Citation for published version (APA): van den Berg, S. M. (2017). A hot interaction between immune cells and adipose tissue.

Transcription

1 UvA-DARE (Digital Academic Repository) A hot interaction between immune cells and adipose tissue van den Berg, S.M. Link to publication Citation for published version (APA): van den Berg, S. M. (217). A hot interaction between immune cells and adipose tissue. General rights It is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), other than for strictly personal, individual use, unless the work is under an open content license (like Creative Commons). Disclaimer/Complaints regulations If you believe that digital publication of certain material infringes any of your rights or (privacy) interests, please let the Library know, stating your reasons. In case of a legitimate complaint, the Library will make the material inaccessible and/or remove it from the website. Please Ask the Library: or a letter to: Library of the University of Amsterdam, Secretariat, Singel 425, 112 WP Amsterdam, The Netherlands. You will be contacted as soon as possible. UvA-DARE is a service provided by the library of the University of Amsterdam ( Download date: 28 Mar 219

2

3 A hot interaction between immune cells and adipose tissue Susanna Maria van den Berg

4 A hot interaction between immune cells and adipose tissue PhD thesis, University of Amsterdam, the Netherlands ISBN: Cover design: Emily van t Wout and Susan van den Berg Lay out: Susan van den Berg Printing: Ridderprint BV Copyright 217 S.M. van den Berg

5 A hot interaction between immune cells and adipose tissue ACADEMISCH PROEFSCHRIFT ter verkrijging van de graad van doctor aan de Universiteit van Amsterdam op gezag van de Rector Magnificus prof. dr. ir. K.I.J. Maex ten overstaan van een door het College voor Promoties ingestelde commissie, in het openbaar te verdedigen in de Agnietenkapel op vrijdag 2 juni 217, te 14: uur door Susanna Maria van den Berg geboren te Leiden

6 Promotiecommissie: Promotores: Prof. Dr. E. Lutgens AMC Universiteit van Amsterdam Prof. Dr. M.P.J. de Winther AMC Universiteit van Amsterdam Copromotor: Prof. Dr. P.C.N. Rensen Universiteit Leiden Overige leden: Prof. Dr. N. Zelcer AMC Universiteit van Amsterdam Prof. Dr. M. Nieuwdorp AMC Universiteit van Amsterdam Prof. Dr. R.P.J. Oude Elferink AMC Universiteit van Amsterdam Prof. Dr. R. Shiri Sverdlov Universiteit van Maastricht Dr. B.G.A. Guigas Universiteit Leiden Dr. R.H.L. Houtkooper AMC Universiteit van Amsterdam Faculteit der Geneeskunde Financial support by the Dutch Heart Foundation for the publication of this thesis is gratefully acknowledged. Further financial support for printing this thesis was kindly provided by Special Diets Services.

7

8

9 Table of contents Chapter 1 General introduction 9 Chapter 2 Immune modulation of brown(ing) adipose tissue in obesity. 15 Chapter 3 Chapter 4 Chapter 5 Diet induced obesity induces rapid inflammatory changes in brown adipose tissue in mice Type 2 inflammatory response by helminth derived antigens induces beiging of white adipose tissue in mice.. 65 Diet induced obesity in mice diminishes hematopoietic stem and progenitor cells in the bone marrow.. 81 Chapter 6 Blocking CD4 TRAF6 interactions by small molecule inhibitor ameliorates the complications of diet induced obesity in mice 11 Chapter 7 General discussion. 121 Appendix I. Summary. 133 II. Samenvatting III. PhD Portfolio Dankwoord. 147

