EPIGENETIC REGULATION of BROWN FAT THERMOGENESIS Fenfen Li

Georgia State University ScholarWorks @ Georgia State University

Biology Dissertations Department of Biology

8-7-2018 EPIGENETIC REGULATION OF BROWN FAT THERMOGENESIS Fenfen Li

Follow this and additional works at: https://scholarworks.gsu.edu/biology_diss

Recommended Citation Li, Fenfen, "EPIGENETIC REGULATION OF BROWN FAT THERMOGENESIS." Dissertation, Georgia State University, 2018. https://scholarworks.gsu.edu/biology_diss/208

This Dissertation is brought to you for free and open access by the Department of Biology at ScholarWorks @ Georgia State University. It has been accepted for inclusion in Biology Dissertations by an authorized administrator of ScholarWorks @ Georgia State University. For more information, please contact [email protected]. i

EPIGENETIC REGULATION OF BROWN FAT THERMOGENESIS

FENFEN LI

Under the Direction of Bingzhong Xue, PhD

ABSTRACT

Diet-induced obesity and its related metabolic diseases including type 2 diabetes are public health concern. While the major function of white adipose tissue (WAT) is to store excess energy as fat, that of brown adipose tissue (BAT) is to dissipate extra energy as heat, suggesting that activating BAT function may be an effective therapy for treating obesity and its related metabolic diseases. Epigenetic modification including histone methylation, histone acetylation and DNA methylation is associated with energy homeostasis. However, the epigenetic modification contributing to BAT reprogramming in high fat diet (HFD)-induced obesity is not known. Here, we found that histone deacetylase 1 (HDAC1) negatively regulated thermogenic program in brown adipocytes through regulation of histone H3 lysine 27 deacetylation and methylation. In addition, we found that HFD induced the reprogramming of brown adipocyte by down-regulating ii

BAT-specific, whereas up-regulating skeletal muscle-specific gene expression program.

Remarkably, we found that knockdown of histone demethylase UTX or DNA methyltransferase DNMT1 mimics HFD-induced brown fat remodeling, leading to BAT dysfunction. In addition, mice with brown fat specific deletion of UTX or DNMT1 displayed obesity and insulin resistance. Mechanistically, we found that UTX promoted brown fat thermogenic gene expression including PRDM16, which then recruited DNA methyltransferase DNMT1 to the promoter of Myod1, an important transcriptional factor for skeletal muscle cell development. DNMT1 increased methylation levels of Myod1 and inhibited its gene transcription in brown adipose tissue. Thus, this pathway is important in maintaining brown adipocyte function; disrupting this pathway in HFD results in reprogramming of brown adipocytes into skeletal muscle-like cells, which contributes to

HFD-induced obesity.

INDEX WORDS: Epigenetics, HDAC1, DNMT1, UTX, Brown adipose tissue. iii

EPIGENETIC REGULATION OF BROWN FAT THERMOGENESIS

FENFEN LI

A Dissertation Submitted in Partial Fulfillment of the Requirements for the Degree of

Doctor of Philosophy

in the College of Arts and Sciences

Georgia State University

2018

EPIGENETIC REGULATION OF BROWN FAT THERMOGENESIS

FENFEN LI

Committee Chair: Bingzhong Xue

Committee: Hang Shi

Ming-Hui Zou

Electronic Version Approved:

Office of Graduate Studies

College of Arts and Sciences

Georgia State University

August 2018 iv

DEDICATION

To Jesus Christ, as He is my strength in everything I do;

To my husband Lizhi Fu for supporting and encouraging me;

To my sons (Ethan and Nathan), because of you I become soft and joyful;

To my beloved grandparents (Gentu Jiang and Yuexian Zhong) for nursing me with patience and love;

To my parents (Tuhua Li and Yunlan Jiang) and my parents in law (Ronghua Fu and Zhiying Zhang) for nursing my sons that I have time to finish my research projects; v

ACKNOWLEDGEMENTS

First and foremost, I would like to express my gratitude to Dr. Bingzhong Xue for being my advisor and my mentor, who is patient, kind and absolutely brilliant. Thank you for providing your important perspective on my projects. I would like to thank my dissertation committee members. To Dr. Hang Shi, thank you for your advice on experimental design, scientific writing and presentation! To Dr. Ming-Hui Zou, I am grateful for your advice and guidance on my projects and dissertation. I am grateful in so many ways to each member of my lab family: Dr. Xin Cui, Dr. Qiang Cao, Dr. Jia

Jing, Dr. Rui Wu, Dr. Xiaosong Yang, Dr. Lin Zha, Lizhi Fu, Yii-shyuan Chen,

Anubama Rajan, Shirong Wang, Neda Movahhed Avval, Ke Li, Shuping Kou,

Mary Schneider, Michael Alex Thomas, Ngoc Ly Nguyen. Everyone deserves such an excellent interpersonal lab environment to work in! I especially thank Professor

Qiang-min Xie, who was the first person I looked up to in graduate school, and has encouraged me in so many ways. I would like to express my thankfulness to Dr. Phang-

Chang Tai for your great recommendation. I also would like to express my gratefulness to Dr. Tim Bartness for giving me a step up, in so many ways, throughout the first three years in Georgia State University (GSU). To my husband Lizhi Fu, thank you for constant companion for six years in GSU. To my parents Tuhua Li and Yunlan Jiang, thank you for your love and support. To my grandparents Gentu Jiang and Yuexian Zhong, thank you for raising me to be patient like you. Finally, I would like to acknowledge the

Second Century Initiative (2CI) University Doctoral Fellowship and NIH for their financial support for my study and research.

TABLE OF CONTENTS

ACKNOWLEDGEMENTS ...... V

LIST OF TABLES ...... XI

LIST OF FIGURES ...... XII

LIST OF ABBREVIATIONS ...... XIV

1 GENERAL INTRODUCTION ...... 1

1.1 Obesity disease and energy hemostasis ...... 1

1.2 Brown adipocyte activation ...... 2

1.3 Epigenetic modification regulates brown adipocyte thermogenesis ...... 3

1.4 Specific aims of the dissertation ...... 5

2 SPECIFIC AIM 1 ...... 13

(Results of SA1 have been published in J Biol Chem. 2016 Feb

26;291(9):4523-36.)...... 13

2.1 Abstract ...... 13

2.2 Introduction ...... 14

2.3 Materials and Methods ...... 17

2.3.1 Mice ...... 17

2.3.2 Cell Culture...... 17

2.3.3 small interfering RNA (siRNA) and plasmid DNA

transfection ...... 18 vii

2.3.4 Quantitative real-time RT-PCR ...... 19

2.3.5 Immunoblotting ...... 19

2.3.6 Chromatin immunoprecipitation (ChIP) assay ...... 20

2.3.7 Statistical Analysis ...... 21

2.4 Results ...... 21

2.5 Discussion of specific aim 1 ...... 31

2.6 Figures of specific aim 1 ...... 39

2.7 Tables of specific aim 1 ...... 47

3 SPECIFIC AIM 2 ...... 51

3.1 Abstract ...... 51

3.2 Introduction ...... 52

3.3 Materials and methods ...... 54

3.3.1 Mice ...... 54

3.3.2 Metabolic measurements ...... 55

3.3.3 HE staining, immunofluorescence (IF) and immunohistochemistry (IHC) ...... 56

3.3.4 Cell Culture ...... 57

3.3.5 Oil-red-O staining ...... 57

3.3.6 small interfering RNA (siRNA) transfection ...... 58

3.3.7 Quantitative real-time RT-PCR ...... 58 viii

3.3.8 RNA sequencing ...... 59

3.3.9 Immunoblotting ...... 59

3.3.10 Chromatin immunoprecipitation (ChIP) assay ...... 59

3.3.11 Bisulfite conversion and pyrosequencing ...... 60

3.3.12 Statistical Analysis ...... 60

3.4 Results ...... 60

3.4.1 High-fat diet (HFD) reprograms brown fat to a muscle-like phenotype and modulates brown fat thermogenic function...... 60

3.4.2 Deficiency of DNMT1 promotes myogenic gene expression and prevents brown thermogenic gene expression in brown adipocytes...... 62

3.4.3 Deletion of DNMT1 in brown fat results in the impairment of cold-induced brown fat thermogenesis...... 63

3.4.4 Deletion of DNMT1 in brown fat results in obese and insulin resistance...... 64

3.4.5 DNMT1 regulates myogenesis and thermogenesis in BATs through targeting Myod1 ...... 66

3.5 Discussion of specific aim 2 ...... 71

3.6 Figures of specific aim 2 ...... 79

3.7 Tables of specific aim 2 ...... 91

4 SPECIFIC AIM 3 ...... 95 ix

4.1 Abstract ...... 95

4.2 Introduction ...... 96

4.3 Materials and methods ...... 98

4.3.1 Mice ...... 98

4.3.2 small interfering RNA (siRNA) transfection ...... 98

4.3.3 RNA sequencing ...... 99

4.3.4 Metabolic measurements, Histological analysis, molecular analysis and Statistical Analysis ...... 99

4.4 Results ...... 100

4.4.1 BAT-specific UTX deletion reprograms brown fat to a muscle-like phenotype and reduces BAT thermogenic function ...... 100

4.4.2 BAT-specific UTX deletion impairs cold-induced thermogenesis ...... 101

4.4.3 Mice with BAT-specific UTX deletion are prone to diet- induced obesity and metabolic disorders...... 102

4.4.4 UTX regulates BAT/skeletal muscle programming via targeting PRDM16 and Myod1...... 104

4.5 Discussion of specific aim 3 ...... 106

4.6 Figures of specific aim 3 ...... 113

5 GENERAL DISCUSSION ...... 120

REFERENCES ...... 126 x

APPENDICES: PUBLICATIONS ...... 152

LIST OF TABLES

Table 1 Antibodies used in immunoblotting and ChIP-qPCR ...... 47

Table 2 Primer/probe sequence for gene expression experiments ...... 48

Table 3 Primer/probe sets for gene expression experiments ...... 49

Table 4 Sequence of siRNA smart pools for individual HDACs ...... 50

Table 5 Primer/probe sets for Myogenic gene expression ...... 91

Table 6 Primers for ChIP assay ...... 92

Table 7 Primer sequences for pyrosequencing ...... 93

Table 8 Primer sequence for Myod1 gRNA ...... 94

xii

LIST OF FIGURES

Figure 1 The expression pattern of HDAC1...... 39

Figure 2 Reducing HDAC expression promotes Brown-specific gene expression in BAT1

brown adipocytes...... 40

Figure 3 HDAC1 overexpression attenuates mRNA expression of Brown-specific genes.

...... 41

Figure 4 HDAC1 negatively regulates Brown-specific gene expression in HIB-1B brown

adipocytes...... 42

Figure 5 HDAC1 mediates the effects of class I HDACi MS-275 on BAT-1 brown

adipocyte gene expression...... 43

Figure 6 HDAC1 regulates H3K27ac at PGC1α and UCP1 promoters...... 44

Figure 7 HDAC1 regulates trimethylation of H3K27me3 at UCP1 and PGC1α promoters

through coordinated regulation of EZH2 and UTX binding to BAT-specific

promoters...... 45

Figure 8 HDAC1 regulates the recruitment of polycomb repressor complexes to UCP1

and PGC1α promoters...... 46

Figure 9 High-fat diet (HFD) induces myogenic gene expression in BATs...... 79

Figure 10 Brown specific-DNMT1 deficiency up-regulates myogenic genes and down-

regulates brown-selective genes in brown adipocytes...... 80

Figure 11 DNMT1 KO in BAT reduces thermogenesis in BAT and reduces cold tolerance.

...... 81

Figure 12 Chow fed DNMT1 KO mice display an obese phenotype...... 82

Figure 13 DNMT1 KO male mice display an obese phenotype ...... 83

Figure 14 DNMT1 KO female mice display High-fat (HF) diet-induced obesity...... 84 xiii

Figure 15 DNMT1 KO male mice display high-fat (HF) diet-induced obesity...... 85

Figure 16 DNMT1 regulates myogenesis and brown fat development by targeting Myod1.

...... 86

Figure 17 DNMT1 regulates myogenic transcription factors...... 87

Figure 18 Myod1 is the target of DNMT1 regulating expression of myogenic genes and

thermogenic genes in brown adipocytes...... 88

Figure 19 DNMT1 regulates methylation levels of Myod1 promoters...... 89

Figure 20 DNMT1 regulates myogenesis and thermogenesis in BATs through targeting

Myod1...... 90

Figure 21 Brown specific-UTX deficiency up-regulates myogenic genes and down-

regulates brown-selective genes in brown adipocytes...... 113

Figure 22 UTX KO in BAT reduces cold-induced thermogenesis in BATs...... 114

Figure 23 UTX KO mice display obese phenotype...... 115

Figure 24 Low metabolic respiration in UTX KO mice...... 116

Figure 25 UTX KO mice display insulin resistance...... 117

Figure 26 UTX siRNA increased enrichment of H3K27me3 in promoters of PRDM16

and UCP1...... 118

Figure 27 Methylation levels of Myod1 promoters were regulated by UTX and PRDM16.

...... 119

xiv

LIST OF ABBREVIATIONS

ACOX1: acyl-CoA oxidase 1

Acta1: actin, alpha 1, skeletal muscle aP2: adipocyte protein 2

Atp2a1: ATPase, Ca++ transporting, cardiac muscle, fast twitch 1

BAT: brown adipose tissue

C/EBPα: CCAAT/enhancer-binding protein α

CIDEA: cell death activator CIDE-A

CKM: creatine kinase, muscle

COX1: cytochrome c oxidase 1

CPT1B: Carnitine palmitoyltransferase I B

CRE: cAMP-response element

CREB: cAMP response element-binding protein

DIO: diet-induced obese

DIO2: Type II iodothyronine deiodinase

DNMTs: DNA methyltransferases

EBF2: early B-cell factor-2

EBF3: early B-cell factor-3

ELOVL3: Elongation of very long chain fatty acids protein 3

EPI: epididymal

EVA1: epithelial V-like antigen 1

EZH2: Enhancer of zeste homolog 2

FGF21: fibroblast growth factor 21 gRNA: guide RNA xv

GTT: Glucose tolerance testing

HDAC: histone deacetylase

HDACi: HDAC inhibitor

H3K4ac: histone H3 Lys-4 acetylation

H3K9ac: histone H3 Lys-9 acetylation

H3K14ac: histone H3 Lys-14 acetylation

H3K27: histone H3 Lys-27

H3K27ac: H3K27 acetylation

H3K27me3: H3K27 trimethylation

HFD: high fat diet

IP: immunoprecipitation

ITT: insulin tolerance testing

Isop: isoproterenol

KD: knockdown

KLF2: Kruppel-like factor 2

KO: knockout

Mfsd2a: major facilitator superfamily domain containing 2A

Myh1: myosin, heavy polypeptide 1, skeletal muscle, adult

Myh4: myosin, heavy polypeptide 4, skeletal muscle

NE: norepinephrine

Nr4a1: nuclear receptor subfamily 4, group A, member 1

Nr4a2: nuclear receptor subfamily 4, group A, member 2

Nr4a3: nuclear receptor subfamily 4, group A, member 3

OTOP1: otopetrin 1 xvi

PGC1α: proliferator-activated receptor γ co-activator 1α

PGC1β: proliferator-activated receptor γ co-activator 1 β

PKA: protein kinase A

PPARγ: peroxisome proliferator-activated receptor γ

PPARα: peroxisome proliferator-activated receptor α

PRDM16: PR domain-containing protein 16

RNF2: ring finger protein 2

SAHA: suberoylanilidehydroxamic acid (vorinostat).

Sik1: salt inducible kinase 1

SREBP1C: Sterol regulatory element-binding transcription factor 1 C

SUZ12: SUZ12 polycomb repressive complex 2 subunit

TETs: Ten-eleven translocation methylcytosine dioxygenases

TG: triglycerides

Tnnt3: troponin T3, skeletal, fast

Ttn: titin

TSA: trichostatin A

UTX: Ubiquitously transcribed tetratricopeptide repeat, X chromosome

UCP1: uncoupling protein 1

WAT: white adipose tissue 1

1 GENERAL INTRODUCTION

1.1 Obesity disease and energy hemostasis Obesity is considered as a severe health issue, reducing life expectancy and leading to various metabolic diseases including type 2 diabetes, dyslipidemia, hypertension and other cardiovascular diseases, and certain types of cancer. Two-thirds of United States adults are obese, and the incidence of obesity continues to grow (Kraschnewski et al.,

2010). Approximately $147 billion is spent to prevent obesity and cure obesity-related diseases in the United States (Couzin-Franke, 2012). It is very urgent to study and control obesity. The persistent imbalance between energy intake and energy expenditure is the cause of obesity (Hill et al., 2012). Thus, understanding mechanisms regulating energy balance can help develop strategies to prevent or treat obesity. High fat diet (HFD) induces obesity and insulin resistance. Recent studies link HFD and metabolic disorders including adipose inflammation, hepatic steatosis, heart dysfunction, muscle atrophy (Birse et al., 2010, Fu et al., 2015a, Gao et al., 2018, Onishi et al., 2018). However, brown adipose tissue (BAT) function during HFD feeding is not well-studied. BAT consumes energy for thermogenesis due to activation of uncoupling protein 1 (UCP1) in the inner mitochondrial membrane (Cannon and Nedergaard,

2004). It is well known now that BAT is a thermogenic organ not only in rodents and human infants but also in adult humans in response to mild cold exposure (Cypess et al., 2009, van Marken Lichtenbelt et al., 2009, Enerback, 2010, Cypess et al., 2013).

The activation of brown fat is positively associated with energy expenditure and negatively associated with obesity (Feldmann et al., 2009). Therefore, activation 2

of brown adipocytes in humans is a promising strategy for the treatment of obesity and its related metabolic diseases.

1.2 Brown adipocyte activation There are two types of brown adipocytes: classical brown adipocytes in BAT and beige/brite adipocytes (brown adipocyte-like) in WAT induced by cold exposure or β3 adrenergic stimulation in rodents (Harms and Seale, 2013, Shabalina et al., 2013). The thermogenic capacity of brown adipocyte is dependent on its unique expression of

UCP1, which dissipates proton electrochemical gradient generated by combustion of fatty acids in the form of heat. Cold exposure or high fat diet (HFD) triggers sympathetic neurons to release neurotransmitter norepinephrine (NE) (Rothwell and

Stock, 1979, Puigserver et al., 1998, Feldmann et al., 2009, Sanchez-Gurmaches et al.,

2018). NE activates β3 adrenoreceptor signaling pathway via cAMP formation, protein kinase A (PKA) activation. On the one hand, activated PKA phosphorylates lipases such as hormone sensitive lipase (HSL) and adipose triglyceride lipase (ATGL) to break down triglycerides (TG) into free fatty acids (FFAs) (Schreiber et al., 2017).

Interestingly, recent studies demonstrate that lipolysis in white adipose tissue (WAT) not BAT is essential for cold-induced brown fat thermogenesis (Schreiber et al., 2017,

Shin et al., 2017). FFAs released from WAT or from diet (HFD) are delivered to BAT and served as fuel and activate UCP1. On the other hand, activated PKA phosphorylates transcription factor cAMP response element-binding protein (CREB) to promote UCP1 gene expression in brown adipocytes. This transcriptional control depends on regulatory elements in the promoter of UCP1 gene, which includes a 3

proximal promoter region (immediately upstream of transcription start site) and enhancer region (about 2.5 kb upstream of transcription start site). The proximal region contains a cAMP response element (CRE) which is a binding site for CRE- binding protein (CREB), and CCAAT-enhancer-binding protein binding site (C/EBP), the binding site for C/EBPs (such as CEBPα and CEBPβ). The enhancer region of UCP1 promoter regulates tissue-specific expression of UCP1. Except for CRE, UCP1 promoters also contain retinoic acid response element (RARE) that binds retinoic acid, thyroid receptor element (TRE) that binds thyroid hormone, and peroxisome proliferator-activated receptor response element (PPRE) that binds PPARα, PPARγ and PPARγ co-activator 1-alpha (PGC1α) (Kozak et al., 1994, Villarroya et al., 2017).

Other transcription factors are also essential for cAMP-mediated induction of UCP1 gene such as PR-domain containing protein 16 (PRDM16), which induces UCP1 gene expression through interacting with CEBPβ, PPARα, PPARγ and PGC1α (Seale et al.,

2007). In conclusion, transcriptional control of UCP1 gene is an essential point in the process of brown adipocyte thermogenesis.

1.3 Epigenetic modification regulates brown adipocyte thermogenesis Except for transcription factor of brown thermogenesis mentioned above, epigenetic modifications have also emerged as important mechanisms regulating brown adipocyte function (Inagaki et al., 2016). Although all cell types share the same genome, epigenetic modification (including DNA methylation and histone modification) results in phenotypic variability correlating with human diseases which is without change of

DNA sequence. Epigenetic modification regulates gene transcription through 4

influencing chromatin structure, making it more or less “compact” to transcriptional machinery (Maunakea et al., 2010, Bannister and Kouzarides, 2011). Recent evidence suggests that epigenetic modification has emerged as an important link between environmental factors and obesity (Wang et al., 2013a, Rosen, 2016, You et al., 2017).

Abnormal DNA methylation is associated with complex diseases such as obesity. Un- methylation of gene promoters facilitates access of transcriptional factors to underlying gene promoters. For example, UCP1 transcription is regulated by alteration of DNA methylation status in its promoter regions (Shore et al., 2010). DNA methylation is catalyzed by DNA methyltransferases including DNMT3A, DNMT3B, and DNMT1. While DNA demethylation is regulated by ten-eleven translocation methylcytosine dioxygenases (TETs) including TET1, TET2, and TET3, which, in a series of oxidation steps, convert 5-methylcytosine (5mC) to 5-hydroxymethylcytosine

(5hmC), 5-formylcytosine (5fC) and 5-carboxylcytosine (5caC). 5-fC and 5-caC can then be converted back to the unmethylated cytosine (Wu and Zhang, 2014). Recent studies indicate the role of DNA methylation in adipocyte development by using DNA methyltransferase inhibitor, 5-aza-2'-deoxycytidine (5azadC) (Chen et al., 2016a, You et al., 2017). DNMT1 silencing accelerates adipocyte differentiation (Londono Gentile et al., 2013).Adipocyte DNMT3A knockout attenuates diet-induced insulin resistance

(You et al., 2017). Diet induces DNMT3b expression, leading to DNA methylation at the PPARγ1 promoter, which may contribute to insulin resistance in obese mice (Yang et al., 2014). However, the role of DNMTs in the regulation of brown adipocyte function is unclear. 5

The major histone modifications include histone acetylation and histone methylation.

Histone acetylation is enriched at enhancer elements and promoter regions near transcriptional start site in genes, where they facilitate the transcriptional factor access.

Histone methylation regulates either gene activation or repression depending on which residues are modified and on the degree of methylation. For example, the histone H3 Lys-9 demethylase KDM3A directly regulates UCP1 expression, and genetic deletion of KDM3A in mice results in obesity (Tateishi et al., 2009, Okada et al., 2010). The histone deacetylase HDAC3 promotors brown fat-specific gene expression including UCP1 and regulates brown thermogenesis in response to cold exposure (Emmett et al., 2017). These findings indicate how some epigenetic modification enzymes influence UCP1 gene expression and metabolic diseases such as obesity, but our knowledge is far from complete.

1.4 Specific aims of the dissertation The overarching hypothesis is that epigenetic modification in brown adipocytes regulates energy homeostasis and obesity development. We down-regulated specific epigenetic enzymes in brown adipocytes by using siRNA in cultured brown adipocyte cell lines in vitro (BAT1/HIB-1B) and using Cre/loxP recombination to specifically disrupt endogenous epigenetic enzymes in mouse brown adipocytes in vivo (UCP1-

Cre), which allows us to study in depth the role of epigenetic modification in brown thermogenesis and whole body energy homeostasis under over-nutrition or cold stress condition. DNA methylation, histone deacetylation, methylation of histone H3 at lysine residue 27 (H3K27) are some of the important epigenetic modifications related 6

to a repressed chromatin state. Therefore, this dissertation is organized into three principal Specific Aims (SAs) that focus on answering the following questions:

SA1: What is the function of Histone Deacetylase 1 (HDAC1) in regulating thermogenic program in brown adipocytes? Recent studies report that pan inhibitors of HDACs attenuate high-fat diet-induced insulin resistance. It is classⅠ

HDACs inhibitor, not classⅡ HDACs inhibitor that facilitates mitochondrial biogenesis in skeletal muscle and adipose tissues (Galmozzi et al., 2013), suggesting that individual HDACs may play different roles in metabolism and energy homeostasis.

However, little is known about the role of individual HDACs in brown thermogenic program. To test it, we separately deleted distinct family members of HDACs in immortalized brown adipocytes (HIB-1B cells) by siRNA. Interestingly, the most dramatically increased level of UCP1 expression was observed in brown adipocytes with HDAC1 deletion (Fig. 1A). This promoted us to investigate the role of HDAC1 in brown thermogenic program. Therefore, the hypothesis for SA1 is that HDAC1 negatively regulates brown adipocyte thermogenic program. It is tested by two following sub-aims: Aim 1.1. How does HDAC1 regulate brown adipocyte thermogenic program? The experiments in Aim 1.1 studied the role of HDAC1 in brown adipocyte thermogenesis. We first measured the HDAC1 expression pattern in brown and white adipocytes and its expression during brown adipocyte differentiation and β adrenergic stimulation. Then, we used loss- or gain- function approaches to measure whether HDAC1 deficiency or overexpression affects brown thermogenic gene expression including UCP1 and PGC1α. Aim 1.2. What is the mechanism underlying 7

HDAC1 regulating brown adipocyte thermogenesis? Considering HDAC1 regulates histone modification, we tested the histone acetylation and methylation in UCP1 and

PGC1α prom0ters.

SA2: Does DNA methyltransferase 1 (DNMT1) mediate brown fat reprogramming and regulate whole-body energy homeostasis in HFD- induced obesity? DNA methylation status is also associated with gene transcriptional repression which plays important roles in cell differentiation. DNA methylation biphasically regulates white adipocyte differentiation, promoting adipocyte differentiation at the early stage whereas inhibiting lipogenesis at the late stage (Yang et al., 2016). Pharmacological inhibition of DNA methylation switches white adipogenesis to osteoblastogenesis by activating Wnt10a (Chen et al., 2016a).

DNA methyltransferase 1 (DNMT1) is a major enzyme for maintenance of DNA methylation (Ko et al., 2005). Deletion of DNMT1 results in global demethylation (Ko et al., 2005, Elliott et al., 2015). DNA methylation is a reversible and dynamic process responding to the changes in nutrient status and physical stimuli which affect systemic energy homeostasis (Ko et al., 2005). DNMT1 gene expression in white adipocytes is upregulated in db/db mice, and wild-type mice fed high-fat diet (Kim et al., 2015).

Moreover, DNMT1 gene expression in human white adipocytes positively correlates with body mass index (Sharma et al., 2015). These findings suggest that DNMT1 may play a crucial role in adipocyte differentiation. However, there is no study to identify the role of DNMT1 in the regulation of brown adipocyte development and activity. In our preliminary studies, we find that DNMT1 deletion in brown fat reduced brown 8

specific gene expression including UCP1 and PRDM16, and increased myogenic gene expression including myogenic differentiation 1 (Myod1), myosin heavy chain 1 (Myh1), and myogenin (Myog). DNMT1 also play important roles in myogenesis (Wang et al.,

2015). Ectopic expression of DNMT1 represses, whereas inhibition of DNMT1 promotes myogenic differentiation in myoblast cells. Therefore, the hypothesis for

SA2 is that brown adipocyte-specific deletion of DNMT1 reprograms brown fat to a muscle-like phenotype, reduces brown fat thermogenic function and results in obesity.

It is addressed by several sub-aims: Aim 2.1. Does brown fat DNMT1 deletion reprograms brown fat to a muscle-like phenotype? In our experiment, we used

Cre/loxp recombination technology to achieve brown adipose tissue-specific gene ablation by crossing the DNMT1f/f mice with mice that express Causes Recombination or Cyclization recombination (Cre) recombinase under the control of UCP1 promoter

(DNMT1f/f-UCP1-Cre+/-). The DNMT1f/f mouse was generated by inserting two loxp sites on both sides of exon 4-5 of DNMT1 (Jackson-Grusby et al., 2001). Cre gene encodes a DNA recombinase, which can recognize 34-bp loxp sites and catalyze recombination between two loxp sites. UCP1 is a BAT-specific marker. Using DNMT1 knockout (KO) mice, we can specifically disrupt endogenous DNMT1 in brown adipose tissues. DNMT1f/f mice are served as control. We measured myogenic and brown thermogenic gene expression in Brown adipose tissues (BATs) from DNMT1 knockout

(KO) mice and control mice. Aim 2.2. Does brown fat DNMT1 deletion results in the impairment of cold-induced thermogenesis? BAT is a thermogenic tissue that can be activated in response to physiological stimuli. Cold temperature triggers sympathetic 9

discharge, leading to the release of norepinephrine (NE) in BATs (Vitali et al., 2012).

