25-hydroxyvitamin D and Biomarkers of … · ii 25-hydroxyvitamin D and Biomarkers of Cardiometabolic Disease Bibiana García Bailo Doctor of Philosophy Department of Nutritional

25-hydroxyvitamin D and Biomarkers of Cardiometabolic

Disease

by

Bibiana García Bailo

A thesis submitted in conformity with the requirements

for the degree of Doctor of Philosophy

Department of Nutritional Sciences

University of Toronto

© Copyright by Bibiana García Bailo, 2013

ii

25-hydroxyvitamin D and Biomarkers of Cardiometabolic Disease

Bibiana García Bailo

Doctor of Philosophy

Department of Nutritional Sciences

University of Toronto

2013

Abstract

Background: Vitamin D may have beneficial effects on cardiometabolic disease, but the

evidence is equivocal. This may be due to unaccounted confounders, such as lifestyle factors and

genetic variation. We examined the association between circulating 25-hydroxyvitamin D

[25(OH)D] and biomarkers of cardiometabolic disease risk, including biomarkers of

inflammation, glycemic dysregulation and lipid metabolism, and a panel of 54 plasma proteomic

biomarkers, and determined whether lifestyle variables and genetic variation modified these

associations.

Methods: Participants were from the Toronto Nutrigenomics and Health Study, an ethnically

diverse population of individuals aged 20-29 years. Anthropometric measurements were

obtained. Participants answered general health and lifestyle and food frequency questionnaires

and provided a fasting blood sample for biochemical measurements and genotyping.

Results: Across ethnic groups, women who used hormonal contraceptives (HC) had higher

25(OH)D and C-reactive protein (CRP) than women HC non-users and men. Circulating

25(OH)D was positively associated with CRP in the entire population in models not accounting

for HC use. However, there was no association after accounting for HC use. 25(OH)D was also

not associated with inflammatory cytokines after adjusting for HC use. 25(OH)D was inversely

iii

associated with insulin, HOMA-IR, and HOMA-Beta among Caucasians and East Asians and

among men and women HC non-users. No biomarkers were associated with 25(OH)D among

South Asians and women HC users, although non-significant inverse trends were observed for

markers of glycemic dysregulation. Only two of the 54 plasma proteomic biomarkers were

associated with 25(OH)D in women HC non-users, and none were associated in men. Among

women HC users, after accounting for hormone dose, only three proteins were associated with

25(OH)D. Finally, 25(OH)D affected the association between rs2239182, a variant in the vitamin

D receptor (VDR) and the pro-inflammatory cytokine interferon γ-induced protein 10 (IP-10).

However, the association was suggestive of heterosis and may have been due to chance.

Conclusions: We identified a confounding effect of HC use on the association between

25(OH)D, biomarkers of inflammation and plasma proteomic biomarkers. In addition, HC use

might also affect the association between 25(OH)D and biomarkers of glycemic dysregulation.

Genetic variation in VDR did not modify any associations.

.

iv

Acknowledgments

In my head, I have been writing this section of my thesis for years. In many ways, the degree feels

like a group effort and throughout the journey I have been grateful to my mentors, peers, friends and

family more times than I can count. For having reached the final stage of my PhD with my sanity and

my love of academic pursuits intact, I am forever indebted to Dr. Ahmed El-Sohemy, my thesis

supervisor. Ahmed's unwavering support, optimism and sense of humour kept me going through

good and bad times. His approachability, patience and open-mindedness made me look forward to all

of our discussions, and he broadened my intellectual horizons, provided me with training and career

opportunities that I never would have been able to pursue on my own, and taught me how to

persevere and have faith in my own scientific abilities. Working with Ahmed was a true privilege,

and I could not have wished for a better supervisor.

I am very grateful to the members of my advisory committee, Drs. David Cole, Mohamed Karmali

and Reinhold Vieth, and my PhD examiners, Drs. Tom Wolever and Marie-Claude Vohl. They

shared many ideas, asked important questions and provided thoughtful, constructive criticism. The

quality of my thesis improved significantly thanks to their input. I am equally grateful to the faculty

of the Department of Nutritional Sciences, and in particular Drs. Tony Hanley and Val Tarasuk, who

taught me a great deal about epidemiology and public health nutrition. Their passion and knowledge

are awe-inspiring, and their doors are always open to students. Many thanks also to the great staff of

the department, and especially Louisa Kung and Emeliana D'Souza, who were always so helpful and

such a joy to interact with.

I am thankful to the funding sources that made my work possible, including the Ontario Graduate

Scholarship program, the Public Health Agency of Canada and the Advanced Foods and Materials

Network. In addition, I was fortunate to be able to attend several conferences thanks to funding from

the University of Toronto's Department of Nutritional Sciences, Banting & Best Diabetes Centre, and

School of Graduate Studies, as well as the American Society for Nutrition.

I am also quite grateful to our collaborators from the University of Victoria/Genome BC Proteomics

Centre, the University of Guelph, and, most importantly, the Office of Biotechnology, Genomics and

Population Health of the Public Health Agency of Canada. The BGPH team, including Drs.

Mohamed Karmali, Alaa Badawi, Suneil Malik and John Nash, as well as Ross Duncan, Rachel

Rasile, Andre Villegas, Paul Arora, Chrissy Wessman, Rowena Mulato and Lindsay Nolan, provided

v

a second home for me, wrote many reference letters, and gave me an invaluable opportunity to

conduct epidemiologic research in a public health setting. It was an honour and a pleasure to be a part

of their group.

My interest in nutrigenomics arose during the time I spent as a technician in Dr. José Ordovás's lab,

at Tufts University in Boston, Massachusetts. I am eternally thankful to José and his team, especially

Drs. Larry Parnell and Chao Lai and Ms. Xian Adiconis, for taking a chance on me, giving me the

opportunity to learn from them and their many collaborators around the world, and acting as referees

numerous times. Had I not spent several excellent formative years with their group, I may never have

embarked upon this PhD.

To my friends around the department, thank you for all the wonderful silliness (and crustiness!) that

made nerve-racking situations seem less daunting. Fellow "El-Sohemys" Laura Da Costa, Drs.

Andrea Josse, Karen Eny and Leah Cahill, Lilli Mauer, Daiva Nielsen, Andre Dias, Dennis Wang,

Ouxi Tian, Joseph Jamnik, Nanci Guest, Cristina Cuda, Christine Asik, Erika Day Tasevski,

Francesca Garofalo and Hyeon-Joo Lee provided endless support and bits of wisdom. I am especially

thankful for the friendship of Laura, Andrea and Lilli, with whom I shared many great moments

across North America. Drs. Darren Brenner and Sheena Kayaniyil were fountains of knowledge and I

cherished all of our conversations. With Julie Mason, Ashleigh Wiggins and Chuck Chen I laughed

up and down the Fitzgerald hallways more times than I can count. Shirin Panahi and Ahmed

Aldughpassi were fellows in home-stretch misery. The DNS Running Rogues, and especially fellow

Executive members Matt Parrott and Lesley Hoyt, kept my body fit and my wits sharp (give or take a

few dozens of recovery pints). Matt deserves a special shout-out for being my EB roommate

extraordinaire, ultimate writing comrade and perennial partner in crime. I will forever wonder if all

that barbed wire really was the Mexican border.

My final acknowledgement goes to my parents, Jeremías and Belén, my brother, Alfonso, and my

partner, Joanne. They fed me, cared for my four-legged creatures and sat patiently through many

practice presentations. They put up with me through brutally early mornings, late nights and long

weekends of writing. They dragged me into the sunshine (both figurative and literal) when I needed it

most, and they felt each one of my frustrations and achievements as their own. Most importantly,

they kept me rooted and made me happy every day. I truly lack the words to express my love and

gratitude. I hope to thank you with my actions for the rest of my life.

Y colorín colorado, este cuento se ha acabado.

vi

Table of Contents

Acknowledgments .......................................................................................................................... iv

Table of Contents ........................................................................................................................... vi

List of Tables ................................................................................................................................. xi

List of Figures .............................................................................................................................. xiii

List of Abbreviations ................................................................................................................... xiv

List of Appendices ....................................................................................................................... xvi

Chapter 1 : Literature Review ......................................................................................................... 1

1.1 Introduction ......................................................................................................................... 1

1.2 Cardiometabolic Disease .................................................................................................... 2

1.2.1 Definition and Prevalence ....................................................................................... 2

1.2.2 Etiology ................................................................................................................... 3

1.2.3 Risk Factors ............................................................................................................ 4

1.2.4 Biomarkers of risk ................................................................................................... 5

1.3 Vitamin D .......................................................................................................................... 10

1.3.1 Background and historical perspective ................................................................. 10

1.3.2 Vitamin D metabolism .......................................................................................... 11

1.3.3 Biological Actions of Vitamin D .......................................................................... 13

1.3.4 Vitamin D Sources ................................................................................................ 13

1.3.5 Assessment of Vitamin D Nutritional Status ........................................................ 15

1.3.6 Dietary Reference Intakes ..................................................................................... 15

1.3.7 Determinants of Vitamin D Status ........................................................................ 17

1.3.8 Measurement of 25(OH)D .................................................................................... 22

vii

1.4 Vitamin D and Cardiometabolic Disease .......................................................................... 23

1.4.1 Relationship with Innate Immunity and Inflammation ......................................... 24

1.4.2 Relationship with Glycemic Regulation ............................................................... 25

1.4.3 Relationship with Lipid Metabolism .................................................................... 27

1.4.4 Relationship with the Plasma Proteome ............................................................... 28

1.5 Summary and Rationale .................................................................................................... 29

1.6 Hypothesis and Objectives ................................................................................................ 30

Chapter 2 : Positive association between 25-hydroxyvitamin D and C-reactive protein is

confounded by hormonal contraceptive use ............................................................................. 31

2.1 Abstract ............................................................................................................................. 32

2.2 Introduction ....................................................................................................................... 33

2.3 Methods ............................................................................................................................. 34

2.3.1 Study Design and Participants .............................................................................. 34

2.3.2 Anthropometrics and Physical Activity ................................................................ 35

2.3.3 Dietary Assessment ............................................................................................... 35

2.3.4 Hormonal Contraceptive Use ................................................................................ 36

2.3.5 Biochemical and 25(OH)D Measurements ........................................................... 36

2.3.6 Statistical Analysis ................................................................................................ 36

2.4 Results ............................................................................................................................... 38

2.5 Discussion ......................................................................................................................... 50

Chapter 3 : Association between circulating 25-hydroxyvitamin D and plasma cytokine

concentrations in young adults ................................................................................................. 54

3.1 Abstract ............................................................................................................................. 55

3.2 Introduction ....................................................................................................................... 56

3.3 Methods ............................................................................................................................. 57


3.3.2 Anthropometrics and physical activity ................................................................. 57

viii



3.4 Results ............................................................................................................................... 59

3.5 Discussion ......................................................................................................................... 63

Chapter 4 : Plasma 25-hydroxyvitamin D, hormonal contraceptive use, and cardiometabolic

disease risk in an ethnically diverse population of young adults ............................................. 66

4.1 Abstract ............................................................................................................................. 67

4.2 Introduction ....................................................................................................................... 68

4.3 Methods ............................................................................................................................. 69






4.4 Results ............................................................................................................................... 71

4.5 Discussion ......................................................................................................................... 85

Chapter 5 : Plasma 25-hydroxyvitamin D, hormonal contraceptive use, and the plasma

proteome in Caucasian, East Asian, and South Asian young adults ........................................ 90

5.1 Abstract ............................................................................................................................. 91

5.2 Introduction ....................................................................................................................... 92

5.3 Methods ............................................................................................................................. 93





5.3.5 Plasma Proteomic Measurements ......................................................................... 94


ix

5.4 Results ............................................................................................................................... 96

5.5 Discussion ....................................................................................................................... 110

Chapter 6 : Genetic variation in the vitamin D receptor, plasma 25-hydroxyvitamin D, and

biomarkers of cardiometabolic disease in Caucasian young adults ....................................... 115

6.1 Abstract ........................................................................................................................... 116

6.2 Introduction ..................................................................................................................... 117

6.3 Methods ........................................................................................................................... 119

6.3.1 Study Design and Participants ............................................................................ 119

6.3.2 Anthropometrics and Physical Activity .............................................................. 119

6.3.3 Hormonal Contraceptive Use .............................................................................. 119

6.3.4 Biochemical and 25(OH)D Measurements ......................................................... 119

6.3.5 Plasma Proteomic Measurements ....................................................................... 119

6.3.6 Genotyping and candidate SNP selection ........................................................... 119

6.3.7 Statistical Analysis .............................................................................................. 120

6.4 Results ............................................................................................................................. 121

6.5 Discussion ....................................................................................................................... 134

Chapter 7 : Summary, Limitations, Future Directions and Implications .................................... 137

7.1 Summary ......................................................................................................................... 138

7.1.1 Relationship with Inflammation ......................................................................... 140

7.1.2 Relationship with Glycemic Dysregulation ........................................................ 141

7.1.3 Relationship with Lipid Metabolism .................................................................. 142

7.1.4 Relationship with the Plasma Proteome ............................................................. 143

7.2 Limitations ...................................................................................................................... 143

7.3 Future Research .............................................................................................................. 144

7.4 Implications ..................................................................................................................... 145

7.5 Thesis Summary .............................................................................................................. 146

x

References ................................................................................................................................... 147

Appendices .................................................................................................................................. 174

xi

List of Tables

Table 1.1. Dietary sources of vitamin D........................................................................................14

Table 1.2. DRIs for vitamin D, by life stage..................................................................................16

Table 2.1. Study participant characteristics...................................................................................41

Table 2.2. Correlation between plasma 25(OH)D and either dietary vitamin D or plasma CRP,

across ethnic groups, by sex and hormone use..............................................................................43

Table 2.3. Association between 25(OH)D and CRP among individuals below and above the

median 25(OH)D concentration.....................................................................................................45


Table 3.2. Mean cytokine concentrations by plasma 25(OH)D tertiles.........................................61


Table 4.2. Distribution of vitamin D status among men, women HC non-users, women HC users,

and different ethnic groups............................................................................................................76

Table 4.3. Association between vitamin D status and biomarkers of cardiometabolic disease risk

across ethnic groups.......................................................................................................................77

Table 4.4. Association between vitamin D status and biomarkers of cardiometabolic disease risk

across men, women HC non-users, and women HC users............................................................80

Table 5.1. Study participant characteristics by vitamin D status..................................................99

Table 5.2. Proteins included in each plasma proteomic profile...................................................101

Table 5.3a-c. Mean plasma concentrations (μmol/L) of profile 1 proteins, by vitamin D status, by

sex and HC use.............................................................................................................................103

xii

Table 5.4. Mean plasma concentrations (μmol/L) of profile 1 proteins that were associated with

25(OH)D in women using <1mg vs. ≥1mg total hormone per day.............................................108

Table 6.1. VDR SNPs selected for analysis.................................................................................123

Table 6.2. Study participant characteristics, by vitamin D tertile................................................125

Table 6.3a-c. Associations between VDR SNPs and biomarkers of inflammation, glucose and

lipid metabolism, and plasma proteomic biomarkers where a significant (p<0.05) SNP x

25(OH)D interaction term existed................................................................................................127

xiii

List of Figures

Figure 2.1. Plasma 25(OH)D concentrations among men, women hormonal contraceptive (HC)

non-users, and women HC users, by ethnicity...............................................................................46

Figure 2.2. CRP concentrations among men, women hormonal contraceptive (HC) non-users,

and women HC users, by ethnicity................................................................................................47

Figure 2.3. Association between 25(OH)D and CRP....................................................................48

Figure 3.1. Correlation analysis between circulating 25(OH)D and cytokines.............................62

Figure 4.1. Seasonal fluctuations in 25(OH)D across ethnic groups.............................................83

Figure 4.2. Seasonal fluctuations in 25(OH)D across men, women HC non-users, and women HC

users...............................................................................................................................................84

Figure 5.1. Mean plasma proteomic profile scores, by vitamin D status, by sex and HC use, in

the population as a whole and among female HC users, female HC non-users, and men...........109

Figure 6.1. The association between rs2239182 and IP-10, an inflammatory cytokine, is modified

by vitamin D status......................................................................................................................133

xiv

List of Abbreviations

T2D - Type 2 diabetes

CVD - Cardiovascular disease

25(OH)D - 25-hydroxyvitamin D

IL - Interleukin

TNF - Tumor necrosis factor

VLDL - Very low-density lipoprotein

HDL - High-density lipoprotein

LDL - Low-density lipoprotein

CRP - C-reactive protein

FTO - Obesity-associated gene

Apo - Apolipoprotein

ABCA - ATP-binding cassette transporter

LPL - Lipoprotein lipase

LIPC - Hepatic lipase

CETP - Cholesterol ester transfer protein

GWAS - Genome-wide association study

TCF7L2 - Transcription factor 7-like 2

RANTES - Regulated upon and secreted

HOMA - Homeostasis model assessment

LC - Liquid chromatography

MS - Mass spectrometry

MRM - Multiple reaction monitoring

Ultraviolet - UV

DBP - Vitamin D binding protein

CYP - Cytochrome P450

1,25(OH)2D - 1,25-dihydroxyvitamin D

VDR - Vitamin D receptor

FGF-23 - Fibroblast growth factor 23

RXR - Retinoid X receptor

VDRE - Vitamin D response element

IU - International units

DRI - Dietary reference intake

EAR - Estimated average requirement

RDA - Recommended dietary intake

UL - Tolerable upper intake

AI - Adequate intake

BMI - Body mass index

HC - Hormonal contraceptive

xv

HRT - Hormone replacement therapy

SNP - Single nucleotide polymorphism

DHCR7 - 7-dehydrocholesterol reductase

NADSYN1 - Nicotinamide adenine

dinucleotide synthetase

RIA - Radioimmunoassay

HPLC - High-performance liquid

chromatography

FDA - Food and Drug Administration

DEQAS - International External Quality

Assessment Scheme for Vitamin D

IFN-γ - Interferon-γ

HbA1c - Hemoglobin A1c

NHANES - National Health and Nutrition

Examination Survey

TNH - Toronto Nutrigenomics and Health

FFQ - Food frequency questionnaire

MET - Metabolic equivalent task

ANOVA - Analysis of variance

ANCOVA - Analysis of covariance

SE - Standard error

IL-1RA - Interleukin-1-receptor antagonist

IP-10 - Interferon-γ-induced protein 10

PDGF-bb - Platelet derived growth factor

BB

RBP4 - Retinol-binding protein

List of Appendices

Table A.1a-c. Mean plasma protein concentrations (μmol/L) by vitamin D status, by sex and

contraceptive use.........................................................................................................................174

1

Chapter 1 : Literature Review

1.1 Introduction

Cardiometabolic diseases, such as type 2 diabetes mellitus (T2D) and cardiovascular disease

(CVD), are a leading cause of death and disability worldwide (1). Metabolic abnormalities

including dyslipidemia, central obesity and glycemic dysregulation are considered risk factors

for cardiometabolic diseases (2). In addition, there is growing evidence that abnormal innate

immune responses and chronic low-grade inflammation play a key role in the pathogenesis of

insulin resistance, T2D and CVD (3-9). Overall dietary patterns, as well as specific foods and

nutrients, have been investigated for their effects on pathophysiological processes related to

cardiometabolic disease (10-16). In particular, in recent years, a great deal of interest and

controversy have arisen around vitamin D as a therapeutic agent against the development of

numerous non-skeletal health conditions, including cardiometabolic disease (17-24).

Beyond its well established role in mineral metabolism and bone health, vitamin D may affect

processes involved in cardiometabolic disease, such as innate immunity and inflammation,

insulin sensitivity and pancreatic β cell function (25-28). Indeed, adequate circulating

concentrations of 25-hydroxyvitamin D (25(OH)D), the vitamin D metabolite used to measure

status, have been associated with improved levels of biomarkers of inflammation, glycemic

dysregulation and lipid metabolism, as well as decreased risk of the metabolic syndrome (29-34).

Recent in vitro evidence also suggests that vitamin D modulates the expression of hundreds of

genes, many of which belong to disease-associated pathways (27). However, whether these

widespread genomic effects translate into observable effects at the level of the proteome remains

unknown. Furthermore, recent systematic reviews of epidemiological studies and clinical trials

have concluded that there is insufficient evidence for a relationship between vitamin D and

cardiometabolic disease-related outcomes (35-37). In addition to differences in methodology, the

widespread inconsistencies between studies may result from unaccounted confounding from

lifestyle and demographic variables, genetic variation across individuals, and a lack of adequate

biomarkers of vitamin D action or disease progression (38).

The overall goal of this thesis was to determine the association between 25(OH)D and

biomarkers of cardiometabolic disease risk, including biomarkers of inflammation, glycemic

2

dysregulation and lipid metabolism and plasma proteomic biomarkers, and to examine whether

lifestyle factors and genetic variation in the vitamin D pathway modify these associations.

1.2 Cardiometabolic Disease

1.2.1 Definition and Prevalence

Cardiometabolic disease is an umbrella term for complex, chronic diseases, in particular T2D

and CVD, which share pathophysiologic abnormalities including hypertension, dyslipidemia,

central obesity, glycemic dysregulation, and inflammation. The burden of these conditions is

very high. Indeed, in Canada alone, T2D and CVD each affect over 3 million and nearly 1.3

million individuals, respectively (39;40). T2D itself is an important risk factor for CVD, and a

staggering 30% of mortality among Canadians is a result of CVD (41). In 2000, CVD cost the

Canadian health care system approximately $22 billion, of which about $15 billion were indirect

costs associated with short- and long-term disability and mortality. In turn, overall health care

costs for T2D in Canada were $3 billion during that year (42).

Recently, the Canadian Cardiometabolic Risk Working Group proposed that the compendium of

risk factors that contribute to the development of both T2D and CVD be referred to as

cardiometabolic risk (2). The definition of cardiometabolic risk is similar to that of the metabolic

syndrome, but it differs in a few key ways. The metabolic syndrome, which was originally

termed "Syndrome X", considers insulin resistance the main underlying disorder of both CVD

and T2D, and it quantifies relative disease risk through criteria established around cut-points for

fasting blood glucose, visceral obesity, hypertension and dyslipidemia (2;43). A recent

harmonized definition of the metabolic syndrome has been proposed (44) that takes into

consideration the criteria for diagnosis issued by various health organizations, such as the

National Cholesterol Education Program/American Heart Association (45;46), the World Health

Organization (47), the European Group for the Study of Insulin Resistance (48), and the

International Diabetes Federation (49).

By contrast, the concept of cardiometabolic risk, while acknowledging the importance of insulin

resistance and traditional risk factors to the development of T2D and CVD, also recognizes that

other, emerging risk factors, such as systemic inflammation and pro-thrombotic states, play an

3

important role in the pathophysiology of these diseases (2). Furthermore, no clinically defined

cut-points have been established to determine cardiometabolic risk, with the focus rather

centered around identifying and understanding the various mechanisms that contribute to the

development of cardiometabolic diseases (2).

1.2.2 Etiology

Insulin resistance in adipose, hepatic and muscle tissues, together with central adiposity and

altered adipokine production, are thought to underlie the pathology of cardiometabolic disorders

(2). Visceral and ectopic adiposity lead to elevated free fatty acids in the circulation, coupled

with increased secretion of pro-inflammatory cytokines, such as interleukin (IL)-6 and tumor

necrosis factor (TNF)-α, and decreased secretion of adiponectin, a cytokine that plays important

roles in the modulation of inflammation, glycemia, and lipid metabolism (41;50). Together, these

actions result in dyslipidemia, insulin resistance, systemic inflammation, and endothelial

dysfunction (41). Each of these metabolic abnormalities contributes to the development of both

T2D and CVD.

A number of studies suggest an association between excess visceral and ectopic adiposity and

overall insulin resistance (50;51). As visceral and ectopic adipose stores increase, these tissues

become insulin-resistant and hyperlipolytic, releasing an excess of free fatty acids into the

circulation (52;53). Exposure of the liver to high concentrations of free fatty acids leads to

hepatic gluconeogenesis and elevated secretion of triglycerides and very low-density lipoprotein

(VLDL). At the same time, high-density lipoprotein (HDL) cholesterol decreases and low-

density lipoprotein (LDL) cholesterol increases in density, a process which is associated with

increased risk for CVD (2;54). Together, these actions result in dysglycemia and dyslipidemia.

In addition, high free fatty acid concentrations also reduce muscular insulin sensitivity and

increase insulin secretion by the pancreas, further exacerbating glycemic dysregulation (2).

Beyond its function as an energy depot, adipose tissue acts as an endocrine organ that

participates in a number of physiologically important processes (54). For example, molecules

originating in adipose tissue, such as lipoprotein lipase (LPL) and cholesterol ester transfer

protein (CETP), are involved in the regulation of adipose tissue size and distribution (2). In

addition, adipose tissue is a major source of adipokines (55). Macrophages, which migrate to the

expanding visceral adipose tissue to remove dead cells and promote angiogenesis, secrete pro-

4

inflammatory cytokines, such as IL-6 and TNF-α (56). Increased levels of these cytokines lead to

hepatic production and secretion of acute-phase proteins such as C-reactive protein (CRP),

amyloid-A, α1-acid glycoprotein, fibrinogen, and haptoglobin. Many of these proteins are

considered either traditional or emerging biomarkers of inflammation and endothelial

dysregulation (57-59), and their circulating concentrations increase throughout the development

of cardiometabolic disease (60). At the same time, visceral obesity is often paired with decreased

concentrations of adiponectin, an anti-inflammatory cytokine with insulin-sensitizing and anti-

atherosclerotic properties that has been inversely associated with cardiometabolic disease

(61;62).

1.2.3 Risk Factors

Risk factors for both T2D and CVD include traditional factors such as sex, age, family history,

dyslipidemia, dysglycemia, and hypertension, as well as more recently recognized ones such as

abdominal obesity, insulin resistance, systemic inflammation and endothelial dysfunction (2;41).

These risk factors seem to vary in prevalence across ethnic groups. Indeed, in Canada, the degree

of cardiometabolic risk, as well as the rates of CVD, are known to differ by ethnicity, with South

Asian individuals presenting both a higher prevalence of CVD and specific risk factors, such as

impaired glucose tolerance, dyslipidemia, systemic inflammation and endothelial dysfunction,

than Caucasian or East Asian individuals (63).

Genetic variation may contribute to the variation in cardiometabolic risk across individuals. For

example, variation in the fat mass and obesity-associated gene (FTO), which is expressed in the

hypothalamus and adipocytes, has been linked to obesity, as well as T2D (64-68). A number of

genes expressed primarily in the liver have also been linked to alterations in lipid metabolism,

such as reduced HDL cholesterol and elevated triglycerides and LDL cholesterol (64;69).

Included among these are genes encoding apolipoproteins (Apo-) AI, AII, AIV, AV, E, CIII,

ATP-binding cassette transporter (ABCA) 1, LPL, hepatic lipase (LIPC), and CETP (64;70). In

the last few years, numerous genome-wide association studies (GWAS) have uncovered over 70

loci associated with small increases in risk of T2D, with the greatest relative risk (approximately

1.5) being conferred by a variant in transcription factor 7-like 2 (TCF7L2) (71-78). Many of

these genes affect β-cell functions, such as insulin secretion or conversion of pro-insulin to

5

insulin (64;71;72). Finally, a large, recent GWAS identified or confirmed over 40 variants

associated with CVD (79). Pathway analysis of the data showed an overrepresentation of genes

involved in lipid metabolism and inflammation, underscoring the importance of these processes

in the etiology of cardiometabolic diseases (79).

In addition to biologic and demographic factors, environmental exposures and lifestyle habits,

such as smoking and a lack of physical activity, also contribute to cardiometabolic risk (41;80).

Nutritional and dietary exposures may play an important role in the development of

cardiometabolic disease. For example, a diet rich in saturated fat may have deleterious effects on

cardiometabolic risk, although a recent meta-analysis concluded that the results were inconsistent

across studies (81). On the other hand, polyunsaturated fatty acid intake may beneficially affect

cardiometabolic risk, although, again, inconsistencies exist across studies (82). A recent clinical

trial reported reduced incidence of CVD among elderly individuals at risk of cardiometabolic

disease who consumed a traditional Mediterranean diet supplemented with olive oil or mixed

nuts, as compared to those consuming a low-fat diet (83). In addition, a nutrient-poor diet lacking

in fruit, vegetables, and a mild to moderate consumption of alcohol was linked to risk of

myocardial infarction in a large multinational study (80).

Of central relevance to this thesis, an increasing body of evidence suggests that vitamin D

deficiency may be an important risk factor for cardiometabolic outcomes (29-31;84-87) and even

overall mortality (88). However, recent systematic reviews have concluded that the results of the

various studies are equivocal (35-37). The widespread inconsistencies between studies may

result partly from unaccounted confounding or genetic differences (38).

1.2.4 Biomarkers of risk

Biomarkers are measurable biological factors, such as genetic variants, circulating hormones,

proteins, and metabolites, that can be used as surrogate indicators of normal physiologic

processes, pathogenic states, or to monitor responses to various exposures, therapeutic or

otherwise (89;90). The following sections describe biomarkers of physiologic processes

associated with cardiometabolic diseases.

6

1.2.4.1 Inflammation

Inflammation results from the activation of the innate immune response – the body’s immediate,

non-specific reaction against environmental insults such as pathogens, chemical or physical

injury (91). Inflammation plays a role in tissue damage prevention, restoration of tissue

homeostasis and destruction of infectious agents. It is the result of a system-wide process known

as the acute-phase response. During the acute-phase response, an array of pro-inflammatory

cytokines, such as TNF-α, IL-1β and IL-6, are released, primarily by macrophages (92). As

discussed earlier, these cytokines can enhance insulin resistance directly in adipocytes, muscle

and hepatic cells, leading to systemic disruption of insulin sensitivity and impaired glucose

homeostasis (93). Furthermore, elevated cytokine concentrations result in hepatic production and

secretion of acute-phase proteins, whose concentrations increase as cardiometabolic disease

progresses (60).

Circulating concentrations of both cytokines and acute phase proteins are commonly used

biomarkers of inflammation. Chronic inflammatory diseases, such as rheumatoid arthritis and

inflammatory bowel disease, are characterized by concentrations of these biomarkers that are

either much higher, if the biomarkers are pro-inflammatory, or much lower, if they are anti-

inflammatory, than those observed in healthy individuals (94). In contrast, the inflammation

associated with cardiometabolic diseases tends to be low-grade, with biomarker concentrations

often remaining no more than two- to four-fold higher than those observed in healthy persons

(95). Despite these modest changes in circulating concentrations, a number of inflammatory

biomarkers have been consistently associated with obesity, T2D and CVD, and therefore may be

considered biomarkers of cardiometabolic risk. These include well established markers, such as

high concentrations of CRP, IL-6 and TNF-α and low concentrations of adiponectin, as well as

less well-established markers, such as regulated upon and secreted (RANTES) and von

Willebrand factor (95).

Some have suggested that predictive power increases with the number of biomarkers being

measured (58;95). For example, one study examining the effects of an anti-inflammatory diet on

the transcriptome, proteome, and metabolome found no changes in individual biomarker

concentrations, but observed system-wide changes indicative of reduced inflammation and

7

oxidative stress (16). These changes became apparent when the authors analyzed the integrated

omics data using pathway analysis.

1.2.4.2 Glycemic Regulation

In general terms, adequate glycemic regulation is achieved when insulin is able to modulate

circulating glucose, stimulating its uptake and utilization after a meal or promoting

gluconeogenesis in the liver when glucose levels in the circulation are low (96). Insulin is

secreted from Β-cells in the pancreas, which both produce and store the hormone (97). The

progression to T2D involves both insulin resistance and β-cell dysfunction (96). In insulin-

resistant individuals, more insulin is needed to elicit glucose uptake in peripheral tissues (97). β-

cells meet this need by producing and secreting more insulin, and insulin-resistant individuals

thus become hyperinsulinemic in order to maintain euglycemia (98). Eventually, β-cells are

unable to meet the increased demand for insulin and β-cell dysfunction ensues, followed by

hyperglycemia and frank disease (99-101).

A number of approaches are used to measure insulin resistance and β-cell dysfunction. The gold

standard methods for measuring each are the hyperinsulinemic euglycemic clamp and the

hyperglycemic clamp, respectively (102). The hyperinsulinemic euglycemic clamp involves

intravenous infusion of insulin to raise its circulating levels, and parallel infusion of glucose to

maintain euglycemia. The rate at which glucose is infused represents glucose uptake by tissues

and, therefore, is indicative of the body's degree of sensitivity to the elevated insulin levels (102).

The hyperglycemic clamp, by contrast, involves a steady intravenous infusion of glucose to

maintain a hyperglycemic state; in this case, the plasma insulin response to the high glucose load

is measured (102). While accurate, these methods are time-consuming and expensive, making

their use in epidemiologic studies difficult.

The homeostasis model assessment of insulin resistance (HOMA-IR) and β-cell dysfunction

(HOMA-Beta) is a method commonly used to assess glycemic dysregulation in epidemiologic

studies. HOMA is derived from fasting plasma insulin and glucose concentrations (103). Under

fasting conditions, the relationship between glucose and insulin is representative of the balance

between insulin secretion and glucose output in the liver (104). The original HOMA method was

based on the following formulas:

8

HOMA-IR = (fasting plasma insulin x fasting plasma glucose) / 22.5

HOMA-Beta = (20 x fasting plasma insulin) / (fasting plasma glucose - 3.5), where insulin was

measured in mU/L and glucose was measured in mmol/L (105). The model was subsequently

updated to a computer algorithm that allows for a greater range of glucose concentrations and is

more appropriate for use in hyperglycemic individuals (103). In validation studies, the HOMA

model correlates well (r >0.7) with the clamp methods described earlier (106;107). Importantly,

HOMA-Beta must be interpreted in the context of insulin sensitivity, given that β-cell insulin

output is dependent on an individual's degree of insulin resistance (104).

1.2.4.3 Lipid Metabolism

Dysregulation in lipid metabolism, often characterized by hypercholesterolemia and

hypertriglyceridemia, is characteristic of cardiometabolic diseases (2). Fasting total cholesterol

levels are associated with cardiometabolic risk, although it is important to point out that it is the

lipoprotein carriers of cholesterol, rather than cholesterol itself, that are involved in pathologic

processes (108). Lipoprotein particles consist of apolipoproteins, phospholipids, cholesterol and

triglycerides, and they are classified based on their relative density and the identity of their

apolipoprotein components (108). The latter confer stability to lipoprotein particles, but they also

serve as ligands for cell surface receptors and they foster specific enzymatic processes related to

lipoprotein metabolism (109).

The main lipoproteins that affect cardiometabolic risk are LDL and HDL. Elevated

concentrations of the LDL cholesterol (particularly consisting of small, dense LDL particles),

coupled with low HDL cholesterol concentrations, promote the formation of atherosclerotic

plaques and are, therefore, an important contributor to the development of CVD (108;110;111).

LDL acts as the main transporter of cholesterol to tissues that express LDL receptors, including

the vascular endothelium (109). The development of atherosclerotic plaques is thought to begin

when circulating LDL enters the arterial intima and becomes oxidized, attracting macrophages

(110). As the macrophages engulf lipid, they become foam cells (111). These activated

macrophages also release cytokines that trigger inflammatory processes and attract yet more

macrophages, perpetuating the process. In addition, oxidized LDL damages the endothelium,

attracting platelets to the site of injury and triggering the proliferation of smooth muscle cells

(109). Smooth muscle cells also uptake oxidized LDL, becoming foam cells. The accumulation

9

of foam and smooth muscle cells leads to the formation of plaques, which, in turn, narrow the

arterial lumen and diminish blood flow (109). The protective role of HDL in atherogenesis

results from its functions in reverse cholesterol transport, both from other lipoprotein particles

and from peripheral tissues, back to the liver for subsequent excretion (110). This process

decreases the deposition of cholesterol in the vascular endothelium, preventing the formation of

atherosclerotic plaques. Indeed, high HDL cholesterol concentrations are associated with better

cardiometabolic outcomes (112).

High triglyceride concentrations have been consistently associated with cardiometabolic diseases

(113;114). However, controversy has existed over the years regarding whether these lipids are

causally associated with disease outcomes (115). Some evidence indicates that triglyceride

concentrations affect the metabolism and composition of LDL and HDL cholesterol. For

example, in insulin-resistant individuals, increased lipolysis in adipocytes results in the release of

free fatty acids and their subsequent repackaging into VLDL in the liver (115). As hepatic VLDL

production increases, CETP is activated and it transfers triglycerides to LDL and HDL

cholesterol. The enzymatic actions of LIPC lead to hydrolysis of triglycerides inside LDL,

resulting in small, dense particles. Small, dense LDL cholesterol is thought to be more

atherogenic than its normal-sized counterpart because it is more readily oxidized (116). In

addition, triglyceride-enriched HDL cholesterol is less able to scavenge cholesterol (117;118).

Numerous health organizations recommend screening various groups of individuals at risk of

CVD for these fasting LDL and HDL cholesterol, as well as triglycerides, and have set clinical

cutoff criteria for diagnosis of dyslipidemia (45;112). Total cholesterol is also measured, as a

proxy indicator of the lipoproteins present in plasma (108).

1.2.4.4 Plasma Proteome

The plasma proteome represents the largest version of the human proteome, with over 3,000

proteins identified (57;119). Among the most abundant, ranging in concentration between <1 and

approximately 800 μmol/L, are many molecules of physiological importance, such as

apolipoproteins, members of the complement system, coagulation factors, carrier proteins, and

protease inhibitors (57). Many of these proteins are acute phase reactants whose levels become

altered during inflammatory processes, and they may represent important biomarkers of

inflammation and endothelial dysregulation (59;120).

10

Recent methodological developments in the field of proteomics have allowed for the

simultaneous measurement of multiple proteins. In particular, a liquid chromatography

(LC)/tandem mass spectrometry (MS)-based, multiple-reaction-monitoring (MRM) proteomics

assay was recently developed that can simultaneously measure the concentrations of 54 high-

abundance plasma proteins (120). The concentrations of these 54 proteins were found to differ

across ethnic groups and individuals with different dietary intakes (119).

1.3 Vitamin D

1.3.1 Background and historical perspective

Because it can be synthesized endogenously after exposure to sunlight, vitamin D is not

technically a vitamin, but rather a pro-hormone (121). Structurally, vitamin D resembles steroid

hormones such as cortisol, aldosterone and estradiol (122). It is a secosteroid, where a carbon-

carbon bond is broken in one of the rings of the typical cyclopentanoperhydrophenanthrene

steroid structure. In the case of vitamin D, the breakage occurs in the 9,10 carbon bond of ring B

(122).

The most apparent manifestations of vitamin D deficiency in children and adults are rickets and

osteomalacia, respectively. The former is a bone disease characterized by poor skeletal

mineralization, slow growth, and various skeletal deformities such as bowed legs or knocked

knees, while the latter is also characterized by poor mineralization as well as muscular weakness

and bone pain (123;124). The beneficial effects of vitamin D on bone metabolism have been

appreciated for over a century (123). It is estimated that towards the late 1800s, as many as 90%

of children living in industrialized European and North American cities suffered from rickets

(125). Several anecdotal accounts reported reduced disease prevalence among children living in

sunny environments, subsequently leading to the discovery that ultraviolet radiation, both from

sunlight and from artificial sources such as mercury lamps, induced an improvement in rickets

(123;126). In the 1930s, the food industry adopted ultraviolet radiation as a method to fortify

foods and beverages, including not only milk but also products as wide-ranging as peanut butter,

hot dogs and even beer (123). Overall, vitamin D fortification in the first half of the 20th century

essentially eradicated rickets in countries that implemented it (123). However, by the 1950s,

11

excessive fortification of milk led to isolated incidents of idiopathic hypercalcemia in children,

particularly in Europe, with an ensuing restriction of vitamin D fortification across many

European countries that lasts into the present (123;127). By contrast, in Canada vitamin D

fortification of cow's milk and margarine is mandatory to this day, and fortification of other

products, such as goat milk, plant-based milk substitutes, and some calcium-fortified juices is

optional (128).

Beyond its well accepted therapeutic effects on bone metabolism, a great deal of interest has

arisen over the past few decades surrounding vitamin D's potential role in numerous non-skeletal

outcomes, including cardiometabolic diseases. The following sections provide an overview of

vitamin D physiology and a summary of the evidence for its involvement in disease-associated

processes.

1.3.2 Vitamin D metabolism

The main natural source of vitamin D in humans is cutaneous production of cholecalciferol

(vitamin D3) after ultraviolet (UV) B (290 - 315 nm) irradiation of 7-dehydrocholesterol (124).

Specifically, the conversion takes place in the plasma membranes of epidermal keratinocytes and

dermal fibroblasts, where 7-dehydrocholesterol is embedded (123). The formation of

cholecalciferol in the plasma membrane is preceded by formation of previtamin D3; the latter

metabolite is thermodynamically unstable and quickly undergoes isomerization to yield the more

stable cholecalciferol (123). Once formed, cholecalciferol is ejected from the plasma membrane

into the extracellular space (129). Plasma vitamin D binding protein (DBP), which is the main

transport protein for vitamin D metabolites, has a strong affinity for cholecalciferol and draws it

from the extracellular space into the capillaries for subsequent transport in the circulation to the

liver (130).

Humans can also obtain vitamin D through dietary consumption of cholecalciferol or its plant

and fungal equivalent, ergocalciferol (vitamin D2), from vitamin D-rich foods and supplements

(121). Cholecalciferol and ergocalciferol are very similar in structure, the only difference being

that the latter has a double bond at carbons 22-23 and a methyl group on carbon 24 (124). Many

vitamin D supplements contain ergocalciferol, although a recent meta-analysis suggests that

cholecalciferol may be more effective at raising vitamin D status than ergocalciferol (131).

Regardless of its molecular structure, dietary vitamin D is absorbed by enterocytes in the small

12

intestine and packed with other lipids into chylomicrons, which are delivered to the liver

(121;132;133).

Once in the liver, cholecalciferol is hydroxylated, primarily by cytochrome P450 (CYP-)2R1, to

25(OH)D, which is the major circulating form of vitamin D but is thought to be biologically inert

(121;134). The mitochondrial enzyme CYP27B1 then hydroxylates 25(OH)D to 1,25-

dihydroxyvitamin D (1,25(OH)2D) in the kidney and other target tissues (121;134). The latter is

considered a steroid hormone and it is the main active form of the vitamin, affecting both the

regulation of gene transcription and the activation of various signal transduction pathways (135)

when it binds to the vitamin D receptor (VDR).

As mentioned earlier, DBP is the main transport protein for vitamin D metabolites. 25(OH)D

bound to DBP enters the kidney by binding to megalin, an endocytic receptor (136;137). In its

unbound form, 25(OH)D can also diffuse passively through cell membranes (136). In turn,

1,25(OH)2D also circulates bound to DBP and can pass through membranes in its unbound form

(124).

Excessive sun exposure does not result in vitamin D intoxication because both previtamin D3 and

cholecalciferol can undergo conversion to a series of metabolically inert molecules (124).

Production of 25(OH)D from cholecalciferol in the liver is not tightly regulated. However, the

production of 1,25(OH)2D in the kidney is a rigidly regulated process affected by circulating

concentrations of calcium, parathyroid hormone, phosphorus, fibroblast growth factor 23 (FGF-

23), as well as 1,25(OH)2D itself (138). Low calcium concentrations result in a release of

parathyroid hormone, leading to increased renal 1,25(OH)2D production (138). This results in

increased intestinal calcium absorption (139). Conversely, high circulating phosphorus results in

production of FGF-23 in bone, leading to decreased 1,25(OH)2D production in the kidney (140).

Finally, high concentrations of 1,25(OH)2D signal a catabolic pathway that hydroxylates the

metabolite to water-soluble calcitroic acid, a step which is mediated by CYP24A1 (134). It is

important to note that the extra-renal regulation of 1,25(OH)2D is independent of calcium and

phosphorus metabolism. For example, in cell types such as macrophages and their precursors,

monocytes, regulation of CYP27B1, and hence 1,25(OH)2D production, is dependent upon

immune stimuli (141;142).

13

1.3.3 Biological Actions of Vitamin D

The majority of the biological actions of vitamin D result from binding of 1,25(OH)2D to the

VDR. The ligand-bound VDR heterodimerizes with retinoid X receptor (RXR) and binds to

vitamin D response elements (VDRE) in target genes, thus regulating their transcription (143).

The unliganded VDR/RXR complex can also bind to genes containing VDREs, but in these

instances transcription is prevented by a co-repressor complex (144). Recent in vitro studies have

examined 1,25(OH)2D action across the genome, using gene expression and chromatin

immunoprecipitation techniques (27;145). Hundreds of genes, belonging mainly to immune and

signaling pathways, have been identified whose expression is affected by liganded VDR binding

to VDREs (27;145). Many of these genes are known to be associated with diseases such as type

1 diabetes, Crohn's disease, multiple sclerosis, and colorectal cancer (27). In addition, CYP27B1

is expressed in numerous tissues beyond the kidney, such as cells of the immune system, brain,

prostate, pancreas, bone and skin (146). This widespread expression of the enzyme that converts

25(OH)D to the biologically active metabolite further suggests potential involvement of vitamin

D in numerous physiologic processes.