10

11 1 Battling obesity General introduction

12 Chapter 1 1

13 General introduction Battling obesity Cookies, candy and chocolate are found at every counter, hamburgers and cola are the cheapest items on the menu and exercising takes too much time and effort; it is so easy to become fat. It has come to a point where obesity rates worldwide are ridiculously high; 39% of the world s adult population is overweight, of which 13% is obese [World Health Organization]. We love to eat, especially when it is cheap, greasy, sweet and fast. Willingness to eat is of course essential to survive but this is becoming a problem when food is everywhere. Why is being overweight or obese a bad thing? Obesity is linked with high risks of type 2 diabetes (T2D), fatty liver disease, cardiovascular disease (CVD), and cancer [1]. Chronic obesity may shorten healthy lifespan by 5 2 years, which results in a tremendous socio economic burden [2]. When energy intake exceeds energy expenditure, white adipose tissue stores excessive lipids. With the accumulation of lipids, adipocytes enlarge and the tissue becomes hypoxic. Dysfunctional adipocytes secrete adipokines, cytokines and chemokines that recruit inflammatory cells, which results in a state of chronic low grade inflammation [3]. In the current era of scientific research, we are stepping in and manipulate our survival mechanisms by finding strategies that limit the damage of our unhealthy eating life style. This thesis describes three approaches that contribute to the battle against obesity associated diseases. First, we hypothesize that obesity induces immunological changes in brown adipose tissue (BAT), which affects brown adipocyte activity. Our second hypothesis is that the chronic low grade inflammatory state in obesity and the continuous recruitment of immune cells affects hematopoietic stem cells in the bone marrow. Lastly, we hypothesize that inhibiting the interaction between the co stimulatory molecule CD4 and its adaptor protein TRAF6 using a small molecule inhibitor will improve white adipose tissue inflammation and the associated metabolic dysfunction in a model of diet induced obesity. Brown adipose tissue burns fat One intriguing target to reduce excessive energy storage in obesity is BAT. BAT is involved in adaptive thermogenesis. Cold exposure activates the production of heat by brown adipocytes using glucose and fat as fuel [4]. Chapter 2 of this thesis elaborates on how BAT is regulated, describes that brown adipocytes can occur within white adipose tissue, describes how different components of the immune system are altered in obese adipose tissue, and extensively discusses recent data on how immune cells contribute to the regulation of brown and beige adipocyte activity. The underlying key component of adverse effects in obesity is inflammation. Although this has been extensively studied in white adipose tissue, effects of obesity on the inflammatory status in BAT are still largely unknown. In chapter 3 we explored what inflammatory changes occur in BAT in the course of obesity. In chapter 4 we apply a model of helminth antigens, which induces a type 2 immune response, to study whether skewing macrophages to a Th2 and M2 anti inflammatory phenotype affects brown(ing) adipose tissue in a setting of high fat diet. 11

14 Chapter 1 Does diet induced chronic inflammation affect hematopoietic stem cells? Hematopoietic stem cells and progenitor cells (HSPCs) are the most primitive precursors of immune cells and are mostly present in the bone marrow. Their most important feature is that they have a self renewal capacity. HSPCs generally remain quiescent but can generate an appropriate immune response when needed [5]. Homeostasis of the system is disturbed after chronic immune stressors, creating a dysbalance in different immune cell lineages. Activation of HSPCs leads to proliferation, differentiation and mobilization which upon prolonged stimuli can even lead to exhaustion of the stem cell pool [6]. Dysregulation of the earliest hematopoietic stem cells can have major effects on later lineages. Obesity is characterized by an ongoing inflammatory response [3]. Due to the chronic nature of obesity, immune cells are continuously recruited from the bone marrow. In chapter 5 we study whether the immune system suffers from chronic activation by looking at HSPCs in the bone marrow after different durations of HFD. Targeting the co stimulatory molecule CD4 to improve metabolism in obesity Co stimulatory molecules have a central role in inflammation in which they help with the activation of immune cells. Antigen presenting cells (APCs) activate T cells by presenting an antigen to the T cell receptor, secondly, the interaction between co stimulatory molecules then provides an additional stimuli to activate the APC and the T cell [7]. Co stimulatory molecules can also directly activate other cell types including monocytes and granulocytes but also adipocytes or endothelial cells. An important co stimulatory dyad is CD4 CD4L, which is involved in the activation of APCs and T cells, the induction of cytokine production, but also B cell isotype switching as well as the activation of endothelial cells and the migration of monocytes [8, 9]. CD4 is involved in mediating a wide variety of immune responses. However, it cannot signal on its own and upon binding of CD4L, it recruits adaptor proteins, TNFR associated factors (TRAFs) for signal transduction. CD4 has different binding sites for TRAF 2/3/5 and for TRAF6. This allows CD4 to activate different signalling pathways depending on which adaptor protein binds, which cell type is involved and other local conditions [9]. The CD4 CD4L dyad has an important role in obesity. Interactions between CD4 CD4L on adipocytes and immune cells promote adipose tissue inflammation [1, 11]. In diet induced obesity, CD4 / mice have increased insulin resistance, more adipose tissue inflammation and enhanced hepatosteatosis compared to wild type mice. CD4 TRAF2/3/5 deficient mice exhibit a similar phenotype, however, deficient CD4 TRAF6 signalling does not result in insulin resistance but reduces adipose tissue inflammation and hepatosteatosis in diet induced obesity. In chapter 6 we approach the diet induced inflammatory response in adipose tissue by targeting CD4 TRAF6 interactions using a small molecule inhibitor, previously designed in our lab. 12

