EPIGENETICS, A SIGNAL SENSOR OF DIETARY COMPONENTS, IS INVOLVED IN THE HEPATIC GENE REGULATION

BY

DAN ZHOU

DISSERTATION

Submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Food Science and Human Nutrition

with a concentration in Human Nutrition in the Graduate College of the

University of Illinois at Urbana-Champaign, 2014

Urbana, Illinois

Doctoral Committee: Professor William Helferich, Chair Associate Professor Yuan-Xiang Pan, Director of Research Professor Nicki Engeseth Associate Professor Hong Chen

ii

Abstract

Epigenetic markers are associated with a broad range of disease symptoms, including

cancer, asthma, metabolic disorders, and various reproductive conditions. Changes in

epigenetics status could be induced by environmental exposures such as malnutrition,

stress, smoking, disease and exposure to air pollutants and organic chemicals. Dietary

factors were extensively acknowledged to be one of the key environmental essences

affecting chromatin structure. It can diversify the patterns of DNA methylation and

histone modifications. The changes may be pathogenic and result in severe

physiological consequences, particularly if the exposure occurs during critical window of

development. This is potentially attributable to the property of epigenetic activities,

which occurs not only in somatic and mitotic cells, but also inherited meiotically,

meaning that the changes are carried over generations [1, 2]. Therefore, maternal

dietary factors may profoundly characterize the phenotype of subsequent generations

and consequently contribute to the early onset of chronic disease and metabolic

disorder [3].

Sustained high fat feeding leads to obesity and metabolic syndrome, accompanied by

low-grade inflammation, insulin resistance, and dyslipidemia. In general, the prevalence

of an inflammatory response is highly associated with obesity incidence, cardiovascular

diseases, non-alcoholic fatty liver disease (NAFLD) and brain damage. In liver, changes

caused by low-grade inflammation may later result in progressive disease and fibrosis.

Early exposure to a fat-enriched diet programs the developmental profile, thus is

associated with disease susceptibility in subsequent generations. Chronic low-grade

inflammation, resulting from high fat diet, is activated in the fetal environment and in

many organs of offspring, including placenta, adipose, liver, vascular system and brain.

It forms the biological basis for many patho-physiological changes and constitutes risk

factors for disease development.

Chronic inflammatory disorders were recently identified as one of the major targets of

dietary-induced epigenetic regulation [4]. Although some transgenerational effects of

iii

high dietary fat intake have been shown in a couple of studies linked to certain types of

epigenetic modulations [5-8]; these modifications and their associated physiological

consequences resulting from high fat induced-chronic inflammation remain elusive.

Should the MHF-resulted epigenetic code persist throughout generations,

characterization in such situations may favor the prediction, early prevention, and

treatment of non-communicable disease in next few generations [9].

Many studies using high fat models have consistently demonstrated the incidence of

such inflammatory reactions. However, the potential contributions of epigenetic

modifications toward the regulation of inflammatory genes and subsequent physiological

outcomes have not been fully revealed in the high fat feeding model. Cyclooxygenase-2

(COX-2) produces prostaglandins that participate in multiple physiological and

pathological processes, including the activation of inflammatory responses. In addition

to many of the transcription factors known for years, it was recently suggested that

COX-2 expression is also subjected to epigenetic modifications, either through histone

remodeling and/or DNA methylation [10]. Therefore, our study aims to investigate the

epigenetic mechanisms by which consumption of a high fat diet at different life stages

influences the inflammatory marker COX-2.

Male Sprague-Dawley rats received a high fat diet at different life stages, including

maternal (HF/C), post-weaning (C/HF), and lifelong (HF/HF). Liver was collected for

analysis at 12 weeks of age. Results showed that the high fat diet induced the

expression of COX-2 mRNA in all three high fat groups. RNA abundance of COX-2 in

HF/HF was significantly higher than that of HF/C and C/HF. Meanwhile, fatty acid

composition showed that the proportion of Linoleic acid and Arachidonic acid, as well as

Δ6Desaturase, were significantly increased in high fat groups, potentially motivating the

biochemical flow as well as providing reaction substrate for COX-2 catalysis. Genome-

wide methylated DNA immunoprecipitation (MeDIP) showed that DNA hypomethylation

occurred in an upstream region of the distal promoter and two coding regions of the

COX-2 gene in all three high fat groups. Site-specific hypomethylation of CpG at 5’ UTR

(untranslated region) of COX-2 was confirmed with bisulfite sequencing. Using in vitro

iv

cloning and Luciferase Reporter assay, this region was identified as a novel enhancer

that produces durable transcriptional activity. In conclusion, high fat intake during

different life stages resulted in a varied induction level of COX-2 gene expression.

Altered DNA methylation at specific gene region, including the 5’UTR enhancer

sequence, may closely associate with COX-2 gene activation.

This study presented a gene-wide illustration of the diet-epigenome interaction. It also

provided evidence of high fat diet-induced region-specific hypomethylation in the liver of

the offspring. Since limited research has been conducted to reveal the epigenetic

regulation of inflammatory markers by high fat diet, this study will not only outline the

dietary outcomes on the epigenetic profile of a specific inflammatory gene, but also

provide insight for future mechanistic investigation and clinical appliance in the area of

epigenetics.

v

Acknowledgements

I would love to express my gratitude and appreciation to Dr. Yuan-Xiang Pan and Dr.

Hong Chen for their selfless support and continued encouragement. They have offered

me the best research environment and resources in the past six years. They also have

given their full efforts to improve my thinking process and to establish myself as a

scientist, a researcher. I am grateful for the experiences, insight and knowledge they

have shared with me. Under their guidance, I slowly learned how to become a well-

developed and qualified Ph.D. student. I would not have accomplished what I have

without them. On my career path, their contribution is immeasurable and imperishable.

I offer my sincere appreciation to my committee members, Dr. William Helferich, Dr.

Nicki Engeseth and Dr. Michael Miller. They have provided continuous help and advice

throughout my Ph.D. study. I am impressed by their knowledge in both basic science

and practical applications. My project could not have been completed without their

guidance and valuable suggestions. And I cannot express enough thanks to all the

members of Pan and Chen groups and my dearest friends in UIUC for sharing the joy

and pain in both research and campus life. Knowing these people and being part of this

lab is one of the best experiences in my life. I would also like to thank the funding

supports from Universiy Library and Research Board.

Finally, I want to thank my beloved parents and all my family members for being

supportive and caring about me all the time. Their encouragement helped me survived

and become stronger when the times got rough.

A heartfelt thanks to all of them for nurturing me, educating me, and eventually helping

me become an independent, honest and responsible person.

vi

TABLE OF CONTENTS

CHAPTER 1. GENERAL INTRODUCTION .................................................................... 1

CHAPTER 2. DIET, EPIGENETICS AND INFLAMMATION ......................................... 3

CHAPTER 3. HIGH FAT DIET INDUCES COX-2 GENE EXPRESSION ..................... 25

CHAPTER 4. HIGH FAT DIET INDUCES EPIGENETIC ALTERATIONS OF COX-2

GENE ........................................................................................................................... 35

CHAPTER 5. AN ENHANCER WAS CHARACTERIZED AT COX-2 UPSTREAM

REGION ........................................................................................................................ 44

CHAPTER 6. TABLES AND FIGURES ........................................................................ 56

CHAPTER 7. REFERENCE .......................................................................................... 88

1

CHAPTER 1. GENERAL INTRODUCTION

Environment is a major factor in disease etiology [11, 12]. A large number of

environmental compounds and toxicants have been shown to promote disease, and are

independent of DNA sequence alterations [13]. Nutrition is a critical environmental

factor that is indispensable for an individual’s well-being. Nutrients provide elements

that are essential for life while sustaining basic and advanced biological functions.

Macronutrients feed human energy needs and also provide elemental building blocks,

while micronutrients maintain proper bodily functions on the cellular level. Under-

consumption and overconsumption of certain nutrients is related to an increased health

risk in both animals and humans, including a higher prevalence of obesity, cancer,

diabetes, and many other non-communicable diseases. The cumulative knowledge of

nutrition and other relevant research points to the importance of being aware of one’s

daily dietary intake.

Non-communicable diseases are initiated through altered gene expression, and are

followed by the aberrant regulation of cellular signaling pathways and physiological

behaviors. The traditional understanding of DNA is that it is expressed by reading its

genetic sequence, which is set in stone during our and our children’s lives. Although the

genetic code is important, it is now widely accepted that DNA is not the dictator of gene

expression. Environmental conditions, particularly nutrition, play a role in how DNA

behaves in our children and ourselves. This is part of the relatively new scientific field of

epigenetics. The so-called epigenome contains huge amounts of information, and

engages in activities on top of the genome by commanding genetic activities.

Epigenetic studies offer a new way of investigating the biological roots of common

diseases. The ability to assess the basis of epigenetic mechanisms has improved

dramatically in the last few years. It is apparent that epigenetics may hold the key to

understanding diseases that scientists have thus far struggled to comprehend, including

metabolic syndrome, cancer, neurodegenerative disorders, asthma, and allergies. Age-

related illnesses, such as Alzheimer’s, are a typical example of how these genes can be

2

turned off and on in a manner that leads to the progression of diseases. Another

exciting breakthrough involves understanding cancer development. Explaining cancer

by relying merely on gene mutations is rare and difficult. Researchers thus hope to

develop drugs that will drive epigenetic changes that will help protect against, or stop,

diseases. However, using an epigenetic drug to switch on or off certain genes may also

alter other genes in the process, which makes the development of epigenetic drugs a

challenging endeavor.

Epigenetic information plays an important role in modifying gene expression and can

potentially affect long-term health. Programming the epigenetic profile, occurs mainly

during the early stages of development, and is of critical importance to researchers.

Obesity, diabetes, cardiovascular disease and hypertension can originate during the

early stages of life and development, and are associated with epigenetic programming.

Therefore, perinatal care is imperative for keeping epigenetic changes in check. Doing

so using a combination of cutting-edge epigenetic knowledge and modern technology

can be used to leverage contributions to public health, life quality, and to reduce the

economic burden of disease treatment.

Based on this information and the current knowledge gap, the overall goal of the

present research was to investigate the effects of early life nutrition on one’s wellbeing

and disease risk in adult life using epigenetic tools. High fat diets are the predominant

dietary pattern in the contemporary western world, and have triggered numerous health

problems, which have drawn the attention of researchers. Therefore, this study intends

to address the epigenetic regulation of inflammatory reactions, which are a major

consequence, and the essential basis of, relevant pathogenic physiologies in the high

fat diet scenario.

3

CHAPTER 2. DIET, EPIGENETICS AND INFLAMMATION

Malnutrition during Early Development

Thrifty Phenotype Hypothesis (Developmental Origins of Adult Disease)

The Developmental Origins of Health and Disease (DOHaD) hypothesis addresses the

“adverse influences during early development, particularly intrauterine life that can result

in permanent changes in physiology and metabolism, which increase the risk of disease

in adulthood” [14]. This was initially proposed in 1992 [15] based on the speculation by

Barker, et al., that low birth weight infants who survived infancy and childhood could

face an increased higher risk of coronary heart disease later in adult life [16]. This

hypothesis has been supported and validated by numerous subsequent studies [17]

[18-22]. One piece of evidence involves the Dutch famine events during 1944 and 1945.

During this year, the food supply to a densely populated area was cut off, and affected

some 4.5 million people. This was a unique event that involved a brief period of intense

nutritional deprivation in a previously well-nourished population whose nutrition source

was promptly restored after the famine [23]. Children exposed to in utero famine were

subsequently well-nourished later in childhood. Several years later, findings from the

Dutch famine cohort studies provided key proofs that in utero undernutrition lead to

impaired glucose tolerance and type 2 diabetes, and an increased risk of many other

diseases such as heart disease, obesity, and cancer [24-26]. In addition, they showed

that dietary intake during pregnancy programs metabolic profiles in the offspring, which

may account for the aforementioned disease development.

The causality between early life nutrition and well-being in adulthood has become

increasingly clear in recent years. Altered fetal nutrition has been shown to produce

many health problems. Using a variety of animal models for maternal malnutrition, such

as protein restriction and unbalanced macronutrients, it has been demonstrated that

early life nutritional deprivation reduced placental weight and birth weight [27-30],

increase blood pressure [31-34], and impair glucose tolerance in the offspring [35-37].

In contrast, higher intake of micronutrients during conception was positively associated

4

with birth size and glucose tolerance [38, 39], and negatively associated with blood

pressure in the offspring [40, 41].

To date, findings from diverse studies using maternal models have led to a better

understanding of pathogenic mechanisms and promoted the development of preventive

approaches, as well as potential therapeutic strategies. The most widely accepted

theory that underlies the developmental origins of disease is known as maternal

programming. This is “a process whereby a stimulus or insult during a sensitive or

critical period has irreversible long-term effects on development” [14]. There are quite a

few underlying mechanisms responsible for the pathophysiology of fetal programming.

Fetal nutrition is evidently an important key regulators during programming because

maternal malnutrition usually fails to support fundamental tissue/organ growth and

development [41]. One consequence is that essential functions associated with

gluconeogenesis, lipid metabolism, insulin and cortisol secretion, endothelial function,

and myogenesis are underperformed, which in turn elevates the risks of diabetes,

hypertension and hyperlipidemia. [14, 41].

Maternal High Fat (MHF) Model

Compromised nutritional environment during early development produces both

immediate health problems in newborns, and leads to continuing consequences during

adulthood. Significant outcomes were revealed in earlier studies that focused on

undernourished pregnancies [42-44]. These changes initially caused tissue structural

modifications, which later led to impaired organ functions and biological systems [45].

In the last decade, a dietary intake greatly shifted towards the consumption of highly

accessible fat-enriched foods, combined with poor eating habits. Such dietary patterns

have contributed to the increased prevalence of chronic diseases. Many studies have

reported higher incidences of obesity, hyperglycemia, and systemic insulin resistance,

as well as other metabolic disturbances as a direct result of the prolonged consumption

of high fat foods [46-48]. More importantly, this both affects individuals who consume

high fat diets, and also predisposes their descendants to similar illnesses. Mounting

5

evidence revealed in animal studies has demonstrated that perinatal exposure to high

fat diets profoundly alters the intrauterine environment, and leads to permanent

phenotypic alterations with adverse outcomes that persist throughout the adulthood of

the progeny. Multiple aspects of physiological outcomes were manifested, including

growth problems, learning disorders, and susceptibility to chronic diseases [49] [50, 51].

Interestingly, postnatal high fat and carbohydrate challenge [52-54], or upon the addition

of a drug treatment [55], can aggravate these symptoms. This phenomenon is known as

the “double hit” hypothesis.

Inflammation: a Patho-Physiological Basis for Compromised Health

Inflammation is a part of the natural defense system that the body maintains against

injury and disease. Inflammation is a response to environmental stimuli that instantly

triggers a series of physiological reactions and non-specific immune responses such as

increased blood flow, cellular metabolism, vasodilatation and fluid influx [56]. Moreover,

chronic low-grade inflammation is often linked to disease and long-term adverse

consequences, and is inevitably associated with the pathogenesis of NAFLD/NASH [57],

cardiovascular diseases [58], renal failure [59], nervous system disorders [60], aging

[61], diabetes [62] and several types of cancer [63]. This type of inflammatory response

is commonly observed in obese individuals and in populations with sustained high fat

intake. It is now know that over consumption of high fat foods before or during

pregnancy also affects the inflammatory response of the next generation, in a tissue-

specific and in a life stage-dependent manner.

The overconsumption of fat before or during pregnancy results in an unfavorable

intrauterine milieu. This means that the concentration of energy, hormones, and blood

supply are suboptimal, and this can lead to abnormal fetal growth and development [50,

64-67]. Multiple physiological outcomes related to inflammation and early life exposure

to high fat diet are described below.

Adiposity

6

Earlier studies have suggested the central and enabling role of inflammation in the

process of cell proliferation, also that it promotes adipogenesis, and simultaneously

represses myogenesis [68, 69]. Increasing the adiposity capacity allows inflammatory

signals to generate additional storage space, which constitutes a risk factor for obesity.

Pro-inflammatory factors, including IL-6, MCP-1 and TNF-α, are closely associated with

body fat deposits [70-73]. Certain cytokines, such as IL-6, are known to have a

characteristic tissue-specific function in the process of energy mobilization. In skeletal

muscle, exogenous IL-6 treatment increased fatty acid oxidation [74, 75] and glucose

uptake [75]. By contrast, adipose-tissue-derived IL-6 and TNF-α promoted the

development of insulin resistance in both adipose tissue [76], and skeletal muscle [74,

77]. Consequently, the above information shows that that maintaining careful control of

inflammatory response in both circulatory and peripheral tissues is vital for maintaining

normal physiological functions.

Some studies show that MHF diets result in increased body fat composition [66, 78],

and elevated growth curve for the offspring [50, 79, 80]. MHF exposure increases

adiposity [81, 82], promotes adipocyte hypertrophy [66], and is accompanied by a

decrease of lean body mass [83]. The unpublished data suggests that consuming a high

fat diet throughout pregnancy and lactation induces mRNA expression of IL-6 and IL-1β

in adipose tissue of the offspring. The programming of these pro-inflammatory cytokines

in the adipocytes can contribute to the growing adiposity capacity and excessive body

weight gain. Additionally, other studies have pointed out that undesired, but normal

physiological changes due to either high fat [64, 79, 84], or high sucrose challenge [50],

can be enhanced when the offspring are receiving MHF. The role of systemic

inflammation in adiposity might account for such phenomena. More importantly, MHF

can induce inflammatory responses in tissues including, but not limited to, the adipose

tissue. This will be discussed below.

Cardiovascular Disease

Biomarkers of Low-grade systemic inflammation, specifically CRP, IL-6 and TNF-α, are

intimately associated with endothelial dysfunction, carotid artery remodeling, and

7

vascular stiffness [85, 86]. Furthermore, endothelial dysfunction and insulin resistance

are derived from elevated inflammation [87], and may contribute to the overall

progression of hypertension [88, 89], which subsequently leads to renal malfunction [90].

Chronic inflammation, in the context of atherosclerosis, also contributes to the lesion

initiation, progression and ultimately clinical complications [91].

Several animal models that used a fat-rich diet throughout pregnancy and suckling have

suggested that endothelial and arterial vascular function were compromised in the

offspring to the extent that blood pressure increased up to a level that impaired

cardiovascular and renal function [66, 92-97]. Recent evidence supports the hypothesis

that MHF produces hypertension. However, it appears to develop in an age-dependent

manner [98]. Chronic inflammatory reactions are potentially responsible for this effect,

given that increased levels of IL-1β and TNF-α in both the circulatory system and

arterial walls were detected in the offspring of high fat-fed dams [99, 100]. Nevertheless,

possibilities that include other contributing factors cannot be excluded. It thus appears

to be the case that MHF-induced systemic inflammation can have a cumulative and long

lasting effect, if one consider the time required to eventually reach the clinical

manifestation.

Brain Damage

In addition to the relevant metabolic disorders, excessive fat supplementation during the

early stages of development also impacts cognitive function in the offspring. Studies

have shown that MHF can cause hypertrophy by altering glial and microglial reactivity

[79]. It is known that activation of microglia can provoke the secretion of cytokines IL-1β,

IL-6 and TNF-α [101, 102]. This ultimately results in hypoxic brain injury and acute liver

failure. In the MHF model, the progeny presented microglial activation and a greater

reactiveness to LPS, as well as induction of IL-1β [67], and IL-6 in the brain [79, 103].

However, the origin of brain cytokines in MHF offspring is still being debated. Some

studies suggest that a certain level of cytokines pass through the maternal-fetal

placenta barrier [104], while others indicate the local production in the fetal brain [105].

8

Taken altogether, these studies propose a new perspective for the underlying

mechanisms that regulate brain inflammation.

It is not surprising to see heightened inflammatory reactions in the brain, but yet the

subsequent outcomes deserve a closer attention, as inflammation now is believed to

significantly partake in the progression of many brain diseases [106, 107]. MHF-induced

cytokine production affects many compartments of the brain. It produces detrimental

effects over time and has become a leading cause of undesired functional decline. IL-6

is typically released from microglia and astrocytes. It is known to accelerate the

inflammatory progression by actively promoting the production of other relevant

cytokines, namely IL-1β and TNF-α [108], both of which have been shown to cause

severe injuries in fetal and postnatal brain [109]. The changes brought about by

activated inflammation are more likely to be mechanistically associated with long-term

adverse outcomes. Meanwhile, IL-6 itself is closely associated with cognitive function,

hypertension, and neurogenesis. The excessive production of IL-6 in the brain was

shown to impair spatial learning [110], increase hypothalamic-pituitary stimulation [111],

and reduce neurogenesis in the hippocampus [112]. Consequently, one of the

mechanisms by which scientists would observe undermined brain functionality in MHF

offspring involves the up-regulation of cytokines such as IL-1β and IL-6 [103]. These

outcomes have not yet been fully described in MHF model. Several recent studies

showed multiple lines of evidences which point out that the offspring exposed to the

maternal and postnatal high fat diet exhibited difficulties in retaining the previous

training in a spatial learning tests [79]. Increased susceptibility to deregulated brain

signaling, including the serotonergic system, can further cause perturbations on

behavioral patterns [113, 114]. Additional studies should be conducted to define the

specific events, but it is possible that environmental challenges, disease and aging

could induce these subjects to underperform in comparison with peers.

Hepatic Energy Mobilization

MHF has a steatogenic and fibrogenic effect in the liver, but does not necessarily cause

hepatic lipid accumulation [115, 116]. However, intrauterine exposure to high fat diets is

9

believed to program the hepatic metabolic patterns, and contributes to NAFLD

progression in the offspring. Severe fat accumulation in liver tissue occurs when a post-

weaning high fat diet is administered, compared with postnatal high fat feeding alone

[84].

Several animal studies have reported that perinatal exposure to a high fat diet triggers

the induction of inflammatory markers in the liver. For instance, in a lactating high fat

mouse model the inflammatory markers IL-6 and TNF-α were significantly induced [115].

Another study using a mouse model observed the stimulation of CRP, Mmd2, TNFsf1

and IL-12b [84]. Some unpublished data concerning rats shows a significant induction of

several other inflammatory markers, including IL-1β, IL-6 and TNF-α, in the liver of MHF

offspring. Although, the exact reason for the species-specificity or feeding-timing-

dependent response remains unclear, a common hepatic inflammatory effect in

response to MHF was observed.