In addition to its genomic actions via binding to the VDR in the nucleus of vitamin D-responsive

cells, 1,25(OH)2D exerts non-genomic actions by activation of signal transduction pathways in

the cell membrane (121). The receptor for these actions is also thought to be the VDR, which,

while previously thought to be expressed only in the nucleus and the cytosol, has now been

shown to also be present in the membranes of various cell types (147). Binding of 1,25(OH)2D to

the cell membrane-associated VDR may lead to activation of several second messenger systems,

such as phospholipase C, protein kinase C, or G protein-coupled receptors, and this affects

several cellular processes such as the opening of voltage-gated calcium or chloride channels and

the release of intracellular calcium (121;148).

1.3.4 Vitamin D Sources

Few foods are naturally high in vitamin D. Among the highest sources of dietary vitamin D are

fatty fish and organ meats (128). In addition, in Canada and the United States, fortified products,

such as milk, are important sources of dietary vitamin D (149). One food with a particularly high

natural concentration of cholecalciferol is cod liver oil, which contains approximately 400 -

1,000 international units (IU) per teaspoon (124). By contrast, a fair-skinned individual exposed

14

to one minimal erithemal dose (MED) of UVB while wearing a bathing suit can produce the

equivalent of approximately 20,000 IUs, highlighting the importance of endogenous synthesis as

the primary source of vitamin D for humans (123). Table 1.1 shows a list of food items high in

vitamin D.

Table 1.1. Dietary sources of vitamin D. Adapted with permission from (124). The original

material is available at http://www.benthamscience.com/cdt/.

Natural sources Content (1 IU = 25 ng) Type of vitamin D

Cod liver oil 400 - 1,000 IU/tsp Cholecalciferol

Salmon, fresh wild caught 600 - 1,000 IU/3.5oz Cholecalciferol

Salmon, fresh farmed 100 - 250 IU/3.5oz Cholecalciferol, ergocalciferol

Salmon, canned 300 - 600 IU/3.5oz Cholecalciferol

Sardines, canned 300 IU/3.5oz Cholecalciferol

Mackerel, canned 250 IU/3.5oz Cholecalciferol

Tuna, canned 236 IU/3.5oz Cholecalciferol

Shiitake mushrooms, fresh 100 IU/3.5oz Ergocalciferol

Shiitake mushrooms, dried 1,600 IU/3.5oz Ergocalciferol

Egg yolk 20 IU/yolk Cholecalciferol or

ergocalciferol

Fortified sources

Fortified milk 100 IU/8oz Usually cholecalciferol

Fortified orange juice 100 IU/8oz Cholecalciferol

Infant formulas 100 IU/8oz Cholecalciferol

Fortified yogurt 100 IU/8oz Usually cholecalciferol

Fortified butter 56 IU/3.5oz Usually cholecalciferol

Fortified margarine 429/3.5oz Usually cholecalciferol

Fortified cheeses 100 IU/3oz Usually cholecalciferol

15

Fortified breakfast cereals 100 IU/serving Usually cholecalciferol

1.3.5 Assessment of Vitamin D Nutritional Status

Vitamin D status is assessed by measuring circulating 25(OH)D, which reflects both endogenous

production and dietary sources of vitamin D. The use of this metabolite for assessment, rather

than the biologically active 1,25(OH)2D, is based on a few factors. First, 25(OH)D is the most

abundant vitamin D metabolite in the circulation, with concentrations in the nmol/L range that

are approximately 1,000 times higher than those of 1,25(OH)2D (121). Second, DBP-bound

25(OH)D has a half-life of approximately 15 days as compared to 1,25(OH)2D, which has a half-

life of less than a day (150). Third, circulating concentrations of 1,25(OH)2D are under tight

control based on concentrations of calcium and phosphorus, and they tend to remain stable, or

even increase slightly, in the presence of 25(OH)D deficiency (151).

1.3.6 Dietary Reference Intakes

In late 2010, the Institute of Medicine released Dietary Reference Intake (DRI) recommendations

for vitamin D and calcium (151). The new DRIs, which include estimated average requirements

(EAR), recommended dietary allowances (RDA) and tolerable upper intake levels (UL) for all

age groups except infants, represent an update from the previous Institute of Medicine report,

released in 1997, which offered recommendations on adequate intakes only (AI).

In updating the DRIs, the Institute of Medicine considered numerous health outcomes potentially

associated with vitamin D, such as bone health, cardiometabolic diseases, autoimmune disorders

and several cancers. The report concluded that well established causality and dose-response

evidence were available only for bone health (rickets and osteomalacia), and this was used as the

health outcome indicator to determine EARs and RDAs in all age groups except for infants, for

whom insufficient data were available. The ULs were determined based on onset of

hypercalcemia (151). The updated DRIs, and the circulating 25OHD level targets on which they

are based, are shown in Table 2. In addition to these reference values, the Institute of Medicine

considered 25(OH)D <30 nmol/L as being associated with risk of clinical vitamin D deficiency

(151;152). In consideration of the potential risk for cancer associated with sunlight exposure, the

Institute of Medicine issued the current DRIs assuming minimal sun exposure. Given that

16

cutaneous synthesis is reduced in dark-skinned individuals (153), by relying only on dietary

sources of vitamin D these DRIs presumably are applicable to all ethnic backgrounds.

Table 1.2. DRIs for vitamin D, by life stage. Adapted from (151).

DRI

IU (μg) per day

AI EAR RDA UL

25(OH)D (nmol/L)a

40 40 50 >125

Life stage

0 - 6 months 400 (10)

1,000 (25)

6 - 12 months 400 (10)

1,500 (38)

1 - 3 years

400 (10) 600 (15) 2,500 (63)

4 - 8 years

400 (10) 600 (15) 3,000 (75)

9 - 70 years

400 (10) 600 (15) 4,000 (100)

71+ years

400 (10) 800 (15) 4,000 (100)

Pregnant / lactating

females

400 (10) 600 (15) 4,000 (100)

a Circulating 25(OH)D concentrations on which the recommended intakes for each category

were based, covering the requirements of 50% of the population (EAR) and 97.5% of the

population (RDA).

A significant amount of controversy accompanied the release of the new Institute of Medicine

recommendations (18;154;155). Many investigators felt that these recommendations were too

conservative and dismissive of substantial evidence for the role of vitamin D in numerous health

outcomes unrelated to bone metabolism. In addition, it was argued that the RDA should have

17

been calculated to maintain serum 25(OH)D at >75 nmol/L, a level proposed by many, including

the Endocrine Society and the Canadian Osteoporosis Society, to be associated with maximum

health benefits (20;23;24;155;156).

1.3.7 Determinants of Vitamin D Status

Circulating concentrations of 25(OH)D are dependent on numerous biologic, demographic and

behavioural/lifestyle factors. An overview of the main contributors to vitamin D status is

presented below.

1.3.7.1 Ultraviolet Exposure

As mentioned previously, cutaneous exposure to UVB in the 290 - 315 nm range results in

conversion of 7-dehydrocholesterol to cholecalciferol, which is subsequently hydroxylated into

25(OH)D (124). Therefore, an individual's degree of exposure to UVB radiation is an important

determinant of vitamin D status. Time of day, season and geographic location all affect the

amount of UVB reaching the earth's surface (125;157). In the early morning and evening, as well

as in winter, sunlight enters the atmosphere at a more oblique angle, resulting in more efficient

absorption of UVB by the ozone layer (123). Similarly, the sun's rays reach the earth at a more

oblique angle at higher latitudes, which results in diminished UVB exposure (124). In general,

those living above 33° north latitude are unable to produce vitamin D from sun exposure

between November and March; in addition, cutaneous production is minimal before 10am and

after 3pm (157;158).

1.3.7.2 Dietary Intake

The correlation between dietary vitamin D intake and circulating levels of 25(OH)D ranges

between approximately r=0.2 and r=0.7, depending on the study (159;160). The observed range

in r values suggests that a significant portion of the variation in circulating concentrations may

be accounted for by factors other than intake, as discussed below.

1.3.7.3 Body Composition

A number of cross-sectional studies report inverse associations between 25(OH)D and obesity,

with a higher prevalence of vitamin D insufficiency among obese than non-obese individuals

(161-164). This is thought to occur due to sequestration of vitamin D, which is a fat-soluble

18

molecule, in adipose tissue (86;164;165). Indeed, intestinal absorption of cholecalciferol derived

from the diet results in its packaging into chylomicrons together with various lipids, such as

triglycerides (132;133). Chylomicrons are delivered to various tissues where some of their lipid

contents are metabolized by LPL, which is particularly expressed in skeletal muscle and adipose

tissue (109;150). Some of the cholecalciferol associated with chylomicrons is absorbed by the

tissues during this process and does not reach the liver; this may partly explain why adipose

storage of vitamin D leads to lower circulating 25(OH)D in obese individuals (150). Other

factors may also contribute to the low vitamin D status observed in obese individuals, such as a

lack of time spent outdoors due to decreased mobility, an increased activity of CYP24A1 in

adipose tissue, or a decreased ability of the lipid-enriched liver to produce 25(OH)D (164).

Recent studies have suggested that a low vitamin D status may predispose to obesity (164;166).

Adipose tissue expresses the VDR as well as CYP27B1, indicating that not only is this tissue

receptive to the biologically active form of vitamin D, but can also synthesize it locally

(146;167). This suggests a potential role for vitamin D in adipose tissue's normal functions.

Some in vitro evidence has shown possible anti-lipogenic and pro-lipolytic effects of

1,25(OH)2D (168-170). In some human studies, a subject's vitamin D status at baseline predicted

later weight loss (171;172); however, other studies reported no effect of vitamin D

supplementation on weight loss (173;174), although in one of those studies supplementation was

associated with favourable effects on triglycerides and TNF-α (174). Most recently, one group

used a Mendelian randomization approach to examine the causal relationship between obesity

and vitamin D status by pooling data from over 20 cohorts, totaling more than 40,000 individuals

(175). The authors examined the association between genetic variants that have been strongly

associated with body mass index (BMI) and 25(OH)D concentrations; at the same time, they also

examined the association between variants associated with 25(OH)D and BMI. Genetic variants

linked to BMI were associated with 25(OH)D levels, but 25(OH)D-related variants were not

associated with BMI, suggesting that a high BMI, which might be indicative of increased

adiposity, may lead to a lower vitamin D status, but not the other way around (175).

1.3.7.4 Ethnicity

A wealth of evidence suggests that ethnicity is associated with vitamin D status (176). Ethnicity

is a proxy measure for skin pigmentation, and it is dark-skinned individuals who are most at risk

19

for being vitamin D-deficient (122;176). Melanin, the pigment responsible for skin colour, acts

as a natural sunscreen that absorbs UVB and prevents the conversion of 7-dehydrocholesterol to

cholecalciferol (177). Indeed, Caucasian individuals tend to have a higher vitamin D status than

other ethnic groups, and deficiency is particularly prevalent among non-Caucasian individuals

living in environments with a seasonal lack of sunlight, such as Canada (152;153). Based on data

from the Canadian Health Measures Survey, the prevalence of vitamin D deficiency (25(OH)D

<30 nmol/L, as defined by the Institute of Medicine) in Canada has been reported to range from

3% among Caucasians to 16% among non-Caucasians (152). In addition, a few small studies

have reported lower circulating 25(OH)D in South Asian and East Asian individuals living in

Canada than those of European ancestry (153;178). Interestingly, seasonal fluctuations in

25(OH)D concentrations were observed among Caucasians and East Asians, but not South

Asians (153).

In parallel with ethnic differences in vitamin D status, disparities in cardiometabolic disease

prevalence have been documented across ethnic groups (63;80). It is possible that the different

rates of cardiometabolic disease may result, in part, from the variation in 25(OH)D

concentrations seen across ethnic groups. In addition, some have suggested that the relationship

between vitamin D and certain outcomes, such as bone mineral density and atherosclerosis,

differs between Caucasians and African Americans, following an inverse direction in the former

group but a positive direction in the latter (179;180). Overall, the relationship between vitamin D

and cardiometabolic disease across ethnic groups remains relatively poorly understood (180-

183).

1.3.7.5 Medications

A number of common medications are known to affect vitamin D metabolism and alter the

circulating concentrations of both 25(OH)D and 1,25(OH)2D. For example, glucocorticoids have

been shown to downregulate VDR expression in osteosarcoma cells in vitro (184), and use of

these medications is associated with decreased 25(OH)D concentrations in large, population-

based studies (185). The mechanism behind this association remains incompletely understood,

but may be partly a result of glucocorticoid upregulation of CYP24A1 and, subsequently,

increased vitamin D catabolism (185;186).

20

In contrast to medications that decrease vitamin D status, sex hormone-containing medications,

such as hormonal contraceptives (HC) or hormone replacement therapy (HRT), are positively

associated with 25(OH)D (187-191). This relationship has been documented in individuals of

European and African ancestry, but the effects of these medications on vitamin D status in other

ethnic groups are less well understood. The effects of these drugs on circulating 25(OH)D are

thought to be mediated by estrogen. In particular, most modern HC formulations consist of a

combination of synthetic estrogen and progestins (192). In vitro and animal studies suggest that

estrogen downregulates CYP24A1 and upregulates CYP27B1 (193;194). In addition, estrogen

appears to upregulate VDR expression (195), and it is associated with elevated DBP

concentrations (196;197). Thus, estrogen seems to play an important role in modulating vitamin

D metabolism, but the clinical relevance of its effects on circulating vitamin D metabolites is

poorly understood.

Understanding the relationship between HC and vitamin D is important because, in addition to

their effects on vitamin D, HC have been associated with numerous adverse health outcomes

(198-203). Therefore, these drugs may be important confounders of the relationship between

25(OH)D and cardiometabolic disease. HC, which are used by over 100 million women

worldwide, are prescribed not only to prevent unwanted pregnancies, but to treat various other

conditions such as acne, polycystic ovarian syndrome and endometriosis (204;205). At the same

time, HC have been linked to widespread effects at the level of the plasma proteome and a higher

risk of adverse health outcomes, such as elevated concentrations of CRP and other inflammatory

biomarkers (198;199;203;206), impaired glucose tolerance and CVD (204;207;208). However,

despite these documented effects on pathways that become dysregulated during the progression

of cardiometabolic disease, no studies have directly investigated whether these widely used

medications modify the association between 25(OH)D and cardiometabolic disease.

1.3.7.6 Genetic Variation

Based on twin studies, heritability estimates for 25(OH)D levels in plasma range between 43%

and 77% (209-211). These high estimates suggest that genetic factors may be an important

contributor to vitamin D status.

A number of candidate gene studies have examined the role of selected variants along the

vitamin D metabolic pathway on circulating 25(OH)D levels (212). Single nucleotide

21

polymorphisms (SNPs) in vitamin D-related genes were found to be associated with vitamin D

status, including CYP27B1 (213;214), CYP2R1 (215;216), GC (the gene that encodes DBP)

(214-220), VDR (218;221), CYP24A1 (216) , and DHCR7 (encoding 7-dehydrocholesterol

reductase, which synthesizes cholesterol from 7-dehydrocholesterol) (216).

Most recently, studies conducted using the GWAS approach have surveyed variants across the

entire human genome in various populations in association with 25(OH)D (222-226). One study

conducted on approximately 4,500 individuals of European ancestry found significant

associations between 25(OH)D concentrations and SNPs in GC, NADSYN1 (encoding

nicotinamide adenine dinucleotide synthetase), which was in high linkage disequilibrium with a

SNP in DHCR7, and CYP2R1 (223). The SNP with the lowest p-value was rs2282679, which is

found within GC. Another study evaluated circulating 25(OH)D and genetic variation in a

sample consisting of about 34,000 individuals of European descent (223). This study also found

significant associations between 25(OH)D levels and variants within or near GC, DHCR7, and

CYP2R1, as well as CYP24A1 (220). As in the previous study, the SNP with the lowest p-value

was rs2282679, suggesting an important role for this variant in modulating 25(OH)D

concentrations. Several smaller (sample sizes of approximately 230 to 1,000) studies have been

conducted on individuals from the Framingham Heart Study (226), as well as Hispanic

individuals (218) and children with asthma (225). These studies identified no variants associated

with 25(OH)D at the Bonferroni-corrected genome-wide level. This may have been due to the

small size of the study populations. However, one of the studies was able to replicate some of the

findings of the two large GWAS conducted on Caucasian individuals (225). In particular, the

authors reported associations between 25(OH)D and rs2282679, further highlighting the

potential importance of this SNP in determining vitamin D status, as well as between 25(OH)D

and rs10741657 (found within CYP2R1) (225).

Despite providing some insights about the genetic determinants of vitamin D status, it is

estimated that, in total, the variants identified through GWAS account for only about 4% of the

variation in vitamin D status (212;227). This is in contrast to the high heritability estimates

provided by twin studies (209-211). Of note, the SNPs identified through GWAS are common

variants with minor allele frequencies >5%, and the GWAS studies conducted thus far have not

considered rare variants, epistasis, gene-environment interactions, or other potential genetic

factors that may play an important role in determining vitamin D status (227;228). However,

22

taken together, the evidence from the various study types indicates that genetic variants affect

vitamin D status and may, therefore, act as important confounders of the association between

vitamin D and specific health outcomes.

1.3.8 Measurement of 25(OH)D

Several methods exist to measure 25(OH)D. The original method, a competitive binding assay,

was developed in the 1970s and it measured 25(OH)D concentrations using DBP as a binding

agent and radioactively labelled 25(OH)D as a reporter (229;230). The assay was cumbersome

and time-consuming and was not suitable for use in environments where a large number of

samples are processed, such as clinical laboratories (231). Other methods introduced since then

include radioimmunoassay (RIA), high-performance liquid chromatography (HPLC) and LC-

MS/MS (230;232). In addition, more recently several automated immunoassays have become

available as well (232).

In general, measuring 25(OH)D is challenging because of its highly hydrophobic nature, its

strong affinity for DBP, and the structural similarities between 25(OH)D3, which is derived from

cholecalciferol, and the ergocalciferol-derived 25(OH)D2 (230). In fact, not every available assay

is able to either measure, or distinguish between, both metabolites. HPLC and LC-MS/MS are

considered the gold standards because they are considered to be the most accurate, precise and

specific methods, they are able to differentiate between 25(OH)D3 and 25(OH)D2, and they can

also measure other vitamin D metabolites, such as 24,25(OH)D (231-234). Some shortcomings

of these assays include the fact that they are technically sophisticated, time-consuming, and

difficult to implement in high-throughput clinical settings (231). RIA is less technically

demanding than the LC-based methods and has been used extensively. Indeed, it was the first

methodology approved by the United States Food and Drug Administration (FDA) for clinical

assessment of 25(OH)D (230). More recently, DiaSorin, the same company that developed the

FDA-approved RIA, introduced a chemiluminescence-based immunoassay method called

LIASON (235). Unlike RIA, which is a manual method, LIAISON is automated and intended for

use in high-throughput settings. Furthermore, while RIA required the use of radioactive

materials, LIAISON does not.

In consideration of the difficulty inherent to measuring vitamin D and in acknowledgement of

problems with inter-laboratory and inter-assay variability, the International External Quality

23

Assessment Scheme for Vitamin D (DEQAS) was launched in 1989 with the goal of assessing

the performance and reliability of the various 25(OH)D methods across laboratories worldwide

(234). At present, approximately 600 groups participate (231). While issues with variability still

exist (233;236;237), it is generally agreed that the existence of DEQAS has greatly improved the

overall performance of the various assays over time (238).

1.4 Vitamin D and Cardiometabolic Disease

Numerous studies have examined the association between vitamin D and markers of

cardiometabolic risk in both adults and children (29;87;239-241). However, recent systematic

reviews have concluded that the results of the various studies are equivocal (35-37;242;243).

Adding to the controversy, a recently published study reported lower circulating 25(OH)D in the

offspring of nonagenarians, suggesting that low vitamin D status may not be causally associated

with increased mortality from age-related diseases (244). Overall, the inconsistencies between

studies have been attributed to small sample sizes and differences in age, sex, ethnicity and

disease status between the populations studied. In addition, unaccounted factors, such as genetic

variation between individuals and HC use among women, may confound the association between

vitamin D and cardiometabolic disease-related processes. Indeed, a large, long-term prospective

study recently identified a novel association between a VDR variant and a composite of risk for

hip fracture, cancer, myocardial infarction, and mortality (38). This association was only

apparent in individuals with a low vitamin D status, highlighting the importance of taking into

account gene-environment interactions. Furthermore, one study using data from the Women's

Health Initiative found that HRT (which, like HC, often contain synthetic estrogen), modified the

effect of calcium and vitamin D supplementation on colorectal cancer (245), and an analysis of

data from a large German cohort showed that HRT use attenuated an inverse association between

25(OH)D and breast cancer (190). While neither of these studies assessed the role of HC, the

trends observed with HRT suggest that HC may also act as important confounders.

The following sections provide a summary of the relationship between vitamin D and specific

pathways that become dysregulated during the development of cardiometabolic disease.

24

1.4.1 Relationship with Innate Immunity and Inflammation

Most immune cell types, including macrophages, express the VDR as well as CYP27B1, and

thus are not only responsive to 1,25(OH)2D but can synthesize it locally from circulating

25(OH)D (28). In vitro studies conducted on macrophages have shown that, in these cells,

pathogen exposure leads to the upregulation of CYP27B1 through signaling by toll-like receptors

(246-248). The enhanced production of 1,25(OH)2D results in its binding to the VDR, and the

complex of the two subsequently upregulates the transcription of genes encoding endogenous

antimicrobial peptides (249). The upregulation of CYP27B1 and ensuing production of

1,25(OH)2D is also triggered by signaling from interferon-γ (IFN-γ) (141;142). In addition to

antimicrobial activity, cytokine production by macrophages and other immune cell types is

directly regulated by 1,25(OH)D2 (28). Indeed, in vitro and animal studies have shown that

1,25(OH)D2 downregulates the production of pro-inflammatory cytokines such as IFN-γ and IL-

2 and it upregulates the transcription of immuno-supressive cytokines such as IL-10 (250-253).

These effects result both through direct binding of the 1,25(OH)2D-VDR complex to VDREs

present in cytokine genes and through modulation of pro-inflammatory intracellular signalling

pathways, such as those downstream of the nuclear transcription factor NFκB and nuclear factor

of activated T cells (NFAT) (28).

Despite abundant experimental evidence for a relationship between vitamin D and inflammation,

epidemiological studies and clinical trials have yielded inconsistent results. On the one hand,

cross-sectional studies carried out in diverse populations have found an inverse association

between circulating concentrations of 25(OH)D and pro-inflammatory cytokines, acute phase

proteins and hemostatic markers (85;254-257). Certain intervention trials also found that oral

supplementation with vitamin D resulted in increased concentrations of 25(OH)D and decreased

concentrations of CRP and other inflammatory markers (31;258-260). On the other hand, some

other cross-sectional and intervention studies found null associations between 25(OH)D and

markers of inflammation (241;261-266), and one recent study reported a positive association

between 25(OH)D and CRP among vitamin D-sufficient adults (267).

Taken together, the evidence from human studies regarding the relationship between vitamin D

and inflammation is inconsistent. The inconsistencies may be due in part to unaccounted

confounding. Indeed, one recent study reported that FokI, a polymorphism of the VDR gene,

25

modifies the response to vitamin D supplementation (1,000 IU per day for 12 weeks) in subjects

with T2D (268). At the end of the treatment, the change in CRP and IL-6 concentrations was

much more pronounced (i.e., the concentrations of both biomarkers were lower) in individuals

who were homozygous for the major allele than those who were risk allele homozygotes. The

latter group may, therefore, be classified as hypo-responders with respect to vitamin D

supplementation (268). It is not known whether other genetic variants modify the effect of

vitamin D on various cardiometabolic outcomes. In addition, it is possible that lifestyle factors

such as HC affect the relationship between 25(OH)D and inflammation. As mentioned

previously, these medications affect circulating 25(OH)D (188) and they are also associated with

increased levels of inflammatory biomarkers (198;199;203;206).

1.4.2 Relationship with Glycemic Regulation

There is evidence for a higher prevalence of diagnosis of T2D during winter, when UVB

radiation is minimal, than in summer (269). This suggests a potential role for vitamin D in

protecting against T2D. Indeed, experimental evidence accumulated over the years indicates that

vitamin D plays a potentially important role in glycemic regulation. The role of vitamin D was

initially suggested from the presence of the VDR in pancreatic β cells (270). In these cells,

1,25(OH)2D enhances insulin production and secretion (271;272). Furthermore, the promoter of

the insulin receptor gene contains a VDRE (25), and 1,25(OH)2D stimulates insulin

responsiveness and glucose update in vitro (26). Another potential mechanism behind the

beneficial actions of vitamin D on glycemic regulation might revolve around this micronutrient's

important immunomodulatory actions, as described in a previous section (273). Through

downregulation of the NFκB pathway, 1,25(OH)2D attenuates pro-inflammatory cytokine

production, thus inhibiting β-cell apoptosis and promoting their survival (274). Finally, a number

of studies have shown associations between SNPs in VDR and GC, glucose intolerance and

insulin secretion, providing additional evidence for a relationship between vitamin D and

glycemic regulation (275-278).

Various cross-sectional studies have shown an inverse association between circulating 25(OH)D

and glycemic status measures, such as fasting plasma glucose, oral glucose tolerance tests,

hemoglobin A1c (HbA1c), and insulin resistance, as well as the metabolic syndrome

(29;87;183;239;279-283). For example, data from NHANES showed an inverse, dose-dependent

26

association between circulating 25(OH)D and diabetes prevalence in non-Hispanic whites and

Mexican Americans, although the association was not observed in non-Hispanic blacks

(183;279). The same inverse trend was observed between 25(OH)D and insulin resistance, but

there was no association with β-cell function (183;279). Data from the same cohort also showed

an inverse association between 25(OH)D and prevalence of the metabolic syndrome (279). More

recently, data from the Women’s Health Initiative Calcium-Vitamin D trial have shown an

association between 25(OH)D and some cardiometabolic risk factors (e.g. triglycerides,

triglyceride:HDL-cholesterol ratio) as well as metabolic syndrome (239). However, other

cardiometabolic risk factors (e.g. LDL-cholesterol, HDL-cholesterol, glucose, insulin, HOMA-

IR, HOMA-Beta) were not associated with 25(OH)D (239). A recent study using data from the

United States National Health and Nutrition Examination Survey (NHANES) reported a

synergistic interaction between low 25(OH)D and abdominal obesity that increases the risk of

insulin resistance (284).

In prospective studies, dietary vitamin D intake has been associated with impaired glycemic

regulation and incidence of T2D. For example, a study using data from the Nurses Health Study

found a significant inverse association between total vitamin D intake and T2D risk (285). More

recently, a study of Canadian individuals reported a reduced risk of progression to impaired

fasting glucose, impaired glucose tolerance and frank T2D over a 3-year period among those

with a higher baseline 25(OH)D status (30). However, intervention studies have shown

conflicting results regarding the effect of vitamin D supplementation on glycemic regulation and

T2D incidence. One study reported that supplementation with 1,25(OH)2D for one week did not

affect fasting glucose or insulin sensitivity in healthy young men (286). Another study found

that, among T2D patients, supplementing with 1,25(OH)2D for three weeks ameliorated insulin

secretion, but did not improve glucose tolerance after a 75 g oral glucose load (271). Yet another

study found that, among middle-aged men with impaired glucose tolerance or mild T2D who had

adequate serum vitamin D levels at baseline, supplementation with 1,25(OH)2D for 3 months had

no effect on glucose tolerance (287). However, in a cross-over design, 20 diabetes patients with

inadequate vitamin D serum levels who were supplemented with 1,25(OH)2D for 4 days had

improved insulin and C-peptide secretion, but showed no changes in fasting or stimulated

glucose, C-peptide, or insulin concentrations (288). The data from a 2-year-long trial designed to

assess bone health in postmenopausal women without diabetes were analyzed a posteriori for

27

effects of vitamin D3 or 1,25(OH)2D supplementation on fasting glucose, and found no

significant effects (289). Furthermore, a post-hoc analysis of data from a three-year-long trial for

bone health found that daily supplementation with 700 IU of cholecalciferol and 500 mg of

calcium citrate malate had no effect on blood glucose or insulin resistance in elderly adults who

were normoglycemic. In contrast, subjects with IGT at baseline presented a significantly

improved fasting glycemic response and insulin sensitivity after three years (273). A

randomized, controlled trial found that daily supplementation with 400 IU of cholecalciferol for

6 months in insulin-resistant, vitamin D-deficient South Asian women living in New Zealand

resulted in improved insulin resistance and sensitivity (290). Another recent trial found that

daily consumption of a vitamin D-fortified yogurt drink (1,000 IU of cholecalciferol per day)

improved glycemic control in Iranian T2D patients (32). Most recently, it was reported that daily

supplementation with 4,000 IU of cholecalciferol for six months resulted in increased circulating

25(OH)D and improved fasting insulin and HOMA-IR among obese adolescents (86).

1.4.3 Relationship with Lipid Metabolism

The VDR is expressed ubiquitously throughout the body, including the liver and adipose tissue,

which are major contributors to lipid metabolism (109;121;291). Therefore, one might

hypothesize that vitamin D plays a role in regulating lipid concentrations. However, the potential

mechanistic link by which vitamin D may affect lipid metabolism remains unclear.

In epidemiologic studies, 25(OH)D has been associated with higher HDL and ApoA-1, which is

the main lipoprotein particle in HDL (292). This has led some investigators to propose that

vitamin D contributes to the maintenance of adequate ApoA-1 and, subsequently, HDL

concentrations (292;293). This would, in turn, lead to a less atherogenic lipid profile because of

the cholesterol scavenging properties of HDL. However, in experimental studies VDR knockout

mice had higher HDL and ApoA-1 than wildtype animals (294). This paradoxical finding was in

agreement with cell culture studies, where 1,25(OH)2D inhibited both Apo-A1 secretion and

mRNA (295). Therefore, it is not clear whether vitamin D affects production of ApoA-1, and

whether this might lead to increases in HDL.

Another potential mechanistic link between vitamin D and lipid metabolism might be through

effects on circulating triglycerides. Some evidence from in vitro studies suggests that LPL

expression is upregulated in adipocytes exposed to 1,25(OH)2D (169). In vivo, this might

28

translate into decreased triglyceride concentrations in the circulation. Indirectly, vitamin D might

also affect triglyceride concentrations via calcium metabolism. Higher circulating 25(OH)D

leads to the increased absorption of dietary calcium (296). The absorbed calcium may form

insoluble complexes with fatty and bile acids, thus inhibiting cholesterol absorption and leading

to its excretion (297;298). Increases in intracellular calcium in the liver following increased

calcium absorption might also lead to decreased hepatic triglyceride production (299). Another

indirect link between 25(OH)D and triglycerides might be through the elevations in parathyroid

hormone concentrations that parallel decreases in 25(OH)D (300). Increased circulating

parathyroid hormone has been linked to hypertriglyceridemia in rats (301).

In general, cross-sectional studies seem to suggest an inverse association between 25(OH)D and

a pro-atherogenic lipid profile, although the results have been somewhat inconsistent

(163;239;280;281;291;293;302-305). Clinical trials have yielded even more inconsistent results,

as summarized by one recent meta-analysis (306). It must be noted that the sample sizes of all

the trials were small, none of the studies was designed to assess the specific relationship between

vitamin D and lipids, the vitamin D dosage differed widely across studies, and some used

cholecalciferol, whereas others used ergocalciferol (291;306). In addition, unaccounted

confounders such as lifestyle variables and genetic variation may account for some of the

variability in results across studies.

1.4.4 Relationship with the Plasma Proteome

Recent in vitro studies have examined vitamin D action across the genome, using gene

expression and chromatin immunoprecipitation techniques (27;145). Hundreds of target genes,

belonging mainly to immune and signaling pathways, have been identified, and many are known

to be associated with diseases such as type 1 diabetes, Crohn's disease, multiple sclerosis, and

colorectal cancer. However, it is not known whether the observed widespread effects of vitamin

D at the genome level in vitro translate into downstream effects at the level of the proteome in

vivo.

29

1.5 Summary and Rationale

Cardiometabolic diseases, including T2D and CVD, represent a significant global public health

burden (1). There is great interest in the beneficial effects of vitamin D on cardiometabolic

disease, given its potential modulation of disease-associated pathways such as inflammation and

glucose and lipid metabolism. However, while vitamin D modulates the expression of hundreds

of genes in vitro (27), whether these widespread genomic effects translate into observable effects

at the level of the proteome in vivo remains unknown. Furthermore, recent systematic reviews

have concluded that there is insufficient evidence for a relationship between vitamin D and

cardiometabolic disease-related outcomes (35-37). The inconsistencies between studies may

result from unaccounted confounding from lifestyle and demographic factors, genetic variation

across individuals, and a lack of adequate biomarkers of vitamin D action or disease progression

(38). Identifying potential confounders is important because it may enhance our understanding of

the relationship between vitamin D and cardiometabolic disease.

As outlined in this literature review, use of HC by women is associated with both elevated

25(OH)D and a worse cardiometabolic profile (188;204;207). However, the potential role of

these medications as confounders of the relationship between 25(OH)D and cardiometabolic

disease remains poorly understood. Genetic variation in VDR represents another potentially

important source of confounding, because vitamin D exerts its biological actions by binding to

the VDR and the latter is expressed in multiple tissues throughout the body (121). Indeed, some

studies suggest that traditional VDR variants such as TaqI or FokI may affect certain aspects of

cardiometabolic disease, but the potential role of other variants in this gene remains poorly

explored. In addition, few studies have considered the interaction between VDR genotype and

vitamin D status. A recent study reported a link between a VDR variant and a composite risk

score for major clinical outcomes, but the association was present only in those with a low

vitamin D status (38). The results of that study highlight the importance of considering gene-

nutrient interactions.

The overall goal of this thesis was to conduct a comprehensive examination of the association

between 25(OH)D and biomarkers of inflammation, biomarkers of glucose and lipid metabolism,

and plasma proteomic biomarkers, and to assess the role of potential lifestyle and genetic

confounders, namely HC use and genetic variation in VDR, in modifying this relationship.

30

1.6 Hypothesis and Objectives

Hypothesis: Common genetic variants and lifestyle factors modify the association between

circulating levels of vitamin D and biomarkers of inflammation, glucose and lipid metabolism,

and plasma proteomic biomarkers in an ethnically diverse population of young adults.

Objectives:

1. To examine the association between 25(OH)D and CRP, and to determine whether HC

use modifies this association.

2. To examine the association between plasma 25(OH)D and five cytokines, and to

determine whether HC use modifies this association.

3. To examine the association between 25(OH)D and biomarkers of glucose and lipid

metabolism, and to determine whether HC use modifies this association.

4. To examine the association between 25(OH)D and plasma proteomic biomarkers, and to

determine whether HC use modifies this association.

5. To examine whether genetic variants in the VDR are associated with biomarkers of

inflammation, biomarkers of glucose and lipid metabolism, and plasma proteomic

biomarkers, and to determine whether vitamin D status modifies these associations.

31

Chapter 2 : Positive association between 25-hydroxyvitamin D and C-

reactive protein is confounded by hormonal contraceptive use

This chapter is adapted with permission from an article published in the Journal of

Women's Health (c) 2013 (copyright Mary Ann Liebert, Inc.); the Journal of

Women's Health is available online at: http://online.liebertpub.com. The original

article is the following:

García-Bailo B, Josse AR, Jamnik J, Badawi A, El-Sohemy A. Positive

Association Between 25-Hydroxyvitamin D and C-Reactive Protein is Confounded

by Hormonal Contraceptive Use. J Womens Health (Larchmt). 2013

May;22(5):417-25. doi: 10.1089/jwh.2012.4046.

32

2.1 Abstract

Background: Studies of the relationship between vitamin D and inflammation are equivocal.

This may be due to unaccounted confounding. Hormonal contraceptive (HC) use is associated

with elevated circulating 25-hydroxyvitamin D (25(OH)D) in Caucasians and African-

Americans, but its effects on 25(OH)D in other ethnicities are unclear. HC use is associated with

elevated C-reactive protein (CRP), an inflammatory biomarker. Our objectives were to assess

the effect of HC use on 25(OH)D across ethnic groups, and to examine the association between

HC, 25(OH)D and CRP in an ethnically diverse population of young adults.

Methods: We recruited Caucasian, East Asian, and South Asian individuals (n=1,403) from

Toronto, Canada. Fasting blood measures of 25(OH)D and CRP were obtained.

Results: Across ethnic groups, women HC users (n=280) had higher 25(OH)D and CRP than

women HC non-users (n=695) and men (n=428) (p<0.008 and <0.0001, respectively).

Circulating 25(OH)D was positively associated with CRP in the entire population in models not

accounting for HC use (β=0.010±0.003; p<0.0001). There was no association when men and

women HC non-users were examined separately. Among women HC users, there was no

association after accounting for hormone dose. A positive association between 25(OH)D and

CRP among individuals above the median 25(OH)D (≥51.9 nmol/L) was not significant after

adjustment for HC use. No association was observed among individuals below the median.

Conclusions: HC use and 25(OH)D were positively associated across ethnic groups. We found

no association between 25(OH)D and CRP when HC use was accounted for. HC use confounds

the association between 25(OH)D and CRP.

33

2.2 Introduction

There is great interest in the potential preventive and therapeutic effects of vitamin D on

numerous chronic diseases, such as T2D, CVD, various cancers and autoimmune disorders, such

as multiple sclerosis and rheumatoid arthritis. Chronic inflammation, characterized by elevated

circulating concentrations of pro-inflammatory cytokines, such as IL-6, and hepatic acute phase

proteins, such as CRP, is a common feature of these diseases (56;307-309).

Growing evidence from in vitro and animal studies supports a role for vitamin D in innate

immune responses and inflammation (144;248). However, epidemiological studies and clinical

trials have yielded inconsistent results. On the one hand, cross-sectional and case-control studies

carried out in diverse populations have found an inverse association between circulating

concentrations of 25(OH)D and CRP (85;255). Certain intervention trials also found that oral

supplementation with vitamin D resulted in increased concentrations of 25(OH)D and decreased

concentrations of CRP and other inflammatory markers (31;258). On the other hand, some other

cross-sectional and intervention studies found a null association between 25(OH)D and markers

of inflammation, including CRP (261;263), and one recent study reported a positive association

between 25(OH)D and CRP among vitamin D-sufficient adults (267). Circulating 25(OH)D, the

metabolite used clinically to determine vitamin D status, captures both dietary vitamin D intake

and endogenous production after exposure to sunlight (121), while CRP is considered a robust

marker of systemic inflammation (56).

The discrepancies between studies have been attributed partially to differences in vitamin D

dosage, as well as nutritional and disease status of the populations surveyed. However, a number

of biological, lifestyle, and demographic factors affect circulating 25(OH)D concentrations, such

as skin colour, sun exposure behaviours, geographical location, seasonality, genetic variation,

and use of certain medications (121;153;191;223;310). These and other factors may play a role

as confounders in the relationship between vitamin D and inflammatory biomarkers.

A number of studies have reported a positive association between estrogen and 25(OH)D in

women who use estrogen-containing HC or HRT (187-191). However, most studies were

conducted in individuals of European or African ancestry, and the effects of these medications

on 25(OH)D in other ethnic groups are less well understood. In addition, use of HC and HRT has

also been associated with deleterious health outcomes, such as increased risk of CVD and cancer,

34

as well as a worsened metabolic profile, thrombosis, and elevated inflammation (198-203). A

recent study examining the association between HC use and the plasma proteome identified a

number of novel pathways that are affected by HC use, suggesting that the effects of HC on

disease-associated metabolic pathways may be widespread (207). Use of these medications may,

therefore, confound the association between 25(OH)D and various health outcomes. Indeed, one

study from the Women's Health Initiative trial found that HRT modified the effect of calcium

and vitamin D supplementation on colorectal cancer (245), and an analysis of data from a large

German cohort showed that HRT use attenuated an inverse association between 25(OH)D and

breast cancer (190). The use of HC in particular has been associated with elevated concentrations

of CRP and other inflammatory biomarkers (198;199;203;206). These effects are apparent after

only three months of continued HC use (199), and are thought to result from first-pass hepatic

metabolism of estrogen-containing HC, which may affect hepatic production of CRP and other

pro-inflammatory acute phase reactants (206). However, despite these well-known effects on

inflammation, no studies have directly investigated whether HC use, which is widespread among

pre-menopausal women worldwide, modifies the association between 25(OH)D and

inflammation. Accounting for this potential confounding effect may help explain the unexpected

positive association between 25(OH)D and CRP observed in a previous study (267).

The objectives of this study were to assess the effect of HC use on 25(OH)D concentrations

across different ethnic groups, and to examine the interaction between HC use, 25(OH)D

concentrations and CRP in a large, ethnically diverse population of young adults.

2.3 Methods

2.3.1 Study Design and Participants

Study participants were from the Toronto Nutrigenomics and Health (TNH) study, which is a

cross-sectional analysis of individuals aged 20-29 years living in Toronto. Recruitment occurred

from the autumn of 2004 to the autumn of 2010. All recruited individuals gave written informed

consent, and the protocol was approved by the Ethics Review Board of the University of

Toronto. All study participants gave an overnight fasting blood sample and completed a food

frequency questionnaire (FFQ), a general health and lifestyle questionnaire, and a physical

35

activity questionnaire. We excluded individuals who were unable to provide a blood sample, as

well as pregnant or breastfeeding women.

The individuals included in this study (n=1,403) were non-smoking men (n=428) and women

(n=975), free of diabetes, who had available information on 25(OH)D, CRP, HC use among

women, and each additional variable considered for analysis at the time that this study was

carried out. Participants were classified as Caucasian (n=702), East Asian (n=536), or South

Asian (n=165) based on self-reported ancestry, as described previously (119). Caucasians were

of European, Middle Eastern or Hispanic descent. East Asians had Chinese, Japanese, Korean,

Filipino, Vietnamese, Thai or Cambodian descent. South Asians were individuals with ancestors

from India, Pakistan, Sri Lanka and Bangladesh. Individuals who were First Nations Canadian,

Afro-Caribbean, or of mixed ancestry were excluded from the analyses reported here because of

the insufficient sample size.

Participants were grouped by season based on the date when they provided blood: winter

(December, January, February), spring (March, April, May), summer (June, July, August), and

autumn (September, October, November).

2.3.2 Anthropometrics and Physical Activity

Anthropometric variables, including height, waist circumference, BMI, and systolic and diastolic

blood pressure, were measured with the participant wearing light clothing and no shoes, as

previously described (119). Participants provided information on physical activity by

questionnaire, and level of physical activity was expressed as metabolic equivalent task (MET)-

hours per week, as described previously (311).

2.3.3 Dietary Assessment

Dietary intake was assessed using the Toronto-modified Willett questionnaire, which is a 196-

item semi-quantitative FFQ that includes questions on dietary supplements, such as

multivitamins and vitamin D supplements (119;312;313). Subject responses to the individual

foods were converted to daily number of servings for each item. Estimates of intake were based

on the United States Department of Agriculture Nutrient Database for Standard Reference (314).

Vitamin D intake is reported as IU per day.

36

2.3.4 Hormonal Contraceptive Use

Women reported use of HC in the general health and lifestyle questionnaire. The questionnaire

asked female participants if they were currently using HC or had used them in the past. Those

who reported current use of any type of HC, regardless of delivery method (oral, transdermal,

vaginal, injection, etc.), were classified as HC users (n=280), while those who were not currently

taking the medications were classified as HC non-users (n=695). The questionnaire also asked

questions about the brand of HC currently used. Using information on dose available from the

specific brand manufacturer's website, HC users were classified as taking <1mg (n=170) vs.

≥1mg (n= 63) total hormone (calculated as the sum of estrogen- and progesterone-derived

ingredients) per day. A total of 47 HC users were excluded from this classification because the

type of HC medication they reported made it difficult to ascertain daily hormonal exposure (e.g.,

receiving HC by injection, since the timing of this mode of delivery ranges from weeks to

months depending on the brand; reporting 'Other'; or reporting HC use but not reporting a type of

medication).

2.3.5 Biochemical and 25(OH)D Measurements

Each participant provided a blood sample after a minimum 12-hour overnight fast. Individuals

with temporary inflammatory conditions, including recent piercings or tattoos, acupuncture,

medical or dental procedures, vaccinations or immunizations, infections, or a fever, gave blood

after a two-week wait. Samples were collected at LifeLabs Medical Laboratory Services

(Toronto, ON, Canada). High-sensitivity CRP was measured at LifeLabs using a latex-enhanced

immunoturbidimetric method with the Siemens Advia® 2400 analyzer (Siemens Healthcare

Diagnostics Inc., Tarrytown, NY, USA). Plasma 25(OH)D was measured by HPLC-MS/MS at

the University Health Network Specialty Lab (Toronto, ON, Canada). The concentration of

25(OH)D is reported as the sum of 25-hydroxycholecalciferol and 25-hydroxyergocalciferol for

each participant.