Through targeting adrenergic receptors, NE facilitates thermogenesis in brown adipocytes (Cannon and Nedergaard, 2004). Cold-induced BAT activity is also observed in adult humans (Cypess et al., 2012, van der Lans et al., 2013). In adult humans, mild cold exposure induces brown adipocyte metabolism without affecting sympathetic activity in other tissues (Cypess et al., 2012). Activated BAT might be a potential target to counterbalance obesity. Thus, this experiment is to investigate whether BAT-specific DNMT1 KO affects cold-induced brown thermogenesis. Body temperature and gene expression in fat tissues were measured during cold exposure in DNMT1 KO and CONTROL mice. Aim 2.3. Does brown fat DNMT1 deletion results in obesity and metabolic disorders? Mice (both male and female) underwent phenotypic characterizations including body weight, body composition, food intake, metabolic phenotypes and glucose metabolism. Phenotyping was performed as follows.

1) Mice Body weight was weekly measured. 2) Body composition was measured by using Bruker Minispec NMR Body Composition Analyzer (Bruker BioSpin

Corporation; Billerica, MA). 3) Food intake was measured over seven consecutive days.

4) Metabolic Phenotyping was performed in the TSE Comprehensive Laboratory

Animal Monitoring System over seven consecutive days. 5) Glucose tolerance testing

(GTT) and insulin tolerance testing (ITT) were performed. 6) At the end of diet treatment, mice were euthanized, and brown and white adipose tissues were collected for measuring the expression of thermogenic genes. Aim 2.4. Does DNMT1 regulate myogenesis and thermogenesis in BATs through targeting Myod1? In our preliminary 10

experiments, we found that brown fat deletion of DNMT1 upregulated expression of

Myod1, which indicates that Myod1 might be the direct target of DNMT1. We first measured the binding of DNMT1 and methylation levels in Myod1 promoters. We then double knockdown (KD) DNMT1 and Myod1 to test whether Myod1 KD reverses

DNMT1 KD-induced decreases in brown thermogenic genes and increases in myogenic genes. Third, we used CRISPR-dCas9-based tools to specifically target Myod1 promoter methylation.

SA3: Does lysine demethylase 6A (UTX) mediate brown fat reprograming and regulate whole-body energy homeostasis in HFD-induced obesity?

DNA methylation cooperates with histone methylation (Kiskinis et al., 2007, Li et al.,

2015). To identify functional epigenetic enzymes that regulate brown thermogenesis activity, we deleted most enzymes that regulate histone methylation including histone methyltransferases and histone demethylases in cultured mature brown adipocytes

(BAT1). We then isolated RNA and measured UCP1 gene expression. We found some enzymes that influence UCP1 gene expression in BAT1 cells, one of which is lysine demethylase 6A (KDM6A, is also known as UTX). UTX deletion decreases brown thermogenic gene expression in cultured mature brown adipocytes (Zha et al., 2015,

Zha et al., 2016). Therefore, our hypothesis for SA3 is that BAT-specific UTX KO reprograms brown fat to a muscle-like phenotype, decreases brown thermogenesis and leads to diet-induced obesity. It is addressed by several sub-aims: Aim3.1. Does BAT- specific UTX deletion reprogram brown fat to a muscle-like phenotype and reduces

BAT thermogenic function? To achieve brown adipose tissue-specific gene ablation of 11

UTX, we crossed UTXf/f mice (Welstead et al., 2012) to UCP1-Cre mice (UTXf/f-

UCP1-Cre+/-) mice and used UTXf/f mice as controls in our experiments. We measured myogenic and brown thermogenic gene expression in Brown adipose tissues

(BATs) from UTX knockout (KO) mice and CONTROL mice. Aim3.2. Does BAT- specific UTX deletion impair cold-induced thermogenesis? Cold exposure induces brown thermogenesis which allows the body to defend cold. We explored the effect of

UTX on the thermogenic activity of brown adipose tissues in response to cold exposure.

We measured body temperature and myogenic- and thermogenic gene expression in

BATs from UTX KO and CONTROL mice during cold exposure. Aim3.3. Are mice with BAT-specific UTX deletion prone to diet-induced obesity and metabolic disorders?

Obesity has been identified as a disease, which increases the healthcare burden and decreases life expectancy (Couzin-Franke, 2012). Given the complexity and long-term process of obesity, it is difficult to understand the exact mechanism of obesity in humans. HFD-induced obesity in mice mimics the pathophysiological change of obesity in humans. The mouse model of HFD-induced obesity is a good tool for studying the development of obesity. In the past thirty years, many studies have identified metabolic derangements in mice fed high-fat diet (Storlien et al., 1986, Luo et al., 2018). Our study is to further investigate whether BAT-specific UTX deficiency prevents brown thermogenesis and promotes HFD-induced obesity. UTX KO and

CONTROL mice were fed high-fat diet (60 kcar% fat diet) or chow diet (10 kcar% fat diet). Phenotypic characterizations including body weight, body composition, food intake, metabolic phenotypes and glucose metabolism were measured as described in 12

SA2. Aim3.4. What is the molecular mechanism underlying UTX regulating muscle/BAT programming? Considering UTX is H3K27me3 demethylase, we measured the enrichment of H3K27me3 in brown thermogenic gene promoters such as UCP1 and PRDM16. It is reported that PRDM16 promotes Myod1 promoter methylation and repression Myod1 gene transcription (Li et al., 2015). However, the molecular mechanism of PRDM16 regulating Myod1 promoter methylation is unknown. We measured whether PRDM16 recruits DNA methyltransferases such as

DNMT1 to bind to Myod1 promoter and regulates its methylation status. 13

2 SPECIFIC AIM 1

(Results of SA1 have been published in J Biol Chem. 2016 Feb 26;291(9):4523-36.)

2.1 Abstract Inhibiting class I histone deacetylases (HDACs) increases energy expenditure, reduces adiposity and improves insulin sensitivity in obese mice. However, the precise mechanism is poorly understood. Here, we demonstrate that HDAC1 is a negative regulator of brown adipocyte thermogenic program. HDAC1 level is lower in mouse brown fat (BAT) than white fat, is suppressed in mouse BAT during cold exposure or

β3-adrenergic stimulation, and is down-regulated during brown adipocyte differentiation. Remarkably, overexpressing HDAC1 profoundly blocks, whereas deleting HDAC1 significantly enhances β-adrenergic activation-induced BAT-specific gene expression in brown adipocytes. β-adrenergic activation in brown adipocytes results in a dissociation of HDAC1 from promoters of BAT-specific genes, including uncoupling protein 1 (UCP1) and peroxisome proliferator-activated receptor γ coactivator 1α (PGC1α), leading to increased acetylation of histone H3 lysine 27 (H3K27), an epigenetic mark of gene activation. This is followed by dissociation of the polycomb repressive complexes, including the H3K27 methyltransferase enhancer of zeste homologue (EZH2), suppressor of zeste 12 (SUZ12), and ring finger protein 2 (RNF2) from, and concomitant recruitment of H3K27 demethylase ubiquitously transcribed tetratricopeptide repeat on chromosome X (UTX) to UCP1 and PGC1α promoters, leading to decreased H3K27 trimethylation, a histone transcriptional repression mark.

Thus, HDAC1 negatively regulates brown adipocyte thermogenic program, and 14

inhibiting HDAC1 promotes BAT-specific gene expression through a coordinated control of increased acetylation and decreased methylation of H3K27, thereby switching transcriptional repressive state to active state at the promoters of UCP1 and

PGC1α. Targeting HDAC1 may be beneficial in prevention and treatment of obesity by enhancing BAT thermogenesis.

2.2 Introduction Obesity develops when a persistent imbalance between energy intake and energy expenditure occurs (Hill et al., 2012). Whereas white adipose tissue (WAT) is involved in energy storage, the role of brown adipose tissue (BAT) is to dissipate energy as heat due to its unique expression of uncoupling protein 1 (UCP1) (Nicholls and Locke, 1984,

Cannon and Nedergaard, 1985, 2004, Oelkrug et al., 2015). In rodents, there exist two types of brown adipocytes. Traditional brown adipocytes are located in discrete areas, whereas “inducible” beige adipocytes are dispersed in WAT (Seale et al., 2008, Wu et al., 2012) and can be induced by cold exposure or β3-adrenergic receptor activation

(Himms-Hagen, 1989, Xue et al., 2005). The ability of brown/beige adipocytes to produce adaptive thermogenesis depends on the unique expression of UCP1 in the inner mitochondrial membrane, which serves to uncouple oxidative phosphorylation from ATP synthesis, thereby profoundly increasing energy expenditure (Nicholls and

Locke, 1984, Cannon and Nedergaard, 1985, 2004, Oelkrug et al., 2015). Recent reports demonstrate that adult humans also possess metabolically active brown fat; the amount of brown fat is inversely correlated with body weight but positively 15

correlated with energy expenditure (Cypess et al., 2009, van Marken Lichtenbelt et al.,

2009, Virtanen et al., 2009). This important discovery provides new insight into the mechanisms regulating energy homeostasis in adult humans and suggests that increasing functional brown/beige adipocytes in humans is a novel and promising target in treating obesity.

Although the genome is fixed and identical in all cells, the epigenome, the combination of all genome-wide DNA and chromatin modifications, is continuously modified in response to developmental, environmental, physiological, and pathological cues (Maunakea et al., 2010, Bannister and Kouzarides, 2011). Epigenetic modifications, including DNA methylation, histone acetylation, and methylation, result in organization of the chromatin structure on different hierarchal levels, which regulate gene expression (Maunakea et al., 2010, Bannister and Kouzarides, 2011).

Recent evidence suggests that epigenetic mechanisms have emerged as an important link between environmental factors and obesity (Maier and Olek, 2002, Ling and

Groop, 2009, Szarc vel Szic et al., 2010). For example, Ucp1 promoter activity is regulated by changes in DNA methylation status (Shore et al., 2010). The histone H3

Lys-9 demethylase JHDM2a directly regulates UCP1 expression, and genetic deletion of Jhdm2a in mice results in obesity (Tateishi et al., 2009). This is a newly emerging research area; however, much remains to be discovered regarding how epigenetic mechanisms regulate metabolism and energy homeostasis.

Histone acetylation and deacetylation are regulated by the balanced action of histone acetyltransferases and histone deacetylases (HDACs) (Backdahl et al., 2009, 16

Maunakea et al., 2010, Bannister and Kouzarides, 2011). HDACs consist of four major classes: class I (HDAC1, -2, -3, and -8), class II (HDAC4, -5, -6, -7, -9, and -10), class

III (SIRT1 to -7), and class IV (HDAC11). The class I, II, and IV HDACs act to remove acetyl groups from lysine residues in histone as well as cellular proteins, thereby regulating gene expression and cellular protein activity, whereas class III HDACs

(Sirt1–Sirt7) form a structurally distinct class of NAD-dependent enzymes and can be inhibited by nicotinamide (Backdahl et al., 2009, Bannister and Kouzarides, 2011).

Recent data suggest that HDACs have emerged as important players in the regulation of energy and glucose homeostasis (Ye, 2013). For example, it is reported that class I

HDAC inhibitor (HDACi) MS-275 ameliorates obesity and diabetes in animal models through stimulation of oxidative phosphorylation and mitochondrial function in muscle and fat (Galmozzi et al., 2013). However, the precise mechanism by which class

I HDACi exerts these effects is poorly understood.

In the current study, we have identified Hdac1 as a prominent epigenetic target in regulating the thermogenic program in brown adipocytes. Using loss- and gain-of- function approaches, we demonstrated that Hdac1 deficiency activated, whereas

Hdac1 overexpression repressed, transcription of brown adipocyte-specific gene expression through regulating the acetylation and methylation status of histone H3 lysine 27 (H3K27) on promoter and enhancer regions of Ucp1 and peroxisome proliferator-activated receptor γ (PPARγ) co-activator 1-α (PGC1α). Thus, our data suggest that epigenetics plays an important role in brown adipocyte thermogenesis, and Hdac1 may be an important regulator during this process. 17

2.3 Materials and Methods

2.3.1 Mice

C57BL/6J (B6) and AJ mice (Jackson Laboratories) were housed with a 12/12 h light–dark cycle in temperature- and humidity-controlled rooms with free access to water and food. To study the role of HDAC1 in the formation of brown adipocytes, 7- week old mice were exposed to cold condition (4°C) or intraperitoneally (i.p.) injected with a β3-adrenergic agonist [CL316243 (Sigma), dose: 1mg/kg] for seven days. After cold exposure or β3-adrenergic agonist stimulation, mice were sacrificed at eight weeks of age. Brown adipose tissue and white adipose tissues were harvested for measuring HDAC1 gene expression. All aspects of animal care were approved by

Georgia State University’s Animal Care and Use Committee.

2.3.2 Cell Culture

All cells were maintained at 37 °C with 5% CO2. Immortalized brown preadipocytes

BAT1 (kindly provided by Dr. Patrick Seale, University of Pennsylvania) (Seale et al.,

2007, Rajakumari et al., 2013) and HIB-1B (Ross et al., 1992, Klaus et al., 1994) were maintained in growth medium (DMEM/F-12 (BAT1) or DMEM (HIB-1B) containing

10% fetal bovine serum and 1% penicillin/streptomycin). Brown adipocyte differentiation was induced as described previously (Seale et al., 2007, Rajakumari et al., 2013). Briefly, cells were grown into 90% confluence in growth medium and were then induced to differentiate with a differentiation medium (growth medium plus

20nM insulin, 1nM triiodothyronine, 125μM indomethacin,

500μM isobutylmethylxanthine, and 0.5μM dexamethasone). After two days, cells 18

were cultured in maintenance medium (growth medium plus 20nM insulin and 1 nM triiodothyronine). At day 6, all of the cells were differentiated as described previously (Seale et al., 2007, Rajakumari et al., 2013). To induce the thermogenic program, brown adipocytes were treated with 1 μM isoproterenol or norepinephrine

(NE) for 3–4 h.

2.3.3 small interfering RNA (siRNA) and plasmid DNA transfection

The MGC fully sequenced mouse cDNA expression plasmids for HDAC1 (Clone ID

4976514), enhancer of zeste homologue (Ezh2) (Clone ID 3586689), suppressor of zeste 12 (Suz12) (Clone ID 6821922), ring finger protein 2 (Rnf2) (Clone ID 4021046), the ON-TARGET plus mouse HDAC1 siRNA-SMART pool (L-040287-02-0005), and HDAC1–HDAC11 siRNAs for our initial screening were purchased from GE

Healthcare. The sequences for HDAC1–HDAC11 siRNAs are listed in Table 4. All other plasmids were in mammalian expression vector pSPORT6 except for Suz12, which was in a non-expressing vector, pXY-Asc. To clone Suz12 into pSPORT6 vector, PacI and

SalI were used to release the Suz12 cDNA inserted from pXY-Asc and subcloned into pSPORT6. The overexpressing plasmids or siRNAs were transfected into BAT1 or HIB-

1B brown adipocytes by Amaxa Nucleofector II Electroporator (Lonza) using Amaxa cell line nucleofector kit L according to the manufacturer's instructions (Lonza).

Briefly, at days 4–6 of differentiation, cells (2 × 106 cells/sample) were trypsinized and centrifuged at 90 × g for 5 min at room temperature. Cell pellet was resuspended in nucleofector solution (100 μl/sample) with 2 μg of plasmid DNA or 20 pmol of siRNA and seeded into 24-well plates. Cells were treated with NE or isoproterenol two 19

days after transfection.

For the co-immunoprecipitation (co-IP) experiment, HEK293T cells were electroporated with pSPORT6 vector or co-transfected with HDAC1, Ezh2, Suz12, and Rnf2 cDNA for two days. Cell lysates from transfected HEK293T cells or endogenous BAT1 brown adipocytes were collected to detect protein interactions by co-IP.

2.3.4 Quantitative real-time RT-PCR

Total RNA was extracted with TRI Reagent according to the manufacturer's instructions (Molecular Research Center, Cincinnati, OH). RNA expression was quantified by real-time RT-PCR with TaqMan one-step RT-PCR Master mix (Life

Technologies, Inc.) using a Stratagene Mx3000p thermocycler (Stratagene, La Jolla,

CA) and normalized to cyclophilin. The primer and probe pairs used in the assays were purchased from Applied Biosystems or designed using Primer Express software (Life

Technologies) (Tables 2 and 3).

2.3.5 Immunoblotting

IP and immunoblotting were performed as we described previously (Xue et al., 2007,

Wang et al., 2011). Briefly, tissue or cell samples were homogenized in a modified radioimmune precipitation assay lysis buffer containing 50 mM Tris-HCl, 1 mM EDTA,

1% Nonidet P-40, 0.5% sodium deoxycholate, 150 mM NaCl, 1 mM sodium fluoride, 1 mMphenylmethylsulfonyl fluoride, 1 mM sodium orthovanadate, 1% protease inhibitor mixture (Sigma), and 1% phosphatase inhibitor mixture (Sigma). Cell homogenates were incubated on ice for 45 min to solubilize all proteins, and insoluble 20

portions were removed by centrifugation at 14,000 × g at 4 °C for 15 min. Two mg of cell lysates was incubated overnight with the appropriate antibodies (Table 1) and protein A/G-agarose (Santa Cruz Biotechnology, Inc.) at 4 °C with constant gentle mixing. Agarose beads were collected by centrifugation, washed with ice-cold radioimmune precipitation assay lysis buffer and then with phosphate-buffered saline, and then boiled in Laemmli sample buffer. Proteins from IP or whole cell lysates were separated by SDS-PAGE and transferred to polyvinylidene difluoride membranes. The transferred membranes were blocked, washed, and incubated with various primary antibodies (Table 1), followed by incubation with Alexa Fluor 680-conjugated secondary antibodies (Life Technologies). The fluorescent signal was visualized with a

LI-COR Imager System (LI-COR Biosciences, Lincoln, NE).

2.3.6 Chromatin immunoprecipitation (ChIP) assay

ChIP assay was performed using a ChIP assay kit (Upstate) as previously described

(Shi et al., 2006). For tissue ChIP assays, tissue samples were cut into small pieces and fixed with 1% of formaldehyde. The samples were then homogenized in cell lysis buffer

(5 mM PIPES, 85 mM KCl, and 0.5% Nonidet P-40, supplemented with protease inhibitors, pH 8.0) using a Dounce homogenizer to isolate nuclei. The nuclei were resuspended in nuclei lysis buffer (50 mM Tris-HCl, 10 mM EDTA, and 1% SDS, supplemented with protease inhibitors, pH 8.1) and sonicated to shear genomic DNA to an average fragment length of 200–1,000 bp with a Diagenode Bioruptor

(Diagenode, Denville, NJ). Lysates were centrifuged, and the supernatants were collected. Fifty μl of each sample was removed as the input control. The supernatants 21

underwent overnight immunoprecipitation, elution, reverse cross-linking, and protease K digestion, according to the manufacturer's manual. A mock immunoprecipitation without antibody was also included for each sample. Eluted

DNA was analyzed by real-time PCR using SYBR Green quantitative PCR (Life

Technologies). Primer sequences used in this study were as follows: UCP1 proximal promoter, 5′-CCCACTAGCAGCTCTTTGGA-3′ and 5′-

CTGTGGAGCAGCTCAAAGGT; UCP1 enhancer region, 5′-

CTCCTCTACAGCGTCACAGAGG-3′ and 5′-AGTCTGAGGAAAGGGTTGA-

3′; PGC1α cAMP-response element (CRE) region, 5′-

CAAAGCTGGCTTCAGTCACA-3′ and 5′-AAAAGTAGGCTGGGCTGTCA-3′.

2.3.7 Statistical Analysis

Data were expressed as mean ±S.E. Statistical tests were performed using SPSS software (version 16.0, SPSS Inc, Chicago, IL, USA). One-way Analysis of Variance

(ANOVA) followed by Student-Newman-Keuls test was used to determine multiple comparisons. Statistical significance was accepted at P<0.05.

2.4 Results 2.4.1 Hdac1 Is Enriched in WAT Versus BAT and Is Down-regulated in BAT during β-Adrenergic Stimulation and during Brown Adipocyte

Differentiation

Recent studies reported that class I but not class II HDACi enhanced whole body energy expenditure and attenuated high fat diet-induced insulin resistance through 22

increased mitochondrial biogenesis in skeletal muscle and adipose tissues (Galmozzi et al., 2013). However, little is known about whether these effects are exerted directly through activating brown fat thermogenesis and which HDAC is responsible for these beneficial effects. Thus, we first knocked down individual HDACs (Hdac1–Hdac11) in brown adipocyte cell line HIB-1B cells using a siRNA approach. Interestingly, although knockdown of other HDACs exerted minimal effects, reducing the expression of the class I HDAC family member Hdac1 by siRNA knockdown significantly enhanced NE- stimulated Ucp1 expression (Fig. 1A). This prompted us to investigate the role of

HDAC1 in regulating the brown adipocyte thermogenic program.

We first measured the HDAC1 expression pattern in brown and white adipose tissues. As expected, UCP1 protein was enriched in BAT but not detectable in epididymal WAT (Epi (WAT)) in adult mice housed at ambient temperature (Fig. 1B).

Interestingly, in contrast to UCP1 expression, HDAC1 protein level was enriched in

WAT but much lower in BAT (Fig. 1B). BAT1 and HIB-1B brown adipocyte differentiation was marked with significantly increased UCP1 mRNA and protein levels (Fig. 1C and D), which were accompanied by decreased mRNA and protein levels of HDAC1 (Fig. 1C and D).

It is well documented that cold temperature triggers sympathetic discharge, leading to the release of NE in BAT and WAT (Lowell and Spiegelman, 2000, Cannon and

Nedergaard, 2004). Thus, we tested whether Hdac1 expression in BAT was regulated by sympathetic stimulation. A/J mice were exposed to cold (5 °C) or intraperitoneally injected with β3-adrenergic agonist (CL-316,243) for up to 7 days. Hdac1 expression 23

was profoundly reduced in BAT after 6 h of cold exposure and tended to stay reduced up to 7 days of cold exposure (Fig. 1E). A similar reduction of Hdac1 expression was also observed in BAT after CL-316,243 injection for 7 days (Fig. 1F). These data indicate that decreased Hdac1 may be a marker of mature brown adipocytes and that

Hdac1 may negatively regulate the BAT thermogenic program.

2.4.2 HDAC1 Regulates Brown Adipocyte Thermogenic Gene Expression

To investigate the role of HDAC1 in the regulation of brown fat-specific gene expression, we performed gain- or loss-of-function experiments in differentiated BAT1 cells. It has been reported that HDAC1 regulates the early steps of adipocyte differentiation (Wiper-Bergeron et al., 2003). Thus, to avoid the confounding effects of HDAC1 on differentiation, we have focused on the role of HDAC1 in mature brown adipocyte gene expression by knocking down or overexpressing HDAC1 in BAT1 cells after 4–6 days of differentiation. HDAC1 mRNA in knockdown cells was reduced by

80%, as measured by real-time RT-PCR, and HDAC1 protein expression was also significantly decreased, as assessed by immunoblotting (Fig. 2A). As expected, knocking down HDAC1 in mature BAT1 cells did not affect the mRNA levels of general adipocyte differentiation markers, including PPARγ, CCAAT/enhancer-binding protein α (CEBPα), sterol regulatory element-binding protein 1C (SREBP1c), adipocyte protein 2 (aP2), and adiponectin (AdipoQ) (Fig. 2B). Remarkably, HDAC1 knockdown in BAT1 cells significantly increased basal and isoproterenol-stimulated expression of brown adipocyte-specific genes, including UCP1, PGC1α (Fig. 2C),

PGC1β, PR domain-containing protein 16 (PRDM16), PPARα, type 2 deiodinase 24

(Dio2), acyl-CoA oxidase 1 (ACOX1), cytochrome c oxidase 1 (COX1), epithelial V-like antigen 1 (EVA1), and otopetrin 1 (Otop1) (Fig. 2D). In addition, isoproterenol significantly stimulated UCP1 protein expression in BAT1 cells, which was further enhanced by HDAC1 knockdown (Fig. 2E).

We then overexpressed HDAC1 in BAT1 brown adipocytes, as shown by increased

HDAC1 protein expression (Fig. 3A). Overexpression of HDAC1 in BAT1 brown adipocytes significantly suppressed basal and/or isoproterenol-stimulated brown adipocyte-specific gene expression, including UCP1, PGC1α, PGC1β, PRDM16, CEBPβ, fibroblast growth factor 21 (FGF21), PPARα, carnitine palmitoyltransferase 1B

(CPT1B), ACOX1, COX1, Otop1, and cell death-inducing DNA fragmentation factor α subunit-like effector A (CIDEA) (Fig. 3B and C).

Similar results were observed in HIB-1B brown adipocyte cell line. In HIB-1B cells with HDAC1 siRNA knockdown, HDAC1 mRNA was reduced by more than 80% (Fig.

4A), and knocking down HDAC1 significantly increased basal and/or NE-stimulated brown adipocyte-specific gene expression, including UCP1 (Fig. 4B), PGC1α, PGC1β,

COX1, ACOX1, CIDEA, PRDM16, CPT1B, and PPARα (Fig. 4C and D). Similarly,

HDAC1 overexpression significantly reduced NE-stimulated UCP1 expression in HIB-

1B cells (Fig. 4E). These data demonstrate that HDAC1 is a negative regulator of the thermogenic program in mature brown adipocytes.

2.4.3 HDAC1 Deficiency Mediates the Beneficial Effects of Class I HDACi on Improving Metabolic Phenotypes

It is reported that the pan-HDACi SAHA and class I HDACi MS-275, but not class II 25

HDACi MC-1568, ameliorate obesity and diabetes in animal models, possibly through stimulation of oxidative phosphorylation and mitochondrial function in muscle and fat (Galmozzi et al., 2013). To our surprise, treating BAT1 brown adipocytes with the pan-HDACi TSA and SAHA resulted in complex effects on brown fat-specific gene expression. Whereas TSA and SAHA enhanced isoproterenol-stimulated expression of genes involved in mitochondria oxidative activity, such as PGC1α and ACOX1, they significantly suppressed isoproterenol-stimulated expression of other BAT-specific genes, including UCP1, PRDM16, PPARγ, and Otop1 (Fig. 5A and B). In contrast, the class I HDACi MS-275 significantly enhanced isoproterenol-stimulated BAT-specific gene expression, including UCP1, ELOVL3, and PGC1α, whereas it had minimal effects on PRDM16 and PPARγ expression (Fig. 5 C).

We also tested whether HDAC1 might be the molecular target of class I HDACi MS-

275 in regulating BAT-specific gene expression in BAT1 cells. As expected, MS275 or

HDAC1 siRNA knockdown enhanced isoproterenol-stimulated gene expression of

UCP1 and PGC1α in BAT1 cells to a similar extent when treated individually; however, they did not exert any additive effects on these gene expressions when treated in combination in BAT1 cells (Fig. 5D). Our data suggest that HDAC1 may be the molecular target of the class I HDACi MS-275 in stimulating brown fat function.

2.4.4 Sympathetic Activation in Brown Adipocytes Reduces HDAC1

Binding and Increases H3K27 Acetylation at UCP1 and PGC1α Promoters

HDACs repress gene expression by removing acetyl groups from lysine residues in histone proteins at specific gene promoters (Bannister and Kouzarides, 2011). We 26

reasoned that if HDAC1 indeed negatively regulates brown fat functions, sympathetic activation in BAT may trigger HDAC1 dissociation from promoter regions of key genes regulating brown fat thermogenesis, which should allow transcriptional activation of these genes and subsequently the full activation of the brown adipocyte thermogenic program. UCP1 is responsible for mitochondrial uncoupling in BAT (Cannon and

Nedergaard, 1985, 2004), whereas PGC1α is the central regulator of brown adipocyte thermogenesis (Puigserver and Spiegelman, 2003, Lin et al., 2005). A ∼220-bp upstream enhancer element along with the proximal promoter that is closer to the transcriptional start site have been identified at the UCP1 promoter that mediates sympathetic stimulated and tissue-specific expression of UCP1 (Kozak et al., 1994,

Cassard-Doulcier et al., 1998, Rim and Kozak, 2002). In addition, PGC1α expression is highly induced via the β-adrenergic receptor-activated protein kinase A (PKA)/CRE- binding protein (CREB) pathway (Herzig et al., 2001, Cao et al., 2004) through the

CRE cis-element on the PGC1α promoter (Handschin et al., 2003, Galmozzi et al.,

2013). We therefore examined the binding of HDAC1 at the enhancer and proximal promoter regions of UCP1 and the CRE cis-element on the PGC1α promoter (Fig. 6A) upon sympathetic activation. As expected, in BAT of mice treated with β3-adrenergic agonist CL-243,316, binding of HDAC1 was significantly reduced at the enhancer and proximal promoter regions of UCP1 (Fig. 6B and C) and the CRE cis-element region at the PGC1α promoter (Fig. 6D). In addition, isoproterenol treatment also significantly reduced HDAC1 binding to these promoter/enhancer regions in differentiated BAT1 brown adipocytes (Fig. 6E). These data indicate that β-adrenergic stimulation in 27

brown adipocytes dissociates HDAC1 from promoters of key genes regulating brown adipocyte thermogenesis, including UCP1 and PGC1α. This may facilitate subsequent chromatin modifications and DNA unwinding, which can eventually lead to increased accessibility to transcription factor binding and gene activation (Maunakea et al., 2010,

Bannister and Kouzarides, 2011).