Inflammation triggered by MHF disturbs lipid and glucose metabolism in the offspring,

and is likely to result in fatty liver disease. For instance, MHF-induced increase of

circulating TNF-α constitutes a risk factors for fatty liver development [99]. Adipocyte

overexpression of TNF-α has been shown to drive local lipolysis and produce free fatty

acids [117], which consequently enhances fat mobilization in other tissues, including the

liver [118]. This phenomenon was indeed observed in the MHF model, and increased

plasma glycerol in the fetus, which suggests up-regulated lipolysis and thus increases in

hepatic fat deposits [65]. Meanwhile, another study observed increased hepatic lipid

synthesis and fat accumulation as results of intraperitoneal injection of TNF-α or LPS

that induced TNF-α [117]. Thus, it is possible that circulating TNF-α, regardless of the

stimulus, is involved to some extent in the master signaling mechanism for energy

mobilization throughout the body. On the other hand, local activation of inflammatory

cytokines in liver tissue, including IL-1β, IL-6 and TNF-α, often worsen insulin signaling,

and also stimulate of gluconeogenesis and lipid synthesis [119]. These phenomena

have been found recently in this particular MHF model. Reduced hepatic insulin

sensitivity in MHF offspring was indicated by the reduced expression of IRS in the liver

10

[115]. Gluconeogenesis is a pathway that helps to maintain plasma glucose, particularly

during periods of starvation. This pathway was also programmed by MHF, and exhibits

a significant up-regulation of hepatic Pck1 and G6Pase in the liver [65, 80]. Previous

research has shown that overexpression of Pck1 is closely associated with

hyperglycemia and the progression of type 2 diabetes [120]. As a consequence,

increased liver weight and glycogen content were reported in a follow-up study in the

offspring of this model [78]. Furthermore, hepatic lipid metabolism is affected by an

MHF diet, as evidenced by the deregulation of genes involved in fatty acid transport and

synthesis, as well as TAGS synthesis [84]. Given the fact that these physiological

outcomes can be attributed to multiple changes throughout the entire system, additional

investigations are necessary to determine whether repression of inflammation in this

model is an effective approach for attenuating these symptoms.

More interestingly, fatty acid composition in neonatal liver is modified by MHF, and may

partially account for the activated inflammatory response. This was exhibited together

with a decrease on DHA [83] and EPA, along with an increased proportion of oleate,

linoleate and arachidonate in the overall pool of both TG and FFA [121]. DHA provides

protection for the cardiovascular system against coronary disease [122], and down-

regulates many inflammatory mediators [123]. Arachidonic acid is a second messenger

for inflammatory signaling, and actively participates in the inflammatory response and its

associated physiological changes [123]. Hence, it appears that the elevated n-6:n-3

ratio in the liver can create a susceptible and enriched environment for inducing the

chain of inflammatory reactions.

Placental Dysfunction

Placenta, the primary organ of the maternal-fetal interface, and transports nutrients from

the mother to the fetus, and is the focus of attention during fetal development. Impaired

functionality of the placenta is one of the major causes of fetal disease. The

maintenance of placental normal function during fetal growth eliminates the

predisposition to preeclampsia [124], preterm labor [125], and stillbirth [126]. Moreover,

perinatal high fat intake impairs the fetal environment, triggering placental inflammation

11

and its associated complications [127]. More particularly, mRNA expression of pro-

inflammatory cytokines IL-1β, MCP-1 and inflammatory mediator TLR4, are induced in

the placenta of high fat-fed mothers [128]. Excessive expression of IL-1β and TNF-α

has been shown to be associated with preeclampsia [129], and TLR4 is associated with

preterm delivery [130]. Increased placental inflammation is evident, such that

compromised placental functions, such as decreased blood flow on the fetal side [128,

131], may result in stillbirth. Unfortunately, a verifiable molecular mechanism for a

causal relationship between placental inflammation and disturbed utero-placental

hemodynamics remains elusive. Current data have suggested that these outcomes, if

any, are independent of NHF feeding, but it has been confirmed that a high fat diet will

certainly aggravate the chemically induced preeclampsia [132].

The fetal consequences of increases in maternal circulating cytokines and placental

inflammation remain elusive. It is easy to observe the increased inflammation in fetal

circulatory system, liver and brain that follow the consumption of a NHF diet [65, 133].

However, it is difficult to pinpoint the origin of these wandering cytokines, given that

bidirectional transfer of pro-inflammatory cytokines is barely detectable in a normal term

placenta [104, 134]. Indeed, there is evidence showing that induced cytokines in the

placenta do not necessarily cross the placental barrier, which has an impact on fetal

inflammatory status [105], despite the altered permeability due to placental inflammation

[135]. Therefore, investigations of the transgenerational impact of up-regulated maternal

cytokines and the fetal origins of inflammation in this model are expected to have

clinical value and allow for the early prevention and development of therapeutic

treatment.

State of the Art Research in Epigenetics

Basis of Epigenetics

Epigenetics is defined as the alterations in gene expression in the absence of

underlying changes in genetic information [136]. It must be heritable from cell to cell,

hence through cell lineage development or even transgenerationally from parent to

12

offspring to grand-offspring [1]. Epigenetic mechanisms facilitate the understanding of

how environmental factors can alter the phenotypic phenomenon without changing DNA

sequence, and hence the origin of some diseases that are not readily explicable in

terms of conventional genetic defects.

The study of epigenetics began 70 years ago, and has continued to evolve since then. It

initially explained the basic process that multicellular organisms reserved for cell

differentiation, which involves a non-genetic factor that guides cell division and fate

regardless of the identical genetic code shared throughout one’s system. Classic

biological actions defined within the realm of epigenetics includes enzymatic

methylation of cytosine bases, also known as DNA methylation, post-translational

modifications of certain amino acid residues on histone tails or histone modifications,

nucleosome positioning, non-coding RNAs, and transcription factor regulatory networks

[137].

DNA Methylation

DNA is usually methylated at the carbon-5 position of cytosine that presents in 5’-CpG-3’

dinucleotides [138]. CpG islands are the regions of more than 500 base pairs in size

and with GC content above 55%. It is highly conserved throughout evolution, because

they are normally kept free of methylation [139]. These stretches of DNA are usually

located at promoter regions and cause stable heritable transcriptional silencing. DNA

methylation is primarily a stable repressive marker, yet its regulation is more dynamic

than what was once believed, and it can be actively removed at specific loci as well as

genome-wide during development [140].

DNA methylation usually, but not always inhibits gene transcription, usually but not

always depending on the position of the methylation change relative to the position of

the transcription start site [141]. Four DNA methyltransferase (DNMTs), namely DNMT1,

DNMT2, DNMT3A and 3B have been found to mediate DNA methylation but perform

different functions [142]. Particularly, DNMT1 maintains the basic needs of DNA

methylation, copying the pattern from parental strand to the newly replicated strand

13

during cell division, whereas DNMT3A and 3B promote the de novo DNA methylation

[142].

Histone Modifications

Unlike DNA methylation, histone tails can be modified in multiple ways as is the case

with acetylation, methylation and phosphorylation [143]. Acetylation of histones at lysine

residues is catalyzed by histone acetyltransferases (HATs), and is associated with the

activation of gene transcription. Deacetylation of histones is associated with the

silencing of transcription, and is catalyzed by histone deacetylases (HDACs) [144].

Overall, the acetylation of histones marks active transcriptionally competent regions,

while hypoacetylated histones are found in transcriptionally inactive euchromatic or

heterochromatic regions. Histone methylation regulates gene transcription depending

on the site of modification. Methylation of lysine 9 on the N terminus of histone H3

(H3K9Me) is typically a hallmark of silent DNA and is globally distributed throughout

heterochromatic regions such as centromeres and telomeres [145]. However,

methylation of lysine 4 of histone 3 (H3K4Me) activates gene transcription and is

primarily found in promoter regions [145]. Lysine can be modified by mono-, di- and

trimethylation. The enormous variation is thus likely to result in a diversity of possible

combinations of different modifications.

Epigenetic Interactions

There is emerging evidence that there are connections between DNA methylation and

histone modifications [146]. Eukaryotic DNA is greatly associated with histone proteins,

and forms a highly ordered and condensed DNA-protein complex termed chromatin

[144]. DNA methylation influences both histone modifications and overall chromatin

structure by guiding the histone deacetylation of certain residues [147], DNA

methylation is intimately connected with histone modifications [148]. DNMTs as well as

three protein complexes that contain methyl-CpG-binding domain [149], were found to

bind HDACs, which represses gene transcription [150]. Interestingly, it was once

believed that histone modification was secondary to DNA methylation. Recent studies

have revealed that histone modifications can take the initiative in chromatin

14

conformation by assisting the process of DNA methylation [146, 151, 152]. There

appears to be a close relationship between DNA methylation and histone modulations,

but many questions concerning the details of such interactions remain to be answered.

Other than these modifications, RNA in various forms, including antisense transcripts,

noncoding RNAs, or RNA interference (RNAi), is also capable of triggering the

transcriptional silencing of genes by aiding histone modification and DNA methylation

[148].

Developmental Programming of Epigenome

Mammalian reproduction involves a single-cell zygote giving rise to a complex living

organism with a large variety of cell types. These highly diversified cellular phenotypes

are produced from the identical genomic sequence by mediating the subset of genes

expressed in each cell type. Thus, cellular differentiation, and tissue and organ

formation, rely largely on the process of extensive erasure and the precise

establishment of lineage-specific epigenetic markers, which are termed epigenetic

reprogramming. The Epigenome is re-established at fertilization, beginning with the

initial wide-ranging loss of epigenetic marks in the zygote and early embryo, followed by

the resetting of DNA methyl marks during many stages of cell division [153].

Developmental plasticity makes the reprogramming process highly susceptible to fetal

and early life environmental exposure due to developmental plasticity. Epigenetic

modifications, such as DNA methylation, are substantially sensitive to perturbation at

the earliest stages of development, and most likely resulting in lifelong and possibly

transgenerational effects. Chromatin states that arise during this period can further

affect the propensity to subsequent epigenetic change. This is evident in the

predisposition of polycomb-repressive complex occupied genes in stem cells to acquire

DNA methylation abnormalities in aging and cancer [154-157].

The “developmental origins of adult disease” hypothesis posits gene-environment

interactions that result in long-lasting effects, and suggests epigenetic inheritance as a

15

prime mechanism. During the early stages of development, a folate (methyl donor)

deficient diet reduces the methylation of offspring on a whole-genome scale [158].

Indeed, rodents are often used to study the subject of developmental programming.

This model has gained recognition, mostly by acknowledging that in utero malnutrition

not only affects gene expression through epigenetic profile editing. This is the case

because the progeny responds to environmental chemicals as well as ultimate adult

phenotype does [159]. At this point, limited information is available from human studies,

although DNA methylation appears to be the culprit for imprinting gene regulation when

early life nutrition was manipulated [160].

Transgenerational Inheritance

Epigenetic marks in somatic cell lineages can be transmitted to daughter cells over the

lifespan of an individual, while marks established in germ cell lineage can be transmitted

to subsequent generations. Determining the transgenerational effects of epigenetic

mechanism requires at least three generations after the exposure. Mounting evidence

has demonstrated the apparent transmission of environmental insult across generations,

although this remains highly controversial [161].

Environmentally induced germline epigenetic modifications occur during the DNA

demethylation and remethylation period [162]. Certain types of chemical environmental

exposures have been shown to promote the epigenetic transgenerational inheritance of

adult onset disease, and the transgenerational epigenetic changes can be used as

biomarkers of exposure and disease [163]. Several studies of the transgenerational

effects of high dietary fat intake have been shown to be linked to certain types of

epigenetic modulations [5-8], direct association between these modifications and their

physiological consequences that result from high fat-induced chronic inflammation

remain elusive. Given the potential persistence of the trans generational epigenetic

code, characterization in such situations in MHF model may favor the prediction, early

prevention, and treatment of non-communicable diseases in the next several

generations [164].

16

Epigenetics and Disease

Epigenetic vs Genetic

Conventionally, it was believed that the inheritable material carried in the primary

sequence of the DNA double helix was the predominant of the determination of

variations in the susceptibility and severity of disease. The understanding of how

germline genetic variations influence disease development was assisted by

technological advances and was expanded in many genome-wide association studies

[165-167]. Changes in the DNA sequence drastically interrupts gene expression, and

can lead to the genesis and progression of several diseases [166]. However, these

studies also support the idea that common genetic variants alone tend not to identify the

causal loci of complex diseases, and do not dictate individual disease risk [165].

Fundamental modification of the genome, also known as epigenetics, thus becomes a

determinant of disease risk and etiology [168, 169], because it allows an individual

system to modify how the genome is read in response to environment cues [168, 170].

Epigenetic events are inheritable and can be long-lasting, just like genetic information.

Nevertheless, epigenetics differs from genetic changes in a small number of unique

characteristics. First, it does not involve alterations of the DNA sequence and is

reversible in principle [171]; secondly, epigenetic events are subject to reprogramming

by endogenous or exogenous factors throughout the life span in a continuous editing

process [172-176]. These unique features allowed of alternative explanations of disease

variability to thrive, including diseases that cannot be fully explained by genetic

variations and changes.

Environmental Epigenetics and Disease Etiology

Environmental epigenetics refers to the phenomenon by which epigenetics can explain

the variability in the risk and severity of environmental diseases. The external

environment, which includes lifestyle factors that are linked to stress, abuse, addiction,

alcoholism and metabolic changes, are comprised in such a concept [177]. Particular

environmental factors, such as endocrine disruptors, infectious pathogens, outdoor

17

pollutants, indoor allergens, and heavy metals, have been shown to yield epigenetic

changes [177].

Environmental factors can trigger complex diseases, including cancer, cardiovascular

disease, stroke, pulmonary diseases, asthma, obesity, and neurodegenerative disorders

[178-181]. More importantly, changes in the epigenetic status due to early-life

reprogramming or experiences during adult life have shown a strong correlation with the

severity and progression of these diseases [177]. Therefore, the behaviors of the

epigenetic code canbe expected to indicate exposure to environmental toxins, as well

as risk and the progression of diseases. Current investigations are focusing on the

mechanistic regulation of epigenetic events, in order to promote a better understanding

of the etiology of environmental disease and future development of preventive

strategies.

Disruption of either the sequence or composition of the epigenetically regulated genes

that result in their aberrant expression, are considered to be the fundamental

mechanism underlying abnormal differentiation and disease development. Epigenetic

factors have been determined to be involved in the pathogenesis of many types of

cancer and common multifactorial disorders [182]. Aberrant de novo methylation of CpG

islands is a hallmark of human cancers and can be found early during carcinogenesis

[183]. Functional epigenetics changes have been observed in obesity [184], diabetes

[185], cardiovascular disease [186] and the constellation of abnormalities known as

metabolic syndrome [187]. Neurological disorders, cancer, reproductive conditions, and

the pathogenesis of asthma have been suggested by emerging evidences that are

attributed to epigenetic alterations [188].

Epigenetic changes are sensitive readouts of the consequences of acute and chronic

exposures to environmental factors. They greatly rely on dose, duration, and life-stage

at the time of exposure. More particularly, intense effects are usually found during the

susceptible window of fetal development [177], thus exposure to environmental factors

during a certain developmental stage will strongly influence disease etiology. During this

18

time window, differentiation of cells and organ systems are susceptible to programming.

Consequentially, altered programming and gene expression due to environmental

factors can lead to abnormal physiology and disease in adult life [162]. Therefore, it is

reasonable to assume that environmental factors predispose phenotypic changes or

disease progression, and this is not limited only to the individual exposed, but also

includes subsequent generations. Normally, exposure to mild environmental factors

such as nutrients, toxicants, or endocrine disruptors, are insufficient in and of

themselves to produce genetic mutations, but can produce alterations in the epigenome,

which are known as epimutations [162]. These epimutations can potentially reside in the

germline of individuals exposed and become inheritable cross generations, and lead to

the early onset of certain types of diseases.

Epigenetics and Inflammation

Inflammatory signaling has recently been identified as an epigenetic mediator [189]. It

has been suggested that histone acetylation is closely associated with inflammation

[190-192]. The pro-inflammatory signaling networks of IL-1β, LPS, and TNF-α-induced

pathways may impact histone modifications through the activation of HDAC1 by

ubiquitination and proteosomal degradation [193, 194]. On the other hand, a growing

body of evidence has shown the power of epigenetics at regulating inflammatory genes,

the incidence of which may eventually result in multiple physiological consequences

[195]. For example, the addition of HDAC inhibitor can reduce the expression of pro-

inflammatory mediators in glial cells, and has the potential to result in improved neural

cell function and survival [196]. Meanwhile, a histone acetyltransferase (HAT) inhibitor

can reduce the acetylation of NF-κB, which in turn suppresses the expression of

cytokines in microglial cells and prevents neuronal cell death [197]. Furthermore, a few

of chromatin-remodeling proteins including Brg1, Brm [198] and TET1 [199], are thought

to participate in pro-inflammatory cytokine gene regulation of IL-1, IL-6 and MCP-1,

possibly through increased activation of histone modifications. However, this leads to

the progression of NASH [198]. Unfortunately, evidence for epigenetic programming of

the inflammatory response by intrauterine environment is scarce, even so for the MHF

model. Although chromatin remodeling in this model has been confirmed for certain

19

functional genes [80, 200], no indication of inflammatory marker programming has been

revealed in any studies involving MHF. Nevertheless, given the reciprocal interaction

between epigenetics and inflammation, MHF-induced chronic inflammation has the

potential to shape the epigenetic profile of other functional genes, and the expression of

inflammatory genes themselves are susceptible to epigenetic modulation. More

investigations are therefore needed to explore the epigenetic mechanisms and the

development of cutting edge therapeutic approaches.

Application of Epigenetics in Pharmacology

Epigenetic changes are intertwined with a variety of diseases. Manipulation of the

epigenetic status is thus likely to expand the array of its applications [201]. Indeed,

epigenetic therapy is believed to be a new development in pharmacology, by virtue of

attempting to correct epigenetic defects, including changes in DNA methylation and/or

histone modulations. It is a medical strategy with great potential, since epigenetic

defects tend to be more reversible than genetic defects when using pharmacological

interventions [202-204]. So far, two clusters of drugs have been developed; one aims to

inhibit DNA methyltransferases (DNMT), and results in the inhibition of DNA methylation,

which blocks one type of causative mechanisms of cancer where tumor suppressor

genes are hypermethylated. The other major group targets the inhibition of HDACs,

which reduces histone deacetylation which may serve to mediate the anticancer effects

of these drugs. Promising results have been observed in drug trials for cancer treatment.

More importantly, epigenetic drugs may be capable of preventing disease, in addition to

the therapeutic agents [205]. Epigenetic therapy has limitations because both DNMT

and HDAC inhibitors lack specificity, which can accelerate tumor progression in certain

cases [206]. A variety of molecules are involved in epigenetic mechanisms, and other

potential targets are plausible for drug development purpose.

Interaction between Nutrition and Epigenetics

Dietary Effects on Epigenetic Profile

20

Many environmental factors can interfere with the epigenetic status and produce either

transient or permanent alterations of chromatin structure. Environmental factors,

smoking, chemical exposure, etc., together with eating patterns, can profoundly affect

biological processes, including oxidative stress and inflammation, which will directly or

indirectly shape the state of the epigenome. Epigenetic mechanisms have recently been

used more frequently used to describe environmental effects on phenotypes,

particularly the nutritional status.

In the 1940s, nutrient supply was initially recognized as an authoritative element, whose

intensity produces long-lasting physiological phenomenon that modify the epigenetic

code during a critical developmental window [160, 207]. The biological role of nutrition in

the process of phenotypic differentiation has been documented in honeybees [208], and

agouti mice [209], and both of them involve DNA methylation [210]. Folate is a typical

nutrient that is often used when studying methylation regulation. Folate is a major

source of the methyl groups for DNA and histones [211]. Moreover, imbalanced dietary

intake of folate is associated with methylation changes and consequently gene

deregulation, and subsequently increased disease risk [212, 213].

Food intake can also manipulate the epigenetic code by integrating miRNA into the

metabolic activities of consumers, with little nutrient content relevance. It was

determined that a plant-origin microRNAs, MIR168a, binds to the coding region of

mammalian genes and decreased protein production [214]. A lot more investigations

are required to confirm this phenomenon.

Developmental Plasticity and Nutrition

Developmental plasticity should ideally allow the organism to monitor its growth process

in an attempt to promote phenotypic changes that better equip it to fit with the

environment. However, in face of an excessive or deficient nutrient supply during the

early stages of life, the progeny are more likely to be programed with greater limitations,

which undermine their defensive capacity against postnatal environmental challenges.

They may consequently experience abnormalities in terms of body size, growth rate,

21

tissue distribution, metabolic profile, and hormonal imbalance and underperformed

intellectually [215, 216]. It is now generally accepted that MHF is an unfavorable

intrauterine environment that is closely associated with the development of metabolic

syndrome and an increased risk of an obese phenotype developing in the offspring

[217]. However, researchers are only beginning to understand the basis of mechanism

that leads to the initiation of those signaling cascades, which in turn produce structural

and functional cellular changes, physiological damage, and ultimately chronic diseases.

Evidence has shown that maternal environment affects gene transcription, partially

through epigenetic modulation, including DNA methylation, multiple forms of histone

modifications [218] and microRNA [219, 220]. Therefore, the concept of epigenetic

regulation can serve as a critical link between maternal diet and progeny phenotype,

which consequently supports the DOHaD hypothesis.

High Fat Diet and Epigenetics

High fat diets impact greatly the epigenetic structure, such as the intensity of DNA

methylation of genes that were highly involved in pathophysiological process. In

general, metabolic and cell cycle relevant genes in liver tissue are the subject of DNA

methylation regulation. For instance, methylation intensity at CpG sites of Esr1 [221],

stearoyl-CoA desaturase 1 (Scd1) [222], promoter of glucokinase, pyruvate kinase [223]

and MTTP [224], was differentially regulated by high fat-diets. Deregulation of these

genes is primarily associated with aberrant hepatic lipid and glucose metabolism, and

result in the development of various types of non-communicable diseases, including

fatty liver disease and obesity. In addition, altered DNA methylation patterns of cell

cycle controlled gene Cdkn1a [225], and the growth hormone secretagogue receptor

[226], were reported in a MHF model, which indicates the transgenerational epigenetic

effects of high fat diets. DNA methylation induced by a high fat diet elicits broad effects

in other tissues besides the liver. For example, the CpG sites of FASN and Leptin in

adipose tissue [227], as well as for the brain μ-opioid receptor [228], tyrosine

hydroxylase and the dopamine transporter [228], leptin [229], and melanocortin-4

receptor (mcr4)[230], were reported to be differentially methylated by high fat diet. A

significant correlation between DNA methylation and gene expression has been

22

validated in several of these studies, which indicates that consuming a high fat diet

affects gene regulation by altering the DNA methylation status of gene. The effects of

high fat diets on the global DNA methylation on the whole-genome level have been

rarely investigated.

Conclusion

The understanding of the epigenetic-based mechanisms for inflammation activation in

the high fat model is insufficient to explain the constellation of body responses to the

environment. Previous research has shown that in utero exposure to high fat diets

programs the increased susceptibility to chronic diseases during development and

growth. There appears to be a causal relationship between low-grade chronic

inflammation and such adverse outcomes. However, it is difficult to pinpoint a single

causative mechanism, or even draw out a conclusion given the essential role that the

inflammatory response plays in many cellular and physiological processes. Epigenetic

mechanisms are associated with a myriad of developmental diseases. It is widely

recognized as a new therapeutic approach nowadays, and unraveling the epigenetic

regulation of inflammatory markers in a high fat model can assist disease prevention

and treatment in both the current generation and in subsequent generations as well.

