Figure 1. Levels of electron transport chain components complex I, II, III, IV and V, and

nicotinamide nucleotide transhydrogenase (NNT) in the post-mortem frontal cortex of patients with

BD, SCZ and non-psychiatric controls (CTL). Results for complexes I-V and NNT are expressed in

median fluorescence intensity (MFI). Inset illustrates levels of NDUFS7 measured by western

blotting and the blot for NDUFS7. NDUFS7 was quantified using β-actin as loading control. This

image was cropped to improve clarity. Differences between groups were analyzed using one-way

ANOVA followed by Least Significant Difference test (*p < 0.05).

55

Figure 2. Levels of NLRP3-inflammasome components NLRP3 (A), ASC (B) and caspase 1 (C) in

the brain homogenate (homogenate), cytoplasmic and crude mitochondrial fractions (mitochondrial)

in the post-mortem frontal cortex of patients with BD, SCZ and non-psychiatric controls (CTL).

Insets in panel A illustrate blots for NLRP3. The images were cropped to improve clarity. Arrows

indicate the NLRP3 band at 110kDa, which was quantified using β-actin as a loading control for the

brain homogenate and cytoplasmic fraction, and total protein (memcode) as loading control for the

crude mitochondrial fraction. Loading controls are shown in red squares below the blots with arrows.

Differences between groups were analyzed using one-way ANOVA followed by Least Significant

Difference test (*p < 0.05; **p < 0.01).

56

Figure 3. Levels of inflammatory factors downstream of NLRP3-inflammasome activation in the

IL-1β pathway, IL-1β, IL-6, & TNF-α, and in the IL-18 pathway, IFNγ & IL-10 in the post-mortem

frontal cortex of patients with BD, schizophrenia (SCZ) and non-psychiatric controls (CTL).

Differences between groups were analyzed using one-way ANOVA followed by Least Significant

Difference test (*p < 0.05; ** p < 0.01).

57

Figure 4. Correlation between NLRP3 inflammasome components in the brain homogenate (whole),

cytoplasmic fraction (cytoplasm) and the crude mitochondrial fraction (mitochondria) against

complex I levels in the post-mortem frontal cortex of patients with bipolar disorder.

58

Statement of significance and impact:

Mitochondrial complex I dysfunction and subsequent production of ROS can have important

consequences when ROS act as signaling molecules to activate downstream pathways through redox

sensors. This study first addressed Aim 2, which was to confirm complex I dysfunction in the post-

mortem brain of patients with BD. By examining levels of all components of the electron transport

chain, this study was able to show that only complex I had between-group differences, where

patients with BD had lower levels of complex I than non-psychiatric controls, suggesting that

complex I dysfunction is specific to BD. Also, this study confirmed the findings of 2 different

studies from our group demonstrating lower levels of NDUFS7, a complex I subunit that is critical

for the efficient transfer of electrons to ubiquinone (Halliwell, 2001). This study also addressed Aim

3, which was to examine the mitochondria as a target of complex I dysfunction by measuring the

activation of the NLRP3-inflammasome, a marker of mitochondrial oxidative stress (Zhou et al.,

2011). Indeed, NLRP3 is activated upon release of mitochondrial ROS from complex I inhibition,

making it a suitable marker for oxidative stress in the mitochondria. To measure the activation of the

NLRP3-inflammasome and not just the immunocontent, I measured the levels of NLRP3, ASC, and

caspase-1 in the crude mitochondrial fraction as well as in the brain homogenate, as the components

of the NLRP3-inflamamsome only associate with the mitochondria when the inflamamsome is

activated. The findings of this study suggested activation of the NLRP3-inflammasome in the frontal

cortex of patients with BD. Since activation of the NLRP3-inflammasome complex is strongly

influenced by mitochondrial production of ROS, and occur as a direct result of complex I inhibition

(Zhou et al., 2011), these results demonstrate that complex I dysfunction may be contributing at least

in part to the activation of the inflammatory system in patients with BD. It is also interesting to note

that inflammatory factors that are less specific to the NLRP3-inflamamsome, such as caspase-1 and

inflammatory cytokines, were found to be increased in both BD and schizophrenia, although

59

evidence supporting NLRP3-inflammasome activation was stronger for BD. These findings suggest

that mitochondria may be a downstream target of complex I dysfunction specifically in BD, as

shown in the activation of the NLRP3-inflammasome, while inflammation may be less specific to

BD.

The next experiment aimed to study the synapse as a target of complex I dysfunction, which

was shown to be a target of oxidative stress in the same study demonstrating oxidative damage in the

mitochondria (Andreazza et al., 2013). We focused on the dopaminergic synapse, as increased

dopamine signaling is known to underlie mania, the defining feature of BD (Berk et al., 2007; H. K.

Kim & Andreazza, 2012; Post, Jimerson, Bunney, & Goodwin, 1980). Hence, the next study

examined the dopamine system as a target of complex I dysfunction and oxidative stress by

measuring oxidation and nitration of two dopaminergic proteins that were shown to be targets of

oxidative modification in in vitro and animal studies, the dopamine transporter and the tyrosine

hydroxylase (Blanchard-Fillion et al., 2001; S. U. Park et al., 2002).

60

CHAPTER 4

Title: Oxidation and nitration in dopaminergic areas of the prefrontal cortex from patients with

bipolar disorder and schizophrenia

Authors: Helena Kyunghee Kim, Ana Cristina Andreazza, Pok Yik Yeung, Cameron Isaacs-

Trepanier, L. Trevor Young

Journal: Journal of Psychiatry & Neuroscience

Volume: 39

Pages: 276-285

Work performed by the student:

The student performed all work included in this chapter, including planning and performing the

experiments, image acquisition, data analysis, and preparation and submission of the manuscript.

61

Oxidation and nitration in dopaminergic areas of the prefrontal cortex from patients with

bipolar disorder and schizophrenia

Helena Kyunghee Kim BSc1, Ana Cristina Andreazza Pharm, PhD1,2, Pok Yik Yeung BSc1,

Cameron Isaacs-Trepanier2, L. Trevor Young MD, PhD1,2

1Departments of Psychiatry and Pharmacology, University of Toronto, Canada

2Campbell Family Mental Health Research Institute, Centre for Addiction and Mental Health

Corresponding author: L. Trevor Young

Center for Addiction and Mental Health, Clarke Site

250 College Street, Rm. 835

Toronto, ON

M5T 1R8

Telephone: 416-979-6948

FAX: 416-979-6928

Email: ltrevor.young@utoronto.ca

62

Abstract

Background: Increased oxidative stress is strongly implicated in bipolar disorder, where protein

oxidation, lipid peroxidation, and oxidative damage to DNA have been consistently reported. High

levels of dopamine in mania are also well-recognized in bipolar disorder, and dopamine produces

reactive oxygen species and electron-deficient quinones that can oxidize proteins when it is

metabolized.

Methods: Using immunohistochemistry and acceptor photobleaching FRET, we examined oxidation

and nitration of areas immunoreactive for the dopamine transporter and tyrosine hydroxylase in the

postmortem prefrontal cortex from patients with bipolar disorder, schizophrenia, major depression,

and non-psychiatric controls.

Results: We found increased oxidation of dopamine transporter-immunoreactive regions in patients

with bipolar disorder (F3,50=7.23, P<0.01; P<0.01), while decreased nitration of tyrosine

hydroxylase-immunoreactive regions was found in both patients with bipolar disorder (F3,46 = 3.59,

P<0.05; P<0.01) and schizophrenia (P<0.05) . On the other hand, global levels of oxidation were

found to be increased in patients with bipolar disorder (F3,46=5.91, P<0.01; P<0.01) and

schizophrenia (P<0.05), although nitration levels did not differ between groups (F3,46 = 1.75;

P=0.17).

Limitations: Limitations of this study include the use of postmortem brain sections, which may be

affected by factors such as postmortem interval and antemortem agonal states, although

demographic factors and postmortem interval were accounted for in statistical analysis.

Conclusion: These findings suggest alterations in levels of protein oxidation and nitration in

dopamine-rich regions of the prefrontal cortex in patients with bipolar disorder and schizophrenia,

but more markedly in bipolar disorder.

63

Introduction

Bipolar disorder (BD) is characterized by recurrent episodes of mania/hypomania and

depression, affects 1.5% of the population, and is associated with high morbidity and mortality

(Kupfer, 2005). The pathophysiology of BD is linked to a number of factors, including

neurotransmitter imbalance, oxidative stress and genetic causes among others (Salvadore et al.,

2010). Increased oxidative stress, which could result in oxidative and nitrosative damage to

biomolecules (Halliwell, 1992), is a consistent finding in BD. For example, mitochondrial

dysfunction and decreased expression of genes of the electron transport chain, particularly that of

complex I, are reported in patients with BD (Andreazza et al., 2010; Scola et al., 2013). Decreased

efficiency of the electron transport chain could result in increased production of reactive oxygen

species (ROS) (Halliwell, 1992). Moreover, increased levels of carbonyl groups, 3-nitrotyrosine and

decreased levels of antioxidants, such as glutathione, are all reported in BD (Andreazza et al., 2009;

Andreazza et al., 2008; Gawryluk et al., 2011).

It is widely held that dysregulation of the dopamine (DA) system is important in this disorder,

where high levels of DA is thought to underlie mania, which is a defining feature of BD (Berk et al.,

2007). Indeed, increasing synaptic DA levels with amphetamine and L-DOPA produces mania-like

behaviour (Berk et al., 2007), and antipsychotics, which act in part by blocking DA receptors, are

among the most effective treatments for acute mania (Cipriani et al., 2011). High levels of DA can

be cytotoxic in part through the generation of ROS and electron-deficient quinones during oxidation

of DA (Hastings et al., 1996). In fact, amphetamine increases markers of oxidative stress such as

lipid peroxidation and protein oxidation in animals (Frey et al., 2006; Valvassori et al., 2010).

Therefore, in this study, we examined oxidative and nitrosative damage in dopamine-rich

regions of the prefrontal cortex (PFC) using acceptor photobleaching forster resonance energy

64

transfer (apFRET). We examined the PFC, as it was previously shown to have increased oxidative

stress in patients with BD (Andreazza et al., 2010). In order to examine dopamine-rich areas, we

labelled the dopamine transporter (DAT), which is a transmembrane protein responsible for the

uptake of synaptic DA into the presynaptic terminal (Akil et al., 1999; Ciliax et al., 1999), and

tyrosine hydroxylase (TH), which converts L-tyrosine to L-DOPA, and is the rate-limiting enzyme

in DA synthesis. Previous studies used TH and DAT labeling with immunohistochemistry

techniques to visualize dopaminergic axons in the PFC, and reported extensive co-localization

between TH and DAT immunoreactive axons (Akil et al., 1999; Ciliax et al., 1999). Here, we report

evidence of oxidative and nitrosative damage in DA-rich regions of the post-mortem PFC of patients

with BD and schizophrenia (SCZ).

Methods

Post-mortem brain samples

Frozen post-mortem PFC (BA9/46) sections (14µm) were from the Stanley Foundation

Neuropathology Consortium (Table 1). The details of these samples have been published elsewhere

(Torrey, Webster, Knable, Johnston, & Yolken, 2000). Briefly, subject groups consisted of patients

with BD, major depressive disorder (MDD), SCZ, and non-psychiatric controls. There were no

differences for sex (F3,55 = 0.11, p = 1.00), age (F3,55 = 0.53, p = 0.67), post-mortem interval (F3,55 =

1.70, p = 0.18; PMI) and pH (F3,55 = 0.63, p = 0.60) between groups (N=15 per group). Diagnoses

were established using DSM-IV criteria. Sections with good structural integrity were used. Samples

were randomly coded numerically, and the experimenters were kept blind to the codes until all the

experiments were completed.

Immunohistochemistry

65

Sections were nissl stained and labelled for NeuN to examine tissue quality prior to

immunohistochemistry experiments (methods in supplementary information; Figure 1A&B). Four

adjacent sections per subject were processed for immunohistochemistry, where 2 sections were

labelled for DAT and free thiols, and the other 2 were labelled for TH and 3NT. FRET analysis and

intensity analysis were performed simultaneously on the same sections. 7-Diethylamino-3-

(4’maleimidylphenyl)-4-methylcoumarin (CPM; Invitrogen, Burlington, ON; D346) is a specific

label for free thiols (Mastroberardino, Orr, Hu, Na, & Greenamyre, 2008). CPM labeling was

performed using a previously published method by Mastroberardino et al. (Mastroberardino et al.,

2008) with slight modifications. We used CPM to directly label free thiols without prior

modification of the tissue to minimize alterations to the sections and introduction of modifications

from the experimental procedure. The indirect method by Mastroberardino et al., in which free thiols

are blocked and disulfide linkages are reduced to be labelled with CPM, was used to increase

sensitivity (Mastroberardino et al., 2008). For our samples, direct labeling method was sufficient to

detect between-individual differences. In the methods described here, lower CPM labeling is

indicative of greater oxidation of thiols. Sections fixed with 4% paraformaldehyde (10min) were

incubated in 0.5mM CPM in Tris-HCl (0.1M pH 6.8) for 1h. Sections were washed extensively and

immunohistochemistry (IHC) was carried out using previously published techniques (Waldvogel,

Curtis, Baer, Rees, & Faull, 2006). Sections were blocked in 4% goat serum for 1h, and incubated

with the DAT antibody (Millipore, Temacula, CA; MAB369; 1:250; 60h at 4°C) in 0.05% Triton-X.

The specificity of this antibody was demonstrated in previous studies (G. W. Miller et al., 1997;

Wills et al., 2010). Sections were then washed, incubated with Alexa Fluor®488 anti-rat (Molecular

Probes, Burlington, ON, Canada A11006) in 0.05% Triton-X for 3h, and mounted in Fluoromount™

(Cedarlane, Burlington, ON, Canada). To create a negative control for CPM labeling,

paraformaldehyde-fixed sections were alkylated with 100mM N-ethylmaleimide (NEM; Bioshop,

66

Burlington, ON; ETM222) and 100mM iodoacetamide (IAA; Sigma-Aldrich, Oakville, ON; I6125)

in Tris-0.1 M, pH 6.8 (15min) prior to CPM labeling. Oxidation of thiol groups with NEM and IAA

diminished CPM labeling, demonstrating specificity of CPM for free thiols. Hence, decreased CPM

labeling is indicative of greater thiol oxidation. These sections were used as controls to account for

non-specific labeling of CPM (Mastroberardino et al., 2008). For 3-nitrotyrosine and TH double

labeling, acetone-fixed sections were blocked with 10% goat serum in PBS-0.3% Triton X-100

(PBS-T) for 30min. Sections were then incubated with primary antibodies in PBS-T (48h, 4˚C): TH

(Millipore; AB152; 1:250) and 3NT (abcam, Cambridge, MA, USA; ab61392; 1:200). Sections were

washed, incubated with secondary antibodies (2h, RT) in PBS-T (1:300): Alexa Fluor®488 anti-

rabbit IgG (Molecular probes®; A11008) and Alexa Fluor®350 anti-mouse IgG (Molecular probes®;

A11045), and mounted in FluoromountTM (Cedarlane). Specificity of the anti-TH antibody was

demonstrated elsewhere (O'Connell, Matthews, Ryan, & Hofmann, 2010). The anti-3NT antibody

was previously used by our laboratory (Andreazza et al., 2009). To test the specificity of the anti-

3NT antibody, sections were treated with peroxynitrite (Millipore, 20-107; 1:100), which is used as a

positive control for 3NT labeling, or degraded peroxynitrite (Millipore, 20-247; 1:100) using a

previously published method (P. Zhang, Wang, Kagan, & Bonner, 2000). An increase in 3NT

labeling was observed only with peroxynitrite, demonstrating the specificity of the anti-3NT

antibody (Figure S1). To minimize the risk of antibody cross-reactivity, different combinations of

primary and secondary antibodies were tested, where only minor and non-specific binding was found,

indicating minimal cross-reactivity (Figure 1C&D).

Intensity Analysis

Intensity analysis was performed using a previously published method from our group (Che et al.,

2010). 60X water immersion objective and a confocal laser scanning microscope (CLSM; Olympus

67

Fluoview FV1000; Olympus America INC, Melville, NY, USA) were used to capture images from 5

fields of each section, which were chosen by dividing the section into 5 equally sized rectangles, and

imaging the middle region of each of the 5 rectangles. For CPM and DAT labelled sections, CPM

was excited with the 405nm laser (15% intensity) and Alexa488 was excited with the 488nm laser

(18% intensity). For 3NT and TH labelled sections, Alexa488 was excited with a 488nm laser (15%

laser intensity) and Alexa350 was excited with a 405nm laser (5% intensity). Detection wavelength

for Alexa488 was 500-600nm, and for Alexa350/CPM was 425-475nm. Confocal aperture diameter

was 110μm, and line-by-line sequential detection mode was used. Scanning parameters were kept

consistent for all experiments. Image analysis was performed using the FV10 – ASW 2.1 software

(Olympus). For analyzing DAT, TH, and 3NT intensity, first, background was subtracted from the

image by selecting 5 regions per field without specific immunoreactivity as background and

measuring the average intensity of the regions. This value was subtracted from the overall intensity

of the field. Then the average intensity of the 5 fields was calculated to represent the intensity for the

section. Therefore, DAT/TH/3NT intensity = (Average intensity of 5 fields from the section) –

(Average background intensity of 5 fields from the section). CPM intensity was determined using

the IAA and NEM-treated sections (negative control) to correct for non-specific labeling

(Mastroberardino et al., 2008) which had significantly decreased CPM labeling, demonstrating its

specificity for free thiols. Therefore, CPM intensity = (Average intensity of 5 fields from the section)

– (Average intensity of 5 fields from negative control).

Forster Resonance Energy Transfer (FRET)

Acceptor photobleaching FRET measures non-radiative energy transfer from a donor fluorophore to

an acceptor fluorophore by observing the fluorescence of the donor in the presence and absence

(photobleaching) of the acceptor (König et al., 2006). apFRET has been successfully used in

68

previous studies with human samples (Hynd, Lewohl, Scott, & Dodd, 2003; Keese et al., 2005;

Sharma et al., 2001). The energy transfer between a donor and an acceptor fluorophore only occurs

when the donor and the acceptor are within ~30-50nm of each other using indirect

immunohistochemistry, which allows us to use this technique to detect the proximity of the

molecules (König et al., 2006). For DAT and CPM FRET, Alexa488 (DAT) was the acceptor

fluorophore, and CPM was the donor fluorophore. For 3NT and TH FRET, Alexa488 (TH) was the

acceptor fluorophore, and Alexa350 (3NT) was the donor fluorophore. Acceptor photobleaching

FRET was performed with a CLSM (Olympus) using previously published techniques with small

modifications (König et al., 2006; Mastroberardino et al., 2008). 60X water immersion objective was

used. Image acquisition parameters were identical to those used for intensity analysis. Intensity of

the donor fluorophore before and after the photobleaching of the acceptor fluorophore was measured.

Photobleaching of the acceptor fluorophore was at 100% laser intensity (10μs/pixel) using the

488nm laser in a single bleaching step (1.5s for TH&3NT, 60ms for DAT&CPM). Scanning

parameters were kept consistent during all experiments. Change in fluorescence was represented as

increase in fluorescence (ΔIF), where ΔIF = intensity of donor after photobleaching – intensity of

donor before photobleaching. To account for non-specific binding of CPM, negative controls were

performed where sections were treated with IAA and NEM prior to CPM staining and

immunohistochemistry as detailed above. Then, ΔIF of the negative controls was subtracted from the

ΔIF determined for the samples to calculate the final ΔIF (Mastroberardino et al., 2008). ΔIF

between CPM and the secondary antibody (Alexa488) without the primary antibody against DAT

was undetectable. To test the effect of antibody cross-reactivity on FRET measurements for 3NT and

TH FRET, sections were labelled with only the primary antibody for TH and both secondary

antibodies, then FRET was measured. A very low level of FRET was detected (Average ΔIF = 12.5),

indicating negligible cross-reactivity. Five regions of interest (ROI) per section were chosen as

69

described in intensity analysis and examined. Therefore, ΔIF = (Average ΔIF of 5 ROIs in section) –

(Average ΔIF of 5 ROIs in negative control). We used ΔIF to measure energy transfer, which was

demonstrated to be more appropriate with tissue sections than FRET efficiency due to its lower

sensitivity to non-specific binding of fluorophores (König et al., 2006). For DAT&CPM FRET,

lower ∆IF indicates lower levels of free thiols in DAT-immunoreactive (IR) regions, and thus likely

increased oxidative stress status in DAT-IR regions. For TH&3NT FRET, higher ∆IF indicates

greater 3NT levels in TH-immunoreactive regions. All calculations were performed using FV10 –

ASW 2.1.

Mouse brain study for the effect of post-mortem interval on oxidation of neurons

Adult male C57/BL6 mice (8-12 weeks) were used to examine the effect of PMI on thiol oxidation

and 3-nitrotyrosine formation in neurons. Use of animals was approved by the University Animal

Care Committee of the University of Toronto (Protocol number: 20009477), and care of animals was

performed according to animal research guidelines of the University of Toronto. Animals were

euthanized by asphyxiation and kept at 4°C until an appropriate PMI was reached (6, 24, 48, 72h) to

mimic the storage of cadavers (Hynd et al., 2003) The brain was then extracted, sagittally bisected,

flash frozen in isopentane and dry ice, and stored at -80°C (Torrey et al., 2000). Each bisected brain

was processed and analyzed independently. 14µm sagittal sections were made using a cryostat, and

using standard immunohistochemistry techniques, acetone and methanol fixed sections were labeled

with NeuN (Millipore, Temacula, CA; MAB377B) and CPM, or 3NT (abcam, Cambridge, MA;

ab61392) and NeuN. The secondary antibodies were streptavidin-Alexa Fluor®568 (Molecular

probes®, Burlington, ON; S11226) and Alexa Fluor®350 anti-mouse IgG (Molecular probes®;

A11045). Controls for CPM and 3NT labeling done as described above. Sections were then used to

obtain images of the frontal cortex with a CLSM (Olympus) using a 60X water immersion objective.

70

CPM and Alexa350 were excited with the 405nm laser (32% intensity) and Alexa568 was excited

with the 543nm laser (38% intensity). Detection wavelength for Alexa568 was 555-655nm, and for

CPM and Alexa350 was 425-475nm. The images were analyzed using co-localization analysis

(Pearson’s correlation coefficient) between NeuN and CPM (free thiols in neurons) or between

NeuN and 3NT (protein nitration in neurons) with FV10-ASW2.1.

Statistical Analysis

Statistical analysis was performed using SPSS version 20. Normal distribution of data was

determined using Kolmogorov-Smirnov test. As the data were normally distributed, parametric tests

were used for further analysis. Effect of covariates (age, gender, PMI, and pH) was assessed using

Pearson’s correlation test. As age was the only variable that significantly correlated with one of our

dependent variables (DAT intensity; Pearson r = 0.34, P<0.05), ANCOVA model was used with age

as a covariate, and Dunnett’s post-hoc test were used to examine the effect of diagnosis. Data are

presented as mean±standard error of the mean (SEM).

Results

Subject characteristics

Subjects consisted of patients with BD (n=15), SCZ (n=15), and MDD (n=15), and non-psychiatric

controls (n=15). Demographic details can be found in Table 1. Age, post-mortem interval and brain

pH did not significantly correlate with CPM intensity, 3NT immunoreactivity, TH immunoreactivity,

∆IF between DAT&CPM, and ∆IF between 3NT&TH (Figure 4). DAT immunoreactivity did not

correlate with PMI or pH, but positively correlated with age (Pearson r = 0.34, P<0.05).

Furthermore, psychiatric patients with a history of substance abuse had lower DAT intensity

(F2,46=6.74; P<0.01) compared to healthy controls (P<0.01). We nissl stained the sections to examine

71

their quality. Sections were found to have good histological quality without large holes or tears

(Figure 1A). Also, the average glia to neuron ratio (1.37) was comparable to what was published

before in previous studies examining the same area (Selemon, Rajkowska, & Goldman-Rakic, 1998,

2004).

DAT and TH Immunoreactivity

We examined DAT immunoreactivity and TH immunoreactivity in the sections by measuring

fluorescence intensity. The groups did not differ from each other for DAT immunoreactivity

(F3,44=1.48, P=0.23; Figure 1C). TH immunoreactivity (F3,46 = 0.86, P=0.47; Figure 1D) also did not

differ between groups.

CPM Intensity and 3-nitrotyrosne (3NT) immunoreactivity

Intensity of CPM was quantified to assess the amount of free thiols in the sections. There was a

significant difference between groups (F3,44=6.74, P<0.01; Figure 2A), where patients with BD

(P<0.01) and SCZ (P<0.05) had lower CPM intensity in the PFC compared to the control group.

Age, gender, PMI, and pH did not correlate with CPM or 3NT intensity. To explore the potential

effect of psychotropic drugs, we compared psychiatric patients who were prescribed lithium,

antipsychotics, or antidepressants at the time of death with patients who were not prescribed drugs

and also with healthy controls. Patients who were prescribed antipsychotics at the time of death were

found to have lower CPM intensity (F2,44=7.69, P<0.01) compared to healthy controls (P<0.01). On

the other hand, patients who were not prescribed antidepressants had lower CPM intensity

(F2,44=8.31, P<0.01) compared to controls (P<0.01). There were no other effects of drug treatment on

either CPM intensity or 3NT immunoreactivity. 3NT immunoreactivity was also assessed by

measuring fluorescence intensity. In contrast to CPM intensity, 3NT immunoreactivity did not differ

72

between groups (F3,45 = 1.80, P=0.16; Figure 2B). Because protein stability may be affected by

longer post-mortem intervals, we decided to examine the effect of longer PMIs on the frontal cortex

using mouse brain. We chose to examine neurons as they have lower antioxidant capacity than glial

cells 3. CPM (F3,12=1.30, P=0.32) and 3NT labeling (F3,12=0.82, P=0.82) in neurons did not change

with increasing PMI up to 72h.

