Genetic and Genomic Analysis of Atp1a2 and Its Role in Alcohol

GENETIC AND GENOMIC ANALYSIS OF ATP1A2 AND ITS ROLE IN ALCOHOL

RELATED BEHAVIORS

STEPHANIE MARIE GRITZ

B.S., Virginia Polytechnic Institute and State University, 2009

A thesis submitted to the

Faculty of the Graduate School of the

University of Colorado in partial fulfillment

of the requirements for the degree of

Doctor of Philosophy

Toxicology Program

2015 This thesis for the Doctor of Philosophy degree by

Stephanie Marie Gritz

has been approved for the

Toxicology Program

Dennis R. Petersen, Chair

Richard A. Radcliffe, Advisor

Manisha Patel

Richard A. Deitrich

Jerry A. Stitzel

Date: __5/4/2015_

ii Gritz, Stephanie Marie (Ph.D., Toxicology)

Genetic and Genomic Analysis of Atp1a2 and its Role in Alcohol Related Behaviors

Thesis directed by Professor Richard Radcliffe.

ABSTRACT

Alcohol abuse is a significant public health issue and understanding genetic factors related to the effects of alcohol in the brain may provide new avenues of treatment. We examined the Na+/K+ exchanger Atpase1a2 (Atp1a2), a sodium/potassium ATPase responsible for establishing and maintaining the electrochemical gradients of sodium and potassium ions across plasma membranes.

This gradient is important in osmoregulation, sodium-coupled transport, and electrical excitability of nerves and muscles. Atp1a2 has been previously studied for anxiety, learning and motor function disorders, and fear. Since Atp1a2 has been shown to be involved in anxiety and this behavior is a known risk factor for developing alcoholism, we investigated Atp1a2 for its potential role in responses to alcohol. Using the Inbred Short

Sleep (ISS) and the Inbred Long Sleep (ILS) mouse strains and the Recombinant Inbred

(RI) strains derived from ISS and ILS (LXS), previous studies by Dr. Radcliffe's lab showed a truncated 3' untranslated region (UTR) located in Atp1a2 in the ILS compared to the ISS and 106 single nucleotide polymorphisms (SNPs) throughout the gene.

Atp1a2 protein expression was measured to determine if the truncated 3' UTR and multiple SNPs throughout the gene affected Atp1a2 expression. Atp1a2 heterozygous mice were tested for behavioral responses to alcohol utilizing loss of righting reflex

(LORR), acute alcohol withdrawal (HIC), drinking in the dark (DID), open field activity

(OFA), and elevated-plus maze (EPM). LORR is a two day test that measures acute

iii alcohol sensitivity, acute alcohol withdrawal is a three day test to measure sensitivity,

DID is a four day test which measures voluntary alcohol consumption, OFA and EPM measure anxiety. The effect of genotype on alcohol metabolism was also examined.

There was no difference in alcohol metabolism or alcohol withdrawal severity. The

Atp1a2 heterozygous consumed more alcohol than wild-type and although results were mixed due to sex differences, there was increased sensitivity, increased rapid tolerance, and increased acute tolerance to alcohol. Alcohol exposure also decreased the measures of anxiety seen in the heterozygous mice. This comprehensive study showed the role of Atp1a2 genetic variance in alcohol related behaviors.

The form and content of this abstract are approved. I recommend its publication.

Approved: Richard A. Radcliffe

iv ACKNOWLEDGMENTS

First, I would like to acknowledge and thank Dr. Richard Radcliffe for accepting me into his laboratory and for proving me with the support and guidance to pursue and develop my scientific skills and knowledge. I would also like to thank the members of my committee, Drs. Dennis Petersen, Manisha Patel, Richard Deitrich and Jerry Stitzel, for their input in regards to my thesis work and the development of my scientific and critical thinking. I would also like to thank Dr. Alison Bauer for all her advice and moral support she has given over the years.

I want to thank all of the past and present members of the Radcliffe laboratory as well as my fellow toxicology students and colleagues for their friendship and scientific support, in particular, Colin Larson, Dr. Molly Millett, Dr. Swetha Inturi, Dr. Sangeeta

Shrotriya, Pallavi Bhuyan, Cameron McElroy, Ross Osgood, Sam Freedman, and Dr.

Carla-Maria Alexander. Thank you for all the moral support, with both life in general and science; I would not have made it through without your encouragement.

Finally, I would like to acknowledge and thank my family, my parents Mark and

Candy and my brother Michael for their unwavering love and support throughout my life and this exciting journey. Mom and Dad, I truly would not have made it this far if it wasn't for the values you instilled in me at an early age and all the sacrifices you made to get me to where I am today. I love you both more than words can describe and I am so thankful you always pushed me to go after what I wanted in life. Michael, or should I say Master Gritz, you will always be my baby brother and I'm so proud of the man you have become and am thankful for the support you have given me over the years, now comes the most exciting adventure of all: adulthood.

v TABLE OF CONTENTS

CHAPTER

I. INTRODUCTION AND BACKGROUND ...... 1

Sodium/Potassium ATPase 1a2 ...... 1

The Mechanism of Atp1a2 ...... 1

The Role of ATP1A2 in FHM2 ...... 5

Atp1a2 Studies in Animal Models ...... 9

Anxiety Disorder ...... 13

Characteristics of Anxiety Disorders ...... 13

Relationship between Anxiety and Atp1a2 ...... 16

Alcoholism ...... 17

Characteristics of Alcoholism ...... 17

Alcoholism in Animals Models ...... 18

Overview of Thesis ...... 21

II. BIOINFORMATIC ANALYSIS AND DIFFERENTIAL EXPRESSION OF ATP1A2 ..23

Introduction ...... 23

Methods ...... 26

Animals ...... 26

SNP Analysis and LXS RI Genotyping ...... 27

Genetic Correlations ...... 27

Western Blotting and Quantification ...... 29

RNA Isolation and qPCR ...... 30

Statistical Methods ...... 30

Results ...... 31 vi

SNP Distriubtion and LXS RI Genotypes ...... 31

Genotype-Phenotype Correlations in the LXS and BXDs ...... 32

Atp1a2 Expression in the LXS RIs ...... 33

Atp1a2 QTL Mapping and Gene Candidates from Chromosome 19 ...... 33

Atp1a2 Correlates with Chromosome 19 Candidate Genes and ATP Binding Proteins ...... 39

GLAST and GLT-1 Gene Expression in the Atp1a2 Heterozygous Mice ...... 39

Discussion ...... 40

III. THE ROLE OF ATP1A2 IN ALCOHOL RELATED BEHAVIORS ...... 48

Introduction ...... 48

Methods ...... 49

Animals ...... 49

Alcohol Metabolism ...... 50

Loss of Righting Reflex ...... 50

Handling Induced Convulsions ...... 51

Drinking in the Dark ...... 52

Open-Field Activity ...... 53

Elevated Plus Maze ...... 53

Statistical Methods ...... 54

Results ...... 54

No Change in Alcohol Metabolism ...... 54

Increased Sensitivity, Rapid Tolerance and Acute Functional Tolerance in Atp1a2 Heterozygous Mice ...... 55

No Difference in Alcohol Withdrawal Measured by Handling Induced Convulsion ...... 57

Voluntary Consumption is Increased in Atp1a2 Heterozygous Mice ...... 57

vii Decrease in Anxiety Seen in Open-Field Activity after Alcohol ...... 58

Elevated Plus Maze Showed Decreased Anxiety After Alcohol Exposure ...... 59

Discussion ...... 59

IV. SUMMARY AND FUTURE DIRECTIONS ...... 68

Summary ...... 68

Future Directions ...... 75

REFERENCES ...... 78

viii LIST OF TABLES

TABLE

1.1. Homology of Atp1a2 ...... 4

1.2. Migraine Criteria ...... 5

1.3. Atp1a2 Known Mutations ...... 8

1.4. Testing in the Atp1a2 Heterozygous Mice ...... 11

2.1. Strain Haplotypes ...... 29

2.2. SNP Distribution in Atp1a2 ...... 32

2.3. LXS RI Genotypes ...... 34

2.4. Behaviors Correlated to Atp1a2 mRNA expression ...... 35

2.5. Candidate Chr. 19 Genes ...... 38

2.6. Atp1a2 Correlations ...... 42

3.1. HIC Score Scale...... 52

4.1. Overall Summary of Behavioral Testing in the Atp1a2 Heterozygous

Mice ...... 72

ix LIST OF FIGURES

FIGURE

1.1. The Na,K-ATPase reaction cycle ...... 3

1.2. Mutations in Atp1a2 in Animal Models ...... 10

1.3. Creation of RI strains ...... 20

2.1. Atp1a2 Microarray ...... 24

2.2. Atp1a2 NextGen RNA Seq Results ...... 26

2.3. Atp1a2 Protein Expression in LXS RIs ...... 36

2.4. Atp1a2 QTL ...... 37

2.5. GO Enrichment on Genes Correlated to Atp1a2 ...... 43

2.6. GLAST and GLT-1 Expression ...... 44

3.1. Alcohol Metabolism in the Atp1a2 Mice ...... 55

3.2. LORR and AFT ...... 56

3.3. Alcohol Withdrawal ...... 58

3.4. Drinking in the Dark ...... 61

3.5. Open Field Activity ...... 62

3.6. Elevated Plus Maze ...... 63

x LIST OF ABBREVIATIONS

Ablim1 Actin-Binding Lim Protein 1 Acsl5 Acyl-CoA Synthetase Long-Chain Family Member 5 Afap1l2 Actin Filament Associated Protein 1-Like 2 AFT Acute Functional Tolerance Atp1a2 Na+/K+ Exchanger Atpase1A2 AUD Alcohol Use Disorder BEC Blood Ethanol Concentrations BXD Recombinant Inbred Strains Derived From C57Bl6 And Dba2 Ca2+ Calcium Cacul1 Cdk2 Associated, Cullin Domain 1 CNS Central Nervous System CSD Cortical Spreading Depression DID Drinking In The Dark EE Ethanol Pre-Treatment EPM Elevated-Plus Maze FHM Familial Hemiplegic Migraine FHM2 Familial Hemiplegic Migraine Type II GABA γ-Amino Butyric Acid GAD Generalized Anxiety Disorder Gfra1 Glial Cell Line Derived Neurotrophic Factor Family Receptor Alpha 1 GLAST Glutamate Aspartate Transporter GLT-1 Solute Carrier Family 1 Member 2 GO Gene Ontology Grk5 G Protein-Coupled Receptor Kinase 5 HET Heterozygous HIC Handing-Induced Convulsions Hspa12a Heat Shock Protein 12A ILS Inbred Long Sleep Ip Intraperitoneally ISS Inbred Short Sleep K+ Potassium LORR Loss Of Righting Reflex LRS Likelihood Ratio Statistic LS Long Sleep LXS Recombinant strains Derived From ISS And ILS Na+ Sodium Nanos1 Nanos Homolog 1 (Drosophila) xi NCX Sodium/Calcium Exchanger Nhlrc2 NHL Repeat Containing 2 Nrap Nebulin-Related Anchoring Protein OCD Obsessive-Compulsive Disorder OFA Open Field Activity Pdzd8 Pdz Domain Containing 8 Pnlip Pancreatic Lipase Pnliprp2 Pancreatic Lipase-Related Protein 2 Prdx3 Peroxiredoxin 3 PTSD Post Traumatic Stress Disorder QTL Quantitative Trait Locus RI Recombinant Inbred SAD Social Anxiety Disorder SE Saline Pre-Treatment Sfxn4 Sideroflexin 4 Slc18a2 Solute Carrier Family 18 (Vesicular Monoamine), Member 2 SNP Single Nucleotide Polymorphisms SS Short Sleep SSRI Selective Serotonin Re-Uptake Inhibitors TGVS Trigeminovascular System UTR Untranslated Region Vesicle Transport Through Interaction With T-Snares Homolog 1A Vti1a (Yeast) WT Wild-type

xii

CHAPTER I

INTRODUCTION AND BACKGROUND1

Sodium-Potassium ATPase1a2

Na+, K+-ATPase alpha 2 (Atp1a2) is an integral plasma membrane protein belonging to the P-type ATPase family that is responsible for maintaining the sodium

(Na+) and potassium (K+) gradients across cellular membranes with hydrolysis of ATP.

Atp1a2 contains two subunits, alpha and beta, with each having various isoforms and differential tissue distribution. In humans, mutations in ATP1A2 are associated with a rare form of hereditary migraines with aura known as familial hemiplegic migraine type II

(FHM2). Genetic studies in mice have revealed other neurological effects of Atp1a2 in mice including anxiety, fear, and learning and motor function disorders.

The Mechanism of Atp1a2

The integral plasma membrane protein Na+, K+-ATPase is a member of the P- type ATPase family. P-type ATPases maintain the essential plasma membrane potential in all eukaryotic cells and are found in all cell type membranes. The plasma membrane potential is upheld by adjusting ion concentrations on the intracellular and extracellular sides of the membrane. This electrochemical gradient fuels central cellular processes,

1Portions of this chapter are a reprint from Human Genomics, 7(8), Gritz SM, Radcliffe, RA. Genetic Effects of ATP1A2 in Familial Hemiplegic Migraine Type II and Animal Models, 2013, with permission from BioMed Central. 1

such as the secondary transport of metabolites, and it also provides the basis for electrical excitation in neurons. The mechanism of action of Na+, K+-ATPase involves a conformation change driven by ATP hydrolysis (figure 1.1) (Kaplan 2002; Vanmolkot,

Kors, et al. 2006; Bassi et al. 2004). While bound to ATP, the pump binds three intracellular sodium ions (fig. 1.1, step 1). ATP is then hydrolyzed to ADP and the pump is phosphorylated at a highly conserved aspartate residue, the target for β-aspartyl phosphorylation (fig. 1.1, step 2) (Bassi et al. 2004). This phosphorylation causes a conformation change, resulting in the release of the sodium ions into the extracellular space and the binding of two potassium ions (fig. 1.1, step 3). This binding of potassium ions causes the pump to dephosphorylate (fig. 1.1, step 4), which returns the pump to its previous conformational state (fig. 1.1, step 5), and transports the potassium ions into the cell (fig. 1.1, step 6). The process then repeats itself when ATP again binds to the pump.

The Na+, K+-ATPases consist of an α subunit and a β subunit, and only in the kidney, a γ subunit (Kaplan 2002; Therien, Karlish, and Blostein 1999). The α and β subunits are synthesized separately, assembled in the endoplasmic reticulum, and are delivered to the plasma membrane (Kaplan 2002; Geering et al. 1996). The α subunit is the catalytic region of the enzyme and contains the binding sites for the sodium and potassium ions, ATP, and cardiac glycosides such as ouabain, while the glycoslyated β subunit is needed for proper folding and function of the catalytic subunit. The x-ray crystal structure of Na+, K+-ATPase was examined from pig kidney, which showed the binding sites in the α subunit and the interactions between the α and β subunits (Morth et al. 2007). Different genes encode each of the multiple α and β isoforms and the γ isoform; four α isoforms (ATP1A1, ATP1A2, ATP1A3, and ATP1A4), four β isoforms

(ATP1B1, ATP1B2, ATP1B3 and ATP1B4), and one γ isoform (FXYD2) have been

identified (table 1.1) (National Library of Medicine 2013a; National Library of Medicine

2013b).

Figure 1.1. The Na,K-ATPase reaction cycle. The sequence of steps involved in the active transport of Na+ and K+ ions with the utilization of ATP. (Kaplan 2002; Vanmolkot, Kors, et al. 2006; Bassi et al. 2004)

The α isoforms differ in tissue distribution and are regulated developmentally

(Orlowski and Lingrel 1988). The α1 subunit is expressed ubiquitously, the α2 isoform is expressed in the brain, heart, and skeletal muscle, the α3 subunit is expressed in the brain and heart, and the α4 isoform is expressed in sperm and its precursor cells (Woo,

James, and Lingrel 1999; Lingrel et al. 2007). The four isoforms have a high degree of amino acid identity, but have differences in kinetic properties and substrate affinity

(Jewell and Lingrel 1991; Segall, Daly, and Blostein 2001). Sequence identity between human and mouse protein is between 83 and 99% and within subunits and species is approximately 80-90% (table 1.1) (Altschul et al. 1997). In the adult brain, the α1 3

isoform is found in multiple central nervous system (CNS) cell types, the α2 isoform is primarily expressed in astrocytes and pyramidal cells in the hippocampus, and α3 is expressed only in neurons (McGrail, Phillips, and Sweadner 1991; Moseley et al. 2003;

Moseley et al. 2007).

Table 1.1. Homology of Atp1a2. Na+, K+ ATPases in the P-type Atpase Family in Human and Mouse. [http://www.ncbi.nlm.nih.gov/gene and http://www.ncbi.nlm.nih.gov/protein]

Location Size Location Size Sequence Type Human Mouse (Chr:bp) (#AA) (Chr:bp) (#AA) Identity α1 ATP1A1 1:116915795 1023 Atp1a1 3:101576219 1023 97% α2 ATP1A2 1:160085520 1020 Atp1a2 1:172271709 1020 99% α3 ATP1A3 19:42470734 1013 Atp1a3 7:24978167 1013 99% α4 ATP1A4 1:160121352 1029 Atp1a4 1:172223508 1032 83% β1 ATP1B1 1:169075947 303 Atp1b1 1:164437267 304 94% β2 ATP1B2 17:7554254 290 Atp1b2 11:69599750 290 97% β3 ATP1B3 3:141595470 279 Atp1b3 9:96332673 278 73% β4 ATP1B4 X:119495940 357 Atp1b4 X:38316267 356 89% γ FXYD2 11:117690790 64 Fxyd2 9:45399709 64 81%

ATP1A2, as well as the other isoforms, is responsible for maintaining the resting membrane potential and for driving nutrient and neurotransmitter uptake. Na+, K+

ATPases are important in clearing extracellular potassium during neuronal activity and are essential in the clearance of released glutamate in the synaptic cleft because re- uptake in astrocytes and neurons is driven by the sodium and potassium gradients

(Ikeda et al. 2003). ATP1A2 is co-localized with other ion transporters, such as the sodium/calcium (Na+/Ca2+) exchanger and the glutamate transporter, which in the CNS are important in clearance of glutamate and potassium from the extracellular space

(James et al. 1999; Rose et al. 2009; Cholet et al. 2002).

The Role of ATP1A2 in FHM2

Familial hemiplegic migraine (FHM) is a rare autosomal dominant form of migraine with aura. FHM attacks are generally longer than the common migraine with aura; however, they share similar symptoms (table 1.2) including visual, sensory, motor, and aphasia (Pietrobon 2007). Aura symptoms are different for each person, but is generally described as disturbances in light perception, such as spots or lines in vision, and changes in sensory perception, such as heightened sensitivity to smells and sounds

(de Vries et al. 2009). There are three types of FHM: type 1 is associated with mutations in the neuronal calcium channel gene and is the most prevalent form of FHM, type 2 is caused by mutations present in ATP1A2, and type 3 is a rare form of FHM related to mutations in the sodium channel gene (de Vries et al. 2009).

