Genetic Factors Regulating Expression of Dopaminergic Genes

DISSERTATION

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

Elizabeth Stofko Barrie

Graduate Program in Integrated Biomedical Science Program

The Ohio State University

2014

Dissertation Committee:

Professor Wolfgang Sadee, Advisor

Professor R. Thomas Boyd

Professor Howard Gu

Professor Amanda Toland

Elizabeth Stofko Barrie

2014

Abstract

Genetic differences are one of the main contributors to diversity in clinical phenotypes.

This project identifies and defines genetic factors affecting dopamine dysregulation focusing on three genes in dopamine signaling: DBH, COMT and SNCA. To study this I measure allelic expression, selecting RNA samples from post-mortem human tissue that are heterozygous for a marker SNP, and quantitate the expression of each allele.

Dopamine β-hydroxylase (DBH) encodes an enzyme which converts dopamine to norepinephrine. A promoter SNP, rs1611115 has been associated with low DBH and high dopamine plasma levels; however, underlying mechanisms remain uncertain. I found a tissue-specific effect of rs1611115 in liver, with up to 11 fold differences in allelic and overall mRNA expression (p<0.0004 to 2x10-7 and p<0.0001 respectively), indicating decreased transcription. Interestingly, locus coeruleus and adrenal gland, the main sources of DBH in the body, did not demonstrate this robust effect; only small AEI ratios were detected in these tissues. More frequent than rs1611115 and in linkage disequilibrium with it, a second SNP, rs1108580 was associated with reduced allelic mRNA expression in all tissues tested. This dual mechanism accounts for the previously described genotype effect on DBH plasma levels, with a novel role for liver as an important source of variability in DBH levels. In combination, rs161115 and rs1108580 contribute to strongly reduced mRNA expression in the liver, reducing transcription in a ii

tissue selective manner. In mice, Dbh mRNA levels in the liver correlated with

cardiovascular risk phenotypes. Using a PheWAS (phenome-wide association study) analysis, the minor alleles of rs1611115 and rs1108580 were associated with sympathetic phenotypes including angina pectoris. Testing the combined effects of rs1611115 and rs1108580 indicated robust protection against myocardial infarction in three clinical cohorts. These results demonstrate profound effects of common DBH variants on expression in sympathetically innervated organs, modulating clinical phenotypes responsive to peripheral sympathetic tone. In a pathway parallel to DBH, catechol-O-

methyltransferase (COMT) converts dopamine to an inactive metabolite. Almost half of

the African-American samples tested demonstrate a significant mRNA fold change, while

only one Caucasian sample demonstrates AEI, indicating the presence of a regulatory

variant which has not yet been described. Extensive sequencing of regions predicted to

harbor this regulatory variant, up to 500 kb from the gene, did not reveal a variant

associated with these instances of AEI. There is a factor, acting predominantly in

African-Americans, regulating allelic expression and I have ruled out hundreds of SNPs

as being the cause. Alpha-synuclein (SNCA) is involved in dopamine regulation, and

implicated in degeneration of dopaminergic neurons. Allelic mRNA analysis indicates

the presence of a regulatory variant. I have found an association with rs17016074, likely

affecting, or marking differential 3′UTR usage. This approach has revealed the presence of frequent regulatory variants in all three genes studied. Functional SNPs contributing to dysregulation of dopamine can be tested for association with clinical phenotypes using large publicly available genome-wide association datasets.

iii

Dedication

This document is dedicated to my family. Thanks to my mother for her many years of

editing, my father for moral support, my brother for his sense of humor and my in-laws for being my home away from home. I’d especially like to thank my husband Mike for being my rock; his encouragement and patience have been vital to my success. Finally,

I’m excited to meet our daughter who has helped me keep everything in perspective.

Acknowledgments

I am grateful to my advisor, Dr. Wolfgang Sadee, for helping me develop and define my research goals while maintaining a forward thinking view on my work and the field. He has challenged me to think critically about my research questions and work through issues that arise. He has undoubtedly shaped my graduate experience for the better and prepared me for a rewarding future scientific career.

I appreciate the time and guidance from my committee members: Drs. R. Thomas Boyd,

Howard Gu, and Amanda Toland. I am thankful to our laboratory manager, Audrey Papp as well as past and present members of the Sadee laboratory for their advice and helpful discussions: Amanda Curtis, Diane Delobel, Katherine Hartmann, Hannah Komar,

Sebastiano Porcu, Jonathan Sanford, Gloria Smith, Adam Suhy, Danielle Sullivan, and

Drs. Sam Handelman, Robert Moyer, Leslie Newman, Julia Pinsonneault, Ryan Smith,

Danxin Wang, and Amy Webb. Pharmacology Department staff members Sherry Ring and Gina Pace were instrumental in managing paperwork and helping navigate the system. Collaborators from outside institutions, Drs. Deborah Mash, Sarah Pendergrass and David Weinshenker, also made vital contributions to this work.

Vita

2004...... Amherst Central High School

2008...... B.S. Biology, Case Western Reserve

University

2008 to present ...... Graduate Research Associate, Department

of Pharmacology, The Ohio State University

Publications

Barrie ES, Smith RM, Sanford JC, and Sadee W. (2012) mRNA Transcript Diversity

Creates New Opportunities for Pharmacological Intervention, Molecular Pharmacology

Barrie ES, Weinshenker D, Pendergrass S, Lange L, Ritchie M, Wilson J, Kuivaniemi H,

Tromp G, Carey D, Gerhard G, Cubells J Sadee W. Regulatory polymorphisms in DBH affect peripheral gene expression and sympathetic phenotypes. Circulation Research. In

Revision.

Barrie ES, Lodder M, Weinreb P, Buss J, Rajab A, Adin C, Mi QS, Hadley GA. Role of

CD103 in the development of autoimmune diabetes in NOD mice. Journal of

Endocrinology. In Revision.

Fields of Study

Major Field: Integrated Biomedical Science Program

vii

Table of Contents

Abstract ...... ii

Dedication ...... iv

Acknowledgments...... v

Vita ...... vi

Publications ...... vi

Fields of Study ...... vii

Table of Contents ...... viii

List of Tables ...... xi

List of Figures ...... xiii

Chapter 1: Introduction ...... 1

1.1 Evolution of the Field of Genetics ...... 1

1.2 Mechanisms of Regulation and Evidence for cis-acting Polymorphisms ...... 3

1.3 Detecting Sequence Variation ...... 5

1.4 Targets of Interest: Dopamine and Norepinephrine ...... 6

1.5 Methods Common to All Projects ...... 7

viii

1.6 Summary of Studies ...... 10

Chapter 2: Regulatory Genetics of Dopamine β-Hydroxylase (DBH) and Effect on

Clinical Phenotypes ...... 11

2.1 Introduction and Background ...... 11

2.2 Known DBH SNPs and Clinical Correlations...... 12

2.3 Myocardial Infarction and Coronary Heart Disease ...... 16

2.4 Study Overview ...... 18

2.5 Materials and Methods ...... 19

2.6 Results ...... 27

2.7 Discussion ...... 72

Chapter 3: Regulation in the 3′ Untranslated Region of Alpha-synuclein (SNCA) ...... 79

3.1 Introduction and Background ...... 79

3.2 Materials and Methods ...... 82

3.3 Results ...... 83

3.4 Discussion ...... 99

Chapter 4: Population-Specific Regulation of Catechol-O-Methyltransferase (COMT) 102

4.1 Introduction and Background ...... 102

4.2 Materials and Methods ...... 105

4.3 Results ...... 105

4.4 Discussion ...... 112

Chapter 5: Conclusions and Discussion ...... 114

References ...... 119

Appendix A: Funding Sources ...... 134

Appendix B: Abbreviations ...... 135

Appendix C: Tables ...... 139

List of Tables

Table 1. SNPs genotyped and genotype association test using the F statistic ...... 40

Table 2. LD values (D′ and R2) for all SNPs genotyped in liver ...... 41

Table 3. Minor allele frequencies and LD values (D′ and R2) ...... 43

Table 4. Demographics for liver samples used in qRT-PCR ...... 47

Table 5. Correlation between phenotype and Dbh liver mRNA expression ...... 53

Table 6. PheWAS analysis of individual SNPs, using an additive model in the Geisinger

Study ...... 57

Table 7. PheWAS analysis using an interaction model, in the Geisinger study ...... 59

Table 8. Genotype association tests with an additive model in JHS ...... 62

Table 9. SNPs genotyped and genotype association with AEI using the F statistic ...... 86

Table 10. Chi squared p value for genotype association with instances of AEI ...... 93

Table 11. LD of SNPs from GWAS Search and rs17016074 ...... 94

Table 12. rs17016074 association with cocaine or ED status ...... 95

Table 13. DBH primers for DNA and RNA amplification efficiency comparison for rs1108580 and rs77905 ...... 139

Table 14. DBH SNaPshot primers ...... 140

Table 15. DBH allele-specific melting curve (GC clamp) genotyping primers ...... 141

Table 16. Primers to amplify DBH gene region for Ion Torrent sequencing ...... 142 xi

Table 17. DBH primers for fluorescent restriction fragment length polymorphism genotyping...... 143

Table 18. Additional primers for DBH experiments ...... 144

Table 19. SNCA primers ...... 145

Table 20. COMT region amplification primers for Ion Torrent sequencing ...... 146

Table 21. COMT primers ...... 149

xii

List of Figures

Figure 1. Allelic expression imbalance assay ...... 9

Figure 2. DBH gene map ...... 11

Figure 3. Example electrocardiogram reading ...... 18

Figure 4. DBH mRNA expression ...... 28

Figure 5. Efficiency comparison of primers for PCR amplification of DNA and RNA

around rs1108580 and rs77905 ...... 31

Figure 6. Allelic mRNA expression ratios in LC or pons measured at rs129882, rs77905,

rs5320 or rs1108580 ...... 33

Figure 7. Allelic mRNA expression ratios measured at rs129882 in liver, LC, or pons. 34

Figure 8. Allelic DBH mRNA expression ratios measured at rs1108580 and rs77905 .. 36

Figure 9. Linear regression of AEI magnitude at two marker SNPs ...... 37

Figure 10. DBH mRNA levels in liver, lung and brain grouped by genotype ...... 46

Figure 11. DBH mRNA expression in livers measured with qRT-PCR grouped by number of minor alleles of rs1108580 and rs1611115 ...... 48

Figure 12. Alignment of liver cDNA with PAH and DBH reference sequence ...... 50

Figure 13. Electrocardiogram readings measured across 14 strains of mice ...... 52

Figure 14. Blood pressure change in response to atenolol treatment ...... 64

Figure 15. Association of rs1611115 genotype with blood glucose response ...... 65 xiii

Figure 16. Linear regression between total number of minor alleles (rs1108580 and

rs1611115) and mean total recall score ...... 67

Figure 17. Sanger sequencing results in chimpanzee gDNA ...... 68

Figure 18. DBH mRNA expression levels in transgenic mice ...... 71

Figure 19. Allelic SNCA mRNA expression ratios measured at marker rs356165...... 84

Figure 20. SNCA mRNA levels grouped by rs17016074 genotype ...... 87

Figure 21. Map of marker SNPs, primer locations and polyA consensus sequences in the

3′UTR of SNCA ...... 87

Figure 22. Allelic SNCA mRNA expression ratios measured at marker rs17016074..... 89

Figure 23. Allelic SNCA mRNA expression ratios measured at marker rs356165 and/or

rs17016074 in PCR product containing both SNPs...... 91

Figure 24. SNCAIP mRNA levels in brain grouped by rs11326 genotype ...... 96

Figure 25. Allelic mRNA expression imbalance for rs11326...... 98

Figure 26. Allelic expression ratios measured at marker SNP rs165599 ...... 107

Figure 27. Allelic expression measured at rs4633 in prefrontal cortex ...... 109

Figure 28. Allelic expression measured at rs4633 in liver ...... 110

xiv

Chapter 1: Introduction

1.1 Evolution of the Field of Genetics

For thousands of years, humans noticed the consequences of heritability and used their observations to breed plants and animals with favorable traits. However, the study of the mechanisms by which these differences are inherited is a relatively recent development, with many unanswered questions.

Gregor Mendel, credited with being the father of genetics, presented his findings in 1865. Through a series of crosses in pea plants he was able to identify dominant and recessive patterns of inheritance by examining traits such as green or yellow peas or smooth or wrinkled ones. While Mendel’s experiments provided the basis for modern genetics, most phenotypes are not as simple as one gene determining one characteristic, nor does one allele always act in a dominant manner. There are also traits, such as height, that manifest over a broad range and cannot be classified into two defined groups.

Many traits are polygenic, or affected by multiple genes. We also recognize the presence of epistasis, where multiple genes have a combined effect on a phenotype.

Genetic information is encoded by deoxyribonucleic acid (DNA), enabling its propagation through multiple generations. This building block for life is composed of four nucleotides: adenine, guanine, cytosine and tyrosine arranged in a double helical

structure, identified in 1953 (Watson et al.). DNA, organized into chromosomes, is

located in the nucleus and divides and replicates as part of the cell cycle. DNA is then

transcribed into ribonucleic acid (RNA). Next the RNA is processed, and transported out of the cell. Using the code from the RNA, proteins are assembled from amino acids during translation, although non-coding RNA species also exist.

Improved visualization of metaphase chromosomes stimulated advances in the field of cytogenetics. Once it was established in 1956 that diploid cells contained 46 chromosomes, (Tijo et al.) reports of aneuploidy (abnormal numbers of chromosomes) surfaced. Several syndromes were described at this time including trisomy 13, trisomy

18, Turner syndrome, Klinefelter syndrome and triple X. Karyotype analysis also

allowed detection of deletions or additions of large regions of the chromosome, such as

Cri du Chat, involving a structural loss of the top of the p arm of chromosome 5.

Technology advanced in the 1980s with the introduction of fluorescent in situ

hybridization (FISH), which made identification of increasingly smaller differences in the

DNA possible. The invention of polymerase chain reaction (PCR) allowed amplification

of targeted regions. Huntington’s disease was the first disease to be mapped to specific

location in the genome in 1983. We now recognize even single base changes can have

effects on phenotypes. Differences in the DNA sequence can manifest as single

nucleotide polymorphisms (SNPs), copy number variation (CNV), repeats, insertions or

deletions.

In order to characterize these genetic differences and ultimately determine their

impact, the Human Genome Project (HGP) was initiated in 1990. The goal was to

sequence and map the entire human genome—an enormous undertaking at the time. The project was completed in 2003, with about 99% of the genome covered. The next step in large scale genetic analysis was the International HapMap Project. This analyzed DNA from different population groups and compared blocks of SNPs, or haplotypes, between

cases and controls to determine whether certain haplotypes were more prevalent in a certain disease state.

Upon completion of the HGP, it was discovered that the majority of the genome consisted of intronic and other non-coding sequences. These were believed to be “junk

DNA,” serving no purpose. Since then, the true complexity of the human genome, as

well as the importance of sequences previously believed to be nonfunctional have

become apparent. A variety of RNA species, that are not translated into proteins have

been identified including microRNA, lncRNA (long non-coding), siRNA (small

interfering), snoRNA (small nucleolar), snRNA (small nuclear), exRNA (extracellular)

and piRNA (piwi-interacting).

The human genome consists of ~20,000 protein-coding genes, but it is now

known that the total repertoire of mRNA sequences and encoded proteins is far greater

due to multiple RNA isoforms generated from each gene.

1.2 Mechanisms of Regulation and Evidence for cis-acting Polymorphisms

SNPs can result in modification of the amino acid sequence, known as non- synonymous variants. However, many heritable polymorphisms have the capacity to regulate gene expression without altering the coding sequence and are more prevalent

than the non-synonymous SNPs. Large-scale studies of regulatory polymorphisms affecting gene expression have demonstrated that these elements are frequent and occur across multiple genes and tissues. As early as 2002, a literature compilation identified cis-regulatory polymorphisms that were experimentally verified in over 100 genes, spanning most chromosomes, with a wide variety of functions (Rockman et al.). In post-

mortem brain tissue, evidence from quantitation of allele-specific expression showed an imbalance in at least one sample for half of the genes surveyed, indicating the presence of regulatory polymorphisms in that tissue (Bray et al. 2003). These results suggest the presence of many more yet to be discovered regulatory polymorphisms. Recent studies searching for relevant genetic variants use exome capture and deep sequencing that target

the transcribed/coding region of the DNA, but do not account for intronic or non-coding

regions.

There are multiple ways that regulatory polymorphisms can affect the resulting

expression of a functional protein. For instance, polymorphisms can affect transcription,

mRNA stability (Duan et al. 2003) or polyadenylation site usage potentially leading to

transcript degradation or instability (Miles et al. 1989) resulting in altered levels of

mRNA and protein. Polymorphisms modifying a splice site resulting in alternative

splicing can cause the inclusion of an intron in the transcript or introduce exon skipping.

Mutations present in the promoter region can affect transcription rate or the transcription

start site (Roberts 1969), while polymorphisms in the 3ʹ UTR can affect mRNA stability

(Kamiyama et al. 2007) or translation.

1.3 Detecting Sequence Variation

A variety of techniques are available to identify and characterize transcripts.

Many studies use a hybridization approach with microarrays by incubation of

fluorescently labeled cDNA on a high density oligonucleotide array. While the number

of genes included in these studies has increased over time, microarrays still do not allow

for the possibility of discovery. Only genes that are included can be detected, and

generally only certain regions of the gene are targeted. Other studies have used exon

array (Kwan et al. 2007), however this assay often has high background due to cross-

hybridization (Okoniewski et al. 2006) and it is difficult to compare results across

different experiments. Sequencing technologies provide an alternative to these

drawbacks by directly determining the sequence and allowing for novel findings.

Full transcriptome sequencing provides a less biased and more complete picture of transcription. It can be used to study tissue-specific splicing, alternate transcripts, post-transcriptional modification, quantitate expression and determine the true 5′ and 3′ transcript boundaries. Briefly, RNA is isolated from target tissues, and double-stranded cDNA is synthesized. The cDNA is fragmented, and adapters ligated to create a library.

High throughput sequencing is used to gather short sequence reads. Reads can be obtained from one end, or both (paired end). Paired end allows for greater accuracy in reads that map to multiple locations in the reference sequence or repetitive regions.

Continuous improvements in read lengths and number of reads are being made.

An effective and more targeted approach to characterizing regulatory genetic

elements is through measurement of mRNA ratios to uncover allelic expression

imbalance (AEI). These allelic differences in the expression of mRNA transcripts at a

given gene locus can be quantitated in mRNA transcribed from each of two alleles. The

presence of AEI can reflect altered transcription, RNA processing and translation. AEI

also indicates the presence of regulatory variants within this gene locus (acting in cis).

To measure AEI, a SNaPshot primer extension assay is used to measure allelic mRNA

ratios in heterozygous carriers of a marker SNP in the transcribed region.

1.4 Targets of Interest: Dopamine and Norepinephrine

Dopamine is a catecholamine neurotransmitter, and can also serve as a

neurohormone. It has both peripheral and central effects, but cannot cross the blood-

brain barrier; therefore its synthesis and action in the periphery is largely independent of

that in the brain.

Mutations in the dopaminergic pathway have been implicated in many diseases,

supported by an association between schizophrenia and aberrant neural development of

the dopamine transporter system (Howes et al. 2009). Deficits in dopamine signaling are

linked to Parkinson’s disease, as it is caused by degradation and death of dopaminergic

neurons in the substantia nigra.

When administered intravenously as a drug, dopamine can act on the sympathetic nervous system, increasing blood pressure and heart rate. It is used in cases of heart failure or shock, especially in infants. As it does not cross the blood-brain barrier,

dopamine cannot be used in treatment of neurocognitive disorders. In these cases, L-

DOPA, the metabolic precursor of dopamine, is the most used treatment for Parkinson’s disease.

Norepinephrine, also referred to as noradrenaline, is chemically similar to dopamine and also acts as a hormone and neurotransmitter throughout the body. It is released from sympathetic neurons and is exclusively produced by the enzyme dopamine

β-hydroxylase (DBH), catalyzing the conversion of dopamine to norepinephrine. When administered as a drug, it increases blood pressure by increasing vascular tone, and can be used in the treatment of hypotension.

Norepinephrine is important in the fight-or-flight response triggering a variety of responses throughout the body. Its release increases heart rate and stimulates the release of glycogen stores from the liver. It also increases blood flow to muscles and oxygen to the brain. Cortical norepinephrine can increase alertness and attention, enhancing learning. Norepinephrine in the central nervous system originates from the locus coeruleus. Neurons project bilaterally from here and innervate multiple sites including the cerebral cortex, limbic system and spinal cord.

1.5 Methods Common to All Projects

Allelic mRNA Expression Analysis

The SNaPshot (Life Technologies) primer extension assay was used to measure allelic mRNA ratios in heterozygous carriers of a marker SNP in the transcribed region

(Wang et al. 2005). A target region containing the SNP is PCR-amplified and then

exonuclease I and bacterial Antarctic phosphatase (New England Biolabs) are used to

remove excess primers and unincorporated dNTPs. A single base extension reaction adds

a fluorescently labeled dideoxynucleotide (ddNTP) complementary to the SNP. Each of the four ddNTPs is labeled with a different color fluorophore allowing detection of each allele in a heterozygous sample (Figure 1). Treatment with calf intestinal alkaline phosphatase degrades unincorporated ddNTPs. The signal is quantitated on a 3730 DNA

Analyzer capillary electrophoresis instrument (Life Technologies) and GeneMapper software is used to analyze the data. The allelic mRNA ratios are calculated from the abundance of major over minor allele of the SNP and the peak height ratios are normalized to the mean of the genomic DNA ratios.

Figure 1. Allelic expression imbalance assay quantitating the expression of each allele in a heterozygous sample. Each peak represents an allele and is proportional to the amount of PCR product amplified. In DNA, both copies are present equally, reflected by similar peak heights with a maternal and a paternal copy. In this case, in the RNA there is a greater abundance of the transcript containing the A allele vs. G. These differences can be detected with the SNaPshot assay, and quantitated by differences in the peak heights.

DNA and RNA Extraction

Genomic DNA was isolated from human frozen tissue. The sample was digested overnight at 55°C in nuclei lysis buffer and proteinase K, followed by sodium chloride treatment to precipitate proteins. The DNA was precipitated with ethanol, and purified with phenol chloroform extraction. To extract total RNA, the frozen tissue was homogenized in TRIzol reagent (Life Technologies) and isolated with chloroform and isopropanol. The product was purified using spin columns and treated with DNase to remove residual gDNA. Final concentrations were measured using the RNA or DNA quantitation reagent on the Qubit Fluorometer (Life Technologies).

1.6 Summary of Studies

The goal of this research is to discover regulatory variants and determine their

influence on regulation and expression in the dopaminergic pathway. I hypothesize that

there are frequent functional variants in our candidate genes (DBH, COMT and SNCA) modulating mRNA expression and in turn, contributing to variability in clinical phenotypes. The objectives of the DBH studies are to: characterize functional variants within DBH, determine the site of action of these SNPs and to identify genetic contributions to clinical phenotypes. These experiments have identified two SNPs that decrease allelic and overall mRNA expression and associate with protection against cardiovascular disorders. Work with SNCA has revealed a variant associated with AEI that marks differential 3′UTR usage. Finally, studies in COMT demonstrate that there is a genetic factor acting predominantly in African-Americans to regulate gene expression that lies outside of a 100 kb region. Overall, I identify evidence of frequent regulatory polymorphisms contributing to differential genetic expression.

Chapter 2: Regulatory Genetics of Dopamine β-Hydroxylase (DBH) and Effect on Clinical Phenotypes

2.1 Introduction and Background

Catalyzing the conversion of dopamine to norepinephrine, DBH is expressed in the adrenal gland, (biogps.gnf.org) and synaptic vesicles of postganglionic sympathetic neurons (Kim et al. 2002). Craig et al. mapped the human DBH gene to chromosome

9q34 (1988) which is composed of 12 exons and spans 23 kb (Kobayashi et al. 1989).

There are two main mRNA transcripts: type A is 2.7-kb and type B is 2.4-kb; they are identical, apart from a 300 base truncation in the 3′UTR of the type B transcript

(Kobayashi et al. 1989). The translated protein can be membrane-bound or soluble, and is 578-amino acids, and 64.8 kD.

Figure 2. DBH gene map indicating the location of two marker SNPs, rs110850 in the last nucleotide of exon 2, and rs77905 in exon 9. The promoter SNP rs1611115 is located upstream of the transcribed region.

DBH protein expression is tissue specific. In the brain, DBH is transcribed in

noradrenergic neurons of the locus coeruleus (LC), and high levels of DBH protein are

found in brain regions innervated by noradrenergic neurons. In the periphery, DBH is

predominantly concentrated in the adrenal glands and sympathetic nerve terminals.

While robust DBH expression occurs in the adrenals, sympathetic nerves are thought to

be the main source of plasma DBH (Weinshilboum 1979). Upon adrenergic stimulation,

both DBH and norepinephrine are released from vesicles via exocytosis (Weinshilboum

et al. 1971) accounting for substantial DBH levels in the circulation. Depletion of

sympathetic nerve terminals by guanethidine treatment reduces serum DBH activity by

50% in rats (Grobecker et al. 1977). Blockade or depletion of DBH also increases the

ratio of dopamine to norepinephrine in the circulation (Ohlstein et al. 1987; Bourdelat-

Parks et al. 2005). DBH-positive nerve fibers have been observed in sympathetically

innervated tissues including the liver (Feher 2004) near portal triads and central veins

(Pongor et al. 2010).

2.2 Known DBH SNPs and Clinical Correlations

Although rare (with less than 15 reported cases), congenital DBH deficiency demonstrates critical functions of DBH. Rare deleterious mutations at DBH appear to cause for DBH deficiency syndrome, (Kim et al. 2002) resulting in no detectable plasma levels of norepinephrine and epinephrine, and a 10-fold increase in dopamine levels

(Robertson et al. 1986; Man in 't Veld et al. 1987). Patients lacking DBH routinely

exhibit severe orthostatic hypotension, fainting episodes and cardiovascular disorders in

12 addition to ptosis, hyperflexible joints, hypoglycemia, and hypothermia; (Robertson et al.

