MIAMI UNIVERSITY The Graduate School

Certificate for Approving the Dissertation

We hereby approve the Dissertation

Of

Minqian Shen

Candidate for the Degree Doctor of Philosophy

______

Dr. Haifei Shi, Advisor

______

Dr. Kathleen Killian, Committee Chair

______

Dr. Jack Vaughn, Committee

______

Dr. James Janik, Committee

______

Dr. Xiaowen Cheng, Graduate School Representative

ABSTRACT

ROLES OF ESTROGEN HORMONES AND ESTROGEN RECEPTORS ON REGULATION OF LIVER AND LIVER CANCER

by

Minqian Shen

Liver is one of the most essential organs involved in the regulation of energy homeostasis. Hepatic steatosis, a major manifestation of metabolic syndrome, is associated with imbalance between lipid formation and breakdown and glucose production and catabolism. Although the most common risk factors of hepatocellular carcinoma (HCC) are hepatic virus infection and alcohol, a rapid increase in obesity has become a prime cause of HCC, outweighing HCC with virus- or alcohol- related etiology. Estrogens affect both cell metabolism and proliferation. Both nuclear estrogen receptors (ERs), ER-α and ER-β, are expressed in the liver. How estrogen hormones regulate liver homeostasis and which receptors play the fundamental role in the metabolic process are not clear. Moreover, whether estrogen hormones and ERs are protective or detrimental in HCC is under debate, and whether estrogens can alter HCC metabolism and interfere with leptin-induced HCC is not known. The goal of my dissertation is to further elucidate the roles of estrogens and ERs in normal liver metabolism as well as HCC metabolism and proliferation using both in vivo and in vitro models. Thus, I hypothesize that 1) exogenous estrogen replacement reverses liver metabolic alteration in estrogen-depleted mice mainly through ER-α receptor; 2) estrogens inhibit leptin-induced HepG2 cell proliferation mainly through ER-β activation and alter HepG2 metabolism mainly through ER-α activation. In this dissertation, previous studies on the functions of estrogen and estrogen receptors in liver and liver cancer are discussed in Chapter 1. Studies using ovariectomized mouse model with hormone-replacement was applied to determine the role of estrogens and estrogen receptors in mouse liver metabolism are discussed in Chapter 2. Human cancer cell line HepG2 was treated with different concentrations of estrogen and estrogen receptor agonists as well as estrogen receptor siRNA to determine the role of estrogen and estrogen receptors on HepG2 cancer cell proliferation, apoptosis and leptin signal pathway in Chapter 3. High performance liquid chromatography (HPLC) was used to determine the role of estrogen and estrogen receptors on HepG2 cancer cell expression and metabolic profiles in Chapter 4. In Chapter Five, I conclude the finding of my current studies and give perspectives for future research directions of estrogen hormones and estrogen receptors in the regulation of liver and liver cancer metabolism.

A DISSERTATION

Presented to the Faculty of

Miami University in partial

fulfillment of the requirements

for the degree of

Doctor of Philosophy

Department of Biology

by

Minqian Shen

The Graduate School Miami University Oxford, Ohio

2017

Dissertation Director: Dr. Haifei Shi

TABLE OF CONTENTS TABLE OF CONTENTS…………………………………………………………………………iii LIST OF TABLES……………………………………………………………………………..….v LIST OF FIGURES……………………………………………………………..………………..vi LIST OF ABBREVIATIONS………………………………………………………………..….viii ACKNOWLEDGEMENTS……………………………………………………………………....x

Chapter 1. Introduction of estrogens’ role in liver metabolism and hepatocellular carcinoma 1.1 Obesity and metabolic syndrome ……………………………………………………..…..…1 1.2 Physiologic function of estrogens in liver…………………………………………...... ……..2 1.3 Glucose and lipid metabolism in liver……………………………………………...……...…4 1.4 Link between obesity and hepatocellular carcinoma…………………………………………5 1.5 Estrogen’s effects on hepatocellular carcinoma………………………………………………6 1.6 Summary…………………………………………………………………………………...…7 References…………………………………………………………………………...……………9 Figure legends……………………………………………………………………………………14

Chapter 2. Estrogens and estrogen receptors regulate energy metabolism in liver in mouse model Abstract…………...…………………………………………………………...…………………16 2.1 Introduction…………………..………………………………………………………………17 2.2 Methods and materials……………………………………………………………………….19 2.3 Results…………………………………………………………………………………….….23 2.4 Discussion.…….………………………………………………………………………….….29 References………………………………………………………………………………………..32 Figure legends……………………………………………………………………………..……..35

Chapter 3. Estrogen and estrogen receptors on HepG2 cancer cell proliferation, apoptosis and leptin signal pathway.

iii

Abstract…………………………………………………………………………………………..53 3.1 Introduction………………………………………………………………………………….54 3.2 Methods and materials……………………………………………………………………….55 3.3 Results………………………………………………………………………………………..58 3.4 Discussion……………………………………………………………………………....……60 References…………………………………………………………………………..……………65 Figure legends……………………………………………………………………………………71

Chapter 4. Estrogen and estrogen receptors on HepG2 cancer cell metabolic profile Abstract…………………………………………………………………………………………..84 4.1 Introduction……………………………………………………………………………….….85 4.2 Methods and materials……………………………………………………………………….86 4.3 Results………………………………………………………………………………………..90 4.4 Discussion……………………………………………………………………………………93 References…………………..……………………………………………………………………95 Figure legends……………………………………………………………………………………97

Chapter 5. Conclusion and perspectives………………………………………………………..119 References………………………………………………………………………………………122

iv

LIST OF TABLES

Table 2.1 Primers for qPCR……….…………………………………………………………….38 Table 4.1 Top biological functions identified by Ingenuity Pathway Analysis regulated by E2, PPT and DPN…………………………………………………………………………………….99 Table 4.2 qPCR primers for RNA sequencing check…………………………………………...107 Table 4.3 qPCR primers for metabolic enzymes……………………………………………….108

v

LIST OF FIGURES

Figure 1.1 Genomic effects of estrogens and androgens via nuclear ERs and ARs……………..12 Figure 1.2 Metabolic effects of estrogens and androgens on regulation of lipid, glucose, and cholesterol in the liver…………..………………………………………………………………..12 Figure 2.1 Body weight, food intake and circulating E2………………………...………………39 Figure 2.2 Body composition, peri-gonadal fat weight and circulating leptin…………………..40 Figure 2.3 Plasma lipid profiles………………………………………………………………….41 Figure 2.4 Liver TG, liver weight, liver/whole body weight ratio and ADP/ATP ratio……….…42 Figure 2.5 Insulin sensitivity test and glucose tolerance test…………………….………………43 Figure 2.6 Insulin sensitivity signal pathway……………………………………………………44 Figure 2.7 levels involved in energy metabolism………………………….…..45 Figure 2.8 Metabolic profiles………………………………………………………………….…47 Figure 3.1 Effects of leptin, 17β-estradiol, and estrogen receptor agonists on HepG2 cell number in cell culture……………………………………………………………………..…………..…..74 Figure 3.2 Effects of leptin, 17β-estradiol, and estrogen receptor agonists on proliferation of HepG2 cells………………………………………………………………………………………75 Figure 3.3 Effects of leptin, 17β-estradiol, and estrogen receptor agonists on apoptosis of HepG2 cells…………………………………….………………………………………………………...76 Figure 3.4 Effects of leptin on estrogen receptor expression in HepG2 cells……………………77 Figure 3.5 Effects of leptin, 17β-estradiol, and estrogen receptor agonists on STAT3 and SOCS3 signaling in HepG2 cells…………………………………………………………………………78 Figure 3.6 Effects of leptin, 17β-estradiol, and estrogen receptor agonists on ERK and p38/MAPK signaling in HepG2 cells……………………………………………………………80 Figure 3.7 Effects of ER siRNAs on cell proliferation and activation of leptin signaling………82 Figure 4.1 Effects of sodium oxamate (OX), oligomycin (OM), 2-deoxy-D-glucose (2-DG), and 17β-estradiol (E2) on HepG2 cell number in cell culture…………………………..…………..109 Figure 4.2 Effects of different doses of sodium oxamate (OX), oligomycin (OM), 2-deoxy-D- vi glucose (2-DG) on HepG2 cytotoxicity, viability and apoptosis……………………………….110 Figure 4.3 Effects of E2, PPT and DPN on gene expression levels involved in energy metabolism……………………………………………………………………………………...111 Figure 4.4 RNA sequencing data shows different gene expression upon E2, PPT and DPN treatments on HepG2 cells……………………………………………………………………...113 Figure 4.5 Evaluation of target identified by RNA sequencing…………………………114 Figure 4.6 Metabolic profiles…………………………………………………………………..115

vii

LIST OF ABBREVIATIONS

E2 17β-Estradiol DPN 2,3-bis(4-hydroxy-phenyl)-propionitrile 2-DG 2-deoxy-D-glucose PPT 4,4’,4’’-(4-propyl-[1H]-pyrazole-1,3,5-triyl) trisphenol 6Pfk 6-phosphofructokinase Acc acetyl co-enzyme carboxylase ArKO aromatase-knockout BMI body mass index ChREBP carbohydrate regulated element-binding Cox6a cytochrome c oxidase ER estrogen receptor ER-α estrogen receptor α ER-β estrogen receptor β ERE estrogen response element FA fatty acid FASN fatty acid synthase FFA free fatty acid GTT glucose tolerance tests GLUT glucose transporter Glut2 glucose transporter 2 Gapdh glyceraldehyde-3-phosphate dehydrogenase Pygl glycogen phosphrylase Gys2 glycogen synthase 2 GPER G-protein coupled estrogen receptor HCC hepatocellular carcinoma HDL-C high-density lipoprotein cholesterol HFD high-fat diet HPLC–MS high-performance liquid chromatography-mass spectrometry IR insulin receptor ITT insulin tolerance test IL-6 Interleukin-6 ip intraperitoneal LPL lipoprotein lipase

viii

LDL-C low-density lipoprotein cholesterol MS metabolic syndrome NAFLD non-alcoholic fatty liver disease OM oligomycin OVX ovariectomy OXPHOS oxidative phosphorylation Pparα peroxisome proliferator-activated receptor alpha Pparγ peroxisome proliferator-activated receptor gamma Pgc-1α peroxisome proliferator-activated receptor gamma 1 a Pepck phosphoenolpyruvate carboxykinase PCA principal component analysis Pk pyruvate kinase BrdU quantification of cell proliferation assay by bromodeoxyuridine STAT3 signal transducers and activators of transcription 3 OX sodium oxamate Srebp-1c sterol-regulatory binding protein-1c SOCS3 suppressor of cytokine signaling 3 TG triglycerides

ix

ACKNOWLEDGEMENTS

The journey of my PhD program would be incomplete without being able to thank the innumerable people who have been an intrinsic part of it. First and foremost, I would like to express my deepest appreciation to my advisor, Dr. Haifei Shi. She always encouraged me to pursue my research interest, and gave me the right direction and right advice I needed. Without her guidance, support and persistent help, this dissertation would not have been possible. I am also truly thankful to my dissertation committee members: Dr. Xiaowen Cheng, Dr. James Janik, Dr. Kathleen Killian, and Dr. Jack Vaughn for their countless advice and edits on my dissertation proposal and helped develop my dissertation research. I would like to express my gratitude to the present and past members of Shi lab, Shiva Senthil Kumar, Zheng Zhu, Xian Liu, Kristen Krolick, and my undergraduates: Anjali, Kenzie, Sean. They helped me a lot in my research. I would like to thank my parents for their constant motivation and encouragement in every phase of my life. My mom and dad have always given me the freedom to follow my heart and gave my dream the wings to fly. I also take this opportunity to express my sincere gratitude towards my dearest husband Jieming, whose presence has added an entirely new meaning to my life. He always stood beside me through the difficult time and shared the burden of these experiments with a cheerful spirit. I am so lucky to have him in my life.

x

Chapter 1. Introduction of estrogens’ role in liver metabolism and hepatocellular carcinoma

1.1 Obesity and metabolic syndrome

1.1.1 Definition of obesity and metabolic syndrome

Body mass index (BMI), which is based on the ratio of height and weight (BMI = kg/m2), is the most commonly used index for “fatness” and estimation for important health outcomes such as heart diseases, diabetes, various types of cancer, and overall mortality. Overweight is defined as a BMI between 25.0 and 29.9; while obesity is defined as a BMI of 30 or higher in adults (https://www.cdc.gov/obesity/adult/index.html). Metabolic syndrome (MS) is commonly considered as a group of conditions, including increased blood pressure, high blood glucose, excess body fat around the waist, insulin resistance, and abnormal cholesterol or triglyceride levels, all of which increase risks of heart disease, stroke and diabetes.

1.1.2 Epidemiology of obesity and metabolic syndrome

Obesity is common, serious and costly in the U.S. Over 60% of American adults are either obese or overweight (1). Diabetes mellitus, hypertension, heart disease, and some forms of cancer are closely associated with obesity (2). More females than males are overweight at any age (3). The risk of overweight/obesity increases with age, and the peak happens at age of 45–54 in men and at age 55–64 in women (4). In 2002, Ford et al reported that the prevalence of metabolic syndrome is 24% for American adults between age of 20 and 74, with no gender differences, but with some modest ethnic differences (5). However, the prevalence of metabolic syndrome depends on the criteria that are used.

1.1.3 Risks of developing obesity and metabolic syndrome

Obesity results from combined causes and contributing factors, such as genetics, unhealthy diets, inactive lifestyles, age, preexisting medical problems, and certain medications, etc. (6). Although obesity is a risk factor for insulin resistance and other metabolic syndrome manifestations, not every obese patient has developed such health problems (7). Actually, people with excess of the intra-abdominal or visceral are at higher risk of developing insulin resistance and metabolic syndrome than people with excess of subcutaneous adipose tissue (8). 1

1.1.4 Side effects of obesity and metabolic syndrome

The net effect of excess body weight and related metabolic disorders is a decrease in life expectancy with an increase in mortality (9). Insulin resistance results from obesity and is a hallmark of metabolic syndrome. Type 2 diabetes mellitus is strongly and positively associated with overweight, as the degree and intensity of overweight and central obesity determine the risk of type 2 diabetes. Actually, weight loss reduces the risk of developing diabetes in those overweight patients (10-12). The lipid overload accumulates in non-fat tissue such as in the liver, heart, muscle, and pancreatic islets. Such excess ectopic lipid triggers non-oxidative pathways and results in cellular dysfunction and cell death (13), leading to many other consequences of obesity including hypertension, heart disease, non-alcoholic fatty liver disease, gallbladder disease and certain cancers (3).

1.1.5 Insulin resistance

Insulin resistance is a pivotal factor in the development of type 2 diabetes and MS, much earlier than hyperglycemia occurrence (14). Insulin resistance is due to the reduced response of peripheral tissues and organs to insulin stimulation. Some theories about the cellular mechanism of insulin resistance include that excessive fatty acids decrease insulin-stimulated glucose transport activity, reduce insulin-stimulated insulin receptor substrate-1 (IRS-1)-associated phosphatidylinositol 3-kinase activity, and cause acquired defects in mitochondrial oxidation functions (15, 16).

1.2 Physiologic function of estrogens in the liver

1.2.1 Production of estrogens

In both males and females, estrogens are derived from the aromatization of testosterones. There are three forms of estrogens: estrone (E1), estradiol (E2), and estriol (E3). In premenopausal women, E2 is mainly synthesized from cholesterol in the ovaries, with E2 concentration being approximately 5 times higher than that in men, while in postmenopausal women E2 is primarily converted from testosterone by aromatase in peripheral tissues, such as adipose tissue, adrenal glands, bones, vascular endothelium, and smooth muscle (17) with E2

2 concentration being similar in men. The level of E2 oscillates during ovarian cycles, varying from 30 pg/ml to 370 pg/ml (http://www.hemingways.org/GIDinfo/hrt_ref.htm).

1.2.2 Estrogen receptors and their signaling

Estrogens act on estrogen receptors (ERs), mainly through classic nuclear receptors ER- 훼� and ER-훽�. All these nuclear ERs are expressed in the livers of male and female humans and rodents, but at a lower level compared with reproductive organs such as the uterus, prostate, testis, ovary, and breast (18-20). ER-훽� is less abundant in almost all tissues and organs than ER- 훼� (21). Upon estrogens binding, classic estrogen nuclear receptors ER-훼� and ER-훽� form homo- or heterodimers and bind to estrogen response element (ERE) in target gene promoters or to other transcription factors, such as activator protein-1 (AP-1) and stimulating protein-1 (22), to induce expression of target genes (Fig 1.1).

1.2.3 Effects of menopause on estrogens

Menopause typically happens in women of age between 45 to 50 years of old that involves a sequence of hormonal changes over several years. The estrogens levels begin to fluctuate unpredictably and eventually decline. In the premenopausal state, E2 is the predominant and the most potent form of estrogens; however, in the postmenopausal state, E1, a much weaker form of estrogens is produced by adipose tissue and liver becomes predominant (23).

1.2.4 Physiological significance of estrogens

Estrogens benefit the liver in many aspects. E2 inhibits fibrogenesis by reducing expression of type I procollagen and increasing deposition of type I and type III collagens in the rat hepatic fibrosis models (24). Estrogens also mediate protection of mitochondrial structure and function via enhanced mitochondrial respiratory chain activity and increased ATP production (25). Women have higher expression of genes associated with immune response and a greater number of innate immune cells due to estrogens altering cell signal pathways and increasing the production of cytokines (26-28); however, these hormonally-mediated advantages may decline in menopause (29, 30). Rat models have demonstrated that administration of E2 to aged rats can restore the activity of anti-oxidative enzymes such as superoxide dismutase and glutathione S- transferases in aged liver (31).

3

1.3 Estrogens regulate energy metabolism in the liver

1.3.1 Estrogens regulate glucose metabolism in the liver

The family of glucose transporters (GLUT) mediates glucose transport into all different tissues, including the liver. GLUT2, the major GLUT in the liver, is responsible for bidirectional transportation of glucose across the liver cell membrane, taking up glucose from the circulation and releasing glucose derived from or glycogenolysis into the circulation. Among all different GLUTs, GLUT 2 is unique with low affinity for glucose (Km~17 mmol/l). Although GLUT2 expression level is regulated by glucose and insulin (32), there is no direct proof of the influence of estrogens on GLUT2 expression. Estrogens have been shown to lower blood glucose by reducing gluconeogenesis and increasing glycogen synthesis and storage in the liver (33, 34). More recent research using rodents with OVX also supports that estrogens lower blood glucose concentration (35, 36), which is further explained by another study showing that the high glucose levels observed in OVX rodents is due to enhanced glucagon signaling caused by an increased amount of glucagon receptor, accompanied with increased gluconeogenic enzymes (37) (Fig 1.2).

1.3.2 Estrogens regulate lipid metabolism in the liver

The liver is a key visceral organ in the regulation of lipid metabolism due to its high capacity for lipid transport, de novo lipogenesis, lipid oxidation, and lipolysis. De novo lipogenesis is creation of fatty acid (FA) from acetyl-CoA and NADPH through the action of enzymes of fatty acid synthases (38); while Neutral hydrolysis of triglycerides (TGs) to FAs and glycerol requires three consecutive steps that involve at least three different enzymes that adipose triglyceride lipase (ATGL) catalyzes the initial step of lipolysis, converting TGs to diacylglycerols (DGs). Hormone-sensitive lipase (HSL) is mainly responsible for the hydrolysis of DGs to monoacylglycerols (MGs) and MG lipase (MGL) hydrolysis. Then the FAs goes through β-oxidation to generate acetyl-CoA, which enters the citric acid cycle to produce ATP and NADH and FADH2 (39). Excessive accumulation of triglyceride (TG) within the hepatocytes results in liver steatosis, also known as non-alcoholic fatty liver disease. The imbalance of high-density lipoprotein cholesterol (HDL-C) and low-density lipoprotein cholesterol (LDL-C) in the circulation is responsible for fat deposition in liver. Sex and age differences in liver fat accumulation are supported by epidemiological data, which has shown

4 higher plasma level of LDL-C and lower plasma level of HDL-C in men and postmenopausal women compared with premenopausal women (40). In estrogen-deficient rodent models, spontaneous obesity and liver steatosis are seen due to impaired fatty acid 훽�-oxidation and elevated fatty acid synthase (FAS) (41, 42). Further evidence also has shown that administration of E2 results in a decrease in lipoproteinlipase activity and increases in expression of hormone- sensitive lipase and adipose TG lipase in the liver (43, 44) (Fig 1.2).

1.4 Link between obesity and hepatocellular carcinoma

1.4.1 The role of insulin in hepatocellular carcinoma

A high level of insulin initially overstimulates insulin receptor (IR) and insulin growth factor 1 (IGF-1) receptor (IGF1R) (45). Both IR and IGF1R are constituted with two α domains for ligand binding and two β domains containing tyrosine kinase activity. The IR has two subtypes, IR-A and IR-B. IR-A is responsible for intracellular signaling, while IR-B is related to glucose uptake directed by insulin (46). IGF-1 is released by the liver; while IGF1R is expressed nearly in all cells and is homologous to IR. Binding to IR-A/IGF1R results in autophosphorylation of tyrosine kinase and subsequently interacts with insulin receptor substrates (IRS) to activate PI3K/AKT, which in turn inhibits Bcl-2 and thus reduces apoptosis (47). Also, the mitogen-activated protein kinase (MAPK) pathway is enhanced by ligand- receptor binding, and promotes sequential activation of RAS/ERK, resulting in cell proliferation and gene expression (48). Of note, insulin resistance has been shown to be significantly associated with HCC development in chronic hepatitis C patients (49).

1.4.2 The role of leptin in hepatocellular carcinoma

Leptin is produced mostly by and its circulating level is positively correlated with body fat mass. Elevated plasma leptin level, usually seen in obesity, attenuates leptin signaling which is known as leptin resistance (50). Besides being known for its role in neuroendocrine function and energy homeostasis in obesity, leptin also plays an important role in regulating immunity and inflammation (51, 52). Additionally, leptin has shown to promote angiogenesis by increasing VEGF secretion via HIF-1α activation (53). In many studies, leptin has been shown to stimulate cancer growth and proliferation through many different signaling

5 pathways, including AKT, STAT3, and mTOR (54-57). HCC patients have increased leptin level in blood as well as in tumor tissue samples (58, 59). In human HCC samples, leptin expression level is positively correlated with HCC proliferation with increased Ki-67, tumor size and local recurrence (60). Depletion of the leptin receptor through shRNA knockdown partially decreased orthotopic tumor growth in obese mice (61).

1.4.3 Other inflammatory hormones in hepatocellular carcinoma

Adiponectin, another hormone secreted by adipose tissue, is decreased in obesity, and has anti-inflammatory and anti-carcinogenic effects (55, 56, 62, 63). Adiponectin has been shown to prevent insulin resistance, reduce hyperinsulinemia, and improve metabolic syndrome (64). Adiponectin induces apoptosis via caspase-3 activation, inhibits cell proliferation via inhibition of mTOR, JNK phosphorylation, and prevents cancer cell growth and survival via NF-κB inhibition (65, 66). In human HCC samples, microarray study shows that adiponectin expression has an inverse correlation with tumor size (66).

Circulating IL-6 increases in subjects with insulin resistance and excessive BMI. It stimulates angiogenesis, promotes cell growth and inhibits apoptosis (67). The overexpression of IL-6 stimulates STAT3 transcription and results in increased gene expression involved in cell proliferation, apoptosis reduction, angiogenesis, metastasis and invasion (68). Another powerful pro-inflammatory adipokine from adipocytes increased in obesity is TNF-α (69, 70), which activates NF-κB, an essential factor for cell survival.

In summary, obesity causes chronic low-grade inflammation in the whole body. Adiponectin is downregulated; while, leptin, IL-6 and TNF-α are upregulated, which reduces anti-inflammatory and tumor-suppressing effects and hence increases pro-inflammatory and tumor-promoting capability.

1.5Effects on hepatocellular carcinoma by estrogens

1.5.1 Gender disparity in hepatocellular carcinoma

The incidence of HCC increases with age, and it occurs in men three to five times more than in women according to recent epidemiological data (71). Also, women with HCC have a higher overall survival rate than men between the ages of 18 years and 64 years; however, this

6 gender disparity in survival rate disappears after age 65 (71). Such gender disparity seen in HCC is considered to be mediated by estrogens. Two types of ERs have been found in HCC, ER-α and ER-β (18). Multiple functions of ERs in liver diseases are achieved by many complicated signal transduction processes, and detailed roles of each subtype of ERs have not been well elucidated.

1.5.2 Protective role of estrogens in hepatocellular carcinoma

In 2007, Naugler et al. reported that estrogens suppress IL-6 production from liver Kupffer cells and reduce diethylnitrosamine-induced hepatocarcinogenesis (72). Moreover, E2 protects IL-6 treated mice from developing liver cancer, suggesting that E2 can inhibit IL-6 downstream signaling (73). These studies provide an explanation for the higher incidence of HCC in males than in females due to anti-IL-6 effects of estrogens. With HCC disease progression, Kupffer cells are replaced by tumor-associated (TAM), and estrogens inhibit TAM activation and JAK1/STAT6 signaling pathway (74). STAT3 signal is the major signal pathway applied by both IL-6 and leptin, and is the central hub in cancer inflammation. Liver tumor size was significantly reduced in STAT-3 deficient mice due to DEN treatment (75). In addition, enhanced JAK/STAT3 pathway due to a suppressor of cytokine signaling-3 (SOCS3) knock-out increases sensitivity of hepatitis-induced HCC (76). Recently, Hou et al.reported that ER-α activation increases the expression of protein tyrosine phosphatase receptor type O (PTPRO), which in turn dephosphorylates STAT3 and then attenuates STAT3 signaling (77). ER- β has strong anti-proliferative and anti-inflammatory features in multiple organs and tissues (78- 80). However, the function of ER-β in HCC is largely unclear.

1.6 Summary

The liver is one of the most essential organs involved in the regulation of energy homeostasis. Hepatic steatosis, a major manifestation of metabolic syndrome, is associated with an imbalance between lipid formation and breakdown as well as glucose production and catabolism. Although the most common risk factors of hepatocellular carcinoma (HCC) are hepatic virus infection and alcohol, a rapid increase in obesity has become a prime cause of HCC, outweighing HCC with virus- or alcohol- related etiology. Estrogens play important roles in both normal and cancer physiology. Both nuclear receptors, ER-α and ER-β, are expressed in the liver. Estrogens regulate liver homeostasis and different ERs and their signal transduction 7 play different roles in metabolic processes. Furthermore, estrogens and ERs are important for HCC proliferation and metabolism. Estrogens contribute to sex disparity in both obesity and HCC development. The roles of estrogens and ERs in the regulation of hepatic metabolism under obesity and liver cancer conditions are the focus of this dissertation.

8

References 1. Flegal KM, Graubard BI, Williamson DF, Gail MH. Excess deaths associated with underweight, overweight, and obesity. Jama. 2005;293(15):1861-7. Epub 2005/04/21. 2. Bray GA. Medical consequences of obesity. The Journal of clinical endocrinology and metabolism. 2004;89(6):2583-9. Epub 2004/06/08. 3. Bray GA, Bellanger T. Epidemiology, trends, and morbidities of obesity and the metabolic syndrome. Endocrine. 2006;29(1):109-17. Epub 2006/04/20. 4. Flegal KM, Graubard BI, Williamson DF, Gail MH. Weight and mortality. Hypertension. 2006;47(2):e6; author reply e-7. Epub 2005/12/29. 5. Ford ES, Giles WH, Dietz WH. Prevalence of the metabolic syndrome among US adults: findings from the third National Health and Nutrition Examination Survey. Jama. 2002;287(3):356-9. Epub 2002/01/16. 6. Eckel RH, Grundy SM, Zimmet PZ. The metabolic syndrome. Lancet. 2005;365(9468):1415-28. Epub 2005/04/20. 7. Roberts CK, Hevener AL, Barnard RJ. Metabolic syndrome and insulin resistance: underlying causes and modification by training. Comprehensive Physiology. 2013;3(1):1-58. Epub 2013/05/31. 8. Hogue JC, Lamarche B, Gaudet D, Tremblay AJ, Despres JP, Gagne C, et al. Genotype of the mutant LDL receptor allele is associated with LDL particle size heterogeneity in familial hypercholesterolemia. Atherosclerosis. 2006;184(1):163-70. Epub 2005/05/19. 9. Tefferi A. A contemporary approach to the diagnosis and management of polycythemia vera. Current hematology reports. 2003;2(3):237-41. Epub 2003/08/07. 10. Colditz GA, Willett WC, Rotnitzky A, Manson JE. Weight gain as a risk factor for clinical diabetes mellitus in women. Annals of internal medicine. 1995;122(7):481-6. Epub 1995/04/01. 11. Chan JM, Rimm EB, Colditz GA, Stampfer MJ, Willett WC. Obesity, fat distribution, and weight gain as risk factors for clinical diabetes in men. Diabetes care. 1994;17(9):961-9. Epub 1994/09/01. 12. Sjostrom CD, Lissner L, Sjostrom L. Relationships between changes in body composition and changes in cardiovascular risk factors: the SOS Intervention Study. Swedish Obese Subjects. Obesity research. 1997;5(6):519-30. Epub 1998/02/04. 13. Kusminski CM, Shetty S, Orci L, Unger RH, Scherer PE. Diabetes and apoptosis: lipotoxicity. Apoptosis : an international journal on programmed cell death. 2009;14(12):1484- 95. Epub 2009/05/08. 14. Perseghin G, Petersen K, Shulman GI. Cellular mechanism of insulin resistance: potential links with inflammation. International journal of obesity and related metabolic disorders : journal of the International Association for the Study of Obesity. 2003;27 Suppl 3:S6-11. Epub 2004/01/06. 15. Shulman GI. Unraveling the cellular mechanism of insulin resistance in humans: new insights from magnetic resonance spectroscopy. Physiology (Bethesda). 2004;19:183-90. Epub 2004/08/12. 16. Shulman GI. Cellular mechanisms of insulin resistance. The Journal of clinical investigation. 2000;106(2):171-6. Epub 2000/07/21. 17. Simpson ER. Sources of estrogen and their importance. The Journal of steroid biochemistry and molecular biology. 2003;86(3-5):225-30. Epub 2003/11/19. 18. Iavarone M, Lampertico P, Seletti C, Francesca Donato M, Ronchi G, del Ninno E, et al.

