PHARMACOGENOMICS, TRANSCRIPTOMICS AND METABOLOMICS FOR THE IDENTIFICATION OF NOVEL BIOMARKERS OF BLOOD PRESSURE RESPONSE TO

ANTIHYPERTENSIVE DRUGS

By

MOHAMED HOSSAM SHAHIN

A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT

OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

2015

© 2015 Mohamed Hossam Shahin

To my precious family, my mother Amal Al-Ashry, my father Hossam Shahin,

my two sisters Noha and Maha, my grandparents, my aunt Azza Younis, my wife Yasmeen and my daughter Jude

4

ACKNOWLEDGMENTS

First, I would like to express my deepest gratitude and appreciation to my

mentor, Dr. Julie Johnson, for her mentorship, training, guidance, help and support over

the past five years. Her truly scientist intuition has made her as a constant oasis of

ideas and passions in science, which exceptionally inspired and enriched my growth as

a student and a scientist. I am indebted to her more than she knows, and will always be

for the rest of my career. I would also like to thank Dr. Taimour Langaee, Dr. Yan Gong,

Dr. Tim Garrett, and Dr. Alberto Riva for serving on my committee and for their valuable

advice, guidance, encouragement and sincere help throughout this work. My utmost

gratitude also goes to Dr. Sherief Khalifa, who provided me with great guidance and

support before joining Dr. Johnson’s lab, and was always encouraging me to seek

graduate studies in the United States. He is indeed one of the great professors who

significantly influenced my character and shaped my career path.

My sincerest gratitude goes to Dr. Rhonda Copper-DeHoff, Dr. Caitrin

McDonough, Dr. Reggie Frye and Dr. Larisa Cavallari for their scientific guidance,

valuable advice, and continuous support during my PhD. I would also like to gratefully

and sincerely thank Dr. Hartmut Derendorf and Dr. William Millard for their great help,

encouragement and support over the past four years. Additionally, I would like to extend

a special thanks to all present and former graduate students and postdocs in the

Department of Pharmacotherapy and Translational Research who made the years of

graduate school enjoyable. Special thanks to Dr. Mohamed Mohamed, Dr. Issam

Hamadeh, Dr. Nihal El-Rouby, Shin-wen Chang, Carol Sa, and Mohamed Solayman for

their great friendship, compassion and kindness which created a family environment

that I will never forget. I would also like to extend many thanks to Ben Burkley, Cheryl

5

Galloway, and Lynda Stauffer, who facilitated part of the research included in this

dissertation.

Last but not least, I would like to deeply thank my amazing wife – Yasmeen – for

her indispensable emotional support, kindness, patience and encouragement. She is

not only the love of my life, but also my best friend and favorite classmate who I always

seek her advice and feedback. I would also like to thank God for blessing me with Jude,

my sweet little daughter, whose pure smiles and giggles soothed the toughness of

graduate school. Additionally, I would like to take the opportunity to extend my deepest

gratitude to my precious family, my parents and my two sisters, for their unconditional

love and support. They have always believed in me, more than I do, and have been fully

supportive of all my decisions. They have been continuously praying for my success

and they were always there for me through the good and bad times. I would like to

dedicate this thesis to them for their endless love, support and self-sacrifices.

6

TABLE OF CONTENTS page

ACKNOWLEDGMENTS .................................................................................................. 4

LIST OF TABLES ............................................................................................................ 9

LIST OF FIGURES ........................................................................................................ 10

ABSTRACT ................................................................................................................... 12

CHAPTER

1 MECHANISMS AND PHARMACOGENETIC SIGNALS UNDERLYING THIAZIDE DIURETICS BLOOD PRESSURE RESPONSE .................................... 14

Hypertension ........................................................................................................... 14 Thiazide Diuretics ................................................................................................... 14

Blood Pressure Lowering Mechanisms of Thiazide Diuretics ................................. 16 Short Term Blood Pressure Lowering Mechanism ........................................... 16

Long Term BP Lowering Mechanism ............................................................... 17 Pharmacogenetics of Thiazide Diuretics BP Response .......................................... 20

Neural Precursor Cell Expressed, Developmentally Down Regulated 4 Like (NEDD4L)...................................................................................................... 20

Protein Kinase C Alpha (PRKCA) ..................................................................... 21 G Protein Alpha Subunit Endothelian-3 (GNAS-EDN3) .................................... 23

YEATS Domain Containing 4 (YEATS4) .......................................................... 23 Summary and Aims of the Project .......................................................................... 24

Significance ............................................................................................................ 26

2 GENOME WIDE PRIORITIZATION AND GENOMICS TRANSCRIPTOMICS INTEGRATION REVEAL NOVEL SIGNATURES ASSOCIATED WITH THIAZIDE DIURETICS BLOOD PRESSURE RESPONSE .................................... 33

Introduction ............................................................................................................. 33 Methods .................................................................................................................. 35

Pharmacogenomic Evaluation of Antihypertensive Response (PEAR) study ... 35 Pharmacogenomic Evaluation of Antihypertensive Response 2 (PEAR-2)

Study ............................................................................................................. 36 Thiazide Blood Pressure Response Measurement .......................................... 36

Genotyping ....................................................................................................... 38 Transcriptomics Profiling .................................................................................. 38

Statistical Analyses .......................................................................................... 39 Genomics analysis ..................................................................................... 39

Genome wide prioritization approach ......................................................... 40 Replication of Genome Wide Prioritized Single Nucleotide Polymorphisms ..... 41

Transcriptomics Analysis and Genomics Integration ........................................ 41

7

Results .................................................................................................................... 42 Characteristics of Study Participants and Thiazide Diuretics Blood Pressure

Response ...................................................................................................... 42 Genome Wide Prioritization Approach.............................................................. 43

Replication of Genome Wide Prioritized Single Nucleotide Polymorphisms ..... 43 Transcriptomics Analysis and Genomics Integration ........................................ 44

Discussion .............................................................................................................. 45

3 INTEGRATING METABOLOMICS AND GENOMICS UNCOVERS NOVEL PATHWAYS AND GENETIC SIGNATURES INFLUENCING HYDROCHLOROTHIAZIDE BLOOD PRESSURE RESPONSE: A GENETIC RESPONSE SCORE FOR HYDROCHLOROTHIAZIDE USE ................................ 60

Introduction ............................................................................................................. 60

Methods .................................................................................................................. 62 Study Participants ............................................................................................ 62

Hydrochlorothiazide Blood Pressure Response Measurement ........................ 63 Untargeted Metabolomics Profiling ................................................................... 64

Genotyping ....................................................................................................... 65 Statistical Analyses .......................................................................................... 66

Metabolomics analysis (step1, figure 3-3) .................................................. 66 Genomics analysis (step2, figure 3-3) ........................................................ 66

Genomics metabolomics integration (step3, figure 3-3) ............................. 67 Replication (step4, figure 3-3) .................................................................... 68

Create a response score (step5, figure 3-3)............................................... 69 Response score replication (step6, figure 3-3) ........................................... 69

Functional validation (step7, figure 3-3) ..................................................... 70 Results .................................................................................................................... 71

Characteristics of Study Participants and Hydrochlorothiazde Blood Pressure Response ....................................................................................... 71

Metabolomics Analysis (Step1, Figure 3-3) ...................................................... 71 Genomics Metabolomics Integration (Step3, Figure 3-3) ................................. 72

Replication (Step4, Figure 3-3) ......................................................................... 73 Create a Response Score (Step5, Figure 3-3) ................................................. 73

Response Score Replication (Step6, Figure 3-3) ............................................. 74 Functional Validation (Step7, Figure 3-3) ......................................................... 74

Discussion .............................................................................................................. 75

4 SPHINGOMYELIN METABOLIC PATHWAY IMPACTS THIAZIDE DIURETIC BLOOD PRESSURE RESPONSE: INSIGHTS FROM GENOMICS, METABOLOMICS AND LIPIDOMICS ANALYSES ................................................. 96

Introduction ............................................................................................................. 96 Methods .................................................................................................................. 97

Pharmacogenomic Evaluation of Antihypertensive Response Study ............... 97 Genetic Epidemiology of Responses to Antihypertensives Study .................... 98

Hydrochlorothiazide Blood Pressure Response Measurement ........................ 98

8

Metabolomics ................................................................................................... 99 Genomics ....................................................................................................... 100

Lipidomics ...................................................................................................... 100 Experimental Approach .................................................................................. 102

Metabolomics pathway analysis (step 1).................................................. 102 Genomics association analysis (step 2) ................................................... 102

Replication (step 3) .................................................................................. 103 Validation (step 4) .................................................................................... 103

Statistical Analyses ........................................................................................ 104 Results .................................................................................................................. 105

Metabolomics Pathway Analysis .................................................................... 105 Replication ...................................................................................................... 106

Validation........................................................................................................ 106 Discussion ............................................................................................................ 108

5 SUMMARY AND CONCLUSIONS ........................................................................ 123

LIST OF REFERENCES ............................................................................................. 131

BIOGRAPHICAL SKETCH .......................................................................................... 154

9

LIST OF TABLES Table page 2-1. Characteristics of PEAR and PEAR-2 participants ............................................. 50

2-2. Characteristics of PEAR European American participants included in the RNA-Seq analysis .............................................................................................. 51

2-3. Characteristics of PEAR-2 European American participants included in the RNA-Seq analysis .............................................................................................. 51

2-4. Genetic signals prioritized according to their potential function using RegulomeDB ...................................................................................................... 52

3-1. Characteristics of participants included in the genomics and metabolomics analyses ............................................................................................................. 81

3-2. Thirteen metabolites significantly associated with hydrochlorothiazide blood pressure response of Whites in the PEAR HCTZ monotherapy study ................ 82

3-3. Genes involved in the synthesis and degradation of arachidonic acid ................ 83

3-4. The effect of the 60 polymorphisms selected from the eleven genes involved in the synthesis and degradation of arachidonic acid on hydrochlorothiazide blood pressure responses .................................................................................. 84

4-1. Characteristics of White PEAR participants involved in the genomics and metabolomics analyses .................................................................................... 112

4-2. Characteristics of White PEAR participants included in the lipidomics analyses ........................................................................................................... 113

4-3. Significant pathways (FDR <0.05) from the metabolomics pathway analysis ... 114

4-4. Canonical genes in the sphingomyelin metabolism pathway which we tested the association between the SNPs located in these genes and hydrochlorothiazide blood pressure response .................................................. 115

4-5. Top signals from testing the correlation between 50 sphingolipids with SPTLC3 rs6078905 SNP .................................................................................. 116

10

LIST OF FIGURES

Figure page 1-1. Known and theoretical blood pressure lowering mechanisms of thiazide

diuretics .............................................................................................................. 27

1-2. Blood pressure response to hydrochlorothiazide by NEDD4L rs4149601 genotype for White participants in PEAR study .................................................. 28

1-3. Blood pressure response to hydrochlorothiazide by PRKCA rs16960228 genotype for White participants from five independent studies........................... 29

1-4. Blood pressure response to hydrochlorothiazide by GNAS-EDN3 rs2273359 genotype for White participants from five independent studies........................... 30

1-5. Blood pressure response to hydrochlorothiazide by rs7297610 genotype in PEAR African Americans. ................................................................................... 31

1-6. The basic flow of genetic information in a cell .................................................... 32

2-1. Represents the overall framework of the experimental approaches used in this study ............................................................................................................ 53

2-2. RegulomeDB scoring scheme. ........................................................................... 54

2-3. Linkage disequilibrium plots between the six prioritized genetic signals from the genome-wide prioritization approach.. .......................................................... 55

2-4. The effect of rs10995 polymorphism on the blood pressure response of Whites treated with thiazide in the PEAR and PEAR-2 studies.. ........................ 56

2-5. Plots showing the difference in the VASP baseline expression levels between thiazide diuretics extreme responders in the PEAR and PEAR-2 studies.. ........ 57

2-6. The expression levels of VASP by rs10995 genotypes in whole blood collected from PEAR White participants at baseline.. ......................................... 58

2-7. Plot showing RhoB and CDC42EP2 baseline expression levels between thiazide responders compared to non-responders in the PEAR and PEAR-2 RNA-Seq analyses.. ........................................................................................... 59

3-1. Represents the study design of the pharmacogenomic evaluation of antihypertensive responses (PEAR) study ......................................................... 86

3-2. Distribution of the systolic blood pressure (SBP) and diastolic blood pressure (DBP) responses to hydrochlorothiazide in PEAR participants included in the metabolomics analysis ....................................................................................... 87

11

3-3. The overall analyses framework of the study...................................................... 88

3-4. Quantile-quantile plots from genome-wide association analysis of blood pressure response to hydrochlorothiazide in Whites in the PEAR study ............ 89

3-5. Netrin signaling pathway generated by integrating genomics and metabolomics data using Ingenuity pathway analysis. ....................................... 90

3-6. The effects of rs2727563 and rs12604940 polymorphisms on the blood pressure response of Whites treated with hydrochlorothiazide in the PEAR HCTZ monotherapy and HCTZ add-on .............................................................. 91

3-7. Correlation between hydrochlorothiazide BP response and arachidonic acid peak height ratio ................................................................................................. 92

3-8. The effects of rs13262930 polymorphism on the blood pressure response of Whites treated with hydrochlorothiazide in the PEAR HCTZ monotherapy and PEAR HCTZ add-on ........................................................................................... 93

3-9. The expression levels of EPHX2 by rs13262930 genotype in whole blood collected from White participants within the PEAR HCTZ monotherapy study at baseline .......................................................................................................... 94

3-10. Hydrochlorothiazide response score in PEAR and GERA studies. ..................... 95

4-1. Overall framework analyses ............................................................................. 117

4-2. Illustrates the thirteen genes involved in the sphingomyelin metabolism canonical pathway which were tested in this study........................................... 118

4-3. The effect of rs6078905 polymorphism on the blood pressure response of Whites and Blacks treated with hydrochlorothiazide in the PEAR study ........... 119

4-4. Illustrates the questions required to be answered to further demonstrate the association between SPTLC3 rs6078905 SNP and HCTZ BP response ......... 120

4-5. The effect of rs6078905 polymorphism on sphingomyelin concentrations of N24:2 and N24:3 in Whites treated with hydrochlorothiazide in the PEAR study ................................................................................................................. 121

4-6. The correlation between Sphingomyelin N24:2 and hydrochlorothiazide BP response ........................................................................................................... 122

5-1. Illustrates the involvement of the thiazide diuretics associated signals identified in this project in the smooth muscle regulation mechanism. ............. 130

12

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy

PHARMACOGENOMICS, TRANSCRIPTOMICS AND METABOLOMICS FOR THE

IDENTIFICATION OF NOVEL BIOMARKERS OF BLOOD PRESSURE RESPONSE TO ANTIHYPERTENSIVE DRUGS

By

Mohamed Hossam Shahin

December 2015

Chair: Julie A. Johnson Major: Pharmaceutical Sciences

Hypertension is a significant public health burden and the most common

cardiovascular disease risk factor worldwide. Adequate control and reduction of blood

pressure (BP) has been associated with significant improvement in cardiovascular

morbidity and mortality. Thiazide diuretics, including hydrochlorothiazide (HCTZ), are

among the most commonly prescribed anti-hypertensives globally. Despite their wide

spread use, the long term anti-hypertensive mechanism of thiazide diuretics is still

poorly understood, and global data have shown that < 50% of thiazide treated patients

achieve BP control. Therefore, we aimed in this project to identify novel pathways and

biomarkers associated with thiazides’ BP response, which could provide more insight in

the BP lowering mechanism of thiazide diuretics and improve their BP control rates.

In this project, we used state of the art approaches to integrate different “omics”

(i.e. genomics, transcriptomics, metabolomics and lipidomics), which helped us identify

VASP, PRKAG2, EPHX2, and DCC as potential determinants of thiazides’ BP

response. We provided multiple levels of replication to our findings, which further

substantiates the importance of these replicated signals to thiazides’ BP response.

13

Additionally, the results of this project shed light on several novel pathways (i.e. actin

nucleation, netrin signaling and sphingomyelin metabolism pathways) that were

significantly associated with thiazides’ BP response. These results strongly support that

thiazides’ long term BP lowering mechanism might be mediated via their effect on

several enzymes and pathways regulating the contraction or relaxation of vascular

smooth muscle.

Collectively, the results of this project highlight the strength of using different

“omics” to identify novel pathways and biomarkers associated with drug response.

Perhaps moving forward, functional studies are highly recommended to confirm the

association of the identified genes with thiazides’ BP response. Additionally, further

replication of thiazides’ BP response biomarkers, identified in this project, should be

done in large well-designed studies to further validate their clinical utility for future use.

Moreover, future investigation of the identified pathways and their relation with the

pathophysiology of hypertension and anti-hypertensive BP response might help identify

new targets of hypertension and facilitate the development of new drugs and

therapeutic approaches to better improve BP control and cardiovascular outcomes.

14

CHAPTER 1 MECHANISMS AND PHARMACOGENETIC SIGNALS UNDERLYING THIAZIDE

DIURETICS BLOOD PRESSURE RESPONSE

Hypertension

Hypertension (HTN) is a pervasive and devastating public health threat affecting

more than one billion individuals worldwide, and about one third of United States (U.S.)

adults [1,2]. Additionally, it has been well acknowledged as a leading contributor to

cardiovascular mortality, and a major modifiable risk factor for stroke, coronary heart

disease, heart failure and end stage renal disease, making its management of critical

importance [2]. Data have shown that reducing the diastolic blood pressure (DBP) by

5 mmHg decreases the risk of stroke by 34%, of ischaemic heart disease by 21%, and

reduces the likelihood of heart failure, dementia, and mortality from CV disease [3].

Moreover, HTN represents a major economic burden in the U.S., according to the

American Heart Association, with estimates of the direct and indirect cost for HTN at

$46.4 billion in 2011 [2]. Thus, using effective anti-hypertensive medications for

controlling blood pressure (BP) is essential for reducing cardiovascular risk and the

overall mortality associated with HTN [4].

Thiazide Diuretics

Over the past five decades, thiazide (TZD) diuretics have been a mainstay in the

treatment of HTN, and currently, they are among the most commonly prescribed anti-

hypertensive medications in the US, with approximately 50 million prescriptions in 2014

[5]. According to the current HTN guidelines in the US, this class of drugs is highly

recommended as first line agents for most patients with uncomplicated essential HTN,

alone or in combination with other anti-hypertensive therapy for BP control [6]. Despite

being recommended as an initial and preferred therapy in most hypertensive patients,

15

the underlying mechanism of BP lowering by TZDs has not been fully elucidated, and

data have shown that only about half of TZD treated patients achieve BP control [7,8].

This reveals that the current approach to TZD use and BP control is suboptimal. And

even with the availability of many anti-hypertensive treatment options, data across the

globe suggest that BP control rates are unsatisfactory (less than 50%) [9].

Since the discovery of the first TZD in 1957, many researchers have sought to

understand the precise mechanism underlying TZD’s BP lowering effects, and to

identify predictors that can be used for identifying those patients who will optimally

benefit from this class of drugs. Identifying patient characteristics associated with BP

response to TZDs could increase the control rates to this class of drugs and represent

an improvement over the current “trial and error” approach for selecting drug therapy for

HTN. Some of the promising predictors that have been identified so far include age,

race and baseline levels of plasma renin activity (PRA) [10,11]. However, beyond these

three predictors, there are limited data on any clinical factors that are predictive to

response to TZDs.

Over the past two decades, pharmacogenomics have also been one of the very

active fields that holds promise for identifying more effective ways of differentiating

responders and non-responders to TZDs and other anti-hypertensive medications. Both

candidate and genome-wide association studies (GWAS) conducted to date have

advanced our understanding of the substantial role of genetics on the variability in

response to TZDs [12-39]. Interpreting the results from these studies and identifying

additional novel genetic signals associated with TZD BP response might provide

insights into the mechanism of BP regulation and facilitate the development of new

16

drugs and therapeutic approaches based on a deeper understanding of the molecular

determinants associated with the BP regulatory mechanism.

Herein, the existing knowledge surrounding the BP lowering mechanisms of

TZDs, and the most compelling data from TZD genetic studies are reviewed. By the end

of this Chapter, we also proposed a project that would help in identifying additional

novel signatures and pathways that could provide us with more insights into the BP

lowering mechanism of TZDs, and hold the promise to discover potential new targets for

antihypertensive drug development.

Blood Pressure Lowering Mechanisms of Thiazide Diuretics

Short Term Blood Pressure Lowering Mechanism

TZDs are well known to mediate their diuretic effects via inhibiting the Na+/Cl-

cotransporter (NCC) in the distal convoluted tubule, which consequently increases fluid

loss, leading to a reduction in the extracellular fluid (ECF), and plasma volume and

eventually a decrease in cardiac output and BP [40]. Therefore, the anti-hypertensive

mechanism of TZDs has long been hypothesized to be attributed to their diuretic effect

and enhancement of sodium excretion. In support of this hypothesis, Bennett et al. have

shown that adding 20 g of salt per day to the diet of HTN patients treated with

hydrochlorothiazide (HCTZ) negated the anti-hypertensive effect of HCTZ. Additionally,

TZDs have been shown to be ineffective in end stage renal disease, which supports the

importance of natriuresis for the anti-hypertensive action of TZD diuretics [41]. However,

other evidence contradicts this hypothesis. Specifically, chlorothiazide, a TZD diuretic,

lowered the BP of patients with severe renal failure [42], suggesting the diuretic effects

of TZDs might not be the driving mechanism underlying their BP lowering action.

Consistent with this suggestion, studies have shown that after 4-6 weeks of TZD diuretic

17

initiation, the ECF and plasma volumes return to their normal levels, yet BP reduction is

maintained [43,44]. Collectively, the results from these studies suggest that TZD BP

lowering effects might be initially related, at least in part, to sodium regulation and

reduction in plasma volume and cardiac output (Figure 1-1). However, it seems unlikely

that this is the central mechanism underlying their chronic anti-hypertensive effects.

Long Term BP Lowering Mechanism

Over the past half a century, researchers have been trying to uncover the

mechanism underlying the chronic BP lowering effects of TZDs (Figure 1-1). Many have

indicated that this long-term mechanism is mediated via the reduction in total peripheral

resistance (TPR) [43,45]. However, the precise mechanism and factors underlying this

reduction have not been fully elucidated [46]. Several studies have suggested that TZDs

reduce TPR via a vasodilation effect [47-49]; yet the mechanism by which TZDs dilate

blood vessels has been perplexing and controversial [50].

One hypothesized mechanism is that TZDs’ vasodilatory effects might be

mediated via the endothelium. This hypothesis was supported by an in vitro study

showing that methaclothiazide, a TZD diuretic, inhibited the vasoconstrictive effect of

norepinephrine and vasopressin in the aorta of spontaneously HTN rats, but not in

Wistar-Kyoto (non-HTN) rats [51]. Additionally, this effect was abolished by the removal

of the endothelium or by using a nitric oxide synthase, suggesting that TZDs'

hypotensive effects might be mediated via a nitric oxide endothelium-dependent

mechanism. On the contrary, another study showed that TZDs, at clinically therapeutic

concentrations, inhibit the vasoconstriction effects of norepinephrine and angiotensin II

in the presence or the absence of the endothelium [52]. This study also reported that

TZD induced vasodilation was associated with a significant reduction in RhoA and Rho

18

Kinase expression in the vascular smooth muscle, but no changes in cellular calcium

levels were observed. The changes in expression observed were independent of the

endothelium, suggesting that TZDs act directly on the vascular smooth muscle and not

the endothelium. The authors of this study suggested that the chronic anti-hypertensive

effects of TZDs might be mediated via calcium desensitization that occurs over long

term use of these medications. However, this hypothesis is based only on this study,

and more research is still needed to confirm this hypothesis.

Others suggest that TZD diuretics cause a reduced vasorelaxant effect via

opening the calcium activated potassium channels (KCA). This hypothesis was

supported by results from an in vitro study showing that HCTZ dilates guinea pig

mesenteric arteries, and this effect was abolished by using charbdotoxin, an inhibitor of

the KCA [53]. Additionally, an in vivo study has also shown that HCTZ caused a

vasodilatory effect when injected into human brachial artery, and this effect was

abolished by using tetraethylammonium, a KCA inhibitor[54]. Although this in vivo study

has shown that the vasodilatory effects of TZDs might be mediated via KCA, the TZD

plasma concentrations measured in this study were ~10 times the plasma

concentrations seen clinically in TZD treated patients [55], which brings into question if

this vasodilatory effect underlies the BP lowering in the clinical setting.

Other researchers have proposed that the long term anti-hypertensive effects of

TZDs might be based on their carbonic anhydrase inhibiting properties that produce

alkalosis in the vascular smooth muscle cells. Consequently, this activates the pH

sensitive KCA channels, reduces voltage gated calcium-channels and causes calcium

fall and eventually vasorelaxation. This hypothesis might be intriguing, nevertheless the

19

carbonic anhydrase inhibiting potency of TZDs varies, and we cannot generalize this

mechanism to all of them. Additionally, as seen with other hypotheses, this hypothesis

was supported by studies that had very high plasma concentrations of TZD, which are

considerably higher than those achieved therapeutically [56,57]. It might be possible

that TZDs accumulate in vascular tissues during chronic use, which could account for

the inconsistency between achieved TZD diuretic plasma concentration and those

reported for vasorelaxation; however, further studies are still needed to support this

hypothesis.

More recently, a study conducted by Fei et al. [58] shed light on

epoxyeicosatrienoic acids (EETs) as an important mediator of TZDs’ hypotensive

effects. EETs are known endothelium derived factors that promote vasodilation via

activation of the KCA, leading to hyperpolarization of vascular smooth muscle and

eventually BP reduction. EETs are known to be catalyzed primarily by an enzyme called

soluble expoxide hydrolase (sEH) to a less active vicinal diol called

dihydroxyeicosatrienoic acid (DHET). Fei et al. have shown that indapamide, a TZD-like

diuretic, and HCTZ decreased the protein expression of sEH in HTN rats after 8 weeks

of treatment. They have also reported that indapamide increased the production of

EETs by increasing the mRNA and protein expression levels of CYP2C23, an enzyme

involved in the synthesis of EETs. Although this hypothesis aligns with previously

proposed mechanisms claiming the involvement of the KCA in the long term mechanism

underlying TZD, nevertheless, more evidence is needed to confirm the involvement of

EETs in the mechanism underlying TZD BP response.

20

Pharmacogenetics of Thiazide Diuretics BP Response

Identifying genetic signals with differential response to TZD holds the potential for

optimizing the use of this class of drugs, but more importantly, may also provide insights

in to TZDs’ BP lowering mechanisms. Hence, in this section, we will highlight genetic

signals that have been associated with differences in the BP response of TZDs, and

have been replicated in several independent cohorts. Insights from these

pharmacogenomics signals may provide insight into BP lowering mechanism of TZDs,

identify potential new targets for antihypertensive drug development and be a tool for

precision medicine approaches to treatment.

Neural Precursor Cell Expressed, Developmentally Down Regulated 4 Like (NEDD4L)

NEDD4L is known to encode a ubiquitin ligase enzyme that interferes with

sodium excretion in the kidneys via reducing the renal tubular expression of epithelial

sodium channel (ENAC) [59]. In knock-out mouse models, NEDD4L has been

associated with higher levels of ENaC expression, and salt-sensitive HTN [60]. A

common synonymous single nucleotide polymorphism (SNP) within NEDD4L,

rs4149601G>A, has been reported as an important predictor of HTN, salt sensitivity,

and TZD-BP response [26,61-63]. This SNP causes alternative splicing which was

associated with A-allele carriers having downregulated ENaC expression compared to

G-allele carriers [64]. In consequence, one would expect that individuals carrying the G-

allele of rs4149601 would respond better to TZDs. Data from NORDIL (Nordic

Diltiazem) study were able to confirm this hypothesis and demonstrated that White G-

allele carriers treated with either a TZD diuretic or a β-blocker, for 6 months, had better

SBP and DBP reduction compared to AA carriers in both groups (SBP: -19.5±16.8 vs -

21

15.0±19.3 mmHg, p<0.001, DBP: -15.4±8.3 vs -14.1±8.4, p=0.02, respectively) [65].

This signal was further confirmed to be a TZD specific signal by replicating it in

European Americans (Whites) treated with a TZD diuretic, for 9 weeks, within the

Pharmacogenetic Evaluation of Antihypertensive Responses (PEAR) study [66] (Figure

1-2). On the contrary, no association was observed between rs4149601 and β-blocker

treated patients within PEAR Whites or diltiazem treated White patients within NORDIL,

suggesting that this SNP influences response to TZDs.

In NORDIL, it was also shown that rs4149601 G-allele carriers, treated with both

β-blocker and TZD diuretic, had better cardiovascular outcomes compared to those with

the AA genotype (OR=0.52, 95% CI (0.36-0.74), p<0.0001) [65]. Additionally, in INVEST

(International Verapamil SR Trandolapril) study, White G-allele carriers, not treated with

HCTZ, had a significant increase in cardiovascular events compared to non-carriers

(OR=10.65, 95% CI (1.18-96.25) [66]. Taken together, these data highlight the

importance of NEDD4L as an important predictor for TZD diuretic BP response.