Acetylation at histone H3 lysine 27 (H3K27ac) results in transcriptional activation

(Ge, 2012). We found that isoproterenol treatment in BAT1 cells significantly increased H3K27ac levels at the enhancer and proximal promoter regions of UCP1 and the CRE region at the PGC1α promoter (Fig. 6F and G). Importantly, reducing HDAC1 expression by siRNA knockdown induced a similar increase in H3K27ac levels and further enhanced isoproterenol-stimulated increase in H3K27ac levels (Fig. 6F), whereas HDAC1overexpression completely blocked isoproterenol-induced increase in

H3K27ac levels at these promoter/enhancer regions (Fig. 6G).

On the other hand, the acetylation levels at histone H3 lysine 14 (H3K14ac) and histone H3 lysine 9 (H3K9ac), two other lysine residues that can be modified by acetylation (An, 2007, Bannister and Kouzarides, 2011), did not exhibit similar responsiveness to HDAC1 knockdown and/or isoproterenol stimulation in BAT1 cells

(Fig. 6H). Thus, our data suggest that H3K27 may be the major target of HDAC1 in brown adipocytes regulating brown-specific gene expression.

2.4.5 HDAC1 Regulates UCP1 and PGC1α Expression by Further Modifying

H3K27 Methylation

Except for acetylation, histone lysine residues can also be mono-, di-, and 28

trimethylated and are associated with either gene repression or activation, depending on the lysine residues that are methylated and the degree of methylation.

Trimethylation of H3K27 (H3K27me3) is a hallmark of gene repression, whereas trimethylation of histone H3 Lys-4 marks transcriptional activation (Maunakea et al.,

2010, Bannister and Kouzarides, 2011). Importantly, histone acetylation and methylation mutually affect each other in the regulation of transcriptional process (An,

2007). We thus tested whether modulation of histone acetylation by HDAC1 led to further alterations in histone methylation. ChIP assay analysis demonstrated that isoproterenol treatment in BAT1 brown adipocytes resulted in a significant decrease in H3K27me3 levels at UCP1 and PGC1α promoters (Fig. 7A and B).

Reducing HDAC1 expression by siRNA knockdown mimicked isoproterenol's effect by reducing H3K27me3 levels at these promoter/enhancer regions to a similar extent as isoproterenol and did not further decrease isoproterenol-suppressed H3K27me3 levels (Fig. 7A). In contrast, HDAC1 overexpression significantly blocked isoproterenol-suppressed H3K27me3 levels at these promoter/enhancer regions (Fig.

7 B).

H3K27 methylation is dynamically regulated by both histone methyltransferases and demethylases (Dambacher et al., 2010, Morey and Helin, 2010, Bannister and

Kouzarides, 2011). EZH2 is a methyltransferase that specifically di- and trimethylates

H3K27 (Dambacher et al., 2010, Morey and Helin, 2010, Bannister and Kouzarides,

2011), whereas lysine-specific demethylase 6A (KDM6A)/ubiquitously transcribed tetratricopeptide repeat on chromosome X (UTX) is a di- and trimethyl-H3K27 29

demethylase (Lee et al., 2007). A ChIP assay demonstrated that isoproterenol treatment in BAT1 brown adipocytes resulted in a significant decrease in EZH2 binding at UCP1 and PGC1α promoters (Fig. 7C and D). Reducing HDAC1expression by siRNA knockdown mimicked isoproterenol's effect by reducing EZH2 binding at these promoter/enhancer regions to a similar extent as isoproterenol and did not further decrease isoproterenol-suppressed EZH2 binding (Fig. 7C), whereas HDAC1 overexpression completely reversed isoproterenol-induced suppression of EZH2 binding (Fig. 7D).

In contrast, isoproterenol treatment in BAT1 brown adipocytes resulted in a significant increase in UTX binding at UCP1 and PGC1α promoters (Fig. 7E and F).

Reducing HDAC1expression by siRNA knockdown also induced a similar increase in

UTX binding to the UCP1 enhancer region and further enhanced isoproterenol- induced binding of UTX to the UCP1 proximal promoter and PGC1α CRE region (Fig.

7E), whereas HDAC overexpression completely prevented the isoproterenol-induced increase in UTX binding to these promoter/enhancer regions (Fig. 7F). Overall, our data demonstrated that HDAC1 may coordinately regulate H3K27ac and H3K27me3 levels at UCP1 and PGC1α promoters through differential recruitment of the H3K27 methyltransferase EZH2 and the H3K27 demethylase UTX, thus regulating the expression of these genes.

EZH2 is a component of the polycomb group proteins, which are known to mediate gene silencing by regulating chromatin structure (Morey and Helin, 2010). Two major polycomb group proteins exist in mammals, namely polycomb repressive complex 1 30

and 2 (PRC1 and PRC2). PRC1 comprises three main components: ring finger protein

1 (RING1), RNF2, and Bmi1 polycomb ring finger oncogene (BMI1). PRC2 comprises three major components: EZH2, SUZ12, and EED (embryonic ectoderm development).

PRC2 promotes H3K27 methylation through the methyltransferase activity of EZH2 and also facilitates the recruitment of PRC1 onto methylated H3K27, which in turn leads to further gene repression (Morey and Helin, 2010). It has been shown that

HDAC1 may be associated with PRC2 (Morey and Helin, 2010). We thus investigated whether HDAC1 regulates the recruitment of components of PRC1/2 to UCP1 and PGC1α promoters. The ChIP assay demonstrated that isoproterenol treatment in BAT1 cells decreased binding of SUZ12 (Fig. 8A and C) and RNF2 (Fig.

8B and D) to UCP1 and PGC1α promoters. Importantly, reducing HDAC1 expression by siRNA knockdown mimicked isoproterenol's effect by inducing a similar decrease in SUZ12 and RNF2 binding and did not further reduce isoproterenol-suppressed

SUZ12 and RNF2 binding at these promoters (Fig. 8A and B), whereas HDAC1 overexpression completely prevented isoproterenol-suppressed

SUZ12 and RNF2 binding to these promoters (Fig. 8C and D).

We further performed co-IP experiments to test whether HDAC1 physically interacts with components of PRC1 and PRC2 complexes. Our data demonstrated that HDAC1 interacts with the PRC2 complex components EZH2 and SUZ12 and the PRC1 component RNF2 in HEK293T cells overexpressing HDAC1, EZH2, SUZ12, and RNF2 (Fig. 8E) and, most importantly, in endogenous BAT1 brown adipocytes

(Fig. 8F). 31

2.5 Discussion of specific aim 1 Recent data suggest that HDACs have emerged as important players in the regulation of energy and glucose homeostasis (Ye, 2013). The pan-HDAC inhibitors sodium butyrate and trichostatin A increase energy expenditure, reduce adiposity, and improve insulin sensitivity in diet-induced obese (DIO) mice (Gao et al., 2009). This is possibly exerted through the inhibition of class I HDACs, because the specific class

I HDACi exerts similar effects in DIO mice, whereas class II HDACi has no effects

(Galmozzi et al., 2013). However, the precise mechanism by which class I HDACi exerts these beneficial effects is poorly understood. Here, we demonstrate that the class I HDAC1 negatively regulates the brown adipocyte thermogenic program. This is based on the following observations. First, in contrast to UCP1 expression, which is usually enriched in BAT, HDAC1 is enriched in WAT, but its expression is much lower in BAT. Second, brown adipocyte differentiation is associated with significantly upregulated UCP1 expression, which is concomitantly associated with reduction of HDAC1 RNA and protein levels. Third, sympathetic activation of BAT is associated with down-regulation of HDAC1 expression. These data suggest that HDAC1 may be a negative regulator of brown adipocyte thermogenic program. Indeed, knocking down HDAC1 in brown adipocytes significantly up-regulates BAT-specific gene expression, whereas overexpressing HDAC1 significantly suppresses BAT-specific gene expression, suggesting a causative link between HDAC1 inhibition and up- regulation of the brown adipocyte thermogenic program. Activating brown/beige 32

adipocyte thermogenesis alleviates obesity and its associated metabolic diseases

(Himms-Hagen et al., 1994, Guerra et al., 1998, Cederberg et al., 2001, Feldmann et al., 2009, Cohen et al., 2014). Thus, our data suggest that HDAC1 may be important in regulating energy homeostasis in animal models, and inhibiting HDAC1 in brown adipocytes may contribute to the beneficial effects of class I HDACi in DIO mice

(Galmozzi et al., 2013). Further studies using genetic animal models with brown fat- specific deletion or overexpression of HDAC1 will be needed to study the role of

HDAC1 in energy homeostasis and metabolism in whole animals.

Interestingly, we found that whereas both the pan-HDACi TSA and SAHA and the class I-specific HDACi MS-275 (Chavan and Somani, 2010) stimulate the expression of genes involved in mitochondrial function and oxidative activity, such as PGC1α, the effect of these different classes of HDACis on the expression of other genes that are important for brown adipocyte differentiation and determination is different. For example, the class I HDACi MS-275 potently stimulates, whereas the pan-HDACis TSA and SAHA significantly suppress, isoproterenol-stimulated UCP1 expression in brown adipocytes. In addition, TSA also significantly inhibited PRDM16 and PPARγ expression, whereas MS-275 had minimal effects on these gene expressions. These

TSA-suppressed genes are either key players in brown adipocyte differentiation and determination (PPARγ and PRDM16) or a marker of brown adipocyte (UCP1). Because

TSA and SAHA inhibit both class I and II HDACs (Chavan and Somani, 2010), and because it has been shown that class II HDACs are mainly involved in the regulation of cellular differentiation (Nebbioso et al., 2010), the difference in the gene regulation 33

between the pan-HDACi TSA and the class I HDACi MS-275 may be mainly due to the different functions of class I and class II HDACs. Thus, further studies are warranted to elucidate the differential effects of class I and II HDACis on brown adipocyte function. Nonetheless, our data suggest that the class I HDACi MS-275 may contribute to its beneficial effect on improving metabolic phenotypes in DIO mice (as reported by

Galmozzi et al. (Galmozzi et al., 2013)) at least partly by directly stimulating brown adipocyte function, and HDAC1 may mediate this effect. Thus, class I HDACis, such as

MS-275, and, more specifically, HDAC1, may serve as promising therapeutic targets in treating obesity and associated metabolic syndrome.

Histone acetylation and methylation are epigenetic mechanisms that regulate gene expression by remodeling chromatin structure (Backdahl et al., 2009, Maunakea et al.,

2010, Bannister and Kouzarides, 2011). Whereas H3K27ac is a histone transcriptional activation mark, H3K27me3 serves as a histone transcriptional repression mark

(Backdahl et al., 2009, Maunakea et al., 2010, Bannister and Kouzarides, 2011, Ge,

2012). Histone acetylation is regulated by the balanced action of histone acetyltransferases and HDACs, whereas histone methylation is regulated by histone methyltransferases and demethylases (Backdahl et al., 2009, Maunakea et al., 2010,

Bannister and Kouzarides, 2011). Specifically, the di- and trimethylation of H3K27 are catalyzed by the PRC2 complex. Within PRC2, EZH2 is the catalytic subunit that possesses methyltransferase activity toward H3K27, and its activity also requires binding to the two other PRC2 protein components, SUZ12 and EED (Morey and Helin,

2010). The di- and trimethylated H3K27 further recruits PRC1 to the target genes, 34

which, through the E3 ubiquitin ligase activity of the PRC1 components RING1 and

RNF2, induces further chromatin compaction through histone 2A ubiquitination

(Morey and Helin, 2010). On the other hand, UTX is a histone demethylase that specifically demethylates di- and trimethylated H3K27 (Lee et al., 2007). Interestingly, we find that HDAC1 physically interacts with the PRC2 components EZH2 and SUZ12 and the PRC1 catalytic subunit RNF2, suggesting a role of HDAC1 in connecting

H3K27 deacetylation to methylation and possibly further chromatin compaction by histone ubiquitination. Indeed, we find that knocking down HDAC1 in brown adipocytes not only results in an increase in H3K27ac; it also leads to reduced

H3K27me3 in promoters of BAT-specific genes, including UCP1 and PGC1α. This is exerted through the dissociation of the PRC2 components EZH2 and SUZ12 and PRC1 catalytic subunit RNF2 from these promoters and a reciprocal recruitment of the

H3K27 demethylase UTX to these promoters. Thus, our data demonstrate that HDAC1 negatively regulates the brown adipocyte thermogenic program through interaction with PRC1/2 complexes, thereby promoting H3K27 deacetylation and methylation at promoters of BAT-specific genes, such as UCP1 and PGC1α. In addition, our data also demonstrate that β-adrenergic activation induces dissociation of HDAC1 along with

PRC complexes from UCP1 and PGC1α promoters, which in turn leads to gene activation.

Sympathetic signaling is essential in brown/beige adipocyte activation (9–13). We find that β-adrenergic activation in BAT1 adipocytes is associated with increased

H3K27ac and decreased H3K27me3 through dissociation of HDAC1 and PRC 35

complexes and concomitant recruitment of UTX on BAT-specific gene promoters, including UCP1 and PGC1α. These data indicate that HDAC1 may be one of the epigenetic modulations triggered by sympathetic signaling in brown adipocytes, which eventually leads to thermogenic activation. It would be interesting to study how β- adrenergic stimulation triggers the dissociation of HDAC1 from BAT-specific gene promoters in brown adipocytes. HDAC1 is a part of the catalytic core of several multimeric corepressor complexes, including SIN3A, NuRD (nucleosome remodeling deacetylase), and CoREST (corepressor of RE1-silencing transcription factor) (Segre and Chiocca, 2011, Kelly and Cowley, 2013, Laugesen and Helin, 2014), and is also a part of the PRC2 complex (Morey and Helin, 2010). The recruitment of these multiprotein complexes is usually triggered by cell-specific transcriptional factors or the histone recognition motifs found within the complex components (Segre and

Chiocca, 2011, Kelly and Cowley, 2013, Laugesen and Helin, 2014). In this context, pRB (retinoblastoma protein) and RIP140 (receptor-interacting protein 140) are potent negative regulators of BAT-specific gene expression (Brehm et al., 1998,

Sharma et al., 2014). It has been reported that pRB and RIP140 silence promoter activity and gene expression through recruitment of HDACs, including HDAC1

(Brehm et al., 1998, Magnaghi-Jaulin et al., 1998, Wei et al., 2000). Thus, it is possible that HDAC1 may be recruited to BAT-specific promoters through association with negative transcriptional regulators of BAT, such as pRB and RIP140.

In addition, HDAC1 itself is subjected to various post-transcriptional modifications.

For example, phosphorylation of HDAC1 by casein kinase II up-regulates its activity, 36

whereas acetylation of HDAC1 by the acetyltransferase p300 suppresses its activity

(Segre and Chiocca, 2011). Interestingly, recent studies using phosphoproteomics have identified casein kinase II as a negative regulator of BAT function through phosphorylating and regulating HDAC1 activity (Shinoda et al., 2015b). Thus, our data fall in line with the results from Shinoda et al.(Shinoda et al., 2015b) and further demonstrate the importance of HDAC1 in the regulation of BAT function. Moreover,

HDAC1 activity or protein levels can also be regulated by ubiquitination, sumoylation, nitrosylation, and carbonylation (Segre and Chiocca, 2011). Further study is required to decipher the cellular signaling cascades and mechanisms that regulate HDAC1 activity and recruitment to BAT-specific promoters in response to sympathetic and other stimuli in brown adipocytes.

In the present study, we have focused on the epigenetic regulation of PGC1α by HDAC1, because it is well established that PGC1α plays a central role in regulating important pathways involved in mitochondrial biogenesis and thermogenesis

(Puigserver and Spiegelman, 2003, Lin et al., 2005). However, we have found that

HDAC1 regulates H3K27 acetylation and methylation at the promoter and enhancer regions of UCP1, a brown adipocyte terminal differentiation marker. Thus, it would be interesting to know whether HDAC1 also regulates H3K27 deacetylation and methylation at other genes' promoters to regulate their transcription. Unbiased approaches, such as ChIP sequencing, will be required to explore the gene profile that

HDAC1 regulates through H3K27 deacetylation. 37

Although both express UCP1 and share striking similarities in morphological and biological properties, traditional brown fat and inducible beige adipocytes are derived from distinct cell origins during embryonic development (Seale et al., 2008, Wu et al.,

2012, Long et al., 2014). In rodents, traditional brown adipocytes are originated from the skeletal muscle lineage (Seale et al., 2008), whereas at least a subset of beige cells arise from the smooth muscle origin (Long et al., 2014). A recent study shows that human brown adipocytes possess molecular features similar to those of rodent beige cells (Shinoda et al., 2015a). We used the brown fat cell BAT1 in this study. Thus, it is not clear whether the role of HDAC1 in regulating brown adipocyte function can be extrapolated to beige cells. Additional studies, involving knockdown or overexpression of HDAC1 in beige lineage cells, will be warranted to determine the role of HDAC1 in the regulation of beige cell function.

In the present study, we investigated the role of HDAC1 in brown fat gene expression at the mature adipocyte stage. It has been reported that HDAC1 inhibits white adipocyte differentiation by deacetylating H4 at the promoter of C/ebpα, an important regulator of adipogenesis (Wiper-Bergeron et al., 2003). In addition, we find that

HDAC1 RNA and protein levels are down-regulated during brown adipocyte differentiation. Thus, it would be interesting to study whether HDAC1 also regulates brown adipogenesis and subsequently affects the brown fat thermogenic program.

In summary, we have identified HDAC1 as a negative regulator of brown adipocyte thermogenic program. Our data show that HDAC1 is down-regulated during brown adipocyte differentiation and is also suppressed by sympathetic activation in BAT. 38

Overexpressing HDAC1 blocks, whereas knocking down HDAC1 further enhances, β- adrenergic agonist-stimulated BAT-specific gene expression in brown adipocytes.

Remarkably, HDAC1 physically interacts with PRC1/2 complexes, and activation of β- adrenergic signaling dissociates HDAC1 along with PRC1/2 complexes from UCP1 and PGC1α promoters and concomitantly recruits UTX to these promoters, leading to increased H3K27 acetylation and decreased H3K27me3 levels in these promoters. These coordinated changes switch the transcriptional repressive state to the transcriptional active state at the promoters of UCP1 and PGC1α, which in turn activates the brown thermogenic program. Thus, our data demonstrate that HDAC1 negatively regulates brown adipocyte gene expression by coordinated regulation of

H3K27 deacetylation and methylation, and inhibiting HDAC1 promotes the brown adipocyte thermogenic program. Targeting HDAC1 may be a novel therapeutic target in the treatment of obesity by promoting brown adipocyte thermogenesis and energy dissipation.

2.6 Figures of specific aim 1

Figure 1 The expression pattern of HDAC1.

A, reducing HDAC1 expression by siRNA knockdown in HIB-1B brown adipocytes up-regulates norepinephrine (NE)-stimulated UCP1 expression. HIB-1B cells were transfected with scramble siRNA or siRNAs targeting individual HDACs. After 48 h, cells were treated with or without 1 µM NE for 4 h, and RNA was isolated for gene expression measurements. Class I HDACs are shown in blue. B, UCP1 and HDAC1 protein levels in BAT and epididymal (Epi) WAT from C57BL/6J (B6) mice. C and D, UCP1 and HDAC1 mRNA (C) and protein (D) levels in BAT1 and HIB-1B brown adipocytes during differentiation. E and F, HDAC1 expression in A/J mouse BAT tissues exposed to 4 °C at the indicated time (E) or treated with β3-agonist CL-316,243 for seven days. Data are expressed as mean ±S.E. (error bars). *p< 0.05 versus undifferentiated samples (C), time 0 (E), or control (Ctrl) (F). 40

Figure 2 Reducing HDAC expression promotes Brown-specific gene expression in BAT1 brown adipocytes.

BAT1 brown adipocytes were transfected with scramble or HDAC1 siRNA. Two days later, cells were treated with isoproterenol (Isop; 1µM) for 3 h, and RNA and protein were isolated for gene expression and protein level measurements. A, HDAC1 mRNA and protein levels in scramble or HDAC1 siRNA-transfected cells. B–D, the mRNA levels of various genes in scramble or HDAC1 siRNA-transfected cells. n=6–8/group. E, UCP1 protein levels in scramble or HDAC1 siRNA-transfected cells. Left, UCP1 immunoblot; right, quantitation of UCP1 protein levels normalized to α- tubulin. n=3/group. Data are expressed as mean ±S.E. (error bars). *p <0.05. n.s., not statistically significant. 41

Figure 3 HDAC1 overexpression attenuates mRNA expression of Brown- specific genes.

BAT1 brown adipocytes were transfected with pSPORT vector or HDAC1 cDNA. Two days later, cells were treated with isoproterenol (isop; 1 µM) for 3 h, and RNA and protein were isolated for gene expression and HDAC1 protein level measurements. A, HDAC1 protein levels in BAT1 cells transfected with pSPORT vector or HDAC1 cDNA. B and C, expression levels of BAT-specific genes in BAT1 cells transfected with pSPORT vector or HDAC1 cDNA. Data are expressed as mean ±S.E. (error bars) n=6– 8/group. *p<0.05. n.s., not statistically significant. 42

Figure 4 HDAC1 negatively regulates Brown-specific gene expression in HIB-1B brown adipocytes.

A–D, reducing HDAC1 expression by siRNA knockdown in HIB-1B cells promotes Brown-specific gene expression. HIB-1B brown adipocytes were transfected with scramble or HDAC1 siRNA. Two days later, cells were treated with 1µM NE for 4 h, and RNA was isolated for gene expression analysis, including HDAC1 (A); UCP1 (B); PGC1α, PGC1β, Cox1, and Acox1 (C); and CIDEA, PRDM16, CPT1B, and PPARα (D). E, HDAC1 overexpression attenuates UCP1 mRNA levels. HIB-1B cells were transfected with pSPORT vector or HDAC1-overexpressing plasmids. Two days later, cells were treated with 1 µM NE for 4 h, and mRNA levels of UCP1 were measured by RT-PCR. Data are expressed as mean ±S.E.(error bars). n = 6–8 in A–D and 4–6 in E. *p <0.05. n.s., not statistically significant. 43

Figure 5 HDAC1 mediates the effects of class I HDACi MS-275 on BAT-1 brown adipocyte gene expression.

A–C, isoproterenol (Isop)-stimulated gene expression in BAT1 cells treated with different concentration of pan-HDACi TSA (A), SAHA (B), and the class I HDACi MS-275 (C). D, isoproterenol-stimulated gene expression in BAT1 cells treated with MS-275 (5 μM) or HDAC1 siRNA individually or in combination. Data are expressed as mean ± S.E. (error bars). n = 6. *p<0.05.

Figure 6 HDAC1 regulates H3K27ac at PGC1α and UCP1 promoters.

A, schematic illustration of promoter/enhancer regions of UCP1 and PGC1α genes. B–D, β3-adrenergic agonist CL-316,243 treatment dissociates HDAC1 from UCP1 and PGC1α promoters in BAT of A/J mice. A/J mice were treated with CL-316.243 for the indicated time, and BAT was collected for a ChIP assay to measure HDAC1 binding to UCP1 and PGC1α promoters as described under “Materials and Methods.” E, isoproterenol treatment dissociates HDAC1 from UCP1 and PGC1α promoters in BAT1 brown adipocytes. BAT1 cells were treated with isoproterenol (Isop; 1 μM) for 3 h. Cells were collected, and ChIP assays were performed to measure HDAC1 binding to UCP1 and PGC1α promoters. F and G, H3K27ac levels in UCP1 and PGC1α promoters in BAT1 brown adipocytes. In F, BAT1 cells were transfected with scramble or HDAC1 siRNA for two days; in G, BAT1 cells were transfected with pSPORT6- or pSPORT6-HDAC1-overexpressing plasmids for two days. Cells were then treated with 1 μM isoproterenol for 3 h. A ChIP assay was performed as described under “Materials and Methods.” H, histone H3 lysine 14 (H3K14ac) and histone H3 lysine 9 (H3K9ac) levels in UCP1 and PGC1α promoters in BAT1 cells transfected with scramble or HDAC1 siRNA. Data are expressed as mean±S.E. (error bars). n = 4–6. *p<0.05. n.s., not statistically significant. 45

Figure 7 HDAC1 regulates trimethylation of H3K27me3 at UCP1 and PGC1α promoters through coordinated regulation of EZH2 and UTX binding to BAT-specific promoters.

A–F, H3K27me3 levels (A and B), EZH2 (C and D), and UTX binding (E and F) at UCP1 and PGC1α promoters in BAR1 cells. BAT1 brown adipocytes were transfected with scramble or HDAC1 siRNA (A, C, and E) or pSPORT6- or pSPORT6-HDAC1-overexpressing plasmids (B, D, and F). Two days later, cells were treated with or without isoproterenol (Isop; 1 μM) for 3 h, and a ChIP assay was performed as described under “Materials and Methods.” Data are expressed as mean ± S.E. (error bars). n = 4–6. *p < 0.05. n.s., not statistically significant. 46

Figure 8 HDAC1 regulates the recruitment of polycomb repressor complexes to UCP1 and PGC1α promoters.

A–D, SUZ12 and RNF2 binding at UCP1 and PGC1α promoters in BAT1 cells. BAT1 brown adipocytes were transfected with scramble or HDAC1 siRNA (A and B) or pSPORT6- or pSPORT-HDAC1-overexpressing plasmids (C and D). Two days later, cells were treated with or without isoproterenol (Isop; 1 μM) for 3 h, and a ChIP assay was performed as described under “Materials and Methods.” E and F, HDAC1 interacts with EZH2, SUZ12, and RNF2. In E, HEK293T cells were transfected with pSPORT vector or co-transfected with HDAC1, EZH2, SUZ12, and RNF2 cDNA. Two days later, cell lysates were collected to detect protein interactions by co-IP. Left, overexpression of HDAC1, EZH2, SUZ12, and RNF2 was verified by immunoblotting using whole cell lysates. Right, IP of HDAC1 pulls down EZH2, SUZ12, and RNF2. In F, the interaction between HDAC1 and components of the polycomb repressor complexes in endogenous BAT1 cells was measured by co-IP. Data are expressed as mean ± S.E. (error bars) n = 4–6 in A–D. Blots in E and F are representative of two or three independent experiments. *p<0.05. n.s., not statistically significant. IB, immunoblotting. 47

2.7 Tables of specific aim 1 Table 1 Antibodies used in immunoblotting and ChIP-qPCR

Table 2 Primer/probe sequence for gene expression experiments

Table 3 Primer/probe sets for gene expression experiments

Table 4 Sequence of siRNA smart pools for individual HDACs

3 SPECIFIC AIM 2

3.1 Abstract High-fat diet induces obesity and related metabolic diseases. Activation of brown adipocyte tissue (BAT) thermogenic function is important for energy homeostasis.

Here, we demonstrated that HFD dynamically remodeled BAT, inducing BAT thermogenesis at the early stage, whereas causing a BAT-to-skeletal muscle switching at the late stage. DNA methyltransferases (DNMTs) are important enzymes regulating

DNA methylation, an epigenetic mark that repress gene expression involved in cell differentiation. Here, we demonstrated that DNMT1 mimicked HFD-induced obesity and BAT remodeling. BAT-specific DNMT1 deletion led to brown fat remodeling and switching to a muscle-like phenotype, which may be a maladaptive mechanism contributing to obesity-induced metabolic diseases including type 2 diabetes. Our data suggest that the normal function of DNMT1 may be to restrict myogenic gene expression and maintain brown fat function in brown adipocytes. As expected, brown adipocyte DNMT1 KO mice exhibited cold intolerance due to BAT dysfunction. In addition, DNMT1 KO mice were obese even on a regular chow diet, and developed

HFD-induced obesity, primarily due to reduced whole-body energy expenditure. This was associated with profound insulin resistance and glucose intolerance. In addition, we found that DNMT1 regulated brown thermogenesis through targeting Myod1, which is an important transcriptional factor involved skeletal muscle cell development.

DNMT1 KO reduced DNA methylation at Myod1 gene promoter and increased its gene transcription. Myod1 knockdown reversed DNMT1 siRNA-mediated inhibition of 52

brown fat-specific gene expression including UCP1. Thus, our results establish brown adipocyte DNMT1 as a positive regulator of brown adipocyte thermogenesis by inhibiting myogenic transcriptional factor Myod1.

3.2 Introduction DNA methylation at cytosine-guanine dinucleotide (CpG) sites is one of the most common epigenetic modification (Paluch et al., 2016). Multiple enzymes regulate DNA methylation and demethylation status. For example, de novo DNA methyltransferases

DNMT3A and DNMT3b initiate DNA methylation, and DNMT1 maintains established DNA methylation. (Ko et al., 2005). It is recently uncovered that TET proteins including Tet1,

Tet2 and Tet3 play essential roles in DNA demethylation through catalyzing conversion of 5-methylcytosine (5mc) to 5-hydroxymethylcytosine (5hmc), 5- formylcytosine (5fC) and 5-carboxylcytosine (5caC) in a step-wise oxidation reactions.

5-fC and 5-caC can then be converted back to the unmethylated cytosine (Tahiliani et al., 2009, Pronier et al., 2011). Hypo-methylated promoter regions of a gene result in its gene activation, whereas hyper-methylated gene promoter leads to gene silence.