1986; Biaggioni et al. 1990; Timmers et al. 2004) however, symptoms involving cognitive function have not been reported, raising the question what consequences result from DBH deficiency in the brain. Mothers of affected individuals have a history of miscarriages and stillbirths, suggesting DBH deficiency is usually lethal (Man in 't Veld et al. 1987).

Likewise, in a Dbh knockout mouse, most homozygous embryos do not survive to term without dihydroxyphenylserine (DOPS, a synthetic amino acid that is converted to norepinephrine) supplementation, due to adverse cardiovascular events (Thomas et al.

1995), demonstrating that DBH is required for normal fetal development. Knockout mice have delayed growth, ptosis, hypotension, and impaired peripheral vasoconstriction

(Thomas et al. 1995; Thomas et al. 1997; Swoap et al. 2004). Glucose levels in knockout mice fluctuate less than wild type in response to an insulin tolerance test (Ste Marie et al.

2003). At baseline, knockout mice do not appear to have any anxiety-related behaviors but are resistant to the anxiogenic effects of stimuli such as stress, caloric restriction, and cocaine (Swoap et al. 2004; Marino et al. 2005; Schank et al. 2008).

Circulating DBH activity has been proposed as an index of sympathetic nervous system activity; however, previous results have been ambiguous. Initial studies suggested acute stressors increase plasma DBH activity (Planz et al. 1973); however, these results were not replicated in subsequent experiments (Stone et al. 1974; Laurian et al. 1982). On the other hand, DBH levels have been proposed to be a useful measure of long-term noradrenergic function (Zabetian et al. 2001).

Significant inter-individual variability in plasma DBH levels has been associated

with genetic DBH variants. The presence of the minor allele (T) of the promoter SNP

rs1611115 (-1021C/T), with a minor allele frequency (MAF) of ~21% in subjects of

European ancestry, is associated with lower DBH serum levels, accounting for 35-52% of

variation in plasma DBH across populations (Zabetian et al. 2001; Garland et al. 2007;

Mustapic et al. 2007; Bhaduri et al. 2008). However, it remains unknown how

rs1611115-T reduces circulating DBH and norepinephrine levels, as reporter gene assays

have yielded opposing results (Chen et al. 2010; Chen et al. 2011; Pasha et al. 2013). It is

also unclear in which tissue(s) rs1611115 exerts its effect.

A multitude of clinical correlations have been found with SNPs in DBH, often

with conflicting results. Multiple clinical association studies have examined the role of

rs1611115 in neurocognitive disorders, for example, risk of addictive behaviors (Freire et

al. 2006; Kalayasiri et al. 2007) and response to addiction therapy (Kosten et al. 2013),

epilepsy (Depondt et al. 2004), schizophrenia (Windemuth et al. 2008) and attention

deficit hyperactivity disorder (Bhaduri et al. 2006). Many studies have found no

association of rs1611115 with Parkinson’s disease (Chun et al. 2007; Ross et al. 2008;

Punia et al. 2010), whereas one study found a protective effect in homozygous T allele

carriers (Healy et al. 2004). The T allele is associated with higher risk avoidance in

healthy females (Kamata et al. 2009), and lower smoking rates (Freire et al. 2006). TT

homozygotes have attentional asymmetry to spatial stimuli (Greene et al. 2010),

decreased risk of migraine (Fernandez et al. 2009), association with impulsive personality

traits (Hess et al. 2009) and increased propensity to paranoia during cocaine use

(Kalayasiri et al. 2007). Other studies have found no associations with vasovagal syncope (Sorrentino et al. 2010), development of cocaine addiction (Guindalini et al.

2008), or alcoholism (Freire et al. 2005).

Peripheral effects of DBH on blood pressure and blood glucose levels have also been examined. Consistent with a lack of DBH causing orthostatic hypotension (Man in

't Veld et al. 1987; Robertson et al. 1993), the major C allele of rs1611115 has been associated with elevated blood pressure, both under normal conditions (Yeh et al. 2010) and in response to stress (Chen et al. 2010), and hypertension in the presence of increased fasting plasma glucose levels (Abe et al. 2005).

Fewer studies have examined rs1108580 (MAF~46% in people of European ancestry), located in the last nucleotide of exon 2. Because of its position directly at a splice junction, rs1108580 has been proposed to alter RNA splicing, although this has not been demonstrated experimentally (Cubells et al. 1998; Wood et al. 2002). The minor allele has been associated with higher DBH in the serum and cerebrospinal fluid (Cubells et al. 1998), reduced risk of paranoid ideation in depressed patients (Wood et al. 2002), superior spatial working memory performance (Parasuraman et al. 2005; Greenwood et al. 2009), and reduced risk of alcoholism (Kohnke et al. 2006). Unique associations were found in an Indian population including an association with orthostatic hypotension

(Punia et al. 2010); however, the lack of a detectable effect on plasma DBH sheds some doubt over these associations (Bhaduri et al. 2008). No associations were found with schizophrenia (Yamamoto et al. 2003), migraine (Fernandez et al. 2009), delirium tremens or alcohol withdrawal seizures (Kohnke et al. 2006). Overall, the evidence for

substantial genetic effects on a variety of phenotypes is strong while the responsible molecular genetic mechanisms are far from clear.

2.3 Myocardial Infarction and Coronary Heart Disease

Accounting for 30% of all deaths (17.3 million in 2008), cardiovascular disease

(CVD) is the leading cause of mortality in the world (Alwan 2011) and is projected to

increase in prevalence, causing 23.3 million deaths by 2030 (Mathers et al. 2006). CVD

is actually a group of disorders including: coronary heart disease, cerebrovascular

disease, peripheral arterial disease, rheumatic heart disease, congenital heart disease and

deep vein thrombosis and pulmonary embolism. Symptoms may include faintness, nausea, shortness of breath, and/or pain in the chest, arms, left shoulder, elbows, jaw or back. Many of these symptoms may be caused by a surge in catecholamines (Little et al.

1986).

Acute myocardial infarction (MI) occurs when there is decreased blood flow to the heart via narrowing of the blood vessels resulting from atherosclerosis as plaque accumulates in the vessel. The acute event is caused by a blockage, a blood clot or plaque, preventing blood flow to the heart. To prevent MI, patients can change lifestyle factors, through exercise, smoking cessation, diet improvement and maintaining a healthy

weight. Drug treatment can also lower risk of recurrence, through a combination of anti-

hypertensives, statins (to lower cholesterol) and aspirin. Following a cardiac event,

interventions include: coronary artery bypass, balloon angioplasty, valve repair, or heart

transplantation.

To diagnose an MI and categorize it, an electrocardiogram (ECG) reading can be performed, which traces the electrical signals in the heart. ECG uses sensors placed on the skin to detect electrical changes in heart activity, following depolarization of cardiac cells. In a resting state, the cell has a negative charge. An influx of positive ions during

depolarization leads to contraction. A normal heart will follow a set pattern, reflected in

the ECG traces (Figure 3). Deviations from baseline can reflect abnormalities. The P

wave is recorded during atrial depolarization, and the upper chambers of the heart

contract. The QRS wave reflects ventricular contraction. In left ventricular hypertrophy,

there is increased force, reflected by a larger than normal R wave. The ST interval is the

straight line between the QRS and T waves, reflecting depolarization of the ventricles.

Elevation or depression of this segment is an indication of heart damage, and helps

categorize the type of myocardial infarction. The T wave is ventricular relaxation.

An elevation of the ST interval, termed an ST elevation MI, or STEMI, is usually

more serious and requires more intense intervention, such as an angioplasty. STEMIs

comprise 25-40% of cases (O'Gara et al. 2013). A non-ST elevation MI, or NSTEMI,

can usually be managed with medication but treatment may include an angioplasty if the

patient is at increased risk. Globally, there are over 3 million STEMIs and 4 million

NSTEMIs per year (White et al. 2008). Blood tests can also be performed to measure

levels of creatine kinase and troponin.

Figure 3. Example electrocardiogram reading, courtesy of Hank van Helvete (http://commons.wikimedia.org/wiki/File:EKG_Complex_en.svg#mediaviewer/File:EKG _Complex_en.svg.)

2.4 Study Overview

This study characterizes two regulatory variants that affect DBH mRNA expression predominantly in peripheral sympathetic neurons, and clinical phenotypes associated with these variants, consistent with their effect on expression in target tissues.

DBH variants have been associated with large changes in circulating DBH and norepinephrine and implicated in multiple disorders; yet causal relationships and tissue-

specific effects remain unresolved. The objective is to characterize regulatory variants in

DBH mRNA, effect on expression in human tissues, and role in modulating sympathetic

tone and disease risk.

2.5 Materials and Methods

Sample Preparation

Liver (125), adrenal (17), kidney (ten) and small bowel (ten) tissue samples were

provided by the Cooperative Human Tissue Network, funded by the National Cancer

Institute. We acknowledge use of human adrenal tissues (18) provided by the National

Disease Research Interchange (NDRI). Lung biopsy specimens (30) were provided by

the Ohio State University Tissue Procurement Pathology Core. Pineal gland samples

(five) were dissected by a trained neuropathologist at the Dartmouth-Hitchcock Medical

Center. Ten brain regions from ten individuals were procured from the Miami-Dade

Brain Endowment Bank. Finally, 20 LC samples were provided via the Neuropathology

Core of the Emory Neuroscience NINDS Core Facilities. Human autopsy/biopsy tissue

samples were flash frozen in liquid nitrogen. A subset of the adrenal samples were

provided fresh in RNAlater (Qiagen). All samples were obtained under protocols approved by the local Institutional Review Boards. RNA (Wang et al. 2005) and DNA

(Miller et al. 1988) were isolated from tissues as previously described, and RNA quality

was assessed using the Bioanalyzer (Agilent Technologies). cDNA was synthesized via

reverse transcription with SuperScript III (Invitrogen), oligo-dT, and gene-specific primers. 19

Genotyping

Fluorescent restriction fragment length polymorphism was used if the SNP of interest was included in a restriction enzyme cut site. If not, SNPs were genotyped with allele-specific melting curve analysis (Papp et al. 2003). A GC clamp of 10-15 bases was added at the 5′ end of one of the allele specific primers, resulting in products that can be differentiated by their melting temperatures via real-time PCR. The SNaPshot primer extension assay was also used in genotype determination. Primer sequences and assay type used are listed in Appendix C.

Allelic mRNA Expression Analysis

AEI was measured as described in Section 1.5. Allelic ratios are calculated as the major/minor allele however, the inverse was taken if the ratio was below one for rs77905.

To test the efficiency of the gDNA and cDNA primers samples homozygous for either the major or minor allele of rs1108580 or rs77905 were selected, and a larger piece of

DBH, ~550bp, was amplified with a different set of primers. This product was purified and quantitated. The products were combined in differing ratios, 1:4, 2:3, 1:1, and 3:2, diluted, used as template for the PCR reaction, and carried through the SNaPshot procedure. Duplicates were run on the 3730 DNA analyzer, and peak area ratios were compared to template ratios.

Quantitative Real-Time reverse transcriptase PCR (qRT-PCR)

DBH mRNA expression was measured by qRT-PCR with a 7500 Fast Real-Time

PCR System (Life Technologies). Reactions were prepared in 10 ul volumes with Fast

SYBR Green Master Mix (Applied Biosystems) and cDNA synthesized from 25 ng of

total RNA. Primers spanning the exon 2-3 junction, including rs1108580 were used for

DBH detection. Threshold values were set to 0.2 using the 7500 Software v2.0.5

(Applied Biosystems) and cycle threshold numbers (Ct) were computed for each well.

The mean PGK1 amplification Ct value was subtracted from the mean DBH Ct value for

each sample to calculate the ∆Ct, and obtain standardized values. To assure specific

detection of mRNA, assays were replicated for a subset of samples using a Taqman probe

spanning the exon 11-12 junction (Life Technologies) with GAPDH as a housekeeping

gene.

The presumed main source of DBH in the body is the adrenal gland, so I used

human adrenal gland total RNA (Clontech) as a positive control in expression assays.

This sample was pooled from 61 male/female Caucasians aged 15-61.

DNA Sequencing Using Ion Torrent

Regions of DBH (~2-3kb) were PCR amplified using NEB (New England

Biolabs) or JumpStart Mastermix (Sigma). Amplicons were combined based on length

and concentration and treated with exonuclease followed by phenol-chloroform extraction. The sample was fragmented using Covaris shearing, and barcoded libraries were prepared using NEBNext Fast DNA Library Prep Set for Ion Torrent, the Ion

Xpress Barcode Adapters Kit and the Ion One Touch Template Kit. The library was

sequenced on the Ion Torrent, using the Ion Torrent Ion PGM Sequencing Kit (Life

Technologies).

Clinical Cohorts

Pharmacogenomic Evaluation of Antihypertensive Responses (PEAR)

In collaboration with Drs. Julie Johnson and Yan Gong, we utilized data from 463 subjects in the Pharmacogenomic Evaluation of Antihypertensive Responses (PEAR) trial. Patients with high blood pressure were treated with the β1-selective adrenergic receptor blocker atenolol, and blood pressure and glucose responses were measured. The data were adjusted for baseline phenotype, age, and gender, and racial groups were analyzed separately.

Health and Retirement Study (HRS)

Data from the Health and Retirement Study, produced and distributed by the

University of Michigan with funding from the National Institute on Aging (grant numbers

U01AG009740, RC2AG036495, and RC4AG039029), Ann Arbor, MI, were used to test the effect of functional variants on clinical phenotypes and biomarkers, with 12,507 subjects genotyped using the Illumina Human Omni2.5-Quad (Omni2.5) platform. This study is a longitudinal survey of Americans 50 years and older from across the country.

The data were downloaded from dbGAP with IRB permission.

Jackson Heart Study (JHS)

Data from the Jackson Heart Study (JHS), which investigates cardiovascular risk factors, were used (Sempos et al. 1999). Participants are African-American men and women from the three counties (Hinds, Madison, and Rankin) surrounding Jackson, MS.

For the current analysis, rs1611115 was genotyped directly using the CARe IBC Array

(Keating et al. 2008) in 2,762 participants, and was imputed in an additional 262 individuals, using Affymetrix 6.0 genotypes and a 1000 Genomes reference panel as described (Fox et al. 2011). Genotypes for rs1108580 were imputed in all 3,024 participants. After excluding participants with coronary heart disease (CHD) at baseline or with missing data for incident myocardial infarction (MI), 2,378 participants were available for analysis. In addition to MI, procedures included in the incident CHD outcome include coronary artery bypass grafts and coronary angioplasty. Genotype association tests were performed with an additive model, counting the number of copies of the T allele for rs1611115 and A allele for rs1108580 and adjusted for age, gender and

10 principal components (to account for potential ethnic stratification). Blood pressure readings were adjusted if the subjects were taking hypertension medication: the systolic blood pressure (SBP) reading was increased by 10 and the diastolic blood pressure (DBP) by 5. All analyses were approved by the Institutional Review Board of the University of

Mississippi Medical Center.

Cell Culture

HepG2 cells were maintained in a 37°C incubator with 5% CO2 in low glucose

DMEM, 10% fetal bovine serum, and Penicillin-Streptomycin. To passage or harvest the cells, 0.25% trypsin was used.

Jackson Laboratories Mouse Phenome Database

This resource is a collection of microarray and phenotype data from different laboratories on standard inbred mouse strains and can be accessed at:

http://phenome.jax.org. DBH expression was measured with a variety of probe sets, with

Cy3 labeling mouse reference mRNA (male C57BL/6 liver, kidney, lung, brain, and

spleen) and Cy5 labeling mouse liver RNA. The analysis generates Cy5:Cy3 ratios and

is normalized for intensity and the log (base 2) is reported.

5′ RACE (Rapid Amplification of cDNA Ends)

5′UTRs (untranslated regions) were captured and sequenced using the FirstChoice

RLM-RACE Kit (Ambion).

Phenome-Wide Association Study (PheWAS)

A PheWAS was performed using data from the Geisinger Clinic MyCode

biorepository, on 3,035 subjects (Gottesman et al. 2013). Clinical and demographic

information was obtained from electronic health records (EHR). We defined case-control

status using International Classification of Diseases, Ninth Edition (ICD-9) diagnosis

codes from the EHR. We required > 3 of the same ICD-9 code per individual and >10 case subjects with 482 phenotypes meeting the inclusion criteria. Using the case/control status and the two SNPs [rs1611115 MAF (T) 0.22; rs1108580 MAF (A) 0.47] we calculated associations using logistic regression via custom script in R, (Team RDC,

2009), adjusting models for age and sex. Associations were calculated using additive, recessive, and dominant genetic encoding, where the minor allele was the coded allele.

We also tested pairwise SNP-SNP interactions. We determined the significance of the interaction via likelihood ratio test (LRT), comparing full versus reduced models, using logistic regression. The full model was: SNP1 + SNP2 + SNP1*SNP2 and the reduced model was: SNP1 + SNP2.

To seek replication for our PheWAS results, we used data on 4,027 individuals

from The Marshfield Personalized Medicine Research Project (PMRP) biobank and

linked EHR (McCarty et al. 2005). ICD-9 codes are organized in a hierarchical manner, with main categories and sub-categories. We first identified the main ICD-9 code category for any results within our discovery MyCode that had an association p < 0.01.

Then we identified all ICD-9 codes within those categories - as different medical centers

often use different ICD-9 codes by habit of practice. We then defined case/control status

for all of those ICD-9 codes using the same criteria used for the discovery MyCode

PheWAS and performed the analysis the same way.

In the Marshfield PMRP dataset, genome-wide genotyping was performed at the

Center for Inherited Disease Research (CIDR) using the Illumina 660W-Quad Beadchip.

Genotyping calls were made at CIDR using BeadStudio version 3.3.7. For the Geisinger

MyCode data set genotyping was performed at the University of Pittsburgh Genomics

and Proteomics Core Laboratories using the Illumina Omni Express Beadchip.

Genotyping calls were made at the University of Pittsburgh using GenomeStudio. Data

for both sets were cleaned using the eMERGE QC pipeline developed by the eMERGE

Genomics Working Group (Zuvich et al. 2011). This process includes evaluation of

sample and marker call rate, gender mismatch, duplicate and HapMap concordance, batch

effects, Hardy-Weinberg equilibrium, sample relatedness, and population stratification.

For the current PheWAS, only the two SNPs of interest were extracted from the GWAS

dataset.

Statistical Analyses

SVS software (Golden Helix) was used to calculate linkage disequilibrium (LD)

and R2 values between SNPs, allele frequencies, Hardy-Weinberg equilibrium, and

perform genotype association tests. ANOVA and t-tests were performed with GraphPad

Prism and p<0.05 was considered to be significant. In SPSS v21 (IBM), stepwise linear regression using age, race and gender was used to determine the effect of subject demographics on DBH mRNA expression.

Humanized DBH mouse model

Mice were housed and treated in accordance with Emory University’s

requirements and regulations for animal care and use. As the site containing rs1611115

is not conserved in the murine version of the gene, a collaborator from Emory University,

Dr. David Weinshenker, generated a transgenic line. These mice were engineered to

have a 175 kb human BAC (RP11-746P3) encompassing DBH, and crossed onto a Dbh

knockout background. The incorporation of the entire, intact BAC was confirmed by

PCR. The human BAC expresses either the C or T allele at rs1611115 and each line was

independently generated twice, yielding the strains C13, C17, 383T and 249T. In

addition to DBH, there are several kb of human sequence both upstream and downstream

of the DBH gene, making it less likely than a conventional transgene to have human and

mouse enhancer elements interacting. The BAC also includes the gene sarcosine

dehydrogenase (SARDH), which is downstream from DBH and can be used as a control.

As mouse Sardh is not knocked out in this model, I designed the primers for SARDH

amplification to have multiple mismatches to Sardh to ensure specific amplification of the human gene.

Following generation of the four strains, mice were sacrificed with carbon

dioxide. Each genotype group consisted of half males and half females. All mice were

between 3 and 5.5 weeks of age at the time of tissue harvest. Each group contained 12 mice, except 383T which had 8. Dr. David Weinshenker, Jason Schroeder, and Katherine

Henry collected adrenal, brain, heart, lung, and liver. Each piece of tissue was dissected, frozen on dry ice and shipped to OSU. For the brain samples, Dr. Ryan Smith isolated the locus coeruleus from surrounding tissue. Then I homogenized the tissue in TRIzol and extracted mRNA as previously described (Anderle et al. 2006). Due to the small

amount of tissue available for the adrenal samples, I added 1 μg of glycogen to the

aqueous phase after the first spin to aid in RNA recovery.

cDNA was synthesized via reverse transcription with SuperScript III (Invitrogen),

oligo-dT, and gene-specific primers for human DBH and SARDH. mRNA expression

was measured by qRT-PCR and the mean SARDH Ct was subtracted from the mean

DBH Ct value for each sample to calculate the ΔCt and obtain standardized values.

2.6 Results

mRNA Levels of DBH in Human Tissues

I measured DBH mRNA expression across a variety of human tissues, including

sympathetically innervated organs devoid of ganglionic nuclei. While DBH protein is

known to be axonally transported to sympathetic nerve terminals (Dahlstrom 1971), I hypothesized that mRNA could also be transported to tissues with sympathetic innervations, supporting local translation at the site of action. The highest expression

levels were detected in the LC, adrenal gland, and liver, followed by lower levels in lung,

kidney, small bowel, and heart (Figure 4A). Because DBH appears to be exclusively

expressed in neuronal tissues, DBH mRNA detected in organs containing

sympathomimetic nerve terminals but no neuronal nuclei was likely transported to nerve

terminals in the target tissues. In particular, DBH mRNA levels were strikingly high in

the liver (Figure 4A), identifying this organ as a possible major source of DBH in the

circulation.

Figure 4. DBH mRNA expression. A. Comparison of DBH mRNA expression across tissues including locus coeruleus (LC), adrenal (AD) and small bowel (SB) using a Taqman assay targeting the exon 11-12 boundary, standardized to GAPDH. Each group contains 6-8 individual samples of each tissue type. B. DBH expression targeting the exon 2-3 boundary, standardized with PGK1 using qRT-PCR for 16 LC, and 41 liver samples. In both cases, the qRT-PCR cycle thresholds are standardized to the house keeping gene; a higher Ct denotes lower expression.

Next I measured DBH mRNA expression using qRT-PCR in an expanded set of

LC and liver tissues with primers spanning the exon 2-3 boundary to avoid the potentially confounding influence of rs1108580 on RNA splicing (Figure 4B). Expression in the liver was again comparable to tissues that strongly express DBH such as the LC.

To determine the source of DBH in the liver, I examined the hepatocellular

carcinoma cell line, HepG2 which is heterozygous for rs1611115, rs77905, and

rs1108580 (Motallebipour et al. 2009). Following cDNA synthesis, I measured DBH

mRNA expression with three different primer sets. Using primers spanning the exon 2-3

boundary: Ct=34.8, exon 9: Ct=32.2, and in 3′UTR: 32.6. The housekeeping gene PGK1

was well expressed (Ct=19.4), indicating successful cDNA preparation, and the quality of

RNA extracted from these cells was high, with a RIN value of 10, as assessed by

Bioanalyzer. Despite sufficient detection of PGK1, all DBH cycle thresholds were above

30, therefore I conclude DBH mRNA expression is very low and almost undetectable in

the cell line HepG2, supporting the notion that DBH measured in the liver was derived

from cell types other than hepatocytes, consistent with immunohistochemistry

experiments (Kaiser et al. 2003; Pongor et al. 2010).

The liver is sympathetically innervated by the celiac ganglia; therefore, any DBH

mRNA detectable in the liver is likely to have undergone axonal transport to hepatic

nerve terminals. I tested for the presence of DBH mRNA in celiac ganglia and found that

it was well expressed. The resulting Cts from measuring mRNA expression in sample 1

were 23.1 ± 0.04 and 24.9 ± 0.2 with primers spanning exon 2-3 or the 3′UTR

respectively, and 18.3 ± 0.06 and 20.1 ± 0.2 in sample 2. The housekeeping gene PGK1

was also well expressed in samples 1 (Ct= 19.5 ± 0.06) and 2 (Ct=18.5 ± 0.2).

Since it is known from previous studies that the LC is a major site of DBH

transcription and translation in the brain, I tested whether DBH mRNA is found in brain

regions that receive noradrenergic innervation, including cortices (Brodmann areas 10,

22, 24, and 46), amygdala, and hippocampus. Other brain regions tested were: insula,

cerebellum, substantia nigra, putamen, and raphe nuclei. Expression was near the limit of

detection or undetectable in most brain regions, with some cortex samples showing Ct

values of 29-31, indicative of low mRNA expression.

The pineal gland receives sympathetic innervation from the superior cervical

ganglion and responds to norepinephrine in the circulation, as it is not protected by the

blood brain barrier. Therefore I also measured mRNA expression in pineal glands, to

determine any regulatory genetic effects in this organ. Expression was measured in five

pineal samples. While all expressed the housekeeping gene, PGK1, four had non-specific

amplification with DBH primers. One sample expressed DBH, with a mean Ct of 26.4 ±

0.9 and was heterozygous for rs1611115 and rs77905. However, the mean allelic ratio

was 1.2 ± 0.2.

SNaPshot Controls

As the two marker SNPs were located near the borders of exons, separate primers

were designed for use with gDNA and cDNA for both rs1108580 and rs77905. In order to ensure there were no differences in amplification, a control experiment was performed

(Figure 5). Samples homozygous for either the major or minor allele were combined in

30 differing ratios, and quantitated using the SNaPshot protocol. I found that the reactions were equally efficient, validating the use of different primers sets for the gDNA and cDNA.

Figure 5. Efficiency comparison of primers for PCR amplification of DNA and RNA around rs1108580 (A) and rs77905 (B). Separate primer sets were used as both SNPs are near the border of an exon. The products were combined in differing ratios, 1:4, 2:3, 1:1, and 3:2, and used in the SNaPshot assay. Peak area ratios were compared to input ratios. A diamond indicates gDNA primers, while a square indicates cDNA primers.

