Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Identification of gene-exposure interactions for risk of cardiovascular disease
(USC Thesis Other)
Identification of gene-exposure interactions for risk of cardiovascular disease
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
IDENTIFICATION OF GENE-EXPOSURE INTERACTIONS
FOR RISK OF CARDIOVASCULAR DISEASE
by
James Raymond Hilser
A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(MOLECULAR MEDICINE)
December 2024
Copyright 2024 James Raymond Hilser
ii
DEDICATION
This dissertation is dedicated to my beautiful wife, Shannon. Thank you for always being you.
iii
ACKNOWLEDGMENTS
I would like to express my deepest gratitude to all those who supported and guided me
throughout this dissertation journey. First and foremost, I am incredibly thankful to my
Dissertation Committee: Pragna Patel, who sat as the chair, Hooman Allayee, and Jaana Hartiala,
for their invaluable mentorship, critical insights, and constant encouragement.
I would especially like to thank my PI, Hooman Allayee, for spending countless days of his life
helping me grow into the scientist I am today. Your guidance, humor, and unwavering support
have been an inspiration, and I truly would not have reached this point without you.
I extend my appreciation to my colleagues and collaborators at USC, including Yi Han, Subarna
Biswas, Neal Spencer, Yuxin Yang, Zhiheng Cai, Janet Gukasyan, and all of the other
individuals who weren’t directly in our lab, whose expertise and contributions greatly enriched
this work. I am equally grateful to the exceptional team at UCLA, particularly Dr. Jake Lusis, for
their guidance and support. Additionally, I would like to acknowledge my collaborators at the
Cleveland Clinic, including Dr. Stanley Hazen for their valuable partnership in advancing this
research.
This work would not have been possible without the resources provided by various biobanks,
including FinnGen, All of Us, GeneBank, and the UK Biobank.
Lastly, I want to thank my family and friends for their unwavering support, and to everyone who
contributed to making this journey a memorable and fulfilling experience.
iv
TABLE OF CONTENTS
DEDICATION………………………………………………………………............... ii
ACKNOWLEDGEMENTS……………………………………………………............ iii
LIST OF TABLES……………………………………………………….……............ viii
LIST OF FIGURES……………………………………………………………............ xi
ABSTRACT………………………………………………………………….............. xiii
INTRODUCTION
Overview of Cardiovascular Disease (CVD)…….……….......……………...... 1
Role of the Environment in Influencing Risk of CVD…….………………….... 2
Current State of CVD Genetics…….………………………………................ 4
Overview of gene-exposure (GxE) Interactions…….………………..........…... 6
Gene-Air Pollution Interactions…….……………………….............………... 7
Gene-Sex Interactions…….…………………………………...............……... 8
Gene-Drug Interactions…….…………………………………...............……. 9
Gene-Pathogen Interactions……………………………………….................. 9
Recent Advances for Investigating GxE Interactions…………….........………. 10
Summary………………………………………………………….................. 11
References……………………………………………………………............ 13
CHAPTER 1: GENE-ENVIRONMENT INTERACTION BETWEEN THE
TRIP4 LOCUS AND AIR POLLUTION EXPOSURE INCREASES
RISK OF CORONARY ARTERY DISEASE
Summary……………………………………………………………................ 20
Title Page………………………………………................………………....... 21
Abstract…………………………………………………….................………. 23
Introduction…………………………………………………................…....... 25
Materials and Methods…………………………………………..............…..... 27
Results………………………………………………………………................ 36
Discussion………………………………………………………….................. 40
References………………………………………………………….................. 45
Acknowledgments………………………………………………...............…... 53
Figure Legends………………………………………………................……... 55
Tables…………………………………………………………….................... 58
Figures………………………………………………………………............... 60
Supplemental Materials……………………………………………….............. 65
CHAPTER 2: COVID-19 IS A CORONARY ARTERY DISEASE RISK
EQUIVALENT AND EXHIBITS A GENETIC INTERACTION WITH
ABO BLOOD TYPE
Summary……………………………………………………………................ 75
Title Page………………………………………................………………....... 76
v
Abstract…………………………………………………….................………. 77
Introduction…………………………………………………................…....... 79
Materials and Methods…………………………………………..............…..... 81
Results………………………………………………………………................ 87
Discussion………………………………………………………….................. 92
References………………………………………………………….................. 97
Acknowledgments………………………………………………...............…... 104
Figure Legends………………………………………………................……... 106
Tables…………………………………………………………….................... 108
Figures………………………………………………………………............... 114
Supplemental Materials……………………………………………….............. 119
CHAPTER 3: GENETICALLY DECREASED CPS1 ACTIVITY
ATTENUATES ATHEROSCLEROSIS IN HUMANS AND MICE
THROUGH SEXUALLY DIMORPHIC PATTERNS
Summary……………………………………………………………................ 136
Title Page………………………………………................………………....... 137
Abstract…………………………………………………….................………. 138
Introduction…………………………………………………................…....... 139
Materials and Methods…………………………………………..............…..... 141
Results………………………………………………………………................ 152
Discussion………………………………………………………….................. 156
References………………………………………………………….................. 160
Acknowledgments………………………………………………...............…... 165
Figure Legends………………………………………………................……... 167
Tables…………………………………………………………….................... 170
Figures………………………………………………………………............... 176
CHAPTER 4: EFFECT OF MENOPAUSAL HORMONE THERAPY
ON METHYLATION LEVELS IN EARLY AND LATE
POSTMENOPAUSAL WOMEN
Summary……………………………………………………………................ 181
Title Page………………………………………................………………....... 182
Abstract…………………………………………………….................………. 183
Introduction…………………………………………………................…....... 185
Materials and Methods…………………………………………..............…..... 187
Results………………………………………………………………................ 192
Discussion………………………………………………………….................. 195
References………………………………………………………….................. 199
Acknowledgments………………………………………………...............…... 204
Figure Legends………………………………………………................……... 206
Tables…………………………………………………………….................... 208
Figures………………………………………………………………............... 213
vi
DISCUSSION
Elucidating Novel Coronary Artery Disease (CAD) Mechanisms
through Gene-Environment Interactions. ……………….................................. 217
Known CAD Variant Contributes to Gene-Exposure Interactions..... ................. 218
Genetic Signal for CAD at 15q22 Loci Linked to TRIP4................................... 219
Air Pollutants regulate systemic TRIP4 expression........................................... 219
GxE Interaction Implicates to TRIP4.............................................................. 221
Severe SARS-CoV-2 Infection Identified as CAD
Risk Equivalent and Potential Genetic Interactor..................................................... 222
Candidate Gene Approach Targeting ABO Locus.................................................... 223
Putative Mechanism Linking ABO and COVID-19................................................. 225
Specificity of ABO in Mediating Thrombotic Risk.................................................. 225
CPS1 Exhibits Interaction with Sex on Risk of CAD........................................ 226
Sexually Dimorphic Effects of CPS1 Variant in Humans Show
Lack of Association in Underrepresented Ancestries........................................ 227
Sexually Dimorphic Hepatic CPS1 Expression Explains
Threshold Association With Risk of CAD............................................................... 228
CpG Sites Associated with Hormone Replacement Therapy are not
Associated with Carotid Intermedial Thickness (cIMT).................................... 229
Blood Cells Serve as Poor Proxies for Vascular
Epigenetic Modifications............................................................................. 230
Methodological and Contextual Limitations................................................... 231
Future Directions...................................................................................................... 233
Conclusions and Implications in Cardiovascular Disease........................................ 234
References................................................................................................. 236
REFERENCES.................................................................................................. 243
APPENDICES
APPENDIX A............................................................................................. 274
Choucair I, Mallela DP, Hilser JR, Hartiala JA, Nemet I, Gogonea V, Li L,
Lusis AJ, Fischbach MA, Tang WHW, Allayee H, Hazen SL. Comprehensive
Clinical and Genetic Analyses of Circulating Bile Acids and Their
Associations With Diabetes and Its Indices. Diabetes. 2024 Aug 1;73(8):1215-
1228. doi: 10.2337/db23-0676. PMID: 38701355; PMCID: PMC11262044.
APPENDIX B............................................................................................. 356
Ferrell M, Wang Z, Anderson JT, Li XS, Witkowski M, DiDonato JA,
Hilser JR, Hartiala JA, Haghikia A, Cajka T, Fiehn O, Sangwan N, Demuth I,
König M, Steinhagen-Thiessen E, Landmesser U, Tang WHW, Allayee H,
Hazen SL. A terminal metabolite of niacin promotes vascular inflammation
vii
and contributes to cardiovascular disease risk. Nat Med. 2024
Feb;30(2):424-434. doi: 10.1038/s41591-023-02793-8. Epub
2024 Feb 19. Erratum in: Nat Med. 2024 Jun;30(6):1791. doi:
10.1038/s41591-024-02899-7. PMID: 38374343.
viii
LIST OF TABLES
INTRODUCTION
Table 1. Table 1. Exposure Variables Associated with CVD Outcomes............................ 2
CHAPTER 1
Table 1. Association of Four Air Pollutants with Risk of CAD in the UK Biobank........... 58
Table 2. Association of PM2.5 with Risk of CAD as a Function of rs6494488
Genotype in the UK Biobank and the GeneBank Cohorts.............................................. 59
Supplemental Table 1. Clinical Characteristics of Participants in the
UK Biobank Cohort............................................................................................................. 65
Supplemental Table 2. Association of Air Pollutant Quintiles with Risk of
CAD in the UK Biobank....................................................................................... 66
Supplemental Table 3. Clinical Characteristics of Participants in the
GeneBank Cohort............................................................................................................... 67
Supplemental Table 4. GxE Interactions for Risk of CAD with PM2.5 and Known
Susceptibility Loci in the UK Biobank and the GeneBank Cohorts.............................. 68
Supplemental Table 5. Cis eQTLs Identified in Blood with rs6494488 in
eQTLGen Consortium and GTEx Project................................................................ 74
CHAPTER 2
Table 1. Clinical Characteristics of Hospitalized COVID-19
Cases and Population Controls.............................................................................. 108
Table 2. Association of Hospitalization for COVID-19 with Increased Risk of MACE... 110
Table 3. Hospitalization for COVID-19 is a CAD (CVD) Risk Equivalent.................... 111
Table 4. Hospitalization for COVID-19 Increases Risk of Thrombotic
Events Through Genetic Interaction with ABO Blood Type....................................... 113
Supplemental Table 1. Validation of COVID-19 Case Status and Severity
Through Comparative Association Patterns of Genetic Variants................................. 120
ix
Supplemental Table 2. Clinical Characteristics of All COVID-19
Cases and Population Controls............................................................................... 121
Supplemental Table 3. Causes of Death Among All COVID-19
Cases and Population Controls Based on ICD-10 Codes.................................................... 122
Supplemental Table 4. Association of COVID-19 at All Levels
of Severity with Increased Risk of MACE............................................................... 124
Supplemental Table 5. Association of COVID-19 with Increased
Risk of MACE Across Different Time Periods......................................................... 125
Supplemental Table 6. Sensitivity Analyses for Risk of Thrombotic Events Among
COVID-19 Cases at All Levels of Severity and All Population Controls.......................... 126
Supplemental Table S7. Sensitivity Analyses for Risk of All-cause Mortality Among
COVID-19 Cases at All Levels of Severity and All Population Controls....................... 127
Supplemental Table 8. Sensitivity Analyses for Risk of MACE Among COVID-19
Cases at All Levels of Severity and All Population Controls............................................. 128
Supplemental Table S9. Sensitivity Analyses for Risk of Thrombotic Events
Among Hospitalized COVID-19 Cases and All Population Controls................................ 129
Supplemental Table S10. Sensitivity Analyses for Risk of All-cause Mortality
Among Hospitalized COVID-19 Cases and All Population Controls............................ 130
Supplemental Table 11. Sensitivity Analyses for Risk of MACE Among
Hospitalized COVID-19 Cases and All Population Controls....................................... 131
Supplemental Table 12. Association of Hospitalization for COVID-19
with Increased Risk of Cardiovascular Mortality and MACE...................................... 132
Supplemental Table S13. Association of COVID-19 with Risk of
Thrombotic Events Stratified by Anti-Platelet Agents................................................ 133
Supplemental Table 14. Hospitalization for COVID-19 Does not Increase
Risk of Thrombotic Events Through Interactions with Genetic Variants
Associated with Risk of CAD, SARS-CoV-2 Infection Susceptibility,
or Hospitalization for COVID-19............................................................................ 135
CHAPTER 3
Table 1. Multi-Ancestry and Sex-Stratified Association of rs715 with Risk of CAD........ 170
x
Table 2. Effect of Cps1 Deficiency on Plasma Metabolite Levels................................. 172
Table 3. Effect of Genetic Perturbation on Fasting Cardiometabolic Traits
and Aortic Lesion Formation in in Cps1+/- Mice...................................................... 174
CHAPTER 4
Table 1. Clinical Characteristics of Study Participants................................................ 208
Table 2. Methylation at Previously Identified CpGs Most Strongly
Affected by Smoking........................................................................................................... 209
Table 3. CpG Sites with Significant Changes in Percent Methylation
as a Function of Menopause and Treatment Groups................................................... 210
Table 3. Methylation Levels in Leukocytes of ELITE Participants at CpGs
Previously Associated with Atherosclerosis........................................................................ 211
xi
LIST OF FIGURES
INTRODUCTION
Figure 1. Conceptual Illustration of Genetic, Environmental, and
GxE Interaction Effects........................................................................................................ 7
CHAPTER 1
Figure 1. Association of air pollution with risk of CAD in the UK Biobank...................... 60
Figure 2. Regional plots show association patterns for main SNP effects
and GxE interactions at CAD locus on chromosome 15..................................................... 61
Figure 3. Multi-tissue eQTL plots for TRIP4...................................................................... 62
Figure 4. Effect of DEP on expression of positional candidate genes at
chromosome 15 locus in vitro.............................................................................................. 63
Figure 5. Effect of DEP on expression of positional candidate genes at
chromosome 15 locus in vivo.............................................................................................. 64
CHAPTER 2
Graphical Abstract............................................................................................................... 114
Figure 1. Overview of Clinical and Genetic Analyses........................................................ 115
Figure 2. Hospitalization for COVID-19 is Associated with Increased Risk of MACE.... 116
Figure 3. COVID-19 Represents a CAD Risk Equivalent.................................................. 117
Figure 4. Hospitalization for COVID-19 Increases Risk of Thrombotic
Events Through a Genetic Interaction with ABO Blood Group......................................... 118
Supplemental Figure 1. Association of COVID-19 with Risk of MACE....................... 119
CHAPTER 3
xii
Figure 1. The genes and intermediates of the urea cycle.................................................... 176
Figure 2. Characterization of genetically modified mouse model for Cps1 Deficiency..... 177
Figure 3. Cps1 Deficiency decreases atherosclerosis in mice through sexually
dimorphic pattern................................................................................................................ 178
Figure 4. Effect of gonadectomy on hepatic Cps1 expression and plasma urea
cycle metabolite levels in mice........................................................................................... 179
Figure 5. Natural hepatic CPS1 expression in humans and mice....................................... 180
CHAPTER 4
Figure 1. Miami plot of EWAS results for association of methylation
levels with smoking............................................................................................................ 213
Figure 2. EWAS results for association of methylation levels with
time-since-menopause and HT........................................................................................... 214
Figure 3. Methylation levels at two significantly associated CpG
sites as a function of treatment group................................................................................. 215
Figure 4. Relationship between changes in methylation levels and
subclinical atherosclerosis..................................................................................... 216
DISCUSSION
Figure 1. Genetic Susceptibility Loci and Exposures Synergize to Increase Risk of
Cardiovascular Diseases..................................................................................................... 217
xiii
ABSTRACT
Coronary artery disease (CAD) is a multifactorial disease characterized by the accumulation of
atherosclerotic plaques (composed of oxidized lipids, fibrous tissue, and inflammatory cells),
within arterial walls. This accumulation narrows the coronary arteries, impeding blood flow to
the heart and potentially leading to plaque rupture, arterial occlusion, coronary ischemia, and
myocardial infarction. Extensive epidemiological research has implicated various biological and
environmental risk factors in development of CAD while genetic studies have identified heritable
susceptibility factors as additional risk contributors. Despite the recognized interplay between
genetic predisposition and environmental exposures in the etiology of CAD, research exploring
their interactions remains limited. Therefore, this dissertation focuses on employing a systems
genetics approach to uncover novel pathways involved in CAD pathogenesis through
interactions between genetic factors and various biological and environmental exposures.
1
INTRODUCTION
Overview of Cardiovascular Disease (CVD). Coronary artery disease (CAD), myocardial
infarction (MI), peripheral artery disease (PAD), and stroke (collectively referred to as
atherosclerotic CVD) are the leading causes of death in Western societies [1], even in the
contemporary era of high-potency lipid-lowering therapy [2]. The pathophysiological basis of
CVD is atherosclerosis, a process that develops slowly over decades without necessarily having
overt manifestations. Thus, individuals with CAD or PAD are typically asymptomatic, with the
first clinical indication often being an adverse clinical event (i.e. MI, stroke, or death) [3]. To
date, elevated plasma lipids, diabetes mellitus, systemic arterial hypertension, inflammation, and
cigarette smoking are established causal pathways for atherothrombotic diseases [4]. While
strategies to lower these traditional risk factors have shown significant success in reducing
cardiovascular morbidity and mortality, a substantial residual risk persists, For example, lipid
and blood pressure lowering therapies only partially mitigate risk and >50% of patients with an
acute cardiovascular event do not exhibit elevated levels of these risk factors. Furthermore,
genetic studies strongly support the role of genetic factors in the progression of CAD. Common
genetic variants, those widely distributed within a population and typically exerting small effects
on CAD risk, consistently highlight the importance of established pathways in atherosclerosis
development. Additionally, rare genetic variants, often unique to individuals or families, exert
more profound and deleterious effects on biological processes known to drive CAD progression.
Despite the identification of hundreds of genetic loci associated with CAD risk, the underlying
mechanisms linking these loci to disease pathogenesis remain largely unexplored. Moreover,
these mechanisms appear to operate independently of traditional risk factors, underscoring a
2
critical gap in knowledge and suggesting the existence of additional unrecognized causal
pathways.
Role of the Environment in Influencing Risk of CVD. It is generally accepted that the risk of
CVD is shaped by heritable susceptibility factors, particularly in the context of lifelong exposure
to the atherogenic environments that are increasingly common in industrialized societies. This
concept is supported by decades of epidemiological research. For example, migration studies
showing that risk of CVD progressively increases in Japanese men as function of migrating from
Japan to Hawaii and California [5] provide clear evidence for the adverse effects of a Western
lifestyle [5].The best understood environmental risk factors for CVD are diet, physical activity,
and smoking (Table 1). It is well-known that a sedentary lifestyle or consumption of high levels
of cholesterol and fats, particularly saturated and trans fats, increase risk of CVD whereas
regular physical activity and diets rich in fruits and vegetables and non-red meat lean protein are
protective [6-8]. The environmental risk factor with the strongest adverse effect on the
development of CVD is smoking, which doubles risk even in light smokers [9]. CVD has also
been estimated to account for over half of the ~500,000 annual premature deaths that are
attributed to smoking [1].
Table 1. Exposure Variables Associated with CVD Outcomes.
Exposure Direction of Association with CVD Reference
Diet Positively associated with saturated and
trans-fat and cholesterol content and
inversely associated with vegetable,
fruit, whole grain, lean protein, and
polyunsaturated fat content. Certain
species of gut bacteria are positively
associated through their nutrients, such [6-8]
3
as choline, L-carnitine, and
phenylalanine.
Physical activity Negatively associated with increased
levels.
[6-8]
Smoking Positively associated with increased
levels and accounts for ~250,000
annual premature deaths due to CVD in
the United States alone.
[1]
Air pollution Positively associated with increased
exposure levels of particulate matter,
aerosols, and volatile organic
chemicals.
[10, 11]
Sex Men are more susceptible than women
until menopause at which point risk
becomes equivalent. [12]
Pharmaceuticals Negatively associated with statin
usage.
[2]
Viral agents Positively associated with infection. [13-16]
A large body of evidence has also shown consistent associations between CVD and short- or
long-term exposure to components of traffic-related air pollution, such as coarse, fine, and
ultrafine particulate matter (PM), or nitrogen oxides [10, 11]. Notably, after smoking and dietary
risk factors, more deaths can be attributed to air pollution than low-density lipoprotein (LDL)
cholesterol levels and obesity [17]. Specifically, fine particulate matter <2.5m in diameter
(PM2.5) contributes to nearly four million premature deaths worldwide each year, underscoring
the importance of understanding and addressing of specific environmental factors [18].
4
Sex can also be considered another variable that can interact with genetic factors since it is well
known that men are at higher risk of CVD than women, at least until menopause due to the
protective effect of estrogen [12]. In terms of interventions with pharmaceuticals, initiation of
hormone replacement therapy (HRT) around the time of menopause leads to protective effects
against CVD compared to those who delay HRT until several years after menopause [19].
Pharmaceutical interventions also play a critical role in modulating cardiovascular risk across
sexes. For example, statins, a class of lipid-lowering medications, have consistently
demonstrated effectiveness in reducing the incidence of major adverse coronary events in both
men and women [2].
Pathogens have also long been recognized as potential contributors to CVD risk, with infectious
diseases linked to excess mortality dating back over a century to the influenza epidemic [13].
Notably, acute systemic respiratory tract infections have been shown to increase MI risk [14].
Recent findings further confirm that laboratory-confirmed influenza infection is associated with
elevated MI risk [15]. More recently, infection with another emerging pathogen, SARS-CoV-2,
has also been linked to increased incident MI and stroke, underscoring the cardiovascular impact
of contemporary viral agents [16].
Current State of CVD Genetics. CVD has a strong genetic component based on heritability
estimates that range between 40-60% [20-23]. These results are corroborated by genome-wide
association studies (GWAS), where in the 10 years after the initial identification of the major
CAD locus on 9p21 [24-26] over 100 additional genomic regions were identified for CVD
outcomes [27]. However, just in the last four years alone, the number of susceptibility loci for
CAD, MI, PAD, and stroke has tripled to >300 [28-37]. These discoveries were made possible,
5
in part, through the establishment of collaborative consortia that have harnessed the cumulative
power of meta-analyzing summary statistics across multiple case-control datasets as well as large
population-based cohorts, such as the UK Biobank [38], FinnGen [39], and Biobank Japan [40].
Notably, the results of these large GWAS meta-analyses provide further evidence that loci
influence risk of CVD via perturbations of lipid metabolism, blood pressure regulation,
inflammation, and coagulation, but an obvious underlying biological mechanism for a majority
of the association signals is not readily apparent. Furthermore, the risk alleles at nearly all
identified loci are common in the population and only explain ~20% of the overall heritability
for CVD outcomes [34]. Thus, it has been postulated that there are other unrecognized
contributions from additional common or rare variants, and/or highly penetrant susceptibility
alleles.
If the results of the last few years are any indication, additional common risk alleles for CVD
will undoubtedly be revealed by even larger GWAS that are already starting to include over one
million individuals. However, the effect sizes of such alleles will be even weaker than those
already identified and therefore, not likely to explain a significant fraction of the remaining socalled missing heritability. In this regard, rare variants have also been previously associated with
CVD risk through exome chip or sequencing analysis, with some exhibiting large effect sizes.
Even so, there have still been far fewer rare variants identified for CVD than common single
nucleotide polymorphisms (SNPs) from GWAS [41]. Even with increased sample sizes in more
recent rare variant studies [42-47], the number of novel genes identified for CVD or related risk
factors has not appreciably increased. These observations suggest that rare variants may not
explain a significant fraction of the genetic risk for CVD, although it is still possible that
expansion of rare variant analysis to the entirety of population-based biobanks or more subjects
6
of non-European ancestry could increase the evidence that this class of variants still has
unrecognized significant contributions to mediating CVD risk.
Overview of gene-exposure (GxE) Interactions. From a statistical perspective, GxE
interactions are defined as those where the combined effect of genotype and exposure differs
significantly from the additive effects of genotype and exposure (Figure 1). A straightforward
example for such synergy would be where the effect of smoking on CAD is significantly greater
among carriers of a risk allele than the effect of smoking observed in non-carriers (Figure 1). To
test for such an interaction, a multiplicative interaction term (GxE) can be included in a standard
statistical model along with genotype (G) and exposure (E). Alternatively, genetic analyses can
be carried out in exposed and non-exposed groups separately, followed by interaction tests to
determine whether effect sizes are significantly different between the two exposure groups. This
exact latter approach was used to investigate gene-smoking interactions for risk of CAD and
identified a functional genetic variant upstream of ADAMTS7 that reduced its expression [48].
Despite the recognition that interactions between genes and the environment or even within
genes themselves are likely to play important roles in development of CVD, their identification
has been hampered by insufficient power due to small sample sizes, weak genetic effects, and
imprecise measurement of relevant exposures [49]. However, the recent availability of large
population-based biobanks that have collected clinical, environmental, and genetic data from the
same individuals, has helped overcome some of the challenges of conducting GxE interaction
studies. Additionally, the formation of consortia that have combined multiple cohorts to
collectively include >1 million subjects have further advanced efforts to identify GxE
interactions.
7
Figure 1. Conceptual Illustration of Genetic, Environmental, and GxE Interaction Effects.
Differences in an outcome (i.e. CAD risk) can be due to main effects of genetic (A) or
environmental (B) factors or the additive effect of both genetic and environmental factors (C).
Interactions are conceptualized as effects that manifest (D), synergize (E), or are antagonized (F)
by an exposure.
Gene-Air Pollution Interactions. Numerous previous studies have evaluated GxE interactions
with various air pollutants, including PM, nitrogen or sulfur containing molecules, and ozone.
Most of these studies involved candidate gene analyses with small sample sizes [50]. However,
more recent GxE interaction studies have attempted to use a GWAS approach for gene-air
pollution interaction studies. Using distance to nearest major road as the exposure, one study
identified two loci (BMP8A and BMP2) that were associated with PAD at the genome-wide
8
significance level [51] and a suggestive locus (PIGR-FCAMR) associated with coronary
atherosclerosis [52]. Another recent GWAS investigated GxE interactions for cardiac function
traits and identified a variant downstream of CXCL12 that was associated with greater QT
prolongation in subjects exposed to high PM10 than those in the low exposure group [53].
Notably, the CXCL12 locus was previously known to influence risk of CVD through main
effects since it was identified in the first series of GWAS for CAD and MI over 10 years ago [54,
55]. However, these GxE interactions were derived from analyses in relatively small numbers of
subjects and have yet to be replicated in larger datasets.
Gene-Sex Interactions. Surprisingly, few studies have investigated GxE interactions with sex
for CVD even though it is known that men are at higher risk than women, at least until
menopause. In this regard, variation at CPS1 was shown to exhibit sexually dimorphic
associations with several choline-related and urea cycle metabolites, which translated into
decreased risk of CAD in women but not men [56]. This finding remains one of the few gene-sex
interaction loci reported in the literature for CAD. Recent large-scale GWAS sought to identify
additional gene-sex interaction loci for CAD by performing meta-analyses separately in males
and females; however, no significant novel findings were reported [37]. Additionally, Huang et
al. used over 300,000 subjects from the UK Biobank to demonstrate that association of a
genome-wide polygenic risk score and a 161-locus genetic risk score was more strongly
associated with CAD in men compared to women [57]. Studies have also examined variation on
the X chromosome for association with CAD as another approach to exploring gene-sex
interactions. One such study found no associations on the X chromosome [58, 59] whereas a
more recent study identified a locus that reduced CAD risk through strong lipid lowering effects
[58, 59]. However, these effects were equivalent in men and women [58, 59]. Thus, it has not
9
been conclusively demonstrated that variation at the X chromosome plays a major role in
mediating sex differences with respect to CVD. This is in contrast to studies that have implicated
Y-chromosome as playing a role in CAD pathogenesis[60] but more recent studies in larger
datasets refute this notion as well [61].
Gene-Drug Interactions. Pharmacogenetic research has identified several gene-drug
interactions that significantly affect CVD treatment response, particularly for medications such
as statins that are used to lower LDL cholesterol levels, and warfarin, an anticoagulant. Variants
in genes involved in drug transport and metabolism, such as solute carrier organic anion
transporter family member 1B1 (SLCO1B1) and cytochrome P450 family 2 subfamily C member
9 (CYP2C9), have been linked to variations in drug efficacy and the risk of adverse effects. For
instance, the C allele of the rs4149056 in SLCO1B1, which encodes the liver transporter
OATP1B1, has been shown to reduce hepatic uptake of statins from circulation. This reduction
leads to higher plasma statin levels and promotes its uptake by other tissues, such as skeletal
muscle, which increases risk of statin-induced myopathy [62]. By comparison, CYP2C9, which
metabolizes warfarin, harbors multiple variants (e.g., rs1799853) that decrease enzyme activity.
Patients carrying these variants experience a higher incidence of elevated international
normalized ratio (INR) >4, indicating a greater risk of bleeding compared to non-carriers [63].
These findings emphasize the importance of genotype-guided drug dosing to enhance efficacy
and reduce side effects.
Gene-Pathogen Interactions. Emerging research supports the role of viral infections as triggers
for inflammatory and autoimmune conditions, which can also be influenced by host genetic
factors. For example, coxsackievirus B3 (CVB3) can induce autoimmune myocarditis through
10
immune responses that lead to tissue damage [64]. However, the observation that only some
individuals develop myocarditis following CVB3 infection suggests that genetic susceptibility
might play a role in the response to this viral exposure. In a more contemporary setting, SARSCoV-2 has emerged as another recent pathogen that increased both short-term and long-term risk
for adverse cardiac events, such as MI and stroke [16]. Given that SARS-CoV-2 viral particles
have been detected within atherosclerotic plaques and shown to trigger pro-atherogenic
inflammatory responses [65], it is possible that infection with this virus may also heighten the
risk of thrombotic events, especially in individuals who are already genetically predisposed to
MI [35, 66].
Recent Advances for Investigating GxE Interactions. The analytical frameworks typically
used for investigating GxE interactions described above poses certain challenges that may
explain why the results of candidate gene GxE interaction studies are often not replicated or why
unbiased GxE interaction studies have, until recently, not been largely pursued. For example,
detecting GxE interactions have been estimated to require ~4-fold larger sample sizes than
detecting main effects of exposures or SNPs alone [67]. Notably, the availability of large
population-based cohorts, such as the UK Biobank [38], FinnGen [39], Biobank Japan [40], and
All of Us Project [68] may overcome such sample size limitations, as has been shown recently
for cardiometabolic traits [69]. Moreover, such large datasets can also facilitate replication of the
GxE interactions since the associations still require validation in independent populations prior to
considering them as true GxE signals.
As an alternative to human studies, mouse models also provide an excellent complementary
approach to investigating GxE interactions since many experimental variables, such as age, sex,
11
and exposures of interest can be tightly controlled. Animal models also have additional
advantages as compared to human studies, including access to tissues, defined genetic
backgrounds, and experimental validation. This is illustrated by studies using ~100 strains from
the Hybrid Mouse Diversity Panel [70], which has led to identification of robust GxE
interactions for diet-induced obesity [71] and responses to inflammatory stimuli [72] or air
pollution-related exposures [73, 74].
Summary. Significant advances in our understanding of GxE interactions in CVD have been
achieved through collaborative efforts leveraging large-scale population-based cohorts and
extensive case-control datasets. These well characterized resources are now being leveraged to
perform nuanced analyses that reveal unidentified mechanisms by which human genes interact
with exposures such as air pollution, infectious diseases, and other environmental factors,
contributing to enhanced CVD risk. Building on advancements in integrating phenotype with
exposures, this dissertation leverages contemporary resources to dissect the intricate interplay
between host genetic variation and environmental factors. Using a systems genetics approach,
these efforts led to the discovery of novel GxE interactions that contribute to CVD and related
atherosclerotic phenotypes. It presents four focused investigations into distinct GxE interactions
with the first chapter examining how PM2.5 exposure modulates CAD risk based on individual
genotypes at previously known CAD loci. The second chapter investigates gene-pathogen
interactions for risk of CAD, focusing on the novel SARS-CoV-2 virus to determine whether
established MI susceptibility loci modulate the risk of future thrombotic events in infected
individuals. The third chapter extends our understanding of a gene-sex interaction by replicating
the sexually dimorphic protective effects of a variant (rs715) of the CPS1 gene on risk of CAD
through a large-scale multi-ancestry meta-analysis, followed by functional validation of the sex-
12
interaction through genetic perturbation of Cps1 in murine models. Lastly, the fourth chapter
introduces an epigenetic framework centered on pharmacogenetic interactions in women to
explore how HRT and menopause timing interact to alter methylation patterns over time.
Together, these chapters aim to elucidate distinct mechanisms by which environmental exposures
intersect with genetic factors, with an overarching goal of shedding light on novel pathways that
drive CVD etiology.
13
References
1. Virani SS, Alonso A, Benjamin EJ, Bittencourt MS, Callaway CW, Carson AP, et al. Heart
Disease and Stroke Statistics-2020 Update: A Report From the American Heart
Association. Circulation. 2020;141(9):e139-e596. doi: 10.1161/CIR.0000000000000757.
2. Ridker PM, Danielson E, Fonseca FA, Genest J, Gotto AM, Jr., Kastelein JJ, et al.
Rosuvastatin to prevent vascular events in men and women with elevated C-reactive
protein. N Engl J Med. 2008;359(21):2195-207. doi: 10.1056/NEJMoa0807646.
3. Hansson GK. Inflammation, atherosclerosis, and coronary artery disease. N Engl J Med.
2005;352(16):1685-95.
4. Grundy SM, Stone NJ, Bailey AL, Beam C, Birtcher KK, Blumenthal RS, et al. 2018
AHA/ACC/AACVPR/AAPA/ABC/ACPM/ADA/AGS/APhA/ASPC/NLA/PCNA
Guideline on the Management of Blood Cholesterol: Executive Summary: A Report of
the American College of Cardiology/American Heart Association Task Force on Clinical
Practice Guidelines. Circulation. 2019;139(25):e1046-e81. doi:
10.1161/CIR.0000000000000624.
5. Worth RM, Kato H, Rhoads GG, Kagan K, Syme SL. Epidemiologic studies of coronary heart
disease and stroke in Japanese men living in Japan, Hawaii and California: mortality. Am
J Epidemiol. 1975;102(6):481-90. doi: 10.1093/oxfordjournals.aje.a112186.
6. Hu FB, Stampfer MJ, Manson JE, Rimm E, Colditz GA, Rosner BA, et al. Dietary fat intake
and the risk of coronary heart disease in women. N Engl J Med. 1997;337(21):1491-9.
doi: 10.1056/NEJM199711203372102.
7. Hu FB, Rimm EB, Stampfer MJ, Ascherio A, Spiegelman D, Willett WC. Prospective study of
major dietary patterns and risk of coronary heart disease in men. Am J Clin Nutr.
2000;72(4):912-21. doi: 10.1093/ajcn/72.4.912.
8. Mozaffarian D, Katan MB, Ascherio A, Stampfer MJ, Willett WC. Trans fatty acids and
cardiovascular disease. N Engl J Med. 2006;354(15):1601-13. doi:
10.1056/NEJMra054035.
9. Schane RE, Ling PM, Glantz SA. Health effects of light and intermittent smoking: a review.
Circulation. 2010;121(13):1518-22. doi: 10.1161/CIRCULATIONAHA.109.904235.
10. Pope CA, 3rd, Turner MC, Burnett RT, Jerrett M, Gapstur SM, Diver WR, et al.
Relationships between fine particulate air pollution, cardiometabolic disorders, and
cardiovascular mortality. Circ Res. 2015;116(1):108-15. doi:
10.1161/CIRCRESAHA.116.305060.
11. Cosselman KE, Navas-Acien A, Kaufman JD. Environmental factors in cardiovascular
disease. Nat Rev Cardiol. 2015;12(11):627-42. doi: 10.1038/nrcardio.2015.152.
14
12. El Khoudary SR, Aggarwal B, Beckie TM, Hodis HN, Johnson AE, Langer RD, et al.
Menopause Transition and Cardiovascular Disease Risk: Implications for Timing of
Early Prevention: A Scientific Statement From the American Heart Association.
Circulation. 2020;142(25):e506-e32. doi: 10.1161/CIR.0000000000000912.
13. Housworth J, Langmuir AD. Excess mortality from epidemic influenza, 1957-1966. Am J
Epidemiol. 1974;100(1):40-8. doi: 10.1093/oxfordjournals.aje.a112007.
14. Smeeth L, Thomas SL, Hall AJ, Hubbard R, Farrington P, Vallance P. Risk of myocardial
infarction and stroke after acute infection or vaccination. N Engl J Med.
2004;351(25):2611-8. doi: 10.1056/NEJMoa041747.
15. Kwong JC, Schwartz KL, Campitelli MA, Chung H, Crowcroft NS, Karnauchow T, et al.
Acute Myocardial Infarction after Laboratory-Confirmed Influenza Infection. N Engl J
Med. 2018;378(4):345-53. doi: 10.1056/NEJMoa1702090.
16. Al-Aly Z, Bowe B, Xie Y. Long COVID after breakthrough SARS-CoV-2 infection. Nat
Med. 2022;28(7):1461-7. doi: 10.1038/s41591-022-01840-0.
17. Brauer M, Casadei B, Harrington RA, Kovacs R, Sliwa K, Group WHFAPE. Taking a Stand
Against Air Pollution-The Impact on Cardiovascular Disease: A Joint Opinion From the
World Heart Federation, American College of Cardiology, American Heart Association,
and the European Society of Cardiology. Circulation. 2021;143(14):e800-e4. doi:
10.1161/CIRCULATIONAHA.120.052666.
18. Fuller R, Landrigan PJ, Balakrishnan K, Bathan G, Bose-O'Reilly S, Brauer M, et al.
Pollution and health: a progress update. Lancet Planet Health. 2022;6(6):e535-e47. doi:
10.1016/S2542-5196(22)00090-0.
19. Hodis HN, Mack WJ, Henderson VW, Shoupe D, Budoff MJ, Hwang-Levine J, et al.
Vascular Effects of Early versus Late Postmenopausal Treatment with Estradiol. N Engl J
Med. 2016;374(13):1221-31. doi: 10.1056/NEJMoa1505241.
20. Wienke A, Holm NV, Skytthe A, Yashin AI. The heritability of mortality due to heart
diseases: a correlated frailty model applied to Danish twins. Twin Res. 2001;4(4):266-74.
doi: 10.1375/1369052012399.
21. Zdravkovic S, Wienke A, Pedersen NL, Marenberg ME, Yashin AI, De Faire U. Heritability
of death from coronary heart disease: a 36-year follow-up of 20 966 Swedish twins. J
Intern Med. 2002;252(3):247-54. doi: 10.1046/j.1365-2796.2002.01029.x.
22. Bevan S, Traylor M, Adib-Samii P, Malik R, Paul NL, Jackson C, et al. Genetic heritability
of ischemic stroke and the contribution of previously reported candidate gene and
genomewide associations. Stroke. 2012;43(12):3161-7. doi:
10.1161/STROKEAHA.112.665760.
15
23. Vinkhuyzen AA, Wray NR, Yang J, Goddard ME, Visscher PM. Estimation and partition of
heritability in human populations using whole-genome analysis methods. Annu Rev
Genet. 2013;47:75-95. doi: 10.1146/annurev-genet-111212-133258.
24. McPherson R, Pertsemlidis A, Kavaslar N, Stewart A, Roberts R, Cox DR, et al. A common
allele on chromosome 9 associated with coronary heart disease. Science.
2007;316(5830):1488-91. doi: 10.1126/science.1142447.
25. Helgadottir A, Thorleifsson G, Manolescu A, Gretarsdottir S, Blondal T, Jonasdottir A, et al.
A common variant on chromosome 9p21 affects the risk of myocardial infarction.
Science. 2007;316(5830):1491-3. doi: 10.1126/science.1142842.
26. Consortium WTCC. Genome-wide association study of 14,000 cases of seven common
diseases and 3,000 shared controls. Nature. 2007;447(7145):661-78. doi:
10.1038/nature05911.
27. Nikpay M, Goel A, Won HH, Hall LM, Willenborg C, Kanoni S, et al. A comprehensive
1,000 Genomes-based genome-wide association meta-analysis of coronary artery disease.
Nat Genet. 2015;47(10):1121-30. doi: 10.1038/ng.3396.
28. Howson JMM, Zhao W, Barnes DR, Ho WK, Young R, Paul DS, et al. Fifteen new risk loci
for coronary artery disease highlight arterial-wall-specific mechanisms. Nat Genet.
2017;49(7):1113-9. doi: 10.1038/ng.3874.
29. Nelson CP, Goel A, Butterworth AS, Kanoni S, Webb TR, Marouli E, et al. Association
analyses based on false discovery rate implicate new loci for coronary artery disease. Nat
Genet. 2017;49(9):1385-91. doi: 10.1038/ng.3913.
30. Klarin D, Zhu QM, Emdin CA, Chaffin M, Horner S, McMillan BJ, et al. Genetic analysis in
UK Biobank links insulin resistance and transendothelial migration pathways to coronary
artery disease. Nat Genet. 2017;49(9):1392-7. doi: 10.1038/ng.3914.
31. van der Harst P, Verweij N. Identification of 64 novel genetic loci provides an expanded
view on the genetic architecture of coronary artery disease. Circ Res. 2018;122(3):433-
43. doi: 10.1161/CIRCRESAHA.117.312086.
32. Malik R, Chauhan G, Traylor M, Sargurupremraj M, Okada Y, Mishra A, et al.
Multiancestry genome-wide association study of 520,000 subjects identifies 32 loci
associated with stroke and stroke subtypes. Nat Genet. 2018;50(4):524-37. doi:
10.1038/s41588-018-0058-3.
33. Klarin D, Lynch J, Aragam K, Chaffin M, Assimes TL, Huang J, et al. Genome-wide
association study of peripheral artery disease in the Million Veteran Program. Nat Med.
2019;25(8):1274-9. doi: 10.1038/s41591-019-0492-5.
34. Koyama S, Ito K, Terao C, Akiyama M, Horikoshi M, Momozawa Y, et al. Populationspecific and trans-ancestry genome-wide analyses identify distinct and shared genetic risk
16
loci for coronary artery disease. Nat Genet. 2020;52(11):1169-77. doi: 10.1038/s41588-
020-0705-3.
35. Hartiala JA, Han Y, Jia Q, Hilser JR, Huang P, Gukasyan J, et al. Genome-wide analysis
identifies novel susceptibility loci for myocardial infarction. Eur Heart J. 2021;42(9):919-
33. doi: 10.1093/eurheartj/ehaa1040.
36. Tcheandjieu C, Zhu X, Hilliard AT, Clarke SL, Napolioni V, Ma S, et al. Large-scale
genome-wide association study of coronary artery disease in genetically diverse
populations. Nat Med. 2022;28(8):1679-92. doi: 10.1038/s41591-022-01891-3.
37. Aragam KG, Jiang T, Goel A, Kanoni S, Wolford BN, Atri DS, et al. Discovery and
systematic characterization of risk variants and genes for coronary artery disease in over
a million participants. Nat Genet. 2022;54(12):1803-15. doi: 10.1038/s41588-022-01233-
6.
38. Bycroft C, Freeman C, Petkova D, Band G, Elliott LT, Sharp K, et al. The UK Biobank
resource with deep phenotyping and genomic data. Nature. 2018;562(7726):203-9. doi:
10.1038/s41586-018-0579-z.
39. Tabassum R, Ramo JT, Ripatti P, Koskela JT, Kurki M, Karjalainen J, et al. Genetic
architecture of human plasma lipidome and its link to cardiovascular disease. Nat
Commun. 2019;10(1):4329. doi: 10.1038/s41467-019-11954-8.
40. Nagai A, Hirata M, Kamatani Y, Muto K, Matsuda K, Kiyohara Y, et al. Overview of the
BioBank Japan Project: Study design and profile. J Epidemiol. 2017;27(3S):S2-S8. doi:
10.1016/j.je.2016.12.005.
41. Hartiala J, Schwartzman WS, Gabbay J, Ghazalpour A, Bennett BJ, Allayee H. The genetic
architecture of coronary artery disease: current knowledge and future opportunities. Curr
Atheroscler Rep. 2017;19(2):6. doi: 10.1007/s11883-017-0641-6.
42. Eicher JD, Chen MH, Pitsillides AN, Lin H, Veeraraghavan N, Brody JA, et al. Whole
exome sequencing in the Framingham Heart Study identifies rare variation in HYAL2
that influences platelet aggregation. Thromb Haemost. 2017;117(6):1083-92. doi:
10.1160/TH16-09-0677.
43. Jian X, Satizabal CL, Smith AV, Wittfeld K, Bis JC, Smith JA, et al. Exome chip analysis
identifies low-frequency and rare variants in mrpl38 for white matter hyperintensities on
brain magnetic resonance imaging. Stroke. 2018;49(8):1812-9. doi:
10.1161/STROKEAHA.118.020689.
44. Paraboschi EM, Khera AV, Merlini PA, Gigante L, Peyvandi F, Chaffin M, et al. Rare
variants lowering the levels of coagulation factor X are protective against ischemic heart
disease. Haematologica. 2020;105(7):e365-e9. doi: 10.3324/haematol.2019.237750.
17
45. Pedicino D, Vinci R, Giglio AF, Pisano E, Porto I, Vergallo R, et al. Alterations of
hyaluronan metabolism in acute coronary syndrome: implications for plaque erosion. J
Am Coll Cardiol. 2018;72(13):1490-503. doi: 10.1016/j.jacc.2018.06.072.
46. Backman JD, Li AH, Marcketta A, Sun D, Mbatchou J, Kessler MD, et al. Exome
sequencing and analysis of 454,787 UK Biobank participants. Nature.
2021;599(7886):628-34. doi: 10.1038/s41586-021-04103-z.
47. Clarke B, Holtkamp E, Öztürk H, Mück M, Wahlberg M, Meyer K, et al. Integration of
variant annotations using deep set networks boosts rare variant association testing. Nat
Genet. 2024;56(10):2271-80. doi: 10.1038/s41588-024-01919-z.
48. Saleheen D, Zhao W, Young R, Nelson CP, Ho W, Ferguson JF, et al. Cardioprotective
effects at the ADAMTS7 locus as a result of gene-smoking interactions. Circulation.
2017;135(24):2336-53. doi: 10.1161/CIRCULATIONAHA.116.022069.
49. Sackton TB, Hartl DL. Genotypic context and epistasis in individuals and populations. Cell.
2016;166(2):279-87. doi: 10.1016/j.cell.2016.06.047.
50. Ward-Caviness CK. A review of gene-by-air pollution interactions for cardiovascular
disease, risk factors, and biomarkers. Hum Genet. 2019;138(6):547-61. doi:
10.1007/s00439-019-02004-w.
51. Ward-Caviness CK, Neas LM, Blach C, Haynes CS, LaRocque-Abramson K, Grass E, et al.
Genetic variants in the bone morphogenic protein gene family modify the association
between residential exposure to traffic and peripheral arterial disease. PLoS One.
2016;11(4):e0152670. doi: 10.1371/journal.pone.0152670.
52. Ward-Caviness CK, Neas LM, Blach C, Haynes CS, LaRocque-Abramson K, Grass E, et al.
A genome-wide trans-ethnic interaction study links the PIGR-FCAMR locus to coronary
atherosclerosis via interactions between genetic variants and residential exposure to
traffic. PLoS One. 2017;12(3):e0173880. doi: 10.1371/journal.pone.0173880.
53. Gondalia R, Avery CL, Napier MD, Mendez-Giraldez R, Stewart JD, Sitlani CM, et al.
Genome-wide association study of susceptibility to particulate matter-associated QT
prolongation. Environ Health Perspect. 2017;125(6):067002. doi: 10.1289/EHP347.
54. Samani NJ, Erdmann J, Hall AS, Hengstenberg C, Mangino M, Mayer B, et al. Genomewide
association analysis of coronary artery disease. N Engl J Med. 2007;357(5):443-53.
55. Myocardial Infarction Genetics C, Kathiresan S, Voight BF, Purcell S, Musunuru K,
Ardissino D, et al. Genome-wide association of early-onset myocardial infarction with
single nucleotide polymorphisms and copy number variants. Nat Genet. 2009;41(3):334-
41. doi: 10.1038/ng.327.
56. Hartiala JA, Tang WH, Wang Z, Crow AL, Stewart AF, Roberts R, et al. Genome-wide
association study and targeted metabolomics identifies sex-specific association of CPS1
with coronary artery disease. Nat Commun. 2016;7:10558. doi: 10.1038/ncomms10558.
18
57. Huang Y, Hui Q, Gwinn M, Hu YJ, Quyyumi AA, Vaccarino V, et al. Sexual differences in
genetic predisposition of coronary artery disease. Circ Genom Precis Med.
2021;14(1):e003147. doi: 10.1161/CIRCGEN.120.003147.
58. Loley C, Alver M, Assimes TL, Bjonnes A, Goel A, Gustafsson S, et al. No Association of
Coronary Artery Disease with X-Chromosomal Variants in Comprehensive International
Meta-Analysis. Sci Rep. 2016;6:35278. doi: 10.1038/srep35278.
59. Natarajan P, Pampana A, Graham SE, Ruotsalainen SE, Perry JA, de Vries PS, et al.
Chromosome Xq23 is associated with lower atherogenic lipid concentrations and
favorable cardiometabolic indices. Nat Commun. 2021;12(1):2182. doi: 10.1038/s41467-
021-22339-1.
60. Charchar FJ, Bloomer LD, Barnes TA, Cowley MJ, Nelson CP, Wang Y, et al. Inheritance of
coronary artery disease in men: an analysis of the role of the Y chromosome. Lancet.
2012;379(9819):915-22. doi: 10.1016/S0140-6736(11)61453-0.
61. Timmers PRHJ, Wilson JF. Limited Effect of Y Chromosome Variation on Coronary Artery
Disease and Mortality in UK Biobank-Brief Report. Arterioscler Thromb Vasc Biol.
2022;42(9):1198-206. doi: 10.1161/ATVBAHA.122.317664.
62. Link E, Parish S, Armitage J, Bowman L, Heath S, Matsuda F, et al. SLCO1B1 variants and
statin-induced myopathy--a genomewide study. N Engl J Med. 2008;359(8):789-99. doi:
10.1056/NEJMoa0801936.
63. Schwarz UI, Ritchie MD, Bradford Y, Li C, Dudek SM, Frye-Anderson A, et al. Genetic
determinants of response to warfarin during initial anticoagulation. N Engl J Med.
2008;358(10):999-1008. doi: 10.1056/NEJMoa0708078.
64. Rose NR. Viral myocarditis. Curr Opin Rheumatol. 2016;28(4):383-9. doi:
10.1097/BOR.0000000000000303.
65. Eberhardt N, Noval MG, Kaur R, Amadori L, Gildea M, Sajja S, et al. SARS-CoV-2
infection triggers pro-atherogenic inflammatory responses in human coronary vessels.
Nat Cardiovasc Res. 2023;2(10):899-916. doi: 10.1038/s44161-023-00336-5.
66. Reilly MP, Li M, He J, Ferguson JF, Stylianou IM, Mehta NN, et al. Identification of
ADAMTS7 as a novel locus for coronary atherosclerosis and association of ABO with
myocardial infarction in the presence of coronary atherosclerosis: two genome-wide
association studies. Lancet. 2011;377(9763):383-92. doi: 10.1016/S0140-6736(10)61996-
4.
67. Smith PG, Day NE. The design of case-control studies: the influence of confounding and
interaction effects. Int J Epidemiol. 1984;13(3):356-65. doi: 10.1093/ije/13.3.356.
68. Denny JC, Rutter JL, Goldstein DB, Philippakis A, Smoller JW, Jenkins G, et al. The "All of
Us" Research Program. N Engl J Med. 2019;381(7):668-76. doi:
10.1056/NEJMsr1809937.
19
69. Wang H, Zhang F, Zeng J, Wu Y, Kemper KE, Xue A, et al. Genotype-by-environment
interactions inferred from genetic effects on phenotypic variability in the UK Biobank.
Sci Adv. 2019;5(8):eaaw3538. doi: 10.1126/sciadv.aaw3538.
70. Lusis AJ, Seldin MM, Allayee H, Bennett BJ, Civelek M, Davis RC, et al. The Hybrid
Mouse Diversity Panel: a resource for systems genetics analyses of metabolic and
cardiovascular traits. J Lipid Res. 2016;57(6):925-42. doi: 10.1194/jlr.R066944.
71. Parks BW, Nam E, Org E, Kostem E, Norheim F, Hui ST, et al. Genetic control of obesity
and gut microbiota composition in response to high-fat, high-sucrose diet in mice. Cell
Metab. 2013;17(1):141-52.
72. Orozco LD, Bennett BJ, Farber CR, Ghazalpour A, Pan C, Che N, et al. Unraveling
inflammatory responses using systems genetics and gene-environment interactions in
macrophages. Cell. 2012;151(3):658-70. doi: 10.1016/j.cell.2012.08.043.
73. Maazi H, Hartiala JA, Suzuki Y, Crow AL, Shafiei Jahani P, Lam J, et al. A GWAS
approach identifies Dapp1 as a determinant of air pollution-induced airway
hyperreactivity. PLoS Genet. 2019;15(12):e1008528. doi: 10.1371/journal.pgen.1008528.
74. Bhetraratana M, Orozco LD, Bennett BJ, Luna K, Yang X, Lusis AJ, et al. Diesel exhaust
particle extract elicits an oxPAPC-like transcriptomic profile in macrophages across
multiple mouse strains. Environ Pollut. 2024;358:124415. doi:
10.1016/j.envpol.2024.124415.
20
CHAPTER 1: GENE-ENVIRONMENT INTERACTION BETWEEN THE TRIP4 LOCUS AND
AIR POLLUTION EXPOSURE INCREASES RISK OF CORONARY ARTERY DISEASE
Summary. In the present study, we investigated gene-air pollution interactions in the risk of
coronary artery disease (CAD). In the UK Biobank, a population-based cohort, we analyzed over
300 known SNPs for interaction on risk of CAD with fine particulate matter <2.5m in diameter
(PM2.5), an air pollutant previously identified as clinically associated with CAD. To validate our
findings, we extended our analysis to the GeneBank, an angiography-based cohort. Our study
identified a significant gene-environment (GxE) interaction between the rs6494488 variant and
ambient PM2.5 levels, demonstrating that genetic variation at this locus modifies CAD risk in
response to air pollution exposure. Expression quantitative locus (eQTL) analysis in humans
revealed that rs6494488 regulates TRIP4 expression in atherosclerosis-relevant tissues, such as
lung, coronary artery, and adipose tissues. Air pollution exposure experiments in humans and
mice confirmed that exposure to air pollution decreased TRIP4 expression, suggesting a
mechanistic link between air pollution and the genetic architecture of CAD. My contributions to
this study included designing and executing the bioinformatics pipeline, implementing the geneenvironment interaction analysis, and integrating publicly available data with cohort-specific
findings. I also performed the analysis of murine model exposure data for functional validation
of the identified locus and contributed to manuscript curation.
21
Gene-Environment Interaction Between the TRIP4 Locus and Air Pollution Exposure
Increases Risk of Coronary Artery Disease
James R. Hilser1,2, Yi Han1,2
, Yuxin Yang1,2
, Hongqiao Zhang3
, Caleb E. Finch3
, William J.
Mack4,5, Guanglin Zhang6,7,8, In Sook Ahn6,7,8, Xia Yang6,7,8,9, W.H. Wilson Tang10,11,12
, Clint L.
Miller13,14,15, Lijiang Ma16, Johan L.M. Björkegren17, Frank D. Gilliland1
, Jesus A. Araujo9,18,19
,
Stanley L. Hazen10,11,12
, Hooman Allayee1,2*, and Jaana A. Hartiala1*
Departments of 1Population and Public Health Sciences, 2Biochemistry & Molecular Medicine,
Keck School of Medicine, University of Southern California, Los Angeles, CA 90033; 3Leonard
Davis School of Gerontology, University of Southern California, Los Angeles, California,
90089; Department of 4Neurological Surgery and 5Zilkha Neurogenetics Institute, Keck School
of Medicine, University of Southern California, Los Angeles, CA 90033; 6Department of
Integrative Biology and Physiology, 7Molecular, Cellular, and Integrative Physiology
Interdepartmental Program, and 8
Institute for Quantitative and Computational Biosciences,
University of California, Los Angeles, Los Angeles, CA 90095; 9Environmental and Molecular
Toxicology Interdepartmental Program, University of California Los Angeles, Los Angeles, CA
90095; 10Department of Cardiovascular and Metabolic Sciences, Lerner Research Institute,
Cleveland Clinic, Cleveland OH 44106; 11Center for Microbiome and Human Health, Cleveland
Clinic, Cleveland OH 44106; 12Department of Cardiovascular Medicine, Heart, Vascular and
Thoracic Institute, Cleveland Clinic, Cleveland OH 44195; Departments of 13Biochemistry &
22
Molecular Genetics and 14Public Health Sciences, and 15Center for Public Health Genomics,
University of Virginia, Charlottesville University of Virginia, Charlottesville, VA 22904.
16Department of Genetics & Genomic Sciences, Institute of Genomics and Multiscale Biology,
Icahn School of Medicine at Mount Sinai, New York, NY, 10029; 17Department of Medicine,
Karolinska Institutet, Karolinska Universitetssjukhuset, 141 57 Huddinge, Sweden;
18Department of Medicine, Division of Cardiology, David Geffen School of Medicine of UCLA,
Los Angeles, CA 90095; 19Department of Environmental Health Sciences, Fielding School of
Public Health at UCLA, Los Angeles, CA 90095;
23
Abstract
Background and Aims: Coronary artery disease (CAD) is characterized by both heritable
susceptibility factors and pro-atherogenic environmental exposures, such as air pollution.
However, progress in identifying gene-environment (GxE) interactions for risk of CAD has been
limited.
Methods: Levels of fine particulate matter <2.5m (PM2.5) or <10m (PM10) in diameter,
nitrogen dioxide (NO2), and nitrogen oxides (NOX) were evaluated for association with risk of
CAD in the UK Biobank (n=377,761) by logistic regression. A meta-analysis of genetic and air
pollution data from the UK Biobank and GeneBank cohorts (total n=381,867) was used to test
320 known CAD susceptibility loci for GxE interactions with PM2.5 levels. Candidate causal
genes were prioritized using bioinformatics analyses, in vitro experiments, and mouse exposure
studies.
Results: Risk of CAD was higher as a function of a 2SD increase in levels of PM2.5 (OR=1.11,
95% CI 1.08-1.14; P=8.2x10-13), NO2 (OR=1.06, 95% CI 1.03-1.09; P=6.8x10-5
), and NOX
(OR=1.07, 95% CI 1.04-1.10; P=7.4x10-7
). A significant GxE interaction was identified with a
variant (rs6494488; A>G) at a chromosome 15 locus (P-int=1.2x10-4
), where PM2.5 exposure
increased risk of CAD to a greater degree in AA homozygotes (OR=1.28, 95% CI 1.25-1.32;
P=8.4x10-60) than AG heterozygotes (OR=1.17, 95% CI 1.11-1.23; P=1.4x10-9
) but not in GG
homozygotes (OR=1.06, 95% CI (0.92-1.22; P=0.42). Of the genes at this locus, eQTL analyses
in cardiometabolic tissues revealed that the A allele was only associated with lower mRNA
levels of thyroid receptor interacting protein 4 (TRIP4). TRIP4 expression localized to major
cell types of atherosclerotic lesions and was ~11% lower in aortas of CAD patients than control
24
subjects (P=2.7x10-4
). Expression of TRIP4, but no other gene at the chromosome 15 locus, was
significantly decreased in human endothelial cells incubated with either plasma from subjects
exposed to whole diesel exhaust (P=2.8x10-4
) or with diesel exhaust particles (DEP) extract
(P=0.02). Long-term exposure of mice to DEP also decreased mRNA levels of Trip4 (P=5.9x10-
3
) in the aorta but no other gene at the chromosome 15 locus.
Conclusions: These results provide evidence for a GxE interaction between a chromosome 15
locus and air pollution exposure that is mechanistically linked to putative athero-protective
properties of TRIP4 at the level of the vessel wall.
25
Introduction
Coronary artery disease (CAD) is characterized by heritable susceptibility factors in the
context of pro-atherogenic environments. For example, CAD has a strong genetic component
based on both heritability estimates of 40-60% [1-4] and the results of gene mapping efforts. In
particular, genome-wide association studies (GWAS) have discovered >300 distinct loci with
main effects that influence risk of CAD and myocardial infarction (MI) through perturbations of
lipid metabolism, blood pressure regulation, inflammation, platelet function, and other
mechanisms that remain unknown [5-10]. Since the risk alleles at these loci, most of which are
common, collectively only account for ~20% of the overall heritability for CAD, it has been
hypothesized that there are other unrecognized contributions from additional common variants
with weaker effect sizes, rare susceptibility alleles with high penetrance, and/or interactions
between genes themselves or with the environment [11].
