Iron Chelation for Treating Obesity and Components of the Metabolic Syndrome

Home , Cellular adaptation

Iron Chelation for Treating Obesity and

Components of the Metabolic

Syndrome

Kenneth Wai Kheong Ho

A submission to the University of New South Wales in

candidature for the degree of Doctor of Philosophy

Diabetes and Transcription Factors Program

Department of Immunology and Inflammation

Garvan Institute of Medical Research

Darlinghurst, Sydney, Australia

August 2011

PLEASE TYPE THE UNIVERSITY OF NEW SOUTH WALES Thesis/Dissertation Sheet

Surname or Family name: HO

First name: KENNETH Other name/s: WAI KHEONG

Abbreviation for degree as given in the University calendar: PhD

School: St Vincent’s Clinical School Faculty: Medicine

Title: Iron Chelation for Treating Obesity and Components of the Metabolic Syndrome

Abstract 350 words maximum: (PLEASE TYPE)

Epidemiological studies suggest a link between iron stores and components of the metabolic syndrome. Human subjects with iron overload syndromes are more prone to obesity, diabetes and metabolic syndrome. Excess iron can generate Reactive Oxygen Species

(ROS) resulting in insulin resistance and cell apoptosis. Iron chelation reduces diabetes in iron overload syndromes but effects on obesity and related metabolic variables are unknown in non-overloaded individuals. Hypoxia Inducible Factor- 1-alpha (HIF-1α) is a transcription factor up-regulated by hypoxia, which is also increased by iron chelation in-vitro. HIF-1α impacts lipid and whole body metabolism and cellular survival by improving glucose uptake, facilitating anaerobic glycolysis, and attenuating ROS.

Hypothesis: That iron chelation may treat or prevent components of the metabolic syndrome via HIF-1 induction. This thesis investigated effects of iron chelators on murine models of obesity and diabetes: C57Bl/6 mice on high fat diet (HFD), ob/ob mice on chow diet, C57Bl/6 mice on chow, and Non-Obese Diabetic (NOD) mice. Two iron chelators were studied: Deferoxamine (DFO) as an intra-peritoneal injection weekly and continuous oral Deferasirox (DFS) mixed with diet. DFS-treatment had important metabolic benefits, including reduction of weight gain in both male and female mice, from as early as 1 week, lasting throughout the period of study (up to 9 months). Other benefits included lower plasma lipids, reduction of hepatic steatosis and improvement in insulin resistance. There was also improved beta cell function, evidenced by relatively preserved first phase insulin secretion and improved glucose tolerance, particularly after prolonged high fat exposure. DFS treatment reduced hepatic iron and increased HIF-1α, associated with down-regulation of hepatic lipogene expression and improved insulin signalling gene expression. Importantly, DFS did not impair appetite and had no apparent toxicity. Treated mice gained weight as per normal chow-fed C57Bl/6 mice and were not rendered anaemic. Rather than reducing appetite, DFS appeared to increase appetite. This, together with increased core body temperature, suggested improved whole body metabolism. Metabolic chamber studies confirmed this, showing higher O2 consumption and CO2 production in DFS-treated mice. These findings suggest iron chelation may be effective in improving the metabolic derangements associated with obesity.

Declaration relating to disposition of project thesis/dissertation

I hereby grant to the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or in part in the University libraries in all forms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all property rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation.

I also authorise University Microfilms to use the 350 word abstract of my thesis in Dissertation Abstracts International (this is applicable to doctoral theses only).

…………………………………………………………… ……………………………………..……………… ……….……………………...…….… Signature Witness Date

The University recognises that there may be exceptional circumstances requiring restrictions on copying or conditions on use. Requests for restriction for a period of up to 2 years must be made in writing. Requests for a longer period of restriction may be considered in exceptional circumstances and require the approval of the Dean of Graduate Research.

FOR OFFICE USE ONLY Date of completion of requirements for Award:

THIS SHEET IS TO BE GLUED TO THE INSIDE FRONT COVER OF THE THESIS

ORIGINALITY STATEMENT

‘I hereby declare that this submission is my own work and to the best of my knowledge it contains no materials previously published or written by another person, or substantial proportions of material which have been accepted for the award of any other degree or diploma at UNSW or any other educational institution, except where due acknowledgement is made in the thesis. Any contribution made to the research by others, with whom I have worked at UNSW or elsewhere, is explicitly acknowledged in the thesis. I also declare that the intellectual content of this thesis is the product of my own work, except to the extent that assistance from others in the project's design and conception or in style, presentation and linguistic expression is acknowledged.’

Signed ……………………………………………......

Date ……………………………………………...... Acknowledgement and Thanks

Tackling a laboratory based PhD program is a formidable task for a clinician with no prior formal laboratory training. As I took the plunge into the unknown to uncover important contributions to the treatment of diabetes and obesity, I met with numerous challenges, not least having had to learn laboratory techniques. Along this 4-year journey, I had lots of failures, but it was the little successes that buoyed me and kept me going. I am indebted to the following people:

My Supervisors

Dr Jenny Gunton, Dr Ross Laybutt and Prof Don Chisholm for their kindness, patience and mentorship in guiding me through the PhD program.

My Family

Weiling, my wife and my children, Elliot, Emma, Ewan for their understanding and generosity. Mom and Dad for their support and encouragement.

The Gunton Laboratory

Bec, Sue Mei, Kim, Chris, Sue Lynn, Matt, Kuan, Tash for their friendship and induction into laboratory life.

Level 10 staff and staff of the Department of Immunology, Garvan Institute

Funding Bodies

NHMRC Post-graduate Medical and Dental Scholarship

Eli Lilly Diabetes Grant

Abstract

Oxygen Species (ROS) resulting in insulin resistance and cell apoptosis. Iron chelation reduces diabetes in iron overload syndromes but effects on obesity and related metabolic variables are unknown in non-overloaded individuals. Hypoxia

Inducible Factor- 1-alpha (HIF-1α) is a transcription factor up-regulated by hypoxia, which is also increased by iron chelation in-vitro. HIF-1α impacts lipid and whole body metabolism and cellular survival by improving glucose uptake, facilitating anaerobic glycolysis, and attenuating ROS.

ii regulation of hepatic lipogenic expression and improved insulin signalling gene expression. Importantly, DFS did not impair appetite and had no apparent toxicity.

Treated mice gained weight as per normal chow-fed C57Bl/6 mice and were not rendered anaemic. Rather than reducing appetite, DFS appeared to increase appetite. This, together with increased core body temperature, suggested improved whole body metabolism. Metabolic chamber studies confirmed this, showing higher

O2 consumption and CO2 production in DFS-treated mice.

These findings suggest iron chelation may be effective in improving the metabolic derangements associated with obesity.

iii

Publication and Presentations Arising from Thesis Work

Co-First-Author-Manuscript

Hypoxia-Inducible Factor-1 regulates  cell function in mouse and human islets.

Cheng K, Ho K, Stokes R, Scott C et al. J Clin Invest 2010. 120(6): 2171-83

Presentations

Oral iron chelation increases metabolism and protects against diet-induced obesity.

Endocrine Society of Australia Annual Scientific Meeting in Adelaide 2009.

(Won the NOVARTIS Junior Scientist Award)

Oral iron chelation down-regulates lipogenesis and reduces hyperlipidaemia and fatty liver of obesity. Australian Diabetes Society Annual Scientific Meeting in

Adelaide 2009 (ORAL)

Increasing HIF-1 prevents weight gain and improves insulin resistance in High Fat

Fed-mice. Australian Diabetes Society Annual Scientific Meeting in Melbourne 2008.

(POSTER)

Over expression of ARNT improves gene expression in β-cells. Australian Diabetes

Society Annual Scientific Meeting in Christchurch 2007. (POSTER)

Table of Contents

Acknowledgement and Thanks……………………………………………..……...... i

Abstract……………………………………………………………………….……...... ii

Publication and Presentations ……………………………………….…………...... iv

Table of Contents…………………………………………………………….………...... v

List of Figures………………………………………………………………….………....xiv

List of Abbreviations………………………………………………………………...... xix

Chapter 1 Literature review

1.1 OBESITY AND THE METABOLIC SYNDROME……………………………...... 2

1.1.1 Obesity is a significant health problem...... 3

1.1.2 Non-alcoholic fatty liver disease (NAFLD) and obesity...... 4

1.1.3 Iron stores and the metabolic syndrome...... 5

1.2 ENERGY HOMEOSTASIS AND HYPOTHALAMUS…………………….….…..8

1.2.1 Weight gain is affected by energy expenditure...... 9

1.2.2 Appetite regulation and energy homeostasis...... 11

1.2.3 The arcuate nucleus: first order integrating centre...... 12

1.2.4 Paraventricular nucleus: second order integrating centre...... 13

1.2.5 Ventral medial hypothalamus: satiety centre...... 14

1.2.6 Lateral hypothalamic area: feeding centre...... 14

1.2.7 Orexin stimulation promotes energy expenditure...... 15

1.2.8 Weight maintenance effect of orexin is via OX2 receptor signalling...... 16

1.3 HYPOXIA INDUCIBLE FACTOR-1 ALPHA AND CELL METABOLISM….....18

1.3.1 HIF-1 and cell metabolism...... 19

1.3.2 The molecular structure of HIF...... 20

1.3.3 Regulation of HIF-1...... 21

1.3.4 HIF-1 and mitochondrial function...... 23

1.3.5 HIF-1 and glucose metabolism...... 25

1.4 STEROL REGULATORY ELEMENT BINDING PROTEINS AND LIPID

METABOLISM…………………………………………………………………...... 27

1.4.1 SREBP proteins in lipid metabolism...... 28

1.4.2 Overview of the SREBP isoforms...... 29

1.4.3 SREBP-1c and fatty liver...... 31

1.4.4 SREBP-1a and dyslipidaemia...... 32

1.4.5 Regulation of SREBP genes...... 33

1.4.6 Liver X-activated receptors (LXR) regulate SREBP- 1c transcription...... 35

1.4.7 Fatty acid synthase (FAS)...... 36

1.4.8 Peroxisome proliferator-activated receptor gamma (PPAR)...... 37

1.5 HYPOXIA INDUCIBLE FACTOR-1ALPHA AND LIPID METABOLISM…...... 38

1.5.1 Hypoxic adaptation...... 39

1.5.2 HIF is involved in hypoxic repression of FAS and SREBP-1c...... 40

1.5.3 Stra13/DEC1 and DEC2 are involved in hypoxic repression of SREBP-1c

and FAS...... 40

1.5.4 Orexin signals via HIF-1 activity...... 42

1.5.5 Orexin-stimulated glucose utilisation is mediated through HIF-1α...... 43

1.6 IRON TRANSPORT, PHYSIOLOGY AND CHELATION…………………...... 45

1.6.1 Iron uptake and transport...... 46

1.6.2 Iron and mitochondrial function...... 47

1.6.3 Iron chelators: pharmacology...... 48

Chapter 2 Research methods and experimental protocols

2.1 ANIMAL CARE……………………………………………………………...... 55

2.2 MOUSE MODELS… ..……………………………………………………...... 55

2.2.1 Male WT C57Bl/6 mice on continuous high fat diet + DFS...... 56

2.2.2 Female WT C57Bl/6 mice on continuous high fat diet + DFS...... 56

2.2.3 Male ob/ob (C57Bl/6) mice on blended normal chow diet + DFS...... 56

2.2.4 Male WT C57Bl/6 mice on blended normal chow diet + DFS...... 56

2.2.5 Male WT C57Bl/6 mice on continuous high fat diet + DFO...... 56

2.2.6 Female NOD mice on unblended normal chow + DFO...... 57

2.3 PREPARATION OF DIETS……………………………………………...... 57

2.3.1 High fat diet…………………………………………………………...... 57

2.3.2 Blended chow diet…………………………………………………...... 58

2.4 IN-VIVO TECHNIQUES AND RATIONALE FOR TIME-POINTS

2.4.1 Glucose tolerance tests...... 59

2.4.2 Insulin tolerance tests...... 59

2.4.3 Glucose stimulated insulin secretion...... 59

2.4.4 Food intake studies...... 61

vii

2.4.5 Oxymax respirator studies...... 62

2.4.6 DEXA scanning...... 62

2.5 HISTOLOGICAL TECHNIQUES

2.5.1 Tissue harvest...... 63

2.5.2 Histological preparation………………………………...... 63

2.5.3 Haematoxylin and eosin staining………………………………...... 63

2.5.4 Cover-slipping…………………………………………………………...... 64

2.5.5 Perl’s protocol………………………………………………………...... 64

2.5.6 Sirius red staining...... 64

2.6 IN-VITRO TECHNIQUES

2.6.1 Plasma, liver and lipid biochemical assays...... 65

2.6.2 Protein extraction and quantitation...... 65

2.6.3 Western immunoblot analysis for HIF-1α...... 66

2.6.4 RNA preparation and real-time quantitative PCR...... 66

2.7 STATISTICAL ANALYSIS…………………………………………………...... 67

2.8 SOLUTIONS AND BUFFERS………………………………………………...... 68

Chapter 3 Effects of deferasirox on high fat-fed wild-type C57Bl/6 mice

3.1 Introduction...... 71

3.2 Experiment time-lines...... 72

3.3 Deferasirox (DFS)-treated mice gained significantly less weight on

continuous High Fat Diet (HFD)……………………………………………...... 77

3.4 DFS treatment did not impair appetite...... ………………………………...... 89

viii

3.5 DFS mice had higher basal metabolic rates……………………...…………….93

3.6 DFS mice had higher core body temperatures…………………....…………..100

3.7 DFS mice had lower fasting hyperinsulinaemia and tended to have better

insulin sensitivity...... 102

3.8 DFS mice had preserved beta cell function and better glucose tolerance

after prolonged HFD feeding...... 111

3.9 DFS mice were less obese...... 122

3.10 DFS-treated mice had lower serum and liver lipids...... 127

3.11 DFS-treated livers had lower iron stores and showed increased levels of

HIF-1. ……………………………………………………………………………135

3.12 DFS treatment resulted in altered gene expression for insulin signalling

and lipid metabolic pathways...... 139

3.13 DFS treatment did not cause anaemia...... 149

3.14 Summary...... 152

Chapter 4 Effects of deferasirox on ob/ob mice and wild-type C57Bl/6 mice

4.1 Introduction...... 155

4.2 Experiment time-lines ...... 156

4.3 DFS-treated ob/ob mice gained less weight...... 158

4.4 Baseline glucose tolerance tests (GTTs)...... 160

4.5 DFS-treated ob/ob mice had higher random blood glucose

concentrations...... 162

4.6 Glucose tolerance tests (GTTs) were not significantly different at Week 8 of

treatment...... 164

4.7 DFS-treated ob/ob mice were not more insulin sensitive despite lower

weight gain………………….………………………………………….…….……166

4.8 DFS-treated ob/ob mice tended to eat more……...………………………...... 168

4.9 DFS did not affect whole body metabolism...... 170

4.10 DFS-treated ob/ob mice did not have higher core body temperatures...... 172

4.11 There were no significant differences in fat depots……….……………….....174

4.12 DFS-treated ob/ob livers were not lighter...... 176

4.13 DFS-treated ob/ob mice have less fatty liver...... 178

4.14 There were no differences between fasted triglyceride levels...... 180

4.15 DFS-treated ob/ob mice livers had less iron staining...... 182

4.16 DFS in chow-fed C57Bl/6 mice did not affect weight gain...... 184

4.17 DFS-treated chow-fed C57Bl/6 mice tended to have lower random blood

glucose concentrations...... 186

4.18 Glucose tolerance did not differ between DFS and CON mice...... 188

4.19 DFS did not significantly improve insulin sensitivities in chow-fed C57Bl/6

mice...... 190

4.20 DFS did not significantly increase appetite of chow-fed C57Bl/6 mice...... 192

4.21 DFS did not affect whole body metabolism of chow-fed C57Bl/6 mice...... 194

4.22 DFS did not affect core body temperatures of chow-fed C57Bl/6 mice...... 196

4.23 DFS-treated chow-fed C57Bl/6 mice had increased pancreatic and muscle

weights but were not different in body weights to CON mice...... 198

4.24 DFS-treated chow-fed C57Bl/6 mice had significantly lower fasted plasma

triglyceride levels...... 200

4.25 DFS-treated chow-fed C57Bl/6 mice were protected from fatty livers...... 202

4.26 DFS treatment resulted in less liver iron...... 204

4.27 DFS treatment did not result in significantly higher liver HIF-1α...... 206

4.28 DFS-treated chow-fed C57Bl/6 mice showed down-regulation of lipolysis

gene expression in livers...... 208

4.29 DFS-treated chow-fed C57Bl/6 mice had reduced gene expression for lipid

synthesis...... 210

4,30 Summary...... 212

Chapter 5 Effects of deferoxamine on high fat-fed wild type C57Bl/6 mice

5.1 Introduction...... 215

5.2 Experiment time-line...... 216

5.3 DFO treatment did not affect weight gain in HFD mice...... 217

5.4 DFO-treated mice had lower random blood glucose………………...... 219

5.5 Glucose Tolerance Tests (GTTs) were not different between groups...... 221

5.6 DFO-treated mice had improved insulin secretion profile……..……………..224

5.7 DFO treatment over 10 weeks resulted in lowering of haemoglobin but not

anaemia...... 226

5.8 Summary...... 228

Chapter 6 Effects of deferoxamine on non-obese diabetic mice

6.1 Introduction...... 230

6.2 Experiment time-lines...... 231

6.3 DFO delayed onset of diabetes in NOD mice...... 233

6.4 DFO did not affect islet infiltrate in diabetic mice at sacrifice...... 235

6.5 No significant differences in plasma lipids with DFO treatment...... 237

6.6 Summary...... 239 xi

Chapter 7 Discussion

7.1 Effects of iron chelation and Hypoxia Inducible Factor-1α on components of

the Metabolic Syndrome...... 241

7.2 Oral iron chelation reduced obesity, hepatic and serum lipids...... 246

7.3 Iron chelation was associated with HIF-1α up-regulation and lipid gene

expression changes...... 245

7.4 Iron chelation protected beta cell function...... 247

7.5 Oral iron chelation apparently improves hepatic insulin resistance but not

whole body insulin resistance...... 248

7.6 Prolonged oral iron chelation protects against glucose intolerance...... 249

7.7 Oral iron chelation affects substrate utilisation and increases whole body

metabolism...... 250

7.8 HIF-1 activity regulates metabolism...... 251

7.9 The efficacy of DFO versus DFS...... 255

7.10 Iron chelation is well-tolerated...... 256

7.11 Conclusion and suggestions for future work...... 257

References...... 259

Publication

Cheng, K., Ho, K et al., Hypoxia-inducible factor-1alpha regulates beta cell function in mouse and human islets. J Clin Invest, 2010. 120(6): p. 2171-83.

(Copy included at the back of thesis)

xii

List of Figures

Figure 1.1 Adaptive thermogenesis and uncoupling proteins.

Figure 1.2 Anatomy of the rat hypothalamus

Figure 1.3 Overview of human obesity and central mechanisms

Figure 1.4 HIF-1 and HIF-1 (ARNT) subunits

Figure 1.5 Regulation of HIF-1α

Figure 1.6 Genes regulated by SREBPs

Figure 3.1 Experimental time-lines: C57Bl/6 mice on HFD experiment 1

Figure 3.2 Experimental time-lines: C57Bl/6 mice on HFD experiment 2

Figure 3.3 Experimental time-lines: C57Bl/6 mice on HFD experiment 3

Figure 3.4 Experimental time-lines: C57Bl/6 mice on HFD experiment 4

Figure 3.5 Experimental time-lines: C57Bl/6 mice on HFD experiment 5

Figure 3.6 Experiment 1: weight curve and overview

Figure 3.7 Experiment 2: weight curve and overview

Figure 3.8 Experiment 3: weight curve and overview

Figure 3.9 Experiment 4: weight curve and overview

Figure 3.10 Experiment 4: a pair of DFS and CON mice at week 19

Figure 3.11 Experiment 5: weight curve and overview

Figure 3.12 Experiment 2: three-day food intake in week 3

Figure 3.13 Experiment 3: three-day food intake in week 4

Figure 3.14 Experiment 4: comparative food intake at weeks 0, 8 and 25.

Figure 3.15 Oxymax studies: experiment 2 at week 4

Figure 3.16 Oxymax studies: experiment 3 at week 5

Figure 3.17 Oxymax studies: experiment 4 at week 4, day 2

Figure 3.18 Oxymax studies: experiment 4 at week 8

xiii

Figure 3.19 Oxymax studies: experiment 4 at week 25

Figure 3.20 DFS mice have higher core body temperature

Figure 3.21 Experiment 1: fasting insulin at week 7

Figure 3.22 Experiment 3: fasting Insulin at week 8

Figure 3.23 Experiment 1: insulin tolerance test at week 6

Figure 3.24 Experiment 2: insulin tolerance test at week 9

Figure 3.25 Experiment 3: insulin tolerance test at week 7

Figure 3.26 Combined experiments 1 to 3: (late) insulin tolerance test

Figure 3.27 Experiment 4: (early) insulin tolerance test at week 2

Figure 3.28 Experiment 1: glucose stimulated insulin secretion at week 7

Figure 3.29 Experiment 3: glucose stimulated insulin secretion at week 8

Figure 3.30 Combined experiments 1 and 3: glucose stimulated insulin

secretion

Figure 3.31 Experiment 1: intra-peritoneal glucose tolerance test at week 5

Figure 3.32 Experiment 2: intra-peritoneal glucose tolerance test at week 8

Figure 3.33 Experiment 3: intra-peritoneal glucose tolerance test at week 6

Figure 3.34 Combined experiments 1, 2, 3: (early) glucose tolerance test

Figure 3.35 Experiment 4: (late) glucose tolerance test at week 21

Figure 3.36 Experiment 5: (late) glucose tolerance test at week 27

Figure 3.37 Experiment 2: fat weights

Figure 3.38 Experiment 4: organ weights

Figure 3.39 Experiment 4: plasma leptin

Figure 3.40 Experiment 5: organ weights

Figure 3.41 Experiment 1: plasma lipids (non-fasted)

Figure 3.42 Experiment 2: plasma lipids (fasted)

xiv

Figure 3.43 Experiment 5: plasma lipids (fasted)

Figure 3.44 Experiment 2: liver triglycerides (fasted)

Figure 3.45 Haematoxylin and eosin slide: experiment 3 liver

Figure 3.46 Haematoxylin and eosin slide: experiment 4 liver

Figure 3.47 Sirius red slide: experiment 3 liver

Figure 3.48 Perls’ stain: experiment 3 liver

Figure 3.49 Western blot: experiment 2 liver

Figure 3.50 Western blot: DFS versus CON liver

Figure 3.51 Liver gene expression 1

Figure 3.52 Liver gene expression 2

Figure 3.53 Liver gene expression 3

Figure 3.54 Liver gene expression 4

Figure 3.55 Visceral adipose tissue (VAT) gene expression 1

Figure 3.56 Visceral adipose tissue (VAT) gene expression 2

Figure 3.57 Gastrocnemius gene expression 1

Figure 3.58 Gastrocnemius gene expression 2

Figure 3.59 DFS effect on haemoglobin

Figure 3.60 DFS effect on serum iron

Figure 4.1 Ob/ob mice experiment time-line

Figure 4.2 C57Bl/6 mice on chow experiment time-line

Figure 4.3 Ob/ob mice experiment weekly weight change

Figure 4.4 Ob/ob mice experiment: glucose tolerance test week 0

Figure 4.5 Ob/ob mice experiment: random blood glucose

Figure 4.6 Ob/ob mice experiment: glucose tolerance test week 8

Figure 4.7 Ob/ob mice experiment: insulin tolerance test

Figure 4.8 Ob/ob mice experiment: 24-hour food intake study

Figure 4.9 Ob/ob mice experiment oxymax study

Figure 4.10 Ob/ob mice experiment rectal temperatures

Figure 4.11 Ob/ob mice organ weights 1

Figure 4.12 Ob/ob mice organ weights 2

Figure 4.13 Ob/ob liver haematoxylin and eosin slide

Figure 4.14 Ob/ob mice fasting triglycerides

Figure 4.15 Ob/ob liver Perls’ stain

Figure 4.16 Chow-fed wild-type C57Bl/6 mice weekly weight change

Figure 4.17 Chow-fed wild-type C57Bl/6 mice random blood glucose

Figure 4.18 Chow-fed wild-type C57Bl/6 mice Glucose Tolerance Test week 8

Figure 4.19 Chow-fed wild-type C57Bl/6 mice Insulin Tolerance Test

Figure 4.20 Chow-fed wild-type C57Bl/6 mice 24-hours food intake study

Figure 4.21 Chow-fed wild type C57Bl/6 mice oxymax study

Figure 4.22 Chow-fed wild-type C57Bl/6 mice weekly rectal temperatures

Figure 4.23 Chow-fed wild-type C57Bl/6 mice organ weights

Figure 4.24 Chow-fed wild-type C57Bl/6 mice fasting plasma triglycerides

Figure 4.25 Chow-fed wild-type C57Bl/6 mice liver Haematoxylin and Eosin slide

Figure 4.26 Chow-fed wild-type C57Bl/6 mice liver Perls’ stain

Figure 4.27 Chow-fed wild-type C57Bl/6 mice liver Western Blot

Figure 4.28 Chow-fed wild-type C57Bl/6 mice liver gene expression 1

Figure 4.29 Chow-fed wild-type C57Bl/6 mice liver gene expression 2

Figure 5.1 C57Bl/6 mice on HFD-DFO experiment time-line

Figure 5.2 DFO treatment did not affect weight gain in HFD mice

Figure 5.3 DFO treated mice have lower random blood glucose concentrations.

xvi

Figure 5.4 Glucose tolerance test at week 6 was not different

Figure 5.5 Glucose tolerance test was not different at week 11 of treatment

Figure 5.6 DFO treated mice had slightly better insulin secretion profile

Figure 5.7 DFO treatment over 10 weeks resulted in lowering of haemoglobin

but not anaemia.

Figure 6.1 Time-line: NOD experiment 1

Figure 6.2 Time-line: NOD experiment 2

Figure 6.3 DFO delayed onset of diabetes in NOD mice

Figure 6.4 DFO had no effect on islet infiltrate in diabetic mice

Figure 6.5 No differences in NOD plasma lipids

Figure 7.1 Hypothalamic HIF-1α increases orexin activity leading to Increased

metabolism

Figure 7.2 Summary of HIF-1 action on glucose and lipid metabolism

xvii

List of Abbreviations

ACC Acetyl-CoA Carboxylase

AGRP Agouti Gene-Related Peptide

Akt2 v-akt murine thymoma viral oncogene homolog 2

Aldo Aldolase

AMPK AMP-activated protein kinase

ANOVA Analysis of Variance

ARC Arcuate Nucleus

ARC Animal Resources Centre

ARNT Aryl hydrocarbon Receptor Nuclear Translocator

ATP Adenosine-5'-triphosphate

AUC Area under the Curve

BAT Brown Adipose Tissue

BMC Bone Mineral Content

BMD Bone Mineral Density

BMI Body Mass Index

BMR Basal Metabolic Rate

β-Hif1a-null C57Bl/6 mice with β-cell specific HIF-1 α deletion

bHLH basic Helix-Loop-Helix

bHLH-Zip basic helix-loop-helix–leucine zipper

BTF Biological Testing Facility

cDNA complementary Deoxyribonucleic Acid

xviii

C2C12 cells Mouse Myoblast Cell Line

C57Bl/6 mouse Common inbred strain of lab mouse

CART Cocaine- and Amphetamine-Regulated Transcript

CBP c AMP response element binding protein

C/EBP alpha CCAAT/enhancer-binding protein alpha

ChIP Chromatin Immunoprecipitation

CHO-7 Chinese Hamster Ovary cells

CNS Central Nervous System

CO2 Carbon Dioxide

CON Control mice on normal HFD or Chow diet, not

receiving iron chelation drug

CPT1 Carnitine palmitoyltransferase I

CSF Cerebrospinal fluid

Cre Type I topoisomerase found in bacteriophage P1 that

catalyses the site-specific recombination of DNA

between loxP sites

CRF Corticotrophin-releasing factor

Ct Cross threshold

DEC2 Differentiated Embryonic Chondrocyte2

DEXA Dual Energy X-ray Absorptiometry

DFO Deferoxamine

DFS Deferasirox

DMH Dorsomedial Hypothalamic nucleus

DMT-1 Divalent Metal Transporter

xix

EDTA Ethylene-Diamine-Tetraacetic Acid

EE Energy Expenditure

eNOS endothelial Nitric Oxide Synthase

ELISA Enzyme Linked Immunosorbent Assay

EMSA Electrophoretic Mobility-Shift Assay

EPO Erythropoietin

ER Endoplasmic Reticulum

Exp Experiments

FAS Fatty Acid Synthase

FFAs Free Fatty Acids

FGF21 Fibroblast growth factor 21

FIH-1 Factor Inhibiting HIF-1

FXR Farnesoid X Receptor

GAPDH Glyceraldehydes-3-phosphate dehydrogenase

GLUT Glucose Transporter

GPCR G-Protein Coupled Receptor

GSIS Glucose Stimulated Insulin Secretion

GTPases Enzymes that hydrolyze guanosine triphosphate

GTT Glucose Tolerance Test

H&E Haematoxylin and Eosin staining

HEK293 cells Human Embryonic Kidney Cell Line 293

Hepa1c1c7 Mouse hepatoma cell line

HIF-1α Hypoxia Inducible Factor- 1 alpha

HIF-1 Hypoxia Inducible Factor- 1 beta

HFD High Fat Diet

HFD+DFS DFS added and mixed into HFD

HFE Most common form of Haemochromatosis gene mutation

HH (Hereditary) Haemochromatosis

HMGred 3-hydroxy-3-methyl-glutaryl-CoA reductase

HMGsyn 3-hydroxy-3-methyl-glutaryl-CoA synthase

HNF4α Hepatocyte Nuclear Factor- 4 alpha

HRE Hypoxia Response Element

HSL Hormone-Sensitive Lipase

IRS Insulin Receptor Substrate

ISC Iron–sulfur cluster (or Fe/S)

ITT Insulin Tolerance Test

KCl Potassium Chloride

L-ARNT KO ARNT deletion specific in liver cells

LDHA Lactate Dehydrogenase

LDL Low Density Lipoprotein xxi

LHA Lateral hypothalamus area

LID Detergent-based cell lysis buffer

LON protease ATP-dependent proteases of misfolded proteins

loxP Locus of Crossover in P1 ( See ‘Cre’)

LPI Labile Plasma Iron

LPL Lipoprotein lipase

LXR Liver X Receptor

MCH Melanin-Concentrating Hormone

ME Median Eminence

MeOH Methanol

MetS Metabolic Syndrome

Min6 Mouse Insulinoma Cell Line

MODY Maturity Onset Diabetes of the Young

mRNAs Messenger ribonucleic acid

α-MSH α-melanocyte-stimulating hormone

MVA Mevalonate

NaCl Sodium Chloride

NADPH Nicotinamide Adenine Dinucleotide Phosphate-Oxidase

NAFLD Non-Alcoholic Fatty Liver Disease

NEFA Non-Esterified Fatty Acids

NOD Non-obese diabetic mouse

NP40 Nonyl phenoxypolyethoxylethanol

NPY Neuropeptide Y

xxii

nSREBP-1c nuclear SREBP-1c

NTBI Non-Transferrin Bound Iron

O2 Oxygen

ob/ob Mouse with homozygous mutations in Ob gene

Ob-Rb Leptin receptors

OX1R Orexin, type 1

OXPHOS Oxidative phosphorylation

PAS Per-ARNT-Sim

PBS Phosphate Buffered Solution

PBST Phosphate Buffered Solution with Tween-20

PCR Polymerase Chain Reaction

PDH Pyruvate Dehydrogenase

PDK Pyruvate Dehydrogenase Kinase

PFK Phosphofructokinase

PGK-1 Phosphoglycerate kinase-1

PHD Prolyl hydroxylases

PMSF Phenylmethanesulfonylfluoride

PPAR Peroxisome Proliferator-Activated Receptor gamma

POMC Pro-opiomelanocortin

PVDF Polyvinylidene Fluoride

PVN Paraventicular nucleus

Redox Reduction-oxidation reaction

xxiii

RER Respiratory exchange ratio

RIPA Radioimmunoprecipitation assay buffer

RLT Cell and Nuclear lysis buffer used in extracting RNA

RNA Ribonucleic acid

ROS Reactive Oxygen Species

S1P Site-1 protease

SCAP SREBP cleavage–activating protein

SCN Suprachiasmatic nucleus

SDS PAGE Sodium dodecyl sulfate polyacrylamide gel

electrophoresis

SEM Standard Error of the Mean

siRNA small interfering RNA

SPA Spontaneous physical activity

SRE Sterol response elements

SREBP Sterol response element-binding protein

Stra13 Stimulated by Retinoic Acid 13 Homologues, also

known as DEC1

3T3-L1 Mouse embryonic fibroblast - adipose like cell line

T2D Type 2 Diabetes

TBP TATA box binding-protein

TBST Tris-Buffered Saline Tween-20

TCA Tricarboxylic acid cycle

TEF Thermic effect of food

xxiv

UCP2 Uncoupling Protein 2

VAT Visceral Adipose Tissue

VEGF Vascular Endothelial Growth Factor

VEGFR3 Vascular Endothelial Growth Factor Receptor 3

VHL von Hippel-Lindau

pVHL von Hippel-Lindau protein

VLDL very-low-density lipoprotein

VMH Ventromedial hypothalamic nucleus

VCO2 Carbon dioxide production

VO2 Oxygen consumption

WAT White adipose Tissue or subcutaneous fat depots

WT Wild type

xxv

CHAPTER 1

Literature review

OBESITY

AND

THE METABOLIC SYNDROME

Obesity is a significant health problem

It is estimated that more than 300 million people worldwide are obese and if current trends continue, the number of overweight persons will increase to 1.5 billion by 2015. Nationally, the number of obese Australians have increased from

2.0 million in 1993 to 3.1 million in 2005 [1]. Obesity not only leads to multi-organ dysfunction but may also affect psychosocial function and can impair work performance. It has been estimated that the cost of obesity in Australia in 2005 was $1,721 million. Of this amount, $1,084 million were direct health costs, and

$637 million indirect health costs due to lost work productivity, absenteeism and unemployment. The prevalence cost per year for each obese adult has been estimated at $554 and the value of an obesity cure is about $6,903 per obese person [1].

Many obese individuals exhibit a cluster of metabolic abnormalities collectively called the Metabolic Syndrome (MetS). These include glucose intolerance

(manifested as type 2 diabetes, impaired glucose tolerance, or impaired fasting glycaemia), insulin resistance, central obesity, dyslipidaemia, and hypertension,

[2] which are all well documented risk factors for cardiovascular disease. Since its original conception in 1988, [3] several definitions of the MetS have evolved

[4]. Most have recognised central obesity as a core component. Although in recent years, there has been controversy regarding the precise definitions of

MetS and its ability to predict cardiovascular mortality accurately, the association

3 of MetS and cardiovascular mortality is well documented, [5-8] and may have better long term prognostic value than short term assessment of cardiovascular risks [9-10].

Non-alcoholic fatty liver disease (NAFLD) and obesity

Many individuals with obesity and insulin resistance have fatty livers [11]. A subset of individuals with fatty liver goes on to develop fibrosis, cirrhosis, and liver failure. Fat accumulation in the liver parenchyma is the result of abnormal fatty acids metabolism including: failure of the synthesis/secretion of apolipoproteins and triglycerides; excessive delivery of free fatty acids (FFAs) to the liver; and decreased mitochondrial lipid oxidation. Liver biopsy studies in obese individuals have demonstrated that 30–40% of patients develop more than a simple steatosis [12]. In patients with NAFLD, 74% will have fibrosis and ~30% will have cirrhosis at the time of their initial liver biopsy [12]. Excessive hepatic fat owing to insulin resistance in patients with the metabolic syndrome often predisposes the liver to chronic inflammation.

Lipotoxic effects of FFAs and lipid intermediates cause dysfunction of liver cell organelles from the production of reactive oxidative species (ROS), the activation of pro-inflammatory mediators, and, ultimately, apoptosis. Toxic lipids and release of cytokines impair insulin signalling, which in turn contributes to diminished very-low-density lipoprotein (VLDL) assembly and liver secretion,

4 involving an insufficient regulation of important transcription factors required for lipogenesis [13]. Impaired regulation activity of several nuclear receptors provides a potential molecular mechanism for the deteriorating metabolic dysfunction in NAFLD [14]. Deregulation of insulin signalling is critical in promoting progression from steatosis to inflammation, lipid peroxidation, and liver injury associated with steatohepatitis. Mitochondria are usually the first organelle to be impaired, with decreased mitochondrial fatty acid oxidation and increased compensatory peroxisomal fatty acid oxidation [15]. Oxidative stress further increases the hepatocyte‟s susceptibility to toxic stimuli and stress. Activation of cellular defense programs, and in particular activation of Kupffer cells, produces inflammatory mediators, which in turn activate hepatic stellate cells to synthesise collagen leading to hepatic fibrosis and cirrhosis.

