Metabolic Responses to High–Fat Diet-Induced Obesity: the Long and Short of It

Metabolic responses to high–fat diet-induced obesity: the long and short of it

Christopher Carmine Meoli

Supervisor: Prof. David E. James

Co-supervisor: Prof Gregory J. Cooney

A thesis in fulfilment of the requirements for the degree of

Doctor of Philosophy

Garvan Institute of Medical Research

St. Vincent’s Medical School

Faculty of Medicine

May 2017

Dedicated to the life of Lorenzo Meoli

(RIP 1952 – 2014)

Thank you for giving me the guidance to be the man I am today.

To my beautiful girls, Lesley and Eliana, and my son Sebastian

I am eternally indebted to you all,

for the love, support and strength

you give me every day

‘I hereby grant the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or 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 proprietary 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 Abstract International (this is applicable to doctoral theses only). I have either used no substantial portions of copyright material in my thesis or I have obtained permission to use copyright material; where permission has not been granted I have applied/will apply for a partial restriction of the digital copy of my thesis or dissertation.'

Signed ……………………………………………......

Date ……………………………………………......

AUTHENTICITY STATEMENT

‘I certify that the Library deposit digital copy is a direct equivalent of the final officially approved version of my thesis. No emendation of content has occurred and if there are any minor variations in formatting, they are the result of the conversion to digital format.’

Signed ……………………………………………......

Date ……………………………………………......

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 ……………………………………………...... THE UNIVERSITY OF NEW SOUTH WALES Thesis/Dissertation Sheet

Surname or Family name: Meoli

First name: Christopher Other name/s: Cannine

Abbreviation for degree as given in the University calendar: PhD

School: St Vincent's clinical School Faculty: Medicine

Title: Metabolic responses to high-fat diet-induced obesity: the long and shortof it

Abstract 350 words maximum: (PLEASE TYPE)

With the increasing incidence of obesity and related diseases there is a great deal of interest in understanding the impact of a western diet on long term health. Moreover, diet and aging are linked to insulin resistance a major risk factor for a constellation of diseases including osteoporosis, Alzheimer's disease and type 2 diabetes. However, the direct contribution of diet versus aging to these processes is not clear. Our investigation begins in chapter 3 where we establish and characterise a model of high-fat diet-induced obesity in the form of the HFD fed CS7BL/6 mouse. Using the established model we then validate the therapeutic potential of VEGF neutralisation on insulin resistance, and discovered that aVEGF therapy was able to reverse whole body glucose intolerance, via an improvement in hepatic insulin sensitivity. Using the knowledge developed in Chapters 3 and 4 we then returned to our original question and designed a study, which could address two aspects: 1) Track the acute and long term metabolic consequences and/or adaptations to a HFD, and 2) quantify the relative contribution of diet and age to disease, beyond glucose metabolism. HFD feeding resulted in significant glucose intolerance within 1 d of feeding and this was sustained for 6 months after which it began to resolve until complete resolution by 12 months on the diet. The resolution of glucose intolerance was due to a striking compensation by the pancreas as a result of beta cell proliferation. Preliminary studies indicate that this may be due to an unexpected stimulatory effect of leptin on insulin secretion. The HFD also caused marked deterioration in bone morphology and brain function, as measured by memory deficit and a change in the Amyloid Beta 40 to 42 ratio. The effect of aging per se on each of these parameters was mild compared to the effect of diet. We conclude that western style diets have profound deleterious effects on multiple organ systems and the effect of diet is much more potent than age alone.

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 dissertationin whole or in part in the University libraries in all forms of media, now or here afterknown, 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). · ······························ ...... ?.-:?:·..--:. .-?. ..-. ..I . .r...... 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 circumstancesand re uire the a roval 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 Table of Contents

Table of Contents ...... i Preface ...... v Abstract ...... vi List of Publications ...... vii List of Oral and Poster Presentations ...... viii List of Figures ...... ix List of Tables ...... x Abbreviations ...... xi Acknowledgements ...... xiii Chapter 1- Introduction ...... 2 1.1 General Introduction ...... 2 1.2 An introduction to diabetes ...... 3 1.2.1 Classifying diabetes ...... 4 1.2.2 Type 1 diabetes ...... 4 1.2.3 Type 2 diabetes ...... 5 1.3 Insulin and the maintenance of glucose homeostasis ...... 6 1.4 Cellular insulin signalling ...... 9 1.5 Insulin Resistance ...... 10 1.5.1 Therapeutics for insulin resistance ...... 12 1.5.2 Mouse models of insulin resistance...... 13 1.5.2.1 Rationale for use of rodent models ...... 13 1.5.2.2 Genetic models of insulin resistance ...... 14 1.6 High-fat diet-induced obesity ...... 17 1.6.1 Experimental design considerations ...... 20 1.6.1.1 Strain ...... 20 1.6.1.2 Age and duration of diets ...... 21 1.7 Whole body metabolic effects of high-fat diet ...... 23 1.7.1 Glucose, insulin and glucose tolerance ...... 23 1.7.2 Altered Adipokine expression ...... 25 1.7.3 Ectopic lipid deposition ...... 26 1.8 The metabolic effects of HFD feeding by tissue ...... 28 1.8.1 Liver ...... 28 1.8.2 Skeletal muscle ...... 30 1.8.3 Adipose tissue...... 31 1.8.3.1 Adipocyte hypoxia and angiogenesis ...... 33 i

1.8.4 Pancreas / β-cell ...... 34 1.8.4.1 Leptin and Leptin resistance ...... 37 1.9 Non-classical tissues affected by IR ...... 41 1.9.1 Brain ...... 41 1.9.2 Bone...... 44 1.10 Concluding remarks ...... 46 1.11 Hypothesis ...... 47 1.12 Aims ...... 47 Chapter 2 – Materials & Methods ...... 49 2.1 Experimental chapters ...... 49 2.2 Animal Research ...... 50 2.3 General methods ...... 50 2.3.1 Animals and Husbandry ...... 50 2.3.2 Experimental diets and food intake ...... 51 2.3.3 Body composition...... 54 2.3.4 Indirect calorimetry ...... 54 2.3.5 In vivo measurements ...... 56 2.3.5.1 Glucose and insulin tolerance tests and blood measurements ...... 56 2.3.5.2 Tracer uptake in to WAT and skeletal muscle ...... 56 2.3.6 Ex vivo measurements...... 57 2.3.6.1 In vitro glucose uptake in adipose ...... 57 2.3.6.2 In vitro glucose uptake in muscle ...... 58 2.3.7 Tissue embedding and serum collection processing ...... 58 2.3.8 Triglyceride and glycerol measurements ...... 59 2.4 Chapter 3 methods ...... 60 2.4.1 Dietary intervention studies...... 60 2.4.2 Statistics...... 60 2.5 Chapter 4 methods ...... 61 2.5.1 VEGF neutralization...... 61 2.5.2 Western blots ...... 61 2.5.3 Hyperinsulinemic Clamps ...... 61 2.5.4 Adipocyte diameter measurements ...... 62 2.5.5 Endothelial cell proliferation ...... 62 2.5.6 qPCR ...... 62 2.5.7 Scanning electron microscopy...... 63 2.5.8 Statistics...... 63 2.6 Chapter 5 methods ...... 64 ii

2.6.1 In vivo assessment of insulin clearance ...... 64 2.6.2 Islet isolation and in vitro secretion ...... 64 2.6.3 Pancreatic immunohistochemistry ...... 64 2.6.4 Micro-computed tomography of bone ...... 66 2.6.5 Aβ 40 and 42 measurements ...... 67 2.6.6 Behavioural testing ...... 67 2.6.6.1 Open field test (Baseline movement measurements) ...... 67 2.6.6.2. Elevated plus maze (Anxiety) ...... 68 2.6.6.4. Y-maze (Short term memory) ...... 68 2.7 Mass spectrometry ...... 69 2.7.1 Proteomics – Sample preparation ...... 69 2.7.2 Mass spectrometry ...... 69 2.7.3 Data processing and analysis ...... 70 2.7.4 Differential expression (DE) Analysis ...... 70 2.7.5 Gene set test (GST ...... 70 2.8 Statistical Analysis ...... 72 Chapter 3 - The metabolic consequences of acute high-fat diet feeding ...... 74 3.1 Abstract ...... 75 3.2 Introduction ...... 77 3.3 Methods ...... 78 3.4 Results ...... 78 3.4.1 Adiposity but not lean mass change with a high fat diet ...... 78 3.4.2 The dynamic progression of IR following ongoing exposure to HFD ...... 80 3.4.3 A temporal analysis of insulin resistance in WAT & skeletal muscle ...... 83 3.4.4 Exploring the metabolic flexibility of mice ...... 87 3.5 Discussion ...... 93 Chapter 4 - Systemic vascular endothelial growth factor-A (VEGF-A) neutralisation ameliorates diet induced metabolic dysfunction...... 102 4.1 Abstract ...... 103 4.2 Introduction ...... 104 4.3 Methods ...... 106 4.4 Results ...... 106 4.4.1 VEGF-A neutralization blocks the onset of diet-induced glucose intolerance ...... 107 4.4.2 VEGF-A neutralisation reverses glucose intolerance following extended exposure to high fat feeding ...... 114 4.4.3 VEGF-A neutralization improves hepatic insulin sensitivity...... 119

iii

4.4.4 VEGF-A neutralisation had no effect on adiposity, ectopic lipid deposition or body weight ...... 124 4.4.5 Evidence for altered lipid uptake with VEGF neutralization ...... 126 4.5 Discussion ...... 129 Chapter 5 – The metabolic and health consequences of long term high-fat diet feeding ...... 136 5.1 Abstract ...... 136 5.2 Introduction ...... 137 5.3 Methods ...... 139 5.4 Results ...... 139 5.4.1 Diet is a more potent regulator of adiposity than age...... 139 5.4.2 Assessing energy substrates and activity...... 140 5.4.3 Metabolic phenotyping reveals an adaptive response to long term HFD feeding ...... 141 5.4.4 Assessing peripheral Insulin resistance ...... 146 5.4.5 Ectopic lipid deposition with HFD feeding does not correlate with glucose intolerance ...... 149 5.4.6 Hyperinsulinemia –In vivo assessment of insulin secretion and clearance ...... 151 5.4.7 An investigation in to pancreas morphology and βcell mass using mosaic microscopy ...... 152 5.4.8 Serum cytokines and adipokines ...... 153 5.4.9 Leptin potentiates glucose stimulated insulin secretion under HFD feeding conditions ...... 159 5.4.10 The effects of diet and age on skeletal architecture ...... 162 5.4.11 High fat diet increases susceptibility to neurodegenerative diseases .... 166 5.4.12 assessing organismal health via the liver proteome ...... 169 5.5 Discussion ...... 176 6.1 General discussion ...... 185 7.1 Appendix 1 ...... 193 7.2 Appendix 2 ...... 199 7.3 References ...... 202

Preface

In my four years as a PhD student, I have had the honour of working on a variety of different projects, with the diet-induced obese mouse model serving as my constant.

Upon joining the lab my efforts centred on characterising the metabolic responses of mice to acute high-fat diet (HFD) feeding. Once we understood this we were then able to use the model to interrogate the effectiveness of a drug based therapy for human insulin resistance. We then went on to characterise the metabolic responses to prolonged

HFD feeding and compared the effects of diet and age on health. Using this long term feeding model I have been able to make some great leaps in knowledge, showing for the first time how this model can best be used to recapitulate the pathophysiology of human insulin resistance, and actually quantify the deleterious effects diet has on long term health, a none too easy task. In the following I will give a brief overview of this thesis.

Abstract

Chapters 3 and 4 we then returned to our original question and designed a study, which could address two aspects: 1) Track the acute and long term metabolic consequences and/or adaptations to a HFD, and 2) quantify the relative contribution of diet and age to disease, beyond glucose metabolism. HFD feeding resulted in significant glucose intolerance within 1 d of feeding and this was sustained for 6 months after which it began to resolve until complete resolution by 12 months on the diet. The resolution of glucose intolerance was due to a striking compensation by the pancreas as a result of beta cell proliferation. Preliminary studies indicate that this may be due to an unexpected stimulatory effect of leptin on insulin secretion. The HFD also caused marked deterioration in bone morphology and brain function, as measured by memory deficit and a change in the Amyloid Beta 40 to 42 ratio. The effect of aging per se on each of these parameters was mild compared to the effect of diet. We conclude that western style diets have profound deleterious effects on multiple organ systems and the effect of diet is much more potent than age alone. vi

List of Publications

Fazakerley, D. J., Naghiloo, S., Chaudhuri, R., Koumanov, F., Burchfield, J. G., Thomas, K. C., Krycer, J. R., Prior, M. J., Parker, B. L., Murrow, B. A., Stockli, J., Meoli, C. C., Holman, G. D., & James, D. E. (2015). Proteomic Analysis of GLUT4 Storage Vesicles Reveals Tumour Suppressor Candidate 5 (TUSC5) as a Novel Regulator of Insulin Action in Adipocytes. J Biol Chem. doi: 10.1074/jbc.M115.657361

King, G. J., Stockli, J., Hu, S. H., Winnen, B., Duprez, W. G., Meoli, C. C., Junutula, J. R., Jarrott, R. J., James, D. E., Whitten, A. E., & Martin, J. L. (2012). Membrane curvature protein exhibits interdomain flexibility and binds a small GTPase. J Biol Chem, 287(49), 40996-41006. doi: 10.1074/jbc.M112.349803

Li, J., Cantley, J., Burchfield, J. G., Meoli, C. C., Stockli, J., Whitworth, P. T., Pant, H., Chaudhuri, R., Groffen, A. J., Verhage, M., & James, D. E. (2014). DOC2 isoforms play dual roles in insulin secretion and insulin-stimulated glucose uptake. Diabetologia, 57(10), 2173-2182. doi: 10.1007/s00125-014-3312-y

Stockli, J., Meoli, C. C., Hoffman, N. J., Fazakerley, D. J., Pant, H., Cleasby, M. E., Ma, X., Kleinert, M., Brandon, A. E., Lopez, J. A., Cooney, G. J., & James, D. E. (2015). The RabGAP TBC1D1 plays a central role in exercise-regulated glucose metabolism in skeletal muscle. Diabetes. doi: 10.2337/db13-1489

Tan, S. X., Fisher-Wellman, K., Fazakerley, D., Ng, Y., Pant, H., Li, J., Meoli, C., Coster, A. C., Stockli, J., & James, D. E. (2015). Selective Insulin resistance in adipocytes. J Biol Chem. doi: 10.1074/jbc.M114.623686

Tan, S. X., Ng, Y., Meoli, C. C., Kumar, A., Khoo, P. S., Fazakerley, D. J., Junutula, J. R., Vali, S., James, D. E., & Stockli, J. (2012). Amplification and demultiplexing in insulin-regulated Akt protein kinase pathway in adipocytes. J Biol Chem, 287(9), 6128-6138. doi: 10.1074/jbc.M111.318238

* Wu, L. E., Meoli, C. C., Mangiafico, S. P., Fazakerley, D. J., Cogger, V. C., Mohamad, M., Pant, H., Kang, M. J., Powter, E., Burchfield, J. G., Xirouchaki, C. E., Mikolaizak, A. S., Stockli, J., Kolumam, G., van Bruggen, N., Gamble, J. R., Le Couteur, D. G., Cooney, G. J., Andrikopoulos, S., & James, D. E. (2014). Systemic VEGF-A neutralization ameliorates diet-induced metabolic dysfunction. Diabetes, 63(8), 2656-2667. doi: 10.2337/db13-1665

* Co-first authorship

vii

List of Oral and Poster Presentations

Christopher C. Meoli, Daniel Fazakerley, Shi.-Xiong Tan, Jacqueline Stoeckli, David E. James. The timeline of high fat diet-induced insulin resistance. The Garvan Institute of Medical Research (2012). Sydney, NSW, Australia (Oral)

Christopher C. Meoli, Daniel Fazakerley, Shi.-Xiong Tan, Jacqueline Stoeckli, David E. James. The timeline of high fat diet-induced insulin resistance. The Australian Diabetes Society (2012). Brisbane, QLD, Australia (Poster)

Christopher C. Meoli, Daniel Fazakerley, Rima Chaudhuri, Sean Humphrey, Tristan Iseli, Sally Coulter, Annabel Minard, Jacqueline Stoeckli, Himani Joshi, James Cantley, Tess Whitworth, David E. James.. Diet is a more potent regulator of the total liver proteome than ageing. The Australian Diabetes Society (2013). Sydney, NSW, Australia (Poster)

Christopher C. Meoli, Daniel Fazakerley, Shi.-Xiong Tan, Jacqueline Stoeckli, David E. James. The consequences of long term high fat feeding. The Garvan Institute of Medical Research (2013). Sydney, NSW, Australia (Oral)

Christopher C. Meoli, Daniel Fazakerley, Tess Whitworth, Rima Chaudhuri, Benjamin Parker, Sean Humphrey, Jacqueline Stoeckli, James Cantley, Natalie Wee, Amanda Wright, Bryce Vissel, Paul Baldock, Gregory Cooney, James Burchfield, David E. James. Long term high-fat-diet feeding reveals diet as a major regulator of age-related diseases. The Australian Diabetes Society (2015). Adelaide, SA, Australia (Oral)

viii

List of Figures

Figure 1.1 - Insulin, Glucagon and glucose homeostasis...... 7 Figure 1.2 - The many actions of insulin...... 9 Figure 1.3 - Dysregulation of the Adipo-Insular axis...... 41 Figure 2.1 - Representative DEXA images...... 55 Figure 3.1 – Body composition in response to HFD feeding...... 79 Figure 3.2 –The effects of HFD on glucose tolerance...... 81 Figure 3.3 – A longitudinal study of insulin in HFD fed mice ...... 82 Figure 3.4 – A temporal analysis of insulin resistance in WAT ...... 84 Figure 3.5 – A temporal analysis of insulin resistance in skeletal muscle...... 86 Figure 3.6 - A temporal analysis of metabolic flexibility in response to a change in diet...... 88 Figure 3.7 - A temporal analysis of WAT adaptability to a change in diet...... 91 Figure 3.8 - Glucose tolerance testing in naïve and HFD exposed mice...... 92 Figure 4.1- Biological activity of VEGF-A neutralising antibody B20-4.1...... 107 Figure 4.2 - Glucose tolerance during VEGF neutralization in acute high fat fed mice...... 110 Figure - 4.3 VEGF neutralization...... 112 Figure 4.4 - VEGF neutralization in long term high fat fed mice...... 116 Figure 4.5 - Fasting glucose levels following VEGF neutralization...... 117 Figure 4.6 - Insulin tolerance during VEGF neutralization in long term high fat fed mice...... 117 Figure 4.7 - Glucose tolerance testing following VEGF antibody re-administration. .. 118 Figure 4.8 Insulin sensitivity and hepatic glucose output during VEGF neutralization...... 120 Figure 4.9. Phosphorylation of insulin signalling intermediates during VEGF neutralization...... 122 Figure 4.10 - Adiposity during VEGF neutralization...... 125 Figure 4.11 - Ectopic lipids in VEGF antibody treated mice...... 127 Figure 4.12 Fenestration of hepatic sinusoidal endothelial cells during VEGF neutralisation...... 128 Figure 5.1 - Assessing body composition and food intake with long term HFD feeding...... 143 Figure 5.2 – Assessing substrate utilisation and activity via indirect calorimetry...... 144 Figure 5.3 - Temporal metabolic phenotyping reveals an adaptive response...... 145 Figure 5.4 – Assessing peripheral insulin resistance in WAT and muscle...... 148 Figure 5.5 – Assessing ectopic lipid deposition in liver and muscle...... 150 Figure 5.6 - Investigating in vivo insulin secretion and clearance...... 155 Figure 5.7 – Pancreata and islet quantification with HFD feeding...... 156 Figure 5.8 – Glucose stimulated insulin secretion in 60 wk mice...... 161 Figure 5.9 – Assessing bone structure and formation...... 164 Figure 5.10 – Assessing bone integrity...... 165 Figure 5.11 - Aβ plaque load and memory testing in long term HFD C57BL6 mice. .. 168 Figure 5.12 – Visual assessment of liver morphology and lipid accumulation...... 172 Figure 5.13 - Differentially expressed proteins with age and diet...... 173 Figure 5.14 - Liver proteomics and gene set enrichment test...... 175

List of Tables

Table 2.3 - Mineral mix recipe for HFD...... 52 Table 2.2 - Dry ingredients recipe for HFD...... 53 Table 2.3 - Wet ingredients recipe for HFD...... 53 Table 2.4 - Antibodies...... 66 Table 4.1. Effects of VEGF-neutralising antibody on metabolic parameters in chow-and high-fat-fed mice, during an acute model of WD feeding...... 113 Table 4.2 – Serum cytokines...... 123 Table 5.1 – Serum panel of cytokines and adipokines ...... 158

Abbreviations

Ab Antibody Aβ Amyloid Beta ACC2 Acetyl CoA carboxylase-2 AD Alzheimer’s Disease AMPK 5’AMP-activated protein kinase AUC Area Under the Curve BCAA Branch Chain Amino Acids BMC Bone Mineral Content BMD Bone Mineral Density BW Body Weight CVD Cardiovascular Disease d Day DAGs Diacylglycerols DEXA Dual-Energy X-ray Absorptiometry DGAT1 Diacylglycerol acyltransferase 2 DIO Diet Induced Obesity ECM Extra Cellular Matrix EDL Extensor Digitorum Longus FABP4 Fatty Acid Binding Protein 4 FFA Free Fatty Acids FFM Fat Free Mass GIR Glucose Infusion Rate GIP Gastric inhibitory peptide GLUT4 Glucose Transporter 4 GLP-1 Glucagon-like peptide 1 GPAT1 Glycerol 3-phosphate acyltransferase-1 GTT Glucose Tolerance Test HDL High-Density Lipoprotein HFD High Fat Diet HGO Hepatic Glucose Output HIF1α Hypoxia-inducible factor 1-α HOMA Homeostatic Model Assessment xi

IMCLs Intramyocellular lipids IR Insulin Resistance IRS Insulin receptor substrate JAK Janus Kinase KO Knockout LCFAs Long Chain Fatty Acids LDL Low-Density Lipoprotein Lep Leptin LFD Low Fat Diet MAPK Mitogen-Activated Protein Kinase MCFAs Medium Chain Fatty Acids MS Mass Spectrometry NAFLD Non-Alcoholic Fatty Liver Disease NEFA Non-Esterified Fatty Acids NIDDM Non-Insulin Dependent Diabetes Mellitus ObRb Leptin Receptor Isoform b OGTT Oral Glucose Tolerance Test PEPCK Phosphoenolpyruvate carboxykinase PI3K Phosphatidylinositol 3 Kinase RBP4 Retinol Binding Protein 4 RER Respiratory Exchange Ratio ROS Reactive Oxygen Species RT Room Temperature SOL Soleus SREPB1 Sterol Response Element Binding Protein 1c STAT Signal Transducers and Activators of Transcription T1D Type 1 Diabetes T2D Type 2 Diabetes TNFα Tumour Necrosis Factor α VDCC voltage-dependent calcium channels WAT White Adipose Tissue WHO World Health Organisation Wk Week

xii

Acknowledgements

First and foremost I would like to thank my supervisor, Prof. David James. Thank you for seeing something in me and giving me the opportunity to participate in the amazing world of research. Thank you for your guidance, support, trust, intellectual input and of course patience, which I am sure I tested at times. David your passion for science is unparalleled and contagious and I hope this body of work lives up to your expectations. Lastly, thank you for creating such a fun and amazing lab, I feel so privileged to have been able to call this place my work for the last 6 years.

I would also like to thank the James lab and its many members, both past and present. Thank you for your friendship, stimulating scientific conversations and support. A special thank you goes out to Daniel Fazakerley. Thank you for the guidance and advice throughout this project, I would not stand where I am today if it were not for your involvement in this project. I see a great scientist in you and I know you will go very far and make great contributions to human health and knowledge. To James Burchfield, thank you for all your help with my project especially the microscopy. You truly are an expert in your field. But most of all thank your friendship (which I am sure will last a lifetime) and support through some very tough times in my life, I hope to one day return the favour. Lastly, to Samantha Oakes and Adelaide Young, thank you for giving me your support, and guidance when I was down and out.

To the many members of the James lab, especially my fellow PhD students; Annabel, Beverley, Dougal and Martin. Never squander the amazing opportunity we have here to contribute knowledge to mankind. No job on earth will ever give us such an amazing opportunity as the one that stands in front of all of you right now.

My PhD took an unexpected turn when my father was diagnosed with an extremely rare cancer in November 2013. I spent many early mornings and late nights by his bedside and true to his wishes I worked very hard not to let it interfere with my studies. Even when he was sick he wanted to know how work was going, because above all else he valued education. On the 11th of June, 2014, my father lost his battle with cancer. Dad I know how proud you were of everything I did, and a PhD was going to be the crowning glory. No doubt you are watching over me, and as I write every word in this thesis I carry your spirit with me. Dad, we did it!

xiii

To my beautiful family, mum (Rosa), Joseph, Angelica, Jillian and of course Nonna Ida. You have all supported me in one way or another, and I am eternally grateful for the love and kindness you have all given me throughout my life, and during my PhD. Without it I would not be where I am today. I hope the completion of this brings you some joy and satisfaction.

During the first year of my PhD I got married to the love of my life, Lesley Castillo Meoli. Lesley, I know the last 5 years have not been easy. Thank you for picking me up when I was down, and for believing in me. Without your love and support none of this would have been possible. Your amazing ability to sacrifice your own needs for mine is the reason I get to complete my PhD. This thesis is as much mine as it is yours. To our baby girl, Eliana, a.k.a. “Cheeky monkey”, I am sorry I have been so absent in the first 6 months of your life. To my son Sebastian, the latest edition to our family, I promise the thesis revisions will be over soon, and daddy will cuddle you to sleep once again. I make these sacrifices to better all our futures. I hope that it will be worth it, and I promise to be there for the three of you, for the rest of my life.

Eternally grateful

Christopher Meoli

xiv

An introduction to insulin resistance and

high-fat diet-induced obesity

Chapter 1- Introduction

1.1 General Introduction

Medical advances over the past hundred years or so, most notably the widespread use of vaccination and antibiotics, have resulted in a shift in the types of diseases impacting human health. In the developed world, arguably the two most important determinants of health and quality of life are age and diet.

Humans evolved in conditions of food scarcity, and as such are genetically adapted to scarcity of nutrients. So called thrifty genes have developed to take advantage of times of food abundance by depositing excess lipid as fat and protecting humans from starvation in times of scarcity. The recent industrialisation of agriculture has succeeded in overcoming many of these challenges (Bonomini, Rodella, & Rezzani, 2015) and

(mis)placed these putative thrifty genes in an environment of constant food abundance.

The result has been an explosion in obesity, diabetes and associated co-morbidities.

The interaction between humans genetically adapted for nutrient scarcity and the pervasiveness of caloric excess in modern western society has resulted in a worldwide explosion in obesity. Indeed, the Global Burden of Disease Study conducted in 2010 identified diet as the number one risk factor for disease in Australasia, followed closely by high body mass index. These two factors are interconnected. In Australia it is currently estimated that 63% of the adult population and 25% of children are overweight or obese (AIHW, 2013). This has profound health and economic implications. Diet and obesity are strongly linked to the development of an array of diseases such as insulin resistance (IR), type 2 diabetes (T2D), cardiovascular disease, urogenital disease and cancer. These diseases now pose the greatest threat to long term

2 health in the developed world. Thus, understanding the pathophysiology of these diseases now represents one of the most significant challenges to human health and longevity.

1.2 An introduction to diabetes

Diabetes is a chronic disorder marked by hyperglycaemia. Until recently, diabetes was considered a low agenda health issue by international health agencies and governments. Since the end of WWII, diabetes incidence has steadily increased due to an unprecedented combination of rising agricultural yields, the mechanisation of labour- intensive work, increased income, urbanisation and the increase in cheap, energy dense foods (Caballero, 2007). In the past two decades alone the number of individuals affected by diabetes has more than doubled worldwide (P. Zimmet, Alberti, & Shaw,

2001) and diabetes has risen to the forefront of international healthcare agendas (P. Z.

Zimmet, Magliano, Herman, & Shaw, 2014). While this helps bring the issue to the forefront, we are still far from eliminating diabetes as a public health issue.

Initially, it was thought that the growing incidence of diabetes was confined to post- industrial nations such as the United States, Europe and Australia. However, epidemiological studies have now shown that diabetes is also growing at alarming rates in developing nations. The International Diabetes Federation estimated that 387 million people globally had diabetes in 2014 (IDF, 2014), and that this number will have increased by 55% to 592 million people worldwide by 2035. Morbidity rates are at an all-time high and it is estimated that currently 1 person dies from diabetes every seven seconds equating to 4.9 million deaths in 2014. As of 2010 global health expenditure towards diabetes reached approximately US$376 billion, and this is expected to reach between $490 and $893 billion by 2030(P. Z. Zimmet et al., 2014). These figures may

3 actually underestimate the true cost as they fail to take in to account indirect expenditure towards the treatment of comorbidities that often follow on as a result of diabetes such as cardiovascular and renal disease (P. Z. Zimmet et al., 2014).

1.2.1 Classifying diabetes

Traditionally diabetes has been broken down in to 3 main categories; Diabetes

Mellitus Type 1 (T1D), non-insulin-dependent diabetes mellitus Diabetes (NIDDM)

Mellitus Type 2 more commonly referred to as T2D and gestational diabetes. Recent studies have shown that diabetes; in particular, T2D is a far more heterogeneous disease than was first appreciated ((Tuomi et al., 2014), (Tonks et al., 2013)), and may be worthy of further division.

1.2.2 Type 1 diabetes

Type 1 diabetes (T1D), previously known as insulin dependent, juvenile or childhood-onset diabetes is characterised by the loss of insulin producing beta cells which reside within the pancreas. This results in a lack of, or severe reduction in insulin secretion. T1D accounts for approximately 5-10% of all diabetes cases (Saltiel, 2001).

The molecular causes of T1D remain unknown and many cases are idiopathic, however in cases where a mechanism is involved it is widely accepted to be due to the auto- immune destruction of the islets of Langerhans, which contain the insulin producing β cells (Mathis, Vence, & Benoist, 2001) (Tuomi et al., 2014). T1D involves several years of gradual worsening in glucose regulation as the beta cells are lost.

Administration of exogenous insulin is required for the rest of individuals’ life in order to maintain normoglycaemia.

1.2.3 Type 2 diabetes

T2D is often referred to as “adult onset” or NIDDM, and typically occurs later in life.

Of all diabetes, T2D accounts for 90-95% of all diagnosed cases worldwide (P. Zimmet et al., 2001; P. Z. Zimmet et al., 2014). Whereas in type 1 diabetes, the sole cause of the disorder is the complete destruction of pancreatic β-cells in a relatively short time, type

2 diabetes is a much slower evolving disorder that includes multiple tissue pathologies

(Saltiel & Kahn, 2001) and is characterised by hyperglycemia, IR, and relative insulin deficiency.

Considered primarily as a lifestyle disorder, the increase in T2D is attributed to an increasingly sedentary lifestyle, combined with an abundance of high energy food sources. Although the pathogenesis is strongly influenced by genetics, environment, ethnicity and age, the perturbation in energy balance is thought to be the greatest driving force behind the dramatic increases in obesity, IR and ultimately the diabetes epidemic

(Chisholm, Campbell, & Kraegen, 1997). It must however be acknowledged that although T2D is often accompanied by obesity, it is not a pre-requisite, and in some instances lean individuals are susceptible to developing IR and T2D (Tonks et al.,

2013).

So what is the pathology that underlies the hyperglycaemia in T2D? While T1Ds are unable to secrete insulin, the insulin deficiency observed in T2D is a relative loss.

Individuals with T2D essentially have a loss in glucose homeostasis as a result of beta cell exhaustion. The prediabetic state is defined by insulin resistance, warranting extra insulin secretion to maintain glycaemia. In susceptible individuals, this level of compensation cannot be sustained and insulin production fails, resulting in frank diabetes.

If T2D remains undiagnosed or mismanaged there is an increased risk of additional complications which include microvascular diseases such as renal failure or diabetic retinopathy resulting in visual impairment and blindness and macro-vascular diseases such as stroke, cardiovascular disease and nerve disease often leading to foot ulcers and lower limb amputation. T2D is clearly associated with the metabolic syndrome, which is characterised by a constellation of coexisting risk factors for cardiovascular disease

(CVD) such as hypertension, dyslipidaemia, glucose intolerance, obesity, and IR. In almost all manifestations of the metabolic syndrome obesity and IR are present (Cefalu,

2006). Given the increased risk for ill health and death, there is a very real need for a sustainable and affordable intervention (P. Z. Zimmet et al., 2014).

Prevention is better than a cure. In order to prevent the development of T2D, intervention must occur while an individual is insulin resistant. Insulin resistance is a highly complex disorder influenced, by diet, age and environment. Therefore the remainder of this chapter will provide an overview of our current understanding. The broad aim of this thesis is to investigate how IR develops over time in mice when exposed to high fat diets and what the long term consequences of this disease may be, even if early intervention is achieved.

1.3 Insulin and the maintenance of glucose homeostasis

Typically blood glucose concentrations are maintained within a relatively narrow range despite fluctuations in glucose uptake and disposal (Gerich, 2000). Inadequate glucose can be quite dangerous as it can affect proper function of the central nervous system, thus resulting in seizures, loss of consciousness, and even death (Shepherd &

Kahn, 1999). In contrast, prolonged increases in blood glucose can result in many

6 complications, including blindness, renal failure, cardiac and peripheral vascular disease, and neuropathy (Shepherd & Kahn, 1999).

The maintenance of glucose homeostasis involves balancing glucose supply from the intestine and output from the liver, with glucose uptake by the peripheral tissues such as skeletal muscle and white adipose tissue (WAT) (Saltiel & Kahn, 2001). To achieve this, the body uses multiple hormones such as glucagon and insulin in concert with one another to moderate energy stores and maintain glucose homeostasis.

Figure 1.1 - Insulin, Glucagon and glucose homeostasis. Schematic demonstrating the separate roles of insulin and glucagon in maintaining glucose homeostasis is the fed and fasted state, respectively. 7

Both insulin and glucagon are secreted from distinct specialised secretory cells that together (with a few other cell types) form the islets of Langerhans, the endocrine tissue of the pancreas. During fasting, α-cells in the islets of Langerhans secrete the hormone glucagon. Glucagon acts to raise blood glucose by stimulating gluconeogenesis or the breakdown of glycogen to glucose in the liver, with the aim of maintaining a constant energy supply to peripheral tissues (Fig 1.1).

In the fed state, insulin is the primary regulator of glucose homeostasis. After a meal, a rise in blood glucose is rapidly sensed by the β-cell (also contained in the islets of Langerhans), which responds by secreting an appropriate level of insulin. Insulin then orchestrates a series of events that together maintain glucose homeostasis (Fig 1.1).

This includes suppression of hepatic gluconeogenesis, thus converting the liver from a glucose producing to glucose storing organelle, and promotion of glucose uptake by insulin responsive tissues such as muscle and fat (Schinner, Scherbaum, Bornstein, &

Barthel, 2005). In addition, insulin promotes a number of other actions including increased glycogen synthesis and storage in liver and muscle, increased triglyceride synthesis and storage in liver and adipose tissue, and inhibition of lipolysis in adipose tissue(Saltiel & Kahn, 2001). With respect to protein metabolism, insulin has an anabolic effect on muscle and adipose tissues by stimulating amino acid uptake and protein synthesis and by inhibiting protein breakdown (Saltiel & Kahn, 2001). See fig

1.2 for a brief summary of insulins actions.

Although insulin initiates many processes and has many sites of action the main site of insulin action in terms of glucose homeostasis occurs in skeletal muscle, which is responsible for 75-85% of all insulin stimulated glucose disposal. In contrast, adipose tissue accounts for less than 10% of whole body glucose uptake (Klip & Paquet, 1990).

Despite the small contribution adipose tissue makes to overall glucose tolerance, mice with an adipose specific knockout (KO) of the insulin sensitive glucose transporter 4

(GLUT4) rapidly develop IR (Abel et al., 2001) and display impaired insulin action in muscle and liver. In contrast, mice with a muscle-specific KO of the insulin receptor have normal glucose tolerance (Bruning et al., 1998). These findings clearly indicate that while adipose tissue may not be the primary site of glucose uptake, its role in regulating whole body glucose homeostasis and IR may be much larger than indicated by its contribution to glucose disposal.

Figure 1.2 - The many actions of insulin. A summary of insulins actions in insulin-responsive metabolic tissues including; liver, muscle and white adipose tissue

1.4 Cellular insulin signalling

Insulin binding to its receptor at the cell surface of insulin responsive cells activates a complex and extensive signalling cascade. The activated insulin receptor, catalyses

Insulin receptor substrate (IRS) tyrosine phosphorylation leading to recruitment of downstream effectors. The most crucial effector for insulin dependent glucose transport in muscle and fat is phosphatidylinositol 3-kinase (PI3K)/Akt. Upon activation, Akt phosphorylates a number of substrates in particular TBC1D4/AS160 and orchestrates a complex metabolic program involving the translocation of the glucose transporter 9

GLUT4 (James, Strube, & Mueckler, 1989) to the plasma membrane leading to glucose uptake into target tissues, thus lowering concentrations of blood glucose.

1.5 Insulin Resistance

T2D usually begins with IR. IR occurs when the ability of insulin sensitive tissues to increase glucose uptake in response to insulin is diminished (Chisholm et al., 1997).

The body compensates for insulin resistance by producing additional amounts of insulin, thus resulting in hyperinsulinemia. When the body’s insulin production fails to keep up with the demand, hyperglycaemia will occur. Over time, hyperglycemia worsens and T2D develops (Saltiel, 2001).

IR is a major risk factor for the development of T2D (Shanik et al., 2008) and is a feature of the metabolic syndrome, which also includes other metabolic conditions such as abdominal obesity, increased blood pressure (hypertension), elevated triglycerides and reduced HDL-cholesterol. It is also a risk factor for other metabolic diseases such as

CVD, polycystic ovarian syndrome, non-alcoholic fatty liver disease (NAFLD), obesity, and also some cancers such as breast, colorectal, liver and pancreas (Moss & Alexander,

1990; Ray, Alalem, & Ray, 2014; Scalera & Tarantino, 2014). As previously mentioned, aging is also an independent risk factor for a number of these diseases.

Indeed, in a prospective study of 208 individuals, Fachinni and colleagues showed that

IR was a powerful predictor for a wide range of age-related diseases including T2D, stroke, CVD, hypertension and cancer.

Impaired glucose tolerance and impaired fasting glycaemia are intermediate conditions in the transition between normality and diabetes, although detecting people with impaired glucose is difficult as the β-cell often adapts to hyperglycaemic conditions by increasing insulin secretion and masking these effects. This adaptive 10 response that results in the upregulation in insulin secretion is termed compensatory hyperinsulinemia. This concept was demonstrated by Vaag and colleagues (Vaag,

Henriksen, & Beck-Nielsen, 1992) who showed that when given a glucose load, first degree relatives of NIDDM patients displayed completely normal glucose tolerance principally by compensatory hyperinsulinemia.

There is still much debate surrounding the exact mechanism by which IR occurs. At the heart of it is the inability to translocate GLUT4 storage vesicles from intracellular pools to the cell surface to facilitate glucose transport across the plasma membrane

(Galuska, Ryder, Kawano, Charron, & Zierath, 1998), which overall is considered the result of defective insulin signalling. Reduced expression or activity, at key points in the signalling cascade such as IRS-1 and Akt have previously been reported as a possible mechanism (Hotamisligil et al., 1996; Y. B. Kim, Nikoulina, Ciaraldi, Henry, & Kahn,

1999). However recent work by Hoehn and colleagues, has cast doubt on whether or not

IR requires a defect in insulin signalling (Hoehn et al., 2008). In addition, it has previously been observed that Akt is reduced during insulin resistance, and complete ablation of Akt blocks insulin-stimulated glucose uptake, but activation of just 5% of the total Akt pool was sufficient for a complete GLUT4 response indicating that a defects in Akt and insulin signalling may not play a role in IR (Ng, Ramm, Lopez, &

James, 2008). In Chapter 4 we will discuss briefly the disconnect between insulin signalling and IR.

Aside from signalling, there are a number of alternate mechanisms proposed as the cause of IR. The first involves the inflammatory cytokine tumour necrosis factorα

(TNFα), which is increased in the adipose tissue of obese rodents and humans

(Hotamisligil, Shargill, & Spiegelman, 1993). Indeed, obesity is often accompanied by a

11 chronic state of low grade inflammation, with elevated serum levels of inflammatory cytokines and macrophage infiltration in to adipose tissue (Hotamisligil et al., 1993).

While IR is often accompanied by inflammation, further examination indicates that IR usually precedes the onset of adipose tissue macrophage infiltration and inflammatory cytokines by several weeks (Xu et al., 2003). Another mechanism involves the increase in circulating free fatty acids (FFAs) derived from adipocytes such as diacylglycerol, fatty acyl-CoA and ceramides, which accumulate in muscle and liver and reduce insulin stimulated IRS-1 phosphorylation (Previs, Withers, Ren, White, & Shulman, 2000).

However, it has also been shown that super-physiological concentrations of FFAs have little effect on adipocyte insulin sensitivity (Hoehn et al., 2008), and call in to question this mechanism. Other mechanisms involve elevated glucocorticoids (Bujalska, Kumar,

& Stewart, 1997) and the production of reactive oxygen species (ROS) in response to nutrient oversupply (Hoehn et al., 2009), which reduce insulin sensitivity. Lastly, a link between diet induced-obesity and IR lies in the altered secretion of adipose derived proteins known as adipokines, which become dysregulated in obesity and consequently lead to insulin resistance (Rabe, Lehrke, Parhofer, & Broedl, 2008). We will discuss one of these, Leptin, in more detail below.

1.5.1 Therapeutics for insulin resistance

Apart from the administration of insulin by injection, conventional therapeutic agents for treating insulin resistance and T2D include metformin (modulator of hepatic glucose production), sulfonylureas (insulin secretagogues), thiazolidinedione (PPARγ agonists), analogues of the incretin hormone glucagon-like peptide 1 (GLP-1), α-glucosidase inhibitors, and dipeptidyl peptidase 4 inhibitors. Some of these agents produce side effects such as hypoglycaemia, oedema, gastrointestinal intolerance and weight gain, as well as having inadequate efficacy and short duration of action (Sheehan, 2003). 12

Treatments often become less effective with the progressive deterioration and loss of β- cell mass that occurs during the natural progression of T2D. There is therefore a crucial need for the development of new treatments through the identification of alternative drug targets as well as a better understanding of the complex molecular networks that control glucose homeostasis.

1.5.2 Mouse models of insulin resistance

1.5.2.1 Rationale for use of rodent models

The laboratory mouse (Mus Musculus) is considered one of the best model organisms for modelling human disease. Because it can have similar characteristics to humans, mouse models provide a unique opportunity to study the onset, development and course of obesity and IR, as well as an opportunity to investigate the underlying mechanisms.

The advantages of mouse models include the complete knowledge of the genome, ease of genetic manipulation, a shorter breeding span, and the availability of an array of physiological and invasive tests of relatively low cost (A. W. Lee & Cox, 2011)

(Schimenti & Bucan, 1998). Animal models therefore provide an opportunity to investigate the pathophysiology as well evaluate potential strategies for treatment and prevention of IR and related complications. For the animal model to have relevance it must mirror the developmental process, pathophysiology and etiology of IR and T2D

(Mansor et al., 2013). More specifically, the model should include a prolonged insulin resistance phase, obesity and in the case of T2D, ultimately β-cell dysfunction, (Accili,

2004; DeFronzo, 1997; C. R. Kahn, 2003a, 2003b; Kitamura et al., 2004; Polonsky,

Sturis, & Bell, 1996), attributes that would take many years to study in humans (Mansor et al., 2013).

Rodent models have been used widely for investigating the pathophysiology underlying IR. One of the major advantages is that the genetic variability in the human population can either be overcome by using genetically identical inbred mice, or mimicked using a series of outbred lines. Further, techniques to manipulate the mouse genome in a highly precise manner using gene-KO and transgenic technologies have improved dramatically. In the field of IR, technical advancements such as the miniaturisation of the clamp technique (Berglund et al., 2008) have made studying glucose homeostasis in metabolic diseases much more precise. As such, mice are increasingly being used to elucidate the mechanisms underlying diseases such as obesity, IR, T2D, CVD and cancer. However, no animal models are perfect and current models used to study IR and T2D are often associated with drawbacks in one form or another (Bugger & Abel, 2009). Although there is currently no one model that mirrors all the disease characteristics of IR (Bugger & Abel, 2009, 2014; Cefalu, 2006), there are a number of appropriate mouse models which can be used to interrogate components of the human disease (e.g. obesity induced IR or the insulin signalling pathway), including naturally occurring mutations, genetically engineered mice and the diet- induced obesity models (Berglund et al., 2008)

1.5.2.2 Genetic models of insulin resistance

The development of new mouse models and the increasing production of genetically modified mouse models have allowed researchers great insight in to the understanding of glucose metabolism, obesity and IR (Berglund et al., 2008), and these have been comprehensively reviewed (Cefalu, 2006; A. W. Lee & Cox, 2011; Neubauer &

Kulkarni, 2006) but are not the main focus of this thesis. In short, there are a number of spontaneous, monogenic and polygenic mutations such as the Lepob/ob and LepRdb/db, as

14 well as a number of transgenic and knockout lines used to recapitulate obesity, IR and interrogate insulin signalling.

The quintessential examples of a naturally occurring monogenic mutation are the leptin-deficient (Lepob/ob) (Ingalls, Dickie, & Snell, 1950) and leptin receptor-deficient

(Lepdb/db) mice (Chung, Power-Kehoe, Chua, Lee, & Leibel, 1996; Zhang et al., 1994).

On the C57BL6/J background, Lepob/ob mice develop IR, hyperinsulinemia, and fasting hyperglycaemia (Coleman & Hummel, 1973; Fellmann, Nascimento, Tibirica, &

Bousquet, 2013). By contrast, on the C57BL/KsJ background, Lepob/ob mice provoked moderate hyperphagia and obesity, which was accompanied by severe hyperglycemia and hyperinsulinemia followed by β-cell failure (Coleman & Hummel, 1967; Hummel,

Dickie, & Coleman, 1966). Similarly Lepdb/db mice on the C57BL/KsJ background develop diabetes, whereas on the C57BL/6J background they do not and only display IR

(Baetens et al., 1978; Hummel et al., 1966). There are a number of other monogenic and polygenic mutations which have been reviewed by Fellman et al, 2013 (Fellmann et al.,

2013) and others (Cefalu, 2006)

To dissect the complex genetics of obesity, IR researchers have generated transgenic and KO mice with mutations in genes required for either insulin action or insulin signalling (Bunner, Chandrasekera, & Barnard, 2014). Successful models that manifest more overt T2D include the insulin receptor KO (Accili et al., 1996), insulin receptor substrate-2 (IRS-2) KO (Withers et al., 1998), glucokinase KO (Grupe et al., 1995), and the insulin receptor in β-cell KO (βIRKO) (Leibiger et al., 2001). Models with less success at modulating glucose homeostasis and creating a model of IR & T2D include the insulin receptor in adipose tissue KO (FIRKO) (Bluher et al., 2002), insulin receptor in muscle KO (MIRKO) (Bruning et al., 1998), Insulin receptor substrate-3 (IRS-3) KO

(Liu, Wang, Lienhard, & Keller, 1999), Insulin receptor substrate-4 (IRS-4) (Fantin,

Wang, Lienhard, & Keller, 2000), Glut4 KO (Katz, Stenbit, Hatton, DePinho, &

Charron, 1995), Akt1 KO (W. S. Chen et al., 2001), PPARγ in β-cell (Rosen et al.,

2003). Several models of impaired glucose tolerance and insulin resistance have been developed. These include: the insulin receptor substrate-1 (IRS-1) KO, insulin receptor in liver (LIRKO) KO, Insulin receptor brain KO (NIRKO), IGF-1 receptor β-cell KO,

Akt2 KO, Glut4 KO in muscle and fat and the PPARγ KO in muscle and fat (A. W. Lee

& Cox, 2011; Neubauer & Kulkarni, 2006).

The use of animal models with naturally occurring mutations or silenced proteins in either a tissue specific or global manner have provided major insight into the roles of proteins in multiple pathways that can modulate glucose homeostasis across multiple tissues. Although much has been learned from the use of genetically modified mice, the exact mechanisms underlying IR remain elusive and effective treatments have not been forthcoming. One significant limitation is that these genetic manipulations do not reflect the polygenic nature of metabolic diseases, particularly, IR and T2D. Furthermore, it has been found that many of the genetic models used to recapitulate IR and T2D do not display the same islet pathology as humans (Cefalu, 2006) and their occurrence in humans are extremely rare and do not reflect current trends. Lastly the majority of human IR is driven by diets high in fat, resulting in obesity and hyperglycaemia in genetically susceptible individuals. Despite some of the obvious advantages of using a genetic mouse models; a significant limitation lies in that these do not demonstrate or replicate the etiology and pathophysiology of human IR. The inbred mouse models fed a HFD might provide a more physiologically and patho-physiologically relevant model as this better recapitulates the human condition, making it easier to extrapolate findings from mice to men.(Bunner et al., 2014; Cefalu, 2006; Neubauer & Kulkarni, 2006). 16

1.6 High-fat diet-induced obesity

Diet and dietary interventions are an extremely potent factor in the development of human health. The macronutrient concentrations mainly carbohydrate, protein and fat have been studied extensively to identify the most optimal combination for health and longevity (Solon-Biet et al., 2014). A healthy diet can help negate an array of diseases while a poor diet such as western diet, high in fats, can have deleterious effects. In predisposed individuals this will result in the development of insulin resistance, which is the basis for a number of diseases, particularly T2D. Dietary fat intake has often been claimed as responsible for the increase in adiposity (J. O. Hill, Melanson, & Wyatt,

2000; Jequier, 2002). Human studies have shown that HFDs (>30% energy from fat) can induce obesity (Hariri & Thibault, 2010), and obesity is a strong modifier of diabetes risk. The model of high-fat diet-induced obesity provides a viable alternative to genetic modifications for the study of IR and glucose metabolism. In almost all model species studied there is a positive relationship between whole body adiposity and the level of dietary fat consumed. This is true in monkeys, most mammals and rodents, and is likely the result of the ‘thrifty gene’ hypothesis, where the scarcity of food has selected for genes that promote the storage of lipid. These genes conferred a benefit historically in allowing the storage of fats for use in times of starvation and this same adaptation now confers a propensity to obesity in the context of caloric excess (Neel,

1999). Comprehensive reviews on the roles of fat in the establishment of obesity and IR have been published (Buettner, Scholmerich, & Bollheimer, 2007; Hariri & Thibault,

2010; West & York, 1998), and so here we will only focus on the diets relating to the studies performed within this thesis.