15 General introduction References [1] Berrington de Gonzalez A, Hartge P, Cerhan JR, Flint AJ, Hannan L, MacInnis RJ et al. 21. Body Mass Index and Mortality among 1.46 Million White Adults. New England Journal of Medicine 363(23): [2] Kanneganti T D, Dixit VD Immunological complications of obesity. Nat Immunol 13(8): [3] Lumeng CN, Saltiel AR Inflammatory links between obesity and metabolic disease. Journal of Clinical Investigation 121(6): [4] Kajimura S, Spiegelman Bruce M, Seale P Brown and Beige Fat: Physiological Roles beyond Heat Generation. Cell Metabolism 22(4): [5] Mendelson A, Frenette PS Hematopoietic stem cell niche maintenance during homeostasis and regeneration. Nat Med 2(8): [6] King KY, Goodell MA Inflammatory modulation of HSCs: viewing the HSC as a foundation for the immune response. Nat Rev Immunol 11(1): [7] Zirlik A, Lutgens E An inflammatory link in atherosclerosis and obesity. Co stimulatory molecules. Hämostaseologie 35(3): [8] Seijkens T, Kusters P, Chatzigeorgiou A, Chavakis T, Lutgens E Immune Cell Crosstalk in Obesity: A Key Role for Costimulation? Diabetes 63(12):3982. [9] Engel D, Seijkens T, Poggi M, Sanati M, Thevissen L, Beckers L et al. 29. The immunobiology of CD154 CD4 TRAF interactions in atherosclerosis. Semin Immunol 21(5): [1] Poggi M, Jager J, Paulmyer Lacroix O, Peiretti F, Gremeaux T, Verdier M et al. 29. The inflammatory receptor CD4 is expressed on human adipocytes: contribution to crosstalk between lymphocytes and adipocytes. Diabetologia 52(6): [11] Poggi M, Engel D, Christ A, Beckers L, Wijnands E, Boon L et al CD4L Deficiency Ameliorates Adipose Tissue Inflammation and Metabolic Manifestations of Obesity in Mice. Arterioscler Thromb Vasc Biol 31(1):

16 Chapter 1 14

17 2 Immune modulation of brown(ing) adipose tissue in obesity Susan M. van den Berg 1, Andrea D. van Dam 2,3, Patrick C.N. Rensen 2,3, Menno P.J. de Winther 1,4,*, Esther Lutgens 1,4,* 1 Department of Medical Biochemistry, Subdivision of Experimental Vascular Biology, Academic Medical Centre, University of Amsterdam, The Netherlands. 2 Department of Medicine, Division of Endocrinology, Leiden University Medical Center, Leiden, The Netherlands. 3 Einthoven Laboratory for Experimental Vascular Medicine, Leiden University Medical Center, Leiden, The Netherlands. 4 Institute for Cardiovascular Prevention (IPEK), Ludwig Maximilians University of Munich, Munich, Germany. *these authors contributed equally to this work. Endocrine Reviews (1): 46 68

18 Chapter 2 Abstract Obesity is associated with a variety of medical conditions such as type 2 diabetes and cardiovascular diseases and is therefore responsible for high morbidity and mortality rates. Increasing energy expenditure by brown adipose tissue (BAT) is a current novel strategy to reduce the excessive energy stores in obesity. Brown adipocytes burn energy to generate heat and are mainly activated upon cold exposure. As prolonged cold exposure is not a realistic therapy, researchers worldwide are searching for novel ways to activate BAT and/or induce beiging of WAT. Recently the contribution of immune cells in the regulation of brown adipocyte activity and beiging of WAT has gained increased attention, with a prominent role for eosinophils and alternatively activated macrophages. This review will discuss the re discovery of BAT, present an overview of modes of activation and differentiation of beige and brown adipocytes and describe the recently discovered immunological pathways that are key in mediating brown/beige adipocyte development and function. Interventions in immunological pathways harbour the potential to provide novel strategies to increase beige and brown adipose tissue activity as therapeutic target for obesity. 16