2.3.6 Statistical Analysis

The software SAS (version 9.2; SAS Institute Inc, Cary, NC, USA) was used to perform all

statistical analyses. The α error was set at 0.05, and all reported p-values are two-sided.

Continuous variables that were not normally distributed were loge- or square root-transformed

prior to analysis, unless otherwise indicated. In each case, the p-values from analyses using the

37

transformed values of these variables are reported, but untransformed means, regression

coefficients, and measures of spread (standard deviations or standard errors) are reported to

facilitate interpretability.

Subject characteristics between men, women HC non-users, and women HC users were

compared using χ2 tests for categorical variables and analysis of variance (ANOVA) for

continuous variables. We then compared circulating 25(OH)D and CRP concentrations among

men, women HC non-users, and women HC users within each ethnic group using analysis of

covariance (ANCOVA) adjusted for age, waist circumference, physical activity, and season of

recruitment. We also conducted correlation analyses to explore the association between

circulating 25(OH)D and total dietary intake of vitamin D in the population as a whole,

independently within each ethnic group, and separately among men, women HC non-users and

women HC users within each ethnic group. Correlations were also conducted between 25(OH)D

and CRP in the same manner. Where indicated, correlation coefficients were compared between

sub-groups using the Fisher's z transformation.

We examined the association between 25(OH)D and CRP using linear regression. First, an

unadjusted linear regression was conducted with CRP as the outcome variable and 25(OH)D as

the predictor variable in the entire population to explore trends in the data. Then, linear

regressions adjusted for age, waist circumference, physical activity, ethnicity, and season of

blood draw were carried out separately in the population as a whole, among men, women HC

non-users, and women HC users. In the analysis of the population as a whole, the model was also

adjusted for sex. We then conducted additional linear regressions to examine the association

between 25(OH)D and CRP among individuals below and above the median 25(OH)D

concentration, in order to determine whether the direction of the association differed between

those with a lower or a higher vitamin D status. In these analyses, Model 1 was unadjusted.

Model 2 was adjusted for age, sex, waist circumference, physical activity, ethnicity, and season.

Model 3 was adjusted for all the covariates included in Model 2, plus HC use among women.

Finally, within women HC users, we examined the association between 25(OH)D and CRP

among those taking <1mg (n=170) vs. ≥1mg (n= 63) total hormone per day, to determine

whether hormone dose played a role in this relationship. The 47 individuals for whom we were

unable to determine a daily hormonal dose (see Section 2.3.4) were excluded from this analysis.

38

We first compared circulating 25(OH)D and CRP concentrations between women HC users

taking <1mg/d vs. ≥1mg/d total hormone using ANCOVA adjusted for age, waist circumference,

physical activity, and season. We then conducted linear regressions adjusted for age, waist

circumference, physical activity, ethnicity, and season separately among women taking <1mg/d

and ≥1mg/d total hormone.

2.4 Results

Subject characteristics are shown in Table 2.1. We observed differences in HC use across ethnic

groups, with a greater percentage of Caucasian women reporting use of HC than East or South

Asian women (p<0.0001). Circulating 25(OH)D concentrations were approximately 25 nmol/L

higher in women HC users than in men and women HC non-users, despite a similar total vitamin

D intake across groups (p<0.0001 and p=0.26, respectively). Women HC users had the highest

circulating CRP, whereas women HC non-users had the lowest CRP concentrations (p<0.0001).

We examined differences in 25(OH)D concentrations between women HC users, women HC

non-users, and men across ethnic groups (Figure 2.1). Across all ethnic groups, women HC users

had significantly higher circulating 25(OH)D than men and women HC non-users after adjusting

for age, waist circumference, physical activity, and season of recruitment (p<0.0001 for

Caucasians and East Asians, and p=0.0079 for South Asians). Concentrations of 25(OH)D were

similar between men and women HC non-users across ethnic groups.

Differences in CRP concentrations among women HC users, women HC non-users, and men

across ethnic groups are shown in Figure 2.2. After statistical adjustment, women HC users had

higher CRP concentrations than men and women HC non-users across ethnic groups (p<0.0001).

Among Caucasians and South Asians, men had CRP concentrations that were intermediate

between women HC users and non-users. In East Asians, we observed no differences in CRP

concentrations between men and women HC non-users.

Table 2.2 shows the results of correlation analyses exploring the relationship between 25(OH)D

and total dietary vitamin D intake. Overall, circulating 25(OH)D and dietary vitamin D were

more strongly correlated among South Asians than Caucasians and East Asians (Pearson's r=0.55

39

[p<0.0001], 0.24 [p<0.0001], and 0.27 [p<0.0001], respectively; Fisher's z transformation

p<0.0001). Within Caucasians, circulating 25(OH)D and dietary vitamin D appeared to be more

strongly correlated among men and women HC non-users than among women HC users (r=0.28

[p<0.0001], 0.26 [p<0.0001], and 0.19 [p<0.006]). However, these correlation coefficients were

not significantly different from each other (Fisher's z transformation p=0.15). Within East

Asians, circulating 25(OH)D and dietary vitamin D were correlated only among men and women

HC non-users, with coefficients of approximately 0.30 (p<0.0001). In South Asians, correlation

coefficients between 25(OH)D and dietary vitamin D were different between men, women HC

non-users, and women HC users (Fisher's z transformation p=0.03), ranging from 0.48

(p<0.0001) among women HC non-users to nearly 0.70 among women HC users (p=0.0025).

Table 2.2 also shows the results of correlation analyses between 25(OH)D and CRP. In the

population as a whole, 25(OH)D and CRP were positively correlated (r=0.19, p<0.0001). When

examined within each ethnic group, 25(OH)D and CRP were positively correlated among

Caucasians (r=0.14, p=0.0001) and East Asians (r=0.17, p=0.0001), but not South Asians. When

men, women HC non-users, and women HC users were examined separately, the two metabolites

were weakly correlated in East Asian women HC non-users (r=0.13, p=0.0131), and not

correlated in any other sub-group.

The association between 25(OH)D and CRP was first explored using linear regression (Figure

2.3). The two metabolites were positively associated in an unadjusted model in the population as

a whole (p<0.0001). The positive association was retained in the adjusted model. However, when

men, women HC non-users, and women HC users were examined separately, we observed no

association between 25(OH)D and CRP among men and women HC non-users. Among women

HC users, 25(OH)D and CRP were positively associated in an adjusted model (p=0.002).

The population was then divided into two groups: those with 25(OH)D below or above the

median (51.9 nmol/L). Linear regressions were conducted separately within each group to

examine the association between 25(OH)D and CRP. These results are shown in Table 2.3.

Among those below the median 25(OH)D concentration, we observed no association between

25(OH)D and CRP. In the group above the median, we observed a positive association

(p<0.0001) between the two metabolites in both unadjusted and adjusted models (Model 1 and

40

2, respectively). However, the association was no longer significant after further adjusting for

HC use among women (Model 3).

Because we observed a positive association between 25(OH)D and CRP among women HC

users (Figure 2.3), we assessed whether total HC dose affected this association. After adjusting

for age, waist circumference, physical activity, ethnicity, and season, women who took <1mg/d

total hormone had lower 25(OH)D concentrations than those who took ≥1 mg/d (mean ±

standard error: <1 mg/d = 76.4 ± 2.7 nmol/L; ≥1 mg/d = 85.1 ± 4.5 nmol/L; p=0.0475), but

circulating CRP concentrations were similar across women HC users regardless of daily

hormone dose (<1 mg/d: 2.8 ± 0.3 mg/L, ≥1 mg/d: 2.8 ± 0.4 mg/L; p=0.8544). In separate

adjusted linear regression models of women taking <1 mg/d and women taking ≥ 1mg/d, we

observed no association between 25(OH)D and CRP in either group.

41

Table 2.1. Study participant characteristics 1-3

.

Men Women

HC non-users

Women

HC users p

n 428 695 280

Ethnicity

Caucasian 218 (31.0) 273 (38.9) 211 (30.1) <0.0001

East Asian 147 (27.3) 337 (62.9) 52 (9.7)

South Asian 63 (38.2) 85 (51.5) 17 (10.3)

Season of recruitment

Spring 111 (31.2) 170 (47.8) 75 (21.1) 0.0265

Summer 141 (34.8) 205 (50.6) 59 (14.6)

Autumn 106 (26.4) 201 (50.0) 95 (23.6)

Winter 70 (29.2) 119 (49.6) 51 (21.2)

Age (years) 22.9 ± 2.5a 22.4 ± 2.5

b 23.0 ± 2.4

a <0.0001

Waist circumference (cm) 80.3 ± 8.5a 70.9 ± 7.1

b 71.6 ± 7.1

b <0.0001

BMI (Kg/m2) 23.7 ± 3.4

a 22.3 ± 3.4

b 22.6 ± 3.1

b <0.0001

Systolic blood pressure (mm Hg) 123.9 ± 10.0a 108.1 ± 9.1

b 111.6 ± 9.4

c <0.0001

Diastolic blood pressure (mm Hg) 71.7 ± 7.6a 67.7 ± 8.0

b 69.7 ± 7.3

c <0.0001

Physical Activity (MET-hours/week) 7.8 ± 3.2a,b

7.5 ± 3.1a 8.0 ± 2.7

b 0.0213

Plasma 25(OH)D (nmol/L) 54.2 ± 26.0a 51.1 ± 24.2

a 78.1 ± 33.9

b <0.0001

Total vitamin D intake (IU/d) 341.0 ± 261.26 331.8 ± 250.9 363.8 ± 276.3 0.2583

Dietary vitamin D intake, excluding

supplements (IU/d) 258.0 ± 182.8 233.7 ± 174.2 235.5 ± 157.4 0.0561

42

Vitamin D supplement use 20 (4.7) 39 (5.6) 17 (6.1) 0.6880

Multivitamin supplement use 93 (21.7) 180 (25.9) 95 (33.9) 0.0010

CRP (mg/L) 1.1 ± 2.7a 0.8 ± 1.9

b 2.7 ± 3.3

c <0.0001

1 Shown are crude means ± standard deviations for continuous variables, and n (%) for

categorical variables.

2 p-values from ANOVA for continuous variables and χ

2 for categorical variables. P-values are

from tests using loge- or square root-transformed variables as necessary to improve normality,

but untransformed means and standard deviations are shown for ease of interpretation.

3 Different superscript letters indicate significant differences between groups (p<0.05). The

Tukey-Kramer procedure was used to adjust for multiple comparisons between groups within

each ANOVA.

43

Table 2.2. Correlation between plasma 25(OH)D and either dietary vitamin D or plasma CRP,

across ethnic groups, by sex and hormone use 1,2

.

25(OH)D - Dietary

vitamin D

25(OH)D -

CRP

n r p r p

Population overall

1,403 0.29 <0.0001 0.19 <0.0001

Caucasian All 702 0.24 <0.0001 0.14 0.0001

Men 218 0.28 <0.0001 -0.02 0.7243

Women - HC non-users 273 0.26 <0.0001 0.03 0.6782

Women - HC users 211 0.19 0.0060 0.06 0.3743

East Asian All 536 0.27 <0.0001 0.17 <0.0001

Men 147 0.32 <0.0001 -0.03 0.7414


Women - HC users 52 0.17 0.2194 0.25 0.0787

South Asian All 165 0.55 <0.0001 0.10 0.1858

Men 63 0.60 <0.0001 0.06 0.666


Women - HC users 17 0.68 0.0025 0.11 0.6688

1 Pearson's (r) correlation coefficients are shown. Within Caucasians, although circulating

25(OH)D and dietary vitamin D appeared to be more strongly correlated among men and women

HC non-users than women HC users, the correlation coefficients were not significantly different

from each other ((Fisher's z transformation p=0.15). Within East Asians, 25(OH)D and dietary

vitamin D were correlated among men and women HC non-users only. In South Asians,

25(OH)D and dietary vitamin D were more strongly correlated than in the other ethnic groups,

and were correlated across men, women HC non-users, and women HC users. In the population

44

as a whole, 25(OH)D and CRP were positively correlated. When examined within each ethnic

group, 25(OH)D and CRP were positively correlated among Caucasians and East Asians, but not

South Asians. When men, women HC non-users, and women HC users were examined

separately, the two metabolites were not significantly correlated in any ethnic group.

2 Variables were loge- or square root-transformed as needed prior to analysis.

45

Table 2.3. Association between 25(OH)D and CRP among individuals below and above the

median 25(OH)D concentration1,2

.

<51.9 nmol/L (n=701) ≥51.9 nmol/L (n=702)

β ± SE p β ± SE p

Model 1

-0.007 ± 0.009 0.8366 0.012 ± 0.004 <0.0001

Model 2

-0.004 ± 0.010 0.5155 0.012 ± 0.004 <0.0001

Model 3 -0.005 ± 0.010 0.6363 0.003 ± 0.004 0.2855

1 Shown are regression coefficients (β) ± standard errors (SE).

2 p-values from linear regression. CRP was loge-transformed prior to analysis to improve

normality. However, untransformed β coefficients and standard errors are shown for ease of

interpretation. Model 1 was unadjusted. Model 2 was adjusted for age, sex, waist circumference,

physical activity, ethnicity, and season. Model 3 was adjusted for the variables included in

Model 2, plus HC use among women. No association between 25(OH)D and CRP was seen

among individuals below the median 25(OH)D concentration. Among individuals above the

median, we observed a positive association between 25(OH)D and CRP in Models 1 and 2, but

the association was no longer significant after adjusting for HC use among women (Model 3).

46

Figure 2.1. Plasma 25(OH)D concentrations among men, women HC non-users, and women HC

users, by ethnicity.

Shown are crude means ± standard errors. P-values were obtained with ANCOVA. Within each

ethnic group, mean 25(OH)D concentrations were compared between men, women HC non-

users, and women HC users after adjusting for age, waist circumference, physical activity, and

season of recruitment. Plasma 25(OH)D was loge-transformed prior to analysis to improve

normality. However, untransformed means and standard errors are shown for ease of

interpretation. Within each ethnic group, different superscript letters indicate significant

differences between men, women HC non-users, and women HC users (p<0.05). The Tukey-

Kramer procedure was used to adjust for multiple comparisons between groups. Across all ethnic

groups, women HC users had significantly higher circulating 25(OH)D than men and women HC

non-users. Concentrations of 25(OH)D were similar between men and women HC non-users

across ethnic groups.

47

Figure 2.2. CRP concentrations among men, women HC non-users, and women HC users, by

ethnicity.

Shown are crude means ± standard errors. P-values were obtained with ANCOVA. Within each

ethnic group, mean CRP concentrations were compared between men, women HC non-users,

and women HC users after adjusting for age, waist circumference, physical activity, and season

of recruitment. Circulating CRP was loge-transformed prior to analysis to improve normality.

However, untransformed means and standard errors are shown for ease of interpretation. Within

each ethnic group, different superscript letters indicate significant differences between men,

women HC non-users, and women HC users (p<0.05). The Tukey-Kramer procedure was used to

adjust for multiple comparisons between groups. Women HC users had higher CRP

concentrations than men and women HC non-users across ethnic groups. Among Caucasians and

South Asians, men had CRP concentrations that were intermediate between women HC users

and non-users. In East Asians, we observed no differences in CRP concentrations between men

and women HC non-users.

48

Figure 2.3. Association between 25(OH)D and CRP.

First, unadjusted, untransformed 25(OH)D and CRP values were plotted on a single graph with

open circles representing men, open triangles representing women HC non-users, and crosses

representing women HC users. An unadjusted linear regression was conducted on the

untransformed values for the entire population to explore the association between 25(OH)D and

CRP (linear regression shown as dotted line). Then, linear regressions adjusted for age, waist

circumference, physical activity, ethnicity, and season, using loge-transformed CRP, were carried

out separately among men, women HC non-users and women HC users, as well as in the

population as a whole (in this case, the model was also adjusted for sex). The table insert shows

regression coefficients (β) ± standard errors and p-values for these adjusted linear regressions.

Untransformed β coefficients and standard errors are shown for ease of interpretation. In the

unadjusted model of the population as a whole, the two metabolites were positively associated.

49

The positive association was retained in an adjusted model. However, when men, women HC

non-users, and women HC users were examined separately, we observed no association between

25(OH)D and CRP among men and women HC non-users. Among women HC users, 25(OH)D

and CRP were positively associated in an adjusted model.

50

2.5 Discussion

The present study assessed the effect of HC use on circulating 25(OH)D across ethnic groups,

and examined whether HC use modifies the association between 25(OH)D and the inflammatory

biomarker CRP in an ethnically diverse population of young adults. Consistent with some

previous research (267), we observed a positive association between 25(OH)D and CRP. Women

HC users had the highest plasma concentrations of 25(OH)D, as well as CRP, suggesting that

HC use might have confounded these associations. Indeed, there was no longer an association

between 25(OH)D and CRP when they were examined separately in men and women HC non-

users. Within women HC users, a positive association between 25(OH)D and CRP was no longer

significant once we examined those who took <1 mg/d and ≥1 mg/d total hormone separately,

suggesting that hormone dose may have confounded the initial association.

Other studies have examined the effect of HRT on the association between vitamin D and

various cancers among post-menopausal women (190;245). However, to our knowledge, the

present study is the first to examine whether HC use confounds the association between

25(OH)D and inflammation. In the present study, a confounding effect of HC use was observed

in a mixed, ethnically diverse population consisting of both men and women. Women HC non-

users and men had similar 25(OH)D and CRP concentrations, whereas women HC users had

higher concentrations of both. These results highlight the importance of accounting for HC use in

studies of the relationship between vitamin D and inflammation-related health outcomes, if these

studies include subjects taking these medications.

Despite a wealth of evidence from in vitro and animal studies suggesting an important role for

vitamin D in modulating inflammation, epidemiologic and clinical studies assessing the effects

of vitamin D on inflammation have yielded inconsistent results (31;85;255;258;261;263;267).

Vitamin D is thought to exert its anti-inflammatory actions by binding to the VDR in target

tissues. Indeed, most immune cell types express the VDR (315), which is a transcription factor

that, upon binding to 1,25(OH)2D, the biologically active vitamin D metabolite, regulates the

transcription of hundreds of genes, including many that are involved in innate immunity (27).

While a number of biological, demographic and lifestyle factors may help explain the

unexpected positive association between 25(OH)D and inflammation in some human studies, HC

use has often been overlooked as a potentially important confounder. Use of HC, which most

51

often consist of a combination of synthetic estrogen and progestins (192), has been associated

with elevated 25(OH)D (188). The potential effects of progesterone-derived HC ingredients on

vitamin D metabolism are not well understood. However, estrogen may upregulate CYP27B1,

the enzyme that converts 25(OH)D to 1,25(OH)2D, as well as the VDR and DBP, which is the

main vitamin D metabolite carrier in the circulation (190;195;316). Estrogen also appears to

downregulate CYP24A1, the main catabolic enzyme in the vitamin D pathway (190;193).

Together, these actions may result in elevated circulating vitamin D metabolites, although the

clinical significance of this effect remains unknown. HC use has also been associated with

changes in numerous physiologic pathways, increased inflammation and a poorer metabolic

profile (198;199;203;206;207). Our finding that adjustment for HC use attenuates a positive

association between 25(OH)D and CRP might partly account for some of the unexpected results

of previous studies.

Recently, a cross-sectional analysis of NHANES data reported a positive association between

25(OH)D and CRP among those with 25(OH)D ≥ 52.5 nmol/L (267). Conversely, the

relationship between 25(OH)D and CRP was inverse among those with lower vitamin D status

(267). In the present study, we found no association between 25(OH)D and CRP in those below

the median 25(OH)D concentration, and an initially positive association between 25(OH)D and

CRP in those above the median was no longer significant after adjustment for HC use. It is

possible that the previously observed (267) positive association between 25(OH)D and CRP

among individuals with a higher vitamin D status may be due to unaccounted-for HC use among

women.

The present study also aimed to assess the effect of HC use on 25(OH)D concentrations across

several ethnic groups. To our knowledge, few studies have examined the effect of HC use on

vitamin D status in non-Caucasian individuals, and those that did focused on individuals of

African ancestry only (189;191). In the present study, we observed that, across all ethnic groups,

women HC users had significantly higher circulating 25(OH)D than women HC non-users and

men, even after adjusting for factors known to influence vitamin D concentrations. These

observations corroborate previous findings that HC use elevates circulating 25(OH)D

(188;191;317). Furthermore, our results suggest that the effect is consistent across ethnic groups.

The magnitude of the effect of HC use on 25(OH)D concentrations across ethnic groups was

similar to that observed in previous studies in Caucasians, with women HC users having values

52

over 20 nmol/L higher than women HC non-users, who had 25(OH)D concentrations similar to

men. Our observation that 25(OH)D concentrations were higher in women taking ≥1 mg/d total

hormone than those taking <1 mg/d suggests that the effect of HC on 25(OH)D may be dose-

dependent.

We observed that the association between 25(OH)D and dietary vitamin D differed across ethnic

groups. Vitamin D can be obtained exogenously from the diet and supplements, or produced

endogenously in the skin after exposure to sunlight (121). Dark-skinned individuals are less able

to produce vitamin D endogenously because of a higher skin melanin content, and exogenous

sources may contribute more to vitamin D status in these individuals (153). In the present study,

while dietary vitamin D contributed to 25(OH)D concentrations in all ethnic groups, the

correlation was weaker among Caucasians and East Asians than among South Asians. The

stronger correlation between 25(OH)D and dietary vitamin D observed among South Asians may

partly reflect a decreased ability to produce vitamin D endogenously because of generally darker

skin, so that dietary vitamin D intake becomes a more important source of 25(OH)D. However,

no information was available on lifestyle variables that may affect sun exposure, such as use of

long-sleeved clothing or sunscreen. Therefore, we cannot determine whether other lifestyle

factors specific to these individuals also contributed to decreased endogenous vitamin D

production.

We noted an initial positive association between 25(OH)D and CRP among women HC users.

However, this association was no longer significant after we examined HC users who took <1

mg/d and ≥1 mg/d total hormone separately. As noted earlier, we observed a dose-dependent

association between HC dose and 25(OH)D among women HC users, which might partly explain

the initial positive association between 25(OH)D and CRP observed in this group. However,

CRP concentrations were equally elevated in women taking <1 mg/d vs. ≥1 mg/d total hormone.

It is possible that the initial positive association observed between 25(OH)D and CRP among

women HC users as a whole was due to residual confounding.

One limitation of the present study is the smaller sample size of women HC users across ethnic

groups, particularly South Asians. Another limitation is the lack of information on menstrual

cycle stage and circulating sex hormone concentrations, which prevented us from investigating

the potential role of natural fluctuations in estrogen and progesterone in modulating the

53

association between 25(OH)D and CRP. We were also unable to explore whether estrogen or

progestin-derived HC components have a greater effect on the observed associations.

In conclusion, in the present study we identified a confounding effect of HC use on the

association between 25(OH)D and CRP in an ethnically diverse population of young adults. An

apparent positive association between 25(OH)D and CRP was no longer significant after

adjusting for HC use. In addition, we noted that HC use increased 25(OH)D concentrations

significantly across ethnic groups. Overall, our results bring to light the importance of

considering lifestyle variables, particularly HC use, as significant confounders in the relationship

between vitamin D and inflammation, as well as, potentially, other health outcomes.

54

Chapter 3 : Association between circulating 25-hydroxyvitamin D and

plasma cytokine concentrations in young adults

Adapted from:

García-Bailo B, Roke K, Mutch DM, El-Sohemy A, Badawi A. Association

between circulating ascorbic acid, α-tocopherol, 25-hydroxyvitamin D, and plasma

cytokine concentrations in young adults: a cross-sectional study. Nutr Metab

(Lond). 2012 Nov 16;9(1):102. doi: 10.1186/1743-7075-9-102.

55

3.1 Abstract

Background: Chronic inflammation is thought to be implicated in the etiology of

cardiometabolic disease. Vitamin D can be synthesized locally by immune system cells and is

thought to modulate inflammation, potentially helping prevent or delay disease development.

However, the relationship between vitamin D and inflammation at the systemic level is

inconsistent, and few studies have examined this relationship in young adults, who may be in the

earliest stages of disease development. Our objective was to examine the association between

circulating 25-hydroxyvitamin D (25(OH)D), the biomarker of vitamin D status, and five

cytokines in an ethnically diverse population of young adults.

Methods: Participants (n=1,007) from the Toronto Nutrigenomics and Health study provided

fasting blood samples for biomarker measurements and were categorized into tertiles based on

their plasma 25(OH)D concentrations. We conducted Pearson's correlation analyses to explore

trends in the relationships between 25(OH)D and each cytokine. We further examined the

association between 25(OH)D and the cytokines using analysis of covariance with models

including age, sex, waist circumference, ethnicity, physical activity, season of blood collection,

and hormonal contraceptive use (women only) as covariates.

Results: We observed correlations between 25(OH)D, IP-10 (r=0.12, p<0.0001), and RANTES

(r=0.22, p<0.0001), but no other cytokines. We noticed strong correlations between specific

cytokines, such as interleukin 1- receptor antagonist (IL-1RA) and interferon-γ (IFN-γ) (r=0.83,

p<0.0001), and IL-1RA and platelet-derived growth factor BB (PDGF-bb) (r=0.80, p<0.0001).

After full covariate adjustment, 25(OH)D was not associated with any cytokine.

Conclusion: 25(OH)D was not associated with systemic biomarkers of inflammation in healthy

young adults.

56

3.2 Introduction

Inflammation plays an important role in the development of cardiometabolic diseases, such as

the metabolic syndrome, T2D and CVD (56;318). The excessive release of pro-inflammatory

cytokines that takes place during chronic inflammation can result in dysregulation of processes

such as glucose and lipid metabolism and vascular function, via the effects of cytokines on

adipocytes, muscle tissue, the liver and blood vessels (60;93;319;320). In addition, chronic

inflammation is closely associated with oxidative stress, which is characterized by an elevated

presence of highly reactive, potentially harmful molecular compounds, such as reactive oxygen

and nitrogen species and free radicals (319;321). These reactive molecules contribute to further

oxidative damage and inflammation through their ability to activate the pro-inflammatory

transcription factor NFқB (322).

Accumulating evidence suggests that, beyond its central function as a regulator of calcium and

bone metabolism, vitamin D may play a role in the prevention of cardiometabolic diseases

(19;20). The beneficial effects of this micronutrient may result, at least in part, from its role in

modulation of innate immunity and inflammation (28). Indeed, circulating concentrations of

25(OH)D, which is considered the biomarker of vitamin D nutritional status, have been inversely

associated with biomarkers of systemic inflammation (31;85). However, the results are equivocal

(261;263;267) and few studies have examined this relationship in young, healthy individuals.

Therefore, the role of vitamin D in the earliest stages of chronic inflammation remains poorly

understood.

Traditional enzyme-linked immunoassay-based methods are restricted to measuring single

inflammatory biomarkers, such as CRP, IL-6, and TNFα (323;324). However, the development

of multiplex assays has made it possible to measure numerous cytokines simultaneously in order

to obtain a more comprehensive picture of an individual’s inflammatory status. In the present

study, we analyzed five common cytokines using the multiplex approach. The cytokines included

were IL-1-receptor antagonist (IL-1RA), IFN-γ, interferon γ-induced protein 10 (IP-10), platelet-

derived growth factor BB (PDGF-bb) and RANTES. Examining the association between

25(OH)D and cytokines involved in inflammatory pathways may contribute to our understanding

of the role that this micronutrient plays in modulating inflammation during the earliest stages of

disease progression. Therefore, the objective of this study was to examine the relationship

57

between circulating concentrations of 25(OH)D and biomarkers of inflammation in an ethnically

diverse population of young adults.

3.3 Methods


The present study included 1,007 non-smoking individuals (300 men and 707 women) from the

TNH study population who were free of diabetes and who had available data for 25(OH)D, the

five cytokines, and every other variable considered for analysis at the time that the study was

carried out. For more details on the TNH study population, recruitment, exclusions, and

classification into ethnic groups and seasons, please refer to Chapter 2, Section 2.3.1.

3.3.2 Anthropometrics and physical activity

Please refer to Chapter 2, Section 2.3.2.


For details on fasting blood collection and how 25(OH)D was measured, please refer to Chapter

2, Section 2.3.5.

A custom multiplex bead assay was designed to measure the cytokines assessed in this study

using a Bio-Plex-200 instrument (Bio-Rad, Mississauga, ON, Canada). As a first step, we used a

commercially available kit to examine 27 cytokines in a subset of 70 individuals in order to

determine which inflammatory biomarkers could be consistently measured in our study

population. Since the population consisted of predominantly healthy young adults, it was

expected that some cytokines would be below the detection limit of our analytical platform. To

avoid drawing erroneous conclusions, we made a decision to examine only the five cytokines

that were detected in all 70 of these subjects and had concentrations in the working range of the

assay, specified in the manufacturer's prospectus for each cytokine and defined as concentrations

between the lower and upper limits of quantification. The five cytokines measured included IL-

1RA, IFN-γ, IP-10, PDGF-bb and RANTES. IL-1RA binds to the IL-1 receptor and, therefore,

inhibits the inflammatory actions of both IL-1α and IL-1β (325), while IFN-γ is a pro-

58

inflammatory cytokine with important immunomodulatory properties that induces the production

of IP-10 (326;327). PDGF-bb is a growth factor involved in angiogenesis (328). RANTES, a

pro-inflammatory cytokine, also plays a role in angiogenesis (329;330).

To measure the five cytokines in the study population, plasma samples (30 μL/sample) were

diluted 1:4 with sample diluent, and the assay was run according to the manufacturer’s

instructions. Once the procedure was complete, beads were read using the Bio-Plex suspension

array system (Bio-Rad) and concentrations (pg/ml) were determined with Bio-Plex Manager

software (version 6.0). Analytical reproducibility was assessed by calculating intra-assay

coefficients of variability (CV) (as the average of three standards within each analytical run), and

resulted in the following CVs: IL-1RA <8%; IFN-γ <11%; IP-10 <9%; PDGF-bb <9%;

RANTES <6%. Additionally, inter-assay CVs for all five cytokines, calculated by measuring the

average of three standards across fifteen assays run on different days, were the following: IL-

1RA <4%; IFN-γ <5%; IP-10 <4%; PDGF-bb <5%; RANTES <3%.


All statistical analyses were performed in SAS. The error was set at 0.05, and reported p-

values are 2-sided. The Bonferroni correction for multiple comparisons (significant p=/number

of independent tests conducted) was used to determine significance, unless indicated otherwise.

The distributions of continuous variables were examined and loge- or square root-transformed as

necessary to improve normality. However, untransformed, unadjusted means and measures of

spread are reported throughout to facilitate interpretability. Subject characteristics were

compared between men and women using t-tests for continuous variables and χ2 tests for

categorical variables. Pearson crude correlation coefficients were calculated between 25(OH)D

and each of the cytokines, as well as across all cytokines with each other, using loge- or square

root-transformed variables as necessary to improve normality. The association between each

cytokine and 25(OH)D was further examined with ANCOVA. Circulating concentrations of

plasma 25(OH)D were categorized into tertiles, and these categorical variables were used as

predictor variables. In each case, three models were constructed. Model 1 was unadjusted. Model

2 included age, sex, waist circumference, ethnicity, physical activity, and season as covariates.

Model 3 included the variables considered in Model 2, plus current HC use among women. This

approach allowed us to examine, in addition to variables known to affect 25(OH)D

59

concentrations such as ethnicity, the specific contribution of HC, which may play an important

role as confounders of the relationship between 25(OH)D and systemic inflammation (331).

3.4 Results

Subject characteristics are shown in Table 3.1. Women had lower anthropometric values but

higher plasma 25(OH)D concentrations than men. Concentrations of IFN-γ and RANTES were

lower in men than women, but we observed no differences between the sexes with respect to IL-

1RA, IP-10, or PDGF-bb concentrations.

Pearson's correlation analyses between 25(OH)D and all the cytokines were conducted to explore

trends in the data, and the results are shown in Figure 3.1. Circulating 25OH)D was weakly

positively correlated with IP-10 (r=0.12) and RANTES (r=0.22), but it was not correlated with

any of the other cytokines. The strongest correlation (r=0.83) was between IL-1RA and IFN-γ.

IL-1RA was also correlated with PDGF-bb (r=0.80), RANTES (r=0.38), and IP-10 (r=0.13).

IFN-γ was correlated with PDGF-bb (r=0.67), IP-10 (r=0.23), and RANTES (r=0.18). IP-10 was

correlated with RANTES (r=0.28) and PDGF-bb (r=0.10). Finally, PDGF-bb was correlated

with RANTES (r=0.38).

The associations between each cytokine and 25(OH)D, after adjusting for relevant covariates, are

shown in Table 3.2. We observed a U-shaped association between IP-10 and 25(OH)D

(p=0.0032) in an unadjusted model; however, the association was no longer significant in either

of the adjusted models. A positive association (p≤0.0001) was also present between RANTES

and 25(OH)D in Model 1 (unadjusted) and Model 2 (adjusted for age, sex, waist circumference,

ethnicity, physical activity and season). However, the two variables were no longer associated

after adjusting for HC use.

60

Table 3.1. Study participant characteristics1.

Sex

Male

(n=300)

Female

(n=707) p†

Age (years) 22.7 ± 2.5 22.6 ± 2.4 0.4363

Ethnicity

Caucasian 137 (28.72) 340 (71.28) 0.0039

East Asian 94 (26.86) 256 (73.14)

South Asian 48 (44.86) 59 (55.14)

Other 21 (28.77) 52 (71.23)

HC use among women

221 (31.26)

BMI (kg/m2) 23.6 ± 3.5 22.4 ± 3.5 <0.0001

Waist circumference (cm) 80.1 ± 8.9 71.2 ± 7.6 <0.0001

Physical Activity (Met-h/wk) 7.7 ± 3.2 7.6 ± 3.0 0.5423

Plasma 25(OH)D (nmol/L) 46.8 ± 20.1 54.5 ± 26.5 <0.0001

IL-1RA (pg/mL) 320.9 ± 206.9 322.8 ± 180.2 0.503

IFN-γ (pg/mL) 185 ± 132.7 201.3 ± 134.7 0.0275

IP-10 (pg/mL) 558.2 ± 290.7 613.1 ± 478.4 0.3515

PDGF-bb (pg/mL) 1316.7 ± 2642.2 1152.8 ± 1729 0.5038

RANTES (pg/mL) 2378.7 ± 3176.2 2465.7 ± 1350.6 0.0023

1: Shown are crude, untransformed means ± standard deviations or n (%)

†: p-value from χ2 analysis for categorical variables and t-test for continuous variables, using

loge- or square-root transformed values as necessary.

61

Table 3.2. Mean cytokine concentrations by plasma 25(OH)D tertiles1.

Plasma 25(OH)D (nmol/L) p†

Biomarker <38.3

(n=335)

38.3 - 59.3

(n=336)

59.4 - 194.0

(n=336) Model 1 Model 2 Model 3

IL-1RA (pg/mL) 324.2 ± 10.0 315.2 ± 10.4 327.2 ± 10.5 0.6341 0.6983 0.5515

IFN-γ (pg/mL) 203.7 ± 8.0 187.2 ± 6.2 198.6 ± 7.7 0.7070 0.5307 0.2369

IP-10 (pg/mL) 593.6 ± 28.8 563.0 ± 18.1 633.6 ± 22.6 0.0032* 0.0862 0.4425

PDGF-bb (pg/mL) 1050.2 ± 81.9 1260.6 ± 148.1 1293.7 ± 92.9 0.0134 0.2336 0.9286

RANTES (pg/mL) 2174.9 ± 63.0 2427.7 ± 167.1 2716.0 ± 77.2 <0.0001* 0.0001* 0.0550

1Shown are crude, untransformed means ± standard errors.

*Significantly different concentrations across tertiles at the Bonferroni level (p<0.0033, where α=0.05 and number of independent

tests=15 [5 cytokines x 3 models]).

†p-value from ANCOVA with loge- or square root-transformed biomarker concentrations where necessary.

Model 1: Unadjusted.

Model 2: Adjusted for age, sex, waist circumference, physical activity, ethnicity, and season.

Model 3: Adjusted for model 2 plus HC use among women.

62

Figure 3.1. Correlation analysis between circulating 25(OH)D and cytokines.

Shown are Pearson's (r) coefficients and p-values only for variables that were significantly

correlated at the Bonferroni level for 36 independent tests (p<0.0014, where α=0.05 and number

of independent tests=36 [6 variables correlated with each other]). Variables were loge- or square

root-transformed as necessary prior to analysis. Plasma 25(OH)D was only weakly correlated

with IP-10 and RANTES. By contrast, several of the cytokines were strongly or moderately

correlated.

63

3.5 Discussion

The present study examined the association between circulating concentrations of 25(OH)D and

five cytokines in healthy young adults. We observed that 25(OH)D was not associated with

concentrations of inflammatory biomarkers in this population after accounting for potentially

confounding factors, including age, sex, waist circumference, physical activity, ethnicity, season

of blood collection, and HC use among women.

A wealth of research using cell lines and animal models suggests a key role for vitamin D in

modulating immune responses (144). The actions of vitamin D result from binding of

1,25(OH)2D, the biologically active vitamin D metabolite, to the VDR, a transcription factor

expressed in tissues throughout the body (121). The ligand-bound VDR then binds to VDRE in

target genes, regulating their transcription (143). Most immune cell types express the VDR, as

well as CYP27B1, which is the main enzyme responsible for conversion of 25(OH)D into

1,25(OH)2D. Thus, immune cells are not only responsive to 1α,25(OH)2D but can synthesize it

locally from circulating 25(OH)D (332). In experimental models, vitamin D modulates

inflammatory processes by downregulating the production of pro-inflammatory cytokines, such

as IFN-γ and IL-2, upregulating the transcription of immuno-supressive cytokines, such as IL-10,

and modulating pro-inflammatory intracellular signalling pathways (28).

Despite the strong mechanistic evidence for an immuno-modulatory effect of vitamin D, human

studies have yielded inconsistent results on the association between 25(OH)D and inflammatory

biomarkers (34;257;267;333). These inconsistencies might reflect differences in disease status

between the populations studied, as well as unaccounted confounding from various biological,

demographic, and lifestyle factors. In particular, the use of HC is often overlooked as an

important confounder in the relationship between vitamin D and inflammation. Indeed, we have

recently reported a confounding effect of HC use on the association between 25(OH)D and CRP

in a population of young adults (331). An initial positive association between 25(OH)D and CRP

was no longer significant after use of these medications was considered.

In the present study, we noted a positive association between RANTES and 25(OH)D. RANTES

is a pro-inflammatory cytokine that plays an important role in recruitment of leukocytes to sites

of inflammation and in mediating T cell and monocyte traffic (334). Given the anti-inflammatory

properties of vitamin D, this initial association was unexpected because elevated RANTES

64

concentrations are observed in several inflammatory conditions, including atherosclerosis

(335;336). However, the association was no longer significant after use of HC among women

was taken into account. These medications, which commonly consist of synthetic estrogen and

progestins (192), have been linked to elevated circulating vitamin D metabolites (188). These

effects are thought to be mediated through the actions of estrogen on upregulation of CYP27B1,

the VDR, and DBP (190;195;316), as well as downregulation of CYP24A1, the main catabolic

enzyme in the vitamin D pathway (190;193). At the same time, ethinyl estradiol, the main

estrogenic component of many HC, affects hepatic production of hemostasis-related proteins and

pro-inflammatory acute phase reactants, such as CRP (206;337). Some evidence indicates that

CRP upregulates the production of RANTES (338). Therefore, the observed confounding effect

of HC on the association between 25(OH)D and RANTES may be partly a result of the effects of

these medications on CRP. Alternately, use of HC could affect RANTES directly. Indeed, we

found that RANTES concentrations, as well as those of the other cytokines, were higher among

women who used these medications than those who did not (data not shown).

We observed no associations between 25(OH)D and any of the other cytokines, except for an

initial U-shaped association with IP-10 that was no longer significant after covariate adjustment.

It is important to note that our study population consisted of healthy young adults. It is possible

that the anti-inflammatory effects of vitamin D may be more apparent in older or more diseased

populations with a greater burden of systemic inflammation that can be more easily measured

through biomarkers circulating in the blood. Indeed, several studies that have reported beneficial

effects of vitamin D on inflammation were conducted in obese individuals or those with T2D

(31;85). Furthermore, it is also possible that 25(OH)D concentrations higher than those found

among our study participants are necessary to observe its effects on immune modulation at the

systemic level.

Finally, while we found weak correlations, and no covariate-adjusted associations between

25(OH) and any cytokine, we observed some strong positive correlations between certain

cytokines, namely IL-1RA and IFN-γ, IL-1RA and PDGF-bb, and IFN-γ and PDGF-bb. These

observations are in agreement with previous in vitro work reporting that IFN-γ up-regulates

production of both IL-1RA and PDGF-bb (339;340).

65

One limitation of the present study is the cross-sectional nature of the analyses, which precludes

making any inferences about causality. In addition, of the 27 cytokines initially measured in a

subset of the population, the concentrations of many were too low to detect accurately, which

prevented us from examining other important inflammatory biomarkers such as IL-6 and TNFα.

While the young age and general good health of the present study population may have

accounted for this shortcoming, studying healthy individuals provides a picture of inflammatory

status before the onset of disease, and understanding how inflammation may be modulated by

various micronutrients at this stage provides the potential for developing nutrition-based

strategies to prevent later disease development.

In conclusion, the present study examined the association between circulating concentrations of

25(OH)D, and inflammatory biomarkers in healthy young adults living in Canada. We observed

that circulating 25(OH)D was not associated with decreased concentrations of any inflammatory

biomarkers. In addition, we identified a confounding effect of HC use on the association between

25(OH)D and RANTES.

66

Chapter 4 : Plasma 25-hydroxyvitamin D, hormonal contraceptive use,

and cardiometabolic disease risk in an ethnically diverse population of

young adults

Adapted with permission from an article published by the Journal of the American

College of Nutrition, which is available online at

http://www.tandfonline.com/loi/uacn20#.Uht-Oz8piRM. The original article is the

following:

García-Bailo B, Karmali M, Badawi A, El-Sohemy A. Plasma 25-hydroxyvitamin

D, hormonal contraceptive use, and cardiometabolic disease risk in an ethnically

diverse population of young adults. J Am Coll Nutr. In press.

67

4.1 Abstract

Background: The relationship between vitamin D and cardiometabolic disease risk across ethnic

groups is unclear, and it is not known whether use of HC, which affect vitamin D metabolism

and are also associated with cardiometabolic disease risk, modify this relationship. Our

objectives were to determine the prevalence of vitamin D deficiency (plasma 25(OH)D <30

nmol/L), to assess seasonal variation in concentrations of 25(OH)D, and to examine whether

25(OH)D is associated with cardiometabolic biomarkers across ethnic groups and across men,

women non-users of hormonal contraceptives (HC), and women HC users in an ethnically

diverse population of young adults living in Canada.

Methods: The study population consisted of Caucasian, East Asian and South Asian individuals

(n=1,384, 69% female) aged 20 – 29 years. Participants provided overnight fasting blood

samples, from which plasma 25(OH)D and cardiometabolic biomarkers were measured. Vitamin

D status distributions were compared using χ2

tests, and ANCOVA was used to examine seasonal

variations in 25(OH)D, as well as the association between 25(OH)D and cardiometabolic

biomarkers, across groups.

Results: Plasma 25(OH)D concentrations fluctuated seasonally among Caucasians and East

Asians and across men, women HC non-users, and women HC users, but they remained low

year-round in South Asians, half of whom were vitamin D-deficient. Vitamin D deficiency was

associated with higher insulin, HOMA-IR, and HOMA-Beta among Caucasians and East Asians

and among men and women HC non-users, and with higher triglycerides among men only. No

biomarkers were associated with 25(OH)D among South Asians and women HC users, although

non-significant trends were observed for higher markers of glycemic dysregulation in those who

were vitamin D-deficient in both groups.

Conclusions: Vitamin D deficiency varies between ethnic groups, being particularly high among

South Asians, and it is associated with biomarkers of glycemic dysregulation; however, HC use

among women may attenuate this association. Given the widespread use of HC by women

throughout the world, a better understanding of the extent to which these medications may

modify the relationship between vitamin D and processes related to disease is warranted.

68

4.2 Introduction

Cardiometabolic diseases, such as T2D and CVD, are a leading cause of death and disability

worldwide (1). Metabolic abnormalities including hypertension, dyslipidemia, central obesity

and glycemic dysregulation, which often cluster together as the metabolic syndrome, are

considered risk factors for cardiometabolic disease (341). Evidence from in vitro studies and

recent human clinical trials suggests that vitamin D may affect processes involved in

cardiometabolic disease, such as insulin sensitivity and pancreatic β cell function (26;32;33).

