wenz iD - Proefschrift Mariëtte E.G. Kranendonk

Adipose tissue dysfunction and cardiometabolic risk Ex vivo, in vitro and clinical studies

MARIĂ‹TTE E.G. KRANENDONK

Adipose tissue dysfunction and cardiometabolic risk Ex vivo, in vitro and clinical studies

Thesis with a summary in Dutch, Utrecht University Š M.E.G.Kranendonk, 2014, Utrecht, the Netherlands

ISBN: 978-90-8891-908-4 Cover: Wouter Scheper Lay-out: Wendy Schoneveld, www.wenz iD.nl Printed by: Proefschriftmaken.nl | Uitgeverij BOXPress Published by: Uitgeverij BOXPress, ‘s Hertogenbosch Financial support by the Dutch Heart Foundation for the publication of this thesis is gratefully acknowledged.

Financial support for printing of this thesis by ChipSoft B.V., NovoNordisk B.V. and Servier Nederland Farma B.V. is gratefully acknowledged.

Adipose tissue dysfunction and cardiometabolic risk Ex vivo, in vitro and clinical studies

Vetweefseldysfunctie en cardiometabool risico Ex vivo, in vitro en klinische studies (met een samenvatting in het Nederlands)

Proefschrift

ter verkrijging van de graad van doctor aan de Universiteit Utrecht op gezag van de rector magnificus, prof. dr. G.J. van der Zwaan ingevolge het besluit van het college voor promoties in het openbaar te verdedigen op donderdag 19 juni 2014 des middags te 4.15 uur

door

MariÍtte Elizabeth Gertruida Kranendonk geboren op 31 mei 1983 te ’s Gravenhage

Promotor:

Prof. dr. F.L.J.Visseren

Co-promotor: Dr. E. Kalkhoven

CONTENTS CHAPTER 1

General introduction and thesis outline

7

PART ONE

Relation of distinct visceral fat depots to metabolic disease

CHAPTER 2

The role of distinct adipose tissue depots and extracellular vesicles in obesity-induced metabolic dysfunction

17

CHAPTER 3

Inflammatory characteristics of distinct abdominal adipose tissue depots relate differently to metabolic risk factors for cardiovascular disease; the ADIPOSE II study

45

PART TWO

Adipose tissue extracellular vesicles in adipose tissue dysfunction and metabolic disease

CHAPTER 4

Human adipocyte extracellular vesicles in reciprocal signaling between adipocytes and macrophages

67

CHAPTER 5

Effect of extracellular vesicles of human adipose tissue on insulin signaling in liver and muscle cells

91

PART THREE Biomarkers for obesity-induced cardiovascular or metabolic

disease CHAPTER 6

Extracellular vesicle markers in relation to obesity and metabolic complications in patients with manifest cardiovascular disease

111

CHAPTER 7

Adiponectin and incident coronary heart disease and stroke.

A systematic review and meta-analysis of prospective studies

135

CHAPTER 8

General discussion

157

APPENDIX

Summary Nederlandse samenvatting Dankwoord Contributing Authors List of publications Curriculum Vitae

176 179 184 190 193 195

CHAPTER 1 General introduction

CHAPTER 1

Obesity, the elephant in the corner “The year is 2055. Socrates is discussing the history of the western obesity epidemic with Panacea, the goddess of healing.”

“Socrates: In the past 50 years, obesity has come to account for more health problems than any other single factor. I can’t help thinking, Panacea, that something could have been done to prevent this. What on earth was going on at the beginning of the 21st century?” David Ogilvie & Neil Hamlet (1).

Worldwide prevalence and problems of obesity Since 1980, the worldwide prevalence of obesity has doubled, with a subsequent increase of the risk for life threatening comorbidities accompanied with a severe economic burden (2,3). In fact, 65% of the world’s population live in countries where obesity kills more people than underweight. Obesity accounts for the fifth leading risk for global death, for 44% of disease burden, 23% of ischemic heart disease and between 7 to 41 % of certain cancers (4-7). In 2008, 35 % of adults aged 20+ worldwide were overweight (body mass index (BMI) 25-30 kg/m2, 35% men and 34% women) and 12 % was obese (BMI > 30 kg/m2; 10% men and 14% women). Moreover, while blood pressure and cholesterol levels are successfully reduced worldwide, the contribution of body mass index (BMI) and (obesityinduced) type 2 diabetes in heart disease risk only rises (Figure 1) (8).

Some scientists postulate that obesity is a maladjusted response to our western society.

Our ancient ancestors developed evolutionary benefits to adapt to alternating phases of scarcity and food abundance. As we have now developed a world with excess of energy supply and a minimal need for energy expenditure, we became a victim of our own created obesogenic environment (9).

Figure 1. Contribution of different risk factors to cardiovascular disease. Adapted from Capewell et al (8).

8

General introduction

Adipocyte

1

Anti-inflammatory macrophage Pro-inflammatory macrophage Cytotoxic T cell Adipokines Free fatty acids Extracellular vesicles

Figure 2. Adipose tissue as an immunometabolic organ.

Obesity is actually a very simple problem: when one consumes more energy than is expended, weight gain is the logical outcome. Some can handle the excess weight better than others without developing inconvenient or even life threatening comorbidities, but most individuals will suffer health consequences. Although the solution is as simple as the problem- to consume less and/or expend more- for most it is simply not possible despite public health measures and individual weight loss programs (10).

As discussed above by ´Socrates and Panacea´, why is obesity so difficult to control? (1) For that question to be answered we need to understand more about the etiology of obesity. With knowledge about the underlying molecular mechanisms, we might develop biomarkers able to predict which obese patient is at risk for subsequent cardiovascular disease. When identified such patient, better care could be provided or a more intensive weight loss program could be advised. Also, with a better understanding of the molecular mechanisms by which adipose tissue contributes to metabolic complications of obesity, drugs could be developed to interfere in those processes.

Adipose tissue as an endocrine organ All agony caused by obesity forced scientists to reconsider the role of adipose tissue. After it was postulated that adipose tissue might actually transmit feedback signals to the hypothalamus to regulate energy expenditure and food intake 60 years ago (11), many bioactive proteins secreted by adipose tissue have been discovered which qualify adipose tissue as an endocrine organ (12-14). These proteins secreted by adipose tissue are termed adipokines, of which one of the best-known is adiponectin which is specifically released by adipocytes (13). More importantly, this adipokine has insulin-sensitizing functions, thereby contributing to metabolic homeostasis. In obese patients, adiponectin levels are lower which is one of the mechanisms by which adipose tissue of obese patients contributes to metabolic comorbidities like insulin resistance. This insight revealed that without proper adipose tissue function whole body metabolism is severely disturbed,

9

CHAPTER 1

a process referred to as adipose tissue dysfunction (15). Adipose tissue dysfunction involves inflammation of adipose tissue, characterized by hypertrophic adipocytes, altered immune cell composition in favor of the pro-inflammatory type and subsequent altered secretion of bio-active factors such as adipokines and free fatty acids (Figure 2) (16-18).

Adipose tissue is distributed throughout the body in different depots that are associated differently with comorbidities such as cardiovascular disease (19-20). Interestingly, this observation was made centuries ago. Back in 1761, Morgagni identified the association between intra-abdominal obesity, hypertension, abnormal metabolism, and extensive atherosclerosis by using a combination of clinical histories and autopsies (21). In the following centuries, more became known about the molecular mechanisms of obesityinduced metabolic and vascular disease. For example, adipose tissue inflammation is more active in visceral adipose tissue compared to subcutaneous adipose tissue and visceral fat releases a higher amount of pro-inflammatory adipokines (22). Furthermore, visceral fat has a direct communication with the liver via the portal vein (23,24), which is why visceral fat is thought to contribute more importantly to obesity-induced comorbidities compared to other fat depots (25).

However, besides soluble proteins and cells, extracellular vesicles (EVs) are important intercellular communicators (Figure 2). EVs are a mix of distinct vesicles released by the cell, including microvesicles or microparticles that bud from the plasma membrane, and exosomes which are released upon fusion of multivesicular bodies with the plasma membrane (26,27). EVs contain a distinct set of proteins (including proteins involved in antigen presentation and adhesion and endosomal proteins), lipids (such as

Lipid rafts

RNAs

Adipokines Figure 3. Extracellular vesicles released by adipose tissue.

10

Adhesion molecules

Antigen presenting proteins

General introduction

phosphatidylserine, ceramide) and nucleic acids (mRNA, miRNA) which reflect the cellular phenotype at the time of release (Figure 3) (26,28). The morphological and structural properties of EVs enable them to interact with and modify target cells, via antigen presentation or transport of receptors or RNA transcripts (29-32). Crucial roles of EVs have been shown in many (patho)physiological processes, such as immune responses, inflam足 mation and cancer (33,34). Based on these observations, it seems plausible that EVs also play an etiological role in immunometabolic processes such as adipose tissue inflammation and subsequent insulin resistance, dyslipidemia and vascular disease. However, so far only few authors reported on the role of EVs in adipose tissue inflammation or obesity-related metabolic diseases, or the potential of EVs as biomarkers for metabolic diseases (35-40).

Therefore, based on current knowledge it can be stated that adipose tissue is an endocrine organ that is able to communicate with visceral organs via different molecular mechanisms, and thereby actively contributes to metabolic homeostasis and also metabolic disturbances.

Adipose tissue dysfunction, as a result of obesity, is a very interesting but also complex phenomenon, challenging doctors and scientists to identify patients at risk and to develop effective treatment strategies. In order to develop more powerful treatment or diagnostic tools for the obese patient at risk for subsequent cardiovascular disease, further insight in the underlying molecular mechanisms is required.

Objective of the thesis The aim of this thesis is to evaluate the link between adipose tissue dysfunction and cardiometabolic disease, by studying (extra)cellular behavior of distinct cell types in vitro, characteristics of human adipose tissue ex vivo combined with clinical characteristics of patients with manifest vascular disease from clinical studies.

Outline of the thesis The first part of this thesis (chapter 2 and 3) focuses on morphological and functional properties of adipose tissue and describes the role of distinct adipose tissue depots. Chapter 2 provides an overview of the various mechanisms by which adipose tissue might contribute to risk factors for cardiovascular disease. Adipose tissue consists of different depots which are anatomically linked to distinct organs. Evidence suggests that intrinsic differences as well as interactions of adipose tissue with surrounding organs accounts for different contributions to cardiometabolic disease. Furthermore, adipose tissue secretes numerous molecules that contribute to vascular damage via distinct pathways. Besides the well-known contribution of adipokines and free fatty acids, the potential role of extracellular vesicles in the development of obesity-induced cardiometabolic disease is described.

As mentioned above, significant differences are observed between subcutaneous adipose tissue and visceral adipose tissue depots and their different contributions to cardiometabolic disease. However, less is known about distinct visceral adipose tissue depots. In chapter 3, a clinical study is described in which adipose tissue biopsies from four different abdominal adipose tissue depots (one subcutaneous and three visceral depots) are characterized ex vivo, obtained from 28 males during abdominal aortic surgery. The morphological and

11

1

CHAPTER 1

functional profile of the four different fat depots is compared between males which are abdominally obese versus the abdominal lean, and adipose tissue characteristics are related to clinical variables such as insulin resistance and the metabolic syndrome.

The second part of this thesis (chapter 4 and 5) focuses on the release of extracellular vesicles by human adipose tissue. In chapter 4, EVs released by human adipocytes and human adipose tissue are characterized, and we investigate the potential of human adipose tissue EVs to contribute to reciprocal signaling between adipocytes and human macrophages in vitro. Local adipocyte-macrophage interplay is considered to be a key mechanism in adipose tissue inflammation, leading to metabolic complications such as insulin resistance.

In chapter 5, we show that adiponectin-positive EVs, presumably of adipocyte origin, can be isolated from human plasma. As this indicates a potential endocrine signaling of EVs from human fat, the putative endocrine effects of ex vivo human adipose tissue EVs on insulin signaling in liver and muscle cells in vitro is investigated. Furthermore, the relation between the number of EVs secreted by either subcutaneous and visceral adipose tissue is related to clinical metabolic characteristics of patients undergoing abdominal aortic surgery.

The final part of this thesis (chapter 6 and 7) describes the relation of adipose tissue-derived bioactive markers with metabolic or cardiovascular disease in large patient populations. In chapter 6, the relation between four extracellular vesicle markers and metabolic complications including type 2 diabetes in 1012 patients with clinically manifest vascular disease is described. An important aim of this study was to investigate the clinical utility of certain extracellular vesicle proteins as biomarkers in obesity-induced metabolic disease.

In chapter 7, the potential prognostic value of adiponectin for cardiovascular disease is studied. Adiponectin is exclusively secreted by adipocytes, has metabolic beneficent properties and is negatively related to obesity, insulin resistance and type 2 diabetes. The aim of chapter 7 was to meta-analyze the relation between plasma concentrations of adiponectin and risk of coronary heart disease and stroke in the general population.

12

General introduction

REFERENCES 1.

2.

Ogilvie D, Hamlet N. Obesity: The elephant in the corner. BMJ. 2005 Dec 24;331(7531):1545-8.

Finucane MM, Stevens GA, Cowan MJ, Danaei G, Lin JK, Paciorek CJ, et al. National, regional, and global trends in body-mass index since 1980: Systematic analysis of health examination surveys and epidemiological studies with 960 country-years and 9.1 million participants. Lancet.

2011 Feb 12;377(9765):557-67.

3. Henry FJ. Obesity prevention: The key to non-communicable disease control. West Indian Med J. 2011 Jul;60(4):446-51.

4. http://www.who.int/healthinfo/global_burden_disease/GlobalHealthRisks_report_part2. pdf?ua=1 5. Yusuf S, Hawken S, Ounpuu S, Bautista L, Franzosi MG, Commerford P, et al. Obesity and the risk of myocardial infarction in 27,000 participants from 52 countries: A case-control study.

Lancet. 2005 Nov 5;366(9497):1640-9.

6. Ford ES, Giles WH, Dietz WH. Prevalence of the metabolic syndrome among US adults: Findings from the third national health and nutrition examination survey. JAMA. 2002 Jan 16;287(3):356-9.

7. Renehan AG, Tyson M, Egger M, Heller RF, Zwahlen M. Body-mass index and incidence of cancer: A systematic review and meta-analysis of prospective observational studies. Lancet.

2008 Feb 16;371(9612):569-78.

8. Capewell S, Ford ES, Croft JB, Critchley JA, Greenlund KJ, Labarthe DR. Cardiovascular risk factor trends and potential for reducing coronary heart disease mortality in the united states of america. Bull World Health Organ. 2010 Feb;88(2):120-30.

9. Speakman JR. If body fatness is under physiological regulation, then how come we have an obesity epidemic? Physiology (Bethesda). 2014 Mar;29(2):88-98.

10. Anderson JW, Konz EC, Frederich RC, Wood CL. Long-term weight-loss maintenance: A metaanalysis of US studies. Am J Clin Nutr. 2001 Nov;74(5):579-84.

11. KENNEDY GC. The role of depot fat in the hypothalamic control of food intake in the rat. Proc R Soc Lond B Biol Sci. 1953 Jan 15;140(901):578-96.

12. Hotamisligil GS, Shargill NS, Spiegelman BM. Adipose expression of tumor necrosis factoralpha: Direct role in obesity-linked insulin resistance. Science. 1993 Jan 1;259(5091):87-91.

13. Scherer PE, Williams S, Fogliano M, Baldini G, Lodish HF. A novel serum protein similar to C1q, produced exclusively in adipocytes. J Biol Chem. 1995 Nov 10;270(45):26746-9.

14. Galic S, Oakhill JS, Steinberg GR. Adipose tissue as an endocrine organ. Mol Cell Endocrinol.

2010 Mar 25;316(2):129-39.

15. Hajer GR, van Haeften TW, Visseren FL. Adipose tissue dysfunction in obesity, diabetes, and vascular diseases. Eur Heart J. 2008 Dec;29(24):2959-71.

16. Weisberg SP, McCann D, Desai M, Rosenbaum M, Leibel RL, Ferrante AW,Jr. Obesity is associated with macrophage accumulation in adipose tissue. J Clin Invest. 2003 Dec;112(12):1796-808.

17. Harte AL, Tripathi G, Piya MK, Barber TM, Clapham JC, Al-Daghri N, et al. NFkappaB as a potent regulator of inflammation in human adipose tissue, influenced by depot, adiposity, T2DM status, and TNFalpha. Obesity (Silver Spring). 2013 Feb 14.

18. Suganami T, Nishida J, Ogawa Y. A paracrine loop between adipocytes and macrophages aggravates inflammatory changes: Role of free fatty acids and tumor necrosis factor alpha.

Arterioscler Thromb Vasc Biol. 2005 Oct;25(10):2062-8.

19. Kanhai DA, Kappelle LJ, van der Graaf Y, Uiterwaal CS, Visseren FL, SMART Study Group.

The risk of general and abdominal adiposity in the occurrence of new vascular events and mortality in patients with various manifestations of vascular disease. Int J Obes (Lond). 2012 May;36(5):695-702.

20. Fox CS, Massaro JM, Hoffmann U, Pou KM, Maurovich-Horvat P, Liu CY, et al. Abdominal visceral and subcutaneous adipose tissue compartments: Association with metabolic risk factors in the framingham heart study. Circulation. 2007 Jul 3;116(1):39-48.

13

1

CHAPTER 1

21. Enzi G, Busetto L, Inelmen EM, Coin A, Sergi G. Historical perspective: Visceral obesity and related comorbidity in joannes baptista morgagni's 'de sedibus et causis morborum per anatomen indagata'. Int J Obes Relat Metab Disord. 2003 Apr;27(4):534-5.

22. Maury E, Ehala-Aleksejev K, Guiot Y, Detry R, Vandenhooft A, Brichard SM. Adipokines oversecreted by omental adipose tissue in human obesity. Am J Physiol Endocrinol Metab.

2007 Sep;293(3):E656-65.

23. Xu H, Barnes GT, Yang Q, Tan G, Yang D, Chou CJ, et al. Chronic inflammation in fat plays a crucial role in the development of obesity-related insulin resistance. J Clin Invest. 2003 Dec;112(12):1821-30.

24. Rytka JM, Wueest S, Schoenle EJ, Konrad D. The portal theory supported by venous drainageselective fat transplantation. Diabetes. 2011 Jan;60(1):56-63.

25. Lee MJ, Wu Y, Fried SK. Adipose tissue heterogeneity: Implication of depot differences in adipose tissue for obesity complications. Mol Aspects Med. 2013 Feb;34(1):1-11.

26. Mathivanan S, Ji H, Simpson RJ. Exosomes: Extracellular organelles important in intercellular communication. J Proteomics. 2010 Sep 10;73(10):1907-20.

27. Thery C, Ostrowski M, Segura E. Membrane vesicles as conveyors of immune responses. Nat Rev Immunol. 2009 Aug;9(8):581-93.

28. Nolte-'t Hoen EN, Wauben MH. Immune cell-derived vesicles: Modulators and mediators of inflammation. Curr Pharm Des. 2012;18(16):2357-68.

29. EL Andaloussi S, Mager I, Breakefield XO, Wood MJ. Extracellular vesicles: Biology and emerging therapeutic opportunities. Nat Rev Drug Discov. 2013 May;12(5):347-57.

30. Mack M, Kleinschmidt A, Bruhl H, Klier C, Nelson PJ, Cihak J, et al. Transfer of the chemokine receptor CCR5 between cells by membrane-derived microparticles: A mechanism for cellular human immunodeficiency virus 1 infection. Nat Med. 2000 Jul;6(7):769-75.

31. Guermonprez P, Valladeau J, Zitvogel L, Thery C, Amigorena S. Antigen presentation and T cell stimulation by dendritic cells. Annu Rev Immunol. 2002;20:621-67.

32. Valadi H, Ekstrom K, Bossios A, Sjostrand M, Lee JJ, Lotvall JO. Exosome-mediated transfer of mRNAs and microRNAs is a novel mechanism of genetic exchange between cells. Nat Cell Biol. 2007 Jun;9(6):654-9.

33. Antonyak MA, Li B, Boroughs LK, Johnson JL, Druso JE, Bryant KL, et al. Cancer cell-derived microvesicles induce transformation by transferring tissue transglutaminase and fibronectin to recipient cells. Proc Natl Acad Sci U S A. 2011 Mar 22;108(12):4852-7.

34. Buschow SI, Nolte-'t Hoen EN, van Niel G, Pols MS, ten Broeke T, Lauwen M, et al. MHC II in dendritic cells is targeted to lysosomes or T cell-induced exosomes via distinct multivesicular body pathways. Traffic. 2009 Oct;10(10):1528-42.

35. Deng ZB, Poliakov A, Hardy RW, Clements R, Liu C, Liu Y, et al. Adipose tissue exosomelike vesicles mediate activation of macrophage-induced insulin resistance. Diabetes. 2009 Nov;58(11):2498-505.

36. Aoki N, Yokoyama R, Asai N, Ohki M, Ohki Y, Kusubata K, et al. Adipocyte-derived microvesicles are associated with multiple angiogenic factors and induce angiogenesis in vivo and in vitro.

Endocrinology. 2010 Jun;151(6):2567-76.

37. Aoki N, Jin-no S, Nakagawa Y, Asai N, Arakawa E, Tamura N, et al. Identification and characterization of microvesicles secreted by 3T3-L1 adipocytes: Redox- and hormonedependent induction of milk fat globule-epidermal growth factor 8-associated microvesicles.

Endocrinology. 2007 Aug;148(8):3850-62.

38. Muller G. Microvesicles/exosomes as potential novel biomarkers of metabolic diseases.

Diabetes Metab Syndr Obes. 2012;5:247-82.

39. Muller G, Jung C, Wied S, Biemer-Daub G. Induced translocation of glycosylphosphatidylinositolanchored proteins from lipid droplets to adiposomes in rat adipocytes. Br J Pharmacol. 2009 Oct;158(3):749-70.

40. Muller G, Schneider M, Biemer-Daub G, Wied S. Microvesicles released from rat adipocytes and harboring glycosylphosphatidylinositol-anchored proteins transfer RNA stimulating lipid synthesis. Cell Signal. 2011 Jul;23(7):1207-23.

14

PART ONE

Relation of distinct visceral fat depots to metabolic disease

CHAPTER 2 The role of distinct adipose tissue depots and extracellular vesicles in obesity-induced metabolic dysfunction Manuscript in preparation

MariĂŤtte E.G. Kranendonk, Eric Kalkhoven, Frank L.J.Visseren

CHAPTER 2

ABSTRACT Obesity has become epidemic in the last decades. With subsequent increased risk for metabolic dysfunction including insulin resistance, metabolic syndrome and type 2 diabetes, obesity is an important cardiovascular risk factor. Emerging insights in the metabolic and endocrine functions of adipose tissue have broadened our understanding how adipose tissue contributes to cardiovascular disease via systemic metabolic changes, and may be the basis for novel cellular-based therapies to prevent obesity-induced cardiovascular disease. Intrinsic differences of distinct fat depots and their anatomical relation with responsive organs, such as the liver, determines for a large part how distinct fat depots contribute to metabolic diseases. Secondly, there are a few common pathways by which adipose tissue derangements affect whole body metabolism, including the adipose tissue immune cell composition, secretion of adipokines and the secretion of free fatty acids.

Furthermore, adipose tissue actively secretes extracellular vesicles, which influences both local and systemic metabolism. Although the function of extracellular vesicles in immunological diseases has been well established, their role in immunometabolism is underexposed. This review summarizes current knowledge of pathophysiological mechanisms linking abdominal adipose tissue to obesity-related metabolic dysfunction, with a special focus on distinct adipose tissue depots and the role of adipose tissue-derived extracellular vesicles.

18

Adipose tissue depots, EVs and metabolic dysfunction

INTRODUCTION Obesity (BMI ≥ 30 kg/m2) has been associated with cardiovascular risk since 400 BC (1).

However, with doubling of the worldwide prevalence of obesity over the last 20 years, obesity has become an epidemic (2,3). Obesity predisposes to metabolic comorbidities including low grade systemic inflammation, insulin resistance and lipotoxicity, subsequently leading to metabolic syndrome and type 2 diabetes (4,5). Therefore, obesity is regarded an important risk factors for cardiovascular disease (6).

Obesity is difficult to treat. While for other cardiovascular risk factors such as hypertension and dyslipidemia powerful treatments can be applied, applied, treatment of obesity is generally not very successful except for bariatric surgery. Weight loss is an effective and easy treatment, although for most people difficult to accomplish and maintain, and metabolic benefits are not always achieved (7). Secondly, several drugs to reduce weight have been developed, though none is very efficient or without serious side effects (8).

Lastly, obesity surgery is effective in the treatment for morbid obesity and obesityassociated diabetes (9,10). However, this option is only applicable for morbidly obese patients (BMI ≥ 40 kg/m2 or BMI ≥ 35 kg/m2 and obesity related metabolic complications), leaving the majority of obese patients without effective treatment options.

Adding up to the complexity of obesity, significant differences in metabolic outcome can exist between patients with the same obesity status, known as the obesity paradox (1113). However, this effect seems to be present especially in smokers, while a linear effect between BMI and mortality was observed in non-smokers (14).

The lack of efficient treatment adds to the complexity of obesity-induced metabolic complications. For a long time adipose tissue has just been recognized as an energy storage organ. Just 60 years ago it was hypothesized that adipose tissue might actually transmit feedback signals to regulate energy expenditure and food intake (15). In the following decades secretion of numerous hormones and cytokines, collectively termed adipokines, have been discovered, and their role in healthy and challenged adipose tissue has been elucidated (16-20). Besides adipokines, also the role of free fatty acids (FFA) in obesityinduced metabolic disease has been partly clarified. Furthermore it has become clear that adipose tissue harbors various immune cells, which either contribute to a healthy immunometabolism or cause derangements when adipose tissue is challenged by excess energy. On top of these piling mediators, ‘novel’ mediators termed extracellular vesicles (EVs) are described to be secreted by adipose tissue and able to influence local and distant immune cell activation and metabolism (21,22). Although the biological effects of EVs triggered much interest in their role in immunological diseases and cancer (23,24), the role of EVs in obesity-induced metabolic disease is underexposed.

In this review, the basics of adipose tissue function and adipose tissue dysfunction will first be described. In the second part, we will focus on different abdominal adipose tissue depots, and their differential contribution in obesity-induced metabolic disease. Finally, we will describe the role of EVs as intercellular communicators in AT inflammation and subsequent metabolic disease.

19

2

CHAPTER 2

Adipose tissue Human adipose tissue is one of the largest organs in the body, and is a key regulatory centre of energy storage and metabolic homeostasis by secretion of metabolic mediators.

Adipose tissue is not a single homogenous compartment, but distributed in specific depots which can be distinguished based on anatomical and intrinsic properties (25-28). There are two types of adipose tissue in the human body, white and brown adipose tissue, which regulate metabolic homeaostasis in opposite ways. White adipose tissue stores energy as triglcerides, while brown adipose tissue is specialized in the dissipation of energy trough the production of heat.

Brown adipose tissue Brown adipose tissue is only present in small depots in the human body, located mainly cervical, supraclavicular, mediastinal and pararenal (29). Brown adipocytes originate from Myf5-positive progenitor cells, are generally smaller than white adipocytes and contain multiple small lipid droplets and numerous mitochondria which gives them a brown appearance. The miochondria are functionalized by uncoupling protein 1 (UCP1), which plays a lead role in thermogenesis (30). Interestingly, brown precursor cells have also been found within white adipose tissue. Although these cells are derived from the Myf5-positive precursors, they do not show the appearance of brown adipocytes. However, upon inducement they are able to differentiate into brown adipocytes and are therefore termed beige adipocytes (31). Further literature about brown adipose tissue development, regulatory functions and the role and future potential in metabolic disease is beyond the scope of this review, but are extensively reviewed in detail (32-35).

White adipose tissue White adipose tissue is the most abundant adipose tissue in the human body, and the main distinction is made between subcutaneous adipose tissue (located immediately below the skin) and visceral adipose tissue (fat surrounding organs) (25-27). Visceral fat depots include omental (OAT), mesenteric AT (MAT) and perivascular adipose tissue (PAT) (Figure 1).

The greater omentum is a prominent peritoneal fold that hangs down like an apron from the greater curvature of the stomach and the proximal part of the duodenum, and is a richly vascularized fat depot in close proximity to the liver. Mesenteric fat lines the mesentery, a double layer of peritoneum originated from the invagination of the peritoneum by the intestines. The mesentery consists of fat, nerves, lymph nodes, blood and lymphatic vessels, and provides a means for communication between the intestines and the posterior abdominal wall. PAT is situated outside the adventitial layer and surrounds most of the systemic blood vessels. Besides these major AT compartments, AT is located among or within numerous other organs: such as the heart (pericardial AT), the muscle (intra-muscular AT), or surrounding kidneys (peri-renal AT).

Human white adipose tissue consists mainly of adipocytes, surrounded by a supportive stroma containing a heterogeneous cell population including smooth muscle cells,

20

Adipose tissue depots, EVs and metabolic dysfunction

2 SAT OAT MAT PAT

SAT

OAT

MAT

PAT

Adiponectin, IL-6 FFA

IP-10, HGF FFA

Leptin FFA

IL-8, MCP-1 TNF-Îą

GLP-1 Microbiota

Macrophages Cytokines

Myostatin IL-15

Insulin resistance Dyslipidemia

Atherosclerosis Figure 1. Inter-organ signaling between abdominal adipose tissue depots and surrounding organs.

21

CHAPTER 2

fibroblasts, pre-adipocytes, mast cells, endothelial cells and immune cells called the stromalvascular fraction (SVF) (36,37). White adipocytes are derived from Myf5-negative progenitor cells, yet derived from the embryonic mesoderm, and are unilocular cells composed of a large lipid droplet (95% of the cell volume) which varies in size (~20-200 Âľm) depending on the cell lipid content (38-40). White adipose tissue undergoes constant remodelling, including hyperplasia (proliferation of adipocytes) and hypertrophy (enlargement of adipocytes), in response to changes in energy supply (41,42).

White adipose tissue functions In the physiological state, resident immune cells, secreted adipokines and FFA conjointly regulate energy homeostasis, and influence other biological processes such as immunity, blood pressure and coagulation (43-45).

Energy storage The main function of white adipose tissue is storage of energy in the form of triglycerides (TG), which are synthesized from FFA esterified to a glyceride backbone under the control of insulin stimulated lipoprotein lipase (LPL); a key regulator in the breakdown of TG circulating in chylomicrons and very low density lipoprotein (VLDL) (46)(47). Upon energy demand, lipolysis is induced by catecholamines, activating adipose triglyceride lipase (ATGL) and hormone-sensitive lipase (HSL), two enzymes involved in hydrolysis of stored TG into FFA (48). Expression of HSL and AGTL is higher in visceral adipose tissue compared to subcutaneous adipose tissue (49,50). In contrast, lipid accumulation capacity is higher in subcutaneous fat compared to visceral fat (51).

Adipokines The regulation of lipid storage requires adequate response of adipocytes to changes in nutritional status. Sixty years ago, it was postulated that adipose tissue might actually transmit feedback signals to the hypothalamus to regulate energy expenditure and food intake (15). Since this first discovery of the endocrine function of white adipose tissue, a multitude of adipokines have been discovered to play a role in the energy regulation of white adipose tissue. Only few of them are specifically produced by adipocytes, while the majority is also produced by many other AT-resident cells from the SVF. As the function and role of adipokines has been extensively reviewed (52-55) have summarized the bestknown adipokines in Table 1.

Immunity in white adipose tissue The multitide of resident immune cells in white adipose tissue even outnumber the adipocytes, though take up only 10% in white adipose tissue mass. These immune cells, both from the innate and adaptive immune system, are important mediators in homeostasis which is recently reviewed (56). In the lean state, anti-inflammatory leukocytes predominate, including anti-inflammatory (M2) macrophages, IgE producing B cells, eosinophils, regulatory T cells, natural killer T cells, and CD4+ T cells producing IL-4, IL-10

22

Adipose tissue depots, EVs and metabolic dysfunction

and IL-13, which are the main drivers of the T helper 2 (TH2) response (57,58). These cells, in reciprocal signaling with adipocytes, promote an anti-inflammatory milieu required for tissue remodeling and immunometabolic homeostasis necessary for preservation of insulin sensitivity (43,59).

White adipose tissue dysfunctions Metabolic complications associated with obesity are a direct consequence of limited white adipose tissue expandability. The threshold for adipocyte expandability varies among individuals and among distinct white adipose tissue depots. Both murine and human studies have shown that subcutaneous pre-adipocytes replicate and differentiate more rapidly than visceral pre-adipocytes, and that subcutaneous adipocytes have a higher capacity for hyperplasia than visceral fat (51,60-62). These features, together with a higher angiogenesis potential of subcutaneous adipose tissue, might explain the difference in expanding capacity upon nutritional changes (63,64) When the storage capacity of adipocytes is surpassed, the hypertrophic adipocyte secretes abundant FFA and pro-inflammatory adipokines which function as ‘danger’ signals, subsequently leading to influx of pro-inflammatory immune cells which now in reciprocal signaling with adipocytes create a pro-inflammatory milieu. This process is referred to as adipose tissue dysfunction described in more detail below (65,66).

The role of free fatty acids in adipose tissue dysfunction In obese subjects, elevated levels of FFA are measured in the circulation, as white adipocytes are in a hyperlipolytic state that is resistant to the antilipolytic effect of insulin.

The mechanisms by which supraphysiological levels of FFA contribute to obesity-induced metabolic disease are complex and far from clarified. Many different FFA are shown to play a role in the development of insulin resistance, via multiple mechanisms including activation of inflammatory pathways in peripheral tissues via innate receptors, endoplasmatic reticulum stress or inflammasome activation (67,68) (Figure 2).

FFA have been proposed to represent self-derived patterns able to stimulate toll like receptors (TLRs), although direct binding of FFA to TLRs is still under debate (69). Nevertheless, TLR4 and to a lesser extent TLR2 are suggested to activate intracellular inflammatory pathways including IκB Kinase (IKK), JNK and NFκB in response to saturated fatty acids (70-72).

Activation of these pathways results in subsequent systemic inflammation and inhibition of IRS-1 phosphorylation which promotes insulin resistance (73,74).

The inflammasome is an intracellular multimeric complex involved in the activation of the NLRP3 (nucleotide-binding domain, leucine-rich-containing family, pyrin domain-containing-3) and caspase-1 pathways (75), inducing maturation of pro-inflammatory cytokines such as IL-1β and IL-8. The inflammasome is activated in obesity, while NLRP3 expression is reduced in adipocytes upon weight loss (76,77) In compbination with LPS (lipopolysaccharide), of which elevated levels are measured in the circulation of obese subjects (78), FFA can activate the inflammasome and thereby contribute to a proinflammatory milieu (75,79).

23

2

24

IL-6 (172,173)

TNF-α (171)

Leptin (20)

Adiponectin (19)

Adipokine

OAT

Adipocytes, SVF, liver, muscle

None

Adipocytes, SVF

MAT

Adipocytes

SAT

Adipocytes

Primary source

Pro-inflammatory: - Inhibits PPARγ and thereby inhibits maturation of M2 macrophages - Binds IL6R → activates NFκB via IKK and AP-1 via JNK Insulin resistant: - Suppresses SOCS3 trough IL6R, which directly inhibits activity of insulin receptor and IRS, inhibiting downstream activation of Akt, and activates JNK and IKK further inhibiting insulin signaling - Inhibits transcription of IRS-1, GLUT-4 and PPARγ - Reduces LPL activity - Pro-inflammatory cytokines transcribed by NFκB and AP-1 interfere with insulin signaling pathway

Pro-inflammatory - Binds TNFR → activates NFκB via IKK and AP-1 via JNK Insulin resistant: - Pro-inflammatory cytokines transcribed by NFκB and AP-1 interfere with insulin signaling pathway - Suppresses SOCS3 trough TNFR, which directly inhibits activity of insulin receptor and IRS, inhibiting downstream activation of Akt, and activates JNK and IKK further inhibiting insulin signaling - Reduces LPL activity - Interferes with insulin receptor - Inhibits IRS-1 phosphorylation and GLUT4 expression

- Regulates food intake and energy expenditure via LEPR in central nervous system and intestine Pro-inflammatory: - Regulates Th1 polarization{{7037 Matarese,G. 2010 Insulin resistant: - Antagonizes insulin-signaling trough Jak-STAT pathway

Anti-inflammatory: - Binds adipoR1 (skeletal muscle), adipoR2 (liver) or T-cadherin (endothelial cells) - Activates PPARα trough AMPK (liver) which suppresses NFκB - Suppression of NFκB inhibits transcription of pro-inflammatory cytokines Insulin sensitizing - Activates Akt-phosphorylation - Increases glucose uptake via Akt-induced translocation of GLUT4 - Increases lipid uptake via CD36 up regulation trough AMPK - Increases fatty acid oxidation in liver trough adipoR2

Role in adipose tissue function

Table 1. Adipokines in metabolic and vascular diseases.

- Positive

relation with obesity, insulin resistance

-P ositive relation with obesity, insulin resistance

Positively with obesity, insulin resistance

- Positive relation with insulin sensitivity -N egative relation with obesity, insulin resistance (19)(169) - I s repressed by TNF-α and IL-6, thereby repressed inhibition of NFκB (170)

Relation to metabolic and vascular diseases

CHAPTER 2

No specific fat depot

Liver, adipocytes, SVF

No specific fat depot

Monocytes and macrophages

No specific fat depot

Adipocytes, SVF

No specific fat depot

- Is induced by Il-6 and TNF-α -P ositive relation with obesity, insulin resistance - Positive relation with atherosclerosis

Pro-inflammatory: - Promotes expression of Il-6, TNF-α, VCAM-1, ICAM1, involved in endothelial dysfunction Insulin resistant: - Decreases insulin sensitivity in muscle and liver - Pro-inflammatory cytokines transcribed by NFκB interfere with insulin signaling pathway - Counteracts adiponectin

-P ositive relation with obesity, insulin resistance

-P ositive relation with obesity, insulin resistance and liver steatosis - Expression is increased by TNF-α

Pro-inflammatory: - Recruits monocytes and T lymphocytes in adipose tissue trough CCR2 - Induces ICAM-1 expression Insulin resistant: - Pro-inflammatory cytokines by monocytes and T lymphocytes interfere with insulin signaling pathway

Carrier protein of retinol Pro-inflammatory: - activates JNK trough TLR4 independent of STRA6 and retinol (179) Insulin resistant: - Inhibits P13K, IRS-1 and insulin receptor

- Positive

relation with obesity, insulin resistance (165) - Positive relation with atherosclerosis

Abbreviations: AdipoR: adiponectin receptor; PPARα/PPARγ: Peroxisome proliferator activated gamma alpha/gamma; AMPK: AMP-activated protein kinase; NFκB: nuclear factor-κB; Akt : protein kinase B; GLUT4: glucose transporter 4; LEPR: leptin receptor; Th1: T helper 1; SVF: stromal vascular fraction, TNF-α: tumor necrosis factor alpha; TNFR: TNF receptor; JNK: c-Jun N-terminal kinases; SOCS-3: suppressor of cytokine signalling-3; IRS-1: insulin receptor substrate 1; IKK = IκB Kinase; AP-1: activator protein 1; LPL: lipoprotein lipase; IL-6R: IL-6 receptor, MIF: macrophage migration inhibitory factor; CRP: c-reactive protein; ICAM1: intercellular adhesion molecule 1; JAB-1: c-Jun activation domain binding protein-1; intracellular receptor for MIF; ERK1/2, extracellular-signal regulated kinase 1/2; CCR2: receptor for MCP-1; VCAM-1: vascular cell adhesion molecule 1; RBP4: retinol binding protein 4; TLR4: toll like receptor 4; STRA6: RBP4 receptor; P13K: phosphatidylinositide 3-kinases

RBP4 (177,178)

Resistin (176)

CCL2 / MCP1 (173,175)

MIF (164,174)

Adipocytes, SVF, liver, skeletal muscle

Anti-inflammatory: - Regulates inflammation: binds JAB1 → inhibits JNK and ERK activation of AP-1(163) - Compromises TNFα-induced JNK phosphorylation Pro-inflammatory: - Macrophage recruitment: MIF knockout mice show reduced macrophages and inflammatory cytokines in white adipose tissue and liver - Stimulates expression of CRP, ICAM-1 Insulin resistant: - MIF knockout mice show reduced JNK levels and improved insulin sensitivity - MIF increases glucose induced insulin-secretion

Adipose tissue depots, EVs and metabolic dysfunction

2

25

CHAPTER 2

The endoplasmic reticulum (ER) is a central cellular organelle involved in synthesis, folding and maturation of transmembrane and secretory proteins. In obesity, the increased demand for protein synthesis under nutrient excess causes ER stress in hepatocytes and adipocytes.

In short, PERK (PRKR-like endoplasmic reticulum kinase), IRE-1a (inositol-requiring enzyme 1a) and ATF6 (activating transcription factor 6) play a major role in this stress response (80,81). Increased phosphorylation of these ER-stress sensor proteins enhance JNK, IKK and NFκB activity downstream of these sensor proteins, inducing inflammation and altered adipokine-secretion in enlarged adipocytes (82-84). Saturated FFA were shown to induce an ER stress response in skeletal muscle, although the precise link with subsequent insulin resistance is less clear (85) (Figure 2).

Adipokines in white adipose tissue dysfunction Adipokines interfere with metabolic and vascular homeostasis via multiple mechanisms, though the most important are induction of the inflammatory pathway (mainly trough activation of NFκB) by pro-inflammatory cytokines such as TNF-α and IL-6 or repressed inhibition by reduced adiponectin (Table 1, Figure 2) and inhibition of the insulin pathway (trough tenuation of the insulin receptor, IRS1 or JNK pathway) summarized in Table 1.

Immune cells in white adipose tissue dysfunction The role of adipose tissue resident immune cells in obesity-induced metabolic dysfunction has gained much interest in the past decades and has been extensively reviewed (56,86).

In brief, a marked alteration in adipose tissue immune cells is observed with obesity. The leukocyte population shifts toward a pro-inflammatory repertoire including pro-inflammatory

TNFα

TNF R

IL-6

IL-6 R

Adiponectin

AdipoR1/2

NFκB

PPARα

Adipokines

Saturated FFA Free fatty acids

RBP4

Extracellular Vesicles

miRNA

TLR 4

Ceramides DAG

ER stress

NFκB JNK / IKK

Inflammasome

Caspase-1

TLR 4 / TRIF

IL-6 TNFα

?

AKTp

Leptin

Lipolysis

Figure 2. Mechanisms by which AT secretome contributes to metabolic complications of obesity.

26

Adipose tissue depots, EVs and metabolic dysfunction

(M1) macrophages, interleukin-2 (IL-2) and TNF-Îą secreting cytotoxic CD8+ T cells, IFN-Îł (interferon gamma) producing CD4+ T cells, mast cells and B cells producing complementfixing IgG antibodies, known as the T helper 1 (TH1) response (87,88). This leukocyte shift induces a pro-inflammatory milieu with lower secretion of anti-inflammatory adipokines such as adiponectin, and higher secretion of pro-inflammatory cytokines preceding insulin resistance (89,90).

The mechanisms underlying the leukocyte shift in obese white adipose tissue is still not clear, though for some time now the idea has grown that white adipose tissues itself controls the immune response, rather than the systemic immune system (91). For example, adipocyte-secreted mediators such as FFA resemble endotoxins and can activate the same pattern recognition receptors such as TLR4 (toll-like receptor 4) (92). Also cytokines such as TNF-Îą, IL-6 and MCP-1, secreted by stressed adipocytes, activate inflammation pathways in resident immune cells (17,26,93). Furthermore, changes in T cells have also been reported to precede changes resident adipose tissue macrophages during obesity (90).

Although it is unclear how those T cells are activated, adipocytes were shown to directly stimulate CD4-T cells, and that adipocytes are equipped with functional MHCII and costimulatory molecules (94). Finally, recent insights have shown that adipocytes themselves can function as antigen presenting cells, presenting (yet unknown) lipid antigens to natural killer T cells (43). As B cell receptors bind small epitopes on proteins, carbohydrates and lipids, it is imaginable that adipocytes present these to the B cell receptors.

Distinct adipose tissues Intrinsic differences between distinct white adipose tissue depots regarding TG storage and lipolysis as well as secretion of distinct adipokines contribute to a large extent to obesity-induced metabolic dysfunction. This has been shown in AT transplantation studies, where transplantation of subcutaneous fat into the visceral cavity partly rescued development of IR in obese mice (95). Also epidemiological studies repeatedly indicate that overabundance of visceral fat is the main contributor to complications of obesity, though no distinction is made in the different VAT depots (25,96.97) Although much is known about differences among subcutaneous and visceral fat, less is known about the intrinsic differences among distinct visceral fat depots. Furthermore, the concept of white adipose tissue as an endocrine organ focused much work on the alterations in signalling pathways in obesity, while less attention has been paid to putative reciprocal signalling of white adipose tissue depots with surrounding organs. A better knowledge of interorgan exchange of molecular mediators between adipose tissue and distinct target organs might reveal novel insights in underlying pathways of obesity-induced metabolic and vascular diseases.

Subcutaneous adipose tissue and skeletal muscle Subcutaneous adipose tissue (SAT; Figure1) is shielded from abdominal organs and vessels by skeletal muscle, and muscle secreted myokines such as myostatin and IL-15 have been shown to modulate subcutaneous, but not visceral AT function and mass and increase

27

2

CHAPTER 2

insulin sensitivity in mice (98-100). Effects of AT-derived mediators, though not specifically SAT-derived, on skeletal muscle has also been frequently demonstrated, such as TNFα, IL-6 and FFA (101,102). Furthermore, subcutaneous fat secretes the highest levels of adiponectin, an insulin sensitive adipokine which is produced at a lower rate in obesity (19).

Skeletal muscle is a major site of adiponectin action via expression of the adiponectin receptor, upon which adiponectin increases the expression of genes involved in fatty acid oxidation via transcriptional activity of PPARα (103). Subsequently, administration of adiponectin was shown to enhance insulin sensitivity via decreased circulating FFA levels (104) (Figure 1).

Omental adipose tissue and liver Visceral fat is located intraperitoneally, and studies suggests that distinct visceral fat depots contribute differently to metabolic and vascular risk due to interaction of their surrounding organs (105). Ex vivo studies regarding secretion of bio-active adipokines revealed that VAT has a higher inflammatory status, higher secretory capacity and is associated with more circulating substances related to inflammation and oxidative stress than subcutaneous fat (18,106,107). Omental adipose tissue (OAT) specifically secretes high numbers of soluble chemokines and cytokines, which are involved in macrophage recruitment and activation such as IL-6, MCP-1, MIF, TNF-α and RBP-4 (17,108,109). The close proximity of OAT to the liver enables all metabolic substrates to drain directly to the liver, known as the portal hypothesis (102,110). OAT inflammation (macrophages, expression of chemerin, IP-10 (interferon gamma-induced protein 10) or TIMP-1 (TIMP metallopeptidase inhibitor 1)) has been associated frequently with liver associated pathology, such as hepatic insulin resistance, dyslipidemia and hepatosteatosis (108, 111-115). Furthermore, hsCRP levels which reflect low-grade systemic inflammation are positively related to increase in visceral adipose tissue (116). However, no distinction in the differential contribution of mesenteric or omental fat could be made. Finally, although signaling from OAT toward the liver has been frequently showed, a signaling rout from the liver specifically towards OAT seems unlikely (Figure 1).

Mesenteric adipose tissue and intestine Due to difficulties in the collection of mesenteric adipose tissue (MAT), less is known about the inflammatory status of MAT. A few reports have shown that lipid accumulation as well as lipolysis capacity is higher in MAT compared to OAT. MAT does not display a very active adipokine secretion profile compared to other visceral white fat depots, though secretion of leptin and inflammatory characteristics such as adipocyte necrosis are more prevalent in MAT (49, 50, 109, 117).

MAT is in close proximity to the intestine and several studies pointed towards a close interaction between AT and the enteric endocrine system. Both mice and humans suffering from colitis showed specific MAT inflammation in contrast to other AT depots (118, 119).

Also changes in diet and gut microbiota have a profound effect on MAT (120-123).

Interestingly, inflammation in MAT specifically has been associated with diabetes in contrast

28

Adipose tissue depots, EVs and metabolic dysfunction

to other VAT depots (50). Furthermore, obesity has been associated with decreased levels of intestinal insulin-sensitizing hormones such as GIP (glucose-dependent insulinotropic polypeptide) and GLP-1 (glucagon-like peptide-1).

The mechanisms by which mesenteric fat might influence insulin resistance remain elusive.

MAT is an active secretor of leptin (Table 1), which is a potent stimulator of GLP-1 whereas leptin resistance has been associated with diminished GLP-1 secretion and subsequent insulin resistance (124). Moreover, following a very low caloric diet, circulating leptin levels decreased while soluble leptin receptor levels increased, together with an increase in insulin sensitivity in obese females (125). These findings suggest an indirect but important role for leptin in insulin signaling via the enteric endocrine system. Although still much is unknown, a reciprocal signaling cascade is possible, where changes in diet and microbiota influence MAT mass and function, and subsequent secretion of adipokines could affect incretin function (Figure 1).

Perivascular adipose tissue and the vascular wall Fat lining the greater vessels and epicardial adipose tissue, PAT, behaves somewhat differently compared to other white adipose tissue depots. Perivascular adipocytes are smaller and display a reduced differentiation capacity compared to other VAT and SAT depots (26,126). Interestingly, markers of brown adipose tissue are expressed in PAT, which is believed to generate warmth during cold exposure, to protect against ventricular arrhythmias (127,128). Furthermore, PAT is suggested to have both protective and pathologic actions through outside- to-inside signaling in an autocrine and paracrine fashion (129,130) From (autopsy) studies it is known that atherosclerotic plaque size is highly related to the quantity of peri-coronary adipose tissue and the amount of PAT-resident macrophages, and that PAT located close to atherosclerosis secreted high amounts of pro-inflammatory and angiogenic genes (126,131,132). PAT secretes high amounts of pro-inflammatory adipokines such as TNF-Îą, IL-8 and MCP-1, which are strongly related to atherosclerosis, but not to systemic metabolic dysfunction (26,105,133) (Figure 1).

New kids on the block: the role of extracellular vesicles in obesity-induced metabolic dysfunction The mechanisms underlying metabolic disease described above include direct cell-cell contact and signaling by soluble molecules. In addition, extracellular vesicles (EVs) are actively secreted by adipose tissue of mice and humans, and seem to play a role in insulin resistance as well (21,22). Recent molecular and conceptual insights into these vesicles have opened new avenues to exploit their exciting features for the development of new therapeutic tools.

Extracellular vesicle biogenesis and composition EVs are small membrane-derived vesicles (50-1000 nm in size) which play an important role in intercellular communication. They are released by most eukaryotic cells and have been sub-divided into exosomes, microvesicles and microparticles based on their synthesis

29

2

CHAPTER 2

pathway, size and content. Exosomes are of endosomal origin, where intraluminal vesicles (ILVs) are formed by inward budding of the limiting membrane of the endosome resulting in multivesicular bodies (MVBs) (134). Upon fusion of MVBs with the plasma membrane, ILVs are released and termed exosomes. Microvesicles and microparticles are generated via outward budding from the plasma membrane and incorporate plasma membrane proteins in that process.

For a long time, this process was suggested to represent a mechanism for reticulocytes to dispose of obsolete proteins (135). However, since the discovery of immunomodulatory effects of EVs (136), intensive research into regulation of their biogenesis, composition and function has emerged. Studies revealed that the biogenesis of EVs and trafficking of specific cargo into EVs is a highly regulated process, driven by numerous lipids and proteins.

Involved lipids include sphingomyelinase which by formation of ceramide from sphingomyeline triggers budding of ILVs into MVBs (137). Important proteins include ESCRT proteins, Rab-proteins and lipid related proteins such as phospholipid scramblase, which play a role in sorting of proteins into ILVs, and trafficking and excretion of exosomes (138,139). Furthermore, flotillins associate with membrane microdomains enriched in cholesterol and sphingolipids and mediate endocytosis, intracellular transport and signalling (140-142). Interestingly, sorting of the antigen presenting protein MHCII into ILVs is independent of MHCII ubuquitination in contrast to lysosomal targeting, and seems dependent on the tetraspanin CD9 (143).

However, the exact mechanisms of biogenesis and regulated composition of EVs is complicated and still largely unknown (137,144,145). Furthermore, besides known differences in the biogenesis and cargo sequestration of distinct extracellular vesicle types, there is overlap as well disabling current discrimination between vesicle types based on specific markers (23,145-148). Therefore, in this review we do not discriminate between vesicle types, but solely describe common concepts by which EVs, and particularly adipose tissue derived EVs, can exert immunomodulatory mechanisms contributing to metabolic disease.

Extracellular vesicle immunomodulatory properties The potential of EVs as unique intercellular messengers is based on several features of EVs. First of all the regulated packaging of cargo ensures signaling of specific biological signals from the donor cell to recipient cells. Secondly, due to their inverse topology, intravesicular components (proteins, mRNAs and microRNAs) are protected from degradation, and components at the vesicular surface (transmembrane, soluble and glycosylphosphatidylinolsitol (GPI)-anchored proteins) can interact with target cells (134,149). Third, fusion of EVs with target cells allow their cargo to enter the target cell, enabling subsequent protein translation of transferred mRNA, or reutilization of membrane proteins, thereby modulating features of the target cell with donor cell material (148,150).

Finally, EVs are stable in plasma, and can therefore affect target tissues in a paracrine and/ or endocrine manner (145).

30

Adipose tissue depots, EVs and metabolic dysfunction

Due to their unique signalling properties, the possible immunomodulatory functions of EVs have gained much interest. By transporting ligands and receptors, EVs can modulate immune responses. EVs derived from antigen presenting cells such as macrophages, B cells and dendritic cells are equipped with antigen presenting and co-stimulatory proteins including MHC class I and II and CD86, and are capable of inducing an antigen-dependent CD8+ T-cell-mediated immune response (136,151,152). Furthermore, modulation of cells by microparticles was observed in HIV-1 infection, where the chemokine receptor CCR5 was transferred to CCR5-negative cells, which may lead to infection of tissues without endogenous CCR5 expression (153).

Extracellular vesicles in metabolic diseases Despite advances made regarding EV function in immunology, the role of EVs in obesityinduced metabolic disease is underexposed. Given the important role of intercellular communication for white adipose tissue in whole body metabolism, and the crucial role for immune cell activation in adipose tissue inflammation preceding obesity-induced metabolic disease, adipose tissue EVs (AT-EVs) could be important mediators in maintaining energy homeostasis and endocrine (patho)physiological signaling in obesity (154).

Epidemiological evidence suggests that levels of EVs or EV-associated proteins may be related to metabolic disease (155-157). Although experimental evidence for this hypothesis is scarce, a role for AT-EVs has been shown in lipid transport, adipose tissue inflammation and systemic insulin resistance.

The first evidence of adipose tissue microvesicles was shown in 1964, were it was postulated that adipocyte microvesicles may represent the mechanism of transport of FFA in fat cells and endothelium (38). More recent reports following appropriate isolation procedures for EVs have shown that EVs are actively secreted by murine adipocytes or adipose tissue (21,158). Detailed study of lipid transport between adipocytes in rats indeed showed that the GPI-anchored microvesicle proteins Gce1 and CD73 together with specific microRNAs were transferred from large adipocytes to intracellular lipid droplets of small adipocytes, and subsequently increased lipolysis in part by increased expression of leptin (159,160) (Figure 2).

We have recently shown that human adipocytes and AT also secrete EVs, which induced upregulation of pro- as well as anti-inflammatory (IL-6, MIP-1Îą and IL-10) proteins in macrophages and subsequent insulin resistance in adipocytes, indicating that AT-EVs could play a role in a reciprocal pro-inflammatory loop underlying adipose tissue dysfunction (22,161). These data are in support of Deng et al. which showed that injection of fat tissuederived EVs from obese mice injected into lean mice (i) were taken up by monocytes, (ii) promoted higher IL-6 and TNF-Îą levels and (iii) induced systemic insulin resistance (21).

Besides insulin resistant effects of activated macrophages, AT-EVs interfere in insulin signaling via other mechanisms. In vitro data showed that pretreatment of hepatocytes with human AT-EVs directly inhibited insulin-induced Akt phosphorylation, and that pretreatment of myocytes with bone marrow cells stimulated with AT-EVs of obese mice inhibited insulin-induced Akt-phosphrylation indicating an indirect effect (Figure 2) (Chapter

31

2

CHAPTER 2

5 and (21). A possible pathway by which AT-EVs induce insulin resistance in liver or muscle cells is via the TLR4/TRIF pathway. Murine AT-derived EVs could no longer induce cytokine secretion in macrophages derived from TLR4- or TRIF-deficient mice, in contrast to macrophages from TLR2-/-, MyD88-/- or wild type mice (21).

Several adipokines, shown to be present on adipocyte EVs, might be involved in AT-EV induced macrophage activation and subsequent insulin resistance (Figure 3). Deng et al suggested that macrophage activation by mouse AT-EVs was at least partly due to EVassociated RBP4 (21). Human adipocyte and AT-EVs contain numerous adipokines either involved in insulin sensitivity (adiponectin) or associated with obesity-related insulin resistance (such as IL-6, MIF, TNF-α (Table 1)). In human adipocyte EVs, relative high amounts of MIF (an adipokine with chemotactic properties) were detected in lysed EVs in contrast to unlysed supernatant (22), suggesting that MIF signals via intravesicular transport, supporting the current understanding that MIF is secreted via a non-classical export route (162). Interestingly, MIF exerts its effects by activating an intracellular receptor c-Jun activation domain binding protein-1 (JAB1), upon endocytosis of circulating MIF (163). As one of the transport mechanisms of EV includes fusion of the vesicles with the target plasma membrane and subsequent intracellular release of donor-specific cargo, stable transport of MIF by EVs might provide a novel mechanism in inter-cellular signaling of MIF. The functional role of MIF in EVs has to be established. MIF is frequently associated with obesity, insulin resistance or diabetes in humans, and MIF knockout mice show reduced JNK levels and improved insulin sensitivity (Table 1) (164-166).

Vesicle associated and adipocyte associated protein content of human or rodent EVs

Lipid rafts

RNAs

Adipokines Membrane transport and fusion

Exosome biogenesis

MIF * RBP-4 * † TNF-α * IL-6 * MCP-1 * Resisitin *

PAI * † Complement C3 † FABP4 * † Adiponectin * † Leptin * †

Vesicle associated proteins Tetraspanins: CD9, CD63 * † MFGE8 † RAB proteins † FLotillin 1 *

Adipokines

* detected on human adipose tissue EVs † detected on murine adipose tissue EVs Figure 3. Vesicle associated and adipocyte associated protein content of human or rodent adipose tissue EVs.