Project Overview

Background

Consumption of high fat increases the risk of obesity and metabolic syndrome, both of

which were suggested to be closely associated with a low-grade chronic inflammation. It

has been shown that high fat feeding specifically increased proinflammatory signaling,

including Cyclooxygenase (COX) -2 expression, in a diet-induced obesity mice model

[231]. However, the molecular mechanisms regarding epigenetic modifications on COX-

2 in this model remains to be elucidated. The COX family comprises two isoforms, 1

and 2, sharing significant sequence homology and catalytic activity. COX-1 is

constitutively expressed to produce prostaglandins (PG) for maintaining homeostatic

23

functions, whereas COX-2 requires additional extracellular and/or intracellular stimuli,

including exogenous and endogenous arachidonic acid [232, 233]. The COX enzymes

catalyze the conversion of arachidonic acid to prostaglandins, from which the substrate

for cell specific prostaglandin and thromboxane synthases are produced [234].

Prostaglandins are involved in numerous physiological and pathological processes

including inflammation, and they are rapidly increased by COX-2 in the scenario of

inflammation [235]. Induction of COX-2 is transient, with a rapid return to baseline within

hours following treatment [236]. Therefore, expression of COX-2 is strictly regulated and

it is only markedly induced during inflammation. The nutritional environment alters

epigenetic profiles through DNA methylation and histone code, and thereby affects gene

transcription, intracellular signaling and ultimately the development of disease. Although

it is known that high fat feeding triggers an inflammatory response including COX-2,

there is little knowledge of the epigenetic mechanisms involved. Our study here propose

a connection between high fat diet, COX-2 transcription and epigenetic modifications,

which has never been tested in previous research. By conducting the present research,

it hopes to improve the understanding of epigenetic-related COX-2 gene regulation.

Objective

The objective of this study is to identify the molecular mechanisms involved in high fat –

induced COX-2 gene expression.

Hypothesis

High fat diet induces inflammatory marker COX-2 expression associated with epigenetic

modifications.

Specific Aims

1. Characterize the physiological outcomes and COX-2 response to high fat feeding

at different life stage

2. Determine if COX-2 expression is associated with epigenetic modifications,

including DNA methylation and histone modifications.

24

3. Identify if there is any regulatory element on COX-2 gene that responses to high

fat signal.

Discussion

In Chapter 3, 4 and 5, this study presented several lines of evidence to demonstrate

that high fat feeding (maternal, HF/C; post-weaning, C/HF and life-long, HF/HF) affects

physiological and pathological phenomena in the liver of pups. The resutls revealed that

COX-2 gene expression increased in pups’ liver along with multi-regional DNA

hypomethylation in high fat diet. 5’UTR hypomethylated region was further identified

and characterized as an enhancer element that potentially contributes to COX-2 gene

activation.

Prolonged high fat dietary feeding instigates an inflammatory response, following which

a variety of proteins, including COX-2, can be significantly induced. Understanding how

dietary components regulate gene expression at epigenetic level provides a novel

insight that connects environment and disease predisposition. Following the completion

of the present study, a link between increased hepatic COX-2 seen with high fat diet

feeding with specific epigenetic modification will be elucidated.

25

CHAPTER 3. HIGH FAT DIET INDUCES COX-2 GENE EXPRESSION

Abstract

Sustained high fat feeding leads to obesity and metabolic syndrome, accompanied by

low-grade chronic inflammation, insulin resistance, and dyslipidemia. Changes caused

by low-grade inflammation in the liver can result in the progression of disease and

fibrosis. COX-2 is an enzyme that produces prostaglandins, which participate in multiple

physiological and pathological processes, including the activation of inflammatory

responses. Male Sprague-Dawley rats received a high fat diet during different life

stages, including maternal (HF/C), post-weaning (C/HF), and lifelong (HF/HF). Liver

samples were collected for analysis at 12 weeks of age. Total fat accumulation and liver

damage evaluation were significantly higher in HF/HF than in the rest of the treatment

groups. Analysis by PCR of the expression of COX-2 mRNA showed an induction in all

of the high fat diet groups, especially in the HF/HF group, which doubled the induction

fold of HF/C and C/HF. Meanwhile, analysis of fatty acid composition showed that

proportion of both Arachidonic acid and Δ6-Desaturase, which produces Arachidonic

acid, significantly increased in the high fat groups, which may have motivated the

biochemical flow as well as provided the reaction substrate for COX-2 catalysis. In

conclusion, induction of COX-2 gene expression in response to a high fat diet is

associated with the period of exposure.

Introduction

Modern western diets are rich in fat content. Prolonged consumption of high fat diets is

the causal agent of numerous health problems, including promotion of hepatic

triglyceride accumulation, insulin resistance, oxidative damage, and chronic low-grade

inflammation, which leads to the progression of liver disease and fibrosis [237, 238]. A

murine study showed that consumption of a high fat diet changed liver morphology,

which created a permissive environment for colon cancer metastasis [239]. A high fat

diet can also change cellular membrane protein composition [240], elevate the

26

abundance of pro-inflammatory molecules, and give rise to hepatic insulin resistance

[241].

Induced pro-inflammatory signaling is usually signified by activated expression of

certain genes, such as Interleukin family members and COX-2 [116, 242]. Activated

COX-2, rather than COX-1, is persistently found during the activation of inflammatory

responses [243]. COX-2 is a key regulatory enzyme involved in the production of

prostaglandins, which are important in mediating many biological processes, including

reproduction and immune function. COX-2 actively produces prostaglandins using

exogenous and endogenous arachidonic acid [233]. This occurs during injuries and

inflammation in the metabolic steatohepatitis scenario [244]. In addition, overexpression

of COX-2 is usually associated with inflammatory responses, many types of cancers

[236], and preterm labor [245].

Exaggerated COX-2 expression has significant tissue-specific consequences:

accumulation of COX-2-derived prostaglandins initiates a multitude of physiological

responses, which generally result in fever and pain. However, in the liver, excessive

expression of COX-2 accounts for the pathogenesis of insulin resistance [246, 247],

fatty liver disease [246, 248], and hepatocellular carcinoma [249]. This study is

dedicated to a better comprehension of the pathogenic properties that result from the

exposure to a high fat diet at different stages of life. Examination will include COX-2

gene expression and other relevant physiological features.

Materials and Methods

Experimental Design

Timed-pregnant Sprague Dawley rats (Charles River Laboratories, Wilmington, MA)

were separated into two dietary groups: either control (C, 16% fat) or high fat diet (HF,

45% fat) throughout gestation and/or lactation. Rats were kept individually in standard

polycarbonate cages in a humidity- and temperature-controlled room on a 12-hour light-

dark cycle, with ad libitum access to food and drinking water. Body weight and food

27

intake were recorded every three days during gestation. Body weight of the offspring

and liter size were recorded right after delivery. Twenty-four hours after birth, both

groups of dams were randomly assigned 10 pups (5 male and 5 female coming from the

same gestational diet) until weaning at day 21. After weaning, pups from both groups

were housed individually and exposed to either C or HF diet until they were sacrificed at

12 weeks of age, generating four groups: C/C, HF/C, C/HF and HF/HF. The left lobe of

the liver was snap-frozen in liquid nitrogen and stored at -70C until use. We certify that

all applicable institutional and governmental regulations regarding the ethical use of

animals were followed during this research (University of Illinois Institutional Animal

Care and Use Committee approval no. 09112).

Liver Histology

Frozen liver samples were embedded in Tissue-Tek OCT compound (VWR, Radnor, PA)

and cut to a thickness of 5 µm in a cryostat at -20°C. All sections were fixed in 70%

ethanol and stained with hematoxylin and eosin (H&E) or Oil red O (ORO) solution

(Newcomer Supply). Histopathological examination was conducted by the certified

pathologist. The grading scale used to assess microscopic damage in liver was 0: none,

1: minimal, 2: mild, 3: moderate, 4: marked and 5: severe.

Biochemical Assays

Frozen liver samples were ground in liquid nitrogen and homogenized in 0.3 mL saline

(0.9% w/v NaCl). Homogenized samples were diluted 5 times with saline. Twenty

microliter of the diluted samples were incubated with 20 μL 1% deoxycholate in 37 ºC

for 5 min and 10 μL of the sample were analyzed using the Thermo Infinity Triglycerides

Liquid Stable Reagent (Thermo Fisher Scientific, Rockford, IL) and a standard reference

kit to determine the TAG content (Verichem Laboratories, Providence, RI). The results

of hepatic TAG were normalized to the amount of total protein in the sample. Serum

samples were analyzed for circulating TAG using the same method.

Gas Chromatography Analysis of Fatty Acid Content in the Liver

28

Total lipids were extracted from frozen liver tissue (~100 mg) with the method described

by Bligh and Dyer [250], using chloroform:methanol (1:2 v/v) and 0.15 N acetic acid.

Each extraction solvent contained 0.05% (w/v) BHT to prevent lipid oxidation during

extraction. An aliquot of total lipid was spiked with heptadecanoic acid as internal

standard, dried under nitrogen and derivatized by heating at 95 °C in 3 N methanolic

HCl (Supelco, Bellefonte, PA) for 40 min. After cooling to room temperature, fatty acid

methyl esters were extracted with hexane. Samples were injected onto a 5890 Series II

gas chromatograph (Hewlett-Packard, Palo Alto, CA) equipped with a flame ionization

detector and a 30 m × 0.32 mm Omegawax 320 column with 0.25 μm film thickness

(Supelco, Bellefonte, PA). Chromatography conditions used were 180 °C for 5 min,

followed by a temperature increment of 6 °C per min, and ended with isothermal at

220 °C for 15 min. Peaksimple software (SRI Instruments, Torrance, CA) was used for

data collection and integration. Each fatty acid was normalized to the total fat content

and presented as percentage of total fatty acids.

RNA Isolation and Two-step Real Time Quantitative PCR (qPCR)

Total RNA samples from tissues and cells were extracted using TRI reagent (Sigma, St.

Louis, MO). cDNA was synthesized using the High Capacity cDNA Reverse

Transcription Kit (Applied Biosystems) in a Thermal Cycler (Applied Biosystems) at

25°C for 10 min, 37°C for 2 hr and 85°C for 5 s. Real time PCR was then performed

using SYBR Green fast master mix (Quanta Biosciences, Gaithersburg, MD) in the

7300 PCR System (Applied Biosystems, Foster City, CA) at 95°C for 10 min, followed

by 35 cycles of 95°C for 15 s and 60°C for 1 min. A serial dilution of a sample was used

as the standard curve to quantify all the samples for relative mRNA expression level.

Ribosomal protein L7a was used as an internal control. Primers used for qPCR are

shown in Table 3.4.

Protein Extraction and Western Blotting

Liver tissues were lysed with 1X Laemmli Buffer [62.5 mmol/L Tris–HCl, pH 6.8, 2%

SDS, 10% Glycerol v/v, 0.01% Bromophenol blue, 5% 2-mercaptoethanol, 1x protease

inhibitors (Roche, Indianapolis, IN) and 1x phosphatase inhibitors (Sigma-Aldrich)]. For

29

longer storage, the protein samples were placed in -70 ºC. Lowry assay was performed

to determine protein concentration. Samples containing 30 μg of protein were resolved

by 10% SDS-PAGE. A wet transfer protocol was performed to complete the transfer. A

PVDF membrane (0.2 μm, Bio-Rad, Hercules, CA) was blocked in 10% milk in TBS/T

(30 mM Tris base pH 7.6, 200 mM NaCl and 0.1% Tween 20) for one hour at room

temperature. The membranes were incubated with antibodies against COX-2 (#4842,

Cell Signaling Technology, MA) in 10% non-fat dry milk at a 1:1000 dilution at room

temperature for 3 hours. SuperSignal West Dura Extended Duration Substrate kit

(Thermo Scientific, Rockford, IL) was used to detect the signal and the images were

captured and analyzed by a Chemi Doc XLR system (Bio-Rad). Actin was used as an

internal control to normalize the protein data.

Immunofluorescence

To identify the localization of COX-2 in liver, immunofluorescence staining of COX-2

was conducted using an Alexa Fluor 647-labeled anti-rabbit secondary antibody (red)

(Invitrogen). Frozen liver samples were embedded in Tissue-Tek OCT compound (VWR,

Radnor, PA) and cut to a thickness of 5 µm in a cryostat at -20 °C. The sections were

fixed in 70% ethanol, permeabilized with 0.1% Triton X-100, blocked with Image-iT FX

signal enhancer (Invitrogen) for 30 min, incubated with primary antibody COX-2 (1:200,

cat # 9587S, Abcam) for 2h, followed by an incubation of the secondary antibody (1:200

dilution) for 45 min in a dark chamber. The samples were counterstained with Hoechst

33342 (Invitrogen) for 15 min and then washed with PBS. Coverslip was mounted on

the sections using Prolong Gold anti-fade reagent (Invitrogen). Pictures were taken

using the Zeiss LSM 700 Confocal.

Statistical analysis

Body weight and food intake were tested using repeated-measures one-way ANOVA.

The synergistic effects of maternal and postweaning high fat diet were analyzed using

Two-Way ANOVA [251]. Post-hoc analysis was performed using Fisher’s least

significant difference (LSD, SAS v. 9.1.2, SAS Institute, Cary, NC) when the P-value of

interaction of main effects (maternal vs postweaning) was less than 0.15. Individual bars

30

with different letters differ (P<0.05). Significance was set at p<0.05 for all comparisons.

Individual bars with different letters differ (P<0.05). Correlation between COX2 mRNA

and COX-2 pre-mRNA or Δ6desaturase mRNA was determined by Spearman's rank

correlation coefficient analysis (Spearman's rho).

Results

Growth Curve and Calorie Intake

An overall dietary effect was observed on both growth curve and calorie intake.

Individual comparison between treatment groups showed that maternal diet did not

significantly impact body weight gain (Figure 3.1A) nor calorie intake (Figure 3.1B) of

offspring fed the same diet. However, there was a clear overall separation by post-

weaning diet (Table 3.1). Offspring fed with the high fat diet exhibited a greater increase

in body weight gain (Figure 3.1A) and calorie consumption (Figure 3.1B) compared to

that of the control group, regardless of maternal diet.

Liver Histology and Lipid Profile

H&E and Oil Red O (Figure 3.2) staining were graded individually to evaluate the

pathological changes corresponding to HF dietary intervention at different life-stages

(Figure 3.2). Compared to C/C, only life-long high fat feeding (HF/HF) remarkably

increased the severity of liver damage at the end of 12 weeks. Severe micro-and/or

micro-hepatocellular vacuolation (Figure 3.3, top) and increased hepatic neutral lipid

accumulation (Figure 3.4, top) resulted. In contrast to HF/HF, early life stage high fat

treatment (HF/C) or post-weaning high fat treatment (C/HF) did not lead to severe liver

damage; post-weaning high fat treatment (C/HF) elevated hepatic fat accumulation

(Figure 3.4, top) whereas a NHF diet (HF/C) significantly increased circulating TAG in

comparison to the remaining groups (Figure 3.4, bottom).

Liver Fatty Acid Profile and Associated Enzymes

Palmitoleic acid (PA, 16:1n-7) content was decreased in post-weaning high fat-exposed

groups (C/HF, HF/HF), while the content of linoleic acid (LA, 18:2n-6) increased in the

31

C/HF group. Arachidonic acid content (AA, 20:4n-6) was significantly greater in high fat-

treated groups, regardless of the life-period of exposure (Table 3.2). Palmitoleic and

linoleic acid are fatty acids that have no direct association to COX-2 induction [252],

however in many circumstances linoleic acid can be converted into arachidonic acid,

which is the primordial substrate for COX-2 catalysis. To investigate the sources of

arachidonic acid in the liver of all high fat groups, mRNA expression of both

Δ6desaturase that converts linoleic to arachidonic acid [253] was measured. HF

treatment at all life stages increased mRNA expression of Δ6desaturase significantly.

Therefore, accumulation of arachidonic acid exerted by all high fat diets is due to the

desaturation reaction actively associated with COX-2 induction.

HF Diet Regulates Gene Expression

COX-2 mRNA expression was notably increased in all high fat-fed groups (HF/C, C/HF

and HF/HF) (Figure 3.5). Statistical analysis revealed that compared to C/C, exposure

to a high fat diet during an early developmental (fetal and suckling) or postnatal period

significantly induced COX-2 mRNA expression by ~10 fold; Life-long feeding of the high

fat diet (HF/HF), however, stimulated COX-2 transcription to a remarkably high level,

which was significantly different than COX-2 levels in both HF/C and C/HF; furthermore,

COX-2 pre-mRNA, as well as Δ6desaturase mRNA in HF/HF, did not differ from HF/C

or C/HF, which may be due to the high variation within the groups. Nevertheless, both

COX-2 pre-mRNA and Δ6desaturase mRNA were positively correlated to COX2 mRNA

level (Figure 3.7), suggesting a synergistic effect of maternal and post-weaning high fat

diet on these two genes.

Protein Abundance of COX-2 Increased by High Fat Treatment

To confirm the protein content of COX-2 in the four treatment groups, both Western blot

(Figure 3.6) and Immunofluorescence (Figure 3.8) were conducted. Western blot

provided an average value of COX-2 protein content in each group and showed that

only the HF/HF had a significantly increased amount of COX-2 compared to the C/C

group. Neither HF/C nor C/HF are statistically different than C/C. Nevertheless, protein

amount was positively and significantly correlated to COX2 mRNA level (Figure 3.7).

32

Immunofluorescence presented both protein localization and quantification of COX-2. A

representative image from each group was presented and showed the relative quantity

of COX-2 on the tissue section. Almost no visualized COX-2 staining could be seen in

C/C. The most abundant COX-2 positive signal is received from HF/HF, followed by

HF/C and C/HF.

Discussion

The results demonstrate for the first time a double-hit effect of a high fat diet on COX-2

gene expression. Exposure to a high fat diet throughout the lifespan leads to a more

severe activation of COX-2 than a simple maternal or post-weaning introduction.

Furthermore, as opposed to maternal or post-weaning diet, alone, life-long high fat

feeding exerts a synergistic effect on COX-2, resulting in a significantly greater induction

of COX-2 in liver tissue.

A previous study that used a mouse model stated that early-life exposure to a high fat

diet programmed susceptibility to NAFLD development in the offspring [84]. Fetal

programming thus may create an environment that is less resistant to the progression of

pathogenic events. The likelihood of developing liver disease dramatically increased in

offspring exposed to high fat diets during fetal growth. This hypothesis was validated by

both physiological and pathological evidence, which showed that the most definitive

progression towards NASH occurred in the HF/HF group, rather than C/HF or HF/C

[84]. This study agreed with this previous observation by showing the evaluation of

hepatic damage and lipid accumulation, as well as the increased gene expression of the

pro-inflammatory marker COX-2, a central reflection of both chronic inflammatory status

as well as liver injury. These measurements reinforced the knowledge that the severity

of liver damage is exaggerated when the offspring are exposed to a high fat diet during

early development. The synergistic effect of maternal and post-weaning high fat diets on

COX-2 gene expression was statistically significant and may biologically account for

greater detrimental outcomes in adulthood.

33

Only certain types of cells can express COX-2 upon being stimulated by inflammatory

signals, and these are mostly non-parenchymal cells (NPC) such as Kupffer and

endothelial cells [254, 255]. For example, it has been suggested that peroxidization of

fatty acids was responsible for activated COX-2 expression in NPC, but not hepatocytes

[254]. Our immunofluorescent images showed that COX-2 protein abundance varied

according to the time of high fat exposure, in agreement with the mRNA level. COX-2

positive staining was seen sporadically at certain sites in HF/C and C/HF. In contrast,

HF/HF treatment produced more COX-2 signal. This once again strongly suggested the

maternal fetal programming effect, meaning that exposure to a high fat diet during early

development primes the liver tissue of male offspring with a more intense response to

post-weaning high fat contact. However, it was difficult to differentiate the cell types in

the liver of HF/HF pups which produced excessive amounts of COX-2. Still, this

outcome could be attributed to a wider range of NPC activation, or to hepatocyte

activation. Further investigation is needed to answer this particular question.

Interestingly, the hepatic fatty acid profile was also modified by high fat dietary

treatment. Neither the control nor high fat diets contain significant amounts of

arachidonic acid, which could account for the animals experiencing such an increase.

However, the proportion of arachidonic acid increased in the liver of all HF-treatment

groups in comparison with the control. Arachidonic acid is mainly distributed throughout

the lipid bilayer, and functions as a major substrate for cyclooxygenase enzymatic

reactions. In the presence of adequate amounts of linoleic acid or fully functional

Δ6desaturase, the conversion of linoleic acid to arachidonic acid is enabled [256]. In our

study, Δ6desaturase mRNA abundance positively correlated with COX-2 gene

expression across high fat treatment groups, which suggests a direct crosstalk between

the enzymes. Previous dietary studies have suggested that hepatic Δ6desaturase can

be induced by low level of dietary intake of essential fatty acids [257]. Control diet (AIN-

93) was used in this study because it provides complete nutrients that optimally meet

animal requirements. Therefore, the induction of Δ6desaturase could be triggered by

other cellular signaling molecules that participate in COX-2 reactions.

34

Conclusion

Based on the results of the present study, it is not surprising to find a synergistic effect

on hepatic physiology as a result of combined maternal and post-weaning high fat

feeding. The most intriguing part resides on the gene programming of COX-2. This is

the first study to date to demonstrate that a “double hit” impact occurred on the gene

expression level. In addition, it was revealed that in high fat dietary models the body

adapted to environmental changes, meaning that it produced more arachidonic acid as

a substrate for the support of COX-2 catalysis. Pre-mRNA expression correlated with an

mRNA level of COX-2, which suggests a transcriptional regulation of hepatic COX-2 by

high fat diets. As regards the current knowledge gap concerning the epigenetic

regulation of COX-2, current study will further investigate DNA methylation and histone

modifications in the next chapter.

35

CHAPTER 4. HIGH FAT DIET INDUCES EPIGENETIC ALTERATIONS OF COX-2 GENE

Abstract

High fat diet is associated with pathological development and the progression of liver

diseases. It is hypothesized that high fat diet changes the genomic landscape of DNA

methylation and histone code, leading to differential expression of COX-2 in hepatic

tissue. Methyl-DNA immunoprecipitation assay in combination with high-throughput

sequencing (MeDIP-seq) and methylation-sensitive restriction enzyme sequencing

(MRE-seq) were applied to illustrate the genomic landscape of DNA methylation

modified by a high fat diet. The results showed that in all three high fat dietary groups,

multiple regions of the COX-2 gene exhibited DNA hypomethylation, including the

5’UTR and multiple coding regions. Chromatin Immunoprecipitation analysis (ChIP)

showed that high fat diets programmed histone methylation and acetylation by

increasing H4Ac, H3Ac and H3K4Me2, which characterize an open chromatin structure

and active gene transcription. In conclusion, activation of COX-2 mRNA in the livers of

high fat treated pups is associated with DNA hypomethylation and histone

modifications.