9

The clinical and pathogenetic significance of estrogen receptor-beta expression in chronic liver diseases and liver carcinoma. Cancer. 2003;98(3):529-34. Epub 2003/07/25. 19. Miceli V, Cocciadiferro L, Fregapane M, Zarcone M, Montalto G, Polito LM, et al. Expression of wild-type and variant estrogen receptor alpha in liver carcinogenesis and tumor progression. Omics : a journal of integrative biology. 2011;15(5):313-7. Epub 2011/02/26. 20. Cui J, Shen Y, Li R. Estrogen synthesis and signaling pathways during aging: from periphery to brain. Trends in molecular medicine. 2013;19(3):197-209. Epub 2013/01/26. 21. Kuiper GG, Carlsson B, Grandien K, Enmark E, Haggblad J, Nilsson S, et al. Comparison of the ligand binding specificity and transcript tissue distribution of estrogen receptors alpha and beta. Endocrinology. 1997;138(3):863-70. Epub 1997/03/01. 22. Ciana P, Raviscioni M, Mussi P, Vegeto E, Que I, Parker MG, et al. In vivo imaging of transcriptionally active estrogen receptors. medicine. 2003;9(1):82-6. Epub 2002/12/17. 23. Doshi SB, Agarwal A. The role of oxidative stress in menopause. Journal of mid-life health. 2013;4(3):140-6. Epub 2014/03/29. 24. Yasuda M, Shimizu I, Shiba M, Ito S. Suppressive effects of estradiol on dimethylnitrosamine-induced fibrosis of the liver in rats. Hepatology. 1999;29(3):719-27. Epub 1999/03/03. 25. Chen JQ, Eshete M, Alworth WL, Yager JD. Binding of MCF-7 cell mitochondrial and recombinant human estrogen receptors alpha and beta to human mitochondrial DNA estrogen response elements. Journal of cellular biochemistry. 2004;93(2):358-73. Epub 2004/09/16. 26. Klein SL, Jedlicka A, Pekosz A. The Xs and Y of immune responses to viral vaccines. The Lancet Infectious diseases. 2010;10(5):338-49. Epub 2010/04/27. 27. Villacres MC, Longmate J, Auge C, Diamond DJ. Predominant type 1 CMV-specific memory T-helper response in humans: evidence for gender differences in cytokine secretion. Human immunology. 2004;65(5):476-85. Epub 2004/06/03. 28. Hepworth MR, Hardman MJ, Grencis RK. The role of sex hormones in the development of Th2 immunity in a gender-biased model of Trichuris muris infection. European journal of immunology. 2010;40(2):406-16. Epub 2009/12/02. 29. Gameiro CM, Romao F, Castelo-Branco C. Menopause and aging: changes in the immune system--a review. Maturitas. 2010;67(4):316-20. Epub 2010/09/04. 30. Gameiro C, Romao F. Changes in the immune system during menopause and aging. Front Biosci (Elite Ed). 2010;2:1299-303. Epub 2010/06/03. 31. Kumar P, Kale RK, Baquer NZ. Estradiol modulates membrane-linked ATPases, antioxidant enzymes, membrane fluidity, lipid peroxidation, and lipofuscin in aged rat liver. Journal of aging research. 2011;2011:580245. Epub 2011/10/19. 32. Stanley JC. The regulation of glucose production. The role of liver glycogen and gluconeogenesis in the liver and kidney cortex. British journal of anaesthesia. 1981;53(2):137- 46. Epub 1981/02/01. 33. Bryzgalova G, Gao H, Ahren B, Zierath JR, Galuska D, Steiler TL, et al. Evidence that oestrogen receptor-alpha plays an important role in the regulation of glucose homeostasis in mice: insulin sensitivity in the liver. Diabetologia. 2006;49(3):588-97. Epub 2006/02/08. 34. Ahmed-Sorour H, Bailey CJ. Role of ovarian hormones in the long-term control of glucose homeostasis, glycogen formation and gluconeogenesis. Annals of nutrition & metabolism. 1981;25(4):208-12. Epub 1981/01/01. 35. Saengsirisuwan V, Pongseeda S, Prasannarong M, Vichaiwong K, Toskulkao C.

10

Modulation of insulin resistance in ovariectomized rats by endurance exercise training and estrogen replacement. Metabolism: clinical and experimental. 2009;58(1):38-47. Epub 2008/12/09. 36. Feigh M, Andreassen KV, Hjuler ST, Nielsen RH, Christiansen C, Henriksen K, et al. Oral salmon calcitonin protects against impaired fasting glycemia, glucose intolerance, and obesity induced by high-fat diet and ovariectomy in rats. Menopause. 2013;20(7):785-94. Epub 2013/06/26. 37. Nigro M, Santos AT, Barthem CS, Louzada RA, Fortunato RS, Ketzer LA, et al. A change in liver metabolism but not in thermogenesis is an early event in ovariectomy-induced obesity in rats. Endocrinology. 2014;155(8):2881-91. Epub 2014/06/11. 38. Postic C, Girard J. Contribution of de novo fatty acid synthesis to hepatic steatosis and insulin resistance: lessons from genetically engineered mice. The Journal of clinical investigation. 2008;118(3):829-38. Epub 2008/03/05. 39. Zechner R, Zimmermann R, Eichmann TO, Kohlwein SD, Haemmerle G, Lass A, et al. FAT SIGNALS--lipases and lipolysis in lipid metabolism and signaling. Cell metabolism. 2012;15(3):279-91. Epub 2012/03/13. 40. Trapani L, Segatto M, Pallottini V. Regulation and deregulation of cholesterol homeostasis: The liver as a metabolic "power station". World journal of hepatology. 2012;4(6):184-90. Epub 2012/07/05. 41. Fisher CR, Graves KH, Parlow AF, Simpson ER. Characterization of mice deficient in aromatase (ArKO) because of targeted disruption of the cyp19 gene. Proceedings of the National Academy of Sciences of the United States of America. 1998;95(12):6965-70. Epub 1998/06/17. 42. D'Eon TM, Souza SC, Aronovitz M, Obin MS, Fried SK, Greenberg AS. Estrogen regulation of adiposity and fuel partitioning. Evidence of genomic and non-genomic regulation of lipogenic and oxidative pathways. J Biol Chem. 2005;280(43):35983-91. 43. Shen M, Kumar SP, Shi H. Estradiol regulates insulin signaling and inflammation in adipose tissue. Hormone molecular biology and clinical investigation. 2014;17(2):99-107. Epub 2014/11/06. 44. Foryst-Ludwig A, Kintscher U. Metabolic impact of estrogen signalling through ERalpha and ERbeta. The Journal of steroid biochemistry and molecular biology. 2010;122(1-3):74-81. Epub 2010/07/06. 45. Gallagher EJ, LeRoith D. Minireview: IGF, Insulin, and Cancer. Endocrinology. 2011;152(7):2546-51. Epub 2011/05/05. 46. Arcidiacono B, Iiritano S, Nocera A, Possidente K, Nevolo MT, Ventura V, et al. Insulin resistance and cancer risk: an overview of the pathogenetic mechanisms. Experimental diabetes research. 2012;2012:789174. Epub 2012/06/16. 47. Cohen DH, LeRoith D. Obesity, type 2 diabetes, and cancer: the insulin and IGF connection. Endocrine-related cancer. 2012;19(5):F27-45. Epub 2012/05/18. 48. Braun S, Bitton-Worms K, LeRoith D. The link between the metabolic syndrome and cancer. International journal of biological sciences. 2011;7(7):1003-15. Epub 2011/09/14. 49. Dai CY, Huang JF, Hsieh MY, Chuang WL, Yu ML. Insulin resistance, viral load and response to peginterferon and ribavirin in patients with chronic hepatitis C virus infection. Gut. 2010;59(3):418. Epub 2010/03/09. 50. Cui H, Lopez M, Rahmouni K. The cellular and molecular bases of leptin and ghrelin resistance in obesity. Nature reviews Endocrinology. 2017. Epub 2017/02/25. 51. Procaccini C, Galgani M, De Rosa V, Carbone F, La Rocca C, Ranucci G, et al. Leptin:

11 the prototypic adipocytokine and its role in NAFLD. Current pharmaceutical design. 2010;16(17):1902-12. Epub 2010/04/08. 52. Matarese G, Procaccini C, De Rosa V, Horvath TL, La Cava A. Regulatory T cells in obesity: the leptin connection. Trends in molecular medicine. 2010;16(6):247-56. Epub 2010/05/25. 53. Endo H, Hosono K, Uchiyama T, Sakai E, Sugiyama M, Takahashi H, et al. Leptin acts as a growth factor for colorectal tumours at stages subsequent to tumour initiation in murine colon carcinogenesis. Gut. 2011;60(10):1363-71. Epub 2011/03/17. 54. Mantzoros CS, Bolhke K, Moschos S, Cramer DW. Leptin in relation to carcinoma in situ of the breast: a study of pre-menopausal cases and controls. International journal of cancer. 1999;80(4):523-6. Epub 1999/02/06. 55. Dalamaga M, Diakopoulos KN, Mantzoros CS. The role of adiponectin in cancer: a review of current evidence. Endocrine reviews. 2012;33(4):547-94. Epub 2012/05/02. 56. Spyridopoulos TN, Petridou ET, Skalkidou A, Dessypris N, Chrousos GP, Mantzoros CS. Low adiponectin levels are associated with renal cell carcinoma: a case-control study. International journal of cancer. 2007;120(7):1573-8. Epub 2007/01/06. 57. Moon HS, Dalamaga M, Kim SY, Polyzos SA, Hamnvik OP, Magkos F, et al. Leptin's role in lipodystrophic and nonlipodystrophic insulin-resistant and diabetic individuals. Endocrine reviews. 2013;34(3):377-412. Epub 2013/03/12. 58. Sadik NA, Ahmed A, Ahmed S. The significance of serum levels of adiponectin, leptin, and hyaluronic acid in hepatocellular carcinoma of cirrhotic and noncirrhotic patients. Human & experimental toxicology. 2012;31(4):311-21. Epub 2012/01/18. 59. Chen MJ, Yeh YT, Lee KT, Tsai CJ, Lee HH, Wang SN. The promoting effect of adiponectin in hepatocellular carcinoma. Journal of surgical oncology. 2012;106(2):181-7. Epub 2012/01/31. 60. Sharma D, Wang J, Fu PP, Sharma S, Nagalingam A, Mells J, et al. Adiponectin antagonizes the oncogenic actions of leptin in hepatocellular carcinogenesis. Hepatology. 2010;52(5):1713-22. Epub 2010/10/14. 61. Mendonsa AM, Chalfant MC, Gorden LD, VanSaun MN. Modulation of the leptin receptor mediates tumor growth and migration of pancreatic cancer cells. Plos One. 2015;10(4):e0126686. Epub 2015/04/29. 62. Kaklamani VG, Wisinski KB, Sadim M, Gulden C, Do A, Offit K, et al. Variants of the adiponectin (ADIPOQ) and adiponectin receptor 1 (ADIPOR1) genes and colorectal cancer risk. Jama. 2008;300(13):1523-31. Epub 2008/10/02. 63. Barb D, Pazaitou-Panayiotou K, Mantzoros CS. Adiponectin: a link between obesity and cancer. Expert opinion on investigational drugs. 2006;15(8):917-31. Epub 2006/07/25. 64. Nkontchou G, Bastard JP, Ziol M, Aout M, Cosson E, Ganne-Carrie N, et al. Insulin resistance, serum leptin, and adiponectin levels and outcomes of viral hepatitis C cirrhosis. Journal of hepatology. 2010;53(5):827-33. Epub 2010/08/24. 65. Kamada Y, Matsumoto H, Tamura S, Fukushima J, Kiso S, Fukui K, et al. Hypoadiponectinemia accelerates hepatic tumor formation in a nonalcoholic steatohepatitis mouse model. Journal of hepatology. 2007;47(4):556-64. Epub 2007/04/27. 66. Saxena NK, Fu PP, Nagalingam A, Wang J, Handy J, Cohen C, et al. Adiponectin modulates C-jun N-terminal kinase and mammalian target of rapamycin and inhibits hepatocellular carcinoma. Gastroenterology. 2010;139(5):1762-73, 73 e1-5. Epub 2010/07/20. 67. Aballay LR, Eynard AR, Diaz Mdel P, Navarro A, Munoz SE. Overweight and obesity: a

12 review of their relationship to metabolic syndrome, cardiovascular disease, and cancer in South America. Nutrition reviews. 2013;71(3):168-79. Epub 2013/03/05. 68. Fan Y, Mao R, Yang J. NF-kappaB and STAT3 signaling pathways collaboratively link inflammation to cancer. Protein & cell. 2013;4(3):176-85. Epub 2013/03/14. 69. Weichhaus M, Broom I, Bermano G. The molecular contribution of TNF-alpha in the link between obesity and breast cancer. Oncology reports. 2011;25(2):477-83. Epub 2010/12/18. 70. Tzanavari T, Giannogonas P, Karalis KP. TNF-alpha and obesity. Current directions in autoimmunity. 2010;11:145-56. Epub 2010/02/23. 71. Yang D, Hanna DL, Usher J, LoCoco J, Chaudhari P, Lenz HJ, et al. Impact of sex on the survival of patients with hepatocellular carcinoma: a Surveillance, Epidemiology, and End Results analysis. Cancer. 2014;120(23):3707-16. Epub 2014/08/02. 72. Naugler WE, Sakurai T, Kim S, Maeda S, Kim K, Elsharkawy AM, et al. Gender disparity in liver cancer due to sex differences in MyD88-dependent IL-6 production. Science. 2007;317(5834):121-4. Epub 2007/07/07. 73. Shi L, Feng Y, Lin H, Ma R, Cai X. Role of estrogen in hepatocellular carcinoma: is inflammation the key? Journal of translational medicine. 2014;12:93. Epub 2014/04/09. 74. Yang W, Lu Y, Xu Y, Xu L, Zheng W, Wu Y, et al. Estrogen represses hepatocellular carcinoma (HCC) growth via inhibiting alternative activation of tumor-associated macrophages (TAMs). The Journal of biological chemistry. 2012;287(48):40140-9. Epub 2012/08/22. 75. He G, Yu GY, Temkin V, Ogata H, Kuntzen C, Sakurai T, et al. Hepatocyte IKKbeta/NF- kappaB inhibits tumor promotion and progression by preventing oxidative stress-driven STAT3 activation. Cancer cell. 2010;17(3):286-97. Epub 2010/03/17. 76. Ogata H, Kobayashi T, Chinen T, Takaki H, Sanada T, Minoda Y, et al. Deletion of the SOCS3 gene in liver parenchymal cells promotes hepatitis-induced hepatocarcinogenesis. Gastroenterology. 2006;131(1):179-93. Epub 2006/07/13. 77. Hou J, Xu J, Jiang R, Wang Y, Chen C, Deng L, et al. Estrogen-sensitive PTPRO expression represses hepatocellular carcinoma progression by control of STAT3. Hepatology. 2013;57(2):678-88. Epub 2012/07/24. 78. Paruthiyil S, Parmar H, Kerekatte V, Cunha GR, Firestone GL, Leitman DC. Estrogen receptor beta inhibits human breast cancer cell proliferation and tumor formation by causing a G2 cell cycle arrest. Cancer research. 2004;64(1):423-8. Epub 2004/01/20. 79. Strom A, Hartman J, Foster JS, Kietz S, Wimalasena J, Gustafsson JA. Estrogen receptor beta inhibits 17beta-estradiol-stimulated proliferation of the breast cancer cell line T47D. Proceedings of the National Academy of Sciences of the United States of America. 2004;101(6):1566-71. Epub 2004/01/28. 80. Cvoro A, Tatomer D, Tee MK, Zogovic T, Harris HA, Leitman DC. Selective estrogen receptor-beta agonists repress transcription of proinflammatory genes. J Immunol. 2008;180(1):630-6. Epub 2007/12/22.

13

Figure legends

Figure 1.1 Genomic effects of estrogens and androgens via nuclear ERs and ARs

Figure 1.2 Metabolic effects of estrogens and androgens on regulation of lipid, glucose, and cholesterol in the liver

14

Figure 1.1

Estrogens

Androgens

Protein

ER-β mRNA ER-α ER ER AR ERE AR AR ARE ER TF AR TF

Hepatocyte Nucleus

Figure 1.2

Liver Estrogens Androgens Lipolysis Insulin receptor Insulin clearance Glycogen synthesis Glycogen storage Cholesterol uptake Cholesterol removal Cholesterol synthesis

Lipogenesis Lipogenesis Lipid uptake Glucose uptake Gluconeogenesis Cholesterol removal Cholesterol synthesis

15

Chapter 2. Estrogen and estrogen receptors regulate energy metabolism in the liver of mice

Minqian Shen, Qi Zhu, Fanyi Zhong, Jiangjiang Zhu, Haifei Shi

Manuscript was written and prepared by Minqian Shen. Haifei Shi served in an advisory capacity. All authors participated in this study and will serve in an editing capacity. Text and figure formatting variations will be due to journal guidelines. This work will be submitted for publication.

Abstract

The liver is a key visceral organ for energy balance by regulating lipid and glucose metabolism. Obesity is associated with hepatic lipid accumulation, non-alcoholic fatty liver disease (NAFLD), and liver insulin insensitivity. Sex steroid hormone estrogen may be protective against high-fat diet (HFD)-induced obesity and insulin resistance. Postmenopausal women with declined estradiol (E2) levels increase risk of metabolic diseases. This has been tested by many previous studies utilizing rodent models of ovariectomy (OVX) and/or treatment of estradiol (E2), the major biologically active form of estrogen. However, the underlying mechanisms, such as detailed metabolic processes affected by estrogen and the role of each estrogen receptors (ERs) in metabolic regulation, are not clear. We tested the hypothesis that E2 and its estrogen receptors (ERs) play protective roles in hepatic glucose and lipid metabolism, using estrogen depleted OVX mice combined with estrogen replacement in HFD-induced obese mice. Five groups of female C57BL/6 mice fed with a HFD were used, including sham-operated mice treated with a placebo (sham+placebo), OVX mice treated with placebo (OVX+placebo), OVX mice treated with E2 (OVX+E2), OVX mice treated with ERα agonist PPT (OVX+PPT), and OVX mice treated with ERβ agonist DPN (OVX+DPN). Body weight and adiposity, food intake, plasma lipid profile, insulin sensitivity, insulin signaling, expression of genes related to glucose and lipid metabolism, liver metabolites and metabolic pathways were compared among these groups for a period of eight weeks. OVX+E2 mice and OVX+PPT mice, but not OVX+DPN mice, had lower body mass and adiposity, food intake, better lipid profile, elevated hepatic insulin signaling compared with OVX+placebo mice. OVX+placebo mice had elevated

16 mRNA levels of genes involved in glucose storage and lipid synthesis, while E2- and PPT- treatment concurrently reduced the expression of genes stimulating fat storage and increased the expression of genes enhancing glucose usage and fat oxidation in OVX mice, compared with sham+placebo mice. Hepatic metabolite analysis using HPLC-MS confirmed lowered glycolysis, fatty acid oxidation, and ADP/ATP ratio, whereas enhanced fatty acid synthesis in OVX+placebo and OVX+DPN mice comparing with sham+placebo, OVX+E2, and OVX+PPT mice. However, no overall metabolic pathway was identified among different treatments. These data from physiological, biochemical, molecular and metabolomic analyses collectively indicated beneficial roles of estrogens in hepatic metabolic regulation of glucose and lipid metabolism mainly through ER-α activation.

Key words:

Estradiol; Estrogen receptors; Insulin sensitivity; Ovariectomy; Metabolism

2.1 Introduction

Insulin released by pancreatic β cells in response to high blood glucose stimulates glucose uptake and glycogen formation at multiple metabolic tissues, including , adipose tissues, and liver. During short-term fasting, glycogen is broken down to glucose and glucose is released into blood to maintain stable a blood glucose level. The liver can contain up to 10% of its volume as glycogen (1). Under extreme conditions such as long-term fasting, glucose is made via gluconeogenesis from ketone in the liver and released into the blood stream. Thus, the whole-body glucose homeostasis depends on whole body glucose uptake and glucose production via glycogenolysis and gluconeogenesis. Glycogen synthase (GYS) is a key, rate- limiting enzyme in glycogen synthesis. There are two isoforms of GYS in humans. GYS-1 is the isoform presented in the skeletal muscle and heart, and GYS-2 is the liver isoform (2). Mutations on the Gys2 gene on 12p12.2 causes Gys2 deficiency and low liver glycogen content, a disease also known as “glycogen storage disease type 0” (3). Glycogen phosphorylase (GYPL) is the key and rate-limiting enzyme that generates glucose 1-phosphate from glycogen hydrolysis. Congenital deficiency of hepatic GYPL, also known as glycogen storage disease type VI or Hers disease, results in lower capability to mobilize glycogen to release glucose in response to fasting and glucagon (4). The liver also can synthesize glucose de novo from lactate,

17 alanine and glycerol via gluconeogenesis. Pyruvate carboxylase, phosphoenolpyruvate carboxykinase (PEPCK), and fructose 1,6-biphosphatase are enzymes responsible for irreversible steps in the gluconeogenesis pathway. Congenital deficiency in any of these enzymes leads to hypoglycemia and lactic acidosis during fasting.

Estrogen regulates glucose homeostasis mainly in the liver and skeletal muscle. ER-α is found to be a positive regulator of GLUT4 expression, while ER-β has a negative effect (5). In addition, estrogen has been found to promote aerobic glycolysis by activation of glycolytic enzymes, such as hexokinase, phosphofructokinase, and pyruvate kinase (PK). Estrogen also increases expression of pyruvate dehydrogenase (PDH), a key enzyme that links glycolysis to the TCA cycle (6). Further, estrogen also has been shown to promote the expression of complex I, complex IV, and complex V/ATP synthase in mitochondria (7).

Glucose regulates cellular metabolism by allosteric and transcriptional effects. Glucose is transformed into glucose-6-phosphate by hexokinase at the first and rate-limiting step of glycolysis. Glucose-6-phosphate is an allosteric activator of glycogen synthase. In the meanwhile, glucose itself inhibits glycogenolysis by activating glycogen phosphorylase kinase, which inactivates glycogen phosphorylase and inhibits glycogen breakdown. Thus, increased blood glucose level stimulates glycogen deposition in the liver by allosteric stimulating glycogen synthesis and inhibiting glycogen breakdown (8). A fraction of the absorbed glucose is metabolized into xylulose 5-phosphate through the pentose shunt pathway, which in turn dephosphorylates carbohydrate- responsive element-binding protein (ChREBP) (9, 10). This dephosphorylated ChREBP translocates into the nucleus and activates the transcription of the gene encoding pyruvate kinase-1, acetyl-CoA carboxylase, and fatty acid synthase. These allosteric and transcriptional effects of glucose are important to control glucose uptake, storage, and conversion into fat in the liver.

Women have a higher percentage of body fat, with more subcutaneous adiposity but less visceral adiposity than men (11). Although body-mass index (BMI) is a strong predictor for body fat storage, body fat distribution is a stronger predictor of metabolic health (12-14). Visceral adiposity is associated with peripheral (primarily skeletal muscle) and hepatic insulin resistance, independent of gender (15). In addition, menopausal women are more vulnerable to develop obesity and insulin resistance, which is affected by decrease in estrogen, aging process, and life

18 style change (16). Consistent with human studies, studies using estrogen-deficient rodent models, such as aromatase-knockout (ArKO) and ovarietcomized (OVX) mice, have showed that hepatic steatosis, liver mitochondrial β-oxidation, and hepatic insulin sensitivity in female mice are improved by estrogen replacement (17, 18).

Most lipid supply is mediated by hydrolysis of circulating triglyceride (TG) to free fatty acids (FFAs) by lipoprotein lipase (LPL). FFAs are stored in the white adipose tissues and liver. Estrogen has been found to prevent TG accumulation by suppression of the target gene of LPL (19). Also, estrogen has been demonstrated to increase lipolysis via enhancing the expression of hormone-sensitive lipase, increase β-oxidation via increasing phosphorylation of AMP-activated protein kinase (AMPK), and upregulate expression of peroxisome proliferator-activated receptor- γ coactivator 1γ (Pgc-1γ) and uncoupling protein 2 (20). Estrogen significantly decreases plasma LDL, “the bad cholesterol”, by enhancing LDL receptor activity, instead of modifying the polymorphisms of apoE, apoB, or cholesterol 7a-hydroxylase (21). Meanwhile, estrogen increases plasma HDL, “the good cholesterol” (22). Interestingly, estrogen also significantly enhances the activity of the hepatic cholesterol biosynthesis enzyme of 3-hydroxy-3- methylglutaryl coenzyme A reductase, which indicates that estrogen increases secretion of cholesterol in bile from the liver (23).

In order to determine the exact effects of different ERs on liver metabolism, I hypothesize that exogenous estrogen replacement reverses liver metabolic alteration in estrogen depleted mice. Additionally, estrogen achieves metabolic effects in the liver mainly through the ER-α receptor, since ER-α has greater effects on liver metabolism than the ER-β receptor.

2.2 Methods and materials

2.2.1 Animals

Fifty 12 weeks-old female C57BL/6 mice (Jackson Laboratory, Bar Harbor, ME) were single housed (12 h light-dark cycle, lights on at 0600) and fed a standard rodent diet (28.5% protein, 58% complex carbohydrate, 13.5% fat; LabDiet, St. Louis MO) during acclimation when their body mass and daily food intake were monitored. Ovarian cycles of these mice were determined by examining predominant cell types of vaginal cytology samples, and all

19 female mice were cycling normally during acclimation.

After acclimation, mice were grouped into five groups (n=10 per group) with matched average body mass and daily food intake. One group received sham surgery and the other four groups received OVX surgeries. After three days of surgery, a high-fat diet (HFD, 4.728 kcal/g; 45% fat; Research Diets, Inc., New Brunswick, NJ) was given to all five groups for eight weeks. Food intake and body mass were measured every week. All procedures were approved by the Institutional Animal Care and Use Committee at Miami University Ohio.

2.2.2 OVX and sham surgeries

OVX was performed to decrease endogenous estrogen and disrupt the ovarian cycles of female mice. Under isofluorane anesthesia (Butler Schein Animal Health, Dublin, OH) bilateral skin and muscle incisions on the dorsolateral flank were made paralleling the spinal column. In the OVX procedure, bilateral ovaries were removed without disturbing the uterus, oviduct, or gonadal parametrial adipose tissue. In the sham procedure, the ovaries were visualized but no tissue was removed. After all surgeries, the muscle was sutured with sterile absorbable vicryl sutures, and the skin was closed with sterile wound clips. Success OVX was confirmed by loss of the estrous phase and reduced plasma E2 levels. Estrogen levels of female mice fluctuate (24). To be consistent, all female mice were analyzed in estrus.

2.2.3 Chemicals and pellet implantation

After sham or OVX surgery, a single longitudinal skin incision was made between the mouse dorsal scapulae midline for subcutaneous pellet insertion. 60-day continuous-release hormone pellets containing 17β-estradiol (E2, 1.5 mg/pellet), ERα agonist PPT (2.5 mg/pellet), ERβ agonist DPN (10 mg/pellet) (25, 26), or placebo vehicle were administered to the sham group and one of OVX groups.

2.2.4 Intraperitoneal glucose and insulin tolerance tests

Intraperitoneal (ip) glucose tolerance tests (GTTs) and insulin tolerance tests (ITTs) were conducted pre-surgery (basal) and 8 weeks post-surgery. For both tests mice were fasted overnight (16-h) and moved to the procedure room (0800 ). After 2 h acclimation at the procedure room, baseline (0 min) blood glucose was assessed in samples obtained in duplicate from the tail vein from freely moving mice using glucometers and glucose strips (Infinity®, US

20

Diagnostics, New York, NY). Mice then received a 1.5 mg/kg bolus of 50% dextrose (Phoenix Pharmaceutical, St. Joseph, MO) via ip injection and blood samples were assessed for glucose concentration using glucometers after 15, 30, 45, 60 and 120 min. 15 min blood samples were collected for measuring glucose-stimulated insulin secretion. For the ITTs, after basal glucose levels were measured, mice were injected ip with 1 U/kg of insulin (Sigma, St. Louis, MO), and blood glucose levels were measured at 15, 30, 45 and 60 minutes in duplicate.

2.2.5 Body Composition

A mouse-specific NMR Echo MRI whole body composition analyzer (EchoMedical Systems, Houston, TX) was used to assess body fat and lean mass in conscious animals pre- surgery (basal) and 8 weeks post-surgery.

2.2.6 Sample collection

Sham+placebo mice were killed on the last day mice were food deprived for 2 hours (0700-0900), so that all mice had similar meal status. Blood glucose levels of mice were measured using blood samples obtained from tip of the tail vein with a glucometer (Infinity®, US Diagnostics, New York, NY). Mice were then injected intraperitoneally with saline or insulin (1 mU/g; Sigma, St. Louis, MO), and sacrificed 15 min after injection. Plasma of blood samples collected from the hepatic portal vein (HPV) were measured via sensitive ELISA for 17-β estradiol (Cayman Chemical, Ann Arbor, Michigan), leptin (CrystalChem, Downers Grove, IL), and lipids (Wako Diagnostics, Richmond, VA), including high and low density lipoprotein cholesterol (HDL-C and LDL-C respectively), FFA, and triglyceride (TG). Liver was collected, weighed, and left lateral lobe of the liver was frozen at −80°C until extractions for liver TG, protein, and total RNA.

2.2.7 Tissue insulin signaling activation using western blot

Protein was extracted by homogenizing the frozen liver tissues using lysis buffer with sodium orthovanadate, phenylmethylsulfonyl fluoride, protease inhibitor (Santa Cruz Biotechnology, Santa Cruz, CA) and phosphatase inhibitor cocktail (Sigma). Protein lysates were resolved in 4–15% tris-glycine gels and transferred to nitrocellulose membrane (Bio-Rad). Activity of kinase Akt indicates stimulated insulin signaling. Phosphorylated and total Akt (pAkt and tAkt, respectively; 1:1000; Cell Signaling, Danvers, MA), Phosphorylated insulin receptor

21 substrate 1 (IRS-1) and total IRS-1 (pIRS-1 and tIRS-1, respectively; 1:1000; Cell Signaling, Danvers, MA) were detected by western immunoblotting via chemiluminescence (Amersham+placebo™ ECL™ Prime, GE Healthcare) and visualized using autoradiography film. Density was quantified using ImageQuant software (Amersham+placebo Biosciences). pAkt measurements were normalized to tAkt (pAkt/tAkt). pIRS-1 measurements were normalized to tIRS-1 (pIRS-1 /tIRS-1). Activation of insulin signaling was indicated by pAkt/tAkt% and pIRS-1/tIRS-1% difference among groups.