On the whole, when high-fat diets are given to rodents, significant weight gain through additional adiposity is observed. The first studies in mice (Fenton & Carr, 1951; 17

Fenton & Dowling, 1953; Lemonnier, Suquet, Aubert, De Gasquet, & Pequignot, 1975) used diets extremely high in fat, comprising approximately 70-80% of total energy from fat sources, a percentage which would now be deemed controversial as this does not reflect the diet consumed by humans. These studies produced mice with extreme obesity. Although there are in excess of 32,000 publications (PubMed search July 2015) using the keyword ‘high fat diet’, in murine models there is currently no consensus diet for either the HFD, low fat diets (LFD) or standard chow. Given the likely significance of relative small changes in the macronutrient composition of diet (Solon-Biet et al.,

2014), these discrepancies between studies may be significant. In general, control diets derive less than 10% of calories from fat, whereas LFD and HFDs derive approximately

30-50% from fat and in rare cases such as those mentioned above greater than 70% calories derived from fat (Fellmann et al., 2013). Additionally, the source of fat used in diets varies widely, from plant oils e.g. corn and safflower oil, to animal derived fats such as lard and beef tallow. The type of fat is important in the development of metabolic disorders, from here in references to HFD will refer to diets containing animal derived fats.

Animal fat is more efficient than vegetable fat in inducing metabolic disorders such as obesity, insulin resistance or glucose intolerance because it contains more saturated fatty acids (Buettner et al., 2006; Fellmann et al., 2013; Wang, Storlien, & Huang,

2002). Addition of sucrose aggravates the metabolic consequences of enriched diets and induces severe dyslipidaemia (Chicco et al., 2003; Surwit et al., 1995). The model of high fat feeding in C57BL/6 mice was first described in 1988 (Surwit, Kuhn, Cochrane,

McCubbin, & Feinglos, 1988) where around 58% of energy was derived from saturated fat, in the form of vegetable oils (coconut and soybean). Coconut oils are high in medium chain fatty acids (MCFAs). Diets rich in MCFAs result in less weight gain 18 and adiposity than diets rich in long chain fatty acids. In contrast, animal fat (lard) based diets contain virtually no (<05%) MCFAs and predominantly consist of

LCFAs. This is of particular relevance as the most common fatty acids consumed in a typical ‘western’ style diet are LCFAs, making a lard based diet more appropriate for mimicking the etiology and pathophysiology of human obesity and IR (De

Vogel-van den Bosch et al., 2011; Omar, Pacini, & Ahren, 2012). Regardless of these shortcomings, the Surwit paper laid the foundation for the use of the high-fat diet- induced obesity model in C57BL/6 mice for the investigation of IR in mice, although variations of the initial design have different amounts of fat (58% vs 45%) and different sources (vegetable oil vs lard) (Hoehn et al., 2009; Hoehn et al., 2010; Montgomery et al., 2013; Turner et al., 2009; N. Turner et al., 2013; Wu, Meoli, et al., 2014). The lard based HFD, results in obesity, hyperinsulinemia and altered glucose homeostasis due to insufficient compensation by the islets (Winzell & Ahren, 2004). Since obesity is induced by environmental manipulation rather than genes, it is thought to model the human situation more accurately than genetic models of obesity induced diabetes (King,

2012). Moreover, mice on a 45% lard develop a degree of IR that is proportional to the length of time on diet, whereas mice on a 58% fat diet develop a rapid deterioration in glucose tolerance, with the onset of IR in all peripheral tissues developing simultaneously. A review by Buettner and colleagues (Buettner et al., 2007; Hariri &

Thibault, 2010) of HFD studies between 1997-2007 determined the most optimal method to induce obesity was to use a semi purified high fat diet deriving approximately 40% of calories from animal fat (such as lard and beef tallow), a low amount of n-3 fatty acids and a low amount of plant oils rich in n-6 and n-9 fatty acids such as those found in safflower oil. The diet used in this thesis is consistent with these guidelines. As there is currently no consensus on the percent of fat used in HFDs a

19 paper in 2012 has since tried to establish minimum guidelines in the field of drug testing using DIO based on HFD (Bagnol, Al-Shamma, Behan, Whelan, & Grottick, 2012).

The diet recommended for C57BL6 mice identifies the diet outlined by Buettner as the most optimal diet.

1.6.1 Experimental design considerations

1.6.1.1 Strain

Any rodent strain given an appropriate diet will develop metabolic changes. Indeed almost all mouse strains, with the acception of the A/J strain (Surwit et al., 1995), when placed on a high-fat diet develop obesity. However, many develop varying degrees of glucose intolerance, insulin resistance, insulin secretion profiles and dyslipidaemia

(Fellmann et al., 2013). We have already discussed the dietary composition but it should be noted that the background strain of the mice can determine the susceptibility to diet- induced metabolic changes, and thus, effects could be missed if a more resistant strain is used (Almind & Kahn, 2004; Bachmanov, Reed, Tordoff, Price, & Beauchamp, 2001;

Surwit et al., 1995). Several studies have compared the appropriateness of the C57BL/6 line to other lines as they have previously been described as an obesity prone strain

(Black et al., 1998). In all cases, the C57BL/6 remained the most optimal and robust strain (Andrikopoulos et al., 2005; Berglund et al., 2008; Montgomery et al., 2013).

Coincidentally the C57BL/6 mouse strain is the most commonly used strain in the field of IR and is generally considered the best strain for most metabolic studies (Neubauer &

Kulkarni, 2006). It should be noted though that there is some heterogeneity in the response of C57BL6 mice indicating that differential responses are not purely driven by genetic background (Burcelin, Crivelli, Dacosta, Roy-Tirelli, & Thorens, 2002).

1.6.1.2 Age and duration of diets

There is also a clear relationship between when an obesity producing diet is initiated, the duration of the diet, and the degree of body fat accumulated. The earlier a feeding regime begins the greater the final body fat content (Peckham & Entenman, 1962), and the longer the duration the greater the increment (West & York, 1998). However is this relationship linear and does it plateau? Studies show that diets have a more gradual and continuous effect so exposure to diet will be an important variable. In one study AKR/J mice were fed 15%, 30% and 45% HFD and showed a linear relationship between adiposity and dietary fat content (West & York, 1998). Overall the results suggest that the relation between the amount of dietary fat consumed and body fat content could be linear, with no threshold effect. Therefore, the duration of exposure is an important variable to consider when inducing insulin resistance. Age should also be considered as alterations in diet too early may negatively impact an animal’s natural development, while manipulations given at later stages may compete with the effects of aging. These variables likely have significant implications in the field of IR research.

Relatively few of the many HFD studies in rats and mice have used longitudinal feeding regimes combined with in vivo measurements (e.g. radiolabelled

GTT/hyperinsulinemic-euglycemic clamp) to evaluate whole body insulin sensitivity and to identify the tissue sites where IR first appears. Many of the current studies take static measurements, which are therefore unable to delineate the order with which tissues develop IR. To parallel human obesity and IR, the prolonged, progressive and temporal nature of the disease must be reflected in the model. Moreover, because the induction of IR following a HFD has been reported to be as early as 1 wk, most studies investigating obesity, IR and glucose metabolism feed mice for a relatively short period

(1-12 wks) (B. Ahren & Pacini, 2002; Y. S. Lee et al., 2011; Surwit, Seldin, Kuhn, 21

Cochrane, & Feinglos, 1991; Winzell & Ahren, 2004). A consequence of these endpoint studies is that it is difficult to determine whether the changes in tissue responses are a cause or a consequence of other ongoing metabolic changes, which alter whole body insulin sensitivity. In addition, the relatively short nature of these studies does not reflect the prolonged periods of IR observed in humans.

To date the majority of long term HFD feeding studies have been done in rats

(Gomez-Perez, Capllonch-Amer, Gianotti, Llado, & Proenza, 2012; Heinrich,

Andersen, Cleasby, & Lawson, 2015; Madsen et al., 2010; Nadal-Casellas, Amengual-

Cladera, Proenza, Llado, & Gianotti, 2010; Tofolo et al., 2014). For reasons outlined earlier, the C57BL6 mouse is increasingly becoming the model of choice for metabolic studies. Thus far there have been relatively few studies that have examined metabolism and health in response to a HFD longitudinally over long periods of time. The most comprehensive study examining metabolism longitudinally was performed by Turner and colleagues in 2013, who phenotyped the metabolism of mice at 1, 3,6 and 16 wks of

HFD feeding using a hyperinsulinemic-euglycemic clamp (N. Turner et al., 2013), while a GTT was used at several other intermediate time points. Consistent with previous studies (Commerford et al., 2001; Pagliassotti et al., 1997) they observed the rapid induction of hepatic IR (1 wk) but extended this by showing that IR next developed in adipose tissue (1wk) and muscle (3 wks). Although hyperinsulinemic- euglycemic clamps are technically and logistically challenging, one shortcoming of this study was that it lacked sufficient intermediate time points to allow delineation of liver and adipose tissue IR induction, and could have been improved by performing more clamps at more time points. In addition while an endpoint of 16 wks is sufficient to study the short term effects of IR on metabolism it does not reflect the progression of human IR. This model would have been greatly improved by extending the study out to 22

>12 months, thus allowing the comparison of diet with the effects of age. Other studies have attempted to model the temporal pathophysiology of human IR by feeding mice for much longer periods of time (2-12 months) but have lacked the temporal resolution to sufficiently characterise the metabolism of mice as they progress on a HFD, or did not examine metabolism choosing to characterise a tissue of interest (e.g. heart, brain, pancreas) either longitudinally or only at the end of the feeding regime (McGillicuddy et al., 2013; Mosser et al., 2015; Park et al., 2005; Toyama et al., 2015; Winzell &

Ahren, 2004).

The consistent problem we have identified throughout the literature of all reported high-fat diet-induced obesity studies is a lack of standardised diets and feeding regimes.

The combination of altered fat sources, as well as length of exposure and age of animals tested complicates the interpretation of data produced from what are collectively termed

HFD studies. There is a real need for a comprehensive study examining the acute and prolonged metabolic changes associated with HFD feeding.

1.7 Whole body metabolic effects of high-fat diet

1.7.1 Glucose, insulin and glucose tolerance

The effects of HFD feeding on blood glucose levels are often discrepant

(Andrikopoulos et al., 2005; Burcelin et al., 2002; Hoehn et al., 2009; Montgomery et al., 2013; Wu, Meoli, et al., 2014). Normoglycaemia, slight hyperglycaemia, and the development of type 2 diabetes have been reported with different diet regimes in

C57BL/6 mice (Winzell & Ahren, 2004), although the classification of T2D in mice remains controversial. Although peripheral IR is one of the hallmarks of T2D, development to T2D requires β-cell dysfunction. Β-cell failure in HFD fed mice is rarely reported, and in one study it was noted that C57BL/6 mice do not develop β-cell

23 failure/mass loss per se, but develop a secretory dysfunction as a result of altered

ATP/Ca2+ and lipid signalling (Peyot et al., 2010). In many studies, diabetes is established in C57BL/6 mice using drugs such as Streptozotocin, which destroy the insulin producing β-cells (Leiter, 1982; Yin et al., 2006).

From the data published so far, one can conclude that prolonged, i.e., several weeks

HFD feeding (1-6 wks), will eventually lead to moderate hyperglycaemia in most strains of mice and rats (particularly the C57BL/6(J/n) and the Wistar/Sprague- Dawley rat) (Buettner et al., 2006). When lard based HFDs are used, the elevation in fasting glucose levels is often accompanied by a moderate to distinct increase in fasting plasma insulin levels (Buettner et al., 2007). The increase in glucose and insulin indicates peripheral IR, and when subjected to a GTT a delay in glucose clearance is observed.

Similarly, when HFD fed mice are subjected to a hyperinsulinemic-euglycemic clamp a reduction in glucose uptake and a reduction in the glucose-infusion rate (GIR) is often observed, again indicating peripheral IR.

Lard based HFD studies often span a maximum time of 1-16 weeks, however the induction of IR has been shown to manifest in as little as 1wk on a lard-based HFD, indicating the rapid dietary induction of this disorder (N. Turner et al., 2013; Winzell &

Ahren, 2004). Since IR occurs so rapidly it is questionable whether studies performed longer than 7 days (d) are studying insulin resistance per se, or other consequences of obesity such as inflammation or an adaptation to IR. Moreover, while this model provides a viable avenue for the investigation of acute IR, it does not entirely mirror the pathophysiology of human IR. In contrast to the HFD fed mouse, when IR humans are subjected to a GTT they appear as normoglycaemic. This is because the measurements are often taken many years after IR is established and compensatory hyperinsulinemia

24 has set in (Vaag et al., 1992). These findings indicate that a more prolonged feeding regime is required in mice to better mirror the pathophysiology of the human condition.

1.7.2 Altered Adipokine expression

Once thought of as an inactive storage compartment, adipose tissue is now recognised as an active endocrine organ (Rosen & Spiegelman, 2014; Scherer, 2006), secreting a variety of bioactive proteins collectively termed adipokines, capable of regulating a wide range of metabolic processes including insulin action, around the body, and locally via a paracrine effect (Galic, Oakhill, & Steinberg, 2010). The first adipokine to be identified was adipsin (complement factor D) (Cook et al., 1987), and although levels were altered in genetic models of obesity (Flier, Cook, Usher, &

Spiegelman, 1987); there was little evidence to suggest that adipsin plays an important endocrine role in fat. Subsequent experiments in Lepob/ob and Lepdb/db revealed that a secreted factor was able to regulate food intake (Coleman, 1973; Coleman & Hummel,

1969). It was later discovered that this factor was the Ob gene encoding leptin (Zhang et al., 1994) and secreted from WAT. It has also been shown that leptin has a profound effect on whole body energy metabolism, by increasing energy expenditure and reducing food intake, thus exposing the adipocyte as an endocrine organ. Since then many other adipokines have been discovered, including apelin, FGF21, adiponectin, resistin, nesfastin-1, eNAMPT/Visfatin, RBP4 and FABP4/aP2 (Antuna-Puente, Feve,

Fellahi, & Bastard, 2008; Cantley, 2014; Wu, Samocha-Bonet, et al., 2014; Q. Yang et al., 2005). Other factors include plasminogen activator inhibitor 1 (PAI-1), lipocalin-2, chemerin, interleukin-6, and tumour necrosis factor alpha (TNFα) (Alessi, Poggi, &

Juhan-Vague, 2007; Galic et al., 2010) Although these are secreted from adipose tissue, they may not necessarily be secreted from adipocytes per se, as adipose tissue contains multiple cell types. These cell types fall under what is termed the stromavascular 25 fraction and include pre-adipocytes, endothelial cells and macrophages, and it has been suggested that these cells and not adipose tissue are responsible for as much as 90% of the protein species secreted (Fain, Madan, Hiler, Cheema, & Bahouth, 2004).

HFD feeding in mice results in the expansion of adipose tissue and the secretion profile of many adipokines mirrors human obesity (Buettner et al., 2007). For example, both leptin and resistin levels tend to be elevated, consistent with an increase in adipose tissue and obesity, while adiponectin levels are down-regulated in dietary obese rodents

(Hariri & Thibault, 2010). Both leptin and adiponectin have been categorised as ‘anti- diabetogenic’ based on their capacity to stimulate β-oxidation, decrease triglyceride synthesis, and enhance insulin action in skeletal muscle and liver. However, caution must be taken when interpreting the results in mice due to the inconsistent reporting of the source of fats, which has been shown to effect the level and speed of adiposity accrued by mice, and this influences the secretory profile of adipokines (Buettner et al.,

2007). In saying this, it can still be concluded that HFD feeding of mice with animal/plant based diets (e.g. lard, tallow and safflower oil) for several weeks will alter the adipokine levels and likely resemble human obesity and IR in susceptible mice

(Buettner et al., 2007). A more in-depth examination of leptin’s role in obesity and insulin secretion will be provided later in this introduction.

1.7.3 Ectopic lipid deposition

Ectopic lipid deposition is defined as excess lipid in the form of triglycerides deposited in non-adipose tissues, mainly liver and skeletal muscle, which usually only contain small amounts of fat (Pagliassotti, Prach, Koppenhafer, & Pan, 1996; Rasouli,

Molavi, Elbein, & Kern, 2007). A growing view is that the accumulation of lipid in tissues not designed for its storage, results in cell dysfunction (lipotoxicity) and lipid

26 induced programmed cell death (lipoapoptosis) (Shoelson, Herrero, & Naaz, 2007) and is a key initial step in the development of IR, which is further associated with the development of T2D, NAFLD, and CVD (Lettner & Roden, 2008).

Human lipodystrophy provides the best example of the metabolic consequences that arise when an individual has inadequate adipose tissue and is unable to partition lipid in to adipose tissue. These individuals have a partial or complete absence of adipose tissue, and are characterised by IR, hepatic steatosis, and dyslipidaemia, with the severity of each inversely related to the level of adiposity (Schott, Scherbaum, & Bornstein, 2004).

Similarly, chronic HFD feeding studies in rats and mice result in hypertriglyceridemia, resulting in an oversupply of lipid to metabolic organs, that is temporally related to weight gain, and is associated with impaired insulin sensitivity and aberrant glucose homeostasis. There are two possible mechanisms whereby ectopic lipid storage may occur. The first is the failure of adipose cells to adequately proliferate, resulting in an increase in circulating free fatty acids (the ‘push’ factor). The cornerstone of this relies on the notion that the adipose depot has reached its limit and it can no longer efficiently store lipid. In mice this is somewhat controversial as IR in tissues such as liver have been shown to be established within 1 wk (N. Turner et al., 2013), but adipose expansion can continue for several weeks or months after the initiation of a HFD. The second is impairment in non-adipose organs to increase fat oxidation to match dietary fat intake (the ‘pull’ factor). The oversupply of lipids to mitochondria produces a mismatch between TCA cycle flux and fat oxidation that results in the depletion of organic acid intermediates, incomplete fatty acid oxidation and accumulation of lipotoxic short chain fatty acylcarnitines. The accumulation of these lipotoxic elements results in defective insulin signalling, and is associated with skeletal muscle IR in rats

(Muoio, Dohm, Fiedorek, Tapscott, & Coleman, 1997). This may occur as a 27 consequence of obesity, or result from alterations in the secretion of adipokines such as leptin, and adiponectin, which increase lipid oxidation (P. Huypens, 2007; Yamauchi et al., 2001). Lastly, in C57BL6 HFD fed mice, chronic systemic inflammation was shown to exacerbate lipolysis in non-adipose tissue, resulting in increased ectopic lipid deposition, which was again linked to IR (Mei et al., 2011).

1.8 The metabolic effects of HFD feeding by tissue

1.8.1 Liver

It is well established that the high-fat obesity-inducing HFD based on animal lard, leads to hepatic steatosis (Yaqoob et al., 1995) and is associate with NAFLD

(Nakamura & Terauchi, 2013). Hyperinsulinemic-euglycemic clamp studies have shown that this condition is associated in vivo with hepatic IR, i.e., an impairment of insulin’s ability to lower hepatic glucose output (Hashemi Kani, Alavian,

Haghighatdoost, & Azadbakht, 2014). Temporal analysis of IR in response to HFD indicates that the liver may be the first site of reduced insulin sensitivity (N. Turner et al., 2013) and this is followed by muscle IR (Commerford et al., 2001; Kraegen et al.,

1991; Pagliassotti et al., 1997). The liver appears to play a unique role in whole body insulin sensitivity. Drugs such as metformin, activate 5’AMP-activated protein kinase

(AMPK) stimulating fatty acid oxidation. This effect is abolished by deletion of LKB1, the upstream activator of AMPK (Shaw et al., 2005). In a series of other studies it would appear that the liver can act as a catabolic sink that defends against hyperlipidaemia (An et al., 2004; Nagle et al., 2007). This leads us to the theory of lipid-induced hepatic IR where lipid species accumulate in the liver as a result of impaired fatty acid oxidation. This results in the misdirection of long-chain acyl CoAs

(LC-CoAs) into ER-localised and cytosolic lipid species, such as diacylglycerols

(DAGs), ceramides and triglycerides. Indeed, HFD feeding in mice leads to accumulation of LC-CoAs, DAGs and ceramides resulting in IR (Chavez, Holland, Bar,

Sandhoff, & Summers, 2005; Chavez & Summers, 2003; Griffin et al., 1999).

Furthermore, suppression of the mitochondrial glycerol 3-phosphate acyltransferase-1

(GPAT1 – the first step in triglyceride synthesis) or acetyl CoA carboxylase-2 (ACC2) activity results in increased fatty acid oxidation, lowered DAG levels and reversal of IR in the liver (Abu-Elheiga, Oh, Kordari, & Wakil, 2003; Neschen et al., 2005; Savage et al., 2006). However, hepatic overexpression of Diacylglycerol acyltransferase 2

(DGAT2) (the final step in conversion of triglycerides from DAGs) in mice, resulted in an increase in DAG content but no change in hepatic insulin sensitivity (Monetti et al.,

2007). These data show conflicting roles for the importance of DAGs in the establishment of hepatic IR.

Interventions designed to improve hepatic fat clearance, e.g., the hepatic overexpression of uncoupling protein-1, have demonstrated improved whole-body insulin sensitivity (Ishigaki et al., 2005). However, insulin signalling studies in HFD fed rats have shown that the classic signalling phenotypes observed in muscle and fat are not necessarily replicated in the liver. IRS-1 and IRS-2 proteins and their phosphorylation are not altered, and PI3K activity associated with IRS-1 and IRS-2 is increased (Anai et al., 1999). Together these studies exemplify the importance of the liver in maintaining insulin sensitivity and the development of IR; and in conclusion,

HFDs can induce hepatic steatosis and symptoms of hepatic IR in the whole animal, and this closely resembles the human obese state.

1.8.2 Skeletal muscle

A hallmark of HFD-induced IR is an impairment of insulin-stimulated glucose uptake in skeletal muscle in the range of 30% to 60% of standard-fed controls. This effect was previously reported as early as 14 d but has since been shown to occur after 7 d in the C57BL/6 HFD fed mouse (N. Turner et al., 2013; Youngren, Paik, & Barnard,

2001), with the same fat types that induce whole body IR. The mechanism underlying the pathophysiology of this effect has been examined closely, but still remains controversial. Excess lipids, altered mitochondrial metabolism, endoplasmic reticulum stress responses, inflammation and the generation of ROS have all been implicated in the development of IR.

Early work demonstrated a reduction in the total amount of insulin receptor without modification of the receptor affinity (Grundleger & Thenen, 1982). Subsequent studies have found a decline in insulin receptor autophosphorylation and insulin receptor substrate-1 (IRS-1) phosphorylation (Taouis et al., 2002; Youngren et al., 2001) and a reduced activation of IRS-1- associated PI3K (Zierath, Houseknecht, Gnudi, & Kahn,

1997). These results suggest that impairment in the early steps of insulin signalling in

HFD animals leads to an alteration in the translocation of the glucose transporter

GLUT4 (Hansen et al., 1998; Zierath et al., 1997) and thereby, causes the impairment of insulin-stimulated glucose uptake in muscle. Fat-enriched diets have been reported to change the muscle cell membrane phospholipid composition, which might influence insulin binding or GLUT4 function (Pan & Storlien, 1993).

Other studies have shown that the accumulation of intramyocellular lipids (IMCLs) is associated with mitochondrial dysfunction and reduced insulin action (Buettner,

Newgard, Rhodes, & O'Doherty, 2000; Storlien et al., 1991). As in the liver,

30 intramuscular levels of lipid such as LC-CoAs, DAGs and ceramides, positively correlate with triglyceride content and negatively with insulin sensitivity (Holland et al.,

2007; Hulver et al., 2003; Petersen et al., 2005). For example, the level of ceramides was increased in skeletal muscle of obese IR patients (Adams et al., 2004). Another mechanism involves the mismatch between β-oxidation and TCA cycle flux. During

HFD feeding, fatty acid influx (e.g. branch chain amino acids – BCAA) and PPARα/δ- mediated activation of target genes promote β-oxidation without a co-ordinated increase in TCA-cycle flux. This leads to the production of metabolic by-products from incomplete oxidation (such as acylcarnitines and ROS), which act to ‘fill’ the mitochondria, a process known as anaplerosis (Newgard, 2012). Under these conditions, glucose is rendered superfluous as an energy substrate, resulting in decreased glucose utilization and glucose intolerance. It is proposed that stresses such as these will alter serine kinases that block insulin stimulated GLUT4 translocation (An et al., 2004;

Koves et al., 2005; Muoio & Newgard, 2006). The altered flux through the catabolic pathways in muscle provides a mechanistic link between the consumption of a HFD and

IR. In summary, HFDs decrease the insulin sensitivity of skeletal muscle by altering early steps of the insulin signal transduction pathway. This closely mirrors the pathogenesis of human obesity-related IR.

1.8.3 Adipose tissue

Adipose tissue is primarily made of adipocytes and acts as a caloric reservoir that can expand in response to excess nutrient availability and release lipids in a negative energy state. The adipocyte is uniquely equipped with the ability to store excess calories in the form of triglycerides without the negative effects of lipotoxicity that are observed in tissues not designed to store lipid such as liver and muscle (Konige, Wang, & Sztalryd,

2014). The use of transgenic mice has revealed that fat storage in adipocytes is central 31 to whole body IR. Animals overexpressing diacylglycerol acyltransferase in adipose tissue, the rate-limiting enzyme of triglyceride synthesis, become obese under an HFD but are resistant to diet-induced IR as examined by glucose and insulin tolerance tests

(H. C. Chen, Stone, Zhou, Buhman, & Farese, 2002).

As discussed earlier, adipocytes also act as an endocrine cell, producing hormones that can act far beyond metabolism. For example, leptin can regulate food intake via the hypothalamus and adiponectin can improve insulin sensitivity via its actions in muscle and liver. Adipokine expression also correlates with a number of pathological states including cancer and osteoporosis (Bluher, 2014)

In both humans and mice, the consequence of over-nutrition is obesity. In response to a HFD the adipose tissue undergoes profound morphological changes. While consensus has not been reached concerning the phenotype established, overall, animal fat derived diets result in an increase in adipocyte size and number (Hausman, Loh, Flatt, & Martin,

1997).

Both brown and WAT from mice fed a HFD are insulin resistant as determined by reduced insulin-stimulated glucose uptake into adipocytes. A number of hypotheses have been put forward to explain the molecular mechanisms underlying adipose IR, such as decreases in insulin receptor auto-phosphorylation (Watarai et al., 1988) but the exact signalling defect that causes IR in this tissue remains unknown. Recent work has shown that the defect lies below the level of IRS-1 (Hoehn et al., 2008) and current studies are searching for novel regulators of GLUT4 translocation such as TUSC5

(Fazakerley et al., 2015). Furthermore, the absence of the downstream effects of insulin signalling permits unregulated lipolysis, leading to ectopic lipid deposition.

Dysregulated adipocyte lipolysis has been shown to be the primary source of non- 32 esterified fatty acids (NEFAs) promoting disease in liver and muscle. Recent studies have shown that IR can be circumvented by either increasing the lipid droplet size and upregulating adipocyte triglyceride storage via overexpression of phosphoenolpyruvate carboxykinase (PEPCK), which ultimately reduces NEFA release and systemic IR

(Forest et al., 2003; Kusminski et al., 2012). Although in this context an expanding adipocyte is advantageous, the expansion is ultimately also associated with ER stress, the unfolded protein response, ROS, increased lipid spillover and ultimately IR (Koves et al., 2008; Kusminski & Scherer, 2012).

Taken together, HFD feeding results in adipocyte hypertrophy, hyperplasia and IR. The adipocyte is vital to whole body insulin sensitivity. This is likely governed by its role in fat storage and communication with the rest of the body via adipokines.

1.8.3.1 Adipocyte hypoxia and angiogenesis

Adipocytes require the ability to rapidly contract and expand in response to nutrient availability. As adipose tissue expands, oxygen tension decreases and this has been observed in both obese rodents and humans (Ye, Gao, Yin, & He, 2007) (Pasarica et al.,

2009). Hypoxia-inducible factor 1-α (HIF1α) is an oxygen sensing transcription factor that is rapidly degraded under normoxic conditions, but stable under hypoxic conditions. It is increased in expanding adipose tissue due to the increased need for tissue vasculature (Halberg et al., 2009; Regazzetti, Bost, Le Marchand-Brustel, Tanti,

& Giorgetti-Peraldi, 2010). The downstream targets of HIF1α should improve vascularisation acting to improve oxygen availability however tissue vascularisation has a complicated interaction with adipocyte health that is both supportive and inhibitory depending on the physiological state (Sun et al., 2012). Furthermore, direct inhibition of

HIF1α has been shown to improve adipose tissue health (Sun, Halberg, Khan,

Magalang, & Scherer, 2013) and deletion of HIF1α limited high-fat diet-induced adipose tissue inflammation and IR (Le Lay, Blouin, Hajduch, & Dugail, 2009).

Although the role of HIF1α in regulating angiogenesis is under question (Halberg et al.,

2009) together these studies suggest that in order to avoid the ill effects of hypoxia, the expansion of the adipose tissue must be inhibited, and interfering with angiogenic pathways may provide the means to achieve this. Indeed inhibition of adipose tissue angiogenesis using anti-angiogenic compounds, has been shown to result in a reduction in adipose tissue mass, and improve insulin sensitivity in HFD fed mice (Brakenhielm et al., 2004; Kolonin, Saha, Chan, Pasqualini, & Arap, 2004; Rupnick et al., 2002), although caution must be taken in interpreting these data as the agents used here also caused a reduction in food intake. However, in the context of a shrinking adipose tissue depot, a reduction in food intake protects animals from ectopic storage of lipid in muscle and liver, avoiding the detrimental metabolic consequences that occur during over nutrition. Given the partial success of the current anti-angiogenic compounds used to date, further research in to alternatives with less off target effects is warranted. In this thesis we will present data on the use of anti-VEGF (vascular endothelial growth factor) antibody, which is known to stimulate angiogenesis and restore oxygen supply to tissues with inadequate blood circulation (Wu, Meoli, et al., 2014).

1.8.4 Pancreas / β-cell

Obesity is associated with IR. The increased requirement for insulin is sensed by the β- cells causing an increase in insulin secretion. If the hyperinsulinemia is adequate, normoglycaemia is restored. However, if the signals to stimulate insulin secretion fail, the hyperinsulinemia will be inadequate and result in glucose intolerance or diabetes.

Therefore, knowledge of the signals between IR and insulin secretion is of importance for the understanding of diabetes pathogenesis. 34

Beta cells are remarkably plastic in their ability to regulate insulin release, but at the same time do so in a very precise manner. The quantity of insulin released varies according to the type of stimulus provided and the prevailing glucose levels. Insulin sensitivity also modulates beta cell function and is almost always decreased in obese individuals. In healthy individuals there is a feedback loop between the demand from liver, muscle and adipose tissue to the beta cell. The ability of the beta cell to adapt to a changing environment relies on two parameters: the functional responsiveness of the cell and β-cell mass. The beta cells response to changes in insulin sensitivity, relies on an integration of signals including cellular glucose metabolism, NEFA signalling, citrate levels, sympathetic and parasympathetic innervation and sensitivity to incretins

(e.g. GLP-1 and GIP) and adipokines. Increased citrate levels share some homology with what we have discussed thus far in liver and muscle (S. E. Kahn, Hull, &

Utzschneider, 2006).

When given a HFD, C57BL/6 mice develop obesity, hyperglycaemia, hyperinsulinemia and this is accompanied by expansion of β-cell mass (Hull et al., 2005) and an impaired glucose-stimulated insulin secretion (GSIS) despite the increase in β-cell mass (Hull et al., 2005; Stamateris, Sharma, Hollern, & Alonso, 2013). As expected, the spectrum of these results can vary greatly due to the numerous types of HFDs, strains and macronutrient and micronutrient combinations used. In a series of short (8-12 wks)

(Ebato et al., 2008; Peyot et al., 2010) and long term (1 year) (J. Ahren, Ahren, &

Wierup, 2010; Hull et al., 2005) HFD feeding studies it was determined that the proliferative expansion of β-cells correlates with body weight gain and the degree of IR, suggesting that β-cell hyperplasia is driven by a complex milieu of obesity and IR and not a derivative of high amounts of dietary fat per se (Steil et al., 2001). Consistent with this, acute infusion of lipids did not increase β-cell replication in mice, and elevated 35

FFAs actually appear to inhibit glucose-stimulated β-cell proliferation (Pascoe et al.,

2012). Furthermore, lipotoxicity (i.e. the detrimental effects of FFAs on the β-cell) results in an impaired first phase insulin secretory response indicating impaired β-cell compensation (B. Ahren & Pacini, 2002). In contrast, serial sectioning of pancreata from HFD fed rats (42% fat from lard) resulted in an increase in both islet and β-cell mass (Lingohr, Buettner, & Rhodes, 2002). However, studies examining lipotoxicity in the islets of HFD fed mice/rats are limited by the fact that animals developed a concomitant hyperglycaemia due to the dietary intervention (glucolipotoxicity instead of lipotoxicity). The β-cell is uniquely designed to sense small changes in physiologic glucose concentrations, and is considered a mediator of β-cell hyperplasia (Linnemann,

Baan, & Davis, 2014). Together these data suggest that a combination of hyperglycaemia and dyslipidaemia leads to a reduction in β-cell replication.

In a recent study examining the effects of HFD feeding on β-cell mass in C57BL/6J mice, Mosser and colleagues question whether β-cell proliferation is the result of HFD feeding (60% calories from fat) or IR (Mosser et al., 2015). Stating that both short (1 wk) and long term (2-12 months) HFD feeding studies result in glucose intolerance, IR, and enhanced β-cell mass due to increased β-cell proliferation, but lack the temporal resolution and study design consistency to clearly establish what drove the increases in

β-cell mass. This is further complicated by the differences in composition and timing of diet, mouse genotypes, age, and sex used to study β-cell proliferation. (B. Ahren &

Pacini, 2002; Berglund et al., 2008; Granger & Kushner, 2009; Posey et al., 2009; N.

Turner et al., 2013). They identified one study in particular that showed that β-cell proliferation occurred in response to 1 wk of HFD feeding and this was believed to have occurred in the absence of IR (Stamateris et al., 2013). In response to this they designed a longitudinal feeding regime spanning 1-11 wks with measurements taken at 1, 3, 5, 7 36 and 11 wks, concluding that HFD-induced β-cell proliferation occurs prior to IR. This was based on the observation that β-cell Ki67 staining was increased after 3 d, however

Mosser and colleagues only measured indicators of IR after 1 wk, and although they observed no change in the GIR from clamp studies at 1 wk, mice did display glucose intolerance via a GTT at the same time, which would almost certainly indicate hepatic

IR at the least.

Taken together, HFD animals seem to be a useful model to validate the physiological relevance of β cellular lipotoxicity, but as such, they have been mostly neglected. Until now, there are a few studies with rodents that, indeed, indicate that HFDs might induce a milieu of lipotoxicity and directly (i.e., independent from concomitant insulin resistance) impair the β cellular insulin metabolism.

1.8.4.1 Leptin and Leptin resistance

Leptin is secreted in proportion to the degree of obesity (Maffei et al., 1995) and is strongly correlated with insulin sensitivity. Leptin levels were identical in a study comparing T2D human subjects to non-diabetics of a similar BMI demonstrating that leptin levels are proportional to adiposity, irrespective of T2D (Hotta et al., 2000). In this context, adipokines provide a direct read out of the degree of adiposity and metabolic status of the adipose tissue. Leptin has multiple actions, playing roles in the endocrine and immune system, fertility, bone formation, tissue remodelling, inflammation, and glucose homeostasis via its effects on insulin and glucagon secretion.

Leptin has been shown to act as a satiety hormone, which negatively regulated food intake via its actions in the ventromedial hypothalamus (Friedman, 2000; Gavrilova,

Marcus-Samuels, Leon, Vinson, & Reitman, 2000), but has also been shown to have a direct insulin sensitising effect on peripheral tissues (Ogawa et al., 1999; Shimomura,

Hammer, Ikemoto, Brown, & Goldstein, 1999). Leptin binds to OB-R receptors, which belong to the class 1 cytokine receptor family and are ubiquitously distributed. Binding of leptin to its receptor activates the Janus kinase (JAK)/signal transducers and activators of transcription (STAT) signal transduction pathway, leading to its numerous functions. There is considerable amount of crosstalk between the leptin signalling and other signalling pathways, including insulin-stimulated phosphatidylinositol-3-kinase

(PI3K)/Akt and mitogen-activated protein kinase (MAPK) signalling. Leptin is a key component of the neuroendocrine circuitry that regulated food intake. Humans and mice with leptin or leptin receptor deficiency are hyperphagic, have lowered energy expenditure and are obese. However diet induced obesity is not associated with an absence of leptin, but rather hyperleptinemia. The inability of these elevated leptin levels to negatively regulate food intake is widely viewed as leptin resistance. The exact mechanism of this remains unknown but potential mechanisms include decreased access of leptin to hypothalamic targets, diminished responsiveness of hypothalamic neurons to leptin, modified translation of leptin-dependant signals in downstream target tissues, or some combination thereof (Morrison, Huypens, Stewart, & Gettys, 2009;

Myers, Cowley, & Munzberg, 2008). The mechanisms of hypothalamic leptin resistance are not the focus of this thesis, however for a complete review of leptin resistance, insulin resistance and their interaction with HFD feeding regimes in mice, please refer to the following reference (Morrison et al., 2009). In general it can be concluded that chronic consumption of HFD in mice (Banks et al., 2004; Banks & Farrell, 2003; El-

Haschimi, Pierroz, Hileman, Bjorbaek, & Flier, 2000; Van Heek et al., 1997) is associated with hyperleptinemia and central leptin resistance.

In addition to its roles centrally, leptin has also been tested extensively for its effects on metabolism. The identification of the leptin receptor on the β-cell indicated that leptin 38 may exert more peripheral actions, and gave rise to a line of communication now referred to as the ‘adipo-insular’ axis (Fig 1.3)(P. R. Huypens, 2007; Kieffer &

Habener, 2000; Kieffer, Heller, & Habener, 1996). Initial studies demonstrated that treatment with leptin in in vitro (Emilsson, Liu, Cawthorne, Morton, & Davenport,

1997) and in vivo (Kulkarni et al., 1999; Kulkarni et al., 1997) systems had an inhibitory effect on insulin secretion via the JAK/STAT pathway. In addition leptin has also been shown to inhibit the transcription of pro-insulin (Seufert, 2004; Seufert, Kieffer, &

Habener, 1999)via the phosphorylation of SOCS3, resulting in a reduction in insulin secretion. Leptin has also been shown to regulate β-cell mass via changes in proliferation, apoptosis, and cell growth (Marroqui et al., 2012).For a comprehensive review of leptin signalling in the β-cell and its inhibitory mechanisms please refer to

(Morioka et al., 2007). More recent studies using β-cell specific leptin receptor deficient mice, showed that deletion of the leptin receptor in the whole pancreas resulted in improved glucose tolerance, enhanced GSIS and increased beta cell size and total mass

(Morioka et al., 2007). This study confirmed the inhibitory effect of leptin on insulin secretion, but also uncovered a potentially important role in β-cell growth (Cantley,

2014). When the same mice were fed a HFD, mice displayed impaired GSIS and reduced β-cell mass, indicating that the function of leptin is altered by the prevailing diet. In an earlier study, deletion of the leptin receptor in neurons and β-cells produced mice that were obese, hyperphagic, glucose intolerant, hyperinsulinemic and had an impaired GSIS response (Covey et al., 2006). Taken together it can be concluded that leptin is a negative regulator of β-cell function in lean mice, whereas in HFD/obese mice, leptin is required to sustain adequate β-cell mass (P. R. Huypens, 2007;

Niswender & Magnuson, 2007), possibly by protecting β-cells from overstimulation and exhaustion or through some anti-apoptotic mechanism which has yet to be elucidated

(Marroqui et al., 2012; Rakatzi, Mueller, Ritzeler, Tennagels, & Eckel, 2004). The increase in leptin under HFD/obese conditions along with the suppressive effect of leptin on insulin transcription and secretion in human islets is difficult to reconcile with the hyperinsulinemia and increased GSIS that occurs during obesity and IR. Since, leptin resistance appears to explain why high leptin levels fail to moderate food intake via the hypothalamus; in the pancreas, leptin resistance may justify how both hyperinsulinemia and hyperleptinemia can coexist under obese condition (See fig 1.3)

(Covey et al., 2006; Marroqui et al., 2012; Morioka et al., 2007). Given that leptin signalling in the β-cell triggers similar pathways as those in the hypothalamus and other tissues, it would be interesting to explore whether leptin resistance occurs in the β-cell during obesity, and the mechanisms involved.

Figure 1.3 - Dysregulation of the Adipo-Insular axis. A) Under normal conditions insulin promotes the secretion of adipokines such as leptin from adipose tissue. Leptin in turn binds to the long form of the leptin receptor (OBRb) and activates OBRb-associated JAK2, which subsequently enhances STAT3 activation and inhibits insulin transcription. Additionally, the application of leptin to pancreatic beta cells also activates K+-ATP channel. Activation of the K+-ATP channel inhibits voltage-dependent calcium channels (VDCC) resulting in reduced cytosolic [Ca2+], and blocking insulin secretion. B) In leptin-resistant overweight individuals, diminished leptin signalling in pancreatic β-cells leads to chronic hypersecretion of insulin (hyperinsulinemia). Elevated insulin levels promote both insulin resistance and increased leptin biosynthesis and secretion from adipose tissue, which may further desensitize leptin signalling in the endocrine pancreas and increase leptin resistance. Chronic hypersecretion of insulin by the pancreatic β-cell because of a lack of tonic inhibition by leptin may contribute to β-cell failure and eventual manifestation of type 2 diabetes in overweight patients (adapted from (Seufert, 2004)).

1.9 Non-classical tissues affected by IR

HFD clearly has a profound effect on metabolic tissues, resulting in obesity and the development of IR. While most studies focus on the effect of diet and resulting IR on metabolic tissues, diet has also been shown to have a profound effect on long term health (Solon-Biet et al., 2014). Here we will discuss two tissues (brain and bone) not classically involved in metabolism, but are profoundly affected by diet and obesity.

1.9.1 Brain

An emerging theme in the literature is that IR in the brain may be a mediator of cognitive impairments and neurodegeneration, particularly for Alzheimer’s disease

(AD) (de la Monte, 2009). It is now widely recognized that T2D and AD share several common abnormalities including impaired glucose metabolism, increased oxidative stress, IR and amyloidogenesis (Zhao & Townsend, 2009). Large scale epidemiological studies consistently link T2D and IR with an increased risk of developing AD

(Arvanitakis, Wilson, Bienias, Evans, & Bennett, 2004; Hofman et al., 1997; Schrijvers

41 et al., 2010). AD is an age dependent heterogeneous neurodegenerative disorder characterised by a gradual decline in memory and cognitive function, and pathologically by progressive deposition of amyloid-β neuritic plaques (X. Li, D. Song, & S. X. Leng,

2015a). The amyloid-β protein precursor is cleaved by γ-secretase into several isoforms ranging from 36-43. The most abundant is the neuroprotective Aβ-40 and pathologically linked Aβ-42 (Gu & Guo, 2013).

Individuals with T2D have up to a 65% increased risk of developing AD (Arvanitakis et al., 2004; Butterfield, Di Domenico, & Barone, 2014). In clinical studies, peripheral IR, dysregulated glucose metabolism and brain specific IR have been found within AD patients (Arvanitakis et al., 2004; Craft et al., 1998; Dineley, Jahrling, & Denner, 2014;

Talbot et al., 2012; van Himbergen et al., 2012) and this has even been proposed to be

‘type 3 diabetes’. In addition to T2D, and IR, obesity and consumption of a high-fat diet are known to increase the risk of AD (Xiaohua Li, Dalin Song, & Sean X Leng, 2015)

(de la Monte, 2009). Diets high in fat also increase disease neuropathology and/or cognitive deficits in AD mouse models (Elysse M Knight, Martins, Gümüsgöz, Allan,

& Lawrence, 2014).

The link between T2D, IR and AD may be derived from the role of insulin in the brain

(Dineley et al., 2014; Kleinridders, Ferris, Cai, & Kahn, 2014). Insulin and insulin-like growth factor play a role in neuronal survival, energy metabolism, and plasticity, which are required for learning and memory (de la Monte, 2009), which may account for the majority of AD-associated abnormalities (de la Monte, 2009). Insulin also has the ability to cross the blood brain barrier to regulate glucose homeostasis and utilisation

(Woods, Seeley, Baskin, & Schwartz, 2003), which is also known to influence amyloid, neuronal survival, energy metabolism and neural network plasticity (Craft & Watson,

2004; Kleinridders et al., 2014). Insulin receptors are found throughout the brain, and are found in highest concentrations at the synapses of the hippocampus and amygdala and to a lesser extent in the cortex and cerebellum (Dineley et al., 2014; Werther et al.,

1987). Insulin has also been shown to regulate Aβ production and clearance (X. Li, D.

Song, & S. X. Leng, 2015b) and treatment with insulin has been shown to improve memory (Craft et al., 2012). Animal models of IR, AD or both have shown an increase in the Aβ42/40 ratio(or decreased when expressed as Aβ40/42) and hyper- phosphorylation of tau are exacerbated by IR (Ho et al., 2004; Lester-Coll et al., 2006;

Masciopinto et al., 2012; Plaschke et al., 2010; Searcy et al., 2012; Spies et al., 2010;

Takeda et al., 2010). However, many of these studies rely on models of accelerated aging such as the SAMP8 mice (Mehla, Chauhan, & Chauhan, 2014) or mutations that induce AD pathology in mice (e.g. 3xTgAD & 5XFAD) (Oakley et al., 2006; Oddo,

Caccamo, Kitazawa, Tseng, & LaFerla, 2003; Vandal & Calon, 2015). Feeding HFD alone has been shown to affect short term memory but was not sufficient to induce AD pathology (E. M. Knight, Martins, Gumusgoz, Allan, & Lawrence, 2014; Takalo et al.,

2014). This indicates that insulin dysfunction is not sufficient to cause AD pathogenesis in HFD fed C57BL/6 mice, unless accompanied by other factors that contribute to AD such as a genetic susceptibility or aging.

Lastly, in humans and rodent studies short term doses of insulin have improved memory function and increased Aβ clearance, while disruption of insulin signalling in the CNS leads to cognitive deficits in rodents. Furthermore, it has been indicated in mice that the insulin sensitising drugs, rosiglitazone, pioglitzone and metformin may improve hippocampus-dependent memory (Denner et al., 2012; Dineley et al., 2014) and restrict

AD progression in its early stages.

1.9.2 Bone

There is a well-established connection between the development of T2D and the onset of bone diseases such as osteoporosis (Devlin et al., 2014; Wongdee &

Charoenphandhu, 2015). HFD feeding negatively impacts on bone and bone health. The underlying pathophysiology between obesity and bone is extremely complex. Mice fed a HFD for 14 wks, show an increase in adipocytes within the marrow, and show defects in trabecular/cancellous bone mass (J. J. Cao, Gregoire, & Gao, 2009; J. J. Cao, Sun, &

Gao, 2010). Defects in cortical bone mass were not detected most likely because the turnover of bone is much faster in cancellous than cortical bone. Furthermore, losses in bone mass in relation to a HFD have been shown to be irreversible (Inzana et al., 2013).

It is believed that increased circulating leptin as a result of expanded adipose mass contributes to this phenotype. In the context of prolonged exposure to high leptin levels, leptin resistance develops and bone formation is reduced (Wee & Baldock, 2014).

Dimitri and colleagues found that the elevated leptin levels in obese children were negatively correlated to radial cortical porosity and tibial trabecular thickness, thus suggesting that leptin may be involved in regulating bone mass in the obese (Dimitri et al., 2015). However, controversially, Turner and colleagues demonstrated that peripherally leptin has a positive effect on bone mass and thus would suggest that increased leptin levels during obesity may increase bone mass (R. T. Turner et al.,

2013). The conflicting studies highlight the complicated interaction between bone and adipokines (i.e. leptin) which can act via multiple routes, and have both central and peripheral effects with potentially opposing outcomes.

In addition to leptin resistance, insulin resistance has also been hypothesised as one factor that may contribute to bone pathophysiology (Xia et al., 2012). This hypothesis has gained traction as it was recently shown that osteoblasts express the insulin 44 receptor. Insulin promotes bone formation and differentiation of osteoblasts and the addition of insulin to osteoblast cultures promotes cell survival (P. A. Hill, Tumber, &

Meikle, 1997) and collagen synthesis (Kream, Smith, Canalis, & Raisz, 1985).