Acceptor photobleaching FRET analysis

Initially, our aim was to examine oxidation and nitration in both TH and DAT immunoreactive

regions using FRET analysis. However, FRET between CPM and Alexa488 labeling TH was not

detectable in our sections. Moreover, FRET between Alexa350 labeling 3NT and Alexa488 labeling

DAT was not detectable. Therefore, we examined free thiols in DAT-immunoreative regions

(DAT&CPM FRET), and nitration of TH-immunoreactive regions (TH&3NT FRET). There was a

significant difference between groups for DAT&CPM FRET (F3,48=6.76, P<0.01), where patients

with BD (P<0.01) were found to have lower energy transfer between CPM and DAT(Alexa488) than

the control group (Figure 3A). Also, subject groups were found to significantly differ from each

other for TH&3NT FRET (F3,45 = 3.10, P<0.05; Figure 3B), where patients with BD (P<0.01) and

SCZ (P<0.05) were found to have lower levels of nitration than the control group. Similarly, to

explore the potential effect of psychotropic drugs, we compared patients who were prescribed

antipsychotics, antidepressants, or lithium with those who were not prescribed the drugs and also

with healthy controls. The only significant effect was that patients who were not prescribed

antidepressants had lower TH&3NT FRET (F2,42=3.26; P<0.05) than healthy controls (P<0.05),

although this effect was not found comparing those who were prescribed antidepressants.

Discussion

73

In this study, we examined the relationship between oxidative stress and dopamine in BD by

measuring oxidative and nitrosative damage in DA-rich areas of the PFC. Using FRET, we found

increased oxidation of DAT-immunoreactive areas in patients with BD, while nitration of TH-

immunoreactive areas was found to be decreased in patients with BD and SCZ. When we examined

global levels of oxidative and nitrosative damage, however, we did not find between-group

differences in levels of nitration, but found thiol oxidation to be increased in both patients with BD

and SCZ. These results suggest alterations in oxidation and nitration of DA-rich areas in the PFC

that are particularly marked in BD, but also marked in SCZ.

When oxidation and nitration of DAT and TH immunoreactive (IR) regions were examined

using FRET, we found increased oxidation in DAT-IR areas in patients with BD while nitration in

TH-IR areas was found to be decreased in both patients with BD and SCZ. These results suggest

oxidative and nitrosative modifications to DA-rich areas in both BD and SCZ, which could be

contributing to the dysregulation of the DA system hypothesized to occur in both of these disorders.

More specifically, hyperactivity of the DA system in mania is well-recognized, where

antipsychotics, which act in part by blocking DA signaling, is one of the most effective treatments

for acute mania (Cipriani et al., 2011), and dopaminergic agonists were found to produce mania-like

behavior in humans and in animals (Berk et al., 2007; H. K. Kim & Andreazza, 2012). The DA

hypothesis for SCZ is more complex, however, as the direction of DA dysregulation may differ

depending on symptoms (positive vs. negative) and brain region (cortical vs. subcortical) (Howes &

Kapur, 2009). However, due to the limitations of using postmortem brain sections, future studies

examining the consequences of these alterations to DA signaling and behavior are required.

Furthermore, it is interesting to note that patients with BD and SCZ were found to have similar

alterations in our study, which is consistent with previous studies showing extensive overlap

74

between these two disorders including cognitive deficits (Schretlen et al., 2007), genetic alterations

(Potash & Bienvenu, 2009), changes in brain morphology (Maier, Zobel, & Wagner, 2006), and

oxidative stress (Ng et al., 2008). Our results also showed that patients who are not prescribed

antidepressants have lower TH&3NT FRET than healthy controls. This effect is difficult to interpret.

While antidepressants have been shown to have neuroprotective effects through other mechanisms

than oxidative stress including regulation of neurotrophin levels (Saarelainen et al., 2003), it is not

known whether the reduction in oxidative stress in patients treated with antidepressants reflect their

neuroprotective effects. A larger issue relates to the difficulty in disentangling the effects of

medication and diagnosis since antidepressants in most cases were prescribed to patients with

depression but not to those with BD or SCZ.

Acceptor photobleaching FRET has a resolution of ~30-50nm, and the effectiveness of FRET

increases as the distance between the fluorophores decreases (König et al., 2006), suggesting that

oxidation and nitration in DAT and TH immunoreactive areas as measured by FRET may in large

part include direct oxidation of the DAT, and nitration of the TH. Indeed, previous studies have

demonstrated that DAT is a target of thiol oxidation but not nitration (S. U. Park et al., 2002), which

may have contributed to why we could not measure FRET between the fluorophores labeling DAT

and 3NT (ΔIF undetectable). As DAT oxidation decreases its ability to uptake DA (H. K. Kim &

Andreazza, 2012; S. U. Park et al., 2002), increased DAT oxidation found in patients with BD may

contribute to increased levels of synaptic DA. Whether TH is primarily a target of nitration or

cysteine oxidation is controversial, with some studies showing evidence for both (S. Park, Geddes,

Javitch, & Kuhn, 2003) and other studies reporting TH nitration and oxidation only when the protein

is unfolded (Ara et al., 1998; Blanchard-Fillion et al., 2001). In our study, we could not detect FRET

between CPM labeling free thiols and TH. Since it is physiologically unlikely that all the cysteine

residues of TH were in the oxidized state, we believe that CPM may not have been able to access the

75

thiol groups of the TH protein as they are highly shielded from the surrounding environment

(Daubner et al., 2011). Decrease in 3NT levels in TH-IR regions was an unexpected finding, as

increased oxidative stress is strongly implicated in BD (Andreazza et al., 2009). This finding may be

specific to the TH protein or to L-DOPA-rich areas, since DA, L-DOPA and DOPAC are favorable

substrates for reactive nitrogen species, and may protect proteins in their proximity that are less

favorable substrates for nitration (S. U. Park et al., 2002). Future studies are required to elucidate

this further. Also, because TH nitration was shown to decrease its ability to produce L-DOPA

(Blanchard-Fillion et al., 2001; S. U. Park et al., 2002), decreased nitration of TH in BD and SCZ

suggest that nitration of TH may be present in normal conditions to limit the activity of this enzyme.

Consequently, decreased nitration of TH in BD and SCZ could act to ‘disinhibit’ TH and result in

higher levels of presynaptic DA. Also, while TH and DAT immunoreactive fibers co-localize in the

human prefrontal cortex (Ciliax et al., 1999), TH is also present in noradrenergic neurons, suggesting

that decreased nitration in TH-IR regions may be occurring in noradrenergic cells as well.

Interestingly, dysregulation of noradrenergic signaling has also been shown in BD (Manji & Lenox,

2000), suggesting a potential role for decreased TH nitration in the noradrenergic system.

Furthermore, while the focus of this study was the DA system, such changes could also occur in

other neuronal populations. Future studies examining these alterations in different neuronal cells

may increase our understanding of the pathophysiology of BD and SCZ.

Alterations in the prefrontal cortex (PFC) is consistently implicated in BD and other

psychiatric disorders, where studies have reported decreased metabolism, morphological changes

(Almeida et al., 2009), and increased oxidative damage (Andreazza et al., 2010). Moreover, the PFC

has a high relative density of dopaminergic axons compared to other cortical regions (Ciliax et al.,

1999), and DA is involved in the regulation of working memory by the PFC (D'Ardenne et al.,

2012), which is impaired in patients with BD and SCZ (Glahn et al., 2006; Thompson et al., 2007).

76

These studies suggest that dysregulation of the DA system in the PFC may be involved in the

pathophysiology of these disorders. In agreement with previous studies, we found that TH and DAT

labeling reveals fibers, which may be dopaminergic axons. We also saw punctate structures that

were occasionally connected by fibers, which may be axonal varicosities and terminals (Akil et al.,

1999; Ciliax et al., 1999) (Figure 1C&D). DAT and TH immunoreactivity examined by measuring

fluorescence intensity did not differ between groups, which is in agreement with previous studies

examining the same samples as those used in our study (data can be found in the Stanley

Neuropathology Consortium Integrative Database: http://sncid.stanleyresearch.org). We also found

that DAT intensity was lower in patients with a history of substance abuse compared to healthy

controls. Although we do not have details on the substances that were used by the patients, this

finding suggests that drug abuse may influence the dopaminergic system in the prefrontal cortex. It

should, however, be noted that fluorescence intensity is not a truly quantitative measure of

dopaminergic innervation. As our purpose was to visualize DA-rich areas of the PFC and our

sections (frozen, 14µm thick) were not amenable to stereological analysis (West, 2012), we did not

pursue this further using more quantitative methods (Akil et al., 1999; West, 2012).

Global levels of free thiols and 3-nitrotyrosine were examined by measuring fluorescence

intensity. We used CPM to label free thiols using a method validated by Mastroberardino et al.

(Mastroberardino et al., 2008). We chose to use thiol oxidation as a marker of protein oxidation as it

occurs more readily than formation of protein carbonyls, and hence may be a more sensitive measure

(Mastroberardino et al., 2008). 3-nitrotyrosine is a marker of protein nitration that has been used by

other studies examining oxidative stress (Andreazza et al., 2008; Andreazza et al., 2010; Machado-

Vieira, Andreazza, et al., 2007). We found increased thiol oxidation in patients with BD and SCZ,

while we did not find between-group differences in protein nitration. Previous studies have

77

consistently reported increased markers of oxidative stress for both SCZ and BD, where increased

levels of protein carbonyl groups and lipid peroxidation have been found (Andreazza et al., 2008).

Oxidation of thiol groups can alter protein structure and function (Winther & Thorpe, 2013), and

may be a contributing factor in the pathophysiology of these disorders. Increased thiol oxidation in

BD and SCZ may occur due to a number of different reasons, including lower levels of glutathione

(Gawryluk et al., 2011; Kulak et al., 2013) and mitochondrial dysfunction (Scola et al., 2013),

particularly that of complex I, which could increase the production of ROS (Halliwell, 1992) that

can oxidize proteins. Indeed, decreased levels of expression of subunits of the mitochondrial electron

transport chain were reported by a number of studies for both BD and SCZ (Scola et al., 2013).

Since cell loss and morphological alterations are found in the prefrontal cortex in patients with BD

and SCZ (Almeida et al., 2009; K. F. Berman, Doran, Pickar, & Weinberger, 1993; Frey et al., 2007;

Lopez-Larson, DelBello, Zimmerman, Schwiers, & Strakowski, 2002; Rajkowska, Halaris, &

Selemon, 2001), cellular dysfunction produced by protein oxidation may be contributing to these

alterations. This may lead to a decline in the proper functioning of the PFC. Indeed, as mentioned

previously, decreased working memory is consistently reported in BD (Glahn et al., 2006;

Thompson et al., 2007), suggesting that cellular dysfunction produced by oxidative damage could be

contributing to this as well. Furthermore, we found an effect of prescribed medications, where

patients who were prescribed antipsychotics had greater oxidative damage than the control group.

Since antipsychotics are D2 receptor antagonists, the effect of these drugs on oxidative damage in

the dopamine system is of potential interest (H. K. Kim & Andreazza, 2012). Also, patients who

were not prescribed antidepressants had greater global oxidative damage, which is consistent with a

potential protective effect of these drugs as described above for TH&3NT FRET results. 3-

nitrotyrosine, which is formed when reactive nitrogen species, such as peroxynitrite, react with

tyrosine residues (Greenacre & Ischiropoulos, 2001), was used as a marker of protein nitration. We

78

chose to label 3-nitrotyrosine to examine protein nitration as it has been successfully used by our

group as well as others to examine oxidative stress in post-mortem brain (Andreazza et al., 2007;

Andreazza et al., 2008; Andreazza et al., 2010). While we did not find global 3NT levels to differ

between groups, a previous study showed increased 3NT levels in the PFC of patients with BD and

SCZ (Andreazza et al., 2010). The difference between these findings may be due to the use of

different techniques, as Andreazza et al. measured 3NT levels using the ELISA technique from

whole tissue homogenates while we used immunohistochemistry to label cells in tissue sections.

Moreover, Andreazza et al. examined BA10 while we examined BA9/46. BA9 and BA10 have

different cytoarchitecture, suggesting that they may differ in the type and abundance of cells

(Semendeferi, Armstrong, Schleicher, Zilles, & Van Hoesen, 2001). Since different cells have

different vulnerabilities to oxidative stress, this may have contributed to the difference between these

two areas (Halliwell, 1992). Nonetheless, the differences in overall 3-NT levels between diagnostic

groups were in the same direction as those reported earlier by our group (Andreazza et al., 2010).

Also, while global 3NT levels did not differ between groups, nitration of TH-IR regions measured

by FRET was decreased in BD and SCZ. This suggests that nitration may be a more specific

phenomenon where increase or decrease in nitration of certain proteins does not always correspond

with global levels of oxidative stress (Greenacre & Ischiropoulos, 2001).

Limitations and Conclusion

The findings of this study must be interpreted in light of their limitations. First, we had a

small sample size and used multiple comparisons for statistical analysis. Second, we measured

fluorescence intensity in order to compare DAT, TH, and 3NT immunoreactivity between groups.

Fluorescence intensity obtained from tissue sections labelled with antibodies is a relative measure

and is limited by factors such as photobleaching and tissue quality. In order to minimize the effect of

79

these factors on our results, we kept the scanning parameters consistent, subtracted background

intensity, and only used sections with good histological quality. Also, because we used human

postmortem brain samples, pre and postmortem factors such as antemortem agonal states may have

affected our results (Chandana et al., 2009; Vawter et al., 2006). To account for this, we examined

the effect of demographic factors and post-mortem alterations in data analysis by correlating our

findings with PMI, pH, and age. Furthermore, in agreement with previous studies that examined

PMIs up to 48h (Chandana et al., 2009; Harish et al., 2011), thiol oxidation and nitration in the

mouse frontal cortex did not change up to 72h of PMI. Although we did find an effect of the history

of substance abuse on one measure, DAT intensity, we do not have detailed information on the

history of substance abuse or dependence to further explore how these factors could have influenced

our results. In addition, while the effects of drug treatment on CPM intensity (antipsychotics and

antidepressants) and TH&3NT FRET (antidepressants) are of potential interest, there are several

important limitations. These include small sample size of subjects, the differences in prescribing

patterns across diagnosis and no information of whether the patients were taking the medications

which were prescribed. Other approaches not solely relying on post-mortem brain may be more

helpful to determine drug treatment effects on these measures. We also only examined the prefrontal

cortex. Examining different areas of the brain may help further elucidate the relationship between

DA and oxidative stress in BD and SCZ. Finally, use of other techniques such as mass spectrometry

to examine oxidative and nitrosative modifications to DAT and TH in postmortem brain and

functional assays to examine the cause and consequences of such modifications may aid in further

elucidating the relationship between oxidative stress and DA in BD and SCZ.

In conclusion, our results demonstrate increased protein oxidation in the PFC of patients with

BD and SCZ. We also found increased thiol oxidation in DAT-IR regions of patients with BD and

80

dysregulation of protein nitration in TH-IR regions of patients with BD and SCZ, suggesting

oxidative and nitrosative modifications to DA-rich areas in both of these disorders. Therefore, our

findings suggest that the interaction between oxidative stress and DA may be important for the

pathophysiology of BD and SCZ, and that future studies examining the role of oxidative

modifications on dopaminergic proteins may contribute to a better understanding of how

dysregulation of the DA system occur in these disorders.

Acknowledgements / Conflict of interest disclosure

We acknowledge Ginnie Ng for her contributions in cell counting of the nissl stained sections, and

Dr. Dennis Grant for donating the mice for the post-mortem interval experiment. The authors also

declare CIHR, Brain and Behaviour Research Foundation (NARSAD), and OGS as sources of

funding, and Stanley Foundation Neuropathology Consortium as source of human post-mortem brain

sections. The authors declare no conflict of interest.

81

Figures

Figure 1. A. Example of a nissl-stained section taken with Nikon Eclipse 80i Microscope with good

section quality. Scale = 44µm. B. Example of a section labelled for NeuN (green). Image was taken

with a Nikon Eclipse 80i Microscope. Scale = 58μm. C. Image is an example of a section labelled

for DAT (green). Image was taken with a confocal laser scanning microscope (CLSM). The inset is

an example of a section labelled with only the secondary antibody, demonstrating minimal labeling,

and therefore specificity of the secondary antibody. D. Image is an example of a section labeled with

the TH antibody (green) taken with a CLSM. The inset is an example of a section labelled with the

primary antibody for 3NT, and secondary antibody for TH. Minimal labeling was observed,

indicating specificity of the secondary antibody. E. Image is an example of a FRET experiment for

DAT&CPM FRET. Red circle indicates the area that has been photobleached. Increase in the

intensity of the donor fluorophore (blue; CPM) can be observed after bleaching the acceptor

82

fluorophore (Alexa488; green), indicating the presence of FRET. F. Image is an example of a FRET

experiment for TH&3NT FRET. Red circle indicates the photobleached area. Increase in the

intensity of the donor fluorophore (Alexa350; blue) can be observed after photobleaching the

acceptor fluorophore (Alexa488; green), indicating presence of FRET.

Figure 2. A. Intensity of CPM labeling after correction for non-specific CPM binding with negative

control (IAA&NEM treated sections). Results are expressed as mean±SEM. * P<0.05; ** P<0.01.

N=50 (Control=13, BD=13, SCZ=11, MDD=13). B. 3NT immunoreactivity after correction for non-

specific signals as measured by fluorescence intensity. Results are expressed as mean±SEM. N=50

(Control=14, BD=10, SCZ=13, MDD=13). All analyses were performed on non-processed, original

images in Original Imaging Format (OIF).

83

Figure 3. A. ΔIF between Alexa488 labeling DAT and CPM labeling free thiols. Higher values

indicate greater amounts of free thiols in DAT-IR regions after correction for background ΔIF

(FRET in oxidized sections). Results are expressed as mean±SEM. *P<0.05 **P<0.01. N=54

(Control=14, BD=14, SCZ=12, MDD=14). B. ΔIF between Alexa488 labeling TH, and Alexa350

labeling 3NT. Greater ΔIF indicates greater nitration. Results are expressed as mean±SEM. *P<0.05;

**P<0.01. N=50 (Control=11, BD=11, SCZ=14, MDD=14). All analyses were performed on non-

processed, original images in Original Imaging Format (OIF).

84

Figure 4. A. Correlation between ∆IF for DAT&CPM and brain pH, age, and PMI. B. Correlation

between ∆IF for TH&3NT and brain pH, age, and PMI. Results were assessed using Pearson’s

correlation test. The correlations were not found to be significant.

85

Table 1. Subject information by group

Control BD SCZ MDD

Age: mean (range) 48 (29-68) 42 (25-61) 44.2 (25-62) 46.4 (30-65)

PMI: mean (range); hr 24 (8-42) 33 (13-62) 34 (12-61) 28 (7-47)

pH: mean (range) 6.3 (5.8-6.6) 6.2 (5.8-6.5) 6.1 (5.8-6.6) 6.2 (5.6-6.5)

Sex 9M & 6F 9M & 6F 9M & 6F 9M & 6F

Lithium 0 4 2 2

Antipsychotics 0 8 12 0

Antidepressants 0 8 5 10

86

Statement of significance and impact:

Our group has recently demonstrated that the synapse is a target of oxidative modifications in

the frontal cortex of patients with BD (Andreazza et al., 2013). As the dopamine system largely

exists in the synapse, and dopamine dysregulation has long been known to contribute to mania, a

defining feature of BD (Cousins et al., 2009), I decided to examine oxidative damage to the

dopaminergic system to examine if the synapse may be a downstream target of complex I

dysfunction in BD, addressing Aim 4. More specifically, as in vitro studies have shown the

dopamine transporter and tyrosine hydroxylase, two proteins critical for the regulation of dopamine

levels, as targets of oxidative modifications (Ara et al., 1998; S. U. Park et al., 2002), oxidation of

the dopamine transporter and nitration of tyrosine hydroxylase were measured. Findings of this study

demonstrated oxidative and nitrosative modifications to dopamine rich areas in the post-mortem

prefrontal cortex of patients with BD, where dopamine transporter-immunoreactive regions were

found to have increased thiol oxidation, and tyrosine hydroxylase-immunoreactive regions were

found to have decreased nitration. These findings suggest that oxidative modifications to

dopaminergic proteins could be contributing to dopamine dysregulation underlying BD,

demonstrating the dopamine system as a possible downstream target of complex I dysfunction.

Oxidation of the DAT would lead to decreased ability to uptake dopamine (S. U. Park et al., 2002),

and since DAT is the primary regulator of synaptic levels of dopamine, this would lead to increased

levels of synaptic dopamine levels. Furthermore, nitration of tyrosine hydroxylase was found to

decrease its ability to synthesize L-DOPA, resulting in lower levels of dopamine (Blanchard-Fillion

et al., 2001). This suggests that decreased nitration of tyrosine hydroxylase could “disinhibit” the

enzyme, causing increased synthesis of L-DOPA, although this has not been demonstrated directly.

Therefore, oxidation of DAT and decreased nitration of tyrosine hydroxylase could be contributing

to increased dopamine signaling in mania. Importantly, while global levels of nitration did not differ

87

between groups, patients with bipolar disorder and schizophrenia had lower levels of nitration in

tyrosine hydroxylase-immunoreactive regions. This suggests that oxidative stress can have specific

targets that are affected differently from global levels of oxidative/ nitrosative damage depending on

the target and surrounding conditions (i.e. presence of dopamine). To summarize, my studies

examining the post-mortem brain extended on previous findings from our group identifying the

mitochondria and synapse as targets of oxidative stress (Andreazza et al., 2013) by demonstrating

the activation of the NLRP3-inflammasome in patients of BD, a marker of complex I dysfunction

and mitochondrial ROS production, and by demonstrating oxidative modifications to the dopamine

transporter and tyrosine hydroxylase, which are markers of the dopaminergic synapse (Akil et al.,

1999).

The next set of studies aimed to further examine complex I dysfunction using cell models,

which allows us to directly inhibit the electron transfer process in complex I by using rotenone

(Turrens & Boveris, 1980). We also examined the effect of lithium because it is uniquely used for

the treatment of BD and was shown to improve complex I function and increase its mRNA levels

(Maurer et al., 2009; Sun et al., 2006), making it a good proxy for studying complex I dysfunction in

BD.

88

CHAPTER 5a

Title: Glutathione-mediated effects of lithium in decreasing protein oxidation induced by

mitochondrial complex I dysfunction

Authors: Helena Kyunghee Kim*, Camila Fernandes Nascimento*, L. Trevor Young, Karina

Martinez Mendonca, Beny Lafer, Lea Tenholz Grinberg, Ana Cristina Andreazza

• *HKK and CFN share first authorship

Journal: Journal of Neural Transmission

Volume: Epub ahead of print

Pages: Epub ahead of print

Work performed by the student:

The student performed all work included in this chapter, including planning and performing the

experiments, image acquisition, data analysis, and preparation and submission of the manuscript.

89

Title: Glutathione-mediated effects of lithium in decreasing protein oxidation induced by

mitochondrial complex I dysfunction

Category: Short communication

Authors:

#Camila Fernandes Nascimento1, #Helena Kyunghee Kim2, L. Trevor Young1,2,3, Karina Martinez

Mendonça5, Lea Tenenholz Grinberg1,4, Beny Lafer1, *Ana Cristina Andreazza2,3

#The first two authors (CFN and HKK) share first authorship.

1University of Sao Paulo Medical School, Sao Paulo, Brazil

2Departments of Pharmacology and Psychiatry, University of Toronto, Ontario, Canada

3Centre of Addiction and Mental Health, Ontario, Canada

4Memory and Aging Center, University of California, San Francisco, USA

5Department of Pharmacy, Federal University of Sao Paulo

Addresses:

CFN and BL: Rua Dr. Ovidio Pires de Campos, 785 - 1º andar - Sala 1C015, Cerqueira César,

05403-903 - Sao Paulo, SP – Brasil

HKK, LTY, KMM, and ACA: RM4204, 1 King’s College Circle Toronto, ON Canada

LTG: Sandler Neurosciences Center, 675 Nelson Rising Lane, Suite 190, San Francisco, California,

USA

90

Corresponding author:

Ana Cristina Andreazza

Address: RM4204, 1 King’s College Circle Toronto, ON Canada M5S 1A8

Phone number: 416-946-5722

E-mail: ana.andreazza@utoronto.ca

91

Abstract

The aim of this study was to elucidate whether glutathione is involved in lithium’s ability to decrease

carbonylation and nitration produced by complex I inhibition, which is consistently found in BD.

Neuroblastoma cells were treated with rotenone, a complex I inhibitor. Our results suggest that

glutathione is essential for lithium’s ability to ameliorate rotenone-induced protein carbonylation,

but not nitration.

Key words: 3-nitrotyrosine, carbonylation, lithium, mitochondrial complex I, rotenone

92

1. Introduction

Oxidative stress occurs when the production of reactive oxygen (ROS) or nitrogen (RNS)

species overwhelms the body’s antioxidant defense system (Halliwell, 2001). Mitochondrial

complex I is a major source of ROS and RNS, which form through the reaction between molecular

oxygen and leaked electrons during the process of adenosine triphosphate (ATP) production (Jeong

& Seol, 2008). Superoxide anions that form during this process can become hydroxyl radicals to

produce protein carbonyl groups, or react with nitric oxide to produce peroxynitrite, which can cause

protein nitration (Jeong & Seol, 2008). Both protein carbonylation and nitration can disrupt the

proper functioning of a variety of different proteins (Naoi et al., 2005; Ray, Huang, & Tsuji, 2012).

Glutathione, which is an endogenous antioxidant, can prevent such damages by directly reacting

with oxidizing substances and neutralizing them (Dringen, 2000). Importantly, glutathione is the

main antioxidant of the brain and was found to be decreased in the prefrontal cortex of patients with

BD (Dringen, 2000; Gawryluk et al., 2011). Glutathione peroxidase, which catalyzes the

consumption of free glutathione to reduce reactive oxygen species, was also found to be greater in

relatives of BD patients, suggesting dysfunction of the glutathione system in BD (Huzayyin et al.,

2014). Moreover, the production of oxidative stress through mitochondrial complex I aberrations

including decreased mRNA expression, activity, and protein levels has been strongly implicated in

bipolar disorder (BD) (Andreazza et al., 2010; Konradi et al., 2004; Scola et al., 2013).

Of significance, lithium was demonstrated to have antioxidant properties (Cui et al., 2007),

while whether lithium can prevent oxidative and nitrosative damage to proteins through modulation

of the glutathione system remains to be explored. Elucidating the exact pathway by which lithium

exerts its antioxidant effects could lead to the discovery of novel targets that can be used to develop

drugs for decreasing oxidative stress in BD. Therefore, the goal of this study was to investigate

whether lithium prevents mitochondrial complex I dysfunction-induced protein oxidation and

93

nitration by increasing glutathione levels. We hypothesized that mitochondrial complex I

dysfunction induced by rotenone will increase protein oxidation and nitration, and that lithium will

prevent both modifications by increasing glutathione levels. We tested our hypothesis in the

neuroblastoma cell line (SH-SY5Y).