It is also well-known that environmental exposures play important roles in the
pathogenesis of CAD. For example, a large body of evidence has shown consistent associations
between risk of CAD and short- or long-term exposure to components of air pollution, such as
fine particulate matter <2.5m in diameter (PM2.5), <10m in diameter (PM10), nitrogen dioxide
(NO2), and nitrogen oxides (NOX) [12-17]. Notably, ambient air pollution has been reported to
be the leading contributor to the global burden of diseases [18], and after smoking and dietary
risk factors, more deaths could be attributed to air pollution than LDL cholesterol and obesity
[19]. These observations are supported by functional experimentation in animal models, which
has also shown that PM2.5 and other pollutants, such as diesel exhaust particles (DEP), promote
the development of atherosclerosis and related cardiometabolic traits [20-25]. Multiple plausible
26
mechanisms have been proposed for these associations, including decreased lipoprotein function,
induction of inflammation and oxidative stress, endothelial dysfunction, thrombosis, and
arrhythmia [26-33].
Despite advances in the field of environmental cardiology, our understanding of how air
pollution promotes atherogenesis and increases risk of CAD remains incomplete. Furthermore,
the interactions between genetic factors and air pollution are also poorly understood [34, 35] and
have only been explored in a small number of narrowly focused candidate gene studies [36-38]
because of limitations in power due to small sample sizes and statistical methodology challenges.
In the present study, we addressed these gaps in knowledge through a large-scale analysis with
data from the UK Biobank and GeneBank cohorts, and functional experiments to identify and
validate novel gene-environment (GxE) interactions between known CAD susceptibility loci and
air pollutants.
27
Materials and Methods
Study Populations. The UK Biobank recruited participants between 40-69 years of age who
were registered with a general practitioner of the UK National Health Service (NHS) [39]. From
2006-2010, a total of 503,325 individuals were included. Extensive data on demographics,
ethnicity, education, and disease-related outcomes were obtained through questionnaires or
health records. All study participants provided informed consent, and the study was approved by
the North West Multi-centre Research Ethics Committee. The Cleveland Clinic GeneBank study
is a single‐site sample repository generated from consecutive patients undergoing elective
diagnostic coronary angiography or elective cardiac computed tomographic angiography with
extensive clinical and laboratory characterization and longitudinal observation. Participant
recruitment occurred between 2001 and 2007, and all patients provided written informed consent
prior to being enrolled. Ethnicity was self‐reported, and information regarding demographics,
medical history, and medication use was obtained by patient interviews and confirmed by chart
reviews at baseline enrollment. All clinical outcome data were verified by source documentation.
The GeneBank Study has been used previously for discovery and replication of novel genes and
risk factors for CAD [40-45]. The present study was approved by the Institutional Review
Boards of the University of Southern California Keck School of Medicine and the Cleveland
Clinic.
Air Pollution Exposure Assessment. Land use regression (LUR)-based estimates for annual
average for 2010 of PM2.5, PM10, NO2 and NOx were generated as part of European Study of
Cohorts for Air Pollution Effects (ESCAPE) project [46, 47] and linked to geocoded residential
addresses of UK Biobank participants. In brief, PM2.5, and PM10 were measured in 20 European
28
study areas at 20 sites per area. The concentrations of NO2 and NOx were measured with Ogawa
passive samplers at 40 or 80 sites in each of the 36 study areas in Europe. The spatial variation
in each area was explained by LUR modelling. Centrally and locally available Geographic
Information System (GIS) variables were used as potential predictors. A leave-one out crossvalidation procedure was used to evaluate the model performance. The model explained
variances (R
2
) of the LUR models yield median of 82%, 78%, 77% and 71% for NO2, NOx,
PM10, and PM2.5, respectively.
Exposure estimates for air pollutants and their associations with CAD outcomes in the
GeneBank cohort have been described previously [48]. Briefly, daily concentrations of PM2.5 in
the United States from 1998 through 2010 were downloaded from the US Environmental
Protection Agency's (EPA) Air Quality System (AQS) database (https://www.epa.gov/aqs).
Hourly PM2.5 levels were averaged into standard daily exposure metrics, and monthly averages
were calculated from the daily average pollutant data. Monthly air‐quality exposure values were
spatially interpolated from the air‐quality monitoring locations of the residential ZIP code
coordinates (based on geographic centroid) of each participant at the time of enrollment into the
GeneBank study. The station‐specific monthly air‐quality data were spatially interpolated using
inverse distance‐squared weighting. The data from up to four air‐quality measurement stations
were included in each interpolation. Because of the regional nature of air pollutant
concentrations, a maximum interpolation radius of 50km was used; however, when a residence
was located within 5km of ≥1 station with valid observations, the interpolation was based solely
on the concentrations from the stations within 5km. Air pollution levels 36 months prior to
enrollment to the study were used in the analyses. Although the majority of the GeneBank
subjects reported residing in Ohio at enrollment, some participants also reported residential
29
addresses in other states. Therefore, air pollution levels were normalized to state means prior to
analysis.
Clinical Definitions. In the UK Biobank, CAD cases were defined using previously described
methods [8] based on International Classification of Diseases version-10 (ICD10) codes
I24.0, I24.8, I24.9, I25.0, I25.1, I25.4, I25.8, and I25.9 for ischemic heart diseases and I21, I22,
I23, I25.2 for myocardial infarction (MI) and complications following acute MI. Additional
criteria to define CAD included Office of Population Censuses and Surveys Classification of
Interventions and Procedures, version 4 (OPCS-4) codes K40-K46, K49, K50 and K75, covering
replacement, transluminal balloon angioplasty, other therapeutic transluminal operations on
coronary artery and percutaneous transluminal balloon angioplasty and insertion of stent into
coronary artery. To avoid any misclassification of atherosclerotic CAD, we also excluded
subjects who were only positive for ICD10 codes I24.1 (Dressler's syndrome), I25.3 (aneurysm
of heart), I25.5 (ischaemic cardiomyopathy), or I25.6 (silent myocardial ischaemia) [8].
Participants with a diagnosis of asthma, defined according to field id 6152_8 (doctor diagnosed
asthma), ICD10 J45 (asthma) or J46 (severe asthma), or self-reported asthma [49], were also
excluded. This strategy identified 26,499 CAD cases and 351,262 controls in the UK Biobank
for whom complete clinical, air pollution exposure, and genotype data were available up to
October 2018. In the GeneBank, CAD was defined at enrollment based on adjudicated
diagnoses of stable or unstable angina, MI (adjudicated definition based on defined
electrocardiographic changes or elevated cardiac enzymes), angiographic evidence of ≥50%
stenosis in ≥1 major epicardial vessel, and/or a history of known CAD (documented MI, CAD,
or history of revascularization), as described previously [8, 40, 44, 45].
30
Statistical Analyses for Air Pollution Exposure and Risk of CAD in UK Biobank. Logistic
regression was used to test the association between four air pollutants (PM2.5, PM10, NO2, and
NOx) individually and risk of CAD in the UK Biobank, with adjustment for age, sex, smoking
status, education level, body mass index (BMI), height, type 2 diabetes (defined based on ICD10
code E11), and the first 20 principal components (PC1-20). Association with CAD was tested
based on a 2-standard deviation (2SD) increase in levels of each pollutant: 2.1μg/m3
for PM2.5;
3.8μg/m3
for PM10; 31.1μg/m3
for NOx; and 15.2μg/m3
for NO2. Association was also tested by
grouping participants into quintiles according to the distribution of the air pollutants and using
logistic regression to calculate odds ratios (ORs) and 95% confidence interval (CI) for subjects in
each quintile compared to the first quintile, with adjustment for age, sex, smoking status,
education level, BMI, height, type 2 diabetes, and PC1-20. P-values were considered significant
at the Bonferroni-corrected threshold for testing four air pollutants (0.05/4=2.5x10-3
). All
analyses were performed with R v4.4.0 (R Core Team, Vienna, Austria).
Genotyping. A candidate gene-based discovery analysis was carried out to identify GxE
interactions using clinical, genetic, and PM2.5 data from both the UK Biobank and GeneBank
cohorts. Quality control of samples, DNA variants and imputation in the UK Biobank were
performed by the Wellcome Trust Centre for Human Genetics [39]. Briefly, ~90 million single
nucleotide polymorphisms (SNPs) imputed from the Haplotype Reference Consortium, UK10K,
and 1000 Genomes imputation were available in 487,164 subjects in the UK Biobank. In the
GeneBank cohort, 1000 Genomes and Haplotype Reference Consortium based imputation was
done separately for 2932 and 1607 samples genotyped on the Affymetrix 6.0 SNP chip or the
Illumina Global Screening Array (GSA), respectively) [8, 45]. Following post-imputation
31
quality control checks, the two GeneBank datasets were merged to create a combined sample of
4,539 subjects and 9,225,854 SNP genotypes with minor allele frequency ≥0.01.
Gene-Environment (GxE) Interaction Analyses. Genotypes for the lead SNPs at 320
previously identified CAD loci were extracted from UK Biobank and GeneBank and tested
individually in each cohort for GxE interactions with PM2.5 levels on risk of CAD using a logistic
regression model that included SNP genotype, PM2.5 levels, and an interaction term (Genotype x
PM2.5). Covariates used in the analyses with the UK Biobank included age, sex, genotyping
array, and PC1-20. For GeneBank, covariates in the model included age, sex, and genotyping
array. Effect sizes from the cohort-specific GxE interaction analyses were then combined
together in a fixed effects meta-analysis using the metagen function in the meta package. Pvalues from the discovery GxE interaction meta-analysis were considered significant at the
Bonferroni-corrected threshold for testing 320 SNPs (0.05/320=1.6x10-4
). For variants
exceeding this significance threshold, a genotype-stratified logistic regression analysis was
carried out to test association of a 2SD increase in PM2.5 levels with risk of CAD separately in
the UK Biobank and GeneBank cohorts. Genotype-specific logistic regression models in the UK
Biobank were adjusted for age, sex, genotyping array, and PC1-20, and for age, sex, and
genotyping array in the GeneBank cohort. This was followed by a genotype-stratified fixedeffects meta-analysis with both cohorts using the metagen function in the meta package. All
analyses were performed with R v4.4.0 (R Core Team, Vienna, Austria). Main effect
associations with CAD for 3,781 SNPs in a 1.1 megabase (Mb) interval at the chromosome 15
locus (chr15:64.3-65.4Mb) were taken from publicly available GWAS summary statistics [10].
GxE interaction analyses for CAD in the same 1.1Mb interval on chromosome 15 were first
carried out separately with 5,644 and 6,346 variants that were available in the UK Biobank and
32
GeneBank cohorts, respectively, using the same analytical approach described above for the 320
known CAD loci. A fixed-effects meta-analysis was carried out with 2,687 SNPs in the
chromosome 15 interval common to both datasets using the metagen function in the meta
package. These results of these analyses were visualized through regional plots generated by
LocusZoom [50] to compare the patterns of association for CAD main effects and GxE
interactions with PM2.5 at the chromosome 15 locus. All analyses were performed with R v4.4.0
(R Core Team, Vienna, Austria).
Effect of Diesel Exhaust on Gene Expression in Endothelial Cells. The effect of air pollution
exposure, in the form of diesel exhaust, on expression of positional candidate genes at the
chromosome 15 locus was first evaluated using publicly available data from a previously
published in vitro study (GSE63095) with human coronary artery endothelial cells (hCAECs)
[51]. Briefly, healthy subjects (n=8) were exposed to whole diesel exhaust for 2hrs with
collection of blood prior to exposure (baseline) and 2 and 24hrs after exposure [51]. Primary
hCAECs (Lonza, Basel, Switzerland) were then cultured for 24hrs in the presence of plasma
obtained at the three exposure timepoints (baseline, 2hrs, and 24hrs). Total RNA was isolated
using RNeasy Mini Prep Kits (Qiagen, Redwood City, CA) and gene expression profiling was
carried out with HumanHT-12 V4 BeadChips (Illumina, San Diego, CA) [51]. After exclusion
of outliers whose values were greater than 2SD outside of the mean, P-values for differences in
expression of positional candidate genes for which data were available at baseline and 2hrs or
24hrs after diesel exhaust exposure in hCAECs were obtained from two-sided t-tests comparing
log2 transformed transcript levels (GraphPad Prism v6.04, Boston, MA). In addition, we also
carried out an in vitro exposure study with a human microvascular endothelial cell line (HMEC-1
cells, CDC, Atlanta, GA) that were cultured with 1% FBS or 5mg/ml diesel exhaust particle
33
(DEP) methanol extract for 4hrs, as described previously [52]. Total RNA was isolated from
HMEC-1 cells using RNeasy kits (Qiagen, Valencia, CA) and profiled on HumanRef-8
Expression BeadChips (Illumina, San Diego, CA) [52]. P-values for differences in expression of
positional candidate genes for which data were available between HMEC-1 cells exposed to
5mg/ml DEP methanol extract for 4hrs and the control group were obtained from two-sided ttests comparing log2 transformed transcript levels (GraphPad Prism v6.04, Boston, MA).
Animal Husbandry. Male and female C57BL/6J mice were purchased from the Jackson
Laboratories (Bar Harbor, Maine). Mice were housed 4-5 per cage at 25ºC on a 12hr dark/12hr
light cycle and maintained on a chow diet (Purina #5053). All methods and procedures were
approved by the Institutional Animal Research Committee of the University of Southern
California and performed in accordance with the relevant guidelines and regulations.
In Vivo DEP Exposure. Mice of both sexes were exposed to aerosolized DEP (100mg/m3
) or
filtered air (FA) for a total of 200 hours (5 hours/day, 5 days/week for 8 weeks) using previously
described procedures [25, 53, 54]. Briefly, standardized DEP obtained from an Industrial forklift
diesel engine (NIST SRM 2975) [55] was purchased from the National Institute of Science and
Technology, sonicated into suspension with ultrapure Milli-Q water (200g/ml) for 30min, and
stored in aliquots at -20°. For exposures, DEP suspensions were re-aerosolized through a HOPE
jet nebulizer (Model 1131; B&B Medical Technologies, Carlsbad, CA) and the aerosolized
stream was mixed with clean HEPA-filtered air (1214; Pall Laboratory, Port Washington, NY).
The re-aerosolized DEP was drawn through a silica gel diffusion dryer (Model 3620, TSI Inc.,
Shoreview, MN) and Po-210 neutralizers (Model 2U500, NRD LLC., Grand Island, NY) to
remove excess water and electrical charges of the particles, respectively. The air stream with re-
34
aerosolized DEP or filtered air alone was then entered into sealed whole-body animal exposure
chambers (52cm long by 32cm wide and 16cm high) at a flow rate of 2.5lpm. Each chamber has
nine separate meshed boxes that could house up to 27 mice (three mice per box) from the same
exposure group. DEP size and concentration were continuously monitored (every five minutes)
via TSI DustTrak (Model 8520, TSI Inc.) in parallel with the exposure chambers. The mean
diameter of the DEP was 65nm with 69% of particles being <100nm in diameter and the
variability in mass concentration was <10% throughout the exposures. The chemical
composition of this form of DEP has been extensively characterized by NIST and we have
previously demonstrated its in vivo biological activity in mouse exposure studies [53, 54].
Effect of DEP Exposure on Gene Expression in the Aorta. Real-time quantitative PCR
(qPCR) was used to evaluate the in vivo effect of DEP on expression of positional candidate
genes at the chromosome 15 locus using previously described methods [56]. At the conclusion
of the exposure studies, mice were euthanized and aortas starting at the arch and extending down
to the iliac bifurcation were harvested and snap frozen in liquid nitrogen. Total RNA was
extracted using RNeasy Mini kits (Qiagen, Valencia, CA) and cDNA was prepared from 500ng
of total RNA using cDNA Archive Reverse Transcription kits (Life Technologies, Gaithersburg,
MD) according to the manufacturer’s protocols. Real-time gene expression reactions was carried
out on an Applied Biosystems TaqMan 7900HT instrument in triplicate for seven positional
candidate genes at the chromosome 15 locus for which TaqMan assays were available: Trip4
(Mm00451187_m1), Rbpms2 (Mm00511032_g1), Snx1 (Mm00444807_m1), Csnk1g1
(Mm00557447_m1), Plekho2 (Mm00462242_m1), Fam96a (Mm01178690_m1), and Zfp609
(Mm00553138_m1). Cycle threshold (Ct) values were determined using RQ Manager 1.2.1 and
relative quantification was calculated using the 2^-ΔΔCt method (SDS 2.4). Transcript levels for
35
each sample were then normalized to Ppia (Mm02342430_g1) as an endogenous control and
calculated relative to a calibrator comprised of a pool of all samples. The replicates for each
individual mouse sample were then averaged and P-values for differences in relative gene
expression levels between filtered air control mice and DEP-exposed animals were obtained
from one-sided t-tests (GraphPad Prism v6.04, Boston, MA).
36
Results
Association of Air Pollution Exposure with Risk of CAD in the UK Biobank. The clinical
characteristics of UK Biobank subjects used for the present analyses are shown in Supplemental
Table 1. A 2SD increase in levels of PM2.5, NO2, and NOX, but not PM10, was significantly
associated with increased risk of CAD in the UK Biobank, of which PM2.5 exhibited the
strongest association (OR=1.11, 95% CI 1.08-1.14, P=1.8x10-12) (Figure 1; Table 1). For
example, subjects in the two highest quintiles of PM2.5 (10.2-21.3μg/m3
) had 12-16% increased
risk of CAD (OR=1.12, 95% CI 1.07-1.17; P=1.8x10-6 for quintile 4; and OR=1.16, 95% CI
1.11-1.22; P=8.0x10-11 for quintile 5) compared to individuals in the first quintile (Figure 1;
Supplemental Table 2). A similar association pattern as a function of quintile was also
observed with NOx levels, although effect sizes for risk of CAD were not as strong as those with
PM2.5 (Figure. 1; Supplemental Table 2). By comparison, risk of CAD was increased to the
same extent (~9%) in quintiles 2-5 of NO2 compared to quintiles 1 (Figure 1; Supplemental
Table 2). Taken together, these observations are consistent with previous environmental
epidemiology studies, including our prior analyses in the GeneBank cohort [48], and validate the
quality of the air pollution exposure data available in the UK Biobank.
Gene-Air Pollution Interaction with a CAD Locus on Chromosome 15. We next carried out
GxE interaction analyses using a candidate gene approach with previously identified CAD
susceptibility loci [5-10]. Since PM2.5 levels exhibited the most robust and dose-dependent
association with risk of CAD in the UK Biobank and in our previous study with the GeneBank
cohort [48], we focused the GxE interaction analyses on this pollutant. GxE interactions for risk
of CAD were first tested with PM2.5 levels and 320 known CAD loci under an additive genetic
37
model separately in the UK Biobank and GeneBank cohorts (Supplemental Tables 1 and 3),
followed by a fixed-effects meta-analysis of the GxE interaction effect estimates. These analyses
revealed 16 loci that exhibited nominally significant GxE interactions (p<0.05) (Supplemental
Table 4). However, the GxE interaction between PM2.5 levels and the lead SNP at a
chromosome 15 locus (rs6494488; A>G) for risk of CAD (P-interaction=1.2x10-4
) exceeded the
Bonferroni-corrected significance threshold for testing 320 variants (P=0.05/320=1.6x10-4
)
(Supplemental Table 4). More specifically, a 2 SD increase in PM2.5 levels increased risk of
CAD by 28% (OR=1.28, 95% CI 1.25-1.32; P=8.4x10-60) among individuals with two copies of
the CAD susceptibility allele (A) and by only 17% (OR=1.17, 95% CI 1.11-1.; P=1.4x10-9
) in
AG heterozygotes (Table 2). By comparison, PM2.5 levels were not associated with increased
risk of CAD in GG homozygotes (OR=1.06, 95% CI (0.92-1.22; P=0.42). Furthermore, the
pattern of association signals at the chromosome 15 locus based on previously reported SNP
main effects for CAD [10] was very similar to the pattern based on effect estimates obtained
from our GxE interaction meta-analyses, where rs6494488 was also the most significantly
associated variant (Figure 2A and B). Rs6494488 is located ~28kb and ~43kb away from the
genes encoding ornithine decarboxylase antizyme 2 (OAZ2) and RNA binding protein with
multiple splicing 2 (RBPMS2), respectively, and is in relatively strong (r2≥0.8) linkage
disequilibrium (LD) with only a few other variants at the chromosome 15 locus (Figure 2).
Bioinformatics Analyses Prioritizes TRIP4 as a Candidate Causal Gene. To prioritize
positional candidate genes and gain insight into the biological mechanisms underlying the
putative GxE interaction at the chromosome 15 locus, we first used publicly available expression
quantitative trait loci (eQTL) databases. Based on the assumption that the effects of air pollution
on atherosclerotic processes at the vascular wall could be mediated through circulating factors,
38
we first explored whether rs6494488 yielded any cis eQTLs in blood. Data from the eQTLGen
Consortium [57] revealed markedly significant eQTLs with rs6494488 in blood for several genes
that localized to a 1Mb interval at this locus, including RBPMS2, TRIP4, ANKDD1A, CSNK1G1,
and SNX1 (Figure 2 and Supplemental Table 5). A subset of these blood eQTLs were observed
in the GTEx Project [58] as well (Supplemental Table 5). However, data from GTEx also
revealed that rs6494488 only yielded eQTLs for TRIP4 in other CAD-relevant tissues, such as
aorta (P=5.7x10-24), coronary artery (P=3.4x10-6
), visceral adipose (P=1.2x10-15), and lung
(P=7.5x10-15), (Figure 3A), but no other gene at this locus. In the STARNET database [59, 60],
nearly identical and directionally consistent eQTLs were also observed for TRIP4, but no other
gene at the chromosome 15 locus, in CAD-relevant tissues such as aorta, adipose, and liver
(Figure 3B). Notably, TRIP4 expression was lower in carriers of the CAD risk allele (A) in all
tissues for which an eQTL was identified (Figure 3A and 3B; Supplemental Table 5).
The genotypic effect of eQTLs observed with rs6494488 and the increased risk of CAD
conferred by the A allele through a GxE interaction and main effects suggested that TRIP4 plays
a protective role in development of CAD at the level of the vessel wall. Consistent with this
notion, data from the STARNET cohort revealed that TRIP4 expression was lower in
atherosclerotic aortas of CAD patients compared to aortas from CAD-free subjects (Figure 3C)
[61]. Based on this observation and that eQTLs for TRIP4 in GTEx and STARNET were
particularly strong in aorta (Figure 3A and 3B), we determined the cell-specific expression
pattern of TRIP4 using single-cell(sc)RNA-seq data derived from atherosclerotic plaques of
human coronary arteries [62] and aortas of hyperlipidemic mice [63]. Visualization of these data
with the PlaqView Single-Cell Portal [64] demonstrated TRIP4 expression to be abundant in
39
subsets of endothelial cells, smooth muscle cells (SMCs), and monocytes/macrophages of
atherosclerotic plaques in both mice and humans (Figure 3D and 3E).
Diesel Exhaust Exposure Decreases TRIP4 Expression. We next reasoned that TRIP4
expression and/or activity could be decreased by air pollution exposure given the inverse clinical
and genetic associations between expression of this gene and risk of CAD. To test this
hypothesis, we first used publicly available data from a short-term in vitro study in which
primary human coronary artery endothelial cells (hCAECs) (GSE63095) [51] were profiled for
global gene expression after being cultured with plasma obtained from healthy subjects who had
undergone a 2hr exposure to air pollution in the form of diesel exhaust (See Methods for details).
This analysis revealed that TRIP4 was the only positional candidate gene at the chromosome 15
locus whose expression was significantly decreased in hCAECs that were cultured with plasma
obtained from subjects two (P=1.9x10-4
) and 24hrs (P=7.9x10-3
) after diesel exhaust exposure
compared to incubation with plasma obtained at baseline (Figure 4A). To validate this
observation, we also carried out a short-term diesel exhaust particle (DEP) exposure experiment
with a human microvascular endothelial cell line (HMEC-1 cells) (GSE6584) [52]. As shown in
Figure 4B, expression of TRIP4, but no other positional candidate gene at the chromosome 15
locus, was also decreased in HMEC-1 cells cultured with DEP extract for 4hrs compared to
control exposure. Finally, we carried out a long-term in vivo exposure study in mice where male
and female animals were exposed to a standardized form of biologically active DEP or filtered
air for 8 weeks [54] (Figure 5A). Consistent with our in vitro observations in endothelial cells,
real-time qPCR analysis of aorta revealed that expression of Trip4, but no other positional gene
at the chromosome 15 locus, was significantly decreased (P=5.9x10-3
) in mice exposed to DEP
compared to the filtered air group (Figure 5B).
40
Discussion
In the present study, we carried out a combination of epidemiological, genetic, and
functional analyses to identify GxE interactions with air pollution for risk of CAD. In the UK
Biobank, higher levels of PM2.5, NOX, and NO2 but not PM10, were associated with increased
risk of CAD, consistent with the well-known adverse relationship between air pollution exposure
and cardiovascular outcomes. A candidate gene approach to further evaluate PM2.5 for GxE
interactions with ~300 known CAD susceptibility loci identified a novel GxE interaction locus
on chromosome 15 where PM2.5 was associated with dose-dependent increased risk of CAD as a
function of the carrying the susceptibility allele (A) of rs6494488. These observations were
bolstered by functional data from previously published in vitro cell culture studies as well as our
own mouse exposure experiments. Taken together, these results point to a novel GxE interaction
at the known CAD locus on chromosome 15 with air pollution exposure. Interestingly, other
recent studies have identified similar GxE interactions between known CAD-associated regions
and smoking or severe COVID-19 [65, 66]. However, our understanding of the overall
contribution of GxE interactions to risk of CAD remains incomplete and requires further
investigation.
Multiple lines of evidence from our analyses converged on TRIP4 was one high priority
candidate causal gene for the GxE interaction identified at the chromosome 15 locus. First,
while eQTLs were observed for several positional candidates in blood, TRIP4 was the only gene
for which rs6494488 yielded directionally consistent eQTLs in cardiometabolic tissues in two
independent datasets (GTEx and STARNET). Second, CAD cases from the STARNET cohort
exhibited decreased TRIP4 expression in atherosclerotic aortic tissue, which is consistent with
41
the decreased expression of TRIP4 in carriers of the CAD risk allele (A) of rs6494488. These
observations were further supported by short-term in vitro cell culture experiments with primary
HCAECs and cultured with plasma from DEP-exposed subjects and immortalized HMEC-1 cells
incubated with DEP extracts, as well as in vivo experiments with mice chronically exposed to
DEP. Notably, all three studies demonstrated that TRIP4 was the only positional candidate gene
at the chromosome 15 locus whose expression was affected by DEP exposure. Furthermore, the
consistency of the results between the two cell culture experiments and the mouse exposure
study, despite the source of DEP differing for each, provide compelling independent
corroborative evidence that decreased expression of TRIP4 in response to DEP exposure
represents a true biological effect. However, it is also important to note that while TRIP4 was
shown to be decreased in vitro specifically in endothelial cells, additional studies will be needed
to determine how DEP exposure affects Trip4 expression in other cell types. Nonetheless, the
collective body of evidence from our study and those from the literature support the notion that
TRIP4 is a putative air pollution-responsive gene that underlies the GxE interaction observed
between the chromosome 15 locus.
As part of our in vivo exposure studies, we also used a standardized form of DEP (SRM
2975), the chemical composition of which has been extensively characterized by NIST, and
which we recently [53, 54], and other groups over the past two decades [55, 67-74], have shown
to be biologically active and have adverse effects under various experimental settings, including
aortic lesion formation in mice [24]. However, we note that the GxE interaction we identified in
humans was based on estimates of exposure to PM2.5, which does not necessarily have the same
chemical composition as DEP or whole diesel exhaust, which is comprised of both gas and
particle components. In this regard, we did recently demonstrate that the standardized DEP used
42
for our mouse exposure experiments are a reliable surrogate for the nanoparticulate fraction of
PM [53]. Moreover, DEP are also rich in particles in the nanosize range, but nanoparticles
collected from the environment can vary by season and may not provide consistent biological
effects even when collected from the same site over time [75]. Lastly, evidence from the
literature has specifically demonstrated that DEP and PM2.5 both promote atherosclerosis and
vascular inflammation in mouse models [20-22, 24, 27, 30, 31, 76].
TRIP4 encodes one of the four subunits of the activating signal cointegrator
ribonucleoprotein complex (ASCC), which is thought to facilitate both transcriptional activation
as well as RNA processing events, including splicing [77]. Loss of function alleles of TRIP4 and
other members of the ASCC result in diverse and rare Mendelian disease traits, such as prenatal
spinal muscular atrophy, congenital bone fractures, and cerebellar hypoplasia [78-80]. However,
the biological mechanisms through which TRIP4 could mediate the effects of air pollutants on
gene expression and/or CAD risk are presently not evident. Interestingly, the ASCC has also
been implicated in DNA damage repair induced by alkylating agents, raising the question of
whether TRIP4 could similarly be involved in protecting against oxidative stress and DNA
damage associated with exposure to air pollutants, such as PM2.5 and DEP [81, 82].
Furthermore, cell-specific expression analyses of atherosclerotic lesions from both humans and
mice revealed that TRIP4 was expressed in multiple cell types known to be involved in
atherogenesis, including endothelial cells, SMCs, and macrophages. Taken together with the
results of the in vitro and in vivo exposure studies, our data suggest that the GxE interaction
identified at the chromosome 15 locus is potentially modulated by putative athero-protective
properties of TRIP4 at the level of the vessel wall, although this hypothesis remains to be proven.
43
While our study offers valuable insight into the complex interplay between genetic
factors and environmental exposure that influence CAD susceptibility, our results should be
taken in the context of certain limitations. For example, the GxE interaction signal at the
chromosome 15 locus with PM2.5 for risk of CAD still requires replication in additional cohorts,
which is challenging since large samples sizes in which genetic, air pollution exposure, and
clinical data are all available will be required. In addition, we focused only on testing known
CAD loci and there are likely additional GxE interactions that exist, including those that could be
identified using more unbiased approaches. In this regard, ApoE alleles that are well-known
genetic risk factors for cardiovascular and neurodegenerative traits have also been shown to
modify association of air pollution with cognitive traits and heart rate variability [83-85].
Furthermore, novel methods have recently been developed for carrying GxE GWAS analyses
using large-scale biobank data [86-88], and could thus be applied to human cohorts, such as UK
Biobank and GeneBank. Finally, of the positional candidate genes at the chromosome 15 locus,
our functional in vitro and in vivo data demonstrated DEP only affected TRIP4 expression in
both human and mouse vascular tissues. However, additional studies will be needed to
determine whether other air pollutants similarly affect TRIP4 expression and to validate whether
direct perturbation of TRIP4 in conjunction with air pollution exposure affects atherosclerosis in
vivo.
In summary, our studies have further elucidated the genetic component of CAD and,
when combined with PM2.5 exposure on a candidate gene basis, identified a novel GxE
interaction that could function through decreased TRIP4 expression in vascular cells that play
important roles in atherogenesis. Future studies will be required to understand the molecular
44
mechanisms underlying the GxE interaction identified at the TRIP4 locus and determining its
translational relevance.
45
References
1. Wienke A, Holm NV, Skytthe A, Yashin AI. The heritability of mortality due to heart
diseases: a correlated frailty model applied to Danish twins. Twin Res. 2001;4(4):266-74.
doi: 10.1375/1369052012399.
2. Zdravkovic S, Wienke A, Pedersen NL, Marenberg ME, Yashin AI, De Faire U. Heritability
of death from coronary heart disease: a 36-year follow-up of 20 966 Swedish twins. J
Intern Med. 2002;252(3):247-54. doi: 10.1046/j.1365-2796.2002.01029.x.
3. Bevan S, Traylor M, Adib-Samii P, Malik R, Paul NL, Jackson C, et al. Genetic heritability of
ischemic stroke and the contribution of previously reported candidate gene and
genomewide associations. Stroke. 2012;43(12):3161-7. doi:
10.1161/STROKEAHA.112.665760.
4. Vinkhuyzen AA, Wray NR, Yang J, Goddard ME, Visscher PM. Estimation and partition of
heritability in human populations using whole-genome analysis methods. Annu Rev
Genet. 2013;47:75-95. doi: 10.1146/annurev-genet-111212-133258.
5. Nikpay M, Goel A, Won HH, Hall LM, Willenborg C, Kanoni S, et al. A comprehensive
1,000 Genomes-based genome-wide association meta-analysis of coronary artery disease.
Nat Genet. 2015;47(10):1121-30. doi: 10.1038/ng.3396.
6. van der Harst P, Verweij N. Identification of 64 novel genetic loci provides an expanded view
on the genetic architecture of coronary artery disease. Circ Res. 2018;122(3):433-43. doi:
10.1161/CIRCRESAHA.117.312086.
7. Koyama S, Ito K, Terao C, Akiyama M, Horikoshi M, Momozawa Y, et al. Populationspecific and trans-ancestry genome-wide analyses identify distinct and shared genetic risk
loci for coronary artery disease. Nat Genet. 2020;52(11):1169-77. doi: 10.1038/s41588-
020-0705-3.
8. Hartiala JA, Han Y, Jia Q, Hilser JR, Huang P, Gukasyan J, et al. Genome-wide analysis
identifies novel susceptibility loci for myocardial infarction. Eur Heart J. 2021;42(9):919-
33. doi: 10.1093/eurheartj/ehaa1040.
9. Tcheandjieu C, Zhu X, Hilliard AT, Clarke SL, Napolioni V, Ma S, et al. Large-scale genomewide association study of coronary artery disease in genetically diverse populations. Nat
Med. 2022;28(8):1679-92. doi: 10.1038/s41591-022-01891-3.
10. Aragam KG, Jiang T, Goel A, Kanoni S, Wolford BN, Atri DS, et al. Discovery and
systematic characterization of risk variants and genes for coronary artery disease in over
a million participants. Nat Genet. 2022;54(12):1803-15. doi: 10.1038/s41588-022-01233-
6.
46
11. Hartiala JA, Hilser JR, Biswas S, Lusis AJ, Allayee H. Gene-environment interactions for
cardiovascular disease. Curr Atheroscler Rep. 2021;23(12):75. doi: 10.1007/s11883-021-
00974-9.
12. Pope CA, 3rd, Turner MC, Burnett RT, Jerrett M, Gapstur SM, Diver WR, et al.
Relationships between fine particulate air pollution, cardiometabolic disorders, and
cardiovascular mortality. Circ Res. 2015;116(1):108-15. doi:
10.1161/CIRCRESAHA.116.305060.
13. Lin Y, Ramanathan G, Zhu Y, Yin F, Rea ND, Lu X, et al. Pro-Oxidative and
Proinflammatory Effects After Traveling From Los Angeles to Beijing: A BiomarkerBased Natural Experiment. Circulation. 2019;140(24):1995-2004. doi:
10.1161/CIRCULATIONAHA.119.042054.
14. Al-Kindi SG, Brook RD, Biswal S, Rajagopalan S. Environmental determinants of
cardiovascular disease: lessons learned from air pollution. Nat Rev Cardiol.
2020;17(10):656-72. doi: 10.1038/s41569-020-0371-2.
15. Kaufman JD, Elkind MSV, Bhatnagar A, Koehler K, Balmes JR, Sidney S, et al. Guidance to
Reduce the Cardiovascular Burden of Ambient Air Pollutants: A Policy Statement From
the American Heart Association. Circulation. 2020;142(23):e432-e47. doi:
10.1161/CIR.0000000000000930.
16. Rajagopalan S, Brauer M, Bhatnagar A, Bhatt DL, Brook JR, Huang W, et al. Personal-Level
Protective Actions Against Particulate Matter Air Pollution Exposure: A Scientific
Statement From the American Heart Association. Circulation. 2020;142(23):e411-e31.
doi: 10.1161/CIR.0000000000000931.
17. Khraishah H, Chen Z, Rajagopalan S. Understanding the cardiovascular and metabolic health
effects of air pollution in the context of cumulative exposomic impacts. Circ Res.
2024;134(9):1083-97. doi: 10.1161/CIRCRESAHA.124.323673.
18. Collaborators GBDRF. Global burden and strength of evidence for 88 risk factors in 204
countries and 811 subnational locations, 1990-2021: a systematic analysis for the Global
Burden of Disease Study 2021. Lancet. 2024;403(10440):2162-203. doi: 10.1016/S0140-
6736(24)00933-4.
19. Brauer M, Casadei B, Harrington RA, Kovacs R, Sliwa K, Group WHFAPE. Taking a stand
against air pollution-the impact on cardiovascular disease: a joint opinion from the World
Heart Federation, American College of Cardiology, American Heart Association, and the
European Society of Cardiology. Circulation. 2021;143(14):e800-e4. doi:
10.1161/CIRCULATIONAHA.120.052666.
20. Sun Q, Wang A, Jin X, Natanzon A, Duquaine D, Brook RD, et al. Long-term air pollution
exposure and acceleration of atherosclerosis and vascular inflammation in an animal
model. JAMA. 2005;294(23):3003-10. doi: 10.1001/jama.294.23.3003.
47
21. Chen LC, Nadziejko C. Effects of subchronic exposures to concentrated ambient particles
(CAPs) in mice. V. CAPs exacerbate aortic plaque development in hyperlipidemic mice.
Inhal Toxicol. 2005;17(4-5):217-24. doi: 10.1080/08958370590912815.
22. Araujo JA, Barajas B, Kleinman M, Wang X, Bennett BJ, Gong KW, et al. Ambient
particulate pollutants in the ultrafine range promote early atherosclerosis and systemic
oxidative stress. Circ Res. 2008;102(5):589-96. doi:
10.1161/CIRCRESAHA.107.164970.
23. Sun Q, Yue P, Deiuliis JA, Lumeng CN, Kampfrath T, Mikolaj MB, et al. Ambient air
pollution exaggerates adipose inflammation and insulin resistance in a mouse model of
diet-induced obesity. Circulation. 2009;119(4):538-46.
24. Miller MR, McLean SG, Duffin R, Lawal AO, Araujo JA, Shaw CA, et al. Diesel exhaust
particulate increases the size and complexity of lesions in atherosclerotic mice. Part Fibre
Toxicol. 2013;10:61. doi: 10.1186/1743-8977-10-61.
25. Woodward NC, Crow AL, Zhang Y, Epstein S, Hartiala J, Johnson R, et al. Exposure to
nanoscale particulate matter from gestation to adulthood impairs metabolic homeostasis
in mice. Sci Rep. 2019;9(1):1816. doi: 10.1038/s41598-018-37704-2.
26. Ruckerl R, Phipps RP, Schneider A, Frampton M, Cyrys J, Oberdorster G, et al. Ultrafine
particles and platelet activation in patients with coronary heart disease--results from a
prospective panel study. Part Fibre Toxicol. 2007;4:1. doi: 10.1186/1743-8977-4-1.
27. Hansen CS, Sheykhzade M, Moller P, Folkmann JK, Amtorp O, Jonassen T, et al. Diesel
exhaust particles induce endothelial dysfunction in apoE-/- mice. Toxicol Appl
Pharmacol. 2007;219(1):24-32. doi: 10.1016/j.taap.2006.10.032.
28. Sun Q, Yue P, Kirk RI, Wang A, Moatti D, Jin X, et al. Ambient air particulate matter
exposure and tissue factor expression in atherosclerosis. Inhal Toxicol. 2008;20(2):127-
37. doi: 10.1080/08958370701821482.
29. Kido T, Bai N, Yatera K, Suzuki H, Meredith A, Mukae H, et al. Diesel exhaust inhalation
induces heat shock protein 70 expression in vivo. Inhal Toxicol. 2011;23(10):593-601.
doi: 10.3109/08958378.2011.595843.
30. Bai N, Kido T, Kavanagh TJ, Kaufman JD, Rosenfeld ME, van Breemen C, et al. Exposure
to diesel exhaust up-regulates iNOS expression in ApoE knockout mice. Toxicol Appl
Pharmacol. 2011;255(2):184-92. doi: 10.1016/j.taap.2011.06.013.
31. Bai N, Tranfield EM, Kavanagh TJ, Kaufman JD, Rosenfeld ME, van Eeden SF. Exposure to
diesel exhaust upregulates COX-2 expression in ApoE knockout mice. Inhal Toxicol.
2012;24(8):518-27. doi: 10.3109/08958378.2012.696221.
32. Steenhof M, Janssen NA, Strak M, Hoek G, Gosens I, Mudway IS, et al. Air pollution
exposure affects circulating white blood cell counts in healthy subjects: the role of
48
particle composition, oxidative potential and gaseous pollutants - the RAPTES project.
Inhal Toxicol. 2014;26(3):141-65. doi: 10.3109/08958378.2013.861884.
33. Pope CA, 3rd, Bhatnagar A, McCracken JP, Abplanalp W, Conklin DJ, O'Toole T. Exposure
to fine particulate air pollution is associated with endothelial injury and systemic
inflammation. Circ Res. 2016;119(11):1204-14. doi:
10.1161/CIRCRESAHA.116.309279.
34. Zanobetti A, Baccarelli A, Schwartz J. Gene-air pollution interaction and cardiovascular
disease: a review. Prog Cardiovasc Dis. 2011;53(5):344-52. doi:
10.1016/j.pcad.2011.01.001.
35. Ward-Caviness CK. A review of gene-by-air pollution interactions for cardiovascular
disease, risk factors, and biomarkers. Hum Genet. 2019;138(6):547-61. doi:
10.1007/s00439-019-02004-w.
36. Peters A, Greven S, Heid IM, Baldari F, Breitner S, Bellander T, et al. Fibrinogen genes
modify the fibrinogen response to ambient particulate matter. Am J Respir Crit Care
Med. 2009;179(6):484-91. doi: 10.1164/rccm.200805-751OC.
37. Panasevich S, Leander K, Ljungman P, Bellander T, de Faire U, Pershagen G, et al.
Interaction between air pollution exposure and genes in relation to levels of inflammatory
markers and risk of myocardial infarction. BMJ Open. 2013;3(9):e003058. doi:
10.1136/bmjopen-2013-003058.
38. Levinsson A, Olin AC, Modig L, Dahgam S, Bjorck L, Rosengren A, et al. Interaction effects
of long-term air pollution exposure and variants in the GSTP1, GSTT1 and GSTCD
genes on risk of acute myocardial infarction and hypertension: a case-control study. PLoS
One. 2014;9(6):e99043. doi: 10.1371/journal.pone.0099043.
39. Bycroft C, Freeman C, Petkova D, Band G, Elliott LT, Sharp K, et al. The UK Biobank
resource with deep phenotyping and genomic data. Nature. 2018;562(7726):203-9. doi:
10.1038/s41586-018-0579-z.
40. Hartiala J, Li D, Conti DV, Vikman S, Patel Y, Tang WH, et al. Genetic contribution of the
leukotriene pathway to coronary artery disease. Hum Genet. 2011;129(6):617-27.
41. Wang Z, Klipfell E, Bennett BJ, Koeth R, Levison BS, Dugar B, et al. Gut flora metabolism
of phosphatidylcholine promotes cardiovascular disease. Nature. 2011;472(7341):57-63.
42. Reiner AP, Hartiala J, Zeller T, Bis JC, Dupuis J, Fornage M, et al. Genome-wide and genecentric analyses of circulating myeloperoxidase levels in the CHARGE and CARe
consortia. Hum Mol Genet. 2013;22(16):3381-93.
43. Hartiala J, Bennett BJ, Tang WH, Wang Z, Stewart AF, Roberts R, et al. Comparative
genome-wide association studies in mice and humans for trimethylamine N-oxide, a
proatherogenic metabolite of choline and L-carnitine. Arterioscler Thromb Vasc Biol.
2014;34(6):1307-13. doi: 10.1161/ATVBAHA.114.303252.
49
44. Hartiala JA, Tang WH, Wang Z, Crow AL, Stewart AF, Roberts R, et al. Genome-wide
association study and targeted metabolomics identifies sex-specific association of CPS1
with coronary artery disease. Nat Commun. 2016;7:10558. doi: 10.1038/ncomms10558.
45. Ferrell M, Wang Z, Anderson JT, Li XS, Witkowski M, DiDonato JA, et al. A terminal
metabolite of niacin promotes vascular inflammation and contributes to cardiovascular
disease risk. Nat Med. 2024;30(2):424-34. doi: 10.1038/s41591-023-02793-8.
46. Eeftens M, Beelen R, de Hoogh K, Bellander T, Cesaroni G, Cirach M, et al. Development of
land use regression models for PM(2.5), PM(2.5) absorbance, PM(10) and PM(coarse) in
20 European study areas; results of the ESCAPE project. Environ Sci Technol.
2012;46(20):11195-205. doi: 10.1021/es301948k.
47. Beelen R, Hoek G, Vienneau D, Eeftens M, Dimakopoulou K, Pedeli X, et al. Development
of NO2 and NOx land use regression models for estimating air pollution exposure in 36
study areas in Europe – The ESCAPE project. Atmos Environ. 2013;72:10-23. doi:
https://doi.org/10.1016/j.atmosenv.2013.02.037.
48. Hartiala J, Breton CV, Tang WH, Lurmann F, Hazen SL, Gilliland FD, et al. Ambient air
pollution is associated with the severity of coronary atherosclerosis and incident
myocardial infarction in patients undergoing elective cardiac evaluation. J Am Heart
Assoc. 2016;5(8). doi: 10.1161/JAHA.116.003947.
49. Han Y, Jia Q, Jahani PS, Hurrell BP, Pan C, Huang P, et al. Genome-wide analysis highlights
contribution of immune system pathways to the genetic architecture of asthma. Nat
Commun. 2020;11(1):1776. doi: 10.1038/s41467-020-15649-3.
50. Pruim RJ, Welch RP, Sanna S, Teslovich TM, Chines PS, Gliedt TP, et al. LocusZoom:
regional visualization of genome-wide association scan results. Bioinformatics.
2010;26(18):2336-7. doi: 10.1093/bioinformatics/btq419.
51. Schisler JC, Ronnebaum SM, Madden M, Channell M, Campen M, Willis MS. Endothelial
inflammatory transcriptional responses to an altered plasma exposome following
inhalation of diesel emissions. Inhal Toxicol. 2015;27(5):272-80. doi:
10.3109/08958378.2015.1030481.
52. Gong KW, Zhao W, Li N, Barajas B, Kleinman M, Sioutas C, et al. Air-pollutant chemicals
and oxidized lipids exhibit genome-wide synergistic effects on endothelial cells. Genome
biology. 2007;8(7):R149. doi: 10.1186/gb-2007-8-7-r149.
53. Shkirkova K, Lamorie-Foote K, Zhang N, Li A, Diaz A, Liu Q, et al. Neurotoxicity of diesel
exhaust particles. J Alzheimers Dis. 2022;89(4):1263-78. doi: 10.3233/JAD-220493.
54. Shkirkova K, Demetriou AN, Sizdahkhani S, Lamorie-Foote K, Zhang H, Morales M, et al.
Microglial TLR4 mediates white matter injury in a combined model of diesel exhaust
exposure and cerebral hypoperfusion. Stroke. 2024;55(4):1090-3. doi:
10.1161/STROKEAHA.124.046412.
50
55. Singh P, DeMarini DM, Dick CA, Tabor DG, Ryan JV, Linak WP, et al. Sample
characterization of automobile and forklift diesel exhaust particles and comparative
pulmonary toxicity in mice. Environ Health Perspect. 2004;112(8):820-5. doi:
10.1289/ehp.6579.
56. Vikman S, Brena RM, Armstrong P, Hartiala J, Stephensen CB, Allayee H. Functional
analysis of 5-Lipoxygenase promoter repeat variants. Hum Mol Genet.
2009;18(23):4521-9.
57. Vosa U, Claringbould A, Westra HJ, Bonder MJ, Deelen P, Zeng B, et al. Large-scale cisand trans-eQTL analyses identify thousands of genetic loci and polygenic scores that
regulate blood gene expression. Nat Genet. 2021;53(9):1300-10. doi: 10.1038/s41588-
021-00913-z.
58. Consortium G. The Genotype-Tissue Expression (GTEx) project. Nat Genet.
2013;45(6):580-5. doi: 10.1038/ng.2653.
59. Franzen O, Ermel R, Cohain A, Akers NK, Di Narzo A, Talukdar HA, et al. Cardiometabolic
risk loci share downstream cis- and trans-gene regulation across tissues and diseases.
Science. 2016;353(6301):827-30. doi: 10.1126/science.aad6970.
60. Koplev S, Seldin M, Sukhavasi K, Ermel R, Pang S, Zeng L, et al. A mechanistic framework
for cardiometabolic and coronary artery diseases. Nat Cardiovasc Res. 2022;1(1):85-100.
doi: 10.1038/s44161-021-00009-1.
61. Mosquera JV, Auguste G, Wong D, Turner AW, Hodonsky CJ, Alvarez-Yela AC, et al.
Integrative single-cell meta-analysis reveals disease-relevant vascular cell states and
markers in human atherosclerosis. Cell Rep. 2023;42(11):113380. doi:
10.1016/j.celrep.2023.113380.
62. Wirka RC, Wagh D, Paik DT, Pjanic M, Nguyen T, Miller CL, et al. Atheroprotective roles
of smooth muscle cell phenotypic modulation and the TCF21 disease gene as revealed by
single-cell analysis. Nat Med. 2019;25(8):1280-9. doi: 10.1038/s41591-019-0512-5.
63. Pan H, Xue C, Auerbach BJ, Fan J, Bashore AC, Cui J, et al. Single-Cell Genomics Reveals
a Novel Cell State During Smooth Muscle Cell Phenotypic Switching and Potential
Therapeutic Targets for Atherosclerosis in Mouse and Human. Circulation.
2020;142(21):2060-75. doi: 10.1161/CIRCULATIONAHA.120.048378.
64. Ma WF, Hodonsky CJ, Turner AW, Wong D, Song Y, Mosquera JV, et al. Enhanced singlecell RNA-seq workflow reveals coronary artery disease cellular cross-talk and candidate
drug targets. Atherosclerosis. 2022;340:12-22. doi:
10.1016/j.atherosclerosis.2021.11.025.
65. Saleheen D, Zhao W, Young R, Nelson CP, Ho W, Ferguson JF, et al. Cardioprotective
effects at the ADAMTS7 locus as a result of gene-smoking interactions. Circulation.
2017;135(24):2336-53. doi: 10.1161/CIRCULATIONAHA.116.022069.
51
66. Hilser JR, Spencer NJ, Afshari K, Gilliland FD, Hu H, Deb A, et al. COVID-19 Is a coronary
artery disease risk equivalent and exhibits a genetic interaction with abo blood type.
Arterioscler Thromb Vasc Biol. 2024. doi: 10.1161/ATVBAHA.124.321001.
67. Farraj AK, Haykal-Coates N, Ledbetter AD, Evansky PA, Gavett SH. Neurotrophin
mediation of allergic airways responses to inhaled diesel particles in mice. Toxicol Sci.
2006;94(1):183-92. doi: 10.1093/toxsci/kfl089.
68. Hartz AM, Bauer B, Block ML, Hong JS, Miller DS. Diesel exhaust particles induce
oxidative stress, proinflammatory signaling, and P-glycoprotein up-regulation at the
blood-brain barrier. FASEB J. 2008;22(8):2723-33. doi: 10.1096/fj.08-106997.
69. Hougaard KS, Jensen KA, Nordly P, Taxvig C, Vogel U, Saber AT, et al. Effects of prenatal
exposure to diesel exhaust particles on postnatal development, behavior, genotoxicity and
inflammation in mice. Part Fibre Toxicol. 2008;5:3. doi: 10.1186/1743-8977-5-3.
70. O'Driscoll CA, Owens LA, Gallo ME, Hoffmann EJ, Afrazi A, Han M, et al. Differential
effects of diesel exhaust particles on T cell differentiation and autoimmune disease. Part
Fibre Toxicol. 2018;15(1):35. doi: 10.1186/s12989-018-0271-3.
71. Bengalli R, Zerboni A, Marchetti S, Longhin E, Priola M, Camatini M, et al. In vitro
pulmonary and vascular effects induced by different diesel exhaust particles. Toxicol
Lett. 2019;306:13-24. doi: 10.1016/j.toxlet.2019.01.017.
72. Shears RK, Jacques LC, Naylor G, Miyashita L, Khandaker S, Lebre F, et al. Exposure to
diesel exhaust particles increases susceptibility to invasive pneumococcal disease. J
Allergy Clin Immunol. 2020;145(4):1272-84 e6. doi: 10.1016/j.jaci.2019.11.039.
73. Liu Y, Wang T, Si B, Du H, Liu Y, Waqas A, et al. Intratracheally instillated diesel PM(2.5)
significantly altered the structure and composition of indigenous murine gut microbiota.
Ecotoxicol Environ Saf. 2021;210:111903. doi: 10.1016/j.ecoenv.2021.111903.
74. Warren CE, Campbell KM, Kirkham MN, Saito ER, Remund NP, Cayabyab KB, et al. The
effect of diesel exhaust particles on adipose tissue mitochondrial function and
inflammatory status. Int J Mol Sci. 2024;25(8). doi: 10.3390/ijms25084322.
75. Zhang H, D'Agostino C, Forman HJ, Cacciottolo M, Thorwald M, Mack WJ, et al. Urban air
pollution nanoparticles from Los Angeles: recently decreased neurotoxicity. J Alzheimers
Dis. 2021;82(1):307-16. doi: 10.3233/JAD-201577.
76. Cassee FR, Campbell A, Boere AJ, McLean SG, Duffin R, Krystek P, et al. The biological
effects of subacute inhalation of diesel exhaust following addition of cerium oxide
nanoparticles in atherosclerosis-prone mice. Environ Res. 2012;115(1):1-10. doi:
10.1016/j.envres.2012.03.004.
77. Jung DJ, Sung HS, Goo YW, Lee HM, Park OK, Jung SY, et al. Novel transcription
coactivator complex containing activating signal cointegrator 1. Mol Cell Biol.
2002;22(14):5203-11. doi: 10.1128/MCB.22.14.5203-5211.2002.
52
78. Knierim E, Hirata H, Wolf NI, Morales-Gonzalez S, Schottmann G, Tanaka Y, et al.
Mutations in subunits of the activating signal cointegrator 1 complex are associated with
prenatal spinal muscular atrophy and congenital bone fractures. Am J Hum Genet.
2016;98(3):473-89. doi: 10.1016/j.ajhg.2016.01.006.
79. Davignon L, Chauveau C, Julien C, Dill C, Duband-Goulet I, Cabet E, et al. The
transcription coactivator ASC-1 is a regulator of skeletal myogenesis, and its deficiency
causes a novel form of congenital muscle disease. Hum Mol Genet. 2016;25(8):1559-73.
doi: 10.1093/hmg/ddw033.
80. Topf A, Pyle A, Griffin H, Matalonga L, Schon K, Solve RDSNViwg, et al. Exome
reanalysis and proteomic profiling identified TRIP4 as a novel cause of cerebellar
hypoplasia and spinal muscular atrophy (PCH1). Eur J Hum Genet. 2021;29(9):1348-53.
doi: 10.1038/s41431-021-00851-8.
81. Risom L, Moller P, Loft S. Oxidative stress-induced DNA damage by particulate air
pollution. Mutat Res. 2005;592(1-2):119-37. doi: 10.1016/j.mrfmmm.2005.06.012.
82. Vilas-Boas V, Chatterjee N, Carvalho A, Alfaro-Moreno E. Particulate matter-induced
oxidative stress - mechanistic insights and antioxidant approaches reported in in vitro
studies. Environ Toxicol Pharmacol. 2024;110:104529. doi: 10.1016/j.etap.2024.104529.
83. Ren C, Baccarelli A, Wilker E, Suh H, Sparrow D, Vokonas P, et al. Lipid and endotheliumrelated genes, ambient particulate matter, and heart rate variability--the VA Normative
Aging Study. J Epidemiol Community Health. 2010;64(1):49-56. doi:
10.1136/jech.2008.083295.
84. Cacciottolo M, Wang X, Driscoll I, Woodward N, Saffari A, Reyes J, et al. Particulate air
pollutants, APOE alleles and their contributions to cognitive impairment in older women
and to amyloidogenesis in experimental models. Transl Psychiatry. 2017;7(1):e1022. doi:
10.1038/tp.2016.280.
85. Finch CE, Thorwald MA. Inhaled pollutants of the gero-exposome and later-life health. J
Gerontol A Biol Sci Med Sci. 2024;79(7). doi: 10.1093/gerona/glae107.
86. Kerin M, Marchini J. Gene-by-environment interactions with a bayesian whole-genome
regression model. Am J Hum Genet. 2020;107(4):698-713. doi:
10.1016/j.ajhg.2020.08.009.
87. Kerin M, Marchini J. A non-linear regression method for estimation of gene-environment
heritability. Bioinformatics. 2020;36(4):5632-9. doi: 10.1093/bioinformatics/btaa1079.
88. Pazokitoroudi A, Liu Z, Dahl A, Zaitlen N, Rosset S, Sankararaman S. A scalable and robust
variance components method reveals insights into the architecture of gene-environment
interactions underlying complex traits. Am J Hum Genet. 2024;111(7):1462-80. doi:
10.1016/j.ajhg.2024.05.015.
53
Acknowledgments
We gratefully acknowledge the UK Biobank Resource for providing access to their data under
Application Number 33307. This study was supported by NIH grants R01HL148110,
R01HL168493, U54HL170326, and Pilot Project Programs of the Southern California
Environmental Health Sciences Center (P30ES007048) and the USC Center for Genetic
Epidemiology. The funders had no role in the design and conduct of the study; collection,
management, analysis, and interpretation of the data; and preparation, review, or approval of the
manuscript.
Data Availability Statement
Individual level data used in the present study are available upon application to the UK Biobank
(https://www.ukbiobank.ac.uk/). All other data supporting the findings of this study are
available either within the article, the Supplementary Information and Supplementary Data files,
or upon reasonable request. Publicly available resources used for this study were the GenotypeTissue Expression Project, http://gtexportal.org/; eQTLGen Consortium,
https://www.eqtlgen.org; STARNET Cohort Browser, http://starnet.mssm.edu/; PlaqView
Single-Cell Portal, https://www.plaqview.com/; NCBI Gene Expression Omnibus,
https://www.ncbi.nlm.nih.gov/geo/
Author Contributions
Concept and design: J.R.H, H.A., and J.A.H. Acquisition, analysis, and interpretation of data:
J.R.H, Y.H., Y.Y., H.Z., C.E.F., W.J.M., G.Z., I.S.H., X.Y., W.H.W.T., C.L.M., L.M., J.L.M.B.,
54
F.D.G., J.A.A., S.L.H., H.A., and J.A.H. Drafting of the manuscript: J.R.H., H.A., and J.A.H.
Critical revision of the manuscript for important intellectual content: All authors.
Competing Interests
Dr. Hazen reports being named as co-inventors on pending and issued patents held by the
Cleveland Clinic relating to cardiovascular diagnostics and therapeutics. Dr. Hazen also reports
having received royalty payments for inventions or discoveries related to cardiovascular
diagnostics or therapeutics from Cleveland Heart Lab, a fully owned subsidiary of Quest
Diagnostics, and Procter & Gamble. Dr. Hazen is a paid consultant for Zehna Therapeutics and
Proctor & Gamble, and has received research funds from Zehna Therapeutics, Proctor &
Gamble, Pfizer, and Roche Diagnostics. Dr. Tang serves as a consultant for Sequana Medical,
Cardiol Therapeutics, Genomics plc, Zehna Therapeutics, Boston Scientific, WhiteSwell,
CardiaTec Biosciences, Bristol Myers Squibb, Alexion Pharmaceuticals, Alleviant Medical,
Salubris Biotherapeutics, BioCardia, and has received honoraria from the American Board of
Internal Medicine, Springer Nature, and Belvoir Media Group. All other authors have no known
competing financial interests or personal relationships to declare related to the work reported in
this paper.
55
Figure Legends
Figure 1. Association of air pollution with risk of CAD in the UK Biobank. The distribution
of fine particulate matter <2.5m (PM2.5) or <10m (PM10), nitrogen dioxide (NO2), and
nitrogen oxides (NOX) in the UK Biobank (A). Increasing levels of PM2.5, NO2, and NOx, but
not PM10, are associated with increased risk of CAD (B). Air pollutant levels in the UK Biobank
are based on land use regression (LUR)-based estimates and shown as quintiles. Analyses were
adjusted for age, sex, smoking, education, BMI, height, type 2 diabetes, and PC1-20.
Figure 2. Regional plots show association patterns for main SNP effects and GxE
interactions at CAD locus on chromosome 15. (A) Rs6494488 is the lead SNP identified for
CAD at the chromosome 15 (indicated by the purple diamond) and yielded a P-value of 1.8x10-9
.