This relationship between diet, obesity and reduced bone health is somewhat counterintuitive as obesity was thought to be protective as overweight individuals have a higher bone mineral density (BMD) (Ferrari, 2013; van Daele et al., 1995), a result of increased loading (J. J. Cao, 2011; Veldhuis-Vlug, Fliers, & Bisschop, 2013). However, it has also been shown that obese and T2D individuals have an increased fracture risk, but this might be site specific (Caffarelli, Alessi, Nuti, & Gonnelli, 2014; Compston et al., 2011; Veldhuis-Vlug et al., 2013) and are prone to developing osteoporosis (Ferron

& Lacombe, 2014) calling in to question the quality of the bone that is formed under these conditions (J. J. Cao, 2011; Veldhuis-Vlug et al., 2013). More recently, Mosca and colleagues identified a negative correlation between percent body fat and bone mass in overweight, obese and extremely obese adolescents (Mosca et al., 2014).

The interaction between diet, metabolism and bone is not unidirectional. Up until recently, bone was viewed as a static organ used for locomotion and calcium storage.

However, it is now clear that the skeleton can also act as an endocrine organ implicated in the regulation of glucose and energy metabolism. This may be in part mediated by the osteoblast derived hormone osteocalcin (Ferron & Lacombe, 2014)). Osteocalcin negatively correlates with obesity and IR and has been shown to stimulate insulin sensitivity, insulin secretion, energy expenditure and regulate glucose homeostasis.

Osteocalcin secretion and activity is in turn regulated by leptin and insulin, among other stimuli such as glucocorticoids and the sympathetic nervous system (Ferron &

Lacombe, 2014). Taken together it is clear that metabolic heath can heavily influence

45 bone health and that bone also has the capacity to communicate with other tissues (such as liver, pancreas and brain) to regulate metabolism (J. J. Cao, 2011; Ferron &

Lacombe, 2014; Veldhuis-Vlug et al., 2013; Wee & Baldock, 2014).

1.10 Concluding remarks

In this introduction we have discussed what IR is and how it can be modelled using the high-fat diet-induced obese mouse. We have identified that within this model there is a lack of knowledge surrounding the acute and longitudinal metabolic changes/adaptations that occur as a consequence of diet. We have also identified the need for better therapies of obesity and IR and this may lie in the form of anti- angiogenic agents. In addition we have discussed in some detail the adipokine leptin and leptin resistance in the brain, bone and pancreas, and the contention that surrounds just what role it plays under lean and obese conditions. Lastly, we discussed some of the mechanisms that underscore the formation of IR in multiple organ systems. While this was by no means an exhaustive list, the idea here was to provide vignettes of the complexity of IR. It is likely that no single mechanism within one tissue is the sole contributor to the disease, but rather disease is a function of a complex interaction of many system failures. Although we have spent some time discussing these mechanisms at the heart of all three results chapters is a desire to understand the whole body physiology of the mice, because although the cellular systems are complex we cannot overlook the fact that at a physiological level the system is no less complex and needs to be thoroughly characterised before we try to understand at a cellular level what is happening.

1.11 Hypothesis

We hypothesise that the metabolism of mice is highly flexibility and a temporal analysis of metabolism in response to HFD feeding will identify key adaptive transition points. In the case of VEGF neutralisation, we propose that key transitions and responses to diet can be targeted to alleviate insulin resistance. Further, we propose that diet is a major determinant of long term health and that adaptations to counter metabolic insults may be detrimental to other systems in the long term.

1.12 Aims

1. To characterise the metabolic responses to acute HFD feeding

2. To determine whether VEGF neutralisation is metabolically beneficial to whole body glucose tolerance

3. To characterise the long term effects of high-fat die-induced obesity on the metabolism of mice, in order to establish a model for the investigation of human obesity and IR.

4. To quantify the effect of diet on long term health over and above the effects of aging alone.

Materials & Methods

Chapter 2 – Materials & Methods

2.1 Experimental chapters

This thesis is comprised of the following three main results chapters.

Chapter 3:

The metabolic consequences of acute high-fat diet feeding

Chapter 4:

Systemic vascular endothelial growth factor-A (VEGF-A) neutralisation ameliorates diet induced metabolic dysfunction

Chapter 5:

The metabolic consequences of long term high-fat diet feeding

2.2 Animal Research

During my time as PhD student, animal research has served as a red line throughout.

Working with mice can be challenging. In my time at the Garvan Institute as a PhD student, I have spent many hours in the animal house, learning from my peers and developing my skills before conducting my own research. There is a strong need to be confident in your handling and application of techniques in order to reduce the stress placed on the mice, and maximise the integrity of the data collected. In planning my experiments I have always utilised the 3Rs in my animal research (Replace, Reduce and

Refine). This is particularly evident in chapter 5 where from a single animal we were able to obtain multiple measurements both in vivo and ex vivo. To achieve this the endpoint of each study was clearly defined before it was commenced, and considerable effort went in to planning the most optimal way to obtain physiological data and to collect and store tissues for both live and future assay without compromising each other’s integrity. By doing this I believe we have best respected the sacrifice these animals made for our pursuit of knowledge and I hope this body of work demonstrates the utility animals have in research.

2.3 General methods

2.3.1 Animals and Husbandry

Male C57BL6 mice (7-8 wk old) were obtained from the Animal Resources Centre

(Perth, WA, Australia) or in chapter 4 from the Australian BioResources (Moss Vale,

NSW Australia) and housed at the Garvan Institute of Medical Research. Upon arrival, mice were given 1 wk of acclimation. Mice were maintained on a 12 h light/dark cycle

(0700/1900 h) and given ad libitum access to food and water. All housing and procedures were carried out in accordance with the National Health and Medical

Research Council (NHMRC; Australia) guidelines for animal research and were approved by Garvan Institute/St Vincent’s Hospital Animal Experimentation Ethics

Committee.

2.3.2 Experimental diets and food intake

Age matched mice were maintained on either a standard lab chow (8% calories from fat, 21% calories from protein and 71% calories from carbohydrate; 2.6 kcal/g,

Gordon’s Specialty Stock Feeds, Yanderra, NSW, Australia) or at 8 wks of age placed on a high-fat diet (HFD)

(Surwit et al., 1988; Winzell & Ahren, 2004) sometimes referred to as a western diet

(WD) (45% calories from fat, of which 87% was lard and 13% safflower oil by weight,

20% calories from protein, and 35% calories from carbohydrates; 4.7 kcal/g; based on rodent diet #D12451 [Research Diets, Inc.]) for specified lengths of time. HFD was prepared in house, see Table 2.1, 2.2 and 2.3 for HFD recipe (Hoehn et al., 2010; Turner et al., 2007). Food intake was measured cumulatively over 5 days by measuring the weight of the food in the hopper before and after the allocated time. Bedding was temporarily replaced with dry chuxx daily. Food spillage was collected, separated from faecal matter, weighed and added to the final hopper weight. To calculate energy consumption the weight was multiplied by the energy density.

Nb: Because of the softer texture, the HFD does not flow down the feed hoppers in the same manner as chow diet. Every second day throughout all feeding trials, food was manually advanced down the hoppers to ensure mice had complete access to food at all times.

Dry Ingredients Source Amount (g)

Mixture 1 (makes 10 g)

Sodium Selenate Thermo Fisher 0.28 (0.14 if anhydrous)

Potassium Iodate Thermo Fisher 0.1

Starch Sigma Up to 10

Mixture 2 (Makes 1 kg)

Mixture 1 1

Chromic Potassium Thermo 0.55 Sulphate (ground)

Mixture 3 (Makes 1 kg)

Manganese Carbonate Thermo Fisher 0.63 (ground)

Iron Sulphate Merck 4.98 Heptahydrate (ground)

Zinc Carbonate Thermo Fisher 1.6

Copper Carbonate Sigma 0.3

Mixture 2 All

Hydroxyethyl Starch Sigma 133.9

Mix thoroughly then add the following

Calcium Carbonate Merck 356.71

Potassium Dihydrogen Merck 402.09 Orthophosphate

Sodium Chloride Sigma 74

Magnesium Oxide Thermo 24

Mix thoroughly

Table 2.3 - Mineral mix recipe for HFD. This table outlines the ingredients and the amounts required to make the mineral mix, one of the ingredients required in the dry components mix.

Dry Ingredients Source Amount (g)

CASEIN MPD Dairy 261

SUCROSE JL Stewart 230

STARCH (corn flour) JL Stewart 193

HOMEMADE MINERAL MIX Made in house 51

TRACE MINERALS MP Biomedicals 14.8

BRAN JL Stewart 57

METHIONINE Sigma 3.4

GELATINE JL Stewart 23

CHOLINE BITARTRATE Sigma 4.6

Table 2.2 - Dry ingredients recipe for HFD. This table outlines the ingredients and the amounts required to make the dry components mix. These will form the base of the HFD and can be made in advance and stored at room temperature (RT).

Wet Ingredients Source Amount (g)

Safflower oil Proteco Gold Pty Ltd 34

Lard Allowrie / York Foods 250

AIN vitamins 76-A MP Biomedicals Australia 14.8

Table 2.3 - Wet ingredients recipe for HFD. This table outlines the wet ingredients and the amounts required to complete the HFD. These are to be added to the dry mix. To facilitate the integration of the lard, warm slightly in a microwave (1-2 mins) and mix through thoroughly. Excess diet can be frozen and stored for an extended period of time.

2.3.3 Body composition

Body composition was assessed at time points indicated in figures in both chow and

HFD mice. Dual-energy x-ray absorptiometry (DEXA) scans using the GE PIXImus

Series Densitometerus2 (GE Healthcare, Little Chalfront, UK) installed with software version 1.46.007 was used to obtain % body fat measurements (GE

Medical Systems Ultrasound and BMD, Bedford, United Kingdom). To ensure correct calibration, a phantom control mouse, provided by the manufacturer, was used as a quality control each day. Mice were placed in an induction chamber filled with isofluorane (5%) until anaesthetised, and then positioned supine, with limbs extended outward from the body. Anaesthesia was maintained with 2% isofluorane using a nose cone. Each scan took approximately 3-4 min to obtain, after which animals were allowed to recover before being returned to their cages.

The region of interest was set to exclude the head and tail during scanning.

2.3.4 Indirect calorimetry

Oxygen consumption rate (VO2) and respiratory exchange ratio (RER) were measured under a consistent environmental temperature (22°C) using an indirect calorimetry system (Oxymax series, Columbus Instruments, Columbus, OH). For mice, studies were commenced after 2 h of acclimation to the metabolic chamber using an air flow of 0.6 l/min. VO2 was measured in individual mice at 27-min intervals over a 24-48 h period.

During the studies, chow and HFD mice had ad libitum access to food and water.

Activity was quantified as the total number of beam breaks that occurred while housed in the calorimetry chamber.

Figure 2.1 - Representative DEXA images. Representative images of chow and HFD fed mice with increasing amounts of adiposity after 6, 24 and 60 wks on diet.

2.3.5 In vivo measurements

2.3.5.1 Glucose and insulin tolerance tests and blood measurements

Mice were fasted for 6 h (0800-1400 h) before being administered an i.p. injection of

10% glucose solution at a dose of 1g/kg (Hoehn et al., 2008; Hoehn et al., 2009; Hoehn et al., 2010). In chapters 3 and 5 dosage was based on FFM to avoid dosing based on adiposity. Blood glucose was measured at times indicated by sampling blood from the tail tip with an Accu-Check II glucometer (Roche Diagnostics, Castle Hill, NSW,

Australia). The trapezoidal formula was used to calculate the area under the curve

(AUC). Fasting measurements were collected at the start of the GTT. Postprandial measurements were collected in the morning (0700 h). Blood insulin samples were obtained via tail bleeds using 5 µl heparinised haematocrit tubes (Drummond) and ejecting samples into a mouse ultrasensitive insulin ELISA (90080, Crystal Chem,

Downers Grove, IL, USA). Insulin tolerance test protocol was identical to glucose tolerance test, using 0.75 U/kg insulin diluted in saline.

2.3.5.2 Tracer uptake in to WAT and skeletal muscle

C57BL6 mice were exposed to different lengths of chow or HFD. To assess glucose uptake into tissue, 2-[3H]deoxyglucose (DOG) (2-5 µCi / 25 g body weight) was co- injected with glucose during the glucose tolerance test as above. Blood samples were obtained from the tail tip at 0, 15, 30, 60 and 90 or in chapter 4 at 120 mins, using 5 µL heparinised hematocrit tubes (Drummond) and immediately added to 100 µL saturated

BaOH solution. Blood samples were obtained from the tail tip at time points indicated, using 5 µL heparinised hematocrit tubes (Drummond) and immediately added to 100 µL saturated BaOH solution. At 90 or 120 min, mice were culled, and tissues were snap frozen. Protein was precipitated from blood with saturated ZnSO4 solution, and

56 radioactivity measured in supernatants by liquid scintillation spectroscopy. To analyse the clearance of 2-[3H] DOG into muscle and fat, frozen tissues were powdered, weighted (30 to 50 mg) and homogenised in 500 ul H2O with a Polytron homogenizer.

The homogenate was centrifuged at 13,000 rpm for 10 min and supernatant was collected. To determine the total counts, 150 μl of supernatant was added to 4 ml of scintillation fluid (Ultima Gold, PerkinElmer). To determine free (non-phosphorylated) glucose counts, anion exchange columns were prepared using AG 1-X8 resin (Bio-Rad) washed extensively in dH2O. Supernatant (150 µL) was applied to anion exchange columns and eluted with three x 1 ml dH2O wash steps. Of this combined elution, 1 ml was diluted into 4 ml scintillation fluid. Vortexed samples were read on a Beckman

LS6500 β-counter. 2-[3H] DOG uptake and phosphorylation in tissues was determined by subtracting AG 1-X8 resin eluate readings from total, homogenised tissue readings, and normalised for tissue weight, blood glucose concentrations during the GTT and radioactive area under the curve determined from blood samples taken during the GTT.

2.3.6 Ex vivo measurements

2.3.6.1 In vitro glucose uptake in adipose

Epididymal adipose depots were removed from mice, and immediately incubated in

DMEM supplemented with 2% BSA and 20 mM HEPES, pH 7.4 at 37°C. Visible non- parenchymal tissue was then removed, and explants were minced in to fine pieces.

Minced explants were washed twice and incubated in DMEM supplemented with 2%

BSA and 20 mM HEPES, pH 7.4 for 2 h. Adipose explants were washed in Krebs-

Ringer Phosphate buffer (KRP) (Burchfield et al., 2013) supplemented with 2% BSA, before stimulation with 0.5 or 10 nM insulin for 20 min at 37°C. During the final 5 min,

50 μM unlabelled 2-[3H] DOG containing 1 μCi/ml of 2-[3H] DOG and 0.14 μCi/ml

[14C]-Mannitol was added (total volume 0.5 ml). Glucose/tracer uptake was stopped with three washes in ice-cold PBS, quantified by liquid scintillation spectroscopy and corrected for extracellular [14C]-mannitol and protein content (Li et al., 2014; Tan et al.,

2015).

2.3.6.2 In vitro glucose uptake in muscle

Extensor Digitorum Longus (EDL) and soleus muscles were excised, mounted and pre- incubated for 30 min at 30°C in Krebs Henseleit buffer containing 5.5 mmol/l glucose,

2 mmol/l pyruvate and 0.1% BSA (KRH) (Aslesen, Engebretsen, Franch, & Jensen,

2001). KRH was gassed with carbogen (95% O2/5% CO2). Glucose uptake was assessed in KRH containing 0.375 µCi/ml 2-[3H] DOG and 0.05 µCi/ml [14C]-Mannitol in a final volume of 2 ml for 20 min at 30°C, with or without 100 nM insulin before muscles were washed in ice-cold PBS and snap frozen. Frozen muscles were incubated in 1 M KOH at 70°C for 20 min. Tracer content was quantified by liquid scintillation spectroscopy (Beckman LS6500 counter) and cellular glucose uptake calculated after correcting the [3H]-2-Deoxyglucose counts for extracellular [14C]-mannitol counts and tissue weight (Li et al., 2014; Stockli et al., 2015).

2.3.7 Tissue embedding and serum collection processing

Tissue was excised, weighed and either freeze clamped or fixed in 10% buffered formaldehyde for 24-48 h before being transferred to 70% ethanol, embedded, cut (5

µm), mounted on glass slides and H&E stained. Brain tissue was carefully dissected from the skull, hemi-dissected and frozen or fixed in 4% PFA overnight before being transferred to sucrose and subsequently embedded and cut.

For serum, whole blood was collected via cardiac puncture at time of tissue collection, spun at 4°C in the presence of EDTA and serum was collected. Serum metabolic

58 hormones were measured using Milliplex mouse metabolic hormone magnetic bead panel (MMHMAG-44K;Millipore) or a Luminex multiplex assay (Bio-Rad Mouse Grp

I 23-plex). Leptin and insulin (90030, 90080; Crystal Chem, Downers Grove, IL, USA) samples were measured via ELISA.

2.3.8 Triglyceride and glycerol measurements

Frozen tissue was powdered, and chloroform: methanol precipitation performed. TAG content was determined using an enzymatic colorimetric technique (Triglycerides GPO-

PAP; Roche Diagnostics, IN, USA). Briefly, the assay uses lipoprotein lipase to break down plasma triglycerides into glycerol and FFA. The glycerol is then measured using a coupled enzymatic reaction system, involving glycerol kinase, glycerol phosphate oxidase, and peroxidase. The final reaction involves the production of Quinoneimine dye which is quantitated by reading the absorbance at 540nm on a microplate reader

(SpectraMax 384Plus, Molecular Devices, CA, USA), and calculating TAG content against a glycerol standard curve (0-200mg/dl). Glycerol was assayed in plasma using a free glycerol determination kit (Sigma FG0100) as per manufacturer’s instructions.

2.4 Chapter 3 methods

2.4.1 Dietary intervention studies

In the first set of experiments, mice were placed on a chow or HFD. Chow mice were maintained on the diet for 21 days whereas at 14 d HFD mice were switched to a chow diet and glucose tolerance testing performed at 1 and 7 d post transition. In a similar experiment mice were place on a HFD for 14 or 28 d before being transitioned to a chow diet. At 3 and 7 days post transition 5 mice per condition were sacrificed,

Epididymal WAT was excised, weighed and an in vitro glucose uptake assay performed as described above ‘In vitro glucose uptake in adipose’.

2.4.2 Statistics

Data analysis was carried out using Prism 6 software (v6.01). Statistical significance was set at P<0.05. P values were calculated by t-test, one-way ANOVA or Two-way

ANOVA with multiple comparisons controlled for as appropriate. Data are expressed as mean ± SEM of the replicates.

2.5 Chapter 4 methods

2.5.1 VEGF neutralization

VEGF-A was neutralized using the antibody (AB) B20-4.1 (Liang et al., 2006) from

Genentech. Control antibody was mouse IgG (Sigma). Antibodies were diluted in physiological saline and administered by i.p. injection at 5 mg/kg body weight.

2.5.2 Western blots

WAT samples were from 6hr fasted or acute insulin stimulated (5 U/kg, 10 min) mice maintained on chow of HFD for 3 days with VEGF or control IgG injection. Liver samples were snap frozen following hyperinsulinemic euglycemic clamps.

Densitometry was performed using Odyssey software (Li-Cor, Inc.). n = 3-5 mice per group. Antibodies used for Western blots were from Santa Cruz Biotechnology, CA

(14-3-3, sc-629), Cell Signalling Technologies, MA (phospho-S6K, 9205; phospho-

S473 Akt, 4051; phospho-T308 Akt, 9275; total Akt, 9272, phospho-HSL, total HSL) and Vala Sciences (phospho-perilipin).

2.5.3 Hyperinsulinemic Clamps

Hyperinsulinemic, euglycemic clamps were performed in 5-hour fasted mice. For detailed methods please refer to Mangiafico ei al 2011(Mangiafico et al., 2011). In the current study an initial priming dose of insulin (Actrapid, Novo Nordisk, Bagsvaerd,

Denmark) was followed by constant infusion at a rate of 10mU/kg/min. Euglycemia was maintained by variable infusion of 2.5-10% glucose solution. Glucose turnover was calculated using Steele’s Steady State Equation.

2.5.4 Adipocyte diameter measurements

WAT was collected and fixed as described in section 2.7. Whole tissue sections where imaged on a Leica DM 6000 and mosaic images stitched together using Leica

Application Suite. Adipocyte area was analysed in a blind, semi-automated fashion using software package ImageJ.

2.5.5 Endothelial cell proliferation

Human Umbilical Vein Endothelial Cells (HUVECs) were plated at 4 x 103 cells per well in 96-well culture plates. 10ng/mL VEGF-A or VEGF-B (R&D Systems) was pre- incubated with 1.5µg/mL anti-VEGF-A Ab for 15 minutes then added to cells. Cell number was assessed with the MTS [3-(4,5-dimethylthiazol-2-yl)-5-(3- carboxymethoxyphenyl)-2-(4-sultophenyl)-2H-tetrazolium] assay (Promega) on days 0 and 3.

2.5.6 qPCR

RNA was extracted using Tri reagent (Sigma). cDNA synthesis (DyNAmo kit,

Thermo) was performed with 1 µg RNA. Gene expression of four housekeeper genes

(tbp, ywhaz, b2m, and hprt) was measured and the geometric means of the two most stable housekeepers, determined using NormFinder, were used to normalize expression of Flt-1 (VEGFR1) and Kdr (VEGFR2). Samples were run in technical triplicate on a

Roche LightCycler 480 using LightCycler 480 SYBR I Green Mastermix. qPCR primers used were:

Flt1 (F: TTGTAAACGTGAAACCTCAG, R: GATTCTTCATTCTCAGTGCAG),

Kdr (F: AATGGTACAGAAATGGAAGG, R: GCATCTCTTTCAGTCACTTC),

B2m (F: GTATGCTATCCAGAAAACCC, R: CTGAAGGACATATCTGACATC),

Hprt (F: AGGGATTTGAATCACGTTTG, R: TTTACTGGCAACATCAACAG),

Tbp (F: GTTCTTAGACTTCAAGATCCAG, R: TTCTGGGTTTGATCATTCTG),

Ywhaz (F: ACTTAACATTGTGGACATCG, R: GGATGACAAATGGTCTACTG).

2.5.7 Scanning electron microscopy

For the quantification of hepatic sinusoidal fenestrations livers were perfused. Briefl, livers were perfused via the portal vein with heparinizedsaline (5,000 IU in 1 L) to ﬂush blood from the liver, then with ﬁxative containing 3% glutaraldehyde, 2.5% paraformaldehyde, 2mmol/L calcium chloride, 2% sucrose, and 0.1 mol/L cacodylate buffer at a constant perfusion pressure of 10 cmH2O for 10 minutes. Liver pieces were then prepared (Cogger et al., 2014) and fenestration density and size were examined using a JEOL 6380 scanning electron microscope

2.5.8 Statistics

Independent one-way analysis of variance was used to compare scores of normally distributed continuous variables. Post-hoc analysis was conducted to determine significant difference between groups. Type 1 errors were controlled for by applying a

Bonferroni adjustment. One-way between group analysis of variance was tested using the Kruskal-Wallis Test where data were not normally distributed. Glucose tolerance test data are shown as median values, error bars represent inter-quartile range. For glucose tolerance tests, total areas under the curve (AUC) were used for analysis. Box plots are shown in Tukey plot format. All analysis was performed using SPSS.

2.6 Chapter 5 methods

2.6.1 In vivo assessment of insulin clearance

An i.p.GTT was used to assess insulin clearance as previously described (Tamaki et al.,

2013) with minor changes. Mice were fasted overnight and loaded with 3g/kg of FFM.

Glucose and insulin were measured as previously described in section 2.3.7.1. C-peptide was measured by mouse c-peptide ELISA kit (90050, Crystal Chem, Downers Grove,

IL, USA). Insulin clearance was measured by calculating the C-peptide/Insulin ratio.

2.6.2 Islet isolation and in vitro secretion

Islets were isolated by pancreatic digestion (Cantley et al., 2009; Cantley et al., 2007), and purified using a Ficoll-paque gradient (GE Healthcare, Chalfont St Giles, UK) and then basalled for 1 h in HEPES-buffered KRB containing 0.1% BSA (Sigma) and

2 mmol/l glucose. For insulin secretion assays, batches of five islets were incubated at

37°C for 1 h in 130 μl KRB containing 0.1% BSA and 2 mmol/l glucose (basal) or supplemented with glucose (11 or 20 mmol/l). Exenatide (5ng/ml - Byetta) was supplemented with 11 mM glucose. KCl (25 mmol/l) treatments were supplemented with 2 or 20 mM glucose. Leptin (PreproTech) was tested at 11mM glucose using two doses (2 and 20 ng/ml). Insulin release was determined by RIA (Linco/Millipore,

Billerica, MA, USA) or ELISA (90080, Crystal Chem, Downers Grove, IL, USA).

2.6.3 Pancreatic immunohistochemistry

Pancreata were removed, cleared of fat and lymph nodes and fixed in 10% buffered formalin, transferred to 70% ethanol, embedded in paraffin and serial sectioned (5μm sections) and mounted on Superfrost Plus slides (Fisher Scientific, Pittsburgh, PA).

Slides were rehydrated, then transferred to water. Antigen retrieval was performed by submerging slides in the Target Retrieval solution (Dako – S1699) and heating to 125°C

64 for 1 min and 95°C for 10 seconds before being cooled. Slides were placed in blocking solution (PBS containing 2% BSA, 5% goat & 5% donkey Serum (Sigma)) for 30 mins at RT, then incubated in a primary Ab solution (Blocking solution + 1°Ab)) containing

1°Ab as specified in table 2.4, overnight at 4°C. The following day slides were washed

(3 x 5 min) with T-TBS (0.1% tween) followed by a 1 hr incubation at RT in secondary

Ab solution (PBS + 2% BSA (Sigma)) containing antibodies specified in table 2.4. For quantification of pancreata we used 3-4 pancreata per diet at each time point using 3 sections at least 150 µm apart from each other. Whole sections were imaged and stitched on a Leica DM6000 Power Mosaic using a 40x PLAN APO objective. Image analysis was performed using custom image analysis pipelines deployed across the ilastik (Sommer, Straehle, Kothe, & Hamprecht, 2011), Fiji (Schindelin et al., 2012) and

WEKA (Hall et al., 2009) platforms. For gross measurements of pancreas, islet and adipose tissues, classifiers generated using ilastik were used to generate pixel based probability maps for features of interest. These were segmented and measured in Fiji using a series of custom macros for complex/difficult segmentation the primary segmentation was cleaned with a second round of machine learning, prior to final analysis. Nuclei and Ki-67 were segmented and counted in previously detected islets using custom macros. In total we imaged and analysed 3567mm2 of pancreas tissue in which we detected 11001 islets that contained 294158 nuclei.

Antibody Source Secondary

Anti-insulin I8510; Sigma-Aldrich Alexa 555 (1:100) anti-glucagon (G2654; Sigma-Aldrich Alexa 488 (1:200)

anti-Ki67 RM-9106-S0; Thermo Fisher Alexa 647 (1:200) Scientific

DAPI (1:500) D1306; Thermo Fisher Scientific N/A

Table 2.4 - Antibodies. Primary and Secondary antibodies used for pancreata immunohistochemistry

2.6.4 Micro-computed tomography of bone

Left and right distal femora were fixed in 4% paraformaldehyde for 16 hrs then transferred to 70% ethanol. Femora length measured using calipers (Mitutoyo, Illinois

USA). Femoral BMC and BMD were measured using the GE PIXImus Series

Densitometerus2 (GE Healthcare, Little Chalfront, UK). The distal end of the femur was scanned using micro-computed tomography (µCT) with a Skyscan 1172 scanner and associated analysis software (Skyscan, Aartselaar, Belgium) (Wong et al., 2013).

Femoral Images were captured at an improved resolution of 4.37µm. The trabecular region of interest (ROI) was defined as 100-800 µm from the growth plate (J. J. Cao et al., 2010) and assessed. The following parameters were generated: total bone volume

(TV), bone volume (BV), trabecular bone thickness, trabecular number and trabecular separation. The Cortical ROI was defined as 2500-3500 µm from the growth plate and cortical bone volume and thickness were assessed. Endosteal and periosteal perimeters were traced and measured from the uppermost slice of the cortical ROI (3500 µm from the growth plate) and the average, γ, maximum and polar moment of inertia (MMI)

(indexes of strength) was calculated. All data was generated using CT-Analyzer software (Skyscan).

2.6.5 Aβ 40 and 42 measurements

Hippocampi from chow and HFD fed mice were weighed and homogenized in 5vol/wt of Tris-buffered saline (TBS) (Tris-HCL 50 mM pH 7.6; NaCl 150 mM; EDTA 2 mM) containing a cocktail of protease inhibitors. Samples were then suspended in 2% SDS containing protease inhibitors and centrifuged at 100,000 g for 60 minutes at 4°C. The supernatant was collected for the soluble Aβ ELISA. The Aβ levels were determined by using the commercially available ELISA kits (Mouse Aβ42 – KMB3441, Mouse Aβ42

– KMB3481 – Invitrogen). Plaque levels were normalised to total protein content.

2.6.6 Behavioural testing

2.6.6.1 Open field test (Baseline movement measurements)

The open field test arena (40 x 40cm) is situated in a large box with clear plexiglass walls, no ceiling, and a white floor. Each chamber is set inside a larger sound- attenuating cubicle with lights illuminating the arena and a fan to eliminate background noise. Mice are placed into the centre of the arena and allowed to explore the test box for 10 min, while a computer software program (Activity Monitor; Med Associates) recorded activity via photobeam detection inside the testing chambers. The total distance traveled over the course of the 10 min is recorded as a measure of general activity levels. Mice are returned to normal caging and the arena is cleaned with 70%

EtOH between each mouse. The protocol is repeated once per day for three consecutive days.

2.6.6.2. Elevated plus maze (Anxiety)

The elevated plus-maze consists of four arms (77 x 10cm) elevated (70cm) above the floor. Two of the arms contain 15cm-high walls (enclosed arms) and the other two consisted of no walls (open arms). Each mouse is placed in the middle of the maze facing a closed arm and allowed to explore the maze for 5 min. A video camera records the mouse and a computer software program (Limelight; Med Associates) is used to measure the time spent in the open arms, as an indication of anxiety-like behaviour.

Mice are returned to normal caging and the maze is cleaned with 70% EtOH between each mouse. The length of time spent in the open arms is used as a measure of anxiety

2.6.6.3. Y-maze (Short term memory)

Testing is conducted in an opaque Plexiglas Y maze consisting of three arms (40 x 4 x

17cm high) diverging at a 120-degree angle. Each mouse is placed in the centre of the

Y-maze and allowed to explore freely through the maze during a 5 min session. The sequence and total number of arms entered is recorded. Arm entry is counted when the hind paws of the mouse have been completely placed in the arm. Percentage alternation is calculated as the number of triads containing entries into all three arms divided by the maximum possible alternations (the total number of arms entered minus 2) × 100. The maze was cleaned between each mouse with 70% EtOH. It is well known that spontaneous alternation is a measure of spatial working memory. The Y-maze can be used as a measure of short term memory, general locomotor activity and stereotypic behaviour.

2.7 Mass spectrometry

2.7.1 Proteomics – Sample preparation

Liver samples (n=6-8) were homogenized separately (See Fig 2.2) in SDT-lysis buffer

(2% SDS, 100mM Tris/HCl pH 7.5 + protease inhibitors [use 10 x weight of tissue in mgs]) and cellular debris removed by centrifugation at [20 000 x g], room temperature,

[15 min]. Following lysis by electronic pestle, Benzonase was added and samples sonicated to shear DNA. Protein concentration was determined using bicinchoninic acid assay (BCA) and 60µg of protein subjected to filter-aided sample preparation (FASP)

(Wisniewski, Zougman, Nagaraj, & Mann, 2009). Proteins were digested with 1.2 µg of

Lys-C and / or trypsin overnight at 37 degrees. Peptides were purified using C18 stage tips and dried by vacuum centrifugation.

2.7.2 Mass spectrometry

Peptides were resuspended in 2% acetonitrile, 0.5% acetic acid and loaded onto a 50 cm x 75 µm inner diameter column packed in-house with 1.9 µm C18AQ particles (Dr

Maisch GmbH HPLC) using an Easy nLC-1000 UHPLC operated in single-column mode with intelligent flow control loading at 950 bar. Peptides were separated using a linear gradient of 5 – 30% Buffer B over 240 min at 250 nl/min (Buffer A = 0.5% acetic acid; Buffer B = 80% acetonitrile, 0.5% acetic acid). The column was maintained at

50ºC using a PRSO-V1 ion-source (Sonation) coupled directly to a Q-Exactive mass spectrometer. A full-scan MS1 was measured at 70,000 resolution at 200 m/z (300 –

1750 m/z; 100 ms injection time; 3e6 AGC target) followed by isolation of up to 20 most abundant precursor ions for MS/MS (2 m/z isolation; 8.3e5 intensity threshold;

30.0 normalized collision energy; 17,500 resolution at 200 m/z; 60 ms injection time;

5e5 AGC target).

2.7.3 Data processing and analysis

Data was processed using MaxQuant against the mouse UniProt database (July 2013;

50,808 entries). The data was searched with methionine oxidation as a variable modification and, carbamidomethylation of cysteine as a fixed modification using a precursor-ion mass tolerance of 20 ppm and product-ion mass tolerance of 6 ppm. All results were filtered to 1% FDR at the peptide and protein level and label-free quantification performed within MaxQuant. MS-based protein intensities were log2 transformed and percentage of missing values per sample was calculated. Quality of data was diagnosed using principal component analysis, correlation plots, dendrograms and boxplots and outliers with a high percentage of missing values removed before further analysis was carried out (only proteins with complete data across each time point was retained). The datasets were normalized using median absolute deviation (MAD) normalization and proteins with high variability (SD) were removed. This reduced the

6-week dataset to 1701 proteins, 6 months to 1121 and 60 weeks to 775 proteins.

2.7.4 Differential expression (DE) Analysis

Differential expression analysis was performed using moderated t-tests on each time point individually to identify proteins altered under high fat feeding compared to chow fed animals using LIMMA package (Ritchie et al., 2015) in the R programming environment. Array weights (Ritchie et al., 2006) were used to down-weigh samples during the linear model fit. Global variance shrinkage was performed using empirical

Bayes method and multiple hypothesis testing was corrected using Benjamini &

Hochberg method (http://www.jstor.org/stable/2346101) controlling for 5% false discovery rate. Significance was defined as adjusted p-value of 0.05.

2.7.5 Gene set test (GST)

Kolmogorov-Smirnov based gene set enrichment analysis (Mootha et al., 2003;

Subramanian et al., 2005) was performed at each time point using the moderated t- statistic as the test statistic. Pathways from the KEGG database as obtained from

Molecular Signature Database, C2 curated gene sets (Subramanian et al., 2005) were tested for up, down and mixed regulation at each time point. A p-value significance of

0.05 was used to identify significantly altered pathways.

2.8 Statistical Analysis

Data analysis was carried out using Prism 6 software (v6.01). Statistical significance was set at P<0.05. P values were calculated by t-test, one-way ANOVA or Two-way

ANOVA with multiple comparisons controlled for as appropriate. Data are expressed as mean ± SEM of the replicates.

The metabolic consequences of acute high-

fat diet feeding

Chapter 3 - The metabolic consequences of acute high-fat diet feeding

Author’s comments: At the commencement of my time in the James Lab, the HFD fed mouse was in its infant stages. Dr Kyle Hoehn placed mice on a 6 wk HFD feeding regime and took key measurements at 0, 5, 14 and 42 d, and obtained preliminary data to suggest that glucose intolerance did not progress in a linear fashion with time on diet and therefore believed mice might be able to adapt, halt and possibly even overcome glucose intolerance. Dr Kyle Hoehn subsequently left the lab and did not follow up these preliminary findings. At the same time I was beginning my PhD and we did not have a complete knowledge of what the metabolic consequences of HFD feeding were in the C57BL/6 mouse. Having very little knowledge, but recognising the potential the model held in modelling human IR, I took it upon myself to fill in these knowledge gaps. I spent the first year of my PhD characterising acute HFD feeding in the C57BL/6 mouse. This chapter is a reflection of that time.

3.1 Abstract

Excess nutrient intake is linked to diseases such as stroke, cancer, cardiovascular disease, hepatic steatosis, non-alcoholic fatty liver disease (NAFLD), IR and T2D. In humans, caloric excess contributes to the progression of IR to T2D. We performed longitudinal studies on mice fed a HFD to assess whole body glucose tolerance and insulin sensitivity of insulin-responsive tissues as the mice become IR. This study design allowed us to pinpoint key transitions in response to chronic calorie overload.

Age matched C57Bl6 mice were placed on HFD for specific periods of time and a number of metabolic parameters including body weight, body composition (by DEXA), glucose tolerance as well as in vivo glucose uptake into adipose and muscle using 3H-2- deoxyglucose and blood insulin was measured. We also performed ex vivo glucose uptake assays in white adipose tissue (WAT), white (EDL) and red muscle (soleus).

Reduced glucose tolerance was observed after 1 d of high fat feeding, which was sustained for up to 42 d without significant worsening. This was accompanied by elevated blood insulin in the fed and fasted states. We measured insulin resistance in muscle and fat at different times following transition to HFD. In vivo glucose clearance rates in WAT were significantly reduced after 3 d HFD exposure. Accordingly, a defect in insulin-stimulated glucose transport was observed as early as 3 d in ex vivo assays.

The defect observed in WAT appeared more striking at supramaximal insulin doses (10 nM) and less obvious or absent at physiological doses (0.5 nM). Defective insulin- stimulated glucose uptake into isolated muscle was only observed after 5 or 7 d HFD ex vivo and in vivo respectively. When (14 d) HFD mice transitioned to a chow diet, whole body glucose tolerance was restored rapidly (1 d), and epididymal fat mass and glucose uptake into WAT was returned to pre-HFD levels after 3 d. Interestingly, restoration of both epididymal fat mass and glucose uptake into WAT was delayed with a more

75 prolonged HFD (28d) compared to 14 d HFD fed mice. Furthermore, prior exposure to a HFD delayed glucose intolerance upon re-exposure to a HFD. Taken together we concluded that mice exposed to a western diet (WD) rapidly develop IR which forms first in the liver (1 d), followed by fat (3-7 d) and muscle (7-14 d). However glucose intolerance induced by relatively short term (14 d) HFD feeding can be quickly reversed, and upon re-exposure to a HFD mice appeared to display a metabolic memory that allowed them to slow the onset of glucose intolerance.

3.2 Introduction

The increasing prevalence of obesity and T2D has reached global epidemic proportions.

As a result of obesity, humans develop insulin resistance, which is the pathological centre for the progression and development of T2D. Intervention at early stages of insulin resistance is key, as both IR and T2D have been implicated in a number of other diseases, such as cancer, osteoporosis, cardiovascular disease, and neurodegenerative disorders such as Alzheimer’s disease. One of the major factors linked to the development of IR is the emergence of a western diet comprised of processed food, and its high dietary fat content (De Vogel-van den Bosch et al., 2011; Omar et al., 2012).

A major finding was that exposure of lab rodents to these western style diets also leads to insulin resistance as measured by glucose tolerance testing or hyperinsulinemic euglycemic clamp. Hence, such models have been invaluable for providing mechanistic insights into the development of IR and many studies have used this model to investigate the effects of various interventions on IR. Despite numerous studies using these models there are relatively few studies examining the long term consequences of

HFD and many studies only focus on metabolism. Moreover most studies measure IR at a single point in the progression of the disease and fail to take in to account the progressive nature of IR. Hence there are a number of questions directly relating to the progression of insulin resistance in mice that remain unanswered. For example, how quickly do mice develop glucose intolerance in response to HFD-feeding? What is the temporal shape of IR development in response to a HFD and do mice display any adaptability? Do all tissues become insulin resistant at the same time? Is IR readily reversible? Here we sought to address these questions in one of the most common

HFDs applied to C57BL/6 mice (45% calories derived from Lard).

3.3 Methods

For all methods please refer to Chapter 2.

3.4 Results

Since insulin resistance is a chronic disorder, physiological measurements taken at a single time point after exposure to a HFD may not be representative of all the features of IR. Of particular interest is to follow the transition from a healthy to an IR state and to determine if the IR continues to worsen throughout the life of the animal or if it simply established a new steady state that persists for the life of the animal. Here we sought to examine the development of IR as mice transition on to a HFD. To obtain a longitudinal image of insulin resistance we fed C57BL/6 male mice a chow or HFD for specific lengths of time (as indicated). Feeding regimens were designed so that diets were initiated with the longest time point first and subsequent time points commenced in descending order such that all mice could be sacrificed simultaneously and would be appropriately age matched.

3.4.1 Adiposity but not lean mass change with a high fat diet

Mice were randomly assigned to a treatment group and body composition (Fig 3.1) was assessed across a time course of HFD feeding. Initially, total body weight remained unchanged until 14 (28.9 g ± 0.4) days (d) where a small but significant increase was observed in comparison to the chow controls (27.16 g ± 0.2), which was sustained at 28

(29.5 g ± 0.6) and 42 d (28.9 g ± 0.5). Although significant, the maximal change was greatest at 28 d however this increase was just 2.3 g, indicating that changes in body composition even over a 6 wk period are minor. Lean mass (Fig 3.1 B) was unchanged from 0 to 42 d of dietary intervention, however adiposity as a percentage (Fig 3.1 C) of total body weight steadily increased, reaching significance at 14 and 42 d of HFD

78 feeding. Epididymal fat pads (Fig 3.1 D) were excised and weighed, similarly reaching significance at 14 and 42 d of HFD feeding. The changes observed in adiposity but not lean mass indicate that changes in total body weight were driven by adiposity.

Figure 3.1 – Body composition in response to HFD feeding. A) Body weight (grams) was measured at indicated time points. B) DEXA scanning was performed to determine fat free mass (grams) and whole body adiposity which was normalised to body weight and represented as C) percent adiposity (%). D) Epididymal fat pads were excised and weighed (grams). A-D) One-Way-ANOVA was performed with all comparisons to chow (0) (*p<0.05, ** p<0.01, ***p<0.001, ****p<0.0001). Data are mean ± SEM, n = 10-15.

3.4.2 The dynamic progression of IR following ongoing exposure to HFD

Whole body glucose tolerance was used as a surrogate measure of insulin sensitivity and to assess the induction of IR. Glucose tolerance testing (GTT) (Fig 3.2 A) was performed at time points indicated and area under the curve (AUC) was calculated (Fig

3.2 B). In response to an i.p. bolus of glucose, blood glucose (Fig 3.2 A) was elevated in all mice at the first time point measured (15 min). The peak height and the rate of blood glucose clearance (i.e. the slope of the curve) indicate the degree of glucose tolerance, where the greater the AUC the more glucose intolerant. Chow mice displayed rapid glucose clearance. Impaired glucose tolerance developed in a stepwise manner following exposure to HFD. Strikingly glucose intolerance was evident as early as 1 d of HFD feeding. The level of impairment was sustained for 5 days before worsening again at 7 d. No further deterioration in glucose tolerance was observed with increased exposure to the HFD. Postprandial glucose levels (Fig 3.2 C) were elevated after 1 d of

HFD feeding with a further rise at 3 d. There was a reduction in postprandial glucose from 5-14 d with no significant difference observed compared to chow fed mice between 7-28 d, with a mild elevation again at 42 d. Fasting glucose (Fig 3.2 D) although not significant in comparison to chow controls, trended upwards after 1 d of

HFD feeding, remaining at this level from 1-42 d HFD feeding.

Figure 3.2 –The effects of HFD on glucose tolerance. A time course of HFD feeding was conducted and (A) GTTs (1g/kg) were performed at time points indicated. B) AUC was calculated and normalised to basal glucose levels. C) Postprandial glucose was measured at 0700 h and D) basal glucose measured following a 6 h fast prior to a GTT. A-D). One-Way- ANOVA with multiple comparisons was performed (A-B corrected for multiple comparisons using Bonferroni: C-D Fishers LSD test), all comparisons are to Chow (0) (*p<0.05, ** p<0.01, ***p<0.001, ****p<0.0001). Data are mean ± SEM, n=5-10.

It is widely believed that IR is accompanied by compensatory hyperinsulinemia. To assess this we examined insulin levels in the fed and fasted states. Insulin levels in both the fed and fasted state (Fig 3.3 A-B) were significantly elevated after just 1 day of

HFD feeding. Intriguingly, continued exposure to the diet was accompanied by a return of insulin levels toward levels seen in chow animals. In the fed state, insulin levels were somewhat lower after 3 d HFD and this degree of hyperinsulinemia was sustained for the duration of the experiment accept at 42 d. Similarly, fasting insulin (Fig 3.3 B) levels were also significantly elevated at early time points (1 & 3 d) with HFD but not at later time points (5-42 d).

Figure 3.3 – A longitudinal study of insulin in HFD fed mice. A) Postprandial and B) Fasting insulin (ng/ml) measured at 0700 h or following a 6 h fast, respectively. One-Way-ANOVA (Fishers’ LSD test). All comparisons are to Chow (0) (*p<0.05, ** p<0.01, ***p<0.001, ****p<0.0001). Data are mean ± SEM, n=5-10

A major role of insulin is to stimulate glucose uptake into muscle and fat, which acts to lower blood glucose, which is raised following a meal. During IR these tissues become less responsive and have reduced glucose uptake in response to the same dose of insulin. We next sought to assess the speed of onset and temporal development of IR in visceral white adipose and muscle tissue. To assess this in vivo we performed a GTT and spiked the glucose bolus with 2-[3H]DOG. Upon phosphorylation, [3H]-2- deoxyglucose-6-phosphate is unable to exit adipose and muscle cells, allowing the measurement of glucose uptake in vivo. This allows the assessment of tissue-specific glucose uptake in a physiological environment. In vivo 2DOG uptake into WAT (Fig

3.4 A) was normal after 1 d but was significantly reduced after 3 d of HFD feeding. The inhibition in adipose tissue 2DOG uptake was observed at all subsequent time points but did not worsen with increasing length of diet, indicating a maximal deficiency was reached after 3 d of HFD and maintained for at least till 42 d.

3.4.3 A temporal analysis of insulin resistance in WAT & skeletal muscle

As an alternate method to assess white adipose tissue (WAT) insulin action we isolated epididymal fat pads from animals and incubated these as explants in vitro. Insulin stimulated 2-[3H]DOG uptake was performed ex vivo with either a sub-maximal (0.5) or maximal dose (10) nM of insulin. This method facilitated the analysis of different doses of insulin and confirms whether IR was intrinsic to the tissue. Exposure of adipose tissue from chow mice to either 0.5 or 10 nM insulin increased glucose uptake by three or four fold, respectively (Fig 3.4 B). With HFD, insulin stimulated 2-[3H]DOG uptake into fat was significantly impaired at the maximal insulin concentration (10 nM) after 3 d of HFD and there was a tendency for this to be impaired with submaximal insulin

(0.5) but this did not reach significance until 7 d of HFD feeding. Intriguingly the level of IR observed at the submaximal insulin concentration remained inhibited to a similar 83 degree at all subsequent time points, while at 10 nM insulin there was steady decline in insulin action which levelled out at 14 d. This degree of IR was maintained for 42 d of

HFD feeding. Furthermore from 14 d HFD feeding and onwards the increase in glucose uptake observed between 0.5 and 10 nM insulin is completely lost indicating a further loss in insulin sensitivity.

Figure 3.4 – A temporal analysis of insulin resistance in WAT A) In vitro 2-[3H]DOG uptake in to epididymal WAT, post GTT was performed with time on HFD as indicated on x-axis. B) Epididymal fat pads were isolated and Insulin stimulated 2-[3H]DOG uptake was performed ex vivo. A) One-way-ANOVA was performed (Fisher’s LSD test). B) Two-Way-ANOVA with (Tukey) multiple comparisons. All comparisons are to Chow (0) (*p<0.05, ** p<0.01, ***p<0.001, ****p<0.0001). Data are mean ± SEM, n=5-10.

In similar studies to those described in WAT were also performed in skeletal muscle.

No significant defect at either 3 or 5 d was observed, however glucose clearance into quadriceps muscle was significantly reduced at 7 d HFD feeding(Fig 3.5 A). Similar to

WAT, the defect in muscle glucose clearance was sustained at the same level at all subsequent time points measured until 42 d HFD feeding.

Soleus (oxidative) and EDL (glycolytic) muscle was isolated and stimulated with 10 nM insulin for 20 min (Fig 3.5 B-C). We examined a limited number of time points as these assays are labour intensive. In addition recent studies have suggested that muscle IR develops subsequent to WAT IR at approximately 21 d of HFD exposure (N. Turner et al., 2013)in mice while our in vivo data suggests that it occurs at approximately 1 wk. In chow but not HFD fed mice insulin resulted in a robust and significant 2.4 fold increase in 2-[3H]DOG uptake in soleus muscle (p < 0.0001) but failed to reach significance across all HFD time points (Fig 3.5 B). In addition insulin stimulated 2-[3H]DOG uptake was significantly reduced at 14 and 42 d which trended downwards at 5 d. White muscle such as the EDL showed smaller fold changes in insulin stimulated 2-[3H]DOG uptake than was observed in soleus (Fig 3.5 C). Insulin stimulated glucose uptake in the

EDL remained significantly elevated above basal across all time points (p<0.001). A downward trend in insulin stimulated 2-[3H]DOG uptake was observed at 5 and 14 d which reached significance at 42 d HFD feeding.

Lastly we measured the triglyceride content in chow and 5, 14 and 42 d HFD fed mice as this has previously been implicated in the development of insulin resistance. No difference in triglyceride content was observed between chow and HFD fed mice out to

14 d, but was significantly elevated after 42 d of HFD exposure. This may indicate that triglyceride content does not play a role in the development of early IR.