32

Adipose tissue depots, EVs and metabolic dysfunction

However, in vitro studies suggest anti-inflammatory properties as intracellular binding of MIF to JAB1 inhibited TNF-Îą-induced JNK activity and repressed the pro-inflammatory transcription factor AP-1 (163,167). Although the mechanistical role of MIF in obesityinduced insulin resistance remains unclear, the stable transport of MIF via EVs might provide another mechanisms by which MIF contributes to disease (168).

Extracellular vesicle differences among adipose tissue depots Given the promising role of AT-EVs in obesity-induced metabolic disease, it would be interesting to know whether AT-EVs of distinct adipose tissue depots contribute differently to metabolic disease. We have shown that the number of EVs from omental adipose tissue, but not subcutaneous adipose tissue, was elevated in insulin resistant state, and positively related to HOMA-IR levels in human subjects (22). Furthermore, EVs derived from omental EVs contained significantly higher levels of IL-6, MIF and MCP-1 compared to subcutaneous derived EVs in a small set of patients (Chapter 5). Although preliminary data showed that omental derived EV-associated levels of IL-6 and MIF were related to lower insulin-induced Akt phosphorylation (Chapter 5), mechanistic studies will be required to further elucidate the role of adipose tissue EV in obesity-induced metabolic and vascular disease.

In conclusion, EVs are intercellular signaling vesicles actively secreted by AT that can play a role in local insulin resistance -putatively via activation of macrophages- and also in systemic insulin resistance. Although the precise mechanisms by which AT EVs induce insulin resistance are not clarified, the capacity of EVs to act in a paracrine and endocrine manner to influence target cells via a multitude of bioactive factors suggests that AT-derived EV might represent novel mediators by which AT contributes to obesity-induced metabolic disease, including insulin resistance.

Conclusions In conclusion, AT is an active endocrine organ, which in reciprocal signaling with adipose tissue resident immune cells contribute to immunometabolic homeostasis necessary for preservation of insulin sensitivity in the lean state. In the obese state, the leukocyte shift, increased lipolysis and altered adipokine secretion results in a pro-inflammatory milieu preceding insulin resistance. Despite this knowledge, the lack of effective treatments to prevent obesity-induced metabolic disease indicates a pressing need for reconsideration of signaling pathways by which adipose tissue of obese patients induces metabolic and vascular disease. Evidence suggests that inter-organ communication of specific adipose tissues with surrounding organs could contribute to obesity induced metabolic disease via distinct mechanisms. Furthermore, besides secretion of soluble biological active proteins and free fatty acids, white adipose tissue secretes EVs in which multiple bio-active factors are integrated. Given the multitude of bio-active factors present in adipose tissue EVs and their proven role in obesity-induced metabolic disease in a soluble form, it can be hypothesized that adipose tissue derived EVs play a role in the development of obesityinduced metabolic and vascular disease in a multifactorial manner.

33

2

CHAPTER 2

REFERENCES 1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

34

Haslam D, Rigby N. A long look at obesity. Lancet. 2010 Jul 10;376(9735):85-6.

Finucane MM, Stevens GA, Cowan MJ, Danaei G, Lin JK, Paciorek CJ, et al. National, regional, and global trends in body-mass index since 1980: Systematic analysis of health examination surveys and epidemiological studies with 960 country-years and 9.1 million participants. Lancet.

2011 Feb 12;377(9765):557-67.

Kral JG, Kava RA, Catalano PM, Moore BJ. Severe obesity: The neglected epidemic. Obes Facts. 2012 Apr 25;5(2):254-69.

Grundy SM, Brewer HB,Jr, Cleeman JI, Smith SC,Jr, Lenfant C, American Heart Association, et al. Definition of metabolic syndrome: Report of the national heart, lung, and blood Institute/ American heart association conference on scientific issues related to definition. Circulation.

2004 Jan 27;109(3):433-8.

Kuller LH, Velentgas P, Barzilay J, Beauchamp NJ, O'Leary DH, Savage PJ. Diabetes mellitus: Subclinical cardiovascular disease and risk of incident cardiovascular disease and all-cause mortality. Arterioscler Thromb Vasc Biol. 2000 Mar;20(3):823-9.

Eckel RH, Krauss RM. American heart association call to action: Obesity as a major risk factor for coronary heart disease. AHA nutrition committee. Circulation. 1998 Jun 2;97(21):2099-100.

Anderson JW, Konz EC, Frederich RC, Wood CL. Long-term weight-loss maintenance: A metaanalysis of US studies. Am J Clin Nutr. 2001 Nov;74(5):579-84.

Shi Y, Burn P. Lipid metabolic enzymes: Emerging drug targets for the treatment of obesity. Nat Rev Drug Discov. 2004 Aug;3(8):695-710.

Shin AC, Berthoud HR. Obesity surgery: Happy with less or eternally hungry? Trends Endocrinol Metab. 2013 Feb;24(2):101-8.

Schauer PR, Kashyap SR, Wolski K, Brethauer SA, Kirwan JP, Pothier CE, et al. Bariatric surgery versus intensive medical therapy in obese patients with diabetes. N Engl J Med. 2012 Apr 26;366(17):1567-76.

Barbarroja N, Lopez-Pedrera R, Mayas MD, Garcia-Fuentes E, Garrido-Sanchez L, MaciasGonzalez M, et al. The obese healthy paradox: Is inflammation the answer? Biochem J. 2010 Aug 15;430(1):141-9.

Phillips CM. Metabolically healthy obesity: Definitions, determinants and clinical implications.

Rev Endocr Metab Disord. 2013 Aug 9.

Carnethon MR, De Chavez PJ, Biggs ML, Lewis CE, Pankow JS, Bertoni AG, et al. Association of weight status with mortality in adults with incident diabetes. JAMA. 2012 Aug 8;308(6):58190.

Tobias DK, Pan A, Jackson CL, O'Reilly EJ, Ding EL, Willett WC, et al. Body-mass index and mortality among adults with incident type 2 diabetes. N Engl J Med. 2014 Jan 16;370(3):233-44.

KENNEDY GC. The role of depot fat in the hypothalamic control of food intake in the rat. Proc R Soc Lond B Biol Sci. 1953 Jan 15;140(901):578-96.

Cook KS, Min HY, Johnson D, Chaplinsky RJ, Flier JS, Hunt CR, et al. Adipsin: A circulating serine protease homolog secreted by adipose tissue and sciatic nerve. Science. 1987 Jul 24;237(4813):402-5.

Hotamisligil GS, Shargill NS, Spiegelman BM. Adipose expression of tumor necrosis factoralpha: Direct role in obesity-linked insulin resistance. Science. 1993 Jan 1;259(5091):87-91.

Wentworth JM, Naselli G, Brown WA, Doyle L, Phipson B, Smyth GK, et al. Pro-inflammatory CD11c+CD206+ adipose tissue macrophages are associated with insulin resistance in human obesity. Diabetes. 2010 Jul;59(7):1648-56.

Scherer PE, Williams S, Fogliano M, Baldini G, Lodish HF. A novel serum protein similar to C1q, produced exclusively in adipocytes. J Biol Chem. 1995 Nov 10;270(45):26746-9.

Zhang Y, Proenca R, Maffei M, Barone M, Leopold L, Friedman JM. Positional cloning of the mouse obese gene and its human homologue. Nature. 1994 Dec 1;372(6505):425-32.

Deng ZB, Poliakov A, Hardy RW, Clements R, Liu C, Liu Y, et al. Adipose tissue exosomelike vesicles mediate activation of macrophage-induced insulin resistance. Diabetes. 2009 Nov;58(11):2498-505.

Adipose tissue depots, EVs and metabolic dysfunction

22. Kranendonk ME, Visseren FL, van Balkom BW, Nolte-'t Hoen EN, van Herwaarden JA, de Jager W, et al. Human adipocyte extracellular vesicles in reciprocal signaling between adipocytes and macrophages. Obesity (Silver Spring). 2013 Dec 12.

23. Nolte-'t Hoen EN, Wauben MH. Immune cell-derived vesicles: Modulators and mediators of inflammation. Curr Pharm Des. 2012;18(16):2357-68.

24. EL Andaloussi S, Mager I, Breakefield XO, Wood MJ. Extracellular vesicles: Biology and emerging therapeutic opportunities. Nat Rev Drug Discov. 2013 May;12(5):347-57.

25. Fox CS, Massaro JM, Hoffmann U, Pou KM, Maurovich-Horvat P, Liu CY, et al. Abdominal visceral and subcutaneous adipose tissue compartments: Association with metabolic risk factors in the framingham heart study. Circulation. 2007 Jul 3;116(1):39-48.

26. Chatterjee TK, Stoll LL, Denning GM, Harrelson A, Blomkalns AL, Idelman G, et al.

Proinflammatory phenotype of perivascular adipocytes: Influence of high-fat feeding. Circ Res. 2009 Feb 27;104(4):541-9.

27. Shen W, Wang Z, Punyanita M, Lei J, Sinav A, Kral JG, et al. Adipose tissue quantification by imaging methods: A proposed classification. Obes Res. 2003 Jan;11(1):5-16.

28. Enerback S. Human brown adipose tissue. Cell Metab. 2010 Apr 7;11(4):248-52.

29. Cypess AM, White AP, Vernochet C, Schulz TJ, Xue R, Sass CA, et al. Anatomical localization, gene expression profiling and functional characterization of adult human neck brown fat. Nat Med. 2013 May;19(5):635-9.

30. Nicholls DG, Locke RM. Thermogenic mechanisms in brown fat. Physiol Rev. 1984 Jan;64(1):164.

31. Wu J, Bostrom P, Sparks LM, Ye L, Choi JH, Giang AH, et al. Beige adipocytes are a distinct type of thermogenic fat cell in mouse and human. Cell. 2012 Jul 20;150(2):366-76.

32. Nedergaard J, Cannon B. The changed metabolic world with human brown adipose tissue: Therapeutic visions. Cell Metab. 2010 Apr 7;11(4):268-72.

33. Kozak LP. Brown fat and the myth of diet-induced thermogenesis. Cell Metab. 2010 Apr 7;11(4):263-7.

34. Enerback S. Human brown adipose tissue. Cell Metab. 2010 Apr 7;11(4):248-52.

35. Wu J, Cohen P, Spiegelman BM. Adaptive thermogenesis in adipocytes: Is beige the new brown? Genes Dev. 2013 Feb 1;27(3):234-50.

36. Pettersson P, Cigolini M, Sjostrom L, Smith U, Bjorntorp P. Cells in human adipose tissue developing into adipocytes. Acta Med Scand. 1984;215(5):447-51.

37. Poulos SP, Hausman DB, Hausman GJ. The development and endocrine functions of adipose tissue. Mol Cell Endocrinol. 2009 Dec 16.

38. WILLIAMSON JR. Adipose tissue. morphological changes associated with lipid mobilization. J Cell Biol. 1964 Jan;20:57-74.

39. Hirsch J, Batchelor B. Adipose tissue cellularity in human obesity. Clin Endocrinol Metab. 1976 Jul;5(2):299-311.

40. Engfeldt P, Arner P. Lipolysis in human adipocytes, effects of cell size, age and of regional differences. Horm Metab Res Suppl. 1988;19:26-9.

41. Spalding KL, Arner E, Westermark PO, Bernard S, Buchholz BA, Bergmann O, et al. Dynamics of fat cell turnover in humans. Nature. 2008 Jun 5;453(7196):783-7.

42. Arner P, Bernard S, Salehpour M, Possnert G, Liebl J, Steier P, et al. Dynamics of human adipose lipid turnover in health and metabolic disease. Nature. 2011 Sep 25;478(7367):110-3.

43. Schipper HS, Rakhshandehroo M, van de Graaf SF, Venken K, Koppen A, Stienstra R, et al.

Natural killer T cells in adipose tissue prevent insulin resistance. J Clin Invest. 2012 Aug 6.

44. Samad F, Pandey M, Loskutoff DJ. Tissue factor gene expression in the adipose tissues of obese mice. Proc Natl Acad Sci U S A. 1998 Jun 23;95(13):7591-6.

45. Yvan-Charvet L, Quignard-Boulange A. Role of adipose tissue renin-angiotensin system in metabolic and inflammatory diseases associated with obesity. Kidney Int. 2011 Jan;79(2):162-8.

46. McQuaid SE, Hodson L, Neville MJ, Dennis AL, Cheeseman J, Humphreys SM, et al.

Downregulation of adipose tissue fatty acid trafficking in obesity: A driver for ectopic fat deposition? Diabetes. 2011 Jan;60(1):47-55.

35

2

CHAPTER 2

47. Fielding B. Tracing the fate of dietary fatty acids: Metabolic studies of postprandial lipaemia in human subjects. Proc Nutr Soc. 2011 Aug;70(3):342-50.

48. Duncan RE, Ahmadian M, Jaworski K, Sarkadi-Nagy E, Sul HS. Regulation of lipolysis in adipocytes. Annu Rev Nutr. 2007;27:79-101.

49. Arner P. Differences in lipolysis between human subcutaneous and omental adipose tissues.

Ann Med. 1995 Aug;27(4):435-8.

50. Yang YK, Chen M, Clements RH, Abrams GA, Aprahamian CJ, Harmon CM. Human mesenteric adipose tissue plays unique role versus subcutaneous and omental fat in obesity related diabetes. Cell Physiol Biochem. 2008;22(5-6):531-8.

51. Tchkonia T, Giorgadze N, Pirtskhalava T, Tchoukalova Y, Karagiannides I, Forse RA, et al. Fat depot origin affects adipogenesis in primary cultured and cloned human preadipocytes. Am J Physiol Regul Integr Comp Physiol. 2002 May;282(5):R1286-96.

52. Antoniades C, Antonopoulos AS, Tousoulis D, Stefanadis C. Adiponectin: From obesity to cardiovascular disease. Obes Rev. 2009 May;10(3):269-79.

53. Antuna-Puente B, Feve B, Fellahi S, Bastard JP. Adipokines: The missing link between insulin resistance and obesity. Diabetes Metab. 2008 Feb;34(1):2-11.

54. Ouchi N, Kihara S, Funahashi T, Matsuzawa Y, Walsh K. Obesity, adiponectin and vascular inflammatory disease. Curr Opin Lipidol. 2003 Dec;14(6):561-6.

55. Tilg H, Moschen AR. Role of adiponectin and PBEF/visfatin as regulators of inflammation: Involvement in obesity-associated diseases. Clin Sci (Lond). 2008 Feb;114(4):275-88.

56. Schipper HS, Prakken B, Kalkhoven E, Boes M. Adipose tissue-resident immune cells: Key players in immunometabolism. Trends Endocrinol Metab. 2012 Aug;23(8):407-15.

57. Cho KW, Morris DL, Lumeng CN. Flow cytometry analyses of adipose tissue macrophages.

Methods Enzymol. 2014;537:297-314.

58. Feuerer M, Herrero L, Cipolletta D, Naaz A, Wong J, Nayer A, et al. Lean, but not obese, fat is enriched for a unique population of regulatory T cells that affect metabolic parameters. Nat Med. 2009 Aug;15(8):930-9.

59. Ilan Y, Maron R, Tukpah AM, Maioli TU, Murugaiyan G, Yang K, et al. Induction of regulatory T cells decreases adipose inflammation and alleviates insulin resistance in ob/ob mice. Proc Natl Acad Sci U S A. 2010 May 25;107(21):9765-70.

60. Djian P, Roncari AK, Hollenberg CH. Influence of anatomic site and age on the replication and differentiation of rat adipocyte precursors in culture. J Clin Invest. 1983 Oct;72(4):1200-8.

61. Baglioni S, Cantini G, Poli G, Francalanci M, Squecco R, Di Franco A, et al. Functional differences in visceral and subcutaneous fat pads originate from differences in the adipose stem cell. PLoS One. 2012;7(5):e36569.

62. Kirkland JL, Hollenberg CH, Gillon WS. Effects of fat depot site on differentiation-dependent gene expression in rat preadipocytes. Int J Obes Relat Metab Disord. 1996 Mar;20 Suppl 3:S102-7.

63. Tchkonia T, Tchoukalova YD, Giorgadze N, Pirtskhalava T, Karagiannides I, Forse RA, et al.

Abundance of two human preadipocyte subtypes with distinct capacities for replication, adipogenesis, and apoptosis varies among fat depots. Am J Physiol Endocrinol Metab. 2005 Jan;288(1):E267-77.

64. Gealekman O, Guseva N, Hartigan C, Apotheker S, Gorgoglione M, Gurav K, et al. Depotspecific differences and insufficient subcutaneous adipose tissue angiogenesis in human obesity. Circulation. 2011 Jan 18;123(2):186-94.

65. Skurk T, Alberti-Huber C, Herder C, Hauner H. Relationship between adipocyte size and adipokine expression and secretion. J Clin Endocrinol Metab. 2007 Mar;92(3):1023-33.

66. Hajer GR, van Haeften TW, Visseren FL. Adipose tissue dysfunction in obesity, diabetes, and vascular diseases. Eur Heart J. 2008 Dec;29(24):2959-71.

67. Mittelman SD, Van Citters GW, Kirkman EL, Bergman RN. Extreme insulin resistance of the central adipose depot in vivo. Diabetes. 2002 Mar;51(3):755-61.

68. Hotamisligil GS. Endoplasmic reticulum stress and the inflammatory basis of metabolic disease.

Cell. 2010 Mar 19;140(6):900-17.

36

Adipose tissue depots, EVs and metabolic dysfunction

69. Kawai T, Akira S. The role of pattern-recognition receptors in innate immunity: Update on tolllike receptors. Nat Immunol. 2010 May;11(5):373-84.

70. Joosten LA, Netea MG, Mylona E, Koenders MI, Malireddi RK, Oosting M, et al. Engagement of fatty acids with toll-like receptor 2 drives interleukin-1beta production via the ASC/caspase 1 pathway in monosodium urate monohydrate crystal-induced gouty arthritis. Arthritis Rheum.

2010 Nov;62(11):3237-48.

71. Holland WL, Bikman BT, Wang LP, Yuguang G, Sargent KM, Bulchand S, et al. Lipid-induced insulin resistance mediated by the proinflammatory receptor TLR4 requires saturated fatty acidinduced ceramide biosynthesis in mice. J Clin Invest. 2011 May;121(5):1858-70.

72. Shi H, Kokoeva MV, Inouye K, Tzameli I, Yin H, Flier JS. TLR4 links innate immunity and fatty acid-induced insulin resistance. J Clin Invest. 2006 Nov;116(11):3015-25.

73. Schenk S, Saberi M, Olefsky JM. Insulin sensitivity: Modulation by nutrients and inflammation.

J Clin Invest. 2008 Sep;118(9):2992-3002.

74. Samuel VT, Shulman GI. Mechanisms for insulin resistance: Common threads and missing links. Cell. 2012 Mar 2;148(5):852-71.

75. Koenen TB, Stienstra R, van Tits LJ, Joosten LA, van Velzen JF, Hijmans A, et al. The inflammasome and caspase-1 activation: A new mechanism underlying increased inflammatory activity in human visceral adipose tissue. Endocrinology. 2011 Oct;152(10):3769-78.

76. Reynolds CM, McGillicuddy FC, Harford KA, Finucane OM, Mills KH, Roche HM. Dietary saturated fatty acids prime the NLRP3 inflammasome via TLR4 in dendritic cells-implications for diet-induced insulin resistance. Mol Nutr Food Res. 2012 Aug;56(8):1212-22.

77. Vandanmagsar B, Youm YH, Ravussin A, Galgani JE, Stadler K, Mynatt RL, et al. The NLRP3 inflammasome instigates obesity-induced inflammation and insulin resistance. Nat Med. 2011 Feb;17(2):179-88.

78. Moreno-Navarrete JM, Ortega F, Serino M, Luche E, Waget A, Pardo G, et al. Circulating lipopolysaccharide-binding protein (LBP) as a marker of obesity-related insulin resistance. Int J Obes (Lond). 2012 Nov;36(11):1442-9.

79. Wen H, Gris D, Lei Y, Jha S, Zhang L, Huang MT, et al. Fatty acid-induced NLRP3-ASC inflammasome activation interferes with insulin signaling. Nat Immunol. 2011 May;12(5):40815.

80. Cnop M, Foufelle F, Velloso LA. Endoplasmic reticulum stress, obesity and diabetes. Trends Mol Med. 2012 Jan;18(1):59-68.

81. Ozcan U, Cao Q, Yilmaz E, Lee AH, Iwakoshi NN, Ozdelen E, et al. Endoplasmic reticulum stress links obesity, insulin action, and type 2 diabetes. Science. 2004 Oct 15;306(5695):45761.

82. Nakamura T, Furuhashi M, Li P, Cao H, Tuncman G, Sonenberg N, et al. Double-stranded RNAdependent protein kinase links pathogen sensing with stress and metabolic homeostasis. Cell.

2010 Feb 5;140(3):338-48.

83. Wang Y, Vera L, Fischer WH, Montminy M. The CREB coactivator CRTC2 links hepatic ER stress and fasting gluconeogenesis. Nature. 2009 Jul 23;460(7254):534-7.

84. Jiao P, Ma J, Feng B, Zhang H, Diehl JA, Chin YE, et al. FFA-induced adipocyte inflammation and insulin resistance: Involvement of ER stress and IKKbeta pathways. Obesity (Silver Spring).

2011 Mar;19(3):483-91.

85. Deldicque L, Van Proeyen K, Francaux M, Hespel P. The unfolded protein response in human skeletal muscle is not involved in the onset of glucose tolerance impairment induced by a fatrich diet. Eur J Appl Physiol. 2011 Jul;111(7):1553-8.

86. Chawla A, Nguyen KD, Goh YP. Macrophage-mediated inflammation in metabolic disease. Nat Rev Immunol. 2011 Oct 10;11(11):738-49.

87. Weisberg SP, McCann D, Desai M, Rosenbaum M, Leibel RL, Ferrante AW,Jr. Obesity is associated with macrophage accumulation in adipose tissue. J Clin Invest. 2003 Dec;112(12):1796-808.

88. Winer DA, Winer S, Shen L, Wadia PP, Yantha J, Paltser G, et al. B cells promote insulin resistance through modulation of T cells and production of pathogenic IgG antibodies. Nat Med. 2011 May;17(5):610-7.

37

2

CHAPTER 2

89. Xu H, Barnes GT, Yang Q, Tan G, Yang D, Chou CJ, et al. Chronic inflammation in fat plays a crucial role in the development of obesity-related insulin resistance. J Clin Invest. 2003 Dec;112(12):1821-30.

90. Nishimura S, Manabe I, Nagasaki M, Eto K, Yamashita H, Ohsugi M, et al. CD8+ effector T cells contribute to macrophage recruitment and adipose tissue inflammation in obesity. Nat Med. 2009 Aug;15(8):914-20.

91. Matzinger P, Kamala T. Tissue-based class control: The other side of tolerance. Nat Rev Immunol. 2011 Mar;11(3):221-30.

92. Eguchi K, Manabe I, Oishi-Tanaka Y, Ohsugi M, Kono N, Ogata F, et al. Saturated fatty acid and TLR signaling link beta cell dysfunction and islet inflammation. Cell Metab. 2012 Apr 4;15(4):518-33.

93. Lovren F, Pan Y, Quan A, Szmitko PE, Singh KK, Shukla PC, et al. Adiponectin primes human monocytes into alternative anti-inflammatory M2 macrophages. Am J Physiol Heart Circ Physiol. 2010 Sep;299(3):H656-63.

94. Meijer K, de Vries M, Al-Lahham S, Bruinenberg M, Weening D, Dijkstra M, et al. Human primary adipocytes exhibit immune cell function: Adipocytes prime inflammation independent of macrophages. PLoS One. 2011 Mar 23;6(3):e17154.

95. Foster MT, Softic S, Caldwell J, Kohli R, de Kloet AD, Seeley RJ. Subcutaneous adipose tissue transplantation in diet-induced obese mice attenuates metabolic dysregulation while removal exacerbates it. Physiol Rep. 2013 Aug;1(2):e00015.

96. Yim JE, Heshka S, Albu JB, Heymsfield S, Gallagher D. Femoral-gluteal subcutaneous and intermuscular adipose tissues have independent and opposing relationships with CVD risk. J Appl Physiol. 2008 Mar;104(3):700-7.

97. Koh-Banerjee P, Wang Y, Hu FB, Spiegelman D, Willett WC, Rimm EB. Changes in body weight and body fat distribution as risk factors for clinical diabetes in US men. Am J Epidemiol. 2004 Jun 15;159(12):1150-9.

98. Feldman BJ, Streeper RS, Farese RV,Jr, Yamamoto KR. Myostatin modulates adipogenesis to generate adipocytes with favorable metabolic effects. Proc Natl Acad Sci U S A. 2006 Oct 17;103(42):15675-80.

99. Pedersen BK, Febbraio MA. Muscles, exercise and obesity: Skeletal muscle as a secretory organ. Nat Rev Endocrinol. 2012 Apr 3;8(8):457-65.

100. Nielsen AR, Hojman P, Erikstrup C, Fischer CP, Plomgaard P, Mounier R, et al. Association between interleukin-15 and obesity: Interleukin-15 as a potential regulator of fat mass. J Clin Endocrinol Metab. 2008 Nov;93(11):4486-93.

101. Taube A, Lambernd S, van Echten-Deckert G, Eckardt K, Eckel J. Adipokines promote lipotoxicity in human skeletal muscle cells. Arch Physiol Biochem. 2012 Jul;118(3):92-101.

102. Nielsen S, Guo Z, Johnson CM, Hensrud DD, Jensen MD. Splanchnic lipolysis in human obesity.

J Clin Invest. 2004 Jun;113(11):1582-8.

103. Yoon MJ, Lee GY, Chung JJ, Ahn YH, Hong SH, Kim JB. Adiponectin increases fatty acid oxidation in skeletal muscle cells by sequential activation of AMP-activated protein kinase, p38 mitogen-activated protein kinase, and peroxisome proliferator-activated receptor alpha.

Diabetes. 2006 Sep;55(9):2562-70.

104. Fruebis J, Tsao TS, Javorschi S, Ebbets-Reed D, Erickson MR, Yen FT, et al. Proteolytic cleavage product of 30-kDa adipocyte complement-related protein increases fatty acid oxidation in muscle and causes weight loss in mice. Proc Natl Acad Sci U S A. 2001 Feb 13;98(4):2005-10.

105. Lee DE, Kehlenbrink S, Lee H, Hawkins M, Yudkin JS. Getting the message across: Mechanisms of physiological cross talk by adipose tissue. Am J Physiol Endocrinol Metab.

2009 Jun;296(6):E1210-29.

106. O'Connell J, Lynch L, Cawood TJ, Kwasnik A, Nolan N, Geoghegan J, et al. The relationship of omental and subcutaneous adipocyte size to metabolic disease in severe obesity. PLoS One. 2010 Apr 1;5(4):e9997.

107. Pou KM, Massaro JM, Hoffmann U, Vasan RS, Maurovich-Horvat P, Larson MG, et al. Visceral and subcutaneous adipose tissue volumes are cross-sectionally related to markers of inflammation and oxidative stress: The framingham heart study. Circulation. 2007 Sep 11;116(11):1234-41.

38

Adipose tissue depots, EVs and metabolic dysfunction

108. Maury E, Ehala-Aleksejev K, Guiot Y, Detry R, Vandenhooft A, Brichard SM. Adipokines oversecreted by omental adipose tissue in human obesity. Am J Physiol Endocrinol Metab.

2007 Sep;293(3):E656-65.

109. Fain JN. Release of inflammatory mediators by human adipose tissue is enhanced in obesity and primarily by the nonfat cells: A review. Mediators Inflamm. 2010;2010:513948.

110. Faber DR, Moll FL, Vink A, van der Waal C, Kalkhoven E, Schipper HS, et al. Adipose tissue quantity and composition contribute to adipokine concentrations in the subclavian vein and the inferior mesenteric vein. Int J Obes (Lond). 2012 Aug;36(8):1078-85.

111. Tordjman J, Poitou C, Hugol D, Bouillot JL, Basdevant A, Bedossa P, et al. Association between omental adipose tissue macrophages and liver histopathology in morbid obesity: Influence of glycemic status. J Hepatol. 2009 Aug;51(2):354-62.

112. Tordjman J, Divoux A, Prifti E, Poitou C, Pelloux V, Hugol D, et al. Structural and inflammatory heterogeneity in subcutaneous adipose tissue: Relation with liver histopathology in morbid obesity. J Hepatol. 2012 May;56(5):1152-8.

113. Rajala MW, Obici S, Scherer PE, Rossetti L. Adipose-derived resistin and gut-derived resistinlike molecule-beta selectively impair insulin action on glucose production. J Clin Invest. 2003 Jan;111(2):225-30.

114. Cancello R, Tordjman J, Poitou C, Guilhem G, Bouillot JL, Hugol D, et al. Increased infiltration of macrophages in omental adipose tissue is associated with marked hepatic lesions in morbid human obesity. Diabetes. 2006 Jun;55(6):1554-61.

115. Sell H, Divoux A, Poitou C, Basdevant A, Bouillot JL, Bedossa P, et al. Chemerin correlates with markers for fatty liver in morbidly obese patients and strongly decreases after weight loss induced by bariatric surgery. J Clin Endocrinol Metab. 2010 Jun;95(6):2892-6.

116. Faber DR, van der Graaf Y, Westerink J, Visseren FL. Increased visceral adipose tissue mass is associated with increased C-reactive protein in patients with manifest vascular diseases.

Atherosclerosis. 2010 Sep;212(1):274-80.

117. Bigornia SJ, Farb MG, Mott MM, Hess DT, Carmine B, Fiscale A, et al. Relation of depot-specific adipose inflammation to insulin resistance in human obesity. Nutr Diabetes. 2012 Mar 5;2:e30.

118. Desreumaux P, Ernst O, Geboes K, Gambiez L, Berrebi D, Muller-Alouf H, et al. Inflammatory alterations in mesenteric adipose tissue in crohn's disease. Gastroenterology. 1999 Jul;117(1):73-81.

119. Gambero A, Marostica M, Abdalla Saad MJ, Pedrazzoli J,Jr. Mesenteric adipose tissue alterations resulting from experimental reactivated colitis. Inflamm Bowel Dis. 2007 Nov;13(11):1357-64.

120. Feng Y, Li Y, Mei S, Zhang L, Gong J, Gu L, et al. Exclusive enteral nutrition ameliorates mesenteric adipose tissue alterations in patients with active crohn's disease. Clin Nutr. 2013 Oct 18.

121. Lam YY, Ha CW, Campbell CR, Mitchell AJ, Dinudom A, Oscarsson J, et al. Increased gut permeability and microbiota change associate with mesenteric fat inflammation and metabolic dysfunction in diet-induced obese mice. PLoS One. 2012;7(3):e34233.

122. Peyrin-Biroulet L, Chamaillard M, Gonzalez F, Beclin E, Decourcelle C, Antunes L, et al.

Mesenteric fat in crohn's disease: A pathogenetic hallmark or an innocent bystander? Gut. 2007 Apr;56(4):577-83.

123. Ludwig T, Worsch S, Heikenwalder M, Daniel H, Hauner H, Bader BL. Metabolic and immunomodulatory effects of n-3 fatty acids are different in mesenteric and epididymal adipose tissue of diet-induced obese mice. Am J Physiol Endocrinol Metab. 2013 Jun 1;304(11):E1140-56.

124. Anini Y, Brubaker PL. Role of leptin in the regulation of glucagon-like peptide-1 secretion.

Diabetes. 2003 Feb;52(2):252-9.

125. Anderlova K, Kremen J, Dolezalova R, Housova J, Haluzikova D, Kunesova M, et al. The influence of very-low-calorie-diet on serum leptin, soluble leptin receptor, adiponectin and resistin levels in obese women. Physiol Res. 2006;55(3):277-83.

126. Fain JN, Sacks HS, Buehrer B, Bahouth SW, Garrett E, Wolf RY, et al. Identification of omentin mRNA in human epicardial adipose tissue: Comparison to omentin in subcutaneous, internal mammary artery periadventitial and visceral abdominal depots. Int J Obes (Lond). 2008 May;32(5):810-5.

39

2

CHAPTER 2

127. Sacks HS. The importance of brown adipose tissue. N Engl J Med. 2009 Jul 23;361(4):418; author reply 418-20.

128. Sacks HS, Fain JN, Bahouth SW, Ojha S, Frontini A, Budge H, et al. Adult epicardial fat exhibits beige features. J Clin Endocrinol Metab. 2013 Sep;98(9):E1448-55.

129. Iacobellis G, di Gioia CR, Cotesta D, Petramala L, Travaglini C, De Santis V, et al. Epicardial adipose tissue adiponectin expression is related to intracoronary adiponectin levels. Horm Metab Res. 2009 Mar;41(3):227-31.

130. Stern N, Marcus Y. Perivascular fat: Innocent bystander or active player in vascular disease? J Cardiometab Syndr. 2006 Spring;1(2):115-20.

131. Sacks HS, Fain JN, Cheema P, Bahouth SW, Garrett E, Wolf RY, et al. Depot-specific overexpression of proinflammatory, redox, endothelial cell, and angiogenic genes in epicardial fat adjacent to severe stable coronary atherosclerosis. Metab Syndr Relat Disord. 2011 Dec;9(6):433-9.

132. Verhagen SN, Vink A, van der Graaf Y, Visseren FL. Coronary perivascular adipose tissue characteristics are related to atherosclerotic plaque size and composition. A post-mortem study. Atherosclerosis. 2012 Nov;225(1):99-104.

133. Barandier C, Montani JP, Yang Z. Mature adipocytes and perivascular adipose tissue stimulate vascular smooth muscle cell proliferation: Effects of aging and obesity. Am J Physiol Heart Circ Physiol. 2005 Nov;289(5):H1807-13.

134. Thery C, Zitvogel L, Amigorena S. Exosomes: Composition, biogenesis and function. Nat Rev Immunol. 2002 Aug;2(8):569-79.

135. Pan BT, Johnstone RM. Fate of the transferrin receptor during maturation of sheep reticulocytes in vitro: Selective externalization of the receptor. Cell. 1983 Jul;33(3):967-78.

136. Raposo G, Nijman HW, Stoorvogel W, Liejendekker R, Harding CV, Melief CJ, et al. B lymphocytes secrete antigen-presenting vesicles. J Exp Med. 1996 Mar 1;183(3):1161-72.

137. Trajkovic K, Hsu C, Chiantia S, Rajendran L, Wenzel D, Wieland F, et al. Ceramide triggers budding of exosome vesicles into multivesicular endosomes. Science. 2008 Feb 29;319(5867):1244-7.

138. Tamai K, Tanaka N, Nakano T, Kakazu E, Kondo Y, Inoue J, et al. Exosome secretion of dendritic cells is regulated by hrs, an ESCRT-0 protein. Biochem Biophys Res Commun. 2010 Aug 27;399(3):384-90.

139. De Gassart A, Trentin B, Martin M, Hocquellet A, Bette-Bobillo P, Mamoun R, et al. Exosomal sorting of the cytoplasmic domain of bovine leukemia virus TM env protein. Cell Biol Int. 2009 Jan;33(1):36-48.

140. Phuyal S, Hessvik NP, Skotland T, Sandvig K, Llorente A. Regulation of exosome release by glycosphingolipids and flotillins. FEBS J. 2014 Mar 8.

141. Morelli AE, Larregina AT, Shufesky WJ, Sullivan ML, Stolz DB, Papworth GD, et al. Endocytosis, intracellular sorting, and processing of exosomes by dendritic cells. Blood. 2004 Nov 15;104(10):3257-66.

142. Bissig C, Gruenberg J. Lipid sorting and multivesicular endosome biogenesis. Cold Spring Harb Perspect Biol. 2013 Oct 1;5(10):a016816.

143. Buschow SI, Nolte-'t Hoen EN, van Niel G, Pols MS, ten Broeke T, Lauwen M, et al. MHC II in dendritic cells is targeted to lysosomes or T cell-induced exosomes via distinct multivesicular body pathways. Traffic. 2009 Oct;10(10):1528-42.

144. ten Broeke T, de Graaff A, van't Veld EM, Wauben MH, Stoorvogel W, Wubbolts R. Trafficking of MHC class II in dendritic cells is dependent on but not regulated by degradation of its associated invariant chain. Traffic. 2010 Mar;11(3):324-31.

145. Raposo G, Stoorvogel W. Extracellular vesicles: Exosomes, microvesicles, and friends. J Cell Biol. 2013 Feb 18;200(4):373-83.

146. Bobrie A, Colombo M, Raposo G, Thery C. Exosome secretion: Molecular mechanisms and roles in immune responses. Traffic. 2011 Jun 6.

147. Burnier L, Fontana P, Kwak BR, Angelillo-Scherrer A. Cell-derived microparticles in haemostasis and vascular medicine. Thromb Haemost. 2009 Mar;101(3):439-51.

148. Thery C, Ostrowski M, Segura E. Membrane vesicles as conveyors of immune responses. Nat Rev Immunol. 2009 Aug;9(8):581-93.

40

Adipose tissue depots, EVs and metabolic dysfunction

149. Brown D, Waneck GL. Glycosyl-phosphatidylinositol-anchored membrane proteins. J Am Soc Nephrol. 1992 Oct;3(4):895-906.

150. Valadi H, Ekstrom K, Bossios A, Sjostrand M, Lee JJ, Lotvall JO. Exosome-mediated transfer of mRNAs and microRNAs is a novel mechanism of genetic exchange between cells. Nat Cell Biol. 2007 Jun;9(6):654-9.

151. Bhatnagar S, Shinagawa K, Castellino FJ, Schorey JS. Exosomes released from macrophages infected with intracellular pathogens stimulate a proinflammatory response in vitro and in vivo.

Blood. 2007 Nov 1;110(9):3234-44.

152. Zitvogel L, Regnault A, Lozier A, Wolfers J, Flament C, Tenza D, et al. Eradication of established murine tumors using a novel cell-free vaccine: Dendritic cell-derived exosomes. Nat Med. 1998 May;4(5):594-600.

153. Mack M, Kleinschmidt A, Bruhl H, Klier C, Nelson PJ, Cihak J, et al. Transfer of the chemokine receptor CCR5 between cells by membrane-derived microparticles: A mechanism for cellular human immunodeficiency virus 1 infection. Nat Med. 2000 Jul;6(7):769-75.

154. Muller G. Microvesicles/exosomes as potential novel biomarkers of metabolic diseases.

Diabetes Metab Syndr Obes. 2012;5:247-82.

155. Goichot B, Grunebaum L, Desprez D, Vinzio S, Meyer L, Schlienger JL, et al. Circulating procoagulant microparticles in obesity. Diabetes Metab. 2006 Feb;32(1):82-5.

156. Diamant M, Nieuwland R, Pablo RF, Sturk A, Smit JW, Radder JK. Elevated numbers of tissuefactor exposing microparticles correlate with components of the metabolic syndrome in uncomplicated type 2 diabetes mellitus. Circulation. 2002 Nov 5;106(19):2442-7.

157. Kranendonk ME, de Kleijn DP, Kalkhoven E, Kanhai DA, Uiterwaal CS, van der Graaf Y, et al.

Extracellular vesicle markers in relation to obesity and metabolic complications in patients with manifest cardiovascular disease. Cardiovasc Diabetol. 2014 Feb 5;13(1):37,2840-13-37.

158. Muller G, Jung C, Straub J, Wied S, Kramer W. Induced release of membrane vesicles from rat adipocytes containing glycosylphosphatidylinositol-anchored microdomain and lipid droplet signalling proteins. Cell Signal. 2009 Feb;21(2):324-38.

159. Muller G, Schneider M, Biemer-Daub G, Wied S. Microvesicles released from rat adipocytes and harboring glycosylphosphatidylinositol-anchored proteins transfer RNA stimulating lipid synthesis. Cell Signal. 2011 Jul;23(7):1207-23.

160. Muller G, Schneider M, Biemer-Daub G, Wied S. Upregulation of lipid synthesis in small rat adipocytes by microvesicle-associated CD73 from large adipocytes. Obesity (Silver Spring).

2011 Aug;19(8):1531-44.

161. Suganami T, Nishida J, Ogawa Y. A paracrine loop between adipocytes and macrophages aggravates inflammatory changes: Role of free fatty acids and tumor necrosis factor alpha.

Arterioscler Thromb Vasc Biol. 2005 Oct;25(10):2062-8.

162. Flieger O, Engling A, Bucala R, Lue H, Nickel W, Bernhagen J. Regulated secretion of macrophage migration inhibitory factor is mediated by a non-classical pathway involving an ABC transporter. FEBS Lett. 2003 Sep 11;551(1-3):78-86.

163. Kleemann R, Hausser A, Geiger G, Mischke R, Burger-Kentischer A, Flieger O, et al. Intracellular action of the cytokine MIF to modulate AP-1 activity and the cell cycle through Jab1. Nature.

2000 Nov 9;408(6809):211-6.

164. Verschuren L, Kooistra T, Bernhagen J, Voshol PJ, Ouwens DM, van Erk M, et al. MIF deficiency reduces chronic inflammation in white adipose tissue and impairs the development of insulin resistance, glucose intolerance, and associated atherosclerotic disease. Circ Res. 2009 Jul 2;105(1):99-107.

165. Koska J, Stefan N, Dubois S, Trinidad C, Considine RV, Funahashi T, et al. mRNA concentrations of MIF in subcutaneous abdominal adipose cells are associated with adipocyte size and insulin action. Int J Obes (Lond). 2009 Aug;33(8):842-50.

166. Kleemann R, Bucala R. Macrophage migration inhibitory factor: Critical role in obesity, insulin resistance, and associated comorbidities. Mediators Inflamm. 2010;2010:610479.

167. Berndt K, Kim M, Meinhardt A, Klug J. Macrophage migration inhibitory factor does not modulate co-activation of androgen receptor by Jab1/CSN5. Mol Cell Biochem. 2008 Jan;307(12):265-71.

41

2

CHAPTER 2

168. Lue H, Kleemann R, Calandra T, Roger T, Bernhagen J. Macrophage migration inhibitory factor (MIF): Mechanisms of action and role in disease. Microbes Infect. 2002 Apr;4(4):449-60.

169. Lin HV, Kim JY, Pocai A, Rossetti L, Shapiro L, Scherer PE, et al. Adiponectin resistance exacerbates insulin resistance in insulin receptor transgenic/knockout mice. Diabetes. 2007 Aug;56(8):1969-76.

170. Degawa-Yamauchi M, Moss KA, Bovenkerk JE, Shankar SS, Morrison CL, Lelliott CJ, et al.

Regulation of adiponectin expression in human adipocytes: Effects of adiposity, glucocorticoids, and tumor necrosis factor alpha. Obes Res. 2005 Apr;13(4):662-9.

171. Hotamisligil GS, Budavari A, Murray D, Spiegelman BM. Reduced tyrosine kinase activity of the insulin receptor in obesity-diabetes. central role of tumor necrosis factor-alpha. J Clin Invest.

1994 Oct;94(4):1543-9.

172. Senn JJ, Klover PJ, Nowak IA, Zimmers TA, Koniaris LG, Furlanetto RW, et al. Suppressor of cytokine signaling-3 (SOCS-3), a potential mediator of interleukin-6-dependent insulin resistance in hepatocytes. J Biol Chem. 2003 Apr 18;278(16):13740-6.

173. Odegaard JI, Ricardo-Gonzalez RR, Goforth MH, Morel CR, Subramanian V, Mukundan L, et al. Macrophage-specific PPARgamma controls alternative activation and improves insulin resistance. Nature. 2007 Jun 28;447(7148):1116-20.

174. Bernhagen J, Mitchell RA, Calandra T, Voelter W, Cerami A, Bucala R. Purification, bioactivity, and secondary structure analysis of mouse and human macrophage migration inhibitory factor (MIF). Biochemistry. 1994 Nov 29;33(47):14144-55.

175. Weisberg SP, Hunter D, Huber R, Lemieux J, Slaymaker S, Vaddi K, et al. CCR2 modulates inflammatory and metabolic effects of high-fat feeding. J Clin Invest. 2006 Jan;116(1):115-24.

176. Bokarewa M, Nagaev I, Dahlberg L, Smith U, Tarkowski A. Resistin, an adipokine with potent proinflammatory properties. J Immunol. 2005 May 1;174(9):5789-95.

177. Yang Q, Graham TE, Mody N, Preitner F, Peroni OD, Zabolotny JM, et al. Serum retinol binding protein 4 contributes to insulin resistance in obesity and type 2 diabetes. Nature. 2005 Jul 21;436(7049):356-62.

178. Ost A, Danielsson A, Liden M, Eriksson U, Nystrom FH, Stralfors P. Retinol-binding protein-4 attenuates insulin-induced phosphorylation of IRS1 and ERK1/2 in primary human adipocytes.

FASEB J. 2007 Nov;21(13):3696-704.

179. Norseen J, Hosooka T, Hammarstedt A, Yore MM, Kant S, Aryal P, et al. Retinol-binding protein 4 inhibits insulin signaling in adipocytes by inducing proinflammatory cytokines in macrophages through a c-jun N-terminal kinase- and toll-like receptor 4-dependent and retinol-independent mechanism. Mol Cell Biol. 2012 May;32(10):2010-9.

42

43

CHAPTER 3 Inflammatory characteristics of distinct abdominal adipose tissue depots relate differently to metabolic risk factors for cardiovascular disease Submitted

MariĂŤtte E.G. Kranendonk, Joost A. van Herwaarden, Tereza Stupkova, Wilco de Jager, Aryan Vink, Frans L. Moll, Eric Kalkhoven, Frank L.J.Visseren

CHAPTER 3

ABSTRACT Aims/hypothesis Abdominal obesity is associated with insulin resistance and dyslipidemia. However, specific contributions of distinct adipose tissue (AT) depots to metabolic diseases is still unclear.

In this study, the inflammatory profile of four distinct abdominal AT-depots and the relation between AT-characteristics and obesity-induced metabolic diseases was evaluated in patients with or without abdominal obesity.

Methods In 28 men undergoing abdominal aortic surgery, biopsies were collected from subcutaneous fat (SAT), and 3 visceral AT-depots: mesenteric (MAT), omental (OAT) and periaortic (PAT).

The AT biopsies were characterized morphologically (adipocyte size, capillary density, CD68+macrophages and crown-like-structures (CLS)) and the ex vivo adipokine secretion profile was determined by multiplex-immunoassay. The relation between depot-specific inflammatory characteristics and clinical parameters (insulin resistance, dyslipidemia and systemic low-grade inflammation) was assessed by multivariable linear regression analysis.

Results All AT-depots of abdominally-obese patients showed larger adipocytes, lower capillary density and a different adipokine secretion profile compared to abdominal-lean patients.

SAT and PAT characteristics did not relate to systemic metabolic dysfunction. MAT adipocyte size (β0.22;95%CI0.13–0.32) and CLS (β0.07;95%CI0.01–0.12) of abdominally-obese patients were related to insulin resistance, in contrast to abdominal-lean patients, while CLS of abdominal-obese OAT related to lower HDL-cholesterol (β-0.05; 95%CI -0.10 – -0.01) and higher triglyceride levels (β0.16; 95%CI 0.01–0.31).

Conclusions/interpretation The AT-inflammatory profile is morphologically and functionally different between abdominallean and obese men, and AT-characteristics of distinct visceral depots relate differently to glucose or lipid metabolism. This suggests a differential contribution of AT-depots to systemic metabolic dysfunction which precedes type 2 diabetes and vascular diseases.

46

Distinct fat depots and vascular risk factors

INTRODUCTION The worldwide prevalence of obesity has doubled over the last 20 years, and with that the risk for type 2 diabetes mellitus, metabolic syndrome and cardiovascular diseases has increased tremendously (1-3). Obesity is the result of chronic energy excess, which challenges adipose tissue (AT) function to regulate metabolic homeostasis (4). Obesity results in AT derangements, characterized by adipocyte hypertrophy, altered AT-resident immune cell composition in AT and consequently an altered adipokine and cytokine secretion profile, often referred to as AT dysfunction (5.6). In addition, AT of obese subjects has less capacity to expand the capillary network surrounding adipocytes, resulting in adipocyte hypoxia and necrosis (7). This process is marked by presence of crown-likestructures (CLS); e.g. macrophages surrounding necrotic adipocytes (5). Presence of these inflammatory characteristics reflecting AT dysfunction is considered a key mechanism leading to systemic metabolic changes such as low-grade inflammation, insulin resistance and the dyslipidemia (8 9).

Intra-abdominal, several AT-depots can be distinguished based upon anatomic location.

Subcutaneous AT (SAT) is located underneath the skin and visceral AT (VAT) surrounds organs, of which mesenteric AT (MAT) lines the surface of the intestine, omental AT (OAT) relates to the greater omentum, and periaortic AT (PAT) surrounds the abdominal aorta (10/11). Epidemiological studies indicate that mainly abundance of VAT (MAT and OAT) is related to development of type 2 diabetes mellitus and cardiovascular disease (12.13). SAT is considered to be the least active AT-depot, with lower macrophage infiltration and adipokine secretion compared to visceral AT-depots (14,15). Discordant, CLS are repeatedly shown to be present in SAT and associated with metabolic complications as well (8,16).

PAT is located just outside the adventitial layer, and capable of secreting adipokines, which can diffuse directly into the vascular wall and thereby contribute to atherogenesis (17,18).

VAT is considered a highly active secretory organ, and the direct release of VAT-derived adipokines into the portal vein might directly affect hepatic glucose and lipid metabolism (19,20). However, less is known about the differential contribution of distinct VAT-depots to obesity-induced metabolic derangements as MAT and OAT are often conjointly termed VAT, and considered one and the same depot. However, intrinsic differences between MAT and OAT indicates that these distinct abdominal VAT-depots might contribute differently to metabolic changes in glucose or lipid regulation (21,22).

In order to unravel the mechanisms whereby AT contributes to risk factors for cardiovascular disease, accurate knowledge of depot-specific contribution is crucial. In this study, the inflammatory profile of abdominally-lean versus abdominally-obese male patients undergoing abdominal aortic surgery was investigated. The primary objective was to directly compare the morphology and the ex vivo adipokine secretion profile of four abdominal AT-depots (SAT, MAT, OAT, PAT) from patients with or without abdominal obesity. The secondary objective was to relate AT-inflammatory characteristics of each depot to metabolic changes known to be a consequence of AT dysfunction such as systemic low-grade inflammation, insulin resistance and dyslipidemia, which often

47

3

CHAPTER 3

precedes the risk for type 2 diabetes, low grade systemic inflammation and recurrent vascular events in the general population as well as patients with established manifest vascular disease (13,23,24).

METHODS Subjects Male patients undergoing elective abdominal aortic surgery for aneurysmatic or stenotic aortic disease at the University Medical Centre Utrecht (UMCU) were eligible for participation in the study in the period from October 2010 to March 2013. Exclusion criteria were hypothyroidism (thyroid-stimulating hormone >5.0mUl-1), elevated liver enzymes (aspartate aminotransferase (ASAT) or alanine aminotransferase (ALAT) >2 times the upper limit of normal), renal failure (MDRD <30 ml/min/1.73m2), known malignancy in the past 2 years, use of thiazolidinediones and a history of liposuction. All patients gave written informed consent and the study was approved by the Medical Ethics Committee of the UMCU. Anthropometric and metabolic measures During a single visit before planned abdominal aortic surgery, height and weight were collected, and the body mass index -the weight in kilograms divided by the square if the height in meters- was computed. Waist circumference was measured as the circumference in centimeters halfway between the lower rib and the iliac crest. A patient with waist circumference above 102 cm was considered abdominally-obese. Further, blood was drawn in a fasting state for biochemical analysis including plasma glucose, insulin, glycated haemoglobin A1c, total cholesterol, triglycerides, HDL-cholesterol, hsCRP, creatinin, ASAT, ALAT and gamma glutamyltransferase. LDL-cholesterol was calculated with the Friedewaldâ&#x20AC;&#x2122;s formula. The value for insulin resistance was assessed by the formula: homeostasis model assessment parameter of insulin resistance (HOMA-IR: fasting glucose (mmol/mL) X insulin (IU/mL))/22.5) and was only performed in patients without antihyperglycaemic drugs (25).

Low-grade inflammation was measured by hsCRP levels, determined by immunonephelometry with a lower detection limit of the test of 0.2 mg/L (Nephelometer Analyser BN II, Dade-Behring, Marburg, Germany). The Adult Treatment Panel (ATP) III criteria were taken for the definition of the metabolic syndrome (3) Adipose tissue collection and handling Approximately 5 gram of SAT, MAT or OAT and approximately 1-2 gram of PAT was obtained during surgery. The AT biopsies were cut into different pieces. One part was incubated in DMEMf12 (supplemented with 50 IU/ml penicillin and 50 mg/ml streptomycin) for 24 hours and weighed afterwards. Culture supernatant was centrifuged for 5 minutes at 500g to remove cells and stored at -80°C until further processing. Another part of AT was fixed in 10% formaldehyde until further processing. Two patients did not have enough PAT for

48

Distinct fat depots and vascular risk factors

sampling. Of one other patient the AT biopsies were not cultured. Altogether, AT supernatant of all 4 depots of 25 patients was eligible for ex vivo adipokine measurement. Histological examination was performed on AT of all patients.

(Immuno)histochemistry Fixed AT samples were embedded in paraffin and 4 Âľm sections were processed for histological staining with hematoxylin & eosin (H&E). To detect macrophages a CD68 immunostain was used and capillary density was analysed using a von Willebrand factor immunostain. Digital images of the H&E stained slides were acquired with Leica Qprodit imaging software (Leica Microsystems, Rijswijk, the Netherlands), after which the adipocyte size of 100 adipocytes was measured in randomly selected areas of the slide at 200x magnification (Figure 1a). Capillary density (%) was determined on digital slide images, acquired by Aperio Scan ScopeTM. Due to technical issues, a digital image of some slides could not be acquired, and those patients were left out of analyses (n=2). Capillary density was calculated by measuring the area of von Willebrand factor-positive staining in five fields per slide at 200x magnification using ImageJ Software (v1.45) as described previously (Figure 1b) (20). Macrophage infiltration was scored by determining the mean number of CD68-positive cells in 10 high power fields (HPF; magnification 400x) per slide as described previously (Figure 1c) (20). CLS were defined as CD68-positive cells surrounding at least 50% of the circumference of an adipocyte (Figure 1d). The number of CLS per AT-depot was determined by counting all CLS at 100x magnification on the whole slide, which was corrected for total surface area using ImageJ Software (v1.45). To validate our histological methods, 10% of the slides were analyzed by two researchers and the intra class correlation (ICC) was calculated. For macrophages, CLS and adipocyte size, the ICC was 0.92 (p<0.001), 0.94 (p<0.001) and 0.85 (p=0.002) respectively.

Multiplex immunoassay 30 adipokines were measured in supernatants from incubated fat biopsies by multiplex immunoassay. interleukin-6 (IL-6) was measured 100x diluted, all other cytokines were measured undiluted. Data was acquired with Bio-Plex Flexmap3D system, using Xponent 4.2 software (Luminex Austin TX, USA). Data analysis was perfomed with Bio-PLex Manager Software version 6.1.1 (Biorad Laboratories, Hercules, CA). Antibody pairs and recombinant proteins used were described earlier (26).

Adipokine secretion profile Concentrations of measured adipokines were corrected for sample weight. Using OmniViz software version 6.2 (OmniViz, Instem, Staffordshire UK) hierarchical clustering of differentially secreted adipokines was performed. For cytokine clustering, data was normalized setting lowest measurement per cytokine to zero and highest level measured set to 1. By combining cytokine and AT clustering, a heat map was created with colors corresponding to the relative expression of the median adipokine levels per AT-depot, where yellow indicates high, and blue indicates low relative expression per gram AT.

49

3

CHAPTER 3

Figure 1. Morphology of four abdominal adipose tissue depots stratified by visceral obesity. (A) Adipocyte size measurement in a HE-stained section (x200 magnification). Measurements were performed in 100 adipocytes in randomly selected fields of the microscopic slide. (B) Anti-von-Willibrand factor staining (x200 magnification) showing endothelial cells. (C) Anti-CD68 staining (x400 magnification) shows macrophages. (D) CD68-positive macrophages organized into crown-like-structures (200x) (E-H). quantitation of AT-characteristics in four abdominal AT-depots of abdominal-lean (white bars) and abdominal-obese men (grey bars). Number of AT-samples of abdominal-lean versus abdominal-obese was 10 versus 16 (E,G,H) or 9 versus 14 per AT-depot (F). HPF: high power field 400x.

50

Distinct fat depots and vascular risk factors

Data analysis Histograms and Q-Q plots were used to evaluate if continuous variables or residuals were normally distributed. Continuous variables are expressed as mean ± standard deviation (SD) when normal distributed or as median (interquartile range) in case of skewed distribution.

Categorical variables are expressed as numbers (percentage). Mean differences in adipocyte size, % capillary density, number of macrophages and number of CLS of SAT, MAT, OAT and PAT in different metabolic groups were calculated by analysis of covariance (ANCOVA), corrected for age, use of insulin therapy and current smoking. Analyses on differences in CLS values were performed on square root transformations to fulfil ANCOVA criteria, but for ease of interpretation values were back transformed to original mean number of CLS per depot. To assure that an equal comparison was performed, only measurements obtained from all four AT-depots within one patient were included in the analyses.

Multivariable linear regression analyses were used to evaluate relations between AT morphology parameters (adipocyte size, capillary density, number of macrophages or number of CLS) and metabolic dysfunction (insulin resistance, low-grade inflammation and parameters of lipid metabolism). Analyses were adjusted for age and current smoking.

Results are expressed as β-regression coefficients and 95% confidence intervals (95%CI) denoting the change in AT-inflammation parameter per unit change in HOMA-IR, lipid parameters or hsCRP. HsCRP and HOMA-IR values were log transformed to fulfill linear regression criteria. Subjects with antihyperglycaemic drugs were excluded from the analysis with HOMA-IR (n=4). We did not adjust for multiple comparisons in these analyses as they were prespecified analyses.

All data analyses were conducted using SPSS version 20 (SPSS Inc. Chicago, IL).

RESULTS Patient characteristics The characteristics of the participants, stratified by abdominal obesity (waist > 102 cm) is shown in Table 1. The mean waist circumference of abdominally-obese patients was 110 ± 5 cm versus 95 ± 6 cm of abdominally-lean patients. Abdominally-obese patients more often had metabolic syndrome, type 2 diabetes, a higher blood pressure and were more insulin resistant (HOMA-IR 5.1 (2.4 – 9.1) versus 4.0 (1.8 – 5.1) in the abdominally-lean patients. Surprisingly, abdominally-lean patients showed higher levels of systemic low grade inflammation (hsCRP 4.0 (1.8 – 6.1)) compared to abdominally-obese patients (hsCRP 2.8 (1.4 – 5.5)).

Morphology of four distinct adipose tissue depots The morphological profile of the four abdominal AT-depots was determined by adipocyte size, capillary density, macrophage infiltration and presence of CLS (Figure 1 a-d). Adipocytes of abdominally-obese patients were larger compared to adipocytes of abdominally-lean patients (Figure 1e).

51

3

CHAPTER 3

Table 1. Baseline characteristics.