19 Immune regulation in adipose tissue Introduction According to the latest statistics of the world health organization, the worldwide prevalence of obesity, defined as a BMI>3, has nearly doubled since 198 and at least 2.8 million people die each year as a result of obesity. This number is expected to further increase over the next decade. Obesity leads to adverse effects on blood pressure, plasma lipid levels and insulin resistance. Health problems related to these adverse effects include coronary artery disease, atherosclerosis, fatty liver disease, type 2 diabetes, cancer and degenerative diseases [1]. The immune system plays a key role in the pathogenesis of obesity, and extensive research is ongoing to identify critical players involved to find novel therapeutic targets to combat obesity. Currently, adequate treatment options are limited. Interestingly, recent studies show that the activation of beige or brown adipocytes and the subsequent increase in energy expenditure can lower body fat mass and potentially lower adipose tissue inflammation [2]. Moreover, there is increasing evidence that brown and beige adipose tissue function is subject to immune regulation [3 6]. Brown and beige adipose tissue are therefore of great interest as a novel therapeutic target for obesity. Brown adipose tissue BAT regulates adaptive thermogenesis by mitochondrial uncoupling Whereas white adipose tissue (WAT) is specialized in the storage of energy, brown adipose tissue (BAT) plays a central role in energy expenditure. Brown adipocytes convert energy from glucose and fatty acids into heat via non shivering thermogenesis, which contributes to the maintenance of body temperature [7]. The regulation of body temperature is crucial to ensure that cellular functions and physiological processes continue in cold environments [8]. This regulation is particularly important in small organisms with a relatively large surface area. Even new borns of large organisms have distinct depots of BAT that regress with increasing age. BAT in adult humans is most commonly present in the supraclavicular and neck region, but also along the vertebrae and aorta and near the kidneys (Figure 1) [9]. In rodents, the major BAT depot is found in the interscapular region, whereas smaller depots include axillary BAT, cervical BAT and perirenal and periaortic BAT (Figure 1) [1, 11]. BAT is highly innervated by the sympathetic nervous system and a well structured vascularization enables the supply of oxygen and transport of heat. Brown adipocytes have numerous small lipid droplets for fast energy supply and a large number of mitochondria that can produce heat via mitochondrial uncoupling. The wealth of mitochondria and the extensive vascularization pattern account for their dark colour and hence its name: brown adipose tissue. Cold activates the hypothalamus, which induces sympathetic outflow towards BAT and results in release of noradrenaline by efferent sympathetic nerve endings. Noradrenaline binds to β adrenergic receptors present on brown, but also on white adipocytes [12, 13]. Cold induced adrenergic receptor stimulation has both acute and chronic effects on BAT [14]. Acute thermogenesis results in lipolysis, degradation of fatty acids, glucose uptake and activation of uncoupling protein 1 (UCP1). Chronic activation leads to increased gene transcription of UCP1 and mitochondrial biogenesis [14]. BAT expresses substantial amounts of UCP1, an inner membrane mitochondrial protein that uncouples oxidative phosphorylation from ATP synthesis, resulting in dissipation of energy into heat (Figure 2) [15]. 17