Main effect associations with CAD for 3,781 SNPs in a 1.1 megabase (Mb) region at the
chromosome 15 locus spanning TRIP4 (chr15:64.3-65.4Mb) were taken from publicly available
GWAS summary statistics [10]. (B) A similar association pattern was observed for GxE
interactions at the chromosome 15 locus with rs6494488 being the most significantly associated
variant (P=1.2x10-4
). GxE interaction P-values are based on the results of a fixed-effects metaanalysis in the same 1.1Mb interval with 2,687 SNPs that were available in both the UK Biobank
and GeneBank cohort. For both regional plots, the degree of linkage disequilibrium (LD)
between rs6494488 and other variants in the region is indicated by r2 values according to the
color-coded legend in the top panel, and genomic coordinates and genes located within the
interval are shown in the bottom panel.
56
Figure 3. Multi-tissue eQTL plots for TRIP4. (A and B) The CAD susceptibility allele (A) of
the lead SNP at chromosome 15 (rs6494488) is associated with decreased TRIP4 expression in
each cardiometabolic tissue examined using eQTL data from The GTEx Project and the
STARNET cohort. (C) TRIP4 expression is decreased in atherosclerotic aortic tissue from CAD
cases compared to normal aortic tissue from controls. (D and E) scRNA-seq data from human
coronary plaques and aortic lesions of hyperlipidemic mice shows abundant expression of TRIP4
in subsets of vascular wall cells, including endothelial cells, SMCs, and monocytes/
macrophages. Data for panel C was based on analyses in the STARNET cohort [59, 60] and
publicly available data from previous studies [62, 63] were used for panels D and E.
Figure 4. Effect of DEP on expression of positional candidate genes at chromosome 15 locus
in vitro. (A) Expression of only TRIP4, but none of the eight other positional candidate genes at
the chromosome 15 locus, was significantly downregulated in hCAECs incubated with plasma
from subjects exposed to DEP for 2hrs compared to pre-exposure plasma. Decreased expression
of TRIP4 and CSNK1G1 was also observed after incubation of HCAECs with plasma obtained
24hrs after DEP exposure ended compared to pre-exposure plasma. Expression data are shown as
log2 transformed values and results are based on publicly available data from a previously
published study with n=7-8 subjects per group [51]. (B) Incubation of HMEC-1 cells with 5mg/ml
of DEP extract for 4hrs decreased expression of only TRIP4 but no other positional candidate at
the chromosome 15 locus. Expression data are shown as log2 transformed values from n=2-3
replicates per group [52]. P-values for differences between exposure groups were derived from 2-
sided t-tests. ru, relative units.
57
Figure 5. Effect of DEP on expression of positional candidate genes at chromosome 15 locus
in vivo. (A) A mouse exposure study was carried where male and female C57BL/6J mice were
exposed to either filtered air (as a control) or 100mg/m3 of DEP for 5hrs/day, 5days/wk for 8 weeks
(total of 200hrs). Gene expression analysis was carried out in the aorta for positional candidate
genes at chromosome 15 locus. (B) Expression of Trip4, but no other positional candidate gene at
the chromosome 15 locus, was decreased in aortas of mice chronically exposed to DEP. mRNA
levels of Trip4 and Ppia, as an endogenous control, were quantified by qPCR in triplicate using
pre-designed Taqman gene expression assays with n=7-11 male and female mice in each exposure
group. Significant differences were derived from one-sided t-tests. ru, relative units.
58
Table 1. Association of Four Air Pollutants with Risk of CAD in the UK Biobank.
Pollutant Cases/Controls OR (95% CI) P-value
PM2.5 25,163/338,503 1.11 (1.08-1.14) 1.8x10-12
PM10 28,629/338,503 1.02 (0.99-1.05) 0.18
NO2 27,250/363,959 1.06 (1.03-1.09) 6.7x10-05
NOX 27,250/363,959 1.07 (1.04-1.10) 3.6x10-07
Odds ratios (ORs), 95% confidence intervals (CIs), and P-values were obtained from logistic
regression models testing association of a 2SD increase in levels of indicated air pollutant with
risk of CAD, with adjustment for age, sex, smoking status, education level, BMI, height, type 2
diabetes, and PC1-20.
PM2.5, fine particulate matter <2.5μm in diameter; PM10, fine particulate matter <10μm in
diameter; NO2, nitrogen dioxide; NOX, nitrogen oxides.
59
Table 2. Association of PM2.5 with Risk of CAD as a Function of rs6494488 Genotype in
the UK Biobank and the GeneBank Cohorts.
Genotype Study n OR (95% CI) P-value P-het
AA
UK Biobank 267,891 1.28 (1.24-1.32) 3.7x10-58
GeneBank 2,791 1.33 (1.09-1.64) 5.6x10-03
Meta-analysis 270,682 1.28 (1.25-1.32) 8.4x10-60 0.71
AG
UK Biobank 96,332 1.17 (1.11-1.23) 3.1x10-09
GeneBank 1,097 1.23 (0.90-1.69) 0.20
Meta-analysis 97,429 1.17 (1.11-1.23) 1.4x10-09 0.75
GG
UK Biobank 13,654 1.06 (0.92-1.23) 0.40
GeneBank 102 0.89 (0.27-2.87) 0.84
Meta-analysis 13,756 1.06 (0.92-1.22) 0.42 0.77
Odds ratios (ORs) and 95% confidence intervals (CIs) were derived from logistic regression
models carried out separately in the UK Biobank and GeneBank cohorts stratified by genotype of
rs6494488. Models were based on a 2SD increase in PM2.5 levels and were adjusted for age, sex,
genotyping array, and the first 20 principal components (PCs) in UK Biobank and age, sex, and
genotyping array in the GeneBank cohort. This was followed by a fixed-effects meta-analysis
with both cohorts.
PM2.5, fine particulate matter <2.5μm in diameter
P-het, p-value for heterogeneity for difference in effect size of PM2.5 between UK Biobank and
GeneBank cohorts derived from the meta-analysis.
60
Figure 1.
61
Figure 2.
62
Figure 3.
63
Figure 4.
64
Figure 5.
65
66
67
68
69
70
71
72
73
74
75
CHAPTER 2: COVID-19 IS A CORONARY ARTERY DISEASE RISK EQUIVALENT AND
EXHIBITS A GENETIC INTERACTION WITH ABO BLOOD TYPE
Summary. In the present study using the data from the UK Biobank, we investigated 1) whether
COVID-19 infection increases the risk of thrombotic events and major adverse coronary events
(MACE), 2) whether this risk is equivalent to that conferred by coronary artery disease (CAD),
and 3) whether there is a potential genetic interaction for risk of thrombotic events with severe
COVID-19 and ABO locus, a known myocardial infarction (MI) risk locus. We observed that
COVID-19 acts as a CAD risk equivalent, which has important implications for clinical care,
particularly in risk stratification and long-term management of patients with prior COVID-19
infections. Mechanistic insights suggest that severe COVID-19 infection interacts with blood
group status to create a prothrombotic environment, potentially exacerbating the risk of
thrombotic events. My contributions included data analysis, data curation, manuscript curation
and organizing findings within the large UK Biobank dataset. This study was published in
Arteriosclerosis, Thrombosis, and Vascular Biology (Hilser JR, Spencer NJ, Afshari K, et al.
COVID-19 Is a Coronary Artery Disease Risk Equivalent and Exhibits a Genetic Interaction
With ABO Blood Type. Arterioscler Thromb Vasc Biol. 2024;44(11):2321-2333.
doi:10.1161/ATVBAHA.124.321001).
76
COVID-19 is a Coronary Artery Disease Risk Equivalent and Exhibits a Genetic
Interaction with ABO Blood Type
James R. Hilser, MPH1,2,, Neal J. Spencer, BS1,2,, Kimia Afshari1,2, Frank D. Gilliland, MD,
PhD1
, Howard Hu1
, MD, PhD1
, Arjun Deb, MD3
, Aldons J. Lusis, PhD3,4,5, W.H. Wilson Tang,
MD6,7,8, Jaana A. Hartiala, PhD1
, Stanley L. Hazen, MD, PhD6,7,8*, and Hooman Allayee, PhD1,2*
Departments of 1Population & Public Health Sciences and 2Biochemistry & Molecular Medicine,
Keck School of Medicine, University of Southern California, Los Angeles, CA 90033;
Departments of 3Medicine, 4Human Genetics, and 5Microbiology, Immunology, & Molecular
Genetics, David Geffen School of Medicine of UCLA, Los Angeles, CA 90095. 6Department of
Cardiovascular and Metabolic Sciences, Lerner Research Institute, Cleveland Clinic, Cleveland,
OH 44195; 7Department of Cardiovascular Medicine, Heart, Vascular, and Thoracic Institute,
Cleveland Clinic, Cleveland, OH 44195; 8Center for Microbiome and Human Health, Cleveland
Clinic, Cleveland, OH 44195.
*These authors contributed equally to this work.
Short Title: COVID-19, Adverse Cardiac Risk, and ABO Blood Type
77
Abstract
Background: COVID-19 is associated with acute risk of major adverse cardiac events (MACE),
including myocardial infarction (MI), stroke, and mortality (all-cause). However, the duration
and underlying determinants of heightened risk of cardiovascular disease (CVD) and MACE
post-COVID-19 are not known.
Methods: Data from the UK Biobank was used to identify COVID-19 cases (n=10,005) who
were PCR+
for SARS-CoV-2 infection (n=8,062) or received hospital-based ICD-10 codes for
COVID-19 (n=1,943) between February 1, 2020 and December 31, 2020. Population controls
(n=217,730) and propensity score-matched controls (n=38,860) were also drawn from the UK
Biobank during the same period. Proportional hazards models were used to evaluate COVID-19
for association with long-term (>1000 days) risk of MACE and as a coronary artery disease
(CAD) risk equivalent. Additional analyses examined whether COVID-19 interacted with
genetic determinants to affect risk of MACE and its components.
Results: Risk of MACE was elevated in COVID-19 cases at all levels of severity (HR=2.09,
95% CI 1.94-2.25; P<0.0005) and to a greater extent in cases hospitalized for COVID-19
(HR=3.85, 95% CI 3.51-4.24; P<0.0005). Hospitalization for COVID-19 represented a CAD risk
equivalent since incident MACE risk among cases without history of CVD was even higher than
that observed in CVD patients without COVID-19 (HR=1.21; 95% CI 1.08-1.37; P<0.005). A
significant genetic interaction was observed between the ABO locus and hospitalization for
COVID-19 (P-int=0.01), with risk of thrombotic events being increased in subjects with non-O
blood types (HR=1.65, 95% CI 1.29-2.09; P=4.8x10-5
) to a greater extent than subjects with
blood type O (HR=0.96, 95% CI 0.66-1.39; P=0.82).
78
Conclusions: Hospitalization for COVID-19 represents a CAD risk equivalent, with post-acute
MI and stroke risk particularly heightened in non-O blood types. These results may have
important clinical implications and represent, to our knowledge, one of the first examples of a
gene-pathogen exposure interaction for thrombotic events.
Key Words: COVID-19; SARS-CoV-2; major adverse cardiac events; myocardial infarction;
stroke; thrombosis; genetics; ABO; blood type; gene-pathogen interaction.
79
Introduction
Over the course of four years, the global pandemic caused by the SARS-CoV-2 virus has
led to nearly 800 million confirmed infections and over seven million deaths [1]. One intriguing
aspect of COVID-19, the severe respiratory illness associated with SARS-CoV-2 infection, is the
markedly increased rate of cardiovascular disease (CVD) complications and post-acute risk of
thrombotic events observed in patients, such as myocardial infarction (MI) and stroke. For
example, rates of major adverse cardiac events (MACE=MI, stroke, or all-cause mortality) in
COVID-19 patients are increased immediately after and during the first 30 days after infection as
well as up to two years after infection [2-14]. However, it is not known how long heightened
risk of MACE persists in COVID-19 patients, or what factors modulate this risk. Furthermore,
whether certain subgroups of COVID-19 patients are at greater increased risk of adverse CVD
events is not entirely clear, and whether or not any of these subgroups rise to the level of a
“coronary artery disease (CAD) risk equivalent” has not been explored. In this regard, CAD risk
equivalence has historically been used as a benchmark for escalation of global CVD prevention
efforts, including lowering of lipid goals and initiation of anti-platelet therapies [15-24]. As a
consequence of these and other unanswered questions, regulatory bodies across the world have
yet to adapt preventive CVD guidelines for patients with prior COVID-19 infection. Thus,
determining the duration of increased CVD risk in COVID-19 patients and whether COVID-19
represents a CAD risk equivalent could have important clinical implications for patient care.
Given the variability in CVD outcomes exhibited by patients, it is also reasonable to
assume that increased risk of MACE associated with post-COVID-19 infection can be attributed,
at least in part, to genetic predisposition. For example, large-scale genetic analyses have
80
identified over 20 loci that are associated with susceptibility to SARS-CoV-2 infection and
COVID-19 severity [25], including regions on chromosome 3p21 and the ABO locus on
chromosome 9q34. Interestingly, ABO, which defines A, B, AB, and O blood types, was
previously identified as a genetic susceptibility factor for plaque rupture and MI in the presence
of CAD [26, 27] and one of the first loci to be identified for SARS-CoV-2 infection [28-32].
Thus, it is plausible that COVID-19 may increase risk of MACE through interactions with shared
genetic determinants, such as ABO. However, a link between incident MACE risks among postCOVID-19 subjects and genetic variants associated with COVID-19-related traits or CVD has
not been reported.
In the present study, we leveraged data from the UK Biobank to investigate long-term
risk of MACE among COVID-19 patients and whether COVID-19 rises to the level of a CAD
risk equivalent. We also tested whether association of COVID-19 with risk of MACE exhibited
genetic interactions with known susceptibility variants that were specific to either risk of CVD
outcomes or COVID-19 traits.
81
Materials and Methods
Data Availability. Individual level data used in the present study are available upon application
to the UK Biobank (https://www.ukbiobank.ac.uk/). All other relevant data are either provided
in the manuscript or available upon request from the authors.
Study Population. Between 2006-2010, the UK Biobank recruited a total of 503,325
participants who were 40-69 years of age and registered with a general practitioner of the UK
National Health Service [33]. At enrollment, extensive data on demographics, medication use,
and disease-related outcomes were obtained through questionnaires, health records. Baseline
blood samples were also collected for measurement of serum biomarkers that are either
established disease risk factors or routinely measured as part of clinical evaluations. All UK
Biobank participants provided informed consent. The study protocol was approved by the North
West Multi-Centre Research Ethics Committee and carried out according the Declaration of
Helsinki principles. The present study was approved by the Institutional Review Board of the
USC Keck School of Medicine.
Definition of COVID-19 Cases and Controls. Subsequent to the emergence of SARS-CoV-2
in early 2020, the UK Biobank released results of PCR-based tests for SARS-CoV-2 infection
(Data-Field 40100) among its participants and data on hospital-based International Classification
of Diseases version-10 (ICD10) codes related to COVID-19 (U07.1 or U07.2). Based on the
information provided for the period between February 1, 2020 and December 31, 2020, a total of
11,115 COVID-19 cases were identified (Figure 1). Of these individuals, 2,868 cases had been
assigned an U07.1 or U07.2 ICD10 code and were therefore hospitalized for COVID-19,
whereas the remaining 8,247 cases only received a positive PCR test for SARS-CoV-2 infection.
82
As a comparison group, we defined “population controls” as subjects who were already part of
the UK Biobank hospital inpatient data system but had never received a positive PCR test for
SARS-CoV-2 or were assigned a COVID-19 ICD10 code between February 1, 2020 and
December 31, 2020. Population controls were first assigned a random number using the rand
function with uniform distribution in SAS (v9.4, SAS Institute, Inc, Cary, NC). Controls were
then ranked based on their random number and reiteratively matched 20 times to each of the
11,115 cases. This strategy resulted in 222,300 controls being assigned, at a ratio of 20:1, a start
date that corresponded to the same date that each matched case received a positive PCR test for
SARS-CoV-2 or was assigned an ICD10 code for COVID-19 hospitalization (Figure 1). After
exclusion of individuals who experienced an MI, stroke, or who died within 30 days of entry into
the analysis, complete demographic and clinical data were available in 8,062 cases with only a
positive PCR test for SARS-CoV-2 infection, 1,943 cases who were hospitalized for COVID-19,
and 217,730 randomly assigned population controls (Figure 1, Table 1, and Supplemental
Table 2). We also used propensity score matching as another analytical method to minimize
potential confounding effects of predisposing risk factors among hospitalized COVID-19 cases.
Propensity scores were generated for hospitalized COVID-19 cases and population controls
based on age, sex, education, self-reported ethnicity, diabetes, asthma, smoking, obesity, CVD
status, lipid-lowering medication use, and anti-hypertension medication using the Twang
package in R [34] (v4.3.0, R Core Team, Vienna, Austria). Controls were then matched to cases
at a ratio of 20:1 based on the proximity of their propensity scores using a K-nearest neighbor
algorithm, as implemented in the Matching package in R [35] (v4.3.0, R Core Team, Vienna,
Austria). This strategy yielded a dataset of 1,943 hospitalized COVID-19 cases and a subset of
38,860 propensity score-matched controls.
83
Clinical and Demographic Definitions. CAD was defined as positive for ICD-10 codes I24
and I25 and MI was defined based on ICD10 codes I21, I22, I23, I25.2, as well as doctordiagnosed and self-reported MI, as described previously [27]. Similarly, stroke was defined
based on ICD10 codes I63 and I64, as well as doctor-diagnosed and self-reported stroke. CVD
was defined as the composite of CAD, MI, or stroke. Peripheral artery disease (PAD) was
defined based on ICD-10 codes I730, I731, I738, and I739. Obesity was defined as a BMI ≥30
and diabetes was defined based on ICD10 codes E10, E11, E12, E13, and E14 [36]. Asthma was
defined using previously reported criteria [37] based on ICD10 codes J45 and J46, as well as
doctor-diagnosed or self-reported asthma. Education was categorized based on data provided by
the UK Biobank: A) college/university/nursing/teaching; B) national vocational qualification
(NVQ)/higher national diploma (HND)/higher national certificate (HNC); C) A level; D) O
level/certificate of secondary education (CSE); or E) none of the above/not available. Selfreported ethnicity was assessed using Data Field 21000. Smoking status was defined as
reporting never or ever smoking using Data Field 20116. Use of lipid-lowering medications,
anti-hypertension medications, or anti-platelet agents (aspirin or clopidogrel) at the time of
enrollment in the UK Biobank was based on self-reported data using Data Fields 6153, 6177,
10004, 6154, or 20003 (1141168318 and 1141168322 for clopidogrel) and classified as never or
ever. The numbers of all COVID-19 cases and population controls for whom these data were
available is summarized in Supplemental Table 2.
Time-to-Event Data. Incident major adverse cardiac events (MACE=MI, stroke, or death) were
based on the number of days from the date of receiving a positive PCR test for SARS-CoV-2
infection or an initial ICD10 code assignment for COVID-19 between February 1, 2020 and
December 31, 2020 up to the assignment of an ICD10 code for MI, stroke, or all-cause mortality
84
until October 31, 2022 (up to ~1000 days of follow-up). The causes of death for all COVID-19
cases and population controls based on ICD-10 code data is shown Supplemental Table 3. A
sub-analysis with cardiovascular mortality was also carried out where death could be attributed
to an ICD10 code for Diseases of the Circulatory System (Chapter IX). Control subjects who
were assigned an ICD10 code for COVID-19 or had a positive PCR test for SARS-CoV-2
infection after December 31, 2020 were censored at that date in the time-to-event analyses.
Statistical Analyses. Differences in demographic and clinical characteristics at the time of
enrollment into the UK Biobank between cases and controls were evaluated using chi-square
tests for categorical variables and two-sample t-tests for continuous traits, respectively. Time-toevent analyses were carried out with Cox proportional hazards models to test whether COVID19 was associated with risk of incident thrombosis (MI or stroke), all-cause mortality, or the
composite MACE trait with all three outcomes (MI, stroke, or all-cause mortality) over 1003
days of follow-up from the date of entry into the analysis. Age at the time of COVID-19
diagnosis, sex, self-reported ethnicity, education, diabetes, asthma, smoking status, lipidlowering medication use, and anti-hypertension medication use were included as covariates.
Sensitivity analyses were also carried out in subjects stratified by CVD status, age, sex, obesity,
diabetes, and smoking. Conditional Cox proportional hazards models were used for time-toevent analyses with hospitalized cases and propensity score-matched population controls, with
inclusion of age at the time of COVID-19 diagnosis, sex, self-reported ethnicity, education,
diabetes, asthma, smoking status, lipid-lowering medication use, and anti-hypertension
medication as covariates.
85
Genetic and Gene-Pathogen Interaction Analyses. Details on genotyping arrays, quality
control metrics, and imputation methods used by the UK Biobank have been described
previously [33]. Briefly, ABO blood types were provided by the UK Biobank using Data Field
23165 and were defined based on genotypes of two variants at the ABO locus (rs8176746 and
rs8176719), as described previously [38]. Genotypes of the lead variants at the loci with the
strongest effect sizes for SARS-CoV-2 infection only (rs73062389) or hospitalized COVID-19
(rs11385942) on chromosome 3p21 [25], and for CAD (rs4977574) on chromosome 9p21 [39-
42] were extracted from the imputed data for ~90 million single nucleotide polymorphisms
(SNPs) that were available in the UK Biobank [33] for the same subset of subjects used in the
clinical analyses. Logistic regression models were used to test association of ABO-derived
blood types (A, B, or AB vs. O), rs73062389, rs11385942 with hospitalized COVID-19 cases vs.
SARS-CoV-2
-
controls; hospitalized COVID-19 cases vs. SARS-CoV-2
+
subjects; and SARSCoV-2
+
cases vs. SARS-CoV-2
-
controls, with adjustment for age at the time of COVID-19
diagnosis, sex, ethnicity based the first 10 principal components, as provided by the UK Biobank
in Data Field 22009, and genotyping array. Cox proportional hazards models were used to
assess whether ABO blood type (non-O [AA, AO, BB, BO, or AB] vs. O) or genotypes at
rs73062389, rs11385942, or rs4977574 were associated with incident MI or stroke, with
adjustment for age at the time of ICD-10 code assignment, sex, first 10 principal components,
genotyping array, education, diabetes, asthma, smoking, lipid medication use, and antihypertension medication use. To determine whether association of COVID-19 with risk of MI or
stroke differed as a function of ABO blood type or genotype, an interaction term was included in
the model, with adjustment for the same covariates. Due to the relatively low frequency of the
minor alleles of rs73062389 and rs11385942 (~10% for each) and since previous studies had
86
demonstrated that increased risk of hospitalized COVID-19 was similar in subjects with one or
two copies of the risk allele (A) at rs11385942 [28], dominant models were used for genetic
analyses with these two variants. Analyses with rs4977574 assumed an additive genetic model.
All clinical and statistical genetics analyses were carried out with SAS 9.4 (SAS Institute, Inc,
Cary, NC).
87
Results
Definition and Clinical Characteristics of COVID-19 Cases and Controls. An overview of
the analytical strategy used to identify COVID-19 cases and controls in the UK Biobank is
presented in Figure 1. We first validated our definitions of COVID-19 cases using the strongest
known genetic determinants of SARS-CoV-2 infection (rs73062389) and hospitalization for
COVID-19 (rs11385942) on chromosome 3p21 [25, 28]. Both rs73062389 and rs11385942
yielded associations specifically with SARS-CoV-2 infection and severe COVID-19,
respectively, (Supplemental Table 1) that were directionally consistent with previous reported
effect sizes [25]. Compared to controls, COVID-19 cases at all levels of severity were more
likely to be male, exhibit differences in demographic and clinical characteristics, and have more
CVD-associated co-morbidities (Supplemental Table 2). These differences were even more
pronounced when population controls were compared to hospitalized COVID-19 cases (Table
1). In addition, the most common causes of death among COVID-19 cases and population
controls were due to diseases of the circulatory system, including CVD, followed by diseases of
the respiratory system and neoplasms (Supplemental Table 3).
Association of COVID-19 with Increased Risk of MACE. Consistent with prior studies,
COVID-19 at all levels of severity was associated with significantly higher risk of MI, stroke, or
all-cause mortality over 1003 days of follow-up (HR=2.09, 95% CI 1.94-2.25; P<0.0005)
(Supplemental Figure 1; Supplemental Table 4). Association of COVID-19 with MACE was
even more pronounced among cases requiring hospitalization (HR=3.85, 95% CI 3.51-4.24;
P<0.0005) (Figure 2A; Table 2). To account for unmeasured confounders, we also used a set of
controls matched to hospitalized cases based on propensity scores (Table 1). As shown in Table
88
2 and Figure 2B, hospitalization for COVID-19 was similarly associated with increased risk of
MACE in comparison to propensity score-matched controls (HR=3.65, 95% CI 3.30-4.05;
P<0.0005).
We next carried out a series of sensitivity analyses to evaluate the consistency of the
associations with COVID-19. Notably, risk of MACE in cases at all levels of severity and
particularly in those requiring hospitalization was consistently elevated during the first, second,
or third year after COVID-19 diagnosis (Supplemental Table 5). The effect sizes for
association of COVID-19 at all levels of severity with risk of MACE and its components were
also comparable between groups, except in men and younger subjects (Supplemental Tables 6-
8). Among hospitalized cases and all population controls, similar differential risk of MACE and
its components was also observed in men, younger subjects, and individuals with pre-existing
obesity or diabetes (Supplemental Tables 9-11). Moreover, hospitalization for COVID-19 was
associated with increased risk of mortality attributable to CVD (Supplemental Table 12).
Collectively, these observations demonstrate that risk of MACE among COVID-19 patients was
elevated out to nearly three years after SARS-CoV-2 infection regardless of the presence or
absence of co-associated CVD risk factors.
Hospitalization for COVID-19 is a CAD (CVD) Risk Equivalent. We next evaluated whether
COVID-19 represented a CAD risk equivalent. Among participants who remained COVID-19
negative throughout the follow-up period, risk of MACE was increased in subjects with diabetes
(HR=1.88; 95% CI 1.73-2.04; P<0.0005), peripheral artery disease (PAD) (HR=5.08; 95% CI
4.62-5.59; P<0.0005), or CVD (HR=5.63; 95% CI 5.36-5.92; P<0.0005) (Figure 3 and Table 3).
However, risk of MACE among hospitalized COVID-19 cases without a history of CVD was
89
also increased (HR=7.04; 95% CI 6.25-7.92; P<0.0005) compared to COVID-19 negative
controls without any of the selected CAD equivalents (Figure 3 and Table 3). More
specifically, hospitalized COVID-19 cases without a history of CVD had ~20% increased risk of
MACE compared to COVID-19 negative subjects with CVD (HR=1.21; 95% CI 1.08-1.37;
P<0.005).
Given the observation that COVID-19 represented a CAD equivalent, we next explored
whether thrombotic risk could be modulated by the use of anti-platelet agents. Thrombotic risk
remained elevated among primary prevention patients without known CVD who were
hospitalized for COVID-19 and not on anti-platelet agents at the time of enrollment in the UK
Biobank (HR=1.98; 95% CI 1.39-2.82; P<0.0005) (Supplemental Table 13). By comparison,
risk of MI or stroke was not significantly increased among primary prevention COVID-19
patients who reported taking anti-platelet agents (Supplemental Table 13). Taken together,
these results demonstrate that, in the context of our UK Biobank dataset, hospitalization for
COVID-19 increased risk of MACE among primary prevention subjects to the same degree as in
COVID-19 negative subjects with pre-existing CAD equivalent risk factors. Furthermore, out
data suggest that thrombotic risk can potentially be mitigated by use of anti-platelet agents.
Hospitalization for COVID-19 Increases Risk of Thrombotic Events through Genetic
Interaction with ABO Blood Type. We next used a candidate gene approach to test whether
increased risk of thrombosis post-COVID-19 could be affected through interactions with genetic
susceptibility factors. Based on its identification as a specific susceptibility locus for MI and
stroke [26, 27, 43] and SARS-CoV-2 infection (but not with hospitalized COVID-19) [25, 28],
we focused our initial analyses on the ABO locus. In our UK Biobank dataset, non-O blood
90
types increased the likelihood of testing positive for SARS-CoV-2 infection (OR=1.16, 95% CI
1.11-1.22; P=2.3x10-10) but did not increase risk of being hospitalized for COVID-19
(Supplemental Table 1). These results are consistent with prior studies demonstrating that nonO blood types are specifically associated with increased susceptibility to being infected with
SARS-CoV-2 but not necessarily with risk of developing its severe post-infection respiratory
complications [44]. We next tested the hypothesis that incident risk of MI and stroke among
hospitalized COVID-19 patients could exhibit a gene-pathogen interaction with ABO blood
types. Consistent with this notion, a significant genetic interaction was observed between
COVID-19 and ABO (P-int=0.011) where hospitalization for COVID-19 increased risk of MI
and stroke to a greater extent in subjects with non-O blood types (HR=1.65, 95% CI 1.29-2.09;
P=4.8x10-5
) than individuals with blood type O (HR=0.96, 95% CI 0.66-1.39; P=0.82) (Figure 4
and Table 4).
To explore whether the gene-pathogen interaction between COVID-19 and ABO was
indirectly due to the association of non-O blood types with COVID-19 traits, we also tested the
same genetic determinants of SARS-CoV-2 infection susceptibility (rs73062389) and severe
COVID-19 (rs11385942) used to verify our definition of COVID-19 cases (Supplemental Table
1). However, there was no evidence that hospitalization for COVID-19 increased risk of MI and
stroke through gene-pathogen interactions with rs73062389 (P-interaction=0.81) or rs11385942
(P-interaction=0.71) (Supplemental Table 14). Finally, we applied the same approach to assess
whether increased risk of thrombosis among hospitalized COVID-19 patients could result from
interactions with genetic susceptibility factors for atherosclerotic CAD. Of the ~300 known
CAD loci [41, 42], we tested the lead SNP at chromosome 9p21 (rs4977574) since it is one of
the most strongly associated loci for CAD itself but not associated specifically with MI or stroke
91
in patients with CAD [26, 45, 46]. Similar to rs73062389 and rs11385942, there was no
evidence (P-interaction=0.96) that thrombosis risk among hospitalized COVID-19 cases differed
as a function of genotype at the chromosome 9p21 locus (Supplemental Table 13).
Collectively, these observations suggest that increased thrombosis risk among hospitalized
COVID-19 patients differs as a function of ABO blood type but not through interactions with
genetic determinants of COVID-19 severity, SARS-CoV-2 infection susceptibility, or
atherosclerotic CAD.
92
Discussion
In the present study, we leveraged the UK Biobank resource to demonstrate that patients
all levels COVID-19 severity (from simply being PCR+
for SARS-CoV-2 to those requiring
hospitalization) are at significantly increased long-term risk of MACE. These associations were
observed amongst subjects stratified by one year intervals after developing COVID-19 or the
presence of CVD, diabetes, obesity, and other comorbidities. Furthermore, the magnitude of the
effect sizes for MACE and its components were consistently more pronounced among COVID19 cases requiring hospitalization and comparable to those reported in two of the largest
published studies for the same outcomes out to two years [6, 14]. Notably, effect sizes for
increased risk of MACE as a result of hospitalization for COVID-19 were similar in analyses
using propensity score-matched controls. Taken together, our data indicate that the elevated risk
of MACE in COVID-19 patients shows no apparent signs of attenuation up to nearly three years
after SARS-CoV-2 infection and suggest that COVID-19 continues to pose a significant public
health burden with lingering adverse cardiovascular risk.
A major finding from our analyses was that risk of MACE among the subset of
hospitalized COVID-19 cases without known CVD (i.e. primary prevention patients) was
comparable to (or even slightly higher than) the risk in patients with CVD, PAD, or diabetes but
without COVID-19. These observations argue that primary prevention subjects who developed
severe COVID-19 carry the same risk as a CAD (CVD) risk equivalent for as long as follow-up
data are available. Clinical guidelines across the world (US, EU, Australia, Asia) employ “CAD
risk equivalence” as a metric to escalate CVD risk reduction as part of primary prevention
efforts, including both lowering of LDL cholesterol goals and initiation of anti-platelet agents
93
[15-23]. For example, subjects with any form of CVD or CAD risk equivalent (e.g. diabetes,
PAD), are routinely treated with heightened preventive treatment goals comparable to those used
for subjects with pre-existing CVD (as if the subject just experienced an MI). Thus, our findings
raise the question of whether hospitalized COVID-19 infection should also be given
consideration as a CAD risk equivalent, which would invoke a discussion of altering preventive
cardiovascular practice in patients previously hospitalized with COVID-19. In this regard,
retrospective observational studies have suggested that antiplatelet therapy reduces risk of
mortality [47, 48]. However, randomized controls trials with antiplatelet therapy among
COVID-19 patients individually did not demonstrate therapeutic benefits [49-53], although a
meta-analysis of data from these same trials did provide evidence for reduced risk of thrombotic
events [54]. It is important to note that the antiplatelet therapy in these clinical trials was
administered during the acute phase of COVID-19 infection in a hospital setting and for
relatively short durations (follow-up of 90 days or less) compared to the observational studies in
which historical use of aspirin and other agents were considered. Thus, whether anti-platelet
therapy provides cardiovascular benefit in subjects who developed severe COVID-19 will need
to be evaluated more thoroughly in larger numbers of subjects and with prospective study
designs of longer duration.
Systemic or localized infections increase the risk of thrombosis ~2-20-fold and are
independent risk factors for thromboembolic diseases, such as deep vein thrombosis, pulmonary
embolism, MI and stroke [55-60]. Since the emergence of SARS-CoV-2, these clinical
associations have extended to COVID-19 as well [61], including the results presented herein.
However, our analyses also revealed that increased thrombosis risk due to COVID-19 differed as
a function of ABO blood type, which represents, at least to our knowledge, one of the first
94
examples of a gene-pathogen exposure interaction for CVD-related outcomes. Specifically,
hospitalization for COVID-19 increased risk of MI or stroke by ~2-fold among patients with
non-O blood types but not in patients with blood type O. Given that non-O blood types comprise
~60% of the global population, our results would indicate that a substantial fraction of
individuals in the world who developed COVID-19 are at increased risk for thrombosis.
Consistent with our data, a retrospective observational analysis also reported increased risk of
MACE in a small number of COVID-19 patients with blood type A versus O [62]. However,
this study only reported outcomes over the first 30 days following SARS-Cov-2 infection, was
underpowered to detect associations with other non-O blood types, and did not test for other
genetic interactions.
Our results also raise important questions with respect to the biological mechanism(s)
through which COVID-19 interacts with host genetic factors. For example, incident thrombotic
events in COVID-19 patients with non-O blood types may have been increased because of
known associations between ABO and COVID-19 traits and/or MI or stroke in the presence of
CAD. In our UK Biobank dataset, ABO was specifically associated with increased susceptibility
to infection by SARS-CoV-2 but not development of hospitalized COVID-19 itself. While our
findings are consistent with those from large-scale genetic analyses [25], association of ABO
with COVID-19 severity has not been uniformly observed in all studies [63, 64]. With respect to
the two variants identified for COVID-19 at the chromosome 3p21 locus, one increases
susceptibility to SARS-CoV-2 infection (rs73062389) whereas the other increases risk of
becoming hospitalized from COVID-19 (rs11385942) [65]. Interestingly, both variants are part
of a 50kb haplotype that was introgressed into the human genome from Neanderthals [66].
However, rs73062389 and rs11385942 are not in linkage disequilibrium with each other (r2=0.0)
95
[67] and thus reflect independent association signals for COVID-19-related traits. This concept
is further supported by functional studies implicating SLC6A20 and CXCR6 as two candidate
causal genes at 3p21 that could mediate SARS-CoV-2 infection and development of severe
COVID-19, respectively [68]. However, we did not obtain evidence for risk of thrombotic
events differing as a result of genetic interactions between COVID-19 and the two variants at the
chromosome 3p21 locus or with the lead variant (rs4977574) at one of the strongest genetic
determinants of atherosclerotic CAD on 9p21 [26, 45]. Taken together, these observations
suggest that interactions between COVID-19 and increased risk of MI and/or stroke may be
specific to genetic factors that influence risk of plaque rupture and thrombus formation (ABO),
but not those that directly increase risk of hospitalization for COVID-19 (3p21), susceptibility to
SARS-CoV-2 infection (3p21), or that promote atherosclerosis (9p21). It should also be noted
that ABO is one of the most pleiotropic loci in the genome and exhibits associations with
numerous cardiometabolic traits [69], including coagulation biomarkers [70]. Thus, increased
thrombotic risk as a result of genetic interaction between ABO and COVID-19 could be due to
synergistic effects of non-O blood types and SARS-CoV-2 infection at the level of the vessel
wall that potentially destabilizes vulnerable plaques and/or renders the endothelium more prone
to thrombus formation.
While our results point to interesting clinical and genetic associations with COVID-19,
we also note certain limitations of our study. First, defining cases and appropriate controls using
data from the UK Biobank requires special consideration, given how the original SARS-CoV-2
strain and its virulence have evolved over time, improvements in patient care since the beginning
of the pandemic, and development of vaccines in early 2021. For these reasons, and due to the
unavailability of complete data on vaccination dates or the SARS-CoV-2 strain that UK Biobank
96
subjects were infected with, we only used cases who developed COVID-19 prior to the
availability of COVID-19 vaccines. Second, it is also possible that severe COVID-19 patients
may have required hospitalization because of underlying undiagnosed CVD, which we were not
able to ascertain given the nature of the UK Biobank as a population-based cohort. Third,
information on medication use in the UK Biobank was not specific to the beginning of the
pandemic in 2020 or the date of being infected with SARS-CoV-2. Furthermore, the number of
subjects on anti-platelet agents may have also limited our power to evaluate the effects of these
medications on thrombosis risk, particularly among primary prevention subjects. Despite these
limitations, the consistency of the results obtained within the framework of our study design,
coupled with extensive prior published evidence that COVID-19 and ABO are each associated
with CVD outcomes, suggest that the clinical and genetic relationships we observed with MACE
outcomes represent true associations. However, additional studies in larger and non-European
ancestry populations will be needed to determine whether risk of MACE among COVID-19
patients remains elevated beyond three years and to replicate interactions with host genetic
factors.
In summary, our findings demonstrate that development of COVID-19 requiring
hospitalization confers a CAD risk equivalent to the subject, with heightened risk of thrombosis
particularly evident among cases with non-O blood types. These observations suggest that more
aggressive cardiovascular risk reduction efforts may be warranted as part of primary prevention
in patients hospitalized for COVID-19 and provide new avenues for understanding the biological
mechanisms underlying CVD-related adverse outcomes of severe SARS-CoV-2 infection.
97
References
1. Dong E, Du H, Gardner L. An interactive web-based dashboard to track COVID-19 in real
time. Lancet Infect Dis. 2020;20(5):533-4. doi: 10.1016/S1473-3099(20)30120-1.
2. Giacca M, Shah AM. The pathological maelstrom of COVID-19 and cardiovascular disease.
Nat Cardiovasc Res. 2022;1:200-10.
3. Aleksova A, Fluca AL, Gagno G, Pierri A, Padoan L, Derin A, et al. Long-term effect of
SARS-CoV-2 infection on cardiovascular outcomes and all-cause mortality. Life Sci.
2022;310:121018. doi: 10.1016/j.lfs.2022.121018.
4. Nappi F, Nappi P, Gambardella I, Avtaar Singh SS. Thromboembolic disease and cardiac
thrombotic complication in COVID-19: a systematic review. Metabolites. 2022;12(10).
doi: 10.3390/metabo12100889.
5. Xie Y, Xu E, Bowe B, Al-Aly Z. Long-term cardiovascular outcomes of COVID-19. Nat Med.
2022;28(3):583-90. doi: 10.1038/s41591-022-01689-3.
6. Knight R, Walker V, Ip S, Cooper JA, Bolton T, Keene S, et al. Association of COVID-19
with major arterial and venous thrombotic diseases: a population-wide cohort study of 48
million adults in England and Wales. Circulation. 2022;146(12):892-906. doi:
10.1161/CIRCULATIONAHA.122.060785.
7. Xu E, Xie Y, Al-Aly Z. Long-term neurologic outcomes of COVID-19. Nat Med.
2022;28:pages2406–15. doi: 10.1038/s41591-022-02001-z.
8. Kumar A, Narayan RK, Kumari C, Faiq MA, Kulandhasamy M, Kant K, et al. SARS-CoV-2
cell entry receptor ACE2 mediated endothelial dysfunction leads to vascular thrombosis
in COVID-19 patients. Med Hypotheses. 2020;145:110320. doi:
10.1016/j.mehy.2020.110320.
9. Long B, Brady WJ, Koyfman A, Gottlieb M. Cardiovascular complications in COVID-19. Am
J Emerg Med. 2020;38(7):1504-7. doi: 10.1016/j.ajem.2020.04.048.
10. Zheng YY, Ma YT, Zhang JY, Xie X. COVID-19 and the cardiovascular system. Nat Rev
Cardiol. 2020;17(5):259-60. doi: 10.1038/s41569-020-0360-5.
11. Bansal M. Cardiovascular disease and COVID-19. Diabetes Metab Syndr. 2020;14(3):247-
50. doi: 10.1016/j.dsx.2020.03.013.
12. Nishiga M, Wang DW, Han Y, Lewis DB, Wu JC. COVID-19 and cardiovascular disease:
from basic mechanisms to clinical perspectives. Nat Rev Cardiol. 2020;17(9):543-58.
doi: 10.1038/s41569-020-0413-9.
98
13. Loffi M, Piccolo R, Regazzoni V, Di Tano G, Moschini L, Robba D, et al. Coronary artery
disease in patients hospitalised with Coronavirus disease 2019 (COVID-19) infection.
Open Heart. 2020;7(2). doi: 10.1136/openhrt-2020-001428.
14. Bowe B, Xie Y, Al-Aly Z. Postacute sequelae of COVID-19 at 2 years. Nat Med.
2023;29(9):2347-57. doi: 10.1038/s41591-023-02521-2.
15. Banks E, Crouch SR, Korda RJ, Stavreski B, Page K, Thurber KA, et al. Absolute risk of
cardiovascular disease events, and blood pressure- and lipid-lowering therapy in
Australia. Med J Aust. 2016;204(8):320. doi: 10.5694/mja15.01004.
16. Chew DP, Scott IA, Cullen L, French JK, Briffa TG, Tideman PA, et al. National Heart
Foundation of Australia & Cardiac Society of Australia and New Zealand: Australian
clinical guidelines for the management of acute coronary syndromes 2016. Heart Lung
Circ. 2016;25(9):895-951. doi: 10.1016/j.hlc.2016.06.789.
17. Grundy SM, Stone NJ, Bailey AL, Beam C, Birtcher KK, Blumenthal RS, et al. 2018
AHA/ACC/AACVPR/AAPA/ABC/ACPM/ADA/AGS/APhA/ASPC/NLA/PCNA
guideline on the management of blood cholesterol: executive summary: A report of the
American College of Cardiology/American Heart Association task force on clinical
practice guidelines. Circulation. 2019;139(25):e1046-e81. doi:
10.1161/CIR.0000000000000624.
18. Arnett DK, Blumenthal RS, Albert MA, Buroker AB, Goldberger ZD, Hahn EJ, et al. 2019
ACC/AHA guideline on the primary prevention of cardiovascular disease: executive
summary: A report of the american college of cardiology/american heart association task
force on clinical practice guidelines. Circulation. 2019;140(11):e563-e95. doi:
10.1161/CIR.0000000000000677.
19. Pearson GJ, Thanassoulis G, Anderson TJ, Barry AR, Couture P, Dayan N, et al. 2021
Canadian Cardiovascular Society Guidelines for the Management of Dyslipidemia for the
Prevention of Cardiovascular Disease in Adults. Can J Cardiol. 2021;37(8):1129-50. doi:
10.1016/j.cjca.2021.03.016.
20. Visseren FLJ, Mach F, Smulders YM, Carballo D, Koskinas KC, Back M, et al. 2021 ESC
Guidelines on cardiovascular disease prevention in clinical practice. Eur Heart J.
2021;42(34):3227-337. doi: 10.1093/eurheartj/ehab484.
21. Tan JWC, Chew DP, Brieger D, Eikelboom J, Montalescot G, Ako J, et al. 2020 Asian
pacific society of cardiology consensus recommendations on antithrombotic management
for high-risk chronic coronary syndrome. Eur Cardiol. 2021;16:e26. doi:
10.15420/ecr.2020.45.
22. Koh N, Ference BA, Nicholls SJ, Navar AM, Chew DP, Kostner K, et al. Asian pacific
society of cardiology consensus recommendations on dyslipidaemia. Eur Cardiol.
2021;16:e54. doi: 10.15420/ecr.2021.36.
99
23. ElSayed NA, Aleppo G, Aroda VR, Bannuru RR, Brown FM, Bruemmer D, et al. 10.
Cardiovascular disease and risk management: Standards of care in diabetes - 2023.
Diabetes Care. 2023;46(Suppl 1):S158-S90. doi: 10.2337/dc23-S010.
24. Haffner SM, Lehto S, Ronnemaa T, Pyorala K, Laakso M. Mortality from coronary heart
disease in subjects with type 2 diabetes and in nondiabetic subjects with and without prior
myocardial infarction. N Engl J Med. 1998;339(4):229-34. doi:
10.1056/NEJM199807233390404.
25. Initiative C-HG. A second update on mapping the human genetic architecture of COVID-19.
Nature. 2023;621(7977):E7-E26. doi: 10.1038/s41586-023-06355-3.
26. Reilly MP, Li M, He J, Ferguson JF, Stylianou IM, Mehta NN, et al. Identification of
ADAMTS7 as a novel locus for coronary atherosclerosis and association of ABO with
myocardial infarction in the presence of coronary atherosclerosis: two genome-wide
association studies. Lancet. 2011;377(9763):383-92.
27. Hartiala JA, Han Y, Jia Q, Hilser JR, Huang P, Gukasyan J, et al. Genome-wide analysis
identifies novel susceptibility loci for myocardial infarction. Eur Heart J. 2021;42(9):919-
33. doi: 10.1093/eurheartj/ehaa1040.
28. Severe Covid GG, Ellinghaus D, Degenhardt F, Bujanda L, Buti M, Albillos A, et al.
Genomewide association study of severe Covid-19 with respiratory failure. N Engl J
Med. 2020;383(16):1522-34. doi: 10.1056/NEJMoa2020283.
29. Goker H, Aladag Karakulak E, Demiroglu H, Ayaz Ceylan CM, Buyukasik Y, Inkaya AC, et
al. The effects of blood group types on the risk of COVID-19 infection and its clinical
outcome. Turk J Med Sci. 2020;50(4):679-83. doi: 10.3906/sag-2005-395.
30. Solmaz I, Arac S. ABO blood groups in COVID-19 patients; Cross-sectional study. Int J Clin
Pract. 2021;75(4):e13927. doi: 10.1111/ijcp.13927.
31. Goel R, Bloch EM, Pirenne F, Al-Riyami AZ, Crowe E, Dau L, et al. ABO blood group and
COVID-19: a review on behalf of the ISBT COVID-19 Working Group. Vox Sang.
2021;116(8):849-61. doi: 10.1111/vox.13076.
32. Kim Y, Latz CA, DeCarlo CS, Lee S, Png CYM, Kibrik P, et al. Relationship between blood
type and outcomes following COVID-19 infection. Semin Vasc Surg. 2021;34(3):125-31.
doi: 10.1053/j.semvascsurg.2021.05.005.
33. Bycroft C, Freeman C, Petkova D, Band G, Elliott LT, Sharp K, et al. The UK Biobank
resource with deep phenotyping and genomic data. Nature. 2018;562(7726):203-9. doi:
10.1038/s41586-018-0579-z.
34. McCaffrey DF, Ridgeway G, Morral AR. Propensity score estimation with boosted
regression for evaluating causal effects in observational studies. Psychol Methods.
2004;9(4):403-25. doi: 10.1037/1082-989X.9.4.403.
100
35. Sekhon JS. Multivariate and Propensity Score Matching Software with Automated Balance
Optimization: The Matching package for R. Journal of Statistical Software. 2011;42(7):1
- 52. doi: 10.18637/jss.v042.i07.
36. Hilser JR, Han Y, Biswas S, Gukasyan J, Cai Z, Zhu R, et al. Association of serum HDLcholesterol and apolipoprotein A1 levels with risk of severe SARS-CoV-2 infection. J
Lipid Res. 2021;62:100061. doi: 10.1016/j.jlr.2021.100061.
37. Han Y, Jia Q, Jahani PS, Hurrell BP, Pan C, Huang P, et al. Genome-wide analysis highlights
contribution of immune system pathways to the genetic architecture of asthma. Nat
Commun. 2020;11(1):1776. doi: 10.1038/s41467-020-15649-3.
38. Groot HE, Villegas Sierra LE, Said MA, Lipsic E, Karper JC, van der Harst P. Genetically
determined ABO blood group and its associations with health and disease. Arterioscler
Thromb Vasc Biol. 2020;40(3):830-8. doi: 10.1161/ATVBAHA.119.313658.
39. McPherson R, Pertsemlidis A, Kavaslar N, Stewart A, Roberts R, Cox DR, et al. A common
allele on chromosome 9 associated with coronary heart disease. Science.
2007;316(5830):1488-91. doi: 10.1126/science.1142447.
40. Helgadottir A, Thorleifsson G, Manolescu A, Gretarsdottir S, Blondal T, Jonasdottir A, et al.
A common variant on chromosome 9p21 affects the risk of myocardial infarction.
Science. 2007;316(5830):1491-3. doi: 10.1126/science.1142842.
41. Tcheandjieu C, Zhu X, Hilliard AT, Clarke SL, Napolioni V, Ma S, et al. Large-scale
genome-wide association study of coronary artery disease in genetically diverse
populations. Nat Med. 2022;28(8):1679-92. doi: 10.1038/s41591-022-01891-3.
42. Aragam KG, Jiang T, Goel A, Kanoni S, Wolford BN, Atri DS, et al. Discovery and
systematic characterization of risk variants and genes for coronary artery disease in over
a million participants. Nat Genet. 2022;54(12):1803-15. doi: 10.1038/s41588-022-01233-
6.
43. Mishra A, Malik R, Hachiya T, Jurgenson T, Namba S, Posner DC, et al. Stroke genetics
informs drug discovery and risk prediction across ancestries. Nature.
2022;611(7934):115-23. doi: 10.1038/s41586-022-05165-3.
44. Pereira E, Felipe S, de Freitas R, Araujo V, Soares P, Ribeiro J, et al. ABO blood group and
link to COVID-19: A comprehensive review of the reported associations and their
possible underlying mechanisms. Microb Pathog. 2022;169:105658. doi:
10.1016/j.micpath.2022.105658.
45. Beutner F, Teupser D, Gielen S, Holdt LM, Scholz M, Boudriot E, et al. Rationale and
design of the Leipzig (LIFE) Heart Study: phenotyping and cardiovascular characteristics
of patients with coronary artery disease. PLoS One. 2011;6(12):e29070. doi:
10.1371/journal.pone.0029070.
101
46. Patel RS, Schmidt AF, Tragante V, McCubrey RO, Holmes MV, Howe LJ, et al. Association
of chromosome 9p21 with subsequent coronary heart disease events. Circ Genom Precis
Med. 2019;12(4):e002471. doi: 10.1161/CIRCGEN.119.002471.
47. Zong X, Wang X, Liu Y, Li Z, Wang W, Wei D, et al. Antiplatelet therapy for patients with
COVID-19: Systematic review and meta-analysis of observational studies and
randomized controlled trials. Frontiers in medicine. 2022;9:965790. doi:
10.3389/fmed.2022.965790.
48. Sokolski M, Reszka K, Adamik B, Kilis-Pstrusinska K, Lis W, Pomorski M, et al.
Antiplatelet therapy prior to COVID-19 infection impacts on patients mortality: a
propensity score-matched cohort study. Sci Rep. 2024;14(1):4832. doi: 10.1038/s41598-
024-55407-9.
49. Connors JM, Brooks MM, Sciurba FC, Krishnan JA, Bledsoe JR, Kindzelski A, et al. Effect
of Antithrombotic Therapy on Clinical Outcomes in Outpatients With Clinically Stable
Symptomatic COVID-19: The ACTIV-4B Randomized Clinical Trial. JAMA.
2021;326(17):1703-12. doi: 10.1001/jama.2021.17272.
50. Investigators R-CWCftR-C, Bradbury CA, Lawler PR, Stanworth SJ, McVerry BJ,
McQuilten Z, et al. Effect of Antiplatelet Therapy on Survival and Organ Support-Free
Days in Critically Ill Patients With COVID-19: A Randomized Clinical Trial. JAMA.
2022;327(13):1247-59. doi: 10.1001/jama.2022.2910.
51. Group RC. Aspirin in patients admitted to hospital with COVID-19 (RECOVERY): a
randomised, controlled, open-label, platform trial. Lancet. 2022;399(10320):143-51. doi:
10.1016/S0140-6736(21)01825-0.
52. Berger JS, Kornblith LZ, Gong MN, Reynolds HR, Cushman M, Cheng Y, et al. Effect of
P2Y12 Inhibitors on Survival Free of Organ Support Among Non-Critically Ill
Hospitalized Patients With COVID-19: A Randomized Clinical Trial. JAMA.
2022;327(3):227-36. doi: 10.1001/jama.2021.23605.
53. Bohula EA, Berg DD, Lopes MS, Connors JM, Babar I, Barnett CF, et al. Anticoagulation
and Antiplatelet Therapy for Prevention of Venous and Arterial Thrombotic Events in
Critically Ill Patients With COVID-19: COVID-PACT. Circulation. 2022;146(18):1344-
56. doi: 10.1161/CIRCULATIONAHA.122.061533.
54. Duo H, Jin M, Yang Y, Baheti R, Feng Y, Fu Z, et al. Effect of antiplatelet therapy after
COVID-19 diagnosis: A systematic review with meta-analysis and trial sequential
analysis. PLoS One. 2024;19(2):e0297628. doi: 10.1371/journal.pone.0297628.
55. Kinasewitz GT, Yan SB, Basson B, Comp P, Russell JA, Cariou A, et al. Universal changes
in biomarkers of coagulation and inflammation occur in patients with severe sepsis,
regardless of causative micro-organism [ISRCTN74215569]. Crit Care. 2004;8(2):R82-
90. doi: 10.1186/cc2459.
102
56. Beristain-Covarrubias N, Perez-Toledo M, Thomas MR, Henderson IR, Watson SP,
Cunningham AF. Understanding infection-induced thrombosis: lessons learned from
animal models. Front Immunol. 2019;10:2569. doi: 10.3389/fimmu.2019.02569.
57. Smeeth L, Thomas SL, Hall AJ, Hubbard R, Farrington P, Vallance P. Risk of myocardial
infarction and stroke after acute infection or vaccination. N Engl J Med.
2004;351(25):2611-8. doi: 10.1056/NEJMoa041747.
58. Smeeth L, Cook C, Thomas S, Hall AJ, Hubbard R, Vallance P. Risk of deep vein
thrombosis and pulmonary embolism after acute infection in a community setting.
Lancet. 2006;367(9516):1075-9. doi: 10.1016/S0140-6736(06)68474-2.
59. Clayton TC, Gaskin M, Meade TW. Recent respiratory infection and risk of venous
thromboembolism: case-control study through a general practice database. Int J
Epidemiol. 2011;40(3):819-27. doi: 10.1093/ije/dyr012.
60. Dalager-Pedersen M, Sogaard M, Schonheyder HC, Nielsen H, Thomsen RW. Risk for
myocardial infarction and stroke after community-acquired bacteremia: a 20-year
population-based cohort study. Circulation. 2014;129(13):1387-96. doi:
10.1161/CIRCULATIONAHA.113.006699.
61. Varga Z, Flammer AJ, Steiger P, Haberecker M, Andermatt R, Zinkernagel AS, et al.
Endothelial cell infection and endotheliitis in COVID-19. Lancet.
2020;395(10234):1417-8. doi: 10.1016/S0140-6736(20)30937-5.
62. Nauffal V, Achanta A, Goldhaber SZ, Piazza G. Association of ABO blood group type with
cardiovascular events in COVID-19. J Thromb Thrombolysis. 2021;51(3):584-6. doi:
10.1007/s11239-020-02364-5.
63. Zietz M, Zucker J, Tatonetti NP. Associations between blood type and COVID-19 infection,
intubation, and death. Nat Commun. 2020;11(1):5761. doi: 10.1038/s41467-020-19623-
x.
64. Latz CA, DeCarlo C, Boitano L, Png CYM, Patell R, Conrad MF, et al. Blood type and
outcomes in patients with COVID-19. Ann Hematol. 2020;99(9):2113-8. doi:
10.1007/s00277-020-04169-1.
65. Velavan TP, Pallerla SR, Ruter J, Augustin Y, Kremsner PG, Krishna S, et al. Host genetic
factors determining COVID-19 susceptibility and severity. EBioMedicine.
2021;72:103629. doi: 10.1016/j.ebiom.2021.103629.
66. Zeberg H, Paabo S. The major genetic risk factor for severe COVID-19 is inherited from
Neanderthals. Nature. 2020;587(7835):610-2. doi: 10.1038/s41586-020-2818-3.
67. Myers TA, Chanock SJ, Machiela MJ. LDlinkR: An R package for rapidly calculating
linkage disequilibrium statistics in diverse populations. Front Genet. 2020;11:157. doi:
10.3389/fgene.2020.00157.
103
68. Kasela S, Daniloski Z, Bollepalli S, Jordan TX, tenOever BR, Sanjana NE, et al. Integrative
approach identifies SLC6A20 and CXCR6 as putative causal genes for the COVID-19
GWAS signal in the 3p21.31 locus. Genome biology. 2021;22(1):242. doi:
10.1186/s13059-021-02454-4.
69. Buniello A, MacArthur JAL, Cerezo M, Harris LW, Hayhurst J, Malangone C, et al. The
NHGRI-EBI GWAS Catalog of published genome-wide association studies, targeted
arrays and summary statistics 2019. Nucleic Acids Res. 2019;47(D1):D1005-D12. doi:
10.1093/nar/gky1120.
70. Vasan SK, Rostgaard K, Majeed A, Ullum H, Titlestad KE, Pedersen OB, et al. ABO blood
group and risk of thromboembolic and arterial disease: A study of 1.5 million blood
donors. Circulation. 2016;133(15):1449-57. doi:
10.1161/CIRCULATIONAHA.115.017563.
104
Acknowledgments
We gratefully acknowledge the UK Biobank Resource for providing access to their data under
Application Number 33307.
Sources of Funding
This work was supported, in part, by NIH Grants R01HL148110, R01HL168493,
U54HL170326, R01DK132735, P01HL147823, and R01HL147883. The sponsors had no role in
study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Disclosures
Dr. Hazen reports being named as co-inventors on pending and issued patents held by the
Cleveland Clinic relating to cardiovascular diagnostics and therapeutics. Dr. Hazen also reports
having received royalty payments for inventions or discoveries related to cardiovascular
diagnostics or therapeutics from Cleveland Heart Lab, a fully owned subsidiary of Quest
Diagnostics, and Procter & Gamble. Dr. Hazen is a paid consultant for Zehna Therapeutics and
Proctor & Gamble, and has received research funds from Zehna Therapeutics, Proctor &
Gamble, Pfizer, and Roche Diagnostics. All other authors have no competing financial interests
or personal relationships to declare related to the work reported in this paper.
Author Contributions
Concept and design: J.R.H., J.A.H., S.L.H., and H.A. Acquisition, analysis, and interpretation of
data: J.R.H., N.J.S., K.A., F.D.G., H.H., A.D., A.J.L., W.H.W.T., J.A.H., S.L.H., and H.A.
105
Drafting of manuscript: J.R.H., S.L.H., and H.A. Critical revision of the manuscript for
important intellectual content: All authors.
106
Figure Legends
Graphical Abstract. SARS-CoV-2 infection and progression to hospitalization for COVID-19
increased risk of incident major adverse cardiac events (MACE), including myocardial infarction
(MI), stroke, and all-cause mortality, and represented a coronary artery disease risk equivalent.
Risk of thrombosis among hospitalized cases and controls also exhibited a genetic interaction
with ABO blood type and was increased in subjects with non-O blood types (A, B, or AB) to a
greater extent (thick arrow) than subjects with blood type O (thin arrow).