For example, UCP1 gene transcription is associated with its reduced methylation levels at CpG sites within the promoter enhancer region (Shore et al., 2010). On the one hand, the repressive effects of DNA methylation on gene transcription is partially due to its interference with the binding of certain transcriptional factors to their gene promoters.

On the other hand, it might be due to methylated CpGs recruiting repressive enzymes such as histone deacetylases (HDACs) binding to active elements of genes. It is 53

reported that DNMT1 activity is associated with HDAC activity, and histone acetylation is associated with DNA methylation inhibition. These data indicate the cross-talk between DNA methylation and histone acetylation.

DNA methylation level at gene promoter is influenced by aging and environmental triggers such as diet and cold exposure (Ko et al., 2005). Through associated chromatin changes, such epigenetic modification can affect human diseases such as obesity. It is important to understand whether alterations of genes encoding enzymes related to epigenetic modification are associated with metabolic diseases such as obesity. It is demonstrated that DNMT1 gene expression in white adipocytes is upregulated in db/db mice and in wild-type mice fed high-fat diet (Xia et al., 2014,

Kim et al., 2015). Moreover, DNMT1 gene expression in human white adipocytes positively correlates with body mass index (Sharma et al., 2015). These findings suggest that DNMT1 may play a crucial role in adipocyte differentiation. However, there is no study to identify the role of DNMT1 in the regulation of brown adipocyte development and activity.

In this study, we found that prolonged HFD feeding induced reprogramming of brown adipocyte into skeletal muscle-like cells. In addition, we identified DNMT1 as an important epigenetic mechanism in regulating the brown fat reprogramming during HFD feeding. Using Cre/lox recombination system, we specifically deleted

DNMT1 in mature brown adipocytes (DNMT1f/f-UCP1-Cre+/-). We found that

DNMT1 deletion mimicked HFD-induced obesity and insulin resistance. We also demonstrated that brown adipocyte-specific DNMT1 deficiency repressed 54

transcription of brown adipocyte-specific genes, whereas activated transcription of myogenic genes through targeting DNA methylation at Myod1 promoter.

Consequently, DNMT1 deletion impaired brown fat thermogenesis and induced obesity and insulin resistance. Overall, our data suggest that DNA methylation plays an important role in brown adipocyte thermogenesis, and DNMT1 might be an important regulator during this process. Understanding the role of DNMT1 in brown adipocyte function and insulin sensitivity will enable us to develop novel and targeted therapeutic interventions for metabolic diseases.

3.3 Materials and methods

3.3.1 Mice

Brown adipocyte-specific DNMT1 knockout mice were created by crossing DNMT1- floxed mice with UCP1-Cre mice (Jackson Laboratories #024670), where UCP1 is specifically expressed in brown adipocytes. DNMT1-floxed mice were generated by inserting two loxP sites on both sides of exon 4-5 of DNMT1 (Jackson-Grusby et al.,

2001).When crossed with a tissue-specific Cre model, Cre-dependent cleavage will result in a truncated DNMT1 protein that lacks catalytic activity. C57BL/6J (B6) mice

(Jackson Laboratories) were used in some experiments. All mice were fed chow diet

(LabDiet 5001, fat content 13.5% by calorie) or HF diet (D12492, 60% calories from fat, Research Diets Inc.). All mice were housed with a 12/12 h light–dark cycle in temperature- and humidity-controlled rooms with free access to water and food. To study the role of DNMT1 in thermogenic activity of brown adipocytes, we exposed 8- 55

week old mice to cold condition (5°C) up to seven days. Body temperature was monitored during cold exposure by a temperature transponder (BioMedic Data

Systems, Seaford, DE) implanted into the peritoneal cavity (Nguyen et al., 2017). After cold exposure, mice were sacrificed. Brown adipose tissues and white adipose tissues were harvested for measuring gene expression or protein levels. All aspects of animal care were approved by Georgia State University’s Animal Care and Use Committee.

3.3.2 Metabolic measurements

Body weight was measured weekly during the whole process of the experiment in both DNMT1 KO and fl/fl control mice fed chow or HFD. Body composition (fat mass and lean mass) were measured by using a Minispec NMR Body Composition Analyzer

(Bruker BioSpin Corporation; Billerica, MA). Food intake was measured over seven consecutive days. Total food intake during seven days was analyzed and normalized per body weight. Metabolic Phenotyping was performed in the TSE Comprehensive

Laboratory Animal Monitoring System (TSE Systems, Chesterfield, MO). Mice were singly housed in metabolic chambers for over seven consecutive days. Tail blood was harvested at fed or after overnight fast for blood glucose measurement by OneTouch

Ultra Glucose meter (LifeScan, Inc., Milpitas, CA) and serum insulin measurement by enzyme-linked immunosorbent assay (ELISA) (CrystalChem, Downers Grove, IL).

Glucose tolerance testing (GTT) and insulin tolerance testing (ITT) were performed as previously described (Wang et al., 2011, Yang et al., 2012). For GTT, KO and

CONTROL mice were overnight fasted , and glucose values from tail blood were measured at time 0, 15, 30, 60, 90, 120 minutes after intraperitoneal glucose 56

administration (1mg/kg body weight) (Andrikopoulos et al., 2008). For ITT, KO and fl/fl mice were fasted for four hours, and then glucose values from tail blood were measured at time 0, 15, 30, 60, 90, 120 minutes after intraperitoneal insulin administration (0.75 U/kg body weight) (McGuinness et al., 2009). At the end of diet treatment, mice were euthanized, and brown and white adipose tissues were collected for measuring the expression of thermogenic genes (such as UCP1, PRDM16, and

PGC1α).

3.3.3 HE staining, immunofluorescence (IF) and immunohistochemistry

(IHC)

Brown adipose tissues (BATs) and white adipose tissues (WATs) were fixed in 10% formalin and embedded in Paraffin. Paraffin-embedded sections were cut into 5 μm sections. Sections were processed for hematoxylin and eosin (HE) staining. For fluorescence detection, Paraffin-embedded sections were incubated with anti-myosin heavy chain (Myh1) antibody (MF20, Developmental Studies Hybridoma Bank

(DSHB), Iowa) overnight at 4˚C and then incubated with anti-mouse secondary antibodies labeled with Alexa fluor 488 (Invitrogen) for 1 hour at room temperature and counterstained with 4′,6-diamidino-2-phenylindole (DAPI). For in vitro study, cell cultures were fixed in 10% formalin. Fixed cells were incubated with anti-Myh1 antibody overnight at 4˚C and then incubated with anti-mouse secondary antibodies labeled with Alexa fluor 594 (Invitrogen) for 1 hour at room temperature and counterstained with DAPI. For IHC, Paraffin-embedded sections of BATs were incubated with anti-UCP1 antibody (abcam 10983) overnight at 4˚C and then 57

incubated with biotin-conjugated anti-rabbit secondary antibody (Jackson

ImmunoResearch, 711-065-152) for 30 min at room temperature. The sections were washed in PBS and incubated with streptavidin-conjugated horseradish Peroxidase

(VECTASTAIN® ABC Kit, PK-6100). Then the sections were washed in PBS and incubated with 3, 3-diaminobenzidine (DAB). Histology images were captured using

Nikon Eclipse E800 Microscopy as previously described (Fu et al., 2015a).

3.3.4 Cell Culture

Immortalized brown fat preadipocytes (BAT1) were maintained in DMEM/F12 containing 10% fetal bovine serum and 1% penicillin/streptomycin at 37°C in the presence of 5% CO2. BAT1 cells were kindly provided by Dr. Patrick Seale, University of Pennsylvania (Seale et al., 2007, Rajakumari et al., 2013). To differentiate brown preadipocytes, 90% confluent cells were cultured in induction medium containing

20nM insulin, 1nM T3, 125μM indomethacin, 500μM isobutylmethylxanthine (IBMX) and 0.5μM dexamethasone. After two days, cells were cultured in maintenance medium containing 20nM insulin and 1nM T3. At day 6, all of cells were differentiated as previously described (Rajakumari et al., 2013). To activate thermogenic genes, brown adipocytes were treated with 1μM isoproterenol for 3 hours.

3.3.5 Oil-red-O staining BAT1 cells were grown in 6-well plate. Differentiated BAT1 cells were washed twice with PBS, then incubated with 10% formalin for 1 hour. Fixed cells were washed twice with ddH2O, and incubated with 0.05% oil red O (Sigma: 00625-25G) working 58

solution for 1 hour. Cells were then washed three times with ddH2O. Samples were visualized using Nikon Eclipse E800 Microscopy.

3.3.6 small interfering RNA (siRNA) transfection

Mouse DNMT1 siRNA-SMART pool (L-056796-01), Myod1 siRNA-SMART pool (L-

041113-00), KLF2 siRNA-SMART pool (L-046974-01), Nr4a1 siRNA-SMART pool (L-

040970-01), Nr4a2 siRNA-SMART pool (L-048281-01), Nr4a3 siRNA-SMART pool

(L-043983-01) and non-targeting scramble siRNA were purchased from GE

Healthcare. siRNAs were transfected into BAT1 brown adipocytes by Amaxa

Nucleofector II Electroporator (Lonza) using Amaxa cell line nucleofector kit L according to the manufacturer's instructions (Lonza). Briefly, at days 4–6 of differentiation, cells (2 × 106 cells/sample) were trypsinized and centrifuged at 90 × g for 5 min at room temperature. Cell pellet was resuspended in nucleofector solution

(100 μl/sample) with 20 pmol of siRNA and seeded into 24-well plates or 100 mm dishes. Cells were treated with isoproterenol for 3 hours two days after transfection.

3.3.7 Quantitative real-time RT-PCR Total RNA was extracted with TRI-Reagent (Molecular Research Center) according to the manufacturer’s instructions. Messenger RNAs (mRNA) were quantified by RT-

PCR using the TaqMan one-step RT-PCR Master mix (Life technologies). Cyclophilin was used as an internal control which is described in Methods sections under SA1. The primer and probe pairs used for myogenic genes were purchased from Applied

Biosystems (Table 5). 59

3.3.8 RNA sequencing

Total RNA was isolated from brown adipose tissues (BATs) of chow-fed DNMT1 KO and fl/fl mice as noted above. Subsequently, total RNA was treated with RNase-free

DNase I to remove residual DNA. mRNA was purified from 2 μg of total RNA using beads containing oligo (dT). mRNA Libraries were prepared according to the

Illumina’s protocol as previously described (Peng et al., 2012, Ren et al., 2012).

Sequencing was performed on the Illumina-HiSeq2000. RNA sequencing data were aligned using TopHat. Clean reads were aligned to the reference genome (UCSC mm9) and transcriptome using SOAP2. Gene sets were considered to have a difference in expression with a fold change of 1.5 or greater, and FDR<0.001.

3.3.9 Immunoblotting

Protein (25μg) was separated by sodium dodecyl sulfate polyacrylamide gel and transferred to polyvinylidene difluoride (PVDF) membranes, which was blocked with

5% fat-free milk (2h, room temperature), as described in Methods sections under SA1.

The Protein expression levels were measured by immunoblotting. Antibodies against

UCP1 were from abcam (ab23841). Antibodies against α-Tubulin were from Advanced

Biochemicals (ABCENT4777).

3.3.10 Chromatin immunoprecipitation (ChIP) assay

ChIP assay was performed using a ChIP assay kit (Upstate), as described in Methods sections under SA1. Eluted DNA was analyzed by real-time PCR using SYBR green.

Primer sequences used in this study were shown in Table 6. 60

3.3.11 Bisulfite conversion and pyrosequencing

Genomic DNA was harvested by phenol/chloroform extraction. Bisulfite conversion was performed by using EpiTech Bisulfite Kit (Qiagen). The primers used to amplify

Myod1 proximal promoter near transcriptional start sites/differentially methylated regions (DMRs) covering CpG sites are shown in Table 7. Bisulfite-converted DNA

(about 2μg) was amplified by PCR and pyrosequencing was performed by EpiGenDx

(Hopkinton, MA).

3.3.12 Statistical Analysis

Data were expressed as mean ±SEM. Statistical tests were performed using SPSS software (version 23, SPSS Inc, Chicago, IL, USA). Two groups comparison was performed by Student-Newman-Keuls test. One-way Analysis of Variance (ANOVA) was used to perform multiple comparisons. Two-way ANOVA with Bonferroni post hoc tests were used when both genotypes and treatments were compared. Statistical significance was accepted at P<0.05. Figures were prepared using GraphPad Prism 5

(GraphPad Software, San Diego, CA).

3.4 Results

3.4.1 High-fat diet (HFD) reprograms brown fat to a muscle-like phenotype and modulates brown fat thermogenic function.

Recent studies have identified metabolic derangements in mice fed high-fat diet

(HFD) (Yoshioka et al., 1992, Putti et al., 2015, Haque and Ansari, 2018). Acute HFD feeding incudes brown fat thermogenesis through increasing circulating energy 61

substrates including free fatty acids, which function as fuels and activate UCP1

(Rothwell et al., 1983, Shi et al., 2017). However, whether prolonged HFD feeding induces BAT reprogram is unknown. Here, we found that compared to chow-diet feeding, HFD feeding significantly increased UCP1 mRNA levels in BATs at the short term (from 1 week to 12 weeks), but the degree of increase gradually reduced (Fig. 9A).

Interestingly, the increase in UCP1 mRNA disappeared after 24-week HFD feeding

(Fig. 9A). To our surprise, RNA-seq data showed that HFD feeding (24-week) significantly upregulated groups of myogenic gene expression in BATs (Fig. 9B), which was further confirmed by real-time PCR (Fig. 9C). In parallel to this increase, Myh1 protein was increased in BAT in HFD-fed mice as measured by Myh1 immunofluorescent staining (Fig. 9D). In addition, we found a reciprocal relationship between UCP1 and myogenic gene expression. While short-term HFD feeding (1-12 week) dramatically reduced Myod1, Myog and Myh1 expression, prolonged HFD feeding (24 weeks - 1 year) significantly increased these gene expression (Fig. 9E-G).

In contrast, the expression pattern of UCP1 is opposite to that of myogenic gene expression, with gradual decrease in UCP1 expression corresponding to the gradual increase of the myogenic gene expression during HFD feeding (Fig. 9H). Moreover, myogenic genes including Myod1, Myog and Myoh1 were negatively correlated with

UCP1 gene (Fig. 9I). Thus, our data suggest that after an initial increase in diet-induced thermogenesis, prolonged HFD feeding might induce brown adipocyte reprogramming to a skeletal muscle-like phenotype, leading to BAT dysfunction. 62

3.4.2 Deficiency of DNMT1 promotes myogenic gene expression and prevents brown thermogenic gene expression in brown adipocytes.

DNA methylation status is associated with gene transcriptional repression which plays important roles in cell differentiation (Yang et al., 2016). DNA methylation status of UCP1 promoter is associated with organ-specific expression of UCP1. DNA methyltransferase 1 (DNMT1) is a major enzyme for maintenance of DNA methylation

(Ko et al., 2005). We found that DNMT1 gene expression increased during brown adipocyte development (data not show). Therefore, we asked whether DNMT1 regulates UCP1 gene expression in brown adipocytes. We used Cre/loxp mouse models to generate mice with brown adipocyte specific deletion of DNMT1 to investigate the effects of DNMT1 on thermogenic activity of brown adipocytes. Using RNA sequencing

(RNA-seq), we compared gene expression in brown adipose tissues (BATs) between fl/fl control and DNMT1 knockout (KO) mice. Consistent with the inhibitory function of DNMT1, many genes were induced by DNMT1 deletion. We identified 98 upregulated genes in BATs from DNMT1 KO mice fed chow diet compared to that of fl/fl mice. To our surprise, among the upregulated genes, most of them are involved in skeletal myogenesis, for example, Myod1, creatine kinase, muscle (CKM), myosin heavy polypeptide 1 (Myh1), Myh4, ATPase Ca++ transporting, cardiac muscle fast twitch 1 (Atp2a1), and myogenin (Myog) (Fig. 10A). DNMT1 deletion-mediated increase in myogenic gene expression was confirmed by using real-time PCR (Fig. 10B).

In parallel to this increase, we observed a decrease in brown thermogenic gene expression in BATs of DNMT1 KO mice including UCP1, PGC1α, PR domain- 63

containing protein 16 (PRDM16), early B-cell factor-2 (EBF2), early B-cell factor-3

(EBF3), epithelial V-like antigen 1 (EVA1), otopetrin 1 (OTOP1), cell death-inducing

DFFA-like effector a (CIDEA), Fibroblast growth factor 21 (FGF21), cytochrome c oxidase 1 (COX1) and acyl-CoA oxidase 1 (ACOX1) (Fig. 10C). In addition, DNMT1 deletion significantly decreased UCP1 protein expression BATs (Fig. 10D and E), whereas increased myh1 protein expression in BATs (Fig. 10F).

Similar results were observed in BAT1 brown adipocyte cells line. DNMT1 knockdown resulted in an increase in myogenic gene expression including Myog, Myod1, CKM

(creatine kinase, muscle), Myh1 (myosin, heavy polypeptide 1, skeletal muscle), Acta1

(actin, alpha 1, skeletal muscle), Mfsd2a (major facilitator superfamily domain containing 2A), Ttn (titin) (Fig. 10G). Similarly, DNMT1 knockdown resulted in decreases in brown specific gene expression including UCP1, EBF2, EBF3 and PGC1α

(Fig. 10H). As expected, by day 5 of brown adipogenic differentiation, cultured BAT1 cells with scrambled siRNA had undergone normal brown adipocyte differentiation, corresponding to significant oil red O staining (Fig. 10I). In contrast, differentiated

BAT1 cells with DNMT1 siRNA had significantly reduced oil red O staining (Fig. 10I) and contained muscle-like cells, which were positively stained with skeletal myocyte marker Myh1 (in red fluorescence) (Fig. 10J). These data strongly suggest that loss of

DNMT1 switch brown thermogenesis to myogenesis leading to dysfunctional BATs.

3.4.3 Deletion of DNMT1 in brown fat results in the impairment of cold- induced brown fat thermogenesis.

Cold stress amplifies the need for UCP1-mediated thermogenesis. To further 64

investigate the role of DNMT1 in the regulation of brown fat thermogenic function,

DNMT1 KO and fl/fl control mice were exposed to cold condition (5 °C) for up to seven days and core body temperature was monitored during the cold exposure. We observed that the temperature of fl/fl mice dropped about 1.7°C, while that of DNMT1

KO mice dropped about 3.5°C during an acute 4-hr cold exposure, indicating KO mice had impaired cold-induced thermogenesis (Fig. 11A and B). As expected, DNMT1 KO reduced cold-induced UCP1 expression and the other brown specific- and beige- specific gene expression in BATs compared to that in fl/fl (Fig. 11C). Consistent with decreased UCP1 mRNA levels, DNMT1 KO mice had a significant decrease in UCP1 protein levels in BATs (Fig. 11D). In contrast, myogenic gene expression such as Myod1,

Myog, CKM, Ttn, Mfsd2a, and Acta1 was elevated in BATs of DNMT1 KO mice relative to that of fl/fl mice (Fig. 11E). Consistent with having hypoactive thermogenesis,

DNMT1 KO mice had more lipid accumulation in BAT (Fig. 11F). These data demonstrated that DNMT1 deletion impaired cold-induced brown fat thermogenesis.

3.4.4 Deletion of DNMT1 in brown fat results in obese and insulin resistance.

To determine effects of DNMT1 on brown adipose thermogenic function and its consequences on whole body physiology and metabolism, DNMT1 KO and fl/fl mice were fed either chow diet or High-fat diet (HFD). In chow diet groups, DNMT1 KO mice on both female and male were more obese than fl/fl mice over 12 weeks after birth (Fig. 12A and Fig. 13A). Mice with deletion of DNMT1 had higher fat mass and lower lean mass compared with fl/fl mice (Fig. 12B and Fig. 13B). The higher 65

percentage of fat mass in DNMT1 KO mice was primarily due to higher WATs depots without change in BAT depots which were measured by dissection at the time of sacrifice (Fig. 12C). The difference of WATs depots remained after adjusting for body weight (Fig. 12D). Morphology analysis showed that the size of adipocytes in BAT and

WATs of DNMT1 KO mice was larger than that of CONTROL mice (Fig. 12E). These data suggested that DNMT1 deletion in brown fat induces obese phenotype.

To determine whether obesity in BAT-specific DNMT1 KO mice is due to changes in food intake or energy expenditure, we put BAT-specific DNMT1 KO and fl/fl mice into the TSE metabolic cage systems. Compared to fl/fl mice, DNMT1 KO mice exhibited decreases in oxygen consumption, CO2 production and energy expenditure in dark cycle without change in light cycle (Fig. 12F-H). There is no change in respiration exchange rate (Fig. 12I) and physical activity (Fig. 12J) between DNMT1 KO and

CONTROL mice. In addition, DNMT1 KO mice ate less than fl/fl mice, possibly secondary to reduced energy expenditure in KO mice (Fig. 12K). These data suggest that mice with BAT-specific DNMT1 deletion are prone to developing obesity even on a regular chow diet. This is primarily due to decreased energy expenditure in KO mice, but not due to any changes in physical activity or food intake.

Consistent with the increase in body weight and decreases in whole body energy expenditure in DNMT1 KO mice, significant change in insulin sensitivity was apparent.

At 5-month age, although there was no difference in fed and fasted blood glucose levels between chow-fed DNMT1 KO and fl/fl mice, BAT-specific DNMT1 KO mice exhibited increases in fasting and fed serum insulin levels, indicating impaired insulin sensitivity 66

(Fig. 12L). Both male and female DNMT1 KO mice showed impaired glucose and insulin tolerance when compared to fl/fl mice (Fig. 12M-N and Fig. 13C-D). These results indicate that DNMT1 deletion in brown fat results in obesity and insulin resistance in vivo.

Similar results were observed on HFD-fed mice. Both female and male DNMT1 KO mice exhibited higher body weight (Fig. 14A and Fig. 15A), higher fat mass, lower lean mass (Fig. 14B), higher weight of WAT depots (Fig. 14C and Fig. 15B-C). DNMT1 KO mice also showed larger size of adipocytes in BAT and WATs (Fig. 15D). Female

DNMT1 KO mice had higher serum insulin level and similar blood glucose levels compared with fl/fl mice (Fig. 14D). In addition, DNMT1 deletion in brown fat resulted in HFD-induced glucose intolerance and insulin resistance (Fig. 14E-F and Fig. 15E-

F). Furthermore, mice with brown fat loss of DNMT1 had decreased oxygen consumption, CO2 production and energy expenditure (Fig. 14G-I and Fig. 15G-I) without change in RER and physical activity (Fig. 14J-K and Fig. 15J-K). Consistent with lower metabolic rate, DNMT1 KO mice had decreased brown thermogenic gene expression including UCP1 (Fig. 15L) and decreased UCP1 protein express in BAT (Fig.

15M). In summary, these data suggest that BAT-specific DNMT1 deletion impaired brown fat thermogenic function and whole body energy metabolism, consequently resulting in insulin resistance.

3.4.5 DNMT1 regulates myogenesis and thermogenesis in BATs through targeting Myod1

Brown adipocytes and skeletal myocytes develop from the same precursors (Seale et 67

al., 2008). There are some important transcriptional factors determining cell fate switching between brown adipocytes and skeletal myocytes. For example, Myod1 is a master regulator for muscle development and a repressor for brown fat development

(Seale et al., 2008). Myod1 gene expression could be induced by 5-Aza (as an inhibitor of DNA methyltransferases) (Constantinides et al., 1977). Nuclear receptors are important transcriptional factors including Nr4a1, Nr4a2 and Nr4a3 which are expressed both in adipose tissues and skeletal muscle. DNMT1 is reported to regulate cell differentiation through directly repressing NR4as (Chen et al., 2016b).

Overexpression of Nr4as suppresses adipocyte differentiation (Chao et al., 2008), but promotes skeletal myocyte differentiation (Kanzleiter et al., 2009). In addition,

Kruppel-like factors (KLFs) regulates critical aspects of cellular growth and differentiation. It is reported that KLF2 promotes skeletal muscle development

(Sunadome et al., 2011). According to RNA-seq, we found some transcriptional factors induced by DNMT1 deletion involved in myogenesis such as Myod1, Nr4a1, Nr4a2,

Nr4a3 and KLF2 (Fig. 10A), which indicates that Myod1, Nr4as or KLF2 might be the direct target of DNMT1. In our experiments, we found that in vivo brown fat development were associated with significantly increased UCP1 mRNA, which was accompanied by decreased Myod1 mRNA (Fig. 16A). In contrast, Nr4a1/2/3 and KLF2 gene expression were increased during brown fat development (Fig. 17A). Cold- induced brown thermogenesis was associated with increased UCP1 gene expression which was accompanied by decreased Myod1 (Fig. 16B), Nr4a1 and KLF2 gene expression (Fig. 17B). Nr4a2 mRNA was too low to detect in brown adipose tissues 68

during cold exposure. Nr4a3 mRNA was dramatically increased at 6 hours after cold exposure and dropped to the basal level at 24 hours after cold exposure (Fig. 17B). To determine the molecular mechanism by which DNMT1 deletion drives expression of myogenic genes, we performed chromatin immunoprecipitation followed by real-time

PCR (ChIP-PCR) in BAT1 for DNMT1. We found that DNMT1 knockdown significantly decreased DNMT1 binding to myod1 promoter in brown preadipocytes and brown adipocytes. In addition, compared to preadipocytes, brown adipocytes exhibited high levels of DNMT1 binding to Myod1 promoter, which was reduced by DNMT1 knockdown (Fig. 16C). Similar results were observed in Nr4as and KLF2 promoters.

Mature brown adipocytes had higher levels of DNMT1 at Nr4as and KLF2 promoters, which was decreased by DNMT1 knockdown (Fig. 17C). We then wondered whether inhibition of Myod1, Nr4as or KLF2 reverses the role of DNMT1 knockdown-induced downregulation of brown thermogenic gene expression. To test this hypothesis, we simultaneously knocked down DNMT1 and individual transcriptional factors (Myod1,

Nr4as and KLF2) in differentiated BAT1 cells. As expected, DNMT1 and Myod1 siRNA treatment resulted in significantly knockdown of DNMT1 and Myod1, respectively (Fig.

16D). In addition, while DNMT1 KD significantly reduced BAT-specific gene expression, Myod1 KD significantly increased BAT-specific gene expression (Fig. 16D).

More importantly, Myod1 knockdown reversed DNMT1 inactivation-induced inhibition of brown thermogenic gene expression including UCP1 and PGC1α (Fig.

16D).As expected, Myod1 knockdown inhibited myogenic gene expression and reversed DNMT1 siRNA-induced myogenic gene expression including Myod1, Myog 69

and Acta1 (Fig. 16D). However, knocking down Nr4a1, Nr4a2, Nr4a3 or KLF2 either did not change UCP1 expression or did not reverse the reduced UCP1 expression by

DNMT1 KD (Fig. 17D). These data indicate that Myod1 might be the direct target of

DNMT1 in the regulation of cell fate switching between brown adipocyte and skeletal myocyte.

To confirm it, we investigated DNA methylation status of Myod1 gene. Considering that DNA methylation status of gene contributes to its organ-specific expression, we tested DNA methylation levels of Myod1 gene in brown adipose tissue (BAT) and skeletal muscle gastrocnemius (GAS). Previous data suggest that there is a differential methylated region (DMR) about 25kb upstream of Myod1 transcriptional start site that controls Myod1 expression (Liu et al., 2016). In addition, we have also identified an extensive CpG island at the proximal promoter, near the transcriptional start site (TSS) on Myod1 (Fig 19A) (by using the GT-Scan web-tool). Bisulfite sequencing showed that compared to GAS, BAT had significantly higher levels of DNA methylation near transcription start site (TTS) of Myod1 gene (Fig. 18A), but lower levels of DNA methylation at DMRs about 25kb upstream region of Myod1 gene (Fig. 19A). We also found that high expression of myogenic genes in GAS including Myod1, Myog and

CKM, and high expression of brown thermogenic genes in BAT including UCP1 and

PRDM16 (Fig. 18B). Therefore, we focused on DNA methylation status at Myod1 proximal promoter.

As expected, inactivation of DNMT1 reduced DNA methylation at Myod1 promoter

(Fig. 18C). We then investigated whether targeting DNA demethylation near TTS of 70

Myod1 gene could induce myogenic gene expression and prevent brown thermogenic gene expression in brown adipocytes. To target specific loci of DNA methylation, we fused a modified endonuclease dead version of CRISPR associated protein 9 (dCas9) with Tet1, an enzyme regulates DNA demethylation (Choudhury et al., 2016, Liu et al.,

2016). Co-expression of Myod1-specific guide RNA (Myod1-gRNA) targeted dCas9-

Tet1 to Myod1 specific locus will result in decreases DNA methylation on Myod1 gene without altering its DNA sequence, as dCas9 lacks endonuclease activity, as illustrated in Fig. 20A. To test whether dCas9-Tet1/Myod1-gRNA could reduce DNA methylation levels of Myod1 promoters and activate its gene expression, we transfected dCas9-Tet1 with Myod1-gRNA or with scrambled gRNA into cultured BAT1 cells, and measured

DNA methylation status of Myod1 and its gene expression. Bisulfite sequencing showed that transduction of dCas9-Tet1-Myod1-gRNA resulted in a significant decrease of DNA methylation in the Myod1 promoter region near its transcription start sites (Fig. 20B). As expected, Myod1 gene expression was increased in dCas9-Tet1-

Myod1-gRNA transfected group (Fig. 20C). In parallel, brown thermogenic genes including UCP1, PGC1α, EBF2 and PRDM16 were repressed in dCas9-Tet1-Myod1- gRNA transfected group (Fig. 20C). These data suggest that demethylation of Myod1 proximal promoter by dCas9-Tet1 with Myod1-gRNA resulted the switch of brown adipocyte thermogenesis to myogenesis.