20 Chapter 2 Interscapular Neck Supraclavicular Periaortic Paravertebral Perirenal Cervical Axillary Interscapular Periaortic Perirenal Figure 1. BAT locations Human new borns have a distinct interscapular BAT depot which regresses with age. Human adults have supraclavicular BAT and BAT depots in the neck region. Smaller depots are found along the aorta, vertebrae and kidneys. In mice, cervical BAT is located underneath the muscles running from the back of the head to the interscapular area. Ventrolateral of the scapulae is an axillary BAT depot. The major BAT depot is found interscapular, but also around the aorta and the hilum of the kidneys. β adrenergic receptors signal via camp, leading to activation of protein kinase A (PKA) and p38 mitogen activated protein kinase (MAPK) [16, 17]. Eventually, this results in phosphorylation of relevant transcription factors, including camp response element binding protein (CREB), which control the expression of genes involved in BAT activation, such as peroxisome proliferator activated receptor γ co activator 1α (PGC 1α). PGC 1α is a transcriptional cofactor that increases mitochondrial biogenesis and induces expression of UCP1 [15, 18, 19]. White, beige and brown adipocytes Currently, two distinct types of brown adipocytes, each of a different origin, have been described. The classical brown adipocyte, which is found at distinct anatomical sites in mice, including the interscapular, perirenal and axillary BAT depots, and the so called brite or beige adipocytes which are found within WAT [2]. Inducing these beige adipocytes in WAT is referred to as browning or beiging. Because humans lack a large classical BAT depot, it is attractive to identify approaches to stimulate formation of beige cells within WAT that share functional characteristics with classical brown adipocytes. An important difference between the two cell types is that classical brown adipocytes constitutively express UCP1, while beige adipocytes only do so upon appropriate stimuli, such as cold and β adrenergic receptor stimulation [21 23]. Increased biogenesis of beige adipocytes may contribute to increased energy expenditure, improved metabolic parameters and improved tolerance to cold. 18

21 Immune regulation in adipose tissue Figure 2. Activation of brown adipocytes by cold. Cold activates the hypothalamus in the brain, which activates the SNS and results in the release of noradrenaline that binds to β adrenergic receptors on brown adipocytes. The acute result is intracellular lipolysis, degradation of fatty acids via beta oxidation and activation of UCP1. Other effects of activated β adrenergic receptors occur via camp, PKA, CREB and p38 mitogen activated protein kinase. Oxidative phosphorylation in mitochondria drives protons from the mitochondrial matrix into the intermembrane space, generating an electrochemical gradient by which protons flow back into the mitochondrial matrix via ATP synthase, activating ATP synthesis in e.g. heart and skeletal muscle. In brown adipocytes, UCP1 is found on the inner mitochondrial membrane where it causes mitochondrial uncoupling. UCP1 increases the permeability of the inner mitochondrial membrane and thus causes a reflux of protons into the mitochondrial matrix, bypassing ATP synthase. This proton leakage leads to dissipation of energy into heat. ADP; adenosine diphosphate, ATP; adenosine triphosphate, camp; cyclic adenosine monophosphate, CREB; camp response element binding protein, FFA; free fatty acids, P38; p38 mitogen activated protein kinase, PKA; protein kinase A, SNS; sympathetic nervous system, UCP1; uncoupling protein 1. 19