Moreover, it suggests that ENaC and sodium regulation in fact play a role in the long-

term BP-lowering mechanism seen in TZDs. It also suggests that the NEDD4L protein

may represent a novel protein target as an antihypertensive drug. Whether NEDD4L

genotype might be used in the future to guide selection of antihypertensive therapy

remains to be seen, but the data in that regard are promising, particularly since it

associates not only with BP-lowering but also long-term cardiovascular outcomes.

Protein Kinase C Alpha (PRKCA)

PRKCA is a member of the PKC family of serine-threonine specific protein

kinases that have been shown as a fundamental regulator of cardiac contractility and

calcium handling in the myocytes [67], and involved in diverse cellular signaling

22

pathways as vascular smooth muscle contraction and vascular endothelial growth factor

signaling pathways [68]. A GWAS meta-analysis between PEAR and GERA (Genetic

Epidemiology of Responses to Antihypertensive) White HTN participants treated with

HCTZ revealed an intronic SNP, rs16960228, in PRKCA as an important predictor of

HCTZ BP response [37]. The results of this study revealed that rs16960228 A-allele

carriers had a greater BP response compared to GG carriers, which was further

replicated in two other studies, NORDIL and GENRES (Genetics of Drug

Responsiveness in Essential Hypertension Study) (Figure 1-3). The meta-analysis p-

value of rs16960228 from the combined four studies achieved GWAS significance

(p=3.3x10-8). Further functional analysis revealed that rs16960228 A-allele carriers (with

better HCTZ BP-response) had significantly higher baseline expression levels

compared to GG carriers in PEAR Whites (p=0.028). Moreover, rs16960228 was also

significantly associated with DBP response in White PEAR participants treated with a β-

blocker, in an opposite direction to its association with HCTZ BP response (which

further validates this signal given the different pharmacologies of these two drug

classes).

Collectively, rs16960228 replication evidence along with biological relevance and

initial functional validation of the PRKCA gene suggests the potential importance of

PRKCA as an important predictor for TZD BP response. Additionally, the involvement of

the PRKCA gene in calcium handling and vascular smooth muscle contraction pathway

suggest that TZDs’ long-term BP-lowering mechanism might be mediated by acting on

the vascular smooth muscles and/or interfering with calcium handling or sensitivity, as

previously hypothesized [52], via PRKCA. More work on this candidate gene might open

23

new avenues for new drug discoveries and new therapeutic approaches for better

clinical outcomes.

G Protein Alpha Subunit Endothelian-3 (GNAS-EDN3)

The GNAS-EDN3 region has been shown in GWAS meta-analyses to be

associated with HTN and BP [69]. A SNP within this region, rs2273359, has been

reported with a consistent significant association to HCTZ SBP-response in Whites

within PEAR, GERA and NORDIL [37] (Figure 1-4). The combined meta-analysis p-

value of this SNP across the three studies almost reached GWAS significant level

(p=5.5x10-8). These data suggest the importance of this region as a potential

determinant of TZD BP-response, however, more functional and biological evidence is

still needed to better elucidate the link between this region and TZDs’ BP effects. The

association between this genetic region and the effect of TZDs on BP further emphasize

the notion that TZD action is mediated via vascular smooth muscle, given the fact that

both GNAS and EDN3 are involved in the vascular smooth muscle contraction pathway

[68]. Nevertheless, more work on this region is needed, which might provide us with

valuable insights into the complex pathophysiological mechanism underlying HTN.

YEATS Domain Containing 4 (YEATS4)

Using a GWAS approach, Turner and colleagues have also identified a haplotype

signal (constructed from rs317689, rs315135, and rs7297610 near LYZ, YEATS4, and

FRS genes on chromosome 12q) associated with TZD DBP response in Blacks within

the GERA study (P=2.39 x 10-7) [38]. They showed that the ATC haplotype was more

prevalent in Black good responders (p=2x10-4), whereas the ACT and ATT were more

prevalent among Black poor responders (p=0.0018 and 0.0219, respectively). This

haplotype signal was further replicated in hypertensive Blacks treated with HCTZ

24

monotherapy in PEAR [70]. Additionally, single SNP analysis, in PEAR Blacks, revealed

that this haplotype association is driven by rs7297610, which has been shown to affect

the expression levels of YEATS4 gene (Figure 1-5) [70]. Additionally, data have shown

that carriers of the CC genotype (associated with better HCTZ BP response) had a

higher baseline expression compared to T-allele carriers (p=0.024). However, whether

this expression differences play a role in the reduced BP-response observed with HCTZ

therapy still unknown. Thus, the lack of functional and biological evidence associated

with this signal make it hard to interpret how it might be associated with TZD-BP

lowering mechanism. Additional studies are needed to confirm the importance of this

signal to TZDs BP-response and their BP-lowering mechanism.

Summary and Aims of the Project

Collectively, it is clear from these results that additional research is needed to

replicate and confirm currently identified signals and functionally validate the biological

association of many of them with TZDs BP response. It is also clear that with more than

two decades of continuous research, we had few reliable replicated genetic predictors

identified. Even with the recent use of GWAS approaches, limited numbers of genetic

signals were discovered. One limitation to the success in identifying additional novel

genetic markers from GWAS is the stringent genome wide significant p-value (5x10-8)

relative to the small sample sizes of the globally available HTN pharmacogenomics

studies. This suggests that the standard GWAS approach will not be able to yield all or

even the majority of the genetic variance that contributes to variability in TZDs BP

response. International collaborative consortiums, such as the International Consortium

for Antihypertensive Pharmacogenomics Studies (ICAPs; https://icaps-htn.org/), may

advance the field of HTN pharmacogenomics and provide more insight into TZD BP

25

response, and other antihypertensive medications, by validating current signals and

identifying additional novel genetic predictors of antihypertensive BP response.

Nevertheless, it should be obvious that information from the genome alone likely will not

explain the complex mechanisms underlying variability in BP response, as decoding the

DNA is considered only the first step towards understanding the complexity of the

system (Figure 1-6).

Therefore, in recent years, research in the fields of transcriptomics and

metabolomics have been very active and successful in identifying novel biomarkers

associated with different diseases and traits, and to bridge the gap between genomics

and phenotype [71-78]. Additionally, significant progress in the correlation of genetic

variation with transcriptomics and/or metabolomics has been made [79-85]. Moreover,

integrative genomics approaches that utilize functional genomics and network biology

have been developed [79,86-89]. These integrative approaches have been successful

in identifying novel key regulators, pathways, and gene networks that underlie GWAS

findings for various diseases and traits [80,85,90-93]. Thus, we sought in this project to

integrate different “omics” datasets (genomics, transcriptomics, and metabolomics) to

identify novel candidate biomarkers of TZD BP response. We hypothesized that

integrating these hierarchial datasets together will give us more power to identify

candidate biomarkers associated with variability in the efficacy of TZD therapy and

provide more insight in the complex mechanism underlying TZD BP response and BP

regulation. We tested our hypothesis through the following specific aims:

Aim 1: Identify genetic predictors associated with BP response in HCTZ treated

participants using a genomics-transcriptomics integrative approach

26

Aim 2: Identify metabolites associated with BP response in HCTZ treated

participants, and use a genomics-metabolomics integrative approach to better elucidate

the complexity of TZD BP response mechanism

Significance

The results of this study may lead to more optimal approaches to anti-

hypertensive treatment selection and better BP control in the future. Additionally, the

knowledge of potential genetic variants, metabolites and candidate pathways

significantly associated with variability in TZD BP response might facilitate the

development of new drugs and therapeutic approaches based on a deeper

understanding of the molecular determinants of the BP response.

27

Figure 1-1. Known and theoretical blood pressure lowering mechanisms of thiazide

diuretics

28

Figure 1-2. Blood pressure response to hydrochlorothiazide by NEDD4L rs4149601

genotype for White participants in PEAR study. Solid gray bars indicate change in systolic blood pressure (SBP), gray and white lined bars indicate change in diastolic blood pressure (DBP). Values are shown as means ± standard error. Add; additive, DOM; dominant. Reprinted with permission [66]

29

Figure 1-3. Blood pressure response to hydrochlorothiazide by PRKCA rs16960228

genotype for White participants from five independent studies. (A) diastolic blood pressure response. (B) systolic blood pressure response. The blood pressure responses are adjusted for pretreatment blood pressure levels, age, and sex. P-values are for contrast of adjusted means between genotype groups. Reprinted with permission [37]

30

Figure 1-4. Blood pressure response to hydrochlorothiazide by GNAS-EDN3 rs2273359

genotype for White participants from five independent studies. (A) diastolic blood pressure response. (B) systolic blood pressure response. The blood pressure responses are adjusted for pretreatment blood pressure levels, age, and sex. P-values are for contrast of adjusted means between genotype groups. Reprinted with permission [37]

31

Figure 1-5. Blood pressure response to hydrochlorothiazide by rs7297610 genotype in

PEAR African Americans. Values are adjusted for age, sex, and baseline blood pressure. Error bars indicate standard error. *P≤0.05 compared with the common C/C genotype. Reprinted with permission [70]

32

Figure 1-6. The basic flow of genetic information in a cell

33

CHAPTER 2 GENOME WIDE PRIORITIZATION AND GENOMICS TRANSCRIPTOMICS

INTEGRATION REVEAL NOVEL SIGNATURES ASSOCIATED WITH THIAZIDE DIURETICS BLOOD PRESSURE RESPONSE

Introduction

Hypertension (HTN) is the most chronic disease and primary cause of

cardiovascular (CV) morbidity and mortality globally [94,95]. Studies have shown that

the risk for coronary diseases and stroke doubles with every 20 mmHg increase in

systolic BP (SBP) or 10 mmHg increase in diastolic BP (DBP) [96-98]. Accordingly,

using anti-hypertensive medications for controlling BP substantially reduces CV risk and

the overall mortality associated with HTN [4]. Thiazide diuretics, including

hydrochlorothiazide (HCTZ), are ranked among the most commonly prescribed drugs

for the treatment of HTN in the U.S.[5], and are highly recommended as first line agents

for most patients with uncomplicated essential HTN [6,99]. Although they have been

used as anti-hypertensives for more than half a century, the mechanism by which

thiazide diuretics chronically lower BP has not been fully elucidated yet [46,50].

Additionally, studies have shown that < 50% of thiazide treated patients achieve BP

control [7,8]. Even with the use of other anti-HTN medications, acting on a variety of BP

regulatory systems, only 44% of HTN treated patients achieve BP control [9,100,101].

Given these facts, it is clear that the current approach for selecting anti-HTN

medications is suboptimal and more work is still needed to optimize the use of these

drugs.

In the past decade, application of genome wide association studies (GWAS) has

advanced our understanding of the potential role of genetics in variable response to

drugs [102-109]. Using GWAS gave us the opportunity to uncover novel genetic regions

34

associated with HCTZ BP response, like YEATS4 (YEATS domain containing 4)[38] and

PRKCA (protein kinase C, alpha) [37]. Despite these successes, these genetic signals

only explain a small proportion of the genetic contributions to the variability associated

with HCTZ BP response and many more remain to be found. We propose that the

stringent genome-wide statistical threshold limits our success in identifying additional

significant single nucleotide polymorphisms (SNPs) influencing HCTZ BP response,

particularly with the small sample sizes of the globally available HTN pharmacogenetics

datasets. Thus, it is critical to leverage other information to effectively prioritize GWAS

signals, increase replication rates and better understand the mechanism underlying

HCTZ BP response.

Recent studies have shown that GWAS SNPs associated with complex traits are

more likely to be expression quantitative trait loci (eQTLs) [110-112]. Additionally,

studies have demonstrated that the majority (~93%) of previously conducted GWAS

findings lay in non-coding regions [113], and that these SNPs are significantly enriched

in the regions that harbor functional elements, such as transcriptional factor binding

sites (TFBSs), histone modification marked regions, DNase I hypersensitive sites

(DHSs) and eQTLs [114-117]. Accordingly, we hypothesized that prioritizing the GWAS

output based on regulatory functional signals that perturb gene expression might

elucidate novel genetic signals affecting HCTZ BP response. Investigating genes and

pathways where these signals are involved might open new avenues for a better

understanding of the molecular determinants of the BP response.

More recently, whole transcriptomics profiling has been promising in identifying

novel genetic markers and revealing valuable mechanistic insights underlying different

35

diseases [118-120], and responses to drugs [121,122]. Additionally, integrating

genomics with transcriptomics has been successful in revealing novel genetic markers

and pathways underlying different traits [123-125]. Thus, we aimed in the current study

to use data from the Encyclopedia of DNA Elements (ENCODE) project, like

transcriptional factor CHIP-seq, histone CHIP-seq, and DNase I hypersensitivity site

data, along with publically available eQTL data, to prioritize and highlight novel genetic

variants affecting the BP response to HCTZ. We also sought to use a pathway analysis

to integrate the prioritized replicated signals from the GWAS with genes of baseline

expression levels that are significantly different between HCTZ extreme responders to

identify novel pathways and additional key regulators involved the mechanism

underlying HCTZ BP response.

Methods

Pharmacogenomic Evaluation of Antihypertensive Response (PEAR) study

Herein, primary analysis included samples and data from participants recruited

as part of the PEAR trial (clinicaltrials.gov # NCT00246519) [126]. In brief, PEAR was a

prospective, randomized, open-label, multi-center study with one of its primary aims

was to evaluate the role of genetics on the BP response of HCTZ treated participants.

PEAR recruited mild to moderate HTN participants, aged 17-65 years, from the

University of Florida (Gainesville, FL), Emory University (Atlanta, GA), and the Mayo

Clinic (Rochester, MN). After enrollment, all participants had an average of 4 weeks

washout period followed by a randomization to either a monotherapy of 12.5 mg/daily

HCTZ or 50 mg/daily atenolol (β-1-selective blocker) for a duration of three weeks, with

the dose titrated upwards for additional 6 weeks (i.e. HCTZ 25 mg/daily or atenolol 100

mg/daily) in participants with BP > 120/70 mmHg. BP responses were assessed after

36

nine weeks and the other drug was added for another nine additional weeks (i.e. HCTZ

for those on atenolol, and vice versa) in participants whose BP remained > 120/70

mmHg after the single drug.

Pharmacogenomic Evaluation of Antihypertensive Response 2 (PEAR-2) Study

We also used clinical data and biological samples from PEAR-2 participants to

validate and replicate our findings from the PEAR primary analysis. In brief, PEAR-2

was a prospective, multi-center, sequential monotherapy study (clinicaltrials.gov #

NCT01203852) [127] with one of its primary aims was to evaluate the role of genetics

on the BP response to chlorthalidone (CLT; a thiazide like diuretic). PEAR-2 recruited

mild to moderate HTN participants, aged 18-65 years, from the University of Florida

(Gainesville, FL), Emory University (Atlanta, GA), and the Mayo Clinic (Rochester, MN).

All participants had an average 4 week washout period then they started on a

monotherapy of 15 mg/daily CLT for a duration of two weeks, with the dose titrated

upwards for additional 6 weeks (i.e. CLT 25 mg/daily) in participants with BP still >

120/70 mmHg. BP response to CLT was evaluated by subtracting BP measured post

those 8 weeks of CLT treatment minus BP measured pre-CLT therapy. Both PEAR and

PEAR-2 studies were approved by the Institutional Review Board at each study site,

and written informed consent was obtained from all participants.

Thiazide Blood Pressure Response Measurement

The primary analysis of the current study included 228 European American

(White) participants from the PEAR study with their BP measured pre-HCTZ (baseline)

and 9 weeks post-HCTZ therapy. In PEAR, home, office and ambulatory daytime and

night-time were measured, as previously described [126]. Briefly, for home BP, PEAR

participants were asked to measure their BP in triplicates, using Microlife model 3AC1-

37

PC BP monitor (Minneapolis, MN), in the morning upon rising and in the evening before

retiring for at least five of seven days prior the randomization visit (baseline) and the

assessment visit after HCTZ monotherapy. Microlife model 3AC1-PC BP monitor

(Minneapolis, MN) was also used to measure the office BP of HCTZ treated PEAR

participants in triplicate pre- and post-HCTZ. Additionally, participants were also asked

and instructed to measure their BP using a 24hr-ambulatory monitor, Spacelabs model

90207 BP monitor (Redmond WA). Using this monitor, we were able to obtain PEAR

participants’ BP four times per hour during the day and twice per hour during the night

(10 PM to 6 AM). For the analysis of the PEAR HCTZ treated participants included in

this study, we used a composite weighted average of the home, office and ambulatory

daytime and night time, which we showed that it represents a more accurate

measurement of BP response with a better signal-to-noise ratio and more power to

identify genetic predictors of BP response[128].

The replication analysis within this study included 186 White participants from the

PEAR-2 study with their BP measured pre-CLT (baseline) and 8 weeks post-CLT

therapy. Both home and office BP were measured, as previously described [127]. In

brief, participants were instructed to measure their BP at home using Microlife model

3AC1-PC BP monitor (Dunedin, FL). Home BP was measured in triplicates during the

morning and evening, and was accepted if at least five morning and evening

measurements were recorded prior to the baseline (pre-CLT) visit and the assessment

visit post-CLT monotherapy. Office BP was also measured in triplicates using the same

monitor. In PEAR-2, We did not have a composite BP phenotype as the one calculated

for PEAR participants, accordingly, we decided to use home BP response as the most

38

suitable phenotype in PEAR-2 participants as it exhibits less variability and superior

reproducibility to office BP in prediction of prognosis [128-130].

Genotyping

PEAR DNA samples underwent genotyping using the Illumina Human Omni-

1Million Quad BeadChip (Illumina, San Diego CA). Genotypes were called using

GenTrain2 Illumina clustering algorithm in the software package GenomeStudio

(Illumina, San Diego CA). MaCH software (version 1.0.16) was used to impute SNPs

that passed quality control filtering in PEAR, based on HapMapIII haplotypes. SNPs

with minor allele frequency (MAF) less than 3 % or imputation r2 less than 0.3 were

excluded from the analysis. While for PEAR-2, genotyping were conducted using

Human Omni2.5 S BeadChip (Illumina, San Diego CA). Patients from PEAR or PEAR-2

were excluded if sample genotype call rates were below 95%. Additionally, SNPs with a

genotype call rates below 95% were also excluded.

Transcriptomics Profiling

We performed RNA-Seq analyses on both PEAR (discovery) and PEAR-2

(replication) White participants treated with HCTZ and CLT, respectively. For PEAR

RNA-Seq analysis, we collected whole blood samples at baseline from 50 White

participants with extreme BP response to HCTZ. Participants were selected from the

upper and lower quartiles of HCTZ home DBP response (25 poor BP responders and

25 good BP responders). Similarly, for PEAR-2 gene expression, whole blood samples

were collected at baseline from 50 White participants with extreme home DBP response

to CLT (25 poor BP responders and 25 good BP responders) to conduct RNA-Seq

analysis on these samples. Home DBP was used for selection of participants for RNA-

39

Seq analyses since it exhibits less variability and superior reproducibility compared to

office BP [128-130].

PEAR and PEAR-2 RNA were extracted using the PAXgene Blood RNA kit IVD

(Qiagen, Valenica, CA). Selection of poly (A) mRNA from total RNA was performed

using Sera-Mag Magnetic Oligo (dT) Beads (Illumina, San Diego,CA) according to the

manufacturer’s protocol. A total of 100 ng of RNA was then used as template for cDNA

synthesis. Libraries were prepared according to instructions for RNA sample

preparation kit (Illumina, San Diego, CA). DNA clusters were generated using the

Illumina cluster station, followed by 38 cycles of paired-end sequencing on the Illumina

HiSeq 2000, performed at Baylor Human Genome Sequencing Center in Texas. For

data quality control purposes, read duplicates removal was implemented using

Picard (http://picard.sourceforge.net) Mark Duplicates option. The 100 bp reads

generated in the paired-end RNA sequencing were uniquely mapped to the human

reference genome (hg19) using TopHat v2.0.10.

Statistical Analyses

Genomics analysis

GWAS analysis was conducted to test the association of ~1.1 million SNPs with

SBP and DBP responses to HCTZ in 228 White participants from the PEAR study.

PLINK software[131] was used to run the GWAS analysis based on an additive genetic

model that included age, sex, pre-HCTZ BP, and population substructure by considering

the first principal component (PC1) in all our analysis.

A Chi square test with one degree of freedom was used to assess the Hardy-

Weinberg Equilibrium of the SNPs included in the GWAS analysis. Characteristics of

the study participants were analysed using descriptive statistics. Numerical variables

40

were represented as mean ± standard deviation, and categorical variables were

presented as percentages. All statistical analyses were carried out with SAS (version

9.3; SAS Institute) and SPSS software (version 17.0 for windows; SPSS Inc., Chicago,

Illinois, USA).

Genome wide prioritization approach

To address the aims of the current study, two approaches including a genome-

wide prioritization approach and a whole transcriptomics profiling were conducted as

shown in Figure 2-1. For the genome-wide prioritization approach, a total of one

hundred and five SNPs, with p-values <5x10-5, from both HCTZ SBP and DBP PEAR

GWAS analyses were selected for the GWAS prioritization analysis. These SNPs were

then prioritized using the RegulomeDB (http://www.regulomedb.org/) which annotates

SNPs based on high-throughput data sets from the ENCODE Project, along with other

publically available Expression Quantitative Trait Loci (eQTL) data sets, computational

predictions and manual annotations[132]. The RegulomeDB has a scoring system that

categorizes SNPs, ranging from 1 to 6, based on the degree of experimental or

computational evidence and the regulatory functional consequences of tested SNPs.

Category 1 includes variants that are known to be associated with the expression of

target genes (i.e. eQTLs), and is further classified in to subcategories, 1a to 1f,

according to the function of the SNPs and their annotations. In the current study, we

prioritized SNPs using RegulomeDB, and then focused only on the SNPs with a score

of 1, since the lower score indicates stronger evidence for a SNP to be located in a

potentially functional region.

41

Replication of Genome Wide Prioritized Single Nucleotide Polymorphisms

After prioritization, SNPs with a RegulomeDB score 1 were then moved forward

for replication in PEAR-2 Whites treated with CLT. A general linear model was used to

run the genetic analysis in PEAR-2 participants based on an additive genetic model that

included age, sex, pre-CLT BP, and population substructure.

For genome-wide prioritized SNPs that were replicated in PEAR-2, we also

tested the difference in the expression levels of the genes (in which replicated SNPs

where located) between PEAR HCTZ responders and non-responders. We further

validated our findings by testing the expression levels of these genes for replication in

PEAR-2 European Americans with extreme BP response to CLT. A logistic regression

analysis was conducted to test the differences in baseline expression levels between

thiazide diuretics extreme responders, in both PEAR and PEAR-2, with adjustment for

baseline variables that had p-values <0.1 in univariate analysis.

Transcriptomics Analysis and Genomics Integration

Parallel to the previous approach, we also conducted whole transcriptomics

profiling to identify genes that were significantly different between HCTZ BP extreme

responders within PEAR. Abundance comparisons between HCTZ BP responders and

non-responders were carried using Cufflinks v2.2.1. Gene expression levels were

reported in fragments per kilobase per million reads (FPKM). To address the problem of

multiple comparisons, we used the false discovery rate (FDR) control statistical method.

FDR-adjusted p-value (Q value) was set at < 0.05 for statistical significance.

Ingenuity pathway analysis software (Ingenuity Systems, www.ingenuity.com,

Redwood City CA) was used to integrate significant genes that were differently

expressed in PEAR RNA-Seq analysis, defined with FDR < 0.05, with the prioritized

42

signals from the genome-wide prioritization analysis. Fischer’s exact test was used to

determine if the significant genes from the PEAR RNA-Seq (FDR<0.05) and the genes

prioritized from the GWAS prioritization analysis were over-represented in significant

canonical pathways (Figure 2-1).

Results

Characteristics of Study Participants and Thiazide Diuretics Blood Pressure Response

Characteristics and thiazide diuretic BP responses of participants, included in the

discovery and replication genetic analyses, are presented in Table 2-1. We found that

baseline characteristics such as age, sex, body mass index (BMI) and baseline BP were

similar among PEAR and PEAR-2 participants. However, we noticed that CLT treated

participants in PEAR-2 had more BP lowering effects compared to HCTZ treated

participants in PEAR. This might be because of the fact that CLT is approximately 1.5 to

2.0 times as potent as HCTZ therapy[133,134].

Characteristics of participants selected from either PEAR or PEAR-2 for the

RNA-Seq analyses are presented in Table 2-2 and 2-3, respectively. For PEAR RNA-

Seq selected participants, baseline characteristics such as age, gender, BMI and

baseline BP were not significantly different between responders and non-responders to

HCTZ. However, in PEAR-2 RNA-Seq selected participants, age was marginally

significant, and gender and baseline BP were both significantly different between CLT

responders and non-responders. Accordingly, we adjusted for age, gender and baseline

DBP in all the analyses conducted using PEAR-2 RNA-Seq data.

43

Genome Wide Prioritization Approach

The Genome-wide prioritization analysis revealed six SNPs with a score of 1

(Table 2-4) according to the RegulomeDB scoring system (Figure 2-2). Out of those six

SNPs, three SNPs (rs10995, rs4802260, and rs4803830) were eQTLs to a gene called

Vasodilator-Stimulated Phosphoprotein gene (VASP), two others (rs6083536, and

rs6083538) were eQTLs to a gene called Protein Tyrosine Phosphatase, receptor type

A (PTPRA), and rs654997 was an eQTL to Chitinase 3 like 2 (CHI3L2) gene. Testing

the linkage disequilibrium (r2) between these 6 SNPs revealed the high LD (r2>0.8)

between rs10995, rs4802260 and rs4083830 SNPs (Figure 2-3 A), as well as the high

LD (r2>0.8) between rs6083536 and rs6083538 SNPs (Figure 2-3 B). Because of the

high LD between the prioritized SNPs, we selected a representative SNP from each

block to move forward for replication in PEAR-2. From the first block, shown in Figure 2-

3 A, rs10995 was selected since it had the highest score, 1d, according to the

RegulomeDB compared to the other two SNPs in the same haplotype block. From the

second block, rs6083536 was selected since it was present on the chip used for

genotyping PEAR-2 participants, while rs6083538 genotyping information was absent.

Lastly, we also moved forward the third SNP, rs654997, to be tested for replication in

PEAR-2.

Replication of Genome Wide Prioritized Single Nucleotide Polymorphisms

Out of the three SNPs that moved forward for replication, only VASP rs10995

SNP replicated in PEAR-2 CLT treated patients with a significant association to SBP

response and trending toward significance with DBP response to CLT (Figure 2-4).

Conversely, rs6083536 or rs654997 did not replicate in PEAR-2. Testing the difference

in the VASP baseline expression levels, the gene in which rs10995 is located, between

44

HCTZ extreme responders in PEAR revealed significantly higher baseline expression

levels in HCTZ responders compared to non-responders (p=8x10-3; Figure 2-5 A).

These results were further replicated in PEAR-2 where we also found that CLT

responders had a significantly higher baseline expression levels of the VASP gene

compared to CLT non-responders (1-sided p=0.02; Figure 2-5 B). Additionally, we also

replicated the effect of the rs10995 SNP on the VASP gene in PEAR White participants

where we found that G allele carriers (with better response to HCTZ) had higher

baseline expression levels of VASP compared to GA and AA carriers (p=3x10-3; Figure

2-6). Collectively, these results highlighted VASP as a potential genetic marker that

might be involved in the mechanism underlying thiazide diuretics anti-hypertensive

effect. Therefore, we sought to integrate this signal with the top significant genes that

are differentially expressed between thiazide diuretics extreme responders in the

transcriptomics analysis, as shown below, to better identify pathways that could help us

understand how this gene might be involved in the BP lowering mechanism of thiazide

diuretics.

Transcriptomics Analysis and Genomics Integration

PEAR RNA-Seq analysis uncovered 14 genes that were significantly different

between HCTZ BP responders and non-responders (FDR<0.05). Integrating these 14

genes with the VASP gene, identified from the previous approach, revealed significant

pathways including the actin-nucleation by ARP-WASP complex pathway (p=4x10-6)

and the Integrin Signalling pathway (p=1x10-4) as the top significant pathways, where

the VASP gene overlapped with the RhoB (RNA-Seq expression p=5x10-5, Figure 2-7

A) and CDC42EP2 (RNA-Seq expression p=5x10-5, Figure 2-7 B) genes in both

pathways. Testing the expression levels of the RhoB and CDC42EP2 in PEAR2

45

revealed a significant difference in the baseline RhoB expression levels (p=0.03, Figure

2-7 C), but not CDC42EP2 (p=0.53, Figure 2-7 D), between CLT extreme BP

responders, with adjustment for age and gender. However, the expression differences

with RhoB in PEAR2 may have been influenced by baseline differences in DBP

between responders and non-responders as adjustment for this variable led to an

increase in the p value (p=0.2).

Discussion

Thiazide diuretics have been the mainstay anti-HTN therapy for years and are

currently ranked among the most commonly prescribed first line anti-HTN in the US.