3.5 Discussion of specific aim 2 Recent studies of obesity and diabetes link high-fat diet (HFD) to hepatic steatosis

(Kang et al., 2010), heart dysfunction (Birse et al., 2010), intramyocellular lipids (Baek et al., 2018), and adipose inflammation (Chalkiadaki and Guarente, 2012). However, its physiological role in brown adipose tissue (BAT) is not clear. Here, we find that

HFD induces the reprogramming of brown adipocyte into skeletal muscle cells in BATs.

For decades, BAT thermogenesis has been considered essential for burning extra energy to maintain metabolic health (Vinter et al., 1988), because BAT plays an important role in blood FFAs clearance and glucose disposal as well as non-thriving thermogenesis (Cannon and Nedergaard, 2004, Bartelt et al., 2011). The protein UCP1 in BAT is capable of dissipating chemical energy such as free fatty acids (FFAs) to produce heat. Thus, the release of FFAs from fat stores (mainly in white adipose tissues) or absorption of FFAs from the diet is considered an important step in brown fat thermogenesis. FFAs from WATs or absorbed from the diet are delivered to BATs via circulation to act as a fuel substrate of thermogenesis (Bartelt et al., 2011). HFD feeding induces circulating energy substrates including blood glucose, plasma FFAs, glycerol and triglyceride (TG) (Bartelt et al., 2011). In C57BL/6 mouse models, we find that short-term HFD feeding (less than 12 weeks) promotes BAT thermogenesis evidenced by increases in UCP1 expression in BATs of HFD fed mice. Thus, our results confirm the observation that HFD feeding increases BAT thermogenesis and metabolic rate at least in the short term (Rothwell and Stock, 1979, Feldmann et al.,

2009, Hibi et al., 2016). However, HFD fed mice still exhibit obesity and insulin 72

resistance (Fu et al., 2015a, b). It is possible that hyperactive BAT may not combat

HFD-induced increase in circulating energy substrates including FFAs, TG, and glucose. This is based on observations of high levels of glucose (Fu et al., 2015b), FFAs and TG (Fu et al., 2015a) in diet-induced obese mice. However, whether prolonged

HFD induces BAT reprogramming is unknown.

Although short term HFD induces UCP1 gene expression compared to chow-diet feeding, this effect diminishes when HFD feeding prolongs. By contrast, RNA-seq shows that numerous myogenic genes in BAT are induced by long-term HFD feeding.

In addition, we find that long-term HFD feeding significantly increases, whereas short-term HFD feeding dramatically inhibits myogenic gene expression including

Myod1, Myh1, and Myog in BATs. An explanation for long-term HFD feeding- mediated decrease in BAT-specific gene expression and increase in myogenic gene expression is that HFD may control switch of brown adipocyte to skeletal muscle cell determination. Brown adipocyte and myocyte share a common Myf5+ lineage origin; whereas PRDM16 promotes brown adipocyte lineage and suppresses myocyte lineage development, Myod1 does the opposite. Thus the interaction between the two transcriptional factors determines brown adipocyte and myocyte lineage development in BAT and skeletal muscle (Wang et al., 2017, Tosic et al., 2018). Thus, obese BAT might create an environment suitable for the cell fate determination switch from brown adipocyte progenitor to skeletal myofibrobasts, which alters BAT plasticity and precipitates the loss of thermogenic function in BAT. Indeed, we find that expression of myogenic genes including Myod1, Myog, and Myh1 is negatively correlated with 73

UCP1 gene expression in BATs during HFD feeding. Thus, our data suggest that HFD may impair brown fat thermogenesis through promoting skeletal muscle cell development in BATs. However, further studies are required to investigate how HFD triggers cell fate switching between myocyte and brown adipocytes in BAT. Overall, our data indicate that prolonged HFD causes BAT dysfunction, suggesting that obese phenotype and insulin resistance in HFD fed mice might be at least partially due to

inactivity of brown adipose tissue.

Recent studies suggest that DNMTs are important epigenetic marks that are associated with the development of obesity (Chen et al., 2016b), and obesity development affects DNMT expression (including DNMT1, DNMT3A, and DNMT3B) and their enzymatic activity (Xia et al., 2014). However, the role of DNMTs in the regulation of brown adipocyte activity has not been studied. DNMT3a and 3b catalyze de novo methylation, whereas DNMT1 is a maintenance enzyme that maintains DNA methylation pattern during cell division (Ko et al., 2005). Here, we are surprised to find that DNMT1 deletion in brown adipocyte mimics HFD-induced brown fat reprogramming, leading to BAT dysfunction and contributing to obesity and insulin resistance. This is evidenced by followed observations. Firstly, obese phenotypes including increased body weight, increased fat mass, increased blood glucose, and insulin resistance are observed in BAT-specific DNMT1 KO mice even in regular chow diet condition. Secondly, BAT-specific DNMT1 KO leads to dysfunctional brown adipocytes with increased lipid accumulation and reduced Brown-specific gene expression. Thirdly, BAT-specific DNMT1 deletion reprograms brown fat to a muscle- 74

like phenotype, indicating the decrease in brown fat thermogenic activity. Our data suggest that DNMT1 is important in maintaining brown fat cell phenotype and thermogenic function, and restricting skeletal myocyte lineage development. Indeed, we find that BAT-specific DNMT1 KO mice display decreased oxygen consumption and energy expenditure compared to their fl/fl littermate control mice. Activating brown fat thermogenesis alleviates metabolic diseases including type 2 diabetes and obesity

(Himms-Hagen, 1979, Jun et al., 2014, Kim and Plutzky, 2016, Viana-Huete et al.,

2018). Thus our data suggest that DNMT1 deletion induces obesity and insulin resistance primarily through attenuating brown adipocyte activity.

Moreover, we observed the cold intolerance of DNMT1 KO mice. Compared to fl/fl control mice, BAT-specific DNMT1 KO mice had significantly more decreased body temperature during acute cold exposure. As BAT thermogenic function is essential for mammalian to maintain their body temperature and survive under cold condition

(Griggio, 1982, Golozoubova et al., 2001, Nedergaard et al., 2001), our data suggest that DNMT1 deletion in brown adipocytes impairs cold-induced brown fat thermogenic function. Indeed, DNMT1 KO in brown adipocytes significantly down- regulates brown fat-specific gene expression, indicating a positive association between

DNMT1 and brown adipocyte thermogenic function. Thus, our data suggest that cold intolerance in DNMT1 KO mice is at least partially caused by BAT dysfunction.

DNA methylation is a reversible and dynamical process responding to the changes in nutrient status and physical stimuli which affect systemic energy homeostasis (Ko et al., 2005). DNMTs catalyze the methylation at 5’ carbon of cytosine, predominantly in 75

the cytosine phosphor-guanine dinucleotide (CpG) sites in the genome (Paluch et al.,

2016). De novo DNMT3a and DNMT3b initiate DNA methylation, and DNMT1 maintains established DNA methylation pattern during cell division. DNA methylation at the promoter regions leads to gene transcription repression. Overexpression of

DNMT3a or DNMT3b promotes expression of genes involved in obesity and inflammation (Kamei et al., 2010, Yang et al., 2014). DNMT1 catalyzes methylation of genes involved in adipogenesis such as adiponectin and leptin (Xia et al., 2014, Kim et al., 2015), and myogenic genes such as FOXO1 and Myod1 (Wang et al., 2015, Yan et al., 2016). In the present study, RNA-seq data shows that Myod1 gene expression is upregulated in BAT in DNMT1 KO mice. Moreover, reduced representation bisulfite sequencing (RRBS) data indicate the hypomethylation in Myod1 promoter in BAT in

DNMT1 KO mice (data not shown). These data suggest that Myod1 might be the direct target of DNMT1 in the regulation of brown adipocyte function. Indeed, sympathetic activation of BAT by cold stress is associated with downregulation of Myod1 gene expression. Our data are consistent with recent studies that Myod1 is a negative regulator of brown adipocyte development and activity (Seale et al., 2008, Borensztein et al., 2012, Yin et al., 2013). Additional mechanistic studies with differentiated brown adipocyte cell lines (BAT1) further demonstrated an important role of Myod1 in

DNMT1-mediated regulation of BAT thermogenic function. We find that DNMT1 binds to myod1 gene promoter in differentiated BAT1 cells, which is significantly reduced by DNMT1 knockdown. Indeed, DNMT1 knockdown reduces DNA methylation levels in Myod1 promoter. CRISPR/dCAS9-mediated delivery of DNA 76

demethyltransferase TET1 at Myod1 promoter significantly decreases its DNA methylation levels and increases its gene transcription, which in turn inhibit brown specific gene expression including UCP1, PGC1α, EBF2 and PRDM16 in BAT1 cells.

Thus, these data suggest that Myod1 may be a direct target of DNMT1-mediated regulation of brown adipocyte thermogenesis. As expected, double knockdown of

DNMT1 and Myod1 reverses DNMT1 siRNA-induced inhibition of brown specific gene expression including UCP1 and PGC1α. Thus, our data suggest that DNMT1 promotes brown adipocyte thermogenesis through inhibiting Myod1. Recent studies demonstrate that Myod1 negatively regulates brown fat development through upregulating a myogenic microRNA miR-133 to repress brown specific mark PRDM16.

Thus, it is possible that DNMT1 may be recruited to Myod1 promoters, leading to down-regulation of Myod1 expression. This will in turn inhibit its downstream cellular signaling cascades and mechanisms including miR-133 and PRDM16. Additional studies using animal models with CRISPR/dCAS9-mediated delivery of TET1 in

Myod1 promoters directly into BAT tissues will be warranted to further demonstrate the important role of Myod1 in DNMT1-induced brown adipocyte thermogenesis.

It is reported that DNMT1 regulates cell differentiation through directly repressing the nuclear receptor subfamily 4 group A members (NR4a receptors) (Chen et al.,

2016b). Nuclear receptors are important transcriptional factors including Nr4a1,

Nr4a2 and Nr4a3 which are expressed both in adipose tissues and skeletal muscle.

Overexpression of Nr4a receptors suppresses adipocyte differentiation (Chao et al.,

2008), but promotes skeletal myocyte differentiation (Kanzleiter et al., 2009). In our 77

study, we find that DNMT1 binds to Nr4a receptors and brown fat deletion of DNMT1 upregulated expression of Nr4a receptors, which indicates that Nr4a receptors might be another targets of DNMT1. In addition, knocking down of Nr4a3, but not Nr4a1 or

Nr4a2 reverses DNMT1 siRNA-mediated decrease in UCP1 gene expression. However,

Nr4a3 gene expression is upregulated during brown fat development. Moreover,

Nr4a3 gene expression is upregulated during acute cold exposure, whereas downregulated during prolonged cold exposure. Thus, it would be interesting to determine whether DNMT1 also regulates DNA methylation levels at other genes’ promoters including Nr4a3 to regulate their gene transcription in brown adipocytes.

To more widely study epigenetic pathways targeted by DNMT1 in BAT, Reduced representation bisulfite sequencing (RRBS) will be needed to analyze the genome- wide methylation profiles regulated by DNMT1 deletion in brown adipocyte.

In summary, our work unveiled a novel epigenetic pathway that regulates brown adipocyte-to-muscle switch. BAT-specific DNMT1 deletion mimicks HFD-induced brown fat remodeling to a skeletal muscle like phenotype, which might be a maladaptive mechanism contributing to obesity-induced metabolic diseases including type 2 diabetes. BAT-specific DNMT1 deletion prevented thermogenic gene expression; whereas promoted myogenic gene expression in brown adipocytes. We also find that

DNMT1 deficiency mice exhibited cold intolerance, which is due to impairment of brown fat thermogenic function. Furthermore, loss of DNMT1 reduced whole-body metabolic rate and induced obese phenotype. As a consequence, insulin resistance and 78

glucose intolerance were observed in BAT-specific DNMT1 deficient mice. In addition, we find that DNMT1 regulated brown adipocyte thermogenesis through targeting

Myod1, which is an important transcriptional factor involved skeletal muscle regeneration. DNMT1 deletion down-regulated DNA methylation of Myod1 gene and increased its gene transcription. Overall, DNMT1 plays important role in the regulation of classical brown fat thermogenesis. We find important role of DNMT1 in switch between brown adipocytes and myocytes. Most of recent studies focus on the role of DNMTs in the regulation of tumor development. Some DNMT inhibitors are under investigation for the treatment of cancers such as Myelodysplastic Syndromes

(Mack, 2010). But our studies shift recent researches to a new focus on epigenetic regulation of DNMT1 in brown fat reprogramming and obesity. Moreover, our data provide novel therapeutic target of epigenetic regulation in the treatment of metabolic diseases such as type 2 diabetes.

3.6 Figures of specific aim 2

Figure 9 High-fat diet (HFD) induces myogenic gene expression in BATs.

(A) UCP1 mRNA levels in BATs during HFD feeding; (B) Heat map comparison of genes that are significantly up-regulated in brown adipose tissues (BATs) from chow fed mice compared to that from HFD fed mice. (C) mRNA levels of myogenic genes in BATs from chow and HFD fed mice. (D) Immunofluorescence of skeletal myosin heavy chain (Myh1) expression (in green) in BATs from chow and HFD fed mice. mRNA levels of Myod1(E), Myog (F) and Myh1 (G) in BATs during HFD feeding; (H) and (I) correlation of UCP1 and myogenic genes including Myod1, Myog and Myh1. Data are expressed as Mean ± SEM. n=8 in A, C and E-H. *p<0.05.

Figure 10 Brown specific-DNMT1 deficiency up-regulates myogenic genes and down-regulates brown-selective genes in brown adipocytes.

(A) Heat map comparison of genes that are significantly up-regulated in DNMT1-deficient BAT relative to the control. mRNA levels of myogenic genes (B) and Brown-specific genes (C) in brown adipose tissues from female chow diet-fed control and DNMT1 KO mice. Mean ± SEM. n=8. *p<0.05. (D) UCP1 protein levels in BAT. Mean ± SEM. *p<0.05. (E) Immunohistochemistry for UCP1 in BAT from female chow diet-fed control and DNMT1 KO mice. (F) Immunofluorescence of skeletal myosin heavy chain (Myh1) expression (in green) in BATs from CONTROL and DNMT1 KO mice. (G) mRNA levels of myogenic genes and (H) Brown- specific genes in mature brown adipocytes (BAT1 cells) with DNMT1- siRNA or scramble-siRNA. (I) BAT1 cells with DNMT1 siRNA or scramble siRNA were stained with Oil-Red-O 5 days after inducing adipocyte differentiation. (J) Immunocytochemistry of skeletal myosin heavy chain (Myh1) expression in differentiated BAT1 cells with DNMT1 siRNA or scramble siRNA. Myh1 fluorescence in red; DAPI fluorescence in blue. Mean ± SEM. n=8. *p<0.05. 81

Figure 11 DNMT1 KO in BAT reduces thermogenesis in BAT and reduces cold tolerance.

(A) Whole body temperature in male chow diet-fed control and DNMT1 KO mice exposed to the cold environment. Mean ± SEM. *p<0.05. (B) Body temperature change of control and DNMT1 KO mice within first 2 hours cold exposure. Mean ± SEM. *p<0.05. (C) mRNA levels of Brown-specific genes and beige-selective genes in brown adipose tissues from control and DNMT1 KO mice exposed to the cold environment for seven days. Mean ± SEM. n=6. *p<0.05. (D) UCP1 protein levels in BATs. Mean ± SEM. *p<0.05. (E) mRNA levels of myogenic genes in BATs from control and DNMT1 KO mice exposed to the cold environment for seven days. Mean ± SEM. n=6. *p<0.05. (F) Hematoxylin and eosin staining of BATs from control and DNMT1 KO mice. 82

Figure 12 Chow fed DNMT1 KO mice display an obese phenotype.

(A) Body weight curves of female chow diet-fed control (DNMT1f/f) and DNMT1 KO mice. Mean ± SEM. n=12. *p<0.05. (B) Body composition of female control and DNMT1 KO mice on chow diet. Mean ± SEM. n=12. *p<0.05. (C-D) Individual fat depots of female control and DNMT1 KO mice on chow diet. Mean ± SEM. n=12. *p<0.05. **p<0.01. (E) Hematoxylin and eosin staining of brown adipose tissue (BAT), subcutaneous fat (SQ), and gonadal fat (Gon) from control and DNMT1 KO mice. (F) Oxygen consumption rate, (G) carbon dioxide production rates, (H) energy expenditure, (I) respiratory exchange ratio (RER), and (J) physical activity in chow diet-fed control and DNMT1 KO mice. Dark cycle in dark bar, light cycle in light bar. (K) Cumulative food intake from day 1 to day 7. Mean ± SEM. n=4. *p<0.05. (L) Glucose levels and amounts of serum insulin at the fasted and fed states from chow diet-fed control and DNMT1 KO mice. (M) Glucose tolerance test in 22-week chow diet-fed mice. (N) Insulin tolerance test in 23-week chow-fed mice. Mean ± SEM. n=6. *p<0.05, **p<0.01, ***p<0.001, n.s., not statistically significant.

Figure 13 DNMT1 KO male mice display an obese phenotype

(A) Body weight curves of male chow diet-fed control and DNMT1 KO mice. (B) Body composition of female control and DNMT1 KO mice on chow diet. (C) Glucose tolerance test in 29-week chow diet-fed mice. (D) Insulin tolerance test in 30-week chow diet-fed mice. Mean ± SEM. n=8. *p<0.05, **p<0.01, n.s., not statistically significant.

Figure 14 DNMT1 KO female mice display High-fat (HF) diet-induced obesity.

(A) Body weight curves of female HF diet-fed control and DNMT1 KO mice. (B) Body composition of female control and DNMT1 KO mice on HF diet. (C) Individual fat depots of female control and DNMT1 KO mice on HF diet. (D) Glucose levels and amounts of serum insulin from HF diet-fed control and DNMT1 KO mice. (E) Glucose tolerance test in 27-week chow diet-fed mice. (F) Insulin tolerance test in 28-week chow diet-fed mice. (G) Oxygen consumption rate, (H) carbon dioxide production rates, (I) energy expenditure, (J) respiratory exchange ratio (RER), and (K) physical activity in female HF diet-fed control and DNMT1 KO mice. Dark cycle in dark bar, light cycle in light bar. Mean ± SEM. n=8. *p<0.05, **p<0.01, n.s., not statistically significant.

Figure 15 DNMT1 KO male mice display high-fat (HF) diet-induced obesity.

(A) Body weight curves of male HF diet-fed control and DNMT1 KO mice. (B and C) Individual fat depots of male control and DNMT1 KO mice on HF diet. (D) Hematoxylin and eosin staining of brown adipose tissue (BAT), subcutaneous fat (SQ), and Epididymal fat (Epi) from control and DNMT1 KO mice. (E) Glucose tolerance test in 23-week chow diet-fed mice. (F) Insulin tolerance test in 24-week chow diet-fed mice. (G) Oxygen consumption rate, (H) carbon dioxide production rates, (I) energy expenditure, (J) respiratory exchange ratio (RER), and (K) physical activity in male HF diet-fed control and DNMT1 KO mice. Dark cycle in dark bar, light cycle in light bar. (L) mRNA levels of Brown-specific genes and white-selective genes in BATs from male HF diet-fed control and DNMT1 KO mice. (M) UCP1 protein levels in BAT. Mean ± SEM. n=6. *p<0.05, **p<0.01, n.s., not statistically significant. 86

Figure 16 DNMT1 regulates myogenesis and brown fat development by targeting Myod1.

(A) UCP1 and Myod1 gene expression in brown fat development. E: Embryo, D: Postnatal. Mean ± SEM. n=4 *p<0.05 vs 17E; (B) UCP1 and Myod1 gene expression in BATs during cold exposure. n=4 *p<0.05 vs cold exposure day 0; (C) BAT1 preadipocytes and adipocytes were treated with scramble siRNA or DNMT1 (D1) siRNA. Cells were collected for a ChIP assay to measure DNMT1 protein binding to myod1 promotors. Different lower letters above the bars represent statistical significance. (D) Mature BAT1 cells were treated with scramble siRNA, DNMT1 siRNA, or Myod1 siRNA. mRNA was collected to measure gene expression of UCP1, PGC1α, DNMT1, Myod1, Myog and Acta1. Mean ± SEM. n=6 *p<0.05 vs Scramble; #p<0.05 vs DNMT1 siRNA.

Figure 17 DNMT1 regulates myogenic transcription factors.

Nr4a1, Nr4a2, Nr4a3 and KLF2 gene expression in BATs during brown fat development (A) and cold exposure (B). E: Embryo, D: Postnatal. Mean ± SEM. In A, n=4 *p<0.05 vs cold exposure day 0; in B, n=4 *p<0.05 vs 17E. (C) BAT1 preadipocytes and adipocytes were treated with scramble siRNA or DNMT1 (D1) siRNA. Cells were collected for a ChIP assay to measure DNMT1 protein binding to promotors of Nr4a1, Nr4a2, Nr4a3 and KLF2. Different lower letters above the bars represent statistical significance. (D) Mature BAT1 cells were treated with scramble siRNA, DNMT1 siRNA, Nr4a1 siRNA, Nr4a2 siRNA, Nr4a3 siRNA, or KLF2 siRNA. mRNA was collected to measure gene expression of UCP1. Mean ± SEM. n=6 *p<0.05 vs Scramble; #p<0.05 vs DNMT1 siRNA.

Figure 18 Myod1 is the target of DNMT1 regulating expression of myogenic genes and thermogenic genes in brown adipocytes.

(A) Methylation levels at CpGs before and after TRANSCRIPTION START SITE sites of myod1 in brown adipose tissue (BAT) and skeletal muscle gastrocnemius (GAS). Mean ± SEM. n=4 *p<0.05. (B) mRNA levels of brown-selective genes and myogenic genes in BAT and GAS. Mean ± SEM. n=4 *p<0.05. (C) Methylation levels at CpGs before and after TRANSCRIPTION START SITE sites of myod1 in BAT1 preadipocytes with scramble siRNA and BAT1 adipocytes with scramble or DNMT1 siRNA. Mean ± SEM. n=4 *p<0.05 vs Preadipocyte: scramble siRNA; #p<0.05 vs Adipocyte: scramble siRNA.

Figure 19 DNMT1 regulates methylation levels of Myod1 promoters.

(A) Methylation levels of CpGs at enhancer region of Myod1 in brown adipose tissue (BAT) and skeletal muscle gastrocnemius (GAS). Mean ± SEM. n=4 *p<0.05. (B) BAT1 cells were infected with lentiviruses expressing dCas9-Tet1 and gRNA target scramble or myod1 promoter region. DNA was collected to measure methylation levels of individual CpGs at enhancer region of Myod1. Mean ± SEM. n=4 *p<0.05. 90

Figure 20 DNMT1 regulates myogenesis and thermogenesis in BATs through targeting Myod1.

(A) BAT1 cells were infected with lentiviruses expressing dCas9-Tet1 and gRNA target scramble or myod1 promoter region. (B) DNA were collected to measure methylation levels of individual CpGs in myod1 promoters. Mean ± SEM. n=4 *p<0.05. (C) mRNA was collected to measure expression of brown-selective genes and myod1. Mean ± SEM. n=4 *p<0.05.

3.7 Tables of specific aim 2

Table 5 Primer/probe sets for Myogenic gene expression

Table 6 Primers for ChIP assay

Table 7 Primer sequences for pyrosequencing

Table 8 Primer sequence for Myod1 gRNA Sequence (5--3)

Myod1-sgRNA 1 F caccG AGCATTTGGGGGCATTTATGGG

Myod1-sgRNA 1 R aaac CCCATAAATGCCCCCAAATGCT c

Myod1-sgRNA 2 F caccG AAGTATCCTCCTCCAGCAGCTGG

Myod1-sgRNA 2 R aaac CCAGCTGCTGGAGGAGGATACTT c

Myod1-sgRNA 3 F caccG ACACAGCCAGTTGGGGGAAGGGG

Myod1-sgRNA 3 R aaac CCCCTTCCCCCAACTGGCTGTGT c

Myod1-sgRNA 4 F caccG CCAGAGTCAGCTGTTCCTGGG

Myod1-sgRNA 4 R aaac CCCAGGAACAGCTGACTCTGG c

Myod1-sgRNA 5 F caccG TAGCCAAGTGCTACCGCGTATGG

Myod1-sgRNA 5 R aaac CCATACGCGGTAGCACTTGGCTA c

Myod1-sgRNA 6 F caccG AGGGCGCCTGGCGCGAAGCTAGG

Myod1-sgRNA 6 R aaac AGGGCGCCTGGCGCGAAGCTAGG c

Myod1-sgRNA 7 F cacc GTACTGTTGGGGTTCCGGAGTGG

Myod1-sgRNA 7 R aaac CCACTCCGGAACCCCAACAGTAC

Myod1-sgRNA 8 F caccG TAAGACGACTCTCACGGCTTGGG

Myod1-sgRNA 8 R aaac CCCAAGCCGTGAGAGTCGTCTTA c

4 SPECIFIC AIM 3

4.1 Abstract Brown adipose tissue is a thermogenic organ that burns energy substrates in the form of heart to defend obesity and hypothermia. Brown adipocytes derive from skeletal muscle cell lineage through the activation of transcription factor PR domain containing 16 (PRDM16). However, the precise molecular mechanism that regulates the determination of brown adipocytes remains unknown. Here we show that lysine- specific demethylase 6a (KDM6a, also known as UTX) is required to activate brown adipose tissue. Mice with brown adipocyte-specific deletion of UTX become cold sensitive, and develop obesity and insulin resistance in response to high fat diet feeding. As an active epigenetic regulator, UTX deletion reduces brown-specific genes including uncoupling protein 1 (UCP1) and PR domain containing 16 (PRDM16) in brown adipocyte. Importantly, brown adipocyte-specific deletion of UTX not only inhibits brown fat-thermogenesis but also switches cell fate towards skeletal muscle- like cells, since UTX deletion increases key myogenic transcription factor genes including Myod1, Myog, myosin heavy chain (Myh1) and creatine kinase muscle

(CKM). UTX prevents skeletal muscle cell determination by up-regulating brown-fat transcriptional factor PRDM16, which recruits DNA methyltransferase 1 (DNMT1) to the promoter of the myocyte-specific transcriptional factor Myod1. DNMT1 increases

DNA methylation levels at Myod1 promoter and represses Myod1 gene transcription.

Myod1 is required for myoblast specification. Elevated levels of Myod1 activate its downstream transcriptional factor Myog that governs muscle cell differentiation. Thus, 96

our results suggest an epigenetic mechanism by which alleviation of obesity-related brown adipocyte dysfunction through UTX-PRDM16-DNMT1-Myod1 pathway inhibits skeletal muscle cell determination and promotes brown adipocyte determination.

4.2 Introduction There is cooperation between DNA methylation and histone methylation (Cedar and

Bergman, 2009, Zhou et al., 2018). DNA methylation blocks other epigenetic modulators binding to specific histone residues and prevent the histone modifications by these epigenetic modulators. For example, lysine demethylase 2A (KDM2A) only binds to tri-methylated H3K9 when the DNA is not methylated. Both histone methylation and DNA methylation are stable modification and play important roles in epigenetic modification of gene expression. Unlike histone acetylation, histone methylation does not directly regulate positively charged histone tails. Histone methylation mediates both gene silence and activation depending on specific histone residues modified and degree of methylation. For example, histone H3 lysine 27

(H3K27) methylated by EZH2 (Li et al., 2016) and H3K9 methylated by SUV39H1

(Peters et al., 2001) are repressive histone marks which repress gene transcription.

H3K4 methylated by SET1 and H3K36 methylated by SET2 are histone activating marks which activate gene expression (Keogh et al., 2005, Kim and Buratowski, 2009).

The enrichment of different histone methylation at UCP1 promoter region mediates tissue-specific UCP1 gene transcription. For example, increased H3K9 dimethylation 97

at UCP1 promoter in white fat lead to UCP1 gene repression (Kiskinis et al., 2007).

Whereas increased H3K4me3 at UCP1 promoter in BAT leads to UCP1 gene activation in BAT (Shore et al., 2010).

We previously found that β adrenergic stimuli decreased the enrichment of H3K27 me3 in UCP1 promoter in differentiated brown adipocytes (Zha et al., 2016).

Ubiquitously transcribed tetratricopeptide repeat X chromosome (UTX) is a histone demethylase that demethylates di- or tri-methylated histone H3 lysine 27

(H3K27me2/3) (Rath et al., 2018). Knockdown of UTX increases the enrichment of

H3K27me3 at UCP1 promoter and leads to a decrease in the UCP1 gene expression in cultured brown adipocytes (Zha et al., 2015). However, it is unknown how UTX regulates brown fat thermogenesis and whole body energy homeostasis.