Determination of Allelic DBH mRNA Ratios as an Indicator of Regulatory

Polymorphisms

Allelic ratios in gDNA are assumed to be 1 for autosomal DBH, while any significant deviation of allelic ratios in mRNA (allelic mRNA expression imbalance or

AEI) reveals the presence of regulatory variants (Wang et al. 2005).

The first tissues used for allelic expression analysis were the LC and pons, as these are presumed to be the main sources of DBH in the brain. Little to no AEI was observed when using the first marker SNP, rs5320. Therefore, three additional marker

SNPs were used to measure AEI. The magnitude of AEI was consistent when measured at multiple marker SNPs within the same sample, confirming the reproducibility of the assay. However, these results appeared to demonstrate that there was little allelic expression variation. As we later determined that rs1611115 was a functional SNP, heterozygous samples are indicated with a star (Figure 6). The AEI does not correlate with rs1611115 genotype in these samples.

Figure 6. Allelic mRNA expression ratios in LC or pons (P) measured at rs129882, rs77905, rs5320 or rs1108580. If samples were heterozygous for multiple markers, AEI was measured at each. The values are plotted as a ratio of the major over the minor allele and standardized to the measured genomic DNA allelic ratio. Samples heterozygous for rs1611115 are indicated with a star. The horizontal lines demark the threshold for significant AEI (0.80-1.26).

Next, I wanted to determine if this lack of AEI was a global phenomenon, or tissue-specific. Therefore I measured AEI at rs129882 in liver and compared the values to those measured in brain. There were two liver samples that displayed significant AEI, and sample #111 had a large fold change, with an average value of 5.1. This sample had the greatest AEI value, and was the only one heterozygous for rs1611115. Pons sample

#99 was also heterozygous for this SNP, but did not display significant AEI (Figure 7).

Figure 7. Allelic mRNA expression ratios measured at rs129882 in liver, LC, or pons (P). Samples heterozygous for rs1611115 are indicated with a star and rs1108580 with a diamond. The horizontal lines demark the threshold for significant AEI (0.6-1.7).

In order to expand the number of samples included, two common SNPs located in the DBH coding region, rs1108580 and rs77905 were selected as markers to measure allelic mRNA ratios in human liver, LC, and adrenals. No sample displayed evidence of 34 copy number variation (gene duplications). With three standard deviations of the gDNA ratio as the threshold, I considered an allelic mRNA ratio of <0.8 and >1.3 as significant evidence of AEI.

With marker SNP rs1108580, most samples demonstrated AEI, with the main allele more abundant than the minor allele (ratios >1) in all tissues (Figure 8). Therefore, rs1108580 is likely a causative variant and/or shares a haplotype with one or more causative variants. However, striking differences were observed in AEI ratios between tissues: all less than 2-fold in LC and adrenals (Figure 8B) but up to 11-fold in the liver

(Figure 8A).

There was no AEI found in celiac ganglia tissue heterozygous for rs1611115 measured at rs1108580 (1.1 ± 0.2), which differs dramatically from the large AEI ratios measured in the liver of compound heterozygotes. A second celiac ganglion tissue was homozygous for the major allele of rs1611115 but heterozygous for rs1108580 and marker SNP rs77905. Allelic imbalance was measured at rs77905 in exon 9 (4.0 ± 0.8, 5 replicates) and rs1108580 in exon 2 (AEI = 1.3 ± 0.3, 5 replicates). These discrepant allelic ratios – well supported by multiple replicates – indicate that different mRNA transcripts may be present in celiac ganglia.

Additional instances of AEI were revealed only with marker rs77905 (Figure 8A).

The large AEI ratios were replicated with marker rs77905, in a subset of livers heterozygous for rs1611115. These results document a remarkable regulatory genetic effect specifically in liver, with much lesser effects in the brain and adrenals (Figure 8B).

DBH mRNA levels in other tissues were too low for accurate AEI analysis.

Figure 8. Allelic DBH mRNA expression ratios measured at marker rs1108580 (black) and rs77905 (grey) in liver (A), locus coeruleus and adrenal gland (B). The values are plotted as a ratio of major over minor allele for rs1108580. As rs77905 is not in LD with rs1108580, AEI ratios ranged from >1 to <1 with near equal distribution in either direction; for comparison, these ratios are shown as absolute values in only one direction (>1). Livers heterozygous for rs1611115 are indicated with *. The horizontal line at 1.3 demarks the threshold for significant AEI. 36

In both liver and brain, samples heterozygous for both marker SNPs yielded AEI ratios of similar magnitude, with a correlation of R2=0.96 (Figure 9). This demonstrates the validity and reproducibility of the assay, as well as a similar effect in exons 2 and 9.

Figure 9. Linear regression of average AEI magnitude measured at two marker SNPs: rs1108580 and rs77905 in any samples heterozygous for both marker SNPs in all tissues. The equation of the line was y = 1.1x - 0.14 with R2=0.96. In order to compare the magnitude of the values, we used the inverse of ratios below 1.

Third Regulatory Variant

The allelic mRNA ratios in liver using the SNP rs77905 deviate significantly from

1 and by as much as tenfold, indicative of the presence of regulatory factors with robust effects (Figure 8). The ratios, however, range both above 1, indicating the transcript containing the major allele is more prevalent, and below 1, indicating the transcript containing the minor allele is more prevalent. The range of values is also more heterogeneous than those measured with rs1108580. Therefore, the SNP rs77905 is unlikely to be functional and any causative variants likely are in low LD with this SNP.

If the functional SNP resides on the same haplotype as rs77905, I would expect AEI values in the same direction, not both above and below 1.

A subset of livers had AEI ratios of 1.5-2.5 fold measured at marker SNP rs77905, but were homozygous for both rs1611115 and rs1108580. None of the SNPs genotyped here fully accounted for these cases of AEI suggesting the existence of another regulatory variant. The entire DBH gene locus was sequenced, beginning 2 kb upstream of the coding region and extending 15 kb downstream from the end of the 3′UTR into the

gene encoding sarcosine dehydrogenase (SARDH). Two SNPs were heterozygous in all

samples sequenced, rs7864658 and rs739398, but both are in high LD with rs77905 (D′

for both is 0.96, and R2 is 0.7–0.85). Since the AEI ratios point in both directions (<1 and >1), these SNPs were excluded as causative.

Any additional regulatory variant would likely be in low LD with the SNP because the allelic mRNA ratios ranged from below to above one. I hypothesize that this

putative third variant is located outside of the gene region sequenced, has a MAF of 15-

20% and is in low LD with rs77905 and rs1108580.

Transcript Characterization

To determine whether full-length DBH mRNA was expressed in the liver, the mRNA was converted to cDNA and sequenced. Sequencing demonstrated that the liver mRNA contained the annotated DBH coding sequence according to the UCSC genome database. To confirm that the DBH mRNA from liver and celiac ganglia represents capped mRNA, I performed 5′ RACE and amplified the 5′ UTR. Sequencing of this product revealed that liver transcripts were capped and the 5′UTR was full-length (39 bases) matching the AceView and UCSC references. Thus, there is bona fide DBH mRNA expression in the liver and celiac ganglia. I hypothesized that rs1611115 could introduce an extended 5′UTR resulting in transcripts with one 5′UTR being preferentially transported, accounting for the allelic imbalance measured in the liver, but not in the celiac ganglia. However I did not find an extended or unannotated 5′UTR sequence in either the liver or celiac ganglia.

Genotype Association with Instances and Degree of AEI in Liver

To identify the DBH variants responsible for the observed AEI, 11 variants previously implicated in clinical association studies were genotyped in all samples.

Table 1 lists the MAF and results of testing each SNP genotype for association with AEI, while Table 2 provides linkage disequilibrium (LD) between the SNPs.

Table 1. SNPs genotyped and genotype association test using the F statistic

Association with Association with Minor rs77905 AEI rs1108580 AEI Marker Location Allele MAF values (F-test p) values (F-test p) rs1079783 upstream G 0.25 0.16 0.92 rs3025343 upstream A 0.10 0.48 0.51 rs141116007 and CA upstream B 0.41 0.047 0.098 repeat rs1076150 upstream T 0.49 0.28 0.60 rs1989787 upstream A 0.34 0.54 0.26 rs1611115 upstream T 0.21 0.0016 2.0E-07 rs2519143* intron 1 A 0.20 0.0016 2.9E-07 rs1108580 exon 2 A 0.46 0.20 x† rs77905 exon 9 T 0.46 x† 0.57 rs6271 exon 11 T 0.06 0.67 0.24 rs129882 exon 12 T 0.25 0.43 0.46

*rs2519143 had similar p values compared to rs1611115 but was excluded as the causative SNP because it was homozygous in a tissue showing strong AEI. †When used as the marker SNP, rs1108580 and rs77905 could not be tested as all samples were heterozygous.

Table 2. LD values (D′ and R2) for all SNPs genotyped in liver

D' R2 D' R2 D' R2 D' R2 D' R2 D' R2 D' R2 D' R2 D' R2 D' R2 D' R2 rs141116007 Marker rs1611115 rs129882 rs1108580 rs77905 rs2519143 rs3025343 rs1989787 rs1079783 rs1076150 & CA rep rs6271 rs1611115 1.00 1.00 ...... rs129882 0.29 0.00 1.00 1.00 ...... rs1108580 0.73 0.15 0.40 0.04 1.00 1.00 ...... rs77905 0.11 0.00 0.06 0.00 0.04 0.00 1.00 1.00 ...... rs2519143 0.91 0.79 0.02 0.00 0.77 0.16 0.44 0.02 1.00 1.00 ...... rs3025343 0.99 0.03 0.98 0.04 0.07 0.00 1.00 0.18 0.99 0.03 1.00 1.00 ...... rs1989787 1.00 0.14 0.63 0.13 1.00 0.40 0.26 0.06 1.00 0.12 0.22 0.01 1.00 1.00 ...... rs1079783 0.50 0.01 0.27 0.06 0.55 0.06 0.58 0.08 0.69 0.02 1.00 0.04 0.70 0.19 1.00 1.00 ...... rs1076150 1.00 0.26 0.27 0.02 0.94 0.80 0.06 0.00 0.87 0.19 0.16 0.00 1.00 0.46 0.39 0.03 1.00 1.00 .... rs141116007

41 & CA rep 0.82 0.27 0.60 0.05 0.38 0.10 0.35 0.04 0.86 0.27 0.75 0.04 0.83 0.21 0.56 0.04 0.40 0.10 1.00 1.00 .. rs6271 1.00 0.02 0.99 0.02 0.49 0.01 1.00 0.09 1.00 0.02 0.32 0.05 0.53 0.03 0.97 0.02 0.54 0.02 1.00 0.04 1.00 1.00

AEI was significantly associated with rs1611115 genotype, measured at

rs1108580 (p=2.0E-07) and at rs77905 (p=0.0016). Of the 17 livers with the highest AEI ratios, all were heterozygous for rs1611115 (Figure 8), providing overwhelming evidence that this promoter SNP has strong effects on DBH mRNA expression. When measured with marker rs1108580 which is in partial LD with rs1611115, the minor allele was consistently associated with lower expression. Of 15 livers heterozygous for marker rs1108580 but not for rs1611115, 11 showed AEI with ratios above 1.3, indicating rs1108580 is causative itself. Moreover, livers with the highest ratios were heterozygous for both, comparing left and right panels in Figure 8. These results support the conclusion that both act to decrease mRNA expression of the variant allele, with rs1611115-T causing ~4-fold and rs1108580-A ~2-fold imbalance.

In contrast to the results in liver, rs1611115 genotype had no discernible effect in the brain or adrenal tissues, while rs1108580-A was associated with 1.2-1.7-fold reduced

DBH mRNA expression (Figure 8B), suggesting a small but potentially relevant effect of the minor allele of rs1108580 on mRNA abundance in these tissues. AEI ratios could not be determined accurately in the other organs tested because of low DBH mRNA expression.

Most other SNPs genotyped failed to achieve significance when tested for an association between genotype and AEI (see Table 1 for all values). The SNP rs2519143 was found to be statistically significant with regard to AEI, but this can be accounted for by high LD with rs1611115 (Table 3).

Table 3. Minor allele frequencies and LD values (D′ and R2)

D′ R2 D′ R2 D′ R2 MAF Marker rs1611115 rs1108580 rs2519143 (T): 0.21 rs1611115 ...... (A): 0.46 rs1108580 0.73 0.15 . . . . (A):0.20 rs2519143 0.91 0.79 0.77 0.16 . .

SNPs in the liver cohort that were significantly associated with instances of AEI.

Effect of Genotype on DBH mRNA Expression in Human Tissues

mRNA levels were measured in all available tissues and tested for association

with genotypes (Figure 10). Expression was measured in liver, adrenal, brain and lung,

while mRNA levels were close to the detection limit in heart and small bowel.

Judged by the AEI ratios, the minor A allele of rs1108580 was associated with lower expression in nearly all tissues tested, albeit with rather small effects in LC and adrenals. Accordingly, mRNA levels were significantly lower (1.7-fold) for AA livers

compared to GG (Figure 10A; p=0.005; two-tailed t-test). The rs1108580 genotype was

not significantly associated with mRNA levels in brain (p=0.6) nor lung tissues (p=0.3),

possibly owing to small effect size (brain), large variability and low sample number

(Figure 10A-C).

In livers, there was no significant difference between the expression in CT and TT

(p=0.2), so these genotypes were combined. A two-tailed t-test showed a significant effect of rs1611115 genotype on total DBH mRNA expression in liver tissue (p=0.0001)

(Figure 10D). The minor T allele was associated with decreased DBH expression. The

Ct difference between the means was 1.1, reflecting a 2-fold change in expression between the groups. Similarly, a significant effect of rs1611115 genotype on DBH expression was observed in the lung (p=0.02) with the T allele again associated with nearly 2-fold lower DBH expression (Figure 10E). Expression was not significantly associated with rs1611115 genotype in brain tissue (p=0.6), as expected from the absence of a discernible effect of rs1611115 on allelic mRNA expression.

Figure 10. DBH mRNA levels in liver, lung and brain grouped by rs1108580 (A-C) or rs1611115 (D-F) genotype. The data represent the mean (n=2) of qRT-PCR cycle thresholds standardized to PGK1 mRNA as the house keeping gene. In liver (A), there was a significant difference between groups (p=0.03) and between AA and GG genotypes for rs1108580 (*p=0.005) but no association in lung (B) (p=0.3) or brain (C) (p=0.6). For rs1611115, there was a significant effect in liver (D) (**p=0.0001), and lung (E) (p=0.02) but no association in brain tissue (F) (p=0.6).

Continued

Figure 10 continued

Since both rs1611115 and rs1108580 affected DBH mRNA expression in liver, I examined the combined effect of both SNPs on overall expression. A stepwise linear regression to test the effect of demographics (race, age and gender) on expression indicated that these variables were not significant (Table 4) and were excluded as covariates.

Table 4. Demographics for liver samples used in qRT-PCR

Demographic Group (n) Mean DBH Pearson ±SD Expression Correlation (p (ΔCt)* value) † Sex Male (25) 6.6 ± 1.0 0.1 Female (32) 6.2 ± 1.0 Race Caucasian (49) 6.4 ± 1.0 0.5 Other (8) 6.4 ± 0.9 Age 62.3 ±13.2 0.3

*The mean PGK1 Ct was subtracted from the mean DBH Ct. †As p values for the Pearson correlation were all >0.05, no variables were included as covariates in the analysis.

Figure 11. DBH mRNA expression in livers measured with qRT-PCR grouped by number of minor alleles of rs1108580 and rs1611115. A one-way ANOVA showed an overall significant difference between the groups (p=0.01).

With increasing number of minor alleles per tissue, DBH mRNA levels decreased stepwise from 1-4 minor alleles, (Figure 11) (ANOVA p=0.01) reflected by increasing

ΔCt values. This result demonstrates that the two variants need to be considered together when addressing DBH associated phenotypes.

Verification of DBH-Specific Amplification

As DBH expression in the liver has not previously been reported, it was important to confirm this mRNA sequence aligned with DBH was not an artifact of non-specific

48 amplification. PAH (phenylalanine hydroxylase) is another copper-dependent mono- oxygenase, and is abundant in the liver. There is some sequence homology between the two. To test the overlap, I amplified a region of liver cDNA, including rs1108580 and sequenced the product. Then I aligned the resulting sequence to reference DBH mRNA sequence and PAH mRNA respectively using Clustal W 2.1

(www.ebi.ac.uk/Tools/msa/clustalw2/). As demonstrated in Figure 12, the product completely aligns with DBH, whereas only small regions align with PAH, and there are large gaps in the alignment. I confirmed that my primers were specifically amplifying

DBH, and not PAH or any other gene.

cDNA GACCCCAGAAGGCCTGACCCTGCTTTTCAAGAGGCCCTTTGGCACCTGCGACCCCAAGGA 103 DBH GACCCCAGAAGGCCTGACCCTGCTTTTCAAGAGGCCCTTTGGCACCTGCGACCCCAAGGA 480 ************************************************************

cDNA TTACCTCATTGANGACGGCACTGTCCACTTGGTCTACGGGATCCTGGAGGAGCCGTTCCG 163 DBH TTACCTCATTGAAGACGGCACTGTCCACTTGGTCTACGGGATCCTGGAGGAGCCGTTCCG 540 ************ ***********************************************

cDNA GTCACTGGAGGCCATCAACGGCTCGGGCCTGCAGATGGGGCTGCAGAGGGTGCAGCTCCT 223 DBH GTCACTGGAGGCCATCAACGGCTCGGGCCTGCAGATGGGGCTGCAGAGGGTGCAGCTCCT 600 ************************************************************

cDNA GAAGCCCAATATCCCCGAACCGGAGTTGCCCTCAGACGCGTGCACCATGGAGGTCCAAGC 283 DBH GAAGCCCAATATCCCCGAACCGGAGTTGCCCTCAGACGCGTGCACCATGGAGGTCCAAGC 660 ************************************************************

cDNA --GGCA------CCTGCGACCCCAA------100 PAH CGGGCAGCGAAGTGGTGCCTCCTGCGTCCCCCACACCCTCCCTCAGCCCCTCCCCTCCGG 240 **** ****** **** *

cDNA ------GGATTACCT----CATT--GANGACGGCACTGTCCACTTG------134 PAH CCCGTCCTGGGCAGGTGACCTGGAGCATCCGGCAGGCTGCCCTGGCCTCCTGCGTCAGGA 300 * * **** *** * * * ** *** ** * **

cDNA ---GTCTACGGGA-----TCCTG------149 PAH CAAGCCCACGAGGGGCGTTACTGTGCGGAGATGCACCACGCAAGAGACACCCTTTGTAAC 360 * * *** * * ***

cDNA ------GAGG------AGCC------GTT------160 PAH TCTCTTCTCCTCCCTAGTGCGAGGTTAAAACCTTCAGCCCCACGTGCTGTTTGCAAACCT 420 **** **** ***

cDNA ------CCGG-----TCACT----GGAGGCCATCA-----AC 182 PAH GCCTGTACCTGAGGCCCTAAAAAGCCAGAGACCTCACTCCCGGGGAGCCAGCATGTCCAC 480 ** * ***** ** **** ** **

cDNA --GGCA------CCTGCGACCCCAA------100 PAH CGGGCAGCGAAGTGGTGCCTCCTGCGTCCCCCACACCCTCCCTCAGCCCCTCCCCTCCGG 240 **** ****** **** *

cDNA ------GGATTACCT----CATT--GANGACGGCACTGTCCACTTG------134 PAH CCCGTCCTGGGCAGGTGACCTGGAGCATCCGGCAGGCTGCCCTGGCCTCCTGCGTCAGGA 300 * * **** *** * * * ** *** ** * **

cDNA ---GTCTACGGGA-----TCCTG------149 PAH CAAGCCCACGAGGGGCGTTACTGTGCGGAGATGCACCACGCAAGAGACACCCTTTGTAAC 360 * * *** * * ***

cDNA ------GAGG------AGCC------GTT------160 PAH TCTCTTCTCCTCCCTAGTGCGAGGTTAAAACCTTCAGCCCCACGTGCTGTTTGCAAACCT 420 **** **** ***

cDNA ------CCGG-----TCACT----GGAGGCCATCA-----AC 182 PAH GCCTGTACCTGAGGCCCTAAAAAGCCAGAGACCTCACTCCCGGGGAGCCAGCATGTCCAC 480 ** * ***** ** **** ** **

cDNA ------GGCTCGGGC------CTGCAGATGGGG------203 PAH TGCGGTCCTGGAAAACCCAGGCTTGGGCAGGAAACTCTCTGACTTTGGACAGGAAACAAG 540 **** **** ** ** **

cDNA ------CTGCAG-----AGGGTGC------AGCTCCT------223 PAH CTATATTGAAGACAACTGCAATCAAAATGGTGCCATATCACTGATCTTCTCACTCAAAGA 600 ***** * ***** * ** **

cDNA -GAAGC------CCAA--TATCCC------CGAACCGGA- 247 PAH AGAAGTTGGTGCATTGGCCAAAGTATTGCGCTTATTTGAGGAGAATGATGTAAACCTGAC 660 **** **** *** * **** **

Figure 12. Alignment of liver cDNA with PAH and DBH reference sequence.

Association of DBH mRNA Levels in the Liver with Phenotypes in Inbred Mouse

Strains

To determine whether DBH expression levels were associated with physiological phenotypes consistent with differences in sympathetic tone, I queried the Jackson

Laboratory database of inbred mouse strains. This database provides mRNA expression data in murine macrophage, brain, and liver tissues, as well as measurement of a wide range of phenotypes. Dbh mRNA was detectable in all three tissue types, and all available phenotypes were included in the analysis (not limited to sympathetic phenotypes). Expression of hepatic Dbh mRNA was significantly associated with several phenotypes at p≤0.001 (Table 5), whereas no significant results were observed at p≤0.001 for Dbh expression in macrophages or brain. High Dbh expression was also associated with increased R wave electrocardiogram amplitude (Berthonneche et al. 2009), a putative marker of increased myocardial hypertrophy in humans (Figure 13).

Figure 13. Electrocardiogram readings measured across 14 strains of mice age 7-13 weeks, using the IOX 1.7.0 and ECG-Auto 1.5.7 software (EMKA Technologies) compared to Dbh expression in liver, measured with probe 10024413494. Each point represents the average of one strain of mice.

Inverse correlations included measures related to increased body size, weight, and length, whereas positive correlations involved lean mass. On the other hand, hepatic Dbh mRNA was positively correlated with increased food and water intake and triglycerides.

Hepatic Dbh mRNA correlated positively with time spent in the center of the elevated plus maze and time near the walls in a fear conditioning chamber.

Table 5. Correlation between phenotype and Dbh liver mRNA expression

Pearson Chr Phenotype p value* R location Grip Strength 0.000030 –0.92 2:27173449 Percent Lean 0.00011 0.79 2:27174217 Percent Fat 0.00011 –0.79 2:27174217 Body Weight day 120 (60d post-exposure) 0.00017 –0.840 2:27174217 Food Intake 0.00019 0.92 2:27173449 Body Weight day 0 0.00020 –0.84 2:27174217 Triglycerides 0.00021 0.87 2:27173449 Total Tissue Mass 0.00028 –0.87 2:27173449 Water Intake 0.00040 0.90 2:27173449 R-wave Amplitude 0.00047 0.81 2:27174217 Body Length 0.00048 –0.81 2:27173449 Protein phosphatase 1 regulatory (inhibitor) subunit 1B (PPP1R1B), relative fluorescence 0.00053 0.89 2:27173449 intensity in cerebral cortex Weight of Fat Tissue 0.00056 –0.73 2:27174217 Percent Fat 0.00059 –0.71 2:27174217 Grip Strength 0.00064 –0.84 2:27173449 Elevated plus maze 0.00070 0.76 2:27174217 CCAAT/enhancer binding protein (C/EBP), alpha 0.00071 0.86 2:27173449 (Cebpa), relative mRNA abundance Total Tissue Mass 0.00081 –0.79 2:27172582 Fucosyltransferase 9 (FUT9), relative 0.00083 0.79 2:27174217 fluorescence intensity in cerebral cortex Total Tissue Mass 0.00087 –0.77 2:27172582 Body Weight 0.00090 –0.77 2:27174217 Total Body Area 0.00091 –0.78 2:27173449 Total Tissue Mass 0.0010 –0.76 2:27173449 Time spent near walls in fear conditioning 0.0010 0.85 2:27172582 chamber *The significance cutoff was set at p≤0.001

rs1108580 and Splicing

It has been hypothesized that rs1108580 affects splicing since it is located in the

last base of exon 2; however this has not been tested experimentally. As a first step, I

used ESEfinder, an in silico prediction tool to determine whether this SNP would be

predicted to change splicing. The sequence containing the ancestral G allele, resulted in a

score of 11.6, while substitution of the A allele decreased the score to 8.2 (Cartegni et al.

2003; Smith et al. 2006) suggesting there may be a decreased propensity for splicing in the presence of the variant. A second prediction method resulted in a score of 1.00 with the G allele and 0.95 with the A allele (Reese et al. 1997). A third method demonstrated that the introduction of the variant A allele resulted in a decrease of the splicing score from 88 to 76 (Shapiro et al. 1987). Other groups have used this method to predict changes in splicing and possible exon skipping for other genes however; these predictions have not been experimentally validated (Arcos-Burgos et al. 2010).

I used an RT-PCR assay to determine whether there was retention of the downstream intron 2. To target this region, a forward primer in exon 2 and a reverse primer in intron 2 were used, with a predicted 235 bp product if the intron were retained.

Whereas this primer set successfully amplified gDNA, in the majority of the cDNA samples, amplification was undetectable, with Ct values above 35. A small percentage of samples did show amplification of the product containing part of intron 2; however these amplicons were not correlated with rs1108580 genotype, and may represent heteronuclear RNA. Alternatively, failure to splice out intron 2 could have caused

54 nonsense mediated RNA decay, resulting in only sporadic detection of splice variants.