Abdominal-lean

Abdominal-obese

n = 12

n = 16

Age (years)

65 ± 14

65 ± 6

Indication surgery: Aneurysm of abdominal or iliac aorta, n (%)

11 (92)

16 (100)

Indication surgery: Stenosis of abdominal aorta, n (%)

1 (3)

1 (3)

Body mass index (kg/m²)

24.2 ± 2.7

28.2 ± 2.3

Waist circumference (cm)

95 ± 6

110 ± 5

Systolic blood pressure (mmHg)

137 ± 16

149 ± 20

Diastolic blood pressure (mmHg)

81 ± 14

89 ± 9

Past smoking, n (%)

7 (58%)

12 (75%)

Current smoking, n (%)

3 (25)

5 (18)

Type 2 diabetes, n (%)

2 (17)

5 (31)

Metabolic syndrome, n (%)a

4 (33)

15 (94)

Overweight ((body mass index > 25 kg/m )

3 (25)

15 (94)

Obesity (body mass index > 30 kg/m2)

0 (12)

4 (25)

Cholesterol (mmol/L)

4.3 ± 0.9

3.9 ± 0.9

LDL-cholesterol (mmol/L)

2.5 ± 0.8

1.9 ± 0.7

HDL-cholesterol (mmol/L)

1.1 (1.0 – 1.2)

1.0 (0.8 – 1.3)

Triglycerides (mmol/L)

1.3 (0.7 – 1.8)

1.5 (1.1 – 2.4)

Fasting glucose (mmol/L)

6.3 ± 0.8

6.6 ± 1.0

Insulin (IU/I)

13.0 (7.0 – 16.8)

17.5 (9.78 – 28.5)

HOMA-IR

4.0 (1.8 – 5.1)

5.1 (2.4 – 9.1)

HbA1c (mmol/mol)

40 ± 4.9

42 ± 5

HsCRP (mg/L)

4.0 (1.8 – 6.1)

2.8 (1.4 – 5.5)

Creatinine (mmol/l)

96.8 ± 20

117 ± 39

Aspartate aminotransferase (IU/L)

20.1 ± 9.25

21.2 ± 11.5

Alanine aminotransferase (IU/L)

21.3 ± 14.3

24.8 ± 11.2

Gamma glutamyltransferase (IU/l)

34.5 (2.01 – 77.5)

33.0 (19.5 – 54.3)

Cerebrovascular disease

0 (0)

3 (19)

Coronary artery disease

6 (50)

4 (25)

Peripheral artery disease

4 (33)

9 (56)

Renal failure (GFR < 60 ml/min)

1 (8)

4 (25)

2

Metabolic parameters

History of vascular disease, n (%)

52

Distinct fat depots and vascular risk factors

Table 1 continued Medication use, n (%) Oral anticoagulants

1 (8)

32 (13)

Platelet-aggregation inhibitors

9 (75)

13 (81)

Blood pressure-lowering agents

12 (100)

16 (100)

Lipid-lowering agents

10 (83)

16 (100)

Metformin

2 (17)

3 (19)

Insulin therapy

0 (0)

1 (6)

Values are expressed in n (%), mean Âą SD or median (interquartile range). a Defined according to the National Cholesterol Education Program ATPIII-revised guidelines b eGFR, estimated glomerular filtration rate, estimated by the modification of diet in renal disease equation (MDRD)

After adjustment for age, insulin treatment and current smoking this difference remained statistically significant for SAT, MAT and OAT adipocytes (Table 2). Capillary density was higher in the visceral AT-depots of abdominal-lean males (Figure 1f). The most pronounced difference in capillary density among abdominally-obese and lean patients was observed in OAT, but after adjustment for confounders only the difference in capillary density of MAT remained statistically significant (Table 2). There was no significant difference in the number of macrophages between abdominally-lean versus abdominally-obese patients (Figure 1g and Table 2). CLS were observed in all four AT-depots in both abdominally-lean and abdominally-obese patients, and striking inter-individual differences were observed (Figure 1h). The mean number of CLS in SAT and PAT was significantly higher in abdominally-obese patients, while the mean number of CLS in OAT was significantly higher in abdominally-lean patients.

Next, the difference in AT-characteristics between men with or without the metabolic syndrome or type 2 diabetes was determined (Table 2). In metabolic syndrome patients, MAT showed statistically significant larger adipocytes and lower capillary density, and higher mean number of CLS was observed in SAT and PAT, compared to patients without metabolic syndrome. Patients with type 2 diabetes showed a similar profile, albeit far less extent compared to abdominally-obese and metabolic syndrome patients. Overall, there were no differences in adipocyte size or number of macrophages, and the capillary density was lower in MAT of diabetic patients. In contrast, the number of CLS in SAT and PAT was higher in non-diabetic patients (Table 2).

To determine the adipokine secretion profile of distinct abdominal AT depots (26), secreted adipokines were measured in the supernatant of ex vivo incubated fat biopsies. Median differences were depicted as extremes in a heat map, enabling us to observe possible differences between abdominally-obese versus abdominally-lean patients in four AT-depots (Figure 2). Marked differences were observed in the adipokine secretion profile of abdominally-lean versus abdominally-obese patients.

53

3

CHAPTER 3

Table 2. Difference in morphologic AT characteristics in metabolically (un)compromised groups.

Adipocyte size (mm) Abdominal obesity SAT No (n= 12) Yes (n= 16)

MAT

p value OAT

PAT Metabolic syndrome SAT No (n= 9) Yes (n=19)

MAT

p value OAT

PAT Type 2 diabetes SAT No (n= 21) Yes (n= 7)

MAT

p value OAT

PAT

cell Capillary density Macrophages (%) (n per HPF)

CLS† (n per cm2)

0.78 ± 0.20

1.21 ± 0.36

4.11 ± 1.40

4.40 ± 7.25

1.06 ± 0.22

1.08 ± 0.39

3.64 ± 1.10

5.44 ± 6.78

0.012

0.716

0.483

0.005

0.91 ± 0.26

1.47 ± 0.80

5.34 ± 1.59

2.13 ± 3.20

1.20 ± 0.25

1.14 ± 0.47

5.09 ± 1.16

4.81 ± 6.02

0.049

0.003

0.662

0.169

0.83 ± 0.19

1.68 ± 0.93

6.15 ± 1.78

4.45 ± 5.99

1.17 ± 0.27

1.03 ± 0.40

5.04 ± 1.06

2.71 ± 3.99

0.016

0.063

0.232

0.015

0.70 ± 0.16

1.90 ± 0.76

5.47 ± 2.71

1.14 ± 1.50

0.98 ± 0.27

1.62 ± 0.92

4.68 ± 1.40

4.34 ± 6.83

0.171

0.643

0.101

0.001

0.79 ± 0.24

1.20 ± 0.35

3.63 ± 1.29

3.56 ± 7.59

1.01 ± 0.23

1.10 ± 0.39

3.94 ± 1.24

5.68 ± 6.61

0.069

0.754

0.607

0.003

0.86 ± 0.26

1.36 ± 0.84

4.47 ± 1.37

0.99 ± 1.07

1.18 ± 0.24

1.26 ± 0.52

5.55 ± 1.20

4.93 ± 5.77

0.035

0.009

0.303

0.106

0.84 ± 0.19

1.39 ± 0.58

5.60 ± 1.78

2.54 ± 5.99

1.11 ± 0.29

1.26 ± 0.81

5.47 ± 1.39

3.89 ± 3.99

0.147

0.454

0.808

0.057

0.77 ± 0.21

1.58 ± 0.23

4.21 ± 2.15

0.49 ± 0.89

0.91 ± 0.28

1.79 ± 0.97

5.27 ± 1.91

4.07 ± 6.29

0.812

0.789

0.225

0.002

0.93 ± 0.31

1.24 ± 0.39

3.93 ± 1.16

6.29 ± 8.52

0.94 ± 0.24

1.09 ± 0.37

3.81 ± 1.29

4.57 ± 6.42

0.198

0.455

0.640

0.003

1.13 ± 0.26

1.41 ± 0.47

5.11 ± 1.06

4.05 ± 5.24

1.06 ± 0.30

1.25 ± 0.69

5.23 ± 1.44

3.53 ± 5.19

0.582

0.014

0.738

0.262

1.13 ± 0.21

1.15 ± 0.40

5.03 ± 0.95

3.63 ± 5.13

0.99 ± 0.31

1.36 ± 0.82

5.67 ± 1.62

3.40 ± 4.99

0.451

0.499

0.624

0.066

0.88 ± 0.20

1.36 ± 0.64

4.99 ± 2.02

4.74 ± 9.81

0.87 ± 0.30

1.88 ± 0.90

4.98 ± 2.04

2.50 ± 3.16

0.989

0.558

0.257

0.002

Abdominal obesity: waist > 102 cm, insulin resistance: HOMA-IR > 2.6. P value is calculated by analysis of covariance, corrected for age, use of insulin and current smoking. CLS = crown like structure. †CLS values were transformed to fulfil ANCOVA criteria, but for ease of interpretation values were back transformed to original mean number of CLS per depot

54

Distinct fat depots and vascular risk factors

Adipokine secretion profile of four distinct adipose tissue depots Overall, PAT of abdominally-lean patients was the most abundant producer of adipokines, with a slight shift in adipokine secretion in abdominally-obese patients. PAT of abdominallylean patients abundantly secreted acute phase proteins (chemerin, resistin and thrombopoietin (TPO)) as well as anti-inflammatory adipokines (IL-4, adiponectin; Figure 2). PAT of abdominaly-obese secreted higher levels of adipokines associated with atherosclerosis (tumor necrosis factor alpha (TNFα), IL-6, rantes, PAI), or involved in recruitment of immune cells to sites of inflammation (macrophage inflammatory protein 1α and 1β (MIP1α, MIP1β), cathepsin S; Figure 2).

Abdominal-lean OAT was the highest secretory depot of pro-inflammatory cytokines and growth factors (IL-18, IL-1β, macrophage migration inhibitory factor (MIF), TNF receptor 2 and hepatic growth factor (HGF); Figure 2), while abdominal-obese OAT secreted relatively higher amounts of adipokines involved in coagulation (PAI-1, TIMP metallopeptidase inhibitor 1(TIMP-1); Figure 2).

MAT showed a mild inflammatory profile in both abdominally-lean and obese patients, though leptin was most prominently secreted by MAT of abdominally-obese patients (Figure 2).

We observed no hotspots for SAT, which generally secreted the lowest amounts of adipokines of all AT-depots (Figure 2). Remarkably, secretion of the anti-inflammatory adipokine IL-10 was higher in all four AT-depots in abdominal-obese versus abdominally-lean patients, which might indicate a counteractive inflammatory response.

Morphologic AT characteristics in relation to systemic metabolic dysfunction To assess whether AT depot-specific differences between abdominally-lean and abdominallyobese patients were clinically relevant, we evaluated the relation between morphological characteristics of distinct abdominal AT depots, insulin resistance and dyslipidemia. In abdominally-lean patients, higher HOMA-IR levels were only associated with larger adipocytes in OAT (β0.23; 95%CI 0.05–0.41). In abdominally-obese patients, larger MAT and OAT adipocytes, and a higher number of CLS in MAT were related to higher HOMA-IR levels (β0.22; 95%CI 0.13–0.32, β0.18; 95%CI 0.06–0.31 and β0.07; 95%CI 0.01–0.12 respectively; Figure 3A). Overall, no relation between the capillary density or number of macrophages and levels of systemic insulin resistance was observed (Figure 3A).

Next the relation with systemic lipid levels was assessed, of which the number of CLS were most strongly related to derangements in lipid metabolism, shown in Figure 3B. No relation of AT characteristics with lipid levels was observed in abdominally-lean patients, while in abdominal-obese patients only CLS of MAT and OAT were related to systemic lipid levels, in contrast to SAT and PAT. Derangements of systemic lipid levels were most strongly related to OAT CLS, as higher number of CLS in OAT related negatively to HDL-cholesterol (β-0.05; 95%CI -0.10 – -0.01) and positively to triglyceride levels (β0.16; 95%CI 0.01–0.31), while CLS in MAT were only related to lower HDL-cholesterol levels (β-0.03; 95%CI -0.06 – -0.03).

In contrast to metabolic complications of obesity, no relation was found between morphologic AT characteristics and low-grade inflammation as measured by circulating hsCRP levels (Table 3).

55

3

56

0.025 (-0.131 – 0.181)

-0.015 (-0.082 – 0.053)

Waist > 102

0.013 (-0.053 – 0.079)

0.126 (-0.091 – 0.343)

MAT CLS

SAT CLS

Waist < 102

-0.003 (-0.377 – 0.372)

-0.004 (-0.374 – 0.366)

0.256 (-0.147 – 0.659)

-0.114 (-0.650 – 0.423)

Waist > 102

MAT macrophages

SAT macrophages

Waist < 102

0.093 (-0.768 – 0.954)

0.080 (-0.921 – 1.082)

-0.324 (-1.515 – 0.868)

-0.948 (-2.945 – 1.049)

Waist > 102

MAT capillary density

SAT capillary density

Waist < 102

0.036 (-0.127 – 0.198)

0.122 (-0.037 – 0.281)

0.080 (-0.206 – 0.366)

0.145 (-0.182 – 0.471)

Waist > 102

β (95%CI)

β (95%CI)

Waist < 102

MAT adipocyte size

SAT adipocyte size

0.032 (-0.072 – 0.137)

0.070 (-0.056 – 0.196)

OAT CLS

0.059 (-0.325 – 0.443)

0.020 (-0.400 – 0.441)

OAT macrophages

-0.525 (-1.560 – 0.511)

0.431 (-0.320 – 1.183)

OAT capillary density

0.067 (-0.090 – 0.224)

0.068 (-0.300 – 0.436)

β (95%CI)

OAT adipocyte size

0.083 (-0.686 – 0.853)

0.411 (0.014 – 0.8)

PAT CLS

0.015 (-0.319 – 0.349)

0.152 (-0.067 – 0.371)

PAT macrophages

0.075 (-0.415 – 0.564)

0.093 (-0.821 – 1.007)

PAT capillary density

0.069 (-0.065 – 0.202)

0.250 (-0.103 – 0.604)

β (95%CI)

PAT adipocyte size

Beta regression coefficients (β) with 95% confidence interval (CI) indicates the difference in log hsCRP per unit increase in adipocyte size, % capillary density, number of macrophages per high power field or number of crown-like-structures (CLS) per mm2. Estimated differences are based on linear regression models adjustment for age and current smoking.

Log hsCRP

Log hsCRP

Log hsCRP

Log hsCRP

Model

Table 3. Relation between morphological AT characteristics and low grade systemic inflammation in men with or without abdominal obesity.

CHAPTER 3

Distinct fat depots and vascular risk factors

Together, these results indicate that morphological characteristics of MAT were most strongly related to obesity-induced derangements in glucose metabolism, while presence of CLS in OAT was most strongly related to obesity-induced derangements in lipid metabolism.

3

Figure 2. Adipokine secretion profile of four distinct adipose tissue depots stratified by abdominal obesity. (A) A total of 28 adipokines measured by multiplex immunoassay.

Results are shown as a heat map, representing the median adipokine secretion levels in AT supernatant from 25 male men. Colors represent median adipokine concentrations relative to the minimum and maximum median adipokine value of four fat tissues, in abdominally-lean (white rows) or abdominally-obese (black rows); showing areas of relatively high (hot spots, yellow) or low (cold spots, blue) adipokine expression per gram fat for each AT-depot.

57

CHAPTER 3

A

Relation with insulin resistance

Waist < 102 cm Waist > 102cm

Adipocyte cell size

Capillary density 2

*

0.4

* *

β (95% CI)

β (95% CI)

0.8

0.0 -0.4 -0.8

SAT

MAT

OAT

1 0 -1 -2

PAT

Macrophages

β (95% CI)

β (95% CI)

0.0 -0.4

SAT

MAT

OAT

0.0 -0.2

SAT

MAT

β (95% CI)

0.0 -0.4

0.0 -0.5 -1.0 -1.5

-0.8

SAT

MAT

OAT

PAT

0.1

0.3

0.0

β (95% CI)

0.6

0.0

SAT

MAT

OAT

PAT

HDL-cholesterol

LDL-cholesterol

-0.3

PAT

Cholesterol

0.5

*

0.4

OAT

Waist < 102 cm Waist > 102cm

Relation with lipid metabolism Triglycerides

β (95% CI)

PAT

*

0.2

-0.4

PAT

0.8

β (95% CI)

OAT

Crown like structures

0.4

B

MAT

0.4

0.8

-0.8

SAT

*

*

MAT

OAT

-0.1 -0.2

-0.6

SAT

MAT

OAT

PAT

-0.3

SAT

PAT

Figure 3. Relation of morphological AT characteristics with metabolic diseases stratified by abdominal obesity. Beta regression coefficients (β) with 95% confidence interval (CI) indicates the difference in log HOMA-IR (A) per unit increase in adipocyte size (SAT n=23, MAT n=23, OAT n=23, PAT n=21), capillary density (SAT n=22, MAT n=22, OAT n=23, PAT n=20), number of macrophages or number of crown-likestructures (SAT n=23, MAT n=23, OAT n=23, PAT n=21), based on linear regression models adjusted for age and current smoking. (B) β(95%CI) indicates the difference in systemic triglyceride, cholesterol, HDLcholesterol and HDL-cholesterol levels per unit increase in number of crown-like-structures (SAT n=28, MAT n=28, OAT n=28, PAT n=26) based on multivariable linear regression models adjusted for age and current smoking. * P<0.05. HOMA-IR, homeostasis model assessment parameter of insulin resistance.

58

Distinct fat depots and vascular risk factors

DISCUSSION The morphology and ex vivo adipokine secretion of four abdominal AT-depots of abdominallyobese versus abdominally-lean patients undergoing abdominal aortic surgery, revealed an unique inflammatory profile for each AT-depot, which was differently related to systemic metabolic dysfunction. PAT was the most active AT-depot as reflected by the highest adipokine secretions, though morphological PAT characteristics were not related to systemic metabolic dysfunction. In contrast, the mild inflammatory profile of abdominal-obese MAT was most strongly related to insulin resistance, while abdominal-obese OAT was most strongly related to altered systemic lipid levels.

Overall, AT of abdominally-obese patients exhibited a higher inflammatory profile compared to abdominally-lean patients, which is in agreement with current literature (5,27-29).

Furthermore, inflammation in the main visceral depots MAT and OAT was related to systemic metabolic dysfunction in abdominal-obese but not abdominal-lean patients. This indicates that even in patients with established vascular disease, the amount of visceral AT-inflammation is indicative for metabolic dysfunction preceding type 2 diabetes and vascular diseases. As these patients are at higher risk for type 2 diabetes, low grade systemic inflammation and recurrent vascular events, it might be wise to monitor abdominally-obese patients even closer than patients with vascular disease but without abdominal obesity. Furthermore, as the data presented here suggest an etiological relation between crown-like-structures and derangements in glucose or lipid metabolism, it might be desirable to aim for development of drugs which specifically reduce AT-inflammation as is shown for rosiglitazone and Îą-lipoic acid in mice (30,31).

The number of macrophages was not informative regarding status of AT-inflammation or clinical metabolic derangements. The CD68 marker used for macrophages does not distinguish between pro-or anti-inflammatory subtypes, which might explain why no differences were observed among abdominally-obese versus lean patients. However, also the number of CD68-macrophages has been reported to be related to other inflammatory markers or insulin resistance (32). Nonetheless, in contrast to macrophage number, aggregation of CD68+ macrophages into CLS was strongly related to metabolic dysfunction in both glucose and lipid metabolism, in particular CLS of MAT and OAT of abdominallyobese patients. Although on average there was no difference in the number of CLS in MAT and OAT among abdominally-lean or obese (Table 2), the significant relation with glucose or lipid dysfunction points toward an important contribution of the amount of CLS at an individual level. CLS are regarded a pathologic hallmark of obesity, which are associated with adipocyte necrosis and a highly active state of the macrophages (27). Therefore, while the role of individual CD68 macrophages in AT-inflammation seems unclear in this study considering men with established vascular disease, the aggregation of these cells into CLS seems to contribute to obesity-associated metabolic dysfunction.

Besides differences in the AT profile among abdominally-lean or obese, interesting differences among distinct abdominal AT-depots were observed. Although in close proximity to other visceral fat tissues, PAT is distinguished by relatively small adipocytes and a rich

59

3

CHAPTER 3

capillary network. Furthermore, PAT secreted abundant amounts of adipokines related to atherosclerosis, such as TNFÎą, TPO and resistin (33,34), and chemokines like MCP-1, MIP1Îą and rantes, involved in recruitment of immune cells to sites of inflammation (35,36). It should be noted that in the present study the PAT biopsies were obtained from regions close to the aortic aneurysm or occlusion, which might have contributed to the inflammatory status of PAT. Interestingly, the morphological characteristics of PAT, including CLS which are active secretors of mentioned adipokines, were not related to systemic metabolic dysfunction in either abdominal-lean or obese patients. This supports the current idea that AT surrounding arteries may act locally, and directly affect atherogenic processes in the vessel wall from outside to inside (17,37-39).

Another interesting finding was that while MAT showed a relatively mild inflammatory profile, MAT-characteristics were related to insulin resistance in contrast to the other depots. Also based on lipolysis studies, MAT was shown to be uniquely related to obesity related diabetes (40). Of particular interest is the contrast with OAT, of which the AT inflammatory profile in abdominal-obese men is almost identical to that of MAT regarding adipocyte size, capillary density and number of macrophages (Figure 1). Therefore, we believe that an explanation for the relation between MAT and metabolic dysfunction should be sought in the pathway by which MAT influences metabolism, rather than the amount of inflammatory markers. Anatomically, MAT is dispersed in between the intestine, a very active endocrine organ (41). Both murine and human studies have shown that MAT mass and function alters in mice suffering from colitis, indicating an active interaction between the intestine and MAT (42-44). Furthermore, diet has a more pronounced influence on MAT compared to other, more distant, AT-depots (45,46). We have observed increased leptin secretion in particular by MAT of abdominally-obese patients, which is often associated with insulin resistance (47). The leptin receptor is expressed in human intestinal endocrine L cells, and mice studies report that physiological levels of leptin stimulate glucagon-like peptide-1 (GLP-1) secretion, whereas leptin resistance was associated with lower GLP-1 levels and a diminished GLP-1 response to glucose (48). Therefore, it can be hypothesized that MAT-inflammation has a direct influence on the enteric endocrine system, interfering with secretion of insulin-sensitizing hormones including GLP-1.

OAT on the other hand was more strongly associated with disturbances in lipid metabolism.

The link between OAT and liver metabolism is far more established, and the secretion of PAI-1, chemerin and TIMP-1, adipokines associated with liver steatosis, may well underlie the strong relation between OAT and changes in lipid metabolism in abdominally-obese men (28,49).

Strengths of the study are that AT biopsies were obtained within one patient which provided the unique opportunity to determine the inflammatory signature of four distinct abdominal AT-depots among abdominally-lean and abdominally-obese patients. Furthermore, we were able to relate this inflammatory profile to changes in metabolic and inflammatory processes, known to be related to development of type 2 diabetes and vascular diseases (13,23,24).

Some limitations of the study have to be considered. First, the study population consisted of men with aortic vascular disease, and therefore generalization of the results to women

60

Distinct fat depots and vascular risk factors

or to men without vascular disease needs to be done with caution. Second, due to the limited amount of elective open abdominal aneurysm surgeries, we could only include a limited number of patients in this study.

In conclusion, we observed differences in the morphological and adipokine profile of four AT-depots between abdominally-lean and abdominally-obese patients, of which morphological characteristics of MAT were most strongly related to obesity-induced derangements in glucose metabolism, while presence of CLS in OAT was most strongly related to obesity-induced derangements in lipid metabolism. Our observations suggest differential contributions of distinct visceral AT-depots in the pathogenesis of obesityinduced metabolic dysfunctions. As there still are few successful treatment options for obesity, we feel that the mechanisms by which distinct AT-depots contribute to systemic metabolic derangements should be studied more closely to aim for a more targeted therapeutic approach for different AT-depots.

Acknowledgments We gratefully acknowledge members of the department of vascular surgery at the UMC Utrecht for collection of adipose tissue biopsies during surgery: P. Berger, R.J. Toorop, E.J. Waasdorp and C.E.V. Hazenberg. Also we like to thank prof. R. Bleys for helpful discussions.

All adipokine measurements were performed at the UMC Utrecht Luminex Core Facility.

61

3

CHAPTER 3

REFERENCES 1.

Finucane MM, Stevens GA, Cowan MJ, Danaei G, Lin JK, Paciorek CJ, et al. National, regional, and global trends in body-mass index since 1980: Systematic analysis of health examination surveys and epidemiological studies with 960 country-years and 9.1 million participants. Lancet.

2011 Feb 12;377(9765):557-67.

2. Guilherme A, Virbasius JV, Puri V, Czech MP. Adipocyte dysfunctions linking obesity to insulin resistance and type 2 diabetes. Nat Rev Mol Cell Biol. 2008 May;9(5):367-77.

3. Grundy SM, Brewer HB,Jr, Cleeman JI, Smith SC,Jr, Lenfant C, American Heart Association, et al. Definition of metabolic syndrome: Report of the national heart, lung, and blood Institute/ American heart association conference on scientific issues related to definition. Circulation.

2004 Jan 27;109(3):433-8.

4. Anghel SI, Wahli W. Fat poetry: A kingdom for PPAR gamma. Cell Res. 2007 Jun;17(6):486-511.

5. Weisberg SP, McCann D, Desai M, Rosenbaum M, Leibel RL, Ferrante AW,Jr. Obesity is associated with macrophage accumulation in adipose tissue. J Clin Invest. 2003 Dec;112(12):1796-808.

6. Hajer GR, van Haeften TW, Visseren FL. Adipose tissue dysfunction in obesity, diabetes, and vascular diseases. Eur Heart J. 2008 Dec;29(24):2959-71.

7. Gealekman O, Guseva N, Hartigan C, Apotheker S, Gorgoglione M, Gurav K, et al. Depotspecific differences and insufficient subcutaneous adipose tissue angiogenesis in human obesity. Circulation. 2011 Jan 18;123(2):186-94.

8. Le KA, Mahurkar S, Alderete TL, Hasson RE, Adam TC, Kim JS, et al. Subcutaneous adipose tissue macrophage infiltration is associated with hepatic and visceral fat deposition, hyperinsulinemia, and stimulation of NF-kappaB stress pathway. Diabetes. 2011 Nov;60(11):2802-9.

9. Wentworth JM, Naselli G, Brown WA, Doyle L, Phipson B, Smyth GK, et al. Pro-inflammatory CD11c+CD206+ adipose tissue macrophages are associated with insulin resistance in human obesity. Diabetes. 2010 Jul;59(7):1648-56.

10. Shen W, Wang Z, Punyanita M, Lei J, Sinav A, Kral JG, et al. Adipose tissue quantification by imaging methods: A proposed classification. Obes Res. 2003 Jan;11(1):5-16.

11. Schlett CL, Massaro JM, Lehman SJ, Bamberg F, O'Donnell CJ, Fox CS, et al. Novel measurements of periaortic adipose tissue in comparison to anthropometric measures of obesity, and abdominal adipose tissue. Int J Obes (Lond). 2009 Feb;33(2):226-32.

12. Nicklas BJ, Penninx BW, Cesari M, Kritchevsky SB, Newman AB, Kanaya AM, et al. Association of visceral adipose tissue with incident myocardial infarction in older men and women: The health, aging and body composition study. Am J Epidemiol. 2004 Oct 15;160(8):741-9.

13. Kanhai DA, Kappelle LJ, van der Graaf Y, Uiterwaal CS, Visseren FL, SMART Study Group.

The risk of general and abdominal adiposity in the occurrence of new vascular events and mortality in patients with various manifestations of vascular disease. Int J Obes (Lond). 2012 May;36(5):695-702.

14. Fain JN. Release of interleukins and other inflammatory cytokines by human adipose tissue is enhanced in obesity and primarily due to the nonfat cells. Vitam Horm. 2006;74:443-77.

15. Harman-Boehm I, Bluher M, Redel H, Sion-Vardy N, Ovadia S, Avinoach E, et al. Macrophage infiltration into omental versus subcutaneous fat across different populations: Effect of regional adiposity and the comorbidities of obesity. J Clin Endocrinol Metab. 2007 Jun;92(6):2240-7.

16. Bremer AA, Devaraj S, Afify A, Jialal I. Adipose tissue dysregulation in patients with metabolic syndrome. J Clin Endocrinol Metab. 2011 Aug 24.

17. Police SB, Thatcher SE, Charnigo R, Daugherty A, Cassis LA. Obesity promotes inflammation in periaortic adipose tissue and angiotensin II-induced abdominal aortic aneurysm formation.

Arterioscler Thromb Vasc Biol. 2009 Oct;29(10):1458-64.

18. Chatterjee TK, Stoll LL, Denning GM, Harrelson A, Blomkalns AL, Idelman G, et al.

Proinflammatory phenotype of perivascular adipocytes: Influence of high-fat feeding. Circ Res. 2009 Feb 27;104(4):541-9.

62

Distinct fat depots and vascular risk factors

19. Abate N, Garg A. Heterogeneity in adipose tissue metabolism: Causes, implications and management of regional adiposity. Prog Lipid Res. 1995;34(1):53-70.

20. Faber DR, Moll FL, Vink A, van der Waal C, Kalkhoven E, Schipper HS, et al. Adipose tissue quantity and composition contribute to adipokine concentrations in the subclavian vein and the inferior mesenteric vein. Int J Obes (Lond). 2012 Aug;36(8):1078-85.

21. Fain JN. Release of inflammatory mediators by human adipose tissue is enhanced in obesity and primarily by the nonfat cells: A review. Mediators Inflamm. 2010;2010:513948.

22. Bigornia SJ, Farb MG, Mott MM, Hess DT, Carmine B, Fiscale A, et al. Relation of depot-specific adipose inflammation to insulin resistance in human obesity. Nutr Diabetes. 2012 Mar 5;2:e30.

23. Wassink AM, van der Graaf Y, van Haeften TW, Spiering W, Soedamah-Muthu SS, Visseren FL, et al. Waist circumference and metabolic risk factors have separate and additive effects on the risk of future type 2 diabetes in patients with vascular diseases. A cohort study. Diabet Med. 2011 Aug;28(8):932-40.

24. Verhagen SN, Wassink AM, van der Graaf Y, Gorter PM, Visseren FL, SMART Study Group.

Insulin resistance increases the occurrence of new cardiovascular events in patients with manifest arterial disease without known diabetes. the SMART study. Cardiovasc Diabetol. 2011 Nov 21;10:100,2840-10-100.

25. Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: Insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia. 1985 Jul;28(7):412-9.

26. Schipper HS, de Jager W, van Dijk ME, Meerding J, Zelissen PM, Adan RA, et al. A multiplex immunoassay for human adipokine profiling. Clin Chem. 2010 Aug;56(8):1320-8.

27. Cinti S, Mitchell G, Barbatelli G, Murano I, Ceresi E, Faloia E, et al. Adipocyte death defines macrophage localization and function in adipose tissue of obese mice and humans. J Lipid Res. 2005 Nov;46(11):2347-55.

28. Sell H, Divoux A, Poitou C, Basdevant A, Bouillot JL, Bedossa P, et al. Chemerin correlates with markers for fatty liver in morbidly obese patients and strongly decreases after weight loss induced by bariatric surgery. J Clin Endocrinol Metab. 2010 Jun;95(6):2892-6.

29. Poitou C, Dalmas E, Renovato M, Benhamo V, Hajduch F, Abdennour M, et al. CD14dimCD16+ and CD14+CD16+ monocytes in obesity and during weight loss: Relationships with fat mass and subclinical atherosclerosis. Arterioscler Thromb Vasc Biol. 2011 Oct;31(10):2322-30.

30. Deiuliis JA, Kampfrath T, Ying Z, Maiseyeu A, Rajagopalan S. Lipoic acid attenuates innate immune infiltration and activation in the visceral adipose tissue of obese insulin resistant mice.

Lipids. 2011 Aug 23.

31. Rossmeisl M, Medrikova D, van Schothorst EM, Pavlisova J, Kuda O, Hensler M, et al. Omega-3 phospholipids from fish suppress hepatic steatosis by integrated inhibition of biosynthetic pathways in dietary obese mice. Biochim Biophys Acta. 2013 Dec 1;1841(2):267-78.

32. Di Gregorio GB, Yao-Borengasser A, Rasouli N, Varma V, Lu T, Miles LM, et al. Expression of CD68 and macrophage chemoattractant protein-1 genes in human adipose and muscle tissues: Association with cytokine expression, insulin resistance, and reduction by pioglitazone.

Diabetes. 2005 Aug;54(8):2305-13.

33. Reilly MP, Lehrke M, Wolfe ML, Rohatgi A, Lazar MA, Rader DJ. Resistin is an inflammatory marker of atherosclerosis in humans. Circulation. 2005 Feb 22;111(7):932-9.

34. Gustafson B. Adipose tissue, inflammation and atherosclerosis. J Atheroscler Thromb. 2010 Feb 3.

35. Dahlman I, Kaaman M, Olsson T, Tan GD, Bickerton AS, Wahlen K, et al. A unique role of monocyte chemoattractant protein 1 among chemokines in adipose tissue of obese subjects. J Clin Endocrinol Metab. 2005 Oct;90(10):5834-40.

36. Karastergiou K, Evans I, Ogston N, Miheisi N, Nair D, Kaski JC, et al. Epicardial adipokines in obesity and coronary artery disease induce atherogenic changes in monocytes and endothelial cells. Arterioscler Thromb Vasc Biol. 2010 Jul;30(7):1340-6.

37. Verhagen SN, Visseren FL. Perivascular adipose tissue as a cause of atherosclerosis.

Atherosclerosis. 2011 Jan;214(1):3-10.

63

3

CHAPTER 3

38. Verhagen SN, Vink A, van der Graaf Y, Visseren FL. Coronary perivascular adipose tissue characteristics are related to atherosclerotic plaque size and composition. A post-mortem study. Atherosclerosis. 2012 Nov;225(1):99-104.

39. Zhang H, Zhang C. Regulation of microvascular function by adipose tissue in obesity and type 2 diabetes: Evidence of an adipose-vascular loop. Am J Biomed Sci. 2009 Apr 1;1(2):133-42.

40. Yang YK, Chen M, Clements RH, Abrams GA, Aprahamian CJ, Harmon CM. Human mesenteric adipose tissue plays unique role versus subcutaneous and omental fat in obesity related diabetes. Cell Physiol Biochem. 2008;22(5-6):531-8.

41. Baggio LL, Drucker DJ. Biology of incretins: GLP-1 and GIP. Gastroenterology. 2007 May;132(6):2131-57.

42. Gambero A, Marostica M, Abdalla Saad MJ, Pedrazzoli J,Jr. Mesenteric adipose tissue alterations resulting from experimental reactivated colitis. Inflamm Bowel Dis. 2007 Nov;13(11):1357-64.

43. Desreumaux P, Ernst O, Geboes K, Gambiez L, Berrebi D, Muller-Alouf H, et al. Inflammatory alterations in mesenteric adipose tissue in crohn's disease. Gastroenterology. 1999 Jul;117(1):73-81.

44. Peyrin-Biroulet L, Chamaillard M, Gonzalez F, Beclin E, Decourcelle C, Antunes L, et al.

Mesenteric fat in crohn's disease: A pathogenetic hallmark or an innocent bystander? Gut. 2007 Apr;56(4):577-83.

45. Feng Y, Li Y, Mei S, Zhang L, Gong J, Gu L, et al. Exclusive enteral nutrition ameliorates mesenteric adipose tissue alterations in patients with active crohn's disease. Clin Nutr. 2013 Oct 18.

46. Ludwig T, Worsch S, Heikenwalder M, Daniel H, Hauner H, Bader BL. Metabolic and immunomodulatory effects of n-3 fatty acids are different in mesenteric and epididymal adipose tissue of diet-induced obese mice. Am J Physiol Endocrinol Metab. 2013 Jun 1;304(11):E114056.

47. Evans J, Goedecke JH, Soderstrom I, Buren J, Alvehus M, Blomquist C, et al. Depot- and ethnic-specific differences in the relationship between adipose tissue inflammation and insulin sensitivity. Clin Endocrinol (Oxf). 2011 Jan;74(1):51-9.

48. Anini Y, Brubaker PL. Role of leptin in the regulation of glucagon-like peptide-1 secretion.

Diabetes. 2003 Feb;52(2):252-9.

49. Mathew M, Tay E, Cusi K. Elevated plasma free fatty acids increase cardiovascular risk by inducing plasma biomarkers of endothelial activation, myeloperoxidase and PAI-1 in healthy subjects. Cardiovasc Diabetol. 2010 Feb 16;9:9.

64

PART TWO

Adipose tissue extracellular vesicles in adipose tissue dysfunction and metabolic disease

CHAPTER 4 Human adipocyte extracellular vesicles in reciprocal signaling between adipocytes and macrophages Obesity (Silver Spring). 2013 Dec 12. doi: 10.1002/oby.20679

MariĂŤtte E. G. Kranendonk, Frank L. J. Visseren, Bas W.M. van Balkom, Esther N.M. Nolte-â&#x20AC;&#x2122;t Hoen, Joost A. van Herwaarden, Wilco de Jager, Henk S. Schipper, Arjan B. Brenkman, Marianne C. Verhaar, Marca H.M. Wauben, Eric Kalkhoven

CHAPTER 4

ABSTRACT Objective We determined whether extracellular vesicles (EVs) released by human adipocytes or ATexplants play a role in the paracrine interaction between adipocytes and macrophages, a key mechanism in adipose tissue (AT) inflammation, leading to metabolic complications like insulin resistance.

Design and Methods EVs released from in vitro differentiated adipocytes and AT-explants ex vivo were characterized by electron microscopy, Western blot, multiplex adipokine-profiling and quantified by flow cytometry. Primary monocytes were stimulated with EVs from adipocytes, subcutaneous (SCAT) or omental-derived AT (OAT), and phenotyped.

Macrophage supernatant was subsequently used to assess the effect on insulin signaling in adipocytes.

Results Adipocyte and AT-derived EVs differentiated monocytes into macrophages characteristic of human adipose tissue macrophages (ATM), defined by release of both pro- and antiinflammatory cytokines. The adiponectin-positive subset of AT-derived EVs, presumably representing adipocyte-derived EVs, induced a more pronounced ATM-phenotype than the adiponectin-negative AT-EVs. This effect was more evident for OAT-EVs versus SCAT-EVs. Furthermore, supernatant of macrophages pre-stimulated with AT-EVs interfered with insulin signaling in human adipocytes. Finally, the number of OAT-derived EVs correlated positively with patients HOMA-IR. Conclusions We demonstrate a possible role for human AT-EVs in a reciprocal pro-inflammatory loop between adipocytes and macrophages, with the potential to aggravate local and systemic insulin resistance. â&#x20AC;&#x192;

68

Human Adipocyte EVs and Macrophages

INTRODUCTION The worldwide prevalence of obesity has doubled over the last 20 years (1). Obesity often engenders insulin resistance (IR), low-grade systemic inflammation, type 2 diabetes and dyslipidemia, conditions that substantially increase the risk for cardiovascular disease (CVD) (2,3). In particular visceral adipose tissue, including omental adipose tissue (OAT), is associated with adverse metabolic consequences of obesity (4). Besides expansion of (O) AT, obesity coincides with AT-inflammation, characterized by hypertrophic adipocytes, increased influx of immune cells and inflammasome activation (5,6). Subsequent disproportionate pro-inflammatory adipokine secretion elicits a local immune response, propagated by communication between adipocytes and macrophages, generating a selfsustaining reciprocal pro-inflammatory loop (7). Soluble chemokines and cytokines, secreted by adipocytes, are involved in macrophage recruitment and activation and thereby important mediators in the cellular crosstalk that help to sustain AT-inflammation (8,9). Furthermore, AT-resident immune cells, both from the innate and adaptive immune system, are important mediators in AT-homeostasis and inflammation (10). The best-studied contributors to ATinflammation are monocytes, which infiltrate AT where they differentiate into macrophages.

While in mice a clear distinction between pro-inflammatory (M1) and anti-inflammatory (M2) macrophages can be made (11), human AT macrophages (ATM) display a mixed phenotype (12,13).

However, besides soluble proteins and cells, extracellular vesicles (EVs) also play important roles in intercellular communication. EVs are a mix of distinct vesicles released by the cell, including microvesicles or microparticles that bud from the plasma membrane, and exosomes which are released upon fusion of multivesicular bodies with the plasma membrane (14). As it is currently not possible to distinguish the different EVs based on specific markers, we define the observed vesicle population simply as EVs. The morphological and structural properties of EVs enable them to interact with and modify target cells, and crucial roles of EVs have been shown in many (patho)physiological processes, like immune responses, inflammation and lipid metabolism (15,16).

The immunomodulatory properties of EVs suggest active participation in AT-homeostasis and inflammation through intercellular communication with resident immune cells.

Understanding communication between adipocytes and immune cells in obese subjects might enable interventions in the process of AT-inflammation and thereby prevent obesityinduced vascular pathology. While a role of adipose tissue EVs on systemic immunometabolism has been described in a murine model (17), the relevance of human adipose tissue EVs and their role in the paracrine signaling of AT inflammation has not been reported so far.

In the present study we characterized EVs secreted by in vitro differentiated human adipocytes and human adipose tissue explants ex vivo, both from subcutaneous and visceral fat. We determined whether human adipocyte and adipose tissue EVs could have immunomodulatory effects on macrophages and further studied the reciprocal effects of adipose EV-stimulated macrophages on insulin signaling in adipocytes. Finally, we evaluated the relation between AT derived EVs and systemic insulin resistance in overweight patients.â&#x20AC;&#x192;

69

4

CHAPTER 4

METHODS AND PROCEDURES In vitro differentiated adipocytes Human SGBS pre-adipocytes were cultured and differentiated into adipocytes as described previously (18).

Ex vivo AT explants Abdominal subcutaneous and omental AT-explants were collected from subjects undergoing surgery for aneurysmatic aortic disease at the University Medical Center Utrecht (UMCU).

The study was approved by the Medical Ethics Committee of the UMCU and all study subjects gave written informed consent. AT-explants were incubated in DMEMf12 supplemented with 50 IU/ml penicillin and 50 mg/ml streptomycin (PS) for 24 hours and weighed afterwards. Culture supernatant was centrifuged for 10 minutes at 500g to remove cells and stored at -80°C until EV isolation.

Isolation of EVs EVs were isolated from 48-72 hour supernatant of 50x106 in vitro differentiated adipocytes or 1-2 gram of ex vivo AT-explants by differential steps of (ultra)centrifugation as described previously (19). In short, collected supernatant was centrifuged sequentially for 10 minutes at 500g, 15 minutes at 1500g, 30 minutes at 10,000g and 2 hours at 100,000g using a Beckman LE-80K centrifuge with SW32-Ti, SW40-Ti and SW60-Ti rotors (Beckman Instruments, Inc., Fullerton, CA, USA.) 100,000g pelleted EVs were washed in PBS and centrifuged again for 2 hours at 100,000g. The final pellet was resuspended in 2.5 M sucrose and subfractioned by sucrose gradient (SG)-ultracentrifugation. A linear SG (2.5â&#x20AC;&#x201C;0.4 M sucrose, 20 mM Tris-HCL pH7,4) was layered on top of EVs and ultracentrifuged for 14-16 hours at 190,000g.

Immunoblotting SG-fractions were mixed 1:1 with lysis buffer (20), equal amounts were subjected to SDSPAGE and transferred to PVDF membranes (21). Proteins were detected with AmershamECL prime Western Blotting Detection reagent and imaged with Imagequant Las 4000.

Primary antibodies against Flotillin-1 (Santa-Cruz; sc-25506), FABP-4 (Santa-Cruz, sc-18661), CD9 (Santa-Cruz, sc-53679), CD63 (Millipore, CBL553) and adiponectin (R&D-systems, AF1065) were used.

Electron Microscopy Transmission electron microscopy was performed as described previously (22). Briefly, carbon-coated formvar filmed grids were incubated with the EV suspension, washed with 0.15% glycine in PBS, fixed in 1% gluteraldehyde in PBS and washed twice in PBS and distilled water. Next, grids were incubated with 1.8% methylcellulose/0.4% uranyl acetate, dried and EVs were visualized using a FEI Tecnai 12 (FEI, Hillsboro, OR, USA) transmission electron microscope.

70

Human Adipocyte EVs and Macrophages

Flow cytometric analysis of EVs EVs were analyzed on the BD InfluxTM flow cytometer (Becton Dickinson, San-Jose, USA) as described previously (23). Briefly, 100,000g pelleted EVs were resuspended in 20 µl PBS with 0.1% BSA (cleared from aggregates by ultracentrifugation), labeled with PKH67 (7.5 µM; Sigma-Aldrich), and subfractioned by SG-ultracentrifugation overnight. One ml fractions were collected from the bottom of the tube and diluted 1:200 (in vitro differentiated adipocyte-EVs) or 1:100 (ex vivo AT-explant EVs) in PBS before measurement. The system was triggered on the fluorescence signal derived from the fluorescently-labeled EVs and thresholding was applied on this fluorescence channel. Fluorescent 100 and 200 nm polystyrene beads (yellow-green-fluorescent FluoSpheres, Invitrogen) were used to calibrate the fluorescence and reduced wide angle-FSC settings on the flow cytometer.

Time-based quantitative measurements were performed. Data was acquired using Spigot software version 6.1 (Becton Dickinson) and analyzed using FCS Express software (De Novo Software).

Multiplex immunoassay Supernatants from pelleted EVs derived from in vitro differentiated adipocytes were measured undiluted. The pelleted adipocyte-derived EVs were lysed in Roche cOmplete lysisM buffer, supplemented with cOmplete protease inhibitors (Roche). Of each sample 10 µg of protein was used for adipokine measurements in triplicate. All multiplex immunoassays were carried out using the Bio-Plex system with Bio-PLex Manager Software version 6.1.1 (Biorad Laboratories, Hercules, CA). Antibody pairs and recombinant proteins used were described earlier (24).

Monocyte isolation and functional assays Human peripheral blood mononuclear cells were isolated from buffy coat bags obtained from healthy blood donors (Sanquin Blood Bank, Amsterdam, The Netherlands) by Ficoll density gradient centrifugation using Ficoll Paque Plus (Amersham Biosciences, Uppsala, Sweden). Monocytes were isolated by negative selection with the Human Monocyte Enrichment Set (BD Imag biosciences), suspended in RPMI 1640, GlutaMAX, 10% heat inactivated (HI) pooled human serum (HS) (Sanquin, Amsterdam, The Netherlands) and PS, and seeded at a density of 2x106 cells/well in 6-well plates coated with HI-pooled-HS. In all experiments, pooled EVs from all SG-fractions except the bottom fractions were added to monocytes at day 2 and 5 (in vitro differentiated adipocyte-EVs) or day 2 (ex vivo AT explant EVs) of culture, and PBS was used as a negative control. In vitro differentiated adipocyte-EVs were added in two different concentrations equivalent to supernatant of 1x106 and 10x106 adipocytes, to monocytes from three different blood donors. Monocyte assays with AT-EVs were performed with monocytes from one blood donor and AT-EVs from 3 different AT-donors per experiment, added separately to monocytes in duplicate.

In each experiment equal EV numbers were used, quantitated by measurement on the BD InfluxTM flow cytometer (Becton Dickinson, San Jose, USA). At day 6 of culture fully matured macrophages and their supernatants were harvested for further analysis.

71

4

CHAPTER 4

Insulin signaling in SGBS-adipocytes In vitro differentiated SGBS-adipocytes were incubated in serum-free DMEMf12 supplemented with 50% macrophage conditioned medium of macrophages pre-treated with SCAT-EVs, OAT-EVs or vehicle (PBS) for 24 hours, after which cells were stimulated with insulin (20 nM) for 10 minutes and lysed in ice cold lysis buffer. Cell lysates were subjected to immunoblot analysis to detect Akt phosphorylation. Images were taken with Imagequant Las 4000 and quantification of Akt and phosphorylated Akt bands was performed using ImageQuant TL Software. Antibodies were from Cell Signaling Technology (anti-Akt #9272; anti-phosphorylated Akt-Ser473, #9271).

RNA extraction and analysis RNA isolation and cDNA synthesis were performed as previously described (21). Primers for quantitative RT-PCR were designed with the universal probe library (Roche), and are described in Table S2. Specificity of the amplification was verified by melt curve analysis and evaluation of efficiency of PCR amplification. The mRNA expression of genes reported was normalized to 36B4 expression.

Data analysis Statistical evaluation was performed using GraphPad Prism software, version 5.03 (GraphPad Software, La Jolla, CA) and SPSS 20.0 for Windows (SPSS, Inc., Chicago, IL).

Data are presented as meanÂąSEM. Basic descriptive statistics were used to describe the subject characteristics. Differences between experimental samples were evaluated by one way ANOVA with Bonferroni or Dunnets post hoc test for parametric data, or Kruskal-Wallis ANOVA with Dunnâ&#x20AC;&#x2122;s post hoc test for non-parametric comparisons. Associations between two variables was measured by Pearson correlation. P-values <0.05 were considered significant.

RESULTS Extracellular vesicles released from human in vitro differentiated adipocytes contain adipocyte-specific and immunomodulatory proteins.

Extracellular vesicles (EVs) released by human SGBS-adipocytes were first visualized by transmission electron microscopy, where EVs of various sizes were observed (Figure 1A).

Next, fluorescence-based high resolution flow cytometry was performed, a recently developed method for analysis of single EVs (23). EVs labeled with the membrane dye PKH67 could well be detected above the fluorescence threshold that distinguishes EVs from noise signals (Figure 1B; (23), and were primarily detected at densities of 1.12-1.14 g/ml (Figure 1C). To prove presence of vesicles rather than soluble proteins, sucrose density gradient experiments were performed, of which the gradient fractions were subjected to immunoblot analysis. The strongest signal was observed at densities between 1.09-1.14 g/ml, and general EV-associated proteins like CD9, CD63 and flotillin-1 (19), as well as

72

Human Adipocyte EVs and Macrophages

A

B

C Number of EVs

PKH67 fluorescence

threshold Reduced wide angle FSC

D

4

1.5×10 1 0

Density ( g/ml )

1.03 1.05 1.07 1.09 1.11 1.12

1.0×10 1 0

5.0×10 0 9

0

1.08 1.10 1.12 1.14 1.16 1.18 1.20 1.22 1.24 1.26

Density (g/ml)

1.13 1.14 1.16 1.17 1.19 1.20

CD9

CD63 FLOT1

FABP4

ADIPOQ

Figure 1. Extracellular vesicles released from human in vitro differentiated adipocytes contain adipocyte-specific and immunomodulatory proteins. (A) Morphological characterization of 100,000g pelleted whole mounted EVs from SGBS-adipocytes, observed by transmission electron microscopy. (B) Dot plot of reduced wide angle-FSC versus PKH67 fluorescence of SGBS-A EVs from density fraction 1.12 g/ml is shown. (C) Quantitative analysis of 100,000g pelleted adipocyte-EVs by flow cytometry. EVs were labeled with PKH67, separated by SG ultracentrifugation overnight, and diluted SG fractions were analyzed by a time based quantification. Indicated are the number of events measured in 30 seconds within each sucrose gradient fraction, using a threshold on PKH67 fluorescence. (D) Immunoblot analysis of pelleted adipocyte-EVs separated by sucrose density gradient (SG) ultracentrifugation. Representative immunoblots for detection of CD9, CD63, flotillin-1, fatty acid binding protein (FABP4) and adiponectin are shown.

73

CHAPTER 4

A

B

50

10

5

concentration (pg/ml)

IL-8 IP-10 Leptin

60

GM-CSF

** ***

40 20 0

MIF MIP-1α

**

NGF

10

concentration (pg/ml)

3000

Resistin TNF- α

**

20

0

2000

104

***

***

1000

Adiponectin

0.8

IL-1β

***

**

0.6 0.4 0.2

RBP-4 Thrombopoietin

Chemerin *** ***

150 100 50

200 150 100 50

80

1400

700

20

105

106

6000 4000 2000 0

***

***

**

20000 15000 10000 5000

30000

6

NGF

**

4

2

80

Resistin

***

**

60 40 20

12

TNFα

8

4

*

600000

Adiponectin

400000

200000

0

800000

RBP-4

600000 400000 200000 0

0

concentration (pg/ml)

104

concentration (pg/ml)

103

8000

**

40

Thrombopoietin

100

0

MCSF

60

25000

200

0

MCP-1

PAI-1

0

Timp-1

**

200

250

MIP-1α

0

**

0

concentration (pg/ml)

concentration (pg/ml)

PAI-1

2100

1

0

0.0

Chemerin

ND ≤10 102 pg/ ml

1.0

*

2

IL-1RA

concentration (pg/ml)

103

concentration (pg/ml)

102

*

3

250

6000

0

IP-10

0

0

101

100

Leptin

HGF concentration (pg/ml)

concentration (pg/ml)

M-CSF

* **

200

4

12000

300

0

concentration (pg/ml)

MCP-1

30

300

MIF

18000

0

IL-8

0

concentration (pg/ml)

80

IL-6

ND ≤ 100 pg/ ml

400

concentration (pg/ml)

Cathepsin S

0

IL-1β

5

concentration (pg/ml)

IL-1RA

10

concentration (pg/ml)

GM-CSF HGF

15

IL-6

0

concentration (pg/ml)

concentration (pg/ml)

0

15

concentration (pg/ml)

***

concentration (pg/ml)

Adipsin Cathepsin S

100

***

concentration (pg/ml)

+

Adipsin

concentration (pg/ml)

150

+

concentration (pg/ml)

+ +

concentration (pg/ml)

Supernatant EVs

TIMP-1

Supernatant + EVs

20000

EVs 10000

Supernatant - EVs

0

Figure 2. Adipokine profiling of in vitro differentiated adipocyte-EVs reveals a specific adipokine profile. (A) A total of 23 adipokines were measured by multiplex immunoassay. Results shown are mean adipokine values in pg/ml of three independent experiments, depicted as a heat map. The scale of adipokine concentrations is displayed in 2 ranges: 1 to 100,000 pg/ml in the upper panel and 100 to 10,000,000 pg/ ml in the lower panel. To allow comparison of EV-associated proteins with soluble secreted proteins in the supernatant, adipocyte-EVs were lysed prior to analysis, while supernatant was left untreated. EVs were lysed with Roche cOmplete lysis M buffer prior to measurement. (B) Graphs of all adipokines measured by luminex, showing different distributions of adipokines in adipocyte supernatant, lysed EVs or EV depleted supernatant. One way ANOVA with Bonferroni post hoc test was used to assess differences between adipokine levels (*P < 0.05, ** P < 0.01, *** P<0.001).

74

Human Adipocyte EVs and Macrophages

adipocyte-specific proteins FABP4 and adiponectin (25), could clearly be detected in adipocyte-EVs (Figure 1D). To further explore the adipokine profile of adipocyte-EVs, a range of adipokines was measured by multiplex immunoassay, which confirmed the presence of adiponectin in adipocyte-EVs (Figure 2). Furthermore some immunomodulatory proteins like tumor necrosis factor alpha (TNF-α), macrophage-colony-stimulating factor (MCSF) and retinol binding protein 4 (RBP-4) were present, while macrophage migration inhibitory factor (MIF) was clearly enriched in adipocyte-EVs (Figure 2).

EVs isolated from human in vitro differentiated adipocytes induce differentiation of monocytes into macrophages with an ATM-phenotype.

To evaluate the effect of adipocyte-derived EVs on macrophage phenotype, primary human monocytes were cultured in the absence or presence of different amounts of in vitro differentiated adipocyte-EVs. The resulting macrophage gene expression profile showed up-regulation of pro-inflammatory (Il-6, TNF-α and monocyte chemoattractant protein 1 (MCP-1); Figure 3A) as well as anti-inflammatory genes (alternative activated macrophage associated CC-chemokine (AMAC-1) and mannose receptor (MR); Figure 3B). Furthermore, CD163 and CD209 were up-regulated upon adipocyte-EV stimulation (Figure 2C), while IL10, CD36 and αVβ5ITG levels were unaltered (Figure 3B,S1). The observed mixed pro- and anti-inflammatory phenotype was confirmed at the protein level, as macrophage inflammatory protein-1-alpha (MIP-1α), TNF-α, IL-6 and IL-10 release by macrophages increased upon incubation with adipocyte-EVs (Figure 3D). The mixed pro- and anti-inflammatory phenotype induced by adipocyte-EVs is characteristic for human ATMs (12,13).

EVs isolated from human AT explants ex vivo also stimulate differentiation of monocytes into ATM. We next studied the effects of EVs derived from ex vivo human AT-explants on monocyte differentiation. We analyzed both subcutaneous (SCAT) and omental AT (OAT)-derived EVs, as these tissues differently associate with metabolic disorders (26). Both SCAT and OATEVs induced a gene expression pattern (Figure 4A-C) and secretion profile (Figure 4E) that was largely similar to the phenotype of matured macrophages stimulated by adipocyte-EVs (Figure 3). However, both SCAT and OAT-EVs upregulated CD51 and down-regulated αVβ5ITG and CD36 (Figure 4C,D), which were unaffected by adipocyte-EVs (Figure S1).

Adipocyte-derived and non-adipocyte derived EVs from ex vivo AT explants both induce different ATM markers.

While both in vitro differentiated adipocyte-EVs and ex vivo AT-explant EVs induced a similar gene expression profile in macrophages, AT-EVs also altered several genes (CD51, CD36, αVβ5ITG) which were unaffected by adipocyte-EVs. These findings suggested the presence and functional effect of EVs from distinct cellular origins within the AT EV pool. Based on flow cytometry analysis, the AT-derived EV pools could be subdivided into FSC-low and FSChigh populations (Figure 5A). FSC-low EVs mainly equilibrated at densities 1.08-1.14 g/ml while the FSC-high EVs mainly equilibrated at higher densities (1.20-1.26 g/ml; Figure 5B).