Table 1.2. Migraine Criteria. International Headache Society migraine criterion (Olsen 2004)

Repeated episodic headaches within 4-72 hours with the following features:

Any two of: Any one of:  Unilateral  Nausea and/or vomiting  Throbbing  Photophobia and  Worsened by movement phonophobia  Moderate or severe

FHM2 patients suffer from migraines with hemiplegia and partial paralysis during the aura phase and in some cases accompanied by seizures or cognitive dysfunction.

Neuroimaging studies have shown that migraine aura is caused by cortical spreading depression (CSD). CSD is a wave of continual strong neuronal depolarization that slowly progresses across the cortex generating a brief intense spike of activity that is followed by long-lasting neural suppression (Pietrobon and Striessnig 2003). CSD has

been shown to activate the trigeminovascular system (TGVS), which is responsible for the headache associated with migraines (de Vries et al. 2009). Inhibition of ATP1A2 leads to high levels of extracellular potassium causing neurons to become depolarized which can cause CSD (Leo et al. 2011).

ATP1A2 was identified as a gene associated with FHM2 in 2003 in two Italian families (De Fusco et al. 2003). The mutations and deficiencies in ATP1A2 that cause

FHM2 are responsible for approximately 20% of FHM in families (Vanmolkot, Kors, et al.

2006). There are over 50 mutations in ATP1A2 that have been identified in association with FHM2. Almost all FHM2 mutations are non-synonymous SNPs, but there are also small deletions (Riant et al. 2005) and a mutation affecting the stop codon causing an extension of the ATP1A2 protein by 27 amino acid residues (Jurkat-Rott et al. 2004).

Most of the mutations are associated with pure FHM without additional clinical symptoms

(De Fusco et al. 2003; Riant et al. 2005; Jurkat-Rott et al. 2004; Kaunisto et al. 2004;

Pierelli et al. 2006). Recently, a number of ATP1A2 mutations were reported to be associated with FHM and cerebellar problems, specifically motor problems (Spadaro et al. 2004), childhood convulsions (Vanmolkot et al. 2003), epilepsy (Jurkat-Rott et al.

2004; Deprez et al. 2008), and mental retardation (Jurkat-Rott et al. 2004; Vanmolkot,

Stroink, et al. 2006). Some ATP1A2 mutations have been shown to be associated with non-hemiplegic migraine phenotypes, such as basilar migraine (Ambrosini et al. 2005) and the common migraine (Todt et al. 2005).

One of the most important experimental drugs to study ATP1A2, as well as

ATP1A1 and ATP1A3, studies is ouabain which is a plant derived steroid that binds to

+ + the E2P form of Na , K -ATPases, as seen in figure 1. Ouabain acts as a non-selective antagonist, and inhibits the enzyme transport activity (Crambert et al. 2004). The three isoforms in humans show similar affinity for ouabain, however ouabain binding is altered

due to changes in extracellular potassium levels, resulting in varying sensitivity

(Crambert et al. 2000). In rats, Atp1a3 is very sensitive to ouabain; Atp1a2 is less sensitive and Atp1a1 is insensitive (Munzer et al. 1994).

Cell based studies have shown that multiple individual mutations in ATP1A2 result in dysfunctional ion pump activity. The functional effects of mutations in ATP1A2 have been studied in HeLa cells and Xenopus oocytes (table 1.3). Recombinant

ATP1A2 subunits containing mutations that cause the pump to be insensitive to ouabain were expressed in HeLa cells; the ouabain insensitive mutants were used to distinguish between endogenous ATP1A2 activity and the effects of other mutations engineered into the recombinant ATP1A2 (De Fusco et al. 2003; Vanmolkot et al. 2003; Vanmolkot,

Stroink, et al. 2006; Segall et al. 2004; Segall et al. 2005). Three mutations in HeLa cells and two mutations in Xenopus oocytes produced severe or complete loss-of- function of pump activity, which leads to cell death (Bassi et al. 2004; De Fusco et al.

2003; Vanmolkot, Kors, et al. 2006; Segall et al. 2004). Five other FHM2 mutants were analyzed and pump activity was reduced, but sufficient to allow survival of the HeLa cells

(Segall et al. 2004; Segall et al. 2005; Koenderink et al. 2005).

ATP1A2 is expressed primarily in astrocytes in the adult, where it appears functionally coupled to various transporters (glutamate transporter and Na+/Ca2+ exchanger), and is essential in the clearance of released glutamate and potassium from the extracellular space during neuronal activity. FHM2 mutations that cause loss-of- function of ATP1A2 may lead to decreased glutamate clearance and an increase of potassium in the synaptic cleft during neuronal activity which could lead to prolonged recovery time after neuronal excitation and may render the brain more susceptible to

CSD (Pietrobon 2007; Leo et al. 2011).

Table 1.3. Atp1a2 Known Mutations. Cell based studies demonstrating the effect of Atp1a2 mutations in Xenopus oocytes and HeLa cells. (Vanmolkot, Kors, et al. 2006; Bassi et al. 2004; De Fusco et al. 2003; Segall et al. 2004; Segall et al. 2005; Koenderink et al. 2005)

Cell Type Mutation Affect on Pump Function HeLa L764P Complete Loss W887R Complete Loss G615R Complete Loss R593W Reduced Rate V628M Reduced Rate T345A Similar to Wild-type: Lower affinity for potassium Similar to Wild-type: decreased catalytic turnover M731T and increased affinity for potassium Similar to Wild-type: decreased catalytic turnover R689Q and increased affinity for potassium Xenopus L764P Complete Loss oocytes W887R Complete Loss

There are two hypothesized mechanisms for the effects of ATP1A2 mutations.

The first hypothesis is that the mutations cause an increase in extracellular potassium, which can result in impaired clearance of potassium ions and therefore induce CSD

(Koenderink et al. 2005). The second hypothesis is since the distribution of ATP1A2 is co-localized with the Na+/Ca2+ exchanger, the mutations to ATP1A2 would cause intracellular sodium to increase, which increases intracellular calcium levels through the

Na+/Ca2+ exchanger, similar to FHM1, resulting in glutamate release and a decrease in glutamate clearance which can also lead to CSD (Pietrobon and Striessnig 2003;

Montagna 2004). Both hypotheses result in making the brain more susceptible to CSD and therefore migraines with aura.

Currently, the medications used to treat FHM2 are standard migraine prophylactic drugs, such as antidepressants, beta blockers, and calcium channel blockers (Jen 2001), which treat the symptoms, not the cause. Further investigation of

ATP1A2 in humans and animal models are needed to better determine treatment options.

Atp1a2 Studies in Animal Models

Over the last twenty years animal studies using either Atp1a2 knock-outs or knock-in mutations have increased our understanding of its effect on behavior. Atp1a2 heterozygous mice have been studied by two separate groups, one at the University of

Cincinnati lead by Dr. Jerry Lingrel and the second at Jichi Medical School in Japan overseen by Dr. Kiyoshi Kawakami. Both groups have shown that modulation of Atp1a2 activity affects neural activity and whole animal behavior. More recently, a group led by

Giorgio Casari at the Vita-Salute San Raffaele University and Center for Translational

Genomics and Bioinformatics in Italy generated the first FHM2 knock-in mouse. The constructs for the gene alterations are shown in figure 1.2.

Dr. Lingrel's lab has been investigating the functional roles of the various Na+, K+-

ATPase isoforms for over twenty years. They developed heterozygous mice for Atp1a2, as well as Atp1a1 and Atp1a3 (James et al. 1999). Atp1a2 heterozygous mice were tested for differences in behavior as well as to investigate the role of Atp1a2 in the heart.

Kawakami's group also created an Atp1a2 heterozygous mouse by deleting a portion of the gene resulting in only one functioning copy (Ikeda et al. 2003). They examined the role of Atp1a2 in neural activity by measuring fear and anxiety in heterozygous mice.

The Atp1a2 heterozygous mice also showed hyperphagia during the light period and suffered from late onset obesity (Kawakami et al. 2005). The homozygous null Atp1a2 mice are neonatal lethal due to lack of synchronized neuronal firing in the breathing center of the brain (Lingrel et al. 2003; Ikeda et al. 2004; Onimaru and Homma 2007)

and the Atp1a2 heterozygous mice have approximately half the protein compared to wild-type (James et al. 1999).

Figure 1.2. Mutations in Atp1a2 in Animal Models. Three separate groups have mutated the Atp1a2 gene in mice. Two groups created knock-outs through disruption of the gene at two different locations and other has created a knock-in mutation to mimic one of the FHM2 mutations in exon 19. Atp1a2 is located in the anti-sense strand of chromosome 1. (Ikeda et al. 2003; James et al. 1999; Leo et al. 2011)

The heterozygous mice from both groups consistently showed an anxiety-related phenotype (table 1.4). The Atp1a2 heterozygous mice were hypo-active compared to wild-type based on measurements of total distance traveled and time spent in the corners of the open-field testing box (Moseley et al. 2007). They also spent less time in and had fewer entries into the open arm of the plus maze (Lingrel et al. 2007; Moseley et al. 2007; Segall et al. 2004). In the light/dark test, the knock-outs spent less time in the light compartment and had fewer transitions between the two compartments (Ikeda et al.

2003). Taken together, these results show that the Atp1a2 heterozygous mice were more anxious than the wild-type.

Table 1.4. Testing in the Atp1a2 Heterozygous Mice. The various tests, parameter measured, and results compared to wild-type. (Moseley et al 2007, Ikeda et al 2003) ↑ = increased, ↓ = decreased

Heterozygous Phenotype Parameters Test compared to Lab Measured Measured Wild-type Hidden Learning and Latency time Lingrel Platform memory Elevated Time spent in Anxiety Lingrel Zero Maze open arms Entries into open quadrant Open-field Time spent in Anxiety Lingrel Activity corners Total Lingrel distance and traveled Kawakami Path length Light/Dark Anxiety in light Kawakami compartment Time spent in light compartment Latency to first entry into

light compartment Number of transitions

between light and dark Elevated Time in open Anxiety Kawakami Plus Maze arms Entries into

open arms Conditioned Fear and Freezing Kawakami Fear Stimuli learning time Cardiac Hypercontractile Contractibility Lingrel Performance heart

Learning and memory behavioral tests were interpreted differently between the two groups. The Lingrel lab examined spatial learning and memory using the Morris water maze and found that the knock out mouse had a longer latency to find the hidden platform suggesting they had impaired learning. The Kawakami lab studied the knock out mouse in the conditioned fear test, which showed that the knock out mouse had increased freezing time which is typically interpreted as improved learning and memory

(Ikeda et al. 2003). However, the authors concluded that the knock out mouse had enhanced fear; other effects such as increased learning and memory and altered sensory perception could explain the results. The authors did not examine or control for these other possibilities.

The role of Atp1a2 has also been examined in the hearts of the heterozygous mice. While the hearts were histologically the same in Atp1a2 heterozygous and wild- type mice, the Atp1a2 heterozygous mice had hypercontractile hearts (James et al.

1999; Segall et al. 2004; Segall et al. 2005). As noted above, the Atp1a2 heterozygous mice have 50% less protein than the wild-type and reduction of Atp1a2 alters calcium levels in cardiomyocytes suggesting that the intracellular concentration of sodium ions would increase. This increased concentration results in an increase in intracellular calcium levels because the sodium-calcium exchanger is inhibited by the high intracellular sodium levels. The excess intracellular calcium causes the contractions in the cardiomyocytes to increase in strength (James et al. 1999; Segall et al. 2004).

These results suggest Atp1a2 regulates calcium levels and therefore contractibility in the heart.

The group at the Vita-Salute San Raffaele University and Center for Translational

Genomics and Bioinformatics in Italy has generated the first FHM2 knock-in mouse.

This mouse model was created by inserting the T2763C mutation, in exon 19 of the

mouse, which is one of the first mutations that was associated with FHM2 (Leo et al.

2011; Segall et al. 2004; Segall et al. 2005). The T2763C mutation causes the amino acid substitution W887R, which affects the β subunit binding site in Atp1a2, resulting in a misfolding of the protein and, in cell based studies, a complete loss of pump function.

The homozygous knock-in mice are neonatal lethal, similar to the knock-out mice described previously, therefore heterozygous knock-in mice were used for experiments.

The authors suggested that because the W887R mutant protein does not translocate efficiently to the plasma membrane, it is degraded via the proteasome, which results in reduced overall protein in the brain. The knock-in mice were tested for CSD by electrical stimulation and results showed that they were more susceptible to CSD: they had a higher threshold and velocity but equal duration of CSD than the wild-type. These results may be due to the impaired clearance of extracellular potassium and glutamate by astrocytes, comparable to the effects of FHM2 mutations in humans (Pietrobon 2007;

Leo et al. 2011). Further testing on this mouse model will be beneficial to understanding the mechanism of at least this mutation in FHM and possible treatment avenues.

Anxiety Disorder

Characteristics of Anxiety Disorders

Anxiety disorders are the most prevalent psychiatric disorders in the United

States, with approximately 18% diagnosed each year (NIMH 2009). Anxiety is described as a normal response to stress, however it can become excessive, resulting in a recurring negative emotional state with feelings of worry and apprehension, along with cognitive and behavioral changes (Nuss 2015). A wide variety of anxiety disorders exist,

including generalized anxiety disorder (GAD), obsessive-compulsive disorder (OCD), panic disorder, post traumatic stress disorder (PTSD) and social phobia also known as social anxiety disorder (SAD) (Association 2013).

GAD is characterized as an excessive, uncontrollable worry pertaining to daily activities and events which lasts for longer than six months and worsens without treatment (Torpy, Burke, and Golub 2011). OCD is a type of anxiety disorder, affecting about one to three percent of the population, with obsessions or compulsions or a combination of both that produce feelings of uneasiness, apprehension, and fear that are time-consuming, distressing or impairing to daily life (Grant 2014). Panic disorder patients suffer from reoccurring panic attacks of extreme anxiety with ongoing worry of having additional attacks because panic attacks cannot be predicted resulting in significant impairment in daily functioning (Herr et al. 2014). PTSD arises from a traumatic experience resulting in disturbing flashbacks, emotional numbness, and hyperarousal (NIMH 2009). SAD is the most prominent anxiety disorder caused by an intense fear of social situations and the scrutiny of others and is often co-morbid with other psychiatric disorders (Stein and Stein 2008). Anxiety disorders are also often co- morbid with other psychiatric disorders such as substance abuse disorders, alcoholism, and other mood disorders, including depression and bi-polar disorder (Olivier, Vinkers, and Olivier 2013).

Anxiety and alcoholism are often co-morbid, but it has been unclear which disease leads to the other (Buckner and Turner 2009). There are two hypotheses for the co-morbidity of alcoholism and anxiety: 1) increased anxiety results in increased alcohol consumption suggests that the pharmacological effects of alcohol decrease anxiety symptoms, which leads to negative reinforcement and increased risk for developing alcoholism (Quitkin et al. 1972); and 2) anxiety symptoms are a consequence of chronic

alcohol abuse or the withdrawal syndrome (George et al. 1990). Previous research has shown anxiety to consistently correlate with alcohol problems and generally anxiety appears to precede alcohol use (Buckner and Turner 2009), but there could be other factors that lead to both diseases, such as genetic predisposition for both, environmental factors, or exposure to pre-natal environmental factors (Kushner, Abrams, and Borchardt

2000). Further investigation into the cause and co-morbidity of alcoholism and anxiety is needed.

The cause of anxiety is not fully understood and as a complex trait disorder, interactions between genetic, environmental, psychological and developmental factors result in the disease state. It is known that many brain regions are involved in the modulation of anxiety, including the limbic system and the hypothalamic–pituitary– adrenal axis (HPA) (Nuss 2015; Harvey and Shahid 2012); more importantly, neurotransmitter imbalance may lead to the development of an anxiety disorder. The balance between inhibitory and excitatory neurotransmission is crucial to normal brain function and behavior, especially between γ-amino butyric acid (GABA) and glutamate as well as serotonin, dopamine, and epinephrine. The roles of GABA and serotonin have been implicated as the main regulators for anxiety (Nuss 2015); however more recently, glutamate has become a new target for mediating anxiety levels and opens new avenues of treatment (Riaza Bermudo-Soriano et al. 2012; Harvey and Shahid

2012).

Anxiety disorders have been primarily treated with benzodiazepines, but these are known to cause dependence and have sedative side effects (Nuss 2015; Lader

2011). More recently, anxiety has been treated with antidepressants, specifically selective serotonin re-uptake inhibitors (SSRIs). Non-pharmaceutical treatment options also exist, with cognitive behavioral therapy being the most popular because it allows the

patient to understand where the anxiety originates, how to cope with the symptoms associated with an attack, and how to modify behavior to relieve anxiety and promote relaxation (NIMH 2009). The best approach to treat anxiety disorders is a combination of both pharmaceuticals and behavioral therapy.

Relationship between Anxiety and Atp1a2

As mentioned previously, Atp1a2 is found in astrocytes and it is often co- localized with glutamate transporters. Glutamate is the major excitatory neurotransmitter found in the brain and extracellular concentration regulation is extremely important to maintain normal neuronal signaling and to avoid excitotoxicity. Astrocytes are responsible for removing approximately 80% of the glutamate in the extracellular space during neuronal activity, mostly through the glutamate transporters solute carrier family 1

(glial high-affinity glutamate transporter), member 3 or the glutamate aspartate transporter (SLC1A3 or GLAST) and solute carrier family 1 member 2 (SLC1A2 or GLT-

1) (Danbolt 2001). Transporter function of the glutamate transporters are dependent on transmembrane sodium concentration, which is regulated by Atp1a2 and the sodium/calcium exchanger (James et al. 1999; Rose et al. 2009; Cholet et al. 2002).

Previous studies in Atp1a2 deficient mice showed impaired glutamate re-uptake and it has been suggested that during neuronal activity Atp1a2 is activated by increased extracellular potassium which is crucial for neurotransmitter re-uptake (Ikeda et al.

2003). As previously described in this chapter, the heterozygous mice were behaviorally tested, for spatial learning and memory, locomotor activity, and anxiety (Moseley et al.

2007; Ikeda et al. 2003). The behavioral testing showed that the heterozygous mice had deficits in spatial learning/memory, reduced locomotor activity, and displayed an anxiety-

related phenotype. The neuronal over-excitation due to reduced neurotransmitter clearance found in the Atp1a2 heterozygous mice might be causing the increased anxiety seen in the behavioral testing measured by hypo-activity.