2.2.8 Gene expression using quantitative PCR

Total RNA was extracted and reverse transcribed into cDNA using 1 μg RNA. Expression of genes related to lipogenesiss (fatty acid synthase, Fas; sterol regulatory element-binding 1, Srebf1), lipolysis (peroxisome proliferator-activated receptor-gamma coactivator, Pgc-1α; acetyl co-enzyme carboxylase, Acc), gluconeogenesis (phosphoenolpyruvate carboxykinase, Pepck), glycogen synthesis (glycogen synthase, Gys), glycogenolysis (glycogen phosphorylase, Pygl), glycolysis (6-phosphofructo-1-kinase, 6Pfk1; pyruvate kinase, Pk), and oxidative phosphorylation (cytochrome c oxidase, Cox) were measured (Table 2.1). Glyceraldehyde-3-phosphate dehydrogenase (Gapdh) mRNA levels were similar among treatment groups and Gapdh was used as a reference gene. Quantitative PCR was run in triplicates using iQ SYBR Green Supermix (Bio-Rad, Hercules, CA) and an iCycler (Bio-Rad) with 40 cycles of amplification (95 °C for 10 s) and annealing (55 °C for 30 s). The amplified products were confirmed via gel electrophoresis and melt curve analysis. Results were calculated by a 2−ΔΔCt method, and presented using sham+placebo group as 100%.

2.2.9 High-performance liquid chromatography-mass spectrometry (HPLC–MS) metabolic profiling

Liver sample (~20mg) from each mouse was homogenized with PBS and the metabolites were extracted by 250 μL cold methanol in −20 °C for 20 min. Then 50 μL isotope-labeled mixture was added as internal standards (Cambridge Isotope Laboratories, Inc.). The mixture was stored at −20 °C for 20 min and centrifuged for 20 min. Then 150 μL of the supernatant was collected and dried on a vacuum concentrator. The dried sample was reconstituted by 50% H2O and 50% acetonitrile and kept in 4 °C autosampler for MS runs. All raw MS data were manually inspected using the Quanbrowser module of Xcalibur version 2.0

22

(Thermo Fisher Scientific). The MS data were normalized by the mass of liver tissue used for metabolite extraction. JMP Pro12 (SAS Institute, Cary, NC) was used for statistical analysis. Principle components analysis (PCA) was applied for the metabolic profiles comparison between different groups. MetaboAnalyst 3.0 (http://www.metaboanalyst.ca/) was used to explore the metabolic pathway impact of E2.

2.2.10 Statistics

Data were presented as mean ± SEM. Prism 5 (La Jolla, CA) was used to perform two- way repeated-measures ANOVA comparing time and treatment followed by Bonferroni posttest to analyze weekly body mass, food intake, accumulative food intake, weekly body mass gain, ITT, and GTT. One-way ANOVA followed by Tukey posttest was used to compare gene expression levels, circulating E2 and leptin level,, plasma lipid profile including LDL, HDL, glycerol, FFA, plasma TG, and HDL/LDL, liver mass and TG, peri-gonadal fat mass, metabolites, gene expression levels, and insulin signaling from different groups. A test with p<0.05 was considered statistically significant.

2.3 Results

2.3.1 Circulating E2, body mass, and food intake

Plasma E2 levels were different among groups. Pairwise comparison indicated that circulating E2 concentration of OVX mice with placebo (OVX+placebo; 8.080±2.032 pg/ml; t=10.21), PPT (OVX+PPT; 8.463±2.631 pg/ml; t=10.45), or DPN (OVX+DPN; 7.244±2.78 pg/ml; t=10.57) were significantly reduced compared with sham+placebo group (118.8±17.20 pg/ml); except for OVX+E2 group that had increased E2 comparing to sham+placebo group (247.5±6.264 pg/ml, t=11.49). Additionally, among the OVX groups, E2 levels were significantly higher in E2-replaced than placebo- (t=22.08), PPT- (t=22.64), or DPN- (t=22.75) groups, indicating that the ERs agonists PPT and DPN did not affect plasma E2 level (Fig 2.1A).

Both time and treatments affected weekly body mass gain (time: F (7,350)=15.41, p<0.001; treatment: F (4,350)=4.04, p=0.0032), but no significant difference was revealed in any pairwise comparison of the weekly body mass gain among groups. OVX+placebo mice gained significantly more body mass than OVX+E2 mice during week 5 (t=2.9, p<0.05) and OVX+PPT

23 mice during week 1 (t=3.1, p<0.05), but no difference was observed between OVX+placebo and OVX+DPN groups. OVX+E2 gained less body mass than OVX+PPT during week 4 (t=3.1, p<0.05), and OVX+DPN during week 4 and week 5 (t=3.5, p<0.05; t=3.8, p<0.05). Body mass gain was not significantly different between OVX+PPT and OVX+DPN groups at any time (Fig 2.1B).

Both treatment (F=19.8, p<0.0001) and time (F=39.16, p<0.0001) affect the weekly body mass. Compared with sham+placebo mice, only OVX+E2 mice among all the other groups, had significantly less body mass at the end of week 8 (t= 3.236, p<0.05). OVX+E2 group, but not OVX+PPT or OVX+DPN group, had significantly less body mass from week 5 to week 8 (t= 3.658, p<0.01; t= 4.208, p<0.001; t= 3.963, p<0.001; t= 4.228, p<0.001), than OVX+placebo mice. There was no significant difference between OVX+E2 and OVX+PPT group mice, while OVX+E2 mice had less body mass than OVX+DPN mice from week 5 to week 8 (t=4.244, p<0.001; t= 4.230, p<0.001; t=4.706, p<0.001; t=4.990, p<0.001). Body mass between OVX+PPT and OVX+DPN mice were not significantly different (Fig 2.1C).

Weekly food intake and accumulative food intake among different groups were affected by both treatment (F=4.54, p=0.0013; F=3.38, p=0.0016) and time (F=13.02, p<0.0001, F=750.33, p<0.0001). Weekly food intake was not different among sham+placebo and OVX groups with different treatments, although OVX+E2 mice ate less than OVX+DPN (t= 3.254, p<0.01) and OVX+PPT mice ate less than OVX+DPN (t=3.137, p<0.05) mice during the first week (Fig 2.1D).

Accumulative food intake was similar among sham+placebo and all OVX groups with different treatments. OVX+E2 mice consumed less accumulative food than OVX+placebo mice since week 7 (t=2.833, p<0.05; t=2.995, p<0.05), and less than OVX+DPN mice since week 6 (t=2.779, p<0.05; t=2.808, p<0.05, t=3.101, p<0.05) (Fig 2.1E).

2.3.2 Body composition, peri-gonadal fat mass and circulating leptin

At the end of 8 weeks experimental treatments, OVX+E2 mice had the least fat/lean ratio compared with sham+placebo (t=3.837, p<0.05), OVX+placebo (t=5.246, p<0.05), OVX+PPT (t=3.880, p<0.05), and OVX+DPN (t=5.581, p<0.05) (Fig 2.2A). Peri-gonadal fat is another indicator of body fat distribution (27). E2 treatment, but not PPT or DPN treatment, had significantly less peri-gonadal fat mass than sham+placebo (t= 3.3, p<0.05) and OVX+placebo

24

(t= 4.5, p<0.05) groups (Fig 2.2B). Circulating leptin is not only a body adiposity indicator, but also an inflammatory cytokine. E2 treatment significantly reduced leptin level compared with all other groups (sham+placebo vs OVX+E2, t=5.3, p<0.05; OVX+placebo vs OVX+E2, t=6.2, p<0.05; OVX+E2 vs OVX+PPT, t=3.6, p<0.05; OVX+E2 vs OVX+DPN, t=5.6, p<0.05) (Fig 2.2C), which indicated that E2 treatment had the strongest effect on body weight control and overall systemic inflammation.

2.3.3 Plasma lipid profiles and liver TG

OVX+placebo mice had higher plasma TG than sham+placebo mice (t= 3.088, p<0.05), OVX+E2 mice had lower plasma TG than sham+placebo mice (t=3.190, p<0.05), while OVX+PPT and OVX+DPN mice had similar TG levels as sham+placebo mice. Of note, OVX+placebo had higher TG compared with OVX+E2 (t=6.195, p<0.05), OVX+PPT (t=5.523, p<0.05), and OVX+DPN (t= 4.026, p<0.05) groups, whose TG levels were comparable among these three treatment groups (Fig 2.3A). Only OVX+E2 treatment significantly reduced plasma FFA compared with OVX+placebo (t=4.1, p<0.05), and there were no differences among other groups (Fig 2.3B). LDL-C, whose level is highly correlated to the risk of cardiovascular diseases, was significantly higher in OVX+placebo group than sham+placebo group (t= 5.0, p<0.05). E2 and PPT treatments significantly rescued OVX’s effects on LDL-C, with OVX+E2 and OVX+PPT having significantly reduced LDL-C level compared with OVX+placebo mice (t= 4.8, p<0.05; t= 3.8, p<0.05). The reduction of LDL-C by DPN treatment failed to reach statistical significance (p>0.05) (Fig 2.3C). HDL-C, which on the other hand is considered as the good cholesterol as it transports LDL-C to the liver to be metabolized, was significantly reduced in OVX+placebo group comparing with sham+placebo group (t= 3.1, p<0.05). E2, PPT and DPN treatments similarly rescued reduced HDL-C in OVX models (t= 3.0, p<0.05; t= 4.9, p<0.05; t= 3.8, p<0.05) (Fig 2.3D). HDL/LDL was also lower in OVX+placebo mice than sham+placebo mice (t= 4.2, p<0.05). E2 and PPT, but not DPN, rescued HDL/LDL in OVX models (t= 4.5, p<0.05; t= 3.5, p<0.05) (Fig 2.3E). Although there was no significant difference of plasma glycerol among all groups, OVX+E2 mice trended to have the lowest glycerol level (Fig 2.3F).

There was no significant difference in liver TG content among any groups, but there was a trend that E2 and PPT treatments, instead of DPN treatment, reduced OVX-induced liver TG accumulation (Fig 2.4A). Interestingly, OVX+E2 mice had heavier liver mass than

25 sham+placebo (t= 3.0, p<0.05), OVX+PPT (t= 3.3, p<0.05), and OVX+DPN (t= 3.2, p<0.05) mice, but not than OVX+placebo mice (Fig 2.4B). Consistent with this, OVX+E2 mice had higher liver/whole body mass ratio than sham+placebo (t= 7.234, p<0.05), OVX+placebo (t= 7.234, p<0.05), and OVX+PPT (t= 6.606, p<0.05) mice (Fig 2.4C). These findings may be explained by some evidence showing that E2 can induce angiogenesis and increased blood supply in liver (28). Although there was no significant difference in ADP/ATP ratio among each group, there was a trend that E2- and PPT-treatment decreased the ADP/ATP ratio (Fig 2.4D), indicating that these mice consumed more ATP and were more energetic than other groups.

2.3.4 Insulin tolerance test (ITT) and glucose tolerance test (GTT)

Groups were matched according to average body mass and area under the curve (AUC) of basal ITT/GTT (Fig 2.5A, Fig 2.5C); thus there were no group differences at baseline. For insulin tolerance, OVX+placebo mice had higher blood glucose levels at 15 min (t=2.746, p<0.05) and 30 min (t= 2.680, p<0.05) than sham+placebo mice, and there was no difference between sham+placebo group and OVX groups that were treated with E2, PPT or DPN. Also, there was no difference among four OVX groups (Fig 2.5B). The glucose level of the OVX+placebo group was significantly higher than that of the sham+placebo mice at 60 min post-glucose injection (t=2.8, p<0.05). Also, the OVX+E2 mice had lower blood glucose level at 30 min (t=4.8, p<0.001), 45 min (t=3.2, p<0.01), and 60 min (t=4.8, p<0.001) post-glucose injection than OVX+placebo mice, while OVX+PPT group had significantly lower glucose level only at 30 min (t=3.1, p< 0.05) than OVX+placebo group. E2 significantly lowered blood glucose level at 60 min than that of PPT and DPN treatments (t=2.8, p< 0.05; t=3.0, p< 0.05). However, there was no difference in glucose level at any time point between PPT and DPN treatments (Fig 2.5D).

2.3.5 Insulin signaling pathway

Insulin sensitivity was indicated by pAkt/tAkt% and pIRS-1/tIRS-1%, as both Akt and IRS-1 are activated in insulin sensitive tissues (Fig 2.6A). OVX+placebo mice had less signal than sham+placebo mice for pAkt/tAkt% (t=12.53, p < 0.05) and pIRS-1/tIRS-1% (t=21.63, p < 0.05), indicating insulin insensitivity resulted from loss of endogenous estrogen hormones. OVX+E2 and OVX+PPT groups had comparable pAkt/tAkt% and pIRS-1/tIRS-1% as sham+placebo group, which were greater than OVX+placebo mice (t=14.46, p < 0.05; t=14.25, p

26

< 0.05), which indicated that insulin sensitivity was reduced by OVX, and E2 and PPT improved insulin sensitivity of OVX mice. DPN-treated OVX mice, however, had similarly reduced pAkt signaling as placebo-treated OVX mice (t=12.77, p < 0.05), which was much less than E2 treatment (t=14.65, p < 0.05) or PPT treatment (t=14.44, p < 0.05) (Fig 2.6B).

pIRS-1 signal was significantly reduced in OVX+placebo mice than sham+placebo group, and E2 and PPT treatments rescued pIRS-1 signal in OVX mice (t=23.28, p < 0.05; t=23.06, p < 0.05). DPN treatment had less pIRS-1 signal than sham+placebo mice (t=16.27, p < 0.05), OVX+E2 (t=17.87, p < 0.05) and OVX+PPT mice (t=17.66, p < 0.05) (Fig 2.6C), indicating E2 preserved insulin sensitivity mainly through ER-α.

2.3.6 Gene expression levels involved in energy metabolism

Glucose transporters mediate the downhill movement of glucose across cell plasma membranes. Glut2 gene expression was not significantly different among all five groups (Fig 2.7A), consistent with increased translocation, rather than synthesis, of glucose transporter following insulin signaling activation that contributes to increased glucose uptake into liver cells.

The expressions of two rate-limiting enzymes involved in glycolysis, 6- phosphofructokinase (6PFK) and pyruvate kinase (PK), were detected to test if glucose usage is changed. 6PFK is the enzyme for the first irreversible reaction unique to the glycolytic pathway, i.e., the phosphorylation of fructose 6-phosphate to fructose 1,6-bisphosphate, thus 6PFK is the prominent regulatory enzyme and primary control site in glycolysis. PK is the enzyme catalyzing the final irreversible step in glycolysis to yield ATP and pyruvate, a central metabolic intermediate that can be oxidized further or used as a building block.

OVX+placebo and OVX+DPN groups had significantly lower 6Pfk mRNA levels than the sham+placebo group (t=4.298, p < 0.05; t= 3.517, p < 0.05). E2 and PPT treatment significantly restored 6pfk expression in OVX mice compared with that of OVX+placebo mice (t=6.335, p < 0.05; t=6.663, p < 0.05). DPN treatment, however, didn’t achieve the similar effects as E2 or PPT (t=5.474, p < 0.05; t= 5.772, p < 0.05) (Fig 2.7B). No difference in Pk expression was seen among all five groups (Fig 2.7C). Additionally, Cox6a, an essential enzyme involved in OXPHOS, was not significantly changed by different treatments (Fig 2.7D). The expression of genes involved in glycogen synthesis was then detected. Similar as 6Pfk expression, expression of Pygl, an enzyme that breaks down glycogen in the liver, was

27 significantly lower in OVX+placebo and OVX+DPN groups than sham+placebo group (t= 5.278, p < 0.05; t= 3.562, p < 0.05). E2 and PPT treatment significantly increased Pygl expression in OVX mice compared with that of OVX+placebo mice (t= 7.374, p < 0.05; t= 7.713, p < 0.05), whereas DPN didn’t achieve the similar effects as E2 (t=5.485, p < 0.05) or PPT (t= 5.757, p < 0.05) (Fig 2.7E). Interestingly, no significant difference of Gys gene expression was revealed among different groups (Fig 2.7F). Expression of Pepck, an enzyme regulating gluconeogenesis, was not significantly different among the groups (Fig 2.7G).

Three important genes regulating fatty acid synthesis, Fasn, Acc, and Srebp-1c were then tested. OVX+placebo mice had significantly higher mRNA levels of Fasn than sham+placebo (t=4.614, p < 0.05), OVX+E2 (t=7.415, p < 0.05), and OVX+PPT (t=8.415, p < 0.05) groups, but not OVX+DPN group. Mice with DPN treatment had greater Fas expression than E2- (t=4.775, p < 0.05) or PPT- (t=5.683, p < 0.05) treated mice (Fig 2.7H). In contrast, no significant difference was detected among any groups in terms of Acc expression (Fig 2.7I). OVX+placebo mice had significantly higher expression of Srebp-1c gene than OVX+E2 (t= 3.851, p < 0.05) and OVX+PPT (t= 3.755, p < 0.05) group, but Srebp-1c expression was not significantly different comparing to sham+placebo or OVX+DPN groups. Mice with DPN treatment had higher Srebp-1c gene expression than E2 (t= 3.886, p < 0.05) or PPT treatments (t= 3.791, p < 0.05) (Fig 2.7J). Pgc-1α and Ppar-α are two genes regulating fatty acid oxidation. No significant difference was found among groups in Pgc1α expression level (Fig 2.7K). In contrast, E2-treated OVX mic had significantly lower expression of Ppar-α than OVX+placebo mice (t= 3.677, p < 0.05), and DPN treatment increased Ppar-α mRNA levels than E2 treatment in OVX mice (t=4.040, p < 0.05) (Fig 2.7L).

2.3.7 Metabolic profiles

The HPLC-MS system detected 13 metabolites directly associated with glucose and lipid metabolism, including glucose-1-phosphate, glucose-6-phosphate, coenzyme A, NAD, FAD, which were analyzed using principal component analysis (PCA), one of the most commonly used multivariate statistical analyses. PCA did not distinguish differences in the metabolic profile induced by these five different treatment groups (Fig 2.8A and Fig 2.8B). The loading plots showed the contributions of different metabolites to the separation. When all 255 metabolites detected by the HPLC were analyzed, no distinguished differences in metabolic profiles were

28 found (Fig 2.8C and Fig 2.8D).

Metabolic pathway impact analyses were conducted that put individual metabolites into the context of connected metabolic pathway networks. All detected metabolites were included in the metabolic pathway analysis, so the broader coverage of extensive metabolic networks can be achieved. Fig 2.8E - Fig 2.8N showed the major metabolic pathways impacted by different treatments, with the x-axis being the metabolic pathway impact value and the y-axis being the statistical significance (represented by p-value) of the impacted pathways. The dot size corresponds to the x-axis value and the dot color corresponds to the y-axis value.

2.4 Discussion

Estrogen plays critical roles in the regulation of glucose metabolism and maintenance of insulin sensitivity, and deficiency of estrogen may lead to the development of insulin resistance (29). Postmenopausal women with deficiency in endogenous estrogen have increased risk of developing T2DM (30), while hormone replacement therapy or treatment with E2 improves insulin sensitivity and lowers blood glucose levels (31, 32) and reduces incidence of diabetes (33). However, function of estrogen receptors in regulating energy homeostasis is unclear. OVX in rodents provides a good model of mimicking low postmenopausal estrogen level, thus the challenge remains in the rodent OVX model to distinguish the role different estrogen receptors play in the regulation of glucose metabolism and insulin sensitivity.

The current study was the first to compare different estrogen receptors in the regulation of hepatic glucose and lipid metabolic pathways. The weekly body mass gain and body mass were significantly lower in OVX+E2 and OVX+PPT groups than OVX+placebo mice (Fig 2.1B and Fig 2.1C), which indicated that estrogen controls body weight mainly through ER-α, rather than ER-β. DPN-replaced OVX mice, but not the OVX+placebo mice, consumed more food during the first week than E2- or PPT-treated OVX mice (Fig 2.1D), which indicated that ER-β activation increased food intake at least transiently. Estrogen improves plasma lipid profiles, increased adiposity, and fat distribution through ER-α, but not ER-β (Fig 2.2 and Fig 2.3), which is consistent with previous study results showing that ob/ob mice and OVX mice treated with 17β-estradiol or selective agonist binding ER-α had improved hepatic lipid content (34, 35). Loss of estrogen impaired insulin sensitivity and glucose tolerance. Administration of E2, PPT or

29

DPN did not repair ITT (Fig 2.5B); interestingly, E2 and PPT restored the GTT of OVX mice (Fig 2.5D). These results pointed out that E2 not only worked on peripheral organs, but also had direct effects on islets to induce insulin release. Some animal studies have evidence showing that estrogen directly improves β-cell function by binding to its receptor in rat islets and thereby stimulating insulin release (36). Insulin signaling of the liver was similar among sham+placebo, E2-treated, and PPT-treated OVX mice, whereas placebo-treated OVX and DPN-treated mice had defective insulin signaling (Fig 2.6). This finding is consistent with previous studies showing that estrogen participates in insulin sensitivity mainly through ER-α (30, 31). Furthermore, a number of enzymes and transcriptional factors regulate multistep metabolic sequences. OVX+placebo mice had higher mRNA levels of several genes involved in glucose storage and lipid synthesis, while E2-treated and PPT-treated mice had higher mRNA levels of several genes involved in glucose expenditure and lipid oxidation (Fig 2.7), consistent with increased energy usage indicated by lowered ADP/ATP ratio in E2- and PPT-treated OVX mice (Fig 2.4D). These results suggested that the effects of estrogen in insulin sensitivity improvement in the liver was attributable at least partially to increased glucose and lipid metabolic enzyme transcriptional levels.

In recent years, metabolomics has been used to study comprehensive characterization of the numerous metabolites found in different organisms (36). Although we did qPCR to detect genes involved in glucose and lipid metabolic pathways, we wondered whether the final products in such metabolic pathways would change or not. Thus metabolomics provide a unique approach for examining the physiological status of livers with different treatments. PCA studies did not differentiate our samples, and no significantly different metabolites were found in our different group samples (Fig 2.8A-D). We also compared all groups and found other metabolic pathways, such as ether lipid, amino acid and folate, affected differently by OVX, E2, PPT or DPN treatments; however, we failed to find any significant difference in glucose or fatty acid metabolic pathways among all five groups (Fig 2.8E-N). Usually, metabolomics are used in studying bacteria, blood samples, feces samples or urine samples, not typically in organ samples, such as liver. Liver samples contain a mixture of liver cells, bile ducts and blood inside of tissue, which made the results difficult and complicated to analyze.

Environmental factors including availability of palatable and high-energy content food greatly impact the prevalence of obesity and its related metabolic syndrome. There are active 30 searches for the mechanisms of estrogen in the regulation of energy homeostasis. We found that ER-α played more important roles than ER-β in regulation of food intake, body mass, lipid content, body fat distribution, insulin sensitivity, and gene expressions of glucose and fatty acid metabolic pathways. Different ERs also induce different metabolites and metabolic pathways, but the answers are not assuring yet due to failure to purify liver cells from liver tissues.

31

References

1. Sherwin RS, Sacca L. Effect of epinephrine on glucose metabolism in humans: contribution of the liver. The American journal of physiology. 1984;247(2 Pt 1):E157-65. Epub 1984/08/01. 2. Cameron JM, Levandovskiy V, MacKay N, Utgikar R, Ackerley C, Chiasson D, et al. Identification of a novel mutation in GYS1 (muscle-specific glycogen synthase) resulting in sudden cardiac death, that is diagnosable from skin fibroblasts. Molecular genetics and metabolism. 2009;98(4):378-82. Epub 2009/08/25. 3. Irgens HU, Fjeld K, Johansson BB, Ringdal M, Immervoll H, Leh S, et al. Glycogenin-2 is dispensable for liver glycogen synthesis and glucagon-stimulated glucose release. The Journal of clinical endocrinology and metabolism. 2015;100(5):E767-75. Epub 2015/03/10. 4. Ogawa A, Ogawa E, Yamamoto S, Fukuda T, Sugie H, Kohno Y. Case of glycogen storage disease type VI (phosphorylase deficiency) complicated by focal nodular hyperplasia. Pediatrics international : official journal of the Japan Pediatric Society. 2010;52(3):e150-3. Epub 2010/08/21. 5. Barros RP, Machado UF, Warner M, Gustafsson JA. Muscle GLUT4 regulation by estrogen receptors ERbeta and ERalpha. Proceedings of the National Academy of Sciences of the United States of America. 2006;103(5):1605-8. Epub 2006/01/21. 6. O'Mahony F, Razandi M, Pedram A, Harvey BJ, Levin ER. Estrogen modulates metabolic pathway adaptation to available glucose in breast cancer cells. Mol Endocrinol. 2012;26(12):2058-70. Epub 2012/10/03. 7. Rettberg JR, Yao J, Brinton RD. Estrogen: a master regulator of bioenergetic systems in the brain and body. Frontiers in neuroendocrinology. 2014;35(1):8-30. Epub 2013/09/03. 8. Adeva-Andany MM, Perez-Felpete N, Fernandez-Fernandez C, Donapetry-Garcia C, Pazos-Garcia C. Liver glucose metabolism in humans. Bioscience reports. 2016;36(6). Epub 2016/10/21. 9. Dentin R, Denechaud PD, Benhamed F, Girard J, Postic C. Hepatic gene regulation by glucose and polyunsaturated fatty acids: a role for ChREBP. The Journal of nutrition. 2006;136(5):1145-9. Epub 2006/04/15. 10. Kabashima T, Kawaguchi T, Wadzinski BE, Uyeda K. Xylulose 5-phosphate mediates glucose-induced lipogenesis by xylulose 5-phosphate-activated protein phosphatase in rat liver. Proceedings of the National Academy of Sciences of the United States of America. 2003;100(9):5107-12. Epub 2003/04/10. 11. Jackson AS, Stanforth PR, Gagnon J, Rankinen T, Leon AS, Rao DC, et al. The effect of sex, age and race on estimating percentage body fat from body mass index: The Heritage Family Study. International journal of obesity and related metabolic disorders : journal of the International Association for the Study of Obesity. 2002;26(6):789-96. Epub 2002/05/31. 12. Whitlock G, Lewington S, Sherliker P, Clarke R, Emberson J, Halsey J, et al. Body-mass index and cause-specific mortality in 900 000 adults: collaborative analyses of 57 prospective studies. Lancet. 2009;373(9669):1083-96. Epub 2009/03/21. 13. Folsom AR, Kaye SA, Sellers TA, Hong CP, Cerhan JR, Potter JD, et al. Body fat distribution and 5-year risk of death in older women. Jama. 1993;269(4):483-7. Epub 1993/01/27. 14. Bluher M. The distinction of metabolically 'healthy' from 'unhealthy' obese individuals. Current opinion in lipidology. 2010;21(1):38-43. Epub 2009/11/17. 15. Miyazaki Y, Glass L, Triplitt C, Wajcberg E, Mandarino LJ, DeFronzo RA. Abdominal 32 fat distribution and peripheral and hepatic insulin resistance in type 2 diabetes mellitus. American journal of physiology Endocrinology and metabolism. 2002;283(6):E1135-43. Epub 2002/11/09. 16. Shi H, Clegg DJ. Sex differences in the regulation of body weight. Physiology & behavior. 2009;97(2):199-204. Epub 2009/03/03. 17. Nemoto Y, Toda K, Ono M, Fujikawa-Adachi K, Saibara T, Onishi S, et al. Altered expression of fatty acid-metabolizing enzymes in aromatase-deficient mice. The Journal of clinical investigation. 2000;105(12):1819-25. Epub 2000/06/23. 18. Camporez JP, Jornayvaz FR, Lee HY, Kanda S, Guigni BA, Kahn M, et al. Cellular mechanism by which estradiol protects female ovariectomized mice from high-fat diet-induced hepatic and muscle insulin resistance. Endocrinology. 2013;154(3):1021-8. Epub 2013/02/01. 19. Jelenik T, Roden M. How estrogens prevent from lipid-induced insulin resistance. Endocrinology. 2013;154(3):989-92. Epub 2013/02/23. 20. Barros RP, Gustafsson JA. Estrogen receptors and the metabolic network. Cell metabolism. 2011;14(3):289-99. Epub 2011/09/13. 21. Karjalainen A, Heikkinen J, Savolainen MJ, Backstrom AC, Kesaniemi YA. Mechanisms regulating LDL metabolism in subjects on peroral and transdermal estrogen replacement therapy. Arteriosclerosis, thrombosis, and vascular biology. 2000;20(4):1101-6. Epub 2000/04/15. 22. Fahraeus L. The effects of estradiol on blood lipids and lipoproteins in postmenopausal women. Obstetrics and gynecology. 1988;72(5 Suppl):18S-22S. Epub 1988/11/01. 23. Faulds MH, Zhao C, Dahlman-Wright K, Gustafsson JA. The diversity of sex steroid action: regulation of metabolism by estrogen signaling. The Journal of endocrinology. 2012;212(1):3-12. Epub 2011/04/23. 24. Wood GA, Fata JE, Watson KL, Khokha R. Circulating hormones and estrous stage predict cellular and stromal remodeling in murine uterus. Reproduction. 2007;133(5):1035-44. Epub 2007/07/10. 25. Frasor J, Barnett DH, Danes JM, Hess R, Parlow AF, Katzenellenbogen BS. Response- specific and ligand dose-dependent modulation of estrogen receptor (ER) alpha activity by ERbeta in the uterus. Endocrinology. 2003;144(7):3159-66. Epub 2003/06/18. 26. Li J, McMurray RW. Effects of estrogen receptor subtype-selective agonists on immune functions in ovariectomized mice. International immunopharmacology. 2006;6(9):1413-23. Epub 2006/07/19. 27. Chusyd DE, Wang D, Huffman DM, Nagy TR. Relationships between Rodent White Adipose Fat Pads and Human White Adipose Fat Depots. Frontiers in nutrition. 2016;3:10. Epub 2016/05/06. 28. Barnabas O, Wang H, Gao XM. Role of estrogen in angiogenesis in cardiovascular diseases. Journal of geriatric cardiology : JGC. 2013;10(4):377-82. Epub 2014/01/24. 29. Bailey CJ, Ahmed-Sorour H. Role of ovarian hormones in the long-term control of glucose homeostasis. Effects of insulin secretion. Diabetologia. 1980;19(5):475-81. Epub 1980/11/01. 30. Ford ES. Prevalence of the metabolic syndrome defined by the International Diabetes Federation among adults in the U.S. Diabetes care. 2005;28(11):2745-9. Epub 2005/10/27. 31. Crespo CJ, Smit E, Snelling A, Sempos CT, Andersen RE. Hormone replacement therapy and its relationship to lipid and glucose metabolism in diabetic and nondiabetic postmenopausal women: results from the Third National Health and Nutrition Examination Survey (NHANES III). Diabetes care. 2002;25(10):1675-80. Epub 2002/09/28.