In this study, we identify UTX as an important epigenetic target in regulating the brown fat reprogram during obesity. Using Cre/lox recombination system, we specifically deleted UTX in mature brown adipocyte (UTXf/f-UCP1-Cre+/-). We demonstrated that brown adipocyte-specific UTX deficiency repressed transcription of brown adipocyte-specific genes through regulating the methylation status of H3K27 on promoters of UCP1 and PRDM16. In contrast, brown adipocyte-specific UTX deficiency activated transcription of myogenic genes in brown adipose tissues. Overall, our data suggest that histone methylation plays an important role in brown adipocyte thermogenesis, and UTX may be an important regulator during this process.

Understanding the role of UTX in brown adipocyte function and insulin sensitivity will enable us to develop novel and targeted therapeutic interventions for metabolic 98

diseases.

4.3 Materials and methods

4.3.1 Mice

Brown adipocyte-specific UTX knockout mice were generated by crossing UTX- floxed mice (JAX stock #021926) with UCP1-Cre mice. The UTX-floxed mice were generated by inserting two loxP sites on both sides of exon 3 of UTX (Welstead et al.,

2012). C57BL/6J (B6) mice (Jackson Laboratories) were used in some experiments.

All mice were fed chow diet (LabDiet 5001, fat content 13.5% by calorie) or HF diet

(D12492, 60% calories from fat, Research Diets Inc.). All mice were housed with a

12/12 h light–dark cycle in temperature- and humidity-controlled rooms with free access to water and food. To study the role of UTX in thermogenic activity of brown adipocytes, 8-week old mice were exposed to cold condition (5°C) up to seven days.

Body temperature was monitored during cold exposure. After cold exposure, mice were sacrificed. Brown adipose tissues and white adipose tissues were harvested for measuring gene expression or protein levels. All aspects of animal care were approved by Georgia State University’s Animal Care and Use Committee.

4.3.2 small interfering RNA (siRNA) transfection

The ON-TARGET plus mouse UTX siRNA-SMART pool (L-042844-01-0005), and

PRDM16 siRNA-SMART pool (L-041318-01-0005) were purchased from GE

Healthcare. The overexpressing plasmid of mouse PRDM16 was obtained from

Addgene (Plasmid #15504) (Kajimura et al., 2008, Gupta et al., 2010). The overexpressing plasmids or siRNAs were transfected into BAT1 brown adipocytes by 99

Amaxa Nucleofector II Electroporator (Lonza) using Amaxa cell line nucleofector kit

L according to the manufacturer's instructions (Lonza). Briefly, at days 4–6 of differentiation, cells (2 × 106 cells/sample) were trypsinized and centrifuged at 90 × g for 5 min at room temperature. The cell pellet was resuspended in nucleofector solution (100 μl/sample) with 2 μg of plasmid DNA or 20 pmol of siRNA and seeded into 24-well plates. Cells were treated with isoproterenol two days after transfection.

4.3.3 RNA sequencing

Total RNA was isolated from brown adipose tissues (BATs) of chow-fed UTX KO and

CONTROL mice as noted above. Subsequently, total RNA was treated with RNase-free

DNase I to remove residual DNA. mRNA was purified from 2 μg of total RNA using beads containing oligo (dT). mRNA Libraries were prepared according to the

Illumina’s protocol as previously described (Peng et al., 2012, Ren et al., 2012).

4.3.4 Metabolic measurements, Histological analysis, molecular analysis and Statistical Analysis

All metabolic phenotyping, the histological analysis including HE staining, IF and IHC, and molecular analysis including ChIP, real-time PCR, Immunoblotting and Bisulfite conversion/pyrosequencing, and Statistical Analysis were performed as described in

Methods section of SA2. 100

4.4 Results

4.4.1 BAT-specific UTX deletion reprograms brown fat to a muscle-like phenotype and reduces BAT thermogenic function

We previously reported that UTX knockdown reduces mRNA levels of brown specific genes including UCP1, while overexpression of UTX does the opposite in cultured brown adipocyte cell lines (BAT1 cells) (Zha et al., 2015). However, it remains unknown whether UTX regulates BAT thermogenic function and whole body energy homeostasis in vivo. We first measured the UTX expression pattern in brown and white adipose tissues. As expected, UTX was highly expressed in BAT, but its expression in subcutaneous WAT (SQ) and epididymal WAT (EPI) was low in adult mice housed at ambient temperature (Fig. 21A). In additional, cold stress stimulated

UTX gene expression in BAT (Fig. 21B). We then analyzed UTX KO mice, which we created by breeding UTXfl/fl mice with UCP1-Cre mice, leading to UTX deletion in

BAT. UTX mRNA was 88% lower in BAT of UTX KO mice compared with that of fl/fl mice (Fig. 21C).

We compared gene expression in BATs between UTX KO and CONTROL mice by using RNA sequencing. We identified 254 upregulated genes in BATs from UTX KO mice compared to that of control mice fed HF diet for three months. To our surprise, among the upregulated genes, most of them are involved in skeletal myogenesis such as Myod1, CKM (creatine kinase, muscle), Myh1 (myosin heavy polypeptide 1), Myh4,

Atp2a1 (ATPase Ca++ transporting, cardiac muscle fast twitch 1), and myogenin 101

(Myog) (Fig. 21D). This finding was confirmed when real-time PCR was performed

(Fig. 21E-F). UTX deletion increased myogenic gene expression in brown adipocytes of mice fed either HF diet (Fig. 21E) or chow diet (Fig. 21F). Similar to increased mRNA levels of myogenic genes, increase in Myh1 protein was observed in BAT of UTX

KO mice (Fig. 21G). In contrast to this increase, we observed a decrease in brown thermogenic gene expression in BATs of UTX KO mice including UCP1, CPT1B, PR domain-containing protein 16 (PRDM16), PGC1α, PGC1β, otopetrin 1 (OTOP1), cell death-inducing DFFA-like effector a (CIDEA), cytochrome c oxidase 1 (COX1), DIO2 and PPARγ (Fig. 21H). These data strongly suggested that loss of UTX switches brown thermogenesis to myogenesis leading to dysfunctional BAT.

4.4.2 BAT-specific UTX deletion impairs cold-induced thermogenesis

As shown above, UTX deletion remodeled brown adipocytes to a skeletal muscle-like phenotype, leading to BAT dysfunction. This finding was confirmed when we determined the cold sensitivity in UTX KO mice. Considering that Cold stress amplifies the need for UCP1-mediated thermogenesis, we placed UTX KO mice and fl/fl mice under cold condition (5 °C) for up to seven days. In contrast to similar mRNA levels (data not shown), UCP1 protein in BAT was significantly decreased in UTX KO mice compared to fl/fl mice under cold condition (Fig. 22A). In contrast, UTX deletion in BAT significantly stimulated transcription of myogenic genes including Myod1,

Myog, CKM, Myh1 and Myh4 in BAT of mice housed at RT (Fig. 22B) or cold condition

(Fig. 22C). These data demonstrated that UTX deletion impaired cold-induced brown fat thermogenesis. 102

4.4.3 Mice with BAT-specific UTX deletion are prone to diet-induced obesity and metabolic disorders.

To further investigate the importance of UTX in the regulation of brown fat thermogenic function and its consequences on whole body metabolism, BAT-specific

UTX KO and fl/fl mice were fed either chow or high-fat diet (HFD) at ambient temperature. Similar body weight was observed between male UTX KO and fl/fl mice fed chow diet (Fig. 23A). However, male UTX KO mice exhibited lower percentage of lean mass and higher percentage of fat mass (Fig. 23B) which was primarily due to increased BAT and WAT depots (Fig. 23C). Morphology analysis showed enlarged adipocytes in BAT and WAT of UTX KO mice compared to that of fl/fl mice (Fig. 23D).

Similar results were observed in HFD-fed mice. Male BAT-specific UTX KO mice were more obese than fl/fl mice after four weeks of HF diet feeding (Fig. 23E). Mice with

UTX deletion had higher percentage of fat mass and lower percentage of lean mass compared with fl/fl mice (Fig. 23F). In addition, BAT and SQ WAT of UTX KO mice had increased mass (Fig. 23G) and contained larger lipid droplets in adipocytes than that of fl/fl mice (Fig. 23H). These data indicate that brown fat UTX is important in regulating energy homeostasis during diet-induced obesity. On the other hand, female

BAT-specific UTX KO mice did not display significant obesity phenotype on HFD compared to fl/fl littermates (data not shown).

The importance of UTX on regulating brown fat thermogenic activity was further determined by measuring whole body metabolic rate using the TSE PhenoMaster metabolic cage system. Compared to male fl/fl mice, male BAT-specific UTX KO mice 103

fed HFD exhibited decreased oxygen consumption, CO2 production and energy expenditure (Fig. 24A-C). There was no change in physical activity (Fig. 24D) and food intake (Fig. 24E) between UTX KO and fl/fl mice. These data show that obesity phenotype observed in BAT-specific UTX deletion is primarily due to decreased energy expenditure, possibly due to dysfunctional brown fat, but not due to changes in physical activity or food intake.

Consistent with increased body weight and decreased whole-body energy expenditure, BAT-specific UTX KO mice also developed insulin resistance. At 5-month age, HFD-fed male UTX KO mice exhibited increased fasting blood glucose and serum insulin levels, whereas there was no difference in glucose and insulin levels in UTX KO mice under fed condition compared to their respective fl/fl mice (Fig. 25A). Mice with

BAT-specific UTX deletion also exhibited severe glucose intolerance and insulin resistance (Fig. 25B and C). Consistent data were shown in chow-diet fed mice. When fed chow diet, UTX KO mice had increased blood glucose under both fasted and fed condition, and increased serum insulin level under fed condition compared to fl/fl mice (Fig. 25D). In addition, BAT-specific UTX KO mice displayed mild glucose intolerance when compared to fl/fl mice (Fig. 25E). The increased blood glucose level in UTX KO mice was due to insulin resistance but not any impairment in glucose- induced insulin secretion, as serum insulin level was significantly higher in KO mice

10 minutes after glucose injection during a GTT test than that of fl/fl mice (Fig. 25F).

These results indicate that UTX deletion in brown fat improves diet-induced obesity and insulin resistance in vivo. 104

4.4.4 UTX regulates BAT/skeletal muscle programming via targeting

PRDM16 and Myod1.

Since UTX catalyzes demethylation of H3K27me3, we asked whether UTX deletion alters methylation status of brown fat marker genes. As expected, ChIP-PCR analysis showed that UTX knockdown dramatically increased the enrichment of H3K27me3 at

UCP1 enhancer region in cultured brown adipocytes (BAT1 cells) treated with isoproterenol (Fig. 26A). This may contribute to the decrease in UCP1 gene transcription in brown adipocytes of UTX KO mice. We identified 1053 down- regulated genes in BAT of the BAT-specific UTX KO mice by using RNA sequencing

(data are not shown). Among these down-regulated genes, there are several brown specific marker genes including PRDM16, which was further confirmed by using real- time PCR (Fig. 21H). In addition, we found that isoproterenol treatment reduced the enrichment of H3K27me3 at PRDM16 proximal promoter, which was reversed by UTX knockdown in BAT1 cells (Fig. 26B). These data suggest that UTX may directly regulate the expression of brown specific genes through modifying histone methylation status on these genes, including PRDM16.

In contrast to decreased mRNA levels of brown specific genes, increased mRNA levels of myogenic genes were observed (Fig. 21) in brown adipocytes of UTX KO mice. This suggests that UTX might indirectly regulate myogenic gene expression. Considering the cooperation between histone methylation and DNA methylation,, and the phenotypic similarity between BAT-specific UTX and DNMT1 KO mice, we then investigated the possible involvement of DNA methylation in this process. PRDM16 105

contains N-terminal PR domain, zinc-fingers and other sites that are essential for protein-protein interaction. PRDM16 recruits transcriptional factors such as PGC1α and CEBPβ to UCP1 gene promoters, which activates UCP1 gene expression. It is reported that PRDM16 deficiency in brown preadipocytes promotes muscle like differentiation, whereas ectopic expression of PRDM16 in myoblasts converts myoblasts to brown adipocytes (Seale et al., 2007, Seale et al., 2008, Seale et al., 2011).

These findings suggest that PRDM16 plays an important role in determining cell fate switching between brown adipocytes and skeletal muscle cells. It is recently reported that PRDM16 increases DNA methylation levels of Myod1 gene. However, the specific upstream and downstream regulators of PRDM16 remains to be investigated in brown adipocyte thermogenesis program. In our experiments, we found that DNMT1 KO and

UTX KO mice exhibited a similar physiological response to cold stress and over nutrition supplement. This suggests that there might be cross-talk between UTX and

DNMT1 in the regulation of cell fat switching between brown adipocytes and skeletal muscle cells. We asked whether PRDM16 is the important key linking UTX and

DNMT1 to Myod1 promoter. We found that PRDM16 knockdown reduced the binding of DNMT1 at Myod1 promoter (Fig. 27A). Isoproterenol treatment dramatically increased DNMT1 binding to Myod1 promoter, which was reduced by PRDM16 knockdown (Fig. 27A). In addition, ectopic expression of PRDM16 significantly increased DNA methylation levels at Myod1 promoter (Fig. 27B and C). Together, these data indicate that in BAT under normal condition, UTX expression results in increased PRDM16 expression, possibly due to reduced H3K27me3 level on PRDM16 106

promoter. PRDM16 in turn recruits DNMT1 to myod1 gene and inhibit myod1 gene expression. Whereas in obese status, reduced UTX expression leads to reduced

PRDM16 expression. This will result in decreased DNMT1 binding to Myod1 promoter, leading to increased Myod1 expression, which may contribute to BAT-skeletal muscle switch in obesity.

4.5 Discussion of specific aim 3 Based on the function of brown adipose tissue (BAT) on energy expenditure, BAT thermogenesis is considered as a potential treatment strategy for metabolic diseases such as obesity and type 2 diabetes (Seale et al., 2007, Rajakumari et al., 2013, Zou et al., 2017). Interestingly, we find that UTX is abundantly expressed in BAT compared to WATs. We previously identified the important role of UTX in the regulation of brown specific gene expression in cultured brown adipocytes (Zha et al., 2015).

However, it is still unknown whether brown fat UTX regulates cold or high-fat diet

(HFD)-induced thermogenesis in animal models. High-fat diet induces 1) brown fat thermogenesis (Feldmann et al., 2009, Jun et al., 2014) and 2) increases circulating energy substrates including free fatty acids, TGs and glucose (Fu et al., 2015a, b).

Obesity and insulin resistance occurs when the increase in brown fat activity cannot compensate for the increase in circulating energy substrates including FFAs, TG and glucose (Bartelt et al., 2011). Moreover, we previously found that prolonged HFD feeding leads to switch of cell fate towards skeletal muscle cells in BAT, which causes

BAT dysfunction. Here, we find that loss of brown adipocyte UTX mimics HFD- 107

induced BAT dysfunction, insulin resistance and obesity. Elevated blood glucose levels are observed in UTX KO mice compared to their control littermates. Insulin resistance and glucose intolerance are consequently observed in BAT-specific UTX deficient mice fed a HFD. In addition, brown fat UTX deletion decreases brown thermogenesis based on the observations as followed: 1) downregulation of brown specific gene expression and 2) white-like unilocular adipocytes in BAT of UTX KO mice. BAT dysfunction contributes to HFD-induced obesity and insulin resistance (Lowell et al., 1993). Thus, our date suggests that UTX deletion exacerbates HFD-induced obesity and insulin resistance through reducing brown adipocyte function.

Cold exposure activates the thermogenic function of brown adipocytes (Seale et al.,

2008, Ishibashi and Seale, 2010, Wu et al., 2012). Here, we find that UTX gene expression is upregulated by cold stress in both BAT and WAT (epi), which is consistent with our previous finding (Zha et al., 2015). UTX deficiency downregulated cold-induced brown fat-specific gene expression including UCP1 and PRDM16 in BATs.

Thus, these data further confirm the importance of UTX in the regulation of brown fat thermogenesis (Zha et al., 2015). Two types of brown adipocytes exist: classical brown adipocytes in BATs and beige cells (brown-like) in WATs induced by β adrenergic agonist or cold stress (Seale et al., 2008, Ishibashi and Seale, 2010, Wu et al., 2012).

Although brown and beige adipocytes exhibit similar morphology and UCP1 expression, each cell has distinct cell lineage origins. Classical brown adipocytes and skeletal muscle cells derive from common myogenic factor 5 (Myf5)-expressing progenitors (Seale et al., 2008, Ohno et al., 2013, Yin et al., 2013), whereas beige 108

adipocytes are derived from a smooth muscle-like origin (Wang et al., 2013b, Long et al., 2014, Berry et al., 2016). In present studies, we focus on brown adipose tissues and brown fat cell line (BAT1) as in vivo and in vitro models to study the molecular mechanism regulating BAT and skeletal muscle switch. It is currently unknown whether UTX also regulates beige beige adipocyte function similarly to its role in brown adipocytes. Further studies are required to study the role of UTX in regulating beige adipocyte function.

Histone methylation is an epigenetic mechanism that mediates both gene silence and gene activation depending on which lysine or arginine residues are modified (Zhang and Reinberg, 2001, Ge, 2012). For example, H3 lysine 4 (H3K4me3) is a histone transcriptional activation mark (Benayoun et al., 2014), whereas H3K27me3 is a histone transcriptional repression mark (Li et al., 2016). Histone methylation is regulated by balanced action of histone methyltransferase and demethylases. For example, UTX is a histone demethylase that targets H3K27me2 and H3K27me3 (Zha et al., 2015). Here, we focus on brown specific mark PRDM16, because it is one of the major transcription factors required for brown adipocyte differentiation. As expected, we find that UTX regulates the enrichment of H3K27me3 at PRDM16 promoter.

Indeed, deletion of UTX increases enrichment of H3K27me3 at PRMD16 gene and inhibits its gene transcription. Thus, our data indicate that UTX promotes PRDM16 gene expression through catalyzing demethylation of H3K27me3 on PRDM16 promoter. Indeed, we find that loss of UTX also increased enrichment of H3K27me3 at UCP1 enhancer. Thus, it is interesting to study whether UTX also regulates 109

H3K27me3 demethylation at other BAT-specific gene promoters to regulate relative gene expression. Further studies are required to study UTX-targeted pathways in BAT using ChIP-sequencing analysis with antibodies against UTX or H3K27me3, which allows us to identify other UTX-targeted pathways in regulating brown adipocyte function.

All cells share the same genome, but each has distinct transcriptional profiles which determine cell fate in different organs. Brown adipocyte and skeletal muscle cell derive from same lineage origin (Yin et al., 2013). Brown adipocyte activation is under control of transcriptional factors. For example, transcriptional factors including PRDM16 and

Myod1 function as reciprocal switches to regulate cell lineage determination between brown adipogenesis and myogenesis. PRDM16 promotes brown adipocyte determination (Seale et al., 2008, Becerril et al., 2013, Ishibashi and Seale, 2015), whereas Myod1 promotes skeletal muscle cell determination (Liu et al., 2013, Yin et al., 2013, Wang et al., 2017). Recent studies discover the important role of epigenetic modification in the regulation of satellite cell differentiation to skeletal muscle cells

(Tosic et al., 2018). However, epigenetic mechanisms regulating cell fate switch between brown adipocyte and skeletal muscle cell determination in brown adipose tissue are not yet fully understood. Interestingly, we find that UTX deletion promotes myogenic gene expression such as Myod1, Myh1 and Myog in brown adipocytes. Thus, our data suggest that UTX deletion may impair brown fat thermogenesis through promoting skeletal muscle cell determination in BAT. Indeed, RNA-seq analysis reveals that majority of upregulated genes induced by UTX deletion in BATs are 110

involved in skeletal muscle cell differentiation. Moreover, HFD induces a similar skeletal muscle gene expression pattern in BAT, which is further increased by brown fat deletion of UTX. Thus, our data suggest that UTX deletion leads to classical brown fat remodeling, which may be a maladaptive mechanism that contributes to obesity and insulin resistance. However, it should also be noted that UTX might also target other biological process to regulate BAT function. For example, we find that at the early stage of HFD feeding (12 weeks), UTX deletion impairs BAT function by inducing transcription of skeletal myocyte maker genes. Whereas at the late stage of HF feeding

(24 weeks), UTX deletion further enhances the HFD-mediated increase in immune cell marker genes (RNA-seq analysis, data not shown), indicating that UTX deficiency in BAT may also cause immune cell infiltration into BAT that may further lead to BAT dysfunction. Additional studies are needed to determine the role of UTX in the regulation of HFD-induced inflammation in BAT.

As an active epigenetic regulator, UTX may not directly regulate key myogenic transcription factor genes in BAT. Here, we find that BAT-specific DNMT1 deficient mice and BAT-specific UTX deficient mice display similar metabolic phenotypes based on our observations, as they both developed: 1) obesity, 2) insulin resistance, 3) cold sensitivity, 4) brown adipocyte remodeling to skeletal muscle-like cells. Our data suggest the presence of a possible cross-talk between UTX and DNMT1. It is reported that ectopic expression of PRDM16 significantly increase Myod1 DNA methylation levels and inhibit Myod1 gene expression in differentiated satellite cells (Li et al., 2015).

However, the precise molecular mechanism remains unclear. In the present study, we 111

find that PRDM16 knockdown reduces binding of DNA methyltransferase DNMT1 to

Myod1 promoter. Thus, our data suggest that PRDM16 may indirectly regulate Mydod1

DNA methylation through cooperating with DNMT1. Additional study using co-IP is needed to determine whether PRDM16 protein cooperates with DNMT1 to regulate

Myod1 expression in brown adipocytes. In the current study, we examined the binding of DNMT1 in Myod1 promoter. It is unclear whether DNMT3A or DNMT3B are also involved in the regulation of Myod1 DNA methylation levels. Additional studies such as ChIP assay are needed to determine whether PRDM16 recruits DNMT3A or

DNMT3B to bind to Myod1 promoters and regulate its DNA methylation. Taken together, our results demonstrate that DNA methylation and histone methylation cooperate and regulate gene transcription (Cedar and Bergman, 2009, Zhou et al.,

2018).

In summary, we have identified UTX as a positive regulator of brown adipocyte thermogenesis. UTX is abundantly expressed in BAT and upregulated by β adrenergic activation in BAT. Brown fat deletion of UTX prevents brown thermogenic gene expression, whereas promotes skeletal muscle-specific gene expression. In addition, we find that UTX increases brown thermogenic gene PRDM16 expression through demethylation of H3K27me3 at PRDM16 promoter. PRDM16 in turn recruits DNA methyltransferase DNMT1 to Myod1 promoter to increase Myod1 DNA methylation levels and inhibit its gene expression. This UTX-PRDM16-DNMT1-Myod1 pathway contributes to the maintenance of brown adipocyte thermogenic pathway and the suppression of myogenic program in BAT. Furthermore, BAT-specific UTX KO mice 112

display obese and insulin resistance on a HFD, including increased body weight and fat mass, dysfunctional BAT and decreased energy expenditure. Thus, our data demonstrate that UTX positively regulates brown adipocyte thermogenesis through directly inducing brown specific gene PRDM16 and indirectly cooperating with

DNMT1 to repress myogenic gene expression including Myod1. Our findings not only bring new insight into epigenetic regulation of brown adipose tissue function but also provide a possible therapeutic target for the prevention and treatment of obesity and diabetes.

113

4.6 Figures of specific aim 3

Figure 21 Brown specific-UTX deficiency up-regulates myogenic genes and down-regulates brown-selective genes in brown adipocytes.

(A) UTX mRNA levels in brown adipose tissues (BATs), subcutaneous (SQ) and epididymal (EPI) WAT from C57BL/6J (B6) mice fed chow; (B) UTX mRNA levels in BATs and EPI of mice housed at room temperature (RT) or under cold condition for seven days. (C) UTX mRNA levels in BATs of UTXf and UTXf-UCP1-Cre+/- mice. (D) Heat map comparison of genes that are significantly up-regulated in UTX-deficient BAT compared to the control in HF diet-fed mice. mRNA levels of myogenic genes in BATs from UTX KO and CONTROL mice fed HF (E) or Chow diet (F). (G) Immunocytochemistry of skeletal myosin heavy chain (Myh1) expression in BATs from UTX KO and CONTROL mice fed chow diet. Myh1 fluorescence in green; mRNA levels of Brown-specific genes (H) in BATs from HF diet-fed CONTROL and UTX KO mice. Mean ± SEM. n=7-9. *p<0.05.

114

Figure 22 UTX KO in BAT reduces cold-induced thermogenesis in BATs.

(A) UCP1 protein levels in BATs from control and UTX KO mice exposed to the cold environment for seven days. Mean ± SEM. (B) mRNA levels of myogenic genes in BATs from control mice housed at room temperature (RT) or exposed to the cold environment for seven days. (C-D) mRNA levels of myogenic genes in BATs from control and UTX KO mice housed at room temperature (RT) (C) or exposed to the cold environment (D) for seven days. Mean ± SEM. n=6. *p<0.05. 115

Figure 23 UTX KO mice display obese phenotype.

(A) Body weight curves of male chow diet-fed control (UTX1f/f) and UTX KO mice; (B) Body composition of male control and UTX KO mice on chow diet. Mean ± SEM; (C) Individual fat depots of male control and UTX KO mice on chow diet; (D) Hematoxylin and eosin staining of brown adipose tissue (BAT), subcutaneous fat (SQ), and epididymal (EPI) from control and UTX KO mice fed chow diet. (E) Body weight curves and image of male HF diet-fed control (UTX1f/f) and UTX KO mice; (F) Body composition of male control and UTX KO mice on HF diet. (G) Individual fat depots of male control and UTX KO mice on HF diet; (H) Hematoxylin and eosin staining of BAT, SQ, and EPI from control and UTX KO mice fed HF diet. Data are expressed as Mean ± SEM. n=7 for HF diet-fed mice, and n=9 for chow diet-fed mice. *p<0.05.

116

Figure 24 Low metabolic respiration in UTX KO mice.

(A) Oxygen consumption rate, (B) carbon dioxide production rates, (C) energy expenditure, (D) respiratory exchange ratio (RER), (E) physical activity in HF diet-fed control and UTX KO mice. Dark cycle in dark bar, light cycle in light bar. Data are expressed as Mean ± SEM. n=4. *p<0.05.

117

Figure 25 UTX KO mice display insulin resistance.

(A) Glucose levels and amounts of serum insulin at the fasted and fed states from control and UTX KO mice fed HF diet. (B) Glucose tolerance test in 21-week HF diet-fed mice. (C) Insulin tolerance test in 22-week HF diet-fed mice. Data are expressed as Mean ± SEM. n=7. *p<0.05. (D) Glucose levels and amounts of serum insulin at the fasted and fed states from control and UTX KO mice fed chow. (E) Glucose tolerance test in 16- week chow diet-fed mice. (F) Glucose levels and amounts of serum insulin 10 min after i.p. injection of 0.75g/kg glucose; Data are expressed as Mean ± SEM. n=9. *p<0.05, n.s. not statistically significant.

118

Figure 26 UTX siRNA increased enrichment of H3K27me3 in promoters of PRDM16 and UCP1.

H3K27me3 levels in UCP1 enhancer region (A) and PRDM16 promoter (B) in BAT1 cells. Differentiated BAT1 brown adipocytes were transfected with scramble or UTX siRNA. Two days later, cells were treated with or without isoproterenol (Isop; 1 μM) for 3 h, and a ChIP assay was performed as described under “Materials and Methods.” Data are expressed as mean ± SEM; n = 4. Different letters above bars represent statistical significance between each other.

119

Figure 27 Methylation levels of Myod1 promoters were regulated by UTX and PRDM16.

(A) Differentiated BAT1 brown adipocytes were transfected with scramble or PRDM16 siRNA. Then cells were treated with PBS or Isoproterenol for 3 hours. Cells were collected for a ChIP assay to measure DNMT1 protein binding to myod1 promotors. Different lower letters above the bars represent statistical significance. (B and C) Differentiated BAT1 brown adipocytes were transfected with pSPORT6- or pSPORT6-PRMD16- overexpressing plasmids DNA was isolated from BAT1 cells and methylation levels in proximal promoters near TSS and DMRs about 25kb upstream of TSS of Myod1 were measured as described under “Materials and Methods.” Data are expressed as Mean ± SEM. n=4 *p<0.05. 120

5 GENERAL DISCUSSION

Obesity epidemic is associated with increased risks for type 2 diabetes, hypertension, hyperlipidemia and other metabolic diseases. Obesity and its related metabolic diseases are associated by BAT dysfunction, suggesting that activating BAT function may be an effective therapy for treating obesity. However, the molecular mechanisms contributing to BAT dysfunction that occurs with high fat diet (HFD)-induced obesity are not clear. The goal of this dissertation was to test our overarching hypothesis that epigenetic modification in brown adipocytes regulates energy homeostasis and obesity development. Specially, my dissertation asked the following questions: 1) what is the function of Histone Deacetylase 1 (HDAC1) in regulating thermogenic program in brown adipocytes? 2) Does DNA methyltransferase 1 (DNMT1) regulate brown fat reprogramming and whole-body energy homeostasis in obesity? 3) Does lysine demethylase 6A (UTX) regulate brown fat reprograming and whole-body energy homeostasis in obesity? Here, we used mouse models and cultured brown adipocyte cell lines (BAT1) to elucidate the significance of epigenetic regulators (HDAC1, DNMT1 and UTX) in brown adipocyte thermogenesis. Our findings further support the newly emerging research area that epigenetic modification plays an important role in brown adipocyte function and energy homeostasis (Ohno et al., 2013, Pan et al., 2015, Lin and Farmer, 2016, You et al., 2017). Our conclusion is based on three fundamental findings: 121

1). HDAC1 negatively regulates brown adipocyte thermogenesis via directly regulating H3K27 deacetylation and indirectly regulating H3K27 methylation

Chromatin organization is dynamically regulated by histone modifications, most of which are acetylation and methylation at lysine residues of histone tails (Bannister and

Kouzarides, 2011). Increased acetylation is associated with gene activation, whereas decreased acetylation is associated with gene repression (Maunakea et al., 2010).