Alcoholism

Characteristics of Alcoholism

Alcoholism is a major health issue in the United States. Approximately eight percent of US adults have been diagnosed with Alcohol Use Disorder (AUD) (NIAAA

2006). AUDs arise from drinking "too much, too fast, or too often" (NIAAA 2013). The estimated cost to treat AUD per year is over two hundred billion dollars arising from increased healthcare costs, legal fees, and lost labor productivity (Dick and Foroud

2002). There are two types of AUDs, alcohol abuse and alcohol dependence. Alcohol abuse is characterized by a recurring pattern of high-risk drinking that creates problems for the drinker, others, and society. Adverse consequences can also arise from a single instance of hazardous alcohol abuse. Alcohol dependence is typically considered to be synonymous with alcoholism (alcohol addiction) and is a persistent and intense alcohol- seeking behavior, which results in a loss of control over drinking, a preoccupation with drinking, compulsion to drink or inability to stop, and the development of alcohol tolerance and dependence. Alcoholism is recognized as a complex disorder that results from a combination of genetic and environmental factors (Whitfield et al. 1998). The environmental factors include cultural influences and prenatal exposure (Ducci and

Goldman 2008). Genetic factors include alcohol metabolism and acute sensitivity

(Whitfield et al. 1998). Factors that are both genetic and environmental include family

history and major psychiatric disorders, such as anxiety and depression (Ducci and

Goldman 2008; Whitfield et al. 1998).

Alcoholism in Animals Models

While alcoholism is fundamentally a human problem, like many other complex disorders, much can be learned through studies in animal models, such as rodents, because they allow research that tightly controls both genetic and environmental factors that contribute to alcohol-related behaviors, including alcohol sensitivity, tolerance, withdrawal, dependence and reinforcement (Crabbe 2014). For example, genetic and environmental factors associated with alcohol-related behaviors are studied in rodents because some responses are not ethical or practical to study in humans.

Animal models cannot perfectly model alcoholism, just certain components including alcohol sensitivity, acute tolerance, withdrawal severity, and voluntary consumption. Relating animal experiments to human alcoholism is difficult, but with examination of specific relevant traits and behavioral tests, the relationship between animal and human disease can be defined (Barkley-Levenson and Crabbe 2012).

Genes implicated in alcoholism can easily be studied in animal models, however with environmental factors and gene-gene interactions in humans, relating results from animal alcoholism to human alcoholism is difficult (Ramsden 2015). For example, mice can voluntarily consume large amounts of alcohol, which suggests they have alcoholic tendencies, but they will not necessarily drink to intoxication as human alcoholics would

(Barkley-Levenson and Crabbe 2012). Animal models for alcoholism play a significant role, but it may not always be possible to perfectly model, study, and apply the results gathered to human alcoholism (Ramsden 2015).

One main advantage of using animal models for genetic studies is the ability to tightly control environmental influences, which makes it easier to identify genetic risk factors. Another advantage of animal studies is the high conservation of linked regions of genetic material across many species; genes that are located close to each other on a given chromosome in humans are often located close to each other on one of the chromosomes in mice (Dick and Foroud 2002). The synteny between mouse and human chromosomes allows for discovery of genes related to alcohol-related behaviors in mice to be approximated on the human chromosome. Animal studies also permit breeding strategies that ethically cannot be performed in humans, such as producing genetically identical subjects. Inbred mouse strains are used because all members are genetically identical within a single strain with the exception of male and female. The use of recombinant inbred (RI) strains allows for genetic studies among a reference genetic population with genotypes that only need to be determined once and phenotypes that are conserved over time.

RI strains have been used in genetic studies for more than 40 years to study complex traits and gene-environment interactions (Williams et al. 2001). RI strains are created by crossing two inbred parental strains to produce the F1 generation, which are heterozygous at all markers because they have one allele from each parent (Figure 1.3).

The F1s are then crossed to form the F2 generation, a true segregating population; each allele has regions from the Parent strains due to homologous recombination. Twenty generations of brother-sister matings are completed to ensure homozygousity at all markers; the mice in each strain are genetically identical (McClearn and Kakihana 1981).

These animals are useful for examining environmental contributions to alcohol related responses because genotypes are maintained across subjects (Barkley-Levenson and

Crabbe 2012). Each strain has a fixed genome and can be studied at different times

and in different locations for comparison and collaboration using the same strains

(Williams et al. 2004). A disadvantage to RI panels has been small number of strains, but this could be solved by creating larger RI panels (Williams et al. 2004).

Figure 1.3. Creation of RI strains. Two parental strains are crossed and the resulting generation is crossed until the F2 generation, where 20 generations of brother-sister mating are performed to ensure each marker is homozygous.

One of the first behavioral genetic mouse models for acute alcohol sensitivity were the Long Sleep and Short Sleep mice (LS, SS) that were selectively bred for differential alcohol responses; specifically, the loss of righting reflex (LORR) (McClearn and Kakihana 1973; McClearn and Kakihana 1981). The difference in LORR between the LS and SS strains demonstrated the effect that genetics has on alcohol responses

(Crabbe 1989). The LS and SS strains are now inbred as the ILS and ISS strains and a large panel of RI strains have been bred from the ILS and the ISS, which are called the

LXS mouse strains. The LXS RIs are a good model for the study of differential alcohol responses related to different genotypes among conserved phenotypes.

Overview of Thesis

The overall guiding hypothesis for this comprehensive study is that genetic variation of Atp1a2 affects alcohol related behaviors. This was accomplished by focusing on two main objectives: 1) determining the contribution of Atp1a2 to genetic variance in behavior through examining Atp1a2 expression in a genetic mapping population and discovering if structural variation of Atp1a2 contributed to behavior and 2) determining the effects of Atp1a2 in alcohol related behaviors using a knock-out model.

The current body demonstrates a link between Atp1a2 and alcohol-related behaviors including anxiety and the findings from this study examining the genetics of Atp1a2 and its associated phenotypes are expected to show that genetic variation in Atp1a2 is a contributing factor to AUDs. The following approaches were used to test this hypothesis.

Genetic variation in Atp1a2 in both humans and animal models has been shown to affect neural activity and whole animal behavior (De Fusco et al. 2003; Vanmolkot et al. 2003; Vanmolkot, Kors, et al. 2006; Segall et al. 2004; Segall et al. 2005; Moseley et al. 2007; Ikeda et al. 2003). In Chapter II, genetic variation of Atp1a2 was examined for changes in protein expression of itself and other known associated proteins, such as glutamate transporters, and if expression differences of Atp1a2 correlated with behaviors. There was differential expression of Atp1a2 in the LXS RIs, however it was not controlled by its own gene, and that glutamate transporter expression was affected by Atp1a2 expression. Atp1a2 mRNA correlated strongly to variations in behavior in the

LXS and C57BL/6 and DBA/2 RI strains (BXDs), however this was not observed with differential protein expression. Based on the findings from this section, it is possible that genetic variation in Atp1a2 is contributing to variance in behavior.

In Chapter III, the role of Atp1a2 in alcohol related behaviors was examined, specifically by measuring alcohol sensitivity, acute functional tolerance, withdrawal severity, voluntary drinking, and if alcohol exposure modifies the anxious behavior seen in the Atp1a2 heterozygous mice. The hypothesis for this study was decreased Atp1a2 expression contributes to variation in alcohol related behaviors. There were mixed behavioral results among the sexes. Atp1a2 heterozygous mice voluntarily drink more alcohol, had increased sensitivity and increased rapid tolerance to alcohol, increased acute tolerance, and decreased anxiety with alcohol exposure; however, there was no difference in alcohol metabolism or withdrawal severity.

In summary, this thesis provides a comprehensive study determining the role of

Atp1a2 genetic variance in behavior.

CHAPTER II

BIOINFORMATIC ANALYSIS AND DIFFERENTIAL EXPRESSION OF ATP1A2

Introduction

Atp1a2 is essential to establishing and maintaining the electrochemical gradients of sodium and potassium ions with the utilization of ATP. Atp1a2 is found in astrocytes and is important in the clearance of extracellular potassium during neuronal activity and released glutamate in the synaptic cleft because reuptake in astrocytes and neurons is driven by the sodium and potassium gradients (Ikeda et al. 2003; James et al.

1999; Rose et al. 2009; Cholet et al. 2002). Atp1a2 has been identified as a gene associated with a rare form of hereditary migraines with aura known as familial hemiplegic migraine type II (FHM2). Modulation of Atp1a2 activity affects neural activity and whole animal behavior. Atp1a2 has been studied in association with increased anxiety, learning and motor function disorders, and fear in Atp1a2 deficient mice.

Homozygous null Atp1a2 mice are neonatal lethal, therefore the function of Atp1a2 is studied in heterozygous mice, which have a single functional copy of the allele. The heterozygous Atp1a2 mice have approximately half the protein amount as the wild- types, have increased heart and skeletal muscle contractions, and have increased vascular tone (James et al. 1999; He et al. 2001; Zhang et al. 2005).

The distal end of chromosome 1, where Atp1a2 is located, has had multiple quantitative trait loci (QTLs) mapped to this region with various alcohol-related phenotypes (Mozhui et al. 2008; Ehlers et al. 2010). These QTLs can assist in

identifying regions of genes and candidate genes that result in phenotypic variation.

Previous work in the lab examined gene expression in the striatum in the ILS and ISS strains as a function of alcohol treatment by microarray. Total RNA was isolated from the striatum and measured using the Affymetrix 430_2 microarray. The probe set used to measure Atp1a2 mRNA was located in the distal end of the 3' UTR. Figure 2.1 presents the results of the microarray analysis with the grey line showing the results for the ILS strain and the black line the measurements for the ISS strain. As shown in this figure the ISS mice had a 3.5 fold greater level of Atp1a2 mRNA in their brain tissue compared to the ILS strain. This figure also shows that alcohol treatment caused an increase of Atp1a2 brain mRNA in both the ISS and the ILS strains with mRNA expression of Atp1a2 returning close to baseline levels after 24 hours post alcohol treatment. These results showed that Atp1a2 is differentially expressed between the two strains and could be a contributing factor in acute alcohol sensitivity.

1000 ILS 800 ISS

600

400 units) 200

0 Expression Expression (arbitrary 0 10 20 Time (hrs)

Figure 2.1. Atp1a2 Microarray. Results of microarray analysis in ILS and ISS after 5 g/kg alcohol. Two-ANOVA showed significant strain and time difference (p<0.0001).

NextGen RNA sequencing was also performed to examine mRNA expression between the ILS and ISS. Poly-A RNA was isolated from whole brain (three animals per strain) and prepared for sequencing on an Illumina Genome Analyzer II. As seen in

Figure 2.2, RNA seq data showed that there was approximately equal expression

throughout Atp1a2 in the ILS and ISS whole brain except in the distal end of the 3' UTR.

There were also 106 putative SNPs detected throughout the gene.

The results from the RNA sequencing and the microarray data showed an interesting strain effect, a clear truncation of the 3' UTR in the ILS. The results from the microarray alone would have lead us to believe the ISS have higher expression than the

ILS because the probe set used to measure mRNA in the microarray was located in the

3' UTR region. Since this region is truncated in the ILS strain, the ISS strain showed higher levels of expression.

The results from the microarray and RNA sequencing data show that the truncated 3' UTR in the ILS strain could be important in Atp1a2 activity and protein expression because microRNAs are known to bind to the 3' UTR and inhibit translation.

The multiple SNPs located throughout the Atp1a2 gene have not been studied and the effects of these SNPs on protein function are also unknown, although none appear to be non-synonymous.

The purpose of this study was to determine if variation of Atp1a2 changed protein expression of itself and other known associated proteins, such as glutamate transporters. First, the 106 SNPs located within the gene and the truncated 3' UTR were examined for their affect on protein expression in the recombinant inbred strains derived from the ILS and ISS (LXS RIs). Based on these results, Atp1a2 protein expression and mRNA expression of Atp1a2, glutamate transporters found on astrocytes, and candidate genes for Atp1a2 regulation were examined for co-regulation and contribution to behavior. Further investigation of glutamate transporter expression by quantitative PCR was done with Atp1a2 heterozygous mice. Based on these findings, it is possible that genetic variation in Atp1a2 is contributing to variance in behavior.

Figure 2.2. Atp1a2 NextGen RNA Seq Results. RNA seq results showing the difference in Atp1a2 sequence variants. Data is plotted 3' to 5' because Atp1a2 is located on the anti-sense strand. SNPs are found in both strains and the truncated 3' UTR is located in the ILS.

Methods

Animals

LXS RI breeders were obtained from The Jackson Laboratory (Bar Harbor, ME) and Atp1a2 breeders, on a mixed background, were obtained from Dr. Jerry Lingrel at the University of Cincinnati (James et al. 1999). All mice were bred in-house in the

University of Colorado Anschutz Medical Campus vivarium, a pathogen-free facility.

Atp1a2 wild-type and heterozygous were bred to avoid homozygous null offspring.

Offspring were weaned and sex-separated at 21 days of age. All experiments were conducted with males, females were also used for qPCR, that were group-housed in

standard housing containing from 2 to 5 mice per cage; the mice were 60 to 101 days of age at the time of brain collection. The mice were maintained in a constant temperature

(22 to 23°C), humidity (20-24%), and light (14L/10D) environment. The procedures described in this report have been established to ensure the absolute highest level of humane care and use of the animals, and have been reviewed and approved by the

UCAMC IACUC.

SNP Analysis and LXS RI Genotyping

106 SNPs located throughout Atp1a2, identified by DNA sequencing, (Bennett et al. 2015) were examined for location and type of SNP using the software Geneious and were confirmed by dbSNP. Genotyping of the LXS RIs was done by examining two

SNPs on either side of Atp1a2 (rs13476239 and rs8259388) and two SNPs within

Atp1a2 (rs30934288 and rs31570902) to determine which haplotype each LXS strain had, either ILS or ISS (Table 2.1). DNA was extracted from tail snips, isolated and amplified by PCR, and gel purified samples were sequenced by Sanger sequencing.

Genotyping the LXS RIs was important to determine if strains with the truncated 3' UTR had differential expression of Atp1a2 due to the truncation. Atp1a2 wild-type and heterozygous mice were also examined for the two SNPs present in exon 19 of Atp1a2 to determine whether they had the full or truncated 3' UTR.

Genetic Correlations

The database GeneNetwork (www.genenetwork.org) was used to determine the phenotypes that correlated with the genotype of Atp1a2, naïve protein expression of

Atp1a2, and for QTL analysis. The search parameters included the species mouse, the strain LXS, the tissue type prefrontal cortex, the database VCU LXS PFC Sal M430A 2.0

(Aug06) RMA, and the keyword Atp1a2. The trait analysis for Record ID 1455136_at contains the truncated 3' UTR probe set that determines the genotype for the LXS as either the ILS (truncated 3' UTR) or ISS. We also examined the phenotypes correlated to LXS hippocampal mRNA (Illumina record ILM110278). To compare phenotypes observed in the LXS mice, we examined the phenotypes that correlate with the BXD strains, which are RI strains bred from C57BL/6 and DBA/2 mice which had the 3' UTR truncation as well (Table 2.1). Naïve protein expression data was uploaded and correlated to published LXS phenotype data. P-values were calculated using Pearson's rank.

The LXS RIs are a mix of the ILS and ISS genotypes across the entire genome and QTL mapping takes advantage of this random recombination of parental genotypes to estimate the covariance of a trait with a specific haplotype, in this case the 3' truncation and SNPs located in Atp1a2 (Williams et al. 2001). Interval mapping across the entire genome was performed to determine if Atp1a2 protein expression was driven by the SNPs within Atp1a2 or by another region of genes. 2000 permutations were used to determine the suggestive and significant likelihood ratio statistic (LRS).

Gene-gene correlations were performed on exon array values in Phenogen

(phenogen.ucdenver.edu) to correlate gene probe sets to Atp1a2 mRNA expression values in the LXS RIs (probe set 6764138). Standard filters and a p-value of 0.05 with correction for multiple testing was used. mRNA expression values from Atp1a2, chromosome 19 genes, and the glutamate transporters were also obtained from

Phenogen. Naive Atp1a2 protein expression was correlated to mRNA expression data from the obtained expression values, with correlation coefficients and p-values

calculated using Spearman's rank. 17 genes with SNPs in the ILS and ISS were selected under the suggestive QTL for analysis. Gene lists from the gene-gene correlations were then analyzed using WebGestalt

(http://bioinfo.vanderbilt.edu/webgestalt/) and a Gene Ontology (GO) enrichment analysis was conducted. Exon array data were generated in the lab of Dr. Boris

Tabakoff (Bhave et al. 2007).

Table 2.1. Strain Haplotypes. The strains used throughout the studies either have a truncated 3’ UTR or a complete 3’ UTR. ILS and C57 were the two strains with the 3’ UTR truncation, which was verified on GeneNetwork with Record ID 1455136_at.

Strain 3' UTR Truncation ILS Yes ISS No C57BL/6 Yes DBA/2 No 129/Sv No

Western Blotting and Quantification

Whole brain tissue was collected, homogenized in TBS with 0.1% Protease

Inhibitor, and hydrophobic proteins were extracted using Mem-PER® Eukaryotic

Membrane Protein Extraction Reagent Kit (Thermo Scientific). Thirty (30) micrograms

(µg) protein samples were prepared and run on 10% Bis-Tris gels, transferred to a nitrocellulose membrane and probed with anti-Na+/K+ ATPase α2 (1:10,000 dilution,

Upstate) and anti-Actin (1:1,000 dilution, Sigma Aldrich). Horse radish peroxidase conjugated goat anti-rabbit IgG (1:15,000, Abcam Inc.) was used as a secondary antibody. Samples were standardized to actin and expressed as a ratio of Atp1a2/Actin density. Protein concentration was measured using Pierce BCA Protein Assay Kit

(Thermo Scientific). 29

RNA Isolation and qPCR

Total RNA was extracted from male and female Atp1a2 heterozygous and wild- type mice using RNeasy Mini Kit (Qiagen) following manufacturer's instructions. RNA concentrations were determined by NanoDrop 2000c (Thermo Scientific). Reverse transcription reactions were performed using SuperScript III First-Strand Synthesis Kit

(Invitrogen). Five (5) µg of total RNA was used as a template to synthesize first-strand complementary DNA using random hexamer primer. QuaniTech SYBR Green PCR kit

(Qiagen) was used to measure expression of a 103 bp GLT-1 cDNA fragment (5'-

CCAAGCTGATGGTGGAGTTC-3' and 5'-CAAATCAAGCAGGCGATACC-3'), a 91 bp

GLAST cDNA fragment (5'-ATGTTCCCTCCCAATCTGGT-3' and 5'-

CGTTGGACTGGATAGGCACT-3'), and as a control a 98 bp GAPDH cDNA fragment

(5'-AGTGCCAGCCTCGTCCCGTA-3' and 5'-GCCACTGCAAATGGCAGCCC-3'). The plates were placed in a Mastercycler ep Realplex4 (Eppendorf) for reading. Samples were internally standardized to GAPDH. The fold changes were calculated by taking the

2-∆∆CT of the heterozygous (Het) group average and dividing it by the wild type (WT) group average per sex.

Statistical Methods

Data were analyzed using SPSS Statistics 22 (IBM). All data are expressed as the mean ± S.E.M. Statistical significance was determined by one-way ANOVA, two- way ANOVA, and student’s t-tests for 2 group comparisons. P-value <0.05 was considered significant.