33

32. Saglam K, Polat Z, Yilmaz MI, Gulec M, Akinci SB. Effects of postmenopausal hormone replacement therapy on insulin resistance. Endocrine. 2002;18(3):211-4. Epub 2002/11/27. 33. Margolis KL, Bonds DE, Rodabough RJ, Tinker L, Phillips LS, Allen C, et al. Effect of oestrogen plus progestin on the incidence of diabetes in postmenopausal women: results from the Women's Health Initiative Hormone Trial. Diabetologia. 2004;47(7):1175-87. Epub 2004/07/15. 34. Lundholm L, Bryzgalova G, Gao H, Portwood N, Falt S, Berndt KD, et al. The estrogen receptor {alpha}-selective agonist propyl pyrazole triol improves glucose tolerance in ob/ob mice; potential molecular mechanisms. J Endocrinol. 2008;199(2):275-86. Epub 2008/09/02. 35. Riant E, Waget A, Cogo H, Arnal JF, Burcelin R, Gourdy P. Estrogens protect against high-fat diet-induced insulin resistance and glucose intolerance in mice. Endocrinology. 2009;150(5):2109-17. Epub 2009/01/24. 36. SutterDub MT. Preliminary report: effects of female sex hormones on insulin secretion by the perfused rat pancreas. Journal de physiologie. 1976;72(6):795-800. Epub 1976/11/01.

34

Figure legends

Figure 2.1 Body weight, food intake and circulating E2.

Circulating E2 after sacrifice (Fig 2.1A) was analyzed by one-way ANOVA. Weekly body weight gain (Fig 2.1B), weekly body mass (Fig 2.1C), weekly food intake (Fig 2.1D) and accumulative food intake (Fig 2.1E) were analyzed by a two-way ANOVA (time x treatment) followed by Bonferroni posttest.

*: Significantly different comparing to sham+placebo groups (p < 0.05); †: Significantly different comparing to OVX+Placebo groups (p < 0.05), ‡: Significantly different comparing to OVX+E2 groups (p < 0.05), §:Significantly different comparing to OVX+PPT groups (p < 0.05).

Figure 2.2 Body composition, peri-gonadal fat weight and circulating leptin

Fat/Lean rations after 8 weeks were analyzed by one-way ANOVA (Fig 2.2A). Peri-gonadal fat pad weight (Fig 2.2B), and circulating leptin (Fig 2.2C) after 8-weeks treatments were analyzed by a one-way ANOVA.

*: Significantly different comparing to sham+placebo groups (p < 0.05); †: Significantly different comparing to OVX+Placebo groups (p < 0.05), ‡: Significantly different comparing to OVX+E2 groups (p < 0.05), §:Significantly different comparing to OVX+PPT groups (p < 0.05).

Figure 2.3 Plasma lipid profiles

Plasma lipid levels, including triglyceride (TG; Fig 2.3A), free fatty acid (FFA; Fig 2.3B); LDL- cholesterol (LDL-C; Fig 2.3C), HDL-cholesterol (HDL-C; Fig 2.3D), HDL/LDL ratio (Fig 2.3E), glycerol (Fig 2.3F) after 8 weeks were analyzed by a one-way ANOVA.

*: Significantly different comparing to sham+placebo groups (p < 0.05); †: Significantly different comparing to OVX+Placebo groups (p < 0.05), ‡: Significantly different comparing to OVX+E2 groups (p < 0.05), §:Significantly different comparing to OVX+PPT groups (p < 0.05).

Figure 2.4 Liver TG, liver weight, liver/whole body weight ratio and ADP/ATP ratio

35

Liver triglyceride content (TG; Fig 2.4A), liver weight (Fig 2.4B), liver/whole body weight ratio (Fig 2.4C), and ADP/ATP ratio (Fig 2.4D) after 8-weeks treatments were analyzed by a one-way ANOVA.

*: Significantly different comparing to sham+placebo groups (p < 0.05); †: Significantly different comparing to OVX+Placebo groups (p < 0.05), ‡: Significantly different comparing to OVX+E2 groups (p < 0.05), §:Significantly different comparing to OVX+PPT groups (p < 0.05).

Figure 2.5 Insulin sensitivity test and glucose tolerance test

Insulin tolerance test at 0, 15, 30, 45 and 60 min post-g injection (ip.) at basal (Fig 2.5A) and 8 (Fig 2.5B) weeks. Glucose tolerance test at 0, 15, 30, 45, 60 and 120 min post-glucose injection (ip.) at basal (Fig 2.5C), and 8 (Fig 2.5D) weeks. All data was analyzed by one-way ANOVA.

*: Significantly different comparing to sham+placebo groups (p < 0.05); †: Significantly different comparing to OVX+Placebo groups (p < 0.05), ‡: Significantly different comparing to OVX+E2 groups (p < 0.05), §:Significantly different comparing to OVX+PPT groups (p < 0.05).

Figure 2.6 Insulin sensitivity signal pathway

Western blot showed pAkt, tAkt, p-IRS-1 and t-IRS-1(Fig 2.6A). Percentage differences in pAkt/tAkt (Fig 2.6B) and p-IRS-1/ t-IRS-1 ratio (Fig 2.6C) at liver among different groups. All data was analyzed by one-way ANOVA.

*: Significantly different comparing to sham+placebo groups (p < 0.05); †: Significantly different comparing to OVX+Placebo groups (p < 0.05), ‡: Significantly different comparing to OVX+E2 groups (p < 0.05), §:Significantly different comparing to OVX+PPT groups (p < 0.05).

Figure 2.7 Gene expression levels involved in energy metabolism

Gene expression levels of sham+placebo group were set at 100%. Liver gene expression levels of glucose transporter 2 (Glut2), Phosphofructokinase-6 (6Pfk), Pyruvate kinase (Pk), Cytochrome c oxidase (Cox6a), Glycogen phosphrylase (Pygl), Phosphoenolpyruvate carboxykinase (Pepck), Glycogen synthase 2 (Gys2), Fatty acid synthase (Fasn), Acetyl co-

36 enzyme carboxylase (Acc), Transcription factor for lipogenic enzymes sterol-regulatory binding protein-1c (Srebp-1c), Peroxisome proliferator-activated receptor gamma coactivator 1 a (Pgc- 1α), Peroxisome proliferator-activated receptor alpha (Ppar-α) were analyzed by one-way ANOVA.

*: Significantly different comparing to sham+placebo groups (p < 0.05); †: Significantly different comparing to OVX+Placebo groups (p < 0.05), ‡: Significantly different comparing to OVX+E2 groups (p < 0.05), §:Significantly different comparing to OVX+PPT groups (p < 0.05).

Figure 2.8 Metabolic profiles

13 metabolites involved in glucose and lipid metabolic pathways of five mice groups were analyzed by principal component analysis (PCA) score plots (Fig 2.8A) and loading plots (Fig 2.8B). Total 255 metabolites of five groups metabolites were analyzed by principal component analysis (PCA) score plots (Fig 2.8C) and loading plots (Fig 2.8D). Major metabolic pathways of were impacted by different treatments during this study, the x-axis is the metabolic pathway impact value, while the y-axis is statistical significance (represented by p-value) of the impacted pathway from paired groups. The dot size corresponds to the x-axis value and the dot color corresponds to the y-axis value. A: Tryptophan metabolism; B: Vitamin B6 metabolism; C: Butanoate metabolism; D: Fructose and mannose metabolism; E: Pyruvate metabolism; F: Glycolysis or Gluconeogenesis; G: Glycine, serine and threonine metabolism; H: Fatty acid metabolism; I: Nicotinate and nicotinamide metabolism; J: TCA cycle; K: Riboflavin metabolism; L: Ether lipid metabolism; M: Primary bile acid biosynthesis; N: Glycerophospholipid metabolism; O: Arginine and proline metabolism; P: beta-Alanine metabolism; Q: Pantothenate and CoA biosynthesis. sham+placebo VS OVX+placebo (Fig 2.8E); sham+placebo VS OVX+E2 (Fig 2.8F); sham+placebo VS OVX+PPT (Fig 2.8G); sham+placebo VS OVX+DPN (Fig 2.8H); OVX+placebo VS OVX+E2 (Fig 2.8I); OVX+placebo VS OVX+PPT (Fig 2.8J); OVX+placebo VS OVX+DPN (Fig 2.8K); OVX+E2 VS OVX+PPT (Fig 2.8L); OVX+E2 VS OVX+DPN (Fig 2.8M); OVX+PPT VS OVX+DPN (Fig 2.8N).

37

Table 2.1 Primers for q-PCR

Genes Forward Reverse Gys2 5’-TTCTATGGTCATCTGGACTTTG-3’ 5’-CACGGTGACGTTACTCTTATG-3’ Pygl 5’-AGACAGAATTGTGGCCTTG-3’ 5’-GATGTCCGAGTGGATCTTTG-3’ 6Pfk 5’-ATTGACCGGCATGGAAAG-3’ 5’-CCATCTTGCTACTCAGGATTC-3’ Pk 5’-TGGCATCGAAAGTGGAAAG-3’ 5’-GGTGCAACTAGGTCAGAAAG-3’ COX 6a 5’-CTCAACGTGTTCCTCAAGTC-3’ 5’-GCCAGGTTCTCTTTACTCATC-3’ Glut2 5’-GGTCACTGTGGGCATAATC-3’ 5’-GGTACCAAAGGCACTCATAC-3’ Fas 5’- TCACCACTGTGGGCTCTGCAGAGAAGCGAG-3’ 5’-TGTCATTGGCCTCAAAAAGGGCGTCCA-3’ Acc 5’-CCCAGCAGAATAAAGCTACTTTGG-3’ 5’-TCCTTTTGTGCAACTAGGAACGT-3’ Srebp- 5’- GGCACTAAGTGCCCTCAACCT-3’ 5’- GCCACATAGATCTCTGCCAGTGT-3’ 1c Pgc-1α 5’- ATGTGTCGCCTTCTTGCTCT-3’ 5’- ATCTACTGCCTGGGGACCTT-3’ Ppar-a 5’- TTTGGATCCATGGTGGACACAGAGAGCCCCATC-3’ 5’-TTTGCGGCCGCTCAGTACATGTCTCTGTAGATCTCTTGC-3’ Pepck 5’- GGCCACAGCTGCTGCAG-3’ 5’- GGTCGCATGGCAAAGGG-3’ Gapdh 5’-GCGACTTCAACAGCAACTC-3’ 5’-GCCTCTCTTGCTCAGTGTCC-3’

38

Figure 2.1 Body weight, food intake and circulating E2.

Fig 2.1A Fig 2.1B

300 * † sham+placebo OVX+placebo OVX+E2 4 200 † OVX+PPT OVX+DPN ‡ ‡

(pg/ml) 100 2 * * ‡ * ‡ 0 ‡ Plasma estrodiol concentration estrodiol Plasma 0 † 1 2 3 4 5 6 7 8 OVX+E2 OVX+PPT OVX+DPN Weeks OVX+placebo sham+placebo -2 (g) gain weight body Weekly Fig 2.1C Fig 2.1D

sham+placebo 30 OVX+placebo § sham+placebo OVX+E2 OVX+placebo 33 ‡ ‡ OVX+PPT ‡ ‡ 27 OVX+E2 30 OVX+DPN OVX+PPT OVX+DPN 27 24

24 ‡ † 21 21 † † † *

weekly food intake (g) intake food weekly

Weekly body weight (g) weight body Weekly 18 18 0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 7 8 Weeks Weeks

Fig 2.1E

sham+placebo † 200 OVX+placebo † OVX+E2 150 OVX+PPT OVX+DPN ‡ ‡ 100 ‡

50

Cumulative food intake (g) intake food Cumulative 0 0 1 2 3 4 5 6 7 8 Weeks

39

Fig 2.2 Body composition, peri-gonadal fat weight and circulating leptin

Fig 2.2A Fig 2.2B

0.8 ‡ 2.5 0.6 ‡ ‡ 2.0

0.4 1.5

Fat/Lean † 1.0 0.2 * † 0.5 *

0.0 (g) weight Perigonadal 0.0

OVX+E2 OVX+E2 OVX+PPT OVX+DPN OVX+PPT OVX+DPN OVX+placebo OVX+placebo sham+placebo sham+placebo

Fig 2.2C

20 ‡

15 ‡

10

† Leptin (ng/ml) Leptin 5 *

0

OVX+E2 OVX+PPT OVX+DPN OVX+placebo sham+placebo

40

Figure 2.3 Plasma lipid profiles

Fig 2.3A Fig 2.3B

800 4000 * 600 3000 † † † † * 400 2000

(nmol/ml)

Plasma TG Plasma FFA (nmol/ml) FFA 200 1000

0 0

OVX+E2 OVX+E2 OVX+PPT OVX+DPN OVX+PPT OVX+DPN OVX+placebo OVX+placebo sham+placebo sham+placebo

Fig 2.3C Fig 2.3D

1000 † † 600 * † 800 * 400 † † 600

400 200 HDL-C (ug/ml) 200

LDL-C (ug/ml) LDL-C

0 0

OVX+E2 OVX+E2 OVX+PPT OVX+DPN OVX+PPT OVX+DPN OVX+placebo OVX+placebo sham+placebo sham+placebo

Fig 2.3E Fig 2.3F

4 † 300 † † 3

200 2 *

HDL/LDL 1 100

0 Plasma glycerol (nmol/ml) glycerol Plasma 0

OVX+E2 OVX+PPT OVX+DPN OOVX+E2 OVX+PPT OVX+DPN OVX+placebo OVX+placebo sham+placebo sham+placebo

41

Figure 2.4 Liver TG, liver weight, liver/body weight ratio, and ADP/ATP ratio

Fig 2.4A Fig 2.4B

2.0 200 * 1.5 ‡ ‡ 150 1.0 100

(nmol/ml)

LiverTG 0.5 50 (g) weight Liver

0.0 0

OVX+E2 OVX+E2 OVX+PPT OVX+PPT OVX+DPN OVX+DPN OVX+placebo OVX+placebo sham+placebo sham+placebo

Fig 2.4C Fig 2.4D

20 0.08 † § * 15 0.06

10

0.04 ADP/ATP 5 0.02

0.00 0

Liver/wholebodyweight

OVX+E2 OVX+E2 OVX+PPT OVX+DPN OVX+PPT OVX+DPN OVX+placebo OVX+placebo sham+placebo sham+placebo

42

Figure 2.5 Insulin sensitivity test and glucose tolerance test

Fig 2.5A Fig 2.5B

ITT after surgery ITT before surgery sham+placebo 150 sham+placebo 180 OVX+placebo OVX+placebo OVX+E2 OVX+E2 150 * OVX+PPT 120 OVX+PPT * OVX+DPN OVX+DPN 120

90

Glucose (mg/dl) Glucose 90

Glucose (mg/dl) Glucose

60 60 0 15 30 45 60 0 15 30 45 60 Time (Minutes) Time (Minutes)

Fig 2.5C Fig 2.5D

GTT after surg GTT before surg sham+placebo 500 OVX+placebo 400 sham+placebo OVX+E2 350 OVX+placebo 400 * OVX+PPT OVX+E2 OVX+DPN 300 OVX+PPT 300 ‡ 250 OVX+DPN ‡ 200 † † † 200 †

Glucose (mg/dl) Glucose 100 150

100 0 0 15 30 45 60 90 120 0 15 30 45 60 90 120 Time (Minutes)

43

Figure 2.6 Insulin sensitivity signal pathway

Fig 2.6A

p-AKT

t-AKT

p-IRS1

t-IRS1

sham+pl OVX+pl OVX+E2 OVX+PPT OVX+DPN

Fig 2.6B Fig 2.6C

0.5 † † 0.25 † † 0.4 0.20 0.3 0.15 0.2 0.10 * † ‡ §

p-AKT/t-AKT 0.1 * * ‡ § p-IRS-1/t-IRS-1 0.05 *

0.0 0.00

OVX+E2 OVX+E2 OVX+PPT OVX+DPN OVX+PPT OVX+DPN OVX+placebo OVX+placebo sham+placebo sham+placebo

44

Figure 2.7 Gene expression levels involved in energy metabolism

Fig 2.7A Fig 2.7B

Glut2 6Pfk 300 150 † †

200 100

-actin -actin   * ‡ §

100 50 6Pfk/

(% of Sham) (% of Sham) Glut2/ *

0 0

OVX+E2 OVX+E2 OVX+PPT OVX+DPN OVX+PPT OVX+DPN OVX+placebo OVX+placebo sham+placebo sham+placebo Fig 2.7C Fig 2.7D

Pk Cox6a 400 150

300

100

-actin 200 -actin  

Pk/ (% of Sham) 100 50

(% of Sham)

Cox6a/

0 0

OVX+E2 OVX+PPT OVX+DPN OVX+placebo OVX+E2 sham+placebo OVX+PPT OVX+DPN OVX+placebo sham+placebo Fig 2.7E Fig 2.7 F

Pygl Gys 150 † † 250

200 100 150

-actin -actin  ‡ § *  100 50

Pygl/

(% of Sham) Gys/ * (% of Sham) 50

0 0

OVX+E2 OVX+E2 OVX+PPT OVX+DPN OVX+PPT OVX+DPN OVX+placebo OVX+placebo sham+placebo sham+placebo

45

Fig 2.7G Fig 2.7H

Pepck Fasn 200 250 *

200 150 ‡ § 150

-actin -actin  100  100

† †

(% of Sham) Fasn/

(% of Sham) Pepck/ 50 50

0 0

OVX+E2 OVX+E2 OVX+PPT OVX+DPN OVX+PPT OVX+DPN OVX+placebo OVX+placebo sham+placebo sham+placebo Fig 2.7I Fig 2.7J

Acc 400 Srebp-1c 250 300 ‡ § 200

-actin -actin 150  200  † † 100

Acc/ (% of Sham) 100

(% of Sham)

Srebp-1c/ 50 0 0

OVX+E2 OVX+E2 OVX+PPT OVX+DPN OVX+PPT OVX+DPN OVX+placebo OVX+placebo sham+placebo sham+placebo Fig 2.7K Fig 2.7L

Ppar-a Pgc-1a 200 150 ‡§

150

100 -actin

-actin   100 †

50 (% of Sham)

(% of Sham) Ppar-a/ 50

Pgc-1a/

0 0

OVX+E2 OVX+E2 OVX+PPT OVX+DPN OVX+PPT OVX+DPN OVX+placebo OVX+placebo sham+placebo sham+placebo 46

Figure 2.8 Metabolic profiles

Fig 2.8A Fig 2.8B

Fig 2.8C Fig 2.8D

47

Fig 2.8E sham+placebo VS OVX+placebo

A

B

C D

G I

H F H E K J

Fig 2.8F sham+placebo VS OVX+E2

L

B

M N O

I G

E F J K H

48

Fig 2.8G sham+placebo VS OVX+PPT

P

B Q

K

L M G J E F H I

Fig 2.8H sham+placebo VS OVX+DPN

B A P

F G I E

D M K

H J

49

Fig 2.8I OVX+placebo VS OVX+E2

L N

M C I

P F J G H Q D E

Fig 2.8J OVX+placebo VS OVX+PPT

P G I K Q

F J B H E

50

Fig 2.8K OVX+placebo VS OVX+DPN

N

K F L E M H J I D

Fig 2.8L OVX+E2 VS OVX+PPT

L Q

B

K N G J

F I E H M

51

Fig 2.8M OVX+E2 VS OVX+DPN

N

L

I

F G

K M Q J H D E

Fig 2.8N OVX+PPT VS OVX+DPN

P

L F E I Q G D B K M J

52

Chapter 3. Estrogen and estrogen receptors on HepG2 cancer cell proliferation, apoptosis and leptin signal pathway.

Minqian Shen, Haifei Shi

This chapter was published in PLoS ONE 11.3 (2016): e0151455.

Abstract Obesity is a significant risk factor for certain cancers, including hepatocellular carcinoma (HCC). Leptin, a hormone secreted by white adipose tissue, precipitates HCC development. Epidemiology data show that men have a much higher incidence of HCC than women, suggesting that estrogen and its receptors may inhibit HCC development and progression. Thus, we hypothesized estrogen antagonized oncogenic action of leptin is uncertain. To investigate potential inhibitory effects of estrogens on leptin-induced HCC development, HCC cell line HepG2 cells were treated with leptin in combination with 17 β-estradiol (E2), estrogen receptor- α (ER-α) selective agonist PPT and ER-β selective agonist DPN. Cell number, proliferation, and apoptosis were determined, and leptin- and estrogen-related intracellular signaling pathways were analyzed. HepG2 cells expressed a low level of ER-β mRNA, and leptin treatment increased ER-β expression. E2 suppressed leptin-induced HepG2 cell proliferation and promoted cell apoptosis in a dose-dependent manner. Additionally E2 reversed leptin-induced STAT3 and leptin-suppressed SOCS3, which was mainly achieved by activation of ER-β. E2 also enhanced ERK via activating ER-α and activated p38/MAPK via activating ER-β. To conclude, E2 and its receptors antagonize the oncogenic actions of leptin in HepG2 cells by inhibiting cell proliferation and stimulating cell apoptosis, which was associated with reversing leptin-induced changes in SOCS3/STAT3 and increasing p38/MAPK by activating ER-β, and increasing ERK by activating ER-α. Identifying roles of different estrogen receptors would provide comprehensive understanding of estrogenic mechanisms in HCC development and shed light on potential treatment for HCC patients.

Key words:

Estradiol; Estrogen receptors; HepG2; p38-ERK; ERK; STAT3; SOCS3

53

3.1 Introduction

Hepatocellular carcinoma (HCC) is the most common primary carcinoma in the liver and the fourth most common cancer worldwide with high malignancy. The incidence and mortality rate of HCC continue to increase in the USA (1). The common risk factors of developing HCC include obesity, nonalcoholic fatty liver disease, chronic alcohol consumption, viral hepatitis infection, cirrhosis, and aflatoxin exposure (2, 3). Among the aforementioned risk factors, the rapid increase in obesity has become the prime cause of HCC, outweighing alcohol- or virus- related etiology (4). Epidemiological and clinical studies indicate that people with a body mass index (BMI) > 35 have greater risk for developing HCC, and obesity can precipitate other risk factors for HCC (5-7). Leptin is a 16-KD protein primarily secreted by white adipose tissue, and its level increases in obese animals including humans. Leptin is involved in the regulation of many physiological functions such as food intake and thermogenesis, as well as development of diseases such as atherosclerosis and carcinogenesis (8, 9). Abnormal level of leptin and dysregulation of leptin signaling have been identified to be crucial players in pathogenesis of HCC (10-13), contributing to the malignant development and progress of obesity-related liver cancer (14-16). Leptin signaling starts with binding to its long form receptor, then mediates intracellular signaling pathways (17). Leptin increases proliferation of many cancer cells mostly via Janus kinase (JAK) / signal transducers and activators of transcription 3 (STAT3) pathway (18), which is a common signal pathway shared by many cytokines and growth factors (19). Following dimerization and nuclear translocation, STAT3 binds to DNA as a transcriptional factor, and promotes cellular proliferation and reduces apoptosis (19). In normal cells STAT3 signal is controlled by suppressor of cytokine signaling proteins 3 (SOCS3), and down-regulation of SOCS3 is responsible for constitutive activation of STAT3 in HCC (20-22). Epidemiological data indicate that men have 3-5 times the risk of developing HCC compared with women, suggesting that sex hormones play a role in such gender disparity in HCC development (23). Whether estrogens play a protective or destructive role in HCC is under debate. Evidence has shown that estrogens suppress progression of fibrosis, tumor growth, and carcinogenesis in HCC (24, 25).

54

Estrogens act on both nuclear and membrane ERs to mediate estrogenic actions. ER-α and ER-β are found in all types liver cancer tissues (26). ER-α is usually considered as a proliferation activator in many reproductive cancer cells, including breast, ovarian, and endometrial cancers in females (27, 28). ER-β is less abundant in liver cells compared with ER-α (29). Decreases in levels of gene expression and protein of ER-β have been found in many cancers, such as breast cancer, prostate cancer, and ovarian cancer (30-32). The biological functions and significance of different subtypes of ERs, especially ER-β in HCC development remain largely unknown. HepG2 cell line is the most commonly used liver cancer cell line in metabolic studies. In general, obesity-related liver cancer does not involve any viral infection. Different from many other liver cancer cell lines, such as Hep3B, Huh7 and HA22T/VGH, HepG2 cells are poor host cells for supporting replication of hepatitis B virus or hepatitis C virus (33-35), and thus is appropriate for studying the interaction of obesity hormone leptin and estrogens in liver cancer cell growth. In this study, we applied HCC cell line HepG2 cells with 17 β-estradiol (E2), the most potent physiological estrogen, and selective agonists of ERs to investigate the roles of E2 and ER subtypes on leptin-induced HCC development.

3.2 Materials and methods

3.2.1 Reagents ER-α selective agonist 4,4’,4’’-(4-propyl-[1H]-pyrazole-1,3,5-triyl) trisphenol (PPT) and ER-β selective agonist 2,3-bis(4-hydroxy-phenyl)-propionitrile (DPN) were purchased from Fisher (Waltham, MA). Water soluble E2 was purchased from Sigma-Aldrich (St. Louis, MO). Human leptin was purchased from National Hormone & Peptide Program (Torrance, CA). 3.2.2 Cell culture and treatments The human hepatocellular cancer-derived cell line HepG2 was obtained from American Type Culture Collection (ATCC; Manassas, VA) and maintained in phenol red-free DMEM supplemented with 10% (v/v) heat-inactivated and charcoal-stripped FBS and 1% antibiotics of 50 U/ml penicillin and 50 µg/ml streptomycin (Invitrogen, Grand Island, NY) in a 37 ºC cell culture incubator. The initial cell concentration was 1 × 105 /ml. When cells were 70%-80% confluent, culture medium was starved in low serum (0.1% v/v FBS) for 16 h prior to experiments. Cells were treated with vehicle (1 μM dimethyl sulphoxide [DMSO]) as control, 55 leptin (100 ng/ml, similar to the circulating leptin level of obese humans (36)), serial concentrations of E2 (from 0 nM to 1000 nM), or combination of constant leptin (100 ng/ml) and serial concentrations of E2 for 48 hours. To examine the roles of different ERs involved in leptin signaling in HepG2 cells, cells were treated with 1 μM DMSO (no agonist), PPT, and DPN respectively. The dose of 1 μM for selective ER agonists is commonly used in liver cancer cells (37), as well as other non- reproductive cancer cells with lower expression of ERs than reproductive cancer cells (29), such as adrenal carcinoma cells (38, 39), medulloblastoma cells (40), thyroid carcinoma cells (41), and colon cancer cells (42). 3.2.3 Cell counting and viability assay HepG2 cell growth was evaluated after being treated for 48 hours in vitro by light microscopy HCC cell number and cell viability were measured using TC10TM automated cell counter (Bio-Rad, Hercules, CA) that counts cells within a 6-50 µm cell diameters range. 3.2.4 RNA interference HepG2 cells were grown to 70% confluency in 6-well plates in phenol-red free DMEM supplemented with 5% charcoal-stripped FBS. HepG2 cells were then transfected with one of small interfering RNAs (siRNA) specific for ER-α and ER-β or negative control siRNA (Santa Cruz Biotechnology, Santa Cruz, CA), according to the manufacturer’s transfection protocol. For all the experiments, transfected HepG2 cells were harvested 48 hours after transfection. 3.2.5 Quantification of cell proliferation assay by bromodeoxyuridine (BrdU) incorporation BrdU incorporation analysis was performed using an enzyme-linked immunosorbent assay (ELISA) kit (Millipore Corporation, Billerica, MA). Approximately 5 × 104 HepG2 cells were seeded in a sterile 96-well tissue culture plate for 24 h. Subsequently, BrdU was added to the HepG2 cultures for 4 h. HepG2 cells were fixed and DNA denatured. The prediluted BrdU detection antibody conjugated with peroxidase at a 2000X concentrate binds to the newly BrdU incorporated cellular DNA. The resultant immune complexes were quantified by spectrophotometer microplate reader set at 450/550 nm double wavelength. Relative light units/second is proportional to amount of DNA synthesis and number of proliferating cells. 3.2.6 Quantitative real-time PCR Total RNA was isolated from cells collected 48 h after vehicle or leptin treatment using a RNeasy Plus kit (Qiagen, Foster City, CA), and was reverse transcribed using a cDNA synthesis

56 kit (Bio-Rad). The primers synthesized by Integrated DNA Technologies (San Jose, CA) were: Era (accession#: NM_000125) forward 5’-GGAGGGCAGGGGTGAA-3’ and reverse 5’- GGAGGGCAGGGGTGAA-3’; Erb (accession#: NM_001214902) forward 5’- TTCCCAGCAATGTCACTAACTT-3’ and reverse 5’-TTGAGGTTCCGCATACAGA-3’; and β-actin (accession#: NM_001101) forward 5’-AGAGCTACGAGCTGCCTGAC-3’ and reverse 5’-AGCACTGTGTTGGCGTACAG-3’. Quantitative real-time PCR was carried out using SYBR green master mixes and an iCycler (Bio-Rad). The amplified products were confirmed via gel electrophoresis and melt curve analysis. Results were generated from triplicate experiments, calculated by a 2-ΔΔCt method, and normalized using a housekeeping gene β-actin. For each subtype of ERs, the expression level of ER from the leptin treatment group was expressed as fold change relative to the vehicle treatment group. 3.2.7 Western blotting Protein was extracted by homogenization using lysis buffer with sodium orthovanadate, phenylmethylsulf inhibitor (Santa Cruz Biotechnology, Santa Cruz, CA), and phosphatase inhibitor cocktail (Sigma-Aldrich). Protein lysates were resolved in 4%-15% tris-glycine gels and transferred to a nitrocellulose membrane (Bio-Rad). Cleaved-caspase 3 and total caspase 3, phosphorylated and total ERK, phosphorylated and total p38/MAPK, phosphorylated and total STAT3, SOCS3, β-actin (1:1000; Cell Signaling, Danvers, MA), and ER-α, ER-β (1:200; Santa Cruz, Dallas, TX) were detected by immunoblotting via chemiluminescence (Amersham™ ECL™ Prime, GE Healthcare). Protein band density was visualized and quantified using an Odyssey Infrared Imaging System (LI-COR, Lincoln, NE). Quantitative densitometric values of each protein of interest were normalized to β-actin or to the non-phosphorylated form of the protein. 3.2.8 Statistical analysis All data were presented as Mean ± SEM. All measurements were repeated for at least three independent experiments. Statistical analyses were performed using Prism5 GraphPad Software (La Jolla, CA). Two-way ANOVA comparing treatments with leptin and E2 or ER agonists followed by Bonferroni posttest was used to analyze cell numbers, proliferation and apoptosis, and intracellular signaling protein levels. One-way ANOVA followed by Tukey posttest was used to compare gene expression of ERs from groups treated with or without leptin.