Figure 1. Overview of Clinical and Genetic Analyses. A study was designed with 416,588 UK
Biobank subjects who had hospital in-patient data and were alive on February 1, 2020 (pink
box). COVID-19 cases at all levels of severity were defined as subjects who had either a
positive PCR test for SARS-CoV-2 infection or received a hospital-based ICD-10 code for
COVID-19 through October 31, 2022 (light green box). “Severe” cases were defined as the
subset of subjects who were hospitalized for COVID-19 (dark green inset box). Population
controls were defined as subjects alive on February 1, 2020 and who did not have a positive PCR
test for SARS-CoV-2 infection or who had ever been assigned a hospital-based ICD-10 code for
COVID-19 through December 31, 2020 (yellow box). Controls were then randomly assigned an
enrollment date based on the start dates of all COVID-19 case and matched to cases at a ratio of
20:1. After exclusion of subjects with thrombotic events or death within 30 days of the date of
entry into analysis (gray box), 10,005 COVID-19 cases and 217,730 population controls (green
and yellow boxes) were used to evaluate association of COVID-19 with major adverse cardiac
events (MACE), defined as myocardial infarction (MI), stroke, or all-cause mortality, up until
107
October 31, 2022 (orange box). Gene-pathogen exposure interactions on risk of thrombotic
events were carried out with previously identified genetic variants (blue box).
Figure 2. Hospitalization for COVID-19 is Associated with Increased Risk of MACE. (A)
Among hospitalized cases and all population controls, COVID-19 increased cumulative
incidence of MACE (MI, stroke, or all-cause mortality) in both subjects with and without CVD.
(B) COVID-19 similarly increased cumulative incidence of MACE when comparing hospitalized
cases to propensity score-matched population controls.
Figure 3. COVID-19 Represents a CAD (CVD) Risk Equivalent. Cumulative incidence of
MACE in hospitalized COVID-19 cases without CVD (red line) was equivalent to that observed
in all population controls with CVD (purple line) or PAD (pink line), and even greater than in
population controls with diabetes (green line). Incidence of MACE was highest among
hospitalized COVID-19 cases with CVD (blue line).
Figure 4. Hospitalization for COVID-19 Increases Risk of Thrombotic Events Through a
Genetic Interaction with ABO Blood Group. COVID-19 increased cumulative incidence of
MI and stroke to a greater extent among hospitalized cases with non-O blood types than blood
type O, leading to a significant gene-pathogen exposure interaction (P-interaction=0.01). No
differences in rate of thrombotic events was observed as a function of ABO blood type among all
population controls.
108
Table 1. Clinical Characteristics of Hospitalized COVID-19 Cases and Population Controls.
Trait
Cases
(n=1,943)
All Controls
(n=217,730)
aP-value Propensity Score-matched
Controls (n=38,860)
bP-value
Age at start of follow-up 70.4 (8.1) 68.4 (8.1) <0.001 70.6 (8.0) 0.32
Male/Female 1,069/874 96,403/121,327 <0.001 21,730/17,130 0.44
Ethnicity <0.001 1.00
White 1,750 (90.1) 205,378 (94.3) 35,068 (90.2)
Asian 65 (3.4) 4,222 (1.9) 1,331 (3.4)
Black 75 (3.9) 3,541 (1.6) 1,390 (3.6)
Chinese 8 (0.4) 706 (0.3) 160 (0.4)
Mixed 11 (0.6) 1,246 (0.6) 234 (0.6)
Other 24 (1.2) 1,920 (0.9) 481 (1.2)
Unknown 10 (0.5) 717 (0.3) 196 (0.5)
Education <0.001 0.59
College/University/Nursing/Teaching
(A)
694 (35.7) 102,402 (47.0) 13,471 (34.7)
NVQ/HND/HNC (Vocational
qualification) (B) 242 (12.5) 27,947 (12.8) 4,705 (12.1)
A level (High School) (C) 79 (4.1) 11,793 (5.4) 1,610 (4.1)
O level/CSE (Less than High
School) (D) 300 (15.4) 36,854 (16.9) 5,816 (15.0)
None of the above/Not available
(E) 628 (32.3) 38,734 (17.8) 13,258 (34.1)
Ever smoker 1,069 (55.0) 96,301 (44.2) <0.001 21,685 (55.8) 0.50
History of MI, stroke, or CAD (CVD) 392 (20.2) 25,236 (11.6) <0.001 8,321 (21.4) 0.19
Diabetes 498 (25.6) 20,433 (9.4) <0.001 10,660 (27.4) 0.08
Asthma 427 (22.0) 32,197 (14.8) <0.001 8,707 (22.4) 0.65
Lipid-lowering medication use 579 (29.8) 41,698 (19.2) <0.001 12,289 (31.6) 0.09
Anti-hypertension medication use 633 (32.6) 47,863 (22.0) <0.001 13,252 (34.1) 0.17
Data are shown as mean (SD) for age and numbers (%) for categorical traits.
109
P-values for comparisons between cases and controls were derived from a two-sample t-test for age and chi-square tests for categorical traits.
Matching of controls was done using propensity scores generated using all variables shown in the Table.
aFor comparisons between cases and all controls.
bFor comparisons between cases and propensity score-matched controls.
NVQ/HND/HNC, national vocational qualification/higher national diploma/higher national certificate
CSE, certificate of secondary education
110
Table 2. Association of Hospitalization for COVID-19 with Increased Risk of MACE.
Outcome
Outcome Among
Hospitalized Cases
(Yes/No)
Outcome Among All
Population Controls
(Yes/No)
Outcome Among
Propensity Scorematched Matched
Controls (Yes/No) aHR (95% CI) bHR (95% CI)
Thrombotic
events
All subjects 102/1,841 (5.3%) 5,466/212,264 (2.5%) 1,813/37,047
(4.7%)
**1.37 (1.13-1.67) *
1.24 (1.01-1.53)
Without CVD 38/1,513 (2.5%) 1,836/190,658 (1.0%) 428/30,111 (1.4%) ***1.85 (1.34-2.55) **1.76 (1.23-2.50)
With CVD 64/328 (16.3%) 3,630/21,606 (14.4%) 1,385/6,936
(16.6%)
1.08 (0.85-1.39) 1.08 (0.82-1.43)
All-cause
mortality
All subjects 395/1,548 (20.3%) 4,297/213,433 (2.0%) 1,278/37,582
(3.3%)
***7.09 (6.38-7.87) ***6.99 (6.19-7.90)
Without CVD 284/1,267 (18.3%) 3,055/189,439 (1.6%) 726/29,813 (2.4%) ***8.30 (7.33-9.39) ***9.22 (7.85-10.8)
With CVD 111/281 (28.3%) 1,242/23,994 (4.9%) 552/7,769 (6.6%) ***4.87 (4.00-5.93) ***5.64 (4.31-7.38)
MACE
All subjects 463/1,480 (23.8%) 9,183/208,547 (4.2%) 2,870/35,990
(7.4%)
***3.85 (3.51-4.24) ***3.65 (3.30-4.05)
Without CVD 307/1,244 (19.8%) 4,672/187,822 (2.4%) 1,092/29,447
(3.6%)
***5.93 (5.28-6.67) ***6.37 (5.52-7.34)
With CVD 156/236 (39.8%) 4,511/20,725 (17.9%) 1,778/6,543
(21.4%)
***2.08 (1.78-2.45) ***2.18 (1.80-2.66)
Data are shown as hazard ratios (HR) and 95% confidence intervals (CI) for association of hospitalized COVID-19 with risk of thrombotic events
(MI or stroke), all-cause mortality, and MACE (MI, stroke, or all-cause mortality).
aHRs and 95% CIs from analyses with all population controls were derived from Cox proportional hazards models adjusted for age, sex, ethnicity,
education, diabetes, asthma, smoking, lipid-lowering medication use, and anti-hypertension medication use.
111
bHRs and 95% CIs from analyses with propensity score-matched controls were derived from conditional Cox proportional hazards models with
adjustment for age, sex, ethnicity, education, diabetes, asthma, smoking, lipid-lowering medication use, and anti-hypertension medication use.
*P<0.05; **P<0.005; ***P<0.0005.
112
Table 3. Hospitalization for COVID-19 is a CAD (CVD) Risk Equivalent.
Hospitalized
COVID-19
CAD Risk
Equivalent Group MACE (Yes/No) HR (95% CI)
No Without CVD 3,817/172,334 (2.2%) 1
No With Diabetes 756/13,984 (5.1%) ***1.88 (1.73-2.04)
No With PAD 545/2,704 (16.8%) ***5.08 (4.62-5.59)
No With CVD 4,511/20,725 (17.9%) ***5.63 (5.36-5.92)
Yes Without CVD 307/1,244 (19.8%) ***7.04 (6.25-7.92)
Yes With CVD 156/236 (39.8%) ***12.2 (10.3-14.3)
Data are shown as hazard ratios (HR) and 95% confidence intervals (CI) for association of hospitalized
COVID-19 and/or indicated CAD equivalent group with risk of MACE (MI, stroke, or all-cause
mortality).
HRs and 95% CIs were derived from Cox proportional hazards models adjusted for age, sex, ethnicity,
education, asthma, smoking, lipid-lowering medication use, and anti-hypertension medication use.
Reference group is population control subjects who did not have cardiovascular disease (CVD) and who
remained SARS-CoV-2 and COVID-19 negative over the entire follow-up period.
PAD, peripheral artery disease.
***P<0.0005.
113
Table 4. Hospitalization for COVID-19 Increases Risk of Thrombotic Events Through Genetic Interaction with ABO Blood Type.
Outcome
ABO Blood
Type
Outcome Among
Hospitalized Cases
(Yes/No)
Outcome Among All
Population Controls
(Yes/No) HR (95% CI) P-value aP-interaction
Thromboti
c events
O 29/758 (3.7%) 2,329/90,332 (2.5%) 0.96 (0.66-1.39) 0.82 0.01
Non-O 69/1,026 (6.3%) 3,023/117,066 (2.5%) 1.65 (1.29-2.09) 4.8x10-5
All-cause
mortality
O 152/635 (19.3%) 1,786/90,875 (1.9%) 6.74 (5.69-7.97) 1.5x10-109 0.56
Non-O 229/866 (20.9%) 2,409/117,680 (2.0%) 7.25 (6.31-8.32) 3.1x10-173
MACE O 173/614 (22.0%) 3,880/ 88,781 (4.2%) 3.51 (3.01-4.10) 5.8x10-58 0.11
Non-O 275/820 (25.1%) 5,101/114,988 (4.3%) 4.07 (3.60-4.60) 2.7x10-111
Data are shown as hazard ratios (HR) and 95% confidence intervals (CI) for association of hospitalization for COVID-19 with risk of thrombotic
events (MI or stroke), all-cause mortality, and MACE (MI, stroke, or all-cause mortality) stratified by ABO blood group.
HRs and 95% CIs were derived from Cox proportional hazards models adjusted for age, sex, first 10 principal components, genotyping array,
education, diabetes, asthma, smoking, lipid-lowering medication use, and anti-hypertension medication use.
aP-values for interaction were obtained from models that included an interaction term between COVID-19 status and ABO blood type (Non-O [A,
B, or AB] vs. O).
114
Graphical Abstract.
115
Figure 1.
116
Figure 2.
117
Figure 3.
118
Figure 4.
119
Figure S1. Association of COVID-19 with Risk of MACE. Among cases at all levels of
severity and all population controls, COVID-19 increased cumulative incidence of MACE (MI,
stroke, or all-cause mortality) over 1003 days of follow-up in both subjects with and without
CVD.
120
Table S1. Validation of COVID-19 Case Status and Severity Through Comparative Association Patterns of Genetic Variants.
Cases PCR+
for SARS-CoV-2 vs. Controls
PCRfor SARS-CoV-2
Hospitalized COVID-19 Cases vs.
Controls PCRfor SARS-CoV-2
Hospitalized COVID-19 Cases vs. Cases
PCR+
for SARS-CoV-2
Genetic
Variant/Factor EA/OA
Cases/
Controls
OR
(95% CI) P-value
Cases/
Controls
OR
(95% CI) P-value
Cases/
Controls
OR
(95% CI) P-value
rs73062389 A/G
7,854/
212,678
1.24
(1.16-1.32)
5.6x10-11
1,882/
212,678
1.16
(1.02-1.32)
0.028
1,882/
7,854
0.93
(0.80-1.08)
0.35
a
rs11385942 A/-
7,825/
211,942
0.93
(0.87-
0.997)
0.041
1,876/
211,942
1.33
(1.18-1.50)
2.6x10-6
1,876/
7,825
1.49
(1.29-1.72)
6.8x10-8
ABO blood type bNon-O/O
7,865/
212,750
1.16
(1.11-1.22)
2.3x10-10
1,882/
212,750
1.08
(0.99-1.19)
0.091
1,882/
7,865
0.94
(0.85-1.05)
0.29
Data are shown as odds ratios (OR) and 95% confidence intervals (CI) for association of indicated variants/blood types with being PCR+
for
SARS-CoV-2 infection or hospitalized due to COVID-19.
ORs and 95% CIs were derived from logistic regression models adjusted for age, sex, PC1-10, and genotyping array.
ORs (95% CIs) and P-values from analyses with rs73062389 and rs11385942 were obtained assuming dominant genetic models.
EA/OA, effect allele/other allele.
a
rs11385942 is an insertion (A) variant.
bBlood types A, B, and AB.
121
121
Table S2. Clinical Characteristics of All COVID-19 Cases and Population Controls.
Trait
All Cases
(n=10,005)
All Controls
(n=217,730)
Pvalue
Age at start of follow-up 65.3 (8.6) 68.4 (8.1) <0.001
Male/Female 4,658/5,347 96,403/121,327 <0.001
Ethnicity <0.001
White 9,069 (90.6) 205,378 (94.3)
Asian 392 (3.9) 4,222 (1.9)
Black 261 (2.6) 3,541 (1.6)
Chinese 22 (0.2) 706 (0.3)
Mixed 80 (0.8) 1,246 (0.6)
Other 146 (1.5) 1,920 (0.9)
Unknown 35 (0.4) 717 (0.3)
Education <0.001
College/University/Nursing/Teachin
g (A)
4,017 (40.2) 102,402 (47.0)
NVQ/HND/HNC (Vocational
qualification) (B) 1,590 (15.9) 27,947 (12.8)
A level (High School) (C) 491 (4.9) 11,793 (5.4)
O level/CSE (Less than High
School) (D) 1,917 (19.2) 36,854 (16.9)
None of the above/Not available
(E) 1,990 (19.9) 38,734 (17.8)
Ever smoker 4,609 (46.1) 96,301 (44.2) <0.001
History of MI, stroke, or CAD
(CVD) 1,147 (11.5) 25,236 (11.6) <0.001
Diabetes 1,233 (12.3) 20,433 (9.4) <0.001
Asthma 1,705 (17.0) 32,197 (14.8) <0.001
Lipid-lowering medication use 1,760 (17.6) 41,698 (19.2) <0.001
Anti-hypertension medication use 1,967 (19.7) 47,863 (22.0) <0.001
Data are shown as mean (SD) for age and numbers (%) for categorical traits.
P-values for comparisons between cases and controls were derived from a two-sample t-test for age and
chi-square tests for categorical traits.
122
Table S3. Causes of Death Among All COVID-19 Cases and Population Controls Based on ICD-10 Codes.
Disease Category (ICD-10 Codes)
All Cases (%)
N=582
All Controls (%)
N=4,297
Total (%)
N=4,879
Chapter I – Infectious and parasitic diseases (A and B) 28 (4.8) 188 (4.4) 216 (4.4)
Chapter II – Neoplasms (C and D) 205 (35.2) 1979 (46.1) 2184 (44.8)
Chapter III – Diseases of the blood and blood-forming organs and certain
disorders involving the immune mechanism (D) 5 (0.9) 68 (1.6) 73 (1.5)
Chapter IV – Endocrine, nutritional and metabolic diseases (E) 74 (12.7) 472 (11.0) 546 (11.2)
Chapter V – Mental and behavioral disorders (F) 86 (14.8) 322 (7.5) 408 (8.4)
Chapter VI – Diseases of the nervous system (G) 120 (20.6) 531 (12.4) 651 (13.3)
Chapter VII – Diseases of the eye and adnexa (H) 0.0 (0) 2 (0.1) 2 (0.04)
Chapter VIII – Diseases of the ear and mastoid process (H) 0.0 (0) 2 (0.1) 2 (0.04)
Chapter IX – Diseases of the circulatory system (I) 226 (38.8) 1840 (42.8) 2066 (42.3)
Chapter X – Diseases of the respiratory system (J) 210 (36.1) 1133 (26.4) 1343 (27.5)
Chapter XI – Diseases of the digestive system (K) 40 (6.9) 365 (8.5) 405 (8.3)
Chapter XII – Diseases of the skin and subcutaneous tissue (L) 6 (1.0) 36 (0.8) 42 (0.9)
Chapter XIII – Diseases of the musculoskeletal system and connective tissue
(M) 17 (2.9) 89 (2.1) 106 (2.2)
Chapter XIV – Diseases of the genitourinary system (N) 59 (10.1) 367 (8.5) 426 (8.7)
Chapter XV – Pregnancy, childbirth and the puerperium (O) 0.0 (0) 0.0 (0) 0.0 (0)
Chapter XVI – Certain conditions originating in the perinatal period (P) 0.0 (0) 0.0 (0) 0.0 (0)
Chapter XVII – Congenital malformations, deformations and chromosomal
abnormalities (Q) 3 (0.5) 19 (0.4) 22 (0.5)
Chapter XVIII – Symptoms, signs and abnormal clinical and laboratory
findings, not elsewhere classified (R) 92 (15.8) 469 (10.9) 561 (11.5)
Chapter XIX – Injury, poisoning and certain other consequences of external
causes (S and T) 22 (3.8) 140 (3.3) 162 (3.3)
Chapter XX – External causes of morbidity and mortality (V, W, X, and Y) 26 (4.5) 195 (4.5) 221 (4.5)
Chapter XXI – Factors influencing health status and contact with health
services (Z) 12 (2.1) 69 (1.6) 81 (1.7)
123
Chapter XXII – Codes for special purposes (U) 98 (16.8) 45 (1.1) 143 (2.9)
Unknown 12 (2.1) 87 (2.0) 99 (2.0)
Causes of death in the UK Biobank are taken from Data-Field 41202. Percentages are calculated based on total number of deaths in each group.
124
Table S4. Association of COVID-19 at All Levels of Severity with Increased Risk of MACE.
Outcome
Outcome
Among All
Cases (Yes/No)
Outcome Among All
Population Controls
(Yes/No) HR (95% CI)
Thrombotic events
All subjects 283/9,722
(2.8%) 5,466/212,264 (2.5%) *
1.16 (1.03-1.31)
Without CVD 95/8,763 (1.1%) 1,836/190,658 (1.0%) 1.17 (0.95-1.44)
With CVD 188/959 (16.4%) 3,630/21,606 (14.4%) 1.12 (0.97-1.30)
All-cause mortality
All subjects 582/9,423
(5.8%) 4,297/213,433 (2.0%)
***3.37 (3.09-
3.68)
Without CVD 415/8,443
(4.7%) 3,055/189,439 (1.6%)
***3.55 (3.20-
3.94)
With CVD 167/980 (14.6%) 1,242/23,994 (4.9%)
***2.91 (2.47-
3.42)
MACE
All subjects 813/9,192
(8.1%) 9,183/208,547 (4.2%)
***2.09 (1.94-
2.25)
Without CVD 492/8,366
(5.6%) 4,672/187,822 (2.4%)
***2.63 (2.39-
2.88)
With CVD 321/826 (28.0%) 4,511/20,725 (17.9%)
***1.55 (1.38-
1.74)
Data are shown as hazard ratios (HR) and 95% confidence intervals (CI) for association of COVID-19 at
all levels of severity with risk of thrombotic events (MI or stroke), all-cause mortality, and MACE (MI,
stroke, or all-cause mortality).
HRs and 95% CIs were derived from Cox proportional hazards models adjusted for age, sex, ethnicity,
education, diabetes, asthma, smoking, lipid-lowering medication use, and anti-hypertension medication
use.
*P<0.05; ***P<0.0005.
125
Table S5. Association of COVID-19 with Increased Risk of MACE Across Different Time Periods.
Group (Time Period)
Outcome Among
Cases (Yes/No)
Outcome Among
Controls (Yes/No) HR (95% CI)
All Cases and All Population Controls
0-365 Days 480/9,525 (4.8%) 4,871/212,859 (2.2%) ***2.41 (2.19-2.65)
365-730 Days 288/9,237 (3.0%) 3,911/199,201 (1.9%) ***2.06 (1.91-2.22)
730-1003 Days 45/4,191 (1.1%) 401/79,556 (0.5%) ***2.75 (2.01-3.75)
Hospitalized Cases and All Population Controls
0-365 Days 307/1,636 (15.8%) 4,871/212,859 (2.2%) ***5.05 (4.50-5.68)
365-730 Days 133/1,503 (8.1%) 3,911/199,201 (1.9%) ***3.98 (3.62-4.39)
730-1003 Days 23/935 (2.4%) 401/79,556 (0.5%) ***2.37 (1.55-3.62)
Data are shown as hazard ratios (HR) and 95% confidence intervals (CI) for association of COVID-19 with risk of MACE (MI, stroke, or all-cause
mortality) during different time periods after entry into the analyses.
HRs and 95% CIs were derived from Cox proportional hazards adjusted for age, sex, ethnicity, education, diabetes, asthma, smoking, lipidlowering medication use, and anti-hypertension medication use.
***P<0.0005.
126
Table S6. Sensitivity Analyses for Risk of Thrombotic Events Among COVID-19 Cases at All Levels of Severity and All Population
Controls.
Stratification Status
Outcome Among All
Cases (Yes/No)
Outcome Among All
Controls (Yes/No) HR (95% CI)
CVD
No 95/8,763 (1.1%) 1,836/190,658 (1.0%) 1.17 (0.95-1.44)
Yes 188/959 (16.4%) 3,630/21,606 (14.4%) 1.12 (0.97-1.30)
Age (years)
<65 66/4,987 (1.3%) 749/69,885 (1.1%) 1.08 (0.84-1.40)
≥65 217/4,735 (4.4%) 4,717/142,379 (3.2%)
*
1.18 (1.03-1.36)
Male
No 112/5,235 (2.1%) 1,773/119,554 (1.5%)
***1.53 (1.26-1.85)
Yes 171/4,487 (3.7%) 3,693/92,710 (3.8%) 1.00 (0.85-1.16)
Obese (BMI≥30)
No 165/6,791 (2.4%) 3,566/161,095 (2.2%) 1.16 (0.99-1.36)
Yes 118/2,931 (3.9%) 1,900/51,169 (3.6%) 1.13 (0.94-1.37)
Diabetes
No 192/8,580 (2.2%) 3,986/193,311 (2.0%)
*
1.20 (1.04-1.39)
Yes 91/1,142 (7.4%) 1,480/18,953 (7.2%) 1.07 (0.87-1.33)
Smoking
No 107/5,289 (2.0%) 2,229/119,200 (1.8%) 1.13 (0.93-1.38)
Yes 176/4,433 (3.8%) 3,237/93,064 (3.4%)
*
1.18 (1.01-1.37)
Data are shown as hazard ratios (HR) and 95% confidence intervals (CI) for association of COVID-19 at all levels of severity with risk of
thrombotic events (MI or stroke), stratified by the indicated variable. HRs and 95% CIs in analyses stratified by CVD status, age, or obesity status
were derived from Cox proportional hazards adjusted for age, sex, ethnicity, education, diabetes, asthma, smoking, lipid-lowering medication use,
and anti-hypertension medication use. All other analyses were adjusted for same covariates except for factor by which subjects were stratified.
*P<0.05; ***P<0.0005
127
Table S7. Sensitivity Analyses for Risk of All-cause Mortality Among COVID-19 Cases at All Levels of Severity and All Population
Controls.
Stratification Status
Outcome Among All
Cases (Yes/No)
Outcome Among All
Controls (Yes/No) HR (95% CI)
CVD
No 415/8,443 (4.7%) 3,055/189,439 (1.6%)
***3.55 (3.20-3.94)
Yes 167/980 (14.6%) 1,242/23,994 (4.9%)
***2.91 (2.47-3.42)
Age (years)
<65 49/5,004 (1.0%) 421/70,213 (0.6%)
***1.49 (1.49-2.00)
≥65 533/4,419 (10.8%) 3,876/143,220 (2.6%)
***3.75 (3.42-4.10)
Male
No 261/5,086 (4.9%) 1,855/119,472 (1.5%)
***3.79 (3.32-4.32)
Yes 321/4,337 (6.9%) 2,442/93,961 (2.5%)
***3.08 (2.74-3.46)
Obese (BMI≥30)
No 379/6,577 (5.5%) 2,936/161,725 (1.8%)
***3.63 (3.26-4.04)
Yes 203/2,846 (6.7%) 1,361/51,708 (2.6%)
***2.92 (2.52-3.39)
Diabetes
No 425/8,347 (4.8%) 3,367/193,930 (1.7%)
***3.53 (3.18-3.90)
Yes 157/1,076 (12.7%) 930/19,503 (4.6%)
***2.99 (2.53-3.55)
Smoking
No 230/5,166 (4.3%) 1,749/119,680 (1.4%)
***3.73 (3.24-4.28)
Yes 352/4,257 (7.6%) 2,548/93,753 (2.7%)
***3.18 (2.84-3.56)
Data are shown as hazard ratios (HR) and 95% confidence intervals (CI) for association of COVID-19 at all levels of severity with risk of allcause mortality, stratified by the indicated variable. HRs and 95% CIs in analyses stratified by CVD status, age, or obesity status were derived
from Cox proportional hazards adjusted for age, sex, ethnicity, education, diabetes, asthma, smoking, lipid-lowering medication use, and antihypertension medication use. All other analyses were adjusted for same covariates except for factor by which subjects were stratified. ***P<0.0005.
128
Table S8. Sensitivity Analyses for Risk of MACE Among COVID-19 Cases at All Levels of Severity and All Population Controls.
Stratification Status
Outcome Among All
Cases (Yes/No)
Outcome Among All
Controls (Yes/No) HR (95% CI)
CVD
No 492/8,366 (5.6%) 4,672/187,822 (2.4%)
***2.63 (2.39-2.88)
Yes 321/826 (28.0%) 4,511/20,725 (17.9%)
***1.55 (1.38-1.74)
Age (years)
<65 112/4,941 (2.2%) 1,131/69,503 (1.6%)
*
1.24 (1.02-1.50)
≥65 701/4,251 (14.2%) 8,052/139,044 (5.5%)
***2.32 (2.15-2.50)
Male
No 353/4,994 (6.6%) 3,437/117,890 (2.8%)
***2.64 (2.37-2.95)
Yes 460/4,198 (9.9%) 5,746/90,657 (6.0%)
***1.79 (1.63-1.97)
Obese (BMI≥30)
No 513/6,443 (7.4%) 6,150/158,511 (3.7%)
***2.21 (2.01-2.41)
Yes 300/2,749 (9.8%) 3,033/50,036 (5.7%)
***1.87 (1.66-2.11)
Diabetes
No 586/8,186 (6.7%) 6,969/190,328 (3.5%)
***2.21 (2.03-2.41)
Yes 227/1,006 (18.4%) 2,214/18,219 (10.8%)
***1.81 (1.57-2.07)
Smoking
No 325/5,071 (6.0%) 3,768/117,661 (3.1%)
***2.22 (1.98-2.49)
Yes 488/4,121 (10.6%) 5,415/90,886 (5.6%)
***2.01 (1.83-2.21)
Data are shown as hazard ratios (HR) and 95% confidence intervals (CI) for association of COVID-19 at all levels of severity with risk of MACE
(MI, stroke, or all-cause mortality), stratified by the indicated variable. HRs and 95% CIs in analyses stratified by CVD status, age, or obesity
status were derived from Cox proportional hazards adjusted for age, sex, ethnicity, education, diabetes, asthma, smoking, lipid-lowering
medication use, and anti-hypertension medication use. All other analyses were adjusted for same covariates except for factor by which subjects
were stratified.*P<0.05; ***P<0.0005
129
Table S9. Sensitivity Analyses for Risk of Thrombotic Events Among Hospitalized COVID-19 Cases and All Population Controls.
Stratification Status
Outcome Among Hospitalized
Cases (Yes/No)
Outcome Among All
Controls (Yes/No) HR (95% CI)
CVD
No 38/1,513 (2.5%) 1,836/190,658 (1.0%)
***1.85 (1.34-2.55)
Yes 64/328 (16.3%) 3,630/21,606 (14.4%) 1.08 (0.85-1.39)
Age (years)
<65 12/467 (2.5%) 749/69,885 (1.1%) 1.26 (0.71-2.24)
≥65 90/1,374 (6.2%) 4,717/142,379 (3.2%)
**1.38 (1.12-1.71)
Male
No 37/837 (4.2%) 1,773/119,554 (1.5%)
**1.74 (1.24-2.41)
Yes 65/1,004 (6.1%) 3,693/92,710 (3.8%) 1.21 (0.95-1.55)
Obese (BMI≥30)
No 59/1,110 (5.1%) 3,566/161,095 (2.2%)
**1.52 (1.17-1.96)
Yes 43/731 (5.6%) 1,900/51,169 (3.6%) 1.18 (0.87-1.60)
Diabetes
No 55/1,390 (3.8%) 3,986/193,311 (2.0%)
*
1.37 (1.05-1.79)
Yes 47/451 (9.4%) 1,480/18,953 (7.2%) 1.33 (0.996-1.78)
Smoking
No 36/838 (4.1%) 2,229/119,200 (1.8%)
*
1.45 (1.05-2.02)
Yes 66/1,003 (6.2%) 3,237/93,064 (3.4%)
*
1.33 (1.04-1.70)
Data are shown as hazard ratios (HR) and 95% confidence intervals (CI) for association of hospitalized COVID-19 with risk of thrombotic events
(MI or stroke), stratified by the indicated variable. HRs and 95% CIs in analyses stratified by CVD status, age, or obesity status were derived from
Cox proportional hazards adjusted for age, sex, ethnicity, education, diabetes, asthma, smoking, lipid-lowering medication use, and antihypertension medication use. All other analyses were adjusted for same covariates except for factor by which subjects were stratified. *P<0.05;
**P<0.005; ***P<0.0005.
130
Table S10. Sensitivity Analyses for Risk of All-cause Mortality Among Hospitalized COVID-19 Cases and All Population Controls.
Stratification Status
Outcome Among Hospitalized
Cases (Yes/No)
Outcome Among All
Controls (Yes/No) HR (95% CI)
CVD
No 284/1,267 (18.3%) 3,055/189,439 (1.6%)
***8.30 (7.33-9.39)
Yes 111/281 (28.3%) 1,242/23,994 (4.9%)
***4.87 (4.00-5.93)
Age (years)
<65 27/452 (5.6%) 421/70,213 (0.6%)
***5.72 (3.84-8.53)
≥65 368/1,096 (25.1%) 3,876/143,220 (2.6%)
***7.17 (6.43-8.00)
Male
No 166/708 (19.0%) 1,855/119,472 (1.5%)
***8.48 (7.21-9.97)
Yes 229/840 (21.4%) 2,442/93,961 (2.5%)
***6.28 (5.48-7.21)
Obese (BMI≥30) No 251/918 (21.5%) 2,936/161,725 (1.8%)
***8.31 (7.29-9.47)
Yes 144/630 (18.6%) 1,361/51,708 (2.6%)
***5.44 (4.57-6.48)
Diabetes No 285/1,160 (19.7%) 3,367/193,930 (1.7%)
***8.56 (7.57-9.67)
Yes 110/388 (22.1%) 930/19,503 (4.6%)
***4.65 (3.81-5.67)
Smoking No 156/718 (17.9%) 1,749/119,680 (1.4%)
***9.02 (7.64-10.7)
Yes 239/830 (22.4%) 2,548/93,753 (2.7%)
***6.20 (5.42-7.10)
Data are shown as hazard ratios (HR) and 95% confidence intervals (CI) for association of hospitalized COVID-19 with risk of all-cause mortality,
stratified by the indicated variable. HRs and 95% CIs in analyses stratified by CVD status, age, or obesity status were derived from Cox
proportional hazards adjusted for age, sex, ethnicity, education, diabetes, asthma, smoking, lipid-lowering medication use, and anti-hypertension
medication use. All other analyses were adjusted for same covariates except for factor by which subjects were stratified.***P<0.0005.
131
Table S11. Sensitivity Analyses for Risk of MACE Among Hospitalized COVID-19 Cases and All Population Controls.
Stratification Status
Outcome Among Hospitalized
Cases (Yes/No)
Outcome Among All
Controls (Yes/No) HR (95% CI)
CVD
No 307/1,244 (19.8%) 4,672/187,822 (2.4%)
***5.93 (5.28-6.67)
Yes 156/236 (39.8%) 4,511/20,725 (17.9%)
***2.08 (1.78-2.45)
Age (years)
<65 37/442 (7.7%) 1,131/69,503 (1.6%)
***2.73 (1.96-3.80)
≥65 426/1,038 (29.1%) 8,052/139,044 (5.5%)
***3.97 (3.60-4.38)
Male
No 194/680 (22.2%) 3,437/117,890 (2.8%)
***5.11 (4.41-5.92)
Yes 269/800 (25.2%) 5,746/90,657 (6.0%)
***3.25 (2.87-3.67)
Obese (BMI≥30) No 287/882 (24.6%) 6,150/158,511 (3.7%)
***4.46 (3.96-5.03)
Yes 176/598 (22.7%) 3,033/50,036 (5.7%)
***3.06 (2.63-3.57)
Diabetes No 319/1,126 (22.1%) 6,969/190,328 (3.5%)
***4.61 (4.12-5.17)
Yes 144/354 (28.9%) 2,214/18,219 (10.8%)
***2.69 (2.27-3.19)
Smoking No 183/691 (20.9%) 3,768/117,661 (3.1%)
***4.69 (4.04-5.45)
Yes 280/789 (26.2%) 5,415/90,886 (5.6%)
***3.45 (3.06-3.89)
Data are shown as hazard ratios (HR) and 95% confidence intervals (CI) for association of hospitalized COVID-19 with risk of MACE (MI,
stroke, or all-cause mortality), stratified by the indicated variable. HRs and 95% CIs in analyses stratified by CVD status, age, or obesity status
were derived from Cox proportional hazards adjusted for age, sex, ethnicity, education, diabetes, asthma, smoking, lipid-lowering medication use,
and anti-hypertension medication use. All other analyses were adjusted for same covariates except for factor by which subjects were stratified.
***P<0.0005.
132
Table S12. Association of Hospitalization for COVID-19 with Increased Risk of Cardiovascular Mortality and MACE.
Outcome
Outcome Among Hospitalized
Cases (Yes/No)
Outcome Among All Population
Controls (Yes/No) HR (95% CI)
Cardiovascular mortality
All subjects 166/1,777 (8.5%) 1,840/215,890 (0.9%) ***6.26 (5.33-7.36)
Without CVD 107/1,444 (6.9%) 1,049/191,445 (0.5%) ***8.27 (6.75-10.1)
With CVD 59/333 (15.1%) 791/24,445 (3.1%) ***3.99 (3.05-5.21)
MACE
All subjects 243/1,700 (12.5%) 6,870/210,860 (3.2%) ***2.59 (2.28-2.95)
Without CVD 133/1,418 (8.6%) 2,719/189,775 (1.4%) ***4.25 (3.56-5.06)
With CVD 110/282 (28.1%) 4,151/21,085 (16.5%) ***1.60 (1.33-1.94)
Data are shown as hazard ratios (HR) and 95% confidence intervals (CI) for association of hospitalized COVID-19 with risk of cardiovascular
mortality and MACE (defined as MI, stroke, and cardiovascular mortality).
HRs and 95% CIs were derived from Cox proportional hazards adjusted for age, sex, ethnicity, education, diabetes, asthma, smoking, lipidlowering medication use, and anti-hypertension medication use.
***P<0.0005
133
Table S13. Association of COVID-19 with Risk of Thrombotic Events Stratified by Anti-Platelet Agents.
Group
Anti-Platelet
Agents
Outcome Among
Cases (Yes/No)
Outcome Among
Controls (Yes/No) HR (95% CI) aP-interaction
All Cases vs. All Population Controls
All
No 164/8,494 (1.9%) 3,053/182,430 (1.7%) *1.18 (1.01-1.38)
0.73
Yes 119/1,228 (8.8%) 2,413/29,834 (7.5%) 1.14 (0.95-1.37)
With CVD
No 82/534 (13.3%) 1,533/11,443 (11.8%) 1.10 (0.88-1.37)
0.84
Yes 106/425 (20.0%) 2,097/10,163 (17.1%) 1.15 (0.94-1.40)
Without CVD
No 82/7,960 (1.0%) 1,520/170,987 (0.9%) 1.20 (0.96-1.51)
0.49
Yes 13/803 (1.6%) 316/19,671 (1.6%) 0.99 (0.57-1.72)
Hospitalized Cases vs. All Population Controls
All Subjects
No 62/1,462 (4.1%) 3,053/182,430 (1.7%) ***1.65 (1.28-2.12)
0.03
Yes 40/379 (9.6%) 2,413/29,834 (7.5%) 1.12 (0.82-1.54)
With CVD
No 30/171 (14.9%) 1,533/11,443 (11.8%) 1.19 (0.83-1.71)
0.59
Yes 34/157 (17.8%) 2,097/10,163 (17.1%) 1.03 (0.74-1.45)
Without CVD No 32/1,291 (2.4%) 1,520/170,987 (0.9%) ***1.98 (1.39-2.82) 0.38
Yes 6/222 (2.6%) 316/19,671 (1.6%) 1.35 (0.60-3.03)
Data are shown as hazard ratios (HR) and 95% confidence intervals (CI) for association of COVID-19 with risk of thrombotic events (MI or
stroke) among subjects with or without CVD, stratified by use of anti-platelet agents. HRs and 95% CIs were derived from Cox proportional
hazards with adjustment for age, sex, education, self-reported ethnicity, diabetes, asthma, smoking, lipid-lowering medication use, and antihypertension medication use. CVD was additionally included as a covariate in the analyses with all subjects.
134
aP-values were obtained from models that included an interaction term between COVID-19 status and reported use of anti-platelet agents.
*P<0.05; ***P<0.0005.
135
Table S14. Hospitalization for COVID-19 Does not Increase Risk of Thrombotic Events Through Interactions with Genetic Variants
Associated with Risk of CAD, SARS-CoV-2 Infection Susceptibility, or Hospitalization for COVID-19.
Locus
Genetic
Variant
Previously
Associated
Trait aGenotype
Thrombotic
Event Among
Hospitalized
Cases
(Yes/No)
Thrombotic Event
Among All
Population
Controls (Yes/No) bHR (95% CI)
Pvalue
cPinteraction
3p21.3
1
rs7306238
9
Susceptibility
to SARS-CoV2 infection
GG 86/1,561
(5.2%)
4,779/183,784
(2.5%) 1.39 (1.12-1.72) 2.5x10-
3
0.81
GA/AA 12/219 (5.2%) 562/23,210 (2.4%) 1.26 (0.71-2.24) 0.44
3p21.3
1
rs1138594
2
Hospitalized
COVID-19
-/-
82/1,450
(5.4%)
4,581/177,265
(2.5%) 1.41 (1.13-1.75) 2.1x10-
3
0.71
-/A and
A/A 16/324 (4.7%) 751/29,005 (2.5%) 1.24 (0.76-2.04) 0.39
9p21 rs4977574 CAD
AA 24/499 (4.6%) 1,260/55,666
(2.2%) 1.37 (0.91-2.05) 0.13
AG 48/867 (5.3%) 0.96 2,651/102,576
(2.5%) 1.36 (1.03-1.81) 0.033
GG 26/414 (5.9%) 1,430/48,752
(2.9%) 1.40 (0.95-2.06) 0.092
Data are shown as hazard ratios (HR) and 95% confidence intervals (CI) for association of hospitalization for COVID-19 with risk of thrombotic
events (MI or stroke) stratified by genotypes of the indicated genetic variants. HRs and 95% CIs were derived from Cox proportional hazards
regression models adjusted for age, sex, PC1-10, genotyping array, education, diabetes, asthma, smoking, lipid-lowering medication use, and antihypertension medication use.
a
rs11385942 is an insertion (A) variant.
bHRs (95% CIs) and P-values derived from analyses with rs73062389 and rs11385942 were obtained assuming dominant genetic models.
cP-values were obtained from models that included an interaction term between COVID-19 status and the indicated genetic variant.
136
CHAPTER 3: GENETICALLY DECREASED CPS1 ACTIVITY ATTENUATES
ATHEROSCLEROSIS IN HUMANS AND MICE THROUGH SEXUALLY DIMORPHIC
PATTERNS
Summary. In the present study, we replicated the cardioprotective effect of the rs715 variant in
CPS1, previously shown to be protective for coronary artery disease (CAD) in women but not in
men. Our multi-ancestry single-SNP meta-analysis of > 1.4 million subjects in the UK Biobank,
GeneBank, FinnGen, CARDIoGRAM+C4D consortium, All of Us, and Biobank Japan cohorts
demonstrated that the protective effect of rs715 was highly robust in females across all
ancestries, while the effect in males was less pronounced. To functionally validate these
observations, heterozygous Cps1 deficient mice were examined. Male mice exhibited decreased
aortic lesion formation compared to wild-type littermates, recapitulating a sexually dimorphic
but opposite pattern observed with the CPS1 decreasing allele of rs715 in humans. This
unexpected result was clarified through mechanistic insights showing that natural basal levels of
CPS1 expression differ by sex. Female humans had naturally lower hepatic CPS1 expression
compared to males, while in male mice, lower baseline Cps1 expression, combined with
heterozygous deficiency, contributed to the observed atheroprotective effect. My contributions
included conducting genetic analyses across large biobanks to identify and confirm the effects of
rs715 in human cohorts. I also analyzed the phenotype data from the murine model and
contributed to preparing and curating the manuscript.
137
Genetically Decreased CPS1 Activity Attenuates Atherosclerosis in Humans and Mice
Through Sexually Dimorphic Patterns
James R. Hilser1,2*, Nicholas C. Woodward1,2*, Zeneng Wang3,4,5, Janet Gukasyan1,2, Ina
Nemet3,4, William S. Schwartzman2,3, Pin Huang2,3, Yi Han2,3, Sarada Charugundla6,7,8, Zachary
Fouladian6,7,8, Calvin Pan6,7,8, W.H. Wilson Tang3,4,5
, Samuli Ripatti 9
, Aarno Palotie9
, Lijiang
Ma11, Johan L.M. Björkegren12,
, Aldons J. Lusis, PhD6,7,8, Stanley L. Hazen3,4,5, Jaana A.
Hartiala2
, and Hooman Allayee1,2
Departments of 1Population & Public Health Sciences, and 2Biochemistry & Molecular
Medicine, Keck School of Medicine, University of Southern California, Los Angeles, CA 90033;
Departments of 3Cardiovascular & Metabolic Sciences, Lerner Research Institute, 4Center for
Microbiome and Human Health, and 5Cardiovascular Medicine, Heart, Vascular and Thoracic
Institute, Cleveland Clinic, Cleveland, OH 44195; Departments of 6Medicine, 7Human Genetics,
and 8Microbiology, Immunology, & Molecular Genetics, David Geffen School of Medicine of
UCLA, Los Angeles, CA 90095; 9
Inst. for Molecular Med. Finland, FIMM, Univ. of Helsinki,
Helsinki, Finland; 11Department of Genetics & Genomic Sciences, Institute of Genomics and
Multiscale Biology, Icahn School of Medicine at Mount Sinai, New York, NY, 10029;
12Department of Medicine, Karolinska Institutet, Karolinska Universitetssjukhuset, 141 57
Huddinge, Sweden.
*These authors contributed equally to this work.
138
Abstract
Over 300 susceptibility loci have been identified for coronary artery disease (CAD) but
the contribution of genetic factors to CAD risk through sexually dimorphic patterns is not well
known. Based on prior metabolomics studies, we evaluated sex-stratified association of
genetically decreased activity of CPS1, the rate-limiting enzyme of the urea cycle, with
atherosclerosis in humans and mice. A strong sexually dimorphic association (Pheterogeneity=2.3x10-5
) was observed between a variant of CPS1 (rs715) and decreased risk of
CAD in women (OR=0.95, 95% CI 0.93-0.96; P=2.6x10-12) compared to men (OR=0.99, 95% CI
0.98-0.998; P=0.011). Male, but not female, heterozygous Cps1 deficient (Cps1+
/
-
) mice had
decreased aortic lesion formation compared to wildtype littermates (452,855±23,550 vs.
343,481±29,752mm2 /section; p=0.009), thus recapitulating a sexually dimorphic but opposite
pattern observed with the CPS1-decreasing allele of rs715 in humans. Regardless of genotype,
CPS1 mRNA levels in livers of women were ~19% lower than men (11.5±1.3 vs. 11.8±1.1 log2
CPM; P=1.1x10-5
) whereas hepatic Cps1 expression was ~28% lower in male mice than in
females (12.2±0.13 vs. 12.5±0.09 log₂ units; P=2.7x10-51). These results provide evidence for
the sexually dimorphic contribution of genetically decreased CPS1 activity to atheroprotective
traits in humans and mice. Further studies will be needed to elucidate the underlying biological
mechanisms for the opposite and differential effects observed in males and females.
139
Introduction
The incidence of coronary artery disease (CAD), myocardial infarction (MI), and sudden
death in women generally lags behind men by 10-20 years leading to the traditional view that
men are more prone to adverse cardiac outcomes. One widely accepted explanation for these
differences has been attributed, at least in part, to the protective effect of female sex hormones
since early menopause and other female endocrine disorders are associated with accelerated
development of CAD [1-4]. Furthermore, the prevalence of cardiometabolic diseases in women
at advanced ages actually outnumbers men, and women generally face a worse prognosis
following a primary cardiovascular event [5]. In this regard, evidence points to sex differences
in the presentation, pathophysiology, and diagnosis of cardiovascular outcomes [6-8]. However,
the sex-specific biological mechanisms that underlie development of atherosclerosis are poorly
understood.
CAD is also known to have a strong genetic component based on both heritability
estimates of 40-60% [9-12] and the results of gene mapping efforts. For example, genome-wide
association studies (GWAS) have discovered >300 loci with effects in both sexes that influence
risk of CAD and MI through perturbations of lipid metabolism, blood pressure regulation,
inflammation, platelet function, and mechanisms that remain unknown [13-18]. However, few
analyses have formally explored sexually dimorphic associations between genetic variants and
risk of CAD. In the most recent GWAS meta-analysis from the CARDIoGRAM+C4D
consortium, 10 loci were reported as exhibiting sex differentiated associations with CAD, of
which the chromosome 9p21 locus was the only one that yield a genome-wide significant
140
association in women [18]. Thus, our understanding of the genetic factors that contribute to risk
of CAD specifically in women also remains incomplete.
By using a systems biology strategy that integrated metabolomics and genetic data, we
identified a functional variant (rs715; T>C) of carbamoyl phosphate synthase 1 (CPS1) that was
associated with reduced risk of CAD in women but not men [19]. CPS1 encodes the ratelimiting enzyme in the urea cycle that detoxifies ammonia to carbamoyl phosphate and is
coupled with ornithine to produce citrulline, followed by arginine, which is then metabolized
further to reform ornithine and simultaneously release urea for excretion by the kidney (Figure
1). Notably, the athero-protective C allele of rs715 was associated with decreased levels of
various urea cycle metabolites [19], suggesting that this variant leads to reduced CPS1 activity.
Thus, one plausible biological model for our previous observations would be that genetically
decreased CPS1 activity reduces flux through the urea cycle and attenuates risk of CAD.
However, CPS1 is one of the most pleiotropic loci in the genome and has been associated with
hundreds of clinical traits, including many cardiometabolic risk factors [20]. Thus, it has been
difficult to determine the precise atheroprotective mechanism of decreased CPS1 activity and
direct functional evidence demonstrating a role for CPS1 in atherogenesis is lacking. In the
present study, we sought to validate the sexually dimorphic genetic association of CPS1 with risk
of CAD in large, multi-ancestry datasets, as well as characterize a mouse model of Cps1
deficiency for atherosclerosis-related phenotypes.
141
Materials and Methods
Study Populations. All human subjects gave written consent for participation in genetic studies
and the protocol of each study was approved by the corresponding local research ethics
committee or institutional review board. The present study was approved by the Institutional
Review Board of USC Keck School of Medicine.
UK Biobank. The UK Biobank is a large, multi-site cohort that recruited participants between
40-69 years of age who were registered with a general practitioner of the UK National Health
Service (NHS) [21]. Between 2006-2010, a total of 503,325 individuals were enrolled through
22 assessment centers in the UK. At enrollment, extensive data on demographics, ethnicity,
education, lifestyle indicators, imaging of the body and brain, and disease-related outcomes were
obtained through questionnaires, health records, and/or clinical evaluations. Blood samples were
also collected at baseline for measurement of serum biomarkers that are either established
disease risk factors or routinely measured as part of clinical evaluations. Subjects were defined
as being of White, African, or Asian/Chinese ancestry using self-reported ethnicity data (Data
Field 21000). CAD cases were defined as positive for International Classification of Diseases,
10th revision (ICD-10) I21, I22, I23, I25.2, I24.0, I24.8, I24.9, I25.0, I25.1, I25.4, I25.8, and
I25.9, which included MI and ischemic heart diseases. Additional CAD cases were identified
using Office of Population Censuses and Surveys Classification of Interventions and Procedures,
version 4 (OPCS-4) codes K40-K46, K49, K50 and K75 covering replacement, transluminal
balloon angioplasty, and other therapeutic transluminal operations on coronary artery and
percutaneous transluminal balloon angioplasty and insertion of stent into coronary
artery. Doctor-diagnosed and self-reported MI was also included in the definition of CAD.
142
Quality control of samples, DNA variants and imputation in the UK Biobank were performed by
the Wellcome Trust Centre for Human Genetics [21]. Briefly, ~90 million single nucleotide
polymorphisms (SNPs) imputed from the Haplotype Reference Consortium, UK10K, and 1000
Genomes imputation were available in 487,164 subjects in the UK Biobank. All study
participants provided informed consent and the study was approved by the North West Multicentre Research Ethics Committee.
Cleveland Clinic GeneBank Cohort. GeneBank is a single site sample repository generated
from ~10,000 consecutive patients who underwent elective diagnostic coronary angiography or
elective cardiac computed tomographic angiography with extensive clinical and laboratory
characterization and longitudinal observation. Written informed consent was obtained from all
participants prior to enrollment and subject recruitment occurred between 2001 and 2007.
Ethnicity was self-reported and information regarding demographics, medical history, and
medication use was obtained by patient interviews and confirmed by chart reviews. All clinical
outcome data were verified by source documentation. Imputation based on 1000 Genomes and
Haplotype Reference Consortium reference panels was done separately for 2,932 and 1,607
samples genotyped on the Affymetrix 6.0 SNP chip or the Illumina Global Screening Array
(GSA), respectively) [16, 22]. Following post-imputation quality control checks, the two
GeneBank datasets were merged to create a combined sample of 4,539 subjects and 9,225,854
SNP genotypes with minor allele frequency (MAF) ≥0.01. The present analysis included 4,478
subjects of northern European ancestry, all of whom had CAD as defined by angiographic
evidence of ≥50% stenosis in one or more major epicardial vessel and/or a documented history
of known CAD. The GeneBank cohort has been used previously for discovery and replication of
novel genes and risk factors for atherosclerotic disease [19, 23-26].
143
All of Us Research Program. All of Us recruited individuals that have been and continue to be
underrepresented in biomedical research due to limited access to health care. Therefore, All of
Us takes demographics, including race, ethnic group, age, sex, gender identity, income,
educational attainment, and geographic location, into account when enrolling participants [27].
All participants completed electronic consent modules and health questionnaires upon
enrollment, and the study protocol has been published previously [28]. Clinical characteristics of
All of Us subjects have been described [28] and we defined CAD for the present analyses using
ICD10 codes I21, I22, I23, I24.8, I24.9, I25.1, I25.2, I25.4, I25.8, I25.9. Six ancestry groups
were defined using principal components (PC) analysis, including European, African, East
Asian, South Asian, Middle Eastern, or Latino/Admixed American ancestry [29]. Due to the
small numbers, subjects of Middle Eastern ancestry were excluded from the present analyses.
The All of Us Research Program provided PLINK 1.9 datasets (.bed, .bim, .fam) for array data
and srWGS SNP and Indel variants over limited genomic regions. PLINK files were generated
from Hail MatrixTables using the export_plink function, preserving all information from the
original MatrixTable. For srWGS data, the ACAF v7.1 callset includes variants with a
population-specific allele frequency >1% or allele count >100 in any ancestry subpopulation
with additional information described [29]. Genotypes for rs715 (chr2:210678331:T:C) was
extracted from the ACAF v7.1 callset using PLINK2.
FinnGen. The FinnGen study (https://www.finngen.fi/en) is an ongoing research project that
utilizes ~500,000 samples from a nationwide network of Finnish biobanks and digital healthcare
data from national health registers [30]. Registry data on all FinnGen participants were collected
and processed from various national health registers. Pseudonymized register data were
combined with the minimum phenotype dataset from the Finnish biobanks (age, sex, year of
144
sampling, height, weight and smoking status). CAD cases were defined using the same 12
ICD10 codes described above for UK Biobank (I21, I22, I23, I25.2, I24.0, I24.8, I24.9, I25.0,
I25.1, I25.4, I25.8, and I25.9). The Finnish ICD version is mostly identical to the international
ICD classification, but has minor modifications. FinnGen participants were genotyped with
either Illumina or Affymetrix arrays and genotype calls were made with GenCall and zCall
algorithms, respectively, as described previously [30]. Quality control steps included removal of
individuals with genetically inferred sex not matching the reported sex in registries, high
genotype missingness (>5%), and excess heterozygosity (±4 standard deviations), and removal
of variants with high missingness (>2%), deviation from Hardy-Weinberg equilibrium (P<1x10-
6
), and minor allele counts <3. Genotype imputation was carried out using a Finnish-specific
reference panel (Sequencing Initiative Suomi v.3) with removal of variants with imputation
INFO scores <0.6 or MAF values <0.0001, as described previously [30].
Statistical Genetics Analyses. Genotypes for rs715 were extracted from the UK Biobank,
GeneBank, All of Us, and FinnGen datasets and tested for association with risk of CAD using
logistic regression models. These analyses were first stratified by sex and ancestry and carried
out in each cohort individually. Covariates used for all analyses in the UK Biobank regardless of
ancestry were age, PC1-20, and genotyping array. For GeneBank, covariates in the model were
age and genotyping array. For All of Us, covariates in the model were age and PC1-16
regardless of ancestry. For FinnGen, covariates included age, genotyping batch, and PC1-10, as
used in prior GWAS analyses [30]. Analyses in UK Biobank, GeneBank, All of Us were
performed with R v4.4.0 (R Core Team, Vienna, Austria) and a custom-GWAS analysis pipeline
was utilized for FinnGen with REGENIE (v2.2.4) [31]. Sex-specific effect sizes for rs715 in
CARDIoGRAM and Biobank Japan were taken from publicly available summary statistics [15,
145
19]. Sex- and ancestry-specific effect sizes from each cohort were combined in fixed-effects
meta-analyses, followed by a combined meta-analysis across both sexes and all ancestries using
METAL [32].
Animal Husbandry. All animal studies were performed with approval and in accordance with
the guidelines of the Institutional Animal Research Committees of the University of Southern
California. Mice were housed in a temperature-controlled facility with a 12-h light/dark cycle
and fed ad libitum with free access to water. Male and female Cps1+/- mice (stock number:
CPS1-DEL1032-EM1-B6N) on a C57BL/6N background were developed by the MRC
Harwell Institute (Oxfordshire, UK) of the International Mouse Phenotyping Consortium
(IMPC). CRISPR-Cas9-mediated genomic modification of the Cps1 locus led to a 1032bp
deletion spanning exon 3 [33]. Mice for the present study were bred in-house and genotyped
by PCR using KAPA2G Fast HotStart PCR kits (KAPA Biosystems, Wilmington, MA) with
the following primers: 5′-GGCATTATGAGGGAGCAACATTT-3′ (forward), 5′-
GCCACCTAGGAAGACAGACAG-3′ (reverse). Amplification was carried out using 1ml of
ear-digested genomic DNA and 0.5mM of each primer for 10 cycles at 94° for 1min and 68°
for 2min, followed by 25 cycles at 95° for 30sec, 55° for 30sec, and 72° for 45sec. This assay
PCR products of 1662bp and 630bp corresponded to the Cps1wildtype and knockout alleles,
respectively. 9-10-week-old age-matched male and female Cps1+/-
and wildtype littermates
were administered a single intraperitoneal injection of 1x1011gc of rAAV8/D377Y-mPCSK9
(AAV-268246, Vector Biolabs, Malvern, PA). One week after injection, mice were placed on
an atherogenic high fat diet containing 1.25% cholesterol (Research Diets # D12108Ci) for 16
weeks. Mice were euthanized after a 12 overnight fast and tissues were harvested and snap
frozen in liquid nitrogen for analysis.
146
Blood Measurements. Following 16 weeks of the atherogenic diet, fasting blood was collected
at the time of euthanization from the submandibular vein. One week prior to euthanization, nonfasting blood was also obtained from a subset of mice two hours after the beginning of the
feeding cycle. Plasma total cholesterol, high-density lipoprotein (HDL) cholesterol, triglyceride,
glucose, and insulin levels were determined by enzymatic colorimetric assays, as described
previously [34, 35]. Combined very low-density lipoprotein (VLDL) cholesterol and lowdensity lipoprotein (LDL) cholesterol levels were calculated by subtracting HDL cholesterol
from total plasma cholesterol levels. Plasma insulin levels were measured in duplicate using
Mouse Ultrasensitive Insulin ELISA kits (Alpco Inc: 80-INSMSU-E10, Salem, NH).
Homeostasis modeling assessment insulin resistance (HOMA-IR) was calculated according to
the formula: [glucose (mmol/L) X insulin (mIU/ml)]/22.5][36]. Inflammatory cytokines were
measured using a multiplexed immunoassay kit (Meso Scale Discovery, K15048G-1, Rockville,
MD).
Metabolomics Measurements. Absolute levels metabolites and amino acids in plasma were
quantified by stable isotope dilution high performance liquid chromatography with online
electrospray ionization tandem mass spectrometry (LC/MS) [37] [46] on an ABI SCIEX QTRAP
5500 mass spectrometer (Applied Biosystems, Foster City, CA) interfaced with a Shimatzu
HPLC (Torrance, CA) equipped with a phenyl column (4.6 × 2505mm, 5μm RexChrom Phenyl;
Regis, Morton Grove, IL). Separation was performed using a gradient starting from 10mM
ammonium formate over 0.5min, then to 5mM ammonium formate, 25% methanol and 0.1%
formic acid over 3min, held for 8min, followed by 100% methanol and water washing for 3min.
Metabolites were monitored in multiple reaction monitoring (MRM) mode using characteristic
parent-daughter ion transitions at m/z ratios for each metabolite. Stable isotope labeled internal
147
standards for each monitored analyte were added to plasma samples prior to protein precipitation
and similarly monitored at the appropriate transitions in MRM mode. Various concentrations of
metabolite standards and a fixed amount of internal standards were spiked into control plasma to
prepare the calibration curves for quantification of plasma analytes.
Gene Expression Analysis. Total RNA was extracted using RNeasy Mini kits (Qiagen,
Valencia, CA) and cDNA was prepared from 500ng of total RNA using cDNA Archive Reverse
Transcription kits (Life Technologies, Gaithersburg, MD) according to the manufacturer’s
protocols. Real-time qPCR gene expression reactions were carried out in triplicate using readymade Taqman assays for Cps1 (Mm01256489_m1) and Ppia (Mm02342430_g1), as an
endogenous control, on an Applied Biosystems 7900HT instrument (Foster City, CA). Cycle
threshold (Ct) values were determined using RQ Manager 1.2.1 and relative quantification was
calculated using the 2^-ΔΔCt method (SDS 2.4). Cps1 transcript levels for each replicate were
then normalized to Ppia and calculated relative to a calibrator comprised of a pool of all
samples. The replicates for each individual mouse sample were then averaged and used for
statistical analysis.
Western Blot Analysis. CPS1 protein levels in neonatal livers were quantified through
Western blot analysis. Briefly, newborn Cps1+/+ (wildtype littermates), heterozygous Cps1+/-
,
and homozygous Cps1-/- pups were euthanized one day after birth and livers were dissected
out. 10mg of tissue was immediately added to RIPA lysis buffer (Thermo Fisher Scientific,
Waltham, MA) supplemented with a protease inhibitor cocktail (Thermo Fisher Scientific)
and homogenized. Lysates were standardized with BSA assay and 5μg of protein was
resolved on 4%–20% TGX gels (BioRad, Hercules CA) and transferred to Immun-Blot PVDF
148
membranes (BioRad). Western blotting was performed using a primary anti-CPS1 antibody
(Abcam, Cambridge, UK) at a 1:50,000 dilution, a loading control anti-beta Actin antibody
(Abcam) at a 1:5,000 dilution, and a secondary HRP-conjugated antibody (Abcam) at a
1:50,000 dilution in 3% BSA-50 (Rockland Immunochemicals, Limerick, PA) in Trisbuffered saline and 0.1% Tween 20. Bands were visualized using chemiluminescent HRP
substrate (Millipore Sigma, Burlington, MA). Densitometric quantification of CPS1 relative
to beta Actin was carried out with ImageJ software (NIH).