Iron stores and the metabolic syndrome

The link between iron stores and the presence of the metabolic syndrome has been demonstrated for more than a decade [16-18]. Individuals with iron overload syndromes such as hereditary haemochromatosis (HH) and transfusion-dependent anaemias are at high risk for developing diabetes mellitus

[19]. It is known that the prevalence of impaired glucose tolerance and or diabetes can be as high as 50% in genetic HH [20]. This is thought to be due to a combination of -cell failure due to deposition of iron in the pancreas as well as insulin resistance due to iron deposition in liver and skeletal muscle [21]. Affected

5 individuals are likely more likely to be overweight [21]. Various studies have shown that elevated ferritin concentrations frequently cluster with well- established risk factors of diabetes including obesity, metabolic syndrome, chronic inflammation, and altered circulating adipokines [22]. This association holds true in many ethnic populations [16, 22] and occurs even in the absence of a genetic mutation for HH and at levels well below that of iron overload syndromes [23]. The adjusted odds ratios are higher for type 2 diabetes and metabolic syndrome for patients in the highest ferritin quartile compared with those in the lowest quartile [24]. These associations remained significant after further adjustment for dietary factors, body mass index, inflammatory markers, and adipokines [22]. Interestingly, serum ferritin was positively correlated with androgen levels in obese premenstrual women with Polycystic Ovarian

Syndrome [25]. Moreover, in a cross-sectional study of Mexican American men aged 20-49 years, serum ferritin concentration was associated with waist-to-hip ratio independent of BMI and age [26].

Treatment of iron overloaded individuals with iron chelators is associated with improved survival [27]. Small studies have also suggested that even in HFE carriers with normal levels of iron stores, there may be an insulin sensitising benefit with modest iron depletion through venesection [28]. Similarly, in a study of overweight individuals with known NAFLD but without HH, phlebotomy improved insulin resistance measured by the HOMA-R index in people with the highest tertile of ferritin [29]. In the laboratory setting, rodents fed low iron diets

6 have lower fat gain and serum lipids, due to reduced lipogenic enzyme activities

[30]. It has been shown that addition of iron chelating agents to cells results in up-regulation of Hypoxia Inducible Factor-1-alpha (HIF-1α) protein, an important transcription factor in cell metabolism, which was more recently shown to be important in modulating glucose [31-34] and lipid [35-36] metabolism. This literature review will discuss how iron chelation therapy may play a role in treating obesity and the metabolic syndrome by effects on whole body metabolism, lipid metabolism, glucose handling and anti-oxidant effects.

ENERGY HOMEOSTASIS

AND

HYPOTHALAMUS

Weight gain is affected by energy expenditure

Figure 1.1 Dullo, A. Adaptive thermogenesis and uncoupling proteins.

Physiology & Behaviour: 83 (2004) 587-602

In simple thermodynamic terms, obesity can be considered as a chronic disorder of energy imbalance, in which long-term excess of energy intake compared to expenditure leads to the storage of that excess energy as adipose tissue [37].

Energy expenditure in the basal state (basal metabolic rate, BMR) or after food

(thermic effect of food, TEF), results in thermogenesis. An increase in body weight will also increase metabolic rate (on the basis of the extra energy cost for synthesis and subsequent maintenance of extra lean and fat tissues), which will produce a negative energy balance and hence a subsequent decline in body weight. Similarly, a reduction in body weight would also be automatically corrected inasmuch, as the resulting diminished metabolic rate due to the loss in weight will produce a positive balance, and hence a subsequent return towards the set or preferred weight. These compensatory changes in energy expenditure

(~15% above or below predicted values) reflect changes in resting energy expenditure that are unaccounted for by changes in body weight and composition. They reflect changes in metabolic efficiency and hence in adaptive changes in thermogenesis.

Heat is produced during muscle contraction and contributes to energy expenditure. The efficiency of muscular contraction during exercise is low

(~25%), but that of spontaneous physical activity (SPA) (including fidgeting, muscle tone and posture maintenance, and other low-level physical activities of everyday life) is even lower since these essentially involuntary activities comprise a larger proportion of isometric work, which is simply thermogenic. In this context, an increase in the amount of SPA in response to overfeeding or a decrease during starvation also constitutes adaptive changes in thermogenesis [38]. SPA plays a substantial role in weight regulation [38-40], but is not likely to be the major component in adaptive thermogenesis. Changes in muscle work efficiency could account for a third of the change in daily energy expended in physical activity [38].

Appetite regulation and energy homeostasis

SCN--Suprachiasmatic nucleus

PVN--Paraventicular nucleus ME--Median Eminence

LHA--Lateral hypothalamus area VMH--ventromedial hypothalamic nucleus ARC--Arcuate Nucleus

Figure 1.2 Anatomy of the rat hypothalamus. Pickup and

Williams. Textbook of diabetes.

2nd ed., vol2, Blackwell, 1997.pp65.1-65.29

Food intake and appetite regulation regulate energy consumption. Feeding behaviour is controlled by a series of short-term hormonal, psychological and neural signals that derive from the gastrointestinal tract, such as cholecystokinin and ghrelin [41]. Other hormones such as insulin and leptin, together with circulating nutrients, indicate long-term energy stores [41]. All these signals act at several central nervous system (CNS) sites but the pathways converge on the hypothalamus, which contains a large number of peptides and other neurotransmitters that influence food intake [41]. The hypothalamus has thus been recognised as a central region of feeding regulation [41]. The basic

11 neuroanatomy of the rat hypothalamus is shown in Figure 1.2. (See List of

Abbreviations for explanations) [42].

The arcuate nucleus: first order integrating centre

Figure 1.3 Overview of human obesity and central mechanisms. Crowley, V. Ann Clin Biochem, 2008. 45(Pt 3): p. 245-55

The arcuate nucleus (ARC), situated around the base of the third ventricle, lies immediately above the median eminence. The ARC is an elongated collection of neuronal cell bodies occupying nearly one-half of the length of the hypothalamus and is subdivided into several functional domains. Neuropeptide Y (NPY) and agouti gene-related protein (AGRP), both potent stimulators of food intake, are co-localised in a population of neurones in the ARC [43], while pro- opiomelanocortin [POMC; the precursor of α-melanocyte-stimulating hormone (α-

MSH)] and cocaine- and amphetamine-regulated transcript (CART), which

12 induce an anorexic response, are co-localised in an adjacent subset of ARC neurones [44]. These two populations interact with each other. The ARC has extensive reciprocal connections with other hypothalamic regions. These include the paraventicular nucleus (PVN), dorsomedial hypothalamic nucleus (DMH), ventromedial hypothalamic nucleus (VMH) and lateral hypothalamus area (LHA).

Capillaries in the underlying median eminence lack tight junctions: this region therefore effectively lies outside the blood–brain barrier [45], so that the ARC neurones are readily accessible to circulating messengers including leptin and insulin including glucose, may also gain access to the ARC by diffusion across the ependyma from the cerebrospinal fluid (CSF) in the third ventricle [46]. The arcuate nucleus of hypothalamus (ARC) is a point of convergence for both orexin and leptin signalling, which modulate the activities of neuropeptide regulators of food intake and metabolism such as neuropeptide Y (NPY), agouti-related peptide (AGRP), and pro-opiomelanocortin (POMC) [47].

Paraventricular nucleus: second order integrating centre

The PVN lies beside the top of third ventricle in the anterior hypothalamus. It receives and integrates axons projecting from the ARC NPY/AGRP and

POMC/CART neurons and from the orexin neurones of the lateral hypothalamus

[48]. The nucleus is rich in terminals containing numerous appetite-modifying neurotransmitters, including NPY, α-MSH, serotonin (5-HT), galanin, noradrenaline and the opioid peptides, and the PVN is particularly sensitive to

13 these neurotransmitters‟ effects on feeding and energy expenditure.

Corticotrophin-releasing factor (CRF) is expressed by neurones in the PVN that project to the median eminence [49] and may act to inhibit the NPY neurones of the ARC-PVN projection.

Ventral medial hypothalamus: satiety centre

Stimulation of the VMH inhibits feeding, whereas a lesion in this region causes overeating and weight gain [50]. Recent studies have shown high abundance of leptin receptors (long form: Ob-Rb) in neurones of the VMH, and evidence indicates that this region may be an important target for circulating leptin [51].

The VMH has direct connections with the PVN, the lateral hypothalamus and the

DMH. The DMH, located immediately dorsal to the VMH, has extensive direct connections with other hypothalamic nuclei such as the PVN, the lateral hypothalamus and the brainstem. The VMH and the lateral hypothalamus have no direct connections but connect indirectly through the DMH and the PVN. The

PVN and the DMH may cooperate functionally as a unit, which is involved in initiating and maintaining food intake [52]. The DMH contains plentiful insulin receptors as well as leptin receptors (Ob-Rb). Some ARC-NPY/AGRP neurones also terminate in the DMH.

Lateral hypothalamic area: feeding centre

The lateral hypothalamic area (LHA) is vaguely defined and comprises a large, diffuse population of neurons including defined subpopulations that express orexins and melanin-concentrating hormone (MCH), both peptides that stimulate food intake. NPY terminals are abundant in the LHA, in contact with orexin and

MCH cells [53], while the perifornical part of the LHA contains a high density of

NPY-„„Y5‟‟ receptors thought to mediate the appetite-stimulating effects of NPY.

The LHA was viewed classically as the „„feeding centre.‟‟ Stimulation of this nucleus increases food intake, while its destruction attenuates feeding and causes weight loss. This nucleus also contains large numbers of glucose- receptive neurones that respond to circulating glucose levels, probably mainly via pathways ascending from the hypothalamus [54].

Orexin stimulation promotes energy expenditure

Orexins (also known as hypocretins) are lateral hypothalamic neuropeptides that are upregulated with fasting and acutely promote appetite when administered into the central nervous system. The two receptors for orexin, type 1 (OX1R) and type 2 (OX2R), show differential affinity for the products of the prepro-orexin gene, orexin-A and orexin-B [55]. OX1R and OX2R exhibit distinct expression patterns, indicating distinct roles in behaviour and metabolism. Central administration of orexin neuropeptides to rodents acutely promotes appetite, and prepro-orexin deficiency or post-gestational ablation of orexin neurons in mice causes modest reductions in food intake. However, orexin-deficient mice also

15 exhibit narcolepsy, inactivity, and obesity, indicating that orexin may exert an overall catabolic influence upon energy balance [56]. Narcoleptic human individuals (the majority of whom are orexin deficient) have also been reported to have greater body mass index and higher incidence of metabolic syndrome [57].

Weight maintenance effect of orexin is via OX2 receptor signalling

The roles of orexin and its receptors in energy expenditure were recently demonstrated using CAG/ orexin transgenic mice. Immunohistochemical localisation of orexin-A in the brain of CAG/orexin mice demonstrates ectopic peptide production in medial, basal, lateral, and suprachiasmatic hypothalamic nuclei, nucleus accumbens, globus pallidus, hippocampal formation, ventral tegmental area, and locus coeruleus. These mice were found to be resistant to diet induced obesity with significantly reduced fat mass and serum leptin when fed on high fat diet but not on low-fat diet [58].

In order to determine whether this effect was due to the OXR1R or OX2R action,

OX1R-/- and OX2R-/- mice were crossed onto the CAG/orexin transgenic mice respectively. This effect was demonstrated to be due to OX2R signalling rather than the OX1R. On a high fat diet for 30 weeks, the OX2R-/- mice put on approximately 8g of weight more than the OX2R-/-; CAG/orexin crossed mice.

The OX1R-/- mice put on weight similarly to the CAG/orexin transgenics whether on high or low fat diets, proving that the weight maintenance effect of orexin was

16 not mediated via the OX1R. When the mice were studied in metabolic chambers, the effective mass-corrected energy expenditures of CAG/orexin male mice and

OX1R-/-; CAG/orexin mice on a high-fat diet were consistently elevated over those of wild-type mice and OX1R-/- mice, respectively, while the energy expenditures of OX2R-/-; CAG/orexin mice resembled those of OX2R-/- mice

[58]. There were no consistent differences in respiratory quotient (RQ; an indirect indicator of lipid versus carbohydrate utilisation) among different genotypic groups on a high-fat diet. The CAG/orexin transgene induced no differences in energy expenditure or RQ among any genotypic groups on a low-fat diet, regardless of the presence or absence of orexin receptors. Importantly,

CAG/orexin transgenic mice did not exhibit hyperactivity, regardless of diet or receptor status. Basal core body temperature in CAG/orexin mice on a high-fat diet tended to be higher than in wild-type controls, but this difference did not reach significance (wild-type low-fat diet: 36.6C ± 0.1C. The weight-adjusted food intake was significantly lower for the CAG/orexin transgenics. Using an

OX2R specific agonist directly reproduced may of these observed orexin effects on metabolism including significantly higher oxygen consumption, confirming that orexin signalling is via OX2R activation.

HYPOXIA INDUCIBLE FACTOR

– 1 ALPHA

AND

CELL METABOLISM

HIF-1 and cell metabolism

All organisms from bacteria to humans possess mechanisms to maintain O2 homeostasis that are essential for survival. Hypoxia can result in a failure to generate sufficient ATP to maintain essential cellular functions, whereas hyperoxia results in the generation of reactive oxygen intermediates and potentially lethal damage to membranes and DNA [59]. Thus cellular O2 concentrations must be tightly regulated within a narrow physiological range. This is achieved in part through a regulatory protein, hypoxia-inducible factor 1 (HIF-

1), transcription factor that appears to play a critical role in cellular metabolism during both development and postnatal life [59].

The identification of a Dnase I hypersensitive site in the 30-flanking region of the human EPO gene [60] led to the characterisation of a 50-bp sequence that functions as a hypoxia response element (HRE). When this DNA sequence was introduced into and transfected into cultured cells, there was a dramatic increase in reporter gene transcription when the cells were incubated at 1% O2 relative to expression at 20% O2. An electrophoretic mobility-shift assay (EMSA) using an oligonucleotide containing the 18 bp of the HRE as a probe revealed a DNA- binding activity, designated HIF-1, that was detected in hypoxic cells, but not in non-hypoxic cells [61]. HIF-1 fulfils several criteria for an activator of EPO transcription in response to hypoxia: (a) Mutations in the HRE eliminate HIF-1 binding and HRE function; (b) inducers of EPO gene expression [hypoxia, cobalt

19 chloride, and desferrioxamine (an iron chelator)] also induce HIF-1 with similar kinetics; and (c) inhibitors of hypoxia-induced EPO expression also inhibit HIF-1 induction [61]. HIF-1 activity and the expression of reporter genes containing the

EPO HRE were induced by hypoxia in all mammalian cell types analysed, including both EPO-expressing and non-expressing cells [62], suggesting that

EPO is not the only target of HIF-1 in hypoxic cells. In addition, HIF-1 deficiency was associated with decreased expression of at least 13 different genes encoding glucose transporters and glycolytic enzymes [63]. HIF-1 thus mediates increased O2 delivery to cells, as well as adaptation to decreased O2 availability

(via glucose transporters and glycolytic enzymes).

The molecular structure of HIF

HIF-1 exists as a heterodimer consisting of two subunits with molecular masses of 120–130 kD and 91–94 kD that are designated HIF-1 and HIF-1, respectively [64]. Both subunits contain basic helix-loop-helix (bHLH) and PAS domains (See figure). The bHLH domain mediates dimerisation and DNA binding in a large number of transcription factors. The members of one subfamily of bHLH factors contain an additional dimerisation motif, the PAS domain. HIF-1 is identical to ARNT (aryl hydrocarbon receptor nuclear translocator), which was previously shown to dimerise with the aryl hydrocarbon receptor after activation

Figure 1.4 HIF-1 and HIF-1 (ARNT) subunits. Semenza GL. Annu. Rev. Cell Dev. Biol.

1999. 15:551-78

of the latter by binding of aryl hydrocarbons such as dioxin [65]. The identification of ARNT as a subunit of HIF-1 thus represents the first demonstration that ARNT is a common subunit of many different bHLH-PAS heterodimers. HIF-1 is highly conserved evolutionally, with >90% amino acid sequence homology [64].

Regulation of HIF-1

HIF-1 levels increase dramatically under hypoxic conditions [66]. Under normoxic conditions, HIF-1 is subjected to ubiquitination and proteasomal degradation [67] due to the binding of the von Hippel-Lindau (VHL) tumour suppressor protein [68], which is the substrate recognition subunit of an E3 ubiquitin-protein ligase complex [69]. VHL binds to HIF-1 when the latter is hydroxylated on proline residues 402 and/or 564 [70]. The hydroxylation reaction 21 is performed by prolyl hydroxylases (PHDs) that utilise O2 and -ketoglutarate as substrates and generate carbon dioxide and succinate as by-products. Under hypoxic conditions, hydroxylation, ubiquitination and degradation are inhibited, leading to the accumulation of HIF- 1. Under normoxic conditions, asparagine residue 803 is also hydroxylated. This reaction, which is mediated by factor inhibiting HIF-1 (FIH-1), prevents the binding of the co-activators CBP and p300 to HIF-1 [71]. Thus, O2-dependent hydroxylation regulates both the stability and transcriptional activity of HIF-1. PHD activities are known to be iron-dependent

[72]. Therefore cobaltous ions or iron chelators can mimic hypoxia, by chelating iron and inhibiting the prolyl and asparagine hydroxylases [72].

Figure1.5 Regulation of HIF-1α

In VHL-defective cells, HIF-1 subunits are constitutively stabilised and HIF-1 is activated. Re-expression of pVHL restored oxygen-dependent instability [68].

22 pVHL and HIF alpha-subunits co-immunoprecipitate, and pVHL is present in the hypoxic HIF-1 DNA-binding complex. In cells exposed to iron chelation or cobaltous ions, HIF-1 is dissociated from pVHL. These findings indicate that the interaction between HIF-1 and pVHL is iron dependent, and that it is necessary for the oxygen-dependent degradation of HIF-1 subunits [68]. Iron chelation drugs such as deferoxamine (DFO) and deferasirox (DFS) are thus seen as

„hypoxia mimics‟ and are known to increase HIF-1α in cells [34, 73-74].

HIF-1 and mitochondrial function

It is now known that the HIF system is widely operating in most cells and not just restricted to EPO regulation [75]. Critical HIF-1α-binding sites were identified in other genes encoding the glycolytic enzymes phosphoglycerate kinase-1 and lactate dehydrogenase A [76]. Successive studies identified more enzymes involved in this metabolic pathway that are up-regulated by hypoxia, as are glucose transporters and also enzymes of the gluconeogenesis. The metabolism of glucose to CO2 and water is oxygen dependent and very energy efficient.

Glucose is transformed into pyruvate in the cytoplasm and, secondarily, pyruvate is catabolised through the tricarboxylic acid (TCA) cycle and oxidative phosphorylation (OXPHOS) in the mitochondria. In contrast to OXPHOS, O2- deprived cells utilise the less energy-efficient metabolism of pyruvate to lactic acid. Interestingly, most cancer cells rely on this source of energy also in normoxia, as described by Otto Warburg et al [77] more than 80 years ago.

Multiple enzymes responsible for shifting the metabolism toward anaerobic glycolysis are directly controlled by HIF-1α. Since HIF-1α overexpression is a frequent feature of many tumours, the contribution of HIF-1α to this characteristic metabolic phenotype of tumours seems very likely.

HIF-1α has a significant influence on mitochondrial respiration. HIF-1α induces the expression of pyruvate dehydrogenase kinase (PDK), which inhibits the enzyme pyruvate dehydrogenase by phosphorylation. Thus, pyruvate is not converted into acetyl-CoA, preventing pyruvate entry into the TCA cycle [78]. As a consequence, the mitochondrial oxygen-consumption is down-regulated and hypoxic ROS generation is attenuated. HIF-1 also fine-tunes mitochondrial respiration by changing the composition of the cytochrome oxidase complex in hypoxia: a more efficient isoform is up-regulated by HIF-1, and the isoform predominant in normoxia is degraded by the HIF-1α-regulated LON protease

[79]. Finally, in VHL-defective renal carcinoma cell lines, mitochondrial biogenesis and metabolism are actively repressed in a HIF-1-dependent manner

[79]. Thus, HIF-1α modulates key metabolic pathway to optimise glucose and O2 utilisation in hypoxia to generate sufficient amounts of ATP without producing excessive amounts of ROS by inhibition of the TCA cycle and mitochondrial respiration.

HIF-1 and glucose metabolism

It is increasingly recognised that the HIF-1 expression is important in the T2 diabetic phenotype. Some evidence came from studies of the role of ARNT in beta cell function [31]. Using oligonucleotide microarrays and real-time PCR of pancreatic islets isolated from humans with type 2 diabetes versus normal glucose-tolerant controls, Gunton et al identified multiple changes in expression of genes known to be important in beta cell function, including major decreases in expression of HNF4α, insulin receptor, IRS2, Akt2, and several glucose- metabolic-pathway genes [31]. There was also a 90% decrease in expression of the transcription factor ARNT. Reducing ARNT levels in Min6 cells with small interfering RNA (siRNA) resulted in markedly impaired glucose-stimulated insulin release and changes in gene expression similar to those in human type 2 islets.

Likewise, beta cell-specific ARNT knockout mice exhibited abnormal glucose tolerance, impaired insulin secretion, and changes in islet gene expression that mimicked those in human diabetic islets.

ARNT is also found reduced in livers of people with T2 diabetes [32]. To study the functional effect of its reduction, mice were created with liver-specific ablation

(L-ARNT KO) using ARNT loxP mice and adenoviral-mediated delivery of Cre. L-

ARNT KO mice had normal blood glucose but increased fed insulin levels. These mice also exhibited features of type 2 diabetes with increased hepatic gluconeogenesis, increased lipogenic gene expression, and low serum beta- hydroxybutyrate. These effects appear to be secondary to increased expression of CCAAT/enhancer-binding protein alpha (C/EBPα), farnesoid X receptor (FXR),

25 and sterol response element-binding protein 1c (SREBP-1c) and a reduction in phosphorylation of AMPK without changes in the expression of enzymes in ketogenesis, fatty acid oxidation, or FGF21. These results demonstrate that a deficiency of ARNT action in the liver, coupled with that in beta cells, could contribute to the metabolic phenotype of human type 2 diabetes.

STEROL REGULATORY ELEMENT

BINDING PROTEINS

AND

LIPID METABOLISM

SREBP proteins in lipid metabolism

Figure 1.6 Genes regulated by SREBPs. Horton JD et al. JCI May 2002

Lipid homeostasis in vertebrate cells is regulated by a family of membrane-bound transcription factors designated sterol regulatory element–binding proteins

(SREBPs)[80]. SREBPs directly activate the expression of more than 30 genes dedicated to the synthesis and uptake of cholesterol, fatty acids, triglycerides, and phospholipids, as well as the NADPH cofactor required to synthesise these molecules [81]. In the liver, three SREBP isoforms regulate the production of lipids for export into the plasma as lipoproteins and into the bile as micelles.

These are designated SREBP-1a, SREBP-1c, and SREBP-2. SREBP-2 is encoded by a gene on human chromosome 22q13. Both SREBP-1a and -1c are derived from a single gene on human chromosome 17p11.2 through the use of

28 alternative transcription start sites that produce alternate forms of exon 1, designated 1a and 1c [81].

Overview of the SREBP isoforms

The SREBPs play related but distinct roles: SREBP-1c, the predominant SREBP-

1 isoform in adult liver, preferentially activates genes required for fatty acid synthesis, while SREBP-2 preferentially activates the LDL receptor gene and various genes required for cholesterol synthesis. SREBP-1a and SREBP-2, but not SREBP-1c, are required for normal embryogenesis [80]. SREBP-1a and

SREBP-2 are the predominant isoforms of SREBP in most cultured cell lines, whereas SREBP-1c and SREBP-2 predominate in the liver and most other intact tissues [82]. When expressed at higher than physiologic levels, each of the three

SREBP isoforms can activate all enzymes in the biosynthetic pathways used to generate cholesterol and fatty acids. However, at normal levels of expression,

SREBP-1c favours the fatty acid biosynthetic pathway and SREBP-2 favours cholesterol biosynthesis. Overexpression studies indicate that both SREBP-1 isoforms show a relative preference for activating fatty acid synthesis, whereas

SREBP-2 favours cholesterol [80]. SREBP-1c and SREBP-2 activate genes required to generate NADPH, which is consumed at multiple stages in these lipid biosynthetic pathways [83].

SREBP-1a is a potent activator of all SREBP-responsive genes, including those that mediate the synthesis of cholesterol, fatty acids, and triglycerides. SREBP-

1a and SREBP-2 have long transcriptional activation domains, but preferentially activates cholesterol synthesis [81]. Overexpression of SREBP-2 in the liver increases the mRNAs encoding all cholesterol biosynthetic enzymes; the most dramatic is a 75-fold increase in HMG-CoA reductase mRNA [84]. mRNAs for fatty acid synthesis enzymes are increased to a lesser extent, consistent with the in vivo observation that the rate of cholesterol synthesis increases 28-fold in these transgenic SREBP-2 livers, while fatty acid synthesis increases only fourfold. SREBP-1a transgenic mice develop a massive fatty liver engorged with both cholesterol and triglycerides [85], with heightened expression of genes controlling cholesterol biosynthesis and, more dramatically, fatty acid synthesis.

The preferential activation of fatty acid synthesis (26-fold increase) relative to cholesterol synthesis (5-fold increase) explains the greater accumulation of triglycerides in their livers.

SREBP-1c preferentially enhances transcription of genes required for fatty acid synthesis but not cholesterol synthesis. Overexpression of SREBP-1c in the liver of transgenic mice produces a triglyceride-enriched fatty liver with no increase in cholesterol [86]. mRNAs for fatty acid synthetic enzymes and rates of fatty acid synthesis are elevated fourfold in this tissue, whereas the mRNAs for cholesterol synthetic enzymes and the rate of cholesterol synthesis are not increased [83].

SREBP-1c and fatty liver

SREBP-1c mediates the lipogenic actions of insulin in the liver. The liver is responsible for the conversion of excess carbohydrates to fatty acids to be stored as triglycerides or burned in muscle. During times of carbohydrate excess, insulin typically stimulates fatty acid synthesis in liver, an effect mediated by an increase in SREBP-1c. This action is opposed by glucagon, which acts by raising cAMP. A period of fasting suppresses insulin and increases glucagon levels, reducing

SREBP-1c in liver and adipose tissue. This situation is reversed by re-feeding

[87]. Overexpression of SREBP-1c in livers of transgenic mice prevents the reduction in lipogenic mRNAs that normally follows a fall in plasma insulin levels

[87]. Conversely, in livers of Scap knockout mice that lack all SREBPs in the liver

[88] or knockout mice lacking either SREBP-1c [89] or both SREBP-1 isoforms

[90], there is a marked decrease in the insulin-induced stimulation of lipogenic gene expression that normally occurs after fasting/ refeeding. Insulin and glucagon also exert a post-translational control of fatty acid synthesis though changes in the phosphorylation and activation of acetyl-CoA carboxylase (ACC).

The posttranslational regulation of fatty acid synthesis persists in transgenic mice that over-express SREBP-1c [86].

Evidence indicates that the fatty liver of insulin resistance is caused by SREBP-

1c, which is elevated in response to the high insulin levels. SREBP-1c levels are elevated in the fatty livers of obese (ob/ob) mice with insulin resistance and

31 hyperinsulinaemia caused by leptin deficiency [91]. Despite the presence of insulin resistance in peripheral tissues, insulin continues to activate SREBP-1c transcription and cleavage in the livers of these insulin-resistant mice. The elevated nSREBP-1c increases lipogenic gene expression, enhances fatty acid synthesis, and accelerates triglyceride accumulation [91]. These metabolic abnormalities are reversed with the administration of leptin, which corrects the insulin resistance and lowers the insulin levels [92]. In addition, metformin, by stimulating AMP-activated protein kinase (AMPK), an enzyme that inhibits lipid synthesis and fatty acid oxidation, lowers lipid accumulation in livers of insulin- resistant ob/ob mice by phosphorylation and inactivation of key lipogenic enzymes, including SREBP-1 and ACC [93-94].

SREBP-1a and dyslipidaemia

The incidence of coronary artery disease increases with increasing plasma LDL- cholesterol levels, which in turn are inversely proportional to the levels of hepatic

LDL receptors. SREBPs stimulate LDL receptor expression, but they also enhance lipid synthesis [81], so their net effect on plasma lipoprotein levels depends on a balance between opposing effects. In mice, the plasma levels of lipoproteins tend to fall when SREBPs are either overexpressed or underexpressed. In transgenic mice that overexpress SREBPs in liver, plasma cholesterol and triglycerides are generally lower than in control mice, even though these mice massively overproduce fatty acids, cholesterol, or both.

Hepatocytes of SREBP-1a transgenic mice overproduce VLDL, but these particles are rapidly removed through the action of LDL receptors, and they do not accumulate in the plasma. Indeed, some nascent VLDL particles are degraded even before secretion by a process that is mediated by LDL receptors

[95]. The high levels of SREBP-1a in these animals support continued expression of the LDL receptor, even in cells whose cholesterol concentration is elevated. In LDL receptor–deficient mice carrying the SREBP-1a transgene, plasma cholesterol and triglyceride levels rise tenfold [96]. Mice that lack all

SREBPs in liver also manifest lower plasma cholesterol and triglyceride levels. In these mice, hepatic cholesterol and triglyceride synthesis is markedly reduced, and this likely causes a decrease in VLDL production and secretion. LDL receptor mRNA and LDL clearance from plasma is also significantly reduced in these mice, but the reduction in LDL clearance is less than the overall reduction in VLDL secretion, the net result being a decrease in plasma lipid levels [97].

However, because humans and mice differ substantially with regard to LDL receptor expression, LDL levels, and other aspects of lipoprotein metabolism, it is difficult to predict whether human plasma lipids will rise or fall when the SREBP pathway is blocked or activated.

Regulation of SREBP genes

SREBPs belong to the basic helix-loop-helix–leucine zipper (bHLH-Zip) family of transcription factors synthesised as inactive precursors bound to the

33 endoplasmic reticulum (ER) [81]. To reach the nucleus, the NH2-terminal domain of each SREBP is cleaved by the SREBP cleavage–activating protein (SCAP),

Site-1 protease (S1P) and Site-2 protease (S2P) [88]. SCAP is both an escort for

SREBPs and a sensor of sterols. When cells become depleted in cholesterol,

SCAP escorts the SREBP from the ER to the Golgi apparatus, where S1P, a membrane-bound serine protease, first cleaves the SREBP in the luminal loop between its two membrane-spanning segments. The NH2-terminal bHLHZip domain is then released from the membrane via a second cleavage mediated by

S2P, a membrane-bound zinc metalloproteinase. The NH2-terminal domain, designated nuclear SREBP (nSREBP), translocates to the nucleus, where it activates transcription by binding to non-palindromic sterol response elements

(SREs) in the promoter/enhancer regions of multiple target genes [81]. When the cholesterol content of cells rises, SCAP changes its conformation so that the

SCAP/SREBP complex is no longer incorporated into ER transport vesicles.

Regulation of SREBPs occurs at transcriptional and post-transcriptional levels

[80]. The post-transcriptional regulation involves the sterol-mediated suppression of SREBP cleavage, which results from sterol-mediated suppression of the movement of the SCAP/SREBP complex from the ER to the Golgi apparatus.

SREBP-1c and SREBP-2 are subject to distinct forms of transcriptional regulation, whereas SREBP-1a appears to be constitutively expressed at low levels [82]. SREBP-1c and SREBP-2 can be regulated by a feed-forward regulation mediated by SREs present in the enhancer/promoters of each gene

[98-99]. Through this 6m -forward loop, SREBPs activate the transcription of their own mRNAs. In contrast, when SREBPs decline, as in Scap or S1p knockout mice, there is a secondary decline in the mRNAs encoding SREBP-1c and

SREBP-2 [88, 97].

Liver X-activated receptors (LXRs) regulate SREBP-1c transcription

LXRα and LXRβ, nuclear receptors that form heterodimers with retinoid X receptors, are activated by a variety of sterols, including oxysterol intermediates that form during cholesterol biosynthesis [100]. An LXR-binding site in the

SREBP-1c promoter activates SREBP-1c transcription in the presence of LXR agonists [101]. Mice that lack both LXRα and LXRβ express reduced levels of

SREBP-1c and its lipogenic target enzymes in liver and respond relatively weakly to treatment with a synthetic LXR agonist [101]. LXR activates SREBP-1c, and induces synthesis of oleate when sterols are in excess [101]. Oleate is the preferred fatty acid for the synthesis of cholesteryl esters, which are necessary for both the transport and the storage of cholesterol. LXR-mediated regulation of

SREBP-1c appears also to be one mechanism by which unsaturated fatty acids suppress SREBP-1c transcription and thus fatty acid synthesis. Rodents fed diets enriched in polyunsaturated fatty acids manifest reduced SREBP-1c mRNA expression and low rates of lipogenesis in liver [102]. In vitro, unsaturated fatty acids, e.g. arachidonic acid, a n-6 polyunsaturated acid, competitively blocked

LXR activation of SREBP-1c expression by antagonising the activation of LXR by

35 its endogenous ligands [103]. In addition to LXR-mediated transcriptional inhibition, polyunsaturated fatty acids lower SREBP-1c levels by accelerating degradation of its mRNA [104]. These combined effects may contribute to the long recognised ability of polyunsaturated fatty acids to lower plasma triglyceride levels.

Fatty acid synthase (FAS)

Whereas cholesterol synthesis depends almost entirely on SREBPs, fatty acid synthesis is only partially dependent on these proteins. In the absence of SREBP processing, as when the Site-2 protease is defective, the levels of mRNAs encoding cholesterol biosynthetic enzymes and the rates of cholesterol synthesis decline nearly to undetectable levels, whereas the rate of fatty acid synthesis is reduced by only 30% [105]. Under these conditions, transcription of the fatty acid biosynthetic genes must be maintained by factors other than SREBPs. In liver, the gene encoding fatty acid synthase (FAS) can be activated transcriptionally by upstream stimulatory factor, which acts in concert with SREBPs [106]. The FAS promoter also contains an LXR element that permits a low-level response to LXR ligands even when SREBPs are suppressed [107]. These two transcription factors may help to maintain fatty acid synthesis in liver when nSREBP-1c is low.

Peroxisome proliferator-activated receptor gamma (PPAR)

PPAR-γ, a member of the nuclear hormone receptor super-family is required for normal adipocyte differentiation and is important in the development of hepatic steatosis [108]. Normally, PPAR-γ is expressed at very low levels in the liver. In animal models with fatty livers, the expression of PPAR-γ is markedly increased.

SREBP-1c can activate transcriptionally PPAR-γ by stimulating production of an activating ligand for the nuclear receptor [109]. The importance of PPAR-γ expression in the development of fatty livers has been demonstrated in insulin- resistant mouse models in which the specific gene deletion of PPAR-γ attenuates the development of hepatic steatosis, independently to the presence of hyperinsulinaemia or hyperglycaemia [110]. In human NAFLD, PPAR-γ is transcriptionally upregulated and consequently activates lipogenic enzymes and exacerbates steatosis [110].

HYPOXIA INDUCIBLE FACTOR

-1 ALPHA

AND

LIPID METABOLISM

Hypoxic adaptation

Studies have hypothesised that HIF-1 could modulate metabolic syndrome by up regulating some proteins deeply involved in cholesterol metabolism [111]. During hypoxia, liver and adipose tissue adapt by reducing protein synthesis and lipogenesis [112]. In adipocytes, the exposure of hypoxia reduces the content of triglyceride and cholesterol. This is due to reduced expression of PPAR [113].

Chinese Hamster Ovary cells (CHO-7) under hypoxic conditions show a higher 3- hydroxy-3-methylglutaryl coenzyme A reductase (HMG-CoAR) degradation, the key enzyme of the cholesterol biosynthetic pathway that catalyses reduction of

HMG-CoA to mevalonate (MVA). Hypoxia reduced FAS mRNA in Hep3B human hepatocytes, L6 mouse skeletal myocytes, C2C12 mouse myoblasts, 3T3-L1 mouse preadipocytes and Hepa1c1c7 mouse hepatoma cells. In contrast, transcripts of hypoxia-inducible genes such as phosphoglycerate kinase-1 (PGK-

1) and glyceraldehydes-3-phosphate dehydrogenase (GAPDH) increased. After

16 h of hypoxic exposure, FAS mRNA was reduced by 20% in Hepa1c1c7 cells, but the ATP level was unaffected [114]. These findings imply that cells can shut down anabolic genes prior to an actual reduction of ATP, ultimately by reducing the levels of anabolic enzymes such as FAS.

HIF is involved in hypoxic repression of FAS and SREBP-1c

Hypoxia failed to reduce mRNA of either FAS or SREBP-1c in the HIF-1β defective cells, indicating that HIF-1β is required for their repression by hypoxia

[35]. Hypoxic treatment also failed to repress FAS mRNA and SREBP-1c protein in the HIF-1α-knockdown 3T3-L1 cells. HIF-1α and 2α induce HRE-dependent expression of same target genes with different temporal patterns [115]. Small inhibitory RNAs (siRNAs) against HIF-1α and HIF-2α were transfected into

Hepa1c1c7 cells. Western analysis showed that HIF-1α siRNA but not HIF-2α siRNA restored the expression of SREBP-1c protein, which was repressed by acute hypoxic exposure (4 h).