Figure 3.5 – A temporal analysis of insulin resistance in skeletal muscle. A) In vitro 2-[3H]DOG uptake in to quadriceps muscle, post GTT was performed with time on HFD as indicated on x- axis. B) Soleus and EDL muscle was isolated and Insulin stimulated 2-[3H]DOG uptake was performed ex vivo. D) triglyceride content in quadriceps of mice fed a chow / HFD was quantified at time points indicated. A/D) One-way-ANOVA was performed (Fisher’s LSD test/Sidaks). B-C) Two-Way-ANOVA with (Tukey) multiple comparisons was performed. All comparisons are to Chow (*p<0.05, ** p<0.01, ***p<0.001, ****p<0.0001). Data are mean ± SEM, n=5-10. 86

3.4.4 Exploring the metabolic flexibility of mice

The above data show that there are time dependant changes in insulin sensitivity in different tissues, and is consistent with previous reports that IR does not occur simultaneously in tissues such as liver, fat and muscle. Thus far we have shown that glucose tolerance develops rapidly (1 d) and this is likely due to defects found in the liver, followed by adipose tissue and muscle. In view of the progressive nature of IR we next investigated whether mice given a HFD develop lasting glucose intolerance with long term systemic effects. To test this we fed mice a HFD for 14 d, and then returned them to a chow diet for 7 d. Glucose tolerance testing was performed as a measure of peripheral insulin sensitivity at time points indicated in figure 3.6. Consistent with previous data (fig 3.2 A-B) mice on a HFD displayed a significant worsening in glucose tolerance at both 7 and 14 d (Fig 3.6). After 14 d of HFD feeding, one day of chow diet intervention was sufficient to reverse all observed impairments, and no significant difference was detected between chow and HFD-Chow mice at subsequent reversal time points measured (1 & 7 d).

Figure 3.6 - A temporal analysis of metabolic flexibility in response to a change in diet. Glucose tolerance testing was performed on C57BL6 mice exposed to a chow or HFD for 14 d. After 14 d on diet, HFD mice were switched back to a chow diet and glucose tolerance testing performed at 1 and 7 days post transition. Total AUC (arbitrary units) was calculated using the trapezoidal method. Two-Way-ANOVA with (Sidaks) multiple comparisons was performed. All comparisons are to chow. *p<0.05, ** p<0.01, ***p<0.001, ****p<0.0001. Data are mean ± SEM, n=5-10

WAT is an important tissue playing a key role as an endocrine organ and regulator of glucose homeostasis. It is well documented that the mass of WAT present is correlated with the level of circulating hormones(Cantley, 2014; Galic et al., 2010), and this acts as an important mechanism to negatively regulate its own mass and provide important signals to other tissues . In order to respond to such signals in a timely manner the adipose depot possess a high degree of structural and metabolic flexibility. One of the largest WAT depots in the male C57BL/6 mouse are the epididymal fat pad.

Here, mice were fed a HFD for 14 or 28 days, at which point mice transitioned to a chow diet. As an indicator of adipose tissue flexibility in response to various diets, epididymal fat pad weight was measured at 3 and 7 d post transition (Fig 3.7). After 3 d

HFD feeding, fat pad weight had significantly increased (0.58g) above chow (0.34g) depots, and by 14 d had almost tripled in size (0.85g). When 14 d HFD mice were transitioned to a chow diet a significant reduction in epididymal weight was observed at both 3 (p<0.01) and 7 (p<0.0001) to the extent that no significant difference could be detected when compared to chow fat pad weight. Similarly to 14 d HFD fed mice, a tripling in epididymal weight was observed after 28 days (0.82g) on HFD. In contrast to

14 day HFD fed mice where a rapid reduction in fat pad weight was observed, epididymal fat pad weight was unchanged after 3 days (0.74g) of reversal, however by 7 days (0.34g) of reversal, epididymal weight had been significantly reduced (p<0.001) such that they had returned again to chow levels.

We next subjected the epididymal fat pads insulin stimulated [3H] 2-deoxyglucose glucose uptake as a measure of tissue health and insulin resistance and sensitivity (Fig

3.7 B-C). Glucose uptake was significantly reduced after 14 d on a HFD. Mice were then placed back on a chow diet and at 3 and 7 days post transition, showed a

89 significant improvement in glucose uptake which surpassed the glucose uptake observed in chow mice (Fig 3.7 B). 28 d HFD treated (Fig 3.7 C) mice displayed a severe impairment in insulin stimulated glucose uptake by 3 d on a HFD, which was further worsened by 28 d such that insulin failed stimulate uptake above basal levels. As described earlier some mice on 28 d HFD transitioned back on to a chow diet and again isolated WAT was subjected to an ex vivo insulin stimulated [3H] 2-deoxyglucose glucose uptake. An increase in insulin stimulated glucose uptake was observed at 1, 3 and 7 days post transition, but in contrast to 14 d treated fat pads, did not equal or surpass uptake levels observed in chow mice (Fig 3.7 C), indicating a delay in the restoration of glucose uptake with prolonged exposure to a HFD.

The above experiments indicate that animals recover rapidly from the dietary insult returning quite quickly to normal GTT (Fig 3.6, 3.7) once reverting back to chow diet.

We next wondered if exposure to HFD for a brief period indices some kind of adaptive change that modulates future exposure of the animal to the HFD. To explore this, mice were divided into 3 groups: chow fed control (black), mice fed a HFD for different periods of time (grey); mice that were fed HFD for 2 weeks, returned to chow for one week and then again exposed to a HFD for different periods of time (Striped) (Fig 3.8).

No significant difference was observed between animals exposed to chow diet for different periods of time. Exposure to HFD for different periods of time caused glucose intolerance consistent with the previous experiments presented in this chapter.

Intriguingly, re-exposure of mice to HFD for 1 d did not cause glucose intolerance as was the case in naïve animals. However, sustained exposure to the HFD for 7 d or longer caused a similar degree of glucose intolerance as observed in naïve animals.

Figure 3.7 - A temporal analysis of WAT adaptability to a change in diet. Mice were placed on a HFD for 14 or 28 d, and then transitioned to a recovery (R) i.e. a chow diet for 1, 3 or 7 days. A) Epididymal fat pads were excised and weighed (g). B-C) Epididymal fat pads were then transferred to DMEM, minced, basalled, washed in Krebs and insulin stimulated 2-[3H]DOG uptake performed at time point indicated. Two-Way-ANOVA with (Sidaks) multiple comparisons was performed. All comparisons are to chow unless specified otherwise. *p<0.05, ** p<0.01, ***p<0.001, ****p<0.0001. Data are mean ± SEM, n=5.

Figure 3.8 - Glucose tolerance testing in naïve and HFD exposed mice. Mice were divided in to three groups, chow (dotted line represents average response), Naïve mice fed a HFD for the first time (grey) or a third group where mice were pre-treated with 14 d HFD and 7 d chow before returning to a HFD for 28 d (Striped). Glucose tolerance tests were performed at each time point and total area under the curve (AUC) calculated using the trapezoidal method. One- Way-ANOVA performed, all comparisons are to the relevant chow at each time point unless specified otherwise. *p<0.05, ** p<0.01, ***p<0.001, ****p<0.0001. Data are mean ± SEM, n=10-15.

3.5 Discussion

We studied the metabolic consequences of short term HFD feeding (up to 42 d) in mice and the reversibility of these phenotypes. The following key observations were made: the onset of glucose intolerance was extremely rapid (Fig. 3.1, 2 & 8) and preceded the onset of obesity; IR developed rapidly in fat and muscle, but is observed later than glucose intolerance; once established, IR was sustained for several weeks without any further deterioration; glucose intolerance following short term (14 d) HFD feeding was rapidly reversible but the rate of restoration was slowed with prolonged exposure (28 d); lastly prior exposure to a HFD provided some protection against development of glucose intolerance.

Metabolic disease is often thought of as a progressive problem starting with obesity and

IR, progressing to elevated triglycerides, dyslipidaemia, hypertension, hyperglycaemia, pancreatic beta cell death, and ultimately a number of diseases including NAFLD,

CVD, and T2D. It has been hypothesised that these diseases are an outgrowth of IR.

However given the range of complications and associated diseases it is likely that the driver of IR is different at different times and that therapies might be different depending on the stage. It is therefore important to understand the temporal shape of IR development when transitioning to a HFD. The progressive nature of these diseases implies that IR worsens over time however most studies examining IR take a snapshot/static measurement and fail to describe the temporal nature of IR. Once exposed to a HFD, mice rapidly developed IR, and this persists for as long as mice were exposed to the diet. However, glucose intolerance did not worsen with increasing time on HFD; rather glucose intolerance progressed in a stepwise manner, increasing immediately at 1 d of HFD feeding with a minor increase at 7 d which was sustained for up to 42 d without any significant worsening. The worsening in glucose tolerance at 7 d 93 corresponded with the development of muscle IR which developed deficits in glucose uptake at either 7 d in vivo or non-significantly at 5 days ex vivo. Interestingly IR was established at 3 d in adipose tissue, but this did not correspond with a similar worsening in glucose tolerance. It is possible that elevated insulins masked the effect of adipose tissue on glucose tolerance, however it is also well known that adipose tissue contributes as little as 5% to whole body glucose uptake, compared to muscles 80-90% and this likely explains the impact muscle but not adipose might have on whole body glucose tolerance. In addition the observation that IR forms in adipose and muscle at 3 and 7 d respectively cannot explain whole body glucose intolerance after 1 d of HFD feeding. It is likely that 1 d glucose intolerance is driven by the development of hepatic

IR, and therefore the failure of insulin to suppress gluconeogenesis. Indeed consistent with recent studies (N. Turner et al., 2013) we and others have confirmed that the development IR is first observed in liver, and this is closely followed by WAT and then muscle. It is also noteworthy that while whole body glucose intolerance plateaus, the formation of insulin resistance in WAT and muscle as tested by ex vivo glucose uptake assays appears to have a more gradual onset that is correlated with length of exposure to a HFD. Interestingly the progressive reduction in muscle insulin sensitivity observed ex vivo was not matched by a progressive accumulation of triglycerides, indicating a disconnect between triglyceride content and insulin sensitivity. Future studies should also measure alternate lipid species such as DAGs and ceramides which have also been implicated in the development of IR (Holland et al., 2007; Hulver et al., 2003; Petersen et al., 2005). These data suggest that defects in glucose tolerance and thus insulin resistance observed with short term HFD feeding was rapidly established and reached a plateau almost immediately after transitioning to the new diet, which did not

94 progressively worsen with prolonged exposure despite a worsening in peripheral insulin resistance.

One interesting question that arises from the examination of glucose tolerance longitudinally is whether the plateau observed in GTT represents the peak of glucose intolerance? We hypothesised that we may observe several inflection points over this short term feeding intervention. We predicted that these inflection points may reveal important transition to disease and studying these time points may uncover drug-gable targets. For example, the worsening in glucose tolerance observed at 7 d may represent the failure of protective mechanisms or a worsening in liver, adipose and muscle IR, and studies targeting these points may hold the key in halting the progression of IR. In another example, insulin levels in the fasting and post-prandial state are significantly elevated over the first few days of HFD feeding (Fig 3.3). After this early period there is no detectable difference in fasting insulinemia between chow and HFD fed mice, but postprandial insulinemia remains significantly elevated in the HFD mice, which can be taken as indicative of hyperinsulinemia and insulin resistance during the overnight feeding period. In mice, it is generally accepted that exposure to a HFD promotes hyperinsulinemia, which is a compensatory mechanism prompted by increasing insulin resistance in muscle, fat and liver. One possible hypothesis for why insulin dependent glucose uptake in muscle and fat decreases with HFD feeding, but was not accompanied by overt hyperinsulinemia, could be due to adaptations in hepatic metabolism. For example, an up-regulation in triglyceride synthesis and export pathways over the first few days of a HFD could reduce fat oxidation in liver and alter the balance of glucose disposal and glucose output in the liver. Overall, in this study we identified several small inflection points, which we aim to study in more detail in future studies. However, it cannot be discounted that our time course may be too acute to properly reveal such 95 transitions. To answer whether such transitions occur beyond 42 d HFD feeding more long term studies are required which we will address in chapter 5.

We also tested the degree to which glucose intolerance was reversible with short term

HFD feeding. Mice fed a HFD for 14 d showed marked improvements in whole body glucose tolerance (Fig 3.6), a significant reduction in epididymal fat mass and improvements in WAT glucose uptake (Fig 3.7) after 1-3 d of respite from a HFD.

Indeed all indices of health measured were restored back to chow levels, indicating that the glucose intolerance and IR can be rapidly reversed after a short term exposure to a

HFD. In contrast, prolonged exposure to a HFD (i.e. 28 d) appeared to inhibit or slow the rate of recovery as 28 d fed mice were unable to reduce fat pad weight or restore glucose uptake in to WAT (Fig 3.7) to the same extent as 14 d HFD treated mice.

Interestingly in 28 d HFD fed mice, the reduction in epididymal fat pad weight correlated with the restoration of glucose uptake indicating that fat pad size and uptake are not independent. Ex vivo muscle glucose uptake indicates that IR is established later at 14 d in red (Sol) and between 14-42 d in white (EDL) muscle. One explanation for the faster rate of recovery in 14 d HFD fed mice may be that muscle IR has just been established, and is readily reversible, whereas muscle IR in 28 d HFD fed mice is more entrenched and may be more resistant to reversal. In addition, it would appear that the liver and possibly fat are more susceptible to reversal as glucose tolerance in 14 d HFD fed mice was completely restored after just 1 d of chow feeding. In future studies we will take mice out to 28 d and beyond on a HFD and reverse them to test whether these hypothesis are true. Taken together these data suggest that IR can be embedded in the molecular systems of the mouse, and that prolonged exposure does not eliminate metabolic flexibility but does increase its rigidity.

In the next set of experiments we examined the metabolic flexibility of mice by exposing them to a HFD for 14 d, and then restored glucose tolerance by placing mice back on a chow diet for 1 wk. Then to test how mice would respond to a second dietary insult, mice were re-exposed to a HFD. We hypothesised that repeated exposure to a

HFD would improve the metabolic flexibility of mice as transitioning between a state of glucose intolerance and tolerance would not allow the system to establish a status quo.

Consistent with our hypothesis mice placed back on to a HFD displayed a delayed onset of glucose tolerance compared to naïve mice fed a HFD for the first time. While glucose intolerance is rapidly reversible following dietary insult, these data hint at the possibility that lasting changes are formed at a molecular level and that these changes may provide a sort of molecular memory and protection against future dietary insults. In addition these experimental regimens represent a valuable resource for the study of rapid onset and reversal of IR, presumably in the liver. Additionally these findings offer a new model of resistance to a HFD, where the factors that drive a protective phenotype represent potential therapeutic targets.

We also examined postprandial and fasting insulin and glucose levels in a longitudinal manner (Fig 3.2-3.3) which was characterised by rapid onset of both hyperglycaemia and hyperinsulinemia. After 1 d of HFD feeding mice were hyperinsulinemic in both the postprandial and fasted state, correlating with hepatic IR. At 3 d of HFD feeding, despite the induction of adipose IR it appears there was a lessening in the phenotype as we observed a reduction in insulin levels and glycaemia is restored beyond 3 d HFD, suggesting that initially hyperinsulinemia may be an adaptive response that can partially restore normoglycaemia following acute exposure to a HFD. In contrast the fasting state was characterised initially by hyperinsulinemia which did not persist beyond 5 d and normoglycaemia across all time points measured. The relatively acute effect of HFD on 97 these parameters may be partially explained by the minimal role insulin plays in the control of glycaemia, particularly in the fasted state where glucose levels are already low and gluconeogenesis is required to maintain normoglycaemia, compared to the fed state, where insulins role is more prolific and is required to suppress hepatic gluconeogenesis and initiate glucose uptake in to muscle and fat.

The persistent elevation of insulin levels in the postprandial state is indicative of peripheral tissue insulin resistance. However, in adipose and muscle tissue we observed discrepancies between in vivo and ex vivo measures, whereby in vivo glucose uptake indicated a defect earlier than ex vivo measures. However these two tests are fundamentally different, in that in vivo measures do not assess insulin action per se but rather insulin plus other circulating factors involved in mounting a response to a glucose challenge. By isolating the tissues we were able to directly assess insulin action in adipose and muscle tissue. In muscle we observed deficits at 7 d in vivo, and at 14 and

42 d ex vivo. This may be because we used different muscles (i.e. quadriceps, soleus,

EDL) for this measure or enforces the notion that there are other factors in vivo which may worsen the degree of tissue IR. However, in adipose tissue the same depot was used for both in vivo and ex vivo measures arguing against the different tissue theory, where we observed tissue IR at 3 d in vivo. Surprisingly we observed differences in the temporal induction of IR between 0.5 and 10 nM insulin in WAT. Consistent with in vivo data we observed a defect in glucose uptake at 3 d with 10 nM insulin and although trending downwards we did not observe a similar defect at 0.5 nM insulin. This may indicate that at 0.5 nM insulin, glucose uptake does not have to cross the threshold set by IR, whereas with 10 nM the ceiling set by IR in glucose uptake cannot be overcome.

Nevertheless, these types of studies stress the importance of examining tissues within and independently of their in vivo setting as results may provide insight in to the 98 development and progression of disease. Although beyond the capacity of this study, in future studies a hyperinsulinemic-euglycemic clamp should also be included so that a direct measure of liver insulin sensitivity can be provided.

These data suggest that insulin resistance develops rapidly in mice placed on a HFD, resulting in changes in adiposity and whole body glucose tolerance. It remains unclear whether whole body glucose intolerance, and tissue specific insulin resistance observed after 42 d HFD feeding represents the peak or if a more prolonged exposure could result in further worsening in both glucose tolerance and glucose uptake. Reductions in glucose uptake in to WAT and subsequently skeletal muscle could not entirely explain the rapid onset of IR in mice and the data presented here and literature support the liver to be the first site of resistance. In addition the metabolic system at least in mice appears to be extremely flexible, as two wks of dietary insult can be almost completely reverted back to control levels after just one day of low fat diet (chow diet), indicating that the phenotype is not engrained with short term feeding. Interestingly although glucose tolerance, fat mass, and WAT glucose uptake in 14 d HFD was completely restored by

1-3 d post transition to chow, it appears longer term feeding (28 d) may delay the rate of recovery out to 7 d, indicating that increased exposure to dietary insults can make a more permanent and lasting change to metabolism. Surprisingly repeated exposure to a

HFD protected mice from the rapid onset of glucose intolerance and presumably IR.

The putative adaptation to diet, therefore offers some protection against HFD induced glucose intolerance, and within this space therapeutic targets may lie.

Overall we can conclude that diet acutely manipulates glucose tolerance and that there is strong evidence for different tissues being differentially sensitive. Understanding the factors that contribute to these in each tissue will be key to identifying how to best treat

IR within each tissue. We have also identified key time points of interest in this model; however it is yet to be seen if whole body glucose tolerance and tissue insulin sensitivity will continue to worsen with longer exposure to HFD. Furthermore, whether

HFD feeding has an impact on systems beyond glucose homeostasis and insulin resistance?

100

Systemic vascular endothelial growth factor- A (VEGF-A) neutralisation ameliorates diet induced metabolic dysfunction

This work was a collaborative effort between Christopher Meoli and Dr Lindsay Wu. Data from this chapter was published in an equal first author manuscript. Wu, L. E., C. C. Meoli, S. P. Mangiafico, D. J. Fazakerley, V. C. Cogger, M. Mohamad, H. Pant, M. J. Kang, E. Powter, J. G. Burchfield, C. E. Xirouchaki, A. S. Mikolaizak, J. Stockli, G. Kolumam, N. van Bruggen, J. R. Gamble, D. G. Le Couteur, G. J. Cooney, S. Andrikopoulos and D. E. James (2014). "Systemic VEGF-A neutralization ameliorates diet-induced metabolic dysfunction." Diabetes 63(8): 2656- 2667. I (Christopher Meoli) have since expanded upon the original manuscript and adapted the writing for this thesis. 101

Chapter 4 - Systemic vascular endothelial growth factor-A (VEGF-A) neutralisation ameliorates diet induced metabolic dysfunction

Author’s comments: In the previous chapter we discussed the rapid onset of IR in mice fed a HFD and the reversibility by dietary intervention. In humans however diet intervention in combination with exercise has failed to curb the current obesity epidemic. In addition there are a number of individuals who either cannot comply with certain diets or cannot perform the necessary levels of exercise required to reduce or maintain a healthy BMI or reverse the effects of IR. For these reasons there is a very real need for intervention that is independent of diet and exercise. VEGF or rather a reduction in VEGF has been proposed as one such factor. In this chapter we investigate the potential use of an anti-VEGF factor for the treatment of obesity, glucose intolerance and insulin resistance.

102

4.1 Abstract

The vascular endothelial growth factor (VEGF) family of cytokines are important regulators of angiogenesis that have emerged as important targets for the treatment of obesity. While serum VEGF levels rise during obesity, recent studies using genetic models provide conflicting evidence as to whether VEGF prevents or accelerates metabolic dysfunction during obesity. In the present study, we sought to identify the effects of vascular endothelial growth factor A (VEGF-A) neutralization on parameters of glucose metabolism and insulin action in a dietary mouse model of obesity. Within only 72 h of administration of the VEGF-A neutralizing monoclonal antibody B.20-4.1, we observed almost complete reversal of high fat diet induced insulin resistance principally due to improved insulin sensitivity in the liver and in adipose tissue. These effects were independent of changes in whole body adiposity or insulin signalling.

These findings show an important and unexpected role for VEGF in liver insulin resistance opening up a potentially novel therapeutic avenue for obesity related metabolic disease.

103

4.2 Introduction

VEGF proteins are a subgroup of the PDGF family and comprise 4 members including

VEGF-A, VEGF-B, VEGF-C and VEGF-D, which bind their cognate receptors Flt-1

(VEGFR1) and Flk-1 (VEGFR2), to promote angiogenesis (Segerstrom et al., 2006).

Though classically studied in the context of angiogenesis stimulation in endothelial cells, VEGF receptors are present in a wide range of cell types and exhibit pleiotropic effects outside of angiogenesis (Claes, Vandevelde, Moons, & Tjwa, 2011; Tjwa,

Luttun, Autiero, & Carmeliet, 2003). For example, VEGF-B was recently shown to regulate lipid transport across endothelial cells (Hagberg et al., 2010). This is mediated by transcriptional induction of fatty acid transporters, leading to enhanced trans- endothelial transport of fatty acids, and promoting their delivery to tissues such as heart and muscle. Deletion of VEGF-B reduces ectopic lipid deposition and improves insulin sensitivity in dietary and genetic models of obesity in mice (Hagberg et al., 2012), and this evidence has been used to suggest a role for VEGF-B in type 2 diabetes and the metabolic syndrome.

The role of VEGF and angiogenesis in obesity and diabetes has become somewhat confused due to a number of recent conflicting studies (Elias, Franckhauser, & Bosch,

2013; Lu & Zheng, 2013; Yilmaz & Hotamisligil, 2013). Serum levels of VEGF-A are raised during obesity (Garcia de la Torre et al., 2008; Loebig et al., 2010; Silha, Krsek,

Sucharda, & Murphy, 2005), and rapidly decrease following bariatric surgery (Garcia de la Torre et al., 2008), suggesting that elevated VEGF is deleterious. In support of this

Lu et al (Lu et al., 2012) showed that VEGF knockdown suppressed obesity and promoted “browning” of white adipose tissue (WAT). In contrast, reports using adipose specific VEGF transgenic or knockout mice suggest that increased expression of VEGF is beneficial during obesity (Elias et al., 2012; Lu et al., 2012; Sun et al., 2012; Sung et 104 al., 2013). This is further complicated by contradictory reports depending on the model system; Sun et al (Sun et al., 2012) found that antibody neutralization of VEGF impaired metabolic homeostasis in a dietary model of obesity but improved glucose tolerance in a genetic (ob/ob) model. These inconsistencies may be due to the following reasons. In two of these studies (Elias et al., 2012; Sung et al., 2013) the Fabp4 promoter was used to achieve adipose specific over-expression or deletion using the

Cre-LoxP system. This promoter is not specific to adipose tissue (Mullican et al., 2013) meaning that the observed effects may be the result of VEGF-A changes in non-adipose tissues. In particular, FABP4 is expressed in microvascular endothelial cells (Elmasri et al., 2009), which are a key target of VEGF and present in tissues throughout the body.

Secondly, in addition to its role as an extracellular signalling factor, VEGF displays intracellular, cell autonomous regulation of cell signalling (Gerber et al., 2002; S. Lee et al., 2007). It may be that the effects observed with genetic VEGF over-expression or deletion may reflect changes in intracellular signalling, rather than changes in extracellular VEGF signalling, which is selectively targeted by neutralizing antibodies.

Lastly, the possibility remains that adipose tissue may not be the most important site of action for VEGF in mediating changes in insulin sensitivity. This would be consistent with studies reporting that systemic administration of anti-angiogenic compounds improves insulin sensitivity (Brakenhielm et al., 2004; Kolonin et al., 2004; Rupnick et al., 2002).

To address the issues described above, we determined the temporal relationship between changes in whole body adiposity and glucose homeostasis upon blockade of extracellular VEGF signalling in mice following administration of a VEGF-A neutralizing antibody (Liang et al., 2006). We show that systemic VEGF-A neutralization is an effective and rapid strategy for preventing as well as reversing diet 105 induced insulin resistance (IR) in short term and long term models of high fat feeding.

These effects occur within a short timeframe (72 h), involving almost complete amelioration of impaired hepatic insulin sensitivity and occur independently of adiposity (Brakenhielm et al., 2004; Y. Cao, 2007, 2010; Kolonin et al., 2004; Rupnick et al., 2002) and insulin signalling.

4.3 Methods

For all methods please refer to Chapter 2.

4.4 Results

To investigate the role of VEGF-A on metabolic activity in vivo, we utilised the selective VEGF-A neutralizing antibody B20-4.1, which has previously been characterised on an in vitro , in vivo and structural basis (Fuh et al., 2006; Liang et al.,

2006). To further test the specificity of B20-4.1, we measured the ability of this antibody to suppress the biological activity of either mouse VEGF-A or VEGF-B in primary endothelial cells. Human umbilical vein endothelial cells (HUVECs) were incubated with either recombinant mouse VEGF-A or VEGF-B (10 ng/ml) in the presence or absence of the VEGF-A neutralising antibody B20-4.1 (1.5 µg/ml) and the number of viable cells in proliferation was quantified by measuring the MTS absorbance, where absorbance is positively correlated with proliferation. To establish a baseline HUVECS were incubated alone for 0 and 3 d where an increase in viable proliferating cells at day 3 compared to d 0 was detected. HUVECs were then incubated with either VEGF-A or VEGF-B with or without the anti-VEGF-A antibody B20-4.1.

Treatment with VEGF-A or VEGF-B stimulated endothelial cell proliferation and consistent with previous findings B20-4.1 significantly blunted proliferation in cells pre-treated with VEGF-A, but had no effect in VEGF-B treated cells (Fig. 4.1).

106

Figure 4.1- Biological activity of VEGF-A neutralising antibody B20-4.1. Human umbilical vein endothelial cells (HUVECs) were incubated as indicated with either recombinant mouse VEGF- A or VEGF-B (10 ng/ml) in the presence or absence of the VEGF-A neutralising antibody B20-4.1 (1.5 µg/ml). *p<0.05, Mann-Whitney test

4.4.1 VEGF-A neutralization blocks the onset of diet-induced glucose intolerance

To investigate the role of VEGF-A in metabolic dysfunction, eight wk. old C57BL6 males were randomly assigned to treatment groups, and an initial baseline glucose tolerance test was performed (Fig. 4.2 A, 4.2 B) in all mice to confirm homogeneity in the ability to handle glucose (i.p. 1g/kg). Following a bolus of glucose, blood glucose was measured over time as an indicator of glucose tolerance, where a greater excursion from basal glucose levels and an increased area under the curve (AUC) represent an increase in glucose intolerance. Across all treatment groups blood glucose peaked at 15 minutes and reduced by ~20-25% by 30 minutes. All mice had returned to basal glucose levels by 120 minutes post injection (Fig 4.3 A-B). Prior to diet/drug intervention all

107 four treatment groups exhibited a similar level of glucose tolerance as no difference in

GTT AUC (data not shown) was observed (Fig 4.2 A). Mice were then given a single i.p. injection with either the VEGF-A neutralizing antibody B20-4.1, or the control antibody (5 mg/kg body weight), and immediately placed on either a chow or HFD.

Three d of high fat feeding was sufficient to cause a pronounced decrease in whole body glucose tolerance in control mice (Fig 4.2 C). HFD IgG mice displayed a greater excursion from basal glucose compared to chow IgG/VEGF mice, as well as a slower rate of clearance (indicated by the slope of the curve), resulting in an increased AUC

(Fig 4.2 C-D, Table 4.1). Glucose intolerance was accompanied by an increase in fasting glucose indicative of peripheral insulin resistance (Table 4.1). In contrast treatment with the VEGF-A antibody almost completely prevented impaired glucose tolerance in HFD animals (Fig 4.2 C-D, Table 4.1) as was indicated by a significant reduction in GTT AUC (Fig 4.3 C-D, Table 4.1). These improvements in glucose tolerance were observed without any significant change in circulating insulin levels in either the basal or glucose stimulated state (Fig. 4.2E, Table 4.1) of VEGF treated animals. In addition both chow and high fat fed animals treated with VEGF-A neutralising antibody displayed a reduction in fasting blood glucose (Table 4.1).

We investigated whether increased glucose uptake in to peripheral tissues, including skeletal muscle and white adipose tissue (WAT) underpinned improvements in glucose tolerance. To determine the contribution of these tissues to the changes in glucose tolerance we administered 3H-2-deoxyglucose (2-DOG) during the glucose tolerance test and measured uptake into skeletal muscle and WAT (Fig. 4.2F-G). We observed increased 2DOG uptake into WAT of mice treated with VEGF-A under both chow and

HFD conditions, but did not observe any improvement in quadriceps of mice treated with VEGF-A during high fat feeding. 108

To investigate whether administration of anti-VEGF-A was having the desired effect we next measured circulating VEGF-A levels. Indeed the VEGF-A neutralizing antibody raised circulating VEGF-A levels in both the chow and HFD groups above the IgG control mice (Fig 4.3 A). This is consistent with the prior observation that binding of

VEGF-A antibody to circulating VEGF blocks clearance and increases circulating levels

(Finley, Engel-Stefanini, Imoukhuede, & Popel, 2011; Segerstrom et al., 2006; Willett et al., 2005; J. C. Yang et al., 2003). Another possibility is that VEGF neutralisation might alter the expression level of the VEGF receptor (VEGFR) expression given that neutralisation prevents clearance and results in an accumulation of antibody bound inactive VEGF in serum that is not see by tissues. We measured VEGFR1 (Flt1) and

VEGFR2 (KDR) mRNA expression in the livers chow and HFD ice +/- VEGF neutralisation. Surprisingly, there was a trend towards decreased VEGF receptor expression for both receptors (Fig. 4.3B-C), which may be the result of reduced binding, as all antibody bound VEGF-A is rendered biologically inactive and remains in circulation.

109

Figure 4.2 - Glucose tolerance during VEGF neutralization in acute high fat fed mice. A) Schema for experiment. Chow fed mice were randomly assigned to treatment groups and B) baseline glucose tolerance was measured. Immediately following baseline GTT, mice were injected (5 mg/kg, i.p.) with either VEGF neutralizing or control antibody (IgG), and placed on

110 chow or HFD. Three d after injection and diet, C) glucose tolerance was again assessed, and quantified by D) total area under curve. E) Blood insulin levels were measured during GTT. Phosphorylated 3H-2-DOG uptake into F) quadriceps and G) epididymal white adipose tissue during GTT was assessed to determine rates of glucose uptake. B) and C) n = 22-27 mice per treatment across four independent cohorts. D-G) n=14-19 mice per treatment across three independent cohorts. ***p<0.001, *p<0.05, Kruskal Wallis test. Glucose and insulin levels are plotted as median values, error bars are interquartile range. Solid black or grey dots represent outliers.

111

Figure - 4.3 VEGF neutralization. A) Serum VEGF levels measured by ELISA following administration with the VEGF neutralizing antibody B.20.4.1. Increased serum levels detected by ELISA are a marker of VEGF neutralization, consistent with decreased clearance rates. Gene expression of B) VEGF receptor 1 and C) VEGF receptor 2 was measured in the livers of mice 3 d after high fat feeding and antibody treatment. **p<0.01, Kruskal Wallis test. Solid black or grey dots represent outliers.

112

Parameter Chow IgG Chow HFD IgG HFD anti-VEGF anti-VEGF

Fasting glucose 10.2 9.0 11.7 10.8 (mmol/L) (9.3-10.8) (8.3-9.8) (10.8-13.1) (9.2-11.9) n=22-27 Glucose tolerance 1557 1425 1915 1667 (total area under (1420-1724) (1375-1492) (1826-2051) (1511-1766) curve) n=22-27 Fasting insulin (pg/ml) 455.1 337.9 847.6 747.2 (275.8-719.2) (250.6-582.4) (600.8-970.9) (538.2-1190.0) Liver triglycerides 14.88 12.85 (10.55- 23.83 18.10 (16.20- (µmol glycerol/ml) (12.05-18.29) 16.96) (15.47-50.42) 33.69) Serum triglycerides 80.12 (66.76- 84.61 89.12 91.39 (79.92- (µg glycerol/ml) 93.47) (70.99-123.2) (66.85-96.77) 124.5) Serum HDL (mmol/L) 1.900 (1.775- 1.820 (1.595- 2.770 (2.480- 2.565 (2.108- n=5-8 1.995) 1.890) 2.900) 2.695) Serum cholesterol 2.6 (2.5-2.8) 2.7 (1.9-2.8) 3.9 (3.8-4.2) 3.8 (3.3-4.0) (mmol/L) n=5-8 Serum glycerol (ng/ml) 74.89 67.94 (56.09- 72.72 84.62 (63.2-84.68) 97.40) (55.34-115.00) (66.46-124.00) WAT glucose 2.77 (1.47- 2.49 (2.11-3.44) 2.16 (1.73- 3.69 (3.04-4.99) clearance (3H-2-DOG 4.08) 3.18) dpm/g/min) Quad 3H-2DOG 13.11 18.13 13.41 14.34 clearance (3H-2-DOG (10.24-17.75) (15.48-23.39) (10.73-14.28) (11.48-17.56) dpm/g/min) Epididymal fat mass 259 (248-354) 324 (299-375) 459 (308-600) 541 (431-605) (mg) Retroperitoneal fat 156 (132-181) 172 (141-184) 240 (156-368) 285 (230-313) mass (mg) Interscapular brown 79.4 (64.75- 74.8 (35-82.2) 72.4 (67.8- 85.3 (59.8-97.3) fat mass (mg) 85.55) 94.7) Body weight (g) 23.1 (22.6- 22.7 (22.0-24.8) 23.2 (22.0- 23.3 (22.4-24.1) 24.1) 25.1) Adipocyte diameter 1839 (1708- 1790 (1700- 1923 (1814- 1914 (1891-1992) (µm2) 1852) 1867) 2009) Glucose infusion rate 86.97 (43.05- n.a. 9.89 (2.86- 51.46 (33.28- (GIR)(clamp) (µmol / 94.69) 20.22) 81.87) kg / min) n=5-7 Rd (clamp) (µmol / kg / 115.3 (84.38- n.a. 71.63 (65.56- 87.65 (79.33- min) 135.8) 85.23) 111.4) n=5-7 Endogenous glucose 41.19 (21.22- n.a. 54.71 (52.47- 33.29 (29.5-47.5) production (clamp) 54.84) 83.16) (µmol / kg / min) n=5-7 Table 4.1. Effects of VEGF-neutralising antibody on metabolic parameters in chow-and high-fat-fed mice, during an acute model of WD feeding. See figures 4.2 and 4.7. Values shown are median (interquartile range) and represent 14-19 animals per group unless otherwise indicated. N.a. - not applicable 113

4.4.2 VEGF-A neutralisation reverses glucose intolerance following extended exposure to high fat feeding

Acute studies, we next wanted to determine if VEGF-A neutralization could reverse IR following a longer exposure to a HFD. Male (8 wk. old) C57BL6 mice were placed on

HFD for 4 wks to establish obesity, glucose intolerance and IR. This was sufficient to trigger more profound glucose intolerance than observed after 3 d of high fat feeding concomitant with a significant increase in fasting hyperglycaemia (Fig. 4.4 C-F). Mice were given an i.p. injection of VEGF-A neutralising antibody (5mg/kg) at d 0 and 3 and maintained on either a chow or HFD. A longitudinal series of GTTs were performed and AUC calculated. At 3 d post injection, glucose tolerance was somewhat improved in HFD VEGF mice (Fig 4.4 A, G) and a significant reduction in AUC was observed at

7 d when compared to HFD IgG mice (Fig 4.4 B, G). Similar to our short term studies

(Fig. 4.2C-D), only two injections of VEGF-A neutralizing antibody were sufficient to improve glucose tolerance in high fat fed animals (Fig. 4.4B) and this improvement was sustained for at least 17 d after the last dose of antibody (Fig. 4.4B-E) despite the increased length of exposure to a HFD. One observation was the substantial reduction in fasting blood glucose, which was maintained until 17 d post VEGF antibody treatment

(Fig. 4.5). Insulin tolerance was also assessed as a direct measure of insulin sensitivity

(Fig. 4.6). The glucose lowering effect of insulin in the VEGF-neutralized mice appeared limited by the significantly lower fasting glucose levels, complicating interpretation of these data. By 21 d post injection, the ameliorating effects of the

VEGF-A neutralizing antibody on glucose tolerance were no longer evident (Fig. 4.4F).

The eventual loss of an effect on glucose tolerance may reflect a decline in the downstream effects of VEGF-A neutralization or alternatively the eventual clearance of the neutralizing antibody. These data suggest that VEGF-A inhibition is an effective and 114 persistent strategy for treating pre-existing metabolic dysfunction. Moreover, in view of the reversibility of the effects on metabolism this suggests that the VEGF antibody is targeting a regulatory parameter that is highly plastic. After this 21 d period, mice were maintained on their respective diets without further intervention for an additional 3 wks.

(Fig. 4.4 A). Mice were again treated with a single dose of VEGF-A or control antibody

(5 mg/kg body weight), and 72 h later subjected to a glucose tolerance test (Fig. 4.7).

Again, this single re-treatment with the VEGF-A neutralizing antibody rapidly improved whole body glucose tolerance in mice exposed to a HFD for an extended period of time. This provides significant evidence in favour of the therapeutic potential of this reagent for the treatment of metabolic disease.

115

Figure 4.4 - VEGF neutralization in long term high fat fed mice. A) Eight week old mice were placed onto chow or HFD for 4 wks, and administered a single injection of either VEGF neutralizing or control antibody at d 0 and 3. Glucose tolerance was assessed B) three d after the first single injection, and again at C) day 7, D) day 11, E) day 17 and F) day 21. G) Glucose

116 tolerance was quantified by measuring areas under the curve. n = 5-8 mice per group. Glucose levels are plotted as median values, error bars are interquartile range.

Figure 4.5 - Fasting glucose levels following VEGF neutralization. Fasting glucose levels in long term high fat fed mice shown in Fig 2A were measured at indicated d following VEGF neutralization. n=5-8 mice per group. Median values with interquartile range shown.

Figure 4.6 - Insulin tolerance during VEGF neutralization in long term high fat fed mice. Long term high fat fed animals were injected with VEGF neutralizing antibody (5 mg/kg) at day 0 and day 3 as shown in Fig. 4.2 A. At day 13, animals were subjected to insulin tolerance test (0.75 U/kg). n = 5-8 mice per group. Values shown are median, error bars are interquartile range.

117

Figure 4.7 - Glucose tolerance testing following VEGF antibody re-administration. Long term high fat fed mice in Fig 4.4 A were kept on diets without treatment for 3 wks from the final GTT at which VEGF neutralization had worn off. After this 3 wk wash-out period, animals were re- injected with VEGF neutralizing antibody, and glucose tolerance assessed 72 hr. later. n = 5-8 mice per group. Median values shown, error bars are interquartile range.

118

4.4.3 VEGF-A neutralization improves hepatic insulin sensitivity

Whole body insulin action in mammals is largely governed by insulin action in muscle and liver. Our analysis of 3H-2-DOG uptake during the GTT revealed no significant effect of VEGF neutralization in muscle and a significant improvement in WAT. During acute high fat feeding, whole body IR is largely due to impaired insulin action in the liver (N. Turner et al., 2013), which precedes insulin resistance in WAT and muscle

(Kraegen et al., 1991). This is consistent with our observation of either no change in 2-

DOG uptake into muscle or WAT with high fat feeding, despite profound glucose intolerance (Fig. 4.2). To investigate this, we used the hyperinsulinemic euglycemic clamp method (Fig. 4.8) where a low glucose infusion rate (GIR) indicates hepatic IR.

After 3 d of high fat feeding, the whole body GIR in response to a 10 mU/kg/min insulin infusion was reduced by >80% compared with the chow fed group. This inhibitory effect was almost completely abolished by one single dose of the VEGF neutralizing antibody prior to commencement of the HFD (Fig 4.8 A). There was a tendency toward increased peripheral glucose disappearance (Rd) in VEGF treated animals but this failed to reach statistical significance (p=0.08) (Fig. 4.8 B). More strikingly, VEGF neutralisation completely prevented HFD-induced hepatic IR as indicated by measurement of endogenous glucose production (EndoRa) (Fig. 4.8 C).

These data are in agreement with previous studies also showing that IR during short term high fat feeding is mediated by decreased suppression of hepatic glucose output, rather than peripheral IR (N. Turner et al., 2013), and suggests that VEGF neutralization prevents HFD induced IR largely through changes in the liver.

119

Figure 4.8 Insulin sensitivity and hepatic glucose output during VEGF neutralization. Hyperinsulinemic euglycemic clamps were performed 72 h. after treatment with VEGF neutralizing or control antibody and chow or high fat feeding as in Fig 4.2 A) GIR, B) rate of disappearance (Rd), C) endogenous glucose output from the liver (EndoRa). n=5-7 mice per group, *p<0.05, **p<0.01 Kruskal Wallis test. Solid black or grey dots represent outliers.

In an effort to further explore the mechanism of VEGF neutralization on hepatic and

WAT insulin action we next examined insulin signalling. As in previous experiments mice were given either IgG or VEGF antibody and placed on a chow or HFD for 3 d

(Fig 4.2 A). Mice were then fasted and half given acute insulin stimulation (5U/kg, 0 min). WAT (Fig 4.9 A) was collected or in another study subjected to a hyperinsulinemic euglycemic clamp where livers (Fig 4.9 B) were collected, and prepared for western blotting. Basal phosphorylation of either Akt or p70 S6K was 120 similar in all groups. With insulin we did not observe any significant difference in insulin-stimulated Akt signalling (T308 or S473) or on p70 S6K (T389) with VEGF neutralization (Fig. 4.9 A). Similarly no difference in the level of phospho-Akt (T308) was observed in livers collected post clamp indicating VEGF was not acting via the insulin signalling pathway.

We then sought to determine whether changes in circulating cytokines could account for improvements in whole body glucose tolerance, and measured a panel of cytokines in the serum of VEGF or control IgG treated mice (Table 4.2). The panel included a standard set of cytokines involved in the pathophysiology of IR including c-peptide,

GIP, glucagon, IL-6, leptin, resistin, IL-1α, IL-1β, IL-3, IL-10, GM-CSF, IFNγ,

RANTES and MCP-1. While several cytokines showed changes with diet such as leptin, which is positively correlated to adiposity, no changes were observed with VEGF, indicating that the cytokines measured do not play a role in mediating metabolic changes with VEGF neutralization.

121

Figure 4.9. Phosphorylation of insulin signalling intermediates during VEGF neutralization. Insulin signalling was assessed using phospho-specific antibodies to signalling intermediates in WAT and liver tissue 72 hr. after treatment with VEGF neutralizing or control antibody and chow or high fat feeding as in Fig. 4.2. A) Phosphorylation of Akt at T308 and S473 and p70 S6K at T389 in epididymal WAT were determined in fasted mice and mice administered with insulin for 10 min. Levels of phospho-T308 Akt are presented relative to levels following acute insulin injection of mice fed a chow diet and treated with IgG antibody (A, right panel), B) phosphorylation of Akt at T308 in liver was determined in mice following a hyperinsulinemic euglycemic clamp. Levels of phospho-T308 Akt are presented relative to levels observed in clamped mice fed a chow diet and treated with IgG antibody (B, right panel).

122

Cytokine Chow IgG Chow HFD IgG HFD anti-VEGF anti-VEGF c-peptide 1033 848 1257 1466 (963-1451) (492-1282) (984-1952) (827-1922) GIP 90.7 73.7 109.7 84.9 (79.4- (47.7-127.6) (23.6-111.9) (89.2-139.0) 159.6) glucagon 153.0 161.2 116.8 (92.7-169.3) 148.9 (124.9- (138.1-156.0) (110.1-189.6) 187.3) IL-6 64.9 85.4 118.2 127.1 (23.6-75.5) (51.4-164.8) (100.9-261.4) (34.8-165.4) leptin 1399 717 (581-1637) 3756 (2969-4754) 4469 (4305- (873-1508) 6800) resistin 7173 9582 18439 (14518- 21361 (17516- (5881-9308) (4236-12484) 19920) 30433) IL-1α 40.0 27.0 (5.4-38.5) 39.6 (26.2-49.0) 32.9 (29.8- (33.2-41.9) .60 37.46 49.5)

IL-1β 229.8 166.1 278.1 238.1 (214.2- (164.6-283.1) (147.6-310.7) (243.9-324.2) 293.6) IL-3 10.0 (7.9-20.4) 16.2 (11.5-21.5) 11.9 (10.3-18.4) 15.6 (9.9-26.5) IL-10 43.4 (36.4-60.4) 48.0 (38.1-70.9) 53.3 (44.3-63.7) 46.3 (32.0- 59.1) GM-CSF 218.7 (198.8- 182.0 (162.6- 299.3 294.8 (199.0- 287.0) 233.4) (267.5-379.4) 346.9) IFNγ 9.9 (8.3-20.0) 10.2 (6.2-13.6) 11.5 (5.3-19.5) 27.8 (21.3- 82.9) RANTES 88.0 (67.3-93.2) 110.6 (64.0- 118.5 (104.0-152.5) 126.7 (108.4- 147.5) 160.9) MCP-1 356.8 (290.7- 350.3 (267.0- 518.7 (348.8-725.5) 395.9 (323.0- 405.3) 387.2) 581.5) Table 4.2 – Serum cytokines. Serum cytokine content (picogram/ml serum) in mice maintained for 3 d on chow or HFD with VEGF neutralization or control antibody (IgG) injection. Values shown are median and inter-quartile range.

123

4.4.4 VEGF-A neutralisation had no effect on adiposity, ectopic lipid deposition or body weight

We next measured the effect of VEGF neutralization on adipocyte diameter, fat depot size and whole body adiposity as changes in these parameters have also been implicated in changes in whole body insulin sensitivity. Chow fed mice showed no change with antibody treatment in all parameters measured. There was a pronounced increase in the size of epididymal and inguinal fat pads, as well as adipocyte diameter, in control (HFD

IgG) animals after 3 d of HFD, which was not affected by VEGF-A neutralisation (Fig.

4.10 A, F, and Table 4.1). We next measured whole body adiposity by DEXA scanning, and again observed an increase in adiposity with high fat feeding, with no reduction following VEGF neutralization (Fig. 10 B). Despite changes in whole body adiposity, neither diet or antibody treatment altered total body weight (Fig. 4.10 E). Using a long term model of high fat feeding (Fig. 4.4 A) we observed a more pronounced increase in epididymal fat pad mass and whole body adiposity as determined by DEXA with diet, but observed no change in animals treated with the VEGF antibody (Fig. 4.10 C-D). In contrast to previous studies (Brakenhielm et al., 2004; Kolonin et al., 2004; Rupnick et al., 2002), these data suggest that improvements in insulin sensitivity and glucose tolerance occur independently of changes in adiposity, and are consistent with a previous report showing no change in adiposity with systemic antibody neutralization of

VEGF-R2 (Lijnen & Scroyen, 2013).

124

Figure 4.10 - Adiposity during VEGF neutralization. A) Fat pad mass and B) percentage whole body fat in an acute model of high fat feeding and VEGF neutralization (Fig. 4.2), n = 14-19 mice per group. C) Epididymal fat pad mass, D) whole body fat during a long term model (Fig. 2) of HFD (Fig. 4.2), n = 5-8 mice per group. E) Body weights before and after acute high fat feeding and VEGF treatment. F) Frequency distribution of adipocyte size in 3 d model of VEGF treatment.

125

4.4.5 Evidence for altered lipid uptake with VEGF neutralization

We next quantified the level of triglyceride in liver and muscle tissues since previous studies have shown that anti-angiogenic compounds reduce ectopic lipid deposition.

This is beneficial as an increase in ectopic lipid deposition such as that observed with hepatic steatosis during increased caloric intake has been reported to play a key role in the development of hepatic IR (Nagle, Klett, & Coleman, 2009). As expected, HFD increased triglyceride content in both liver (Fig. 4.11 A) and muscle (Fig. 4.11 B), and was reduced by VEGF neutralization in muscle only. Intriguingly, VEGF neutralization raised serum triglyceride levels under both chow and HFD conditions (Fig. 4.11 C); while there was no change in circulating free glycerol levels (Fig. 4.11 D). These data support the idea that muscle triglyceride levels are reduced with VEGF neutralization due to decreased lipid uptake, rather than increased β oxidation, which would have decreased serum triglycerides and increased free glycerol in the blood. One possibility for the increase in serum triglycerides is decreased uptake of lipid particles from the bloodstream into the liver, which occurs via fenestrations in the hepatic sinusoid, whose formation is VEGF dependent (Carpenter et al., 2005). We next performed scanning electron microscopy (SEM) of hepatic sinusoidal endothelial cells (Fig. 4.12 A-G), and observed no trend in fenestration frequency, diameter or porosity with either diet or antibody, suggesting some other mechanism is at play.