Aortic Lesion and En Face Analyses. After euthanization, hearts were perfused through the
left ventricle with approximately 15ml of phosphate-buffered saline (PBS) and characterized for
atherosclerotic lesion formation, as described previously [38]. The hearts were then removed,
placed in 10% formalin solution, and transferred to 30% sterile sucrose for 48hrs before being
embedded in OCT compound (Fisher Scientific, Hampton, NH). Serial interrupted 10mm thick
aortic cryosections were cut starting at the origins of the aortic valve leaflets and stained with Oil
Red O and hematoxylin for quantification of atherosclerotic lesion area and with Masson’s
Trichrome for quantification of collagen content and necrotic core area. For each mouse, total
aortic lesion size or necrotic area was determined in blinded fashion by summing the lesion areas
of 10 sections using ImageJ. En face analysis was carried out on a different set of euthanized
mice after first perfusing the heart and aorta with 15ml of PBS, followed by a formal sucrose
solution (4% paraformaldehyde/7.5% sucrose/10mM sodium phosphate buffer/2mM
EDTA/20mM butylated hydroxytoluene) for 15mins, and rinsing again with 10ml of PBS.
Aortas from the aortic arch to iliac bifurcation were carefully dissected out under a microscope
(Richter Optica S6-RLT) and the surrounding adventitial fat tissue was removed. The aorta was
then opened longitudinally from the aortic root to iliac bifurcation and pinned on a black rubber
149
plate filled with PBS. The aortas were then incubated in 70% ethanol for 5mins, stained with
Sudan-IV solution (5mg/ml Sudan-IV in 70% ethanol and 100% acetone) for 15mins, and destained with 80% ethanol for 3mins. Aortas were then briefly rinsed under running tap water to
remove any residual ethanol, then submerged in PBS for image capture. Images were taken with
a digital camera and Sudan-IV stained atherosclerotic lesion area was calculated using ImageJ.
Lesion area along the aortic arch, descending aorta, and abdominal aorta was calculated as the
percentage of total area. All image capture and quantitation for the en face analyses were done
in a blinded fashion.
Gonadectomy. Gonadectomy was performed as previously described [39]. Briefly, male and
female C57BL/6J mice were purchased from The Jackson Laboratory (Bar Harbor, ME),
maintained on a chow diet, gonadectomy was performed at 6 weeks of under isoflurane
anesthesia. Scrotal regions of male mice were bilaterally incised, testes removed, and the
incisions closed with wound clips. Ovaries of female mice were removed through an incision
below the ribcage (about midway between the hip and the bottom of the ribcage, and about 2/3 of
the way from ventrum to dorsum). The muscle layer was sutured, and the incision closed with
wound clips. In sham-operated control mice, incisions were made and closed as described above
but the gonads remained intact. At 16 weeks of age, mice were fasted for 4hrs and euthanized
for collection of blood and tissues. Total RNA was isolated from liver and RNA libraries were
prepared by the sequencing facility at the UCLA Neurosciences Genomics Core using Illumina
TruSeq Stranded kits v2. Paired-end RNA sequencing was carried and reads were aligned using
STAR 2.5.2b, mm10 version of the mouse genome, and GENCODE M11 transcript annotation,
as described previously [40]. Reads-per-gene tables were generated as part of the STAR output
150
and Cps1 expression levels were log2 transformed. Metabolomics analyses in plasma for urea
cycle metabolites were carried out by LC/MS as described above.
Hepatic CPS1 Gene Expression in Humans and Mice. CPS1 expression was evaluated in
livers of 403 men and 196 women using data from the STARNET cohort [41]. Briefly, 502
CAD patients and 97 controls undergoing open-thorax surgery were recruited at the Department
of Cardiac Surgery, Tartu University Hospital, Estonia [41] after providing written informed
consent. At the time of surgery, biopsies were obtained from liver (n=599 and single-end RNA
sequencing was carried out at a length of 50-100 bp to a depth of 20-30 million reads on an
Illumina HiSeq sequencer (Illumina), as described previously [41]. Genotyping was performed
using the Illumina Infinium assay with the human OmniExpressExome-8v1 bead chip [41].
Expression data were adjusted for age, BMI, CAD status, library protocols, sequencing
laboratory, and four genotyping multidimensional scaling components using linear regression.
Data were analyzed as counts per million (CPM) with log2 transformation. Cps1 expression in
livers of male and female mice was evaluated among 89 inbred mouse strains from the Hybrid
Mouse Diversity Panel [42]. Briefly, gene expression profiles were determined by hybridization
of total RNA to Affymetrix HT-MG_430 PM microarrays (n=1-3 mice of each sex per strain).
After quality control steps, gene expression data were background corrected and normalized
using the affy package (from bioconductor) using rma, pmonly, and median-polish normalization
methods. Data were analyzed using strain-averaged log2 transformed values.
Statistical Analyses. Sex-stratified differences were determined by unpaired Student’s t-tests
(PRISM v6.04, GraphPad Software, Boston, MA). Values are expressed as mean ± SD or mean
151
± SE and differences and were considered statistically significant at P<0.05 or at the significance
threshold after correcting for multiple comparisons.
152
Results
Sexually Dimorphic Association of CPS1 Variant rs715 with Decreased Risk of CAD in
Humans. We leveraged access to primary- and summary-level data from several cohorts to test
association of rs715 with risk CAD using a large-scale, sex-stratified, and multi-ancestry metaanalysis approach (Table 1). Among subjects of European ancestry (n=1,140,340), rs715 was
significantly associated with decreased risk of CAD in women (OR=0.95, 95% CI 0.93-0.96;
P=2.7x10⁻10; n=633,354) but not men (OR=0.99, 95% CI 0.98-1.00; P=0.056; n=506,986)
(Table 1). By comparison, in East/South Asian ancestry subjects, rs715 was significantly
associated with decreased risk of CAD in both women (OR=0.92, 95% CI 0.88-0.97; P=7.5x10⁻4
;
n=88,288) and men (OR=0.96, 95% CI 0.93-0.99; P=0.011; n=99,132), although the effect size
in men was attenuated compared to women (Table 1). Although association of rs715 with risk
of CAD did not achieve statistical significance in women or men of Black/African or
Latino/Admixed American ancestry (Table 1), a combined meta-analysis across all ancestry
groups revealed a striking sexually dimorphic association of rs715 with decreased risk of CAD
(P-heterogeneity=6.7x10-5
) in women (OR=0.95, 95% CI 0.93-0.96; P=7.3x10-12; n=786,952)
and men (OR=0.99, 95% CI 0.98-0.998; P=0.011; n=647,318) (Table 1).
Genetic Deficiency of Cps1 Attenuates Atherosclerosis in Mice. We next sought to
functionally validate the sexually dimorphic impact of decreased Cps1 activity on aortic lesion
development in mice using a genetically modified mouse model. A 1032bp deletion spanning
exon 3 generated by CRISPR-Cas9-mediated genomic modification of the Cps1 locus was
confirmed at the DNA level in homozygous Cps1 knockout (Cps1-/-
) and heterozygous Cps1+/-
newborn pups (Figure 1A). As expected, this led to reduced CPS1 protein in the livers of
153
Cps1+/-
animals and its complete absence in Cps1-/- newborn pups (Figure 1B). However, since
complete Cps1 deficiency in mice results in neonatal mortality within a few days of birth [43]
and mirrors the clinical severity observed in humans with homozygous CPS1 mutations when
left untreated [44], it precluded us from studying Cps1-/- mice. Therefore, we focused our
subsequent studies on characterizing adult heterozygous Cps1+/- mice and wildtype littermates
for atherosclerosis-related traits. In addition to reduced CPS1 protein levels, Cps1+/- mice also
exhibited significantly lower plasma levels of urea cycle metabolites, including citrulline and
ornithine but not arginine, and with a stronger effect in males compared to female mice (Figure
1C-E; Table 2). By comparison, differences in urea cycle metabolites in the non-fasted state
were less pronounced with only citrulline being significantly decreased in both male and female
Cps1+/- mice (Table 2).
To evaluate aortic lesion formation in Cps1+/- mice and wildtype littermates, we induced
hyperlipidemia through a one-time intraperitoneal injection of an adeno-associated virus (AAV)
to overexpress murine proprotein convertase subtilisin/kexin type 9 (AAV-Pcsk9), followed by
feeding of an atherogenic diet for 16 weeks. Consistent with prior studies, administration of
AAV-Pcsk9 led to increased fasting lipid levels and insulin resistance (Table 3). Compared to
wildtype littermates, male Cps1+/- mice had decreased atherosclerotic lesion area at the aortic
root (452,855±23,550 vs. 343,481±29,752μm2
/section; P=0.009) and along the entire aorta by en
face analysis (6.4±1.2 vs 2.3±0.6%; P=0.01) as assessed by Oil-Red O staining (Figure 2A and
B). Quantification of collagen content in aortic lesions using Masson trichrome staining further
revealed decreased necrotic core area in male Cps1+/- mice compared to wildtype littermates
(128,871±9,663μm2
/section; P=0.009) (Figure 2C). These phenotypic differences were
accompanied by significantly lower fasting plasma triglyceride levels in male Cps1+/- mice but
154
not body weight, cholesterol levels, or glucose-related metabolic traits (Table 3). By
comparison, there were no significant differences in atherosclerosis and metabolic traits between
female Cps1+/- mice and wildtype littermates, with the exception of decreased fasting insulin
levels and homeostasis modeling of insulin resistance (HOMA-IR) (Figure 2A-C; Table 3). We
also measured levels of several atherosclerosis-related inflammatory biomarkers, which revealed
significantly decreased interleukin-6 (IL-6) levels in male Cps1+/- mice compared to wildtype
littermates but no differences in female Cps1+/- mice (Table 3).
Effect of Cps1 Deficiency on Amino Acid Metabolism. We next explored whether alterations
in choline and amino acid metabolism could contribute to the sexually dimorphic
atheroprotective phenotype of Cps1+/- mice. With the exception of fasting betaine and
phenylalanine levels in males and dimethylglycine levels in females, these analyses did not
reveal an obvious sexually dimorphic metabolite profile that differed between male or female
Cps1+/- mice and wildtype littermates (Table 2). There were also few metabolomics differences
between Cps1+/- mice and wildtype littermates under non-fasting conditions (Table 2). Thus, the
atheroprotective phenotype observed in male Cps1+/- mice was likely not due to differences in
amino acid or choline metabolism.
Sexually Dimorphic Effects of Cps1 Deficiency in Mice are Independent of Sex Hormone.
To investigate whether the cardioprotective effect Cps1 deficiency was hormone-dependent, we
performed gonadectomy in wildtype C57Bl/6J mice. Surgical removal of the testes and ovaries
in male and female mice, respectively, did not change hepatic Cps1 expression or circulating
levels of urea cycle metabolites (citrulline, arginine, and ornithine) compared to sham-operated
mice (Figure 3). These observations suggest that differences between male and female Cps1+/-
155
mice could be due to genetic differences between the sexes rather than hormonal differences. To
explore this possibility, we examined whether males and females exhibited natural differences in
CPS1 expression. Among 600 CAD patients from the STARNET cohort [41], hepatic mRNA
levels of CPS1 women were ~19% lower compared to men (11.5±1.3 vs. 11.8±1.1 log2 CPM;
P=1.1x10-5
) regardless of rs715 genotype (Figure 4A). We next determined whether the same
sexually dimorphic Cps1 expression pattern existed in wildtype mice of diverse genetic
backgrounds. Among 89 inbred strains from the Hybrid Mouse Diversity Panel (HMDP) [45],
hepatic Cps1 mRNA levels also exhibited a sexually dimorphic pattern but opposite to that
observed in humans. For example, male mice had ~22% reduced expression compared to
females (12.2±0.13 vs. 12.5±0.09 log₂ units; P=2.7x10-51) (Figure 4B). We next carried out the
same analysis in Cps1+/- mice and wildtype littermates. Consistent with the observations in the
HMDP, male wildtype mice had ~50% reduced Cps1 expression compared to wildtype female
mice (Figure 4C). As expected, genetic ablation of one Cps1 allele further reduced hepatic
expression by another 50% in males and females. The same differences in gene expression
between male and female Cps1+/- mice and wildtype littermates were also observed at the protein
level (Figure 4D). These data thus provide an explanation for the sexually dimorphic
association patterns observed as a result of genetically decreased CPS1 activity in both humans
and mice.
156
Discussion
In the present study, we confirmed the sex-differentiated association of CPS1 variant
rs715 with decreased risk of CAD in large multi ancestry meta-analysis with >1.4 million
subjects. These observations provide evidence for one of the few known sexually dimorphic
genetic associations for CAD that exceeds the genome-wide significance threshold in women.
Functional validation studies with a Cps1 knockdown murine model also recapitulated the
sexually dimorphic effects on urea cycle metabolites (lower citrulline and ornithine levels) and
atherogenesis of humans carrying the CPS1-reducing allele of rs715 [19]. However,
gonadectomy experiments in mice did not provide evidence that the sexually dimorphic effects
of genetically decreased CPS1 activity were modulated by sex hormones. These findings suggest
that intrinsic sex differences in hepatic CPS1 expression and/or activity and its downstream
metabolic effects were more likely to underlie the observed associations. For example, naturally
lower basal CPS1 expression coupled with genetic variants that also reduce enzyme activity
could amplify the cardioprotective effects of decreased CPS1 function in women. By contrast,
our data suggest this biological model is opposite in mice, where inherently lower Cps1
expression in males synergizes with heterozygous Cps1 deficiency to confer an atheroprotective
phenotype.
Our findings also provide further insight into the association of rs715 with decreased risk
of CAD. For example, we previously observed this association as being sex-specific and only
among subjects of European ancestry [19]. However, we now demonstrate that this association
is sexually dimorphic in nature since the protective effect of the C allele on CAD risk was
stronger in women compared to men in whom it was still significant but much weaker. In this
regard, prior GWAS analyses identified 10 other sexually dimorphic loci for CAD but none were
157
associated at the genome-wide significance threshold (P=5.0x10-8
) in women and all but one had
stronger associations in men [18]. While sexually dimorphic association of chromosome 9p21
with CAD was previously shown to be genome-wide significant in both men and women [18], to
our knowledge, CPS1 would represent the only known sexually dimorphic locus that is
significantly associated with CAD in women but not men at the genome-wide threshold.
By incorporating participants of Asian, Black/African, and Latino/Admixed American
ancestries, our study further expanded the generalizability of the associations observed at the
CPS1 locus. For example, the sexually dimorphic association of rs715 with decreased CAD was
also evident in subjects of East/South Asian ancestry. Notably, a recent sex-stratified GWAS
analysis in East Asians identified a CPS1 variant (rs1047891) tightly linked to rs715 (r2=0.87)
that was also associated with decreased risk of type 2 diabetes in only women but not men [46].
Although these observations provides a potentially biological link between cardiometabolic
pathways and CPS1 activity, it remains to be determined whether the association of rs715 with
decreased risk of CAD in women is related to its sex-specific association with reduced risk of
type 2 diabetes. Consistent with this sexually dimorphic metabolic association in humans,
heterozygous Cps1 deficiency also improved insulin resistance in female mice. By comparison,
rs715 did not yield association with CAD in analyses with Black-African ancestry subjects. It is
possible that the relatively smaller sample size of this ancestry group in combination with the
lower MAF of rs715 in subjects of Black-African ancestry compared to European ancestry
subjects (20% vs. 31%) limited power to detect association. However, another explanation
could be that rs715 had a weaker effect size in Black-African ancestry subjects. This notion is
supported by previous observations with the major CAD locus on chromosome 9p21 locus where
158
it was shown to not be as strong a predictor of CAD risk in African populations as in White
populations [47, 48].
Our observations in humans also raise the question of the underlying molecular
mechanism of association of rs715 with decreased risk of CAD. In all major ancestry groups,
rs715 is in strong LD with rs1047891 (r2~0.90), which encodes a Thr1405Asn substitution
(ACC>AAC) that localizes to a domain of CPS1 that has been shown to modulate its activation
by N-acetylglutamate (NAG) [49, 50]. Thus, it is possible that rs1047891 is the functional
variant responsible for association of the CPS1 locus with risk of CAD. However, rs715 and
rs1047891 are in relatively weak LD (r2=0.32) in subjects of Black/African ancestry, raising the
possibility that testing rs715 may not have captured association of the causal variant with CAD
in this ancestry group. While this may provide yet another explanation for the lack of a
significant association between rs715 with CAD in Black/African ancestry subjects, other
functional CPS1 variants that are either at higher frequency or only found in this ancestry groups
could also exist that we did not evaluate.
To functionally validate a role of Cps1 in atherosclerosis, we also characterized the
metabolomics and atherosclerosis profile of a genetically modified mouse model. Unexpectedly,
Cps1 deficiency in mice manifested with decreased atherosclerotic plaque burden in males rather
than females. However, this finding was consistent with observations that urea cycle metabolites
were lower in both women carrying the CAD-protective C allele of rs715 and in male Cps1+/-
mice, with the strongest effects on citrulline and increasingly weaker effects on more distal
metabolites. To explore a putative mechanism for these findings, we performed gonadectomy
experiments but did not obtain evidence that the sexually dimorphic associations observed in
159
humans and mice were due to sex hormones. Therefore, we postulated that there could be
natural differences in CPS1 expression/activity between males and females. This notion was
confirmed through hepatic gene expression analyses revealing that women and male mice have
lower CPS1 expression regardless of rs715 genotype or heterozygous Cps1 deficiency,
respectively. For example, among Cps1+/+ and Cps1+/- mice, wildtype females showed the
highest Cps1 expression, followed by Cps1+/-
females and Cps1+/+ males, with Cps1+/- males
showing the lowest expression. This hierarchical pattern thus provides a plausible explanation
for female Cps1+/- mice not manifesting an atheroprotective phenotype since Cps1 expression
may have not been decreased to a sufficiently low level. Finally, while comprehensive profiling
of amino acid metabolism did not provide a putative biological mechanism for why decreased
CPS1 activity would reduce development of atherosclerosis, we note that male Cps1+/- mice did
exhibit lower levels of pro-atherogenic biomarkers, such as plasma triglycerides and IL-6.
In conclusion, a large-scale genetics analysis in humans and functional studies with mice
provided evidence that reduced CPS1 activity protects against development of atherosclerosis.
While the observed sexually dimorphic patterns were opposite between humans and mice, the
directional consistency of the metabolomics and atherosclerosis effects support the concept that
decreased CPS1 has atheroprotective properties. Given that r715 (rs1047891) have been
associated with numerous cardiometabolic risk factors, identifying a precise single mechanism
underlying the association of CPS1 with decreased atherosclerosis in humans or mice may prove
difficult and will require additional studies.
160
References
1. Hodis HN, Mack WJ, Lobo R. Antiatherosclerosis interventions in women. Am J Cardiol.
2002;90(1A):17F-21F.
2. Lobo RA, Davis SR, De Villiers TJ, Gompel A, Henderson VW, Hodis HN, et al. Prevention
of diseases after menopause. Climacteric. 2014;17(5):540-56. doi:
10.3109/13697137.2014.933411.
3. Appelman Y, van Rijn BB, Ten Haaf ME, Boersma E, Peters SA. Sex differences in
cardiovascular risk factors and disease prevention. Atherosclerosis. 2015;241(1):211-8.
doi: 10.1016/j.atherosclerosis.2015.01.027.
4. Martin SS, Aday AW, Almarzooq ZI, Anderson CAM, Arora P, Avery CL, et al. 2024 Heart
Disease and Stroke Statistics: A Report of US and Global Data From the American Heart
Association. Circulation. 2024;149(8):e347-e913. doi: 10.1161/CIR.0000000000001209.
5. Go AS, Mozaffarian D, Roger VL, Benjamin EJ, Berry JD, Blaha MJ, et al. Heart disease and
stroke statistics--2014 update: a report from the American Heart Association. Circulation.
2014;129(3):e28-e292. doi: 10.1161/01.cir.0000441139.02102.80.
6. Shaw LJ, Bugiardini R, Merz CN. Women and ischemic heart disease: evolving knowledge. J
Am Coll Cardiol. 2009;54(17):1561-75. doi: 10.1016/j.jacc.2009.04.098.
7. Wenger NK. Women and coronary heart disease: a century after Herrick: understudied,
underdiagnosed, and undertreated. Circulation. 2012;126(5):604-11. doi:
10.1161/CIRCULATIONAHA.111.086892.
8. Anand SS, Islam S, Rosengren A, Franzosi MG, Steyn K, Yusufali AH, et al. Risk factors for
myocardial infarction in women and men: insights from the INTERHEART study. Eur
Heart J. 2008;29(7):932-40. doi: 10.1093/eurheartj/ehn018.
9. Wienke A, Holm NV, Skytthe A, Yashin AI. The heritability of mortality due to heart
diseases: a correlated frailty model applied to Danish twins. Twin Res. 2001;4(4):266-74.
doi: 10.1375/1369052012399.
10. Zdravkovic S, Wienke A, Pedersen NL, Marenberg ME, Yashin AI, De Faire U. Heritability
of death from coronary heart disease: a 36-year follow-up of 20 966 Swedish twins. J
Intern Med. 2002;252(3):247-54. doi: 10.1046/j.1365-2796.2002.01029.x.
11. Bevan S, Traylor M, Adib-Samii P, Malik R, Paul NL, Jackson C, et al. Genetic heritability
of ischemic stroke and the contribution of previously reported candidate gene and
genomewide associations. Stroke. 2012;43(12):3161-7. doi:
10.1161/STROKEAHA.112.665760.
12. Vinkhuyzen AA, Wray NR, Yang J, Goddard ME, Visscher PM. Estimation and partition of
heritability in human populations using whole-genome analysis methods. Annu Rev
Genet. 2013;47:75-95. doi: 10.1146/annurev-genet-111212-133258.
161
13. Nikpay M, Goel A, Won HH, Hall LM, Willenborg C, Kanoni S, et al. A comprehensive
1,000 Genomes-based genome-wide association meta-analysis of coronary artery disease.
Nat Genet. 2015;47(10):1121-30. doi: 10.1038/ng.3396.
14. van der Harst P, Verweij N. Identification of 64 novel genetic loci provides an expanded
view on the genetic architecture of coronary artery disease. Circ Res. 2018;122(3):433-
43. doi: 10.1161/CIRCRESAHA.117.312086.
15. Koyama S, Ito K, Terao C, Akiyama M, Horikoshi M, Momozawa Y, et al. Populationspecific and trans-ancestry genome-wide analyses identify distinct and shared genetic risk
loci for coronary artery disease. Nat Genet. 2020;52(11):1169-77. doi: 10.1038/s41588-
020-0705-3.
16. Hartiala JA, Han Y, Jia Q, Hilser JR, Huang P, Gukasyan J, et al. Genome-wide analysis
identifies novel susceptibility loci for myocardial infarction. Eur Heart J. 2021;42(9):919-
33. doi: 10.1093/eurheartj/ehaa1040.
17. Tcheandjieu C, Zhu X, Hilliard AT, Clarke SL, Napolioni V, Ma S, et al. Large-scale
genome-wide association study of coronary artery disease in genetically diverse
populations. Nat Med. 2022;28(8):1679-92. doi: 10.1038/s41591-022-01891-3.
18. Aragam KG, Jiang T, Goel A, Kanoni S, Wolford BN, Atri DS, et al. Discovery and
systematic characterization of risk variants and genes for coronary artery disease in over
a million participants. Nat Genet. 2022;54(12):1803-15. doi: 10.1038/s41588-022-01233-
6.
19. Hartiala JA, Tang WH, Wang Z, Crow AL, Stewart AF, Roberts R, et al. Genome-wide
association study and targeted metabolomics identifies sex-specific association of CPS1
with coronary artery disease. Nat Commun. 2016;7:10558. doi: 10.1038/ncomms10558.
20. Sollis E, Mosaku A, Abid A, Buniello A, Cerezo M, Gil L, et al. The NHGRI-EBI GWAS
Catalog: knowledgebase and deposition resource. Nucleic Acids Res. 2023;51(D1):D977-
D85. doi: 10.1093/nar/gkac1010.
21. Bycroft C, Freeman C, Petkova D, Band G, Elliott LT, Sharp K, et al. The UK Biobank
resource with deep phenotyping and genomic data. Nature. 2018;562(7726):203-9. doi:
10.1038/s41586-018-0579-z.
22. Ferrell M, Wang Z, Anderson JT, Li XS, Witkowski M, DiDonato JA, et al. A terminal
metabolite of niacin promotes vascular inflammation and contributes to cardiovascular
disease risk. Nat Med. 2024;30(2):424-34. doi: 10.1038/s41591-023-02793-8.
23. Hartiala J, Li D, Conti DV, Vikman S, Patel Y, Tang WH, et al. Genetic contribution of the
leukotriene pathway to coronary artery disease. Hum Genet. 2011;129(6):617-27.
24. Wang Z, Klipfell E, Bennett BJ, Koeth R, Levison BS, Dugar B, et al. Gut flora metabolism
of phosphatidylcholine promotes cardiovascular disease. Nature. 2011;472(7341):57-63.
162
25. Reiner AP, Hartiala J, Zeller T, Bis JC, Dupuis J, Fornage M, et al. Genome-wide and genecentric analyses of circulating myeloperoxidase levels in the CHARGE and CARe
consortia. Hum Mol Genet. 2013;22(16):3381-93.
26. Hartiala J, Bennett BJ, Tang WH, Wang Z, Stewart AF, Roberts R, et al. Comparative
genome-wide association studies in mice and humans for trimethylamine N-oxide, a
proatherogenic metabolite of choline and L-carnitine. Arterioscler Thromb Vasc Biol.
2014;34(6):1307-13. doi: 10.1161/ATVBAHA.114.303252.
27. Mayo KR, Basford MA, Carroll RJ, Dillon M, Fullen H, Leung J, et al. The All of Us Data
and Research Center: creating a secure, scalable, and sustainable ecosystem for
biomedical research. Annu Rev Biomed Data Sci. 2023;6:443-64. doi: 10.1146/annurevbiodatasci-122120-104825.
28. All of Us Research Program I, Denny JC, Rutter JL, Goldstein DB, Philippakis A, Smoller
JW, et al. The "All of Us" Research Program. N Engl J Med. 2019;381(7):668-76. doi:
10.1056/NEJMsr1809937.
29. All of Us Research Program Genomics I. Genomic data in the All of Us research program.
Nature. 2024;627(8003):340-6. doi: 10.1038/s41586-023-06957-x.
30. Kurki MI, Karjalainen J, Palta P, Sipila TP, Kristiansson K, Donner KM, et al. FinnGen
provides genetic insights from a well-phenotyped isolated population. Nature.
2023;613(7944):508-18. doi: 10.1038/s41586-022-05473-8.
31. Mbatchou J, Barnard L, Backman J, Marcketta A, Kosmicki JA, Ziyatdinov A, et al.
Computationally efficient whole-genome regression for quantitative and binary traits. Nat
Genet. 2021;53(7):1097-103. doi: 10.1038/s41588-021-00870-7.
32. Willer CJ, Li Y, Abecasis GR. METAL: fast and efficient meta-analysis of genomewide
association scans. Bioinformatics. 2010;26(17):2190-1.
33. Birling MC, Yoshiki A, Adams DJ, Ayabe S, Beaudet AL, Bottomley J, et al. A resource of
targeted mutant mouse lines for 5,061 genes. Nat Genet. 2021;53(4):416-9. doi:
10.1038/s41588-021-00825-y.
34. Mehrabian M, Schulthess FT, Nebohacova M, Castellani LW, Zhou Z, Hartiala J, et al.
Identification of ALOX5 as a gene regulating adiposity and pancreatic function.
Diabetologia. 2008;51(6):978-88.
35. Woodward NC, Crow AL, Zhang Y, Epstein S, Hartiala J, Johnson R, et al. Exposure to
nanoscale particulate matter from gestation to adulthood impairs metabolic homeostasis
in mice. Sci Rep. 2019;9(1):1816. doi: 10.1038/s41598-018-37704-2.
36. Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis
model assessment: insulin resistance and beta-cell function from fasting plasma glucose
and insulin concentrations in man. Diabetologia. 1985;28(7):412-9.
163
37. Wang Z, Levison BS, Hazen JE, Donahue L, Li XM, Hazen SL. Measurement of
trimethylamine-N-oxide by stable isotope dilution liquid chromatography tandem mass
spectrometry. Anal Biochem. 2014;455:35-40. doi: 10.1016/j.ab.2014.03.016.
38. Mehrabian M, Allayee H, Wong J, Shi W, Wang XP, Shaposhnik Z, et al. Identification of 5-
lipoxygenase as a major gene contributing to atherosclerosis susceptibility in mice. Circ
Res. 2002;91(2):120-6.
39. Parks BW, Sallam T, Mehrabian M, Psychogios N, Hui ST, Norheim F, et al. Genetic
architecture of insulin resistance in the mouse. Cell Metab. 2015;21(2):334-47. doi:
10.1016/j.cmet.2015.01.002.
40. Norheim F, Hasin-Brumshtein Y, Vergnes L, Chella Krishnan K, Pan C, Seldin MM, et al.
Gene-by-Sex Interactions in Mitochondrial Functions and Cardio-Metabolic Traits. Cell
Metab. 2019;29(4):932-49 e4. doi: 10.1016/j.cmet.2018.12.013.
41. Franzen O, Ermel R, Cohain A, Akers NK, Di Narzo A, Talukdar HA, et al. Cardiometabolic
risk loci share downstream cis- and trans-gene regulation across tissues and diseases.
Science. 2016;353(6301):827-30. doi: 10.1126/science.aad6970.
42. Bennett BJ, Davis RC, Civelek M, Orozco L, Wu J, Qi H, et al. Genetic architecture of
atherosclerosis in mice: A systems genetics analysis of common inbred strains. PLoS
Genet. 2015;11(12):e1005711. doi: 10.1371/journal.pgen.1005711.
43. Schofield JP, Cox TM, Caskey CT, Wakamiya M. Mice deficient in the urea-cycle enzyme,
carbamoyl phosphate synthetase I, die during the early neonatal period from
hyperammonemia. Hepatology. 1999;29(1):181-5. doi: 10.1002/hep.510290112.
44. Gelehrter TD, Snodgrass PJ. Lethal neonatal deficiency of carbamyl phosphate synthetase. N
Engl J Med. 1974;290(8):430-3. doi: 10.1056/NEJM197402212900804.
45. Lusis AJ, Seldin MM, Allayee H, Bennett BJ, Civelek M, Davis RC, et al. The Hybrid
Mouse Diversity Panel: a resource for systems genetics analyses of metabolic and
cardiovascular traits. J Lipid Res. 2016;57(6):925-42. doi: 10.1194/jlr.R066944.
46. Spracklen CN, Horikoshi M, Kim YJ, Lin K, Bragg F, Moon S, et al. Identification of type 2
diabetes loci in 433,540 East Asian individuals. Nature. 2020;582(7811):240-5. doi:
10.1038/s41586-020-2263-3.
47. Tcheandjieu C, Zhu X, Hilliard AT, Clarke SL, Napolioni V, Ma S, et al. Large-scale
genome-wide association study of coronary artery disease in genetically diverse
populations. Nat Med. 2022;28(8):1679-92. doi: 10.1038/s41591-022-01891-3.
48. Lettre G, Palmer CD, Young T, Ejebe KG, Allayee H, Benjamin EJ, et al. Genome-wide
association study of coronary heart disease and its risk factors in 8,090 African
Americans: the NHLBI CARe Project. PLoS Genet. 2011;7(2):e1001300.
164
49. Rodriguez-Aparicio LB, Guadalajara AM, Rubio V. Physical location of the site for Nacetyl-L-glutamate, the allosteric activator of carbamoyl phosphate synthetase, in the 20-
kilodalton COOH-terminal domain. Biochemistry (Mosc). 1989;28(7):3070-4.
50. Britton HG, Garcia-Espana A, Goya P, Rozas I, Rubio V. A structure-reactivity study of the
binding of acetylglutamate to carbamoyl phosphate synthetase I. Eur J Biochem.
1990;188(1):47-53.
165
Acknowledgments
We gratefully acknowledge the UK Biobank Resource for providing access to their data under
Application Number 33307. This study was supported by NIH grants R01HL133169,
R01HL148110, R01HL168493, and U54HL170326. The funders had no role in the design and
conduct of the study; collection, management, analysis, and interpretation of the data; and
preparation, review, or approval of the manuscript.
Data Availability Statement
Individual level data used in the present study are available upon application to the UK Biobank
(https://www.ukbiobank.ac.uk/). All other data supporting the findings of this study are
available either within the article, the Supplementary Information and Supplementary Data files,
or upon reasonable request.
Author Contributions
Concept and design: J.R.H, N.C.W., J.A.H., and H.A. Acquisition, analysis, and interpretation
of data: J.R.H, N.C.W., Z.W., J.G., I.N., W.S.W., P.A., Y.H., S.C., Z.F., C.P., W.H.W.T., S.R.,
A.P., L.M., J.L.M.B., A.J.L., S.L.H., J.A.H, and H.A. Drafting of the manuscript: J.R.H.,
N.C.W., and H.A.. Critical revision of the manuscript for important intellectual content: All
authors.
Competing Interests
Dr. Hazen reports being named as co-inventors on pending and issued patents held by the
Cleveland Clinic relating to cardiovascular diagnostics and therapeutics. Dr. Hazen also reports
166
having received royalty payments for inventions or discoveries related to cardiovascular
diagnostics or therapeutics from Cleveland Heart Lab, a fully owned subsidiary of Quest
Diagnostics, and Procter & Gamble. Dr. Hazen is a paid consultant for Zehna Therapeutics and
Proctor & Gamble, and has received research funds from Zehna Therapeutics, Proctor &
Gamble, Pfizer, and Roche Diagnostics. Dr. Tang serves as a consultant for Sequana Medical,
Cardiol Therapeutics, Genomics plc, Zehna Therapeutics, Boston Scientific, WhiteSwell,
CardiaTec Biosciences, Bristol Myers Squibb, Alexion Pharmaceuticals, Alleviant Medical,
Salubris Biotherapeutics, BioCardia, and has received honoraria from the American Board of
Internal Medicine, Springer Nature, and Belvoir Media Group. All other authors have no known
competing financial interests or personal relationships to declare related to the work reported in
this paper.
167
Figure Legends
Figure 1. The genes and intermediates of the urea cycle. Ammonia is detoxified to
carbamoyl phosphate, which enters the urea cycle through the rate-limiting reaction catalyzed by
carbamoyl-phosphate synthase 1 (CPS1). Carbamoyl phosphate is metabolized by ornithine
transcarbamylase (OTC) to form citrulline and subsequently argininosuccinate through a reaction
catalyzed by argininosuccinate synthetase (ASS). This is followed by the formation of L-arginine
by arginosuccinate lyase (ASL). L-arginine is used as a substrate for the production of nitric
oxide or metabolized by arginase (ARG1) to form urea for excretion and ornithine for re-entry
back into the cycle. Metabolites that were available for analysis are shown in black whereas
unmeasured metabolites are shown in gray.
Figure 2. Characterization of genetically modified mouse model for Cps1 Deficiency. (A)
DNA analysis validates gene targeting of Cps1 locus where a 1032bp deletion spanning exon 3
in results in smaller products in heterozygote (+/-) and homozygote (-/-) mutant newborn pups.
(B) CPS1 protein levels are reduced and absent in livers of mutant newborn pups that are
heterozygous and homozygous for the 1032bp deletion fragment, respectively. (C-E) Adult
Cps1+/
- mice exhibit lower fasting plasma levels of the urea cycle metabolites, citrulline and
ornithine, with a more pronounced effect on the latter in males. Metabolites were measured by
LC/MS as described in the Methods. Data are represented as mean ± SE. n=21-32 mice for data
in panels C-E. P-values were derived from 2-sided t-tests carried out between Cps1+/
- mice and
wildtype (Cps1+/+) littermates separately in males and females.
168
Figure 3. Cps1 Deficiency decreases atherosclerosis in mice through sexually dimorphic
pattern. Compared to male wildtype Cps1+/
+ littermates, male Cps1+/
- mice had significantly
decreased atherosclerotic lesion formation, as assessed through serial cryosections at the aortic
arch (A) or along the entire aorta by en face analysis (B). Representative sections of aortic
lesions (A) and en face aortas stained for lipid content (B) are shown for male Cps1+/
+ and
Cps1+/
- mice. (C) Atherosclerotic lesions of male Cps1+/
- mice also had significantly smaller
necrotic core area, as assessed by Masson trichrome staining, compared to wildtype littermates.
No differences in aortic lesion size or necrotic area were observed in female mice (A-C). Data
are represented as mean ± SE. n=14-22 mice for data in panels A and C, and n=4-7 for data in
panel B. P-values were derived from 2-sided t-tests carried out between Cps1+/
- mice and
wildtype (Cps1+/+) littermates separately in males and females.
Figure 4. Effect of gonadectomy on hepatic Cps1 expression and plasma urea cycle
metabolite levels in mice. Surgical removal of testes in males and ovaries in females did not
alter expression of Cps1 in the liver (A) or fasting plasma levels of citrulline (B), arginine (C),
and ornithine (D). Metabolite levels were measured by LC/MS as described in the Methods.
Data are represented as mean ± SE. n=4 mice for each group in panels A-D. P-values were
derived from 2-sided t-tests carried out between sham-operated and gonadectomized mice
separately in males and females.
Figure 5. Natural hepatic expression of CPS1 differs between male and females in humans
and mice. (A) Liver CPS1 mRNA levels in CAD cases (n=502) and controls (n=97) from the
STARNET cohort [41] were ~19% lower (11.5±1.3 vs. 11.8±1.1 log2 CPM; P=1.1x10-5
) in
women (n=196) compared to men (n=403) regardless of rs715 genotype. Expression levels were
169
adjusted for age, BMI, CAD status, and RNAseq experimental parameters. Data are represented
as mean ± SD of log2 transformed CPM levels. P-values were derived from 2-sided t-tests
carried out between sexes and within each genotype group. (B) An opposite sexually dimorphic
Cps1 expression pattern was observed among inbred mouse strains from the Hybrid Mouse
Diversity Panel (HMDP) [45], with male mice having ~22% reduced expression compared to
females (12.2±0.13 vs. 12.5±0.09 log₂ units; P=2.7x10-51). Expression data are from n=1-3 mice
of each sex for each of the 89 strains. Data are represented as mean ± SE of strain-averaged log₂
transformed values. P-value for difference between males and females was derived from a 2-
sided t-test. (C) Real-time qPCR gene expression analysis demonstrated that hepatic Cps1
mRNA levels were ~50% lower in male Cps1+/+ (wildtype littermates) compared to females, and
further reduced through a sexually dimorphic pattern by genetic ablation of one Cps1 allele in
Cps1+/- mice. n=9-10 mice of each sex and genotype. (D) Western blot analysis revealed the
same sexually dimorphic pattern of differences in hepatic CPS1 protein levels between male
(n=5) and female (n=5) Cps1+/+ (wildtype littermates) and Cps1+/- mice. For panels C and D,
Data are represented as mean ± SE and P-values for differences between male and female
Cps1+/+ (wildtype littermates) and with Cps1+/
- mice for each sex were derived from 2-sided ttests.
170
Table 1: Multi-Ancestry and Sex-Stratified Association of rs715 with Risk of CAD.
Female Male Combined
Ancestry Cohort EAF N
OR
(95% CI)
P-value N
OR
(95% CI)
P-value N
OR
(95% CI)
P-value
UK Biobank 0.31 245,279
0.96
(0.93-
0.99)
0.02 206,706
1.00
(0.98-1.03)
0.67 451,985
0.99
(0.97-
1.01)
0.34
CARDIoGRA
M+CD4
0.31 26,905
0.88
(0.83-
0.94)
6.3x10-5 26,772
1.00
(0.96-1.05)
0.95 53,677
0.95
(0.80-
1.03)
0.12
European GeneBank 0.31 1,438
0.83
(0.69-
0.99)
0.04 3,040
0.98
(0.83-1.16)
0.79 4,478
0.99
(0.98-
1.01)
0.22
FinnGen 0.31 282,051
0.94
(0.92-
0.96)
6.7x10-7 218,277
0.98
(0.96-0.99)
6.9x10-3 500,328
0.95
(0.93-
0.98)
2.4x10-
04
All of Us 0.31 77,681
0.98
(0.94-
1.02)
0.42 52,191
0.99
(0.95-1.02)
0.47 129,872
0.99
(0.96-
1.01)
0.28
MetaAnalysis
0.31 633,354
0.95
(0.93-
0.96)
2.7x10-
10 506,986
0.99
(0.98-1.00)
0.056
1,140,34
0
0.97
(0.96-
0.98)
1.3x10-7
Biobank Japan 0.16 78,008
0.91
(0.87-
0.95)
3.0x10-4 90,220
0.97
(0.94-1.00)
0.04 168,228
0.96
(0.95-
0.98)
5.6x10-7
East/South
Asian
UK Biobank 0.28 5,044
1.01
(0.83-
1.24)
0.91 5,413
0.91
(0.81-1.03)
0.13 10,457
0.94
(0.84-
1.04)
0.21
All of Us 0.23 5,236
1.05
(0.80-
1.39)
0.71 3,499
0.82
(0.65-1.03)
0.09 8,735
0.90
(0.76-
1.08)
0.27
171
MetaAnalysis
0.22 88,288
0.92
(0.88-
0.97)
7.5x10-4 99,132
0.96
(0.93-0.99)
0.011 187,420
0.95
(0.92-
0.97)
6.7x10-5
UK Biobank 0.18 4,033
0.79
(0.57-
1.10)
0.17 3,063
1.06
(0.78-1.44)
0.69 7,096
0.92
(0.74-
1.15)
0.46
Black/
African
All of Us 0.21 31,910
1.02
(0.95-
1.10)
0.61 23,457
1.06
(0.97-1.15)
0.19 55,367
1.03
(0.98-
1.09)
0.23
MetaAnalysis
0.20 35,943
1.01
(0.94-
1.08)
0.85 26,520
1.06
(0.98-1.15)
0.17 62,463
1.03
(0.97-
1.08)
0.33
Latino/
Admixed
American
All of Us 0.29 29,367
0.98
(0.90-
1.07)
0.66 14,680
0.97
(0.89-1.07)
0.58 44,047
0.98
(0.92-
1.04)
0.49
Multiancestry
Combined
meta-analysis
P-het=6.7x10-
5
N/A 786,952
0.95
(0.93-
0.96)
7.3x10-
12 647,318
0.99
(0.97-
0.998)
0.011
1,434,27
0
0.97
(0.96-
0.98)
8.1x10-
10
Data are shown as odds ratios (OR) and 95% confidence intervals (95% CI). Effect sizes refer to C allele of rs715; EAF, effect allele frequency.
172
Table 2. Effect of Cps1 Deficiency on Plasma Metabolite Levels in Mice.
Fasting Non-Fasting
Males Females Males Females
Metabolite (M)
Cps1+/+
(n=21)
Cps1+/-
(n=26) P-value
Cps1+/+
(n=32)
Cps1+/-
(n=27) P-value
Cps1+/+
(n=6)
Cps1+/-
(n=4) P-value
Cps1+/+
(n=11)
Cps1+/-
(n=7) P-value
Citrulline 68 ± 3 46 ± 2 1.4x10-8
91 ± 3 62 ± 2 1.9x10-12 80 ± 5 50 ± 3 2.3x10-
3 117 ± 6 78 ± 9 1.1x10-3
Arginine 26.6 ± 3.2 25 ± 3 0.71 52 ± 5 46 ± 5 0.42 29 ± 9 49 ± 8 0.16 84 ± 11 61 ± 8 0.15
Ornithine 210 ± 11 163 ± 7 5.2x10-4 176 ± 6 169 ± 7 0.41 242 ± 35 138 ± 19 0.053 237 ±
19 197 ± 32 0.26
Choline 25 ± 2 20 ± 1 1.3x10-3 21 ± 1 20 ± 1 0.5588 17 ± 2 10 ± 1 0.031 17± 2 15 ± 1 0.50
TMAO 2.5 ± 0.2 2.4 ± 0.2 0.72 10.6 ± 0.7 11.4 ± 0.6 0.35 2.9 ± 0.3 2.6 ± 0.7 0.72 11.4 ±
1.9
11.5 ±
1.8 0.98
Betaine 52 ± 2 46 ± 1 0.011 66 ± 2 62 ± 3 0.22 38 ± 2 24 ± 2 3.3x10-
3 54 ± 4 47 ± 4 0.28
Dimethylglycine 5.4 ± 0.3 4.7 ± 0.3 0.13 7.2 ± 0.2 6.3 ± 0.2 2.6x10-3 3.0 ± 0.5 1.5 ± 0.2 0.069 3.3± 0.1 3.1 ± 0.3 0.60
Glycine 389 ± 16 384 ± 13 0.78 421 ± 13 410 ± 13 0.52 301 ± 14 303 ± 25 0.94 437 ±
21 368 ± 24 0.056
Creatinine 9.3 ± 0.3 8.7 ± 0.3 0.098 10.9 ± 0.3 10.3 ± 0.3 0.18 8.7 ± 1.3 6.2 ± 0.1 0.14 9.1 ±
0.4 9.2 ± 0.8 0.91
Lysine 101 ± 5 98 ± 4 0.59 134 ± 5 141 ± 6 0.35 151 ± 5 197 ± 22 0.037 295 ±
20 276 ± 18 0.53
Histidine 132 ± 4 131 ± 3 0.80 139 ± 3 137 ± 3 0.68 150 ± 6 152 ± 10 0.85 175 ±
13 158 ± 8 0.37
173
Tryptophan 106 ± 3 112 ± 4 0.28 131 ± 3 136 ± 3 0.32 94 ± 11 71 ± 6 0.16 133 ±
10 138 ± 12 0.78
Serine 129 ± 6 124 ± 4 0.49 136 ± 5 139 ± 5 0.76 138 ± 5 140 ± 14 0.88 214± 13 179 ± 13 0.092
Proline 135 ± 6 123 ± 3 0.088 155 ± 5 152 ± 6 0.72 275 ± 40 322 ± 34 0.42 457 ±
46 385 ± 45 0.31
Valine 111 ± 5 109 ± 3 0.76 122 ± 6 121 ± 6 0.92 165 ± 16 201 ± 14 0.16 218 ±
16 210 ± 19 0.75
Phenylalanine 90 ± 4 80 ± 2 9.7x10-3 96 ± 3 95 ± 2 0.82 95 ± 9 91 ± 4 0.78 108 ±
10 98 ± 7 0.45
Metabolites were measured by LC/MS in plasmas obtained after either a 4 hour fast (fasting) or after two hours into the feeding cycle (nonfasting). P-values between Cps1+/+ and Cps1+/- mice were derived from 2-sided unpaired t-tests, separately in males and females and in fasted
and non-fasting groups. Differences at P<0.05 are highlighted in bold. TMAO, trimethylamine N-oxide
174
Table 3. Effect of Cps1 Deficiency on Fasting Cardiometabolic Traits and Aortic Lesion Formation in Mice.
Males Females
Trait Cps1+/+ Cps1+/- P-value Cps1+/+ Cps1+/- P-value
Body weight (g) 39.1 ± 0.8
(n=20)
37.7 ± 0.9
(n=24) 0.26 27.6 ± 0.7 (n=33) 27.4 ± 0.7
(n=29) 0.88
Total cholesterol (mg/dL) 1404 ± 68
(n=23)
1272 ± 79
(n=29) 0.22 851 ± 34
(n=34)
764 ± 43
(n=29) 0.11
LDL/VLDL cholesterol
(mg/dL)
1312 ± 70
(n=23)
1171 ± 84
(n=29) 0.22 779± 34 (n=34) 693 ± 46
(n=29) 0.13
HDL cholesterol (mg/dL) 93 ± 4
(n=23)
101 ± 7
(n=29) 0.35 73 ± 3 (n=34) 71 ± 4
(n=29) 0.72
Triglycerides (mg/dL) 180 ± 24
(n=23)
123 ± 11
(n=29) 0.027 45 ± 9 (n=34) 30 ± 2
(n=29) 0.14
Glucose (mg/dL) 222 ± 11 (n=23) 208 ± 6.0 (n=29) 0.23 173± 7 (n=34) 170 ± 6 (n=29) 0.70
Insulin (pg/mL) 1829 ± 141 (n=23) 1615 ± 132 (n=29) 0.28 727 ± 78 (n=34) 453 ± 49 (n=29) 5.6x10-3
aHOMA-IR 29.1 ± 2.3 (n=23) 24.6 ± 2.4 (n=29) 0.20 8.9 ± 1.0 (n=34) 5.4 ± 0.6 (n=29) 4.0x10-3
Il-6 (pg/mL) 77 ± 16 (n=14) 28.1 ± 5.4 (n=14) 7.6x10-3 26 ± 7.5 (n=10) 23 ± 4.3 (n=10) 0.70
Il1-b (pg/mL) 1.4 ± 0.7 (n=13) 0.39 ± 0.11 (n=10) 0.23 0.96 ± 0.37 (n=3) 0.69 ± 0.23 (n=6) 0.53
TNF-a (pg/mL) 31.5 ± 3.7 (n=14) 25.7 ± 2.8 (n=14) 0.22 20.6 ± 4.0 (n=10) 20.2 ± 1.5 (n=10) 0.95
INF-g (pg/mL) 0.29 ± 0.05 (n=13) 0.41 ± 0.18
(n=12) 0.47 0.34 ± 0.07 (n=9) 0.39 ± 0.06
(n=9) 0.59
175
Aortic lesion area
(m2
/section)
452,855 ± 23,550
(n=15)
343,481 ± 29,752
(n=18)
8.7x10-3
582,311 ± 29,748
(n=19)
550,103 ± 36,020
(n=14)
0.49
En Face lesion size (%) 6.4 ± 1.2 (n=4) 2.3 ± 0.6
(n=6)
0.011 3.7 ± 0.3 (n=7)
2.9 ± 0.6
(n=4)
0.19
Necrotic core area
(m2
/section) 12,8871 ± 9,663 (n=15) 76,711 ± 10,442
(n=19)
1.1x10-3
142,508 ± 12,215
(n=22)
153,353 ± 13,266
(n=18)
0.55
dCalculated according to the formula: [glucose mg/dL) X insulin (mIU/ml)]/405.
176
Figure 1
177
Figure 2
178
Figure 3
179
Figure 4
180
Figure 5
181
CHAPTER 4: EFFECT OF MENOPAUSAL HORMONE THERAPY ON METHYLATION LEVELS
IN EARLY AND LATE POSTMENOPAUSAL WOMEN
Summary. In the present study, we investigated the role of DNA methylation in explaining the
cardiovascular protective effects of hormone therapy (HT) in women who were within six years
of menopause compared to those who were 10 or more years from menopause Using data from
the Early versus Late Intervention Trial with Estradiol (ELITE), we conducted an epigenomewide association study (EWAS) to investigate whether changes in DNA methylation percentages
could provide mechanistic insights into the differences in CAD risk observed between women
who initiated hormone therapy (HT) within six years of menopause compared to those who
started HT 10 or more years post-menopause. Over a 36-month period, we analyzed methylation
changes at over 700,000 CpG sites and identified two differentially methylated CpG sites
associated with the use of HT in women who were within six years of menopause compared to
those who were 10 or more years post-menopause. Notably, these CpG sites did not correlate
with changes in carotid intima-media thickness, suggesting that their methylation changes are not
directly linked to measurable subclinical atherosclerosis. My contributions included preparing
the buffy coat samples for methylation analysis, curating, and processing the methylation data,
performing statistical analyses, and contributing to manuscript preparation. This study was
published in the Clinical Epigenetics (Hilser, J.R., Hartiala, J.A., Sriprasert, I. et al. Effect of
menopausal hormone therapy on methylation levels in early and late postmenopausal women.
Clin Epigenet 14, 90 (2022). https://doi.org/10.1186/s13148-022-01311-w).
182
Effect of Menopausal Hormone Therapy on Methylation Levels in Early and Late
Postmenopausal Women
James R. Hilser1,2, Jaana A. Hartiala1
, Intira Sriprasert3
, Naoko Kono1,5, Zhiheng Cai1,2, Roksana
Karim1,5, Joseph DeYoung6
, Wendy J. Mack1,5, Howard N. Hodis1,4,5, and Hooman Allayee1,2*
Departments of 1Population and Public Health Sciences, 2Biochemistry & Molecular Medicine,
3Obstetrics and Gynecology, and 4Medicine; 5Atherosclerosis Research Unit; Keck School of
Medicine, University of Southern California, Los Angeles, CA 90033; 6Center for
Neurobehavioral Genetics, Semel Institute for Neuroscience and Human Behavior, Department
of Psychiatry and Biobehavioral Sciences, David Geffen School of Medicine of UCLA, Los
Angeles, CA 90095.
Keywords: methylation, hormone therapy, menopause, atherosclerosis.
183
Abstract
Background. Cardiovascular disease (CVD) remains the leading cause of death among
postmenopausal women but standard primary prevention strategies in women are not as effective
as in men. By comparison, the Early versus Late Intervention Trial with Estradiol (ELITE)
study demonstrated that hormone therapy (HT) was associated with significant reduction in
atherosclerosis progression in women who were within six years of menopause compared to
those who were 10 or more years from menopause. These findings are consistent with other
studies showing significant reductions in all-cause mortality and CVD with HT, particularly
when initiated in women younger than 60 years of age or within 10 years since menopause. To
explore the biological mechanisms underlying the age-related atheroprotective effects of HT, we
investigated changes in methylation of blood cells of postmenopausal women who participated in
ELITE.
Results. We first validated the epigenetic data generated from blood leukocytes of ELITE
participants by replicating previously known associations between smoking and methylation
levels at previously identified CpG sites, such as cg05575921 at the AHRR locus. An
epigenome-wide association study (EWAS) evaluating changes in methylation through
interactions with time-since-menopause and HT revealed two significantly associated CpG sites
on chromosomes 12 (cg19552895; p=1.1x10-9
) and 19 (cg18515510; p=2.4x10-8
). Specifically,
HT resulted in modest, but significant, increases in methylation levels at both CpGs but only in
women who were 10 or more years since menopause and randomized to HT. Changes in carotid
artery intima-media thickness (CIMT) from baseline to 36 months after HT were not
significantly correlated with changes in methylation levels at either cg19552895 or cg18515510.
184
Evaluation of other previously identified CpG sites at which methylation levels in either blood or
vascular tissue were associated with atherosclerosis also did not reveal any differences in
methylation as a function of HT and time-since-menopause or with changes in CIMT.
Conclusions. We identified specific methylation differences in blood in response to HT among
women who were 10 or more years since menopause. The functional consequence of these
change with respect to atherosclerosis progression and protective effects of HT remains to be
determined and will require additional studies.
185
Introduction
Cardiovascular disease (CVD) remains the leading cause of death among postmenopausal
women [1]. Although standard primary prevention with statins, aspirin, and ACE inhibitors
significantly reduces CVD risk in men, the cardioprotective effects in women are less certain
with no reduction in all-cause mortality [2-8]. By comparison, hormone therapy (HT) has been
shown to reduce all-cause mortality and CVD in primary prevention, when initiated in women
younger than 60 years or who are less than 10 years since menopause [9-11]. These metaanalyses of randomized controlled trials are consistent with results from the randomized, doubleblinded, placebo-controlled Early versus Late Intervention Trial with Estradiol (ELITE) that
specifically tested the effect of HT on subclinical atherosclerosis as a function of time-sincemenopause [12]. ELITE confirmed the HT timing hypothesis by demonstrating that, compared
with placebo, carotid artery intima media thickness (CIMT) progression was significantly
lowered by HT when initiated within 6 years of menopause but had no effect on CIMT
progression when initiated 10 years or more after menopause [12]. Thus, results of ELITE
clearly indicated a sex-specific and age-related opportunity for reducing CVD and all-cause
mortality trends in women.
Despite the benefits of HT shown in ELITE and other studies, biological mechanisms
underlying the age-related atheroprotective effects of HT remain unknown and cannot be
completely explained by effects on known risk factors. Possible explanations for these
observations at the molecular level may be related to epigenetic modification, expression, and/or
signaling of estrogen receptors (ESRs) in atherosclerosis-related tissues as a function of aging
and/or time-since-menopause. For example, in women, estradiol has been shown to upregulate
ESR1 and ESR2 mRNA levels in leukocytes, such as macrophages and neutrophils [13, 14], and
186
age-related increases in methylation of CpG islands in the promoters of both ESR1 and ESR2 has
been observed in atherosclerotic and normal vascular tissue [15, 16] as well as in proliferating
smooth muscle cells that are characteristically found in atherosclerotic lesions [17]. The
promoter regions of genes in other atherosclerosis relevant genes, such as the pro-inflammatory
enzyme 15-lipoxygenase, have also exhibited significantly decreased methylation in advanced
human atherosclerotic lesions compared with fatty streaks, which was accompanied by abundant
15-lipoxygenase mRNA levels [18].
In the present study, we sought to explore the potential molecular mechanisms by which
HT decreased subclinical atherosclerosis progression among participants of ELITE. An unbiased
epigenome-wide association study (EWAS) was carried out to identify CpG sites at which
methylation levels changed in response to HT as function of time-since-menopause. Candidate
loci were further evaluated bioinformatically and for association with CIMT progression.
187
Materials and Methods
Study Population. ELITE was a single-center, randomized, double-blind, placebo-controlled
clinical trial (ClinicalTrials.gov number NCT00114517) testing effects of HT on progression of
subclinical atherosclerosis as a function of time-since-menopause. Participants were healthy
postmenopausal women without diabetes and clinical evidence of CVD who had no menses for
at least 6 months or who had surgically induced menopause, as well as a serum estradiol level
lower than 25pg/mL (92pmol/L). Women in whom time-since-menopause could not be
determined, or who had fasting plasma TG levels >500 mg/dL, diabetes mellitus or fasting serum
glucose levels >140 mg/dL, serum creatinine level >2.0 mg/dL, uncontrolled hypertension,
untreated thyroid disease, life-threatening disease with prognosis <5 years, a history of deep vein
thrombosis, pulmonary embolism, breast cancer, or current use of postmenopausal HT within 1
month of screening were excluded. A total of 643 women were stratified according to timesince-menopause (<6 years [early] or >10 years [late]) and randomized to receive either HT or
placebo using a 1:1 ratio of stratified blocked randomization, resulting in four treatment groups:
early/placebo, early/HT, late/placebo, and late/HT. HT consisted of oral micronized 17bestradiol 1mg/day with 4% vaginal micronized progesterone gel 45mg/day for 10 days each
month (among women with intact uterus). ELITE demonstrated that, compared with placebo,
HT reduced CIMT progression in women who were within six years of menopause but not
women who were 10 or more years from menopause [12]. Additional details on the design,
methods, and results of the trial have been described previously [12, 19]. ELITE was approved
by the University of Southern California institutional review board and all participants provided
written informed consent.
188
Whole-genome DNA Methylation Profiling. To maximize the likelihood of identifying
associations between CpG sites and HT as a function of time-since-menopause, we selected
subset of 48 women from the early/HT with the lowest rate of CIMT progression over the course
of the trial for methylation profiling and an equivalent number of women from each of the
early/placebo, late/placebo, and late/HT groups with the highest CIMT progression. Genomic
DNA was extracted from buffy coats of this subset of 192 ELITE participants obtained at
baseline and 36 months following randomization using DNeasy kits (Qiagen, Valencia, CA) and
bisulfite treated with the Zymo EZ DNA Methylation Kit (Zymo Research, Orange, CA).
Quantitative levels of DNA methylation were obtained for >850,000 CpG sites using the
Infinium Human Methylation EPIC BeadChip (Illumina, San Diego, CA) according to the
manufacturer’s protocols.