Stra13/DEC1 and DEC2 are involved in hypoxic repression of SREBP-1c and FAS

Yun et al. [36] showed that Stra13, a hypoxia-induced transcription repressor family, represses PPARγ2 promoter and functions as a mediator of hypoxic inhibition of adipogenesis. Stra13 is also referred to as Differentiated embryo chondrocyte 1 (DEC1). Stra13/DEC1 and its isoform DEC2 are class B type homodimeric bHLH proteins. Both Stra13 and DEC2 homodimers are able to bind the E-box sequences [116] playing key roles in cell differentiation, circadian rhythms, immune regulation and carcinogenesis. Northern analyses confirmed that hypoxia increases mRNA level of Stra13/DEC1 prior to maximum decrease

40 of FAS expression. By using HIF-1β defective cells, Stra13 induction was confirmed to be HIF-1-dependent. Consistent with the case of SREBP-1c, in acute (4 h) hypoxic exposure, siRNA against HIF-1α reduced the hypoxic induction of Stra13, whereas in prolonged hypoxic exposure (24 h) HIF-2α siRNA was more effective than HIF-1α siRNA. Even in normoxic cells, forced expression of either HIF-1α or Stra13 reduces the amount of the endogenous

SREBP-1c protein and transcriptional activity [115]. HIF-1α and Stra13 repress the activity of FAS promoter. ChIP assays confirmed that SREBP-1c interacts with Stra13 in vivo, but not as much with HIF-1α [116].

Therefore, HIF-1 induces Stra13, with Stra13 then interacting with the E-box sequence in the SREBP-1c promoter and/or with the SREBP-1c protein itself, preventing SREBP-1c from binding to its recognition site. The presence of both

Stra13 and HIF-1α/β prevents SREBP-1c from binding to the FAS promoter. Both

Stra13/DEC1 and DEC2 are induced by hypoxia. The mRNA expression of

Stra13/DEC1 gradually increased and reached the maximum after 8-h exposure to hypoxia, while that of DEC2 instantly and temporarily increased during acute hypoxic exposure (1–2 h). Li et al. [117] had showed that Stra13/DEC1 represses the expression of DEC2 through binding to E-box in DEC2 promoter. Hypoxic induction of DEC2 also depends on the HIF-1 [116]. Over-expression of DEC2 reduced SREBP-1c promoter activity, FAS promoter activity, and the expression level of the endogenous SREBP-1c protein. Similar to Stra13, ChIP assay showed that DEC2 also interacts with SREBP-1c. Treatment of DEC2 siRNA

41 recovers the hypoxic repression of both FAS and SREBP-1c. Accordingly, hypoxic repression of FAS and SREBP- 1c is regulated by both Stra13 and

DEC2; however, DEC2 plays more pivotal role in hypoxic repression of SREBP-

1c and FAS.

Orexin signals via HIF-1 activity

Array analysis of transfected HEK293 cells expressing the orexin 1 receptor

(OXR1) treated with Orexin A revealed that among the transcripts that were up- regulated were genes that encoded transcription factors, repressors and corepressors, GPCRs, GTPases, endocytosis machinery ion channels, proteins involved in cell cycle and cell proliferation, apoptosis, signal transduction, and metabolism [118]. More than half of the responsive genes fell into two categories: cell growth (30%) and metabolism (27%). These are followed closely by genes that play a role in cell–cell communication. HIF-1α mRNA, in particular, was up- regulated by 15 fold. Western blot analysis showed that induction of HIF-1α protein by addition of orexin to HEK cells but not when the cells were incubated with anti-Orexin R1 antibody, confirming that HIF-1 induction was mediated via orexin binding to its receptor [118]. When mice hypothalamic cells were subjected to hypoxia for 6 h, or treated with orexin for 2 h, and the levels of several HIF-1-regulated transcripts were upregulated four- to eightfold, notably

VEGF, GLUT1, VEGFR3, eNOS genes. However in addition to up-regulating

HIF-1α gene transcript, it appears that orexin also increases HIF-1 activity by

42 decreasing the rate of protein degradation through reduction of VHL protein

[118]. This was demonstrated using quantitative RT–PCR and Western blotting, showing reduction in VHL mRNA levels and protein when orexin was added to the hypothalamic cells. These changes were abolished when the cells were pre- treated with anti-Orexin R1 antibody.

Orexin-stimulated glucose utilisation is mediated through HIF-1α

Orexin increases glucose uptake in HEK293 cells. Using small interfering RNA

(siRNA) to knock down HIF-1α expression abolishes this effect. Orexin increases mRNA of GLUT1, and glycolysis gene expression, all known to be also increased by HIF-1α [118]. However, the orexin-mediated increase in glucose utilisation resulting in increased cellular energy production occurs via shunting pyruvate preferentially into the tricarboxylic acid (TCA) cycle independent of HIF-1 activity.

Glucose enters glycolysis pathway, generating pyruvate as an end product.

Pyruvate can either be reduced to lactate by NADH and lactate dehydrogenase

(LDHA), thus facilitating anaerobic glycolysis, or be converted irreversibly to

Acetyl CoA by the pyruvate dehydrogenase (PDH) complex, priming the TCA cycle and oxidative phosphorylation in the mitochondria. In hypoxic cells, HIF-1 plays a major role in the pyruvate-to-lactate conversion. LDHA is known to be a

HIF-1 target gene [119]. Moreover, it has been found recently that PDH kinase

(PDK1), a negative regulator of PDH activity, is also a HIF-1-regulated gene [78,

120]. Thus, in hypoxic cells, HIF-1 actively promotes LDHA function and suppresses PDH activity to shift metabolic flux from oxidative phosphorylation to anaerobic glycolysis [121]. However in normoxic neurons, oxidative phosphorylation is a more efficient way to generate ATP from glucose. It was shown that orexin treatment of hypothalamic neuronal cells results in a modest down-regulation of LDHA and a larger decrease in PDK1 expression, as well as a 3.5-fold increase in PDH message levels resulting in a situation favouring shunting pyruvate through the TCA cycle. Similarly, hypothalamic cells derived from the knockout mice have much higher levels of PDK1 and LDHA mRNA than the wild-type mice, consistent with the idea that the presence of the orexin receptor down-regulates the expression of these genes even without added exogenous orexin.

IRON TRANSPORT,

PHYSIOLOGY

AND

CHELATION

Iron uptake and transport

Iron is taken up in the duodenum via the Divalent Metal Transporter (DMT-1)

[122]. The DMT-1 protein is a classic transmembrane pump, increased in iron deficiency. It transports a number of divalent metals, including ferrous, but trivalent ions such as ferric iron are not transported. The ferric iron in the diet needs reduction to ferrous iron prior to transport by DMT-1, and this may be accomplished by a membrane cytochrome containing ferrireductase (reducing ferric iron, Fe(III), to ferrous iron, Fe(II)). This enzyme is increased in iron deficiency, but it has not been shown to have a direct role in transport. Ferric iron also requires reduction to ferrous iron to serve as a substrate for ferrochelatase.

This may be accomplished by a ferrireductase found in a large protein complex called paraferritin[122-123]. Under normal conditions DMT-1 is largely found in the cytoplasm and lamina propria, but is brought to the luminal surface in iron deficiency. In iron deficiency, DMT-1 is localised in the apical membrane, but on re-feeding with iron, it internalises as vesicles. The transport proteins scavenge metals in the lumenal space and return to the membrane along a mucin track

[124].

Once across the lumenal membrane of the duodenum, iron transits the cytoplasm via a carrier protein chelated in the ferric form. At the basolateral membrane, the iron is incorporated into transferrin and released to the plasma.

Hephaestin, a ceruloplasm-like ferro-oxidase, and a transport protein ferroportin

46 is involved in the exodus of iron from the intestinal mucosa [125]. Hephastin is a multi-copper enzyme that can oxidise ferrous(II) to ferric(III) for incorporation into transferrin that binds only the ferric form. Another protein, ferroportin [126], appears to be the actual transmembrane transporter. Hephaestin and ferroportin appear to respond to systemic rather than local (i.e. duodenal) iron levels and may exercise control of iron uptake at that stage.

Iron and mitochondrial function

Mitochondria supply energy for all cell compartments. Besides this important function, these organelles exhibit many other activities that are fundamental for cell life. They generate the regulators of cellular redox potential, such as superoxide anions, hydrogen peroxide, nitric oxide, peroxynitrite etc., which control proteolysis, activation of transcription, cell metabolism, and differentiation

[127]. They are responsible for the synthesis of important compounds like steroids, heme, and iron–sulfur clusters [128]. With their filamentous structure they transport, in addition to energy, signalling molecules and lipophilic compounds into the cell [128]. Mitochondria also play an important role in calcium signalling and regulation of apoptotic cell death [129]. Thus, they are able to control the development or death of the cell.

The mitochondrion is the site where iron is transformed into its bioactive form by the heme and iron–sulfur cluster (ISC or Fe/S) biosynthetic pathways [128].

These cofactors are responsible for the activity of several enzymes involved in many metabolic reactions [127]. Thus, this organelle is a major user of cellular iron, plays a central role in iron metabolism and, similarly to the cell, relies on iron transport, storage, and regulatory proteins to maintain iron homeostasis [130].

Iron transport mechanisms across mitochondrial membranes have evolved to satisfy mitochondrial iron requirements and preserve the balance of the cytosolic compartment [130]. Furthermore, since mitochondria are the major site of oxygen consumption and iron is also a potent inducer of ROS formation, the simultaneous presence of high concentrations of oxygen and iron may be detrimental to the organelle.

Free iron is known to enhance ROS production, as it is a substrate of the Fenton reaction. Thus, it is important that mitochondrial iron be maintained in a bio- available and safe form to limit oxidative damage. A form of labile iron, defined as

“chelatable iron”, is present in the matrix of mitochondria. This form of iron is potentially redox-active, and thus its level is expected to be extremely low in mitochondria. This can be achieved by the tight coordination of the rate of influx with the rate of incorporation into heme and iron–sulfur clusters.

Iron chelators: pharmacology

The principal goal of chelation therapy involves preventing the accumulation of iron reaching harmful levels in the body [131]. The chances of achieving this goal

48 can be maximised by starting chelation treatment before iron accumulation is excessive [132]. Iron (III) has six coordination sites for iron chelation, and for complete coordination requires either one molecule possessing six coordination sites (hexadentate chelation e.g. deferoxamine), two molecules each possessing three coordination sites (tridentate chelation e.g. deferasirox), or three molecules each possessing two coordination sites (bidentate chelation e.g. deferiprone)

[131]. Storage iron present as ferritin or hemosiderin is not available for direct chelation over a biologically useful timescale [133] so chelators must act so as to bind rapidly chelatable iron pools that are typically rapid-turnover pools. The most important of these pools, quantitatively, are iron released from macrophages after the breakdown of old red blood cells [134], and iron released by the breakdown of ferritin in lysosomes [134]. These chelated pools are primarily responsible for iron excreted by chelators in the urine or feces. With deferoxamine, urine iron accounting for about half of excreted iron, is derived from red blood cell breakdown, whereas faecal iron is derived from hepatocellular chelation of lysosomally degraded ferritin [134]. With deferiprone, the chelated pools are the same, but both are excreted in the urine with little faecal excretion

[135]. With deferasirox, the chelated turnover pools are again the same, but iron is excreted almost entirely in faeces [136].

An important goal of chelation treatment involves preventing tissue damage from labile iron pools that may be present in plasma as Non-Transferrin Bound Iron

(NTBI) species[137], or as labile iron within cells. As these pools are constantly

49 being turned over, continuous exposure to chelation is theoretically desirable, both in terms of maximising the efficiency of chelation therapy, as well as minimising the exposure to harmful free radicals generated by labile iron species.

The overall goal is to find a balance between the effects of too much iron and too much chelator [138]. Different chelators may have distinct modes of access to labile iron pools in certain cell types not usually critical to iron turnover but where iron-mediated toxicity may occur. This may result in variable rates of removal between such tissues, as well as different patterns of potential toxicity from over- chelation.

Chelators may also act by preventing the uptake of iron into cells. As there is an increased NTBI uptake into parenchymal cells in transfusional iron overload

[139], this is an important target for chelation therapy. NTBI is constantly being turned over, reappears rapidly after chelators are cleared from the circulation

[137], and is incompletely removed by standard chelation regimes with subcutaneous deferoxamine. A chelation regime that provides 24-hr cover from

NTBI uptake is desirable [131]. Although the conventional NTBI assay is applicable to measuring NTBI in the presence of deferoxamine, because the iron complex of deferoxamine is stable under conditions of the assay, this is not the case with deferiprone or deferasirox. A modified assay has therefore been developed [140] that detects a subfraction of NTBI that is redox active, termed labile plasma iron (LPI). Although it is not yet known whether the remaining NTBI species (that are not redox active) are taken up as rapidly into cells as the

50 species detected as LPI, the assay has the advantage of being usable in the presence of chelators such as deferiprone and deferasirox. Using this assay, this fraction is effectively removed by standard 8–10 hr deferoxamine subcutaneously at night, but LPI is present during the day [140]. Conversely, with deferiprone LPI is partially removed during the day, rebounding between doses with high levels of

LPI at night. By giving deferoxamine at night and deferiprone during the day, LPI is efficiently removed over a 24-hr period [140]. This is a demanding regime however, and 24-hr cover can be more simply achieved using once-daily deferasirox. With repeated once-daily dosing of deferasirox at 20 mg/kg, mean trough levels of deferasirox are more than 20 µmol/L [141]. A progressive removal of LPI with time using this regime has been reported [131].

The efficiency of deferoxamine is about 13%, deferiprone is about 4%, and deferasirox is about 27% [142]. The efficiency of chelation therapy can be calculated from the change in liver iron concentration over time (and hence the change in total body iron) [143], from the transfusional iron loading rate, and from the iron binding equivalents of the chelator that were given over the time period studied. The high efficiency of chelation with deferasirox is the result of several factors including the continuous chelation achieved with once-daily oral treatments [141], the high-affinity constant for iron, and the favourable biodistribution properties [144]. A subanalysis of the pivotal phase III studies

[145] on deferasirox, and analysis of the efficiency of chelation over the dose- ranges studied [5–30 mg/(kg day)] was undertaken [131]. It was found that the

51 efficiency of chelation was more or less constant over this range of doses and over the range of liver iron concentration (LIC) values that were studied.

In summary, while whole body iron is important for metabolism, and in particular, mitochondrial function, it is unclear what the optimum iron level for normal cell physiology and metabolism is. Certainly, individuals with iron overload syndromes are more likely to develop insulin resistance and beta cell failure with subsequent development of diabetes[146] . Iron deposition in other organs such as the heart and pituitary gland can lead to organ dysfunction [147]. Iron chelation prolongs survival and improves organ function in these individuals

[147]. Moreover, iron, as measured using serum ferritin, is shown to be independently associated in individuals with obesity and the metabolic syndrome

[22] . What is unclear, however, is the role of iron in non-iron overloaded individuals with obesity and the metabolic syndrome.

The following experiments in this thesis will examine the effect of iron chelation on murine whole-body physiology, specifically looking at glucose and lipid metabolism. Two forms of iron chelators were used: oral deferasirox (DFS) and intra-peritoneal deferoxamine (DFO) injections, with differing pharmacology and bioavailability (as discussed in the literature review). The results suggest that modest iron chelation has beneficial effects on limiting weight gain from continuous high fat feeding, associated with reduction of hepatic steatosis, dyslipidaemia, and glucose intolerance. Iron chelation is shown to increase HIF-

1α levels and preliminary results support the role of HIF-1α in mediating the beneficial “Iron Chelator” effects from reduced lipid gene expression and improved gene expression for glucose handling. Importantly, iron chelation is shown to be well tolerated in the mice, as they did not develop anaemia or other forms of toxicities.

CHAPTER 2

Research methods

and experimental protocols

Animal care

All animals were housed and cared for in the Biological Testing Facility (BTF) at the Garvan Institute of Medical Research, Sydney. Animal procedures were carried out in accordance with the guidelines set out by the St Vincent‟s Hospital and Garvan animal ethics committee. Mice were in-bred within the BTF or purchased from Animal Resources Centre (Western Australia) and housed 1 to 5 per cage in rooms maintained at constant temperature (25C) with a 12hour dark cycle each day. Upon completion of studies, the mice were sacrificed for tissue harvesting.

Mice models

The effect of iron chelation therapy was studied in 5 mice models. C57Bl/6 mice are well-known models for the study of insulin resistance. When they are fed continuously on high fat diet, they serve as models of diet-induced obesity [148].

The murine models of insulin resistance is affected by sexual dimorphism: males have a stronger phenotype compared to females. Non-obese diabetic (NOD) mice are a well-known model for the study of autoimmune Type1 diabetes. The iron chelating drugs used were deferasirox (DFS), an oral drug or another injectable iron chelator, deferoxamine (DFO). Initial „power‟ calculations estimated that thirty mice in each group (sixty in total) would enable significant treatment effects in glucose and weight changes. However subsequent

55 experiments using much smaller numbers gave significant results, so smaller numbers were used instead.

1. Male WT C57Bl/6 mice on continuous high fat diet +/-DFS

These mice were studied for 10 weeks in 3 experiments (Exp 1 to 3) and to 25 weeks in Exp4. These mice were randomised to receive either normal HFD or

HFD with DFS mixed-in.

2. Female WT C57Bl/6 mice on continuous high fat diet +/-DFS

Corresponding female littermates were studied in another experiment (Exp5) over 34 weeks. These mice were also randomised to receive either normal HFD or HFD with DFS mixed-in.

3. Male ob/ob (C57Bl/6) mice on blended normal chow diet +/- DFS

These mice were studied over 25 weeks. They were randomised to receive either normal chow or chow with DFS powder mixed in.

4. Male WT C57Bl/6 mice on blended normal chow diet +/-DFS

These mice were studied over 25 weeks. They were randomised to receive either normal chow or chow with DFS powder mixed in.

5. Male WT C57Bl/6 mice on continuous high fat diet +/-DFO

These mice were studied for 10 weeks in 2 experiments. They were randomised to receive either DFO intra-peritoneal injections or control saline intra-peritoneal injections.

6. Female NOD mice on unblended normal chow +/- DFO.

These mice were studied for 10 weeks in 2 experiments. They were randomised to receive either DFO intra-peritoneal injections or control saline intra-peritoneal injections.

Preparation of mice diets

High fat diet

The high fat diet was prepared according to Rodent Diet D12451 from Research

Diets Incorporated. The ingredients of the dry mineral mix are as follows: casein

261 grams, sucrose 230 grams, starch 193 grams, bulking mineral mix 51 grams, trace minerals (including iron, iodine, zinc, copper, magnesium, manganese, etc)

14.8 grams, bran 57 grams, methionine 3.4 grams, gelatine 23 grams, choline bitartrate 4.6 grams. 29.6 grams of AIN Vitamin Mix 76A (ICN Biomedicals Inc) was added and mixed in thoroughly, before further addition of 68grams of safflower oil and 500 grams of melted Allowrie lard. The whole preparation was then left to harden overnight in the refrigerator to form high-fat diet, This forms the normal control HFD (CON). Mice were fed the assigned diet ad libitum, delivering 45% of calories from fat. The treatment HFD diet consists of powdered

57 deferasirox (DFS) (Exjade, Novartis Pharmaceuticals) added to normal HFD. The treatment dose used was similar to the maximum daily human dose for therapeutic iron chelation, i.e. 30mg/kg/day [149]. In order for the mice to receive such a dose, the mice were initially fed normal HFD and the rate of consumption of HFD per cage of five mice per day was first determined. This worked out that approximately one and a half Exjade tablets (750 mg) had to be were finely ground using a mortar and pestle and mixed in thoroughly with the dry mineral/vitamin preparation prior to the addition of melted lard, to achieve the desired dose of 30mg/kg/day. When this process was repeated in 4 weeks, it was noted that the daily consumption of HFD per cage was similar. Therefore the

DFS (treatment diet) was prepared in the same way throughout the study. The mice were fed ad libitum with blocks of HFD (normal and HFD+DFS). As mice are untidy feeders, there would be some diet littered among the bedding.

Measures were taken to ensure that mice were never without feed and as most of the HFD were consumed per cage of nice within 3 days, fresh HFD blocks were put in every 3 days.

Blended chow diet

Normal mice chow pellets were ground finely using a large blender and powdered DFS was mixed in thoroughly. The daily consumption rate per cage was determined in the beginning of the study, as was in the HFD studies. To achieve a dose of at least 30 mg/kg/day, 220 mg of powdered Exjade was added to 280 g of blended chow and the mice were fed ad libitum. Control mice were

58 fed normal blended chow. Blended chow powder were placed in large petri dishes and placed in the cages., the petri dishes were changed and fresh feed was replaced on a daily basis, to ensure that there were adequate feeding and dosing of DFS powder.

In-vivo techniques and rationale for time-points

Tests were carried out after a certain period of observation, to allow mice time to adjust to their diet and to allow for metabolic changes to take place. This often manifests as significant weight changes. The various in-vivo tests were carried in weekly succession, to allow time for the mice to recover in-between tests. Some tests were carried out at time points prior to significant weight change to serve as comparison. In ob/ob mice, glucose tolerance tests were carried out soon after randomisation to serve as baseline, against which a later GTT can be compared with. In some experiments, mice were assessed between 5 to 6 months to study the late effects of treatment.

Glucose tolerance tests (GTT)

Glucose tolerance testing was performed in conscious un-restrained mice following a 4-hour fast. For fasting, mice were placed into clean cages to avoid any powdered food mixed in their bedding and water was available as usual.

They were weighed at study commencement. At time 0, a tail nick was made using a scalpel and blood glucose was measured using an Accucheck

59 glucometer (Roche, Basel, Switzerland). Immediately after that, mice received an intra-peritoneal injection of 2 g/kg glucose delivered as 20% dextrose. Glucose was then measured at time-points 15, 30, 60, 90 and 120 minutes. Plasma glucose concentrations were recorded as mmol/L and when the glucometer read

“Hi”, it was recorded as 32 mmol/L. At the completion of the study, the study diet was re-introduced.

Insulin tolerance test s(ITT)

Insulin tolerance testing was performed in conscious un-restrained mice following a 4-hour fast. Mice were placed into clean cages to avoid any powdered food mixed in their bedding and water was available as usual. They were weighed at study commencement. At time 0, a tail nick was made using a scalpel and blood glucose was measured using an Accucheck glucometer (Roche, Basel,

Switzerland). Immediately after that, mice received an intra-peritoneal injection of human short acting insulin (Actrapid, NovoNordisk). Because of differences in insulin sensitivity in the different studies, different doses of insulin were used, as indicated where the results are discussed. Glucose was then measured at time- points 10, 20, 30, 45 and 60 minutes. If glucose concentration fell to <2.0 mmol/L, 0.5 mls of intra-peritoneal 20% dextrose was given and no further values were obtained for that animal. At the completion of the study, the study diet was re-introduced.

Glucose stimulated insulin secretion (GSIS)

Glucose stimulated insulin secretion was performed in conscious un-restrained mice following a 4-hour fast. Mice were placed into clean cages to avoid any powdered food mixed in their bedding and water was available as usual. They were weighed at study commencement. At time 0, a tail nick was made using a scalpel and 3-4 drops of blood were collected into a 1.5 ml eppendorf test-tube with 10 µL of EDTA and placed on ice. Immediately after that, mice received an intra-peritoneal injection of 3 g/kg glucose delivered as 20% dextrose. 3-4 drops of blood were again collected similarly at time-points of 2, 5, and 20 minutes.

Samples were placed on ice and centrifuged as soon as possible to obtain the supernatant, which was stored at -80 degrees C for insulin assay. At the completion of the study, the study diet was re-introduced.

Food intake studies

These were carried over 3 days. Mice were housed separately in clean cages without bedding. Instead 2 clean protowels were used to line the bottom. Water was provided as usual. HFD with and without DFS added or chow with or without DFS were studied separately. Blended chow powder was placed in large

Petri-dishes. The weight of the supplied diet was noted at the beginning of study and residual food was weighed at 24 hr intervals including food droppings. The mice and their faecal droppings were carefully weighed at 24-hour intervals.

Intake was calculated as follows:

(Original diet weight - final diet weight + diet droppings) / mouse weight at beginning of 24 hr interval

Oxymax respirator studies

Indirect calorimetry was performed using the Oxymax System (Columbus

Instruments, Columbus, OH). Measurements were taken with a 12-hour light cycle / 12-hour dark cycle. Temperature was kept constant at 25C. Mice were housed separately in closed circuit chambers with provision of oxygen and carbon dioxide removal. The chambers were fitted with sensors to measure differences in oxygen (consumption used to calculate VO2) and carbon dioxide

(production used to calculate VCO2), and activity (beam interruption in 3- dimensions). Usual diet (HFD or chow) was provided according to the diet for that experiment along with water. Animals were allowed to acclimatize for 4 hours prior to measurement. VO2 and VCO2 were measured in ml/kg/min. The respiratory exchange ratio (RER) was calculated from the ratio of VCO2 and VO2.

Heat production or energy expenditure (EE) in Kcal was calculated using the formula: 3.9VO2 +1.1VCO2. VO2, CO2 and was adjusted for both total body mass as well as lean body mass.

DEXA scanning

Body composition was assessed using dual energy x-ray absorptiometry (DEXA) densitometer (GE Lunar PIXImus, Madison, USA) housed in the BTF. Mice were anaesthetiised with 2% Avertin (Tribromoethanol) prior to diagnostic testing.

Scanning was performed as per manufacturer‟s protocol. The following data was obtained: lean mass, fat mass, bone mineral content (BMC) and bone mineral density (BMD). Mice were allowed to recover on heated pads for up to 4 hours prior to placing back in usual cages.

Histological techniques

Tissue harvest

Mice were usually fasted for 4-6 hours from early morning prior to sacrifice.

Cardiac punctures were performed after the mice were deeply anaesthetised with intra-peritoneal Avertin. At least 500L of whole blood was collected. Whole blood was obtained fresh for full blood count and haemoglobin analysis. The rest was spun down and supernatant was aspirated and stored at -80C. Liver, pancreas, fat and muscle tissues were obtained via sharps dissection and stored in formalin or snap frozen for later analysis.

Histological preparation

Tissues were fixed in 10% buffered formalin, embedded in paraffin and sections were cut at 6 micron widths. Slides were allowed to dry overnight at room temperature prior to staining. They were then cut in 5-micrometer widths and stained using a standard Haematoxylin and Eosin protocol

Haematoxylin and Eosin (H&E) staining

Slides were initially de-waxed through 30 minutes of heating in an oven at 70C followed by immersion in xylene for 3 to 5 minutes. They were then progressively re-hydrated with 1 minute exposure to 95% ethanol, 70% ethanol and finally in de-ionized water (dH2O) respectively. They were then placed in Mayer‟s haematoxylin solution for 30 seconds followed by 1 minute dH2O wash. After a quick 70% ethanol wash (10 seconds) the sides were placed in eosin-Y for 1 minute.

Cover-slipping

Slides were progressively dehydrated by being placed in sequential 95% and

100% ethanol washes, followed by xylene treatment for 5 minutes. A glass cover slip was carefully placed over the sections and left in a fume hood overnight.

100X and 200X magnification views were obtained from the Leica microscope.

Perls‟ protocol

To visualize hepatic iron, Perl‟s staining was used. Mounted, de-paraffinised slides were soaked for 20 minutes in a solution consisting of equal volumes of

20% HCL and 10% K4Fe(CN)6 and counter-stained with Nuclear Fast Red

Solution (DAKO), followed by cover-slipping.

Sirius Red staining

To visualise liver collagen, liver sections were prepared in the usual way ( see

Histological Preparation) and mounted onto slides. They were then placed in a

64 glass dish containing saturated solution of picric acid in distilled water with 0.1%

Fast green FCF and 0.1% Sirius red. This was then covered with aluminum foil and incubated at room temperature for 30 minutes in a rotary shaker. After incubation, the slides were carefully rinsed with distilled water, air-dried and cover-slipped as above.

In-vitro techniques

Plasma, liver and lipid biochemical assays

Plasma insulin was measured with ELISA kits (Crystal Chem, Chicago, IL) according to the Manufacturer‟s instructions. Liver transaminases and plasma lipids were performed by Sydpath Pathology, St Vincent‟s Hospital. Liver triglycerides were extracted using methanol / chloroform (2:1) and detected using a colorimetric assay (Roche, Basel, Switzerland). Leptin was measured by ELISA

(Crystal Chem, Chicago, IL) according to the manufacturer‟s instructions.

Protein extraction and quantitation

Fresh tissues were removed from anaesthetised animals and snap-frozen in liquid nitrogen and stored at -80C. They were then homogenised in lysis buffer and centrifuged. The pellets were sonified in cold RIPA buffer to obtain a nuclear- enriched extract. Total protein was quantitated using DC (Detergent Compatible) assay (Biorad, Berley, California) as per manufacturer‟s instructions.

Western immunoblot analysis for HIF-1α

30 to 120 g of protein sample was mixed with 5μL of protein sample buffer and separated by electrophoresis on an 8% SDS PAGE gel in Western running buffer using a Biorad Protean 3 apparatus. Proteins were transferred onto PVDF membrane over one hour in transfer buffer using the semi-dry method and a

Biorad transfer apparatus. Primary antibodies for HIF-1 (Novus Biologicals,

Littleton, Colorado, USA) and -tubulin as house-keeping protein (Abcam,

Cambridge UK) were prepared in 5% milk antibody buffer at concentrations of

1:500 and 1:2000 respectively and the membrane was incubated on a gentle shaker overnight at 4C. On the next morning, the membranes were washed 3 times with PBST. They were then incubated in secondary mouse antibodies

1:1000 (conjugated with Horseradish Peroxidase) (Pierce, Rockford, USA) prepared in a milk antibody buffer for 1 hour at room temperature. After that, the membranes were again washed 3 times with PBST. 7mls of Western Blot

Luminol Reagent (Santa Cruz Biotechnology, Santa Cruz, California, USA), and bands were visualised via chemiluminescence using ChemiDoc XRS (Biorad,

Berkeley, California, USA).

RNA preparation and real-time quantitative PCR.

Tissues were homogenised in RLT extraction buffer using the SilentCrusher homogeniser (Heidolph, Schwabach, Germany) followed by passing the lysates through Qiashredder columns (Qiagen, Venlo, The Netherlands). RNA was isolated using the RNeasy Kit (Qiagen) according to the manufacturer‟s protocol.

First strand cDNA was achieved with Superscript cDNA synthesis kit (Invitrogen,

Carlsbad, California) using random hexamers according to the manufacturer‟s instructions. For real-time PCR, 384 well plates were loaded using an EpMotion robot (Eppendorf, Hamburg, Germany), with primers, cDNAs and Sybergreen

(Roche, Basel, Switzerland). Amplification was measured by an ABI 7900 qPCR

Sequence Detection System according to protocols provided by the manufacturer

(Applied Biosystems, Foster City, California, USA). The cross threshold (ct) level of mRNA for each gene was normalised to the level of the housekeeping gene

TATA box binding-protein (TBP) mRNA in triplicates for each sample in every plate. The average of the triplicates was converted to CT fold change and presented as bar graphs.

Statistical analysis

Data are expressed as means + SEM. Statistical analysis was conducted using

Student‟s t test except where stated otherwise. ANOVA was employed where there were multiple repeated measures and time-points. Statistical significance was defined as p < 0.05. Non-parametric analyses were performed using the

SPSS package v 15.0 (IBM Corporation, Nov 2006).

Solutions and buffer

LID cell lysis buffer

10mmol/L Tris-HCl, 1% Triton-X-100, 0.5% NP40, 150 mmol/L NaCl, 10 mmol/L

Na orthophosphate, 10 mmol/L Na pyrophosphate, 100 mmol/L NaF, 1 mmol/L

EDTA, 1 mmol/L EGTA, 10 mmol/L Na orthovanadate, 1 protease inhibitor cocktail tablet per 50 ml.

RIPA buffer

50 nM Tris pH 7.4, 1% NP40, 0.25% sodium deoxycholate, 150 nM NaCl, 1 mM

EDTA, 1 mM PMSF and protease inhibitor cocktail 1 tablet per 50 ml.

Phosphate buffered solution (PBS)

3.6% Disodium hydrogen orthophosphate, 0.2% KCl, 0.24% potassium dihydrogen orthophosphate, 8% sodium chloride.

PBST

PBS plus 0.05% Tween-20.

Tris-buffered saline tween-20 (TBST)

Dissolve the following in 800 ml of distilled H2O: 8.8 g of NaCl, 0.2g of KCl, 3g of

Tris base. Add 500ul of Tween-20. Adjust the pH to 7.4 and add distilled water to

1L. Sterilise by filtration or autoclaving.

Protein sample buffer (reducing)

75% water, 2% SDS, 0.0625 mol/L Tris-HCl pH 6.8, 25% glycerol, 0.5% bromophenol blue and 0.5ml 2MeOH. Add 10 µL of mercaptoethanol to each ml of buffer prior to use.

Western Blot running buffer (5X)

1.5% Tris, 7.2% glycine, 5% SDS 10%. Dilute to 1X for use

Western Blot transfer buffer (10X)

2.25% Tris, 10.5% glycine. Add 10% of 10X buffer to 70% dH2O and 20% MeOH.

Milk antibody buffer

Add 5% blotting grade dry milk powder (Bio-Rad laboratories) to PBST.

CHAPTER 3

Effects of deferasirox on

high fat-fed

wild-type C57Bl/6 mice

INTRODUCTION

In Chapter 3, the effects of iron chelation using deferasirox (DFS) were studied on C57Bl/6 mice physiology when these mice were fed continuously on high fat diet (HFD). This is a well-known obese mice model with manifestations of the metabolic syndrome such as hepatic steatosis, hyperlipidaemia and insulin resistance, adiposity [150] and are prone to accelerated atherosclerosis [148]. In this model, there is sexual dimorphism in that male mice are more susceptible to the deleterious effects of HFD than female mice [151] . Therefore DFS effects were studied separately in both sexes.

Altogether, 148 male mice were studied in 4 separate experiments and 10 female mice in another separate experiment. The starting age for the mice in all experiments was 7 to 8 weeks, except for experiment 3, where they were older at

12 to 14 weeks. 60 male mice were each studied in each of experiments 1 and 2 for 10 weeks. 8 male mice in experiment 3 and 20 male mice in Experiment 4 were studied for 25 weeks. 10 female mice were studied in experiment 5 for 34 weeks. All mice were randomised to receive HFD containing 45% of calories from fat (see materials and methods for composition), or to the same HFD mixed with DFS (HFD+DFS). Mice were weighed weekly, and physiological testing was performed according to the experimental timelines. At the conclusion of the studies, the mice were sacrificed and tissues weighed and kept for histology and molecular analyses. These include western blots for protein, and real-time PCR

71 for gene expression. Blood was collected for biochemistry and haemoglobin assays. An overview of the findings is also displayed at the beginning with each finding discussed separately in more detail along with supporting figures and graphs. Finally a summary appears at the end of the chapter.

Figure 3.1: Experimental time-line C57Bl/6 mice on HFD experiment 1

Starting age: 7 to 8 weeks Experimental weeks

-1 0 5 6 7 8 9

60 Start HFD+DFS GTT ITT GSIS GTT CULL Male 2g/kg 0.5U/kg 3g/kg 1g/kg mice Check weekly arrive weights and from blood glucose ARC concentrations

ARC = Animal Resource Centre, Perth WA DFS = Deferasirox HFD = High fat diet GTT = Glucose tolerance test ITT = Insulin tolerance test GSIS = Glucose stimulated insulin secretion

Figure 3.2: Experimental time-line C57Bl/6 mice on HFD experiment 2

Starting age: 7 to 8 weeks Experimental weeks

-1 0 3 4 8 9 10

60 Start HFD+DFS Food Oxymax GTT ITT CULL Male Intake Study 2g/kg 0.25u/kg mice Check weekly Study arrive weights and from blood glucose ARC concentrations

ARC = Animal resource centre, Perth WA DFS = Deferasirox HFD = High fat diet Food Intake=Average amount of HFD eaten over 3 days Oxymax=Metabolic chamber studies GTT = Glucose tolerance test ITT = Insulin tolerance test

Figure 3.3: Experimental time-line C57Bl/6 mice on HFD experiment 3

Starting age: 12 to 14 weeks Experimental weeks

-1 0 4 5 6 7 8 10

8 Start Food Oxymax GTT ITT GSIS CULL Male HFD+DFS Intake Study 2g/kg 0.25u/kg 3g/kg mice Study from Check weekly Garvan weights and BTF blood glucose concentrations

DFS = Deferasirox BTF = Biological testing facility HFD = High fat diet Food Intake=Average amount of HFD eaten over 3 days Oxymax=Metabolic chamber studies GTT = Glucose tolerance test ITT = Insulin tolerance test GSIS = Glucose stimulated insulin secretion

Figure 3.4: Experimental time-line C57Bl/6 mice on HFD experiment 4

Starting age: 7 to 8 weeks Experimental weeks

-1 0 0.5 2 8 21 25

20 Start Food ITT Food GTT Food male HFD+DFS Intake & 0.25u/kg Intake & 2g/kg Intake & mice Oxymax Oxymax Oxymax Study 1 Study 2 Study 3 from Check weekly Garvan weights and BTF blood glucose CULL concentrations

Figure 3.5: Experimental time-line C57Bl/6 mice on HFD experiment 5

Starting age: 7 to 8 weeks Experimental weeks

-1 0 28 34

10 Start GTT CULL female HFD+DFS 2g/kg mice from Check weekly Garvan weights and BTF blood glucose concentrations

DFS = Deferasirox BTF = Biological testing facility HFD = High fat diet GTT = Glucose tolerance test

Deferasirox (DFS)-treated mice gained significantly less weight on continuous high fat diet (HFD)

The weight curves for the 5 experiments are shown in Figures 3.6 to 3.9, and

3.11. All show that mice that receiving HFD+DFS (DFS) did not gain as much weight as their littermates receiving control HFD (CON). Physiological tests were performed throughout the experiments and are shown in red arrows below the curves.