While rs1108580, located at the exon 2/intron 2 boundary, affected DBH mRNA expression in all organs tested, I was unable to ascertain whether rs1108580 affects splicing as previously suggested (Cubells et al. 1998; Wood et al. 2002).

Clinical Association Studies with Human Phenotypes and Diseases

To explore phenotypic associations in human subjects over a wide array of clinical diagnoses for rs1611115 and rs1108580, we performed a PheWAS with these two SNPs. PheWAS studies to date have been performed with different collections of data with comprehensive phenotypic and outcome data, including the use of de-identified electronic health record (EHR) data (Denny et al. 2010; Pendergrass et al. 2011;

Hebbring et al. 2013). For this PheWAS study we had clinical and genotype information on 3,035 subjects from the Geisinger Clinic MyCode biorepository (Gottesman et al.

2013) and a replication sample of 4,027 from the Marshfield Personalized Medicine

Research Project (PMRP).

The Geisinger Study. Of the 482 phenotypes tested for each SNP in the Geisinger dataset, the single SNP analysis using an additive model revealed four significant associations (p < 0.01) with rs1108580 and three for rs1611115 at a cutoff of p<0.01

(Table 6). Notably, the minor allele of both SNPs, rs1611115-T (p=0.0002, OR=0.43,

95% CI=0.28-0.66) and rs1108580-A (p=0.007, OR=0.67, 95% CI=0.50-0.90) was associated with reduced risk of angina pectoris (code 413.9). In addition, rs1611115-T was significantly associated as a risk allele for diabetes (code 250.6) (p=0.002, OR=1.9,

95% CI=1.28-2.92). The most significant association observed for SNP 1108580-A was

with intervertebral disc disorders (p=3.1E-05) with a negative direction of effect. This may be related to clinical observations in which patients with congenital DBH deficiency have hyperflexible joints.

Table 6. PheWAS analysis of individual SNPs, using an additive model in the Geisinger Study

SNP Coded Beta SE Odds 95% CI pval ICD-9 ICD-9 Description Cases Ctrls Allele Ratio Code (minor) rs1108580 A -0.78 0.19 0.46 0.32-0.66 3.1E-05 722.52 Intervertebral disc disorders: 71 2964 degeneration lumbar or lumbosacral intervertebral disc rs1611115 T -0.84 0.22 0.43 0.28-0.67 0.0002 413.9 Angina pectoris: unspecified angina 103 2932 pectoris rs1611115 T 0.66 0.21 1.9 1.3-2.9 0.002 250.6 Diabetes mellitus: diabetes with 48 2987 neurological complications rs1108580 A -0.69 0.24 0.50 0.32-0.80 0.004 780.39 Convulsions 43 2992 rs1108580 A -0.36 0.13 0.70 0.54-0.90 0.007 784 General symptoms: symptoms 131 2904

57 involving head and neck rs1108580 A -0.40 0.15 0.67 0.50-0.89 0.007 413.9 Angina pectoris: unspecified 103 2932 Angina pectoris rs1611115 T 0.60 0.23 1.8 1.2-2.9 0.008 789.01 Other symptoms involving 44 2991 abdomen and pelvis: abdominal pain, right upper quadrant

Data is sorted by p value (unadjusted), filtered at a cutoff of 0.01. The following abbreviations are used: SE, standard error; CI, confidence interval; and Ctrls, controls.

In the dominant model, only SNP rs1611115-T was associated with reduced risk

of angina pectoris, for the same diagnosis code (413.9), and in the same direction as in

the additive model (p=0.001, OR=0.44, 95% CI=0.26-0.73). Also using a dominant

model, rs1611115-T emerged as a risk factor for asthma (code 493) with a large

estimated effect size (p=0.007, OR=3.1, 95% CI=1.36-7.16). These results need to be

understood in the context of MAF (rs1611115 ~20% and rs1108580~45%, respectively)

and effect size (4-fold reduction of mRNA compared to ~2-fold, respectively, as judged

from the AEI ratios in liver).

The molecular genetics results in this study indicated that rs1611115 and

rs1108580 need to be considered together. In the Geisinger study, seven phenotypes

were associated with statistically significant interactions between the two SNPs (LRT

p≤0.01) (Table 7). The combined effects of both SNPs were significantly associated

(LRT p=0.001) with substantially reduced risk of myocardial infarction (code 410.1)

even though the number of cases was low. At a lower level of significance, a potential

risk effect with chronic ischemic heart disease (code 414.9) (full model p=0.01) was

demonstrated for the interaction term for the two SNPs. Additional risk effects were

observed for metabolic syndrome and hypoglycemia (Table 7). Significant interaction was also observed for obesity (code 278.1) (full model p=0.01) with a potentially protective effect. Although opposite observations regarding liver Dbh mRNA effects on body mass and obesity were observed in the mouse model, this may be a difference between localized adiposity and general fat distribution.

Table 7. PheWAS analysis using an interaction model, in the Geisinger study fullV1 fullV2 fullV1.V2 fullMod LRT ICD-9 ICD-9 Description Cases Ctrls Beta Beta Beta pval pval Code 2.3 0.80 -3.5 0.0008 0.001 410.1 Acute myocardial 11 3024 infarction: acute myocardial infarction of other anterior wall 1.2 0.43 -1.3 1.5E-04 0.004 110.1 Dermatophytosis: 47 2988 dermatophytosis of nail -3.5 -0.39 2.3 1.1E-05 0.01 706.1 Diseases of 16 3019 sebaceous glands: other acne -3.1 -0.21 1.7 1.8E-06 0.01 277.7 Other and 28 3007 unspecified disorders of metabolism: dysmetabolic syndrome X -1.9 0.031 0.98 2.8E-44 0.01 251.1 Other disorders of 78 2957 pancreatic internal secretion: other specified hypoglycemia 2.3 1.3 -1.7 3.1E-06 0.01 278.1 Overweight 18 2017 obesity: localized adiposity -0.48 -0.32 0.42 5.3E-50 0.01 414.9 Ischemic heart 283 2752 disease: chronic ischemic heart disease

The model uses logistic regression and an interaction term to test the interaction of the two SNPs. V1 is rs1611115 and V2 is rs1108580. This allowed us to identify if the full model including the interaction was significant, as well as if the full model results were significantly different when compared to the reduced model using a likelihood ratio test (LRT). Results with p≤0.01 are reported, and sorted by increasing LRT p-values. The following abbreviations are used: fullMod, full model; and Ctrls, controls.

The Marshfield Study. To replicate these findings, all ICD-9 codes within any ICD-9 category that showed an association p < 0.01 in the initial PheWAS analysis were also tested in the Marshfield EHR. It is feasible to register replication across EHR data within an ICD-9 category, even if the exact same ICD-9 code does not show replication of an association. While the two SNPs did not replicate with the same direction of effect for the main effect models (additive, dominant, and recessive), use of the two-SNP interaction model yielded similar results, with reduced risk for myocardial infarction,

(code 410.9) (full model p=0.0033) showing a strong protective effect with 13 cases and

4,015 controls. We also found a protective effect with angina pectoris (code 413.9) (full model p=0.038) in 309 cases and 3,719 controls.

The Jackson Heart Study (JHS). To specifically test cardiovascular events, and increase cohort diversity, we used data from the JHS. Among the 2,378 JHS participants, 62 incident MIs and 91 total CHD events were observed. Association of rs1611115 and rs1108580 with incident MI and CHD was tested by Cox Proportional Hazards (survival) analysis. Neither SNP on its own was significantly associated with incident MI

(rs1611115 p=0.11, rs1108580 p=0.077); however when considered together, there was a significant association between MI and the two SNPs (p=0.039). Both SNPs were also significantly associated with CHD when considered together (p=0.033). As with the

PheWAS analysis, there was a protective effect of the minor allele for both SNPs, yielding a relative risk of 0.73 (95% CI 0.54-0.99) for each additional copy of a minor allele at either SNP. This results in a cumulative relative risk for four copies of 0.28.

Available data did not allow analysis of the relationship of these variants with either incident or prevalent angina.

We also tested the association of blood pressure, hypertension status, BMI and

RR interval (ECG reading) with the SNPs in the JHS; however neither SNP was significant on its own, or in the additive model (Table 8).

Table 8. Genotype association tests with an additive model in JHS

rs1611115 rs1108580 Sum of rs1611115 T alleles and rs1108580 A Continuous alleles outcomes N p value Beta (SE) p value p value for N Beta (SE) N sum of SNPs Beta (SE) ln(BMI) 2750 0.45 0.0058 (0.0077) 3011 0.93 0.00061 (0.0067) 2749 0.61 0.0022 (0.0042) ln(SBP) 2587 0.20 0.0064 (0.0050) 2838 0.55 0.0026 (0.0043) 2586 0.32 0.0028 (0.0027) DBP 2587 0.91 0.046 (0.40) 2838 0.98 0.010 (0.35) 2586 0.95 0.013 (0.22) RR interval 2744 0.21 -0.0070 (0.0056) 3007 0.78 -0.0013 (0.0048) 2744 0.34 -0.0029 (0.0031) Dichotomous outcomes N (cases) p value OR (95% CI) N (cases) p value OR (95% CI) N (cases) p value for OR (95% CI) sum of SNPs Hypertension 2731 (1747) 0.22 1.11 (0.94-1.31) 2993 (1907) 0.091 1.13 (0.98-1.31) 2730 (1746) 0.15 1.07 (0.98-1.17) status CHF 2416 (125) 0.31 0.90 (0.66-1.23) 2655 (136) 0.33 0.90 (0.70-1.17) 2415 (125) 0.30 0.95 (0.81-1.12) CHD _all 2408 (91) 0.079 0.67 (0.42-1.05) 2641 (99) 0.063 0.71 (0.49-1.02) 2407 (91) 0.033 0.77 (0.61-0.98) CHD_MI 2379 (62) 0.11 0.63 (0.36-1.12) 2611 (69) 0.077 0.67 (0.43-1.04) 2378 (62) 0.039 0.73 (0.54-0.99) 62

Pharmacogenomic Evaluation of Antihypertensive Responses (PEAR)

As these SNPs were demonstrating an effect in liver, we examined the role of norepinephrine in this organ. Norepinephrine is known to stimulate glucose release from the liver, and result in increased heart rate and blood pressure. We performed an analysis in the PEAR cohort to test the association of DBH SNPs with glucose levels and blood pressure in response to treatment with atenolol. Although no associations were significant in either racial group with either SNP, there was a trend toward a greater blood pressure response in subjects with the minor allele of rs1611115 (Figure 14).

These results suggest that subjects with lower DBH expression have a greater change in blood pressure in response to treatment. For rs1108580, in Caucasians the association with DBP was p=0.93 and SBP, p=0.74; and in African-Americans, DBP, p=0.57 and

SBP, p=0.74. A model accounting for the total number of minor alleles did not improve the association with SBP (Caucasians p=0.44, African-Americans p=0.29) or DBP

(Caucasians p=0.38, African-Americans p=0.61).

Figure 14. Blood pressure change in response to atenolol treatment in Caucasians (A) and African-Americans (B) grouped by rs1611115 genotype. In Caucasians the association with DBP, p=0.19 and SBP, p=0.22; and in African-Americans, DBP, p=0.13 and SBP, p=0.49.

We also tested whether these SNPs were associated with a change in blood glucose levels however; again the results were not significant for either group or genotype (Figure 15). For rs1108580 the p value was 0.90 in Caucasians and 0.83 in

African-Americans.

Figure 15. Association of rs1611115 genotype with blood glucose response following atenolol treatment in Caucasians (p=0.52) (A) and African-Americans (p=0.25) (B).

Health and Retirement Study (HRS)

We were able to test the effects of these SNPs on blood pressure and glucose regulation (assessed by HbA1c levels) in a second cohort, the Health and Retirement

Study (HRS). In this population race significantly correlated with both outcomes

(HbA1c and BP); therefore, African-Americans and Caucasians were analyzed separately. In Caucasians there was a significant difference in systolic blood pressure between 0 and 1 vs. 2 or more minor alleles, after normalizing the data (natural log) and using a t-test assuming equal variances (p=0.028). After converting the difference between the groups from log form, the difference is 1 mm Hg, in the direction consistent

65 with our hypothesis with more minor alleles associated with lower blood pressure; however this is not a biologically significant difference.

I also tested the association between HbA1c levels and DBH SNPs. In 8,527

Caucasian subjects, there was not a significant effect of rs1611115 on HbA1c (p=0.38).

There was a combined effect of both SNPs (p=0.02), however this was due to the effect of rs1108580 (p=0.002). Although the effect was statistically significant, the differences between the genotype groups were small, with each additional copy of the minor allele adding 0.04 to the HbA1c level and therefore unlikely to be biologically relevant.

As body mass associated with Dbh expression in mouse liver, I tested the effect of

DBH genotypes on BMI in this cohort. In 240 Caucasian subjects, there was no effect of rs1108580 (p=0.33) or rs1611115 (p=0.97). Also, the total number of minor alleles was not associated with BMI (p=0.51) using a regression analysis.

As DBH SNPs have previously been implicated in neurocognitive phenotypes, I tested their effect on a total recall score in a memory task. In 1,658 African-American subjects, there was not a significant effect of each SNP on its own (rs1108580 p=0.55, or rs1611115 p=0.31) or combined (p=0.38). However, in a larger sample, of 10,810

Caucasian subjects, there was a significant effect of both SNPs combined, using a regression analysis (p=0.001). The equation of the line was y = -0.12x + 10.1, with R² =

0.91 (Figure 16). This result demonstrates that lower DBH expression (increasing number of minor alleles) associates with a decrease in memory score.

Figure 16. Linear regression between total number of minor alleles (rs1108580 and rs1611115) and mean total recall score on a memory task in 10,810 Caucasian subjects. The equation of the line was y = -0.12x + 10.1, with R² = 0.91

Genotyping in Chimpanzees to Test Conservation

To determine whether these SNPs were conserved I genotyped them in chimpanzee gDNA. All 41 samples were homozygous for the major (ancestral) allele: rs1108580, GG; rs1611115, CC; and rs2519143 (in high LD with rs1611115), GG. The results of the rs1611115 genotyping assay were confirmed via Sanger sequencing (Figure

17). Another study genotyped rs1108580 and rs2519143 in 19 chimpanzees and also found that all animals were homozygous for the major allele (Healy et al. 2004). This suggests that these SNPs are unique to humans.

Figure 17. Sanger sequencing results in chimpanzee gDNA using a reverse primer. The SNP rs1611115 is marked with a vertical line indicating genotype GG (or CC on the forward strand).

Dbh Knockout Mouse Model with Human BAC

To test the effect of rs1611115 under a more controlled set of conditions, we used a transgenic mouse model generated by David Weinshenker. This line incorporates a human BAC encompassing DBH, expressing either the C or T allele at rs1611115 crossed onto a Dbh knockout background.

Following RNA extraction, I measured the RNA integrity via Bioanalyzer

(Agilent Technologies) for a subset of 12 samples for each tissue type and genotype group. The mRNA was of high quality, reflected by an average value of 8.7 (on a scale from 1 to 10). The average RINs per tissue type were: brain, 9.1 ± 0.4; adrenal, 9.7 ± 0.2; liver, 8.0 ± 1.1; heart, 9.0 ± 0.1; and lung 7.7 ± 1.7. 68

In the cDNA preparations, I tested for gDNA contamination in a random subset of

8 samples from each tissue type, using human DBH primers designed to amplify gDNA.

In a positive gDNA control, the Ct was 21.8, but the lowest Ct measured in these mouse

cDNA samples was 32, with a range from 32-36, with undetectable expression in the rest of the samples. I can conclude there is little to no gDNA contamination. If there is gDNA present in the cDNA preparation, it is at least 1000 fold less than in a gDNA sample.

I measured DBH mRNA expression with qRT-PCR, normalizing expression to

SARDH expression; both measured in duplicate. Then relative DBH expression levels

were compared between each of the lines containing the C or T allele. In the brain

samples, we also measured murine Slc6a2 (norepinephrine transporter) expression in

duplicate to confirm isolation of the LC.

I found a statistically significant difference in mRNA expression between the C

and T groups in all tissue types (p<0.02) with higher expression in the T group compared

to the C group. There were unequal variances between groups in liver, lung and adrenal samples. Using a two-sided t-test, the p values were: liver, p=0.019; heart, p=1.3x10-5;

brain, p=2.1x10-5; lung, p=0.0032; and adrenal, p=0.00017. Of note, these results are in

the opposite direction from what is observed in humans. In patients, the T allele is

associated with lower levels of plasma DBH and in my studies in human tissues, I

observed lower levels of DBH mRNA expression in samples containing the T allele. In

this mouse model the DBH cycle thresholds are lower (indicative of higher expression) in

the T group than the C group in all tissues (Figure 18). This difference may be due to

69 differential expression of transcription factors in mice vs. humans. Another possibility is that we are detecting an artifact of the construct. In 2010, the O’Connor group reported higher luciferase activity in PC-12 cells transfected with the construct containing the T allele (Chen et al. 2010) similar to what we observed in the mouse model, however, they later reported the opposite effect, with constructs containing the T allele having lower luciferase activity (Pasha et al. 2013). The effect of rs1611115 genotype on mRNA in these mice was confirmed via in situ hybridization by Dr. Patti Szot, again reporting decreased DBH in the C allele group.

Figure 18. DBH mRNA expression levels in different tissues harvested from transgenic mice. 71

2.7 Discussion

With a gene such as DBH, in which multiple common SNPs influence variation in

sympathetic (and possibly central) nervous system function, it is critical that we fully

understand the underlying genetics. This area requires further clarification. These results

demonstrate profound effects of common DBH variants on expression in sympathetically

innervated organs, modulating clinical phenotypes responsive to peripheral sympathetic

tone. The molecular genetics results of this study indicate that rs1611115 and rs1108580

both need to be considered in clinical association studies.

I have identified two key regulatory DBH genetic variants, rs1611115 and

rs1108580, that robustly reduce DBH mRNA expression in sympathetically innervated

organs, the liver and the lung, whereas in adrenals and brain the effect was undetectable

(rs1611115) or less pronounced (rs1180580), suggesting the mode of action is

predominantly peripheral. Acting together, rs1611115-T and rs1108580-A reduce allelic

DBH mRNA expression up to 11-fold in the liver, potentially accounting for previous

observations that rs1611115 genotype accounts for 35-52% of interindividual variability

in circulating DBH levels (Zabetian et al. 2001; Garland et al. 2007; Mustapic et al. 2007;

Bhaduri et al. 2008). Similarly, a genome-wide association study in human livers found a

significant association between rs2519143 (owing to its high LD with the promoter SNP

rs1611115 (Table 3; see also (Johnson et al. 2008)) and DBH expression in the liver

(p=2.31x10-18) (Schadt et al. 2008), supporting a strong effect of rs1611115 on DBH expression in the liver. The tissue specific effect of promoter region rs1611115 on transcription is similar to observations with promoter/enhancer variants in CETP (Papp et al. 2012) and CYP3A4 (Wang et al. 2011), while an effect of the exonic rs1108580

parallels results with TPH2 (Lim et al. 2007). The substantial level of DBH mRNA in

the liver, and to a lesser extent in the lung, indicates that DBH mRNA is transported from

sympathetic ganglia to nerve terminals in the target organ where DBH protein is released

together with norepinephrine. Initial experiments in human celiac ganglia indeed

revealed a high level of DBH mRNA expression. Preliminary results from others have

also demonstrated the presence of DBH mRNA in distal axons of sympathetic neuron

cultures (N. Gervasi et al., 2012, 10th Int. Catecholamine Symposium, abstract).

Combined, these results are consistent with rs1611115-T and rs1108580-A

affecting DBH mRNA abundance in the liver. While the AEI ratio results with SNP

rs77905 suggested the presence of at least one other regulatory variant affecting DBH

mRNA expression, I was unable to identify such variants by sequencing the immediate

DBH gene locus and 15 kb downstream of the 3′UTR and comparing the results to the

allelic expression data. It is possible that a distant regulatory variant exists, such as in

CYP2D6 (Wang et al. 2014), resulting in a long-range effect on transcription, a

hypothesis that needs to be further studied. Another possibility is an epigenetic effect.

Our approach of sequencing the gene region focused on detecting a cis-acting variant; however it would not provide information on differential methylation.

These results resolve a long-standing debate on the process by which the minor allele of rs1611115 is associated with strongly reduced circulating DBH activity. They also account for previous failures to resolve the promoter activity of rs1611115 in reporter gene assays in cell lines, out of the context of the appropriate tissue environment.

Pasha et al. (2013) examined the impact of DBH haplotypes on reporter-construct transcriptional activity driven by the DBH proximal promoter in PC-12 cells, a rat

pheochromocytoma line derived from the adrenal medulla. However, some of the data

presented in the report appear either to be inconsistent with prior reports by the same

laboratory, also in PC-12 cells, (Chen et al. 2010; Chen et al. 2011), or internally

inconsistent. The haplotypes are composed of different alleles at rs1989787 (C-2073T)

and rs1611115 (C-970T), both of which have previously been shown to exert functional

effects on transcription (Chen et al. 2010; Chen et al. 2011). In the most recent report,

the rank order of transcriptional activity associated with the haplotypes is C-C > C-T > T-

C > T-T. However in both prior reports (Chen et al. 2010; Chen et al. 2011), the rank order was C-T > C-C > T-T > T-C, regardless of various haplotype backgrounds defined by other SNPs.

Although DBH expression in the liver is lower than in the adrenals, genetic effects in the liver are multiplied because of the large organ size. The liver further plays a key role in filtering norepinephrine from the blood during its passage but also releases norepinephrine, so that a portion of circulating norepinephrine stems from the liver

(Aneman et al. 1996). Release from other sympathetically innervated organs, including the lung, may further account for the large genetic effect on circulating DBH and norepinephrine levels.

The absent or small effect of rs1611115 and rs1108580 in brain is critical in interpreting clinical association results regarding CNS related phenotypes. Tissue- specific differences are frequently observed for promoter/enhancer variants that rely on the expression of requisite transcription factors (Wang et al. 2011; Papp et al. 2012).

While norepinephrine and dopamine are reported not to penetrate the blood-brain-barrier substantially, it is possible that these circulating hormones have an indirect effect on the

CNS through the pineal gland, which is outside the barrier. They may also act at sites

directly connected to the CNS, such as the vagus nerve, as its activity is regulated by

peripheral catecholamines, travels to the brain, and has profound effects on the LC

(Fukuda et al. 1987; Ruffoli et al. 2011).

Because of the profound effects of rs1611115 and rs1108580 on peripheral sympathetically innervated neurons, I have performed an initial survey study on associations of DBH genotype and activity with phenotypes in mice and human clinical studies. Previous clinical association studies have largely focused on either variant alone

(Wood et al. 2002; Freire et al. 2006; Kohnke et al. 2006) but need to be reexamined in light of the predominantly peripheral site of action and two-SNP interactions. I present here evidence for the involvement of DBH in a range of peripheral sympathetic phenotypes, including metabolic disorders, cardiovascular disease, and asthma.

Effect of Dbh mRNA Levels in Mouse Liver on Phenotypes

I found a significant association between Dbh mRNA expression in the liver and body size, metabolic processes, and cardiovascular phenotypes. While such broad phenotypic associations are expected from the multifaceted role of norepinephrine throughout the body, the association of increased hepatic Dbh mRNA levels with R wave amplitude, a prognostic indicator of myocardial hypertrophy, was of particular interest.

Significant negative correlation of high hepatic Dbh mRNA on body mass and percent

fat, and an inverse effect on percent lean mass, indicates a strong impact on metabolic

activity. Here, Dbh activity could be related to adverse outcomes with beta blockers,

such as metabolic syndrome and type II diabetes. The finding that measures of fear

conditioning and anxiety also ranked significantly suggest the possibility that peripheral

Dbh activity might affect CNS phenotypes. They may also indicate effects reflected indirectly by the levels of expression.

Effect of DBH Genotype on Clinical Phenotypes

A PheWAS approach exploiting electronic medical records was employed to survey multiple phenotypes that could be affected by DBH variants. In the Geisinger study, the dominant model revealed increased asthma risk associated with rs1611115-T.

Drugs that mimic the action of norepinephrine, such as albuterol, are useful in asthma treatment because they relax bronchial smooth muscle. With decreased levels of DBH, and less norepinephrine, rs1611115-T is a plausible risk factor for asthma, and should be investigated for responsiveness to adrenergic asthma medication. Moreover, rs1611115-

T is a potential risk factor for type II diabetes and the interaction model showed a significant effect on hypoglycemia. This is noteworthy because Dbh knockout mice exhibit resistance to changes in glucose levels in an insulin tolerance test (Ste Marie et al.

2003). Also, the use of beta blockers has been associated with increased risk of type II diabetes (Cooper-DeHoff et al. 2013).

Consistent associations were observed with cardiovascular phenotypes. Angina pectoris was the most significantly associated phenotype with both rs1611115-T and rs1108580-A. Again along expectations for a reduced sympathetic tone, the minor alleles of both DBH variants were protective, paralleling the effect of beta-blocker therapy.

We then tested phenotype associations stipulating an interaction between rs1611115 and rs1108580. This approach revealed a possible protective effect of the minor alleles against myocardial infarction in the Geisinger MyCode study, a result that was replicated in the Marshfield PMRP study. We speculate that the combined effect of

76 two or more minor variants in rs1611115 and rs1108580 is needed to reach a threshold effect and thereby reveal this association. In an ethnically distinct population, the

Jackson Heart Study, we tested the effects of rs1611115 and rs1108580 in African-

Americans. While each SNP was not significant on its own, an increasing protective effect against coronary heart disease was attributable to an increasing number of the minor alleles.

Limitations of the Study

While we consider the molecular genetics findings on the effects of rs1611115 and rs1108580 on expression of transcripts encoding DBH to be strongly supported by the presented results, extrapolation to cellular function and clinical phenotypes is more exploratory. Confounding factors include the distinct biological effects of norepinephrine and dopamine that vary inversely with DBH genotype. However, it is noted that most of the phenotypes observed in Dbh-/- mice are attributed to a lack of norepinephrine rather than increased dopamine. This is demonstrated by DOPS rescue, which restores norepinephrine without affecting dopamine levels (Thomas et al. 1998).