57

Statistically significance was set at p < 0.05. Exact probabilities and test values were omitted for simplicity and clarity of the presentation of the results.

3.3 Results 3.3.1 E2 and ER agonists opposed leptin-induced increase in cell number. Leptin increased whereas the highest dose of E2 tested (1000 nM) decreased the numbers of HepG2 cells compared with control treatment (Fig 3.1A, 3.1B). Additionally cell numbers were greater in presence of leptin comparing to vehicle, except for the 100 nM E2 groups, indicating that the increase in cell number by leptin was blocked when the cells were treated with a combination of leptin and E2 at 100 nM (Fig 3.1B). Further investigation of the effects of different ER agonists showed that cell numbers were decreased by PPT and DPN compared with control (Fig 3.1A, 3.1C). Cells numbers were similar between the vehicle and leptin groups for the cultures treated with DPN, indicating that ER-β selective agonist DPN blocked leptin- induced increase in cell number (Fig 3.1C). 3.3.2 E2 and ER agonists decreased leptin-induced cell proliferation. Cell proliferation affects (contributes to) cell number and thus was assessed. E2 at 1000 nM but not 1 nM reduced cell proliferation, indicated by lower BrdU incorporation at 1000 nM E2 than control. Leptin promoted BrdU incorporation and thus HepG2 cell proliferation. Leptin- induced cell proliferation was reduced when cells were co-treated with E2 at 1000 nM or E2 at 1 nM (Fig 3.2A). Additionally, comparing to vehicle groups, cell proliferation was greater when cells were co-treated with leptin and 1 nM E2, but was less when cells were co-treated with leptin and 1000 nM E2 (Fig 3.2A), indicating that E2 at a high dose not only inhibited cell proliferation when treated alone but also opposed leptin-induced cell proliferation. Cell proliferation was inhibited by ER-α agonist PPT and ER-β agonist DPN. The increase in cell proliferation by leptin was reduced when cells were treated with leptin in combination with each of the ER agonists (Fig 3.2B). Thus, treatment of each of ER agonists alone suppressed cell proliferation, and each of ER agonists was able to oppose leptin-induced cell proliferation. 3.3.3 E2 and ER agonists promoted cell apoptosis. Number of cells is a net result of cell proliferation and apoptosis, thus apoptosis was assessed next. Leptin treatment alone did not affect protein level of cleaved-caspase 3, a critical

58 executioner in apoptotic cells responsible for proteolytic cleavage of many key proteins (43). In contrast, either with or without the presence of leptin, caspase 3 cleavage was increased following E2 treatment at 1000 nM or 1 nM, with a greater increase induced by 1000 nM E2 than 1 nM E2. Furthermore, for both dose of E2 tested here, E2-induced increase in caspase 3- dependent apoptosis was less evident with the presence of leptin than without leptin (Fig 3.3A). Cleaved-caspase 3 level was increased by treatment of PPT and DPN alone, or by co- treatment with leptin and each of these ER agonists (Fig 3.3B). Additionally either with or without the presence of leptin, DPN has greater apoptotic effects than PPT (Fig 3.3B), indicating that ER-β play more important roles in regulating cell apoptosis than ER-α (Fig 3.3B). Thus, E2 and its receptor agonists promoted caspase 3-dependent apoptosis in HepG2 cells. 3.3.4 Leptin treatment increased expression of ER-β in HepG2 cells. To begin to understand the effects of E2 and its receptors in HCC development, the expressions of different subtypes of ER in HepG2 cells with or without leptin treatment were examined. Quantitative real-time PCR analysis demonstrated that the mRNA levels of ER-α was similar between vehicle- and leptin- treated HepG2 cells, whereas ER-β expression of HepG2 cells was significantly higher following leptin treatment than vehicle treatment (Fig 3.4A-B). Furthermore, Western blot analysis demonstrated that protein levels of ER-α were similar between vehicle- and leptin-treated HepG2 cells, whereas ER-β protein level of leptin-treated cells was significantly higher than vehicle-treated cells (Fig 3.4C). 3.3.5 E2 and ER agonists diminished activation of leptin signaling pathway. SOCS3/STAT3, a common signaling pathway shared by many cytokines including leptin to promote cell growth and oncogenesis (44, 45), was assessed. Leptin alone increased p-STAT3 and decreased SOCS3, an upstream inhibitor of p-STAT3. E2 at 1 nM alone did not affect protein levels of pSTAT3 or SOCS3 in comparison to the vehicle-treated HepG2 cells, whereas E2 at 1000 nM increased SOCS3 without changing p-STAT3. Combined treatment of leptin and E2 at 1 nM or 1000 nM significantly reduced leptin-induced p-STAT3 and increased leptin- suppressed SOCS3 (Fig 3.5A-C). Among the three different ER selective agonists, p-STAT3 level was decreased by each of ER agonists, PPT and DPN, whereas SOCS3 signaling was increased only by DPN but not by PPT either with or without leptin treatment (Fig 3.5D-F). Although leptin-stimulated p-STAT3 was lessened by each of ER agonists PPT and DPN, p-STAT3 level of leptin-treated groups was

59 greater than their respective vehicle treated groups (Fig 3.5E). Leptin-suppressed SOCS3 was blocked only by DPN but not by PPT (Fig 3.5F). 3.3.6 E2 and ER agonists activated ERK and p38/MAPK signaling pathways. ERK and p38/MAPK are highly regulated in HCC development (46-48) and thus were assessed. Leptin alone did not affect pERK or p-p38/MAPK. In contrast, E2 at dose of 1 nM or 1000 nM activated both ERK and p38/MAPK. Interestingly, combination of leptin and E2 at a dose of 1 nM or 1000 nM activated ERK and p38/MAPK to a greater extent than E2 treatment alone (Fig 3.6A-C). Further study using selective ER agonists revealed that PPT activated ERK, while DPN activated p38/MAPK, either with or without leptin treatment (Fig 3.6D-F). Comparing to the groups without leptin treatment, the combination treatment of PPT with leptin significantly increased ERK activation, while the combination treatment of DPN with leptin significantly decreased p38/MAPK activation. In contrast, p-ERK level was similar between groups treated with PPT and DPN with or without leptin, and p-p38/MAPK level was similar between groups treated with PPT in presence of leptin or not (Fig 3.6D-F). 3.3.7 Loss of ER-β enhanced activation of leptin signaling pathway Since leptin signaling pathway (SOCS3/STAT3) was affected by ER agonists with or without leptin treatment (Fig 3.5D–5F), siRNA transfection to specific ER subtypes was used to selectively knock down each of respective ER subtype to determine role of each ER subtypes in SOCS3/STAT3 signaling pathway. Specific siRNA transfections to ER-α and ER-β successfully inhibited both RNA transcription (Fig 3.7A) and protein expression (Fig 3.7B) of each respective ER subtype with or without leptin treatment. Additionally, ER-β siRNA, but not ER-α siRNA, enhanced STAT3 signaling and inhibited SOCS3 signaling to a greater extent than control siRNA in leptin-treated HepG2 cells (Fig 3.7C).

3.4 Discussion Leptin stimulates liver cancer development and progress (14-16), which could contribute to the high incidence of liver cancer in obese population. Estrogens and their receptors have been implicated to promote some types of cancers in women, including breast, ovarian and endometrial cancers (27, 28). Paradoxically, women have a much lower rate of liver cancer than men (23). Although this lower incidence of HCC in women may be attributed to female sex hormone estrogens, the protective role of estrogens in leptin-induced HCC development has not

60 been investigated yet to our knowledge. In the present study, we investigated the effects of E2 and selective agonists of ER subtypes on oncogenic action of leptin. E2 is known to either induce cell proliferation or induce cell apoptosis, by stimulating either oncogenes or tumor suppressor genes, depending on whether cell types are estrogen sensitive or insensitive (49-53). The mechanisms underlying these opposite E2 effects could be partially explained by genomic and non-genomic estrogenic action via different ER isoforms, which then either modulates estrogen target gene transcription or rapidly activates intracellular signaling pathways, respectively (54-56). Therefore investigating the effects of ER-α and ER-β in HepG2 cell growth to understand genomic and non-genomic estrogenic actions in liver cancer development is of great interest. Cell number is a net outcome between cell proliferation and cell apoptosis. Leptin administration increased HepG2 cell number via enhancing cell proliferation. E2 not only increased cell apoptosis indicated by increased caspase 3 cleavage and decreased cell proliferation indicated by reduced BrdU incorporation when treated alone, but it also blocked leptin-induced increase in cell number, via stimulating apoptosis and blocking pro-proliferative effect of leptin (Figs 3.1-3.3). It is noteworthy that, although E2 at 1 μM has been used previously in vitro studies of liver culture cells that display specific binding of E2 and with similar properties of ERs as in reproductive tissues (57) but at a much lower expression level (29), caution should be taken into account that E2 greater than in vivo physiological concentration would potentially impact cell growth and intracellular molecules of HepG2 cells. Selective activation of different subtypes of ERs elicited distinct effects. The decrease in HepG2 cell number and blocking leptin-induced cell increase were more evident by ER-β selective agonist DPN than ER-α selective agonist PPT (Fig 3.1C). It is noteworthy that E2 or each of ER selective agonists suppressed proliferation and stimulated apoptosis when treated alone or co-treated with leptin in HepG2 cells (Figs. 3.2 and 3.3). It has been reported that estrogen and estrogen receptor agonists can stimulate cell apoptosis in a number of cancers, including ovarian cancer (58, 59). The apoptotic effect of estrogen or its receptor agonists was not tested in hepatocellular carcinoma. In this study, we demonstrated that ER agonists alone can induce apoptosis and suppress cell proliferation for the first time. Between the two nuclear receptors ER-α and ER-β, the dominant form of ER expressed in HepG2 cells is ER-α (Fig 3.4). Additionally, E2 binding affinity for ER-α is higher than for ER-β (60). Thus, E2 acts mainly

61 via ER-α and provokes similar effects as ER-α selective agonist PPT. In contrast, DPN is an ER- β selective agonist and has a 70-fold higher binding affinity for ER-β than for ER-α. DPN had far greater effects in regulating cell number than E2 (Fig 3.1C), suggesting that ER-β, instead of ER- α, plays major roles in regulating cell viability, including suppressing cell proliferation (Fig 3.2B) and inducing cell apoptosis (Fig 3.3B). Despite suppressing leptin-induced cell proliferation (Fig 3.2A), 1nM E2 did not significantly affect cell number in leptin-treated group (Fig 3.1B). Two possibilities may cause such discrepancy. First, BrdU incorporation into replicating DNA of the cells in S phase of the cell cycle was measured during a 4-hour period to indicate cell proliferation, which suggests, but is not necessary to correlated total number of cells grown during a 48-hour period. Second, although typical mature liver cells have diameters of 20-30 µm, within the measuring range of 6- 50 µm, some newly divided cells may have smaller sizes with diameters less than this measuring range and thus were not counted. We observed much lower expression of ER-β than other ER subtypes of vehicle-treated HepG2 cells, consistent with previous finding from other studies in the literature (61, 62), and increased expression of ER-β when HepG2 cells were treated with leptin for 48 h (Fig 3.4), a novel finding that was demonstrated for the first time. Albeit not clear whether this increased ER-β expression is unique to leptin treatment or is related to cell proliferation, such induction of ER-β expression in HepG2 cells may potentially play protective roles in leptin-induced HCC development in the presence of E2 or ER-β agonist. The effects of E2 and selective agonists of ER isoforms on leptin-related intracellular signaling pathways were investigated to understand potential mechanisms of estrogenic protection in leptin-induced HepG2 cell growth and thus obesity-related HCC. Obesity has become a pandemic disease and results in many health problems including some types of cancer (63). Leptin, an obesity hormone, suppresses SOCS3 and activates STAT3 signaling, which is one of the key points involved in a number of signaling pathways to promote cell growth and oncogenesis (44, 45). E2 treatment inhibited leptin-stimulated STAT3 and increased leptin- suppressed SOCS3, both of which were mostly evident via ER-β activation (Fig 3.5). This was further supported by the siRNA experiment that selective knockdown of ER-β, but not ER-α, enhanced leptin-induced stimulation of STAT3 (Fig 3.7). Some previous studies have reported activation of STAT3 up-regulates anti-apoptotic Bcl-xL family (64, 65), increase G1 to S phase

62 transition by increasing inappropriate expression of cell cycle-related proteins, including cyclin D1, cyclin-dependent kinase 4, cyclin E, cyclin A (64-66). SOCS3 gene has been found to be silenced by promotor methylation in human lung cancer, and restoration of SOCS3 suppresses cell growth and promotes cell apoptosis (67), indicating that increase in SOCS3 signaling enhances apoptosis. Inhibition of leptin STAT3 signaling, achieved by administration of leptin antagonists (68), by leptin receptor monoclonal antibodies (69), or by upregulating SOCS3 (70) would facilitate apoptosis and impede cancer cell growth. To summarize, suppression of STAT3 and activation of SOCS3 would inhibit cancer cell growth. Consistent with the literature, E2 and ER selective agonists inhibited leptin’s effects on cell growth, which was associated with suppression of STAT3 and activation of SOCS3 signaling. Some studies have suggested that ER-α is able to increase SOCS3 in endothelial cells of the cardiovascular system, a different type of cells from HCC cells with dominant ER-α expression but little ER-β expression (71, 72). In the current study, selective agonist of ER-β, but not ER-α increased SOCS3 protein level when treated alone or together with leptin. Thus, the elevated ER-β expression following leptin treatment would facilitate E2 or ER-β selective agonist DPN to promote apoptosis. It is noteworthy that PPT also enhanced apoptosis (Fig 3.3), but may be in an SOCS3-independent manner (Fig 3.5). Therefore, this study suggests that ER-β is the main ER subtype responsible for estrogen-mediated inhibition of leptin-induced changes in STAT3 signaling pathway, at least partially by enhancing SOCS3, the STAT3 inhibitor. ERK and p38/MAPK are highly regulated in HCC development (46-48). We found that selective agonists of ER-α activated ERK while selective agonist of ER-β activated p38/MAPK (Fig 3.6). There is evidence showing that ERK/MAPK pathway is regulated by multiple factors in HepG2 cells. For example, protein phosphatase 5 dephosphorylates Ser 338 of Raf-1, the upstream regulator of ERK, inactivates ERK (73), while inactivates ERK and activates p38/MAPK to regulate cell death (74). ER-α and ER-β share more than 96% similarity in DNA-binding region, but only 53% similarity in ligand-binding region (75), consequently inducing different signal pathways activated by ER-α and ER-β. ER-α selective agonist PPT activated ERK, but it failed to increase HepG2 cell proliferation (Fig 3.3B). In general, high p38 MAPK/ERK ratio induces dormancy and stops tumor cell growth, whereas high ERK/p38 MAPK ratio favors tumor cell growth (76, 77). Previous studies have reported that p38/MAPK significantly slows cell proliferation, induces dormancy, and induces

63 cell apoptosis (76, 77). Particularly, p38/MAPK increases Fas/CD-95 and Bax expression and subsequently activates caspase cascade to induce apoptosis (78). In addition, activation of p38/MAPK leads to dephosphorylation of ERK downstream molecule MEK and subsequent apoptosis (79). Findings from this study showed that ER-β selective agonist DPN was associated with increase in p38 MAPK signaling without changing ERK signaling, thus increased p38/MAPK / ERK ratio and inhibited cancer cell growth. Furthermore, p38/MAPK can enhance SOCS3 protein by stabilization of SOCS3 mRNA level (80), which could at least partially explain the protective role of ER-β in HCC progression. Inhibition of p38/MAPK and activation of ERK, on the other hand, promotes myoblast proliferation (81). Activation of ERK has been shown to inhibit apoptosis by suppressing the functions of pro-apoptotic proteins, enhancing the activity of anti-apoptotic molecules (82). Under certain conditions however, ERK activation actually induces apoptosis (83). For example, some DNA-damaging agents such as doxorubicin and antitumor compounds such as benzopyrene activate ERK and promote cell death in HepG2 cell line (84, 85). Whether ERK activation by PPT in HepG2 cells would lead to pro-apoptosis or anti-apoptosis has not been well established. Findings from the current study also demonstrated that PPT was associated with decreased cell growth and enhanced ERK phosphorylation without changing p38 activation, which indicated that ERK activation by E2 via PPT was pro-apoptosis in HepG2 cell line. To summarize, findings from the present study supported that estrogens attenuated leptin- induced HepG2 cell growth by facilitating apoptosis, which was associated with increase in SOCS3 and p38/MAPK signaling proteins. More importantly, we found that in HepG2 cells none of ERs was pro-proliferative, as they are in breast, endometrial and ovarian cancer cells; instead, all selective agonists of ER-α and ER-β stimulated cell apoptosis. The increased ER-β expression followed by leptin treatment is possibly due to a self-protection mechanism, although such protection from HCC development is evident with presence of E2 or agonists for ER-β. In future, the effects of ERs can be further confirmed using knock-down of respective receptor. Taken together, our data provided a better understanding of the protective role of estrogen in HCC development, and suggested an attractive target of estrogen receptor in the prevention and/or treatment of leptin-induced HCC. To conclude, this study provides better understanding of estrogenic protective role in obesity related HCC development and indicates that ER-β agonists may have implications in potential HCC treatment.

64

References

1. Njei B, Rotman Y, Ditah I, Lim JK. Emerging trends in hepatocellular carcinoma incidence and mortality. Hepatology. 2015;61(1):191-9. Epub 2014/08/22. 2. Gomaa AI, Khan SA, Toledano MB, Waked I, Taylor-Robinson SD. Hepatocellular carcinoma: epidemiology, risk factors and pathogenesis. World J Gastroenterol. 2008;14(27):4300-8. Epub 2008/07/31. 3. Smedile A, Bugianesi E. Steatosis and hepatocellular carcinoma risk. Eur Rev Med Pharmacol Sci. 2005;9(5):291-3. Epub 2005/10/20. 4. Khan FZ, Perumpail RB, Wong RJ, Ahmed A. Advances in hepatocellular carcinoma: Nonalcoholic steatohepatitis-related hepatocellular carcinoma. World journal of hepatology. 2015;7(18):2155-61. Epub 2015/09/04. 5. Calle EE, Rodriguez C, Walker-Thurmond K, Thun MJ. Overweight, obesity, and mortality from cancer in a prospectively studied cohort of U.S. adults. N Engl J Med. 2003;348(17):1625-38. Epub 2003/04/25. 6. Vanni E, Bugianesi E. Obesity and liver cancer. Clin Liver Dis. 2014;18(1):191-203. Epub 2013/11/28. 7. Yuan JM, Govindarajan S, Arakawa K, Yu MC. Synergism of alcohol, diabetes, and viral hepatitis on the risk of hepatocellular carcinoma in blacks and whites in the U.S. Cancer. 2004;101(5):1009-17. Epub 2004/08/27. 8. Garofalo C, Surmacz E. Leptin and cancer. J Cell Physiol. 2006;207(1):12-22. Epub 2005/08/20. 9. Janeckova R. The role of leptin in human physiology and pathophysiology. Physiol Res. 2001;50(5):443-59. Epub 2001/11/13. 10. Duan XF, Tang P, Li Q, Yu ZT. Obesity, adipokines and hepatocellular carcinoma. Int J Cancer. 2013;133(8):1776-83. 11. Polyzos SA, Kountouras J, Zavos C, Deretzi G. The potential adverse role of leptin resistance in nonalcoholic fatty liver disease: a hypothesis based on critical review of the literature. J Clin Gastroenterol. 2011;45(1):50-4. Epub 2010/08/19. 12. Ramani K, Yang H, Xia M, Ara AI, Mato JM, Lu SC. Leptin's mitogenic effect in human liver cancer cells requires induction of both methionine adenosyltransferase 2A and 2 beta. Hepatology. 2008;47(2):521-31. 13. Stefanou N, Papanikolaou V, Furukawa Y, Nakamura Y, Tsezou A. Leptin as a critical regulator of hepatocellular carcinoma development through modulation of human telomerase reverse transcriptase. BMC Cancer. 2010;10. 14. Vansaun MN, Mendonsa AM, Lee Gorden D. Hepatocellular proliferation correlates with inflammatory cell and cytokine changes in a murine model of nonalchoholic fatty liver disease. PLoS One. 2013;8(9):e73054. Epub 2013/09/17. 15. Wieser V, Moschen AR, Tilg H. Adipocytokines and hepatocellular carcinoma. Dig Dis. 2012;30(5):508-13. Epub 2012/10/31. 16. Xiong Y, Zhang J, Liu M, An M, Lei L, Guo W. Human leptin protein activates the growth of HepG2 cells by inhibiting PERKmediated ER stress and apoptosis. Mol Med Rep. 2014;10(3):1649-55. Epub 2014/07/16. 17. Tartaglia LA. The leptin receptor. Journal of Biological Chemistry. 1997;272(10):6093-6. 18. Saxena NK, Sharma D, Ding XK, Lin SB, Marra F, Merlin D, et al. Concomitant activation of the JAK/STAT, PI3K/AKT, and ERK signaling is involved in leptin-mediated

65 promotion of invasion and migration of hepatocellular carcinoma cells. Cancer research. 2007;67(6):2497-507. 19. Sweeney G. Leptin signalling. Cell Signal. 2002;14(8):655-63. Epub 2002/05/22. 20. Niwa Y, Kanda H, Shikauchi Y, Saiura A, Matsubara K, Kitagawa T, et al. Methylation silencing of SOCS-3 promotes cell growth and migration by enhancing JAK/STAT and FAK signalings in human hepatocellular carcinoma. Oncogene. 2005;24(42):6406-17. Epub 2005/07/12. 21. Wei RC, Cao X, Gui JH, Zhou XM, Zhong D, Yan QL, et al. Augmenting the antitumor effect of TRAIL by SOCS3 with double-regulated replicating oncolytic adenovirus in hepatocellular carcinoma. Hum Gene Ther. 2011;22(9):1109-19. Epub 2011/03/03. 22. Wu WY, Kim H, Zhang CL, Meng XL, Wu ZS. Loss of suppressors of cytokine signaling 3 promotes aggressiveness in hepatocellular carcinoma. J Invest Surg. 2014;27(4):197- 204. Epub 2014/01/31. 23. Bosch FX, Ribes J, Diaz M, Cleries R. Primary liver cancer: Worldwide incidence and trends. Gastroenterology. 2004;127(5):S5-S16. 24. Shimizu I, Yasuda M, Mizobuchi Y, Ma YR, Liu F, Shiba M, et al. Suppressive effect of oestradiol on chemical hepatocarcinogenesis in rats. Gut. 1998;42(1):112-9. Epub 1998/03/20. 25. Yasuda M, Shimizu I, Shiba M, Ito S. Suppressive effects of estradiol on dimethylnitrosamine-induced fibrosis of the liver in rats. Hepatology. 1999;29(3):719-27. Epub 1999/03/03. 26. Mazzarella L. Why does obesity promote cancer? Epidemiology, biology, and open questions. Ecancermedicalscience. 2015;9:554. Epub 2015/08/19. 27. Anbalagan M, Rowan BG. Estrogen receptor alpha phosphorylation and its functional impact in human breast cancer. Mol Cell Endocrinol. 2015. Epub 2015/01/20. 28. Persson I. Estrogens in the causation of breast, endometrial and ovarian cancers - evidence and hypotheses from epidemiological findings. J Steroid Biochem Mol Biol. 2000;74(5):357-64. Epub 2001/02/13. 29. Kuiper GG, Carlsson B, Grandien K, Enmark E, Haggblad J, Nilsson S, et al. Comparison of the ligand binding specificity and transcript tissue distribution of estrogen receptors alpha and beta. Endocrinology. 1997;138(3):863-70. Epub 1997/03/01. 30. Chan KK, Wei N, Liu SS, Xiao-Yun L, Cheung AN, Ngan HY. Estrogen receptor subtypes in ovarian cancer: a clinical correlation. Obstet Gynecol. 2008;111(1):144-51. Epub 2008/01/01. 31. Dondi D, Piccolella M, Biserni A, Della Torre S, Ramachandran B, Locatelli A, et al. Estrogen receptor beta and the progression of prostate cancer: role of 5alpha-androstane- 3beta,17beta-diol. Endocr Relat Cancer. 2010;17(3):731-42. Epub 2010/06/22. 32. Lazennec G, Bresson D, Lucas A, Chauveau C, Vignon F. ER beta inhibits proliferation and invasion of breast cancer cells. Endocrinology. 2001;142(9):4120-30. Epub 2001/08/23. 33. Clementi M, Testa I, Festa A, Bagnarelli P, Chang CM, Carloni G. Differential response of the human hepatoma-derived cell line HA22T/VGH to polypeptide mitogens. FEBS Lett. 1987;221(1):11-7. Epub 1987/08/31. 34. Jammart B, Michelet M, Pecheur EI, Parent R, Bartosch B, Zoulim F, et al. Very-low- density lipoprotein (VLDL)-producing and hepatitis C virus-replicating HepG2 cells secrete no more lipoviroparticles than VLDL-deficient Huh7.5 cells. J Virol. 2013;87(9):5065-80. Epub 2013/02/22.

66

35. Lin SJ, Shu PY, Chang C, Ng AK, Hu CP. IL-4 suppresses the expression and the replication of hepatitis B virus in the hepatocellular carcinoma cell line Hep3B. J Immunol. 2003;171(9):4708-16. Epub 2003/10/22. 36. Considine RV, Sinha MK, Heiman ML, Kriauciunas A, Stephens TW, Nyce MR, et al. Serum immunoreactive-leptin concentrations in normal-weight and obese humans. N Engl J Med. 1996;334(5):292-5. Epub 1996/02/01. 37. Hayashida K, Shoji I, Deng L, Jiang DP, Ide YH, Hotta H. 17beta-estradiol inhibits the production of infectious particles of hepatitis C virus. Microbiol Immunol. 2010;54(11):684-90. Epub 2010/11/04. 38. Altiok N, Ersoz M, Koyuturk M. Estradiol induces JNK-dependent apoptosis in glioblastoma cells. Oncol Lett. 2011;2(6):1281-5. Epub 2012/08/01. 39. Prieto LM, Brown JW, Perez-Stable C, Fishman LM. High dose 17 beta-estradiol and the alpha-estrogen agonist PPT trigger apoptosis in human adrenal carcinoma cells but the beta- estrogen agonist DPN does not. Horm Metab Res. 2008;40(5):311-4. Epub 2008/05/21. 40. Ciucci A, Meco D, De Stefano I, Travaglia D, Zannoni GF, Scambia G, et al. Gender effect in experimental models of human medulloblastoma: does the estrogen receptor beta signaling play a role? PLoS One. 2014;9(7):e101623. Epub 2014/07/08. 41. Chu R, van Hasselt A, Vlantis AC, Ng EK, Liu SY, Fan MD, et al. The cross-talk between estrogen receptor and peroxisome proliferator-activated receptor gamma in thyroid cancer. Cancer. 2014;120(1):142-53. Epub 2013/10/12. 42. Motylewska E, Stasikowska O, Melen-Mucha G. The inhibitory effect of diarylpropionitrile, a selective agonist of estrogen receptor beta, on the growth of MC38 colon cancer line. Cancer Lett. 2009;276(1):68-73. Epub 2008/12/23. 43. Shalini S, Dorstyn L, Dawar S, Kumar S. Old, new and emerging functions of caspases. Cell Death Differ. 2015;22(4):526-39. Epub 2014/12/20. 44. Epling-Burnette PK, Liu JH, Catlett-Falcone R, Turkson J, Oshiro M, Kothapalli R, et al. Inhibition of STAT3 signaling leads to apoptosis of leukemic large granular lymphocytes and decreased Mcl-1 expression. J Clin Invest. 2001;107(3):351-62. Epub 2001/02/13. 45. Li XP, Li CY, Li X, Ding Y, Chan LL, Yang PH, et al. Inhibition of human nasopharyngeal carcinoma growth and metastasis in mice by adenovirus-associated virus- mediated expression of human endostatin. Mol Cancer Ther. 2006;5(5):1290-8. Epub 2006/05/30. 46. Liu W, Ning R, Chen RN, Huang XF, Dai QS, Hu JH, et al. Aspafilioside B induces G2/M cell cycle arrest and apoptosis by up-regulating H-Ras and N-Ras via ERK and p38 MAPK signaling pathways in human hepatoma HepG2 cells. Mol Carcinog. 2015. Epub 2015/02/17. 47. Tong Y, Huang H, Pan H. Inhibition of MEK/ERK activation attenuates and potentiates pemetrexed-induced activity against HepG2 hepatocellular carcinoma cells. Biochem Biophys Res Commun. 2015;456(1):86-91. Epub 2014/12/03. 48. Zhao LY, Zhang J, Guo B, Yang J, Han J, Zhao XG, et al. MECP2 promotes cell proliferation by activating ERK1/2 and inhibiting p38 activity in human hepatocellular carcinoma HEPG2 cells. Cell Mol Biol (Noisy-le-grand). 2013;Suppl 59:OL1876-81. Epub 2013/11/10. 49. Ho SM. Estrogen, progesterone and epithelial ovarian cancer. Reprod Biol Endocrinol. 2003;1:73. Epub 2003/10/28.