HDAC1 is a histone deacetylase that removes the acetyl group from lysine in the amino termini of histone. Here, we demonstrate that the mRNA and protein levels of HDAC1 are lower in mice brown compared to white adipose tissues. In brown adipose tissue, the reduced levels of HDAC1 are also observed during β3-adrenergic stimulation and during brown adipocyte differentiation. HDAC1 deletion in brown adipocytes induces transcription of brown thermogenic genes, including uncoupling protein 1 (UCP1) and peroxisome proliferator-activated receptor gamma co-activator 1-alpha (PGC1α). This effect is concurrent with an increase in acetylation at lysine 27 of histone H3 (H3K27) and a decrease in H3K27 methylation at brown thermogenic gene locus though interacting with demethylase UTX, and polycomb repressive complexes, including

EZH2, SUZ12 and RNF2, which together are indicative of increased thermogenic gene expression. These findings identify histone deacetylase HDAC1 as a negative regulator of brown thermogenic genes, suggesting that HDAC1 may be a novel therapeutic target in the treatment of metabolic syndrome, including diabetes and obesity. 122

2). DNA methyltransferase 1 (DNMT1) mediates brown fat reprogramming and regulates whole-body energy homeostasis in obesity.

HDAC1 cooperates with enzymes related to DNA methylation (Russo et al., 2013), which is another epigenetic modification that regulates gene transcription (Kiskinis et al., 2007). Our work unveils new insight about DNMT1 deletion causing classical brown fat remodeling to skeletal muscle-like cells, which may be a maladaptive mechanism contributing to obesity and insulin resistance. DNMT1 deletion prevents brown fat-specific gene expression; whereas promotes myogenic gene expression in brown adipocytes. We find that DNMT1 deficiency mice exhibit cold intolerance, which is due to an impairment of classical brown fat thermogenic function.

Furthermore, loss of DNMT1 in BAT reduced energy expenditure and induced an obese phenotype. As a consequence, insulin resistance and glucose intolerance are observed in BAT-specific DNMT1 deficient mice. In addition, we find that DNMT1 regulates brown thermogenesis through targeting Myod1, which is an important transcriptional factor involved in skeletal myogenesis. DNMT1 deletion in BAT down- regulates DNA methylation of Myod1 gene and increases its gene transcription. Overall,

DNMT1 deletion mimics high fat diet-induced reprogramming of brown fat to a muscle-like phenotype, which suppresses brown fat thermogenic function and results in obesity and metabolic disorders even in mice fed a regular chow.

3). lysine demethylase 6A (UTX) mediates brown fat reprograming and regulates whole-body energy homeostasis in obesity. 123

Histone methylation is also a key epigenetic modification that regulates brown fat cell thermogenesis (Abe et al., 2015, Zhuang et al., 2018). Here, we find that histone demethylase UTX is important in maintaining brown fat cell phenotype; deletion of

UTX in BAT mimics HFD-induced reprogramming of brown fat to a muscle-like phenotype, suppresses brown fat thermogenic function, and results in diet-induced obesity and metabolic disorders. Interestingly, PRDM16 knockdown dramatically reduces the binding of DNMT1 to Myod1 proximal promoters, leading to an increase in Myod1 gene expression in brown adipocytes. This suggests that PRDM16 may be a key link between histone methylation and DNA methylation at Myod1 promoter to regulate brown adipocyte determination. Previous data suggest that lysine-specific demethylase KDM1a, catalyzing the demethylation of H3K9me3, promotes satellite cells differentiation to skeletal muscle cells (Tosic et al., 2018). Here, we demonstrated that the lysine-spedific demethylase KDM6a/UTX, through a pathway involving

PRDM16-DNMT1-Myod1, maintains brown adipocyte thermogenic program and suppresses skeletal muscle myogenic program in BAT.

In summary, our experiments lead to novel findings. Firstly, HDAC1 knockdown in brown adipocytes increases transcription of brown fat-specific genes. HDAC1 directly regulates deacetylation status and indirectly regulates methylation status of UCP1 and

PGC1α gene via recruiting polycomb repressor complexes including EZH2, SUZ12 and

RNF2 to these genes’ promoters. Secondly, brown adipocyte DNMT1 KO mimics HFD- induced obesity and insulin resistance, which is at least partly due to BAT dysfunction.

BAT-specific DNMT1 deletion remodels BAT through promoting skeletal muscle cell 124

determination in BAT. Myod1 is the direct target of DNMT1 in the regulation of brown adipocyte thermogenesis. Thirdly, brown adipocyte UTX KO mimics HFD-induced obesity and insulin resistance, which is also primarily due to BAT dysfunction. UTX deletion remodels BAT through promoting skeletal muscle cell determination in BAT.

Myod1 is the direct target of DNMT1 in the regulation of brown adipocyte thermogenesis. On the one hand, UTX directly decreases the enrichment of

H3K27me3 in PRDM16 gene promoter and promote its gene expression. In turn,

PRDM16 recruits DNMT1 to Myod1 promoter and increase Myod1 DNA methylation level, which inhibits its gene expression. This pathway (UTX-PRDM16-DNMT1-

Myod1) in turn promotes brown adipocyte determination and prevents skeletal muscle cell determination in BAT.

It is undoubtable that high fat diet (HFD) triggers obesity and insulin resistance

(Matsubara et al., 2012). Most of studies focus on diet-induced hepatic steatosis (Lee et al., 2013), adipose inflammation (Matsubara et al., 2012), (Birse et al., 2010) and muscle dysfunction (Wessels et al., 2015). However, it is not well studied how HFD affects brown adipocyte function. Our experiments show that short-term HFD feeding increases brown fat thermogenesis, which is consistent with previous studies (Kozak,

2010, Kazak et al., 2017). Interestingly, we find that prolonged HFD feeding causes the reprogramming of brown adipocyte to skeletal muscle-like cell in brown adipose tissue

(BAT), which may result in brown adipocyte dysfunction. Our finding suggests that

BAT dysfunction may be at least in part due to prolonged HFD-induced BAT-to- muscle reprogramming. Additionally, brown fat-specific deletion of DNMT1 or UTX 125

mimics HFD-induced BAT-to-muscle reprogramming in obesity. Our data highlight the crucial importance of DNA methylation and histone modification in brown fat reprogramming, which indicates that understanding the role of epigenetic modification in energy regulation and obesity has significant impact in developing ways to prevent and reverse obesity.

126

REFERENCES

Abe Y, Rozqie R, Matsumura Y, Kawamura T, Nakaki R, Tsurutani Y, Tanimura-

Inagaki K, Shiono A, Magoori K, Nakamura K, Ogi S, Kajimura S, Kimura H,

Tanaka T, Fukami K, Osborne TF, Kodama T, Aburatani H, Inagaki T, Sakai J

(2015) JMJD1A is a signal-sensing scaffold that regulates acute chromatin

dynamics via SWI/SNF association for thermogenesis. Nat Commun 6:7052.

An W (2007) Histone acetylation and methylation: combinatorial players for

transcriptional regulation. Subcell Biochem 41:351-369.

Andrikopoulos S, Blair AR, Deluca N, Fam BC, Proietto J (2008) Evaluating the

glucose tolerance test in mice. Am J Physiol Endocrinol Metab 295:E1323-1332.

Backdahl L, Bushell A, Beck S (2009) Inflammatory signalling as mediator of

epigenetic modulation in tissue-specific chronic inflammation. Int J Biochem

Cell Biol 41:176-184.

Baek KW, Cha HJ, Ock MS, Kim HS, Gim JA, Park JJ (2018) Effects of regular-

moderate exercise on high-fat diet-induced intramyocellular lipid

accumulation in the soleus muscle of Sprague-Dawley rats. J Exerc Rehabil

14:32-38.

Bannister AJ, Kouzarides T (2011) Regulation of chromatin by histone modifications.

Cell Res 21:381-395.

Barbatelli G, Murano I, Madsen L, Hao Q, Jimenez M, Kristiansen K, Giacobino JP,

De Matteis R, Cinti S (2010) The emergence of cold-induced brown adipocytes

in mouse white fat depots is determined predominantly by white to brown 127

adipocyte transdifferentiation. Am J Physiol Endocrinol Metab 298:E1244-

1253.

Bartelt A, Bruns OT, Reimer R, Hohenberg H, Ittrich H, Peldschus K, Kaul MG,

Tromsdorf UI, Weller H, Waurisch C, Eychmuller A, Gordts PL, Rinninger F,

Bruegelmann K, Freund B, Nielsen P, Merkel M, Heeren J (2011) Brown

adipose tissue activity controls triglyceride clearance. Nat Med 17:200-205.

Bartness TJ, Ryu V (2015) Neural control of white, beige and brown adipocytes. Int J

Obes Suppl 5:S35-39.

Becerril S, Gomez-Ambrosi J, Martin M, Moncada R, Sesma P, Burrell MA, Fruhbeck

G (2013) Role of PRDM16 in the activation of brown fat programming.

Relevance to the development of obesity. Histol Histopathol 28:1411-1425.

Benayoun BA, Pollina EA, Ucar D, Mahmoudi S, Karra K, Wong ED, Devarajan K,

Daugherty AC, Kundaje AB, Mancini E, Hitz BC, Gupta R, Rando TA, Baker JC,

Snyder MP, Cherry JM, Brunet A (2014) H3K4me3 breadth is linked to cell

identity and transcriptional consistency. Cell 158:673-688.

Berry DC, Jiang Y, Graff JM (2016) Mouse strains to study cold-inducible beige

progenitors and beige adipocyte formation and function. Nat Commun 7:10184.

Betz MJ, Enerback S (2018) Targeting thermogenesis in brown fat and muscle to treat

obesity and metabolic disease. Nat Rev Endocrinol 14:77-87.

Birse RT, Choi J, Reardon K, Rodriguez J, Graham S, Diop S, Ocorr K, Bodmer R,

Oldham S (2010) High-fat-diet-induced obesity and heart dysfunction are

regulated by the TOR pathway in Drosophila. Cell Metab 12:533-544. 128

Borensztein M, Viengchareun S, Montarras D, Journot L, Binart N, Lombes M,

Dandolo L (2012) Double Myod and Igf2 inactivation promotes brown adipose

tissue development by increasing Prdm16 expression. FASEB J 26:4584-4591.

Brehm A, Miska EA, McCance DJ, Reid JL, Bannister AJ, Kouzarides T (1998)

Retinoblastoma protein recruits histone deacetylase to repress transcription.

Nature 391:597-601.

Cannon B, Nedergaard J (1985) The biochemistry of an inefficient tissue: brown

adipose tissue. Essays Biochem 20:110-164.

Cannon B, Nedergaard J (2004) Brown adipose tissue: function and physiological

significance. Physiol Rev 84:277-359.

Cao W, Daniel KW, Robidoux J, Puigserver P, Medvedev AV, Bai X, Floering LM,

Spiegelman BM, Collins S (2004) p38 mitogen-activated protein kinase is the

central regulator of cyclic AMP-dependent transcription of the brown fat

uncoupling protein 1 gene. Mol Cell Biol 24:3057-3067.

Cassard-Doulcier AM, Gelly C, Bouillaud F, Ricquier D (1998) A 211-bp enhancer of

the rat uncoupling protein-1 (UCP-1) gene controls specific and regulated

expression in brown adipose tissue. Biochem J 333 ( Pt 2):243-246.

Cedar H, Bergman Y (2009) Linking DNA methylation and histone modification:

patterns and paradigms. Nat Rev Genet 10:295-304.

Cederberg A, Gronning LM, Ahren B, Tasken K, Carlsson P, Enerback S (2001) FOXC2

is a winged helix gene that counteracts obesity, hypertriglyceridemia, and diet-

induced insulin resistance. Cell 106:563-573. 129

Chalkiadaki A, Guarente L (2012) High-fat diet triggers inflammation-induced

cleavage of SIRT1 in adipose tissue to promote metabolic dysfunction. Cell

Metab 16:180-188.

Chao LC, Bensinger SJ, Villanueva CJ, Wroblewski K, Tontonoz P (2008) Inhibition

of adipocyte differentiation by Nur77, Nurr1, and Nor1. Mol Endocrinol

22:2596-2608.

Chavan AV, Somani RR (2010) HDAC inhibitors - new generation of target specific

treatment. Mini Rev Med Chem 10:1263-1276.

Chen YS, Wu R, Yang X, Kou S, MacDougald OA, Yu L, Shi H, Xue B (2016a) Inhibiting

DNA methylation switches adipogenesis to osteoblastogenesis by activating

Wnt10a. Sci Rep 6:25283.

Chen YT, Liao JW, Tsai YC, Tsai FJ (2016b) Inhibition of DNA methyltransferase 1

increases nuclear receptor subfamily 4 group A member 1 expression and

decreases blood glucose in type 2 diabetes. Oncotarget.

Choudhury SR, Cui Y, Lubecka K, Stefanska B, Irudayaraj J (2016) CRISPR-dCas9

mediated TET1 targeting for selective DNA demethylation at BRCA1 promoter.

Oncotarget 7:46545-46556.

Cohen P, Levy JD, Zhang Y, Frontini A, Kolodin DP, Svensson KJ, Lo JC, Zeng X, Ye

L, Khandekar MJ, Wu J, Gunawardana SC, Banks AS, Camporez JP, Jurczak

MJ, Kajimura S, Piston DW, Mathis D, Cinti S, Shulman GI, Seale P,

Spiegelman BM (2014) Ablation of PRDM16 and beige adipose causes

metabolic dysfunction and a subcutaneous to visceral fat switch. Cell 156:304- 130

316.

Constantinides PG, Jones PA, Gevers W (1977) Functional striated muscle cells from

non-myoblast precursors following 5-azacytidine treatment. Nature 267:364-

366.

Couzin-Franke J (2012) Tackling America's eating habits, one store at a time. Science

337:1473-1475.

Cowherd RM, Lyle RE, Miller CP, McGehee RE, Jr. (1997) Developmental profile of

homeobox gene expression during 3T3-L1 adipogenesis. Biochem Biophys Res

Commun 237:470-475.

Cypess AM, Chen YC, Sze C, Wang K, English J, Chan O, Holman AR, Tal I, Palmer

MR, Kolodny GM, Kahn CR (2012) Cold but not sympathomimetics activates

human brown adipose tissue in vivo. Proc Natl Acad Sci U S A 109:10001-10005.

Cypess AM, Lehman S, Williams G, Tal I, Rodman D, Goldfine AB, Kuo FC, Palmer EL,

Tseng YH, Doria A, Kolodny GM, Kahn CR (2009) Identification and

importance of brown adipose tissue in adult humans. N Engl J Med 360:1509-

1517.

Cypess AM, White AP, Vernochet C, Schulz TJ, Xue R, Sass CA, Huang TL, Roberts-

Toler C, Weiner LS, Sze C, Chacko AT, Deschamps LN, Herder LM, Truchan N,

Glasgow AL, Holman AR, Gavrila A, Hasselgren PO, Mori MA, Molla M, Tseng

YH (2013) Anatomical localization, gene expression profiling and functional

characterization of adult human neck brown fat. Nat Med 19:635-639.

Dambacher S, Hahn M, Schotta G (2010) Epigenetic regulation of development by 131

histone lysine methylation. Heredity (Edinb) 105:24-37.

Elliott EN, Sheaffer KL, Schug J, Stappenbeck TS, Kaestner KH (2015) Dnmt1 is

essential to maintain progenitors in the perinatal intestinal epithelium.

Development 142:2163-2172.

Emmett MJ, Lim HW, Jager J, Richter HJ, Adlanmerini M, Peed LC, Briggs ER, Steger

DJ, Ma T, Sims CA, Baur JA, Pei L, Won KJ, Seale P, Gerhart-Hines Z, Lazar

MA (2017) Histone deacetylase 3 prepares brown adipose tissue for acute

thermogenic challenge. Nature 546:544-548.

Enerback S (2010) Human brown adipose tissue. Cell Metab 11:248-252.

Feldmann HM, Golozoubova V, Cannon B, Nedergaard J (2009) UCP1 ablation

induces obesity and abolishes diet-induced thermogenesis in mice exempt from

thermal stress by living at thermoneutrality. Cell Metab 9:203-209.

Fu L, Bruckbauer A, Li F, Cao Q, Cui X, Wu R, Shi H, Zemel MB, Xue B (2015a)

Interaction between metformin and leucine in reducing hyperlipidemia and

hepatic lipid accumulation in diet-induced obese mice. Metabolism 64:1426-

1434.

Fu L, Bruckbauer A, Li F, Cao Q, Cui X, Wu R, Shi H, Zemel MB, Xue B (2015b) Leucine

amplifies the effects of metformin on insulin sensitivity and glycemic control in

diet-induced obese mice. Metabolism 64:845-856.

Galmozzi A, Mitro N, Ferrari A, Gers E, Gilardi F, Godio C, Cermenati G, Gualerzi A,

Donetti E, Rotili D, Valente S, Guerrini U, Caruso D, Mai A, Saez E, De Fabiani

E, Crestani M (2013) Inhibition of class I histone deacetylases unveils a 132

mitochondrial signature and enhances oxidative metabolism in skeletal muscle

and adipose tissue. Diabetes 62:732-742.

Gao J, Song J, Du M, Mao X (2018) Bovine alpha-Lactalbumin Hydrolysates (alpha-

LAH) Ameliorate Adipose Insulin Resistance and Inflammation in High-Fat

Diet-Fed C57BL/6J Mice. Nutrients 10.

Gao Z, Yin J, Zhang J, Ward RE, Martin RJ, Lefevre M, Cefalu WT, Ye J (2009)

Butyrate improves insulin sensitivity and increases energy expenditure in mice.

Diabetes 58:1509-1517.

Ge K (2012) Epigenetic regulation of adipogenesis by histone methylation. Biochim

Biophys Acta 1819:727-732.

Golozoubova V, Hohtola E, Matthias A, Jacobsson A, Cannon B, Nedergaard J (2001)

Only UCP1 can mediate adaptive nonshivering thermogenesis in the cold.

FASEB J 15:2048-2050.

Griggio MA (1982) The participation of shivering and nonshivering thermogenesis in

warm and cold-acclimated rats. Comp Biochem Physiol A Comp Physiol 73:481-

484.

Guerra C, Koza RA, Yamashita H, Walsh K, Kozak LP (1998) Emergence of brown

adipocytes in white fat in mice is under genetic control. Effects on body weight

and adiposity. J Clin Invest 102:412-420.

Gupta RK, Arany Z, Seale P, Mepani RJ, Ye L, Conroe HM, Roby YA, Kulaga H, Reed

RR, Spiegelman BM (2010) Transcriptional control of preadipocyte

determination by Zfp423. Nature 464:619-623. 133

Handschin C, Rhee J, Lin J, Tarr PT, Spiegelman BM (2003) An autoregulatory loop

controls peroxisome proliferator-activated receptor gamma coactivator 1alpha

expression in muscle. Proc Natl Acad Sci U S A 100:7111-7116.

Haque MR, Ansari HS (2018) Anti-Obesity Effect of Arq Zeera and Its Main

Components Thymol and Cuminaldehyde in High Fat Diet Induced Obese Rats.

Drug Res (Stuttg).

Harms M, Seale P (2013) Brown and beige fat: development, function and therapeutic

potential. Nat Med 19:1252-1263.

Herzig S, Long F, Jhala US, Hedrick S, Quinn R, Bauer A, Rudolph D, Schutz G, Yoon

C, Puigserver P, Spiegelman B, Montminy M (2001) CREB regulates hepatic

gluconeogenesis through the coactivator PGC-1. Nature 413:179-183.

Hibi M, Oishi S, Matsushita M, Yoneshiro T, Yamaguchi T, Usui C, Yasunaga K,

Katsuragi Y, Kubota K, Tanaka S, Saito M (2016) Brown adipose tissue is

involved in diet-induced thermogenesis and whole-body fat utilization in

healthy humans. Int J Obes (Lond) 40:1655-1661.

Hill JO, Wyatt HR, Peters JC (2012) Energy balance and obesity. Circulation 126:126-

132.

Himms-Hagen J (1979) Obesity may be due to a malfunctioning of brown fat. Can Med

Assoc J 121:1361-1364.

Himms-Hagen J (1989) Brown adipose tissue thermogenesis and obesity. Prog Lipid

Res 28:67-115.

Himms-Hagen J, Cui J, Danforth E, Jr., Taatjes DJ, Lang SS, Waters BL, Claus TH 134

(1994) Effect of CL-316,243, a thermogenic beta 3-agonist, on energy balance

and brown and white adipose tissues in rats. Am J Physiol 266:R1371-1382.

Inagaki T, Sakai J, Kajimura S (2016) Transcriptional and epigenetic control of brown

and beige adipose cell fate and function. Nat Rev Mol Cell Biol 17:480-495.

Ishibashi J, Seale P (2010) Medicine. Beige can be slimming. Science 328:1113-1114.

Ishibashi J, Seale P (2015) Functions of Prdm16 in thermogenic fat cells. Temperature

(Austin) 2:65-72.

Jackson-Grusby L, Beard C, Possemato R, Tudor M, Fambrough D, Csankovszki G,

Dausman J, Lee P, Wilson C, Lander E, Jaenisch R (2001) Loss of genomic

methylation causes p53-dependent apoptosis and epigenetic deregulation. Nat

Genet 27:31-39.

Jun HJ, Joshi Y, Patil Y, Noland RC, Chang JS (2014) NT-PGC-1alpha activation

attenuates high-fat diet-induced obesity by enhancing brown fat thermogenesis

and adipose tissue oxidative metabolism. Diabetes 63:3615-3625.

Kajimura S, Seale P, Tomaru T, Erdjument-Bromage H, Cooper MP, Ruas JL, Chin S,

Tempst P, Lazar MA, Spiegelman BM (2008) Regulation of the brown and

white fat gene programs through a PRDM16/CtBP transcriptional complex.

Genes Dev 22:1397-1409.

Kamei Y, Suganami T, Ehara T, Kanai S, Hayashi K, Yamamoto Y, Miura S, Ezaki O,

Okano M, Ogawa Y (2010) Increased expression of DNA methyltransferase 3a

in obese adipose tissue: studies with transgenic mice. Obesity (Silver Spring)

18:314-321. 135

Kang JH, Goto T, Han IS, Kawada T, Kim YM, Yu R (2010) Dietary capsaicin reduces

obesity-induced insulin resistance and hepatic steatosis in obese mice fed a

high-fat diet. Obesity (Silver Spring) 18:780-787.

Kanzleiter T, Wilks D, Preston E, Ye J, Frangioudakis G, Cooney GJ (2009) Regulation

of the nuclear hormone receptor nur77 in muscle: influence of exercise-

activated pathways in vitro and obesity in vivo. Biochim Biophys Acta 1792:777-

782.

Kazak L, Chouchani ET, Lu GZ, Jedrychowski MP, Bare CJ, Mina AI, Kumari M, Zhang

S, Vuckovic I, Laznik-Bogoslavski D, Dzeja P, Banks AS, Rosen ED, Spiegelman

BM (2017) Genetic Depletion of Adipocyte Creatine Metabolism Inhibits Diet-

Induced Thermogenesis and Drives Obesity. Cell Metab 26:693.

Kelly RD, Cowley SM (2013) The physiological roles of histone deacetylase (HDAC) 1

and 2: complex co-stars with multiple leading parts. Biochem Soc Trans 41:741-

749.

Keogh MC, Kurdistani SK, Morris SA, Ahn SH, Podolny V, Collins SR, Schuldiner M,

Chin K, Punna T, Thompson NJ, Boone C, Emili A, Weissman JS, Hughes TR,

Strahl BD, Grunstein M, Greenblatt JF, Buratowski S, Krogan NJ (2005)

Cotranscriptional set2 methylation of histone H3 lysine 36 recruits a repressive

Rpd3 complex. Cell 123:593-605.

Kim AY, Park YJ, Pan X, Shin KC, Kwak SH, Bassas AF, Sallam RM, Park KS, Alfadda

AA, Xu A, Kim JB (2015) Obesity-induced DNA hypermethylation of the

adiponectin gene mediates insulin resistance. Nat Commun 6:7585. 136

Kim SH, Plutzky J (2016) Brown Fat and Browning for the Treatment of Obesity and

Related Metabolic Disorders. Diabetes Metab J 40:12-21.

Kim T, Buratowski S (2009) Dimethylation of H3K4 by Set1 recruits the Set3 histone

deacetylase complex to 5' transcribed regions. Cell 137:259-272.

Kiskinis E, Hallberg M, Christian M, Olofsson M, Dilworth SM, White R, Parker MG

(2007) RIP140 directs histone and DNA methylation to silence Ucp1 expression

in white adipocytes. EMBO J 26:4831-4840.

Klaus S, Choy L, Champigny O, Cassard-Doulcier AM, Ross S, Spiegelman B, Ricquier

D (1994) Characterization of the novel brown adipocyte cell line HIB 1B.

Adrenergic pathways involved in regulation of uncoupling protein gene

expression. J Cell Sci 107 ( Pt 1):313-319.

Ko YG, Nishino K, Hattori N, Arai Y, Tanaka S, Shiota K (2005) Stage-by-stage change

in DNA methylation status of Dnmt1 locus during mouse early development. J

Biol Chem 280:9627-9634.

Kozak LP (2010) Brown fat and the myth of diet-induced thermogenesis. Cell Metab

11:263-267.

Kozak UC, Kopecky J, Teisinger J, Enerback S, Boyer B, Kozak LP (1994) An upstream

enhancer regulating brown-fat-specific expression of the mitochondrial

uncoupling protein gene. Mol Cell Biol 14:59-67.

Kraschnewski JL, Boan J, Esposito J, Sherwood NE, Lehman EB, Kephart DK,

Sciamanna CN (2010) Long-term weight loss maintenance in the United States.

Int J Obes (Lond) 34:1644-1654. 137

Laugesen A, Helin K (2014) Chromatin repressive complexes in stem cells,

development, and cancer. Cell Stem Cell 14:735-751.

Lee KM, Yang SJ, Kim YD, Choi YD, Nam JH, Choi CS, Choi HS, Park CS (2013)

Disruption of the cereblon gene enhances hepatic AMPK activity and prevents

high-fat diet-induced obesity and insulin resistance in mice. Diabetes 62:1855-

1864.

Lee MG, Villa R, Trojer P, Norman J, Yan KP, Reinberg D, Di Croce L, Shiekhattar R

(2007) Demethylation of H3K27 regulates polycomb recruitment and H2A

ubiquitination. Science 318:447-450.

Li F, Wu R, Cui X, Zha L, Yu L, Shi H, Xue B (2016) Histone Deacetylase 1 (HDAC1)

Negatively Regulates Thermogenic Program in Brown Adipocytes via

Coordinated Regulation of Histone H3 Lysine 27 (H3K27) Deacetylation and

Methylation. J Biol Chem 291:4523-4536.

Li X, Wang J, Jiang Z, Guo F, Soloway PD, Zhao R (2015) Role of PRDM16 and its PR

domain in the epigenetic regulation of myogenic and adipogenic genes during

transdifferentiation of C2C12 cells. Gene 570:191-198.

Lim YC, Chia SY, Jin S, Han W, Ding C, Sun L (2016) Dynamic DNA methylation

landscape defines brown and white cell specificity during adipogenesis. Mol

Metab 5:1033-1041.

Lin J, Handschin C, Spiegelman BM (2005) Metabolic control through the PGC-1

family of transcription coactivators. Cell Metab 1:361-370.

Lin JZ, Farmer SR (2016) LSD1-a pivotal epigenetic regulator of brown and beige fat 138

differentiation and homeostasis. Genes Dev 30:1793-1795.

Ling C, Groop L (2009) Epigenetics: a molecular link between environmental factors

and type 2 diabetes. Diabetes 58:2718-2725.

Liu W, Bi P, Shan T, Yang X, Yin H, Wang YX, Liu N, Rudnicki MA, Kuang S (2013)

miR-133a regulates adipocyte browning in vivo. PLoS Genet 9:e1003626.

Liu XS, Wu H, Ji X, Stelzer Y, Wu X, Czauderna S, Shu J, Dadon D, Young RA, Jaenisch

R (2016) Editing DNA Methylation in the Mammalian Genome. Cell 167:233-

247 e217.