Introduction

High fat dietary feeding results in various long-term health problems, and

overproduction of COX-2 plays a central role. More particularly, COX-2 induction is

known to actively participate in body weight and fat mass gains, blood pressure

elevation, glucose and lipid level increases, glucose tolerance impairments and

steatohepatitis development [258-260] COX-2 performs critical biological functions, so

its expression should be closely monitored.

Epigenetic modification dynamically orchestrates gene expression. It actively

participates in the processes of cell differentiation, proliferation, and apoptosis [261]

primarily through transcriptional and post-transcriptional manipulation of gene activity

36

[262]. The epigenetic profile is subject to re-modulation through inter- or intracellular

signaling[263]. Therefore, the management of epigenetics is a novel contemporary

therapeutic approach for the treatment of certain diseases.

It has recently been recognized that dietary fat molecules have the potential to stimulate

cellular epigenetic signals, profoundly affecting chromatin structure as well as gene

expression, signaling pathways, and subsequent biological processes [264]. Consuming

a high fat diet increases the risk of obesity and metabolic syndrome, which are often

accompanied by low-grade chronic inflammation [265]. Moreover, high fat diets have

been previously shown to program the epigenetic profile across numerous tissues [225,

266]. Reprogramming the histone code [200], together with altered DNA methylation

[225], have been documented for differentially expressed genes in the high fat model.

Although gene deregulation can be attributed to aberrant DNA methylation and/or

histone modifications, it has been considered to play a role in the production of

pathophysiological outcomes [267]. However, it remains unclear whether, first, the

alteration of DNA methylation and/or histone modifications are solely accountable for

high fat-induced COX-2, and secondly, whether the timing of exposure to a high fat diet

differentially shapes the epigenetic profile.

A wide range of stimuli, including pathogens, cytokines, nitric oxide, and growth factors,

are known to quickly but transiently activate COX-2 transcription [268]. Depending upon

the cell type and stimulus, transcriptional control of COX-2 depends on a combination of

transcription factors that bind to specific sites on the 5’-UTR (untranslated region) of the

gene. More particularly, binding motifs of NF-κb (nuclear factor κb), AP-1 (activator

protein 1), and CREs (cAMP-response elements), were found at gene 5’UTRs [269].

Transcription of COX-2 is also influenced by the acetylation status of histone and non-

histone proteins. Upregulation of p300, a histone acetyltransferase (HAT), activates

COX-2 by acetylating NF-kB components [270]. By contrast, HDAC inhibitors suppress

the activation of COX-2 expression in human primary myometrial cells [271] and cancer

cell lines [272] by preventing the binding of transcription factor c-Jun to the COX-2

promoter.

37

DNA methylation status may likewise affect COX-2 expression because two CpG

islands are located at the core promoter of COX-2, the hypermethylation of which

usually results in transcriptional silencing [273]. On the other hand, post-transcriptional

regulation of COX-2 can occur through changes in its mRNA stability [274]; the 3’-UTR

of COX-2 contains numerous AU-rich elements (AREs), which are bound by specific

trans-acting ARE-binding factors. Many of these AREs are located within the first 150

nucleotides of the COX-2 3’-UTR. CUGBP2 is a RNA binding protein. When it is bound

to specific AREs within the first 60 nucleotides of the 3’-UTR, that stabilizes COX-2

mRNA while inhibiting its translation. Consequently, there are many known checkpoints

for the study of COX-2 gene expression. However, there is a lack of up-to-date

understanding of the epigenetic-based molecular mechanisms by which high fat diet

activates COX-2 gene expression. Our study aims to investigate the epigenetic

regulation of COX-2 using a high fat diet, and seeks to determine whether the timing of

high fat exposure generates a differential DNA methylation pattern and histone

modifications.

Materials and Methods

Genomic DNA Isolation

Ten milligram of liver tissue was ground in liquid nitrogen and genomic DNA was

extracted in 600µL of Extraction Buffer (50mM Tris, pH 8.0, 1mM EDTA, pH 8.0, 0.5%

SDS, 1mg/ml Proteinase K) at 55 ºC overnight. Lysates were centrifuged and

supernatant was transferred to 2mL phase lock gel (PLG) (5’-Prime, Fisher Scientific

Company, LLC). Phenol/chloroform extraction (PCI) was performed to purify DNA.

Briefly, equal amount of aqueous sample and organic extraction solvent were gently

mixed in PLG tube and centrifuged at 16,000 g for 5 min to separate the phases. Upper

phase was transferred and 1µL of RNase (Roche, 10mg/mL) was added with, followed

by 1 hour of incubation at 37 ºC. Another PCI and 1Xcholorofom extraction was

performed. Purified DNA was precipitated with 1/10 volume of 3M Sodium Acetate (pH

38

5.2), and 2.5 volumes 100% EtOH. DNA pellet was washed with 70% EtOH and

resuspended in TE.

Methyl-DNA Immunoprecipitation Assay (MeDIP) in Combination with High-

Throughput Sequencing (HTS)

DNA methylation status in liver samples of rats was examined using sequencing-

coupled methylated DNA immunoprecipitation (MeDIP-seq). Genomic DNA from each

liver sample was purified and subjected to sonication to ~100–500 bp. Sonicated DNA

was end-repaired, A-tailed and ligated to adapters following the standard Illumina

protocol. Agarose size-selection was used to remove unligated adapters. A mouse

monoclonal anti-methylcytidine antibody (Eurogentec) was then applied to adaptor-

ligated DNA for each immunoprecipitation. Fifteen cycles of PCR were performed on

the immunoprecipitated DNA using the single-end Illumina PCR primers. Final size

selection (220–420 bp) was performed in a 2% agarose electrophoresis, and the

resulting reactions were purified with Qiagen MinElute columns. Quality was checked by

spectrophotometry and Agilent DNA Bioanalyzer, which indicated an average fragment

size of 150 bp. An aliquot of each library was diluted and used as template in four

independent PCR reactions to confirm enrichment for methylated and de-enrichment for

unmethylated sequences, compared to 5 ng of input (sonicated DNA). Each library was

diluted to 8 nM for sequencing on an Illumina Genome Analyzer following the

recommended manufacturer's protocol at the core facility of the Center for Comparative

and Functional Genomics [275, 276].

Chromatin Immunoprecipitation [275]

ChIP analysis was employed according to a modified protocol [277]. Briefly, 200 mg

frozen liver samples from each pup were ground and resuspened in phosphate buffered

saline (PBS). Cross-linking was performed in 37% formaldehyde for 10 min on a rotator

at room temperature. The pellet was resuspended in nuclei swelling buffer and lysed in

SDS lysis buffer, both of which contained protease inhibitors. The chromatin was

sonicated (Fisher Scientific, model 100 Sonic Dismembrator) on ice with 7 bursts for 40

s and 2 min cooling interval between each burst with power set at 5. After removing cell

39

debris, sheared chromatin was diluted in ChIP dilution buffer. Diluted lysate (1 ml) was

incubated with 2 μg of primary antibody of interest (Table 4.1) overnight at 4 C on a

hematology mixer (Model 346, Fisher Scientific). Pre-blocked salmon sperm

DNA/protein G agarose beads (60 μL, 33% slurry; Millipore) were then incubated with

each chromatin sample for 2 hours, followed by centrifugation for 1 min at 4 C.

Supernatant from incubation with normal rabbit IgG was saved as input. The pellets

containing immunoprecipitated complexes were washed by a series of salt buffers, and

antibody/protein/DNA complexes were eluted from Protein G agarose beads. After

incubation at 65 C for 5 hours with 20 μL 5 M NaCl and 1 μg of RNase A to reverse the

cross-linking, protein was removed by Proteinase K, and DNA fragments were

separated from chromatin structure. DNA was purified by a Wizard SV Gel and PCR

Clean up System (Promega). Purified DNA with a dilution series as a standard curve

was quantified by real time PCR reaction with primers designed for COX-2 promoter

(Table 3.4).

Statistical Analysis

ChIP results are expressed as mean±SEM. Mean differences in promoter protein

binding, and histone modification between C/C and HF/HF groups were determined by

SAS statistical analysis using one-way ANOVA. Significance level was set at P<0.05 for

all comparisons.

Results

HF Diet Induced Hypomethylation of COX-2 Gene

MeDIP-seq results showed a decreased DNA methylation intensity at the COX-2 5’UTR

upstream and three coding regions (Figure 4.1, 4.2), whereas little difference was found

at any β-actin regions (Figure 4.3, 4.4), suggesting that the exposure to high fat diet

induced region-specific hypomethylation of COX-2 in rat liver. No DNA methylation

differences were observed at the two CpG islands located at the proximal promoter

region of COX-2. Decreased methylation intensity consistently occurred in all HF groups

at 1KB, 2KB and 4KB of the coding region and 1.2 KB of 5’-flanking region.

40

Histone Modifications

Chromatin Immunoprecipitation [275] analysis showed that histone acetylation (histone

H3 and H4) and methylation at lysine residues (K4, 9 and 27) of histone H3 in HF/C and

C/HF were not different from that of C/C group (data not shown). However, life-long high

fat feeding (HF/HF) modulated histone tails of COX-2. A wide range of reduced

association of HDAC3 to the DNA sequence was observed, which may correlate to the

increased acetylated histone H3 (H3Ac) and H4 (H4Ac) at coding regions (Figure 4.5).

Also, increased di-methylation of histone H3 lysine 4 (H3K4Me2) occurred across

multiple sites at both upstream and coding regions. Increased tri-methylation of histone

H3 lysine 9 (H3K9Me3) was detected at -1.2kb and +3kb. The overall biding levels of

H4Ac, HDAC3 and H3K4Me2 are remarkably higher than IgG, whereas H3Ac,

H3K9Me3 and H3K27Me3 are relatively low. Increased acetylation of histones H3 and

H4 often suggests an open chromatin structure and thus greater accessibility of DNA

transcription regulators [278]. Hence, life-long feeding of a high fat diet programs

histone modifications of the COX-2 gene in the liver, which in turn matches the activated

transcription patterns.

Discussion

This is the first report concerning how high fat diet-induced COX-2 gene expression is

associated with persistent decreased DNA methylation at 5’ flanking and multiple coding

regions. However, histone reprogramming, unlike DNA methylation, can be detected

only in HF/HF pups, meaning that the modifications were dependent on the stage of life

during which exposure to a high fat diet occurred.

The interaction between nutritional and epigenetic influences on gene regulation has

been of great interest to scientists who deal with chronic diseases and the improvement

of human health. For decades, the limited experimental technology for the detection of

DNA methylation constrained most studies to focus exclusively upon the impact of the

diet on the methylation profile within a limited region of a single gene, and its

41

association with gene expression. The combination of MeDIP-seq and MRE-seq provide

a detection method that provides a comprehensive high-resolution genome-wide

coverage. This allows researchers to map out a slight shift in DNA methylation in the

human genome [279].

High fat diets have been shown to mediate DNA methylation on the gene promoter and

affect gene expression, which thereby leads to physiological alterations. The genome is

composed of gene and non-gene associated regions. This study investigated the effects

of high fat diets on the DNA methylation pattern within the associated genetic region,

including the promoter, coding, and downstream regions. A study reported that high fat

diets induced greater DNA methylation in the leptin promoter, which parallels the

decrease in the expression of leptin [229]. Using a genome-wide promoter to analyze

DNA methylation has shown that fatty acids increase the methylation of PPARγ

coactivator-1 (PGC-1α) in skeletal muscle cells [280]. A recent study, which used

Illumina’s Infinium Bead Array, showed that a short period of high fat overfeeding

induced genome-wide DNA methylation changes [281]. Still, there is no evidence on

how maternal and post-weaning diet, alone or combined, can impact DNA methylation

profile of COX-2 in the livers of the offspring. Combining two advanced sequencing

methods, namely MeDIP-seq and MRE-seq, allowed us to be the first to present

evidence that high fat diets induce region-specific hypomethylation of COX-2 in the liver

of rats.

DNA methylation that occurs at genomic locations other than the promoter region can

affect gene expression as well. It have been previously showed that a high fat diet was

able to induce COX-2 gene transcription, independent of the feeding period. The

present study observed that high fat feeding resulted in DNA hypomethylation at the 5’

upstream and coding region of COX-2. Generally speaking, DNA hypomethylation is

closely associated with gene repression or silencing. Our results thus suggest that high

fat dietary intake delivers a strong signal to the epigenome, generating epigenetic

behaviors, which in turn cause DNA hypomethylation in regions other than the

promoter: this might be related to activation of COX-2. In addition, the existence of

42

positive correlations between DNA methylation in regions other than the promoter

sequence and COX-2 gene enhancement might lead to the discovery of novel

regulatory regions.

High fat diets enabled alterations of DNA methylation profiles, which led to further

investigation of histone modifications. This study investigated several histone variations

of interest, and found that the histone patterns for HF/C and C/HF were quite different

than HF/HF. Comparing to the C/C group, no sound evidence of altered histone

modifications was found in HF/C or C/HF. It thus remains difficult to define any possible

participation of these histone modifications by the activation of COX-2, or DNA

hypomethylation at any genetic region in HF/C and C/HF groups. Nevertheless, there

appears to be a direct link to the HF/HF group. In most cases, histone acetylation,

namely H3Ac and H4Ac, and H3K4Me2, are associated with transcriptional activation.

Unlike DNA methylation, a particular configuration of histones depicts a very unique

pattern in response to exposure of a high fat during different stages of life. It is worth

noting that the effects of double-hit theory were reflected on the gene expression level

of COX-2, as well as some physiological measurements, but neither for DNA

methylation nor for histone modifications. Although changes in histone modifications

were only detected in HF/HF, we cannot rule out the possibility that other histone

modifications were taking place in these treatment groups, given the limited

modifications tested and current knowledge of the histone code.

Fatty acids have been described as signal transducers that can be sensed by the

epigenome [264]. Our results revealed a full landscape of DNA methylation profiles of

COX-2, and showed that high fat diets removed methyl groups from CpG sites in

multiple regions. The length of the traditional basic promoter of COX-2 is limited to 1kb,

which comprises two detectable CpG islands. In our study, none of the changes, or only

minimal changes, were identified by MeDIP-seq within this promoter region or the CpG

islands, while a substantial reduction in methylation intensity was observed in the

promoter upstream and coding regions, which correspond to the activation of COX-2.

Interestingly, the location and grade of reduction is similar among all HF groups,

43

regardless of the stage of life during which exposure to a high fat diet occurs. Thus,

DNA hypomethylation of the COX-2 gene is less likely to be connected to a double-hit

effect on COX-2 mRNA abundance in this particular model.

Conclusion

Our findings provide insight into nutritional modulation of the epigenome, which is

involved in liver health and function. For the first time anywhere, the impact of a high fat

diet on the full landscape of DNA methylation of COX-2 was showed in the livers of rats.

In our high fat feeding models, hepatic COX-2 activation is persistently accompanied by

DNA hypomethylation, but is not accompanied by histone modifications. These region-

specific hypomethylations induced by high fat diet are positively associated with

changes in gene expression. Future site-specific validation research should be

conducted in order to confirm DNA methylation change and to identify the transcriptional

functions of these regions.

44

CHAPTER 5. AN ENHANCER WAS CHARACTERIZED AT COX-2 UPSTREAM REGION

Abstract

High fat diets, regardless of the feeding period, significantly induced COX-2 gene

expression in the livers of male offspring pups. Genome-wide methylated DNA

immunoprecipitation (MeDIP) showed that multiple regions of COX-2 gene sequences

were hypomethylated, which corresponds to the activation of gene expression. Bisulfite

sequencing was therefore performed to confirm the MeDIP results, and showed that

hypomethylation occurred on a specific CpG site within the 5’UTR of COX-2. This

region was further cloned to test the transcriptional activity using a luciferase reporter

assay. The results show that this hyomethylated region produced strong enhancer

activity. Further characterization showed that the binding elements HOXD8, RAF and

CTCF contribute to the enhancer’s activity at different levels. In addition, the

transcriptional activity of this enhancer cannot be further induced by the presence of

arachidonic acid. In conclusion, high fat induced hypomethylation of this enhancer may

be responsible for triggering gene activation of COX-2.

Introduction

COX-2 is present in cells in minute amounts under normal conditions [282], which

justifies why induction of the gene is necessary and under exquisite regulation. Many

experiments have been conducted to study COX-2 gene regulation, such as the

Luciferase reporter assay, a widely used lab tool that is used to determine promoter

strength and enhancer activity. Various cis-elements in the human COX-2 promoter

were revealed that operate to control luciferase reporter activity. These core promoter

elements, including nuclear factor kappaB (NF-kB), CCAAT/enhancer-binding protein

(C/EBP) and cyclic AMP-response element (CRE) sites, were shown to be modulated in

response to environmental stimuli, such as sulforaphane [283] and 6-

(Methylsulfinyl)hexyl isothiocyanate [284]. These critical binding elements are

exclusively located within 300 base pairs upstream of the transcription starting site

45

(TSS). Highly conservative binding domains for CREB, NF-Kb, AP-1, and C/EBP can

also be found within the promoter region of murine COX-2 [285]. Despite the fact that a

number of transcription factors are commonly found in humans, murine, and rats, each

of them may exhibit unique responses toward different environmental stimuli, and these

require further inspection and specification. Nevertheless, the investigation of the

5’UTR’s regulatory role of COX-2 has never gone beyond its promoter, ranging from -

900 to TSS.

In the previous chapter, an important aspect was presented that high fat diet actively

manipulates COX-2 gene expression by modifying the landscape of its epigenome.

Three hypomethylated regions in the high fat feeding models were identified, two of

which situate in the gene body, and the other in the 5’UTR. The hypomethylation of

DNA fragments in this particular dietary model perfectly correlates with the transcription

level, and may or may not account for the gene activation of COX-2. Little research has

been done on these particular regions to determine whether or not they exhibit any

regulatory functions or transcriptional activities. In addition to the single CpG sites that

are spread over the entire gene body and 5’UTR, two CpG islands are primarily

distributed in the proximal promoter region. Methylation of CpG sites within the promoter

can lead to gene silencing. However, these two CpG islands are barely methylated.

Alteration of them is independent of high fat treatment as opposed to the single CpG

sites, the methylation of which can be reduced by high fat diet. Even though the MeDIP-

seq provides an overview of the DNA methylation pattern of COX-2, it is quite difficult to

pinpoint the CpG site(s) that were truly modified. In addition, we cannot rule out the

possibility that the hypomethylated CpG sites varied in the three high fat groups.

The traditionally defined rat COX-2 promoter does not cover this particular region

ranging from -1.2kb to -1.1kb. This short DNA fragment contains multiple TF binding

elements, several of which have not been studied in the context of COX-2 gene

regulation. Moreover, in contrast to the default condition, which is highly methylated, the

DNA methylation level in this region was nearly evened out by using high fat diet. In

other words, the fold of reduction of methylation intensity in the 5’UTR region exceeds

46

those in the coding region. On the other hand, methylation of CpG regions downstream

of the transcription starting site doesn’t block elongation, which means it may not

critically be involved in transcription initiation [286]. Thus, this study is particularly

interested in exploring this hypomethylated region in the vicinity of the COX-2 core

promoter, because it may also represent an undefined regulatory region (enhancers,

silencers, boundary elements/insulators) working in concert with the promoter. In order

to answer the questions proposed above regarding methylation of which CpG sites were

altered by high fat diet and the degree to which this region contributes to COX-2 gene

transcription, bisulfite sequencing and a series of luciferase reporter assays were

conducted.

Materials and Methods

Bisulfite Sequencing

To confirmation DNA methylation at 5’ upstream region (region 1), bisulfite conversion

of genomic DNA was performed using EZ Methylation-Gold kit (Zymo Research,

Orange, CA) according to the manufacturer’s instructions. Converted DNA was

amplified in 5X LongAmp Taq Buffer (XX), 10 μM dNTPs, 5 μM forward and reverse

primers (-1533 to -1109), and 1 U of Long Amp Taq polymerase. The cycling conditions

used were as follows: 45 cycles of 95°C for 30 seconds, followed by 95°C for 10

seconds, 48°C 30 seconds and 65°C for 30 seconds, then 65°C for 10 minutes. PCR

products were elongated with 5X Go Taq Flexi Buffer, 25 mM MgCl, 10 mM dATP, and

Go Taq Flexi DNA polymerase at 72°C for 10 minutes. The PCR product was

concentrated with a DNA Clean & Concentrator -5 kit (Zymo Research), ligated

overnight at 16°C into the TOPO PCR 4.0 TA (Invitrogen) in accordance with the

manufacturer's instructions and transformed into DH5α E.coli (New England Biolabs).

Plasmid DNA was isolated with the QiaPrep Spin Miniprep kit (Qiagen XX) and at least

50 independent recombinant clones per sample sequenced at the University of Illinois

Core Sequencing Facility (Urbana-Champaign, IL) and the methylation data was

compiled using the BISMA (Bisulfite Sequencing DNA Methylation Analysis) software

(http://biochem.jacobs-university.de/BDPC/BISMA/). The lower threshold conversion

47

rate was set to 90%, while all the other analysis parameters were set to the BISMA

default values (Mamrut, Harony et al. 2013).

Cell Culture and Arachidonic Acid Treatment

The human prostate cancer cell lines (PC3) was obtained from the American Type

Culture Collection (Manassas, VA) and was cultured in Roswell Park Memorial Institute

(RPMI) medium supplemented with 5% (v/v) fetal bovine serum (Mediatech; Herndon,

VA) and 100 IU/mL Penicillin, and 100 μg/mL Streptomycin (Mediatech; Herndon, VA).

Cells were maintained at 37 °C in 95% humidity and 5% CO2. For Arachidonic acid (AA)

treatment, cells were first seeded at a density of 0.5×10^6 cells in a 6-well plate, grown

for 16 h, and then starved with 0.3% FBS supplemented with 1.25mg/mL fatty acid free

BSA (Sigma) for 48 hours before treatments. AA was neutralized with 1N NaOH and

dissolved in the starvation media. Cells were incubated with 10μg/ μL AA for 30min, 1h,

2h and 3h to induce COX-2 expression.

Plasmid Construction and Methylation

Wild type COX-2 upstream fragments -1234/-1199, -1204/-1160, -1164/-1131, -1234/

+30, -1234/ -1029, -1234/ -917 and -1234/ -814 were amplified by a Thermal Cycler

(Applied Biosystems) using LongAmp Taq DNA polymerase (New England Biolabs).

The primers/fragments can be found in Table 3. 1. The reaction was performed at 94°C

for 30 sec, followed by 40 cycles of 94°C for 30 s, 67°C for 60 sec and 65°C for 50 sec

to reach maximum amplification, and ended with 10 min of 65°C for final extension. The

restriction enzymes used below are all acquired from New England Biolabs. The

fragment of -1234 to -+30 was cut with NheI, gel-purified, and cloned into the NheI site

of a luciferase reporter vector PGL3-basic (PGL3-b) (Promega Corporation). This

generated forward and reverse- orientated products. The forward product was cut with

SacI, deriving a partially truncated promoter region (-774/+30) and upstream region (-

1234/-774), and SmaI generating a fragment of -1234/-1131. The upstream fragment (-

1234/-774) was cloned into the SacI site of a PGL3-promoter (PGL3-p) vector and

generated forward and reverse- orientated products. This forward product was further

cut with SmaI and generated two upstream regions (-1234/-1131 and -1131/-774) in

48

PGL3-p vector. The reverse product of PGL3-b (+30/-1234) was cut with SmaI and

derived a partially truncated region (-1131/-1234), which was cloned into PGL3-p vector.