Methylation Data Processing and Normalization. Prior to analysis, the meffil package in R
[20] was used to carry out several quality control (QC) steps, including filtering of samples and
CpG sites, identifying batch effects, and normalizing sample quantiles. Meffil is a
comprehensive and integrated toolkit that utilizes multiple previously developed R packages for
methylation analysis, such as minfi [21], illuminaio [22], and noob [23]. Background and dyebias correction was first performed using raw probe signal intensities as the input. The noob
background normalization method [23] was used to account for technical variation in
background fluorescence signal, which capitalizes on a new use for the Infinium I design bead
types to measure nonspecific fluorescence in the color channel opposite of their design
(Cy3/Cy5). Poor quality CpGs were removed using the illuminaio R package [22] and the
ChAMP R package [24-26] was used to identify and exclude SNP-related CpG probes based on
previously reported annotations [27]. Low quality samples were removed if they were outliers
189
for methylated/unmethylated levels or control probe means, had too many undetected probes or
low bead number probes [20]. Functional normalization (FN) as implemented with the minfi R
package [21] was used to minimize technical variation based on control probes present in the
EPIC BeadChip that do not exhibit biological variation and whose only source of variation is due
to technical artifacts. FN was also used to identify the number of principal components (PCs) of
methylation matrix to include in the normalization that minimizes the residual variance
unexplained by the given number of PCs, and to remove technical artifacts by normalizing
sample quantiles using additional fixed and random effects [20]. Quantile normalization was
performed using meffil where slide, plate, and array were treated as random effects, and the first
10 PCs were included as fixed effects. The Houseman algorithm [28] as implemented in meffil
was used to estimate fractions of six different white blood cell populations (B cells, CD4 T cells,
CD8 T cells, granulocytes, monocytes, and natural killer cells) using GSE35069 as the cell type
reference [29]. Leukocyte fraction estimates were subsequently used as covariates in the EWAS
analyses. Methylation levels (β values) at each CpG site were determined by calculating the
ratio of fluorescence intensities between methylated (signal A) and un-methylated (signal B)
sites using the formula β = Max(M,0)/[Max(M,0) + Max(U,0) + 100]. Thus, β values range from
0 (completely un-methylated) to 1 (completely methylated). Prior to final analysis, β values
were transformed to M-values (log2 ratio of methylated vs. unmethylated probe) using ‘beta2m’
function in the lumi package in R [30]. The final dataset included 186 ELITE participants in
whom methylation data at 748,567 CpG sites were available at both visits (total of 372
methylation profiles at baseline and 36 months after trial randomization).
Differential Methylation Analysis. EWAS analyses with smoking were carried out at the
baseline and 36-month visits out using linear regression models that were fitted using limma [31]
190
as implemented in meffil [20]. Participants were categorized as never, former, and current
smokers (coded as 0, 1, 2) and methylation M-values were compared across categories using an
analysis pipeline in meffil with adjustment for age, ethnicity, and estimated blood cell fractions.
Longitudinal EWAS analyses with epigenetic data at both the baseline and 36-month visits were
used to investigate the effect of HT (treated vs. placebo) and time-since-menopause (<6 years
[early] or ≥10 years [late]) on methylation levels. Changes in methylation at CpG sites were
tested using lmrse package in R [32] designed to fit linear models with cluster robust standard
errors across high-dimensional data to evaluate methylation trajectories. Participants for
repeated measures analysis were categorized into four groups based on early or late postmenopause and randomized to HT or placebo. P-values for methylation changes at each CpG
site were obtained from tests of interaction between these latter four categories and a time
variable (baseline vs. 36 months) with adjustment for age, ethnicity, and estimated blood cell
fractions.
Measurement of Subclinical Atherosclerosis and Free Estradiol Levels. Rate of change in
far wall intima–media thickness of the right distal common carotid artery was assessed by
computer image processing of B-mode ultrasonograms. At baseline, two examinations were
conducted (averaged to obtain baseline CIMT values) and every 6 months during trial follow-up
[12]. High-resolution B-mode ultrasonographic imaging and CIMT measurements were
performed with use of standardized procedures and in-house technology that was specifically
developed for longitudinal measurements of changes in atherosclerosis [19]. Coefficient of
variation for baseline CIMT measurements was 0.69% [12]. Plasma estradiol levels were
measured at baseline and 36 months after treatment by radioimmunoassay with preceding
organic solvent extraction and Celite column partition chromatography, as described previously
191
[33]. The relationship between changes in CIMT progression or free estradiol levels [34] and
methylation levels (M-values) at the two CpGs identified through interactions with HT and timesince-menopause was assessed in the late/HT group using partial Spearman’s correlation with
adjustment for age and estimated blood cell fractions at baseline and at 36 months after
treatment.
192
Results
Characteristics of the Study Population. Based on the findings of ELITE [12], we designed a
study to maximize the likelihood of identifying differentially methylated CpG sites as a function
of time-since-menopause and HT. We selected subset of 48 women from the early/HT with the
lowest 36-month rate of CIMT progression and an equivalent number of women from each of the
early/placebo, late/placebo, and late/HT groups with the highest CIMT progression for the
present analysis. As shown in Table 1, there were no significant differences in baseline clinical
or demographic characteristics across the four study groups apart from expected differences in
age, years since menopause, and progression of CIMT after 36 months of HT.
Association of Methylation with Smoking. As an initial step in our analyses, we validated
methylation data generated from buffy coat-derived blood leukocytes of all selected ELITE
participants by carrying out an EWAS for smoking. We chose this exposure as a representative
outcome since previous studies have identified strong and reproducible methylation signals in
blood DNA at multiple sites associated with smoking [35]. Consistent with prior studies,
comparisons between never, former, and current smokers revealed associations between smoking
and methylation levels at both baseline and 36 months post HT with several of the five CpG sites
previously reported to be most strongly affected by smoking (Table 2). At these CpGs, smoking
was associated with decreased methylation levels, which is directionally consistent with the
observed effects of smoking in numerous previous studies [35]. In particular, the association
signals at cg05575921 in intron 3 of the aryl hydrocarbon receptor repressor gene (AHRR) on
chromosome 5 exceeded the Bonferroni-corrected genome-wide significance threshold for
testing 748,567 CpGs (p=0.05/748,567=6.7x10-8
) (Figure 1; Table 2). Nominally significant
193
(p<0.05) associations between methylation levels and smoking were also observed at three of the
four other selected smoking-associated CpG sites at baseline and 36 months after treatment,
although only cg19859270 and cg03636183 are considered significant at a Bonferroni threshold
for testing five CpGs (p=0.05/5=0.01; Table 2). Thus, these data validate the quality and
suitability of the methylation data generated from buffy coats of ELITE participants for EWAS
analyses.
Differential Methylation as Function of Time-Since-Menopause and HT. We next carried
out an EWAS analysis to identify differentially methylated regions of the genome associated
with time-since-menopause and HT. Two CpG sites located on chromosomes 12 (cg19552895;
p-int=1.1x10-9
) and 19 (cg18515510; p-int=2.4x10-8
) yielded genome-wide significant p-values
(p<6.7x10-8
) for interaction of time and the time-since-menopause/HT groups, indicating
differences in the 36-month changes in methylation in the four groups (Figure 2A).
Cg19552895 maps to a shelf region downstream of WNT1 whereas cg18515510 is located in the
3’ UTR of CLEC4M (Figure 2B and C). The interactions between time-since-menopause and
HT for methylation differences between baseline and 36 months after treatment at both CpGs
were based on a ~2% increase after treatment in only the late postmenopausal HT group (Table
3; Figure 3). In addition to cg19552895 and cg18515510, numerous other CpG sites distributed
throughout the genome also yielded suggestive (p-int<6.7x10-6
) interactions with time-sincemenopause and HT (Figure 2A).
Association Between Differentially Methylated CpGs and CIMT Progression. We next
evaluated whether the increased methylation at cg19552895 and cg18515510 as a result of HT
was associated with CIMT progression. Among women in the late/HT group, changes in CIMT
194
from baseline to 36 months after HT were not significantly correlated with differences in
methylation levels at either CpG (Figure 4). A similar analysis with changes in free estradiol
levels from baseline to 36 months post-treatment also did not reveal a relationship with changes
in methylation levels at cg19552895 (r=-0.18; P=0.29). Although there was a modest positive
correlation with cg18515510 (r=0.36; P=0.033), it would not be considered significant at a
Bonferroni-corrected p-value for testing 2 CpGs (p=0.05/2=0.025). Lastly, we also evaluated
other previously identified CpG sites at which methylation levels in either blood or vascular
tissue were associated with atherosclerosis, including those related to smoking [36-38].
However, none of the selected regions exhibited differences in methylation among women in the
four treatment groups that would be considered significant for the number of CpG sites tested
(Table 4).
195
Discussion
In the present study, we sought to determine whether changes in methylation levels could
represent at least one molecular mechanism for the protective effects of HT on subclinical
atherosclerosis as a function of time-since-menopause observed in ELITE [12]. A longitudinal
EWAS analysis among a subset of ELITE participants with the lowest and highest 36-month
CIMT progression identified two CpG sites on chromosomes 12 and 19 that exhibited highly
significant interactions between methylation levels, time-since-menopause, and HT.
Specifically, HT resulted in modest, but significant, increases of methylation at both CpGs in the
late/HT group but had no effect in the other three groups of women. This was a somewhat
surprising finding since we hypothesized that methylation changes in response to HT would be
more pronounced in the early/HT group, given that the atheroprotective effects of HT on CIMT
progression were only observed among these women [12]. However, we previously observed
that estradiol levels were differentially associated with atherosclerosis progression according to
timing of HT initiation. For example, CIMT progression rate was decreased with higher
estradiol levels among women in early post-menopause but increased among women in the late
post-menopause treatment group [33], suggesting potentially adverse effects of HT in this latter
group. Alternatively, it is possible that the effects of HT on methylation of leukocytes blood are
stronger in women who are 10 or more years since menopause. In this regard, studies have
suggested that HT can have certain adverse biological effects in women when initiated in those
who are further from menopause [10], which could be reflected at the molecular level by
epigenetic modifications in blood cells. Follow up studies will be required to address this
possibility.
196
The two CpG sites at which methylation levels increased in response to HT in the late
post-menopause group were located near WNT1 (cg19552895) and in CLEC4M (cg18515510).
While an obvious connection between these genes, HT, atherosclerosis, and methylation changes
at either CpG site in blood cells is not presently evident, prior studies have shown that 17bestradiol can both induce and decrease methylation at CpG sites through ESR1-mediated
mechanisms [39]. These observations have mostly been in the context of breast cancer and it is
not known whether ESR1-mediated effects of 17b-estradiol on methylation occurs in blood
leukocytes as well. However, the WNT family of signaling molecules have been implicated in
various aspects of CVD, including cellular cholesterol homeostasis [40], and circulating levels of
WNT1 protein have been reported to be lower in premature myocardial infarction patients than
controls during both the acute and stable phases [41]. By comparison, CLEC4M is a member of
the C-type lectin gene family expressed primarily on endothelial cells in liver and lymph nodes,
and plays a role in promoting cellular entry of various viruses [42]. Despite these observations,
it is not clear whether the increased methylation observed at cg19552895 and cg18515510 in
response to HT would alter expression of WNT1 and CLEC4M, respectively, in blood cells and,
more broadly, how the biological function of each protein is directly related to the effects of HT
in ELITE participants.
Although strong interactions between HT and time-since-menopause were observed with
cg19552895 or cg18515510, we did not obtain evidence for a statistically significant relationship
between methylation changes at these CpG sites and changes in free estradiol levels or CIMT
progression in the late/HT group. One possible explanation for these observations is that the
effects of free estradiol and HT on methylation directly at the level of the vessel wall may not be
reflected by epigenetic changes in blood cells. For example, a large EWAS analysis with >6400
197
individuals also did not identify any genome-wide significant associations between CIMT and
blood cell-derived methylation levels at any CpG site except for cg05575921 at the well-known
smoking-associated AHRR locus [38]. By comparison, epigenetic analyses with tissue obtained
from atherosclerotic and normal carotid, aortic, mammary, or femoral artery samples have
identified thousands of associations with atherosclerosis [36-38, 43-45]. Interestingly, several
studies observed hypomethylation of CpG sites in atherosclerotic tissue compared to normal
arteries and upregulation of multiple pathways that could potentially be causal drivers of plaque
development [43, 45]. These findings, taken together with our data, suggest that circulating
leukocytes may not be an appropriate surrogate tissue in which to associate methylation
modifications with vascular wall phenotypes, even in response to HT. Interestingly, efforts are
underway to determine whether methylation profiling of peripheral tissues, such as blood, can
provide insight into epigenetic patterns in other tissues [46], which could be applied to epigenetic
studies of atherosclerosis-related traits.
While revealing potentially novel epigenetic associations with HT among
postmenopausal women, our study should be considered in the context of its limitations. For
example, our intention by selecting subsets of women with the lowest CIMT progression from
early/HT group, in which the atheroprotective effects of HT were only observed, was to
maximize potential molecular differences in response to HT compared to participants from the
other three treatment groups. However, this strategy, while efficient, may not have provided
sufficient power to detect associations with CIMT due to weak biological effects, particularly
since our analyses specifically sought to identify interactions between methylation levels, HT,
and time-since-menopause. In addition, although blood cells are the most readily accessible
198
tissue from humans for methylation analyses, they may not, as noted above, adequately reflect
atherosclerotic mechanisms at the level of the vessel wall.
In summary, we evaluated the effects of HT on epigenetic modifications of blood cells in
women as a function of time-since-menopause. In addition to replicating previously described
associations between smoking and methylation, our study also identified a small number CpG
sites at which methylation levels increased in response to HT but only among women who were
10 or more years from menopause. These data represent, to our knowledge, one of the first
descriptions of the effects of HT on methylation profiles of blood cells and provide a unique
hypothesis-generating dataset that can be leveraged for future studies.
199
References
1. Virani SS, Alonso A, Benjamin EJ, Bittencourt MS, Callaway CW, Carson AP, et al. Heart
Disease and Stroke Statistics-2020 Update: A Report From the American Heart
Association. Circulation. 2020;141(9):e139-e596. doi: 10.1161/CIR.0000000000000757.
2. Shekelle PG, Rich MW, Morton SC, Atkinson CS, Tu W, Maglione M, et al. Efficacy of
angiotensin-converting enzyme inhibitors and beta-blockers in the management of left
ventricular systolic dysfunction according to race, gender, and diabetic status: a metaanalysis of major clinical trials. J Am Coll Cardiol. 2003;41(9):1529-38. doi:
10.1016/s0735-1097(03)00262-6.
3. Wing LM, Reid CM, Ryan P, Beilin LJ, Brown MA, Jennings GL, et al. A comparison of
outcomes with angiotensin-converting--enzyme inhibitors and diuretics for hypertension
in the elderly. N Engl J Med. 2003;348(7):583-92. doi: 10.1056/NEJMoa021716.
4. Walsh JM, Pignone M. Drug treatment of hyperlipidemia in women. JAMA.
2004;291(18):2243-52. doi: 10.1001/jama.291.18.2243.
5. Ridker PM, Cook NR, Lee IM, Gordon D, Gaziano JM, Manson JE, et al. A randomized trial
of low-dose aspirin in the primary prevention of cardiovascular disease in women. N
Engl J Med. 2005;352(13):1293-304. doi: 10.1056/NEJMoa050613.
6. Berger JS, Roncaglioni MC, Avanzini F, Pangrazzi I, Tognoni G, Brown DL. Aspirin for the
primary prevention of cardiovascular events in women and men: a sex-specific metaanalysis of randomized controlled trials. JAMA. 2006;295(3):306-13. doi:
10.1001/jama.295.3.306.
7. Petretta M, Costanzo P, Perrone-Filardi P, Chiariello M. Impact of gender in primary
prevention of coronary heart disease with statin therapy: a meta-analysis. Int J Cardiol.
2010;138(1):25-31. doi: 10.1016/j.ijcard.2008.08.001.
8. Brugts JJ, Yetgin T, Hoeks SE, Gotto AM, Shepherd J, Westendorp RG, et al. The benefits of
statins in people without established cardiovascular disease but with cardiovascular risk
factors: meta-analysis of randomised controlled trials. BMJ. 2009;338:b2376. doi:
10.1136/bmj.b2376.
9. Salpeter SR, Walsh JM, Greyber E, Salpeter EE. Brief report: Coronary heart disease events
associated with hormone therapy in younger and older women. A meta-analysis. J Gen
Intern Med. 2006;21(4):363-6. doi: 10.1111/j.1525-1497.2006.00389.x.
10. Boardman HM, Hartley L, Eisinga A, Main C, Roque i Figuls M, Bonfill Cosp X, et al.
Hormone therapy for preventing cardiovascular disease in post-menopausal women.
Cochrane Database Syst Rev. 2015(3):CD002229. doi:
10.1002/14651858.CD002229.pub4.
200
11. Salpeter SR, Walsh JM, Greyber E, Ormiston TM, Salpeter EE. Mortality associated with
hormone replacement therapy in younger and older women: a meta-analysis. J Gen Intern
Med. 2004;19(7):791-804. doi: 10.1111/j.1525-1497.2004.30281.x.
12. Hodis HN, Mack WJ, Henderson VW, Shoupe D, Budoff MJ, Hwang-Levine J, et al.
Vascular effects of early versus late postmenopausal treatment with estradiol. N Engl J
Med. 2016;374(13):1221-31. doi: 10.1056/NEJMoa1505241.
13. Molero L, Garcia-Duran M, Diaz-Recasens J, Rico L, Casado S, Lopez-Farre A. Expression
of estrogen receptor subtypes and neuronal nitric oxide synthase in neutrophils from
women and men: regulation by estrogen. Cardiovasc Res. 2002;56(1):43-51. doi:
10.1016/s0008-6363(02)00505-9.
14. Murphy AJ, Guyre PM, Wira CR, Pioli PA. Estradiol regulates expression of estrogen
receptor ERalpha46 in human macrophages. PLoS One. 2009;4(5):e5539. doi:
10.1371/journal.pone.0005539.
15. Post WS, Goldschmidt-Clermont PJ, Wilhide CC, Heldman AW, Sussman MS, Ouyang P, et
al. Methylation of the estrogen receptor gene is associated with aging and atherosclerosis
in the cardiovascular system. Cardiovasc Res. 1999;43(4):985-91. doi: 10.1016/s0008-
6363(99)00153-4.
16. Kim J, Kim JY, Song KS, Lee YH, Seo JS, Jelinek J, et al. Epigenetic changes in estrogen
receptor beta gene in atherosclerotic cardiovascular tissues and in-vitro vascular
senescence. Biochim Biophys Acta. 2007;1772(1):72-80. doi:
10.1016/j.bbadis.2006.10.004.
17. Ying AK, Hassanain HH, Roos CM, Smiraglia DJ, Issa JJ, Michler RE, et al. Methylation of
the estrogen receptor-alpha gene promoter is selectively increased in proliferating human
aortic smooth muscle cells. Cardiovasc Res. 2000;46(1):172-9. doi: 10.1016/s0008-
6363(00)00004-3.
18. Hiltunen MO, Turunen MP, Hakkinen TP, Rutanen J, Hedman M, Makinen K, et al. DNA
hypomethylation and methyltransferase expression in atherosclerotic lesions. Vasc Med.
2002;7(1):5-11. doi: 10.1191/1358863x02vm418oa.
19. Hodis HN, Mack WJ, Shoupe D, Azen SP, Stanczyk FZ, Hwang-Levine J, et al. Methods and
baseline cardiovascular data from the Early versus Late Intervention Trial with Estradiol
testing the menopausal hormone timing hypothesis. Menopause. 2015;22(4):391-401.
doi: 10.1097/GME.0000000000000343.
20. Min JL, Hemani G, Davey Smith G, Relton C, Suderman M. Meffil: efficient normalization
and analysis of very large DNA methylation datasets. Bioinformatics. 2018;34(23):3983-
9. doi: 10.1093/bioinformatics/bty476.
21. Fortin JP, Labbe A, Lemire M, Zanke BW, Hudson TJ, Fertig EJ, et al. Functional
normalization of 450k methylation array data improves replication in large cancer
studies. Genome biology. 2014;15(12):503. doi: 10.1186/s13059-014-0503-2.
201
22. Smith ML, Baggerly KA, Bengtsson H, Ritchie ME, Hansen KD. illuminaio: An open source
IDAT parsing tool for Illumina microarrays. F1000Res. 2013;2:264. doi:
10.12688/f1000research.2-264.v1.
23. Triche TJ, Jr., Weisenberger DJ, Van Den Berg D, Laird PW, Siegmund KD. Low-level
processing of Illumina Infinium DNA Methylation BeadArrays. Nucleic Acids Res.
2013;41(7):e90. doi: 10.1093/nar/gkt090.
24. Aryee MJ, Jaffe AE, Corrada-Bravo H, Ladd-Acosta C, Feinberg AP, Hansen KD, et al.
Minfi: a flexible and comprehensive Bioconductor package for the analysis of Infinium
DNA methylation microarrays. Bioinformatics. 2014;30(10):1363-9. doi:
10.1093/bioinformatics/btu049.
25. Morris TJ, Butcher LM, Feber A, Teschendorff AE, Chakravarthy AR, Wojdacz TK, et al.
ChAMP: 450k Chip Analysis Methylation Pipeline. Bioinformatics. 2014;30(3):428-30.
doi: 10.1093/bioinformatics/btt684.
26. Fortin JP, Triche TJ, Jr., Hansen KD. Preprocessing, normalization and integration of the
Illumina HumanMethylationEPIC array with minfi. Bioinformatics. 2017;33(4):558-60.
doi: 10.1093/bioinformatics/btw691.
27. Zhou W, Laird PW, Shen H. Comprehensive characterization, annotation and innovative use
of Infinium DNA methylation BeadChip probes. Nucleic Acids Res. 2017;45(4):e22. doi:
10.1093/nar/gkw967.
28. Houseman EA, Accomando WP, Koestler DC, Christensen BC, Marsit CJ, Nelson HH, et al.
DNA methylation arrays as surrogate measures of cell mixture distribution. BMC
Bioinformatics. 2012;13:86. doi: 10.1186/1471-2105-13-86.
29. Kennedy DW, White NM, Benton MC, Fox A, Scott RJ, Griffiths LR, et al. Critical
evaluation of linear regression models for cell-subtype specific methylation signal from
mixed blood cell DNA. PLoS One. 2018;13(12):e0208915. doi:
10.1371/journal.pone.0208915.
30. Du P, Kibbe WA, Lin SM. lumi: a pipeline for processing Illumina microarray.
Bioinformatics. 2008;24(13):1547-8. doi: 10.1093/bioinformatics/btn224.
31. Ritchie ME, Phipson B, Wu D, Hu Y, Law CW, Shi W, et al. limma powers differential
expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res.
2015;43(7):e47. doi: 10.1093/nar/gkv007.
32. Staley JR, Suderman M, Simpkin AJ, Gaunt TR, Heron J, Relton CL, et al. Longitudinal
analysis strategies for modelling epigenetic trajectories. Int J Epidemiol. 2018;47(2):516-
25. doi: 10.1093/ije/dyy012.
33. Sriprasert I, Hodis HN, Karim R, Stanczyk FZ, Shoupe D, Henderson VW, et al. Differential
Effect of Plasma Estradiol on Subclinical Atherosclerosis Progression in Early vs Late
202
Postmenopause. J Clin Endocrinol Metab. 2019;104(2):293-300. doi: 10.1210/jc.2018-
01600.
34. Rinaldi S, Geay A, Dechaud H, Biessy C, Zeleniuch-Jacquotte A, Akhmedkhanov A, et al.
Validity of free testosterone and free estradiol determinations in serum samples from
postmenopausal women by theoretical calculations. Cancer Epidemiol Biomarkers Prev.
2002;11(10 Pt 1):1065-71.
35. Gao X, Jia M, Zhang Y, Breitling LP, Brenner H. DNA methylation changes of whole blood
cells in response to active smoking exposure in adults: a systematic review of DNA
methylation studies. Clin Epigenetics. 2015;7:113. doi: 10.1186/s13148-015-0148-3.
36. Zaina S, Heyn H, Carmona FJ, Varol N, Sayols S, Condom E, et al. DNA methylation map
of human atherosclerosis. Circ Cardiovasc Genet. 2014;7(5):692-700. doi:
10.1161/CIRCGENETICS.113.000441.
37. Siemelink MA, van der Laan SW, Haitjema S, van Koeverden ID, Schaap J, Wesseling M, et
al. Smoking is associated to dna methylation in atherosclerotic carotid lesions. Circ
Genom Precis Med. 2018;11(9):e002030. doi: 10.1161/CIRCGEN.117.002030.
38. Portilla-Fernandez E, Hwang SJ, Wilson R, Maddock J, Hill WD, Teumer A, et al. Metaanalysis of epigenome-wide association studies of carotid intima-media thickness. Eur J
Epidemiol. 2021;36(11):1143-55. doi: 10.1007/s10654-021-00759-z.
39. Kovacs T, Szabo-Meleg E, Abraham IM. Estradiol-induced epigenetically mediated
mechanisms and regulation of gene expression. Int J Mol Sci. 2020;21(9). doi:
10.3390/ijms21093177.
40. Boucher P, Matz RL, Terrand J. atherosclerosis: gone with the Wnt? Atherosclerosis.
2020;301:15-22. doi: 10.1016/j.atherosclerosis.2020.03.024.
41. Goliasch G, Wiesbauer F, Kastl SP, Katsaros KM, Blessberger H, Maurer G, et al. Premature
myocardial infarction is associated with low serum levels of Wnt-1. Atherosclerosis.
2012;222(1):251-6. doi: 10.1016/j.atherosclerosis.2012.02.017.
42. Khoo US, Chan KY, Chan VS, Lin CL. DC-SIGN and L-SIGN: the SIGNs for infection. J
Mol Med (Berl). 2008;86(8):861-74. doi: 10.1007/s00109-008-0350-2.
43. Aavik E, Lumivuori H, Leppanen O, Wirth T, Hakkinen SK, Brasen JH, et al. Global DNA
methylation analysis of human atherosclerotic plaques reveals extensive genomic
hypomethylation and reactivation at imprinted locus 14q32 involving induction of a
miRNA cluster. Eur Heart J. 2015;36(16):993-1000. doi: 10.1093/eurheartj/ehu437.
44. Yamada Y, Horibe H, Oguri M, Sakuma J, Takeuchi I, Yasukochi Y, et al. Identification of
novel hyper- or hypomethylated CpG sites and genes associated with atherosclerotic
plaque using an epigenome-wide association study. Int J Mol Med. 2018;41(5):2724-32.
doi: 10.3892/ijmm.2018.3453.
203
45. Li J, Zhang X, Yang M, Yang H, Xu N, Fan X, et al. DNA methylome profiling reveals
epigenetic regulation of lipoprotein-associated phospholipase A2 in human vulnerable
atherosclerotic plaque. Clin Epigenetics. 2021;13(1):161. doi: 10.1186/s13148-021-
01152-z.
46. Husby A. On the Use of Blood Samples for Measuring DNA Methylation in Ecological
Epigenetic Studies. Integr Comp Biol. 2020;60(6):1558-66. doi: 10.1093/icb/icaa123.
204
Acknowledgments
We gratefully acknowledge the women who participated in ELITE
Ethics approval and consent to participate
All ELITE participants provided written informed consent at enrollment and underwent clinical
protocols for sample collection that were designed to minimize discomfort and inconvenience.
The Institutional Review Board of Keck School of Medicine of USC reviewed and approved all
study protocols.
Availability of data and materials
Datasets generated and analyzed during the current study are not publicly available since consent
for such public release of epigenetic data from ELITE participants was not obtained. However,
raw data to generate figures and tables presented in this study are available from the
corresponding author upon reasonable request, with the appropriate permission from ELITE
Research Group Committee, and with institutional review board approval.
Competing interests
The authors declare that they have no competing interests.
Funding
This work was supported, in part, by National Institutes of Health grants R01AG024154,
R01AG059690, R01HL133169, and R01HL148110.
Author Contributions
205
Concept and design: WJM, RK, HNH, and HA. Acquisition, analysis, or interpretation of data:
JR H, JAH, IS, NK, ZC, RK, JD, WJM, HNH, and HA. Drafting of the manuscript: JRH, JAH,
and HA. All authors contributed to critical revision of the manuscript for important intellectual
content. All authors read and approved the final manuscript.
206
Figure Legends
Figure 1. Miami plot of EWAS results for association of methylation levels with smoking.
Methylation levels at a CpG site on chromosome 5 (cg05575921) were significantly associated
with smoking at baseline (top panel) and 36 months after hormone treatment (bottom panel).
Genome-wide methylation was assessed across 748,567 CpG sites and P-values for differences
between never, former, and current smokers, as determined by linear regression using M-values
for methylation with adjustment for age, ethnicity, time since menopause, treatment, and
estimated blood cell fractions, are plotted as a function of genomic location. The solid red and
blue lines indicate the significant (P=6.7x10-8
) and suggestive (P=6.7x10-6
) thresholds for
significance, respectively.
Figure 2. EWAS results for association of methylation levels with time-since-menopause
and HT. (A) Manhattan plot shows two CpG sites on chromosomes 12 (cg19552895) and 19
(cg18515510) at which the difference in methylation levels between baseline and 36 months
after treatment were significantly associated with time-since-menopause and treatment. Genomewide methylation was assessed across 748,567 CpG sites and interaction P-values between timesince-menopause and HT for changes in methylation (M-values) from baseline to 36 months post
treatment, with adjustment for age, ethnicity, and estimated blood cell fractions, are plotted as a
function of genomic location. The solid red and blue lines indicate the significant (P=6.7x10-8
)
and suggestive (P=6.7x10-6
) thresholds for significance, respectively. Regional plots show 400kb
intervals on chromosomes 12 and 19 centered on cg19552895 (B) and cg18515510 (C),
respectively. Genes located within the 400kb intervals are shown in the bottom panels.
207
Figure 3. Methylation levels at two significantly associated CpG sites as a function of
treatment group. Methylation levels (%) for cg19552895 (A) and cg18515510 (B) at baseline
and 36 months after treatment are shown in the four treatment groups of women. Significant
increases in methylation were observed from baseline (pre) to 36 months after treatment (post)
for both CpGs in the late (>6 years from menopause) HT group but not any of the other three
groups. P-values are based on linear regression using M-values for methylation, with adjustment
for age, ethnicity, and estimated blood cell fractions.
Figure 4. Relationship between changes in methylation levels and subclinical
atherosclerosis. The correlation between changes in CIMT and changes in methylation levels
(%) of cg19552895 (A) and cg18515510 (B) from baseline to 36 months after treatment
(calculated as post-pre for CIMT and methylation) is shown for women in the late/HT group. Pvalues are based on Spearman correlations using M-values for methylation with adjustment for
age and estimated blood cell fractions at baseline and at 36 months after treatment.
208
Table 1. Clinical Characteristics of Study Participants.
Trait (N or %) Early/Placebo
(n=48)
Early/Treatment
(n=48)
aLate/Placebo
(n=42)
Late/Treatment
(n=48)
bp-value
Age 55.5 (5.2) 55.3 (4.7) 65.1 (7.8) 66.7 (8.1) <0.0001
Years since menopause 3.3 (3.1) 3.5 (3.5) 14.6 (7.9) 14 (6.7) <0.0001
Ethnicity 0.68
White Non-Hispanic 34 [71] 39 [81] 30 [71] 33 [44]
Black Non-Hispanic 8 [17] 2 [4] 4 [45] 6 [13]
Hispanic 3 [6] 5 [45] 6 [14] 5 [45]
Asian 3 [6] 2 [4] 2 [46] 4 [8]
Smoking 0.70
Never 24 [50] 28 [58] 24 [57] 33 [44]
Former 22 [46] 19 [46] 17 [46] 14 [29]
Current 2 [4] 1 [2] 1 [2] 1 [2]
Body mass index 27.0 (6.6) 26.1 (5.3) 27.1 (5.5) 26.2 (7.8) 0.90
Total cholesterol, mg/dL 233 (46) 231 (30) 227 (30) 222 (53) 0.64
LDL, mg/dL 144 (41) 147 (37) 137 (42) 133 (52) 0.30
HDL, mg/dL 64 (26) 60 (20) 63 (29) 59 (22.5) 0.73
SBP, mmHg 120 (21) 117 (15) 121 (19) 122 (13) 0.09
DBP, mmHg 77 (10) 77 (9) 75 (7) 75 (11) 0.47
*Change in free estradiol (pg/mL) 0.015 (0.09) 0.38 (0.45) 0.01 (0.12) 0.35 (0.70) <0.0001
*Change in CIMT, mm 0.048 (0.017) -0.011 (0.029) 0.057 (0.014) 0.059 (0.014) <0.0001
Data are shown as median (IQR) or as n [%].
*Change is shown as the difference between baseline and 36 months after treatment (calculated as post-pre).
aSix subjects in the late/placebo group whose methylation data did not pass QC steps were excluded from all analyses.
bP-values for differences between groups were derived from Kruskal-Wallis tests for continuous variables or chi-square tests for
dichotomous/categorical traits, respectively.
209
Table 2. Methylation at Previously Identified CpGs Most Strongly Affected by Smoking.
Baseline 36-month visit
CpG aChr:pos Nearest Gene Effect (SE) bp-value Effect (SE) bp-value
cg09935388 1:92947588 GFI1 -0.094 (0.063) 0.14 -0.063 (0.057) 0.27
cg19859270 3:98251294 GPR15 -0.078 (0.023) 9.8x10-4
-0.045 (0.023) 0.047
cg23576855 5:373299 AHRR -0.24 (0.12) 0.04 -0.25 (0.12) 0.04
cg05575921 5:373378 AHRR -0.46 (0.07) 3.9x10-10
-0.44 (0.072) 4.0x10-9
cg03636183 19:17000585 F2RL3 -0.13 (0.033) 2.4x10-4
-0.12 (0.030) 1.4x10-4
Data are shown as the effect of smoking on methylation (M-values) at the baseline visit and 36 months after treatment across all treatment groups.
aChromosome and position (base pair) are based on build 37 (hg19) of the human genome reference sequence.
bP-values were derived from linear regression analyses for smoking status (never, former, or current; coded as 0, 1, 2) with adjustment for age,
ethnicity, time-since-menopause, randomized treatment, and estimated blood cell fractions.
210
Table 3. CpG Sites with Significant Changes in Percent Methylation as a Function of Menopause and Treatment Groups.
CpG aChr:pos Nearest Gene (location)
Early/Placebo
(n=48)
Early/Treatment
(n=48)
Late/Placebo
(n=42)
Late/Treatmen
t (n=48)
bP-int
cg19552895 12:49379205 WNT1 (S_Shelf) -0.8 -0.6 -1.2 1.7 1.1x10-9
cg18515510 19:7831896 CLEC4M (3' UTR) 0.4 -0.2 0.6 1.9 2.4x10-8
Data are shown as the mean change in methylation levels (%) between baseline and 36 months after treatment (calculated as
post-pre) for the four treatment groups.
aChromosome and position (base pair) are based on build 37 (hg19) of the human genome reference sequence.
bP-values are derived from tests of interaction between time-since-menopause and HT using M-values for methylation with
adjustment for age, ethnicity, and estimated blood cell fractions.
S_Shelf, southern shelf region that is directly adjacent to a southern shore, which is directly downstream of a CpG island.
211
Table 4. Methylation Levels in Leukocytes of ELITE Participants at CpGs Previously Associated with Atherosclerosis.
Tissue CpG aChr:pos Nearest
Gene
Early/Placeb
o (n=48)
Early/Treatme
nt (n=48)
Late/Placeb
o (n=42)
Late/Treatme
nt (n=48)
bP-int
Aortic
plaque cg07608848 2:1,647,185 PXDN 0.4 0.1 0.1 0.2 0.43
Blood cg23079012 2:8,343,662 LINC0029
9
-0.2 0.3 0.2 0.1 0.66
Blood cg21566642 2:233,284,61
3
ALPI -0.7 0.3 0.1 0.01 0.08
Blood cg03358636 3:197,473,95
8
RUBCN 0.5 1.3 -0.4 0.1 0.73
Blood/Caroti
d plaque cg12806681 5:368,346 AHRR 0.1 0.1 0.3 0.3 0.82
Blood cg23916896 5:368,756 AHRR 1.0 0.9 0.1 0.7 0.16
Blood/Caroti
d plaque cg05575921 5:373,378 AHRR -0.1 -0.2 0.6 -0.2 0.63
Blood cg26703534 5:377,358 AHRR 0.5 0.1 -0.6 -0.3 0.81
Blood cg21161138 5:399,312 AHRR 0.6 0.1 -0.5 -0.1 0.97
Aortic
plaque cg23979631 7:27,142,427 HOXA2 0.2 1.0 0.0 0.0 0.36
Aortic
plaque cg19816811 7:27,188,364 HOXA6 1.0 0.7 0.1 -0.5 0.28
Aortic
plaque cg03217995 7:27,203,430 HOXA9 3.2 2.1 -1.4 0.8 0.45
Aortic
plaque cg16913789 7:27,204,005 HOXA9 0.8 -0.2 -0.1 -0.2 0.37
Aortic
plaque cg25188395 7:27,204,052 HOXA9 0.6 0.8 -0.2 0.9 0.70
Aortic
plaque cg17466857 7:27,225,528 HOXA11A
S
0.1 0.0 -0.1 0.2 0.03
Aortic
plaque cg01419713 8:42038135 PLAT 0.0 -1.0 0.8 0.2 0.49
212
Blood cg03450842 10:80,834,94
7
ZMIZ1 0.3 0.8 -0.9 -0.2 0.26
Blood cg11660018 11:86,510,91
5
OR7E2P 0.2 0.9 -0.4 0.3 0.16
Blood cg03371962 12:1,772,275 MIR3649 0.6 0.8 0.1 -0.5 0.87
Aortic
plaque cg23395715 12:54,369,51
4
HOXC11 0.8 0.6 0.9 0.7 0.68
Aortic
plaque cg02384661 12:54,369,63
8
HOXC11 0.3 0.0 -0.2 0.3 0.47
Aortic
plaque cg03146625 12:54,448,72
9
HOXC4 1.1 1.9 0.3 1.8 0.42
Aortic
plaque cg15648389 12:54,448,76
9
HOXC4 2.5 0.9 0.9 2.5 0.48
Blood/Caroti
d plaque cg05284742 14:93,552,08
0
ITPK1 0.2 1.2 0.2 -0.2 0.96
Blood cg17295878 17:77,924,66
5
TBC1D16 0.1 -0.3 -0.2 0.1 0.01
Blood cg03636183 19:17,000,53
7
F2RL3 0.4 0.5 1.0 -0.5 0.80
Data are shown as the mean change in methylation levels (%) between baseline and 36 months after treatment (calculated as post-pre) for the four
groups of women.
aChromosome and position (basepair) are based on build 37 (hg19) of the human genome reference sequence.
bP-values are derived from tests of interaction between time-since-menopause and HT using m-values for methylation, with adjustment for age,
ethnicity, and estimated blood cell fractions.
213
Figure 1.
214
Figure 2.
215
Figure 3.
216
Figure 4
217
Discussion
Elucidating Novel Coronary Artery Disease (CAD) Mechanisms through GeneEnvironment Interactions. CAD is a multifactorial disease with well-established causal
pathways [1] driven by genetic predisposition [2-5] and environmental factors [6-11]. Despite
the existence of a large body of knowledge regarding the individual contributions of genetic and
environmental factors to CAD, a comprehensive exploration of the synergistic relationship
between them is lacking in the literature. Therefore, this dissertation highlights various pathways
by which genetic variation and exposures interact to shape CAD risk (Figure 1). These
innovative findings presented herein offer new insights into cardiovascular disease (CVD)
pathogenesis to uncover novel mechanisms underlying the unique etiology of atherosclerosis.
Figure 1. Genetic Susceptibility Loci and Exposures Synergize to Increase Risk of CVD.
One key discovery was that TRIP4 is the likely causal gene driving the observed gene-air
pollution interaction at a CAD locus on chromosome 15, where genetic variation and fine
particulate matter <2.5m in diameter (PM2.5) interacted to regulate TRIP4 expression and
218
amplify CAD risk. Additionally, a second discovery was identification of the ABO locus
heightening thrombotic event risk in individuals with non-O blood type who were hospitalized
for COVID-19 following infection with SARS-CoV-2, thus illustrating a novel gene-pathogen
interaction. Third, a functional variant in CPS1 was independently validated to decrease CAD
risk in women but not men in the largest sex stratified multiethnic meta-analysis for CAD risk
known to date. This sexually dimorphic pattern was supported by analogous observations in male
Cps1 knockdown mice, thus validating one of the few examples of a gene-sex interaction for
CAD in humans. Finally, in postmenopausal women, hormone replacement therapy (HRT) was
shown to influence DNA methylation patterns, revealing how gene-drug interactions can affect
epigenetic regulation with potential implications for CAD risk. Collectively, these findings
underscore how diverse genetic interactions (e.g. gene-air pollution, gene-pathogen, gene-sex,
and gene-drug) can contribute to CAD risk, setting the stage for a deeper exploration of their
underlying mechanisms (Figure 1).
Known CAD Variant Contributes to Gene-Exposure Interactions. We conducted a
comprehensive analysis combining epidemiological data, genetic studies, and functional
experiments to explore gene-exposure (GxE) interaction between air pollution and CAD risk.
Our analysis of the UK Biobank cohort revealed that elevated levels of PM2.5, nitrogen oxides
(NOx), and nitrogen dioxide were associated with an increased risk of CAD. These findings align
with established links between air pollution and adverse cardiovascular outcomes [12]. To
investigate potential GxE interactions, we examined over 300 known CAD susceptibility loci. Of
these, we identified a novel interaction at a locus on chromosome 15 where PM2.5 exposure was
linked to increased CAD risk among carriers of the susceptibility allele (A) of the lead single
nucleotide polymorphism (SNP), rs6494488, at this locus.
219
Genetic Signal for CAD at 15q22 Loci Linked to TRIP4. At the functional level, rs6494488
significantly influenced the expression of TRIP4 (thyroid hormone receptor interactor 4), with
carriers of the risk allele (A) exhibiting reduced TRIP4 expression in several cardiometabolic
tissues. These findings were observed in independent datasets, including the Genotype-Tissue
Expression (GTEx) Project [13] and the Stockholm-Tartu Atherosclerosis Reverse Network
Engineering Task (STARNET) study [14], and were most significant for TRIP4 despite
rs6494488 affecting expression of a few other genes at the chromosome 15 locus. The genotypic
influence of rs6494488 as an expression quantitative trait locus (eQTL), along with the
association of the A allele with increased CAD risk through main effects, indicates that TRIP4
may have a protective function in maintaining vascular health. Supporting this hypothesis,
analysis of the STARNET cohort showed reduced TRIP4 expression in atherosclerotic aortas
from CAD patients compared to aortas from individuals without CAD. Taken together, these
data point to TRIP4 as one high probability candidate causal gene influencing CAD
independently of air pollution exposure.
Air Pollutants regulate systemic TRIP4 expression. Our in vitro analysis further demonstrated
that TRIP4 was the sole positional candidate gene at the chromosome 15 locus whose expression
in both human coronary artery and microvascular endothelial cells was affected following
exposure to various diesel exhaust particle (DEP) equivalents. Consistent with these
observations, the in vivo murine model also demonstrated downregulation of Trip4 in response to
chronic DEP exposure, suggesting a reactive vascular response to pollutants implicated in
atherosclerosis. Thus, these data collectively provide compelling evidence that TRIP4 is the
likely candidate causal gene and is affected by both short-term and long-term exposure to DEPrelated air pollution in humans and mice, respectively.
220
These findings highlight the need to determine whether the response to air pollution
influences gene expression exclusively in the aorta or extends to other cardiometabolic-relevant
tissues. In support of this notion, previous studies have shown differential gene expression
profiles in macrophages from DEP-exposed mice [15]. Additionally, our observations underscore
the need to determine whether the gene expression changes are confined to atherosclerosisrelated tissues or represent a broader systemic environmental stress response. For example,
previous studies have demonstrated differential transcriptomic and metabolomic dysregulation in
human alveolar cells exposed to PM2.5 [16] highlighting the systemic impact of ambient air
pollutants on gene regulation and metabolic pathways.
Even with systemic effects, it may still be difficult to determine what component of air
pollution triggers upregulation of stress-related pathways due to the heterogeneous composition
of air pollution. DEP, for example, contains high levels of elemental and organic carbon [17],
creating a complex mixture of pro-oxidant and pro-inflammatory components [18]. Expanding
studies to isolate specific fractions of pollutants through controlled, homogenous exposures
could refine our understanding of the air pollutants driving these effects. To address the
challenges posed, our in vivo murine exposure study utilized a well-characterized and
standardized form of DEP (NIST SRM 2975), which we [19, 20] and others over the past two
decades [21-29] have demonstrated to be biologically active and capable of inducing adverse
effects, including aortic lesion formation in mice [30]. Although it was reassuring to observe that
Trip4 expression was downregulated in mice exposed to DEP, the GxE interaction observed in
the UK Biobank and GeneBank cohorts was specifically with exposure to PM2.5. Therefore,
studies examining TRIP4 expression responses to PM2.5, as well as to other pollutants associated
with CAD, such as NOx, could also be carried out to provide additional insight into the
221
mechanisms underlying our observations. These additional studies would be instrumental in
elucidating the precise environmental contributions to the observed GxE effects and their
implications for CAD pathogenesis.
An essential aspect of ambient air pollution is that the timing and duration of exposures
significantly influences their impact on disease risk. Our in vitro analysis of endothelial cells
demonstrated that TRIP4 is acutely downregulated in response to DEP exposure, highlighting the
immediate and reactive nature of TRIP4 regulation to air pollution. Complementing these
findings, our in vivo experiment, designed to model chronic exposure similar to the cumulative
effects observed in humans, revealed that prolonged DEP exposure for 8 weeks led to decreased
Trip4 expression, potentially contributing to vascular dysfunction and atherosclerosis
progression. This temporal consistency, observed across both acute and chronic exposure
contexts, raises the possibility that TRIP4 plays as a central role in mediating the vascular
response to air pollution, irrespective of exposure duration.
GxE Interaction Implicates to TRIP4. TRIP4 encodes one of the four subunits of the activating
signal cointegrator ribonucleoprotein complex (ASCC), which is thought to facilitate both
transcriptional activation as well as RNA processing events, including splicing [31]. These
events may have broad implications for CAD pathogenesis by influencing splicing and gene
expression regulation in vascular and cardiometabolic contexts. For example, the rs688 variant in
the low-density lipoprotein receptor gene (LDLR) is associated with alternative splicing of its
transcripts, resulting in altered exon 12 usage that increases nonsense-mediated decay of the
alternatively spliced LDLR mRNA, thereby influencing circulating LDL cholesterol levels [32].
Furthermore, the ASCC has also been implicated in DNA damage repair induced by alkylating
222
agents, raising the question of whether TRIP4 could similarly be involved in protecting against
oxidative stress and DNA damage associated with exposure to air pollutants, such as PM2.5 and
DEP [33, 34]. Taken together, our GxE interaction and functional data strongly support the
hypothesis that genetic predisposition and environmental exposures converge on TRIP4 to
modulate its expression and potentially influence cardiovascular outcomes. The parallel
reduction in TRIP4 expression observed among carriers of the risk allele of rs6494488 and in
various human endothelial cells exposed to DEP suggests a potentially shared mechanistic
pathway. This highlights how genetic susceptibility, combined with pollution exposure, could
amplify CAD risk.
Severe SARS-CoV-2 Infection Identified as CAD Risk Equivalent and Potential Genetic
Interactor. Another important highlight from our studies was the significant impact severe
COVID-19 was observed to have on incident cardiovascular risk. Notably, hospitalized COVID19 patients without pre-existing CVD experienced rates of major adverse cardiac events (MACE)
comparable to, or even exceeding, those observed in individuals with established CVD, PAD, or
diabetes but without COVID-19. These findings suggest that severe COVID-19 may function as
a “CAD risk equivalent,” similar to diabetes or PAD, which traditionally warrant comprehensive
and aggressive CVD prevention measures, including lipid-lowering, antihypertensive, and
antiplatelet therapies.
Beyond epidemiological associations, our study also documented a gene-pathogen
interaction with severe COVID-19 and thrombotic outcomes. We showed hospitalized COVID19 patients with non-O blood types had nearly two-fold increased risk of MI or stroke while risk
remained neutral in those with blood type O, underscoring the critical role that host genetics
223
plays in the cardiovascular manifestations of infectious diseases. Therefore, future research
should build on these findings with additional large-scale epidemiological studies. Ultimately,
the results presented in this dissertation underscore that while genetic and environmental factors
can independently influence disease risk, their synergy creates amplified effects.
Candidate Gene Approach Targeting ABO Locus. It has long been known that systemic or
localized infections are known to elevate thrombosis risk post infection [35-40] and COVID-19
studies have further emphasized these associations [41]. However, few studies have sought to
investigate whether host genome exacerbates this risk. Therefore, we used a candidate gene
approach to test whether increased risk of thrombosis post-COVID-19 could be affected through
interactions with genetic susceptibility factors based on the identification of ABO as a
susceptibility locus specifically for MI and stroke [42-44]. Although a previous study similarly
reported heightened MACE risk in COVID-19 patients with blood type A compared to type O,
that analysis was limited by its small sample size, observational nature of short duration, and
inability to explore confounding genetic interactions [45]. Therefore, we conducted a formal
longitudinal analysis to evaluate the presence of an interaction, which demonstrated that the
increased incident risk of thrombosis associated with COVID-19 varied depending on ABO
blood type.
Our observations with ABO blood type raised the question of whether other blood groups,
such as Rhesus (Rh) factor, might also contribute to post-infection thrombotic risk through
genetic interactions, given published clinical data suggesting that being Rh positive increases
risk of being hospitalized for COVID-19 [46]. Therefore, we extended our analytical framework
to investigate a potential interaction between Rh factor blood type and COVID-19 as well. Since
224
serological data on Rh factor was not available in the UK Biobank, we utilized genetic variants
that were previously identified as proxies [47]. Specifically, we focused on two common variants
associated with being Rh negative: rs590787 and rs3927482. Rh negative individuals were
classified as those who were homozygous for the minor allele of either rs590787 (CC) or
rs3927482 (GG) or carriers of the minor allele of each variant (AC and TG; i.e. compound
heterozygotes). Rh factor positive subjects were those who were homozygous or heterozygous
for the common alleles at both variants (rs590787/rs3927482 AA/TG or AC/TT). Despite this
rigorous classification, genetic interaction analyses did not reveal significant evidence that Rh
factor modifies the risk of thrombotic events in the context of COVID-19 (data not shown). It is
also noteworthy that Rh factor has not been linked to venous or arterial thrombosis [42-44].
Although we employed a candidate gene approach, future research could benefit from
expanding to unbiased searches for identifying GxE interactions with COVID-19. This broader
strategy has the potential to identify novel mechanisms and genes implicated in post-COVID
thrombotic events and other adverse long-term outcomes. However, even with the size of the UK
Biobank cohort, statistical power will still remain a challenge due the large sample sizes needed
for GxE interaction analyses [48]. Such limitations could be addressed through a collaborative
meta-analysis combining results from similar studies to enhance detection of subtle interactions
and uncover novel pathways driving future thrombotic risk in severe COVID-19 cases.
Putative Mechanism Linking ABO and COVID-19. Large-scale genetic studies have shown
that the ABO locus is associated with increased susceptibility to SARS-CoV-2 infection but not
with hospitalization for COVID-19 itself [49]. However, our follow-up genetic analyses
suggested that the underlying mechanism for how thrombotic risk was increased as a function of
225
ABO blood type was not related to SARS-CoV-2 infection susceptibility, COVID-19 severity, or
atherosclerotic plaque development. For example, two independent variants at a chromosome
3p21 locus [50] have also been implicated in COVID-19 susceptibility. One SNP,
rs73062389,increases the likelihood of being infected with SARS-CoV-2 [49], whereas
rs11385942 heightens risk of hospitalization after infection [51]. However, we did not obtain
evidence that rs73062389 or rs11385942 influenced thrombotic risk through an interaction with
severe COVID-19. Similarly, we also tested for an interaction with the well-known CAD locus
on chromosome 9p21 [42, 52]. This analysis also did not yield significant evidence for an
interaction with severe COVID-19. Collectively, these results suggest that increased risk of MI
or stroke post severe COVID-19 infection is specific to genetic factors that influence plaque
rupture and thrombus formation (e.g., ABO) rather than those associated with susceptibility to
infection/severity (3p21) or atherosclerotic plaque development (9p21).
Specificity of ABO in Mediating Thrombotic Risk. ABO is one of the most pleiotropic loci in
the genome and has been associated with numerous cardiometabolic traits [53], including
coagulation biomarkers [54]. These observations thus further support the interaction observed
between ABO blood type and thrombotic risk in severe COVID-19 patients. Furthermore,
SARS-CoV-2 has been detected within atherosclerotic plaques, where its replication induces a
pro-atherogenic inflammatory response [55]. The heightened thrombotic risk observed in non-O
blood type individuals may thus result from synergistic effects between SARS-CoV-2 infection
and non-O blood types, which could destabilize plaques or render the endothelium more prone to
thrombus formation. In this case, the genetic determinants of ABO blood type may shape an
individual's thrombotic response to the inflammatory and prothrombotic milieu induced by
SARS-CoV-2. Notably, these findings would also represent one of the first documented gene-
226
pathogen exposure interactions for CVD-related outcomes in a long-term study. Considering that
non-O blood types comprise a substantial part of the global population (>50%), these findings
further suggest a substantial worldwide increase in thrombosis risk among individuals with nonO blood types who develop severe COVID-19.
CPS1 Exhibits Interaction with Sex on Risk of CAD. In another candidate gene approach, we
also evaluated a functional variant, rs715, in the CPS1 gene for association with CAD through an
interaction with sex. This single SNP meta-analysis comprising >1.4 million cases and controls
provided compelling evidence that rs715 was associated with reduced risk of CAD, with stronger
protective effects observed in women compared to men. These analyses also revealed that the
nature of this association was sexually dimorphic rather than strictly sex-specific as was initially
observed.. In vivo functional studies with murine models to validate this association also
demonstrated that Cps1 knockdown mice exhibited a similar metabolic profile for urea cycle
metabolites as humans with the CPS1 decreasing allele of rs715, including decreased levels of
citrulline and ornithine. However, whereas the human association studies demonstrated that
rs715 had a stronger protective effect with decreased CPS1 activity and risk of CAD in women,
male Cps1 knockdown mice exhibited significantly reduced atherosclerotic lesions at the aortic
root and along the aorta compared to wild-type controls.
A central hypothesis of this dissertation is that genetic factors and environmental
exposures can independently contribute to disease risk, but their interaction often produces a
synergistic effect. We initially proposed two possible mechanisms for these GxE interactions:
first, that a specific genotype is necessary for the exposure to exert an effect, and second, that the
exposure alone has a weak independent effect, but genetic susceptibility amplifies its impact.
227
Evidence from our CPS1 study supports the latter hypothesis, as genetically decreased CPS1
activity in both humans and mice produced sexually dimorphic protective effects against CAD
development. These results also underscore that even subtle GxE interactions can be robustly
detected with adequate statistical power, despite variability across genetic backgrounds and
sexes.
Sexually Dimorphic Effects of CPS1 Variant in Humans Show Lack of Association in
Underrepresented Ancestries. While our single SNP meta-analysis provided evidence for a
sexually dimorphic association of rs715 with CAD risk in subjects of European ancestry,
replication of this association across multiple ancestries proved challenging. For example, among
individuals of European ancestry, decreased CPS1 enzymatic activity significantly reduced CAD
risk in women but not in men, a pattern that was also mirrored in East and South Asian groups.
However, association of rs715 with risk of CAD was not statistically significant in Black/African
(n=62,463) or admixed American cohorts (n=44,047), although a combined meta-analysis across
all ancestries did increase the strength of the sexually dimorphic association.
The lack of replication in certain ancestries raises several interesting points. First, the
lack of association in the Black/African and admixed American cohorts could be due to the 10-
fold smaller sample sizes of these groups that were available in our meta-analysis compared to
European ancestry subjects. Non-Asian and non-European populations are notoriously
underrepresented in genetic analyses where only 2.4% of public genome-wide association study
(GWAS) results are derived from African ancestry individuals [56]. Another possibility is that
rs715 had a minor allele frequency (MAF) of 31% in Whites and 20% in Africans, which would
result in decreased power to detect an association in the latter group, assuming effect sizes are
228
similar in both ancestries. It is also possible that the effect size of rs715 on risk of CAD was
weaker in both male and female African ancestry subjects. In this regard, the same lead variants
at chromosome 9p21 have been shown to yield disparate effect sizes in European and African
subjects, suggesting that this locus may not be as strong a predictor of CAD risk in African
populations as in White populations [57, 58]. Nevertheless, identification of genetic associations
and biological mechanisms influencing risk of CAD more strongly in one ethnicity may still
have therapeutic implications for all groups, as has been demonstrated for PCSK9 variants and
inhibitors [59].
Sexually Dimorphic Hepatic CPS1 Expression Explains Threshold Association With Risk
of CAD. The sexually dimorphic protective effects of rs715 on risk of CAD suggests that the
association could be driven by sex hormones. To test this hypothesis, we conducted
gonadectomy experiments in mice to directly evaluate the role of male and female sex hormones.
These experiments revealed that Cps1 expression was not modulated by testosterone or estrogen,
suggesting that intrinsic sex differences in hepatic CPS1 expression and its downstream
metabolic effects could be another explanation for the sexually dimorphic association with risk
of CAD. In this regard, another notable finding highlighted the importance of baseline hepatic
CPS1 expression in modulating CAD risk. For example, in humans, the lower baseline
expression of CPS1 in women could amplify the cardioprotective effects of reduced CPS1
enzymatic activity. Despite the opposite pattern, the same concept would also apply to mice
where inherently lower baseline Cps1 expression in males synergizes with heterozygous genetic
Cps1 deficiency to confer an atheroprotective phenotype. These findings suggest that lower basal
levels of CPS1 expression, independent of sex hormones, in conjunction with decreased
enzymatic function, exceed a threshold that elicits protective effects in both species.
229
The broader concept of gene expression thresholds influencing disease manifestation is
supported by previous studies. For instance, high-density lipoprotein cholesterol (HDL) does not
demonstrate a linear clinical association with mortality across the biologically relevant range. In
men, HDL cholesterol levels between 58–96 mg/dL showed no association with mortality risk,
whereas levels below 58 mg/dL were clearly associated with increased mortality risk [60].
Similarly, in women, HDL cholesterol levels between 77–134 mg/dL were not associated with
mortality risk, but levels below 77 mg/dL exhibited a clear link to increased mortality risk [60].
These findings highlight the importance of achieving specific thresholds of proteins or
metabolites before atherosclerotic effects manifest clinically, a principle that may also apply to
CPS1. Within such a framework, reduced CPS1 activity could act through a threshold-dependent
mechanism as well, where optimal suppression mitigates CAD risk. This mechanism could
involve dysregulation of the urea cycle and related downstream effects on cardiovascular
phenotypes, representing a multifaceted interaction of genetic and metabolic networks.
Alternatively, other cardiometabolic pathways affected by CPS1 could also contribute to
modifying risk of CAD.
CpG Sites Associated with Hormone Replacement Therapy are not Associated with
Carotid Intermedial Thickness (cIMT). Finally, our studies also evaluated the interactions
between pharmacological interventions and epigenetic modifications of DNA rather genetic
variation per se. Specifically, a longitudinal epigenome-wide association study (EWAS)
conducted among a subset of ELITE participants with the lowest and highest 36-month cIMT
progression identified two CpG sites on chromosomes 12 and 19 exhibiting significant
interactions between methylation levels, time since menopause, and hormone replacement
therapy (HRT). These CpG sites, located near WNT1 (cg19552895) and within CLEC4M
230
(cg18515510), showed increased methylation in response to HRT specifically in the late postmenopause group, consistent with studies demonstrating that 17β-estradiol can modulate CpG
methylation via ESR1-mediated pathways [61]. However, despite the strong interactions
observed at these CpG sites, no statistically significant association was found between
methylation changes at cg19552895 or cg18515510 and changes in free estradiol levels or cIMT
progression in the late/HRT group [62].
Blood Cells Serve as Poor Proxies for Vascular Epigenetic Modifications. The results of our
EWAS analyses in ELITE, when considered in the context of other data, suggest that circulating
leukocytes may not serve as suitable proxy tissues for exploring the relationship between
methylation changes and vascular phenotypes. The lack of significant associations in blood cells
indicates that epigenetic modifications relevant to atherosclerosis may be specific to other
relevant tissues beyond peripheral blood leukocytes. Therefore, it could be that the effects of free
estradiol and HRT on methylation at the vascular wall may not be reflected by epigenetic
modifications in circulating leukocytes. Supporting this notion, a large-scale EWAS with over
6,400 individuals failed to detect genome-wide significant associations between cIMT and
methylation levels in blood cells, apart from cg05575921 at a well-established smoking-related
AHRR locus [63]. In contrast, epigenetic analyses of arterial tissues, including atherosclerotic
and normal carotid, aortic, mammary, or femoral arteries, have identified CpG sites associated
with atherosclerosis [63-68]. Notably, these studies often report hypomethylation of CpG sites in
atherosclerotic tissue compared to normal arteries, alongside upregulation of pathways
potentially contributing to plaque development [65, 68]. However, emerging efforts are
investigating whether methylation profiling of peripheral tissues, such as blood, can infer
epigenetic patterns in other tissues [69]. Such approaches hold promise for extending epigenetic
231
studies to atherosclerosis-related traits and their interplay with systemic interventions like HRT.