75

4

CHAPTER 4

B fold induction

30 20 10

15 10 5

5

3 2 1

6 4 2 0

4 3 2 1 0

2.0

MCP-1 fold induction

4 3 2 1 0

1x

PBS

CD51

**

1.5 1.0 0.5 0.0

5

IL-10

1.5 1.0 0.5 0.0

10x

CD209

8

2.0

**

MR

fold induction

TNFα

0

fold induction

10

**

0

4

5

AMAC-1

20

fold induction

0 5

fold induction

25

*

IL-6

40

fold induction

fold induction

50

C fold induction

A

1x

PBS

CD163

4 3 2 1 0

10x

1x

PBS

Adipocyte EVs

Adipocyte EVs

10x

Adipocyte EVs

D concentration (pg/ml)

5000

MIP1 α

4000

*

3000

**

100

1500

TNF

40

1000

20

0

0

1x

10x

Adipocyte EVs

**

IL10

300

1000

60

2000

PBS

400

IL6

80

200 500

100 0

0

PBS

1x

10x

Adipocyte EVs

PBS

1x

10x

Adipocyte EVs

PBS

1x

10x

Adipocyte EVs

Figure 3. Vesicles isolated from in vitro differentiated adipocytes induce differentiation of macrophages with an ATM-phenotype. (A) mRNA levels of pro-inflammatory macrophage markers IL-6, TNF-α, MCP1, (B) anti-inflammatory macrophage markers AMAC-1, MR, IL-10 and (C) adipose tissue macrophage (ATM) markers: CD51, CD209, and CD163; in resting macrophages (RM, white bar) or macrophages in the presence of SGBS-adipocyte vesicles in two concentrations (first and second black bar resp.) were measured by real-time PCR with 36B4 acting as the housekeeping gene. Each bar shows the mean value ± SEM of three experiments regarding different monocyte donors with two replicates. (D) MIP-1α, TNF-α, IL-6 and IL-10 levels were quantified by luminex in macrophage supernatants. Each bar shows the mean value ± SEM of three experiments regarding different monocyte donors with two replicates. One way ANOVA with Dunnets post hoc test was used to compare means of A vesicles versus resting macrophages, and significant differences are indicated (*P < 0.05, ** P < 0.01).

76

Human Adipocyte EVs and Macrophages

B **

60

**

40

20

TNF-α

*

*

3 2

Fold induction

*

2

1

**

4

2

0

PBS

SCAT EVs

OAT EVs

D

***

***

20 10

CD51

3

2

1

4

4

1.5

1.0

0.5

0.0

PBS

SCAT EVs

CD163

3

2

1

0

OAT EVs

PBS

SCAT EVs

OAT EVs

E 1.5

αVβ5ITG 8000

1.0

** 0.5

***

MIP-1α

***

***

6000 4000 2000

2000

1500

1.0

0.5

***

***

0.0

PBS SCAT EVs

OAT EVs

1000

TNF-α

***

1500

***

1000

500

2000

*** **

***

500

0

PBS SCAT EVs

OAT EVs

Concentration (pg/ml)

CD36

Concentration (pg/ml)

1.5

IL-6

0

0.0

Fold induction

**

0

IL-10 Fold induction

6

2.0

Fold induction

**

Concentration (pg/ml)

Fold induction

*

0

MCP-1

**

30

4

3

1

CD209

40

0

MR

Concentration (pg/ml)

Fold induction

0 4

0

Fold induction

*

100 50

4

8

50

**

150

0

5

AMAC-1

200

Fold induction

IL-6

C

250

Fold induction

Fold induction

80

Fold induction

A

IL-10

***

***

1500 1000 500 0

PBS SCAT EVs

OAT EVs

Figure 4. EVs isolated from human AT explants ex vivo also stimulate differentiation of monocytes into ATM. (A) mRNA levels of pro-inflammatory macrophage markers IL-6, TNF-α, MCP-1, (B) anti-inflammatory macrophage markers AMAC-1, MR, IL-10 and (C) adipose tissue macrophage (ATM) markers: CD51, CD209, CD163 and (D) ATM markers unaffected by A EVs: αVβ5integrin and CD36; in resting macrophages (RM, white bar) or macrophages in the presence of SCAT or OAT-EVs (first and second black bar resp.) were measured by real-time PCR with 36B4 acting as the housekeeping gene. Each bar shows the mean value ± SEM of three experiments regarding EVs from three different AT donors with two replicates. (E) MIP1α, TNF-α, IL-6 and IL-10 levels, quantified by luminex in macrophage supernatants. Each bar shows the mean value ± SEM of three experiments regarding EVs from three different AT donors with two replicates.

One way ANOVA with Dunnets post hoc test was used to compare means of A vesicles versus resting macrophages, and significant differences are indicated (*P < 0.05, ** P < 0.01).

77

CHAPTER 4

Further immunoblot analysis revealed that the adipocyte-specific marker adiponectin was mainly present in EVs floating at densities 1.20-1.25 g/ml, corresponding to the FSC-high EVs (Figure 5C). In agreement with other studies, a density difference was observed between the ex vivo AT explant EVs and the in vitro differentiated adipocyte-EVs (Figure 1C,D (27).

A

SCAT EVs PKH67 fluorescence

PKH67 fluorescence

OAT EVs

FSC low

FSC high

FSC low

Reduced wide angle FSC

Reduced wide angle FSC

B

FSC low EVs

FSC high EVs SCAT EVs

80 60 40 20 0

1,08 1,10 1,12 1,14 1,16 1,18 1,20 1,22 1,24 1,25

Density (g/ml)

C Density (g /ml )

100

% of total vesicles

% of total vesicles

100

FSC high

OAT EVs

80 60 40 20 0

1,08 1,10 1,12 1,14 1,16 1,18 1,20 1,22 1,24 1,25

Density (g/ml)

SCAT EVs

OAT EVs

1.08 1.10 1.12 1.14 1.16 1.18 1.20 1.22 1.24 1.25

1.08 1.10 1.12 1.14 1.16 1.18 1.20 1.22 1.24 1.25

CD9 ADIPOQ

Figure 5. EVs isolated from AT explants ex vivo are segregated into two EV populations. (A) Dot plots of reduced wide angle-FSC versus PKH67 fluorescence representing PKH67 labeled SCAT-EVs (left panel) or OAT-EVs (right panel). (B) Gates were set around EV subpopulations (FSC-low or FSC-high) to quantify the number of EVs in each subpopulation detected in each gradient fraction. The bottom two gradient fractions (1.26 and 1.27 g/ml) were left out of the analysis. Indicated are the means Âą SEM of FSC-low EVs (grey bars) or FSC-high EVs (black bars) derived from 6 different SCAT or OAT donors, presented as percentage EVs per total number of EVs in each gradient fraction. (C) Representative immunoblot analysis for detection of CD9 and adiponectin are shown of SCAT and OAT-EVs.

78

Human Adipocyte EVs and Macrophages

A

B Fold induction Fold induction

200

100

*

50

*** Fold induction

Fold induction

**

1

2.0

2 1

SCAT EVs

0.6

4

CD36

*** *

1.0

0.5

2.0

CD51

***

1.5

1.5 1.0 0.5

PBS

OAT EVs

***

0.0

MR

0.0

PBS

*** *

1.2

1.5

0

3

0

*** **

2

0

MCP-1

α Vβ5ITG

0.0

CD163

Fold induction

Fold induction

Fold induction

4

*

3

4

1.8

0

5

5

100

0

10

10

0

Fold induction

Fold induction

150

*

TNF-α

*** *

15

AMAC-1

Fold induction

IL-6

15

CD209

20

FSC low EVs FSC high EVs

300

25

0.5 0.0

OAT EVs

SCAT EVs

***

1.0

PBS

SCAT EVs

OAT EVs

C Concentration (pg/ml)

1000

MIP-1α

20

800 600 400 200 0

TNF-α

*

600

10

400

5

200

SCAT EVs

OAT EVs

IL-6

*

PBS

SCAT EVs

OAT EVs

1500

IL-10

***

1000

500

0

0

PBS

800

15

0

PBS

SCAT EVs

OAT EVs

PBS

SCAT EVs

OAT EVs

Figure 6. Adipocyte-derived and non-adipocyte derived EVs from ex vivo AT explants both induce different ATM markers. (A) mRNA levels of IL-6, TNF-α, MCP-1, AMAC-1, CD163 and MR and (B) CD209, αVβ5integrin, CD36 and CD51 were measured by real-time PCR with 36B4 acting as the housekeeping gene. White bar = resting macrophages (RM), grey bar = macrophages in the presence of FSC-low EVs of SCAT or OAT-EVs, black bar = macrophages in the presence of FSC-high EVs of SCAT or OAT-EVs. Each bar shows the mean value ± SEM of one experiment using EVs from three different AT donors with two replicates. (C) MIP-1α, TNF-α, IL-6 and IL-10 levels quantified by luminex in macrophage supernatants. Each bar shows the mean value ± SEM of three experiments regarding EVs from three different AT donors with two replicates. One way ANOVA with Bonferroni post hoc test was used to compare means of FSC high versus FSC low vesicles, and significant differences are indicated (*P < 0.05, ** P < 0.01).

79

CHAPTER 4

A 0.3

-

Insulin

PBS

-

+

+

OAT EVs

SCAT EVs

+

+

+

+

+

pAkt / Akt ratio

PBS

MCM

+

pAkt s473 Akt

0.2

**

0.1

***

0.0

MCM Insulin

B

C Number of EVs per gram AT

2.4×10

9

1.6×10 9

8.0×10 8

0

SCAT

D Number of EVs per gram OAT

4.0×10 9

1.0×10 9

-1.0×10 9

HOMA-IR

SCAT EVs OAT EVs

+

+

*

2.4×10 9

1.6×10 9

8.0×10 8

1.5×10 9

2.0×10 9

10

+

HOMA-IR < 2.59 HOMA-IR > 2.59

3.2×10 9

E

r: 0.62; p = 0.042

5

PBS

-

SCAT

3.0×10 9

0

PBS

0

OAT

15

Number of EVs per gram SCAT

Number of EVs per gram AT

3.2×10 9

***

OAT

r: -0.27 ; p = 0.43

1.0×10 9

5.0×10 8

0

-5.0×10 8

5

10

15

HOMA-IR

Figure 7. AT-EVs are quantitatively and qualitatively related to insulin resistance. (A) Effect of 50% macrophage conditioned medium (MCM) from macrophages stimulated with vehicle (PBS) or with SCAT or OAT vesicles (both from three patients) on insulin stimulated (20 nM) phosphorylation of Akt in SGBS adipocytes, analyzed by Western blot. Shown is a representative western blot, quantified by densitometric analysis presented as bar graph. One way ANOVA with Dunnets post hoc test was used to compare means of EV-treated MCM versus vehicle treated MCM. Significant differences are indicated (**P < 0.01, *** P < 0.001 versus MCM PBS+ insulin). (B) EVs from SCAT or OAT were quantified by flow cytometry analysis as described in Figure 1. Indicated are the medians with interquartile range of the number of EVs released by SCAT or OAT-EVs per gram AT of 11 patients. (C) Number of EVs derived from SCAT or OAT-EVs in patients stratified by HOMA-IR (insulin sensitive (IS; HOMA-IR < 2.59, n=5),insulin resistant (IR; HOMA-IR > 2.59, n=6) Kruskal-Wallis ANOVA with Dunn’s post hoc test was used to compare medians of EVs from IS versus IR patients, and significant differences are indicated (*P < 0.05). (E) and (F) Pearson correlation coefficients (r) and multiple regression analysis were calculated for the number of EVs per gram OAT (C) or SCAT-EVs (D) with donors’ HOMA-IR (n=11). Confidence intervals of 95% are indicated by the dashed arcs.

80

Human Adipocyte EVs and Macrophages

1

4

Insulin Receptor

4

2

3

M1

M2

IL-6 TNF-α MIP-1α

IL-10 IL1-RA

Figure 8. Adipose tissue-EVs in paracrine signaling between adipocytes and macrophages.

Adipocytes secrete extracellular vesicles (1), which have immunomodulatory effects on monocytes (2). Adipocyte-derived EVs differentiate monocytes into macrophages with a pro-inflammatory (M1) and anti-inflammatory (M2) phenotype (3), with reciprocal effects on insulin signaling in adipocytes (4).

To further explore the differential attribution of distinct vesicle populations on macrophage phenotype, the effect of separate EV populations on monocytes was studied. As the FSChigh EV pool was enriched for the adipocyte marker adiponectin, the majority of FSC-high EVs appeared to be from adipocyte origin, while the adiponectin-negative FSC-low pool presumably represents a collection of EVs from non-adipocyte cells. Monocytes were cocultured with pooled EVs from densities 1.20-1.26 g/ml (majority FSC-high) or with pooled EVs from densities 1.08-1.14 g/ml (majority FSC-low). The FSC-high, adiponectin-positive EV pool induced higher mRNA expression of IL-6, TNF-α, MCP-1, AMAC-1, CD163 and MR in macrophages compared to the adiponectin-negative EVs (Figure 6A). Interestingly, the stronger effect of FSC-high EVs appeared more pronounced for OAT than for SCAT-EVs. These differences were confirmed on protein level (Figure 6C). The adiponectin-negative EV pool induced higher mRNA expression of CD209, αVβ5ITG and CD51 compared to the adiponectin-positive EV pool (Figure 6B). However, this only holds true for the SCAT-derived EVs, while the OAT-derived EVs downregulated αVβ5ITG, CD36 and CD51 (Figure 6B).

Together, these data suggest that the immunomodulatory effects of adipose tissue EVs on macrophages observed in Figure 4 were mainly caused by adipocyte-derived EVs.

81

CHAPTER 4

AT-EVs are quantitatively and qualitatively related to insulin resistance It is known that ATMs are a major source of pro-inflammatory cytokines, which in a paracrine fashion can induce insulin resistance in adipocytes (28). As both SCAT and OAT-derived EVs differentiated monocytes into macrophages with an ATMâ&#x20AC;&#x201C;phenotype (Figure 4), we wondered whether AT-EV-induced macrophage inflammation would actually have an effect on insulin signaling in human adipocytes. We stimulated differentiated SGBS-adipocytes with 50% conditioned medium of macrophages which were pre-treated with equal numbers of SCAT or OAT-EVs (Figure 4), and determined the effect on insulin stimulated phosphorylation of Akt, a central player in insulin signaling (29). We observed a robust inhibition of Akt-phosphorylation by both SCAT- and OAT-EV treated macrophage CM, compared to vehicle treatment, in which the effect of OAT-EV stimulated macrophage secretome was slightly stronger than that of SCAT-EV macrophage secretome (Figure 7A).

Having established that both SCAT and OAT-derived EVs differentiated monocytes into macrophages with an ATMâ&#x20AC;&#x201C;phenotype (Figure 4), and that these ATMs secreted factors that inhibit insulin signaling in adipocytes in a paracrine fashion (Figure 7A), we next questioned whether the number of AT-EVs might be also related to systemic insulin resistance. AT-EVs were quantified by flow cytometry, and determined the number of EVs per gram fat. In 11 patients, the median number of EVs was not different between SCAT or OAT, however we observed a wide distribution in the number of OAT-EVs opposed to SCAT-EVs (Figure 7B).

This difference in number of OAT-EVs appeared to be based on a difference in insulin resistance, measured by HOMA-IR (Figure 7C) and correlated positively with HOMA-IR (Pearson r0.62; p=0.042; Figure 7D). In contrast, the number of SCAT-EVs did not (Pearson r-0.27; p=0.43; Figure 7E). Subject characteristics revealed that the BMI and waist circumference were also different amongst IR and IS subjects (Table S1), suggesting that the observed effect might be driven by a higher amount of AT in the IR subjects. However, the number of SCAT or OAT-EVs did not correlate with BMI (Pearson r0.01, p=0.98 and r0.16, p=0.65 respectively) or hsCRP levels (Pearson r0.03, p=0.92 and r0.24, p= 0.48 respectively).

DISCUSSION In the present study, we demonstrated that (i) AT-EVs from in vitro differentiated human adipocytes contain immunomodulatory proteins, (ii) human adipocyte- and human AT-EVs have immunomodulatory effects on macrophages, and (iii) macrophages pre-treated with human AT-EVs secreted factors that interfered with insulin signaling in human adipocytes (Figure 8). Furthermore, we observed a specific increase of OAT-EVs in overweight patients, associated with systemic insulin resistance. Together, these data suggest a possible role for human AT-EVs in reciprocal signaling between adipocytes and macrophages, with the potential to aggravate local and systemic insulin resistance.

82

Human Adipocyte EVs and Macrophages

So far, the role of EVs in AT-inflammation has been studied with rodent adipocyte-EVs (30), or rodent AT-EVs (16,17). However, processes related to AT-homeostasis and AT-dysfunction differ between rodents and humans (31), hindering extrapolation to human pathophysiology.

We are the first to show that also human adipocytes, both in vitro and ex vivo, produce EVs with immunomodulatory properties, which can contribute to the development of local insulin resistance, a key element in the mechanistic link between obesity and adverse metabolic complications (32).

We focused on the interaction between EVs and monocytes as the influx of monocytes, differentiation into macrophages and subsequent release of pro-inflammatory cytokines are crucial steps in AT-inflammation (28). Furthermore, Deng et al showed that AT-EVs injected into mice were taken up mainly by monocytes in contrast to other immune cells (17). Furthermore, though T cells also serve important roles in AT-inflammation (33), we observed no effect of human adipocyte-EVs on na誰ve T-cells (data not shown), in contrast to their effect on monocytes. While many other immune cells also play a role in ATinflammation (10), macrophages are still key players in that process and, based on the aforementioned findings, probably most responsive to adipocyte-EVs as well. While the effect of adipocyte-EVs on other immune cells might be limited, AT is composed of many different cell types, which are all capable of secreting functional EVs. The FSC-low EVs, putatively from non-adipocyte origin, were capable of inducing other macrophage genes than adipocyte-EVs (Figure 6), indicating that non-adipocyte-EVs may also contribute to alterations in macrophage phenotype. Our observation that monocytes stimulated with adipocyte-EVs differentiated into ATM suggests an important role for adipocyte-EVs in AT-inflammation, which is in line with previous studies. Deng et al. demonstrated that ATEVs from obese mice could activate monocytes (17), however that was not narrowed down to adipocyte-EVs specifically. We do provide proof of principle that indeed adipocytes secrete functional EVs, which can influence ATM differentiation.

Although the mechanisms need to be established, a number of adipokines present on adipocyte-EVs have previously been shown to play a role in either pro- or anti-inflammatory macrophage differentiation. Deng et al. previously showed that macrophage differentiation

by AT-EVs was at least partly due to EV-associated RBP4 (17). Human adipocyte-EVs contain RBP4 as well (Figure 2), leaving the possibility that this is one of the factors contributing to macrophage differentiation. Second, MIF, an adipokine with chemotactic properties and regulator of macrophage accumulation, is secreted mainly via EVs (Figure 2) and acts intracellularly when encountering the target cell (34,35). Last, soluble adiponectin was shown to differentiate monocytes into anti-inflammatory macrophages (36). However, whether adiponectin is located on the outside or inside of adipocyte-EVs is currently unknown. Therefore, EV-associated adiponectin can serve as a marker for adipocyte-EVs, but its specific biological relevance remains to be established. Nevertheless, activation of both the pro- and anti-inflammatory macrophage machinery suggests the influence of multiple factors. Moreover, besides proteins, RNAs as well as lipids have been shown to play a role in the effect of adipocyte-EVs on target cells (16,30). Therefore, it is more likely that a combination of compounds is responsible for the ATM phenotype, underscoring the

83

4

CHAPTER 4

importance of EVs harboring different sets of compounds as opposed to single factors.

It is known that macrophage driven AT-inflammation decreases insulin signaling in the adipocyte, resulting in increased lipolysis and release of free fatty acids, consequently leading to systemic IR (37,38). Mice studies showed that obesity-related IR is initiated in AT, partly by macrophage derived IL-6, TNF-α and MIP-1α (39). However, the role of human EVs in this paracrine signaling between adipocytes and macrophages has not been shown before. As we have shown that adipocyte-EVs stimulate macrophages to secrete IL-6, TNF-α and MIP-1α, and that EV-stimulated macrophage supernatant subsequently impaired insulin signaling in adipocytes, we postulate that AT-EVs might play a role in the development of local IR via macrophage activation, which might further elucidate the mechanisms by which AT-inflammation leads to insulin resistance. Furthermore, our finding that OAT-EVs were increased in IR subjects and positively correlated with IR suggests a possible link between OAT-EVs and systemic IR. Interestingly, adipocyte-EVs contain RBP4, MIF and adiponectin, three factors that are associated with IR as well (9,40).

Together, our data suggest a possible role for human AT-EVs in in a reciprocal proinflammatory loop between adipocytes and macrophages, with the potential to aggravate local and systemic insulin resistance.

Acknowledgements The authors thank members of the Kalkhoven laboratory and Department of Vascular Medicine for helpful discussions, we thank D.R. Faber, I. M. Schrover and S.N. Verhagen from the Department of Vascular Medicine for subject recruitment and study design and we thank W.A. Scheper from the Department of Hematology and Immunology for help with the macrophage assays. We further thank G.J.A Arkesteijn for technical assistance with the flow cytometry measurements, and members of the Wauben laboratory for assistance with analyzing the flow cytometry data.

84

Human Adipocyte EVs and Macrophages

REFERENCES 1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

Finucane MM, Stevens GA, Cowan MJ, Danaei G, Lin JK, Paciorek CJ et al. National, regional, and global trends in body-mass index since 1980: systematic analysis of health examination surveys and epidemiological studies with 960 country-years and 9.1 million participants. Lancet 2011; 377: 557-567.

Yusuf S, Hawken S, Ounpuu S, Bautista L, Franzosi MG, Commerford P et al. Obesity and the risk of myocardial infarction in 27,000 participants from 52 countries: a case-control study.

Lancet 2005; 366: 1640-1649.

Ford ES, Giles WH and Dietz WH. Prevalence of the metabolic syndrome among US adults: findings from the third National Health and Nutrition Examination Survey. JAMA 2002; 287: 356-359.

Kanhai DA, Kappelle LJ, van der Graaf Y, Uiterwaal CS, Visseren FL and SMART Study Group.

The risk of general and abdominal adiposity in the occurrence of new vascular events and mortality in patients with various manifestations of vascular disease. Int J Obes (Lond) 2012; 36: 695-702.

Weisberg SP, McCann D, Desai M, Rosenbaum M, Leibel RL and Ferrante AW,Jr. Obesity is associated with macrophage accumulation in adipose tissue. J Clin Invest 2003; 112: 17961808.

Harte AL, Tripathi G, Piya MK, Barber TM, Clapham JC, Al-Daghri N et al. NFkappaB as a potent regulator of inflammation in human adipose tissue, influenced by depot, adiposity, T2DM status, and TNFalpha. Obesity (Silver Spring) 2013; .

Suganami T, Nishida J and Ogawa Y. A paracrine loop between adipocytes and macrophages aggravates inflammatory changes: role of free fatty acids and tumor necrosis factor alpha.

Arterioscler Thromb Vasc Biol 2005; 25: 2062-2068.

Keophiphath M, Rouault C, Divoux A, Clement K and Lacasa D. CCL5 promotes macrophage recruitment and survival in human adipose tissue. Arterioscler Thromb Vasc Biol 2010; 30: 39-45.

Norseen J, Hosooka T, Hammarstedt A, Yore MM, Kant S, Aryal P et al. Retinol-binding protein 4 inhibits insulin signaling in adipocytes by inducing proinflammatory cytokines in macrophages through a c-Jun N-terminal kinase- and toll-like receptor 4-dependent and retinol-independent mechanism. Mol Cell Biol 2012; 32: 2010-2019.

Schipper HS, Prakken B, Kalkhoven E and Boes M. Adipose tissue-resident immune cells: key players in immunometabolism. Trends Endocrinol Metab 2012; 23: 407-415.

Murray PJ and Wynn TA. Protective and pathogenic functions of macrophage subsets. Nat Rev Immunol 2011; 11: 723-737.

Zeyda M, Farmer D, Todoric J, Aszmann O, Speiser M, Gyori G et al. Human adipose tissue macrophages are of an anti-inflammatory phenotype but capable of excessive pro-inflammatory mediator production. Int J Obes (Lond) 2007; 31: 1420-1428.

Dalmas E, Clement K and Guerre-Millo M. Defining macrophage phenotype and function in adipose tissue. Trends Immunol 2011; 32: 307-314.

Thery C, Ostrowski M and Segura E. Membrane vesicles as conveyors of immune responses.

Nat Rev Immunol 2009; 9: 581-593.

Buschow SI, Nolte-'t Hoen EN, van Niel G, Pols MS, ten Broeke T, Lauwen M et al. MHC II in dendritic cells is targeted to lysosomes or T cell-induced exosomes via distinct multivesicular body pathways. Traffic 2009; 10: 1528-1542.

Muller G, Jung C, Straub J, Wied S and Kramer W. Induced release of membrane vesicles from rat adipocytes containing glycosylphosphatidylinositol-anchored microdomain and lipid droplet signalling proteins. Cell Signal 2009; 21: 324-338.

Deng ZB, Poliakov A, Hardy RW, Clements R, Liu C, Liu Y et al. Adipose tissue exosome-like vesicles mediate activation of macrophage-induced insulin resistance. Diabetes 2009; 58: 2498-2505.

Wabitsch M, Brenner RE, Melzner I, Braun M, Moller P, Heinze E et al. Characterization of a human preadipocyte cell strain with high capacity for adipose differentiation. Int J Obes Relat Metab Disord 2001; 25: 8-15.

85

4

CHAPTER 4

19. Thery C, Amigorena S, Raposo G and Clayton A. Isolation and characterization of exosomes from cell culture supernatants and biological fluids. Curr Protoc Cell Biol 2006; Chapter 3: Unit 3.22.

20. de Jong, O., Verhaar, M., Chen, Y., Vader, P., Gremmels, H., Posthuma, G., Schiffelers, R., Gucek, M., van Balkom,B. Cellular stress conditions are reflected in the protein and RNA content of endothelial cell-derived exosomes. Journal of Extracellular Vesicles 2012; 1: .

21. Jeninga EH, van Beekum O, van Dijk AD, Hamers N, Hendriks-Stegeman BI, Bonvin AM et al.

Impaired peroxisome proliferator-activated receptor gamma function through mutation of a conserved salt bridge (R425C) in familial partial lipodystrophy. Mol Endocrinol 2007; 21: 1049-1065.

22. Slot JW and Geuze HJ. Cryosectioning and immunolabeling. Nat Protoc 2007; 2: 2480-2491.

23. van der Vlist EJ, Nolte-'t Hoen EN, Stoorvogel W, Arkesteijn GJ and Wauben MH. Fluorescent labeling of nano-sized vesicles released by cells and subsequent quantitative and qualitative analysis by high-resolution flow cytometry. Nat Protoc 2012; 7: 1311-1326.

24. Schipper HS, de Jager W, van Dijk ME, Meerding J, Zelissen PM, Adan RA et al. A multiplex immunoassay for human adipokine profiling. Clin Chem 2010; 56: 1320-1328.

25. Scherer PE, Williams S, Fogliano M, Baldini G and Lodish HF. A novel serum protein similar to C1q, produced exclusively in adipocytes. J Biol Chem 1995; 270: 26746-26749.

26. Preis SR, Massaro JM, Robins SJ, Hoffmann U, Vasan RS, Irlbeck T et al. Abdominal subcutaneous and visceral adipose tissue and insulin resistance in the Framingham heart study.

Obesity (Silver Spring) 2010; 18: 2191-2198.

27. Caby MP, Lankar D, Vincendeau-Scherrer C, Raposo G and Bonnerot C. Exosomal-like vesicles are present in human blood plasma. Int Immunol 2005; 17: 879-887.

28. Xu H, Barnes GT, Yang Q, Tan G, Yang D, Chou CJ et al. Chronic inflammation in fat plays a crucial role in the development of obesity-related insulin resistance. J Clin Invest 2003; 112: 1821-1830.

29. Manning BD and Cantley LC. AKT/PKB signaling: navigating downstream. Cell 2007; 129: 12611274.

30. Ogawa R, Tanaka C, Sato M, Nagasaki H, Sugimura K, Okumura K et al. Adipocyte-derived microvesicles contain RNA that is transported into macrophages and might be secreted into blood circulation. Biochem Biophys Res Commun 2010; 398: 723-729.

31. Muniyappa R, Chen H, Muzumdar RH, Einstein FH, Yan X, Yue LQ et al. Comparison between surrogate indexes of insulin sensitivity/resistance and hyperinsulinemic euglycemic clamp estimates in rats. Am J Physiol Endocrinol Metab 2009; 297: E1023-9.

32. Dandona P, Aljada A and Bandyopadhyay A. Inflammation: the link between insulin resistance, obesity and diabetes. Trends Immunol 2004; 25: 4-7.

33. Nishimura S, Manabe I, Nagasaki M, Eto K, Yamashita H, Ohsugi M et al. CD8+ effector T cells contribute to macrophage recruitment and adipose tissue inflammation in obesity. Nat Med 2009; 15: 914-920.

34. Kleemann R, Hausser A, Geiger G, Mischke R, Burger-Kentischer A, Flieger O et al. Intracellular action of the cytokine MIF to modulate AP-1 activity and the cell cycle through Jab1. Nature 2000; 408: 211-216.

35. Bernhagen J, Calandra T and Bucala R. Regulation of the immune response by macrophage migration inhibitory factor: biological and structural features. J Mol Med (Berl) 1998; 76: 151-161.

36. Lovren F, Pan Y, Quan A, Szmitko PE, Singh KK, Shukla PC et al. Adiponectin primes human monocytes into alternative anti-inflammatory M2 macrophages. Am J Physiol Heart Circ Physiol 2010; 299: H656-63.

37. Olefsky JM and Glass CK. Macrophages, inflammation, and insulin resistance. Annu Rev Physiol 2010; 72: 219-246.

38. Solinas G, Vilcu C, Neels JG, Bandyopadhyay GK, Luo JL, Naugler W et al. JNK1 in hematopoietically derived cells contributes to diet-induced inflammation and insulin resistance without affecting obesity. Cell Metab 2007; 6: 386-397.

39. Hotamisligil GS, Shargill NS and Spiegelman BM. Adipose expression of tumor necrosis factoralpha: direct role in obesity-linked insulin resistance. Science 1993; 259: 87-91.

40. Koska J, Stefan N, Dubois S, Trinidad C, Considine RV, Funahashi T et al. mRNA concentrations of MIF in subcutaneous abdominal adipose cells are associated with adipocyte size and insulin action. Int J Obes (Lond) 2009; 33: 842-850.

86

SUPPLEMENTARY INFORMATION Supporting table 1. Subject characteristics.

Subject characteristics HOMA-IR < 2.59

HOMA-IR > 2.59

N

5

6

Age

62.6 ± 0.57

65.3 ± 6.3

Sex (men/women)

4/1

6/0

Body weight (kg)

77.1 ± 15.4

93.8 ± 16.8

BMI (kg/m2)

23.7 ± 3.9

27.8 ± 3.1

Waist circumference (cm)

96 ± 16

108 ± 8

Fasting glucose (mmol/L)

6.1 ± 1.2

6.5 ± 0.1

Fasting insulin (mIU/L)

6.0 (5.0 – 9.5)

19.5 (12.8 – 32.3)

HOMA-IR

1.8 (1.3 – 2.5)

5.8 (4.2 – 9.6)

Fasting hsCRP (mg/L)

2.9 (0.5 – 21.9)

7.2 (2.4 – 8.8)

4

Data are presented as means ± standard deviation or medians (interquartile range)

Supporting table 2. Primers used for qPCR analysis.

Gene

Forward primer

Reverse primer

36B4

CGGGAAGGCTGTGGTGCTG

GTGAACACAAAGCCCACATTCC

AMAC-1

AGCTCTGCTGCCTCGTCTAT

CCCACTTCTTATTGGGGTCA

MR

ACACCAAAACCTGAGCCAAC

CCACCCATCTTCAGTAACTGGT

IL10

AACCTGCCTAACATGCTTCGA

TGTCCAGCTGATCCTTCATTTG

CD163

AATGGGAATTTATAACCCAGTGAG

GGTGAATTTCTGCTCCATTCA

CD36

CCTCCTTGGCCTGATAGAAA

GTTTGTGCTTGAGCCAGGTT

TNFα

CAGAGGGCCTGTACCTCATC

GGAAGACCCCTCCCAGATAG

IL6

AGTGCCTCTTTGCTGCTTTCAC

TGACAAACAAATTCGGTACATCCT

MCP1

CAGCCAGATGCAATCAATGCC

TGGAATCCTGAACCCACTTCT

CD51

GCCGTGGATTTCTTCGTG

CCAGTCACATTTGAGGACCTG

CD209

TTCACCTGGATGGGACTTTC

GGGCTCTCCTCTGTTCCAAT

αVβ5Integrin

CACAGGAGATTGCCGTGAA

GGGAGAGGTCCATCAGGTAGT

87

CHAPTER 4

Fold induction

2.0

aVb5ITG

1.5 1.0 0.5 0.0

Fold induction

1.5

CD36

1.0

0.5

0.0

RM

1x

10x

Adipocyte vesicles

88

Supporting Figure 1. Effect of EVs isolated from in vitro differentiated adipocytes on αVβ5ITG and CD36. mRNA levels of αVβ5ITG and CD36 in resting macrophages (RM, white bar) or macrophages in the presence of SGBS-A EVs in two concentrations (first and second black bar resp.) were measured by real-time PCR with 36B4 acting as the housekeeping gene. Each bar shows the mean value ± SEM of three experiments regarding different monocyte donors with two replicates.

89

CHAPTER 5 Effect of extracellular vesicles of human adipose tissue on insulin signaling in liver and muscle cells In revision, Obesity

MariĂŤtte E.G. Kranendonk, Frank L.J. Visseren, Joost A. van Herwaarden, Esther N.M. Nolte-â&#x20AC;&#x2122;t Hoen, Wilco de Jager, Marca H.M. Wauben, Eric Kalkhoven

CHAPTER 5

ABSTRACT Objective Insulin resistance (IR) is a key mechanism in obesity-induced cardiovascular disease. To unravel mechanisms whereby human adipose tissue (AT) contributes to systemic IR, the effect of human AT-extracellular vesicles (EVs) on insulin signaling in liver and muscle cells was determined.

Design and Methods EVs released from human subcutaneous (SAT) and omental AT (OAT)-explants ex vivo were used for stimulation of hepatocytes and myotubes in vitro. Subsequently, insulin-induced Akt phosphorylation and expression of gluconeogenic genes (G6P, PEPCK) was determined.

AT-EV adipokine levels were measured by multiplex immunoassay, and AT-EVs were quantified by high-resolution flow cytometry.

Results In hepatocytes, AT-EVs from the majority of patients inhibited insulin-induced Akt phosphorylation, while EVs from some patients stimulated insulin-induced Akt phosphorylation. In myotubes AT-EVs exerted an ambiguous effect on insulin signaling.

Hepatic Akt phosphorylation related negatively to G6P-expression by both SAT-EVs (r=0.60,p=0.01) and OAT-EVs (r=-0.73,p=0.001). MCP-1, IL-6 and MIF concentrations were higher in OAT-EVs compared to SAT-EVs and differently related to lower Akt phosphorylation in hepatocytes. Finally, the number of OAT-EVs correlated positively with liver enzymes indicative for liver dysfunction.

Conclusions Human AT-EVs can stimulate or inhibit insulin signaling in hepatocytes- possibly depending on their adipokine content- and may thereby contribute to systemic IR.

92

Adipose vesicles and insulin signaling

INTRODUCTION The worldwide prevalence of obesity has doubled over the last 20 years (1). Obesity often engenders insulin resistance and type 2 diabetes, conditions that substantially increase the risk for cardiovascular disease (2,3). Mechanistic studies suggest that insulin resistance is a key process in obesity-induced cardiovascular disease (4,5). In the insulin-resistant state, tissues such as liver, fat and muscle show a reduced response to actions of the hormone insulin. This results in a range of effects, for example impaired regulation of gluconeogenesis in the liver and impaired glucose uptake in skeletal muscle (6).

In obesity, distinct adipose tissue (AT) depots contribute differently to cardiometabolic risk.

Visceral AT (VAT) rather than subcutaneous AT (SAT) is most strongly related to insulin resistance and cardiovascular disease (7,8). The molecular mechanisms underlying the difference in disease contribution between VAT and SAT are not fully understood. The underlying pathophysiology is likely to be based on inflammation of adipose tissue (AT) as a consequence of chronic energy excess, which is more pronounced in VAT compared to SAT (9).

As a consequence of AT-inflammation, several mediators can contribute to systemic insulin resistance. First of all, dysfunctional lipid metabolism results in increased levels of systemic free fatty acids (FFA) (10). Via stimulation of toll-like receptors, FFA can activate the innate immune signaling pathway in muscle and liver cells, resulting in lipid-induced insulin resistance by accumulation of diacylglycerols (DAG) and ceramides (11,12). Secondly, ATinflammation is characterized by increased influx of immune cells, which contribute to disproportionate pro-inflammatory adipokine secretion, resulting in both local and systemic metabolic imbalance (13). Several adipokines secreted by AT are described as potent inducers of insulin resistance, including tumor necrosis factor alpha (TNFÎą) (14), retinolbinding protein-4 (RBP-4) (15), resistin and macrophage migration inhibitory factor (MIF) (16). Also a decrease in the plasma concentrations of adiponectin, an anti-inflammatory adipocyte-specific adipokine with insulin sensitizing properties, is related to insulin resistance (17,18).

However, besides FFA and soluble adipokines, AT of obese mice also secretes extracellular vesicles (EVs), which are associated with systemic IR (19). We have recently shown that also human AT secretes EVs, which are able to effect local insulin resistance in adipocytes via macrophage activation, hence indicating a paracrine effect of AT-EVs (20). Furthermore, EVs derived from human visceral AT were associated with systemic IR (20). EVs harbor proteins, RNA and lipids reflecting the pathophysiological state of the donor cell, which are able to influence the function of target cells. Furthermore, EVs are stable in plasma, and can therefore interact with target cells at a distance from the donor tissue (21). Therefore, the unique composition and stability of EVs allows for transfer of signalling molecules to a wide variety of target tissues, indicating that AT-EVs could be important mediators in endocrine immune signaling (22).

In the present study we investigated putative endocrine effects of human AT-derived EVs and sought to determine whether they could directly influence insulin signaling in liver or skeletal muscle cells, two of the primary peripheral tissues contributing to systemic IR.

93

5

CHAPTER 5

RESEARCH DESIGN AND METHODS Subjects Abdominal subcutaneous and omental AT-explants were collected from subjects undergoing surgery for aneurysmatic aortic disease at the University Medical Center Utrecht (UMCU).

Patients with auto-immune diseases, use of insulin therapy and hypothyroidism (TSH>5.0 MU/l) were excluded. The study was approved by the Medical Ethics Committee of the UMCU and all study subjects gave written informed consent. Anthropometric and biochemical analysis were performed in the fasting state. The Adult Treatment Panel (ATP) III criteria were taken for the definition of the metabolic syndrome (23).

Ex vivo AT explants AT explants were cut into two pieces. One part was incubated in DMEMf12 supplemented with 50 IU/ml penicillin and 50 mg/ml streptomycin (PS) for 24 hours and weighed afterwards. Culture supernatant was centrifuged for 10 minutes at 500g to remove cells and stored at -80°C until EV isolation. The second part was fixed in 10% formaldehyde until further processing for immunohistochemistry.

AT characteristics The number of macrophages and adipocyte cell size was determined by immunohistochemistry. Fixed AT samples were embedded in paraffin and 4 μm sections were processed for histological staining with hematoxylin & eosin (H&E). Digital images of the H&E stained slides were acquired with Leica Qprodit imaging software (Leica Microsystems, Rijswijk, the Netherlands), after which the adipocyte size of 100 adipocytes was measured in randomly selected areas of the slide at 200x magnification. To detect macrophages a CD68 immunostain was used and capillary density was analyzed using a von Willebrand factor immunostain. Macrophage infiltration was scored by determining the mean number of CD68-positive cells in 10 high power fields (HPF; magnification 400x) per slide. Crown like structures (CLS) were defined as CD68-positive cells surrounding at least 50% of the circumference of an adipocyte. The number of CLS per AT depot was determined by counting all CLS at 100x magnification on the whole slide, which was corrected for total surface area using ImageJ Software (v1.45).

Isolation of EVs Plasma EVs were isolated from human plasma as described previously (24). Blood was collected from a healthy donor in Na-heparin tubes, centrifuged for 20 minutes at 900g and diluted 1:1 with PBS. Diluted plasma was further centrifuged for 30 minutes at 2000g, and 2 hours at 100,000g using a Beckman LE-80K centrifuge with SW32-Ti rotors (Beckman Instruments, Inc., Fullerton, CA, USA.). The 100,000 g pellet was resuspended in 50 ml PBS, filtered through a 0.22-µm filter (Steritop, Millipore) and ultracentrifuged again for 2 hours at 100,000g. The final pellet was resuspended in 2.5 M sucrose and subfractioned by sucrose gradient (SG)-ultracentrifugation. A linear SG (2.5–0.4 M sucrose, 20 mM Tris-

94

Adipose vesicles and insulin signaling

HCL pH7,4) was layered on top of EVs and ultracentrifuged for 16 hours at 190,000g in a SW60-Ti rotor.

AT-EVs were isolated from 24 hour supernatant of 1-2 gram of ex vivo AT-explants by differential steps of (ultra)centrifugation as described previously (24). In short, collected supernatant was centrifuged sequentially for 10 minutes at 500g, 15 minutes at 1500g, 30 minutes at 10,000g and 2 hours at 100,000g using a Beckman LE-80K centrifuge with SW60-Ti rotors (Beckman Instruments, Inc., Fullerton, CA, USA.). 100,000g pelleted EVs were washed in PBS and centrifuged again for 2 hours at 100,000g after which the EV pellets were resuspended in PBS and immediately used for functional experiments.

Immunoblotting plasma EVs EV SG-fractions were mixed 1:1 with lysis buffer (25). Equal amounts of lysed SG-fractions were subjected to SDS-PAGE and transferred to polyvinyl difluoride membranes (26).

Proteins were detected with Amersham ECL prime Western Blotting Detection reagent and imaged with the Imagequant Las 4000. Primary antibodies against CD9 (Santa-Cruz, sc-53679) and adiponectin (R&D-systems, AF1065) were used.

Cell culture HepG2 cells were maintained in DMEM-low glucose, supplemented with L-glutamine, heat inactivated fetal bovine serum (HI-FBS) and PS. C2C12 myoblasts were maintained in DMEM-high glucose, supplemented with HI-FBS and PS. For differentiation into myotubes, C2C12 cells were grown to confluence and medium was replaced with DMEM containing 2% horse serum and PS, changed daily. Myotubes were used for experiments on day 6 of differentiation.

Stimulation of cells with EVs and insulin After washing twice with PBS, cells were incubated in serum-free low glucose-DMEM supplemented with EVs, TNF-α 100 ng/ml or vehicle (PBS). The amount of EVs used to stimulate liver and muscle cells was obtained from supernatant derived from 0,1 gram of AT, e.g. 100,000g pelleted EVs derived from supernatant of 0,5 gram AT were resuspended in 50 µl PBS, of which 10 µl EVs was used for stimulation of HepG2 and 10 µl for stimulation of C2C12 cells. AT EVs from identical individuals were used for both liver and muscle cell stimulation, enabling equal comparison of effect of AT-EVs on insulin signaling between both tissues. After 24 hour incubation, cells intended for Western blot analysis were stimulated with insulin (20 nM) for 10 minutes, after which cells were washed with ice cold PBS and lysed in ice cold RIPA lysis buffer (50mM Tris HCL ph 7.4, 1mM EDTA, 150mM NACL, 1%NP40, 5mM NaF, 0.25% Na deoxycholate, 2mM NaVO3 and 1x protease inhibitors (Roche)).

HepG2 cells intended for semi-quantitative reverse transcriptase-polymerase chain reaction (RT-PCR) analyses were stimulated with insulin (20 NM) for 1 hour after which total RNA was isolated using TRIzol reagent (Invitrogen, Carsbad, USA).

95

5

CHAPTER 5

Immunoblotting HepG2 and C2C12 lysates Cell lysates were incubated on ice for 20 minutes and centrifuged for 5 min, after which supernatants (20 µl of HepG2 and 15 µl of C2C12 lysates) were subjected to immunoblot analysis to detect Akt phosphorylation. Images were taken with Imagequant Las 4000 and quantification of Akt and phosphorylated Akt bands was performed using ImageQuant TL Software. Antibodies were from Cell Signaling Technology (anti-Akt #9272; antiphosphorylated Akt-Ser473, #9271).

RNA extraction and analysis RNA isolation and cDNA synthesis were performed as previously described (26). Primers for quantitative RT-PCR were designed with the universal probe library (Roche), and are described in Table S1. Specificity of the amplification was verified by melt curve analysis and evaluation of efficiency of PCR amplification. The mRNA expression of genes reported was normalized to 36B4 expression.

Multiplex immunoassay Pelleted EVs derived subcutaneous and visceral AT were lysed in Roche cOmplete lysisM buffer, supplemented with cOmplete protease inhibitors (Roche). Of each sample, EVs derived from 0.1 gram AT were used for adipokine profiling. Samples were diluted 1:1 with PBS prior to measurement, to reduce some of the high background signals of the lysis buffer.

All multiplex immunoassays were carried out using the Bio-Plex system FlexMAP3D. Acquisition was performed with xPonent 4.2 and data analysis with Bio-PLex Manager Software version 6.1.1 (Biorad Laboratories, Hercules, CA) (27). Antibody pairs and recombinant proteins used were described earlier (28).

High-resolution flow cytometric analysis of EVs EVs were analyzed on the BD InfluxTM flow cytometer (Becton Dickinson, San-Jose, USA) as described previously (29). Briefly, 100,000g pelleted EVs were resuspended in 20 µl PBS with 0.1% BSA (cleared from aggregates by ultracentrifugation), labeled with PKH67 (7.5 µM; Sigma-Aldrich), and subfractioned by SG-ultracentrifugation overnight. One ml fractions were collected from the bottom of the tube and diluted 1:100 in PBS before measurement. Due to contamination of unbound PKH-67 dye, the two bottom fractions were not used for measurement of vesicles (29). Time-based quantitative measurements were performed. Data was acquired using Spigot software version 6.1 (Becton Dickinson) and analyzed using FCS Express software (De Novo Software). From the absolute number of EVs measured in 30 seconds, the number of EVs per gram fat was calculated based on the weight of the AT-biopsies.

Data analysis Statistical evaluation was performed using GraphPad Prism software, version 5.03 (GraphPad Software, La Jolla, CA) and SPSS 20.0 for Windows (SPSS, Inc., Chicago, IL).

96

Adipose vesicles and insulin signaling

Data are presented as mean ± standard deviation (SD) when normal distributed or as median with interquartile range (IQR) in case of skewed distribution. Basic descriptive statistics were used to describe the subject characteristics. Differences between > 2 experimental samples were evaluated by one way ANOVA with Dunnets post hoc test.

Differences between adipokine concentrations were analyzed by Wilcoxon matched-pairs signed rank test. Concentrations of adipokines were log transformed to fulfill regression criteria. Associations between two variables with normal distribution were measured by Pearson correlation, and variables with skewed distribution by Spearman correlation tests.

P-values <0.05 were considered significant.

RESULTS Patient characteristics The majority of study participants was male (n=15, 93%) and the mean age of all patients was 63.9±7.1 years. Three patients were obese (BMI ≥30 kg/m2), and five patients were overweight (BMI 25-30 kg/m2). The mean waist circumference of men was 102±10 cm.

The metabolic syndrome was present in 8 patients (50%) (Table 1). Three patients had type 2 diabetes, for which they were treated with the glucose-lowering drug metformin.

Adipocyte EVs in circulation To investigate whether adipocyte EVs are present in the circulation, sucrose gradient fractions of plasma EVs were analyzed by Western blot. A general EV marker CD9 (24) as well as the adipocyte-specific marker adiponectin (17) were clearly detectable on EVs isolated from human plasma (Figure 1). These data indicate that adipocyte EVs are present in plasma and may therefore be systemic communicators between adipose tissue and liver or skeletal muscle.

Density g/ml

1.04 1.06 1.08 1.10

1.12

1.14

1.15

1.16

1.18

1.20

1.22

1.23

Adiponectin CD9 Figure 1. Adipocyte EVs are detectable in the circulation. Immunoblot analysis of pelleted adipocyte EVs separated by sucrose density gradient (SG) ultracentrifugation. Representative Western blots for detection of adiponectin and CD9 are shown.

97

5

CHAPTER 5

Table 1. Baseline characteristics. n = 16 Age (years)

63.9 ± 7.1

Male gender, n (%)

15 (94)

Body mass index (kg/m²)

25.8 ± 4.0

Waist circumference (cm)

Men

102 ± 10

Systolic blood pressure (mmHg)

Women

74

Diastolic blood pressure (mmHg)

141 ± 22

Current smoking, n (%)

85 ± 10

Type 2 diabetes, n (%)

6 (38)

Metabolic syndrome, n (%)a

3 (19)

Metabolic parameters

9 (50)

Fasting glucose (mmol/L)

6.2 ± 0.7

Insulin (IU/I)

13.5 (6.8 – 27.0)

HOMA-IR

3.9 (1.6 – 7.6)

HbA1c (mmol/mol)

40 ± 4

HsCRP (mg/L)

4.3 (1.4 – 7.4)

Aspartate aminotransferase (IU/L)

22 ± 12

Alanine aminotransferase (IU/L)

24 ± 13

Gamma glutamyltransferase (IU/l)

49 (31 – 65)

Cholesterol (mmol/L)

4.1 ± 0.8

LDL-cholesterol (mmol/L)

2.2 ± 0.6

HDL-cholesterol (mmol/L)

1.1 (0.9 – 1.0)

Triglycerides (mmol/L)

1.4 (0.9 – 2.0)

History of vascular disease, n (%) Cerebrovascular disease

1 (6)

Coronary artery disease

6 (38)

Peripheral artery disease

6 (38)

Renal failure (GFR < 60 ml/min)

2 (12)

Type 2 diabetes Mellitus

3 (19)

Medication use, n (%) Oral anticoagulants

1 (6)

Thrombocyte-aggregation inhibitors

12 (75)

Blood pressure-lowering agents

16 (100)

Lipid-lowering agents

13 (81)

Metformin

3 (19)

Insulin therapy

0 (0)

Values are expressed in n (%), mean±SD or median (interquartile range). a Defined according to the National Cholesterol Education Program ATPIII-revised guidelines.

98

Adipose vesicles and insulin signaling

Effect of AT-EVs on hepatic insulin signaling We first determined the effect of human ex vivo AT-EVs on insulin signaling in human liver cells. HepG2 cells were treated with either SAT or OAT-EVs, and TNFα was used as a positive control. After insulin stimulation, phosphorylation of the serine/threonine kinase Akt, a central player in the insulin signaling pathway, was determined. We observed considerable variability in individual response of insulin-induced Akt phosphorylation by both SAT and OAT-EV stimulation of HepG2 hepatocytes (Figure 2 A and B). The majority of AT-EVs (cluster 1, red dots) inhibited insulin-induced Akt phosphorylation compared to insulin stimulation alone (mean pAkt/Akt ratio of SAT-EVs: 0.20±0.07, p<0.05, of OAT-EVs 0.17±0.07 p<0.01; versus pAkt/Akt ratio of insulin 0.30±0.05). Remarkably, AT-EVs of cluster 2 (blue dots) showed a profound stimulation of Akt phosphorylation (mean pAkt/ Akt ratio SAT-EVs: 0.45±0.06; OAT-EVs 0.50±0.08 versus insulin 0.36±0.05, not significantly different), suggesting an insulin sensitizing effect (Figure 2, blue dots). We further analyzed gene expression levels of glucose-6-phosphatase (G6P) and (phosphoenolpyruvate carboxykinase) PEPCK, two gluconeogenic genes that are normally repressed by insulin.

5 C.

HepG2 hepatocytes

3

Fold induction

s TN Fα

s

EV T A

A

O

T

EV

EV s O

SA

T

T SA

PB S

PB S

EV

s

+ Insulin

pAkt

HepG2 hepatocytes

D.

2

1

Akt O AT

TN Fα

S

SA T

3

PEPCK

2

1

T OA

TN Fα

SA T

PB S

0 PB S

TN Fα

s

O

A

T

EV

EV s SA

T

PB S

0.0

PB S

1

0 0.0

Relative fold induction G6P

Fold induction

**

+ Insulin

2

0.2

0.4

pAKT / AKT ratio

0.6

OAT EVs 3

0.4

0.2

r: -0.60; p = 0.01

+ Insulin

HepG2 hepatocytes 0.6

pAkt / Akt ratio

PB

PB S

0

B.

SAT EVs 3

G6P

Relative fold induction G6P

A.

r: -0.74; p = 0.001

2

1

0 0.0

0.2

0.4

pAKT / AKT ratio

0.6

+ Insulin

Figure 2. Effect of EVs derived from subcutaneous AT (SAT) or visceral AT (OAT) on Akt phosphorylation of insulin induced hepatocytes. (A) Representative immunoblot showing the effect of SAT and OAT-EVs from 2 subjects on Akt phosphorylation in cultured HepG2 hepatocytes. (B) pAkt/Akt ratios of densitometric analysis of Akt phosphorylation of six independent experiments, including 16 patients. Data are presented as mean with standard deviation (SD). One way ANOVA with Dunnets post hoc test was used to compare means of vehicle-induced, TNFα-induced or AT-EV-induced insulin-stimulated Akt phosphorylation versus insulin-induced Akt phosphorylation alone. Significant differences are indicated (**P<0.01 versus insulin treatment). (C) Relative fold induction of G6P and PEPCK. Data are presented as median with interquartile range. (D) Spearman correlation coefficients (r) were calculated for the relation between the pAKT/AKT ratio and gene expression of G6P from SAT-EVs (upper panel) and OAT-EVs (lower panel).

99

CHAPTER 5

EVs of the 4 patients that induced Akt phosphorylation (Figure 2B, blue dots), also showed the most pronounced down regulation of G6P and PEPCK (Figure 2C, blue dots).

Furthermore there was a strong negative relation between the pAkt/Akt ratio and relative expression of G6P (SAT-EVs r = -0.60, p=0.01; OAT-EVs r = -0.73, p=0.001; Figure 2D).

There were no significant differences in clinical characteristics between patients of cluster 1 versus cluster 2. OAT of patients from cluster 2 (insulin resistant effect) contained more macrophages compared to patients from cluster 1 (Supporting table 2).

Effect of AT-EVs on insulin signaling in muscle cells Next, the effect of AT-EVs on insulin-induced Akt phosphorylation in skeletal muscle cells was investigated. C2C12 myotubes were treated with AT-EVs from the same individuals as used for hepatocytes, but in contrast to hepatocytes we observed no difference in insulin-induced Akt phosphorylation of either SAT or OAT-EVs versus insulin treatment alone (mean pAKT/AKT ratio SAT-EVs: 0.25±0.11, OAT-EVs 0.26±0.13 versus insulin 0.25±0.08). Furthermore, no clusters of insulin-stimulation or inhibition of Akt phosphorylation in C2C12 myotubes could be distinguished. As indicated by the colors, in muscle cells AT-EVs from different patients stimulated Akt phosphorylation, while the EVs that induced insulin-stimulated Akt phosphorylation in HepG2 hepatocytes (blue dots) showed both inhibiting and stimulating effects in muscle cells (Figure 3). These data indicate that besides an individual variability in effects of AT-EVs on insulin signaling (Akt phosphorylation and target gene expression), identical AT-EVs also exert distinct effects on insulin signaling in liver and muscle cells.

A.

B.

C2C12 myotubes

C2C12 myotubes 0.6

TN Fα

SAT EVs

OAT EVs

pAkt / Akt ratio

pAkt

PB S

PB S

+ Insulin 0.4

0.2

*

Akt

s EV A T

TN Fα

s EV O

T

SA

In su lin

PB S

0.0

Figure 3. Effect of EVs from subcutaneous AT (SAT) or visceral AT (OAT) on Akt phosphorylation of insulin induced skeletal muscle cells. (A) Representative immunoblot showing the effect of SAT and OAT-EVs from 3 subjects on Akt phosphorylation in cultured C2C12 myotubes. (B) pAkt/Akt ratios of densitometric analysis of Akt phosphorylation of six independent experiments, including 16 patients.

Data are presented as mean with standard deviation (SD). One way ANOVA with Dunnets post hoc test was used to compare means of vehicle-induced, TNFα-induced or AT-EV-induced insulin-stimulated Akt phosphorylation versus insulin-induced Akt phosphorylation alone. Significant differences are indicated (*P<0.05 versus insulin treatment).

100

Adipose vesicles and insulin signaling

Table 2. Adipokine profile of subcutaneous and visceral EVS. SAT-EVs

OAT-EVs

P value

IL6 (pg/ml)

66.2 (36.7 – 137.9)

288.9 (116.5 – 873.2)

0.024

MIF (ng/ml)

53.6 (144.7 – 470.7)

286.4 (144.7 – 470.7)

0.005

MCP-1 (pg/ml)

37.5 (31.3 – 69.0)

55.2 (23.53 – 296.3)

0.034

Adiponectin (ng/ml)

82.1 (39.6 – 111.9)

82.8 (48.5 – 155.8)

0.909

Resistin (pg/ml)

13.4 (5.94 – 21.3)

16.9 (7.25 – 68.7)

0.831

RBP-4 (ng/ml)

35.5 (20.36 – 55.1)

31.4 (25.9 – 41.5)

0.470

Adipokine values are presented as median with interquartile range. Differences in median adipokine concentrations in SAT versus OAT-EVs are calculated by Wilcoxon matched-pairs signed rank test.

Table 3. Correlations between AT-EV adipokine concentration and Akt phosphorylation.

HepG2 hepatocytes

C2C12 myotubes

pAKT / AKT ratio SAT

pAKT / AKT ratio OAT

pAKT / AKT ratio SAT

pAKT / AKT ratio OAT

Log IL6

-0.46

-0.64*

0.42

-0.13

Log MIF

-0.30

-0.59*

-0.01

-0.27

-0.75**

-0.48

-0.03

-0.03

Log adiponectin

0.30

-0.32

0.03

-0.13

Log Resistin

-0.43

-0.50

0.57

0.05

Log RBP-4

-0.41

-0.31

0.43

-0.42

Log MCP-1

Pearson correlation coefficients between degree of Akt phosphorylation and log adipokine concentration in SAT or OAT-EVs (n=14). * P value < 0.05, ** P value < 0.001.

Adipokine profile of SAT and OAT-EVs Having previously shown that adipocyte derived EVs contain numerous adipokines (20), we wondered whether there would be a difference in the adipokine profile between SAT and OAT-EVs. In both SAT and OAT-derived EVs insulin-sensitive (adiponectin) and insulinresistant adipokines (interleukin-6 (IL-6), monocyte chemoattractant protein-1 (MCP-1), macrophage migration inhibitory factor (MIF), retinol-binding protein 4 (RBP-4) and resistin) could be detected. The concentration of IL-6, MIF and MCP-1 was significantly higher in OAT-EVs compared to SAT-EVs (median concentration of IL-6 was 288.9 (116.5 – 873.2) pg/ml versus 66.2 (36.7 – 137.9) pg/ml, of MIF 286.4 (144.7 – 470.7) ng/ml versus 53.6 (144.7 – 470.7) ng/ml and of MCP-1 55.2 (23.53 – 296.3) pg/ml versus 37.5 (31.3 – 69.0) pg/ml (Table 2). TNF-α and leptin were below detection limit in our assay.