The fragments of -1234/ -1029, -1234/ -917 and -1234/ -814 were cut with SacI, gel-

purified, and cloned into the SacI site of PGL3-p, from where a forward- orientated

product was generated for each fragment. Selected plasmids were methylated using

CpG Methyltransferase(M.SssI) (New England Biolabs).

Transfection and Luciferase Reporter Assay

PC3 cells were plated in 24-well plate for 24hr prior to transfection. Each plasmid (0.5

ug) were transfected into each well using Superfect transfection reagent (Qiagen;

Valencia, CA) following manufacturer's instructions. Meanwhile, an empty reporter-

gene plasmid was transfected following the same procedure, and used as a control.

Cells were maintained in transfection for duration of 3 hr. Cells were then lysed with

Passive Lysis Buffer (PLB) and subjected to one round of freeze-thaw. Luciferase

activity was measured using Single-Luciferase Assay (Promega) with a luminometer

(Fermomaster FB12, Zylux Corporation; Huntsville, AL). For each experimental

condition, six replicates were performed.

Prediction of Transcription Factor Binding Sites and Site-directed Mutagenesis of

CTCF, RAF and HOXD8 Binding Element

Transcription Element Search System (TESS) was used to predict transcription factor

binding sites in DNA sequences. It identifies binding sites using site or consensus

strings and positional weight matrices from the TRANSFAC, IMD, and CBIL-GibbsMat

database. Mutant -1234/ -1131 construct were prepared using QuickChange Lighting

Multi Site-Directed Mutagenesis Kit (Agilent Technologies, Santa Clara, CA). Primers

contained mutations in putative CTCF, RAF and HOXD8 binding site (Table 3.1). PCR

products were digested with DpnI restriction enzyme (Agilent Technologies) to remove

the template DNA. These DNA constructs were sequenced to confirm the mutation site.

Statistical Analysis

49

Percent methylation patterns of all CpG sites were extracted and for each site the

probability of the observed pattern was calculated assuming a model of stochastic

binomial distribution for the calculation of the p-value. A parallel comparison using one-

way ANOVA was performed in between different transfections.

Results

Bisulfite Sequencing Showed a Site-specific CpG Hypomethylation

After MeDIP analysis, the region upstream of the hypomethylated promoter was located.

The specific upstream region from -1533 to -1109 was fully analyzed through bisulfite

sequencing, and the results for the methylation pattern for all 9 CpG sites are presented

on Figure 5.1. The graphical display of the average methylation at each CpG site

reveals a differential distribution of methylated cytosines between the control group (C/C)

and the high fat groups (HF/HF, HF/C, C/HF); The analysis of the average methylation

percentage (Figure 5.2)(Table 5.1) for each clone indicates greater site-specific CpG

methylation for the healthy control (C/C), compared to the high fat control group

(HF/HF), specifically on the 8th CpG site (-1169). Hence, the 8th CpG site is persistently

hypomethylated by high fat diet in this region.

Identification of An Enhancer Element Located at the 5’ UTR

To investigate the transcription function of the HF- induced hypomethylated fragment at

the 5’ region upstream of COX-2, several constructs were prepared from the rat gene

that contain this fragment (-1234/-1131, the hypomethylated region from MeDIP result)

as well as downstream genomic sequences through +30 and -774. These sequences

were cloned into the PGL3-B luciferase reporter vector and their ability to drive

regulated transcription was measured (Figure 5.3). The basic promoter of COX-2 (-

774/+30) has greater transcription activity, but not as great as the full fragment does (-

1234/+30). Therefore, the fragment of -1234/ -774 was cloned into Pgl3-p luciferase

reporter for further investigation. This construct containing the portion of the 3’ upstream

(-1234/-774) region exhibited a relatively small but reproducible increase in transcription.

Deletion in the 3’ direction yielded a construct (-1131/-774) that lost transcription activity

50

and another construct (-1234/-1131) that had significantly increased transcription

activity. The fragment containing the -1234/-1131 region, regardless of the orientation,

produced the highest transcription activity. When this DNA sequence was placed in

front of the Pgl3-b vector, it resulted in an even higher transcription induction than the

traditional promoter of COX-2 (Figure 5.3). Therefore, the DNA fragment of -1234/-1131

exhibits transcriptional enhancer activity, and it appears that this fragment can serve as

an alternative promoter to COX-2.

Effects of in vitro CpG Methylation on Enhancer Activity

In vitro methylation at CG dinucleotides in a transfecting plasmid can inhibit gene

expression in mammalian cells. In order to test if the methylated enhancer lost its

transcriptional activity, PGL3-P and PGL3-P containing the COX-2 enhancer were

methylated (Figure 5.5). The results showed that methylation of both plasmids have

deprived their original transcriptional activities. PGL3-P containing the COX-2 enhancer

exhibited 3 times greater ability to drive regulated transcription than PGL3-P. After

methylation, the transcriptional activity driven by the COX-2 enhancer was only

increased by less than 1 time. Therefore, in vitro methylation of the plasmid resulted in a

significant reduction of the transcriptional activity driven by COX-2 enhancer. This result

also suggested that the transcriptional activity of this COX-2 enhancer was not fully

dependent on methylation status. Other factors, such as the binding of TFs that are not

affected by DNA methylation, may sustain the remaining activity of the enhancer.

Transcriptional Activity of Different Fragment of COX-2 Enhancer

In order to locate the possible critical binding elements within this enhancer, three

shorter DNA fragments (Fragment 1, -1234/-1199; Fragment 2, -1204/-1160; Fragment

3, -1163/-1131) (Figure 5.6) were cloned into PGL3-P in both forward and reverse

position and were checked for their transcription activity (Figure 5.7). All of them

produced significantly higher luciferase signal than PGL3-P. In comparison to the

control plasmid, fragment 1 exhibited 3 times higher activity while fragment 2 and 3

exhibited 7 to 8 times higher activities in forward position. Even though the activity was

lower when these fragments were placed in reverse orientation, they were still able to

51

generate 3 to 4 times higher activity than the control plasmid did. These results

suggested that each fragment was sufficient to produce transcriptional enhancer activity,

and that fragment 2 and 3 might comprise stronger TFs binding elements to drive the

transcription initiation.

Site-directed Mutagenesis Revealed Crucial Transcription Factors

In order to characterize the critical and novel binding factors to this enhancer, binding

sites for predicted transcription factors were mutated and tested them with a luciferase

reporter assay. The predicted TFs have been mapped out in Figure 5.6 based on the

TESS analysis. The binding of these TFs tend be clustered. Therefore, mutation of the

DNA sequences that bind to one TF may also affect the binding of the ones in the

nearby vicinity. Several TFs have been shown to be associated with COX-2

transcriptional activity in a variety of cell types, including NF-1 [287], Ets-1 [288], c-Myb

[289], YY1 [290], PU.1 [291], HSP1 [292] and AP-2 [293]. Mutation of the binding

elements for these TFs is prone to alter transcriptional activity. In contrast, there have

not been any studies showing that COX-2 gene transcription is subject to the regulation

of POU1F1α, NF-1/L, TEF, HOXD8, HSTF, GAL4, GCN4, RAF, CTCF and MAZ. Given

the fact that POU1F1α, NF1-L, TEF, HSTF and GAL4 share consensus binding

sequences with TFs that were known to regulate COX-2 transcription, we decide to

begin the tests with RC2/HOXD8, RAF and CTCF. CTCF is of greater interest in the test

due to its insulator properties [294, 295] and its close relevance to DNA methylation

status [296]. The individual site-directed mutations were performed for each of them,

generating three sets of enhancers that carried one mutation site at a time (Figure 5.8

and 5.9). The Luciferase assay showed that mutated CTCF and RAF sites lessen the

transcriptional activity of the original enhancer by 25%, whereas mutated HOXD8 sites

entirely deplete transcriptional activity (Figure 5.9). These results, taken together,

suggested that RC2/HOXD8, RAF and CTCF are potential contributors to enhancer

activity.

Enhancer Activity Cannot be Further Induced by Arachidonic Acid

52

In accordance to the previous publication, AA was used to induce COX-2 expression in

PC3 cells. Samples were harvested at different time points. Compared to BSA control

cells, the greatest induction of COX-2 mRNA was acquired at a time of 3 hours (Figure

5.10). However, using the same treatment condition, luciferase activity of the enhancer

element in both orientations cannot be further induced by AA treatment (Figure 5.11).

Discussion

To our knowledge, these results are the first to identify a novel enhancer element at

5’UTR that is independent of the promoter sequence of the COX-2 gene. In high fat

treatments, this enhancer element was persistently hypomethylated, which may

contribute to the gene activation of COX-2 in the liver. Bisulfite analysis confirmed the

MeDIP-seq map and pinpointed the possible hypomethylated CpG sites on this

enhancer fragment.

The MeDIP-seq results in Chapter 4 led to the identification of a novel enhancer at the 5’

UTR of COX-2. Rat model-based COX-2 gene analysis has not been as extensive as

human and mouse analysis. This study preceded the previous research of rat COX-2

gene regulation. It was able to target a specific region based on the fully revealed DNA

methylation landscape of COX-2. In both humans and mice, a luciferase assay

established the predominant regulatory role of the COX-2 core promoter. Promoter

activities in these two species are quite significant. Mutations of these given TF binding

elements can result in a dramatically reduced transcriptional activity toward LPS

challenge [285, 297]. However, there appears to be a different regulation scenario for

rat hepatic COX-2, if one consider that the promoter location of this gene is

conservative across species. These results showed that the promoter activity of COX-2

in rats was not as great as what had been seen in both humans and mice. Instead,

when the tested fragment was extended and harbored the hypomethylated region at the

5’ upstream region of the promoter, this yielded stronger activity. This is an essential

piece of evidence that demonstrated that COX-2 activation in rat livers may require the

DNA sequence at the region which is 5’ upstream of the promoter. A further set of

experiments proved this point of view by revealing the enhancer activity of this particular

53

DNA fragment. Thus, it is speculate that consuming a high fat diet induces COX-2 at

least partially by motivating the enhancer activity, and the hypomethylation that

occurred on this enhancer could be an underlying mechanism which bridges the high fat

signal and enhancer stimulation. This enhancer also presents activity which resembles

that of the promoter. Additional investigations will be required to further characterize its

properties.

This study also provided the first evidence in rats that high fat diet undermethylated

single CpG sites within this transcriptional enhancer. DNA methylation on both CpG

islands and single/multiple CpG sites can be closely associated with gene expression

and disease development. Many genes contain more than one CpG island on their

genomes, and these are specifically associated with the start of the gene when they are

within promoter regions. Likewise, methylation of CpG sites within the promoters of

genes can lead to gene silencing, whereas reduced methylation intensity of CpG sites

has been associated with gene activation. This “on and off” phenomenon has been

found frequently in cancer [298, 299] and has also been observed in numerous other

types of diseases, such as heart disease [300], chronic kidney disease [301] and

NAFLD [302]. Activation of the human COX-2 gene was previously found to be inversely

correlated with the DNA methylation of a single CpG site within the promoter of patients

with chronic periodontitis [303]. In contrast to this study, rat COX-2 was induced when

methylation of a single CpG site within the enhancer was decreased, which means it is

with fewer relevancies to promoter methylation. This observation supports the

increasing amount of emerging evidence regarding mice [304] and human studies [305,

306] that DNA methylation state could be closely associated with gene expression

alteration. Furthermore, these cell type- [307] and lineage-specific [308] DNA

methylations of a given enhancer were expected to produce gradients of expression in a

wide range of tissues. Thus, DNA hypomethylation of this enhancer of hepatic COX-2

might be a very particular and unique consequence of a high fat diet.

Nevertheless, it was wondered how DNA hypomethylation of this enhancer might be

associated with COX-2 induction. Distal methylation sites that were associated with

54

expression were in a particular class of enhancers [306]. These enhancers were bound

to transcription factors depending on the DNA methylation state [306]. In other words,

altered DNA methylation at a given CpG site in these enhancers may affect the

downstream gene transcription by changing the interaction between DNA structure and

transcription factors. TESS analysis revealed a number of TFs that can potentially bind

to this enhancer sequence. Among them, CTCF generally binds to the hypomethylated

region [296], which partially protects surrounding genetic sequences from methylation

[309]. It can form boundaries by binding to the regions that are between enhancers and

promoters, and thereby block enhancer function [294, 295]. CTCF shares consensus

sequences with many other TFs, such as Sp1 and AP-2. Therefore, even though

mutation of CTCF binding sequence reduced transcriptional enhancer activity, it is not

assured that CTCF truly interacted with the enhancer. Additional sets of experiments

and analysis may allow revealing this point. The DNA sequence containing the 8th CpG

that is hypomethylated in high fat groups constitutes a binding spot for RAF. Mutation of

this binding site undermined the transcriptional activity of COX-2 enhancer, although it

is not clear as to whether DNA methylation of this CpG site affected RAF binding. There

hasn’t been a solid conclusion concerning the regulatory function of HOXD8 during

COX-2 activation. Our research is the first to reveal that removal of HOXD8 in COX-2

enhancer not only reduced the enhancer activity, but also diminished SV40 promoter

activity carried by PGL3-P. However, there hasn’t been a reasonable explanation

provided for this phenomenon. Although these DNA sequences appear to contain

elements that are essential to enhancer activity, further characterizations are

indispensable for fully defining this methylation-sensitive enhancer of COX-2.

In addition to characterizing the transcriptional function of hypomethylation region, the

current study is interested in finding a possible physiological cause that can explain why

and how it happened in the first place. In chapter 3, increased arachidonic acid content

of the livers of high fat treated pups was showed. Given that it was the sole substrate of

COX-2 enzymatic reaction, we want to determine whether arachidonic acid can promote

enhancer activity in order to induce COX-2 gene expression. However, the supply of

arachidonic acid did not further elevate enhancer activity in the PGL3-P construct.

55

There are multiple explanations why this might happen. First, arachidonic acid does not

affect this enhancer. Second, the induction condition for human PC3 cells cannot be

applied to rat COX-2. Third, arachidonic acid might induce enhancer activity by reducing

the methylation intensity, which cannot be fulfilled in this plasmid because the inserted

DNA sequence is demethylated. Given these concerns, an alternative experimental

design, such as using a rat cell line, might offer better insights that can reveal the

physiological causes.

Conclusion

In summary, the current study showed that a single CpG site at 5’UTR was persistently

hypomethylated by high fat diets. This study also identified the enhancer property of this

hypomethylated region by comparing its transcriptional activity to other DNA sequences.

It is speculated that altered DNA methylation is associated with enabled enhancer

activity, which triggers COX-2 transcription. Additional characterization and exploration

of physiological motivators of this enhancer will be conducted in future research.

Undermethylated genomic regions, specifically promoters or enhancers, have been

associated with the elevated expression of downstream genes. However, previous work

may have overlooked the roles that enhancer DNA methylation, and single CpG site

demethylation, has played in regulating gene expression. These regions may be

actively involved in other scenarios that provoke COX-2 transcription, so correcting

misconstructed epigenomes on these enhancers/ regions may efficiently reduce

transcription and subsequent physiological responses. COX-2 inhibitors have long been

used in the anti-inflammatory drug category, but produce a significant increase in heart

attacks and strokes [310]. Studies using epigenetic tools may offer a different

perspective in inhibiting overproduction of COX-2 in such diseases by correcting

transcription levels using epigenetic therapeutic strategies.

56

CHAPTER 6. TABLES AND FIGURES

Figure 3.1 Growth curve (A) and weekly caloric intake (B) of control (C/C) offspring and offspring fed a high fat diet during gestation and lactation (HF/C), postweaning (C/HF) or throughout life (HF/HF). Values are mean ± SEM, n=10 animals. Some error bars are too small to be seen.

57

Table 3.1 Body weight (BW) and caloric intake analyzed by Repeated ANOVA1, 2

BW HF/C C/HF

C/C 0.5 0.05 HF/HF 0.009* 0.11

Kcal HF/C C/HF

C/C 0.79 <0.0001* HF/HF <0.0001* 0.64

1 Comparison are made between each two treatment groups 2 *P<0.05

58

Figure 3.2. Representative images of H&E staining (upper panel) and ORO staining (lower panel) in livers of C/C, HF/C, C/HF and HF/HF offspring. Objective, 20X, Scale bar, 200µm.

59

Figure 3.3. Scattered plot of hepatocellular vacuolation counting (upper panel) and ORO evaluation (lower panel). Each symbol represents one individual sample. Horizontal lines represent mean value. n=8 individual animals. Individual treatment with different letters differ (P<0.05)

60

Figure 3.4. Hepatic (top) and plasma (bottom) TAG levels were measured in four treatment groups. Data are presented as means ± SEM, n= 6. Letters are assigned by LSD. Individual bars with different letter differ (p<0.05).

61

Figure 3.5 Expression of COX-2 mRNA, pre- mRNA and Δ6Desaturase mRNA were measured by qPCR and normalized to internal control gene β-actin. Data are presented as means ± SEM, n= 8. Letters are assigned by LSD. Individual bar with different letter differ (p<0.05).

62

Table 3.2. Gas Chromatograph analysis of fatty acid composition in liver1,2

Diet

Fatty Acid Composition CC (%) HFC (%) CHF (%) HFHF (%)

(Lauric acid)12:0 44.6±0.03 46.8±0.03 49.8±0.03 57.0±0.03

(palmitic acid)16:0 14.5±0.006 15.3±0.01 9.6±0.01 5.9±0.01

(palmitoleic acid, n-7)16:1 6.7±0.004a

3.2±0.005a,b

0.3±0.001b

0.2±0.001b

(stearic acid)18:0 4.7±0.004 4.7±0.003 3.9±0.2 3.5±0.01*

(oleic acid, n-9)18:1 16.7±0.02 13.4±0.02 15.8±0.008 15.1±0.01

(linoleic acid, n-6)18:2 8.7±0.004b

10.0±0.004a,b

12.2±0.007a

11.1±0.01b

(Arachidonic acid, n-6 )20:4 4.3±0.005b 6.4±0.003

a 8.3±0.005

a 7.3±0.1

a

1Values are means ± SEM, n=6.

2Letters were assigned to indicate significant different between treatment groups.

63

Table 3.3. Two-Way Factorial Interaction1

Maternal Postweaning Maternal * Postweaning

COX-2 mRNA <0.0001 <0.0001 0.06

COX-2 Protein 0.004 0.12 0.01

Δ6Desaturase 0.0008 0.0014 0.09

TAG 0.22 <0.0001 0.03

Vacuolation 0.39 0.0012 0.1

ORO 0.55 <0.0001 0.1

(Lauric acid)12:0 0.19 0.07 0.6

(palmitic acid)16:0 0.26 0.0003 0.17

(palmitoleic acid, n-7)16:1 0.22 0.01 0.34

(stearic acid)18:0 0.93 0.25 0.54

(oleic acid, n-9)18:1 0.4 0.81 0.59

(linoleic acid, n-6)18:2 0.53 0.13 0.02

(Arachidonic acid, n-6 )20:4 0.26 0.003 0.04 1Interaction of main effects (maternal X postweaning)

64

Figure 3.6 Expression of COX-2 protein content in the liver of male offspring by Western blot (n=6). The bands presented the average level of OCX-2 protein in control and all high fat groups. Data are presents as means ±SEM, n=6. Letters are assigned by LSD. Individual bar with different letter differ (p<0.05). Actin was used as loading control to indicate total protein quantity.

65

Figure 3.7 Correlation of Δ6Desaturase mRNA (top), COX-2 pre-mRNA (middle) and protein (bottom). Each symbol represents one individual sample. Regression line was calculated after the analysis.

66

Figure 3.8. Representative images of COX-2 protein content in livers of C/C, HF/C, C/HF and HF/HF offspring. Protein content ofCOX-2 in liver was analyzed by immunofluorescent staining using an antibody against COX-2 protein and an Alexa Fluor 647-labeled secondary antibody (red, middle panel). Nuclei were counterstained with Hoechst 33342 fluorescent stain (blue, top panel). The two pictures were merged to show the cytoplasmic location and distribution of COX-2. Objective, 63x. Scale bar, 20 μm.

67

Table 3.4. Primer sequences used in real-time PCR analysis of genes

Gene Name (Ensembl ID #) Sequences

rActb mRNA (ENSRNOE00000310180) Forward (+451), 5’- GAG ACC TTC AAC ACC CCA GC-3’ Reverse (+526), 5’- CAG TGG TAC GAC CAG AGG CA -3’

rΔ6D mRNA (ENSRNOE00000222342) Forward (+165), 5’- TGA CAG GCT CAG GAA ATG TC-3’ Reverse (+257), 5’- TTG CGA TCA ATC ACC AGC-3’

rCOX-2 mRNA (ENSRNOT00000003567) Forward (+1048), 5’- ATTACTGCTGAAGCCCACC -3’ Reverse (+1115), 5’- TGTTCCAGACTCCCTTGAAG -3’

rCOX-2 pre-mRNA

Forward (+2972), 5’- CAGGCATCGGACTCTGCTAT -3’ Reverse (+3052), 5’- AGCAGCACAATCAGGGGA -3’

rCOX-2 ChIP

Forward (-1664), 5’- CAGTCTGTGCACTCAGTTGAGATGG -3’ Reverse (-1598), 5’- TTCCGTGTATGGCACTCCCAA -3

rCOX-2 ChIP

Forward (-1320), 5’- TTGCCCAGACTGCTTCAAAC -3’ Reverse (-1217), 5’- CGTTCTTGGCATGAATGGAG -3’

rCOX-2 ChIP

Forward (-135), 5’- AACTGAGCGGAGAGCTTCAGGAG -3’ Reverse (+68), 5’- TGGCAGGACACAGAGCTGAGTTC -3

rCOX-2 ChIP

Forward (+1048), 5’- ATTACTGCTGAAGCCCACC -3’ Reverse (+1115), 5’- TGTTCCAGACTCCCTTGAAG -3

rCOX-2 ChIP

Forward (+1847), 5’- TCTCCAACCTCTCCTACTACAC -3’ Reverse (+1934), 5’- TGTACTCACCTTTCACACCC -3

rCOX-2 ChIP

Forward (+2972), 5’- CAGGCATCGGACTCTGCTAT -3’ Reverse (+3052), 5’- AGCAGCACAATCAGGGGA -3

rCOX-2 ChIP

Forward (+3967), 5’- CTCAAACAGGAGCATCCTG -3’ Reverse (+4031), 5’- ATCAGTATGAGCCTGCTGG -3

68

Figure 4.1 MeDIP-seq results of COX-2. Displayed tracks include DNA methylation (MeDIP-seq) for rat COX-2 in liver tissue from C/C, HF/C, C/HF and HF/HF offspring. The dashed box highlights the hypomethylated region at COX-2 5’ upstream in all HF treated groups. Blue line illustrates COX-2 gene structure, with each square box represents an exon. Two CpG islands are marked at the bottom at their relative location.