Understanding the tissue-specific nature of epigenetic changes could provide additional insight
for developing accurate biomarkers and therapeutic targets for cardiovascular diseases.
Our study highlights the complexity of epigenetic responses to pharmaceutical
interventions like HRT and underscores the importance of considering tissue specificity in
epigenetic research. The identification of CpG sites that interact with HRT and time-sincemenopause offers potential avenues for exploring the mechanisms underlying the vascular
effects of HRT. Nevertheless, it remains uncertain whether the observed methylation changes at
cg19552895 and cg18515510 influence the expression of WNT1 and CLEC4M in blood cells or,
more broadly, how the function of these genes align with the vascular effects of HRT in ELITE
participants.
Methodological and Contextual Limitations. Despite the robust nature of our results, our
studies also highlight several limitations that must be considered when interpreting the findings.
First, novel genetic interactions, such as those we observe between the TRIP4 locus and PM2.5
exposure or ABO blood type and severe SARS-CoV-2 infection, still require replication in
independent cohorts. A significant constraint is the limited diversity in ancestry representation,
with most analyses relying on European and East/South Asian populations. This lack of
representation limits the generalizability of results and risks missing population-specific loci or
interactions, particularly in underrepresented groups such as indigenous populations or subjects
of mixed or African ancestry.
In addition, environmental exposures, including air pollution, were assessed using
population-level data or self-reported information, which may not adequately capture individual-
232
level variations or exposure timing. This limitation can obscure critical temporal dynamics, such
as latency effects or cumulative exposure histories, which are crucial for understanding timesensitive interactions. Advances in wearable technology [70] or satellite-based geospatial data
[71] could address this gap.
Although in vitro and in vivo models were employed to validate findings, such as TRIP4
regulation by DEP and Cps1 knockdown in mice, these experimental systems still have inherent
limitations. For example, murine physiology and environmental contexts may not fully replicate
human biology, particularly for sexually dimorphic or epigenetic traits.
The pleiotropic effects of CPS1 and ABO further complicate causal inference. These loci
each influence hundreds of phenotypes relevant to CVD, such as urea cycle metabolites,
thrombotic risk, and numerous clinical biomarkers that make it difficult to isolate the specific
causal factor contributing to CAD risk. Advanced statistical approaches, including Mendelian
randomization methods that take into account pleiotropy [72] and multi-omics mediation
analyses [73], will be needed to disentangle overlapping pathways.
Finally, epigenetic insights, such as those mediated by DNA methylation in response to
HRT, remain largely associative. Challenges like tissue heterogeneity and lack of longitudinal
data limit the interpretability of these findings. Emerging methods, such as single-cell
methylation sequencing [74] or epigenetic profiling of atherosclerosis-related tissues, could
provide more precise insights.
Addressing these limitations requires a multi-faceted approach, including expanded
cohort diversity, precise exposure assessment, advanced functional studies, and longitudinal
233
designs. Integrating these strategies will enhance the robustness and translational potential of
future research, paving the way for discovery of novel mechanisms implicated in CAD.
Future Directions. Although the data we have described has significantly advanced the field,
our future experiments will prioritize investigating the mechanisms underlying sexually
dimorphic differences in CPS1 expression between humans and mice. This will involve both
transcriptomic and proteomic analyses to uncover sex-specific regulatory pathways of CPS1.
Furthermore, we will investigate how decreased expression of CPS1 protects against
atherosclerosis, potentially through the modulation of metabolic or inflammatory pathways.
These studies will employ a combination of human datasets and experimental animal models,
leveraging our current Cps1 murine model to assess causal relationships. Building on our
findings at the TRIP4 locus, we plan to validate the GxE association through expanded human
datasets, integrating both rare and common genetic variant analyses. This effort will also
leverage genetically engineered Trip4 mouse models to experimentally test its function and its
role in atherosclerosis. Beyond these efforts, we aim to explore the feasibility of carrying out
unbiased genome-wide GxE analyses to identify additional loci and interactions that could
provide novel insights into the genetic architecture of CAD. Additionally, we will replicate
previously observed associations of CAD risk and genetic-pathogen interactions from the UK
Biobank in independent datasets to ensure the robustness and generalizability of these findings.
This includes examining how infectious exposures modulate genetic risk. Lastly, we aim to
uncover the mechanisms through which the ABO blood group influences thrombotic risk,
particularly in the context of severe COVID-19. These studies will combine human genetic data,
functional assays, and potentially in vivo models to elucidate how ABO contributes to enhance
the risk of adverse cardiovascular outcomes in severe COVID-19 infection. Together, these
234
directions will provide a comprehensive understanding of genetic and environmental interactions
in CAD risk and uncover novel targets for therapeutic intervention.
Conclusions and Implications in Cardiovascular Disease. In summary, this dissertation
demonstrates how genes interact with environmental and biological exposures to modify risk of
CAD through gene-pollution, gene-pathogen, gene-sex, and gene-drug interactions. First, a genepollution interaction in genetically susceptible individuals exacerbated atherogenesis and
established TRIP4 as a likely key mediator of the vascular response to fine particulate matter
exposure. Additionally, SARS-CoV-2 infection was shown to act as an acute environmental
stressor, where a gene-pathogen interaction involving the ABO locus increased thrombotic risk.
Third, a gene-sex interaction with CPS1 was associated with a protective effect on risk of CAD,
which was more pronounced in women than in men. This effect was independent of hormonal
regulation and possibly related to differences between men and women in baseline hepatic CPS1
expression. Finally, HRT, a pharmaceutical exposure, revealed a gene-drug interaction by
modulating DNA methylation patterns at specific CpG sites, particularly in late postmenopausal
women. However, the lack of significant associations between these epigenetic changes and
cIMT progression underscores the tissue-specific nature of vascular epigenetic modifications.
Overall, the exposures explored in this dissertation, including air pollution, SARS-CoV-2
infection, intrinsic metabolic differences mediated by CPS1 variants, and pharmaceutical
interventions such as HRT, illustrate the diverse mechanisms through which GxE interactions
can influence cardiovascular disease risk. The findings from these studies underscore the
importance of understanding the temporal dynamics, exposure types, and genetic mechanisms
235
unique to each setting for developing more tailored strategies to mitigate cardiovascular risk
across diverse populations and environments.
236
References
1. Grundy SM, Stone NJ, Bailey AL, Beam C, Birtcher KK, Blumenthal RS, et al. 2018
AHA/ACC/AACVPR/AAPA/ABC/ACPM/ADA/AGS/APhA/ASPC/NLA/PCNA
guideline on the management of blood cholesterol: executive summary: A report of the
American College of Cardiology/American Heart Association task force on clinical
practice guidelines. Circulation. 2019;139(25):e1046-e81. doi:
10.1161/CIR.0000000000000624.
2. Wienke A, Holm NV, Skytthe A, Yashin AI. The heritability of mortality due to heart
diseases: a correlated frailty model applied to Danish twins. Twin Res. 2001;4(4):266-74.
doi: 10.1375/1369052012399.
3. Zdravkovic S, Wienke A, Pedersen NL, Marenberg ME, Yashin AI, De Faire U. Heritability
of death from coronary heart disease: a 36-year follow-up of 20 966 Swedish twins. J
Intern Med. 2002;252(3):247-54. doi: 10.1046/j.1365-2796.2002.01029.x.
4. Bevan S, Traylor M, Adib-Samii P, Malik R, Paul NL, Jackson C, et al. Genetic heritability of
ischemic stroke and the contribution of previously reported candidate gene and
genomewide associations. Stroke. 2012;43(12):3161-7. doi:
10.1161/STROKEAHA.112.665760.
5. Vinkhuyzen AA, Wray NR, Yang J, Goddard ME, Visscher PM. Estimation and partition of
heritability in human populations using whole-genome analysis methods. Annu Rev
Genet. 2013;47:75-95. doi: 10.1146/annurev-genet-111212-133258.
6. Hu FB, Stampfer MJ, Manson JE, Rimm E, Colditz GA, Rosner BA, et al. Dietary fat intake
and the risk of coronary heart disease in women. N Engl J Med. 1997;337(21):1491-9.
doi: 10.1056/NEJM199711203372102.
7. Hu FB, Rimm EB, Stampfer MJ, Ascherio A, Spiegelman D, Willett WC. Prospective study of
major dietary patterns and risk of coronary heart disease in men. Am J Clin Nutr.
2000;72(4):912-21. doi: 10.1093/ajcn/72.4.912.
8. Mozaffarian D, Katan MB, Ascherio A, Stampfer MJ, Willett WC. Trans fatty acids and
cardiovascular disease. N Engl J Med. 2006;354(15):1601-13. doi:
10.1056/NEJMra054035.
9. Pope CA, 3rd, Turner MC, Burnett RT, Jerrett M, Gapstur SM, Diver WR, et al. Relationships
between fine particulate air pollution, cardiometabolic disorders, and cardiovascular
mortality. Circ Res. 2015;116(1):108-15. doi: 10.1161/CIRCRESAHA.116.305060.
10. Cosselman KE, Navas-Acien A, Kaufman JD. Environmental factors in cardiovascular
disease. Nat Rev Cardiol. 2015;12(11):627-42. doi: 10.1038/nrcardio.2015.152.
11. Schane RE, Ling PM, Glantz SA. Health effects of light and intermittent smoking: a review.
Circulation. 2010;121(13):1518-22. doi: 10.1161/CIRCULATIONAHA.109.904235.
237
12. Brook RD, Rajagopalan S, Pope CA, Brook JR, Bhatnagar A, Diez-Roux AV, et al.
Particulate matter air pollution and cardiovascular disease: An update to the scientific
statement from the American Heart Association. Circulation. 2010;121(21):2331-78. doi:
10.1161/CIR.0b013e3181dbece1.
13. Consortium G. The Genotype-Tissue Expression (GTEx) project. Nat Genet.
2013;45(6):580-5. doi: 10.1038/ng.2653.
14. Franzen O, Ermel R, Cohain A, Akers NK, Di Narzo A, Talukdar HA, et al. Cardiometabolic
risk loci share downstream cis- and trans-gene regulation across tissues and diseases.
Science. 2016;353(6301):827-30. doi: 10.1126/science.aad6970.
15. Bhetraratana M, Orozco LD, Bennett BJ, Luna K, Yang X, Lusis AJ, et al. Diesel exhaust
particle extract elicits an oxPAPC-like transcriptomic profile in macrophages across
multiple mouse strains. Environ Pollut. 2024;358:124415. doi:
10.1016/j.envpol.2024.124415.
16. Shukla A, Timblin C, BeruBe K, Gordon T, McKinney W, Driscoll K, et al. Inhaled
particulate matter causes expression of nuclear factor (NF)-kappaB-related genes and
oxidant-dependent NF-kappaB activation in vitro. Am J Respir Cell Mol Biol.
2000;23(2):182-7. doi: 10.1165/ajrcmb.23.2.4035.
17. Saxon A, Diaz-Sanchez D. Air pollution and allergy: you are what you breathe. Nat
Immunol. 2005;6(3):223-6. doi: 10.1038/ni0305-223.
18. Romieu I, Moreno-Macias H, London SJ. Gene by environment interaction and ambient air
pollution. Proc Am Thorac Soc. 2010;7(2):116-22. doi: 10.1513/pats.200909-097RM.
19. Shkirkova K, Lamorie-Foote K, Zhang N, Li A, Diaz A, Liu Q, et al. Neurotoxicity of diesel
exhaust particles. J Alzheimers Dis. 2022;89(4):1263-78. doi: 10.3233/JAD-220493.
20. Shkirkova K, Demetriou AN, Sizdahkhani S, Lamorie-Foote K, Zhang H, Morales M, et al.
Microglial TLR4 mediates white matter injury in a combined model of diesel exhaust
exposure and cerebral hypoperfusion. Stroke. 2024;55(4):1090-3. doi:
10.1161/STROKEAHA.124.046412.
21. Singh P, DeMarini DM, Dick CA, Tabor DG, Ryan JV, Linak WP, et al. Sample
characterization of automobile and forklift diesel exhaust particles and comparative
pulmonary toxicity in mice. Environ Health Perspect. 2004;112(8):820-5. doi:
10.1289/ehp.6579.
22. Farraj AK, Haykal-Coates N, Ledbetter AD, Evansky PA, Gavett SH. Neurotrophin
mediation of allergic airways responses to inhaled diesel particles in mice. Toxicol Sci.
2006;94(1):183-92. doi: 10.1093/toxsci/kfl089.
23. Hartz AM, Bauer B, Block ML, Hong JS, Miller DS. Diesel exhaust particles induce
oxidative stress, proinflammatory signaling, and P-glycoprotein up-regulation at the
blood-brain barrier. FASEB J. 2008;22(8):2723-33. doi: 10.1096/fj.08-106997.
238
24. Hougaard KS, Jensen KA, Nordly P, Taxvig C, Vogel U, Saber AT, et al. Effects of prenatal
exposure to diesel exhaust particles on postnatal development, behavior, genotoxicity and
inflammation in mice. Part Fibre Toxicol. 2008;5:3. doi: 10.1186/1743-8977-5-3.
25. O'Driscoll CA, Owens LA, Gallo ME, Hoffmann EJ, Afrazi A, Han M, et al. Differential
effects of diesel exhaust particles on T cell differentiation and autoimmune disease. Part
Fibre Toxicol. 2018;15(1):35. doi: 10.1186/s12989-018-0271-3.
26. Bengalli R, Zerboni A, Marchetti S, Longhin E, Priola M, Camatini M, et al. In vitro
pulmonary and vascular effects induced by different diesel exhaust particles. Toxicol
Lett. 2019;306:13-24. doi: 10.1016/j.toxlet.2019.01.017.
27. Shears RK, Jacques LC, Naylor G, Miyashita L, Khandaker S, Lebre F, et al. Exposure to
diesel exhaust particles increases susceptibility to invasive pneumococcal disease. J
Allergy Clin Immunol. 2020;145(4):1272-84 e6. doi: 10.1016/j.jaci.2019.11.039.
28. Liu Y, Wang T, Si B, Du H, Liu Y, Waqas A, et al. Intratracheally instillated diesel PM(2.5)
significantly altered the structure and composition of indigenous murine gut microbiota.
Ecotoxicol Environ Saf. 2021;210:111903. doi: 10.1016/j.ecoenv.2021.111903.
29. Warren CE, Campbell KM, Kirkham MN, Saito ER, Remund NP, Cayabyab KB, et al. The
effect of diesel exhaust particles on adipose tissue mitochondrial function and
inflammatory status. Int J Mol Sci. 2024;25(8). doi: 10.3390/ijms25084322.
30. Miller MR, McLean SG, Duffin R, Lawal AO, Araujo JA, Shaw CA, et al. Diesel exhaust
particulate increases the size and complexity of lesions in atherosclerotic mice. Part Fibre
Toxicol. 2013;10:61. doi: 10.1186/1743-8977-10-61.
31. Jung DJ, Sung HS, Goo YW, Lee HM, Park OK, Jung SY, et al. Novel transcription
coactivator complex containing activating signal cointegrator 1. Mol Cell Biol.
2002;22(14):5203-11. doi: 10.1128/MCB.22.14.5203-5211.2002.
32. Gao F, Ihn HE, Medina MW, Krauss RM. A common polymorphism in the LDL receptor
gene has multiple effects on LDL receptor function. Hum Mol Genet. 2013;22(7):1424-
31. doi: 10.1093/hmg/dds559.
33. Risom L, Moller P, Loft S. Oxidative stress-induced DNA damage by particulate air
pollution. Mutat Res. 2005;592(1-2):119-37. doi: 10.1016/j.mrfmmm.2005.06.012.
34. Vilas-Boas V, Chatterjee N, Carvalho A, Alfaro-Moreno E. Particulate matter-induced
oxidative stress - mechanistic insights and antioxidant approaches reported in in vitro
studies. Environ Toxicol Pharmacol. 2024;110:104529. doi: 10.1016/j.etap.2024.104529.
35. Kinasewitz GT, Yan SB, Basson B, Comp P, Russell JA, Cariou A, et al. Universal changes
in biomarkers of coagulation and inflammation occur in patients with severe sepsis,
regardless of causative micro-organism [ISRCTN74215569]. Crit Care. 2004;8(2):R82-
90. doi: 10.1186/cc2459.
239
36. Beristain-Covarrubias N, Perez-Toledo M, Thomas MR, Henderson IR, Watson SP,
Cunningham AF. Understanding infection-induced thrombosis: lessons learned from
animal models. Front Immunol. 2019;10:2569. doi: 10.3389/fimmu.2019.02569.
37. Smeeth L, Thomas SL, Hall AJ, Hubbard R, Farrington P, Vallance P. Risk of myocardial
infarction and stroke after acute infection or vaccination. N Engl J Med.
2004;351(25):2611-8. doi: 10.1056/NEJMoa041747.
38. Smeeth L, Cook C, Thomas S, Hall AJ, Hubbard R, Vallance P. Risk of deep vein
thrombosis and pulmonary embolism after acute infection in a community setting.
Lancet. 2006;367(9516):1075-9. doi: 10.1016/S0140-6736(06)68474-2.
39. Clayton TC, Gaskin M, Meade TW. Recent respiratory infection and risk of venous
thromboembolism: case-control study through a general practice database. Int J
Epidemiol. 2011;40(3):819-27. doi: 10.1093/ije/dyr012.
40. Dalager-Pedersen M, Sogaard M, Schonheyder HC, Nielsen H, Thomsen RW. Risk for
myocardial infarction and stroke after community-acquired bacteremia: a 20-year
population-based cohort study. Circulation. 2014;129(13):1387-96. doi:
10.1161/CIRCULATIONAHA.113.006699.
41. Varga Z, Flammer AJ, Steiger P, Haberecker M, Andermatt R, Zinkernagel AS, et al.
Endothelial cell infection and endotheliitis in COVID-19. Lancet.
2020;395(10234):1417-8. doi: 10.1016/S0140-6736(20)30937-5.
42. Reilly MP, Li M, He J, Ferguson JF, Stylianou IM, Mehta NN, et al. Identification of
ADAMTS7 as a novel locus for coronary atherosclerosis and association of ABO with
myocardial infarction in the presence of coronary atherosclerosis: two genome-wide
association studies. Lancet. 2011;377(9763):383-92.
43. Hartiala JA, Han Y, Jia Q, Hilser JR, Huang P, Gukasyan J, et al. Genome-wide analysis
identifies novel susceptibility loci for myocardial infarction. Eur Heart J. 2021;42(9):919-
33. doi: 10.1093/eurheartj/ehaa1040.
44. Mishra A, Malik R, Hachiya T, Jurgenson T, Namba S, Posner DC, et al. Stroke genetics
informs drug discovery and risk prediction across ancestries. Nature.
2022;611(7934):115-23. doi: 10.1038/s41586-022-05165-3.
45. Nauffal V, Achanta A, Goldhaber SZ, Piazza G. Association of ABO blood group type with
cardiovascular events in COVID-19. J Thromb Thrombolysis. 2021;51(3):584-6. doi:
10.1007/s11239-020-02364-5.
46. Anderson JL, May HT, Knight S, Bair TL, Horne BD, Knowlton KU. Association of Rhesus
factor blood type with risk of SARS-CoV-2 infection and COVID-19 severity. Br J
Haematol. 2022;197(5):573-5. doi: 10.1111/bjh.18086.
240
47. Shi J, Luo Y. Effects of RHD gene polymorphisms on distinguishing weak D or DEL from
RhD- in blood donation in a Chinese population. Mol Genet Genomic Med.
2019;7(6):e00681. doi: 10.1002/mgg3.681.
48. Smith PG, Day NE. The design of case-control studies: the influence of confounding and
interaction effects. Int J Epidemiol. 1984;13(3):356-65. doi: 10.1093/ije/13.3.356.
49. Initiative C-HG. A second update on mapping the human genetic architecture of COVID-19.
Nature. 2023;621(7977):E7-E26. doi: 10.1038/s41586-023-06355-3.
50. Myers TA, Chanock SJ, Machiela MJ. LDlinkR: An R package for rapidly calculating
linkage disequilibrium statistics in diverse populations. Front Genet. 2020;11:157. doi:
10.3389/fgene.2020.00157.
51. Velavan TP, Pallerla SR, Ruter J, Augustin Y, Kremsner PG, Krishna S, et al. Host genetic
factors determining COVID-19 susceptibility and severity. EBioMedicine.
2021;72:103629. doi: 10.1016/j.ebiom.2021.103629.
52. Beutner F, Teupser D, Gielen S, Holdt LM, Scholz M, Boudriot E, et al. Rationale and
design of the Leipzig (LIFE) Heart Study: phenotyping and cardiovascular characteristics
of patients with coronary artery disease. PLoS One. 2011;6(12):e29070. doi:
10.1371/journal.pone.0029070.
53. Buniello A, MacArthur JAL, Cerezo M, Harris LW, Hayhurst J, Malangone C, et al. The
NHGRI-EBI GWAS Catalog of published genome-wide association studies, targeted
arrays and summary statistics 2019. Nucleic Acids Res. 2019;47(D1):D1005-D12. doi:
10.1093/nar/gky1120.
54. Vasan SK, Rostgaard K, Majeed A, Ullum H, Titlestad KE, Pedersen OB, et al. ABO blood
group and risk of thromboembolic and arterial disease: A study of 1.5 million blood
donors. Circulation. 2016;133(15):1449-57. doi:
10.1161/CIRCULATIONAHA.115.017563.
55. Eberhardt N, Noval MG, Kaur R, Amadori L, Gildea M, Sajja S, et al. SARS-CoV-2
infection triggers pro-atherogenic inflammatory responses in human coronary vessels.
Nat Cardiovasc Res. 2023;2(10):899-916. doi: 10.1038/s44161-023-00336-5.
56. Morales J, Welter D, Bowler EH, Cerezo M, Harris LW, McMahon AC, et al. A standardized
framework for representation of ancestry data in genomics studies, with application to the
NHGRI-EBI GWAS Catalog. Genome Biol. 2018;19(1):21. doi: 10.1186/s13059-018-
1396-2.
57. Tcheandjieu C, Zhu X, Hilliard AT, Clarke SL, Napolioni V, Ma S, et al. Large-scale
genome-wide association study of coronary artery disease in genetically diverse
populations. Nat Med. 2022;28(8):1679-92. doi: 10.1038/s41591-022-01891-3.
241
58. Lettre G, Palmer CD, Young T, Ejebe KG, Allayee H, Benjamin EJ, et al. Genome-wide
association study of coronary heart disease and its risk factors in 8,090 African
Americans: the NHLBI CARe Project. PLoS Genet. 2011;7(2):e1001300.
59. Sabatine MS, Giugliano RP, Keech AC, Honarpour N, Wiviott SD, Murphy SA, et al.
Evolocumab and Clinical Outcomes in Patients with Cardiovascular Disease. N Engl J
Med. 2017;376(18):1713-22. doi: 10.1056/NEJMoa1615664.
60. Madsen CM, Varbo A, Nordestgaard BG. Extreme high high-density lipoprotein cholesterol
is paradoxically associated with high mortality in men and women: two prospective
cohort studies. Eur Heart J. 2017;38(32):2478-86. doi: 10.1093/eurheartj/ehx163.
61. Kovacs T, Szabo-Meleg E, Abraham IM. Estradiol-induced epigenetically mediated
mechanisms and regulation of gene expression. Int J Mol Sci. 2020;21(9). doi:
10.3390/ijms21093177.
62. Hilser JR, Hartiala JA, Sriprasert I, Kono N, Cai Z, Karim R, et al. Effect of menopausal
hormone therapy on methylation levels in early and late postmenopausal women. Clin
Epigenetics. 2022;14(1):90. doi: 10.1186/s13148-022-01311-w.
63. Portilla-Fernandez E, Hwang SJ, Wilson R, Maddock J, Hill WD, Teumer A, et al. Metaanalysis of epigenome-wide association studies of carotid intima-media thickness. Eur J
Epidemiol. 2021;36(11):1143-55. doi: 10.1007/s10654-021-00759-z.
64. Zaina S, Heyn H, Carmona FJ, Varol N, Sayols S, Condom E, et al. DNA methylation map
of human atherosclerosis. Circ Cardiovasc Genet. 2014;7(5):692-700. doi:
10.1161/CIRCGENETICS.113.000441.
65. Aavik E, Lumivuori H, Leppanen O, Wirth T, Hakkinen SK, Brasen JH, et al. Global DNA
methylation analysis of human atherosclerotic plaques reveals extensive genomic
hypomethylation and reactivation at imprinted locus 14q32 involving induction of a
miRNA cluster. Eur Heart J. 2015;36(16):993-1000. doi: 10.1093/eurheartj/ehu437.
66. Siemelink MA, van der Laan SW, Haitjema S, van Koeverden ID, Schaap J, Wesseling M, et
al. Smoking is associated to dna methylation in atherosclerotic carotid lesions. Circ
Genom Precis Med. 2018;11(9):e002030. doi: 10.1161/CIRCGEN.117.002030.
67. Yamada Y, Horibe H, Oguri M, Sakuma J, Takeuchi I, Yasukochi Y, et al. Identification of
novel hyper- or hypomethylated CpG sites and genes associated with atherosclerotic
plaque using an epigenome-wide association study. Int J Mol Med. 2018;41(5):2724-32.
doi: 10.3892/ijmm.2018.3453.
68. Li J, Zhang X, Yang M, Yang H, Xu N, Fan X, et al. DNA methylome profiling reveals
epigenetic regulation of lipoprotein-associated phospholipase A2 in human vulnerable
atherosclerotic plaque. Clin Epigenetics. 2021;13(1):161. doi: 10.1186/s13148-021-
01152-z.
242
69. Husby A. On the use of blood samples for measuring DNA methylation in ecological
epigenetic studies. Integr Comp Biol. 2020;60(6):1558-66. doi: 10.1093/icb/icaa123.
70. Perez AJ, Zeadally S. Recent Advances in Wearable Sensing Technologies. Sensors (Basel).
2021;21(20). doi: 10.3390/s21206828.
71. Anenberg SC, Bindl M, Brauer M, Castillo JJ, Cavalieri S, Duncan BN, et al. Using Satellites
to Track Indicators of Global Air Pollution and Climate Change Impacts: Lessons
Learned From a NASA-Supported Science-Stakeholder Collaborative. Geohealth.
2020;4(7):e2020GH000270. doi: 10.1029/2020GH000270.
72. Burgess S, Thompson SG. Interpreting findings from Mendelian randomization using the
MR-Egger method. Eur J Epidemiol. 2017;32(5):377-89. doi: 10.1007/s10654-017-0255-
x.
73. Sanderson E. Multivariable Mendelian Randomization and Mediation. Cold Spring Harb
Perspect Med. 2021;11(2). doi: 10.1101/cshperspect.a038984.
74. Nichols RV, O'Connell BL, Mulqueen RM, Thomas J, Woodfin AR, Acharya S, et al. Highthroughput robust single-cell DNA methylation profiling with sciMETv2. Nat Commun.
2022;13(1):7627. doi: 10.1038/s41467-022-35374-3.
243
REFERENCES
Aavik E, Lumivuori H, Leppanen O, Wirth T, Hakkinen SK, Brasen JH, et al. Global DNA
methylation analysis of human atherosclerotic plaques reveals extensive genomic
hypomethylation and reactivation at imprinted locus 14q32 involving induction of a
miRNA cluster. Eur Heart J. 2015;36(16):993-1000. doi: 10.1093/eurheartj/ehu437.
Aavik E, Lumivuori H, Leppanen O, Wirth T, Hakkinen SK, Brasen JH, et al. Global DNA
methylation analysis of human atherosclerotic plaques reveals extensive genomic
hypomethylation and reactivation at imprinted locus 14q32 involving induction of a
miRNA cluster. Eur Heart J. 2015;36(16):993-1000. doi: 10.1093/eurheartj/ehu437.
Al-Aly Z, Bowe B, Xie Y. Long COVID after breakthrough SARS-CoV-2 infection. Nat Med.
2022;28(7):1461-7. doi: 10.1038/s41591-022-01840-0.
Aleksova A, Fluca AL, Gagno G, Pierri A, Padoan L, Derin A, et al. Long-term effect of SARSCoV-2 infection on cardiovascular outcomes and all-cause mortality. Life Sci.
2022;310:121018. doi: 10.1016/j.lfs.2022.121018.
Al-Kindi SG, Brook RD, Biswal S, Rajagopalan S. Environmental determinants of
cardiovascular disease: lessons learned from air pollution. Nat Rev Cardiol.
2020;17(10):656-72. doi: 10.1038/s41569-020-0371-2.
All of Us Research Program Genomics I. Genomic data in the All of Us research program.
Nature. 2024;627(8003):340-6. doi: 10.1038/s41586-023-06957-x.
All of Us Research Program I, Denny JC, Rutter JL, Goldstein DB, Philippakis A, Smoller JW,
et al. The "All of Us" Research Program. N Engl J Med. 2019;381(7):668-76. doi:
10.1056/NEJMsr1809937.
Anand SS, Islam S, Rosengren A, Franzosi MG, Steyn K, Yusufali AH, et al. Risk factors for
myocardial infarction in women and men: insights from the INTERHEART study. Eur
Heart J. 2008;29(7):932-40. doi: 10.1093/eurheartj/ehn018.
Anderson JL, May HT, Knight S, Bair TL, Horne BD, Knowlton KU. Association of Rhesus
factor blood type with risk of SARS-CoV-2 infection and COVID-19 severity. Br J
Haematol. 2022;197(5):573-5. doi: 10.1111/bjh.18086.
Anenberg SC, Bindl M, Brauer M, Castillo JJ, Cavalieri S, Duncan BN, et al. Using Satellites to
Track Indicators of Global Air Pollution and Climate Change Impacts: Lessons Learned
From a NASA-Supported Science-Stakeholder Collaborative. Geohealth.
2020;4(7):e2020GH000270. doi: 10.1029/2020GH000270.
Appelman Y, van Rijn BB, Ten Haaf ME, Boersma E, Peters SA. Sex differences in
cardiovascular risk factors and disease prevention. Atherosclerosis. 2015;241(1):211-8.
doi: 10.1016/j.atherosclerosis.2015.01.027.
244
Aragam KG, Jiang T, Goel A, Kanoni S, Wolford BN, Atri DS, et al. Discovery and systematic
characterization of risk variants and genes for coronary artery disease in over a million
participants. Nat Genet. 2022;54(12):1803-15. doi: 10.1038/s41588-022-01233-6.
Aragam KG, Jiang T, Goel A, Kanoni S, Wolford BN, Atri DS, et al. Discovery and systematic
characterization of risk variants and genes for coronary artery disease in over a million
participants. Nat Genet. 2022;54(12):1803-15. doi: 10.1038/s41588-022-01233-6.
Aragam KG, Jiang T, Goel A, Kanoni S, Wolford BN, Atri DS, et al. Discovery and systematic
characterization of risk variants and genes for coronary artery disease in over a million
participants. Nat Genet. 2022;54(12):1803-15. doi: 10.1038/s41588-022-01233-6.
Aragam KG, Jiang T, Goel A, Kanoni S, Wolford BN, Atri DS, et al. Discovery and systematic
characterization of risk variants and genes for coronary artery disease in over a million
participants. Nat Genet. 2022;54(12):1803-15. doi: 10.1038/s41588-022-01233-6.
Araujo JA, Barajas B, Kleinman M, Wang X, Bennett BJ, Gong KW, et al. Ambient particulate
pollutants in the ultrafine range promote early atherosclerosis and systemic oxidative
stress. Circ Res. 2008;102(5):589-96. doi: 10.1161/CIRCRESAHA.107.164970.
Arnett DK, Blumenthal RS, Albert MA, Buroker AB, Goldberger ZD, Hahn EJ, et al. 2019
ACC/AHA guideline on the primary prevention of cardiovascular disease: executive
summary: A report of the american college of cardiology/american heart association task
force on clinical practice guidelines. Circulation. 2019;140(11):e563-e95. doi:
10.1161/CIR.0000000000000677.
Aryee MJ, Jaffe AE, Corrada-Bravo H, Ladd-Acosta C, Feinberg AP, Hansen KD, et al. Minfi: a
flexible and comprehensive Bioconductor package for the analysis of Infinium DNA
methylation microarrays. Bioinformatics. 2014;30(10):1363-9. doi:
10.1093/bioinformatics/btu049.
Backman JD, Li AH, Marcketta A, Sun D, Mbatchou J, Kessler MD, et al. Exome sequencing
and analysis of 454,787 UK Biobank participants. Nature. 2021;599(7886):628-34. doi:
10.1038/s41586-021-04103-z.
Bai N, Kido T, Kavanagh TJ, Kaufman JD, Rosenfeld ME, van Breemen C, et al. Exposure to
diesel exhaust up-regulates iNOS expression in ApoE knockout mice. Toxicol Appl
Pharmacol. 2011;255(2):184-92. doi: 10.1016/j.taap.2011.06.013.
Bai N, Tranfield EM, Kavanagh TJ, Kaufman JD, Rosenfeld ME, van Eeden SF. Exposure to
diesel exhaust upregulates COX-2 expression in ApoE knockout mice. Inhal Toxicol.
2012;24(8):518-27. doi: 10.3109/08958378.2012.696221.
Banks E, Crouch SR, Korda RJ, Stavreski B, Page K, Thurber KA, et al. Absolute risk of
cardiovascular disease events, and blood pressure- and lipid-lowering therapy in
Australia. Med J Aust. 2016;204(8):320. doi: 10.5694/mja15.01004.
245
Bansal M. Cardiovascular disease and COVID-19. Diabetes Metab Syndr. 2020;14(3):247-50.
doi: 10.1016/j.dsx.2020.03.013.
Beelen R, Hoek G, Vienneau D, Eeftens M, Dimakopoulou K, Pedeli X, et al. Development of
NO2 and NOx land use regression models for estimating air pollution exposure in 36
study areas in Europe – The ESCAPE project. Atmos Environ. 2013;72:10-23. doi:
https://doi.org/10.1016/j.atmosenv.2013.02.037.
Bengalli R, Zerboni A, Marchetti S, Longhin E, Priola M, Camatini M, et al. In vitro pulmonary
and vascular effects induced by different diesel exhaust particles. Toxicol Lett.
2019;306:13-24. doi: 10.1016/j.toxlet.2019.01.017.
Bennett BJ, Davis RC, Civelek M, Orozco L, Wu J, Qi H, et al. Genetic architecture of
atherosclerosis in mice: A systems genetics analysis of common inbred strains. PLoS
Genet. 2015;11(12):e1005711. doi: 10.1371/journal.pgen.1005711.
Berger JS, Kornblith LZ, Gong MN, Reynolds HR, Cushman M, Cheng Y, et al. Effect of
P2Y12 Inhibitors on Survival Free of Organ Support Among Non-Critically Ill
Hospitalized Patients With COVID-19: A Randomized Clinical Trial. JAMA.
2022;327(3):227-36. doi: 10.1001/jama.2021.23605.
Berger JS, Roncaglioni MC, Avanzini F, Pangrazzi I, Tognoni G, Brown DL. Aspirin for the
primary prevention of cardiovascular events in women and men: a sex-specific metaanalysis of randomized controlled trials. JAMA. 2006;295(3):306-13. doi:
10.1001/jama.295.3.306.
Beristain-Covarrubias N, Perez-Toledo M, Thomas MR, Henderson IR, Watson SP,
Cunningham AF. Understanding infection-induced thrombosis: lessons learned from
animal models. Front Immunol. 2019;10:2569. doi: 10.3389/fimmu.2019.02569.
Beutner F, Teupser D, Gielen S, Holdt LM, Scholz M, Boudriot E, et al. Rationale and design of
the Leipzig (LIFE) Heart Study: phenotyping and cardiovascular characteristics of
patients with coronary artery disease. PLoS One. 2011;6(12):e29070. doi:
10.1371/journal.pone.0029070.
Beutner F, Teupser D, Gielen S, Holdt LM, Scholz M, Boudriot E, et al. Rationale and design of
the Leipzig (LIFE) Heart Study: phenotyping and cardiovascular characteristics of
patients with coronary artery disease. PLoS One. 2011;6(12):e29070. doi:
10.1371/journal.pone.0029070.
Bevan S, Traylor M, Adib-Samii P, Malik R, Paul NL, Jackson C, et al. Genetic heritability of
ischemic stroke and the contribution of previously reported candidate gene and
genomewide associations. Stroke. 2012;43(12):3161-7. doi:
10.1161/STROKEAHA.112.665760.
246
Bevan S, Traylor M, Adib-Samii P, Malik R, Paul NL, Jackson C, et al. Genetic heritability of
ischemic stroke and the contribution of previously reported candidate gene and
genomewide associations. Stroke. 2012;43(12):3161-7. doi:
10.1161/STROKEAHA.112.665760.
Bevan S, Traylor M, Adib-Samii P, Malik R, Paul NL, Jackson C, et al. Genetic heritability of
ischemic stroke and the contribution of previously reported candidate gene and
genomewide associations. Stroke. 2012;43(12):3161-7. doi:
10.1161/STROKEAHA.112.665760.
Bevan S, Traylor M, Adib-Samii P, Malik R, Paul NL, Jackson C, et al. Genetic heritability of
ischemic stroke and the contribution of previously reported candidate gene and
genomewide associations. Stroke. 2012;43(12):3161-7. doi:
10.1161/STROKEAHA.112.665760.
Bhetraratana M, Orozco LD, Bennett BJ, Luna K, Yang X, Lusis AJ, et al. Diesel exhaust
particle extract elicits an oxPAPC-like transcriptomic profile in macrophages across
multiple mouse strains. Environ Pollut. 2024;358:124415. doi:
10.1016/j.envpol.2024.124415.
Bhetraratana M, Orozco LD, Bennett BJ, Luna K, Yang X, Lusis AJ, et al. Diesel exhaust
particle extract elicits an oxPAPC-like transcriptomic profile in macrophages across
multiple mouse strains. Environ Pollut. 2024;358:124415. doi:
10.1016/j.envpol.2024.124415.
Birling MC, Yoshiki A, Adams DJ, Ayabe S, Beaudet AL, Bottomley J, et al. A resource of
targeted mutant mouse lines for 5,061 genes. Nat Genet. 2021;53(4):416-9. doi:
10.1038/s41588-021-00825-y.
Boardman HM, Hartley L, Eisinga A, Main C, Roque i Figuls M, Bonfill Cosp X, et al. Hormone
therapy for preventing cardiovascular disease in post-menopausal women. Cochrane
Database Syst Rev. 2015(3):CD002229. doi: 10.1002/14651858.CD002229.pub4.
Bohula EA, Berg DD, Lopes MS, Connors JM, Babar I, Barnett CF, et al. Anticoagulation and
Antiplatelet Therapy for Prevention of Venous and Arterial Thrombotic Events in
Critically Ill Patients With COVID-19: COVID-PACT. Circulation. 2022;146(18):1344-
56. doi: 10.1161/CIRCULATIONAHA.122.061533.
Boucher P, Matz RL, Terrand J. atherosclerosis: gone with the Wnt? Atherosclerosis.
2020;301:15-22. doi: 10.1016/j.atherosclerosis.2020.03.024.
Bowe B, Xie Y, Al-Aly Z. Postacute sequelae of COVID-19 at 2 years. Nat Med.
2023;29(9):2347-57. doi: 10.1038/s41591-023-02521-2.
Brauer M, Casadei B, Harrington RA, Kovacs R, Sliwa K, Group WHFAPE. Taking a stand
against air pollution-the impact on cardiovascular disease: a joint opinion from the World
247
Heart Federation, American College of Cardiology, American Heart Association, and the
European Society of Cardiology. Circulation. 2021;143(14):e800-e4. doi:
10.1161/CIRCULATIONAHA.120.052666.
Brauer M, Casadei B, Harrington RA, Kovacs R, Sliwa K, Group WHFAPE. Taking a Stand
Against Air Pollution-The Impact on Cardiovascular Disease: A Joint Opinion From the
World Heart Federation, American College of Cardiology, American Heart Association,
and the European Society of Cardiology. Circulation. 2021;143(14):e800-e4. doi:
10.1161/CIRCULATIONAHA.120.052666.
Britton HG, Garcia-Espana A, Goya P, Rozas I, Rubio V. A structure-reactivity study of the
binding of acetylglutamate to carbamoyl phosphate synthetase I. Eur J Biochem.
1990;188(1):47-53.
Brook RD, Rajagopalan S, Pope CA, Brook JR, Bhatnagar A, Diez-Roux AV, et al. Particulate
matter air pollution and cardiovascular disease: An update to the scientific statement from
the American Heart Association. Circulation. 2010;121(21):2331-78. doi:
10.1161/CIR.0b013e3181dbece1.
Brugts JJ, Yetgin T, Hoeks SE, Gotto AM, Shepherd J, Westendorp RG, et al. The benefits of
statins in people without established cardiovascular disease but with cardiovascular risk
factors: meta-analysis of randomised controlled trials. BMJ. 2009;338:b2376. doi:
10.1136/bmj.b2376.
Buniello A, MacArthur JAL, Cerezo M, Harris LW, Hayhurst J, Malangone C, et al. The
NHGRI-EBI GWAS Catalog of published genome-wide association studies, targeted
arrays and summary statistics 2019. Nucleic Acids Res. 2019;47(D1):D1005-D12. doi:
10.1093/nar/gky1120.
Buniello A, MacArthur JAL, Cerezo M, Harris LW, Hayhurst J, Malangone C, et al. The
NHGRI-EBI GWAS Catalog of published genome-wide association studies, targeted
arrays and summary statistics 2019. Nucleic Acids Res. 2019;47(D1):D1005-D12. doi:
10.1093/nar/gky1120.
Burgess S, Thompson SG. Interpreting findings from Mendelian randomization using the MREgger method. Eur J Epidemiol. 2017;32(5):377-89. doi: 10.1007/s10654-017-0255-x.
Bycroft C, Freeman C, Petkova D, Band G, Elliott LT, Sharp K, et al. The UK Biobank resource
with deep phenotyping and genomic data. Nature. 2018;562(7726):203-9. doi:
10.1038/s41586-018-0579-z.
Cacciottolo M, Wang X, Driscoll I, Woodward N, Saffari A, Reyes J, et al. Particulate air
pollutants, APOE alleles and their contributions to cognitive impairment in older women
and to amyloidogenesis in experimental models. Transl Psychiatry. 2017;7(1):e1022. doi:
10.1038/tp.2016.280.
248
Cassee FR, Campbell A, Boere AJ, McLean SG, Duffin R, Krystek P, et al. The biological
effects of subacute inhalation of diesel exhaust following addition of cerium oxide
nanoparticles in atherosclerosis-prone mice. Environ Res. 2012;115(1):1-10. doi:
10.1016/j.envres.2012.03.004.
Charchar FJ, Bloomer LD, Barnes TA, Cowley MJ, Nelson CP, Wang Y, et al. Inheritance of
coronary artery disease in men: an analysis of the role of the Y chromosome. Lancet.
2012;379(9819):915-22. doi: 10.1016/S0140-6736(11)61453-0.
Chen LC, Nadziejko C. Effects of subchronic exposures to concentrated ambient particles
(CAPs) in mice. V. CAPs exacerbate aortic plaque development in hyperlipidemic mice.
Inhal Toxicol. 2005;17(4-5):217-24. doi: 10.1080/08958370590912815.
Chew DP, Scott IA, Cullen L, French JK, Briffa TG, Tideman PA, et al. National Heart
Foundation of Australia & Cardiac Society of Australia and New Zealand: Australian
clinical guidelines for the management of acute coronary syndromes 2016. Heart Lung
Circ. 2016;25(9):895-951. doi: 10.1016/j.hlc.2016.06.789.
Clarke B, Holtkamp E, Öztürk H, Mück M, Wahlberg M, Meyer K, et al. Integration of variant
annotations using deep set networks boosts rare variant association testing. Nat Genet.
2024;56(10):2271-80. doi: 10.1038/s41588-024-01919-z.
Clayton TC, Gaskin M, Meade TW. Recent respiratory infection and risk of venous
thromboembolism: case-control study through a general practice database. Int J
Epidemiol. 2011;40(3):819-27. doi: 10.1093/ije/dyr012.
Collaborators GBDRF. Global burden and strength of evidence for 88 risk factors in 204
countries and 811 subnational locations, 1990-2021: a systematic analysis for the Global
Burden of Disease Study 2021. Lancet. 2024;403(10440):2162-203. doi: 10.1016/S0140-
6736(24)00933-4.
Connors JM, Brooks MM, Sciurba FC, Krishnan JA, Bledsoe JR, Kindzelski A, et al. Effect of
Antithrombotic Therapy on Clinical Outcomes in Outpatients With Clinically Stable
Symptomatic COVID-19: The ACTIV-4B Randomized Clinical Trial. JAMA.
2021;326(17):1703-12. doi: 10.1001/jama.2021.17272.
Consortium G. The Genotype-Tissue Expression (GTEx) project. Nat Genet. 2013;45(6):580-5.
doi: 10.1038/ng.2653.
Consortium WTCC. Genome-wide association study of 14,000 cases of seven common diseases
and 3,000 shared controls. Nature. 2007;447(7145):661-78. doi: 10.1038/nature05911.
Cosselman KE, Navas-Acien A, Kaufman JD. Environmental factors in cardiovascular disease.
Nat Rev Cardiol. 2015;12(11):627-42. doi: 10.1038/nrcardio.2015.152.
249
Dalager-Pedersen M, Sogaard M, Schonheyder HC, Nielsen H, Thomsen RW. Risk for
myocardial infarction and stroke after community-acquired bacteremia: a 20-year
population-based cohort study. Circulation. 2014;129(13):1387-96. doi:
10.1161/CIRCULATIONAHA.113.006699.
Dalager-Pedersen M, Sogaard M, Schonheyder HC, Nielsen H, Thomsen RW. Risk for
myocardial infarction and stroke after community-acquired bacteremia: a 20-year
population-based cohort study. Circulation. 2014;129(13):1387-96. doi:
10.1161/CIRCULATIONAHA.113.006699.
Davignon L, Chauveau C, Julien C, Dill C, Duband-Goulet I, Cabet E, et al. The transcription
coactivator ASC-1 is a regulator of skeletal myogenesis, and its deficiency causes a novel
form of congenital muscle disease. Hum Mol Genet. 2016;25(8):1559-73. doi:
10.1093/hmg/ddw033.
Denny JC, Rutter JL, Goldstein DB, Philippakis A, Smoller JW, Jenkins G, et al. The "All of Us"
Research Program. N Engl J Med. 2019;381(7):668-76. doi: 10.1056/NEJMsr1809937.
Dong E, Du H, Gardner L. An interactive web-based dashboard to track COVID-19 in real time.
Lancet Infect Dis. 2020;20(5):533-4. doi: 10.1016/S1473-3099(20)30120-1.
Du P, Kibbe WA, Lin SM. lumi: a pipeline for processing Illumina microarray. Bioinformatics.
2008;24(13):1547-8. doi: 10.1093/bioinformatics/btn224.
Duo H, Jin M, Yang Y, Baheti R, Feng Y, Fu Z, et al. Effect of antiplatelet therapy after
COVID-19 diagnosis: A systematic review with meta-analysis and trial sequential
analysis. PLoS One. 2024;19(2):e0297628. doi: 10.1371/journal.pone.0297628.
Eberhardt N, Noval MG, Kaur R, Amadori L, Gildea M, Sajja S, et al. SARS-CoV-2 infection
triggers pro-atherogenic inflammatory responses in human coronary vessels. Nat
Cardiovasc Res. 2023;2(10):899-916. doi: 10.1038/s44161-023-00336-5.
Eeftens M, Beelen R, de Hoogh K, Bellander T, Cesaroni G, Cirach M, et al. Development of
land use regression models for PM(2.5), PM(2.5) absorbance, PM(10) and PM(coarse) in
20 European study areas; results of the ESCAPE project. Environ Sci Technol.
2012;46(20):11195-205. doi: 10.1021/es301948k.
Eicher JD, Chen MH, Pitsillides AN, Lin H, Veeraraghavan N, Brody JA, et al. Whole exome
sequencing in the Framingham Heart Study identifies rare variation in HYAL2 that
influences platelet aggregation. Thromb Haemost. 2017;117(6):1083-92. doi:
10.1160/TH16-09-0677.
El Khoudary SR, Aggarwal B, Beckie TM, Hodis HN, Johnson AE, Langer RD, et al.
Menopause Transition and Cardiovascular Disease Risk: Implications for Timing of
Early Prevention: A Scientific Statement From the American Heart Association.
Circulation. 2020;142(25):e506-e32. doi: 10.1161/CIR.0000000000000912.
250
ElSayed NA, Aleppo G, Aroda VR, Bannuru RR, Brown FM, Bruemmer D, et al. 10.
Cardiovascular disease and risk management: Standards of care in diabetes - 2023.
Diabetes Care. 2023;46(Suppl 1):S158-S90. doi: 10.2337/dc23-S010.
Farraj AK, Haykal-Coates N, Ledbetter AD, Evansky PA, Gavett SH. Neurotrophin mediation of
allergic airways responses to inhaled diesel particles in mice. Toxicol Sci.
2006;94(1):183-92. doi: 10.1093/toxsci/kfl089.
Ferrell M, Wang Z, Anderson JT, Li XS, Witkowski M, DiDonato JA, et al. A terminal
metabolite of niacin promotes vascular inflammation and contributes to cardiovascular
disease risk. Nat Med. 2024;30(2):424-34. doi: 10.1038/s41591-023-02793-8.
Finch CE, Thorwald MA. Inhaled pollutants of the gero-exposome and later-life health. J
Gerontol A Biol Sci Med Sci. 2024;79(7). doi: 10.1093/gerona/glae107.
Fortin JP, Labbe A, Lemire M, Zanke BW, Hudson TJ, Fertig EJ, et al. Functional normalization
of 450k methylation array data improves replication in large cancer studies. Genome
biology. 2014;15(12):503. doi: 10.1186/s13059-014-0503-2.
Fortin JP, Triche TJ, Jr., Hansen KD. Preprocessing, normalization and integration of the
Illumina HumanMethylationEPIC array with minfi. Bioinformatics. 2017;33(4):558-60.
doi: 10.1093/bioinformatics/btw691.
Franzen O, Ermel R, Cohain A, Akers NK, Di Narzo A, Talukdar HA, et al. Cardiometabolic
risk loci share downstream cis- and trans-gene regulation across tissues and diseases.
Science. 2016;353(6301):827-30. doi: 10.1126/science.aad6970.
Fuller R, Landrigan PJ, Balakrishnan K, Bathan G, Bose-O'Reilly S, Brauer M, et al. Pollution
and health: a progress update. Lancet Planet Health. 2022;6(6):e535-e47. doi:
10.1016/S2542-5196(22)00090-0.
Gao F, Ihn HE, Medina MW, Krauss RM. A common polymorphism in the LDL receptor gene
has multiple effects on LDL receptor function. Hum Mol Genet. 2013;22(7):1424-31.
doi: 10.1093/hmg/dds559.
Gao X, Jia M, Zhang Y, Breitling LP, Brenner H. DNA methylation changes of whole blood
cells in response to active smoking exposure in adults: a systematic review of DNA
methylation studies. Clin Epigenetics. 2015;7:113. doi: 10.1186/s13148-015-0148-3.
Gelehrter TD, Snodgrass PJ. Lethal neonatal deficiency of carbamyl phosphate synthetase. N
Engl J Med. 1974;290(8):430-3. doi: 10.1056/NEJM197402212900804.
Giacca M, Shah AM. The pathological maelstrom of COVID-19 and cardiovascular disease. Nat
Cardiovasc Res. 2022;1:200-10.
251
Go AS, Mozaffarian D, Roger VL, Benjamin EJ, Berry JD, Blaha MJ, et al. Heart disease and
stroke statistics--2014 update: a report from the American Heart Association. Circulation.
2014;129(3):e28-e292. doi: 10.1161/01.cir.0000441139.02102.80.
Goel R, Bloch EM, Pirenne F, Al-Riyami AZ, Crowe E, Dau L, et al. ABO blood group and
COVID-19: a review on behalf of the ISBT COVID-19 Working Group. Vox Sang.
2021;116(8):849-61. doi: 10.1111/vox.13076.
Goker H, Aladag Karakulak E, Demiroglu H, Ayaz Ceylan CM, Buyukasik Y, Inkaya AC, et al.
The effects of blood group types on the risk of COVID-19 infection and its clinical
outcome. Turk J Med Sci. 2020;50(4):679-83. doi: 10.3906/sag-2005-395.
Goliasch G, Wiesbauer F, Kastl SP, Katsaros KM, Blessberger H, Maurer G, et al. Premature
myocardial infarction is associated with low serum levels of Wnt-1. Atherosclerosis.
2012;222(1):251-6. doi: 10.1016/j.atherosclerosis.2012.02.017.
Gondalia R, Avery CL, Napier MD, Mendez-Giraldez R, Stewart JD, Sitlani CM, et al. Genomewide association study of susceptibility to particulate matter-associated QT prolongation.
Environ Health Perspect. 2017;125(6):067002. doi: 10.1289/EHP347.
Gong KW, Zhao W, Li N, Barajas B, Kleinman M, Sioutas C, et al. Air-pollutant chemicals and
oxidized lipids exhibit genome-wide synergistic effects on endothelial cells. Genome
biology. 2007;8(7):R149. doi: 10.1186/gb-2007-8-7-r149.
Groot HE, Villegas Sierra LE, Said MA, Lipsic E, Karper JC, van der Harst P. Genetically
determined ABO blood group and its associations with health and disease. Arterioscler
Thromb Vasc Biol. 2020;40(3):830-8. doi: 10.1161/ATVBAHA.119.313658.
Group RC. Aspirin in patients admitted to hospital with COVID-19 (RECOVERY): a
randomised, controlled, open-label, platform trial. Lancet. 2022;399(10320):143-51. doi:
10.1016/S0140-6736(21)01825-0.
Grundy SM, Stone NJ, Bailey AL, Beam C, Birtcher KK, Blumenthal RS, et al. 2018
AHA/ACC/AACVPR/AAPA/ABC/ACPM/ADA/AGS/APhA/ASPC/NLA/PCNA
guideline on the management of blood cholesterol: executive summary: A report of the
American College of Cardiology/American Heart Association task force on clinical
practice guidelines. Circulation. 2019;139(25):e1046-e81. doi:
10.1161/CIR.0000000000000624.
Grundy SM, Stone NJ, Bailey AL, Beam C, Birtcher KK, Blumenthal RS, et al. 2018
AHA/ACC/AACVPR/AAPA/ABC/ACPM/ADA/AGS/APhA/ASPC/NLA/PCNA
Guideline on the Management of Blood Cholesterol: Executive Summary: A Report of
the American College of Cardiology/American Heart Association Task Force on Clinical
Practice Guidelines. Circulation. 2019;139(25):e1046-e81. doi:
10.1161/CIR.0000000000000624.
252
Grundy SM, Stone NJ, Bailey AL, Beam C, Birtcher KK, Blumenthal RS, et al. 2018
AHA/ACC/AACVPR/AAPA/ABC/ACPM/ADA/AGS/APhA/ASPC/NLA/PCNA
guideline on the management of blood cholesterol: executive summary: A report of the
American College of Cardiology/American Heart Association task force on clinical
practice guidelines. Circulation. 2019;139(25):e1046-e81. doi:
10.1161/CIR.0000000000000624.
Haffner SM, Lehto S, Ronnemaa T, Pyorala K, Laakso M. Mortality from coronary heart disease
in subjects with type 2 diabetes and in nondiabetic subjects with and without prior
myocardial infarction. N Engl J Med. 1998;339(4):229-34. doi:
10.1056/NEJM199807233390404.
Han Y, Jia Q, Jahani PS, Hurrell BP, Pan C, Huang P, et al. Genome-wide analysis highlights
contribution of immune system pathways to the genetic architecture of asthma. Nat
Commun. 2020;11(1):1776. doi: 10.1038/s41467-020-15649-3.
Hansen CS, Sheykhzade M, Moller P, Folkmann JK, Amtorp O, Jonassen T, et al. Diesel exhaust
particles induce endothelial dysfunction in apoE-/- mice. Toxicol Appl Pharmacol.
2007;219(1):24-32. doi: 10.1016/j.taap.2006.10.032.
Hansson GK. Inflammation, atherosclerosis, and coronary artery disease. N Engl J Med.
2005;352(16):1685-95.
Hartiala J, Bennett BJ, Tang WH, Wang Z, Stewart AF, Roberts R, et al. Comparative genomewide association studies in mice and humans for trimethylamine N-oxide, a
proatherogenic metabolite of choline and L-carnitine. Arterioscler Thromb Vasc Biol.
2014;34(6):1307-13. doi: 10.1161/ATVBAHA.114.303252.
Hartiala J, Bennett BJ, Tang WH, Wang Z, Stewart AF, Roberts R, et al. Comparative genomewide association studies in mice and humans for trimethylamine N-oxide, a
proatherogenic metabolite of choline and L-carnitine. Arterioscler Thromb Vasc Biol.
2014;34(6):1307-13. doi: 10.1161/ATVBAHA.114.303252.
Hartiala J, Breton CV, Tang WH, Lurmann F, Hazen SL, Gilliland FD, et al. Ambient air
pollution is associated with the severity of coronary atherosclerosis and incident
myocardial infarction in patients undergoing elective cardiac evaluation. J Am Heart
Assoc. 2016;5(8). doi: 10.1161/JAHA.116.003947.
Hartiala J, Li D, Conti DV, Vikman S, Patel Y, Tang WH, et al. Genetic contribution of the
leukotriene pathway to coronary artery disease. Hum Genet. 2011;129(6):617-27.
Hartiala J, Schwartzman WS, Gabbay J, Ghazalpour A, Bennett BJ, Allayee H. The genetic
architecture of coronary artery disease: current knowledge and future opportunities. Curr
Atheroscler Rep. 2017;19(2):6. doi: 10.1007/s11883-017-0641-6.
253
Hartiala JA, Han Y, Jia Q, Hilser JR, Huang P, Gukasyan J, et al. Genome-wide analysis
identifies novel susceptibility loci for myocardial infarction. Eur Heart J. 2021;42(9):919-
33. doi: 10.1093/eurheartj/ehaa1040.
Hartiala JA, Hilser JR, Biswas S, Lusis AJ, Allayee H. Gene-environment interactions for
cardiovascular disease. Curr Atheroscler Rep. 2021;23(12):75. doi: 10.1007/s11883-021-
00974-9.
Hartiala JA, Tang WH, Wang Z, Crow AL, Stewart AF, Roberts R, et al. Genome-wide
association study and targeted metabolomics identifies sex-specific association of CPS1
with coronary artery disease. Nat Commun. 2016;7:10558. doi: 10.1038/ncomms10558.
Hartz AM, Bauer B, Block ML, Hong JS, Miller DS. Diesel exhaust particles induce oxidative
stress, proinflammatory signaling, and P-glycoprotein up-regulation at the blood-brain
barrier. FASEB J. 2008;22(8):2723-33. doi: 10.1096/fj.08-106997.
Helgadottir A, Thorleifsson G, Manolescu A, Gretarsdottir S, Blondal T, Jonasdottir A, et al. A
common variant on chromosome 9p21 affects the risk of myocardial infarction. Science.
2007;316(5830):1491-3. doi: 10.1126/science.1142842.
Hilser JR, Han Y, Biswas S, Gukasyan J, Cai Z, Zhu R, et al. Association of serum HDLcholesterol and apolipoprotein A1 levels with risk of severe SARS-CoV-2 infection. J
Lipid Res. 2021;62:100061. doi: 10.1016/j.jlr.2021.100061.
Hilser JR, Hartiala JA, Sriprasert I, Kono N, Cai Z, Karim R, et al. Effect of menopausal
hormone therapy on methylation levels in early and late postmenopausal women. Clin
Epigenetics. 2022;14(1):90. doi: 10.1186/s13148-022-01311-w.
Hilser JR, Spencer NJ, Afshari K, Gilliland FD, Hu H, Deb A, et al. COVID-19 Is a coronary
artery disease risk equivalent and exhibits a genetic interaction with abo blood type.
Arterioscler Thromb Vasc Biol. 2024. doi: 10.1161/ATVBAHA.124.321001.
Hiltunen MO, Turunen MP, Hakkinen TP, Rutanen J, Hedman M, Makinen K, et al. DNA
hypomethylation and methyltransferase expression in atherosclerotic lesions. Vasc Med.
2002;7(1):5-11. doi: 10.1191/1358863x02vm418oa.
Hodis HN, Mack WJ, Henderson VW, Shoupe D, Budoff MJ, Hwang-Levine J, et al. Vascular
Effects of Early versus Late Postmenopausal Treatment with Estradiol. N Engl J Med.