In Experiment 1, the weight gain for DFS+HFD mice became statistically significant after 3 weeks, and by 10 weeks, mean weight gain for the DFS group was 6.2+0.3g compared to 7.9+0.3g for CON (p<0.0001) (Figure 3.6).

Figure 3.6 Experiment 1 weight curves and overview of male C57Bl/6 mice fed

HFD (CON) or HFD+DFS (DFS)

n= 30 in each group ** from week 3 9

Weight 6 change 5 (g) dfs 4 con 3

0 0 1 2 3 4 5 6 7 8 9 10

Time on HFD (Weeks)

GTT ITT GSIS GTT

** = p <0.01 by t-test

DFS = Deferasirox CON = Control HFD = High fat diet GTT = Glucose tolerance test ITT = Insulin tolerance test GSIS = Glucose stimulated insulin Secretion

In experiment 2, mice receiving DFS were also observed to gain less weight compared to CON mice. In this experiment, the difference became significant after week 5. This may be because these mice had more interventions in the early period, with feeding and Oxymax studies. The difference in weight gain at the end of the study was similar, at 10.0+0.3g versus 11.5+0.4g (p=0.004)

(Figure 3.7).

Figure 3.7 Experiment 2 weight curves and overview

n= 30 in each group * from week 5

Weight 8 change (g) 7 Time on HFD (weeks) 6 dfs 5 con 4

0 0 1 2 3 4 5 6 7 8 9 10 11 Time on HFD (Weeks)

Food Oxymax GTT ITT Intake DFS = Deferasirox * = p <0.05 by t-test CON = Control HFD = High fat diet Food Intake=Average amount of HFD eaten over 3 days Oxymax=Metabolic chamber studies GTT = Glucose tolerance test ITT = Insulin tolerance test

In experiment 3, 8 wild-type male C57Bl/6 mice were subjected to the same treatment protocol (Figure 3.8). The effect of DFS was greater in this experiment: the weight curves became significantly different after only one week of DFS treatment. DFS-treated mice did not put on much weight, unlike CON mice. By week 10, the mean weight gain for DFS mice was 0.6+2.4g while that of CON mice was 8.1+1.7g (p<0.05).

The mice from this experiment were different from previous cohorts. Firstly they were obtained from the in-house colony at the Garvan, rather than ARC (Western

Australia). Secondly, they were slightly older at the start of the study: 12-14 weeks rather than 7-8 weeks for experiments 1 and 2. These differences may account for the increased effect of DFS on weight gain. Older rats are known to have different insulin sensitivities compared to younger rats [152]. It is also known that when different populations of mice from the same genetic line are inbred, they may show genetic drift [153]. Therefore it is possible that the Garvan

C57Bl/6 mice colony may have been altered genetically such that the effects of

DFS are more pronounced.

Figure 3.8 Experiment 3 weight curves and overview

n= 4 in each group **f rom week 1

12 dfs 10 con 8 Weight 6 change (g) 4

0 0 1 2 3 4 5 6 7 8 9 10 -2

-4 Time on HFD (Weeks)

Food Oxymax GTT ITT GSIS Intake

** = p <0.01 by t-test DFS = Deferasirox CON = Control HFD = High fat diet Food Intake=Average amount of HFD eaten over 3 days Oxymax=Metabolic chamber studies GTT = Glucose tolerance test ITT = Insulin tolerance test GSIS = Glucose stimulated insulin secretion

In experiment 4, twenty mice were studied for 25 weeks to determine if there were any adverse effects with longer treatment (Figure 3.9). Feeding and oxymax studies were carried out after two days on HFD. ITT was performed after two weeks. Again, possibly because of early physiological testing, the DFS effect on weight gain became significant after week 5. The mean weight gain from baseline for DFS mice compared to CON mice after week 10 was 3.3+1.6g versus 6.6+2.8g (p<0.01). At 26 weeks, the difference was larger, at 9.1+0.8g versus 16.8+1.7g (p=0.002) after week 25. There were small dips in weight gain later on in both groups in association with the intermittent fasting and blood collections associated with physiological testing.

Figure 3.9 Experiment 4 weight curves and overview

n= 10 in each group

* from week 5 19

Weight change

(g) 9

dfs 4 con

-1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

Time on HFD (Weeks)

Oxymax ITT Oxymax GTT Oxymax + Food + Food + Food Intake Intake Intake

* = p <0.05 by t-test DFS = Deferasirox CON = Control HFD = High fat diet Oxymax=Metabolic chamber studies ITT = Insulin tolerance test Food Intake=Average amount of HFD eaten over 3 days

Some CON mice were obviously larger than the DFS mice. Figure 3.10 shows a pair of male C57Bl/6 mice from experiment 4. These mice were aged 6 weeks at the beginning of the experiment and both weighed 22.1g. At the end of week 19 on continuous HFD, the CON mouse weighed 49.3g while the DFS mouse only weighed 27.3g.

Figure 3.10 Experiment 4 -- one DFS and one CON mouse at week 19 of HFD

In experiment 5, a group of female C57Bl/6 mice was studied over 34 weeks

(Figure 3.11). The weight gain of female C57Bl/6 mice on HFD was less than the weight gain in male mice in the other four experiments. In this study, the effect of

DFS on weight gain was smaller. However, the difference became statistically significant after week 23 of treatment. By week 34, the mean weight gain of DFS mice compared to CON mice was 0.3+0.5g versus 5.7+2.2g (p<0.05).

Figure 3.11 Experiment 5 weight curves and overview – female C57Bl/6 mice *

n= 10 in each group * from week 23 10

Weight dfs change 4 (g) con 2

0 0 5 10 15 20 25 30 35 40

-2 Time on HFD (Weeks)

GTT

* = p <0.05 by t-test

DFS = Deferasirox CON = Control HFD = High fat diet GTT = Glucose tolerance test

DFS treatment did not impair appetite

Smaller weight gain can be caused by a lower caloric intake, higher energy expenditure, or a combination of these effects. To determine if the addition of

DFS altered food intake, or made food unpalatable, food -intake studies were carried out in mice from experiments 2, 3 and 4, as described in materials and methods. Mice (eight each from DFS+HFD and HFD groups) were housed separately over 3 days. Residual HFD including food dropped onto the cage floor was subtracted from the weight of food put in at baseline to determine food intake. Faecal droppings were also weighed.

In experiment 2, three-day food -intake (adjusted for body weight) was not different when measured on week 3. This was done prior to any significant change in body weight between DFS and CON groups (Figure 3.12).

Again in experiment 3, there was no difference in week 4, even though at that time, DFS mice weighed significantly less than CON mice (Figure 3.13).

In experiment 4, there were no significant differences in food intake at baseline.

However, at weeks 8 and 25, DFS mice consumed significantly more food per gram of body weight than CON mice (Figure 3.14). There was also a trend to mice consuming more food per mouse (data not shown). This occurred despite the observation that DFS mice were smaller in size.

Figure 3.12 Experiment 2--three-day food intake during week 3

3 day food intake in week 3

0.12

0.10

HFD (g) 0.08 eaten per DFS gram 0.06 mouse CON weight 0.04

0.02

0.00 Day1 Day2 Day3

Days on HFD

HFD = High fat diet

Figure 3.13 Experiment 3--three-day food intake in week 4

3 day food intake in Week 4

0.12 P=0.12

0.10

HFD (g) 0.08 eaten per DFS gram 0.06 mouse CON weight 0.04

0.02

0.00 Day1 Day2 Day3

Days on HFD

HFD = High fat diet

Figure 3.14 Experiment 4--comparative food intake at weeks 0, 8 and 25.

Food intake at weeks 0, 8 and 25

0.14

0.12 * *

0.10 DFS HFD (g) CON eaten per 0.08 gram mouse 0.06 weight 0.04

0.02

0.00 Week 0 Week 8 Week 25

Days on HFD

*= P<0.05

HFD = High fat diet

DFS mice had higher basal metabolic rates

To determine if whole body metabolism could account for the differences in body weight, mice were housed separately in closed circuit metabolic chambers. Rates of oxygen consumption (VO2), carbon dioxide production (VCO2), and 3D movement (activity) were recorded using the Oxymax Columbus system.

Respiratory exchange ratio (RER) and heat were calculated from VO2 and VCO2 readings (see methods-Oxymax respirator studies).

When these studies were carried out in experiment 2 in week 4, prior to significant weight change in week 5, there were no differences in the Oxymax readings (Figure 3.15). Experiment 3, at week 5, significant oxymax differences were present at time when there was a weight difference (Figure 3.16). DFS mice had higher VO2 and VCO2 rates during the night time, suggesting higher metabolic rates. Physical activity levels were not different. Their RER curve had lowered significantly during the day, demonstrating that the DFS mice had higher fat oxidation than the CON mice.

Figure 3.15 Oxymax studies: experiment 2 at week 4

6600 1.1 Respiratory exchange ratio 6200 O2 consumption 5800 Ml 5400 1 / 5000 4600 RER kg 4200 0.9 3800 / 3400 hr 3000 0.8 2600 2200 1800 0.7 0 2 4 6 8 10 12 14 16 18 20 22 24 0 2 4 6 8 10 12 14 16 18 20 22 24 Time Time

Energy expenditure per body Activity X+Yamb 30 mass 1000 Kcal No. of 20 / events Kg 500 10 / hr

0 0 0 2 4 6 8 10 12 14 16 18 20 22 24 0 2 4 6 8 10 12 14 16 18 20 22 24 Time Time

DFS CON

Figure 3.16 Oxymax studies: experiment 3 at week 5

Respiratory exchange ratio O2 consumption 1.1 6400 6000 5600 1 Ml 5200 * * * / 4800 RER 4400 0.9 kg 4000 / 3600 3200 0.8

hr 2800 2400 2000 0.7 0 2 4 6 8 10 12 14 16 18 20 22 24 0 2 4 6 8 10 12 14 16 18 20 22 24 Time Time

Energy expenditure per body mass 2000 Activity X+Yamb 30 * * * Kcal No. of / 20 events Kg 1000 / 10 hr

0 0 0 2 4 6 8 10 12 14 16 18 20 22 24 0 2 4 6 8 10 12 14 16 18 20 22 24 Time Time

DFS *=p<0.05 CON

To determine if the metabolic changes were present early in the course of DFS treatment, sixteen mice in experiment 4 were investigated with oxymax studies at day 2, week 8 and week 25.

At baseline, mice weights were not different. Oxymax study at day 2 of HFD feeding did not reveal any detectable effect of DFS on metabolism (Figure 3.17).

By week 8 in experiment 4, DFS mice were already significantly lower in body weight compared to CON mice. Oxymax studies at this stage showed that VO2 and VCO2 were actually higher in DFS mice compared to CON mice. There were no differences in RER or activity (Figures 3.18). These differences were again observed in week 25 (Figure 3.19). They occurred during daytime hours, suggesting that DFS treatment may improve resting metabolic rate in association with lower weight gain on HFD.

Therefore it appears that DFS acts to increase basal metabolism and this effect is apparent after a few weeks of treatment, and associated with significantly less weight gain.

Figure 3.17 Oxymax studies: experiment 4 at week 0, day 2

O2 concumption 1.1 6600 Respiratory exchange ratio 6200 Ml 5800 1 5400 / 5000 RER 4600 0.9 kg 4200 3800 / 3400 0.8 hr 3000 2600 2200 1800 0.7 0 2 4 6 8 10 12 14 16 18 20 22 24 0 2 4 6 8 10 12 14 16 18 20 22 24

Time Time

Energy expenditure per body mass Activity X+Yamb 2000 40

Kcal No. of 30 / events Kg 20 1000 /

hr 10

0 0 0 2 4 6 8 10 12 14 16 18 20 22 24 0 2 4 6 8 10 12 14 16 18 20 22 24 Time Time

DFS CON

Figure 3.18 Oxymax studies: experiment 4 at week 8

O2 consumption Respiratory exchange ratio 6600 1.1 Ml 6200 5800 / 5400 * 1 5000 RER kg 4600 4200 0.9 / 3800 hr 3400 3000 0.8 2600 2200 1800 0.7 0 2 4 6 8 10 12 14 16 18 20 22 24 0 2 4 6 8 10 12 14 16 18 20 22 24 Time Time

Energy expenditure per body mass Activity X + Y amb 30 * Kcal 20 1000 / No. of Kg events / 10 hr

0 0 0 2 4 6 8 10 12 14 16 18 20 22 24 0 2 4 6 8 10 12 14 16 18 20 22 24 Time Time

*=p<0.05 DFS CON

Figure 3.19 Oxymax studies: experiment 4 at week 25

O consumption 6600 2 1.1 Respiratory exchange ratio 6200 5800 P=0.05 Ml 5400 1 / 5000 4600 RER kg 4200 0.9 3800 / 3400 hr 3000 0.8 2600 2200 1800 0.7 0 2 4 6 8 10 12 14 16 18 20 22 24 0 2 4 6 8 10 12 14 16 18 20 22 24 Time Time

Energy expenditure per body mass Activity X+Yamb 30 2000 * No. of Kcal events / 20 Kg 1000

/ 10 hr

0 0 0 2 4 6 8 10 12 14 16 18 20 22 24 0 2 4 6 8 10 12 14 16 18 20 22 24 Time Time

*=p<0.05

DFS CON

DFS mice had higher core body temperatures

Core body temperatures were compared between both groups of mice in

Experiment 4. Using a thermometer probe, rectal temperatures were obtained at noon in both groups of mice at weeks 7, 9, 10, 11, 13 and 16. DFS mice had temperature readings 0.5-1.0 degrees Celsius higher when compared to CON mice (p<0.01 by ANOVA for repeated measures) (Figure 3.20). Again these results suggest that DFS treatment increased metabolic rates.

100

Figure 3.20 DFS mice have higher core body temperature

Rectal temperatures DFS P=0.01 by ANOVA for repeated measures 38.0 CON

37.0

Degree 36.0 celcius 35.0

34.0

33.0

32.0 Week 7 Week 9 Week 10 Week 11 Week 13 Week 16

101

DFS mice had lower fasting hyperinsulinaemia and tended to have better insulin sensitivity

Fasted insulin levels were measured in weeks 7 to 8, after there were significant weight changes between both groups of mice. In experiment 1, DFS mice had significantly lower fasting insulin levels compared to CON mice (2837+874 pmol/L versus 4703+1741 pmol/L) (p=0.005) (Figure 3.21) and again tended to be lower in experiment 3 (Figure 3.22). This suggested that the DFS mice were less insulin resistant.

102

Figure 3.21 Experiment 1 fasting insulin at week 7

n= 30 in each group

Fasting insulin

6000 5000 ** 4000

3000 pmol/L 2000

1000

0 DFS CON

**=p<0.01

103

Figure 3.22 Experiment 3 fasting insulin at week 8

n=4 mice in each group

Fasting insulin

P=0.08 5000 4500 4000 3500 3000 DFS 2500 CON pmol/L 2000 1500 1000 500 0

104

Intra-peritoneal insulin tolerance testing (ITT) was performed to test for whole body insulin sensitivities. While there was only a slight lowering of glucose in experiment 1at week 6 of DFS treatment, this observation was not seen in experiments 2 and 3 at weeks 9 and 7respectively (Figures 3.23-25). The combination of these ITT, results termed „late ITT‟, showed only a trend to be lower (Figure 3.26). When an ITT was performed in week 2 (termed “early” ITT) in experiment 4 (Figure 3.27), it showed that DFS treatment did not confer early insulin-sensitising effects, but if anything, resulted in a delay in the insulin- induced fall in glucose. This was prior to the occurrence of significant weight changes.

105

Figure 3.23 Experiment 1insulin tolerance test (ITT) at week 6

n=30 mice in each group

ITT (0.5units/kg) in week 6 DFS 120 CON 100 % blood 80 glucose * * 60 from baseline 40

0 0 10 20 30 40 50 60

Minutes

*=p<0.05

106

Figure 3.24 Experiment 2 insulin tolerance test (ITT) at week 9

n=30 mice in each group

ITT (0.25units/kg) in week 9

120 DFS 100 % blood CON glucose 80

from 60

baseline 40

0 0 10 20 30 40 50 60

Minutes

107

Figure 3.25 Experiment 3 insulin tolerance test (ITT) at week 7

n=4 mice in each group

ITT (0.25units/kg) week 7 DFS CON 120

100 % blood glucose 80

from 60 baseline 40

0 0 10 20 30 40 50 60

Minutes

108

Figure 3.26 Combined experiments 1 to 3 (late) insulin tolerance test (ITT)

n=64 mice in each group

Combined late ITT result DFS

120 CON

100 % blood p=0.17 glucose 80

from 60 baseline 40

0 0 10 20 30 40 50 60

Minutes

109

Figure 3.27 Experiment 4 (early) insulin tolerance test (ITT) at week 2

n=10 mice in each group

ITT (0.25units/kg) in week 2

DFS 120 * * CON 100 % blood 80 glucose 60 from baseline 40

0 0 10 20 30 40 50 60

Minutes

*=p<0.05

110

DFS mice had preserved beta cell function and better glucose tolerance after prolonged HFD feeding

Mice were injected with intra-peritoneal glucose to stimulate insulin secretion during weeks 7 and 8 of HFD. In experiment 1, first-phase insulin secretory response was present in DFS mice but was completely abolished in CON mice

(Figure 3.28). The second phase secretion in DFS mice was also stronger. When repeated in experiment 3 (n=8 mice), this trend was again present (Figure 3.29).

Combining both results showed significant first-phase response at 2 minutes

(117+9% versus 94+6%, p=0.02) and a stronger second-phase response at 20 minutes (177+25% versus 126+10%, p=0.07) (Figure 3.30).

111

Figure 3.28 Experiment 1 glucose stimulated insulin secretion (GSIS) at week 7

n=30 mice in each group

GSIS (3g/kg) in week 7 250 *

200 % insulin 150 * secretion

from 100 baseline DFS 50 CON

0 0 5 10 15 20

Minutes after glucose injection

*=p<0.05

112

Figure 3.29 Experiment 3 glucose stimulated insulin secretion (GSIS) at week 8

n=4 mice in each group

GSIS (3g/kg) in week 8

300 DFS 250 % insulin CON secretion 200

from 150

baseline 100

0 0 5 10 15 20 Minutes

113

Figure 3.30 Combined experiments 1 and 3 glucose stimulated insulin secretion

n=34 mice in each group

GSIS (3g/kg) week 7 and 8

250 DFS *

200 CON % insulin secretion 150 * from 100 baseline 50

0 0 5 10 15 20

Minutes

*=p<0.05

114

When glucose tolerance testing (GTT) was examined, there was a trend to better glucose tolerance for DFS mice compared to CON mice at week 5 of experiment

1 (Figure 3.31). However, this was not found when GTTs were repeated in experiments 2 (week 8) and 3 (week 6). GTT‟s from weeks 5 , 6 and 8 were combined and analysed as „early‟ and the combined GTT result found no differences (Figures 3.32-34). It was noted, however, the CON mice in experiments 1 and 2 did not have the expected deteriorations in glucose tolerance on the HFD.

In experiment 4, when the mice were fed for up to 25 weeks, DFS treated mice demonstrated significantly better glucose tolerance (late GTT) when compared with CON mice at week 21 (AUC=2565+140 versus 3321+156, p=0.002) (Figure

3.35). Therefore it appears that prolonged treatment with DFS may improve GTT, probably secondary to reduction in weight gain.

In experiment 5, ten female mice were subjected to GTT at week 27 (Figure

3.36). There was a tendency towards better glucose tolerance in the mice receiving DFS compared to CON mice, with significantly lower blood glucose at the 120min time-point. Female C57Bl/6 mice are known to be more resistant to the deleterious effects of HFD upon glucose tolerance compared to males. This may account for the smaller benefit.

115

Figure 3.31 Experiment 1 intra-peritoneal glucose tolerance test (GTT)

at week 5

n=30 mice in each group

GTT (2g/kg) in week 5

25 P=0.095 DFS

20 CON Blood 15 glucose * (mmol/L) 10

0 0 20 40 60 80 100 120 Minutes

*=p<0.05

116

Figure 3.32 Experiment 2 intra-peritoneal glucose tolerance test (GTT)

at week 8

n=30 mice in each group

GTT (2g/kg) in week 8

25 DFS

20 CON Blood

glucose 15 (mmol/L) 10

0 0 20 40 60 80 100 120

Minutes

117

Figure 3.33 Experiment 3 intra-peritoneal glucose tolerance test (GTT)

at week 6

n=4 mice in each group

GTT (2g/kg) in week 6

30 DFS 25 CON Blood 20 glucose 15 (mmol/L) 10

0 0 20 40 60 80 100 120

Minutes

118

Figure 3.34 Combined experiments 1, 2, 3 (early) glucose tolerance test (GTT)

n=63 mice in each group

Combined early GTT

DFS 25.0 CON 20.0 Blood

glucose 15.0 P=0.09 (mmol/L) 10.0

5.0

0.0 0 20 40 60 80 100 120

Minutes

119

Figure 3.35 Experiment 4 (late) glucose tolerance test (GTT) at week 21

n=10 mice in each group

GTT (2g/kg) in week 21 DFS ** CON 35 ** 30 ** ** Blood 25 glucose 20 *

(mmol/L) 15

0 0 20 40 60 80 100 120

Minutes

*=p<0.05

**=p<0.01

120

Figure 3.36 Experiment 5 (late) glucose tolerance test (GTT) at week 27

n=5 mice in each group

GTT (2g/kg) in week 27

DFS

35 CON 30

Blood 25 glucose 20 *

(mmol/L) 15

0 0 20 40 60 80 100 120

Minutes

*=p<0.05

121

DFS mice were less obese

The differences in weight gain between the DFS and CON mice were thought to be due to differences in fat gain. This was confirmed upon sacrifice at completion of the various experiments. In Experiment 2, after 10 weeks of continuous HFD,

DFS mice had significantly lighter visceral fat depots (VAT) compared to CON mice: 0.80+0.04g versus 1.31+0.08g (p<0.00001) and subcutaneous fat depots

(WAT): 0.25+0.01 g versus 0.38+0.03 g (p<0.00001) (Figure 3.37).

When fed for 25 weeks in Experiment 4, DFS mice had significantly less VAT

(1.4+0.2 g versus 2.1+0.2 g) (p=0.02) and WAT (0.7+0.2 g versus 1.8+0.3 g)

(p=0.009) compared to CON mice. There were no differences in the weights of other organs (Figure 3.38). This was associated with significantly reduced plasma leptin levels in DFS mice (7.3+3.2 ng/ml versus 15.3+5.4 ng/ml)

(p=0.034) (Figure 3.39)

Likewise, for the female mice in Experiment 5, there was a trend towards lighter fat depots in the DFS mice (Figure 3.40).

122

Figure 3.37 Experiment 2 fat weights

n=30 mice in each group

Fat weights 1.6 DFS 1.4 ********

1.2 CON

1.0 Mass 0.8 ******** (g) 0.6

0.4

0.2

0.0 VAT WAT

********=p<0.00001

VAT = Visceral adipose tissue WAT = White adipose tissue

123

Figure 3.38 Experiment 4 organ weights

n=10 mice in each group

Organ weights

2.5 ** ** DFS

2.0 CON

1.5 Mass (g) 1.0

0.5

0.0 Liver Pancreas Brain VAT WAT BAT Gastrocnemius

**=p<0.01

VAT = Visceral adipose tissue WAT = White adipose tissue BAT = Brown adipose tissue

124

Figure 3.39 Experiment 4 plasma leptin

n=10 mice in each group

Plasma leptin

25.0 * DFS 20.0 CON Leptin

(ng/ml) 15.0

10.0

5.0

0.0

*=p<0.05

125

Figure 3.40 Experiment 5 organ weights

n=5 mice in each group

Organ weights DFS CON 2.5 P=0.12 2.0 Mass P=0.08 1.5 (g)

1.0

0.5

0.0 Liver Pancreas VAT WAT

VAT = Visceral adipose tissue WAT = White adipose tissue

126

DFS-treated mice had lower serum and liver lipids

In experiment 1, mice were not fasted at sacrifice. DFS mice had significantly lower plasma cholesterol compared to CON mice: 1.86+0.05g versus

2.19.1+0.08g (p=0.03) (Figure 3.41). Plasma triglycerides were not different.

The fasted plasma in experiment 2 (male mice) and experiment 5 (female mice) showed similar trends (Figures 3.42-43).

In experiment 2, fasted DFS mice livers had significantly lower hepatic triglyceride content compared to CON livers (Figure 3.44).

Liver histological examination showed extensive macroscopic steatosis in CON mice, but the DFS mice were relatively spared (Figures 3.45-46).

However, there were no obvious differences in hepatic fibrosis when stained using sirius red staining (Figure 3.47).

127

Figure 3.41 Experiment 1 plasma lipids (non-fasted)

n=5 mice in each group

Plasma lipids (non-fasted)

DFS 2.5 * CON 2

1.5 mmol/L 1

0.5

0 Triglycerides Cholesterol

*=p<0.05

128

Figure 3.42 Experiment 2 plasma lipids (fasted)

n=5 mice in each group

Plasma lipids (fasted) DFS P=0.08 3.5 CON

2.5

2 mmol/L 1.5

0.5

0 Triglycerides Cholesterol

129

Figure 3.43 Experiment 5 plasma lipids (fasted)

n=5 mice in each group

Serum lipids (fasted)

3.0 DFS

2.5 CON

2.0 mmol/L 1.5

1.0

0.5

0.0 Triglycerides Cholesterol

130

Figure 3.44 Experiment 2 liver triglycerides (fasted)

n=5 mice in each group

Liver triglycerides

1000 DFS 900 * 800 CON

700 ng per 30mg 600 liver 500 400

300

200

100

*=p<0.05

131

Figure 3.45 Haematoxylin and eosin slide experiment 3 Liver

132

Figure 3.46 Haematoxylin and eosin slide experiment 4 liver

133

Figure 3.47 Sirius red slide experiment 3 liver

134

DFS-treated livers had lower iron stores and showed increased levels of

HIF-1

In order to assess liver iron content, liver sections were stained using Perls‟ staining protocol. Iron deposits react with the reagents to stain Prussian blue colour. DFS treated livers showed fewer blue deposits, demonstrating lower iron levels compared to CON livers (Figure 3.48).

Western blots of experiment 2 liver lysates suggest increased HIF-1 in DFS treated mice (Figure 3.49). This was confirmed in another liver Western Blot,

(Figure 3.50) with four mice fed DFS for 3 days and injected with intra-peritoneal deferoxamine (DFO) 45 minutes prior to sacrifice. CON mice received normal

HFD and saline injections.

135

Figure 3.48 Perls‟ stain experiment 3 liver

136

Figure 3.49 Western blot experiment 2 liver

Liver HIF - 1α

CON DFS DFS

HIF-1α

α-tubulin

HIF= Hypoxia inducible factor

137

Figure 3.50 Western blot DFS versus CON liver

n=4 mice in each group

Liver HIF - 1α

CON DFS

30000 * 25000

20000 CON

15000 DFS 10000 Vol density Vol

5000

0 1

HIF= Hypoxia inducible factor

*=p<0.05

138

DFS treatment resulted in altered gene expression for insulin signalling and lipid metabolic pathways

Tissue gene expression changes were examined using real-time PCR analyses experiment 2 mice. Mice from each treatment group (six from each) were analysed separately, for gene expression in livers, visceral adipose tissue and gastrocnemius muscle.

Livers

In livers (Figures 3.51-54) it was found that DFS treatment was associated with significantly increased gene expression for AKT2, IRS1, IRS2 (insulin signalling), and PFK, but not HNF4a (MODY). There was a trend for improved GLUT2 expression in the liver but in one experiment, GLUT1 was significantly decreased

(Figure 3.54). Lipoprotein lipase (LPL) was consistently and significantly increased in all liver samples but hormone-sensitive lipase (HSL) changes were inconsistent, as were HMG CoA reductase.

VAT

There were no significant gene expression (Figure 3.55-56) changes present in visceral adipose tissue except for leptin expression, which was significantly down regulated in DFS mice. This is likely to be secondary to reduced fat depots in these animals.

139

Gastrocnemius muscle

Significant changes were also seen in the muscle, with increased GLUT1,

GLUT2 and LPL gene expression. However, IRS2 and CPT1 expression were significantly down regulated (Figure 3.57-58).

Thus it is possible that DFS, through up-regulation of HIF-1α, may increase insulin signalling, muscle glucose transport and lipase expression resulting in improved insulin sensitivity and weight loss.

140

Figure 3.51 Liver gene expression 1

Experiment 2

P=0.07 DFS 3.5 CON 3.0 * * 2.5 * 2.0 * * 1.5 Fold Change Fold 1.0

0.5

0.0 AKT2 ARNT GLUT2 HIF-1a HNF4a HSL IRS1 IRS2 PFK

*=p<0.05

AKT2 = Protein kinase B ARNT = Aryl hydrocarbon receptor nuclear translocator GLUT2 = Glucose transporter 2 HIF-1a = Hypoxia inducible factor-1 alpha HNF-4a = Hepatocyte nuclear factor-4 alpha HSL = Hormone sensitive lipase IRS = Insulin receptor substrate PFK = Phosphofructose kinase

141

Figure 3.52 Liver gene expression 2

Experiment 2

3.5 DFS

CON 3.0 ** 2.5 2.0 *

Fold Change Fold 1.5

1.0

0.5

0.0

*=p<0.05 **=p<0.01

GK = Glucokinase GLUT1 = Glucose transporter 1 HNF-4a = Hepatocyte nuclear factor-4 alpha HSL = Hormone sensitive lipase IRS1 = Insulin receptor substrate 1 LPL = Lipoprotein lipase PPARg = Peroxisome proliferator-activated receptors-gamma SREBP-1c = Sterol regulatory element binding protein-1c VEGF = Vascular endothelial growth factor

142

Figure 3.53 Liver gene expression 3

Experiment 2

3.5 **

3.0 DFS

CON 2.5 * 2.0 *

1.5 Fold Change Fold

1.0

0.5

0.0

*=p<0.05 **=p<0.01

ACC1 = Acetyl-CoA carboxylase 1 ALDO = Aldolase CPT1 = Carnitine palmitoyltransferase 1 FAS = Fatty acid synthase HMGred = HMG-CoA reductase HMGsyn = HMG-CoA Synthase HSL = Hormone sensitive lipase LPL = Lipoprotein lipase UCP2 = Uncoupling protein 2

143

Figure 3.54 Liver gene expression 4

Experiment 2 DFS 3.5 CON 3.0

2.5 ** 2.0 * *

Fold Change Fold 1.5

1.0

0.5

0.0

*=p<0.05 **=p<0.01

AKT2 = Protein kinase B GLUT1 = Glucose transporter 1 GLUT2 = Glucose transporter 2 HMGsyn = HMG-CoA synthase HSL = Hormone sensitive lipase IRS1 = Insulin receptor substrate 1 IRS2 = Insulin receptor substrate 2 LPL = Lipoprotein lipase PFK = Phosphofructose kinase

144

Figure 3.55 Visceral adipose tissue (VAT) gene expression 1

Experiment 2

3.5

3.0 DFS CON 2.5

2.0

Fold Change Fold 1.5

1.0

0.5

0.0

ACC1 = Acetyl-CoA carboxylase 1 ACC2 = Acetyl-CoA carboxylase 2 CPT1 = Carnitine palmitoyltransferase 1 FAS = Fatty acid synthase GLUT1 = Glucose transporter 1 HSL = Hormone sensitive lipase LPL = Lipoprotein lipase PPARg = Peroxisome proliferator-activated receptors-gamma SREBP-1c = Sterol regulatory element binding protein-1c

145

Figure 3.56 Visceral adipose tissue (VAT) gene expression 2

Experiment 2 5.0

4.5 DFS

4.0 CON 3.5

3.0 change

Fold 2.5 2.0 *** 1.5

1.0

0.5

0.0

*** = p<0.005

ACC1 = Acetyl-CoA carboxylase 1 ACC2 = Acetyl-CoA carboxylase 2 IRS2 = Insulin receptor substrate 2 PPARg = Peroxisome proliferator-activated receptors-gamma SREBP-1c = Sterol regulatory element binding protein-1c UCP2 = Uncoupling protein 2

146

Figure 3.57 Gastrocnemius gene expression 1

Experiment 2

5.0 *** 4.5 DFS 4.0 CON 3.5

3.0

change 2.5

Fold 2.0 *** ***

1.5 ***

1.0

0.5

0.0 AKT2 GLUT1 GLUT2 GLUT4 HSL IRS1 IRS2 LPL PFK

*** = p<0.005

AKT2 = Protein kinase B GLUT1 = Glucose transporter 1 GLUT2 = Glucose transporter 2 GLUT4 = Glucose transporter 4 HSL = Hormone sensitive lipase IRS1 = Insulin receptor substrate 1 IRS2 = Insulin receptor substrate 2 LPL = Lipoprotein lipase PFK = Phosphofructose kinase

147

Figure 3.58 Gastrocnemius gene expression 2

Experiment 2 5.0

4.5 DFS

4.0 CON

3.5

3.0 Change 2.5 Fold 2.0 ***

1.5

1.0

0.5

0.0

*** = p<0.005

ACC1 = Acetyl-CoA carboxylase 1 ACC2 = Acetyl-CoA carboxylase 2 CPT1 = Carnitine palmitoyltransferase 1 FAS = Fatty acid synthase HMGred = HMG-CoA reductase HMGsyn = HMG-CoA synthase PPARg = Peroxisome proliferator-activated receptors-gamma SREBP-1c = Sterol regulatory element binding protein-1c VEGF = Vascular endothelial growth factor

148

DFS treatment did not cause anaemia

It was important to establish if DFS treatment resulted in iron deficiency anaemia, since these mice were not iron overloaded, and were eating a normal iron content diet. Mice in experiment 2 had received DFS for 10 weeks, and those in experiment 5 had been treated for 35 weeks. Blood were collected at sacrifice to check for haemoglobin and plasma iron levels. As shown in Figure 3.59, there were no significant differences in haemoglobin between DFS and CON mice at

10 and 35 weeks. However, DFS mice had significantly higher plasma iron levels after 10 weeks of DFS: 22.6+01.2g versus 17.9+1.2g (p=0.03). DFS treatment did not result in significantly higher plasma iron after 35 weeks of treatment, even though there may have been a tendency (Figure 3.60). These results are consistent with the reports of DFS increasing serum iron levels and decreasing tissue iron content (see literature review).

149

Figure 3.59 DFS effect on haemoglobin

Haemoglobin on DFS treatment

DFS

CON 140

120

100

80 g/L 60

0 10 weeks 35 weeks

150

Figure 3.60 DFS effect on serum iron

Plasma iron on DFS treatment

DFS 40 * CON 35

umol/L 20

0 10 weeks 35 weeks

*=p<0.05

151

Summary

Chapter 3 details five separate experiments that examined the effects of an oral iron chelation agent (deferasirox, DFS) on an obese mouse model, C57Bl/6 on continuous high fat diet (HFD). The majority of the mice were young male mice (4 to 6 weeks of age) studied for up to 10 weeks ( approximately 3 months) of continuous HFD feeding. A small number of male mice were studied for up to 25 weeks (approximately 6 months). There were a very small number of female mice which were studied for up to 35 weeks (9 months). Thus, the effects of DFS were examined in both male and female mice in prolonged HFD feeding as well as in shorter periods of feeding.

The findings show obvious benefits of DFS in mitigating the deleterious effects of continuous high fat feeding in C57Bl/6 mice. These include consistent reduction of weight gain in both male and female mice, from as early as after 1 week, lasting throughout the whole period of study (up to 9 months). Other beneficial effects include lower plasma lipids, reduction of hepatic steatosis and improvement in insulin resistance, as shown by lower fasting hyperinsulinaemia in DFS treated mice. There is also improvement in beta cell dysfunction, evidenced by a relatively preserved first phase insulin secretion and improved glucose tolerance, especially with prolonged high fat exposure. DFS treatment is associated with reduced hepatic iron stores and increased HIF-1α levels. This is accompanied by down-regulation of hepatic lipogenic gene expression and

152 improved insulin signalling gene expression. Importantly, DFS does not impair food intake and does not appear to exert toxic effects as the treated mice gained weight as per normal chow fed C57Bl/6 mice and are not rendered anaemic. It is also interesting to note that rather than reducing appetite, DFS appears to increase appetite. This, together with increased core body temperature, suggests improved whole body metabolism. This was confirmed with metabolic chamber studies showing higher oxygen consumption, carbon dioxide production and energy expenditure in DFS treated mice.

153

CHAPTER 4

Effects of deferasirox on

ob/ob mice

and

wild-type C57Bl/6 mice

154

INTRODUCTION

Having noted the beneficial effects of deferasirox on C57Bl/6 mice on high fat diet (HFD) in Chapter 3, it became important to examine these effects in other obese mouse models. Simultaneous experiments were carried out in ob/ob mice, a well-known obese mouse model, as well as in normal chow-fed wild-type (WT)

C57Bl/6 mice. Ob/ob mice also have a propensity (albeit variable and unpredictable) to develop beta cell failure.