2. Methods

2.1 Cell culture

Human neuroblastoma cells (SH-SY5Y) were grown in Dulbecco Modified Eagle’s Media (DMEM)

containing 10% fetal bovine serum (FBS) at 5% CO2, 37oC. To induce mitochondrial complex I

impairment, SH-SY5Y cells were treated with different concentrations of rotenone (16h; 5nM, 10nM,

15nM, 20nM and 30nM; Sigma-Aldrich, Oakville, ON). Rotenone was chosen as it has been

successfully used to model complex I dysfunction in neuroblastoma cells (Watabe & Nakaki, 2004).

From here on, rotenone-induced effects will be described as being due to mitochondrial complex I

dysfunction. The effect of lithium was assessed by pre-treating SH-SY5Y cells with a therapeutic

dose of lithium carbonate (72h; 0.75mM; Li; Bioshop, Burlington, ON) prior to rotenone treatment.

Buthionine sulfoximine (BSO; 1mM; Bioshop) was used to deplete glutathione and analyze its role

in the effect of lithium. For this, the cells were pre-treated with lithium (72 hours), followed by BSO

treatment (12 hours) and rotenone treatment (5nM, 15nM, 30nM only; 16 hours). The three different

series of treatments described so far will be referred throughout the text as rotenone alone,

Li+rotenone and Li+BSO+rotenone. All the experiments were performed in starving media

(DMEM+0.5% FBS) for all conditions as suggested by previous studies (Krämer, Bouzakri,

Holmqvist, Al-Khalili, & Krook, 2005; van der Valk et al., 2010; X. Zheng et al., 2006). Starving

media was not shown to decrease mitochondrial activity within 1 week in fibroblasts (Takeda et al.,

2002). In our experiments, mitochondrial function was shown to be preserved in cells treated with

94

starving media as observed by positive MitoTracker Red CMXRos staining, which only accumulates

in actively respiring mitochondria (Takeda et al., 2002).

2.2 Complex I activity

Activity of complex I was analyzed using a colorimetric kit (Abcam, Cambridge, UK) according to

manufacturer’s protocol.

2.3 Cell viability

Cell viability was measured using the CellTiter-blue assay (Promega, Madison, WI, USA) according

to manufacturer’s protocol.

2.4 Immunocytochemistry

Immunocytochemistry was carried out using a previously published method with slight

modifications (Watabe & Nakaki, 2007) to observe 3-nitrotyrosine (3NT). Briefly, cells were

incubated with MitoTracker Red CMXRos for 20min at 37oC (100nM; Life Technologies,

Burlington, ON). Cells were fixed with 4% paraformaldehyde and permeabilized with 0.2% Triton-

X for 5min. Following blocking (4% goat serum, 30min), cells were incubated with anti-3-

nitrotyrosine antibody (16h, 4 oC, 1:200; Abcam, Toronto, ON), followed by Alexa488 anti-mouse

(1h, Life Technologies), and mounted with ProLong Gold antifade reagent with DAPI (Life

Technologies). Carbonylation was visualized by using the OxyICC Oxidized Protein Detection Kit

(Millipore, Temecula, CA) according to manufacturer’s instructions. Mitochondria were visualized

using MitoTracker Red and the cells were mounted in ProLong Gold antifade reagent with DAPI. A

confocal laser scanning microscope (CLSM; Olympus FV1000; Olympus, Center Valley, PA) and a

60X objective were used to image the cells. Alexa488 was observed using a 488nm laser,

95

MitoTracker Red was observed using a 543nm laser, and DAPI was observed using a 405nm laser.

3NT and carbonyl fluorescence intensities were measured from five regions of interest (ROIs) per N

using the FV10-ASW2.1 software (Olympus), and the average value of the ROIs was determined.

All experiments were performed in triplicates.

2.5 Data analyses

Statistical analyses were performed using IBM SPSS 20 software (Chicago, USA). Differences

between the treatment groups (rotenone alone vs Li+rotenone or Li+rotenone vs Li+BSO+rotenone)

were assessed by two-way analysis of variance (ANOVA). Differences between rotenone treatments

against the appropriate control were assessed using one-way ANOVA followed by Dunnett post-hoc

analysis.

3. Results

3.1 Lithium prevents decrease of complex I activity and cell viability induced by mitochondrial

complex I dysfunction

Treating neuroblastoma cells with rotenone produced a decrease in complex I activity (F5,18 = 3.77,

P<0.05; P<0.05 for 30nM; Figure 1A). Lithium pre-treatment was able to prevent this decrease, such

that rotenone no longer produced a decrease in its activity. Dysfunction of mitochondrial complex I

induced by rotenone also demonstrated a dose-dependent effect in decreasing cell viability (F5,24 =

4.09, P<0.01, P<0.05 for 30nM), and lithium pre-treatment prevented this decrease. The addition of

BSO inhibited the ability of lithium to prevent decreased cell viability induced by mitochondrial

complex I dysfunction (F3,16 = 20.64, P<0.01, P<0.01 for all concentrations of rotenone; Figure 1B).

3.2 Lithium’s ability to decrease rotenone-induced protein carbonylation is dependent on glutathione.

96

Mitochondrial complex I inhibition through rotenone treatment produced an increase in protein

carbonylation (F5, 24 = 15.38, P<0.01; P<0.01 for all concentrations; Figure 2). Although lithium pre-

treatment did not result in complete prevention of this increase in carbonylation (F5,24 = 28, P<0.01;

P<0.01 for all concentrations), it produced a decrease in carbonylation compared to cells treated with

rotenone without lithium pre-treatment (F1,48 = 22.23, P<0.01). The addition of BSO inhibited the

ability of lithium to decrease carbonylation such that carbonylation levels became similar or higher

than cells treated with rotenone alone (F3,16 = 25.10, P<0.01; P<0.01 for all).

3.3 Lithium prevents protein nitration induced by mitochondrial complex I dysfunction

independently of glutathione.

Protein nitration was increased by mitochondrial complex I inhibition with rotenone treatment (F5,36

= 9.58, P<0.01; P<0.01 for all rotenone concentrations; Figure 2). Lithium pre-treatment was able to

completely prevent this increase in nitration such that rotenone no longer produced an increase. The

ability of lithium to decrease nitration did not change with the addition of BSO.

4. Discussion

Mitochondrial dysfunction and increased oxidative stress are two of the most consistent

alterations reported in patients with BD (Andreazza et al., 2010; Clay et al., 2010; Scola et al., 2013)

and lithium, which is considered a gold standard of mood stabilizers (Goodwin & Malhi, 2007), was

shown to have antioxidant-like effects (Cui et al., 2007). Yet, the mechanisms underlying these

effects are poorly understood. In agreement with previous studies, we found that lithium prevented

rotenone-induced cell death and decreased complex I activity (Allagui et al., 2009; Maurer et al.,

2009). The findings of our study are the first to demonstrate that mitochondrial complex I inhibition

is able to increase protein oxidation and nitration and that lithium can ameliorate both alterations in

97

SH-SY5Y cells. Furthermore, our findings suggest that while the ability of lithium to reduce

carbonylation may be mediated by glutathione, its ability to prevent nitration may not be. This

demonstrates that multiple systems are involved in lithium’s antioxidant effects, and suggests that

elucidation of such pathways could contribute to the development of specific drug interventions to

target oxidative stress in BD.

In agreement with our results, many studies have demonstrated that lithium has

neuroprotective effects (Bachmann et al., 2009) such as decreasing apoptotic signaling (Q. Li et al.,

2010), improving energy metabolism (Bachmann et al., 2009; Maurer et al., 2009), and decreasing

oxidative damage (Cui et al., 2007; Shao et al., 2008). Although the mechanisms by which lithium

exerts its antioxidant effects are not clear, it has been hypothesized to work in part through the

glutathione system (Shao et al., 2008) and by improving mitochondrial functioning (Bachmann et al.,

2009). Since complex I inhibition can lead to increased leakage of electrons, resulting in the

production of the superoxide anion that can produce protein carbonylation and nitration (Kamat,

2006), lithium may be ameliorating these alterations by improving complex I functioning.

Glutathione, which is the most abundant antioxidant in the brain, acts by directly reacting

with ROS/RNS to neutralize them (Halliwell, 2001). In our study, glutathione depletion with BSO

increased carbonylation in the absence of rotenone as well, suggesting that glutathione alone is

important for decreasing the formation of carbonyl groups. BSO reduces the levels of glutathione,

thus our experiments suggest that lithium requires glutathione to prevent formation of carbonyl

groups. Indeed, glutathione is important to decrease levels of H2O2, which could alone contribute to

the formation of carbonyl groups, again suggesting the relevance of glutathione for prevention of

carbonyl groups (Dringen, 2000). Previous studies have shown that lithium can increase levels of

glutathione (Cui et al., 2007) and glutathione transferase, which conjugates glutathione to radicals

(Shao et al., 2008). While our results support the involvement of glutathione in the ability of lithium

98

to decrease carbonylation, the effect of lithium on nitration may occur through a different pathway,

such as improving complex I activity (N. Li et al., 2003). The effect of lithium on nitric oxide

production is controversial (Karimollah, Ghasemi, Ghahremani, & Dehpour, 2009; Wegener et al.,

2004). Further studies are required to examine the mechanism by which lithium prevents protein

nitration.

Limitations of this study include the use of a single technique, immunocytochemistry (ICC),

to quantify protein nitration and oxidation. Although measures were taken to minimize bias, such as

background subtraction and normalization by control, ICC is limited by factors such as

photobleaching. Furthermore, because we only used SH-SY5Y cells, which are well-characterized

neuronal cells (Cui et al., 2007; Watabe & Nakaki, 2004, 2007), using animal models may help to

expand the generalizability of these findings.

Acknowledgements

The authors have no conflict of interest to disclose. The authors thank Brain & Behavior Research

Foundation (Formerly NARSAD, ACA, BL), Coordenação de Aperfeiçoamento de Pessoal de Nível

Superior (CAPES-DFAIT) Program /Brazil and Canada (BL, LTG and CFN) and Canadian

Institutes of Health Research (CIHR) as sources of funding in support of this report (LTY and ACA).

99

Figures

Fig. 1 Lithium prevents decrease in complex I activity and cell viability induced by mitochondrial

complex I dysfunction. A. Complex I activity. Rotenone treatment (F5,36 = 3.67, P<0.01, *P<0.05

against DMSO control; green line); lithium pre-treatment (F1,36 = 4.68, P<0.05, red line). N =4. B.

Cell viability. Rotenone treatment (F5,44 = 3.83, P<0.01, *P<0.05 against DMSO control; green line);

lithium pre-treatment (F1,44=32.50, P<0.01, red line); BSO following lithium pre-treatment (F3,16 =

20.64, P<0.01; ++P<0.01 against lithium+BSO alone; blue line). N = 5. Statistics shown in brackets

are results of two-way ANOVA followed by Dunnett post-hoc analysis. Symbols represent Dunnett

post-hoc analysis against appropriate control

100

Fig. 2 Glutathione is important for lithium’s ability to decrease protein carbonylation, but not

nitration. A. Protein carbonyl levels. Rotenone treatment (F5,76 = 9.93, P<0.01, **P<0.01; green

line); lithium pre-treatment (F1,48 = 22.23, P<0.01, ++P<0.01, red line). BSO following lithium pre-

treatment (F1,32 = 56.39, P<0.01, ##P<0.01. blue line). N = 5. B. 3-nitrotyrosine levels. Rotenone

treatment (F5,48 = 19.23, P<0.01, **P<0.01, green line); lithium pre-treatment (F1,76 = 217.64,

P<0.01, red line). BSO following lithium pre-treatment (F1,43 = 64.07, P<0.01, blue line). N = 4-9.

Statistics shown in brackets are results of two-way ANOVA followed by Dunnett post-hoc analysis.

Symbols represent Dunnett post-hoc analysis against appropriate control. Insets are examples of

cells treated with different combinations of lithium, rotenone, and BSO as indicated on the image.

Mitochondria were labelled with MitoRed, shown in red (Mito), and carbonyl groups or 3NT were

101

labelled with Alexa488 (green). Images were taken with a 60X objective and a confocal laser

scanning microscope (Olympus). Scale bar = 20um

102

CHAPTER 5b

Title: Lithium reduces the effects of rotenone-induced complex I dysfunction on DNA methylation

and hydroxymethylation in rat cortical neurons.

Authors: Helena K. Kim*, Gustavo Scola*, L. Trevor Young, Mirian Salvador, Ana C. Andreazza

• HKK and GS share first authorship

Journal: Psychopharmacology

Volume: 21

Pages: 4189-4198

Work performed by the student:

The student performed all work included in this chapter, including planning and performing the

experiments, image acquisition, data analysis, and preparation and submission of the manuscript.

103

Title: Lithium reduces the effects of rotenone-induced complex I dysfunction on DNA

methylation and hydroxymethylation in rat cortical primary neurons.

Gustavo Scola1,3,4*, Helena K. Kim2,3*, L. Trevor Young1,2,3 Mirian Salvador4, Ana C.

Andreazza1,2,3#

1Departments of Psychiatry and 2Pharmacology, University of Toronto, ON, Canada

3Centre for Addiction and Mental Health, Toronto, ON, Canada

4Institute of Biotechnology, University of Caxias do Sul, RS, Brazil

*both authors contribute equally for this study

#Corresponding Author

Ana Andreazza, Pharm, PhD

University of Toronto

Medical Science Building, Room 4204.

1 king’s College Circle

Toronto, ON, M5S 1A8

Phone: 416-978-6042

e-mail: ana.andreazza@utoronto.ca.

104

Abstract

Rationale Mitochondrial complex I dysfunction and alterations in DNA methylation levels are

consistently reported in bipolar disorder (BD) and are regulated by lithium. One of the mechanisms

by which lithium may exert its effects in BD is by improving mitochondrial complex I function.

Therefore, we examined whether complex I dysfunction induces methylation and

hydroxymethylation of DNA and whether lithium alters these effects in rat primary cortical neurons.

Methods Rotenone was used to induce mitochondrial complex I dysfunction. Cell viability was

measured by MTT assay and ATP levels were assessed by Cell-Titer-Glo®. Complex I activity was

measured using an ELISA based assay. Apoptosis, DNA methylation and hydroxymethylation levels

were measured by immunocytochemistry.

Results Rotenone decreased complex I activity and ATP production, but increased cell death and

apoptosis. Rotenone treatment increased levels of 5-methylcytosine (5mc) and

hydroxymethylcytosine (5hmc), suggesting a possible association between complex I dysfunction

and DNA alterations. Lithium prevented rotenone-induced changes in mitochondrial complex I

function, cell death and changes to DNA methylation and hydroxymethylation.

Conclusions These findings suggest that decreased mitochondrial complex I activity may increase

DNA methylation and hydroxymethylation in rat primary cortical neurons, and that lithium may

prevent these effects.

Keywords: lithium, complex I, methylation, hydroxymethylation, bipolar disorder.

105

Introduction

Mitochondrial complex I dysfunction and increased oxidative stress have been consistently

demonstrated in bipolar disorder (BD) (Clay et al., 2010; Scola et al., 2013). Complex I dysfunction

has a number of consequences, including production of reactive oxygen species and subsequent

oxidative damage to proteins, lipids, and DNA (Andreazza et al., 2008; Halliwell, 1992). Indeed,

increased 8-hydroxy-2-deoxyguanosine (8-OhdG) has been reported in lymphocytes (Che et al.,

2010; Soeiro-de-Souza et al., 2013) and post-mortem hippocampus (Che et al., 2010; Soeiro-de-

Souza et al., 2013) from patients with BD, suggesting that oxidative modifications to DNA may be a

factor in this disorder. Of significance, recent studies have shown that oxidative stress can modulate

the levels of DNA methylation in cancer cells (Campos et al., 2007; Chestnut et al., 2011; K. A.

Kang et al., 2012; Lim et al., 2008; R. Zhang et al., 2013); however, the role of mitochondrial

complex I dysfunction in the regulation of DNA methylation levels, especially in neurons, is still

unclear.

There is growing interest in the role of epigenetic modifications in BD (Abdolmaleky et al.,

2006; D'Addario et al., 2012; Nohesara et al., 2011; Petronis, 2003). For example, increased levels

of 5-methylcytosine (5mc) have been reported at the BDNF promoter and the COMT gene in

patients with BD (Abdolmaleky et al., 2006; D'Addario et al., 2012). 5mc plays an important role in

cell dynamics, such as regulation of gene expression, X-chromosome inactivation, and maintenance

of epigenetic memory, where increase in methylation levels at promoter regions will decrease gene

expression (Bird, 2002). 5mc can be oxidized by the ten-eleven translocation (TET; EC 1.14.11.n2)

enzyme, producing 5-hydroxymethylcytosine (5hmc) (Jin et al., 2010; Wu et al., 2010). Although

the role of 5hmc in gene expression is unclear, it was shown to increase transcription of methylated

DNA and has been suggested to be an important factor in regulating the balance between

methylation and demethylation (Grayson & Guidotti, 2013; Jin et al., 2010; Wu et al., 2010). Of

106

significance, increased levels of 5hmc were found in the inferior parietal lobule of patients with

psychosis when compared to healthy controls (Dong, Gavin, Chen, & Davis, 2012).

Lithium, which is the most commonly prescribed drug for the treatment of BD (Kupfer,

2005), has a number of specific cellular targets including: activation of the Wnt signaling pathway,

inhibition of glycogen synthase kinase 3β, changes in cytoplasmic β-catenin (Boyadjieva &

Varadinova, 2012; Malhi et al., 2013), among others in addition to an effect on mitochondrial

function. More recently, lithium was reported to increase the antioxidant capacity of the cell by

increasing glutathione content (Cui et al., 2007; Shao et al., 2008) and improving mitochondrial

function, particularly that of complex I (Bachmann et al., 2009; Quiroz et al., 2008; Sun et al., 2006).

Another pathway by which lithium could be exerting its effects in BD is through the regulation of

DNA oxidation and methylation, as shown in in vitro studies and in patients with BD (Boyadjieva &

Varadinova, 2012; D'Addario et al., 2012). For example, decreased global levels of 5mc were

reported in patients with BD who respond to lithium (Huzayyin et al., 2013).

The aforementioned studies are in support of the critical roles that mitochondrial complex I

dysfunction and alterations in methylation levels play in BD. Lithium, the most prescribed

medication for patients with BD, has also been demonstrated to regulate antioxidant activity, cell

survival response and to induce changes in methylation patterns in patients with BD. However, it is

still unclear whether neuronal mitochondrial complex I dysfunction can directly induce changes in

DNA methylation and hydroxymethylation. To study this phenomenon, we treated rat cortical

primary neurons with rotenone, a direct mitochondrial complex I inhibitor (Halliwell 2001). We

also tested whether alterations in DNA methylation and hydroxymethylation levels require

mitochondrial dysfunction or can also occur in a high oxidative stress environment, such as that

produced by hydrogen peroxide (H2O2) treatment. Lastly, we examined the ability of lithium to

prevent mitochondrial complex I dysfunction, and any alterations to global 5mc and 5hmc levels

107

which may occur. Here, we report that rotenone decreases complex I activity leading to increased

levels of DNA methylation and hydroxymethylation. Treatment with H2O2 decreased 5mc levels, but

did not change the levels of 5hmc, suggesting that oxidative stress alone is insufficient to change

DNA hydroxymethylation levels. Lithium treatment was found to prevent mitochondrial complex I

dysfunction and consequent changes in 5mc and 5hmc levels. These results suggest that the effect of

lithium to regulate 5mc and 5hmc levels in rat primary cortical neurons may be through its ability to

enhance mitochondrial complex I function.

Methods

Cell culture

Rat E18 cortical neurons (PC35102 Neuromics, US) were purchased from Neuromics and

cultured according to instructions from Neuromics in Neurobasal® medium (Gibco, Burlington,

ON) supplemented with B27® (Gibco, Burlington, ON) and 0.5 mM glutamine. Cells were cultured

in plates/coverslips coated with poly-D-lysine (Sigma, St. Louis, MO) for 7 days and treated with

0.75mM lithium (BioShop, Burlington, ON) for 7 more days. Next, cells were treated with different

concentrations of rotenone (5nM, 10nM, and 50nM; Sigma, St. Louis, MO) or H2O2 (5µM, 10µM

and 50µM; Sigma, St. Louis, MO) for 30 minutes. After washout, cells were cultured additionally

for 24 hours in the same conditions.

Complex I activity assay

Cells (3 x 105 per well in 6 well plate) were treated with lithium, and then exposed to

different concentrations of rotenone (5, 10, 50nM) for 30 minutes. After 24 hours in media without

B27, cells were assayed for complex I activity using the Complex I Enzyme Activity Microplate

108

Assay Kit (Abcam, Cambridge, MA, USA) according to manufacturer’s instructions. Data are

presented as percent of non-lithium treated DMSO control.

ATP measurement assay

Cells (4 x 104 per well in 96 well plate) were treated with lithium, followed by exposure to

different concentrations of rotenone (5, 10, and 50nM) for 30 minutes. Cells were grown in media

without B27 for 24 hours and assayed for ATP levels using Cell-Titer-Glo® assay (Promega,

Madison, WI) according to manufacturer’s instructions. Data are presented as percent of DMSO

treated control cells.

Annexin V assay

Annexin V labels phosphatidylserine that has been exposed at the outer leaflet of the plasma

membrane during early apoptosis, and propidium iodide (PI) enters the cell when membrane

integrity is lost during necrosis or late stage apoptosis. Therefore, labeling the cells simultaneously

with Annexin V and PI allows for the discrimination of cells in the early apoptotic stages with the

integrity of the membrane still intact (annexin V+ only) and cells in the late apoptotic stage with

exposure of phosphatidylserine and compromised membrane integrity (positive for both annexin V

and PI) (Schutte, Nuydens, Geerts, & Ramaekers, 1998). ApoDETECT Annexin V-FITC kit

(Invitrogen) was used according to manufacturer’s instructions. Briefly, cells were seeded in 24

wells (2 x 105 cells per well) and pre-treated with media or lithium followed by treatment with

different concentrations of rotenone as described above. Cells were viewed using a confocal laser

scanning microscope. Number of annexin V only positive cells (AV+ cells), cells positive for both

Annexin V and PI (AVPI+ cells), and total number of cells were counted by two examiners based on

previously published methods (Farinacci, 2007; Schutte et al., 1998). Cells positive only for Annexin

109

V were considered early apoptotic, cells positive for AV and PI were considered late apoptotic (B.

Weber et al., 2004).

Cell mortality

The MTT (3-[4,5-dimethylthiazol-2-yl]-2,5 diphenyl tetrazolium bromide) assay (Sigma, St.

Louis, MO) was used to determine cell mortality. Cells (4 x 104 per well in 96 well plate) were

treated with lithium, and then exposed to different concentrations of rotenone (5, 10, and 50nM) for

30 minutes. After 24 hours, 1mg/mL of MTT solution was added per well and the cells were

cultured for another 2 h. Following incubation, supernatant fluid was discarded, and precipitates

were dissolved in 100 µl of DMSO per well. After 20 min of agitation, optical density of the

resultant solution was measured with a microplate reader (BioTek® Instruments, Inc.) at 517 nm.

The increase or decrease in mortality of cells in each well was expressed as the percentage of non-

lithium treated DMSO control.

Immunocytochemistry

For immunocytochemistry, cells (2 x 105) were seeded in 12mm coverslips, pre-treated with

lithium or media, and exposed to different concentrations of rotenone (5, 10, and 50nM) and H2O2

(5, 10, and 50µM) for 30 minutes. After the removal of treatment media and 24 hours in media

without B27, cells were prepared for immunocytochemistry procedures as previously described (Ficz

et al., 2011; Guo, Su, Zhong, Ming, & Song, 2011; Wossidlo et al., 2011) with slight modifications.

Briefly, cells were fixed in pre-cooled methanol for 10 min at -20°C, washed 3 times with

Dulbecco's Phosphate Buffered Saline (DPBS), and permeabilized with 0.2% Triton X-100. Next,

samples were treated with 2N HCl for 30 min at RT, neutralized with 0.1M Tris-HCl (pH 8) for 10

min, and washed in DPBS 3 times. Cells were blocked with 1% bovine serum albumin for 1 hour

110

and washed 2 times in DPBS. Cells were then incubated with primary antibodies for 24 hours at 4°C

diluted in DPBS-0.4% Triton-X: anti-5-hydroxymethylcytosine (Millipore, Temecula, CA;

MABE176; 1:100) and anti-5-methylcytosine (Millipore, Temecula, CA; MABE146; 1:200). Slides

were washed 3 times and incubated with secondary antibodies for 1 hour at RT in DPBS-0.4%

Triton-X: Alexa Fluor®488 goat-anti rat IgG (Molecular Probes, Burlington, ON; A11006; 1:200)

and Alexa Fluor®568 goat-anti mouse IgG (Molecular Probes, Burlington, ON; A11031; 1:200).

Next, slides were washed 3 times and incubated in 300nM 4',6-diamidino-2-phenylindole (DAPI)

(Molecular Probes, Burlington, ON; D21490) in DPBS for 10 min. Finally, the cells were mounted

in Fluoromount™ (Cedarlane, Burlington, ON; CLSG80116). Cross-reactivity for antibodies was

assayed by adding different combinations of the primary and secondary antibodies for assessing

resultant fluorescence intensity (Online Resource 1).

Image Analysis

All images were taken using a confocal laser scanning microscope (Olympus Fluoview1000).

A 40X objective was used for the Annexin V assay, and Annexin V and PI were visualized using

excitation with a 488nm laser at 28% laser intensity. Detection wavelength was 500-530nm for

Annexin V, and 560-660nm for PI. At least two hundred cells were counted in each experiment per

condition. A 60X water-immersion objective was used for visualizing 5mc and 5hmc. Alexa568

labeling 5mc was excited with the 543nm laser at 34% laser intensity, Alexa488 labeling 5hmc was

excited with the 488nm laser at 33% laser intensity, and DAPI was excited with the 405nm laser at

20% laser intensity. Detection wavelength was 560-660nm for Alexa568, 500nm-530nm for

Alexa488, and 425-475nm for DAPI. Scanning parameters were kept consistent for all images.

Image analysis was done using the FV10-ASW2.1 software connected to the microscope with

images in Original Imaging Format (OIF). For measuring fluorescence intensity, individual cells

111

were selected using DAPI labeling as individual regions of interest (ROI) in order to analyze 5mc

and 5hmc levels within the nucleus. At least 100 cells were included in each experiment per

condition.