69

Figure 4.2 Landscape of DNA methylation patterns (A) and gene structure (B) of COX-2. A. Y axils represents the MeDIP reads, X axils represents the relative location to TSS. Individual comparison was made between each high fat group (HF/C, top; C/HF, middle, HF/HF, bottom) and control. Hypomethylated region in 5’UTR was highlighted in the dashboxed as it may closely associated with gene transcription of COX-2. B. Ten exons were labeled according to the relative location. Two CpG islands were marked with arrows in gene structure illustration. Each hypomethlated region was tagged with hyphen.

A.

B.

70

Figure 4.3. MeDIP-seq results of β-actin. Displayed tracks include DNA methylation (MeDIP-seq) for rat β-actin in liver tissue from C/C, HF/C, C/HF and HF/HF offspring. Blue line illustrates β-actin gene structure, with each square box represents an exon. Two CpG islands are marked at the bottom at their relative location.

71

Me

DIP

-se

q R

ea

din

g

Figure 4.4 Landscape of DNA methylation patterns (A) and gene structure (B) of β-actin. A. Y axils represents the MeDIP reads, X axils represents the relative location to TSS. Individual comparison was made between each high fat group (HF/C, top; C/HF, middle, HF/HF, bottom) and control. B. Six exons were labeled according to the relative location. One CpG island was marked with arrows in gene structure illustration.

72

Figure 4.5. Mapping of the histone structure of COX-2 gene in the liver of C/C (n=8) and HF/HF (n=8) groups. Acetylation of histone H4 (top), H3 (middle) and binding of HDAC3 (bottom) at upstream, promoter and coding regions of COX-2 were detected. Data are represented as the ratio to the value obtained with input DNA. The normal rabbit IgG was used as negative control to mark the non-specific association background for each tested region. Values are means ± SEM. *P<0.05

73

Figure 4.6 Mapping of the histone structure of COX-2 gene in the liver of C/C (n=8) and HF/HF (n=8) groups. Di-methylation of histone H3 Lysine 4 (top), tri-methylation of H3 Lysine 9 (middle) and Lysine 27 (bottom) at upstream, promoter and coding regions of COX-2 were detected. Data are represented as the ratio to the value obtained with input DNA. The normal rabbit IgG was used as negative control to mark the non-specific association background for each tested region. Values are means ± SEM. *P<0.05

74

Table 4.1 Antibody Information

Antibody Name

Catalog Number Source

Name Modification

COX-2 H-300 Sc-32879 Santa Cruz

H3Ac Ac-K9,14 06-599 Millipore

H4Ac Ac-K5,8,12,16 06-866 Millipore

HDAC3 H-99 sc-11417 Santa Cruz

H3K4Me2 M-K4 07-030 Millipore

H3K9Me3 M-K9 07-442 Millipore

H3K27Me3 M-K27 07-449 Millipore

IgG N/A sc-2027 Santa Cruz

75

Figure 5.1 CpG methylation within -1533 to -1109 upstream region of COX-2 coding region in liver of offspring from high fat fed groups. Distribution of methylation levels of individual CG sites within the analyzed region. *Indicates significant difference compared to the C/C group (P≤0.05).

* * *

76

Figure 5.2 Methylation statuses of the nine CpG sites. Each column corresponds to one CpG site in the studied region, and each row represents average methylation of 50 individual clones.

77

Table 5.1 Percentage of Methylation at each CpG site

CpG site# 1 2 3 4 5 6 7 8 9

CpG position -1481 -1457 -1441 -1431 -1410 -1217 -1213 -1169 -1134

Me

thy

lati

o

n (

%)

C/C 97 85.3 91.2 93.8 76.5 84.8 76.5 64.5 87.9

HF/C 96.4 82.8 89.3 92.9 58.6 72.4 86.2 53.6* 82.1

C/HF 96.3 96.3 92.6 88.9 81.5 77.8 77.8 44.4* 88.9

HF/HF 89.6 91.5 93.8 89.6 70.2 69.6 75 45.7* 78.3 *p<0.05 comparing to C/C

78

Figure 5.3 Identification of the essential elements in rat COX-2 by luciferase assay. The diagrams on the top show the firefly luciferase reporter constructs driven by various rat COX-2 fragments of -774/ +30, -1234/+30 and -1234/-1131 in Pgl3-basic. The luciferase activities from firefly were assayed and the relative luciferase activity was obtained by normalizing the firefly luciferase activity of constructed plasmids against the control plasmid luciferase activity. The results are expressed as mean ± the SEM of six replicated experiments.

79

Figure 5.4 Identification of the essential elements in rat COX-2 by luciferase assay. The diagrams on the top show the firefly luciferase reporter constructs driven by various rat COX-2 fragments of -1234/ -774, -774/-1234, -1131/-774, -774/-1131, -1234/-1131 and -1131/-1234 in Pgl3-promoter. The luciferase activities from firefly were assayed and the relative luciferase activity was obtained by normalizing the firefly luciferase activity of constructed plasmids against the control plasmid luciferase activity. The results are expressed as mean ± the SEM of six replicated experiments.

80

Figure 5.5. Luciferase activity of methylated constructs. Values of control are normalized to PGL3-p, values of +Me are normalized to PGL3-p+Me. The results are expressed as mean ± the SEM of six replicated experiments.

81

Figure 5.6. Transcription Element Search System (TESS) analysis of COX-2 enhancer. Three segments representing the left, middle and right part of this enhancer was constructed. The starting and ending nucleotide position was noted. The potential binding transcription factors are labeled at its predicted site.

82

Figure 5.7 Characterization of COX-2 enhancer by luciferase assay. The diagrams on the top show the firefly luciferase reporter constructs driven by various rat COX-2 fragments of -1234/-1199 (forward and reverse), -1204/-1160 (forward and reverse), -1163/-1131 (forward and reverse) in Pgl3-promoter. The luciferase activities from firefly were assayed and the relative luciferase activity was obtained by normalizing the firefly luciferase activity of constructed plasmids against the control plasmid luciferase activity. The results are expressed as mean ± the SEM of six replicated experiments.

83

Figure 5.8. Diagram shows the site-directed mutagenic sites of COX-2 enhancer. Three different mutations were conducted separately. From left to right: ΔHOXD8, TTA to GGG; ΔRAF, CGG to ATT; ΔCTCF, CCCT to TTTC

-1234

-1131

84

Figure 5.9 Characterization of COX-2 enhancer by luciferase assay. The diagrams on the top show the firefly luciferase reporter constructs driven by various rat COX-2 fragments that contains different mutation sites. Relative location of each mutation site is marked. The luciferase activities from firefly were assayed and the relative luciferase activity was obtained by normalizing the firefly luciferase activity of constructed plasmids against the control plasmid luciferase activity. The results are expressed as mean ± the SEM of six replicated experiments.

85

Figure 5.10 Time course of the Arachidonic acid treatment. PC3 cells were grown in RPMI-1640 medium and starved with 0.3% FBS supplemented with 1.25mg/mL fatty acid free BSA for 48 hours before treatment. Cells were then incubated with 10μg/ μL AA for 30min, 1h, 2h and 3h to induce COX-2 expression. Values are means ± SEM. *P<0.05

Hours

86

Figure 5.11 Luciferase activities of enhancer constructs (Forward, F; Reverse, R) in response to Arachidonic acid. Values are normalized to PGL3-p. The results are expressed as mean ± the SEM of six replicated experiments.

87

Table 5.2 Primer sequences used in cloning and sequencing COX-2 (ENSRNOT00000003567)

Sequences

-1234 to -1199, +SmaI top, GGG CCATTCATGCCAAGAACGTACGGTTTAATTGAATGT CCC bottom, GGG ACATTCAATTAAACCGTACGTTCTTGGCATGAATGG CCC

-1204 to -1160, +SmaI top, GGG GAATGTTTTAGTTTCCTCATTTTCTTGTTTTACTCGGTTTTTCAC CCC bottom, GGG GTGAAAAACCGAGTAAAACAAGAAAATGAGGAAACTAAAACATTC CCC

-1163 to -1131, +SmaI top, GGG CACTATTCCATCCTCAGATCCCCCTCCCGGGC CCC bottom, GGG GCCCGGGAGGGGGATCTGAGGATGGAATAGTG CCC

-1234, +SacI Forward, ATTATA GAGCTC CCATTCATGCCAAGAACGTACGGT

-1029, +SacI Reverse, ATTATA GAGCTC ATCCA CTTCT TTCTG AAACA TTGCT TGA

-906, +SacI Reverse, ATTATA GAGCTC GACTTATTTTTGTCCACGTTTACGTTTACAAG

-814, +SacI Reverse, ATTATA GAGCTC GGTATTTTCCTCCCTGCTGTTACATAGC

-1234, +NheI Forward, ATTATAGCTAGCCCATTCATGCCAAGAACGTACGGT

+30, +NheI Reverse, ATTATAGCTAGCGCGCAGAGCAGCACAGCTCGGAAGAGC

deltaRAF, cgg to att Forward, TTTAGTTTCCTCATTTTCTTGTTTTACTATTTTTTTCACTATTCCATCCTCAGATCCCC Reverse, GGGGATCTGAGGATGGAATAGTGAAAAAAATAGTAAAACAAGAAAATGAGGAAACTAAA

deltaHOXD8, tta to ggg Forward, TCCCATTCATGCCAAGAACGTACGGTGGGATTGAATGTTTTAGTTTCCTCATTTTC Reverse, GAAAATGAGGAAACTAAAACATTCAATCCCACCGTACGTTCTTGGCATGAATGGGA

deltaCTCF, ccct to tttc Forward, CTATTCCATCCTCAGATCCTTTCCCCGGGCTCGAGATCTGCG Reverse, CGCAGATCTCGAGCCCGGGGAAAGGATCTGAGGATGGAATAG

-1533, bisulfite seq Forward, AGAATTAAATTAGTATGTATATGAAGTAAA

-1109, bisulfite seq Reverse, TTCCCACTATATTTAATATTAACCC

88

CHAPTER 7. REFERENCES

1. Youngson, N.A. and E. Whitelaw, Transgenerational epigenetic effects. Annu Rev Genomics Hum Genet, 2008. 9: p. 233-57. 2. Kaati, G., et al., Transgenerational response to nutrition, early life circumstances and longevity. Eur J Hum Genet, 2007. 15(7): p. 784-90. 3. Lupu, D.S., D. Tint, and M.D. Niculescu, Perinatal epigenetic determinants of cognitive and metabolic disorders. Aging Dis, 2012. 3(6): p. 444-53. 4. Szarc vel Szic, K., et al., Nature or nurture: let food be your epigenetic medicine in chronic inflammatory disorders. Biochem Pharmacol, 2010. 80(12): p. 1816-32. 5. Braunschweig, M., et al., Investigations on transgenerational epigenetic response down the male line in F2 pigs. PLoS One, 2012. 7(2): p. e30583. 6. Li, J., et al., Accumulation of endoplasmic reticulum stress and lipogenesis in the liver through generational effects of high fat diets. J Hepatol, 2012. 56(4): p. 900-7. 7. Dunn, G.A. and T.L. Bale, Maternal high-fat diet effects on third-generation female body size via the paternal lineage. Endocrinology, 2011. 152(6): p. 2228-36. 8. Ding, Y., et al., DNA hypomethylation of inflammation-associated genes in adipose tissue of female mice after multigenerational high fat diet feeding. Int J Obes (Lond), 2014. 38(2): p. 198-204. 9. Feinberg, A.P., et al., Personalized epigenomic signatures that are stable over time and covary with body mass index. Sci.Transl.Med., 2010/9/15. 2(49): p. 49ra67-. 10. Romagnolo, D.F., A.J. Papoutsis, and O. Selmin, Nutritional targeting of cyclooxygenase-2 for colon cancer prevention. Inflamm Allergy Drug Targets, 2010. 9(3): p. 181-91. 11. Jirtle, R.L. and M.K. Skinner, Environmental epigenomics and disease susceptibility. Nat Rev Genet, 2007. 8(4): p. 253-62. 12. Szyf, M., The dynamic epigenome and its implications in toxicology. Toxicol Sci, 2007. 100(1): p. 7-23. 13. Edwards, T.M. and J.P. Myers, Environmental exposures and gene regulation in disease etiology. Environ Health Perspect, 2007. 115(9): p. 1264-70. 14. de Boo, H.A. and J.E. Harding, The developmental origins of adult disease (Barker) hypothesis. Aust.N.Z.J.Obstet.Gynaecol., 2006/2. 46(1): p. 4-14. 15. Hales, C.N. and D.J. Barker, Type 2 (non-insulin-dependent) diabetes mellitus: the thrifty phenotype hypothesis. Diabetologia., 1992/7. 35(7): p. 595-601. 16. Barker, D.J. and C. Osmond, Infant mortality, childhood nutrition, and ischaemic heart disease in England and Wales. Lancet., 1986/5/10. 1(8489): p. 1077-1081. 17. Barker, D.J., et al., The relation of small head circumference and thinness at birth to death from cardiovascular disease in adult life. BMJ., 1993/2/13. 306(6875): p. 422-426. 18. Barker, D.J., C. Osmond, and C.M. Law, The intrauterine and early postnatal origins of cardiovascular disease and chronic bronchitis. J Epidemiol Community Health, 1989. 43(3): p. 237-40. 19. Rich-Edwards, J.W., et al., Birth weight and risk of cardiovascular disease in a cohort of women followed up since 1976. BMJ., 1997/8/16. 315(7105): p. 396-400.

89

20. Frankel, S., et al., Birthweight, body-mass index in middle age, and incident coronary heart disease. Lancet., 1996/11/30. 348(9040): p. 1478-1480. 21. Stein, C.E., et al., Fetal growth and coronary heart disease in south India. Lancet., 1996/11/9. 348(9037): p. 1269-1273. 22. Osmond, C., et al., Early growth and death from cardiovascular disease in women. BMJ., 1993/12/11. 307(6918): p. 1519-1524. 23. Hales, C.N. and D.J. Barker, The thrifty phenotype hypothesis. Br Med Bull, 2001. 60: p. 5-20. 24. Portrait, F., E. Teeuwiszen, and D. Deeg, Early life undernutrition and chronic diseases at older ages: the effects of the Dutch famine on cardiovascular diseases and diabetes. Soc.Sci.Med., 2011/9. 73(5): p. 711-718. 25. Elias, S.G., et al., The 1944-1945 Dutch famine and subsequent overall cancer incidence. Cancer Epidemiol.Biomarkers Prev., 2005/8. 14(8): p. 1981-1985. 26. Painter, R.C., T.J. Roseboom, and O.P. Bleker, Prenatal exposure to the Dutch famine and disease in later life: an overview. Reprod.Toxicol., 2005/9. 20(3): p. 345-352. 27. Widdowson, E.M. and R.A. McCance, A review: new thoughts on growth. Pediatr.Res., 1975/3. 9(3): p. 154-156. 28. Desai, M., et al., Organ-selective growth in the offspring of protein-restricted mothers. Br.J.Nutr., 1996/10. 76(4): p. 591-603. 29. Godfrey, K., et al., Maternal nutrition in early and late pregnancy in relation to placental and fetal growth. BMJ., 1996/2/17. 312(7028): p. 410-414. 30. Moore, V.M., et al., Dietary composition of pregnant women is related to size of the baby at birth. J.Nutr., 2004/7. 134(7): p. 1820-1826. 31. Woodall, S.M., et al., Chronic maternal undernutrition in the rat leads to delayed postnatal growth and elevated blood pressure of offspring. Pediatr.Res., 1996/9. 40(3): p. 438-443. 32. Langley-Evans, S.C., et al., Weanling rats exposed to maternal low-protein diets during discrete periods of gestation exhibit differing severity of hypertension. Clin.Sci.(Lond). 1996/11. 91(5): p. 607-615. 33. Kind, K.L., et al., Effect of maternal feed restriction on blood pressure in the adult guinea pig. Exp.Physiol., 2002/7. 87(4): p. 469-477. 34. Gopalakrishnan, G.S., et al., Programming of adult cardiovascular function after early maternal undernutrition in sheep. Am.J.Physiol Regul.Integr.Comp Physiol., 2004/7. 287(1): p. R12-R20. 35. Ozanne, S.E. and C.N. Hales, The long-term consequences of intra-uterine protein malnutrition for glucose metabolism. Proc.Nutr.Soc., 1999/8. 58(3): p. 615-619. 36. Kind, K.L., et al., Effect of maternal feed restriction during pregnancy on glucose tolerance in the adult guinea pig. Am.J.Physiol Regul.Integr.Comp Physiol., 2003/1. 284(1): p. R140-R152. 37. Gardner, D.S., et al., Programming of glucose-insulin metabolism in adult sheep after maternal undernutrition. Am.J.Physiol Regul.Integr.Comp Physiol., 2005/10. 289(4): p. R947-R954. 38. Rao, S., et al., Intake of micronutrient-rich foods in rural Indian mothers is associated with the size of their babies at birth: Pune Maternal Nutrition Study. J.Nutr., 2001/4. 131(4): p. 1217-1224.

90

39. Yajnik, C.S., Early life origins of insulin resistance and type 2 diabetes in India and other Asian countries. J.Nutr., 2004/1. 134(1): p. 205-210. 40. Gillman, M.W., et al., Maternal calcium intake and offspring blood pressure. Circulation., 2004/10/5. 110(14): p. 1990-1995. 41. Belizan, J.M., et al., Long-term effect of calcium supplementation during pregnancy on the blood pressure of offspring: follow up of a randomised controlled trial. BMJ., 1997/8/2. 315(7103): p. 281-285. 42. Van Assche, F.A., K. Holemans, and L. Aerts, Long-term consequences for offspring of diabetes during pregnancy. Br Med Bull, 2001. 60: p. 173-82. 43. Petry, C.J., S.E. Ozanne, and C.N. Hales, Programming of intermediary metabolism. Mol Cell Endocrinol, 2001. 185(1-2): p. 81-91. 44. Holemans, K., L. Aerts, and F.A. Van Assche, Fetal growth and long-term consequences in animal models of growth retardation. Eur J Obstet Gynecol Reprod Biol, 1998. 81(2): p. 149-56. 45. Symonds, M.E., S.P. Sebert, and H. Budge, The impact of diet during early life and its contribution to later disease: critical checkpoints in development and their long-term consequences for metabolic health. Proc Nutr Soc, 2009. 68(4): p. 416-21. 46. Tschop, M. and M.L. Heiman, Rodent obesity models: an overview. Exp Clin Endocrinol Diabetes, 2001. 109(6): p. 307-19. 47. Oakes, N.D., et al., Mechanisms of liver and muscle insulin resistance induced by chronic high-fat feeding. Diabetes, 1997. 46(11): p. 1768-74. 48. Buettner, R., J. Scholmerich, and L.C. Bollheimer, High-fat diets: modeling the metabolic disorders of human obesity in rodents. Obesity (Silver Spring), 2007. 15(4): p. 798-808. 49. Napoli, C., et al., Maternal hypercholesterolemia during pregnancy promotes early atherogenesis in LDL receptor-deficient mice and alters aortic gene expression determined by microarray. Circulation, 2002. 105(11): p. 1360-7. 50. Srinivasan, M., et al., Maternal high-fat diet consumption results in fetal malprogramming predisposing to the onset of metabolic syndrome-like phenotype in adulthood. Am J Physiol Endocrinol Metab, 2006. 291(4): p. E792-9. 51. Elahi, M.M., et al., Long-term maternal high-fat feeding from weaning through pregnancy and lactation predisposes offspring to hypertension, raised plasma lipids and fatty liver in mice. Br J Nutr, 2009. 102(4): p. 514-9. 52. Shankar, K., et al., Maternal obesity at conception programs obesity in the offspring. Am J Physiol Regul Integr Comp Physiol, 2008. 294(2): p. R528-38. 53. Plagemann, A., et al., Perinatal elevation of hypothalamic insulin, acquired malformation of hypothalamic galaninergic neurons, and syndrome x-like alterations in adulthood of neonatally overfed rats. Brain Res, 1999. 836(1-2): p. 146-55. 54. Levin, B.E. and E. Govek, Gestational obesity accentuates obesity in obesity-prone progeny. Am J Physiol, 1998. 275(4 Pt 2): p. R1374-9. 55. Karnik, H.B., et al., High dietary fat feeding during perinatal development of rats alters hepatic drug metabolism of progeny. Dev Pharmacol Ther, 1989. 14(2): p. 135-40. 56. Ferrero-Miliani, L., et al., Chronic inflammation: importance of NOD2 and NALP3 in interleukin-1beta generation. Clin Exp Immunol, 2007. 147(2): p. 227-35.

91

57. Andreozzi, P., et al., [Predictors of liver fibrosis in patients with non-alcoholic fatty liver disease. The role of metabolic syndrome, insulin-resistance and inflammation]. Recenti Prog Med, 2012. 103(12): p. 570-4. 58. Taube, A., et al., Inflammation and metabolic dysfunction: links to cardiovascular diseases. Am J Physiol Heart Circ Physiol, 2012. 302(11): p. H2148-65. 59. Manabe, I., Chronic inflammation links cardiovascular, metabolic and renal diseases. Circ J, 2011. 75(12): p. 2739-48. 60. Sartori, A.C., et al., The impact of inflammation on cognitive function in older adults: implications for healthcare practice and research. J Neurosci Nurs, 2012. 44(4): p. 206-17. 61. Jenny, N.S., Inflammation in aging: cause, effect, or both? Discov Med, 2012. 13(73): p. 451-60. 62. Romeo, G.R., J. Lee, and S.E. Shoelson, Metabolic syndrome, insulin resistance, and roles of inflammation--mechanisms and therapeutic targets. Arterioscler Thromb Vasc Biol, 2012. 32(8): p. 1771-6. 63. Chaturvedi, A.K., S.C. Moore, and A. Hildesheim, Invited commentary: circulating inflammation markers and cancer risk--implications for epidemiologic studies. Am J Epidemiol, 2013. 177(1): p. 14-9. 64. Chen, H., et al., Maternal and postnatal overnutrition differentially impact appetite regulators and fuel metabolism. Endocrinology, 2008. 149(11): p. 5348-56. 65. McCurdy, C.E., et al., Maternal high-fat diet triggers lipotoxicity in the fetal livers of nonhuman primates. J Clin Invest, 2009. 119(2): p. 323-35. 66. Samuelsson, A.M., et al., Diet-induced obesity in female mice leads to offspring hyperphagia, adiposity, hypertension, and insulin resistance: a novel murine model of developmental programming. Hypertension, 2008. 51(2): p. 383-92. 67. Bilbo, S.D. and V. Tsang, Enduring consequences of maternal obesity for brain inflammation and behavior of offspring. FASEB J, 2010. 24(6): p. 2104-15. 68. Tong, J.F., et al., Maternal obesity downregulates myogenesis and beta-catenin signaling in fetal skeletal muscle. Am J Physiol Endocrinol Metab, 2009. 296(4): p. E917-24. 69. Bayol, S.A., et al., Offspring from mothers fed a 'junk food' diet in pregnancy and lactation exhibit exacerbated adiposity that is more pronounced in females. J Physiol, 2008. 586(13): p. 3219-30. 70. Kubaszek, A., et al., Promoter polymorphisms of the TNF-alpha (G-308A) and IL-6 (C-174G) genes predict the conversion from impaired glucose tolerance to type 2 diabetes: the Finnish Diabetes Prevention Study. Diabetes, 2003. 52(7): p. 1872-6. 71. Yudkin, J.S., Adipose tissue, insulin action and vascular disease: inflammatory signals. Int J Obes Relat Metab Disord, 2003. 27 Suppl 3: p. S25-8. 72. Kirwan, J.P., et al., TNF-alpha is a predictor of insulin resistance in human pregnancy. Diabetes, 2002. 51(7): p. 2207-13. 73. Sell, H. and J. Eckel, Monocyte chemotactic protein-1 and its role in insulin resistance. Curr Opin Lipidol, 2007. 18(3): p. 258-62. 74. Wolsk, E., et al., IL-6 selectively stimulates fat metabolism in human skeletal muscle. Am J Physiol Endocrinol Metab, 2010. 299(5): p. E832-40.