Results

SNP Distribution and LXS RI Genotypes

Each of the 106 SNPs was examined for the SNP location and type of mutation

(Table 2.2) and were confirmed on dbSNP. 50% of the SNPs were located in the 3'

UTR, most of those in the distal end where the truncation occurs and 31% were in introns. There were 20 SNPs in exons, 16 of them were synonymous and 4 were non- synonymous. The 4 non-synonymous SNPs were located in low expressed Atp1a2 transcripts determined by DNA sequencing of the ILS and ISS (Bennett et al. 2015;

Radcliffe, unpublished). Synonymous SNPs do not change the amino acid, but they can affect transcription, splicing, mRNA transport or translation, which can all result in altered activity and expression (Goymer 2007). Both types of SNPs are known to change the protein structure, affect expression and the SNPs located in the UTRs and introns could affect micro RNA binding sites, splice sites, and non-coding RNAs.

The LXS RIs were genotyped based on two of the synonymous SNPs located in exon 19. Approximately 50% of the LXS RIs had the ILS haplotype at Atp1a2, the truncated 3' UTR (Table 2.3). This means the RIs have that specific genotype (ILS or

ISS) at Atp1a2, not the ILS or ISS phenotype.

The Atp1a2 wild-type and heterozygous mice were also genotyped to determine what alleles they had at Atp1a2. We found that the wild-type Atp1a2 mice had the

C57BL/6 haplotype, whereas the heterozygous mice had one allele from the C57BL/6 and one of the 129 allele, which contains the Atp1a2 knock-out mutation and is similar to the DBA allele with the full 3’ UTR. The wild-type mice had two functioning C57BL/6 alleles with the 3' UTR truncation and the heterozygous mice had one functioning

C57BL/6 allele with the 3' UTR truncation; the second allele, with the full 3' UTR was not functional, so any phenotypic differences measured were due to the loss of Atp1a2 protein because the functional alleles all had the 3' UTR truncation.

Table 2.2. SNP Distribution in Atp1a2. 106 putative SNPs were detected in Atp1a2 mRNA via RNA-Seq. 20 SNPs were present in exons, with 4 non-synonymous, however not in highly expressed transcripts.

Region Amount

3' UTR 53

Synonymous 16

Non-Synonymous 4

Intronic 33

5' UTR 0

Genotype-Phenotype Correlations in the LXS and BXDs

The phenotypes that correlated with Atp1a2 pre-frontal cortex mRNA expression were examined in the LXS and BXD mice using GeneNetwork (Record ID 1455136_at).

As seen in table 2.4 anxiety behaviors, seizure susceptibility, alcohol drinking preference, and seizures due to withdrawal were significantly correlated to Atp1a2 expression in the pre-frontal cortex of both the LXS and BXDs. The phenotypes that also correlated with Atp1a2 hippocampal mRNA expression (Illumina record ILM110278) were examined in the LXS and BXDs. Anxiety behaviors, learning and memory, and locomotion were significantly correlated to Atp1a2 mRNA expression (Table 2.4). Both the pre-frontal cortex and hippocampal mRNA expression correlations show that Atp1a2 and anxiety behaviors were significantly correlated.

Atp1a2 Protein Expression in the LXS RIs

Naive brains from 45 LXS RIs, ISS and ILS were measured for Atp1a2 protein expression. There was a two-fold difference between the lowest expressing and the highest expressing LXS RIs (Figure 2.3a). ISS had lower expression than ILS, however this was not significant (Figure 2.3b). The ISS-like LXS RIs also had lower expression of

Atp1a2 than the ILS haplotype LXS RIs, but not significant (Figure 2.3c).

Atp1a2 QTL Mapping and Gene Candidates from Chromosome 19

QTL mapping across the entire genome was performed to determine if Atp1a2 protein expression was driven by variation in its own gene (cis regulation) and SNPs or by another region of genes (trans regulation). The results of the QTL mapping showed there was a suggestive QTL over the distal end of chromosome 19. This means that the genes in this region could be driving the expression of Atp1a2 (Figure 2.4), at least in part. Under the suggestive QTL, there were 17 genes with coding region SNPs found in the ILS and ISS that could be regulating Atp1a2 expression and there were 17 genes that were cis regulated on chromosome 19 that could also be regulating Atp1a2 protein expression; five genes were found to both contain SNPs in the ILS and ISS and were shown to be cis regulated. In total there were 29 genes under the suggestive QTL on chromosome 19 that could be regulating protein expression of Atp1a2 (Table 2.5).

Table 2.3. LXS RI Genotypes. All the LXS RIs were genotyped to determine which haplotype they had at Atp1a2. Each haplotype was approximately equal between all the LXS RIs.

ISS-Like ILS-Like

LXS 5 LXS 46 LXS 89 LXS 2 LXS 48 LXS 88

LXS 7 LXS 52 LXS 90 LXS 3 LXS 49 LXS 92

LXS 9 LXS 54 LXS 93 LXS 8 LXS 50 LXS 98

LXS 13 LXS 56 LXS 94 LXS 10 LXS 51 LXS 99

LXS 23 LXS 59 LXS 96 LXS 14 LXS 55 LXS 101

LXS 25 LXS 60 LXS 97 LXS 16 LXS 61 LXS 102

LXS 26 LXS 64 LXS 98 LXS 19 LXS 62 LXS 112

LXS 31 LXS 66 LXS 100 LXS 22 LXS 76 LXS 117

LXS 32 LXS 68 LXS 103 LXS 24 LXS 78 LXS 122

LXS 34 LXS 70 LXS 107 LXS 28 LXS 80 LXS 123

LXS 36 LXS 72 LXS 110 LXS 35 LXS 84 LXS 124

LXS 38 LXS 73 LXS 114 LXS 39 LXS 86

LXS 42 LXS 75 LXS 115 LXS 41 LXS 87

LXS 43 LXS 79

Table 2.4. Behaviors Correlated to Atp1a2 mRNA expression. Various behaviors, including alcohol sensitivity, anxiety, and fear correlated to pre-frontal cortex and hippocampus Atp1a2 expression in both the LXS and the BXDs using Record ID 1455136_at and Illumina record ILM110278. PFC = pre-frontal cortex, HIP = hippocampus

Parameter # RI Tissue Phenotype r p-value Measured strains Strain Type

LORR Sleeptime 40 -0.34 0.03 LXS PFC

Activity in Plus Maze Closed 15 0.52 0.05 LXS PFC Quadrants Audiogenic Seizure Seizure 8 0.96 8.00E-06 LXS PFC Score Severity Activity in Plus Maze Closed 24 0.67 2.30E-04 LXS PFC Quadrants Alcohol Handling Withdrawal Induced 6 0.94 2.50E-03 BXD PFC Induced Convulsions Convulsions Fear Freezing 6 -0.82 4.20E-02 BXD PFC Conditioning Time

% Time Plus Maze Open 33 -0.46 0.01 LXS HIP Quadrants

Obesity Weight 18 -0.69 9.70E-04 BXD HIP

Morris Spatial 24 0.94 2.30E-03 BXD HIP Water Maze Navigation

Locomotion Activity 6 -0.88 1.50E-02 BXD HIP

1.5

1.0

Atp1a2: Actin Denisty Actin Atp1a2: 0.5

0.0 7 70 98 114 8 94 99 100123 86 96 32 66 112122 5 73 14 103 87 92 25 76 48 43 72 80 24 84 3 101 90 52 41 93 13 26 115 89 23 9 102107 19 22 28 Strain

B C 1.5 1.5

1.0 1.0

0.5

Atp1a2:Actin Denisty 0.5 Atp1a2:ActinDenisty

0.0 0.0 ILS ISS ILS-like ISS-like Strain Genotype

Figure 2.3. Atp1a2 Protein Expression in LXS RIs. A) Differential expression of Atp1a2 was found in the LXS RIs. There was a two-fold difference between the lowest and highest expressing strains. 1-way ANOVA p<0.0001 B) ILS express more Atp1a2 than ISS but not significant. p = 0.21 C) The LXS RIs with the ILS haplotype had higher Atp1a2 than the ISS haplotypes, however it was not significant. p = 0.061

Figure 2.4. Atp1a2 QTL. A) QTL map over the entire mouse genome showed a suggestive peak over the distal end of chromosome 19. B) Close-up of the suggestive QTL over distal chromosome 19, with SNPs [colored yellow]. Figures created on GeneNetwork.

Table 2.5. Candidate Chr. 19 Genes. 29 genes under the suggestive QTL on chromosome 19 that could be regulating protein expression of Atp1a2.

Possible Gene Description Regulation 2700078E11Rik RIKEN cDNA 2700078E11 gene cis regulation 6430537H07Rik RIKEN cDNA 6430537H07 gene cis regulation 6720468P15Rik RIKEN cDNA 6720468P15 gene cis regulation 9930023K05Rik RIKEN cDNA 9930023K05 gene cis regulation Ablim1 actin-binding LIM protein 1 SNPs acyl-CoA synthetase long-chain family Acsl5 SNPs member 5 Add3 adducin 3 (gamma) cis regulation Adrb1 adrenergic receptor, beta 1 cis regulation Afap1l2 actin filament associated protein 1-like 2 SNPs AI450540 expressed sequence AI450540 cis regulation Cacul1 CDK2 associated, cullin domain 1 SNPs E330013P04Rik RIKEN cDNA E330013P04 gene cis regulation empty spiracles homolog 2 (Drosophila) Emx2os cis regulation opposite strand glial cell line derived neurotrophic factor Gfra1 SNPs family receptor alpha 1 Grk5 G protein-coupled receptor kinase 5 SNPs SNPs, cis Hspa12a heat shock protein 12A regulation Nanos1 nanos homolog 1 (Drosophila) SNPs Nhlrc2 NHL repeat containing 2 SNPs Nrap nebulin-related anchoring protein SNPs SNPs, cis Pdzd8 PDZ domain containing 8 regulation SNPs, cis Pnlip pancreatic lipase regulation Pnliprp2 pancreatic lipase-related protein 2 SNPs Prdx3 peroxiredoxin 3 SNPs SNPs, cis Sfxn4 sideroflexin 4 regulation soc-2 (suppressor of clear) homolog (C. Shoc2 cis regulation elegans) solute carrier family 18 (vesicular Slc18a2 SNPs monoamine), member 2 Smndc1 survival motor neuron domain containing 1 cis regulation transcription factor 7-like 2, T-cell specific, Tcf7l2 cis regulation HMG-box vesicle transport through interaction with t- SNPs, cis Vti1a SNAREs homolog 1A (yeast) regulation

Atp1a2 Correlates with Chromosome 19 Candidate Genes and ATP Binding Proteins

The correlation of naive Atp1a2 protein expression, Atp1a2 mRNA expression, glutamate transporters mRNA, and chromosome 19 candidate genes is shown in Table

2.6. Atp1a2 protein and mRNA expression were not significantly correlated to each other (Table 2.6a). Naive protein expression was positively correlated to GLT-1 mRNA expression (p<0.01) (Table 2.6b). Four of the chromosome 19 candidate genes were significantly correlated to Atp1a2 protein expression, two were negatively correlated and two were positively correlated. Atp1a2 mRNA significantly correlated to both GLAST and GLT-1 positively (p<0.01 and p<0.05, respectively). Nine of the candidate genes correlated to Atp1a2 mRNA, two were negatively correlated and seven were positively correlated. Two of the chromosome 19 genes correlated to both the protein and mRNA of Atp1a2. Pdzd8 mRNA expression values were not on Phenogen, and therefore that gene was not included.

Gene enrichment analysis was performed to better understand the biological significance of the genes that correlated to Atp1a2 gene expression. 240 genes were significantly correlated to Atp1a2 gene expression and 10 molecular function categories were enriched in the GO analysis (Figure 2.5). Further analysis showed ATP binding proteins to be significantly correlated to Atp1a2, these genes were biologically significant suggesting that they could be co-regulated with Atp1a2 and affect its function.

GLAST and GLT-1 Gene Expression in Atp1a2 Heterozygous Mice

Solute carrier family 1 (glial high-affinity glutamate transporter), member 3 or the glutamate aspartate transporter (SLC1A3 or GLAST) and solute carrier family 1 member

2 (SLC1A2 or GLT-1) were examined by qPCR to determine if expression was modified 39

by changes in Atp1a2 expression (Figure 2.6). In males, GLAST had equal expression and GLT-1 had higher expression in the heterozygous compared to wild-type. In females, however, both GLAST and GLT-1 were down-regulated in the heterozygous mice. When compared across sex and genotypes (F/M), GLAST was equally expressed in the wild-types and GLT-1 was higher in the female wild-types. In the heterozygous mice, GLAST and GLT-1 expression were increased in males compared to the females.

There was no change in GLAST expression however there was a slight reduction in

GLT-1 expression, but it was not significant (p>0.05), and this finding of reduced expression supports previous research that glutamate transporters are coupled with Na+,

K+ Atpases (Rose et al 2009). The number of mice used for this study was slightly under-powered due to poor breeding; more animals perhaps would have shown a greater genotype effect.

Discussion

In this study, the focus was on the genetic variation of Atp1a2 in the LXS RIs and the heterozygous Atp1a2 mice. First it was determined if changes in Atp1a2 resulted in modified protein structure via non-synonymous SNPs, altered Atp1a2 protein expression and modified expression of known associated proteins. Examination of the SNPs found in Atp1a2 in the ILS and ISS were all synonymous except for four that were present in low expressed transcripts. Although the SNPs were synonymous, they could be affecting protein structure and activity by altering transcription, splicing, mRNA transport or translation (Goymer 2007). The truncation and SNPs in the 3' UTR were thought to affect protein expression because microRNAs are known to bind to the 3' UTR and inhibit translation (Mu and Zhang 2012); therefore ISS and ISS-like LXS RIs would have

lower Atp1a2 protein expression because it has a complete 3' UTR. There are four microRNAs that could potentially bind to the 3' UTR of Atp1a2, two of which bind in the truncation region in the ILS, however there was no difference in protein expression between the ILS and the ISS. Many behaviors, such as increased anxiety and fear, correlated to Atp1a2 mRNA in both the pre-frontal cortex and the hippocampus in the

LXS and the BXD RI panels. Although the 3' UTR truncation and SNPs had no difference on Atp1a2 protein expression, the variation in behavior could be due to their effect on the protein structure.

There was differential protein expression of Atp1a2 in the LXS RIs but there was no statistically significant difference between the ILS and ISS or between the ILS and

ISS haplotypes. There also were no significant correlations between naive Atp1a2 expression and the behaviors previously correlated to the mRNA data in GeneNetwork which could be due to the small number of strains with available behavioral data.

Further examination into the differential expression of Atp1a2 showed this was not due cis-regulation, but perhaps to another region of genes. There was a suggestive QTL mapped to the distal end of chromosome 19, many of which had SNPs located in the ILS and ISS, and these genes along with cis-regulated genes on chromosome 19 could possibly be regulating Atp1a2 expression. Utilizing the Phenogen database and the mRNA expression data from the LXS RIs, there were significant correlations between

Atp1a2 and a few of the genes within the chromosome 19 QTL confidence interval.

An enrichment analysis classified the genes correlated to Atp1a2 to the molecular function and cellular component categories based on their functional characteristics. From the 240 genes significantly correlated to Atp1a2 gene expression,

10 molecular function categories were enriched, in particular ATP binding proteins.

These genes were biologically significant suggesting that they could be co-regulated with Atp1a2 and affect its function.

Table 2.6. Atp1a2 Correlations. A.) Atp1a2 mRNA and protein levels did not correlate to each other, however there were correlations between both the Atp1a2 protein and mRNA with the genes found in chromosome 19. For gene descriptions refer to Table 2.5. B.) Atp1a2 protein and mRNA correlate with glutamate transporter expression. ns-not significant, * p<0.05, ** p<0.01, *** p<0.001.

A) Genes within the 95% CI of QTL on Chromosome 19 Atp1a2 Protein Atp1a2 mRNA

Gene r P-value r P-value Ablim1 0.075 ns 0.2158 ns Acsl5 0.34 ** 0.3467 ** Afap1l2 0.0741 ns 0.2288 ns Atp1a2 -0.0072 ns -0.0072 ns Cacul1 0.0014 ns 0.4604 *** Gfra1 0.4471 *** 0.103 ns Grk5 -0.0443 ns 0.4283 ** Hspa12a -0.0275 ns 0.4168 ** Nanos1 0.274 * 0.0528 ns Nhlrc2 0.0222 ns 0.1753 ns Nrap -0.1233 ns -0.3769 ** Pdzd8 n/a n/a n/a n/a Pnlip -0.0949 ns -0.1698 ns Pnliprp2 -0.2476 * -0.2806 * Prdx3 0.1616 ns 0.3973 ** Sfxn4 -0.0676 ns 0.4636 *** Slc18a2 0.0624 ns 0.1953 ns Vti1a -0.1234 ns 0.4104 **

B) Glutamate Transporter Genes Atp1a2 Protein Atp1a2 mRNA

Gene r P-value r P-value Glast 0.0846 ns 0.3573 ** Glt-1 0.3917 ** 0.2813 *

Figure 2.5. GO Enrichment on Genes Correlated to Atp1a2. The GO enrichment analysis showed that the genes significantly correlated to Atp1a2 were significantly enriched in the molecular function categories.

A GLAST 0.04 Wild-Type Heterozygous

0.03

Ct 

0.02 Relative Expression 2 Expression Relative 0.01

0.00 Female Male

B GLT-1 0.20 Wild-Type Heterozygous

0.15

Ct 

0.10 Relative Expression 2 Expression Relative 0.05

0.00 Female Male

Figure 2.6. GLAST and GLT-1 Expression. mRNA expression of GLAST and GLT-1 in the Atp1a2 heterozygous and wild-type mice. A) GLAST is equally expressed in the males and across wild-types, however in females GLAST is down-regulated in the heterozygous and compared to males. B) GLT-1 is slightly up-regulated in heterozygous males and down-regulated in heterozygous females, but is not significant (p>0.05).

The relationships between Atp1a2, GLAST and GLT-1 were explored because

Atp1a2 is known to be co-localized with the glutamate transporters that are important in the clearance of glutamate and potassium from the extracellular space (Rose et al.

2009; Cholet et al. 2002). mRNA expression data showed both glutamate transporters positively correlated to Atp1a2 mRNA expression. To look further, we performed qPCR with the heterozygous Atp1a2 mice. The reduction in Atp1a2 resulted in a reduction of

GLT-1, but only in females. This reduction in glutamate transporters may possibly lead to decreased glutamate clearance, which is known to alter synaptic transmission and whole animal behavior.

As mentioned in Chapter I, it has been well published that glutamate transporters are coupled to Na+, K+-ATPases (Rose et al. 2009). The ion gradient that Atp1a2 and the other isoforms maintain is essential for establishing the resting membrane potential and for driving nutrient and neurotransmitter uptakes. The clearance of extracellular potassium and released glutamate in the synaptic cleft is dependent on Atp1a2 because reuptake in astrocytes and neurons is driven by the sodium and potassium gradients

(Ikeda et al. 2003). Increased extracellular glutamate coupled with decreased glutamate re-uptake as well as increased extracellular potassium are known to cause cortical spreading depression (CSD) which can result in migraines with aura (Pietrobon 2007;

Koenderink et al. 2005; Montagna 2004). These two factors are hypothesized to be the effect of Atp1a2 mutations that are known to cause FHM2.