67

50. Hsu I, Yeh CR, Slavin S, Miyamoto H, Netto GJ, Tsai YC, et al. Estrogen receptor alpha prevents bladder cancer via INPP4B inhibited akt pathway in vitro and in vivo. Oncotarget. 2014;5(17):7917-35. Epub 2014/10/04. 51. Marino M. Xenoestrogens challenge 17beta-estradiol protective effects in colon cancer. World J Gastrointest Oncol. 2014;6(3):67-73. Epub 2014/03/22. 52. Russo J, Russo IH. The role of estrogen in the initiation of breast cancer. J Steroid Biochem Mol Biol. 2006;102(1-5):89-96. Epub 2006/11/23. 53. Wei Q, Guo P, Mu K, Zhang Y, Zhao W, Huai W, et al. Estrogen suppresses hepatocellular carcinoma cells through ERbeta-mediated upregulation of the NLRP3 inflammasome. Lab Invest. 2015. Epub 2015/05/26. 54. Komyod W, Bohm M, Metze D, Heinrich PC, Behrmann I. Constitutive suppressor of cytokine signaling 3 expression confers a growth advantage to a human melanoma cell line. Mol Cancer Res. 2007;5(3):271-81. Epub 2007/03/22. 55. Matsuda Y, Ichida T. p16 and p27 are functionally correlated during the progress of hepatocarcinogenesis. Med Mol Morphol. 2006;39(4):169-75. Epub 2006/12/26. 56. Matsuda Y, Ichida T, Genda T, Yamagiwa S, Aoyagi Y, Asakura H. Loss of p16 contributes to p27 sequestration by cyclin D(1)-cyclin-dependent kinase 4 complexes and poor prognosis in hepatocellular carcinoma. Clin Cancer Res. 2003;9(9):3389-96. Epub 2003/09/10. 57. van Beurden-Lamers WM, Brinkmann AO, Mulder E, van der Molen HJ. High-affinity binding of oestradiol-17beta by cytosols from testis interstitial tissue, pituitary, adrenal, liver and accessory sex glands of the male rat. Biochem J. 1974;140(3):495-502. Epub 1974/06/01. 58. Ignatov T, Modl S, Thulig M, Weissenborn C, Treeck O, Ortmann O, et al. GPER-1 acts as a tumor suppressor in ovarian cancer. J Ovarian Res. 2013;6(1):51. Epub 2013/07/16. 59. Wang C, Lv X, He C, Hua G, Tsai MY, Davis JS. The G-protein-coupled estrogen receptor agonist G-1 suppresses proliferation of ovarian cancer cells by blocking tubulin polymerization. Cell Death Dis. 2013;4:e869. Epub 2013/10/19. 60. Kuiper GG, Lemmen JG, Carlsson B, Corton JC, Safe SH, van der Saag PT, et al. Interaction of estrogenic chemicals and phytoestrogens with estrogen receptor beta. Endocrinology. 1998;139(10):4252-63. Epub 1998/09/29. 61. Miceli V, Cocciadiferro L, Fregapane M, Zarcone M, Montalto G, Polito LM, et al. Expression of wild-type and variant estrogen receptor alpha in liver carcinogenesis and tumor progression. Omics : a journal of integrative biology. 2011;15(5):313-7. Epub 2011/02/26. 62. Porter LE, Elm MS, Van Thiel DH, Eagon PK. Hepatic estrogen receptor in human liver disease. Gastroenterology. 1987;92(3):735-45. Epub 1987/03/01. 63. Song RX, Santen RJ. Apoptotic action of estrogen. Apoptosis. 2003;8(1):55-60. Epub 2003/01/03. 64. Liang X, He M, Chen T, Wu Y, Tian Y, Zhao Y, et al. 17beta-estradiol suppresses the foam cell formation associated with SOCS3. Horm Metab Res. 2013;45(6):423-9. Epub 2013/02/23. 65. Mendelsohn ME, Karas RH. Molecular and cellular basis of cardiovascular gender differences. Science. 2005;308(5728):1583-7. Epub 2005/06/11. 66. Canfield S, Lee Y, Schroder A, Rothman P. Cutting edge: IL-4 induces suppressor of cytokine signaling-3 expression in B cells by a mechanism dependent on activation of p38 MAPK. J Immunol. 2005;174(5):2494-8. Epub 2005/02/25.

68

67. He B, You L, Uematsu K, Zang K, Xu Z, Lee AY, et al. SOCS-3 is frequently silenced by hypermethylation and suppresses cell growth in human lung cancer. Proc Natl Acad Sci U S A. 2003;100(24):14133-8. Epub 2003/11/18. 68. Otvos L, Jr., Kovalszky I, Scolaro L, Sztodola A, Olah J, Cassone M, et al. Peptide-based leptin receptor antagonists for cancer treatment and appetite regulation. Biopolymers. 2011;96(2):117-25. Epub 2010/06/22. 69. Fazeli M, Zarkesh-Esfahani H, Wu Z, Maamra M, Bidlingmaier M, Pockley AG, et al. Identification of a monoclonal antibody against the leptin receptor that acts as an antagonist and blocks human monocyte and T cell activation. J Immunol Methods. 2006;312(1-2):190-200. Epub 2006/05/13. 70. Xu Z, Huang G, Gong W, Zhou P, Zhao Y, Zhang Y, et al. FXR ligands protect against hepatocellular inflammation via SOCS3 induction. Cell Signal. 2012;24(8):1658-64. Epub 2012/05/09. 71. Acharya KD, Veney SL. Characterization of the G-protein-coupled membrane-bound estrogen receptor GPR30 in the zebra finch brain reveals a sex difference in gene and protein expression. Dev Neurobiol. 2012;72(11):1433-46. 72. Jin J, Yuan F, Shen MQ, Feng YF, He QL. Vascular endothelial growth factor regulates primate choroid-retinal endothelial cell proliferation and tube formation through PI3K/Akt and MEK/ERK dependent signaling. Mol Cell Biochem. 2013;381(1-2):267-72. 73. von Kriegsheim A, Pitt A, Grindlay GJ, Kolch W, Dhillon AS. Regulation of the Raf- MEK-ERK pathway by protein phosphatase 5. Nat Cell Biol. 2006;8(9):1011-6. Epub 2006/08/08. 74. Liu WH, Cheng YC, Chang LS. ROS-mediated p38alpha MAPK activation and ERK inactivation responsible for upregulation of Fas and FasL and autocrine Fas-mediated cell death in Taiwan cobra phospholipase A(2)-treated U937 cells. J Cell Physiol. 2009;219(3):642-51. Epub 2009/01/31. 75. Signoretti S, Loda M. Estrogen receptor beta in prostate cancer: brake pedal or accelerator? The American journal of pathology. 2001;159(1):13-6. Epub 2001/07/05. 76. Aguirre-Ghiso JA, Estrada Y, Liu D, Ossowski L. ERK(MAPK) activity as a determinant of tumor growth and dormancy; regulation by p38(SAPK). Cancer research. 2003;63(7):1684-95. Epub 2003/04/03. 77. Zhang W, Liu HT. MAPK signal pathways in the regulation of cell proliferation in mammalian cells. Cell Res. 2002;12(1):9-18. Epub 2002/04/11. 78. Chan QKY, Lam HM, Ng CF, Lee AYY, Chan ESY, Ng HK, et al. Activation of GPR30 inhibits the growth of prostate cancer cells through sustained activation of Erk1/2, c-jun/c-fos- dependent upregulation of p21, and induction of G(2) cell-cycle arrest. Cell Death Differ. 2010;17(9):1511-23. 79. Li SP, Junttila MR, Han J, Kahari VM, Westermarck J. p38 Mitogen-activated protein kinase pathway suppresses cell survival by inducing dephosphorylation of mitogen-activated protein/extracellular signal-regulated kinase kinase1,2. Cancer research. 2003;63(13):3473-7. Epub 2003/07/04. 80. Ehlting C, Lai WS, Schaper F, Brenndorfer ED, Matthes RJ, Heinrich PC, et al. Regulation of suppressor of cytokine signaling 3 (SOCS3) mRNA stability by TNF-alpha involves activation of the MKK6/p38MAPK/MK2 cascade. J Immunol. 2007;178(5):2813-26. Epub 2007/02/22.

69

81. Lee J, Hong F, Kwon S, Kim SS, Kim DO, Kang HS, et al. Activation of p38 MAPK induces cell cycle arrest via inhibition of Raf/ERK pathway during muscle differentiation. Biochem Biophys Res Commun. 2002;298(5):765-71. Epub 2002/11/07. 82. Lu Z, Xu S. ERK1/2 MAP kinases in cell survival and apoptosis. IUBMB Life. 2006;58(11):621-31. Epub 2006/11/07. 83. Cagnol S, Chambard JC. ERK and cell death: mechanisms of ERK-induced cell death-- apoptosis, autophagy and senescence. FEBS J. 2010;277(1):2-21. Epub 2009/10/22. 84. Alexia C, Fallot G, Lasfer M, Schweizer-Groyer G, Groyer A. An evaluation of the role of insulin-like growth factors (IGF) and of type-I IGF receptor signalling in hepatocarcinogenesis and in the resistance of hepatocarcinoma cells against drug-induced apoptosis. Biochem Pharmacol. 2004;68(6):1003-15. Epub 2004/08/18. 85. Lin T, Mak NK, Yang MS. MAPK regulate p53-dependent cell death induced by benzo[a]pyrene: involvement of p53 phosphorylation and acetylation. Toxicology. 2008;247(2- 3):145-53. Epub 2008/04/15.

70

Figure legends Figure 3.1 Effects of leptin, 17β-estradiol (E2) and estrogen receptor (ER) agonists on HepG2 cell number in cell culture. (A) HepG2 cells with different treatments, including vehicle DMSO (1μM) as control, leptin (100 ng/ml), E2 (1000 nM), ER-α selective agonist PPT (1 μM) and ER- β selective agonist DNP (1 μM), were evaluated after being treated for 48 hours using light microscopy (10 × magnification). (B, C) HepG2 cell numbers with different treatments, determined by TC10™ automated cell counter. *: Significantly different comparing to E0 or no agonist groups within the same leptin treatment (p < 0.05); †: Significantly different comparing to vehicle groups within the same E2 or ER agonist treatment (p < 0.05).

Figure 3.2 Effects of leptin, 17β-estradiol (E2), and estrogen receptor (ER) agonists on BrdU incorporation to determine proliferation of HepG2 cells. (A) BrdU incorporation assay was performed on HepG2 cells that were treated with vehicle DMSO (1 μM), leptin (100 ng/ml), E2 (1000 nM and 1 nM), or combinations of leptin 100 ng/ml and E2 1000 nM (L+E1000), or leptin 100 ng/ml and E2 1 nM (L+E1) for 48 hours. (B) BrdU incorporation assay was performed on HepG2 cells that were treated with vehicle DMSO (1 μM), leptin (100 ng/ml), ERα agonist PPT (1 μM), ERβ agonist DNP (1 μM), or combinations of leptin with PPT (L+PPT), leptin with DPN (L+DPN) for 48 hours. *: Significantly different comparing to E0 or no agonist groups within the same leptin treatment (p < 0.05); †: Significantly different comparing to vehicle groups within the same E2 or ER agonist treatment (p < 0.05).

Figure 3.3 Effects of leptin, 17β-estradiol (E2), and estrogen receptor (ER) agonists on protein levels of cleaved-caspase 3 and total caspase 3 to indicate apoptosis of HepG2 cells. (A) Western blot of protein extracts from HepG2 cells that were treated for 48 hours with vehicle DMSO (1 μM), leptin (100 ng/ml), E2 (1000 nM and 1 nM), or combinations of leptin 100 ng/ml and E2 1000 nM (L+E1000), or leptin 100 ng/ml and E2 1 nM (L+E1). (B) Western blot of protein extracts from HepG2 cells that were treated for 48 hours with vehicle DMSO (1 μM), leptin (100 ng/ml), ERα agonist PPT (1 μM), ERβ agonist DNP (1 μM), or combinations of leptin with PPT (L+PPT), leptin with DPN (L+DPN).

71

*: Significantly different comparing to E0 or no agonist groups within the same leptin treatment (p < 0.05); †: Significantly different comparing to vehicle groups within the same E2 or ER agonist treatment (p < 0.05).

Figure 3.4 Effects of leptin on estrogen receptor expression in HepG2 cells. HepG2 cells were treated without leptin or with leptin (100 ng/ml) for 48 hours. Quantitative real time-PCR and western blot were performed to measure mRNA levels and protein levels of ER-α and ER-β (A,B) Quantitative real time-PCR products were verified with 1.5% agarose running gel. (C) Estrogen receptor expressions were normalized to β-actin. Leptin treatment increased the mRNA level of ER-β, but not ER-α, in HepG2 cells, comparing to vehicle treatment. The data are presented as changes relative to the vehicle group. (D) Western blot data of estrogen receptors expression. *: Significantly different mRNA levels comparing to ER-α within the same leptin treatment (p < 0.05); †: Significantly different mRNA level comparing to vehicle groups (p < 0.05).

Figure 3.5 Effects of leptin, 17β-estradiol (E2) and estrogen receptor agonists on STAT3 and SOCS3 signaling in HepG2 cells. (A) Western blot results of different treatments, including vehicle DMSO (1 μM), leptin (100 ng/ml), E2 1000 nM (E1000), E2 1 nM (E1), and combinations of leptin and E2 1000 nM (L+E1000) or leptin and E2 1 nM (L+E1), for 48 hours. (D) Western blot results of different treatments, including vehicle DMSO (1 μM), leptin (100 ng/ml), ERα agonist PPT (1 μM), ERβ agonist DNP (1 μM), or combinations of leptin with PPT (L+PPT), leptin with DPN (L+DPN) for 48 hours. (B, E) Relative p-STAT3/t-STAT3 ratio normalized with control. (C, F) Relative SOCS3/β-actin ratio normalized with control. *: Significantly different comparing to E0 or no agonist groups within the same leptin treatment (p < 0.05); †: Significantly different comparing to vehicle groups within the same E2 or ER agonist treatment (p < 0.05).

Figure 3.6 Effects of leptin, 17β-estradiol (E2) and estrogen receptor agonists on ERK and p38/MAPK signaling in HepG2 cells. (A) Western blot results of different treatments, including vehicle DMSO (1 μM), leptin (100 ng/ml), E2 1000 nM (E1000), E2 1 nM (E1), and

72 combinations of leptin and E2 1000 nM (L+E1000) or leptin and E2 1 nM (L+E1), for 48 hours. (D) Western blot results of different treatments, including vehicle DMSO (1 μM), leptin (100 ng/ml), ERα agonist PPT (1 μM), ERβ agonist DNP (1 μM), or combinations of leptin with PPT (L+PPT), leptin with DPN (L+DPN) for 48 hour. (B, E) Relative p-ERK/t-ERK ratio normalized with control. (C, F) Relative p-p38/MAPK/t-p38/MAPK ratio normalized with control. *: Significantly different comparing to E0 or no agonist groups within the same leptin treatment (p < 0.05); †: Significantly different comparing to vehicle groups within the same E2 or ER agonist treatment (p < 0.05).

Figure 3.7 Effects of knockdown of each of specific ER subtypes using siRNAs on respective ER subtype gene expression (A), protein levels (B), and leptin signaling pathway (C). (A) Quantitative real time-PCR was performed to measure mRNA levels of ER-α and ER-β, normalized to β-actin and presented as fold change relative to the control siRNA-treated vehicle group. (B) Western blot analysis of protein levels of ER-α and ER-β, normalized to β-actin. (C) Western blot analysis of SOCS3/STAT3 signaling pathway, with relative p-STAT3/t STAT3 ratio and relative SOCS3/β-actin ratio of cells with different treatments. *: Significantly different comparing to control siRNA groups within the same leptin treatment (p < 0.05); †: Significantly different comparing to vehicle groups within the same siRNA treatment (p < 0.05).

73

Figure 3.1 Effects of leptin, 17β-estradiol (E2), and estrogen receptor agonists on HepG2 cell number in cell culture

Fig 3.1A

Control Leptin E2

PPT DPN

Fig 3.1B Fig 3.1C

E0 E1 E10 No agonist 8.010 5 † PPT † E100 † † 8.010 5 E1000 * DPN 6.010 5 * † 5 * 6.010 † * * * 4.010 5 4.010 5

Cell number Cell 5 2.010 number Cell 2.010 5 * * 0 0 Vehicle Leptin Vehicle Leptin

74

Figure 3.2 Effects of leptin, 17β-estradiol, and estrogen receptor (ER) agonists on proliferation of HepG2 cells, indicated by BrdU incorporation

Fig 3.2A Fig 3.2B

2.0 No agonist 2.0 † † E0 PPT E1000 DPN 1.5 † 1.5 E1 * † 1.0 * 1.0 * * † † * * *

0.5 0.5

BrdU incorporation (relative light units) light (relative

BrdU incorporation

(relative light units) light (relative

0.0 0.0 Vehicle Leptin Vehicle Leptin

75

Figure 3.3 Effects of leptin, 17β-estradiol (E2), and estrogen receptor (ER) agonists on apoptosis of HepG2 cells, indicated by protein levels of cleaved-caspase 3 and total caspase 3

Fig 3.3A

Cleaved-caspase 3

Caspase 3

Ctrl Lep E1000 E1 L+E1000 L+E1

0.6 E0 * E1000 E1 0.4

* *† 0.2 *†

0.0 cleaved-caspase 3/caspase 3 3/caspase cleaved-caspase Vehicle Leptin

Fig 3.3B

Cleaved-Caspase 3

Caspase 3

Ctrl Lep PPT DPN L+PPT L+DPN

1.0 No agonist PPT 0.8 DPN

0.6

0.4 * * 0.2 * * 0.0 cleaved-caspase 3/caspase 3 3/caspase cleaved-caspase Vehicle Leptin

76

Figure 3.4 Effects of leptin on estrogen receptor expression in HepG2 cells

Fig 3.4A

ER-α ER-β ER-α ER-β

β-actin β-actin

Fig 3.4B

150 1000 *

100 800

-actin

 600

-actin 

/

 /

50  400

(% of vehicle)

ER- (% of vehicle) ER- 200

0 0 Vehicle Leptin Vehicle Leptin

Fig 3.4C

ER-α ER-β

β-actin β-actin

Vehicle Leptin Vehicle Leptin

77

Figure 3.5 Effects of leptin, 17β-estradiol, and estrogen receptor agonists on STAT3 and SOCS3 signaling in HepG2 cells Fig 3.5A

p-STAT3

t-STAT3

SOCS3

β-actin

Ctrl Lep E1000 E1 L+E1000 L+E1

Fig 3.5B Fig 3.5C

1.0 E0 † 0.25 E0 0.8 E1000 * E1000 0.20 E1 E1 † 0.6 *

† -actin 0.15

† *  0.4 * †

0.10 * SOCS3/ p-STAT3/t-STAT3 0.2 0.05 † 0.0 0.00 Vehicle Leptin Vehicle Leptin

78

Fig 3.5D

p-STAT3

t-STAT3

SOCS3

β-actin

Ctrl Lep PPT DPN L+PPT L+DPN

Fig 3.5E Fig 3.5F

No agonist PPT 0.8 No agonist † DPN 0.15 * PPT *† * 0.6 DPN *† 0.10

0.4 -actin 

†

0.2 0.05 † SOCS3/ p-STAT3/t-STAT3 * * 0.0 0.00 Vehicle Leptin Vehicle Leptin

79

Figure 3.6 Effects of leptin, 17β-estradiol, and estrogen receptor agonists on ERK and p38/MAPK signaling in HepG2 cells Fig 3.6A

p-ERK

t-ERK

p-p38/MAPK

t-p38/MAPK

Ctrl Lep E1000 E1 L+E1000 L+E1

Fig 3.6B Fig 3.6C

† † 0.7 0.04 * E0 † E0 * * 0.6 E1000 E1000 E1 † 0.03 E1 0.5 * 0.4 * * * * 0.02 0.3 0.2 p-ERK/t-ERK 0.01 0.1

0.00 p-p38/MAPK/t-p38/MAPK 0.0 Vehicle Leptin Vehicle Leptin

80

Fig 3.6D

p-ERK

t-ERK

p-p38/MAPK t-p38/MAPK

Ctrl Lep PPT DPN L+PPT L+DPN

Fig 3.6E Fig 3.6F

No agonist 0.9 * * 0.5 PPT No agonist * † PPT * DPN 0.4 0.6 DPN 0.3

0.2 0.3

p-ERK/t-ERK 0.1

0.0 p-p38/MAPK/t-p38/MAPK 0.0 Vehicle Leptin Vehicle Leptin

81

Figure 3.7 Effects of ER siRNAs on cell proliferation and activation of leptin signaling pathways

Fig 3.7A

Control siRNA 5 1.5 ER- siRNA Control siRNA † ER- siRNA ER- siRNA 4 ER- siRNA 1.0 3

-actin

-actin

/

/

 2

 0.5 *

(% of vehicle)

ER-

(% of vehicle) ER- 1 * * * * 0.0 0 Vehicle Leptin Vehicle Leptin

Fig 3.7B

ER-α

ER-β

β-actin

C ERα siRNA ERβsiRNA L L+ ERα siRNA L+ ERβ siRNA

Control siRNA 0.15 0.8 ER- siRNA Control siRNA ER- siRNA ER- siRNA † 0.6 ER- siRNA

0.10 -actin

-actin 

 0.4 /

/ * 

0.05 † ER- ER- 0.2 * * * * 0.0 0.00 Vehicle Leptin Vehicle Leptin

82

Fig 3.7C

p-STAT3

t-STAT3

SOCS3

β-actin

C ERα siRNA ERβsiRNA L L+ ERα siRNA L+ ERβ siRNA

1.0 0.15 Control siRNA Control siRNA † * 0.8 ER- siRNA * ER- siRNA † ER- siRNA ER- siRNA † 0.10

0.6 -actin 

0.4

0.05 † † SOCS3/

p-STAT3/t-STAT3 0.2 *† 0.0 0.00 Vehicle Leptin Vehicle Leptin

83

Chapter 4. Effects of estrogen and estrogen receptors on HepG2 cancer cell metabolic profile

Minqian Shen, Fanyi Zhong, Jiangjiang Zhu, Haifei Shi

Manuscript was written and prepared by Minqian Shen. Haifei Shi served in an advisory capacity. All authors participated this research. This chapter will be submitted for publication. Text and figure formatting variations were due to journal guidelines.

Abstract

Increased glucose uptake and metabolism are common features of many cancers. Different from normal cells, cancer cells frequently have higher rates of glycolysis followed by lactate fermentation over oxidative phosphorylation (OXPHOS) in mitochondria. Additionally, 17-β estradiol (E2), a primary female sex hormone, has been shown to have various effects on increase of cell proliferation through ER-α activation followed by enhancing expression of genes that glucose transporter and glycolytic enzymes in breast cancer. However, it is unclear if liver cancer HepG2 cell metabolism favors glycolysis or OXPHOS, and the role of E2 in liver cancer metabolism is unknown. We hypothesized that HepG2 cells were more glycolysis- dependent than OXPHOS-dependent, and E2 inhibited glucose uptake and glycolysis mainly through ER-α activation in HepG2 cells. In this study, we applied an inhibitor for glycolysis 2- deoxy-D-glucose (2-DG), an inhibitor for lactate dehydrogenase sodium oxamate (OX), or an inhibitor for mitochondrial ATP synthase oligomycin (OM) to HepG2 cells, then cell viability, cytotoxicity and apoptosis were measured. We also applied E2, PPT or DPN to HepG2 cells to test changes in RNA expression across the transcriptome using RNA sequencing (RNA-Seq), to identify expression of genes involved in metabolic pathways using quantitative RT-PCR, and to analyze metabolomes of treated HepG2 cells using HPLC-MS technique. We found that HepG2 cells were more sensitive to 2-DG and OX treatments, with higher cell cytotoxicity and apoptosis and lower cell viability. We also identified different gene expression profiles and metabolic pathways due to different E2, PPT, and DPN treatments. We concluded that E2 directly affected liver cancer HepG2 cell metabolism mainly through ER-α activation.

Key words

Estradiol; Estrogen receptors; Glycolysis, OXPHOS; Metabolites

84

4.1 Introduction

Hepatocytes are highly specialized for their essential functions such as detoxification and glucose/lipid metabolism. Hepatocellular carcinoma (HCC) is one of the most common and deadly cancers worldwide (1). Different from normal, healthy cells that produce energy through oxidative phosphorylation (OXPHOS) in the mitochondria as the most common energy source, cancer cells have characterized glucose metabolism with increased glucose uptake and increased aerobic and anaerobic glycolysis. Cancer cells grow more rapidly than the blood vessels that nourish them, consequently obtain inadequate oxygen and experience hypoxia. In the absence of oxygen, cancer cells tend to use anaerobic glycolysis, i.e. lactic acid fermentation, as a primary source of ATP; along with reduction of pyruvate to lactic acid catalyzed by lactate dehydrogenase. Cancer cells also increase aerobic glycolysis even in the presence of oxygen, predominantly producing energy through glycolysis followed by lactic acid formation in the cytosol, rather than through OXPHOS in normal cells, a phenomenon known as the Warburg effect (2, 3). In addition, it has been shown that a high level of lactate, the end product of glycolysis, is associated with more aggressive cancer cells, such as drug-resistant and metastatic cancer cells (4). Although glycolysis yields less ATP, its rate is faster than mitochondrial OXPHOS, providing sufficient energy to meet cancer cell demand (5). Furthermore, glycolysis accumulates more glycolytic intermediates such as NADPH and ribose-5-phosphate for lipid and nucleic acids biosynthesis (6, 7).

Although the Warburg effect is widely recognized, not all types of cancer rely primarily on glycolysis for ATP production. Glycolysis actually contributes 1-64% ATP source in different cancers (8). For example, among four leukemia cell lines, NB4 cells are found more sensitive to 2-deoxy-D-glucose (2-DG), an inhibitor of glycolysis, than other leukemia cell lines, such as THP-1. NB4 cells are regarded as “glycolytic” leukemia cells. In contrast, THP-1 cells are resistant to 2-DG, but sensitive to oligomycin, an inhibitor of OXPHOS; thus THP-1 cells are considered as “OXPHOS” leukemia cell line (9). Therefore energy metabolic pathways utilized vary in different cancer cells, and specific pathways involved in energy metabolism could be used as a target for cancer therapy.

The unique change in HCC metabolism is the switch from glucose production to glucose usage. The activities of glucose-6-phosphatase, phosphoenolpyruvate carboxykinase, and

85 fructose-diphosphatase involved in gluconeogenesis are decreased (10-12) leading to reduced gluconeogenesis in HCC. In addition glycogenesis is decreased in liver cancer (13, 14). On the contrary, the enzymes involved in glucose catabolism, including hexokinase-2, glucose-6- phosphate dehydrogenase, and pyruvate kinase-M2, have increased activities in liver cancer (15) to increase proliferation of cancer cells.

Lipid de novo synthesis also increases in liver cancer. Evidence has shown increases in both gene and protein expression levels of fatty acid synthase, ATP citrate lyase, acetyl-CoA carboxylase, malic enzyme, sterol-CoA desaturase 1, 3-hydroxy-3-methylglutaryl-CoA reductase, mevalonate kinase, and squalene synthase, sterol regulatory element-binding protein 1 and 2, liver X receptors α and β, and carbohydrate-responsive element-binding protein in liver cancer (16). In addition, the expression of sterol-CoA desaturase is induced to catalyze saturated fatty acids into monounsaturated fatty acids to fulfill the need of new cell membrane production (16).

We first identified major metabolic pathways utilized by HepG2 cells, using an inhibitor for glycolysis 2-deoxy-D-glucose (2-DG), an inhibitor for lactate dehydrogenase sodium oxamate (OX), and an inhibitor for mitochondrial ATP synthase oligomycin (OM). 2-DG and OX inhibit glycolysis, whereas OM inhibits mitochondrial OXPHOS. Since we identified that estradiol (E2) and its receptors regulated HCC development (Chapter 3), we then tested the hypothesis that E2 and ER agonists PPT and DPN regulate HCC development, at least partially via affecting glucose/lipid metabolic pathways.

4.2 Methods and materials

4.2.1 Cell line and reagents

The human hepatocellular cancer-derived cell line HepG2 was obtained from American Type Culture Collection (ATCC; Manassas, VA) and maintained in phenol red-free DMEM supplemented with 10% (v/v) heat-inactivated and charcoal-stripped FBS and 1% antibiotics of 50 U/ml penicillin and 50 µg/ml streptomycin (Invitrogen, Grand Island, NY) in a 37 ºC cell culture incubator. The initial cell concentration was 1 × 105 /ml. When cells were 70%-80% confluent, culture medium was starved in low serum (0.1% v/v FBS) for 16 h prior to experiments. Cells were treated with 1μM vehicle, 1μM 17β-estradiol (E2, Sigma Aldrich, St

86

Louis, MO) , 1μM PPT (Fisher,Waltham, MA), 1μM DPN (Fisher,Waltham, MA), 10-50 mM oxamate (OX; Santa Cruz, Dallas, TX), 0.5-1.0 µg/ml oligomycin (OM; Santa Cruz, Dallas, TX), and 1-10mM 2-DG (Santa Cruz, Dallas, TX) (17, 18). 4.2.2 ApoTox-Glo™Triplex Assay

ApoTox-Glo Triplex Assay from Promega, Madison, WI, combines three assay chemistries to assess viability, cytotoxicity and caspase activation events within a single assay well. In the first part of the assay, it measures two protease activities simultaneously; one being a marker of cell viability and the other being a marker of cytotoxicity. Peptide substrate (glycylphenylalanyl-aminofluorocoumarin; GF-AFC) enters intact cells where it is cleaved by the live-cell protease to generate a fluorescent signal proportional to the number of living cells. This live-cell protease becomes inactive upon loss of cell membrane integrity and leaks into the surrounding culture medium. Peptide substrate (bis-alanylalanyl-phenylalanyl-rhodamine 110; bis-AAF-R110) is used to measure dead-cell protease activity, which is released from cells that have lost membrane integrity. Bis-AAF-R110 is not cell-permeable, so no signal from this substrate is generated by intact, viable cells. The live- and dead-cell proteases produce different products, AFC and R110, which have different excitation and emission spectra, allowing them to be detected simultaneously. In the second part of the assay, the Caspase-Glo® 3/7 Reagent, added in an "add-mix-measure" format, results in cell lysis, followed by caspase cleavage of the substrate and generation of a "glow-type" luminescent signal produced by luciferase (Manufactory protocol).