Despite their wide spread use, a wide inter-individual variability in response to thiazide

diuretics has been reported, which has stimulated interest in identifying predictors that

can be used for optimizing the BP response of this therapy. Over the past years, results

from GWAS have advanced our understanding of the potential role of genetics on the

inter-individual variability in response to different drugs[125,135], including thiazide

diuretics[38],[37]. However, GWAS stringent statistical thresholds hinder the discovery of

additional true genetic variants that are difficult to ascertain statistically, particularly with

the small sample sizes of the globally available HTN pharmacogenetic studies. Thus, in

the current study, we sought to identify additional novel genetic variants associated with

thiazide diuretic BP response by leveraging functional data generated from the

ENCODE project and other publically available eQTL datasets. We hypothesized that

this approach will help us prioritize the genetic signals from the GWAS and increase our

chances to refine and identify the true signals that might be missing.

We showed that using this genome-wide prioritization approach helped us to

identify six eQTLs SNPs affecting three different genetic regions (VASP, PTPRA and

46

CHI3L2). Out of these SNPs, only rs10995, affecting the VASP expression gene, was

replicated in PEAR-2 participants. Testing the baseline expression levels of this gene in

PEAR thiazide diuretics extreme responders revealed significantly higher baseline

expression levels in HCTZ BP good responders compared to non-responders.

Additionally, we replicated this finding in PEAR-2, where we showed that CLT BP

responders had significantly higher baseline expression levels of the VASP gene

compared to non-responders.

VASP protein is a member of the ENA/VASP protein family that is involved in

actin polymerization, actin cytoskeleton regulation and intracellular pathways of integrin-

extracellular matrix (ECM) interaction[136] - pathways that are involved in the

mechanism underlying smooth muscle contraction and BP regulation[137,138]. VASP

has also been well characterized as a substrate for cAMP- and cGMP-dependent

protein kinases[139], which are known for their important role in regulating the

contraction of vascular smooth muscle[140]. Additionally, VASP phosphorylation has

been implicated in various cellular responses ranging from endothelial cell permeability

and angiogenesis[141,142] to platelet aggregation and secretion[143,144]. Moreover, it

has been well known as a biochemical marker for monitoring nitric oxide stimulated

soluble guanylyl cyclase/cGMP-dependent protein kinase type I pathway[145], which in

involved in BP regulation, vascular remodelling and platelet, cardiac and kidney

function[146-148]. Collectively, the results of this study along with these other data

suggest that VASP might be involved in the BP lowering mechanism of thiazides and

might be an important determinant of the thiazide diuretic BP response.

47

The initial mechanism of HCTZ BP lowering is known to involve inhibiting the

Na+/Cl- co-transporter (NCC) in the distal convoluted tubule within the kidney. This

inhibition initially contracts plasma volume and decreases cardiac output leading to BP

lowering, however, the plasma volume and cardiac output returns to normal after 4-6

weeks of thiazide initiation. This reveals that the long term BP lowering effects of HCTZ

might be controlled by other unknown mechanisms. In the current study, the results

from genome-wide prioritization approach, the transcriptomics profiling, and the

integration of RNA-Seq top significant genes with the VASP gene highlighted the actin-

nucleation by ARP-WASP complex pathway and the integrin signaling pathways as

potential pathways that might be involved in the mechanism underlying HCTZ BP

response. In PEAR HCTZ treated participants, we were able to identify the VASP, RhoB

and the CDC42EP2 genes, which are all known to be involved in the actin cytoskeleton

dynamics within the smooth muscle, as potential genes affecting thiazide diuretics BP

response. These results might support the hypothesis that thiazide diuretics long term

BP lowering mechanism might be performed via their direct vasodilatory effect on the

vascular smooth muscle[52]. Of note, there were multiple levels of replication for VASP,

but not for RhoB or CDC42EP2 in PEAR2. However, this does not negate the potential

importance of the actin-nucleation pathway by ARP-WASP complex and the integrin

signalling pathways as potential pathways that might be involved in the BP lowering

mechanism of thiazide diuretics.

Our study has several limitations. First, the small sample sizes used for the

genetic or transcriptomics analyses limited our power to identify additional novel

markers and replicate some of our genetic and transcriptomics signals. Secondly, we

48

used whole blood based RNA for the RNA-Seq analyses, which might not be the perfect

tissue. However, response to anti-HTN drugs might arise from a variety of target

tissues, thus, it is very difficult to select a specific tissue for testing the expression of BP

genes as BP genes might be expressed in blood, heart, brain, kidney or any other

specific tissue. Additionally, such tissues are essentially impossible to obtain from

otherwise healthy hypertensive individuals. Lastly, sodium intake data were not

collected in PEAR or GERA participants. Since salt intake is known to play a major role

in HCTZ response and HTN progression, this may have been a confounding variable for

BP response.

Our study has also several strengths. First, the genome-wide prioritization

approach used in this study was successful in identifying novel genetic variants that we

were not able to identify using traditional approaches for analysing the GWAS output.

Our findings from this approach support the hypothesis that not all SNPs are equal[149],

and mandates the importance of utilizing functional annotations to better prioritize the

GWAS output and increase the chances of identifying and replicating true signals

associated with variability in drug response. We believe that using such an approach

could help us to take forward the large investment in GWAS and convert the output of

this approach to identify additional genetic variants, and biologically relevant pathways

associated with drug response. Secondly, integrating the results from the RNA-Seq

transcriptomics profiling with the VASP finding further confirmed the importance of the

VASP and its involvement with thiazide diuretic BP response. Additionally, it highlighted

the actin-nucleation by ARP-WASP complex and the integrin signalling pathways as

significant pathways that might be involved in the mechanism underlying HCTZ

49

antihypertensive effect. Future research on those pathways might help better

understanding the long term mechanism underlying HCTZ BP lowering effect and might

identify targets for novel anti-hypertensives.

In summary, up to our knowledge, this is considered the first study to highlight

the importance of the VASP gene in thiazide BP response. Additionally, the results of

this study highlight several novel pathways significantly associated with thiazide

diuretics BP response. Moreover, this study illustrates the power of utilizing the

RegulomeDB, ENCODE data and eQTL available datasets to prioritize the GWAS

output and increase the probability of identifying novel genetic variants underlying drug

response. Future use of these tools might give us more insight about the mechanism

underlying BP response, which might facilitate the development of new drugs and

therapeutic approaches that can be utilized for optimizing anti-hypertensive BP

response.

50

Table 2-1. Characteristics of PEAR and PEAR-2 participants

Characteristics PEAR HCTZ monotherapy

(N=228)

PEAR-2 CLT monotherapy

(N=186)

Age, mean (SD) years 50 ± 9.5 51.1 ± 8.9

Women, N (%) 91 (40) 79 (42.5)

BMI, mean (SD) kg*m-2 30.30 ± 4.90 30.66 ± 4.95

Pre-treatment home SBP, mean (SD) mmHg

146 ± 9.96 147.43 ± 10.31

Pre-treatment home DBP, mean (SD) mmHg

93.61 ± 5.59 90.28 ± 5.04

Home SBP response, mean (SD) mmHg

-7.68 ± 8.1 -12.25 ± 9.1

Home DBP response, mean (SD) mmHg

-4.23 ± 5.32 -6.83 ± 5.43

Composite SBP response, mean (SD) mmHg

-8.50 ± 7.02 NA

Composite DBP response, mean (SD) mmHg

-4.68 ± 4.79 NA

Continuous variables are presented as mean ± standard deviation (SD); categorical variables are presented as numbers and percentage. PEAR, Pharmacogenomic Evaluation of Antihypertensive Responses; PEAR-2, Genetic Epidemiology of Responses to Antihypertensives; BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure. For PEAR, composite BP response was generated using the weighted average of the home, office, ambulatory daytime and night time BP responses.

51

Table 2-2. Characteristics of PEAR European American participants included in the RNA-Seq analysis

Characteristics PEAR Responders

(N=24)

PEAR Non-responders

(N=25)

P-value

Age, mean (SD) years 48 ± 11.94 47.5 ± 8.70 0.85

Women, N (%) 9 (36.0) 12 (50.0) 0.46

BMI, mean (SD) kg*m-2 29.3 ± 5.13 31.8 ± 5.90 0.12

Pre-treatment home SBP, mean (SD) mmHg

145.6 ± 10.2

144.1 ± 9.82 0.60

Pre-treatment home DBP, mean (SD) mmHg

93.8 ± 5.0 94.1 ± 4.37 0.81

Home SBP response, mean (SD) mmHg -11.80 ± 6.85

-0.9 ± 5.85 3x10-7

Home DBP response, mean (SD) mmHg -8.36 ± 5.57 0.08 ± 3.65 1x10-7

Composite SBP response, mean (SD) mmHg

-12 ± 6.44 -4 ± 5.60 2x10-5

Composite DBP response, mean (SD) mmHg

-7.56 ± 4.84 -1.6 ± 3.69 1x10-5

Table 2-3. Characteristics of PEAR-2 European American participants included in the

RNA-Seq analysis

Characteristics PEAR-2 Responders

(N=25)

PEAR-2 Non-responders

(N=24)

P-value

Age, mean (SD) years 53.40 ± 7.80 48.38 ± 10.40 0.064

Women, N (%) 15 (60) 5 (20.8) 9x10-3

BMI, mean (SD) kg*m-2 32.60 ± 5.14 30.69 ± 5.31 0.21

Pre-treatment home SBP, mean (SD) mmHg

152.23 ± 10.9 143.97 ± 9.2 6x10-3

Pre-treatment home DBP, mean (SD) mmHg

96.78 ± 6.59 92.93 ± 5.12 0.027

Home SBP response, mean (SD) mmHg

-21.66 ± 7.67 -1.21 ± 4.85 1x10-14

Home DBP response, mean (SD) mmHg

-14.25 ± 4.22 -0.27 ± 2.28 8x10-19

52

Table 2-4. Genetic signals prioritized according to their potential function using RegulomeDB

Chromosome SNP ID RegulomeDB Score

eQTL gene

DBP p-value

SBP p-value

Chr19 rs10995 1d VASP 1.03E-05 6.21E-05

Chr19 rs4802260 1f VASP 1.03E-05 6.21E-05

Chr19 rs4803830 1f VASP 7.11E-06 4.49E-05

Chr20 rs6083536 1f PTPRA 7.52E-07 2.19E-07

Chr20 rs6083538 1f PTPRA 1.11E-06 2.07E-07

Chr1 rs654997 1f CHI3L2 2.49E-03 3.85E-05

Insilico analysis was conducted using the RegulomeDB (http://www.regulomedb.org/). SNP: single nucleotide polymorphism, VASP: vasodilator stimulated phosphoprotein, PTPRA: Protein Tyrosine Phosphatase, receptor type A, CHI3L2: Chitinase 3 like 2, eQTL: expression quantitative trait loci, SBP: systolic blood pressure, DBP: diastolic blood pressure

53

Figure 2-1. Represents the overall framework of the experimental approaches used in

this study

54

Figure 2-2. RegulomeDB scoring scheme.

55

Figure 2-3. Linkage disequilibrium plots between the six prioritized genetic signals from

the genome-wide prioritization approach. Linkage disequilibrium is presented in r2 values using Haploview[150]. A) Linkage disequilibrium between rs10995, rs4802260, and rs4803830. B) Linkage disequilibrium between rs10995, rs4802260, and rs4803830.

56

A/A (n=125) G/A (n=89) G/G (n=14)-15

-10

-5

0

P=6x10 -5

VASP rs10995 Genotypes

HC

TZ S

BP

res

pons

e (m

mH

g)

A/A (n=99) G/A (n=76) G/G (n=10)-20

-15

-10

-5

0

P=0.015*

VASP rs10995 Genotypes

CLT

SB

P r

espo

nse

(mm

Hg)

A/A (n=125) G/A (n=89) G/G (n=14)-10

-8

-6

-4

-2

0

P=1x10 -5

VASP rs10995 Genotypes

HC

TZ D

BP

res

pons

e (m

mH

g)

A/A (n=99) G/A (n=76) G/G (n=10)-10

-8

-6

-4

-2

0

P=0.15*

VASP rs10995 Genotypes

CLT

DB

P r

espo

nse

(mm

Hg)

PEAR PEAR-2

A)

B)

C)

D)

Figure 2-4. The effect of rs10995 polymorphism on the blood pressure response of

Whites treated with thiazide in the PEAR and PEAR-2 studies. Blood pressure responses were adjusted for baseline blood pressure, age, sex, and population substructure, and p-values represented are for contrast of adjusted means between different genotype groups. Error bars represent standard error of the mean. *One sided p-value based on a one-sided hypothesis tested in the replication study. A) systolic blood pressure response to hydrochlorothiazide in the PEAR study. B) diastolic blood pressure response in the PEAR study. C) systolic blood pressure response to chlorothalidone in the PEAR-2 study. D) diastolic blood pressure response to chlorthalidone in the PEAR-2 study.

57

Non-r

esponder

s (n

=25)

Res

ponders

(n=2

4)

0

100

200

300

P=8x10-3

HCTZ Response in PEAR European Americans

VA

SP

Baselin

e E

xp

ressio

n (

FP

KM

)A) B)

PEAR PEAR-2

Non-r

esponder

s (n

=24)

Res

ponders

(n=2

5)

0

50

100

150

200

250P=0.02*

CLT Response in PEAR-2 European AmericansV

AS

P B

aselin

e E

xp

ressio

n (

FP

KM

)

Figure 2-5. Plots showing the difference in the VASP baseline expression levels

between thiazide diuretics extreme responders in the PEAR and PEAR-2 studies. P-values were generated using logistic regression adjusting for age, gender and baseline DBP. *One sided p-value based on a one-sided hypothesis tested in the replication study. A) PEAR study. B) PEAR-2 study.

58

A/A (n=25) G/A (n=21) G/G (n=3)0

100

200

300

P=3x10-3

VASP rs10995 Genotypes

VA

SP

Baselin

e E

xp

ressio

n (

FP

KM

)

Figure 2-6. The expression levels of VASP by rs10995 genotypes in whole blood

collected from PEAR White participants at baseline. P-value was adjusted for baseline blood pressure, age and gender. Error bars indicate standard error of the mean. FPKM: fragments per kilobase per million reads.

59

Non-r

esponder

s (n

=25)

Res

ponders

(n=2

4)

0

20

40

60

80

P=5x10-5

HCTZ Response in PEAR Whites

Rh

oB

baselin

e

exp

ressio

n levels

(F

PK

M)

Non-r

esponder

s (n

=24)

Res

ponders

(n=25

)

0

10

20

30

40

50 P=0.033*

CLT Response in PEAR-2 Whites

Rh

oB

baselin

e

exp

ressio

n levels

(F

PK

M)

Non-R

esponder

s (N

=25)

Res

ponders

(N=2

4)

0

10

20

30

40P=5x10-5

HCTZ Response in PEAR Whites

CD

C42E

P2

baseli

ne e

xp

ressio

n l

evels

(F

PK

M)

Non-R

esponder

s (N

=24)

Res

ponders

(N=2

5)

0

5

10

15

20

25P=0.534

CLT Response in PEAR-2 Whites

CD

C42E

P2

baseli

ne e

xp

ressio

n l

evels

(F

PK

M)

PEAR PEAR-2A)

B)

C)

D)

Figure 2-7. Plot showing RhoB and CDC42EP2 baseline expression levels between

thiazide responders compared to non-responders in the PEAR and PEAR-2 RNA-Seq analyses. A) RhoB in PEAR. B) CDC42EP2 in PEAR. C) RhoB in PEAR-2. D) CDC42EP2 in PEAR-2. Abundance comparisons between hydrochlorothiazide BP responders and non-responders were carried using Cufflinks v2.2.1. P-value was adjusted for age and gender. Error bars indicate standard error of the mean. HCTZ: hydrochlorothiazide, FPKM: fragments per kilobase per million reads.

60

CHAPTER 3 INTEGRATING METABOLOMICS AND GENOMICS UNCOVERS NOVEL PATHWAYS

AND GENETIC SIGNATURES INFLUENCING HYDROCHLOROTHIAZIDE BLOOD PRESSURE RESPONSE: A GENETIC RESPONSE SCORE FOR

HYDROCHLOROTHIAZIDE USE

Introduction

Hypertension (HTN) is a major public health burden affecting more than 1 billion

individuals’ worldwide [94], and about one third of American adults [2]. It is a significant

modifiable risk factor for myocardial infarction, stroke, heart failure, and kidney failure,

making its control of critical importance. Hydrochlorothiazide (HCTZ), a thiazide diuretic,

is among the most commonly prescribed anti-hypertensive medications in the U.S, with

over 50 million prescriptions annually[5]. It is highly recommended as first line treatment

for most patients with uncomplicated essential HTN, and for patients requiring more

than one anti-hypertensive therapy for blood pressure (BP) control[99]. Despite its

importance, patients’ response to HCTZ varies widely, and studies have shown that less

than 50% of HCTZ treated patients achieve BP control[7,8]. This wide inter-individual

variability in response to HCTZ and other anti-hypertensive medications reveals that the

current approach for therapy selection and BP control is suboptimal. Thus, identifying

predictors of BP response to HCTZ and other anti-hypertensive medications, which

could be utilized in therapy selection, would help optimize anti-hypertensive treatment

selection and improve BP control. Additionally, the knowledge of novel biomarkers and

pathways significantly associated with HCTZ BP response might enhance our

understanding of HTN and anti-hypertensive drug mechanisms, which might facilitate

the development of new drugs and therapeutic approaches based on a deeper

understanding of the determinants of the BP response.

61

In the past decade, HTN pharmacogenetic studies have advanced our

understanding of the potential role of genetics in variable response to anti-hypertensive

medications[151]. However, most of these studies focused on candidate genes, which

revealed few reliable predictors of anti-hypertensive efficacy[12]. More recently, genome

wide association studies (GWAS) have been successful in identifying novel genetic

variants associated with the variability in blood pressure (BP) lowering effect of HCTZ

therapy[37,38]. Nevertheless, we believe that the GWAS stringent threshold (p<5x10-8)

limits our success in identifying additional relevant single nucleotide polymorphisms

(SNPs) that might be associated with drug response, particularly with the small sample

sizes of the globally available HTN pharmacogenetics studies. This suggests that the

standard GWAS approach will not be able to yield all or even the majority of the genetic

variance associated with variability in drug response.

In recent years, metabolomics approaches have been successfully employed to

identify novel biomarkers associated with different diseases and traits, and bridging the

gap between genomics and phenotype[71,73,74,78,152]. Additionally, integrating

metabolomics with genomics has been successful in identifying novel key regulators,

pathways, and gene networks for various diseases and traits, including drug

response[80,85,153]. Thus, we aimed in this study to (A) identify metabolites that

significantly influence the BP response to HCTZ, and (B) use a metabolomics-genomics

integrative approach to identify novel genetic variants with significant impact on HCTZ

BP response.

62

Methods

Study Participants

The primary analysis of the current study included clinical data and biological

samples from European American (White) participants (n=228) recruited as part of the

Pharmacogenomic Evaluation of Antihypertensive Response (PEAR) trial

(clinicaltrials.gov # NCT00246519). The study design and objectives of the PEAR study

have been previously described [126]. In brief, PEAR was a prospective study that

recruited mild to moderate hypertensive participants, aged 17-65 years, at the

University of Florida (Gainesville, FL), Emory University (Atlanta, GA), and the Mayo

Clinic (Rochester, MN). After enrollment, all participants had an average 4 weeks

washout period of any anti-HTN therapies, followed by collection of baseline BP data,

along with collection of biological samples. Study participants were then randomized to

receive 12.5 mg of HCTZ daily or 50 mg of atenolol (β-1-selective blocker) daily for

three weeks (Figure 3-1). HCTZ dose was then increased to 25 mg/daily and atenolol to

100 mg/daily for six additional weeks in those with BP >120/70 mmHg. BP response

was assessed, and biological samples were collected, after the nine weeks total

treatment and then the other drug was added for nine additional weeks (i.e. HCTZ for

those on atenolol, and vice versa) for participants with BP still above 120/70 mmHg,

with a similar dose titration step occurred during this add on therapy.

The discovery analysis of this study included PEAR Whites treated with HCTZ

monotherapy (n=228), which will be referred to as HCTZ monotherapy. PEAR Whites

who started HCTZ after atenolol will be referred to as HCTZ add-on. Data from the latter

group of participants (n=214) were used for replication efforts as described under the

validation section.

63

A total of 196 White participants, from the Genetic Epidemiology of Responses to

Antihypertensives (GERA) study (clinicaltrials.gov # NCT00005520), were also used for

the replication efforts within this study. The study design and objectives of the GERA

study have been previously described[154]. In brief, GERA was a prospective study that

recruited hypertensive participants, aged 30 to 59 years, at Emory University (Atlanta,

GA), and the Mayo Clinic (Rochester, MN). After enrollment, all participants had an

average 4 weeks washout period of any anti-HTN therapies followed by a BP

assessment. Participants then started taking 25mg of HCTZ daily for four weeks

followed by another BP assessment.

The PEAR and GERA studies were approved by the Institutional Review Board

at each study site. All participants provided voluntary written informed consent prior to

participation in the study.

Hydrochlorothiazide Blood Pressure Response Measurement

PEAR White participants had their BP measured pre-HCTZ (baseline) and after 9

weeks of HCTZ therapy. Home, office and ambulatory daytime and night-time BP were

measured, as previously described[126]. In brief, the Microlife model 3AC1-PC BP

monitor (Minneapolis, MN) was used to measure home BP in triplicates for at least five

out of seven days prior to participants’ BP assessment visit. Participants were instructed

to measure their BP in the morning upon rising and in the evening before retiring. The

same monitor was used to measure Office BP in triplicate. A 24hr-ambulatory BP was

also measured using Spacelabs model 90207 BP monitor (Redmond WA). For the

analysis of HCTZ monotherapy and HCTZ add-on participants included in this study, we

used a weighted composite BP of the home, office and ambulatory daytime and night

time BP data, which we have shown to be a more accurate measurement of BP

64

response with a better signal-to-noise ratio and more power to identify genetic

predictors of BP response[128].

GERA White participants had their BP measured in triplicate by a trained

assistant using a random zero sphygmomanometer (Hawksley and Sons, Ltd.; West

Sussex, England)[154]. HCTZ BP response was measured by calculating the difference

between post- and pre-HCTZ BP readings.

Untargeted Metabolomics Profiling

Baseline fasting plasma samples from 123 PEAR Whites treated with HCTZ were

used for the metabolomics analysis. Samples were selected based on participants with

a large waist circumference (men ≥ 40 in. and women ≥ 35 in.), and there was a good

representation of HCTZ BP response among participants within the metabolomics

dataset (Figure 3-2). Untargeted metabolite profiling was conducted using Gas

Chromatography-Time-of-Flight Mass Spectroscopy (GC-TOF MS).

Plasma samples were prepared for analysis using a two-step

methoximation/silylation protocol[155]. Briefly, plasma samples stored at -80 °C were

thawed and 15 µL aliquots were extracted using 1 mL of degassed extraction solvent

consisting of acetonitrile:isopropanol:water (3:3:2) at -20 °C, centrifuged, removed the

supernatant and solvents evaporated to dryness under reduced pressure. Membrane

lipids and triglycerides were further removed through a clean-up step where dried

samples were reconstituted with acetonitrile/water (1:1), decanted and taken to dryness

under reduced pressure. Internal standards, C8-C30 fatty acid methyl esters (FAMEs),

were then added to samples and with methoxyamine hydrochloride in pyridine and

subsequently by MSTFA (Sigma-Aldrich) for trimethylsilylation of acidic protons.

65

GC-TOF analysis was conducted using a 6890 gas chromatograph (Agilent

Technologies, Santa Clara, CA) with a CIS4 temperature programmable injector and a

Gerstel MPS2 automatic linear exchange system, which was used to inject 1 µL of

sample at 50 °C (ramped to 250 °C) in splitless mode with a 25 sec splitless time.

Chromatographic separation was performed on a Rtx5Sil-MS column with a 10 m

integrated guard column [95% dimethyl/5%diphenylpolysiloxae film; 30 m x0.25 mm

(inside diameter) x 0.25 µm diphenyl film (Restek, Bellefonte, PA). Chromatography was

performed at a constant flow of 1 ml/min, ramping the oven temperature from 50 °C to

330 °C over 22 min. Mass spectrometry was conducted using a Leco Pegasus IV time

of flight mass spectrometer with 280 °C transfer line temperature, electron ionization at -

70 V and an ion source temperature of 250 °C. Mass spectra were acquired at 20

scans/sec with a mass range of m/z 85-500 and a detector voltage of 1750 V. All

samples were analysed in one batch, and acquired spectra were exported and filtered

for consistency using the UC Davis Metabolomics BinBase database. All database

entries in BinBase were matched against the Fiehn mass spectral library of 1,200

authentic metabolite spectra using retention index and mass spectrum information in

addition to the NIST05 commercial library. Quantitative data were normalized based on

the sum intensities of all structurally identified metabolites.

Genotyping

PEAR DNA samples were genotyped for more than one million SNPs using the

Illumina Human Omni-Quad BeadChip (Illumina, San Diego CA). Genotypes were

called using GenTrain2 Illumina clustering algorithm in the software package

GenomeStudio (Illumina, San Diego CA). For GERA samples, DNA was genotyped for

about five hundred thousand SNPs using Affymetrix GeneChip® Human Mapping 500K

66

Array set. Genotypes were called using Birdseed and Dynamic Modeling

algorithms[156]. Participants from PEAR or GERA were excluded if sample genotype

call rates were below 95%. Additionally, SNPs with a genotype call rates below 95%

were also excluded. MaCH software (version 1.0.16) was used to impute SNPs, in

PEAR and GERA, based on HapMapIII haplotypes[157]. SNPs with minor allele

frequency (MAF) less than 3 % or imputation r2 less than 0.3 were excluded from the

analysis.

Statistical Analyses

The overall analyses framework used in this study is illustrated in Figure 3-3. Our

analyses included seven steps, as described below.

Metabolomics analysis (step1, figure 3-3)

A linear regression analysis was conducted to test the association between the

baseline levels of each structurally identified metabolite (n=212) and HCTZ BP

response of PEAR White participants (n=123), with adjustment for age, sex and

baseline BP. False discovery rate (FDR) with a significance threshold < .05 was used to

account for multiple comparisons. Significant metabolites from this analysis were then

moved forward to be integrated with top signals from the GWAS analysis as described

below.

Genomics analysis (step2, figure 3-3)

A linear regression analysis was used to test the association of approximately 1.1

million SNPs with HCTZ SBP and DBP responses in 228 White participants from the

PEAR study. The GWAS was conducted using PLINK software[131], and the analysis

was based on an additive genetic model that included age, sex and baseline BP as

adjustment variables. A principal component (PC) analysis was conducted where we

67

found no population substructure among the studied participants; however, we forced

the first and second PCs into all analyses. A total of 105 SNPs from both HCTZ SBP

and DBP GWAS analyses, at p<5x10-5, were selected for the genomics-metabolomics

integration analysis, as described below. Of note, the cut off p-value we selected

(p<5x10-5) was based on the quantile-quantile (Q-Q) plots, Figure 3-4, which reveal that

SNPs with p-values <5x10-5 deviated above the diagonal, that is, deviated from the

expectation under the null hypothesis of no relationship between SNPs and HCTZ BP

response in PEAR White treated participants.

A Chi square test with one degree of freedom was used to assess the Hardy-

Weinberg Equilibrium of SNPs at p-value <5x10-5. Characteristics of the study

participants were analysed using descriptive statistics. Numerical variables were

represented as mean ± standard deviation or standard error as described, and

categorical variables were presented as percentages. All statistical analysis was carried

out with SAS (version 9.3; SAS Institute) and SPSS software (version 17.0 for windows;

SPSS Inc., Chicago, Illinois, USA).

Genomics metabolomics integration (step3, figure 3-3)

Ingenuity pathway analysis software (Ingenuity Systems, www.ingenuity.com,

Redwood City CA) was used to integrate the metabolites significantly associated with

HCTZ BP response, defined with FDR less than 0.05, with the top genome wide

association (GWA) analysis (p-value<5x10-5) signals for both systolic BP (SBP) and

diastolic BP (DBP). From this analysis, we focused on significant SNPs/genes and

metabolites converging within the top significant pathway and further confirmed their

association with HCTZ BP response by testing them for replication in PEAR HCTZ add-

on.

68

Our genomics-metabolomics pathway analysis shed light on arachidonic acid

association with HCTZ BP response and its involvement in the top significant pathway.

Since arachidonic acid and its metabolites are well known for their influence on

cardiovascular traits and BP regulation[158,159], we speculated the arachidonic acid

association with HCTZ BP response might also be influenced by SNPs within genes

involved in the arachidonic acid metabolic pathway. Accordingly, we investigated

genetic variants within eleven genes directly involved in the synthesis and degradation

of arachidonic acid and that have been previously reported to be associated with BP

regulation37,38,43,45-59. A total of 60 SNPs, within the candidate gene regions of those

eleven genes, were extracted after linkage disequilibrium SNP pruning. Candidate gene

regions were defined as the full transcript +/- 2kb. Genetic association analyses were

conducted between those 60 SNPs and HCTZ BP response in PEAR HCTZ

monotherapy, with adjustment for age, sex, baseline BP, and population substructure.