126

Figure 4.11 - Ectopic lipids in VEGF antibody treated mice. Triglyceride content in A) liver, B) quadriceps (*p<0.05, Mann-Whitney test) and C) serum (p<0.05, antibody effect, 2-way ANOVA), and D) free glycerol content in serum in mice 72 hr. after administration with VEGF neutralizing antibody and HFD as described in Figs. 4.2 & 4.4. n = 14-19 mice per group

127

Figure 4.12 Fenestration of hepatic sinusoidal endothelial cells during VEGF neutralisation. Scanning electron microscopy was used to determine A) porosity, the proportion of surface area covered by fenestrations, B) frequency of fenestrations, and C) fenestration diameter. Representative images from D) chow IgG, E) chow VEGF, F) HFD IgG, and G) HFD VEGF treated animals. Mean of 5-10 images used as single value for each animal, n=5-12 animals per group.

128

4.5 Discussion

In the present study, we have shown that systemic VEGF neutralization is a rapid and effective strategy for improving glucose tolerance and insulin sensitivity under both chow and HFD conditions. VEGF neutralization not only prevented glucose intolerance upon induction of high fat feeding, but reversed glucose intolerance in long term high fat fed mice. These improvements persisted for 17 d after administration of the VEGF neutralizing antibody, though were ameliorated by 21 d. The rapid and eventually reversible nature suggests that these effects are mediated through a persistent, ongoing maintenance process, such as vascular remodelling. The vasculature of adipose tissue was previously identified as a target for reducing obesity and improving metabolic homeostasis (Brakenhielm et al., 2004; Y. Cao, 2007, 2010; Kolonin et al., 2004;

Rupnick et al., 2002). Interestingly studies have shown that metabolic benefits can be obtained by either upregulating angiogenesis through adipose specific VEGF overexpression or by targeting adipose vasculature through compounds such as TNP-

470, angiostatin, endostatin, Bay 12-9566, thalidomide and the pro-apoptotic peptide targeting adipose endothelium. From these studies it is clear that the ability of the vasculature to rapidly restructure itself is key to maintaining the metabolic health of adipose tissue and likely underlies the flexibility of adipose tissue to expand or contract in response to diet, which may have flow on effects to whole body glucose tolerance (as was observed in Chapter 3). However in this study we observed no change in adipocyte diameter, fat pad mass or overall adiposity (Fig. 4.10) with VEGF-A neutralisation indicating that WAT remodelling via vascularisation did not play a role in improving glucose tolerance. In addition our data showing a lack of effect on VEGF-A neutralisation on adipose mass are consistent with other studies showing that treatment with a pro-apoptotic peptide designed to target adipose vasculature can improve glucose

129 tolerance in a short period of time, independent of changes in fat and body mass (D. H.

Kim et al., 2012).

In general muscle (~80-85%) and fat (~5-10) are responsible for the majority of glucose disposal acting as sites of storage and utilisation. In HFD mice, improvements in glucose uptake following VEGF antibody treatment were observed only in adipose tissue. Although this is indicative of changes in adipose tissue metabolism, the minor improvements in glucose uptake observed in the WAT of HFD-VEGF treated mice is unlikely to be responsible for improvements in whole body glucose tolerance due to the small relative contribution of WAT to whole body glucose homeostasis. In addition there was no improvement in skeletal muscle glucose uptake in HFD VEGF-A treated mice, despite the fact that these animals displayed improved glucose tolerance. Recent studies have shown that upon induction of HFD feeding the liver rapidly develops insulin resistance (<1 week), and precedes development of IR in WAT and muscle tissue by one and three weeks, respectively(N. Turner et al., 2013). In our hands we observed glucose intolerance after 1 day of HFD feeding by GTT, and defective glucose uptake in to WAT after 3 d on diet, and defects in muscle glucose uptake from 5-7 d

(Chapter 3). In either case one can conclude that immediate defects in glucose tolerance are derived from hepatic insulin resistance, and that any improvement in glucose tolerance observed immediately after the diet is initiated will coincide with improved hepatic insulin sensitivity. Indeed VEGF neutralization resulted in profound amelioration of hepatic insulin resistance following 3 d of high fat feeding (Fig. 4.8) as indicated by an improvement in the GIR. These effects of VEGF neutralization on the liver are in agreement with two recent studies (Taniguchi et al., 2013; Wei et al., 2013) which also showed improved glucose homeostasis with VEGF inhibition. The lack of

130 effect in peripheral glucose disposal (Rd) further highlights the importance of the liver over that of muscle or fat.

Interestingly, we observed a trend towards increased serum triglycerides (Fig. 4.11 C) with VEGF-A antibody treatment. This increase in serum triglycerides may not reflect a change in angiogenesis per se, but rather an alternate action of VEGF-A. For example,

VEGF-B, another VEGF family member with partial homology to VEGF-A, regulates lipid uptake across the endothelium into tissues (Hagberg et al., 2010). Recently, it was shown that genetic ablation or antibody neutralization of VEGF-B protected against metabolic dysfunction in diet or genetic models of insulin resistance (Hagberg et al.,

2012), due to reduced ectopic lipid deposition into skeletal and cardiac muscle (Hagberg et al., 2012). It is possible that VEGF-A has similar effects to VEGF-B in blocking lipid uptake into tissues, which might explain increased serum triglycerides and decreased ectopic lipid deposition into muscle of chow fed mice. Consistent with this hypothesis, in a recent study from Sun et al, antibody neutralization of VEGF impaired lipid uptake into tissues of high fat fed mice, as measured by a lipid tolerance test (Sun et al., 2012).

It has been shown that VEGF positively regulates metabolic homeostasis. These findings were based on adipocyte specific over-expression or deletion of VEGF (Elias et al., 2012; Sun et al., 2012; Sung et al., 2013), and so may reflect a long term local effect of VEGF in adipose tissue. Conversely, systemic modulation of VEGF function as used here and in other studies (Taniguchi et al., 2013; Wei et al., 2013) that describe positive effects of neutralizing VEGF function, likely reflect effects of VEGF in alternate tissues, such as the liver. However, we also observed positive effects of neutralizing VEGF on insulin action in adipose tissue (Fig. 4.2). This leaves open the

131 possibility that genetic manipulation of VEGF in adipose tissue gives rise to some chronic change in adipose function, possibly even developmental, which is not observed with the shorter term systemic administration of VEGF neutralizing antibody.

We examined insulin signalling at the level of Akt to determine whether improvements in glucose transport in WAT could be attributed to enhanced signal transduction. There was no change in Akt phosphorylation following VEGF neutralization. Although Akt activation is both necessary and sufficient to transduce the insulin signal to enhance glucose transport in adipocytes (Ng et al., 2008), we cannot rule out the involvement of alternate signalling pathways, such as atypical PKCs (Standaert et al., 1997).

There are several possible mechanisms by which VEGF neutralisation might improve insulin sensitivity in the liver. Given that we observed increased glucose uptake in to

WAT with VEGF neutralisation, it is possible that a fat derived cytokine may influence hepatic glucose homeostasis. We next examined a panel of cytokines in the serum of

VEGF or control IgG treated mice (Table 4.2). One cytokine in the panel of particular interest was glucagon. In the fasted state an increase in glucagon encourages glucose production by the liver, preventing hypoglycaemia. While an increase in glucagon was observed with diet, no change was observed with VEGF-A treatment despite a reduction in fasting glucose in VEGF-A treated mice. Indeed administration of VEGF-A neutralising antibody had no effect on the levels of any circulating cytokines measured.

This may indicate that the VEGF-A neutralising antibody regulates the secretion of an alternate molecule not measured here or that acts directly on the liver.

One key question regarding these findings is how VEGF neutralization so rapidly modulates insulin sensitivity in the liver. It is conceivable that this observation has significant bearing on the mechanism by which dietary manipulations such as HFD 132 rapidly induce IR in the liver (N. Turner et al., 2013). Vascular remodelling in the liver following VEGF neutralization has been reported to activate the HIF2α pathway, induce

IRS2 expression and improve insulin signalling (Taniguchi et al., 2013; Wei et al.,

2013). In contrast, we did not observe any detectable change in downstream Akt signalling following VEGF neutralization (Fig. 4.9). We have previously shown that defects in IRS signalling are unlikely to contribute to IR (Hoehn et al., 2008). Unlike the partial restoration of IRS-2 observed in these recent studies (Taniguchi et al., 2013;

Wei et al., 2013), complete genetic deletion of IRS2 in the liver has no measurable effect on hepatic glucose homeostasis (Simmgen et al., 2006). Given this, there is no plausible explanation as to why an increase in IRS2 levels would be sufficient to improve insulin action in the absence of other changes. We therefore believe that some other mechanism most likely accounts for the effects of VEGF on liver metabolism. It is notable that microvascular blood flow is impaired in the liver of obese rodents due to a combination of ballooning of hepatocytes, distorting hepatic sinusoidal endothelial cells, collagen deposition in the spaces of Disse, and recruitment of pro-inflammatory, non-parenchymal cell types to the microvasculature, which could impair blood flow

(Farrell, Teoh, & McCuskey, 2008). VEGF is a potent vasodilator, and it is likely that

VEGF neutralization would further impair vasodilation, which is required for insulin sensitivity. Changes in vasodilation are therefore unlikely to account for the improved insulin sensitivity observed during VEGF neutralization. An alternative mechanism also involving microvascular looks specifically at changes in the fenestrae of hepatic sinusoidal endothelial cells. Fenestrations are openings in the cell membrane of endothelial cells which facilitate the rapid exchange of molecules between sinusoid blood vessels and hepatocytes (Svistounov et al., 2012). The structure of these can tell one much about the function of the liver. VEGF plays an important role in the formation

133 of hepatic fenestrations, and subsequent lipid uptake away from the bloodstream in the form of lipoprotein endocytosis (Carpenter et al., 2005). Compromised hepatic fenestrations through the neutralisation of VEGF might be expected to improve overall glucose homeostasis by blocking the uptake of lipid. Consistent with this possibility, we observed increased plasma triglycerides, and decreased hepatic triglyceride content.

However, we observed no changes in fenestration density, diameter or porosity with administration of VEGF reinforcing that neither changes in microvascular blood flow or structure underlie the positive effects on glucose tolerance observed in VEGF-A treated mice.

Several studies have linked positive metabolic affects from either upregulating or downregulating angiogenesis in WAT (Elias et al., 2013; Lu et al., 2012; Lu & Zheng,

2013; Yilmaz & Hotamisligil, 2013) . Here we have shown that VEGF neutralization results in a rapid and sustained improvement in overall glucose homeostasis, an effect primarily mediated by changes in hepatic insulin sensitivity which was unaccompanied by any major changes in the WAT. In view of the untoward side effects associated with systemic manipulation of VEGF action it is unlikely that this will provide a practical therapeutic approach for the management of insulin resistance. It is therefore essential to pinpoint the mode of action of VEGF neutralization in the liver that so effectively reverses IR in HFD mice.

134

The metabolic & health consequences of long term high-fat diet feeding

135

Chapter 5 – The metabolic and health consequences of long term high- fat diet feeding

5.1 Abstract

With the increasing incidence of obesity and related disease there is a great deal of interest in fully understanding the impact of a western diet (WD) on long term health.

Moreover diet and aging are linked to insulin resistance, a major risk factor for a constellation of diseases including osteoporosis, Alzheimer’s disease (AD) and type 2 diabetes (T2D). However, the direct contribution of diet versus aging to these processes is not clear. We conducted a comprehensive temporal analysis of WD feeding by examining the long term consequences of a chow versus high-fat-diet (HFD) on various health indices in C57BL6 mice. Mice fed a chow diet for 60 wks displayed no change in glucose tolerance, postprandial or fasting glucose and maintained a highly stable degree of health in all systems examined. Intriguingly, this precise metabolic homeostasis occurred in the face of a 25% increase in whole body adiposity. In striking contrast, we observed major impairments in every system examined, including metabolism, skeletal architecture, neurological and hepatic health with prolonged HFD feeding. Most notably, HFD resulted in significant glucose intolerance within 1 d of feeding and this was sustained for 6 months, after which it began to resolve until complete resolution by

12 months on the diet. The resolution of glucose intolerance was due to a striking compensation by the pancreas as a result of β-cell proliferation and resultant hyperinsulinemia. Preliminary studies indicate that this may be due to an unexpected stimulatory effect of leptin on insulin secretion. The HFD also caused marked deterioration in bone morphology and brain function, as measured by memory deficit and a change in the Amyloid β-40/42 ratio. The effect of aging per se on each of these

136

parameters was mild compared to the effect of diet. We conclude that western style diets have profound deleterious effects on multiple organ systems and the effect of diet is more potent than age alone.

5.2 Introduction

Throughout evolution infectious disease and malnutrition posed the greatest threats to longevity. However, improvements in medicine have led to an increase in lifespan in the last 50 years while exposing a new set of age related diseases such as osteoporosis and

Alzheimers disease (AD). Separately obesity has also been linked to diseases such as osteoporosis and AD. Obesity is often associated with insulin resistance (IR), which is a major risk factor for T2D, and there is growing evidence that it is a risk factor for other obesity related disorders such as AD, cardiovascular disease (CVD), cancer, and osteoporosis. While both aging and diet can clearly influence the risk for an individual to develop one of these chronic diseases, it is unclear what the relative contribution of diet and age is towards their formation. In the context of an obesity epidemic and an aging populaion it would be interesting to understand the impact of both on long term health outcomes.

Diet is known to affect metabolism (Surwit et al., 1995; Surwit et al., 1988; N. Turner et al., 2013; Winzell & Ahren, 2004) and play a role in determining longevity in rodent models (Solon-Biet et al., 2014). The mouse model of diet-induced obesity has become one of the most important tools for understanding the interplay between high-fat western diets and the development of obesity and IR as it does not require any genetic modification. In particular the HFD utilised in this study closely mimics the increasing availability of the high-fat/high-density foods driving the current obesity epidemic.

Indeed, when fed ad libitum with a high-fat diet, C57BL6 mice develop obesity,

137

hyperinsulinemia, hyperglycemia, and become glucose intolerant, but when fed ad libitum chow, they remain lean without metabolic derangement (Surwit et al., 1988).

There have been a number of recent reports that HFD can detrimentally affect the health of a number of tissues. For example it was shown that HFD feeding could increase the level of tau expression (Takalo et al., 2014), a marker of AD, in the brains of mice, and in another study induce memory impairment (E. M. Knight et al., 2014). In bone, HFD feeding has been shown to decrease bone volume and to interfere with intestinal calcium absorption (J. J. Cao et al., 2009; J. J. Cao et al., 2010). In fact, in almost every tissue we examined in the literature (e.g. heart, brain ,bone & liver), HFD feeding was linked to a disease process within that organ (Aroor, Mandavia, & Sowers, 2012;

Butterfield et al., 2014; Calligaris et al., 2013; J. J. Cao et al., 2009; J. J. Cao et al.,

2010; Nakamura & Terauchi, 2013; Schwingshackl & Hoffmann, 2014; Yaqoob et al.,

1995).

Since many studies that have characterised the effect of HFD feeding in this mouse model have been assessing insulin resistance, they have been mostly limited to relatively short-term dietary interventions (i.e. 6-20 wks) and analysis has been performed at a single time point. In cases where longitudinal studies have been performed they have either focused on the metabolic derangements caused by HFD feeding (B. Ahren & Pacini, 2002; Y. S. Lee et al., 2011; Surwit et al., 1991; N. Turner et al., 2013; Winzell & Ahren, 2004) or the effect of diet on a certain disease process

(McGillicuddy et al., 2013; Mosser et al., 2015; Park et al., 2005; Toyama et al., 2015;

Winzell & Ahren, 2004). Indeed there are no studies that have examined both or that have assessed multiple disease processes in the same mice. Therefore, the connection

138

between certain metabolic variables and the extent of disease in specific tissues and general health is not clear.

An obesity epidemic and an aging population are driving the increasing incidence of non-communicative diseases in the western world. However, there is a lack of longitudinal studies that access both the progressive nature of insulin resistance and relative contribution of diet and age to disease. To this end, we performed a longitudinal feeding study with specific intervention points (6, 24 and 60 wks) and addressed the following questions; 1) how well maintained is insulin sensitivity and glucose tolerance throughout life and how is this influenced by a HFD? 2) What is the relative impact of age per se versus diet on long term metabolic and general health? 3) How pervasive is diet on health vis a vis the different systems that comprise a complex organism.

5.3 Methods

For all methods please refer to Chapter 2.

5.4 Results

5.4.1 Diet is a more potent regulator of adiposity than age

Age and diet both impact organismal health and disease. To assess the relative and combined effects of aging and diet on health, we fed mice either a standard chow diet or a HFD for 60 weeks (wks), which represents approximately half their natural life. At 8 wks of age, mice were either maintained on a chow diet or transitioned to a HFD. Mice placed on a chow diet for 60 wks initially gained weight gradually (Fig 5.1 A), which tapered off by 24 wks, whereas mice placed on a HFD displayed a rapid and sustained increase in BW up to 24 wks, which slowed beyond this point.

139

DEXA scanning was conducted to assess changes in body composition. Both chow and high fat fed mice displayed an equivalent increase in FFM at 6, 24 and 60 wks of diet

(Fig 5.1 B). In contrast to FFM measurements, which increased with age, at sacrifice, some specific muscles displayed a reduction in weight with prolonged HFD feeding (6-

60 wks) but not in chow mice with age (Fig 5.1 C). DEXA scanning also revealed no significant change in adiposity in chow mice with increasing age (0-60 wks), whereas adiposity worsened increasingly with HFD at all points measured (Fig 5.1 D). We then examined the epididymal fat, a visceral depot and observed no increase in chow mice between 6 and 24 wks, however there was a significant enlargement between 24 and 60 wks (Fig 5.1 E). Consistent with increased adiposity as measured by DEXA (Fig 5.1 D), we observed the same increasing trend with HFD feeding. On average, fat pads from

HFD mice were two to three times heavier than chow mice at 6, 24 and 60 wks.

We next examined whether changes in adiposity were driven by differences in energy density or a change in amount of food consumed as both have been implicated as contributors to the obesity epidemic. Food intake was similar in both chow and HFD fed mice at 60 wks (fig 5.1 F), however when we took into account the energy density of the diets, high fat fed mice consumed ~30% more energy than chow mice (Fig 5.1 G).

5.4.2 Assessing energy substrates and activity

The respiratory exchange ratio (RER) is the ratio between the amount of oxygen consumed and carbon dioxide produced, which can be utilised to calculate the respiratory quotient. The respiratory quotient is an indicator of which fuel is being metabolised to supply the body with energy. Age had no effect on substrate utilisation as chow mice consistently displayed an RER close to 1 indicating that they predominantly utilised carbohydrates in both the light and dark cycle (Fig 5.2 A). High

140

fat diet mice showed a reduced RER after 6 wks which was further reduced at 24 and 60 wks indicating that with increasing exposure to a HFD, mice utilise increasing amounts energy from fat. Simultaneously, activity was measured in the light and dark cycles (Fig

5.2 B). Mice are nocturnal and therefore relatively inactive during the day. Consistent with this, we observed a similarly low level of activity in the light cycle in chow and

HFD mice, which was unaffected by age. In the dark cycle, mice on a HFD displayed a reduced level of activity when compared to chow mice at 6 wks, which was maintained at 24 and 60 wks on diet.

5.4.3 Metabolic phenotyping reveals an adaptive response to long term HFD feeding

Human obesity is often accompanied by a dysregulation in glucose homeostasis. To assess metabolic health we monitored glucose homeostasis in mice of different ages fed either a chow or HFD. To do this, mice were challenged with a 6 h fast followed by a glucose tolerance test to examine any defects in glucose handling while under these conditions. Mice arrived at 7 wks of age, and were given 1 wk of acclimation. Prior to the commencement of diets all mice were subjected to a baseline GTT (data no shown), before being separated into treatment groups. The incremental trapezoidal formula was used to calculate the GTT AUC. No difference in glucose tolerance between treatment groups was observed prior to dietary intervention as measured by AUC (Fig 5.3 A). In humans a mild decline in glucose homeostasis is associated with age. In contrast, we observed no deterioration in glucose tolerance with age in chow mice (Fig 5.3 A). HFD mice displayed significant glucose intolerance after just one wk on HFD and this degree of intolerance was maintained from 1-6 wks. At 12 wks there was a further deterioration in glucose tolerance, which was maintained until 24 wks. Strikingly, there was a significant improvement in glucose tolerance at 32 wks of HFD feeding, which 141

continued to improve until no difference between chow and HFD mice at 60 wks could be detected.

Human insulin resistance is accompanied by hyperglycaemia. Sustained hyperglycaemia impairs insulin-stimulated glucose utilisation and glycogen synthesis in human and rodent skeletal muscle, a phenomenon referred to clinically as glucose toxicity (Tomas et al., 2002). Interestingly, postprandial glucose, an indicator of glucose disposal efficiency was unchanged in chow and HFD mice over 60 wks, indicating no effect from either age or diet (Fig 5.3 B). In contrast, fasting glucose levels reflected changes in glucose tolerance (Fig 5.3 C) remaining unchanged in chow mice but significantly elevated with HFD feeding from 12 wks onwards before a significant improvement was observed at 60 wks. Taken together these data indicate that a HFD is deleterious for metabolic health while relatively stable with age.

Insulin resistance is often characterised by normal fasting glucose or GTT in the face of hyperinsulinemia (Shanik et al., 2008). To determine whether changes in insulin levels could account for the improvements in glucose tolerance, we examined insulin in the postprandial and fasting state in mice. Postprandial insulin was unchanged in chow mice before a mild increase with age was observed from 32 wks onwards but failed to reach statistical significance (Fig 5.3 D). With HFD feeding, a sharp increase in insulin was observed at 16 wks, which continued to worsen such that by 24 wks, insulin levels were

~3 fold higher than chow controls, a difference which persisted for the remainder of the time course. Fasting insulin levels in chow mice were unchanged for the first 24 wks, before an age related incline was observed from 30-60 wks (Fig 5.3 E). In contrast, fasting insulin levels were significantly increased after 12 wks of high fat feeding, and continued to increase out to 60 wks.

142

Figure 5.1 - Assessing body composition and food intake with long term HFD feeding. A) C57BL6 mice were fed a chow or HFD over a 60 wk period and total BW (g) assessed at time points indicated. DEXA scanning was performed and B) fat free mass (g) D) adiposity (expressed as % of BW) was measured at time points indicated. In addition C) Quadricep (mg) and E) epididymal fat (g) were excised and weighed after either 6, 24 or 60 wks on a chow or HFD. F) Food intake (g) was measured after 60 wks of either chow or HFD feeding and G) daily energy intake (Kcal/day/mouse) calculated. Significance was calculated with T-tests and Two- way ANOVAs with Sidaks correction for multiple comparisons was performed; comparisons are

143

to chow within a time point unless indicated otherwise in figure, * = p <0.05, ** = p <0.01, *** = p <0.001, **** = p < 0.0001). BW presented as mean ± SD, all other data presented as mean ± SEM (n = 10-20 mice per group).

Figure 5.2 – Assessing substrate utilisation and activity via indirect calorimetry. After either 6, 24 and 60 wks on a chow or HFD mice were placed in an indirect calorimetry system (Oxymax series, Columbus Instruments, Columbus, OH) where A) respiratory exchange ratio and B) activity (number of beam crosses) was measured, as described in materials and methods. Data was then separated in to light and dark cycles. Two-way ANOVAs with Sidaks multiple comparisons was performed; comparisons are to chow within a time point unless indicated otherwise in figure, * = p <0.05, ** = p <0.01, *** = p <0.001, **** = p < 0.0001), data presented as mean ± SEM (n = 10 mice per group).

144

Figure 5.3 - Temporal metabolic phenotyping reveals an adaptive response. Mice were fed a chow or HFD for 0-60 wks. A) Glucose tolerance testing (GTT) was performed at time points indicated and AUC calculated (0-90 mins, arbitrary units). At 0700hrs or following a 6hr fast B) postprandial or C) fasting glucose (mmol/l) was measured at time points indicated. Similarly D)

145

Postprandial or E) fasting insulins (ng/ml) were measured at time points indicated. Significance was calculated using Two-way ANOVAs with Sidaks correction for multiple comparisons; comparisons are to chow within a time point unless indicated otherwise in figure, * = p <0.05, ** = p <0.01, *** = p <0.001, **** = p < 0.0001). Data presented as mean ± SEM (n = 15-20 mice per group).

5.4.4 Assessing peripheral Insulin resistance

To obtain a metric of insulin sensitivity at each time point we calculated an insulin sensitivity index by multiplying glucose levels by their respective insulin levels in the postprandial state, providing a descriptor of how much insulin is required to clear a single unit of glucose. This is similar to a HOMA-IR but is not divided by the constant, as this constant does not relate to mice (Matthews et al., 1985; Wallace, Levy, &

Matthews, 2004). In chow mice, the insulin sensitivity index did not significantly change as mice aged (Fig 5.4 A). In HFD mice, the insulin sensitivity index was unchanged for the first 6 wks on diet. At 12 wks the index was significantly higher in

HFD mice, remaining significantly elevated and on an upward trend from 12 -60 wks of

HFD feeding. Similar results for the insulin sensitivity index were observed in the fasted state of chow and HFD mice (data not shown). These results indicate that per unit of glucose, relatively more insulin is required in HFD mice than chow fed mice to restore glycaemia back to pre-prandial levels.

Despite an improvement in glucose tolerance, the index indicates that long term HFD mice are still insulin resistant. One of the major drivers of glucose clearance following an i.p. GTT is the uptake of glucose into muscle, and to a lesser extent, adipose tissue.

We next sought to determine the insulin sensitivity of muscle and adipose tissue.

Epididymal WAT and muscle (Soleus and EDL) were isolated from mice and ex vivo 2-

[3H]deoxyglucose (DOG) uptake tested to assess insulin responses independent of circulating factors that may influence tissue uptake. 146

In chow fed mice, 0.5 and 10 nM insulin-stimulated 2-[3H]DOG uptake was increased by 3-4 and 4-6 fold over basal, respectively in WAT (Fig 5.4 B). Further, insulin stimulated glucose uptake reached a similar level in chow mice between 6 and 60 wks.

After 6 wks on a HFD, insulin-stimulated 2-[3H]DOG uptake at 0.5 nM insulin was reduced by ~38%, and ~45% with 10 nM insulin. The extent of this defect in insulin- stimulated 2-[3H]DOG transport at both doses remained similar at 24 and 60 wks of

HFD feeding.

Similarly soleus and EDL muscle were isolated and incubated with 10 nM insulin (Fig

5.4 C-H). High basal uptake is not uncommon with this technique, as isolated muscles are susceptible to stretching, which may promote glucose uptake, as was observed in 24 wk chow soleus and EDL. Therefore, each experiment has its own basal control and comparisons are made within a time point. Insulin stimulated 2-[3H]DOG uptake was significantly increased in all chow mice across all time points in both soleus and EDL muscle. With HFD feeding, IR was established in both soleus and EDL muscle by 6 wks, which persisted or worsened out to 60 wks. Specifically, in HFD treated soleus muscle, insulin stimulated glucose transport was significantly blunted (p<0.05*) by

~31% at 6 wks and this was further reduced to, 52% and 45% at 24 and 60 wks, respectively. Similarly, insulin stimulated 2-[3H]DOG uptake was significantly reduced

(p<0.05*) by ~20% in EDL muscle isolated from 6, 24 and 60 wk HFD mice .

147

Figure 5.4 – Assessing peripheral insulin resistance in WAT and muscle. Insulin sensitivity index in the postprandial state was calculated (Glucose x insulin) and describes the relative amount of insulin required to clear glucose. Epididymal WAT, soleus and EDL muscle were excised immediately after sacrifice from mice fed either a chow or HFD for 6, 24 or 60 wks. B) 2-[3H]DOG uptake in to WAT was performed in response to either 0.5nM or 10nM insulin. Glucose uptake data was normalised to total protein content and expressed as rate of glucose 148

transport, relative to chow basal. 2-[3H]DOG uptake in to (C-E) Soleus and (F-H) EDL was performed in response to 10 nM insulin and expressed as µmol/g/hr. Significance was calculated using Two-way ANOVAs with Sidaks correction for multiple comparisons; comparisons are to chow (A) or insulin to basal (B-H) within a time point, * = p <0.05, ** = p <0.01, *** = p <0.001, **** = p < 0.0001), data presented as mean ± SEM (n = 10-15 mice per group).

5.4.5 Ectopic lipid deposition with HFD feeding does not correlate with glucose intolerance

Ectopic fat is defined by the deposition of triglycerides within cells of non-adipose origin that normally contain only small amounts of fat. This occurs when expansion of the adipose depots is insufficient and fails to act as an ‘energy sink’. A western diet is associated with a positive energy balance and lipid overload. Ectopic lipid deposition in the liver has been proposed as a mechanism driving hepatic IR and is associated with the formation and progression of liver diseases such as Hepatic steatosis and NAFLD

(Nakamura & Terauchi, 2013). We have previously shown that IR in the liver develops rapidly (~1 d) with HFD feeding (Chapter 3 and 4). In addition, ectopic lipid deposition has been implicated as a contributor to the development of IR in liver (Hashemi Kani et al., 2014; N. Turner et al., 2013). In chow mice, liver triglycerides were unchanged with increasing age (Fig 5.5 A). However, a profound increase in the accumulation of triglycerides was observed in the liver with HFD from 24 wks onwards, indicating that triglyceride accumulation was secondary to the establishment of IR.

Similarly accumulation of lipid in muscle is believed to be a driver of IR in both skeletal and cardiac muscle (Lettner & Roden, 2008). In skeletal muscle a reduction in

2-[3H] DOG uptake was observed at 6 wks, and we have previously shown this to occur in vivo as early as 7 d (Chapter 3). Triglyceride content was unchanged in chow mice between 6 and 60 wks indicating age had no impact on triglyceride accumulation in

149

skeletal and cardiac muscle. In contrast, with a HFD, triglyceride content was significantly increased at 6, 24 and 60 wks in skeletal muscle and at 60 wks but not prior in cardiac muscle was observed (Fig 5.1 C).

Figure 5.5 – Assessing ectopic lipid deposition in liver and muscle. A) Liver, B) quadricep and B) heart tissue were excised from chow and HFD mice after 6, 24 and 60 wks of feeding. Tissue was snap frozen and powdered. Chloroform-methanol precipitation was then performed on a weighed amount (~30-40mgs). Triglycerides were resuspended in ethanol and measured using Roche Diagnostics Triglyceride Assay (#450032). Significance was calculated using Two-way ANOVAs with Sidaks correction for multiple comparisons; comparisons are to chow within a

150

time point unless indicated otherwise in figure, * = p <0.05, ** = p <0.01, *** = p <0.001, **** = p < 0.0001), data presented as mean ± SEM (n = 10 mice per group).

5.4.6 Hyperinsulinemia –In vivo assessment of insulin secretion and clearance

The observed elevated insulin levels may be the result of two mechanisms; an increase in insulin secretion or a reduction in insulin clearance. To examine insulin secretion, 60 wk chow and HFD mice were challenged with a single bolus of glucose (2g/kg FFM).

Glucose was given orally, which promotes the secretion of gut derived incretins, which in turn act on the pancreas, driving a robust insulin secretory response (Andrikopoulos,

Blair, Deluca, Fam, & Proietto, 2008; Fam et al., 2012). Blood glucose increased in both chow and HFD fed mice in response to a bolus of glucose (Fig 5.6 A), but peaked and remained higher in HFD mice indicating impaired glucose tolerance.

Simultaneously, blood was collected for measurement of insulin in serum (5.6 B). We detected a robust insulin secretory response in both chow and HFD mice. Overall, HFD mice secreted more insulin than chow mice, but once again had higher fasting levels.

Interestingly the level of response as indicated by the change in insulin between 0-15 mins was similar in both groups (Fig 5.6 C). From these studies we concluded that there was no defect in insulin secretion, and given that this was in response to a glucose load, that the pancreatic islets retain their glucose sensing abilities.

Elevated insulin levels in the HFD mice could also be explained by a reduction in the rate of insulin clearance. To assess whether there was a decrease in the rate of insulin clearance we measured the C-peptide/insulin ratio. C-peptide is secreted together with insulin in equimolar quantities. C-peptide has negligible hepatic clearance and can be used as a measure of total endogenous insulin secretion in the periphery. Conversely, the liver rapidly clears insulin and so peripheral blood measurements are representative of this clearance mechanism. The ratio of the two (C-peptide/insulin) provides a 151

measure of clearance (Tamaki et al., 2013), where an increase in the ratio represents an accelerated rate of clearance and a decrease indicates defective insulin clearance. Chow and HFD mice were challenged with an i.p. GTT (3g/kg FFM) and glucose, insulin and c-peptide were measured (Fig 5.6 D-G). Upon stimulation, glucose levels increased rapidly in chow and HFD mice, but were significantly greater in HFD mice (Fig 5.6 D).

Consistent with the OGTT, HFD mice secreted significantly more insulin (Fig 5.6 C) and therefore more c-peptide (Fig 5.6 D) than chow controls across all time points. We then calculated the c-peptide/insulin ratio and observed no significant difference in the ratio (Fig 5.6 E) between treatment groups. Based on these data we conclude that hyperinsulinemia in HFD mice was not the result of impaired insulin clearance.

5.4.7 An investigation in to pancreas morphology and βcell mass using mosaic microscopy

Because we observed such marked increases in circulating insulin we next sought to examine the state of the pancreas. To identify whether β cell expansion may underlie the

HFD induced hyperinsulinemia and to determine a time frame for any expansion we performed immunohistochemistry on pancreata at 6, 24 and 60 wks. In addition, we examined a small panel of adipokines and cytokines known to influence pancreas morphology and physiology.

We first examined the gross anatomy of the pancreas and found no difference in pancreata weight between chow and HFD fed mice across the time course (data not shown). Pancreata were then fixed and serial sectioned for morphological analysis (See methods chapter 2 for a detailed description). In addition, no difference in total pancreas area was detected between diet groups (Fig 5.7 A). While we measured no change in total number of islets (Fig 5.7 B) in HFD fed mice, average islets size was increased

152

with HFD feeding (fig 5.7 C) independent of total area, and this increase widened when normalised to total pancreas area (Fig 5.7 D). The increase in β-cell mass at 24 wks corresponded with an increase in islets positively stained for Ki67 (Fig 5.7 E) and an increase in circulating insulin (Fig 5.3 D-E), and although Ki67 staining was not observed at later time points the expanded cell mass was maintained.

The inability of fat to store excess energy may cause fat to accumulate in undesirable locations. Consistent with this we observed ectopic lipid deposition in skeletal and cardiac muscle. Because pancreata were fixed for immunohistochemistry we were unable to obtain triglyceride measurements, however we were able to quantify the infiltration of adipose tissue in pancreata of chow and HFD fed mice. Adipose tissue as a proportion of total area was unchanged in chow mice but was significantly increase with HFD from 24 wks onwards. Interestingly, in some cases we found islets were completely encircled by adipose tissue, and had no visible cell-cell contact with acinar tissue (Nb: For representative images of the mosaic microscopy and individual islets please see Appendix 1).

5.4.8 Serum cytokines and adipokines

The secretion profile of a number of secreted factors was altered during the progression of insulin resistance. We collected serum from 6, 24 and 60 wk chow and HFD fed mice for analysis (Table 5.1). Originally, we ran an extensive panel of cytokines & adipokines (MMHMAG-44K; Millipore). However, many of the analytes fell above or below the detectable range and we were unable to obtain sufficient sample for a second run. We did however successfully obtain information on Gastric inhibitory polypeptide

(GIP), c-peptide, Interleukin-6 (IL-6), Resistin, and Leptin. HFD had no effect on IL-6 levels until 60 wks where a significant increase was observed. C-peptide and GIP

153

serum levels were significantly increased at 24 and 60 wks of HFD feeding, but unchanged at 6 wks. In contrast, HFD resulted in a significant increase in both Resistin and Leptin serum levels across all time points which paralleled increases in adiposity

(Fig 5.1).

154

Figure 5.6 - Investigating in vivo insulin secretion and clearance. Sixty wk chow and HFD mice were fasted for 6 h, and then challenged with a 1hr oral glucose challenge (2g/kg FFM). A) Blood glucose and B) insulin were monitored at time points indicated and the C) change in insulin from 0-15 mins calculated. C peptide and insulin are secreted in equimolar quantities. C peptide has negligible hepatic insulin clearance and represents total endogenous insulin secretion, while insulin undergoes rapid hepatic clearance. Peripheral measurements of insulin therefore represent post cleared insulin levels. 60 wk chow and HFD fed mice were challenged with 3g/kg FFM and E) glucose (mM), F) Insulin (ng/ml) and G) C-peptide (ng/ml) were measured from blood or serum and the H) C- peptide/insulin ratio calculated. Significance was calculated using two-way ANOVAs with Sidaks correction for multiple comparisons; comparisons are to chow within a time point, * = p <0.05, ** = p <0.01, *** = p <0.001, **** = p < 0.0001), data presented as mean ± SEM (n = 10 mice per group). 155

Figure 5.7 – Pancreata and islet quantification with HFD feeding. Pancreata were excised and weighed at time points indicated, then fixed, embedded and mounted and stained with Dapi, insulin (β cells), glucagon (α cells), and Ki67. We then performed microscopy of the entire section and quantified A) total area (mm2). Images were then analysed and B) total number of islets per pancreas counted. C) Average islet area and E) islet area as a percentage of total pancreas area were calculated as an indication of β-cell mass. In addition F) number of islets 156

with positive Ki67 positive staining was counted as a measure of proliferation. H) Adipose tissue as a percentage of total area was measured. Significance was calculated using a Two- way ANOVAs with Sidaks correction for multiple comparisons; comparisons are to chow within a time point, * = p <0.05, ** = p <0.01, *** = p <0.001, **** = p < 0.0001), data presented as mean ± SEM (n = 3-4 mice per group).

157

6 weeks 24 weeks 60 weeks Chow HFD Chow HFD Chow HFD GIP 20.97 (±3.7) 62.03 (±12.2) * 21.35 (±4.2) 88.40 (±21.0) ** 57.36 (±7.5) 170.0 (±34.1) ** + C-peptide 688.6 (±111.0) 796.2 (±110.0) 1007 (±111.0) 1765 (±356.1) * 637.6 (±82.3) 1535 (±402.6) * IL-6 10.45 (±2.4) 16.58 (±4.2) 22.20 (±5.8) 15.38 (±3.3) 25.61 (±4.0) 77.5 (±15.7) ** ++ Resistin 9071 (±662) 12658 (±1073)* 7761 (±500) 12141 (±1593) * 6831 (±732) 13479 (±1721) *** Leptin 432.0 (±54.3) 2887 (±367.5)**** 1461 (±519.8) 14402 (±3449) **** +++ 1830 (±739.1) 15408 (±1228) **** +++

Table 5.1 – Serum panel of cytokines and adipokines . A panel of serum cytokine and adipokine content (pictograms/ml serum) from mice maintained on either a chow or HFD for six, twenty four, or sixty wks. Measurements included, GIP, C-peptide, IL-6, Resistin and Leptin. T tests compared chow to HFD at a single time point (*). Two-way ANOVAs with multiple comparisons (Sidaks) were performed to compare between time points. (+) compares diet to previous time point. * = p <0.05, ** = p <0.01, *** = p <0.001, **** = p < 0.0001, data presented as mean ± SEM, (n=5-10 mice per group).

158

5.4.9 Leptin potentiates glucose stimulated insulin secretion under HFD feeding conditions

The increase in circulating leptin paralleled an increase in β-cell mass (fig 5.7) and insulin secretion (Fig 5.6). While an increase in β-cell mass is associated with high leptin levels, leptin role in insulin secretion is inhibitory (Morioka et al., 2007). So we next set out to test the overall secretory capacity of HFD conditioned islets, and to clarify whether leptins inhibitory capacity was remained intact, rationalising that the co- observation of hyperleptinemia and increased insulin secretion may indicate leptin resistance within the β-cell.

In response to increasing concentrations of glucose (2-20 mM) we observed a robust secretory response which trended higher in HFD conditioned islets compared to chow mice (Fig 5.8 A). To further characterise the insulin secretory profile of chow and HFD conditioned islets, we also treated islets with glucose (11 mM) plus exenatide, a GLP-1 analogue. GLP-1 is an incretin secreted by the intestinal L-cells. Consistent with the increased insulin secretion observed during the OGTT (Fig 5.6 B) we detected a significant increase in secretion with HFD. We next stimulated islets with a first phase insulin secretion secratogue, KCl in the presence of glucose. KCl artificially depolarises the β cell membrane and mimics the glucose trigger via the amplification pathway. At

2 mM glucose + KCL, GSIS was only minimally stimulated providing a baseline, while the combination of 20 mM glucose + KCl resulted in maximal activation of both insulin secretory pathways (i.e. GSIS and amplification pathway), providing a measure of total insulin secretion which was similar in chow and HFD mice (Fig 5.8 A). These data indicate that HFD feeding did not significantly alter GSIS or insulin secretion via the amplification pathway, but the consistently higher insulin response in HFD conditioned islets may indicate a heightened responsiveness to glucose, which may contribute to an 159

overall hyperinsulinemic phenotype. Lastly we tested two concentrations of leptin, a low dose (2ng/ml) and a physiological dose consistent with circulating levels on 60 wk

HFD fed mice (20ng/ml). At the lowest dose, the insulin secretory response in chow and

HFD mice was similar to 11mM glucose without leptin, indicating leptin did not act to suppress insulin secretion. At a physiological dose (20ng/ml), consistent with current literature (Cantley, 2014) (Emilsson et al., 1997), leptin suppressed insulin secretion in chow mice by approximately 50% (Fig 5.8 A, B), but strikingly resulted in significant increase in insulin secretion with HFD. Improving insulin secretion by ~150% (Fig 8A,

B). Together these data indicate that HFD feeding resulted in an improved response to glucose, and that insulin secretion was neither supressed or resistant to the effects of leptin but rather leptin acted in a stimulatory manner.

160

Figure 5.8 – Glucose stimulated insulin secretion in 60 wk mice. A) Pancreatic islets were isolated from chow or HFD fed mice, and stimulated with either 2, 11 or 20 mM glucose and insulin secretion [(pg/ng)hr] was measured at 60 wks dietary intervention. To test the effect of GLP-1 on insulin secretion islets were incubated with 11mM glucose plus exenatide (Ex – 5ng/ml). Islets were also stimulated with potassium chloride (KCl- 25mM) at either 2 or 20 mM glucose. KCl artificially depolarises the β cell membrane and mimics the glucose trigger via the amplification pathway. Lastly to test the effect of leptin on insulin secretion islets were stimulated with 11mmol/l glucose and recombinant leptin at either 2 or 20 ng/ml. All secretion data was normalised to islet DNA content. B) Percent change in insulin secretion in response to 11 mM glucose & leptin (20 ng/ml) was calculated. Significance was calculated with T-tests and Two-way ANOVAs with Sidaks correction for multiple comparisons; comparisons are to chow, * = p <0.05, **** = p < 0.0001, data presented as mean ± SEM (n = 5 mice per treatment).

161

5.4.10 The effects of diet and age on skeletal architecture

It is well established that aging is strongly associated with a number of bone diseases.

Additionally individuals with insulin resistance or diabetes (Type I and II) have a higher incidence of bone diseases such as osteoporosis when compared to the population

(Devlin et al., 2014; Wongdee & Charoenphandhu, 2015). DEXA scanning was applied to isolated femoral bones to assess at a whole organ level bone mineral content (BMC), bone mineral density (BMD) and bone area (Fig 5.9 A-C). No significant difference in

BMC or BMD (Fig 5.9 A-B) was detected between chow and HFD mice prior to 24 wks. Both diet groups showed an age related decline in BMD and BMC between 24 to

60 wks, which was significantly worsened in mice on HFD. The decline in BMC and

BMD can be influenced by alterations in either bone area (Fig 5.9 C) or femur length

(Fig 5.9 D). However, both bone area and femur length were unchanged throughout the time course with either diet.

We then utilised Micro CT scanning to assess the internal structure of the femur.

Trabecular thickness was unchanged in chow mice indicating age had no effect (5.9 E).

In response to a HFD, trabecular thickness was reduced at 6 wks, increased at 24 wks and returned to baseline levels after 60 wks. In contrast to thickness, trabecular bone volume (Fig 5.9 F) and trabecular number (5.9 G) both showed a significant reduction with age in chow fed mice, however this reduction was significantly worsened in HFD fed mice at 6 (Fig 5.9 M-N), 24 (Fig 5.9 O-P) and 60 (Fig 5.9 Q-R) wks on diet, indicating that diet had a greater impact on bone degradation than aging alone.

Trabecular bone is organised into a web like structure and so we measured the space between the internal structures (i.e. trabecular separation) and found this to be significantly increased with age in chow mice, but again the distance between structures was significantly worsened in HFD mice, indicating that diet has an additional effect 162

above age (Fig 5.9 H). Representative images showing a loss in trabecular bone volume, trabecular number, trabecular separation and changes in cortical thickness when comparing chow and HFD fed mice at 6, 24 and 60 wks are provided (Fig 5.9 M-R,

Refer to Appendix 2 for enlarged images).

Micro CT scanning was also used to assess cortical bone (Fig 5.9 I-L). It is known that cortical bone is influenced by both activity and load, which become altered under HFD conditions. Cortical bone volume (Fig 5.9 I) and thickness (Fig 5.9 J) were altered similarly throughout the time course. Interestingly, both trabecular volume and thickness declined at between 24 and 60 wks in both diet groups. Because cortical bone can also change in diameter, we next measured periosteal and endosteal perimeter length. Periosteal and Endosteal perimeter length increased to a similar extent in both chow and HFD animals indicating the increase was driven by age (Fig 5.9 K-L). The external radius of bone is linked to bending strength, whereby the larger the perimeter the greater the strength. To see if the increase in perimeter resulted in a compensatory increase in bone strength we examined the average Moment of Inertia x (MMIx),

Moment of Inertia y (MMIy), the Maximum principle Moment of Inertia (MMI(max)) and the Polar Moment of Inertia (MMIp) (Fig 5.10 A-D), all descriptors of radius taken on various axis. Although we observed fluctuations in these measurements over time, in all cases and across all time points no difference between chow and high fat fed mice was detected. These data indicate that bending strength is similar between chow and

HFD mice.

163

Figure 5.9 – Assessing bone structure and formation. Femora were isolated from 6, 24 and 60 wk chow and HFD fed mice. DEXA scanning was utilised to measure femora A) BMC (g), B) BMD (g) and C) bone area (mm2). Callipers were used to generate D) femur length (mm). Micro-CT scanning was utilised to characterise internal bone structure. Cancellous bone was scanned and E) trabecular thickness (mm), F) trabecular bone volume/total volume (check units), G) 164

trabecular number and H) trabecular separation were assessed. Cortical bone was analysed and I) cortical bone volume (mm3) and J) cortical thickness were assessed. Additionally cortical K) Periosteal and L) Endosteal perimeters (mm) were measured. Representative images showing the loss in trabecular bone volume, and trabecular number and changes in cortical thickness when comparing chow and HFD fed mice at M-N) 6, O-P) 24 and Q-R) 60 wks, respectively. T-tests and Two-way ANOVAs with Sidaks correction for multiple comparisons; comparisons are to chow within a time point unless indicated otherwise in figure. * = p <0.05, ** = p <0.01, *** = p <0.001, **** = p < 0.0001, data presented as mean ± SEM, (n=10 mice per group).

Figure 5.10 – Assessing bone integrity. The moment of inertia is a measure of external bending strength of bone A) Average moments of inertia x, B) moment of inertia y, C) maximum principle moment of inertia, D) polar moment of inertia (r4) were calculated using the Skyscan CT analyser software as an indicator of bone strength in isolated femurs. Two-way ANOVAs with multiple comparisons (Sidaks) was performed; comparisons are to chow within a time point unless indicated otherwise in figure, * = p <0.05, ** = p <0.01, *** = p <0.001, **** = p < 0.0001) data presented as mean ± SEM, (n=10).

165

5.4.11 High fat diet increases susceptibility to neurodegenerative diseases

Endogenous Aβ plaque 40 and 42 were measured in 6, 24 and 60 wk chow and HFD fed mice. The accumulation of plaques, in particular Aβ 42 in humans is associated with the development of age-related neurodegenerative diseases such as Alzheimer’s disease

(AD), while the accumulation or maintenance of Aβ-40 levels is believed to be neuroprotective (Gu & Guo, 2013). Aβ-42 was unchanged in chow fed mice with increasing age (Fig 5.11 A) but was reduced in HFD mice after 6 wks on diet. The reduction was sustained for 24 wks on the diet but was significantly reduced at 60 wks on diet.

Similarly the neuro-protective Aβ-40 was unchanged in chow fed mice with increasing age (Fig 5.11 B). Aβ-40 levels in HFD mice trended downwards with increasing increments at 6, 24 and 60 wks on diet, reaching a statistically significant reduction at

60 wks compared to either 60 wk chow fed mice or to 6 wk HFD fed mice. Recently it was shown that the Aβ-40/42 ratio is an indicator of an individual’s susceptibility to

AD, where a reduction in the ratio increases the chance of developing AD (Spies et al.,

2010). The Aβ-40/42 is unchanged in chow mice indicating no change in disease susceptibility with age (Fig 5.11 C). Conversely a significant reduction in the Aβ-40/42 ratio between 6 and 60 wk HFD fed mice was observed, indicating that these mice are more prone to developing neurological disorders such as AD.