The ob/ob mice were created on the C57Bl/6 background. They were bred from two Ob heterozygotes, producing a litter of mixed ob/ob, ob/WT and WT/WT mice. The heterozygotes were considered to be functional wild-type C57Bl/6 mice. Only male mice were studied. Altogether, ten ob/ob mice and eight WT male mice were studied. They were housed five per cage, in varying proportions of ob/ob and WT mice and fed continuously on normal mice chow diet. Cages were randomised to receive blended chow with or without deferasirox (DFS) powder mixed in.

Mice were weighed weekly for 23 weeks. Glucose and insulin tolerance tests were carried out at week 8 and 9 respectively and an oxymax respirator study was done at week 22. The results for ob/ob and WT mice were analysed and presented separately in this chapter.

155

Figure 4.1: Experimental time-lines: ob/ob mice experiment

Starting age: 18-19 weeks Experimental weeks

-2 -0.5 0 8 9 22 23

10 GTT Start GTT ITT Food CULL Male 2g/kg Chow+DFS 2g/kg 2.0u/ Intake & Oxymax mice kg Study arrive Weekly from weights and ARC blood glucose concentrations

DFS = Deferasirox Food Intake=Average amount of HFD eaten over 3 days Oxymax=Metabolic chamber studies GTT = Glucose tolerance test ITT = Insulin tolerance test

156

Figure 4.2: Experimental time-lines: C57Bl/6 mice on chow experiment

Starting age: 18-19 weeks Experimental weeks

-2 -0.5 0 8 9 22 23

8 GTT Start GTT ITT Food CULL Male 2g/kg Chow+DFS 2g/kg 0.3u/ Intake & Oxymax mice kg Study arrive Weekly from weights and ARC blood glucose concentrations

DFS = Deferasirox Food Intake=Average amount of HFD eaten over 3 days Oxymax=Metabolic chamber studies GTT = Glucose tolerance test ITT = Insulin tolerance test

157

DFS-treated ob/ob gained less weight

The ob/ob mice were randomised to receive either blended chow diet with mixed- in deferasirox powder (DFS) or control blended chow diet (CON). DFS and CON mice were well matched in age (20 weeks) and in mean starting weight

(46.8g+2.9 vs 45.4g+1.5, p=0.695).

Because of the large variation in weight in the DFS group (39.3g to 53.5g), weight change from week 0 was noted, rather than absolute weight. From week 7 onwards, DFS mice gained significantly less weight than CON mice, such that at the end of week 23, the mean weight gain for DFS mice was 22.4+1.2g vs

27.8+0.9g for CON mice (p=0.01) (Figure 4.3).

158

Figure 4.3: Ob/ob weekly weight change

n= 5 in each group

*** from week 7

Weight 20 change (g) 15 DFS 10 CON 5

0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

Time on Chow (Weeks) GTT GTT ITT Oxymax + food intake

*** = p<0.005

DFS = Deferasirox CON = Control GTT = Glucose tolerance test ITT = Insulin tolerance test

159

Baseline glucose tolerance tests (GTTs)

Because ob/ob mice have a varying propensity to develop beta cell failure and therefore diabetes, it was important to compare their glucose tolerance at baseline, after randomisation, prior to DFS treatment (Figure 4.4).

This figure shows that, while all the ob/ob mice were moderately glucose intolerant, the ob/ob that have been randomised to receive DFS later have significantly prolonged hyperglycaemia, suggesting worse glucose tolerance at baseline (Figure 4.4).

160

Figure 4.4 Ob/ob experiment glucose tolerance test (GTT) week 0

n= 5 in each group

35 GTT at week 0 DFS 30 CON

25 * Blood 20 glucose (mmol/L) 15

0 0 20 40 60 80 100 120

Minutes

*=p<0.05

DFS = Deferasirox CON = Control GTT = Glucose tolerance test

161

DFS-treated ob/ob had higher random blood glucose concentrations

Despite less weight gain, it appeared that DFS-treated ob/ob mice have higher random blood glucose concentrations compared with CON mice generally. These were significant on some weeks (Figure 4.5). This result was unexpected and may be due to worse beta cell function at baseline in the DFS group prior to commencement of DFS.

162

Figure 4.5 Ob/ob experiment random blood glucose

n= 5 in each group

P=0.08 16 * DFS 14 * CON 12

Random 10 * blood 8 glucose (mmol/L) 6

0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

Time on Chow (Weeks) GTT GTT

*=p<0.05

DFS = Deferasirox CON = Control GTT = Glucose tolerance test

163

Glucose tolerance tests (GTTs) were not significantly different at week 8 of treatment

By week 8 of DFS treatment, DFS mice have gained significantly less weight compared to CON mice. However, at this stage, GTT was not significantly different between both DFS and CON mice (Figure 4.6).

164

Figure 4.6 Ob/ob experiment glucose tolerance test (GTT) week 8

n= 5 in each group

GTT at week 8

35 DFS

30 CON

25 Blood glucose 20

(mmol/L) 15

0 0 20 40 60 80 100 120

Minutes

DFS = Deferasirox CON = Control GTT = Glucose tolerance test

165

DFS-treated ob/ob were not more insulin sensitive despite lower weight gain

Insulin tolerance testing (ITT) was carried out at week 9 using an insulin dose of

2 units/kg given as an intra-peritoneal injection (Figure 4.7). This was a higher dose as the ob/ob mice were severely insulin resistant. Blood glucose was recorded at intervals as a percentage of baseline glucose. Percent glucose rose in both groups at 10 minutes and fell quickly back to baseline levels in the CON group. In the DFS group, the fall was more gradual. Even though this difference was not statistically significant it suggested that DFS-treated ob/ob mice were not more insulin sensitive.

166

Figure 4.7 Ob/ob experiment insulin tolerance test (ITT) week 9

n= 5 in each group

ITT at week 9 DFS 200 CON % blood 150 glucose from 100 baseline

0 0 10 20 30 40 50 60

Minutes

DFS = Deferasirox CON = Control GTT = Glucose tolerance test

167

DFS-treated ob/ob tended to eat more

When food intake study was examined at week 22 of DFS treatment, there was a trend for DFS-treated ob/ob mice to eat slightly more chow diet, when adjusted for body weight (Figure 4.8).

168

Figure 4.8 Ob/ob 24-hour food intake study

P=0.08 0.10 DFS

CON 0.08 Chow (g) Eaten Per 0.06 Gram Mouse Weight 0.04

0.02

0.00

DFS = Deferasirox CON = Control

169

DFS did not affect whole body metabolism

Oxymax studies were performed on 8 ob/ob mice in week 22. Contrary to the high fat fed C57Bl/6 mice data, DFS did not affect metabolism in chow-fed ob/ob mice (Figure 4.9). There were no significant differences in oxygen consumption, energy expenditure, respiratory exchange ratio and activity, despite the observation of reduced weight gain on DFS treatment.

170

Figure 4.9 Ob/ob experiment oxymax study at week 22

O2 consumption Respiratory Exchange Ratio

1.1 2400

1 Ml 2000 / RER 0.9 kg 1600

/ 1200 0.8 hr 800 0.7 0 2 4 6 8 10 12 14 16 18 20 22 24 0 2 4 6 8 10 12 14 16 18 20 22 24

Time Time

30 Energy Expenditure per Body Mass 250 Activity X+Yamb

200 Kcal 20 No. of / events 150

Kg 100 10 / 50 hr

0 0 0 2 4 6 8 10 12 14 16 18 20 22 24 0 2 4 6 8 10 12 14 16 18 20 22 24

Time Time

DFS CON

171

DFS-treated ob/ob did not have higher core body temperatures

Rectal temperatures were not significantly different between DFS and CON ob/ob mice throughout the study (Figure 4.10).

172

Figure 4.10 Ob/ob rectal temperatures

n= 5 in each group

Rectal temperature DFS 38.0 CON 37.0

Degree 36.0

Celcius 35.0

34.0

33.0

32.0

31.0 Week 6 Week 12 Week 15 Week 18

DFS = Deferasirox CON = Control

173

There were no significant differences in fat depots

While there were significant differences in weight gain between DFS and CON ob/ob mice, there were no significant differences in visceral (VAT) and subcutaneous (WAT) fat depots (Figure 4.11). There were also no significant differences when adjusted as a percentage of their total body weights.

174

Figure 4.11 Ob/ob organ weights 1

P=0.16

80 DFS

CON 60 Mass

(g) 40

0 Whole body VAT WAT

DFS = Deferasirox CON = Control VAT = Visceral adipose tissue WAT = White adipose tissue

175

DFS-treated ob/ob livers were not lighter

Other organ weights were compared. There were no differences in the pancreas, gastrocnemius or brain weights. Whilst livers from DFS-treated mice were not significantly lighter than those from CON mice, there was a trend to be so (Figure

4.12).

176

Figure 4.12 Ob/ob organ weights 2

P=0.08 4.0

3.5 DFS

3.0 CON

2.5 Mass (g) 2.0

1.5

1.0

0.5

0.0 Liver Pancreas Gastroc Brain

DFS = Deferasirox CON = Control

177

DFS-treated ob/ob have less fatty liver

The CON livers appeared paler compared to DFS livers. When sectioned and stained with routine Haematoxylin and eosin reagents CON livers have more fat globules compared with the DFS treated livers. This suggests that DFS treatment may protect ob/ob mice from severe fatty liver change.

178

Figure 4.13 Ob/ob liver haematoxylin & eosin slide

DFS = Deferasirox CON = Control

179

There were no differences between fasted triglyceride levels

Despite the differences in fatty liver, there were no significant differences in fasting plasma triglycerides. These mice were fasted for four hours prior to sacrifice.

180

Figure 4.14 Ob/ob fasted triglyceride levels

0.6 DFS

Fasted CON Plasma 0.4 Triglycerides (ng per 30mg whole liver) 0.2

0.0

DFS = Deferasirox CON = Control

181

DFS-treated ob/ob livers had less iron staining

As DFS is an iron chelator, it was important to compare iron staining between

DFS and CON mice. The liver is most likely to be affected by iron chelation because of its portal circulation. Liver Kupffer cells are known to ingest iron and store them as ferritin.

Five-micron wide liver sections were made and mounted on slides. Dilute hydrochloric acid was added, releasing ferric ions from tissue binding, and allowing them to bind with ferrocyanide ions instead, to give a Prussian blue reaction. This is the basis of the Perl‟s stain. The sections are then counter stained with nuclear fast staining.

The livers that were treated with DFS had less blue staining, indicating lower iron stores compared with the CON livers.

182

Figure 4.15 Ob/ob liver Perls‟ stain

183

DFS in chow-fed C57Bl/6 did not affect weight gain

Wild-type (WT) C57Bl/6 mice were randomised to receive either blended chow diet with mixed-in deferasirox powder (DFS) or control blended chow diet (CON).

DFS and CON mice were well-matched in age (14 weeks) and in mean starting weight (21.2g+0.5 vs 19.8.5g+0.5, p=0.10). Weight gain from week0 (baseline) was recorded against weeks of chow feeding. DFS treatment did not affect weight change in WT C57Bl/6 mice on chow diet (Figure 4.16).

184

Figure 4.16 Chow-fed wild-type C57Bl/6 weekly weight change

n= 4 in each group

6 Weight 5 Change in Grams 4 DFS 3

2 CON

0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

Time on Chow (Weeks) GTT GTT ITT Oxymax + food intake

DFS = Deferasirox CON = Control GTT = Glucose tolerance test ITT = Insulin tolerance test

185

DFS-treated chow-fed C57Bl/6 tended to have lower random blood glucose concentrations

Tail-nicks were done each week between 12noon to 2pm in order to record random blood glucose concentrations. There was a trend for DFS mice to have lower random glucose readings at week 11 (p=0.08) and significantly lower random blood glucose concentrations at weeks 18 and 20 (p<0.05) (Figure 4.17).

186

Figure 4.17 Chow-fed wild-type C57Bl/6 random blood glucose

n= 4 in each group

14 DFS * 12 CON P=0.08 10 Random * * blood 8 glucose

(mmol/L) 6

0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

Time on Chow (Weeks) GTT GTT

DFS = Deferasirox CON = Control GTT = Glucose tolerance test

187

Glucose tolerance did not differ between DFS and CON mice

However, when these mice were challenged with an intra-peritoneal injection of

20% dextrose at 2g/kg dose, the glucose excursions were not different between

DFS and CON mice (Figure 4.18).

188

Figure 4.18 Chow-fed wild-type C57Bl/6 glucose tolerance test (GTT) week 8

n= 4 in each group

GTT at week 8

25 DFS

20 CON Blood glucose 15

(mmol/L) 10

0 0 20 40 60 80 100 120 minutes

DFS = Deferasirox CON = Control GTT = Glucose tolerance test

189

DFS did not significantly improve insulin sensitivities in chow-fed C57Bl/6 mice

In order to examine whole body insulin sensitivity between both groups,

0.33units/kg insulin was given intra-peritoneally and percentage drop of blood glucose from baseline was recorded. There was a slightly greater lowering of blood glucose in DFS treated mice but this did not reach significance (Figure

4.19). This could be due to the small number of mice tested (N=8).

190

Figure 4.19 Chow-fed wild-type C57Bl/6 insulin tolerance test (ITT)

n= 4 in each group

ITT at week 9 DFS CON 120

100 P=0.052 % blood glucose 80

from 60 baseline 40

0 0 10 20 30 40 50 60 Minutes

DFS = Deferasirox CON = Control ITT = Insulin tolerance test

191

DFS did not significantly increase appetite of chow-fed C57Bl/6

While DFS was shown in Chapter 3 to increase appetite of high fat fed C57Bl/6 mice, it was interesting to note, however, that DFS did not significantly increase food intake in chow-fed C57Bl/6 mice (Figure 4.20). At week 22 of the study, total food intake was measured over 24 hours and adjusted for total body weight.

There was no significant difference between DFS and CON mice.

192

Figure 4.20 Chow-fed wild-type C57Bl/6 24hr food intake study

Chow eaten per body mass

DFS 0.20 CON Chow (g) eaten per gram mouse 0.10 weight

0.00

DFS = Deferasirox CON = Control

193

DFS did not affect whole body metabolism of chow-fed C57Bl/6

At the same time as the food intake study was being carried out in week 22, these mice were put in the Oxymax Respirator system. This consists of closed circuit chambers allowing the measurement of oxygen consumption (VO2), carbon dioxide production (VCO2), as well as activity. The respiratory exchange ratio was calculated from the VO2 and VCO2 (see Methods section).

Unlike the C57Bl/6 mice fed on high fat diet shown in chapter 3, there were no significant differences between DFS and CON mice (Figure 4.21). This data may explain why there were no differences in weight gain.

194

Figure 4.21 Chow-fed wild-type C57Bl/6 oxymax study at week 22

O2 consumption Respiratory exchange ratio

6600 1.2 6200 5800 1.1 Ml 5400 5000 / 4600 RER 1 kg 4200 3800 0.9 / 3400 3000 0.8 hr 2600 2200 1800 0.7 0 2 4 6 8 10 12 14 16 18 20 22 24 0 2 4 6 8 10 12 14 16 18 20 22 24 Time Time

Activity X+Yamb 0.05 Energy expenditure Kcal 8000 7000 / 0.04 No. of 6000 Kg 0.03 events 5000 / 4000 0.02 hr 3000 0.01 2000 1000 0 0 0 2 4 6 8 10 12 14 16 18 20 22 24 Time 0 2 4 6 8 10 12 14 16 18 20 22 24 Time

DFS

CON

195

DFS did not affect core body temperatures of chow-fed C57Bl/6

When rectal temperatures were measured throughout the study, there were no significant differences between both groups of mice (Figure 4.22). Again this suggests that DFS and CON Chow-fed C57Bl/6 mice did not have different whole body metabolism.

196

Figure 4.22 Chow-fed wild-type C57Bl/6 weekly rectal temperatures

n= 5 in each group

Rectal temperatures DFS

CON 38.0

37.0

36.0 Degree Celcius 35.0

34.0

33.0

32.0

31.0 Week 6 Week 12 Week 15 Week 18

DFS = Deferasirox CON = Control

197

DFS-treated chow-fed C57Bl/6 had increased pancreatic and muscle weights but were not different in body weights to CON mice

Upon completion of study at 23 weeks, the mice were sacrificed and their organs examined (Figure 4.23). There were no significant differences in the weights of the visceral fat depots (VAT), subcutaneous fat (WAT) or liver. However, DFS pancreases tended to be slightly heavier (0.21+0.02g vs 0.17+0.01g, p=0.07), and the gastrocnemius muscle was significantly heavier (0.26+0.01g vs

0.23+0.01g, p=0.04).

198

Figure 4.23 Chow-fed wild-type C57Bl/6 organ weights

n= 5 in each group

DFS

CON 1.8 1.6 1.4

Organ 1.2 1.0 weights 0.8 P=0.07 (grams) 0.6 * 0.4 0.2 0.0

DFS = Deferasirox CON = Control VAT = Visceral adipose tissue WAT = White (subcutaneous) adipose tissue BAT = Brown adipose tissue

199

DFS-treated chow-fed C57Bl/6 had significantly lower fasted plasma triglycerides

Mice were fasted for 4-6 hours from early morning prior to sacrifice. Blood was obtained via cardiac puncture and assayed (see methods). These mice were aged 43 weeks approximately.

DFS treated mice had significantly lower plasma triglycerides compared to CON mice: 0.37+0.03ng vs 0.55+0.05ng, p=0.03 (Figure 4.24).

200

Figure 4.24 Chow-fed wild-type C57Bl/6 fasted plasma triglyceride levels

DFS

CON 0.8 0.7 * Fasted plasma 0.6 triglycerides 0.5 (ng per 30mg 0.4 whole liver) 0.3 0.2 0.1 0.0

DFS = Deferasirox CON = Control

201

DFS-treated chow-fed C57Bl/6 were protected from fatty livers

CON livers showed normal age-related steatosis with large fat globules, while the

DFS livers had improved architecture and decreased lipids.

202

Figure 4.25 Chow-fed wild-type C57Bl/6 liver haematoxylin and eosin slide

203

DFS treatment resulted in less liver iron

Iron staining was done using the Perl‟s staining protocol and compared between

DFS treated and CON mice livers. This demonstrated reduced iron staining in the

DFS livers (Figure 4.26), as was shown in the livers of the DFS--treated high fat fed C57Bl/6 mice in Experiment 3.

204

Figure 4.26 Chow-fed wild-type C57Bl/6 liver Perls‟ stain

205

DFS treatment did not result in significantly higher liver HIF-1α

Despite histology showing reduction in iron deposits in DFS-treated livers, HIF-1α protein was not demonstrated to be significantly increased, although there was a slight trend to being so.

206

Figure 4.27 Chow-fed wild-type C57Bl/6 liver western blot

CON DFS

HIF-1α

α tubulin

P=0.08

CON DFS

DFS = Deferasirox CON = Control HIF-1α = Hypoxia inducible factor-1alpha

207

DFS-treated chow-fed C57Bl/6 showed down-regulation of lipolysis gene expression in livers

Gene expression for insulin sensitivity (AKT2, IRS1 and 2), glucose transport

(GLUT1 and 2), MODY (HNF4), glycolysis (PFK) and lipase were examined by extracting RNA and subsequent real-time PCR analyses (Figure 4.28).

There was a trend for higher AKT2 gene expression in the DFS treated livers but significantly reduced gene expression in hormone-sensitive lipase (HSL) and lipoprotein lipase (LPL). The changes in the lipase expression may be secondary to lower levels of plasma triglycerides, as shown earlier.

208

Figure 4.28 Chow-fed wild-type C57Bl/6 liver gene expression1

Liver gene expression1 P=0.09 DFS 2.0 CON

1.5 * * Fold change 1.0

0.5

0.0 AKT2 GLUT1 GLUT2 HNF4α HSL IRS1 IRS2 LPL PFK

AKT2 = Protein kinase B GLUT1 = Glucose transporter 1 GLUT2 = Glucose transporter 2 HNF4α = Hepatocyte nuclear factor-4 alpha HSL = Hormone sensitive lipase IRS1 = Insulin receptor substrate 1 IRS2 = Insulin receptor substrate 2 LPL = Lipoprotein lipase PFK = Phosphofructose kinase

209

DFS-treated chow-fed C57Bl/6 had reduced gene expression for lipid synthesis

Gene expression for lipid synthesis was examined in Figure 4.29. There was highly significant reduction in gene expression in DFS treated livers for ACC1,

FAS, HMG CoAred, HMG CoAsyn, PPARγ and SREBP1c. Whilst CPT1 is not a lipid synthesis gene, it was found to be significantly reduced as well.

These results suggest that decreased lipid synthesis may account for the presence of reduced plasma triglycerides and fatty liver in DFS treated mice.

210

Figure 4.29 Chow-fed wild-type C57Bl/6 liver gene expression2

Liver gene expression2 DFS 2.0 CON

1.5 ** ** ** ** * ** Fold 1.0 change

0.5

0.0

ACC1 = Acetyl-CoA carboxylase 1 ACC2 = Acetyl-CoA carboxylase 2 CPT1 = Carnitine palmitoyltransferase 1 FAS = Fatty acid synthase HMGCoARed = HMG-CoA reductase HMGCoASyn = HMG-CoA synthase PPARγ = Peroxisome proliferstor-activiated receptors-gamma SREBP1c = Sterol regulatory element binding protein-1c

*=p<0.05 ** = P<0.01

211

Summary

In chapter 4, the effects of deferasirox (DFS) were studied in ob/ob mice, a well- known obese mouse model with severe insulin resistance and varying degrees of beta cell failure, resulting in diabetes mellitus. These mice were bred on C57Bl/6 background with ob homozygotes developing the phenotype and the ob heterozygotes behaving like wild –type C57Bl/6 mice.

The findings show that ob/ob mice receiving DFS mixed into their chow diet gained significantly less weight from week 7 of treatment onwards to six months, even though the weight difference was only about 5g. However, unlike the findings from high fat-fed C57Bl/6 mice (as reported in Chapter 3), there were insignificant differences in fat depots, plasma triglyceride levels and whole body metabolism. In addition, this weight difference was not associated with lower random blood glucose levels or improved insulin sensitivity and glucose tolerance. However, the DFS-treated livers tended to be lighter and on histology, noted to have less hepatic iron and fatty change.

In chow-fed C57Bl/6 mice, DFS treatment not only did not affect weight gain but there were also no differences in random blood glucose levels, glucose tolerance or insulin sensitivity. Moreover, DFS did not affect whole body metabolism, with no differences in Oxymax data, appetite or rectal temperatures in the chow-fed

C57Bl/6 mice. However, DFS treatment was again demonstrated to give

212 protection from hepatic steatosis. This was shown to be related to reduced hepatic iron leading to increased HIF-1α protein and reduced gene expression for lipid synthesis.

213

CHAPTER 5

Effects of deferoxamine

on high fat-fed C57Bl/6 mice

214

INTRODUCTION

In Chapters 3 and 4, it was shown that the oral iron chelator, deferasirox (DFS) had significant metabolic benefits in both C57Bl/6 and ObOb when mixed in with their food (high fat diet and normal chow respectively). In this chapter, the effect of an injectable iron chelator, deferoxamine (DFO) was examined on metabolism of wild-type C57Bl/6 mice on high fat diet (HFD). Sixty male mice were housed in

12 cages of 5 mice each. They were fed continuously on HFD.

These mice were weighed weekly and random blood glucose concentrations were recorded for 10 weeks. At week 0, they were randomised to receive either weekly intra-peritoneal DFO adjusted for their body weight (10uL or 125nmol per gram body weight) or to receive equivalent normal saline doses. Glucose tolerance tests (GTT) were performed at week 6 and again at week 10. Glucose stimulated insulin secretion (GSIS) was carried out at week 8.

215

Figure 5.1: Experimental time-line: C57Bl/6 on HFD-DFO experiment

Starting age: 7 to 8 weeks Experimental weeks

-1 0 6 8 10

60 Start HFD GTT GSIS GTT Male 2g/kg 3g/kg 2g/kg mice Check weekly arrive weights and from blood glucose CULL ARC concentrations

Mice randomized to weekly DFO or Saline injections

ARC = Animal resource centre, Perth WA DFO = Deferoxamine HFD = High fat diet GTT = Glucose tolerance test GSIS = Glucose stimulated insulin secretion

216

DFO treatment did not affect weight gain in HFD mice

The C57Bl/6 mice randomised to receive either weekly intra-peritoneal injections

(DFO) or control saline injections (CON) were well matched in starting age (6 weeks) and in mean starting weight (17.2g+0.2 vs 17.2.g+0.2, p=0.956).

Both groups of mice put on weight with HFD. There were slight dips in weight gain at weeks 6 and 8, due to fasting for in-vivo testing. However there were no significant differences in weight gain between both groups (Figure 5.2).

217

Figure 5.2: Weekly weight change

n= 30 in each group

Weight 6 change in grams DFO 4 CON

0 0 2 4 6 8 10

Weeks on HFD

HFD = High fat diet DFO = Deferoxamine CON = Control mice receiving saline injections

218

DFO-treated mice had lower random blood glucose concentrations

Random blood glucose concentrations were obtained once a week at midafternoon via tail nicks. There was a tendency for DFO treated mice to have lower random blood glucose concentrations, with significance at week 1 and 5 of the study (Figure 5.3).

219

Figure 5.3 Random blood glucose

14 DFO 13 * ** CON 12 Random 11 blood 10 glucose concentration 9 (mmol/L) 8

5 1 2 3 4 5 6 7 8 9 10 11

Week on HFD

HFD = High fat diet DFO = Deferoxamine CON = Control mice receiving saline injections

*=p<0.05 ** = P<0.01

220

Glucose tolerance tests (GTTs) were not different between groups

Glucose tolerance testing was performed at week 6 using an intra-peritoneal dose of glucose at 2g/kg body weight. This showed that there were no differences between both groups of mice (Figure 5.4). There were also no differences in body weight. When the glucose tolerance test was repeated at the end of the study (week 10), there was again no difference between both treatment groups (Figure 5.5).

221

Figure 5.4 Glucose tolerance test (GTT) week 6

30 DFO

CON 25

Blood 20 glucose concentration 15 (mmol/L) 10

0 0 20 40 60 80 100 120 Minutes

DFO = Deferoxamine CON = Control mice receiving saline injections GTT = Glucose tolerance test

222

Figure 5.5 Glucose tolerance test (GTT) week 10

DFO 30 CON

Blood 20 glucose concentration 15 (mmol/L) 10

0 0 20 40 60 80 100 120 Minutes

DFO = Deferoxamine CON = Control mice receiving saline injections GTT = Glucose tolerance test

223

DFO-treated mice had improved insulin secretion profile

To evaluate beta cell function in the DFO treated group, glucose stimulated insulin secretion was performed at week 8 of the study. An intra-peritoneal bolus of glucose of 3g/kg was given. Insulin was quantitated using an ELISA. Insulin secretion is shown as percentage of baseline (0 minutes) (Figure 5.6). There was a tendency for increased insulin secretion in the DFO-treated mice, with nonsignificant 1st phase response and for 2nd phase response (P=0.051).

224

Figure 5.6 Glucose stimulated insulin secretion

DFO

500 CON P=0.051

400 % P=NS insulin 300 secretion 200

100

0 0 2 4 6 8 10 12 14 16 18 20

Minutes

NS= Non-significant

DFO = Deferoxamine CON = Control mice receiving saline injections GSIS = Glucose stimulated insulin secretion

225

DFO treatment over 10 weeks resulted in lowering of haemoglobin but not anaemia

Because DFO is an effective intravenous iron chelator, it was important to know if the mice were adversely affected by iron deficiency. Upon sacrifice at week 10, blood was collected and analysed for haemoglobin. There was a slight but significant lowering of haemoglobin in the DFO treated mice but it did not reach anaemic levels (<100g/L) (Figure 5.7). The mice had been observed to be healthy and active prior to sacrifice. Iron levels were not assessed.

226

Figure 5.7 Haemoglobin at end of study

140 *

DFO 135 CON

130 Haemoglobin g/L 125

120

115

110

DFO = Deferoxamine CON = Control mice receiving saline injections

227

Summary

In this chapter, the metabolic effects of deferoxamine (DFO), an intra-peritoneal iron chelator were examined on high fat-fed C57Bl/6 mice. Results contrasted with responses to deferasirox in Chapters 3 and 4 in that there were no differences in weight gain, or glucose tolerance. However, there was a trend to stronger first and second phase insulin secretion to glucose challenge. This suggests a protective effect DFO on beta cells in the presence of continuous high fat feeding.

The relative lack of benefit of this form of iron chelation may possibly be attributed to inadequate iron chelation from the once weekly intra-peritoneal injection. HIF-1α may not be up-regulated a sufficient proportion of overall time to result in important metabolic benefits. However the DFO-treated mice did have a significant lowering of haemoglobin, unlike in deferasirox (DFS). This may be explained by the understanding that DFO is more potent than DFS in reducing the labile plasma (chelatable) iron pool.

228

CHAPTER 6

Effects of deferoxamine

on non-obese diabetic (NOD) mice

229

INTRODUCTION

In Chapter 3, it was shown that iron chelation therapy through oral iron chelation

(deferasirox, DFS) improved glucose tolerance in C57Bl/6 mice fed on high fat diet. This effect is likely to be mediated through reduction of obesity but because there was a small but consistent trend to better insulin secretion profile in treatment with both deferasirox (DFS, Chapter 3) and deferoxamine (DFO,

Chapter 5), it is possible that iron chelation in general, may benefit beta cell function. In this chapter, the effect of deferoxamine (DFO) an injectable iron chelator was examined in Non-Obese Diabetic (NOD) mice. Non-Obese Diabetic

(NOD) mice are a model of Type 1 diabetes, as they spontaneously develop autoimmune destruction of their pancreatic islets. Moreover, they do not become obese when fed continuously on normal mice chow diet. Onset of diabetes usually occurs after 14-15 weeks in about 70% of female mice. 120 mice were randomised over 2 separate experiments, to receive weekly DFO injections, dose adjusted for their body weight (125nmol per gram body weight) or to receive equivalent normal saline doses. The study ran for 47 weeks of treatment

(experiment 1) or 40 weeks (experiment 2). See Figures 6.1 and 6.2. All injections were ceased after 35 weeks and the mice were observed for a further

5 to 12 weeks. Mice were weighed weekly and their random blood glucose concentrations checked through tail nicks. Mice with blood glucose more than

24mmol/L on two consecutive days were considered to have developed diabetes

230 and were sacrificed. Plasma was collected for biochemistry and pancreases for histology.

Figure 6.1: Experimental time-line: NOD experiment 1

Starting age: 7 to 8 weeks Experimental weeks

-1 0 35 47

DFO or Saline injections Stop injections

60 Start CHOW Male diet mice CULL all arrive Check weekly remaining from weights and “Normal” mice ARC blood glucose concentrations

Cull mice with random blood glucose >24mmol/L

ARC = Animal resource centre, Perth WA DFO = Deferoxamine

231

Figure 6.2: Experimental time-line: NOD experiment 2

Starting age: 7 to 8 weeks Experimental weeks

-1 0 35 40

DFO or Saline injections Stop injections

60 Start CHOW Male diet mice CULL all arrive Check weekly remaining from weights and “Normal” mice ARC blood glucose concentrations

Cull mice with random blood glucose >24mmol/L

ARC = Animal resource centre, Perth WA DFO = Deferoxamine

232

DFO delayed onset of diabetes in NOD mice

Results from both experiments were combined and are shown below. Onset of diabetes started approximately 14 weeks into the study in both cohorts. By approximately 23 weeks, the percentage of mice without diabetes was slightly higher in the DFO treated group, reaching statistical significance after 29 weeks

(Figure 6.3).

However with the cessation of DFO injections after week 35, the significance was lost, consistent with the hypothesis that DFO delays onset of diabetes in NOD mice. Approximately 30% of NOD mice in both groups remained non-diabetic.

233

Figure 6.3: DFO delayed onset of diabetes in NOD mice

n= 60 in each group

*=p<0.05

DFO = Deferoxamine CON = Control mice receiving saline injections

234

DFO did not affect islet infiltrate in diabetic mice at sacrifice

Pancreases were collected from mice that had developed diabetes prior to cessation of injections (before week 35). These islets showed varying degrees of immune cell (likely plasma cell) infiltration. Their islets were scored for degree of infiltrate. Score 0 = no infiltrate. Score 1 = surrounding infiltrate. Score 2 = infiltrate inside islet. Score 3 = islet fully infiltrated, but islet cells present. Score 4

= islet fully infiltrated, no islet cells visible.

The results shown in Figure 6.4 were grouped according to low or no infiltrate

(Score 0-1) or high infiltrate (Score 2-4). There were no significant differences between DFO treated islets and control.

235

Figure 6.4 NOD islet infiltrate score

n= 17 mice in each group

DFO

CON 70 P=NS P=NS

60 % 50 of 40 islets 30

0 score 0-1 score 2-4

NS= non-significant

DFO = Deferoxamine CON = Control mice receiving saline injections

236

No significant differences in plasma lipids with DFO treatment

Unlike deferasirox (DFS), treatment with weekly DFO injections did not significantly reduce plasma triglycerides or total cholesterol as shown in Figure

6.5. There was wide inter-individual variability in triglycerides in NOD controls, consistent with previous reports.

237

Figure 6.5 No difference in NOD plasma lipids

NOD plasma lipids DFO

CON 3.5

3.0 P=NS P=NS

2.5

2.0

1.5

1.0

0.5

0.0 Triglycerides Total Cholesterol

NS= non-significant

DFO = Deferoxamine CON = Control mice receiving saline injections

238

Summary

In this final results chapter, the effects of deferoxamine (DFO) were examined on

Non-Obese Diabetic (NOD) mice. From obese mice models of diabetes, it was suggested that iron chelation may have protective effects on beta cell function possibly by mitigating the lipotoxic effects of high fat diet. In NOD mice there is autoimmune mediated destruction of pancreatic islets with an immune cellular infiltrate leading to onset of diabetes in approximately 70% of mice.

In the case of Non-Obese Diabetic mice, once weekly intra-peritoneal injection of

DFO led to a significant delay in the onset of diabetes. This significance was lost within several weeks after DFO injections were ceased. Islet infiltration was similar in both groups that developed diabetes. It is possible that the mechanism of islet preservation is via decreased apoptosis when the drug is being administered, a well known HIF-1α effect. There were no differences in weight gain or plasma lipids.

239

CHAPTER 7

Discussion

240

Effects of iron chelation and hypoxia inducible factor-1 alpha on components of the metabolic syndrome

Individuals with the metabolic syndrome are often obese, have deranged serum lipids, exhibit insulin resistance and glucose intolerance, including frank Type 2 diabetes [4]. There is evidence that increased whole body iron, in particular liver iron, is associated with the metabolic syndrome [154]. While iron is important for cellular metabolism, in particular mitochondrial function (heme synthesis, redox reaction), supra-physiological iron levels are detrimental to cellular function [131].

Iron is a potent inducer of reactive oxygen species (ROS) in the presence of increased oxygen concentration [127]. This, in turn can lead to apoptosis, organ dysfunction, and may lead to obesity and the metabolic syndrome [127].

Hypoxia Inducible Factor-1 alpha (HIF-1α) is a transcription factor that allows cells to survive under hypoxic stress by promoting anaerobic glycolysis, reducing mitochondrial oxygen consumption and therefore reducing ROS generation [79].

Upon hydroxylation by prolyl hydroxylases (PHD) at key proline residues, and in the presence of oxygen and iron, HIF-1α binds with the Von-Hippel Lindau (VHL) protein; becomes ubiquitinated, leading to degradation by proteasomes [70].

When cells are deprived of oxygen or iron, this binding is significantly reduced, resulting in up-regulation of HIF-1α levels [70]. Application of iron chelation drugs to cells lead to increased HIF-1α [34]. This may be achieved by interference of the PHD, which are known to be iron-dependent [72]. For HIF-1α to function optimally, it needs to be associated with transcriptional co-activators CBP and

241 p300 [70]. This activation enhances binding of the HIF-1α+ARNT (HIF-1β) heterodimer (called HIF-1) to hypoxia response elements in gene promoters.

Under normoxic conditions, this co-activation step is inhibited by Factor Inhibiting

HIF-1 (FIH-1) which hydroxylates HIF-1α on an asparagine residue [70].

This thesis presents evidence that iron chelation improves murine glucose, lipid and energy metabolism. The evidence suggests that HIF-1α is responsible for the metabolic benefits. Two types of iron chelators, deferoxamine (DFO) and deferasirox (DFS) were studied. DFO is normally administered as an infusion over several hours for treatment of human iron overload syndromes [131], but in the mice studies, it was given as weekly intra-peritoneal injections. Given it has an intravenous half-life of only 15 minutes, it is not likely that it was an optimal dosing regime, but more frequent dosing was not considered ethically appropriate for the mice. With DFS, it was mixed in with the mice diet and fed to the mice ad libitum. It did not affect their appetite adversely and because it has a longer half-life allowing daily dosing in people, is a better method of administration. Although the results varied slightly across the different mice models, they are consistent with a beneficial effect of iron chelation therapy on modifying the metabolic syndrome. Mice that received DFO injections showed improvements in glucose metabolism, a trend to improved insulin secretion and lower random blood glucose levels albeit with no difference in glucose tolerance.