An FDR with a significance threshold less than .05 was used to account for multiple

comparisons. From this analysis, we moved SNPs that had an FDR of less than .05, in

either SBP or DBP HCTZ responses, for testing their association in PEAR HCTZ add-

on.

Replication (step4, figure 3-3)

SNPs identified from the genomics-metabolomics integration approach were then

tested for replication in Whites treated with HCTZ within PEAR HCTZ add-on group.

Replication was considered significant if the SNP tested had p-value of <.05 with effects

in the same direction to the original finding. Additionally, to confirm the specificity of the

genetic replicated signals to HCTZ, we also tested their association in Whites treated

with atenolol monotherapy in PEAR study (n=214), looking for either absence of

69

association or an association in the opposite direction, given the different

pharmacological effects between β-1-blockers and thiazide diuretics drug classes.

Create a response score (step5, figure 3-3)

To assess the effect of multiple response alleles on HCTZ BP response, and to

examine the relative contribution of our genetic findings toward our phenotype, we

constructed a genetic response score based on replicated SNPs. The genetic response

score was created based on three replicated SNPs (rs2727563, rs12604940, and

rs13262930) that were identified using the genomics-metabolomics integrative

approach. Points were given for the genotypes of the replicated SNPs in which the

homozygous genotype of each SNP with the greatest BP lowering effect had 2 points,

while heterozygous genotype had 1 point, and homozygous genotype associated with

the worst BP lowering effect had zero, as follows: (A) rs2727563 (PRKAG2) C/C=2

points, T/C= 1 point, T/T= zero point (B) rs12604940 (DCC) A/A=2, G/A=1, G/G=zero

and (C) rs13262930 (EPHX2) C/C=2, C/G=1, G/G=zero. Alleles with BP lowering effect

were then summed up for inclusion in a regression model, with adjustment for age,

gender and baseline BP.

Response score replication (step6, figure 3-3)

To replicate the association of this score with HCTZ BP response, we tested this

score in data from HCTZ treated participants (n=196) within the GERA study. Points

were given for each SNP as before, and alleles with BP lowering effect were summed.

A linear regression model was used to test the association between the response score

and HCTZ BP response, with adjustment for age, sex, baseline BP, and PC1 and 2. A

one sided p-value of less than 0.05, with effects in the same direction to the PEAR

HCTZ monotherapy response score, was considered significant.

70

Functional validation (step7, figure 3-3)

To further understand how the replicated SNPs affect the function of the genes,

we tested the effect of the replicated SNPs on the gene expression of the genes in

which they are located. We did this validation using gene expression data generated

from PEAR HCTZ monotherapy participants and by conducting in silico analyses using

publically available databases to identify the effect of these SNPs in different tissues

(http://gtexportal.org/home/, http://genenetwork.nl/bloodeqtlbrowser/, and

http://www.broadinstitute.org/mammals/haploreg/).

For testing gene expression in PEAR Whites HCTZ monotherapy group, whole

blood samples were collected during the baseline study period using PAXgene Blood

RNA tube IVD (Qiagen, Valenica, CA, USA). Samples were selected for RNA

sequencing analysis based on extreme of HCTZ BP response (25 good responders and

25 poor responders). PAXgene Blood RNA extraction kit IVD (Qiagen, Valenica, CA,

USA) was used to isolate RNA from those samples. Illumina© HiSeq 2000 was used to

conduct whole RNA sequencing, then sequencing reads were aligned to the reference

genome (homo sapiens Hg19) with TopHat2, and gene expression levels were

calculated using cufflinks/cuffdiff and reported as fragments per kilobase per million

reads (FPKM). Out of the 50 samples, one sample failed the quality control of the RNA-

seq and another failed the quality control of the GWAS analysis, ending up with 48

samples (24 good responders and 24 poor responders) included in the analysis. The

association between genes’ baseline expression levels and SNP genotypes were then

tested using a simple linear regression analysis, with adjustment for age, gender and

baseline BP. The genotype groups were coded as follows: (A) rs2727563 (PRKAG2;

protein kinase, AMP-activated, gamma 2 non-catalytic subunit) C/C=2, T/C=1, T/T=0

71

and (B) rs13262930 (EPHX2; Epoxide hydrolase 2) C/C=2, C/G=1, G/G=0. We were not

able to test the effect of the rs12604940 on the DCC (Deleted in Colorectal Cancer)

expression levels since the DCC gene was not expressed in blood.

Results

Characteristics of Study Participants and Hydrochlorothiazde Blood Pressure Response

Baseline characteristics and HCTZ BP responses of participants, included in the

genomics and metabolomics analyses, are presented in Table 3-1. Age, sex and body

mass index (BMI) baseline characteristics were similar between PEAR HCTZ

monotherapy (genomics and metabolomics datasets), PEAR HCTZ add-on and GERA

HCTZ studies. Pre-treatment SBP and DBP were lower within the PEAR HCTZ add-on

participants compared to the PEAR HCTZ monotherapy and GERA participants, due to

atenolol treatment before starting HCTZ therapy, whereas all other groups were

untreated at baseline. Because we have previously shown that baseline BP is the most

significant predictor of BP response[154], we adjusted for baseline BP in all the

analyses. Of note, HCTZ produced greater BP lowering when used as monotherapy in

PEAR HCTZ and GERA HCTZ compared to its use when added to atenolol as HCTZ

add-on therapy.

Metabolomics Analysis (Step1, Figure 3-3)

Using a GC-TOF MS platform, we were able to identify 212 structurally known

and 272 unknown metabolites in fasting plasma samples, from PEAR HCTZ

monotherapy participants, collected at baseline. In our analyses, we only focused on the

known metabolites because unknown metabolites could not be assigned to pathways,

72

and since we used pathway analyses to integrate metabolomics with genomics signals,

we realized that adding the unknowns would not add to our analyses.

Our analyses identified thirteen metabolites, out of the 212, that were significantly

associated with both DBP and SBP responses to HCTZ (FDR<.05), after adjustment for

age, sex and baseline BP (Table 3-2). Those thirteen metabolites were then integrated

with PEAR HCTZ monotherapy GWAS top signals (p<5x10-5) using a pathway

integrative approach as shown below.

Genomics Metabolomics Integration (Step3, Figure 3-3)

A total of 105 SNPs were selected, from PEAR HCTZ monotherapy SBP and

DBP GWAS analyses, based on our suggestive cut off p-value (i.e. p<5x10-5).

Integrating those 105 SNPs with the thirteen significant metabolites identified the netrin

signaling pathway as the top significant pathway (p=1.54x10-5) from this pathway

integrative analysis, with rs2727563 in PRKAG2 and rs12604940 in DCC converging

with the arachidonic acid metabolite in the same pathway (Figure 3-5). We found that

carriers of the PRKAG2 rs2727563 C allele had better responses to HCTZ in a manner

consistent with an additive genetic model (p=2x10-5, Figure 3-6 A). We also found that

DCC rs12604940 carriers of the CC genotypes had a better response to HCTZ

compared to participants with CG and GG genotypes (p=2x10-5, Figure 3-6 B).

We also showed that arachidonic acid is involved in the netrin signaling pathway

and had a significant association with HCTZ BP response (SBP adjusted-p=1x10-4, DBP

adjusted-p=7x10-4; Figure 3-7), after adjustment for age, gender and baseline BP. Since

arachidonic acid and its metabolites have been associated with cardiovascular traits

and BP regulation[158,159], we hypothesized that the arachidonic acid association with

HCTZ BP response might also be mediated via polymorphisms within genes involved in

73

the arachidonic acid metabolic pathway. Therefore, we tested our hypothesis by

investigating SNPs within eleven genes directly involved in the synthesis and

degradation of arachidonic acid and have been previously reported to be associated

with BP regulation[160-171] (Table 3-3). From this analysis, we were able to identify

rs324425, within the candidate genetic region of the FAAH (fatty acid amide hydrolase)

gene, and rs7816586 and rs13262930 in the EPHX2 gene, with statistical significant

association with HCTZ BP response (FDR< .05, Table 3-4). Those three SNPs along

with the PRKAG2 rs2727563 and DCC rs12604940 SNPs were then moved for

replication in PEAR HCTZ add-on participants as shown below.

Replication (Step4, Figure 3-3)

Three SNPs (PRKAG2 rs2727563, DCC rs12604940 and EPHX2 rs13262930),

out of the five tested SNPs, were replicated in the same direction as shown in Figure 3-

6 C, 3-6 D, and 3-8, respectively. The specificity of these three signals to HCTZ BP

response was further confirmed by testing them in Whites treated with another

antihypertensive agent, atenolol, within the PEAR study. We found that none of these

SNPs were significantly associated with atenolol BP response (rs2727563 SBP P=0.24,

DBP P=0.47; rs12604940 SBP P=0.96, DBP P=0.79; rs13262930 SBP P=0.56, DBP

P=0.93), suggesting that these signals might be important determinants of HCTZ BP

response in particular.

Create a Response Score (Step5, Figure 3-3)

Linear regression analysis adjusting for age, sex, baseline BP, and PCs 1 and 2,

revealed that individuals with a higher score had a better HCTZ SBP (p=1x10-8) and

DBP (p=3x10-9) responses compared to lower score participants (Figure 3-10 A, and 3-

74

10 B, respectively). We found that this genetic response score, by itself, explained

11.3% and 11.9% of HCTZ SBP and DBP responses, respectively.

Response Score Replication (Step6, Figure 3-3)

This response score was validated in Whites treated with HCTZ in GERA. Our

analyses showed a significant association with DBP response (1-sided p-value=0.03,

Figure 3-10 C), and a marginally significant association with SBP response (1-sided p-

value=0.07, Figure 3-10 D).

Functional Validation (Step7, Figure 3-3)

Functional validation of the three replicated signals, PRKAG2 rs2727563, DCC

rs12604940 and EPHX2 rs13262930, using gene expression data generated from

PEAR HCTZ monotherapy and by conducting in silico analyses.

Using PEAR expression data, we were able to test the effect of only two SNPs

(rs2727563 and rs13262930 on PRKAG2 and EPHX2 expression levels, respectively.

We were not able to test the effect of rs12604940 on DCC since DCC was not

expressed in the whole blood RNA. Among the two tested SNPs, rs13262930 was the

only SNP that showed a significant association with the baseline expression levels of

EPHX2 gene. We found that individuals carrying the C allele had higher baseline

expression levels of EPHX2 in an additive genetic model (p=0.01, Figure 3-9). In silico

expression quantitative trait loci (eQTL) analyses, http://gtexportal.org/home/, also

suggest rs13262930 significantly affects the expression levels of EPHX2 in blood

(n=338, p=2x10-8) and in other tissues including the left ventricle of the heart (n=190,

p=5x10-14), skeletal muscle (n=361, p=5x10-21), heart arterial appendage (n=159,

p=1x10-6) and aorta (n=197, p=4x10-6) where the C allele carriers had higher EPHX2

expression levels.

75

We did not observe any association between rs2727563 SNP and PRKAG2

baseline expression in PEAR HCTZ monotherapy expression dataset. However,

performing in silico eQTL analysis in large dataset (n=5,311), using http://genenetwork.

nl/bloodeqtlbrowser/, revealed a significant association between rs2727563 and

PRKAG2 expression levels in blood, in which the CC genotype carriers had lower

expression levels than TC and TT genotypes (p=1.2x10-4). This association was also

confirmed in other tissues http://www.broadinstitute.org/mammals/haploreg/ haploreg

.php.

Discussion

Thiazide diuretics, including HCTZ, have been the mainstay anti-HTN therapy for

decades and are currently ranked among the most commonly prescribed medications in

the US. Despite their wide spread use, data across the globe have shown a wide inter-

individual variability in response to this class of drugs, highlighting the need for

identifying predictors that can be used for improving the BP response of this therapy. In

the past decade, results from both candidate gene and GWAS studies have advanced

our understanding of the potential role of genetics in HCTZ BP response. However, only

a few signals, explaining a small percent of the variability in the BP lowering effects of

HCTZ, have been replicated to date. In recent years, metabolomics was successful in

bridging the gap between the genomics and the phenotype[71,73,74,78] and was

powerful in identifying novel biomarkers influencing patients’ variability in response to

different drugs[152,172]. Thereby, we aimed in this study to use this promising tool

along with genomics to identify novel biomarkers influencing the BP response to HCTZ

by investigating both the metabolomics and the genomics profiles of patients treated

with HCTZ.

76

The genomics metabolomics integrative approach used in this study helped us

identify 3 signals PRKAG2 rs2727563, DCC rs12604940, and EPHX2 rs13262930,

significantly associated with HCTZ BP response, replicated in a second cohort and

shown to have functional effects on the expression levels of the genes where they are

located. Using these three replicated signals, we constructed a genetic response score

with a stronger association with HCTZ BP response compared to individual SNPs. This

is not surprising for a complex phenotype, as antihypertensive response, since it is

known to be affected by multiple genetic contributors. This response score, by itself,

explained 11.3% and 11.9% of HCTZ SBP and DBP responses, respectively, and was

further validated in a third independent study, which emphasizes the importance of this

response score and its signals to be considered in future models for guiding the

selection of HCTZ therapy.

HCTZ is known to inhibit the Na+/Cl- co-transporter (NCC) in the distal

convoluted tubule within the kidney. This inhibition initially contracts plasma volume and

decreases cardiac output leading to BP lowering, however, the plasma volume and

cardiac output returns to normal after 4-6 weeks of thiazide initiation. This suggests that

the long term BP lowering effects of HCTZ might be controlled by other unknown

mechanisms. The genomics-metabolomics pathway analysis performed in this study

highlighted the netrin signalling pathway as a significant pathway, including metabolic

and genetic signatures associated with HCTZ BP response. This pathway is activated

by netrins, a class of proteins that play a crucial role in neuronal migration and in axon

guidance. Netrin-1 is the most studied member of the family and has been shown as a

potent endothelial mitogen stimulating the production of nitric oxide via a DCC-ERK1/2

77

dependent mechanism[173]. Additionally, a recent study has shown that netrin-1 and its

receptor, DCC, control sympathetic arterial innervation and play an important role in the

regulation of the blood flow to peripheral organs[174]. Moreover, netrin-1 binding to

specific receptors like DCC has been shown to activate multiple pathways including

MAPKs, PKC, src, Rac and Rho kinase, and focal adhesion kinase[175-178], which all

have been previously reported to be associated with HTN and BP regulation[179-182].

Furthermore, a recent study has demonstrated that netrin-1 activates PRKC alpha, and

FAK/Fyn, which are important for the activation of the ERK, JNK and NF-kB [183]. Of

note, we recently identified and replicated a signal within the PRKC alpha gene with

clinically significant influence on the BP response of HCTZ treated participants in our

GWAS analysis [37]. Collectively, this highlights the importance of the netrin signalling

pathway and suggests that it might be a novel and substantial pathway in which HCTZ

produces its long term antihypertensive effects.

The genomics metabolomics integrative analyses have also identified rs2727563

SNP within the PRKAG2 with a significant association to HCTZ BP response. PRKAG2

has been shown as an important regulator of cellular energy metabolism including de

novo biosynthesis of fatty acids, and also acts as a regulator of cellular polarity by

remodelling the actin cytoskeleton[184]. Additionally, PRKAG2 has previously shown to

be significantly associated with BP[185], ventricular pre-excitation (Wolff-Parkinson-

White syndrome), urate levels[186], chronic kidney disease[187], and left ventricular

hypertrophy resembling cardiomyopathy[188]. Altogether, the literature evidence

supporting the association of the PRKAG2 with BP and cardiovascular diseases, and

the evidence from our results which included the identification and replication of

78

PRKAG2 rs2727563 association with HCTZ BP response suggest PRKAG2 as a

potential determinant of HCTZ BP response. Future work still needed to demonstrate

the mechanistic relation between this gene and HCTZ BP response mechanism.

Our results have also revealed arachidonic acid, within the netrin signalling

pathway, as a significant metabolomic signature influencing the BP response to HCTZ

therapy. Arachidonic acid and its metabolites have been well known for their role in the

regulation of renal vascular tone, BP and sodium transport[159,189]. Testing genetic

variants within genes, directly involved in the synthesis and degradation of arachidonic

acid, revealed EPHX2 rs13262930 SNP, which was further replicated, to be significantly

associated with HCTZ BP response. EPHX2 is well known for encoding the soluble

epoxide hydrolase (sEH) enzyme, which converts epoxyeicosatrienoic acid (EET), a

strong vasodilator and anti-inflammatory compound, to the biologically less active

compound, dihydroxyeicosatrienoic acid (DHET)[190,191]. Studies have shown the

expression of the sEH enzyme is positively correlated with BP and inhibiting this

enzyme increases the production of the EETs and ultimately reduces BP[192]. Our

results revealed that participants carrying the C allele (had better HCTZ BP response)

had higher expression levels of the EPHX2 at baseline. Interestingly, Ma et.al. recently

reported that HCTZ might be mediating its antihypertensive BP response through the

inhibition of the sEH[58]. Accordingly, we propose that the better HCTZ BP response

observed in rs13262930 C allele carriers might be because of the inhibitory effect of

HCTZ to the high baseline expression levels of the EPHX2, encoding sEH. However, in

those participants with lower EPHX2 baseline expression levels, HCTZ had a poor BP

response.

79

Our study has several strengths. To our knowledge, this is the first study to use a

genomics-metabolomics integrative approach to identify novel biomarkers associated

with HCTZ BP response. This integrative approach was successful in identifying novel

genetic variants that we were not able to identify using GWAS data alone. We believe

that using such an approach can help us to take forward the large investment in GWAS

and convert the output of this approach to identify additional genetic variants, and

biologically relevant pathways associated with drug response. Secondly, replicating our

genetic signals and further validating them in another independent study, as a combined

alleles response score, emphasize the importance of our findings and their significant

influential effect on HCTZ BP response.

Our study also has several limitations. First, our samples size was relatively

small which limited our power to identify additional SNPs within the GWAS analysis.

However, integrating the metabolomics and the genomics profiles of participants treated

with HCTZ added to the breadth and the depth of our analyses and helped us to

overcome this limitation and to identify novel genetic signals that we were not able to

identify using GWAS data alone. Secondly, we used whole blood based RNA for testing

the expression levels of EPHX2, which is not the target tissue for HCTZ action.

However, response to anti-HTN drugs might arise from a variety of highly inaccessible

target tissues, which might include vasculature, endothelium, heart, brain or kidney.

Additionally, we were able to replicate our expression results in different tissues (i.e.

EPHX2 rs13262930) using publicly available databases. Lastly, sodium intake data

were not collected in PEAR or GERA participants. Since salt intake is known to play a

major role in HCTZ response and HTN progression, this may have been a confounding

80

variable for BP response. However, to minimize this potential confounder, the

participants were instructed not to change their sodium intake during their participation

in the study.

In summary, to our knowledge, this is the first study to highlight the importance of

the netrin signalling pathway on HCTZ BP response. Future work on this pathway might

provide more insights in the mechanism underlying HCTZ antihypertensive effect, and

help in identifying novel drug targets of new antihypertensive medications. The results

of this study have also shed light on DCC, PRKAG2 and EPHX2 genes as important

determinants of HCTZ BP response. The response score created using SNPs within

these genes should be further tested in other independent cohorts to further confirm its

utility in guiding the selection of HCTZ therapy.

In conclusion, this study illustrates the power of integrating different types of

omics data to identify novel genetic variants underlying drug response. Future use of

this approach would improve the breadth and depth of studying complex phenotypes, as

antihypertensive response, and might provide more knowledge and insight in to the

mechanism underlying BP response. This knowledge might facilitate the development of

new drugs and therapeutic approaches based on a deeper understanding of the

determinants of the BP response.

81

Table 3-1. Characteristics of participants included in the genomics and metabolomics analyses

Characteristics PEAR HCTZ monotherapy (Genomics,

N=228)

PEAR HCTZ monotherapy

(Metabolomics, N=123)

PEAR HCTZ add-on therapy

(Genomics, N=214)

GERA HCTZ monotherapy (Genomics,

N=196)

Age, mean (SD) years 50 ± 9.5 50.7 ± 8.9 50.2 ± 9.2

48.5 ± 7.3

Women, N (%) 91 (40) 57 (46.7%) 98 (42)

84 (43)

BMI, mean (SD) kg*m-2 30.30 ± 4.90 33 ± 4.90 30.23 ± 5.50

31.30 ± 5.57

Pre-treatment office SBP, mean (SD) mmHg

151.80 ± 12.40 153.46 ± 12.24 136.22 ± 14.15

142.70 ± 12.60

Pre-treatment office DBP, mean (SD) mmHg

98.10 ± 5.80 98.12 ± 6.30 86.31 ± 8.74

95.60 ± 5.70

Office SBP response, mean (SD) mmHg

-11.00 ± 12.80 -10.80 ± 12.94

-7.23 ± 12.82 -10.90 ± 13.00

Office DBP response, mean (SD) mmHg

-5.01 ± 7.17 -4.39 ± 6.97 -3.47 ± 8.69 -6.26 ± 8.83

Composite SBP response, mean (SD) mmHg

-8.50 ± 7.02 -9.3 ± 6.90 -6.68 ± 6.54

NA

Composite DBP response, mean (SD) mmHg

-4.68 ± 4.79 -5.11 ± 4.87 -3.79 ± 4.40 NA

Continuous variables are presented as mean ± standard deviation (SD); categorical variables are Continuous variables are presented as mean ± standard deviation (SD); categorical variables are presented as numbers and percentage. PEAR, Pharmacogenomic Evaluation of Antihypertensive Responses; GERA, Genetic Epidemiology of Responses to Antihypertensives; BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure. For PEAR and GERA office BP, hydrochlorothiazide BP response was calculated by subtracting the BP measured post-hydrochlorothiazide minus the BP measured pre-hydrochlorothiazide. For PEAR and PEAR HCTZ add-on composite BP, BP response was generated using the weighted average of the home, office, and ambulatory daytime and night time BP responses.

82

Table 3-2. Thirteen metabolites significantly associated with hydrochlorothiazide blood pressure response of Whites in the PEAR HCTZ monotherapy study

Metabolite name§

DBP p-value

DBP q-value

SBP p-value

SBP q-value

Classification#

Glycolic acid 0.000051 0.0108 0.00021 0.0036

Organic acids

Fumaric acid 0.00049 0.0177 0.00018 0.0036

Organic acids

Arachidonic acid 0.0007 0.0177 0.00017 0.0036

Lipids

Caprylic acid 0.0007 0.0177 0.000018 0.0029

Lipids

Dodecanol 0.001 0.0177 0.000064 0.0029

Lipid

Iminodiacetic acid 0.001 0.0177 0.000068 0.0029

Organic acids

Trihydroxypyrazine NIST 0.001 0.0177 0.000063 0.0029

Organoheterocyclic

Pyrazine 2,5 dihydroxy NIST 0.0008 0.0177 0.000067 0.0029

Organoheterocyclic

2 hydroxyvaleric acid 0.0005 0.0177 0.000107 0.0036

Lipids

Dihydroabietic acid 0.0007 0.0177 0.00022 0.0036

Lipids

Phytol 0.001 0.0177 0.00017 0.0036

Lipids

2 hydroxybutanoic Acid 0.001 0.0177 0.005 0.0236

Lipids

Arabinose 0.002 0.0326 0.00021 0.0036 Organooxygen Compounds

P-values were generated based on linear regression analysis of each metabolite with hydrochlorothiazide blood pressure response, with adjustment for baseline blood pressure, age, and sex. * Correlation coefficient generated from partial correlation analysis of each metabolite False discovery rate (q-value) with a significant threshold less than .05 was used to account for multiple comparisons in both systolic and diastolic blood pressure. SBP, systolic blood pressure; DBP, diastolic blood pressure.

#Metabolites

were classified based on the human metabolome database superclass classification http://www.hmdb.ca/classyfication.

83

Table 3-3. Genes involved in the synthesis and degradation of arachidonic acid

Gene name Gene symbol Evidence

Fatty acid amide hydrolase FAAH

[193,194]

Cytochrome P450, 4A11 CYP4A11

[160,161]

Cytochrome P450, 4F2 CYP4F2

[162,163]

Cytochrome P450, 2J2 CYP2J2

[164,165]

Cytochrome P450, 2C8 CYP2C8

[195,196]

Cytochrome P450, 2C9 CYP2C9

[166,167]

Epoxide hydrolase 2 EPHX2

[168,169]

Arachidonate 12-lipoxygenase, 12S-Type ALOX12

[170,171]

Prostacyclin synthase PTGIS

[197,198]

Prostaglandin E synthase PTGES

[199,200]

Prostaglandin-endoperoxide synthase 2 PTGS2 [201,202]

84

Table 3-4. The effect of the 60 polymorphisms selected from the eleven genes involved in the synthesis and degradation of arachidonic acid on hydrochlorothiazide blood pressure responses

CHR SNP DBP p-value DBP q-value SBP p-value SBP q-value

17 rs312470 0.176 0.5429 0.181 0.7229

17 rs12325805 0.352 0.6697 0.503 0.8383

17 rs1042356 0.493 0.7395 0.495 0.8383

10 rs11572127 0.047 0.348 0.211 0.7229

10 rs1934956 0.058 0.348 0.128 0.5908

10 rs11572181 0.169 0.5429 0.122 0.5908

10 rs11572177 0.412 0.6697 0.775 0.949

10 rs1934985 0.441 0.6963 0.386 0.7959

10 rs10882517 0.506 0.7405 0.277 0.7555

10 rs17110453 0.707 0.8657 0.831 0.9588

10 rs11188148 0.831 0.9521 0.635 0.8583

10 rs1934967 0.014 0.21 0.044 0.36

10 rs2860905 0.046 0.348 0.042 0.36

10 rs4918758 0.057 0.348 0.008 0.16

10 rs12772884 0.148 0.5429 0.198 0.7229

10 rs9332092 0.147 0.5429 0.076 0.4145

10 rs2475376 0.413 0.6697 0.349 0.7959

10 rs9332104 0.932 0.9651 0.299 0.78

10 rs9332113 0.916 0.9651 0.315 0.7875

1 rs11207538 0.084 0.42 0.231 0.7229

1 rs11572311 0.19 0.5429 0.024 0.36

1 rs10789082 0.231 0.6026 0.245 0.7229

1 rs3754205 0.291 0.6207 0.96 0.9763

1 rs7515289 0.297 0.6207 0.632 0.8583

1 rs1155002 0.408 0.6697 0.683 0.86

1 rs11572191 0.841 0.9521 0.931 0.9641

1 rs10493270 0.915 0.9651 0.451 0.7959

1 rs9332982 0.28 0.6207 0.168 0.72

1 rs3890011 0.67 0.8375 0.409 0.7959

19 rs3093089 0.388 0.6697 0.932 0.9641

19 rs3093200 0.552 0.7886 0.596 0.8583

19 rs1558139 0.639 0.8375 0.651 0.8583

19 rs2074901 0.748 0.8976 0.657 0.8583

19 rs3093207 0.829 0.9521 0.903 0.9641

19 rs3093088 0.857 0.9522 0.634 0.8583

19 rs3093153 0.939 0.9651 0.448 0.7959

8 rs7816586 0.001 0.03 0.041 0.36

85

Table 3-4. Continued

CHR SNP DBP p-value DBP q-value SBP p-value SBP q-value

8 rs13262930 0.002 0.04 0.003 0.09

8 rs4149253 0.21 0.5727 0.355 0.7959

8 rs2741334 0.261 0.6207 0.978 0.978

8 rs7341557 0.363 0.6697 0.253 0.7229

1 rs324425 0.00008 0.0048 0.0003 0.018

1 rs324419 0.025 0.3 0.053 0.36

1 rs4141964 0.3 0.6207 0.6 0.8583

1 rs6662982 0.385 0.6697 0.904 0.9641

1 rs11576941 0.638 0.8375 0.592 0.8583

9 rs12004095 0.034 0.34 0.054 0.36

9 rs2302821 0.123 0.5271 0.395 0.7959

9 rs4837404 0.186 0.5429 0.447 0.7959

20 rs927068 0.073 0.3982 0.067 0.402

20 rs11700258 0.112 0.5169 0.2485 0.7229

20 rs491490 0.468 0.72 0.688 0.86

20 rs5602 0.606 0.8375 0.623 0.8583

20 rs6090996 0.662 0.8375 0.81 0.9529

20 rs522962 0.659 0.8375 0.89 0.9641

20 rs476496 0.949 0.9651 0.8 0.9529

20 rs501908 0.98 0.98 0.372 0.7959

1 rs4648276 0.183 0.5429 0.658 0.8583

1 rs2745557 0.287 0.6207 0.439 0.7959

1 rs5275 0.357 0.6697 0.914 0.9641

86

Figure 3-1. Represents the study design of the pharmacogenomic evaluation of

antihypertensive responses (PEAR) study

87

Figure 3-2. Distribution of the systolic blood pressure (SBP) and diastolic blood

pressure (DBP) responses to hydrochlorothiazide in PEAR participants included in the metabolomics analysis

88

Figure 3-3. The overall analyses framework of the study

89

Figure 3-4. Quantile-quantile plots from genome-wide association analysis of blood

pressure response to hydrochlorothiazide in Whites in the PEAR study

90

Figure 3-5. Netrin signaling pathway generated by integrating genomics and

metabolomics data using Ingenuity pathway analysis. *Genes including polymorphisms of p-values less than 5x10-5 from the genome wide association analysis of hydrochlorothiazide blood pressure response. #Metabolites with false discovery rate less than 0.05 from the metabolomics analyses

91

C/C (n=77) T/C (n=113) T/T (n=38)-15

-10

-5

0

P=0.00002

PRKAG2 rs2727563 Genotypes

HC

TZ S

BP

resp

on

se (

mm

Hg

)

C/C (n=62) T/C (n=110) T/T (n=42)-10

-8

-6

-4

-2

0

P=0.015*

PRKAG2 rs2727563 Genotypes

HC

TZ S

BP

resp

on

se (

mm

Hg

)

A/A (n=194) G/A (n=31) G/G (n=3)-10

-5

0

5

10

DCC rs12604940 Genotypes

P=0.00002

HC

TZ D

BP

resp

on

se (

mm

Hg

)

A/A (n=164) G/A (n=50)

-4

-2

0

2

4

P=0.015*

DCC rs12604940 Genotypes

HC

TZ D

BP

resp

on

se (

mm

Hg

)

A)

B)

C)

D)

PEAR HCTZ PEAR HCTZ add-on

Figure 3-6. The effects of rs2727563 and rs12604940 polymorphisms on the blood

pressure response of Whites treated with hydrochlorothiazide in the PEAR HCTZ monotherapy and HCTZ add-on. A): rs2727563 on the systolic blood pressure response in the PEAR monotherapy. B) rs12604940 on the diastolic blood pressure response in the PEAR monotherapy. C) Replicating the effect of the rs2727563 on the systolic blood pressure response in the PEAR HCTZ add-on. D) Replicating the effect of the rs12604940 on the diastolic blood pressure response in the PEAR HCTZ add-on. Blood pressure responses were adjusted for baseline blood pressure, age, sex, and population substructure, and p-values represented are for contrast of adjusted means between different genotype groups. Error bars represent standard error of the mean. *One sided p-value based on a one-sided hypothesis tested in the replication study.