2016;374(13):1221-31. doi: 10.1056/NEJMoa1505241.
Hodis HN, Mack WJ, Lobo R. Antiatherosclerosis interventions in women. Am J Cardiol.
2002;90(1A):17F-21F.
Hodis HN, Mack WJ, Shoupe D, Azen SP, Stanczyk FZ, Hwang-Levine J, et al. Methods and
baseline cardiovascular data from the Early versus Late Intervention Trial with Estradiol
254
testing the menopausal hormone timing hypothesis. Menopause. 2015;22(4):391-401.
doi: 10.1097/GME.0000000000000343.
Hougaard KS, Jensen KA, Nordly P, Taxvig C, Vogel U, Saber AT, et al. Effects of prenatal
exposure to diesel exhaust particles on postnatal development, behavior, genotoxicity and
inflammation in mice. Part Fibre Toxicol. 2008;5:3. doi: 10.1186/1743-8977-5-3.
Hougaard KS, Jensen KA, Nordly P, Taxvig C, Vogel U, Saber AT, et al. Effects of prenatal
exposure to diesel exhaust particles on postnatal development, behavior, genotoxicity and
inflammation in mice. Part Fibre Toxicol. 2008;5:3. doi: 10.1186/1743-8977-5-3.
Houseman EA, Accomando WP, Koestler DC, Christensen BC, Marsit CJ, Nelson HH, et al.
DNA methylation arrays as surrogate measures of cell mixture distribution. BMC
Bioinformatics. 2012;13:86. doi: 10.1186/1471-2105-13-86.
Housworth J, Langmuir AD. Excess mortality from epidemic influenza, 1957-1966. Am J
Epidemiol. 1974;100(1):40-8. doi: 10.1093/oxfordjournals.aje.a112007.
Howson JMM, Zhao W, Barnes DR, Ho WK, Young R, Paul DS, et al. Fifteen new risk loci for
coronary artery disease highlight arterial-wall-specific mechanisms. Nat Genet.
2017;49(7):1113-9. doi: 10.1038/ng.3874.
Hu FB, Rimm EB, Stampfer MJ, Ascherio A, Spiegelman D, Willett WC. Prospective study of
major dietary patterns and risk of coronary heart disease in men. Am J Clin Nutr.
2000;72(4):912-21. doi: 10.1093/ajcn/72.4.912.
Hu FB, Stampfer MJ, Manson JE, Rimm E, Colditz GA, Rosner BA, et al. Dietary fat intake and
the risk of coronary heart disease in women. N Engl J Med. 1997;337(21):1491-9. doi:
10.1056/NEJM199711203372102.
Huang Y, Hui Q, Gwinn M, Hu YJ, Quyyumi AA, Vaccarino V, et al. Sexual differences in
genetic predisposition of coronary artery disease. Circ Genom Precis Med.
2021;14(1):e003147. doi: 10.1161/CIRCGEN.120.003147.
Husby A. On the Use of Blood Samples for Measuring DNA Methylation in Ecological
Epigenetic Studies. Integr Comp Biol. 2020;60(6):1558-66. doi: 10.1093/icb/icaa123.
Initiative C-HG. A second update on mapping the human genetic architecture of COVID-19.
Nature. 2023;621(7977):E7-E26. doi: 10.1038/s41586-023-06355-3.
Investigators R-CWCftR-C, Bradbury CA, Lawler PR, Stanworth SJ, McVerry BJ, McQuilten Z,
et al. Effect of Antiplatelet Therapy on Survival and Organ Support-Free Days in
Critically Ill Patients With COVID-19: A Randomized Clinical Trial. JAMA.
2022;327(13):1247-59. doi: 10.1001/jama.2022.2910.
255
Jian X, Satizabal CL, Smith AV, Wittfeld K, Bis JC, Smith JA, et al. Exome chip analysis
identifies low-frequency and rare variants in mrpl38 for white matter hyperintensities on
brain magnetic resonance imaging. Stroke. 2018;49(8):1812-9. doi:
10.1161/STROKEAHA.118.020689.
Jung DJ, Sung HS, Goo YW, Lee HM, Park OK, Jung SY, et al. Novel transcription coactivator
complex containing activating signal cointegrator 1. Mol Cell Biol. 2002;22(14):5203-11.
doi: 10.1128/MCB.22.14.5203-5211.2002.
Kasela S, Daniloski Z, Bollepalli S, Jordan TX, tenOever BR, Sanjana NE, et al. Integrative
approach identifies SLC6A20 and CXCR6 as putative causal genes for the COVID-19
GWAS signal in the 3p21.31 locus. Genome biology. 2021;22(1):242. doi:
10.1186/s13059-021-02454-4.
Kaufman JD, Elkind MSV, Bhatnagar A, Koehler K, Balmes JR, Sidney S, et al. Guidance to
Reduce the Cardiovascular Burden of Ambient Air Pollutants: A Policy Statement From
the American Heart Association. Circulation. 2020;142(23):e432-e47. doi:
10.1161/CIR.0000000000000930.
Kennedy DW, White NM, Benton MC, Fox A, Scott RJ, Griffiths LR, et al. Critical evaluation
of linear regression models for cell-subtype specific methylation signal from mixed blood
cell DNA. PLoS One. 2018;13(12):e0208915. doi: 10.1371/journal.pone.0208915.
Kerin M, Marchini J. A non-linear regression method for estimation of gene-environment
heritability. Bioinformatics. 2020;36(4):5632-9. doi: 10.1093/bioinformatics/btaa1079.
Kerin M, Marchini J. Gene-by-environment interactions with a bayesian whole-genome
regression model. Am J Hum Genet. 2020;107(4):698-713. doi:
10.1016/j.ajhg.2020.08.009.
Khoo US, Chan KY, Chan VS, Lin CL. DC-SIGN and L-SIGN: the SIGNs for infection. J Mol
Med (Berl). 2008;86(8):861-74. doi: 10.1007/s00109-008-0350-2.
Khraishah H, Chen Z, Rajagopalan S. Understanding the cardiovascular and metabolic health
effects of air pollution in the context of cumulative exposomic impacts. Circ Res.
2024;134(9):1083-97. doi: 10.1161/CIRCRESAHA.124.323673.
Kido T, Bai N, Yatera K, Suzuki H, Meredith A, Mukae H, et al. Diesel exhaust inhalation
induces heat shock protein 70 expression in vivo. Inhal Toxicol. 2011;23(10):593-601.
doi: 10.3109/08958378.2011.595843.
Kim J, Kim JY, Song KS, Lee YH, Seo JS, Jelinek J, et al. Epigenetic changes in estrogen
receptor beta gene in atherosclerotic cardiovascular tissues and in-vitro vascular
senescence. Biochim Biophys Acta. 2007;1772(1):72-80. doi:
10.1016/j.bbadis.2006.10.004.
256
Kim Y, Latz CA, DeCarlo CS, Lee S, Png CYM, Kibrik P, et al. Relationship between blood
type and outcomes following COVID-19 infection. Semin Vasc Surg. 2021;34(3):125-31.
doi: 10.1053/j.semvascsurg.2021.05.005.
Kinasewitz GT, Yan SB, Basson B, Comp P, Russell JA, Cariou A, et al. Universal changes in
biomarkers of coagulation and inflammation occur in patients with severe sepsis,
regardless of causative micro-organism [ISRCTN74215569]. Crit Care. 2004;8(2):R82-
90. doi: 10.1186/cc2459.
Kinasewitz GT, Yan SB, Basson B, Comp P, Russell JA, Cariou A, et al. Universal changes in
biomarkers of coagulation and inflammation occur in patients with severe sepsis,
regardless of causative micro-organism [ISRCTN74215569]. Crit Care. 2004;8(2):R82-
90. doi: 10.1186/cc2459.
Klarin D, Lynch J, Aragam K, Chaffin M, Assimes TL, Huang J, et al. Genome-wide association
study of peripheral artery disease in the Million Veteran Program. Nat Med.
2019;25(8):1274-9. doi: 10.1038/s41591-019-0492-5.
Klarin D, Zhu QM, Emdin CA, Chaffin M, Horner S, McMillan BJ, et al. Genetic analysis in UK
Biobank links insulin resistance and transendothelial migration pathways to coronary
artery disease. Nat Genet. 2017;49(9):1392-7. doi: 10.1038/ng.3914.
Knierim E, Hirata H, Wolf NI, Morales-Gonzalez S, Schottmann G, Tanaka Y, et al. Mutations
in subunits of the activating signal cointegrator 1 complex are associated with prenatal
spinal muscular atrophy and congenital bone fractures. Am J Hum Genet.
2016;98(3):473-89. doi: 10.1016/j.ajhg.2016.01.006.
Knight R, Walker V, Ip S, Cooper JA, Bolton T, Keene S, et al. Association of COVID-19 with
major arterial and venous thrombotic diseases: a population-wide cohort study of 48
million adults in England and Wales. Circulation. 2022;146(12):892-906. doi:
10.1161/CIRCULATIONAHA.122.060785.
Koh N, Ference BA, Nicholls SJ, Navar AM, Chew DP, Kostner K, et al. Asian pacific society
of cardiology consensus recommendations on dyslipidaemia. Eur Cardiol. 2021;16:e54.
doi: 10.15420/ecr.2021.36.
Koplev S, Seldin M, Sukhavasi K, Ermel R, Pang S, Zeng L, et al. A mechanistic framework for
cardiometabolic and coronary artery diseases. Nat Cardiovasc Res. 2022;1(1):85-100.
doi: 10.1038/s44161-021-00009-1.
Kovacs T, Szabo-Meleg E, Abraham IM. Estradiol-induced epigenetically mediated mechanisms
and regulation of gene expression. Int J Mol Sci. 2020;21(9). doi: 10.3390/ijms21093177.
Koyama S, Ito K, Terao C, Akiyama M, Horikoshi M, Momozawa Y, et al. Population-specific
and trans-ancestry genome-wide analyses identify distinct and shared genetic risk loci for
257
coronary artery disease. Nat Genet. 2020;52(11):1169-77. doi: 10.1038/s41588-020-
0705-3.
Koyama S, Ito K, Terao C, Akiyama M, Horikoshi M, Momozawa Y, et al. Population-specific
and trans-ancestry genome-wide analyses identify distinct and shared genetic risk loci for
coronary artery disease. Nat Genet. 2020;52(11):1169-77. doi: 10.1038/s41588-020-
0705-3.
Koyama S, Ito K, Terao C, Akiyama M, Horikoshi M, Momozawa Y, et al. Population-specific
and trans-ancestry genome-wide analyses identify distinct and shared genetic risk loci for
coronary artery disease. Nat Genet. 2020;52(11):1169-77. doi: 10.1038/s41588-020-
0705-3.
Kumar A, Narayan RK, Kumari C, Faiq MA, Kulandhasamy M, Kant K, et al. SARS-CoV-2 cell
entry receptor ACE2 mediated endothelial dysfunction leads to vascular thrombosis in
COVID-19 patients. Med Hypotheses. 2020;145:110320. doi:
10.1016/j.mehy.2020.110320.
Kurki MI, Karjalainen J, Palta P, Sipila TP, Kristiansson K, Donner KM, et al. FinnGen provides
genetic insights from a well-phenotyped isolated population. Nature.
2023;613(7944):508-18. doi: 10.1038/s41586-022-05473-8.
Kwong JC, Schwartz KL, Campitelli MA, Chung H, Crowcroft NS, Karnauchow T, et al. Acute
Myocardial Infarction after Laboratory-Confirmed Influenza Infection. N Engl J Med.
2018;378(4):345-53. doi: 10.1056/NEJMoa1702090.
Latz CA, DeCarlo C, Boitano L, Png CYM, Patell R, Conrad MF, et al. Blood type and
outcomes in patients with COVID-19. Ann Hematol. 2020;99(9):2113-8. doi:
10.1007/s00277-020-04169-1.
Lettre G, Palmer CD, Young T, Ejebe KG, Allayee H, Benjamin EJ, et al. Genome-wide
association study of coronary heart disease and its risk factors in 8,090 African
Americans: the NHLBI CARe Project. PLoS Genet. 2011;7(2):e1001300.
Levinsson A, Olin AC, Modig L, Dahgam S, Bjorck L, Rosengren A, et al. Interaction effects of
long-term air pollution exposure and variants in the GSTP1, GSTT1 and GSTCD genes
on risk of acute myocardial infarction and hypertension: a case-control study. PLoS One.
2014;9(6):e99043. doi: 10.1371/journal.pone.0099043.
Li J, Zhang X, Yang M, Yang H, Xu N, Fan X, et al. DNA methylome profiling reveals
epigenetic regulation of lipoprotein-associated phospholipase A2 in human vulnerable
atherosclerotic plaque. Clin Epigenetics. 2021;13(1):161. doi: 10.1186/s13148-021-
01152-z.
Li J, Zhang X, Yang M, Yang H, Xu N, Fan X, et al. DNA methylome profiling reveals
epigenetic regulation of lipoprotein-associated phospholipase A2 in human vulnerable
258
atherosclerotic plaque. Clin Epigenetics. 2021;13(1):161. doi: 10.1186/s13148-021-
01152-z.
Lin Y, Ramanathan G, Zhu Y, Yin F, Rea ND, Lu X, et al. Pro-Oxidative and Proinflammatory
Effects After Traveling From Los Angeles to Beijing: A Biomarker-Based Natural
Experiment. Circulation. 2019;140(24):1995-2004. doi:
10.1161/CIRCULATIONAHA.119.042054.
Link E, Parish S, Armitage J, Bowman L, Heath S, Matsuda F, et al. SLCO1B1 variants and
statin-induced myopathy--a genomewide study. N Engl J Med. 2008;359(8):789-99. doi:
10.1056/NEJMoa0801936.
Liu Y, Wang T, Si B, Du H, Liu Y, Waqas A, et al. Intratracheally instillated diesel PM(2.5)
significantly altered the structure and composition of indigenous murine gut microbiota.
Ecotoxicol Environ Saf. 2021;210:111903. doi: 10.1016/j.ecoenv.2021.111903.
Liu Y, Wang T, Si B, Du H, Liu Y, Waqas A, et al. Intratracheally instillated diesel PM(2.5)
significantly altered the structure and composition of indigenous murine gut microbiota.
Ecotoxicol Environ Saf. 2021;210:111903. doi: 10.1016/j.ecoenv.2021.111903.
Lobo RA, Davis SR, De Villiers TJ, Gompel A, Henderson VW, Hodis HN, et al. Prevention of
diseases after menopause. Climacteric. 2014;17(5):540-56. doi:
10.3109/13697137.2014.933411.
Loffi M, Piccolo R, Regazzoni V, Di Tano G, Moschini L, Robba D, et al. Coronary artery
disease in patients hospitalised with Coronavirus disease 2019 (COVID-19) infection.
Open Heart. 2020;7(2). doi: 10.1136/openhrt-2020-001428.
Loley C, Alver M, Assimes TL, Bjonnes A, Goel A, Gustafsson S, et al. No Association of
Coronary Artery Disease with X-Chromosomal Variants in Comprehensive International
Meta-Analysis. Sci Rep. 2016;6:35278. doi: 10.1038/srep35278.
Long B, Brady WJ, Koyfman A, Gottlieb M. Cardiovascular complications in COVID-19. Am J
Emerg Med. 2020;38(7):1504-7. doi: 10.1016/j.ajem.2020.04.048.
Lusis AJ, Seldin MM, Allayee H, Bennett BJ, Civelek M, Davis RC, et al. The Hybrid Mouse
Diversity Panel: a resource for systems genetics analyses of metabolic and cardiovascular
traits. J Lipid Res. 2016;57(6):925-42. doi: 10.1194/jlr.R066944.
Ma WF, Hodonsky CJ, Turner AW, Wong D, Song Y, Mosquera JV, et al. Enhanced single-cell
RNA-seq workflow reveals coronary artery disease cellular cross-talk and candidate drug
targets. Atherosclerosis. 2022;340:12-22. doi: 10.1016/j.atherosclerosis.2021.11.025.
Maazi H, Hartiala JA, Suzuki Y, Crow AL, Shafiei Jahani P, Lam J, et al. A GWAS approach
identifies Dapp1 as a determinant of air pollution-induced airway hyperreactivity. PLoS
Genet. 2019;15(12):e1008528. doi: 10.1371/journal.pgen.1008528.
259
Madsen CM, Varbo A, Nordestgaard BG. Extreme high high-density lipoprotein cholesterol is
paradoxically associated with high mortality in men and women: two prospective cohort
studies. Eur Heart J. 2017;38(32):2478-86. doi: 10.1093/eurheartj/ehx163.
Malik R, Chauhan G, Traylor M, Sargurupremraj M, Okada Y, Mishra A, et al. Multiancestry
genome-wide association study of 520,000 subjects identifies 32 loci associated with
stroke and stroke subtypes. Nat Genet. 2018;50(4):524-37. doi: 10.1038/s41588-018-
0058-3.
Martin SS, Aday AW, Almarzooq ZI, Anderson CAM, Arora P, Avery CL, et al. 2024 Heart
Disease and Stroke Statistics: A Report of US and Global Data From the American Heart
Association. Circulation. 2024;149(8):e347-e913. doi: 10.1161/CIR.0000000000001209.
Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis
model assessment: insulin resistance and beta-cell function from fasting plasma glucose
and insulin concentrations in man. Diabetologia. 1985;28(7):412-9.
Mayo KR, Basford MA, Carroll RJ, Dillon M, Fullen H, Leung J, et al. The All of Us Data and
Research Center: creating a secure, scalable, and sustainable ecosystem for biomedical
research. Annu Rev Biomed Data Sci. 2023;6:443-64. doi: 10.1146/annurev-biodatasci122120-104825.
Mbatchou J, Barnard L, Backman J, Marcketta A, Kosmicki JA, Ziyatdinov A, et al.
Computationally efficient whole-genome regression for quantitative and binary traits. Nat
Genet. 2021;53(7):1097-103. doi: 10.1038/s41588-021-00870-7.
McCaffrey DF, Ridgeway G, Morral AR. Propensity score estimation with boosted regression
for evaluating causal effects in observational studies. Psychol Methods. 2004;9(4):403-
25. doi: 10.1037/1082-989X.9.4.403.
McPherson R, Pertsemlidis A, Kavaslar N, Stewart A, Roberts R, Cox DR, et al. A common
allele on chromosome 9 associated with coronary heart disease. Science.
2007;316(5830):1488-91. doi: 10.1126/science.1142447.
Mehrabian M, Allayee H, Wong J, Shi W, Wang XP, Shaposhnik Z, et al. Identification of 5-
lipoxygenase as a major gene contributing to atherosclerosis susceptibility in mice. Circ
Res. 2002;91(2):120-6.
Mehrabian M, Schulthess FT, Nebohacova M, Castellani LW, Zhou Z, Hartiala J, et al.
Identification of ALOX5 as a gene regulating adiposity and pancreatic function.
Diabetologia. 2008;51(6):978-88.
Miller MR, McLean SG, Duffin R, Lawal AO, Araujo JA, Shaw CA, et al. Diesel exhaust
particulate increases the size and complexity of lesions in atherosclerotic mice. Part Fibre
Toxicol. 2013;10:61. doi: 10.1186/1743-8977-10-61.
260
Min JL, Hemani G, Davey Smith G, Relton C, Suderman M. Meffil: efficient normalization and
analysis of very large DNA methylation datasets. Bioinformatics. 2018;34(23):3983-9.
doi: 10.1093/bioinformatics/bty476.
Mishra A, Malik R, Hachiya T, Jurgenson T, Namba S, Posner DC, et al. Stroke genetics informs
drug discovery and risk prediction across ancestries. Nature. 2022;611(7934):115-23. doi:
10.1038/s41586-022-05165-3.
Molero L, Garcia-Duran M, Diaz-Recasens J, Rico L, Casado S, Lopez-Farre A. Expression of
estrogen receptor subtypes and neuronal nitric oxide synthase in neutrophils from women
and men: regulation by estrogen. Cardiovasc Res. 2002;56(1):43-51. doi: 10.1016/s0008-
6363(02)00505-9.
Morales J, Welter D, Bowler EH, Cerezo M, Harris LW, McMahon AC, et al. A standardized
framework for representation of ancestry data in genomics studies, with application to the
NHGRI-EBI GWAS Catalog. Genome Biol. 2018;19(1):21. doi: 10.1186/s13059-018-
1396-2.
Morris TJ, Butcher LM, Feber A, Teschendorff AE, Chakravarthy AR, Wojdacz TK, et al.
ChAMP: 450k Chip Analysis Methylation Pipeline. Bioinformatics. 2014;30(3):428-30.
doi: 10.1093/bioinformatics/btt684.
Mosquera JV, Auguste G, Wong D, Turner AW, Hodonsky CJ, Alvarez-Yela AC, et al.
Integrative single-cell meta-analysis reveals disease-relevant vascular cell states and
markers in human atherosclerosis. Cell Rep. 2023;42(11):113380. doi:
10.1016/j.celrep.2023.113380.
Mozaffarian D, Katan MB, Ascherio A, Stampfer MJ, Willett WC. Trans fatty acids and
cardiovascular disease. N Engl J Med. 2006;354(15):1601-13. doi:
10.1056/NEJMra054035.
Murphy AJ, Guyre PM, Wira CR, Pioli PA. Estradiol regulates expression of estrogen receptor
ERalpha46 in human macrophages. PLoS One. 2009;4(5):e5539. doi:
10.1371/journal.pone.0005539.
Myers TA, Chanock SJ, Machiela MJ. LDlinkR: An R package for rapidly calculating linkage
disequilibrium statistics in diverse populations. Front Genet. 2020;11:157. doi:
10.3389/fgene.2020.00157.
Myocardial Infarction Genetics C, Kathiresan S, Voight BF, Purcell S, Musunuru K, Ardissino
D, et al. Genome-wide association of early-onset myocardial infarction with single
nucleotide polymorphisms and copy number variants. Nat Genet. 2009;41(3):334-41. doi:
10.1038/ng.327.
261
Nagai A, Hirata M, Kamatani Y, Muto K, Matsuda K, Kiyohara Y, et al. Overview of the
BioBank Japan Project: Study design and profile. J Epidemiol. 2017;27(3S):S2-S8. doi:
10.1016/j.je.2016.12.005.
Nappi F, Nappi P, Gambardella I, Avtaar Singh SS. Thromboembolic disease and cardiac
thrombotic complication in COVID-19: a systematic review. Metabolites. 2022;12(10).
doi: 10.3390/metabo12100889.
Natarajan P, Pampana A, Graham SE, Ruotsalainen SE, Perry JA, de Vries PS, et al.
Chromosome Xq23 is associated with lower atherogenic lipid concentrations and
favorable cardiometabolic indices. Nat Commun. 2021;12(1):2182. doi: 10.1038/s41467-
021-22339-1.
Nauffal V, Achanta A, Goldhaber SZ, Piazza G. Association of ABO blood group type with
cardiovascular events in COVID-19. J Thromb Thrombolysis. 2021;51(3):584-6. doi:
10.1007/s11239-020-02364-5.
Nelson CP, Goel A, Butterworth AS, Kanoni S, Webb TR, Marouli E, et al. Association analyses
based on false discovery rate implicate new loci for coronary artery disease. Nat Genet.
2017;49(9):1385-91. doi: 10.1038/ng.3913.
Nichols RV, O'Connell BL, Mulqueen RM, Thomas J, Woodfin AR, Acharya S, et al. Highthroughput robust single-cell DNA methylation profiling with sciMETv2. Nat Commun.
2022;13(1):7627. doi: 10.1038/s41467-022-35374-3.
Nikpay M, Goel A, Won HH, Hall LM, Willenborg C, Kanoni S, et al. A comprehensive 1,000
Genomes-based genome-wide association meta-analysis of coronary artery disease. Nat
Genet. 2015;47(10):1121-30. doi: 10.1038/ng.3396.
Nishiga M, Wang DW, Han Y, Lewis DB, Wu JC. COVID-19 and cardiovascular disease: from
basic mechanisms to clinical perspectives. Nat Rev Cardiol. 2020;17(9):543-58. doi:
10.1038/s41569-020-0413-9.
Norheim F, Hasin-Brumshtein Y, Vergnes L, Chella Krishnan K, Pan C, Seldin MM, et al. Geneby-Sex Interactions in Mitochondrial Functions and Cardio-Metabolic Traits. Cell Metab.
2019;29(4):932-49 e4. doi: 10.1016/j.cmet.2018.12.013.
O'Driscoll CA, Owens LA, Gallo ME, Hoffmann EJ, Afrazi A, Han M, et al. Differential effects
of diesel exhaust particles on T cell differentiation and autoimmune disease. Part Fibre
Toxicol. 2018;15(1):35. doi: 10.1186/s12989-018-0271-3.
Orozco LD, Bennett BJ, Farber CR, Ghazalpour A, Pan C, Che N, et al. Unraveling
inflammatory responses using systems genetics and gene-environment interactions in
macrophages. Cell. 2012;151(3):658-70. doi: 10.1016/j.cell.2012.08.043.
262
Pan H, Xue C, Auerbach BJ, Fan J, Bashore AC, Cui J, et al. Single-Cell Genomics Reveals a
Novel Cell State During Smooth Muscle Cell Phenotypic Switching and Potential
Therapeutic Targets for Atherosclerosis in Mouse and Human. Circulation.
2020;142(21):2060-75. doi: 10.1161/CIRCULATIONAHA.120.048378.
Panasevich S, Leander K, Ljungman P, Bellander T, de Faire U, Pershagen G, et al. Interaction
between air pollution exposure and genes in relation to levels of inflammatory markers
and risk of myocardial infarction. BMJ Open. 2013;3(9):e003058. doi: 10.1136/bmjopen2013-003058.
Paraboschi EM, Khera AV, Merlini PA, Gigante L, Peyvandi F, Chaffin M, et al. Rare variants
lowering the levels of coagulation factor X are protective against ischemic heart disease.
Haematologica. 2020;105(7):e365-e9. doi: 10.3324/haematol.2019.237750.
Parks BW, Nam E, Org E, Kostem E, Norheim F, Hui ST, et al. Genetic control of obesity and
gut microbiota composition in response to high-fat, high-sucrose diet in mice. Cell
Metab. 2013;17(1):141-52.
Parks BW, Sallam T, Mehrabian M, Psychogios N, Hui ST, Norheim F, et al. Genetic
architecture of insulin resistance in the mouse. Cell Metab. 2015;21(2):334-47. doi:
10.1016/j.cmet.2015.01.002.
Patel RS, Schmidt AF, Tragante V, McCubrey RO, Holmes MV, Howe LJ, et al. Association of
chromosome 9p21 with subsequent coronary heart disease events. Circ Genom Precis
Med. 2019;12(4):e002471. doi: 10.1161/CIRCGEN.119.002471.
Pazokitoroudi A, Liu Z, Dahl A, Zaitlen N, Rosset S, Sankararaman S. A scalable and robust
variance components method reveals insights into the architecture of gene-environment
interactions underlying complex traits. Am J Hum Genet. 2024;111(7):1462-80. doi:
10.1016/j.ajhg.2024.05.015.
Pearson GJ, Thanassoulis G, Anderson TJ, Barry AR, Couture P, Dayan N, et al. 2021 Canadian
Cardiovascular Society Guidelines for the Management of Dyslipidemia for the
Prevention of Cardiovascular Disease in Adults. Can J Cardiol. 2021;37(8):1129-50. doi:
10.1016/j.cjca.2021.03.016.
Pedicino D, Vinci R, Giglio AF, Pisano E, Porto I, Vergallo R, et al. Alterations of hyaluronan
metabolism in acute coronary syndrome: implications for plaque erosion. J Am Coll
Cardiol. 2018;72(13):1490-503. doi: 10.1016/j.jacc.2018.06.072.
Pereira E, Felipe S, de Freitas R, Araujo V, Soares P, Ribeiro J, et al. ABO blood group and link
to COVID-19: A comprehensive review of the reported associations and their possible
underlying mechanisms. Microb Pathog. 2022;169:105658. doi:
10.1016/j.micpath.2022.105658.
263
Perez AJ, Zeadally S. Recent Advances in Wearable Sensing Technologies. Sensors (Basel).
2021;21(20). doi: 10.3390/s21206828.
Peters A, Greven S, Heid IM, Baldari F, Breitner S, Bellander T, et al. Fibrinogen genes modify
the fibrinogen response to ambient particulate matter. Am J Respir Crit Care Med.
2009;179(6):484-91. doi: 10.1164/rccm.200805-751OC.
Petretta M, Costanzo P, Perrone-Filardi P, Chiariello M. Impact of gender in primary prevention
of coronary heart disease with statin therapy: a meta-analysis. Int J Cardiol.
2010;138(1):25-31. doi: 10.1016/j.ijcard.2008.08.001.
Pope CA, 3rd, Bhatnagar A, McCracken JP, Abplanalp W, Conklin DJ, O'Toole T. Exposure to
fine particulate air pollution is associated with endothelial injury and systemic
inflammation. Circ Res. 2016;119(11):1204-14. doi:
10.1161/CIRCRESAHA.116.309279.
Pope CA, 3rd, Turner MC, Burnett RT, Jerrett M, Gapstur SM, Diver WR, et al. Relationships
between fine particulate air pollution, cardiometabolic disorders, and cardiovascular
mortality. Circ Res. 2015;116(1):108-15. doi: 10.1161/CIRCRESAHA.116.305060.
Portilla-Fernandez E, Hwang SJ, Wilson R, Maddock J, Hill WD, Teumer A, et al. Meta-analysis
of epigenome-wide association studies of carotid intima-media thickness. Eur J
Epidemiol. 2021;36(11):1143-55. doi: 10.1007/s10654-021-00759-z.
Post WS, Goldschmidt-Clermont PJ, Wilhide CC, Heldman AW, Sussman MS, Ouyang P, et al.
Methylation of the estrogen receptor gene is associated with aging and atherosclerosis in
the cardiovascular system. Cardiovasc Res. 1999;43(4):985-91. doi: 10.1016/s0008-
6363(99)00153-4.
Pruim RJ, Welch RP, Sanna S, Teslovich TM, Chines PS, Gliedt TP, et al. LocusZoom: regional
visualization of genome-wide association scan results. Bioinformatics.
2010;26(18):2336-7. doi: 10.1093/bioinformatics/btq419.
Rajagopalan S, Brauer M, Bhatnagar A, Bhatt DL, Brook JR, Huang W, et al. Personal-Level
Protective Actions Against Particulate Matter Air Pollution Exposure: A Scientific
Statement From the American Heart Association. Circulation. 2020;142(23):e411-e31.
doi: 10.1161/CIR.0000000000000931.
Reilly MP, Li M, He J, Ferguson JF, Stylianou IM, Mehta NN, et al. Identification of
ADAMTS7 as a novel locus for coronary atherosclerosis and association of ABO with
myocardial infarction in the presence of coronary atherosclerosis: two genome-wide
association studies. Lancet. 2011;377(9763):383-92.
Reilly MP, Li M, He J, Ferguson JF, Stylianou IM, Mehta NN, et al. Identification of
ADAMTS7 as a novel locus for coronary atherosclerosis and association of ABO with
myocardial infarction in the presence of coronary atherosclerosis: two genome-wide
264
association studies. Lancet. 2011;377(9763):383-92. doi: 10.1016/S0140-6736(10)61996-
4.
Reilly MP, Li M, He J, Ferguson JF, Stylianou IM, Mehta NN, et al. Identification of
ADAMTS7 as a novel locus for coronary atherosclerosis and association of ABO with
myocardial infarction in the presence of coronary atherosclerosis: two genome-wide
association studies. Lancet. 2011;377(9763):383-92.
Reiner AP, Hartiala J, Zeller T, Bis JC, Dupuis J, Fornage M, et al. Genome-wide and genecentric analyses of circulating myeloperoxidase levels in the CHARGE and CARe
consortia. Hum Mol Genet. 2013;22(16):3381-93.
Ren C, Baccarelli A, Wilker E, Suh H, Sparrow D, Vokonas P, et al. Lipid and endotheliumrelated genes, ambient particulate matter, and heart rate variability--the VA Normative
Aging Study. J Epidemiol Community Health. 2010;64(1):49-56. doi:
10.1136/jech.2008.083295.
Ridker PM, Cook NR, Lee IM, Gordon D, Gaziano JM, Manson JE, et al. A randomized trial of
low-dose aspirin in the primary prevention of cardiovascular disease in women. N Engl J
Med. 2005;352(13):1293-304. doi: 10.1056/NEJMoa050613.
Ridker PM, Danielson E, Fonseca FA, Genest J, Gotto AM, Jr., Kastelein JJ, et al. Rosuvastatin
to prevent vascular events in men and women with elevated C-reactive protein. N Engl J
Med. 2008;359(21):2195-207. doi: 10.1056/NEJMoa0807646.
Rinaldi S, Geay A, Dechaud H, Biessy C, Zeleniuch-Jacquotte A, Akhmedkhanov A, et al.
Validity of free testosterone and free estradiol determinations in serum samples from
postmenopausal women by theoretical calculations. Cancer Epidemiol Biomarkers Prev.
2002;11(10 Pt 1):1065-71.
Risom L, Moller P, Loft S. Oxidative stress-induced DNA damage by particulate air pollution.
Mutat Res. 2005;592(1-2):119-37. doi: 10.1016/j.mrfmmm.2005.06.012.
Ritchie ME, Phipson B, Wu D, Hu Y, Law CW, Shi W, et al. limma powers differential
expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res.
2015;43(7):e47. doi: 10.1093/nar/gkv007.
Rodriguez-Aparicio LB, Guadalajara AM, Rubio V. Physical location of the site for N-acetyl-Lglutamate, the allosteric activator of carbamoyl phosphate synthetase, in the 20-kilodalton
COOH-terminal domain. Biochemistry (Mosc). 1989;28(7):3070-4.
Romieu I, Moreno-Macias H, London SJ. Gene by environment interaction and ambient air
pollution. Proc Am Thorac Soc. 2010;7(2):116-22. doi: 10.1513/pats.200909-097RM.
Rose NR. Viral myocarditis. Curr Opin Rheumatol. 2016;28(4):383-9. doi:
10.1097/BOR.0000000000000303.
265
Ruckerl R, Phipps RP, Schneider A, Frampton M, Cyrys J, Oberdorster G, et al. Ultrafine
particles and platelet activation in patients with coronary heart disease--results from a
prospective panel study. Part Fibre Toxicol. 2007;4:1. doi: 10.1186/1743-8977-4-1.
Sabatine MS, Giugliano RP, Keech AC, Honarpour N, Wiviott SD, Murphy SA, et al.
Evolocumab and Clinical Outcomes in Patients with Cardiovascular Disease. N Engl J
Med. 2017;376(18):1713-22. doi: 10.1056/NEJMoa1615664.
Sackton TB, Hartl DL. Genotypic context and epistasis in individuals and populations. Cell.
2016;166(2):279-87. doi: 10.1016/j.cell.2016.06.047.
Saleheen D, Zhao W, Young R, Nelson CP, Ho W, Ferguson JF, et al. Cardioprotective effects at
the ADAMTS7 locus as a result of gene-smoking interactions. Circulation.
2017;135(24):2336-53. doi: 10.1161/CIRCULATIONAHA.116.022069.
Salpeter SR, Walsh JM, Greyber E, Ormiston TM, Salpeter EE. Mortality associated with
hormone replacement therapy in younger and older women: a meta-analysis. J Gen Intern
Med. 2004;19(7):791-804. doi: 10.1111/j.1525-1497.2004.30281.x.
Salpeter SR, Walsh JM, Greyber E, Salpeter EE. Brief report: Coronary heart disease events
associated with hormone therapy in younger and older women. A meta-analysis. J Gen
Intern Med. 2006;21(4):363-6. doi: 10.1111/j.1525-1497.2006.00389.x.
Samani NJ, Erdmann J, Hall AS, Hengstenberg C, Mangino M, Mayer B, et al. Genomewide
association analysis of coronary artery disease. N Engl J Med. 2007;357(5):443-53.
Sanderson E. Multivariable Mendelian Randomization and Mediation. Cold Spring Harb
Perspect Med. 2021;11(2). doi: 10.1101/cshperspect.a038984.
Saxon A, Diaz-Sanchez D. Air pollution and allergy: you are what you breathe. Nat Immunol.
2005;6(3):223-6. doi: 10.1038/ni0305-223.
Schane RE, Ling PM, Glantz SA. Health effects of light and intermittent smoking: a review.
Circulation. 2010;121(13):1518-22. doi: 10.1161/CIRCULATIONAHA.109.904235.
Schisler JC, Ronnebaum SM, Madden M, Channell M, Campen M, Willis MS. Endothelial
inflammatory transcriptional responses to an altered plasma exposome following
inhalation of diesel emissions. Inhal Toxicol. 2015;27(5):272-80. doi:
10.3109/08958378.2015.1030481.
Schofield JP, Cox TM, Caskey CT, Wakamiya M. Mice deficient in the urea-cycle enzyme,
carbamoyl phosphate synthetase I, die during the early neonatal period from
hyperammonemia. Hepatology. 1999;29(1):181-5. doi: 10.1002/hep.510290112.
266
Schwarz UI, Ritchie MD, Bradford Y, Li C, Dudek SM, Frye-Anderson A, et al. Genetic
determinants of response to warfarin during initial anticoagulation. N Engl J Med.
2008;358(10):999-1008. doi: 10.1056/NEJMoa0708078.
Sekhon JS. Multivariate and Propensity Score Matching Software with Automated Balance
Optimization: The Matching package for R. Journal of Statistical Software. 2011;42(7):1
- 52. doi: 10.18637/jss.v042.i07.
Severe Covid GG, Ellinghaus D, Degenhardt F, Bujanda L, Buti M, Albillos A, et al.
Genomewide association study of severe Covid-19 with respiratory failure. N Engl J
Med. 2020;383(16):1522-34. doi: 10.1056/NEJMoa2020283.
Shaw LJ, Bugiardini R, Merz CN. Women and ischemic heart disease: evolving knowledge. J
Am Coll Cardiol. 2009;54(17):1561-75. doi: 10.1016/j.jacc.2009.04.098.
Shears RK, Jacques LC, Naylor G, Miyashita L, Khandaker S, Lebre F, et al. Exposure to diesel
exhaust particles increases susceptibility to invasive pneumococcal disease. J Allergy
Clin Immunol. 2020;145(4):1272-84 e6. doi: 10.1016/j.jaci.2019.11.039.
Shekelle PG, Rich MW, Morton SC, Atkinson CS, Tu W, Maglione M, et al. Efficacy of
angiotensin-converting enzyme inhibitors and beta-blockers in the management of left
ventricular systolic dysfunction according to race, gender, and diabetic status: a metaanalysis of major clinical trials. J Am Coll Cardiol. 2003;41(9):1529-38. doi:
10.1016/s0735-1097(03)00262-6.
Shi J, Luo Y. Effects of RHD gene polymorphisms on distinguishing weak D or DEL from RhDin blood donation in a Chinese population. Mol Genet Genomic Med. 2019;7(6):e00681.
doi: 10.1002/mgg3.681.
Shkirkova K, Demetriou AN, Sizdahkhani S, Lamorie-Foote K, Zhang H, Morales M, et al.
Microglial TLR4 mediates white matter injury in a combined model of diesel exhaust
exposure and cerebral hypoperfusion. Stroke. 2024;55(4):1090-3. doi:
10.1161/STROKEAHA.124.046412.
Shkirkova K, Demetriou AN, Sizdahkhani S, Lamorie-Foote K, Zhang H, Morales M, et al.
Microglial TLR4 mediates white matter injury in a combined model of diesel exhaust
exposure and cerebral hypoperfusion. Stroke. 2024;55(4):1090-3. doi:
10.1161/STROKEAHA.124.046412.
Shkirkova K, Lamorie-Foote K, Zhang N, Li A, Diaz A, Liu Q, et al. Neurotoxicity of diesel
exhaust particles. J Alzheimers Dis. 2022;89(4):1263-78. doi: 10.3233/JAD-220493.
Shukla A, Timblin C, BeruBe K, Gordon T, McKinney W, Driscoll K, et al. Inhaled particulate
matter causes expression of nuclear factor (NF)-kappaB-related genes and oxidantdependent NF-kappaB activation in vitro. Am J Respir Cell Mol Biol. 2000;23(2):182-7.
doi: 10.1165/ajrcmb.23.2.4035.
267
Siemelink MA, van der Laan SW, Haitjema S, van Koeverden ID, Schaap J, Wesseling M, et al.
Smoking is associated to dna methylation in atherosclerotic carotid lesions. Circ Genom
Precis Med. 2018;11(9):e002030. doi: 10.1161/CIRCGEN.117.002030.
Singh P, DeMarini DM, Dick CA, Tabor DG, Ryan JV, Linak WP, et al. Sample characterization
of automobile and forklift diesel exhaust particles and comparative pulmonary toxicity in
mice. Environ Health Perspect. 2004;112(8):820-5. doi: 10.1289/ehp.6579.
Smeeth L, Cook C, Thomas S, Hall AJ, Hubbard R, Vallance P. Risk of deep vein thrombosis
and pulmonary embolism after acute infection in a community setting. Lancet.
2006;367(9516):1075-9. doi: 10.1016/S0140-6736(06)68474-2.
Smeeth L, Thomas SL, Hall AJ, Hubbard R, Farrington P, Vallance P. Risk of myocardial
infarction and stroke after acute infection or vaccination. N Engl J Med.
2004;351(25):2611-8. doi: 10.1056/NEJMoa041747.
Smith ML, Baggerly KA, Bengtsson H, Ritchie ME, Hansen KD. illuminaio: An open source
IDAT parsing tool for Illumina microarrays. F1000Res. 2013;2:264. doi:
10.12688/f1000research.2-264.v1.
Smith PG, Day NE. The design of case-control studies: the influence of confounding and
interaction effects. Int J Epidemiol. 1984;13(3):356-65. doi: 10.1093/ije/13.3.356.
Sokolski M, Reszka K, Adamik B, Kilis-Pstrusinska K, Lis W, Pomorski M, et al. Antiplatelet
therapy prior to COVID-19 infection impacts on patients mortality: a propensity scorematched cohort study. Sci Rep. 2024;14(1):4832. doi: 10.1038/s41598-024-55407-9.
Sollis E, Mosaku A, Abid A, Buniello A, Cerezo M, Gil L, et al. The NHGRI-EBI GWAS
Catalog: knowledgebase and deposition resource. Nucleic Acids Res. 2023;51(D1):D977-
D85. doi: 10.1093/nar/gkac1010.
Solmaz I, Arac S. ABO blood groups in COVID-19 patients; Cross-sectional study. Int J Clin
Pract. 2021;75(4):e13927. doi: 10.1111/ijcp.13927.
Spracklen CN, Horikoshi M, Kim YJ, Lin K, Bragg F, Moon S, et al. Identification of type 2
diabetes loci in 433,540 East Asian individuals. Nature. 2020;582(7811):240-5. doi:
10.1038/s41586-020-2263-3.
Sriprasert I, Hodis HN, Karim R, Stanczyk FZ, Shoupe D, Henderson VW, et al. Differential
Effect of Plasma Estradiol on Subclinical Atherosclerosis Progression in Early vs Late
Postmenopause. J Clin Endocrinol Metab. 2019;104(2):293-300. doi: 10.1210/jc.2018-
01600.
Staley JR, Suderman M, Simpkin AJ, Gaunt TR, Heron J, Relton CL, et al. Longitudinal analysis
strategies for modelling epigenetic trajectories. Int J Epidemiol. 2018;47(2):516-25. doi:
10.1093/ije/dyy012.
268
Steenhof M, Janssen NA, Strak M, Hoek G, Gosens I, Mudway IS, et al. Air pollution exposure
affects circulating white blood cell counts in healthy subjects: the role of particle
composition, oxidative potential and gaseous pollutants - the RAPTES project. Inhal
Toxicol. 2014;26(3):141-65. doi: 10.3109/08958378.2013.861884.
Sun Q, Wang A, Jin X, Natanzon A, Duquaine D, Brook RD, et al. Long-term air pollution
exposure and acceleration of atherosclerosis and vascular inflammation in an animal
model. JAMA. 2005;294(23):3003-10. doi: 10.1001/jama.294.23.3003.
Sun Q, Yue P, Deiuliis JA, Lumeng CN, Kampfrath T, Mikolaj MB, et al. Ambient air pollution
exaggerates adipose inflammation and insulin resistance in a mouse model of dietinduced obesity. Circulation. 2009;119(4):538-46.
Sun Q, Yue P, Kirk RI, Wang A, Moatti D, Jin X, et al. Ambient air particulate matter exposure
and tissue factor expression in atherosclerosis. Inhal Toxicol. 2008;20(2):127-37. doi:
10.1080/08958370701821482.
Tabassum R, Ramo JT, Ripatti P, Koskela JT, Kurki M, Karjalainen J, et al. Genetic architecture
of human plasma lipidome and its link to cardiovascular disease. Nat Commun.
2019;10(1):4329. doi: 10.1038/s41467-019-11954-8.
Tan JWC, Chew DP, Brieger D, Eikelboom J, Montalescot G, Ako J, et al. 2020 Asian pacific
society of cardiology consensus recommendations on antithrombotic management for
high-risk chronic coronary syndrome. Eur Cardiol. 2021;16:e26. doi:
10.15420/ecr.2020.45.
Tcheandjieu C, Zhu X, Hilliard AT, Clarke SL, Napolioni V, Ma S, et al. Large-scale genomewide association study of coronary artery disease in genetically diverse populations. Nat
Med. 2022;28(8):1679-92. doi: 10.1038/s41591-022-01891-3.
Timmers PRHJ, Wilson JF. Limited Effect of Y Chromosome Variation on Coronary Artery
Disease and Mortality in UK Biobank-Brief Report. Arterioscler Thromb Vasc Biol.
2022;42(9):1198-206. doi: 10.1161/ATVBAHA.122.317664.
Topf A, Pyle A, Griffin H, Matalonga L, Schon K, Solve RDSNViwg, et al. Exome reanalysis
and proteomic profiling identified TRIP4 as a novel cause of cerebellar hypoplasia and
spinal muscular atrophy (PCH1). Eur J Hum Genet. 2021;29(9):1348-53. doi:
10.1038/s41431-021-00851-8.
Triche TJ, Jr., Weisenberger DJ, Van Den Berg D, Laird PW, Siegmund KD. Low-level
processing of Illumina Infinium DNA Methylation BeadArrays. Nucleic Acids Res.
2013;41(7):e90. doi: 10.1093/nar/gkt090.
van der Harst P, Verweij N. Identification of 64 novel genetic loci provides an expanded view on
the genetic architecture of coronary artery disease. Circ Res. 2018;122(3):433-43. doi:
10.1161/CIRCRESAHA.117.312086.
269
Varga Z, Flammer AJ, Steiger P, Haberecker M, Andermatt R, Zinkernagel AS, et al.
Endothelial cell infection and endotheliitis in COVID-19. Lancet.
2020;395(10234):1417-8. doi: 10.1016/S0140-6736(20)30937-5.
Vasan SK, Rostgaard K, Majeed A, Ullum H, Titlestad KE, Pedersen OB, et al. ABO blood
group and risk of thromboembolic and arterial disease: A study of 1.5 million blood
donors. Circulation. 2016;133(15):1449-57. doi:
10.1161/CIRCULATIONAHA.115.017563.
Velavan TP, Pallerla SR, Ruter J, Augustin Y, Kremsner PG, Krishna S, et al. Host genetic
factors determining COVID-19 susceptibility and severity. EBioMedicine.
2021;72:103629. doi: 10.1016/j.ebiom.2021.103629.
Vikman S, Brena RM, Armstrong P, Hartiala J, Stephensen CB, Allayee H. Functional analysis
of 5-Lipoxygenase promoter repeat variants. Hum Mol Genet. 2009;18(23):4521-9.
Vilas-Boas V, Chatterjee N, Carvalho A, Alfaro-Moreno E. Particulate matter-induced oxidative
stress - mechanistic insights and antioxidant approaches reported in in vitro studies.
Environ Toxicol Pharmacol. 2024;110:104529. doi: 10.1016/j.etap.2024.104529.
Vilas-Boas V, Chatterjee N, Carvalho A, Alfaro-Moreno E. Particulate matter-induced oxidative
stress - mechanistic insights and antioxidant approaches reported in in vitro studies.
Environ Toxicol Pharmacol. 2024;110:104529. doi: 10.1016/j.etap.2024.104529.
Vinkhuyzen AA, Wray NR, Yang J, Goddard ME, Visscher PM. Estimation and partition of
heritability in human populations using whole-genome analysis methods. Annu Rev
Genet. 2013;47:75-95. doi: 10.1146/annurev-genet-111212-133258.
Virani SS, Alonso A, Benjamin EJ, Bittencourt MS, Callaway CW, Carson AP, et al. Heart
Disease and Stroke Statistics-2020 Update: A Report From the American Heart
Association. Circulation. 2020;141(9):e139-e596. doi: 10.1161/CIR.0000000000000757.
Visseren FLJ, Mach F, Smulders YM, Carballo D, Koskinas KC, Back M, et al. 2021 ESC
Guidelines on cardiovascular disease prevention in clinical practice. Eur Heart J.
2021;42(34):3227-337. doi: 10.1093/eurheartj/ehab484.
Vosa U, Claringbould A, Westra HJ, Bonder MJ, Deelen P, Zeng B, et al. Large-scale cis- and
trans-eQTL analyses identify thousands of genetic loci and polygenic scores that regulate
blood gene expression. Nat Genet. 2021;53(9):1300-10. doi: 10.1038/s41588-021-00913-
z.
Walsh JM, Pignone M. Drug treatment of hyperlipidemia in women. JAMA. 2004;291(18):2243-
52. doi: 10.1001/jama.291.18.2243.
270
Wang H, Zhang F, Zeng J, Wu Y, Kemper KE, Xue A, et al. Genotype-by-environment
interactions inferred from genetic effects on phenotypic variability in the UK Biobank.
Sci Adv. 2019;5(8):eaaw3538. doi: 10.1126/sciadv.aaw3538.
Wang Z, Klipfell E, Bennett BJ, Koeth R, Levison BS, Dugar B, et al. Gut flora metabolism of
phosphatidylcholine promotes cardiovascular disease. Nature. 2011;472(7341):57-63.
Wang Z, Levison BS, Hazen JE, Donahue L, Li XM, Hazen SL. Measurement of
trimethylamine-N-oxide by stable isotope dilution liquid chromatography tandem mass
spectrometry. Anal Biochem. 2014;455:35-40. doi: 10.1016/j.ab.2014.03.016.
Ward-Caviness CK, Neas LM, Blach C, Haynes CS, LaRocque-Abramson K, Grass E, et al. A
genome-wide trans-ethnic interaction study links the PIGR-FCAMR locus to coronary
atherosclerosis via interactions between genetic variants and residential exposure to
traffic. PLoS One. 2017;12(3):e0173880. doi: 10.1371/journal.pone.0173880.
Ward-Caviness CK, Neas LM, Blach C, Haynes CS, LaRocque-Abramson K, Grass E, et al.
Genetic variants in the bone morphogenic protein gene family modify the association
between residential exposure to traffic and peripheral arterial disease. PLoS One.
2016;11(4):e0152670. doi: 10.1371/journal.pone.0152670.
Ward-Caviness CK. A review of gene-by-air pollution interactions for cardiovascular disease,
risk factors, and biomarkers. Hum Genet. 2019;138(6):547-61. doi: 10.1007/s00439-019-
02004-w.
Warren CE, Campbell KM, Kirkham MN, Saito ER, Remund NP, Cayabyab KB, et al. The
effect of diesel exhaust particles on adipose tissue mitochondrial function and
inflammatory status. Int J Mol Sci. 2024;25(8). doi: 10.3390/ijms25084322.
Wenger NK. Women and coronary heart disease: a century after Herrick: understudied,
underdiagnosed, and undertreated. Circulation. 2012;126(5):604-11. doi:
10.1161/CIRCULATIONAHA.111.086892.
Wienke A, Holm NV, Skytthe A, Yashin AI. The heritability of mortality due to heart diseases: a
correlated frailty model applied to Danish twins. Twin Res. 2001;4(4):266-74. doi:
10.1375/1369052012399.
Willer CJ, Li Y, Abecasis GR. METAL: fast and efficient meta-analysis of genomewide
association scans. Bioinformatics. 2010;26(17):2190-1.
Wing LM, Reid CM, Ryan P, Beilin LJ, Brown MA, Jennings GL, et al. A comparison of
outcomes with angiotensin-converting--enzyme inhibitors and diuretics for hypertension
in the elderly. N Engl J Med. 2003;348(7):583-92. doi: 10.1056/NEJMoa021716.
271
Wirka RC, Wagh D, Paik DT, Pjanic M, Nguyen T, Miller CL, et al. Atheroprotective roles of
smooth muscle cell phenotypic modulation and the TCF21 disease gene as revealed by
single-cell analysis. Nat Med. 2019;25(8):1280-9. doi: 10.1038/s41591-019-0512-5.
Woodward NC, Crow AL, Zhang Y, Epstein S, Hartiala J, Johnson R, et al. Exposure to
nanoscale particulate matter from gestation to adulthood impairs metabolic homeostasis
in mice. Sci Rep. 2019;9(1):1816. doi: 10.1038/s41598-018-37704-2.
Worth RM, Kato H, Rhoads GG, Kagan K, Syme SL. Epidemiologic studies of coronary heart
disease and stroke in Japanese men living in Japan, Hawaii and California: mortality. Am
J Epidemiol. 1975;102(6):481-90. doi: 10.1093/oxfordjournals.aje.a112186.
Xie Y, Xu E, Bowe B, Al-Aly Z. Long-term cardiovascular outcomes of COVID-19. Nat Med.
2022;28(3):583-90. doi: 10.1038/s41591-022-01689-3.
Xu E, Xie Y, Al-Aly Z. Long-term neurologic outcomes of COVID-19. Nat Med.
2022;28:pages2406–15. doi: 10.1038/s41591-022-02001-z.
Yamada Y, Horibe H, Oguri M, Sakuma J, Takeuchi I, Yasukochi Y, et al. Identification of
novel hyper- or hypomethylated CpG sites and genes associated with atherosclerotic
plaque using an epigenome-wide association study. Int J Mol Med. 2018;41(5):2724-32.
doi: 10.3892/ijmm.2018.3453.
Yamada Y, Horibe H, Oguri M, Sakuma J, Takeuchi I, Yasukochi Y, et al. Identification of
novel hyper- or hypomethylated CpG sites and genes associated with atherosclerotic
plaque using an epigenome-wide association study. Int J Mol Med. 2018;41(5):2724-32.
doi: 10.3892/ijmm.2018.3453.
Ying AK, Hassanain HH, Roos CM, Smiraglia DJ, Issa JJ, Michler RE, et al. Methylation of the
estrogen receptor-alpha gene promoter is selectively increased in proliferating human
aortic smooth muscle cells. Cardiovasc Res. 2000;46(1):172-9. doi: 10.1016/s0008-
6363(00)00004-3.
Zaina S, Heyn H, Carmona FJ, Varol N, Sayols S, Condom E, et al. DNA methylation map of
human atherosclerosis. Circ Cardiovasc Genet. 2014;7(5):692-700. doi:
10.1161/CIRCGENETICS.113.000441.
Zanobetti A, Baccarelli A, Schwartz J. Gene-air pollution interaction and cardiovascular disease:
a review. Prog Cardiovasc Dis. 2011;53(5):344-52. doi: 10.1016/j.pcad.2011.01.001.
Zdravkovic S, Wienke A, Pedersen NL, Marenberg ME, Yashin AI, De Faire U. Heritability of
death from coronary heart disease: a 36-year follow-up of 20 966 Swedish twins. J Intern
Med. 2002;252(3):247-54. doi: 10.1046/j.1365-2796.2002.01029.x.
Zeberg H, Paabo S. The major genetic risk factor for severe COVID-19 is inherited from
Neanderthals. Nature. 2020;587(7835):610-2. doi: 10.1038/s41586-020-2818-3.
272
Zhang H, D'Agostino C, Forman HJ, Cacciottolo M, Thorwald M, Mack WJ, et al. Urban air
pollution nanoparticles from Los Angeles: recently decreased neurotoxicity. J Alzheimers
Dis. 2021;82(1):307-16. doi: 10.3233/JAD-201577.
Zheng YY, Ma YT, Zhang JY, Xie X. COVID-19 and the cardiovascular system. Nat Rev
Cardiol. 2020;17(5):259-60. doi: 10.1038/s41569-020-0360-5.
Zhou W, Laird PW, Shen H. Comprehensive characterization, annotation and innovative use of
Infinium DNA methylation BeadChip probes. Nucleic Acids Res. 2017;45(4):e22. doi:
10.1093/nar/gkw967.
Zietz M, Zucker J, Tatonetti NP. Associations between blood type and COVID-19 infection,
intubation, and death. Nat Commun. 2020;11(1):5761. doi: 10.1038/s41467-020-19623-
x.
Zong X, Wang X, Liu Y, Li Z, Wang W, Wei D, et al. Antiplatelet therapy for patients with
COVID-19: Systematic review and meta-analysis of observational studies and
randomized controlled trials. Frontiers in medicine. 2022;9:965790. doi:
10.3389/fmed.2022.965790.
273
APPENDICES
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
6
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
Abstract (if available)
Abstract
Coronary artery disease (CAD) is a multifactorial disease characterized by the accumulation of atherosclerotic plaques (composed of oxidized lipids, fibrous tissue, and inflammatory cells), within arterial walls. This accumulation narrows the coronary arteries, impeding blood flow to the heart and potentially leading to plaque rupture, arterial occlusion, coronary ischemia, and myocardial infarction. Extensive epidemiological research has implicated various biological and environmental risk factors in development of CAD while genetic studies have identified heritable susceptibility factors as additional risk contributors. Despite the recognized interplay between genetic predisposition and environmental exposures in the etiology of CAD, research exploring their interactions remains limited. Therefore, this dissertation focuses on employing a systems genetics approach to uncover novel pathways involved in CAD pathogenesis through interactions between genetic factors and various biological and environmental exposures.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Genetic studies of inflammation and cardiovascular disease
PDF
The contribution of arachidonate 5-lipoxygenase to inflammatory response in coronary artery disease
PDF
The influence of DNA repair genes and prenatal tobacco exposure on childhood acute lymphoblastic leukemia risk: a gene-environment interaction study
PDF
Study of genetic variation in Asian Indians -- applications to complex diseases and endogamic exogamy
PDF
Identification of SLC44A3 as a novel susceptibility gene for myocardial infarction
PDF
Gene expression and genetic variation of ERG is associated with inflammation in endothelial cells and risk of coronary artery disease in humans
PDF
Neuroimaging in complex polygenic disorders
PDF
Genetic and dietary determinants of nonalcoholic fatty liver disease in Hispanic children
PDF
Towards identification of proteins interacting with wild-type or mutant PMP22 protein
PDF
Improving the sensitivity and spatial coverage of cardiac arterial spin labeling for assessment of coronary artery disease
PDF
An overview of Alzheimer’s disease and the potential for APP-beta secretase-interaction as a therapeutic target
PDF
Dietary fats, atherosclerosis, and cardiovascular disease
PDF
Genetic and environmental risk factors for childhood cancer
PDF
Bayesian model averaging methods for gene-environment interactions and admixture mapping
PDF
Cardiovascular disease risk factors and cognitive function
PDF
Identification and characterization of cancer-associated enhancers
PDF
Application of genetic association methods in mice to understand phenotypes with a complex etiology
PDF
Multivariate methods for extracting genetic associations from correlated data
PDF
Prostate cancer: genetic susceptibility and lifestyle risk factors
PDF
Expression pattern analysis of coronary and lymphatic vessel genes in zebrafish and mice
Asset Metadata
Creator
Hilser, James Raymond
(author)
Core Title
Identification of gene-exposure interactions for risk of cardiovascular disease
School
Keck School of Medicine
Degree
Doctor of Philosophy
Degree Program
Molecular Medicine
Degree Conferral Date
2024-12
Publication Date
12/16/2024
Defense Date
12/09/2024
Publisher
Los Angeles, California
(original),
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
coronary artery disease,gene-exposure interactions,genetics,OAI-PMH Harvest
Format
theses
(aat)
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Patel, Pragna (
committee chair
), Allayee, Hooman (
committee member
), Hartiala, Jaana (
committee member
)
Creator Email
hilser@usc.edu,jamesrhfinancial@gmail.com
Unique identifier
UC11399EVNA
Identifier
etd-HilserJame-13696.pdf (filename)
Legacy Identifier
etd-HilserJame-13696
Document Type
Dissertation
Format
theses (aat)
Rights
Hilser, James Raymond
Internet Media Type
application/pdf
Type
texts
Source
20241217-usctheses-batch-1229
(batch),
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright.
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA
Repository Email
cisadmin@lib.usc.edu
Tags
coronary artery disease
gene-exposure interactions
genetics