FHM2 is a rare form of migraines that are longer than a traditional migraine and patients suffer from hemiplegia and partial paralysis during the aura phase (Gritz and

Radcliffe 2013). Atp1a2 was identified as the gene causing FHM2 in 2003 and accounts for about 20% of the FHM present in families (De Fusco et al. 2003; Vanmolkot, Kors, et al. 2006). Mutations in Atp1a2 are known to cause motor problems, childhood

convulsions, seizures, and cognitive dysfunction (Spadaro et al. 2004; Vanmolkot et al.

2003; Jurkat-Rott et al. 2004; Deprez et al. 2008). Genetic studies of Atp1a2 mutations found in humans have been studied in HeLa and Xenopus oocytes. It was shown that some of the known mutations caused complete loss of pump activity that eventually lead to cell death and others led to decreased pump activity but enough to survive (Bassi et al. 2004; De Fusco et al. 2003; Vanmolkot, Stroink, et al. 2006; Segall et al. 2004; Segall et al. 2005; Koenderink et al. 2005). Genetic studies have also been conducted in mice with either a knock-out or a knock-in mutation. The Atp1a2 knock-out mice, heterozygous because complete knock-out mice were neo-natal lethal due to non- synchronized neuronal firing in the breathing center brain (Moseley et al. 2003; Ikeda et al. 2004; Onimaru and Homma 2007), were found to have increased fear and anxiety as well as hypo-activity and learning dysfunction. The knock-in mouse, a mutation known to cause FHM2, had increased susceptibility to CSD, which is known to cause migraines with aura (Leo et al. 2011). The results from both the knock-out and knock-in studies showed there are both neuronal activity and whole animal behavior effects, which could be due to increased extracellular glutamate.

Glutamate is the major excitatory neurotransmitter found in the brain and extracellular concentration regulation is extremely important for normal neurotransmission and to avoid excitotoxicity. Astrocytes are responsible for removing approximately 80% of the glutamate in the extracellular space during neuronal activity, mostly through the glutamate transporters GLAST and GLT-1 (Danbolt 2001).

Transporter function is dependent on transmembrane sodium concentration, which is regulated by Atp1a2 and the sodium/calcium exchanger (NCX) (James et al. 1999; Rose et al. 2009; Cholet et al. 2002). Previous studies in Atp1a2 deficient mice showed impaired glutamate re-uptake and it was suggested that during neuronal activity Atp1a2

is activated by increased extracellular potassium and is crucial for neurotransmitter re- uptake (Ikeda et al. 2003). Anxiety disorders result from an imbalance of neurotransmitters, predominately glutamate and gamma amino-butyric acid (GABA), and when this equilibrium shifts to more excitatory than inhibitory it causes a state of hyper- excitation. This state of hyper-excitation can lead to excitotoxicity, which can cause neurodegeneration and eventually cell death via calcium influx (Meldrum 2000). Based on the qPCR results, a slight decrease in GLT-1 mRNA was seen when Atp1a2 was decreased which supports the previous findings. The number of mice used for this study was slightly under-powered, it was recommended for an n=6, but only an n=3 was used due to poor breeding; more animals perhaps would have shown a greater genotype effect. This decrease in glutamate clearance, resulting in increased extracellular glutamate concentrations may be the cause of the increased anxiety seen in the heterozygous Atp1a2 mice.

In summary, the data show that Atp1a2 is differentially expressed in the LXS RIs, protein expression is not controlled by the 3' UTR truncation or SNPs within Atp1a2, and that glutamate transporter, GLT-1, expression was slightly affected by Atp1a2 expression. Decreased Atp1a2 in animal models results in increased anxiety which is a known risk factor for many diseases. Atp1a2 mRNA correlated strongly to variations in behavior in the LXS and BXDs, however this is not observed with differential protein expression. Candidate genes with similar functions were correlated to Atp1a2, suggesting they could regulate Atp1a2 expression. Further investigation into regulation of Atp1a2 protein expression and its effect on behaviors need to be performed.

CHAPTER III

THE ROLE OF ATP1A2 IN ALCOHOL RELATED BEHAVIORS

Introduction

Atp1a2 is an integral plasma membrane protein responsible for maintaining the sodium and potassium gradients across cellular membranes with utilization of ATP and is necessary for central cellular processes, such as the secondary transport of metabolites, provides the basis for electrical excitation in neurons, and drives nutrient and neurotransmitter uptake. Previous research using Atp1a2 knock-out mutations have increased our understanding of its effect on neural activity and whole animal behavior.

The homozygous null Atp1a2 mice are neonatal lethal due to lack of synchronized neuronal firing in the breathing center of the brain (Moseley et al. 2003; Ikeda et al.

2004; Onimaru and Homma 2007) and the Atp1a2 heterozygous mice have approximately half the protein compared to wild-type (James et al. 1999).

Multiple behavioral tests in mice demonstrated the Atp1a2 heterozygous mice were more anxious than the wild-type (Lingrel et al. 2007; Moseley et al. 2007; Ikeda et al. 2003). Atp1a2 heterozygous mice were hypo-active compared to wild-type based on measurements of total distance traveled and time spent in the corners of the open-field testing box (Moseley et al. 2007), spent less time in and had fewer entries into the open arm of the plus maze (Lingrel et al. 2007; Moseley et al. 2007; Segall et al. 2004), and spent less time in the light compartment and had fewer transitions between the two compartments (Ikeda et al. 2003). The Atp1a2 heterozygous mice also showed

hyperphagia during the light period and suffered from late onset obesity (Kawakami et al.

2005).

Since Atp1a2 has been shown to be involved in anxiety and anxiety is a known risk factor for developing alcoholism (Olivier, Vinkers, and Olivier 2013), we have been investigating Atp1a2 for its potential role in responses to alcohol. Alcoholism is a growing problem in the United States and is recognized as a complex disorder that results from a combination of genetic and environmental factors (Whitfield et al. 1998).

Environmental factors include cultural influences and prenatal exposure, whereas genetic factors include alcohol metabolism and acute sensitivity (Whitfield et al. 1998).

Factors that are both genetic and environmental include family history and major psychiatric disorders, such as anxiety and depression. Understanding genetic factors related to the effects of alcohol in the brain may provide new avenues of treatment.

To test our hypothesis that Atp1a2 is related to alcohol effects, we studied various alcohol related behaviors in heterozygous Atp1a2 mice. Atp1a2 heterozygous mice have one functional copy of Atp1a2 and have been shown to have approximately half the protein compared to wild-type. The aim of this work was to study the effects of

Atp1a2 in five behavioral areas: acute functional tolerance (AFT), acute alcohol sensitivity, withdrawal severity, voluntary drinking, and the anxiolytic effects of alcohol.

Methods

Animals

Atp1a2 breeders, on a mixed background, were obtained from Dr. Jerry Lingrel at the University of Cincinnati (James et al. 1999) and bred in-house in the University of

Colorado Anschutz Medical Campus (UCAMC) vivarium, a pathogen-free facility.

Atp1a2 wild-type and heterozygous mice were bred to avoid homozygous null offspring, have both genotypes present in each litter, and to avoid any maternal effects. Offspring were weaned and sex-separated at 21 days of age. All experiments were conducted with males and females that were group-housed in standard housing containing from 2 to 5 mice per cage; the mice were 61 to 102 days of age at the time of testing. The mice were maintained in a constant temperature (22 to 23°C), humidity (20-24%), and light

(14L/10D) environment. The procedures described in this report have been established to ensure the absolute highest level of humane care and use of the animals, and have been reviewed and approved by the UCAMC IACUC.

Alcohol Metabolism

Male and female Atp1a2 heterozygous and Atp1a2 wild-type mice were injected intraperitoneally (ip) with alcohol (5 g/kg, 16% w/v in normal saline) and blood samples were drawn at 0.5, 1, 1.5 and 2 hours after administration. Blood ethanol concentrations

(BEC) values were determined by spectrophotometry (Lundquist 1959).

Loss of Righting Reflex

This was a two-day test to determine rapid tolerance, acute functional tolerance and alcohol sensitivity (Radcliffe, Floyd, and Lee 2006). On day one, mice were injected

(ip) with either 0.9% normal saline (0.01 ml/g) or alcohol (5 g/kg, 16% w/v in normal saline), a pretreatment dose to elicit a rapid tolerance (Radcliffe et al. 2005). No behavioral testing was done on day one, mice were injected and placed back in the

home cage with littermates. Twenty four hours following this single administration, all groups were tested for duration of the loss of righting reflex following a single acute dose of alcohol (ip; 3 g/kg, 16% w/v in normal saline). Immediately after injection, the mouse was placed into a small Plexiglas cylinder fitted with square end caps. The cylinder was rotated 90 degrees every 2-3 seconds until the time at which the mouse became and remained supine for at least 5 seconds; this is defined as the loss of righting reflex

(LORR). A 20 micro-L retro-orbital blood sample (BEC1) was drawn at this point for determination of blood ethanol concentration (BEC) at loss of righting (initial sensitivity).

The animals were tested for recovery of LORR every 3-6 minutes thereafter. A second blood sample (BEC2) was drawn when the animal is able to right itself within a 5 second period after being placed in a supine position or could not be placed on its back after 8 successive 90 deg turns of the cylinder. Duration of LORR (sleeptime) was defined as the time between the loss and regain of the righting reflex. Rapid tolerance was calculated by a decrease in sleeptime between pre-treatments. An increase in BEC2 from BEC1 was interpreted as development of acute functional tolerance (AFT) which is quantitatively expressed as the difference between BEC2 and BEC1. BEC values were determined by spectrophotometry (Lundquist 1959).

Handling Induced Convulsions

Ethanol withdrawal was measured by handing-induced convulsions (HIC) scores

(Crabbe, Merrill, and Belknap 1991; Metten and Crabbe 1994). Mice were picked up by the tail, observed for 10 seconds, and convulsions were scored; if no convulsion occurred, the mice were rotated in a 180 degree arc for 5 seconds and observed for

HICs. HICs are measured on a scale of 0-7 with 0 being no seizure activity at all and 7

being full spontaneous tonic-clonic seizures observed in the home cage (Table 3.1).

Baseline HIC scores were determined, and mice were given a 4 g/kg intraperitoneal injection of alcohol (16% w/v in saline) once a day for 3 consecutive days. The HICs were recorded hourly for 12 consecutive hours and at 24 hours starting immediately after the first injection. Scoring for all animals was the same rater, all tests were recorded, and re-analyzed.

Table 3.1. HIC Score Scale. Mice were observed and convulsions scored for severity.

Score Symptom 0 No convulsion, even after tail spin 1 Only facial grimace after gentle 180° spin No convulsion when lifted by tail, but tonic convulsion with gentle 180° 2 spin 3 Tonic-clonic convulsion after gentle 180° spin 4 Tonic convulsion when lifted by the tail Tonic-clonic convulsion when lifted by the tail, often onset delayed by 5 as much as 1-2 seconds Severe tonic-clonic convulsion when lifted by tail, with quick onset and 6 long duration, often continuing after release Severe tonic-clonic convulsion with quick onset and long duration; 7 spontaneous or elicited by mild environmental stimulus, in home cage

Drinking in the Dark

The procedure for drinking in the dark (DID) was from Rhodes et al 2005 and used because the short ethanol access for the first three days (training days) led to greater alcohol consumption and BECs on day four (Rhodes et al. 2005). Mice were switched to a reverse light/dark schedule two weeks prior to testing, lights off at 12:00pm and lights on at 10:00pm. One week before testing began, the mice were individually housed with ball bearing sipper water bottles. Mice were weighed one hour before lights out on testing days. Three hours after the lights out, the water bottles were replaced with a 25 mL glass tube with a ball bearing sipper containing 20% alcohol (v/v in tap 52

water). Fluid amounts were measured at time zero and again at time two hours. The alcohol was removed at the two hour time point and replaced with water. The procedure was repeated for three days. On day four, the alcohol bottles were available for four hours and the fluid levels are measured at time zero, two hours, and four hours. Alcohol intake was recorded and calculated in g/kg. After the last reading, a 20 micro-L blood sample was collected to measure blood ethanol concentrations. BEC values were determined by spectrophotometry (Lundquist 1959).

Open-field Activity

Open field activity was used to determine anxiety by measuring locomotor activity and position preference and was modified from J C Crabbe, Johnson, Gray, Kosobud, &

Young, 1982. Each mouse was injected (ip) with either 0.9% normal saline (0.01 ml/g) or alcohol (1 or 2 g/kg, 16% w/v in normal saline). Immediately after injection, the mouse was placed into the center of a large white plexi-glass open box (44 cm x 44 cm x

20 cm box) with the room lights on. Locomotor activity and position preference for time spent in the corners, center, and borders were measured automatically over a twenty minute time period by digital recording equipment, Noldus Information Technology -

EthoVision XT.

Elevated Plus Maze

Elevated plus maze was used to determine anxiety by time spent in the open and closed arms of the maze as well as how many entries were made and was modified from

Stinchcomb, Bowers, & Wehner, 1989. The elevated plus maze consists of two open

arms (30 cm x 5 cm), a center platform (5 cm x 5 cm) and two closed arms with non- transparent walls (30 cm x 5 cm x 14 cm) elevated 37 cm off the ground. Mice were injected (ip) with either 0.9% normal saline (0.01 ml/g) or alcohol (0.5 or 1 g/kg, 16% w/v in normal saline). Fifteen minutes after injection, the mouse was placed into the central platform of the maze with its head facing a closed arm. The frequency of entry into the open and closed arms as well as the amount of time spent in the open and closed arms was measured over a fifteen minute time period by EthoVision XT (Noldus Information

Technology) digital equipment.

Statistical Methods

Data were analyzed using SPSS Statistics 22 (IBM). All data are expressed as the mean ± S.E.M. Statistical significance was determined by three-way ANOVA, two- way ANOVA, logistical regressions, and student’s t-tests for 2 group comparisons.

Results

No Difference in Alcohol Metabolism

Multiple blood samples were taken to determine if alcohol metabolism was affected by the Atp1a2 partial knock-out. The rate of alcohol metabolism in the wild-type was 58.88 ± 13.2 % mg/hr and the heterozygous was 72.30 ± 10.63 % mg/hr. There was no difference in alcohol metabolism between the sexes, so they were combined and analyzed for genotypic differences. No difference in alcohol metabolism between the

Atp1a2 heterozygous and Atp1a2 wild-type was seen (Figure 3.1).

650

Wild-Type Heterozygous 600

550

BEC mg%BEC 500

450

400 0.0 0.5 1.0 1.5 2.0 2.5 Time (hrs)

Figure 3.1. Alcohol Metabolism in the Atp1a2 Mice. There is no difference in alcohol metabolism between genotypes or sex. n = 3 per sex and genotype

Increased Sensitivity, Rapid Tolerance and Acute Functional Tolerance in Atp1a2

Heterozygous Mice

Acute sensitivity and AFT in male and female Atp1a2 wild-type and Atp1a2 heterozygous mice was examined using LORR (Figure 3.2). Sleeptime was significantly decreased in females compared with males (p<0.001) and there was a genotype effect seen between the saline pre-treated females (p<0.05), but no effect was seen in males.

Heterozygous females also showed rapid tolerance. There was no difference between pretreatment with alcohol or saline between the genotypes however there was an interaction of sex-genotype-pretreatment (p<0.01) (Figure 3.2a). The interaction between sex-genotype-pretreatment was primarily driven by the strong sex effect seen and a genotype effect with between the saline pre-treated females.

A LORR SE 80 EE

Gender *** Interaction ** ______

60 *

40 Sleeptime (mins)

0 Females WT Females HET Males WT Males HET

B AFT SE 100 EE ______Interaction *

______** 80

40 AFT (BEC2-BEC1, mg%) 20

0 Females WT Females HET Males WT Males HET

Figure 3.2. LORR and AFT. A) Females had a shorter LORR duration than males, heterozygous females showed increased sensitivity to alcohol and rapid tolerance. There was an interaction of sex, genotype and treatment. B) AFT demonstrated that Atp1a2 heterozygous females had increased acute tolerace to alcohol, no effect was seen in males. There was an interaction of sex, genotype, and treatment. n = 20 per sex and genotype, *p<0.05, **p<0.01, ***p<0.001. SE = Saline pre-treatment, EE = Ethanol pre-treatment

AFT showed no difference between sex and genotype, however there was a difference in pretreatment between female Atp1a2 heterozygous mice (p<0.01) and

there was a sex-genotype-pretreatment interaction (p<0.05) (Figure 3.2b). This interaction of sex-genotype-pretreatment was driven by the strong pre-treatment effect in the female Atp1a2 heterozygous mice and a genotype effect between the EE treated females. This demonstrates the Atp1a2 heterozygous females had increased sensitivity to alcohol, increased rapid tolerance, and increased acute tolerance.

No Difference in Alcohol Withdrawal Measured by Handling Induced Convulsions

Handling induced convulsions were studied to determine if Atp1a2 expression affected alcohol withdrawal. Male and female Atp1a2 wild-type and Atp1a2 heterozygous mice showed no difference in withdrawal severity across any of the three days (Figure 3.3). There was no difference between sexes, so they were combined and then analyzed. There was no difference in withdrawal due genotype and there was no interaction of sex and genotype.

Voluntary Consumption Increased in Atp1a2 Heterozygous Mice

Atp1a2 heterozygous mice have previously been shown to have increased anxiety (Moseley et al 2007, Ikeda et al 2003) and because anxiety is a risk factor for alcoholism this led to the hypothesis that the heterozygous mice would voluntary consume more alcohol. Females of both genotypes consumed more alcohol than the males (p<0.001) however there was no difference in their drinking between genotypes

(Figure 3.4a). On day four, during the first two hours Atp1a2 heterozygous male mice drank more than their wild-type controls (p<0.05). Drinking behavior on day four was significantly correlated between the first and last two hours (p<0.01) (Figure 3.4b).

BECs and total alcohol consumed on day 4 were significantly correlated (p<0.001), however there was no difference between sexes and genotypes. There was an interaction between genotype and sex (p<0.05) which was driven by the genotype difference in the males and the sex effect seen in the females by increased consumption

(Figure 3.4c).

0.5

Wild-Type Heterozygous 0.4

0.3

HIC Score 0.2

0.1

0.0 0 20 40 60 Time (hrs)

Figure 3.3. Alcohol Withdrawal. There was no difference in withdrawal between sexes or genotypes. n = 10 per sex and genotype

Decreased Anxiety Seen in Open-field Activity with Alcohol

The total distance travelled, position preference, and rearings were measured to determine the anxiolytic effects of alcohol. There was no overall difference in the total distance travelled between genotypes, but there was a dose effect (p<0.001), which was expected and there was an interaction of sex-genotype-dose (p<0.05) caused by the genotype difference in the 1 g/kg dose in males (Figure 3.5a). The 1 g/kg dose caused

an increase in the amount travelled in the male Atp1a2 heterozygous mice (p<0.05).