92

0 500 1000 1500 2000 2500-30

-20

-10

0

10

20r=0.30

P=9x10-4

Arachidonic acid peak height ratio

HC

TZ S

BP

resp

on

se (

mm

Hg

)

0 500 1000 1500 2000 2500-20

-10

0

10

r=0.26

P=3x10-3

Arachidonic acid peak height ratioH

CT

Z D

BP

resp

on

se (

mm

Hg

)

A) B)

Figure 3-7. Correlation between hydrochlorothiazide BP response and arachidonic acid

peak height ratio. P-values and correlation coefficient (r-values) were generated using Pearson correlation test. A) systolic blood pressure. B) diastolic blood pressure

93

C/C (n=18) C/G (n=79) G/G (n=131)-20

-15

-10

-5

0

P=0.003

EPHX2 rs13262930 Genotypes

HC

TZ

SB

P r

esp

on

se (

mm

Hg

)

C/C (n=12) C/G (n=63) G/G (n=139)-15

-10

-5

0

P=0.039*

EPHX2 rs13262930 Genotypes

HC

TZ

SB

P r

esp

on

se (

mm

Hg

)

C/C (n=18) C/G (n=79) G/G (n=131)-10

-8

-6

-4

-2

0

EPHX2 rs13262930 Genptypes

P=0.002

HC

TZ

DB

P r

esp

on

se (

mm

Hg

)

C/C (n=12) C/G (n=63) G/G (n=139)-8

-6

-4

-2

0

P=0.065*

EPHX2 rs13262930 Genotypes

HC

TZ

DB

P r

esp

on

se (

mm

Hg

)

A)

B)

C)

D)

PEAR HCTZ PEAR HCTZ add-on

Figure 3-8. The effects of rs13262930 polymorphism on the blood pressure response of

Whites treated with hydrochlorothiazide in the PEAR HCTZ monotherapy and PEAR HCTZ add-on. A) Systolic blood pressure response in the PEAR HCTZ monotherapy. B) Diastolic blood pressure response in the PEAR HCTZ monotherapy. C) Replicating the effect on systolic blood pressure response in the PEAR HCTZ add-on. D) Replicating the effect on diastolic blood pressure response in the PEAR HCTZ add-on. Blood pressure responses were adjusted for baseline blood pressure, age, sex and population substructure, and p-values represented are for contrast of adjusted means between different genotype groups. Error bars represent standard error of the mean. *One sided p-value based on a one-sided hypothesis tested in the replication study.

94

C/C (n=5) C/G (n=16) G/G (n=27)0

5

10

15P=0.015

EPHX2 rs13262930 GenotypesEP

HX

2 B

aselin

e E

xp

ressio

n (

FP

KM

)

Figure 3-9. The expression levels of EPHX2 by rs13262930 genotype in whole blood

collected from White participants within the PEAR HCTZ monotherapy study at baseline. P-value was generated using linear regression analysis, adjusting for age and gender. Error bars indicate standard error of the mean. FPKM: fragments per kilobase per million reads.

95

1 (n=6)

2 (n=34)

3 (n=72)

4 (n=75)

5 (n=35)

6 (n=6)

-15

-10

-5

0

5

Number of BP lowering alleles

P=3x10 -9

HCTZ

DBP

resp

onse

(mm

Hg)

0+1 (n=7)

2 (n=26)

3 (n=66)

4 (n=71)

5 (n=21)

6 (n=5)

-20

-15

-10

-5

0

P=0.031*

Number of BP lowering alleles

HCTZ

DBP

resp

onse

(mm

Hg)

1 (n=6)

2 (n=34)

3 (n=72)

4 (n=75)

5 (n=35)

6 (n=6)

-20

-15

-10

-5

0

P=1x10-8

Number of BP lowering alleles

HCTZ

SBP

resp

onse

(mm

Hg)

0+1 (n=7)

2 (n=26)

3 (n=66)

4 (n=71)

5 (n=21)

6 (n=5)

-25

-20

-15

-10

-5

0

P=0.075*

Number of BP lowering alleles

HCTZ

SBP

resp

onse

(mm

Hg)

A)

B)

C)

D)

PEAR HCTZ GERA HCTZ

Figure 3-10. Hydrochlorothiazide response score in PEAR and GERA studies. A)

Tested against diastolic blood pressure response in the PEAR HCTZ monotherapy. B) Tested against systolic blood pressure response in the PEAR HCTZ monotherapy. C) Tested against diastolic blood pressure response in the GERA study. D) Tested against systolic blood pressure response in the GERA study. Blood pressure responses were adjusted for baseline blood pressure, age, sex and population substructure, and p-values represented are for contrast of adjusted means between different genotype groups. Error bars represent standard error of the mean. *One sided p-value based on a one-sided hypothesis tested in the replication study.

96

CHAPTER 4 SPHINGOMYELIN METABOLIC PATHWAY IMPACTS THIAZIDE DIURETIC BLOOD

PRESSURE RESPONSE: INSIGHTS FROM GENOMICS, METABOLOMICS AND LIPIDOMICS ANALYSES

Introduction

Cardiovascular disease – including heart disease and stroke – is the leading

cause of death globally [203]. According to the 2011 U.S. death rate data, more than

2,150 Americans die of cardiovascular disease each day [2]. Hypertension (HTN) has

long been recognized as one of the leading causes of cardiovascular diseases

worldwide [1], and reduction of high blood pressure (BP) has been associated with

significant improvement in cardiovascular outcomes [204]. Despite the availability of

multiple drug classes for treating HTN, data across the globe suggest that BP control

rates, to any given anti-hypertensive medication, are far from optimal (<50%) [9]. This

fact is likely influenced, in part, by the empirical “trial and error” approach currently used

for selecting anti-hypertensive medications. Thus, researchers have been working for

years to identify new therapeutic approaches, pathways and biomarkers that can be

utilized to better predict the best anti-hypertensive therapy for each patient to optimize

their BP response.

Thiazide diuretics, including hydrochlorothiazide (HCTZ), are one of the most

commonly prescribed anti-hypertensive classes that have long been used as first line

therapy in most patients with uncomplicated essential HTN [6]. Despite the wide spread

use of this class of drugs, there is still a lack of understanding their long term BP

lowering mechanism (Chapter 1). Additionally, global data have shown that only about

half of thiazide diuretics treated patients achieve BP control [7,8]. Thus, more work is

97

still needed to better understand the mechanism underlying this class of drugs and

identify predictors that can be used for optimizing their BP lowering effect.

Over the past decade, pharmacogenomics studies have identified several

promising genetic signatures associated with variability in response to thiazide diuretics

[37,65]. Additionally, metabolomics and lipidomics have been promising innovative

approaches that identified novel pathways and biomarkers of drug response and

provided mechanistic insights for several drugs [153,205,206]. Moreover, integrating

different omics has been shown as a powerful approach that revealed novel signatures,

key regulators and pathways associated with different traits, including drug response

[79,80,207]. Therefore, in this Chapter, we conducted metabolomics pathway analysis

to identify significant pathways associated with HCTZ BP response. We also leveraged

our analyses with genomics and lipidomics data to provide more insight in the

mechanism underlying HCTZ BP response, and validate our findings, respectively.

Methods

Pharmacogenomic Evaluation of Antihypertensive Response Study

Biological samples and clinical data used for genomics, metabolomics or

lipidomics analyses were collected as part of the Pharmacogenomic Evaluation of

Antihypertensive Response (PEAR) trial (clinicaltrials.gov # NCT00246519). The design

and objectives of the PEAR study has been previously described [126]. In brief, PEAR

was a prospective study that recruited mild to moderate hypertensive participants, aged

17-65 years, at the University of Florida (Gainesville, FL), Emory University (Atlanta,

GA), and the Mayo Clinic (Rochester, MN). All participants had approximately 4 weeks

washout period of any anti-hypertensive therapies, and then were randomized to

receive 12.5 mg/daily of HCTZ or 50 mg/daily of atenolol (β-1-selective blocker)

98

monotherapy for three weeks. The HCTZ dose was then increased to 25 mg/daily and

atenolol to 100 mg/daily for six additional weeks if the BP was greater than 120/70

mmHg.

Genetic Epidemiology of Responses to Antihypertensives Study

White participants treated with HCTZ, from the Genetic Epidemiology of

Responses to Antihypertensives (GERA) study (clinicaltrials.gov # NCT00005520),

were used to replicate our genetic finding from the PEAR primary analysis. The study

design and objectives of the GERA study have been previously described [154]. In brief,

GERA was a prospective study that recruited hypertensive participants, aged 30 to 59

years, at Emory University (Atlanta, GA), and the Mayo Clinic (Rochester, MN). After

enrolment, all participants had an average 4 weeks washout period of any anti-HTN

therapies followed by a BP assessment. Participants then started taking 25mg of HCTZ

daily for four weeks followed by another BP assessment.

Both PEAR and GERA studies were approved by the Institutional Review Board

at each study site. All participants provided written informed consent prior to

participation in the study.

Hydrochlorothiazide Blood Pressure Response Measurement

PEAR BP was measured pre-HCTZ (at baseline) and 9 weeks after HCTZ

monotherapy treatment. BP data were obtained from home, office and ambulatory

daytime and night-time BP measurements, as previously described [126]. In brief,

Microlife model 3AC1-PC BP monitor (Minneapolis, MN) was used to measure home

BP in triplicate for at least five out of seven days prior to participants’ BP assessment

visit. Microlife model 3AC1-PC BP monitor (Minneapolis, MN) was also used to

measure Office BP in triplicate. The 24 hr-ambulatory BP measurements was obtained

99

using Spacelabs model 90207 BP monitor (Redmond WA). The BP used in the current

study is a composite weighted average of the home, office and ambulatory daytime and

night-time data, which has been shown to be a more accurate measurement of BP

response with a better signal-to-noise ratio and more power to identify genetic

predictors of BP response [128].

GERA White participants had their BP measured in triplicate by a trained

assistant using a random zero sphygmomanometer (Hawksley and Sons, Ltd.; West

Sussex, England) [154]. HCTZ BP response was measured by calculating the

difference between post- and pre-HCTZ BP readings.

Metabolomics

A total of 123 PEAR Whites treated with HCTZ BP response were included in the

metabolomics analyses. Participants were selected based on high waist circumference

(men ≥ 40 in. and women ≥ 35 in.) with a good representation of the HCTZ BP response

phenotype among selected individuals (Chapter 3). Metabolite profiling was conducted

on plasma samples collected in the fasting state during baseline studies, using Gas

Chromatography-Time-of-Flight Mass Spectroscopy (GC-TOF MS). Plasma samples

were prepared for analysis using a two-step methoximation/silylation protocol [155].

Briefly, 30 µl aliquots were extracted with 1 ml of degassed

acetonitrile:isopropanol:water (3:3:2) at −20°C, centrifuged, aliquoted into two portions

and evaporated to complete dryness. Acetonitrile/water (1:1) was used to remove

membrane lipids and triglycerides and the supernatant was again dried down. Internal

standards C8–C30 FAMEs were added and the sample was derivatized using

methoxyamine hydrochloride in pyridine and subsequently by MSTFA (Sigma-Aldrich)

100

for trimethylsilylation of acidic protons. GC-TOF MS data acquisition and processing

were conducted as previously described [206].

Genomics

A total of 228 White participants treated with HCTZ in the PEAR study were

included the in the primary genetic analysis. Additionally, we used data from 148 African

Americans (Blacks) participants treated with HCTZ therapy, as one of the two

independent cohorts used for the replication efforts in this study. PEAR DNA samples

were genotyped using the Illumina Human Omni-Quad BeadChip (Illumina, San Diego

CA). Genotypes were called using GenTrain2 Illumina clustering algorithm in the

software package GenomeStudio (Illumina, San Diego CA). For GERA samples, DNA

was genotyped using Affymetrix GeneChip® Human Mapping 500K Array set.

Genotypes were called using Birdseed and Dynamic Modeling algorithms [156].

Participants from PEAR or GERA were excluded if sample genotype call rates were

below 95%. Additionally, SNPs with a genotype call rates below 95% were also

excluded. MaCH software (version 1.0.16) was used to impute SNPs, in PEAR and

GERA, based on HapMapIII haplotypes [157]. SNPs with minor allele frequency (MAF)

less than 3 % or imputation r2 less than 0.3 were excluded from the analysis.

Lipidomics

Participants for the lipidomics analyses (n=40) were selected from each quartile

of BP response, defined as the difference in BP after HCTZ treatment from the baseline

BP. Lipidomics profiling was conducted on fasting baseline studies plasma samples

using multi-dimensional mass spectrometry-based shotgun lipidomics (MDMS-SL), as

previously described [208,209]. In brief, a protein assay on each plasma sample was

performed by using the BCA method with bovine albumin as the standard. After 200 µl

101

of plasma from each plasma sample was transferred to a disposable culture borosilicate

glass tube (166100 mm), a premixed lipid solution used as internal standards for

quantification of lipid species was added to each plasma sample based on its protein

concentration. Lipid extracts were prepared by using a modified procedure of Bligh and

Dyer as previously described [209] and each was resuspended in 500 µl of

dichloromethane/methanol (1:1, v/v) which corresponded to a concentration of 3

nmol/µl. A portion of each individual lipid extract (approximately 100 µl) was treated with

LiOMe and followed by being washed with hexane as previously described [210].

A triple-quadrupole mass spectrometer (Thermo Fisher TSQ Vantage, San Jose,

CA, USA) equipped with an automated nanospray apparatus (i.e., Nanomate HD,

Advion Bioscience Ltd., Ithaca, NY) and Xcalibur system software was then utilized as

previously described [211]. Each lipid solution prepared after treatment with LiOMe was

also properly diluted prior to infusion to the mass spectrometer for the analyses of

sphingolipids. The diluted lipid extract was directly infused through the nanomate

device. Typically, a 1-min period of signal averaging in the profile mode was employed

for each survey scan. For tandem mass spectrometry, a collision gas pressure was set

at 1.0 mTorr but the collision energy was varied with the classes of lipids as described

previously [209]. Typically, a 2 to 5-min period of signal averaging in the profile mode

was employed for acquisition of each tandem MS spectrum. All the MS spectra and

tandem MS spectra were automatically acquired by a customized sequence subroutine

operated under Xcalibur software. Mass spectra in survey scanning mode were

acquired after intrasource separation of each prepared and properly diluted lipid solution

as previously described [212]. Ceramide and sphingomyelins species were identified

102

and quantified directly from lipid solutions after treatment with LiOMe or hexane

washing [209,213]. The identified species were quantified using a two-step approach as

previously described [209]. Although this platform measured 9 lipid classes including

choline glycerophospholipid (PC), lysoPC (LPC), ethanolamine glycerophospholipid

(PE), phosphatidylinositol, sphingomyelin (SM), ceramide (CER), triacylglycerol (TAG),

cholesterol and cholesterol esters, our analyses focused only on sphingomyelins and

ceramides, since our metabolomics pathway analyses highlight their metabolic pathway

as a significant pathway associated with HCTZ BP response.

Experimental Approach

Metabolomics pathway analysis (step 1)

The overall analysis approach used in this Chapter consists of four steps that are

presented in Figure 4-1. First, we selected the 13 metabolites that we previously

reported, in Chapter 3, to be significantly associated with HCTZ BP response (FDR

<0.05; step 1). These 13 metabolites were then entered in to a pathway analysis based

on data from Humancyc http://humancyc.org/. Pathway analysis was conducted using

an R-based tool (http://cran.r-project.org/web/packages/MPINet/). A false discovery rate

(FDR) with a significant threshold less than 0.05 was used to account for multiple

comparisons in the pathway analysis. The top significant pathway from the pathway

analysis was then selected and moved forward to step 2.

Genomics association analysis (step 2)

We selected SNPs within the thirteen genes involved in the top significant

pathway (sphingomyelin metabolism pathway; Figure 4-2), identified in step1. Genes’

regions were defined as the full transcript +/- 2kb. A total of 83 SNPs were extracted

after excluding SNPs with MAF<3% and after linkage disequilibrium (LD) pruning. LD

103

pruning was conducted using the PLINK software option (--indep-pairwise 50 5 0.5),

which is based on removing SNPs within a 50-SNP sliding window that shifts 5 SNPs

along with each move, and considering an r2 threshold greater than 0.5. Genetic

analyses were then conducted to test the association between those 83 SNPs and

HCTZ BP response in PEAR Whites. Association analysis was conducted using PLINK

software [131], based on an additive genetic model with age, sex, baseline BP and

population substructure as adjustment variables. A Bonferroni correction with a p-value

less than 6 x 10-4 (0.05/83) was defined as a significant threshold for this analysis. A Chi

square test with one degree of freedom was used to assess the hardy-Weinberg

equilibrium of SNPs included.

Replication (step 3)

SNPs that were significantly associated with HCTZ BP response, in step 2,

where then tested for replication in two independent cohorts of participants treated with

HCTZ. The first group included 148 Blacks treated with HCTZ in PEAR study. The other

group included 196 White participants treated with HCTZ BP response in the GERA

study. Replication was considered significant if the SNP tested had p-value of <.05 with

effects in the same direction to the original finding. Additionally, to confirm the specificity

of the genetic replicated signals to HCTZ, we also tested their association in Whites

treated with atenolol monotherapy in PEAR study (n=214), looking for either absence of

association or an association in the opposite direction, given the different

pharmacological effects between β-1-blockers and thiazide diuretics drug classes.

Validation (step 4)

The SNPs that are identified in this step 2 and replicated in step 3 are located in

genes within the sphingomyelin metabolic pathway, and are also significantly

104

associated with HCTZ BP response. Therefore, we hypothesized that the association

between these significant SNPs and HCTZ BP response might be mediated via their

effect on either sphingomyelins or ceramides, which are involved in the sphingomyelin

metabolic pathway (top significant pathway identified in step1). Thus, to test our

hypothesis and to confirm the association of the sphingomyelin metabolic pathway to

HCTZ BP response, we tested the effect of the replicated SNPs, in step 3, on

sphingomyelins and ceramides, as discussed below.

First, normality of each sphingolipid was tested using Shapiro-Wilk and

Kolmogorov-Smirnov tests. Samples were considered outliers if they were > 4 standard

deviations from the mean and were subsequently removed from the analysis.

Association between each sphingolipid (sphingomyelins or ceramides) and rs6078905

SNP were then performed. For normally distributed sphingolipids, ANOVA test was

used to test the association between each sphingolipid and rs6078905 SNP. On the

other hand, Kruskal-Wallis was used for non-normally distributed sphingolipids. Multiple

linear regression was also used to test the association between each lipid and

rs6078905 SNP, with adjustment for age. Pearson’s correlation was used to assess the

correlation between significant sphingolipids, identified from this analysis, with HCTZ BP

responses. Partial correlation coefficient was also used to test the correlation coefficient

between sphingolipids and HCTZ BP response, with adjustment for age.

Statistical Analyses

Characteristics of the study participants were analysed using descriptive

statistics. Numerical variables were represented as mean ± standard deviation, and

categorical variables were presented as percentages. All statistical analyses were

105

carried out with SAS (version 9.3; SAS Institute) and SPSS software (version 17.0 for

windows; SPSS Inc., Chicago, Illinois, USA).

Results

Baseline characteristics and HCTZ BP responses of PEAR and GERA

participants included in the genomics analyses, and PEAR participants included in the

metabolomics analyses in this study, are described in Table 4-1. In Table 4-2, we also

described the baseline characteristics and HCTZ BP response of the subset of PEAR

participants who were included in the lipidomics analyses. Given the fact that sex

hormones have been shown to exhibit gender associated differences in sphingolipids

levels as sphingomyelins [214,215], thus, we selected only female samples for the

primary analysis of the lipidomics data.

Metabolomics Pathway Analysis

The metabolomics pathway analysis (step 1, Figure 4-1), for the thirteen

metabolites significantly associated with HCTZ BP response, revealed sphingomyelin

metabolism as the top pathway (FDR p-value = 9x10-4, Table 4-3). We then extracted

83 SNPs in thirteen genes directly involved in the sphingomyelin metabolic pathway

(Figure 4-2, Table 4-4), and tested their association with HCTZ BP response, with

adjustment for age, gender, baseline BP and population substructure (step2). After

adjustment for multiple comparisons, we found the rs6078905 SNP associated with

HCTZ SBP (p=4x10-4) and DBP response (p=5x10-4), as shown in Figure 4-3. This SNP

is located within the Serine Palmitoyltransferase, Long Chain Base Subunit 3 (SPTLC3)

gene, which is involved in the rate lim0iting step of sphingolipids synthesis. Patients

carrying the CC genotype of rs6078905 SNP had better responses (SBP/DBP = -11.4/-

6.8 mmHg) compared to those carrying the CT (SBP/DBP= -9/-4.9 mmHg) and TT

106

genotypes (SBP/DBP= -6.7/-3.5 mmHg) (Figure 4-3). Additionally, in silico analysis

using transformed fibroblast cells in the Genotype-Tissue Expression (GTEx) project

http://www.gtexportal.org/home/ also revealed that CC genotype carriers (with better

HCTZ BP response in PEAR) had higher SPTLC3 expression levels than CT and TT

genotypes (p=6x10-11).

Replication

We were also able to replicate the association of this SNP with HCTZ BP

response in PEAR Blacks in which we found a significant association with DBP

response (1-sided p=0.04) and a trend toward significance with SBP response (1-sided

p=0.14; Figure 4-3). Moreover, we found no significant association between this SNP

and β-blocker (atenolol) BP response in White participants treated in PEAR (SBP=0.6,

DBP=0.8), which suggests that the SPTLC3 rs6078905 SNP and the sphingomyelin

pathway might be specific for the BP response of thiazides.

Collectively, these results further support the association between the

sphingomyelin metabolic pathway and HCTZ BP response, however it did not show how

the effect of SPTLC3 rs6078905 SNP on the sphingomyelin metabolic pathway is

mediated. Additionally, these results did not show if sphingomyelins have any effects on

HCTZ BP response (Figure 4-4). Therefore, to answer these questions and further

confirm the association between the sphingomyelin metabolic pathway and HCTZ BP

response, we tested the genetic influence of rs6078905 SNP on sphingomyelins and

ceramides involved in the sphingomyelin metabolic pathway, as shown below.

Validation

First, using a multi-dimensional mass spectrometry-based shotgun lipidomics

approach, we were able to identify 9 lipid classes, as discussed in the methods section.

107

However, we focused only on the sphingomyelins and ceramides measured (27

Sphingomeylins and 23 ceramides), since they are the sphingolipids involved in the

sphingomyelin metabolic pathway. Testing the association between the SPTLC3

rs6078905 SNP and the baseline levels of each sphingolipid revealed a significant

association between SPTLC3 rs6078905 and the baseline levels of sphingomyelin

N24:2 (p=0.0005), and sphingomyelin N24:3 (p=0.0008) (Step 3, Table 4-5), after

adjustment for age. We found that participants carrying the CC genotypes (with better

response to HCTZ) have higher baseline sphingomyelin N24:2 and sphingomyelin

N24:3 compared to CT and TT carriers (Figure 4-5).

To further confirm whether sphingomyelins N24:2, and N24:3 are associated with

HCTZ BP response, we tested their association with HCTZ BP response. This analysis

revealed a significant association between sphingomyelin N24:2 baseline levels and

HCTZ BP response (DBP: r=-0.42, p=0.007, SBP: r=-0.36, p=0.026; Figure 4-6) and a

trend toward significance between sphingomyelin N24:3 baseline levels and HCTZ DBP

response (r=-0.27, p=0.1) and SBP (r=-0.26, p=0.11). These results further support the

importance of the sphingomyelin metabolic pathway in the mechanism underlying HCTZ

BP response and suggest that genetic variants within this pathway might have an

influential effect on the BP response to HCTZ therapy.

Of note, SPTLC3 rs6078905 SNP did not replicate in Whites treated with HCTZ

in GERA. However, this lack of replication in GERA does not negate the importance of

the sphingomyelin metabolism pathway and SPTLC3 rs6078905 association with HCTZ

BP response. We believe that the lack of replication observed in GERA might be due to

108

different reasons including the differences in the BP response phenotypes used in

PEAR (composite BP) compared to GERA (office BP).

Discussion

Thiazide diuretics, including HCTZ, have been a corner stone in treating

hypertensive patients for more than five decades, and currently, they are ranked among

the most commonly prescribed first line anti-hypertensives globally. However, the

mechanism underlying the long term BP lowering effect of this class of drugs is still not

well understood. It is known that thiazides’ BP effect is initially mediated via inhibiting

the Na+/Cl- co-transporter (NCC) in the distal convoluted tubule. In consequence, this

inhibition contracts plasma volume and decreases cardiac output leading to BP lowering

[40]. However, the plasma volume and cardiac output returns to normal after 4-6 weeks

of thiazide initiation [43,44], which reveals that the long-term BP lowering effects of

thiazides might be mediated via other unknown mechanisms.

Herein, we conducted a metabolomics pathway analysis that highlighted the

sphingomyelin metabolism pathway as a pathway that might be involved in the long-

term mechanism underlying HCTZ BP response. Based on the pathway analysis of the

metabolomics data, we selected SNPs from thirteen genes involved in sphingomyelin

metabolism. Our analyses revealed rs6078905 SNP in the SPTLC3 gene as being

significantly associated with HCTZ response. Moreover, leveraging our analyses with

lipidomics data further confirmed the influence of the rs6078905 SNP on HCTZ BP

response phenotype via the sphingomyelin pathway, and shed light on the association

between sphingomyelins and inter-individual variability in BP response to HCTZ.

Sphingomyelin and its metabolites have an influential effect on the vascular tone

[216-219] and have previously been reported to be involved in the mechanism

109

underlying BP regulation [220-223]. Data from both animal and human studies have

shown that disruption in membrane lipids, including sphingomyelins, is closely linked

with impaired ion transport and cytosolic calcium concentrations in various forms of

HTN [224-226]. Moreover, in the vasculature, biologically active sphingomyelin

metabolites, such as sphingosine-1-phosphate (S1P), have been reported as acute

vasoconstrictors in most vessels [227-232]. S1P is a lipid mediator formed by the

metabolism of sphingomyelins [233]. In the kidney, the target organ of thiazides,

intravenous and intrarenal arterial administration of S1P caused renal vasoconstriction

[227,232]. Additionally, studies have shown that S1P, acting via S1P1 receptors,

regulates sodium excretion by affecting transport mechanisms in the renal medulla,

possibly via modulating the activity of the epithelial sodium channel (ENaC) [234].