A pathological hallmark of Alzheimer’s disease is the accumulation of Aβ plaques in the hippocampus, and in its early stages is associated with short term memory loss (X.

Li et al., 2015a). We next sought to obtain a direct measure of brain function in mice, and conducted a series of behavioural tests that would examine locomotion, anxiety, and short term memory. We first performed an open field test as a measure of general activity allowing us to establish a baseline for movement. We observed no difference in total distance travelled between 60 wk chow and HFD fed mice (Fig 5.11 D). As a 166

measure of anxiety during the open field test we also recorded time spent in the centre of the apparatus (Fig 5.11 E). Although there was a downward trend in HFD fed mice, indicating HFD mice are more anxious this failed to reach statistically significance.

Mice were then placed in an elevated plus maze which is designed to test specifically for anxiety (Fig 5.11 F) by measuring the time spent in the open arm versus the closed arms. A significant reduction in time spent in the open arm was detected indicating a heightened level of anxiousness with HFD feeding but not age.

As previously mentioned, short term memory is a hallmark of dementias such as AD

(X. Li et al., 2015a). The Y-Maze is specifically designed to measure short term working memory. Mice are placed at the centre of the maze, and allowed to explore.

The sequence and total number of arms entered is recorded and the total number of arms entered and the percentage of alterations calculated (see Chapter 2 - materials and methods for calculations). No significant difference was observed in the total number of arm entries between chow and HFD fed mice (Fig 5.11 G) however a significant reduction in the number of spontaneous alterations was observed in HFD mice. It is well known that spontaneous alternation is a measure of spatial working memory and these results indicate a short term memory deficit in HFD fed mice (Fig 5.11 H).

167

Figure 5.11 - Aβ plaque load and memory testing in long term HFD C57BL6 mice. Endogenous Aβ plaque A) 42 and B) 40 were measured via ELISA and the C) Aβ-40/42 ratio calculated by dividing Aβ-40 by Aβ-42. To assess if a pathological change in plaque load resulted in altered brain function we next performed behavioral testing on 60 wk mice. The open field test provided a measure of D) locomotion (total distance travelled (cm)) and E) anxiety (Time in center (Sec). F) Elevated plus maze directly measured anxiety (time in open arm (secs)) and lastly the Y-maze provided a measured of G) movement and H) short term spatial memory. Two-way ANOVAs with multiple comparisons (Sidaks) was performed; comparisons are to

168

chow within a time point unless indicated otherwise in figure, * = p <0.05, ** = p <0.01, *** = p <0.001, **** = p < 0.0001) data presented as mean ± SEM, (n=5-10 mice per group).

5.4.12 assessing organismal health via the liver proteome

As previously discussed the liver is one of the first tissues to develop IR under HFD conditions. The liver is highly susceptible to a number of diseases including cancer

(Healy et al., 2015), hepatic steatosis and NAFLD of which obesity, IR and age are major risk factors (Nakamura & Terauchi, 2013). As IR develops from a very early age and persists for the duration of the study this also makes the liver an ideal tissue to compare the effects of HFD to age.

We first took a macro perspective and as a measure of liver health first performed histology on whole liver sections. Consistent with triglyceride data previously shown

(Fig 5), histological examination using H&E staining revealed a significant accumulation of lipid at 24 and 60 wks with HFD feeding, while changes in lipid accumulation or morphology with age in chow fed mice was completely absent (Fig

5.12).

To assess global liver function and health we next performed proteomics in chow and

HFD mice at 6, 24 and 60 wks, and performed a differential expression analysis, to identify proteins that were significantly altered with either diet or age (Fig 5.13).

Strikingly when we compared the liver proteome of 6 wk Chow to 60 wk Chow mice to test for the effect of age we observed only 14 differentially expressed (DE) proteins (Fig

5.13 A). However, when we compared the proteome of 60 wk chow mice to 60 wk HFD mice, 155 proteins were differentially expressed (Fig 5.13 B). In view of these data we decided to focus on the effect of diet at each time point. Differential expression analysis is useful for identifying relatively large changes in individual proteins. However, it is

169

known that most proteins are co-regulated and move in clusters. In addition, relatively small changes cumulatively could alter pathways in a significant manner; hence we used a gene set enrichment test (GST) believing this will provide a better readout for overall health. We used the KEGG database as this already contains an extensive list of highly curated proteins that are organised in to pathways containing proteins that are known to be co-regulated in a range of biological output, but in particularly disease. Significantly altered pathways in response to a HFD for 6, 24 and 60 wks were then used as a readout for hepatic and organismal health (Fig 5.14)

Within the list of DE proteins we observed changes in proteins that belonged to a range of pathways. These included pathways altered across all time points, pathways transiently altered at 6 wks or 6 month and those pathways only altered after prolonged periods on HFD (Fig 5.14). Six pathways were altered at all-time points in response to

HFD: extracellular matrix (ECM) receptor interaction and ubiquitin mediated proteolysis pathways were upregulated, lysine degradative proteins were differentially changed over the time course, while pathways associated with Lysosome, drug and xenobiotic metabolism, cytochrome p450 and Glutathione metabolism were downregulated.

One advantage of temporal data is the opportunity to reveal pathways that were transiently altered in response to a perturbation. Here, we observed pathways only changed at 6 wks, only changed at 24 wks or changed at both time points but not at 60 wks. These included a number of pathways involved in processes such as glucose/fatty acid metabolism (fatty acid metabolism, pentose phosphate pathway, citrate cycle, glycolysis/gluconeogenesis, pyruvate metabolism), amino acid metabolism

(phenylalanine, alanine, glutamate, aspartate, arginine, proline, glycine, serine,

170

threonine, histidine, valine, leucine, isoleuncine, lycine), folate metabolism (folate biosynthesis, and one carbon pool by folate), primary bile acid biosynthesis, and

DNA/RNA processing (spliceosome, RNA degradation, nucleotide excision repair, pyrimidine/purine metabolism).

Strikingly the largest number of pathways were altered at 60 wks, and many of these were unique to this time point. We examined these pathways for indices of liver health.

We observed changes in pathways involved in ceramide biosynthesis and proteins associated with neurodegenerative diseases (Alzheimers, Parkinsons and Huntingtons disease), cardiovascular disease (Vascular smooth muscle contraction, hypertrophic cardiomyopathy, dilated cardiomyopathy and arrhythmogenic right ventricular cardiomyopathy) and Diabetes. In addition we also observed the downregulation in important signaling pathways including MAPK and JAK/STAT.

The most extraordinary feature was the number of pathways associated with inflammation. These included a number of pathways involved in the immune response

(T and B cell receptor signaling and natural killer cell mediated cytotoxicity) and inflammatory diseases (Asthma, Leishmania infection, staphylococcus aureus infection and Lupus). Moreover these inflammatory related pathways were specific to the 60 wk time point and consistent with this we observed a significant increase in spleen size

(data not shown) that was unique to 6 wk HFD fed mice.

171

Figure 5.12 – Visual assessment of liver morphology and lipid accumulation. Livers from 6, 24 and 60 wk (A-C) chow and (D-E) HFD mice were fixed in 10% buffered formaldehyde for 24 hours and transferred to ethanol before being wax embedded. 5µm sections were cut and mounted on glass slides before H&E staining was performed to assess liver integrity and lipid content. Entire liver sections were examined using mosaic microscopy and processed in Image J.

172

Figure 5.13 - Differentially expressed proteins with age and diet. Single shot proteomics was combined with filter aided sample preparation (FASP) on livers from chow and HFD fed mice at 6, 24 and 60 wks. MS spectra were analysed by MaxQuant. Differential expression (DE) analysis was performed using moderated t-tests to identify altered proteins altered. Data was visualised using volcano plots displaying the number DE proteins that changed with A) age (Comparison: 60w Chow – 6w Chow, purple) and B) diet (Comparison: 60w HFD – 60w Chow, red). Proteins were designated DE if they had a Log fold change >1.5(left of 0=down, right of 0= up) and a P-value < 0.05.

173

174

Figure 5.14 - Liver proteomics and gene set enrichment test. Livers were excised from mice treated with a chow or HFD for 6, 24 or 60 wks. Tissue was snap frozen, powdered, weighed and proteins extracted using the SDT solution (2% SDS, 100mM Tris/HCl pH 7.5 + protease inhibitors). Protein was digested with trypsin and peptides were isolated using FASP. Peptides were then desalted using C18 Stage tipping. Single shot proteomics was combined with filter aided sample preparation (FASP) on livers from chow and HFD fed mice at 6, 24 and 60 wks). MS spectra were analysed by MaxQuant. Differential expression (DE) analysis was performed using moderated t-tests. Proteins that were significantly changed (P<0.05) and had a fold change >1.5 with with diet at each time point were subjected to a gene set enrichment analysis. Pathways from the KEGG database were tested for unchanged (grey), up (blue), down (Red) and mixed (Green) regulation. A p-value significance of 0.05 was used to identify significantly altered pathways.

175

5.5 Discussion

Obesity and related diseases are increasing at an alarming rate, and we do not yet fully understand the impact of a western diet on long term health. Furthermore, human obesity is often a lifelong phenomenon (Babu, Sonnenblick, & Gulati, 1991; Birch &

Ventura, 2009; Sinha & Kling, 2009) and it is therefore important to understand the long-term metabolic characteristics as the obesity syndrome develops. Lastly, obesity may actually be laying the foundation for a number of other diseases which are often labelled as age associated. Here we have designed a study to investigate the relative contribution of diet and age by conducting a comprehensive temporal analysis of normal chow or HFD feeding on a range of parameters that reflect many of the diseases of the modern world for up to 60 wks. Consistent with previous reports (Winzell &

Ahren, 2004), we show that C57BL6 mice placed on a HFD display rapid and consistent increases in adiposity, stable hyperglycaemia, and glucose intolerance.

Intriguingly, a similar degree of glucose intolerance was maintained for at least 24 wks on the HFD. However, underpinning this relative steady state of glucose intolerance was a steady increase in fasting and postprandial insulin levels. Beyond 24 wks on the diet, adiposity and hyperinsulinemia continued to increase while glucose tolerance began to resolve. Strikingly, after mice had been fed the diet for 52 wks glucose tolerance was indistinguishable from the chow fed animals, which was primarily driven by increased β-cell proliferation and a further increase in circulating insulin.

Preliminary studies indicate that this may be due to an unexpected stimulatory effect of leptin on insulin secretion. Interestingly, improvements in glucose tolerance were not accompanied by increased glucose disposal as peripheral tissues remained insulin resistant. In addition, the deleterious effects of long term HFD feeding extended beyond

176

metabolism and resulted in marked deterioration in skeletal architecture, brain function and liver health; deficits that were not observed with age.

The deleterious effects of a HFD on metabolism are well documented (Buettner et al.,

2007; Hariri & Thibault, 2010; West & York, 1998). In rodents, HFD causes a rapid induction of IR and glucose intolerance. In the current study, we tracked glucose tolerance over different periods of high fat feeding revealing several novel observations.

Firstly, consistent with previous reports (N. Turner et al., 2013), mice rapidly became glucose intolerant, occurring as early as 1 d on the diet (Chapter 3). Secondly, although we observed a minor worsening in glucose tolerance after 6-12 wks on the diet, glucose tolerance remained relatively stable until 24 wks. An increase in adiposity and lipid in muscle and liver was observed over the same time frame but did not precede the establishment of IR. Lastly, from 24 wks onwards we observed a resolution in glucose intolerance, which was accompanied by a reduction in fasting glycaemia and a further increase in hyperinsulinemia. These observations are interesting for a number of reasons. First, in humans it has been suggested that the progression of disease follows a path where the initial phase involves IR with compensatory hyperinsulinemia and completely normal glucose tolerance. The second phase involves β-cell loss followed by glucose intolerance and ultimately diabetes. It has always been somewhat surprising that many animal models of IR such as the HFD mouse are characterised initially by glucose intolerance which seems at odds with the proposed schema in humans. Based on our data it is tempting to speculate that possibly the initial phase of glucose intolerance due to pancreatic insufficiency is a phase that has been overlooked in humans. Regardless of whether this is a mouse specific phenotype these data suggest that mice fed HFD from 24 wks onwards is better aligned with the preclinical stages of

T2D in humans and this may provide a better working model to study IR. Secondly, 177

these data emphasise the potential limitations of using GTT as a sole measure of whole body insulin sensitivity because measurement of GTT alone in mice fed HFD for 1 year could be construed as indicating that these animals display normal insulin sensitivity.

However, this ‘normal’ glucose tolerance was achieved in the face of a 2 fold increase in fasting insulin and a 4 fold increase in postprandial insulin. Thus, it is clear that these animals are still significantly insulin resistant but the IR is overcome by a profound increase in insulin secretion. Consistent with this, we observed that insulin stimulated glucose uptake into both adipose tissue and skeletal muscle was significantly impaired in animals fed HFD for 60 wks and this defect was similar to that observed after 6 wks of HFD. Thus it is unlikely that the compensatory hyperinsulinemia is related to some kind of adaptation in peripheral insulin sensitivity as this appeared to remain fairly constant between 6 and 60 wks.

The primary question arising from the above observations is what triggered the compensatory increase in circulating insulin levels? We observed no difference in insulin clearance between long term (60 wk) HFD or chow fed animals, suggesting that this effect was likely due to changes in insulin secretion per se. Consistent with this, we observed an increase in β-cell mass and circulating insulin which paralleled improvements in glucose tolerance . However, what triggered the increase in β-cell mass after animals had been fed the HFD for such a long period of time? A number of systemic factors are known to modulate β-cell hyperplasia and insulin secretion

(Cantley, 2014; W. Kim & Egan, 2008). These include, FABP4 (Wu, Samocha-Bonet, et al., 2014), Adiponectin (Okamoto et al., 2008), Resistin (Nakata, Okada, Ozawa, &

Yada, 2007), Nesfastin-1 (Gonzalez et al., 2011; Nakata, Manaka, Yamamoto, Mori, &

Yada, 2011), glucagon-like peptide-1 (GLP-1) (Vasu, Moffett, Thorens, & Flatt, 2014),

Gastric inhibitory peptide (GIP) and glucose (Weir & Bonner-Weir, 2007). Indeed, both 178

GIP, and glucose were elevated at 24 wks. We also observed a large increase in circulating leptin levels at 24 wks, when we first observed increased β-cell mass, increased insulin secretion and improved glucose tolerance.

Leptin is an adipokine secreted in proportion to adiposity (Amitani, Asakawa, Amitani,

& Inui, 2013) and has been reported to regulate of β-cell function and mass (Covey et al., 2006; Marroqui et al., 2012). Consistent with our data in chow fed mice, much of the literature points towards leptin being a negative regulator of beta cell mass and

GSIS and insulin gene expression (Covey et al., 2006; Emilsson et al., 1997), (J. Chen et al., 2013; Kulkarni et al., 1999; Kulkarni et al., 1997; Morioka et al., 2007; Seufert et al., 1999). In contrast, leptin potentiated GSIS in mice fed on a HFD, suggesting that leptin may play a role in driving hyperinsulinemia in our model. Although unexpected, the context/diet-dependence of the leptin response in beta cells has previously been alluded to in the β-cell specific leptin receptor KO mouse (Morioka et al., 2007). Here loss of the leptin R from beta cells resulted in enhanced insulin secretion under chow conditions, but blunted insulin secretion in HFD fed mice. Taken together these data highlight a novel role for leptin in driving insulin secretion under specific dietary conditions and understanding the differences in insulin and leptin signalling (Marroqui et al., 2012) in the β-cell under these different conditions may further our understanding of the signals driving hyperinsulinemia.

While it is clear that metabolic defects can extend to multiple systems the role of diet per se is not clear. Large scale epidemiological studies have established that individuals with T2D have an increased predisposition to a number of diseases including cardiovascular disease, some cancers, osteoporosis and dementia (e.g. Alzheimer’s disease (AD)) above unaffected population. However a direct link between metabolic

179

disease and associated diseases has been difficult to establish as it has been difficult to separate diet from age as well as other factors including genetics (ethnicity and familial history) and environment. In the current study design we were able to directly quantify the direct contribution of diet and age towards the diseases listed above. We first examined the skeletal architecture of chow and HFD mice and observed significant loss in trabecular (cancellous) bone with age that was further compromised with the addition of a HFD, findings consistent with the pathology of osteoporosis in humans. These findings are consistent with previous reports demonstrating that HFD feeding leads to a loss of trabecular but not cortical bone (J. J. Cao et al., 2009; J. J. Cao et al., 2010) after

14 wks, and it has further been shown that losses such as these are irreversible (Inzana et al., 2013). However, in the current study we also show that the loss of bone continues despite a restoration of glucose tolerance, indicating a disconnect between glucose homeostasis and bone formation. Both insulin and leptin contribute to bone formation and extensive investigations have shown that both insulin and leptin resistance play a role in the loss of bone associated with diet-induced obesity (Hamrick, Pennington,

Newton, Xie, & Isales, 2004; Klein, 2014; Wauman & Tavernier, 2011). It is noteworthy that obesity was thought to be protective as the increased load in overweight individuals leads to an increase BMD (Ferrari, 2013; van Daele et al., 1995). However, these same individuals have an increased fracture risk calling in to question the quality and strength of bone (Caffarelli et al., 2014; Compston et al., 2011; Veldhuis-Vlug et al., 2013). In HFD fed mice we observed no improvement in bone strength (as measured by the MMI(polar)), and a reduction in BMD and BMC in comparison to chow mice.

These findings in combination with a severe loss in trabecular bone indicate a highly compromised and weakened skeletal architecture. Moreover, it would appear that HFD feeding results in the formation of pathology consistent with osteoporosis and in part

180

may explain why obese (de Paula, Horowitz, & Rosen, 2010) and IR individuals are predisposed to osteoporosis more so than the general population (Caffarelli et al., 2014;

Xia et al., 2012).

In keeping with the theme that IR extends beyond glucose homeostasis we also explored the effect of HFD feeding on neurological disorders, as IR and dysregulated glucose homeostasis have been associated with development of AD (Arvanitakis et al., 2004;

Craft et al., 1998; Dineley et al., 2014; Talbot et al., 2012; van Himbergen et al., 2012).

To date, studies examining the interaction between AD and metabolic disease using

HFDs have utilised transgenic lines (e.g. 3xTgAD) designed to induce dementia via the accumulation of plaques, allowing for a more distinct pathology and enlarged window of detection (Vandal & Calon, 2015). Using the non-transgenic C57BL/6 HFD fed mice developed an altered Aβ 40/42 ratio consistent with the pathology of AD (Spies et al.,

2010). This was functionally represented by a loss of short term memory, and increased anxiety, both hallmarks of the cognitive impairment that accompanies AD. While one other study has shown that HFD feeding can lead to a reduction in short term memory after 15 months in C57BL/6 mice, they were unable to relate this to either glucose tolerance, IR, or plaque load, and used a HFD (60% kcal from fat) that was not representative of a western diet (E. M. Knight et al., 2014). Here we show that HFD exacerbates the susceptibility and pathology to diseases such as dementia independently of glucose tolerance (Takalo et al., 2014), and despite reports indicating that treatment with insulin can restore cognitive function and reverse the effects of HFD feeding (Craft et al., 2012; Vandal et al., 2014) these effects occurred in the presence of hyperinsulinemia, peripheral and most likely central IR (Arnold et al., 2014).

181

We lastly examined the liver as a surrogate measure of organismal health. Despite the fact that metabolism had established a new steady state, there were clear indices of disease progression within the liver. Histologically we observed a significant amount of lipid accumulation in the liver, a precursor to the formation of NAFLD, and a significantly altered proteome that was driven by diet but not age. More specifically, there were a number of changes in the proteome that were unique to the 6, 24 and 60 wk time points, indicating that the proteome is extremely flexible and dynamic.

Interestingly, only a few pathways were observed to be changed across all time points, and these by-and-large appeared to be stress related such as Glutathione metabolism, lysososome, cytochrome P450, proteasome, and the ECM, although it could be argued that given these change across all time points they are more related to IR. There were also a number of changes that are more strongly linked to IR which changed at 6 and

24, but were resolved by 60 wks, such as, primary bile acid biosynthesis, amino acid and fatty acid metabolism. Given that these resolve by 60 wks, it is unlikely they are involved in late disease processes and may be a response to a relative lack of insulin prior to 24 wks or are involved in the onset of disease. Strikingly, the largest number of altered pathways occurred at 60 wks with HFD, in diseases associated with IR including neurodegeneration, CVD, inflammatory diseases and diabetes. These only appeared at the 60 wk time point, indicating that prolonged HFD feeding is a precursor that exacerbates the development of disease above the effect of aging, and may in part explain the association between IR, T2D and a number of age related diseases. Future studies will need to examine whether these proteomic changes translate in to functional outputs leading to the formation of disease.

In this study we observed improvements in glucose tolerance, which was accompanied by compensatory hyperinsulinemia and persistent peripheral tissue IR. For the first time 182

to our knowledge we present a model of diet-induced obesity that closely mimics the pathology and etiology of human obesity and IR. The hyperinsulinemia was the result of a marked increase in β cell mass which underlined the improvements in glucose tolerance, with leptin implicated as a potentiating factor. Moreover we set out to quantify the effect of diet on long term health compared to aging and showed that the effects of HFD feeding extends beyond glucose metabolism and participates in the formation of disease processes in bone, brain and liver, despite improvements in glucose tolerance. The implication for an obese, aging western population will be a large increase in disease burden and brings to the forefront the need for effective therapies targeting obesity and IR.

183

General discussion

184

6.1 General discussion

Obesity has reached worldwide pandemic proportions. This rise in obesity rates is attributed to sedentary lifestyles and the availability of foods that are rich in fat. Obesity is a risk factor for a number of diseases including, cancer, NAFLD, atherosclerosis,

CVD, AD, nephropathy, osteoporosis, IR and T2D.

It is widely held that insulin resistance is one of the earliest defects that occur on the road to many of these diseases but the dynamic interplay between changes in lifestyle, the onset of insulin resistance followed by disease is not known. Our goal was to follow the dynamic change in metabolism as well as other health parameters in response to a dietary perturbation to better understand disease progression. The most striking observation was that over the course of one year exposure to HFD the metabolic landscape was constantly shifting. This approach enabled us to unravel a range of metabolic adaptations that have previously gone undetected due to the use of more short term static approaches. Most notably, animals developed glucose intolerance almost immediately upon shifting to the new diet and a similar degree of glucose intolerance was maintained for 24 wks of dietary intervention. This suggests that a new set point for gluco-regulation was achieved in C57BL/6 mice fed this diet, which was well tolerated for a considerable period of time. So unlike the situation in ‘at risk’ humans, these animals displayed an insulin secretory defect whereby insulin release was unable to maintain normal glucose tolerance in the face of reduced peripheral IR. Importantly, from about 6-60 wks there was a steady (almost linear) increase in whole body adiposity, as well as circulating levels of insulin and leptin. This was somewhat surprising because one would have assumed that these animals would have eventually reached a new steady state in terms of some of these parameters but this was not the

185

case. Notably, in parallel long term HFD animals also exhibited a significant reduction in activity. As to whether this is also related to some sort of neurological adaptation remains to be determined but what is most interesting is that this would likely contribute to a reduction in energy expenditure. This is completely counterintuitive, at least as far as obtaining a beneficial metabolic outcome on this new diet. Thus in this case the diet is deleterious to health in at least two ways because of the increased caloric intake as well as the concomitant reduction in energy expenditure. It will be of major interest to determine the molecular basis for the reduced activity, the contribution of reduced activity per se to overall metabolic health and as to whether this effect is also seen with different kinds of diets that also cause IR. Not surprisingly we observed an increase in leptin, which was directly proportional to the degree of adiposity. The gradual increase in insulin however was more surprising and for this reason we focused more on what might be driving this adaptation.

One of the more surprising observations from this study was the profound resolution of glucose intolerance due for the most part to an adaptive increase in β-cell function and insulin secretion. Both fasting and postprandial insulin levels were elevated by 2-4 fold in long term HFD animals. The basal increase was of major interest because this occurred in the face of normal fasting glucose levels so this is likely not a function of glucose-stimulated secretion. We were also able to verify these changes in vitro by measuring GSIS in isolated islets. These studies revealed that insulin secretion was remarkably similar between chow and HFD fed animals in response to 2, 11 and 20 mM glucose. Using this same system we next assessed the role of leptin in insulin secretion and surprisingly observed a potentiating effect of leptin on insulin secretion only in

HFD fed animals. This indicates that prolonged exposure to HFD leads to some kind of adaptation that primes the ability of leptin to stimulate insulin release. This was 186

interesting as it is widely believed that obesity in most individuals is accompanied by a state of leptin resistance and this is one of the main reasons leptin has failed to be therapeutically useful at a population level. In fact previous studies of leptin action in the β-cell have shown that leptin acts to inhibit insulin secretion, insulin transcription and restrict β-cell mass expansion, (Kulkarni et al., 1999; Kulkarni et al., 1997; Morioka et al., 2007; Seufert, 2004; Seufert et al., 1999). Our data, however suggest that as is the case for IR (Tan et al., 2015) obesity may be characterised by ‘selective’ leptin resistance whereby the central effects of leptin are blocked while some of its peripheral effects such as those on the β-cell may in fact be augmented. This is potentially a very important observation that requires further investigation. As to why this differs from previous studies of leptin action in β-cells is perhaps not surprising as these studies were based on the use of leptin receptor knock out mice, whereas our studies were performed in a non-transgenic and more physiologically relevant model. Future work will be required to examine the prolonged effects of HFD on β-cell leptin receptor KO mice as well as examining leptin signalling in β-cells in chow versus HFD fed mice to determine the molecular nature of the adaptation. It was notable that the pancreata of long term HFD (24 and 60 wk) fed mice were characterised by the infiltration of fat. It might be of interest to characterise these adipocytes to determine if they represent visceral or subcutaneous fat cell or even some other kind of fat cell. Moreover, it may be of interest to characterise the adipocyte associated islets to determine whether their contact with adipocytes has altered their responsiveness to adipokines. Finally, it may be interesting to repeat this long term feeding regime in a heterozygous Ob mouse, and test whether the β-cell expansion and hyperinsulinemia would still occur and what the effect of leptin administration might have in these mice.

187

In addition to examining metabolism throughout the thesis, we also assessed other health outcomes. The effect of HFD has been shown to affect multiple tissues (e.g. bone, brain, liver, heart) and be involved in multiple disease processes (e.g. cancer, diabetes, low-grade inflammation) but no studies have examined the effect on metabolism and health in the same mouse. We assessed the health of the liver, bone and brain in response to HFD feeding. Despite an improvement in glucose tolerance and a relatively static peripheral IR, we found that prolonged HFD feeding resulted in significant bone loss, pathological hallmarks of AD, cognitive impairment, and alterations in the liver proteome that would indicate the initiation of a number of disease processes, which were not observed in old chow mice. Based on these studies we conclude that diet has a significant impact on health over and above that of aging alone.

This is particularly concerning with the aging population now intersecting with an obesity epidemic. Given the importance of the liver in the formation of IR and potentially disease, future studies should focus on obtaining more functional measures of the disease indicated by the liver proteome, to examine whether these will truly manifest. In addition, it would be interesting to investigate what role insulin plays in the development of these diseases, as elevated insulin and IR have previously been implicated in the formation of a whole range of diseases. One way to test this would be to make chow fed mice hyperinsulinemic using osmotic mini-pumps and again examine the effects on bone, brain and liver. Alternatively, we could make mice insulin deficient using a low dose of Streptozotocin and repeat the same feeding regime and examine whether the same defects in bone, brain and liver occur again.

There are many models of diet-induced obesity (Buettner et al., 2007; Bunner et al.,

2014; Hariri & Thibault, 2010; A. W. Lee & Cox, 2011)and a number of diets that influence metabolism and long term health (Solon-Biet et al., 2014). In terms of HFD 188

feeding, many studies are performed using different percentages of fat as there is currently no standardised protocol for establishing IR in mice. Indeed, many of the studies implicating diets high in fat to disease (e.g. cancer, diabetes, low-grade inflammation) use extremely high percentage fat diets (i.e. >60%), whereas we used a much lower percentage (45%). Because of the variability in diets used it is difficult to find a unified consensus on the role HFD plays in disease, and whether or not these effects would be observed on lower percentage diets. In light of the recent studies that have utilised different mouse strains and varied diets (Berglund et al., 2008;

Montgomery et al., 2013; Solon-Biet et al., 2014; West & York, 1998), in order to fully elucidate the influence of diet on long term health it would be of interest to study both metabolism and a range of disease processes in different strains of mice fed different styles of diets. This would enable us to correlate the diverse metabolic responses to disease outcomes so that we can start to identify the key drivers of the interactions between diet and aging diseases.

Despite the fact that glucose tolerance gradually improved in the mice fed HFD for 60 wks, which may be indicative of improved health, underlying this was a plethora of changes and persistent IR. We observed significant deficits in the skeletal architecture of bones from long term HFD fed mice. In bone, insulin acts as an anabolic hormone, promoting osteoblast differentiation, and enhancing the production of osteocalcin, which is involved in bone building (Ferron & Lacombe, 2014; Veldhuis-Vlug et al.,

2013). Osteocalcin’ secretion is further regulated by leptin, and both leptin resistance and IR have been implicated in the loss of bone (Klein, 2014; Wee & Baldock, 2014).

We also observed a striking neurological phenotype, which mimicked the pathology of

AD. Again, dysfunction of insulin and IR has been implicated in AD (De Felice,

Lourenco, & Ferreira, 2014). Although we were unable to directly link these findings in 189

brain and bone to circulating insulin levels in our mice, it would be of interest to speculate that some of these changes may be due to the compensatory hyperinsulinemia, given that the insulin receptor is expressed ubiquitously (Desbuquois et al., 1993). Thus, elevated insulin may drive many of these changes we observed via IR or through other essential pathways regulated by insulin to maintain cellular order (e.g. regulation of autophagy). To address this, as mentioned above, it would interesting in the future to repeat these studies in animals that have been treated with a low dose of Streptozotocin to prevent the compensatory hyperinsulinemia. We propose that these animals would still develop IR but be spared from many other health defects.

Many of the findings in this study were the result of incorporating a temporal element in to the study design. In chapter 3, this allowed us to identify the speed and order at which tissues developed IR (liver 1d, WAT 3 d, muscle 5 d). In chapter 4, it allowed us to identify how long the effects of VEGF neutralisation could be sustained, and in chapter

5, to pinpoint the moment at which the phenotype transitioned from glucose intolerance to tolerance. More importantly however, is that we can align these adaptations with the causes of IR. One such cause of IR is ectopic lipid deposition in muscle and liver

(Lettner & Roden, 2008). In all three studies, we observed a disconnect between the initiation of IR and the accumulation of lipid in tissues, as increases in triglycerides were detected subsequent to the onset/initiation of IR. This indicates that there may be alternate factors also driving IR. Other factors such as inflammation (Lumeng &

Saltiel, 2011), hypoxia (Rutkowski, Stern, & Scherer, 2015) and alternate lipid species including ceramides and DAG have also been implicated in the formation of IR(Chavez et al., 2005; Chavez & Summers, 2003; Griffin et al., 1999; Muoio & Newgard, 2008), which should also be measured in future studies.

190

In the context of a HFD, an expanding adipose tissue would aid in avoiding the deposition of lipid in undesired tissues, however this has also been linked with an increase in IR driven by insufficient angiogenesis and hypoxia. We tested whether limiting the expansion of adipose tissue and thus hypoxia using an anti-angiogenic antibody could inhibit the formation of IR. However, we found that improvements in glucose tolerance were principally due to improved hepatic insulin sensitivity, and not an improvement in WAT. It was beyond the scope of our work, but it would be interesting to identify the mechanism by which VEGF acted in the liver. Additional data presented in this thesis highlights the importance of the liver, which appears to be the initial site for development of IR in the HFD fed mouse. It would also be valuable to incorporate the use of the hyperinsulinemic-euglycemic clamp into the longitudinal studies so that a measure of liver IR could be paralleled with the changing metabolic landscape such as the improvement in glucose tolerance we observed. Perhaps moving forward more therapeutic studies should take aim at mitigating the formation of hepatic

IR as the first line of defence against metabolic diseases.

In this thesis we have presented a model, which mimics the pathology and etiology of human IR. Moreover, it provides a useful tool that can be used to determine the drivers of hyperinsulinemia and to assess the effects of hyperinsulinemia and IR on health and wellbeing. This model provides extraordinary insights into the pervasiveness of the effects of diet on health in bone, brain and liver above that of aging alone. Based on these associations it will be important to extend these studies into other systems as we have only just begun to scratch the surface of how diet impacts on organismal health.

191

Appendices & References

192

7.1 Appendix 1

Figure I – Representative Pancreata Sections and Islets from six week Chow-Fed mice Pancreata from six week chow –fed mice were fixed and serial sectioned for morphological analysis. A) And B) Representative pancreas sections from 2 individual six week chow-fed mice stained for Nuclei (Dapi-Blue), Insulin (Red) and Glucagon (Green). Scale bar = 1mm. C) To E) Examples of islet morphology showing Nuclei (Blue), Insulin (Red) and Glucagon (Green); Scale Bar = 50mm. F) And G) Examples of islet morphology - Nuclei (blue), Insulin (green) and KI-67 (red); Scale Bar = 50mm.

193

Figure II – Representative Pancreata Sections and Islets from six week HFD-Fed mice Pancreata from six week HFD –fed mice were fixed and serial sectioned for morphological analysis. A) And B) Representative pancreas sections from 2 individual Six week HFD-fed mice stained for Nuclei (Dapi-Blue), Insulin (Red) and Glucagon (Green). Scale bar = 1mm. C) And D) Examples of islet morphology showing Nuclei (Blue), Insulin (Red) and Glucagon (Green); Scale Bar = 50mm. E) And F) Examples of islet morphology - Nuclei (blue), Insulin (green) and KI-67 (red); Scale Bar = 50mm. G)Pancreas associated adipose tissue. -Nuclei (blue), Insulin (green) and KI-67 (red).

194

Figure III– Representative Pancreata Sections and Islets from twenty-four week Chow-Fed mice Pancreata from twenty-four week chow –fed mice were fixed and serial sectioned for morphological analysis. A) And B) Representative pancreas sections from 2 individual twenty-four week chow-fed mice stained for Nuclei (Dapi-Blue), Insulin (Red) and Glucagon (Green). Scale bar = 1mm. C) And D) Examples of islet morphology showing Nuclei (Blue), Insulin (Red) and Glucagon (Green); Scale Bar = 50mm. E) And F) Examples of islet morphology - Nuclei (blue), Insulin (green) and KI-67 –Red); Scale Bar = 50mm. G) Pancreas associated adipose tissue. -Nuclei (blue), Insulin (green) and KI-67 (red).

195

Figure IV – Representative Pancreata Sections and Islets from twenty-four week HFD-Fed mice Pancreata from twenty-four week HFD –fed mice were fixed and serial sectioned for morphological analysis. A) And B) Representative pancreas sections from 2 individual Twenty-four week HFD-fed mice stained for Nuclei (Dapi-Blue), Insulin (Red) and Glucagon (Green). Scale bar = 1mm. C) Example of islet morphology showing Nuclei (Blue), Insulin (Red) and Glucagon (Green); Scale Bar = 50mm. D) Example of islet morphology (Large Islet) - Nuclei (blue), Insulin (green) and KI-67 (red); Scale Bar = 50mm. E) Example of adipose associated Islet. -Nuclei (blue), Insulin (green) and KI-67 (red). F) Example of islet morphology showing Nuclei (Blue), Insulin (Red) and Glucagon (Green); Scale Bar = 50mm.

196

Figure V – Representative Pancreata Sections and Islets from sixty week Chow-Fed mice Pancreata from sixty week chow –fed mice were fixed and serial sectioned for morphological analysis. A) And B) Representative pancreas sections from 2 individual sixty week chow-fed mice stained for Nuclei (Dapi-Blue), Insulin (Red) and Glucagon (Green). Scale bar = 1mm. C) And D) Examples of islet morphology showing Nuclei (Blue), Insulin (Red) and Glucagon (Green); Scale Bar = 50mm. E) And G) Examples of islet morphology - Nuclei (blue), Insulin (green) and KI-67 –Red); Scale Bar = 50mm. F)Pancreas associated adipose tissue. -Nuclei (blue), Insulin (green) and KI-67 (red).

197

Figure VI – Representative Pancreata Sections and Islets from sixty week HFD-Fed mice Pancreata from sixty week HFD –fed mice were fixed and serial sectioned for morphological analysis. A) and B) Representative pancreas sections from 2 individual Sixty week HFD-fed mice stained for Nuclei (Dapi- Blue), Insulin (Red) and Glucagon (Green). Scale bar = 1mm. C) Adipose associated Islets. Nuclei (Blue), Insulin (Red) and Glucagon (Green); Scale Bar = 50mm. D) Examples of islet morphology -Nuclei (blue), Insulin (green) and KI-67 (red)); Scale Bar = 50mm. E) Examples of islet morphology showing Nuclei (Blue), Insulin (Red) and Glucagon (Green); Scale Bar = 50mm.

198

7.2 Appendix 2

Figure I – Representative images of bone from 6 wk fed mice. The distal end of the femur from 6 wk A) chow or B) HFD fed mice was scanned using micro-computed tomography (µCT) with a Skyscan 1172 scanner.

199

Figure I – Representative images of bone from 24 wk fed mice. The distal end of the femur from 24 wk A) chow or B) HFD fed mice was scanned using micro-computed tomography (µCT) with a Skyscan 1172 scanner.

200

Figure I – Representative images of bone from 60 wk fed mice. The distal end of the femur from 60 wk A) chow or B) HFD fed mice was scanned using micro-computed tomography (µCT) with a Skyscan 1172 scanner.

201

7.3 References

Abel, E. D., Peroni, O., Kim, J. K., Kim, Y. B., Boss, O., Hadro, E., Minnemann, T., Shulman, G. I., & Kahn, B. B. (2001). Adipose-selective targeting of the GLUT4 gene impairs insulin action in muscle and liver. Nature, 409(6821), 729-733. doi: 10.1038/35055575 Abu-Elheiga, L., Oh, W., Kordari, P., & Wakil, S. J. (2003). Acetyl-CoA carboxylase 2 mutant mice are protected against obesity and diabetes induced by high-fat/high- carbohydrate diets. Proc Natl Acad Sci U S A, 100(18), 10207-10212. doi: 10.1073/pnas.1733877100 Accili, D. (2004). Lilly lecture 2003: the struggle for mastery in insulin action: from triumvirate to republic. Diabetes, 53(7), 1633-1642. Accili, D., Drago, J., Lee, E. J., Johnson, M. D., Cool, M. H., Salvatore, P., Asico, L. D., Jose, P. A., Taylor, S. I., & Westphal, H. (1996). Early neonatal death in mice homozygous for a null allele of the insulin receptor gene. Nat Genet, 12(1), 106-109. doi: 10.1038/ng0196-106 Adams, J. M., 2nd, Pratipanawatr, T., Berria, R., Wang, E., DeFronzo, R. A., Sullards, M. C., & Mandarino, L. J. (2004). Ceramide content is increased in skeletal muscle from obese insulin-resistant humans. Diabetes, 53(1), 25-31. Ahren, B., & Pacini, G. (2002). Insufficient islet compensation to insulin resistance vs. reduced glucose effectiveness in glucose-intolerant mice. Am J Physiol Endocrinol Metab, 283(4), E738-744. doi: 10.1152/ajpendo.00199.2002 Ahren, J., Ahren, B., & Wierup, N. (2010). Increased beta-cell volume in mice fed a high-fat diet: a dynamic study over 12 months. Islets, 2(6), 353-356. doi: 10.4161.islets.2.6.13619 AIHW. (2013). Who is overweight? , from http://www.aihw.gov.au/who-is-overweight/ Alessi, M. C., Poggi, M., & Juhan-Vague, I. (2007). Plasminogen activator inhibitor-1, adipose tissue and insulin resistance. Curr Opin Lipidol, 18(3), 240-245. doi: 10.1097/MOL.0b013e32814e6d29 Almind, K., & Kahn, C. R. (2004). Genetic determinants of energy expenditure and insulin resistance in diet-induced obesity in mice. Diabetes, 53(12), 3274-3285. Amitani, M., Asakawa, A., Amitani, H., & Inui, A. (2013). The role of leptin in the control of insulin-glucose axis. Front Neurosci, 7, 51. doi: 10.3389/fnins.2013.00051 An, J., Muoio, D. M., Shiota, M., Fujimoto, Y., Cline, G. W., Shulman, G. I., Koves, T. R., Stevens, R., Millington, D., & Newgard, C. B. (2004). Hepatic expression of malonyl- CoA decarboxylase reverses muscle, liver and whole-animal insulin resistance. Nat Med, 10(3), 268-274. doi: 10.1038/nm995 Anai, M., Funaki, M., Ogihara, T., Kanda, A., Onishi, Y., Sakoda, H., Inukai, K., Nawano, M., Fukushima, Y., Yazaki, Y., Kikuchi, M., Oka, Y., & Asano, T. (1999). Enhanced insulin-stimulated activation of phosphatidylinositol 3-kinase in the liver of high-fat- fed rats. Diabetes, 48(1), 158-169. Andrikopoulos, S., Blair, A. R., Deluca, N., Fam, B. C., & Proietto, J. (2008). Evaluating the glucose tolerance test in mice. Am J Physiol Endocrinol Metab, 295(6), E1323-1332. doi: 10.1152/ajpendo.90617.2008 Andrikopoulos, S., Massa, C. M., Aston-Mourney, K., Funkat, A., Fam, B. C., Hull, R. L., Kahn, S. E., & Proietto, J. (2005). Differential effect of inbred mouse strain (C57BL/6, DBA/2, 129T2) on insulin secretory function in response to a high fat diet. J Endocrinol, 187(1), 45-53. doi: 10.1677/joe.1.06333 Antuna-Puente, B., Feve, B., Fellahi, S., & Bastard, J. P. (2008). Adipokines: the missing link between insulin resistance and obesity. Diabetes Metab, 34(1), 2-11. doi: 10.1016/j.diabet.2007.09.004 202

Arnold, S. E., Lucki, I., Brookshire, B. R., Carlson, G. C., Browne, C. A., Kazi, H., Bang, S., Choi, B. R., Chen, Y., McMullen, M. F., & Kim, S. F. (2014). High fat diet produces brain insulin resistance, synaptodendritic abnormalities and altered behavior in mice. Neurobiol Dis, 67, 79-87. doi: 10.1016/j.nbd.2014.03.011 Aroor, A. R., Mandavia, C. H., & Sowers, J. R. (2012). Insulin resistance and heart failure: molecular mechanisms. Heart Fail Clin, 8(4), 609-617. doi: 10.1016/j.hfc.2012.06.005 Arvanitakis, Z., Wilson, R. S., Bienias, J. L., Evans, D. A., & Bennett, D. A. (2004). Diabetes mellitus and risk of Alzheimer disease and decline in cognitive function. Arch Neurol, 61(5), 661-666. doi: 10.1001/archneur.61.5.661 Aslesen, R., Engebretsen, E. M., Franch, J., & Jensen, J. (2001). Glucose uptake and metabolic stress in rat muscles stimulated electrically with different protocols. J Appl Physiol (1985), 91(3), 1237-1244. Babu, A., Sonnenblick, E., & Gulati, J. (1991). Calcium and strontium activation characteristics of muscle fibres: a comment. J Muscle Res Cell Motil, 12(1), 125; discussion 125-126. Bachmanov, A. A., Reed, D. R., Tordoff, M. G., Price, R. A., & Beauchamp, G. K. (2001). Nutrient preference and diet-induced adiposity in C57BL/6ByJ and 129P3/J mice. Physiol Behav, 72(4), 603-613. Baetens, D., Stefan, Y., Ravazzola, M., Malaisse-Lagae, F., Coleman, D. L., & Orci, L. (1978). Alteration of islet cell populations in spontaneously diabetic mice. Diabetes, 27(1), 1-7. Bagnol, D., Al-Shamma, H. A., Behan, D., Whelan, K., & Grottick, A. J. (2012). Diet- induced models of obesity (DIO) in rodents. Curr Protoc Neurosci, Chapter 9, Unit 9 38 31-13. doi: 10.1002/0471142301.ns0938s59 Banks, W. A., Coon, A. B., Robinson, S. M., Moinuddin, A., Shultz, J. M., Nakaoke, R., & Morley, J. E. (2004). Triglycerides induce leptin resistance at the blood-brain barrier. Diabetes, 53(5), 1253-1260. Banks, W. A., & Farrell, C. L. (2003). Impaired transport of leptin across the blood-brain barrier in obesity is acquired and reversible. Am J Physiol Endocrinol Metab, 285(1), E10-15. doi: 10.1152/ajpendo.00468.2002 Berglund, E. D., Li, C. Y., Poffenberger, G., Ayala, J. E., Fueger, P. T., Willis, S. E., Jewell, M. M., Powers, A. C., & Wasserman, D. H. (2008). Glucose metabolism in vivo in four commonly used inbred mouse strains. Diabetes, 57(7), 1790-1799. doi: 10.2337/db07-1615 Birch, L. L., & Ventura, A. K. (2009). Preventing childhood obesity: what works? Int J Obes (Lond), 33 Suppl 1, S74-81. doi: 10.1038/ijo.2009.22 Black, B. L., Croom, J., Eisen, E. J., Petro, A. E., Edwards, C. L., & Surwit, R. S. (1998). Differential effects of fat and sucrose on body composition in A/J and C57BL/6 mice. Metabolism, 47(11), 1354-1359. Bluher, M. (2014). Adipokines - removing road blocks to obesity and diabetes therapy. Mol Metab, 3(3), 230-240. doi: 10.1016/j.molmet.2014.01.005 Bluher, M., Michael, M. D., Peroni, O. D., Ueki, K., Carter, N., Kahn, B. B., & Kahn, C. R. (2002). Adipose tissue selective insulin receptor knockout protects against obesity and obesity-related glucose intolerance. Dev Cell, 3(1), 25-38. Bonomini, F., Rodella, L. F., & Rezzani, R. (2015). Metabolic syndrome, aging and involvement of oxidative stress. Aging Dis, 6(2), 109-120. doi: 10.14336/AD.2014.0305 Brakenhielm, E., Cao, R., Gao, B., Angelin, B., Cannon, B., Parini, P., & Cao, Y. (2004). Angiogenesis inhibitor, TNP-470, prevents diet-induced and genetic obesity in mice. Circ Res, 94(12), 1579-1588. doi: 10.1161/01.RES.0000132745.76882.70

203

01.RES.0000132745.76882.70 [pii] Bruning, J. C., Michael, M. D., Winnay, J. N., Hayashi, T., Horsch, D., Accili, D., Goodyear, L. J., & Kahn, C. R. (1998). A muscle-specific insulin receptor knockout exhibits features of the metabolic syndrome of NIDDM without altering glucose tolerance. Mol Cell, 2(5), 559-569. Buettner, R., Newgard, C. B., Rhodes, C. J., & O'Doherty, R. M. (2000). Correction of diet- induced hyperglycemia, hyperinsulinemia, and skeletal muscle insulin resistance by moderate hyperleptinemia. Am J Physiol Endocrinol Metab, 278(3), E563-569. Buettner, R., Parhofer, K. G., Woenckhaus, M., Wrede, C. E., Kunz-Schughart, L. A., Scholmerich, J., & Bollheimer, L. C. (2006). Defining high-fat-diet rat models: metabolic and molecular effects of different fat types. J Mol Endocrinol, 36(3), 485- 501. doi: 10.1677/jme.1.01909 Buettner, R., Scholmerich, J., & Bollheimer, L. C. (2007). High-fat diets: modeling the metabolic disorders of human obesity in rodents. Obesity (Silver Spring), 15(4), 798- 808. doi: 10.1038/oby.2007.608 Bugger, H., & Abel, E. D. (2009). Rodent models of diabetic cardiomyopathy. Dis Model Mech, 2(9-10), 454-466. doi: 10.1242/dmm.001941 Bugger, H., & Abel, E. D. (2014). Molecular mechanisms of diabetic cardiomyopathy. Diabetologia, 57(4), 660-671. doi: 10.1007/s00125-014-3171-6 Bujalska, I. J., Kumar, S., & Stewart, P. M. (1997). Does central obesity reflect "Cushing's disease of the omentum"? Lancet, 349(9060), 1210-1213. doi: 10.1016/S0140- 6736(96)11222-8 Bunner, A. E., Chandrasekera, P. C., & Barnard, N. D. (2014). Knockout mouse models of insulin signaling: Relevance past and future. World J Diabetes, 5(2), 146-159. doi: 10.4239/wjd.v5.i2.146 Burcelin, R., Crivelli, V., Dacosta, A., Roy-Tirelli, A., & Thorens, B. (2002). Heterogeneous metabolic adaptation of C57BL/6J mice to high-fat diet. Am J Physiol Endocrinol Metab, 282(4), E834-842. doi: 10.1152/ajpendo.00332.2001 Burchfield, J. G., Lu, J., Fazakerley, D. J., Tan, S. X., Ng, Y., Mele, K., Buckley, M. J., Han, W., Hughes, W. E., & James, D. E. (2013). Novel systems for dynamically assessing insulin action in live cells reveals heterogeneity in the insulin response. Traffic, 14(3), 259-273. doi: 10.1111/tra.12035 Butterfield, D. A., Di Domenico, F., & Barone, E. (2014). Elevated risk of type 2 diabetes for development of Alzheimer disease: a key role for oxidative stress in brain. Biochim Biophys Acta, 1842(9), 1693-1706. doi: 10.1016/j.bbadis.2014.06.010 Caballero, B. (2007). The global epidemic of obesity: an overview. Epidemiol Rev, 29, 1-5. doi: 10.1093/epirev/mxm012 Caffarelli, C., Alessi, C., Nuti, R., & Gonnelli, S. (2014). Divergent effects of obesity on fragility fractures. Clin Interv Aging, 9, 1629-1636. doi: 10.2147/CIA.S64625 Calligaris, S. D., Lecanda, M., Solis, F., Ezquer, M., Gutierrez, J., Brandan, E., Leiva, A., Sobrevia, L., & Conget, P. (2013). Mice long-term high-fat diet feeding recapitulates human cardiovascular alterations: an animal model to study the early phases of diabetic cardiomyopathy. PLoS One, 8(4), e60931. doi: 10.1371/journal.pone.0060931 Cantley, J. (2014). The control of insulin secretion by adipokines: current evidence for adipocyte-beta cell endocrine signalling in metabolic homeostasis. Mamm Genome, 25(9-10), 442-454. doi: 10.1007/s00335-014-9538-7 Cantley, J., Burchfield, J. G., Pearson, G. L., Schmitz-Peiffer, C., Leitges, M., & Biden, T. J. (2009). Deletion of PKCepsilon selectively enhances the amplifying pathways of glucose-stimulated insulin secretion via increased lipolysis in mouse beta-cells. Diabetes, 58(8), 1826-1834. doi: 10.2337/db09-0132