Position preference measured by time spent in the border, time spent in the center and the amount of time spent in the corners as well as the number of rearings showed no genotype effect in either sex, only dose-dependent effects (Figure 3.5b-e).

Elevated Plus Maze Showed Decreased Anxiety after Alcohol Exposure

Time spent in the open and closed arms, as well as the number of entries into the open and closed arms were measured. There was no statistically significant difference between sex, genotype and dose in percent open arm entries (Figure 3.6a).

Three-way ANOVA analysis showed a genotype-dose interaction (p<0.05) in percent time in open arms due to the genotype effect in the 1 g/kg dose, however there were no overall sex, genotype, or dose effects seen (Figure 3.6b). The 1 g/kg dose caused heterozygous females to spend more time in the open arms (p<0.05) (Figure 3.6b).

Discussion

This was the first study to examine the relationship between Atp1a2 and alcohol.

The goal of this project was to determine if Atp1a2 was related to the behavioral effects of alcohol by examining alcohol sensitivity, acute functional tolerance, withdrawal severity, voluntary drinking, and anxiety-like behaviors. First, it was determined that

Atp1a2 heterozygous knock-out did not affect alcohol metabolism in either sex.

Sleeptime was significantly decreased in females compared with males and there was increased rapid tolerance with the heterozygous females, but there was no difference between pre-treatments in the other groups. There was a genotype effect between the

saline pre-treatment females which showed the heterozygous mice had increased sensitivity. AFT also showed no difference between sex and genotype, however there was a difference in pretreatment between female Atp1a2 heterozygous mice. This showed the Atp1a2 heterozygous females were more sensitive to alcohol, had increased rapid tolerance and had greater acute tolerance.

Atp1a2 mutations in humans have been shown to cause seizures (Jurkat-Rott et al. 2004; Vanmolkot et al. 2003; Deprez et al. 2008) and seizures are a side effect of alcohol withdrawal. Seizures scores after chronic withdrawal showed no difference between sexes or genotype. Seizures were also not detected with naive mice (baseline score on day 1). The seizure scores, other than zeros, that were measured were only ones or a facial grimace. These 1-scored seizures occurred in both sexes and genotypes but as shown, there was no difference. There may not have been any withdrawal effects seen because this was an acute time period of alcohol treatment, perhaps with increased alcohol treatment there would be a difference in withdrawal severity. Anxiety is frequently co-morbid with alcohol use disorders (AUD) and it is believed alcohol is used to manage distress associated with the disease (Buckner and

Turner 2009). This correlation to increased drinking led to the hypothesis that the anxious Atp1a2 heterozygous mice would drink more alcohol voluntarily. Females of both genotypes consumed more alcohol than the males however there was no difference in their drinking between genotypes. There was a difference in drinking behavior during the first two hours on day four, Atp1a2 heterozygous male mice drank more than their wild-type controls. Drinking behavior on day four was significantly correlated between the first and second two hour blocks and there was a strong correlation between BECs and total alcohol consumed on day 4. There was an

interaction between sex and genotype, however there was no difference between sexes and genotypes.

4 Wild-Type Heterozygous

______Gender *** 3

2 ______*

Ethanol Comsumed g/kg 1

0 Female (0-2hr) Female (2-4hr) Male (0-2hr) Male (2-4hr)

B C

6 200

150

100 4 hr g/kg

2 BEC mg% 50

0 2 4 6 8 10 0 g/kg 0 2 4 6 8 2 hr g/kg -50

Figure 3.4. Drinking in the Dark. A.) Atp1a2 heterozygous males drank more in the first two hours on day 4 and females of both genotypes drank more than the males. B.) Day 4 hours 0-2 and hours 2-4 consumption is significantly correlated (r2= 0.3, p<0.01). C.) Total g/kg on day four and BEC values were significantly correlated (r2= 0.2, p<0.001) but no difference between genotype or sex was seen, there was an interaction between sex and genotype (p<0.05). n = 18-29 per sex and genotype, *p<0.05, ** p<0.01, ***p<0.001

12000 Wild-Type Heterozygous ______Dose *** Interaction * 10000 ______*

8000

6000

4000 Total Distance Travelled, cm

2000

0 Female 0 g/kg Female 1 g/kg Female 2 g/kg Male 0 g/kg Male 1 g/kg Male 2 g/kg B C

800 Wild-Type 500 Heterozygous Wild-Type ______Dose *** Heterozygous ______Dose *** 400 600

300

400

200 Time Spent in Center, s Center, in Spent Time Time Spent in Corners, s in Spent Time 200 100

0 0 Female 0 g/kg Female 1 g/kg Female 2 g/kg Male 0 g/kg Male 1 g/kg Male 2 g/kg Female 0 g/kg Female 1 g/kg Female 2 g/kg Male 0 g/kg Male 1 g/kg Male 2 g/kg D E Wild-Type Heterozygous Wild-Type 1500 150 Heterozygous ______Dose *** ______Dose ***

1000 100 Rearings

500 50 Time Spent in Border, s in Spent Time

0 0 Female 0 g/kg Female 1 g/kg Female 2 g/kg Male 0 g/kg Male 1 g/kg Male 2 g/kg Female 0 g/kg Female 1 g/kg Female 2 g/kg Male 0 g/kg Male 1 g/kg Male 2 g/kg Figure 3.5. Open Field Activity. A.)Total distance travelled measured over twenty minutes. There was no overall genotype effect, only a dose effect and an interaction between sex, genotype, and dose. There was a genotype effect seen in 1 g/kg in the males. B-E.) Time spent in the corners, center, and borders and the number of rearings showed no difference between sexes and genotypes, there was a significant dose effect seen (p<0.001). n = 10 per genotype and sex, *p<0.05, **p<0.01, ***p<0.001 62

40 Wild-Type Heterozygous

20 % open arms % entries

0 Female 0 g/kg Female 0.5 g/kg Female 1 g/kg Male 0 g/kg Male 0.5 g/kg Male 1 g/kg

30 Wild-Type Heterozygous ______Genotype-Dose Interaction *

______* 20

10 % time spent in open armsin spent time %

0 Female 0 g/kg Female 0.5 g/kg Female 1 g/kg Male 0 g/kg Male 0.5 g/kg Male 1 g/kg

Figure 3.6. Elevated Plus Maze. A.) The number of entries into the open and closed arms was measured for fifteen minutes. No effect of sex, genotype or dose was seen. B.) The amount of time spent in the open and closed arms was measured. There was no overall genotype or dose effect, only an interaction between genotype and dose (p<0.05), however there was a genotype effect in females at the 1 g/kg dose (p<0.05). n = 6-10 per genotype and sex, *p<0.05

To further investigate whether alcohol affected anxious behaviors, open field activity and elevated plus maze were performed. There was no difference in genotypes or sex in open field activity but there was a strong dose effect. There was a genotype difference in the 1 g/kg dose in males, showing the Atp1a2 heterozygous were less anxious with alcohol exposure. This was confirmed after results from elevated plus maze showed a decrease in anxiousness in female Atp1a2 heterozygous mice by an increase in the amount of time spent in the open arms. There were no other significant differences between sexes and genotypes. There was a trend towards a dose response, however this was not significant. Previous OFA and EPM behavioral testing with alcohol have shown that the 1 g/kg dose significantly increases exploratory activity and the 2 g/kg dose initially increases activity with a reduction of activity, approximately ten minutes after injection, due to the sedation effects of alcohol (Crabbe et al. 1982;

Stinchcomb, Bowers, and Wehner 1989; Palmer et al. 2002).

Interestingly, the genotype effect that has been previously published with open field activity and elevated plus maze (Moseley et al. 2007; Ikeda et al. 2003) was not observed. This lack of phenotype could be due to several factors. First, all our mice were weighed and injected. This handling could have affected their behavior, whereas the mice used by Ikeda et al 2003 were completely naive. The mice used in the study by

Moseley et al 2007 were previously tested in the elevated zero maze prior to testing in the open field; this too could be a factor in why the phenotype was more pronounced.

Another possibility could be a genotype-environment interaction; for example, the two labs a mile apart in elevation and the animal facilities and care staff were different which could all affect behavior. Our sample size was also small, 6-10 per group, perhaps the phenotype would be more pronounced with a larger sample size. The Atp1a2

heterozygous mice were slightly more anxious compared to the wild-type in OFA and

EPM, but this difference was not statistically significant.

As previously discussed in Chapter II, there was a slight reduction GLT-1 in the

Atp1a2 heterozygous female mice. It has been well published that glutamate transporters are coupled to Na+, K+-ATPases (Rose et al. 2009) and the clearance of extracellular potassium and released glutamate in the synaptic cleft is dependent on

Atp1a2 and the sodium and potassium gradients it maintains (Ikeda et al. 2003). The increase of extracellular glutamate along with its decreased removal as well as an increase in extracellular potassium are known to cause cortical spreading depression

(CSD) which can result in migraines with aura (Pietrobon 2007; Koenderink et al. 2005;

Montagna 2004) and have been hypothesized to be the effect of Atp1a2 mutations that are known to cause FHM2. This decrease in glutamate clearance which results in increased extracellular glutamate concentrations may be the driving factor for the increase anxiety seen in the heterozygous Atp1a2 mice in the previous behavioral studies (Moseley et al. 2007; Ikeda et al. 2003). The anxiety effect was not seen in these studies, which could have been due to the small amount of animals tested and the increased handling.

Acute alcohol exposure results in an increase of inhibitory neurotransmission and decreases the function of excitatory neurotransmitters, such as glutamate (Valenzuela

1997). It is possible that in the heterozygous Atp1a2 mice, the alcohol causes decreased glutamate transmission, which decreases the amount of glutamate released.

This decrease in released glutamate and therefore lower extracellular glutamate might allow the glutamate transporters to clear the glutamate that is released without the side effects associated with increased extracellular concentrations, thus decreasing anxiety.

However, it has been shown that alcohol inhibits Na+, K+-ATPases, which may increase

extracellular glutamate concentrations, but it may not be consistently causing inhibition or not causing inhibition at all on Atp1a2 (Botta et al. 2010). The authors compared the inhibition of Na+, K+-ATPases by alcohol to the inhibition of the pump by ouabain and found similar results, concluding alcohol inhibits Na+, K+-ATPase function. Ouabain is a non-selective antagonist which inhibits the enzyme transport activity (Crambert et al.

2004) but is non-specific in its binding of Na+, K+ Atpases and Atp1a2 has been shown to be slightly insensitive to ouabain (Munzer et al. 1994). This study possibly only inhibited the neuronal Na+, K+-ATPase, Atp1a1, and further research into how Na+, K+-ATPases are inhibited by alcohol need to be conducted.

Chronic alcohol exposure results in increased excitatory neurotransmission to restore the equilibrium between inhibitory and excitatory neurotransmission (Valenzuela

1997). During withdrawal, the increase in excitatory neurotransmission can result in seizures and increased anxiety, however the alcohol withdrawal study did not show increased seizures. This suggests that the increased anxiety seen might not be due to increased extracellular glutamate concentrations from the decrease in glutamate transporters, but other genes and proteins that are affected by the Atp1a2 partial knock- out. It would be interesting to further investigate if there is increased anxiety in chronic alcohol treated Atp1a2 heterozygous mice utilizing open field activity and elevated plus or zero maze.

In summary, the data show that Atp1a2 heterozygous mice show increased sensitivity, increased rapid tolerance, increased acute functional tolerance, and voluntarily drink more however there was no difference in alcohol metabolism or withdrawal severity. Atp1a2 heterozygous mice have decreased anxiety with alcohol exposure, which could be due to decreased glutamate release and therefore decreased extracellular glutamate. There were interactions between genotype, sex, and alcohol

treatment, suggesting that Atp1a2 slightly contributes to alcohol related behaviors.

There were differences between genotypes in the voluntary drinking study with the heterozygous mice consuming more alcohol and the anxiety models studied with the attenuation of anxiety after alcohol exposure. Further investigation into the cause of the increased anxiety found in the Atp1a2 heterozygous mice as well as discovering if other proteins are modified by the partial knock-out of Atp1a2 are needed.

CHAPTER IV

SUMMARY AND FUTURE DIRECTIONS

Summary

Atp1a2, an integral plasma membrane protein responsible for maintaining the sodium and potassium gradients across cellular membranes with utilization of ATP, fuels central cellular processes, such as the secondary transport of metabolites, provides the basis for electrical excitation in neurons, and drives nutrient and neurotransmitter uptake. Atp1a2 is found in astrocytes and is essential in clearing extracellular potassium during neuronal activity and necessary for clearance of released glutamate in the synaptic cleft. Previous research using Atp1a2 knock-out and knock-in mutations have increased our understanding of its effect on neural activity and whole animal behavior.

Specifically, in humans, mutations in Atp1a2 result in a familial FHM2 and in mice a partial knock-out results in increased anxiety.

The link between Atp1a2 and anxiety and the knowledge that anxiety is known risk factor for developing alcoholism, led to the investigation of Atp1a2 for its potential role in behavioral responses to alcohol. Alcohol abuse is a significant public health issue and understanding the genetic factors related to the effects of alcohol in the brain may provide new treatments and since anxiety is often co-morbid with alcoholism, further investigation into the relationship between these two diseases is needed. The studies presented here determined if the anxiety previously examined in the Atp1a2

heterozygous mice (Moseley et al. 2007; Ikeda et al. 2003) had any effect on alcohol related behaviors.

The overall hypothesis for this comprehensive study was that genetic variation of

Atp1a2 affects behavior. The two main objectives of the project were: 1) to determine the contribution of Atp1a2 to genetic variance in behavior by examining Atp1a2 expression in a genetic mapping population and discovering if structural variation of

Atp1a2 contributed to behavior and 2) to determine if a partial knock-out of Atp1a2 affected responses in alcohol related behaviors.

In Chapter II, the focus was on the genetic variation of Atp1a2 in the LXS RIs and the heterozygous Atp1a2 mice. It was determined if changes in Atp1a2 resulted in modified protein structure via non-synonymous SNPs, altered Atp1a2 protein expression and modified expression of known associated proteins. This was accomplished by first examining the SNPs found in Atp1a2 in the ILS and ISS, which were all synonymous except for four that were present in poorly expressed truncated transcripts. The truncation and SNPs in the 3' UTR were thought to affect protein expression because microRNAs are known to bind to the 3' UTR and inhibit translation (Mu and Zhang 2012).

Many behaviors, such as increased anxiety and fear, correlated to Atp1a2 mRNA in both the pre-frontal cortex and the hippocampus in the LXS and the BXD RI panels and the variation in behavior could be due to the 3' UTR truncation and how it affects the protein expression of Atp1a2. There was differential expression of Atp1a2 in the LXS RIs but there was no difference between the ILS and ISS or between the ILS and ISS haplotypes. There also was no correlation between naive Atp1a2 expression and the behaviors previously correlated to the mRNA data in GeneNetwork.

Further examination into the differential expression of Atp1a2 showed this was not due to the SNPs within its gene but perhaps to another region of genes because a

suggestive QTL mapped to the distal end of chromosome 19, many of which had SNPs located in the ILS and ISS, and these genes could possibly be regulating Atp1a2 expression. These candidate genes located in the suggestive QTL and the genes significantly correlated to Atp1a2 would not account for all the genetic variation if they were indeed regulating Atp1a2 protein expression. Protein expression can be modulated by both genetic and environmental variation. In some cases, the variation seen could be due to only environmental variation, such as diet, drug treatments and stress, whereas in some cases it is a combination of environmental and genetic variation, which includes random mutations and other gene-gene interactions (Griffiths,

Miller, and Suzuki 2000).

Significant correlations between mRNA from Atp1a2 and a few of the genes under chromosome 19 were seen. The relationship between Atp1a2, GLAST and GLT-1 was also explored because Atp1a2 is known to be co-localized with the glutamate transporters. A reduction in Atp1a2 resulted in a slight reduction of GLT-1, but only in females. This reduction in glutamate transporters may possibly lead to decreased glutamate clearance, which is known to alter synaptic transmission and whole animal behavior.

In Chapter III, the effect of Atp1a2 was determined in five behavioral areas: alcohol tolerance, alcohol sensitivity, withdrawal severity, voluntary drinking, and the anxiolytic effects of alcohol. It was first determined that that Atp1a2 heterozygous knock-out did not affect alcohol metabolism in either sex. Acute sensitivity, rapid tolerance, and AFT were examined and showed the Atp1a2 heterozygous females were less sensitive to alcohol, showed rapid tolerance, and had greater acute tolerance.

Atp1a2 mutations in humans have been shown to cause seizures (Jurkat-Rott et al.

2004; Vanmolkot et al. 2003; Deprez et al. 2008) and seizures are a side effect of

alcohol withdrawal, however when Atp1a2 heterozygous mice were tested for alcohol withdrawal severity, there was no difference seen between sexes or genotypes. As mentioned previously, anxiety is frequently co-morbid with AUD and we determined that females of both genotypes consumed more alcohol than the males however there was no difference in their drinking between genotypes. During the first two hours Atp1a2 heterozygous male mice drank more than their wild-type controls. Further investigation into whether alcohol affected anxious behaviors was performed by utilizing open field activity and elevated plus maze. There was no difference in genotypes or sex in open field activity but there was a strong dose effect with a genotype difference in the 1 g/kg dose in males, showing the Atp1a2 heterozygous were less anxious with alcohol exposure. This was confirmed after results from elevated plus maze showed a decrease in anxiousness in female Atp1a2 heterozygous mice measured by an increase in the amount of time spent in the open arms. All behavioral testing results are summarized in

Table 4.1.

The behavioral testing in the Atp1a2 heterozygous mice showed interesting interactions between sex, genotype, and dose and with further analysis it was determined how these interactions were driven. There were strong sex effects with sleeptime in LORR and the amount of alcohol consumed in DID, it has been previously published that female mice consume higher doses of alcohol and have been shown to have increased BECs compared to males (Cozzoli et al. 2014). There were genotype effects with sleeptime in the females between saline pre-treatment and in AFT between the ethanol pre-treatment, heterozygous males significantly drank more alcohol, in OFA we saw a genotype effect in the males with the 1 g/kg dose, and with EPM we also saw a genotype effect in the females with the 1 g/kg dose. Dose effects were seen in all the

OFA parameters measured and there was a trend towards a dose effect in EPM.