Given the individual differences in effect of AT-EVs on hepatic and muscle insulin signaling, we further explored whether concentrations of adipokines with either insulin-resistant or insulin-sensitizing properties might be related to inhibition or stimulation of hepatic Akt phosphorylation. In SAT-EVs, a higher concentration of log transformed MCP-1 was related to a lower Akt phosphorylation (Pearson r -0.75, p<0.001; Table 3). In OAT-EVs, log

101

5

CHAPTER 5

transformed IL-6, MIF and to a lesser extend MCP-1 were negatively correlated with Akt phosphorylation in hepatocytes (pearson r -0.64 p=0.014; r -0.59, p=0.026 and r -0.48, p=0.079 respectively (Table 3). The concentrations of the other adipokines were not related to Akt phosphorylation in hepatocytes. In agreement with the absence of an overall functional effect on Akt phosphorylation (Figure 3), none of the adipokines was related to Akt phosphorylation in muscle cells (Table 3).

Number of SAT-EVs and OAT-EVs in relation to clinical characteristics, metabolic features and liver enzymes Having previously shown that the number of AT-EVs are associated with systemic insulin resistance (20), we wondered whether OAT or SAT-EVs are related to clinical characteristics, metabolic features and liver enzymes. EVs were quantified by fluorescence-based highresolution flow cytometry (20,29) and the amount of measured EVs is expressed as the number of EVs per gram AT. The number of SAT-EVs was inversely related to waist circumference (Pearson r -0.61, p=0.049) and the metabolic syndrome (Pearson r -0.64, p=0.034) (Table 4). In contrast, the number of OAT-EVs were related to elevated systemic concentrations of the liver enzymes aspartate aminotransferase (AST) (Pearson r 0.62, p=0.043), alanine aminotransferase (ALT) (Pearson r 0.61, p=0.045) and gamma glutamyltransferase (ÎłGT) (Pearson r 0.83, p=0.001). Neither SAT or OAT-EVs were related to circulating cholesterol or triglyceride levels (Table 4), nor did AT-EV associated adipokine levels relate to circulating liver enzymes of lipids (data not shown).

Table 4. Correlations between number of AT-EVs and clinical metabolic characteristics.

Number of SAT-EVs

Number of OAT-EVs

Waist circumference

-0.61*

-0.01

Metabolic Syndrome

-0.64*

-0.14

Triglycerides

-0.15

0.49

Cholesterol

-0.38

0.51

HDL-cholesterol

-0.41

-0.012

LDL-cholesterol

-0.19

0.32

AST

-0.19

0.62*

ALT

-0.23

0.61*

Gamma GT

0.18

0.83**

Pearson correlation coefficients (r) were calculated for the number of EVs per gram SAT (A) or OAT-EVs (B) with clinical characteristics (n=11).

102

Adipose vesicles and insulin signaling

DISCUSSION In the present study, we have shown that AT-EVs can directly interfere with insulin signaling in liver and muscle cells. The degree of Akt phosphorylation in hepatocytes was associated with MCP-1 levels of SAT-EVs, and with IL-6 and MIF levels of OAT-EVs. Furthermore, the number of OAT-EVs was related to elevated liver enzymes, while adipokine content was not. From these observations we speculate that human AT-EVs can affect insulin signaling in hepatocytes- possibly depending on their adipokine content- and may thereby contribute to systemic IR. In this study, AT-derived EVs exerted different effects on insulin signaling in muscle and liver cells. Overall, the effect of AT-EVs on hepatocytes was more pronounced compared to skeletal muscle cells, but also the effects of EVs from a single individual on muscle cells were quite different from the effects on hepatocytes. In skeletal muscle, impaired insulin signaling results in impaired glucose uptake. In liver, mediators which interfere with insulin signaling impair the ability of insulin to regulate gluconeogenesis as well as storage of glucose, lipids and glycogen in hepatocytes (30). Based on the difference in insulin signaling pathways between skeletal muscle and liver, distinct effects of AT-EVs from the same individuals on liver versus skeletal muscle might not be surprising (31,32).

The mechanisms by which EVs can interfere with insulin signaling in muscle and liver are currently unknown. Distinct mechanisms can cause disruption of insulin signaling, depending on the responsible mediators. Pro-inflammatory adipokines induce insulin resistance via activation of the inflammatory IKKβ-NFκB pathway, inhibiting IRS1 and AKT signaling (15,33). We have shown that AT-EVs contain multiple adipokines, of which a high concentration of adipokines with insulin-resistant properties (IL6, MIF and MCP-1) is related to inhibition of Akt phosphorylation in hepatocytes. These findings might suggest that the hepatic insulin resistance caused by AT-EVs might be dependent on adipokine content. Of particular interest is MIF, of which we previously reported to be abundantly present in EVs derived from in vitro differentiated adipocytes (20). The observations in the present study support our hypothesis that AT-EV associated MIF might play a role in obesity-induced insulin resistance (20). Conversely, we did not find a relation between insulin-sensitizing adipokines and higher levels of Akt phosphorylation (Table 2). This indicates that AT-EVs contain other insulin sensitizing mediators contributing to the insulin sensitive effect on both liver and muscle cells. However, as only a few patients showed insulin-sensitive effects, the lack of a relation with EV-associated adiponectin might also be due to the low patient number we could assess in this study.

Besides a direct effect of OAT-EVs on hepatic insulin resistance, we observed a positive relation between the number of EVs secreted by OAT with elevated circulating liver enzymes. The liver enzymes (AST, ALT and γGT) have been related to non-alcoholic fatty liver disease (NAFLD), intrahepatic cholestasis and steatosis. It is hypothesized that NAFLD is caused by (obesity-induced) insulin resistance, followed by reactive oxygen species ultimately leading to inflammation (34). Of interest, OAT-EVs contained higher levels of IL-6 and MIF compared to SAT-EVs, which were also related to lower Akt phosphorylation.

103

5

CHAPTER 5

Furthermore, these adipokines are associated with liver pathology such as steatohepatitis and NAFLD as well. However, in the present study we only observed a relation between elevated liver enzymes and the number of OAT-EVs per gram AT, and not with their specific adipokine content (data not shown). Besides adipokines, EV associated lipids might interfere with liver function. FFA are important mediators in obesity-induced insulin resistance and NAFLD by increasing intracellular DAG and ceramides. All of these lipids have been described to be present in human EVs as well, and are therefore possible mediators responsible for disruptive insulin signaling and liver pathophysiology (35,36).

In conclusion, human AT-EVs can directly interfere with insulin signaling in liver and muscle cells, which is associated with MCP-1 levels of SAT-EVs, and with IL-6 and MIF levels of OAT-EVs in HepG2 hepatocytes. Furthermore, the number of OAT-EVs was related to elevated liver enzymes, while EV adipokine content was not. From these observations we speculate that human AT-EVs can positively or negatively affect insulin signaling in hepatocytes- possibly depending on their adipokine content- and may thereby contribute to systemic IR. Acknowledgements We would like to acknowledge everyone in the van Mil and Kalkhoven group for helpful discussions. In particular, we would like to thank A. Milona for expert advice and helpful discussions. We further thank G.J.A Arkesteijn for technical assistance with the high resolution flow cytometry measurements, and members of the Wauben laboratory for assistance with analyzing the flow cytometry data. Finally, we gratefully acknowledge the help of dr. A. Vink from the department of pathology with histological analyses.

104

Adipose vesicles and insulin signaling

REFERENCES 1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

Finucane MM, Stevens GA, Cowan MJ, Danaei G, Lin JK, Paciorek CJ, et al. National, regional, and global trends in body-mass index since 1980: Systematic analysis of health examination surveys and epidemiological studies with 960 country-years and 9.1 million participants. Lancet.

2011 Feb 12;377(9765):557-67.

Kahn BB, Flier JS. Obesity and insulin resistance. J Clin Invest. 2000 Aug;106(4):473-81.

Ford ES, Giles WH, Dietz WH. Prevalence of the metabolic syndrome among US adults: Findings from the third national health and nutrition examination survey. JAMA. 2002 Jan 16;287(3):356-9.

Juhan-Vague I, Alessi MC, Vague P. Increased plasma plasminogen activator inhibitor 1 levels. A possible link between insulin resistance and atherothrombosis. Diabetologia. 1991 Jul;34(7):457-62.

Howard G, O'Leary DH, Zaccaro D, Haffner S, Rewers M, Hamman R, et al. Insulin sensitivity and atherosclerosis. the insulin resistance atherosclerosis study (IRAS) investigators. Circulation.

1996 May 15;93(10):1809-17.

Saltiel AR, Kahn CR. Insulin signalling and the regulation of glucose and lipid metabolism.

Nature. 2001 Dec 13;414(6865):799-806.

Fox CS, Massaro JM, Hoffmann U, Pou KM, Maurovich-Horvat P, Liu CY, et al. Abdominal visceral and subcutaneous adipose tissue compartments: Association with metabolic risk factors in the framingham heart study. Circulation. 2007 Jul 3;116(1):39-48.

Kanhai DA, Kappelle LJ, van der Graaf Y, Uiterwaal CS, Visseren FL, SMART Study Group.

The risk of general and abdominal adiposity in the occurrence of new vascular events and mortality in patients with various manifestations of vascular disease. Int J Obes (Lond). 2012 May;36(5):695-702.

Sell H, Eckel J. Adipose tissue inflammation: Novel insight into the role of macrophages and lymphocytes. Curr Opin Clin Nutr Metab Care. 2010 Jul;13(4):366-70.

Jensen MD, Haymond MW, Rizza RA, Cryer PE, Miles JM. Influence of body fat distribution on free fatty acid metabolism in obesity. J Clin Invest. 1989 Apr;83(4):1168-73.

Shi H, Kokoeva MV, Inouye K, Tzameli I, Yin H, Flier JS. TLR4 links innate immunity and fatty acid-induced insulin resistance. J Clin Invest. 2006 Nov;116(11):3015-25.

Holland WL, Bikman BT, Wang LP, Yuguang G, Sargent KM, Bulchand S, et al. Lipid-induced insulin resistance mediated by the proinflammatory receptor TLR4 requires saturated fatty acidinduced ceramide biosynthesis in mice. J Clin Invest. 2011 May;121(5):1858-70.

Suganami T, Nishida J, Ogawa Y. A paracrine loop between adipocytes and macrophages aggravates inflammatory changes: Role of free fatty acids and tumor necrosis factor alpha.

Arterioscler Thromb Vasc Biol. 2005 Oct;25(10):2062-8.

Hotamisligil GS, Shargill NS, Spiegelman BM. Adipose expression of tumor necrosis factoralpha: Direct role in obesity-linked insulin resistance. Science. 1993 Jan 1;259(5091):87-91.

Norseen J, Hosooka T, Hammarstedt A, Yore MM, Kant S, Aryal P, et al. Retinol-binding protein 4 inhibits insulin signaling in adipocytes by inducing proinflammatory cytokines in macrophages through a c-jun N-terminal kinase- and toll-like receptor 4-dependent and retinol-independent mechanism. Mol Cell Biol. 2012 May;32(10):2010-9.

Verschuren L, Kooistra T, Bernhagen J, Voshol PJ, Ouwens DM, van Erk M, et al. MIF deficiency reduces chronic inflammation in white adipose tissue and impairs the development of insulin resistance, glucose intolerance, and associated atherosclerotic disease. Circ Res. 2009 Jul 2;105(1):99-107.

Scherer PE, Williams S, Fogliano M, Baldini G, Lodish HF. A novel serum protein similar to C1q, produced exclusively in adipocytes. J Biol Chem. 1995 Nov 10;270(45):26746-9.

Lin HV, Kim JY, Pocai A, Rossetti L, Shapiro L, Scherer PE, et al. Adiponectin resistance exacerbates insulin resistance in insulin receptor transgenic/knockout mice. Diabetes. 2007 Aug;56(8):1969-76.

105

5

CHAPTER 5

19. Deng ZB, Poliakov A, Hardy RW, Clements R, Liu C, Liu Y, et al. Adipose tissue exosomelike vesicles mediate activation of macrophage-induced insulin resistance. Diabetes. 2009 Nov;58(11):2498-505.

20. Kranendonk ME, Visseren FL, van Balkom BW, Hoen EN, van Herwaarden JA, de Jager W, et al. Human adipocyte extracellular vesicles in reciprocal signaling between adipocytes and macrophages human adipocyte EVs and macrophages. Obesity (Silver Spring). 2013 Dec 12.

21. Raposo G, Stoorvogel W. Extracellular vesicles: Exosomes, microvesicles, and friends. J Cell Biol. 2013 Feb 18;200(4):373-83.

22. Muller G. Microvesicles/exosomes as potential novel biomarkers of metabolic diseases.

Diabetes Metab Syndr Obes. 2012;5:247-82.

23. Grundy SM, Brewer HB,Jr, Cleeman JI, Smith SC,Jr, Lenfant C, American Heart Association, et al. Definition of metabolic syndrome: Report of the national heart, lung, and blood Institute/ American heart association conference on scientific issues related to definition. Circulation.

2004 Jan 27;109(3):433-8.

24. Thery C, Amigorena S, Raposo G, Clayton A. Isolation and characterization of exosomes from cell culture supernatants and biological fluids. Curr Protoc Cell Biol. 2006 Apr;Chapter 3:Unit 3.22.

25. Jong O.G. de, Verhaar M.C., Chen Y., Vader P., Gremmels H., Posthuma G., Schifferelers R.M., Gucek M., Balkom B.W.M. van. Cellular stress conditions are reflected in the protein and RNA content of endothelial cell-derived exosomes. Journal of Extracellular Vesicles. 2012;1(18396).

26. Jeninga EH, van Beekum O, van Dijk AD, Hamers N, Hendriks-Stegeman BI, Bonvin AM, et al. Impaired peroxisome proliferator-activated receptor gamma function through mutation of a conserved salt bridge (R425C) in familial partial lipodystrophy. Mol Endocrinol. 2007 May;21(5):1049-65.

27. de Jager W, Rijkers GT. Solid-phase and bead-based cytokine immunoassay: A comparison.

Methods. 2006 Apr;38(4):294-303.

28. Schipper HS, de Jager W, van Dijk ME, Meerding J, Zelissen PM, Adan RA, et al. A multiplex immunoassay for human adipokine profiling. Clin Chem. 2010 Aug;56(8):1320-8.

29. van der Vlist EJ, Nolte-'t Hoen EN, Stoorvogel W, Arkesteijn GJ, Wauben MH. Fluorescent labeling of nano-sized vesicles released by cells and subsequent quantitative and qualitative analysis by high-resolution flow cytometry. Nat Protoc. 2012 Jun 14;7(7):1311-26.

30. Matsumoto M, Han S, Kitamura T, Accili D. Dual role of transcription factor FoxO1 in controlling hepatic insulin sensitivity and lipid metabolism. J Clin Invest. 2006 Sep;116(9):2464-72.

31. Muoio DM, Newgard CB. Mechanisms of disease: Molecular and metabolic mechanisms of insulin resistance and beta-cell failure in type 2 diabetes. Nat Rev Mol Cell Biol. 2008 Mar;9(3):193-205.

32. Samuel VT, Shulman GI. Mechanisms for insulin resistance: Common threads and missing links. Cell. 2012 Mar 2;148(5):852-71.

33. Olefsky JM, Glass CK. Macrophages, inflammation, and insulin resistance. Annu Rev Physiol.

2010;72:219-46.

34. Tiniakos DG, Vos MB, Brunt EM. Nonalcoholic fatty liver disease: Pathology and pathogenesis.

Annu Rev Pathol. 2010;5:145-71.

35. Wubbolts R, Leckie RS, Veenhuizen PT, Schwarzmann G, Mobius W, Hoernschemeyer J, et al.

Proteomic and biochemical analyses of human B cell-derived exosomes. potential implications for their function and multivesicular body formation. J Biol Chem. 2003 Mar 28;278(13):10963-72.

36. Laulagnier K, Motta C, Hamdi S, Roy S, Fauvelle F, Pageaux JF, et al. Mast cell- and dendritic cellderived exosomes display a specific lipid composition and an unusual membrane organization.

Biochem J. 2004 May 15;380(Pt 1):161-71.

106

Adipose vesicles and insulin signaling

SUPPLEMENTARY INFORMATION Supporting table 1. Primers used for qPCR analysis.

Gene

Forward primer

Reverse primer

36B4

CGGGAAGGCTGTGGTGCTG

GTGAACACAAAGCCCACATTCC

G6P

TACGTCCTCTTCCCCATCTG

TCCCTGGTCCAGTCTCACA

PEPCK

GATGAGCCGCTAGCTTCA

TTGCCGAAGTTGTAGCCA

Supporting table 2. Baseline characteristics of insulin sensitizing EVs versus insulin resistant EVs. EVs with insulin sensitizing effects n=4

EVs with insulin resistant effects n = 12

P value

Age (years)

62

65 ± 7.5

0.436

Male gender, n (%)

4 (100)

11 (92)

0.339

Body mass index (kg/m²)

27.7 ± 2.0

25.1 ± 4.3

Waist circumference (cm)

105 ± 5

Men

Women (1) 74

(male subjects)

Systolic blood pressure (mmHg)

136 ± 23

141 ± 22

0.952

Diastolic blood pressure (mmHg)

81 ± 8

85 ± 10

0.123

Current smoking, n (%)

1 (25)

4 (33)

0.590

Type 2 diabetes, n (%)

1 (25)

3 (25)

1.000

3 (75)

5 (42)

0.301

Fasting glucose (mmol/L)

6.6 ± 0.4

6.2 ± 0.7

0.137

Insulin (IU/I)

18.0 (6.8 – 28.5)

12.5 (6.8 – 26.0)

0.955

HOMA-IR

5.1 (2.0 – 8.5)

3.4 (1.6 – 7.5)

0.866

HbA1c (mmol/mol)

41 ± 1.4

40 ± 4.6

0.529

HsCRP (mg/L)

2.9 (1.3 – 4.7)

4.5 (1.9 – 8.7)

0.160

AST (IU/L)

17 ± 4

24 ± 14

0.130

ALT (IU/L)

24 ± 9

24 ± 15

0.990

γGT (IU/l)

48 (25 – 84)

49 (31 – 65)

0.775

Cholesterol (mmol/L)

3.7 ± 1.2

4.3 ± 0.7

0.447

LDL-cholesterol (mmol/L)

2.0 ± 0.8

2.3 ± 0.6

0.333

HDL-cholesterol (mmol/L)

1.0 (0.8 – 1.2)

1.1 (1.0 – 1.4)

0.451

Triglycerides (mmol/L)

1.4 (0.9 – 2.3)

1.4 (1.0 – 1.9)

0.801

Number of CD68 macrophages

4.0 ± 0.7

3.9 ± 1,4

0.858

Number of crown like structures

3.2 (0.7 – 10.5)

0.8 (0.0 – 2.1)

0.401

Adipocyte cell size (mm2)

8.4 ± 1.9

9.0 ± 2.8

0.704

Number of CD68 macrophages

4.2 ± 0.2

5.7 ± 1.5

0.008*

Number of crown like structures

0.9 (0.0 – 6.5)

1.5 (0.5 – 10.3)

0.440

Adipocyte cell size (mm2)

10.7 ± 1.5

9.6 ± 4.0

0.624

Metabolic syndrome, n (%)

a

0.124

(11) 101 ± 11 0.272

Metabolic parameters

OAT

SAT

Adipose tissue characteristics

Values are expressed in n (%), mean ± SD or median (interquartile range). a Defined according to the National Cholesterol Education Program ATPIII-revised guidelines. HOMA: homeostatic model assessment; AST: aspartate aminotransferase; ALT: alanine aminotransferase; γGT: gamma glutamyltransferase.

107

5

PART THREE

Biomarkers for obesity-induced cardiovascular or metabolic disease

CHAPTER 6 Extracellular vesicle markers in relation to obesity and metabolic complications in patients with manifest cardiovascular disease Cardiovasc Diabetol. 2014 Feb 5;13:37. doi: 10.1186/1475-2840-13-37

MariĂŤtte E.G. Kranendonk, Dominique P.V. de Kleijn, Eric Kalkhoven, Danny A. Kanhai, Cuno S.P.M. Uiterwaal, Yolanda van der Graaf, Gerard Pasterkamp, Frank L.J.Visseren

CHAPTER 6

ABSTRACT Background Alterations in extracellular vesicles (EVs), including exosomes and microparticles, contribute to cardiovascular disease. We hypothesized that obesity could favour enhanced release of EVs from adipose tissue, and thereby contribute to cardiovascular risk via obesity-induced metabolic complications. The objectives of this study were: 1) to investigate the relation between the quantity, distribution and (dys)function of adipose tissue and plasma concentrations of atherothrombotic EV-markers; 2) to determine the relation between these EV-markers and the prevalence of the metabolic syndrome; and 3) to assess the contribution of EV markers to the risk of incident type 2 diabetes.

Methods In 1012 patients with clinically manifest vascular disease, subcutaneous and visceral fat thickness was measured ultrasonographically. Plasma EVs were isolated and levels of cystatin C, serpin G1, serpin F2 and CD14 were measured, as well as fasting metabolic parameters, hsCRP and adiponectin. The association between adiposity, EV-markers, and metabolic syndrome was tested by multivariable linear and logistic regression analyses.

As sex influences body fat distribution, sex-stratified analyses between adipose tissue distribution and EV-markers were performed. The relation between EV-markers and type 2 diabetes was assessed with Cox regression analyses.

Results Higher levels of hsCRP (β 5.59; 95%CI 3.00–8.18) and lower HDL-cholesterol levels (β11.26; 95%CI -18.39 – -4.13) were related to increased EV-cystatin C levels, and EV-cystatin C levels were associated with a 57% higher odds of having the metabolic syndrome (OR 1.57; 95%CI 1.19–2.27). HDL-cholesterol levels were positively related to EV-CD14 levels (β 5.04; 95%CI 0.07–10.0), and EV-CD14 levels were associated with a relative risk reduction of 16% for development of type 2 diabetes (HR 0.84, 95%CI 0.75–0.94), during a median follow up of 6.5 years in which 42 patients developed type 2 diabetes.

Conclusions In patients with clinically manifest vascular disease, EV-cystatin C levels were positively related, and EV-CD14 levels were negatively related to metabolic complications of obesity.

112

EV-markers, obesity and metabolic complications

BACKGROUND Obese individuals are at increased risk of developing cardiovascular disease (CVD), a consequence of adipose tissue (AT) expansion and subsequent dysfunction (1,2). In particular the expansion of visceral AT (VAT), rather than subcutaneous AT (SAT), is an independent risk factor for cardiovascular morbidity and mortality (3,4). VAT expansion results in local inflammation, characterized by hypertrophic adipocytes and increased influx of pro-inflammatory macrophages and cytotoxic T cells, contributing to elevated plasma levels of interleukin-6 (IL-6) and high sensitive C-reactive protein (hsCRP) (5,6,7).

Simultaneously, expression of the anti-inflammatory adipokine adiponectin is downregulated in adipocytes (8). This inflammatory milieu leads to metabolic complications that predispose to the development of CVD, including low-grade systemic inflammation, insulin resistance (9) and development of the metabolic syndrome and type 2 diabetes (10,11). Further emphasizing the link between VAT and CVD are sex differences in body fat distribution, where males are both prone to develop abdominal obesity due to accumulation of visceral fat and suffer from a higher incidence of metabolic and cardiovascular disease (12).

Nevertheless, the pathophysiological mechanisms underlying the development of obesityinduced CVD are still poorly understood, and a biomarker indicating the obese individuals at risk would be extremely useful (13). Extracellular vesicles (EVs) or EV-associated molecules are promising biomarkers in a variety of pathological settings (14,15). EVs include microvesicles, microparticles and exosomes which are nanometre sized membrane vesicles secreted by all eukaryotic cells (16). These vesicles reflect the state of a cell, and serve as messenger vehicles containing cell-specific cytosolic and membrane-bound proteins and RNA. As such, EVs can modify and activate target cells in a paracrine or endocrine fashion (17). Important biological functions of EVs have been reported in a variety of (patho) physiological processes, including immune responses, inflammation, tumorigenesis and endothelial dysfunction (14).

Our group has previously shown that four EV-associated proteins cystatin C, serpin G1, serpin F2 and CD14 are associated with atherosclerotic plaque formation in patients with clinically manifest vascular disease, and that three of these (cystatin C, serpin F2 and CD14) related to an increased risk for cardiovascular morbidity and mortality in these patients (18).

Increased levels of cystatin C, a cysteine protease, are observed in obese subjects and associate with CVD (19,20). Serpin G1,better known as C1 inhibitor, is an inhibitor of kallikrein and factor IIa (21). Serpin F2 or α-2-antiplasmin, is the primary inhibitor of plasmin in the circulation (22). CD14, a monocyte marker, is present on monocyte-derived EVs and such EVs have been shown to be capable of stimulating endothelial cells (23). Interestingly, there appears to be considerable overlap between the inflammatory processes underlying AT inflammation and those that underlie atherosclerosis, since inflammatory cytokines such as hsCRP and IL-6, as well as immune cells such as macrophages, play crucial roles in the onset and progression of both conditions (24,25). Furthermore, AT inflammation contributes to a pro-thrombotic state by secretion of adipokines such as TNF-α, leptin and IL-1β, which triggers blood coagulation by inducing the synthesis of tissue factor (26,27).

113

6

CHAPTER 6

Therefore, it can be hypothesized that adiposity contributes to an elevated risk of developing CVD via increased production of EVs or EV-related molecules such as EV-associated cystatin C, serpin G1, serpin F2 and CD14 .

While we previously reported on possible determinants for protein levels of these four EV-markers (18), we now aimed to further investigate the potential etiologic relation of AT quantity, AT distribution and metabolic parameters of AT (dys)function with EV-associated plasma protein levels of cystatin C, serpin G1, serpin F2 and CD14. Furthermore, the relation of these EV-markers to metabolic syndrome and incident type 2 diabetes in patients with clinically manifest vascular disease was investigated.

METHODS Study design and patient population The study cohort consists of patients participating in the Second Manifestations of ARTerial disease (SMART) study, an ongoing prospective single-centre cohort study in patients with manifest arterial disease or cardiovascular risk factors at the University Medical Centre Utrecht (UMCU) which started in September 1996 (28). The study was approved by the Ethics Committee of the UMCU and all patients gave their written informed consent. All patients included in the SMART-study were asked to complete a health questionnaire covering medical history, risk factors, smoking habits and medical treatment. A standardized diagnostic protocol was followed consisting of physical examination and laboratory testing in a fasting state. A detailed study rationale and description are published elsewhere (28).

For the present study, a total of 1062 patients with clinically manifest vascular disease at study inclusion, who enrolled in the SMART cohort between May 2001 until December 2005 were included. In this subset of patients, concentrations of extracellular vesicle (EV)associated cystatin C, serpin G1, serpin F2 and CD14 were determined, as described previously (18). Of the eligible 1062 patients, 2 had insufficient material for measurements of the four EV-markers, and 48 patients with hsCRP levels above 15 mg/L were excluded, leaving a final number of 1012 patients eligible for cross-sectional analysis. To reduce bias and improve statistical efficiency, missing values for smoking status (n=6), estimated glomerular filtration rate ((eGFR), based on the modification of diet in renal disease (MDRD) formula) (n=12), body mass index (n=1), waist circumference (n=42), visceral fat thickness (n=8) and subcutaneous fat thickness (n=19) were completed in the dataset by single imputation. There were no substantial differences in outcomes when we compared the results with complete case analysis (i.e. exclusion of cases with missing values). For analyses with HOMA-IR included in the model, only subjects included in the cohort after August 2003 were considered (n = 530), because plasma insulin levels were measured since that date.

Extracellular vesicle marker measurements EVs were isolated using Exoquickâ&#x201E;˘ (SBI) according to the manufacturer's protocol described previously (18). Briefly, 150 Âľl EDTA plasma was centrifuged for 15 min at

114

EV-markers, obesity and metabolic complications

3000xg. The supernatant was filtered over a 0.45 µm Spin-X filter (Corning), which was flushed with preheated PBS (37 °C) and 38 μl ExoQuick™ solution was added to the filtrate.

After vortexing, the sample was stored overnight at 4 °C. The following day, the sample was centrifuged at 1500×g for 30 min at room temperature, and the pelletwas lysed in 100 µl Roche Complete Lysis‐M with protease inhibitors (EDTA free). The sample was filtered over a 0.22 µm Spin-X filter (Corning) and protein concentration was determined using a Pierce® BCA Protein Assay Kit (Pierce Biotechnology, Rockford, USA), in order to correct the amount of measured EV-marker for the total amount of protein present in the EVs. Samples were stored at -80 °C. After thawing, the lysed sample was diluted 20x with Roche complete Lysis-M buffer, and 50 µl was analysed in a multiplex immunoassay on levels of cystatin C, serpin G1, serpin F2 and CD14 using a Biorad Bioplex 200 system as described before (29). Capture antibody, biotinylated detection antibody and antigen of all 4 proteins were purchased from R&D systems. A full description of preceding biomarkers proteomics discovery work is provided previously (18).

Measurements of AT quantity Visceral adipose tissue (VAT) and subcutaneous adipose tissue (SAT) thickness were quantified by ultrasonographic intra-abdominal fat measurement, performed by well-trained registered vascular technologists in a certified vascular laboratory. Ultrasonographic measurements were made in supine position using an ATL HDI 3000 (Philips Medical Systems, Eindhoven, The Netherlands) with a C 4-2 transducer (30). An inter-observer coefficient of variation of 5.4% was found for ultrasound measurements of intra-abdominal fat, indicating good reproducibility (30). Waist circumference (WC) was measured as the circumference in centimetres halfway between the lower rib and the iliac crest. Body Mass Index (BMI), the weight in kilograms divide by the square if the height in meters, was computed after a standardized anthropometric measurement protocol.

Measurements of metabolic parameters of adipose tissue (dys)function Serum concentration of adiponectin was measured by Luminex immunoassay (Biorad, Munich, Germany) as described previously (31). Plasma insulin was measured with an immunometric technique on an IMMULITE 1000 Analyzer (Diagnostic Products Corporation, Los Angeles, USA). Insulin measurements below the lower limit of detection of 2 mIU/L (n = 1) were left out of the analysis. The value for insulin resistance was assessed by the formula: homeostasis model assessment parameter of insulin resistance (HOMA-IR) = fasting serum glucose (mmol/L) x fasting serum insulin (mIU/L)) / 22.5 (32), and was only performed in patients without antihyperglycaemic drugs. High-density lipoprotein cholesterol (HDL-C) in plasma was determined using a commercial enzymatic kit (Boehringer-Mannheim) after precipitation of low density lipoprotein cholesterol (LDL-C) and very low density lipoprotein cholesterol (VLDL-C) with sodiumphosphotungstate­ magnesium chloride. hsCRP levels were determined by immunonephelometry (Nephelometer Analyser BN II, Dade-Behring, Marburg, Germany), with a lower detection limit of the test of 0.2 mg/L. As high hsCRP levels may have a different pathophysiological

115

6

CHAPTER 6

origin than low-grade inflammation as seen in obesity and vascular diseases, subjects with hsCRP levels > 15mg/L were excluded from analysis.

Follow up To assess the incidence of diabetes, all patients that had been included until December 2005 without diabetes at baseline received a questionnaire in the period between June and December 2006 to assess the incidence of type 2 diabetes after study inclusion. After 2006, all patients were biannually asked to complete this questionnaire. Patients were asked whether they had diabetes and if ‘yes’, they received a supplementary questionnaire regarding date of diagnosis, initial and current treatment (oral medication or insulin). All diabetes cases were audited and classified by two independent physicians. Cross-validation with the hospital diagnosis registry revealed that none of the patients who reported not to have diabetes had a physician’s diagnosis of diabetes. Duration of follow-up was defined as the period between the date of study inclusion and the date of incident type 2 diabetes, date of loss to follow-up or the preselected closing date of 1 March 2010. From 1996 until 1 March 2010, 35 of 937 patients (3.5%) were lost to follow-up.

Data analyses Continuous variables are expressed as mean ± standard deviation (SD) when normal distributed or as median (interquartile range) in case of skewed distribution. Categorical variables are expressed as numbers (percentage). Variables with skewed distributions (cystatin C, serpin G1, serpin F2, CD14, HOMA-IR, adiponectin and hsCRP) were transformed to fulfil linear regression criteria.

Differences in baseline concentrations of EV-markers in different metabolic groups were calculated by analysis of covariance (ANCOVA), corrected for age, sex and eGFR. Data in table 2 are displayed in the original scale of measurement. However, to fulfil ANCOVA’s assumptions a logarithmic transformation was applied to EV-cystatin C, EV-serpin G1 and EV-serpin F2 data, and a square root transformation to EV-CD14 data prior to formal analysis (33). No interaction between covariates and independent variables was observed. For several metabolic groups (waist circumference, visceral adipose tissue or subcutaneous adipose tissue), different distribution or cut off values are known for males and females.

Therefore, male and female patients were divided separately into two groups based on the median value or appropriate cut off value for that sex and then combined into sex-pooled groups.

Adipose tissue quantity and EV-markers As body fat distribution differs between males and females, and is differently related to CVD (34), potential effect modification on the relationship between adiposity and EV markers was investigated by entering an interaction term for sex to the most complete adjusted model. Effect modification of sex could be demonstrated in the relation between SAT and BMI with serpin F2 (p-values for interaction were 0.023 and 0.014 respectively),

116

EV-markers, obesity and metabolic complications

and between WC and CD14 (p-value for interaction was 0.016). Therefore, separate analyses were performed for males and females. Multivariable linear regression analysis was used to evaluate relations between parameters of AT quantity (VAT, SAT, WC and BMI) and EV-markers (EV-associated protein levels of cystatin C, serpin G1, serpin F2 and CD14), expressed as beta (β) regression coefficients and 95% confidence intervals (95%CI).

Analyses were adjusted for age, current smoking, eGFR, type 2 diabetes, blood pressure lowering medication, lipid lowering medication and year of inclusion in SMART. As subjects were included over a time period of 6 years, baseline measurements might be slightly different, and adjustment for year of inclusion was added to the final model.

Metabolic parameters of adipose tissue (dys)function and EV-markers The relation between metabolic parameters of AT (dys)function (plasma levels of adiponectin, hsCRP and HDL-cholesterol and HOMA-IR) and EV-markers was assessed by multivariable linear regression analysis adjusted for age, sex, current smoking, eGFR, type 2 diabetes, blood pressure lowering medication, lipid lowering medication, platelet aggregation inhibitors and year of inclusion in SMART. No interaction for sex could be demonstrated in the relation between metabolic parameters and EV-markers. For presentation purposes, log or square root transformed EV marker values were multiplied by 100. We did not correct for multiple testing, as all analysis performed were strictly hypothesis driven.

EV-markers and metabolic syndrome or incident type 2 diabetes Multivariable logistic regression analyses were used to assess the relation between the EV markers and the metabolic syndrome. The Adult Treatment Panel (ATP) III criteria were taken for the definition of the metabolic syndrome (10). No interaction for sex could be demonstrated in the relation between the metabolic syndrome and the EV-markers. Results are expressed as odds ratios (OR) with corresponding 95%CI. Analyses were adjusted for age, sex, current smoking, eGFR and year of inclusion in SMART. The relation between EV-markers and incident type 2 diabetes was quantified with Cox proportional hazards analysis, and results are expressed as hazard ratios (HR) with corresponding 95%CI. Analyses were adjusted for age, sex, current smoking, eGFR, year of inclusion in SMART, metabolic syndrome and HOMA-IR. The proportional hazards assumptions were formally tested with the Schoenfeld test. No significant nonproportionality (p<0.05) was observed. No interaction for sex could be demonstrated in the relation between these metabolic parameters and the EV-markers.

Analyses were performed in SPSS version 20 (SPSS, Chicago, Illinois, USA) and R version 2.15.2.

117

6

CHAPTER 6

Table 1. Baseline characteristics.

N = 1012 Age (years)

59 ± 10

Male sex, n (%)

804 (79)

Obesity (BMI >30 (kg/m²); n (%))

178 (18)

P value

BMI (kg/m²) Males (n=804)

26.8 ± 3.6

Females (n=208)

27.1 ± 4.7

0.479

Waist circumference (cm) Males (n=804)

97 ± 10

Females (n=208)

88 ± 12

< 0.001

Intra-abdominal fat (cm) Males (n=804)

9.9 ± 2.6

Females (n=208)

8.3 ± 2.4

< 0.001

Subcutaneous fat (cm) Males (n=804)

2.3 ± 1.3

Females (n=208)

3.0 ± 1.5

Type 2 diabetes, n (%)

158 (15)

Metabolic syndrome, n (%)

†

517 (51)

Blood pressure (mmHg) Systolic

144 ± 22

Diastolic

83 ± 12

Smoking, n (%) Never

171 (17)

Ever

464 (46)

Current

377 (37)

Pack years smoking

20.8 (6.4 – 35.0)

Metabolic parameters Cholesterol (mmol/L)

4.9 ± 1.0

LDL-cholesterol (mmol/L)

2.8 ± 0.9

HDL-cholesterol (mmol/L)

1.3 (1.0 – 1.5)

Triglycerides (mmol/L)

1.5 (1.1 – 2.1)

Glucose (mmol/L)

6.3 ± 1.80

Insulin (mU/L)

10.0 (7.0 – 14.0)

HOMA-IR

2.5 (1.6 – 3.9)

hsCRP (mg/L)

1.9 (0.9 – 3.8)

Adiponectin (µg/ml)

19.7 (12.5 – 33.2)

eGFR (ml/min/1.73 m²)

77 ± 18

Creatinine (mmol/l)

93 ± 44

‡

118

< 0.001

EV-markers, obesity and metabolic complications

Table 1 continued History of vascular disease, n (%) Cerebrovascular disease

265 (26)

Coronary artery disease

597 (59)

Peripheral artery disease

242 (24)

Aneurysm of the abdominal aorta

99 (10)

Medication use, n (%) Platelet-aggregation inhibitors

766 (76)

Blood pressure-lowering agents

734 (73)

Lipid-lowering agents

685 (68)

Oral anticoagulants

78 (8)

Anti-hyperglycaemic agents

96 (10)

Values are expressed in n (%), mean ± SD or median (interquartile range). †Defined according to the National Cholesterol Education Program ATPIII-revised guidelines. ‡eGFR, estimated glomerular filtration rate, estimated by the modification of diet in renal disease equation (MDRD). BMI, body mass index; LDL, low-density lipoprotein; HDL, high-density lipoprotein; HOMA-IR, homeostasis model assessment parameter of insulin resistance; hsCRP, high sensitive C-reactive protein.

6

RESULTS Baseline characteristics The patient characteristics are summarized in Table 1. The average age was 59±10 years and 79% were males. In total 18% of the patients was obese. Mean VAT thickness and WC were higher in males than females (VAT males: 9.9±2.6 cm, females: 8.3±2.4 cm, WC males: 97±10 cm, females: 88±12 cm), and mean SAT thickness was higher in females than in males (males: 2.3±1.3 cm, females: 3.0±1.5 cm). 51% of the patients had metabolic syndrome, defined by the Adult Treatment Panel (ATP) III criteria (10), of which 36% had central obesity, 94% were hypertensive, 39% had hypertriglyceridemia, 29% had low HDL-cholesterol levels and 63% had an impaired fasting glucose.

In table 2, baseline levels of circulating EV-cystatin C, EV-serpin G1, EV-serpin F2 and EVCD-14 are presented among different metabolically (un)comprised groups. EV-cystatin C levels were higher in patients with metabolic syndrome compared to patients without metabolic syndrome (9.84 pg/µg versus 9.97 pg/µg, p = 0.010). EV-serpin G1 levels were lower in patients with baseline type 2 diabetes compared to patients without type 2 diabetes at baseline (111.0 pg/µg versus 124.9 pg/µg, p = 0.010). EV-serpin F2 levels were lower in patients with more subcutaneous fat (36.81 pg/µg versus 37.37 pg/µg, p=0.002) and higher in patients suffering from hyperlipidaemia compared to patients without hyperlipidaemia (40.16 pg/µg versus 35.44 pg/µg, p = 0.013). EV-CD14 levels were higher in patients suffering from hyperlipidaemia (12.43 pg/µg versus 10.90 pg/µg, p < 0.001) and lower in patients who developed incident type 2 diabetes compared to patients who did not develop incident type 2 diabetes (11.04 pg/µg versus 11.26 pg/µg, p < 0.001).

119

CHAPTER 6

Relation between adipose tissue quantity and EV-markers In figure 1, the results of four different measures of adiposity in relation to EV-markers are presented. None of the measures of adiposity were significantly related to increased EVlevels in the fully adjusted multivariable regression model. In contrast, fat thickness was related to decreased plasma EV-CD14 levels in both males and females. Interestingly, in males an increase in VAT thickness was significantly related to lower EV-CD14 levels (β -0.95; 95%CI -1.73 – -0.17) while in females SAT thickness was significantly related to lower EV-CD14 levels (β -2.8; 95%CI -5.42 – -0.26, Figure 1).

Supplementary table 1 includes linear regression coefficients with 95%CI of unadjusted, partially adjusted and fully adjusted multivariable models for the relation between measures of obesity and EV-markers.

Men

Women Serpin G1

4

7.0

2

3.5

β (95% CI)

β (95% CI)

Cystatin C

0

-3.5

-2 -4

0.0

VAT

SAT

WC

BMI

-7.0

VAT

35

2

20

0

0

-20

VAT

SAT

WC

BMI

WC

BMI

CD14

β (95% CI)

β (95% CI)

Serpin F2

SAT

WC

BMI

-2 -4 -6

VAT

SAT

Figure 1. Relation between visceral fat, subcutaneous fat, waist circumference and BMI and EVmarkers in men and females with manifest cardiovascular disease. Beta regression coefficients (β) with 95% confidence interval (CI) indicates the difference in log EV-cystatin C, log EV-serpin G1, square root EV-serpin F2 or log EV-CD14 levels per unit increase in AT parameter, adjusted for age, current smoking, eGFR, type 2 diabetes, blood pressure lowering medication, lipid lowering medication and year of inclusion in SMART.

120

EV-markers, obesity and metabolic complications

Relation between metabolic parameters of adipose tissue (dys)function and EV-markers Linear regression coefficients with 95%CI of relations between metabolic parameters of AT (dys)function and EV-markers are shown in figure 2 (fully adjusted model) and supplementary table 2 (including models with and without adjustment for confounding variables). Higher hsCRP levels were strongly related to higher EV-marker levels (cystatin C β 5.59; 95%CI 3.00–8.18, serpin G1 β 7.96; 95%CI 4.76–11.16, serpin F2 β 36.52; 95%CI 21.49–51.55, CD14 β 5.72; 95%CI 3.94–7.49; Figure 2, supplementary table 2). Additional adjustment for dyslipidaemia, history of vascular disease, blood pressure and the metabolic syndrome did not affect the determinant-outcome relation (data not shown). Plasma adiponectin concentration and HOMA-IR levels were not significantly related to any of the EV-markers.

SerpinG1

CD14

SerpinF2 15

5

10

60

10

5

40

5

0 -5 -10

0 -5

β (95% CI)

80

β (95% CI)

15

β (95% CI)

β (95% CI)

Cystatin C 10

20 0

0 -5

-10

-20

-10

-15

-15

-40

-15

-20

-20

-60

-20

hsCRP AdipoQ HOM A HDL

hsCRP AdipoQ HOM A HDL

hsCRP AdipoQ HOM A HDL

6 hsCRP AdipoQ HOM A HDL

Figure 2. Relation between metabolic parameters of adipose tissue (dys)function and EV-markers in patients with manifest cardiovascular disease. β with 95%CI indicates the difference in log EVcystatin C, log EV-serpin G1, square root EV-serpin F2 or log EV-CD14 levels per unit increase in log hsCRP, log adiponectin, log HOMA-IR or HDL-cholesterol, adjusted for age, sex, current smoking, eGFR, type 2 diabetes, blood pressure lowering medication, lipid lowering medication, platelet aggregation inhibitors and year of inclusion in SMART. AdipoQ: adiponectin.

Odds for metabolic syndrome Cystatin C Serpin G1 Serpin F2 CD14 0.0

0.5

1.0

1.5

2.0

2.5

Odds ratio (95% confidence interval)

Figure 3. Relation between EVmarkers and metabolic syndrome in patients with manifest cardio­ vascular disease.

Odds ratios with 95%CI indicate the odds for metabolic syndrome per increase in log EV-cystatin C, log EV-serpin G1, square root EV-serpin F2 or log EV-CD14 concentration, adjusted for age, sex, current smoking, eGFR and year of inclusion in SMART.

121

CHAPTER 6

Table 2. Baseline concentrations of EV-markers (pg/µg) in metabolically (un)compromised groups.

EV-cystatin C

EV-serpin G1

EV-serpin F2

EV-CD14

BMI < 30 kg/m2 (n= 829)

9.5 (7.4 – 12.1)

122.7 (86.0 – 167.1)

36.8 (21.9 – 56.7)

11.56(9.6 – 14.0)

BMI > 30 kg/m2 (n= 178)

9.1 (7.1 – 12.1)

121.5 (83.4 – 176.1)

38.1 (21.4 – 61.2)

11.1 (9.3 – 13.4)

p value

0.776

0.081

0.831

0.139

No (n= 645)

9.1 (7.2 – 11.5)

120.5 (82.5 – 165.7)

36.2 (21.9 – 56.1)

11.4 (9.3 – 13.9)

Yes (n= 364)

10.0 (7.5 – 13.1)

126.4 (89.3 – 173.7)

38.6 (22.1 – 61.1)

11.8 (9.7 – 13.9)

p value

0.614

0.503

0.224

0.066

VAT < median (n= 494)

9.1 (7.2 – 11.7)

121.4 (85.9 – 166.1)

38.7 (23.2 – 57.9)

11.7 (9.7 – 14.2)

VAT > median (n= 515)

9.7 (7.4 – 12.8)

123.8 (85.1 – 172.9)

35.6 (19.7 – 57.4)

11.4 (9.4 – 13.7)

p value

0.540

0.765

0.129

0.257

Obesity

Visceral obesity*

Visceral adipose tissue

†

Subcutaneous adipose tissue

‡

SAT < median (n= 484)

9.8 (7.6 – 12.6)

124.3 (87.6 – 170.6)

37.6 (22.5 – 56.8)

11.8 (9.6 – 14.5)

SAT > median (n= 525)

9.1 (7.1 – 11.7)

121.0 (84.0 – 167.8)

36.8 (21.4 – 58.3)

11.3 (9.5 – 13.4)

p value

0.744

0.335

0.002

0.512

No (565)

9.4 (7.3 – 12.1)

120.9 (87.4 – 166.8)

35.4 (20.7 – 56.7)

10.9 (9.1 – 13.1)

Yes (417)

9.4 (7.2 – 12.1)

124.2 (82.5 – 174.4)

40.2 (23.8 – 60.9)

12.4 (10.3 – 14.8)

p value

0.981

0.344

0.013

< 0.001

No (n= 492)

9.0 (7.1 – 11.4)

121.0 (81.7 – 167.2)

36.5 (21.9 – 56.6)

11.5 (9.4 – 13.9)

Yes (n=517)

9.8 (9.1 – 12.8)

124.1 (88.99 – 170.4)

37.6 (22.0 – 58.7)

11.6 (9.6 – 13.8)

p value

0.010

0.450

0.572

0.499

No (n= 850)

9.3 (7.3 – 11.9)

124.9 (87.9 – 172.9)

37.1 (21.9 – 57.4)

11.5 (9.4 – 13.9)

Yes (n= 158)

10.1 (7.4 – 13.2)

111.0 (78.4 – 158.5)

36.9 (22.2 – 59.8)

12.0 (9.9 – 13.8)

p value

0.050

0.010

0.909

0.111

No (n= 745)

9.0 (7.1 – 11.5)

122.3 (86.9 – 168.4)

36.0 (21.6 – 55.7)

11.3 (9.3 – 13.5)

Yes (n= 86)

9.5 (7.3 – 12.7)

122.9 (88.7 – 163.9)

34.7 (20.9 – 55.6)

11.0 (11.1 – 13.8)

p value

0.768

0.768

0.889

< 0.001

Hyperlipidaemia**

Metabolic syndrome

Baseline type 2 diabetes

Development of type 2 diabetes

P value is calculated by analysis of covariance, corrected for age, sex and eGFR. P values are calculated on log or square root transformed EV-marker values, but for ease of interpretation, the results of calculations are transformed back to original baseline levels of EV markers in pg/µg. *Visceral obesity is defined as waist circumference above 88 cm for females, or above 102 cm for males. **Hyperlipidaemia is defined as total cholesterol > 5.0 mmol/L, LDL-cholesterol > 3.2 mmol/L and /or use of lipid lowering agents. †Visceral adipose tissue values are sex pooled. For males median = 9.9 cm, for females median = 8.3 cm. ‡ Subcutaneous adipose tissue values are sex pooled. For males median = 2.3 cm, for females median = 3.0 cm.

122

EV-markers, obesity and metabolic complications

Low HDL-cholesterol was significantly related to higher EV-cystatin C levels (β -11.26; 95%CI -18.39 – -4.13), while high HDL cholesterol was significantly related to higher EVCD14 levels (β 5.04; 95%CI 0.07 – 10.00; Figure 2, Supplementary table 2).

EV-markers in relation to incident type 2 diabetes During a median follow up of 6.5 years (interquartile range 5.8-7.1 years), 42 patients developed type 2 diabetes. As shown in figure 4, higher levels of EV-markers were not related to an increased risk for type 2 diabetes. In fact, in patients with high EV-CD14 levels at baseline, a relative risk reduction of 16% for development of type 2 diabetes was observed (HR 0.84, 95%CI 0.75–0.94; Figure 4). Unadjusted and partially adjusted multivariable models are provided in supplementary table 4.

EV-markers in relation to the metabolic syndrome Higher EV-cystatin C levels were significantly related to a 57% higher prevalence of the metabolic syndrome (OR 1.57, 95%CI 1.19–2.27) while no relation was observed between the other EV markers and prevalence of the metabolic syndrome (Figure 3). Unadjusted and partially adjusted multivariable models are provided in supplementary table 3.

6

Incident type 2 diabetes

# of events

Cystatin C

42

Serpin G1

42

Serpin F2

42 42

CD14 0.7

0.8

0.9

1.0

1.1

1.2

Hazard ratio (95% confidence interval)

Figure 4. EV-markers and the risk of new onset type 2 diabetes in patients with manifest cardio­ vascular disease.

Hazard ratios with 95%CI indicate the relative risk for incident type 2 diabetes per increase in EV-cystatin C, EV-serpin G1, EV-serpin F2 or EVCD14 concentration during 6.5 years (interquartile range 5.8–7.1 years) follow-up, adjusted for age, sex, current smoking, hsCRP, eGFR and year of inclusion in SMART.

123

CHAPTER 6

DISCUSSION In the present study, we assessed the relation between AT quantity, AT distribution and metabolic parameters of AT (dys)function and plasma levels of four CVD-associated EVmarkers. Furthermore, the relation between these EV-markers and metabolic syndrome or incident type 2 diabetes in patients with clinically manifest vascular disease was investigated. We show that EV-cystatin C was positively related to metabolic complications of obesity, including low-grade systemic inflammation, low HDL-cholesterol levels and metabolic syndrome. In contrast, EV-CD14 was inversely related to AT abundance and dyslipidaemia, and was moreover related to a relative risk reduction for the development of type 2 diabetes.

Cystatin C has previously been shown to be elevated in obese subjects (35), to be secreted

by AT in vitro (36) and to be associated with the metabolic syndrome (37). Furthermore, cystatin C has been associated with prediabetes and cardiovascular disease, independent of renal function (18,38,39). In contrast to other studies performed in patients without known cardiovascular disease, we could not demonstrate a relation between obesity and circulating EV-cystatin C-levels, nor were HOMA-IR levels significantly related to EV-cystatin C levels. As we have specifically studied a cohort of patients with clinically manifest vascular disease, a difference in cohort characteristics between these studies may account for discordant results. Nonetheless, we did find strong relations between alternative parameters of adipose tissue dysfunction with EV-cystatin C levels, such as low-grade systemic inflammation and low HDL levels. These findings may suggest that AT dysfunction rather than AT abundance is a more important determinant of EV-cystatin C levels, at least in patients with cardiovascular disease. Furthermore, in concordance with studies performed in healthy individuals (37), we observed a strong relation between EV-cystatin C and the metabolic syndrome in patients with clinically manifest vascular disease. Thus, the EV-marker cystatin C may be an important biomarker for CVD not only in healthy individuals but importantly also in patients with manifest vascular disease.

Since obesity is associated with a pro-thrombotic state, we hypothesized that obesity could contribute to circulating EV-serpin G1 and EV-serpin F2, both pro-coagulant markers. Serpin G1, better known as C1 inhibitor, is primarily involved in the inhibition of coagulation and atherosclerotic plaque formation (40), though its role in obesity has not been investigated.

Serpin F2, or Îą-2-antiplasmin, is a major inhibitor of plasmin and thereby controls the coagulation system (22). Previous studies reported a negative relation of VAT thickness and soluble plasma serpin F2 levels (41). In our study, obesity was not related to EV-serpin F2 nor to EV-serpin G1, and both markers were not related to metabolic complications of obesity. However, EV-serpin G1 and EV-serpin F2 did show a strong positive relation with low-grade inflammation. These data suggest that in patients with manifest vascular disease, these markers might contribute to low grade inflammation, though both EV-serpin G1 and EV-serpin F2 appear to play no role in the pathophysiology of metabolic complications.

Surprisingly, we observed an inverse relation between EV-CD14 levels with obesity and obesity-related metabolic complications in patients with clinically manifest vascular disease.

124

EV-markers, obesity and metabolic complications

CD14 is expressed primarily by monocytes, which play important roles in obesity, obesityinduced AT inflammation and insulin resistance (5,42). CD14 has furthermore been associated with the development of atherosclerosis and the recurrence of vascular events (18,43). EVs secreted by monocytes express CD14, and these EVs are capable of inducing endothelial damage in vitro (23). However, conflicting results have been reported by others, as soluble CD14 did not relate to endothelial damage in type 2 diabetic subjects (44), and a recent study showed that lower soluble CD14 levels were associated with an increase in BMI in both obese and non-obese patients (45). Possibly, two different forms of CD14 studied (soluble CD14 versus membrane-bound CD14) may account for differences in observed relations between obesity and CD14. Although both soluble and membrane bound CD14 are involved in inflammatory signalling pathways, CD14 is only part of a receptor complex and soluble CD14 needs binding to a cellular signal-transducing receptor in order for cell activation to occur (46). Furthermore, an excess of circulating soluble CD14 is believed to inhibit monocyte responses to inflammatory signals for membrane CD14 (42,46). It is unclear whether EVs contain soluble or membrane bound CD14, and whether EV-associated CD14 is functional. Nonetheless, as EV-CD14 levels were associated with low grade inflammation mirrored by circulating hsCRP levels, it is tempting to speculate that high EV-CD14 levels might contribute to vascular risk via inflammatory pathways, but not via obesity-induced metabolic complications.

The mechanisms by which EV-markers could influence the development of type 2 diabetes remain elusive. EVs are regarded as tailor-made messengers for intercellular communication, as their unique composition allows transfer of signalling molecules to a wide variety of target tissues (17). However, even though a negative relation was observed between EVCD14 and incident type 2 diabetes in this study, this does not necessarily imply a direct role for CD14 in reduced progression of development of type 2 diabetes. Considering that CD14 is only part of a receptor complex, the functional role of either soluble or membrane CD14 in EVs remains elusive. It could be hypothesized that EVs containing high levels of CV14 are also enriched for insulin sensitizing molecules like adiponectin, as circulating adiponectin levels were positively related to circulating EV-CD14 levels in this study (supplementary table 2). Further studies are needed to evaluate the pathophysiological role of EV-CD14 in the development of metabolic complications of obesity.

As levels of EV-cystatin C and EV-CD14 can be partly explained by AT abundance and AT (dys)function, we questioned whether AT could actively secrete both EV-markers or whether AT (dysfunction) triggers other tissues for their release. Cell types potentially involved in the release of EV-cystatin C and EV-CD14 include activated monocytes, endothelial cells and platelets, which are all present in high numbers in atherosclerosis lesions (43,47).

However, evidence suggests that these cell types also play active roles in AT dysfunction, in which hypertrophic adipocytes induce endothelial stress and the recruitment of monocytes (5). Furthermore, production of both CD14 and cystatin C is increased in AT of obese compared to lean subjects (36,42). Therefore, AT itself could be capable of secreting these markers due to adipocyte hypertrophy, hypoxia and increased influx of proinflammatory cells such as macrophages. As AT is capable of secreting functional EVs as

125

6

CHAPTER 6

shown in mice and humans (48,49), it will be interesting to study whether these AT EVs contain the EV-markers assessed in the present study.

Strengths of this study include the large sample size of a well-characterized and relevant patient population, which allowed for adjustment of multiple relevant potential confounding factors. Furthermore, measurement of different fat compartments with ultrasound allowed for the assessment of the contribution of the different adipose tissues depots to the levels of circulating EV-markers, known to be associated with CVD. Limitations of this study include the fact that, due to the cross-sectional design, causality in the relationships remain unknown. Second, the study population consisted solely of patients with clinically manifest vascular disease, which may limit the generalization of the results to other cohorts.

Conclusions In patients with clinically manifest vascular disease, EV-cystatin C positively relates to metabolic complications of obesity, and may thus contribute to an increased cardiovascular risk through obesity associated metabolic dysfunction. In contrast, EV-CD14 levels were inversely related to visceral obesity in males and associated with a relative risk reduction for the development of type 2 diabetes.

Acknowledgments We gratefully acknowledge the members of the SMART study group of UMC Utrecht: P.A. Doevendans, MD, PhD, Department of Cardiology; A. Algra, MD, PhD; Y. van der Graaf, MD, PhD; D.E. Grobbee, MD, PhD, G.E.H.M. Rutten, MD, PhD, Julius Center for Health Sciences and Primary Care; L.J. Kappelle, MD, PhD, Department of Neurology; W.P.T.M. Mali, MD, PhD, Department of Radiology; F.L. Moll, MD, PhD, Department of Vascular Surgery; F.L.J. Visseren, MD, PhD, Department of Vascular Medicine. We further thank W.A. Scheper for critically reviewing the manuscript, and M. Weijmans for help with statistical analyses.

126

EV-markers, obesity and metabolic complications

REFERENCES 1.

Hajer GR, van Haeften TW, Visseren FL: Adipose tissue dysfunction in obesity, diabetes, and vascular diseases. Eur Heart J 2008, 29(24):2959-2971.

2. Yusuf S, Hawken S, Ounpuu S, Bautista L, Franzosi MG, Commerford P, Lang CC, Rumboldt Z, Onen CL, Lisheng L, Tanomsup S, Wangai P,Jr, Razak F, Sharma AM, Anand SS, INTERHEART Study Investigators: Obesity and the risk of myocardial infarction in 27,000 participants from 52 countries: a case-control study. Lancet 2005, 366(9497):1640-1649.

3. Nicklas BJ, Penninx BW, Cesari M, Kritchevsky SB, Newman AB, Kanaya AM, Pahor M, Jingzhong D, Harris TB, Health, Aging and Body Composition Study: Association of visceral adipose tissue with incident myocardial infarction in older men and women: the Health, Aging and Body Composition Study. Am J Epidemiol 2004, 160(8):741-749.