204

Cantley, J., Choudhury, A. I., Asare-Anane, H., Selman, C., Lingard, S., Heffron, H., Herrera, P., Persaud, S. J., & Withers, D. J. (2007). Pancreatic deletion of insulin receptor substrate 2 reduces beta and alpha cell mass and impairs glucose homeostasis in mice. Diabetologia, 50(6), 1248-1256. doi: 10.1007/s00125-007-0637-9 Cao, J. J. (2011). Effects of obesity on bone metabolism. J Orthop Surg Res, 6, 30. doi: 10.1186/1749-799X-6-30 Cao, J. J., Gregoire, B. R., & Gao, H. (2009). High-fat diet decreases cancellous bone mass but has no effect on cortical bone mass in the tibia in mice. Bone, 44(6), 1097-1104. doi: 10.1016/j.bone.2009.02.017 Cao, J. J., Sun, L., & Gao, H. (2010). Diet-induced obesity alters bone remodeling leading to decreased femoral trabecular bone mass in mice. Ann N Y Acad Sci, 1192, 292-297. doi: 10.1111/j.1749-6632.2009.05252.x Cao, Y. (2007). Angiogenesis modulates adipogenesis and obesity. J Clin Invest, 117(9), 2362-2368. doi: 10.1172/JCI32239 Cao, Y. (2010). Adipose tissue angiogenesis as a therapeutic target for obesity and metabolic diseases. Nat Rev Drug Discov, 9(2), 107-115. doi: nrd3055 [pii]

10.1038/nrd3055 Carpenter, B., Lin, Y., Stoll, S., Raffai, R. L., McCuskey, R., & Wang, R. (2005). VEGF is crucial for the hepatic vascular development required for lipoprotein uptake. Development, 132(14), 3293-3303. doi: 10.1242/dev.01902 Cefalu, W. T. (2006). Animal models of type 2 diabetes: clinical presentation and pathophysiological relevance to the human condition. ILAR J, 47(3), 186-198. Chavez, J. A., Holland, W. L., Bar, J., Sandhoff, K., & Summers, S. A. (2005). Acid ceramidase overexpression prevents the inhibitory effects of saturated fatty acids on insulin signaling. J Biol Chem, 280(20), 20148-20153. doi: 10.1074/jbc.M412769200 Chavez, J. A., & Summers, S. A. (2003). Characterizing the effects of saturated fatty acids on insulin signaling and ceramide and diacylglycerol accumulation in 3T3-L1 adipocytes and C2C12 myotubes. Arch Biochem Biophys, 419(2), 101-109. Chen, H. C., Stone, S. J., Zhou, P., Buhman, K. K., & Farese, R. V., Jr. (2002). Dissociation of obesity and impaired glucose disposal in mice overexpressing acyl coenzyme a:diacylglycerol acyltransferase 1 in white adipose tissue. Diabetes, 51(11), 3189- 3195. Chen, J., Fu, R., Cui, Y., Li, Y. S., Pan, J. R., Liu, J. L., Luo, H. S., Yin, J. D., Li, D. F., & Cui, S. (2013). LIM-homeodomain transcription factor Isl-1 mediates the effect of leptin on insulin secretion in mice. J Biol Chem, 288(17), 12395-12405. doi: 10.1074/jbc.M113.450536 Chen, W. S., Xu, P. Z., Gottlob, K., Chen, M. L., Sokol, K., Shiyanova, T., Roninson, I., Weng, W., Suzuki, R., Tobe, K., Kadowaki, T., & Hay, N. (2001). Growth retardation and increased apoptosis in mice with homozygous disruption of the Akt1 gene. Genes Dev, 15(17), 2203-2208. doi: 10.1101/gad.913901 Chicco, A., D'Alessandro, M. E., Karabatas, L., Pastorale, C., Basabe, J. C., & Lombardo, Y. B. (2003). Muscle lipid metabolism and insulin secretion are altered in insulin- resistant rats fed a high sucrose diet. J Nutr, 133(1), 127-133. Chisholm, D. J., Campbell, L. V., & Kraegen, E. W. (1997). Pathogenesis of the insulin resistance syndrome (syndrome X). Clin Exp Pharmacol Physiol, 24(9-10), 782-784. Chung, W. K., Power-Kehoe, L., Chua, M., Lee, R., & Leibel, R. L. (1996). Genomic structure of the human OB receptor and identification of two novel intronic microsatellites. Genome Res, 6(12), 1192-1199. Claes, F., Vandevelde, W., Moons, L., & Tjwa, M. (2011). Another angiogenesis-independent role for VEGF: SDF1-dependent cardiac repair via cardiac stem cells. Cardiovasc Res, 91(3), 369-370. doi: 10.1093/cvr/cvr184 205

Cogger, V. C., Svistounov, D., Warren, A., Zykova, S., Melvin, R. G., Solon-Biet, S. M., O'Reilly, J. N., McMahon, A. C., Ballard, J. W., De Cabo, R., Le Couteur, D. G., & Lebel, M. (2014). Liver aging and pseudocapillarization in a Werner syndrome mouse model. J Gerontol A Biol Sci Med Sci, 69(9), 1076-1086. doi: 10.1093/gerona/glt169 Coleman, D. L. (1973). Effects of parabiosis of obese with diabetes and normal mice. Diabetologia, 9(4), 294-298. Coleman, D. L., & Hummel, K. P. (1967). Studies with the mutation, diabetes, in the mouse. Diabetologia, 3(2), 238-248. Coleman, D. L., & Hummel, K. P. (1969). Effects of parabiosis of normal with genetically diabetic mice. Am J Physiol, 217(5), 1298-1304. Coleman, D. L., & Hummel, K. P. (1973). The influence of genetic background on the expression of the obese (Ob) gene in the mouse. Diabetologia, 9(4), 287-293. Commerford, S. R., Bizeau, M. E., McRae, H., Jampolis, A., Thresher, J. S., & Pagliassotti, M. J. (2001). Hyperglycemia compensates for diet-induced insulin resistance in liver and skeletal muscle of rats. Am J Physiol Regul Integr Comp Physiol, 281(5), R1380- 1389. Compston, J. E., Watts, N. B., Chapurlat, R., Cooper, C., Boonen, S., Greenspan, S., Pfeilschifter, J., Silverman, S., Diez-Perez, A., Lindsay, R., Saag, K. G., Netelenbos, J. C., Gehlbach, S., Hooven, F. H., Flahive, J., Adachi, J. D., Rossini, M., Lacroix, A. Z., Roux, C., Sambrook, P. N., Siris, E. S., & Glow, Investigators. (2011). Obesity is not protective against fracture in postmenopausal women: GLOW. Am J Med, 124(11), 1043-1050. doi: 10.1016/j.amjmed.2011.06.013 Cook, K. S., Min, H. Y., Johnson, D., Chaplinsky, R. J., Flier, J. S., Hunt, C. R., & Spiegelman, B. M. (1987). Adipsin: a circulating serine protease homolog secreted by adipose tissue and sciatic nerve. Science, 237(4813), 402-405. Covey, S. D., Wideman, R. D., McDonald, C., Unniappan, S., Huynh, F., Asadi, A., Speck, M., Webber, T., Chua, S. C., & Kieffer, T. J. (2006). The pancreatic beta cell is a key site for mediating the effects of leptin on glucose homeostasis. Cell Metab, 4(4), 291- 302. doi: 10.1016/j.cmet.2006.09.005 Craft, S., Baker, L. D., Montine, T. J., Minoshima, S., Watson, G. S., Claxton, A., Arbuckle, M., Callaghan, M., Tsai, E., Plymate, S. R., Green, P. S., Leverenz, J., Cross, D., & Gerton, B. (2012). Intranasal insulin therapy for Alzheimer disease and amnestic mild cognitive impairment: a pilot clinical trial. Arch Neurol, 69(1), 29-38. doi: 10.1001/archneurol.2011.233 Craft, S., Peskind, E., Schwartz, M. W., Schellenberg, G. D., Raskind, M., & Porte, D., Jr. (1998). Cerebrospinal fluid and plasma insulin levels in Alzheimer's disease: relationship to severity of dementia and apolipoprotein E genotype. Neurology, 50(1), 164-168. Craft, S., & Watson, G. S. (2004). Insulin and neurodegenerative disease: shared and specific mechanisms. Lancet Neurol, 3(3), 169-178. doi: 10.1016/S1474-4422(04)00681-7 De Felice, F. G., Lourenco, M. V., & Ferreira, S. T. (2014). How does brain insulin resistance develop in Alzheimer's disease? Alzheimers Dement, 10(1 Suppl), S26-32. doi: 10.1016/j.jalz.2013.12.004 de la Monte, S. M. (2009). Insulin resistance and Alzheimer's disease. BMB Rep, 42(8), 475- 481. de Paula, F. J., Horowitz, M. C., & Rosen, C. J. (2010). Novel insights into the relationship between diabetes and osteoporosis. Diabetes Metab Res Rev, 26(8), 622-630. doi: 10.1002/dmrr.1135 De Vogel-van den Bosch, J., van den Berg, S. A., Bijland, S., Voshol, P. J., Havekes, L. M., Romijn, H. A., Hoeks, J., van Beurden, D., Hesselink, M. K., Schrauwen, P., & van Dijk, K. W. (2011). High-fat diets rich in medium- versus long-chain fatty acids

206

induce distinct patterns of tissue specific insulin resistance. J Nutr Biochem, 22(4), 366-371. doi: 10.1016/j.jnutbio.2010.03.004 DeFronzo, R. A. (1997). Insulin resistance: a multifaceted syndrome responsible for NIDDM, obesity, hypertension, dyslipidaemia and atherosclerosis. Neth J Med, 50(5), 191-197. Denner, L. A., Rodriguez-Rivera, J., Haidacher, S. J., Jahrling, J. B., Carmical, J. R., Hernandez, C. M., Zhao, Y., Sadygov, R. G., Starkey, J. M., Spratt, H., Luxon, B. A., Wood, T. G., & Dineley, K. T. (2012). Cognitive enhancement with rosiglitazone links the hippocampal PPARgamma and ERK MAPK signaling pathways. J Neurosci, 32(47), 16725-16735a. doi: 10.1523/JNEUROSCI.2153-12.2012 Desbuquois, B., Tozzo, E., Collinet, M., Lopez, S., Bortoli, S., & Amessou, M. (1993). [Regulation of insulin receptor expression and its gene]. Ann Endocrinol (Paris), 54(6), 373-384. Devlin, M. J., Van Vliet, M., Motyl, K., Karim, L., Brooks, D. J., Louis, L., Conlon, C., Rosen, C. J., & Bouxsein, M. L. (2014). Early-onset type 2 diabetes impairs skeletal acquisition in the male TALLYHO/JngJ mouse. Endocrinology, 155(10), 3806-3816. doi: 10.1210/en.2014-1041 Dimitri, P., Jacques, R. M., Paggiosi, M., King, D., Walsh, J., Taylor, Z. A., Frangi, A. F., Bishop, N., & Eastell, R. (2015). Leptin may play a role in bone microstructural alterations in obese children. J Clin Endocrinol Metab, 100(2), 594-602. doi: 10.1210/jc.2014-3199 Dineley, K. T., Jahrling, J. B., & Denner, L. (2014). Insulin resistance in Alzheimer's disease. Neurobiol Dis, 72 Pt A, 92-103. doi: 10.1016/j.nbd.2014.09.001 Ebato, C., Uchida, T., Arakawa, M., Komatsu, M., Ueno, T., Komiya, K., Azuma, K., Hirose, T., Tanaka, K., Kominami, E., Kawamori, R., Fujitani, Y., & Watada, H. (2008). Autophagy is important in islet homeostasis and compensatory increase of beta cell mass in response to high-fat diet. Cell Metab, 8(4), 325-332. doi: 10.1016/j.cmet.2008.08.009 El-Haschimi, K., Pierroz, D. D., Hileman, S. M., Bjorbaek, C., & Flier, J. S. (2000). Two defects contribute to hypothalamic leptin resistance in mice with diet-induced obesity. J Clin Invest, 105(12), 1827-1832. doi: 10.1172/JCI9842 Elias, I., Franckhauser, S., & Bosch, F. (2013). Response to Comment on: Elias et al. Adipose Tissue Overexpression of Vascular Endothelial Growth Factor Protects Against Diet- Induced Obesity and Insulin Resistance. Diabetes 2012;61:1801-1813. Diabetes, 62(1), e4. doi: 10.2337/db12-1274 Elias, I., Franckhauser, S., Ferre, T., Vila, L., Tafuro, S., Munoz, S., Roca, C., Ramos, D., Pujol, A., Riu, E., Ruberte, J., & Bosch, F. (2012). Adipose tissue overexpression of vascular endothelial growth factor protects against diet-induced obesity and insulin resistance. Diabetes, 61(7), 1801-1813. doi: 10.2337/db11-0832 Elmasri, H., Karaaslan, C., Teper, Y., Ghelfi, E., Weng, M., Ince, T. A., Kozakewich, H., Bischoff, J., & Cataltepe, S. (2009). Fatty acid binding protein 4 is a target of VEGF and a regulator of cell proliferation in endothelial cells. FASEB J, 23(11), 3865-3873. doi: fj.09-134882 [pii]

10.1096/fj.09-134882 Emilsson, V., Liu, Y. L., Cawthorne, M. A., Morton, N. M., & Davenport, M. (1997). Expression of the functional leptin receptor mRNA in pancreatic islets and direct inhibitory action of leptin on insulin secretion. Diabetes, 46(2), 313-316. Fain, J. N., Madan, A. K., Hiler, M. L., Cheema, P., & Bahouth, S. W. (2004). Comparison of the release of adipokines by adipose tissue, adipose tissue matrix, and adipocytes from visceral and subcutaneous abdominal adipose tissues of obese humans. Endocrinology, 145(5), 2273-2282. doi: 10.1210/en.2003-1336

207

Fam, B. C., Rose, L. J., Sgambellone, R., Ruan, Z., Proietto, J., & Andrikopoulos, S. (2012). Normal muscle glucose uptake in mice deficient in muscle GLUT4. J Endocrinol, 214(3), 313-327. doi: 10.1530/JOE-12-0032 Fantin, V. R., Wang, Q., Lienhard, G. E., & Keller, S. R. (2000). Mice lacking insulin receptor substrate 4 exhibit mild defects in growth, reproduction, and glucose homeostasis. Am J Physiol Endocrinol Metab, 278(1), E127-133. Farrell, G. C., Teoh, N. C., & McCuskey, R. S. (2008). Hepatic microcirculation in fatty liver disease. Anatomical record, 291(6), 684-692. doi: 10.1002/ar.20715 Fazakerley, D. J., Naghiloo, S., Chaudhuri, R., Koumanov, F., Burchfield, J. G., Thomas, K. C., Krycer, J. R., Prior, M. J., Parker, B. L., Murrow, B. A., Stockli, J., Meoli, C. C., Holman, G. D., & James, D. E. (2015). Proteomic Analysis of GLUT4 Storage Vesicles Reveals Tumour Suppressor Candidate 5 (TUSC5) as a Novel Regulator of Insulin Action in Adipocytes. J Biol Chem. doi: 10.1074/jbc.M115.657361 Fellmann, L., Nascimento, A. R., Tibirica, E., & Bousquet, P. (2013). Murine models for pharmacological studies of the metabolic syndrome. Pharmacol Ther, 137(3), 331- 340. doi: 10.1016/j.pharmthera.2012.11.004 Fenton, P. F., & Carr, C. J. (1951). The nutrition of the mouse. XI. Response of four strains to diets differing in fat content. J Nutr, 45(2), 225-233. Fenton, P. F., & Dowling, M. T. (1953). Studies on obesity. I. Nutritional obesity in mice. J Nutr, 49(2), 319-331. Ferrari, S. (2013). [Diabetes and osteoporosis]. Rev Med Suisse, 9(390), 1256, 1258-1259. Ferron, M., & Lacombe, J. (2014). Regulation of energy metabolism by the skeleton: osteocalcin and beyond. Arch Biochem Biophys, 561, 137-146. doi: 10.1016/j.abb.2014.05.022 Finley, S. D., Engel-Stefanini, M. O., Imoukhuede, P. I., & Popel, A. S. (2011). Pharmacokinetics and pharmacodynamics of VEGF-neutralizing antibodies. BMC Syst Biol, 5, 193. doi: 10.1186/1752-0509-5-193 Flier, J. S., Cook, K. S., Usher, P., & Spiegelman, B. M. (1987). Severely impaired adipsin expression in genetic and acquired obesity. Science, 237(4813), 405-408. Forest, C., Tordjman, J., Glorian, M., Duplus, E., Chauvet, G., Quette, J., Beale, E. G., & Antoine, B. (2003). Fatty acid recycling in adipocytes: a role for glyceroneogenesis and phosphoenolpyruvate carboxykinase. Biochem Soc Trans, 31(Pt 6), 1125-1129. doi: 10.1042/ Friedman, J. M. (2000). Obesity in the new millennium. Nature, 404(6778), 632-634. doi: 10.1038/35007504 Fuh, G., Wu, P., Liang, W. C., Ultsch, M., Lee, C. V., Moffat, B., & Wiesmann, C. (2006). Structure-function studies of two synthetic anti-vascular endothelial growth factor Fabs and comparison with the Avastin Fab. J Biol Chem, 281(10), 6625-6631. doi: 10.1074/jbc.M507783200 Galic, S., Oakhill, J. S., & Steinberg, G. R. (2010). Adipose tissue as an endocrine organ. Mol Cell Endocrinol, 316(2), 129-139. doi: 10.1016/j.mce.2009.08.018 Galuska, D., Ryder, J., Kawano, Y., Charron, M. J., & Zierath, J. R. (1998). Insulin signaling and glucose transport in insulin resistant skeletal muscle. Special reference to GLUT4 transgenic and GLUT4 knockout mice. Adv Exp Med Biol, 441, 73-85. Garcia de la Torre, N., Rubio, M. A., Bordiu, E., Cabrerizo, L., Aparicio, E., Hernandez, C., Sanchez-Pernaute, A., Diez-Valladares, L., Torres, A. J., Puente, M., & Charro, A. L. (2008). Effects of weight loss after bariatric surgery for morbid obesity on vascular endothelial growth factor-A, adipocytokines, and insulin. The Journal of clinical endocrinology and metabolism, 93(11), 4276-4281. doi: 10.1210/jc.2007-1370

208

Gavrilova, O., Marcus-Samuels, B., Leon, L. R., Vinson, C., & Reitman, M. L. (2000). Leptin and diabetes in lipoatrophic mice. Nature, 403(6772), 850; discussion 850-851. doi: 10.1038/35002663 Gerber, H. P., Malik, A. K., Solar, G. P., Sherman, D., Liang, X. H., Meng, G., Hong, K., Marsters, J. C., & Ferrara, N. (2002). VEGF regulates haematopoietic stem cell survival by an internal autocrine loop mechanism. Nature, 417(6892), 954-958. doi: 10.1038/nature00821 Gerich, J. E. (2000). Physiology of glucose homeostasis. Diabetes Obes Metab, 2(6), 345- 350. Gomez-Perez, Y., Capllonch-Amer, G., Gianotti, M., Llado, I., & Proenza, A. M. (2012). Long-term high-fat-diet feeding induces skeletal muscle mitochondrial biogenesis in rats in a sex-dependent and muscle-type specific manner. Nutr Metab (Lond), 9, 15. doi: 10.1186/1743-7075-9-15 Gonzalez, R., Reingold, B. K., Gao, X., Gaidhu, M. P., Tsushima, R. G., & Unniappan, S. (2011). Nesfatin-1 exerts a direct, glucose-dependent insulinotropic action on mouse islet beta- and MIN6 cells. J Endocrinol, 208(3), R9-R16. doi: 10.1530/JOE-10-0492 Granger, A., & Kushner, J. A. (2009). Cellular origins of beta-cell regeneration: a legacy view of historical controversies. J Intern Med, 266(4), 325-338. doi: 10.1111/j.1365- 2796.2009.02156.x Griffin, M. E., Marcucci, M. J., Cline, G. W., Bell, K., Barucci, N., Lee, D., Goodyear, L. J., Kraegen, E. W., White, M. F., & Shulman, G. I. (1999). Free fatty acid-induced insulin resistance is associated with activation of protein kinase C theta and alterations in the insulin signaling cascade. Diabetes, 48(6), 1270-1274. Grundleger, M. L., & Thenen, S. W. (1982). Decreased insulin binding, glucose transport, and glucose metabolism in soleus muscle of rats fed a high fat diet. Diabetes, 31(3), 232- 237. Grupe, A., Hultgren, B., Ryan, A., Ma, Y. H., Bauer, M., & Stewart, T. A. (1995). Transgenic knockouts reveal a critical requirement for pancreatic beta cell glucokinase in maintaining glucose homeostasis. Cell, 83(1), 69-78. Gu, L., & Guo, Z. (2013). Alzheimer's Abeta42 and Abeta40 peptides form interlaced amyloid fibrils. J Neurochem, 126(3), 305-311. doi: 10.1111/jnc.12202 Hagberg, C. E., Falkevall, A., Wang, X., Larsson, E., Huusko, J., Nilsson, I., van Meeteren, L. A., Samen, E., Lu, L., Vanwildemeersch, M., Klar, J., Genove, G., Pietras, K., Stone- Elander, S., Claesson-Welsh, L., Yla-Herttuala, S., Lindahl, P., & Eriksson, U. (2010). Vascular endothelial growth factor B controls endothelial fatty acid uptake. Nature, 464(7290), 917-921. doi: 10.1038/nature08945 Hagberg, C. E., Mehlem, A., Falkevall, A., Muhl, L., Fam, B. C., Ortsater, H., Scotney, P., Nyqvist, D., Samen, E., Lu, L., Stone-Elander, S., Proietto, J., Andrikopoulos, S., Sjoholm, A., Nash, A., & Eriksson, U. (2012). Targeting VEGF-B as a novel treatment for insulin resistance and type 2 diabetes. Nature, 490(7420), 426-430. doi: 10.1038/nature11464 Halberg, N., Khan, T., Trujillo, M. E., Wernstedt-Asterholm, I., Attie, A. D., Sherwani, S., Wang, Z. V., Landskroner-Eiger, S., Dineen, S., Magalang, U. J., Brekken, R. A., & Scherer, P. E. (2009). Hypoxia-inducible factor 1alpha induces fibrosis and insulin resistance in white adipose tissue. Mol Cell Biol, 29(16), 4467-4483. doi: 10.1128/MCB.00192-09 Hall, Mark, Frank, Eibe, Holmes, Geoffrey, Pfahringer, Bernhard, Reutemann, Peter, & Witten, Ian H. (2009). The WEKA data mining software: an update. SIGKDD Explor. Newsl., 11(1), 10-18. doi: 10.1145/1656274.1656278

209

Hamrick, M. W., Pennington, C., Newton, D., Xie, D., & Isales, C. (2004). Leptin deficiency produces contrasting phenotypes in bones of the limb and spine. Bone, 34(3), 376-383. doi: 10.1016/j.bone.2003.11.020 Hansen, P. A., Han, D. H., Marshall, B. A., Nolte, L. A., Chen, M. M., Mueckler, M., & Holloszy, J. O. (1998). A high fat diet impairs stimulation of glucose transport in muscle. Functional evaluation of potential mechanisms. J Biol Chem, 273(40), 26157- 26163. Hariri, N., & Thibault, L. (2010). High-fat diet-induced obesity in animal models. Nutr Res Rev, 23(2), 270-299. doi: 10.1017/S0954422410000168 Hashemi Kani, A., Alavian, S. M., Haghighatdoost, F., & Azadbakht, L. (2014). Diet macronutrients composition in nonalcoholic Fatty liver disease: a review on the related documents. Hepat Mon, 14(2), e10939. doi: 10.5812/hepatmon.10939 Hausman, D. B., Loh, M. Y., Flatt, W. P., & Martin, R. J. (1997). Adipose tissue expansion and the development of obesity: influence of dietary fat type. Asia Pac J Clin Nutr, 6(1), 49-55. Healy, M. E., Chow, J. D., Byrne, F. L., Breen, D. S., Leitinger, N., Li, C., Lackner, C., Caldwell, S. H., & Hoehn, K. L. (2015). Dietary effects on liver tumor burden in mice treated with the hepatocellular carcinogen diethylnitrosamine. J Hepatol, 62(3), 599- 606. doi: 10.1016/j.jhep.2014.10.024 Heinrich, L. F., Andersen, D. K., Cleasby, M. E., & Lawson, C. (2015). Long-term high fat feeding of rats results in increased numbers of circulating microvesicles with proinflammatory effects on endothelial cells. Br J Nutr, 113(11), 1704-1711. doi: 10.1017/S0007114515001117 Hill, J. O., Melanson, E. L., & Wyatt, H. T. (2000). Dietary fat intake and regulation of energy balance: implications for obesity. J Nutr, 130(2S Suppl), 284S-288S. Hill, P. A., Tumber, A., & Meikle, M. C. (1997). Multiple extracellular signals promote osteoblast survival and apoptosis. Endocrinology, 138(9), 3849-3858. doi: 10.1210/endo.138.9.5370 Ho, L., Qin, W., Pompl, P. N., Xiang, Z., Wang, J., Zhao, Z., Peng, Y., Cambareri, G., Rocher, A., Mobbs, C. V., Hof, P. R., & Pasinetti, G. M. (2004). Diet-induced insulin resistance promotes amyloidosis in a transgenic mouse model of Alzheimer's disease. FASEB J, 18(7), 902-904. doi: 10.1096/fj.03-0978fje Hoehn, K. L., Hohnen-Behrens, C., Cederberg, A., Wu, L. E., Turner, N., Yuasa, T., Ebina, Y., & James, D. E. (2008). IRS1-independent defects define major nodes of insulin resistance. Cell Metab, 7(5), 421-433. doi: S1550-4131(08)00115-0 [pii]

10.1016/j.cmet.2008.04.005 Hoehn, K. L., Salmon, A. B., Hohnen-Behrens, C., Turner, N., Hoy, A. J., Maghzal, G. J., Stocker, R., Van Remmen, H., Kraegen, E. W., Cooney, G. J., Richardson, A. R., & James, D. E. (2009). Insulin resistance is a cellular antioxidant defense mechanism. Proc Natl Acad Sci U S A, 106(42), 17787-17792. doi: 10.1073/pnas.0902380106 Hoehn, K. L., Turner, N., Swarbrick, M. M., Wilks, D., Preston, E., Phua, Y., Joshi, H., Furler, S. M., Larance, M., Hegarty, B. D., Leslie, S. J., Pickford, R., Hoy, A. J., Kraegen, E. W., James, D. E., & Cooney, G. J. (2010). Acute or chronic upregulation of mitochondrial fatty acid oxidation has no net effect on whole-body energy expenditure or adiposity. Cell Metab, 11(1), 70-76. doi: S1550-4131(09)00372-6 [pii]

10.1016/j.cmet.2009.11.008 Hofman, A., Ott, A., Breteler, M. M., Bots, M. L., Slooter, A. J., van Harskamp, F., van Duijn, C. N., Van Broeckhoven, C., & Grobbee, D. E. (1997). Atherosclerosis, apolipoprotein E, and prevalence of dementia and Alzheimer's disease in the Rotterdam Study. Lancet, 349(9046), 151-154. doi: 10.1016/S0140-6736(96)09328-2 210

Holland, W. L., Brozinick, J. T., Wang, L. P., Hawkins, E. D., Sargent, K. M., Liu, Y., Narra, K., Hoehn, K. L., Knotts, T. A., Siesky, A., Nelson, D. H., Karathanasis, S. K., Fontenot, G. K., Birnbaum, M. J., & Summers, S. A. (2007). Inhibition of ceramide synthesis ameliorates glucocorticoid-, saturated-fat-, and obesity-induced insulin resistance. Cell Metab, 5(3), 167-179. doi: 10.1016/j.cmet.2007.01.002 Hotamisligil, G. S., Peraldi, P., Budavari, A., Ellis, R., White, M. F., & Spiegelman, B. M. (1996). IRS-1-mediated inhibition of insulin receptor tyrosine kinase activity in TNF- alpha- and obesity-induced insulin resistance. Science, 271(5249), 665-668. Hotamisligil, G. S., Shargill, N. S., & Spiegelman, B. M. (1993). Adipose expression of tumor necrosis factor-alpha: direct role in obesity-linked insulin resistance. Science, 259(5091), 87-91. Hotta, K., Funahashi, T., Arita, Y., Takahashi, M., Matsuda, M., Okamoto, Y., Iwahashi, H., Kuriyama, H., Ouchi, N., Maeda, K., Nishida, M., Kihara, S., Sakai, N., Nakajima, T., Hasegawa, K., Muraguchi, M., Ohmoto, Y., Nakamura, T., Yamashita, S., Hanafusa, T., & Matsuzawa, Y. (2000). Plasma concentrations of a novel, adipose-specific protein, adiponectin, in type 2 diabetic patients. Arterioscler Thromb Vasc Biol, 20(6), 1595-1599. Hull, R. L., Kodama, K., Utzschneider, K. M., Carr, D. B., Prigeon, R. L., & Kahn, S. E. (2005). Dietary-fat-induced obesity in mice results in beta cell hyperplasia but not increased insulin release: evidence for specificity of impaired beta cell adaptation. Diabetologia, 48(7), 1350-1358. doi: 10.1007/s00125-005-1772-9 Hulver, M. W., Berggren, J. R., Cortright, R. N., Dudek, R. W., Thompson, R. P., Pories, W. J., MacDonald, K. G., Cline, G. W., Shulman, G. I., Dohm, G. L., & Houmard, J. A. (2003). Skeletal muscle lipid metabolism with obesity. Am J Physiol Endocrinol Metab, 284(4), E741-747. doi: 10.1152/ajpendo.00514.2002 Hummel, K. P., Dickie, M. M., & Coleman, D. L. (1966). Diabetes, a new mutation in the mouse. Science, 153(3740), 1127-1128. Huypens, P. (2007). Adipokines regulate systemic insulin sensitivity in accordance to existing energy reserves. Med Hypotheses, 69(1), 161-165. doi: 10.1016/j.mehy.2006.10.053 Huypens, P. R. (2007). Leptin and adiponectin regulate compensatory beta cell growth in accordance to overweight. Med Hypotheses, 68(5), 1134-1137. doi: 10.1016/j.mehy.2006.09.046 IDF. (2014). IDF DIABETES ATLAS. from http://www.idf.org/worlddiabetesday/toolkit/gp/facts-figures Ingalls, A. M., Dickie, M. M., & Snell, G. D. (1950). Obese, a new mutation in the house mouse. J Hered, 41(12), 317-318. Inzana, J. A., Kung, M., Shu, L., Hamada, D., Xing, L. P., Zuscik, M. J., Awad, H. A., & Mooney, R. A. (2013). Immature mice are more susceptible to the detrimental effects of high fat diet on cancellous bone in the distal femur. Bone, 57(1), 174-183. doi: 10.1016/j.bone.2013.08.003 Ishigaki, Y., Katagiri, H., Yamada, T., Ogihara, T., Imai, J., Uno, K., Hasegawa, Y., Gao, J., Ishihara, H., Shimosegawa, T., Sakoda, H., Asano, T., & Oka, Y. (2005). Dissipating excess energy stored in the liver is a potential treatment strategy for diabetes associated with obesity. Diabetes, 54(2), 322-332. James, D. E., Strube, M., & Mueckler, M. (1989). Molecular cloning and characterization of an insulin-regulatable glucose transporter. Nature, 338(6210), 83-87. doi: 10.1038/338083a0 Jequier, E. (2002). Pathways to obesity. Int J Obes Relat Metab Disord, 26 Suppl 2, S12-17. doi: 10.1038/sj.ijo.0802123

211

Kahn, C. R. (2003a). The Gordon Wilson Lecture. Lessons about the control of glucose homeostasis and the pathogenesis of diabetes from knockout mice. Trans Am Clin Climatol Assoc, 114, 125-148. Kahn, C. R. (2003b). Knockout mice challenge our concepts of glucose homeostasis and the pathogenesis of diabetes. Exp Diabesity Res, 4(3), 169-182. doi: 10.1155/EDR.2003.169 Kahn, S. E., Hull, R. L., & Utzschneider, K. M. (2006). Mechanisms linking obesity to insulin resistance and type 2 diabetes. Nature, 444(7121), 840-846. doi: 10.1038/nature05482 Katz, E. B., Stenbit, A. E., Hatton, K., DePinho, R., & Charron, M. J. (1995). Cardiac and adipose tissue abnormalities but not diabetes in mice deficient in GLUT4. Nature, 377(6545), 151-155. doi: 10.1038/377151a0 Kieffer, T. J., & Habener, J. F. (2000). The adipoinsular axis: effects of leptin on pancreatic beta-cells. Am J Physiol Endocrinol Metab, 278(1), E1-E14. Kieffer, T. J., Heller, R. S., & Habener, J. F. (1996). Leptin receptors expressed on pancreatic beta-cells. Biochem Biophys Res Commun, 224(2), 522-527. doi: 10.1006/bbrc.1996.1059 Kim, D. H., Sartor, M. A., Bain, J. R., Sandoval, D., Stevens, R. D., Medvedovic, M., Newgard, C. B., Woods, S. C., & Seeley, R. J. (2012). Rapid and weight-independent improvement of glucose tolerance induced by a Peptide designed to elicit apoptosis in adipose tissue endothelium. Diabetes, 61(9), 2299-2310. doi: 10.2337/db11-1579 Kim, W., & Egan, J. M. (2008). The role of incretins in glucose homeostasis and diabetes treatment. Pharmacol Rev, 60(4), 470-512. doi: 10.1124/pr.108.000604 Kim, Y. B., Nikoulina, S. E., Ciaraldi, T. P., Henry, R. R., & Kahn, B. B. (1999). Normal insulin-dependent activation of Akt/protein kinase B, with diminished activation of phosphoinositide 3-kinase, in muscle in type 2 diabetes. J Clin Invest, 104(6), 733- 741. doi: 10.1172/JCI6928 King, A. J. (2012). The use of animal models in diabetes research. Br J Pharmacol, 166(3), 877-894. doi: 10.1111/j.1476-5381.2012.01911.x Kitamura, T., Kitamura, Y., Nakae, J., Giordano, A., Cinti, S., Kahn, C. R., Efstratiadis, A., & Accili, D. (2004). Mosaic analysis of insulin receptor function. J Clin Invest, 113(2), 209-219. doi: 10.1172/JCI17810 Klein, G. L. (2014). Insulin and bone: Recent developments. World J Diabetes, 5(1), 14-16. doi: 10.4239/wjd.v5.i1.14 Kleinridders, A., Ferris, H. A., Cai, W., & Kahn, C. R. (2014). Insulin action in brain regulates systemic metabolism and brain function. Diabetes, 63(7), 2232-2243. doi: 10.2337/db14-0568 Klip, A., & Paquet, M. R. (1990). Glucose transport and glucose transporters in muscle and their metabolic regulation. Diabetes Care, 13(3), 228-243. Knight, E. M., Martins, I. V., Gumusgoz, S., Allan, S. M., & Lawrence, C. B. (2014). High- fat diet-induced memory impairment in triple-transgenic Alzheimer's disease (3xTgAD) mice is independent of changes in amyloid and tau pathology. Neurobiol Aging, 35(8), 1821-1832. doi: 10.1016/j.neurobiolaging.2014.02.010 Knight, Elysse M, Martins, Isaura VA, Gümüsgöz, Sarah, Allan, Stuart M, & Lawrence, Catherine B. (2014). High-fat diet-induced memory impairment in triple-transgenic Alzheimer's disease (3xTgAD) mice is independent of changes in amyloid and tau pathology. Neurobiology of aging, 35(8), 1821-1832. Kolonin, M. G., Saha, P. K., Chan, L., Pasqualini, R., & Arap, W. (2004). Reversal of obesity by targeted ablation of adipose tissue. Nat Med, 10(6), 625-632. doi: 10.1038/nm1048 nm1048 [pii]

212

Konige, M., Wang, H., & Sztalryd, C. (2014). Role of adipose specific lipid droplet proteins in maintaining whole body energy homeostasis. Biochim Biophys Acta, 1842(3), 393- 401. doi: 10.1016/j.bbadis.2013.05.007 Koves, T. R., Li, P., An, J., Akimoto, T., Slentz, D., Ilkayeva, O., Dohm, G. L., Yan, Z., Newgard, C. B., & Muoio, D. M. (2005). Peroxisome proliferator-activated receptor- gamma co-activator 1alpha-mediated metabolic remodeling of skeletal myocytes mimics exercise training and reverses lipid-induced mitochondrial inefficiency. J Biol Chem, 280(39), 33588-33598. doi: 10.1074/jbc.M507621200 Koves, T. R., Ussher, J. R., Noland, R. C., Slentz, D., Mosedale, M., Ilkayeva, O., Bain, J., Stevens, R., Dyck, J. R., Newgard, C. B., Lopaschuk, G. D., & Muoio, D. M. (2008). Mitochondrial overload and incomplete fatty acid oxidation contribute to skeletal muscle insulin resistance. Cell Metab, 7(1), 45-56. doi: 10.1016/j.cmet.2007.10.013 Kraegen, E. W., Clark, P. W., Jenkins, A. B., Daley, E. A., Chisholm, D. J., & Storlien, L. H. (1991). Development of muscle insulin resistance after liver insulin resistance in high- fat-fed rats. Diabetes, 40(11), 1397-1403. Kream, B. E., Smith, M. D., Canalis, E., & Raisz, L. G. (1985). Characterization of the effect of insulin on collagen synthesis in fetal rat bone. Endocrinology, 116(1), 296-302. doi: 10.1210/endo-116-1-296 Kulkarni, R. N., Bruning, J. C., Winnay, J. N., Postic, C., Magnuson, M. A., & Kahn, C. R. (1999). Tissue-specific knockout of the insulin receptor in pancreatic beta cells creates an insulin secretory defect similar to that in type 2 diabetes. Cell, 96(3), 329-339. Kulkarni, R. N., Wang, Z. L., Wang, R. M., Hurley, J. D., Smith, D. M., Ghatei, M. A., Withers, D. J., Gardiner, J. V., Bailey, C. J., & Bloom, S. R. (1997). Leptin rapidly suppresses insulin release from insulinoma cells, rat and human islets and, in vivo, in mice. J Clin Invest, 100(11), 2729-2736. doi: 10.1172/JCI119818 Kusminski, C. M., Holland, W. L., Sun, K., Park, J., Spurgin, S. B., Lin, Y., Askew, G. R., Simcox, J. A., McClain, D. A., Li, C., & Scherer, P. E. (2012). MitoNEET-driven alterations in adipocyte mitochondrial activity reveal a crucial adaptive process that preserves insulin sensitivity in obesity. Nat Med, 18(10), 1539-1549. doi: 10.1038/nm.2899 Kusminski, C. M., & Scherer, P. E. (2012). Mitochondrial dysfunction in white adipose tissue. Trends Endocrinol Metab, 23(9), 435-443. doi: 10.1016/j.tem.2012.06.004 Le Lay, S., Blouin, C. M., Hajduch, E., & Dugail, I. (2009). Filling up adipocytes with lipids. Lessons from caveolin-1 deficiency. Biochim Biophys Acta, 1791(6), 514-518. doi: 10.1016/j.bbalip.2008.10.008 Lee, A. W., & Cox, R. D. (2011). Use of mouse models in studying type 2 diabetes mellitus. Expert Rev Mol Med, 13, e1. doi: 10.1017/S1462399410001729 Lee, S., Chen, T. T., Barber, C. L., Jordan, M. C., Murdock, J., Desai, S., Ferrara, N., Nagy, A., Roos, K. P., & Iruela-Arispe, M. L. (2007). Autocrine VEGF signaling is required for vascular homeostasis. Cell, 130(4), 691-703. doi: 10.1016/j.cell.2007.06.054 Lee, Y. S., Li, P., Huh, J. Y., Hwang, I. J., Lu, M., Kim, J. I., Ham, M., Talukdar, S., Chen, A., Lu, W. J., Bandyopadhyay, G. K., Schwendener, R., Olefsky, J., & Kim, J. B. (2011). Inflammation is necessary for long-term but not short-term high-fat diet- induced insulin resistance. Diabetes, 60(10), 2474-2483. doi: 10.2337/db11-0194 Leibiger, B., Leibiger, I. B., Moede, T., Kemper, S., Kulkarni, R. N., Kahn, C. R., de Vargas, L. M., & Berggren, P. O. (2001). Selective insulin signaling through A and B insulin receptors regulates transcription of insulin and glucokinase genes in pancreatic beta cells. Mol Cell, 7(3), 559-570. Leiter, E. H. (1982). Multiple low-dose streptozotocin-induced hyperglycemia and insulitis in C57BL mice: influence of inbred background, sex, and thymus. Proc Natl Acad Sci U S A, 79(2), 630-634.

213

Lemonnier, D., Suquet, J. P., Aubert, R., De Gasquet, P., & Pequignot, E. (1975). Metabolism of the mouse made obese by a high-fat diet. Diabete Metab, 1(2), 77-85. Lester-Coll, N., Rivera, E. J., Soscia, S. J., Doiron, K., Wands, J. R., & de la Monte, S. M. (2006). Intracerebral streptozotocin model of type 3 diabetes: relevance to sporadic Alzheimer's disease. J Alzheimers Dis, 9(1), 13-33. Lettner, A., & Roden, M. (2008). Ectopic fat and insulin resistance. Curr Diab Rep, 8(3), 185- 191. Li, J., Cantley, J., Burchfield, J. G., Meoli, C. C., Stockli, J., Whitworth, P. T., Pant, H., Chaudhuri, R., Groffen, A. J., Verhage, M., & James, D. E. (2014). DOC2 isoforms play dual roles in insulin secretion and insulin-stimulated glucose uptake. Diabetologia, 57(10), 2173-2182. doi: 10.1007/s00125-014-3312-y Li, X., Song, D., & Leng, S. X. (2015a). Link between type 2 diabetes and Alzheimer's disease: from epidemiology to mechanism and treatment. Clin Interv Aging, 10, 549- 560. doi: 10.2147/CIA.S74042 Li, X., Song, D., & Leng, S. X. (2015b). Link between type 2 diabetes and Alzheimer's disease: from epidemiology to mechanism and treatment. Clin Interv Aging, 10, 549- 560. doi: 10.2147/CIA.S74042 Li, Xiaohua, Song, Dalin, & Leng, Sean X. (2015). Link between type 2 diabetes and Alzheimer’s disease: from epidemiology to mechanism and treatment. Clinical interventions in aging, 10, 549. Liang, W. C., Wu, X., Peale, F. V., Lee, C. V., Meng, Y. G., Gutierrez, J., Fu, L., Malik, A. K., Gerber, H. P., Ferrara, N., & Fuh, G. (2006). Cross-species vascular endothelial growth factor (VEGF)-blocking antibodies completely inhibit the growth of human tumor xenografts and measure the contribution of stromal VEGF. The Journal of biological chemistry, 281(2), 951-961. doi: 10.1074/jbc.M508199200 Lijnen, R. H., & Scroyen, I. (2013). Effect of Vascular Endothelial Growth Factor-Receptor 2 antagonism on adiposity in obese mice. J Mol Endocrinol. doi: 10.1530/JME-12-0244 Lingohr, M. K., Buettner, R., & Rhodes, C. J. (2002). Pancreatic beta-cell growth and survival--a role in obesity-linked type 2 diabetes? Trends Mol Med, 8(8), 375-384. Linnemann, A. K., Baan, M., & Davis, D. B. (2014). Pancreatic beta-cell proliferation in obesity. Adv Nutr, 5(3), 278-288. doi: 10.3945/an.113.005488 Liu, S. C., Wang, Q., Lienhard, G. E., & Keller, S. R. (1999). Insulin receptor substrate 3 is not essential for growth or glucose homeostasis. J Biol Chem, 274(25), 18093-18099. Loebig, M., Klement, J., Schmoller, A., Betz, S., Heuck, N., Schweiger, U., Peters, A., Schultes, B., & Oltmanns, K. M. (2010). Evidence for a relationship between VEGF and BMI independent of insulin sensitivity by glucose clamp procedure in a homogenous group healthy young men. PLoS One, 5(9), e12610. doi: 10.1371/journal.pone.0012610 Lu, X., Ji, Y., Zhang, L., Zhang, Y., Zhang, S., An, Y., Liu, P., & Zheng, Y. (2012). Resistance to obesity by repression of VEGF gene expression through induction of brown-like adipocyte differentiation. Endocrinology, 153(7), 3123-3132. doi: 10.1210/en.2012-1151 Lu, X., & Zheng, Y. (2013). Comment on: Elias et al. Adipose Tissue Overexpression of Vascular Endothelial Growth Factor Protects Against Diet-Induced Obesity and Insulin Resistance. Diabetes 2012;61:1801-1813. Diabetes, 62(1), e3. doi: 10.2337/db12-1130 Lumeng, C. N., & Saltiel, A. R. (2011). Inflammatory links between obesity and metabolic disease. J Clin Invest, 121(6), 2111-2117. doi: 10.1172/JCI57132 Madsen, A. N., Hansen, G., Paulsen, S. J., Lykkegaard, K., Tang-Christensen, M., Hansen, H. S., Levin, B. E., Larsen, P. J., Knudsen, L. B., Fosgerau, K., & Vrang, N. (2010). Long-term characterization of the diet-induced obese and diet-resistant rat model: a

214

polygenetic rat model mimicking the human obesity syndrome. J Endocrinol, 206(3), 287-296. doi: 10.1677/JOE-10-0004 Maffei, M., Halaas, J., Ravussin, E., Pratley, R. E., Lee, G. H., Zhang, Y., Fei, H., Kim, S., Lallone, R., Ranganathan, S., & et al. (1995). Leptin levels in human and rodent: measurement of plasma leptin and ob RNA in obese and weight-reduced subjects. Nat Med, 1(11), 1155-1161. Mangiafico, S. P., Lim, S. H., Neoh, S., Massinet, H., Joannides, C. N., Proietto, J., Andrikopoulos, S., & Fam, B. C. (2011). A primary defect in glucose production alone cannot induce glucose intolerance without defects in insulin secretion. J Endocrinol, 210(3), 335-347. doi: 10.1530/JOE-11-0126 Mansor, L. S., Gonzalez, E. R., Cole, M. A., Tyler, D. J., Beeson, J. H., Clarke, K., Carr, C. A., & Heather, L. C. (2013). Cardiac metabolism in a new rat model of type 2 diabetes using high-fat diet with low dose streptozotocin. Cardiovasc Diabetol, 12, 136. doi: 10.1186/1475-2840-12-136 Marroqui, L., Gonzalez, A., Neco, P., Caballero-Garrido, E., Vieira, E., Ripoll, C., Nadal, A., & Quesada, I. (2012). Role of leptin in the pancreatic beta-cell: effects and signaling pathways. J Mol Endocrinol, 49(1), R9-17. doi: 10.1530/JME-12-0025 Masciopinto, F., Di Pietro, N., Corona, C., Bomba, M., Pipino, C., Curcio, M., Di Castelnuovo, A., Ciavardelli, D., Silvestri, E., Canzoniero, L. M., Sekler, I., Pandolfi, A., & Sensi, S. L. (2012). Effects of long-term treatment with pioglitazone on cognition and glucose metabolism of PS1-KI, 3xTg-AD, and wild-type mice. Cell Death Dis, 3, e448. doi: 10.1038/cddis.2012.189 Mathis, D., Vence, L., & Benoist, C. (2001). beta-Cell death during progression to diabetes. Nature, 414(6865), 792-798. doi: 10.1038/414792a Matthews, D. R., Hosker, J. P., Rudenski, A. S., Naylor, B. A., Treacher, D. F., & Turner, R. C. (1985). Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia, 28(7), 412-419. McGillicuddy, F. C., Reynolds, C. M., Finucane, O., Coleman, E., Harford, K. A., Grant, C., Sergi, D., Williams, L. M., Mills, K. H., & Roche, H. M. (2013). Long-term exposure to a high-fat diet results in the development of glucose intolerance and insulin resistance in interleukin-1 receptor I-deficient mice. Am J Physiol Endocrinol Metab, 305(7), E834-844. doi: 10.1152/ajpendo.00297.2013 Mehla, J., Chauhan, B. C., & Chauhan, N. B. (2014). Experimental induction of type 2 diabetes in aging-accelerated mice triggered Alzheimer-like pathology and memory deficits. J Alzheimers Dis, 39(1), 145-162. doi: 10.3233/JAD-131238 Mei, M., Zhao, L., Li, Q., Chen, Y., Huang, A., Varghese, Z., Moorhead, J. F., Zhang, S., Powis, S. H., Li, Q., & Ruan, X. Z. (2011). Inflammatory stress exacerbates ectopic lipid deposition in C57BL/6J mice. Lipids Health Dis, 10, 110. doi: 10.1186/1476- 511X-10-110 Monetti, M., Levin, M. C., Watt, M. J., Sajan, M. P., Marmor, S., Hubbard, B. K., Stevens, R. D., Bain, J. R., Newgard, C. B., Farese, R. V., Sr., Hevener, A. L., & Farese, R. V., Jr. (2007). Dissociation of hepatic steatosis and insulin resistance in mice overexpressing DGAT in the liver. Cell Metab, 6(1), 69-78. doi: 10.1016/j.cmet.2007.05.005 Montgomery, M. K., Hallahan, N. L., Brown, S. H., Liu, M., Mitchell, T. W., Cooney, G. J., & Turner, N. (2013). Mouse strain-dependent variation in obesity and glucose homeostasis in response to high-fat feeding. Diabetologia, 56(5), 1129-1139. doi: 10.1007/s00125-013-2846-8 Mootha, V. K., Lindgren, C. M., Eriksson, K. F., Subramanian, A., Sihag, S., Lehar, J., Puigserver, P., Carlsson, E., Ridderstrale, M., Laurila, E., Houstis, N., Daly, M. J., Patterson, N., Mesirov, J. P., Golub, T. R., Tamayo, P., Spiegelman, B., Lander, E. S.,