DBH mRNA expression was too low in kidney, small intestines and heart, to obtain an accurate estimate of genotype effects. Whether mRNA transport into these tissues is lower, or less DBH protein is expressed in nerve terminals, remains to be determined. The Human Protein Atlas shows medium to high protein staining in liver, testis, lung, stomach, colon and brain. In brain tissues, little DBH mRNA is found except in the LC, indicating that DBH protein is transported to noradrenergic nerve terminals, as reported (Dahlstrom 1971). Based on these results, we hypothesize that rs1611115 and

rs1108580 affect DBH expression in all peripheral sympathetically innervated organs, but

this remains to be studied.

We are aware of the multiple hypothesis testing burden incurred from the number

of tests in the PheWAS analyses. However we reason the stringency of a Bonferroni correction is inappropriate here. First, unlike a GWAS using an unbiased set of SNPs, we performed this PheWAS using two well characterized SNPs with established effects in target tissues. Further, as many of the diagnoses used in the PheWAS are correlated, not all outcomes here are independent, reducing the number of independent variables. The

PheWAS results require follow up and evaluation through additional association testing and experimental work for full validation. The main limitation is that the clinical trials are observational and not designed a priori to test the specific associations observed here, with relatively small number of cases. Therefore, all observations require follow-up studies but highlight the potential of assessing sympathetic tone through DBH variants in pathophysiology and therapy of sympathetic dysregulation.

Chapter 3: Regulation in the 3′ Untranslated Region of Alpha-synuclein (SNCA)

3.1 Introduction and Background

Aggregation of alpha-synuclein, encoded by the gene SNCA also referred to as

NACP (non-amyloid component of plaques) is a main component of Lewy bodies, the hallmark of Parkinson’s disease (PD) and also implicated in Alzheimer’s disease.

However, it is also expressed in non-disease states in various brain regions. It regulates dopamine homeostasis through modulation of the dopamine transporter (Lee et al. 2001), regulation of dopamine synthesis (Perez et al. 2002) and presynaptic regulation of dopamine storage and release from vesicles (Yavich et al. 2004). It is expressed in dopaminergic neurons and enriched in presynaptic terminals (Lavedan 1998) and is involved in neuronal development, differentiation, maturation, and neuroplasticity

(Maroteaux et al. 1991).

Dopamine plays a key role in drug abuse as it is involved in the reward pathway critical to addiction. The causes of addiction are complex, involving both environmental and genetic factors. Between 40 and 60% of an individual’s risk for drug addiction can be attributed to genetic factors (Kendler et al. 1994; Bierut et al. 1998; Kendler et al.

2000; Agrawal et al. 2008). Genome-wide association studies (GWAS) have identified possible SNPs associated with drug addiction; however, these results are often difficult to

replicate, and relative risk factors are small (Johnson et al. 2009). Therefore more in- depth studies of target genes are vital.

Use of animal models has generated evidence for SNCA involvement in cocaine and alcohol addiction. SNCA mRNA expression is increased in alcoholic monkeys

(Walker et al. 2006) and rats exposed to cocaine (Brenz Verca et al. 2003). Alcohol preferring rats have higher SNCA mRNA expression in the hippocampus and a SNP in the 3'UTR resulted in longer mRNA half-life (Liang et al. 2003; Liang et al. 2006). In alcohol preferring rats, two SNPs in the 3ʹUTR, (+679 and +807), were associated with higher SNCA gene and protein expression in the hippocampus (Liang et al. 2003).

Humans also have elevated mRNA and protein SNCA levels following alcohol

(Bonsch et al. 2004; Bonsch et al. 2005) or cocaine abuse (Qin et al. 2005; Mash et al.

2008). However SNCA mRNA and protein levels are not increased in subjects with excited delirium (Mash et al. 2003).

Excited delirium is a condition triggered by drug use, mental illness or head trauma and can result in sudden death. Symptoms include: anxiety, paranoia, disorientation, aggressive or violent behavior, hyperthermia, surprising physical strength and insensitivity to pain (Samuel et al. 2009). Cocaine induced psychoses was first described by Post et al. in 1975, with excited delirium resulting from cocaine intoxication

defined ten years later (Wetli et al. 1985).

Specific genetic mutations and variants in SNCA have been studied in relation to neurocognitive disorders. One SNP commonly implicated in genetic association studies is rs356165, located in the 3ʹUTR. In case-control comparisons in European populations, the G allele was found to be more frequent in patients with PD (Myhre et al. 2008;

Schmitt et al. 2012) and associated with age of onset (Cardo et al. 2012). However, no association was found between genotype and PD in Chinese subjects (Hu et al. 2010).

The Rep1 variant is a polymorphic microsatellite repeat located approximately 10 kb upstream of the SNCA gene. The presence of the 259 bp Rep1 allele is reported to be protective, and was associated with lower SNCA mRNA levels in the brain (Linnertz et al. 2009). Patients homozygous for this allele had lower blood levels of SNCA protein

(Fuchs et al. 2008) and this allele was associated with a reduced risk of PD (OR=0.86)

(Maraganore et al. 2006). The risk allele for this variant may either be the 261 bp repeat or the 263 bp repeat. Individuals with the Rep1 261/261 genotype had an increased risk of PD (Mamah et al. 2005) and the 263 bp allele has also been associated with PD

(Maraganore et al. 2006; Myhre et al. 2008). Other studies found no association between

Rep1 and PD (Mueller et al. 2005; Hu et al. 2012), serum SNCA levels (Hu et al. 2012),

or alcoholism (in American Indians) (Clarimon et al. 2007).

The SNP rs356219 is used as a haplotype tagging SNP in the 3ʹ region. It has

been associated with PD (Goris et al. 2007; Myhre et al. 2008), higher plasma levels of

SNCA (Mata et al. 2010) and changes in mRNA levels in substantia nigra and cerebellum

(Fuchs et al. 2008). The minor allele of rs17016074, located between two polyA signals,

has been associated with lower SNCA mRNA levels in human tissues (Linnertz et al.

2009). However, the minor allele caused increased luciferase expression in transfection

experiments with SHSY5Y cells (Sotiriou et al. 2009).

Lewy body dementia (LBD) is a neurodegenerative disorder characterized by the

progressive death of dopaminergic and cholinergic neurons and the accumulation of

Lewy bodies. Lewy bodies are also found in brains of patients with Parkinson’s disease

and Alzheimer’s disease. SNCAIP, located at 5q23.2 with 10 exons, encodes α-

synuclein-interacting protein (SNCAIP, synphilin1, Sph1) containing several protein

interaction domains and binds to α-synuclein (SNCA) (Engelender et al. 1999). While

this makes SNCAIP a candidate gene in Lewy body dementias, few studies have

addressed this issue. No frequent mutations have been detected in the coding sequence in

patients with Parkinson’s disease (Bandopadhyay et al. 2001; Farrer et al. 2001) and the

only coding variants are rare, for example rs372883371 (Val44Ala, T131C) (Satoh et al.

2002). Similarly, neither SNCAIP haplotypes (Farrer et al. 2001) nor intronic variants rs6875859 (intron 2) (Bandopadhyay et al. 2001), rs2242224 (intron 3) and rs2290987

(intron 5) significantly associated with Parkinson’s disease (Maraganore et al. 2003).

SNCAIP variants did not affect the incidence of familial (Bandopadhyay et al. 2001;

Marx et al. 2003; Wirdefeldt et al. 2003) or sporadic (Satoh et al. 2002; Maraganore et al.

2003; Marx et al. 2003; Myhre et al. 2008) Parkinson’s; nevertheless, regulatory variants can reside at large distance from the gene locus (Wang et al. 2014) and often evade detection. In this study I have measured allelic mRNA expression in human post-mortem brain tissue to determine whether a variant present within the gene locus is preferentially affecting transcription or mRNA processing.

3.2 Materials and Methods

Cohort

DNA and RNA were extracted from human postmortem brain tissue samples.

Human prefrontal cortex (Brodmann area 46) samples were provided by the Miami Brain

Endowment Bank (University of Miami, Miami, FL). DNA was provided for ~450

82 samples total, a combination of subjects who abused cocaine, those with excited delirium, and controls. Cases were confirmed using toxicology reports and other drug-related measures such as arrest and hospital records. Controls have no history of psychiatric disorders or drug use prior to death. A total of 12 anterior cingulate cortex samples from subjects with Lewy body dementia were provided by the Ohio State University

Neurodegenerative Disease Brain Tissue Repository (Buckeye Brain Bank).

Quantitative Real-Time reverse transcriptase PCR (qRT-PCR)

SNCA and SNCAIP mRNA expression were measured on the 7500 Fast Real-

Time PCR system, using Fast SYBR Green Master Mix. Samples were run in duplicate and normalized to the housekeeping gene PGK1.

DNA Sequencing Using Ion Torrent

Gene regions were PCR amplified, (total of ~3 kb) covering the entire annotated

3′UTR, beginning ~50 bp upstream and ending ~300 bp downstream. The sample was fragmented using Covaris shearing, and barcoded libraries were prepared using NEBNext

Fast DNA Library Prep Set for Ion Torrent, the Ion Xpress Barcode Adapters Kit and the

Ion One Touch Template Kit. The library was sequenced on the Ion Torrent, using the

Ion Torrent Ion PGM Sequencing Kit (Life Technologies).

3.3 Results

Determination of Allelic SNCA mRNA Ratios as an Indicator of Regulatory

Polymorphisms

I measured allelic expression ratios as aforementioned using the SNaPshot assay and marker SNP rs356165 located in the 3′UTR with the red primers indicated in Figure

21. The first four samples in Figure 19 demonstrated AEI measured at rs356165. Cutoffs for significance were determined by three standard deviations of the genomic DNA (>1.3 and <0.8, indicated by the solid lines). Most samples (except two, far right) had ratios less than one, indicating that the functional SNP is likely in LD with this marker.

Figure 19. Allelic SNCA mRNA expression ratios measured at marker rs356165 in brain cDNA samples. The values are plotted as a ratio of the major (G) over the minor (A) allele peak heights and standardized to the measured genomic DNA allelic ratio. Samples marked with an * are heterozygous for rs17016074, also located in the 3ʹUTR. The horizontal lines demark the threshold for significant AEI (0.8-1.3). Each bar is the mean AEI measurement from one subject. The brain sample number includes LBD if the subjects had Lewy body dementia.

Genotype Association with AEI

To identify SNCA variants responsible for the observed AEI, 25 variants were genotyped in all prefrontal cortex samples and tested for association with AEI. The resulting p values are reported in Table 9. The only SNP demonstrating a significant association following Bonferroni correction was rs17016074. The four samples with the largest AEI values were heterozygous for this SNP (Figure 19) and all of the other samples were homozygous, making it a potential functional SNP. I tested the effect of

additional variables on AEI ratios, and found no effect of sex (p=0.44), age (p=0.87), or

case status (p=0.13).

Table 9. SNPs genotyped and genotype association with AEI using the F statistic

Marker F-Test P F-Test Bonf. P rs17016074 0.000031 0.0008 rs7678651 0.018 0.48 rs11097234 0.018 0.48 rs3857059 0.056 1 rs3775446 0.23 1 rs356221 0.25 1 rs2583958 0.27 1 rs2583985 0.27 1 rs2583988 0.27 1 rs2619364 0.27 1 rs2736995 0.27 1 rs2737006 0.27 1 rs2736990 0.42 1 rs356200 0.48 1 rs974711 0.48 1 rs1812923 0.48 1 rs2301134 0.48 1 rs7684318 0.48 1 rs7687945 0.48 1 rs2583988 0.57 1 rs356219 0.71 1 rs748849 0.71 1 rs2197120 0.71 1 rs2737030 0.71 1 rs2619363 0.73 1

rs17016074 Genotype Association with mRNA Expression

As rs17016074 was associated with allelic expression, I sought to determine whether it was also associated with overall SNCA mRNA expression (Figure 20).

However, no significant associations were found between overall SNCA expression and genotype; however, the number of heterozygous carriers was relatively small, with a

MAF of 0.018 in Caucasians and 0.11 in African-Americans in this cohort.

Figure 20. SNCA mRNA levels measured in the 3′UTR, grouped by rs17016074 genotype. There was not a significant difference (p=0.46) between samples homozygous for the major allele (GG) and minor allele carriers.

AEI Measured at Putative Functional SNP (rs17016074)

As rs17016074 is located in the 3'UTR, it could also be used as a marker SNP to measure AEI. If this SNP were functional, we expect AEI in all heterozygous samples.

However, none of the samples demonstrated AEI indicated in Figure 22 with black bars.

Therefore a second primer set, amplifying SNCA from the opposite strand was used to

87 measure AEI (white bars). Again, none of the samples demonstrated AEI measured at rs17016074 (Figure 22). Note that samples MB63, 87 and 96 demonstrated significant

AEI when measured at rs356165 but not when measured at marker rs17016074 using the green primers indicated in Figure 21.

Figure 21. Map of marker SNPs, primer locations and polyA consensus sequences in the 3′UTR of SNCA.

Figure 22. Allelic SNCA mRNA expression ratios measured at marker rs17016074 in brain cDNA samples. The values are plotted as a ratio of the major (G) over the minor (A) allele peak heights and standardized to the measured genomic DNA allelic ratio. All samples are heterozygous for rs17016074. The horizontal lines demark the threshold for significant AEI (0.8-1.2). Each bar is the mean AEI measurement from one subject; black bars indicate amplification from the forward strand, and white bars from the reverse.

AEI Measurement Using PCR Product Containing Both rs356165 and rs17016074

Next, I sought to determine whether these differences in AEI were due to the region amplified, or the marker SNP used to measure AEI. Therefore, I amplified a 667 bp region containing both rs356165 and rs17016074. AEI was measured at both SNPs if the sample was a compound heterozygote (middle section), rs356165 alone (far left) or rs17016074 alone (far right) (Figure 23). The magnitude of AEI is comparable when measured at two different SNPs in the same sample (R2=0.65). All AEI ratios measured at rs356165 were less than 1 (with the exception of sample 212) so the inverse value is plotted in Figure 23, for ease of comparison to AEI measured at rs17016074. Significant

AEI, with ratios above 1, was measured in all samples at rs17016074, with the exception of sample 229. These data demonstrate that rs17016074 is either the functional SNP, or in high LD with it. The PCR product used for measurement of AEI in Figure 22 is closer to the start of the 3′UTR and likely encompasses both the long and short forms and does not show an imbalance, whereas the PCR product used in Figure 23 amplifies a region that differs between the two UTR lengths, thus demonstrating AEI.

Figure 23. Allelic SNCA mRNA expression ratios measured at marker rs356165 (white bars) and/or rs17016074 (black bars) in PCR product containing both SNPs. The inverse was taken if the ratio was below 1, indicated with a *. The threshold for significant AEI is 1.3 and each bar is the mean AEI measurement from one subject (4 replicates).

Sequencing of Outlier

One sample (#229) was heterozygous for rs17016074 but not did demonstrate significant AEI. To determine if there were additional variants responsible for this discrepancy, I sequenced 2 kb of the 3′UTR, beginning ~60 bases upstream via Sanger sequencing. The sample was confirmed to be heterozygous for rs17016074. I found that it was also heterozygous for rs6842093, (MAF=17% in YRI, and in high LD with

91 rs17016074) as well as the rare SNP rs144511886 (MAF=0.6% in AFR). This rare SNP may lie on the opposite haplotype and cancel out the effect of the functional SNP.

Sequencing for Additional Variants

To test for the presence of additional variants affecting AEI, I selected five AEI positive samples (55, 63, 87, 96, and 148) and two AEI negative samples (43, 79) and sequenced the 3′UTR. The resulting sequences revealed eight variants not previously genotyped, and I tested the association of AEI with genotype (Table 10). The SNP rs17016074 again showed a significant association with AEI measured at rs356165, as well as AEI measured with either rs17016074 or rs356165. A second SNP, rs6842093, was also significant in this analysis, at the same level as rs17016074 which can be explained by their perfect LD (D′=1 and R2=1) in African populations (Johnson et al.

2008). Located in the last intron, about 50 bp from the start of the last exon and 587 bp from rs17016074, it was heterozygous in all samples with AEI, and homozygous in all sequenced samples without AEI. However, genotyping of samples from Figures 19 and

23 revealed that a sample with AEI (#181) was homozygous CC for rs6842093.

Therefore rs17016074 remains the strongest candidate functional variant as it is heterozygous is the most AEI positive samples. In populations of European ancestry, rs17016126 is the only SNP reported to be in LD with rs6842093 (D′=1 and R2=1, and in

Africans, D′=0.96 and R2=0.80) and it is 45,326 bases away in a long intron. This long haplotype suggests rapid rise in the population, a possible indicator of positive evolutionary selection. This SNP, rs6842093, was also heterozygous sample #229, the outlier from Figure 23 discussed previously that was heterozygous for rs17016074 but lacking AEI.

Table 10. Chi squared p value for genotype association with instances of AEI

AEI at AEI at SNP either SNP rs356165 rs17016074 0.0047 0.014 rs6842093 0.0047 0.014 rs35214126 (INS) 0.064 0.12 rs1045722 0.17 0.12 rs3857053 0.17 0.12 rs377356638 0.21 0.22 rs356165 0.35 1 rs182722435 0.54 1 rs17016072 1 1 rs201351609 1 0.54

Search GWAS Catalog for Association Results

To determine whether there were phenotypic effects of any variants in SNCA and potentially distinguish the effect of rs17016074 from other SNPs in LD with it, I performed a query of the Catalog of Published Genome-Wide Association Studies

(http://www.genome.gov/gwastudies/) (Table 11). Five SNPs were reported to have a significant association with PD (rs356219, rs6532194, rs356220, rs2736990, rs11931074). The SNP rs356219 is a known 3′UTR haplotype tagging SNP. All of the

SNPs except rs6532194 had a D′ value above 0.9 with each other, and with rs17016074

(although this was not a GWAS result) despite being up to 40 kb apart. This suggests that a long haplotype block is having an effect or a single SNP in LD with others is the functional SNP. The R2 value with rs17016074 is 0.1 or lower, reflecting the differences in minor allele frequencies between these SNPs.

Table 11. LD of SNPs from GWAS Search and rs17016074

HG18 SNP Proxy R2 D′ Distance coordinate Major Minor MAF rs17016074 rs6532194 0.099 0.78 133624 90999925 C T 0.49 rs17016074 rs11931074 0.084 1 7763 90858538 T G 0.35 rs17016074 rs356219 0.058 1 9677 90856624 G A 0.27 rs17016074 rs356220 0.056 1 5938 90860363 T C 0.26 rs17016074 rs2736990 0.026 1 31263 90897564 G A 0.14 rs356219 rs356220 0.96 1 3739 90860363 T C 0.26 rs356219 rs11931074 0.70 1 1914 90858538 T G 0.35 rs356219 rs2736990 0.45 1 40940 90897564 G A 0.14 rs356219 rs6532194 0.30 0.92 143301 90999925 C T 0.49 94 rs6532194 rs356220 0.29 0.91 139562 90860363 T C 0.26

rs6532194 rs11931074 0.20 0.62 141387 90858538 T G 0.35 rs6532194 rs2736990 0.10 0.79 102361 90897564 G A 0.14 rs356220 rs11931074 0.67 1 1825 90858538 T G 0.35 rs356220 rs2736990 0.47 1 37201 90897564 G A 0.14 rs2736990 rs11931074 0.32 1 39026 90858538 T G 0.35

Genotype Association with Case Status

Several SNPs were genotyped in the cocaine/excited delirium/control cohort

(Table 9) and tested for association between genotype and case status. In the analysis

using all subjects, a multinomial logistic regression was run, but genotype was not

significant in any of the genetic models. In Caucasians the MAF for rs17016074 was

0.031 and 0.14 for African-Americans. There was no association between case status in

African-Americans and rs17016074 genotype (p=0.48) nor an association in the

combined racial groups with either cocaine addicts, or those with excited delirium (Table

12).

Table 12. rs17016074 association with cocaine or ED status

Genotype Control Cocaine ED GG 162 (89.5%) 157 (85.79%) 90 (84.91%) GA 18 (10%) 25 (13.66%) 16 (15.09%) AA 1 (0.55%) 1 (0.55%) 0 (0.00%) p‐value 0.32 0.33 adj p‐value* 0.86 0.95 *adjusted for race and gender

Alpha-synuclein Interacting Protein (SNCAIP)

SNCAIP mRNA levels were measured by qRT-PCR with primers spanning an exon boundary to avoid amplification of gDNA. The mean PGK1 cycle threshold (Ct) was subtracted from the mean SNCAIP Ct, yielding a ΔCt value. In the TT genotype

95 group (n=12) the mean ΔCt was 7.3 ± 0.3, and the CT group (n=26) 7.5 ± 0.3. Figure 24 shows rs11326 was not significantly associated with ΔCt and genotype (p=0.66). The mean ΔCt for all samples was 7.4 ± 1.2, indicating robust expression sufficient for AEI analysis.

Figure 24. SNCAIP mRNA levels in brain grouped by rs11326 genotype. The data represent the mean (n=2) of qRT-PCR cycle thresholds standardized to PGK1 as the housekeeping gene. There was no significant difference (p=0.66) between the TT genotype group (n=12, ΔCt = 7.3 ± 0.3) and the CT group (n=26, ΔCt = 7.5 ± 0.3).

In the dbSNP database, most coding region variants in SNCAIP have a MAF less than 1%, or the MAF is not reported. The only common variant is rs11326, located in the 96

3′UTR, with a MAF of 10%. For measurement of allelic mRNA ratios, rs11326 served as the marker SNP. With three standard deviations of the gDNA ratio as the threshold, I considered an allelic mRNA ratio of <0.75 and >1.33 as significant evidence of AEI.

One tissue, sample #80, displayed significant AEI with a ratio of 0.74 (Figure 25) close to the set threshold.

Figure 25. Allelic mRNA expression imbalance quantified in nineteen postmortem brain tissue samples heterozygous for rs11326. The values are plotted as a ratio of T (major allele) over C (minor allele) and standardized to the measured genomic DNA allelic ratio. The horizontal lines demark the threshold for significant AEI (0.75-1.33). Each bar is the mean AEI measurement from one subject. Donors were either controls, or suffered from Lewy body dementia (LBD) or Alzheimer’s (Alz).

3.4 Discussion

The results of this study identify the presence of a cis-acting regulatory variant in the SNCA gene region, as identified by measurement of AEI in the 3′UTR. I demonstrate that this AEI is associated with the presence of rs17016074. There are differences in allelic expression depending on the region amplified, suggesting that this introduces the use, or marks the presence of different length 3′UTRs. When AEI was measured at the 5′ end of the 3′UTR, there was no imbalance, however a PCR product encompassing both rs17016074 and rs356165, did demonstrate AEI.

Allelic mRNA ratios were used as an indicator of regulatory polymorphisms, and four of the 39 samples (10%) demonstrated robust AEI. Genotyping 25 SNPs revealed a significant association with one, rs17016074. Almost all of the ratios (major/minor) were in the same direction, indicating that the functional SNP is in LD with the marker SNP rs356165. Indeed, the LD between rs17016074 and rs356165 is D′=0.99 and R2=0.05-

0.09 (using the 1000 genomes data, depending on population). No other variables were associated with AEI, including cocaine use, sex, or age.

While there was no association of rs17016074 genotype with mRNA expression measured at the 5′ end of the 3′UTR, it is likely that this SNP affects the expression of transcripts with an extended 3′UTR. I did not detect AEI measured at rs17016074 in the

5′ end of the 3′UTR, however when measured using a longer PCR product, encompassing both rs17016074 and rs35615, I did detect AEI. This indicates the presence of multiple

3′UTRs, with the end between rs17016074 and rs35615.

Differences in AEI detection depending on primer location have important implications for measurement of allelic expression in other genes as well. Results from studies in which no AEI is detected may be confounded by alternate UTR usage, although this is less likely to be a factor in the coding region. Therefore, it is important to measure AEI with multiple primer sets using different marker SNPs. For example, although only small DBH allelic ratios for were observed in brain samples, we can have confidence in the results as comparable measurements were obtained using four different marker SNPs throughout the gene region (exons 2, 3, 9 and the 3′UTR.)

Sequencing samples with AEI for the presence of additional variants yielded an association with rs6842093, in high LD with rs17016074. SNPs that were significantly associated with PD phenotypes in the GWAS catalog were also in LD with rs17016074.

I can conclude that rs17016074 itself is functional, or is part of a functional regulatory haplotype.

Another potential mechanism for regulation in the 3′UTR is modification of microRNA binding. To assess this, I first searched all microRNAs predicted to bind to

SNCA; however none of the results from targetscan.org (Lewis et al. 2005), nor microrna.org (Betel et al. 2008) overlapped with rs17016074. MiRanda

(http://www.mirbase.org/) allows the input of a region of interest, so I searched for regions that would have altered binding sites with the introduction of the minor allele of rs17016074. However, there was no difference in predicted microRNAs between the minor or major allele and the microRNAs in that region (miR-450 and miR-513) contain a mismatch at rs17016074 for both alleles. In mice, two miRs, miR-7 and miR-153, have

100 been shown to bind to the 3′UTR and decrease expression of SNCA. Primarily expressed in neurons, miR-7 represses translation of SNCA (Junn et al. 2009) while miR-153 induces mRNA degradation (Doxakis 2010).

Our results demonstrate that it is unlikely that a frequent variant in SNCAIP robustly regulates the level of mRNA expression in the brain. This is reflected in the lack of significant deviation from one, when measuring allelic mRNA ratios. However, there are several tissues close to the cutoff value of 1.33; a small ratio can indicate effects acting by other mechanisms, such as the introduction of splice variants.

Evidence suggests that functional genetic variants are unlikely to be found in the coding region of this gene. In addition to previous studies citing intronic variants, the only two SNPs with significant associations in the Phenotype-Genotype Integrator

(PheGenI) database, (a compilation of genome-wide association study data) are located in intronic (rs10519701) and intergenic regions (rs10519705).