Londono Gentile T, Lu C, Lodato PM, Tse S, Olejniczak SH, Witze ES, Thompson CB,

Wellen KE (2013) DNMT1 is regulated by ATP-citrate lyase and maintains

methylation patterns during adipocyte differentiation. Mol Cell Biol 33:3864-

3878.

Long JZ, Svensson KJ, Tsai L, Zeng X, Roh HC, Kong X, Rao RR, Lou J, Lokurkar I,

Baur W, Castellot JJ, Jr., Rosen ED, Spiegelman BM (2014) A smooth muscle-

like origin for beige adipocytes. Cell Metab 19:810-820.

Lowell BB, Spiegelman BM (2000) Towards a molecular understanding of adaptive

thermogenesis. Nature 404:652-660.

Lowell BB, V SS, Hamann A, Lawitts JA, Himms-Hagen J, Boyer BB, Kozak LP, Flier

JS (1993) Development of obesity in transgenic mice after genetic ablation of

brown adipose tissue. Nature 366:740-742.

Luo H, Jiang M, Lian G, Liu Q, Shi M, Li TY, Song L, Ye J, He Y, Yao L, Zhang C, Lin

ZZ, Zhang CS, Zhao TJ, Jia WP, Li P, Lin SY, Lin SC (2018) AIDA Selectively 139

Mediates Downregulation of Fat Synthesis Enzymes by ERAD to Retard

Intestinal Fat Absorption and Prevent Obesity. Cell Metab 27:843-853 e846.

Mack GS (2010) To selectivity and beyond. Nat Biotechnol 28:1259-1266.

Magnaghi-Jaulin L, Groisman R, Naguibneva I, Robin P, Lorain S, Le Villain JP,

Troalen F, Trouche D, Harel-Bellan A (1998) Retinoblastoma protein represses

transcription by recruiting a histone deacetylase. Nature 391:601-605.

Maier S, Olek A (2002) Diabetes: a candidate disease for efficient DNA methylation

profiling. J Nutr 132:2440S-2443S.

Matsubara T, Mita A, Minami K, Hosooka T, Kitazawa S, Takahashi K, Tamori Y, Yokoi

N, Watanabe M, Matsuo E, Nishimura O, Seino S (2012) PGRN is a key

adipokine mediating high fat diet-induced insulin resistance and obesity

through IL-6 in adipose tissue. Cell Metab 15:38-50.

Maunakea AK, Chepelev I, Zhao K (2010) Epigenome mapping in normal and disease

States. Circ Res 107:327-339.

McGuinness OP, Ayala JE, Laughlin MR, Wasserman DH (2009) NIH experiment in

centralized mouse phenotyping: the Vanderbilt experience and

recommendations for evaluating glucose homeostasis in the mouse. Am J

Physiol Endocrinol Metab 297:E849-855.

Morey L, Helin K (2010) Polycomb group protein-mediated repression of

transcription. Trends Biochem Sci 35:323-332.

Nebbioso A, Dell'Aversana C, Bugge A, Sarno R, Valente S, Rotili D, Manzo F, Teti D,

Mandrup S, Ciana P, Maggi A, Mai A, Gronemeyer H, Altucci L (2010) HDACs 140

class II-selective inhibition alters nuclear receptor-dependent differentiation. J

Mol Endocrinol 45:219-228.

Nedergaard J, Golozoubova V, Matthias A, Asadi A, Jacobsson A, Cannon B (2001)

UCP1: the only protein able to mediate adaptive non-shivering thermogenesis

and metabolic inefficiency. Biochim Biophys Acta 1504:82-106.

Ng Y, Tan SX, Chia SY, Tan HY, Gun SY, Sun L, Hong W, Han W (2017) HOXC10

suppresses browning of white adipose tissues. Exp Mol Med 49:e292.

Nguyen NL, Barr CL, Ryu V, Cao Q, Xue B, Bartness TJ (2017) Separate and shared

sympathetic outflow to white and brown fat coordinately regulates

thermoregulation and beige adipocyte recruitment. Am J Physiol Regul Integr

Comp Physiol 312:R132-R145.

Nicholls DG, Locke RM (1984) Thermogenic mechanisms in brown fat. Physiol Rev

64:1-64.

Oelkrug R, Polymeropoulos ET, Jastroch M (2015) Brown adipose tissue:

physiological function and evolutionary significance. J Comp Physiol B

185:587-606.

Ohno H, Shinoda K, Ohyama K, Sharp LZ, Kajimura S (2013) EHMT1 controls brown

adipose cell fate and thermogenesis through the PRDM16 complex. Nature

504:163-167.

Okada Y, Tateishi K, Zhang Y (2010) Histone demethylase JHDM2A is involved in

male infertility and obesity. J Androl 31:75-78.

Onishi S, Ishino M, Kitazawa H, Yoto A, Shimba Y, Mochizuki Y, Unno K, Meguro S, 141

Tokimitsu I, Miura S (2018) Green tea extracts ameliorate high-fat diet-induced

muscle atrophy in senescence-accelerated mouse prone-8 mice. PLoS One

13:e0195753.

Paluch BE, Naqash AR, Brumberger Z, Nemeth MJ, Griffiths EA (2016) Epigenetics:

A primer for clinicians. Blood Rev 30:285-295.

Pan D, Huang L, Zhu LJ, Zou T, Ou J, Zhou W, Wang YX (2015) Jmjd3-Mediated

H3K27me3 Dynamics Orchestrate Brown Fat Development and Regulate White

Fat Plasticity. Dev Cell 35:568-583.

Peng Z, Cheng Y, Tan BC, Kang L, Tian Z, Zhu Y, Zhang W, Liang Y, Hu X, Tan X, Guo

J, Dong Z, Bao L, Wang J (2012) Comprehensive analysis of RNA-Seq data

reveals extensive RNA editing in a human transcriptome. Nat Biotechnol

30:253-260.

Periasamy M, Herrera JL, Reis FCG (2017) Skeletal Muscle Thermogenesis and Its

Role in Whole Body Energy Metabolism. Diabetes Metab J 41:327-336.

Peters AH, O'Carroll D, Scherthan H, Mechtler K, Sauer S, Schofer C, Weipoltshammer

K, Pagani M, Lachner M, Kohlmaier A, Opravil S, Doyle M, Sibilia M, Jenuwein

T (2001) Loss of the Suv39h histone methyltransferases impairs mammalian

heterochromatin and genome stability. Cell 107:323-337.

Pronier E, Almire C, Mokrani H, Vasanthakumar A, Simon A, da Costa Reis Monte

Mor B, Masse A, Le Couedic JP, Pendino F, Carbonne B, Larghero J, Ravanat

JL, Casadevall N, Bernard OA, Droin N, Solary E, Godley LA, Vainchenker W,

Plo I, Delhommeau F (2011) Inhibition of TET2-mediated conversion of 5- 142

methylcytosine to 5-hydroxymethylcytosine disturbs erythroid and

granulomonocytic differentiation of human hematopoietic progenitors. Blood

118:2551-2555.

Puigserver P, Spiegelman BM (2003) Peroxisome proliferator-activated receptor-

gamma coactivator 1 alpha (PGC-1 alpha): transcriptional coactivator and

metabolic regulator. Endocr Rev 24:78-90.

Puigserver P, Wu Z, Park CW, Graves R, Wright M, Spiegelman BM (1998) A cold-

inducible coactivator of nuclear receptors linked to adaptive thermogenesis.

Cell 92:829-839.

Putti R, Migliaccio V, Sica R, Lionetti L (2015) Skeletal Muscle Mitochondrial

Bioenergetics and Morphology in High Fat Diet Induced Obesity and Insulin

Resistance: Focus on Dietary Fat Source. Front Physiol 6:426.

Rajakumari S, Wu J, Ishibashi J, Lim HW, Giang AH, Won KJ, Reed RR, Seale P (2013)

EBF2 determines and maintains brown adipocyte identity. Cell Metab 17:562-

574.

Rath BH, Waung I, Camphausen K, Tofilon PJ (2018) Inhibition of the Histone H3K27

Demethylase UTX Enhances Tumor Cell Radiosensitivity. Mol Cancer Ther.

Ren S, Peng Z, Mao JH, Yu Y, Yin C, Gao X, Cui Z, Zhang J, Yi K, Xu W, Chen C, Wang

F, Guo X, Lu J, Yang J, Wei M, Tian Z, Guan Y, Tang L, Xu C, Wang L, Tian W,

Wang J, Yang H, Sun Y (2012) RNA-seq analysis of prostate cancer in the

Chinese population identifies recurrent gene fusions, cancer-associated long

noncoding RNAs and aberrant alternative splicings. Cell Res 22:806-821. 143

Rim JS, Kozak LP (2002) Regulatory motifs for CREB-binding protein and Nfe2l2

transcription factors in the upstream enhancer of the mitochondrial uncoupling

protein 1 gene. J Biol Chem 277:34589-34600.

Roh HC, Tsai LTY, Shao M, Tenen D, Shen Y, Kumari M, Lyubetskaya A, Jacobs C,

Dawes B, Gupta RK, Rosen ED (2018) Warming Induces Significant

Reprogramming of Beige, but Not Brown, Adipocyte Cellular Identity. Cell

Metab 27:1121-1137 e1125.

Rosell M, Kaforou M, Frontini A, Okolo A, Chan YW, Nikolopoulou E, Millership S,

Fenech ME, MacIntyre D, Turner JO, Moore JD, Blackburn E, Gullick WJ, Cinti

S, Montana G, Parker MG, Christian M (2014) Brown and white adipose tissues:

intrinsic differences in gene expression and response to cold exposure in mice.

Am J Physiol Endocrinol Metab 306:E945-964.

Rosen ED (2016) Epigenomic and transcriptional control of insulin resistance. J

Intern Med 280:443-456.

Ross SR, Choy L, Graves RA, Fox N, Solevjeva V, Klaus S, Ricquier D, Spiegelman BM

(1992) Hibernoma formation in transgenic mice and isolation of a brown

adipocyte cell line expressing the uncoupling protein gene. Proc Natl Acad Sci

U S A 89:7561-7565.

Rothwell NJ, Stock MJ (1979) A role for brown adipose tissue in diet-induced

thermogenesis. Nature 281:31-35.

Rothwell NJ, Stock MJ, Warwick BP (1983) The effect of high fat and high

carbohydrate cafeteria diets on diet-induced thermogenesis in the rat. Int J 144

Obes 7:263-270.

Russo V, Bernabo N, Di Giacinto O, Martelli A, Mauro A, Berardinelli P, Curini V,

Nardinocchi D, Mattioli M, Barboni B (2013) H3K9 trimethylation precedes

DNA methylation during sheep oogenesis: HDAC1, SUV39H1, G9a, HP1, and

Dnmts are involved in these epigenetic events. J Histochem Cytochem 61:75-

89.

Sanchez-Gurmaches J, Tang Y, Jespersen NZ, Wallace M, Martinez Calejman C, Gujja

S, Li H, Edwards YJK, Wolfrum C, Metallo CM, Nielsen S, Scheele C, Guertin

DA (2018) Brown Fat AKT2 Is a Cold-Induced Kinase that Stimulates ChREBP-

Mediated De Novo Lipogenesis to Optimize Fuel Storage and Thermogenesis.

Cell Metab 27:195-209 e196.

Schreiber R, Diwoky C, Schoiswohl G, Feiler U, Wongsiriroj N, Abdellatif M, Kolb D,

Hoeks J, Kershaw EE, Sedej S, Schrauwen P, Haemmerle G, Zechner R (2017)

Cold-Induced Thermogenesis Depends on ATGL-Mediated Lipolysis in Cardiac

Muscle, but Not Brown Adipose Tissue. Cell Metab 26:753-763 e757.

Seale P, Bjork B, Yang W, Kajimura S, Chin S, Kuang S, Scime A, Devarakonda S,

Conroe HM, Erdjument-Bromage H, Tempst P, Rudnicki MA, Beier DR,

Spiegelman BM (2008) PRDM16 controls a brown fat/skeletal muscle switch.

Nature 454:961-967.

Seale P, Conroe HM, Estall J, Kajimura S, Frontini A, Ishibashi J, Cohen P, Cinti S,

Spiegelman BM (2011) Prdm16 determines the thermogenic program of

subcutaneous white adipose tissue in mice. J Clin Invest 121:96-105. 145

Seale P, Kajimura S, Yang W, Chin S, Rohas LM, Uldry M, Tavernier G, Langin D,

Spiegelman BM (2007) Transcriptional control of brown fat determination by

PRDM16. Cell Metab 6:38-54.

Segre CV, Chiocca S (2011) Regulating the regulators: the post-translational code of

class I HDAC1 and HDAC2. J Biomed Biotechnol 2011:690848.

Shabalina IG, Petrovic N, de Jong JM, Kalinovich AV, Cannon B, Nedergaard J (2013)

UCP1 in brite/beige adipose tissue mitochondria is functionally thermogenic.

Cell Rep 5:1196-1203.

Sharma BK, Patil M, Satyanarayana A (2014) Negative regulators of brown adipose

tissue (BAT)-mediated thermogenesis. J Cell Physiol 229:1901-1907.

Sharma NK, Varma V, Ma L, Hasstedt SJ, Das SK (2015) Obesity Associated

Modulation of miRNA and Co-Regulated Target Transcripts in Human Adipose

Tissue of Non-Diabetic Subjects. Microrna 4:194-204.

Shi H, Kokoeva MV, Inouye K, Tzameli I, Yin H, Flier JS (2006) TLR4 links innate

immunity and fatty acid-induced insulin resistance. The Journal of clinical

investigation 116:3015-3025.

Shi LL, Fan WJ, Zhang JY, Zhao XY, Tan S, Wen J, Cao J, Zhang XY, Chi QS, Wang

DH, Zhao ZJ (2017) The roles of metabolic thermogenesis in body fat regulation

in striped hamsters fed high-fat diet at different temperatures. Comp Biochem

Physiol A Mol Integr Physiol 212:35-44.

Shin H, Ma Y, Chanturiya T, Cao Q, Wang Y, Kadegowda AKG, Jackson R, Rumore D,

Xue B, Shi H, Gavrilova O, Yu L (2017) Lipolysis in Brown Adipocytes Is Not 146

Essential for Cold-Induced Thermogenesis in Mice. Cell Metab 26:764-777 e765.

Shinoda K, Luijten IH, Hasegawa Y, Hong H, Sonne SB, Kim M, Xue R, Chondronikola

M, Cypess AM, Tseng YH, Nedergaard J, Sidossis LS, Kajimura S (2015a)

Genetic and functional characterization of clonally derived adult human brown

adipocytes. Nat Med 21:389-394.

Shinoda K, Ohyama K, Hasegawa Y, Chang HY, Ogura M, Sato A, Hong H, Hosono T,

Sharp LZ, Scheel DW, Graham M, Ishihama Y, Kajimura S (2015b)

Phosphoproteomics Identifies CK2 as a Negative Regulator of Beige Adipocyte

Thermogenesis and Energy Expenditure. Cell Metab 22:997-1008.

Shore A, Karamitri A, Kemp P, Speakman JR, Lomax MA (2010) Role of Ucp1

enhancer methylation and chromatin remodelling in the control of Ucp1

expression in murine adipose tissue. Diabetologia 53:1164-1173.

Singh S, Rajput YS, Barui AK, Sharma R, Datta TK (2016) Fat accumulation in

differentiated brown adipocytes is linked with expression of Hox genes. Gene

Expr Patterns 20:99-105.

Storlien LH, James DE, Burleigh KM, Chisholm DJ, Kraegen EW (1986) Fat feeding

causes widespread in vivo insulin resistance, decreased energy expenditure,

and obesity in rats. Am J Physiol 251:E576-583.

Sunadome K, Yamamoto T, Ebisuya M, Kondoh K, Sehara-Fujisawa A, Nishida E (2011)

ERK5 regulates muscle cell fusion through Klf transcription factors. Dev Cell

20:192-205.

Szarc vel Szic K, Ndlovu MN, Haegeman G, Vanden Berghe W (2010) Nature or 147

nurture: let food be your epigenetic medicine in chronic inflammatory

disorders. Biochem Pharmacol 80:1816-1832.

Tahiliani M, Koh KP, Shen Y, Pastor WA, Bandukwala H, Brudno Y, Agarwal S, Iyer

LM, Liu DR, Aravind L, Rao A (2009) Conversion of 5-methylcytosine to 5-

hydroxymethylcytosine in mammalian DNA by MLL partner TET1. Science

324:930-935.

Tateishi K, Okada Y, Kallin EM, Zhang Y (2009) Role of Jhdm2a in regulating

metabolic gene expression and obesity resistance. Nature 458:757-761.

Tosic M, Allen A, Willmann D, Lepper C, Kim J, Duteil D, Schule R (2018) Lsd1

regulates skeletal muscle regeneration and directs the fate of satellite cells. Nat

Commun 9:366. van der Lans AA, Hoeks J, Brans B, Vijgen GH, Visser MG, Vosselman MJ, Hansen J,

Jorgensen JA, Wu J, Mottaghy FM, Schrauwen P, van Marken Lichtenbelt WD

(2013) Cold acclimation recruits human brown fat and increases nonshivering

thermogenesis. J Clin Invest 123:3395-3403. van Marken Lichtenbelt WD, Vanhommerig JW, Smulders NM, Drossaerts JM,

Kemerink GJ, Bouvy ND, Schrauwen P, Teule GJ (2009) Cold-activated brown

adipose tissue in healthy men. N Engl J Med 360:1500-1508.

Viana-Huete V, Guillen C, Garcia G, Fernandez S, Garcia-Aguilar A, Kahn CR, Benito

M (2018) Male Brown Fat-Specific Double Knockout of IGFIR/IR: Atrophy,

Mitochondrial Fission Failure, Impaired Thermogenesis, and Obesity.

Endocrinology 159:323-340. 148

Villarroya F, Peyrou M, Giralt M (2017) Transcriptional regulation of the uncoupling

protein-1 gene. Biochimie 134:86-92.

Vinter J, Hull D, Batt RA, Tyler DD (1988) The effect of limit feeding on thermogenesis

and thermoregulation in genetically obese (ob/ob) mice during cold exposure.

Int J Obes 12:111-117.

Virtanen KA, Lidell ME, Orava J, Heglind M, Westergren R, Niemi T, Taittonen M,

Laine J, Savisto NJ, Enerback S, Nuutila P (2009) Functional brown adipose

tissue in healthy adults. N Engl J Med 360:1518-1525.

Vitali A, Murano I, Zingaretti MC, Frontini A, Ricquier D, Cinti S (2012) The adipose

organ of obesity-prone C57BL/6J mice is composed of mixed white and brown

adipocytes. J Lipid Res 53:619-629.

Wang C, Liu W, Nie Y, Qaher M, Horton HE, Yue F, Asakura A, Kuang S (2017) Loss

of MyoD Promotes Fate Transdifferentiation of Myoblasts Into Brown

Adipocytes. EBioMedicine 16:212-223.

Wang L, Xu S, Lee JE, Baldridge A, Grullon S, Peng W, Ge K (2013a) Histone H3K9

methyltransferase G9a represses PPARgamma expression and adipogenesis.

EMBO J 32:45-59.

Wang L, Zhao Y, Bao X, Zhu X, Kwok YK, Sun K, Chen X, Huang Y, Jauch R, Esteban

MA, Sun H, Wang H (2015) LncRNA Dum interacts with Dnmts to regulate

Dppa2 expression during myogenic differentiation and muscle regeneration.

Cell Res 25:335-350.

Wang QA, Tao C, Gupta RK, Scherer PE (2013b) Tracking adipogenesis during white 149

adipose tissue development, expansion and regeneration. Nat Med 19:1338-

1344.

Wang X, Yang Z, Xue B, Shi H (2011) Activation of the cholinergic antiinflammatory

pathway ameliorates obesity-induced inflammation and insulin resistance.

Endocrinology 152:836-846.

Wei LN, Hu X, Chandra D, Seto E, Farooqui M (2000) Receptor-interacting protein

140 directly recruits histone deacetylases for gene silencing. J Biol Chem

275:40782-40787.

Welstead GG, Creyghton MP, Bilodeau S, Cheng AW, Markoulaki S, Young RA,

Jaenisch R (2012) X-linked H3K27me3 demethylase Utx is required for

embryonic development in a sex-specific manner. Proc Natl Acad Sci U S A

109:13004-13009.

Wessels B, van den Broek NM, Ciapaite J, Houten SM, Wanders RJ, Nicolay K,

Prompers JJ (2015) Carnitine supplementation in high-fat diet-fed rats does

not ameliorate lipid-induced skeletal muscle mitochondrial dysfunction in vivo.

Am J Physiol Endocrinol Metab 309:E670-678.

Wiper-Bergeron N, Wu D, Pope L, Schild-Poulter C, Hache RJ (2003) Stimulation of

preadipocyte differentiation by steroid through targeting of an HDAC1 complex.

EMBO J 22:2135-2145.

Wu H, Zhang Y (2014) Reversing DNA methylation: mechanisms, genomics, and

biological functions. Cell 156:45-68.

Wu J, Bostrom P, Sparks LM, Ye L, Choi JH, Giang AH, Khandekar M, Virtanen KA, 150

Nuutila P, Schaart G, Huang K, Tu H, van Marken Lichtenbelt WD, Hoeks J,

Enerback S, Schrauwen P, Spiegelman BM (2012) Beige adipocytes are a

distinct type of thermogenic fat cell in mouse and human. Cell 150:366-376.

Xia L, Wang C, Lu Y, Fan C, Ding X, Fu H, Qi K (2014) Time-specific changes in DNA

methyltransferases associated with the leptin promoter during the

development of obesity. Nutr Hosp 30:1248-1255.

Xue B, Coulter A, Rim JS, Koza RA, Kozak LP (2005) Transcriptional synergy and the

regulation of Ucp1 during brown adipocyte induction in white fat depots. Mol

Cell Biol 25:8311-8322.

Xue B, Kim YB, Lee A, Toschi E, Bonner-Weir S, Kahn CR, Neel BG, Kahn BB (2007)

Protein-tyrosine phosphatase 1B deficiency reduces insulin resistance and the

diabetic phenotype in mice with polygenic insulin resistance. J Biol Chem

282:23829-23840.

Yan C, Chen J, Li M, Xuan W, Su D, You H, Huang Y, Chen N, Liang X (2016) A

decrease in hepatic microRNA-9 expression impairs gluconeogenesis by

targeting FOXO1 in obese mice. Diabetologia 59:1524-1532.

Yang X, Wang X, Liu D, Yu L, Xue B, Shi H (2014) Epigenetic regulation of macrophage

polarization by DNA methyltransferase 3b. Mol Endocrinol 28:565-574.

Yang X, Wu R, Shan W, Yu L, Xue B, Shi H (2016) DNA Methylation Biphasically

Regulates 3T3-L1 Preadipocyte Differentiation. Mol Endocrinol 30:677-687.

Yang Z, Hulver M, McMillan RP, Cai L, Kershaw EE, Yu L, Xue B, Shi H (2012)

Regulation of insulin and leptin signaling by muscle suppressor of cytokine 151

signaling 3 (SOCS3). PLoS One 7:e47493.

Ye J (2013) Improving insulin sensitivity with HDAC inhibitor. Diabetes 62:685-687.

Ye L, Fan Z, Yu B, Chang J, Al Hezaimi K, Zhou X, Park NH, Wang CY (2012) Histone

demethylases KDM4B and KDM6B promotes osteogenic differentiation of

human MSCs. Cell Stem Cell 11:50-61.

Yin H, Pasut A, Soleimani VD, Bentzinger CF, Antoun G, Thorn S, Seale P, Fernando

P, van Ijcken W, Grosveld F, Dekemp RA, Boushel R, Harper ME, Rudnicki MA

(2013) MicroRNA-133 controls brown adipose determination in skeletal muscle

satellite cells by targeting Prdm16. Cell Metab 17:210-224.

Yoshioka K, Yoshida T, Kondo M (1992) Brown adipose tissue thermogenesis and

metabolic rate contribute to the variation in obesity among rats fed a high fat

diet. Jpn J Physiol 42:673-680.

You D, Nilsson E, Tenen DE, Lyubetskaya A, Lo JC, Jiang R, Deng J, Dawes BA, Vaag

A, Ling C, Rosen ED, Kang S (2017) Dnmt3a is an epigenetic mediator of

adipose insulin resistance. Elife 6.

Zha L, Cao Q, Cui X, Li F, Liang H, Xue B, Shi H (2016) Epigenetic regulation of E-

cadherin expression by the histone demethylase UTX in colon cancer cells. Med

Oncol 33:21.

Zha L, Li F, Wu R, Artinian L, Rehder V, Yu L, Liang H, Xue B, Shi H (2015) The

Histone Demethylase UTX Promotes Brown Adipocyte Thermogenic Program

Via Coordinated Regulation of H3K27 Demethylation and Acetylation. J Biol

Chem 290:25151-25163. 152

Zhang Y, Reinberg D (2001) Transcription regulation by histone methylation:

interplay between different covalent modifications of the core histone tails.

Genes Dev 15:2343-2360.

Zhou Z, Zhang HS, Liu Y, Zhang ZG, Du GY, Li H, Yu XY, Huang YH (2018) Loss of

TET1 facilitates DLD1 colon cancer cell migration via H3K27me3-mediated

down-regulation of E-cadherin. J Cell Physiol 233:1359-1369.

Zhuang L, Jang Y, Park YK, Lee JE, Jain S, Froimchuk E, Broun A, Liu C, Gavrilova O,

Ge K (2018) Depletion of Nsd2-mediated histone H3K36 methylation impairs

adipose tissue development and function. Nat Commun 9:1796.

Zou T, Chen D, Yang Q, Wang B, Zhu MJ, Nathanielsz PW, Du M (2017) Resveratrol

supplementation of high-fat diet-fed pregnant mice promotes brown and beige

adipocyte development and prevents obesity in male offspring. J Physiol

595:1547-1562.

APPENDICES: PUBLICATIONS

1. Li, F., Wu, R., Cui, X., Zha, L., Yu, L., Shi, H., Xue, B. Histone Deacetylase 1

(HDAC1) Negatively Regulates Thermogenic Program in Brown Adipocytes via

Coordinated Regulation of H3K27 Deacetylation and Methylation (2016) J

Biol Chem 2016, 291(9):4523-4536.

2. Zha, L., Li, F., Wu, R., Artinian, L., Rehder, V., Yu, L., Liang, H., Xue, B., Shi,

H. The Histone Demethylase UTX Promotes Brown Adipocyte Thermogenic

Program Via Coordinated Regulation of H3K27 Demethylation and Acetylation

(2015) Journal of Biological Chemistry 08/2015; 290(41). 153

3. Fu, L., Li, F., Bruckbauer, A., Cao, Q., Cui, X., Wu, R., Shi, H., Xue, B., Zemel,

M.B. Interaction between leucine and phosphodiesterase 5 inhibition in

modulating insulin sensitivity and lipid metabolism (2015) Diabetes,

Metabolic Syndrome and Obesity: Targets and Therapy 05/2015;

8:227-39.

4. Fu, L., Bruckbauer, A., Li, F., Cao, Q., Cui, X., Wu, R., Shi, H., Zemel, M.B.,

Xue, B., Leucine Amplifies the Effects of Metformin on Insulin Sensitivity and

Glycemic Control in Diet-Induced Obese Mice (2015) Metabolism: clinical

and experimental 03/2015; 64(7).

5. Fu, L., Bruckbauer, A., Li, F., Cao, Q., Cui, X., Wu, R., Shi, H., Zemel, M.B.,

Xue, B., Interaction between Metformin and Leucine in Reducing

Hyperlipidemia and Hepatic Lipid accumulation in Diet-Induced Obese Mice

(2015) Metabolism: clinical and experimental 07/2015; 64(11).

6. Zha, L., Cao, Q., Cui, X., Li, F., Liang, H., Xue, B., Shi, H. Epigenetic regulation

of E-cadherin expression by the histone demethylase UTX in colon cancer cells

(2016) Med Oncol 2016, 33(3):21

7. Bruckbauer, A., Banerjee, J., Fu, L., Li, F., Cao, Q., Cui, X., Wu, R ., Shi, H.,

Xue, B. A Combination of Leucine, Metformin, and Sildenafil Treats

Nonalcoholic Fatty Liver Disease and Steatohepatitis in Mice (2016) Int J

Hepatol 2016:9185987. 154

8. Cui X, Nguyen NL, Zarebidaki E, Cao Q, Li F, Zha L, Bartness T, Shi, H., Xue,

B.: Thermoneutrality decreases thermogenic program and promotes adiposity

in high-fat diet-fed mice. Physiol Rep 2016, 4(10).

9. Zha L, Cao Q, Cui X, Li F, Liang H, Xue B, Shi H: Epigenetic regulation of E-

cadherin expression by the histone demethylase UTX in colon cancer cells.

Med Oncol 2016, 33(3):21.

10. Bruckbauer A, Banerjee J, Cao Q, Cui X, Jing J, Zha L, Li F, Xue B, Shi

H, Zemel MB.: Leucine-nicotinic acid synergy stimulates AMPK/Sirt1 signaling

and regulates lipid metabolism and lifespan in Caenorhabditis elegans, and

hyperlipidemia and atherosclerosis in mice. Am J Cardiovasc Dis. 2017,

7(2):33-47.