4. Kanhai DA, Kappelle LJ, van der Graaf Y, Uiterwaal CS, Visseren FL, SMART Study Group: The risk of general and abdominal adiposity in the occurrence of new vascular events and mortality in patients with various manifestations of vascular disease. Int J Obes (Lond) 2012, 36(5):695702.

5. Xu H, Barnes GT, Yang Q, Tan G, Yang D, Chou CJ, Sole J, Nichols A, Ross JS, Tartaglia LA, Chen H: Chronic inflammation in fat plays a crucial role in the development of obesity-related insulin resistance. J Clin Invest 2003, 112(12):1821-1830.

6. Kintscher U, Hartge M, Hess K, Foryst-Ludwig A, Clemenz M, Wabitsch M, Fischer-Posovszky P, Barth TF, Dragun D, Skurk T, Hauner H, Bluher M, Unger T, Wolf AM, Knippschild U, Hombach V, Marx N: T-lymphocyte infiltration in visceral adipose tissue: a primary event in adipose tissue inflammation and the development of obesity-mediated insulin resistance. Arterioscler Thromb Vasc Biol 2008, 28(7):1304-1310.

7. Faber DR, van der Graaf Y, Westerink J, Visseren FL: Increased visceral adipose tissue mass is associated with increased C-reactive protein in patients with manifest vascular diseases.

Atherosclerosis 2010, 212(1):274-280.

8. Berg AH, Combs TP, Scherer PE: ACRP30/adiponectin: an adipokine regulating glucose and lipid metabolism. Trends Endocrinol Metab 2002, 13(2):84-89.

9. Olefsky JM, Glass CK: Macrophages, inflammation, and insulin resistance. Annu Rev Physiol 2010, 72:219-246.

10. Grundy SM, Brewer HB,Jr, Cleeman JI, Smith SC,Jr, Lenfant C, American Heart Association, National Heart, Lung, and Blood Institute: Definition of metabolic syndrome: Report of the National Heart, Lung, and Blood Institute/American Heart Association conference on scientific issues related to definition. Circulation 2004, 109(3):433-438.

11. Kuller LH, Velentgas P, Barzilay J, Beauchamp NJ, O'Leary DH, Savage PJ: Diabetes mellitus: subclinical cardiovascular disease and risk of incident cardiovascular disease and all-cause mortality. Arterioscler Thromb Vasc Biol 2000, 20(3):823-829.

12. Karastergiou K, Smith SR, Greenberg AS, Fried SK: Sex differences in human adipose tissues the biology of pear shape. Biol Sex Differ 2012, 3(1):13-6410-3-13.

13. Musaad S, Haynes EN: Biomarkers of obesity and subsequent cardiovascular events. Epidemiol Rev 2007, 29:98-114.

14. Liu ML, Williams KJ: Microvesicles: potential markers and mediators of endothelial dysfunction.

Curr Opin Endocrinol Diabetes Obes 2012, 19(2):121-127.

15. Muller G: Microvesicles/exosomes as potential novel biomarkers of metabolic diseases.

Diabetes Metab Syndr Obes 2012, 5:247-282.

16. Thery C, Ostrowski M, Segura E: Membrane vesicles as conveyors of immune responses. Nat Rev Immunol 2009, 9(8):581-593.

17. Nolte-'t Hoen EN, Wauben MH: Immune cell-derived vesicles: modulators and mediators of inflammation. Curr Pharm Des 2012, 18(16):2357-2368.

18. Kanhai DA, Visseren FL, van der Graaf Y, Schoneveld AH, Catanzariti LM, Timmers L, Kappelle LJ, Uiterwaal CS, Lim SK, Sze SK, Pasterkamp G, de Kleijn DP, SMART Study Group: Microvesicle protein levels are associated with increased risk for future vascular events and mortality in patients with clinically manifest vascular disease. Int J Cardiol 2013, 168(3):2358-2363.

127

6

CHAPTER 6

19. Lafarge JC, Naour N, Clement K, Guerre-Millo M: Cathepsins and cystatin C in atherosclerosis and obesity. Biochimie 2010, 92(11):1580-1586.

20. Fu Z, Xue H, Guo J, Chen L, Dong W, Gai L, Liu H, Sun Z, Chen Y: Long-term prognostic impact of cystatin c on acute coronary syndrome octogenarians with diabetes mellitus. Cardiovasc Diabetol 2013, 12(1):157.

21. Schousboe I: Binding of activated Factor XII to endothelial cells affects its inactivation by the C1-esterase inhibitor. Eur J Biochem 2003, 270(1):111-118.

22. Collen D, Lijnen HR: The fibrinolytic system in man. Crit Rev Oncol Hematol 1986, 4(3):249-301.

23. Aharon A, Tamari T, Brenner B: Monocyte-derived microparticles and exosomes induce procoagulant and apoptotic effects on endothelial cells. Thromb Haemost 2008, 100(5):878885.

24. Fantuzzi G, Mazzone T: Adipose tissue and atherosclerosis: exploring the connection.

Arterioscler Thromb Vasc Biol 2007, 27(5):996-1003.

25. Rocha VZ, Folco EJ, Sukhova G, Shimizu K, Gotsman I, Vernon AH, Libby P: Interferon-gamma, a Th1 cytokine, regulates fat inflammation: a role for adaptive immunity in obesity. Circ Res 2008, 103(5):467-476.

26. Napoleone E, DI Santo A, Amore C, Baccante G, di Febbo C, Porreca E, de Gaetano G, Donati MB, Lorenzet R: Leptin induces tissue factor expression in human peripheral blood mononuclear cells: a possible link between obesity and cardiovascular risk? J Thromb Haemost 2007, 5(7):1462-1468.

27. Lorenzet R, Napoleone E, Cutrone A, Donati MB: Thrombosis and obesity: cellular bases.

Thromb Res 2012, 129(3):285-289.

28. Simons PC, Algra A, van de Laak MF, Grobbee DE, van der Graaf Y: Second manifestations of ARTerial disease (SMART) study: rationale and design. Eur J Epidemiol 1999, 15(9):773-781.

29. de Jager W, te Velthuis H, Prakken BJ, Kuis W, Rijkers GT: Simultaneous detection of 15 human cytokines in a single sample of stimulated peripheral blood mononuclear cells. Clin Diagn Lab Immunol 2003, 10(1):133-139.

30. Stolk RP, Meijer R, Mali WP, Grobbee DE, van der Graaf Y, Secondary Manifestations of Arterial Disease Study Group: Ultrasound measurements of intraabdominal fat estimate the metabolic syndrome better than do measurements of waist circumference. Am J Clin Nutr 2003, 77(4):857-860.

31. Schipper HS, de Jager W, van Dijk ME, Meerding J, Zelissen PM, Adan RA, Prakken BJ, Kalkhoven E: A multiplex immunoassay for human adipokine profiling. Clin Chem 2010, 56(8):1320-1328.

32. Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC: Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia 1985, 28(7):412-419.

33. Bland JM, Altman DG: Transformations, means, and confidence intervals. BMJ 1996, 312(7038):1079.

34. Fantuzzi G: Adiponectin and inflammation: consensus and controversy. J Allergy Clin Immunol 2008, 121(2):326-330.

35. Muntner P, Winston J, Uribarri J, Mann D, Fox CS: Overweight, obesity, and elevated serum cystatin C levels in adults in the United States. Am J Med 2008, 121(4):341-348.

36. Naour N, Fellahi S, Renucci JF, Poitou C, Rouault C, Basdevant A, Dutour A, Alessi MC, Bastard JP, Clement K, Guerre-Millo M: Potential contribution of adipose tissue to elevated serum cystatin C in human obesity. Obesity (Silver Spring) 2009, 17(12):2121-2126.

37. Magnusson M, Hedblad B, Engstrom G, Persson M, Nilsson P, Melander O: High levels of cystatin C predict the metabolic syndrome: the prospective Malmo Diet and Cancer Study. J Intern Med 2013, 274(2):192-199.

38. Sabanayagam C, Wong TY, Xiao J, Shankar A: Serum cystatin C and prediabetes in non-obese US adults. Eur J Epidemiol 2013, 28(4):311-316.

39. Qing X, Furong W, Yunxia L, Jian Z, Xuping W, Ling G: Cystatin C and asymptomatic coronary artery disease in patients with metabolic syndrome and normal glomerular filtration rate.

Cardiovasc Diabetol 2012, 11:108-2840-11-108.

128

EV-markers, obesity and metabolic complications

40. Cai S, Davis AE,3rd: Complement regulatory protein C1 inhibitor binds to selectins and interferes with endothelial-leukocyte adhesion. J Immunol 2003, 171(9):4786-4791.

41. Kozek E, Katra B, Malecki M, Sieradzki J: Visceral obesity and hemostatic profile in patients with type 2 diabetes: the effect of gender and metabolic compensation. Rev Diabet Stud 2004, 1(3):122-128.

42. Fernandez-Real JM, Perez del Pulgar S, Luche E, Moreno-Navarrete JM, Waget A, Serino M, Sorianello E, Sanchez-Pla A, Pontaque FC, Vendrell J, Chacon MR, Ricart W, Burcelin R, Zorzano A: CD14 modulates inflammation-driven insulin resistance. Diabetes 2011, 60(8):2179-2186.

43. Poitou C, Dalmas E, Renovato M, Benhamo V, Hajduch F, Abdennour M, Kahn JF, Veyrie N, Rizkalla S, Fridman WH, Sautes-Fridman C, Clement K, Cremer I: CD14dimCD16+ and CD14+CD16+ monocytes in obesity and during weight loss: relationships with fat mass and subclinical atherosclerosis. Arterioscler Thromb Vasc Biol 2011, 31(10):2322-2330.

44. Fernandez-Real JM, Lopez-Bermejo A, Castro A, Broch M, Penarroja G, Vendrell J, Vazquez G, Ricart W: Opposite relationship between circulating soluble CD14 concentration and endothelial function in diabetic and nondiabetic subjects. Thromb Haemost 2005, 94(3):615-619.

45. Koethe JR, Dee K, Bian A, Shintani A, Turner M, Bebawy S, Sterling TR, Hulgan T: Circulating interleukin-6, soluble CD14, and other inflammation biomarker levels differ between obese and nonobese HIV-infected adults on antiretroviral therapy. AIDS Res Hum Retroviruses 2013, 29(7):1019-1025.

46. Kirkland TN, Viriyakosol S: Structure-function analysis of soluble and membrane-bound CD14.

Prog Clin Biol Res 1998, 397:79-87.

47. Yin W, Ghebrehiwet B, Peerschke EI: Expression of complement components and inhibitors on platelet microparticles. Platelets 2008, 19(3):225-233.

48. Deng ZB, Poliakov A, Hardy RW, Clements R, Liu C, Liu Y, Wang J, Xiang X, Zhang S, Zhuang X, Shah SV, Sun D, Michalek S, Grizzle WE, Garvey T, Mobley J, Zhang HG: Adipose tissue exosome-like vesicles mediate activation of macrophage-induced insulin resistance. Diabetes 2009, 58(11):2498-2505.

49. Kranendonk ME, Visseren FL, van Balkom BW, Hoen EN, van Herwaarden JA, de Jager W, Schipper HS, Brenkman AB, Verhaar MC, Wauben MH, Kalkhoven E: Human adipocyte extracellular vesicles in reciprocal signaling between adipocytes and macrophages Human adipocyte EVs and macrophages. Obesity (Silver Spring) 2013, doi: 10.1002/oby.20679.

129

6

130

-0.35 (-1.18 – 0.49)

-0.90 (-1.81 – 0.01)

I

IV

0.08 (-0.22 – 0.38)

IV

-0.40 (-1.23 – 0.42)

0.11 (-0.18 – 0.41)

III

-0.39 (-1.26 – 0.48)

0.10 (-0.22 – 0.41)

II

III

0.17 (-0.16 – 0.49)

I

II

-0.11 (-2.49 – 2.26)

IV

-2.17 (-4.75 – 0.42)

I

0.59 (-1.77 – 2.94)

0.25 (-0.93 – 1.44)

IV

0.10 (-2.42 – 2.62)

0.41 (-0.77 – 1.58)

III

III

0.79 (-0.46 – 2.03)

II

II

1.00 (-0.30 – 2.31)

-0.05 (-1.06 – 0.97)

-0.09 (-0.91 – 1.09)

-0.14 (-1.14 – 0.86)

-0.14 (-1.14 – 0.86)

0.05 (-0.31 – 0.41)

0.15 (-0.21 – 0.50)

0.06 (-0.29 – 0.28)

0.10 (-0.26 – 0.45)

-2.26 (-5.15 – 0.62)

-1.63 (-4.49 – 1.24)

-1.77 (-4.66 – 1.11)

-1.56 (-4.41 – 1.29)

0.87 (-0.56 – 2.29)

1.24 (-0.18 – 2.65)

1.05 (-0.37 – 2.46)

1.08 (-0.35 – 2.50)

β (95%CI)

-0.52 (-1.08 – 0.03)

-0.79 (-1.35 – -0.23)* -4.04 (-8.58 – 0.51)

-0.74 (-1.30 – -0.18)* -5.30 (-9.80 – -0.80)*

-0.95 (-1.52 – -0.37)* -5.38 (-9.83 – -0.93)*

0.05 (-0.24 – 0.15) -5.61 (-10.08 – -1.13)*

-0.09 (-0.29 – 0.11)

-0.08 (-0.28 – 0.12)

0.02 (-0.23 – 0.19)

-0.11 (-1.70 – 1.48)

0.53 (-1.08 – 2.15)

0.43 (-1.19 – 2.05)

-0.51 (-2.17 – 1.14)

-0.95 (-1.73 – -0.17)*

-1.02 (-1.82– -0.23)*

-0.90 (-1.69 – -0.11)*

-0.90 (-2.52 – 0.72)

-1.07 (-2.67 – 0.54)

-1.12 (-2.70 – 0.47)

-0.90 (-2.51 – 0.71)

-2.74 (-15.76 – 10.28)

0.78 (-12.19 – 13.76)

0.61 (-12.34 – 13.55)

-0.41 (-13.23 – 12.42)

-3.28 (-9.71 – 3.15)

-3.69 (-10.09 – 2.70)

-3.74 (-10.07 – 2.59)

-0.78 (-1.61 – 0.05)

β (95%CI)

β (95%CI) -3.43 (-9.85 – 2.99)

Mean = 2.43

N = 802

Log CD14 (pg/µg)

Mean = 5.98

N = 797

Sqrt serpin F2 (pg/µg)

Linear regression coefficients (β) with 95% confidence interval (CI) indicates the difference in log cystatin C, serpin G1, CD14 or square root EV serpin F2 concentration per unit increase in adipose tissue parameter. Estimated differences are based on linear regression models. Model I: univariable model; Model II: adjustment for age and current smoking; Model III: model II with additional adjustment for eGFR and type 2 diabetes; Model IV: model III with additional adjustment for blood pressure lowering medication, lipid lowering medication and year of inclusion. Bold values indicate significance, *p<0.05

BMI Mean = 26.8 kg/m2

Waist circumference Mean = 97 cm

Subcutaneous fat thickness Mean = 2.3 cm

Visceral fat thickness Mean = 9.9 cm

β (95%CI)

Mean = 4.82

Mean = 2.22

I

N = 801

N = 799

Model

Log serpin G1(pg/µg)

Log cystatin C (pg/µg)

Supplementary Table 1a. Relation between visceral fat, subcutaneous fat, waist circumference or BMI and EV-markers in males with clinically manifest arterial disease.

CHAPTER 6

SUPPLEMENTARY INFORMATION

1.44 (-1.79 – 4.68)

-0.77 (-3.91 – 2.37)

0.44 (-0.07 – 0.95)

0.30 (-0.18 – 0.77)

0.24 (-0.19 – 0.66)

0.10 (-0.31 – 0.51)

0.94 (-0.34 – 2.21)

0.74 (-0.45 – 1.92)

0.39 (-0.65 – 1.43)

0.26 (-0.74 – 1.26)

IV

I

II

III

IV

I

II

III

IV

-1.12 (-5.09 – 2.85)

I

III

1.41 (-0.62 – 3.44)

IV

0.50 (-3.22 – 4.22)

1.51 (-0.65 – 3.68)

III

II

2.06 (-0.32 – 4.43)

II

0.81 (-0.59 – 2.22)

0.82 (-0.55 – 2.18)

0.85 (-0.50 – 2.20)

0.84 (-0.52 – 2.20)

0.38 (-0.19 – 0.95)

0.38 (-0.18 – 0.93)

0.35 (-0.19 – 0.89)

0.33 (-0.21 – 0.87)

-1.99 (-6.42 – 2.45)

-1.76 (-6.01 – 2.49)

-1.20 (-6.22 – 2.23)

-2.34 (-6.56 – 1.88)

2.56 (-0.31 – 5.42)

2.59 (-0.25 – 5.43)

Log CD14 (pg/µg) N = 207 Mean = 2.48 β (95%CI) 0.40 (-1.29 – 2.09) -0.12 (-1.78 – 1.55) -0.35 (-2.09 – 1.39) -0.34 (-2.04 – 1.35) -2.47 (-5.09 – 0.14) -1.76 (-4.34 – 0.81) -1.67 (-4.25 – 0.92) -2.84 (-5.42 – -0.26)* -0.10 (-0.44 – 0.24) -0.16 (-0.48 – 0.18) -0.19 (-0.53 – 0.15) -0.23 (-0.57 – 0.10) -0.40 (-1.25 – 0.45) -0.47 (-1.30 – 0.35) -0.56 (-1.39 – 0.27) -0.54 (-1.37 – 0.28)

Sqrt serpin F2 (pg/µg) N = 206 Mean = 6.51 β (95%CI) -0.37 (-14.91 – 14.17) 3.17 (-11.19 – 17.54) -0.30 (-15.35 – 14.75) -0.66 (-15.31 – 14.00) 25.28 (2.91 – 47.66)* 22.62 (0.54 – 44.69)* 20.90 (-1.29 – 43.09) 10.91 (-11.59 – 33.41) 1.54 (-1.36 – 4.44) 2.18 (-0.66 – 5.01) 1.75 (-1.16 – 4.66) 1.10 (-1.80 – 4.01) 4.24 (-3.04 – 11.51) 5.08 (-2.01 – 12.17) 4.54 (-2.61 – 11.69) 4.04 (-3.08 – 11.16)

Linear regression coefficients (β) with 95% confidence interval (CI) indicates the difference in log cystatin C, serpin G1, CD14 or square root EV serpin F2 concentration per unit increase in adipose tissue parameter. Estimated differences are based on linear regression models. Model I: univariable model; Model II: adjustment for age and current smoking; Model III: model II with additional adjustment for eGFR and type 2 diabetes; Model IV: model III with additional adjustment for blood pressure lowering medication, lipid lowering medication and year of inclusion. Bold values indicate significance, *p<0.05.

BMI Mean = 26.8 kg/m2

Waist circumference Mean = 97 cm

Subcutaneous fat thickness Mean = 2.3 cm

Visceral fat thickness Mean = 9.9 cm

2.47 (-0.24 – 5.19)

3.20 (0.69 – 5.71)*

I 2.38 (-0.34 – 5.10)

β (95%CI)

Mean = 4.73

Mean = 2.28

β (95%CI)

N = 206

N = 207

Model

Log serpin G1(pg/µg)

Log cystatin C (pg/µg)

Supplementary Table 1b. Relation between visceral fat, subcutaneous fat, waist circumference or BMI and EV-markers in females with clinically manifest arterial disease.

EV-markers, obesity and metabolic complications

131

6

132 -5.40 (-12.45 – 1.64) -6.80 (-15.18 – 1.58) -2.10 (-10.87 – 6.67) -2.55 (-11.43 – 6.33) -0.34 (-9.26 – 8.59)

2.31 (-3.20 – 7.82) -10.45 (-18.07 – -2.82)* -18.49 (-26.06 – -10.91)** -13.10 (-20.27 – -5.94)** -11.26 (-18.39 – -4.13)*

I

II

III

IV

-5.50 (-12.45 – 1.45)

2.61 (-2.83 – 8.05)

IV

-4.56 (-11.27 – 2.15)

-3.51 (-10.22 – 3.21)

III

7.10 (1.12– 13.07)*

I

0.58 (-3.93 – 2.76)

-0.74 (-4.09 – 2.61)

6.85 (1.18 – 12.52)*

0.61 (-2.08 – 3.31)

IV

-0.01 (-3.35 – 3.33)

II

0.65 (-2.08 – 3.38)

III

2.12 (-0.88 – 5.11)

I

-1.46 (-4.73 – 1.82)

7.96 (4.76 – 11.16)**

5.59 (3.00 – 8.18)**

IV

0.21 (-2.71 – 3.13)

8.58 (5.41 – 11.75)**

6.18 (3.59 – 8.77)**

III

II

8.83 (5.66 – 12.00)**

10.00 (6.90 – 13.11)**

7.82 (5.04 – 10.59)**

9.33 (6.50 – 12.17)**

II

I

28.96 (-12.54 – 70.46)

22.16 (-19.46 – 63.78)

20.41 (-20.35 –61.17)

20.83 (-18.34 – 60.01)

-20.23 (-54.81 – 14.35)

-21.60 (-55.52 – 12.33)

-13.21 (-45.85 – 19.44)

-10.97 (-43.67 – 21.73)

-0.108 (-16.00 – 15.38)

1.90 (-13.75 – 17.54)

1.47 (-13.97 – 16.91)

1.73 (-13.50 – 16.96)

36.52 (21.49 – 51.55)**

37.05 (22.07 – 52.03)**

36.71 (21.82 – 51.60)**

5.04 (0.07 – 10.00)*

4.09 (-0.98 – 9.16)

2.44 (-2.56 – 7.44)

4.43 (-0.50 – 9.37)

-3.71 (-8.16 – 0.74)

-4.73 (-9.18 – -0.28)

-3.15 (-7.45 – 1.16)

-2.78 (-7.18 – 1.61)

1.49 (-0.36 – 3.35)

1.97 (0.07– 3.87)*

1.72 (-0.17 – 3.61)

2.25 (0.33 – 4.17)*

5.72 (3.94 – 7.49)**

6.08 (4.28 – 7.87)**

6.34 (4.54 – 8.13)**

7.72 (5.91 – 9.53)**

β (95%CI)

42.18 (27.49 – 56.86)**

Mean = 2.44

β (95%CI)

N = 1009

Log CD14 (pg/µg)

Mean = 6.09

N = 1003

Sqrt serpin F2 (pg/µg)

Relation between metabolic parameters of adipose tissue (dys)function and EV-markers in patients with manifest cardiovascular disease. Linear regression coefficients (β) with 95% CI indicates the difference in log EV-cystatin C, log EV-serpin G1, square root EV-serpin F2 or log EV-CD14 concentration per unit increase in log hsCRP, log adiponectin, log HOMA-IR or HDL-cholesterol. Model I: univariable model, Model II: adjusted for age, sex and current smoking, Model III: additional adjustment for eGFR, type 2 diabetes, Model IV: additional adjustment for blood pressure lowering medication, lipid lowering medication, platelet aggregation inhibitors and year of inclusion in SMART. Bold values indicate significance, *p<0.05; ** p<0.0001.NB The relation between HOMA-IR and EVcystatin C was only performed in patients without use of anti-hyperglycaemic drugs.

HDL-C N= 1012 Median = 1.3 (1.0 – 1.5)

Log HOMA-IR N = 464 Mean log= 0.87

Log adiponectin N = 987 Mean log= 16.93

Log hsCRP N = 992 Mean log = 0.61

β (95%CI)

Mean = 4.80

Mean = 2.24 β (95%CI)

N = 1007

N = 1006

Model

Log serpin G1 (pg/µg)

Log cystatin C (pg/µg)

Supplementary Table 2. Relation between plasma concentrations of hsCRP and EV-markers in patients with clinically manifest arterial disease.

CHAPTER 6

1.10 (0.87 – 1.41) 1.05 (0.82 – 1.35) 1.02 (0.79 – 1.30)

1.55 (1.18 - 2.05)* 1.61 (1.20 – 2.17)* 1.55 (1.12 – 2.14)*

II

III

Log CD14 (pg/µg) N = 1009 Mean = 2.44 OR (95%CI) 1.13 (0.75 – 1.72) 1.02 (0.66 – 1.58) 0.87 (0.56 – 1.37)

Sqrt serpin F2 (pg/µg) N = 1003 Mean = 6.09 OR (95%CI) 1.02 (0.96 – 1.07) 1.01 (0.95 – 1.06) 0.99 (0.94 – 1.05)

0.90 (0.81 – 1.00)* 0.87 (0.78 – 0.97)* 0.84 (0.75 – 0.94)*

0.99 (0.97 – 1.00) 0.99 (0.97 – 0.99)*

1.01 (0.94 – 1.09) 0.98 (0.90 – 1.07)

II

III

0.99 (0.99 – 1.00)

1.00 (0.99 – 1.00)

Hazard ratios with 95%CI indicate the relative risk for incident type 2 diabetes per increase in EV-cystatin C, EV-serpin G1, EV-serpin F2 or EV-CD14 concentration during 6.5 years (interquartile range 5.8–7.1 years) follow-up, based on multivariable Cox proportional hazard regression. Model I: univariable model, Model II: adjusted for age, sex, current smoking, Model II: additional adjustment for hsCRP, eGFR, year of inclusion in SMART, metabolic syndrome and HOMA-IR. Bold values indicate significance, *p<0.05.

Type 2 diabetes # of events 42

HR (95%CI)

1.00 (0.99 – 1.00)

HR (95%CI)

0.99 (0.98 – 1.00)

1.03 (0.96- 1.10)

HR (95%CI)

HR (95%CI)

Median 11.0 (9.1-13.2)

Median 31.1 (20.6-51.1)

Median 113.0 (82.4-155.5)

Median 8.9 (7.0-11.0)

I

N = 530

N = 524

N = 530

N = 528

Model

CD14 (pg/µg)

serpin F2 (pg/µg)

serpin G1 (pg/µg)

cystatin C (pg/µg)

Supplementary Table 4. Relation between EV-markers and risk of onset type 2 diabetes in patients with clinically manifest arterial disease.

Odds ratios with 95% CI indicate the odds for metabolic syndrome per increase in log cystatin C, serpin G1, CD14 or square root EV serpin F2 concentration.

Estimated differences are based on multivariable logistic regression models. Model I: univariable model; Model II: adjustment for age, sex and current smoking; Model III: model II with additional adjustment for hsCRP, eGFR and year of inclusion. Bold values indicate significance, *p<0.05.

Metabolic Syndrome No = 492 Yes = 517

OR (95%CI)

OR (95%CI)

Mean = 4.80

Mean = 2.24

I

N = 1007

N = 1006

Model

Log serpin G1 (pg/µg)

Log cystatin C (pg/µg)

Supplementary Table 3. Relation between EV-markers and the metabolic syndrome in patients with clinically manifest arterial disease.

EV-markers, obesity and metabolic complications

133

6

CHAPTER 7 Adiponectin and incident coronary heart disease and stroke.

A systematic review and meta-analysis of prospective studies Obesity Reviews, 2013;14(7):555-567

Danny A. Kanhai, MariĂŤtte E.G. Kranendonk, Cuno S.P.M. Uiterwaal, Yolanda van der Graaf, L. Jaap Kappelle, Frank L.J. Visseren

CHAPTER 7

ABSTRACT The plasma concentration of adiponectin, an adipokine that has anti-inflammatory, antiatherogenic and insulin sensitizing properties, is lower in obese subjects and could therefore be a target for therapy. In order to review and meta-analyse prospective cohort studies investigating adiponectin concentration and the risk for incident coronary heart disease (CHD) or stroke, a systematic search of MEDLINE, EMBASE and Cochrane databases was performed. Two independent reviewers selected prospective cohort studies investigating the relationship between adiponectin level and incident CHD or stroke using ‘adiponectin’ and ‘cardiovascular disease’ or ‘stroke’ and their synonyms, excluding patients with clinically manifest vascular disease. Random-effects models were used to calculate pooled relative risks (RRs) and 95% confidence intervals (95% CI). Generalized least squares regression was used to assess dose–response relationships for adiponectin concentrations from studies that provided RRs solely based upon categorical data regression. In total, 16 prospective cohort studies, comprising 23,919 patients and 6,870 CHD or stroke outcome events, were included in the meta-analyses. An increase of 1 standard deviation in logtransformed adiponectin did not lower the risk for CHD (RR 0.97; 95% CI 0.86–1.09). A 10 µg mL–1 increase in adiponectin conferred a RR of 0.91 (95% CI 0.80–1.03) for CHD and a RR 1.01 (95% CI 0.97–1.06) for stroke. In conclusion, plasma adiponectin is not related to the risk for incident CHD or stroke

136

Adiponectin and risk for future CHD/stroke

INTRODUCTION Cardiovascular disease is a major health problem worldwide, with coronary heart disease (CHD) and stroke as leading causes of burden of disease (1).

CHD and stroke are the leading causes of mortality globally. In 2008, an estimated 7.3 million people died from CHD and 6.2 million due to stroke, representing 24% of all global deaths (2). Therefore, it is of upmost importance to identify and characterize potential risk factors and prevention strategies for CHD as well as stroke. Adiponectin is a 244-aminoacid long polypeptide, which is abundantly secreted by adipocytes. Unlike other adipokines, plasma concentrations of adiponectin are lower in obesity (3) and adiponectin appears to have contra regulatory effects on inflammation, fibrosis, insulin resistance and atherosclerosis (4â&#x20AC;&#x201C;8). Adiponectin has antiatherogenic effects by down-regulating NFkBregulated endothelial adhesion molecule expression (4) and by inhibition of the transformation of macrophages into foam cells (5), which is an important step in the pathogenesis of atherosclerosis. Adiponectin has anti-inflammatory properties as it directly increases IL-10 mRNA expression, inducing the production of IL-10 (7), an important antiinflammatory mediator.

Even though the results of in vitro studies suggest a protective role for adiponectin in the development and progression of atherosclerosis (5,9,10), evidence from clinical studies with cardiovascular outcomes yielded conflicting results, varying from reduced risk (11,12) to an elevated risk for cardiovascular events (13) with increasing adiponectin plasma levels.

These inconsistent results might relate to the divergent vascular risk associated with obesity. The risk for myocardial infarction is more outspoken in obesity in relation to the risk for stroke (14). Additionally, heterogeneity in the patient groups studied, such as patients with or without pre-existent cardiovascular disease, most likely attribute to these inconsistent results. Elevated adiponectin in patients with existent cardiovascular disease could be the result of compensatory up-regulation that represents a defense mechanism of the body against metabolic changes including insulin resistance and a pro-inflammatory state associated with cardiovascular disease, adiponectin resistance or from intentional or unwanted weight loss (15).

The objective of this systematic review was to assess the relation between adiponectin plasma levels and the risk of CHD and stroke in prospective cohort studies in patients without manifest cardiovascular disease and to summarize the results in a meta-analysis.

METHODS Search strategy The meta-analysis was conducted according to the Meta-analysis of Observational Studies in Epidemiology guidelines (16). Two investigators (D.A.K, M.E.K.) independently identified articles through a systematic search of MEDLINE (PubMed), EMBASE and Cochrane databases up to 1 January 2013. For the MEDLINE search, the following terms were used:

137

7

CHAPTER 7

(adiponectin [MeSH Terms] OR adiponectin [All Fields] OR ADIPOQ[All Fields] OR ACDC [All Fields] OR ADPN [All Fields] OR APM1[All Fields] OR ACRP30 [All Fields] OR GBP28 [All Fields] OR (C1q [All Fields] AND (collagen [MeSH Terms] OR collagen [All Fields]) AND domain-containing [All Fields] AND (proteins [MeSH Terms] OR proteins [All Fields])) OR (30 [All Fields] AND kDa [All Fields] AND (adipocytes [MeSH Terms] OR adipocytes [All Fields]) AND complement-related [All Fields] AND (proteins [MeSH Terms] OR proteins [All Fields])) OR ‘adipose most abundant gene transcript 1’[All Fields] OR (Adipose [All Fields] AND specific [All Fields] AND collagen-like [All Fields] AND factor [All Fields])) AND ((‘cardiovasculardiseases’[MeSH Terms] OR (cardiovascular AND diseases [All Fields]) OR (coronary [all fields] AND (diseases OR ischemia OR ischaemia[all fields]))) OR (stroke [MeSH Terms] OR stroke [all fields] OR ((cerebrovascular [all fields] OR cerebral [all fields]) AND (event [all fields] OR accident [all fields] OR stroke [all fields] OR disease [all fields])) OR ((ischaemic [all fields] OR ischemic [all fields] OR ischemic [MESH terms] OR hemorrhagic [all fields]) AND stroke [all fields]) OR ‘brain infarction’[all fields] OR vascular brain accident [all fields] or CVA [all fields])). Similar search terms were used for the EMBASE and Cochrane search. All searches were conducted without restrictions.

Inclusion criteria Only prospective cohort studies reporting on the association between human plasma total adiponectin concentration and CHD or stroke were considered eligible. For CHD and stroke, a full endpoint-criterion description had to be presented, or referred to in previously published articles. As hemorrhagic stroke is pathophysiologically distinct from ischaemic stroke, we only included ischaemic stroke, which is more associated with atherosclerotic disease. Crosssectional studies, literature reviews and studies assessing adiponectin isoforms, such as high molecular weight (HMW) adiponectin, were excluded. To avoid clinical heterogeneity, studies performed in specific cardiovascular high-risk patient populations as patients with type 2 diabetes or patients with pre-existent cardiovascular disease were excluded.

Data extraction Data extraction was conducted by two authors separately. In case of discrepancy, a third author (F.L.J.V.) was consulted. Final data for the meta-analyses were chosen by majority of decision between the three authors. For each included article, information was extracted for authors, year of publication, study design, population characteristics (including sample size, country, predefined cardiovascular risk, gender and age), mean follow-up, type of adiponectin assay, mean and standard deviation (SD) of adiponectin level, length of blood sample freezer stay, endpoint definition, statistical methods and the use of covariates in multivariate models. The authors of the included original articles were not contacted to acquire missing data in order not to introduce potential bias.

138

Adiponectin and risk for future CHD/stroke

Data analyses Only fully adjusted hazard ratios or odds ratios were used for meta-analyses. Although the rare disease assumption might not hold within higher age strata for the outcome CHD (as the prevalence exceeds 10%), extracted odds ratios (ORs) of the nested case-control studies were still considered to be valid estimations of relative risks (RRs) as they were sufficiently close towards 1.0, according to previously published methods (17) or as risk set sampling was used (18–20).

For our main analyses, RRs are presented that are based upon continuous modelling per 1 SD increase in log adiponectin, as the majority of studies assessing this occurrence relation used RRs for reporting the results. We also used the SD log adiponectin models in order to be able to pool the results of studies using various types of bioassayssuch as Multiplex, enzyme-linked immunosorbent assay or radioimmunoassay. RRs reported per increase in log adiponectin were transformed into RRs per increase in SD log adiponectin when sufficient data were provided. For studies that solely reported categorical analyses, we constructed continuous models by using generalized least-squares (GLS) trend estimation. This method estimates dose response association based upon median dose, sample size, RRs and standard errors (21,22). Median values of adiponectin values were chosen as corresponding ‘doses’ in the GLS-models. If median values were not reported, mean values were considered an estimation of the corresponding median. For studies that described open-ended categories of adiponectin concentration, a normal distribution density function was used in order to compute the median of that specific category. Our secondary meta-analysis incorporated continuous models per 10 µg mL-1 adiponectin, to visualize a potential dose-response fashion and since 1 increase in SD log-transformed adiponectin is difficult to translate to clinical practice. Random-effects estimates models were used to incorporate between-study heterogeneity. In addition, the I2 statistic was used to evaluate between-study heterogeneity throughout the random effects meta-analysis. To assess the contribution of the individual studies, a sensitivity analyses was conducted by omitting one study at a time in a stepwise procedure and calculate the pooled RR with 95% confidence intervals (95% CI) for the remainder of the studies. Other sensitivity analyses were metaanalyses using fixed-effects modeling instead of random-effects modelling.

The Begg adjusted rank correlation test, Egger regression test and visual inspection of funnel plots, in which the log relative risks were plotted against their standard errors, was used to assess potential publication bias. The quality of included studies was assessed by the Newcastle-Ottawa Scale (23), a validated technique in evaluating the quality of observational studies in meta-analyses. Statistical analyses were conducted using STATA version 10 (Stata-Corp, College Station, TX), Comprehensive Meta-Analysis version 2 (Biostat, Englewood, NJ) and StaTable (Cytel Inc, Cambridge, MA).

139

7

CHAPTER 7

8783 Potentially relevant articles 5122 EMBASE 3013 MEDLINE 648 COCHRANE

2640 Duplicates removed

6143 Potentially relevant articles evaluated

5987 Excluded 4010 Adiponectin concentration was not the exposure or CHD/Stroke not the outcome 1141 Review, editorial, lecture, protocol or guideline 836 Non-human study (animal study, in vitro study)

156 Articles evaluated in detail

129 CHD articles evaluated in detail*

115 Excluded 48 Cross-sectional study 17 Specific study domain (high risk population) 39 CHD part of a pooled clinical outcome 2 Only HMW adiponectin studied 4 Duplicate publication 3 No abstract or full-text available 2 Insufficient data provided for analyses

CHD: 14 Articles included in meta-analysis

66 Stroke articles evaluated in detail*

64 Excluded 19 Cross-sectional study 2 Specific study domain (high risk population) 39 Stroke part of a pooled clinical outcome 1 Only HMW adiponectin studied 0 Duplicate publication 0 No abstract or full-text available 3 Insufficient data provided for analyses

Stroke: 2 Articles included in meta-analysis

Figure 1. Flow Chart of Identification of Relevant Articles *Numbers do not add up to 156, as 39 studies (part of pooled clinical outcome) overlapped.

140

Adiponectin and risk for future CHD/stroke

RESULTS Selection of studies evaluating the risk of adiponectin on coronary heart disease After selection of 6,143 potentially relevant articles, 129 articles were evaluated in detail.

Twenty articles (19,20,24–41) were considered eligible for meta-analyses as they met the inclusion criteria. Of these articles, four publications (36–39) were additionally excluded since results based upon the same data were published elsewhere (25,28,30).

For duplicate publications, only the first published manuscript was included in analyses.

Two articles (40,41) provided insufficient data, leaving a final number of 14 articles which were used for meta-analyses as is shown in Figure 1. Of these 14 included articles, two articles were based upon data derived from the same cohort (33,35). Both studies were included as they were used for separate analyses. Study characteristics of all 14 included studies are displayed in Table 1. Descriptions of the study outcomes and effect estimates are presented in Tables 2 and 3, respectively.

Included articles were published in the period 2004– 2012 and were either nested casecontrol studies (18,19,24–26,30–33,35) or cohort studies (27–29,34). In total, data of 21,272 patients with valid adiponectin measurements were available in which 5,790 CHD outcome events occurred. The mean age varied from 53 to 75 years and patients originated from various (mostly Western) countries.

Risk of adiponectin on coronary heart disease events No statistically significant associations between adiponectin concentration and incident CHD were found as 1 SD increase in log-transformed adiponectin was accompanied by a RR of 0.97 (95% CI 0.87–1.09) for CHD (Figure 2). The I2 was 64%, indicating moderate to substantial between study heterogeneity. With regard to this heterogeneity, stratified analyses were performed, which did not identify specific effect modifiers (Table 4). All studies were considered of high quality as they scored _6 on the Newcastle- Ottawa Scale. The funnel plot was symmetrically shaped (Figure not shown), and neither the Begg test (P = 0.25) nor the Egger test (P = 0.35) suggested publication bias.

Using GLS, including the data of eight studies (18, 19,25,26,28,30,31,35), which led to 10 data points, comprising 3,184 CHD events among 11,452 patients, 10 µg mL-1 adiponectin increase effectuated a RR of 0.91 (95% CI 0.80–1.03). The corresponding forest plot is displayed in Figure 3. The Begg adjusted rank correlation test (P = 0.93) nor Egger regression test (P = 0.76) suggested publication bias.

Sensitivity analyses for the relation between adiponectin and coronary heart disease risk In sensitivity analyses, none of the individual studies substantially altered the pooled RRs, which varied from 0.94 (95%C I 0.84–1.06) to 1.01 (95% CI 0.91–1.11) per 1 SD increase in log-transformed adiponectin. The sensitivity analyses were done by omitting one study at a time in a stepwise procedure and calculating the pooled RR for the remainder of the studies. In addition, a fixed-effects model instead of a random-effects model resulted comparable pooled RRs for CHD (RR 1.01; 95% CI 0.95–1.06 vs. RR 0.97;

141

7

CHAPTER 7

Study

Year

RR (95% CI)

Lindsay Kanaya1 Kanaya2 3 Laughlin Laughlin4 Frystyk Kizer Hatano Luc Rana3 Rana4 Ai 3 4 Ai Overall

2005 2006 2006 2006 2006 2007 2008 2009 2009 2011 2011 2011 2011

0.82 (0.54 - 1.25) 5.03 (1.03 - 24.45) 0.46 (0.13 - 1.57) 0.97 (0.79 - 1.18) 1.30 (1.03 - 1.65) 0.81 (0.66 - 0.99) 1.19 (1.02 - 1.38) 1.00 (0.55 - 1.80) 0.99 (0.88 - 1.12) 1.02 (0.91 - 1.14) 1.01 (0.86 - 1.18) 0.48 (0.30 - 0.76) 0.58 (0.31 - 1.08) 0.97 (0.86 - 1.08) 0.1

0.2

0.5

1

2

5

10

Figure 2. Relative Risk of coronary heart disease with 1 standard deviation increase in log adiponectin. CI, confidence interval; RR, relative risk; Overall: based upon random-effects model. Stratified analyses: 1estimates given for blacks; 2estimates given for whites; 3estimates given for males; 4estimates given for females.

CHD Study

Year

RR (95% CI)

Pischon 1 Sattar Kanaya 3 Kanaya 4 Koenig Kizer Hatano Côté 1 Côté 2 Pischon 2 Overall

2004 2006 2006 2006 2006 2008 2009 2011 2011 2011

0.85 (0.71 - 1.01) 0.97 (0.79 - 1.18) 1.61 (1.04 - 2.50) 0.95 (0.78 - 1.16) 0.64 (0.42 - 0.98) 0.98 (0.89 - 1.07) 0.80 (0.21 - 3.03) 0.90 (0.09 - 8.62) 1.97 (0.09 - 40.91) 0.69 (0.53 - 0.88) 0.91 (0.80 - 1.03)

Stroke Study

Year

RR (95% CI)

Matsumoto Rajpathak Overall

2008 2011

1.01 (0.95 - 1.07) 1.01 (0.95 - 1.09) 1.01 (0.97 - 1.06)

0.01

0.1

1

10

100

0.01

0.1

1

10

100

Figure 3. Relative risk of coronary heart disease (CHD) and stroke with 10 µg mL-1 increase in adiponectin. CI, confidence interval; RR, relative risk; Overall: based upon random-effects model; Stratified analyses: 1estimates given for males; 2estimates given for females; 3estimates given for blacks; 4estimates given for whites.

142

Adiponectin and risk for future CHD/stroke

95% CI 0.86–1.09) per increase in SD log transformed adiponectin, indicating that heterogeneity is a not to be overlooked issue.

Selection of studies evaluating the risk of adiponectin on stroke Sixty-six articles assessing the relationship between adiponectin and stroke were evaluated in detail. Six studies (20,41–45) were considered eligible to meta-analyse the effect of adiponectin on future stroke. Three studies were additionally excluded (41,43,44), as insufficient data were provided. Eventually, two studies were eligible for final meta-analysis, which included 2,674 patients and 1,080 stroke outcome events. Baseline characteristics of these two studies are displayed in Table 1, outcome event definitions in Table 2 and effect estimates in Table 3.

Table 2. Endpoint definitions of included studies Author CHD

Pischon

Endpoint definition

2004

Fatal CHD, non-fatal myocardial infarction

2005

Fatal or non-fatal CHD, not otherwise specified

Sattar

2006

Fatal CHD (ICD-9 codes 410-414), non-fatal myocardial infarction

Kanaya26

2006

Sudden death due to a CHD history or in the absence of an alternative cause based review of medical records. Fatal myocardial infarction, or any overnight hospitalization for non-fatal myocardial infarction

Laughlin27

2006

Fatal myocardial infarction or non-fatal CHD (ICD-9 codes 410-414)

Koenig28

2006

Fatal or non-fatal acute myocardial infarction

Frystyk29

2007

Fatal CHD or first-time hospitalization due to CHD (ICD-9 codes 410414 and ICD-10 codes I20-25)

Kizer30

2008

Fatal myocardial infarction, sudden cardiac death, procedure related death, non-fatal myocardial infarction, angina pectoris, coronary bypass surgery or percutaneous revascularisation

Hatano31

2009

Fatal myocardial infarction of non-fatal myocardial infarction

Luc32

2009

Fatal or non-fatal myocardial infarction, angina pectoris or unstable angina

Rana33

2011

Fatal or non-fatal CHD (ICD-9 codes 410-414)

34

Ai

2011

Fatal or non fatal myocardial infarction, angina pectoris or coronary insufficiency

Côté35

2011

Fatal or non-fatal CHD (ICD-9 codes 410-414)

Pischon19

2011

Fatal or non-fatal CHD

Matsumoto

2008

The presence of a focal and nonconvulsive neurological deficit, of clear onset, lasting for 24 hours or longer. Stroke subtypes (including ischemic stroke) were confirmed based on CT or MRI images.

Rajpathak20

2011

The rapid onset of a persistent neurological deficit attributed to an obstruction lasting over 24 hours without evidence of other causes unless death supervened or there was a demonstrable lesion compatible with acute stroke on CT or MRI images.

25

Stroke

Year

Lindsay24

18

42

7

143

144

CHD

18

Study Design

Population

2006 Nested Health ABC study case-control Rancho Bernardo Study MONICA/KORA ULSAM Cardiovascular Health Study

2006 Cohort

2006 Cohort

2007 Cohort

2008 Cohort

Kanaya26

Laughlin27

28

Frystyk29

2009 Nested PRIME case-control

2011 Nested EPIC-Norfolk case-control

2011 Nested Framingham Offspring Study case-control

2011 Cohort

2011 Nested Nurses’ Health Study case-control

Luc32

Rana33, §

Ai34

Côté34, §

Pischon19

EPIC-Norfolk

2009 Nested Jichi Medical School Cohort case-control

Hatano31

30

Kizer

Koenig

2006 Nested British Regional Heart Study case-control

Stong Heart Study

2005

Sattar25

2004 Nested Health Professsionals Followcase-control up Study

Year

Lindsay24

Pischon

Author

53 (5)

71 (1)

54 (6)

65 (8)

65 (8) United States 60 (6)

United Kingdom

NS

99 (78)

648 (47)

832 (100)

976 (100)

835 (55)

1106 (45)

8

(median)

8

18

20

6

1820 (100) 16

NS

6

0 (0)

1869 (63)

1332 (45)

1834 (64)

8

7.5

10

10

NS

NS

NS

NS

10

12-15

13

NS

NS

20

6-7

NS

NS

NS

38

604

116

126

252

262

398

66

266

ELISA

ELISA

ELISA

ELISA

1215

89

782

716

811

1100

2211

1012

66

532

No

2760

455

911

1035 1920

177

1002 1859

Multiplex 617

ELISA

ELISA

ELISA

ELISA

RIA

RIA

ELISA

RIA

RIA

Yes

Mean Mean Type Outcome Follow Up freezer bioassay event (years) stay (n) (years)

50-59 (NS) 1832 (100) 9

65 (8)

United States 57(10)

United Kingdom

France, Ireland

Japan

United States 75 (5)

Sweden

Germany

United States 72 (NS)

United States 74 (3)

United Kingdom

United States 61 (8)

798 (100)

Mean Men (%) age, (SD)

United States 65 (8)

Country

Table 1. Characteristics of included studies assessing the relationship between adiponectin and coronary heart disease (chd) or stroke.

CHAPTER 7

Nested Study case-control

Studies by Rana et al. and Côté et al published on same data. NS: not specified; ELISA: enzyme-linked immunosorbent assay; RIA: radioimmunoassay §

NS 15-20

964 Multiplex 964 14 9.7 0 (0) United States 69 (6) 2011 Nested Woman’s Health case-control Rajpathak20

116 ELISA Stroke Matsumoto42 2008 Nested Jichi Medical School Cohort case-control

Table 1 continued

Japan

66 (9)

396 (49)

14

630

Adiponectin and risk for future CHD/stroke

Risk of adiponectin on stroke It was not feasible to meta-analyse the relation between 1 SD increase in logtransformed adiponectin and stroke as only one study (42) provided sufficient data (RR 0.98; 95% CI 0.76–1.27); hence, only metaanalysis per 10 µg mL-1 increase was performed based on the data of two studies revealing no risk of adiponectin on the occurrence of stroke (RR 1.01 95% CI 0.97– 1.06) as also displayed in Figure 3. The I2 was 0%, indicating that between-study heterogeneity was no issue.

Sensitivity analysis for the relation between adiponectin and stroke risk The use of fixed-effects meta-analysis instead of random effects modelling resulted in exactly the same result: RR 1.01 (95% CI 0.97–1.06).

DISCUSSION

7

In this meta-analysis of prospective studies, it is shown that plasma adiponectin levels are not related to the risk of CHD or stroke in patients without clinical manifest cardiovascular disease and across strata of gender, age, race, body mass index (BMI), study design, geographical location of the study, follow-up time, type of adiponectin assay and number of outcome events.

A previous meta-analysis in 2006 concluded that there was no significant relationship between adiponectin and CHD risk (25).

Nevertheless, there were some methodological issues as top vs. bottom tertile adiponectin concentrations were used for pooling the individual effect estimates of the various studies. Moreover, it was insinuated that this meta-analysis was based

145

CHAPTER 7

upon insufficient number of patients and was based on a heterogeneous population (45).

These methodological issues were overcome in the present meta-analysis, in which the majority of included studies were published after 2006. We assume to give now a valid overview of current knowledge on the risk of adiponectin on incident CHD and stroke.

A meta-analysis on the relationship between adiponectin and future stroke has never been published before. Although analyses were based on 1,080 stroke events and heterogeneity in these analyses was low, with only two studies included the added value of a metaanalysis remains questionable.

Obesity is accompanied by a lower concentration of adiponectin and concomitantly results in higher risk for metabolic syndrome, type 2 diabetes mellitus and insulin resistance, which all have been associated with increased cardiovascular risk. The results of our meta-analysis are rather surprising as higher concentrations of adiponectin do not relate to a decreased risk for cardiovascular events. Adiponectin has been thought to have a protective role in the development of atherosclerosis and it is widely hypothesized that adiponectin could therefore reduce the risk for future CHD or stroke. One can only speculate why there is such a difference in the results from on one hand, in vitro studies showing beneficial effects of adiponectin on pathophysiological processes involved in atherogenesis, arterial thrombosis and arterial plaque rupture and on the other hand, clinical studies with cardiovascular endpoints showing no relation between plasma adiponectin and occurrence of CHD or stroke. A reason why this association could not be found might lie in adiponectin isoforms. In the studies presented in this meta-analysis, total adiponectin was measured, while several studies now consider HMW to be the biological active form of adiponectin (46,47) as HMW-adiponectin selectively suppresses endothelial cell apoptosis, whereas the other adiponectin isoforms do not have this effect on the endothelium (48). However, only few cohort studies have prospectively associated HMW adiponectin with CHD (48â&#x20AC;&#x201C;50) or stroke (51). These studies could not identify HMW adiponectin as independent predictor of CHD or stroke. Several cohort studies show that higher total plasma adiponectin is associated with lower rates of incident type 2 diabetes according to a large meta-analysis of prospective studies (52). Type 2 diabetes mellitus is a well-known risk factor for cardiovascular events and it could be speculated that the follow-up time in the cohort studies used in the present meta-analysis, was not long enough to register a reduced cardiovascular risk as a consequence of a lower risk for type 2 diabetes mellitus.

Our study has several strengths. First, the fair number of both patients included (23,919) and outcome events (6,870), makes conclusions based upon this meta-analysis well-funded.

Second, the design of the included studies is prospective and performed in non-high-risk patients, making reversed causality (altered circulating adiponectin concentration as a representation of a recent â&#x20AC;&#x201C; e.g. cardiovascular â&#x20AC;&#x201C; event) and even selection bias less likely.

Third, the pooled effect estimate with GLS-analyses was essentially similar compared to the effect with the main analysis, indicating the low risk of bias due to missing data as GLS allowed us to use all data.

Several study limitations need to be considered. First, the assumption was made that ORs and hazard ratios are both valid representatives of a RR, while ORs can overestimate the

146

Adiponectin and risk for future CHD/stroke

RR if the rare disease assumption is not met (17). The latter is most likely the case with increasing age, especially in male subpopulations where the incidence of CHD and CHD mortality is high (2). However, extracted ORs were sufficiently low to be valid estimators of the RR (17). In addition, stratified analyses between nested case-control studies and cohort studies yielded no different results. Secondly, the wide duration of freezer stay range could be considered as a limitation. It is shown that length of frozen state of biomaterial may influence the levels of various metabolites, also originating from adipose tissue (53). However, adiponectin is considered to be a stable adipokine with stable concentrations even after long frozen storage (54). Hence, we do not consider the differences in freezer stay to interfere with study outcomes. Thirdly, a potential limitation of our study could be residual confounding. As shown in Table 3, many of the included studies adjusted for a wide variety of potential confounders. It is nearly impossible to acquire and meta-analyse effect estimates extracted from homogeneous models. By choosing the most appropriate adjusted models available and by additional checking of missing essential confounders, we believe that the effects of residual confounding on the individual effect estimates are negligible. Fourthly, clinical heterogeneity could be a limitation. By restricting the present meta-analysis only to patients without clinically manifest cardiovascular disease, we believe that clinical heterogeneity is unlikely. However, we do acknowledge that slightly different endpoint definitions of CHD (Table 1) may induce some inaccuracy. In addition, the relationship between adiponectin and CHD may be different for men and women. As the majority of the studies did not report gender differences, we chose not to distinguish between males or females in our main analysis.

Comparable results for males and females (Table 4) supported this decision. Finally, publication bias can affect the result of every meta-analysis. Nevertheless, tests for publication bias did not indicate this bias has affected our results substantially.

In conclusion, plasma adiponectin concentrations are not associated with risk for future CHD or stroke events in subjects without clinical manifest cardiovascular disease. There was no heterogeneity across strata of gender, age, race, BMI, study design, geographical location of the study, follow-up time, type of adiponectin assay and number of outcome events. Although plasma levels of adiponectin are strongly related to obesity and insulin resistance, there is no relation between adiponectin and the risk for cardiovascular events, limiting the potential for adiponectin in intervention studies in order to reduce the risk for cardiovascular events.

147

7

148 RR (95% CI)

Highest quartile (≥12 μg/ml; median ≈ 14.68) vs. lowest quartile (<5 μg/ml; median ≈ 2.99)

Recalculated Per 1 SD increase in log μg/ml; SD ≈ 4.80

Model 5* 2.13 (0.91 – 4.95)

Recalculated Per 1 SD increase in log μg/ml; SD ≈ 5.13

Kanaya26

Whites

Model 1 Model 2 Model 3 Model 4

Highest tertile (>13.33 μg/ml; median ≈ 5.22) vs. lowest tertile (<8.32 μg/ml; median ≈ 16.42)

Sattar25

Model 1 Model 2 Model 3 Model 4 Model 5*

Model 1 Model 2 Model 3 Model 4*

0.28 (N.S.) 0.30 (N.S.) 0.36 (N.S.) 0.41 (N.S.) 0.46 (0.13 – 1.57)

1.16 (N.S.) 1.48 (N.S.) 1.87 (N.S.) 3.01 (N.S.) 5.03 (1.03 – 24.45)

0.79 (0.58 – 1.06) 0.81 (0.60 – 1.10) 0.90 (0.65 – 1.25) 0.88 (0.63 – 1.24)

Model 1* 0.82 (0.54 – 1.25)

Per 1 SD increase in log μg/ml

Lindsay24

Model 1 0.39 (0.23 – 0.64) Model 2 0.41 (0.24 – 0.70) Model 3* 0.56 (0.32 – 0.99)

Model

Highest quintile (24.9-56.1 μg/ml; median 29.2) vs. lowest quintile (2.4 – 10.5 μg/ml; median 7.9)

Blacks

Subgroup Comparison

Pischon18 (2004)

CHD

Author

Age, sex and study site Model 1 covariates smoking, hypertension, LDL, BMI, abdominal visceral fat and medication use Model 2 covariates + glucose subgroup Model 2 covariates + HDL Model 2 covariates + CRP and fasting insulin

Age, sex, study site, current smoking, hypertension, LDL, , BMI, abdominal visceral fat, medication use, glucose subgroup, HDL, CRP and fasting insulin

Age, sex and study site Model 1 covariates for smoking, hypertension, LDL, BMI, abdominal visceral fat and medication use Model 2 covariates + glucose subgroup Model 2 covariates + HDL Model 2 covariates + CRP and fasting insulin

Age, town Additional adjustment for BMI Model 2 covariates + total cholesterol, HDL and log triglycerides Model 3 covariates + smoking, alcohol, physical activity, social class and systolic blood pressure

Age, sex, for waist circumference, percentage fat, systolic blood pressure, smoking and albumin/creatinin ratio

Age, smoking status and month of blood draw BMI, family history of myocardial infarction before age 60 years, history of diabetes, history op hypertension, alcohol intake and physical activity Model 2 covariates + HDL and LDL

Covariates

Table 3. Relative risk of coronary heart disease (CHD) or stroke according to serum total adiponectin concentration

CHAPTER 7

Age and sex Model 1 covariates + HDL, triglycerides, BMI, diabetes, smoking, systolic blood pressure and CRP Age and sex Model 1 covariates + HDL, triglycerides, BMI, diabetes, smoking, systolic blood pressure and CRP

Model 1 1.03 (0.68 – 1.59) Model 2* 1.00 (0.55 – 1.80) Model 1 1.12 (0.37 – 3.40) Model 2* 0.97 (0.20 – 4.69)

Per 1 SD increase in log μg/ml

Lowest quartile (< 5.30 μg/ml; median ≈ 3.24) vs. highest quartile (<11.60 μg/ ml; median ≈ 14.45)

Hatano31

Age, sex, race, subclinical disease and clinic Model 1 covariates + WHR Model 2 covariates + systolic blood pressure, LDL, smoking, alcohol, serum creatinine and Leptin Model 3 covariates + heath status and measured weight loss or gain >10lb in past 3 years

N.S. N.S. 1.19 (1.02 - 1.38) N.S.