Statistical Analysis

Statistical analysis was performed using SPSS version 20 (Illinois, USA). Parametric

distribution of data was assessed using the Kolmogorov-Smirnov test. As majority of the data was

found to be normally distributed, parametric tests were used for subsequent analyses. One-way

ANOVA was used to examine the effect of different concentrations of rotenone, and the effect of

lithium in combination with different concentrations of rotenone. LSD post-hoc analysis was used to

examine the difference between treatment groups against the appropriate control. Independent t-test

was used to identify differences between lithium pre-treatment and rotenone at each concentration

(5, 10, 50nM).

Results

Lithium reduces rotenone-induced complex I dysfunction, decreased ATP production and

cellular mortality in rat cortical primary neurons

Rotenone produced a dose-dependent decrease in complex I activity (F3,20 = 38.75; p < 0.01;

Figure 1). Rotenone treatment at all concentrations (5nM, 10nM, 50nM) decreased complex I

activity when compared to DMSO (F3,10 = 34.33; p < 0.01; LSD post-hoc p< 0.01 for all). Lithium

pre-treatment decreased the extent of rotenone-induced decline on complex I activity such that the

different concentrations of rotenone were no longer significantly different from control (Figure 1).

Furthermore, rotenone decreased the production of ATP levels in a dose dependent manner (Figure

112

2; F3,37 = 2083.62; p < 0.01), which was prevented by lithium (Figure 2; R5 t=-21.95; p<0.001; R10 t

=-13.49; p<0.001; R50 t =-4.75; p<0.001). Rotenone treatment significantly increased cell mortality

(F3,20 = 393.47; p < 0.01) at all concentrations when compared to DMSO (Figure 3; p < 0.01 for all).

Lithium pre-treatment was able to decrease cell mortality, where rotenone at 5nM did not

statistically differ from control (R5 t =0.88; p=0.401; R10 t =-13.70; p<0.001; R50 t =-25.18;

p<0.001).

Lithium decreases apoptosis induced by complex I dysfunction in rat cortical primary neurons

Cells in early stage of apoptosis are represented by Annexin V+ and those in late stage of

apoptosis are represented by Annexin V and PI positive staining (AVPI+; Figure 4). Rotenone

treatment increased Annexin V+ cells at all concentrations when compared to DMSO (Figure 4A;

F3,18 = 17.07; p < 0.01; LSD p < 0.05 for 5 and 10nM; p < 0.01 for 50nM). Lithium pre-treatment

was able to decrease the levels of Annexin V+ cells in comparison to rotenone treatment alone

(Figure 4A; R5 t =3.20; p=0.013; R10 t =4.21; p=0.002; R50 t =5.11; p=0.001). Mitochondrial

complex I dysfunction induced by rotenone increased the presence of AVPI+ cells at all

concentrations (Figure 4B; F3,18 = 4.52; p < 0.05; LSD p < 0.05 for 5 and 10nM; p < 0.01 for 50nM).

However, lithium pre-treatment did not change rotenone-induced increase in AVPI+ cells (Figure

4B).

Lithium prevents increased DNA methylation and hydroxymethylation levels induced by

mitochondrial complex I dysfunction in rat cortical primary neurons

Rotenone treatment increased DNA methylation levels at concentrations of 10nM and 50nM

in comparison to DMSO (Figure 5A; F3,16 = 4.19; p < 0.05; LSD p < 0.05 for both). A similar pattern

was observed for DNA hydroxymethylation (Figure 5B; F3,16 = 5.87; p < 0.01), where rotenone at

113

10nM (p < 0.01) and 50nM (p < 0.05) significantly increased hydroxymethylation levels compared

to DMSO. Lithium pre-treatment decreased rotenone-induced DNA methylation (Figure 5A; R5 t

=7.50; p<0.001; R10 t =5.65; p<0.001; R50 t =16.48; p<0.001) and hydroxymethylation (Figure 5A;

R5 t =4.49; p<0.001; R10 t =2.51; p<0.013; R50 t =9.46; p<0.001) at all concentrations.

Hydrogen peroxide decreases methylation of DNA in rat cortical primary neurons

In order to examine if the effect of rotenone is specific to mitochondrial impairment or due to

the general induction of oxidative stress, we treated the cells with H2O2. A different pattern was

obtained when cells were treated with H2O2, where decreased levels of DNA methylation (Figure

6A; F3,16 = 8.95; p < 0.01) at all concentrations (LSD 5µM p < 0.05, 10µM p < 0.01, and 50µM p <

0.01) were observed. However, a lack of difference in DNA hydroxymethylation levels showed that

the groups do not significantly differ from each other (Figure 6B; F3,16 = 1.27; p = 0.32).

Discussion

In rat primary cortical neurons, rotenone treatment decreased complex I activity while

increasing DNA methylation and hydroxymethylation. Notably, treatment of neurons with H2O2

decreased 5mc levels without changing the levels of 5hmc. Lithium pre-treatment was able to

prevent the effects of rotenone on complex I activity and simultaneously decreased rotenone-induced

changes in global methylation and hydroxymethylation levels. Decreased expression of complex I

subunits and its activity are consistent findings in BD, while lithium is one of the most commonly

prescribed and effective drugs to treat this disease. Our findings suggest that mitochondrial complex

I dysfunction may lead to changes in DNA methylation and hydroxymethylation levels.

Furthermore, these results suggest that lithium, through regulation of mitochondrial complex I

activity, may normalize the levels of DNA methylation and hydroxymethylation. Since DNA

114

methylation and hydroxymethylation may be critical processes involved in the pathophysiology of

BD and other major psychiatric disorders (Grayson & Guidotti, 2013), the present results may

suggest another role of mitochondrial dysfunction in BD and its implications for the development of

novel treatments.

Mitochondrial electron transport chain is the main source of ATP for cells (Brandt, 2006;

Scola et al., 2013). Importantly, decreased mRNA (Sun et al., 2006) and protein levels (Andreazza et

al., 2010; Andreazza et al., 2013) of mitochondrial complex I, particularly that of NDUFS7, have

been described in patients with BD (Brandt, 2006). In this study, we chose to use rotenone because it

inhibits complex I activity by obstructing the transfer of electrons to ubiquinone, resulting in

decreased efficiency of the electron transport chain, which could potentially decrease ATP

production (Eckert et al., 2003; N. Li et al., 2003). Here, we demonstrated that rotenone decreases

complex I activity and ATP levels, while lithium pre-treatment prevents these changes, suggesting

that lithium may improve complex I function and consequently other rotenone-induced alterations.

These findings are in support of previous studies reporting that lithium increases complex I activity

in post-mortem brain (Maurer et al., 2009), protects against mitochondrial damage produced by high

calcium concentrations (Shalbuyeva et al., 2007) and enhances mitochondrial function (Bachmann et

al., 2009).

Using the Annexin V assay, we demonstrated that rotenone produces an increase in the

number of cells in the early and late apoptotic stages. Annexin V binds to phosphatidylserine that

has been exposed to the outer leaflet of the membrane during apoptosis, which is a calcium-

dependent process (Schutte et al., 1998). Our results are in agreement with previous studies that have

consistently shown that rotenone increases apoptosis by a number of different mechanisms,

including disruption of ATP levels and elevation of intracellular free calcium levels that activate the

apoptotic pathway. In addition, lithium prevented rotenone-induced increase in early apoptosis.

115

Indeed, studies have shown that lithium is able to decrease pro-apoptotic factors such as p53 and bax

(Quiroz et al., 2008), normalize intracellular calcium levels, and regulate ATP production, which

could combat the activation of apoptotic machinery (Q. Li et al., 2010).

In this study, decreasing mitochondrial complex I activity by rotenone increased DNA

methylation in primary cortical neurons. DNA methylation occurs when DNA methyltransferase

catalyzes the methylation of cytosine, resulting in the formation of 5mC (Branco, Ficz, & Reik,

2012; K. A. Kang et al., 2012). Consistent with these results, a previous study reported increased

DNA methylation in healthy adult subjects with mitochondrial J haplotype and lower levels of ATP

(Bellizzi, D'Aquila, Giordano, Montesanto, & Passarino, 2012). In our study, we also demonstrated

an increase in hydroxymethylation, as measured by an increase in 5hmc levels. 5hmc formation has

been suggested as possible compensatory response to overcome high levels of methylation (Lister et

al., 2013; Wu et al., 2010). Studies have shown that high levels of both 5mc and 5hmc in promoters

lead to increased gene transcription (Ficz et al., 2011; Jin et al., 2010; Wu et al., 2010). Furthermore,

we examined if the effect of rotenone on DNA methylation/hydroxymethylation was specific to

mitochondrial inhibition or was a result of general increase in oxidative stress using H2O2 (Halliwell

& Gutteridge, 2007). H2O2 treatment decreased methylation while not altering 5hmc levels,

suggesting that the effect of rotenone on DNA methylation/hydroxymethylation may differ from that

of a general increase in oxidative stress induced by H2O2. This suggests that while decreases in

mitochondrial complex I activity may increase both DNA methylation and hydroxymethylation,

general oxidative stress may only decrease methylation levels. Furthermore, lithium was found to

prevent rotenone-induced increase in 5mc and 5hmc levels. These findings are in support of previous

studies suggesting that lithium promotes hypomethylation in embryonic stem cells and in neural

stem cells (Guo, Ma, et al., 2011), and patients with BD who were taking lithium were shown to

have reduced methylation of the BDNF promoter compared to BD subjects who were not prescribed

116

lithium (D'Addario et al., 2012). Moreover, induction of apoptosis in neurons was found to increase

methylation levels (Chestnut et al., 2011; Hernandez & Singleton, 2012) in vitro, and lithium was

consistently shown in previous studies to prevent apoptotic processes, possibly through the

inhibition of GSK3β (King, Bijur, & Jope, 2001; Quiroz et al., 2008).

The findings of the present study should be interpreted in light of its limitations. First, this

study only examined a limited number of factors that are affected by or involved in the modulation

of 5mc and 5hmc levels. Examining a broader range of factors may help further elucidate the role of

mitochondrial dysfunction in epigenetic modifications. Moreover, we used primary neurons, which

do not allow for the examination of the effect of cell division on 5mc and 5hmc levels (Ito et al.,

2010; Tahiliani et al., 2009). Examining the effect of rotenone and lithium using dividing cells such

as embryonic stem cells may aid in furthering our understanding of the interaction between

mitochondrial dysfunction and epigenetic aberrations. Third, we used immunocytochemistry to

measure the amount of 5mc and 5hmc in cells. Although we used previously published methods and

tested for cross-reactivity, immunofluorescence is limited in that it only allows for the assessment of

global levels. Using more specific techniques such as MeDIP-Seq and hMeDIP-Seq in future studies

to verify the findings of the present study may aid in further elucidating the effect of rotenone and

lithium on methylation and hydroxymethylation (Matarese, Carrillo-de Santa Pau, & Stunnenberg,

2011).

In conclusion, the findings of our study are the first to demonstrate the ability of

mitochondrial complex I dysfunction induced by rotenone to increase methylation and

hydroxymethylation in primary neurons, and to show that lithium is able to ameliorate these

alterations. Moreover, our findings showed that in contrast to rotenone, H2O2 decreased 5mc levels

without changing the levels of 5hmc. This suggests that changes in methylation and

hydroxymethylation levels following rotenone-induced suppression of complex I activity may not be

117

due to a general increase in oxidative stress. Although preliminary, these results suggest a potential

mechanism by which lithium reverts aberrations in methylation and hydroxymethylation levels in

patients with BD, and shed new light for further studies investigating the role of methylation/

hydroxymethylation pattern in BD.

Acknowledgement

The authors have no conflict of interest to disclose. The authors thank Brain & Behavior Research

Foundation (Formerly NARSAD, ACA), CNPq/Brazil (GS) and Canadian Institutes of Health

Research (CIHR) as sources of funding in support of this report (LTY and ACA).

118

Figures

Figure 1. Complex I activity measurement in rat cortical primary neurons with lithium pre-treatment

(0.75mM) and rotenone treatment at different concentrations (5nM, 10nM, and 50nM). Li(-)

indicates pre-treatment with media alone, and Li(+) indicates pre-treatment with lithium. Values are

expressed as % DMSO control, mean±SEM. **p < 0.01 against DMSO control; #p < 0.05 lithium

versus rotenone. N=4.

119

Figure 2. The effect of lithium pre-treatment (0.75mM) and rotenone treatment at different

concentrations (5nM, 10nM, 50nM) on ATP levels in primary cortical neurons. Li(-) indicates pre-

treatment with media alone, and Li(+) indicates pre-treatment with lithium. Values are expressed as

% DMSO control, mean±SEM. **p < 0.01 against DMSO control; ##p < 0.01 lithium versus

rotenone . N=4-8.

120

Figure 3. The effect of lithium pre-treatment (0.75mM) and rotenone treatment at different

concentrations (5nM, 10nM, 50nM) on cell mortality in primary cortical neurons measured by the

MTT assay. Li(-) indicates pre-treatment with media alone, and Li(+) indicates pre-treatment with

lithium. Values are expressed as % DMSO control, mean±SEM. **p < 0.01 against DMSO control;

##p < 0.01 lithium versus rotenone. N=5-8.

121

Figure 4. The effect of lithium pre-treatment (0.75mM) and rotenone treatment at different

concentrations (5nM, 10nM, 50nM) on apoptosis and necrosis in primary cortical neurons. Li(-)

indicates pre-treatment with media alone, and Li(+) indicates pre-treatment with lithium. Values are

expressed as ratio of Annexin V+ (AV+), and Propidium iodide (PI+) positive cells to the total

number of cells, mean±SEM. A. Annexin V+ cells B. PI+ Cells. *p< 0.05 **p< 0.01 against DMSO

control; #p < 0.05 ##p < 0.01 lithium versus rotenone. N=5-6

122

Figure 5. Changes in methylation and hydroxymethylation levels with 7 days of lithium pre-

treatment (0.75mM) and 30 minutes of rotenone treatment (5nM, 10nM, or 50nM) in primary

cortical neurons. Values are expressed as mean±SEM. A. Effect of lithium pre-treatment and

rotenone treatment on methylation levels. B. Effect of lithium pre-treatment and rotenone treatment

on hydroxymethylation levels. *p < 0.05 **p < 0.01 against DMSO control. N = 5. C. Images of

DAPI, 5-methylcytosine (5mc), and 5-hydroxymethylcytosine (5hmc) labeling with a confocal laser

scanning microscope and a 60X water immersion objective. DAPI is shown in blue, 5mc is shown in

red, and 5hmc is shown in green. Images are shown for DMSO control, rotenone treatment at 50nM,

and lithium pre-treatment (0.75mM) followed by rotenone treatment at 50nM. Images of 5mc and

5hmc for rotenone and lithium treatment in other doses are shown in Online Resource 2.

123

Figure 6. Changes in methylation and hydroxymethylation levels of 30 minutes of H2O2 treatment

(5µM, 10µM or 50µM) in primary cortical neurons. Values are expressed as mean±SEM. A. Effect

of H2O2 treatment on methylation levels. B. Effect of H2O2 treatment on hydroxymethylation levels.

*p < 0.05 **p < 0.01 against control. N = 5. C. Images of DAPI, 5-methylcytosine (5mc), and 5-

hydroxymethylcytosine (5hmc) labeling with a confocal laser scanning microscope and a 60X water

immersion objective. DAPI is shown in blue, 5mc is shown in red, and 5hmc is shown in green.

Images are shown for control, H2O2 treatment at 50µM. Images of 5mc and 5hmc for H2O2 treatment

in other doses are shown in Online Resource 3.

124

Schema 1. Mitochondrial complex I dysfunction induced by rotenone and it effect on cell viability

and DNA methylation and hydroxymethylation: the role of lithium.

125

Online Resource

126

127

Statement of significance and impact (Chapter 5a&b):

Downstream targets of complex I dysfunction was further examined in cell models, which

allow us to directly inhibit the electron transfer process in complex I using rotenone. Also, by

examining if lithium can decrease rotenone-induced alterations, we can observe if alterations

occurring from complex I defect are relevant to BD pathology. To examine downstream targets of

complex I dysfunction, we decided to measure protein oxidation and nitration, and methylation and

hydroxymethylation of DNA, addressing Aim 5. These factors were chosen because oxidative

damage to proteins and changes in methylation levels were reported in patients with BD (Andreazza

et al., 2010; D'Addario et al., 2012), and our lab has experience in measuring these factors using

different samples (Andreazza et al., 2010; Andreazza et al., 2013; Huzayyin et al., 2014). To

examine the effects of complex I dysfunction and lithium pre-treatment, we used one-way ANOVA.

We chose this statistical model because one-way ANOVA allows us to compare between the cells

that received the same pre-treatment (lithium vs. media only). For example, lithium was found to

increase cell viability on its own, which may lead to a significant main effect of lithium in a two-way

ANOVA model, even if lithium is unable to prevent the decrease in cell viability produced by an

assault (Allagui et al., 2009). By using one-way ANOVA, cells pre-treated with lithium followed by

rotenone will be compared against cells pre-treated with lithium only without rotenone treatment,

allowing us to specifically examine the protective effects of lithium. T-tests were used to compare

cells pre-treated with lithium against cells that received no pre-treatment.

First, we wanted to examine if complex I dysfunction is sufficient to produce oxidation and

nitration to proteins using SH-SY5Y cells. This study demonstrated that disrupting the electron

transfer process in complex I with rotenone increases oxidative and nitrosative modifications to

proteins. Previous studies have shown that lithium increases levels of complex I subunits and its

activity, suggesting that lithium may be able to prevent oxidative damage to proteins by improving

128

complex I function (Maurer et al., 2009; Sun et al., 2006). On the other hand, lithium was also

shown to increase glutathione levels as well as enzymes related to glutathione function, such as

glutathione peroxidase, suggesting that multiple factors are involved in lithium’s antioxidant

capabilities (Cui et al., 2007). The findings of this study demonstrated that lithium can decrease

carbonylation and nitration induced by rotenone treatment, suggesting that protein oxidation and

nitration from complex I dysfunction may be a part of BD pathology. Glutathione was found to be

critical in lithium’s ability to prevent cell death and carbonylation, suggesting that improvement of

complex I activity alone may not be sufficient to prevent against damages produced by complex I

inhibition.

The next cell study aimed to examine another potential downstream target of complex I

dysfunction, methylation and hydroxymethylation of DNA. To examine this, we pre-treated rat

cortical primary neurons with lithium followed by rotenone treatment, and measured levels of

methylation and hydroxymethylation using immunocytochemistry. The results demonstrated that

complex I inhibition with rotenone increases methylation as well as hydroxymethylation, suggesting

that methylation and hydroxymethylation may also be downstream targets of mitochondrial complex

I dysfunction. Surprisingly, it was also shown that complex I inhibition is most likely causing these

changes through a different pathway than oxidative stress, as hydrogen peroxide produced a different

pattern of alteration, where it decreased levels of methylation while levels of 5-

hydroxymethylcytosine were not affected. Complex I inhibition has other consequences in addition

to superoxide anion production, such as decreased ATP production and activation of apoptotic

mechanisms as demonstrated in this study, which could be contributing to alterations in 5-

methylcytosine and 5-hydroxymethylcytosine levels. Lithium was found to protect against the

effects of rotenone, suggesting that methylation and hydroxymethylation of DNA occurring as a

downstream effect of complex I dysfunction may have a role in BD pathology.

129

6. SUMMARY AND GENERAL DISCUSSION

Complex I dysfunction and oxidative stress are consistently reported in BD, where lower

protein and mRNA levels of complex I subunits, decreased complex I activity, and increased

oxidative modifications to proteins, lipids, and DNA have been found (Andreazza et al., 2010;

Andreazza et al., 2013; Sun et al., 2006). Complex I is the first member of the electron transport

chain that can produce oxidative stress through the leakage of electrons when dysfunctional

(Halliwell, 1992). Therefore, the overall aim of my PhD was to examine downstream targets of

complex I dysfunction in BD.

My first project aimed to examine if complex I dysfunction in the electron transport process

is specific to BD by reviewing microarray studies that measured expression levels of complex I

subunits in patients with BD or schizophrenia. Our findings revealed that the differentiating factor

between BD and schizophrenia is that complex I subunits that are specifically involved in the

electron transfer process are downregulated in BD, while those altered in schizophrenia are involved

in other functions such as proton pumping or structural support. Therefore, while other aspects of

mitochondrial dysfunction, such as lower production of ATP and activation of apoptotic pathways

may be common to BD and schizophrenia, mitochondrial generation of ROS through leakage of

electrons from complex I may be more specific to BD (Scola et al., 2013).

As complex I dysfunction and subsequent production of mitochondrial ROS is most likely

unique to this disorder and may be an important contributor to its pathophysiology, my first study in

post-mortem brain aimed to confirm complex I dysfunction in the frontal cortex. To do this, I

measured levels of all five complexes of the electron transport chain to test if decrease in complex I

is unique to this complex, or is a part a general downregulation of electron transport chain

components. The results showed that while complexes II, III, IV and V had no between-group

differences, complex I was lower in patients with BD compared to non-psychiatric controls,

130

suggesting that complex I dysfunction in BD is specific. To further examine the potential of this

complex I defect to lead to increased generation of ROS, I measured levels of NDUFS7, a complex I

subunit that is critical for the fidelity of the electron transfer process (Andreazza et al., 2010; Lebon

et al., 2007). NDUFS7 levels were lower in patients with BD compared to non-psychiatric controls

and patients with schizophrenia, adding support to the aforementioned review that complex I

dysfunction most likely contributes to oxidative stress in BD. Moreover, two previous studies from

our lab demonstrated lower NDUFS7 levels in patients with BD (Andreazza et al., 2010; Andreazza

et al., 2013), suggesting that this is a very consistent finding.

Following the confirmation of complex I dysfunction in the post-mortem brain of patients

with BD, I aimed to extend on the findings of a recent study from our group demonstrating the

mitochondria and the synapse as targets of oxidative stress in the frontal cortex of patients with BD

(Andreazza et al., 2013). For this, I aimed to examine the mitochondria and the synapse as potential

downstream targets of complex I dysfunction using post-mortem brains from patients with BD. First,

I examined the mitochondria as a downstream target by measuring the activation of the NLRP3-

inflamamsome, which is a marker of mitochondrial oxidative stress (Zhou et al., 2011). Furthermore,

activation of the NLRP3-inflammasome results in the production of inflammatory cytokines

(Schroder et al., 2010; Tschopp & Schroder, 2010; Zhong et al., 2013), making it an interesting

target to examine as increased inflammation is consistently found in patients with BD (Goldstein,

Kemp, et al., 2009; Leboyer et al., 2012; O'Brien et al., 2006; Rao et al., 2010). Using post-mortem

frontal cortex from patients with BD, schizophrenia and non-psychiatric controls, I found increased

levels of NLRP3 and ASC only in the crude mitochondrial fraction in patients with BD, suggesting

increased activation of the NLRP3-inflamamsome and identifying the mitochondria as a potential

target of complex I dysfunction in BD (Zhou et al., 2011). On the other hand, levels of caspase-1 and

cytokines IL-1β, IL-6, TNF-α, and IL-10 were increased in BD and schizophrenia, suggesting

131

increased inflammation in both disorders. This also suggests that while complex I dysfunction and

mitochondrial damage as a downstream target of complex I defect may be specific to BD,

inflammation may be common to both BD and schizophrenia.

The next study examined the synapse as a potential downstream target of complex I

dysfunction in BD, as the synapse was identified as another target of oxidative stress in the post-

mortem brain of patients with BD (Andreazza et al., 2013). Because increased dopamine signaling is

known to underlie mania (Berk et al., 2007) and the dopamine system mainly exists at the synapse

(Elsworth & Roth, 1997), we decided to examine oxidative and nitrosative modifications to areas

rich in dopamine as a marker of synaptic damage due to complex I defect. For this, we examined the

post-mortem prefrontal cortex of patients with BD, schizophrenia, major depression and non-

psychiatric individuals. Dopamine transporter (DAT) and tyrosine hydroxlase (TH) were chosen to

label dopamine rich areas because they have been used by previous studies to label dopaminergic

synapses along axons (Akil et al., 1999; Ciliax et al., 1999), and both have been demonstrated to be

vulnerable to oxidative modifications in vitro (Ara et al., 1998; Blanchard-Fillion et al., 2001; S. U.

Park et al., 2002). Findings of this study revealed that oxidation of DAT-immunoreactive regions are

increased patients with BD as measured by acceptor photobleaching FRET, which has a resolution

of 30-50nm with the immunohistochemistry technique that was used in this study (Kenworthy, 2001;

König et al., 2006), allowing us to examine oxidative modifications to the dopaminergic synapase.

Using the same technique, nitration in areas immunoreactive for TH was found to be decreased in

patients with BD and schizophrenia. This was contrary to my hypothesis, as complex I dysfunction

in BD was expected to contribute to greater levels of oxidation and nitration of dopaminergic

synapses. Patients with BD and schizophrenia can both exhibit psychosis, which is thought to be

caused in part by increased levels of dopamine at the synapse (Howes & Kapur, 2009; Pearlson et

al., 1995; Wong et al., 1997). Since dopamine and L-DOPA are favorable substrates for

132

peroxynitrite (S. Park et al., 2003), which is a strong nitrating agent (Greenacre & Ischiropoulos,

2001; Kamat, 2006), patients with BD and SCZ may have lower levels of nitration in dopamine-rich

areas as peroxynitrite may be consumed by its reaction with dopamine at the synaptic terminal.

Therefore, the findings of this study suggest that the synapse may be another potential downstream

target of complex I dysfunction in BD.

We also aimed to further examine downstream targets of complex I dysfunction by using cell

models, allowing us to directly inhibit the electron transfer process in complex I using rotenone

(Turrens & Boveris, 1980). Also, lithium, which is uniquely used for the treatment of BD and was

shown to improve complex I function (J. S. Lai et al., 2006; Maurer et al., 2009; Sun et al., 2006),

was used as a proxy to understand if downstream alterations caused by complex I inhibition may be

relevant to BD pathology. To examine the consequences of complex I dysfunction, we measured

protein oxidation and nitration, and methylation and hydroxymethylation of DNA as oxidative

modifications and methylation alterations were found in patients with BD, and our group has

examined these factors in patients with BD using different samples (Abdolmaleky et al., 2006;

Andreazza et al., 2010; Andreazza et al., 2013; D'Addario et al., 2012; Huzayyin et al., 2014). First,

using human neuroblastoma cells (SH-SY5Y), we examined if complex I inhibition produced by

rotenone is sufficient to cause protein oxidation and nitration, and if lithium can prevent these

changes by improving complex I activity or through the glutathione system, which is the most

important antioxidant system in the brain (Gawryluk et al., 2011). Rotenone, as expected, decreased

complex I activity, and increased protein carbonylation and nitration in a dose-dependent manner.