Mice receiving DFS were protected from high fat- induced obesity, fatty liver change and hyperlipidaemia. DFS-treated livers had evidence of lower hepatic

242 iron levels. Gene expression studies of liver, and muscle showed significant changes in lipid metabolism genes, with reduction of fat synthesis genes and increased lipolytic mRNAs. In this final chapter, the various effects of iron chelation are summarised and discussed. Possible mechanisms for the „Iron

Chelation Effect‟ are discussed and suggestions for future work in this area are put forth.

Oral iron chelation reduced obesity, hepatic and serum lipids

In the C57Bl/6 mice on continuous high fat diet (HFD) model of obesity, it was clear that deferasirox (DFS) reduced weight gain early on in the experiments.

This was due to reduced subcutaneous and visceral fat depots with no differences in muscle weights. The effect tended to occur after 3 weeks of treatment. Mice that were given DFS did not lose weight; rather, they put on less weight compared to the HFD control mice and similarly to the chow-fed controls.

The effect was present in C57Bl/6 mice fed a high fat diet and in the ob/ob mice.

DFS did not alter weight gain in the wild type (WT) C57Bl/6 mice on normal chow diet. The reason for this may be that chow-fed WT mice are not a model of obesity; they have little adipose tissue mass. Therefore the benefit of DFS in reducing fat gain was not apparent. The DFS effect on weight gain was also less apparent in female mice on HFD (experiment 5), with the difference becoming significant from week 23 onwards. It is well known that female C57Bl/6 mice are

243 not as prone to HFD induced obesity was their male littermates [151], and also this study used much smaller numbers than were studied for the males.

In addition to lessening obesity, DFS treatment had effects on reducing serum triglycerides and cholesterol. This effect varied across the three models. HFD-fed

C57Bl/6 mice had a trend to lower total plasma cholesterol but not triglycerides, in non-fasted mice. Fasting plasma triglycerides were significantly lower in chow- fed DFS C57Bl/6 mice but not in DFS ob/ob mice. The reason for this inconsistency is unclear, and may be related to the differences in fasting status and fat handling in these models. However it does suggest that DFS may have an effect on plasma lipids, independent of weight gain effects, since the clearest effect was in the WT mice which had no difference in weight gain.

As the liver is the first solid organ that all oral drugs pass through after absorption

[155], it is likely to be exposed to the highest concentrations of the orally administered DFS [156]. DFS consistently reduced the severity of hepatic steatosis in C57Bl/6 mice on HFD as well as in the ob/ob mice and also decreased accumulation of hepatic lipids in aged wild type C57Bl/6 mice on normal chow diet. This finding was confirmed with reduced liver triglycerides in the C57Bl/6 mice on HFD and normal chow diet.

244

Iron chelation was associated with HIF-1α up-regulation and lipid gene expression changes

Consistent with its known function as an iron chelator, there were reduced but still detectable iron stores in the livers of C57Bl/6 mice receiving DFS. Western blot analyses of the livers showed increased HIF-1α protein, consistent with the known effect of DFS in up-regulating HIF-1α through iron chelation. It is possible that the mechanism of reduced hepatic steatosis and triglycerides is via HIF-1α repression of fat synthesis. In order to explain these findings of improved lipid handling, gene expression studies were carried out. As the liver is an important organ for lipid synthesis and storage, liver gene expression studies were especially important but fat and muscle gene expression were also examined.

There were highly significant lipid gene changes, particularly in the chow-fed

C57Bl/6 mice. In DFS-treated mice, liver gene expression for Acetyl-CoA

Carboxylase 1 (ACC1), Fatty Acid Synthase (FAS), HMG-CoA synthase,

Peroxisome Proliferator-Activated Receptor gamma (PPARγ) Sterol Regulatory

Element Binding Protein-1c (SREBP-1c) were significantly reduced to approximately half that of the control (CON) mice. Because this is associated with significantly reduced plasma lipids and fatty liver in the DFS- treated mice, this suggests that the gene expression changes exert a direct effect on the lipids.

This is an important observation as SREBP-1c is the pivotal transcription factor in regulating lipid synthesis in the liver [80].

245

There are suggestions in the existing literature that up-regulation of HIF-1 may affect lipid metabolism. By applying hypoxic stimuli to up-regulate HIF-1 expression, Choi et al showed that FAS and SREBP-1c mRNAs were reduced in adipocytes, Hep3B human hepatocytes, L6 myocytes and Hepa1c1c7 mouse hepatoma cells [35]. This was associated with reduction of triglyceride and cholesterol content. Other investigators have found that this is achieved via a transcription factor repressor protein called Stra13, also called Differentiated

Embryo Chondrocyte1 (DEC1)[35] . Stra13/DEC1 is shown to be dependent on

HIF-1 activity because when siRNA was used against HIF-1α or ARNT, the hypoxic induction of Stra13/DEC1 expression was reduced. Moreover it was shown that even in normoxic cells, forced expression of HIF-1α or Stra13/DEC1 reduces the expression and activity of SREBP1c [35].

Triglycerides are transported in the plasma component either by chylomicrons from the small intestine to the liver or as Very Low Density Lipoprotein (VLDL) particles from the liver to the adipose tissue. Lipoprotein lipase (LPL) is an essential enzyme for hydrolysing serum triglycerides in chylomicrons or VLDL particles in order to release Non-Esterified Fatty Acids (NEFA) and 2 – monoacylglycerol for tissue utilisation [157]. Its activity is known to be up- regulated by insulin. LPL is found abundantly in vascular endothelium and therefore may be found in vascular organs such as liver, muscle and adipose tissue [157]. In DFS-treated HFD-fed C57Bl/6 mice, liver and muscle but not adipose tissue, LPL expression was consistently increased. This finding may

246 contribute to the observations of reduced hepatic and serum triglycerides in the

DFS treated mice and may reflect improved insulin sensitivity in the liver.

Hormone-sensitive lipase (HSL) hydrolyses adipose tissue triglycerides into diacylglycerols and monoacylglycerols [157]. It is so named because it is highly sensitive to catecholamines and insulin. Catecholamines stimulate HSL by activating G-protein coupled receptors, leading to phosphorylation while insulin de-phosphorylates via phosphatases [158]. In the liver, HSL expression was variably affected by DFS treatment, but there was no significant effect in muscle and adipose tissue. This suggests that HSL plays a lesser role in the „anti-lipid‟ effect of DFS.

Iron chelation protects beta cell function

Beta cell dysfunction is important in the development of type 2 diabetes and is present in obese pre-diabetes states. It is shown in severely obese individuals that lipotoxicity can diminish beta cell reserve [159]. In the high fat-fed C57Bl/6 mice experiments, it was consistently shown that upon glucose stimulation, the first phase insulin secretion was abolished and the second phase blunted.

However, when treated with DFS, the first and second phases were relatively preserved. This is may be contributed in part, to reduction of weight gain in these animals when treated with DFS. However, there was also a trend to better insulin

247 secretion profiles when C57Bl/6 mice were injected with weekly DFO, which did not alter weight gain.

In the non-obese diabetic (NOD) mice model of autoimmune Type 1 diabetes, weekly DFO injection had a significant benefit to prevent diabetes development.

From Week 23 of treatment, it was shown that more mice receiving DFO remained non-diabetic, reaching statistical significance by week 29. This effect was gradually lost after the cessation of the DFO injections. The degree of islet infiltrate was similar in both CON and DFO groups at diabetes diagnosis, suggesting that the protection was not due to immuno-modulation. It is possible that HIF-1α up-regulation may lead to an improved anti-apoptotic response [160].

Oral iron chelation apparently improves hepatic insulin resistance but not whole body insulin resistance

During a period of fasting, the liver maintains euglycaemia via gluconeogenesis and glycogenolysis, processes regulated by insulin action. Thus, fasting hyperinsulinaemia would suggest hepatic insulin resistance. As DFS treatment reduced hepatic steatosis, and therefore probably improved hepatic insulin action, this may explain the lower fasting insulin levels in high fat-fed C57Bl/6 mice treated with DFS.

248

Resistance to insulin action also occurs in other organs such as muscle and adipose tissue, and to test for whole body insulin resistance, intraperitoneal insulin tolerance tests were carried out. These did not consistently show improved lowering of blood glucose concentrations in mice that were treated with

DFS. This suggests that the effects of DFS are not mediated by improved whole- body insulin sensitivity.

Prolonged oral iron chelation protects against glucose intolerance

Despite the observation that DFS may benefit beta cell function, random fed blood glucose levels were not consistently lower in the DFS-treated mice.

Glucose tolerance testing in chow-fed wild type C57Bl/6 mice and ob/ob mice did not show any differences. In the high fat-fed C57Bl/6 mice, early glucose tolerance testing (GTT) carried out from weeks 5 to 8 did not demonstrate any benefit of DFS despite significant weight differences between DFS and CON groups. However, as the weight differences became more exaggerated with prolonged DFS administration, there was a highly significant improvement in glucose tolerance (late) in the DFS-treated group. Therefore DFS-treated mice are protected from developing glucose intolerance or diabetes during prolonged high fat feeding. This is associated with, and may be explained by weight gain prevention and lesser adiposity in the presence of prolonged high fat feeding. In addition, DFS treatment was associated with significantly increased liver gene

249 expression for AKT2, IRS1, IRS2 (insulin signalling), and PFK, but not HNF4a.

There was a trend for improved GLUT2 expression in the liver but not GLUT1.

Using C57Bl/6 mice with a β cell-specific HIF-1α disruption (β-Hif1α-null), Cheng et al demonstrated that these mice have glucose intolerance and β cell dysfunction, which further deteriorated on high fat feeding [33]. When DFS was added to the high fat diets, glucose tolerance improved in control mice but not in

β-Hif1α-null mice. This suggested that HIF-1α was responsible for the improvement in β cell function [33]. When deferoxamine (DFO) was cultured with human islets, glucose-stimulated insulin secretion (GSIS) was significantly increased, compared with control-cultured islets. DFO treated islets had improved gene expression for GLUT1, GLUT2, AKT2, IRS2 and HNF4A [33].

Oral iron chelation affects substrate utilisation and increases whole body metabolism

Mice were put into closed circuit Oxymax metabolic chambers. These experiments yield information such as oxygen consumption (VO2) and carbon dioxide production (VCO2), both of which when increased, would indicate higher whole body metabolism. The ratio of VCO2 and VO2 is used to calculate the

Respiratory Exchange Ratio (RER). A lower RER indicates preferential fat utilization, which is typical for high fat-fed mice. Oxymax studies were performed in high fat-fed C57Bl/6 mice (experiments 2, 3 and 4) as well as in chow-fed ob/ob and normal C57Bl/6 mice. In the high fat-fed C57Bl/6 mice, early studies

250

(weeks 0 and 4), prior to significant weight differences, did not show any differences in the metabolic profile of DFS treated and CON mice. However, after significant weight differences had occurred (weeks 8 and 25), the DFS mice had higher VO2 and VCO2, indicating that DFS treatment may be associated with increased whole body metabolism. Food intake studies were initially carried out to assess the palatability of DFS and the possibility that DFS may diminish the appetite of treated mice, leading to weight loss. Assessments done from weeks 0 to 4 did not show any significant differences, but at weeks 8 and 25, the DFS- treated mice appeared to have eaten more high fat diet (adjusted per body weight). These studies were averaged over three days and adjusted for mice body weights. Whist the studies may not be truly representative of the mice‟s appetite, they do provide a snapshot of their appetite over the period assessed.

At the same time, these mice demonstrated higher core body temperatures as measured using a rectal probe. These findings are consistent with increased whole body metabolic rates. In experiment 3, DFS mice had significantly lower

RER curves, indicating better fat utilisation or improved „fat burning‟. Therefore, it appears that the anti-obesity effect of DFS may be partly explained by its effect on whole body metabolism and fat metabolism. There were no metabolic differences in the ob/ob and normal C57Bl/6 mice when treated with DFS, but their numbers were small. The effect of parenteral DFO on whole body metabolism was not examined.

HIF-1 activity regulates metabolism

251

Recent publications have supported the role of HIF-1α activity on whole body and lipid metabolism. Zhang et al developed a FIH null mice which have increased

HIF-1α activity; this model exhibited increased metabolic rate, improved glucose and lipid metabolism and resisted high fat-induced weight gain when FIH was globally deleted [161]. The FIH null mice gained significantly less weight compared with their control litter-mates when fed a 60% high fat diet. There was no weight difference, however, when they were placed on a normal chow diet.

FIH global deletion mice have higher metabolic rates, significantly lower serum lipids, and less hepatic steatosis [161]. Other investigators have shown that

Orexin induces HIF-1 activity and its activity is HIF-1 dependent [118]. Up- regulation of Orexin in Orexin transgenic mice increased their metabolism and energy expenditure [58].

Ob/ob mice are leptin deficient, leading to un-regulated increase in appetite and extreme weight gain [162]. Because DFS treatment in ob/ob mice led to reduced weight gain, it would seem that this increase in whole body metabolism and energy expenditure via supposed hypothalamic HIF-1up-regulation is not dependent on leptin activity. Therefore it s possible that by directly increasing hypothalamic HIF-1 levels via iron chelation, whole body metabolism and energy expenditure can be subsequently increased (Figure 7.1).

252

Increased Energy Expenditure

Higher CNS centres Anorexigenic Autonomic NS Orexigenic Pathways Pathways

HIF-1 CRH MSH Second TRH Orexin A, B PVN LHA Order GLP-1 Neurons

- + + -

NPY & POMC & First Order AGRP ARC CART Neurons neurons neurons

- Leptin + Insulin

CNS = Central nervous system MSH = Melanocyte stimulating hormone

HIF = Hypoxia inducible factor LHA = Lateral hypothalamus area PVN = Paraventricular neurons NPY = Neuro - peptide Y CRH = Corticotropin - releasing hormone AGRP = Agouti gene related peptide TRH = Thyrotropin - releasing hormone ARC = Arcuate nucleus

GLP-1 = Glucagon - like peptide-1 POMC = Pro - opiomelanocortin CART = Cocaine and amphetamine-regulated transcript

Figure 7.1 Proposed mechanism whereby up-regulation of hypothalamic HIF-1α

may increase Orexin activity leading to increased metabolism.

253

Figure 7.2 Summary of HIF-1 actions on glucose and lipid metabolism

Iron Hypoxia Chelation

 Iron  Oxygen  HIF-1

Pancreatic Liver Muscle Hypothalamus islets

 Lipogenic  Anti- Protein Energy apoptosis Synthesis Expenditure

   β cell Serum Fat Gain, survival Triglycerides, Obesity Fatty Liver

 Insulin Sensitivity

254

The efficacy of DFO versus DFS

It was interesting to note that oral deferasirox (DFS) but not parenteral deferoxamine (DFO) increased whole body metabolism of the high fat fed

C57Bl/6 mice. These observations may be explained by the possibility that when

DFS is mixed in with high fat diet and fed continuously to the mice, this represents the most efficient drug delivery and thus method of iron chelation.

Also, as explained previously, there is increased exposure to the liver when oral iron chelators are administered. On the other hand, weekly intra-peritoneal DFO injections may not be adequate to maintain sustained drug levels to achieve adequate effect. Daily intra-peritoneal injections were deemed too invasive and therefore not ethically possible to use in the mice experiments. For future experiments, it may be possible to insert subcutaneous mini-osmotic pumps into the mice to achieve more sustained DFO levels, but this would require repeated re-insertions.

For the ob/ob and Chow-fed C57Bl/6 mice, DFS powder was mixed in with blended chow and the mixture was placed in Petri dishes on the cage floor. The dishes were changed daily, as the mice often played with the food, spilling them.

Therefore it is possible that the DFS doses received by were more variable. The ob/ob mice did achieve a highly significant reduction in weight gain when treated with DFS. However, no improvement in whole body metabolic rate occurred in the DFS-treated ob/ob or Chow-fed C57Bl/6 mice. There were no weight changes in the chow-fed C57Bl/6 mice. Because there was significant down- 255 regulation of lipogenic gene expression in the muscles of chow-fed mice, it is possible that peripheral mechanisms were more important in reducing obesity and related complications in non-high fat-fed mice.

Iron chelation is well-tolerated

Because iron is important to cellular metabolism and in particular, vital to heme synthesis and erythropoiesis, it was crucial to determine whether any toxicities were associated with the use of iron chelation in animals without iron overload.

The main concern was that of iron deficiency anaemia. In the high fat-fed

C57Bl/6 mice that received DFO, full blood count was obtained upon sacrifice at

10 weeks of treatment. This showed a slight lowering of haemoglobin, but did not result in anaemia. When full blood count was obtained from a cohort of mice that received DFS at 10 weeks and another that received DFS for 35 weeks, there were no significant differences between the groups.

These observations are interesting because the existing literature suggests that while DFO is efficient at removing the labile plasma iron pool [163], DFS is more efficient at removing tissue iron. DFS may act to re-distribute tissue iron into the serum „chelatable‟ pool [164]; so that serum iron concentrations actually increase with DFS-mediated iron chelation. This effect is beneficial as iron moved from vital organs, such as the liver, (where iron-generated ROS may cause oxidative damage), can be distributed to the plasma compartment, where iron may be available to the bone marrow for erythropoiesis.

256

Conclusion and suggestions for future work

Through a series of experiments on mice models of obesity and insulin resistance, as well as in autoimmune diabetes, the results in this thesis have shown that iron chelation is both beneficial and well tolerated. The benefits included better weight control and reduced hepatic steatosis from reduced fat synthesis and improved lipid handling, ultimately resulting in improved glucose tolerance and less (hepatic) insulin resistance.

The results have shown that oral iron chelation resulted in reduced hepatic iron levels and up-regulation of Hypoxia-inducible Factor-1 α, which in turn led to reduced gene expression for fat synthesis and glucose handling. In addition there may be improvement in whole body metabolism, possibly mediated via increasing hypothalamic HIF-1α levels. Parenteral iron chelation had beneficial effects on conserving beta cell function in Non-Obese Diabetic (NOD) mice. More studies are needed to delineate the role of hypothalamic HIF-1α in increasing whole body metabolism. These may involve deleting hypothalamic HIF-1α in

HFD-fed mice using adenoviral Cre, randomising them to DFS treatment and performing metabolic studies.

Iron chelation is well tolerated and while parenteral iron chelation results in lowering of haemoglobin concentration, it does not result in iron deficiency anaemia over ten weeks of treatment. Oral iron chelation did not affect

257 haemoglobin concentration over nine months (35 weeks) of treatment and does not adversely affect appetite. Therefore iron chelation is a promising therapy for improving the metabolic syndrome in humans. As both forms of iron chelation are currently approved for use in humans, it may be possible to carry out Phase 2 studies to determine the optimum, non-toxic dose for treatment in non-iron overload obese and/ or Type 2 diabetic individuals and to determine the extent of the metabolic benefit in phase 3 studies.

258

References

1. Kouris-Blazos, A. and M.L. Wahlqvist, Health economics of weight management: evidence and cost. Asia Pac J Clin Nutr, 2007. 16 Suppl 1: p. 329-38. 2. Eckel, R.H., S.M. Grundy, and P.Z. Zimmet, The metabolic syndrome. Lancet, 2005. 365(9468): p. 1415-28. 3. Reaven, G.M., Banting lecture 1988. Role of insulin resistance in human disease. Diabetes, 1988. 37(12): p. 1595-607. 4. Cornier, M.A., et al., The metabolic syndrome. Endocr Rev, 2008. 29(7): p. 777- 822. 5. Isomaa, B., et al., Cardiovascular morbidity and mortality associated with the metabolic syndrome. Diabetes Care, 2001. 24(4): p. 683-9. 6. Lakka, H.M., et al., The metabolic syndrome and total and cardiovascular disease mortality in middle-aged men. Jama, 2002. 288(21): p. 2709-16. 7. Girman, C.J., et al., The metabolic syndrome and risk of major coronary events in the Scandinavian Simvastatin Survival Study (4S) and the Air Force/Texas Coronary Atherosclerosis Prevention Study (AFCAPS/TexCAPS). Am J Cardiol, 2004. 93(2): p. 136-41. 8. Malik, S., et al., Impact of the metabolic syndrome on mortality from coronary heart disease, cardiovascular disease, and all causes in United States adults. Circulation, 2004. 110(10): p. 1245-50. 9. Sundstrom, J., et al., Clinical value of the metabolic syndrome for long term prediction of total and cardiovascular mortality: prospective, population based cohort study. Bmj, 2006. 332(7546): p. 878-82. 10. Cameron, A.J., et al., The metabolic syndrome: in need of a global mission statement. Diabet Med, 2009. 26(3): p. 306-9. 11. Malaguarnera, M., et al., Molecular mechanisms involved in NAFLD progression. J Mol Med, 2009. 12. McCullough, A.J., Pathophysiology of nonalcoholic steatohepatitis. J Clin Gastroenterol, 2006. 40 Suppl 1: p. S17-29. 13. Bastard, J.P., et al., Recent advances in the relationship between obesity, inflammation, and insulin resistance. Eur Cytokine Netw, 2006. 17(1): p. 4-12. 14. George, J. and C. Liddle, Nonalcoholic fatty liver disease: pathogenesis and potential for nuclear receptors as therapeutic targets. Mol Pharm, 2008. 5(1): p. 49-59. 15. Begriche, K., et al., Mitochondrial dysfunction in NASH: causes, consequences and possible means to prevent it. Mitochondrion, 2006. 6(1): p. 1-28. 16. Sun, L., et al., Ferritin concentrations, metabolic syndrome, and type 2 diabetes in middle-aged and elderly chinese. J Clin Endocrinol Metab, 2008. 93(12): p. 4690-6. 17. Sharifi, F., N.M. Nasab, and H.J. Zadeh, Elevated serum ferritin concentrations in prediabetic subjects. Diab Vasc Dis Res, 2008. 5(1): p. 15-8. 18. Ruivard, M., [Genetic iron overloads and hepatic insulin-resistance iron overload syndrome: an update]. Rev Med Interne, 2009. 30(1): p. 35-42. 19. Wilson, J.G., Iron and glucose homeostasis: new lessons from hereditary haemochromatosis. Diabetologia, 2006. 49(7): p. 1459-61.

259

20. McClain, D.A., et al., High prevalence of abnormal glucose homeostasis secondary to decreased insulin secretion in individuals with hereditary haemochromatosis. Diabetologia, 2006. 49(7): p. 1661-9. 21. Syrovatka, P., et al., Relationship between increased body iron stores, oxidative stress and insulin resistance in healthy men. Ann Nutr Metab, 2009. 54(4): p. 268-74. 22. Forouhi, N.G., et al., Elevated serum ferritin levels predict new-onset type 2 diabetes: results from the EPIC-Norfolk prospective study. Diabetologia, 2007. 50(5): p. 949-56. 23. Acton, R.T., et al., Relationships of serum ferritin, transferrin saturation, and HFE mutations and self-reported diabetes in the Hemochromatosis and Iron Overload Screening (HEIRS) study. Diabetes Care, 2006. 29(9): p. 2084-9. 24. Salonen, J.T., et al., Relation between iron stores and non-insulin dependent diabetes in men: case-control study. Bmj, 1998. 317(7160): p. 727. 25. Martinez-Garcia, M.A., et al., Body iron stores and glucose intolerance in premenopausal women: role of hyperandrogenism, insulin resistance and genomic variants related to inflammation, oxidative stress and iron metabolism. Diabetes Care, 2009. 26. Gillum, R.F., Association of serum ferritin and indices of body fat distribution and obesity in Mexican American men--the Third National Health and Nutrition Examination Survey. Int J Obes Relat Metab Disord, 2001. 25(5): p. 639-45. 27. Alexander, J. and K.V. Kowdley, HFE-associated hereditary hemochromatosis. Genet Med, 2009. 11(5): p. 307-13. 28. Equitani, F., et al., Bloodletting ameliorates insulin sensitivity and secretion in parallel to reducing liver iron in carriers of HFE gene mutations. Diabetes Care, 2008. 31(1): p. 3-8. 29. Valenti, L., et al., Iron depletion by phlebotomy improves insulin resistance in patients with nonalcoholic fatty liver disease and hyperferritinemia: evidence from a case-control study. Am J Gastroenterol, 2007. 102(6): p. 1251-8. 30. Stangl, G.I. and M. Kirchgessner, Different degrees of moderate iron deficiency modulate lipid metabolism of rats. Lipids, 1998. 33(9): p. 889-95. 31. Gunton, J.E., et al., Loss of ARNT/HIF1beta mediates altered gene expression and pancreatic-islet dysfunction in human type 2 diabetes. Cell, 2005. 122(3): p. 337- 49. 32. Wang, X.L., et al., Ablation of ARNT/HIF1beta in liver alters gluconeogenesis, lipogenic gene expression, and serum ketones. Cell Metab, 2009. 9(5): p. 428-39. 33. Cheng, K., et al., Hypoxia-inducible factor-1alpha regulates beta cell function in mouse and human islets. J Clin Invest. 34. Dongiovanni, P., et al., Iron depletion by deferoxamine up-regulates glucose uptake and insulin signaling in hepatoma cells and in rat liver. Am J Pathol, 2008. 172(3): p. 738-47. 35. Choi, S.M., et al., Stra13/DEC1 and DEC2 inhibit sterol regulatory element binding protein-1c in a hypoxia-inducible factor-dependent mechanism. Nucleic Acids Res, 2008. 36(20): p. 6372-85.

260

36. Yun, Z., et al., Inhibition of PPAR gamma 2 gene expression by the HIF-1- regulated gene DEC1/Stra13: a mechanism for regulation of adipogenesis by hypoxia. Dev Cell, 2002. 2(3): p. 331-41. 37. Arora, S. and Anubhuti, Role of neuropeptides in appetite regulation and obesity- -a review. Neuropeptides, 2006. 40(6): p. 375-401. 38. Levine, J.A., N.L. Eberhardt, and M.D. Jensen, Role of nonexercise activity thermogenesis in resistance to fat gain in humans. Science, 1999. 283(5399): p. 212-4. 39. Miller, D.S., P. Mumford, and M.J. Stock, Gluttony. 2. Thermogenesis in overeating man. Am J Clin Nutr, 1967. 20(11): p. 1223-9. 40. Ravussin, E., et al., Determinants of 24-hour energy expenditure in man. Methods and results using a respiratory chamber. J Clin Invest, 1986. 78(6): p. 1568-78. 41. Wilding, J.P., Neuropeptides and appetite control. Diabet Med, 2002. 19(8): p. 619-27. 42. Williams, G., et al., The hypothalamus and the control of energy homeostasis: different circuits, different purposes. Physiol Behav, 2001. 74(4-5): p. 683-701. 43. Hahn, T.M., et al., Coexpression of Agrp and NPY in fasting-activated hypothalamic neurons. Nat Neurosci, 1998. 1(4): p. 271-2. 44. Kristensen, P., et al., Hypothalamic CART is a new anorectic peptide regulated by leptin. Nature, 1998. 393(6680): p. 72-6. 45. Broadwell, R.D. and M.W. Brightman, Entry of peroxidase into neurons of the central and peripheral nervous systems from extracerebral and cerebral blood. J Comp Neurol, 1976. 166(3): p. 257-83. 46. Elmquist, J.K., et al., Unraveling the central nervous system pathways underlying responses to leptin. Nat Neurosci, 1998. 1(6): p. 445-50. 47. Crowley, V.E., Overview of human obesity and central mechanisms regulating energy homeostasis. Ann Clin Biochem, 2008. 45(Pt 3): p. 245-55. 48. Elmquist, J.K., C.F. Elias, and C.B. Saper, From lesions to leptin: hypothalamic control of food intake and body weight. Neuron, 1999. 22(2): p. 221-32. 49. Arase, K., et al., Effects of corticotropin-releasing factor on food intake and brown adipose tissue thermogenesis in rats. Am J Physiol, 1988. 255(3 Pt 1): p. E255-9. 50. Stellar, E., The physiology of motivation. Psychol Rev, 1954. 61(1): p. 5-22. 51. Meister, B., et al., Neurotransmitters, neuropeptides and binding sites in the rat mediobasal hypothalamus: effects of monosodium glutamate (MSG) lesions. Exp Brain Res, 1989. 76(2): p. 343-68. 52. Christophe, J., Is there appetite after GLP-1 and PACAP? Ann N Y Acad Sci, 1998. 865: p. 323-35. 53. Horvath, T.L., S. Diano, and A.N. van den Pol, Synaptic interaction between hypocretin (orexin) and neuropeptide Y cells in the rodent and primate hypothalamus: a novel circuit implicated in metabolic and endocrine regulations. J Neurosci, 1999. 19(3): p. 1072-87. 54. Bernardis, L.L. and L.L. Bellinger, The lateral hypothalamic area revisited: ingestive behavior. Neurosci Biobehav Rev, 1996. 20(2): p. 189-287.

261

55. Sakurai, T., et al., Orexins and orexin receptors: a family of hypothalamic neuropeptides and G protein-coupled receptors that regulate feeding behavior. Cell, 1998. 92(5): p. 1 page following 696. 56. Hara, J., M. Yanagisawa, and T. Sakurai, Difference in obesity phenotype between orexin-knockout mice and orexin neuron-deficient mice with same genetic background and environmental conditions. Neurosci Lett, 2005. 380(3): p. 239- 42. 57. Nishino, S., Clinical and neurobiological aspects of narcolepsy. Sleep Med, 2007. 8(4): p. 373-99. 58. Funato, H., et al., Enhanced orexin receptor-2 signaling prevents diet-induced obesity and improves leptin sensitivity. Cell Metab, 2009. 9(1): p. 64-76. 59. Semenza, G.L., HIF-1 and human disease: one highly involved factor. Genes Dev, 2000. 14(16): p. 1983-91. 60. Semenza, G.L., et al., Hypoxia-inducible nuclear factors bind to an enhancer element located 3' to the human erythropoietin gene. Proc Natl Acad Sci U S A, 1991. 88(13): p. 5680-4. 61. Semenza, G.L. and G.L. Wang, A nuclear factor induced by hypoxia via de novo protein synthesis binds to the human erythropoietin gene enhancer at a site required for transcriptional activation. Mol Cell Biol, 1992. 12(12): p. 5447-54. 62. Wang, G.L. and G.L. Semenza, General involvement of hypoxia-inducible factor 1 in transcriptional response to hypoxia. Proc Natl Acad Sci U S A, 1993. 90(9): p. 4304-8. 63. Iyer, N.V., et al., Cellular and developmental control of O2 homeostasis by hypoxia-inducible factor 1 alpha. Genes Dev, 1998. 12(2): p. 149-62. 64. Wang, G.L., et al., Hypoxia-inducible factor 1 is a basic-helix-loop-helix-PAS heterodimer regulated by cellular O2 tension. Proc Natl Acad Sci U S A, 1995. 92(12): p. 5510-4. 65. Hoffman, E.C., et al., Cloning of a factor required for activity of the Ah (dioxin) receptor. Science, 1991. 252(5008): p. 954-8. 66. Jiang, B.H., et al., Hypoxia-inducible factor 1 levels vary exponentially over a physiologically relevant range of O2 tension. Am J Physiol, 1996. 271(4 Pt 1): p. C1172-80. 67. Salceda, S. and J. Caro, Hypoxia-inducible factor 1alpha (HIF-1alpha) protein is rapidly degraded by the ubiquitin-proteasome system under normoxic conditions. Its stabilization by hypoxia depends on redox-induced changes. J Biol Chem, 1997. 272(36): p. 22642-7. 68. Maxwell, P.H., et al., The tumour suppressor protein VHL targets hypoxia- inducible factors for oxygen-dependent proteolysis. Nature, 1999. 399(6733): p. 271-5. 69. Kamura, T., et al., Activation of HIF1alpha ubiquitination by a reconstituted von Hippel-Lindau (VHL) tumor suppressor complex. Proc Natl Acad Sci U S A, 2000. 97(19): p. 10430-5. 70. Jaakkola, P., et al., Targeting of HIF-alpha to the von Hippel-Lindau ubiquitylation complex by O2-regulated prolyl hydroxylation. Science, 2001. 292(5516): p. 468-72.

262

71. Lando, D., et al., FIH-1 is an asparaginyl hydroxylase enzyme that regulates the transcriptional activity of hypoxia-inducible factor. Genes Dev, 2002. 16(12): p. 1466-71. 72. Epstein, A.C., et al., C. elegans EGL-9 and mammalian homologs define a family of dioxygenases that regulate HIF by prolyl hydroxylation. Cell, 2001. 107(1): p. 43-54. 73. Roy, A., et al., Effects of iron-chelators on ion-channels and HIF-1alpha in the carotid body. Respir Physiol Neurobiol, 2004. 141(2): p. 115-23. 74. Hamrick, S.E., et al., A role for hypoxia-inducible factor-1alpha in desferoxamine neuroprotection. Neurosci Lett, 2005. 379(2): p. 96-100. 75. Weidemann, A. and R.S. Johnson, Biology of HIF-1alpha. Cell Death Differ, 2008. 15(4): p. 621-7. 76. Firth, J.D., et al., Oxygen-regulated control elements in the phosphoglycerate kinase 1 and lactate dehydrogenase A genes: similarities with the erythropoietin 3' enhancer. Proc Natl Acad Sci U S A, 1994. 91(14): p. 6496-500. 77. Bartrons, R. and J. Caro, Hypoxia, glucose metabolism and the Warburg's effect. J Bioenerg Biomembr, 2007. 39(3): p. 223-9. 78. Papandreou, I., et al., HIF-1 mediates adaptation to hypoxia by actively downregulating mitochondrial oxygen consumption. Cell Metab, 2006. 3(3): p. 187-97. 79. Fukuda, R., et al., HIF-1 regulates cytochrome oxidase subunits to optimize efficiency of respiration in hypoxic cells. Cell, 2007. 129(1): p. 111-22. 80. Horton, J.D., J.L. Goldstein, and M.S. Brown, SREBPs: activators of the complete program of cholesterol and fatty acid synthesis in the liver. J Clin Invest, 2002. 109(9): p. 1125-31. 81. Brown, M.S. and J.L. Goldstein, The SREBP pathway: regulation of cholesterol metabolism by proteolysis of a membrane-bound transcription factor. Cell, 1997. 89(3): p. 331-40. 82. Shimomura, I., et al., Differential expression of exons 1a and 1c in mRNAs for sterol regulatory element binding protein-1 in human and mouse organs and cultured cells. J Clin Invest, 1997. 99(5): p. 838-45. 83. Shimomura, I., et al., Nuclear sterol regulatory element-binding proteins activate genes responsible for the entire program of unsaturated fatty acid biosynthesis in transgenic mouse liver. J Biol Chem, 1998. 273(52): p. 35299-306. 84. Horton, J.D., et al., Activation of cholesterol synthesis in preference to fatty acid synthesis in liver and adipose tissue of transgenic mice overproducing sterol regulatory element-binding protein-2. J Clin Invest, 1998. 101(11): p. 2331-9. 85. Shimano, H., et al., Overproduction of cholesterol and fatty acids causes massive liver enlargement in transgenic mice expressing truncated SREBP-1a. J Clin Invest, 1996. 98(7): p. 1575-84. 86. Shimano, H., et al., Isoform 1c of sterol regulatory element binding protein is less active than isoform 1a in livers of transgenic mice and in cultured cells. J Clin Invest, 1997. 99(5): p. 846-54. 87. Horton, J.D., et al., Regulation of sterol regulatory element binding proteins in livers of fasted and refed mice. Proc Natl Acad Sci U S A, 1998. 95(11): p. 5987- 92.

263

88. Matsuda, M., et al., SREBP cleavage-activating protein (SCAP) is required for increased lipid synthesis in liver induced by cholesterol deprivation and insulin elevation. Genes Dev, 2001. 15(10): p. 1206-16. 89. Liang, G., et al., Diminished hepatic response to fasting/refeeding and liver X receptor agonists in mice with selective deficiency of sterol regulatory element- binding protein-1c. J Biol Chem, 2002. 277(11): p. 9520-8. 90. Shimano, H., et al., Sterol regulatory element-binding protein-1 as a key transcription factor for nutritional induction of lipogenic enzyme genes. J Biol Chem, 1999. 274(50): p. 35832-9. 91. Shimomura, I., Y. Bashmakov, and J.D. Horton, Increased levels of nuclear SREBP-1c associated with fatty livers in two mouse models of diabetes mellitus. J Biol Chem, 1999. 274(42): p. 30028-32. 92. Shimomura, I., et al., Leptin reverses insulin resistance and diabetes mellitus in mice with congenital lipodystrophy. Nature, 1999. 401(6748): p. 73-6. 93. Lin, H.Z., et al., Metformin reverses fatty liver disease in obese, leptin-deficient mice. Nat Med, 2000. 6(9): p. 998-1003. 94. Zhou, G., et al., Role of AMP-activated protein kinase in mechanism of metformin action. J Clin Invest, 2001. 108(8): p. 1167-74. 95. Gillian-Daniel, D.L., et al., Endoplasmic reticulum localization of the low density lipoprotein receptor mediates presecretory degradation of apolipoprotein B. Proc Natl Acad Sci U S A, 2002. 99(7): p. 4337-42. 96. Horton, J.D., et al., Disruption of LDL receptor gene in transgenic SREBP-1a mice unmasks hyperlipidemia resulting from production of lipid-rich VLDL. J Clin Invest, 1999. 103(7): p. 1067-76. 97. Yang, J., et al., Decreased lipid synthesis in livers of mice with disrupted Site-1 protease gene. Proc Natl Acad Sci U S A, 2001. 98(24): p. 13607-12. 98. Sato, R., et al., Sterol-dependent transcriptional regulation of sterol regulatory element-binding protein-2. J Biol Chem, 1996. 271(43): p. 26461-4. 99. Amemiya-Kudo, M., et al., Promoter analysis of the mouse sterol regulatory element-binding protein-1c gene. J Biol Chem, 2000. 275(40): p. 31078-85. 100. Janowski, B.A., et al., Structural requirements of ligands for the oxysterol liver X receptors LXRalpha and LXRbeta. Proc Natl Acad Sci U S A, 1999. 96(1): p. 266-71. 101. Repa, J.J., et al., Regulation of mouse sterol regulatory element-binding protein- 1c gene (SREBP-1c) by oxysterol receptors, LXRalpha and LXRbeta. Genes Dev, 2000. 14(22): p. 2819-30. 102. Xu, J., et al., Sterol regulatory element binding protein-1 expression is suppressed by dietary polyunsaturated fatty acids. A mechanism for the coordinate suppression of lipogenic genes by polyunsaturated fats. J Biol Chem, 1999. 274(33): p. 23577-83. 103. Ou, J., et al., Unsaturated fatty acids inhibit transcription of the sterol regulatory element-binding protein-1c (SREBP-1c) gene by antagonizing ligand-dependent activation of the LXR. Proc Natl Acad Sci U S A, 2001. 98(11): p. 6027-32. 104. Xu, J., et al., Polyunsaturated fatty acids suppress hepatic sterol regulatory element-binding protein-1 expression by accelerating transcript decay. J Biol Chem, 2001. 276(13): p. 9800-7.