Adequate circulating concentrations of 25(OH)D, the vitamin D metabolite used to measure

status, which captures both dietary intake and endogenous synthesis of vitamin D, have been

associated with improved levels of biomarkers of cardiometabolic disease and decreased risk of

the metabolic syndrome (29;30). However, recent systematic reviews have concluded that the

results of the various studies are equivocal (35-37). In addition to differences in methodology,

the widespread inconsistencies between studies may result partly from unaccounted confounding

or genetic differences (38).

While vitamin D can be obtained through the diet, the primary source is endogenous cutaneous

production after exposure to UV radiation (121). Thus, a seasonal lack of sunlight and darker

skin increase the risk of vitamin D deficiency.The prevalence of vitamin D deficiency, as defined

by the Institute of Medicine, in Canada has been reported to range from 3% among Caucasians to

16% among non-Caucasians (152), but only a few small studies have examined the differences

between specific ethnic groups, such as East Asians and South Asians, within Canada (153;178).

Disparities in cardiometabolic disease prevalence have been documented across ethnic groups

(63;80), but few studies have assessed whether these ethnic differences in cardiometabolic

disease are associated with differences in vitamin D status (182;342). Elucidating the

relationship between vitamin D and cardiometabolic disease across ethnic groups may partly

explain the inconsistencies between studies.

In addition to skin colour, other demographic, lifestyle, and biological factors can affect

25(OH)D concentrations and may confound the relationship between this vitamin and various

health outcomes (223;317). In particular, the use of estrogen-containing medications, such as

HC, among women is associated with elevated circulating 25(OH)D (188;191). At the same

time, HC, which are used by over 100 million women worldwide, have been linked to

69

widespread effects on disease-associated pathways and a higher risk of adverse health outcomes,

such as impaired glucose tolerance and CVD (204;207;208). Taken together, this evidence

suggests that HC use may confound the association between 25(OH)D and disease-associated

pathways. Indeed, we have recently characterized confounding effects of HC use on the

association between 25(OH)D and CRP, a biomarker of systemic inflammation, as well as

numerous disease-associated plasma proteins, in young women across ethnic groups (331;343).

However, it is not known whether the association between vitamin D and other biomarkers of

cardiometabolic disease is different in women HC users than men and women HC non-users.

Given the widespread use of these medications by women throughout the world, it is important

to characterize their potential effects on the relationship between vitamin D and cardiometabolic

disease.

The objectives of the present study were to examine whether 25(OH)D is associated with

cardiometabolic biomarkers across ethnic groups and across men, women HC non-users, and

women HC users in an ethnically diverse population of young adults living in Canada. We also

determined the prevalence of vitamin D deficiency and assessed seasonal variation in 25(OH)D

concentrations across groups.

4.3 Methods


The present study included 1,384 non-smoking men (n=427) and women (n=957), free of

diabetes, with available data for 25(OH)D and the remaining variables considered for analysis.

We excluded individuals who were self-reported as Afro-Caribbean, Aboriginal, or of mixed

descent (n=118) because of the low number of persons recruited from each of these groups. For

more details on the TNH study population, recruitment, exclusions, and classification into ethnic

groups and seasons, please refer to Chapter 2, Section 2.3.1.



70





2, Section 2.3.5.

All measures of glucose and lipid metabolism were performed on-site at LifeLabs Medical

Laboratory Services using standard procedures, as previously described (311). Insulin resistance

and % β cell function were assessed through HOMA, with values calculated using the HOMA2

Calculator version 2.2 (Diabetes Trials Unit, University of Oxford, Oxford, United Kingdom)

(103).

Vitamin D status categories were created based on 25(OH)D concentrations and in consideration

of the recommendations of the Institute of Medicine (151), the Endocrine Society (23), and the

Canadian Osteoporosis Society (24). Deficiency was defined as 25(OH)D <30 nmol/L.

Insufficiency was defined as 30 – 49.9 nmol/L. Adequacy was equivalent to 50 – 74.9 nmol/L,

and optimal vitamin D status was defined as ≥75 nmol/L.


All statistical analyses were performed using SAS. The error was set at 0.05 and reported p-

values are 2-sided. Where indicated, the Bonferroni correction for multiple testing was applied.

Continuous variables with non-normal distributions were loge- or square root-transformed prior

to analysis. In these instances, the p-values from analyses using the transformed values are

reported, but untransformed means and measures of spread are shown for ease of interpretation.

Subject characteristics were compared using t-tests, ANOVA, and χ2 tests. Vitamin D status

distributions across men, women HC non-users, and women HC users, as well as across ethnic

groups, were compared using χ2 tests. Concentrations of 25(OH)D across ethnic groups and

across men, women HC non-users, and women HC users were compared between seasons using

ANCOVA. When comparing ethnic groups, the analyses were adjusted for age, sex, waist

circumference, physical activity, and HC use among women. When comparing men, women HC

non-users, and women HC users, the analyses were adjusted for age, waist circumference,

71

physical activity, and ethnicity. Finally, the association between specific cardiometabolic

biomarkers and vitamin D status across ethnic groups and across men, women HC non-users, and

women HC users was explored using ANCOVA. To examine the association across ethnic

groups, we first conducted exploratory analyses in models including an interaction term for

vitamin D status x ethnicity, as well as age, sex, waist circumference, physical activity, season,

and HC use among women as covariates. We also conducted analyses stratified by ethnicity

using ANCOVA models adjusted for age, sex, waist circumference, physical activity, season,

and HC use among women. To examine the association across men, women HC non-users, and

women HC users, as previously, we first conducted exploratory analyses in models including an

interaction term for vitamin D status x HC use, as well as age, sex, waist circumference, physical

activity, season, and ethnicity as covariates. Subsequently, analyses were stratified by sex and

HC use and adjusted for age, waist circumference, ethnicity, physical activity, and season.

Across all ethnicities and sex groups, analyses examining the association between HOMA-Beta

and vitamin D status were also adjusted for the ratio of circulating triglycerides to HDL

cholesterol, as a surrogate measure of insulin sensitivity (104;344).

4.4 Results

The characteristics of the study population are shown in Table 4.1. Men had higher values of

most anthropometric measures than women as a whole, but women had higher plasma 25(OH)D,

insulin, HOMA-IR, HOMA-Beta, total and HDL cholesterol than men. Nearly 45% of Caucasian

women reported HC use, whereas approximately 15% of East Asian and South Asian women

were HC users. Plasma 25(OH)D was highest in Caucasians and lowest in South Asians.

Biomarkers of glycemic dysregulation, LDL cholesterol, and total:HDL ratio were higher among

South Asians than the other groups, whereas HDL cholesterol was lowest in this group.

The distribution of vitamin D status differed across men, women HC non-users, and women HC

users, and it also differed across ethnocultural groups (Table 4.2). Half of women HC users were

in the optimal vitamin D status category (25(OH)D ≥75 nmol/L), while only 20% of men and

14% of women HC non-users had an optimal vitamin D status. Only about 5% of Caucasians

were vitamin D deficient, whereas 17% of East Asians and 45% of South Asians were

categorized as deficient. In contrast, nearly 40% of Caucasians had an optimal vitamin D status,

72

but less than 10% of the individuals from the other two ethnic groups fell in the optimal status

category.

Mean 25(OH)D concentrations fluctuated seasonally among Caucasians and East Asians, but not

South Asians (Figure 4.1). Caucasians and East Asians recruited during summer and fall had

higher average concentrations than those recruited during winter and spring. Among South

Asians, 25(OH)D concentrations did not fluctuate seasonally. Across seasons, Caucasians had

the highest 25(OH)D concentrations. Mean 25(OH)D concentrations also fluctuated seasonally

across men, women HC non-users, and women HC users (Figure 4.2), with concentrations being

higher for those recruited in summer and fall than winter and spring. Regardless of the season,

women HC users had the highest plasma 25(OH)D.

The association between cardiometabolic risk biomarkers and vitamin D status was examined

separately within each ethnic group, as shown in Table 4.3. When examined as a whole, there

were no 25(OH)D x ethnicity interaction effects in the overall population. When examined

separately, among Caucasians, individuals with optimal vitamin D status had the lowest mean

fasting insulin, HOMA-IR, and HOMA-Beta scores. Among East Asians, fasting insulin and

HOMA-IR were highest in vitamin D-deficient individuals. In this ethnic group, HOMA-Beta

was higher in those who were vitamin D-deficient than in those who had an adequate or optimal

vitamin D status. There were no significant associations between any biomarkers and vitamin D

status among South Asians, although we observed non-significant trends for lower fasting

insulin, HOMA-IR, and HOMA-Beta in those with optimal vitamin D status.

Finally, we examined the association between cardiometabolic risk biomarkers and vitamin D

status separately within men, women HC non-users, and women HC users (Table 4.4). When

examined as a whole, there was an interaction effect of 25(OH)D and HC use on triglycerides in

the overall population. When examined separately, among men, fasting insulin and HOMA-IR

were lower in those with an adequate or optimal vitamin D status than those who were vitamin

D-deficient or insufficient. Men with an optimal vitamin D status had the lowest HOMA-Beta

scores. Triglyceride concentrations were higher in men who were vitamin D-deficient or

insufficient than those with an adequate vitamin D status. Women HC non-users with an optimal

vitamin D status had lower insulin concentrations, HOMA-IR and HOMA-Beta scores than those

who were vitamin D-deficient. Among women HC users, we observed no associations between

73

any biomarkers and vitamin D status, although we noted a non-significant trend for measures of

glycemic dysregulation to be higher in the vitamin D-deficient category.

74

Table 4.1. Study participant characteristics 1,2

.

Sex

Men

(n=427)

Women

(n=957) p

Caucasian

(n=695)

East Asian

(n=525)

South Asian

(n=164) p

% Female

68.6 72.2 61.6 0.0340

% HC use among women

29.3

44.2 13.7 16.8 <0.0001

Age (years) 22.9 ± 2.6 22.6 ± 2.4 0.0091 23.2 ± 2.6 a 22.1 ± 2.2

b 22.2 ± 2.4

b <0.0001

BMI (kg/m2) 23.7 ± 3.4 22.4 ± 3.3 <0.0001 23.4 ± 3.5

a 21.8 ± 2.7

b 23.5 ± 4

a <0.0001

Waist circumference (cm) 80.4 ± 8.6 71.1 ± 7.1 <0.0001 75.8 ± 8.8 a 71.2 ± 7.3

b 75.1 ± 10.2

a <0.0001

Systolic blood pressure (mm Hg) 123.9 ± 10.1 109.1 ± 9.3 <0.0001 115.8 ± 11.4 a 111.1 ± 11.7

b 112.9 ± 11.6

b <0.0001

Diastolic blood pressure (mm Hg) 71.7 ± 7.6 68.3 ± 7.8 <0.0001 69.7 ± 7.8 68.7 ± 8.1 70.0 ± 7.6 0.06

Physical Activity (Met-h/wk) 7.8 ± 3.2 7.6 ± 3 0.39 8.2 ± 3.1 a 6.9 ± 3.0

b 7.8 ± 3.0

a <0.0001

Plasma 25(OH)D (nmol/L) 54.3 ± 26.1 59.1 ± 30.1 0.0100 70.8 ± 31 a 46.6 ± 19

b 36.4 ± 17.8

c <0.0001

Glucose (mmol/L) 4.9 ± 0.4 4.7 ± 0.3 <0.0001 4.8 ± 0.3 a 4.8 ± 0.4

a 4.9 ± 0.4

b <0.0001

75

Fasting insulin (pmol/L) 42.3 ± 24.5 48.0 ± 36.1 0.0033 44.7 ± 35.6 a 44.0 ± 26.8

a 60.4 ± 36.7

b <0.0001

HOMA-IR 0.8 ± 0.5 0.9 ± 0.6 0.0092 0.8 ± 0.6 a 0.8 ± 0.5

a 1.1 ± 0.7

b <0.0001

HOMA-Beta 81.3 ± 29.8 95.1 ± 40.3 <0.0001 89.8 ± 38.7 a 88.0 ± 34.5

a 104.3 ± 41.8

b <0.0001

Triglycerides (mmol/L) 1.0 ± 0.5 0.9 ± 0.4 0.06 1.0 ± 0.5 0.9 ± 0.4 1.0 ± 0.5 0.24

Total cholesterol (mmol/L) 4.1 ± 0.8 4.3 ± 0.8 <0.0001 4.2 ± 0.8 4.3 ± 0.7 4.2 ± 0.8 0.65

HDL cholesterol (mmol/L) 1.3 ± 0.3 1.6 ± 0.4 <0.0001 1.6 ± 0.4 a 1.6 ± 0.4

a 1.4 ± 0.4

b <0.0001

LDL cholesterol (mmol/L) 2.3 ± 0.7 2.23 ± 0.6 0.30 2.3 ± 0.6 a 2.3 ± 0.6

ab 2.4 ± 0.7

b 0.0406

Total: HDL cholesterol ratio 3.2 ± 0.9 2.7 ± 0.6 <0.0001 2.9 ± 0.7 a 2.8 ± 0.7

a 3.2 ± 1.0

b <0.0001

1 Shown are means ± standard deviations for continuous variables, and percentages for categorical variables. P-values from tests using

with loge- or square-root transformed variables as necessary to improve fit.

2 Differences between groups were tested using t-tests for binary variables, ANOVA for variables with more than 2 categories, and χ

2 tests

for categorical variables.

a,b,c Mean values within a row with different superscript letters were significantly different (p<0.05). Comparisons between groups were

made using the Tukey-Kramer procedure.

76

Table 4.2. Distribution of vitamin D status among men, women HC non-users, women HC users,

and different ethnic groups1.

Plasma 25(OH)D (nmol/L)

<30 30 - 49.9 50 - 74.9 ≥75 p

% % % %

Men (n=427) 15.7 36.3 27.6 20.4

<0.0001 Women HC non-users

(n=677) 17.1 37.5 31.2 14.2

Women HC users (n=280) 5.0 15.7 29.3 50.0

Caucasian (n=695) 4.8 22.9 33.0 39.4

<0.0001 East Asian (n=525) 17.1 45.5 29.7 7.6

South Asian (n=164) 45.1 33.5 15.9 5.5

1 Differences in distributions between groups were tested using χ

2 tests.

77

Table 4.3. Association between vitamin D status and biomarkers of cardiometabolic disease risk across ethnic groups 1,2

.

25(OH)D (nmol/L)

< 30 30 - 49.9 50 - 74.9 ≥ 75 p

Caucasians

n 33 159 229 274

Glucose (mmol/L) 4.9 ± 0.1 4.8 ± 0.1 4.7 ± 0.1 4.7 ± 0.1 0.55

Insulin (pmol/L) § 64.3 ± 8.9 a

50.6 ± 4.3 ab

43.4 ± 1.7 b 39.9 ± 1.5

c <0.0001

HOMA-IR § 1.2 ± 0.2 a 0.9 ± 0.1

ab 0.8 ± 0.1

b 0.7 ± 0.1

c <0.0001

Homa-Beta § 109.8 ± 9.7 a 95.7 ± 3.6

a 89.0 ± 2.2

a 84.6 ± 2.1

b <0.0001

Triglycerides (mmol/L) 1.1 ± 0.1 1.0 ± 0.1 0.9 ± 0.1 1 ± 0.1 0.0413

Total cholesterol (mmol/L) 4.1 ± 0.1 4.1 ± 0.1 4.3 ± 0.1 4.3 ± 0.1 0.27

HDL cholesterol (mmol/L) 1.4 ± 0.1 1.5 ± 0.1 1.6 ± 0.1 1.6 ± 0.1 0.0260

LDL cholesterol (mmol/L) 2.2 ± 0.1 2.2 ± 0.1 2.3 ± 0.1 2.3 ± 0.1 0.65

Total:HDL cholesterol 3 ± 0.1 3.0 ± 0.1 2.8 ± 0.1 2.8 ± 0.1 0.52

East Asians

78

n 90 239 156 40

Glucose (mmol/L) 4.8 ± 0.1 4.8 ± 0.1 4.8 ± 0.1 4.8 ± 0.1 0.76

Insulin (pmol/L) § 50.4 ± 2.8 a

45.1 ± 1.9 b 39.5 ± 1.8

c 40.7 ± 3.2

bc 0.0002

HOMA-IR § 0.9 ± 0.1 a 0.8 ± 0.1

b 0.7 ± 0.1

c 0.8 ± 0.1

bc 0.0002

Homa-Beta § 96.1 ± 3.3 a 89.7 ± 2.6

ab 81.7 ± 2.1

b 84.8 ± 4.9

b 0.0010

Triglycerides (mmol/L) 0.9 ± 0.1 0.9 ± 0.1 0.9 ± 0.1 1.0 ± 0.1 0.42




Total:HDL cholesterol 2.8 ± 0.1 2.8 ± 0.1 2.8 ± 0.1 2.8 ± 0.1 0.99

South Asians

n 74 55 26 9

Glucose (mmol/L) 5.0 ± 0.1 4.9 ± 0.1 4.8 ± 0.1 4.8 ± 0.1 0.35

Insulin (pmol/L) 65.1 ± 4.7 56.1 ± 4.9 59.2 ± 5.5 50.3 ± 9.6 0.53

HOMA-IR 1.2 ± 0.1 1.0 ± 0.1 1.1 ± 0.1 0.9 ± 0.2 0.51

Homa-Beta 106.5 ± 4.9 101.2 ± 5.9 108.3 ± 8 92.7 ± 8.8 0.7

79






1 Shown are crude means ± standard errors. P-values from tests using with loge- or square-root transformed variables as necessary to

improve fit.

2 The potential interaction effects of 25(OH)D and ethnicity on cardiometabolic disease biomarkers were explored in the overall

population in models including an interaction term for vitamin D status x ethnicity, as well as age, sex, waist circumference, physical

activity, season, and HC use among women as covariates. No interaction terms were statistically significant (p<0.05). As well, the

association between 25(OH)D and each biomarker was examined specifically within each ethnic group using ANCOVA with the

following covariates: age, sex, waist circumference, physical activity, season, and HC use among women. In addition, models examining

the association between vitamin D and HOMA-Beta were also adjusted for trigylcerides:HDL cholesterol as a surrogate measure for

insulin resistance.

§ Significantly associated with vitamin D status at the Bonferroni level [p<0.002; α=0.05, 27 independent tests (9 tests per ethnicity x 3

ethnicities)].

a,b,c: Mean values within a row with different superscript letters were significantly different from each other (p<0.05). Comparisons

between groups were made using the Tukey-Kramer procedure.

80

Table 4.4. Association between vitamin D status and biomarkers of cardiometabolic disease risk across men, women HC non-users, and

women HC users 1,2

.

25(OH)D (nmol/L)

< 30 30 - 49.9 50 - 74.9 ≥ 75 p

Men

n 67 155 118 87

Glucose (mmol/L) 5.0 ± 0.1 4.9 ± 0.1 4.9 ± 0.1 4.9 ± 0.1 0.50

Insulin (pmol/L) § 55.8 ± 3.4 a 44.8 ± 1.8

a 38.1 ± 2.0

b 33.1 ± 2.4

b <0.0001

HOMA-IR § 1.0 ± 0.1 a 0.8 ± 0.1

a 0.7 ± 0.1

b 0.6 ± 0.1

b <0.0001

Homa-Beta § 94.3 ± 3.7 a 85.8 ± 2.4

a 76.6 ± 2.5

a 69.9 ± 3.0

b 0.0001

Triglycerides (mmol/L) § 1.1 ± 0.1 a 1.1 ± 0.1

a 0.9 ± 0.1

b 1.0 ± 0.1

ab 0.0002





Women HC non-users

81

n 116 254 211 96

Glucose (mmol/L) 4.8 ± 0.1 4.7 ± 0.1 4.7 ± 0.1 4.7 ± 0.1 0.36

Insulin (pmol/L) § 58.8 ± 4.0 a 50.1 ± 3.1

ab 41.5 ± 1.6

bc 38.7 ± 2.5

c 0.0002

HOMA-IR § 1.1 ± 0.1 a 0.9 ± 0.1

ab 0.8 ± 0.1

bc 0.7 ± 0.1

c 0.0002

Homa-Beta § 104.8 ± 4.2 a 96.5 ± 3.1

ab 86.3 ± 1.9

bc 81.3 ± 3.4

c 0.0003






Women HC users

n 14 44 82 140

Glucose (mmol/L) 4.7 ± 0.1 4.7 ± 0.1 4.7 ± 0.1 4.6 ± 0.1 0.96

Insulin (pmol/L) 65.2 ± 8.3 50.9 ± 4.5 53.4 ± 3.2 45.9 ± 2.0 0.11

HOMA-IR 1.2 ± 0.1 0.9 ± 0.1 1.0 ± 0.1 0.8 ± 0.1 0.11

Homa-Beta 119.4 ± 11.1 99.9 ± 5.8 105.9 ± 4.1 96.6 ± 2.9 0.09

82






1 Shown are crude means ± standard errors. P-values from tests using with loge- or square-root transformed variables as necessary to

improve fit.

2 The potential interaction effects of 25(OH)D and HC use on cardiometabolic disease biomarkers were explored in the overall population

in models including an interaction term for vitamin D status x HC use, as well as age, sex, waist circumference, physical activity, season,

and ethnicity as covariates. An interaction term for 25(OH)D x HC use existed for triglycerides (p for interaction = 0.0147), but no other

biomarkers. As well, the association between 25(OH)D and each biomarker was examined specifically within men, women HC non-users

and women HC users using ANCOVA with the following covariates: age, waist circumference, ethnicity, physical activity, and season. In

addition, models examining the association between vitamin D and HOMA-Beta were also adjusted for trigylcerides:HDL cholesterol as a

surrogate measure for insulin resistance.

§ Significantly associated with vitamin D status at the Bonferroni level [p<0.002; α=0.05, 27 independent tests (9 tests per group x 3

groups)].

a,b,c: Mean values within a row with different superscript letters were significantly different from each other (p<0.05). Comparisons

between groups were made using the Tukey-Kramer procedure.

83

Figure 4.1. Seasonal fluctuations in 25(OH)D across ethnic groups.

Values are crude means, with their respective standard errors shown as vertical bars. Differences

between groups were tested using ANCOVA adjusted for age, sex, waist circumference, physical

activity, and HC use among women. Letters indicate significant (p<0.05) differences between

ethnic groups within a given season. Comparisons between groups were made using the Tukey-

Kramer procedure. Asterisks indicate significant (p<0.05) differences within a given ethnic

group across seasons. Across seasons, Caucasians had the highest 25(OH)D concentrations.

Caucasians and East Asians recruited during summer and fall had higher average concentrations

than those recruited during winter and spring. Among South Asians, 25(OH)D concentrations did

not fluctuate seasonally.

84

Figure 4.2. Seasonal fluctuations in 25(OH)D across men, women HC non-users, and women HC

users.

Values are crude means, with their respective standard errors shown as vertical bars. Differences

between groups were tested using ANCOVA adjusted for age, waist circumference, ethnicity,

and physical activity. Letters indicate significant (p<0.05) differences between groups within a

given season. Comparisons between groups were made using the Tukey-Kramer procedure.

Asterisks indicate significant (p<0.05) differences within a given group across seasons. Mean

25(OH)D concentrations fluctuated seasonally across men, women HC non-users, and women

HC users, with concentrations being higher for those recruited in summer and fall than winter

and spring. Women HC users had the highest 25(OH)D concentrations across seasons.

85

4.5 Discussion

The present study explored the relationship between ethnicity, HC use, vitamin D status, and

biomarkers of cardiometabolic disease risk in a population of young adults in Canada. In line

with previous studies, we identified inverse associations between biomarkers of glycemic

dysregulation and 25(OH)D, potentially suggesting a beneficial role of vitamin D in glucose

metabolism (30;182;290). Of note, we observed these associations in men and women HC non-

users, but not women HC users. Similarly, we noted inverse associations between biomarkers of

glycemic dysregulation and 25(OH)D in Caucasian and East Asian, but not South Asian,

individuals. However, the lack of association observed in women HC users and South Asians

may have been due to the smaller sample size of these groups.

Experimental evidence suggests that vitamin D may improve insulin secretion and also

upregulate the expression of insulin receptors, leading to increased insulin sensitivity (26;272).

In the present study, fasting insulin, HOMA-IR and HOMA-β were inversely associated with

25(OH)D, but fasting glucose measures were similar across vitamin D status categories in men,

women HC non-users, Caucasians, and East Asians. Sustained elevated production of insulin by

the pancreas to counteract high circulating glucose may, over time, result in both insulin

resistance and β-cell dysfunction, which are major steps in the progression of T2D (101). The

inverse association between 25(OH)D, insulin and HOMA indices observed here should be

interpreted with caution because vitamin D deficiency is often paralleled by visceral adiposity

and fatty liver disease, conditions which independently contribute to impaired glucose

metabolism. Furthermore, the cross-sectional study design precludes making inferences about

causality. However, while the participants of the present study were young and generally healthy,

it is possible that the onset of insulin resistance and β-cell dysfunction may already be underway

among vitamin D-deficient young adults.

In agreement with previous research, we observed elevated circulating 25(OH)D among women

HC users in the TNH study population (188;191). As shown in the present study, half of women

HC users, but one fifth or less of men and women HC non-users, were in the vitamin D optimal

status category, and the observed association between HC use and elevated 25(OH)D was

consistent across seasons. Most modern HC preparations include a combination of ethinyl

estradiol, a synthetic estrogen, and progestins (192). In human cell lines and in animal models,

86

estrogen has been shown to downregulate CYP24A1, the main enzyme responsible for catalysis

of 25(OH)D, and to upregulate CYP27B1, which converts 25(OH)D to 1,25(OH)2D, the primary

biologically active vitamin D metabolite (193;194). In addition, estrogen upregulates expression

of the VDR in various animal tissues and human cell lines (195), and it is associated with

elevated circulations of DBP (196;197). Thus, estrogen appears to play an important role in

modulating vitamin D metabolism, but the clinical relevance of its effects on circulating vitamin

D metabolites is poorly understood.

To our knowledge, no other study has directly examined the association between vitamin D and

cardiometabolic disease biomarkers separately across women HC users, women HC non-users

and men. In the present study, while vitamin D was inversely associated with markers of

glycemic dysregulation in men and women HC non-users, no significant associations existed

between vitamin D and any cardiometabolic disease biomarkers among women HC users. The

use of HC is associated with decreased insulin sensitivity in healthy, non-obese young women,

although the potential mechanism remains unclear (204;337;345). Overall, the results of the

present study may suggest that, among those who use HC, the potential benefits of elevated

circulating vitamin D on glycemic regulation may be attenuated by the deleterious effects of HC

on this pathway (204;207). We recently reported a confounding effect of HC on the association

between 25(OH)D and CRP, a biomarker of systemic inflammation (331), as well as a similar

confounding effect of these medications on the relationship between 25(OH)D and the plasma

proteome (343). Along parallel lines, an examination of data from the Women's Health Initiative

trial found that estrogen-containing HRT modified the effect of vitamin D and calcium

supplementation on colorectal cancer (245), and these medications were found to attenuate an

inverse association between 25(OH)D and breast cancer in a large population of German post-

menopausal women (190). In the present study, we did observe a non-significant trend for

decreased concentrations of biomarkers of glycemic dysregulation in women HC users with

increasing vitamin D status, similar to the significant inverse associations seen in men and

women HC non-users. It is possible that, among women who use sex hormone-containing

medications, any beneficial effects of vitamin D may only be realized at concentrations of

25(OH)D higher than those observed in the present study. However, it is also important to note

that no statistical interaction existed between 25(OH)D and HC use for any biomarkers other

than triglycerides. Therefore, the lack of association between vitamin D and any biomarkers in

87

women HC users may have been due to the smaller sample size of this group. Future studies with

larger sample sizes are needed to further examine the potential effects of HC use on the

association between 25(OH)D and cardiometabolic disease.

The results of the present study corroborate previous findings showing that the prevalence of

vitamin D deficiency among young Canadians varies by ethnic group, with circulating 25(OH)D

being particularly low in South Asians (153;178). Indeed, nearly half of South Asians in our

study population were vitamin D-deficient (<30 nmol/L), compared to only 5% of Caucasians.

The lower 25(OH)D concentrations observed in South Asians are likely due to reduced

endogenous vitamin D synthesis that may be partly a result of having darker skin and/or adopting

a sun-avoiding lifestyle. Our data also support previous findings that seasonal effects on

25(OH)D are absent among South Asians (153), suggesting that endogenous vitamin D synthesis

from casual sun exposure in summer and fall is low in this group.

While we noted trends towards improved insulin sensitivity and pancreatic β-cell function with

higher vitamin D status in all ethnocultural groups, these associations were not statistically

significant in South Asians. The lack of statistical interaction between vitamin D and ethnicity

suggests that the smaller sample size and generally low 25(OH)D concentrations observed in the

South Asian group may have prevented identifying existing links between vitamin D status and

any cardiometabolic markers. Indeed, in a previous study, 25(OH)D partly explained differences

in insulin sensitivity between Caucasians and African Americans (182). Furthermore, a recent

intervention study found improvements in insulin sensitivity with vitamin D supplementation in

South Asian women (290).

We observed that triglycerides were associated with vitamin D status in men, with individuals in

the adequate status category having lower concentrations than those who were vitamin D

insufficient or deficient. We observed no association between vitamin D status and any other

marker of lipid metabolism in women or any ethnocultural group. Previous studies have reported

an inverse relationship between vitamin D and triglycerides, although the results have been

inconsistent (306;346). The underlying mechanism for the association between 25(OH)D and

triglycerides may involve vitamin D upregulation of lipoprotein lipase activity in fat cells,

leading to decreased circulating triglyceride concentrations (169). Our study population consists

of young adults, and it is possible that any potential deleterious effects of vitamin D deficiency

88

on triglycerides in women may manifest later in life. However, the present study is not designed

to determine whether the observed sex differences in the association between 25(OH)D and

triglycerides are biologically based or a result of unaccounted confounding.

Numerous recent studies have examined the association between vitamin D and markers of

cardiometabolic risk in both adults and children (29;87;239-241). However, the results are

inconsistent, possibly due to small sample sizes, differences in age, sex, ethnicity, and disease

status between the populations studied, and diverse environmental and lifestyle-related factors. A

strength of the present study is the ethnocultural diversity of the study population, which allowed

for assessing the association between 25(OH)D and cardiometabolic biomarkers across various

ethnic groups. In addition, circulating 25(OH)D is considered a better biomarker of vitamin D

status than dietary vitamin D intake, since 25(OH)D captures both endogenous and exogenous

sources of this micronutrient. Finally, assessing the relationship between 25(OH)D and

biomarkers of cardiometabolic risk in health young adults provides an opportunity to better

understand the potential role of this micronutrient in the early stages of cardiometabolic disease

development.

One limitation of the present study is the relatively low number of South Asian participants, as

well as women HC users. Although subjects were recruited from a university campus and may

not be representative of all young adults across Canada, our results corroborate those of a

previous study conducted in a similar population (153). We were unable to assess whether

estrogen or progesterone-derived HC components have different effects on the associations

examined here, and we also lacked information on endogenous sex hormone concentrations.

Finally, unaccounted genetic variation and residual confounding from unidentified factors, such

as time spent outdoors or use of sunscreen, may confound the results presented here.

In summary, the present study found that vitamin D status is inversely associated with

biomarkers of glycemic dysregulation in men and women HC non-users and certain ethnic

groups, but not in women HC users. Our results suggest that HC use among women may

attenuate the association between vitamin D and glycemic dysregulation, but future studies with

larger sample sizes are needed to further examine this potential relationship. Indeed, given the

widespread use of HC by women throughout the world, a better understanding of the extent to

89

which these medications may modify the relationship between vitamin D and processes related

to disease is warranted. In addition, we observed that the prevalence of vitamin D deficiency

varies greatly among young adults living in Canada, depending on ethnicity. Considering the

ethnocultural diversity of the Canadian population, further efforts should be directed to

characterizing the relationship between vitamin D and cardiometabolic disease risk and to

evaluate its preventive efficacy in various ethnic groups.

90

Chapter 5 : Plasma 25-hydroxyvitamin D, hormonal contraceptive use,

and the plasma proteome in Caucasian, East Asian, and South Asian

young adults

Adapted with permission from:

García-Bailo B, Jamnik J, Da Costa LA, Borchers CH, Badawi A, El-Sohemy A.

Plasma 25-hydroxyvitamin D, hormonal contraceptive use, and the plasma

proteome in Caucasian, East Asian, and South Asian young adults. J Proteome

Res. 2013. Feb;12(4):1797-1807. doi: 10.1021/pr4001042. Copyright (2013)

American Chemical Society.

91

5.1 Abstract

Background: Vitamin D affects gene expression, but its downstream effects on the proteome are

unclear. Hormonal contraceptives (HC), which affect vitamin D metabolism and have

widespread effects on the plasma proteome, may confound the association between vitamin D

and the proteome. We determined whether HC use modified the association between 25-

hydroxyvitamin D (25(OH)D) and a panel of 54 high-abundance plasma proteins.

Methods: Cross-sectional analyses were conducted in healthy, non-smoking women HC users

(n=216), women HC non-users (n=502), and men (n=301) from Toronto, Canada. Plasma

25(OH)D was measured by HPLC-MS/MS, and proteins were measured by LC-multiple-

reaction-monitoring (MRM)-MS.

Results: The 54 proteins clustered into four distinct proteomic profiles. A positive association

was observed between Profile 1, containing positive acute phase proteins, and 25(OH)D. In

women HC users, a J-shaped association existed between Profile 1 and 25(OH)D, but no

associations existed in women HC non-users and men. Twelve proteins were individually

associated with 25(OH)D in women HC users, but only two were associated with 25(OH)D in

women HC non-users and no associations were observed in men. After accounting for hormone

dose, only three proteins were associated with 25(OH)D.

Conclusions: In summary, HC use is an important confounder of the association between

circulating 25(OH)D and numerous plasma proteins.

92

5.2 Introduction

Cardiometabolic disease, including T2D and CVD, represents a global health problem (1).

Vitamin D has been examined extensively as a potentially preventive factor in the development

of cardiometabolic and other chronic diseases (151). However, as summarized by recent

systematic reviews, the evidence from epidemiologic studies and clinical trials is inconclusive

(35-37). The inconsistencies between studies may result not only from differences in

methodology, but also from a lack of adequate biomarkers, unaccounted confounding from

various demographic and lifestyle factors, and genetic differences both within and between

populations (38).

The protective actions of vitamin D on cardiometabolic disease are thought to occur through

modulation of physiologic processes that become dysregulated during disease progression, such

as inflammation and pancreatic β-cell function (26;32;272;290). Vitamin D, obtained through

cutaneous sunlight exposure or through the diet, is hydroxylated in the liver to 25(OH)D (121).

This metabolite is the most abundant circulating form of vitamin D, and it is the accepted

functional biomarker of vitamin D status (151). Conversion of 25(OH)D to 1,25(OH)2D, the

main biologically active form of the vitamin, occurs through hydroxylation in the kidney and

other target tissues (121). The latter form of vitamin D is a steroid hormone that affects both the

regulation of target gene transcription and the activation of various signal transduction pathways

when it binds to the VDR, a transcription factor with effects on numerous tissues (121).

Recent in vitro studies have examined vitamin D action across the genome, using gene




colorectal cancer (27). However, it is not known whether the observed widespread effects of

vitamin D at the genome level in vitro translate into downstream effects at the level of the

proteome in vivo. Over 3,000 proteins have been identified in the human plasma proteome

(57;119). Among the most abundant, ranging in concentration between <1 and 800 μmol/L, are

many molecules of physiological importance, such as apolipoproteins, members of the

complement system, coagulation factors, carrier proteins, and protease inhibitors (57). Many of

these proteins are acute phase reactants whose levels become altered during inflammatory

93

processes, and they may represent important biomarkers of disease progression (59;120). A

recently developed LC-MS/MS-based, MRM proteomics assay can simultaneously measure the

concentrations of 54 high-abundance plasma proteins (120). The concentrations of these 54

proteins were found to differ across ethnic groups and individuals with different dietary intakes

(119).

Use of HC has been linked to elevated 25(OH)D concentrations (188;191). HC use by women is

common worldwide, not only to prevent unwanted pregnancies, but also to treat a number of

dermatological and gynecological conditions (204;205). At the same time, these medications are

associated with elevated risk of CVD and impaired glucose tolerance (204;208). Furthermore, we

recently reported a widespread effect of HC use on the plasma proteome (207). Therefore, HC

use may have significant effects on the relationship between vitamin D and numerous diseases.

Determining whether vitamin D modulates the concentrations of plasma proteomic biomarkers,

and understanding whether HC use modifies the association between vitamin D and the plasma

proteome, may help identify novel pathways affected by this micronutrient, as well as elucidate

some of the inconsistencies between studies on the relationship between vitamin D and

cardiometabolic disease. The objective of this study was to examine the association between

25(OH)D and plasma proteomic biomarkers in young adults of Caucasian, East Asian, and South

Asian ancestry, and to determine whether HC use modifies this association.

5.3 Methods


The individuals included in this study (n=1,019) were healthy, non-smoking women HC users

(n=216), women HC non-users (n=502), and men (n=301) who, at the time that the study was

conducted, had available measurements for 25(OH)D, the proteomics assay, and each of the

covariates considered in the analyses. Individuals of Aboriginal Canadian, Afro-Caribbean, or

mixed ancestry were excluded from the analyses reported here because of insufficient sample

size to make adequately powered comparisons. For more details on the TNH study population,

recruitment, exclusions, and classification into ethnic groups and seasons, please refer to Chapter

2, Section 2.3.1.

94







2, Section 2.3.5.

For details on how vitamin D status categories were created, please refer to Chapter 4, Section

4.3.4.

5.3.5 Plasma Proteomic Measurements

The concentrations of high-abundance plasma proteins were measured at the University of

Victoria – Genome British Columbia Proteomics Centre (Victoria, BC, Canada), using a MS-

MRM assay as described previously (59;119;120;207). In brief, the assay involves quantifying

proteins against stable isotope labelled standard proteotypic peptides, representative of

endogenous proteins, that are added to tryptic digests of plasma samples. The protein

concentration measurements obtained with this MS-MRM assay are comparable to reported

values (57). The original assay was able to measure 45 proteins (120). Since being developed,

the list of measurable proteins has expanded to a total of 63. Of the 63 proteins included in the

assay, the present study assessed 54 proteins that had inter-assay coefficients of variation of

<15%, as previously described (119).


All statistical analyses were performed using SAS. Subject characteristics across the four vitamin

D status categories were compared using χ2 tests and ANOVA. Continuous variables that were

not normally distributed were loge- or square root-transformed prior to analysis. Except where

indicated, the p-values from models using the transformed values of these variables are reported,

but untransformed means and measures of spread are reported to facilitate interpretability.

95

Principal components analysis was used to identify principal components representative of

independent plasma proteomic profiles, based on the concentrations of the 54 proteins included

in the proteomics assay, as described previously (119). The principal components representative

of the proteomic profiles were obtained through orthogonal rotation with the Varimax rotation,

which provided a component structure with independent patterns (347). Components were

chosen using the consideration parameters of eigenvalues >1 and the Scree test. Each protein

received a loading score for each one of the identified components, which was representative of

the degree of correlation of that protein with that principal component. A loading score criterion

of ≥0.5 was used to determine inclusion of a given protein into a specific component,

representative of specific proteomic profiles.

In the present study, we examined the association between each of the identified proteomic

profiles and circulating 25(OH)D using ANCOVA, first in the population as a whole and then

separately in women HC users, women HC non-users, and men. The covariates included in the

models were age, waist circumference, physical activity, season, ethnicity, and, in the analysis

for the whole population, sex and HC use as well. These analyses allowed us to identify groups

of proteins to explore further.

When a significant association between a specific proteomic profile and 25(OH)D was observed

in any group, we then examined the association between each individual protein included in that

profile and 25(OH)D separately among women HC users, HC non-users, and men, so as to better

characterize the individual relationships between these proteins and vitamin D across groups.

ANCOVA was conducted in each group with age, waist circumference, physical activity, season,

and ethnicity as covariates. In consideration of the large number of individual tests conducted,

these analyses were adjusted for multiple testing using the Bonferroni correction.

In women HC users, we examined the association between proteins identified in previous

analyses and 25(OH)D separately among those taking <1mg (n= 132) vs. ≥1mg (n= 46) daily

total hormone, in order to assess the potential role of hormone dose in the relationship between

vitamin D and plasma proteins. Thirty-eight subjects were excluded from the analyses because

the self-reported type of HC made ascertainment of daily hormonal dose difficult, as described

previously (207). The association between each of the identified proteins and 25(OH)D was

examined separately in users of <1mg vs. ≥1mg total hormone per day, using ANCOVA

96

adjusted for age, waist circumference, physical activity, season, and ethnicity. As before, these

analyses were adjusted for multiple testing using the Bonferroni correction.

5.4 Results

Study participant characteristics are shown in Table 5.1. We observed significant differences in

the distribution of vitamin D status between men and women and across ethnic groups, with

more women and Caucasians being represented in the optimal category of vitamin D and more

men and East and South Asians being vitamin D deficient. Physical activity levels were higher

among those with adequate or optimal vitamin D status than those who were vitamin D

insufficient or deficient. Use of HC differed across vitamin D status categories, with a higher

proportion of women reporting use of contraceptives in the optimal vitamin D status category

than the other categories. Finally, we also observed significant differences in the distribution of

vitamin D status by season, with a greater proportion of individuals being vitamin D deficient in

winter and spring than in summer and fall.

The 54 plasma proteins clustered into four main proteomic profiles. The proteins included in

each profile are listed in Table 5.2. In brief, profile 1 included primarily positive acute phase

proteins with roles in various physiological systems, such as hemostasis, lipid metabolism, signal

transduction, and transmembrane transport (119;348). Profile 2 consisted of several negative

acute phase proteins also involved in various physiological processes (119;348). Profile 3

comprised proteins involved in innate immunity and regulation of the complement cascade

(119;348). Finally, profile 4 included proteins involved exclusively in coagulation (119;348).

While most proteins loaded onto a single proteomic profile, four proteins (clusterin, α2-

antiplasmin, apolipoprotein C-I, and complement C-3) had loading scores ≥0.5 for two

proteomic profiles and, therefore, were included in both. Four proteins (α1B-glycoprotein, zinc-

α2-glycoprotein, apolipoprotein D, and adiponectin) did not load significantly onto any principal

component and, therefore, were not included in any.

We examined the association between 25(OH)D and each proteomic profile first in the

population as a whole, and then separately in women HC users, women HC non-users, and men

(Figure 5.1). In the overall population, profile 1 was positively associated with 25(OH)D, with

97

higher mean profile scores observed among individuals in the optimal vitamin D status category

than the other categories (p<0.0001). No other proteomic profiles were associated with 25(OH)D

in the whole population. When women HC users were examined separately, we observed a J-

shaped association between profile 1 and 25(OH)D, with individuals in the optimal vitamin D

status category having higher mean scores than those who had an insufficient or adequate

vitamin D status (p<0.0001). No other profiles were associated with 25(OH)D among women

HC users, although we noted a non-significant inverse association between profile 3 and

25(OH)D. Finally, we observed no associations between any proteomic profile and 25(OH)D in

women HC non-users or men.

Because proteomic profile 1 was associated with 25(OH)D both in the population as a whole and

among women HC users, we further evaluated the relationship between the proteins comprising

that profile and vitamin D. To do so, we examined the association between each individual

protein included in the profile and 25(OH)D separately in women HC users, women HC non-

users, and men (Table 5.3a-c). Twelve of the 25 proteins included in profile 1, such as ApoC-III,

DBP, and retinol-binding protein 4 (RBP4), were associated with 25(OH)D at the Bonferroni

level (p<0.0007) in women HC users (Table 5.3a). In general, within this group, women in the

optimal vitamin D status category had higher mean concentrations of these proteins than those in

the insufficient and adequate categories, whereas women with the lowest vitamin D status had

intermediate concentrations that did not differ significantly from the other vitamin D status

categories. For example, as shown in Table 5.3a, mean ApoC-III concentrations were over 3.3

μmol/L in women with optimal vitamin D status, but ranged between 2.4 and 2.8 μmol/L in

women in lower vitamin D status categories (p<0.0001).

While no associations had been observed between any proteomic profile and 25(OH)D among

women HC non-users, two of the proteins included in profile 1 were associated with 25(OH)D

when examined individually. RBP4 and coagulation factor XIIa heavy chain were both

positively associated with 25(OH)D at the Bonferroni level in women HC non-users (Table

5.3b). As previously mentioned, RBP4 was associated with 25(OH)D among women HC users as

well (Table 5.3a), but coagulation factor XIIa heavy chain was not associated with 25(OH)D in

that group, although a similar non-significant trend was observed. Finally, when the relationships

between proteins included in proteomic profile 1 and 25(OH)D were examined separately in

men, no significant associations were observed (Table 5.3c). Non-significant positive trends

98

were noted, however, for several proteins, including ApoC-III, RBP4, DBP, and coagulation

factor XIIa heavy chain.