101

Chapter 4: Population-Specific Regulation of Catechol-O-Methyltransferase (COMT)

4.1 Introduction and Background

The enzyme catechol-O-methyltransferase, encoded by COMT, catalyzes the conversion of dopamine, and other catecholamines to an inactive metabolite via the addition of a methyl group. It may be present in either a soluble or membrane bound form. The membrane-bound form is more prevalent in the brain (Matsumoto et al. 2003),

however most clinical studies focus on the soluble protein, detectable in blood samples.

While COMT has been well-studied, not all of the genetic variability has been explained. Hundreds of studies investigating polymorphisms in COMT linked with neurocognitive disorders have focused on the coding SNP rs4680 (G>A; Val158Met) in exon 4. For thorough reviews see (Hosak 2007; Dickinson et al. 2009; Tunbridge 2010;

Witte et al. 2012), as space does not permit enumeration of these studies. A transition from G (major allele) to A (minor allele) results in the conversion of valine to methionine.

The transcripts with the minor allele produce a heat-labile soluble protein leading to a three- to four-fold reduction in enzymatic activity in liver and peripheral blood (Lotta et al. 1995; Lachman et al. 1996). However, the impact of rs4680 on the full-length membrane-bound COMT protein – the dominant form in the brain – remains unclear.

102

In search of additional genetic factors in COMT, several studies have implicated

additional SNPs either alone, or in combination with rs4680, including rs737865 (A>G)

in intron 1 of the membrane-bound form, and rs165599 (A>G) in the 3′UTR. A COMT

haplotype consisting of the major allele G (Val) at rs4680, the minor allele C at rs737865,

and the minor allele G at rs165599 was found to be significantly correlated with

schizophrenia (Shifman et al. 2002). Risperidone treatment had increased efficacy in

both African-American and Caucasian schizophrenic patients with the GG genotype for

rs165599 however; there was no effect of rs4680 (Fijal et al. 2009). COMT has

transcripts with multiple 3′UTR lengths, and this SNP, rs165599 resides in one of the

extended 3′UTRs with lower expression downstream of the second of three polyA sites,

and is therefore only present in a subset of the transcripts (Jugurnauth et al. 2011).

However, there is no difference in luciferase activity between the A and G alleles for

rs165599 (Jugurnauth et al. 2011).

Overall, minorities exhibit higher COMT activity than Caucasians. In Asian

populations 43% demonstrate high RBC-COMT activity versus 22% of Caucasians

(Rivera-Calimlim et al. 1984). This may partially be explained by differences in rs4680 frequency between the populations. While the frequency of the low activity Met allele is

~50% in Caucasians, it is 18-25% in Chinese (Li et al. 1997; Xie et al. 1997), 27% in

Taiwanese (Chen et al. 1997), and 29% in Japanese (Kunugi et al. 1997). African-

Americans also have higher COMT activity than Caucasians (McLeod et al. 1994) but not in all studies (Chen et al. 2004). Again, this may be partially explained by a lower

103

frequency of the Met/Met genotype (9% in Kenyans, 31% in Caucasians) (McLeod et al.

1998).

Genetic variation in COMT accounts for some cases of phenotypic differences between African-Americans and Caucasians. For example, African-American women

are at higher risk of developing uterine leiomyomas. This may partly be explained by an

association between the Val/Val genotype and these estrogen-dependent tumors (Al-

Hendy et al. 2006). African-Americans also have higher rates of preterm birth and the C

allele of rs4818 (promoter region) is associated with increased risk in African-Americans, but not Caucasians, with no effect of rs4680 (Thota et al. 2012). Cocaine use also shows a genetic association. A haplotype consisting of the minor A (Met) allele for rs4680 and the major allele A in rs737865 was associated with cocaine abuse (Lohoff et al. 2008) and cocaine induced paranoia (CIP) (Ittiwut et al. 2011) in African-Americans.

As there are many clinical phenotypes, as well as protein effects of rs4680, it could be assumed that this SNP also has an effect at the mRNA level. However, this does not appear to be the case. Allelic expression was measured using the marker SNPs rs4633 and rs4680 resulting in small ratios (average ~1.2 fold). There was no difference in AEI between males and females (Bray et al. 2003). It is important to note that the majority of samples in this study were Caucasian. When measured in lymphoblastoid cell lines, there was no significant AEI (Yan et al. 2002). Human COMT mRNA expression in the dorsolateral prefrontal cortex did not associate with rs4680 genotype

(Matsumoto et al. 2003; Chen et al. 2004; Tunbridge et al. 2004). One study on COMT mRNA expression in the brain did find a significant effect of a haplotype comprised of C

104

at rs737865, G at rs165599 and G at rs4680. They demonstrated that transcripts with this

haplotype, including the major G allele at rs4680, were expressed at a lower level than

those containing the Met allele (Bray et al. 2003), contrary to expectations since the G

allele has been associated with higher protein activity.

Taken together, the molecular genetics of COMT are unresolved even while

numerous clinical association studies utilize the aforementioned COMT SNPs. It is

critical to determine whether there are genetic differences between population groups,

and which polymorphism(s) affects the membrane-bound form, with the goal of

improving the power of clinical associations and enabling gene-gene interaction tests, likely to yield enhanced predictive accuracy.

4.2 Materials and Methods

See earlier chapters for techniques used in these studies; all experiments are performed as previously described.

4.3 Results

I selected rs4633 (in high LD with rs4680) and rs165599 (3′UTR) as marker

SNPs for allelic mRNA ratio measurements. First I measured AEI using rs165599 as the

marker SNP (Figure 26). Of the 20 samples analyzed, 14 demonstrated significant AEI,

with ratios above the threshold for significance (3 standard deviations of the gDNA

ratios, 1.2). All of the samples with the largest AEI values are African-American (right half, indicated with a *) with ratios ranging from 1.55-2.62. In the Caucasian samples, ratios ranged from 1.05-1.36. In all cases, the allelic ratios, calculated as major/minor

105 allele, were above one, indicating there were more transcripts containing the major allele in all of the samples. This also suggests that any functional variant is likely in LD with this SNP however; rs165599 itself is not functional, as significant AEI is not observed in all of the heterozygous samples.

106

Figure 26. Allelic expression ratios measured at marker SNP rs165599 in cDNA samples synthesized from RNA from prefrontal cortex samples. A * is used to indicate African- American samples.

As the first assay suggested that race may play a role, I used a second marker

SNP, rs4633, to measure AEI in both African-Americans and Caucasians. Almost half of the African-American samples demonstrated significant AEI (Figure 27A), while only

107 one Caucasian sample did (Figure 27B). AEI measured at rs4633 significantly associates with race (p=0.015). In African-Americans, samples with significant AEI ranged from

0.61-0.72, reflecting a 1.39-1.67 fold difference between the alleles. Overall, in both groups, there was a bias toward higher expression of the minor allele, with most of the ratios below one.

As rs4680 has been widely studied and implicated in a host of clinical associations, I hypothesized that this variant would be causing the differences in allelic expression. However, as rs4680 is in high LD with our marker SNP, it was heterozygous in most samples. Despite being heterozygous in all of the Caucasian samples, only one demonstrated AEI. Two of the African-American samples were not heterozygous for rs4680, (indicated by an x, Figure 27A) one had a ratio of 0.69 and the other, 0.87, while the rest of the samples were heterozygous. Of the 30 total samples heterozygous for rs4680, only 7 of them displayed AEI; therefore rs4680 is not the functional SNP affecting allelic mRNA expression.

108

Figure 27. Allelic expression measured at rs4633 in African-Americans (A) and Caucasians (B) in prefrontal cortex cDNA. Samples not heterozygous for rs4680 are indicated with an ‘x’.

The next question I wanted to address was whether this phenomenon, of AEI in

African-Americans, was exclusive to the brain. I measured AEI in the liver, using marker SNP rs4633 and found no robust AEI (Figure 28). Samples 18, 49 and 107 are

African-American (indicated with *) but do not demonstrate AEI. This result suggests that this mechanism is tissue-specific, and may be due to differences between the membrane bound and soluble forms of COMT.

109

Figure 28. Allelic expression measured at rs4633 in liver cDNA. African-American samples are indicated with a *.

As I determined liver samples did not demonstrate significant AEI, and rs4680 was not associated with instances of AEI in the brain, the next question was which 110 variant was causing the differences I did observe. To determine the functional SNP causing this AEI, I sequenced gDNA from AEI positive and AEI negative brain samples to determine genetic differences between the two groups. A functional SNP would be heterozygous in all of the AEI positive samples, and homozygous in all of the AEI negative samples. I narrowed the screen to African-American samples and selected the six samples with highest AEI, (far left, Figure 27A) as well as four samples that did not demonstrate significant AEI for the sequencing experiments.

Following PCR amplification and library preparation, gene regions were sequenced and variants called using Ion Reporter Software. Initially I sequenced intron

1, as this differs between the membrane bound and soluble forms and therefore a variant acting preferentially in the brain would likely reside in this region. However, no variants in this region associated with AEI and sequencing was expanded to the entire COMT gene region to include all introns and exons. Again finding no significant variants, I expanded the region to include all of TXNRD2, the gene immediately upstream. It had histone markings indicating the potential to act as an enhancer region based on the

ENCODE data. I also included part of ARVCF (downstream). In all, I fully sequenced a

100 kb region. No SNP fit the expected pattern of being heterozygous in all of the AEI positive samples and homozygous in AEI negative samples. The best candidates, rs3788316, rs5993863, and rs9606176, were heterozygous in three samples in the AEI positive group, one heterozygote in the AEI negative group, and the rest of the samples were homozygous.

111

As it was possible that the regulatory variant was located even farther away, I prioritized the next regions to sequence assessing SNPs within 500 kb for viability as candidates based on several criteria. First, I required that the SNP have a reported MAF in YRI (African) of 10-30%, as the estimated MAF based on the proportion of samples with AEI was ~25%. The SNP must also have a MAF less than 10% in CEU, as only one

Caucasian sample demonstrated AEI. Another criteria was LD with the marker SNP, as most of the AEI ratios were in the same direction, so I set the cutoff for D' to be greater than 0.7. Filtering the candidate SNPs by these parameters reduced the number of target

SNPs within 500 kb to 75. I was able to directly sequence, or sequence SNPs in LD, 57 of these candidates. Regions with evidence of histone modifications (enhancer/promoter markings) from the ENCODE database were also included when possible. Sequencing did not reveal any SNPs that fit the pattern of being heterozygous in all AEI positive samples, and homozygous in AEI negative samples.

4.4 Discussion

The results of this study demonstrate that COMT is under regulation and there is a genetic factor affecting allelic mRNA expression predominantly in African-Americans.

While I was unable to identify the specific variant causing this change, it is important to note that it is not well studied ValMet SNP rs4680, nor is it located in a 100 kb region surrounding the gene. Through the sequencing results, I was able to exclude hundreds of

SNPs as being the functional variant.

112

Several possible factors could account for the failure to identify the functional

variant. First, it is possible that multiple SNPs are acting together to cause the imbalance,

and our analysis focused on the identification of a single variant. There could be a

multiple SNPs, or an entire haplotype block causing the effect. While the Ion Torrent

variant caller identified a few regions with insertions or deletions, this is not its strongest

feature. The strength of the algorithm is in SNP identification. Therefore if an insertion

or deletion was causing this change, it may have been overlooked in the analysis.

Although a large gene region was sequenced, it is still possible that the functional

variant lies outside of the targeted region. A variant in CYP2D6 was identified in a

distant downstream enhancer region more than 100 kb away that increased transcription

(Wang et al. 2014). Theoretically, any SNP on chromosome 22 could be causing AEI.

Another explanation is differential methylation.

While AEI has previously been measured in brain RNA, our study found larger

fold changes (up to 2.6) compared to previous results of 1.2 fold (Bray et al. 2003). The

main difference between the studies was the sample demographics. The Bray cohort was

predominantly Caucasians, while I also included African-American subjects, and

measured the highest AEI values in those samples. Our AEI ratios in Caucasians were

comparable to those previously reported.

113

Chapter 5: Conclusions and Discussion

This project characterizes genetic variation in DBH, an enzyme critical to the conversion of dopamine to norepinephrine. Because of the broad importance of this pathway both centrally and peripherally, DBH genetics has been studied in great detail; yet, causative relationships and related phenotypes have remained enigmatic. The promoter SNP rs1611115 accounts for up to half of the variability in circulating plasma

DBH levels but the site of action of rs1611115 and contribution of other DBH variants remain unknown. Clinical association studies have focused on neurocognitive disorders, assuming the genetic effects on DBH plasma levels replicate throughout the body. In vitro studies have failed to establish a consistent mechanism for this variant, leaving this field of investigation on uncertain grounds. Further evidence for additional regulatory variants also remains unclear; clarifying these issues is imperative, given the pervasive biological effects of norepinephrine.

I have determined that rs1611115 and rs1108580 unexpectedly have primary effects in peripheral sympathetically innervated organs, particularly in the liver where

DBH mRNA appears to be transported from the celiac ganglia. Both SNPs have either no or little effect in brain and adrenal tissue. This result accounts for the strong effect of rs1611115 on circulating DBH protein levels and points to numerous phenotypes

114 associated with peripheral sympathetic tone. In mice, Dbh mRNA levels in the liver were linked with cardiovascular phenotypes; a demonstration of how clinically relevant phenotypes can be prioritized using an animal model. Application of validated causative variants to human GWAS data can establish significant association alleviating the burden of excessive multiple hypotheses testing, which I exploited in two PheWAS and one

GWAS dataset. Among the top results were cardiovascular phenotypes, diabetes, and asthma, consistent with the expectation of the influence on sympathetic tone. A further advance emerging from our detailed molecular genetics study is the ability for testing targeted interactions between two causative SNPs, indicating a strong protective effect against myocardial infarction in all three clinical studies, a rare example of potential SNP interactions with clinical implications. I contend that my study explores novel avenues, combining molecular genetics and exploiting large scale mouse and human genomics datasets. Lastly, this work contributes to the question of rare versus frequent variants with strong effects on intermediate phenotypes, with disease risk observable only under specific circumstances.

AEI ratios in SNCA reflect the presence of a cis-acting regulatory variant. While undetectable at the 5′ end of the 3′UTR, a larger product, likely encompassing a longer

3′UTR did show AEI. I demonstrate that this AEI is associated with the presence of rs17016074 and introduces the use, or marks the presence of different length 3′UTRs.

Also, I have demonstrated that a factor acting predominantly in African-Americans is regulating COMT and affecting allelic mRNA expression. The ValMet SNP rs4680 is

115

not responsible for these differences, nor are hundreds of other SNPs in a 100 kb region

around the gene.

This study resolves important issues that will guide future studies, as well as opening new avenues of inquiry, with the potential for developing a predictive clinical biomarker test. Evidence from the DBH study suggests the use of rs1108580 and rs1611115 as markers for myocardial infarction, with other sympathetic phenotypes requiring further study. SNCA SNPs are likely to have implications in Parkinson’s disease, and findings from this study could aid in disease prediction. Given the current structure of the American health care system, it is difficult to obtain an exact price for any health care costs. However, many of these costs can be estimated, leading to a range of dollar amounts, depending on the source. The estimated cost for genotyping a single

SNP in a clinically approved setting, ranges from about $100 (Compagni et al. 2008) to

$500 (Dubinsky et al. 2005).

While these costs may appear prohibitive, these biomarkers will have utility

beyond a single test, or even a biomarker panel, as advances are made in sequencing

technology and the healthcare field evolves. This progress is leading to a drop in prices,

making the inclusion of a patient’s entire genomic sequence in his/her electronic medical

record a real possibility. It is important to study and define the genetic risk factors now,

before whole genome sequencing becomes the standard of care. Establishing solid

evidence for the effect of these variants will aid in clinical implementation by physicians.

The human genome project and subsequent GWAS were expected to provide

evidence for the genetic basis of complex diseases however; genetic heritability for many

116

common diseases still is unexplained. Simply collecting the nucleotide sequence is not

sufficient, we must be able to interpret its meaning and identify where the risk alleles lie.

GWAS analyses examine millions of SNPs in hundreds or thousands of subjects,

comparing groups with and without a certain disease leading to a high threshold for

significance. Resulting effect sizes are usually small and not consistently validated in

subsequent studies. In addition, these results do not explain all of the genetic factors

leading to “missing heritability.” For example, cardiovascular disease shows a strong

genetic component, but the variants that define the inheritance pattern are still

incompletely described (Go et al. 2013).

There are several potential explanations for missing heritability. Some of these

phenotypes may be caused by rare variants, which can only be detected by vastly

increasing the sample size. Another factor is examination of heterogeneous populations.

It is possible that one disease is actually a collection of diseases. Refinement of

phenotype characterization could aid in more sensitive detection of genetic variants

causing that condition. Another explanation is epigenetic effects such as methylation that

would not be detected by GWAS, as they do not change the sequence. Studies that use

exome-capture are overlooking a host of polymorphisms with the potential to contribute

to disease. As I have demonstrated with rs1611115 in DBH and rs17016074 in SNCA,

regulatory variants are not required to change the coding sequence to have an effect.

Finally, I believe the field is advancing with the study of SNP-SNP and gene-gene interactions, moving beyond polygenic models that are simply additive. In the case of

DBH, I found novel associations with cardiovascular phenotypes by studying the effects

117

of the two variants together, as opposed to single SNP analysis that is commonly

performed in clinical association studies. Studies examining epistasis are also promising,

as evidenced by an interaction between SNPs in DAT and DRD2 contributing to risk of

cocaine overdose, with an odds ratio of 7, whereas the odds ratio for DRD2 alone was 3

(Sullivan et al. 2013). The interaction between genes and the environment is another aspect that must be considered.

I propose that the study of RNA expression, and variants that contribute to its regulation, is vital. The use of surrogate markers may hinder detection of a strong effect if the SNP is not in perfect LD with the functional SNP, and the effect becomes further diluted when added into a genetic model and combined with other variants. Identification of these variants can begin to form the basis to explain complex human disease.

Establishing evidence of these variants as biomarkers will be essential to effectively using results of whole genome sequencing in patients once this becomes a widespread practice.

118

References

Abe, M., Z. Wu, et al. (2005). "Association of dopamine beta-hydroxylase polymorphism with hypertension through interaction with fasting plasma glucose in Japanese." Hypertens. Res. 28(3): 215-221.

Agrawal, A. and M. T. Lynskey (2008). "Are there genetic influences on addiction: evidence from family, adoption and twin studies." Addiction 103(7): 1069-1081.

Al-Hendy, A. and S. A. Salama (2006). "Catechol-O-methyltransferase polymorphism is associated with increased uterine leiomyoma risk in different ethnic groups." J. Soc. Gynecol. Investig. 13(2): 136-144.

Alwan, A. (2011). Global status report on noncommunicable diseases 2010. Geneva, World Health Organization.

Anderle, P., C. U. Nielsen, et al. (2006). "Genetic variants of the human dipeptide transporter PEPT1." J. Pharmacol. Exp. Ther. 316(2): 636-646.

Aneman, A., G. Eisenhofer, et al. (1996). "Sympathetic discharge to mesenteric organs and the liver. Evidence for substantial mesenteric organ norepinephrine spillover." J. Clin. Invest. 97(7): 1640-1646.

Arcos-Burgos, M., M. Jain, et al. (2010). "A common variant of the latrophilin 3 gene, LPHN3, confers susceptibility to ADHD and predicts effectiveness of stimulant medication." Mol. Psychiatry 15(11): 1053-1066.

Bandopadhyay, R., R. de Silva, et al. (2001). "No pathogenic mutations in the synphilin-1 gene in Parkinson's disease." Neurosci. Lett. 307(2): 125-127.

Berthonneche, C., B. Peter, et al. (2009). "Cardiovascular response to beta-adrenergic blockade or activation in 23 inbred mouse strains." PloS one 4(8): e6610.

Betel, D., M. Wilson, et al. (2008). "The microRNA.org resource: targets and expression." Nucleic Acids Res. 36(Database issue): D149-153.

119

Bhaduri, N. and K. Mukhopadhyay (2006). "Lack of significant association between - 1021C-->T polymorphism in the dopamine beta hydroxylase gene and attention deficit hyperactivity disorder." Neurosci. Lett. 402(1-2): 12-16.

Bhaduri, N. and K. Mukhopadhyay (2008). "Correlation of plasma dopamine beta- hydroxylase activity with polymorphisms in DBH gene: a study on Eastern Indian population." Cell. Mol. Neurobiol. 28(3): 343-350.

Biaggioni, I., D. S. Goldstein, et al. (1990). "Dopamine-beta-hydroxylase deficiency in humans." Neurology 40(2): 370-373.

Bierut, L. J., S. H. Dinwiddie, et al. (1998). "Familial transmission of substance dependence: alcohol, marijuana, cocaine, and habitual smoking: a report from the Collaborative Study on the Genetics of Alcoholism." Arch. Gen. Psychiatry 55(11): 982-988.

Bonsch, D., V. Greifenberg, et al. (2005). "Alpha-Synuclein Protein Levels Are Increased in Alcoholic Patients and Are Linked to Craving." Alcohol. Clin. Exp. Res. 29(5): 763-765.

Bonsch, D., U. Reulbach, et al. (2004). "Elevated alpha synuclein mRNA levels are associated with craving in patients with alcoholism." Biol. Psychiatry 56(12): 984-986.

Bourdelat-Parks, B. N., G. M. Anderson, et al. (2005). "Effects of dopamine beta- hydroxylase genotype and disulfiram inhibition on catecholamine homeostasis in mice." Psychopharmacology (Berl). 183(1): 72-80.

Bray, N. J., P. R. Buckland, et al. (2003). "Cis-acting variation in the expression of a high proportion of genes in human brain." Hum. Genet. 113(2): 149-153.

Bray, N. J., P. R. Buckland, et al. (2003). "A haplotype implicated in schizophrenia susceptibility is associated with reduced COMT expression in human brain." Am. J. Hum. Genet. 73(1): 152-161.

Brenz Verca, M. S., A. Bahi, et al. (2003). "Distribution of alpha- and gamma-synucleins in the adult rat brain and their modification by high-dose cocaine treatment." Eur. J. Neurosci. 18(7): 1923-1938.

Cardo, L. F., E. Coto, et al. (2012). "A search for SNCA 3' UTR variants identified SNP rs356165 as a determinant of disease risk and onset age in Parkinson's disease." J. Mol. Neurosci. 47(3): 425-430.

Cartegni, L., J. Wang, et al. (2003). "ESEfinder: A web resource to identify exonic splicing enhancers." Nucleic Acids Res. 31(13): 3568-3571.

120

Chen, C. H., Y. R. Lee, et al. (1997). "Association study of NlaIII and MspI genetic polymorphisms of catechol-O-methyltransferase gene and susceptibility to schizophrenia." Biol. Psychiatry 41(9): 985-987.

Chen, J., B. K. Lipska, et al. (2004). "Functional analysis of genetic variation in catechol- O-methyltransferase (COMT): effects on mRNA, protein, and enzyme activity in postmortem human brain." Am. J. Hum. Genet. 75(5): 807-821.

Chen, Y., G. Wen, et al. (2010). "Human dopamine beta-hydroxylase (DBH) regulatory polymorphism that influences enzymatic activity, autonomic function, and blood pressure." J. Hypertens. 28(1): 76-86.

Chen, Y., K. Zhang, et al. (2011). "Human Dopamine beta-Hydroxylase Promoter Variant Alters Transcription in Chromaffin Cells, Enzyme Secretion, and Blood Pressure." Am. J. Hypertens. 24(1): 24-32.

Chun, L. S., A. Samii, et al. (2007). "DBH -1021C-->T does not modify risk or age at onset in Parkinson's disease." Ann. Neurol. 62(1): 99-101.

Clarimon, J., R. R. Gray, et al. (2007). "Linkage disequilibrium and association analysis of alpha-synuclein and alcohol and drug dependence in two American Indian populations." Alcohol. Clin. Exp. Res. 31(4): 546-554.

Compagni, A., S. Bartoli, et al. (2008). "Avoiding adverse drug reactions by pharmacogenetic testing: a systematic review of the economic evidence in the case of TPMT and AZA-induced side effects." Int. J. Technol. Assess. Health Care 24(3): 294-302.

Cooper-DeHoff, R. M., S. T. Bird, et al. (2013). "Antihypertensive drug class interactions and risk for incident diabetes: a nested case-control study." J. Am. Heart Assoc. 2(3): e000125.

Craig, S. P., V. J. Buckle, et al. (1988). "Localization of the human dopamine beta hydroxylase (DBH) gene to chromosome 9q34." Cytogenet. Cell Genet. 48(1): 48-50.

Cubells, J. F., D. P. van Kammen, et al. (1998). "Dopamine beta-hydroxylase: two polymorphisms in linkage disequilibrium at the structural gene DBH associate with biochemical phenotypic variation." Hum. Genet. 102(5): 533-540.

Dahlstrom, A. (1971). "Axoplasmic transport (with particular respect to adrenergic neurons)." Philos. Trans. R. Soc. Lond. B. Biol. Sci. 261(839): 325-358.

121

Denny, J. C., M. D. Ritchie, et al. (2010). "PheWAS: demonstrating the feasibility of a phenome-wide scan to discover gene-disease associations." Bioinformatics 26(9): 1205-1210.

Depondt, C., H. R. Cock, et al. (2004). "The -1021C->T DBH gene variant is not associated with epilepsy or antiepileptic drug response." Neurology 63(8): 1497- 1499.

Dickinson, D. and B. Elvevag (2009). "Genes, cognition and brain through a COMT lens." Neuroscience 164(1): 72-87.

Doxakis, E. (2010). "Post-transcriptional regulation of alpha-synuclein expression by mir-7 and mir-153." J. Biol. Chem. 285(17): 12726-12734.

Duan, J., M. S. Wainwright, et al. (2003). "Synonymous mutations in the human dopamine receptor D2 (DRD2) affect mRNA stability and synthesis of the receptor." Hum. Mol. Genet. 12(3): 205-216.

Dubinsky, M. C., E. Reyes, et al. (2005). "A cost-effectiveness analysis of alternative disease management strategies in patients with Crohn's disease treated with azathioprine or 6-mercaptopurine." Am. J. Gastroenterol. 100(10): 2239-2247.