215

Hirschhorn, J. N., Altshuler, D., & Groop, L. C. (2003). PGC-1alpha-responsive genes involved in oxidative phosphorylation are coordinately downregulated in human diabetes. Nat Genet, 34(3), 267-273. doi: 10.1038/ng1180 Morioka, T., Asilmaz, E., Hu, J., Dishinger, J. F., Kurpad, A. J., Elias, C. F., Li, H., Elmquist, J. K., Kennedy, R. T., & Kulkarni, R. N. (2007). Disruption of leptin receptor expression in the pancreas directly affects beta cell growth and function in mice. J Clin Invest, 117(10), 2860-2868. doi: 10.1172/JCI30910 Morrison, C. D., Huypens, P., Stewart, L. K., & Gettys, T. W. (2009). Implications of crosstalk between leptin and insulin signaling during the development of diet-induced obesity. Biochim Biophys Acta, 1792(5), 409-416. doi: 10.1016/j.bbadis.2008.09.005 Mosca, L. N., Goldberg, T. B., da Silva, V. N., da Silva, C. C., Kurokawa, C. S., Bisi Rizzo, A. C., & Corrente, J. E. (2014). Excess body fat negatively affects bone mass in adolescents. Nutrition, 30(7-8), 847-852. doi: 10.1016/j.nut.2013.12.003 Moss, E. G., & Alexander, J. (1990). Noninvasive diagnosis of deep vein thrombosis. N J Med, 87(12), 1005-1007. Mosser, R. E., Maulis, M. F., Moulle, V. S., Dunn, J. C., Carboneau, B. A., Arasi, K., Pappan, K., Poitout, V., & Gannon, M. (2015). High-fat diet-induced beta-cell proliferation occurs prior to insulin resistance in C57Bl/6J male mice. Am J Physiol Endocrinol Metab, 308(7), E573-582. doi: 10.1152/ajpendo.00460.2014 Mullican, S. E., Tomaru, T., Gaddis, C. A., Peed, L. C., Sundaram, A., & Lazar, M. A. (2013). A novel adipose-specific gene deletion model demonstrates potential pitfalls of existing methods. Mol Endocrinol, 27(1), 127-134. doi: 10.1210/me.2012-1267 Muoio, D. M., Dohm, G. L., Fiedorek, F. T., Jr., Tapscott, E. B., & Coleman, R. A. (1997). Leptin directly alters lipid partitioning in skeletal muscle. Diabetes, 46(8), 1360-1363. Muoio, D. M., & Newgard, C. B. (2006). Obesity-related derangements in metabolic regulation. Annu Rev Biochem, 75, 367-401. doi: 10.1146/annurev.biochem.75.103004.142512 Muoio, D. M., & Newgard, C. B. (2008). Mechanisms of disease:Molecular and metabolic mechanisms of insulin resistance and beta-cell failure in type 2 diabetes. Nat Rev Mol Cell Biol, 9(3), 193-205. doi: 10.1038/nrm2327 Myers, M. G., Cowley, M. A., & Munzberg, H. (2008). Mechanisms of leptin action and leptin resistance. Annu Rev Physiol, 70, 537-556. doi: 10.1146/annurev.physiol.70.113006.100707 Nadal-Casellas, A., Amengual-Cladera, E., Proenza, A. M., Llado, I., & Gianotti, M. (2010). Long-term high-fat-diet feeding impairs mitochondrial biogenesis in liver of male and female rats. Cell Physiol Biochem, 26(3), 291-302. doi: 10.1159/000320552 Nagle, C. A., An, J., Shiota, M., Torres, T. P., Cline, G. W., Liu, Z. X., Wang, S., Catlin, R. L., Shulman, G. I., Newgard, C. B., & Coleman, R. A. (2007). Hepatic overexpression of glycerol-sn-3-phosphate acyltransferase 1 in rats causes insulin resistance. J Biol Chem, 282(20), 14807-14815. doi: 10.1074/jbc.M611550200 Nagle, C. A., Klett, E. L., & Coleman, R. A. (2009). Hepatic triacylglycerol accumulation and insulin resistance. J Lipid Res, 50 Suppl, S74-79. doi: 10.1194/jlr.R800053-JLR200 Nakamura, A., & Terauchi, Y. (2013). Lessons from mouse models of high-fat diet-induced NAFLD. Int J Mol Sci, 14(11), 21240-21257. doi: 10.3390/ijms141121240 Nakata, M., Manaka, K., Yamamoto, S., Mori, M., & Yada, T. (2011). Nesfatin-1 enhances glucose-induced insulin secretion by promoting Ca(2+) influx through L-type channels in mouse islet beta-cells. Endocr J, 58(4), 305-313. Nakata, M., Okada, T., Ozawa, K., & Yada, T. (2007). Resistin induces insulin resistance in pancreatic islets to impair glucose-induced insulin release. Biochem Biophys Res Commun, 353(4), 1046-1051. doi: 10.1016/j.bbrc.2006.12.134

216

Neel, J. V. (1999). Diabetes mellitus: a "thrifty" genotype rendered detrimental by "progress"? 1962. Bull World Health Organ, 77(8), 694-703; discussion 692-693. Neschen, S., Morino, K., Hammond, L. E., Zhang, D., Liu, Z. X., Romanelli, A. J., Cline, G. W., Pongratz, R. L., Zhang, X. M., Choi, C. S., Coleman, R. A., & Shulman, G. I. (2005). Prevention of hepatic steatosis and hepatic insulin resistance in mitochondrial acyl-CoA:glycerol-sn-3-phosphate acyltransferase 1 knockout mice. Cell Metab, 2(1), 55-65. doi: 10.1016/j.cmet.2005.06.006 Neubauer, N., & Kulkarni, R. N. (2006). Molecular approaches to study control of glucose homeostasis. ILAR J, 47(3), 199-211. Newgard, C. B. (2012). Interplay between lipids and branched-chain amino acids in development of insulin resistance. Cell Metab, 15(5), 606-614. doi: 10.1016/j.cmet.2012.01.024 Ng, Y., Ramm, G., Lopez, J. A., & James, D. E. (2008). Rapid activation of Akt2 is sufficient to stimulate GLUT4 translocation in 3T3-L1 adipocytes. Cell Metab, 7(4), 348-356. doi: 10.1016/j.cmet.2008.02.008 Niswender, K. D., & Magnuson, M. A. (2007). Obesity and the beta cell: lessons from leptin. J Clin Invest, 117(10), 2753-2756. doi: 10.1172/JCI33528 Oakley, H., Cole, S. L., Logan, S., Maus, E., Shao, P., Craft, J., Guillozet-Bongaarts, A., Ohno, M., Disterhoft, J., Van Eldik, L., Berry, R., & Vassar, R. (2006). Intraneuronal beta-amyloid aggregates, neurodegeneration, and neuron loss in transgenic mice with five familial Alzheimer's disease mutations: potential factors in amyloid plaque formation. J Neurosci, 26(40), 10129-10140. doi: 10.1523/JNEUROSCI.1202- 06.2006 Oddo, S., Caccamo, A., Kitazawa, M., Tseng, B. P., & LaFerla, F. M. (2003). Amyloid deposition precedes tangle formation in a triple transgenic model of Alzheimer's disease. Neurobiol Aging, 24(8), 1063-1070. Ogawa, Y., Masuzaki, H., Hosoda, K., Aizawa-Abe, M., Suga, J., Suda, M., Ebihara, K., Iwai, H., Matsuoka, N., Satoh, N., Odaka, H., Kasuga, H., Fujisawa, Y., Inoue, G., Nishimura, H., Yoshimasa, Y., & Nakao, K. (1999). Increased glucose metabolism and insulin sensitivity in transgenic skinny mice overexpressing leptin. Diabetes, 48(9), 1822-1829. Okamoto, M., Ohara-Imaizumi, M., Kubota, N., Hashimoto, S., Eto, K., Kanno, T., Kubota, T., Wakui, M., Nagai, R., Noda, M., Nagamatsu, S., & Kadowaki, T. (2008). Adiponectin induces insulin secretion in vitro and in vivo at a low glucose concentration. Diabetologia, 51(5), 827-835. doi: 10.1007/s00125-008-0944-9 Omar, B., Pacini, G., & Ahren, B. (2012). Differential development of glucose intolerance and pancreatic islet adaptation in multiple diet induced obesity models. Nutrients, 4(10), 1367-1381. doi: 10.3390/nu4101367 Pagliassotti, M. J., Horton, T. J., Gayles, E. C., Koppenhafer, T. A., Rosenzweig, T. D., & Hill, J. O. (1997). Reduced insulin suppression of glucose appearance is related to susceptibility to dietary obesity in rats. Am J Physiol, 272(4 Pt 2), R1264-1270. Pagliassotti, M. J., Prach, P. A., Koppenhafer, T. A., & Pan, D. A. (1996). Changes in insulin action, triglycerides, and lipid composition during sucrose feeding in rats. Am J Physiol, 271(5 Pt 2), R1319-1326. Pan, D. A., & Storlien, L. H. (1993). Dietary lipid profile is a determinant of tissue phospholipid fatty acid composition and rate of weight gain in rats. J Nutr, 123(3), 512-519. Park, S. Y., Cho, Y. R., Kim, H. J., Higashimori, T., Danton, C., Lee, M. K., Dey, A., Rothermel, B., Kim, Y. B., Kalinowski, A., Russell, K. S., & Kim, J. K. (2005). Unraveling the temporal pattern of diet-induced insulin resistance in individual organs and cardiac dysfunction in C57BL/6 mice. Diabetes, 54(12), 3530-3540.

217

Pasarica, M., Sereda, O. R., Redman, L. M., Albarado, D. C., Hymel, D. T., Roan, L. E., Rood, J. C., Burk, D. H., & Smith, S. R. (2009). Reduced adipose tissue oxygenation in human obesity: evidence for rarefaction, macrophage chemotaxis, and inflammation without an angiogenic response. Diabetes, 58(3), 718-725. doi: 10.2337/db08-1098 Pascoe, J., Hollern, D., Stamateris, R., Abbasi, M., Romano, L. C., Zou, B., O'Donnell, C. P., Garcia-Ocana, A., & Alonso, L. C. (2012). Free fatty acids block glucose-induced beta-cell proliferation in mice by inducing cell cycle inhibitors p16 and p18. Diabetes, 61(3), 632-641. doi: 10.2337/db11-0991 Peckham, S. C., & Entenman, C. (1962). The influence of a hypercaloric diet on gross body and adipose tissue composition in the rat. Res Dev Tech Rep, 23. Petersen, K. F., Dufour, S., Befroy, D., Lehrke, M., Hendler, R. E., & Shulman, G. I. (2005). Reversal of nonalcoholic hepatic steatosis, hepatic insulin resistance, and hyperglycemia by moderate weight reduction in patients with type 2 diabetes. Diabetes, 54(3), 603-608. Peyot, M. L., Pepin, E., Lamontagne, J., Latour, M. G., Zarrouki, B., Lussier, R., Pineda, M., Jetton, T. L., Madiraju, S. R., Joly, E., & Prentki, M. (2010). Beta-cell failure in diet- induced obese mice stratified according to body weight gain: secretory dysfunction and altered islet lipid metabolism without steatosis or reduced beta-cell mass. Diabetes, 59(9), 2178-2187. doi: 10.2337/db09-1452 Plaschke, K., Kopitz, J., Siegelin, M., Schliebs, R., Salkovic-Petrisic, M., Riederer, P., & Hoyer, S. (2010). Insulin-resistant brain state after intracerebroventricular streptozotocin injection exacerbates Alzheimer-like changes in Tg2576 AbetaPP- overexpressing mice. J Alzheimers Dis, 19(2), 691-704. doi: 10.3233/JAD-2010-1270 Polonsky, K. S., Sturis, J., & Bell, G. I. (1996). Seminars in Medicine of the Beth Israel Hospital, Boston. Non-insulin-dependent diabetes mellitus - a genetically programmed failure of the beta cell to compensate for insulin resistance. N Engl J Med, 334(12), 777-783. doi: 10.1056/NEJM199603213341207 Posey, K. A., Clegg, D. J., Printz, R. L., Byun, J., Morton, G. J., Vivekanandan-Giri, A., Pennathur, S., Baskin, D. G., Heinecke, J. W., Woods, S. C., Schwartz, M. W., & Niswender, K. D. (2009). Hypothalamic proinflammatory lipid accumulation, inflammation, and insulin resistance in rats fed a high-fat diet. Am J Physiol Endocrinol Metab, 296(5), E1003-1012. doi: 10.1152/ajpendo.90377.2008 Previs, S. F., Withers, D. J., Ren, J. M., White, M. F., & Shulman, G. I. (2000). Contrasting effects of IRS-1 versus IRS-2 gene disruption on carbohydrate and lipid metabolism in vivo. J Biol Chem, 275(50), 38990-38994. doi: 10.1074/jbc.M006490200 Rabe, K., Lehrke, M., Parhofer, K. G., & Broedl, U. C. (2008). Adipokines and insulin resistance. Mol Med, 14(11-12), 741-751. doi: 10.2119/2008-00058.Rabe Rakatzi, I., Mueller, H., Ritzeler, O., Tennagels, N., & Eckel, J. (2004). Adiponectin counteracts cytokine- and fatty acid-induced apoptosis in the pancreatic beta-cell line INS-1. Diabetologia, 47(2), 249-258. doi: 10.1007/s00125-003-1293-3 Rasouli, N., Molavi, B., Elbein, S. C., & Kern, P. A. (2007). Ectopic fat accumulation and metabolic syndrome. Diabetes Obes Metab, 9(1), 1-10. doi: 10.1111/j.1463- 1326.2006.00590.x Ray, A., Alalem, M., & Ray, B. K. (2014). Insulin signaling network in cancer. Indian J Biochem Biophys, 51(6), 493-498. Regazzetti, C., Bost, F., Le Marchand-Brustel, Y., Tanti, J. F., & Giorgetti-Peraldi, S. (2010). Insulin induces REDD1 expression through hypoxia-inducible factor 1 activation in adipocytes. J Biol Chem, 285(8), 5157-5164. doi: 10.1074/jbc.M109.047688 Ritchie, M. E., Diyagama, D., Neilson, J., van Laar, R., Dobrovic, A., Holloway, A., & Smyth, G. K. (2006). Empirical array quality weights in the analysis of microarray data. BMC Bioinformatics, 7, 261. doi: 10.1186/1471-2105-7-261

218

Ritchie, M. E., Phipson, B., Wu, D., Hu, Y., Law, C. W., Shi, W., & Smyth, G. K. (2015). limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res, 43(7), e47. doi: 10.1093/nar/gkv007 Rosen, E. D., Kulkarni, R. N., Sarraf, P., Ozcan, U., Okada, T., Hsu, C. H., Eisenman, D., Magnuson, M. A., Gonzalez, F. J., Kahn, C. R., & Spiegelman, B. M. (2003). Targeted elimination of peroxisome proliferator-activated receptor gamma in beta cells leads to abnormalities in islet mass without compromising glucose homeostasis. Mol Cell Biol, 23(20), 7222-7229. Rosen, E. D., & Spiegelman, B. M. (2014). What we talk about when we talk about fat. Cell, 156(1-2), 20-44. doi: 10.1016/j.cell.2013.12.012 Rupnick, M. A., Panigrahy, D., Zhang, C. Y., Dallabrida, S. M., Lowell, B. B., Langer, R., & Folkman, M. J. (2002). Adipose tissue mass can be regulated through the vasculature. Proc Natl Acad Sci U S A, 99(16), 10730-10735. doi: 10.1073/pnas.162349799

162349799 [pii] Rutkowski, J. M., Stern, J. H., & Scherer, P. E. (2015). The cell biology of fat expansion. J Cell Biol, 208(5), 501-512. doi: 10.1083/jcb.201409063 Saltiel, A. R. (2001). New perspectives into the molecular pathogenesis and treatment of type 2 diabetes. Cell, 104(4), 517-529. Saltiel, A. R., & Kahn, C. R. (2001). Insulin signalling and the regulation of glucose and lipid metabolism. Nature, 414(6865), 799-806. doi: 10.1038/414799a Savage, D. B., Choi, C. S., Samuel, V. T., Liu, Z. X., Zhang, D., Wang, A., Zhang, X. M., Cline, G. W., Yu, X. X., Geisler, J. G., Bhanot, S., Monia, B. P., & Shulman, G. I. (2006). Reversal of diet-induced hepatic steatosis and hepatic insulin resistance by antisense oligonucleotide inhibitors of acetyl-CoA carboxylases 1 and 2. J Clin Invest, 116(3), 817-824. doi: 10.1172/JCI27300 Scalera, A., & Tarantino, G. (2014). Could metabolic syndrome lead to hepatocarcinoma via non-alcoholic fatty liver disease? World J Gastroenterol, 20(28), 9217-9228. doi: 10.3748/wjg.v20.i28.9217 Scherer, P. E. (2006). Adipose tissue: from lipid storage compartment to endocrine organ. Diabetes, 55(6), 1537-1545. doi: 10.2337/db06-0263 Schimenti, J., & Bucan, M. (1998). Functional genomics in the mouse: phenotype-based mutagenesis screens. Genome Res, 8(7), 698-710. Schindelin, J., Arganda-Carreras, I., Frise, E., Kaynig, V., Longair, M., Pietzsch, T., Preibisch, S., Rueden, C., Saalfeld, S., Schmid, B., Tinevez, J. Y., White, D. J., Hartenstein, V., Eliceiri, K., Tomancak, P., & Cardona, A. (2012). Fiji: an open- source platform for biological-image analysis. Nat Methods, 9(7), 676-682. doi: 10.1038/nmeth.2019 Schinner, S., Scherbaum, W. A., Bornstein, S. R., & Barthel, A. (2005). Molecular mechanisms of insulin resistance. Diabet Med, 22(6), 674-682. doi: 10.1111/j.1464- 5491.2005.01566.x Schott, M., Scherbaum, W. A., & Bornstein, S. R. (2004). Acquired and inherited lipodystrophies. N Engl J Med, 351(1), 103-104; author reply 103-104. Schrijvers, E. M., Witteman, J. C., Sijbrands, E. J., Hofman, A., Koudstaal, P. J., & Breteler, M. M. (2010). Insulin metabolism and the risk of Alzheimer disease: the Rotterdam Study. Neurology, 75(22), 1982-1987. doi: 10.1212/WNL.0b013e3181ffe4f6 Schwingshackl, L., & Hoffmann, G. (2014). Comparison of the long-term effects of high-fat v. low-fat diet consumption on cardiometabolic risk factors in subjects with abnormal glucose metabolism: a systematic review and meta-analysis. Br J Nutr, 111(12), 2047- 2058. doi: 10.1017/S0007114514000464 Searcy, J. L., Phelps, J. T., Pancani, T., Kadish, I., Popovic, J., Anderson, K. L., Beckett, T. L., Murphy, M. P., Chen, K. C., Blalock, E. M., Landfield, P. W., Porter, N. M., & 219

Thibault, O. (2012). Long-term pioglitazone treatment improves learning and attenuates pathological markers in a mouse model of Alzheimer's disease. J Alzheimers Dis, 30(4), 943-961. doi: 10.3233/JAD-2012-111661 Segerstrom, L., Fuchs, D., Backman, U., Holmquist, K., Christofferson, R., & Azarbayjani, F. (2006). The anti-VEGF antibody bevacizumab potently reduces the growth rate of high-risk neuroblastoma xenografts. Pediatr Res, 60(5), 576-581. doi: 10.1203/01.pdr.0000242494.94000.52 Seufert, J. (2004). Leptin effects on pancreatic beta-cell gene expression and function. Diabetes, 53 Suppl 1, S152-158. Seufert, J., Kieffer, T. J., & Habener, J. F. (1999). Leptin inhibits insulin gene transcription and reverses hyperinsulinemia in leptin-deficient ob/ob mice. Proc Natl Acad Sci U S A, 96(2), 674-679. Shanik, M. H., Xu, Y., Skrha, J., Dankner, R., Zick, Y., & Roth, J. (2008). Insulin resistance and hyperinsulinemia: is hyperinsulinemia the cart or the horse? Diabetes Care, 31 Suppl 2, S262-268. doi: 10.2337/dc08-s264 Shaw, R. J., Lamia, K. A., Vasquez, D., Koo, S. H., Bardeesy, N., Depinho, R. A., Montminy, M., & Cantley, L. C. (2005). The kinase LKB1 mediates glucose homeostasis in liver and therapeutic effects of metformin. Science, 310(5754), 1642-1646. doi: 10.1126/science.1120781 Sheehan, M. T. (2003). Current therapeutic options in type 2 diabetes mellitus: a practical approach. Clin Med Res, 1(3), 189-200. Shepherd, P. R., & Kahn, B. B. (1999). Glucose transporters and insulin action--implications for insulin resistance and diabetes mellitus. N Engl J Med, 341(4), 248-257. doi: 10.1056/NEJM199907223410406 Shimomura, I., Hammer, R. E., Ikemoto, S., Brown, M. S., & Goldstein, J. L. (1999). Leptin reverses insulin resistance and diabetes mellitus in mice with congenital lipodystrophy. Nature, 401(6748), 73-76. doi: 10.1038/43448 Shoelson, S. E., Herrero, L., & Naaz, A. (2007). Obesity, inflammation, and insulin resistance. Gastroenterology, 132(6), 2169-2180. doi: 10.1053/j.gastro.2007.03.059 Silha, J. V., Krsek, M., Sucharda, P., & Murphy, L. J. (2005). Angiogenic factors are elevated in overweight and obese individuals. International journal of obesity, 29(11), 1308- 1314. doi: 10.1038/sj.ijo.0802987 Simmgen, M., Knauf, C., Lopez, M., Choudhury, A. I., Charalambous, M., Cantley, J., Bedford, D. C., Claret, M., Iglesias, M. A., Heffron, H., Cani, P. D., Vidal-Puig, A., Burcelin, R., & Withers, D. J. (2006). Liver-specific deletion of insulin receptor substrate 2 does not impair hepatic glucose and lipid metabolism in mice. Diabetologia, 49(3), 552-561. doi: 10.1007/s00125-005-0084-4 Sinha, A., & Kling, S. (2009). A review of adolescent obesity: prevalence, etiology, and treatment. Obes Surg, 19(1), 113-120. doi: 10.1007/s11695-008-9650-4 Solon-Biet, S. M., McMahon, A. C., Ballard, J. W., Ruohonen, K., Wu, L. E., Cogger, V. C., Warren, A., Huang, X., Pichaud, N., Melvin, R. G., Gokarn, R., Khalil, M., Turner, N., Cooney, G. J., Sinclair, D. A., Raubenheimer, D., Le Couteur, D. G., & Simpson, S. J. (2014). The ratio of macronutrients, not caloric intake, dictates cardiometabolic health, aging, and longevity in ad libitum-fed mice. Cell Metab, 19(3), 418-430. doi: 10.1016/j.cmet.2014.02.009 Sommer, C., Straehle, C., Kothe, U., & Hamprecht, F. A. (2011, March 30 2011-April 2 2011). Ilastik: Interactive learning and segmentation toolkit. Paper presented at the Biomedical Imaging: From Nano to Macro, 2011 IEEE International Symposium on. Spies, P. E., Slats, D., Sjogren, J. M., Kremer, B. P., Verhey, F. R., Rikkert, M. G., & Verbeek, M. M. (2010). The cerebrospinal fluid amyloid beta42/40 ratio in the

220

differentiation of Alzheimer's disease from non-Alzheimer's dementia. Curr Alzheimer Res, 7(5), 470-476. Stamateris, R. E., Sharma, R. B., Hollern, D. A., & Alonso, L. C. (2013). Adaptive beta-cell proliferation increases early in high-fat feeding in mice, concurrent with metabolic changes, with induction of islet cyclin D2 expression. Am J Physiol Endocrinol Metab, 305(1), E149-159. doi: 10.1152/ajpendo.00040.2013 Standaert, M. L., Galloway, L., Karnam, P., Bandyopadhyay, G., Moscat, J., & Farese, R. V. (1997). Protein kinase C-zeta as a downstream effector of phosphatidylinositol 3- kinase during insulin stimulation in rat adipocytes. Potential role in glucose transport. J Biol Chem, 272(48), 30075-30082. Steil, G. M., Trivedi, N., Jonas, J. C., Hasenkamp, W. M., Sharma, A., Bonner-Weir, S., & Weir, G. C. (2001). Adaptation of beta-cell mass to substrate oversupply: enhanced function with normal gene expression. Am J Physiol Endocrinol Metab, 280(5), E788- 796. Stockli, J., Meoli, C. C., Hoffman, N. J., Fazakerley, D. J., Pant, H., Cleasby, M. E., Ma, X., Kleinert, M., Brandon, A. E., Lopez, J. A., Cooney, G. J., & James, D. E. (2015). The RabGAP TBC1D1 plays a central role in exercise-regulated glucose metabolism in skeletal muscle. Diabetes. doi: 10.2337/db13-1489 Storlien, L. H., Jenkins, A. B., Chisholm, D. J., Pascoe, W. S., Khouri, S., & Kraegen, E. W. (1991). Influence of dietary fat composition on development of insulin resistance in rats. Relationship to muscle triglyceride and omega-3 fatty acids in muscle phospholipid. Diabetes, 40(2), 280-289. Subramanian, A., Tamayo, P., Mootha, V. K., Mukherjee, S., Ebert, B. L., Gillette, M. A., Paulovich, A., Pomeroy, S. L., Golub, T. R., Lander, E. S., & Mesirov, J. P. (2005). Gene set enrichment analysis: a knowledge-based approach for interpreting genome- wide expression profiles. Proc Natl Acad Sci U S A, 102(43), 15545-15550. doi: 10.1073/pnas.0506580102 Sun, K., Halberg, N., Khan, M., Magalang, U. J., & Scherer, P. E. (2013). Selective inhibition of hypoxia-inducible factor 1alpha ameliorates adipose tissue dysfunction. Mol Cell Biol, 33(5), 904-917. doi: 10.1128/MCB.00951-12 Sun, K., Wernstedt Asterholm, I., Kusminski, C. M., Bueno, A. C., Wang, Z. V., Pollard, J. W., Brekken, R. A., & Scherer, P. E. (2012). Dichotomous effects of VEGF-A on adipose tissue dysfunction. Proc Natl Acad Sci U S A, 109(15), 5874-5879. doi: 10.1073/pnas.1200447109 Sung, H. K., Doh, K. O., Son, J. E., Park, J. G., Bae, Y., Choi, S., Nelson, S. M., Cowling, R., Nagy, K., Michael, I. P., Koh, G. Y., Adamson, S. L., Pawson, T., & Nagy, A. (2013). Adipose Vascular Endothelial Growth Factor Regulates Metabolic Homeostasis through Angiogenesis. Cell Metab, 17(1), 61-72. doi: 10.1016/j.cmet.2012.12.010 Surwit, R. S., Feinglos, M. N., Rodin, J., Sutherland, A., Petro, A. E., Opara, E. C., Kuhn, C. M., & Rebuffe-Scrive, M. (1995). Differential effects of fat and sucrose on the development of obesity and diabetes in C57BL/6J and A/J mice. Metabolism: clinical and experimental, 44(5), 645-651. Surwit, R. S., Kuhn, C. M., Cochrane, C., McCubbin, J. A., & Feinglos, M. N. (1988). Diet- induced type II diabetes in C57BL/6J mice. Diabetes, 37(9), 1163-1167. Surwit, R. S., Seldin, M. F., Kuhn, C. M., Cochrane, C., & Feinglos, M. N. (1991). Control of expression of insulin resistance and hyperglycemia by different genetic factors in diabetic C57BL/6J mice. Diabetes, 40(1), 82-87. Svistounov, D., Warren, A., McNerney, G. P., Owen, D. M., Zencak, D., Zykova, S. N., Crane, H., Huser, T., Quinn, R. J., Smedsrod, B., Le Couteur, D. G., & Cogger, V. C. (2012). The Relationship between fenestrations, sieve plates and rafts in liver

221

sinusoidal endothelial cells. PLoS One, 7(9), e46134. doi: 10.1371/journal.pone.0046134 Takalo, M., Haapasalo, A., Martiskainen, H., Kurkinen, K. M., Koivisto, H., Miettinen, P., Khandelwal, V. K., Kemppainen, S., Kaminska, D., Makinen, P., Leinonen, V., Pihlajamaki, J., Soininen, H., Laakso, M., Tanila, H., & Hiltunen, M. (2014). High-fat diet increases tau expression in the brain of T2DM and AD mice independently of peripheral metabolic status. J Nutr Biochem, 25(6), 634-641. doi: 10.1016/j.jnutbio.2014.02.003 Takeda, S., Sato, N., Uchio-Yamada, K., Sawada, K., Kunieda, T., Takeuchi, D., Kurinami, H., Shinohara, M., Rakugi, H., & Morishita, R. (2010). Diabetes-accelerated memory dysfunction via cerebrovascular inflammation and Abeta deposition in an Alzheimer mouse model with diabetes. Proc Natl Acad Sci U S A, 107(15), 7036-7041. doi: 10.1073/pnas.1000645107 Talbot, K., Wang, H. Y., Kazi, H., Han, L. Y., Bakshi, K. P., Stucky, A., Fuino, R. L., Kawaguchi, K. R., Samoyedny, A. J., Wilson, R. S., Arvanitakis, Z., Schneider, J. A., Wolf, B. A., Bennett, D. A., Trojanowski, J. Q., & Arnold, S. E. (2012). Demonstrated brain insulin resistance in Alzheimer's disease patients is associated with IGF-1 resistance, IRS-1 dysregulation, and cognitive decline. J Clin Invest, 122(4), 1316- 1338. doi: 10.1172/JCI59903 Tamaki, M., Fujitani, Y., Hara, A., Uchida, T., Tamura, Y., Takeno, K., Kawaguchi, M., Watanabe, T., Ogihara, T., Fukunaka, A., Shimizu, T., Mita, T., Kanazawa, A., Imaizumi, M. O., Abe, T., Kiyonari, H., Hojyo, S., Fukada, T., Kawauchi, T., Nagamatsu, S., Hirano, T., Kawamori, R., & Watada, H. (2013). The diabetes- susceptible gene SLC30A8/ZnT8 regulates hepatic insulin clearance. J Clin Invest, 123(10), 4513-4524. doi: 10.1172/JCI68807 Tan, S. X., Fisher-Wellman, K., Fazakerley, D., Ng, Y., Pant, H., Li, J., Meoli, C., Coster, A. C., Stockli, J., & James, D. E. (2015). Selective Insulin resistance in adipocytes. J Biol Chem. doi: 10.1074/jbc.M114.623686 Taniguchi, C. M., Finger, E. C., Krieg, A. J., Wu, C., Diep, A. N., LaGory, E. L., Wei, K., McGinnis, L. M., Yuan, J., Kuo, C. J., & Giaccia, A. J. (2013). Cross-talk between hypoxia and insulin signaling through Phd3 regulates hepatic glucose and lipid metabolism and ameliorates diabetes. Nat Med, 19(10), 1325-1330. doi: 10.1038/nm.3294 Taouis, M., Dagou, C., Ster, C., Durand, G., Pinault, M., & Delarue, J. (2002). N-3 polyunsaturated fatty acids prevent the defect of insulin receptor signaling in muscle. Am J Physiol Endocrinol Metab, 282(3), E664-671. doi: 10.1152/ajpendo.00320.2001 Tjwa, M., Luttun, A., Autiero, M., & Carmeliet, P. (2003). VEGF and PlGF: two pleiotropic growth factors with distinct roles in development and homeostasis. Cell Tissue Res, 314(1), 5-14. doi: 10.1007/s00441-003-0776-3 Tofolo, L. P., da Silva Ribeiro, T. A., Malta, A., Miranda, R. A., Gomes, R. M., de Oliveira, J. C., Abdennebi-Najar, L., de Almeida, D. L., Trombini, A. B., da Silva Franco, C. C., Pavanello, A., Fabricio, G. S., Rinaldi, W., Barella, L. F., de Freitas Mathias, P. C., & Palma-Rigo, K. (2014). Short-term moderate exercise provides long-lasting protective effects against metabolic dysfunction in rats fed a high-fat diet. Eur J Nutr. doi: 10.1007/s00394-014-0816-7 Tomas, E., Lin, Y. S., Dagher, Z., Saha, A., Luo, Z., Ido, Y., & Ruderman, N. B. (2002). Hyperglycemia and insulin resistance: possible mechanisms. Ann N Y Acad Sci, 967, 43-51. Tonks, K. T., Ng, Y., Miller, S., Coster, A. C., Samocha-Bonet, D., Iseli, T. J., Xu, A., Patrick, E., Yang, J. Y., Junutula, J. R., Modrusan, Z., Kolumam, G., Stockli, J., Chisholm, D. J., James, D. E., & Greenfield, J. R. (2013). Impaired Akt

222

phosphorylation in insulin-resistant human muscle is accompanied by selective and heterogeneous downstream defects. Diabetologia, 56(4), 875-885. doi: 10.1007/s00125-012-2811-y Toyama, K., Koibuchi, N., Hasegawa, Y., Uekawa, K., Yasuda, O., Sueta, D., Nakagawa, T., Ma, M., Kusaka, H., Lin, B., Ogawa, H., Ichijo, H., & Kim-Mitsuyama, S. (2015). ASK1 is involved in cognitive impairment caused by long-term high-fat diet feeding in mice. Sci Rep, 5, 10844. doi: 10.1038/srep10844 Tuomi, T., Santoro, N., Caprio, S., Cai, M., Weng, J., & Groop, L. (2014). The many faces of diabetes: a disease with increasing heterogeneity. Lancet, 383(9922), 1084-1094. doi: 10.1016/S0140-6736(13)62219-9 Turner, N., Bruce, C. R., Beale, S. M., Hoehn, K. L., So, T., Rolph, M. S., & Cooney, G. J. (2007). Excess lipid availability increases mitochondrial fatty acid oxidative capacity in muscle: evidence against a role for reduced fatty acid oxidation in lipid-induced insulin resistance in rodents. Diabetes, 56(8), 2085-2092. doi: 10.2337/db07-0093 Turner, N., Hariharan, K., TidAng, J., Frangioudakis, G., Beale, S. M., Wright, L. E., Zeng, X. Y., Leslie, S. J., Li, J. Y., Kraegen, E. W., Cooney, G. J., & Ye, J. M. (2009). Enhancement of muscle mitochondrial oxidative capacity and alterations in insulin action are lipid species dependent: potent tissue-specific effects of medium-chain fatty acids. Diabetes, 58(11), 2547-2554. doi: db09-0784 [pii]

10.2337/db09-0784 Turner, N., Kowalski, G. M., Leslie, S. J., Risis, S., Yang, C., Lee-Young, R. S., Babb, J. R., Meikle, P. J., Lancaster, G. I., Henstridge, D. C., White, P. J., Kraegen, E. W., Marette, A., Cooney, G. J., Febbraio, M. A., & Bruce, C. R. (2013). Distinct patterns of tissue-specific lipid accumulation during the induction of insulin resistance in mice by high-fat feeding. Diabetologia, 56(7), 1638-1648. doi: 10.1007/s00125-013-2913-1 Turner, R. T., Kalra, S. P., Wong, C. P., Philbrick, K. A., Lindenmaier, L. B., Boghossian, S., & Iwaniec, U. T. (2013). Peripheral leptin regulates bone formation. J Bone Miner Res, 28(1), 22-34. doi: 10.1002/jbmr.1734 Vaag, A., Henriksen, J. E., & Beck-Nielsen, H. (1992). Decreased insulin activation of glycogen synthase in skeletal muscles in young nonobese Caucasian first-degree relatives of patients with non-insulin-dependent diabetes mellitus. J Clin Invest, 89(3), 782-788. doi: 10.1172/JCI115656 van Daele, P. L., Stolk, R. P., Burger, H., Algra, D., Grobbee, D. E., Hofman, A., Birkenhager, J. C., & Pols, H. A. (1995). Bone density in non-insulin-dependent diabetes mellitus. The Rotterdam Study. Ann Intern Med, 122(6), 409-414. Van Heek, M., Compton, D. S., France, C. F., Tedesco, R. P., Fawzi, A. B., Graziano, M. P., Sybertz, E. J., Strader, C. D., & Davis, H. R., Jr. (1997). Diet-induced obese mice develop peripheral, but not central, resistance to leptin. J Clin Invest, 99(3), 385-390. doi: 10.1172/JCI119171 van Himbergen, T. M., Beiser, A. S., Ai, M., Seshadri, S., Otokozawa, S., Au, R., Thongtang, N., Wolf, P. A., & Schaefer, E. J. (2012). Biomarkers for insulin resistance and inflammation and the risk for all-cause dementia and alzheimer disease: results from the Framingham Heart Study. Arch Neurol, 69(5), 594-600. doi: 10.1001/archneurol.2011.670 Vandal, M., & Calon, F. (2015). Response to Comment on Vandal et al. Insulin Reverses the High-Fat Diet-Induced Increase in Brain Abeta and Improves Memory in an Animal Model of Alzheimer Disease. Diabetes 2014;63:4291-4301. Diabetes, 64(7), e18. doi: 10.2337/db15-0386 Vandal, M., White, P. J., Tremblay, C., St-Amour, I., Chevrier, G., Emond, V., Lefrancois, D., Virgili, J., Planel, E., Giguere, Y., Marette, A., & Calon, F. (2014). Insulin reverses the high-fat diet-induced increase in brain Abeta and improves memory in an 223

animal model of Alzheimer disease. Diabetes, 63(12), 4291-4301. doi: 10.2337/db14- 0375 Vasu, S., Moffett, R. C., Thorens, B., & Flatt, P. R. (2014). Role of endogenous GLP-1 and GIP in beta cell compensatory responses to insulin resistance and cellular stress. PLoS One, 9(6), e101005. doi: 10.1371/journal.pone.0101005 Veldhuis-Vlug, A. G., Fliers, E., & Bisschop, P. H. (2013). Bone as a regulator of glucose metabolism. Neth J Med, 71(8), 396-400. Wallace, T. M., Levy, J. C., & Matthews, D. R. (2004). Use and abuse of HOMA modeling. Diabetes Care, 27(6), 1487-1495. Wang, H., Storlien, L. H., & Huang, X. F. (2002). Effects of dietary fat types on body fatness, leptin, and ARC leptin receptor, NPY, and AgRP mRNA expression. Am J Physiol Endocrinol Metab, 282(6), E1352-1359. doi: 10.1152/ajpendo.00230.2001 Watarai, T., Kobayashi, M., Takata, Y., Sasaoka, T., Iwasaki, M., & Shigeta, Y. (1988). Alteration of insulin-receptor kinase activity by high-fat feeding. Diabetes, 37(10), 1397-1404. Wauman, J., & Tavernier, J. (2011). Leptin receptor signaling: pathways to leptin resistance. Front Biosci (Landmark Ed), 16, 2771-2793. Wee, N. K., & Baldock, P. A. (2014). The hunger games of skeletal metabolism. Bonekey Rep, 3, 588. doi: 10.1038/bonekey.2014.83 Wei, K., Piecewicz, S. M., McGinnis, L. M., Taniguchi, C. M., Wiegand, S. J., Anderson, K., Chan, C. W., Mulligan, K. X., Kuo, D., Yuan, J., Vallon, M., Morton, L. C., Lefai, E., Simon, M. C., Maher, J. J., Mithieux, G., Rajas, F., Annes, J. P., McGuinness, O. P., Thurston, G., Giaccia, A. J., & Kuo, C. J. (2013). A liver Hif-2alpha-Irs2 pathway sensitizes hepatic insulin signaling and is modulated by Vegf inhibition. Nat Med, 19(10), 1331-1337. doi: 10.1038/nm.3295 Weir, G. C., & Bonner-Weir, S. (2007). A dominant role for glucose in beta cell compensation of insulin resistance. J Clin Invest, 117(1), 81-83. doi: 10.1172/JCI30862 Werther, G. A., Hogg, A., Oldfield, B. J., McKinley, M. J., Figdor, R., Allen, A. M., & Mendelsohn, F. A. (1987). Localization and characterization of insulin receptors in rat brain and pituitary gland using in vitro autoradiography and computerized densitometry. Endocrinology, 121(4), 1562-1570. doi: 10.1210/endo-121-4-1562 West, D. B., & York, B. (1998). Dietary fat, genetic predisposition, and obesity: lessons from animal models. Am J Clin Nutr, 67(3 Suppl), 505S-512S. Willett, C. G., Boucher, Y., Duda, D. G., di Tomaso, E., Munn, L. L., Tong, R. T., Kozin, S. V., Petit, L., Jain, R. K., Chung, D. C., Sahani, D. V., Kalva, S. P., Cohen, K. S., Scadden, D. T., Fischman, A. J., Clark, J. W., Ryan, D. P., Zhu, A. X., Blaszkowsky, L. S., Shellito, P. C., Mino-Kenudson, M., & Lauwers, G. Y. (2005). Surrogate markers for antiangiogenic therapy and dose-limiting toxicities for bevacizumab with radiation and chemotherapy: continued experience of a phase I trial in rectal cancer patients. J Clin Oncol, 23(31), 8136-8139. doi: 10.1200/JCO.2005.02.5635 Winzell, M. S., & Ahren, B. (2004). The high-fat diet-fed mouse: a model for studying mechanisms and treatment of impaired glucose tolerance and type 2 diabetes. Diabetes, 53 Suppl 3, S215-219. Wisniewski, J. R., Zougman, A., Nagaraj, N., & Mann, M. (2009). Universal sample preparation method for proteome analysis. Nat Methods, 6(5), 359-362. doi: 10.1038/nmeth.1322 Withers, D. J., Gutierrez, J. S., Towery, H., Burks, D. J., Ren, J. M., Previs, S., Zhang, Y., Bernal, D., Pons, S., Shulman, G. I., Bonner-Weir, S., & White, M. F. (1998). Disruption of IRS-2 causes type 2 diabetes in mice. Nature, 391(6670), 900-904. doi: 10.1038/36116

224

Wong, I. P., Nguyen, A. D., Khor, E. C., Enriquez, R. F., Eisman, J. A., Sainsbury, A., Herzog, H., & Baldock, P. A. (2013). Neuropeptide Y is a critical modulator of leptin's regulation of cortical bone. J Bone Miner Res, 28(4), 886-898. doi: 10.1002/jbmr.1786 Wongdee, K., & Charoenphandhu, N. (2015). Update on type 2 diabetes-related osteoporosis. World J Diabetes, 6(5), 673-678. doi: 10.4239/wjd.v6.i5.673 Woods, S. C., Seeley, R. J., Baskin, D. G., & Schwartz, M. W. (2003). Insulin and the blood- brain barrier. Curr Pharm Des, 9(10), 795-800. Wu, L. E., Meoli, C. C., Mangiafico, S. P., Fazakerley, D. J., Cogger, V. C., Mohamad, M., Pant, H., Kang, M. J., Powter, E., Burchfield, J. G., Xirouchaki, C. E., Mikolaizak, A. S., Stockli, J., Kolumam, G., van Bruggen, N., Gamble, J. R., Le Couteur, D. G., Cooney, G. J., Andrikopoulos, S., & James, D. E. (2014). Systemic VEGF-A neutralization ameliorates diet-induced metabolic dysfunction. Diabetes, 63(8), 2656- 2667. doi: 10.2337/db13-1665 Wu, L. E., Samocha-Bonet, D., Whitworth, P. T., Fazakerley, D. J., Turner, N., Biden, T. J., James, D. E., & Cantley, J. (2014). Identification of fatty acid binding protein 4 as an adipokine that regulates insulin secretion during obesity. Mol Metab, 3(4), 465-473. doi: 10.1016/j.molmet.2014.02.005 Xia, J., Zhong, Y., Huang, G., Chen, Y., Shi, H., & Zhang, Z. (2012). The relationship between insulin resistance and osteoporosis in elderly male type 2 diabetes mellitus and diabetic nephropathy. Ann Endocrinol (Paris), 73(6), 546-551. doi: 10.1016/j.ando.2012.09.009 Xu, H., Barnes, G. T., Yang, Q., Tan, G., Yang, D., Chou, C. J., Sole, J., Nichols, A., Ross, J. S., Tartaglia, L. A., & Chen, H. (2003). Chronic inflammation in fat plays a crucial role in the development of obesity-related insulin resistance. J Clin Invest, 112(12), 1821-1830. doi: 10.1172/JCI19451 Yamauchi, T., Kamon, J., Waki, H., Terauchi, Y., Kubota, N., Hara, K., Mori, Y., Ide, T., Murakami, K., Tsuboyama-Kasaoka, N., Ezaki, O., Akanuma, Y., Gavrilova, O., Vinson, C., Reitman, M. L., Kagechika, H., Shudo, K., Yoda, M., Nakano, Y., Tobe, K., Nagai, R., Kimura, S., Tomita, M., Froguel, P., & Kadowaki, T. (2001). The fat- derived hormone adiponectin reverses insulin resistance associated with both lipoatrophy and obesity. Nat Med, 7(8), 941-946. doi: 10.1038/90984 Yang, J. C., Haworth, L., Sherry, R. M., Hwu, P., Schwartzentruber, D. J., Topalian, S. L., Steinberg, S. M., Chen, H. X., & Rosenberg, S. A. (2003). A randomized trial of bevacizumab, an anti-vascular endothelial growth factor antibody, for metastatic renal cancer. N Engl J Med, 349(5), 427-434. doi: 10.1056/NEJMoa021491 Yang, Q., Graham, T. E., Mody, N., Preitner, F., Peroni, O. D., Zabolotny, J. M., Kotani, K., Quadro, L., & Kahn, B. B. (2005). Serum retinol binding protein 4 contributes to insulin resistance in obesity and type 2 diabetes. Nature, 436(7049), 356-362. doi: 10.1038/nature03711 Yaqoob, P., Sherrington, E. J., Jeffery, N. M., Sanderson, P., Harvey, D. J., Newsholme, E. A., & Calder, P. C. (1995). Comparison of the effects of a range of dietary lipids upon serum and tissue lipid composition in the rat. Int J Biochem Cell Biol, 27(3), 297-310. Ye, J., Gao, Z., Yin, J., & He, Q. (2007). Hypoxia is a potential risk factor for chronic inflammation and adiponectin reduction in adipose tissue of ob/ob and dietary obese mice. Am J Physiol Endocrinol Metab, 293(4), E1118-1128. doi: 10.1152/ajpendo.00435.2007 Yilmaz, M., & Hotamisligil, G. S. (2013). Damned If You Do, Damned If You Don't: The Conundrum of Adipose Tissue Vascularization. Cell Metab, 17(1), 7-9. doi: 10.1016/j.cmet.2012.12.014

225

Yin, D., Tao, J., Lee, D. D., Shen, J., Hara, M., Lopez, J., Kuznetsov, A., Philipson, L. H., & Chong, A. S. (2006). Recovery of islet beta-cell function in streptozotocin- induced diabetic mice: an indirect role for the spleen. Diabetes, 55(12), 3256-3263. doi: 10.2337/db05-1275 Youngren, J. F., Paik, J., & Barnard, R. J. (2001). Impaired insulin-receptor autophosphorylation is an early defect in fat-fed, insulin-resistant rats. J Appl Physiol (1985), 91(5), 2240-2247. Zhang, Y., Proenca, R., Maffei, M., Barone, M., Leopold, L., & Friedman, J. M. (1994). Positional cloning of the mouse obese gene and its human homologue. Nature, 372(6505), 425-432. doi: 10.1038/372425a0 Zhao, W. Q., & Townsend, M. (2009). Insulin resistance and amyloidogenesis as common molecular foundation for type 2 diabetes and Alzheimer's disease. Biochim Biophys Acta, 1792(5), 482-496. doi: 10.1016/j.bbadis.2008.10.014 Zierath, J. R., Houseknecht, K. L., Gnudi, L., & Kahn, B. B. (1997). High-fat feeding impairs insulin-stimulated GLUT4 recruitment via an early insulin-signaling defect. Diabetes, 46(2), 215-223. Zimmet, P., Alberti, K. G., & Shaw, J. (2001). Global and societal implications of the diabetes epidemic. Nature, 414(6865), 782-787. doi: 10.1038/414782a Zimmet, P. Z., Magliano, D. J., Herman, W. H., & Shaw, J. E. (2014). Diabetes: a 21st century challenge. Lancet Diabetes Endocrinol, 2(1), 56-64. doi: 10.1016/S2213- 8587(13)70112-8

226