In an effort to prevent overlooking potential associations between individual proteins not

included in profile 1 and 25(OH)D, we examined the association between each of the 54

individual proteins and 25(OH)D across women HC users, women HC non-users, and men

(Appendix Table A.1a-c). None of the proteins included in profiles 2, 3, and 4 were significantly

associated with 25(OH)D in any group when examined individually. In addition, none of the four

proteins that did not load into any proteomic profile (α1B-glycoprotein, zinc-α2-glycoprotein,

apolipoprotein D, and adiponectin) were associated with 25(OH)D in any group either.

We have previously reported a dose effect of HC use on the association between 25(OH)D and

CRP (331). Women HC users who took ≥1mg total hormone per day had higher concentrations

of both metabolites than those taking <1mg per day, and an apparent positive association

between 25(OH)D and CRP in women HC users overall was no longer significant after taking

hormone dose into account (331). Therefore, we examined whether HC dose affected the

association between 25(OH)D and plasma proteins included in profile 1 (Table 5.4). Among

women taking <1mg total daily hormone, only three of the 25 proteins included in the profile

(DBP, ApoC-III, and RBP4) remained associated with 25(OH)D at the Bonferroni level. We

observed a positive association between DBP and 25(OH)D. However, the association between

ApoC-III, RBP4, and 25(OH)D was J-shaped, with women in the optimal vitamin D status

category having higher concentrations of these proteins than women with an insufficient or

adequate vitamin D status. In women taking ≥1mg total hormone per day, no proteins were

associated with 25(OH)D at the Bonferroni level, although a non-significant J-shaped trend

existed for ApoC-III.

99

Table 5.1. Study participant characteristics by vitamin D status1

Plasma 25(OH)D (nmo/L)

Total

population

(n=1,019)

<30

(n=179)

30 to 49.9

(n=355)

50 to 74.9

(n=310)

≥75

(n=175) p

Age (years) 22.7 ± 2.5 22.5 ± 2.4a,b

22.3 ± 2.4a 23.0 ± 2.6

b 23.1 ± 2.4

b 0.0001

Sex

Male 301 (29.5) 62 (20.6) 118 (39.2) 87 (28.9) 34 (11.3) 0.0037

Female 718 (70.5) 117 (16.3) 237 (33.0) 223 (31.1) 141 (19.6)

Ethnocultural group

Caucasian 519 (50.9) 34 (6.6) 134 (25.8) 197 (38.0) 154 (29.7) <0.0001

East Asian 381 (37.5) 85 (22.3) 185 (48.4) 96 (25.1) 16 (4.2)

South Asian 118 (11.6) 60 (50.9) 36 (30.5) 17 (14.4) 5 (4.2)

BMI (kg/m2) 22.7 ± 3.4 23.0 ± 4.2 22.6 ± 3.2 22.8 ± 3.2 22.4 ± 3.0 0.3305

Waist circumference (cm) 73.7 ± 8.6 74.4 ± 11.0 73.7 ± 8.6 73.9 ± 7.7 72.6 ± 7.2 0.3366

Systolic blood pressure (mm

Hg) 113.6 ± 11.3 113.8 ± 11.9 113.4 ± 11.1 114.4 ± 11.9

112.15 ±

10.0 0.2105

Diastolic blood pressure (mm

Hg) 68.9 ± 7.9 69.2 ± 8.0 68.8 ± 7.8 69.3 ± 8.5 68.2 ± 7.2 0.5379

Physical Activity (Met-h/wk) 7.6 ± 3.1 6.9 ± 3.0a 7.3 ± 3.0

a 7.9 ± 3.0

b 8.2 ± 2.7

b <0.0001

100

Hormonal contraceptive use

among women

No 502 (69.9) 103 (88.0) 202 (85.2) 149 (66.8) 48 (34.0) <0.0001

Yes 216 (30.1) 14 (12.0) 35 (14.8) 74 (33.2) 93 (66.0)

Season of blood draw

Spring 281 (27.6) 72 (25.6) 103 (36.7) 67 (23.8) 39 (13.9) <0.0001

Summer 296 (29.1) 24 (8.1) 100 (33.8) 109 (36.8) 63 (21.3)

Fall 261 (25.6) 31 (11.9) 87 (33.4) 85 (32.6) 58 (22.2)

Winter 181 (17.8) 52 (28.7) 65 (35.9) 49 (27.1) 15 (8.3)

Plasma 25(OH)D (nmol/L) 52.7 ± 25.2 22.4 ± 5.2a 39.9 ± 5.8

b 61.3 ± 7.1

c 94.6 ± 18.9

d <0.0001

1 Values shown are crude means ± standard deviations or n (%). P-values calculated from χ

2 analysis for categorical variables and

ANOVA for continuous variables, with loge-transformed values where necessary. Different superscript letters indicate significant

differences between groups (p<0.05). The Tukey-Kramer procedure was used to adjust for multiple comparisons between groups within

each ANOVA.

101

Table 5.2. Proteins included in each plasma proteomic profile1

Profile 1 Profile 2 Profile 3 Profile 4

Angiotensinogen Clusterin Complement C4 β chain Fibrinogen α2 chain

Kininogen-1 Albumin Complement C4 γ chain Fibrinogen β chain

Vitamin D binding protein Antithrombin-III α1-Acid glycoprotein 1 Fibrinogen γ chain

Ceruloplasmin Complement C1 inactivator Complement factor B Fibrinopeptide A

Vitronectin Gelsolin, isoform 1 Complement C9 Fibronectin

Plasminogen Histidine-rich glycoprotein α1-Antichymotrypsin

Apolipoprotein A-II precursor Apolipoprotein E Haptoglobin β chain

Retinol-binding protein 4 α2-Antiplasmin Complement C3

α1-Antitrypsin Apolipoprotein A-IV Complement factor H

Coagulation factor XIIa HC Apolipoprotein C-I Serum amyloid p-component

Heparin cofactor II Transthyretin

Apolipoprotein L1 β2-Glycoprotein I

Apolipoprotein A-I α2-Macroglobulin

Transferrin

102

Apolipoprotein C-III

α2-HS-glycoprotein

Afamin

Hemopexin

Inter- α-trypsin inhibitor HC

Clusterin

Prothrombin

Apolipoprotein B-100

α2-Antiplasmin

Apolipoprotein C-I

Complement C3

1 Proteomic profiles were derived by principal components analysis using orthogonal rotation with the Varimax rotation. Independent

principal components were determined using the criterion of eigenvalues >1, as well as the Scree test. Included within a given component

were individual proteins with a loading score ≥0.5 for that component. Proteins with loading scores ≥0.5 for two components were

included in both. Four proteins (α1B-glycoprotein, zinc-α2-glycoprotein, apolipoprotein D, and adiponectin) did not load significantly onto

any component and, therefore, were not included in any. Proteins within each profile are listed in descending order, based on loading

scores. Abbreviations: HC, Heavy chain; HS, Heremans-Schmid.

103

Table 5.3a-c. Mean plasma concentrations (μmol/L) of profile 1 proteins, by vitamin D status, by

sex and HC use1

a) Women HC users


<30

(n=14)

30 to 49.9

(n=35)

50 to 74.9

(n=74)

≥75

(n=93) p

Afamin 0.29 ± 0.02a,b

0.25 ± 0.01a 0.28 ± 0.01

b 0.30 ± 0.01

b <0.0001 §

Angiotensinogen 2.10 ± 0.23a,b

1.67 ± 0.15a 1.88 ± 0.10

a 2.36 ± 0.09

b <0.0001 §

Apolipoprotein C-III 2.84 ± 0.21a,b,c

2.37 ± 0.14a 2.84 ± 0.09

b 3.36 ± 0.08

c <0.0001 §

Ceruloplasmin 4.32 ± 0.31a,b

3.28 ± 0.20a 3.35 ± 0.14

a 3.86 ± 0.12

b <0.0001 §

Retinol-binding protein 1.13 ± 0.07a,b

1.02 ± 0.05a 1.09 ± 0.03

a 1.26 ± 0.03

b <0.0001 §

Vitamin D binding protein 3.55 ± 0.21a,b,c

3.16 ± 0.13a 3.54 ± 0.09

b 3.93 ± 0.08

c <0.0001 §

Apolipoprotein A-I 52.7 ± 2.77a 45.16 ± 1.75

b 49.2 ± 1.2

a,b 53.75 ± 1.07

a,c 0.0002 §

Transferrin 15.75 ± 0.83a,b

13.32 ± 0.53a 13.5 ± 0.36

a 15.24 ± 0.32

b 0.0003 §

Apolipoprotein C-I 3.43 ± 0.21a,b

3.06 ± 0.13a 3.44 ± 0.09

b 3.68 ± 0.08

b 0.0004 §

Heparin cofactor II 0.94 ± 0.06a,b

0.79 ± 0.04a 0.81 ± 0.02

a 0.91 ± 0.02

b 0.0004 §

Kininogen-1 2.81 ± 0.16a,b

2.49 ± 0.1a 2.62 ± 0.07

a 2.91 ± 0.06

b 0.0004 §

Apolipoprotein A-II

precursor 28.96 ± 1.55

a,b 27.41 ± 0.98

a 29.27 ± 0.67

a 31.71 ± 0.60

b 0.0005 §

α2-HS-glycoprotein 11.53 ± 0.61 9.39 ± 0.38 9.92 ± 0.26 10.53 ± 0.23 0.0035

Coagulation factor XIIa

HC

0.38 ± 0.03 0.34 ± 0.02 0.36 ± 0.01 0.41 ± 0.01 0.0042

Inter- α-trypsin inhibitor

HC

0.76 ± 0.03 0.63 ± 0.02 0.66 ± 0.01 0.69 ± 0.01 0.0070

α2-Antiplasmin 2.17 ± 0.10 1.84 ± 0.07 1.95 ± 0.05 2.03 ± 0.04 0.0076

104

α1-Antitrypsin 15.20 ± 0.91 13.13 ± 0.57 13.76 ± 0.4 14.80 ± 0.35 0.0141

Hemopexin 11.90 ± 0.52 10.57 ± 0.33 11.00 ± 0.23 11.68 ± 0.20 0.0143

Complement C3 23.67 ± 1.18 20.18 ± 0.75 21.35 ± 0.51 21.98 ± 0.46 0.0444

Clusterin 1.75 ± 0.08 1.53 ± 0.05 1.62 ± 0.04 1.68 ± 0.03 0.0503

Plasminogen 1.67 ± 0.08 1.42 ± 0.05 1.48 ± 0.04 1.52 ± 0.03 0.0537

Apolipoprotein B-100 0.91 ± 0.06 0.82 ± 0.04 0.89 ± 0.03 0.95 ± 0.02 0.0836

Vitronectin 5.13 ± 0.26 4.45 ± 0.17 4.49 ± 0.11 4.70 ± 0.10 0.0842

Apolipoprotein L1 0.70 ± 0.05 0.57 ± 0.03 0.56 ± 0.02 0.59 ± 0.02 0.0892

Prothrombin 0.62 ± 0.03 0.58 ± 0.02 0.62 ± 0.01 0.63 ± 0.01 0.1304

b) Women HC non-users


<30

(n=103)

30 to 49.9

(n=202)

50 to 74.9

(n=149)

≥75

(n=48) p

Afamin 0.25 ± 0.01 0.25 ± 0.01 0.25 ± 0.01 0.23 ± 0.01 0.3589

Angiotensinogen 0.68 ± 0.02 0.66 ± 0.02 0.7 ± 0.02 0.77 ± 0.04 0.0866

Apolipoprotein C-III 2.16 ± 0.07 2.28 ± 0.05 2.37 ± 0.06 2.41 ± 0.11 0.0317

Ceruloplasmin 2.06 ± 0.06 1.94 ± 0.04 2.08 ± 0.05 2.18 ± 0.09 0.0543

Retinol-binding protein 0.76 ± 0.02a 0.79 ± 0.02

a,b 0.85 ± 0.02

b,c 0.93 ± 0.03

c 0.0001 §

Vitamin D binding protein 2.54 ± 0.05 2.57 ± 0.04 2.69 ± 0.04 2.78 ± 0.07 0.0325

Apolipoprotein A-I 42.76 ± 0.94 43.08 ± 0.67 43.82 ± 0.78 44.69 ± 1.38 0.3769

Transferrin 12.61 ± 0.3 12.14 ± 0.21 12.48 ± 0.25 12.43 ± 0.44 0.3461

Apolipoprotein C-I 3.20 ± 0.09 3.22 ± 0.06 3.24 ± 0.07 3.42 ± 0.13 0.6789

105

Heparin cofactor II 0.67 ± 0.02 0.65 ± 0.01 0.67 ± 0.01 0.67 ± 0.03 0.6493

Kininogen-1 2.00 ± 0.04 1.95 ± 0.03 2.03 ± 0.03 2.09 ± 0.06 0.3359

Apolipoprotein A-II

precursor 23.58 ± 0.50 23.38 ± 0.35 23.87 ± 0.41 23.69 ± 0.73 0.9470



HC

0.21 ± 0.01a 0.21 ± 0.01

a 0.26 ± 0.01

b 0.27 ± 0.01

b <0.0001 §


HC

0.62 ± 0.01 0.61 ± 0.01 0.61 ± 0.01 0.61 ± 0.02 0.7522

α2-Antiplasmin 1.96 ± 0.04 1.89 ± 0.03 1.94 ± 0.04 1.93 ± 0.06 0.4369

α1-Antitrypsin 10.90 ± 0.22 10.16 ± 0.16 10.38 ± 0.19 10.58 ± 0.33 0.0290

Hemopexin 10.30 ± 0.22 9.73 ± 0.16 10.35 ± 0.18 10.22 ± 0.32 0.0348

Complement C3 19.13 ± 0.46 18.42 ± 0.33 19.01 ± 0.38 18.25 ± 0.67 0.1247

Clusterin 1.49 ± 0.03 1.49 ± 0.02 1.52 ± 0.03 1.54 ± 0.05 0.5858

Plasminogen 1.18 ± 0.02 1.16 ± 0.02 1.16 ± 0.02 1.19 ± 0.03 0.6600


Vitronectin 3.66 ± 0.07 3.49 ± 0.05 3.54 ± 0.06 3.41 ± 0.1 0.0412


Prothrombin 0.58 ± 0.01 0.57 ± 0.01 0.57 ± 0.01 0.55 ± 0.02 0.3779

c) Men


<30

(n=62)

30 to 49.9

(n=118)

50 to 74.9

(n=87)

≥75

(n=34) p

Afamin 0.26 ± 0.01 0.25 ± 0.01 0.24 ± 0.01 0.23 ± 0.01 0.0603

106



Ceruloplasmin 1.98 ± 0.07 1.81 ± 0.05 1.84 ± 0.06 1.91 ± 0.1 0.1901

Retinol-binding protein 0.90 ± 0.03 0.95 ± 0.02 0.97 ± 0.03 1.02 ± 0.04 0.0741

Vitamin D binding protein 2.57 ± 0.06 2.55 ± 0.04 2.68 ± 0.05 2.81 ± 0.08 0.1406


Transferrin 12.01 ± 0.34 11.64 ± 0.25 11.36 ± 0.29 11.47 ± 0.47 0.5161



Kininogen-1 2.11 ± 0.05 2.02 ± 0.04 2.00 ± 0.04 2.01 ± 0.07 0.3258

Apolipoprotein A-II

precursor 24.39 ± 0.62 24.38 ± 0.45 24.84 ± 0.53 23.64 ± 0.84 0.7895

α2-HS-glycoprotein 9.14 ± 0.24 8.42 ± 0.17 8.47 ± 0.20 8 ± 0.32 0.0547


HC

0.23 ± 0.01 0.24 ± 0.01 0.25 ± 0.01 0.28 ± 0.02 0.0046


HC

0.64 ± 0.02 0.60 ± 0.01 0.61 ± 0.01 0.59 ± 0.02 0.279

α2-Antiplasmin 1.97 ± 0.05 1.89 ± 0.04 1.89 ± 0.04 1.83 ± 0.07 0.6426

α1-Antitrypsin 10.63 ± 0.27 9.99 ± 0.20 9.71 ± 0.23 10.16 ± 0.37 0.0706

Hemopexin 9.82 ± 0.26 9.37 ± 0.19 9.24 ± 0.22 9.64 ± 0.35 0.4086

Complement C3 21.20 ± 0.60 19.46 ± 0.43 18.81 ± 0.50 19.69 ± 0.81 0.0348

Clusterin 1.49 ± 0.03 1.46 ± 0.03 1.44 ± 0.03 1.45 ± 0.05 0.6359

Plasminogen 1.22 ± 0.03 1.15 ± 0.02 1.12 ± 0.02 1.12 ± 0.04 0.0767


Vitronectin 3.64 ± 0.08 3.43 ± 0.06 3.27 ± 0.07 3.36 ± 0.11 0.0509

107


Prothrombin 0.59 ± 0.01 0.56 ± 0.01 0.55 ± 0.01 0.57 ± 0.02 0.2624

1 Values shown are crude means ± standard errors. Proteins are listed alphabetically and in order

of statistical significance among women HC users. P-values from ANCOVA with loge- or

square-root transformed plasma protein concentrations where necessary, adjusted for age,

ethnicity, waist circumference, physical activity, and season. Different superscript letters indicate

significant differences between groups (p<0.05). The Tukey-Kramer procedure was used to

adjust for multiple comparisons between groups within each ANCOVA.

§: These p-values meet the Bonferroni level of significance (p<0.0007; α=0.05, 25 independent

tests x 3 test groups).

108

Table 5.4. Mean plasma concentrations (μmol/L) of profile 1 proteins that were associated with 25(OH)D in women using <1mg vs.

≥1mg total hormone per day1,2


Hormone dose <1mg/day Hormone dose ≥1mg/day

<30

(n=11)

30 to 49.9

(n=20)

50 to 74.9

(n=44)

≥75

(n=57) p

<30

(n=3)

30 to 49.9

(n=8)

50 to 74.9

(n=13)

≥75

(n=22) p

Vitamin D binding protein 3.33 ± 0.21a 3.15 ± 0.13a 3.61 ± 0.12a 4.07 ± 0.11b 0.0002§ 4.34 ± 0.90 3.64 ± 0.17 3.46 ± 0.23 3.82 ± 0.16 0.4897

Apolipoprotein C-III 2.71 ± 0.17a,b 2.44 ± 0.16a 2.94 ± 0.12b 3.27 ± 0.11b 0.0004§ 3.28 ± 0.38 2.41 ± 0.15 2.49 ± 0.23 3.74 ± 0.19 0.0046

Retinol-binding protein 1.10 ± 0.06a,b 1.01 ± 0.04a 1.13 ± 0.04a 1.28 ± 0.04b 0.0006§ 1.27 ± 0.37 1.07 ± 0.10 0.95 ± 0.07 1.23 ± 0.05 0.1191

1 Values shown are crude means ± standard errors. Proteins are listed in order of statistical significance among women using <1mg

total hormone per day.

2 The association between 25(OH)D and each of the 25 proteins included in proteomic profile 1 was examined in both groups of HC

users, but shown are only proteins that were significantly associated with 25(OH)D in either group. P-values from ANCOVA with

loge- or square-root transformed plasma protein concentrations where necessary, adjusted for age, ethnicity, waist circumference,

physical activity, and season. Different superscript letters indicate significant differences between groups (p<0.05). The Tukey-

Kramer procedure was used to adjust for multiple comparisons between groups within each ANCOVA. § These p-values meet the

Bonferroni level of significance (p<0.001; α=0.05, 25 independent tests x 2 test groups).

109

Figure 5.1. Mean plasma proteomic profile scores, by vitamin D status, by sex and HC use, in

the population as a whole and among women HC users, women HC non-users, and men.

Profile 1 consisted primarily of acute phase proteins. Profile 2 included primarily negative acute

phase proteins. Profile 3 comprised proteins involved in innate immunity and regulation of the

complement cascade. Finally, profile 4 consisted of proteins involved in coagulation. Values

shown are covariate-adjusted means ± standard errors. P-values from ANCOVA adjusted for

age, ethnicity, waist circumference, physical activity, season, and, in the population as a whole,

sex and HC use. Different superscript letters indicate significant differences between groups

(p<0.05). The Tukey-Kramer procedure was used to adjust for multiple comparisons between

groups within each ANCOVA.

110

5.5 Discussion

There is considerable interest in the potential protective role of vitamin D in cardiometabolic

disease (349). The beneficial effects of vitamin D are thought to occur partly through regulation

of gene transcription by the VDR. Two recent studies identified hundreds of genetic targets of

vitamin D action across the human genome (27;145). However, it is not known whether this

widespread genomic signature of vitamin D translates into downstream effects at the proteome

level. The present study is the first to examine the association between circulating 25(OH)D and

the plasma proteome. We identified a number of associations between positive acute phase

proteins and 25(OH)D; however, the majority of these associations were present only in women

HC users. Furthermore, after accounting for total daily hormone dose, only three proteins (DBP,

ApoC-III, and RBP4) were still significantly associated with 25(OH)D. We have previously

reported a confounding effect of HC use on the association between 25(OH)D and CRP, a well

established biomarker of systemic inflammation (331). In addition, use of HRT in post-

menopausal women has been found to modify the association between 25(OH)D and various

cancers (190;245). Together with these findings, the results of the present study suggest that sex

hormone-based medications may play an important role as confounders or effect modifiers of the

association between 25(OH)D and various disease-related outcomes. Given the widespread use

of these medications by women throughout the world, efforts to fully elucidate their effects on

the relationship between vitamin D and disease are warranted.

High-abundance components of the plasma proteome include a number of physiologically

important proteins, such as protease inhibitors, complement system members, apolipoproteins,

carrier proteins, and coagulation factors (57). The concentrations of many of these proteins

become altered during inflammation, and they are considered acute phase reactants that may be

useful as biomarkers of disease risk (59;120). Indeed, many of the 54 proteins included in the

proteomics panel used in the present study are acute phase reactants that represent both

traditional and novel disease biomarkers (207), and they belong to pathways that become

dysregulated during disease progression, such as hemostasis, lipid metabolism, and innate

immunity (119;348).

A large body of evidence suggests that vitamin D may have important anti-inflammatory and

immunomodulatory properties (248). The active form of the vitamin, 1,25(OH)2D, binds to the

111

VDR, which then binds to VDRE in cytokine-encoding genes, downregulating their transcription

and attenuating the pro-inflammatory cascades that trigger hepatic production of acute phase

proteins (28;319). Indeed, a number of studies have reported inverse associations between

circulating 25(OH)D and acute phase proteins, such as CRP and RBP4 (31;85). Given these

findings, the positive association between 25(OH)D and a plasma proteomic profile consisting of

positive acute phase proteins in the overall population is unexpected. However, as our analyses

stratified by sex and HC use indicate, the observed association seemed to be driven by HC use.

Use of HC is associated with high circulating concentrations of 25(OH)D (188;191;331). Most

contemporary HC consist of a combination of synthetic estrogen and progestins (192). The

effects of these medications on vitamin D metabolite concentrations are thought to be mediated

by estrogen via downregulation of CYP24A1, a 24-hydroxylase that plays a key role in vitamin

D metabolism by initiating the catabolism of both 25(OH)D and 1,25(OH)2D, and upregulation

of CYP27B1, the main 1-α-hydroxylase responsible for conversion of 25(OH)D to 1,25(OH)2D

(193;194). Decreased catabolism of both 25(OH)D and 1,25(OH)2D and increased conversion

from 25(OH)D to 1,25(OH)2D by estrogen may result in elevated vitamin D metabolite

concentrations in the blood. At the same time, HC use is associated with a number of metabolic

perturbations, including increased inflammation, impaired glucose metabolism, and altered blood

lipid and coagulation factor concentrations (337). In particular, ethinyl estradiol, the synthetic

estrogenic component of many HC, affects hepatic production of hemostasis-related proteins

involved in coagulation and fibrinolysis, as well as pro-inflammatory acute phase reactants

(206;337). These effects are thought to be, at least in part, the result of hepatic metabolism of

ethinyl estradiol, combined with its long half-life and chemical potency (337). We have recently

reported a widespread effect of HC use on the same 54 proteins surveyed in the present study,

with women HC users having higher concentrations of many of them than women HC non-users

(207). These effects of HC on multiple plasma proteins, combined with their effect on 25(OH)D

concentrations, are the likely explanation for the initially unexpected positive association

between 25(OH)D and positive acute phase proteins in women HC users. Indeed, after adjusting

for hormone dose, most of these associations were no longer significant.

Of the twelve proteins associated with 25(OH)D in women HC users as a whole, only three

(DBP, ApoC-III, and RBP4) were associated with 25(OH)D in women HC users taking <1mg

total hormone per day, and none of them were associated in HC users taking ≥1mg total hormone

112

per day. In all three cases, in the <1mg group, the protein concentrations were higher in women

with a higher vitamin D status. ApoC-III is a component of triglyceride-rich particles, such as

chylomicrons and VLDL, that inhibits triglyceride-rich particle hydrolysis (350), preventing their

uptake and contributing to the development of hypertriglyceridemia (351). In addition, levels of

ApoC-III are altered during the acute phase response, and ApoC-III has been shown to

upregulate pro-inflammatory cascades (352). DBP, which is produced hepatically, is the main

carrier of vitamin D metabolites in the circulation (121). RBP4 is secreted hepatically and

transports the main vitamin A metabolite, retinol, from the liver to target tissues (353). RBP4 is

also secreted from adipocytes and is considered a pro-inflammatory adipokine (353). HC use is

generally associated with high concentrations of all three of these proteins, and we recently

reported a dose-dependent association between HC use and 25(OH)D concentrations

(197;331;353;354). Therefore, the observed associations between these proteins and 25(OH)D in

HC users taking <1mg total hormone may still be driven by dose-dependent effects within that

group that our study design was unable to capture.

Although no association was observed between any proteomic profile and 25(OH)D among

women HC non-users, analysis of individual proteins revealed positive associations between

RBP4 and coagulation factor XIIa heavy chain and 25(OH)D. No association existed between

coagulation factor XIIa heavy chain and 25(OH)D in women HC users. Among men, we

identified no significant associations between any proteins and 25(OH)D. The lack of association

between 25(OH)D and DBP in women HC non-users and men may seem surprising, because

genetic polymorphisms in GC, the gene encoding this protein, are associated with 25(OH)D

concentrations (222;223). However, while nearly 90% of circulating 25(OH)D is bound to DBP,

the concentration of the latter greatly exceeds that of 25(OH)D (133), so that changes in vitamin

D status may not necessarily be paralleled by fluctuations in DBP concentration.

The observed positive association between 25(OH)D and RBP4 across all women, regardless of

HC use, is inconsistent with the results of a previous trial where vitamin D supplementation in

individuals with T2D resulted in decreased concentrations of RBP4. In both adults and

adolescents, circulating RBP4 is positively associated with whole body adiposity, BMI, insulin

resistance and T2D, and it may predict CVD (355). In light of these findings, our observation of

a sex-specific positive association between 25(OH)D and RBP4 is unexpected and warrants

further study.

113

To our knowledge, the present study is the first to report an association between 25(OH)D and

coagulation factor XIIa heavy chain. The latter is a principal member of the intrinsic pathway,

which triggers the formation of fibrin (356). The link between vitamin D and coagulation has

been poorly explored. A small study found no improvements in coagulation-related parameters

with vitamin D supplementation (357). Despite our finding of a positive association between

coagulation factor XIIa heavy chain and 25(OH)D, we observed no association between

25(OH)D and other coagulation-related proteins included in the proteomics panel, suggesting an

overall lack of important effects of 25(OH)D on this pathway in healthy young adults.

Finally, we found no associations between 25(OH)D and either negative acute phase reactants

(proteomic profile 2) or innate immunity-related proteins (proteomic profile 3). We also

observed no association between 25(OH)D and any of the four proteins that did not load onto

any specific proteomic profiles, such as adiponectin, which is an anti-inflammatory cytokine. It

is possible that the anti-inflammatory and immunomodulatory effects of circulating vitamin D

are more evident in older populations with a greater degree of chronic inflammation, which may

be more easily captured through measurement of plasma biomarkers. In young, healthy

individuals, the anti-inflammatory effects of vitamin D may be localized to specific sites of

inflammation and, therefore, only be measurable in specific tissue regions.

One limitation of the present study is that, while the overall study population was large in size,

there were relatively few women HC users. The small sample size of this sub-group precluded

examining whether the effect of HC use on the association between 25(OH)D and plasma

proteins differs across different ethnic groups that have varying degrees of vitamin D deficiency,

as well as different concentrations of the assessed proteins (119;153). In addition, the present

study was not designed to examine whether the estrogenic or progesterone-related compounds in

HC drive the associations reported here, nor whether endogenous differences in sex hormones

play a role in the relationship between vitamin D and the plasma proteome. Despite these

limitations, our findings suggest that sex hormone-based medications represent a potentially

important confounder that should be considered in any studies examining the association

between vitamin D and biomarkers of cardiometabolic and other diseases. Adequately powered

prospective studies and clinical trials that examine the effect of various types of sex hormone-

based medications, as well as endogenous hormones, on the relationship between vitamin D and

specific outcomes are needed to elucidate how sex hormones may modulate the effects of

114

vitamin D. Such studies will allow for a better understanding of whether the confounding effects

of these drugs are specific to certain ethnic groups and whether certain components of the

medication (e.g. estrogen or progesterone-derived components), as well as an individual's

endogenous hormonal status, play a more important confounding role.

In summary, the present study examined the association between circulating 25(OH)D and the

plasma proteome and determined whether HC use modifies this association. While a number of

positive associations between 25(OH)D and positive acute phase proteins involved in various

pathways that become dysregulated during disease progression were identified, the majority were

no longer significant after accounting for HC use. These results suggest that HC use may play an

important role as a confounder in the association between 25(OH)D and proteomic biomarkers of

disease risk. Given the widespread use of these medications by women throughout the world,

efforts to fully elucidate their effects on the relationship between vitamin D and disease are

warranted.

115

Chapter 6 : Genetic variation in the vitamin D receptor, plasma 25-

hydroxyvitamin D, and biomarkers of cardiometabolic disease in

Caucasian young adults

Adapted with permission from:

García-Bailo B, Jamnik J, Da Costa L, Badawi A, El-Sohemy A. Genetic variation

in the vitamin D receptor, plasma 25-hydroxyvitamin D, and biomarkers of

cardiometabolic disease in Caucasian young adults. J Nutrigenet Nutrigenomics. In

press.

116

6.1 Abstract

Background: Vitamin D regulates gene transcription through binding of 1,25-dihydroxyvitamin

D (1,25(OH)2D) to the vitamin D receptor (VDR), potentially modulating physiologic processes

related to cardiometabolic disease. However, studies of the association between 25-

hydroxyvitamin D (25(OH)D) (which is the accepted biomarker of vitamin D status and is

converted to 1,25(OH)2D) and cardiometabolic disease have been inconsistent. These

inconsistencies may result from unaccounted gene-nutrient interactions between VDR and

circulating 25(OH)D. Our objective was to examine the effect of 25(OH)D on the association

between VDR variants and cardiometabolic disease biomarkers, including inflammation, glucose

and lipid metabolism, and the plasma proteome, in young adults.

Methods: The relationship between 27 VDR variants and each biomarker (biomarkers of

inflammation n=6; biomarkers of glucose and lipid metabolism n=9; plasma proteomic

biomarkers n=54) was examined in 488 Caucasians aged 20-29 years. General linear models

were used to examine the effect of the SNP x 25(OH)D interaction on each biomarker. For any

SNP-biomarker associations where SNP x 25(OH)D interaction p <0.05, the effect of vitamin D

status on that association was further examined with analysis of covariance, stratified by tertiles

of vitamin D.

Results: 25(OH)D affected the association between rs2239182 and interferon γ-induced protein

(IP-10) (p=0.0005). Among individuals in the lowest vitamin D tertile, heterozygotes had higher

IP-10 concentrations than minor and major allele homozygotes, suggesting a heterosis effect.

Conclusion: An association between rs2239182 and IP-10 was modified by 25(OH)D. The

association seemed indicative of heterosis, making its biological interpretation difficult, and it

may have been due to chance. Overall, in this population, vitamin D status does not seem to

interact with VDR variants to affect biomarkers of cardiometabolic disease.

117

6.2 Introduction

Interest in vitamin D as a therapeutic agent has grown exponentially over the last decade.

Beyond its central function as a regulator of calcium and bone metabolism, vitamin D may play a

role in the prevention of cardiometabolic diseases, such as CVD and T2D (19;20). Metabolic

abnormalities such as dyslipidemia and dysglycemia, which often cluster together with

hypertension and central adiposity, are a major risk factor for cardiometabolic disease (341).

Growing evidence also indicates that abnormal innate immune responses and chronic low-grade

inflammation play a key role in the development of cardiometabolic disease (3;5;7;8).

Evidence from in vitro and animal models indicates that vitamin D may play an important role as

a mediator of innate immunity and inflammatory responses (144;248). In addition, experimental

models suggest that vitamin D improves insulin sensitivity and secretion, regulates adipocyte

differentiation, and upregulates LPL activity in these cells (26;169;358;359). Through its actions

on these pathways, vitamin D may decrease inflammation and improve glucose and lipid

metabolism, thus contributing to a reduced risk of disease. A number of human studies have

reported inverse associations between 25(OH)D, which is the biomarker of vitamin D status, and

traditional biomarkers of inflammation, glycemic dysregulation, and dyslipidemia

(30;31;85;302). However, recent systematic reviews examining the association between vitamin

D and various cardiometabolic disease-related outcomes have yielded inconsistent findings

(35;36;360). These inconsistencies may be due to differences in methodology across studies, as

well as confounding from lifestyle and demographic factors, a lack of adequate biomarkers, and

unaccounted genetic variation between individuals.

25(OH)D is the most abundant vitamin D metabolite in the circulation and is representative of

vitamin D produced endogenously and obtained from dietary sources (121). Hydroxylation of

25(OH)D results in conversion to the main biologically active vitamin D metabolite, 1,25(OH)2D

(361). The latter exerts its biological effects by binding to the VDR, a transcription factor present

in tissues throughout the body (121). Together with other molecules, 1,25(OH)2D and the VDR

form a complex that binds to VDREs in target genes, regulating their transcription (361). Recent

in vitro work examining VDR binding sites throughout the genome has identified hundreds of

potential vitamin D target genes, primarily along innate immune and signaling pathways, many

of which are associated with diseases such as type 1 diabetes, Crohn's disease, colorectal cancer,

118

and multiple sclerosis (27;145). However, despite this widespread genomic signature, the

downstream effects of vitamin D at the level of the proteome remain poorly explored. The

plasma proteome is considered the largest version of the human proteome, with over 3,000

plasma proteins identified (58). Of these, a small number represent 99% of the total plasma

protein mass (57). These high-abundance proteins include a number of physiologically important

molecules, such as coagulation factors, complement system components, apolipoproteins, carrier

proteins, and protease inhibitors, many of which are acute phase reactants whose levels become

altered during inflammatory processes (59;120). Given the potential importance of plasma

proteins as biomarkers of disease, understanding their relationship with vitamin D may lead to a

better understanding of this vitamin's role in biological pathways that become dysregulated

during disease progression.

A number of variants exist within the VDR gene, and they may modify the effects of vitamin D

on disease-associated pathways across individuals. Traditionally, most studies have assessed

only a few polymorphisms (e.g. BsmI, FokI) (362;363), and the potential effects of other VDR

variants on disease-related outcomes remain poorly explored. Furthermore, few studies have

considered the vitamin D status of the study subjects. Recently, a large, long-term prospective

study identified a novel association between a VDR polymorphism, rs7968585, and a composite

of risk for hip fracture, cancer, myocardial infarction, and mortality (38). This association was

only apparent in individuals with a low vitamin D status, highlighting the importance of taking

into account gene-environment interactions.

Investigating the relationship between circulating 25(OH)D, VDR genetic variants, and

biomarkers of cardiometabolic disease may yield a better understanding of the biological roles of

vitamin D and help explain some of the inconsistent results across studies of the relationship

between this vitamin and cardiometabolic disease. The objective of the present study was to

examine the effect of vitamin D status on the association between variants across the VDR gene

and biomarkers of inflammation, biomarkers of glucose and lipid metabolism, and plasma

proteomic biomarkers in healthy young adults.

119

6.3 Methods


For details on the TNH study population, recruitment, exclusions, and classification into ethnic

groups and seasons, please refer to Chapter 2, Section 2.3.1.

The present study consisted of 488 non-smoking Caucasian men and women without diabetes,

with available measurements for circulating 25(OH)D and plasma proteomic biomarkers, and

data from genome-wide scans at the time that the study was carried out.






Please refer to Chapter 2, Section 2.3.5; Chapter 3, Section 3.3.3; and Chapter 4, Section 4.3.4.

6.3.5 Plasma Proteomic Measurements


6.3.6 Genotyping and candidate SNP selection

Genome-wide scans were carried out by Gene Logic (Houston, TX, United States), using the

Affymetrix (Santa Clara, CA, United States) 6.0 chip. After initial quality control measures

(excluding SNPs with call rate <95%, Hardy-Weinberg equilibrium p <10 x 10-8

, associated with

sex, monoallelic in the study sample), a total of 38 VDR SNPs were available on the chip. Of

these, 26 tag SNPs representative of variation across the VDR gene were selected based on the

following criteria: minor allele frequency (MAF)>0.01, Hardy-Weinberg equilibrium p>0.001,

and linkage disequilibrium r2<0.8. The selected SNPs are shown in Table 6.1. In addition, the

candidate SNP rs7968585, which was recently found to modify the association between

25(OH)D and a composite of several disease outcomes (38), was genotyped in the study

120

population using the Sequenom MassArray multiplex method, as described previously (364).

This SNP was not in linkage disequilibrium with any of the other 26 VDR SNPs.


Statistical analyses were carried out using SAS version 9.2. Alpha was set at 0.05 and p-values

are two-sided. Non-normally distributed continuous variables were loge- or square root-

transformed prior to analysis, so as to improve normality. In such cases, p-values from models

using transformed values are reported, but untransformed means and measures of spread are

shown for ease of interpretation.

Subjects were categorized into plasma 25(OH)D tertiles, and subject characteristics were

compared across tertiles using χ2 tests and ANOVA. We then examined the effect of 25(OH)D

on the association between each of the 27 VDR SNPs and each biomarker. The analyses were

conducted separately for each group of biomarkers (inflammation, glucose and lipid metabolism,

and plasma proteomic biomarkers). First, general linear models were constructed to examine the

effect of the interaction between 25(OH)D and each SNP on each biomarker. In each case, the

model included the biomarker as the outcome variable and an interaction term for SNP x

25(OH)D as the predictor, with age, sex, waist circumference, physical activity, season, and HC

use among women as covariates. In addition, models examining HOMA-Beta also were adjusted

for the ratio of circulating triglycerides to HDL cholesterol, as a surrogate measure of insulin

sensitivity (104;344). Biomarker-SNP associations where the models yielded p<0.05 for the SNP

x 25(OH)D interaction term were carried forward to the second step of the analysis. No multiple

testing criteria were applied at this point because these analyses were considered exploratory and

their purpose was to generate candidate associations for further analysis in the second step.

For each of the identified candidate associations, we then conducted analyses stratified by

vitamin D tertile that compared mean levels of the specific biomarker across genotypes for the

SNP with which it was associated. ANCOVA adjusted for age, sex, waist circumference,

physical activity, season, and HC use among women was used. As before, models examining

HOMA-Beta also were adjusted for the ratio of circulating triglycerides to HDL cholesterol. At

this stage, the Bonferroni (=α/number of tests) method was used as a multiple testing correction

threshold. Finally, for any SNP-biomarker associations identified in this second stage that were

modified by vitamin D status, we also compared mean levels of the biomarker across vitamin D

121

categories within each genotype. ANCOVA, with the same covariates listed earlier, was used to

conduct these analyses.

6.4 Results

Subject characteristics are shown in Table 6.2. A greater percentage of men were in the lowest

25(OH)D tertile than women, and a greater proportion of women who reported HC use were in

the highest vitamin D tertile than women who did not use these medications. Individuals with the

highest vitamin D status had lower BMI, waist circumference, and systolic blood pressure, and

reported higher physical activity, than those in the lowest vitamin D tertile.

As shown in Table 6.3a, we identified a total of 10 associations between seven SNPs and six

biomarkers of inflammation where the interaction term for SNP x 25(OH)D was significant

(p=0.0054 to p=0.0461). A total of 162 potential associations (27 SNPs and 6 inflammatory

biomarkers) were examined. One SNP, rs2239186, had a significant interaction with 25(OH)D

for three biomarkers, and another, rs2248098, had a significant interaction for two biomarkers.

An additional five SNPs each had significant interactions with 25(OH)D for one biomarker each.

We identified 30 associations between 13 SNPs and eight biomarkers of glucose and lipid

metabolism with significant SNP x 25(OH)D interaction terms (p=0.0018 to p=0.0474) (Table

6.3b). A total of 243 potential associations (27 SNPs and 9 biomarkers of glucose and lipid

metabolism) were examined. The interaction terms with the lowest p-values were for the

associations between rs3819545 and fasting insulin, HOMA-Beta and HOMA-IR (p=0.0018,

0.0019 and 0.0021, respectively). No SNPs were associated with triglycerides.

Finally, as shown in Table 6.3c, we observed a total of 52 associations between 16 SNPs and 31

plasma proteomic biomarkers with significant SNP x 25(OH)D interaction terms (p=0.0008 to

p=0.0499). A total of 1,458 potential associations (27 SNPs and 54 plasma proteomic

biomarkers) were examined. The interaction term with the lowest p-value was for rs2228570,

commonly referred to as FokI, and circulating levels of retinol binding protein. This SNP also

had the greatest number of significant interactions with 25(OH)D, for a total of nine.

122

For each of the associations identified in Tables 6.3a-c, we further investigated the effect of

25(OH)D on the relationship between each SNP and biomarker by comparing mean levels of the

biomarker across genotypes, separately within each vitamin D tertile group. Of the candidate

associations between VDR SNPs and biomarkers of inflammation identified in Table 6.3a, after

correcting for multiple testing, we found an association significant at the Bonferroni level

(p<0.0017, based on α=0.05 and 30 independent tests [10 SNP-inflammatory biomarker

associations x 3 vitamin D tertiles]) between rs2239182 and the inflammatory biomarker IP-10

among individuals in the lowest vitamin D tertile (Figure 6.1). Heterozygotes had higher IP-10

concentrations than either group of homozygotes. No association between rs2239182 and IP-10

was seen, however, among those in the intermediate and high vitamin D tertiles. In these

analyses stratified by vitamin D tertile, none of the candidate associations listed in Tables 6.3b-c

were significant after adjustment for multiple comparisons.

For the association between rs2239182 and IP-10 identified in Figure 6.1, we then examined

potential differences in concentrations of this cytokine across vitamin D tertiles within each

genotype. We noted that IP-10 concentrations were higher in the lowest vitamin D tertile than the

middle and high tertiles among heterozygotes (p=0.0259). Minor allele homozygotes had higher

IP-10 concentrations in the middle than the lowest vitamin D tertile (p=0.0396). Among major

allele homozygotes, we saw no differences in IP-10 between vitamin D categories (p=0.0887).

123

Table 6.1. VDR SNPs selected for analysis1

Position (Assembly:

GRCh37) rs ID

n MAF

48295133 rs12581281 487 0.01

48272515 rs11574049 486 0.03

48235738 rs2853563 488 0.04

48273814 rs11574046 484 0.09

48256046 rs2239180 488 0.12

48231430 rs12721364 487 0.15

48255859 rs2283342 488 0.15

48270596 rs6580642 488 0.16

48269410 rs2239186 484 0.20

48272275 rs11168275 487 0.23

48294626 rs7136534 488 0.24

48255570 rs2107301 488 0.27

48277713 rs2238136 488 0.31

48296486 rs4760658 488 0.32

48273714 rs2254210 488 0.33

48269650 rs10875693 487 0.35

48282805 rs2853559 485 0.35

48263828 rs2189480 488 0.35

48272895 rs2228570 487 0.36

48265006 rs3819545 487 0.38

124

48259126 rs12717991 483 0.38

48239835 rs1544410 485 0.39

48232747 rs2525046 487 0.42

48231898 rs10783215 487 0.49

48255411 rs2239182 488 0.49

48231843 rs7968585 482 0.49

48253356 rs2248098 488 0.50

1 SNPs are listed in order based on minor allele frequency, from lowest to highest. 26 tag SNPs

representative of variation across the VDR gene were selected based on the following criteria:

minor allele frequency (MAF)>0.01, Hardy-Weinberg equilibrium p>0.001, and linkage

disequilibrium r2<0.8. An additional candidate VDR SNP, rs7968585, previously associated

with a composite of risk for hip fracture, cancer, myocardial infarction, and mortality (38) was

also selected for analysis.