HepG2 cells of approximately 500 cells/well were seeded in a flat 96-well micro-plate in triplicates. After 24 days of exposure to different treatments, 20 μl of Viability/Cytotoxicity reagent containing both GF-AFC and bis-AAF-R110 substrates was added to each well, briefly mixed by orbital shaking at 300-500 rpm for 30 seconds, and then incubated at 37°C for 30 minutes. Fluorescence was measured at 400Ex/505Em (Viability) or 485Ex/520Em (Cytotoxicity). Next, 100 μl of Caspase-Glo 3/7 reagent was added into each well, briefly mixed by orbital shaking at 300-500 rpm for 30 seconds, and incubated at room temperature for 30 minutes. Luminance was measured using a BMG Labtech plate reader and analyzed using NOVO star software.

4.2.3 RNA isolation

87

Total RNA from HepG2 was obtained using RNeasy Mini Kit (QIAGEN, US) following the manufacturer's instructions. Briefly, a number of 106 HepG2 cells were lysed and homogenized in 350 µl RLT buffer. Each flow-through was transferred to a fresh 1.5 ml Eppendorf tube and mixed with 300 µl of 70% EtOH by pipetting. Each lysate solution was then applied to an RNeasy mini column in a 2 ml collection tube and spun for 15 seconds at 10,000×g, discarding the flow-through afterwards, and washed twice with RW1 buffer provided in kit to remove biomolecules such as carbohydrates, proteins, fatty acids etc., that are non-specifically bound to the silica membrane, while RNA molecules larger than 200 bases remain bound to the column. The columns were then washed with buffer RPE to wash membrane-bound RNA. Finally high-quality RNA was eluted in 30 µL of ribonuclease-free water. Concentrations of the RNA samples were monitored by gel electrophoresis and quantified using Nanodrop 2000 Spectrophotometer (Thermo Scientific, Wilmington, DE, USA).

4.2.4 RNA sequencing

Total RNA samples of HepG2 cultured cells were sent for mRNA sequencing using the Illumina HiSeq2000 sequencing system at the Genomics, Epigenomic and Sequencing Core at the University of Cincinnati. The Illumina HiSeq2000 system is a high-throughput sequencing technology based on massive parallel sequencing of millions of fragments using proprietary reversible terminator-based sequencing chemistry, the same sequencing principle used in the Illumina Genome Analyzer II system. The RNA sequencing process was performed according to the manufacturer's protocols as well as standardized protocols developed by the core laboratory at the University of Cincinnati.

4.2.5 Quantitative RT-PCR

Total RNA was extracted and reverse transcribed into cDNA using 1 μg RNA. Expression of genes related to lipogenesiss (fatty acid synthase, Fasn; sterol regulatory element- binding transcription factor 1, Srebp-1c), lipolysis (peroxisome proliferator-activated receptor- gamma coactivator, Pgc-1α; acetyl co-enzyme carboxylase, Acc), gluconeogenesis (phosphoenolpyruvate carboxykinase, Pepck), glycogen synthesis gene expression (glycogen synthase, Gys), glycogenolysis (glycogen phosphorylase, Pygl), glycolysis (6-phosphofructo-1- kinase, 6Pfk1; pyruvate kinase, Pk), and oxidative phosphorylation (cytochrome c oxidase, Cox) were measured (Table 2.1). Glyceraldehyde-3-phosphate dehydrogenase (Gapdh) was used as a

88 reference gene Quantitative PCR was run in triplicates using iQ SYBR Green Supermix (Bio- Rad, Hercules, CA) and an iCycler (Bio-Rad) with 40 cycles of amplification (95 °C for 10 s) and annealing (55 °C for 30 s). The amplified products were confirmed via gel electrophoresis and melt curve analysis. Results were calculated by a 2−ΔΔCt method, and presented using sham groups as 100%.

4.2.6 HPLC-MS metabolite analysis

Cells ~10^5 were transferred into a 2 mL sample preparation tube. Tissues were homogenized with PBS and the metabolites were extracted by 250 μL cold methanol in −20 °C for 20 min. Then 50 μL isotope labeled amino acid mixture was added as internal standards (Cambridge Isotope Laboratories, Inc.) and then mixed one more time. The mixture was stored at −20 °C for 20 min and centrifuged for 20 min. Then 150 μL of the supernatant was collected and dried on a vacuum concentrator. The dried sample was later reconstituted in 50% H2O and 50% acetonitrile mixture and kept in 4 °C autosampler for HPLC-MS runs. Briefly, all raw data were manually inspected using the Quanbrowser module of Xcalibur version 2.0 (Thermo Fisher Scientific). The mass spectrometry data normalized by the liver tissue weight at the point of metabolite extraction. JMP Pro12 (SAS Institute, Cary, NC) was used for statistical analysis. Principle components analysis (PCA) was applied to compare the metabolic profiles between different groups. MetaboAnalyst 3.0 (http://www.metaboanalyst.ca/) was used to explore the metabolic pathway impact of each treatment.

Metabolic pathway impact analyses were conducted to put individual metabolite into the context of connected metabolic pathway networks. All detected metabolites were included in the metabolic pathway analysis, so the broader coverage of extensive metabolic networks can be achieved. The x-axis is the metabolic pathway impact value, and the y-axis is the statistical significance (represented by log of p-value) of the impacted pathways. The dot size corresponds to the x-axis value and the dot color corresponds to the y-axis value.

4.4.7 Statistics

Data were presented as mean ± SEM. Prism 5 (La Jolla, CA) was used to perform one- way ANOVA followed by Tukey posttest was used to compare cell growth number, viability, cytotoxicity, apoptosis and gene expression levels, A test with p<0.05 was considered statistically significant.

89

4.3 Results

4.3.1 Effects of OX, OM, 2-DG on E2-treated HepG2 cell number.

Consistent with our previous finding, E2 reduced HepG2 cell number. 2-DG at all doses tested and high dose of OX decreased the numbers of HepG2 cells compared with control treatment (Fig 4.1A, 4.1B, 4.1D). Treatment with up to 1 µg/ml of OM did not reduce HepG2 cells significantly (Fig 4.1A, Fig 4.1C). Thus, HepG2 cells were resistant to OM and sensitive to OX and 2-DG, thus indicating that this cell line was more dependent on glycolysis, instead of OXPHOS in mitochondria for ATP production.

4.3.2 Effects of different doses of sodium oxamate (OX), oligomycin (OM), 2-deoxy-D-glucose (2-DG) on HepG2 cytotoxicity, viability and apoptosis.

Cytotoxicity, viability and apoptosis were measured in HepG2 cells treated with arranges of different doses of OX, OM and 2-DG, using an APOTOX-Glo Triplex assay (Fig 4.2). Cells treated with OX or 2-DG showed increased cytotoxicity, reduced viability, and increased apoptosis in a dose-dependent manner. However, OM treatment did not significantly affect cytotoxicity, viability, or apoptosis. These results confirmed that HepG2 cell line was primarily dependent on glycolysis, but independent of OXPHOS, for ATP production.

4.3.3 Effects of E2, PPT and DPN on gene expression involved in energy metabolism

HepG2 cells display enhanced rates of glucose usage. E2 and ERα agonist suppress HepG2 cell growth (Chapter 3; (19)). Inhibition of glycolysis, but not inhibition of OXPHOS, suppressed HepG2 cell growth. We tested the hypothesis that E2 and ERα agonist could suppress glycolysis and thus glucose usage to achieve anti-cancer effect, by measuring the effects of E2 and ER agonists on expression of factors and metabolites of glucose and lipid metabolism.

Glut2 gene expression was increased by E2 (t=5.400, p < 0.05) or PPT (t=3.911, p < 0.05) compared with control treatment (Fig 4.3A), indicating that E2 and PPT increased glucose uptake. Gene expressions of 6Pfk and Pk, which codes for two rate-limiting enzymes involved in glycolysis, were measured. E2, PPT, and DPN treatments all significantly lowered 6Pfk expression comparing with the control treatment group, with E2 having the most prominent effects (Fig 4.3B). E2 and DPN treatment, but not PPT treatment, significantly reduced Pk expression (Fig 4.3C). Expression of Cox6b, an essential gene involved in OXPHOS, and Gys, a

90 critical gene for the formation of glycogen, however, were not significantly changed by any E2 or ER agonist treatment (Fig 4.3D, Fig 4.3E). The mRNA levels of Gypl, coding the enzyme that breaks down glycogen in the liver, was significantly increased by E2 treatment (Fig 4.3F). Interestingly, PPT induced Pepck, a gene regulating gluconeogenesis comparing with E2 (t=3.936, p < 0.05) and DPN (t=3.617, p < 0.05) groups, although no difference was found when compared with the control group (t=3.449, p > 0.05) (Fig 4.3G). Pgc-1α and Pparγ are two genes regulating lipolysis, and both E2 and DPN treatments induced Pcg1α compared with control treatment, with E2 having more pronounced effect (Fig 4.3H). All E2, PPT and DPN treatments significantly induced Pparγ expression compared with the control treatment, and E2-treated cells had higher Pparγ expression level than PPT and DPN treatments (Fig 4.3I). Furthermore, three important genes regulating fatty acid synthesis, Fasn, Acc, and Srebp-1c, were measured. E2 and PPT treatments significantly increased Acc expression compared with control and DPN-treated groups (Fig 4.3L), and no significant difference in Fasn or Srebp-1c was detected among any groups (Fig 4.3J, Fig 4.3K).

4.3.4 Effects of E2, PPT and DPN on HepG2 transcriptome

The target genes up- or downregulated in response to E2, PPT and DPN treatments were identified using RNA-seq, a method for genome-wide expression profiling. More than 4000 genes were significantly affected by different treatments, and a few gene networks involved in variety of physiological functions, such as cell cycle and cell metabolism, were defined (Table 4.1A). In brief, 1736 unique genes were upregulated by E2 treatment, 233 unique genes were upregulated by PPT treatment, and 209 unique genes were upregulated by DPN treatment. Among these genes, 81 common genes were upregulated by E2 and PPT, 115 common genes were upregulated by E2 and DPN, and 229 common genes were upregulated by PPT and DPN. Additionally, there were 118 genes upregulated by all three treatments with E2, PPT and DPN (Fig 4.4A). We also detected 752 unique genes downregulated by E2 treatment, 345 unique genes downregulated by PPT treatment, and 221 unique genes downregulated by DPN treatment. Among these down-regulated genes, 109 common genes were downregulated by E2 and PPT, 75 common genes were downregulated by E2 and DPN, and 189 common genes were downregulated by PPT and DPN. Additionally there were 247 common genes downregulated by E2, PPT and DPN treatments (Fig 4.4B).

91

We validated gene expression of Socs3, Elovl3, Cdc20, Notch1, Pparg, and Il6r by qPCR. These genes were chosen due to different expression patterns according to RNA-seq results, and their involvment in different cellular functions of STAT3 signal regulation (Socs3), fatty acid elongation (Elovl3), cell cycle division cycle (Cdc20), cancer development (Notch1), lipid metabolism (Pgc-1α), and inflammation (Il6r). Activation of Notch1 signaling contributes HCC (20) and downregulation of Notch1 signal inhibits HCC (21). The mRNA level of Notch1 was significantly lower in DPN-treated groups than other groups (Fig 4.5A). Activation of Pparg inhibits proliferation of HCC cells (22). E2 and DPN, not PPT, significantly increased Pparg expression comparing with control group (Fig 4.5B). Instead, PPT treatment significantly reduced Cdc20 expression level compared with all other groups (Fig 4.5C). Socs3 regulates STAT3 signal (23). E2 and DPN, but not PPT, induced Socs3 expression by 6-fold and 5-fold respectively (Fig 4.5D). This is consistent with an increase of Socs3 by 2.04 log2 fold from the RNA-seq data. Interestingly, PPT and DPN, but not E2 increased Il6r expression level (Fig 4.5E), and there was no significantly difference in Elovl3 expression level among different treatments (Fig 4.5F), which is consistent with that Elovl3 gene change was not detected from RNA-Seq data.

4.3.5 Metabolic profiles

The HPLC-MS system was used to detect ADP/ATP ratio among different treatments. Although the difference did not reach statistical significance, E2, PPT and DPN treatments intended to increase the ratio (Fig 4.6A). Twelve metabolites directly associated with glucose and lipid metabolism, including citrate, 2,3-diphosphoglyceric acid, glucose 6-phosphate, glutaric acid, 3-hydroxybutyric acid, glyceric acid, lactic acid, nicotinate, 6-hydroxynicotinic acid, FAD, glyceraldehyde, 2-hydroxybutyric acid were then detected by the HPLC-MS (Fig 4.6B, Fig 4.6C). The PCA score plots indicated that E2, PPT and DPN treatments displayed unique metabolic fingerprints on HepG2 glucose and lipid , which were different from control treatment (Fig 4.6B). The loading plots (Fig 4.6C) showed the contributions of different metabolites to the separation of metabolic pathways.

HPLC-MS didn’t distinguish any difference in the total 255 metabolites involved in multiple metabolic pathways (Fig 4.6D, Fig 4.6E). All detected metabolites were analyzed in the metabolic pathway analysis using metabolic pathway impact analyses (Fig 4.6F - Fig 4.7K).

92

4.4 Discussion

We previously reported that E2 suppressed HepG2 proliferation mainly through ER-β activation (Chapter 3), and this study explored the potential underlying mechanism involving differential activation of metabolic pathways by E2 and its receptor agonists. Reprogrammed energy metabolism is critical for the survival of tumor cells. The well-known metabolic change in cancer cells is the Warburg effect, allowing cancer cells to be more dependent on glycolysis to produce ATP instead of OXPHOS, even in the presence of oxygen. Inhibition of hexokinase 2, pyruvate kinase isozymes muscle type 2, or lactate dehydrogenase profoundly suppresses tumor progression in HCC, suggesting the critical role of the Warburg effect in cancer physiology (24- 26). In this study, we first showed that HepG2 cell growth was suppressed by a glycolysis inhibitor 2-DG and a lactate dehydrogenase inhibitor sodium oxamate, but not by a mitochondrial ATP synthase inhibitor oligomycin (Fig 4.1 and Fig 4.2), which suggested that HepG2 cells more dominantly relied on glycolysis other than OXPHOS for ATP production. Since E2 and its agonists suppressed HepG2 cell growth, it is possible that E2 and its agonists inhibit glycolysis. We then examined whether E2 affected energy metabolism and which ER subtype was more relevant to HepG2 metabolism (Fig 4.3). qPCR results showed that E2, PPT and DPN all reduced 6Pfk and Pk mRNA levels, but not mRNA level of Cox6b, an OXPHOS enzyme, which indicated that E2 could suppress glycolysis via either ER-α or ER-β. Glut2 expression and glycogen breakdown (indicated by Gypl expression) were enhanced by E2 treatment, and gluconeogenesis (indicated by Pepck expression) was enhanced by PPT treatment, which indicated that increased glucose uptake and production by E2 via ER-α activation were also achieved by other pathways, such as the pentose phosphorylation pathway, besides by glycolysis and OXPHOS (12). As cancer cells shift from OXPHOS to glycolysis, intermediate metabolites, such as pyruvate, accumulate and drive into de novo fatty acid synthesis to meet the need of lipid-rich membrane production (27-29). Our results showed that E2 treatment, ER-α and ER-β activation simultaneously increased lipogenesis (indicated by increased expression of Acc) and lipolysis (indicated by increased expression of Pparg and Pgc-1α), which leaves estrogen’s effects on lipid metabolism unclear. Thus we performed extensive gene analysis using RNA-seq.

RNA-seq results unveiled different genes affected by E2 and its receptor agonists. A set of target genes were analyzed. Two estrogen receptors demonstrated both unique and shared 93 gene regulation, including down-regulating cell cycle genes and changing different metabolic genes (Fig 4.4, Fig 4.5 and Table 4.1). One interesting finding was that Socs3 expression level was increased by E2 through ER-β activation, which was consistent with our previous findings (Chapter 3). These data implied that inhibition of HepG2 growth by E2 was achieved by inhibition of cell metabolism and cell proliferation, and induction of cell apoptosis by ER-α and ER-β.

Non-proliferating normal cells generate most of their cellular ATP via OXPHOS in mitochondria, and a low cytosolic ADP/ATP ratio is essential to inhibit glycolysis. By contrast, proliferating cancer cells are characterized by enhanced aerobic glycolysis and suppression of mitochondrial metabolism, leading to lower production of ATP and hence higher cytosolic ADP/ATP ratios that favor enhanced glycolysis. HPLC data showed that E2, PPT and DPN did not affect ADP/ATP ratios (Fig 4.6A), suggesting that lipid catabolism induced by E2 might also contribute to ATP production and accumulation. HPLC data also showed that E2, PPT, and DPN treatments had significantly different effects on 12 major metabolites in glucose and lipid metabolism, and many other metabolic pathways, such as vitamin B6 metabolism, primary bile acid biosynthesis, and amino acid metabolism (Fig 4.6).

In summary, one major conclusion from RNA-seq is that cell cycle, cell proliferation, apoptosis, and cell metabolism of HepG2 are changed by E2 through both ER-α and ER-β activation. Cell metabolism, including glucose metabolism (uptake, glycolysis and OXPHOS) and and fatty acid metabolism (lipogenesis and lipolysis), plays a fundamental role in cancer cell survival. Certain metabolic genes and products identified in this study could be the next generation of tumor treatment targets. Further research is required to recognize the ultimate functions of metabolic pathways as unexpected drivers of progression of human normal cells toward malignancy.

94

References

1. El-Serag HB. Hepatocellular carcinoma. The New England journal of medicine. 2011;365(12):1118-27. Epub 2011/10/14. 2. Warburg O. On respiratory impairment in cancer cells. Science. 1956;124(3215):269-70. Epub 1956/08/10. 3. Weinhouse S. On respiratory impairment in cancer cells. Science. 1956;124(3215):267-9. Epub 1956/08/10. 4. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell. 2011;144(5):646-74. Epub 2011/03/08. 5. Pfeiffer T, Schuster S, Bonhoeffer S. Cooperation and competition in the evolution of ATP-producing pathways. Science. 2001;292(5516):504-7. Epub 2001/04/03. 6. Traverso N, Ricciarelli R, Nitti M, Marengo B, Furfaro AL, Pronzato MA, et al. Role of glutathione in cancer progression and chemoresistance. Oxidative medicine and cellular longevity. 2013;2013:972913. Epub 2013/06/15. 7. Riganti C, Gazzano E, Polimeni M, Aldieri E, Ghigo D. The pentose phosphate pathway: an antioxidant defense and a crossroad in tumor cell fate. Free radical biology & medicine. 2012;53(3):421-36. Epub 2012/05/15. 8. Zu XL, Guppy M. Cancer metabolism: facts, fantasy, and fiction. Biochemical and biophysical research communications. 2004;313(3):459-65. Epub 2003/12/31. 9. Suganuma K, Miwa H, Imai N, Shikami M, Gotou M, Goto M, et al. Energy metabolism of leukemia cells: glycolysis versus oxidative phosphorylation. Leukemia & lymphoma. 2010;51(11):2112-9. Epub 2010/09/24. 10. Weber G, Allard C, De Lamirande G, Cantero A. Liver glucose-6-phosphatase activity and intracellular distribution after cortisone administration. Endocrinology. 1956;58(1):40-50. Epub 1956/01/01. 11. Harding JW, Jr., Pyeritz EA, Morris HP, White HB, 3rd. Proportional activities of glycerol kinase and glycerol 3-phosphate dehydrogenase in rat hepatomas. The Biochemical journal. 1975;148(3):545-50. Epub 1975/06/01. 12. Wang B, Hsu SH, Frankel W, Ghoshal K, Jacob ST. Stat3-mediated activation of microRNA-23a suppresses gluconeogenesis in hepatocellular carcinoma by down-regulating glucose-6-phosphatase and peroxisome proliferator-activated receptor gamma, coactivator 1 alpha. Hepatology. 2012;56(1):186-97. Epub 2012/02/10. 13. Weber G, Cantero A. Glucose-6-phosphatase activity in normal, pre-cancerous, and neoplastic tissues. Cancer research. 1955;15(2):105-8. Epub 1955/02/01. 14. Weber G, Morris HP. Comparative Biochemistry of Hepatomas. Iii. Carbohydrate Enzymes in Liver Tumors of Different Growth Rates. Cancer research. 1963;23:987-94. Epub 1963/08/01. 15. Shonk CE, Morris HP, Boxer GE. Patterns of Glycolytic Enzymes in Rat Liver and Hepatoma. Cancer research. 1965;25:671-6. Epub 1965/06/01. 16. Calvisi DF, Wang C, Ho C, Ladu S, Lee SA, Mattu S, et al. Increased lipogenesis, induced by AKT-mTORC1-RPS6 signaling, promotes development of human hepatocellular carcinoma. Gastroenterology. 2011;140(3):1071-83. Epub 2010/12/15. 17. Pike Winer LS, Wu M. Rapid analysis of glycolytic and oxidative substrate flux of cancer cells in a microplate. PloS one. 2014;9(10):e109916. Epub 2014/11/02.

95

18. Ding Y, Liu Z, Desai S, Zhao Y, Liu H, Pannell LK, et al. Receptor tyrosine kinase ErbB2 translocates into mitochondria and regulates cellular metabolism. Nature communications. 2012;3:1271. Epub 2012/12/13. 19. Shen M, Shi H. Estradiol and Estrogen Receptor Agonists Oppose Oncogenic Actions of Leptin in HepG2 Cells. PloS one. 2016;11(3):e0151455. Epub 2016/03/18. 20. Villanueva A, Alsinet C, Yanger K, Hoshida Y, Zong Y, Toffanin S, et al. Notch signaling is activated in human hepatocellular carcinoma and induces tumor formation in mice. Gastroenterology. 2012;143(6):1660-9 e7. Epub 2012/09/15. 21. Hu YJ, Li HY, Qiu KJ, Li DC, Zhou JH, Hu YH, et al. Downregulation of Notch1 inhibits the invasion of human hepatocellular carcinoma HepG2 and MHCC97H cells through the regulation of PTEN and FAK. International journal of molecular medicine. 2014;34(4):1081- 6. Epub 2014/08/12. 22. Borbath I, Horsmans Y. The Role of PPARgamma in Hepatocellular Carcinoma. PPAR research. 2008;2008:209520. Epub 2008/05/30. 23. Carow B, Rottenberg ME. SOCS3, a Major Regulator of Infection and Inflammation. Frontiers in immunology. 2014;5:58. Epub 2014/03/07. 24. Sheng SL, Liu JJ, Dai YH, Sun XG, Xiong XP, Huang G. Knockdown of lactate dehydrogenase A suppresses tumor growth and metastasis of human hepatocellular carcinoma. The FEBS journal. 2012;279(20):3898-910. Epub 2012/08/18. 25. Iansante V, Choy PM, Fung SW, Liu Y, Chai JG, Dyson J, et al. PARP14 promotes the Warburg effect in hepatocellular carcinoma by inhibiting JNK1-dependent PKM2 phosphorylation and activation. Nature communications. 2015;6:7882. Epub 2015/08/11. 26. Dai W, Wang F, Lu J, Xia Y, He L, Chen K, et al. By reducing hexokinase 2, induces apoptosis in HCC cells addicted to aerobic glycolysis and inhibits tumor growth in mice. Oncotarget. 2015;6(15):13703-17. Epub 2015/05/06. 27. Jackowski S. Coordination of membrane phospholipid synthesis with the cell cycle. The Journal of biological chemistry. 1994;269(5):3858-67. Epub 1994/02/04. 28. Costello LC, Franklin RB. Tumor cell metabolism: the marriage of molecular genetics and proteomics with cellular intermediary metabolism; proceed with caution! Molecular cancer. 2006;5:59. Epub 2006/11/09. 29. Menendez JA, Lupu R. Fatty acid synthase and the lipogenic phenotype in cancer pathogenesis. Nature reviews Cancer. 2007;7(10):763-77. Epub 2007/09/21.

96

Figure legends

Figure 4.1 Effects of sodium oxamate (OX), oligomycin (OM), 2-deoxy-D-glucose (2-DG), and 17β-estradiol (E2) on HepG2 cell number in cell culture. (A) HepG2 cells with different treatments, including vehicle DMSO (1μM) as control, oxamate (5mM), oxamate (50mM), oligomycin (1µg/ml), 2-DG (1mM) and E2 (1000 nM) were evaluated after being treated for 48 hours using light microscopy (10 × magnification). (B) HepG2 cell numbers with different doses of oxamate (0-50mM). (C) HepG2 cell numbers with different doses of OM (0-1µg/ml). (D) HepG2 cell numbers with different doses of 2-DG treatments (0-10mM). Cell numbers were analyzed by one-way ANOVA analysis.

*: Significantly different comparing to control groups (p < 0.05).

Figure 4.2 Effects of different doses of sodium oxamate (OX), oligomycin (OM), 2-deoxy-D- glucose (2-DG) on HepG2 cytotoxicity (A), viability (B) and apoptosis (C) after 24 hours treatment. Light signals were analyzed by one-way ANOVA analysis.

*: Significantly different comparing to control groups (p < 0.05).

Figure 4.3 Effects of E2 (1 μM), PPT (1 μM) and DPN (1 μM) on gene expression levels involved in energy metabolism. Gene expression levels of control group were set at 100%. HepG2 gene expression levels of glucose transporter 2 (Glut2), Phosphofructokinase-6 (6Pfk), Pyruvate kinase (Pk), Cytochrome c oxidase (Cox6b), Glycogen phosphrylase (Pygl), Phosphoenolpyruvate carboxykinase (Pepck), Glycogen synthase (Gys), Fatty acid synthase (Fasn), Acetyl co-enzyme carboxylase (Acc), Transcription factor for lipogenic enzymes sterol- regulatory binding protein-1c (Srebp-1c), Peroxisome proliferator-activated receptor gamma (Pparg) and Peroxisome proliferator-activated receptor gamma coactivator 1 a (Pgc1α) were analyzed by one-way ANOVA.

*: Significantly different comparing to control groups (p < 0.05); †: Significantly different comparing to E2 groups (p < 0.05), ‡: Significantly different comparing to PPT groups (p < 0.05).

97

Figure 4.4 RNA sequencing data shows different gene expression upon E2 (1 μM), PPT (1 μM) and DPN (1 μM) treatments on HepG2 cells. The Venn diagram shows unique and shared gene groups of different treatments.

Figure 4.5 Evaluation of target genes identified by RNA sequencing. Relative mRNA expression levels of Socs3, Elovl3, Cdc20, Notch1, Pparg, and Il6r were determined and normalized to the housekeeping gene β-actin. The results were analyzed by one-way ANOVA.

*: Significantly different comparing to control groups (p < 0.05); †: Significantly different comparing to E2 groups (p < 0.05), ‡: Significantly different comparing to PPT groups (p < 0.05).

Figure 4.6 Metabolic profiles

ADP/ATP ratio among different treatments were analyzed by one-way ANOVA (Fig 4.6A) 11 metabolites involved in glucose and lipid metabolic pathways of four groups were analyzed by principal component analysis (PCA) score plots (Fig 4.6B) and loading plots (Fig 4.6C). Total 255 metabolites of four groups of metabolites were analyzed by principal component analysis (PCA) score plots (Fig 4.6D) and loading plots (Fig 4.6E). Major metabolic pathways were impacted by different treatments during this study, the x-axis is the metabolic pathway impact value, while the y-axis is statistical significance (represented by p-value) of the impacted pathway from paired groups. The dot size corresponds to the x-axis value and the dot color corresponds to the y-axis value. A: Tryptophan metabolism; B: Vitamin B6 metabolism; C: Butanoate metabolism; D: Fructose and mannose metabolism; E: Pyruvate metabolism; F: Glycolysis or Gluconeogenesis; G: Glycine, serine and threonine metabolism; H: Fatty acid metabolism; I: Nicotinate and nicotinamide metabolism; J: TCA cycle; K: Riboflavin metabolism; L: Ether lipid metabolism; M: Primary bile acid biosynthesis; N: Glycerophospholipid metabolism; O: Arginine and proline metabolism; P: beta-Alanine metabolism; Q: Pantothenate and CoA biosynthesis. Control VS E2 (Fig 4.6F); Control VS PPT (Fig 4.6G); Control VS DPN (Fig 4.6H); E2 VS PPT (Fig 4.6I); E2 VS DPN (Fig 4.6J); PPT VS DPN (Fig 4.6K).