Engelender, S., Z. Kaminsky, et al. (1999). "Synphilin-1 associates with alpha-synuclein and promotes the formation of cytosolic inclusions." Nat. Genet. 22(1): 110-114.

Farrer, M., A. Destee, et al. (2001). "Genetic analysis of synphilin-1 in familial Parkinson's disease." Neurobiol. Dis. 8(2): 317-323.

Feher, E. (2004). "[Neuropeptide containing nerve fibres in the liver]." Orv. Hetil. 145(34): 1759-1763.

Fernandez, F., N. Colson, et al. (2009). "Association between migraine and a functional polymorphism at the dopamine beta-hydroxylase locus." Neurogenetics 10(3): 199-208.

Fijal, B. A., B. J. Kinon, et al. (2009). "Candidate-gene association analysis of response to risperidone in African-American and white patients with schizophrenia." Pharmacogenomics J. 9(5): 311-318.

Fox, E. R., J. H. Young, et al. (2011). "Association of genetic variation with systolic and diastolic blood pressure among African Americans: the Candidate Gene Association Resource study." Hum. Mol. Genet. 20(11): 2273-2284.

Freire, M. T., M. H. Hutz, et al. (2005). "The DBH -1021 C/T polymorphism is not associated with alcoholism but possibly with patients' exposure to life events." J. Neural Transm. 112(9): 1269-1274. 122

Freire, M. T., F. Z. Marques, et al. (2006). "Polymorphisms in the DBH and DRD2 gene regions and smoking behavior." Eur. Arch. Psychiatry Clin. Neurosci. 256(2): 93- 97.

Fuchs, J., A. Tichopad, et al. (2008). "Genetic variability in the SNCA gene influences alpha-synuclein levels in the blood and brain." FASEB J. 22(5): 1327-1334.

Fukuda, A., T. Minami, et al. (1987). "The effects of noradrenaline on neurones in the rat dorsal motor nucleus of the vagus, in vitro." J. Physiol. 393: 213-231.

Garland, E. M., B. K. Black, et al. (2007). "Dopamine-beta-hydroxylase in postural tachycardia syndrome." Am. J. Physiol. Heart Circ. Physiol. 293(1): H684-690.

Go, A. S., D. Mozaffarian, et al. (2013). "Executive summary: heart disease and stroke statistics--2013 update: a report from the American Heart Association." Circulation 127(1): 143-152.

Goris, A., C. H. Williams-Gray, et al. (2007). "Tau and alpha-synuclein in susceptibility to, and dementia in, Parkinson's disease." Ann. Neurol. 62(2): 145-153.

Gottesman, O., H. Kuivaniemi, et al. (2013). "The Electronic Medical Records and Genomics (eMERGE) Network: past, present, and future." Genet. Med. 15(10): 761-771.

Greene, C. M., I. H. Robertson, et al. (2010). "Dopaminergic genotype influences spatial bias in healthy adults." Neuropsychologia 48(9): 2458-2464.

Greenwood, P. M., R. Sundararajan, et al. (2009). "Both a nicotinic single nucleotide polymorphism (SNP) and a noradrenergic SNP modulate working memory performance when attention is manipulated." J. Cogn. Neurosci. 21(11): 2139- 2153.

Grobecker, H., M. F. Roizen, et al. (1977). "Effect of prolonged treatment with adrenergic neuron blocking drugs on sympathoadrenal reactivity in rats." Eur. J. Pharmacol. 46(2): 125-133.

Guindalini, C., R. Laranjeira, et al. (2008). "Dopamine-beta hydroxylase polymorphism and cocaine addiction." Behav. Brain Funct. 4: 1.

Healy, D. G., P. M. Abou-Sleiman, et al. (2004). "A functional polymorphism regulating dopamine beta-hydroxylase influences against Parkinson's disease." Ann. Neurol. 55(3): 443-446.

Hebbring, S. J., S. J. Schrodi, et al. (2013). "A PheWAS approach in studying HLA- DRB1*1501." Genes Immun. 14(3): 187-191.

123

Hess, C., A. Reif, et al. (2009). "A functional dopamine-beta-hydroxylase gene promoter polymorphism is associated with impulsive personality styles, but not with affective disorders." J. Neural Transm. 116(2): 121-130.

Hosak, L. (2007). "Role of the COMT gene Val158Met polymorphism in mental disorders: a review." Eur. Psychiatry 22(5): 276-281.

Howes, O. D. and S. Kapur (2009). "The dopamine hypothesis of schizophrenia: version III--the final common pathway." Schizophr. Bull. 35(3): 549-562.

Hu, F. Y., W. B. Hu, et al. (2010). "Lack of replication of a previously reported association between polymorphism in the 3'UTR of the alpha-synuclein gene and Parkinson's disease in Chinese subjects." Neurosci. Lett. 479(1): 31-33.

Hu, Y., B. Tang, et al. (2012). "Variant in the 3' region of SNCA associated with Parkinson's disease and serum alpha-synuclein levels." J. Neurol. 259(3): 497- 504.

Ittiwut, R., J. B. Listman, et al. (2011). "Association between polymorphisms in catechol- O-methyltransferase (COMT) and cocaine-induced paranoia in European- American and African-American populations." Am. J. Med. Genet. B Neuropsychiatr. Genet. 156B(6): 651-660.

Johnson, A. D., R. E. Handsaker, et al. (2008). "SNAP: a web-based tool for identification and annotation of proxy SNPs using HapMap." Bioinformatics 24(24): 2938-2939.

Johnson, A. D. and C. J. O'Donnell (2009). "An open access database of genome-wide association results." BMC medical genetics 10: 6.

Jugurnauth, S. K., C. K. Chen, et al. (2011). "A COMT gene haplotype associated with methamphetamine abuse." Pharmacogenet. Genomics 21(11): 731-740.

Junn, E., K. W. Lee, et al. (2009). "Repression of alpha-synuclein expression and toxicity by microRNA-7." Proc. Natl. Acad. Sci. U. S. A. 106(31): 13052-13057.

Kaiser, M. J., G. Tiegs, et al. (2003). "Close apposition of dynorphin-positive nerve fibres to lymphocytes in the liver suggests opioidergic neuroimmunomodulation." Histochem. Cell Biol. 120(3): 213-221.

Kalayasiri, R., A. Sughondhabirom, et al. (2007). "Dopamine beta-hydroxylase gene (DbetaH) -1021C-->T influences self-reported paranoia during cocaine self- administration." Biol. Psychiatry 61(11): 1310-1313.

124

Kamata, M., A. Suzuki, et al. (2009). "Association study between the -1021C/T polymorphism of the dopamine-beta-hydroxylase gene promoter and personality traits in healthy subjects." Neurosci. Lett. 462(1): 54-57.

Kamiyama, M., M. Kobayashi, et al. (2007). "Polymorphisms in the 3' UTR in the neurocalcin delta gene affect mRNA stability, and confer susceptibility to diabetic nephropathy." Hum. Genet. 122(3-4): 397-407.

Keating, B. J., S. Tischfield, et al. (2008). "Concept, design and implementation of a cardiovascular gene-centric 50 k SNP array for large-scale genomic association studies." PloS one 3(10): e3583.

Kendler, K. S., L. M. Karkowski, et al. (2000). "Illicit psychoactive substance use, heavy use, abuse, and dependence in a US population-based sample of male twins." Arch. Gen. Psychiatry 57(3): 261-269.

Kendler, K. S., M. C. Neale, et al. (1994). "A twin-family study of alcoholism in women." Am. J. Psychiatry 151(5): 707-715.

Kim, C. H., C. P. Zabetian, et al. (2002). "Mutations in the dopamine beta-hydroxylase gene are associated with human norepinephrine deficiency." Am. J. Med. Genet. 108(2): 140-147.

Kobayashi, K., Y. Kurosawa, et al. (1989). "Human dopamine beta-hydroxylase gene: two mRNA types having different 3'-terminal regions are produced through alternative polyadenylation." Nucleic Acids Res. 17(3): 1089-1102.

Kohnke, M. D., W. Kolb, et al. (2006). "DBH*444G/A polymorphism of the dopamine- beta-hydroxylase gene is associated with alcoholism but not with severe alcohol withdrawal symptoms." J. Neural Transm. 113(7): 869-876.

Kosten, T. R., G. Wu, et al. (2013). "Pharmacogenetic Randomized Trial for Cocaine Abuse: Disulfiram and Dopamine beta-Hydroxylase." Biol. Psychiatry 73(3): 219-224.

Kunugi, H., S. Nanko, et al. (1997). "High and low activity alleles of catechol-O- methyltransferase gene: ethnic difference and possible association with Parkinson's disease." Neurosci. Lett. 221(2-3): 202-204.

Kwan, T., D. Benovoy, et al. (2007). "Heritability of alternative splicing in the human genome." Genome Res. 17(8): 1210-1218.

Lachman, H. M., D. F. Papolos, et al. (1996). "Human catechol-O-methyltransferase pharmacogenetics: description of a functional polymorphism and its potential application to neuropsychiatric disorders." Pharmacogenetics 6(3): 243-250.

125

Laurian, L., Z. Oberman, et al. (1982). "Changes in serum dopamine-beta-hydroxylase activity during cold pressor test in subjects with high and low basal activity of this enzyme." J. Neural Transm. 53(2-3): 127-132.

Lavedan, C. (1998). "The synuclein family." Genome Res. 8(9): 871-880.

Lee, F. J., F. Liu, et al. (2001). "Direct binding and functional coupling of alpha- synuclein to the dopamine transporters accelerate dopamine-induced apoptosis." FASEB J. 15(6): 916-926.

Lewis, B. P., C. B. Burge, et al. (2005). "Conserved seed pairing, often flanked by adenosines, indicates that thousands of human genes are microRNA targets." Cell 120(1): 15-20.

Li, T., H. Vallada, et al. (1997). "Catechol-O-methyltransferase Val158Met polymorphism: frequency analysis in Han Chinese subjects and allelic association of the low activity allele with bipolar affective disorder." Pharmacogenetics 7(5): 349-353.

Liang, T. and L. G. Carr (2006). "Regulation of alpha-synuclein expression in alcohol- preferring and -non preferring rats." J. Neurochem. 99(2): 470-482.

Liang, T., J. Spence, et al. (2003). "alpha-Synuclein maps to a quantitative trait locus for alcohol preference and is differentially expressed in alcohol-preferring and - nonpreferring rats." Proc. Natl. Acad. Sci. U. S. A. 100(8): 4690-4695.

Lim, J. E., J. Pinsonneault, et al. (2007). "Tryptophan hydroxylase 2 (TPH2) haplotypes predict levels of TPH2 mRNA expression in human pons." Mol. Psychiatry 12(5): 491-501.

Linnertz, C., L. Saucier, et al. (2009). "Genetic regulation of alpha-synuclein mRNA expression in various human brain tissues." PloS one 4(10): e7480.

Little, R. A., K. N. Frayn, et al. (1986). "Plasma catecholamines in the acute phase of the response to myocardial infarction." Arch. Emerg. Med. 3(1): 20-27.

Lohoff, F. W., A. E. Weller, et al. (2008). "Association between the catechol-O- methyltransferase Val158Met polymorphism and cocaine dependence." Neuropsychopharmacology 33(13): 3078-3084.

Lotta, T., J. Vidgren, et al. (1995). "Kinetics of human soluble and membrane-bound catechol O-methyltransferase: a revised mechanism and description of the thermolabile variant of the enzyme." Biochemistry (Mosc). 34(13): 4202-4210.

Mamah, C. E., T. G. Lesnick, et al. (2005). "Interaction of alpha-synuclein and tau genotypes in Parkinson's disease." Ann. Neurol. 57(3): 439-443. 126

Man in 't Veld, A. J., F. Boomsma, et al. (1987). "Congenital dopamine-beta-hydroxylase deficiency. A novel orthostatic syndrome." Lancet 1(8526): 183-188.

Maraganore, D. M., M. de Andrade, et al. (2006). "Collaborative analysis of alpha- synuclein gene promoter variability and Parkinson disease." JAMA 296(6): 661- 670.

Maraganore, D. M., M. J. Farrer, et al. (2003). "Case-control study of the alpha-synuclein interacting protein gene and Parkinson's disease." Mov. Disord. 18(11): 1233- 1239.

Marino, M. D., B. N. Bourdelat-Parks, et al. (2005). "Genetic reduction of noradrenergic function alters social memory and reduces aggression in mice." Behav. Brain Res. 161(2): 197-203.

Maroteaux, L. and R. H. Scheller (1991). "The rat brain synucleins; family of proteins transiently associated with neuronal membrane." Brain Res. Mol. Brain Res. 11(3-4): 335-343.

Marx, F. P., C. Holzmann, et al. (2003). "Identification and functional characterization of a novel R621C mutation in the synphilin-1 gene in Parkinson's disease." Hum. Mol. Genet. 12(11): 1223-1231.

Mash, D. C., N. Adi, et al. (2008). "Alpha synuclein protein levels are increased in serum from recently abstinent cocaine abusers." Drug Alcohol Depend. 94(1-3): 246- 250.

Mash, D. C., Q. Ouyang, et al. (2003). "Cocaine abusers have an overexpression of alpha-synuclein in dopamine neurons." J. Neurosci. 23(7): 2564-2571.

Mata, I. F., M. Shi, et al. (2010). "SNCA variant associated with Parkinson disease and plasma alpha-synuclein level." Arch. Neurol. 67(11): 1350-1356.

Mathers, C. D. and D. Loncar (2006). "Projections of global mortality and burden of disease from 2002 to 2030." PLoS Med 3(11): e442.

Matsumoto, M., C. S. Weickert, et al. (2003). "Catechol O-methyltransferase (COMT) mRNA expression in the dorsolateral prefrontal cortex of patients with schizophrenia." Neuropsychopharmacology 28(8): 1521-1530.

McCarty, C. A., R. A. Wilke, et al. (2005). "Marshfield Clinic Personalized Medicine Research Project (PMRP): design, methods and recruitment for a large population-based biobank." Per. Med. 2(1): 49-79.

127

McLeod, H. L., L. Fang, et al. (1994). "Ethnic differences in erythrocyte catechol-O- methyltransferase activity in black and white Americans." J. Pharmacol. Exp. Ther. 270(1): 26-29.

McLeod, H. L., A. C. Syvanen, et al. (1998). "Ethnic differences in catechol O- methyltransferase pharmacogenetics: frequency of the codon 108/158 low activity allele is lower in Kenyan than Caucasian or South-west Asian individuals." Pharmacogenetics 8(3): 195-199.

Miles, J. S., A. W. McLaren, et al. (1989). "Alternative splicing in the human cytochrome P450IIB6 gene generates a high level of aberrant messages." Nucleic Acids Res. 17(20): 8241-8255.

Miller, S. A., D. D. Dykes, et al. (1988). "A simple salting out procedure for extracting DNA from human nucleated cells." Nucleic Acids Res. 16(3): 1215.

Motallebipour, M., A. Ameur, et al. (2009). "Differential binding and co-binding pattern of FOXA1 and FOXA3 and their relation to H3K4me3 in HepG2 cells revealed by ChIP-seq." Genome Biol. 10(11): R129.

Mueller, J. C., J. Fuchs, et al. (2005). "Multiple regions of alpha-synuclein are associated with Parkinson's disease." Ann. Neurol. 57(4): 535-541.

Mustapic, M., N. Pivac, et al. (2007). "Dopamine beta-hydroxylase (DBH) activity and - 1021C/T polymorphism of DBH gene in combat-related post-traumatic stress disorder." Am. J. Med. Genet. B Neuropsychiatr. Genet. 144B(8): 1087-1089.

Myhre, R., H. Klungland, et al. (2008). "Genetic association study of synphilin-1 in idiopathic Parkinson's disease." BMC medical genetics 9: 19.

Myhre, R., M. Toft, et al. (2008). "Multiple alpha-synuclein gene polymorphisms are associated with Parkinson's disease in a Norwegian population." Acta Neurol. Scand. 118(5): 320-327.

O'Gara, P. T., F. G. Kushner, et al. (2013). "2013 ACCF/AHA guideline for the management of ST-elevation myocardial infarction: executive summary: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines." Circulation 127(4): 529-555.

Ohlstein, E. H., L. I. Kruse, et al. (1987). "Cardiovascular effects of a new potent dopamine beta-hydroxylase inhibitor in spontaneously hypertensive rats." J. Pharmacol. Exp. Ther. 241(2): 554-559.

128

Okoniewski, M. J. and C. J. Miller (2006). "Hybridization interactions between probesets in short oligo microarrays lead to spurious correlations." BMC bioinformatics 7: 276.

Papp, A. C., J. K. Pinsonneault, et al. (2003). "Single nucleotide polymorphism genotyping using allele-specific PCR and fluorescence melting curves." Biotechniques 34(5): 1068-1072.

Papp, A. C., J. K. Pinsonneault, et al. (2012). "Cholesteryl Ester Transfer Protein (CETP) polymorphisms affect mRNA splicing, HDL levels, and sex-dependent cardiovascular risk." PloS one 7(3): e31930.

Parasuraman, R., P. M. Greenwood, et al. (2005). "Beyond heritability: neurotransmitter genes differentially modulate visuospatial attention and working memory." Psychol. Sci. 16(3): 200-207.

Pasha, D. N., J. T. Davis, et al. (2013). "Heritable Influence of DBH on Adrenergic and Renal Function: Twin and Disease Studies." PloS one 8(12): e82956.

Pendergrass, S. A., K. Brown-Gentry, et al. (2011). "The use of phenome-wide association studies (PheWAS) for exploration of novel genotype-phenotype relationships and pleiotropy discovery." Genet. Epidemiol. 35(5): 410-422.

Perez, R. G., J. C. Waymire, et al. (2002). "A role for alpha-synuclein in the regulation of dopamine biosynthesis." J. Neurosci. 22(8): 3090-3099.

Planz, G. and D. Palm (1973). "Acute enhancement of dopamine-beta-hydroxylase activity in human plasma after maximum work load." Eur. J. Clin. Pharmacol. 5: 255-258.

Pongor, E., N. Ledo, et al. (2010). "Distribution and possible origin of neuropeptide- containing nerve elements in the mammalian liver." Acta Vet. Hung. 58(2): 177- 187.

Post, R. M. (1975). "Cocaine psychoses: a continuum model." Am. J. Psychiatry 132(3): 225-231.

Punia, S., M. Das, et al. (2010). "Role of polymorphisms in dopamine synthesis and metabolism genes and association of DBH haplotypes with Parkinson's disease among North Indians." Pharmacogenet. Genomics 20(7): 435-441.

Qin, Y., Q. Ouyang, et al. (2005). "Cocaine abuse elevates alpha-synuclein and dopamine transporter levels in the human striatum." Neuroreport 16(13): 1489-1493.

Reese, M. G., F. H. Eeckman, et al. (1997). "Improved splice site detection in Genie." J. Comput. Biol. 4(3): 311-323. 129

Rivera-Calimlim, L. and D. K. Reilly (1984). "Difference in erythrocyte catechol-O- methyltransferase activity between Orientals and Caucasians: difference in levodopa tolerance." Clin. Pharmacol. Ther. 35(6): 804-809.

Roberts, J. W. (1969). "Promoter mutation in vitro." Nature 223(5205): 480-482.

Robertson, D. and E. M. Garland (1993). Dopamine Beta-Hydroxylase Deficiency. GeneReviews. R. A. Pagon, M. P. Adam, T. D. Birdet al. Seattle WA, University of Washington, Seattle.

Robertson, D., M. R. Goldberg, et al. (1986). "Isolated failure of autonomic noradrenergic neurotransmission. Evidence for impaired beta-hydroxylation of dopamine." N. Engl. J. Med. 314(23): 1494-1497.

Rockman, M. V. and G. A. Wray (2002). "Abundant raw material for cis-regulatory evolution in humans." Mol. Biol. Evol. 19(11): 1991-2004.

Ross, O. A., M. G. Heckman, et al. (2008). "Dopamine beta-hydroxylase -1021C>T association and Parkinson's disease." Parkinsonism Relat. Disord. 14(7): 544-547.

Ruffoli, R., F. S. Giorgi, et al. (2011). "The chemical neuroanatomy of vagus nerve stimulation." J. Chem. Neuroanat. 42(4): 288-296.

Samuel, E., R. B. Williams, et al. (2009). "Excited delirium: Consideration of selected medical and psychiatric issues." Neuropsychiatr. Dis. Treat. 5: 61-66.

Satoh, J. I. and Y. Kuroda (2002). "A putative polymorphic Val44Ala variation in the synphilin-1 gene is undetectable in Japanese sporadic Parkinson's disease patients." Eur. J. Neurol. 9(1): 15-18.

Schadt, E. E., C. Molony, et al. (2008). "Mapping the genetic architecture of gene expression in human liver." PLoS Biol 6(5): e107.

Schank, J. R., L. C. Liles, et al. (2008). "Norepinephrine signaling through beta- adrenergic receptors is critical for expression of cocaine-induced anxiety." Biol. Psychiatry 63(11): 1007-1012.

Schmitt, I., U. Wullner, et al. (2012). "Variants in the 3'UTR of SNCA do not affect miRNA-433 binding and alpha-synuclein expression." Eur. J. Hum. Genet. 20(12): 1265-1269.

Sempos, C. T., D. E. Bild, et al. (1999). "Overview of the Jackson Heart Study: a study of cardiovascular diseases in African American men and women." Am. J. Med. Sci. 317(3): 142-146.

130

Shapiro, M. B. and P. Senapathy (1987). "RNA splice junctions of different classes of eukaryotes: sequence statistics and functional implications in gene expression." Nucleic Acids Res. 15(17): 7155-7174.

Shifman, S., M. Bronstein, et al. (2002). "A highly significant association between a COMT haplotype and schizophrenia." Am. J. Hum. Genet. 71(6): 1296-1302.

Smith, P. J., C. Zhang, et al. (2006). "An increased specificity score matrix for the prediction of SF2/ASF-specific exonic splicing enhancers." Hum. Mol. Genet. 15(16): 2490-2508.

Sorrentino, S., C. Forleo, et al. (2010). "Lack of association between genetic polymorphisms affecting sympathetic activity and tilt-induced vasovagal syncope." Auton. Neurosci. 155(1-2): 98-103.

Sotiriou, S., G. Gibney, et al. (2009). "A single nucleotide polymorphism in the 3'UTR of the SNCA gene encoding alpha-synuclein is a new potential susceptibility locus for Parkinson disease." Neurosci. Lett. 461(2): 196-201.

Ste Marie, L. and R. D. Palmiter (2003). "Norepinephrine and epinephrine-deficient mice are hyperinsulinemic and have lower blood glucose." Endocrinology 144(10): 4427-4432.

Stone, R. A., N. Kirshner, et al. (1974). "Changes of plasma dopamine-beta-hydroxylase activity and other plasma constituents during the cold pressor test." Life Sci. 14(9): 1797-1805.

Sullivan, D., J. K. Pinsonneault, et al. (2013). "Dopamine transporter DAT and receptor DRD2 variants affect risk of lethal cocaine abuse: a gene-gene-environment interaction." Transl. Psychiatry 3: e222.

Swoap, S. J., D. Weinshenker, et al. (2004). "Dbh(-/-) mice are hypotensive, have altered circadian rhythms, and have abnormal responses to dieting and stress." Am. J. Physiol. Regul. Integr. Comp. Physiol. 286(1): R108-113.

Team RDC (2009) R: A language and environment for statistical computing. Vienna, Austria: The R Foundation for Statistical Computing. 2.12.0 edition.

Thomas, S. A., B. T. Marck, et al. (1998). "Restoration of norepinephrine and reversal of phenotypes in mice lacking dopamine beta-hydroxylase." J. Neurochem. 70(6): 2468-2476.

Thomas, S. A., A. M. Matsumoto, et al. (1995). "Noradrenaline is essential for mouse fetal development." Nature 374(6523): 643-646.

131

Thomas, S. A. and R. D. Palmiter (1997). "Thermoregulatory and metabolic phenotypes of mice lacking noradrenaline and adrenaline." Nature 387(6628): 94-97.

Thota, C., R. Menon, et al. (2012). "A single-nucleotide polymorphism in the fetal catechol-O-methyltransferase gene is associated with spontaneous preterm birth in African Americans." Reprod. Sci. 19(2): 135-142.

Tijo, J. H. and A. Levan (1956). "The chromosome number of man." Hereditas 42: 1-6.

Timmers, H. J., J. Deinum, et al. (2004). "Congenital dopamine-beta-hydroxylase deficiency in humans." Ann. N. Y. Acad. Sci. 1018: 520-523.

Tunbridge, E., P. W. Burnet, et al. (2004). "Catechol-o-methyltransferase (COMT) and proline dehydrogenase (PRODH) mRNAs in the dorsolateral prefrontal cortex in schizophrenia, bipolar disorder, and major depression." Synapse 51(2): 112-118.

Tunbridge, E. M. (2010). "The catechol-O-methyltransferase gene: its regulation and polymorphisms." Int. Rev. Neurobiol. 95: 7-27.

Walker, S. J. and K. A. Grant (2006). "Peripheral blood alpha-synuclein mRNA levels are elevated in cynomolgus monkeys that chronically self-administer ethanol." Alcohol 38(1): 1-4.

Wang, D., Y. Guo, et al. (2011). "Intronic polymorphism in CYP3A4 affects hepatic expression and response to statin drugs." Pharmacogenomics J. 11(4): 274-286.

Wang, D., A. D. Johnson, et al. (2005). "Multidrug resistance polypeptide 1 (MDR1, ABCB1) variant 3435C>T affects mRNA stability." Pharmacogenet. Genomics 15(10): 693-704.

Wang, D., M. J. Poi, et al. (2014). "Common CYP2D6 polymorphisms affecting alternative splicing and transcription: long-range haplotypes with two regulatory variants modulate CYP2D6 activity." Hum. Mol. Genet. 23(1): 268-278.

Watson, J. D. and F. H. Crick (1953). "Molecular structure of nucleic acids; a structure for deoxyribose nucleic acid." Nature 171(4356): 737-738.