Studies have also shown that S1P can regulate the activity of various ion

channels, including potassium channels [235-237], which have been previously

proposed to be of importance in the mechanism underlying thiazide diuretics BP

response [53,54]. S1P has also been shown to be involved in the mobilization of

calcium from intracellular stores, influx of extracellular calcium via L-type calcium

channels [238,239] and activation of rho-kinase [219,240,241]. Interestingly, rho-kinase

was previously shown to be involved in the pathogenesis of HTN [242], cardiovascular

diseases [243-245] and hypothesized to be involved in the long term mechanism

underlying thiazides’ BP response [52]. Zhu et. al. have shown that thiazide diuretics

induced vasodilation by reducing the expression of rho-kinase significantly in the

vascular smooth muscle [52]. Therefore, we suggest that sphingomyelins and their

active biological metabolites (i.e. S1P) might be involved in the long term mechanism

110

underlying thiazide diuretics BP response via the rho-kinase pathway. Thus, further

work on sphingomyelins, S1P and their metabolic signaling pathways might provide

more insights into the mechanism underlying BP regulation and facilitate the

development of new anti-hypertensive drugs by identifying new targets of BP regulation.

Our study has several limitations. First, our sample size was relatively small

which limited our power to identify additional novel signals associated with HCTZ BP

response. However, using genomics, metabolomics and lipidomics data from

participants treated with HCTZ added to the breadth and the depth of our analyses and

helped us to identify and confirm the importance of the sphingomyelins metabolic

pathway as a potential pathway associated with HCTZ BP response. Secondly, our

lipidomics data analysis included only females to validate the effect of the rs6078905 on

sphingomyelin metabolic pathway. We selected only one single sex in our lipdiomics

analyses since there are well known differences in sphingomyelins levels attributed to

sex [214,215], that we would not have been able to overcome with our limited sample

size. Therefore, future work should test our findings in males to confirm the role of

sphingomyelins on HCTZ BP response in this gender as well. Lastly, we found a

significant association between SPTLC3 rs6078905 and HCTZ BP response in PEAR,

and replicated this finding in PEAR Blacks, however, this finding did not replicate in

Whites treated with HCTZ in GERA. This lack of replication in GERA might have several

explanations. First, the BP response phenotype used in PEAR analysis was based on a

composite of office, home and ambulatory BP measurements, which we have shown

previously to be a more accurate measurement of BP response with a better signal to

noise ratio [128]. On the other hand, the BP response in GERA participants was based

111

on office measurements, which might have more signal to noise ratio compared to the

composite BP used in PEAR. This discrepancy in measuring BP might be one of the

reasons that contributed to the failure of replicating the rs6078905 signal in GERA,

especially with the small sample size used. Additionally, the rs6078905 might not be the

causal signal; presumably it might be in an LD with a rare causal signal that might be

driving its effect on HCTZ BP response. Therefore, more work is still needed to test if

this SNP or another SNP in LD with this SNP can be used as a predictor for HCTZ BP

response.

Our study also has several strengths. Using metabolomics, genomics and

lipidomics data to identify pathways and markers associated with drug response is an

innovative and powerful approach. We believe that using multiple “omics” approaches,

similar to the one presented here, can help in uncovering novel pathways and

biomarkers that were not identified using GWAS data alone. These pathways and

biomarkers hold the promise to provide more insight in drug response mechanisms and

facilitate the development of new drugs based on deeper understanding of determinants

of drug response phenotypes.

In summary, to our knowledge, this is the first study to highlight the association

between sphingomyelin metabolism pathway and HCTZ BP response. We showed that

this association might be mediated via the effect of polymorphisms within the SPTLC3

gene that influence the production of sphingomyelins, in which we showed a significant

association between the latter and HCTZ BP response. In conclusion, this study

illustrates the importance of sphingomyelin metabolic pathway in HCTZ BP response.

Additional research on this pathway may open new avenues for new drug development

112

and provide us with more insights into the mechanism underlying the long-term BP

lowering effects of thiazide diuretics.

Table 4-1. Characteristics of White PEAR participants involved in the genomics and metabolomics analyses

Characteristics PEAR HCTZ monotherapy (Genomics,

N=228 Whites)

PEAR HCTZ monotherapy

(Metabolomics, N=123 Whites)

PEAR HCTZ Monotherapy (Genomics,

N=148 Blacks)

GERA HCTZ monotherapy (Genomics,

N=196)

Age, mean (SD) years 50 ± 9.5 50.7 ± 8.9 47.4 ± 8.8 48.5 ± 7.3

Women, N (%) 91 (40) 57 (46.7) 92 (62.2) 84 (43)

BMI, mean (SD) kg*m-2 30.30 ± 4.90 33 ± 4.90 31.53 (5.41) 31.30 ± 5.57

Pre-treatment office SBP, mean (SD) mmHg

151.80 ± 12.40 153.46 ± 12.24 151.37 ± 13.44 142.70 ± 12.60

Pre-treatment office DBP, mean (SD) mmHg

98.10 ± 5.80 98.12 ± 6.30 99.23 ± 6.16 95.60 ± 5.70

Office SBP response, mean (SD) mmHg

-11.00 ± 12.80 -10.80 ± 12.94

-15.6 ± 14.37 -10.90 ± 13.00

Office DBP response, mean (SD) mmHg

-5.01 ± 7.17 -4.39 ± 6.97 -9.27 ± 8.67 -6.26 ± 8.83

Composite SBP response, mean (SD) mmHg

-8.50 ± 7.02 -9.3 ± 6.90 -12.61 ± 7.81 NA

Composite DBP response, mean (SD) mmHg

-4.68 ± 4.79 -5.11 ± 4.87 -7.56 ± 5.32 NA

113

Table 4-2. Characteristics of White PEAR participants included in the lipidomics

analyses Characteristics PEAR_HCTZ monotherapy (N=40)

Age, mean (SD) years

49.5 ± 10.20

BMI, mean (SD) kg*m-2

29.34 ± 5.20

Pre-treatment home SBP, mean (SD) mmHg 145.41 ± 10.40

Pre-treatment home DBP, mean (SD) mmHg 92.85 ± 5.75

Composite SBP response, mean (SD) mmHg -11.37 ± 6.04

Composite DBP response, mean (SD) mmHg -5.95 ± 4.36

114

Table 4-3. Significant pathways (FDR <0.05) from the metabolomics pathway analysis Pathway Name P-value Q-value

Sphingomyelin Metabolism 3.94 x 10-5 9 x 10

-4

Visual Phototransduction 9.58 x10

-5 9 x 10

-4

Phospholipases 1.24 x 10

-4 9 x10

-4

Fatty Acid Alpha Oxidation 1.56 x 10

-4 9x10

-4

Ceramide Degradation 2.11 x 10

-4 9.75 x 10

-4

Triacylglycerol Degradation 7.20 x 10

-4 2.76 x 10

-3

Sphingosine and Sphingosine-1-phosphate metabolism

1.29 x 10-3 4.25 x 10

-3

115

Table 4-4. Canonical genes in the sphingomyelin metabolism pathway which we tested the association between the SNPs located in these genes and hydrochlorothiazide blood pressure response

Gene Symbol Gene Name

ASAH1 N-Acylsphingosine Amidohydrolase 1

CERS1 Ceramide synthase 1

DEGS2 Delta(4)-desaturase, sphingolipid 2

KDSR 3-ketodihydrosphingosine reductase

SGMS1 Sphingomyelin synthase 1

SGMS2 Sphingomyelin synthase 2

SGPL1 Sphingosine-1-Phosphate Lyase 1

SMPD1 Sphingomyelin phosphodiesterase 1

SMPD2 Sphingomyelin phosphodiesterase 2

SPHK1 Sphingosine Kinase 1

SPTLC1 Serine palmitoyltransferase, long chain base subunit 1

SPTLC2 Serine palmitoyltransferase, long chain base subunit 2

SPTLC3 Serine palmitoyltransferase, long chain base subunit 3

116

Table 4-5. Top signals from testing the correlation between 50 sphingolipids with SPTLC3 rs6078905 SNP

Lipids Correlation# P-value

§

SM N24:2 0.532 0.0005

SM N24:3 0.519 0.0008

SM N16:1 0.506 0.0011

SM N22:1 0.455 0.0040

SM N22:2 0.437 0.0061

SM N24:1 0.410 0.0105

SM N23:1 0.403 0.0120

SM N20:1 0.387 0.0163

117

Figure 4-1. Overall framework analyses

118

Figure 4-2. Illustrates the thirteen genes involved in the sphingomyelin metabolism

canonical pathway which were tested in this study

119

C/C (n=31) C/T (n=110) T/T (n=87)-10

-8

-6

-4

-2

0

P=5.09x10-4

SPTLC3 rs6078905 Genotypes

HC

TZ D

BP

resp

on

se (

mm

Hg

)

C/C (n=33) T/C (n=83) T/T (n=32)-10

-8

-6

-4

-2

0

P=0.04*

SPTLC3 rs6078905 Genotypes

HC

TZ D

BP

resp

on

se (

mm

Hg

)

C/C (n=31) C/T (n=110) T/T (n=87)-15

-10

-5

0

P=4.06 x 10 -4

SPTLC3 rs6078905 Genotypes

HC

TZ S

BP

resp

on

se (

mm

Hg

)

C/C (n=33) T/C (n=83) T/T (n=32)-20

-15

-10

-5

0

P=0.14*

SPTLC3 rs6078905 Genotypes

HC

TZ S

BP

resp

on

se (

mm

Hg

)

A)

B)

PEAR WhitesC)

D)

PEAR Blacks

Figure 4-3. The effect of rs6078905 polymorphism on the blood pressure response of

Whites and Blacks treated with hydrochlorothiazide in the PEAR study. Blood pressure responses were adjusted for baseline blood pressure, age, sex, and population substructure, and p-values represented are for contrast of adjusted means between different genotype groups. Error bars represent standard error of the mean. *One sided p-value based on a one-sided hypothesis tested in the replication study

120

Figure 4-4. Illustrates the questions required to be answered to further demonstrate the

association between SPTLC3 rs6078905 SNP and HCTZ BP response

121

T/T (n=14) C/T (n=21) C/C (n=4)0

200

400

600 P=0.0005

SPTLC3 rs6078905 Genotypes

N24:2

Co

ncen

trati

on

(n

mo

l/g

)

T/T (n=14) C/T (n=21) C/C (n=4)0

20

40

60

80 P=0.0008

SPTLC3 rs6078905 Genotypes

N24:3

Co

ncen

trati

on

(n

mo

l/g

)

Figure 4-5. The effect of rs6078905 polymorphism on sphingomyelin concentrations of

N24:2 and N24:3 in Whites treated with hydrochlorothiazide in the PEAR study. P-values were generated using a linear regression model adjusted for age

122

0 200 400 600-30

-20

-10

0

r= -0.36, P=0.026

N24:2 Concentration (nmol/g)

HC

TZ S

BP

resp

on

se (

mm

Hg

)

0 200 400 600-20

-15

-10

-5

0

r= -0.42, P=0.007

N24:2 Concentration (nmol/g)

HC

TZ D

BP

resp

on

se (

mm

Hg

)

A) B)

Figure 4-6. The correlation between Sphingomyelin N24:2 and hydrochlorothiazide BP

response. P-values and r values were generated using partial correlation with adjustment for age

123

CHAPTER 5 SUMMARY AND CONCLUSIONS

Thiazide diuretics are cornerstone in the treatment of hypertension (HTN) and

are considered one of the most commonly prescribed anti-hypertensives globally.

However, as we discussed in Chapter 1 of this dissertation, the chronic mechanism

underlying the blood pressure (BP) lowering effect of thiazide diuretics is still not fully

understood. Additionally, data across the globe have shown that only about half of

thiazide (i.e. hydrochlorothiazide; HCTZ) treated patients achieve blood pressure (BP)

control. This lack of response might be influenced, in part, by the empirical “trial and

error” approach currently used for selecting thiazide diuretics and other anti-

hypertensives. Even for other anti-hypertensives, data across the globe suggest that BP

control rates are far from optimal (<50 %), which reveals that the current approach for

anti-hypertensive therapy selection and BP control is suboptimal. Therefore, the overall

goal of this research project was to use state of the art approaches for integrating

different “omics” (i.e. genomics, transcriptomics, metabolomics and lipidomics) to

identify novel biomarkers that can help select thiazide diuretics to patients who will most

likely benefit from this therapy. Additionally, we sought to identify novel pathways that

can provide more insight in the long term mechanism underlying thiazide diuretics BP

lowering effects.

In Chapter 2, we hypothesized that prioritizing genetic variants from genome-

wide association (GWA) analysis based on regulatory functional properties (i.e. variants

affecting gene expression) might help identify novel genetic markers affecting HCTZ BP

response. Our primary analysis included testing the association between more than one

million single nucleotide polymorphisms (SNPs) and HCTZ BP response. Afterwards,

124

we leveraged our analyses using data from the Encyclopedia of DNA Elements

(ENCODE) project (i.e. transcriptional factor CHIP-seq, histone CHIP-seq, and DNase I

hypersensitivity site data) along with publically available expression quantitative trait loci

(eQTL) data to prioritize genetic signals from the HCTZ GWA analysis. Using this

approach, we were able to identify a significant association between rs10995 SNP,

within the Vasodilator Stimulated Phosphoprotein (VASP) gene, and HCTZ BP

response. This association was further replicated in another independent study (PEAR-

2). We also found that those participants carrying rs10995 G-allele (with better response

to HCTZ) had higher baseline expression levels of VASP compared to GA and AA

carriers. Moreover, we found that PEAR White participants with a good response to

HCTZ had a significantly higher VASP baseline expression levels compared to HCTZ

poor responders. This finding was further replicated in White participants treated with

Chlorothalidone in PEAR-2. All these pieces of evidence and multiple levels of

replication shed light on the importance of the VASP gene as a potential marker

associated with HCTZ BP response.

We also sought to integrate this well replicated signal with fourteen genes that

were differentially expressed between PEAR thiazide diuretics extreme responders to

identify pathways that could help us understand how the VASP gene might be involved

in the BP lowering mechanism underlying HCTZ BP effect. From this pathway

integrative approach, we were able to identify the actin nucleation pathway and the

integrin signaling pathway, as top significant pathways, in which the VASP gene

overlapped with two other genes (RhoB and CDC42EP2). These results highlight the

actin nucleation and the integrin signaling pathway as important pathways in which

125

HCTZ might be acting on. Additionally, the signals we identified from this approach

(VASP, RhoB and CDC42EP2) are known of their effects on the vascular smooth

muscle contraction, which suggests that thiazide diuretics BP lowering mechanism

might be mediated via their vasodilatory effect on the vascular smooth muscle, as

previously hypothesized (Chapter 1). Of note, we were able to provide multiple levels of

replication for the VASP, but not for RhoB or CDC42EP2. Our lack of replication for the

RhoB or the CDC42EP2 might be related to the small sample size used in the

expression analyses, therefore additional studies with larger sample size might be able

to replicate the RhoB and CDC42EP2 and confirm their association with HCTZ BP

response.

The research described in Chapter 3 focused on analyzing the metabolomics

profiles of HCTZ treated patients to identify metabolites that significantly influence the

BP response to HCTZ. In Chapter 3, we also sought to use a metabolomics-genomics

integrative approach to identify novel pathways and genetic variants with significant

impact on HCTZ BP response. Our analyses revealed thirteen novel metabolites that

were significantly associated with HCTZ BP response. Additionally, using the genomics-

metabolomics integrative approach, we identified the netrin signaling pathway as a

significant pathway that might be involved in the BP lowering mechanism underlying

HCTZ BP effect. Moreover, we were able to identify three signals (PRKAG2 rs2727563,

DCC rs12604940, and EPHX2 rs13262930) significantly associated with HCTZ BP

response, which were further replicated in an independent cohort. To examine the

relative contribution of these three replicated genetic signals toward our phenotype, we

constructed a genetic response score based on summing the BP lowering alleles of

126

these three signals. This response score explained 11.3% and 11.9% of HCTZ SBP and

DBP responses in PEAR monotherapy, respectively, and the association of this

response score with HCTZ BP response was further validated in an independent study.

Of note, the three potential candidate genes identified in Chapter 3 have been

either known to have a direct effect on the vascular smooth muscle or involved in

pathways regulating vascular smooth muscle contraction or relaxation. For instance,

PRKAG2, an AMP-activated protein kinase, has been known to attenuate smooth

muscle contraction by phosphorylating myosin light chain kinase (MLCK) at ser815 and

thus leading to MLCK inhibition[246]. Additionally, EPHX2 gene is well known for coding

the soluble epoxide hydrolase (sHE) enzyme, which converts epoxyeicosatrienoic acid

(EET), a strong vasodilator and anti-inflammatory compound, to the biologically less

active compound, dihydroxyeicosatrienoic acid (DHET) 58,59. EETs are known to

modulate vascular smooth muscle tone by activating large conductance, calcium

activated potassium channels, hence generating membrane hyperpolarization and

relaxation of the vascular smooth muscle[247]. In addition to PRKAG2, and EPHX2, the

netrin-1 receptor (DCC) has also been shown to be involved in regulating multiple

enzymes (PKC, src, Rac, and Rho Kinase) that have been known of their influential

effect on the vascular smooth muscle function[175-178] (Chapter 3). Altogether, having

signals identified in this Chapter and in Chapter 2 with multiple levels of replication and

high level of literature evidence strongly suggest that thiazide diuretics chronic BP

lowering effects might be mediated via their effect on vascular smooth muscle functions.

In Chapter 4, we sought to identify additional significant pathways associated

with HCTZ BP lowering effect. We started our approach by running a metabolomics

127

pathway analysis using the thirteen metabolites that were significantly associated with

HCTZ BP response in Chapter 3. Then, we leveraged our analyses with genomics and

lipidomics data to provide more insight in the BP lowering mechanism underlying HCTZ,

and to further validate our findings. The results of the metabolomics pathway analysis

shed light on the sphingomyelin metabolism pathway as the top significant pathway

associated with HCTZ BP response. Testing the association between SNPs within

thirteen genes in the sphingomyelin metabolism pathway and HCTZ BP response

uncovered a significant association between rs6078905 SNP, in the SPTLC3 gene, and

HCTZ BP response. We further validated the effect of the SPTLC3 rs6078905 SNP on

the sphingomyelin metabolism pathway by revealing a significant association between

this SNP and sphingomyelin levels, which we later showed their significant effects on

HCTZ BP response.

Interestingly, sphingomyelin and its biologically active metabolites (i.e.

sphingosine 1-phosphate; S1P) have been shown to exert an influential effect on the

vascular smooth muscle tone [216,219] (Chapter 4). Additionally, studies have shown

that S1P has a vasoconstrictive effect on the vascular smooth muscles in most tissues

which might be mediated via their effect on calcium mobilization from intracellular stores

or their activation to rho-kinase [219,240,241]. Collectively, the results from this Chapter

further support that thiazide diuretics long term BP lowering effects are mediated via

their vasodilatory effect on the vascular smooth muscles; however, this vasodilatory

effect seems to be complex and mediated via the effect of HCTZ on different pathways.

In summary, in this project, we used several innovative approaches to integrate

different “omics” of HCTZ treated participants with the aim of identifying pathways and

128

biomarkers associated with HCTZ BP response. The results of this project shed light on

novel pathways and markers associated with thiazide diuretics BP response, which

provided more insight in the mechanism underlying this class of drugs, and strongly

suggest that thiazide diuretics long term BP lowering mechanism might be mediated via

their effect on enzymes and pathways involved in the regulation of smooth muscle

function (Figure 5-1). One of the strengths of this project is that most of its results are

supported by multiple level of replication, which adds toward the validity and the

promise for future utility of these findings for guiding the selection of thiazide diuretics.

Future replication of these signals across multiple, appropriately-designed and well

powered studies might open new avenues for understanding the complex mechanism of

BP regulation and discovery of new therapeutic approaches to better optimize the BP

response in thiazide treated patients. Additionally, the results of this project highlighted

that the mechanism underlying thiazide diuretics BP lowering effect is complex;

therefore, we hypothesize that personalizing the use of HCTZ will need an algorithm

consisting of several genes to cover the complex signaling pathways that might be

involved in the BP lowering mechanism of this drug. The genes identified in this project

and the response score proposed should be considered in future models and algorithms

aiming to optimize the BP lowering effects of thiazide diuretics and improving the

therapeutic approaches for selecting this class of anti-hypertensives. Moreover, our

findings provided strong evidence that thiazide diuretics BP lowering effects might be

mediated via their regulatory effect on smooth muscle function. Future well designed in

vivo studies to test this hypothesis are recommended, which might identify additional

novel anti-hypertensive drug targets by fully understanding the mediators involved in

129

this mechanism. Future consideration for studying the link between the pathways

identified in this project (i.e. the actin nucleation, the netrin signaling, and the

sphingomyelin metabolism pathways) and the pathophysiology of HTN, or other anti-

hypertensive BP response, might provide insight in the mechanism underlying HTN and

BP regulation and eventually facilitate the development of new regimens and

therapeutic approaches for HTN control.

In conclusion, the results of this project highlight the strength of using different

“omics” for identifying novel pathways and biomarkers associated with drug response.

Using such approaches holds the promise to identify novel markers associated with the

variability in the efficacy or safety of pharmacotherapies and could improve the

discovery and development of new drugs by discovering novel determinants of the

studied phenotypes.

130

Figure 5-1. Illustrates the involvement of the thiazide diuretics associated signals

identified in this project in the smooth muscle regulation mechanism. Green boxes represent the signals that have been identified in this project to be significantly associated with thiazide diuretics blood pressure response. AA: arachidonic acid; AMPK: 5' adenosine monophosphate-activated protein kinase; Ca2+: calcium; CaM: Calmodulin; cAMP: cyclic adenosine monophosphate; DCC: Deleted in Colorectal Cancer; DHETE: dihydroxyeicosatetraenoic acids; EET: epoxyeicosatrienoic acids; EPHX2: Epoxide Hydrolase 2; IP3: inositol trisphosphate; K+: potassium; MLC: myosin light chain; MLCK: myosin light chain kinase; MLCP: myosin light chain phosphatase; P-MLC: phosphorylated myosin light chain; PLC: Phospholipase C; P-VASP: phosphorylated vasodilator-stimulated phosphoprotein; S1P: sphingosine-1-phosphate; VASP: vasodilator-stimulated phosphoprotein.

131

LIST OF REFERENCES

1. Kearney PM, Whelton M, Reynolds K, Muntner P, Whelton PK, He J: Global burden of hypertension: analysis of worldwide data. Lancet 2005, 365:217-223.

2. Mozaffarian D, Benjamin EJ, Go AS, Arnett DK, Blaha MJ, Cushman M, de Ferranti S, Despres JP, Fullerton HJ, Howard VJ, et al.: Heart disease and stroke statistics--2015 update: a report from the American Heart Association. Circulation 2015, 131:e29-322.

3. Law M, Wald N, Morris J: Lowering blood pressure to prevent myocardial infarction and stroke: a new preventive strategy. Health Technol Assess 2003, 7:1-94.

4. Lewington S, Clarke R, Qizilbash N, Peto R, Collins R, Collaboration PS: Age-specific relevance of usual blood pressure to vascular mortality: a meta-analysis of individual data for one million adults in 61 prospective studies. Lancet 2002, 360:1903-1913.

5. IMS Institue for Healthcare Informatics. Medicines Use and Spending Shifts: A Review of the Use of Medicines in the U.S. in 2014. Accessed June 1st, 2015.

6. Weber MA, Schiffrin EL, White WB, Mann S, Lindholm LH, Kenerson JG, Flack JM, Carter BL, Materson BJ, Ram CV, et al.: Clinical practice guidelines for the management of hypertension in the community a statement by the American Society of Hypertension and the International Society of Hypertension. J Hypertens 2014, 32:3-15.

7. Materson BJ: Variability in response to antihypertensive drugs. Am J Med 2007, 120:S10-20.

8. Materson BJ, Reda DJ, Cushman WC, Massie BM, Freis ED, Kochar MS, Hamburger RJ, Fye C, Lakshman R, Gottdiener J, et al.: Single-drug therapy for hypertension in men. A comparison of six antihypertensive agents with placebo. The Department of Veterans Affairs Cooperative Study Group on Antihypertensive Agents. N Engl J Med 1993, 328:914-921.

9. Thoenes M, Neuberger HR, Volpe M, Khan BV, Kirch W, Bohm M: Antihypertensive drug therapy and blood pressure control in men and women: an international perspective. J Hum Hypertens 2010, 24:336-344.

10. Blaufox MD, Lee HB, Davis B, Oberman A, Wassertheil-Smoller S, Langford H: Renin predicts diastolic blood pressure response to nonpharmacologic and pharmacologic therapy. JAMA 1992, 267:1221-1225.

132

11. Calhoun DA, Jones D, Textor S, Goff DC, Murphy TP, Toto RD, White A, Cushman WC, White W, Sica D, et al.: Resistant hypertension: diagnosis, evaluation, and treatment: a scientific statement from the American Heart Association Professional Education Committee of the Council for High Blood Pressure Research. Circulation 2008, 117:e510-526.

12. Johnson JA: Advancing management of hypertension through pharmacogenomics. Ann Med 2012, 44 Suppl 1:S17-22.

13. Turner ST, Schwartz GL, Chapman AB, Boerwinkle E: C825T polymorphism of the G protein beta(3)-subunit and antihypertensive response to a thiazide diuretic. Hypertension 2001, 37:739-743.

14. Schwartz GL, Turner ST, Chapman AB, Boerwinkle E: Interacting effects of gender and genotype on blood pressure response to hydrochlorothiazide. Kidney Int 2002, 62:1718-1723.

15. Turner ST, Chapman AB, Schwartz GL, Boerwinkle E: Effects of endothelial nitric oxide synthase, alpha-adducin, and other candidate gene polymorphisms on blood pressure response to hydrochlorothiazide. Am J Hypertens 2003, 16:834-839.

16. Cusi D, Barlassina C, Azzani T, Casari G, Citterio L, Devoto M, Glorioso N, Lanzani C, Manunta P, Righetti M, et al.: Polymorphisms of alpha-adducin and salt sensitivity in patients with essential hypertension. Lancet 1997, 349:1353-

1357.

17. Duarte JD, Lobmeyer MT, Wang Z, Chapman AB, Gums JG, Langaee TY, Boerwinkle E, Turner ST, Johnson JA: Lack of association between polymorphisms in STK39, a putative thiazide response gene, and blood pressure response to hydrochlorothiazide. Pharmacogenet Genomics 2010, 20:516-519.

18. Duarte JD, Zineh I, Burkley B, Gong Y, Langaee TY, Turner ST, Chapman AB, Boerwinkle E, Gums JG, Cooper-Dehoff RM, et al.: Effects of genetic variation in H3K79 methylation regulatory genes on clinical blood pressure and blood pressure response to hydrochlorothiazide. J Transl Med 2012, 10:56.

19. Frazier L, Turner ST, Schwartz GL, Chapman AB, Boerwinkle E: Multilocus effects of the renin-angiotensin-aldosterone system genes on blood pressure response to a thiazide diuretic. Pharmacogenomics J 2004, 4:17-23.

20. Glorioso N, Manunta P, Filigheddu F, Troffa C, Stella P, Barlassina C, Lombardi C, Soro A, Dettori F, Parpaglia PP, et al.: The role of alpha-adducin polymorphism in blood pressure and sodium handling regulation may not be excluded by a negative association study. Hypertension 1999, 34:649-654.

133

21. Glorioso N, Filigheddu F, Cusi D, Troffa C, Conti M, Natalizio M, Argiolas G, Barlassina C, Bianchi G: alpha-Adducin 460Trp allele is associated with erythrocyte Na transport rate in North Sardinian primary hypertensives. Hypertension 2002, 39:357-362.

22. Han YF, Fan XH, Wang XJ, Sun K, Xue H, Li WJ, Wang YB, Chen JZ, Zhen YS, Zhang WL, et al.: Association of intergenic polymorphism of organic anion transporter 1 and 3 genes with hypertension and blood pressure response to hydrochlorothiazide. Am J Hypertens 2011, 24:340-346.

23. Huang CC, Chung CM, Hung SI, Leu HB, Wu TC, Huang PH, Lin SJ, Pan WH, Chen JW: Genetic variation in renin predicts the effects of thiazide diuretics. Eur J Clin Invest 2011, 41:828-835.

24. Li Y, Zhou Y, Yang P, Niu JQ, Wu Y, Zhao DD, Wu SL: Interaction of ACE and CYP11B2 genes on blood pressure response to hydrochlorothiazide in Han Chinese hypertensive patients. Clin Exp Hypertens 2011, 33:141-146.

25. Lobmeyer MT, Wang L, Zineh I, Turner ST, Gums JG, Chapman AB, Cooper-DeHoff RM, Beitelshees AL, Bailey KR, Boerwinkle E, et al.: Polymorphisms in genes coding for GRK2 and GRK5 and response differences in antihypertensive-treated patients. Pharmacogenet Genomics 2011, 21:42-49.

26. Luo F, Wang Y, Wang X, Sun K, Zhou X, Hui R: A functional variant of NEDD4L is associated with hypertension, antihypertensive response, and orthostatic hypotension. Hypertension 2009, 54:796-801.

27. Manunta P, Lavery G, Lanzani C, Braund PS, Simonini M, Bodycote C, Zagato L, Delli Carpini S, Tantardini C, Brioni E, et al.: Physiological interaction between alpha-adducin and WNK1-NEDD4L pathways on sodium-related blood pressure regulation. Hypertension 2008, 52:366-372.