Model 1 Model 2 Model 3* Model 4

Per 1 SD increase in log μg/ml

Kizer30

Crude Age, systolic blood pressure, total cholesterol, HDL, smoking and BMI

Model 1 0.77 (0.64 - 0.92) Model 2* 0.81 (0.66 - 0.99)

Per 1 SD increase in log μg/ml

0.56 (0.36 – 0.87) 0.62 (0.39 – 0.98) 0.62 (0.39 – 0.98) 0.71 (0.44 – 1.15)

Age Model 1 covariate+ smoking, alcohol, BMI, physical activity, actual hypertension and history of diabetes Model 2 covariates + total cholesterol Model 2 covariates + HDL

Age and sex Model 1 covariates + waist girth Model 2 covariates + HDL, triglycerides and fasting plasma glucose

Model 1 Model 2 Model 3

0.99 (0.86 - 1.21) 1.09 (0.88 - 1.33) 1.30 (1.03 - 1.65)

Age and sex Model 1 covariates + waist girth Model 2 covariates+ HDL, triglycerides and fasting plasma glucose

Model 1 0.83 (0.69 - 0.98) Model 2 0.87 (0.71 - 1.05) Model 3* 0.97 (0.79 - 1.18)

Frystyk29

Per 1 SD increase in log μg/ml

Women

Age, sex, study site, current smoking, hypertension, LDL, , BMI, abdominal visceral fat, medication use, glucose subgroup, HDL, CRP and fasting insulin

Model 5* 0.90 (0.54 – 1.58)

Highest tertile (IQR 8.94 – 13.00 μg/ Model 1 ml; median 10.4) vs. lowest tertile (IQR Model 2 3.06 – 4.44 μg/ml; median 3.8) Model 3* Model 4

Per 1 SD increase in log μg/ml

Men

Highest quartile (≥16 μg/ml; median ≈ 18.68) vs. lowest quartile( <5 μg/ml; median ≈ 5.29)

Koenig28

Laughlin27

Table 3 continued

Adiponectin and risk for future CHD/stroke

7

149

150

Per 1 SD increase in log μg/ml

Per 1 SD increase in log μg/ml

Women

Per 1 SD increase in log μg/ml

Women

Men

Per 1 SD increase in log μg/ml

Men

Rana33,§

Ai34

Per 1 SD increase in log μg/ml

Subgroup Comparison

Luc32

CHD

Author

Table 3 continued RR (95% CI)

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6* Model 7

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6* Model 7

0.40 0.54 0.55 0.69 0.60 0.58 (0.31 – 1.07) 0.64

0.41 0.47 0.46 0.49 0.47 0.49 (0.31 – 0.77) 0.51

Model 1* 1.01 (0.86 – 1.18)

Model 1* 1.02 (0.91 – 1.14)

Model 1* 0.99 (0.88 - 1.12)

Model

Age and smoking status Model 1+ systolic blood pressure, hypertensive treatment and diabetes Model 2 + cholesterol lowering medication Model 3 + total cholesterol and HDL Model 4 + BMI Model 5 + CRP Age, smoking status, systolic blood pressure, hypertensive treatment, cholesterol lowering medication, total cholesterol, HDL, CRP, HOMA-IR and waist circumference

Age and smoking status Model 1+ systolic blood pressure, hypertensive treatment and diabetes Model 2 + cholesterol lowering medication Model 3 + total cholesterol and HDL Model 4 + BMI Model 5 + CRP Age, smoking status, systolic blood pressure, hypertensive treatment, cholesterol lowering medication, total cholesterol, HDL, CRP, HOMA-IR and waist circumference

Waist circumference, physical activity, smoking, diabetes, systolic blood pressure, LDL, HDL and hormone replacement therapy

Waist circumference, physical activity, smoking, diabetes, systolic blood pressure, LDL and HDL

Total cholesterol, HDL, triglycerides, type 2 diabetes, hypertension and smoking

Covariates

CHAPTER 7

Per 1 µg/ml increase

Women

Crude BMI, waist circumference, age, gender, blood pressure, HDL, LDL, total cholesterol, smoking and type 2 diabetes Matching factors: Age, smoking status, date of blood draw, fasting status and reported problems with blood draw Model 1 covariates + parental history of myocardial infarction, hormone replacement therapy, alcohol consumption, physical activity, BMI, history of hypertension and LDL

Model 1 0.73 (0.55 – 0.96) Model 2* 1.07 (0.79 – 1.45) Model 1

Age and sex Model 1 covariates + HDL, triglycerides and BMI Model 2 covariates + current smoking, systolic blood pressure and CRP Age and ethnicity Model 1 covariates + BMI Model 2 covariates + current smoking, physical activity, nonsteroidal anti-inflammatory drug use, hypertension medication use, systolic blood pressure, history of coronary and artery diseases, HDL, triglycerides, diabetes and waist circumference

Model 1 2.04 (1.09 – 3.80) Model 2 1.94 (0.96 – 3.91) Model 3* 1.60 (0.78 – 3.31) Model 1 0.77 (0.59 – 1.01) Model 2 0.81 (0.61 – 1.08) Model 3* 1.16 (0.82 – 1.63)

Lowest quartile (<5.6 μg/ml; median ≈ 1.5) vs. Highest quartile (>12.4 μg/ml; median ≈ 15.1)

Highest quartile (median = 46.0) vs.

Lowest quartile (median =14.8)

Age and sex Model 1 covariates + HDL, triglycerides and BMI Model 2 covariates + current smoking, systolic blood pressure and CRP

0.88 (0.71 – 1.09) 0.92 (0.72 – 1.18) 0.98 (0.76 – 1.27)

Model 1 Model 2 Model 3

Model 2* 0.50 (0.33 – 0.75)

0.35 (0.24 – 0.51)

Crude BMI, waist circumference, age, gender, blood pressure, HDL, LDL, total cholesterol, smoking and type 2 diabetes

Model 1 0.78 (0.63 – 0.96) Model 2* 0.99 (0.79 – 1.24)

Per 1 SD increase in log μg/ml

Highest quintile (range 12.39-28.45 μg/ ml; median 14.69) vs. Lowest quintile (range 1.57 – 5.97 μg/ml; median 4.83)

Per 1 µg/ml increase

Men

* Models that were chosen for meta-analysis. § Publications by Rana et al. and Côté et al. are based upon data derived from the same cohort. N.S: Not Specified; BMI: body mass index; LDL: low density lipoprotein; HDL: high density lipoprotein; CRP: C-reactive protein; IQR: Interquartile range

Rajpathak20 (2011)

Matsumoto42 (2008)

Stroke

Pischon19 (2011)

Côté35,§

Table 3 continued

Adiponectin and risk for future CHD/stroke

7

151

CHAPTER 7

Table 4. Stratified meta-analyses per 1 SD increase in log adiponectin and risk of CHD Characteristic

Data points

CHD

RR (95%CI)

P-value I² (%) heterogeneity

Yes

No

All studies

13

3134

10798

0.97 (0.86 – 1.09)

0.001

64

Male

3

Female

3

1341

2538

0.86 (0.65 – 1.14)

0.007

80

533

2738

1.02 (0.76 – 1.37)

0.032

71

Blacks Whites

1

96

948

5.03 (1.03 – 24.45)

1

166

1263

0.46 (0.13 – 1.57)

American Indians

1

66

66

0.82 (0.54 – 1.25)

North-America

8

1361

6919

0.92 (0.71 – 1.18)

Europe

4

1735

4691

0.98 (0.90 – 1.06)

0.000

75

Asia

1

38

89

1.00 (0.55 – 1.80)

0.257

26

Nested case-control

6

2327

4011

1.03 (0.97 – 1.10)

0.398

2

Cohort

7

807

6787

0.85 (0.63 – 1.14)

0.000

77

50 ≤ 60

3

794

3975

0.68 (0.40 – 1.16)

0.003

82

61 ≤ 70

4

1106

2014

1.01 (0.92 – 1.10)

0.808

0

71 ≤ 80

6

1184

4809

1.05 (0.85 – 1.31)

0.003

72

<100

4

260

2648

0.91 (0.56 – 1.49)

0.084

55

100-500

6

1007

4965

0.90 (0.73 – 1.11)

0.001

75

>500

3

1867

3185

1.05 (0.95 – 1.16)

0.150

47

ELISA

7

1937

6206

0.92 (0.78 – 1.08)

0.001

73

RIA

4

580

3377

1.05 (0.77 – 1.41)

0.032

62

Multiplex

1

617

1215

0.99 (0.88 – 1.12)

0 ≤ 10

5

1020

3798

1.01 (0.72 – 1.41)

0.006

72

>10 – ≤20

5

1796

5834

0.92 (0.78 – 1.07)

0.012

69

>20

2

252

1100

1.12 (0.84 – 1.49)

0.064

71

20.0 ≤ 25.0

1

38

89

1.00 (0.55 – 1.81)

25.1 ≤ 30.0

11

3030

10643

0.97 (0.86 – 1.10)

0.000

69

3134

10798

0.97 (0.86 – 1.09)

0.001

64

Gender

Race

Study continent

Study Design

Mean Age, years

Outcome events, n

Type of assay

Follow-up, years

BMI

Quality score <6

0

≥6

13

BMI, body mass index; CHD, coronary heart disease; CI, confidence interval; ELISA, enzyme-linked immunosorbent assay; RIA, radioimmunoassay; RR, relative risk.

152

Adiponectin and risk for future CHD/stroke

REFERENCES 1. Lopez AD, Murray CC. The global burden of disease, 1990– 2020. Nat Med 1998; 4: 1241–1243.

2. World Health Organization. [WWW document]. URL http:// www.who.int/mediacentre/ factsheets/fs317/en/index.html (accessed 29 January 2013).

3. Hajer GR, van Haeften TW, Visseren FL. Adipose tissue dysfunction in obesity, diabetes, and vascular diseases. Eur Heart J 2008; 29: 2959–2971.

4. Ouchi N, Kihara S, Arita Y et al. Adiponectin, an adipocytederived plasma protein, inhibits endothelial NF-kappaB signaling through a cAMP-dependent pathway. Circulation 2000; 102: 1296–1301.

5. Ouchi N, Kihara S, Arita Y et al. Adipocyte-derived plasma protein, adiponectin, suppresses lipid accumulation and class A scavenger receptor expression in human monocyte-derived macrophages. Circulation 2001; 103: 1057–1063.

6. Shibata R, Ouchi N, Murohara T. Adiponectin and cardiovascular disease. Circ J 2009; 73: 608– 614.

7. Kumada M, Kihara S, Ouchi N et al. Adiponectin specifically increased tissue inhibitor of metalloproteinase-1 through interleukin-10 expression in human macrophages. Circulation 2004; 109: 2046–2049.

8. Berg AH, Combs TP, Du X, Brownlee M, Scherer PE. The adipocyte-secreted protein Acrp30 enhances hepatic insulin action. Nat Med 2001; 7: 947–953.

9. Okamoto Y, Kihara S, Ouchi N et al. Adiponectin reduces atherosclerosis in apolipoprotein E-deficient mice. Circulation 2002; 106: 2767–2770.

10. Kubota N, Terauchi Y, Yamauchi T et al. Disruption of adiponectin causes insulin resistance and neointimal formation. J Biol Chem 2002; 277: 25863–25866.

11. Costacou T, Zgibor JC, Evans RW et al. The prospective association between adiponectin and coronary artery disease among individuals with type 1 diabetes. The Pittsburgh Epidemiology of Diabetes Complications Study. Diabetologia 2005; 48: 41–48.

12. Ho DY, Cook NR, Britton KA et al. High-molecular-weight and total adiponectin levels and incident symptomatic peripheral artery disease in women: a prospective investigation. Circulation 2011; 124: 2303–2311.

13. Hajer GR, van der Graaf Y, Olijhoek JK, Edlinger M, Visseren FL. Low plasma levels of adiponectin are associated with low risk for future cardiovascular events in patients with clinical evident vascular disease. Am Heart J 2007; 154: 750.e1–7.

14. Emerging Risk Factors Collaboration, Wormser D, Kaptoge S et al. Separate and combined associations of body-mass index and abdominal adiposity with cardiovascular disease: collaborative analysis of 58 prospective studies. Lancet 2011; 377: 1085–1095.

15. Dekker JM, Funahashi T, Nijpels G et al. Prognostic value of adiponectin for cardiovascular disease and mortality. J Clin Endocrinol Metab 2008; 93: 1489–1496.

16. Stroup DF, Berlin JA, Morton SC et al. Meta-analysis of observational studies in epidemiology: a proposal for reporting. Meta-analysis Of Observational Studies in Epidemiology (MOOSE) group. JAMA 2000; 283: 2008–2012.

17. Zhang J, Yu KF. What’s the relative risk? A method of correcting the odds ratio in cohort studies of common outcomes. JAMA 1998; 280: 1690–1691.

18. Pischon T, Girman CJ, Hotamisligil GS, Rifai N, Hu FB, Rimm EB. Plasma adiponectin levels and risk of myocardial infarction in men. JAMA 2004; 291: 1730–1737.

19. Pischon T, Hu FB, Girman CJ et al. Plasma total and high molecular weight adiponectin levels and risk of coronary heart disease in women. Atherosclerosis 2011; 219: 322–329.

20. Rajpathak SN, Kaplan RC, Wassertheil-Smoller S et al. Resistin, but not adiponectin and leptin, is associated with the risk of ischemic stroke among postmenopausal women: results from the Women’s Health Initiative. Stroke 2011; 42: 1813–1820.

21. Orsini N, Bellocco R, Greenland S. Generalized least squares for trend estimation of summarized doseresponse data. Stata J 2006; 6: 40–57.

22. Greenland S, Longnecker MP. Methods for trend estimation from summarized dose-response data, with applications to metaanalysis. Am J Epidemiol 1992; 135: 1301–1309.

153

7

CHAPTER 7

23. Wells GA, Shea B, O’Connell D et al. The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses. [WWW document]. URL http://www. ohri.ca/ programs/clinical_epidemiology/oxford.htm (accessed 29 January 2013). 24. Lindsay RS, Resnick HE, Zhu J et al. Adiponectin and coronary heart disease: the Strong Heart Study.

Arterioscler Thromb Vasc Biol 2005; 25: e15–e16.

25. Sattar N, Wannamethee G, Sarwar N et al. Adiponectin and coronary heart disease: a prospective study and meta-analysis. Circulation 2006; 114: 623–629.

26. Kanaya AM, Wassel Fyr C, Vittinghoff E et al. Serum adiponectin and coronary heart disease risk in older Black and White Americans. J Clin Endocrinol Metab 2006; 91: 5044–5050.

27. Laughlin GA, Barrett-Connor E, May S, Langenberg C. Association of adiponectin with coronary heart disease and mortality: the Rancho Bernardo study. Am J Epidemiol 2007; 165: 164–174.

28. Koenig W, Khuseyinova N, Baumert J, Meisinger C, Löwel H. Serum concentrations of adiponectin and risk of type 2 diabetes mellitus and coronary heart disease in apparently healthy middle-aged men: results from the 18-year follow-up of a large cohort from southern Germany.

J Am Coll Cardiol 2006; 48: 1369–1377.

29. Frystyk J, Berne C, Berglund L, Jensevik K, Flyvbjerg A, Zethelius B. Serum adiponectin is a predictor of coronary heart disease: a population-based 10-year follow-up study in elderly men.

J Clin Endocrinol Metab 2007; 92: 571–576.

30. Kizer JR, Barzilay JI, Kuller LH, Gottdiener JS. Adiponectin and risk of coronary heart disease in older men and women. J Clin Endocrinol Metab 2008; 93: 3357–3364.

31. Hatano Y, Matsumoto M, Ishikawa S, Kajii E. Plasma adiponectin level and myocardial infarction: the JMS Cohort Study. J Epidemiol 2009; 19: 49–55.

32. Luc G, Empana JP, Morange P et al. Adipocytokines and the risk of coronary heart disease in healthy middle aged men: the PRIME Study. Int J Obes (Lond) 2010; 34: 118–126. 33. Rana JS, Arsenault BJ, Després JP et al. Inflammatory biomarkers, physical activity, waist circumference, and risk of future coronary heart disease in healthy men and women. Eur Heart J 2011; 32: 336–344.

34. Ai M, Otokozawa S, Asztalos BF et al. Adiponectin: an independent risk factor for coronary heart disease in men in the Framingham offspring Study. Atherosclerosis 2011; 217: 543– 548.

35. Côté M, Cartier A, Reuwer AQ et al. Adiponectin and risk of coronary heart disease in apparently healthy men and women (from the EPIC-Norfolk Prospective Population Study). Am J Cardiol 2011; 108: 367–373.

36. Karakas M, Zierer A, Herder C et al. Leptin, adiponectin, their ratio and risk of Coronary Heart Disease: results from the MONICA/KORA Augsburg Study 1984–2002. Atherosclerosis 2010; 209: 220–225.

37. Wannamethee SG, Welsh P, Whincup PH et al. High adiponectin and increased risk of cardiovascular disease and mortality in asymptomatic older men: does NT-proBNP help to explain this association? Eur J Cardiovasc Prev Rehabil 2011; 18: 65–71.

38. Herder C, Baumert J, Zierer A et al. Immunological and cardiometabolic risk factors in the prediction of type 2 diabetes and coronary events: MONICA/KORA Augsburg case-cohort study. PLoS ONE 2011; 6: e19852.

39. Kizer JR, Benkeser D, Arnold AM et al. Total and highmolecular- weight adiponectin and risk of coronary heart disease and ischemic stroke in older adults. J Clin Endocrinol Metab 2013; 98: 255–263.

40. Lawlor DA, Davey Smith G, Ebrahim S, Thompson C, Sattar N. Plasma adiponectin levels are associated with insulin resistance, but do not predict future risk of coronary heart disease in women. J Clin Endocrinol Metab 2005; 90: 5677–5683.

41. Gardener H, Goldberg R, Mendez AJ et al. Adiponectin and risk of vascular events in the Northern Manhattan Study. Atherosclerosis 2013; 226: 483–489.

42. Matsumoto M, Ishikawa S, Kajii E. Association of adiponectin with cerebrovascular disease: a nested case-control study. Stroke 2008; 39: 323–328.

43. Khalili P, Flyvbjerg A, Frystyk J et al. Total adiponectin does not predict cardiovascular events in middle-aged men in a prospective, long-term follow-up study. Diabetes Metab 2010; 36: 137– 143.

154

Adiponectin and risk for future CHD/stroke

44. Prugger C, Luc G, Haas B et al. Adipocytokines and the risk of ischemic stroke: the PRIME Study. Ann Neurol 2012; 71: 478– 486.

45. Pischon T, Schulze MB, Rimm EB. Letter by Pischon et al. regarding article, ‘Adiponectin and coronary heart disease: a prospective study and meta-analysis. Circulation 2007; 115: e322; author reply e323.

46. Waki H, Yamauchi T, Kamon J et al. Impaired multimerization of human adiponectin mutants associated with diabetes. J Biol Chem 2003; 278: 40352–40363.

47. Goldstein BJ, Scalia RG, Ma XL. Protective vascular and myocardial effects of adiponectin. Nat Clin Pract Cardiovasc Med 2009; 6: 27–35.

48. Kobayashi H, Ouchi N, Kihara S et al. Selective suppression of endothelial cell apoptosis by the high molecular weight form of adiponectin. Circ Res 2004; 94: e27–e31.

49. Sattar N, Watt P, Cherry L, Ebrahim S, Davey Smith G, Lawlor DA. High molecular weight adiponectin is not associated with incident coronary heart disease in older women: a nested prospective case-control study. J Clin Endocrinol Metab 2008; 93: 1846–1849.

50. Beassler A, Schlossbauer S, Stark K et al. Adiponectin multimeric forms but not total adiponectin levels are associated with myocardial infarction in non-diabetic men. J Atheroscler Thromb 2011; 18: 616–627.

51. Ogorodnikova AD, Wassertheil-Smoller S, Mancuso P et al. High-molecular-weight adiponectin and incident ischemic stroke in postmenopausal women: a Women’s Health Initiative Study.

Stroke 2010; 41: 1376–1381.

52. Li S, Shin HJ, Ding EL, van Dam RM. Adiponectin levels and risk of type 2 diabetes: a systematic review and meta-analysis. JAMA 2009; 302: 179–188.

53. de Jager W, Bourcier K, Rijkers GT, Prakken BJ, Seyfert-Margolis V. Prerequisites for cytokine measurements in clinical trials with multiplex immunoassays. BMC Immunol 2009; 10: 52.

54. Shand B, Elder P, Scott R, Frampton C, Willis J. Biovariability of plasma adiponectin. Clin Chem Lab Med 2006; 44: 1264– 1268.

7

155

CHAPTER 8 General discussion

CHAPTER 8

Obesity is the result of the expansion of adipose tissue as a consequence of chronic energy excess. The relation between abundance of (visceral) fat and death has been established for millennia, as well as the observation that obesity is the result of excessive food intake or lack of exercise (1-3).

Patient: The problem is that obesity runs in our family.

Doctor: No, the problem is that no one runs in your family.

In the 21th century we are well aware that obesity is a complex condition leading to multiple comorbidities including type 2 diabetes, the metabolic syndrome and coronary heart disease (4,5). Despite intensive weight loss programs and obesity prevention programs the obesity epidemic is not controlled (6,7). Therefore, there is a pressing need for reconsideration of current anti-obesity strategies to prevent obesity-induced comorbidities and death. In order to develop more powerful treatment or diagnostic tools for the obese patient at risk, further insights in the underlying molecular mechanisms are required. In this thesis, we aimed to explore current and novel pathways in adipose tissue dysfunction, as a result of obesity, and investigated how they might contribute to metabolic and cardiovascular disease.

PART ONE Differential contribution of distinct visceral fat depots to metabolic disease: does reciprocal signaling with surrounding organs play a role?

Adipose tissue is a very interesting but complex tissue, which is heavily challenged in the state of obesity. The role of adipose tissue in obesity-induced comorbidities has been the subject of extensive study in the last decades. However, due to difficulties of studying human adipose tissue in vivo or ex vivo, many scientists focus on imaging fat tissue with MRI or CT, enabling quantification of fat or even qualitative assessment of metabolic activity by enhanced imaging methods such as spectroscopy (8-10). Although this yields some valuable information, it only provides information on the amount or location of fat and FFA, but not about the subsequent effects of FFA or any other molecules, or how they are mechanistically involved in obesity-induced metabolic disease. Furthermore, with current imaging techniques it is very difficult to distinguish distinct visceral fat depots, which is why most scientists focus on subcutaneous and visceral fat only. The study of ex vivo human biopsies reveals more insights about morphology of adipose tissue, the expression and secretion of molecules by distinct adipose tissue depots, and their subsequent effect

158

General discussion

on target cells in vitro or relation with clinical metabolic characteristics (11-15). However, due to difficulty of obtaining and handling multiple visceral fat depots, and in particular mesenteric fat, most scientists again only focus on the difference between subcutaneous fat and one other visceral depot, of which omental fat is most often studied. These approaches therefore really undervalue the role of distinct visceral fat depots, and most likely will result in incomplete or puzzling results as findings from omental as well as mesenteric fat are used interchangeably under the term visceral fat.

Visceral fat can be subdivided into multiple different depots as is described in chapter 2.

The most important depots are mesenteric, omental and perivascular fat, which differ in metabolic functions such as lipolysis and adipokine secretion (12,16-19). In chapter 2 we have reviewed current literature about intrinsic differences between distinct abdominal fat depots, and hinted at some mechanistic differences by which distinct depots might contribute to metabolic dysfunction via interaction with different target organs. In chapter 3 we have studied ex vivo human adipose tissue biopsies from four different regions in the abdomen: subcutaneous adipose tissue and three visceral fat depots: mesentery, omentum and periaortic adipose tissue. In this clinical study, fat biopsies were characterized ex vivo, based on morphology and secretion of various adipokines. The most striking finding was the differential contribution of distinct visceral fat depots with metabolic dysfunction.

Perivascular fat was highly active as noted by abundant presence of crown like structures (activated macrophages surrounding a death adipocyte) and adipokine secretion, although the morphological characteristics were not related to systemic metabolic dysfunction.

These data suggest that, by equal comparison with other visceral fat depots, perivascular fat might act in a paracrine manner by interaction with the vascular wall (20,21), but its effects on systemic derangements seem to be limited. In contrast, adipocyte size and crown like structures of both omental and mesenteric fat were related to metabolic derangements. Interestingly, mesenteric fat was mainly related to insulin resistance, whereas omental fat was more strongly related to higher triglyceride and lower HDLcholesterol levels. To our knowledge, this differential contribution of distinct AT depots to either glucose or lipid derangements has not been frequently reported previously. While the relation between omental fat and dyslipidemia has been regularly reported to be associated with the secretion of PAI-1, chemerin and TIMP-1, adipokines associated with liver steatosis (22,23), the relation between mesenteric adipose tissue and insulin resistance is less well known.

Mesenteric fat and insulin resistance A specific relation has been observed between mesenteric fat and insulin resistance in both animal and human studies (24-26) and chapter 3 of this thesis. However, the mechanisms by which mesenteric fat might influence insulin resistance remain elusive.

Because of the close proximity of mesenteric fat with the intestine it is tempting to speculate about reciprocal signaling between mesenteric fat and the intestinal endocrine system. Some indications underscore this hypothesis, such as specific inflammation of mesenteric fat -but not other AT depots- in mice and humans with colitis (27), and influence

159

8

CHAPTER 8

of change in diet or microbiota on mesenteric fat (26,28). However, it is uncertain whether fat-derived molecules might directly affect the enteric endocrine system, for example by interaction of adipokines with the insulin sensitizing enteric hormones GIP (glucosedependent insulinotropic polypeptide) and GLP-1 (glucagon-like peptide-1). The adipokine leptin is a potential mediator of such effects, which is a protein hormone released by adipocytes and is a key regulator of appetite, metabolism and immunology (29). We have observed that mesenteric fat of obese men secreted the highest levels of the adipokine leptin compared to three other fat depots studied (chapter 3), which has been shown on mRNA level before as well (30). Furthermore, increased circulating leptin levels are consistently related to systemic insulin resistance (31,32). The leptin receptor is expressed in human intestinal enteroendocrine L cells, and binding of leptin to this receptor has been shown to stimulate GLP-1 secretion, whereas leptin resistance was associated with lower GLP-1 levels and a diminished GLP-1 response to glucose in mice (33). Moreover, following a very low caloric diet, circulating leptin levels decreased while soluble leptin receptor levels increased, together with an increase in insulin sensitivity in obese females (34). These findings suggest an indirect but important role for leptin in insulin signaling via the enteric endocrine system. Whether these observations can explain the main mechanisms by which mesenteric adipose tissue contributes to insulin resistance remains to be determined.

Furthermore, it would be interesting to study whether mesenteric adipose tissue indeed has a more important role in glucose metabolism than omental adipose tissue. In favor of this hypothesis is the observation that surgical removal of omental fat does not improve insulin sensitivity in obese adults, while lipid levels did improve after 1 month (35-37).

Nevertheless, reports on omentectomy are inconclusive, and do not provide information about the role of mesenteric fat in obesity-induced metabolic dysfunction (35,36).

PART TWO Underlying mechanisms of obesity-induced metabolic disease. What signaling pathways are we missing?

In adipose tissue, numerous immune cells and molecules are involved to regulate metabolic homeostasis via different pathways (chapter 2). In obesity, these pathways are disturbed in favor of pro-inflammatory immune cells and cytokines, which induce low grade systemic inflammation and insulin resistance in peripheral tissues. However, besides soluble molecules and cells, adipocytes secrete functional extracellular vesicles as well (38,39).

In the past decades it became clear that extracellular vesicles (EVs), once thought to be just a disposal route for obsolete proteins (40), are actually active immunomodulatory signaling vesicles (41,42). EVs include microvesicles, microparticles and exosomes which are nanometer-sized membrane vesicles secreted by all eukaryotic cells (43). These vesicles reflect the (patho)physiological state of the donor tissue they are derived from, and serve as messenger vehicles containing cell-specific cytosolic and membrane-bound proteins, mRNA and miRNA (42). The components expressed in EVs influence cellular signaling and

160

General discussion

transport processes as well as regulate epigenetic modulations, metabolic memory and immunometabolism in cells and organs (41-43). Compared to single molecules, EVs are more stable in plasma and transfer of cargo to target cells is conducted via multiple mechanisms, enabling transport of proteins that might otherwise be difficult to achieve.

Furthermore, the distinct set of proteins, lipids and RNAs packaged together in one vesicle might have a cumulative effect when hitting the target cell (44). As such, EVs can modify and activate target cells in a paracrine or endocrine fashion. The mechanisms by which EVs can transfer specific cargo into target cells include 1) interaction of EV-membrane proteins with receptors on a target cell and activate intracellular signaling, 2) fusion of the EVs with the plasma membrane of the target cell, and subsequent release of their contents into the recipient cell whereupon proteins can be incorporated in the plasma membrane and RNA transcripts could be translated into protein and 3) cleavage of membrane receptor bound proteins by proteases in the extracellular space, enabling cleaved fragments to act as a soluble ligand and bind to target cell surface receptor (44-46).

The immunomodulatory properties of EVs may lead to active participation in adipose tissue inflammation, a key process underlying obesity-induced metabolic disease. However, so far the role of adipocyte EVs in adipose tissue inflammation or obesity-induced insulin resistance was only studied in mice (38,47-49). As this ‘novel’ pathway may also explain some missing links in human adipose tissue dysfunction and metabolic disease, we have studied the role of EVs secreted by adipocytes and human adipose tissue in chapter 4 and 5.

In chapter 4 we have shown that EVs released by human adipocytes or adipose tissue explants play a role in the paracrine interaction between adipocytes and macrophages, a key mechanism in adipose tissue inflammation, leading to metabolic complications like insulin resistance. We have shown that adipose tissue EVs can stimulate macrophages to secrete pro- (IL-6, TNF-α, MIP-1α) and anti-inflammatory (IL-10, IL-1RA) cytokines, which subsequently (or other unmeasured cytokines) induced insulin resistance in in vitro differentiated adipocytes (chapter 4; Figure 1). Adipokine profiling of adipocyte-EVs revealed the presence of numerous adipokines including macrophage migration inhibitory factor (MIF) and adiponectin. Besides a paracrine insulin resistant effect, adipose tissue EVs were also quantitatively related to systemic insulin resistance. To follow up on the association of adipose tissue secreted EVs and systemic insulin resistance, we have studied the presence of adipose tissue EVs in human plasma, and the potential of human adipose tissue EVs to interfere with insulin signaling of liver and skeletal muscle cells which is described in chapter 5. As adiponectin is regarded an adipocyte-specific marker (50), the presence of adiponectin-positive EVs in human plasma encouraged us to hypothesize that human adipose tissue EVs might very well reach the circulation and could therefore be systemic communicators between adipose tissue and liver or skeletal muscle. When studying the effect of adipose tissue EVs, derived from 16 different patients, on insulin signaling in hepatocytes, the adipose tissue EVs of the majority of patients (12 out of 16) showed insulin inhibitory effects, which was related to higher levels of extracellular vesicle associated IL-6, MIF and monocyte chemotactic protein-1 (MCP-1). Of these adipokines, MIF, an adipokine with chemotactic properties that has been associated with IR (51), is a

161

8

CHAPTER 8

very interesting candidate protein by which adipose tissue EVs could activate macrophages (chapter 4), as well as interfere with insulin signaling in hepatocytes (chapter 5).

Adipokine profiling of adipocyte EVs revealed that MIF was detected primarily in lysed adipocyte EVs and not in unlysed supernatant (chapter 4), indicating that its secretion by adipocytes occurs mainly via EVs. This interesting finding complements the current understanding that MIF is secreted via a non-classical export route (52). Given that (i) MIF is secreted via EVs, (ii) EVs can activate surrounding cells via fusion with the target plasma membrane, and (iii) MIF exerts its effects by activating an intracellular protein (53), adipocyte EV-associated MIF may very well play a role in ATM activation preceding IR. However, studies involving the intracellular actions of MIF showed that coupling of MIF to its intracellular receptor JAB1 inhibits transcription of pro-inflammatory genes by AP-1, and inhibits the inhibition of c-Jun N-terminal kinases (JNK) (Figure 1) indicating anti-inflammatory effects of MIF (53). As MIF deficiency has repeatedly been associated with adipose tissue inflammation and insulin resistance (51,54,55), these are confusing findings and it remains unclear whether (EV-associated) MIF is a causal factor in insulin resistance.

Another mechanism by which adipose tissue EVs could contribute to insulin resistance might be via IL-6, which was shown to negatively impact insulin signaling by promoting serine phosphorylation of insulin-receptor substrate (IRS) in human hepatocytes (56).

Upon insulin binding to its receptor in liver, two major pathways are activated involving IRS1 and IRS 2, which via phosphorylation of AKT shut down gluconeogenesis (by inhibition of transcription of gluconeogenesis genes by FOXO1) and fatty acid oxidation (via Foxa2) (57). In our experiments, the adipose tissue EVs that inhibited phosphorylation of AKT, as well induced transcription of genes involved in gluconeogenesis, of which the degree of AKT phosphorylation was related to EV IL-6 levels. As a direct inhibitory role of IL-6 on AKT phosphorylation was shown in HepG2 hepatocytes (56), EV-derived IL-6 might as well exert such effects. It remains unknown how EV-associated IL-6 would be able to bind its receptor on the recipient cell, however as the IL-6 receptor is anchored to the membrane, EV-IL6 would putatively be presented on the outside of the vesicle (Figure 1).

In contrast to the findings described above, the adipose tissue EVs of a minority of patients (3 out of 16) induced insulin sensitizing effects in liver and muscle cells. As adiponectin is present in vesicles derived from in vitro differentiated human adipocytes (chapter 4), ex vivo incubated human adipose tissue and circulating human plasma (chapter 5), it is a tempting to hypothesize that adipocyte EVs could transport adiponectin to target tissues thereby providing an alternative endocrine route for adiponectin in metabolism and insulin sensitivity. However, we could not detect a relation between adiponectin content of EVs and AKT phosphorylation which might be due to the low number of patients we were able to analyze. Furthermore, it remains unknown whether adiponectin is packaged into EVs via regulated pathways, and in what isoform and location adiponectin is present in these vesicles. For EV-associated adiponectin to have a biological function, it should be able to bind to the adiponectin receptors in its soluble form. If

162

General discussion

adiponectin is packaged intravesicular, could it have an endocrine function? Theoretically, adipocyte EVs could as well be equipped with the adiponectin receptor in the vesicular membrane, by which adiponectin could be bound as well either during the endosomal route or by scavenging extracellular or circulating adiponectin, which might bind adiponectin receptor on its target cells upon cleavage from the vesicle receptor by proteases (Figure 1).

In conclusion, in chapter 4 and 5 we provide evidence for the secretion of EVs by human adipocytes and adipose tissue, which might contribute to local insulin resistance in reciprocal signaling with macrophages (chapter 4), and systemic insulin resistance by direct interference with insulin signaling in liver cells (chapter 5).

PART THREE Biomarkers for obesity-induced metabolic or cardiovascular disease The perception of the role of adipose tissue changed dramatically since the 1950s, when Kennedy provided the first hypothesis that adipose tissue might secrete bioactive factors able to communicate with the hypothalamus to regulate food control, which up till then was believed to be regulated by temperature (58). However, it took another 40 years until the actual bioactive factor (leptin) responsible for that communication was discovered (59).

Since then, the number of fat-derived bioactive factors shown to play a role in metabolic homeostasis and disturbances have piled up (60-63). As a result of obesity, supraphysiological levels of these metabolic substrates, including adipokines and free fatty acids, were shown to have deleterious effects on peripheral tissues such as liver and muscle causing insulin resistance and low grade systemic inflammation, which subsequently contribute to atherosclerosis (64,65).

The discovery of bioactive factors linking obesity to cardiovascular disease held great promises for diagnostics. Over the past decades, biomarkers are increasingly utilized to improve patient care for prediction of disease outcome, diagnosis of disease or evaluation of treatment efficacy (66,67). Results from knock out mouse models raised promising expectations for development of biomarkers in obesity, though only few have been successfully developed so far (68,69). Adiponectin was regarded a very promising biomarker for obesity-induced cardiovascular disease (70), as this protein is almost exclusively secreted by adipocytes (50) and exerts inulin sensitizing and anti-inflammatory effects. In obesity, expression of adiponectin is down-regulated, possibly via inhibitory effects of TNF-Îą of which levels are elevated in obesity (71,72). Lower systemic adiponectin levels have been associated with insulin resistance, atherosclerosis and type 2 diabetes (73-75).

To investigate the potential role of adiponectin as a biomarker for cardiovascular disease, we reviewed current literature describing the association of plasma levels of adiponectin with risk for incident coronary heart disease (CHD) or ischemic stroke in the general population (chapter 7). This meta-analysis, including 14 articles for CHD and 2 for ischemic stroke, revealed that there is no relation between circulating levels of adiponectin and

163

8

CHAPTER 8

Chapter 4

M1

M2

IL-6 TNF-α MIP-1α

IL-10 IL-1RA

Insulin Receptor

Low grade systemic inflammation

Chapter 5 CD14

Chapter 6

LPS

Insulin resistance IL-6 IL6R

IRS AKT

MIF JAB1

JNK

AP-1

164

Insulin Receptor

AdipoR2

PPARα NFκB FOXO

Inflammatory genes Gluconeogenesis

General discussion

incident CHD or stroke. A more recent meta-analysis exploring the association between adiponectin and total stroke (ischemic and hemorrhagic) confirmed the lack of a direct association between circulating adiponectin and incident cardiovascular disease (76). These results indicate that the etiologic role of adiponectin in development of cardiovascular disease might be indirect, putatively via development of type 2 diabetes. It has been shown that adiponectin levels are inversely related with the risk of type 2 diabetes in an enhanced meta-analysis (75). However, also in diabetic patients, there was no relation between systemic adiponectin levels and incident cardiovascular disease (77). It could be possible that different isoforms of adiponectin are differently related to incident cardiovascular disease. High molecular weight adiponectin is considered to be the biological active form of adiponectin, which has been shown to suppress endothelial cell apoptosis, in contrast to other adiponectin isoforms (78). However, cohort studies which have prospectively associated HMW adiponectin with incident cardiovascular disease could neither identify HMW adiponectin as independent predictor of CHD or stroke (79,80). Therefore, despite all anti-inflammatory, insulin sensitizing and endothelial function effects of adiponectin observed in vivo or in vitro, systemic levels of adiponectin only do not seem to cover enough of the underlying pathological activity to be used as a biomarker for obesity-induced cardiovascular disease. Furthermore, increased understanding of underlying mechanisms points towards a multifactorial and interdependent process involving genetic, epigenetic and molecular markers (81). Conjointly with a reduction in adiponectin, numerous other molecules are up regulated including tumor necrosis factor alpha (TNFÎą), C-reactive protein (CRP), free fatty acids and microRNAs (82-84). To provide diagnostic tools for such a complex disease as obesity, we might need to search for a multifactorial approach which may be found in EVs. As described above, EVs are tailor-made packages composed of lipids, proteins and mRNAs which reflect the pathological state of the donor cells they are derived from. These transport vehicles are able to communicate complex biological signals from one cell to another, over short as well as long distances. Therefore, EVs offer novel opportunities as biomarkers for obesity-induced cardiovascular diseases (81). The general concept of EVs as biomarkers for diseases, and the potential of EVs or EV-associated proteins as biomarker for cardiovascular diseases has been shown and discussed multiple times (85-87). However, the potential of adipose tissue EVs as diagnostic tool for obesity-induced metabolic disease is a largely unexplored area. In chapter 6, we have explored the potential of EV-associated markers, which were positively related to recurrent vascular events (87), to serve as biomarkers for obesity-induced metabolic disease in a cohort of 1012 patients with manifest cardiovascular disease. Two of these markers, cystatin C and CD14, both secreted by adipose tissue (88), are repeatedly associated with obesity, diabetes and metabolic syndrome in the general population (89-93). In contrast with other studies, neither obesity nor insulin resistance was related to higher levels of vesicle-associated cystatin C in our cohort of patients with manifest vascular disease. However, EV-cystatin C levels were strongly related to low grade systemic inflammation, lower HDL-cholesterol levels and metabolic syndrome. Therefore, we propose that also in patients with manifest

165

8

CHAPTER 8

cardiovascular disease, cystatin C might be a biomarker for metabolic dysfunction, which often precedes recurrent events (94). In contrast, visceral obesity was negatively related to EV-CD14 levels in men with clinically manifest vascular disease. Furthermore, higher EV-CD14 levels were related to higher HDL-cholesterol levels, higher adiponectin levels, lower insulin resistance and a reduced risk for incident type 2 diabetes. However, in the same patients these EV-CD14 levels were related to low grade systemic inflammation and an increased risk for recurrent myocardial infarction, ischemic stroke and mortality (87).

These data indicate that vesicle-associated CD14 might be a protective biomarker for obesity-induced metabolic disease, as well as non-metabolically induced cardiovascular disease. Interestingly, CD14 circulates in a soluble and membrane bound form. Membrane bound CD14 is expressed mainly in myeoloid cells and macrophages, and transduces inflammatory signals via interaction with protein tyrosine kinases (95). Soluble CD14 is expressed in hepatocytes and adipocytes and lacks the glycosylphosphatidylinositol anchor, therefore unable to transduce signals trough the membrane. However, soluble CD14 does interact with the same inflammatory signals for membrane CD14 such as LPS, and the observation has been made that high levels of soluble CD14 buffers these signals avoiding their exposure with membrane bound (macrophage)-CD14 (96). This phenomenon might explain how adipose tissue released soluble CD14 could be etiologically related to a lower incidence of diabetes (Figure 1). A protective effect of CD14 for insulin resistance has also been showed in mice injected with soluble CD14 (97), while mice lacking membrane bound CD14 were protected from insulin resistance by diet-induced obesity as well (93,97).

Although adipocyte secreted soluble CD14 might explain a protective effect on diabetes, it is unclear how soluble CD14 can buffer circulating inflammatory signals when it is incorporated in extracellular vesicles, unless the vesicle membrane-bound CD14 also exerts a buffering function when it is not able to transduce the signals via intracellular pathways.

However, this would not explain how the same EV-CD14 levels might then be etiologically related to cardiovascular disease via inflammation. Therefore, these data are puzzling to say the least, and it would be very interesting to study whether vesicle-associated CD14 is present in soluble form, membrane-bound or both, and subsequently investigate its functional effects in an in vitro or in vivo setting.

Closing remarks and future perspectives Considerable progress has been made since the first discovery of the endocrine functions of adipose tissue. However, while the obesity epidemic develops at an alarming rate, scientifically we are still far behind with regard to diagnostic and therapeutic actions.

Therefore, there is a pressing need to reconsider our current anti-obesity strategies to prevent obesity-induced comorbidities and death, which requires further insights in the underlying molecular mechanisms.

In this chapter, some potential mechanisms underlying our findings have been presented.

The specific relation between inflamed mesenteric fat and systemic insulin resistance might be due to reciprocal signaling between mesenteric fat secreted factors and the intestinal endocrine system. Although direct clinical application of our findings is still a few

166

General discussion

steps away, our data should urge scientists involved in obesity research to further investigate the potential different contributions of omental and mesenteric fat to obesityinduced metabolic disease in order to develop targeted biomarkers or drugs. The inter-organ communication between mesenteric fat and the intestine should be further investigated to identify insulin resistance inducing factors secreted by mesenteric adipose tissue. A valuable approach would be to study the effect of distinct fractions of conditioned medium from inflamed mesenteric fat on intestinal cells in a stepwise manner. By separating conditioned medium fractions (proteins, lipids or extracellular vesicles) based on size, polarity and density, responsible factors could be identified. Potential factors might be adipose tissue-derived extracellular vesicles, which can contribute to local insulin resistance in a paracrine manner, and are able to induce insulin resistance in hepatocytes as shown in chapter 4 and 5. A very interesting outstanding question is whether extracellular vesicles released from omental and mesenteric fat differ from each other, either by number or composition, and if they might play a role in inter-organ communication with liver and the intestine.

EVs hold a great potential for clinical applications, both at diagnostic as well as therapeutic levels. In this chapter, we have proposed a few signaling pathways through which EVassociated adipokines might influence insulin signaling, though these are highly speculative.

However, although much is unknown and the possibilities to study the contribution of EVs and selective proteins/RNAs on EVs are limited and complex, the etiological role as well as biomarker potential of adipose tissue derived EVs in obesity-induced metabolic disease should be further explored. The concentration of specific adipokines of EVs, and possibly the combination of a set of adipokines within the EVs does seem to matter in the functional outcome of EVs (chapter 5). Proteomic studies comparing the membranous versus the intravesicular compartments of EVs could indicate the location of candidate proteins and thereby their putative mechanism of action. Alongside, the role of distinct EV-associated candidate proteins (Il-6, MIF, adiponectin, leptin) in hepatic insulin resistance could be studied by stimulation of hepatocytes with EVs isolated from adipocyte-knock down cell lines, lacking different candidate adipokines. Furthermore, recipient receptors such as TLR4 or IL6R in hepatocytes or GLP1R on enteroendocrine cells could be blocked as well to study different signaling routes by which EVs induce insulin resistance.

More mechanistic insights will also be crucial in finding a powerful biomarker for obese patients at risk. As it is generally believed that the balance between pro-and antiinflammatory factors in adipose tissue determines the pro-or anti-inflammatory milieu and thereby the onset for adipose tissue dysfunction, it is imaginable that within adipose tissue EVs the same balance takes place, only the message of this balance is delivered both in a paracrine and endocrine fashion. Therefore, more efforts should be undertaken in defining the composition of adipose tissue EVs (from distinct adipose tissue depots) and the relation of distinct compositions to various disease outcomes such as insulin resistance. This way, EVs could serve as biomarkers in which not just 1 marker is used in attempt to predict disease outcome, but a set of markers all packaged into 1 vesicle. Furthermore, once a detrimental EV composition is known, there might be ways to manipulate that composition,

167

8

CHAPTER 8

by preventing packaging of pro-inflammatory factors such as IL-6 or MCP-1 and stimulation of packaging anti-inflammatory factors such as adiponectin. Furthermore, by engineering the composition of EVs, the delivery of EVs to target cells might as well be adjusted, enabling targeted delivery of specific cargo.

Unfortunately, there still are huge challenges to overcome before it is possible to interfere with packaging of proteins or with vesicle release of EVs from human ex vivo material, and thereby study the specific role of lipids or adipokines in EV-stimulated insulin resistance.

This is a limitation in EV research at this point, and still hampers identification of their physiological relevance in vivo (42).

Nevertheless, exploiting the composition of adipose tissue EVs would enable multiparameter biomarker development which seems crucial for complex diseases such as obesity-induced cardiometabolic disease. Furthermore, manipulation of the contents and binding specificity of adipose tissue-EVs hold true therapeutic potential for targeted delivery in vivo.

168

General discussion

REFERENCES 1. Enzi G, Busetto L, Inelmen EM, Coin A, Sergi G. Historical perspective: Visceral obesity and related comorbidity in joannes baptista morgagni's 'de sedibus et causis morborum per anatomen indagata'. Int J Obes Relat Metab Disord. 2003 Apr;27(4):534-5.

2. Haslam D, Rigby N. A long look at obesity. Lancet. 2010 Jul 10;376(9735):85-6.

3. Speakman JR. If body fatness is under physiological regulation, then how come we have an obesity epidemic? Physiology (Bethesda). 2014 Mar;29(2):88-98.

4. Yusuf S, Hawken S, Ounpuu S, Bautista L, Franzosi MG, Commerford P, et al. Obesity and the risk of myocardial infarction in 27,000 participants from 52 countries: A case-control study.

Lancet. 2005 Nov 5;366(9497):1640-9.

5. Ford ES. Prevalence of the metabolic syndrome defined by the international diabetes federation among adults in the U.S. Diabetes Care. 2005 Nov;28(11):2745-9.

6. Kral JG, Kava RA, Catalano PM, Moore BJ. Severe obesity: The neglected epidemic. Obes Facts. 2012 Apr 25;5(2):254-69.

7. Ogilvie D, Hamlet N. Obesity: The elephant in the corner. BMJ. 2005 Dec 24;331(7531):1545-8.

8. Wald D, Teucher B, Dinkel J, Kaaks R, Delorme S, Boeing H, et al. Automatic quantification of subcutaneous and visceral adipose tissue from whole-body magnetic resonance images suitable for large cohort studies. J Magn Reson Imaging. 2012 Aug 21.

9. Fonvig CE, Bille DS, Chabanova E, Nielsen TR, Thomsen HS, Holm JC. Muscle fat content and abdominal adipose tissue distribution investigated by magnetic resonance spectroscopy and imaging in obese children and youths. Pediatr Rep. 2012 Jan 2;4(1):e11.

10. Ren J, Dimitrov I, Sherry AD, Malloy CR. Composition of adipose tissue and marrow fat in humans by 1H NMR at 7 tesla. J Lipid Res. 2008 Sep;49(9):2055-62.

11. Fischer-Posovszky P, Wang QA, Asterholm IW, Rutkowski JM, Scherer PE. Targeted deletion of adipocytes by apoptosis leads to adipose tissue recruitment of alternatively activated M2 macrophages. Endocrinology. 2011 Aug;152(8):3074-81.

12. Gealekman O, Guseva N, Hartigan C, Apotheker S, Gorgoglione M, Gurav K, et al. Depotspecific differences and insufficient subcutaneous adipose tissue angiogenesis in human obesity. Circulation. 2011 Jan 18;123(2):186-94.

13. Bigornia SJ, Farb MG, Mott MM, Hess DT, Carmine B, Fiscale A, et al. Relation of depot-specific adipose inflammation to insulin resistance in human obesity. Nutr Diabetes. 2012 Mar 5;2:e30.

14. Fain JN, Sacks HS, Bahouth SW, Tichansky DS, Madan AK, Cheema PS. Human epicardial adipokine messenger RNAs: Comparisons of their expression in substernal, subcutaneous, and omental fat. Metabolism. 2010 Jan 29.

15. Fain JN, Madan AK, Hiler ML, Cheema P, Bahouth SW. 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. 2004 May;145(5):2273-82.

16. Tchkonia T, Giorgadze N, Pirtskhalava T, Tchoukalova Y, Karagiannides I, Forse RA, et al. Fat depot origin affects adipogenesis in primary cultured and cloned human preadipocytes. Am J Physiol Regul Integr Comp Physiol. 2002 May;282(5):R1286-96.

17. Villaret A, Galitzky J, Decaunes P, Esteve D, Marques MA, Sengenes C, et al. Adipose tissue endothelial cells from obese human subjects: Differences among depots in angiogenic, metabolic, and inflammatory gene expression and cellular senescence. Diabetes. 2010 Nov;59(11):2755-63.

18. Tchkonia T, Tchoukalova YD, Giorgadze N, Pirtskhalava T, Karagiannides I, Forse RA, et al.

Abundance of two human preadipocyte subtypes with distinct capacities for replication, adipogenesis, and apoptosis varies among fat depots. Am J Physiol Endocrinol Metab. 2005 Jan;288(1):E267-77.

19. Arner P. Differences in lipolysis between human subcutaneous and omental adipose tissues.

Ann Med. 1995 Aug;27(4):435-8.

20. Chatterjee TK, Stoll LL, Denning GM, Harrelson A, Blomkalns AL, Idelman G, et al.

Proinflammatory phenotype of perivascular adipocytes: Influence of high-fat feeding. Circ Res. 2009 Feb 27;104(4):541-9.

169

8

CHAPTER 8

21. Rajsheker S, Manka D, Blomkalns AL, Chatterjee TK, Stoll LL, Weintraub NL. Crosstalk between perivascular adipose tissue and blood vessels. Curr Opin Pharmacol. 2010 Apr;10(2):191-6.

22. Sell H, Divoux A, Poitou C, Basdevant A, Bouillot JL, Bedossa P, et al. Chemerin correlates with markers for fatty liver in morbidly obese patients and strongly decreases after weight loss induced by bariatric surgery. J Clin Endocrinol Metab. 2010 Jun;95(6):2892-6.

23. Mathew M, Tay E, Cusi K. Elevated plasma free fatty acids increase cardiovascular risk by inducing plasma biomarkers of endothelial activation, myeloperoxidase and PAI-1 in healthy subjects. Cardiovasc Diabetol. 2010 Feb 16;9:9.

24. Catalano KJ, Stefanovski D, Bergman RN. Critical role of the mesenteric depot versus other intra-abdominal adipose depots in the development of insulin resistance in young rats. Diabetes.

2010 Jun;59(6):1416-23.

25. de Oliveira CC, Acedo SC, Pedrazzoli J,Jr, Saad MJ, Gambero A. Depot-specific alterations to insulin signaling in mesenteric adipose tissue during intestinal inflammatory response. Int Immunopharmacol. 2009 Apr;9(4):396-402.

26. Lam YY, Ha CW, Campbell CR, Mitchell AJ, Dinudom A, Oscarsson J, et al. Increased gut permeability and microbiota change associate with mesenteric fat inflammation and metabolic dysfunction in diet-induced obese mice. PLoS One. 2012;7(3):e34233.

27. Peyrin-Biroulet L, Chamaillard M, Gonzalez F, Beclin E, Decourcelle C, Antunes L, et al.

Mesenteric fat in crohn's disease: A pathogenetic hallmark or an innocent bystander? Gut. 2007 Apr;56(4):577-83.

28. Ludwig T, Worsch S, Heikenwalder M, Daniel H, Hauner H, Bader BL. Metabolic and immunomodulatory effects of n-3 fatty acids are different in mesenteric and epididymal adipose tissue of diet-induced obese mice. Am J Physiol Endocrinol Metab. 2013 Jun 1;304(11):E114056.

29. Conde J, Scotece M, Gomez R, Gomez-Reino JJ, Lago F, Gualillo O. At the crossroad between immunity and metabolism: Focus on leptin. Expert Rev Clin Immunol. 2010 Sep;6(5):801-8.

30. Yang YK, Chen M, Clements RH, Abrams GA, Aprahamian CJ, Harmon CM. Human mesenteric adipose tissue plays unique role versus subcutaneous and omental fat in obesity related diabetes. Cell Physiol Biochem. 2008;22(5-6):531-8.

31. Rabe K, Lehrke M, Parhofer KG, Broedl UC. Adipokines and insulin resistance. Mol Med. 2008 Nov-Dec;14(11-12):741-51.

32. Evans J, Goedecke JH, Soderstrom I, Buren J, Alvehus M, Blomquist C, et al. Depot- and ethnic-specific differences in the relationship between adipose tissue inflammation and insulin sensitivity. Clin Endocrinol (Oxf). 2011 Jan;74(1):51-9.

33. Anini Y, Brubaker PL. Role of leptin in the regulation of glucagon-like peptide-1 secretion.

Diabetes. 2003 Feb;52(2):252-9.

34. Anderlova K, Kremen J, Dolezalova R, Housova J, Haluzikova D, Kunesova M, et al. The influence of very-low-calorie-diet on serum leptin, soluble leptin receptor, adiponectin and resistin levels in obese women. Physiol Res. 2006;55(3):277-83.

35. Herrera MF, Pantoja JP, Velazquez-Fernandez D, Cabiedes J, Aguilar-Salinas C, Garcia-Garcia E, et al. Potential additional effect of omentectomy on metabolic syndrome, acute-phase reactants, and inflammatory mediators in grade III obese patients undergoing laparoscopic roux-en-Y gastric bypass: A randomized trial. Diabetes Care. 2010 Jul;33(7):1413-8.

36. Fabbrini E, Tamboli RA, Magkos F, Marks-Shulman PA, Eckhauser AW, Richards WO, et al.

Surgical removal of omental fat does not improve insulin sensitivity and cardiovascular risk factors in obese adults. Gastroenterology. 2010 Aug;139(2):448-55.

37. Lima MM, Pareja JC, Alegre SM, Geloneze SR, Kahn SE, Astiarraga BD, et al. Visceral fat resection in humans: Effect on insulin sensitivity, beta-cell function, adipokines, and inflammatory markers. Obesity (Silver Spring). 2013 Mar;21(3):E182-9.

38. Aoki N, Jin-no S, Nakagawa Y, Asai N, Arakawa E, Tamura N, et al. Identification and characterization of microvesicles secreted by 3T3-L1 adipocytes: Redox- and hormonedependent induction of milk fat globule-epidermal growth factor 8-associated microvesicles.

Endocrinology. 2007 Aug;148(8):3850-62.

39. Muller G. Take-over: Multiple mechanisms of inter-adipocyte communication. J Mol Cell Biol.

2011 Apr;3(2):81-90.

170

General discussion

40. Pan BT, Johnstone RM. Fate of the transferrin receptor during maturation of sheep reticulocytes in vitro: Selective externalization of the receptor. Cell. 1983 Jul;33(3):967-78.

41. Nolte-'t Hoen EN, Wauben MH. Immune cell-derived vesicles: Modulators and mediators of inflammation. Curr Pharm Des. 2012;18(16):2357-68.

42. Raposo G, Stoorvogel W. Extracellular vesicles: Exosomes, microvesicles, and friends. J Cell Biol. 2013 Feb 18;200(4):373-83.

43. Thery C. Exosomes: Secreted vesicles and intercellular communications. F1000 Biol Rep. 2011;3:15.

44. Thery C, Zitvogel L, Amigorena S. Exosomes: Composition, biogenesis and function. Nat Rev Immunol. 2002 Aug;2(8):569-79.

45. Thery C, Amigorena S. The cell biology of antigen presentation in dendritic cells. Curr Opin Immunol. 2001 Feb;13(1):45-51.

46. Mathivanan S, Ji H, Simpson RJ. Exosomes: Extracellular organelles important in intercellular communication. J Proteomics. 2010 Sep 10;73(10):1907-20.

47. Deng ZB, Poliakov A, Hardy RW, Clements R, Liu C, Liu Y, et al. Adipose tissue exosomelike vesicles mediate activation of macrophage-induced insulin resistance. Diabetes. 2009 Nov;58(11):2498-505.

48. Ogawa R, Tanaka C, Sato M, Nagasaki H, Sugimura K, Okumura K, et al. Adipocyte-derived microvesicles contain RNA that is transported into macrophages and might be secreted into blood circulation. Biochem Biophys Res Commun. 2010 Aug 6;398(4):723-9.

49. Muller G, Schneider M, Biemer-Daub G, Wied S. Microvesicles released from rat adipocytes and harboring glycosylphosphatidylinositol-anchored proteins transfer RNA stimulating lipid synthesis. Cell Signal. 2011 Jul;23(7):1207-23.

50. Scherer PE, Williams S, Fogliano M, Baldini G, Lodish HF. A novel serum protein similar to C1q, produced exclusively in adipocytes. J Biol Chem. 1995 Nov 10;270(45):26746-9.

51. Koska J, Stefan N, Dubois S, Trinidad C, Considine RV, Funahashi T, et al. mRNA concentrations of MIF in subcutaneous abdominal adipose cells are associated with adipocyte size and insulin action. Int J Obes (Lond). 2009 Aug;33(8):842-50.

52. Flieger O, Engling A, Bucala R, Lue H, Nickel W, Bernhagen J. Regulated secretion of macrophage migration inhibitory factor is mediated by a non-classical pathway involving an ABC transporter. FEBS Lett. 2003 Sep 11;551(1-3):78-86.