Weinshilboum, R. M. (1979). "Serum dopamine beta-hydroxylase." Pharmacol. Rev. 30(2): 133-166.

Weinshilboum, R. M., N. B. Thoa, et al. (1971). "Proportional release of norepinephrine and dopamine-beta-hydroxylase from sympathetic nerves." Science 174(4016): 1349-1351.

Wetli, C. V. and D. A. Fishbain (1985). "Cocaine-induced psychosis and sudden death in recreational cocaine users." J. Forensic Sci. 30(3): 873-880. 132

White, H. D. and D. P. Chew (2008). "Acute myocardial infarction." Lancet 372(9638): 570-584.

Windemuth, A., V. D. Calhoun, et al. (2008). "Physiogenomic analysis of localized FMRI brain activity in schizophrenia." Ann. Biomed. Eng. 36(6): 877-888.

Wirdefeldt, K., C. E. Burgess, et al. (2003). "A linkage study of candidate loci in familial Parkinson's Disease." BMC neurology 3: 6.

Witte, A. V. and A. Floel (2012). "Effects of COMT polymorphisms on brain function and behavior in health and disease." Brain Res. Bull. 88(5): 418-428.

Wood, J. G., P. R. Joyce, et al. (2002). "A polymorphism in the dopamine beta- hydroxylase gene is associated with "paranoid ideation" in patients with major depression." Biol. Psychiatry 51(5): 365-369.

Xie, T., S. L. Ho, et al. (1997). "G/A1947 polymorphism in catechol-O-methyltransferase (COMT) gene in Parkinson's disease." Mov. Disord. 12(3): 426-427.

Yamamoto, K., J. F. Cubells, et al. (2003). "Dopamine beta-hydroxylase (DBH) gene and schizophrenia phenotypic variability: a genetic association study." Am. J. Med. Genet. B Neuropsychiatr. Genet. 117B(1): 33-38.

Yan, H., W. Yuan, et al. (2002). "Allelic variation in human gene expression." Science 297(5584): 1143.

Yavich, L., H. Tanila, et al. (2004). "Role of alpha-synuclein in presynaptic dopamine recruitment." J. Neurosci. 24(49): 11165-11170.

Yeh, T. K., T. C. Yeh, et al. (2010). "Association of polymorphisms in genes involved in the dopaminergic pathway with blood pressure and uric acid levels in Chinese females." J. Neural Transm. 117(12): 1371-1376.

Zabetian, C. P., G. M. Anderson, et al. (2001). "A quantitative-trait analysis of human plasma-dopamine beta-hydroxylase activity: evidence for a major functional polymorphism at the DBH locus." Am. J. Hum. Genet. 68(2): 515-522.

Zuvich, R. L., L. L. Armstrong, et al. (2011). "Pitfalls of merging GWAS data: lessons learned in the eMERGE network and quality control procedures to maintain high data quality." Genet. Epidemiol. 35(8): 887-898.

133

Appendix A: Funding Sources

This work was supported in part by NIH grant U01 GM092655 (W.S.) and a

Distinguished University Fellowship from The Ohio State University (E.S.B.) Tissues were provided under NIH grant 5 U42 RR006042-20, UL1TR000427 and

1U01HG006389-01 and Emory P30NS055077. The Jackson Heart Study is supported by contracts HHSN268201300046C, HHSN268201300047C, HHSN268201300048C,

HHSN268201300049C, HHSN268201300050C from the National Heart, Lung, and

Blood Institute and the National Institute on Minority Health and Health Disparities. The

PheWAS analyses were partially supported by NIH grants UL1TR000427, NIH

1U01HG006389-01, and NIH U01HL065962. The Lewy body dementia project was supported by a grant from the Harry T. Mangurian, Jr. Foundation.

134

Appendix B: Abbreviations

AEI allelic expression imbalance

ANOVA analysis of variance

ARVCF armadillo repeat deleted in velocardiofacial syndrome

BAC bacterial artificial chromosome

BMI body mass index bp base pairs cDNA complementary deoxyribonucleic acid

CEU HapMap population from Utah with ancestry from northern and western

Europe

CHD coronary heart disease

CI confidence interval

CIDR Center for Inherited Disease Research

CIP cocaine induced paranoia

CNS central nervous system

CNV copy number variation

COMT catechol-O-methyltransferase

Ct cycle threshold dbGAP database of Genotypes and Phenotypes

135

DBH dopamine β-hydroxylase

DBP diastolic blood pressure

DNA deoxyribonucleic acid

dNTP deoxynucleotide triphosphate

DOPS dihydroxyphenylserine

ECG electrocardiogram

EHR electronic health record

ENCODE Encyclopedia of DNA Elements

FISH fluorescent in situ hybridization

gDNA genomic deoxyribonucleic acid

GSP gene-specific primer

GWAS genome-wide association study

HGP human genome project

HRS Health and Retirement Study

ICD-9 International Classification of Diseases, Ninth Edition

IRB Institutional Review Board

JHS Jackson Heart Study

kb kilobases

LBD Lewy body dementia

LC locus coeruleus

LD linkage disequilibrium

LRT likelihood ratio test

136

MAF minor allele frequency

MI myocardial infarction mRNA messenger ribonucleic acid

NEB New England Biolabs

OR odds ratio

PAH phenylalanine hydroxylase

PCR polymerase chain reaction

PD Parkinson’s disease

PEAR Pharmacogenomic Evaluation of Antihypertensive Responses

PGK1 phosphoglycerate kinase 1

PheGenI Phenotype-Genotype Integrator

PheWAS phenome-wide association study

PMRP Marshfield Personalized Medicine Research Project qRT-PCR quantitative real-time reverse transcriptase PCR

RACE rapid amplification of cDNA ends

RIN RNA integrity number

RNA ribonucleic acid

SARDH sarcosine dehydrogenase

SBP systolic blood pressure

SE standard error

SNCA alpha-synuclein

SNCAIP alpha-synuclein interacting protein

137

SNP single nucleotide polymorphism

TXNRD2 thioredoxin reductase 2

UCSC University of California, Santa Cruz

UTR untranslated region

YRI HapMap population from Yoruba in Ibadan, Nigeria

138

Appendix C: Tables

Table 13. DBH primers for DNA and RNA amplification efficiency comparison for rs1108580 and rs77905

SNP F/R Primer Sequence F GCTGGCAATGAATGCGG rs1108580 DNA R CAAGGTCTAGTCCAATGGGAGAG F GTCCTGTTTGGGATGTCCG rs1108580 RNA R GGAAGACTTCCATGTGGTGGA F TTGTAGTGGACGACAGGGACTG rs77905 DNA R TCTTGTCCTACAGCTTCTCCCC F GATCCACATCTTCGCCTCTCA rs77905 RNA R AGATGACCTTGGGCAGGG

139

Table 14. DBH SNaPshot primers

SNP F/R Primer Sequence F TTCCAAATTGGATGATATTATAAACCC rs1079783 PEF GGGAAAAATCTAAACACAGAGCCAA R CCCTTGTAGGTCATGCAGCC F GACGCCTGGAGTGACCAGAA rs1108580 PEF CCCAAGGATTACCTCATTGA R_RNA CAGTGACCGGAACGGCTC R_DNA GCACAGGATGGCAGCATG F TCTGCGACGATCCCCATGG rs129882 PEF CCATGGAACAGCCCTGCA PER GCCCCTTCATCCTGGGC R TGAGGCGGAACAGGGCTG F CAAGCAGAATGTCCTGAAGGC rs1611115 PER CTCCCTCCTGTCCTCTCCC R TCCTCTTGCCCAGAGCAGAT F TGTCCTTGTCATCCAACTTCCTG rs1989787 PER CTGGCGTCCTCAGAGAAGC R TGCTCCGCCCTCCTGAA F CATGCCCGCCCTCAGTC rs5318 PEF GGCCAGCCTGCCC R CCAGGATGACCAGGAAGATGG F ACGGCTCGGGCCTGC rs5320 PEF CGGAGTTGCCCTCAGAC PER GACCTCCATGGTGCACG R CTCCTTAATGTAGCACCAGTACGTG F_RNA GCGTCCGTGTCTCAGCAGT rs6271 PER AGGGCCTTCAGTACGTCGC R_RNA TTGCCCCCACCAATGC F ATGTGCTCATCACCTCCTGCA rs77905 PER CCAGCTCCCGGTCTTC R_DNA GGATTGCCCGATGCCAC R_RNA GGGCGCGAAGCTGTACAG

140

Table 15. DBH allele-specific melting curve (GC clamp) genotyping primers

SNP F/R Primer Sequence F CCAGGTGGCAGGTGCTGA rs77905 R_C GCCAGCTCCCGGTCTACC R_T GGCCAGCTCCCGGTCTACT F_A ACCCCAAGGATTACCTCATTCAA F_Agc GCGCGCGCCGCGCCCCCCAAGGATTACCTCATTTAA rs1108580 F_G CCCCAAGGATTACCTCATCGAG F_Ggc CCGCGCGGCCACCCCAAGGATTACCTCATCGAG R TGATACACGCAGCCCACG R CCAGACTTATCAGGGACAGGAAAG F AAGCAGAATGTCCTGAAGGCAG rs1611115 R_Cgc CCCCGGCGCGCGGCCCCCTCTCCCTCCTGTCCTCTCACG R_T CTCTCCCTCCTGTCCTCTGCCA F_Cgc CCCCGCCCCGCCCGCCCCACAGAAACGCAAGAGCCC rs1611122 F_G CCACAGAAACGCAAGTTCCG R CCACCTGTACACAAAACCCTGAC F_Agc CGCCCCGCGCGCGCCGTGTCTGCACCTGCTCAATT rs2519143 F_G GTGTCTGCACCTGCTCTGTC R TGGGATCCAGCCTGACCTAC

141

Table 16. Primers to amplify DBH gene region for Ion Torrent sequencing

Gene Region F/R Primer Sequence DBH upstream 1 F GGTGCTCGTTTAGCCTGTTGT R CCTTAGTCCCAGCCCTGCT DBH upstream 1 nest F AGGCTGCTGGGAGGGCT R AGAAGGTAAACTGAGGCCCAGG DBH 2 F GGCAGAGGTCACTGATCCAGAT R TCGGGTGTGGAAAGTGCCT DBH 3 F GGCAAGGGTGAGCAGAGAAC R GTGGCGGCAGTAGTTGAGG DBH 4 F CACCAAGGGCAATGAGGC R CTGGGTGCCAAAAATGCTCT DBH 5 F TTAAAGGAAACAAGTAGGGAGATGGT R CTGTGCCTGCCACCCTTC DBH 6 F CACGGACAAGTGCACCCAG R AGAGCTCAAGGTCCCGACAG DBH 7 F GCTGCTTGGGAGCGAGAG R CTGAGAGGCGAAGATGTGGAT DBH 8 F GCTCCCTACCGGGTCCTG R GCTTGTTGTTAGCAGCAGGTG DBH 9 F CACTCAGCCTGTGACCTTCG R CAAACACTGGCTGGGTGCT DBH 10 F CAGCACCAGGACACGGGCAG R GCAACCGTCCTGCCACCTCC DBH 10 nest F GATGCTGTTCACAATTACCTGGGT R CCCCAGATGCTGCCTTCAC DBH 11 F GGAGATGTGTGTCAACTACGTGC R CCTCACCTCAGTAAATGCCCC DBH 12 F GGTGAATGATTAAATTGGGTGG R CACCAGAAAGCACAGCCCT DBH 12 nest F TGGGGACAGAGGCGGG R ACTTCTCTCAACACTGATGGGCA SARDH 1 F GTGCCTGCCCGACTGG R TGTCGCAGAAAAAGCTTGGTT SARDH 2 F GCTTTTTCTGCGACAGGAGC R GAAGATGCTGGCAGTGGAGT SARDH 3 F GCAGTGTGGCAGGGCTGT R GCTGCAAACACTGCCACTTC SARDH 4 F GAAGTGGCAGTGTTTGCAGC R CAGAGACACAAAGGGGAGGC SARDH 5 F CTCCCCTTTGTGTCTCTGTCTTC R CGAATCCCTCCTTTCAATTCC

142

Table 17. DBH primers for fluorescent restriction fragment length polymorphism genotyping

Restriction SNP F/R Primer Sequence Enzyme F GGGTGGGCCAGTGCC rs1076150 DdeI R CAGGAGGAGAGCTGACTGTTGC F CCCAGTGCTTGTCTCTGCAG rs1108580 BslI R CCATCCTCCTTGGCTTTCTCT F CTGTATTTGGAACTTGGCATC rs2519152 TaqI R AGGCATTTTACTACCCAGAGG F GGTGCCAGGCCAAACTATTG rs3025343 MboI R GACACAGGTTCAGTGAAAACGG F GGTGGCCTGGCGGAG rs6271 BstUI R ACCAGAGAGGGTGCCTGC

143

Table 18. Additional primers for DBH experiments

Region/Purpose F/R Primer Sequence F GCAAAAGTCAGGCACATGCACC rs141116007 and R GTCAGCGAGATGGGGAGGTGGA CA repeat R-HEX [HEX]GTCAGCGAGATGGGGAGGTGGA exon 1-12 PCR F CATGCCCGCCCTCAGTC R TGAGGCGGAACAGGGCTG F CTGCTGAGTGGCTCGTGTTT 3'UTR qPCR R GAATGGAGCAATTTCCCAGG 5'RACE-adapter R TGCAGTGCGGCCACC 5'RACE-isoform A F TTGATGAAAACCCCAGCCA 5'RACE-isoform B F TTGATGAAACCCGCCCTC 5'RACE-isoform C F TTTGATGAAAGCCCTCAGTCG F AGTCTCCACACCCGCGAC SARDH qPCR R CTTCTGCCCGAGGAGCAC F GCTGCTGCTCTGTCTGATGGT Slc6a2 qPCR R CCAGACAACCTTTCCTGATGTCTT DBH GSP1 GACCTCCATGGTGCACG DBH GSP2 CCAGCTCCCGGTCTTC DBH GSP3 ATGCCTGAGGAGTCGTTTC DBH GSP4 AGGAGTAGGTCCCCCCTCAG DBH GSP5 CGACGGCCCGGAGTG DBH GSP6 GGTGCCAGCCCAAGTGTG SARDH GSP1 TGCAACTTGTCTTTGAGTGCAG SARDH GSP2 GTACAAGAGGCTCATGTCGCTG SARDH GSP3 GGAGCGACCGGGATGG SARDH GSP4 CACCACCGTGCTGCAGG SARDH GSP5 ACCTATGGTGCCCAGGCTC

144

Table 19. SNCA primers

Purpose F/R Primer Sequence GSP1 AGACTACGAACCTGAAGCCTAAGAA GSP2 ATGTTGGAGGAGCAGTGGTGA GSP3 GGAGTGGCCATTCGACGAC 3′UTR F CTCCTTCCTTCCTCACCAGC sequencing R GCATTTCGGTGCTTCCCTT 3′UTR F AAAAGATGACCCTAGCACCCCT sequencing R GAGAACCAGACAGTCCAACTTATTTGT 3′UTR F CCATGAATTTAAGGATTTATGTGGATAC sequencing R CTCACCATTTATATACAAACACAAGTGAAT 3′UTR F CAAAATCATCTTCTACACTGCTTAGTTCC sequencing R GTATCTGTACCTGCCCCCACTC F AATACTTAAAAATATGTGAGCATGAAACTATG qPCR R TTTATTTTAATTCTCACCATTTATATACAAACAC F_FAM [6FAM]CCTGGCATATTTGATTGCAA Rep 1 F CCTGGCATATTTGATTGCAA genotyping R GACTGGCCCAAGATTAACCA F CAGCATTCACACCAATATCAGACA rs356165 R GAATTCCCTGAAGCAACACTGC RFLP R_FAM [6FAM]GAATTCCCTGAAGCAACACTGC F_HEX [HEX]TCAGTGAAAGGGAAGCACCG rs6842093 F TCAGTGAAAGGGAAGCACCG RFLP R GTAAGTGGGGAGCCATTTCCT F CAGCATTCACACCAATATCAGACA rs356165 R GAATTCCCTGAAGCAACACTGC SNaPshot PEF AACAACAGTTCCCCAAAATAC F CTCACCATTTATATACAAACACAAGTGAAT rs17016074 R TCTTTTAATGATACTGTCTAAGAATAATGACG SNaPshot R_alt GTATCTGTACCTGCCCCCACTC PEF AACACAAGTGAATAAAACACATC SNaPshot F CAGCATTCACACCAATATCAGACA both SNPs R TCTTTTAATGATACTGTCTAAGAATAATGACG rs17016074 F_G CAAACACAAGTGAATAAAACACTTCG allele- F_Agc GCCCGCCGCCCCCGCAAACACAAGTGAATAAAACAC specific R AATACTTAAAAATATGTGAGCATGAAACTATG

145

Table 20. COMT region amplification primers for Ion Torrent sequencing

Primers F/R Sequence F AGAGTAGAAGAAAAGCACTGGAATGC 1 R CCACCACACCTGGCCAAT F TCCATCTGGCCTGAAGGTCT 2 R GCTTAGGTGCCTCTGGTCTC F CACCCAGGGCCACAGGT 3 R GGTTGAATGCCAGACACTTGAA F CCCAATGCCCTTCCTGTG 4 R GCCACCGCCACCAATGT F TAAATTTAACCCAAAAAGTGAGAGACTTG 5 R TGCCATGCAGCTGGGACT F GCAGCTTCCTCACCTCTCTC 6 R AAGCTAGACCTCTTGCACCA F CGGCTCAAGCAAGGAGAGTG 7 R CTGTAGCAAGGCACCCCAA F ATCCAGGAATGGCATTCAACC 8 R TGGGTCCAAACAGGGAGGT F GATGGACCAAGAGGTAAGCGG 9 R TGCAGCACGGTCCTTTTGT F TCCACAACCTGCTCATGGGT 10 R GCTGAGGCTGACTGAATGGG F AGAGCTAGGGAATGAGTTATTAGACTTGA 11 R TGATGCCATCTTGAAACACTGTATTA F AGAGGTGGCCGAACTGACC 12 R AGGGATGCGCCTTGCTG F TGCGGAGACCAGTAGTGGC 13 R TCTCCCTGTGATGCTGTGCTTA F GTCTACAGCTCAGCGCCCAT 14 R GCCTCCACAGCGCCTG F AGGGCTCTTTCTCTGCAGCTT 15 R GCCGTCAGACCACATGAATG F TACCATGCTAGAAGGAGCGCA 16 R GACTGGAGTTGATTCATTTTCAGTTC F CGAGGGAGTTTGGCTGGA 17 R CGTGCCCCCAAGTAACCC F GTATCCACTTCAGGTCCACTAAAGG 18 R CATGCTTCTGTGCCAGGTCA F AAAATGAGGCTCCCAGGTCA 19 R TTTCCTTCTGTCCCTTGCTTCT F GATCGCTGAACAGTACTCCTCAAAT 20 R GACAGAGTTTCTCTTTAGGGTGATGAA 21 F AGCTGGGTGGTTTGATGCC Continued 146

Table 20 continued R GGAACACAGTACAGCCCATAACAA F ACGCCTTGTGACTCAGCTCC 22 R CACCGCACACCTCCTGCT F GTCCAGAGATGAGCGCGG 23 R GAAACTGACTTTGTTCTCTTTGCAAA F CTTTTAAAGAAAGATGAGACCAGATGC 24 R TCAGCCCCCCCAAGTAGC F GTGCTACTGGCTGACAACGTG 25 R CCTGCAGCTCTCCCAGAAACT F GGCCTTCTCCAGGCCGT 26 R CAGGCGCTGACGTTACCACTA F CCAACAAACTGGTAAATGACAGACA 27 R CTCCAACCTCAGCCCCCTA F GATCTGGATGGGAACCACCA 28 R CAGGCAGAGTGGAGATAACACG F CGCTGGACTGAGCCCG 29 R GGCTTTGTTTACAGTACTCAGCAGC F GCGCTCATCTCTGGACCC 30 R TCCAGGCTGAGCGACAGAGT F AAACCCTAACCTGGCGATGAC 31 R CCAGGTTAGGGTTTATGTTGATGAT F CCCAAAGAGCAGGAGCCC 32 R CACCTCTCCTCCGTCCCC F GCTACCCTGTCTTGGCCTGTAG 33 R AGGCCCGACACAAGGGA F GAAAGTAAGTGGTCATGTGGCTGT 34 R CCACAGCCACATGACCACTTA F ACGGGAGCAGGAAGGTCC 35 R GGGTGTTATATTTCTCCACTGAAGC F AACCTAGGAGAAACCAGGCCAT 36 R ACGACATGGTGCAATCATGG F CTGGTCATTCCTTTAGATTCCCA 37 R GGTGCAGGCCGGAGG F GGAAAGCTGAGGTTGAAGATCTGT 38 R TTGAGACGCAGTCTTGCTGTG F GCTCCAGATAGCGGTGAGAGAT 39 R GACGCAGTCTTGCTGTGTCG F CCACCCTCCCTCTCATCCTC 40 R GTCCCTGAAAGCACCAGGG F TTTCTGGGTAGAGTGAGATTCATGG 41 R GATGTTGTGTGCTGTGCGC F ACTGGAGTCGGCCTTAGACT 42 R TTGGTGGTGATGGACAGCTC Continued 147

Table 20 continued F ACTGGAGTCGGCCTTAGACT 43 R TGTCCAGCTGTCATGCCTTT F TCTGGGAAGTCGGTGGGAAT 44 R AGGTGCCCTTGATGTAGCTG F GCGCATAGGAAGCCCCA 45 R GTCTATCACGCCCATTATAAACCA F CCCACCTTTGGCAACGC 46 R GTGGAAGAGGAAGAGGAGGCTTAT F TGTGCACATGAAGCAGGAGG 47 R CGCTTGGCTTGAGGGCTAGT F CAACCAAAAATTACCAGTCACACAA 48 R GCATGTACCTTAACTTGCCAGTCTAC F TAAAGCAACTCAAAATAAATCAAAGACC 49 R TCTCTAAATACAGTCACATTCTGAAGTGC F ACATGCCTGGTTGCCTAGGAT 50 R GCAGAGTAGACACGATGACACCTT F GGCTGGTTGTGGTGTCGC 51 R GTTGGGACCAGTTCTCGGAA F GGTGCAAAATTCCGAGAACTG 52 R ATCCATCCAGGTGAGAGGTCG F TCCTCCACTTATTCGACCTCTCA 53 R TCAGCTTTCCACAAGGAGTCTG F ACTGCCCAATGTCACTCAGG 54 R CAGGTGTGGGGCTTCCTATG F GTCAGCACCATGTCCGGG 55 R CGCGCTTAGCACTCTTTGATG F GAGCCCAGAAATCATACAGGTCA 56 R CAAGTACTTTAACATCAAAGCCAGCTT F CTGGTGGGTAAGCACAGGG 57 R TCAGCACATAAAAATGACACAGAGT F TGGAAAGGCCTGCAAAATGC 58 R CTGCAGAAGTGTTAGCCCCA F CTCACCCCAGCTCACACATT 59 R GGGAGGCGTACAGTGAAAAC F GCTCGTGTGGAGGGTCATTT 60 R TGTCTCCCTAGCAAGGTGGC F CACAGGGTTGCAGAGGAGC 61 R TACGCTTTACAAGGGGAGCTG F GCAGGCAGAGTACACATGGA 62 R CCTCCACCAGCACATCTTGT F CCGTGTGCTCGTCAACAAAG 63 R CGTGGCATTCCAGTGCTTTT

148

Table 21. COMT primers

Region/Purpose F/R Primer Sequence GSP1 GCGGACGCAGGCCG GSP2 TGCCTCAGCCTCCCAAGTA GSP3 CTCCTGGCCTCAAGTGATTCTC GSP4 GGAGGTCATGCGGACACC GSP5 GGTTGTGGATGGGCTGCA GSP6 TGTGGGTGTCTTTCCTGCC GSP7 GAAAGAAGGCAGCACCTGTCC GSP8 GTTACTGGACACGCAGCCATAG GSP9 AGCATGTCAGGGCGCAG GSP10 ACTCAGGTCTGGCACAAGTGG qPCR F CCGCCTGCTGTCACCAG R ACGCTCCAACCACAAGGG rs165599 F GAAGGAGATGCTTCCACTCTGT R ACATTCAAAGCTCCCCTTGAC SNaPshot PEF ATGGGGACGACTGCC rs4633 F CCCCGCCTCTGCTGTT R TTTTTCTTTCTTGTCGCCCA SNaPshot PER CGCATGCTGCAGCAC F GAAGGAGATGCTTCCACTCTGT rs165599 RFLP R ACATTCAAAGCTCCCCTTGAC R HEX [HEX]ACATTCAAAGCTCCCCTTGAC F GCCCGCCTGCTGTCACC rs4633 RFLP F FAM [6FAM]GCCCGCCTGCTGTCACC R CTGAGGGGCCTGGTGATAGTG rs2020917 F GAACTGTCCTGAGTGACCAGACAC F FAM [6FAM]GAACTGTCCTGAGTGACCAGACAC RFLP R TGTCTAAGGTCCTGCTGTGCTG rs116474580 F TGGCCTCTGTGAGGTGTTAGC R GCCTCCATGCTCTCAGGGT RFLP R HEX [HEX]GCCTCCATGCTCTCAGGGT Deletion F CCCAGGCACTGCATTGTG R CCGGCCCACTCCACTCT rs13306278 F Cgc CCGCCGCCCGCCCGCCCGCACCCAGCCCCAGTATCC allele-specific FT CACCCAGCCCCAGCTTCT genotyping R TTCTAGCCACAAGTAGCCCCC rs4633 allele- F Cgc GGCGGCCCCGGCGGCGCAGCGCATCCTGAATCAC specific FT CAGCGCATCCTGGACCAT genotyping R AGTAGGTGTCAATGGCCTCCA rs737865 allele- F CGTGGGAATGTTAGAGAAAGGG specific RT TCGCCAACAGGACACAAAGAT genotyping R Cgc CGCGGGCCCGGCGCCAACAGGACACAATAAC

149