Table 4.1. Overall Summary of Behavioral Testing in the Atp1a2 Heterozygous Mice. The various behavioral tests, the parameters measured, and results compared to wild- type. ↑ = increased, ↓ = decreased, ↔ = the same Heterozygous Heterozygous Phenotype Parameters Test compared to compared to Measured Measured Wild-type Wild-type Alcohol Metabolism BEC Metabolism

Alcohol LORR Sleeptime Sensitivity

Rapid Sleeptime Tolerance

AFT BEC

HIC Withdrawal Convulsions

Voluntary Alcohol DID Drinking Consumed

Total Open Field Anxiety Distance Activity Travelled

Time Spent in Corners Time Spent in Center Time Spent in Borders Rearings

Elevated % Time in Anxiety Plus Maze open arms

% Entries into open

arms 72

Interestingly, the genotype effect that has been previously published with open field activity and elevated plus maze (Moseley et al. 2007; Ikeda et al. 2003) was not observed. As discussed in Chapter III, the lack of phenotype could be due to several factors: 1) all our mice were weighed and injected, this handling could have affected their behavior, whereas the mice used by Ikeda et al were completely naive, 2) the mice used in the study by Moseley et al (2007) were previously tested in the elevated zero maze prior to testing in the open field, 3) there could be a genotype-environment interaction, with the two labs a mile apart in elevation. The Atp1a2 heterozygous mice were slightly more anxious compared to the wild-type, but this difference was not statistically significant. Both males and females were behaviorally tested, whereas the other groups only tested males. Some of the phenotypes measured were only significantly seen in one of the sexes, such as males with DID and females with sleeptime and AFT, however it is possible that with larger numbers of animals tested, significance could appear in both sexes.

Although the behavioral results were mixed, there was clearly a connection between anxiety and alcohol. It has been well published that alcohol exposure decreases anxiety (Acevedo et al. 2014) and the 1 g/kg dose in OFA resulted in decreased levels of anxiety measured by locomotor activity in the heterozygous males.

DID also showed male heterozygous mice voluntarily drank more alcohol than wild- types, but the interaction with the anxiety caused by Atp1a2 is still unknown.

The variations present in Atp1a2 were shown to be associated with increased anxiety in mice, but in humans, mutations in Atp1a2 cause increased migraines. It has been very difficult to determine whether mice suffer from headaches or migraines using behavioral testing, however recently there is evidence of a connection between migraines and deficits in learning and memory (Vanmolkot, Stroink, et al. 2006). The

only study that has studied the human mutation in Atp1a2 in mice was done by a group at the Vita-Salute San Raffaele University and Center for Translational Genomics and

Bioinformatics in Italy (Leo et al. 2011) and they discovered that the knock-in mice suffer from increased CSD leading to activation of the trigeminovascular system, which is responsible for the headache associated with migraines (de Vries et al. 2009). The two groups that created the Atp1a2 heterozygous mice did examine them for learning and memory deficits, but they interpreted them differently. The Lingrel lab examined spatial learning and memory using the Morris water maze and found the heterozygous mouse to have increased latency finding the hidden platform, suggesting they had impaired learning (Moseley et al. 2007) whereas the Kawakami lab found the heterozygous mouse had increased freezing time in the conditioned fear test, which is generally interpreted as improved learning and memory (Ikeda et al. 2003). Based on these learning and memory results, it is unable to determine whether the Atp1a2 heterozygous suffer from migraines without further investigation.

Currently, there are no studies on alcohol use and patients suffering from FHM2 or in Atp1a2 knock-out mice. In human studies, those who suffer from migraines have been shown to have decreased alcohol consumption compared to non-migraine sufferers due to the belief that symptoms will worsen (Panconesi 2008), but results on this subject tend to be varied with no definitive proof that alcohol consumption results in an increased number of headaches or migraines (Rist, Buring, and Kurth 2014). While this study demonstrated a connection between the anxiety caused by Atp1a2 variation and alcohol behaviors, the Atp1a2 heterozygous mice were not examined for learning and memory deficits and if alcohol exposure further affected behavior. Understanding the relationship between migraines and alcohol use is an important field of study and needs to be further examined.

In summary, Atp1a2 was differentially expressed in the LXS RIs, expression was not controlled by its own gene, and that glutamate transporter expression was affected by Atp1a2 expression. Atp1a2 mRNA correlated strongly to variations in behavior in the

LXS and BXDs, however this was not observed with differential protein expression.

Atp1a2 heterozygous mice showed increased sensitivity, increased rapid tolerance, increased acute functional tolerance, and voluntarily drank more alcohol; however there was no difference in alcohol metabolism or withdrawal severity. The Atp1a2 heterozygous mice have decreased anxiety with alcohol exposure, which could be due to decreased glutamate release and therefore decreased extracellular glutamate. There were interactions between genotype, sex, and alcohol treatment, suggesting that Atp1a2 slightly contributes to alcohol related behaviors. There were differences between genotypes in the voluntary drinking study with the heterozygous mice consuming more alcohol and the anxiety models studied with the attenuation of anxiety after alcohol exposure.

Future Directions

Although the experiments contained in this dissertation determined that genetic variation of Atp1a2 affected alcohol related behaviors, further investigation is certainly warranted. Behavioral effects due to the Atp1a2 knock-out were seen, however they were inconsistent. Additional studies with varying alcohol doses, increased treatment durations, and further investigation in the effect of the mixed background strain are needed.

There is little known about other modifications in the genome in the Atp1a2 heterozygous mice, therefore future studies should include determining what other

genes are modified because of the Atp1a2 knock-out are needed. This could be accomplished by Next Generation sequencing of the heterozygous and wild-type mice and examining for differential sequences that may result in protein modifications, interactions, and activities of other proteins are affected by the Atp1a2 knock-out.

In Chapter II it was shown that Atp1a2 was differentially expressed but expression is not controlled by its own gene and with the presence of a suggestive QTL on chromosome 19, there were genes that could be regulating Atp1a2 expression. To test the possibility that one of these genes are controlling Atp1a2 expression, the candidate genes could be individually knocked-out and then measured for Atp1a2 expression. Atp1a2 also correlated to various phenotypes other than anxiety and alcohol related behaviors, so further investigation into the phenotypes associated with differential expression of Atp1a2 are needed in more mouse strains. Similarly, a phenotype that did significantly correlate with differential expression of Atp1a2 was not seen.

In Chapter III there was no overall increase in seizures in the heterozygous mice during acute alcohol withdrawal, perhaps with chronic alcohol exposure an increase in seizure severity in the heterozygous mice would be seen. Some of the error bars were large, even with ten to twenty animals, so increasing the number of animals tested might show more significant results. The inconsistencies measured in our behavioral tests could also be explained by the mixed background of our mice. Further investigation into how the backgrounds affect behavior needs to be examined. The two groups who created the heterozygous mice interpreted the learning and memory deficits differently; therefore additional behavioral testing needs to be done, as well as examining whether alcohol exposure affects these behaviors. Finally, further investigation into the cause of the increased anxiety found in the Atp1a2 heterozygous mice needs to be performed.

There was slight down-regulation of glutamate transporters in the heterozygous mice, but there could be other genes and proteins that result in the altered phenotype, more notably GABA and serotonin. This could be accomplished by measuring neurotransmitter levels in the brain and, as mentioned earlier, through examination of sequence or protein modifications.

Overall, additional examination into the effects of variation in Atp1a2 in mice, either through knock-outs or knock-ins with mutations associated with FHM2, are needed to better understand Atp1a2 with regards to FHM2, as a potential risk factor for the development of alcoholism due to the increased anxiety, and possible treatment options for both diseases.

REFERENCES

Acevedo, María Belén, Michael E. Nizhnikov, Juan C. Molina, and Ricardo Marcos Pautassi. 2014. “Relationship between Ethanol-Induced Activity and Anxiolysis in the Open Field, Elevated plus Maze, Light-Dark Box, and Ethanol Intake in Adolescent Rats.” Behavioural Brain Research 265: 203–15. doi:10.1016/j.bbr.2014.02.032.

Altschul, Stephen F., Thomas L. Madden, Alejandro A. Schäffer, Jinghui Zhang, Zheng Zhang, Webb Miller, and David J. Lipman. 1997. “Gapped BLAST and PSI-BLAST: A New Generation of Protein Database Search Programs.” Nucleic Acids Research 25: 3389–3402.

Ambrosini, A, M D’Onofrio, G S Grieco, A Di Mambro, G Montagna, D Fortini, F Nicoletti, et al. 2005. “Familial Basilar Migraine Associated with a New Mutation in the ATP1A2 Gene.” Neurology 65 (11): 1826–28. doi:10.1212/01.wnl.0000187072.71931.c0.

Association, American Psychiatric. 2013. Diagnostic and Statistical Manual of Mental Disorders: DSM-5. Washington, D.C.: American Psychiatric Association. http://dsm.psychiatryonline.org/doi/book/10.1176/appi.books.9780890425596. Access Date: April 1, 2015

Barkley-Levenson, Amanda M, and John C Crabbe. 2012. “Bridging Animal and Human Models: Translating From (and To) Animal Genetics.” Alcohol Research : Current Reviews 34 (3): 325–35.

Bassi, M T, N Bresolin, a Tonelli, K Nazos, F Crippa, C Baschirotto, C Zucca, et al. 2004. “A Novel Mutation in the ATP1A2 Gene Causes Alternating Hemiplegia of Childhood.” Journal of Medical Genetics 41 (8): 621–28. doi:10.1136/jmg.2003.017863.

Bennett, B, C Larson, PA Richmond, AT Odell, LM Saba, B Tabakoff, R Dowell, and RA Radcliffe. 2015. “QTL Mapping of Acute Functional Tolerance in the LXS Recombinant Inbred Strains.” Alcoholism: Clinical and Experimental Research in press.

Bhave, Sanjiv V, Cheryl Hornbaker, Tzu L Phang, Laura Saba, Razvan Lapadat, Katherina Kechris, Jeanette Gaydos, et al. 2007. “The PhenoGen Informatics Website: Tools for Analyses of Complex Traits.” BMC Genetics 8: 59. doi:10.1186/1471-2156-8-59.

Botta, Paolo, Fabio M Simões de Souza, Thomas Sangrey, Erik De Schutter, and C Fernando Valenzuela. 2010. “Alcohol Excites Cerebellar Golgi Cells by Inhibiting the Na+/K+ ATPase.” Neuropsychopharmacology : Official Publication of the American College of Neuropsychopharmacology 35: 1984–96. doi:10.1038/npp.2010.76. 78

Buckner, Julia D., and R. Jay Turner. 2009. “Social Anxiety Disorder as a Risk Factor for Alcohol Use Disorders: A Prospective Examination of Parental and Peer Influences.” Drug and Alcohol Dependence 100: 128–37. doi:10.1016/j.drugalcdep.2008.09.018.

Cholet, N, L Pellerin, P J Magistretti, and E Hamel. 2002. “Similar Perisynaptic Glial Localization for the Na+,K+-ATPase Alpha 2 Subunit and the Glutamate Transporters GLAST and GLT-1 in the Rat Somatosensory Cortex.” Cerebral Cortex (New York, N.Y. : 1991) 12: 515–25. doi:10.1093/cercor/12.5.515.

Cozzoli, DK, MA Tanchuck-Nipper, MN Kaufman, CB Horowitz, and DA Finn. 2014. “Environmental Stressors Influence Limited-Access Ethanol Consumption by C57BL/6J Mice in a Sex-Dependent Manner.” Alcohol 48 (8): 741–54. doi:10.1016/j.alcohol.2014.07.015.

Crabbe, JC. 1989. “Genetic Animal Models in the Study of Alcoholism.” Alcoholism: Clinical and Experimental Research 13 (1): 120–27.

Crabbe, JC. 2014. “Use of Animal Models of Alcohol-Related Behavior.” Handbook of Clinical Neurology 125: 71–86.

Crabbe, JC, N A Johnson, D K Gray, A Kosobud, and E R Young. 1982. “Biphasic Effects of Ethanol on Open-Field Activity: Sensitivity and Tolerance in C57BL/6N and DBA/2N Mice.” Journal of Comparitive and Physiological Psychology 96 (3): 440–51.

Crabbe, JC, C D Merrill, and J K Belknap. 1991. “Effects of Convulsants on Handling- Induced Convulsions in Mice Selected for Ethanol Withdrawal Severity.” Brain Research 550 (1): 1–6.

Crambert, G, U Hasler, A T Beggah, C Yu, N N Modyanov, J D Horisberger, L Lelièvre, and K Geering. 2000. “Transport and Pharmacological Properties of Nine Different Human Na, K-ATPase Isozymes.” The Journal of Biological Chemistry 275 (3): 1976–86.

Crambert, G, Daniele Schaer, Sophie Roy, and Käthi Geering. 2004. “New Molecular Determinants Controlling the Accessibility of Ouabain to Its Binding Site in Human Na,K-ATPase Alpha Isoforms.” Molecular Pharmacology 65 (2): 335–41. doi:10.1124/mol.65.2.335.

Danbolt, Niels C. 2001. “Glutamate Uptake.” Progress in Neurobiology 65: 1–105. doi:10.1016/S0301-0082(00)00067-8.

De Fusco, Maurizio, Roberto Marconi, Laura Silvestri, Luigia Atorino, Luca Rampoldi, Letterio Morgante, Andrea Ballabio, Paolo Aridon, and Giorgio Casari. 2003. “Haploinsufficiency of ATP1A2 Encoding the Na+/K+ Pump alpha2 Subunit Associated with Familial Hemiplegic Migraine Type 2.” Nature Genetics 33 (2): 192–96. doi:10.1038/ng1081.

De Vries, Boukje, Rune R Frants, Michel D Ferrari, and Arn M J M van den Maagdenberg. 2009. “Molecular Genetics of Migraine.” Human Genetics 126 (1): 115–32. doi:10.1007/s00439-009-0684-z.

Deprez, Liesbet, Sarah Weckhuysen, Katelijne Peeters, Tine Deconinck, Kristl G Claeys, Lieve R F Claes, Arvid Suls, et al. 2008. “Epilepsy as Part of the Phenotype Associated with ATP1A2 Mutations.” Epilepsia 49 (3): 500–508. doi:10.1111/j.1528- 1167.2007.01415.x.

Dick, Danielle M, and Tatiana Foroud. 2002. “Genetic Strategies to Detect Genes Involved in Alcoholism and Alcohol-Related Traits.” Alcohol Research & Health : The Journal of the National Institute on Alcohol Abuse and Alcoholism 26 (3): 172– 80.

Ducci, Francesca, and David Goldman. 2008. “Genetic Approaches to Addiction: Genes and Alcohol.” Addiction. doi:10.1111/j.1360-0443.2008.02203.x.

Ehlers, Cindy L., Nicole A R Walter, Danielle M. Dick, Kari J. Buck, and John C. Crabbe. 2010. “A Comparison of Selected Quantitative Trait Loci Associated with Alcohol Use Phenotypes in Humans and Mouse Models.” Addiction Biology. doi:10.1111/j.1369-1600.2009.00195.x.

Geering, K, a Beggah, P Good, S Girardet, S Roy, D Schaer, and P Jaunin. 1996. “Oligomerization and Maturation of Na,K-ATPase: Functional Interaction of the Cytoplasmic NH2 Terminus of the Beta Subunit with the Alpha Subunit.” The Journal of Cell Biology 133 (6): 1193–1204.

George, D T, D J Nutt, B A Dwyer, and M Linnoila. 1990. “Alcoholism and Panic Disorder: Is the Comorbidity More than Coincidence?” Acta Psychiatrica Scandinavica 81 (2): 97–107. doi:10.1111/j.1600-0447.1990.tb06460.x.

Goymer, Patrick. 2007. “Synonymous Mutations Break Their Silence.” Nature Reviews Genetics. doi:10.1038/nrg2056.

Grant, Jon E. 2014. “Clinical Practice: Obsessive-Compulsive Disorder.” N Engl J Med. doi:10.1056/NEJMcp1402176.

Griffiths, AJF, JH Miller, and DT Suzuki. 2000. An Introduction to Genetic Analysis. Edited by W.H. Freeman. 7th Editio. New York. http://www.ncbi.nlm.nih.gov/books/NBK22007/. Access Date: April 1, 2015.

Gritz, Stephanie M, and Richard A Radcliffe. 2013. “Genetic Effects of ATP1A2 in Familial Hemiplegic Migraine Type II and Animal Models.” Human Genomics 7: 8. doi:10.1186/1479-7364-7-8.

Harvey, Brian H., and Mohammed Shahid. 2012. “Metabotropic and Ionotropic Glutamate Receptors as Neurobiological Targets in Anxiety and Stress-Related Disorders: Focus on Pharmacology and Preclinical Translational Models.” Pharmacology Biochemistry and Behavior. doi:10.1016/j.pbb.2011.06.014. 80

He, Suiwen, Daniel A Shelly, Amy E Moseley, Paul F James, J Howard James, Richard J Paul, and Jerry B Lingrel. 2001. “The alpha(1)- and alpha(2)-Isoforms of Na-K- ATPase Play Different Roles in Skeletal Muscle Contractility.” American Journal Of Physiology 281 (3): R917–25.

Herr, Nathaniel R, John W Williams Jr, Sophiya Benjamin, and Jennifer Mcduffie. 2014. “Does This Patient Have Generalized Anxiety or Panic Disorder ? The Rational Clinical Examination Systematic Review.” JAMA : The Journal of the American Medical Association 312: 78–84. doi:10.1001/jama.2014.5950.

Ikeda, Keiko, Tatsushi Onaka, Makoto Yamakado, Junichi Nakai, Tomo-o Ishikawa, Makoto M Taketo, and Kiyoshi Kawakami. 2003. “Degeneration of the Amygdala/piriform Cortex and Enhanced Fear/anxiety Behaviors in Sodium Pump alpha2 Subunit (Atp1a2)-Deficient Mice.” The Journal of Neuroscience : The Official Journal of the Society for Neuroscience 23 (11): 4667–76.

Ikeda, Keiko, Hiroshi Onimaru, Junko Yamada, Koichi Inoue, Shinya Ueno, Tatsushi Onaka, Hiroki Toyoda, et al. 2004. “Malfunction of Respiratory-Related Neuronal Activity in Na+, K+-ATPase alpha2 Subunit-Deficient Mice Is Attributable to Abnormal Cl- Homeostasis in Brainstem Neurons.” The Journal of Neuroscience : The Official Journal of the Society for Neuroscience 24 (47): 10693–701. doi:10.1523/JNEUROSCI.2909-04.2004.

James, P F, I L Grupp, G Grupp, a L Woo, G R Askew, M L Croyle, R a Walsh, and J B Lingrel. 1999. “Identification of a Specific Role for the Na,K-ATPase Alpha 2 Isoform as a Regulator of Calcium in the Heart.” Molecular Cell 3 (5): 555–63.

Jen, Joanna C. 2001. “Familial Hemiplegic Migraine.” In GeneReviews (Internet), edited by RA Pagon, TD Bird, CR Dolan, K Stephens, and MP Adam. Seattle (WA): University of Washington, Seattle. http://www.ncbi.nlm.nih.gov/books/NBK1388/. Access Date: April 1, 2015

Jewell, EA, and JB Lingrel. 1991. “Comparison of the Substrate Dependence Properties of the Rat Na, K-ATPase Alpha 1, Alpha 2, and Alpha 3 Isoforms Expressed in HeLa Cells.” Journal of Biological Chemistry 266 (25). ASBMB: 16925.

Jurkat-Rott, K, T Freilinger, JP Dreier, and J Herzog. 2004. “Variability of Familial Hemiplegic Migraine with Novel A1A2 Na+/K+-ATPase Variants.” Neurology 62: 1857–61. doi:10.1212/01.WNL.0000127310.11526.FD.