This suggests that complex I dysfunction in the electron transfer process is an important contributor

to oxidative stress, resulting in increased oxidation and nitration of proteins. Furthermore, lithium

pre-treatment was able to improve complex I activity, and decrease protein nitration and

carbonylation produced by complex I inhibition from rotenone. The results also suggested that

133

lithium may be doing this by improving complex I function to offset the inhibitory effects of

rotenone, as cells pre-treated with lithium showed higher complex I activity compared to cells pre-

treated with only media regardless of the presence of rotenone. This suggests that protein oxidation

and nitration from complex I dysfunction may be contributing to the pathology of BD (Maurer et al.,

2009). The ability of lithium to protect against rotenone induced cell death was abolished with the

addition of buthionine sulfoximine to deplete glutathione, again demonstrating that the main

mechanism of toxicity exerted by blocking electron transfer in complex I is through the generation

of oxidative stress. Also, lithium was no longer able to decrease protein carbonylation produced by

rotenone after blocking glutathione synthesis, suggesting that lithium’s ability to increase glutathione

levels are also important for its antioxidant effects. Surprisingly, buthionine sulfoximine did not

affect lithium’s ability to prevent nitration, suggesting that lithium’s ability to decrease nitration does

not significantly depend on glutathione. As superoxide anion is required for the formation of

peroxynitrite, a powerful nitrating reagent, lithium may be decreasing nitration by inhibiting

superoxide production from complex I. While this may be through decreasing nitric oxide

production, findings of previous studies examining lithium’s effect on nitric oxide synthase are

inconsistent (Karimollah et al., 2009; Nahman, Belmaker, & Azab, 2012; Wegener et al., 2004).

The next study further examined downstream effects of complex I dysfunction using cell

models by measuring methylation and hydroxymethylation of DNA following rotenone treatment.

Also, as our previous study demonstrated that lithium is able to protect against oxidative damage

produced by complex I inhibition, it was of interest to see if lithium could protect against another

downstream target of complex I dysfunction. Findings of our study showed that inhibiting complex I

with rotenone increased global levels of methylation and hydroxymethylation, suggesting that these

alterations may also be downstream targets of complex I dysfunction. In order to examine if these

changes were occurring as a consequence of oxidative stress produced by complex I inhibition, we

134

measured methylation and hydroxymethylation in cells treated with hydrogen peroxide. Surprisingly,

hydrogen peroxide produced a decrease in methylation and produced no changes in

hydroxymethylation, suggesting that changes that were observed with rotenone were due to other

effects of complex I inhibition other than generation of oxidative stress. Results from this study as

well as previous studies have shown that complex I dysfunction causes activation of apoptotic

pathways (Eckert et al., 2003), and inducing apoptosis in neurons was found to increase methylation

(Chestnut et al., 2011), suggesting that this could be the mechanism by which complex I dysfunction

is producing changes in methylation. Hence, lithium may be preventing rotenone-induced changes in

methylation and hydroxylation by improving complex I function, and therefore decreasing complex I

inhibition-induced apoptosis. In fact, lithium, in addition to improving complex I function, is also

known to decrease apoptosis (Chuang, 2005; Q. Li et al., 2010), possibly contributing to its ability to

prevent methylation changes. Mechanisms that underlie changes in hydroxymethylation patterns are

less well-known as it was discovered more recently (Song et al., 2011; Tahiliani et al., 2009).

Together, the findings of this study suggest that complex I dysfunction may cause alterations in

methylation and hydroxymethylation, and that this may be relevant to the pathophysiology of BD, as

lithium pre-treatment was able to prevent rotenone-induced alterations. Hence, by inducing complex

I dysfunction in the electron transfer process using rotenone and measuring its effects on methylation

and hydroxymethylation, this study extended the examination of downstream effects of complex I

dysfunction in BD.

Our findings suggest that complex I dysfunction may have an important role in the

pathophysiology of BD. For example, complex I dysfunction, by increasing the production of

oxidative stress, may contribute to oxidative modifications to enzymes and transporters, such as

those in the dopamine system. Dopamine dysregulation theory of BD states that increased dopamine

levels underlie mania, which is followed by a subsequent downregulation of dopaminergic signaling,

135

resulting in depression (Berk et al., 2007). Oxidative and nitrosative modifications to dopaminergic

proteins with cysteine or tyrosine residues, such as the dopamine transporter or tyrosine hydroxylase,

may contribute to this imbalance and to the cyclic nature of BD. For example, accumulating

oxidative modifications to dopaminergic proteins could result in increased production or decreased

uptake of dopamine from the synapse, resulting in increased dopamine signaling, which may

contribute to mania (H. K. Kim & Andreazza, 2012). According to Berk and colleagues (Berk et al.,

2007), excessive activation of the dopamine system can be followed by downregulation of

dopaminergic proteins, which may result in depression. Furthermore, according to kindling theory of

BD (Weiss et al., 2015), due to the increased vulnerability of the system from various neurological

defects including oxidative stress, the brain may be more vulnerable to experiencing a similar

hyperactivation upon facing a trigger, such as stress, resulting in another episode of mania and

contributing to the cyclical nature of BD. Dysregulation of the dopamine system in the prefrontal

and frontal cortices may also contribute to the cognitive decline in BD, as it is a part of the

mesocortical pathway critical for working memory, emotional processing, and attention, which have

been shown to be impaired in BD (Nikolaus, Antke, & Muller, 2009; Seamans & Yang, 2004).

Complex I dysfunction can also activate downstream signaling pathways through redox sensors,

such as the NLRP3-inflammasome (Zhou et al., 2011). Our study found increased activation of the

NLRP3-inflamamsome in patients with BD, suggesting that complex I dysfunction may be

contributing to increased inflammation in the frontal cortex of patients with BD as well.

Inflammation can lead to pyroptosis and apoptosis through the release of cytokines that activate the

NF-κB pathway, possibly contributing to neuronal and glial cell death and cognitive decline

(Shimada et al., 2012; Sollberger et al., 2014; Takuma, Baba, & Matsuda, 2004). Inflammation is

also hypothesized to play an important role in the progressive decline that is experienced by many

patients with BD (Berk et al., 2011). The strongest predictor of response to treatment is the number

136

of previous episodes, and cognitive performance in various tasks decreases with disease duration, all

suggesting that BD is a progressive disorder (Anderson et al., 2012; Berk et al., 2013; Kupfer, 2005).

One hypothesis states that since cells of monocytic origin, such as the microglia, react stronger and

faster to various assaults upon multiple activations due to immune memory, patients with BD may

become increasingly vulnerable to inflammatory activation with increasing number of episodes

(Barbosa et al., 2014), contributing to the progressive decline in this disease. Hence, these findings

suggest that complex I dysfunction may be contributing to the cyclical nature of BD by causing

dopamine dysregulation through oxidative modifications to dopaminergic proteins, and contributing

to the progressive nature of this disease by increasing the vulnerability of the brain through the

production of oxidative stress and inflammatory activation. Interestingly, while evidence suggesting

complex I dysfunction has been reported in other brain areas as well, such as the hippocampus

(Konradi et al., 2004), accumulating studies are suggesting that complex I defect and oxidative stress

are very consistent findings in the frontal and prefrontal cortices of patients with BD (Andreazza et

al., 2010; Andreazza et al., 2013; Frey et al., 2007; Gawryluk et al., 2011; Sun et al., 2006).

BD is a chronic and progressive disorder (Berk et al., 2011) where the development of more

efficacious treatments is of great interest (Aubry, Ferrero, & Schaad, 2007; Berk et al., 2011). Our

studies included in this thesis strongly suggest that complex I dysfunction may be an important

contributor to BD, suggesting that drugs that ameliorate complex I defect may be effective on their

own or as adjunct treatments. N-acetyl cysteine, which increases glutathione levels, was shown to

improve depression in patients with BD (Berk et al., 2008), suggesting that targeting oxidative stress,

which could occur as a result of complex I defect, may be useful for the treatment of BD as well.

Coenzyme Q10 or ubiquinone, which receives electrons from complex I – a process disturbed by

rotenone – was shown to decrease depressive symptoms in older adult patients with BD (Forester et

al., 2012), further suggesting that improving the electron transfer process in complex I may be

137

beneficial for BD. To my knowledge, while drugs commonly used to treat BD such as valproate or

lithium were shown to improve complex I activity and decrease oxidative stress (Cui et al., 2007;

Frey et al., 2006; Jornada et al., 2011; Sun et al., 2006), drugs that directly improve complex I

function have yet to be tested in patients with BD. Melatonin, which first received attention in BD to

correct irregular sleep patterns (Salvadore et al., 2010; Srinivasan et al., 2006), is well known for its

antioxidant properties and was shown to increase expression levels of complex I genes (Acuña-

Castroviejo, Escames, León, Carazo, & Khaldy, 2003; Lowes, Almawash, Webster, Reid, & Galley,

2011; Srinivasan et al., 2006; Srinivasan, Spence, Pandi-Perumal, Brown, & Cardinali, 2011). These

findings suggest that melatonin may be efficacious as an adjunct in BD by improving complex I

function and decreasing oxidative stress. In fact, patients with BD were found to have genetic

variations in the melatonin gene (Etain et al., 2012), further suggesting that melatonin may have

positive effects on the treatment of BD. Moreover, melatonin was found to decrease metabolic

effects of antipsychotic medications such as increased blood pressure or weight gain in patients with

BD, demonstrating that melatonin is safe to use as an adjunct for the treatment of BD and that it may

have multiple beneficial effects (Romo-Nava et al., 2014). Hence, a randomized clinical trial

examining the effect of melatonin as an adjunct on symptom severity, cognitive decline, complex I

dysfunction and oxidative stress may be beneficial for the development of better treatments for BD.

To summarize, I first re-examined microarray studies measuring mRNA levels of complex I

subunits in patients with BD or SCZ, and determined that patients with BD have lower levels of

subunits that are involved in the process of electron transfer, suggesting that they may be more

susceptible to increased leakage of electrons and greater generation of ROS. I also confirmed lower

levels of complex I and its subunit, NDUFS7, in the post-mortem brain of patients with BD. After

confirming that complex I dysfunction in BD is likely to contribute to increased oxidative stress, I

aimed to extend on a previous study identifying the mitochondria and the synapse as two targets of

138

oxidative stress in the frontal cortex of patients with BD. To examine the mitochondria as a target of

complex I dysfunction, I examined activation of the NLRP3-inflammasome, a marker of complex I

dysfunction and mitochondrial production of ROS, and found that the activation of the NLRP3-

inflammasome is increased in the post-mortem frontal cortex of patients with BD, demonstrating the

mitochondria as a potential downstream target of complex I dysfunction in BD. The next study

aimed to examine the second potential target, the synapse, by focusing on the dopaminergic synapse

as dopamine dysregulation is known to underlie mania. This study demonstrated the synapse as a

potential target of complex I dysfunction in BD, as oxidative modifications to the dopaminergic

synapse were found to be altered in the post-mortem prefrontal cortex from patients with BD.

Downstream targets of complex I dysfunction were further examined using cell models, allowing us

to directly inhibit the electron transfer process in complex I with rotenone. Using cell models, I

could also examine the effect of lithium, which is uniquely used for the treatment of BD and can be

used to examine if alterations produced by rotenone may be relevant to BD pathology. Complex I

dysfunction produced by rotenone increased protein nitration and oxidation, and increased

methylation and hydroxymethylation of DNA. These alterations were decreased with lithium pre-

treatment, suggesting that downstream effects of complex I dysfunction may have important

pathological roles in BD. While much remains to be explored in the role of complex I dysfunction in

the pathophysiology of BD, these studies suggest that examining complex I defect and its

downstream pathways may increase our understanding of the different systems that are involved in

BD, and potentially reveal novel targets that can be used to improve therapeutic interventions.

139

7. CONCLUSIONS

The overall aim of my PhD was to elucidate downstream targets of complex I dysfunction in

BD. Our review of microarray studies examining complex I subunits in psychiatric disorders

revealed that the differentiating factor between BD and schizophrenia is the downregulation of

complex I subunits specifically involved in the process of electron transfer in BD, suggesting that

patients with BD may be more vulnerable to the generation of oxidative stress. We then confirmed

complex I dysfunction in post-mortem brains of patients with BD, where patients with BD were

found to have lower levels of complex I, but not other members of the electron transport chain, and

also lower levels of NDUFS7, a complex I subunit that is critical for the electron transfer process.

After confirming complex I dysfunction in the frontal cortex of patients with BD, we aimed to

extend on a recent study published from our group, identifying the mitochondria and the synapse as

targets of oxidative stress in the post-mortem brain of patients with BD. I first examined the

mitochondria as a potential downstream target of complex I dysfunction in BD by measuring a

marker of mitochondrial oxidative stress, the NLRP3-inflamamsome, in the post-mortem frontal

cortex. Patients with BD were found to have increased activation of the NLRP3-inflammasome,

demonstrating the mitochondria as a potential downstream target of complex I dysfunction in BD. I

then examined the synapse as a downstream target of complex I defect by examining oxidative

modifications to the dopaminergic synapse, as dopamine dysregulation is known to underlie mania.

Results showed altered levels of oxidation and nitration in areas immunoreactive for markers of the

dopamine synapse, suggesting that complex I dysfunction may also target the synapse in BD. To

further examine downstream effects of complex I dysfunction, we used cell models, allowing us to

directly inhibit the electron transfer process in complex I using rotenone. We also examined the

effect of lithium, as it is uniquely used for the treatment of BD and can aid us in determining if

alterations occurring due to complex I inhibition with rotenone may be relevant to BD pathology.

140

Results revealed that complex I dysfunction produced by rotenone causes an increase in protein

oxidation and nitration and methylation and hydroxymethylation of DNA, suggesting that complex I

defect in the electron transfer process can directly cause potentially pathological alterations

downstream. Importantly, lithium pre-treatment was able to decrease these alterations caused by

rotenone, suggesting that consequences of complex I dysfunction may have pathological

implications in BD.

Our findings in post-mortem brain and cell models suggest that complex I dysfunction and its

downstream targets have important roles in the pathophysiology of BD. In vitro and animal studies

have shown that complex I dysfunction can lead to other downstream alterations as well (Beal, 2002;

Halliwell, 1992). Of interest, oxidative stress from complex I defect can increase intracellular

calcium levels (Persson et al., 2013; L. S. Zheng, Ishii, Zhao, Kondo, & Sasahara, 2013), which has

been consistently reported in patients with BD (Dubovsky et al., 1994; Perova et al., 2010; Perova,

Wasserman, Li, & Warsh, 2008), suggesting that complex I dysfunction may have multiple effects in

BD. Future directions include confirmation of these findings using different techniques. For

example, while DAT and TH immunoreactive regions were found to be targets of oxidative

modifications using FRET, most likely representing alterations of DAT and TH themselves,

confirming this using a biochemical technique, such as co-immunoprecipitation would help solidify

the dopaminergic synapse as a downstream target of complex I dysfunction. Furthermore, as

identification of biomarkers in BD is of great interest, examining whether markers of complex I

dysfunction, such as the NLRP3-inflammasome, can be measured in peripheral cells from patients

with BD may yield interesting results. This will also allow us to confirm our findings using a larger

sample, increasing the generalizability of the results. Also, using animal models to confirm the

findings of our studies using neuroblastoma cells and primary neurons would allow for the validation

of the identified mechanisms in a more complex system. Lastly, as lithium was demonstrated to be

141

effective in preventing downstream effects of complex I inhibition through its ability to improve

complex I functioning, examining the effect of other reagents that were shown to improve complex I

activity as an adjunct may contribute to developing better treatments for BD.

142

References

Abdolmaleky, H. M., Cheng, K. H., Faraone, S. V., Wilcox, M., Glatt, S. J., Gao, F., . . . Thiagalingam, S. (2006). Hypomethylation of MB-COMT promoter is a major risk factor for schizophrenia and bipolar disorder. Hum Mol Genet, 15(21), 3132-3145. doi: 10.1093/hmg/ddl253

Acuña-Castroviejo, D., Escames, G., León, J., Carazo, A., & Khaldy, H. (2003). Mitochondrial regulation by melatonin and its metabolites. Adv Exp Med Biol, 527, 549-557.

Agostini, L., Martinon, F., Burns, K., McDermott, M. F., Hawkins, P. N., & Tschopp, J. (2004). NALP3 forms an IL-1beta-processing inflammasome with increased activity in Muckle-Wells autoinflammatory disorder. Immunity, 20(3), 319-325.

Ahluwalia, P., Grewaal, D. S., & Singhal, R. L. (1981). Brain gabaergic and dopaminergic systems following lithium treatment and withdrawal. Prog Neuropsychopharmacol, 5(5-6), 527-530.

Akil, M., Pierri, J. N., Whitehead, R. E., Edgar, C. L., Mohila, C., Sampson, A. R., & Lewis, D. A. (1999). Lamina-specific alterations in the dopamine innervation of the prefrontal cortex in schizophrenic subjects. Am J Psychiatry, 156(10), 1580-1589.

Allagui, M. S., Nciri, R., Rouhaud, M. F., Murat, J. C., El Feki, A., Croute, F., & Vincent, C. (2009). Long-term exposure to low lithium concentrations stimulates proliferation, modifies stress protein expression pattern and enhances resistance to oxidative stress in SH-SY5Y cells. Neurochem Res, 34(3), 453-462. doi: 10.1007/s11064-008-9804-8

Almeida, J. R., Akkal, D., Hassel, S., Travis, M. J., Banihashemi, L., Kerr, N., . . . Phillips, M. L. (2009). Reduced gray matter volume in ventral prefrontal cortex but not amygdala in bipolar disorder: significant effects of gender and trait anxiety. Psychiatry Res, 171(1), 54-68.

Anand, A., Verhoeff, P., Seneca, N., Zoghbi, S. S., Seibyl, J. P., Charney, D. S., & Innis, R. B. (2000). Brain SPECT imaging of amphetamine-induced dopamine release in euthymic bipolar disorder patients. Am J Psychiatry, 157(7), 1108-1114.

Anderson, I. M., Haddad, P. M., & Scott, J. (2012). Bipolar disorder. BMJ, 345, e8508. Andreazza, A. C., Cassini, C., Rosa, A. R., Leite, M. C., de Almeida, L. M., Nardin, P., . . .

Goncalves, C. A. (2007). Serum S100B and antioxidant enzymes in bipolar patients. J Psychiatr Res, 41(6), 523-529.

Andreazza, A. C., Kapczinski, F., Kauer-Sant'Anna, M., Walz, J. C., Bond, D. J., Goncalves, C. A., . . . Yatham, L. N. (2009). 3-Nitrotyrosine and glutathione antioxidant system in patients in the early and late stages of bipolar disorder. J Psychiatry Neurosci, 34(4), 263-271.

Andreazza, A. C., Kauer-Sant'anna, M., Frey, B. N., Bond, D. J., Kapczinski, F., Young, L. T., & Yatham, L. N. (2008). Oxidative stress markers in bipolar disorder: a meta-analysis. J Affect Disord, 111(2-3), 135-144.

Andreazza, A. C., Shao, L., Wang, J. F., & Young, L. T. (2010). Mitochondrial complex I activity and oxidative damage to mitochondrial proteins in the prefrontal cortex of patients with bipolar disorder. Arch Gen Psychiatry, 67(4), 360-368.

Andreazza, A. C., Wang, J. F., Salmasi, F., Shao, L., & Young, L. T. (2013). Specific subcellular changes in oxidative stress in prefrontal cortex from patients with bipolar disorder. J Neurochem. doi: 10.1111/jnc.12316

Ara, J., Przedborski, S., Naini, A. B., Jackson-Lewis, V., Trifiletti, R. R., Horwitz, J., & Ischiropoulos, H. (1998). Inactivation of tyrosine hydroxylase by nitration following exposure to peroxynitrite and 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP). Proc Natl Acad Sci U S A, 95(13), 7659-7663.

143

Asanuma, M., Miyazaki, I., Diaz-Corrales, F. J., & Ogawa, N. (2004). Quinone formation as dopaminergic neuron-specific oxidative stress in the pathogenesis of sporadic Parkinson's disease and neurotoxin-induced parkinsonism. Acta Med Okayama, 58(5), 221-233.

Aubry, J.-M., Ferrero, F.-P., & Schaad, N. (2007). Pharmacotherapy of bipolar disorders (pp. xiii, 277 p.). Retrieved from http://myaccess.library.utoronto.ca/login?url=http://books.scholarsportal.info/viewdoc.html?id=/ebooks/ebooks2/wiley/2011-12-13/1/9780470510421 Retrieved from http://myaccess.library.utoronto.ca/login?url=http://onlinelibrary.wiley.com/book/10.1002/9780470510421

Bachmann, R. F., Schloesser, R. J., Gould, T. D., & Manji, H. K. (2005). Mood stabilizers target cellular plasticity and resilience cascades: implications for the development of novel therapeutics. Mol Neurobiol, 32(2), 173-202.

Bachmann, R. F., Wang, Y., Yuan, P., Zhou, R., Li, X., Alesci, S., . . . Manji, H. K. (2009). Common effects of lithium and valproate on mitochondrial functions: protection against methamphetamine-induced mitochondrial damage. Int J Neuropsychopharmacol, 12(6), 805-822. doi: S1461145708009802 [pii]

10.1017/S1461145708009802

Barbosa, I. G., Huguet, R. B., Mendonça, V. A., Sousa, L. P., Neves, F. S., Bauer, M. E., & Teixeira, A. L. (2011). Increased plasma levels of soluble TNF receptor I in patients with bipolar disorder. Eur Arch Psychiatry Clin Neurosci, 261(2), 139-143. doi: 10.1007/s00406-010-0116-z

Barbosa, I. G., Morato, I. B., de Miranda, A. S., Bauer, M. E., Soares, J. C., & Teixeira, A. L. (2014). A preliminary report of increased plasma levels of IL-33 in bipolar disorder: further evidence of pro-inflammatory status. J Affect Disord, 157, 41-44. doi: 10.1016/j.jad.2013.12.042

Barbosa, I. G., Rocha, N. P., Bauer, M. E., de Miranda, A. S., Huguet, R. B., Reis, H. J., . . . Teixeira, A. L. (2013). Chemokines in bipolar disorder: trait or state? Eur Arch Psychiatry Clin Neurosci, 263(2), 159-165. doi: 10.1007/s00406-012-0327-6

Barnett, J. H., & Smoller, J. W. (2009). The genetics of bipolar disorder. Neuroscience, 164(1), 331-343. doi: S0306-4522(09)00576-4 [pii]

10.1016/j.neuroscience.2009.03.080

Barzilai, A., & Yamamoto, K. (2004). DNA damage responses to oxidative stress. DNA Repair (Amst), 3(8-9), 1109-1115.

Bashkatova, V., Alam, M., Vanin, A., & Schmidt, W. J. (2004). Chronic administration of rotenone increases levels of nitric oxide and lipid peroxidation products in rat brain. Exp Neurol, 186(2), 235-241. doi: 10.1016/j.expneurol.2003.12.005

Bauer, M. (2004). Review: lithium reduces relapse rates in people with bipolar disorder. Evid Based Ment Health, 7(3), 72.

Beal, M. F. (2002). Oxidatively modified proteins in aging and disease. Free Radic Biol Med, 32(9), 797-803.

Bearden, C. E., Thompson, P. M., Dalwani, M., Hayashi, K. M., Lee, A. D., Nicoletti, M., . . . Soares, J. C. (2007). Greater cortical gray matter density in lithium-treated patients with bipolar disorder. Biol Psychiatry, 62(1), 7-16. doi: 10.1016/j.biopsych.2006.10.027

Bellizzi, D., D'Aquila, P., Giordano, M., Montesanto, A., & Passarino, G. (2012). Global DNA methylation levels are modulated by mitochondrial DNA variants. Epigenomics, 4(1), 17-27. doi: 10.2217/epi.11.109

144

Benes, F. M., Matzilevich, D., Burke, R. E., & Walsh, J. (2006). The expression of proapoptosis genes is increased in bipolar disorder, but not in schizophrenia. Mol Psychiatry, 11(3), 241-251.

Berk, M., Berk, L., Dodd, S., Cotton, S., Macneil, C., Daglas, R., . . . Malhi, G. S. (2013). Stage managing bipolar disorder. Bipolar Disord. doi: 10.1111/bdi.12099

Berk, M., Copolov, D. L., Dean, O., Lu, K., Jeavons, S., Schapkaitz, I., . . . Bush, A. I. (2008). N-acetyl cysteine for depressive symptoms in bipolar disorder--a double-blind randomized placebo-controlled trial. Biol Psychiatry, 64(6), 468-475.

Berk, M., Dodd, S., Kauer-Sant'anna, M., Malhi, G. S., Bourin, M., Kapczinski, F., & Norman, T. (2007). Dopamine dysregulation syndrome: implications for a dopamine hypothesis of bipolar disorder. Acta Psychiatr Scand Suppl(434), 41-49.

Berk, M., Kapczinski, F., Andreazza, A. C., Dean, O. M., Giorlando, F., Maes, M., . . . Malhi, G. S. (2011). Pathways underlying neuroprogression in bipolar disorder: focus on inflammation, oxidative stress and neurotrophic factors. Neurosci Biobehav Rev, 35(3), 804-817. doi: 10.1016/j.neubiorev.2010.10.001

Berman, K. F., Doran, A. R., Pickar, D., & Weinberger, D. R. (1993). Is the mechanism of prefrontal hypofunction in depression the same as in schizophrenia? Regional cerebral blood flow during cognitive activation. Br J Psychiatry, 162, 183-192.

Berman, S. B., & Hastings, T. G. (1999). Dopamine oxidation alters mitochondrial respiration and induces permeability transition in brain mitochondria: implications for Parkinson's disease. J Neurochem, 73(3), 1127-1137.

Berman, S. B., Zigmond, M. J., & Hastings, T. G. (1996). Modification of dopamine transporter function: effect of reactive oxygen species and dopamine. J Neurochem, 67(2), 593-600.