22 Chapter 2 Microarray analysis as well as in vivo lineage tracing studies showed a common developmental ancestry between the classical brown adipocyte and skeletal muscle cells, in which the common progenitor is myogenic factor 5 (Myf5) expressing muscle precursor cell [2, 24]. Whereas myogenin stimulates myocyte development, transcriptional regulator PR domain zinc finger protein 16 (Prdm16) regulates the developmental switch to brown adipocytes. PRDM16 forms a transcriptional complex with activated transcription factor CCAAT/enhancer binding protein β (C/EBPβ), which enables the switch from a myogenic precursor to a brown adipocyte [25]. This in turn leads to the expression of Peroxisome proliferator activated receptor γ (Pparγ) and Pgc 1α, which are key regulators of brown adipocyte differentiation, and BAT associated genes such as Ucp1, Cidea, Cox7α1, Cox8β, Elovl3 and Cpt1b. PPARγ itself can stimulate brown adipocyte differentiation as well. Indeed, continuous treatment with a PPARγ agonist, rosiglitazone, enhances brown adipocyte differentiation in vitro [26]. Although the Myf5 expressing precursors were assumed to be precursors of brown adipocytes and muscle cells only, recent lineage studies contradict this. It was found that Myf5 + precursor cells can also give rise to unilocular white adipocytes in subcutaneous and retroperitoneal WAT [21, 27 29]. In classic myogenic transcription, Myf5 is thought to function downstream or simultaneously with paired box 3 (Pax3) and upstream of myogenic differentiation 1 (MyoD1). Indeed, expression of Myf5 and Pax3, not MyoD1, overlaps in adipocyte progenitors. This holds true for most adipose tissue depots, but not in perigonadal WAT, indicating depot and gender differences [28]. Although all brown adipocytes in interscapular BAT descend from Myf5 + precursors, only a subset of classical brown adipocytes in cervical and no brown adipocytes in perirenal or periaortic BAT could be traced back to a Myf5 + precursor, suggesting a large heterogeneity within and between adipose tissue depots [28]. Other markers for committed brown adipocyte precursors that differentiate into mature brown adipocytes are early B cell factor 2 (EBF2) and platelet derived growth factor receptor α (PDGFRα) (Figure 3) [3]. Beige cells are present within WAT and, when present in sufficient quantity, significantly increase energy expenditure and thereby are thought to contribute to a reduction in body fat [31 34]. In mice, cold exposure and β adrenergic receptor agonist treatment increases Ucp1 gene expression and mitochondrial biogenesis in WAT, which increases the presence of beige cells within WAT [35]. Beige cells are mostly present in subcutaneous WAT. Adipocyte specific PRDM16 transgenic mice are able to induce the development of beige cells in subcutaneous WAT and have an increased energy expenditure, are resistant to weight gain and show improved glucose tolerance on a HFD [36]. Other stimuli that can induce beige adipogenesis include noradrenaline [5, 37], lactate [38], irisin [39], fibroblast growth factor (FGF) 21 [4], BMP (bone morphogenetic protein) 4 [41] and BMP7 [42]. Beige adipocytes arise from white (i.e. unilocular and Ucp1 negative) adipocytes, via transdifferentiation. The transdifferentiation hypothesis is based on experiments in rodents in which beige adipogenesis was induced and examined by microscopy [33, 43, 44], analysis of DNA content and labelling with bromodeoxyuridine [32, 45, 46] and UCP1 Cre reporter mice, which allows inducible permanent labelling of Ucp1 + cells and tracing of white, beige and brown adipocytes [47]. However, other studies show that beige adipocytes can arise de novo from specific precursor cells. Using AdipoChaser mice, which have an inducible adipocyte tagging system with an adiponectin promoter driven tetracycline on (Tet) transcription factor, a Tet responsive Cre (activated by doxycycline) and a Rosa26 promoter driven loxp stop loxp β galactosidase. Upon doxycycline treatment, all adipocytes become positive for β galactosidase (β gal). Upon induction of beige 2

23 Immune regulation in adipose tissue EBF2 + PDGFRα + Preadipocyte Myf5 + Pax3 + Precursor Myogenin Myf5 + Pax3 + Precursor Myf5 - Pax3 - Precursor De novo adipogenesis EBF2 + PDGFRα + Preadipocyte Myocyte White adipocyte Transdifferentiation PRDM16 PPARγ Brown adipocyte Dedifferentiation Whitening Beige adipocyte Inguinal subcutaneous white adipose tissue Interscapular brown adipose tissue Figure 3. Beige and brown adipogenesis. Model showing beige and brown adipocyte development in inguinal subcutaneous adipose tissue and interscapular brown adipose tissue. White adipocytes can derive from both Myf5 + Pax3 + as well as Myf5 Pax3 precursors. Beige adipocytes can either transdifferentiate from mature white adipocytes or directly differentiate from EBF2 + PDGFRα + preadipocytes, called de novo adipogenesis. EBF2 is a selective marker for brown and beige preadipocytes.in interscapular BAT, brown adipocytes are derived from a multipotent Myf5 + Pax3 + expressing precursor population. When these precursors are exposed to myogenin they will develop into myocytes. PRDM16 and PPARy promote brown adipocyte differentiation. Brown adipocytes in BAT can undergo whitening upon exposure to thermoneutrality, obesity, ageing or sympathetic denervation. EBF2; early B cell factor 2, Myf5; myogenic factor 5, MyoD1; myogenic differentiation 1, Pax3; paired box 3, PDGFRα; platelet derived growth factor receptor α, PPARγ; peroxisome proliferator activated receptor γ, PRDM16; transcriptional regulator PR domain zinc finger protein