125

Table 6.2. Study participant characteristics, by vitamin D tertile1

Plasma 25(OH)D

Tertile 1 Tertile 2 Tertile 3 p

25(OH)D range (nmol/L) 13.3 - 50.7 50.8 - 72.3 72.6 - 194.0

n 162 163 163

Sex

Male (n=142) 63 (44.4) 49 (34.5) 30 (21.1) 0.0002

Female (n=346) 99 (28.6) 114 (33.0) 133 (38.4)

Hormonal contraceptive use among women

No (n=192) 70 (36.5) 74 (38.5) 48 (25.0) <0.0001

Yes (n=154) 29 (18.8) 40 (26.0) 85 (55.2)

Age (years) 23.1 ± 2.7 23.3 ± 2.6 23.3 ± 2.3 0.7769

BMI (kg/m2) 23.9 ± 4

a 23.1 ± 3.3

a,b 22.6 ± 2.9

b 0.0023

Waist circumference (cm) 77.8 ± 10a 75.2 ± 7.4

b 73.2 ± 7

b <0.0001

126

Physical Activity (met-h/wk) 7.6 ± 3a 8.2 ± 3.3

a,b 8.5 ± 2.8

b 0.0496

Systolic blood pressure (mm Hg) 116.6 ± 11.4a 116.4 ± 12

a 112.8 ± 9.7

b 0.0023

Diastolic blood pressure (mm Hg) 69.7 ± 7.9 69.7 ± 8.5 68.3 ± 7.2 0.2024

1 Values shown are crude means ± standard deviations or n (%). P-values calculated from χ

2 analysis for categorical variables and

ANOVA for continuous variables, with loge- or square root-transformed values where necessary. Different superscript letters indicate

significant differences between groups (p<0.05). The Tukey-Kramer procedure was used to adjust for multiple comparisons between

groups within each ANOVA.

127

Table 6.3a. Associations between VDR SNPs and biomarkers of inflammation where a

significant (p<0.05) SNP x 25(OH)D interaction term existed1

VDR SNP Biomarkers of inflammation p for SNP x 25(OH)D

interaction

rs2239186 RANTES 0.0054

PDFG-bb 0.0105

IL1RA 0.0278

rs2248098 IFN-γ 0.0115

CRP 0.0247

rs2239182 IP-10 0.0175

rs2853563 PDFG-bb 0.0210

rs3819545 IFN-γ 0.0245

rs2228570 IP-10 0.0415

rs11574046 IP-10 0.0461

Total number of significant (p<0.05) interaction terms = 10

1 Results are listed by number of significant interactions per SNP and in order of statistical

significance. Only associations where p for interaction between SNP and 25(OH)D<0.05 are

shown. P-values are from ANCOVA with a SNP x 25(OH)D interaction term and age, sex, waist

circumference, physical activity, season, and HC use as covariates. Multiple testing corrections

were not applied at this stage. Any SNP-biomarker associations where the interaction term was

p<0.05 were carried forward as candidates for subsequent analysis. A total of 162 potential

associations (27 SNPs x 6 biomarkers of inflammation) were examined.

128

Table 6.3b. Associations between VDR SNPs and biomarkers of glucose and lipid metabolism

where a significant (p<0.05) SNP x 25(OH)D interaction term existed1

VDR SNP Biomarkers of glucose and

lipid metabolism

p for SNP x 25(OH)D

interaction

rs12721364 Total cholesterol 0.0150

Insulin 0.0346

HOMA-IR 0.0359

HOMA-Beta 0.0436

LDL 0.0474


Insulin 0.0120

HOMA-IR 0.0124

HOMA-Beta 0.0068

rs3819545 Insulin 0.0018

HOMA-Beta 0.0019

HOMA-IR 0.0021

rs1544410 LDL 0.0042

HOMA-Beta 0.0166

Total cholesterol 0.0166

rs12581281 Glucose 0.0025

HOMA-Beta 0.0429

rs6580642 HOMA-IR 0.0046

Insulin 0.0052

129

rs10783215 LDL 0.0083

Total:HDL cholesterol ratio 0.0158

rs7968585 Total:HDL cholesterol ratio 0.0143

LDL 0.0056

rs2189480 HOMA-IR 0.0190

Insulin 0.0186

rs2248098 HOMA-Beta 0.0265

LDL 0.0433

rs12717991 Glucose 0.0063

rs11168275 HOMA-Beta 0.0369






circumference, physical activity, season, and HC use as covariates. In n the case of HOMA-Beta,

triglyceride/HDL cholesterol ratio was also included as a surrogate measure of insulin

sensitivity. Multiple testing corrections were not applied at this stage. Any SNP-biomarker

associations where the interaction term was p<0.05 were carried forward as candidates for

subsequent analysis. A total of 243 potential associations (27 SNPs x 9 biomarkers of glucose

and lipid metabolism) were examined.

130

Table 6.3c. Associations between VDR SNPs and plasma proteomic biomarkers where a

significant (p<0.05) SNP x 25(OH)D interaction term existed1

VDR SNP Plasma proteomic biomarkers p for SNP x 25(OH)D

interaction

rs2228570 Retinol binding protein 0.0008

Afamin 0.0036

Apolipoprotein AII 0.0092

Transthyretin 0.0112

Apolipoprotein CIII 0.0113

Apolipoprotein CI 0.0195

Complement C3 0.0353

α2-HS-glycoprotein 0.0425

Antithrombin-III 0.0473

rs2238136 Apolipoprotein E 0.0065

Transferrin 0.0076


Plasminogen 0.0122

Albumin 0.0152


Afamin 0.0289

α1-Antichymotrypsin 0.0345

Prothrombin 0.0499

rs12581281 Fibronectin 0.0211

131

Fibrinogen α2 chain 0.0246

Fibrinogen β chain 0.0353

Fibrinopeptide A 0.0371

β2-glycoprotein 0.0416

rs2239182 α1B-glycoprotein 0.0155

Apolipoprotein AII 0.0174



rs10783215 β2-glycoprotein 0.0224

Gelsolin, isoform 1 0.0330


α1B-glycoprotein 0.0439

rs6580642 Apolipoprotein D 0.0410

Histidine-rich glycoprotein 0.0442

Fibronectin 0.0446

Fibrinogen β chain 0.0478

rs1544410 Apolipoprotein AI 0.0109

Complement C1 inactivator 0.0165

Apolipoprotein B100 0.0385

rs3819545 Albumin 0.0200

Transthyretin 0.0224


rs7968585 Gelsolin, isoform 1 0.0318

132



rs10875693 α2-macroglobulin 0.0027

Complement C1 inactivator 0.0275

rs2239180 Fibronectin 0.0078

rs2283342 β2-glycoprotein 0.0122

rs2853563 Ceruloplasmin 0.0290

rs2239186 Coagulation factor XIIa HC 0.0297

rs12721364 Apolipoprotein B100 0.0326

rs2189480 Heparin cofactor II 0.0372





circumference, physical activity, season, and HC use as covariates. Multiple testing corrections

were not applied at this stage. Any SNP-biomarker associations where the interaction term was

p<0.05 were carried forward as candidates for subsequent analysis. A total of 1,458 potential

associations (27 SNPs x 54 plasma proteomic biomarkers) were examined.

133

Figure 6.1. The association between rs2239182 and IP-10, an inflammatory cytokine, is modified

by vitamin D status.

Shown are crude means and standard errors across genotypes, stratified by 25(OH)D tertile. P-

values from ANCOVA, stratified by vitamin D tertile and adjusted for age, sex, waist

circumference, season, physical activity, and HC use among women.

134

6.5 Discussion

To our knowledge, the present study is the first to conduct a comprehensive examination of the

interplay between vitamin D status and VDR genetic variation along several cardiometabolic

disease-associated pathways simultaneously. In particular, we identified an association between

rs2239182 and IP-10 that appeared to be mediated by vitamin D status. Among individuals with

a low vitamin D status, heterozygotes had higher values of IP-10 than both major and minor

allele homozygotes. However, concentrations of IP-10 were similar across genotypes in the

higher vitamin D status categories. Among heterozygotes, those in the lowest vitamin D tertile

had higher IP-10 concentrations. However, minor allele homozygotes had higher concentrations

in the middle than the lowest vitamin D tertile, and major allele homozygotes had similar IP-10

concentrations regardless of their vitamin D status.

IP-10 is a chemoattractant for cells of the immune system whose transcription is upregulated by

IFN-γ (365). In vitro, 1,25(OH)D decreases expression of IP-10 (366). This is in line with

experimental evidence suggesting that 1,25(OH)2D downregulates the production of pro-

inflammatory cytokines such as IFN-γ and IL-2 and upregulates the transcription of

immunosuppressive cytokines such as IL-10 (250-253). Previously, the variant rs2239182 has

been associated with myopia (367). This SNP, which yields an A to G nucleotide substitution,

resides within intron 7 of the VDR gene, located on chromosome 12q13.1 and comprising 14

exons (367;368). The minor allele frequency observed in our population, 0.49, is similar to that

previously reported in Caucasians (367). It is not known whether rs2239182 is a functional

variant.

The results of the present study suggest a heterosis effect of rs2239182 on IP-10 that is apparent

only in individuals with a low vitamin D status, where heterozygotes with low circulating

25(OH)D may have greater inflammation, as indicated by IP-10 levels, than either group of

homozygotes. At the molecular level, heterosis takes place when heterozygotes for a particular

variant exhibit a greater or lesser phenotypic effect than either group of homozygotes for a

specific quantitative trait (369). This may seem counterintuitive, as one might expect

heterozygotes to exhibit a phenotype intermediate between the two homozygote groups. Indeed,

the existence of molecular heterosis has been debated; however, some instances of molecular

heterosis have been reported in humans (369). The mechanisms by which heterosis might occur

135

at the molecular level remain poorly understood, but may involve factors such as protein subunit

interaction, lethality associated with being homozygous for the risk allele, a greater range of

adaptive responsiveness in heterozygotes, or the influence of an independent third factor (369).

In order to determine the biological plausibility of the heterosis effect observed in the present

study, further research would be needed to examine the potential molecular mechanisms through

which this effect might act. However, it is also possible that the observed result is due to chance.

Accumulating evidence over the past decade suggests that, beyond its traditional role in bone

metabolism and calcium homeostasis, vitamin D has important functions as a modulator of

immune function and inflammation (144;248). Previous studies have also reported inverse

associations between biomarkers of glycemic dysregulation and 25(OH)D (30;182;290),

suggesting a role for vitamin D in maintaining or improving glucose control. Some evidence also

indicates a potential role for this micronutrient in adipocyte differentiation and regulation of LPL

activity, suggesting an involvement in lipid metabolism (169;358). Finally, two recent studies

identified hundreds of genetic targets of vitamin D action across the human genome, suggesting

widespread effects across numerous physiologic pathways, in particular those associated with

innate immunity (27;145). However, despite examining a large number of biomarkers, the

present study found little evidence for a relationship between circulating 25(OH)D, VDR genetic

variants and these physiologic pathways. It is possible that the observed lack of effects is due to

the young age and relative good health of our study population. These relationships may be more

easily observed in older adults with overt cardiometabolic disease, where circulating plasma

biomarkers may better reflect metabolic dysregulation.

A recent study identified a VDR variant, rs7968585, that predicted risk for hip fracture, cancer,

myocardial infarction, and mortality among older Caucasian individuals with a low vitamin D

status (38). We observed no relationships between this SNP, circulating 25(OH)D and any of the

biomarkers examined. It is possible that the effects of this variant are only apparent in the

context of severe outcomes such as those examined in (38). Nonetheless, studies such as that

(38) highlight the importance of considering gene-environment interactions in the context of

vitamin D and disease. Indeed, one recent study reported an effect of FokI, a traditional VDR

restriction fragment length polymorphism, on the association between vitamin D

supplementation and improved insulin sensitivity in a small cohort of South Asian women living

in New Zealand (370). Another study reported that FokI modifies the response to vitamin D

136

supplementation in subjects with T2D, where those who were homozygous for the risk allele saw

lesser improvements in levels of inflammatory biomarkers than heterozygotes and non-risk allele

homozygotes (268). Given these findings, one important strength of the present study is its

consideration of the joint effects of genetic variation and vitamin D status on the assessed

biomarkers. Furthermore, we examined variants across VDR, rather than focusing exclusively on

a few commonly assessed polymorphisms. This approach allows for the identification of

potential non-traditional VDR SNPs that may be implicated in disease risk.

In addition to the limitations already mentioned, the present study has several other drawbacks.

The VDR gene encodes the transcription factor that mediates the majority of the biological

effects of vitamin D and, as such, variants within this gene may play important roles in the

relationship between vitamin D and disease-related processes. However, it is possible that

variation in other genes along the vitamin D pathway may affect some of the relationships

examined here. The genetic variants chosen for analysis were drawn from those available in the

Affymetrix 6.0 chip. Thus, other potentially important SNPs not present in that chip may have

been excluded. The study population consisted only of Caucasian individuals and was relatively

small in size, especially considering the large number of SNPs and biomarkers examined. While

we corrected for multiple testing so as to minimize the chances of reporting spurious

associations, it is possible that the observed results are due to chance.

In summary, we examined the effect of 25(OH)D on the association between 27 variants across

VDR and cardiometabolic disease biomarkers, including inflammation, glucose and lipid

metabolism, and the plasma proteome, in young Caucasian adults. We identified an association

between the VDR variant rs2239182 and the inflammatory biomarker IP-10 that was modified by

25(OH)D. Among individuals with a low vitamin D status, heterozygotes had higher IP-10

concentrations than either group of homozygotes. This observation seemed indicative of

heterosis, making its biological interpretation difficult, and it may have been due to chance. No

other associations were significantly modified by vitamin D status. These results suggest that, in

this population of healthy young adults, vitamin D status does not seem to interact with VDR

variants to affect biomarkers of cardiometabolic disease.

137

Chapter 7 : Summary, Limitations, Future Directions and Implications

138

7.1 Summary

The overall goal of this thesis was to determine the association between 25(OH)D and

biomarkers of cardiometabolic disease risk, including biomarkers of inflammation, glycemic

dysregulation and lipid metabolism and plasma proteomic biomarkers, and to examine whether

lifestyle factors (specifically, HC use among women) and genetic variation in the vitamin D

pathway (in particular, genetic variation in VDR) modify these associations.

Objective 1: To examine the association between 25(OH)D and CRP, and to determine whether

HC use modifies this association.

Results: Across ethnic groups, women HC users had higher 25(OH)D and CRP than women HC

non-users and men (p<0.008 and <0.0001, respectively) (thesis chapter 2). Circulating 25(OH)D

was positively associated with CRP in the entire population in models not accounting for HC use

(β=0.010±0.003; p<0.0001). However, there was no association when men and women HC non-

users were examined separately. Among women HC users, there was no association after

accounting for hormone dose. A positive association between 25(OH)D and CRP among

individuals above the median 25(OH)D (≥51.9 nmol/L) was not significant after adjustment for

HC use. No association was observed among individuals below the median. These results

suggest that HC use confounds a positive association between 25(OH)D and CRP.

Objective 2: To examine the association between plasma 25(OH)D and pro-inflammatory

cytokines, and to determine whether HC use modifies this association.

Results: There were crude correlations between 25(OH)D, IP-10 (Pearson's r=0.12, p<0.0001),

and RANTES (r=0.22, p<0.0001), but no other cytokines (thesis chapter 3). However, after full

covariate adjustment, 25(OH)D was not associated with any cytokine. These results suggest that

25(OH)D is not associated with systemic biomarkers of inflammation in healthy young adults.

Objective 3: To examine the association between 25(OH)D and biomarkers of glucose and lipid

metabolism, and to determine whether HC use modifies this association.

Results: Plasma 25(OH)D concentrations fluctuated seasonally among Caucasians and East

Asians and across men, women HC non-users, and women HC users, but they remained low

year-round in South Asians, half of whom were vitamin D-deficient (thesis chapter 4). Vitamin

139

D deficiency was associated with higher insulin, HOMA-IR, and HOMA-Beta among

Caucasians and East Asians and among men and women HC non-users, and with higher

triglycerides among men only. No biomarkers were associated with 25(OH)D among South

Asians and women HC users, although non-significant trends were observed for higher markers

of glycemic dysregulation in those who were vitamin D-deficient. These results suggest that

vitamin D deficiency is associated with biomarkers of glycemic dysregulation; however, HC use

among women may attenuate this association.

Objective 4: To examine the association between 25(OH)D and plasma proteomic biomarkers,

and to determine whether HC use modifies this association.

Results: The 54 proteins included in the proteomics panel used in this study clustered into four

distinct plasma proteomic profiles (thesis chapter 5). A positive association was observed

between Profile 1, containing positive acute phase proteins, and 25(OH)D. In women HC users, a

J-shaped association existed between Profile 1 and 25(OH)D, but no associations existed in

women HC non-users and men. Twelve proteins were individually associated with 25(OH)D in

women HC users, but only two were associated with 25(OH)D in women HC non-users and no

associations were observed in men. After accounting for hormone dose among women HC users,

only three proteins were associated with 25(OH)D. These results suggest that HC use may be an

important confounder of the association between circulating 25(OH)D and numerous plasma

proteins.

Objective 5: To examine whether genetic variants in the VDR are associated with biomarkers of

inflammation, biomarkers of glucose and lipid metabolism, and plasma proteomic biomarkers,

and to determine whether vitamin D status modifies these associations.

Results: 25(OH)D affected the association between rs2239182 and IP-10 (p=0.0005) (thesis

chapter 6). Among individuals in the lowest vitamin D tertile, heterozygotes had higher IP-10

concentrations than minor and major allele homozygotes. The association was suggestive of

heterosis and may have been due to chance. No other vitamin D status x VDR interactions were

statistically significant. These results suggest that, in this population, 25(OH)D does not seem to

interact with VDR variants to affect biomarkers of cardiometabolic disease.

140

7.1.1 Relationship with Inflammation

Overall, the results of the present thesis suggest that 25(OH)D is not associated with biomarkers

of systemic inflammation, including CRP and various cytokines, in young adults. Furthermore,

initially we observed unexpected positive associations between 25(OH)D and two inflammatory

biomarkers, namely CRP and RANTES. As indicated by the results of chapters 2 and 3, these

associations were confounded by HC use among women, but genetic variation across VDR does

not appear to modify the association between 25(OH)D and any inflammatory biomarker

(chapter 6).

A wealth of experimental evidence supports a role for vitamin D in modulating inflammatory

processes. Indeed, in vitro and animal studies have shown that 1,25(OH)D2 downregulates the

production of pro-inflammatory cytokines such as IFN-γ and IL-2 and it upregulates the

transcription of immuno-supressive cytokines such as IL-10 (250-253). However,

epidemiological studies and clinical trials have yielded inconsistent results, with some suggesting

an inverse association between circulating concentrations of (25(OH)D) and pro-inflammatory

cytokines, acute phase proteins and hemostatic markers (85;254;256-259), while others found

null associations between 25(OH)D and markers of inflammation (241;261-266). Furthermore,

using NHANES data, one recent study reported a positive association between 25(OH)D and

CRP among vitamin D-sufficient adults (267). These inconsistencies between studies may be due

to unaccounted confounders.

The present work is the first to examine the modifying effect of HC use on the relationship

between 25(OH)D and inflammation. These medications have often been overlooked as

potentially important confounders of the relationship between 25(OH)D and various disease-

related outcomes. Use of HC has been linked to elevated 25(OH)D (188), a relationship thought

to be mediated, at least in part, by estrogen's downregulation of CYP24A1, the main catabolic

enzyme in the vitamin D pathway (190;193). However, most studies were conducted in

individuals of European or African ancestry, and the effects of these medications on 25(OH)D in

other ethnic groups have been less well understood. The present thesis found that women who

used HC had higher 25(OH)D across ethnic groups. At the same time, HC use has also been

associated with changes in numerous physiologic pathways, increased inflammation and a poorer

metabolic profile (198;199;203;206;207). Our finding that adjustment for HC use attenuates a

141

positive association between 25(OH)D and CRP, as well as RANTES, might partly account for

some of the unexpected results of previous studies.

Considering the wealth of experimental evidence for anti-inflammatory effects of vitamin D, the

lack of evidence for an inverse association between 25(OH)D and biomarkers of inflammation

may seem surprising. It is important to note that our study population consisted of healthy young

adults. It is possible that the anti-inflammatory effects of vitamin D may be more apparent in

older or more diseased populations with a greater burden of systemic inflammation that can be

more easily measured through biomarkers circulating in the blood. Indeed, several studies that

have reported beneficial effects of vitamin D on inflammation were conducted in obese

individuals or those with T2D (31;85). Furthermore, it is also possible that 25(OH)D

concentrations higher than those found among our study participants are necessary to observe its

effects on immune modulation at the systemic level.

7.1.2 Relationship with Glycemic Dysregulation

The present thesis shows an inverse association between 25(OH)D and biomarkers of glycemic

dysregulation, in particular insulin, HOMA-IR and HOMA-Beta (chapter 4). However, the

association was only significant in men and in women who did not use HC, as well as

Caucasians and East Asians. In women who used HC and South Asians, no significant

association was present. Furthermore, VDR variants did not appear to modify the relationship

between 25(OH)D and biomarkers of glycemic dysregulation (chapter 6).

The results of this thesis are in line with previous cross-sectional and prospective studies

showing an inverse relationship between circulating 25(OH)D and glycemic status measures

(30;183;280;281;285). Indeed, experimental evidence suggests that vitamin D may improve

insulin secretion and also upregulate the expression of insulin receptors, leading to increased

insulin sensitivity (26;272). In the present study, fasting insulin and HOMA-IR were inversely

associated with 25(OH)D, but fasting glucose measures were similar regardless of vitamin D

status. This observation suggests that, in this population of primarily healthy young adults,

individuals with low circulating 25(OH)D may be able to compensate for a reduced insulin

sensitivity through elevated insulin production, in order to maintain normal glucose

concentrations. However, as these individuals age, sustained elevated production of insulin by

the pancreas to counteract high circulating glucose may result in both insulin resistance and β-

142

cell dysfunction. Our results suggest that the onset of these conditions may already be underway

among vitamin D-deficient young adults.

To our knowledge, no other study has directly examined the association between vitamin D and

cardiometabolic disease biomarkers separately across women HC users, women HC non-users

and men. In addition to their previously mentioned effects on 25(OH)D concentrations, these

medications are associated with decreased insulin sensitivity in healthy, non-obese young women

(204;337;345). Overall, the results of this thesis suggest that, among women who use HC, the

potential benefits of elevated circulating 25(OH)D on glucose metabolism may be attenuated by

potential deleterious effects of HC on this pathway (204;207). However, it is important to note

that no statistical interaction existed between 25(OH)D and HC use for biomarkers of glucose

metabolism (chapter 4). It is possible that the lack of association between 25(OH)D and any

biomarkers in women HC users may have been due to the smaller sample size of this group.

Future studies with larger sample sizes are needed to further examine the potential effects of HC

use on the association between 25(OH)D and glucose metabolism.

The lack of association between 25(OH)D and markers of glucose metabolism in South Asians

may be a result of the small size of this subgroup, combined with their overall low vitamin D

status (chapter 4). Indeed, the results of the present thesis confirm previous observations that the

prevalence of vitamin D deficiency in individuals of this ethnic group living in Canada is very

high and their vitamin D status remains low year-round (153).

7.1.3 Relationship with Lipid Metabolism

Overall, the present thesis found no association between 25(OH)D and any markers of lipid

metabolism, except for an inverse relationship with triglycerides that was apparent in men only

(chapter 4). Genetic variation in VDR did not explain the observed lack of association between

25(OH)D and biomarkers of lipid metabolism (chapter 6).

Although some studies have noted inverse trends between 25(OH)D and a pro-atherogenic lipid

profile, the results have been inconsistent (239;291;306). Furthermore, while some mechanisms

have been proposed, the biological link between vitamin D and lipid metabolism remains

unclear. With respect to triglycerides, it is possible that 1,25(OH)2D, the biologically active

vitamin D metabolite, upregulates LPL activity in adipocytes, leading to decreased circulating

143

triglyceride concentrations (169). Whether biological differences between the sexes account for

the observed lack of association between 25(OH)D and triglycerides in women remains to be

ascertained.

7.1.4 Relationship with the Plasma Proteome

After accounting for HC use, 25(OH)D was not associated with most of the 54 high-abundance

plasma proteins surveyed (chapter 5). Furthermore, genetic variation in VDR did not modify any

associations (chapter 6).

Recent in vitro studies have examined 1,25(OH)2D action across the genome, using gene




colorectal cancer. The research in this thesis is the first to examine whether these widespread

effects of vitamin D at the level of the genome translate into downstream effects at the level of

the plasma proteome. Our results suggest that HC use is an important confounder of the

relationship between 25(OH)D and the plasma proteome. Many of the proteins included in the

proteomics panel are acute phase reactants that belong to disease-associated pathways, such as

innate immunity, hemostasis and lipid metabolism. The observed overall lack of association

between 25(OH)D and the assessed proteins may reflect the young age and general good health

of the study population. Any effects of vitamin D may be more apparent in older or more

diseased populations, where there may be more pronounced differences in concentrations of the

assessed biomarkers across vitamin D status categories.

7.2 Limitations

The present thesis has a number of limitations. First, the cross-sectional study design precludes

drawing any conclusions about causality based on the observed associations. Analyses within

specific subgroups, in particular women HC users and South Asians, may be limited by the small

sample size of these groups. Alternatively, given the large number of independent tests

conducted in some of the thesis chapters, the observed associations may have been due to

144

chance. We attempted to minimize this possibility by using the Bonferroni correction for

multiple testing whenever necessary.

We were unable to assess whether estrogen or progesterone-derived HC components have

different effects on the associations examined here, and we also lacked information on

endogenous sex hormone concentrations. While we measured a comprehensive set of biomarkers

along several cardiometabolic disease-associated pathways, it is possible that 25(OH)D may be

associated with other biomarkers not assessed in the present thesis. Furthermore, in addition to

VDR, it is possible that variation in other genes along the vitamin D pathway may affect some of

the relationships examined here. Residual confounding from unidentified factors may also have

confounded the results presented here.

Finally, certain associations may have been more apparent in an older, more diseased population

with higher circulating concentrations of the assessed biomarkers. However, studying healthy

individuals provides a picture of physiologic status before the onset of disease, and

understanding how specific disease-associated pathways may be modulated by various

micronutrients at this stage provides the potential for developing nutrition-based strategies to

prevent later disease development.

7.3 Future Research

Our findings suggest that HC use among women may represent a potentially important

confounder that should be considered in any epidemiologic studies examining the association

between vitamin D and cardiometabolic disease. Adequately powered longitudinal studies and

clinical trials that examine the effect of various types of sex hormone-based medications, as well

as endogenous hormones, on the relationship between vitamin D and specific outcomes are

needed to further document how sex hormones modulate the effects of vitamin D. In addition,

experimental studies conducted in cell culture and animal models are needed to better understand

the mechanisms through which estrogen, and potentially other sex hormones, mediates the

relationship between vitamin D and various physiologic processes. An increasing number of

recent studies suggests that genetic variants along the vitamin D pathway may affect certain

disease outcomes (38;212;268). Indeed, a large, long-term prospective study recently identified a

145

novel association between a VDR variant and a composite of risk for hip fracture, cancer,

myocardial infarction, and mortality that was only apparent in individuals with a low vitamin D

status (38). Studies such as (38) highlight the importance of taking into account gene-

environment interactions when examining the relationship between vitamin D and disease. While

the present thesis found that no VDR variants, including the one assessed in (38), modified the

association between 25(OH)D and any biomarkers, large studies are needed to examine the

potential effects of variants in VDR and other vitamin D pathway genes on the relationship

between vitamin D status and various disease-related outcomes across ethnic groups.

7.4 Implications

The most important implication of this work is that it highlights the importance of accounting for

HC use among women in any studies examining the relationship between vitamin D and

biomarkers of cardiometabolic disease, as well as potentially other diseases. Our finding that HC

use may confound or attenuate several of the assessed relationships might help explain some of

the inconsistencies observed across previous studies.

HC are used by over 100 million women throughout the world, not only to prevent unwanted

pregnancies but also to treat various other conditions such as acne, polycystic ovarian syndrome

and endometriosis (204;205). Therefore, understanding the extent to which they might modify

the relationship between vitamin D and disease outcomes is of potentially great public health

importance.

The present thesis also lends further support to the increasingly accepted view that vitamin D

may play a beneficial role in certain non-skeletal health outcomes, namely glycemic regulation.

Furthermore, our work highlights the widespread prevalence of vitamin D deficiency in certain

ethnocultural groups in Canada, particularly South Asians, as many of 50% of whom were

vitamin D-deficient based on the IOM's definition of deficiency, which is arguably quite

conservative. Given the potential benefits of vitamin D on glucose metabolism that may already

be present in young adults, efforts to improve vitamin D status in sub-groups at risk of deficiency

are warranted and may have a potentially large public health impact.

146

7.5 Thesis Summary

In summary, the present thesis examined the relationship between 25(OH)D and biomarkers of

inflammation, glucose and lipid metabolism, and the plasma proteome, and assessed the potential

effects of HC use and genetic variants across VDR on these relationships. We identified a

confounding effect of HC use on the association between 25(OH)D, biomarkers of inflammation

and plasma proteomic biomarkers. In addition, HC use appeared to attenuate an inverse

association between 25(OH)D and biomarkers of glycemic dysregulation. Genetic variation in

VDR did not modify any associations.

147

References

1. Kraushaar LE, Kramer A. Are we losing the battle against cardiometabolic

disease? The case for a paradigm shift in primary prevention. BMC Public Health

2009;9:64.

2. Leiter LA, Fitchett DH, Gilbert RE, Gupta M, Mancini GB, McFarlane PA, Ross

R, Teoh H, Verma S, Anand S et al. Cardiometabolic risk in Canada: a detailed analysis

and position paper by the cardiometabolic risk working group. Can J Cardiol 2011;27:e1-

e33.

3. Barzilay JI, Abraham L, Heckbert SR, Cushman M, Kuller LH, Resnick HE,

Tracy RP. The relation of markers of inflammation to the development of glucose

disorders in the elderly: the Cardiovascular Health Study. Diabetes 2001;50:2384-9.

4. Adameova A, Xu YJ, Duhamel TA, Tappia PS, Shan L, Dhalla NS. Anti-

atherosclerotic molecules targeting oxidative stress and inflammation. Curr Pharm Des

2009;15:3094-107.

5. Hu FB, Meigs JB, Li TY, Rifai N, Manson JE. Inflammatory markers and risk of

developing type 2 diabetes in women. Diabetes 2004;53:693-700.

6. Krogh-Madsen R, Plomgaard P, Moller K, Mittendorfer B, Pedersen BK.

Influence of TNF-alpha and IL-6 infusions on insulin sensitivity and expression of IL-18

in humans. Am J Physiol Endocrinol Metab 2006;291:E108-E114.

7. Laaksonen DE, Niskanen L, Nyyssonen K, Punnonen K, Tuomainen TP,

Valkonen VP, Salonen R, Salonen JT. C-reactive protein and the development of the

metabolic syndrome and diabetes in middle-aged men. Diabetologia 2004;47:1403-10.

8. Jenny NS, Arnold AM, Kuller LH, Tracy RP, Psaty BM. Serum amyloid P and

cardiovascular disease in older men and women: results from the Cardiovascular Health

Study. Arterioscler Thromb Vasc Biol 2007;27:352-8.

9. Hage FG, Szalai AJ. C-reactive protein gene polymorphisms, C-reactive protein

blood levels, and cardiovascular disease risk. J Am Coll Cardiol 2007;50:1115-22.

10. Lopez-Garcia E, Schulze MB, Fung TT, Meigs JB, Rifai N, Manson JE, Hu FB.

Major dietary patterns are related to plasma concentrations of markers of inflammation

and endothelial dysfunction. Am J Clin Nutr 2004;80:1029-35.

11. Nettleton JA, Schulze MB, Jiang R, Jenny NS, Burke GL, Jacobs DR, Jr. A

priori-defined dietary patterns and markers of cardiovascular disease risk in the Multi-

Ethnic Study of Atherosclerosis (MESA). Am J Clin Nutr 2008;88:185-94.

12. Nettleton JA, Matijevic N, Folsom AR, Boerwinkle E. Associations between

dietary patterns and flow cytometry-measured biomarkers of inflammation and cellular

activation in the Atherosclerosis Risk in Communities (ARIC) Carotid Artery MRI

Study. Atherosclerosis 2010;212:260-7.

13. Brighenti F, Valtuena S, Pellegrini N, Ardigo D, Del RD, Salvatore S, Piatti P,

Serafini M, Zavaroni I. Total antioxidant capacity of the diet is inversely and

148

independently related to plasma concentration of high-sensitivity C-reactive protein in

adult Italian subjects. Br J Nutr 2005;93:619-25.

14. Detopoulou P, Panagiotakos DB, Antonopoulou S, Pitsavos C, Stefanadis C.

Dietary choline and betaine intakes in relation to concentrations of inflammatory markers

in healthy adults: the ATTICA study. Am J Clin Nutr 2008;87:424-30.

15. Esposito K, Marfella R, Ciotola M, Di PC, Giugliano F, Giugliano G, D'Armiento

M, D'Andrea F, Giugliano D. Effect of a mediterranean-style diet on endothelial

dysfunction and markers of vascular inflammation in the metabolic syndrome: a

randomized trial. JAMA 2004;292:1440-6.

16. Bakker GC, van Erk MJ, Pellis L, Wopereis S, Rubingh CM, Cnubben NH,

Kooistra T, van OB, Hendriks HF. An antiinflammatory dietary mix modulates

inflammation and oxidative and metabolic stress in overweight men: a nutrigenomics

approach. Am J Clin Nutr 2010;91:1044-59.

17. Lavie CJ, Lee JH, Milani RV. Vitamin D and cardiovascular disease will it live

up to its hype? J Am Coll Cardiol 2011;58:1547-56.

18. Maxmen A. Nutrition advice: the vitamin D-lemma. Nature 2011;475:23-5.

19. Rosen CJ, Adams JS, Bikle DD, Black DM, Demay MB, Manson JE, Murad MH,

Kovacs CS. The nonskeletal effects of vitamin D: an Endocrine Society scientific

statement. Endocr Rev 2012;33:456-92.

20. Vieth R, Bischoff-Ferrari H, Boucher BJ, Dawson-Hughes B, Garland CF,

Heaney RP, Holick MF, Hollis BW, Lamberg-Allardt C, McGrath JJ et al. The urgent

need to recommend an intake of vitamin D that is effective. Am J Clin Nutr 2007;85:649-

50.

21. Shapses SA, Manson JE. Vitamin D and prevention of cardiovascular disease and

diabetes: why the evidence falls short. JAMA 2011;305:2565-6.

22. Manson JE, Mayne ST, Clinton SK. Vitamin D and prevention of cancer--ready

for prime time? N Engl J Med 2011;364:1385-7.

23. Holick MF, Binkley NC, Bischoff-Ferrari HA, Gordon CM, Hanley DA, Heaney

RP, Murad MH, Weaver CM. Evaluation, treatment, and prevention of vitamin D

deficiency: an Endocrine Society clinical practice guideline. J Clin Endocrinol Metab

2011;96:1911-30.

24. Hanley DA, Cranney A, Jones G, Whiting SJ, Leslie WD, Cole DE, Atkinson SA,

Josse RG, Feldman S, Kline GA et al. Vitamin D in adult health and disease: a review

and guideline statement from Osteoporosis Canada. CMAJ 2010;182:E610-E618.

25. Maestro B, D+ívila N, Carranza MC, Calle C. Identification of a Vitamin D

response element in the human insulin receptor gene promoter. Journal of Steroid

Biochemistry and Molecular Biology 2003;84:223-30.

26. Maestro B, Campion J, Davila N, Calle C. Stimulation by 1,25-dihydroxyvitamin

D3 of insulin receptor expression and insulin responsiveness for glucose transport in U-

937 human promonocytic cells. Endocr J 2000;47:383-91.

149

27. Ramagopalan SV, Heger A, Berlanga AJ, Maugeri NJ, Lincoln MR, Burrell A,

Handunnetthi L, Handel AE, Disanto G, Orton SM et al. A ChIP-seq defined genome-

wide map of vitamin D receptor binding: associations with disease and evolution.

Genome Res 2010;20:1352-60.

28. van Etten E, Mathieu C. Immunoregulation by 1,25-dihydroxyvitamin D3: basic

concepts. J Steroid Biochem Mol Biol 2005;97:93-101.

29. Brenner DR, Arora P, Garcia-Bailo B, Wolever TM, Morrison H, El-Sohemy A,

Karmali M, Badawi A. Plasma vitamin D levels and risk of metabolic syndrome in

Canadians. Clin Invest Med 2011;34:E377.

30. Kayaniyil S, Retnakaran R, Harris SB, Vieth R, Knight JA, Gerstein HC, Perkins

BA, Zinman B, Hanley AJ. Prospective associations of vitamin D with beta-cell function

and glycemia: the PROspective Metabolism and ISlet cell Evaluation (PROMISE) cohort

study. Diabetes 2011;60:2947-53.

31. Neyestani TR, Nikooyeh B, avi-Majd H, Shariatzadeh N, Kalayi A, Tayebinejad

N, Heravifard S, Salekzamani S, Zahedirad M. Improvement of vitamin D status via daily

intake of fortified yogurt drink either with or without extra calcium ameliorates systemic

inflammatory biomarkers, including adipokines, in the subjects with type 2 diabetes. J

Clin Endocrinol Metab 2012;97:2005-11.

32. Nikooyeh B, Neyestani TR, Farvid M, avi-Majd H, Houshiarrad A, Kalayi A,

Shariatzadeh N, Gharavi A, Heravifard S, Tayebinejad N et al. Daily consumption of

vitamin D- or vitamin D + calcium-fortified yogurt drink improved glycemic control in

patients with type 2 diabetes: a randomized clinical trial. Am J Clin Nutr 2011;93:764-71.

33. Mitri J, wson-Hughes B, Hu FB, Pittas AG. Effects of vitamin D and calcium

supplementation on pancreatic beta cell function, insulin sensitivity, and glycemia in

adults at high risk of diabetes: the Calcium and Vitamin D for Diabetes Mellitus

(CaDDM) randomized controlled trial. Am J Clin Nutr 2011;94:486-94.

34. Bucharles S, Barberato SH, Stinghen AE, Gruber B, Piekala L, Dambiski AC,

Custodio MR, Pecoits-Filho R. Impact of cholecalciferol treatment on biomarkers of

inflammation and myocardial structure in hemodialysis patients without

hyperparathyroidism. J Ren Nutr 2012;22:284-91.

35. Pittas AG, Chung M, Trikalinos T, Mitri J, Brendel M, Patel K, Lichtenstein AH,

Lau J, Balk EM. Systematic review: Vitamin D and cardiometabolic outcomes. Ann

Intern Med 2010;152:307-14.

36. Elamin MB, bu Elnour NO, Elamin KB, Fatourechi MM, Alkatib AA, Almandoz

JP, Liu H, Lane MA, Mullan RJ, Hazem A et al. Vitamin D and cardiovascular outcomes:

a systematic review and meta-analysis. J Clin Endocrinol Metab 2011;96:1931-42.

37. Mitri J, Muraru MD, Pittas AG. Vitamin D and type 2 diabetes: a systematic

review. Eur J Clin Nutr 2011;65:1005-15.

38. Levin GP, Robinson-Cohen C, de B, I, Houston DK, Lohman K, Liu Y,

Kritchevsky SB, Cauley JA, Tanaka T, Ferrucci L et al. Genetic Variants and

Associations of 25-Hydroxyvitamin D Concentrations With Major Clinical Outcomes.

JAMA 2012;308:1898-905.

150

39. Lee DS, Chiu M, Manuel DG, Tu K, Wang X, Austin PC, Mattern MY, Mitiku

TF, Svenson LW, Putnam W et al. Trends in risk factors for cardiovascular disease in

Canada: temporal, socio-demographic and geographic factors. CMAJ 2009;181:E55-E66.

40. Canadian Diabetes Association. The prevalence and costs of diabetes. 2012.

http://www.diabetes.ca/documents/about-diabetes/PrevalanceandCost_09.pdf. Accessed

February 2013.

41. Chatterjee A, Harris SB, Leiter LA, Fitchett DH, Teoh H, Bhattacharyya OK.

Managing cardiometabolic risk in primary care: summary of the 2011 consensus

statement. Can Fam Physician 2012;58:389-201.

42. Public Health Agency of Canada. Economic Burden of Illness in Canada, 2000.

http://www.phac-aspc.gc.ca/cd-mc/cvd-mcv/cvd_ebic-mcv_femc-eng.php. Accessed

February 2013.

43. Reaven GM. Role of insulin resistance in human disease (syndrome X): an

expanded definition. Annu Rev Med 1993;44:121-31.

44. Alberti KG, Eckel RH, Grundy SM, Zimmet PZ, Cleeman JI, Donato KA,

Fruchart JC, James WP, Loria CM, Smith SC, Jr. Harmonizing the metabolic syndrome:

a joint interim statement of the International Diabetes Federation Task Force on

Epidemiology and Prevention; National Heart, Lung, and Blood Institute; American

Heart Association; World Heart Federation; International Atherosclerosis Society; and

International Association for the Study of Obesity. Circulation 2009;120:1640-5.

45. Executive Summary of The Third Report of The National Cholesterol Education

Program (NCEP) Expert Panel on Detection, Evaluation, And Treatment of High Blood

Cholesterol In Adults (Adult Treatment Panel III). JAMA 2001;285:2486-97.

46. Grundy SM, Brewer HB, Jr., Cleeman JI, Smith SC, Jr., Lenfant C. Definition of

metabolic syndrome: Report of the National Heart, Lung, and Blood Institute/American

Heart Association conference on scientific issues related to definition. Circulation

2004;109:433-8.

47. Alberti KG, Zimmet PZ. Definition, diagnosis and classification of diabetes

mellitus and its complications. Part 1: diagnosis and classification of diabetes mellitus

provisional report of a WHO consultation. Diabet Med 1998;15:539-53.

48. Balkau B, Charles MA. Comment on the provisional report from the WHO

consultation. European Group for the Study of Insulin Resistance (EGIR). Diabet Med

1999;16:442-3.

49. International Diabetes Federation. The IDF consensus global definition of the

metabolic syndrome. 2006. http://www.idf.org/webdata/docs/IDF_Meta_def_final.pdf.

Accessed February 2013.

50. Despres JP, Lemieux I, Bergeron J, Pibarot P, Mathieu P, Larose E, Rodes-Cabau

J, Bertrand OF, Poirier P. Abdominal obesity and the metabolic syndrome: contribution

to global cardiometabolic risk. Arterioscler Thromb Vasc Biol 2008;28:1039-49.

51. Despres JP, Lemieux I. Abdominal obesity and metabolic syndrome. Nature

2006;444:881-7.

151

52. Cornier MA, Dabelea D, Hernandez TL, Lindstrom RC, Steig AJ, Stob NR, Van

Pelt RE, Wang H, Eckel RH. The metabolic syndrome. Endocr Rev 2008;29:777-822.

53. Eckel RH, Grundy SM, Zimmet PZ. The metabolic syndrome. Lancet

2005;365:1415-28.

54. Tchernof A, Despres JP. Pathophysiology of human visceral obesity: an update.

Physiol Rev 2013;93:359-404.

55. Shoelson SE, Lee J, Goldfine AB. Inflammation and insulin resistance. J Clin

Invest 2006;116:1793-801.

56. Badawi A, Klip A, Haddad P, Cole DEC, Garcia Bailo B, El-Sohemy A, Karmali

M. Type 2 Diabetes Mellitus and Inflammation: Prospects for Biomarkers of Risk and

Nutritional Intervention. Diabetes, Metabolic Syndrome and Obesity: Targets and

Therapy 2010;3:173-86.

57. Hortin GL, Sviridov D, Anderson NL. High-abundance polypeptides of the

human plasma proteome comprising the top 4 logs of polypeptide abundance. Clin Chem

2008;54:1608-16.

58. Anderson L. Candidate-based proteomics in the search for biomarkers of

cardiovascular disease. J Physiol 2005;563:23-60.

59. Fischer K, Da Costa LA, Garcia-Bailo B, Borchers CH, El-Sohemy A.

Glutathione S-Transferase (GST) M1, but Not GSTT1, Genotype Influences Plasma

Proteomic Profiles in Caucasian and East Asian Young Adults. J Proteome Res

2012;11:5022-33.

60. Fernandez-Real JM, Pickup JC. Innate immunity, insulin resistance and type 2

diabetes. TRENDS Endocrin Metabol 2007;19:10-6.

61. Lau DC, Dhillon B, Yan H, Szmitko PE, Verma S. Adipokines: molecular links

between obesity and atheroslcerosis. Am J Physiol Heart Circ Physiol 2005;288:H2031-

H2041.

62. Szmitko PE, Teoh H, Stewart DJ, Verma S. Adiponectin and cardiovascular

disease: state of the art? Am J Physiol Heart Circ Physiol 2007;292:H1655-H1663.

63. Anand SS, Yusuf S, Vuksan V, Devanesen S, Teo KK, Montague PA, Kelemen

L, Yi C, Lonn E, Gerstein H et al. Differences in risk factors, atherosclerosis, and

cardiovascular disease between ethnic groups in Canada: the Study of Health Assessment

and Risk in Ethnic groups (SHARE). Lancet 2000;356:279-84.

64. Blackett PR, Sanghera DK. Genetic determinants of cardiometabolic risk: a

proposed model for phenotype association and interaction. J Clin Lipidol 2013;7:65-81.