264

105. Pai, J.T., et al., Differential stimulation of cholesterol and unsaturated fatty acid biosynthesis in cells expressing individual nuclear sterol regulatory element- binding proteins. J Biol Chem, 1998. 273(40): p. 26138-48. 106. Latasa, M.J., et al., Nutritional regulation of the fatty acid synthase promoter in vivo: sterol regulatory element binding protein functions through an upstream region containing a sterol regulatory element. Proc Natl Acad Sci U S A, 2000. 97(19): p. 10619-24. 107. Joseph, S.B., et al., Direct and indirect mechanisms for regulation of fatty acid synthase gene expression by liver X receptors. J Biol Chem, 2002. 277(13): p. 11019-25. 108. Schadinger, S.E., et al., PPARgamma2 regulates lipogenesis and lipid accumulation in steatotic hepatocytes. Am J Physiol Endocrinol Metab, 2005. 288(6): p. E1195-205. 109. Fajas, L., et al., Regulation of peroxisome proliferator-activated receptor gamma expression by adipocyte differentiation and determination factor 1/sterol regulatory element binding protein 1: implications for adipocyte differentiation and metabolism. Mol Cell Biol, 1999. 19(8): p. 5495-503. 110. Gavrilova, O., et al., Liver peroxisome proliferator-activated receptor gamma contributes to hepatic steatosis, triglyceride clearance, and regulation of body fat mass. J Biol Chem, 2003. 278(36): p. 34268-76. 111. Li, J., et al., Altered metabolic responses to intermittent hypoxia in mice with partial deficiency of hypoxia-inducible factor-1alpha. Physiol Genomics, 2006. 25(3): p. 450-7. 112. Hochachka, P.W., et al., Unifying theory of hypoxia tolerance: molecular/metabolic defense and rescue mechanisms for surviving oxygen lack. Proc Natl Acad Sci U S A, 1996. 93(18): p. 9493-8. 113. Gentil, C., et al., Is oxygen a key factor in the lipodystrophy phenotype? Lipids Health Dis, 2006. 5: p. 27. 114. Seagroves, T.N., et al., Transcription factor HIF-1 is a necessary mediator of the pasteur effect in mammalian cells. Mol Cell Biol, 2001. 21(10): p. 3436-44. 115. Holmquist-Mengelbier, L., et al., Recruitment of HIF-1alpha and HIF-2alpha to common target genes is differentially regulated in neuroblastoma: HIF-2alpha promotes an aggressive phenotype. Cancer Cell, 2006. 10(5): p. 413-23. 116. Miyazaki, K., et al., Identification of functional hypoxia response elements in the promoter region of the DEC1 and DEC2 genes. J Biol Chem, 2002. 277(49): p. 47014-21. 117. Li, Y., et al., DEC1 negatively regulates the expression of DEC2 through binding to the E-box in the proximal promoter. J Biol Chem, 2003. 278(19): p. 16899-907. 118. Sikder, D. and T. Kodadek, The neurohormone orexin stimulates hypoxia- inducible factor-1 activity. Genes Dev, 2007. 21(22): p. 2995-3005. 119. Semenza, G.L., Oxygen-dependent regulation of mitochondrial respiration by hypoxia-inducible factor 1. Biochem J, 2007. 405(1): p. 1-9. 120. Kim, J.W., et al., HIF-1-mediated expression of pyruvate dehydrogenase kinase: a metabolic switch required for cellular adaptation to hypoxia. Cell Metab, 2006. 3(3): p. 177-85.

265

121. Simon, M.C., Coming up for air: HIF-1 and mitochondrial oxygen consumption. Cell Metab, 2006. 3(3): p. 150-1. 122. Umbreit, J., Iron deficiency: a concise review Am J Hematol, 2005. 78(3): p. 225- 31. 123. Powell, J.J., R. Jugdaohsingh, and R.P. Thompson, The regulation of mineral absorption in the gastrointestinal tract. Proc Nutr Soc, 1999. 58(1): p. 147-53. 124. Simovich, M., et al., Localization of the iron transport proteins Mobilferrin and DMT-1 in the duodenum: the surprising role of mucin. Am J Hematol, 2003. 74(1): p. 32-45. 125. Aisen, P., C. Enns, and M. Wessling-Resnick, Chemistry and biology of eukaryotic iron metabolism. Int J Biochem Cell Biol, 2001. 33(10): p. 940-59. 126. McKie, A.T., et al., A novel duodenal iron-regulated transporter, IREG1, implicated in the basolateral transfer of iron to the circulation. Mol Cell, 2000. 5(2): p. 299-309. 127. Levi, S. and E. Rovida, The role of iron in mitochondrial function. Biochim Biophys Acta, 2009. 1790(7): p. 629-36. 128. Zorov, D.B., et al., Mitochondria revisited. Alternative functions of mitochondria. Biosci Rep, 1997. 17(6): p. 507-20. 129. Pozzan, T., et al., Molecular and cellular physiology of intracellular calcium stores. Physiol Rev, 1994. 74(3): p. 595-636. 130. Napier, I., P. Ponka, and D.R. Richardson, Iron trafficking in the mitochondrion: novel pathways revealed by disease. Blood, 2005. 105(5): p. 1867-74. 131. Porter, J.B., Concepts and goals in the management of transfusional iron overload. Am J Hematol, 2007. 82(12 Suppl): p. 1136-9. 132. Brittenham, G.M., et al., Efficacy of deferoxamine in preventing complications of iron overload in patients with thalassemia major. N Engl J Med, 1994. 331(9): p. 567-73. 133. Crichton, R.R., F. Roman, and F. Roland, Ferritin iron mobilisation by chelating agents. FEBS Lett, 1980. 110(2): p. 271-4. 134. Hershko, C. and E.A. Rachmilewitz, Mechanism of desferrioxamine-induced iron excretion in thalassaemia. Br J Haematol, 1979. 42(1): p. 125-32. 135. Collins, A.F., et al., Iron-balance and dose-response studies of the oral iron chelator 1,2-dimethyl-3-hydroxypyrid-4-one (L1) in iron-loaded patients with sickle cell disease. Blood, 1994. 83(8): p. 2329-33. 136. Nisbet-Brown, E., et al., Effectiveness and safety of ICL670 in iron-loaded patients with thalassaemia: a randomised, double-blind, placebo-controlled, dose-escalation trial. Lancet, 2003. 361(9369): p. 1597-602. 137. Porter, J.B., et al., Kinetics of removal and reappearance of non-transferrin- bound plasma iron with deferoxamine therapy. Blood, 1996. 88(2): p. 705-13. 138. Glickstein, H., et al., Intracellular labile iron pools as direct targets of iron chelators: a fluorescence study of chelator action in living cells. Blood, 2005. 106(9): p. 3242-50. 139. Oudit, G.Y., et al., L-type Ca2+ channels provide a major pathway for iron entry into cardiomyocytes in iron-overload cardiomyopathy. Nat Med, 2003. 9(9): p. 1187-94.

266

140. Cabantchik, Z.I., et al., LPI-labile plasma iron in iron overload. Best Pract Res Clin Haematol, 2005. 18(2): p. 277-87. 141. Piga, A., et al., Randomized phase II trial of deferasirox (Exjade, ICL670), a once-daily, orally-administered iron chelator, in comparison to deferoxamine in thalassemia patients with transfusional iron overload. Haematologica, 2006. 91(7): p. 873-80. 142. Hoffbrand, A.V., A. Cohen, and C. Hershko, Role of deferiprone in chelation therapy for transfusional iron overload. Blood, 2003. 102(1): p. 17-24. 143. Angelucci, E., et al., Hepatic iron concentration and total body iron stores in thalassemia major. N Engl J Med, 2000. 343(5): p. 327-31. 144. Glickstein, H., et al., Action of chelators in iron-loaded cardiac cells: Accessibility to intracellular labile iron and functional consequences. Blood, 2006. 108(9): p. 3195-203. 145. Cappellini, M.D., et al., A phase 3 study of deferasirox (ICL670), a once-daily oral iron chelator, in patients with beta-thalassemia. Blood, 2006. 107(9): p. 3455-62. 146. Jehn, M., J.M. Clark, and E. Guallar, Serum ferritin and risk of the metabolic syndrome in U.S. adults. Diabetes Care, 2004. 27(10): p. 2422-8. 147. Rund, D. and E. Rachmilewitz, Beta-thalassemia. N Engl J Med, 2005. 353(11): p. 1135-46. 148. Schreyer, S.A., D.L. Wilson, and R.C. LeBoeuf, C57BL/6 mice fed high fat diets as models for diabetes-accelerated atherosclerosis. Atherosclerosis, 1998. 136(1): p. 17-24. 149. Taher, A., et al., Efficacy and safety of deferasirox, an oral iron chelator, in heavily iron-overloaded patients with beta-thalassaemia: the ESCALATOR study. Eur J Haematol, 2009. 150. Anstee, Q.M. and R.D. Goldin, Mouse models in non-alcoholic fatty liver disease and steatohepatitis research. Int J Exp Pathol, 2006. 87(1): p. 1-16. 151. Bryzgalova, G., et al., Mechanisms of antidiabetogenic and body weight-lowering effects of estrogen in high-fat diet-fed mice. Am J Physiol Endocrinol Metab, 2008. 295(4): p. E904-12. 152. Barzilai, N. and L. Rossetti, Age-related changes in body composition are associated with hepatic insulin resistance in conscious rats. Am J Physiol, 1996. 270(6 Pt 1): p. E930-6. 153. Wolfer, D.P., W.E. Crusio, and H.P. Lipp, Knockout mice: simple solutions to the problems of genetic background and flanking genes. Trends Neurosci, 2002. 25(7): p. 336-40. 154. Sundaram, S.S., P. Zeitler, and K. Nadeau, The metabolic syndrome and nonalcoholic fatty liver disease in children. Curr Opin Pediatr, 2009. 155. Klotz, U., Pharmacokinetics and drug metabolism in the elderly. Drug Metab Rev, 2009. 41(2): p. 67-76. 156. Berdoukas, V., et al., The efficacy of iron chelator regimes in reducing cardiac and hepatic iron in patients with thalassaemia major: a clinical observational study. J Cardiovasc Magn Reson, 2009. 11: p. 20. 157. Nilsson-Ehle, P., A.S. Garfinkel, and M.C. Schotz, Lipolytic enzymes and plasma lipoprotein metabolism. Annu Rev Biochem, 1980. 49: p. 667-93.

267

158. Large, V. and P. Arner, Regulation of lipolysis in humans. Pathophysiological modulation in obesity, diabetes, and hyperlipidaemia. Diabetes Metab, 1998. 24(5): p. 409-18. 159. Summers, S.A., Ceramides in insulin resistance and lipotoxicity. Prog Lipid Res, 2006. 45(1): p. 42-72. 160. Semenza, G.L., Regulation of cancer cell metabolism by hypoxia-inducible factor 1. Semin Cancer Biol, 2009. 19(1): p. 12-6. 161. Zhang, N., et al., The asparaginyl hydroxylase factor inhibiting HIF-1alpha is an essential regulator of metabolism. Cell Metab. 11(5): p. 364-78. 162. van den Hoek, A.M., et al., Leptin deficiency per se dictates body composition and insulin action in ob/ob mice. J Neuroendocrinol, 2008. 20(1): p. 120-7. 163. Daar, S., et al., Reduction in labile plasma iron during treatment with deferasirox, a once-daily oral iron chelator, in heavily iron-overloaded patients with beta- thalassaemia. Eur J Haematol, 2009. 82(6): p. 454-7. 164. Sohn, Y.S., et al., Redistribution of accumulated cell iron: a modality of chelation with therapeutic implications. Blood, 2008. 111(3): p. 1690-9.

268

Research article Hypoxia-inducible factor-1α regulates β cell function in mouse and human islets Kim Cheng,1 Kenneth Ho,1 Rebecca Stokes,1 Christopher Scott,1 Sue Mei Lau,1 Wayne J. Hawthorne,2 Philip J. O’Connell,2 Thomas Loudovaris,3 Thomas W. Kay,3 Rohit N. Kulkarni,4 Terumasa Okada,4 Xiaohui L. Wang,4 Sun Hee Yim,5 Yatrik Shah,5,6 Shane T. Grey,7 Andrew V. Biankin,8,9 James G. Kench,8 D. Ross Laybutt,10 Frank J. Gonzalez,5 C. Ronald Kahn,4 and Jenny E. Gunton1,11,12,13

1Diabetes and Transcription Factors Group, Garvan Institute of Medical Research (GIMR), Sydney, New South Wales, Australia. 2Centre for Transplantation and Renal Research, Westmead Research Institute, University of Sydney at Westmead Hospital, Sydney, New South Wales, Australia. 3St. Vincent’s Institute, Melbourne, Victoria, Australia. 4Joslin Diabetes Center and Harvard Medical School, Boston, Massachusetts, USA. 5Laboratory of Metabolism, National Cancer Institute, Bethesda, Maryland, USA. 6University of Michigan, Department of Internal Medicine, Division of Gastroenterology, Ann Arbor, Michigan, USA. 7Gene Therapy and Autoimmunity Group and 8Cancer Research Program, GIMR, Sydney, New South Wales, Australia. 9Department of Surgery, Bankstown Hospital, Sydney, New South Wales, Australia. 10Diabetes and Obesity Research Program, GIMR, Sydney, New South Wales, Australia. 11Faculty of Medicine, University of Sydney, Sydney, New South Wales, Australia. 12St. Vincent’s Clinical School, University of New South Wales, Sydney, New South Wales, Australia. 13Department of Diabetes and Endocrinology, Westmead Hospital, Sydney, New South Wales, Australia.

Hypoxia-inducible factor-1α (HIF-1α) is a transcription factor that regulates cellular stress responses. While the levels of HIF-1α protein are tightly regulated, recent studies suggest that it can be active under normoxic conditions. We hypothesized that HIF-1α is required for normal β cell function and reserve and that dysregula- tion may contribute to the pathogenesis of type 2 diabetes (T2D). Here we show that HIF-1α protein is present at low levels in mouse and human normoxic β cells and islets. Decreased levels of HIF-1α impaired glucose- stimulated ATP generation and β cell function. C57BL/6 mice with β cell–specific Hif1a disruption (referred to herein as β-Hif1a-null mice) exhibited glucose intolerance, β cell dysfunction, and developed severe glucose intolerance on a high-fat diet. Increasing HIF-1α levels by inhibiting its degradation through iron chelation markedly improved insulin secretion and glucose tolerance in control mice fed a high-fat diet but not in β-Hif1a-null mice. Increasing HIF-1α levels markedly increased expression of ARNT and other genes in human T2D islets and improved their function. Further analysis indicated that HIF-1α was bound to the Arnt promoter in a mouse β cell line, suggesting direct regulation. Taken together, these findings suggest an important role for HIF-1α in β cell reserve and regulation of ARNT expression and demonstrate that HIF-1α is a potential therapeutic target for the β cell dysfunction of T2D.

Introduction Because of its role in the regulation of glycolysis and other bio- The transcription factor HIF-1α is important for a range of func- logical processes in other tissues (24, 25), we hypothesized that (a) tions, including cellular responses to hypoxia and other stressors, HIF-1α might be the important partner for ARNT in β cells, (b) angiogenesis, and fetal development (1–6). It has strong antiapop- that decreasing HIF-1α would impair β cell reserve and thus lead totic effects (7–11) and is implicated in the pathogenesis of cardio- to diabetes under conditions of β cell stress, and (c) that increasing vascular diseases and some cancers (12–20). HIF-1α in a nontoxic way would improve β cell function. HIF-1α is a member of the bHLH-PAS family (reviewed in refs. 2, Consistent with its role in regulating a number of important 18, 21) and functions as an obligate dimer with other family mem- biological processes, HIF-1α protein is tightly regulated (reviewed bers, including aryl hydrocarbon receptor (AhR) nuclear transloca- in refs. 2, 17, 19, 21, 25, 26). In the basal state, it is hydroxylated tor (ARNT). We previously reported that ARNT was decreased in on proline residues and becomes competent to associate with von islets isolated from patients with type 2 diabetes (T2D) and that Hippel-Lindau (VHL) protein, leading to ubiquitination and rapid decreasing ARNT in Min6 cells or disrupting it in mouse β cells proteolysis, giving a half-life of minutes (19, 27, 28). Oxygen, iron, caused changes in gene expression and glucose-stimulated insu- and 2-oxoglutarate are required for hydroxylation (29–32). Thus, lin secretion (GSIS) similar to those seen in islets isolated from hypoxia inhibits degradation, leading to a rapid increase. In addi- humans with T2D (22). Recently, we reported a loss of ARNT tion, HIF-1α protein can be increased by genetic inactivation of expression in the livers of people with T2D, affecting dysregula- VHL or the hydroxylases, treatment with heavy metals such as tion of gluconeogenesis (23). Though the specific ARNT partner cobalt chloride, or iron chelation with deferoxamine (DFO) or which is important for its actions in β cells (or liver) is not known, deferasirox (DFS) (20, 29). An additional layer of regulation is candidates include AhR, HIF-1α, HIF-2α, HIF-3α, and circadian added by asparaginyl-hydroxylation, which inhibits association rhythm molecules, e.g., BMAL. with transcriptional cofactors, including p300 (21). Until recently, it was thought that HIF-1α did not function under Conflict of interest: Rohit N. Kulkarni declares that his laboratory received research normoxic conditions. However, the presence of HIF-1α protein in funding from Novartis for an unrelated project. brain, kidney, liver, embryonic stem cells, trophoblastic cells, and Citation for this article: J Clin Invest doi:10.1172/JCI35846. others (5, 6, 33) is now recognized. It is stabilized by inflammation,

The Journal of Clinical Investigation http://www.jci.org research article

Figure 1 In contrast, in T2D islets, HIF-1α staining was not higher than HIF-1α is present in normoxic β cells, associates with ARNT, and is that in the acinar pancreas (Figure 1C). decreased in T2D. (A) HIF1A mRNA was decreased in islets of people HIF-1α associated with ARNT. To determine whether the HIF-1α in with T2D (n = 6) compared with people with normal glucose tolerance β cells and islets was associated with ARNT and therefore poten- (n = 12). ***P < 0.001. (B) HIF-1α protein was present in β cells in tially transcriptionally active in the basal (nonhypoxic) state, we floxed control mice (horizontal arrows) but was decreased in β-Hif1a- null mice. In both genotypes, HIF-1α was present in blood vessels used ARNT affinity purification and tandem MALDI-TOF mass (vertical arrows). Scale bar: 20 μm. (C) HIF-1α protein was higher than spectrometry. The Coomassie-stained gel showed a band at the background in people with normal glucose tolerance (top panels) but appropriate size for HIF-1α in the untreated nuclear fraction and was decreased in T2D pancreata (bottom panels). Scale bar: 50 μm. in the DFO-treated positive control (Figure 1D, black arrow). Mass (D) HIF-1α protein (arrow) was associated with ARNT by affinity purifi- spectrometry revealed 4 peptide sequences derived from HIF-1α. cation. Cyto, cytoplasm; Nuc, nucleus; Prot, protein. (E) HIF-1α protein HIF-2α was also associated with ARNT but only in DFO-treated associated with ARNT in the basal state in Min6 cells, and nuclear cells. AhR protein was not detected. HIF-1α increased with DFO. (F) ARNT protein associated with HIF-1α by coimmunoprecipitation. (G) HIF-1α protein was increased by DFO To confirm that ARNT was bound to HIF-1α, we performed coim- treatment of isolated mouse islets and was decreased in islets from a munoprecipitation studies in Min6 cells. Immunoprecipitation β-Hif1a-null mouse (Cre+). with ARNT antibodies revealed low levels of associated HIF-1α in the cytoplasm and nucleus, under normoxic conditions (Figure 1E). As expected, DFO treatment increased nuclear HIF-1α. The con- TGF, PDGF, EGF, and IL-1β (20, 34, 35) and by increased levels verse was also true; immunoprecipitation with HIF-1α antibodies of ROS (36–38). Of potential relevance to β cells, insulin increases purified ARNT, under normoxic conditions (Figure 1F, left lane). HIF-1α activity in liver, muscle, breast carcinoma, prostate carcino- AhR antibodies were able to “pull-down” ARNT from whole-cell ma, and retinal epithelial-derived cells (39–42). PI3K-Akt pathway lysates (Figure 1F, right lane). However, without exogenous ligand, activation is required for the insulin-induced increase (43). AhR antibodies did not purify detectable amounts of ARNT from The role of HIF-1sα in islets i not fully understood. Pancre- nuclear extracts, suggesting that it was not functionally active (data atic islets are normally exposed to relatively low oxygen tension not shown). As shown in Figure 1G, ARNT immunoprecipitation (20–37 mmHg) (44, 45) and to locally secreted insulin. These fac- and HIF-1α immunoblotting was performed in whole-cell lysates tors suggest a possible role for HIF-1α in islets and the possibility from primary mouse islets. Lanes 1 and 2 are islets from mice treated for decreased HIF-1α in the setting of insulin resistance. with DFO and then cultured in DFO. Lanes 3–8 were islets cultured This study found that targeted disruption of HIF-1α in β cells in normal media. The Cre+ lane shows islets from a β-Hif1a-null of C57BL/6 mice (referred to herein as β-Hif1a-null mice) led to mouse in which there was decreased HIF-1α. glucose intolerance with impaired ATP generation and GSIS in β cell–specific Hif1α-null mice have impaired β cell function. Using the isolated islets. Conversely, increasing HIF-1α using iron chelation Cre-lox system, with Cre under control of the rat insulin promoter with DFO or DFS caused significant changes in gene expression, (RIP-Cre), and mice with a floxed Hif1a gene (floxed controls) which differed from severe hypoxia or VHL deletion (46–48). DFS (49), we generated β-Hif1a-null mice. HIF-1α immunostaining is significantly improved glucose tolerance in mice receiving a high- shown in Figure 1B. fat diet (HFD) but had no effect in β-Hif1a-null mice, demonstrat- Importantly, in our colony, RIP-Cre mice have normal glucose tol- ing that β cell HIF-1α was required for its effect. Importantly, erance (Figure 2A). β-Hif1a-null mice were fertile and did not differ DFO treatment of T2D islets normalized expression of ARNT in size or weight (data not shown). Fasting glucose did not differ and downstream genes and improved GSIS. HIF-1α bound to the among groups (Figure 2A); however, levels after glucose loading were ARNT promoter, as revealed by ChIP, and increasing HIF-1α levels significantly higher in β-Hif1a-null mice than in floxed control mice increased ARNT expression. Taken together, these findings sug- or in RIP-Cre mice. Disruption of HIF-1α in β cells did not alter gest that decreased HIF-1α levels impair β cell reserve and that fasting insulin. However, as shown in Figure 2B, β-Hif1a-null mice iron chelation, which increases HIF-1α activity in β cells, may be a had significantly impaired first-phase GSIS. Consistent with these therapeutic strategy for the treatment of human T2D. in vivo effects, islets isolated from β-Hif1a-null mice had impaired GSIS (>80% reduction at 3.3 and 11 mmol/l glucose; Figure 2C). The Results difference was not statistically significant at 22 mmol/l. HIF-1α was present at low levels in islets and was decreased in humans with Gene expression was assessed in isolated islets using real-time T2D. HIF1α levels were assessed using real-time PCR in isolated PCR. In β-Hif1a-null islets, there was a more than 40% decrease human T2D and control islets and using immunohistochemistry in Glut2, glucokinase (Gck), glucose-6-phosphoisomerase (G6pi), of pancreatic sections collected during partial pancreatectomy. phosphofructokinase (Pfk), hepatocyte nuclear factor 4α (Hnf4a), HIF1A mRNA was decreased by 90% in T2D islets (P < 0.0001; Fig- and others (Figure 2D). As expected, control islets had higher ATP ure 1A). Immunohistochemistry revealed HIF-1α protein in some content after exposure to high glucose (60% increase; Figure 2E). normal murine β cells (Figure 1B, horizontal arrows), and stain- Islets from β-Hif1a-null mice had significantly lower basal ATP lev- ing was present in blood vessels (Figure 1B, vertical arrows). In els (60% decrease) and the increase in ATP levels with 25 mmol/l β-Hif1a-null islets, staining was present in vascular cells (Figure 1B, glucose was severely blunted. This was despite similar islet insulin vertical arrows) but not β cells, demonstrating antibody specificity. content (Figure 2F) and nonsignificant, higher β cell content in In human partial pancreatectomy tissue, there was mild, diffuse β-Hif1a-null mice (Figure 2G). No differences were observed for HIF-1α staining, consistent with the operative procedure in which CD31 staining, indicating similar islet vascularity (data not shown). vessels are ligated prior to tissue removal, thus inducing hypox- Hif1a knockdown in Min6 cells impaired β cell function. To confirm ia. Subjects with normal glucose tolerance showed more intense the β-Hif1a-null mice results in a βe cell line, w used RNAi in Min6 HIF-1α staining in islets than in the acinar pancreas (Figure 1C). cells. RNAi achieved approximately 70% knockdown of Hif1a

The Journal of Clinical Investigation http://www.jci.org research article

Figure 2 β cell deletion of Hif1a in mice causes glucose intolerance, impaired gene expression, ATP generation, and insulin secretion. (A) β-Hif1a-null mice were glucose intolerant compared with either floxed controls or RIP-Cre alone mice. n = 15, 15, and 12, respectively. (B) GSIS was decreased in β-Hif1a-null mice. (C) GSIS was decreased in isolated β-Hif1a-null islets. (D) Expression of several genes was decreased in β-Hif1a-null islets. (E) ATP concentrations were significantly decreased in β-Hif1a-null islets at both basal and high glucose levels. (F) Insulin content did not differ between floxed control and β-Hif1a-null islets. (G) β cell mass did not differ between groups. *P < 0.05, **P < 0.01, and ***P < 0.001. mRNA (Figure 3A). This was accompanied by markedly impaired rate cohort of mice (Figure 4A), we performed glucose tolerance GSIS (Figure 3B). There was a milder (approximately 25%) impair- testing (GTT) and replicated the finding of impaired glucose toler- ment in KCl-stimulated insulin release (Figure 3B, right). In sepa- ance, with preserved fasting levels (P = 0.007, ANOVA for repeated rate experiments examining combined RNAi treatments, slightly measures). For β-Hif1a-null mice, weight was 24.7 ± 0.6 g, and for more severe impairment occurred with Arnt RNAi and with Hif1a controls, weight was 25.2 ± 0.6 g. plus Hif2a plus AhR knockdown, suggesting small additional roles To examine β cell compensation in the setting of insulin resis- for Hif-2α and AhR (Figure 3C). KCl-stimulated insulin secretion tance, we placed mice on a HFD (45% of calories from fat) for was again impaired by approximately 25% (Figure 3C, right), sug- 3 weeks. Weight increased to 29.4 ± 0.9 g for β-Hif1a-null mice and gesting a partially glucose-specific effect. to 29.0 ± 0.8 g for controls. GTTs deteriorated in floxed controls Decreased expression of glucose transporter and glycolytic (Figure 4B, dotted line) but more severely in β-Hif1a-null mice mRNAs was found, including Gck, Glut2, G6pi, Aldolase (AldoB), (Figure 4C, dotted line). AUCs are shown in Figure 4D (P = 0.047 and Pfk (40%–60%; Figure 3, D and E). These changes were similar for floxed control HFD-fed versus β-Hif1a-null HFD-fed mice). to those for β-Hif1a-null islets (Figure 2E). As shown in Figure 3F, Following 3 weeks of HFD, all mice were changed to HFD admixed Hif1a knockdown impaired ATP generation basally (25% decrease) with DFS (HFD plus DFS) to increase HIF-1α. After 3 weeks, weight and following glucose stimulation (only 90% of basal control). was 29.0 ± 0.8 g in β-Hif1a-null mice versus 29.0 ± 0.8 g in controls. Increasing HIF-1α improved glucose tolerance in HFD-fed C57BL/6 As shown in Figure 4C, despite continuing HFD, floxed controls mice. As noted above, β-Hif1a-null mice exhibited mild glucose had highly significantly improved glucose tolerance. In contrast, intolerance, with preserved fasting glucose (Figure 2A). In a sepa- there was no improvement in β-Hif1a-null mice, demonstrating that

The Journal of Clinical Investigation http://www.jci.org research article

Figure 3 Decreasing Hif1a by RNAi-impaired β cell function, gene expression, and ATP generation. (A) RNAi decreased Hif1a mRNA. (B) Hif1a RNAi decreased GSIS in Min6 cells and caused a small decrease in KCl-stimulated insulin release. (C) Combination RNAi treatment caused slightly more severe impairment in insulin release. (D) Hif1a RNAi decreased expression of genes from the MODY family and (E) glucose-uptake and glycolysis genes. (F) Hif1a RNAi decreased basal and glucose-stimulated ATP concentrations. *P < 0.05, **P < 0.01, and ***P < 0.001.

Apoptosis rates were less than 1% in both floxed control and β-Hif1a-null groups with this short- term treatment (data not shown). We confirmed the effects of DFS on HFD-induced glucose intolerance in a separate cohort of normal (not genetically modified) C57BL/6 mice fed HFD or HFD plus DFS for 26 weeks. Glucose tolerance was significantly better in mice receiving HFD plus DFS (Figure 4F, solid line) compared with mice receiving HFD alone (Figure 4F, dotted line). In this 26 week experiment, weight was lower in the HFD plus DFS– fed group (HFD-fed, 38.8 ± 4.2 g versus HFD plus DFS–fed, 33.0 ± 3.2 g; P = 0.037), which may have contributed to the difference in GTTs. However, weight and fasting glucose were not significantly correlated (P > 0.13; Figure 4G) and neither were weight and AUC of GTTs (data not shown). Mul- tivariate regression using weight and DFS as uni- variate predictors showed that only DFS independently predicted fasting glucose (P = 0.010 for DFS, P = 0.811 for weight) and AUC of GTTs (P = 0.039 for DFS, P = 0.433 for weight). Mice were not ane- mic (hemoglobin, 120 ± 7 mg/dl in DFS mice versus 124 ± 3 mg/dl in controls; P > 0.6). Insulin tolerance testing (Supplemental Figure 1; supplemental material available online with this article; doi:10.1172/ JCI35846DS1) showed that the HFD plus DFS–fed mice had similar percentages of decreases in blood glucose after insulin administration. Increasing HIF-1α improved glucose tolerance in HFD- fed Balb/c mice. Because the above experiments were all performed in C57BL/6 mice, we sought to confirm the effects in another mouse strain. Balb/c mice were weighed (20.8 ± 0.4 g), underwent GTT (Figure 4H, dashed line), and were placed on HFD. After 2 weeks, mice weighed 24.5 ± 0.5 g. Repeat GTT showed marked deterioration (dotted line). The diet was then changed to HFD plus DFS. After β cell HIF-1α was required for DFS to improve glucose tolerance. 2 weeks, mice weighed 24.7 ± 0.4 g. GTTs were repeated and showed AUCs are shown in Figure 4D. Interestingly, β-Hif1a-null mice had significant improvement (2.1 mmol/l in peak glucose), confirming 69% greater β cell mass than controls at the end of the study (Fig- an effect of DFS in another mouse line. ure 4E), despite worse glucose tolerance, suggesting attempted and Increasing HIF-1α in human islets improved gene expression. Given unsuccessful β cell compensation. the deleterious effects of β cell HIF-1α disruption and the benefi- There was some 4-hydroxynonenal staining in HFD-fed mice, cial effects of DFS in vivo, we examined the effects of increasing consistent with increased ROS, but it was without obvious dif- HIF-1α with DFO upon cultured human islets and compared this ferences between genotypes (data not shown). Similarly, β cell with the effects of hypoxia. Islets were cultured with DFO at the apoptosis did not differ, as assessed by cleaved caspase-3 staining. doses shown (μt mol/l) or a normoxia (21% oxygen) or hypoxia (1%

The Journal of Clinical Investigation http://www.jci.org research article

Figure 4 Lack of β cell HIF-1α leads to severe deterioration in glucose tolerance on a HFD and increasing HIF-1α levels with DFS improves glucose tolerance on a HFD. (A) β-Hif1a-null mice (n = 10) had worse glucose tolerance than floxed control littermates (n = 17). (B) On HFD, glucose tolerance deteriorated in floxed controls and improved following DFS. (C) On HFD, glucose tolerance deteriorated markedly in β-Hif1a-null mice, and there was no improvement with DFS. (D) Glucose tolerance AUC for mice at completion of the HFD and HFD plus DFS stages. (E) β cell mass was increased in β-Hif1a-null mice at study completion. (F) Glucose tolerance was significantly better in C57BL/6 mice receiving HFD plus DFS versus mice receiving HFD alone (n = 10 per group). (G) Weight and fasting glucose were not significantly correlated in the mice. Rectangles indicate mice receiving HFD plus DFS, and triangles indicate mice receiving HFD alone. (H) Balb/c mice had deterioration in glucose tolerance on HFD (dotted line) compared with chow (dashed line). Their glucose tolerance improved significantly on HFD plus DFS (n = 12). *P < 0.05 and **P < 0.01.

oxygen). The human therapeutic dose is approximately 125 μmol/l. Hif1a regulated Arnt. Having observed a small increase in ARNT DFO caused dose-dependent increases in HNF4A, GLUT1, GLUT2, expression with DFO treatment in normal human islets, we exam- AKT2, and IRS2 mRNAs (Figure 5A). In contrast, hypoxia increased ined the effects of manipulating HIF-1α upon Arnt expression. GLUT1 but not GLUT2, AKT2, or the other genes shown (Figure 5B). RNAi-mediated knockdown of Hif1a in Min6 cells caused a more Interestingly, there was a small but significant increase in ARNT than 80% decrease in Arnt expression (Figure 6A; P < 0.01), and mRNA with DFO (Figure 5C), which was not seen with hypoxia islets from β-Hif1a-null mice had a 50% decrease in Arnt expres- (Figure 5D). There was a significant increase in GSIS at moderate sion (Figure 6B; P < 0.01), which was similar to the reduction in hyperglycemia in DFO-cultured human islets compared with con- Hif1a itself. This was paralleled by a decrease in ARNT protein in trol-cultured islets from the same donors (Figure 5E). The approxi- immunostained pancreas of β-Hif1a-null versus floxed control mately 60% increase at high glucose was usual for recently isolated mice (Figure 6C, compare bottom and top panels). Together these human islets. As expected, GSIS was not improved in hypoxic experiments show that decreasing HIF-1 α in islets and β cells led human islets and declined with longer exposure (Figure 5F). to decreased ARNT.

The Journal of Clinical Investigation http://www.jci.org research article

Figure 5 DFO improves gene expression and insulin secretion from human islets. (A) DFO increased expression of several genes in isolated human islets. (B) Hypoxic culture increased GLUT1 expression but decreased that of HNF4A and AKT2. (C) DFO increased expression of ARNT and HIF1A. (D) Hypoxia did not alter ARNT or HIF1A expression. (E) DFO increased insulin secretion in isolated human islets. (F) GSIS was not improved in hypoxic islets and declined with longer exposure. *P < 0.05, **P < 0.01, and ***P < 0.001.