Bestor, T. (2000). The DNA methyltransferases of mammals. Human Molecular Genetics, 9(16), 2395-2402. doi: 10.1093/hmg/9.16.2395

Bhagwagar, Z., & Goodwin, G. M. (2002). Role of mood stabilizers in bipolar disorder. Expert Rev Neurother, 2(2), 239-248. doi: 10.1586/14737175.2.2.239

Bird, A. (2002). DNA methylation patterns and epigenetic memory. Genes Dev, 16(1), 6-21. doi: 10.1101/gad.947102

Blanchard-Fillion, B., Souza, J. M., Friel, T., Jiang, G. C., Vrana, K., Sharov, V., . . . Ischiropoulos, H. (2001). Nitration and inactivation of tyrosine hydroxylase by peroxynitrite. J Biol Chem, 276(49), 46017-46023. doi: M105564200 [pii]

10.1074/jbc.M105564200

Boufidou, F., Nikolaou, C., Alevizos, B., Liappas, I. A., & Christodoulou, G. N. (2004). Cytokine production in bipolar affective disorder patients under lithium treatment. J Affect Disord, 82(2), 309-313. doi: 10.1016/j.jad.2004.01.007

Boveris, A., & Chance, B. (1973). The mitochondrial generation of hydrogen peroxide. General properties and effect of hyperbaric oxygen. Biochem J, 134(3), 707-716.

Boyadjieva, N., & Varadinova, M. (2012). Epigenetics of psychoactive drugs. J Pharm Pharmacol, 64(10), 1349-1358. doi: 10.1111/j.2042-7158.2012.01475.x

Branco, M. R., Ficz, G., & Reik, W. (2012). Uncovering the role of 5-hydroxymethylcytosine in the epigenome. Nat Rev Genet, 13(1), 7-13. doi: 10.1038/nrg3080

Brandt, U. (2006). Energy converting NADH:quinone oxidoreductase (complex I). Annu Rev Biochem, 75, 69-92. doi: 10.1146/annurev.biochem.75.103004.142539

Bresee, C., & Rapaport, M. H. (2009). Persistently increased serum soluble interleukin-2 receptors in continuously ill patients with schizophrenia. Int J Neuropsychopharmacol, 12(6), 861-865. doi: 10.1017/S1461145709000315

145

Brietzke, E., Stabellini, R., Grassis-Oliveira, R., & Lafer, B. (2011). Cytokines in Bipolar Disorder: Recent Findings, Deleterious Effects But Promise for Future Therapeutics. CNS Spectr.

Brown, N. C., Andreazza, A. C., & Young, L. T. (2014). An updated meta-analysis of oxidative stress markers in bipolar disorder. Psychiatry Res, 218(1-2), 61-68. doi: 10.1016/j.psychres.2014.04.005

Bryan, N. B., Dorfleutner, A., Kramer, S. J., Yun, C., Rojanasakul, Y., & Stehlik, C. (2010). Differential splicing of the apoptosis-associated speck like protein containing a caspase recruitment domain (ASC) regulates inflammasomes. J Inflamm (Lond), 7, 23. doi: 10.1186/1476-9255-7-23

Bryant, C., & Fitzgerald, K. A. (2009). Molecular mechanisms involved in inflammasome activation. Trends Cell Biol, 19(9), 455-464. doi: 10.1016/j.tcb.2009.06.002

Burgess, S., Geddes, J., Hawton, K., Townsend, E., Jamison, K., & Goodwin, G. (2001). Lithium for maintenance treatment of mood disorders. Cochrane Database Syst Rev(3), CD003013. doi: 10.1002/14651858.CD003013

Campos, A. C., Molognoni, F., Melo, F. H., Galdieri, L. C., Carneiro, C. R., D'Almeida, V., . . . Jasiulionis, M. G. (2007). Oxidative stress modulates DNA methylation during melanocyte anchorage blockade associated with malignant transformation. Neoplasia, 9(12), 1111-1121.

Chandana, R., Mythri, R. B., Mahadevan, A., Shankar, S. K., & Srinivas Bharath, M. M. (2009). Biochemical analysis of protein stability in human brain collected at different post-mortem intervals. Indian J Med Res, 129(2), 189-199.

Chanput, W., Mes, J. J., & Wichers, H. J. (2014). THP-1 cell line: An in vitro cell model for immune modulation approach. Int Immunopharmacol. doi: 10.1016/j.intimp.2014.08.002

Che, Y., Wang, J. F., Shao, L., & Young, T. (2010). Oxidative damage to RNA but not DNA in the hippocampus of patients with major mental illness. J Psychiatry Neurosci, 35(5), 296-302.

Chen, C. H., Suckling, J., Lennox, B. R., Ooi, C., & Bullmore, E. T. (2011). A quantitative meta-analysis of fMRI studies in bipolar disorder. Bipolar Disord, 13(1), 1-15. doi: 10.1111/j.1399-5618.2011.00893.x

Cheng, K. C., Cahill, D. S., Kasai, H., Nishimura, S., & Loeb, L. A. (1992). 8-Hydroxyguanine, an abundant form of oxidative DNA damage, causes G----T and A----C substitutions. J Biol Chem, 267(1), 166-172.

Chestnut, B. A., Chang, Q., Price, A., Lesuisse, C., Wong, M., & Martin, L. J. (2011). Epigenetic regulation of motor neuron cell death through DNA methylation. J Neurosci, 31(46), 16619-16636. doi: 10.1523/JNEUROSCI.1639-11.2011

Choi, K. H., Higgs, B. W., Wendland, J. R., Song, J., McMahon, F. J., & Webster, M. J. (2011). Gene expression and genetic variation data implicate PCLO in bipolar disorder. Biol Psychiatry, 69(4), 353-359. doi: 10.1016/j.biopsych.2010.09.042

Chuang, D. M. (2005). The antiapoptotic actions of mood stabilizers: molecular mechanisms and therapeutic potentials. Ann N Y Acad Sci, 1053, 195-204.

Ciliax, B. J., Drash, G. W., Staley, J. K., Haber, S., Mobley, C. J., Miller, G. W., . . . Levey, A. I. (1999). Immunocytochemical localization of the dopamine transporter in human brain. J Comp Neurol, 409(1), 38-56.

Cipriani, A., Barbui, C., Salanti, G., Rendell, J., Brown, R., Stockton, S., . . . Geddes, J. R. (2011). Comparative efficacy and acceptability of antimanic drugs in acute mania: a multiple-treatments meta-analysis. Lancet, 378(9799), 1306-1315. doi: 10.1016/S0140-6736(11)60873-8

Clay, H. B., Sillivan, S., & Konradi, C. (2010). Mitochondrial dysfunction and pathology in bipolar disorder and schizophrenia. Int J Dev Neurosci, 29(3), 311-324.

146

Connor, C. M., & Akbarian, S. (2008). DNA methylation changes in schizophrenia and bipolar disorder. Epigenetics, 3(2), 55-58.

Cordeiro, M. L., Gundersen, C. B., & Umbach, J. A. (2002). Lithium ions modulate the expression of VMAT2 in rat brain. Brain Res, 953(1-2), 189-194.

Cousins, D. A., Butts, K., & Young, A. H. (2009). The role of dopamine in bipolar disorder. Bipolar Disord, 11(8), 787-806. doi: 10.1111/j.1399-5618.2009.00760.x

Craddock, N., & Sklar, P. (2013). Genetics of bipolar disorder. Lancet, 381(9878), 1654-1662. doi: 10.1016/S0140-6736(13)60855-7

Cui, J., Shao, L., Young, L. T., & Wang, J. F. (2007). Role of glutathione in neuroprotective effects of mood stabilizing drugs lithium and valproate. Neuroscience, 144(4), 1447-1453.

Cunha, A. B., Frey, B. N., Andreazza, A. C., Goi, J. D., Rosa, A. R., Gonçalves, C. A., . . . Kapczinski, F. (2006). Serum brain-derived neurotrophic factor is decreased in bipolar disorder during depressive and manic episodes. Neurosci Lett, 398(3), 215-219. doi: 10.1016/j.neulet.2005.12.085

D'Addario, C., Dell'Osso, B., Palazzo, M. C., Benatti, B., Lietti, L., Cattaneo, E., . . . Altamura, A. C. (2012). Selective DNA methylation of BDNF promoter in bipolar disorder: differences among patients with BDI and BDII. Neuropsychopharmacology, 37(7), 1647-1655. doi: 10.1038/npp.2012.10

D'Ardenne, K., Eshel, N., Luka, J., Lenartowicz, A., Nystrom, L. E., & Cohen, J. D. (2012). Role of prefrontal cortex and the midbrain dopamine system in working memory updating. Proc Natl Acad Sci U S A, 109(49), 19900-19909. doi: 10.1073/pnas.1116727109

Daubner, S. C., Le, T., & Wang, S. (2011). Tyrosine hydroxylase and regulation of dopamine synthesis. Arch Biochem Biophys, 508(1), 1-12. doi: S0003-9861(10)00514-X [pii]

10.1016/j.abb.2010.12.017

Dean, B., Gibbons, A. S., Tawadros, N., Brooks, L., Everall, I. P., & Scarr, E. (2013). Different changes in cortical tumor necrosis factor-α-related pathways in schizophrenia and mood disorders. Mol Psychiatry, 18(7), 767-773. doi: 10.1038/mp.2012.95

Dean, O. M., Bush, A. I., & Berk, M. (2012). Translating the rosetta stone of N-acetylcysteine. Biol Psychiatry, 71(11), 935-936. doi: 10.1016/j.biopsych.2012.04.001

Deicken, R. F., Weiner, M. W., & Fein, G. (1995). Decreased temporal lobe phosphomonoesters in bipolar disorder. J Affect Disord, 33(3), 195-199.

Dheen, S. T., Kaur, C., & Ling, E. A. (2007). Microglial activation and its implications in the brain diseases. Curr Med Chem, 14(11), 1189-1197.

Dixon, J. F., & Hokin, L. E. (1998). Lithium acutely inhibits and chronically up-regulates and stabilizes glutamate uptake by presynaptic nerve endings in mouse cerebral cortex. Proc Natl Acad Sci U S A, 95(14), 8363-8368.

Dong, E., Gavin, D. P., Chen, Y., & Davis, J. (2012). Upregulation of TET1 and downregulation of APOBEC3A and APOBEC3C in the parietal cortex of psychotic patients. Transl Psychiatry, 2, e159. doi: 10.1038/tp.2012.86

Dringen, R. (2000). Glutathione metabolism and oxidative stress in neurodegeneration. Eur J Biochem, 267(16), 4903.

Dubovsky, S. L., Thomas, M., Hijazi, A., & Murphy, J. (1994). Intracellular calcium signalling in peripheral cells of patients with bipolar affective disorder. Eur Arch Psychiatry Clin Neurosci, 243(5), 229-234.

Dunigan, C. D., & Shamoo, A. E. (1995). Li+ stimulates ATP-regulated dopamine uptake in PC12 cells. Neuroscience, 65(1), 1-4.

147

Eckert, A., Keil, U., Marques, C. A., Bonert, A., Frey, C., Schüssel, K., & Müller, W. E. (2003). Mitochondrial dysfunction, apoptotic cell death, and Alzheimer's disease. Biochem Pharmacol, 66(8), 1627-1634.

Eder, C. (2009). Mechanisms of interleukin-1beta release. Immunobiology, 214(7), 543-553. doi: 10.1016/j.imbio.2008.11.007

El Khoury, A., Petterson, U., Kallner, G., Aberg-Wistedt, A., & Stain-Malmgren, R. (2002). Calcium homeostasis in long-term lithium-treated women with bipolar affective disorder. Prog Neuropsychopharmacol Biol Psychiatry, 26(6), 1063-1069.

Elsworth, J. D., & Roth, R. H. (1997). Dopamine synthesis, uptake, metabolism, and receptors: relevance to gene therapy of Parkinson's disease. Exp Neurol, 144(1), 4-9. doi: 10.1006/exnr.1996.6379

Elvidge, G., Jones, I., McCandless, F., Asherson, P., Owen, M. J., & Craddock, N. (2001). Allelic variation of a BalI polymorphism in the DRD3 gene does not influence susceptibility to bipolar disorder: results of analysis and meta-analysis. Am J Med Genet, 105(4), 307-311.

Etain, B., Dumaine, A., Bellivier, F., Pagan, C., Francelle, L., Goubran-Botros, H., . . . Jamain, S. (2012). Genetic and functional abnormalities of the melatonin biosynthesis pathway in patients with bipolar disorder. Hum Mol Genet, 21(18), 4030-4037. doi: 10.1093/hmg/dds227

Farinacci, M. (2007). Improved apoptosis detection in ovine neutrophils by annexin V and carboxyfluorescein diacetate staining. Cytotechnology, 54(3), 149-155. doi: 10.1007/s10616-007-9086-z

Fariss, M. W., Chan, C. B., Patel, M., Van Houten, B., & Orrenius, S. (2005). Role of mitochondria in toxic oxidative stress. Mol Interv, 5(2), 94-111. doi: 10.1124/mi.5.2.7

Felger, J. C., Li, L., Marvar, P. J., Woolwine, B. J., Harrison, D. G., Raison, C. L., & Miller, A. H. (2013). Tyrosine metabolism during interferon-alpha administration: association with fatigue and CSF dopamine concentrations. Brain Behav Immun, 31, 153-160. doi: 10.1016/j.bbi.2012.10.010

Felger, J. C., & Miller, A. H. (2012). Cytokine effects on the basal ganglia and dopamine function: the subcortical source of inflammatory malaise. Front Neuroendocrinol, 33(3), 315-327. doi: 10.1016/j.yfrne.2012.09.003

Ferreira, M. A., O'Donovan, M. C., Meng, Y. A., Jones, I. R., Ruderfer, D. M., Jones, L., . . . Craddock, N. (2008). Collaborative genome-wide association analysis supports a role for ANK3 and CACNA1C in bipolar disorder. Nat Genet, 40(9), 1056-1058.

Ferrie, L., Young, A. H., & McQuade, R. (2006). Effect of lithium and lithium withdrawal on potassium-evoked dopamine release and tyrosine hydroxylase expression in the rat. Int J Neuropsychopharmacol, 9(6), 729-735.

Ficz, G., Branco, M. R., Seisenberger, S., Santos, F., Krueger, F., Hore, T. A., . . . Reik, W. (2011). Dynamic regulation of 5-hydroxymethylcytosine in mouse ES cells and during differentiation. Nature, 473(7347), 398-402. doi: 10.1038/nature10008

Fineberg, A. M., & Ellman, L. M. (2013). Inflammatory cytokines and neurological and neurocognitive alterations in the course of schizophrenia. Biol Psychiatry, 73(10), 951-966. doi: 10.1016/j.biopsych.2013.01.001

Fleckenstein, A. E., Metzger, R. R., Beyeler, M. L., Gibb, J. W., & Hanson, G. R. (1997). Oxygen radicals diminish dopamine transporter function in rat striatum. Eur J Pharmacol, 334(1), 111-114.

Forester, B. P., Zuo, C. S., Ravichandran, C., Harper, D. G., Du, F., Kim, S., . . . Renshaw, P. F. (2012). Coenzyme Q10 effects on creatine kinase activity and mood in geriatric bipolar depression. J Geriatr Psychiatry Neurol, 25(1), 43-50. doi: 10.1177/0891988712436688

148

Fountoulakis, K. N., Gonda, X., Siamouli, M., & Rihmer, Z. (2009). Psychotherapeutic intervention and suicide risk reduction in bipolar disorder: a review of the evidence. J Affect Disord, 113(1-2), 21-29. doi: 10.1016/j.jad.2008.06.014

Frey, B. N., Stanley, J. A., Nery, F. G., Monkul, E. S., Nicoletti, M. A., Chen, H. H., . . . Soares, J. C. (2007). Abnormal cellular energy and phospholipid metabolism in the left dorsolateral prefrontal cortex of medication-free individuals with bipolar disorder: an in vivo 1H MRS study. Bipolar Disord, 9 Suppl 1, 119-127.

Frey, B. N., Valvassori, S. S., Réus, G. Z., Martins, M. R., Petronilho, F. C., Bardini, K., . . . Quevedo, J. (2006). Effects of lithium and valproate on amphetamine-induced oxidative stress generation in an animal model of mania. J Psychiatry Neurosci, 31(5), 326-332.

Furlong, R. A., Coleman, T. A., Ho, L., Rubinsztein, J. S., Walsh, C., Paykel, E. S., & Rubinsztein, D. C. (1998). No association of a functional polymorphism in the dopamine D2 receptor promoter region with bipolar or unipolar affective disorders. Am J Med Genet, 81(5), 385-387.

Gambarana, C., Ghiglieri, O., Masi, F., Scheggi, S., Tagliamonte, A., & De Montis, M. G. (1999). The effects of long-term administration of rubidium or lithium on reactivity to stress and on dopamine output in the nucleus accumbens in rats. Brain Res, 826(2), 200-209.

Gawryluk, J. W., Wang, J. F., Andreazza, A. C., Shao, L., & Young, L. T. (2011). Decreased levels of glutathione, the major brain antioxidant, in post-mortem prefrontal cortex from patients with psychiatric disorders. Int J Neuropsychopharmacol, 14(1), 123-130.

Gerner, R. H., Post, R. M., & Bunney, W. E., Jr. (1976). A dopaminergic mechanism in mania. Am J Psychiatry, 133(10), 1177-1180.

Giros, B., el Mestikawy, S., Godinot, N., Zheng, K., Han, H., Yang-Feng, T., & Caron, M. G. (1992). Cloning, pharmacological characterization, and chromosome assignment of the human dopamine transporter. Mol Pharmacol, 42(3), 383-390.

Glahn, D. C., Bearden, C. E., Cakir, S., Barrett, J. A., Najt, P., Serap Monkul, E., . . . Soares, J. C. (2006). Differential working memory impairment in bipolar disorder and schizophrenia: effects of lifetime history of psychosis. Bipolar Disord, 8(2), 117-123. doi: 10.1111/j.1399-5618.2006.00296.x

Goldberg, J. F., Burdick, K. E., & Endick, C. J. (2004). Preliminary randomized, double-blind, placebo-controlled trial of pramipexole added to mood stabilizers for treatment-resistant bipolar depression. Am J Psychiatry, 161(3), 564-566.

Goldstein, B. I., Fagiolini, A., Houck, P., & Kupfer, D. J. (2009). Cardiovascular disease and hypertension among adults with bipolar I disorder in the United States. Bipolar Disord, 11(6), 657-662.

Goldstein, B. I., Kemp, D. E., Soczynska, J. K., & McIntyre, R. S. (2009). Inflammation and the phenomenology, pathophysiology, comorbidity, and treatment of bipolar disorder: a systematic review of the literature. J Clin Psychiatry, 70(8), 1078-1090. doi: 10.4088/JCP.08r04505

Goodwin, G. M., & Malhi, G. S. (2007). What is a mood stabilizer? Psychol Med, 37(5), 609-614. doi: 10.1017/S0033291706009305

Gracie, J. A., Robertson, S. E., & McInnes, I. B. (2003). Interleukin-18. J Leukoc Biol, 73(2), 213-224.

Graham, D. G. (1978). Oxidative pathways for catecholamines in the genesis of neuromelanin and cytotoxic quinones. Mol Pharmacol, 14(4), 633-643.

Graham, D. G., Tiffany, S. M., Bell, W. R., Jr., & Gutknecht, W. F. (1978). Autoxidation versus covalent binding of quinones as the mechanism of toxicity of dopamine, 6-

149

hydroxydopamine, and related compounds toward C1300 neuroblastoma cells in vitro. Mol Pharmacol, 14(4), 644-653.

Grayson, D. R., & Guidotti, A. (2013). The dynamics of DNA methylation in schizophrenia and related psychiatric disorders. Neuropsychopharmacology, 38(1), 138-166. doi: 10.1038/npp.2012.125

Green, D. R., & Kroemer, G. (2004). The pathophysiology of mitochondrial cell death. Science, 305(5684), 626-629.

Greenacre, S. A., & Ischiropoulos, H. (2001). Tyrosine nitration: localisation, quantification, consequences for protein function and signal transduction. Free Radic Res, 34(6), 541-581.

Guo, J. U., Ma, D. K., Mo, H., Ball, M. P., Jang, M. H., Bonaguidi, M. A., . . . Song, H. (2011). Neuronal activity modifies the DNA methylation landscape in the adult brain. Nat Neurosci, 14(10), 1345-1351.

Guo, J. U., Su, Y., Zhong, C., Ming, G. L., & Song, H. (2011). Hydroxylation of 5-methylcytosine by TET1 promotes active DNA demethylation in the adult brain. Cell, 145(3), 423-434. doi: 10.1016/j.cell.2011.03.022

Haddad, J. (2002). Antioxidant and prooxidant mechanisms in the regulation of redox(y)-sensitive transcription factors. Cellular Signalling, 14(11), 879-897. doi: 10.1016/S0898-6568(02)00053-0

Hallahan, B., Newell, J., Soares, J. C., Brambilla, P., Strakowski, S. M., Fleck, D. E., . . . McDonald, C. (2011). Structural magnetic resonance imaging in bipolar disorder: an international collaborative mega-analysis of individual adult patient data. Biol Psychiatry, 69(4), 326-335. doi: 10.1016/j.biopsych.2010.08.029

Halliwell, B. (1992). Reactive oxygen species and the central nervous system. J Neurochem, 59(5), 1609-1623.

Halliwell, B. (2001). Role of free radicals in the neurodegenerative diseases: therapeutic implications for antioxidant treatment. Drugs Aging, 18(9), 685-716.

Halliwell, B. (2006). Oxidative stress and neurodegeneration: where are we now? J Neurochem, 97(6), 1634-1658.

Halliwell, B. (2007). Biochemistry of oxidative stress. Biochem Soc Trans, 35(Pt 5), 1147-1150. Halliwell, B., & Gutteridge, J. M. (1984). Free radicals, lipid peroxidation, and cell damage. Lancet,

2(8411), 1095. Halliwell, B., & Gutteridge, J. M. C. (2007). Free Radicals in Biology and Medicine (4 ed.). Oxford:

Clarendon Press. Hanamsagar, R., Torres, V., & Kielian, T. (2011). Inflammasome activation and IL-1β/IL-18

processing are influenced by distinct pathways in microglia. J Neurochem, 119(4), 736-748. doi: 10.1111/j.1471-4159.2011.07481.x

Hardan, A. Y., Fung, L. K., Libove, R. A., Obukhanych, T. V., Nair, S., Herzenberg, L. A., . . . Tirouvanziam, R. (2012). A randomized controlled pilot trial of oral N-acetylcysteine in children with autism. Biol Psychiatry, 71(11), 956-961. doi: 10.1016/j.biopsych.2012.01.014

Harish, G., Venkateshappa, C., Mahadevan, A., Pruthi, N., Srinivas Bharath, M. M., & Shankar, S. K. (2011). Glutathione metabolism is modulated by postmortem interval, gender difference and agonal state in postmortem human brains. Neurochem Int, 59(7), 1029-1042. doi: 10.1016/j.neuint.2011.08.024

Hashimoto, R., Takei, N., Shimazu, K., Christ, L., Lu, B., & Chuang, D. M. (2002). Lithium induces brain-derived neurotrophic factor and activates TrkB in rodent cortical neurons: an essential step for neuroprotection against glutamate excitotoxicity. Neuropharmacology, 43(7), 1173-1179.

150

Hastings, T. G. (2009). The role of dopamine oxidation in mitochondrial dysfunction: implications for Parkinson's disease. J Bioenerg Biomembr, 41(6), 469-472.

Hastings, T. G., Lewis, D. A., & Zigmond, M. J. (1996). Role of oxidation in the neurotoxic effects of intrastriatal dopamine injections. Proc Natl Acad Sci U S A, 93(5), 1956-1961.

Hastings, T. G., & Zigmond, M. J. (1994). Identification of catechol-protein conjugates in neostriatal slices incubated with [3H]dopamine: impact of ascorbic acid and glutathione. J Neurochem, 63(3), 1126-1132.

Hernandez, D. G., & Singleton, A. B. (2012). Using DNA methylation to understand biological consequences of genetic variability. Neurodegener Dis, 9(2), 53-59. doi: 10.1159/000333097

Hope, S., Dieset, I., Agartz, I., Steen, N. E., Ueland, T., Melle, I., . . . Andreassen, O. A. (2011). Affective symptoms are associated with markers of inflammation and immune activation in bipolar disorders but not in schizophrenia. J Psychiatr Res, 45(12), 1608-1616. doi: 10.1016/j.jpsychires.2011.08.003

Howes, O. D., & Kapur, S. (2009). The dopamine hypothesis of schizophrenia: version III--the final common pathway. Schizophr Bull, 35(3), 549-562. doi: 10.1093/schbul/sbp006

Huzayyin, A. A., Andreazza, A. C., Turecki, G., Cruceanu, C., Rouleau, G. A., Alda, M., & Young, L. T. (2013). Decreased global methylation in patients with bipolar disorder who respond to lithium. Int J Neuropsychopharmacol, 1-9. doi: 10.1017/S1461145713001569

Huzayyin, A. A., Andreazza, A. C., Turecki, G., Cruceanu, C., Rouleau, G. A., Alda, M., & Young, L. T. (2014). Decreased global methylation in patients with bipolar disorder who respond to lithium. Int J Neuropsychopharmacol, 17(4), 561-569. doi: 10.1017/S1461145713001569

Hynd, M., Lewohl, J., Scott, H., & Dodd, P. (2003). Biochemical and molecular studies using human autopsy brain tissue. Journal of Neurochemistry, 85(3), 543-562. doi: 10.1046/j.1471-4159.2003.01747.x

Ito, S., D'Alessio, A., Taranova, O., Hong, K., Sowers, L., & Zhang, Y. (2010). Role of Tet proteins in 5mC to 5hmC conversion, ES-cell self-renewal and inner cell mass specification. Nature, 466(7310), 1129-U1151. doi: 10.1038/nature09303

Iwamoto, K., Bundo, M., & Kato, T. (2005). Altered expression of mitochondria-related genes in postmortem brains of patients with bipolar disorder or schizophrenia, as revealed by large-scale DNA microarray analysis. Hum Mol Genet, 14(2), 241-253.

Iwamoto, K., Kakiuchi, C., Bundo, M., Ikeda, K., & Kato, T. (2004). Molecular characterization of bipolar disorder by comparing gene expression profiles of postmortem brains of major mental disorders. Mol Psychiatry, 9(4), 406-416.