24 Chapter 2 adipogenesis by cold exposure or β 3 adrenergic receptor agonist treatment, β gal negative beige adipocytes appear in subcutaneous WAT [48, 49]. As in interscapular BAT, Ebf2 expression marks beige adipocytes, regardless of their developmental origin [3] (Figure 3). These discrepant insights might result from technical restraints, or the fact that transdifferentiation and de novo adipogenesis occur simultaneously, maybe depending on whether adipocytes have experienced beige stimuli before [47]. Formerly beige, dedifferentiated white adipocytes might be capable of transdifferentiation into beige adipocytes whereas a first encounter with beige stimuli induces de novo development. Alternatively, pre encoded beige adipocytes may exist disguised as white adipocytes and only develop their beige phenotype upon the adequate stimuli. The vasculature may be another important mediator of beiging, as BAT is highly vascularized to enable a fast supply of oxygen and nutrients and transport of generated heat [5]. In WAT, exercise, as well as treatment with a β 3 adrenergic receptor agonist or a PPARγ ligand such as rosiglitazone, not only induces beiging but also increases angiogenesis. Overexpression of Vegf increases WAT vascularization and induces increased gene expression of Ucp1 and Pgc 1a in retroperitoneal WAT [51]. The importance of VEGF as an essential downstream target in the process of beiging was further confirmed by administration of an anti VEGF antibody, which reduced both angiogenesis and beiging in WAT [51]. Contrary to beiging of WAT, whitening of BAT also occurs. Sympathetic denervation, thermoneutrality, ageing and excessive energy supplies in obesity causes BAT to accumulate lipid droplets, resulting in whitening of the tissue [52, 53]. This is accompanied by decreased vascularization, hypoxia and mitochondrial dysfunction. A whitened phenotype of BAT in obesity is therefore predictive of decreased BAT activity and is associated with complications such as insulin resistance [5]. Although the mechanisms of BAT whitening are largely unknown, VEGF mediated vascularization seems to play an important role; obesity reduces Vegfa expression and deletion of Vegfa in adipose tissue results in whitened BAT [5]. 22

25 Immune regulation in adipose tissue BAT activity and obesity: what we have learned from mouse models Much of our current knowledge on the role of BAT in obesity is derived from experimental animal studies, often including genetically modified mice that were subjected to different models of obesity. In these studies, decreased BAT activity was associated with aggravated metabolic dysfunction and an increase in obesity, whereas increased BAT activity was associated with improved metabolism. Over the years, several experimental animal studies have been published in which the relation between obesity and BAT was studied (Table 1). BAT activity is dependent on environmental temperature and therefore, housing temperature has major metabolic consequences. Mice housed at room temperature (18 22 C) are thermally stressed, leading to increased metabolism and activated BAT to maintain their body temperature. Elimination of thermal stress is accomplished by thermoneutral housing conditions of 3 C. It is important to realize that in some experimental setups, it is thus appropriate to house mice under thermoneutral conditions, for example when comparing studies to thermoneutral conditions in humans (i.e. 25 C) or to distinguish central mediated effects from direct activation of BAT. Genetic models of decreased BAT activity Energy expenditure and heat production in BAT occurs via UCP1 and mice lacking UCP1 are indeed unable to maintain their body temperature when exposed to acute cold. However, at room temperature, the decreased ability to dissipate energy into heat does not result in increased obesity neither on chow nor upon high fat diet (HFD) feeding [54], however, UCP1 knockout mice are more susceptible to develop obesity with age [55]. Strikingly, when UCP1 knockout mice are housed under thermoneutral conditions (3 C), they do develop more severe obesity on a HFD and also gain slightly more weight when on chow diet [56]. Other studies in mice have revealed that besides UCP1 deficiency, additional models with decreased BAT activity also suffer from cold intolerance and display an obesity sensitive phenotype, as listed in Table 1. Genetic ablation of BAT function via expression of a diphtheria toxin A chain driven by regulatory elements of the Ucp1 gene (UCP DTA) decreased the UCP1 content in BAT, which resulted in a reduction in non shivering thermogenesis and obesity. Consequently, these mice did not show a thermogenic response upon treatment with a β 3 adrenergic receptor agonist [57]. β adrenergic receptor knockout mice lacking all three known β adrenergic receptors, as well as dopamine β hydroxylase deficient mice, which fail to convert dopamine into noradrenaline, both present inactive BAT and are sensitive to cold [58, 59]. Furthermore, β adrenergic receptor knockout mice are slightly obese on a low fat diet and develop massive obesity on a HFD [59]. Leptin deficient ob/ob mice also have reduced Ucp1 expression in BAT and, similar to UCP1 knockout mice, are cold sensitive, but will survive when they are gradually exposed to extreme cold (4 C). However, mice lacking both UCP1 and leptin do not survive temperatures below 12 C. These models show that alternative mechanisms for maintaining body temperatures do exist, but that these are only effective within certain limits [6]. Increasing BAT activity improves metabolism in mice The disruption of BAT activation results in an obesity prone phenotype in mice. Conversely, increasing the amount and/or function of BAT can induce a healthy metabolic phenotype. Cold exposure and β 3 adrenergic receptor agonist treatment increase Ucp1 expression, mitochondrial 23