Kaplan, Jack H. 2002. “Biochemistry of Na,K-ATPase.” Annual Review of Biochemistry 71 (January): 511–35. doi:10.1146/annurev.biochem.71.102201.141218.

Kaunisto, M a, H Harno, K R J Vanmolkot, J J Gargus, G Sun, E Hämäläinen, E Liukkonen, et al. 2004. “A Novel Missense ATP1A2 Mutation in a Finnish Family with Familial Hemiplegic Migraine Type 2.” Neurogenetics 5 (2): 141–46. doi:10.1007/s10048-004-0178-z.

Kawakami, Kiyoshi, Tatsushi Onaka, Michiko Iwase, Ikuo Homma, and Keiko Ikeda. 2005. “Hyperphagia and Obesity in Na,K-ATPase alpha2 Subunit-Defective Mice.” Obesity Research 13 (10): 1661–71. doi:10.1038/oby.2005.204.

Koenderink, Jan B, Giovanni Zifarelli, Li Yan Qiu, Wolfgang Schwarz, Jan Joep H H M De Pont, Ernst Bamberg, and Thomas Friedrich. 2005. “Na,K-ATPase Mutations in Familial Hemiplegic Migraine Lead to Functional Inactivation.” Biochimica et Biophysica Acta 1669 (1): 61–68. doi:10.1016/j.bbamem.2005.01.003.

Kushner, Matt G., Kenneth Abrams, and Carrie Borchardt. 2000. “The Relationship between Anxiety Disorders and Alcohol Use Disorders: A Review of Major Perspectives and Findings.” Clinical Psychology Review. doi:10.1016/S0272- 7358(99)00027-6.

Lader, Malcolm. 2011. “Benzodiazepines Revisited-Will We Ever Learn?” Addiction 106: 2086–2109. doi:10.1111/j.1360-0443.2011.03563.x.

Leo, Loredana, Lisa Gherardini, Virginia Barone, Maurizio De Fusco, Daniela Pietrobon, Tommaso Pizzorusso, and Giorgio Casari. 2011. “Increased Susceptibility to Cortical Spreading Depression in the Mouse Model of Familial Hemiplegic Migraine Type 2.” PLoS Genetics 7 (6): e1002129. doi:10.1371/journal.pgen.1002129.

Lingrel, J., A. Moseley, I. Dostanic, M. Cougnon, S. He, P. James, A. Woo, K. O’CONNOR, and J. Neumann. 2003. “Functional Roles of the Alpha Isoforms of the Na, K-ATPase.” Annals of the New York Academy of Sciences 986 (1). Wiley Online Library: 354–59.

Lingrel, J., Michael T Williams, Charles V Vorhees, and Amy E Moseley. 2007. “Na,K- ATPase and the Role of Alpha Isoforms in Behavior.” Journal of Bioenergetics and Biomembranes 39 (5-6): 385–89. doi:10.1007/s10863-007-9107-9.

Lundquist, F. 1959. “The Determination of Ethyl Alcohol in Blood and Tissue.” Methods of Biochemical Analysis 7: 217–51.

McClearn, G. E., and R Kakihana. 1973. “Selective Breeding for Ethanol Sensitivity in Mice.” Behavior Genetics 3: 409–10.

McClearn, G. E., and R. Kakihana. 1981. Selective Breeding for Ethanol Sensitivity: Short-Sleep and Long-Sleep Mice. Edited by G. E. McClearn, R. A. Deitrich, and V. G. Erwin. Development of Animal Models as Pharmacogenetic Tools. Washington, D.C.: U.S. Govt. Printing Office.

McGrail, K M, J M Phillips, and K J Sweadner. 1991. “Immunofluorescent Localization of Three Na,K-ATPase Isozymes in the Rat Central Nervous System: Both Neurons and Glia Can Express More than One Na,K-ATPase.” The Journal of Neuroscience : The Official Journal of the Society for Neuroscience 11 (2): 381–91.

Meldrum, B S. 2000. “Glutamate as a Neurotransmitter in the Brain: Review of Physiology and Pathology.” The Journal of Nutrition 130: 1007S – 15S. 82

Metten, P, and JC Crabbe. 1994. “Common Genetic Determinants of Severity of Acute Withdrawal from Ethanol, Pentobarbital and Diazepam in Inbred Mice.” Behavioural Pharmacology 5: 533–47.

Montagna, P. 2004. “The Physiopathology of Migraine: The Contribution of Genetics.” Neurological Sciences : Official Journal of the Italian Neurological Society and of the Italian Society of Clinical Neurophysiology 25 Suppl 3 (October): S93–96. doi:10.1007/s10072-004-0261-0.

Morth, J Preben, Bjørn P Pedersen, Mads S Toustrup-Jensen, Thomas L-M Sørensen, Janne Petersen, Jens Peter Andersen, Bente Vilsen, and Poul Nissen. 2007. “Crystal Structure of the Sodium-Potassium Pump.” Nature 450 (7172): 1043–49. doi:10.1038/nature06419.

Moseley, Amy E, Steve P Lieske, Randall K Wetzel, Paul F James, Suiwen He, Daniel a Shelly, Richard J Paul, et al. 2003. “The Na,K-ATPase Alpha 2 Isoform Is Expressed in Neurons, and Its Absence Disrupts Neuronal Activity in Newborn Mice.” The Journal of Biological Chemistry 278 (7): 5317–24. doi:10.1074/jbc.M211315200.

Moseley, Amy E, Michael T Williams, Tori L Schaefer, Cynthia S Bohanan, Jon C Neumann, Michael M Behbehani, Charles V Vorhees, and Jerry B Lingrel. 2007. “Deficiency in Na,K-ATPase Alpha Isoform Genes Alters Spatial Learning, Motor Activity, and Anxiety in Mice.” The Journal of Neuroscience : The Official Journal of the Society for Neuroscience 27 (3): 616–26. doi:10.1523/JNEUROSCI.4464- 06.2007.

Mozhui, Khyobeni, Daniel C. Ciobanu, Thomas Schikorski, Xusheng Wang, Lu Lu, and Robert W. Williams. 2008. “Dissection of a QTL Hotspot on Mouse Distal Chromosome 1 That Modulates Neurobehavioral Phenotypes and Gene Expression.” PLoS Genetics 4 (11). doi:10.1371/journal.pgen.1000260.

Mu, Wenbo, and Wei Zhang. 2012. “Bioinformatic Resources of microRNA Sequences, Gene Targets, and Genetic Variation.” Frontiers in Genetics 3 (March): 31. doi:10.3389/fgene.2012.00031.

Munzer, J S, S E Daly, E a Jewell-Motz, J B Lingrel, and R Blostein. 1994. “Tissue- and Isoform-Specific Kinetic Behavior of the Na,K-ATPase.” The Journal of Biological Chemistry 269 (24): 16668–76.

National Library of Medicine. 2013a. “Gene Database.” http://www.ncbi.nlm.nih.gov/gene. Access Date: April 1, 2015.

National Library of Medicine. 2013b. “Protein Database.” http://www.ncbi.nlm.nih.gov/protein/. Access Date: April 1, 2015.

NIAAA. 2006. “Alcohol Use and Alcohol Use Disorders in the United States: Main Findings from the 2001-2002 National Epidemiological Survey on Alcohol and Related Conditions (NESARC).” U.S. Alcohol Epidemiologic Data Reference Manual 8 (1).

NIAAA. 2013. “Alcohol Use Disorders.” http://www.niaaa.nih.gov/alcohol- health/overview-alcohol-consumption/alcohol-use-disorders. Access Date: April 1, 2015.

NIMH. 2009. “Anxiety Disorders.” http://www.nimh.nih.gov/health/topics/anxiety- disorders/index.shtml. Access Date: April 1, 2015.

Nuss, Philippe. 2015. “Anxiety Disorders and GABA Neurotransmission : A Disturbance of Modulation,” 165–75.

Olivier, Jocelien D A, Christiaan H. Vinkers, and Berend Olivier. 2013. “The Role of the Serotonergic and GABA System in Translational Approaches in Drug Discovery for Anxiety Disorders.” Frontiers in Pharmacology. doi:10.3389/fphar.2013.00074.

Olsen, J. 2004. “The International Classification of Headache Disorders (second Edition).” Cephalalgia 24 (Suppl 1): 1–160.

Onimaru, H., and I. Homma. 2007. “Spontaneous Oscillatory Burst Activity in the Piriform-Amygdala Region and Its Relation to in Vitro Respiratory Activity in Newborn Rats.” Neuroscience 144: 387–94. doi:10.1016/j.neuroscience.2006.09.033.

Orlowski, J, and J B Lingrel. 1988. “Tissue-Specific and Developmental Regulation of Rat Na,K-ATPase Catalytic Alpha Isoform and Beta Subunit mRNAs.” The Journal of Biological Chemistry 263 (21): 10436–42.

Palmer, Abraham A, Carrie S McKinnon, Hadley C Bergstrom, and Tamara J Phillips. 2002. “Locomotor Activity Responses to Ethanol, Other Alcohols, and GABA-A Acting Compounds in Forward- and Reverse-Selected FAST and SLOW Mouse Lines.” Behavioral Neuroscience 116 (6): 958–67. doi:10.1037/0735- 7044.116.6.958.

Panconesi, Alessandro. 2008. “Alcohol and Migraine: Trigger Factor, Consumption, Mechanisms. A Review.” Journal of Headache and Pain. doi:10.1007/s10194-008- 0006-1.

Pierelli, F, G S Grieco, F Pauri, C Pirro, G Fiermonte, A Ambrosini, A Costa, et al. 2006. “BRIEF REPORT A Novel ATP1A2 Mutation in a Family with FHM Type II.” Cephalalgia, no. 10: 324–28.

Pietrobon, Daniela. 2007. “Familial Hemiplegic Migraine.” Neurotherapeutics : The Journal of the American Society for Experimental NeuroTherapeutics 4 (2): 274–84. doi:10.1016/j.nurt.2007.01.008.

Pietrobon, Daniela, and Jörg Striessnig. 2003. “Neurobiology of Migraine.” Nature Reviews. Neuroscience 4 (5): 386–98. doi:10.1038/nrn1102.

Quitkin, F M, A Rifkin, J Kaplan, and D F Klein. 1972. “Phobic Anxiety Syndrome Complicated by Drug Dependence and Addiction. A Treatable Form of Drug Abuse.” Archives of General Psychiatry 27 (2): 159–62. doi:10.1001/archpsyc.1972.01750260013002.

Radcliffe, Richard A, Kirsten L Floyd, Joseph A Drahnak, and Richard A Deitrich. 2005. “Genetic Dissociation between Ethanol Sensitivity and Rapid Tolerance in Mouse and Rat Strains Selectively Bred for Differential Ethanol Sensitivity.” Alcoholism, Clinical and Experimental Research 29 (9): 1580–89.

Radcliffe, Richard A, Kirsten L Floyd, and Michael J Lee. 2006. “Rapid Ethanol Tolerance Mediated by Adaptations in Acute Tolerance in Inbred Mouse Strains.” Pharmacology, Biochemistry, and Behavior 84 (3): 524–34. doi:10.1016/j.pbb.2006.06.018.

Ramsden, Edmund. 2015. “MAKING ANIMALS ALCOHOLIC: SHIFTING LABORATORY MODELS OF ADDICTION.” Journal of the History of the Behavioral Sciences, n/a – n/a. doi:10.1002/jhbs.21715.

Rhodes, Justin S., Karyn Best, John K. Belknap, Deborah A. Finn, and John C. Crabbe. 2005. “Evaluation of a Simple Model of Ethanol Drinking to Intoxication in C57BL/6J Mice.” Physiology and Behavior 84: 53–63. doi:10.1016/j.physbeh.2004.10.007.

Riant, Florence, Maurizio De Fusco, Paolo Aridon, Anne Ducros, Claire Ploton, Florence Marchelli, Jacqueline Maciazek, Marie Germaine Bousser, Giorgio Casari, and Elisabeth Tournier-Lasserve. 2005. “ATP1A2 Mutations in 11 Families with Familial Hemiplegic Migraine.” Human Mutation 26 (3): 281. doi:10.1002/humu.9361.

Riaza Bermudo-Soriano, Carlos, M. Mercedes Perez-Rodriguez, Concepcion Vaquero- Lorenzo, and Enrique Baca-Garcia. 2012. “New Perspectives in Glutamate and Anxiety.” Pharmacology Biochemistry and Behavior. doi:10.1016/j.pbb.2011.04.010.

Rist, P. M., J. E. Buring, and T Kurth. 2014. “Dietary Patterns according to Headache and Migraine Status: A Cross-Sectional Study.” Cephalalgia. doi:10.1177/0333102414560634.

Rose, Erin M, Joseph C P Koo, Jordan E Antflick, Syed M Ahmed, Stephane Angers, and David R Hampson. 2009. “Glutamate Transporter Coupling to Na,K-ATPase.” The Journal of Neuroscience : The Official Journal of the Society for Neuroscience 29: 8143–55. doi:10.1523/JNEUROSCI.1081-09.2009.

Segall, L, S E Daly, and R Blostein. 2001. “Mechanistic Basis for Kinetic Differences between the Rat Alpha 1, Alpha 2, and Alpha 3 Isoforms of the Na,K-ATPase.” The Journal of Biological Chemistry 276 (34): 31535–41. doi:10.1074/jbc.M103720200.

Segall, L, Alessandra Mezzetti, Rosemarie Scanzano, J Jay Gargus, Enrico Purisima, and Rhoda Blostein. 2005. “Alterations in the alpha2 Isoform of Na,K-ATPase Associated with Familial Hemiplegic Migraine Type 2.” Proceedings of the National Academy of Sciences of the United States of America 102 (31): 11106–11. doi:10.1073/pnas.0504323102.

Segall, L, Rosemarie Scanzano, Mari a Kaunisto, Maija Wessman, Aarno Palotie, J Jay Gargus, and Rhoda Blostein. 2004. “Kinetic Alterations due to a Missense Mutation in the Na,K-ATPase alpha2 Subunit Cause Familial Hemiplegic Migraine Type 2.” The Journal of Biological Chemistry 279 (42): 43692–96. doi:10.1074/jbc.M407471200.

Spadaro, Maria, Simona Ursu, Frank Lehmann-Horn, Liana Veneziano, Veneziano Liana, Giovanni Antonini, Antonini Giovanni, et al. 2004. “A G301R Na+/K+ - ATPase Mutation Causes Familial Hemiplegic Migraine Type 2 with Cerebellar Signs.” Neurogenetics 5 (3): 177–85. doi:10.1007/s10048-004-0183-2.

Stein, Murray B., and Dan J. Stein. 2008. “Social Anxiety Disorder.” The Lancet. doi:10.1016/S0140-6736(08)60488-2.

Stinchcomb, A, B J Bowers, and J M Wehner. 1989. “The Effects of Ethanol and Ro 15- 4513 on Elevated plus-Maze and Rotarod Performance in Long-Sleep and Short- Sleep Mice.” Alcohol (Fayetteville, N.Y.) 6 (5): 369–76.

Therien, A G, S J Karlish, and R Blostein. 1999. “Expression and Functional Role of the Gamma Subunit of the Na, K-ATPase in Mammalian Cells.” The Journal of Biological Chemistry 274 (18): 12252–56.

Todt, Unda, Martin Dichgans, Karin Jurkat-Rott, Axel Heinze, Giovanni Zifarelli, Jan B Koenderink, Ingrid Goebel, et al. 2005. “Rare Missense Variants in ATP1A2 in Families with Clustering of Common Forms of Migraine.” Human Mutation 26 (4): 315–21. doi:10.1002/humu.20229.

Torpy, Janet M, Alison E Burke, and Robert M Golub. 2011. “JAMA Patient Page. Generalized Anxiety Disorder.” JAMA : The Journal of the American Medical Association. doi:10.1001/jama.305.5.522.

Valenzuela, C F. 1997. “Alcohol and Neurotransmitter Interactions.” Alcohol Health and Research World 21: 144–48.

Vanmolkot, K, Esther E Kors, Jouke-Jan Hottenga, Gisela M Terwindt, Joost Haan, Wil a J Hoefnagels, David F Black, et al. 2003. “Novel Mutations in the Na+, K+-ATPase Pump Gene ATP1A2 Associated with Familial Hemiplegic Migraine and Benign Familial Infantile Convulsions.” Annals of Neurology 54 (3): 360–66. doi:10.1002/ana.10674.

Vanmolkot, K, Esther E Kors, Ulku Turk, Dylsad Turkdogan, Antoine Keyser, Ludo a M Broos, Sima Kheradmand Kia, et al. 2006. “Two de Novo Mutations in the Na,K- ATPase Gene ATP1A2 Associated with Pure Familial Hemiplegic Migraine.” European Journal of Human Genetics : EJHG 14 (5): 555–60. doi:10.1038/sj.ejhg.5201607.

Vanmolkot, K, H Stroink, J B Koenderink, E E Kors, J J M W van den Heuvel, E H van den Boogerd, a H Stam, et al. 2006. “Severe Episodic Neurological Deficits and Permanent Mental Retardation in a Child with a Novel FHM2 ATP1A2 Mutation.” Annals of Neurology 59 (2): 310–14. doi:10.1002/ana.20760.

Whitfield, John B., Brian N. Nightingale, Martin E. O’Brien, Andrew C. Heath, Andrew J. Birley, and Nicholas G. Martin. 1998. “Molecular Biology of Alcohol Dependence, a Complex Polygenic Disorder.” In Clinical Chemistry and Laboratory Medicine, 36:633–36. doi:10.1515/CCLM.1998.111.

Williams, R W, J Gu, S Qi, and L Lu. 2001. “The Genetic Structure of Recombinant Inbred Mice: High-Resolution Consensus Maps for Complex Trait Analysis.” Genome Biology 2 (11): RESEARCH0046.

Williams, Robert W, Beth Bennett, Lu Lu, Jing Gu, John C DeFries, Phyllis J Carosone- Link, Brad a Rikke, John K Belknap, and Thomas E Johnson. 2004. “Genetic Structure of the LXS Panel of Recombinant Inbred Mouse Strains: A Powerful Resource for Complex Trait Analysis.” Mammalian Genome : Official Journal of the International Mammalian Genome Society 15 (8): 637–47. doi:10.1007/s00335-004- 2380-6.

Woo, A L, P F James, and J B Lingrel. 1999. “Membrane Biology Characterization of the Fourth Alpha Isoform of the Na , K-ATPase.” Journal of Membrane Biology, no. 169: 39–44.

Zhang, Jin, Moo Yeol Lee, Maurizio Cavalli, Ling Chen, Roberto Berra-Romani, C William Balke, Giuseppe Bianchi, et al. 2005. “Sodium Pump alpha2 Subunits Control Myogenic Tone and Blood Pressure in Mice.” The Journal of Physiology 569 (Pt 1): 243–56. doi:10.1113/jphysiol.2005.091801.