Jaber, M., Jones, S., Giros, B., & Caron, M. G. (1997). The dopamine transporter: a crucial component regulating dopamine transmission. Mov Disord, 12(5), 629-633. doi: 10.1002/mds.870120502

Janssen, R. J., Nijtmans, L. G., van den Heuvel, L. P., & Smeitink, J. A. (2006). Mitochondrial complex I: structure, function and pathology. J Inherit Metab Dis, 29(4), 499-515. doi: 10.1007/s10545-006-0362-4

Jarskog, L. F. (2006). Apoptosis in schizophrenia: pathophysiologic and therapeutic considerations. Curr Opin Psychiatry, 19(3), 307-312. doi: 10.1097/01.yco.0000218603.25346.8f

Jeong, S. Y., & Seol, D. W. (2008). The role of mitochondria in apoptosis. BMB Rep, 41(1), 11-22. Jin, S., Kadam, S., & Pfeifer, G. (2010). Examination of the specificity of DNA methylation

profiling techniques towards 5-methylcytosine and 5-hydroxymethylcytosine. Nucleic Acids Research, 38(11). doi: 10.1093/nar/gkq223

Jones, D. P. (2010). Redox sensing: orthogonal control in cell cycle and apoptosis signalling. J Intern Med, 268(5), 432-448. doi: 10.1111/j.1365-2796.2010.02268.x

151

Jornada, L. K., Valvassori, S. S., Steckert, A. V., Moretti, M., Mina, F., Ferreira, C. L., . . . Quevedo, J. (2011). Lithium and valproate modulate antioxidant enzymes and prevent ouabain-induced oxidative damage in an animal model of mania. J Psychiatr Res, 45(2), 162-168. doi: S0022-3956(10)00163-9 [pii] 10.1016/j.jpsychires.2010.05.011

Jurata, L. W., Bukhman, Y. V., Charles, V., Capriglione, F., Bullard, J., Lemire, A. L., . . . Altar, C. A. (2004). Comparison of microarray-based mRNA profiling technologies for identification of psychiatric disease and drug signatures. J Neurosci Methods, 138(1-2), 173-188. doi: 10.1016/j.jneumeth.2004.04.002

Kamat, J. P. (2006). Peroxynitrite: a potent oxidizing and nitrating agent. Indian J Exp Biol, 44(6), 436-447.

Kang, H. J., Voleti, B., Hajszan, T., Rajkowska, G., Stockmeier, C. A., Licznerski, P., . . . Duman, R. S. (2012). Decreased expression of synapse-related genes and loss of synapses in major depressive disorder. Nat Med, 18(9), 1413-1417. doi: nm.2886 [pii]

10.1038/nm.2886

Kang, K. A., Zhang, R., Kim, G. Y., Bae, S. C., & Hyun, J. W. (2012). Epigenetic changes induced by oxidative stress in colorectal cancer cells: methylation of tumor suppressor RUNX3. Tumour Biol, 33(2), 403-412. doi: 10.1007/s13277-012-0322-6

Karimollah, A. R., Ghasemi, M., Ghahremani, M. H., & Dehpour, A. R. (2009). Inhibition by lithium of the nitrergic relaxation of rat anococcygeus muscle. Nitric Oxide, 20(1), 31-38. doi: 10.1016/j.niox.2008.08.007

Kato, T. (2005). [Mitochondrial dysfunction in bipolar disorder]. Nihon Shinkei Seishin Yakurigaku Zasshi, 25(2), 61-72.

Kato, T. (2008). Role of mitochondrial DNA in calcium signaling abnormality in bipolar disorder. Cell Calcium, 44(1), 92-102.

Kato, T., & Kato, N. (2000). Mitochondrial dysfunction in bipolar disorder. Bipolar Disord, 2(3 Pt 1), 180-190.

Kato, T., Murashita, J., Kamiya, A., Shioiri, T., Kato, N., & Inubushi, T. (1998). Decreased brain intracellular pH measured by 31P-MRS in bipolar disorder: a confirmation in drug-free patients and correlation with white matter hyperintensity. Eur Arch Psychiatry Clin Neurosci, 248(6), 301-306.

Kato, T., Takahashi, S., Shioiri, T., & Inubushi, T. (1993). Alterations in brain phosphorous metabolism in bipolar disorder detected by in vivo 31P and 7Li magnetic resonance spectroscopy. J Affect Disord, 27(1), 53-59.

Kauer-Sant'Anna, M., Kapczinski, F., Andreazza, A. C., Bond, D. J., Lam, R. W., Young, L. T., & Yatham, L. N. (2009). Brain-derived neurotrophic factor and inflammatory markers in patients with early- vs. late-stage bipolar disorder. Int J Neuropsychopharmacol, 12(4), 447-458. doi: 10.1017/S1461145708009310

Keese, M., Magdeburg, R. J., Herzog, T., Hasenberg, T., Offterdinger, M., Pepperkok, R., . . . Bastiaens, P. I. (2005). Imaging epidermal growth factor receptor phosphorylation in human colorectal cancer cells and human tissues. J Biol Chem, 280(30), 27826-27831. doi: M504485200 [pii]

10.1074/jbc.M504485200

Kempton, M. J., Geddes, J. R., Ettinger, U., Williams, S. C., & Grasby, P. M. (2008). Meta-analysis, database, and meta-regression of 98 structural imaging studies in bipolar disorder. Arch Gen Psychiatry, 65(9), 1017-1032. doi: 10.1001/archpsyc.65.9.1017

Kenworthy, A. K. (2001). Imaging protein-protein interactions using fluorescence resonance energy transfer microscopy. Methods, 24(3), 289-296. doi: S1046-2023(01)91189-2 [pii]

152

10.1006/meth.2001.1189

Kilbourne, A. M., Cornelius, J. R., Han, X., Pincus, H. A., Shad, M., Salloum, I., . . . Haas, G. L. (2004). Burden of general medical conditions among individuals with bipolar disorder. Bipolar Disord, 6(5), 368-373. doi: 10.1111/j.1399-5618.2004.00138.x

Kim, H. K., & Andreazza, A. C. (2012). The relationship between oxidative stress and post-translational modification of the dopamine transporter in bipolar disorder. Expert Rev Neurother, 12(7), 849-859. doi: 10.1586/ern.12.64

Kim, H. K., Andreazza, A. C., Yeung, P. Y., Isaacs-Trepanier, C., & Young, L. T. (2014). Oxidation and nitration in dopaminergic areas of the prefrontal cortex from patients with bipolar disorder and schizophrenia. J Psychiatry Neurosci, 39(4), 276-285.

Kim, H. W., Rapoport, S. I., & Rao, J. S. (2010). Altered expression of apoptotic factors and synaptic markers in postmortem brain from bipolar disorder patients. Neurobiol Dis, 37(3), 596-603.

Kim, H. W., Rapoport, S. I., & Rao, J. S. (2011). Altered arachidonic acid cascade enzymes in postmortem brain from bipolar disorder patients. Mol Psychiatry, 16(4), 419-428. doi: 10.1038/mp.2009.137

Kim, Y. K., Jung, H. G., Myint, A. M., Kim, H., & Park, S. H. (2007). Imbalance between pro-inflammatory and anti-inflammatory cytokines in bipolar disorder. J Affect Disord, 104(1-3), 91-95. doi: 10.1016/j.jad.2007.02.018

King, T. D., Bijur, G. N., & Jope, R. S. (2001). Caspase-3 activation induced by inhibition of mitochondrial complex I is facilitated by glycogen synthase kinase-3beta and attenuated by lithium. Brain Res, 919(1), 106-114.

Konradi, C., Eaton, M., MacDonald, M. L., Walsh, J., Benes, F. M., & Heckers, S. (2004). Molecular evidence for mitochondrial dysfunction in bipolar disorder. Arch Gen Psychiatry, 61(3), 300-308.

Konradi, C., Sillivan, S. E., & Clay, H. B. (2012). Mitochondria, oligodendrocytes and inflammation in bipolar disorder: evidence from transcriptome studies points to intriguing parallels with multiple sclerosis. Neurobiol Dis, 45(1), 37-47. doi: 10.1016/j.nbd.2011.01.025

Krämer, D. K., Bouzakri, K., Holmqvist, O., Al-Khalili, L., & Krook, A. (2005). Effect of serum replacement with plysate on cell growth and metabolismin primary cultures of human skeletal muscle. Cytotechnology, 48(1-3), 89-95. doi: 10.1007/s10616-005-4074-7

Kulak, A., Steullet, P., Cabungcal, J. H., Werge, T., Ingason, A., Cuenod, M., & Do, K. Q. (2013). Redox dysregulation in the pathophysiology of schizophrenia and bipolar disorder: insights from animal models. Antioxid Redox Signal, 18(12), 1428-1443. doi: 10.1089/ars.2012.4858

Kuloglu, M., Ustundag, B., Atmaca, M., Canatan, H., Tezcan, A. E., & Cinkilinc, N. (2002). Lipid peroxidation and antioxidant enzyme levels in patients with schizophrenia and bipolar disorder. Cell Biochem Funct, 20(2), 171-175.

Kupfer, D. J. (2005). The increasing medical burden in bipolar disorder. Jama, 293(20), 2528-2530. König, P., Krasteva, G., Tag, C., König, I. R., Arens, C., & Kummer, W. (2006). FRET-CLSM and

double-labeling indirect immunofluorescence to detect close association of proteins in tissue sections. Lab Invest, 86(8), 853-864. doi: 3700443 [pii]

10.1038/labinvest.3700443

Lai, C. T., & Yu, P. H. (1997). Dopamine- and L-beta-3,4-dihydroxyphenylalanine hydrochloride (L-Dopa)-induced cytotoxicity towards catecholaminergic neuroblastoma SH-SY5Y cells. Effects of oxidative stress and antioxidative factors. Biochem Pharmacol, 53(3), 363-372.

153

Lai, J. S., Zhao, C., Warsh, J. J., & Li, P. P. (2006). Cytoprotection by lithium and valproate varies between cell types and cellular stresses. Eur J Pharmacol, 539(1-2), 18-26. doi: 10.1016/j.ejphar.2006.03.076

LaVoie, M. J., & Hastings, T. G. (1999). Peroxynitrite- and nitrite-induced oxidation of dopamine: implications for nitric oxide in dopaminergic cell loss. J Neurochem, 73(6), 2546-2554.

Lebon, S., Minai, L., Chretien, D., Corcos, J., Serre, V., Kadhom, N., . . . Rotig, A. (2007). A novel mutation of the NDUFS7 gene leads to activation of a cryptic exon and impaired assembly of mitochondrial complex I in a patient with Leigh syndrome. Mol Genet Metab, 92(1-2), 104-108.

Leboyer, M., Soreca, I., Scott, J., Frye, M., Henry, C., Tamouza, R., & Kupfer, D. J. (2012). Can bipolar disorder be viewed as a multi-system inflammatory disease? J Affect Disord, 141(1), 1-10. doi: 10.1016/j.jad.2011.12.049

Lenaz, G. (2001). The mitochondrial production of reactive oxygen species: mechanisms and implications in human pathology. IUBMB Life, 52(3-5), 159-164.

Lertratanangkoon, K., Wu, C. J., Savaraj, N., & Thomas, M. L. (1997). Alterations of DNA methylation by glutathione depletion. Cancer Lett, 120(2), 149-156.

Lewis, D. A., Melchitzky, D. S., Sesack, S. R., Whitehead, R. E., Auh, S., & Sampson, A. (2001). Dopamine transporter immunoreactivity in monkey cerebral cortex: regional, laminar, and ultrastructural localization. J Comp Neurol, 432(1), 119-136.

Li, N., Ragheb, K., Lawler, G., Sturgis, J., Rajwa, B., Melendez, J. A., & Robinson, J. P. (2003). Mitochondrial complex I inhibitor rotenone induces apoptosis through enhancing mitochondrial reactive oxygen species production. J Biol Chem, 278(10), 8516-8525. doi: 10.1074/jbc.M210432200

Li, Q., Li, H., Roughton, K., Wang, X., Kroemer, G., Blomgren, K., & Zhu, C. (2010). Lithium reduces apoptosis and autophagy after neonatal hypoxia-ischemia. Cell Death Dis, 1, e56. doi: 10.1038/cddis.2010.33

Libíková, H., Breier, S., Kocisová, M., Pogády, J., Stünzner, D., & Ujházyová, D. (1979). Assay of interferon and viral antibodies in the cerebrospinal fluid in clinical neurology and psychiatry. Acta Biol Med Ger, 38(5-6), 879-893.

Lim, S. O., Gu, J. M., Kim, M. S., Kim, H. S., Park, Y. N., Park, C. K., . . . Jung, G. (2008). Epigenetic changes induced by reactive oxygen species in hepatocellular carcinoma: methylation of the E-cadherin promoter. Gastroenterology, 135(6), 2128-2140, 2140.e2121-2128. doi: 10.1053/j.gastro.2008.07.027

Lister, R., Mukamel, E. A., Nery, J. R., Urich, M., Puddifoot, C. A., Johnson, N. D., . . . Ecker, J. R. (2013). Global epigenomic reconfiguration during mammalian brain development. Science, 341(6146), 1237905. doi: 10.1126/science.1237905

Liu, H. C., Yang, Y. Y., Chou, Y. M., Chen, K. P., Shen, W. W., & Leu, S. J. (2004). Immunologic variables in acute mania of bipolar disorder. J Neuroimmunol, 150(1-2), 116-122. doi: 10.1016/j.jneuroim.2004.01.006

Lohoff, F. W., Sander, T., Ferraro, T. N., Dahl, J. P., Gallinat, J., & Berrettini, W. H. (2005). Confirmation of association between the Val66Met polymorphism in the brain-derived neurotrophic factor (BDNF) gene and bipolar I disorder. Am J Med Genet B Neuropsychiatr Genet, 139B(1), 51-53. doi: 10.1002/ajmg.b.30215

Lopez-Larson, M. P., DelBello, M. P., Zimmerman, M. E., Schwiers, M. L., & Strakowski, S. M. (2002). Regional prefrontal gray and white matter abnormalities in bipolar disorder. Biol Psychiatry, 52(2), 93-100.

Lowes, D. A., Almawash, A. M., Webster, N. R., Reid, V. L., & Galley, H. F. (2011). Melatonin and structurally similar compounds have differing effects on inflammation and mitochondrial

154

function in endothelial cells under conditions mimicking sepsis. Br J Anaesth, 107(2), 193-201. doi: 10.1093/bja/aer149

Lyoo, I. K., Kim, M. J., Stoll, A. L., Demopulos, C. M., Parow, A. M., Dager, S. R., . . . Renshaw, P. F. (2004). Frontal lobe gray matter density decreases in bipolar I disorder. Biol Psychiatry, 55(6), 648-651.

López-Armada, M. J., Riveiro-Naveira, R. R., Vaamonde-García, C., & Valcárcel-Ares, M. N. (2013). Mitochondrial dysfunction and the inflammatory response. Mitochondrion, 13(2), 106-118. doi: 10.1016/j.mito.2013.01.003

Machado-Vieira, R., Andreazza, A. C., Viale, C. I., Zanatto, V., Cereser, V., Jr., da Silva Vargas, R., . . . Gentil, V. (2007). Oxidative stress parameters in unmedicated and treated bipolar subjects during initial manic episode: a possible role for lithium antioxidant effects. Neurosci Lett, 421(1), 33-36.

Machado-Vieira, R., Dietrich, M. O., Leke, R., Cereser, V. H., Zanatto, V., Kapczinski, F., . . . Gentil, V. (2007). Decreased plasma brain derived neurotrophic factor levels in unmedicated bipolar patients during manic episode. Biol Psychiatry, 61(2), 142-144. doi: 10.1016/j.biopsych.2006.03.070

Maes, M., Bosmans, E., Calabrese, J., Smith, R., & Meltzer, H. Y. (1995). Interleukin-2 and interleukin-6 in schizophrenia and mania: effects of neuroleptics and mood stabilizers. J Psychiatr Res, 29(2), 141-152.

Magalhães, P. V., Dean, O. M., Bush, A. I., Copolov, D. L., Malhi, G. S., Kohlmann, K., . . . Berk, M. (2011). N-acetyl cysteine add-on treatment for bipolar II disorder: a subgroup analysis of a randomized placebo-controlled trial. J Affect Disord, 129(1-3), 317-320. doi: 10.1016/j.jad.2010.08.001

Maier, W., Zobel, A., & Wagner, M. (2006). Schizophrenia and bipolar disorder: differences and overlaps. Curr Opin Psychiatry, 19(2), 165-170. doi: 10.1097/01.yco.0000214342.52249.82

Maker, H. S., Weiss, C., Silides, D. J., & Cohen, G. (1981). Coupling of dopamine oxidation (monoamine oxidase activity) to glutathione oxidation via the generation of hydrogen peroxide in rat brain homogenates. J Neurochem, 36(2), 589-593.

Malhi, G. S., Tanious, M., Das, P., Coulston, C. M., & Berk, M. (2013). Potential mechanisms of action of lithium in bipolar disorder. Current understanding. CNS Drugs, 27(2), 135-153. doi: 10.1007/s40263-013-0039-0

Manji, H. K., & Lenox, R. H. (2000). The nature of bipolar disorder. J Clin Psychiatry, 61 Supp 13, 42-57.

Maragos, W. F., Zhu, J., Chesnut, M. D., & Dwoskin, L. P. (2002). Mitochondrial toxin inhibition of [(3)H]dopamine uptake into rat striatal synaptosomes. Biochem Pharmacol, 63(8), 1499-1505.

Martinez-Aran, A., Vieta, E., Reinares, M., Colom, F., Torrent, C., Sanchez-Moreno, J., . . . Salamero, M. (2004). Cognitive function across manic or hypomanic, depressed, and euthymic states in bipolar disorder. Am J Psychiatry, 161(2), 262-270.

Mastroberardino, P. G., Orr, A. L., Hu, X., Na, H. M., & Greenamyre, J. T. (2008). A FRET-based method to study protein thiol oxidation in histological preparations. Free Radic Biol Med, 45(7), 971-981.

Matarese, F., Carrillo-de Santa Pau, E., & Stunnenberg, H. G. (2011). 5-Hydroxymethylcytosine: a new kid on the epigenetic block? Mol Syst Biol, 7, 562. doi: 10.1038/msb.2011.95

Maurer, I. C., Schippel, P., & Volz, H. P. (2009). Lithium-induced enhancement of mitochondrial oxidative phosphorylation in human brain tissue. Bipolar Disord, 11(5), 515-522. doi: 10.1111/j.1399-5618.2009.00729.x

155

Merikangas, K. R., Jin, R., He, J. P., Kessler, R. C., Lee, S., Sampson, N. A., . . . Zarkov, Z. (2011). Prevalence and correlates of bipolar spectrum disorder in the world mental health survey initiative. Arch Gen Psychiatry, 68(3), 241-251.

Messiha, F. S., Agallianos, D., & Clower, C. (1970). Dopamine excretion in affective states and following Li2Co3 therapy. Nature, 225(5235), 868-869.

Miller, B. J., Buckley, P., Seabolt, W., Mellor, A., & Kirkpatrick, B. (2011). Meta-analysis of cytokine alterations in schizophrenia: clinical status and antipsychotic effects. Biol Psychiatry, 70(7), 663-671. doi: 10.1016/j.biopsych.2011.04.013

Miller, G. W., Staley, J. K., Heilman, C. J., Perez, J. T., Mash, D. C., Rye, D. B., & Levey, A. I. (1997). Immunochemical analysis of dopamine transporter protein in Parkinson's disease. Ann Neurol, 41(4), 530-539. doi: 10.1002/ana.410410417

Mirnics, K., Levitt, P., & Lewis, D. A. (2006). Critical appraisal of DNA microarrays in psychiatric genomics. Biol Psychiatry, 60(2), 163-176. doi: 10.1016/j.biopsych.2006.02.003

Mitchell, P., Selbie, L., Waters, B., Donald, J., Vivero, C., Tully, M., & Shine, J. (1992). Exclusion of close linkage of bipolar disorder to dopamine D1 and D2 receptor gene markers. J Affect Disord, 25(1), 1-11.

Miyazaki, I., Asanuma, M., Diaz-Corrales, F., Fukuda, M., Kitaichi, K., Miyoshi, K., & Ogawa, N. (2006). Methamphetamine-induced dopaminergic neurotoxicity is regulated by quinone formation-related molecules. Faseb Journal, 20(1), 571-+. doi: 10.1096/fj.05-4996fje

Modabbernia, A., Taslimi, S., Brietzke, E., & Ashrafi, M. (2013). Cytokine alterations in bipolar disorder: a meta-analysis of 30 studies. Biol Psychiatry, 74(1), 15-25. doi: 10.1016/j.biopsych.2013.01.007

Muir, W. J., Thomson, M. L., McKeon, P., Mynett-Johnson, L., Whitton, C., Evans, K. L., . . . Blackwood, D. H. (2001). Markers close to the dopamine D5 receptor gene (DRD5) show significant association with schizophrenia but not bipolar disorder. Am J Med Genet, 105(2), 152-158.

Muller-Oerlinghausen, B., Berghofer, A., & Bauer, M. (2002). Bipolar disorder. Lancet, 359(9302), 241-247.

Munkholm, K., Braüner, J. V., Kessing, L. V., & Vinberg, M. (2013). Cytokines in bipolar disorder vs. healthy control subjects: a systematic review and meta-analysis. J Psychiatr Res, 47(9), 1119-1133. doi: 10.1016/j.jpsychires.2013.05.018

Müller, N., Myint, A. M., & Schwarz, M. J. (2012). Inflammation in schizophrenia. Adv Protein Chem Struct Biol, 88, 49-68. doi: 10.1016/B978-0-12-398314-5.00003-9

Nahman, S., Belmaker, R. H., & Azab, A. N. (2012). Effects of lithium on lipopolysaccharide-induced inflammation in rat primary glia cells. Innate Immun, 18(3), 447-458. doi: 10.1177/1753425911421512

Naoi, M., Maruyama, W., Shamoto-Nagai, M., Yi, H., Akao, Y., & Tanaka, M. (2005). Oxidative stress in mitochondria: decision to survival and death of neurons in neurodegenerative disorders. Mol Neurobiol, 31(1-3), 81-93.

Nascimento, C., Kim, H. K., Young, L. T., Mendonça, K. M., Grinberg, L. T., Lafer, B., & Andreazza, A. C. (2014). Glutathione-mediated effects of lithium in decreasing protein oxidation induced by mitochondrial complex I dysfunction. J Neural Transm. doi: 10.1007/s00702-014-1318-8

Nery, F. G., Monkul, E. S., Hatch, J. P., Fonseca, M., Zunta-Soares, G. B., Frey, B. N., . . . Soares, J. C. (2008). Celecoxib as an adjunct in the treatment of depressive or mixed episodes of bipolar disorder: a double-blind, randomized, placebo-controlled study. Hum Psychopharmacol, 23(2), 87-94. doi: 10.1002/hup.912

156

Ng, F., Berk, M., Dean, O., & Bush, A. I. (2008). Oxidative stress in psychiatric disorders: evidence base and therapeutic implications. Int J Neuropsychopharmacol, 11(6), 851-876.

Nikolaus, S., Antke, C., & Muller, H. W. (2009). In vivo imaging of synaptic function in the central nervous system: II. Mental and affective disorders. Behav Brain Res, 204(1), 32-66.

Nohesara, S., Ghadirivasfi, M., Mostafavi, S., Eskandari, M. R., Ahmadkhaniha, H., Thiagalingam, S., & Abdolmaleky, H. M. (2011). DNA hypomethylation of MB-COMT promoter in the DNA derived from saliva in schizophrenia and bipolar disorder. J Psychiatr Res, 45(11), 1432-1438.

Nöthen, M. M., Erdmann, J., Körner, J., Lanczik, M., Fritze, J., Fimmers, R., . . . Propping, P. (1992). Lack of association between dopamine D1 and D2 receptor genes and bipolar affective disorder. Am J Psychiatry, 149(2), 199-201.

O'Brien, S. M., Scully, P., Scott, L. V., & Dinan, T. G. (2006). Cytokine profiles in bipolar affective disorder: focus on acutely ill patients. J Affect Disord, 90(2-3), 263-267. doi: 10.1016/j.jad.2005.11.015

O'Connell, L. A., Matthews, B. J., Ryan, M. J., & Hofmann, H. A. (2010). Characterization of the dopamine system in the brain of the túngara frog, Physalaemus pustulosus. Brain Behav Evol, 76(3-4), 211-225. doi: 10.1159/000321715

Offen, D., Ziv, I., Sternin, H., Melamed, E., & Hochman, A. (1996). Prevention of dopamine-induced cell death by thiol antioxidants: possible implications for treatment of Parkinson's disease. Exp Neurol, 141(1), 32-39.

Ortiz-Domínguez, A., Hernández, M. E., Berlanga, C., Gutiérrez-Mora, D., Moreno, J., Heinze, G., & Pavón, L. (2007). Immune variations in bipolar disorder: phasic differences. Bipolar Disord, 9(6), 596-602. doi: 10.1111/j.1399-5618.2007.00493.x

Padmos, R. C., Hillegers, M. H., Knijff, E. M., Vonk, R., Bouvy, A., Staal, F. J., . . . Drexhage, H. A. (2008). A discriminating messenger RNA signature for bipolar disorder formed by an aberrant expression of inflammatory genes in monocytes. Arch Gen Psychiatry, 65(4), 395-407. doi: 10.1001/archpsyc.65.4.395

Papiol, S., Molina, V., Desco, M., Rosa, A., Reig, S., Sanz, J., . . . Fañanás, L. (2008). Gray matter deficits in bipolar disorder are associated with genetic variability at interleukin-1 beta gene (2q13). Genes Brain Behav, 7(7), 796-801. doi: 10.1111/j.1601-183X.2008.00421.x

Park, S., Geddes, T. J., Javitch, J. A., & Kuhn, D. M. (2003). Dopamine prevents nitration of tyrosine hydroxylase by peroxynitrite and nitrogen dioxide: is nitrotyrosine formation an early step in dopamine neuronal damage? J Biol Chem, 278(31), 28736-28742. doi: M304362200 [pii]

10.1074/jbc.M304362200

Park, S. U., Ferrer, J. V., Javitch, J. A., & Kuhn, D. M. (2002). Peroxynitrite inactivates the human dopamine transporter by modification of cysteine 342: potential mechanism of neurotoxicity in dopamine neurons. J Neurosci, 22(11), 4399-4405.

Pearlson, G. D., Wong, D. F., Tune, L. E., Ross, C. A., Chase, G. A., Links, J. M., . . . Wagner, H. N. (1995). In vivo D2 dopamine receptor density in psychotic and nonpsychotic patients with bipolar disorder. Arch Gen Psychiatry, 52(6), 471-477.