98

Table 4.1 Top biological functions identified by Ingenuity Pathway Analysis regulated by E2, PPT and DPN

Table 4.1A Top biological functions identified by Ingenuity Pathway Analysis up-regulated by E2 treatment

Cellular function p-value #Genes Example genes log2 fold change Signal transduction 5.5E-5 129 RAS p21 protein activator 3(RASA3) Inf Caspase recruitment domain family member Inf 11(CARD11) Scavenger receptor class F member 2.33 1(SCARF1) Mitogen-activated protein kinase 7(MAPK7) 1.30 Notch1 1.86 Response to hypoxia 4.1E-6 33 Uncoupling protein 2(UCP2) 1.297 ATPase Na+/K+ transporting subunit beta 1.41 1(ATP1B1) Dipeptidyl peptidase 4(DPP4) 1.22 Erythropoietin(EPO) 3.28 Negative regulation of 3.2E-3 48 B-cell CLL/lymphoma 6(BCL6) 0.72 cell proliferation Notch 1(NOTCH1) 1.86 Ras association domain family member 2.34 5(RASSF5) Regulation of 6.0E-5 35 B-cell CLL/lymphoma 3(BCL3) 1.87 apoptotic process B-cell CLL/lymphoma 6(BCL6) 0.72 BCL2 like 2(BCL2L2) 1.17 N-myc downstream regulated 1(NDRG1) 4.06 RELT tumor necrosis factor receptor(RELT) 1.10 Cell adhesion 2.1E-5 64 Cadherin 4(CDH4) 2.4 Hepatic and glial cell adhesion 1.10 molecule(HEPACAM) Integrin subunit alpha 2(ITGA2) 1.66 Growth factor 2.5E-3 21 Epidermal growth factor receptor(EGFR) 2.95 Fibroblast growth factor 11(FGF11) 3.59 Insulin like growth factor 1 receptor(IGF1R) 2.16 Insulin signaling 1.6E-2 20 PPARG coactivator 1 alpha(PPARGC1A) 2.14 pathway Glycogen synthase 1(GYS1) 1.84 Phosphoenolpyruvate carboxykinase 1(PCK1) 3.03 Hexokinase 2(HK2) 1.13 Insulin receptor substrate 2(IRS2) 1.46 Negative regulation of 3.3e-3 15 Suppressor of cytokine signaling 3(SOCS3) 2.04 inflammatory response Kruppel like factor 4(KLF4) 3.96 Suppressor of cytokine signaling 5(SOCS5) 1.06 Glycolysis 1.8e-6 12 6-phosphofructo-2-kinase/fructose-2,6- 1.36

99

biphosphatase 1(PFKFB1) 6-phosphofructo-2-kinase/fructose-2,6- 3.77 biphosphatase 3(PFKFB3) Aldolase, fructose-bisphosphate a(ALDOA) 1.19 Hexokinase 1(HK1) 3.36 Glucose-6-phosphatase catalytic 3.61 subunit(G6PC) Phosphoglycerate kinase 1(PGK1) 3.1 Glucose metabolism 5.5E-6 13 Amylase, alpha 1B (salivary)(AMY1B) 5.7 Amylase, alpha 2A (pancreatic)(AMY2A) 4.63 Glycogen synthase 1(GYS1) 1.84 Solute carrier family 2 member 3(SLC2A3) 1.69 Lipid transport 6.9e-5 18 Apolipoprotein a4(APOA4) 1.04 Apolipoprotein c3(APOC3) 1.75 Ether lipid metabolism 1.8E-4 13 Phospholipase A2 group IIA(PLA2G2A) 5.07

Gluconeogenesis 4.7E-4 12 Pyruvate carboxylase(PC) 1.78 Phosphoglycerate kinase 1(PGK1) 1.3 Phosphoenolpyruvate carboxykinase 1(PCK1) 3.03 Cholesterol 3.9e-3 13 Ldl receptor related protein 5(LRP5) 1.16 homeostasis Cholesteryl ester transfer protein(CETP) 1.44 Scavenger receptor class b member 1.27 1(SCARB1) Lipid metabolism 3.8E-2 45 Peroxisome proliferator activated receptor 2.48 gammar(PPARG) Fatty acid desaturase 6(FADS6) 2.38 Phospholipase A2 group IIA(PLA2G2A) 5.07 Sterol regulatory element binding 1.02 transcription factor 2(SREBF2)

100

Table 4.1B Top biological functions identified by Ingenuity Pathway Analysis down-regulated by E2 treatment

Cellular function p-value #Genes Example genes log2 fold change Cell cycle 4.4E-39 127 DNA replication and sister chromatid -1.05 cohesion 1(DSCC1) E2F transcription factor 2(E2F2) -1.22 Cell division cycle 20(CDC20) -2.25 Cyclin A2(CCNA2) -1.86 DNA repair 3.4E-9 31 Chromatin assembly factor 1 subunit -1.18 B(CHAF1B) Cyclin dependent kinase 1(CDK1) -1.86 Single stranded DNA binding protein 1(SSBP1) -1.14 Glutathione 6.1E-4 9 Glutathione S-transferase alpha 1(GSTA1) -1.77 transferase activity Microsomal glutathione S-transferase -1.16 1(MGST1) NADPH 3.8E-5 25 NADPH oxidase 1(NOX1) -2.97 Glucose-6-phosphate dehydrogenase(G6PD) -1.10 NAD(P)H quinone dehydrogenase 1(NQO1) -1.12 Metabolic pathways 7.6E-5 105 UDP glucuronosyltransferase family 1 -6.00 member A1(UGT1A1) Glycerol-3-phosphate dehydrogenase -2.15 1(GPD1) Fatty acid metabolism 2.0E-3 16 ELOVL fatty acid elongase 4(ELOVL4) -6.00 Peroxisomal trans-2-enoyl-CoA -1.56 reductase(PECR) Malonyl-CoA-acyl carrier protein -1.13 transacylase(MCAT) Glycerol-3-phosphate acyltransferase, -1.70 mitochondrial(GPAM)

101

Table 4.1C Top biological functions identified by Ingenuity Pathway Analysis up-regulated by PPT treatment

Cellular function p-value #Genes Example genes log2 fold change Alternative splicing 5.0E-6 342 ATP binding cassette subfamily A member 1.218 5(ABCA5) BCL2 like 2(BCL2L2) 1.06 LDL receptor related protein 10(LRP10) 1.28 Membrane 6.1E-6 256 5-hydroxytryptamine receptor 1D (HTR1D) 1.67 DnaJ heat shock protein family (Hsp40) 1.35 member B9(DNAJB9) Transport 9.5E-6 87 Aquaporin 8(AQP8) 2.27 Solute carrier family 38 member 6(SLC38A6) 1.46 growth factor activity 6.7E-3 12 VGF nerve growth factor inducible(VGF) 3.39 Oxidative stress induced growth inhibitor 1.07 1(OSGIN1) Platelet derived growth factor subunit 1.14 A(PDGFA) oxidoreductase activity 1.7E-5 19 Aldehyde dehydrogenase 1 family member 0.94 L1(ALDH1L1) NAD(P)H quinone dehydrogenase 2(NQO2) 1.04 Metabolic pathways 2.1E-2 53 Adenosine monophosphate deaminase 2.69 3(AMPD3) Glycerol kinase(GK) 1.36 Phosphogluconate dehydrogenase(PGD) 1.13 Peroxisome proliferator activated receptor 0.91 gammar(PPARG) Fructose and mannose 1.9E-2 5 6-phosphofructo-2-kinase/fructose-2,6- 1.68 metabolism biphosphatase 1(PFKFB1) Aldo-keto reductase family 1 member 2.85 B10(AKR1B10) Phosphofructokinase, platelet(PFKP) 2.74 lipid metabolic process 6.1E-3 12 Low density lipoprotein receptor(LDLR) 1.17 Fatty acid desaturase 3(FADS3) 1.39 glutamate metabolic 5.6E-3 4 N-acetylglutamate synthase(NAGS) 2.73 process Glutamate-cysteine ligase catalytic 2.07 subunit(GCLC) Glutathione 5.8E-3 7 Glucose-6-phosphate dehydrogenase(G6PD) 1.27 metabolism Glutathione S-transferase alpha 2(GSTA2) 1.04 Phosphogluconate dehydrogenase(PGD) 1.13 Lipoprotein 2.8E-2 34 LDL receptor related protein 8(LRP8) 1.73 Toll like receptor adaptor molecule 2(TICAM2) 1.05 Gamma-aminobutyric acid type A receptor 1.27 gamma1 subunit(GABRG1) 102

Carbohydrate 4.4E-2 7 Amylase, alpha 1A (salivary)(AMY1A) 3.0 metabolism Starch binding domain 1(STBD1) 1.50

103

Table 4.1D Top biological functions identified by Ingenuity Pathway Analysis down-regulated by PPT treatment

Cellular function p-value #Genes Example genes log2 fold change Cell cycle 1.7E-46 119 Rac GTPase activating protein 1(RACGAP1) -1.9 Cell division cycle 20(CDC20) -1.53 Cyclin dependent kinase 2(CDK2) -1.26 Mitosis 5.0E-36 68 BUB1 mitotic checkpoint serine/threonine -1.66 kinase B(BUB1B) Abnormal spindle microtubule -1.39 assembly(ASPM) Centromere protein A(CENPA) -2.30 DNA replication 1.1E-17 29 DNA polymerase delta 1, catalytic -1.06 subunit(POLD1) DNA ligase 1(LIG1) -1.38 Replication factor C subunit 3(RFC3) -1.30 Signal 1.8E-9 232 TNF receptor superfamily member -2.42 19(TNFRSF19) Fc fragment of IgG binding protein(FCGBP) -1.37 metabolic process 8.6E-6 22 UDP glucuronosyltransferase family 1 -3.95 member A1(UGT1A1) Enolase superfamily member 1(ENOSF1) -1.23 Lipid metabolism 1.8E-3 31 Lipase C, hepatic type(LIPC) -1.33 Apolipoprotein A4(APOA4) -2.04 LDL receptor related protein 2(LRP2) -2.10 Diacylglycerol O-acyltransferase 2(DGAT2) -1.93 Acyl-CoA synthetase medium-chain family -2.04 member 1(ACSM1)

104

Table 4.1E Top biological functions identified by Ingenuity Pathway Analysis up-regulated by DPN treatment

Cellular function p-value #Genes Example genes log2 fold change Membrane 1.2E-8 277 FIC domain containing(FICD) 2.13 LDL receptor related protein 8(LRP8) 1.08 A-kinase anchoring protein 12(AKAP12) 3.21 Signal 1.3E-5 160 G protein-coupled receptor 155(GPR155) 1.05 Interleukin 6 receptor(IL6R) 2.57 Suppressor of cytokine signaling 3 (SOCS3) 2.32 Transport 9.2E-4 80 Potassium voltage-gated channel modifier 1.88 subfamily F member 1(KCNF1) Calcium voltage-gated channel subunit 2.10 alpha1 H(CACNA1H) Sodium voltage-gated channel alpha subunit 4.34 5(SCN5A) oxidation-reduction 3.9E-3 31 Aldehyde dehydrogenase 1 family member 1.64 process L2(ALDH1L2) Fatty acid desaturase 3(FADS3) 1.08 Lactate dehydrogenase B(LDHB) FoxO signaling 6.3E-3 12 Tumor necrosis factor superfamily member 1.83 pathway 10(TNFSF10) Mitogen-activated protein kinase 2.28 11(MAPK11) Kruppel like factor 2(KLF2) 1.90 Glucagon signaling 2.0E-2 9 PPARG coactivator 1 alpha(PPARGC1A) 1.08 pathway 6-phosphofructo-2-kinase/fructose-2,6- 1.21 biphosphatase 1(PFKFB1) lipid metabolic process 2.1E-2 11 Peroxisome proliferator activated receptor 1.12 gamma(PPARG) Glycerophosphocholine phosphodiesterase 1.82 1(GPCPD1) LDL receptor related protein 10(LRP10) 1.75

105

Table 4.1F Top biological functions identified by Ingenuity Pathway Analysis down-regulated by DPN treatment

Cellular function p-value #Genes Example genes log2 fold change Cell cycle 1.1E-71 132 Cyclin B2(CCNB2) -2.49 G2 and S-phase expressed 1(GTSE1) -3.08 BCL2 binding component 3(BBC3) -1.36 Notch1 -2.99 Cell division cycle 20 (CDC20) -2.85 DNA replication 1.7E-25 33 DNA ligase 1(LIG1) -1.64 DNA polymerase delta 1, catalytic -1.61 subunit(POLD1) Cell division cycle 45(CDC45) -1.95 Lipid metabolism 1.5E-2 7 Carboxyl ester lipase(CEL) -1.52 Acyl-CoA synthetase medium-chain family -1.73 member 1(ACSM1) Acyl-CoA synthetase long-chain family -1.20 member 6(ACSL6) LDL receptor related protein 2(LRP2) -2.30 fatty acid metabolism 6.5E-2 3 Diacylglycerol O-acyltransferase 1(DGAT1) -1.12 Bile acid-CoA:amino acid N- acyltransferase(BAAT) -1.40 lipoprotein lipase 3.9E-3 2 Apolipoprotein A5(APOA5) -1.05 activator activity Apolipoprotein H(APOH) -1.33

106

Table 4.2 qPCR primers for RNA sequencing check

Genes Forward Reverse Socs3 5’- GGAGTTCCTGGACCAGTACG-3′ 5’-TTCTTGTGCTTGTGCCATGT-3′ Elovl3 5'-GACATGAGGCCCTTTTTCGA-3' 5'-CCCACAGCGATGAGAACCA-3' Cdc20 5’-CTACAGCCAAAAGGCCACTC-3’ 5’-GATCCAGGCCACAGAGGATA-3’ Notch1 5’-CAATGTGGATGCCGCAGTTGTG-3’ 5’-CAGCACCTTGGCGGTCTCGTA-3’ Pparg 5’-TCTGGCCCACCAACTTTGGG-3′ 5’-CTTCACAAGCATGAACTCCA-3′ Il6r 5’-TAGCCGCCCCACACAGACAG-3’ 5’-GGCTGGCATTTGTGGTTGGG-3’

107

Table 4.3 qPCR primers for metabolic enzymes

Genes Forward Reverse Gys2 5’- GCCAGACACCTGACATTAAG -3’ 5’- CTCCACTTCATCTTCCACATC -3’ Pygl 5’- CCTGTGATGAGGCCATTTAC -3’ 5’- GTATCCATAGGCTGCAAGTC -3’ 6Pfk 5’- CGAGGACCCTTTCAACATC -3’ 5’- CGAGGACCCTTTCAACATC -3’ Pk 5’- TCGTCTTTGCCTCCTTTG -3’ 5’- CTCACCTCCAGGATTTCATC -3’ Cox6b 5’- CTCAACGTGTTCCTCAAGTC -3’ 5’- ATGGAGGACAGAGGAAAGG -3’ Glut2 5′- AGTTAGATGAGGAAGTCAAAGCAA -3’ 5′- TAGGCTGTCGGTAGCTGG -3′ Fas 5’- GAAACTGCAGGAGCTGTC -3’ 5’- CACGGAGTTGAGGCGGAT -3’ Acc 5’- GCTGCTCGGATCACTAGTGAA -3’ 5’- GCTGCTCGGATCACTAGTGAA -3’ Srebp-1c 5’- CTTTGCCCACCCTGGTGAGT -3’ 5’- CTTTGCCCACCCTGGTGAGT -3’ Pgc-1α 5’- GACGACGAAGCAGACAAG -3’ 5’- CCAAGGGTAGCTCAGTTTATC -3’ Pepck 5’- CCAGGCAGTGAGGGAGTTTCT -3’ 5’- ACTGTGTCTCTTTGCTCTTGG -3’ Ppar-g 5’- GAAATGACCATGGTTGAC -3’ 5’- CCGCTAGTACAAGTCCTTGTA -3’ β-actin 5’- AGAGCTACGAGCTGCCTGAC -3’ 5’- AGCACTGTGTTGGCGTACAG -3’

108

Figure 4.1 Effects of sodium oxamate (OX), oligomycin (OM), 2-deoxy-D-glucose (2-DG), and 17β-estradiol (E2) on HepG2 cell number in cell culture

Fig 4.1A

Ctr Oxamate 5mM Oxamate 50mM

Oligomycin 1µg/ml 2-DG 10mM E2 1mM

Fig 4.1B Fig 4.1C

15000 15000

* 10000 10000 * * 5000 5000

Cell number Cell Cell number Cell

0 0

Control OM 1 OM 0.1 OM 0.5 Oxamate 5 Oxmate 10 Oxmate 20 Oxmate 50 Control Fig 4.1D

15000

10000

5000

Cell number Cell * * * 0

2-DG 1 2-DG 5 Control 2-DG 10 109

Figure 4.2 Effects of different doses of sodium oxamate (OX), oligomycin (OM), 2-deoxy-D- glucose (2-DG) on HepG2 cytotoxicity, viability and apoptosis

Fig 4.2A

20000 * * 15000 * * * 10000 *

Cytotoxicity 5000

(Fluorenscence, RFU) (Fluorenscence, 0

C OX-5 DG-1 DG-5 OX-10OX-20OX-50 DG-10 OM-1 OM-0.1OM-0.5 Fig 4.2B

400

300 * 200

Viability * 100 * (Fluoresence, RFU) (Fluoresence, * * * 0

C OX-5 DG-1 DG-5 OX-10OX-20OX-50 DG-10 OM-1 OM-0.1OM-0.5 Fig 4.2C

20000 * *

15000 * * 10000 *

Apoptosis 5000

(Luminiscence, RLU) (Luminiscence, 0

C OX-5 DG-1 DG-5 OX-10OX-20OX-50 DG-10 OM-1 OM-0.1OM-0.5

110

Figure 4.3 Effects of E2, PPT and DPN on gene expression levels involved in energy metabolism

Fig 4.3A Fig 4.3B

6Pfk 150 Glut2 200 * 100 † †

150 † * * -actin

-actin  100 50

6pfk/ * (% of Control)

Glut2/

(% of (% of Control) 50

0 0

Ctr E2 Ctr E2 PPT PPT DPN DPN Fig 4.3C Fig 4.3D

Pk

150 Cox6b 150 † 100 100

-actin -actin  * ‡ 

Pk/ 50

(% of Control) * 50

Cox6b/ (% of Control)

0 0

Ctr E2 Ctr E2 PPT DPN PPT DPN Fig 4.3E Fig 4.3F

Gys 150 Gypl 200 *

100 150

-actin †

 † -actin

 100 50

Gys/

(% of Control) Gypl/

(% of Control) 50

0 0

Ctr E2 Ctr E2 PPT DPN PPT DPN

111

Fig 4.3G Fig 4.3H

Pepck Pgc-1a 200 500 † * 400 150

‡ 300 † -actin -actin *  †  100

200 Pgc1a/

Pepck/ (% of Control) (% of Control) 50 100

0 0

Ctr E2 Ctr E2 PPT DPN PPT DPN

Fig 4.3I Fig 4.3J

Srebp-1c Pparg 250 500 * 200 400

-actin 150 †  300 * -actin †  * 100

200 (% of Control)

Pparg/ Srebp1c/ (% of Control) 50 100

0 0

E2 Ctr E2 Ctr PPT PPT DPN DPN

Fig 4.3K Fig 4.3L

Fas Acc 200 200 * 150 * 150 † ‡

-actin

100 -actin   100

Fas/ Acc/

(% of Control) 50 (% of Control) 50

0 0

Ctr E2 Ctr E2 PPT DPN PPT DPN

112

Figure 4.4 RNA sequencing data shows different gene expression upon E2, PPT and DPN treatments on HepG2 cells

Figure 4.4A up-regulated genes

Figure 4.4B down-regulated genes

113

Figure 4.5 Evaluation of target genes identified by RNA sequencing

Fig 4.5A Fig 4.5B

Pparg Notch1 1000 150

800 *

100 600 -actin

-actin

  400 † ‡ 50 *

Pparg/ † (% of Control)

(% of Control)

Notch1/ † * ‡ 200

0 0

Ctr E2 Ctr E2 PPT DPN PPT DPN Fig 4.5C Fig 4.5D

Cdc20 Socs3 150 800 * ‡ ‡ 600 *

100 -actin -actin   400 †

50 * Cdc20/ Socs3/ (% of Control) (% of Control) 200 †

0 0

Ctr E2 Ctr E2 PPT DPN PPT DPN Fig 4.5E Fig 4.5F

Elovl3 Il6r 150 800 † † *

600 * 100 -actin

-actin 400  50

Il6r/

Elovl3/ (% of Control)

(% of Control) 200

0 0

E2 Ctr E2 Ctr PPT PPT DPN DPN

114

Figure 4.6 HPLC-MS Metabolic profiles

Fig 4.6A

150

100

ADP/ATP 50

0

Ctr E2 PPT DPN Fig 4.6B Fig 4.6C

Fig 4.6D Fig 4.6E

115

Fig 4.6F E2 VS CTR

H

L

D

M B E

C G I F K Q

Fig 4.6G PPT VS CTR

B H

M I E F C D

J K

116

Fig 4.6H DPN VS CTR

B L E G F M Q J K

Fig 4.6I E2 vs PPT

B D

I Q

E H G F M P J K

117

Fig 4.6J E2 VS DPN

A B M E F O J H P K

Fig 4.6K PPT VS DPN

H O G L A P

M E B J K D

118

Chapter 5. Conclusion and perspectives Estrogens play important roles in normal physiological functions and cancers, which require a more in-depth knowledge of the effects of E2 and its mechanisms of action.

Our first study has demonstrated that E2, acting via ERα, regulates glucose and lipid homeostasis mainly by modulating hepatic insulin sensitivity, which can be due to the upregulation of lipolysis genes, glycogenlysis genes and glycolysis genes, as well as downregulation of lipogenic genes. Although estrogens mainly activate nuclear estrogen receptors, membrane estrogen receptors (mERs), including membrane-associated ERα (mERα) and ERβ (mERβ), G-protein coupled estrogen receptor (GPER) (1, 2), are all cell surface receptors which rapidly alter cell signaling via modulation of intracellular signaling cascades, including AMPK, MAPK, PI3 K, PKC, and Ca2+ mobilization (3, 4). Female GPER-deficient mice had hyperglycemia and impaired glucose tolerance and reduced serum IGF-I levels, which were associated with decreased insulin expression and release in vivo and in vitro in isolated pancreatic islets (5, 6). Interestingly, a recent study using specific plasma membrane ER-훼 knockout has demonstrated that it is the membrane-localized ER-훼, but not nuclear ER-훼, that is responsible for protection from hyperlipidemia by decreasing expressions of many hepatic genes involved in lipid synthesis, at least in female mice with OVX (7). In our study, we treated the OVX mice with E2, PPT and DPN, but we were not sure about whether mERs play any significant roles in the results. Future studies will be focused on exploring the effects of GPER agonist and specific mERs agonists on metabolic homeostasis and gene expressions. Also, we used an OVX mouse model, which caused deficiency of both estrogens and progesterone. Interestingly, a recent paper pointed out that disrupted liver glucose homeostasis following OVX is not merely caused by deficiency of endogenous E2 (8), but could be caused by deficiency of other ovarian hormones such as progesterone, which inspires us that in future studies we should also add progesterone replacement alone and with E2.

Secondly, we identified that E2 suppressed leptin-induced HepG2 growth mainly through ER-β activation and p38/MAPK signal pathway. A group of scientists found that liver tumor growth in ovariectomized (OVX) female mice was increased compared with that in control female mice and that the addition of 17β-estradiol (E2) reduced tumor growth in the OVX mice (9). They didn’t find a difference in liver tumor size between control and castrated male mice;

119 however, E2 reduced tumor size in the castrated mice (9). Estrogens suppression of liver tumor mainly by suppressing inflammation and IL-6-dependent STAT3 activation is considered a key event in inflammation-induced liver cancer. The anti-inflammation effect of estrogen is well documented (10-12). Interestingly, membrane ERα had unique expression pattern that ERα-66 had high expression in normal liver, but had low expression in HCC; while ERα-36 had low expression in normal liver, but had high expression in HCC (13). GPER also has demonstrated important roles in cancers that GPER stimulates caspase dependent and independent programmed cell death (14, 15), and suppresses cancer cell proliferation via blocking tubulin polymerization and disrupting spindle formation of ovarian cancer cells (16), and inhibiting cell cycle progression in G2/M phase and thus arresting cells at G2 phase of mitosis of ovarian cancer cells (17) and prostate cancer cells (18). In our study, we only used HepG2 cell lines. In order to get a more universal and convincing conclusion about E2’s effects on liver cancer, we will apply more HCC cell lines, such as Huh-7 and Hep3B as well as normal liver cells. We applied PPT and DPN to specifically activate ER-α and ER-β and siRNA to specifically block ER-α and ER-β; however, we should also treat the cells with specific chemical antagonists to confirm siRNA results. Moreover, we should identify mERs, GPER expression levels before and after leptin and E2 treatments, and investigate GPER’s functions on HepG2 growth by applying specific GPER agonists and antagonists.

In our third study, we demonstrated HepG2 cell was more dependent on glycolysis and tested different roles of ERs on HepG2 metabolism. Cancer cells have different metabolic processes from normal cells. For example, fatty acid synthase (FASN) gene expression is usually low in normal tissues (except liver and adipose tissue) (19), but it is up-regulated in many cancers, such as breast cancer and prostate cancer (20). Interestingly, a recent paper pointed out that FASN regulated E2/ERα signaling in breast cancer and this finding may represent a promising strategy for anticancer treatment involving a new generation of FASN inhibitors (21). In our results, we didn’t find E2, PPT or DPN changed FASN expression level, however, according to the paper just mentioned, we will apply FASN specific inhibitor to HepG2 to test how block of FASN can affect HepG2 growth, estrogen receptor expressions and signal pathways. Although, we applied HPLC-mass spectrum to detect more than 200 metabolites of HepG2, we still left some critical metabolites, such as NADH and FADH that we failed to get

120

NAD/NADH or FAD/FADH ratios. Thus in the future, we will apply NAD/NADH and FAD/FADH assay kits.

In summary, my overall dissertation study found that the estrogens regulate liver metabolism, liver cancer proliferation/apoptosis and liver cancer metabolism. ER-α activation mainly regulates both normal and cancer hepatocytes metabolism, while ER-β activation mainly regulates liver cancer proliferation and apoptosis. In the future, more study should be done to further explore the genomic and non-genomic effects of ER-α and ER-β activation as well as other membrane ERs.

121

References 1. Soltysik K, Czekaj P. Membrane estrogen receptors - is it an alternative way of estrogen action? Journal of physiology and pharmacology : an official journal of the Polish Physiological Society. 2013;64(2):129-42. Epub 2013/06/13. 2. Prabhushankar R, Krueger C, Manrique C. Membrane estrogen receptors: their role in blood pressure regulation and cardiovascular disease. Current hypertension reports. 2014;16(1):408. Epub 2013/12/18. 3. Prossnitz ER, Barton M. Estrogen biology: new insights into GPER function and clinical opportunities. Molecular and cellular endocrinology. 2014;389(1-2):71-83. Epub 2014/02/18. 4. Sharma G, Mauvais-Jarvis F, Prossnitz ER. Roles of G protein-coupled estrogen receptor GPER in metabolic regulation. The Journal of steroid biochemistry and molecular biology. 2017. Epub 2017/02/23. 5. Martensson UE, Salehi SA, Windahl S, Gomez MF, Sward K, Daszkiewicz-Nilsson J, et al. Deletion of the G protein-coupled receptor 30 impairs glucose tolerance, reduces bone growth, increases blood pressure, and eliminates estradiol-stimulated insulin release in female mice. Endocrinology. 2009;150(2):687-98. Epub 2008/10/11. 6. Sharma G, Hu C, Brigman JL, Zhu G, Hathaway HJ, Prossnitz ER. GPER deficiency in male mice results in insulin resistance, dyslipidemia, and a proinflammatory state. Endocrinology. 2013;154(11):4136-45. Epub 2013/08/24. 7. Pedram A, Razandi M, O'Mahony F, Harvey H, Harvey BJ, Levin ER. Estrogen reduces lipid content in the liver exclusively from membrane receptor signaling. Science signaling. 2013;6(276):ra36. Epub 2013/05/23. 8. Nigro M, Santos AT, Barthem CS, Louzada RA, Fortunato RS, Ketzer LA, et al. A change in liver metabolism but not in brown adipose tissue thermogenesis is an early event in ovariectomy-induced obesity in rats. Endocrinology. 2014;155(8):2881-91. Epub 2014/06/11. 9. Yang W, Lu Y, Xu Y, Xu L, Zheng W, Wu Y, et al. Estrogen represses hepatocellular carcinoma (HCC) growth via inhibiting alternative activation of tumor-associated macrophages (TAMs). The Journal of biological chemistry. 2012;287(48):40140-9. Epub 2012/08/22. 10. Rogers A, Eastell R. The effect of 17beta-estradiol on production of cytokines in cultures of peripheral blood. Bone. 2001;29(1):30-4. Epub 2001/07/27. 11. Polan ML, Loukides J, Nelson P, Carding S, Diamond M, Walsh A, et al. Progesterone and estradiol modulate interleukin-1 beta messenger ribonucleic acid levels in cultured human peripheral monocytes. The Journal of clinical endocrinology and metabolism. 1989;69(6):1200-6. Epub 1989/12/01. 12. Kawasaki T, Ushiyama T, Inoue K, Hukuda S. Effects of estrogen on interleukin-6 production in rheumatoid fibroblast-like synoviocytes. Clinical and experimental rheumatology. 2000;18(6):743-5. Epub 2001/01/04. 13. Miceli V, Cocciadiferro L, Fregapane M, Zarcone M, Montalto G, Polito LM, et al. Expression of wild-type and variant estrogen receptor alpha in liver carcinogenesis and tumor progression. Omics : a journal of integrative biology. 2011;15(5):313-7. Epub 2011/02/26. 14. Cregan SP, Dawson VL, Slack RS. Role of AIF in caspase-dependent and caspase- independent cell death. Oncogene. 2004;23(16):2785-96. Epub 2004/04/13. 15. Filardo EJ, Thomas P. Minireview: G protein-coupled estrogen receptor-1, GPER-1: its mechanism of action and role in female reproductive cancer, renal and vascular physiology. Endocrinology. 2012;153(7):2953-62. Epub 2012/04/13.

122

16. Wang C, Lv X, He C, Hua G, Tsai MY, Davis JS. The G-protein-coupled estrogen receptor agonist G-1 suppresses proliferation of ovarian cancer cells by blocking tubulin polymerization. Cell death & disease. 2013;4:e869. Epub 2013/10/19. 17. Ignatov T, Modl S, Thulig M, Weissenborn C, Treeck O, Ortmann O, et al. GPER-1 acts as a tumor suppressor in ovarian cancer. Journal of ovarian research. 2013;6(1):51. Epub 2013/07/16. 18. Chan QK, Lam HM, Ng CF, Lee AY, Chan ES, Ng HK, et al. Activation of GPR30 inhibits the growth of prostate cancer cells through sustained activation of Erk1/2, c-jun/c-fos- dependent upregulation of p21, and induction of G(2) cell-cycle arrest. Cell death and differentiation. 2010;17(9):1511-23. Epub 2010/03/06. 19. Menendez JA, Lupu R. Fatty acid synthase and the lipogenic phenotype in cancer pathogenesis. Nature reviews Cancer. 2007;7(10):763-77. Epub 2007/09/21. 20. Flavin R, Peluso S, Nguyen PL, Loda M. Fatty acid synthase as a potential therapeutic target in cancer. Future Oncol. 2010;6(4):551-62. Epub 2010/04/09. 21. Menendez JA, Lupu R. Fatty acid synthase regulates estrogen receptor-alpha signaling in breast cancer cells. Oncogenesis. 2017;6(2):e299. Epub 2017/02/28.

123