65. Frayling TM, Timpson NJ, Weedon MN, Zeggini E, Freathy RM, Lindgren CM,

Perry JR, Elliott KS, Lango H, Rayner NW et al. A common variant in the FTO gene is

associated with body mass index and predisposes to childhood and adult obesity. Science

2007;316:889-94.

66. Rees SD, Islam M, Hydrie MZ, Chaudhary B, Bellary S, Hashmi S, O'Hare JP,

Kumar S, Sanghera DK, Chaturvedi N et al. An FTO variant is associated with Type 2

152

diabetes in South Asian populations after accounting for body mass index and waist

circumference. Diabet Med 2011;28:673-80.

67. Li H, Kilpelainen TO, Liu C, Zhu J, Liu Y, Hu C, Yang Z, Zhang W, Bao W, Cha

S et al. Association of genetic variation in FTO with risk of obesity and type 2 diabetes

with data from 96,551 East and South Asians. Diabetologia 2012;55:981-95.

68. Pausova Z, Syme C, Abrahamowicz M, Xiao Y, Leonard GT, Perron M, Richer

L, Veillette S, Smith GD, Seda O et al. A common variant of the FTO gene is associated

with not only increased adiposity but also elevated blood pressure in French Canadians.

Circ Cardiovasc Genet 2009;2:260-9.

69. Kathiresan S, Manning AK, Demissie S, D'Agostino RB, Surti A, Guiducci C,

Gianniny L, Burtt NP, Melander O, Orho-Melander M et al. A genome-wide association

study for blood lipid phenotypes in the Framingham Heart Study. BMC Med Genet

2007;8 Suppl 1:S17.

70. Kathiresan S, Melander O, Guiducci C, Surti A, Burtt NP, Rieder MJ, Cooper

GM, Roos C, Voight BF, Havulinna AS et al. Six new loci associated with blood low-

density lipoprotein cholesterol, high-density lipoprotein cholesterol or triglycerides in

humans. Nat Genet 2008;40:189-97.

71. Schafer SA, Machicao F, Fritsche A, Haring HU, Kantartzis K. New type 2

diabetes risk genes provide new insights in insulin secretion mechanisms. Diabetes Res

Clin Pract 2011;93 Suppl 1:S9-24.

72. Sanghera DK, Blackett PR. Type 2 Diabetes Genetics: Beyond GWAS. J Diabetes

Metab 2012;3.

73. Zeggini E, Groves CJ, Parkinson JR, Halford S, Owen KR, Frayling TM, Walker

M, Hitman GA, Levy JC, O'Rahilly S et al. Large-scale studies of the association

between variation at the TNF/LTA locus and susceptibility to type 2 diabetes.

Diabetologia 2005;48:2013-7.

74. Saxena R, Voight BF, Lyssenko V, Burtt NP, de Bakker PI, Chen H, Roix JJ,

Kathiresan S, Hirschhorn JN, Daly MJ et al. Genome-wide association analysis identifies

loci for type 2 diabetes and triglyceride levels. Science 2007;316:1331-6.

75. Scott LJ, Mohlke KL, Bonnycastle LL, Willer CJ, Li Y, Duren WL, Erdos MR,

Stringham HM, Chines PS, Jackson AU et al. A genome-wide association study of type 2

diabetes in Finns detects multiple susceptibility variants. Science 2007;316:1341-5.

76. Sladek R, Rocheleau G, Rung J, Dina C, Shen L, Serre D, Boutin P, Vincent D,

Belisle A, Hadjadj S et al. A genome-wide association study identifies novel risk loci for

type 2 diabetes. Nature 2007;445:881-5.

77. Genome-wide association study of 14,000 cases of seven common diseases and

3,000 shared controls. Nature 2007;447:661-78.

78. Zeggini E, Scott LJ, Saxena R, Voight BF, Marchini JL, Hu T, de Bakker PI,

Abecasis GR, Almgren P, Andersen G et al. Meta-analysis of genome-wide association

data and large-scale replication identifies additional susceptibility loci for type 2 diabetes.

Nat Genet 2008;40:638-45.

153

79. Deloukas P, Kanoni S, Willenborg C, Farrall M, Assimes TL, Thompson JR,

Ingelsson E, Saleheen D, Erdmann J, Goldstein BA et al. Large-scale association analysis

identifies new risk loci for coronary artery disease. Nat Genet 2013;45:25-33.

80. Yusuf S, Hawken S, Ounpuu S, Dans T, Avezum A, Lanas F, McQueen M, Budaj

A, Pais P, Varigos J et al. Effect of potentially modifiable risk factors associated with

myocardial infarction in 52 countries (the INTERHEART study): case-control study.

Lancet 2004;364:937-52.

81. Siri-Tarino PW, Sun Q, Hu FB, Krauss RM. Meta-analysis of prospective cohort

studies evaluating the association of saturated fat with cardiovascular disease. Am J Clin

Nutr 2010;91:535-46.

82. Siri-Tarino PW, Sun Q, Hu FB, Krauss RM. Saturated fat, carbohydrate, and

cardiovascular disease. Am J Clin Nutr 2010;91:502-9.

83. Estruch R, Ros E, Salas-Salvado J, Covas MI, Pharm D, Corella D, Aros F,

Gomez-Gracia E, Ruiz-Gutierrez V, Fiol M et al. Primary Prevention of Cardiovascular

Disease with a Mediterranean Diet. N Engl J Med 2013.

84. Kayaniyil S, Vieth R, Retnakaran R, Knight JA, Qi Y, Gerstein HC, Perkins BA,

Harris SB, Zinman B, Hanley AJ. Association of vitamin D with insulin resistance and

beta-cell dysfunction in subjects at risk for type 2 diabetes. Diabetes Care 2010;33:1379-

81.

85. Bellia A, Garcovich C, D'Adamo M, Lombardo M, Tesauro M, Donadel G,

Gentileschi P, Lauro D, Federici M, Lauro R et al. Serum 25-hydroxyvitamin D levels are

inversely associated with systemic inflammation in severe obese subjects. Intern Emerg

Med 2013;8:33-40.

86. Belenchia AM, Tosh AK, Hillman LS, Peterson CA. Correcting vitamin D

insufficiency improves insulin sensitivity in obese adolescents: a randomized controlled

trial. Am J Clin Nutr 2013; 97:774-81.

87. Kayaniyil S, Vieth R, Harris SB, Retnakaran R, Knight JA, Gerstein HC, Perkins

BA, Zinman B, Hanley AJ. Association of 25(OH)D and PTH with metabolic syndrome

and its traditional and nontraditional components. J Clin Endocrinol Metab 2011;96:168-

75.

88. Schottker B, Haug U, Schomburg L, Kohrle J, Perna L, Muller H, Holleczek B,

Brenner H. Strong associations of 25-hydroxyvitamin D concentrations with all-cause,

cardiovascular, cancer, and respiratory disease mortality in a large cohort study. Am J

Clin Nutr 2013;97:782-93.

89. Cummings J, Ward TH, Greystoke A, Ranson M, Dive C. Biomarker method

validation in anticancer drug development. Br J Pharmacol 2008;153:646-56.

90. Biomarkers and surrogate endpoints: preferred definitions and conceptual

framework. Clin Pharmacol Ther 2001;69:89-95.

91. Pickup JC. Inflammation and activated innate immunity in the pathogenesis of

type 2 diabetes. Diabetes Care 2004;27:813-23.

92. Medzhitov R, Janeway C. Innate immunity. N Engl J Med 2000;343:338-44.

154

93. Hotamisligil GS, Arner P, Caro JF, Atkinson RL, Spiegelman BM. Increased

adipose tissue expression of tumor necrosis factor-a in human obesity and insulin

resistance. J Clin Invest 1995;95:2409-15.

94. Calder PC, Albers R, Antoine JM, Blum S, Bourdet-Sicard R, Ferns GA, Folkerts

G, Friedmann PS, Frost GS, Guarner F et al. Inflammatory disease processes and

interactions with nutrition. Br J Nutr 2009;101 Suppl 1:S1-45.

95. Calder PC, Ahluwalia N, Albers R, Bosco N, Bourdet-Sicard R, Haller D, Holgate

ST, Jonsson LS, Latulippe ME, Marcos A et al. A consideration of biomarkers to be used

for evaluation of inflammation in human nutritional studies. Br J Nutr 2013;109 Suppl

1:S1-34.

96. Stumvoll M, Goldstein B, van Haeften T. Type 2 diabetes: principles of

pathogenesis and therapy. Lancet 2005;365:1333-46.

97. Gastaldelli A. Role of beta-cell dysfunction, ectopic fat accumulation and insulin

resistance in the pathogenesis of type 2 diabetes mellitus. Diabetes Res Clin Pract

2011;93 Suppl 1:S60-S65.

98. Bremer AA, Mietus-Snyder M, Lustig RH. Toward a unifying hypothesis of

metabolic syndrome. Pediatrics 2012;129:557-70.

99. Maedler K. Beta cells in type 2 diabetes - a crucial contribution to pathogenesis.

Diabetes Obes Metab 2008;10:408-20.

100. Lencioni C, Lupi R, Del PS. Beta-cell failure in type 2 diabetes mellitus. Curr

Diab Rep 2008;8:179-84.

101. Prentki M, Nolan CJ. Islet beta cell failure in type 2 diabetes. J Clin Invest

2006;116:1802-12.

102. DeFronzo RA, Tobin JD, Andres R. Glucose clamp technique: a method for

quantifying insulin secretion and resistance. Am J Physiol 1979;237:E214-E223.

103. Levy JC, Matthews DR, Hermans MP. Correct homeostasis model assessment

(HOMA) evaluation uses the computer program. Diabetes Care 1998;21:2191-2.

104. Wallace TM, Levy JC, Matthews DR. Use and abuse of HOMA modeling.

Diabetes Care 2004;27:1487-95.

105. Matthews DR, Hosker JP, Rudenski AS. Homeostasis model assessment: Insulin

resistance and beta-cell function from fasting plasma glucose and insulin concentrations

in man. Diabetologia 1985;28:412-9.

106. Bonora E, Targher G, Alberiche M, Bonadonna RC, Saggiani F, Zenere MB,

Monauni T, Muggeo M. Homeostasis model assessment closely mirrors the glucose

clamp technique in the assessment of insulin sensitivity: studies in subjects with various

degrees of glucose tolerance and insulin sensitivity. Diabetes Care 2000;23:57-63.

107. Emoto M, Nishizawa Y, Maekawa K, Hiura Y, Kanda H, Kawagishi T, Shoji T,

Okuno Y, Morii H. Homeostasis model assessment as a clinical index of insulin

resistance in type 2 diabetic patients treated with sulfonylureas. Diabetes Care

1999;22:818-22.

155

108. Brunzell JD, Davidson M, Furberg CD, Goldberg RB, Howard BV, Stein JH,

Witztum JL. Lipoprotein management in patients with cardiometabolic risk: consensus

statement from the American Diabetes Association and the American College of

Cardiology Foundation. Diabetes Care 2008;31:811-22.

109. Gropper SS, Smith JL, Groff JL. Advanced nutrition and human metabolism, 5th

edition. Belmont, CA: Wadsworth, 2005.

110. Steinberg D, Witztum JL. Lipoproteins and atherogenesis. Current concepts.

JAMA 1990;264:3047-52.

111. Durrington P. Dyslipidaemia. Lancet 2003;362:717-31.

112. Anderson TJ, Gregoire J, Hegele RA, Couture P, Mancini GB, McPherson R,

Francis GA, Poirier P, Lau DC, Grover S et al. 2012 update of the Canadian

Cardiovascular Society guidelines for the diagnosis and treatment of dyslipidemia for the

prevention of cardiovascular disease in the adult. Can J Cardiol 2013;29:151-67.

113. Sarwar N, Danesh J, Eiriksdottir G, Sigurdsson G, Wareham N, Bingham S,

Boekholdt SM, Khaw KT, Gudnason V. Triglycerides and the risk of coronary heart

disease: 10,158 incident cases among 262,525 participants in 29 Western prospective

studies. Circulation 2007;115:450-8.

114. Ninomiya JK, L'Italien G, Criqui MH, Whyte JL, Gamst A, Chen RS. Association

of the metabolic syndrome with history of myocardial infarction and stroke in the Third

National Health and Nutrition Examination Survey. Circulation 2004;109:42-6.

115. Miller M, Stone NJ, Ballantyne C, Bittner V, Criqui MH, Ginsberg HN, Goldberg

AC, Howard WJ, Jacobson MS, Kris-Etherton PM et al. Triglycerides and cardiovascular

disease: a scientific statement from the American Heart Association. Circulation

2011;123:2292-333.

116. Chait A, Brazg RL, Tribble DL, Krauss RM. Susceptibility of small, dense, low-

density lipoproteins to oxidative modification in subjects with the atherogenic lipoprotein

phenotype, pattern B. Am J Med 1993;94:350-6.

117. Greene DJ, Skeggs JW, Morton RE. Elevated triglyceride content diminishes the

capacity of high density lipoprotein to deliver cholesteryl esters via the scavenger

receptor class B type I (SR-BI). J Biol Chem 2001;276:4804-11.

118. Skeggs JW, Morton RE. LDL and HDL enriched in triglyceride promote

abnormal cholesterol transport. J Lipid Res 2002;43:1264-74.

119. Garcia-Bailo B, Brenner DR, Nielsen D, Lee HJ, Domanski D, Kuzyk M,

Borchers CH, Badawi A, Karmali MA, El-Sohemy A. Dietary patterns and ethnicity are

associated with distinct plasma proteomic groups. Am J Clin Nutr 2012;95:352-61.

120. Kuzyk MA, Smith D, Yang J, Cross TJ, Jackson AM, Hardie DB, Anderson NL,

Borchers CH. Multiple reaction monitoring-based, multiplexed, absolute quantitation of

45 proteins in human plasma. Mol Cell Proteomics 2009;8:1860-77.

121. Norman AW. From vitamin D to hormone D: fundamentals of the vitamin D

endocrine system essential for good health. Am J Clin Nutr 2008;88:491S-9S.

156

122. Norman AW. Sunlight, season, skin pigmentation, vitamin D, and 25-

hydroxyvitamin D: Integral components of the vitamin D endocrine system. American

Journal of Clinical Nutrition 1998;67:1108-10.

123. Holick MF. Sunlight and vitamin D for bone health and prevention of

autoimmune diseases, cancers, and cardiovascular disease. Am J Clin Nutr

2004;80:1678S-88S.

124. Holick MF. Vitamin D: evolutionary, physiological and health perspectives. Curr

Drug Targets 2011;12:4-18.

125. Holick MF. Vitamin D: A millenium perspective. J Cell Biochem 2003;88:296-

307.

126. Hess AF, Unger LJ, Pappenheimer AM. Experimental rickets in rats: VII. The

prevention of rickets by sunlight, by the rays of the mercury vapor lamp, and by the

carbon arc lamp. J Exp Med 1922;36:427-46.

127. Schlingmann KP, Kaufmann M, Weber S, Irwin A, Goos C, John U, Misselwitz J,

Klaus G, Kuwertz-Broking E, Fehrenbach H et al. Mutations in CYP24A1 and idiopathic

infantile hypercalcemia. N Engl J Med 2011;365:410-21.

128. Calvo MS, Whiting SJ, Barton CN. Vitamin D fortification in the United States

and Canada: current status and data needs. Am J Clin Nutr 2004;80:1710S-6S.

129. Holick MF, Tian XQ, Allen M. Evolutionary importance for the membrane

enhancement of the production of vitamin D3 in the skin of poikilothermic animals. Proc

Natl Acad Sci U S A 1995;92:3124-6.

130. Haddad JG, Matsuoka LY, Hollis BW, Hu YZ, Wortsman J. Human plasma

transport of vitamin D after its endogenous synthesis. J Clin Invest 1993;91:2552-5.

131. Tripkovic L, Lambert H, Hart K, Smith CP, Bucca G, Penson S, Chope G,

Hypponen E, Berry J, Vieth R et al. Comparison of vitamin D2 and vitamin D3

supplementation in raising serum 25-hydroxyvitamin D status: a systematic review and

meta-analysis. Am J Clin Nutr 2012;95:1357-64.

132. Reboul E, Borel P. Proteins involved in uptake, intracellular transport and

basolateral secretion of fat-soluble vitamins and carotenoids by mammalian enterocytes.

Prog Lipid Res 2011;50:388-402.

133. Speeckaert MM, Taes YE, De Buyzere ML, Christophe AB, Kaufman JM,

Delanghe JR. Investigation of the potential association of vitamin D binding protein with

lipoproteins. Ann Clin Biochem 2010;47:143-50.

134. Prosser DE, Jones G. Enzymes involved in the activation and inactivation of

vitamin D. Trends Biochem Sci 2004;29:664-73.

135. Haussler MR, Jurutka PW, Mizwicki M, Norman AW. Vitamin D receptor

(VDR)-mediated actions of 1alpha,25(OH)vitamin D: genomic and non-genomic

mechanisms. Best Pract Res Clin Endocrinol Metab 2011;25:543-59.

136. Nykjaer A, Dragun D, Walther D, Vorum H, Jacobsen C, Herz J, Melsen F,

Christensen EI, Willnow TE. An endocytic pathway essential for renal uptake and

activation of the steroid 25-(OH) vitamin D3. Cell 1999;96:507-15.

157

137. Hilpert J, Wogensen L, Thykjaer T, Wellner M, Schlichting U, Orntoft TF,

Bachmann S, Nykjaer A, Willnow TE. Expression profiling confirms the role of

endocytic receptor megalin in renal vitamin D3 metabolism. Kidney Int 2002;62:1672-

81.

138. Wacker M, Holick MF. Vitamin D - effects on skeletal and extraskeletal health

and the need for supplementation. Nutrients 2013;5:111-48.

139. Christakos S, Dhawan P, Liu Y, Peng X, Porta A. New insights into the

mechanisms of vitamin D action. J Cell Biochem 2003;88:695-705.

140. Shimada T, Hasegawa H, Yamazaki Y, Muto T, Hino R, Takeuchi Y, Fujita T,

Nakahara K, Fukumoto S, Yamashita T. FGF-23 is a potent regulator of vitamin D

metabolism and phosphate homeostasis. J Bone Miner Res 2004;19:429-35.

141. Overbergh L, Stoffels K, Waer M, Verstuyf A, Bouillon R, Mathieu C. Immune

regulation of 25-hydroxyvitamin D-1alpha-hydroxylase in human monocytic THP1 cells:

mechanisms of interferon-gamma-mediated induction. J Clin Endocrinol Metab

2006;91:3566-74.

142. Stoffels K, Overbergh L, Giulietti A, Verlinden L, Bouillon R, Mathieu C.

Immune regulation of 25-hydroxyvitamin-D3-1alpha-hydroxylase in human monocytes. J

Bone Miner Res 2006;21:37-47.

143. Zhang Y, Leung DY, Richers BN, Liu Y, Remigio LK, Riches DW, Goleva E.

Vitamin d inhibits monocyte/macrophage proinflammatory cytokine production by

targeting MAPK phosphatase-1. J Immunol 2012;188:2127-35.

144. Bouillon R, Carmeliet G, Verlinden L, van EE, Verstuyf A, Luderer HF, Lieben

L, Mathieu C, Demay M. Vitamin D and human health: lessons from vitamin D receptor

null mice. Endocr Rev 2008;29:726-76.

145. Heikkinen S, Vaisanen S, Pehkonen P, Seuter S, Benes V, Carlberg C. Nuclear

hormone 1alpha,25-dihydroxyvitamin D3 elicits a genome-wide shift in the locations of

VDR chromatin occupancy. Nucleic Acids Res 2011;39:9181-93.

146. Zehnder D, Bland R, Williams MC, McNinch RW, Howie AJ, Stewart PM,

Hewison M. Extrarenal expression of 25-hydroxyvitamin d(3)-1 alpha-hydroxylase. J


147. Huhtakangas JA, Olivera CJ, Bishop JE, Zanello LP, Norman AW. The vitamin D

receptor is present in caveolae-enriched plasma membranes and binds 1 alpha,25(OH)2-

vitamin D3 in vivo and in vitro. Mol Endocrinol 2004;18:2660-71.

148. Fleet JC. Rapid, membrane-initiated actions of 1,25 dihydroxyvitamin D: what

are they and what do they mean? J Nutr 2004;134:3215-8.

149. Baz-Hecht M, Goldfine AB. The impact of vitamin D deficiency on diabetes and

cardiovascular risk. Curr Opin Endocrinol Diabetes Obes 2010;17:113-9.

150. Jones G. Pharmacokinetics of vitamin D toxicity. Am J Clin Nutr 2008;88:582S-

6S.

151. Institute of Medicine. Dietary Reference Intakes for Calcium and Vitamin D.

Washington, DC: National Academies Press, 2011.

158

152. Whiting SJ, Langlois KA, Vatanparast H, Greene-Finestone LS. The vitamin D

status of Canadians relative to the 2011 Dietary Reference Intakes: an examination in

children and adults with and without supplement use. Am J Clin Nutr 2011;94:128-35.

153. Gozdzik A, Barta JL, Weir A, Cole DE, Vieth R, Whiting SJ, Parra EJ. Serum 25-

hydroxyvitamin d concentrations fluctuate seasonally in young adults of diverse ancestry

living in toronto. J Nutr 2010;140:2213-20.

154. Schwalfenberg GK, Whiting SJ. A Canadian response to the 2010 Institute of

Medicine vitamin D and calcium guidelines. Public Health Nutr 2011;14:746-8.

155. Vieth R. Why the minimum desirable serum 25-hydroxyvitamin D level should be

75 nmol/L (30 ng/ml). Best Pract Res Clin Endocrinol Metab 2011;25:681-91.

156. Bischoff-Ferrari HA, Giovannucci E, Willett WC, Dietrich T, Dawson-Hughes B.

Estimation of optimal serum concentrations of 25-hydroxyvitamin D for multiple health

outcomes. Am J Clin Nutr 2006;84:18-28.

157. Webb AR, Kline L, Holick MF. Influence of season and latitude on the cutaneous

synthesis of vitamin D3: exposure to winter sunlight in Boston and Edmonton will not

promote vitamin D3 synthesis in human skin. J Clin Endocrinol Metab 1988;67:373-8.

158. Holick MF, Chen TC, Lu Z, Sauter E. Vitamin D and skin physiology: a D-

lightful story. J Bone Miner Res 2007;22 Suppl 2:V28-V33.

159. Willett WC. Nutritional Epidemiology. New York, NY: Oxford University Press,

1998.

160. Wu H, Gozdzik A, Barta JL, Wagner D, Cole DE, Vieth R, Parra EJ, Whiting SJ.

The development and evaluation of a food frequency questionnaire used in assessing

vitamin D intake in a sample of healthy young Canadian adults of diverse ancestry. Nutr

Res 2009;29:255-61.

161. Aasheim ET, Hofso D, Hjelmesaeth J, Birkeland KI, Bohmer T. Vitamin status in

morbidly obese patients: a cross-sectional study. Am J Clin Nutr 2008;87:362-9.

162. Alemzadeh R, Kichler J, Babar G, Calhoun M. Hypovitaminosis D in obese

children and adolescents: relationship with adiposity, insulin sensitivity, ethnicity, and

season. Metabolism 2008;57:183-91.

163. Cheng S, Massaro JM, Fox CS, Larson MG, Keyes MJ, McCabe EL, Robins SJ,

O'Donnell CJ, Hoffmann U, Jacques PF et al. Adiposity, cardiometabolic risk, and

vitamin D status: the Framingham Heart Study. Diabetes 2010;59:242-8.

164. Earthman CP, Beckman LM, Masodkar K, Sibley SD. The link between obesity

and low circulating 25-hydroxyvitamin D concentrations: considerations and

implications. Int J Obes (Lond) 2012;36:387-96.

165. Blum M, Dolnikowski G, Seyoum E, Harris SS, Booth SL, Peterson J, Saltzman

E, wson-Hughes B. Vitamin D(3) in fat tissue. Endocrine 2008;33:90-4.

166. Gonzalez-Molero I, Rojo-Martinez G, Morcillo S, Gutierrez C, Rubio E, Perez-

Valero V, Esteva I, Ruiz de Adana MS, Almaraz MC, Colomo N et al. Hypovitaminosis

D and incidence of obesity: a prospective study. Eur J Clin Nutr 2013.

159

167. Wood RJ. Vitamin D and adipogenesis: new molecular insights. Nutr Rev

2008;66:40-6.

168. Kong J, Li YC. Molecular mechanism of 1,25-dihydroxyvitamin D3 inhibition of

adipogenesis in 3T3-L1 cells. Am J Physiol Endocrinol Metab 2006;290:E916-E924.

169. Querfeld U, Hoffmann MM, Klaus G, Eifinger F, Ackerschott M, Michalk D,

Kern PA. Antagonistic effects of vitamin D and parathyroid hormone on lipoprotein

lipase in cultured adipocytes. J Am Soc Nephrol 1999;10:2158-64.

170. Rayalam S, la-Fera MA, Ambati S, Yang JY, Park HJ, Baile CA. Enhanced

effects of 1,25(OH)(2)D(3) plus genistein on adipogenesis and apoptosis in 3T3-L1

adipocytes. Obesity (Silver Spring) 2008;16:539-46.

171. Ortega RM, Lopez-Sobaler AM, Aparicio A, Bermejo LM, Rodriguez-Rodriguez

E, Perea JM, Andres P. Vitamin D status modification by two slightly hypocaloric diets

in young overweight/obese women. Int J Vitam Nutr Res 2009;79:71-8.

172. Teegarden D, White KM, Lyle RM, Zemel MB, Van L, Matkovic V, Craig BA,

Schoeller DA. Calcium and dairy product modulation of lipid utilization and energy

expenditure. Obesity (Silver Spring) 2008;16:1566-72.

173. Sneve M, Figenschau Y, Jorde R. Supplementation with cholecalciferol does not

result in weight reduction in overweight and obese subjects. Eur J Endocrinol

2008;159:675-84.

174. Zittermann A, Frisch S, Berthold HK, Gotting C, Kuhn J, Kleesiek K, Stehle P,

Koertke H, Koerfer R. Vitamin D supplementation enhances the beneficial effects of

weight loss on cardiovascular disease risk markers. Am J Clin Nutr 2009;89:1321-7.

175. Vimaleswaran KS, Berry DJ, Lu C, Tikkanen E, Pilz S, Hiraki LT, Cooper JD,

Dastani Z, Li R, Houston DK et al. Causal relationship between obesity and vitamin D

status: bi-directional Mendelian randomization analysis of multiple cohorts. PLoS Med

2013;10:e1001383.

176. Prentice A. Vitamin D deficiency: a global perspective. Nutr Rev 2008;66:S153-

S164.

177. Clemens TL, Adams JS, Henderson SL, Holick MF. Increased skin pigment

reduces the capacity of skin to synthesise vitamin D3. Lancet 1982;1:74-6.

178. Sulistyoningrum DC, Green TJ, Lear SA, Devlin AM. Ethnic-specific differences

in vitamin d status is associated with adiposity. PLoS One 2012;7:e43159.

179. Freedman BI, Register TC. Effect of race and genetics on vitamin D metabolism,

bone and vascular health. Nat Rev Nephrol 2012.

180. Freedman BI, Wagenknecht LE, Hairston KG, Bowden DW, Carr JJ, Hightower

RC, Gordon EJ, Xu J, Langefeld CD, Divers J. Vitamin d, adiposity, and calcified

atherosclerotic plaque in african-americans. J Clin Endocrinol Metab 2010;95:1076-83.

181. Harris SS. Does vitamin D deficiency contribute to increased rates of

cardiovascular disease and type 2 diabetes in African Americans? Am J Clin Nutr

2011;93:1175S-8S.

160

182. Alvarez JA, Ashraf AP, Hunter GR, Gower BA. Serum 25-hydroxyvitamin D and

parathyroid hormone are independent determinants of whole-body insulin sensitivity in

women and may contribute to lower insulin sensitivity in African Americans. Am J Clin

Nutr 2010;92:1344-9.

183. Scragg R, Sowers M, Bell C. Serum 25-hydroxyvitamin D, diabetes, and ethnicity

in the Third National Health and Nutrition Examination Survey. Diabetes Care

2004;27:2813-8.

184. Godschalk M, Levy JR, Downs RW, Jr. Glucocorticoids decrease vitamin D

receptor number and gene expression in human osteosarcoma cells. J Bone Miner Res

1992;7:21-7.

185. Skversky AL, Kumar J, Abramowitz MK, Kaskel FJ, Melamed ML. Association

of glucocorticoid use and low 25-hydroxyvitamin D levels: results from the National

Health and Nutrition Examination Survey (NHANES): 2001-2006. J Clin Endocrinol

Metab 2011;96:3838-45.

186. Kurahashi I, Matsunuma A, Kawane T, Abe M, Horiuchi N. Dexamethasone

enhances vitamin D-24-hydroxylase expression in osteoblastic (UMR-106) and renal

(LLC-PK1) cells treated with 1alpha,25-dihydroxyvitamin D3. Endocrine 2002;17:109-

18.

187. Sowers MR, Wallace RB, Hollis BW, Lemke JH. Parameters related to 25-OH-D

levels in a population-based study of women. Am J Clin Nutr 1986;43:621-8.

188. Harris SS, Dawson-Hughes B. The association of oral contraceptive use with

plasma 25-hydroxyvitamin D levels. J Am Coll Nutr 1998;17:282-4.

189. Bouillon R, van BH, de MP. The measurement of the vitamin D-binding protein

in human serum. J Clin Endocrinol Metab 1977;45:225-31.

190. Abbas S, Linseisen J, Slanger T, Kropp S, Mutschelknauss EJ, Flesch-Janys D,

Chang-Claude J. Serum 25-hydroxyvitamin D and risk of post-menopausal breast cancer-

-results of a large case-control study. Carcinogenesis 2008;29:93-9.

191. Nesby-O'Dell S, Scanlon KS, Cogswell ME, Gillespie C, Hollis BW, Looker AC,

Allen C, Doughertly C, Gunter EW, Bowman BA. Hypovitaminosis D prevalence and

determinants among African American and white women of reproductive age: third

National Health and Nutrition Examination Survey, 1988-1994. Am J Clin Nutr

2002;76:187-92.

192. Williams JK. Rationale for new oral contraceptive dosing. Int J Fertil Womens

Med 2004;49:30-5.

193. Lechner D, Bajna E, Adlercreutz H, Cross HS. Genistein and 17beta-estradiol, but

not equol, regulate vitamin D synthesis in human colon and breast cancer cells.

Anticancer Res 2006;26:2597-603.

194. Tanaka Y, Castillo L, DeLuca HF. Control of renal vitamin D hydroxylases in

birds by sex hormones. Proc Natl Acad Sci U S A 1976;73:2701-5.

195. Schwartz B, Smirnoff P, Shany S, Liel Y. Estrogen controls expression and

bioresponse of 1,25-dihydroxyvitamin D receptors in the rat colon. Mol Cell Biochem

2000;203:87-93.

161

196. van Hoof HJ, de Sevaux RG, van BH, Swinkels LM, Klipping C, Ross HA,

Sweep CG. Relationship between free and total 1,25-dihydroxyvitamin D in conditions of

modified binding. Eur J Endocrinol 2001;144:391-6.

197. Aarskog D, Aksnes L, Markestad T, Rodland O. Effect of estrogen on vitamin D

metabolism in tall girls. J Clin Endocrinol Metab 1983;57:1155-8.

198. Kwok S, Canoy D, Ashton WD, Lowe GD, Wood D, Humphries SE, Charlton-

Menys V, Durrington PN. Increased C-reactive protein levels in overweight and obese

women taking exogenous hormones: the United Kingdom Women's Heart Study

(UKWHS). Clin Endocrinol (Oxf) 2009;71:727-32.

199. Soska V, Fiala J, Nebeska K, Jarkovsky J, Hruba D. The atherogenic index of

plasma is increased by hormonal contraception. Scand J Clin Lab Invest 2011;71:94-100.

200. Katayama H, Paczesny S, Prentice R, Aragaki A, Faca VM, Pitteri SJ, Zhang Q,

Wang H, Silva M, Kennedy J et al. Application of serum proteomics to the Women's

Health Initiative conjugated equine estrogens trial reveals a multitude of effects relevant

to clinical findings. Genome Med 2009;1:47.

201. Pitteri SJ, Hanash SM, Aragaki A, Amon LM, Chen L, Busald BT, Paczesny S,

Katayama H, Wang H, Johnson MM et al. Postmenopausal estrogen and progestin effects

on the serum proteome. Genome Med 2009;1:121.

202. Chlebowski RT, Anderson GL. Changing concepts: Menopausal hormone therapy

and breast cancer. J Natl Cancer Inst 2012;104:517-27.

203. Dreon DM, Slavin JL, Phinney SD. Oral contraceptive use and increased plasma

concentration of C-reactive protein. Life Sci 2003;73:1245-52.

204. Piltonen T, Puurunen J, Hedberg P, Ruokonen A, Mutt SJ, Herzig KH, Nissinen

A, Morin-Papunen L, Tapanainen JS. Oral, transdermal and vaginal combined

contraceptives induce an increase in markers of chronic inflammation and impair insulin

sensitivity in young healthy normal-weight women: a randomized study. Hum Reprod

2012;27:3046-56.

205. Haider Z, D'Souza R. Non-contraceptive benefits and risks of contraception. Best

Pract Res Clin Obstet Gynaecol 2009;23:249-62.

206. Kluft C, Leuven JA, Helmerhorst FM, Krans HM. Pro-inflammatory effects of

oestrogens during use of oral contraceptives and hormone replacement treatment. Vascul

Pharmacol 2002;39:149-54.

207. Josse AR, Garcia-Bailo B, Fischer K, El-Sohemy A. Novel effects of hormonal

contraceptive use on the plasma proteome. PLoS One 2012;7:e45162.

208. Lidegaard O, Lokkegaard E, Jensen A, Skovlund CW, Keiding N. Thrombotic

stroke and myocardial infarction with hormonal contraception. N Engl J Med

2012;366:2257-66.

209. Hunter D, De Lange M, Snieder H, MacGregor AJ, Swaminathan R, Thakker RV,

Spector TD. Genetic contribution to bone metabolism, calcium excretion, and vitamin D

and parathyroid hormone regulation. J Bone Miner Res 2001;16:371-8.

162

210. Orton SM, Morris AP, Herrera BM, Ramagopalan SV, Lincoln MR, Chao MJ,

Vieth R, Sadovnick AD, Ebers GC. Evidence for genetic regulation of vitamin D status in

twins with multiple sclerosis. Am J Clin Nutr 2008;88:441-7.

211. Karohl C, Su S, Kumari M, Tangpricha V, Veledar E, Vaccarino V, Raggi P.

Heritability and seasonal variability of vitamin D concentrations in male twins. Am J Clin

Nutr 2010;92:1393-8.

212. Dastani Z, Li R, Richards B. Genetic Regulation of Vitamin D Levels. Calcif

Tissue Int 2013;92:106-17.

213. Hypponen E, Berry DJ, Wjst M, Power C. Serum 25-hydroxyvitamin D and IgE -

a significant but nonlinear relationship. Allergy 2009;64:613-20.

214. Signorello LB, Shi J, Cai Q, Zheng W, Williams SM, Long J, Cohen SS, Li G,

Hollis BW, Smith JR et al. Common variation in vitamin D pathway genes predicts

circulating 25-hydroxyvitamin D Levels among African Americans. PLoS One

2011;6:e28623.

215. Bu FX, Armas L, Lappe J, Zhou Y, Gao G, Wang HW, Recker R, Zhao LJ.

Comprehensive association analysis of nine candidate genes with serum 25-hydroxy

vitamin D levels among healthy Caucasian subjects. Hum Genet 2010;128:549-56.

216. Cooper JD, Smyth DJ, Walker NM, Stevens H, Burren OS, Wallace C, Greissl C,

Ramos-Lopez E, Hypponen E, Dunger DB et al. Inherited variation in vitamin D genes is

associated with predisposition to autoimmune disease type 1 diabetes. Diabetes

2011;60:1624-31.

217. Lu L, Sheng H, Li H, Gan W, Liu C, Zhu J, Loos RJ, Lin X. Associations

between common variants in GC and DHCR7/NADSYN1 and vitamin D concentration

in Chinese Hans. Hum Genet 2012;131:505-12.

218. Engelman CD, Fingerlin TE, Langefeld CD, Hicks PJ, Rich SS, Wagenknecht

LE, Bowden DW, Norris JM. Genetic and environmental determinants of 25-

hydroxyvitamin D and 1,25-dihydroxyvitamin D levels in Hispanic and African

Americans. J Clin Endocrinol Metab 2008;93:3381-8.

219. Sinotte M, Diorio C, Berube S, Pollak M, Brisson J. Genetic polymorphisms of

the vitamin D binding protein and plasma concentrations of 25-hydroxyvitamin D in

premenopausal women. Am J Clin Nutr 2009;89:634-40.

220. Fu L, Yun F, Oczak M, Wong BY, Vieth R, Cole DE. Common genetic variants

of the vitamin D binding protein (DBP) predict differences in response of serum 25-

hydroxyvitamin D [25(OH)D] to vitamin D supplementation. Clin Biochem

2009;42:1174-7.

221. d'Alesio A, Garabedian M, Sabatier JP, Guaydier-Souquieres G, Marcelli C,

Lemacon A, Walrant-Debray O, Jehan F. Two single-nucleotide polymorphisms in the

human vitamin D receptor promoter change protein-DNA complex formation and are

associated with height and vitamin D status in adolescent girls. Hum Mol Genet

2005;14:3539-48.

163

222. Ahn J, Yu K, Stolzenberg-Solomon R, Simon KC, McCullough ML, Gallicchio

L, Jacobs EJ, Ascherio A, Helzlsouer K, Jacobs KB et al. Genome-wide association study

of circulating vitamin D levels. Hum Mol Genet 2010;19:2739-45.

223. Wang TJ, Zhang F, Richards JB, Kestenbaum B, van Meurs JB, Berry D, Kiel

DP, Streeten EA, Ohlsson C, Koller DL et al. Common genetic determinants of vitamin

D insufficiency: a genome-wide association study. Lancet 2010;376:180-8.

224. Engelman CD, Meyers KJ, Ziegler JT, Taylor KD, Palmer ND, Haffner SM,

Fingerlin TE, Wagenknecht LE, Rotter JI, Bowden DW et al. Genome-wide association

study of vitamin D concentrations in Hispanic Americans: the IRAS family study. J

Steroid Biochem Mol Biol 2010;122:186-92.

225. Lasky-Su J, Lange N, Brehm JM, Damask A, Soto-Quiros M, Avila L, Celedon

JC, Canino G, Cloutier MM, Hollis BW et al. Genome-wide association analysis of

circulating vitamin D levels in children with asthma. Hum Genet 2012;131:1495-505.

226. Benjamin EJ, Dupuis J, Larson MG, Lunetta KL, Booth SL, Govindaraju DR,

Kathiresan S, Keaney JF, Jr., Keyes MJ, Lin JP et al. Genome-wide association with

select biomarker traits in the Framingham Heart Study. BMC Med Genet 2007;8 Suppl

1:S11.

227. Hiraki LT, Major JM, Chen C, Cornelis MC, Hunter DJ, Rimm EB, Simon KC,

Weinstein SJ, Purdue MP, Yu K et al. Exploring the genetic architecture of circulating

25-hydroxyvitamin D. Genet Epidemiol 2013;37:92-8.

228. Manolio TA, Collins FS, Cox NJ, Goldstein DB, Hindorff LA, Hunter DJ,

McCarthy MI, Ramos EM, Cardon LR, Chakravarti A et al. Finding the missing

heritability of complex diseases. Nature 2009;461:747-53.

229. Haddad JG, Chyu KJ. Competitive protein-binding radioassay for 25-

hydroxycholecalciferol. J Clin Endocrinol Metab 1971;33:992-5.

230. Hollis BW. Measuring 25-hydroxyvitamin D in a clinical environment: challenges

and needs. Am J Clin Nutr 2008;88:507S-10S.

231. Hollis BW. Assessment and interpretation of circulating 25-hydroxyvitamin D

and 1,25-dihydroxyvitamin D in the clinical environment. Endocrinol Metab Clin North

Am 2010;39:271-86.

232. Wagner D, Hanwell HE, Schnabl K, Yazdanpanah M, Kimball S, Fu L, Sidhom

G, Rousseau D, Cole DE, Vieth R. The ratio of serum 24,25-dihydroxyvitamin D(3) to

25-hydroxyvitamin D(3) is predictive of 25-hydroxyvitamin D(3) response to vitamin

D(3) supplementation. J Steroid Biochem Mol Biol 2011;126:72-7.

233. Roth HJ, Schmidt-Gayk H, Weber H, Niederau C. Accuracy and clinical

implications of seven 25-hydroxyvitamin D methods compared with liquid

chromatography-tandem mass spectrometry as a reference. Ann Clin Biochem

2008;45:153-9.

234. Snellman G, Melhus H, Gedeborg R, Byberg L, Berglund L, Wernroth L,

Michaelsson K. Determining vitamin D status: a comparison between commercially

available assays. PLoS One 2010;5:e11555.

164

235. Wagner D, Hanwell HE, Vieth R. An evaluation of automated methods for

measurement of serum 25-hydroxyvitamin D. Clin Biochem 2009;42:1549-56.

236. Binkley N, Krueger D, Cowgill CS, Plum L, Lake E, Hansen KE, DeLuca HF,

Drezner MK. Assay variation confounds the diagnosis of hypovitaminosis D: a call for

standardization. J Clin Endocrinol Metab 2004;89:3152-7.

237. Glendenning P, Taranto M, Noble JM, Musk AA, Hammond C, Goldswain PR,

Fraser WD, Vasikaran SD. Current assays overestimate 25-hydroxyvitamin D3 and

underestimate 25-hydroxyvitamin D2 compared with HPLC: need for assay-specific

decision limits and metabolite-specific assays. Ann Clin Biochem 2006;43:23-30.

238. Carter GD. Accuracy of 25-hydroxyvitamin D assays: confronting the issues. Curr

Drug Targets 2011;12:19-28.

239. Chacko SA, Song Y, Manson JE, Van HL, Eaton C, Martin LW, McTiernan A,

Curb JD, Wylie-Rosett J, Phillips LS et al. Serum 25-hydroxyvitamin D concentrations in

relation to cardiometabolic risk factors and metabolic syndrome in postmenopausal

women. Am J Clin Nutr 2011;94:209-17.

240. Delvin EE, Lambert M, Levy E, O'Loughlin J, Mark S, Gray-Donald K, Paradis

G. Vitamin D status is modestly associated with glycemia and indicators of lipid

metabolism in French-Canadian children and adolescents. J Nutr 2010;140:987-91.

241. Ganji V, Zhang X, Shaikh N, Tangpricha V. Serum 25-hydroxyvitamin D

concentrations are associated with prevalence of metabolic syndrome and various

cardiometabolic risk factors in US children and adolescents based on assay-adjusted

serum 25-hydroxyvitamin D data from NHANES 2001-2006. Am J Clin Nutr

2011;94:225-33.

242. Zittermann A, Gummert JF, Borgermann J. The role of vitamin D in dyslipidemia

and cardiovascular disease. Curr Pharm Des 2011;17:933-42.

243. Parker J, Hashmi O, Dutton D, Mavrodaris A, Stranges S, Kandala NB, Clarke A,

Franco OH. Levels of vitamin D and cardiometabolic disorders: systematic review and

meta-analysis. Maturitas 2010;65:225-36.

244. Noordam R, de Craen AJ, Pedram P, Maier AB, Mooijaart SP, van PJ, Feskens

EJ, Streppel MT, Slagboom PE, Westendorp RG et al. Levels of 25-hydroxyvitamin D in

familial longevity: the Leiden Longevity Study. CMAJ 2012.

245. Ding EL, Mehta S, Fawzi WW, Giovannucci EL. Interaction of estrogen therapy

with calcium and vitamin D supplementation on colorectal cancer risk: reanalysis of

Women's Health Initiative randomized trial. Int J Cancer 2008;122:1690-4.

246. Liu PT, Stenger S, Li H, Wenzel L, Tan BH, Krutzik SR, Ochoa MT, Schauber J,

Wu K, Meinken C et al. Toll-like receptor triggering of a vitamin D-mediated human

antimicrobial response. Science 2006;311:1770-3.

247. Hewison M. An update on vitamin D and human immunity. Clin Endocrinol

(Oxf) 2012;76:315-25.

248. Lagishetty V, Liu NQ, Hewison M. Vitamin D metabolism and innate immunity.

Mol Cell Endocrinol 2011;347:97-105.

165

249. Adams JS, Ren S, Liu PT, Chun RF, Lagishetty V, Gombart AF, Borregaard N,

Modlin RL, Hewison M. Vitamin d-directed rheostatic regulation of monocyte

antibacterial responses. J Immunol 2009;182:4289-95.

250. Cippitelli M, Santoni A. Vitamin D3: a transcriptional modulator of the

interferon-gamma gene. Eur J Immunol 1998;28:3017-30.