28. Matayoshi T, Kamide K, Takiuchi S, Yoshii M, Miwa Y, Takami Y, Tanaka C, Banno M, Horio T, Nakamura S, et al.: The thiazide-sensitive Na(+)-Cl(-) cotransporter gene, C1784T, and adrenergic receptor-beta3 gene, T727C, may be gene polymorphisms susceptible to the antihypertensive effect of thiazide diuretics. Hypertens Res 2004, 27:821-833.

29. Psaty BM, Smith NL, Heckbert SR, Vos HL, Lemaitre RN, Reiner AP, Siscovick DS, Bis J, Lumley T, Longstreth WT, et al.: Diuretic therapy, the alpha-adducin gene variant, and the risk of myocardial infarction or stroke in persons with treated hypertension. JAMA 2002, 287:1680-1689.

30. Schelleman H, Stricker BH, Verschuren WM, de Boer A, Kroon AA, de Leeuw PW, Kromhout D, Klungel OH: Interactions between five candidate genes and antihypertensive drug therapy on blood pressure. Pharmacogenomics J 2006, 6:22-26.

134

31. Schelleman H, Klungel OH, Witteman JC, Breteler MM, Hofman A, van Duijn CM, de Boer A, Stricker BH: Diuretic-gene interaction and the risk of myocardial infarction and stroke. Pharmacogenomics J 2007, 7:346-352.

32. Schelleman H, Klungel OH, Witteman JC, Hofman A, van Duijn CM, de Boer A, Stricker BH: The influence of the alpha-adducin G460W polymorphism and angiotensinogen M235T polymorphism on antihypertensive medication and blood pressure. Eur J Hum Genet 2006, 14:860-866.

33. Sciarrone MT, Stella P, Barlassina C, Manunta P, Lanzani C, Bianchi G, Cusi D: ACE and alpha-adducin polymorphism as markers of individual response to diuretic therapy. Hypertension 2003, 41:398-403.

34. Suonsyrjä T, Hannila-Handelberg T, Fodstad H, Donner K, Kontula K, Hiltunen TP: Renin-angiotensin system and alpha-adducin gene polymorphisms and their relation to responses to antihypertensive drugs: results from the GENRES study. Am J Hypertens 2009, 22:169-175.

35. Turner ST, Schwartz GL, Chapman AB, Boerwinkle E: WNK1 kinase polymorphism and blood pressure response to a thiazide diuretic. Hypertension 2005, 46:758-765.

36. van Wieren-de Wijer DB, Maitland-van der Zee AH, de Boer A, Kroon AA, de Leeuw PW, Schiffers P, Janssen RG, Psaty BM, van Duijn CM, Stricker BH, et al.: Interaction between the Gly460Trp alpha-adducin gene variant and diuretics on the risk of myocardial infarction. J Hypertens 2009, 27:61-68.

37. Turner ST, Boerwinkle E, O'Connell JR, Bailey KR, Gong Y, Chapman AB, McDonough CW, Beitelshees AL, Schwartz GL, Gums JG, et al.: Genomic association analysis of common variants influencing antihypertensive response to hydrochlorothiazide. Hypertension 2013, 62:391-397.

38. Turner ST, Bailey KR, Fridley BL, Chapman AB, Schwartz GL, Chai HS, Sicotte H, Kocher JP, Rodin AS, Boerwinkle E: Genomic association analysis suggests chromosome 12 locus influencing antihypertensive response to thiazide diuretic. Hypertension 2008, 52:359-365.

39. Chittani M, Zaninello R, Lanzani C, Frau F, Ortu MF, Salvi E, Fresu G, Citterio L, Braga D, Piras DA, et al.: TET2 and CSMD1 genes affect SBP response to hydrochlorothiazide in never-treated essential hypertensives. J Hypertens 2015, 33:1301-1309.

40. Hropot M, Fowler N, Karlmark B, Giebisch G: Tubular action of diuretics: distal effects on electrolyte transport and acidification. Kidney Int 1985, 28:477-489.

41. Bennett WM, McDonald WJ, Kuehnel E, Hartnett MN, Porter GA: Do diuretics have antihypertensive properties independent of natriuresis? Clin Pharmacol Ther 1977, 22:499-504.

135

42. Jones B, Nanra RS: Double-blind trial of antihypertensive effect of chlorothiazide in severe renal failure. Lancet 1979, 2:1258-1260.

43. van Brummelen P, Man in 't Veld AJ, Schalekamp MA: Hemodynamic changes during long-term thiazide treatment of essential hypertension in responders and nonresponders. Clin Pharmacol Ther 1980, 27:328-336.

44. Tarazi RC, Dustan HP, Frohlich ED: Long-term thiazide therapy in essential hypertension. Evidence for persistent alteration in plasma volume and renin activity. Circulation 1970, 41:709-717.

45. Wilson IM, Freis ED: Relationship between plasma and extracellular fluid volume depletion and the antihypertensive effect of chlorothiazide. Circulation 1959, 20:1028-1036.

46. Hughes AD: How do thiazide and thiazide-like diuretics lower blood pressure? J Renin Angiotensin Aldosterone Syst 2004, 5:155-160.

47. Shah S, Khatri I, Freis ED: Mechanism of antihypertensive effect of thiazide diuretics. Am Heart J 1978, 95:611-618.

48. Aleksandrow D, Wysznacka W, Gajewski J: Influence of chlorothiazide upon arterial responsiveness to nor-epinephrine in hypertensive subjects. N Engl J Med 1959, 261:1052-1055.

49. Freis ED, Wanko A, Schnaper HW, Frohlich ED: Mechanism of the altered blood pressure responsiveness produced by chlorothiazide. J Clin Invest 1960, 39:1277-1281.

50. Duarte JD, Cooper-DeHoff RM: Mechanisms for blood pressure lowering and metabolic effects of thiazide and thiazide-like diuretics. Expert Rev Cardiovasc Ther 2010, 8:793-802.

51. Colas B, Slama M, Collin T, Safar M, Andrejak M: Mechanisms of methyclothiazide-induced inhibition of contractile responses in rat aorta. Eur J Pharmacol 2000, 408:63-67.

52. Zhu Z, Zhu S, Liu D, Cao T, Wang L, Tepel M: Thiazide-like diuretics attenuate agonist-induced vasoconstriction by calcium desensitization linked to Rho kinase. Hypertension 2005, 45:233-239.

53. Calder JA, Schachter M, Sever PS: Potassium channel opening properties of thiazide diuretics in isolated guinea pig resistance arteries. J Cardiovasc Pharmacol 1994, 24:158-164.

54. Pickkers P, Hughes AD, Russel FG, Thien T, Smits P: Thiazide-induced vasodilation in humans is mediated by potassium channel activation. Hypertension 1998, 32:1071-1076.

136

55. Beermann B, Groschinsky-Grind M: Pharmacokinetics of hydrochlorothiazide in man. Eur J Clin Pharmacol 1977, 12:297-303.

56. Pickkers P, Hughes AD: Relaxation and decrease in [Ca2+]i by hydrochlorothiazide in guinea-pig isolated mesenteric arteries. Br J Pharmacol 1995, 114:703-707.

57. Pickkers P, Garcha RS, Schachter M, Smits P, Hughes AD: Inhibition of carbonic anhydrase accounts for the direct vascular effects of hydrochlorothiazide. Hypertension 1999, 33:1043-1048.

58. Ma F, Lin F, Chen C, Cheng J, Zeldin DC, Wang Y, Wang DW: Indapamide lowers blood pressure by increasing production of epoxyeicosatrienoic acids in the kidney. Mol Pharmacol 2013, 84:286-295.

59. Harvey KF, Dinudom A, Cook DI, Kumar S: The Nedd4-like protein KIAA0439 is a potential regulator of the epithelial sodium channel. J Biol Chem 2001, 276:8597-8601.

60. Shi PP, Cao XR, Sweezer EM, Kinney TS, Williams NR, Husted RF, Nair R, Weiss RM, Williamson RA, Sigmund CD, et al.: Salt-sensitive hypertension and cardiac hypertrophy in mice deficient in the ubiquitin ligase Nedd4-2. Am J Physiol Renal Physiol 2008, 295:F462-470.

61. Russo CJ, Melista E, Cui J, DeStefano AL, Bakris GL, Manolis AJ, Gavras H, Baldwin CT: Association of NEDD4L ubiquitin ligase with essential hypertension. Hypertension 2005, 46:488-491.

62. Dahlberg J, Nilsson LO, von Wowern F, Melander O: Polymorphism in NEDD4L is associated with increased salt sensitivity, reduced levels of P-renin and increased levels of Nt-proANP. PLoS One 2007, 2:e432.

63. Dahlberg J, Sjogren M, Hedblad B, Engstrom G, Melander O: Genetic variation in NEDD4L, an epithelial sodium channel regulator, is associated with cardiovascular disease and cardiovascular death. J Hypertens 2014, 32:294-299.

64. Dunn DM, Ishigami T, Pankow J, von Niederhausern A, Alder J, Hunt SC, Leppert MF, Lalouel JM, Weiss RB: Common variant of human NEDD4L activates a cryptic splice site to form a frameshifted transcript. J Hum Genet 2002, 47:665-676.

65. Svensson-Farbom P, Wahlstrand B, Almgren P, Dahlberg J, Fava C, Kjeldsen S, Hedner T, Melander O: A functional variant of the NEDD4L gene is associated with beneficial treatment response with beta-blockers and diuretics in hypertensive patients. J Hypertens 2011, 29:388-395.

137

66. McDonough CW, Burbage SE, Duarte JD, Gong Y, Langaee TY, Turner ST, Gums JG, Chapman AB, Bailey KR, Beitelshees AL, et al.: Association of variants in NEDD4L with blood pressure response and adverse cardiovascular outcomes in hypertensive patients treated with thiazide diuretics. J Hypertens 2013, 31:698-704.

67. Braz JC, Gregory K, Pathak A, Zhao W, Sahin B, Klevitsky R, Kimball TF, Lorenz JN, Nairn AC, Liggett SB, et al.: PKC-alpha regulates cardiac contractility and propensity toward heart failure. Nat Med 2004, 10:248-254.

68. Kanehisa M, Goto S, Sato Y, Furumichi M, Tanabe M: KEGG for integration and interpretation of large-scale molecular data sets. Nucleic Acids Res 2012, 40:D109-114.

69. Ehret GB, Munroe PB, Rice KM, Bochud M, Johnson AD, Chasman DI, Smith AV, Tobin MD, Verwoert GC, Hwang SJ, et al.: Genetic variants in novel pathways influence blood pressure and cardiovascular disease risk. Nature 2011, 478:103-109.

70. Duarte JD, Turner ST, Tran B, Chapman AB, Bailey KR, Gong Y, Gums JG, Langaee TY, Beitelshees AL, Cooper-Dehoff RM, et al.: Association of chromosome 12 locus with antihypertensive response to hydrochlorothiazide may involve differential YEATS4 expression. Pharmacogenomics J 2013, 13:257-263.

71. Kaddurah-Daouk R, Kristal BS, Weinshilboum RM: Metabolomics: a global biochemical approach to drug response and disease. Annu Rev Pharmacol Toxicol 2008, 48:653-683.

72. Suhre K, Meisinger C, Döring A, Altmaier E, Belcredi P, Gieger C, Chang D, Milburn MV, Gall WE, Weinberger KM, et al.: Metabolic footprint of diabetes: a multiplatform metabolomics study in an epidemiological setting. PLoS One 2010, 5:e13953.

73. Wang TJ, Larson MG, Vasan RS, Cheng S, Rhee EP, McCabe E, Lewis GD, Fox CS, Jacques PF, Fernandez C, et al.: Metabolite profiles and the risk of developing diabetes. Nat Med 2011, 17:448-453.

74. Wang-Sattler R, Yu Y, Mittelstrass K, Lattka E, Altmaier E, Gieger C, Ladwig KH, Dahmen N, Weinberger KM, Hao P, et al.: Metabolic profiling reveals distinct variations linked to nicotine consumption in humans--first results from the KORA study. PLoS One 2008, 3:e3863.

75. Bauer JW, Bilgic H, Baechler EC: Gene-expression profiling in rheumatic disease: tools and therapeutic potential. Nat Rev Rheumatol 2009, 5:257-265.

76. Marshall A, Lukk M, Kutter C, Davies S, Alexander G, Odom DT: Global gene expression profiling reveals SPINK1 as a potential hepatocellular carcinoma marker. PLoS One 2013, 8:e59459.

138

77. Chow A, Amemiya Y, Sugar L, Nam R, Seth A: Whole-transcriptome analysis reveals established and novel associations with TMPRSS2:ERG fusion in prostate cancer. Anticancer Res 2012, 32:3629-3641.

78. Trupp M, Zhu H, Wikoff WR, Baillie RA, Zeng ZB, Karp PD, Fiehn O, Krauss RM, Kaddurah-Daouk R: Metabolomics reveals amino acids contribute to variation in response to simvastatin treatment. PLoS One 2012, 7:e38386.

79. Ji Y, Hebbring S, Zhu H, Jenkins GD, Biernacka J, Snyder K, Drews M, Fiehn O, Zeng Z, Schaid D, et al.: Glycine and a glycine dehydrogenase (GLDC) SNP as citalopram/escitalopram response biomarkers in depression: pharmacometabolomics-informed pharmacogenomics. Clin Pharmacol Ther 2011, 89:97-104.

80. Inouye M, Kettunen J, Soininen P, Silander K, Ripatti S, Kumpula LS, Hamalainen E, Jousilahti P, Kangas AJ, Mannisto S, et al.: Metabonomic, transcriptomic, and genomic variation of a population cohort. Mol Syst Biol 2010, 6:441.

81. Suhre K, Gieger C: Genetic variation in metabolic phenotypes: study designs and applications. Nat Rev Genet 2012, 13:759-769.

82. Suhre K, Shin SY, Petersen AK, Mohney RP, Meredith D, Wagele B, Altmaier E, Deloukas P, Erdmann J, Grundberg E, et al.: Human metabolic individuality in biomedical and pharmaceutical research. Nature 2011, 477:54-60.

83. Gieger C, Geistlinger L, Altmaier E, Hrabe de Angelis M, Kronenberg F, Meitinger T, Mewes HW, Wichmann HE, Weinberger KM, Adamski J, et al.: Genetics meets metabolomics: a genome-wide association study of metabolite profiles in human serum. PLoS Genet 2008, 4:e1000282.

84. Wikoff WR, Frye RF, Zhu H, Gong Y, Boyle S, Churchill E, Cooper-Dehoff RM, Beitelshees AL, Chapman AB, Fiehn O, et al.: Pharmacometabolomics reveals racial differences in response to atenolol treatment. PLoS One 2013, 8:e57639.

85. Homuth G, Teumer A, Volker U, Nauck M: A description of large-scale metabolomics studies: increasing value by combining metabolomics with genome-wide SNP genotyping and transcriptional profiling. J Endocrinol 2012, 215:17-28.

86. Song IS, Lee dY, Shin MH, Kim H, Ahn YG, Park I, Kim KH, Kind T, Shin JG, Fiehn O, et al.: Pharmacogenetics meets metabolomics: discovery of tryptophan as a new endogenous OCT2 substrate related to metformin disposition. PLoS One 2012, 7:e36637.

87. Xiong Q, Ancona N, Hauser ER, Mukherjee S, Furey TS: Integrating genetic and gene expression evidence into genome-wide association analysis of gene sets. Genome Res 2012, 22:386-397.

139

88. Bebek G, Koyutürk M, Price ND, Chance MR: Network biology methods integrating biological data for translational science. Brief Bioinform 2012, 13:446-459.

89. Hsu YH, Zillikens MC, Wilson SG, Farber CR, Demissie S, Soranzo N, Bianchi EN, Grundberg E, Liang L, Richards JB, et al.: An integration of genome-wide association study and gene expression profiling to prioritize the discovery of novel susceptibility Loci for osteoporosis-related traits. PLoS Genet 2010, 6:e1000977.

90. Curtis C, Shah SP, Chin SF, Turashvili G, Rueda OM, Dunning MJ, Speed D, Lynch AG, Samarajiwa S, Yuan Y, et al.: The genomic and transcriptomic architecture of 2,000 breast tumours reveals novel subgroups. Nature 2012, 486:346-352.

91. Kasarskis A, Yang X, Schadt E: Integrative genomics strategies to elucidate the complexity of drug response. Pharmacogenomics 2011, 12:1695-1715.

92. Nibbe RK, Koyutürk M, Chance MR: An integrative -omics approach to identify functional sub-networks in human colorectal cancer. PLoS Comput Biol 2010, 6:e1000639.

93. Li Q, Seo JH, Stranger B, McKenna A, Pe'er I, Laframboise T, Brown M, Tyekucheva S, Freedman ML: Integrative eQTL-based analyses reveal the biology of breast cancer risk loci. Cell 2013, 152:633-641.

94. Lozano R, Naghavi M, Foreman K, Lim S, Shibuya K, Aboyans V, Abraham J, Adair T, Aggarwal R, Ahn SY, et al.: Global and regional mortality from 235 causes of death for 20 age groups in 1990 and 2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet 2012, 380:2095-2128.

95. Chow CK, Teo KK, Rangarajan S, Islam S, Gupta R, Avezum A, Bahonar A, Chifamba J, Dagenais G, Diaz R, et al.: Prevalence, awareness, treatment, and control of hypertension in rural and urban communities in high-, middle-, and low-income countries. JAMA 2013, 310:959-968.

96. Lewington S, Clarke R, Qizilbash N, Peto R, Collins R, Prospective Studies C: Age-specific relevance of usual blood pressure to vascular mortality: a meta-analysis of individual data for one million adults in 61 prospective studies. Lancet 2002, 360:1903-1913.

97. Psaty BM, Furberg CD, Kuller LH, Borhani NO, Rautaharju PM, O'Leary DH, Bild DE, Robbins J, Fried LP, Reid C: Isolated systolic hypertension and subclinical cardiovascular disease in the elderly. Initial findings from the Cardiovascular Health Study. JAMA 1992, 268:1287-1291.

140

98. Chobanian AV, Bakris GL, Black HR, Cushman WC, Green LA, Izzo JL, Jr., Jones DW, Materson BJ, Oparil S, Wright JT, Jr., et al.: The Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure: the JNC 7 report. JAMA 2003, 289:2560-

2572.

99. James PA, Oparil S, Carter BL, Cushman WC, Dennison-Himmelfarb C, Handler J, Lackland DT, Lefevre ML, Mackenzie TD, Ogedegbe O, et al.: 2014 Evidence-Based Guideline for the Management of High Blood Pressure in Adults: Report From the Panel Members Appointed to the Eighth Joint National Committee (JNC 8). JAMA 2013.

100. Ma J, Stafford RS: Screening, treatment, and control of hypertension in US private physician offices, 2003-2004. Hypertension 2008, 51:1275-1281.

101. Lloyd-Jones D, Adams RJ, Brown TM, Carnethon M, Dai S, De Simone G, Ferguson TB, Ford E, Furie K, Gillespie C, et al.: Heart disease and stroke statistics--2010 update: a report from the American Heart Association. Circulation 2010, 121:e46-e215.

102. Ge D, Fellay J, Thompson AJ, Simon JS, Shianna KV, Urban TJ, Heinzen EL, Qiu P, Bertelsen AH, Muir AJ, et al.: Genetic variation in IL28B predicts hepatitis C treatment-induced viral clearance. Nature 2009, 461:399-401.

103. Suppiah V, Moldovan M, Ahlenstiel G, Berg T, Weltman M, Abate ML, Bassendine M, Spengler U, Dore GJ, Powell E, et al.: IL28B is associated with response to chronic hepatitis C interferon-alpha and ribavirin therapy. Nat Genet 2009, 41:1100-1104.

104. Tanaka Y, Nishida N, Sugiyama M, Kurosaki M, Matsuura K, Sakamoto N, Nakagawa M, Korenaga M, Hino K, Hige S, et al.: Genome-wide association of IL28B with response to pegylated interferon-alpha and ribavirin therapy for chronic hepatitis C. Nat Genet 2009, 41:1105-1109.

105. Shuldiner AR, O'Connell JR, Bliden KP, Gandhi A, Ryan K, Horenstein RB, Damcott CM, Pakyz R, Tantry US, Gibson Q, et al.: Association of cytochrome P450 2C19 genotype with the antiplatelet effect and clinical efficacy of clopidogrel therapy. JAMA 2009, 302:849-857.

106. Cooper GM, Johnson JA, Langaee TY, Feng H, Stanaway IB, Schwarz UI, Ritchie MD, Stein CM, Roden DM, Smith JD, et al.: A genome-wide scan for common genetic variants with a large influence on warfarin maintenance dose. Blood 2008, 112:1022-1027.

107. Takeuchi F, McGinnis R, Bourgeois S, Barnes C, Eriksson N, Soranzo N, Whittaker P, Ranganath V, Kumanduri V, McLaren W, et al.: A genome-wide association study confirms VKORC1, CYP2C9, and CYP4F2 as principal genetic determinants of warfarin dose. PLoS Genet 2009, 5:e1000433.

141

108. Teichert M, Eijgelsheim M, Rivadeneira F, Uitterlinden AG, van Schaik RH, Hofman A, De Smet PA, van Gelder T, Visser LE, Stricker BH: A genome-wide association study of acenocoumarol maintenance dosage. Hum Mol Genet 2009, 18:3758-3768.

109. Treviño LR, Shimasaki N, Yang W, Panetta JC, Cheng C, Pei D, Chan D, Sparreboom A, Giacomini KM, Pui CH, et al.: Germline genetic variation in an organic anion transporter polypeptide associated with methotrexate pharmacokinetics and clinical effects. J Clin Oncol 2009, 27:5972-5978.

110. Nicolae DL, Gamazon E, Zhang W, Duan S, Dolan ME, Cox NJ: Trait-associated SNPs are more likely to be eQTLs: annotation to enhance discovery from GWAS. PLoS Genet 2010, 6:e1000888.

111. Stranger BE, Stahl EA, Raj T: Progress and promise of genome-wide association studies for human complex trait genetics. Genetics 2011, 187:367-383.

112. Li L, Kabesch M, Bouzigon E, Demenais F, Farrall M, Moffatt MF, Lin X, Liang L: Using eQTL weights to improve power for genome-wide association studies: a genetic study of childhood asthma. Front Genet 2013, 4:103.

113. Maurano MT, Humbert R, Rynes E, Thurman RE, Haugen E, Wang H, Reynolds AP, Sandstrom R, Qu H, Brody J, et al.: Systematic localization of common disease-associated variation in regulatory DNA. Science 2012, 337:1190-

1195.

114. Dunham I, Kundaje A, Aldred SF, Collins PJ, Davis CA, Doyle F, Epstein CB, Frietze S, Harrow J, Kaul R, et al.: An integrated encyclopedia of DNA elements in the human genome. Nature 2012, 489:57-74.

115. Schaub MA, Boyle AP, Kundaje A, Batzoglou S, Snyder M: Linking disease associations with regulatory information in the human genome. Genome Res 2012, 22:1748-1759.

116. Hoffman MM, Ernst J, Wilder SP, Kundaje A, Harris RS, Libbrecht M, Giardine B, Ellenbogen PM, Bilmes JA, Birney E, et al.: Integrative annotation of chromatin elements from ENCODE data. Nucleic Acids Res 2013, 41:827-841.

117. Hardison RC: Genome-wide epigenetic data facilitate understanding of disease susceptibility association studies. J Biol Chem 2012, 287:30932-

30940.

118. Kim WJ, Lim JH, Lee JS, Lee SD, Kim JH, Oh YM: Comprehensive Analysis of Transcriptome Sequencing Data in the Lung Tissues of COPD Subjects. Int J Genomics 2015, 2015:206937.

142

119. Liu Y, Morley M, Brandimarto J, Hannenhalli S, Hu Y, Ashley EA, Tang WH, Moravec CS, Margulies KB, Cappola TP, et al.: RNA-Seq identifies novel myocardial gene expression signatures of heart failure. Genomics 2015, 105:83-89.

120. Pflueger D, Terry S, Sboner A, Habegger L, Esgueva R, Lin PC, Svensson MA, Kitabayashi N, Moss BJ, MacDonald TY, et al.: Discovery of non-ETS gene fusions in human prostate cancer using next-generation RNA sequencing. Genome Res 2011, 21:56-67.

121. Rathe SK, Moriarity BS, Stoltenberg CB, Kurata M, Aumann NK, Rahrmann EP, Bailey NJ, Melrose EG, Beckmann DA, Liska CR, et al.: Using RNA-seq and targeted nucleases to identify mechanisms of drug resistance in acute myeloid leukemia. Sci Rep 2014, 4:6048.

122. Khatoon Z, Figler B, Zhang H, Cheng F: Introduction to RNA-Seq and its applications to drug discovery and development. Drug Dev Res 2014, 75:324-330.

123. Lappalainen T, Sammeth M, Friedlander MR, t Hoen PA, Monlong J, Rivas MA, Gonzalez-Porta M, Kurbatova N, Griebel T, Ferreira PG, et al.: Transcriptome and genome sequencing uncovers functional variation in humans. Nature 2013, 501:506-511.

124. Conde L, Bracci PM, Richardson R, Montgomery SB, Skibola CF: Integrating GWAS and expression data for functional characterization of disease-associated SNPs: an application to follicular lymphoma. Am J Hum Genet 2013, 92:126-130.

125. Westra HJ, Peters MJ, Esko T, Yaghootkar H, Schurmann C, Kettunen J, Christiansen MW, Fairfax BP, Schramm K, Powell JE, et al.: Systematic identification of trans eQTLs as putative drivers of known disease associations. Nat Genet 2013, 45:1238-1243.

126. Johnson JA, Boerwinkle E, Zineh I, Chapman AB, Bailey K, Cooper-DeHoff RM, Gums J, Curry RW, Gong Y, Beitelshees AL, et al.: Pharmacogenomics of antihypertensive drugs: rationale and design of the Pharmacogenomic Evaluation of Antihypertensive Responses (PEAR) study. Am Heart J 2009, 157:442-449.

127. Hamadeh IS, Langaee TY, Dwivedi R, Garcia S, Burkley BM, Skaar TC, Chapman AB, Gums JG, Turner ST, Gong Y, et al.: Impact of CYP2D6 polymorphisms on clinical efficacy and tolerability of metoprolol tartrate. Clin Pharmacol Ther 2014, 96:175-181.

128. Turner ST, Schwartz GL, Chapman AB, Beitelshees AL, Gums JG, Cooper-Dehoff RM, Boerwinkle E, Johnson JA, Bailey KR: Power to identify a genetic predictor of antihypertensive drug response using different methods to measure blood pressure response. J Transl Med 2012, 10:47.

143

129. Ragot S, Genes N, Vaur L, Herpin D: Comparison of three blood pressure measurement methods for the evaluation of two antihypertensive drugs: feasibility, agreement, and reproducibility of blood pressure response. Am J Hypertens 2000, 13:632-639.

130. Stergiou GS, Baibas NM, Gantzarou AP, Skeva, II, Kalkana CB, Roussias LG, Mountokalakis TD: Reproducibility of home, ambulatory, and clinic blood pressure: implications for the design of trials for the assessment of antihypertensive drug efficacy. Am J Hypertens 2002, 15:101-104.

131. Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MA, Bender D, Maller J, Sklar P, de Bakker PI, Daly MJ, et al.: PLINK: a tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet 2007, 81:559-575.

132. Boyle AP, Hong EL, Hariharan M, Cheng Y, Schaub MA, Kasowski M, Karczewski KJ, Park J, Hitz BC, Weng S, et al.: Annotation of functional variation in personal genomes using RegulomeDB. Genome Res 2012, 22:1790-1797.

133. Ernst ME, Carter BL, Goerdt CJ, Steffensmeier JJ, Phillips BB, Zimmerman MB, Bergus GR: Comparative antihypertensive effects of hydrochlorothiazide and chlorthalidone on ambulatory and office blood pressure. Hypertension 2006, 47:352-358.

134. Bowlus WE, Langford HG: A Comparison of the Antihypertensive Effect of Chlorthalidone and Hydrochlorthiazide. Clin Pharmacol Ther 1964, 5:708-711.

135. Link E, Parish S, Armitage J, Bowman L, Heath S, Matsuda F, Gut I, Lathrop M, Collins R: SLCO1B1 variants and statin-induced myopathy--a genomewide study. N Engl J Med 2008, 359:789-799.

136. Krause M, Dent EW, Bear JE, Loureiro JJ, Gertler FB: Ena/VASP proteins: regulators of the actin cytoskeleton and cell migration. Annu Rev Cell Dev Biol 2003, 19:541-564.

137. Yamin R, Morgan KG: Deciphering actin cytoskeletal function in the contractile vascular smooth muscle cell. J Physiol 2012, 590:4145-4154.

138. Gunst SJ, Zhang W: Actin cytoskeletal dynamics in smooth muscle: a new paradigm for the regulation of smooth muscle contraction. Am J Physiol Cell Physiol 2008, 295:C576-587.

139. Walter U, Eigenthaler M, Geiger J, Reinhard M: Role of cyclic nucleotide-dependent protein kinases and their common substrate VASP in the regulation of human platelets. Adv Exp Med Biol 1993, 344:237-249.