92

75. Carey, A.L., et al., Interleukin-6 increases insulin-stimulated glucose disposal in humans and glucose uptake and fatty acid oxidation in vitro via AMP-activated protein kinase. Diabetes, 2006. 55(10): p. 2688-97. 76. Hotamisligil, G.S., et al., Increased adipose tissue expression of tumor necrosis factor-alpha in human obesity and insulin resistance. J Clin Invest, 1995. 95(5): p. 2409-15. 77. Sell, H., et al., Skeletal muscle insulin resistance induced by adipocyte-conditioned medium: underlying mechanisms and reversibility. Am J Physiol Endocrinol Metab, 2008. 294(6): p. E1070-7. 78. Franco, J.G., et al., Maternal high-fat diet induces obesity and adrenal and thyroid dysfunction in male rat offspring at weaning. J Physiol, 2012. 590(Pt 21): p. 5503-18. 79. White, C.L., et al., Effects of high fat diet on Morris maze performance, oxidative stress, and inflammation in rats: contributions of maternal diet. Neurobiol Dis, 2009. 35(1): p. 3-13. 80. Strakovsky, R.S., et al., Gestational high fat diet programs hepatic phosphoenolpyruvate carboxykinase gene expression and histone modification in neonatal offspring rats. J Physiol, 2011. 589(Pt 11): p. 2707-17. 81. Taylor, P.D., et al., Impaired glucose homeostasis and mitochondrial abnormalities in offspring of rats fed a fat-rich diet in pregnancy. Am J Physiol Regul Integr Comp Physiol, 2005. 288(1): p. R134-9. 82. Sen, S. and R.A. Simmons, Maternal antioxidant supplementation prevents adiposity in the offspring of Western diet-fed rats. Diabetes, 2010. 59(12): p. 3058-65. 83. Grant, W.F., et al., Maternal high fat diet is associated with decreased plasma n-3 fatty acids and fetal hepatic apoptosis in nonhuman primates. PLoS One, 2011. 6(2): p. e17261. 84. Bruce, K.D., et al., Maternal high-fat feeding primes steatohepatitis in adult mice offspring, involving mitochondrial dysfunction and altered lipogenesis gene expression. Hepatology, 2009. 50(6): p. 1796-808. 85. van Bussel, B.C., et al., Endothelial dysfunction and low-grade inflammation are associated with greater arterial stiffness over a 6-year period. Hypertension, 2011. 58(4): p. 588-95. 86. van Bussel, B.C., et al., Low-grade inflammation, but not endothelial dysfunction, is associated with greater carotid stiffness in the elderly: the Hoorn Study. J Hypertens, 2012. 30(4): p. 744-52. 87. Thompson, A.M., et al., Association of inflammation and endothelial dysfunction with metabolic syndrome, prediabetes and diabetes in adults from Inner Mongolia, China. BMC Endocr Disord, 2011. 11: p. 16. 88. Li, H., et al., Co-effect of insulin resistance and biomarkers of inflammation and endothelial dysfunction on hypertension. Hypertens Res, 2012. 35(5): p. 513-7. 89. Mateos-Caceres, P.J., et al., New and old mechanisms associated with hypertension in the elderly. Int J Hypertens, 2012. 2012: p. 150107. 90. Kes, P., et al., [The role of arterial hypertension in developement of chronic renal failure]. Acta Med Croatica, 2011. 65 Suppl 3: p. 78-84. 91. Libby, P., Inflammation in atherosclerosis. Arterioscler Thromb Vasc Biol, 2012. 32(9): p. 2045-51.

93

92. Taylor, P.D., et al., Uterine artery function in pregnant rats fed a diet supplemented with animal lard. Exp Physiol, 2003. 88(3): p. 389-98. 93. Khan, I.Y., et al., Gender-linked hypertension in offspring of lard-fed pregnant rats. Hypertension, 2003. 41(1): p. 168-75. 94. Armitage, J.A., et al., Developmental programming of aortic and renal structure in offspring of rats fed fat-rich diets in pregnancy. J Physiol, 2005. 565(Pt 1): p. 171-84. 95. Khan, I.Y., et al., A high-fat diet during rat pregnancy or suckling induces cardiovascular dysfunction in adult offspring. Am J Physiol Regul Integr Comp Physiol, 2005. 288(1): p. R127-33. 96. Palinski, W., et al., Maternal hypercholesterolemia and treatment during pregnancy influence the long-term progression of atherosclerosis in offspring of rabbits. Circ Res, 2001. 89(11): p. 991-6. 97. Taylor, P.D., et al., Impaired EDHF-mediated vasodilatation in adult offspring of rats exposed to a fat-rich diet in pregnancy. J Physiol, 2004. 558(Pt 3): p. 943-51. 98. Liang, C., M.E. Oest, and M.R. Prater, Intrauterine exposure to high saturated fat diet elevates risk of adult-onset chronic diseases in C57BL/6 mice. Birth Defects Res B Dev Reprod Toxicol, 2009. 86(5): p. 377-84. 99. Ashino, N.G., et al., Maternal high-fat feeding through pregnancy and lactation predisposes mouse offspring to molecular insulin resistance and fatty liver. J Nutr Biochem, 2012. 23(4): p. 341-8. 100. Fan, L., et al., Maternal high-fat diet impacts endothelial function in nonhuman primate offspring. Int J Obes (Lond), 2012. 101. Deng, Y.Y., et al., Role of microglia in the process of inflammation in the hypoxic developing brain. Front Biosci (Schol Ed), 2011. 3: p. 884-900. 102. Butterworth, R.F., Neuroinflammation in acute liver failure: mechanisms and novel therapeutic targets. Neurochem Int, 2011. 59(6): p. 830-6. 103. Sasaki, A., et al., Perinatal high fat diet alters glucocorticoid signaling and anxiety behavior in adulthood. Neuroscience, 2013. 240: p. 1-12. 104. Aaltonen, R., et al., Transfer of proinflammatory cytokines across term placenta. Obstet Gynecol, 2005. 106(4): p. 802-7. 105. Ashdown, H., et al., The role of cytokines in mediating effects of prenatal infection on the fetus: implications for schizophrenia. Mol Psychiatry, 2006. 11(1): p. 47-55. 106. Galea, J. and D. Brough, The role of inflammation and interleukin-1 in acute cerebrovascular disease. J Inflamm Res, 2013. 6: p. 121-128. 107. Theoharides, T.C., S. Asadi, and A.B. Patel, Focal brain inflammation and autism. J Neuroinflammation, 2013. 10: p. 46. 108. Chiang, C.S., et al., Reactive gliosis as a consequence of interleukin-6 expression in the brain: studies in transgenic mice. Dev Neurosci, 1994. 16(3-4): p. 212-21. 109. Elovitz, M.A., et al., Intrauterine inflammation, insufficient to induce parturition, still evokes fetal and neonatal brain injury. Int J Dev Neurosci, 2011. 29(6): p. 663-71. 110. Samuelsson, A.M., et al., Prenatal exposure to interleukin-6 results in inflammatory neurodegeneration in hippocampus with NMDA/GABA(A) dysregulation and impaired spatial learning. Am J Physiol Regul Integr Comp Physiol, 2006. 290(5): p. R1345-56.

94

111. Samuelsson, A.M., et al., Prenatal exposure to interleukin-6 results in hypertension and increased hypothalamic-pituitary-adrenal axis activity in adult rats. Endocrinology, 2004. 145(11): p. 4897-911. 112. Vallieres, L., et al., Reduced hippocampal neurogenesis in adult transgenic mice with chronic astrocytic production of interleukin-6. J Neurosci, 2002. 22(2): p. 486-92. 113. Sullivan, E.L., L. Nousen, and K. Chamlou, Maternal high fat diet consumption during the perinatal period programs offspring behavior. Physiol Behav, 2012. 114. Sullivan, E.L., et al., Chronic consumption of a high-fat diet during pregnancy causes perturbations in the serotonergic system and increased anxiety-like behavior in nonhuman primate offspring. J Neurosci, 2010. 30(10): p. 3826-30. 115. Oben, J.A., et al., Maternal obesity during pregnancy and lactation programs the development of offspring non-alcoholic fatty liver disease in mice. J Hepatol, 2010. 52(6): p. 913-20. 116. Zhang, X., et al., A maternal high-fat diet represses the expression of antioxidant defense genes and induces the cellular senescence pathway in the liver of male offspring rats. J Nutr, 2011. 141(7): p. 1254-9. 117. Endo, M., et al., TNF-alpha induces hepatic steatosis in mice by enhancing gene expression of sterol regulatory element binding protein-1c (SREBP-1c). Exp Biol Med (Maywood), 2007. 232(5): p. 614-21. 118. Bugianesi, E., A.J. McCullough, and G. Marchesini, Insulin resistance: a metabolic pathway to chronic liver disease. Hepatology, 2005. 42(5): p. 987-1000. 119. Cintra, D.E., et al., Interleukin-10 is a protective factor against diet-induced insulin resistance in liver. J Hepatol, 2008. 48(4): p. 628-37. 120. Valera, A., et al., Transgenic mice overexpressing phosphoenolpyruvate carboxykinase develop non-insulin-dependent diabetes mellitus. Proc Natl Acad Sci U S A, 1994. 91(19): p. 9151-4. 121. Ghebremeskel, K., et al., Maternal diet high in fat reduces docosahexaenoic acid in liver lipids of newborn and sucking rat pups. Br J Nutr, 1999. 81(5): p. 395-404. 122. Albert, C.M., et al., Fish consumption and risk of sudden cardiac death. JAMA, 1998. 279(1): p. 23-8. 123. Lands, W.E., Biosynthesis of prostaglandins. Annu Rev Nutr, 1991. 11: p. 41-60. 124. Roberts, D.J. and M.D. Post, The placenta in pre-eclampsia and intrauterine growth restriction. J Clin Pathol, 2008. 61(12): p. 1254-60. 125. Salafia, C.M., et al., Preterm delivery: correlations of fetal growth and placental pathology. Am J Perinatol, 1992. 9(3): p. 190-3. 126. Kidron, D., J. Bernheim, and R. Aviram, Placental findings contributing to fetal death, a study of 120 stillbirths between 23 and 40 weeks gestation. Placenta, 2009. 30(8): p. 700-4. 127. Liang, C., K. DeCourcy, and M.R. Prater, High-saturated-fat diet induces gestational diabetes and placental vasculopathy in C57BL/6 mice. Metabolism, 2010. 59(7): p. 943-50. 128. Frias, A.E., et al., Maternal high-fat diet disturbs uteroplacental hemodynamics and increases the frequency of stillbirth in a nonhuman primate model of excess nutrition. Endocrinology, 2011. 152(6): p. 2456-64. 129. Lockwood, C.J., et al., Decidual hemostasis, inflammation, and angiogenesis in pre-eclampsia. Semin Thromb Hemost, 2011. 37(2): p. 158-64.

95

130. Li, L., J. Kang, and W. Lei, Role of Toll-like receptor 4 in inflammation-induced preterm delivery. Mol Hum Reprod, 2010. 16(4): p. 267-72. 131. Abdulkadir, A.A., et al., Placental inflammation and fetal hemodynamics in a rat model of chorioamnionitis. Pediatr Res, 2010. 68(6): p. 513-8. 132. Sun, M.N., Z. Yang, and R.Q. Ma, Effect of high-fat diet on liver and placenta fatty infiltration in early onset preeclampsia-like mouse model. Chin Med J (Engl), 2012. 125(19): p. 3532-8. 133. Grayson, B.E., et al., Changes in melanocortin expression and inflammatory pathways in fetal offspring of nonhuman primates fed a high-fat diet. Endocrinology, 2010. 151(4): p. 1622-32. 134. Zaretsky, M.V., et al., Transfer of inflammatory cytokines across the placenta. Obstet Gynecol, 2004. 103(3): p. 546-50. 135. Assadi, F., Neonatal nephrotic syndrome associated with placental transmission of proinflammatory cytokines. Pediatr Nephrol, 2011. 26(3): p. 469-71. 136. Barker, D.J., et al., Fetal origins of adult disease: strength of effects and biological basis. Int J Epidemiol, 2002. 31(6): p. 1235-9. 137. Ptashne, M., On the use of the word 'epigenetic'. Curr Biol, 2007. 17(7): p. R233-6. 138. Costello, J.F. and C. Plass, Methylation matters. J Med Genet, 2001. 38(5): p. 285-303. 139. Takai, D. and P.A. Jones, Comprehensive analysis of CpG islands in human chromosomes 21 and 22. Proc Natl Acad Sci U S A, 2002. 99(6): p. 3740-5. 140. Wu, S.C. and Y. Zhang, Active DNA demethylation: many roads lead to Rome. Nat Rev Mol Cell Biol, 2010. 11(9): p. 607-20. 141. Jones, P.A. and D. Takai, The role of DNA methylation in mammalian epigenetics. Science, 2001. 293(5532): p. 1068-70. 142. Brueckner, B. and F. Lyko, DNA methyltransferase inhibitors: old and new drugs for an epigenetic cancer therapy. Trends Pharmacol Sci, 2004. 25(11): p. 551-4. 143. Jenuwein, T. and C.D. Allis, Translating the histone code. Science, 2001. 293(5532): p. 1074-80. 144. Hake, S.B., A. Xiao, and C.D. Allis, Linking the epigenetic 'language' of covalent histone modifications to cancer. Br J Cancer, 2004. 90(4): p. 761-9. 145. Lachner, M. and T. Jenuwein, The many faces of histone lysine methylation. Curr Opin Cell Biol, 2002. 14(3): p. 286-98. 146. Dobosy, J.R. and E.U. Selker, Emerging connections between DNA methylation and histone acetylation. Cell Mol.Life Sci., 2001/5. 58(5-6): p. 721-727. 147. Bender, C.M., J.M. Zingg, and P.A. Jones, DNA methylation as a target for drug design. Pharm.Res., 1998/2. 15(2): p. 175-187. 148. Egger, G., et al., Epigenetics in human disease and prospects for epigenetic therapy. Nature, 2004. 429(6990): p. 457-63. 149. Prokhortchouk, E. and B. Hendrich, Methyl-CpG binding proteins and cancer: are MeCpGs more important than MBDs? Oncogene., 2002/8/12. 21(35): p. 5394-5399. 150. Burgers, W.A., F. Fuks, and T. Kouzarides, DNA methyltransferases get connected to chromatin. Trends Genet, 2002. 18(6): p. 275-7. 151. Tamaru, H. and E.U. Selker, A histone H3 methyltransferase controls DNA methylation in Neurospora crassa. Nature., 2001/11/15. 414(6861): p. 277-283.

96

152. Selker, E.U., Trichostatin A causes selective loss of DNA methylation in Neurospora. Proc.Natl.Acad.Sci.U.S.A., 1998/8/4. 95(16): p. 9430-9435. 153. Faulk, C. and D.C. Dolinoy, Timing is everything: the when and how of environmentally induced changes in the epigenome of animals. Epigenetics, 2011. 6(7): p. 791-7. 154. Ohm, J.E., et al., A stem cell-like chromatin pattern may predispose tumor suppressor genes to DNA hypermethylation and heritable silencing. Nat Genet, 2007. 39(2): p. 237-42. 155. Schlesinger, Y., et al., Polycomb-mediated methylation on Lys27 of histone H3 pre-marks genes for de novo methylation in cancer. Nat Genet, 2007. 39(2): p. 232-6. 156. Teschendorff, A.E., et al., Age-dependent DNA methylation of genes that are suppressed in stem cells is a hallmark of cancer. Genome Res, 2010. 20(4): p. 440-6. 157. Widschwendter, M., et al., Epigenetic stem cell signature in cancer. Nat Genet, 2007. 39(2): p. 157-8. 158. McKay, J.A., et al., Folate depletion during pregnancy and lactation reduces genomic DNA methylation in murine adult offspring. Genes Nutr, 2011. 6(2): p. 189-96. 159. Gluckman, P.D., et al., Metabolic plasticity during mammalian development is directionally dependent on early nutritional status. Proc Natl Acad Sci U S A, 2007. 104(31): p. 12796-800. 160. Heijmans, B.T., et al., Persistent epigenetic differences associated with prenatal exposure to famine in humans. Proc Natl Acad Sci U S A, 2008. 105(44): p. 17046-9. 161. Bollati, V. and A. Baccarelli, Environmental epigenetics. Heredity (Edinb), 2010. 105(1): p. 105-12. 162. Skinner, M.K., M. Manikkam, and C. Guerrero-Bosagna, Epigenetic transgenerational actions of environmental factors in disease etiology. Trends Endocrinol Metab, 2010. 21(4): p. 214-22. 163. Manikkam, M., et al., Transgenerational actions of environmental compounds on reproductive disease and identification of epigenetic biomarkers of ancestral exposures. PLoS One, 2012. 7(2): p. e31901. 164. Feinberg, A.P., et al., Personalized epigenomic signatures that are stable over time and covary with body mass index. Sci Transl Med, 2010. 2(49): p. 49ra67. 165. Gibson, G., Rare and common variants: twenty arguments. Nat Rev Genet, 2011. 13(2): p. 135-45. 166. Hartman, M., et al., Molecular epidemiology and its current clinical use in cancer management. Lancet Oncol, 2010. 11(4): p. 383-90. 167. Sivakumaran, S., et al., Abundant pleiotropy in human complex diseases and traits. Am J Hum Genet, 2011. 89(5): p. 607-18. 168. Feil, R. and M.F. Fraga, Epigenetics and the environment: emerging patterns and implications. Nat Rev Genet, 2011. 13(2): p. 97-109. 169. Petronis, A., Epigenetics as a unifying principle in the aetiology of complex traits and diseases. Nature, 2010. 465(7299): p. 721-7. 170. Crews, D., Epigenetics, brain, behavior, and the environment. Hormones (Athens), 2010. 9(1): p. 41-50. 171. Goldberg, A.D., C.D. Allis, and E. Bernstein, Epigenetics: a landscape takes shape. Cell, 2007. 128(4): p. 635-8.

97

172. Bell, J.T. and T.D. Spector, A twin approach to unraveling epigenetics. Trends Genet, 2011. 27(3): p. 116-25. 173. Foley, D.L., et al., Prospects for epigenetic epidemiology. Am J Epidemiol, 2009. 169(4): p. 389-400. 174. Fraga, M.F., et al., Epigenetic differences arise during the lifetime of monozygotic twins. Proc Natl Acad Sci U S A, 2005. 102(30): p. 10604-9. 175. McGowan, P.O., et al., Epigenetic regulation of the glucocorticoid receptor in human brain associates with childhood abuse. Nat Neurosci, 2009. 12(3): p. 342-8. 176. Szyf, M. and J. Bick, DNA methylation: a mechanism for embedding early life experiences in the genome. Child Dev, 2013. 84(1): p. 49-57. 177. Ho, S.M., et al., Environmental epigenetics and its implication on disease risk and health outcomes. ILAR J, 2012. 53(3-4): p. 289-305. 178. Irigaray, P., et al., Lifestyle-related factors and environmental agents causing cancer: an overview. Biomed Pharmacother, 2007. 61(10): p. 640-58. 179. Lorenzen, J.M., F. Martino, and T. Thum, Epigenetic modifications in cardiovascular disease. Basic Res Cardiol, 2012. 107(2): p. 245. 180. Mathers, J.C., G. Strathdee, and C.L. Relton, Induction of epigenetic alterations by dietary and other environmental factors. Adv Genet, 2010. 71: p. 3-39. 181. Nise, M.S., P. Falaturi, and T.C. Erren, Epigenetics: origins and implications for cancer epidemiology. Med Hypotheses, 2010. 74(2): p. 377-82. 182. Bjornsson, H.T., M.D. Fallin, and A.P. Feinberg, An integrated epigenetic and genetic approach to common human disease. Trends Genet, 2004. 20(8): p. 350-8. 183. Jones, P.A. and S.B. Baylin, The fundamental role of epigenetic events in cancer. Nat Rev Genet, 2002. 3(6): p. 415-28. 184. Campion, J., F.I. Milagro, and J.A. Martinez, Individuality and epigenetics in obesity. Obes Rev, 2009. 10(4): p. 383-92. 185. Reddy, M.A. and R. Natarajan, Epigenetic mechanisms in diabetic vascular complications. Cardiovasc Res, 2011. 90(3): p. 421-9. 186. Wierda, R.J., et al., Epigenetics in atherosclerosis and inflammation. J Cell Mol Med, 2010. 14(6A): p. 1225-40. 187. Bruce, K.D. and M.A. Hanson, The developmental origins, mechanisms, and implications of metabolic syndrome. J Nutr, 2010. 140(3): p. 648-52. 188. Cortessis, V.K., et al., Environmental epigenetics: prospects for studying epigenetic mediation of exposure-response relationships. Hum Genet, 2012. 131(10): p. 1565-89. 189. Backdahl, L., A. Bushell, and S. Beck, Inflammatory signalling as mediator of epigenetic modulation in tissue-specific chronic inflammation. Int J Biochem Cell Biol, 2009. 41(1): p. 176-84. 190. Villagra, A., E.M. Sotomayor, and E. Seto, Histone deacetylases and the immunological network: implications in cancer and inflammation. Oncogene, 2010. 29(2): p. 157-73. 191. Enya, K., et al., The interaction with Sp1 and reduction in the activity of histone deacetylase 1 are critical for the constitutive gene expression of IL-1 alpha in human melanoma cells. J Leukoc Biol, 2008. 83(1): p. 190-9. 192. Lu, J., et al., Interleukin-12 p40 promoter activity is regulated by the reversible acetylation mediated by HDAC1 and p300. Cytokine, 2005. 31(1): p. 46-51.