Differential effects with different methods of increasing HIF-1α. We examined the effects of hypoxia, Vhl knockdown, and DFO in isolated islets and Min6 cells with hypoxia in order to examine the mechanisms behind the different effects of DFO/DFS compared with hypoxia and to investigate the different results compared with those reported with Vhl knockouts (46–48). As shown in Figure 5F and Figure 7A, 1% oxygen did not promote GSIS. In human islets (Figure 7B) and Min6 cells (Figure 7C), 1% oxygen also did not increase expression of GLUT2 or AKT2. As expected, 1% oxygen increased Glut1 expression in Min6 cells. Interestingly, changes in gene expression differed between 5% oxygen and 1% oxygen. In cells exposed to 5% oxygen, Glut1, Glut2, and Akt2 increased (Figure 7D). RNAi-induced Vhl knockdown in Min6 cells decreased Vhl mRNA by 33%. This led to a small but significant increase in Glut2 (Figure 7E) and a nonsignificant increase in GSIS (Figure 7F). By dou- bling RNAi concentrations, approximately 5 5% Vhl knockdown was achieved. This led to decreased Hnf4a and Akt2 expression, and the increase in Glut2 was lost (Figure 7G). This was accompanied by a nonsignificant impairment in GSIS (Figure 7H). Increasing HIF-1 α levels by transfecting a pro- Increasing HIF-1α in human T2D islets increased ARNT, HNF4A, line-to-alanine mutant caused significant impairment in GSIS (Figure and G6PI expression. Human islets isolated from a new cohort 7I). This was associated with the expected increases in Hif1a (>29-fold) of 3 people with T2D had significantly decreased HIF1A and and Glut1 (>3-fold). However, there was also a significant decrease in ARNT expression (Figure 1A and Figure 6D). We examined the expression of Gck (Figure 7J). Interestingly, there was significantly effect of culturing T2D islets with DFO. In this new group of decreased total insulin content in the mutant-HIF-1α–transfected T2D donors, we also found a more than 80% decrease in ARNT cells (approximately 50% of vector transfected; Figure 7K). expression (Figure 6D). HNF4A and G6PI were also decreased, Removing iron by chelation improved HIF-1α activity and β cell consistent with our previous report (Figure 6, E and F). Culture function. The effect of adding iron in the form of ferric citrate was of islets from the same T2D donors with DFO increased ARNT studied. At high doses, there was obvious cell death. At lower doses, to near-normal levels (Figure 6D; P < 0.01). DFO increased there were significant decreases in Hif1a, Hnf4a, Glut1, and Glut2 HNF4A and G6PI expression to near-normal levels (Figure 6, E expression (Figure 7L). This identifies iron as a potential regula- and F; P < 0.01 for both genes). Similar results were also seen for tor of HIF1A in β cells. There was also significantly impaired GSIS, AKT2 (data not shown). with only a nonsignificant 13% increase, following high glucose in To determine whether HIF-1α bound directly to the Arnt pro- iron-treated cells (P = 0.017 versus control high glucose). moter, we performed ChIP assays. In Min6 cells, HIF-1α antibodies DFO treatment at 16-times the therapeutic dose (2,000 μmol/l) pulled down the proximal Arnt promoter (Figure 6G). The ampli- also decreased GLUT2 by 34% (Figure 8). Changes in GLUT2 expres- fied sequence (primers in Methods) contains a potential hypoxia- sion with various treatments are compiled in Figure 8. Glut2 gener- response element, GCGTG. ally corresponded to GSIS, with lower Glut2 expression and lower

The Journal of Clinical Investigation http://www.jci.org research article

Figure 6 In other cell types, HIF-1α regulates PDK1 (10, 11), COX4.1, HIF-1α regulates expression of ARNT and downstream genes. (A) COX4.2, and LON protease changes associated with increased ATP Hif1a RNAi in Min6 cells decreased Arnt expression. (B) β-Hif1a-null (58). Consistent with these studies, we observed that decreasing mice had decreased Arnt mRNA compared with floxed controls. (C) HIF-1α decreased ATP. ATP generation is required in β cells for ARNT protein was decreased in β-Hif1a-null islets versus floxed con- sensing of glucose, which in turn triggers insulin release. Further- trols. Scale bar: 50 μm. (D) ARNT mRNA was decreased in islets from people with T2D. DFO increased ARNT expression to levels that did more, impaired ATP generation in our models was associated with not differ significantly from normal. (E) HNF4A mRNA was decreased impaired insulin release, consistent with HIF-1α being a regulator in islets from people with T2D and was increased by DFO. (F) G6PI of β cell energy homeostasis and insulin release. Thus, decreased expression was decreased in islets from people with T2D and was HIF-1α impaired glucose-stimulated ATP generation, providing increased by DFO. (G) HIF-1α associated with the proximal Arnt pro- the mechanism by which decreased availability of a transcription moter by ChIP. **P < 0.01. factor can cause β cell dysfunction. Recently, 3 groups reported clear, adverse effects of homozygous deletion of VHL upon β cell function (46–48). In these models, insulin secretion in β-Hif1a-null islets after HIF-1α knockdown there was a massive increase in HIF-1α protein. Disruption of VHL with ferric citrate, reported for genetic VHL deletion (46, 47), and was accompanied by adverse gene expression changes, increased with toxic doses of DFO. lactate, and severely impaired GSIS. These findings are interesting but were unexpected as heterozygous whole-body VHL knockout Discussion mice appeared grossly normal (59), with 5%–25% of mice develop- People with T2D characteristically have pronounced impairment ing abnormal vascular lesions in later life on some genetic back- of first-phase insulin secretion (50–52). This defect is intrinsic, as grounds (60). People with VHL syndrome may develop endocrine it persists in isolated islets (53, 54) and is relatively glucose spe- pancreatic tumors, pancreatic cysts, and occasional insulinomas cific (54, 55) in the earlier stages of the disease. In previous stud- (16, 61–63). However, less than 3% of patients are reported to have ies, we demonstrated that T2D islets had decreased ARNT expres- abnormal glucose tolerance, despite frequently requiring steroids sion and, using gene inactivation approaches, showed that this and/or pancreatic surgery (16, 61, 62, 64). This suggests that a could contribute to altered β cell function. In the present study, heterozygous germline mutation, as occurs in VHL patients, may we show that T2D islets had decreased HIF-1α, which we show is not cause an increased risk of diabetes. Mutations in subunits of also important for islet function. Deletion of HIF-1α in C57BL/6 the succinate dehydrogenase complex (65, 66) and the HIF-1α pro- mice resulted in impaired ATP generation and impaired glucose lyl hydroxylases are also associated with increased HIF-1α protein, tolerance, accompanied by altered gene expression. Similar results but there are no reported alterations in diabetes incidence. were found in Min6 cells using RNAi. The β cell defect was rela- Loss of VHL was associated with decreased Glut2 mRNA (46, 47). tively glucose specific, with only approximately 25% impairment We also found decreased Glut2 with DFO at 16-times the therapeu- in KCl-stimulated insulin release. tic dose, exposure to 1% oxygen, high-dose VHL RNAi, and supple- Although HIF-1α protein is tightly regulated, several methods of mental iron. In contrast, therapeutic levels of DFO and 5% oxygen increasing it exist. These include hypoxia, decreasing VHL protein, treatment both caused different changes in gene expression, and mutation or decreased expression of the prolyl hydroxylases, treat- in particular, Glut2 was increased. The different changes in gene ment with heavy metals such as cobalt chloride, and iron chelation. expression with different methods of increasing HIF-1α were in Severe hypoxia and cobalt chloride are toxic. Genetic modifica- accordance with the changes in β cell function (Figure 8). tion is not usually a therapeutic option for humans, although the Thus, there appears to be a dose-response curve for HIF-1α (Fig- future possibility exists with antisense RNA strategies. Thus, we ure 8). Deletion is deleterious in C57BL/6 mice and Min6 cells. studied the effects of iron chelation with DFS or DFO. Mild increases are beneficial for β cell function and glucose toler- Treating mice made diabetic by high-fat feeding with DFS ance but very high levels, such as those achieved with homozygous improved glucose tolerance. DFS was also effective in C57BL/6 VHL deletion, severe hypoxia, or overexpression of a degradation and Balb/c wild-type mice but was completely ineffective in mice resistant mutant, are clearly deleterious for β cell function. lacking β cell HIF-1α, demonstrating that HIF-1α is required for Hydroxylation and proteolysis of HIF-1α requires iron, which the benefit. Our HFD had 45% of calories from fat, compared with is chelated by DFO and DFS. Iron overload due to transfusion 12% in normal chow. While this is high, the average fat intake in dependency or hemochromatosis can cause β cell dysfunction and the American diet is more than 30%. The top 20% of the popula- increases diabetes incidence (67, 68). It is perhaps less widely recog- tion consume 46% of calories from fat (56, 57). DFS was effective nized that in the absence of transfusion-dependent iron overload despite continuing the HFD, suggesting that it may be effective in or hemochromatosis, increases in serum ferritin or transferrin people with T2D, in whom high fat-intake is common. saturation are associated with increased risk of diabetes and the Surprisingly, DFO treatment normalized ARNT and other genes metabolic syndrome (69–74). High dietary iron intake is also asso- in T2D islets. HIF-1α is predominantly regulated at the protein ciated with diabetes (71, 75). Conversely, venesection and blood level, and DFO treatment was apparently sufficient to normalize donation can improve β cell function in people with diabetes (70, HIF-1α function, as assessed by expression of downstream genes. 76). Regular blood donation has been reported to protect against The magnitude of effect on ARNT, HNF4A, and G6PI was large diabetes, as does a vegetarian diet (67, 70, 76). Disruption of the in T2D islets (>10-fold), in which basal HIF-1α was low. This is HIF-1α–partner ARNT in endothelial cells leads to pronounced the first time that a strong regulator of ARNT expression has been iron accumulation in the liver (77), suggesting the intriguing identified. In contrast, the change in ARNT expression in normal potential for a vicious cycle of decreased HIF-1α, decreased ARNT, islets, in which HIF-1α was not low at baseline was modest. Hif1a increased iron accumulation, and decreased HIF-1α. Based on our itself was decreased by treatment with iron. data and the absence of a DFS effect in mice lacking β cell HIF-1α,

The Journal of Clinical Investigation http://www.jci.org research article

Figure 7 The differing effects on gene expression and insulin secretion with increasing HIF-1α levels using hypoxia or VHL RNAi. (A) Insulin secretion was not improved in human islets, mouse islets, or Min6 cells cultured under hypoxic conditions. (B) One percent oxygen did not increase GLUT2 expression in human islets. (C) One percent oxygen did not increase Glut2 expression in Min6 cells. (D) Moderate hypoxia (5% oxygen) increased Glut2 expression. (E) Modest Vhl knockdown (35%) increased expression of Glut2 and was associated with a nonsignificant increase in insulin release (F). (G) High-dose Vhl knockdown achieved a 55% decrease in Vhl and did not increase Glut2 expression. (H) High- dose Vhl RNAi lowered insulin secretion non- significantly. (I) Transfection with proline-to- alanine mutant HIF-1α significantly impaired insulin secretion. (J) Hif1a expression was increased more than 29-fold and was accompanied by increased Glut1 expression and decreased Gck. (K) Total insulin content was decreased in the proline mutant HIF–overex- pressing cells. (L) Ferric citrate treatment significantly decreased Hif1a expression and was accompanied by decreased expression of Hnf4a, Akt2, Glut1, and Glut2. *P < 0.05, **P < 0.01, and ***P < 0.001.

we postulate that a decrease in HIF-1α may be a mechanism con- Methods tributing to the increased risk of diabetes with increased iron. Human islet studies were approved by St. Vincent’s Clinical School Human In summary, these studies demonstrate that β cell HIF-1α Research Ethics Committee. All participants gave informed consent. Animal is important for β cell reserve, and increasing HIF-1α by iron studies were approved by the Garvan Institute Animal Ethics Committee. chelation markedly improved glucose tolerance on a HFD. Increas- Human islets were purified using the modified Ricordi method as previously ing HIF-1α normalized gene expression in T2D islets. Therefore, described (22). RNA was isolated using Qiagen RNeasy kits. Gene expres- we propose that increasing HIF-1α by iron chelation may be a valid sion was measured by real-time PCR using the Invitrogen RT-for-PCR kit. therapeutic strategy for the treatment of human T2D. The second step was performed in an ABI Prism 7700 Sequence Detection

10 The Journal of Clinical Investigation http://www.jci.org research article

For mass spectrometry, gel slices were digested with 5 ng/ml sequencing grade-modified trypsin (Promega) in 25 mM ammonium bicarbonate containing 0.01% n-octylglucoside for 18 hours at 37°C. Peptides were eluted from the gel slices with 80% acetonitrile and 1% formic acid. Tryptic digests were separated by capillary HPLC (C18, 75 mM i.d.; Picofrit column, New Objective), using a flow rate of 100 nl/ min over a 3-hour reverse phase gradient, and analyzed using a LTQ linear Ion Trap LC/MSn system (Thermo Electron). Resultant MS/MS spectra were searched against the NCBI Refseq database (http://www. ncbi.nlm.nih.gov/refseq/) (TurboSequest, BioWorks 3.1, Thermo Electron), with cross-correlation scores of greater than 1.5, 2.0, and 2.5 for charge states U′, u′, and ℧, respectively, more than 30% frag- ment ions, and a ranking of primary score (RsP) value of <3. Proteins were identified with more than 2 unique peptide matches. Coimmunoprecipitation studies were performed using 2 μg of the indicated antibody and protein A/G beads and by incubating overnight with the indicated cell lysate, followed by washing, elution with reducing sample buffer, and separation by 10% SDS-PAGE. Proteins were detected with the indicated antibody, followed by the appropri- Figure 8 ate HRP-conjugated secondary antibody, and detection by enhanced Modest increases in HIF-1α improve insulin secretion. Changes in HIF-1α, which were associated with decreased GLUT2, were associated with impaired chemiluminescence. For each antibody, species-matched nonimmune insulin secretion. *P < 0.05, **P < 0.01, and ***P < 0.001. immunoglobulin and antibody-alone lanes were tested as negative controls to confirm antibody specificity. Islet isolation from mice. Islets were isolated from mice as previously described (22). All mice except for the Balb/c mice were inbred C57BL/6 System (Applied Biosciences) with LightCycler-RNA Master SYBR Green I for at least 12 generations. (Roche). Primers are in Supplemental Table 1. Every plate included a con- DFO and hypoxia treatment. DFO treatment was at 125 μM for 4 hours, trol gene (TATA-box binding protein/TBP) for every subject. unless otherwise specified. Hypoxia treatments were for 2 hours, unless Immunohistochemistry and antibodies. Slides were cut from paraffin-embed- otherwise indicated. Hypoxia was achieved with a hypoxic chamber and an ded pancreata. Antibodies were purchased from Novus Biologicals (HIF-1α, oxygen sensor to confirm levels. HIF2α/EPAS1), Orbigen (AhR), Cell Signaling Technology (insulin), or BD Alanine HIF-1α mutant. Proline residues 402 and 577, in the murine HIF-1α Biosciences (ARNT). Primary antibodies were applied overnight at 4°C. Sec- cDNA, were mutated by site-directed mutagenesis to alanine and the con- ondary antibodies were Cy2, Cy3, or Cy5 conjugated and applied for 1-hour struct was cloned into the pcDNA3 vector and sequenced. The construct at room temperature. Slides were viewed on a Zeiss inverted microscope and and the vector were transfected into Min6 cells using Lipofectamine 2000, images were taken with AxioVision software. For each figure, the images were according to the manufacturer’s instructions, and selected using geneticin taken in the same session with identical camera settings. For each antibody, for 1 week. Total insulin content was measured and corrected for total pro- species-matched nonimmune immunoglobulin and secondary antibody tein, which was measured by DC Bradford assay. alone were tested as negative controls. HIF-1α antibody specificity was addi- In vivo testing. GTTs, GSIS, in vitro GSIS, β cell mass, and mRNA expres- tionally supported by the lack of β cell HIF-1α staining in knockout mice. sion in islets were assessed as previously described (22). AUC for the GTTs Generation of β-Hif1a-null mice. β-Hif1a-null mice were generated using the was calculated using the trapezoidal method. Insulin tolerance tests were Cre-lox system. Mice, with floxed HIF-1α (49), were bred with mice express- performed by injecting insulin at 0.5 U/kg and measuring glucose at ing Cre-recombinase, under control of the rat insulin promoter (RIP-Cre the times shown. mice). In our colony, RIP-Cre mice did not have abnormal glucose tolerance Measurement of intracellular ATP concentrations. ATP concentrations were (Figure 2A). Recombination efficiency was estimated by semiquantitative measured in islets and in Min6 cells following basal culture in 1 mM glu- PCR, using the genotyping primers (49) at 60%–80%. Anti-mouse and anti- cose for 1 hour, followed by washing and exposure to 1 mM or 25 mM glu- rabbit secondary antibodies were from Santa Cruz Biotechnology Inc. cose for 15 minutes. Cells were then placed on ice, washed twice in ice-cold ARNT affinity purification was done by binding ARNT antibody (12 μg) PBS, and lysed. ATP was measured using the Roche Bioluminescence kit. to 1 ml of packed protein A/G beads in 5 ml columns. Unbound anti- Results were corrected for total protein. body was removed by washing with 20 ml of PBST. Min6 cells (a gift from RNAi treatment of Min6 cells and insulin release. Using Min6 cells, HIF-1α, J. Miyazaki, Physiological Chemistry, Osaka University, Osaka, Japan; ref. HIF-2α, AhR, and ARNT were decreased by treatment with smartpool RNAi 78) were grown to 80%–90% confluence, washed twice in PBS, and placed (Dharmacon) and transfected using Lipofectamine 2000 (Invitrogen), in serum-free high glucose DMEM for 4 hours with or without DFO. Cells according to the respective manufacturers’ protocols. Scrambled-sequence were scraped into LID lysis buffer with protease inhibitors as previously RNAi was used as a control in all experiments. Total RNAi concentrations described (79). Cytoplasmic extracts were collected after centrifugation. were the same for the combination experiments (e.g., 3-times control ver- The nuclear-containing pellet was disrupted by sonication. Extracts were sus HIF-1α plus AhR plus HIF-2α versus 2-times control plus HIF-1α). applied to columns, and the flow-through was reapplied twice to obtain Cy3-labelled RNAi and FACS sorting were used to determine transfection maximal binding. After this, the columns were washed twice with 20 ml efficiency, which was more than 75% (data not shown). Experiments were of LID buffer, followed by 2 washes with 20 ml of PBST. Bound proteins performed 48 hours after transfection. GSIS was assessed in triplicate wells were eluted with reducing sample buffer and were size separated by 10% in 3 separate experiments and corrected for total insulin content. In sepa- SDS-PAGE, followed by protein staining with Coomassie blue (Figure 1D). rate experiments, RNA was isolated for real-time PCR.

The Journal of Clinical Investigation http://www.jci.org 11 research article

HFD studies. Male floxed control (n = 17) or β-Hif1a-null mice (n = 10) the L’Oreal Australian For Women in Science fellowship. W.J. Haw- had GTTs as described above. They were then placed on a HFD, based on thorne, P.J. O’Connell, T. Loudovaris, and T.W. Kay were funded by Rodent Diet no. D12451 from Research Diets Incorporated, which con- JDRF and NHMRC. C.R. Kahn was supported by the Mary K. Iacocca tained 45% of calories from fat (lard). DFS was thoroughly mixed into the Professorship and NIH grants RO1 DK33201 and DK60837-02. R.N. vitamin mix during diet formulation to achieve a 30 mg/kg/d dose. This Kulkarni was supported by NIH grants K08, DK02885, and R01 was calculated by measuring food intake of separate C57BL/6 mice on DK67536. A.V. Biankin is supported by a Cancer Institute New South HFD (HFD [g]/mouse weight [g]/d) and calculating accordingly. Wales fellowship. A.V. Biankin and J.G. Kench are supported by an T2D culture with DFO. Islets were freshly isolated from 3 individuals with NHMRC program grant. We would like to thank Andrew Dwyer, Sof T2D. Islets were cultured overnight, in either control medium or control Andrikopoulos, Cecile King, James Cantley, and Don Chisholm for medium with 125 μM DFO, prior to RNA isolation. helpful comments; Amber Johns for assistance with the human pan- ChIP. ChIP was performed using the Active Motif kit (Carlsbad), accord- creatic slides; Alice Boulghourjian from the Garvan histology-core; ing to the manufacturer’s instructions. The ARNT promoter primers were Will Hughes from the Garvan microscope-core; Ed Feener from the GCTTCCTAGCTCAGGCTTCC and AAGAGCCACTCCGCAGATTA, Joslin Proteomics Core Facility DERC; staff at BTF for maintaining which produce a 250-bp band, which incorporates a GCGTG sequence. the mice; and Tina Patel and Lindy Williams from Westmead and Lina Statistics. Statistics were calculated in Excel or in SPSS version 14. Unless Mariana from St. Vincent’s Melbourne for human islet isolations. otherwise specified, Student’s t test with unequal variance was used to compare groups. For all figures, error bars indicate ± SEM. P values of less Received for publication November 4, 2009, and accepted in than 0.05 were considered significant. revised form March 10, 2010.

Acknowledgments Address correspondence to: Jenny E. Gunton, Diabetes and Tran- J.E. Gunton was funded by the National Health and Medical Research scription Factors Group, Garvan Institute of Medical Research, 384 Council of Australia (NHMRC), Diabetes Australia Research Trust, Victoria St., Darlinghurst, Sydney, New South Wales 2010, Austra- Juvenile Diabetes Research Foundation (JDRF), the Royal Australasian lia. Phone: 011.61.2.9295.8474; Fax: 011.61.2.9295.8404; E-mail: College of Physicians Pfizer and Servier postdoctoral fellowships, and [email protected].

1. Hofer T, Wenger H, Gassmann M. Oxygen sensing, a positive factor in solid tumor growth. Cancer changes. J Biol Chem. 1997;272(36):22642–22647. HIF-1α stabilization and potential therapeutic Research. 2000;60(15):4010–4015. 29. Triantafyllou A, Liakos P, Tsakalof A, Georgatsou strategies. Pflugers Arch. 2002;443(4):503–507. 15. Semenza GL. HIF-1 and human disease: one highly E, Simos G, Bonanou S. Cobalt induces hypoxia- 2. Semenza GL. Regulation of mammalian O2 involved factor. Genes Dev. 2000;14(16):1983–1991. inducible factor-1alpha (HIF-1alpha) in HeLa cells homeostasis by hypoxia-inducible factor 1. Annu 16. Woodward ER, Maher ER. Von Hippel-Lindau dis- by an iron-independent, but ROS-, PI-3K- and Rev Cell Dev Biol. 1999;15:551–578. ease and endocrine tumour susceptibility. Endocr MAPK-dependent mechanism. Free Radic Res. 2006; 3. Kotch LE, Iyer NV, Laughner E, Semenza GL. Defective Relat Cancer. 2006;13(2):415–425. 40(8):847–856. vascularization of HIF-1α-null embryos is not associ- 17. Guillemin K, Krasnow MA. The hypoxic response: 30. Maxwell PH, et al. The tumour suppressor pro- ated with VEGF deficiency but with mesenchymal cell Huffing and HIFing. Cell. 1997;89(1):9–12. tein VHL targets hypoxia-inducible factors for death. Dev Biol. 1999;209(2):254–267. 18. Kewley RJ, Whitelaw ML, Chapman-Smith A. The oxygen-dependent proteolysis. Nature. 1999; 4. Iyer NV, et al. Cellular and developmental control mammalian basic helix-loop-helix/PAS family of 399(6733):271–275. of O2 homeostasis by hypoxia inducible factor 1α. transcriptional regulators. Int J Biochem Cell Biol. 31. Epstein ACR, et al. C. elegans EGL-9 and mamma- Genes Dev. 1998;12(2):149–162. 2004;36(2):189–204. lian homologs define a family of dioxygenases that 5. Cowden Dahl KD, et al. Hypoxia-inducible factors 19. Bunn HF, Poyton RO. Oxygen sensing and molec- regulate HIF by prolyl hydroxylation. Cell. 2001; 1{alpha} and 2{alpha} regulate trophoblast differ- ular adoption to hypoxia. Physiological Reviews. 107(1):43–54. entiation. Mol Cell Biol. 2005;25(23):10479–10491. 1996;76(3):839–885. 32. Jaakkola P, et al. Targeting of HIF-alpha to the 6. Ryan HE, Lo J, Johnson RS. HIF-1[alpha] is required 20. Lee JW, Bae SH, Jeong JW, Kim SH, Kim KW. Hypoxia- von Hippel-Lindau ubiquitylation complex by for solid tumor formation and embryonic vascular- inducible factor (HIF-1alpha): its protein stability and O2-regulated prolyl hydroxylation. Science. 2001; ization. EMBO J. 1998;17(11):3005–3015. biological functions. Exp Mol Med. 2004;36(1):1–12. 292(5516):468–472. 7. Brown LM, et al. Reversing hypoxic cell chemore- 21. Schofield CJ, Ratcliffe PJ. Oxygen sensing by HIF 33. Stroka DM, et al. HIF-1 is expressed in normoxic tis- sistance in vitro using genetic and small molecule hydroxylases. Nat Rev Mol Cell Biol. 2004;5(5):343–354. sue and displays an organ-specific regulation under approaches targeting hypoxia inducible factor-1. 22. Gunton JE, et al. Loss of ARNT/HIF1beta medi- systemic hypoxia. FASEB J. 2001;15(13):2445–2453. Mol Pharmacol. 2006;69(2):411–418. ates altered gene expression and pancreatic-islet 34. Pugh CW, Ratcliffe PJ. Regulation of angiogenesis 8. Buchler P, et al. Hypoxia-inducible factor 1 regu- dysfunction in human type 2 diabetes. Cell. 2005; by hypoxia: role of the HIF system. Nat Med. 2003; lates vascular endothelial growth factor expres- 122(3):337–349. 9(6):677–684. sion in human pancreatic cancer. Pancreas. 2003; 23. Wang XL, et al. Ablation of ARNT/HIF1β in liver 35. Cramer T, et al. HIF-1α is essential for myeloid cell– 26(1):56–64. alters gluconeogenesis, lipogenic gene expression, and mediated inflammation. Cell. 2003;112(5):645–657. 9. Walmsley SR, et al. Hypoxia-induced neutrophil serum ketones. Cell Metabolism. 2009;9(5):428–439. 36. Bonello S, et al. Reactive oxygen species activate the survival is mediated by HIF-1alpha-dependent NF- 24. Lum JJ, et al. The transcription factor HIF-1alpha HIF-1alpha promoter via a functional NFkappaB site. kappaB activity. J Exp Med. 2005;201(1):105–115. plays a critical role in the growth factor-dependent Arterioscler Thromb Vasc Biol. 2007;27(4):755–761. 10. Papandreou I, Cairns RA, Fontana L, Lim AL, regulation of both aerobic and anaerobic glycoly- 37. Biswas S, Gupta MK, Chattopadhyay D, Mukho- Denko NC. HIF-1 mediates adaptation to hypoxia sis. Genes Dev. 2007;21(9):1037–1049. padhyay CK. Insulin-induced activation of hypox- by actively downregulating mitochondrial oxygen 25. Ratcliffe PJ. HIF-1 and HIF-2: working alone or togeth- ia-inducible factor-1 requires generation of reac- consumption. Cell Metab. 2006;3(3):187–197. er in hypoxia. J Clin Invest. 2007;117(4):862–865. tive oxygen species by NADPH oxidase. Am J Physiol 11. Kim JW, Tchernyshyov I, Semenza GL, Dang CV. HIF- 26. Ruas JL, Poellinger L. Hypoxia–dependent activa- Heart Circ Physiol. 2007;292(2):H758–H766. 1-mediated expression of pyruvate dehydrogenase tion of HIF into a transcriptional regulator. Semin 38. Kaelin WG Jr. ROS: really involved in oxygen sens- kinase: a metabolic switch required for cellular adap- Cell Dev Biol. 2005;16(4–5):514–522. ing. Cell Metabolism. 2005;1(6):357–358. tation to hypoxia. Cell Metab. 2006;3(3):177–185. 27. Huang LE, Gu J, Schau M, Bunn HF. Regulation 39. Zelzer E, Levy Y, Kahana C, Shilo BZ, Rubinstein 12. Hopfl G, et al. Rescue of hypoxia-inducible fac- of hypoxia-inducible factor 1α is mediated by an M, Cohen B. Insulin induces transcription of target tor-1α-deficient tumor growth by wild-type cells is O2-dependent degradation domain via the ubiqui- genes through the hypoxia inducible factor HIF- independent of vascular endothelial growth factor. tin-proteasome pathway. Proc Natl Acad Sci U S A. 1alpha/ARNT. EMBO J. 1998;17(17):5085–5094. Cancer Research. 2002;62(10):2962–2970. 1998;95(14):7987–7992. 40. Jiang BH, Jiang G, Zheng JZ, Lu Z, Hunter T, Vogt 13. Maxwell PH, Pugh CW, Ratcliffe PJ. Activation of the 28. Salceda S, Caro J. Hypoxia-inducible factor 1α (HIF- PK. Phosphatidylinositol 3-kinase signalling HIF pathway in cancer. Curr Opin Genet Dev. 2001; 1α) protein is rapidly degraded by the ubiquitin- controls levels of hypoxia-inducible factor 1. Cell 11(3):293–299. proteasome system under normoxic conditions. Its Growth Differ. 2001;12(7):363–369. 14. Ryan HE, et al. Hypoxia-inducible factor-1α is stabilization by hypoxia depends on redox induced 41. Treins C, Giorgetti-Peraldi S, Murdaca J, Semenza

12 The Journal of Clinical Investigation http://www.jci.org research article

GL, Van Obberghen E. Insulin stimulates hypoxia- 54. Del Guerra S, et al. Functional and molecular Stampfer MJ. Blood donations and risk of coronary inducible factor 1 through a phosphatidylinositol defects of pancreatic islets in human type 2 diabe- heart disease in men. Circulation. 2001;103(1):52–57. 3-kinase/target of rapamycin-dependent signalling tes. Diabetes. 2005;54(3):727–735. 68. De Sanctis V, Zurlo MG, Senesi E, Boffa C, Cavallo pathway. J Biol Chem. 2002;277(31):27975–27981. 55. Deng S, et al. Structural and functional abnormali- L, Di Gregorio F. Insulin dependent diabetes in 42. Doronzo G, Russo I, Mattiello L, Riganti C, Anfossi ties in the islets isolated from type 2 diabetic sub- thalassaemia. Arch Dis Child. 1988;63(1):58–62. G, Trovati M. Insulin activates hypoxia-inducible jects. Diabetes. 2004;53(3):624–632. 69. Ford ES, Cogswell ME. Diabetes and serum ferritin factor-1alpha in human and rat vascular smooth 56. Hu FB, et al. Dietary fat intake and the risk of coro- concentration among US adults. Diabetes Care. 1999; muscle cells via phosphatidylinositol-3 kinase and nary heart disease in women. N Engl J Med. 1997; 22(12):1978–1983. mitogen-activated protein kinase pathways: impair- 337(21):1491–1499. 70. Fernandez-Real JM, Lopez-Bermejo A, Ricart W. ment in insulin resistance owing to defects in insu- 57. Miller WC, Niederpruem MG, Wallace JP, Linde- Iron stores, blood donation, and insulin sensitivity lin signalling. Diabetologia. 2006;49(5):1049–1063. man AK. Dietary fat, sugar, and fiber predict body and secretion. Clin Chem. 2005;51(7):1201–1205. 43. Laughner E, Taghavi P, Chiles K, Mahon PC, fat content. J Am Diet Assoc. 1994;94(6):612–615. 71. Jiang R, Manson JE, Meigs JB, Ma J, Rifai N, Hu Semenza GL. HER2 (neu) signaling increases the 58. Fukuda R, Zhang H, Kim J, Shimoda L, Dang CV, FB. Body iron stores in relation to risk of type 2 rate of hypoxia-inducible factor 1α (HIF-1α) syn- Semenza GL. HIF-1 regulates cytochrome oxidase diabetes in apparently healthy women. JAMA. 2004; thesis: novel mechanism for HIF-1-mediated vas- subunits to optimize efficiency of respiration in 291(6):711–717. cular endothelial growth factor expression. Mol Cell hypoxic cells. Cell. 2007;129(1):111–122. 72. Wrede CE, Buettner R, Bollheimer LC, Scholmerich Biol. 2001;21(12):3995–4004. 59. Gnarra JR, et al. Defective placental vasculogenesis J, Palitzsch KD, Hellerbrand C. Association between 44. Carlsson PO, Liss P, Andersson A, Jansson L. Measure- causes embryonic lethality in VHL-deficient mice. serum ferritin and the insulin resistance syndrome ments of oxygen tension in native and transplanted Proc Natl Acad Sci U S A. 1997;94(17):9102–9107. in a representative population. Eur J Endocrinol. rat pancreatic islets. Diabetes. 1998;47(7):1027–1032. 60. Kleymenova E, Everitt JI, Pluta L, Portis M, Gnarra 2006;154(2):333–340. 45. Carlsson PO, Palm F. Oxygen tension in isolated JR, Walker CL. Susceptibility to vascular neoplasms 73. Salonen JT, Tuomainen TP, Nyyssonen K, Lakka transplanted rat islets and in islets of rat whole-pan- but no increased susceptibility to renal carcino- HM, Punnonen K. Relation between iron stores creas transplants. Transpl Int. 2002;15(11):581–585. genesis in Vhl knockout mice. Carcinogenesis. 2004; and non–insulin–dependent diabetes in men: 46. Cantley J, et al. Deletion of the von Hippel-Lindau 25(3):309. case–control study. BMJ. 1999;317(7160):727. gene in pancreatic beta cells impairs glucose homeo- 61. Chetty R, Kennedy M, Ezzat S, Asa SL. Pancreatic 74. Fernández-Real JM, et al. Serum ferritin as a com- stasis in mice. J Clin Invest. 2009;119(1):125–135. endocrine pathology in von Hippel-Lindau disease: ponent of the insulin resistance syndrome. Diabetes 47. Puri S, Cano DA, Hebrok M. A role for von Hip- an expanding spectrum of lesions. Endocr Pathol. Care. 1998;21(1):62–68. pel-Lindau protein in pancreatic beta-cell function. 2004;15(2):141–148. 75. Lee DH, Folsom AR, Jacobs DRJ. Dietary iron Diabetes. 2009;58(2):433–441. 62. Perigny M, et al. Pancreatic endocrine microadeno- intake and type 2 diabetes incidence in postmeno- 48. Zehetner J, et al. pVHL is a regulator of glucose matosis in patients with von Hippel-Lindau disease: pausal women: the Iowa Women’s Health Study. metabolism and insulin secretion in pancreatic characterization by VHL/HIF pathway proteins Diabetologia. 2004;47(2):185–194. beta-cells. Genes Dev. 2008;22(22):3135–3146. expression. Am J Surg Pathol. 2009;33(5):739–748. 76. Fernandez-Real JM, Penarroja G, Castro A, Garcia- 49. Tomita S, et al. Defective brain development in 63. Lott ST, et al. High frequency loss of heterozygos- Bragado F, Hernandez-Aguado I, Ricart W. Blood let- mice lacking the Hif-1alpha gene in neural cells. ity in von Hippel-Lindau (VHL)-associated and spo- ting in high-ferritin type 2 diabetes: effects on insu- Mol Cell Biol. 2003;23(19):6739–6749. radic pancreatic islet cell tumors evidence for a step- lin sensitivity and beta-cell function. Diabetes. 2002; 50. Gerich JE. The genetic basis of type 2 diabetes mel- wise mechanism for malignant conversion in VHL 51(4):1000–1004. litus: impaired insulin secretion versus impaired Tumorigenesis 1. Cancer Res. 2002;62(7):1952–1955. 77. Yim SH, et al. Disruption of the Arnt gene in insulin sensitivity. Endocr Rev. 1998;19(4):491–503. 64. Hammel PR, et al. Pancreatic involvement in von endothelial cells causes hepatic vascular defects and 51. Butler AE, Janson J, Bonner-Weir S, Ritzel R, Rizza Hippel-Lindau disease. Gastroenterology. 2000; partial embryonic lethality in mice. Hepatology. 2006; RA, Butler PC. Beta-cell deficit and increased beta- 119(4):1087–1095. 44(3):550–560. cell apoptosis in humans with type 2 diabetes. Dia- 65. Pollard PJ, et al. Accumulation of Krebs cycle inter- 78. Miyazaki J, et al. Establishment of a pancreatic betes. 2003;52(1):102–110. mediates and over-expression of HIF1{alpha} in beta cell line that retains glucose inducible insu- 52. Bagdade JD, Bierman EL, Porte D Jr. The signifi- tumours which result from germline FH and SDH lin secretion: Special reference to expression of cance of basal insulin levels in the evaluation of the mutations. Hum Mol Genet. 2005;14(15):2231–2239. glucose transporter isoforms. Endocrinology. 1990; insulin response to glucose in diabetic and nondia- 66. Ciara MM, Diana EB, Deborah JM, Bruce GR, Mar- 127(1):126–132. betic subjects. J Clin Invest. 1967;46(10):1549–1557. garet RZ. K40E: a novel succinate dehydrogenase 79. Gunton JE, Delhanty PJ, Takahashi S, Baxter RC. 53. Marchetti P, et al. Pancreatic islets from type 2 dia- (SDH)B mutation causing familial phaeochromo- Metformin rapidly increases insulin receptor acti- betic patients have functional defects and increased cytoma and paraganglioma. Clin Endocrinol (Oxf). vation in human liver and signals preferentially apoptosis that are ameliorated by metformin. J Clin 2004;61(4):510–514. through insulin-receptor substrate-2. J Clin Endo- Endocrinol Metab. 2004;89(11):5535–5541. 67. Ascherio A, Rimm EB, Giovannucci E, Willett WC, crinol Metab. 2003;88(3):1323–1332.

The Journal of Clinical Investigation http://www.jci.org 13