140. Hofmann F, Feil R, Kleppisch T, Schlossmann J: Function of cGMP-dependent protein kinases as revealed by gene deletion. Physiol Rev 2006, 86:1-23.

144

141. Comerford KM, Lawrence DW, Synnestvedt K, Levi BP, Colgan SP: Role of vasodilator-stimulated phosphoprotein in PKA-induced changes in endothelial junctional permeability. FASEB J 2002, 16:583-585.

142. Chen H, Levine YC, Golan DE, Michel T, Lin AJ: Atrial natriuretic peptide-initiated cGMP pathways regulate vasodilator-stimulated phosphoprotein phosphorylation and angiogenesis in vascular endothelium. J Biol Chem 2008, 283:4439-4447.

143. An D, Rodrigues B: Role of changes in cardiac metabolism in development of diabetic cardiomyopathy. Am J Physiol Heart Circ Physiol 2006, 291:H1489-1506.

144. Aszodi A, Pfeifer A, Ahmad M, Glauner M, Zhou XH, Ny L, Andersson KE, Kehrel B, Offermanns S, Fassler R: The vasodilator-stimulated phosphoprotein (VASP) is involved in cGMP- and cAMP-mediated inhibition of agonist-induced platelet aggregation, but is dispensable for smooth muscle function. EMBO J 1999, 18:37-48.

145. Oelze M, Mollnau H, Hoffmann N, Warnholtz A, Bodenschatz M, Smolenski A, Walter U, Skatchkov M, Meinertz T, Munzel T: Vasodilator-stimulated phosphoprotein serine 239 phosphorylation as a sensitive monitor of defective nitric oxide/cGMP signaling and endothelial dysfunction. Circ Res 2000, 87:999-1005.

146. Feil R, Lohmann SM, de Jonge H, Walter U, Hofmann F: Cyclic GMP-dependent protein kinases and the cardiovascular system: insights from genetically modified mice. Circ Res 2003, 93:907-916.

147. Warner TD, Mitchell JA, Sheng H, Murad F: Effects of cyclic GMP on smooth muscle relaxation. Adv Pharmacol 1994, 26:171-194.

148. Buechler WA, Ivanova K, Wolfram G, Drummer C, Heim JM, Gerzer R: Soluble guanylyl cyclase and platelet function. Ann N Y Acad Sci 1994, 714:151-157.

149. Schork AJ, Thompson WK, Pham P, Torkamani A, Roddey JC, Sullivan PF, Kelsoe JR, O'Donovan MC, Furberg H, Schork NJ, et al.: All SNPs are not created equal: genome-wide association studies reveal a consistent pattern of enrichment among functionally annotated SNPs. PLoS Genet 2013, 9:e1003449.

150. Barrett JC, Fry B, Maller J, Daly MJ: Haploview: analysis and visualization of LD and haplotype maps. Bioinformatics 2005, 21:263-265.

151. Johnson JA: Pharmacogenomics of antihypertensive drugs: past, present and future. Pharmacogenomics 2010, 11:487-491.

145

152. Cooper-Dehoff RM, Hou W, Weng L, Baillie RA, Beitelshees AL, Gong Y, Shahin MH, Turner ST, Chapman A, Gums JG, et al.: Is Diabetes Mellitus-Linked Amino Acid Signature Associated With beta-Blocker-Induced Impaired Fasting Glucose? Circ Cardiovasc Genet 2014, 7:199-205.

153. Yerges-Armstrong LM, Ellero-Simatos S, Georgiades A, Zhu H, Lewis JP, Horenstein RB, Beitelshees AL, Dane A, Reijmers T, Hankemeier T, et al.: Purine pathway implicated in mechanism of resistance to aspirin therapy: pharmacometabolomics-informed pharmacogenomics. Clin Pharmacol Ther 2013, 94:525-532.

154. Chapman AB, Schwartz GL, Boerwinkle E, Turner ST: Predictors of antihypertensive response to a standard dose of hydrochlorothiazide for essential hypertension. Kidney Int 2002, 61:1047-1055.

155. Fiehn O, Wohlgemuth G, Scholz M, Kind T, Lee do Y, Lu Y, Moon S, Nikolau B: Quality control for plant metabolomics: reporting MSI-compliant studies. Plant J 2008, 53:691-704.

156. Korn JM, Kuruvilla FG, McCarroll SA, Wysoker A, Nemesh J, Cawley S, Hubbell E, Veitch J, Collins PJ, Darvishi K, et al.: Integrated genotype calling and association analysis of SNPs, common copy number polymorphisms and rare CNVs. Nat Genet 2008, 40:1253-1260.

157. Kushner FG, Hand M, Smith SC, Jr., King SB, 3rd, Anderson JL, Antman EM, Bailey SR, Bates ER, Blankenship JC, Casey DE, Jr., et al.: 2009 focused updates: ACC/AHA guidelines for the management of patients with ST-elevation myocardial infarction (updating the 2004 guideline and 2007 focused update) and ACC/AHA/SCAI guidelines on percutaneous coronary intervention (updating the 2005 guideline and 2007 focused update) a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines. J Am Coll Cardiol 2009, 54:2205-2241.

158. Rahman M, Wright JT, Jr., Douglas JG: The role of the cytochrome P450-dependent metabolites of arachidonic acid in blood pressure regulation and renal function: a review. Am J Hypertens 1997, 10:356-365.

159. Sarkis A, Roman RJ: Role of cytochrome P450 metabolites of arachidonic acid in hypertension. Curr Drug Metab 2004, 5:245-256.

160. Laffer CL, Elijovich F, Eckert GJ, Tu W, Pratt JH, Brown NJ: Genetic variation in CYP4A11 and blood pressure response to mineralocorticoid receptor antagonism or ENaC inhibition: an exploratory pilot study in African Americans. J Am Soc Hypertens 2014, 8:475-480.

146

161. Mayer B, Lieb W, Gotz A, Konig IR, Aherrahrou Z, Thiemig A, Holmer S, Hengstenberg C, Doering A, Loewel H, et al.: Association of the T8590C polymorphism of CYP4A11 with hypertension in the MONICA Augsburg echocardiographic substudy. Hypertension 2005, 46:766-771.

162. Ward NC, Tsai IJ, Barden A, van Bockxmeer FM, Puddey IB, Hodgson JM, Croft KD: A single nucleotide polymorphism in the CYP4F2 but not CYP4A11 gene is associated with increased 20-HETE excretion and blood pressure. Hypertension 2008, 51:1393-1398.

163. Ward NC, Croft KD, Puddey IB, Phillips M, van Bockxmeer F, Beilin LJ, Barden AE: The effect of a single nucleotide polymorphism of the CYP4F2 gene on blood pressure and 20-hydroxyeicosatetraenoic acid excretion after weight loss. J Hypertens 2014, 32:1495-1502; discussion 1502.

164. Wu SN, Zhang Y, Gardner CO, Chen Q, Li Y, Wang GL, Gao PJ, Zhu DL: Evidence for association of polymorphisms in CYP2J2 and susceptibility to essential hypertension. Ann Hum Genet 2007, 71:519-525.

165. King LM, Gainer JV, David GL, Dai D, Goldstein JA, Brown NJ, Zeldin DC: Single nucleotide polymorphisms in the CYP2J2 and CYP2C8 genes and the risk of hypertension. Pharmacogenet Genomics 2005, 15:7-13.

166. Donner KM, Hiltunen TP, Suonsyrja T, Hannila-Handelberg T, Tikkanen I, Antikainen M, Hirvonen A, Kontula K: CYP2C9 genotype modifies activity of the renin-angiotensin-aldosterone system in hypertensive men. J Hypertens 2009, 27:2001-2009.

167. Taddei S, Versari D, Cipriano A, Ghiadoni L, Galetta F, Franzoni F, Magagna A, Virdis A, Salvetti A: Identification of a cytochrome P450 2C9-derived endothelium-derived hyperpolarizing factor in essential hypertensive patients. J Am Coll Cardiol 2006, 48:508-515.

168. Imig JD, Zhao X, Capdevila JH, Morisseau C, Hammock BD: Soluble epoxide hydrolase inhibition lowers arterial blood pressure in angiotensin II hypertension. Hypertension 2002, 39:690-694.

169. Koeners MP, Wesseling S, Ulu A, Sepulveda RL, Morisseau C, Braam B, Hammock BD, Joles JA: Soluble epoxide hydrolase in the generation and maintenance of high blood pressure in spontaneously hypertensive rats. Am J Physiol Endocrinol Metab 2011, 300:E691-698.

170. Gonzalez-Nunez D, Claria J, Rivera F, Poch E: Increased levels of 12(S)-HETE in patients with essential hypertension. Hypertension 2001, 37:334-338.

171. Xie C, Wang DH: Inhibition of renin release by arachidonic acid metabolites, 12(s)-HPETE and 12-HETE: role of TRPV1 channels. Endocrinology 2011, 152:3811-3819.

147

172. Kaddurah-Daouk R, Weinshilboum R: Metabolomic Signatures for Drug Response Phenotypes-Pharmacometabolomics Enables Precision Medicine. Clin Pharmacol Ther 2015.

173. Zhang J, Cai H: Netrin-1 prevents ischemia/reperfusion-induced myocardial infarction via a DCC/ERK1/2/eNOS s1177/NO/DCC feed-forward mechanism. J Mol Cell Cardiol 2010, 48:1060-1070.

174. Brunet I, Gordon E, Han J, Cristofaro B, Broqueres-You D, Liu C, Bouvree K, Zhang J, del Toro R, Mathivet T, et al.: Netrin-1 controls sympathetic arterial innervation. J Clin Invest 2014, 124:3230-3240.

175. Lai Wing Sun K, Correia JP, Kennedy TE: Netrins: versatile extracellular cues with diverse functions. Development 2011, 138:2153-2169.

176. Nishiyama M, Hoshino A, Tsai L, Henley JR, Goshima Y, Tessier-Lavigne M, Poo MM, Hong K: Cyclic AMP/GMP-dependent modulation of Ca2+ channels sets the polarity of nerve growth-cone turning. Nature 2003, 423:990-995.

177. Liu G, Beggs H, Jurgensen C, Park HT, Tang H, Gorski J, Jones KR, Reichardt LF, Wu J, Rao Y: Netrin requires focal adhesion kinase and Src family kinases for axon outgrowth and attraction. Nat Neurosci 2004, 7:1222-1232.

178. Forcet C, Stein E, Pays L, Corset V, Llambi F, Tessier-Lavigne M, Mehlen P: Netrin-1-mediated axon outgrowth requires deleted in colorectal cancer-dependent MAPK activation. Nature 2002, 417:443-447.

179. Buchholz RA, Dundore RL, Cumiskey WR, Harris AL, Silver PJ: Protein kinase inhibitors and blood pressure control in spontaneously hypertensive rats. Hypertension 1991, 17:91-100.

180. Newton-Cheh C, Johnson T, Gateva V, Tobin MD, Bochud M, Coin L, Najjar SS, Zhao JH, Heath SC, Eyheramendy S, et al.: Genome-wide association study identifies eight loci associated with blood pressure. Nat Genet 2009, 41:666-

676.

181. Andre G, Sandoval JE, Retailleau K, Loufrani L, Toumaniantz G, Offermanns S, Rolli-Derkinderen M, Loirand G, Sauzeau V: Smooth muscle specific Rac1 deficiency induces hypertension by preventing p116RIP3-dependent RhoA inhibition. J Am Heart Assoc 2014, 3:e000852.

182. Logan EM, Aileru AA, Shaltout HA, Averill DB, Diz DI: The functional role of PI3K in maintenance of blood pressure and baroreflex suppression in (mRen2)27 and mRen2.Lewis rat. J Cardiovasc Pharmacol 2011, 58:367-373.

183. Lee SJ, Jung YH, Oh SY, Yong MS, Ryu JM, Han HJ: Netrin-1 induces MMP-12-dependent E-cadherin degradation via the distinct activation of PKCalpha and FAK/Fyn in promoting mesenchymal stem cell motility. Stem Cells Dev 2014, 23:1870-1882.

148

184. Kemp BE, Mitchelhill KI, Stapleton D, Michell BJ, Chen ZP, Witters LA: Dealing with energy demand: the AMP-activated protein kinase. Trends Biochem Sci 1999, 24:22-25.

185. Tragante V, Barnes MR, Ganesh SK, Lanktree MB, Guo W, Franceschini N, Smith EN, Johnson T, Holmes MV, Padmanabhan S, et al.: Gene-centric meta-analysis in 87,736 individuals of European ancestry identifies multiple blood-pressure-related loci. Am J Hum Genet 2014, 94:349-360.

186. Kottgen A, Albrecht E, Teumer A, Vitart V, Krumsiek J, Hundertmark C, Pistis G, Ruggiero D, O'Seaghdha CM, Haller T, et al.: Genome-wide association analyses identify 18 new loci associated with serum urate concentrations. Nat Genet 2013, 45:145-154.

187. Kottgen A, Pattaro C, Boger CA, Fuchsberger C, Olden M, Glazer NL, Parsa A, Gao X, Yang Q, Smith AV, et al.: New loci associated with kidney function and chronic kidney disease. Nat Genet 2010, 42:376-384.

188. Arad M, Maron BJ, Gorham JM, Johnson WH, Jr., Saul JP, Perez-Atayde AR, Spirito P, Wright GB, Kanter RJ, Seidman CE, et al.: Glycogen storage diseases presenting as hypertrophic cardiomyopathy. N Engl J Med 2005, 352:362-372.

189. Lasker JM, Chen WB, Wolf I, Bloswick BP, Wilson PD, Powell PK: Formation of 20-hydroxyeicosatetraenoic acid, a vasoactive and natriuretic eicosanoid, in human kidney. Role of Cyp4F2 and Cyp4A11. J Biol Chem 2000, 275:4118-4126.

190. Fleming I: Vascular cytochrome p450 enzymes: physiology and pathophysiology. Trends Cardiovasc Med 2008, 18:20-25.

191. Spector AA, Fang X, Snyder GD, Weintraub NL: Epoxyeicosatrienoic acids (EETs): metabolism and biochemical function. Prog Lipid Res 2004, 43:55-

90.

192. Imig JD, Hammock BD: Soluble epoxide hydrolase as a therapeutic target for cardiovascular diseases. Nat Rev Drug Discov 2009, 8:794-805.

193. Godlewski G, Alapafuja SO, Batkai S, Nikas SP, Cinar R, Offertaler L, Osei-Hyiaman D, Liu J, Mukhopadhyay B, Harvey-White J, et al.: Inhibitor of fatty acid amide hydrolase normalizes cardiovascular function in hypertension without adverse metabolic effects. Chem Biol 2010, 17:1256-1266.

194. Sarzani R, Bordicchia M, Salvi F, Cola G, Franchi E, Battistoni I, Mancinelli L, Giovagnoli A, Dessi-Fulgheri P, Rappelli A: A human fatty acid amide hydrolase (FAAH) functional gene variant is associated with lower blood pressure in young males. Am J Hypertens 2008, 21:960-963.

149

195. Lee CR, Imig JD, Edin ML, Foley J, DeGraff LM, Bradbury JA, Graves JP, Lih FB, Clark J, Myers P, et al.: Endothelial expression of human cytochrome P450 epoxygenases lowers blood pressure and attenuates hypertension-induced renal injury in mice. FASEB J 2010, 24:3770-3781.

196. Edin ML, Wang Z, Bradbury JA, Graves JP, Lih FB, DeGraff LM, Foley JF, Torphy R, Ronnekleiv OK, Tomer KB, et al.: Endothelial expression of human cytochrome P450 epoxygenase CYP2C8 increases susceptibility to ischemia-reperfusion injury in isolated mouse heart. FASEB J 2011, 25:3436-3447.

197. Yamada Y, Matsuo H, Segawa T, Watanabe S, Kato K, Hibino T, Yokoi K, Ichihara S, Metoki N, Yoshida H, et al.: Assessment of the genetic component of hypertension. Am J Hypertens 2006, 19:1158-1165.

198. Iwai N, Katsuya T, Ishikawa K, Mannami T, Ogata J, Higaki J, Ogihara T, Tanabe T, Baba S: Human prostacyclin synthase gene and hypertension : the Suita Study. Circulation 1999, 100:2231-2236.

199. Facemire CS, Griffiths R, Audoly LP, Koller BH, Coffman TM: The impact of microsomal prostaglandin e synthase 1 on blood pressure is determined by genetic background. Hypertension 2010, 55:531-538.

200. Jia Z, Guo X, Zhang H, Wang MH, Dong Z, Yang T: Microsomal prostaglandin synthase-1-derived prostaglandin E2 protects against angiotensin II-induced hypertension via inhibition of oxidative stress. Hypertension 2008, 52:952-959.

201. Valencia DM, Naranjo CA, Parra MV, Caro MA, Valencia AV, Jaramillo CJ, Bedoya G: [Association and interaction of AGT, AGTR1, ACE, ADRB2, DRD1, ADD1, ADD2, ATP2B1, TBXA2R and PTGS2 genes on the risk of hypertension in Antioquian population]. Biomedica 2013, 33:598-614.

202. Huan T, Esko T, Peters MJ, Pilling LC, Schramm K, Schurmann C, Chen BH, Liu C, Joehanes R, Johnson AD, et al.: A meta-analysis of gene expression signatures of blood pressure and hypertension. PLoS Genet 2015, 11:e1005035.

203. Global, regional, and national age-sex specific all-cause and cause-specific mortality for 240 causes of death, 1990-2013: a systematic analysis for the Global Burden of Disease Study 2013. Lancet 2015, 385:117-171.

204. Law MR, Morris JK, Wald NJ: Use of blood pressure lowering drugs in the prevention of cardiovascular disease: meta-analysis of 147 randomised trials in the context of expectations from prospective epidemiological studies. BMJ 2009, 338:b1665.

150

205. Kaddurah-Daouk R, Baillie RA, Zhu H, Zeng ZB, Wiest MM, Nguyen UT, Watkins SM, Krauss RM: Lipidomic analysis of variation in response to simvastatin in the Cholesterol and Pharmacogenetics Study. Metabolomics 2010, 6:191-201.

206. Wikoff WR, Frye RF, Zhu H, Gong Y, Boyle S, Churchill E, Cooper-Dehoff RM, Beitelshees AL, Chapman AB, Fiehn O, et al.: Pharmacometabolomics reveals racial differences in response to atenolol treatment. PLoS One 2013, 8:e57639.

207. Song IS, Lee do Y, Shin MH, Kim H, Ahn YG, Park I, Kim KH, Kind T, Shin JG, Fiehn O, et al.: Pharmacogenetics meets metabolomics: discovery of tryptophan as a new endogenous OCT2 substrate related to metformin disposition. PLoS One 2012, 7:e36637.

208. Han X, Gross RW: Shotgun lipidomics: electrospray ionization mass spectrometric analysis and quantitation of cellular lipidomes directly from crude extracts of biological samples. Mass Spectrom Rev 2005, 24:367-412.

209. Yang K, Cheng H, Gross RW, Han X: Automated lipid identification and quantification by multidimensional mass spectrometry-based shotgun lipidomics. Anal Chem 2009, 81:4356-4368.

210. Jiang X, Cheng H, Yang K, Gross RW, Han X: Alkaline methanolysis of lipid extracts extends shotgun lipidomics analyses to the low-abundance regime of cellular sphingolipids. Anal Biochem 2007, 371:135-145.

211. Wang M, Han X: Multidimensional mass spectrometry-based shotgun lipidomics. Methods Mol Biol 2014, 1198:203-220.

212. Han X, Yang J, Cheng H, Ye H, Gross RW: Toward fingerprinting cellular lipidomes directly from biological samples by two-dimensional electrospray ionization mass spectrometry. Anal Biochem 2004, 330:317-

331.

213. Han X: Characterization and direct quantitation of ceramide molecular species from lipid extracts of biological samples by electrospray ionization tandem mass spectrometry. Anal Biochem 2002, 302:199-212.

214. Merrill AH, Jr., Wang E, Innis WS, Mullins R: Increases in serum sphingomyelin by 17 beta-estradiol. Lipids 1985, 20:252-254.

215. Nikkila J, Sysi-Aho M, Ermolov A, Seppanen-Laakso T, Simell O, Kaski S, Oresic M: Gender-dependent progression of systemic metabolic states in early childhood. Mol Syst Biol 2008, 4:197.

216. Hemmings DG: Signal transduction underlying the vascular effects of sphingosine 1-phosphate and sphingosylphosphorylcholine. Naunyn Schmiedebergs Arch Pharmacol 2006, 373:18-29.

151

217. Igarashi J, Thatte HS, Prabhakar P, Golan DE, Michel T: Calcium-independent activation of endothelial nitric oxide synthase by ceramide. Proc Natl Acad Sci U S A 1999, 96:12583-12588.

218. Li H, Junk P, Huwiler A, Burkhardt C, Wallerath T, Pfeilschifter J, Forstermann U: Dual effect of ceramide on human endothelial cells: induction of oxidative stress and transcriptional upregulation of endothelial nitric oxide synthase. Circulation 2002, 106:2250-2256.

219. Ohmori T, Yatomi Y, Osada M, Kazama F, Takafuta T, Ikeda H, Ozaki Y: Sphingosine 1-phosphate induces contraction of coronary artery smooth muscle cells via S1P2. Cardiovasc Res 2003, 58:170-177.

220. Fenger M, Linneberg A, Jorgensen T, Madsbad S, Sobye K, Eugen-Olsen J, Jeppesen J: Genetics of the ceramide/sphingosine-1-phosphate rheostat in blood pressure regulation and hypertension. BMC Genet 2011, 12:44.

221. Spijkers LJ, van den Akker RF, Janssen BJ, Debets JJ, De Mey JG, Stroes ES, van den Born BJ, Wijesinghe DS, Chalfant CE, MacAleese L, et al.: Hypertension is associated with marked alterations in sphingolipid biology: a potential role for ceramide. PLoS One 2011, 6:e21817.

222. Zheng H, Xie X, Xie N, Xu H, Huang J, Luo M: Sphingomyelin levels in nondipper and dipper hypertensive patients. Exp Ther Med 2014, 7:599-603.

223. Fenger M, Linneberg A, Jeppesen J: Network-based analysis of the sphingolipid metabolism in hypertension. Front Genet 2015, 6:84.

224. Rutherford PA, Thomas TH, Laker MF, Wilkinson R: Plasma lipids affect maximum velocity not sodium affinity of human sodium-lithium countertransport: distinction from essential hypertension. Eur J Clin Invest 1992, 22:719-724.

225. Chi Y, Mota de Freitas D, Sikora M, Bansal VK: Correlations of Na+-Li+ exchange activity with Na+ and Li+ binding and phospholipid composition in erythrocyte membranes of white hypertensive and normotensive individuals: a nuclear magnetic resonance investigation. Hypertension 1996, 27:456-464.

226. Zicha J, Kunes J, Devynck MA: Abnormalities of membrane function and lipid metabolism in hypertension: a review. Am J Hypertens 1999, 12:315-331.

227. Bischoff A, Czyborra P, Meyer Zu Heringdorf D, Jakobs KH, Michel MC: Sphingosine-1-phosphate reduces rat renal and mesenteric blood flow in vivo in a pertussis toxin-sensitive manner. Br J Pharmacol 2000, 130:1878-1883.

152

228. Bischoff A, Finger J, Michel MC: Nifedipine inhibits sphinogosine-1-phosphate-induced renovascular contraction in vitro and in vivo. Naunyn Schmiedebergs Arch Pharmacol 2001, 364:179-182.

229. Hedemann J, Fetscher C, Michel MC: Comparison of noradrenaline and lysosphingolipid-induced vasoconstriction in mouse and rat small mesenteric arteries. Auton Autacoid Pharmacol 2004, 24:77-85.

230. Salomone S, Yoshimura S, Reuter U, Foley M, Thomas SS, Moskowitz MA, Waeber C: S1P3 receptors mediate the potent constriction of cerebral arteries by sphingosine-1-phosphate. Eur J Pharmacol 2003, 469:125-134.

231. Coussin F, Scott RH, Nixon GF: Sphingosine 1-phosphate induces CREB activation in rat cerebral artery via a protein kinase C-mediated inhibition of voltage-gated K+ channels. Biochem Pharmacol 2003, 66:1861-1870.

232. Guan Z, Singletary ST, Cook AK, Hobbs JL, Pollock JS, Inscho EW: Sphingosine-1-phosphate evokes unique segment-specific vasoconstriction of the renal microvasculature. J Am Soc Nephrol 2014, 25:1774-1785.

233. Spiegel S, Milstien S: Sphingosine-1-phosphate: an enigmatic signalling lipid. Nat Rev Mol Cell Biol 2003, 4:397-407.

234. Zhu Q, Xia M, Wang Z, Li PL, Li N: A novel lipid natriuretic factor in the renal medulla: sphingosine-1-phosphate. Am J Physiol Renal Physiol 2011, 301:F35-41.

235. Himmel HM, Meyer Zu Heringdorf D, Graf E, Dobrev D, Kortner A, Schuler S, Jakobs KH, Ravens U: Evidence for Edg-3 receptor-mediated activation of I(K.ACh) by sphingosine-1-phosphate in human atrial cardiomyocytes. Mol Pharmacol 2000, 58:449-454.

236. Xu SZ, Muraki K, Zeng F, Li J, Sukumar P, Shah S, Dedman AM, Flemming PK, McHugh D, Naylor J, et al.: A sphingosine-1-phosphate-activated calcium channel controlling vascular smooth muscle cell motility. Circ Res 2006, 98:1381-1389.

237. Kim MY, Liang GH, Kim JA, Kim YJ, Oh S, Suh SH: Sphingosine-1-phosphate activates BKCa channels independently of G protein-coupled receptor in human endothelial cells. Am J Physiol Cell Physiol 2006, 290:C1000-1008.

238. Birchwood CJ, Saba JD, Dickson RC, Cunningham KW: Calcium influx and signaling in yeast stimulated by intracellular sphingosine 1-phosphate accumulation. J Biol Chem 2001, 276:11712-11718.

239. Mattie M, Brooker G, Spiegel S: Sphingosine-1-phosphate, a putative second messenger, mobilizes calcium from internal stores via an inositol trisphosphate-independent pathway. J Biol Chem 1994, 269:3181-3188.

153

240. Ishizawa S, Takahashi-Fujigasaki J, Kanazawa Y, Matoba K, Kawanami D, Yokota T, Iwamoto T, Tajima N, Manome Y, Utsunomiya K: Sphingosine-1-phosphate induces differentiation of cultured renal tubular epithelial cells under Rho kinase activation via the S1P2 receptor. Clin Exp Nephrol 2014, 18:844-852.

241. Garcia JG, Liu F, Verin AD, Birukova A, Dechert MA, Gerthoffer WT, Bamberg JR, English D: Sphingosine 1-phosphate promotes endothelial cell barrier integrity by Edg-dependent cytoskeletal rearrangement. J Clin Invest 2001, 108:689-701.

242. Wirth A: Rho kinase and hypertension. Biochim Biophys Acta 2010, 1802:1276-1284.

243. Do e Z, Fukumoto Y, Sugimura K, Miura Y, Tatebe S, Yamamoto S, Aoki T, Nochioka K, Nergui S, Yaoita N, et al.: Rho-kinase activation in patients with heart failure. Circ J 2013, 77:2542-2550.

244. Kataoka C, Egashira K, Inoue S, Takemoto M, Ni W, Koyanagi M, Kitamoto S, Usui M, Kaibuchi K, Shimokawa H, et al.: Important role of Rho-kinase in the pathogenesis of cardiovascular inflammation and remodeling induced by long-term blockade of nitric oxide synthesis in rats. Hypertension 2002, 39:245-250.

245. Satoh K, Fukumoto Y, Shimokawa H: Rho-kinase: important new therapeutic target in cardiovascular diseases. Am J Physiol Heart Circ Physiol 2011, 301:H287-296.

246. Horman S, Morel N, Vertommen D, Hussain N, Neumann D, Beauloye C, El Najjar N, Forcet C, Viollet B, Walsh MP, et al.: AMP-activated protein kinase phosphorylates and desensitizes smooth muscle myosin light chain kinase. J Biol Chem 2008, 283:18505-18512.

247. Pfister SL, Gauthier KM, Campbell WB: Vascular pharmacology of epoxyeicosatrienoic acids. Adv Pharmacol 2010, 60:27-59.

154

BIOGRAPHICAL SKETCH

Mohamed Hossam Shahin was born in Egypt. He received his bachelor’s degree

in pharmaceutical sciences in May 2007, from Misr International University in Cairo.

After graduation, he worked at Misr International University for four years as a teaching

assistant in the Department of Pharmacy Practice and Clinical Pharmacy and

completed his master’s degree. Soon after, he joined the clinical pharmaceutical PhD.

program in the Department of Pharmacotherapy and Translational Research at College

of Pharmacy, University of Florida. During his PhD, Mohamed has authored multiple

peer-reviewed manuscripts, presented his research at multiple national meetings and

received several research awards. Mohamed received his PhD. degree from the

University of Florida in the fall of 2015, and started a postdoctoral fellowship with Dr.

Julie Johnson at College of Pharmacy, University of Florida.