251. Takeuchi A, Reddy GS, Kobayashi T, Okano T, Park J, Sharma S. Nuclear factor

of activated T cells (NFAT) as a molecular target for 1alpha,25-dihydroxyvitamin D3-

mediated effects. J Immunol 1998;160:209-18.

252. Boonstra A, Barrat FJ, Crain C, Heath VL, Savelkoul HF, O'Garra A. 1alpha,25-

Dihydroxyvitamin d3 has a direct effect on naive CD4(+) T cells to enhance the

development of Th2 cells. J Immunol 2001;167:4974-80.

253. Overbergh L, Decallonne B, Waer M, Rutgeerts O, Valckx D, Casteels KM,

Laureys J, Bouillon R, Mathieu C. 1alpha,25-dihydroxyvitamin D3 induces an

autoantigen-specific T-helper 1/T-helper 2 immune shift in NOD mice immunized with

GAD65 (p524-543). Diabetes 2000;49:1301-7.

254. Codoner-Franch P, Tavarez-Alonso S, Simo-Jorda R, Laporta-Martin P,

Carratala-Calvo A, onso-Iglesias E. Vitamin D status is linked to biomarkers of oxidative

stress, inflammation, and endothelial activation in obese children. J Pediatr

2012;161:848-54.

255. Salekzamani S, Neyestani TR, avi-Majd H, Houshiarrad A, Kalayi A,

Shariatzadeh N, Gharavi A. Is vitamin D status a determining factor for metabolic

syndrome? A case-control study. Diabetes Metab Syndr Obes 2011;4:205-12.

256. Kim M, Na W, Sohn C. Correlation between vitamin D and cardiovascular

disease predictors in overweight and obese Koreans. J Clin Biochem Nutr 2013;52:167-

71.

257. Hypponen E, Berry D, Cortina-Borja M, Power C. 25-Hydroxyvitamin D and pre-

clinical alterations in inflammatory and hemostatic markers: a cross sectional analysis in

the 1958 British Birth Cohort. PLoS One 2010;5:e10801.

258. Shab-Bidar S, Neyestani TR, Djazayery A, Eshraghian MR, Houshiarrad A,

Kalayi A, Shariatzadeh N, Khalaji N, Gharavi A. Improvement of vitamin D status

resulted in amelioration of biomarkers of systemic inflammation in the subjects with type

2 diabetes. Diabetes Metab Res Rev 2012;28:424-30.

259. Schleithoff SS, Zittermann A, Tenderich G, Berthold HK, Stehle P, Koerfer R.

Vitamin D supplementation improves cytokine profiles in patients with congestive heart

failure: a double-blind, randomized, placebo-controlled trial. Am J Clin Nutr

2006;83:754-9.

260. Wu CC, Chang JH, Chen CC, Su SB, Yang LK, Ma WY, Zheng CM, Diang LK,

Lu KC. Calcitriol treatment attenuates inflammation and oxidative stress in hemodialysis

patients with secondary hyperparathyroidism. Tohoku J Exp Med 2011;223:153-9.

261. Jorde R, Sneve M, Torjesen PA, Figenschau Y, Goransson LG, Omdal R. No

effect of supplementation with cholecalciferol on cytokines and markers of inflammation

in overweight and obese subjects. Cytokine 2010;50:175-80.

166

262. Barnes MS, Horigan G, Cashman KD, Hill TR, Forsythe LK, Lucey AJ,

McSorley EM, Kiely M, Bonham MP, Magee PJ et al. Maintenance of wintertime

vitamin D status with cholecalciferol supplementation is not associated with alterations in

serum cytokine concentrations among apparently healthy younger or older adults. J Nutr

2011;141:476-81.

263. Gepner AD, Ramamurthy R, Krueger DC, Korcarz CE, Binkley N, Stein JH. A

prospective randomized controlled trial of the effects of vitamin d supplementation on

cardiovascular disease risk. PLoS One 2012;7:e36617.

264. Vilarrasa N, Vendrell J, Maravall J, Elio I, Solano E, San JP, Garcia I, Virgili N,

Soler J, Gomez JM. Is plasma 25(OH) D related to adipokines, inflammatory cytokines

and insulin resistance in both a healthy and morbidly obese population? Endocrine

2010;38:235-42.

265. Shea MK, Booth SL, Massaro JM, Jacques PF, D'Agostino RB, Sr., wson-Hughes

B, Ordovas JM, O'Donnell CJ, Kathiresan S, Keaney JF, Jr. et al. Vitamin K and vitamin

D status: associations with inflammatory markers in the Framingham Offspring Study.

Am J Epidemiol 2008;167:313-20.

266. Yiu YF, Yiu KH, Siu CW, Chan YH, Li SW, Wong LY, Lee SW, Tam S, Wong

EW, Lau CP et al. Randomized controlled trial of vitamin D supplement on endothelial

function in patients with type 2 diabetes. Atherosclerosis 2013;227:140-6.

267. Amer M, Qayyum R. Relation between serum 25-hydroxyvitamin D and C-

reactive protein in asymptomatic adults (from the continuous National Health and

Nutrition Examination Survey 2001 to 2006). Am J Cardiol 2012;109:226-30.

268. Neyestani TR, Djazayery A, Shab-Bidar S, Eshraghian MR, Kalayi A,

Shariatzadeh N, Khalaji N, Zahedirad M, Gharavi A, Houshiarrad A et al. Vitamin D

Receptor Fok-I Polymorphism Modulates Diabetic Host Response to Vitamin D Intake:

Need for a nutrigenetic approach. Diabetes Care 2012.

269. Doro P, Benko R, Matuz M, Soos G. Seasonality in the incidence of type 2

diabetes: a population-based study. Diabetes Care 2006;29:173.

270. Holick MF. Diabetes and the vitamin d connection. Curr Diab Rep 2008;8:393-8.

271. Inomata S, Kadowaki S, Yamatani T, Fukase M, Fujita T. Effect of 1 alpha (OH)-

vitamin D3 on insulin secretion in diabetes mellitus. Bone Miner 1986;1:187-92.

272. Bourlon PM, Billaudel B, Faure-Dussert A. Influence of vitamin D3 deficiency

and 1,25 dihydroxyvitamin D3 on de novo insulin biosynthesis in the islets of the rat

endocrine pancreas. J Endocrinol 1999;160:87-95.

273. Pittas AG, Harris SS, Stark PC, Dawson-Hughes B. The effects of calcium and

vitamin D supplementation on blood glucose and markers of inflammation in nondiabetic

adults. Diabetes Care 2007;30:980-6.

274. Riachy R, Vandewalle B, Kerr CJ, Moerman E, Sacchetti P, Lukowiak B, Gmyr

V, Bouckenooghe T, Dubois M, Pattou F. 1,25-dihydroxyvitamin D3 protects RINm5F

and human islet cells against cytokine-induced apoptosis: implication of the antiapoptotic

protein A20. Endocrinology 2002;143:4809-19.

167

275. Baier LJ, Dobberfuhl AM, Pratley RE, Hanson RL, Bogardus C. Variations in the

vitamin D-binding protein (Gc locus) are associated with oral glucose tolerance in

nondiabetic Pima Indians. J Clin Endocrinol Metab 1998;83:2993-6.

276. Szathmary EJ. The effect of Gc genotype on fasting insulin level in Dogrib

Indians. Hum Genet 1987;75:368-72.

277. Hirai M, Suzuki S, Hinokio Y, Hirai A, Chiba M, Akai H, Suzuki C, Toyota T.

Variations in vitamin D-binding protein (group-specific component protein) are

associated with fasting plasma insulin levels in Japanese with normal glucose tolerance. J


278. Wang Q, Xi B, Reilly KH, Liu M, Fu M. Quantitative assessment of the

associations between four polymorphisms (FokI, ApaI, BsmI, TaqI) of vitamin D

receptor gene and risk of diabetes mellitus. Mol Biol Rep 2012;39:9405-14.

279. Ford ES, Ajani UA, McGuire LC, Liu S. Concentrations of serum vitamin D and

the metabolic syndrome among U.S. adults. Diabetes Care 2005;28:1228-30.

280. Chiu KC, Chu A, Go VL, Saad MF. Hypovitaminosis D is associated with insulin

resistance and beta cell dysfunction. Am J Clin Nutr 2004;79:820-5.

281. Hypponen E, Power C. Vitamin D status and glucose homeostasis in the 1958

British birth cohort: the role of obesity. Diabetes Care 2006;29:2244-6.

282. Need AG, O'Loughlin PD, Horowitz M, Nordin BE. Relationship between fasting

serum glucose, age, body mass index and serum 25 hydroxyvitamin D in postmenopausal

women. Clin Endocrinol (Oxf) 2005;62:738-41.

283. Hurskainen AR, Virtanen JK, Tuomainen TP, Nurmi T, Voutilainen S.

Association of serum 25-hydroxyvitamin D with type 2 diabetes and markers of insulin

resistance in a general older population in Finland. Diabetes Metab Res Rev

2012;28:418-23.

284. Kabadi SM, Lee BK, Liu L. Joint effects of obesity and vitamin D insufficiency

on insulin resistance and type 2 diabetes: results from the NHANES 2001-2006. Diabetes

Care 2012;35:2048-54.

285. Pittas AG, Dawson-Hughes B, Li T, Van Dam RM, Willett WC, Manson JE, Hu

FB. Vitamin D and calcium intake in relation to type 2 diabetes in women. Diabetes Care

2006;29:650-6.

286. Fliser D, Stefanski A, Franek E, Fode P, Gudarzi A, Ritz E. No effect of calcitriol

on insulin-mediated glucose uptake in healthy subjects. Eur J Clin Invest 1997;27:629-

33.

287. Ljunghall S, Lind L, Lithell H, Skarfors E, Selinus I, Sorensen OH, Wide L.

Treatment with one-alpha-hydroxycholecalciferol in middle-aged men with impaired

glucose tolerance--a prospective randomized double-blind study. Acta Med Scand

1987;222:361-7.

288. Orwoll E, Riddle M, Prince M. Effects of vitamin D on insulin and glucagon

secretion in non-insulin-dependent diabetes mellitus. Am J Clin Nutr 1994;59:1083-7.

168

289. Nilas L, Christiansen C. Treatment with vitamin D or its analogues does not

change body weight or blood glucose level in postmenopausal women. Int J Obes

1984;8:407-11.

290. von Hurst PR, Stonehouse W, Coad J. Vitamin D supplementation reduces insulin

resistance in South Asian women living in New Zealand who are insulin resistant and

vitamin D deficient - a randomised, placebo-controlled trial. Br J Nutr 2010;103:549-55.

291. Jorde R, Grimnes G. Vitamin D and metabolic health with special reference to the

effect of vitamin D on serum lipids. Prog Lipid Res 2011;50:303-12.

292. Jaimungal S, Wehmeier K, Mooradian AD, Haas MJ. The emerging evidence for

vitamin D-mediated regulation of apolipoprotein A-I synthesis. Nutr Res 2011;31:805-

12.

293. Johnson MD, Nader NS, Weaver AL, Singh R, Kumar S. Relationships between

25-hydroxyvitamin D levels and plasma glucose and lipid levels in pediatric outpatients.

J Pediatr 2010;156:444-9.

294. Wang JH, Keisala T, Solakivi T, Minasyan A, Kalueff AV, Tuohimaa P. Serum

cholesterol and expression of ApoAI, LXRbeta and SREBP2 in vitamin D receptor

knock-out mice. J Steroid Biochem Mol Biol 2009;113:222-6.

295. Wehmeier K, Beers A, Haas MJ, Wong NC, Steinmeyer A, Zugel U, Mooradian

AD. Inhibition of apolipoprotein AI gene expression by 1, 25-dihydroxyvitamin D3.

Biochim Biophys Acta 2005;1737:16-26.

296. Barger-Lux MJ, Heaney RP, Lanspa SJ, Healy JC, DeLuca HF. An investigation

of sources of variation in calcium absorption efficiency. J Clin Endocrinol Metab

1995;80:406-11.

297. Welberg JW, Monkelbaan JF, de Vries EG, Muskiet FA, Cats A, Oremus ET,

Boersma-van EW, van RH, van der MR, Mulder NH et al. Effects of supplemental

dietary calcium on quantitative and qualitative fecal fat excretion in man. Ann Nutr

Metab 1994;38:185-91.

298. Vaskonen T, Mervaala E, Sumuvuori V, Seppanen-Laakso T, Karppanen H.

Effects of calcium and plant sterols on serum lipids in obese Zucker rats on a low-fat diet.

Br J Nutr 2002;87:239-45.

299. Cho HJ, Kang HC, Choi SA, Ju YC, Lee HS, Park HJ. The possible role of Ca2+

on the activation of microsomal triglyceride transfer protein in rat hepatocytes. Biol

Pharm Bull 2005;28:1418-23.

300. Lips P. Vitamin D deficiency and secondary hyperparathyroidism in the elderly:

consequences for bone loss and fractures and therapeutic implications. Endocr Rev

2001;22:477-501.

301. Lacour B, Basile C, Drueke T, Funck-Brentano JL. Parathyroid function and lipid

metabolism in the rat. Miner Electrolyte Metab 1982;7:157-65.

302. Karhapaa P, Pihlajamaki J, Porsti I, Kastarinen M, Mustonen J, Niemela O,

Kuusisto J. Diverse associations of 25-hydroxyvitamin D and 1,25-dihydroxy-vitamin D

with dyslipidaemias. J Intern Med 2010;268:604-10.

169

303. Liu E, Meigs JB, Pittas AG, McKeown NM, Economos CD, Booth SL, Jacques

PF. Plasma 25-hydroxyvitamin D is associated with markers of the insulin resistant

phenotype in nondiabetic adults. Journal of Nutrition 2009;139:329-34.

304. Pinelli NR, Jaber LA, Brown MB, Herman WH. Serum 25-hydroxy vitamin d and

insulin resistance, metabolic syndrome, and glucose intolerance among Arab Americans.

Diabetes Care 2010;33:1373-5.

305. Botella-Carretero JI, Alvarez-Blasco F, Villafruela JJ, Balsa JA, Vazquez C,

Escobar-Morreale HF. Vitamin D deficiency is associated with the metabolic syndrome

in morbid obesity. Clin Nutr 2007;26:573-80.

306. Wang H, Xia N, Yang Y, Peng DQ. Influence of vitamin D supplementation on

plasma lipid profiles: a meta-analysis of randomized controlled trials. Lipids Health Dis

2012;11:42.

307. Multhoff G, Radons J. Radiation, inflammation, and immune responses in cancer.

Front Oncol 2012;2:58.

308. Choy E. Understanding the dynamics: pathways involved in the pathogenesis of

rheumatoid arthritis. Rheumatology (Oxford) 2012;51 Suppl 5:v3-11.

309. Comabella M, Khoury SJ. Immunopathogenesis of multiple sclerosis. Clin

Immunol 2012;142:2-8.

310. Wang Y, Jacobs EJ, McCullough ML, Rodriguez C, Thun MJ, Calle EE, Flanders

WD. Comparing methods for accounting for seasonal variability in a biomarker when

only a single sample is available: insights from simulations based on serum 25-

hydroxyvitamin d. Am J Epidemiol 2009;170:88-94.

311. Cahill L, Corey PN, El-Sohemy A. Vitamin C deficiency in a population of young

Canadian adults. Am J Epidemiol 2009;170:464-71.

312. Fontaine-Bisson B, Wolever TM, Connelly PW, Corey PN, El-Sohemy A. NF-

kappaB -94Ins/Del ATTG polymorphism modifies the association between dietary

polyunsaturated fatty acids and HDL-cholesterol in two distinct populations.

Atherosclerosis 2009;204:465-70.

313. Eny KM, Corey PN, El-Sohemy A. Dopamine D2 receptor genotype (C957T) and

habitual consumption of sugars in a free-living population of men and women. J

Nutrigenet Nutrigenomics 2009;2:235-42.

314. USDA National Nutrient Database for Standard Reference, Release 20. US

Department of Agriculture Agricultural Research Service 2007.

http://www.ars.usda.gov/main/site_main.htm?modecode=12-35-45-00. Accessed March

2010.

315. Wang Y, Zhu J, DeLuca HF. Where is the vitamin D receptor? Arch Biochem

Biophys 2012;523:123-33.

316. Buchanan JR, Santen R, Cauffman S, Cavaliere A, Greer RB, Demers LM. The

effect of endogenous estrogen fluctuation on metabolism of 25-hydroxyvitamin D. Calcif

Tissue Int 1986;39:139-44.

170

317. Gagnon C, Baillargeon JP, Desmarais G, Fink GD. Prevalence and predictors of

vitamin D insufficiency in women of reproductive age living in northern latitude. Eur J

Endocrinol 2010;163:819-24.

318. Hotamisligil GS. Inflammation and metabolic disorders. Nature 2006;444:860-7.

319. Garcia-Bailo B, El-Sohemy A, Haddad PS, Arora P, BenZaied F, Karmali M,

Badawi A. Vitamins D, C, and E in the prevention of type 2 diabetes mellitus:

modulation of inflammation and oxidative stress. Biologics 2011;5:7-19.

320. Pickup JC, Crook MA. Is Type II diabetes mellitus a disease of the innate immune

system? Diabetologia 1998;41:1241-8.

321. Da Costa LA, Garcia-Bailo B, Badawi A, El-Sohemy A. Genetic determinants of

dietary antioxidant status. Prog Mol Biol Transl Sci 2012;108:179-200.

322. Evans JL, Goldfine ID, Maddux BA, Grodsky GM. Oxidative stress and stress-

activated signaling pathways: a unifying hypothesis of type 2 diabetes. Endocr Rev

2002;23:599-622.

323. Clendenen TV, Koenig KL, Arslan AA, Lukanova A, Berrino F, Gu Y, Hallmans

G, Idahl A, Krogh V, Lokshin AE et al. Factors associated with inflammation markers, a

cross-sectional analysis. Cytokine 2011;56:769-78.

324. Leng SX, McElhaney JE, Walston JD, Xie D, Fedarko NS, Kuchel GA. ELISA

and multiplex technologies for cytokine measurement in inflammation and aging

research. J Gerontol A Biol Sci Med Sci 2008;63:879-84.

325. Perrier S, Darakhshan F, Hajduch E. IL-1 receptor antagonist in metabolic

diseases: Dr Jekyll or Mr Hyde? FEBS Lett 2006;580:6289-94.

326. Gysemans C, Callewaert H, Overbergh L, Mathieu C. Cytokine signalling in the

beta-cell: a dual role for IFNgamma. Biochem Soc Trans 2008;36:328-33.

327. Rotondi M, Lazzeri E, Romagnani P, Serio M. Role for interferon-gamma

inducible chemokines in endocrine autoimmunity: an expanding field. J Endocrinol

Invest 2003;26:177-80.

328. Hellberg C, Ostman A, Heldin CH. PDGF and vessel maturation. Recent Results

Cancer Res 2010;180:103-14.

329. Suffee N, Richard B, Hlawaty H, Oudar O, Charnaux N, Sutton A. Angiogenic

properties of the chemokine RANTES/CCL5. Biochem Soc Trans 2011;39:1649-53.

330. Navarro-Gonzalez JF, Mora-Fernandez C, Muros de FM, Garcia-Perez J.

Inflammatory molecules and pathways in the pathogenesis of diabetic nephropathy. Nat

Rev Nephrol 2011;7:327-40.

331. Garcia-Bailo B, Josse AR, Jamnik J, Badawi A, El-Sohemy A. Positive

Association Between 25-Hydroxyvitamin D and C-Reactive Protein is Confounded by

Hormonal Contraceptive Use. J Womens Health (Larchmt ) 2013;22:417-25.

332. Yamshchikov AV, Desai NS, Blumberg HM, Ziegler TR, Tangpricha V. Vitamin

D for treatment and prevention of infectious diseases: a systematic review of randomized

controlled trials. Endocr Pract 2009;15:438-49.

171

333. Sacheck J, Goodman E, Chui K, Chomitz V, Must A, Economos C. Vitamin D

deficiency, adiposity, and cardiometabolic risk in urban schoolchildren. J Pediatr

2011;159:945-50.

334. Jin Y, Cui X, Singh UP, Chumanevich AA, Harmon B, Cavicchia P, Hofseth AB,

Kotakadi V, Stroud B, Volate SR et al. Systemic inflammatory load in humans is

suppressed by consumption of two formulations of dried, encapsulated juice concentrate.

Mol Nutr Food Res 2010;54:1506-14.

335. Appay V, Rowland-Jones SL. RANTES: a versatile and controversial chemokine.

Trends Immunol 2001;22:83-7.

336. Virani SS, Nambi V, Hoogeveen R, Wasserman BA, Coresh J, Gonzalez F,

Chambless LE, Mosley TH, Boerwinkle E, Ballantyne CM. Relationship between

circulating levels of RANTES (regulated on activation, normal T-cell expressed, and

secreted) and carotid plaque characteristics: the Atherosclerosis Risk in Communities

(ARIC) Carotid MRI Study. Eur Heart J 2011;32:459-68.

337. Sitruk-Ware R, Nath A. Metabolic effects of contraceptive steroids. Rev Endocr

Metab Disord 2011;12:63-75.

338. Baer PC, Gauer S, Wegner B, Schubert R, Geiger H. C-reactive protein induced

activation of MAP-K and RANTES in human renal distal tubular epithelial cells in vitro.

Clin Nephrol 2006;66:177-83.

339. Kovalovsky D, Paez PM, Sauer J, Perez CC, Nahmod VE, Stalla GK, Holsboer F,

Arzt E. The Th1 and Th2 cytokines IFN-gamma and IL-4 antagonize the inhibition of

monocyte IL-1 receptor antagonist by glucocorticoids: involvement of IL-1. Eur J

Immunol 1998;28:2075-85.

340. Badgett A, Bonner JC, Brody AR. Interferon-gamma modulates lung macrophage

production of PDGF-BB and fibroblast growth. J Lipid Mediat Cell Signal 1996;13:89-

97.

341. Aschner P. Metabolic syndrome as a risk factor for diabetes. Expert Rev

Cardiovasc Ther 2010;8:407-12.

342. Alvarez JA, Gower BA, Calhoun DA, Judd SE, Dong Y, Dudenbostel T, Scholl J,

Ashraf AP. Serum 25-hydroxyvitamin D and Ethnic Differences in Arterial Stiffness and

Endothelial Function. J Clin Med Res 2012;4:197-205.

343. Garcia-Bailo B, Jamnik J, Da Costa LA, Borchers CH, Badawi A, El-Sohemy A.

Plasma 25-Hydroxyvitamin D, Hormonal Contraceptive Use, and the Plasma Proteome in

Caucasian, East Asian, and South Asian Young Adults. J Proteome Res 2013;12:1797-

807.

344. McLaughlin T, Abbasi F, Cheal K, Chu J, Lamendola C, Reaven G. Use of

metabolic markers to identify overweight individuals who are insulin resistant. Ann

Intern Med 2003;139:802-9.

345. Cheang KI, Essah PA, Sharma S, Wickham EP, III, Nestler JE. Divergent effects

of a combined hormonal oral contraceptive on insulin sensitivity in lean versus obese

women. Fertil Steril 2011;96:353-9.

172

346. Ponda MP, Huang X, Odeh MA, Breslow JL, Kaufman HW. Vitamin D may not

improve lipid levels: a serial clinical laboratory data study. Circulation 2012;126:270-7.

347. Kim J, Mueller C. Factor analysis: statistical methods and practical issues.

Newbury Park, CA: Sage Publications, Inc., 1978.

348. Reactome. 2012. http://www.reactome.org. Accessed February 2011.

349. Pittas AG, Dawson-Hughes B. Vitamin D and diabetes. J Steroid Biochem Mol

Biol 2010;121:425-9.

350. Starck M, Schiele F, Herbeth B, Vincent-Viry M, Beaud B, Siest G, Visvikis S.

Apolipoproteins E and C-III in apo B- and non-apo B-containing lipoproteins in middle-

aged women from the Stanislas cohort: effect of oral contraceptive use and common

apolipoprotein E polymorphism. Atherosclerosis 2001;155:509-16.

351. Zugaro A, Pandolfi C, Barbonetti A, Vassallo MR, D'Angeli A, Necozione S,

Colangeli MS, Francavilla S, Francavilla F. Retinol binding protein 4, low birth weight-

related insulin resistance and hormonal contraception. Endocrine 2007;32:166-9.

352. Kotnik P, Fischer-Posovszky P, Wabitsch M. RBP4: a controversial adipokine.

Eur J Endocrinol 2011;165:703-11.

353. Maas C, Renne T. Regulatory mechanisms of the plasma contact system. Thromb

Res 2012;129 Suppl 2:S73-S76.

354. Stricker H, Tosi BF, Guidicelli-Nicolosi S, Limoni C, Colucci G. Effect of a

single, oral, high-dose vitamin D supplementation on endothelial function in patients with

peripheral arterial disease: a randomised controlled pilot study. Eur J Vasc Endovasc

Surg 2012;44:307-12.

355. Nimitphong H, Holick MF, Fried SK, Lee MJ. 25-hydroxyvitamin d(3) and 1,25-

dihydroxyvitamin d(3) promote the differentiation of human subcutaneous preadipocytes.

PLoS One 2012;7:e52171.

356. Bland R, Markovic D, Hills CE, Hughes SV, Chan SLF, Squires PE, Hewison M.

Expression of 25-hydroxyvitamin D3-1-alpha-hydroxylase in pancreatic islets. Journal of

Steroid Biochemistry and Molecular Biology 2004;89-90:121-5.

357. Brondum-Jacobsen P, Benn M, Jensen GB, Nordestgaard BG. 25-hydroxyvitamin

d levels and risk of ischemic heart disease, myocardial infarction, and early death:

population-based study and meta-analyses of 18 and 17 studies. Arterioscler Thromb

Vasc Biol 2012;32:2794-802.

358. DeLuca HF. Overview of general physiologic features and functions of vitamin

D. The American Journal of Clinical Nutrition 2004;80:1698S-96S.

359. Valdivielso JM, Fernandez E. Vitamin D receptor polymorphisms and diseases.

Clin Chim Acta 2006;371:1-12.

360. Reis AF, Hauache OM, Velho G. Vitamin D endocrine system and the genetic

susceptibility to diabetes, obesity and vascular disease. A review of evidence. Diabetes

Metab 2005;31:318-25.

173

361. Josse AR, Da Costa LA, Campos H, El-Sohemy A. Associations between

polymorphisms in the AHR and CYP1A1-CYP1A2 gene regions and habitual caffeine

consumption. Am J Clin Nutr 2012;96:665-71.

362. Schroder K, Hertzog PJ, Ravasi T, Hume DA. Interferon-gamma: an overview of

signals, mechanisms and functions. J Leukoc Biol 2004;75:163-89.

363. Selvaraj P, Harishankar M, Singh B, Banurekha VV, Jawahar MS. Effect of

vitamin D3 on chemokine expression in pulmonary tuberculosis. Cytokine 2012;60:212-

9.

364. Mutti DO, Cooper ME, Dragan E, Jones-Jordan LA, Bailey MD, Marazita ML,

Murray JC, Zadnik K. Vitamin D receptor (VDR) and group-specific component (GC,

vitamin D-binding protein) polymorphisms in myopia. Invest Ophthalmol Vis Sci

2011;52:3818-24.

365. Sung CC, Liao MT, Lu KC, Wu CC. Role of vitamin D in insulin resistance. J

Biomed Biotechnol 2012;2012:634195.

366. Comings DE, MacMurray JP. Molecular heterosis: a review. Mol Genet Metab

2000;71:19-31.

367. Jain R, von Hurst PR, Stonehouse W, Love DR, Higgins CM, Coad J. Association

of vitamin D receptor gene polymorphisms with insulin resistance and response to

vitamin D. Metabolism 2012;61:293-301.

174

Appendices

Table A.1. Mean plasma protein concentrations (μmol/L) by vitamin D status, by sex and

contraceptive use1

a) Women HC users

Plasma 25D (nmol/L)

<30

(n=14)

30 to 49.9

(n=35)

50 to 74.9

(n=74)

≥75

(n=93) p

Afamin 0.29 ± 0.02a,b

0.25 ± 0.01a 0.28 ± 0.01

b 0.3 ± 0.01

b <0.0001 §

Angiotensinogen 2.1 ± 0.23a,b

1.67 ± 0.15a 1.88 ± 0.1

a 2.36 ± 0.09

b <0.0001 §

Apolipoprotein C-III 2.84 ± 0.21a,b,c

2.37 ± 0.14a 2.84 ± 0.09

b 3.36 ± 0.08

c <0.0001 §

Ceruloplasmin 4.32 ± 0.31a,b

3.28 ± 0.2a 3.35 ± 0.14

a 3.86 ± 0.12

b <0.0001 §

Vitamin D binding

protein 3.55 ± 0.21

a,b,c 3.16 ± 0.13

a 3.54 ± 0.09

b 3.93 ± 0.08

c <0.0001 §

Retinol-binding protein 1.13 ± 0.07a,b

1.02 ± 0.05a 1.09 ± 0.03

a 1.26 ± 0.03

b <0.0001 §

Apolipoprotein A-I 52.7 ± 2.77a 45.16 ± 1.75

b 49.2 ± 1.2

a,b 53.75 ± 1.07

a,c 0.0002 §

Transferrin 15.75 ± 0.83a,b

13.32 ± 0.53a 13.5 ± 0.36

a 15.24 ± 0.32

b 0.0003

Apolipoprotein C-I 3.43 ± 0.21a,b

3.06 ± 0.13a 3.44 ± 0.09

b 3.68 ± 0.08

b 0.0004

Heparin cofactor II 0.94 ± 0.06a,b

0.79 ± 0.04a 0.81 ± 0.02

a 0.91 ± 0.02

b 0.0004

Kininogen-1 2.81 ± 0.16a,b

2.49 ± 0.1a 2.62 ± 0.07

a 2.91 ± 0.06

b 0.0004

Apolipoprotein A-II

precursor 28.96 ± 1.55

a,b 27.41 ± 0.98

a 29.27 ± 0.67

a 31.71 ± 0.6

b 0.0005



HC 0.38 ± 0.03 0.34 ± 0.02 0.36 ± 0.01 0.41 ± 0.01 0.0042

β2-Glycoprotein I 2.75 ± 0.18 2.46 ± 0.11 2.68 ± 0.08 2.88 ± 0.07 0.0050


HC 0.76 ± 0.03 0.63 ± 0.02 0.66 ± 0.01 0.69 ± 0.01 0.0070

α2-Antiplasmin 2.17 ± 0.1 1.84 ± 0.07 1.95 ± 0.05 2.03 ± 0.04 0.0076

α1-Antitrypsin 15.2 ± 0.91 13.13 ± 0.57 13.76 ± 0.4 14.8 ± 0.35 0.0141

Hemopexin 11.9 ± 0.52 10.57 ± 0.33 11 ± 0.23 11.68 ± 0.2 0.0143

α2-Macroglobulin 6.48 ± 0.46 5.22 ± 0.29 5.76 ± 0.2 6.14 ± 0.18 0.0172

Fibronectin 0.48 ± 0.34 0.62 ± 0.22 0.85 ± 0.15 0.49 ± 0.13 0.0207

Adiponectin 0.08 ± 0.01 0.07 ± 0.01 0.07 ± 0.01 0.08 ± 0.01 0.0286

Complement C3 23.67 ± 1.18 20.18 ± 0.75 21.35 ± 0.51 21.98 ± 0.46 0.0444

175

Transthyretin 6.25 ± 0.3 5.64 ± 0.19 5.78 ± 0.13 6.11 ± 0.12 0.0445

Clusterin 1.75 ± 0.08 1.53 ± 0.05 1.62 ± 0.04 1.68 ± 0.03 0.0503

Plasminogen 1.67 ± 0.08 1.42 ± 0.05 1.48 ± 0.04 1.52 ± 0.03 0.0537

Albumin 958.17 ± 32.97 859.59 ± 20.85 894.33 ± 14.34 880.46 ± 12.79 0.0670

Antithrombin-III 3.54 ± 0.14 3.20 ± 0.09 3.31 ± 0.06 3.43 ± 0.05 0.0755


Vitronectin 5.13 ± 0.26 4.45 ± 0.17 4.49 ± 0.11 4.70 ± 0.10 0.0842


Prothrombin 0.62 ± 0.03 0.58 ± 0.02 0.62 ± 0.01 0.63 ± 0.01 0.1304

α1B-Glycoprotein 1.72 ± 0.16 1.72 ± 0.1 1.92 ± 0.07 1.89 ± 0.06 0.1639

Histidine-rich

glycoprotein 0.93 ± 0.1 0.96 ± 0.06 1.1 ± 0.04 1.05 ± 0.04 0.1642

Complement C4 β chain 1.51 ± 0.13 1.31 ± 0.08 1.51 ± 0.06 1.38 ± 0.05 0.1997

Fibrinopeptide A 7.82 ± 0.88 7.29 ± 0.56 8.2 ± 0.38 7.40 ± 0.34 0.2267

Serum amyloid p-

component 0.59 ± 0.04 0.50 ± 0.03 0.54 ± 0.02 0.50 ± 0.02 0.2288

Complement C4 γ chain 1.74 ± 0.15 1.47 ± 0.09 1.64 ± 0.06 1.49 ± 0.06 0.2297

Complement factor B 1.72 ± 0.11 1.48 ± 0.07 1.59 ± 0.05 1.58 ± 0.04 0.2343

Fibrinogen γ chain 11.39 ± 1.35 9.41 ± 0.85 11.04 ± 0.59 10.24 ± 0.52 0.2476

Apolipoprotein A-IV 1.29 ± 0.11 1.21 ± 0.07 1.31 ± 0.05 1.40 ± 0.04 0.2544

α1-Antichymotrypsin 3.54 ± 0.21 3.09 ± 0.13 3.29 ± 0.09 3.36 ± 0.08 0.2550

L-Selectin 0.08 ± 0.01 0.07 ± 0.01 0.07 ± 0.01 0.07 ± 0.01 0.2744

Fibrinogen α2 chain 13.24 ± 2.01 11.82 ± 1.27 14.38 ± 0.87 12.95 ± 0.78 0.2790

Fibrinogen β chain 10.62 ± 1.23 9.51 ± 0.78 11.04 ± 0.53 10.25 ± 0.48 0.3256

Complement C9 2.86 ± 0.22 2.55 ± 0.14 2.77 ± 0.1 2.73 ± 0.09 0.3868

Complement C1

inactivator 4.12 ± 0.33 3.93 ± 0.21 3.8 ± 0.15 3.74 ± 0.13 0.4445

Gelsolin, isoform 1 1.15 ± 0.07 1.04 ± 0.04 1.09 ± 0.03 1.12 ± 0.03 0.4491

α1-Acid glycoprotein 1 1.47 ± 0.15 1.48 ± 0.1 1.6 ± 0.07 1.54 ± 0.06 0.5674

Haptoglobin β chain 11.55 ± 1.32 10.6 ± 0.83 9.65 ± 0.57 10.16 ± 0.51 0.7292

Zinc- α2-glycoprotein 1.04 ± 0.10 1.00 ± 0.06 1.05 ± 0.04 1.11 ± 0.04 0.7834

Apolipoprotein E 0.42 ± 0.04 0.43 ± 0.02 0.43 ± 0.02 0.45 ± 0.01 0.8376

Complement factor H 0.66 ± 0.04 0.63 ± 0.02 0.64 ± 0.02 0.64 ± 0.01 0.8441

Apolipoprotein D 0.32 ± 0.02 0.33 ± 0.01 0.33 ± 0.01 0.34 ± 0.01 0.9052

176

b) Women HC non-users

Plasma 25D (nmol/L)

<30

(n=103)

30 to 49.9

(n=202)

50 to 74.9

(n=149)

≥75

(n=48) p

Afamin 0.25 ± 0.01 0.25 ± 0 0.25 ± 0.01 0.23 ± 0.01 0.3589



Ceruloplasmin 2.06 ± 0.06 1.94 ± 0.04 2.08 ± 0.05 2.18 ± 0.09 0.0543

Vitamin D binding

protein 2.54 ± 0.05 2.57 ± 0.04 2.69 ± 0.04 2.78 ± 0.07 0.0325

Retinol-binding protein 0.76 ± 0.02a 0.79 ± 0.02

a,b 0.85 ± 0.02

b,c 0.93 ± 0.03

c 0.0001 §


Transferrin 12.61 ± 0.3 12.14 ± 0.21 12.48 ± 0.25 12.43 ± 0.44 0.3461



Kininogen-1 2 ± 0.04 1.95 ± 0.03 2.03 ± 0.03 2.09 ± 0.06 0.3359

Apolipoprotein A-II

precursor 23.58 ± 0.5 23.38 ± 0.35 23.87 ± 0.41 23.69 ± 0.73 0.9470



HC 0.21 ± 0.01

a 0.21 ± 0.01

a 0.26 ± 0.01

b 0.27 ± 0.01

b <0.0001 §

β2-Glycoprotein I 2.69 ± 0.06 2.67 ± 0.04 2.82 ± 0.05 2.86 ± 0.09 0.1446


HC 0.62 ± 0.01 0.61 ± 0.01 0.61 ± 0.01 0.61 ± 0.02 0.7522

α2-Antiplasmin 1.96 ± 0.04 1.89 ± 0.03 1.94 ± 0.04 1.93 ± 0.06 0.4369

α1-Antitrypsin 10.9 ± 0.22 10.16 ± 0.16 10.38 ± 0.19 10.58 ± 0.33 0.0290

Hemopexin 10.3 ± 0.22 9.73 ± 0.16 10.35 ± 0.18 10.22 ± 0.32 0.0348

α2-Macroglobulin 6.4 ± 0.16 6.21 ± 0.11 6.07 ± 0.13 6.05 ± 0.24 0.7979

Fibronectin 0.66 ± 0.1 0.54 ± 0.07 0.64 ± 0.08 0.65 ± 0.15 0.2095

Adiponectin 0.06 ± 0 0.07 ± 0 0.07 ± 0 0.08 ± 0 0.0012

Complement C3 19.13 ± 0.46 18.42 ± 0.33 19.01 ± 0.38 18.25 ± 0.67 0.1247

Transthyretin 5.17 ± 0.12 5.39 ± 0.09 5.43 ± 0.1 5.47 ± 0.18 0.3791

Clusterin 1.49 ± 0.03 1.49 ± 0.02 1.52 ± 0.03 1.54 ± 0.05 0.5858

Plasminogen 1.18 ± 0.02 1.16 ± 0.02 1.16 ± 0.02 1.19 ± 0.03 0.6600

Albumin 966.57 ± 15.54 969.09 ± 11.1 969.77 ± 12.92 965.85 ± 22.76 0.9966



177

Vitronectin 3.66 ± 0.07 3.49 ± 0.05 3.54 ± 0.06 3.41 ± 0.1 0.0412


Prothrombin 0.58 ± 0.01 0.57 ± 0.01 0.57 ± 0.01 0.55 ± 0.02 0.3779

α1B-Glycoprotein 1.69 ± 0.05 1.6 ± 0.03 1.73 ± 0.04 1.71 ± 0.07 0.1000

Histidine-rich

glycoprotein 1.42 ± 0.04 1.41 ± 0.03 1.38 ± 0.03 1.42 ± 0.06 0.7065



Serum amyloid p-

component 0.38 ± 0.01 0.38 ± 0.01 0.38 ± 0.01 0.37 ± 0.02 0.3564






L-Selectin 0.07 ± 0 0.07 ± 0 0.08 ± 0 0.08 ± 0 0.0838



Complement C9 2.91 ± 0.08 2.75 ± 0.06 2.87 ± 0.07 2.79 ± 0.12 0.1766

Complement C1

inactivator 4.97 ± 0.11 4.86 ± 0.08 4.85 ± 0.09 4.7 ± 0.17 0.2838








c) Men

Plasma 25D (nmol/L)

<30

(n=62)

30 to 49.9

(n=118)

50 to 74.9

(n=87)

≥75

(n=34) p

Afamin 0.26 ± 0.01 0.25 ± 0.01 0.24 ± 0.01 0.23 ± 0.01 0.0603


178


Ceruloplasmin 1.98 ± 0.07 1.81 ± 0.05 1.84 ± 0.06 1.91 ± 0.1 0.1901

Vitamin D binding

protein 2.57 ± 0.06 2.55 ± 0.04 2.68 ± 0.05 2.81 ± 0.08 0.1406

Retinol-binding protein 0.9 ± 0.03 0.95 ± 0.02 0.97 ± 0.03 1.02 ± 0.04 0.0741


Transferrin 12.01 ± 0.34 11.64 ± 0.25 11.36 ± 0.29 11.47 ± 0.47 0.5161

Apolipoprotein C-I 2.99 ± 0.1 3.06 ± 0.07 3.07 ± 0.08 3 ± 0.14 0.889


Kininogen-1 2.11 ± 0.05 2.02 ± 0.04 2 ± 0.04 2.01 ± 0.07 0.3258

Apolipoprotein A-II

precursor 24.39 ± 0.62 24.38 ± 0.45 24.84 ± 0.53 23.64 ± 0.84 0.7895

α2-HS-glycoprotein 9.14 ± 0.24 8.42 ± 0.17 8.47 ± 0.2 8 ± 0.32 0.0547


HC 0.23 ± 0.01 0.24 ± 0.01 0.25 ± 0.01 0.28 ± 0.02 0.0046

β2-Glycoprotein I 3 ± 0.09 2.91 ± 0.06 2.88 ± 0.07 2.99 ± 0.12 0.692


HC 0.64 ± 0.02 0.6 ± 0.01 0.61 ± 0.01 0.59 ± 0.02 0.279

α2-Antiplasmin 1.97 ± 0.05 1.89 ± 0.04 1.89 ± 0.04 1.83 ± 0.07 0.6426

α1-Antitrypsin 10.63 ± 0.27 9.99 ± 0.2 9.71 ± 0.23 10.16 ± 0.37 0.0706

Hemopexin 9.82 ± 0.26 9.37 ± 0.19 9.24 ± 0.22 9.64 ± 0.35 0.4086

α2-Macroglobulin 5.81 ± 0.22 5.19 ± 0.16 5.3 ± 0.18 5.72 ± 0.29 0.13

Fibronectin 0.61 ± 0.09 0.57 ± 0.06 0.47 ± 0.07 0.68 ± 0.12 0.0608

Adiponectin 0.05 ± 0.01 0.05 ± 0.01 0.06 ± 0.01 0.07 ± 0.01 0.0248

Complement C3 21.2 ± 0.6 19.46 ± 0.43 18.81 ± 0.5 19.69 ± 0.81 0.0348

Transthyretin 6.1 ± 0.16 6.27 ± 0.12 6.36 ± 0.14 6.36 ± 0.22 0.6208

Clusterin 1.49 ± 0.03 1.46 ± 0.03 1.44 ± 0.03 1.45 ± 0.05 0.6359

Plasminogen 1.22 ± 0.03 1.15 ± 0.02 1.12 ± 0.02 1.12 ± 0.04 0.0767

Albumin 1004.59 ± 19.57 1006.76 ± 14.18 1006.72 ± 16.52 1009.5 ± 26.43 0.9994



Vitronectin 3.64 ± 0.08 3.43 ± 0.06 3.27 ± 0.07 3.36 ± 0.11 0.0509


Prothrombin 0.59 ± 0.01 0.56 ± 0.01 0.55 ± 0.01 0.57 ± 0.02 0.2624

α1B-Glycoprotein 1.59 ± 0.06 1.51 ± 0.04 1.48 ± 0.05 1.65 ± 0.08 0.1647

Histidine-rich

glycoprotein 1.48 ± 0.05 1.4 ± 0.04 1.38 ± 0.05 1.38 ± 0.07 0.833

179



Serum amyloid p-

component 0.53 ± 0.02 0.5 ± 0.01 0.48 ± 0.02 0.49 ± 0.03 0.2087






L-Selectin 0.07 ± 0 0.07 ± 0 0.07 ± 0 0.07 ± 0 0.1191



Complement C9 2.68 ± 0.11 2.47 ± 0.08 2.37 ± 0.09 2.78 ± 0.15 0.0626

Complement C1

inactivator 5.2 ± 0.13 5.05 ± 0.09 5.03 ± 0.11 5.04 ± 0.17 0.5623








1 Values shown are crude means ± standard errors. Proteins are listed alphabetically and in order

of statistical significance among women HC users. P-values from ANCOVA with loge or square-

root transformed plasma protein concentrations where necessary, adjusted for age, ethnicity,

waist circumference, physical activity, and season. Different superscript letters indicate

significant differences between groups (p<0.05). The Tukey-Kramer procedure was used to

adjust for multiple comparisons between groups within each ANCOVA.

§: These p-values meet the Bonferroni level of significance (p<0.0003; α=0.05, 54 independent

tests x 3 test groups).

Documents

25-hydroxyvitamin D and Biomarkers of … · ii 25-hydroxyvitamin D and Biomarkers of Cardiometabolic Disease Bibiana García Bailo Doctor of Philosophy Department of Nutritional