Perova, T., Kwan, M., Li, P. P., & Warsh, J. J. (2010). Differential modulation of intracellular Ca2+ responses in B lymphoblasts by mood stabilizers. Int J Neuropsychopharmacol, 13(6), 693-702.

Perova, T., Wasserman, M. J., Li, P. P., & Warsh, J. J. (2008). Hyperactive intracellular calcium dynamics in B lymphoblasts from patients with bipolar I disorder. Int J Neuropsychopharmacol, 11(2), 185-196.

157

Perregaux, D., & Gabel, C. A. (1994). Interleukin-1 beta maturation and release in response to ATP and nigericin. Evidence that potassium depletion mediated by these agents is a necessary and common feature of their activity. J Biol Chem, 269(21), 15195-15203.

Persson, A. K., Kim, I., Zhao, P., Estacion, M., Black, J. A., & Waxman, S. G. (2013). Sodium channels contribute to degeneration of dorsal root ganglion neurites induced by mitochondrial dysfunction in an in vitro model of axonal injury. J Neurosci, 33(49), 19250-19261. doi: 10.1523/JNEUROSCI.2148-13.2013

Petronis, A. (2003). Epigenetics and bipolar disorder: new opportunities and challenges. Am J Med Genet C Semin Med Genet, 123C(1), 65-75. doi: 10.1002/ajmg.c.20015

Post, R. M., Jimerson, D. C., Bunney, W. E., Jr., & Goodwin, F. K. (1980). Dopamine and mania: behavioral and biochemical effects of the dopamine receptor blocker pimozide. Psychopharmacology (Berl), 67(3), 297-305.

Potash, J. B., & Bienvenu, O. J. (2009). Neuropsychiatric disorders: Shared genetics of bipolar disorder and schizophrenia. Nat Rev Neurol, 5(6), 299-300. doi: 10.1038/nrneurol.2009.71

Prickaerts, J., Moechars, D., Cryns, K., Lenaerts, I., van Craenendonck, H., Goris, I., . . . Steckler, T. (2006). Transgenic mice overexpressing glycogen synthase kinase 3beta: a putative model of hyperactivity and mania. J Neurosci, 26(35), 9022-9029. doi: 10.1523/JNEUROSCI.5216-05.2006

Quiroz, J. A., Gould, T. D., & Manji, H. K. (2004). Molecular effects of lithium. Mol Interv, 4(5), 259-272.

Quiroz, J. A., Gray, N. A., Kato, T., & Manji, H. K. (2008). Mitochondrially mediated plasticity in the pathophysiology and treatment of bipolar disorder. Neuropsychopharmacology, 33(11), 2551-2565.

Rajkowska, G. (2000). Postmortem studies in mood disorders indicate altered numbers of neurons and glial cells. Biol Psychiatry, 48(8), 766-777.

Rajkowska, G., Halaris, A., & Selemon, L. D. (2001). Reductions in neuronal and glial density characterize the dorsolateral prefrontal cortex in bipolar disorder. Biol Psychiatry, 49(9), 741-752.

Rao, J. S., Harry, G. J., Rapoport, S. I., & Kim, H. W. (2010). Increased excitotoxicity and neuroinflammatory markers in postmortem frontal cortex from bipolar disorder patients. Mol Psychiatry, 15(4), 384-392.

Rao, J. S., Kellom, M., Reese, E. A., Rapoport, S. I., & Kim, H. W. (2012). Dysregulated glutamate and dopamine transporters in postmortem frontal cortex from bipolar and schizophrenic patients. J Affect Disord, 136(1-2), 63-71. doi: S0165-0327(11)00489-7 [pii] 10.1016/j.jad.2011.08.017

Ray, P. D., Huang, B. W., & Tsuji, Y. (2012). Reactive oxygen species (ROS) homeostasis and redox regulation in cellular signaling. Cell Signal, 24(5), 981-990. doi: 10.1016/j.cellsig.2012.01.008

Reik, W., Dean, W., & Walter, J. (2001). Epigenetic reprogramming in mammalian development. Science, 293(5532), 1089-1093. doi: 10.1126/science.1063443

Romo-Nava, F., Alvarez-Icaza González, D., Fresán-Orellana, A., Saracco Alvarez, R., Becerra-Palars, C., Moreno, J., . . . Buijs, R. M. (2014). Melatonin attenuates antipsychotic metabolic effects: an eight-week randomized, double-blind, parallel-group, placebo-controlled clinical trial. Bipolar Disord, 16(4), 410-421. doi: 10.1111/bdi.12196

Saarelainen, T., Hendolin, P., Lucas, G., Koponen, E., Sairanen, M., MacDonald, E., . . . Castrén, E. (2003). Activation of the TrkB neurotrophin receptor is induced by antidepressant drugs and is required for antidepressant-induced behavioral effects. J Neurosci, 23(1), 349-357.

158

Salvadore, G., Quiroz, J. A., Machado-Vieira, R., Henter, I. D., Manji, H. K., & Zarate, C. A. (2010). The neurobiology of the switch process in bipolar disorder: a review. J Clin Psychiatry, 71(11), 1488-1501. doi: 10.4088/JCP.09r05259gre

Savoye, C., Laurent, C., Amadeo, S., Gheysen, F., Leboyer, M., Lejeune, J., . . . Mallet, J. (1998). No association between dopamine D1, D2, and D3 receptor genes and manic-depressive illness. Biol Psychiatry, 44(7), 644-647.

Schretlen, D. J., Cascella, N. G., Meyer, S. M., Kingery, L. R., Testa, S. M., Munro, C. A., . . . Pearlson, G. D. (2007). Neuropsychological functioning in bipolar disorder and schizophrenia. Biol Psychiatry, 62(2), 179-186. doi: 10.1016/j.biopsych.2006.09.025

Schroder, K., Zhou, R., & Tschopp, J. (2010). The NLRP3 inflammasome: a sensor for metabolic danger? Science, 327(5963), 296-300. doi: 10.1126/science.1184003

Schutte, B., Nuydens, R., Geerts, H., & Ramaekers, F. (1998). Annexin V binding assay as a tool to measure apoptosis in differentiated neuronal cells. J Neurosci Methods, 86(1), 63-69.

Schäfer, M., Goodenough, S., Moosmann, B., & Behl, C. (2004). Inhibition of glycogen synthase kinase 3 beta is involved in the resistance to oxidative stress in neuronal HT22 cells. Brain Res, 1005(1-2), 84-89. doi: S0006899304001544 [pii] 10.1016/j.brainres.2004.01.037

Schütt, F., Aretz, S., Auffarth, G. U., & Kopitz, J. (2012). Moderately reduced ATP levels promote oxidative stress and debilitate autophagic and phagocytic capacities in human RPE cells. Invest Ophthalmol Vis Sci, 53(9), 5354-5361. doi: 10.1167/iovs.12-9845

Scola, G., Kim, H. K., Young, L. T., & Andreazza, A. C. (2012). A Fresh Look at Complex I in Microarray Data: Clues to Understanding Disease-Specific Mitochondrial Alterations in Bipolar Disorder. Biol Psychiatry. doi: S0006-3223(12)00583-5 [pii]

10.1016/j.biopsych.2012.06.028

Scola, G., Kim, H. K., Young, L. T., & Andreazza, A. C. (2013). A fresh look at complex I in microarray data: clues to understanding disease-specific mitochondrial alterations in bipolar disorder. Biol Psychiatry, 73(2), e4-5. doi: 10.1016/j.biopsych.2012.06.028

Seamans, J. K., & Yang, C. R. (2004). The principal features and mechanisms of dopamine modulation in the prefrontal cortex. Prog Neurobiol, 74(1), 1-58. doi: 10.1016/j.pneurobio.2004.05.006

Selemon, L. D., Rajkowska, G., & Goldman-Rakic, P. S. (1998). Elevated neuronal density in prefrontal area 46 in brains from schizophrenic patients: application of a three-dimensional, stereologic counting method. J Comp Neurol, 392(3), 402-412. doi: 10.1002/(SICI)1096-9861(19980316)392:3<402::AID-CNE9>3.0.CO;2-5 [pii]

Selemon, L. D., Rajkowska, G., & Goldman-Rakic, P. S. (2004). Evidence for progression in frontal cortical pathology in late-stage Huntington's disease. J Comp Neurol, 468(2), 190-204. doi: 10.1002/cne.10938

Semendeferi, K., Armstrong, E., Schleicher, A., Zilles, K., & Van Hoesen, G. W. (2001). Prefrontal cortex in humans and apes: a comparative study of area 10. Am J Phys Anthropol, 114(3), 224-241. doi: 10.1002/1096-8644(200103)114:3<224::AID-AJPA1022>3.0.CO;2-I

Serretti, A., Cusin, C., Lattuada, E., Lilli, R., Lorenzi, C., Di Bella, D., . . . Smeraldi, E. (1999). No interaction between serotonin transporter gene and dopamine receptor D4 gene in symptomatology of major psychoses. Am J Med Genet, 88(5), 481-485.

Shalbuyeva, N., Brustovetsky, T., & Brustovetsky, N. (2007). Lithium desensitizes brain mitochondria to calcium, antagonizes permeability transition, and diminishes cytochrome C release. J Biol Chem, 282(25), 18057-18068. doi: 10.1074/jbc.M702134200

Shao, L., Cui, J., Young, L. T., & Wang, J. F. (2008). The effect of mood stabilizer lithium on expression and activity of glutathione s-transferase isoenzymes. Neuroscience, 151(2), 518-524.

159

Sharma, N., Hewett, J., Ozelius, L. J., Ramesh, V., McLean, P. J., Breakefield, X. O., & Hyman, B. T. (2001). A close association of torsinA and alpha-synuclein in Lewy bodies: a fluorescence resonance energy transfer study. Am J Pathol, 159(1), 339-344. doi: S0002-9440(10)61700-2 [pii]

Sherer, T. B., Betarbet, R., Testa, C. M., Seo, B. B., Richardson, J. R., Kim, J. H., . . . Greenamyre, J. T. (2003). Mechanism of toxicity in rotenone models of Parkinson's disease. J Neurosci, 23(34), 10756-10764.

Shimada, K., Crother, T. R., Karlin, J., Dagvadorj, J., Chiba, N., Chen, S., . . . Arditi, M. (2012). Oxidized mitochondrial DNA activates the NLRP3 inflammasome during apoptosis. Immunity, 36(3), 401-414. doi: 10.1016/j.immuni.2012.01.009

Shungu, D. C. (2012). N-acetylcysteine for the treatment of glutathione deficiency and oxidative stress in schizophrenia. Biol Psychiatry, 71(11), 937-938. doi: 10.1016/j.biopsych.2012.03.025

Sies, H. (1991). Oxidative stress: from basic research to clinical application. Am J Med, 91(3C), 31S-38S.

Silverstone, T. (1984). Response to bromocriptine distinguishes bipolar from unipolar depression. Lancet, 1(8382), 903-904.

Soeiro-de-Souza, M. G., Andreazza, A. C., Carvalho, A. F., Machado-Vieira, R., Young, L. T., & Moreno, R. A. (2013). Number of manic episodes is associated with elevated DNA oxidation in bipolar I disorder. Int J Neuropsychopharmacol, 1-8. doi: 10.1017/S1461145713000047

Sollberger, G., Strittmatter, G. E., Garstkiewicz, M., Sand, J., & Beer, H. D. (2014). Caspase-1: the inflammasome and beyond. Innate Immun, 20(2), 115-125. doi: 10.1177/1753425913484374

Song, C. X., Szulwach, K. E., Fu, Y., Dai, Q., Yi, C., Li, X., . . . He, C. (2011). Selective chemical labeling reveals the genome-wide distribution of 5-hydroxymethylcytosine. Nat Biotechnol, 29(1), 68-72. doi: 10.1038/nbt.1732

Sorbara, M. T., & Girardin, S. E. (2011). Mitochondrial ROS fuel the inflammasome. Cell Res, 21(4), 558-560. doi: 10.1038/cr.2011.20

Srinivasan, V., Smits, M., Spence, W., Lowe, A. D., Kayumov, L., Pandi-Perumal, S. R., . . . Cardinali, D. P. (2006). Melatonin in mood disorders. World J Biol Psychiatry, 7(3), 138-151. doi: 10.1080/15622970600571822

Srinivasan, V., Spence, D. W., Pandi-Perumal, S. R., Brown, G. M., & Cardinali, D. P. (2011). Melatonin in mitochondrial dysfunction and related disorders. Int J Alzheimers Dis, 2011, 326320. doi: 10.4061/2011/326320

Suhara, T., Nakayama, K., Inoue, O., Fukuda, H., Shimizu, M., Mori, A., & Tateno, Y. (1992). D1 dopamine receptor binding in mood disorders measured by positron emission tomography. Psychopharmacology (Berl), 106(1), 14-18.

Sun, X., Wang, J. F., Tseng, M., & Young, L. T. (2006). Downregulation in components of the mitochondrial electron transport chain in the postmortem frontal cortex of subjects with bipolar disorder. J Psychiatry Neurosci, 31(3), 189-196.

Söderlund, J., Olsson, S. K., Samuelsson, M., Walther-Jallow, L., Johansson, C., Erhardt, S., . . . Engberg, G. (2011). Elevation of cerebrospinal fluid interleukin-1ß in bipolar disorder. J Psychiatry Neurosci, 36(2), 114-118. doi: 10.1503/jpn.100080

Tahiliani, M., Koh, K., Shen, Y., Pastor, W., Bandukwala, H., Brudno, Y., . . . Rao, A. (2009). Conversion of 5-Methylcytosine to 5-Hydroxymethylcytosine in Mammalian DNA by MLL Partner TET1. Science, 324(5929), 930-935. doi: 10.1126/science.1170116

Takeda, K., Akagi, S., Takahashi, S., Onishi, A., Hanada, H., & Pinkert, C. A. (2002). Mitochondrial activity in response to serum starvation in bovine (Bos taurus) cell culture. Cloning Stem Cells, 4(3), 223-229. doi: 10.1089/15362300260339502

160

Takuma, K., Baba, A., & Matsuda, T. (2004). Astrocyte apoptosis: implications for neuroprotection. Prog Neurobiol, 72(2), 111-127.

Tan, H., Young, L., Shao, L., Che, Y., Honer, W., & Wang, J. (2011). Mood stabilizer lithium inhibits amphetamine-increased 4-hydroxynonenal-protein adducts in rat frontal cortex. The Int J of Neuropsychopharmacol, 1-11.

Thompson, J. M., Gray, J. M., Hughes, J. H., Watson, S., Young, A. H., & Ferrier, I. N. (2007). Impaired working memory monitoring in euthymic bipolar patients. Bipolar Disord, 9(5), 478-489. doi: 10.1111/j.1399-5618.2007.00470.x

Torrey, E. F., Webster, M., Knable, M., Johnston, N., & Yolken, R. H. (2000). The stanley foundation brain collection and neuropathology consortium. Schizophr Res, 44(2), 151-155.

Tschopp, J., & Schroder, K. (2010). NLRP3 inflammasome activation: The convergence of multiple signalling pathways on ROS production? Nat Rev Immunol, 10(3), 210-215. doi: 10.1038/nri2725

Tse, D. C., McCreery, R. L., & Adams, R. N. (1976). Potential oxidative pathways of brain catecholamines. J Med Chem, 19(1), 37-40.

Tunbridge, E., Burnet, P. W., Sodhi, M. S., & Harrison, P. J. (2004). Catechol-o-methyltransferase (COMT) and proline dehydrogenase (PRODH) mRNAs in the dorsolateral prefrontal cortex in schizophrenia, bipolar disorder, and major depression. Synapse, 51(2), 112-118. doi: 10.1002/syn.10286

Turrens, J. F., & Boveris, A. (1980). Generation of superoxide anion by the NADH dehydrogenase of bovine heart mitochondria. Biochem J, 191(2), 421-427.

Valvassori, S. S., Rezin, G. T., Ferreira, C. L., Moretti, M., Goncalves, C. L., Cardoso, M. R., . . . Quevedo, J. (2010). Effects of mood stabilizers on mitochondrial respiratory chain activity in brain of rats treated with d-amphetamine. J Psychiatr Res, 44(14), 903-909.

van der Valk, J., Brunner, D., De Smet, K., Fex Svenningsen, A., Honegger, P., Knudsen, L. E., . . . Gstraunthaler, G. (2010). Optimization of chemically defined cell culture media--replacing fetal bovine serum in mammalian in vitro methods. Toxicol In Vitro, 24(4), 1053-1063. doi: 10.1016/j.tiv.2010.03.016

Van Kammen, D. P., & Murphy, D. L. (1975). Attenuation of the euphoriant and activating effects of d- and l-amphetamine by lithium carbonate treatment. Psychopharmacologia, 44(3), 215-224.

Vawter, M. P., Tomita, H., Meng, F., Bolstad, B., Li, J., Evans, S., . . . Bunney, W. E. (2006). Mitochondrial-related gene expression changes are sensitive to agonal-pH state: implications for brain disorders. Mol Psychiatry, 11(7), 615, 663-679.

Versace, A., Andreazza, A. C., Young, L. T., Fournier, J. C., Almeida, J. R., Stiffler, R. S., . . . Phillips, M. L. (2013). Elevated serum measures of lipid peroxidation and abnormal prefrontal white matter in euthymic bipolar adults: toward peripheral biomarkers of bipolar disorder. Mol Psychiatry. doi: 10.1038/mp.2012.188

Waldvogel, H. J., Curtis, M. A., Baer, K., Rees, M. I., & Faull, R. L. (2006). Immunohistochemical staining of post-mortem adult human brain sections. Nat Protoc, 1(6), 2719-2732. doi: nprot.2006.354 [pii]

10.1038/nprot.2006.354

Wang, J. F., Shao, L., Sun, X., & Young, L. T. (2009). Increased oxidative stress in the anterior cingulate cortex of subjects with bipolar disorder and schizophrenia. Bipolar Disord, 11(5), 523-529.

Washizuka, S., Iwamoto, K., Kakiuchi, C., Bundo, M., & Kato, T. (2009). Expression of mitochondrial complex I subunit gene NDUFV2 in the lymphoblastoid cells derived from patients with bipolar disorder and schizophrenia. Neurosci Res, 63(3), 199-204.

161

Washizuka, S., Kakiuchi, C., Mori, K., Kunugi, H., Tajima, O., Akiyama, T., . . . Kato, T. (2003). Association of mitochondrial complex I subunit gene NDUFV2 at 18p11 with bipolar disorder. Am J Med Genet B Neuropsychiatr Genet, 120B(1), 72-78. doi: 10.1002/ajmg.b.20041

Watabe, M., & Nakaki, T. (2004). Rotenone induces apoptosis via activation of bad in human dopaminergic SH-SY5Y cells. J Pharmacol Exp Ther, 311(3), 948-953. doi: jpet.104.071381 [pii]

10.1124/jpet.104.071381

Watabe, M., & Nakaki, T. (2007). Mitochondrial complex I inhibitor rotenone-elicited dopamine redistribution from vesicles to cytosol in human dopaminergic SH-SY5Y cells. J Pharmacol Exp Ther, 323(2), 499-507. doi: jpet.107.127597 [pii]

10.1124/jpet.107.127597

Watabe, M., & Nakaki, T. (2008). Mitochondrial complex I inhibitor rotenone inhibits and redistributes vesicular monoamine transporter 2 via nitration in human dopaminergic SH-SY5Y cells. Mol Pharmacol, 74(4), 933-940. doi: mol.108.048546 [pii]

10.1124/mol.108.048546

Weber, A., Wasiliew, P., & Kracht, M. (2010). Interleukin-1beta (IL-1beta) processing pathway. Sci Signal, 3(105), cm2. doi: 10.1126/scisignal.3105cm2

Weber, B., Serafin, A., Michie, J., Van Rensburg, C., Swarts, J. C., & Bohm, L. (2004). Cytotoxicity and cell death pathways invoked by two new rhodium-ferrocene complexes in benign and malignant prostatic cell lines. Anticancer Res, 24(2B), 763-770.

Weerasinghe, G. R., Rapoport, S. I., & Bosetti, F. (2004). The effect of chronic lithium on arachidonic acid release and metabolism in rat brain does not involve secretory phospholipase A2 or lipoxygenase/cytochrome P450 pathways. Brain Res Bull, 63(6), 485-489. doi: 10.1016/j.brainresbull.2004.04.005

Wegener, G., Bandpey, Z., Heiberg, I. L., Volke, V., Trabace, L., Rosenberg, R., & Harvey, B. H. (2004). Combined chronic treatment with citalopram and lithium does not modify the regional neurochemistry of nitric oxide in rat brain. J Physiol Pharmacol, 55(3), 575-586.

Weiss, R. B., Stange, J. P., Boland, E. M., Black, S. K., LaBelle, D. R., Abramson, L. Y., & Alloy, L. B. (2015). Kindling of life stress in bipolar disorder: comparison of sensitization and autonomy models. J Abnorm Psychol, 124(1), 4-16. doi: 10.1037/abn0000014

West, M. J. (2012). Introduction to stereology. Cold Spring Harb Protoc, 2012(8). doi: 10.1101/pdb.top070623

Wieckowski, M. R., Giorgi, C., Lebiedzinska, M., Duszynski, J., & Pinton, P. (2009). Isolation of mitochondria-associated membranes and mitochondria from animal tissues and cells. Nat Protoc, 4(11), 1582-1590. doi: 10.1038/nprot.2009.151

Wills, J., Jones, J., Haggerty, T., Duka, V., Joyce, J. N., & Sidhu, A. (2010). Elevated tauopathy and alpha-synuclein pathology in postmortem Parkinson's disease brains with and without dementia. Exp Neurol, 225(1), 210-218. doi: 10.1016/j.expneurol.2010.06.017

Winther, J. R., & Thorpe, C. (2013). Quantification of thiols and disulfides. Biochim Biophys Acta. doi: 10.1016/j.bbagen.2013.03.031

Wong, D. F., Pearlson, G. D., Tune, L. E., Young, L. T., Meltzer, C. C., Dannals, R. F., . . . Gjedde, A. (1997). Quantification of neuroreceptors in the living human brain: IV. Effect of aging and elevations of D2-like receptors in schizophrenia and bipolar illness. J Cereb Blood Flow Metab, 17(3), 331-342. doi: 10.1097/00004647-199703000-00010

162

Wossidlo, M., Nakamura, T., Lepikhov, K., Marques, C. J., Zakhartchenko, V., Boiani, M., . . . Walter, J. (2011). 5-Hydroxymethylcytosine in the mammalian zygote is linked with epigenetic reprogramming. Nat Commun, 2, 241. doi: 10.1038/ncomms1240

Wu, H., Coskun, V., Tao, J., Xie, W., Ge, W., Yoshikawa, K., . . . Sun, Y. (2010). Dnmt3a-Dependent Nonpromoter DNA Methylation Facilitates Transcription of Neurogenic Genes. Science, 329(5990), 444-448. doi: 10.1126/science.1190485

Yao, J. K., & Keshavan, M. S. (2011). Antioxidants, redox signaling, and pathophysiology in schizophrenia: an integrative view. Antioxid Redox Signal, 15(7), 2011-2035.

Zhang, J., Echeverry, S., Lim, T. K., Lee, S. H., Shi, X. Q., & Huang, H. (2015). Can Modulating Inflammatory Response be a Good Strategy to Treat Neuropathic Pain? Curr Pharm Des, 21(7), 831-839.

Zhang, P., Wang, Y. Z., Kagan, E., & Bonner, J. C. (2000). Peroxynitrite targets the epidermal growth factor receptor, Raf-1, and MEK independently to activate MAPK. J Biol Chem, 275(29), 22479-22486. doi: 10.1074/jbc.M910425199

Zhang, R., Kang, K. A., Kim, K. C., Na, S. Y., Chang, W. Y., Kim, G. Y., . . . Hyun, J. W. (2013). Oxidative stress causes epigenetic alteration of CDX1 expression in colorectal cancer cells. Gene, 524(2), 214-219. doi: 10.1016/j.gene.2013.04.024

Zheng, L. S., Ishii, Y., Zhao, Q. L., Kondo, T., & Sasahara, M. (2013). PDGF suppresses oxidative stress induced Ca2+ overload and calpain activation in neurons. Oxid Med Cell Longev, 2013, 367206. doi: 10.1155/2013/367206

Zheng, X., Baker, H., Hancock, W. S., Fawaz, F., McCaman, M., & Pungor, E. (2006). Proteomic analysis for the assessment of different lots of fetal bovine serum as a raw material for cell culture. Part IV. Application of proteomics to the manufacture of biological drugs. Biotechnol Prog, 22(5), 1294-1300. doi: 10.1021/bp060121o

Zhong, Z., Zhai, Y., Liang, S., Mori, Y., Han, R., Sutterwala, F. S., & Qiao, L. (2013). TRPM2 links oxidative stress to NLRP3 inflammasome activation. Nat Commun, 4, 1611. doi: 10.1038/ncomms2608

Zhou, R., Yazdi, A. S., Menu, P., & Tschopp, J. (2011). A role for mitochondria in NLRP3 inflammasome activation. Nature, 469(7329), 221-225. doi: 10.1038/nature09663

163

Copyright releases

164

165

January 15, 2015

Helena K. Kim

Ph.D. Candidate

Department of Pharmacology and Toxicology

University of Toronto

SUBJECT: J Psychiatry Neurosci 2014; 39(4):276-85 DOI: 10.1503/jpn.130155

Oxidation and nitration in dopaminergic areas of the prefrontal cortex from patients with bipolar disorder and schizophrenia

Dear Helena,

I am pleased to confirm that the Canadian Medical Association as the copyright holder, grants you permission to reproduce this material as requested in your thesis. We do ask that you use the final, published text as opposed to the originally submitted manuscript.

The Royalty fee has been waived provided that this material will be used for research/education purposes only and that it will not be resold or used for commercial purposes in any way.

Please credit the original publication of the material as follows:

Oxidation and nitration in dopaminergic areas of the prefrontal cortex from patients with bipolar disorder and schizophrenia — Reprinted from, J Psychiatry Neurosci; 2014; 39(4):276-85 DOI: 10.1503/jpn.130155 by permission of the publisher.

© 2014 Canadian Medical Association

If you require further information or assistance, please do not hesitate to contact the undersigned.

Sincerely,

Janis Murrey Program Manager Journals

1867 Alta Vista Drive, Ottawa ON K1G 5W8

166

T: 613 731-8610 x2110

Permissions@cma.ca

167

168