98

193. Gopal, Y.N. and M.W. Van Dyke, Depletion of histone deacetylase protein: a common consequence of inflammatory cytokine signaling? Cell Cycle, 2006. 5(23): p. 2738-43. 194. Aung, H.T., et al., LPS regulates proinflammatory gene expression in macrophages by altering histone deacetylase expression. FASEB J, 2006. 20(9): p. 1315-27. 195. Shanmugam, M.K. and G. Sethi, Role of epigenetics in inflammation-associated diseases. Subcell Biochem, 2012. 61: p. 627-57. 196. Faraco, G., et al., Histone deacetylase (HDAC) inhibitors reduce the glial inflammatory response in vitro and in vivo. Neurobiol Dis, 2009. 36(2): p. 269-79. 197. Kim, M.J., et al., Gallic acid, a histone acetyltransferase inhibitor, suppresses beta-amyloid neurotoxicity by inhibiting microglial-mediated neuroinflammation. Mol Nutr Food Res, 2011. 55(12): p. 1798-808. 198. Tian, W., et al., Brahma related gene 1 (Brg1) bridges epigenetic regulation of pro-inflammatory cytokine production to steatohepatitis. Hepatology, 2012. 199. Neves-Costa, A. and L.F. Moita, TET1 is a negative transcriptional regulator of IL-1beta in the THP-1 cell line. Mol Immunol, 2013. 54(3-4): p. 264-270. 200. Zheng, S., et al., Histone deacetylase 3 (HDAC3) participates in the transcriptional repression of the p16 (INK4a) gene in mammary gland of the female rat offspring exposed to an early-life high-fat diet. Epigenetics, 2012. 7(2): p. 183-90. 201. Peedicayil, J., Epigenetic therapy--a new development in pharmacology. Indian J Med Res, 2006. 123(1): p. 17-24. 202. Issa, J.P., Epigenetic variation and human disease. J Nutr, 2002. 132(8 Suppl): p. 2388S-2392S. 203. Zelent, A., et al., State of the translational science: summary of Baltimore workshop on gene re-expression as a therapeutic target in cancer January 2003. Clin Cancer Res, 2004. 10(14): p. 4622-9. 204. Miyamoto, K. and T. Ushijima, Diagnostic and therapeutic applications of epigenetics. Jpn J Clin Oncol, 2005. 35(6): p. 293-301. 205. Gilbert, J., et al., The clinical application of targeting cancer through histone acetylation and hypomethylation. Clin Cancer Res, 2004. 10(14): p. 4589-96. 206. Sato, N., et al., Frequent hypomethylation of multiple genes overexpressed in pancreatic ductal adenocarcinoma. Cancer Res, 2003. 63(14): p. 4158-66. 207. Hughes, L.A., et al., Early life exposure to famine and colorectal cancer risk: a role for epigenetic mechanisms. PLoS One, 2009. 4(11): p. e7951. 208. Kamakura, M., Royalactin induces queen differentiation in honeybees. Nature, 2011. 473(7348): p. 478-83. 209. Dolinoy, D.C., et al., Maternal genistein alters coat color and protects Avy mouse offspring from obesity by modifying the fetal epigenome. Environ Health Perspect, 2006. 114(4): p. 567-72. 210. Kucharski, R., et al., Nutritional control of reproductive status in honeybees via DNA methylation. Science, 2008. 319(5871): p. 1827-30. 211. Zhou, H.R., et al., Folate polyglutamylation is involved in chromatin silencing by maintaining global DNA methylation and histone H3K9 dimethylation in Arabidopsis. Plant Cell., 2013/7. 25(7): p. 2545-2559.

99

212. McKay, J.A., et al., Folate depletion during pregnancy and lactation reduces genomic DNA methylation in murine adult offspring. Genes Nutr., 2011/5. 6(2): p. 189-196. 213. McKay, J.A., E.A. Williams, and J.C. Mathers, Effect of maternal and post-weaning folate supply on gene-specific DNA methylation in the small intestine of weaning and adult apc and wild type mice. Front Genet., 2011/5/23. 2:23. doi: 10.3389/fgene.2011.00023. eCollection;%2011.: p. 23-. 214. Zhang, L., et al., Exogenous plant MIR168a specifically targets mammalian LDLRAP1: evidence of cross-kingdom regulation by microRNA. Cell Res, 2012. 22(1): p. 107-26. 215. Pollitt, E., et al., Nutrition in early life and the fulfillment of intellectual potential. J Nutr, 1995. 125(4 Suppl): p. 1111S-1118S. 216. Zimmermann, M.B., Iodine deficiency in pregnancy and the effects of maternal iodine supplementation on the offspring: a review. (1938-3207 (Electronic)). 217. Muhlhausler, B.S. and Z.Y. Ong, The fetal origins of obesity: early origins of altered food intake. Endocr Metab Immune Disord Drug Targets, 2011. 11(3): p. 189-97. 218. Lillycrop, K.A., Effect of maternal diet on the epigenome: implications for human metabolic disease. Proc Nutr Soc, 2011. 70(1): p. 64-72. 219. Pan, S., et al., MicroRNA-130b and microRNA-374b mediate the effect of maternal dietary protein on offspring lipid metabolism in Meishan pigs. Br J Nutr, 2013. 109(10): p. 1731-8. 220. Zhang, J., et al., Maternal high fat diet during pregnancy and lactation alters hepatic expression of insulin like growth factor-2 and key microRNAs in the adult offspring. BMC Genomics, 2009. 10: p. 478. 221. Cordero, P., et al., Transcriptomic and epigenetic changes in early liver steatosis associated to obesity: effect of dietary methyl donor supplementation. Mol Genet Metab, 2013. 110(3): p. 388-95. 222. Schwenk, R.W., et al., Diet-dependent alterations of hepatic Scd1 expression are accompanied by differences in promoter methylation. Horm Metab Res, 2013. 45(11): p. 786-94. 223. Jiang, M., et al., Hypermethylation of hepatic glucokinase and L-type pyruvate kinase promoters in high-fat diet-induced obese rats. Endocrinology, 2011. 152(4): p. 1284-9. 224. Chang, X., et al., Berberine reduces methylation of the MTTP promoter and alleviates fatty liver induced by a high-fat diet in rats. J Lipid Res, 2010. 51(9): p. 2504-15. 225. Dudley, K.J., et al., Offspring of mothers fed a high fat diet display hepatic cell cycle inhibition and associated changes in gene expression and DNA methylation. PLoS One, 2011. 6(7): p. e21662. 226. Dunn, G.A. and T.L. Bale, Maternal high-fat diet promotes body length increases and insulin insensitivity in second-generation mice. Endocrinology, 2009. 150(11): p. 4999-5009. 227. Uriarte, G., et al., Shifting to a control diet after a high-fat, high-sucrose diet intake induces epigenetic changes in retroperitoneal adipocytes of Wistar rats. J Physiol Biochem, 2013. 69(3): p. 601-11.

100

228. Vucetic, Z., J. Kimmel, and T.M. Reyes, Chronic high-fat diet drives postnatal epigenetic regulation of mu-opioid receptor in the brain. Neuropsychopharmacology, 2011. 36(6): p. 1199-206. 229. Milagro, F.I., et al., High fat diet-induced obesity modifies the methylation pattern of leptin promoter in rats. J Physiol Biochem, 2009. 65(1): p. 1-9. 230. Widiker, S., et al., High-fat diet leads to a decreased methylation of the Mc4r gene in the obese BFMI and the lean B6 mouse lines. J Appl Genet, 2010. 51(2): p. 193-7. 231. Peng, Y., et al., Diet-induced obesity associated with steatosis, oxidative stress, and inflammation in liver. Surg Obes Relat Dis, 2012. 8(1): p. 73-81. 232. Wu, K.K., Cyclooxygenase 2 induction: molecular mechanism and pathophysiologic roles. J Lab Clin Med, 1996. 128(3): p. 242-5. 233. Morita, I., Distinct functions of COX-1 and COX-2. Prostaglandins Other Lipid Mediat, 2002. 68-69: p. 165-75. 234. Smith, W.L., R.M. Garavito, and D.L. DeWitt, Prostaglandin endoperoxide H synthases (cyclooxygenases)-1 and -2. J Biol Chem, 1996. 271(52): p. 33157-60. 235. Crofford, L.J., COX-1 and COX-2 tissue expression: implications and predictions. J Rheumatol Suppl, 1997. 49: p. 15-9. 236. Williams, C.S., M. Mann, and R.N. DuBois, The role of cyclooxygenases in inflammation, cancer, and development. Oncogene, 1999. 18(55): p. 7908-16. 237. D'Souza A, M., et al., Consumption of a high-fat diet rapidly exacerbates the development of fatty liver disease that occurs with chronically elevated glucocorticoids. Am J Physiol Gastrointest Liver Physiol, 2012. 302(8): p. G850-63. 238. Yuzefovych, L.V., et al., Mitochondrial DNA damage and dysfunction, and oxidative stress are associated with endoplasmic reticulum stress, protein degradation and apoptosis in high fat diet-induced insulin resistance mice. PLoS One, 2013. 8(1): p. e54059. 239. VanSaun, M.N., et al., High fat diet induced hepatic steatosis establishes a permissive microenvironment for colorectal metastases and promotes primary dysplasia in a murine model. Am J Pathol, 2009. 175(1): p. 355-64. 240. Uyy, E., et al., High-fat diet alters protein composition of detergent-resistant membrane microdomains. Cell Tissue Res, 2013. 354(3): p. 771-81. 241. Barbuio, R., et al., Infliximab reverses steatosis and improves insulin signal transduction in liver of rats fed a high-fat diet. J Endocrinol, 2007. 194(3): p. 539-50. 242. Peng, M., et al., The cyclooxygenase-2 inhibitor parecoxib inhibits surgery-induced proinflammatory cytokine expression in the hippocampus in aged rats. J Surg Res, 2012. 178(1): p. e1-8. 243. Smith, W.L., D.L. DeWitt, and R.M. Garavito, Cyclooxygenases: structural, cellular, and molecular biology. Annu Rev Biochem, 2000. 69: p. 145-82. 244. Yu, J., et al., COX-2 induction in mice with experimental nutritional steatohepatitis: Role as pro-inflammatory mediator. Hepatology, 2006. 43(4): p. 826-36. 245. Allport, V.C., et al., Human labour is associated with nuclear factor-kappaB activity which mediates cyclo-oxygenase-2 expression and is involved with the 'functional progesterone withdrawal'. Mol Hum Reprod, 2001. 7(6): p. 581-6.

101

246. Zhao, J.S., et al., Pioglitazone ameliorates nonalcoholic steatohepatitis by down-regulating hepatic nuclear factor-kappa B and cyclooxygenases-2 expression in rats. Chin Med J (Engl), 2012. 125(13): p. 2316-21. 247. Henkel, J., et al., Oncostatin M produced in Kupffer cells in response to PGE2: possible contributor to hepatic insulin resistance and steatosis. Lab Invest, 2011. 91(7): p. 1107-17. 248. Nunez Martinez, O., G. Clemente Ricote, and C. Garcia Monzon, [Role of cyclooxygenase-2 in the pathogenesis of chronic liver diseases]. Med Clin (Barc), 2003. 121(19): p. 743-8. 249. Cervello, M. and G. Montalto, Cyclooxygenases in hepatocellular carcinoma. World J Gastroenterol, 2006. 12(32): p. 5113-21. 250. Bligh, E.G. and W.J. Dyer, A rapid method of total lipid extraction and purification. Can J Biochem Physiol, 1959. 37(8): p. 911-7. 251. Slinker, B.K., The statistics of synergism. J Mol Cell Cardiol, 1998. 30(4): p. 723-31. 252. Kadotani, A., et al., Different impacts of saturated and unsaturated free fatty acids on COX-2 expression in C(2)C(12) myotubes. Am J Physiol Endocrinol Metab, 2009. 297(6): p. E1291-303. 253. Innis, S.M., Essential fatty acid requirements in human nutrition. Can J Physiol Pharmacol, 1993. 71(9): p. 699-706. 254. Bohm, T., et al., Food-derived peroxidized fatty acids may trigger hepatic inflammation: a novel hypothesis to explain steatohepatitis. J Hepatol, 2013. 59(3): p. 563-70. 255. Keller, S.A., et al., Kupffer cell ablation attenuates cyclooxygenase-2 expression after trauma and sepsis. J Surg Res, 2005. 124(1): p. 126-33. 256. Sprecher, H., Biochemistry of essential fatty acids. Prog.Lipid Res., 1981. 20:13-22.: p. 13-22. 257. Hofacer, R., et al., Omega-3 fatty acid deficiency selectively up-regulates delta6-desaturase expression and activity indices in rat liver: prevention by normalization of omega-3 fatty acid status. Nutr Res, 2011. 31(9): p. 715-22. 258. Liang, Y., et al., Constitutive activation of AMPK alpha1 in vascular endothelium promotes high-fat diet-induced fatty liver injury: role of COX-2 induction. Br J Pharmacol, 2014. 171(2): p. 498-508. 259. Cheng, Q., et al., Cyclooxygenase-2 promotes hepatocellular apoptosis by interacting with TNF-alpha and IL-6 in the pathogenesis of nonalcoholic steatohepatitis in rats. Dig Dis Sci, 2013. 58(10): p. 2895-902. 260. Chen, J., et al., Celecoxib attenuates liver steatosis and inflammation in non-alcoholic steatohepatitis induced by high-fat diet in rats. Mol Med Rep, 2011. 4(5): p. 811-6. 261. Cantone, I. and A.G. Fisher, Epigenetic programming and reprogramming during development. Nat Struct Mol Biol, 2013. 20(3): p. 282-9. 262. Gibney, E.R. and C.M. Nolan, Epigenetics and gene expression. Heredity (Edinb), 2010. 105(1): p. 4-13. 263. Arzate-Mejia, R.G., D. Valle-Garcia, and F. Recillas-Targa, Signaling epigenetics: novel insights on cell signaling and epigenetic regulation. IUBMB Life, 2011. 63(10): p. 881-95.

102

264. Burdge, G.C. and K.A. Lillycrop, Fatty acids and epigenetics. Curr Opin Clin Nutr Metab Care, 2013. 265. Iyer, A., et al., Inflammatory lipid mediators in adipocyte function and obesity. Nat Rev Endocrinol, 2010. 6(2): p. 71-82. 266. Vucetic, Z., et al., Epigenetic dysregulation of the dopamine system in diet-induced obesity. J Neurochem, 2012. 120(6): p. 891-8. 267. Iyer, A., D.P. Fairlie, and L. Brown, Lysine acetylation in obesity, diabetes and metabolic disease. Immunol Cell Biol, 2012. 90(1): p. 39-46. 268. Harper, K.A. and A.J. Tyson-Capper, Complexity of COX-2 gene regulation. Biochem.Soc.Trans., 2008/6. 36(Pt 3): p. 543-545. 269. Appleby, S.B., et al., Structure of the human cyclo-oxygenase-2 gene. Biochem.J., 1994/9/15. 302(Pt 3): p. 723-727. 270. Deng, W.G., Y. Zhu, and K.K. Wu, Up-regulation of p300 binding and p50 acetylation in tumor necrosis factor-alpha-induced cyclooxygenase-2 promoter activation. J.Biol.Chem., 2003/2/14. 278(7): p. 4770-4777. 271. Tyson-Capper, A.J., et al., Characterization of cellular retinoid-binding proteins in human myometrium during pregnancy. Mol Hum Reprod, 2006. 12(11): p. 695-701. 272. Yamaguchi, K., et al., Histone deacetylase inhibitors suppress the induction of c-Jun and its target genes including COX-2. J.Biol.Chem., 2005/9/23. 280(38): p. 32569-32577. 273. Song, S.H., et al., Transcriptional silencing of Cyclooxygenase-2 by hyper-methylation of the 5' CpG island in human gastric carcinoma cells. Cancer Res., 2001/6/1. 61(11): p. 4628-4635. 274. Dixon, D.A., et al., Post-transcriptional control of cyclooxygenase-2 gene expression. The role of the 3'-untranslated region. J.Biol.Chem., 2000/4/21. 275(16): p. 11750-11757. 275. Harris, R.A., et al., Comparison of sequencing-based methods to profile DNA methylation and identification of monoallelic epigenetic modifications. Nat.Biotechnol., 2010/10. 28(10): p. 1097-1105. 276. Xie, M., et al., DNA hypomethylation within specific transposable element families associates with tissue-specific enhancer landscape. Nat.Genet., 2013/7. 45(7): p. 836-841. 277. Zhou, D. and Y.X. Pan, Gestational low protein diet selectively induces the amino acid response pathway target genes in the liver of offspring rats through transcription factor binding and histone modifications. Biochim Biophys Acta, 2011. 1809(10): p. 549-56. 278. Gorisch, S.M., et al., Histone acetylation increases chromatin accessibility. J Cell Sci, 2005. 118(Pt 24): p. 5825-34. 279. Stevens, M., et al., Estimating absolute methylation levels at single-CpG resolution from methylation enrichment and restriction enzyme sequencing methods. Genome Res, 2013. 23(9): p. 1541-53. 280. Barres, R., et al., Non-CpG methylation of the PGC-1alpha promoter through DNMT3B controls mitochondrial density. Cell Metab, 2009. 10(3): p. 189-98. 281. Jacobsen, S.C., et al., Effects of short-term high-fat overfeeding on genome-wide DNA methylation in the skeletal muscle of healthy young men. Diabetologia, 2012. 55(12): p. 3341-9.

103

282. Mitchell, J.A. and T.D. Warner, Cyclo-oxygenase-2: pharmacology, physiology, biochemistry and relevance to NSAID therapy. Br J Pharmacol, 1999. 128(6): p. 1121-32. 283. Woo, K.J. and T.K. Kwon, Sulforaphane suppresses lipopolysaccharide-induced cyclooxygenase-2 (COX-2) expression through the modulation of multiple targets in COX-2 gene promoter. Int.Immunopharmacol., 2007/12/15. 7(13): p. 1776-1783. 284. Uto, T., M. Fujii, and D.X. Hou, Inhibition of lipopolysaccharide-induced cyclooxygenase-2 transcription by 6-(methylsulfinyl) hexyl isothiocyanate, a chemopreventive compound from Wasabia japonica (Miq.) Matsumura, in mouse macrophages. Biochem.Pharmacol., 2005/12/5. 70(12): p. 1772-1784. 285. Kang, Y.J., et al., Cyclooxygenase-2 gene transcription in a macrophage model of inflammation. J.Immunol., 2006/12/1. 177(11): p. 8111-8122. 286. Jones, P.A., The DNA methylation paradox. Trends Genet., 1999/1. 15(1): p. 34-37. 287. Adam, R.M., et al., Mechanical stretch is a highly selective regulator of gene expression in human bladder smooth muscle cells. Physiol Genomics, 2004. 20(1): p. 36-44. 288. Zhang, X., et al., Several transcription factors regulate COX-2 gene expression in pancreatic beta-cells. Mol.Biol.Rep., 2007/9. 34(3): p. 199-206. 289. Ramsay, R.G., et al., Cyclooxygenase-2, a colorectal cancer nonsteroidal anti-inflammatory drug target, is regulated by c-MYB. Cancer Res, 2000. 60(7): p. 1805-9. 290. Joo, M., et al., Yin Yang 1 enhances cyclooxygenase-2 gene expression in macrophages. Am J Physiol Lung Cell Mol Physiol, 2007. 292(5): p. L1219-26. 291. Azim, A.C., et al., Regulation of cyclooxygenase-2 expression by small GTPase Rac2 in bone marrow macrophages. Am J Physiol Lung Cell Mol Physiol, 2007. 293(3): p. L668-73. 292. Rossi, A., et al., Regulation of cyclooxygenase-2 expression by heat: a novel aspect of heat shock factor 1 function in human cells. PLoS One, 2012. 7(2): p. e31304. 293. Asting, A.G., et al., COX-2 gene expression in colon cancer tissue related to regulating factors and promoter methylation status. BMC Cancer, 2011. 11: p. 238. 294. Bell, A.C., A.G. West, and G. Felsenfeld, The protein CTCF is required for the enhancer blocking activity of vertebrate insulators. Cell., 1999/8/6. 98(3): p. 387-396. 295. Giles, K.E., et al., Chromatin boundaries, insulators, and long-range interactions in the nucleus. Cold Spring Harb.Symp.Quant.Biol., 2010. 75:79-85. doi: 10.1101/sqb.2010.75.006. Epub;%2010 Nov 3.: p. 79-85. 296. Wang, H., et al., Widespread plasticity in CTCF occupancy linked to DNA methylation. Genome Res., 2012/9. 22(9): p. 1680-1688. 297. Sharma-Walia, N., et al., NFAT and CREB regulate Kaposi's sarcoma-associated herpesvirus-induced cyclooxygenase 2 (COX-2). J.Virol., 2010/12. 84(24): p. 12733-12753. 298. Jones, P.A. and P.W. Laird, Cancer epigenetics comes of age. Nat.Genet., 1999/2. 21(2): p. 163-167. 299. Wilson, A.S., B.E. Power, and P.L. Molloy, DNA hypomethylation and human diseases. Biochim.Biophys.Acta., 2007/1. 1775(1): p. 138-162.

104

300. Sheng, W., et al., Association of promoter methylation statuses of congenital heart defect candidate genes with Tetralogy of Fallot. J.Transl.Med., 2014/1/31. 12:31. doi: 10.1186/1479-5876-12-31.: p. 31-12. 301. Smyth, L.J., et al., DNA hypermethylation and DNA hypomethylation is present at different loci in chronic kidney disease. Epigenetics., 2013/11. %19;9(3).): p. -. 302. Ahrens, M., et al., DNA methylation analysis in nonalcoholic fatty liver disease suggests distinct disease-specific and remodeling signatures after bariatric surgery. Cell Metab., 2013/8/6. 18(2): p. 296-302. 303. Zhang, S., et al., Alteration of PTGS2 promoter methylation in chronic periodontitis. J.Dent.Res., 2010/2. 89(2): p. 133-137. 304. Kaaij, L.T., et al., DNA methylation dynamics during intestinal stem cell differentiation reveals enhancers driving gene expression in the villus. Genome Biol., 2013/5/28. 14(5): p. R50-. 305. Aran, D. and A. Hellman, DNA methylation of transcriptional enhancers and cancer predisposition. Cell., 2013/7/3. 154(1): p. 11-13. 306. Aran, D., S. Sabato, and A. Hellman, DNA methylation of distal regulatory sites characterizes dysregulation of cancer genes. Genome Biol., 2013/3/12. 14(3): p. R21-. 307. Wiench, M., et al., DNA methylation status predicts cell type-specific enhancer activity. EMBO J., 2011/6/24. 30(15): p. 3028-3039. 308. Schmidl, C., et al., Lineage-specific DNA methylation in T cells correlates with histone methylation and enhancer activity. Genome Res., 2009/7. 19(7): p. 1165-1174. 309. Cohen, N.M., E. Kenigsberg, and A. Tanay, Primate CpG islands are maintained by heterogeneous evolutionary regimes involving minimal selection. Cell., 2011/5/27. 145(5): p. 773-786. 310. Antman, E.M., et al., Use of nonsteroidal antiinflammatory drugs: an update for clinicians: a scientific statement from the American Heart Association. Circulation, 2007. 115(12): p. 1634-42.