53. Kleemann R, Hausser A, Geiger G, Mischke R, Burger-Kentischer A, Flieger O, et al. Intracellular action of the cytokine MIF to modulate AP-1 activity and the cell cycle through Jab1. Nature.

2000 Nov 9;408(6809):211-6.

54. Verschuren L, Kooistra T, Bernhagen J, Voshol PJ, Ouwens DM, van Erk M, et al. MIF deficiency reduces chronic inflammation in white adipose tissue and impairs the development of insulin resistance, glucose intolerance, and associated atherosclerotic disease. Circ Res. 2009 Jul 2;105(1):99-107.

55. Kleemann R, Bucala R. Macrophage migration inhibitory factor: Critical role in obesity, insulin resistance, and associated comorbidities. Mediators Inflamm. 2010;2010:610479.

56. Senn JJ, Klover PJ, Nowak IA, Mooney RA. Interleukin-6 induces cellular insulin resistance in hepatocytes. Diabetes. 2002 Dec;51(12):3391-9.

57. Montminy M, Koo SH. Diabetes: Outfoxing insulin resistance? Nature. 2004 Dec 23;432(7020):958-9.

58. KENNEDY GC. The role of depot fat in the hypothalamic control of food intake in the rat. Proc R Soc Lond B Biol Sci. 1953 Jan 15;140(901):578-96.

59. Zhang Y, Proenca R, Maffei M, Barone M, Leopold L, Friedman JM. Positional cloning of the mouse obese gene and its human homologue. Nature. 1994 Dec 1;372(6505):425-32.

60. Ouchi N, Parker JL, Lugus JJ, Walsh K. Adipokines in inflammation and metabolic disease. Nat Rev Immunol. 2011 Feb;11(2):85-97.

61. Shah A, Mehta N, Reilly MP. Adipose inflammation, insulin resistance, and cardiovascular disease. JPEN J Parenter Enteral Nutr. 2008 Nov-Dec;32(6):638-44.

62. Nishimura S, Manabe I, Nagai R. Adipose tissue inflammation in obesity and metabolic syndrome. Discov Med. 2009 Aug;8(41):55-60.

171

8

CHAPTER 8

63. RANDLE PJ, GARLAND PB, HALES CN, NEWSHOLME EA. The glucose fatty-acid cycle. its role in insulin sensitivity and the metabolic disturbances of diabetes mellitus. Lancet. 1963 Apr 13;1(7285):785-9.

64. Samuel VT, Shulman GI. Mechanisms for insulin resistance: Common threads and missing links. Cell. 2012 Mar 2;148(5):852-71.

65. Nakamura K, Fuster JJ, Walsh K. Adipokines: A link between obesity and cardiovascular disease. J Cardiol. 2013 Dec 16.

66. Musaad S, Haynes EN. Biomarkers of obesity and subsequent cardiovascular events. Epidemiol Rev. 2007;29:98-114.

67. Apple FS, Murakami MM. The diagnostic utility of cardiac biomarkers in detecting myocardial infarction. Clin Cornerstone. 2005;7 Suppl 1:S25-30.

68. Johansson K, Neovius M, Hemmingsson E. Effects of anti-obesity drugs, diet, and exercise on weight-loss maintenance after a very-low-calorie diet or low-calorie diet: A systematic review and meta-analysis of randomized controlled trials. Am J Clin Nutr. 2014 Jan;99(1):14-23.

69. Shi Y, Burn P. Lipid metabolic enzymes: Emerging drug targets for the treatment of obesity. Nat Rev Drug Discov. 2004 Aug;3(8):695-710.

70. Watanabe T, Sato K, Itoh F, Wakabayashi K, Shichiri M, Hirano T. Endogenous bioactive peptides as potential biomarkers for atherosclerotic coronary heart disease. Sensors (Basel).

2012;12(4):4974-85.

71. Degawa-Yamauchi M, Moss KA, Bovenkerk JE, Shankar SS, Morrison CL, Lelliott CJ, et al.

Regulation of adiponectin expression in human adipocytes: Effects of adiposity, glucocorticoids, and tumor necrosis factor alpha. Obes Res. 2005 Apr;13(4):662-9.

72. Ahn J, Lee H, Kim S, Ha T. Resveratrol inhibits TNF-alpha-induced changes of adipokines in 3T3L1 adipocytes. Biochem Biophys Res Commun. 2007 Dec 28;364(4):972-7.

73. Ouchi N, Kihara S, Funahashi T, Matsuzawa Y, Walsh K. Obesity, adiponectin and vascular inflammatory disease. Curr Opin Lipidol. 2003 Dec;14(6):561-6.

74. Okamoto Y, Kihara S, Ouchi N, Nishida M, Arita Y, Kumada M, et al. Adiponectin reduces atherosclerosis in apolipoprotein E-deficient mice. Circulation. 2002 Nov 26;106(22):2767-70.

75. Li S, Shin HJ, Ding EL, van Dam RM. Adiponectin levels and risk of type 2 diabetes: A systematic review and meta-analysis. JAMA. 2009 Jul 8;302(2):179-88.

76. Arregui M, Buijsse B, Fritsche A, di Giuseppe R, Schulze MB, Westphal S, et al. Adiponectin and risk of stroke: Prospective study and meta-analysis. Stroke. 2014 Jan;45(1):10-7.

77. Wu Z, Cheng Y, Aung LH, Li B. Association between adiponectin concentrations and cardiovascular disease in diabetic patients: A systematic review and meta-analysis. PLoS One. 2013 Nov 4;8(11):e78485.

78. Waki H, Yamauchi T, Kamon J, Ito Y, Uchida S, Kita S, et al. Impaired multimerization of human adiponectin mutants associated with diabetes. molecular structure and multimer formation of adiponectin. J Biol Chem. 2003 Oct 10;278(41):40352-63.

79. Ogorodnikova AD, Wassertheil-Smoller S, Mancuso P, Sowers MR, Rajpathak SN, Allison MA, et al. High-molecular-weight adiponectin and incident ischemic stroke in postmenopausal women: A women's health initiative study. Stroke. 2010 Jul;41(7):1376-81.

80. Sattar N, Watt P, Cherry L, Ebrahim S, Davey Smith G, Lawlor DA. High molecular weight adiponectin is not associated with incident coronary heart disease in older women: A nested prospective case-control study. J Clin Endocrinol Metab. 2008 May;93(5):1846-9.

81. Muller G. Microvesicles/exosomes as potential novel biomarkers of metabolic diseases.

Diabetes Metab Syndr Obes. 2012;5:247-82.

82. Kinet V, Halkein J, Dirkx E, Windt LJ. Cardiovascular extracellular microRNAs: Emerging diagnostic markers and mechanisms of cell-to-cell RNA communication. Front Genet. 2013 Nov 12;4:214.

83. Faber DR, van der Graaf Y, Westerink J, Visseren FL. Increased visceral adipose tissue mass is associated with increased C-reactive protein in patients with manifest vascular diseases.

Atherosclerosis. 2010 Sep;212(1):274-80.

84. Hotamisligil GS, Shargill NS, Spiegelman BM. Adipose expression of tumor necrosis factoralpha: Direct role in obesity-linked insulin resistance. Science. 1993 Jan 1;259(5091):87-91.

172

General discussion

85. Baron M, Boulanger CM, Staels B, Tailleux A. Cell-derived microparticles in atherosclerosis: Biomarkers and targets for pharmacological modulation? J Cell Mol Med. 2012 Jul;16(7):136576.

86. Pant S, Hilton H, Burczynski ME. The multifaceted exosome: Biogenesis, role in normal and aberrant cellular function, and frontiers for pharmacological and biomarker opportunities.

Biochem Pharmacol. 2012 Jun 1;83(11):1484-94.

87. Kanhai DA, Visseren FL, van der Graaf Y, Schoneveld AH, Catanzariti LM, Timmers L, et al.

Microvesicle protein levels are associated with increased risk for future vascular events and mortality in patients with clinically manifest vascular disease. Int J Cardiol. 2013 Oct 3;168(3):2358-63.

88. Naour N, Fellahi S, Renucci JF, Poitou C, Rouault C, Basdevant A, et al. Potential contribution of adipose tissue to elevated serum cystatin C in human obesity. Obesity (Silver Spring). 2009 Dec;17(12):2121-6.

89. Muntner P, Winston J, Uribarri J, Mann D, Fox CS. Overweight, obesity, and elevated serum cystatin C levels in adults in the united states. Am J Med. 2008 Apr;121(4):341-8.

90. Magnusson M, Hedblad B, Engstrom G, Persson M, Nilsson P, Melander O. High levels of cystatin C predict the metabolic syndrome: The prospective malmo diet and cancer study. J Intern Med. 2013 Aug;274(2):192-9.

91. Sahakyan K, Lee KE, Shankar A, Klein R. Serum cystatin C and the incidence of type 2 diabetes mellitus. Diabetologia. 2011 Jun;54(6):1335-40.

92. Poitou C, Dalmas E, Renovato M, Benhamo V, Hajduch F, Abdennour M, et al. CD14dimCD16+ and CD14+CD16+ monocytes in obesity and during weight loss: Relationships with fat mass and subclinical atherosclerosis. Arterioscler Thromb Vasc Biol. 2011 Oct;31(10):2322-30.

93. Roncon-Albuquerque R,Jr, Moreira-Rodrigues M, Faria B, Ferreira AP, Cerqueira C, Lourenco AP, et al. Attenuation of the cardiovascular and metabolic complications of obesity in CD14 knockout mice. Life Sci. 2008 Sep 26;83(13-14):502-10.

94. Vlek AL, van der Graaf Y, Spiering W, Visseren FL, SMART study group. Effect of metabolic syndrome or type II diabetes mellitus on the occurrence of recurrent vascular events in hypertensive patients. J Hum Hypertens. 2008 May;22(5):358-65.

95. Stefanova I, Horejsi V, Ansotegui IJ, Knapp W, Stockinger H. GPI-anchored cell-surface molecules complexed to protein tyrosine kinases. Science. 1991 Nov 15;254(5034):1016-9.

96. Kirkland TN, Viriyakosol S. Structure-function analysis of soluble and membrane-bound CD14.

Prog Clin Biol Res. 1998;397:79-87.

97. Fernandez-Real JM, Perez del Pulgar S, Luche E, Moreno-Navarrete JM, Waget A, Serino M, et al. CD14 modulates inflammation-driven insulin resistance. Diabetes. 2011 Aug;60(8):2179-86.

173

8

APPENDIX Summary Nederlandse samenvatting Dankwoord Contributing authors List of Publications Curriculum Vitae

APPENDIX

SUMMARY Obesity is the result of the expansion of adipose tissue as a consequence of chronic energy excess. Considerable progress has been made since the first discovery of the endocrine functions of adipose tissue. However, while the obesity epidemic develops at an alarming rate, scientifically we are still far behind with regard to diagnostic and therapeutic tools.

Therefore, there is a pressing need to reconsider our current anti-obesity strategies to prevent obesity-induced comorbidities and death, which requires further insights in the underlying molecular mechanisms. In this thesis we aimed to explore current and novel pathways in adipose tissue dysfunction, as a result of obesity, and investigated how they might contribute to metabolic and cardiovascular disease.

Adipose tissue is an interesting but complex tissue, which is heavily challenged in the state of obesity. In Chapter 1, the obesity epidemic is explained, and the concept of adipose tissue as an endocrine organ is introduced. Adipose tissue communicates with other organs through the release of hormones, proteins and free fatty acids. However, extracellular vesicles are signaling entities as well and crucial for communication between cells in our body. Extracellular vesicles include exosomes and microvesicles which are vesicles formed within each cell, and released into the extracellular space upon blebbing of the membrane, or upon fusion of multivesicular bodies with the cell membrane. These vesicles carry lipids, soluble and transmembrane proteins and various RNA species, through which they can activate cells from other organs (target cells). Owing to these properties, extracellular vesicles can be considered as targets for disease biomarkers as well as therapeutic intervention. The potential of extracellular vesicles in adipose tissue dysfunction is introduced in chapter 1. Furthermore, chapter 1 provides the objective and outline of this thesis.

In the first part of this thesis, we focus on the role of distinct adipose tissue depots in obesity-induced metabolic dysfunction. In chapter 2, current knowledge of patho足 physiological mechanisms linking abdominal adipose tissue to obesity-related metabolic dysfunction is reviewed, with a special focus on distinct adipose tissue depots and the role of adipose tissue-derived extracellular vesicles. Adipose tissue consists of different depots which are anatomically linked to distinct organs. Evidence suggests that intrinsic differences as well as interactions of adipose tissue with surrounding organs accounts for different contributions to cardiometabolic disease. Adipose tissue communicates with surrounding and distant organs via multiple mechanisms. Besides the well-known contribution of adipokines and free fatty acids, the potential role of adipose tissue-derived extracellular vesicles in the development of obesity-induced cardiometabolic disease is described in this review. In chapter 3 we explored the inflammatory profile of four different abdominal adipose tissue depots and their relation with parameters of metabolic dysfunction in a clinical study. The study cohort consisted of abdominally lean versus abdominally obese male patients with clinical manifest vascular disease. The most striking finding was the differential contribution of distinct visceral fat depots with metabolic dysfunction.

Perivascular fat was highly active as noted by abundant presence of crown like structures

176

APPENDIX

(activated macrophages surrounding a death adipocyte) and adipokine secretion. However, the number of crown like structures and other morphological characteristics were not related to systemic metabolic dysfunction. In contrast, adipocyte size and crown like structures of both omental and mesenteric fat were related to metabolic derangements.

Interestingly, mesenteric fat morphology was mainly related to insulin resistance, whereas omental fat morphology was more strongly related to higher triglyceride and lower HDLcholesterol levels. These results indicate that distinct visceral adipose tissue depots are intrinsically different and differently related to metabolic diseases. Because of the close proximity of mesenteric fat with the intestine it is tempting to speculate about reciprocal signaling between mesenteric fat and the intestinal endocrine system. The findings in chapter 3 show the necessity to further investigate the potential different contributions of omental and mesenteric fat to obesity-induced metabolic disease in order to develop targeted biomarkers or drugs.

In the second part of this thesis we have studied the ability of human adipocytes and adipose tissue to secrete extracellular vesicles (EVs), which are active signaling vesicles crucial for communication between cells. In obesity, communication between adipocytes and immune cells such as macrophages is a key mechanism in adipose tissue inflammation, subsequently leading to metabolic complications such as insulin resistance. With knowledge of the underlying mechanisms of macrophage activation in adipose tissue, specific drugs may be developed to prevent metabolic complications. In chapter 4 we have shown that EVs released by either human in vitro differentiated adipocytes or ex vivo human adipose tissue explants play a role in the paracrine interaction between adipocytes and macrophages.

We have shown that human adipose tissue EVs can stimulate macrophages to secrete pro- (IL-6, TNF-Îą, MIP-1Îą) and anti-inflammatory (IL-10, IL-1RA) cytokines, which subsequently induced insulin resistance in in vitro differentiated adipocytes. Adipokine profiling of adipocyte-EVs revealed the presence of numerous adipokines including macrophage migration inhibitory factor (MIF) and adiponectin. Besides paracrine signaling causing insulin resistance, the numbers of EVs of visceral but not subcutaneous adipose tissue were also related to systemic insulin resistance. This observation might indicate endocrine signaling of adipose tissue EVs involved in systemic insulin resistance. To follow up on this theory we have studied the presence of adipose tissue EVs in human plasma, and the potential of human adipose tissue EVs to interfere with insulin signaling of liver and skeletal muscle cells, which is described in chapter 5. Adiponectin-positive EVs were isolated from human plasma demonstrating presence of adipocyte-derived EVs in human plasma enabling endocrine functions of AT EVs. Therefore, we studied whether adipose tissue EVs, derived from ex vivo human fat explants, could directly interfere with insulin signaling in liver and muscle cells. Adipose tissue EVs of the majority of patients inhibited insulin signaling in liver cells in vitro, which was related to pro-inflammatory adipokines present in EVs of visceral adipose tissue.

In conclusion, in the second part we have provided evidence for the secretion of EVs by human adipocytes and adipose tissue, which might contribute to local insulin resistance by reciprocal signaling with macrophages (chapter 4), and systemic insulin resistance by

177

A

APPENDIX

direct interference with insulin signaling in liver cells, possibly depending on their adipokine content (chapter 5).

The final part of this thesis describes the relation of adipose tissue-derived bioactive markers with metabolic or cardiovascular disease in large patient populations. The general concept of EVs as biomarkers for diseases, and the potential of EVs or EV-associated proteins as biomarker for cardiovascular diseases have been shown and discussed multiple times. However, the potential of adipose tissue EVs as diagnostic tool for obesity-induced metabolic disease is a largely unexplored area. In chapter 6, we have explored the potential of EV-associated markers, associated with cardiovascular disease, to serve as biomarkers for obesity-induced metabolic disease in a cohort of 1012 patients with manifest cardiovascular disease. We have observed that EV-cystatin C levels were positively related to metabolic complications of obesity, including low-grade systemic inflammation, low HDL-cholesterol levels and metabolic syndrome. Therefore, we propose that the EV-marker cystatin C may be an important biomarker for CVD not only in healthy individuals but importantly also in patients with manifest vascular disease. In contrast, EV-CD14 levels were inversely related to visceral obesity in males and associated with a relative risk reduction for the development of type 2 diabetes. These data indicate that vesicleassociated CD14 might be a protective biomarker for obesity-induced metabolic disease.

In chapter 7, the potential prognostic value of adiponectin for cardiovascular disease is studied. Adiponectin is a protein which is exclusively secreted by adipocytes, has metabolic beneficent properties and is negatively related to obesity, insulin resistance and type 2 diabetes. The aim of chapter 7 was to meta-analyze the relation between plasma concentrations of adiponectin and risk of coronary heart disease and stroke in the general population. This meta-analysis, including 14 articles for CHD and 2 for ischemic stroke, revealed that there is no relation between circulating levels of adiponectin and incident CHD or stroke. Therefore, despite all anti-inflammatory, insulin sensitizing and endothelial function effects of adiponectin observed in vivo or in vitro, systemic levels of adiponectin only do not seem to cover enough of the underlying pathological activity to be used as a biomarker for obesity-induced cardiovascular disease.

Overall, the mechanisms underlying obesity-induced metabolic complications described in this thesis point towards multifactorial and interdependent communication between adipocytes, the immune system and several organs including the liver. In order to develop diagnostic tools for a complex disease as obesity, we might need to establish a multifactorial approach in which the contribution of EVs should also be evaluated. Exploiting the composition of adipose tissue EVs would enable multi-parameter biomarker development which seems crucial for complex diseases such as obesity-induced cardiometabolic disease.

Furthermore, manipulation of the contents and binding specificity of adipose tissue-EVs hold true therapeutic potential for targeted delivery in vivo.

178

APPENDIX

NEDERLANDSE SAMENVATTING Een goed functionerend lichaam kan niet zonder vetweefsel. Zonder dit weefsel raken diverse lichaamssystemen ontregeld, waaronder de opslag van vetten, de glucosehuishouding en tal van hormonale processen. Goed functionerend vetweefsel beschermt tegen het ontstaan van diabetes, metabool syndroom en uiteindelijk hart- en vaatziekten. Vetweefsel wordt dus niet langer beschouwd als orgaan wat alleen vet opslaat, maar als een belangrijk endocrien orgaan. Echter, een teveel aan vetweefsel door overmatige vetzucht (obesitas) kan zorgen voor ontregeling van deze metabole processen.

Het vet raakt dan dysfunctioneel. Het aantal mensen met obesitas is de laatste jaren enorm gegroeid, waardoor wordt gesproken van een obesitas-epidemie. Tegenwoordig overlijden zelfs meer mensen aan de gevolgen van overvoeding dan aan de gevolgen van ondervoeding. Het blijkt voor veel mensen moeilijk om de inname van voeding te beperken of om voeding te verbranden. Daarom moeten we op zoek naar manieren om tekenen van dysfunctionerend vet vroegtijdig te signaleren (voor diagnostische doeleinden) en te behandelen (voor therapeutische doeleinden). Hiervoor is begrip van de biologische processen die leiden tot cardiometabole ziekten als gevolg van dysfunctionerend vet essentieel. Echter, terwijl de obesitas-epidemie zich in een alarmerend tempo ontwikkelt zijn we wetenschappelijk gezien nog steeds ver achter met de ontwikkeling van diagnostische en therapeutische middelen.

In dit proefschrift hebben we vetweefsel en de gevolgen van vetweefseldysfunctie op drie niveaus bestudeerd: middels klinische studies, analyse van gekweekte vetcellen (in vitro) en analyse van menselijke vetweefselbiopten (ex vivo). Hierbinnen presenteren we een nieuw mechanisme in obesitas-geïnduceerde comorbiditeit: de rol van extracellulaire membraanblaasjes.

In hoofdstuk 1 wordt de endocriene rol van vetweefsel geïntroduceerd. Vetweefsel communiceert met andere organen door middel van uitscheiding van vetcel specifieke eiwitten (adipokines) en vrije vetzuren, maar ook door middel van extracellulaire membraanblaasjes. Deze kleine blaasjes worden in de (vet)cel gevormd en, door versmelting met het celmembraan, uitgescheiden in de extracellulaire ruimte. Extracellulaire membraanblaasjes zijn cruciaal in de communicatie tussen cellen in ons lichaam. Ze vervoeren meerdere moleculen tegelijk, zoals talloze lipiden, eiwitten en genetisch materiaal, waarmee de blaasjes cellen van andere organen kan activeren. Deze eigenschappen maken membraanblaasjes interessante kandidaten voor diagnostische en therapeutische doeleinden. Zo kunnen membraanblaasjes belangrijk zijn voor de ontwikkeling van medicijnen, omdat ze een essentiële rol spelen in de communicatie tussen vetcellen en andere cellen die betrokken zijn bij vetweefseldysfunctie. Ten tweede kunnen membraanblaasjes afkomstig van vetweefsel iets vertellen over de staat van de dysfunctie van het vet. Hierdoor kunnen we mogelijk het risico van een patiënt met obesitas beter kunnen monitoren.

In hoofdstuk 2 wordt de huidige kennis over de onderliggende mechanismen van obesitas en metabole dysfunctie besproken, met een speciale focus op verschillende vetweefsel-

179

A

APPENDIX

gebieden en de rol van het vetweefsel afkomstige membraanblaasjes. Vetcellen vervullen een belangrijke taak in de opslag van vetten (triglyceriden), maar ook is er een limiet in de hoeveelheid vetten die een vetcel kan opslaan. In het geval van overmatige vetzucht (obesitas) is er een langdurig te hoog aanbod van energie die moet worden opgeslagen.

Die opslag kan in bestaande vetcellen die gaan uitzetten (hypertrofie) of in nieuw aangemaakte vetcellen (hyperplasie). Als beide opslagmogelijkheden in vetweefsel verzadigd zijn (een grens die bij ieder persoon anders is), wordt de ingenomen energie noodgedwongen in andere organen zoals de lever of in de bloedvaten opgeslagen. Dit zorgt voor een verstoring van de functie van deze andere organen, zoals de leverfalen en vernauwing van de bloedvaten. Daarnaast raken de andere functies van de oververzadigde vetcellen verstoord. Dit wordt vetweefseldysfunctie genoemd. Een gezonde vetcel is gevoelig voor het hormoon insuline, waardoor glucose uit de bloedbaan opgenomen kan worden voor tijden waarin energie nodig is. Dysfunctionerende vetcellen zijn niet langer gevoelig voor insuline (insulineresistentie), met als gevolg dat de bloedglucose stijgt. Dit kan uiteindelijk leiden tot suikerziekte (diabetes). Tot slot produceert gezond vet tal van hormonen en eiwitten (collectief adipokines genoemd), zoals het hongerhormoon leptine en het anti-diabetische adiponectine. Deze laatste regelt doorcommunicatie met andere organen verschillende processen in het lichaam. Dysfunctioneel vet trekt tal van ontstekingscellen aan zoals macrofagen en T cellen, die samen met de vetcellen tal van ontsteking (pro-inflammatoire) eiwitten uitscheiden zoals interleukines (IL-1, IL-6) tumor necrose factor alfa (TNF-α) en resistine. De verschillende typen pro-inflammatoire eiwitten kunnen zorgen voor ontstekingsreacties en bijdragen aan vaatschade, hypertensie en insuline resistentie. Deze combinatie van risicofactoren wordt het metabool syndroom genoemd. Naast de rol van pro-inflammatoire eiwitten en immuun cellen wordt ook de potentiële rol van membraanblaasjes in de ontwikkeling van obesitas geïnduceerde cardiometabool risico beschreven in dit review.

De verstoring van vetweefsel is verschillend op verschillende plekken in het lichaam. In hoofdstuk 3 hebben we ontstekingseigenschappen van vier verschillende vetweefsel gebieden uit de buik onderzocht. Het vetweefsel onder de huid (subcutaan vetweefsel) wordt beschouwd als gezond vet. De vetcellen van dit type vet zijn in staat tot grotere opslag van vetten, zonder hierbij dysfunctioneel te raken. Daarbij scheiden ze veel metabool gunstige hormonen uit. Het vet rondom organen (visceraal vet) is het ‘ongezonde vet’ omdat het minder capaciteit heeft om te anticiperen op toename van energie aanvoer en daarbij meer ontstekingseiwitten uitscheidt. Echter, de verschillende ontstekings­ eigenschappen van verschillende viscerale vetweefselregio’s is minder goed bekend.

Tijdens een operatie aan de grote buikslagader zijn stukjes van het onderhuidse vetweefsel en drie viscerale vetweefsels verzameld. Karakteristieken van die vier vetstukjes werden vergeleken tussen mannen met en mannen zonder overgewicht. De meest opvallende bevinding was dat verschillende viscerale vetweefsel gebieden gerelateerd zijn aan verschillende kenmerken van het metabool syndroom. Vet rondom de aorta bevat veel ‘crown like structures’ (macrofagen die een doodgaande vetcel opruimen, een teken van vetweefseldysfunctie) en scheidt hoge concentraties adipokines uit, maar de morfologische

180

APPENDIX

kenmerken hielden geen verband met systemische metabole dysfunctie. Het vetweefsel tussen de darmen (mesenteriaal vet) en het vetschort (omentum) waren beide wel gerelateerd aan metabole verstoringen. De hoeveelheid ‘crown like structures’ van mesenteriaal vet was hierbij gerelateerd aan insuline resistentie, terwijl ‘crown like structures’ in omentaal vet gerelateerd waren aan hogere triglyceride waarden en lager HDL-cholesterol. Deze resultaten wijzen erop dat verschillende viscerale vetweefsel gebieden anders aan metabole ziekten gerelateerd zijn. Aangezien mesenteriaal vet tussen de darmen ligt, is het interessant om te speculeren over een mogelijk samenspel tussen de darmen en mesenteriaal vet. Wellicht verstoren stoffen die geproduceerd zijn door dysfunctioneel mesenteriaal vet de uitscheiding van insuline stimulerende hormonen van de darm. De bevindingen beschreven in hoofdstuk 3 geeft de noodzaak aan om de potentieel verschillende bijdrage van omentaal en mesenteriaal vet aan obesitasgeïnduceerde metabole stoornissen verder te onderzoeken, zodat specifiekere medicijnen ontwikkeld kunnen worden.

In het tweede deel van dit proefschrift beschrijven we studies over de rol van extracellulaire membraanblaasjes in vetweefsel, vetweefseldysfunctie en insuline resistentie. In hoofdstuk 4 hebben we onderzocht of membraanblaasjes afkomstig van menselijke in vitro gekweekte vetcellen of ex vivo gekweekte vetbiopten een rol kunnen spelen in de activatie van macrofagen. Activatie van macrofagen in dysfunctioneel vetweefsel is cruciaal in de ontwikkeling van insuline resistentie. In dit hoofdstuk hebben wij aangetoond dat membraanblaasjes afkomstig van vet een rol spelen in de paracriene communicatie tussen vetcellen en macrofagen. Membraanblaasjes afkomstig van vet konden macrofagen activeren, die vervolgens insulineresistentie induceerden in in vitro gedifferentieerde vetcellen. Het is nog niet bekend hoe de membraanblaasjes macrofagen precies activeren.

Membraanblaasjes kunnen bijvoorbeeld ontstekingseiwitten bevatten die daar een rol in zouden kunnen spelen. Karakterisering van de ontstekingseiwitten aanwezig op membraanblaasjes en afkomstig van vetweefsel, toonde aanwezigheid aan van talrijke adipokines, waaronder ‘macrophage migration inhibitory factor’ (MIF) en adiponectine.

Naast een lokaal effect van membraanblaasjes op insuline resistentie, was het aantal membraanblaasjes van visceraal vet (maar niet van het onderhuidse vet) ook gerelateerd aan systemische insulineresistentie. Om de relatie tussen membraanblaasjes afkomstig van vetweefsel en systemische insuline resistentie beter te onderzoeken, hebben we vervolgens bestudeerd of de membraanblaasjes ook kunnen interfereren met de signaalroute van insuline in de lever en spiercellen. Dit is beschreven in hoofdstuk 5. Hierbij toonden we eerst aan dat adiponectine-positieve membraanblaasjes aanwezig zijn in het bloed van mensen. Aangezien adiponectine alleen door vetcellen wordt gemaakt, suggereert deze bevinding dat membraanblaasjes van vetweefsel ook in de bloedsomloop terecht komen, en zo op afstand effect op andere organen kunnen hebben. Vervolgens zagen we dat membraanblaasjes van ex vivo vet biopten rechtstreeks de insuline signaalroute van levercellen kunnen verstoren. Het effect op spiercellen was echter minder duidelijk. Ten slotte bleek dat de mate van insuline resistentie in de levercellen afhankelijk is van de concentraties van ontstekingseiwitten, waaronder MIF en MCP-1. Op basis van

181

A

APPENDIX

deze resultaten concludeerden we dat membraanblaasjes afkomstig van vetweefsel kunnen bijdragen aan insuline-resistentie, zowel lokaal (via activatie van macrofagen) als systemisch (via verstoring van de insuline signaalroute in levercellen). Dit is een nieuw mechanisme waarop vetweefsel bij kan dragen aan metabole stoornissen.

In het laatste deel van dit proefschrift is de relatie tussen vetcel specifieke eiwitten en metabole stoornissen en hart- en vaatziekten in grote patiënten populaties onderzocht.

Dysfunctioneel vet kan bijvoorbeeld eiwitten uitscheiden die de mate van vetweefsel­ dysfunctie weerspiegelen, en ook het risico op complicaties voorspellen zoals diabetes en hart en vaatziekten. Zulke eiwitten worden dan biomarkers genoemd. Maar ook membraanblaasjes zouden geschikte biomarkers kunnen zijn, wat al voor veel ziektes aangetoond is. Biomarkers voor obesitas-geïnduceerde ziekte zijn echter nog schaars. Een potentieel interessante biomarker is adiponectine. De concentratie van adiponectine in het bloed is verlaagd in mensen met obesitas, en lage adiponectine waardes zijn weer geassocieerd met het ontstaan van diabetes. Aangezien obesitas en diabetes beide belangrijke risicofactoren zijn voor hart en vaatziekten hebben we in hoofdstuk 7 de potentiële prognostische waarde van adiponectine voor hart en vaatziekten bestudeerd.

In deze meta-analyse (een studie waarin de onderzoeksresultaten van meerdere studies wordt gecombineerd) blijkt dat de adiponectine waardes van in totaal 23.919 patiënten niet zijn gerelateerd aan het ontstaan van hart en vaatziekten. Deze resultaten impliceren dat ondanks alle gunstige effecten van adiponectine op ontstekingen, de insuline gevoeligheid en de vaatwand die zijn waargenomen in vivo (in muisproeven) of in vitro (in een kweekschaaltje), dit hormoon niet geschikt is als biomarker voor obesitas geïnduceerde hart en vaatziekten. Een verklaring hiervoor kan liggen in het feit dat obesitas geïnduceerde hart en vaatziekten multifactorieel bepaald worden (via meerdere en verschillende manieren), en dat een eiwit wat slechts één van de onderliggende routes weerspiegeld niet voldoende informatie geeft over alle processen die geactiveerd zijn in obesitas en die bijdragen aan diabetes en hart en vaatziekten. Om een goede biomarker voor een complexe ziekte als obesitas te vinden, zullen we dus waarschijnlijk toe moeten naar een multifactorieel signaal, wat te vinden is in membraanblaasjes. Zoals hierboven beschreven bevatten membraanblaasjes meerdere moleculen, zoals adiponectine en tal van andere ontstekingseiwitten en vetten. Wellicht dat de combinatie van meerdere typen moleculen, verpakt in één membraanblaasje, een betere afspiegeling geeft van de ziekte staat van een patiënt. De rol van membraanblaasjes als biomarker voor hart en vaatziekten is al meerdere keren aangetoond, maar de potentie van membraanblaasjes voor metabole stoornissen is een grotendeels onontgonnen gebied. In hoofdstuk 6 hebben we de mogelijkheid van membraanblaasjes als biomarker voor metabole stoornissen onderzocht.

Van vier eiwitten, aanwezig in membraanblaasjes uit het bloed, is de relatie met obesitas, het metabool syndroom en het ontstaan van suikerziekte onderzocht in een cohort van 1012 patiënten met hart en vaatziekten. Het eiwit cystatine C op membraanblaasjes was geassocieerd met het hebben van het metabool syndroom en een lager HDL-cholesterol, terwijl CD14 waardes juist verlaagd waren in mannen met obesitas en hogere CD14 geassocieerd met een lagere kans op de ontwikkeling van diabetes type 2. Uit deze

182

APPENDIX

resultaten blijkt dat eiwitten van membraanblaasjes als biomarker zouden kunnen fungeren voor obesitas-geĂŻnduceerde diabetes.

Een beter begrip van de onderliggende mechanismen waarop obesitas kan leiden tot diabetes en hart en vaatziekten zal ons moeten helpen in de ontwikkeling van de juiste medicijnen. De huidige kennis wijst op een multifactorieel en onderling afhankelijk proces, waarin naast lagere hoeveelheid hormonen die gunstig zijn voor je stofwisseling, talrijke andere moleculen juist worden geactiveerd zoals ontstekingseiwitten en vrije vetzuren.

Een dergelijk complexe ziekte vraagt om een multifactoriĂŤle aanpak die kan worden gevonden in membraanblaasjes. Karakterisering van de samenstelling van vetweefselblaasjes zou ons in staat stellen om een multi-parameter biomarker te ontwikkelen voor obesitas geĂŻnduceerde ziekte. Daarnaast zou manipulatie van de inhoud en de bindingscapaciteit van vetweefsel-membraanblaasjes aan andere cellen therapeutische potentie kunnen hebben.

A

183

APPENDIX

DANKWOORD Wow, het is gewoon af! In al mijn naïviteit en idealisme had ik 4 jaar geleden niet verwacht dat de jaren van mijn promotietraject zo intens en leerzaam zouden zijn, zowel op wetenschappelijk als persoonlijk gebied. Ik ben ongelofelijk blij met de mogelijkheden die hebben geleid tot het boekje wat nu voor u ligt. Graag wil ik iedereen bedanken die heeft bijgedragen aan het tot stand komen hiervan!

Allereerst wil ik graag de patiënten die hebben deelgenomen aan de klinische studies, op basis waarvan ik mijn onderzoeksresultaten heb gebaseerd, expliciet bedanken. Zonder u was het onderzoek beschreven in dit proefschrift niet mogelijk geweest. Bedankt!

Daarnaast wil ik graag een aantal groepen en personen noemen, die in de afgelopen 4 jaar een belangrijke rol hebben gespeeld in de totstandkoming van dit boek, de ontwikkeling van mij als wetenschapper en van mij als persoon.

Promotor Prof. Visseren, beste Frank. Ontzettend bedankt voor je begeleiding van de afgelopen jaren.

Je bent een erg betrokken promotor en was altijd bereikbaar voor overleg, waar ook ter wereld! Ook bezit je de gave om mensen te enthousiasmeren en om alles positief in te zien. Ik heb daar heel veel aan gehad als ik het even niet zag zitten! Met een flinke dosis oneliners en metaforen kan ik nu verder zelf op pad. Ik wil je danken voor het vertrouwen wat je in me had en de vrijheid die je me gaf, wetende dat je deur altijd voor me open stond. Dankjewel dat je van mij een zelfbewuste onderzoeker hebt gemaakt.

Copromotor Dr. Kalkhoven, beste Eric. Ik heb een kleine transformatie ondergaan sinds ik voor het eerst als student bij jou het lab binnen kwam. Van enigszins timide ben ik uitgegroeid tot zelfstandige onderzoeker, wat ik zeker aan jou te danken heb! Ik vond het heel erg fijn dat ik je altijd kon storen om te sparren, om de politiek achter de wetenschap te begrijpen of om tóch nog even iets te laten controleren. Je enorme enthousiasme voor vetweefsel, leuke samenwerkingen en hockey zijn aanstekelijk (die voor vogels nog altijd niet ).

Ontzettend bedankt voor je betrokken begeleiding.

Samenwerkingen Prof. Dr. Frans Moll, dank voor de prettige samenwerking en voor uw inspanningen voor de ADIPOSE II studie.

Dr. van Herwaarden, beste Joost, dank voor je inzet bij de ADIPOSE studie, en je enthousiasme over wat we allemaal voor informatie uit de vetjes konden halen! Het was leuk om met je te sparren over de resultaten, en ik wil je danken voor de prettige samenwerking.

184

APPENDIX

Dr. Vink, beste Aryan, ik vond het erg prettig met je samen te werken en de histologische wereld van vetweefsel te ontdekken. Dank voor je kritische begeleiding, en veel dank dat je mijn ogen hebt geopend voor de mooie wereld van de pathologie!

Dr. van Balkom, beste Bas. Jij bent de basis geweest van mijn EV-avontuur. Jij hebt me alles geleerd over het isoleren van EVs, en me gestimuleerd om telkens een stap verder te denken. Dankjewel!

Dr. de Jager, beste Wilco, luminex is this thesis’ middle name! Dank voor je inzet en expertise bij de luminex die bijna in elk hoofdstuk prijkt, je enorme enthousiasme en het meedenken bij de manuscripten.

Prof. Wauben, beste Marca en Dr. Nolte ‘t Hoen, beste Esther. Erg dankbaar ben ik voor de zeer welkome samenwerking en de specifieke kennis op het gebied van EVs die jullie me gebracht hebben. De mogelijkheid om de vesicles te kwantificeren op de high-resolution flow cytometer heeft het vet-EV manuscript een enorme lift gegeven. Het was erg prettig om met jullie samen te werken en gebruik te mogen maken van het walhalla aan flow cytometers op de afdeling celbiologie!

Prof. van der Graaf, beste Yolanda, dank voor je kritische en levendige commentaar op manuscripten en tijdens research besprekingen.

Dr. Uiterwaal, beste Cuno, dank voor je snelle en gedetailleerde beoordeling van mijn manuscripten, en het begeleiden van het Preventie project.

Prof. Pasterkamp, beste Gerard, dank voor je gastvrijheid op je lab waar ik het materiaal van mijn ADIPOSE studie mocht verwerken, en dank voor de prettige samenwerking.

De Vascu De vasculaire onderzoekers, dat is toch een vak apart. Verspreid over meerdere kamers, maar altijd verbonden via de vascu-app (en nog iets meer in huize Alla…).

Lichting 1: Daniel, Jan, Sandra, Joris, Remy, Ilse, Jannick, Melvin, Anton, Rob en Joep; het waren gezellige eerste jaren ondanks de vele (pogingen tot) pesterijtjes van toch vooral Jan en Daniel. Melvin, Jannick en Anton dank voor jullie tips en hulp bij alle rompslomp rond het boekje! Danny, het was leuk samenwerken, ik kon los op biologische vraagstukken terwijl jij me hielp met epi-vraagstukken. Dank voor al je hulp voor mijn data analyses.

Vissekommaatjes Remy en Ilse, we hebben het er goed gehad. Remy, als een stille rots in de branding, door niets gek te maken, altijd in voor advies en thee. Ilse, een kamergenootje door dik en dun. Dank voor al je hulp met mijn studie, hardlooptraining, opstellen van mooie zinnen en vooral ook met ons maandelijkse: “ik heb een sociaal probleem: wat zou jij doen?” Ik kijk er waardevol op terug! Sandra, wat is het jammer dat ik niet net een jaar eerder was begonnen. Hoe fijn was het geweest als we al die weken coupes scoren samen

185

A

APPENDIX

hadden doorgebracht! Ik bewonder je daadkracht en hoe je alles zelf hebt uitgevogeld.

Heel veel plezier en succes in de weide wereld!

Lichting 2: Maaike, Lotte 1, Lotte 2 (sorry, blijft er toch in), Manon, Bas en nieuwkomers Shahnam en Johanneke. Superleuk dat er zo’n jonge storm door de groep is gaan waaien!

Al was ik niet altijd meer even enthousiast als jullie in mijn ietwat cynische laatste uren, ik heb genoten van jullie energie, de borrels en gezellige etentjes. Heel veel succes nog de komende jaren bij de Vascu!

Nog een speciaal woordje voor mijn kamergenootjes van het laatste uur: Joep, Rob en Maaike. Ik vond het super om niet alleen een kamer met jullie te delen, maar ook onze verhitte discussies over wetenschap (hoe het écht moet), over lekker eten en over prangende levensvragen. Dank dat ik jullie al mijn epi-vragen mocht stellen, en vooral bedankt voor de cadeautjes . Maaikeeeee!! Hoe gezellig dat jij het Vascu team kwam versterken! Ik mis onze koffiemomenten, maar gelukkig heb je in Shahnam een waardige Micafe-opvolger gevonden. Ik ben blij een nieuw lichtpunt in je leven te hebben kunnen introduceren (limoncello-prosecco) en dank je voor je lessen in de Nederlandse en Engelse taal (an MD!) waar ik er toch nog wel meer van kan gebruiken.

Stafleden en fellows van de vascu: Wilko, Stan, Jan, Gideon, Hendrina en Gerben, dank voor jullie input tijdens researchbesprekingen en collegialiteit.

Inge, Corina en Corien, jullie enthousiasme voor patiënten en klinische studies was inspirerend.

Labs Het waardevolle van werken op een lab is de sfeer. Altijd hetzelfde: gestress, gevloek, radio 3FM, biertjes in de gang en altijd gezelligheid en gelach. Ik heb wat labs mee mogen maken tijdens mijn promotieonderzoek, en het heeft me vooral geleerd dat iedereen vooral zijn eigen recept toepast. Als het werkt dan werkt het en dan vooral niets veranderen! Dit heeft me gelukkig iets minder perfectionistisch gemaakt. Ik wil iedereen van al deze afdelingen enorm bedanken voor jullie aanwezigheid, voor de sfeer en voor het gebruik mogen maken van jullie spullen en kennis.

In het bijzonder wil ik noemen: Adipo-groep Henk, Arjen, Ismayil, Nicole, Yuan, Maryam, Annemarie en Inkie. Dank voor de superleuke tijd die ik als AIO in onze adipo-groep heb gehad! Ik heb genoten van de grappen die ik niet altijd begreep (Arjen, Henk), de aanstekelijke lach (Maryam), wijze gesprekken (Annemarie), indrukwekkende pumps en eppenrijtjes (Inkie), jullie BBQs, jullie hulp en expertise (allemaal). Nicole, dank voor je flexibiliteit, je hulp met het kweken van 100 x 150 mm schalen SGBS adipocyten, en je onvermoeibare hulp met regelen van promoties, kaartjes en cadeautjes voor onze groep!

Metabole ziekten Stan, Ellen, Rina, Rafaella, Willianne, Stiaan, Wouter V., Ingrid, Noortje, Danielle, Edwin, Samira, Sophia, Niels, Vittoria, Alexandra, Truus, Arjan, Jo, Joost, Monique A. en Monique de G. Dank voor de afgelopen jaren. Ik heb veel leuke momenten beleefd met Sinterklaas, labuitjes of gewoon in de kweek. Dank voor de inspirerende gesprekken,

186

APPENDIX

gezelligheid en prettige sfeer op het lab. Danielle, ik vond in jou een gezellige koffiepartner ook als minder leuke dingen besproken moesten worden, dankjewel! Alex, thank you for our nice collaboration on Akt phosphorylation in HepG2 cells! Boudewijn, Saskia en de ‘lever groep’, dank voor jullie input bij de werkbesprekingen.

Collega’s van de Burgering/Bos/Timmers/Vermeulen/Holstege groep wil ik bedanken voor hun behulpzaamheid.

Experimentele Nefro & Experimentele Cardio Ik heb nooit helemaal begrepen wie nou bij welke afdeling hoorde en welke voorraadkast dus bij welke groep, dus excuus dat ik het door elkaar benoem . Ik ben dankbaar voor jullie gastvrijheid, zowel voor het gebruik van spullen als voor de vrijdagmiddagborrels waar ik helaas niet heel vaak bij kon zijn.

Olivier, Hendrik, Krijn en Vera dank voor het gebruiken van jullie tips en lab equipment, en wat een verademing om als intimi te kunnen zeuren over de !@#$!& van sucrose gradient EV-isolatie. Sander en Marten, vooral bedankt voor de leuke in interessante discussies onder het genot van een Paulaner of bokbiertje! Loes, Sander (vdW) en Arjan dank voor protocollen, expertise en gezelligheid.

Overige EV-collegae Genoveva, Niek, Jeroen, Els, Marijke, Susan, ik leerde jullie vooral kennen op de ISEV in Götenborg, en de voor mij helaas wat laat maar oh zo nodige UMCbrede ‘EV-wat moeten we ermee’-meetings. Susan, Els(tar) en Marijke, dank voor jullie kennis over EVs, flow-data en heerlijke enthousiasme! Genoveva, lots of luck with adipose tissue EVs! Studenten Jelmer en Tereza het was leuk om met jullie samen te werken. Dank voor jullie inzet en veel succes in jullie wetenschappelijke toekomst!

Boston Even though my Boston period did not directly contribute to the contents of this thesis, my first steps in fundamental research at the Breault lab inspired me to pursue a PhD. The months I spent learning IHC, microscopy and fundamental biology in the inspiring atmosphere of Boston were one of the best experiences of my life. I would like to thank David, Diana, Robert, Dana, Michael and Loredana for including me in team mTert! I also want to thank team Tert and Maria, Masato, Haben, Lulu and Stacey for making me feel so welcome and letting me be inspired by you.

Familie Lieve papa, mama, Michiel en Daphne, Charlotte, en Irene en Xavier, onvoorwaardelijke liefde is een groot goed. Papa en mama dank voor jullie vertrouwen in mij, en dank voor de wijze levenslessen in combinatie met mij vrij laten in de keuzes die ik wil maken.

Het gevoel dat jullie onvoorwaardelijk achter me staan en zo trots op me zijn geeft mij ongelofelijk veel steun! Reen, hoe verschillend kunnen levenspaden van een tweeling verlopen! Ik bewonder je creativiteit en kracht, en als ik van iemand heb geleerd om door te zetten is het van jou. Lot, komend jaar heb ik eindelijk tijd voor een weekendje weg!

Michiel, heel gezellig dat jullie nu in Utrecht komen wonen!

187

A

APPENDIX

Lieve Oom Toon en tante Tip, dank voor jullie interesse in mijn studie en promotieonderzoek.

Oom Toon, onze gesprekken inspireren me enorm en het doet me veel goed dat je bij de hoogtepunten van mijn medische en wetenschappelijke carrière betrokken bent!

Lieve schoonfamilie, het is een feest om jullie te hebben leren kennen! Onze uitjes naar Twente zijn altijd gezellig, en zaterdagavonden met rode wijn en detectives op TV in huize Scheper zijn voor mij de rustmomenten van het jaar. Dank voor jullie betrokkenheid en gezelligheid, en superleuk dat jullie (helemaal vanuit Twente en zelfs Mallorca op en neer) allemaal komen om deze bijzondere dag met me te vieren!

Vrienden Lieve meisjes van Fame en van de Malie. Dank voor jullie gezelligheid en afleiding van de soms toch pittige promotiemomenten.

Arbaz en vrienden, als enige niet Arbaz-ingewijde ben ik trots wel tot jullie te behoren. Elsa, Stefanie, Maaike, Nicoline, Merel, Romy en Selinde dank voor alle jullie steun, raad, luisterend oor, kaartjes, bioscoop- en wijnbarbezoekjes! Romy en Selinde, zo lief hoe jullie altijd bewonderend roepen dat mijn kledingsstijl zo leuk is en ik zo hard kan werken, terwijl dat voor jullie zelf minstens even waar is. Ik vind jullie ook fantastisch!

Claar en Marique, heerlijk onze koffiemomenten in het UMCU. Dank voor jullie steun en tips. Supergezellig dat we samen in Utrecht werken, wat in de nabije toekomst hopelijk echt samenwerken wordt!

Paranimfen Ismayil, jij was mijn steun en toeverlaat de eerste 3 jaren van mijn promotie. Je verbeterde mijn labhumeur aanzienlijk met je grappen, je gelach en het altijd wel kunnen regelen van de spullen die ik nodig had! Maar ook sociale dingen konden we goed samen bespreken, waar ik veel steun aan had. Jouw interesse voor je medemens, reizen, andere culturen en google translate maken jou een bijzonder persoon. Dankjewel voor je vriendschap en voor het delen van je mantra’s!

Max, al meer dan 10 jaar ben jij mijn vriendin waar ik altijd bij terecht kan. Al zat je aan de andere kant van de oceaan, je was er voor mij. Ik met mijn Excel-bestanden en jij met je “go with the flow” maken samen een mooi team, en we groeien alleen maar meer naar elkaar toe. Ik ben verzot op je cashew-pindakaas en weet zeker dat we nog veel meer Van Max gaan zien! Ik ben echt supertrots op al je creativiteit en je moed om je hart te volgen.

Dank voor je optimisme, heerlijke eten en mooie ideeën.

Lieve paranimfen, dank dat jullie tijdens dit bijzondere moment aan mijn zijde willen staan!

188

APPENDIX

Wouter Ik kan niets bedenken waar ik je nĂ­et voor wil bedanken. Op alle niveaus ben je er voor mij geweest. Niet alleen met praktische hulp (dank dat ik 200 ml bloed bij je mocht tappen waaruit je mij vervolgens leerde om PBMCs te isoleren, dank voor het kritisch lezen van mijn hoofdstukken, dank voor het maken van de voorkant van mijn proefschrift!) Maar ook oneindig veel dank voor je steun, geduld en liefde. Weten dat jij bij mij hoort is het fijnste gevoel wat er is.

A

189

APPENDIX

CONTRIBUTING AUTHORS Bas W. van Balkom Department of Nephrology and Hypertension University Medical Center Utrecht, Utrecht, The Netherlands Arjan Brenkman Netherlands Metabolomics Center, the Netherlands.

Yolanda van der Graaf Julius Center for Health Sciences and Primary Care University Medical Center Utrecht, Utrecht, The Netherlands Joost A. van Herwaarden Department of Vascular Surgery University Medical Center Utrecht, Utrecht, The Netherlands Wilco de Jager Laboratory for Translational Immunology, Department of pediatric immunology University Medical Center Utrecht, Utrecht, The Netherlands Eric Kalkhoven Molecular Cancer Research, Center for Molecular Medicine, Section Metabolic Diseases University Medical Center Utrecht, Utrecht, The Netherlands Danny Kanhai Department of Vascular Medicine University Medical Center Utrecht, Utrecht, The Netherlands L. Jaap Kapelle Department of Neurology, Rudolf Magnus Institue of Neuroscience, Utrecht Stroke Center University Medical Center Utrecht, Utrecht, The Netherlands Dominique P.V. de Kleijn Experimental Cardiology Laboratory University Medical Center Utrecht, Utrecht, The Netherlands Cardiovascular Research Insitute & SUrgery National University Hospital, Singapore Frans L. Moll Department of Vascular Surgery University Medical Center Utrecht, Utrecht, The Netherlands

190

APPENDIX

Esther N.M. Nolte â&#x20AC;&#x2122;t Hoen Department of Biochemistry and Cell Biology Faculty Veterinary Medicine, Utrecht University, Utrecht, The Netherlands Gerard Pasterkamp Experimental Cardiology Laboratory University Medical Center Utrecht, Utrecht, The Netherlands Henk Schipper Department of Metabolic Diseases Department of pediatric immunology, Wilhelmina Childrenâ&#x20AC;&#x2122;s Hospital Utrecht University Medical Center Utrecht, Utrecht, The Netherlands Tereza Stupkova Department of Vascular Medicine University Medical Center Utrecht, Utrecht, The Netherlands Cuno Uiterwaal Julius Center for Health Sciences and Primary Care University Medical Center Utrecht, Utrecht, The Netherlands Marianne C. Verhaar Department of Nephrology and Hypertension University Medical Center Utrecht, Utrecht, The Netherlands Aryan Vink Department of Pathology University Medical Center Utrecht, Utrecht, The Netherlands

A

Frank L.J. Visseren Department of Vascular Medicine University Medical Center Utrecht, Utrecht, the Netherlands Marca H.M. Wauben Department of Biochemistry and Cell Biology Faculty Veterinary Medicine, Utrecht University, Utrecht, the Netherlands SMART Study Group of the University Medical Center Utrecht Prof. Dr. A. Algra, Julius Center for Health Sciences and Primary Care Prof. Dr. P.A. Doevendans, Department of Cardiology Prof. Dr. Y. van der Graaf, Julius Center for Health Sciences and Primary Care Prof. de. D.E.Grobbee, Julius Center for Health Sciences and Primary Care

191

APPENDIX

Prof. L.J.Kapelle, Department of Neurology Prof. dr. W.P.T.M. Mali, Department of Radiology Prof. dr. F.L.Moll, Department of Vascular Surgery Prof. dr. G.E.H.M. Rutten, Julius Center for Health Sciences and Primary Care Prof. dr. F.L.J.Visseren, Department of Vascular Medicine

192

APPENDIX

LIST OF PUBLICATIONS Mariëtte E. Kranendonk, Dominique P.V. de Kleijn, Eric Kalkhoven, Danny A. Kanhai, Cuno S.P.M. Uiterwaal, Yolanda van der Graaf, Gerard Pasterkamp, Frank L.J. Visseren; on behalf of the SMART Study Group. “The relation between adiposity and extracellular vesicle proteins in patients with manifest cardiovascular disease.” Cardiovascular Diabetology, 2014 Feb 5;13:37. doi: 10.1186/1475-2840-13-37.

Jan Westerink, MD, Gideon R Hajer, MD PhD, Mariëtte E. Kranendonk, MD, Henk S. Schipper, MD,Houshang Monajemi MD PhD, Eric Kalkhoven PhD, Frank L.J. Visseren, MD PhD. “An oral mixed fat load is followed by a modest anti-inflammatory adipocytokine response in overweight patients with metabolic syndrome.” Lipids, 2014 Mar;49(3):247-54. doi: 10.1007/s11745-014-3877-8. Epub 2014 Jan 21.

Mariëtte E. Kranendonk, Frank L. Visseren, Bas W. van Balkom, Esther N.M. Nolte-’t Hoen, Joost A. van Herwaarden, Henk S. Schipper, Marianne C. Verhaar, Marca H.M. Wauben, Eric Kalkhoven. “Human adipocyte extracellular vesicles in reciprocal signaling between adipocytes and macrophages”. Obesity, accepted manuscript online: 12 DEC 2013. DOI: 10.1002/oby.20679.

Danny A. Kanhai MD, MSc; Mariëtte E. Kranendonk MD; Cuno S.P.M. Uiterwaal MD, PhD; Yolanda van der Graaf MD, PhD; L. Jaap Kappelle MD, PhD; Frank L.J. Visseren MD, PhD. “Adiponectin and incident coronary heart disease and stroke. A systematic review and meta-analysis of prospective studies.” Obesity Reviews 2013 Jul;14(7):555-67.

Visser ME, Kropman E, Kranendonk ME, Koppen A, Hamers N, Stroes ES, Kalkhoven E, Monajemi H. “Characterization of non-obese diabetic patients with marked insulin resistance identifies a novel familial partial lipodystrophy-associated PPARγ mutation (Y151C).” Diabetologia. 2011 Jul;54(7):1639-44. Epub 2011 Apr 9.

Montgomery RK, Carlone DL, Richmond CA, Farilla L, Kranendonk ME, Henderson DE, Baffour-Awuah NY, Ambruzs DM, Fogli LK, Algra S, Breault DT. “Mouse telomerase reverse transcriptase (mTert) expression marks slowly cycling intestinal stem cells.” Proc Natl Acad Sci U S A. 2011 Jan 4;108(1):179-84. Epub 2010 Dec 20.

Heidema J, Rossen JW, Lukens MV, Ketel MS, Scheltens E, Kranendonk ME, van Maren WW, van Loon AM, Otten HG, Kimpen JL, van Bleek GM. “Dynamics of human respiratory virus-specific CD8+ T cell responses in blood and airways during episodes of common cold.” J Immunol. 2008 Oct 15;181(8):5551-9.

193

A

APPENDIX

IN PREPARATION Mariëtte E. Kranendonk, Joost van Herwaarden, Tereza Stupkova, Wilco de Jager, Aryan Vink, Frans L. Moll, Eric Kalkhoven, Frank L.J. Visseren. “Inflammatory characteristics of distinct abdominal adipose tissue depots relate differently to metabolic risk factors for cardiovascular disease.”

In submission Mariëtte E. Kranendonk, Frank L.J. Visseren, Joost A. van Herwaarden, Esther N.M. Nolte-’t Hoen, Wilco de Jager, Marca H.M. Wauben, Eric Kalkhoven. “Effect of extracellular vesicles of human adipose tissue on insulin signaling in hepatocytes and muscle cells.” In revision, Obesity Arjen Koppen, David Cassiman, Marjoleine Broekema, Mariëtte E. Kranendonk, Marian Groot Koerkamp, Nicole Hamers, Frank Holstege, Houshang Monajemi and Eric Kalkhoven.

"The novel FPLD3 PPARγE379K mutation results in gene selective reduction in transcription and DNA binding." Manuscript in preparation.

194

APPENDIX

CURRICULUM VITAE Mariëtte Kranendonk was born the 31st of May in 1983 in The Hague, the Netherlands.

In 2002 she completed grammar school at “Aloysius college” in The Hague. Subsequently, she studied medicine at the University of Utrecht. She completed various rotations abroad: at ‘s Lands Hospital at Paramaribo, Surinam and St Luke’s Hospital in Malta. She conducted her first research project at Harvard Medical School, at the department of Endocrinology of Boston Children’s Hospital, which strongly encouraged her interest in fundamental metabolic research. Her graduation research project was performed at the laboratory of metabolic diseases at the UMCU under supervision of dr. E. Kalkhoven. In February 2010 she started her PhD studies at the departments of Vascular Medicine and Metabolic Diseases, under supervision of prof. F.L.J.Visseren and Dr. E.Kalkhoven. Results are presented in this thesis, entitled “Adipose tissue dysfunction and cardiometabolic risk.

In vitro, ex vivo and clinical studies”. In 2014, she started working as postdoc at the department of Pathology of the University of Utrecht, and from January 2015 she will start her residency in Pathology at the University Medical Centre Utrecht.

A

195