26 Chapter 2 Table 1. Experimental animal studies showing the metabolic effects of increasing or decreasing BAT activity. Model Effect Reference Decreased BAT activity UCP1 / UCP DTA Room temperature (18 22 C) Room temperature (18 22 C) Thermoneutral housing (3 C) β adrenergic receptor / Dopamine β hydroxylase / ob/ob UCP1 / ob/ob FABP4/5 / Adipocyte specific p62 Increased BAT activity Cold exposure β 3 adrenergic receptor agonist treatment Chow diet HFD Chow diet Chow diet HFD Cold sensitive Enerbäck 1997 [54] Increased susceptibility to obesity with age Kontani 25 [55] Obese Feldmann 29 [56] Reduction in NST Obese phenotype Slightly obese Severe obesity Inactive BAT Cold sensitive Slightly obese Obese Cold sensitive Obese Extremely cold sensitive Cold sensitive Decreased BAT activity Obese Increases UCP1 expression Increases thermogenic capacity Accelerates atherosclerosis Increases UCP1 expression Increases thermogenic capacity Reduces body weight Reduces atherosclerosis Lowell 1993 [57] Bachman 22 [59] Thomas 1997 [58] Ukropec 26 [6] Ukropec 26 [6] Syamsunarno 214 [61] Muller 213 [62] Ravussin 214 [63] Bartelt 211 [64] Dong 213 [65] Himms Hagen 1994 [66] Barbatelli 21 [43] Berbée 215 [35] Constitutive UCP1 expression in adipocytes Prevents obesity Kopecky 1995 [67] Increased expression of UCP1 by FOXC2 overexpression on HFD Less adiposity Improved glucose tolerance Cederberg 21 [68] Kim 25 [69] Increased BAT activity in Cidea / Prevents obesity Zhou 23 [7] BAT transplantation Reversal obese phenotype Gunawarda 212 [71] Stanford 213 [72] Zhu 214 [73] P53 Prevents obesity Increases energy expenditure Al Massadi 216 [74] USF1 Prevents obesity Laurila 216 [75] 24

27 Immune regulation in adipose tissue Table 1. continued Model Effect Reference Immune cell models 16 weeks HFD Increases inflammatory markers in BAT Roberts Toler 215 [76] Cancer associated cachexia Increases M2 macrophage markers in beiging WAT Petruzelli 214 [77] Macrophage depletion Cold sensitive Nguyen 211 [3] Cold exposure Increases M2 macrophages in WAT and BAT Nguyen 211 [3] Il 4/IL 13 / STAT6 / IL 4R / 4get ΔdblGATA / (lacking Impairs cold induced beiging Qiu 214 [5] eosinophils) CCR2 / IL 4 injections Reduces body weight Improves insulin sensitivity Qiu 214 [5] Increases beiging IL 33 / Obesity prone on a HFD Less beige adipocytes in WAT Brestoff 215 [6] IL 33 injections Increases energy expenditure Increases beiging Lee 215 [4] Treg depletion Increased inflammation in BAT Cold sensitive Medrikova 215 [78] Adiponectin / Impairs cold induced beiging Hui 215 [79] Germ free mice Antibiotic depleted gut microbiota Macrophage depletion Increases Eosinophils, type 2 cytokines and M2 macrophages Improves insulin sensitivity Increases beiging Decreases inflammation in WAT Lowers body weight Decreases glucose levels Suarez Zamorano 215 [8] Chevalier 215 [81] Sakamoto 216 [82] IL 6 injections Increases Ucp1 in WAT Knudsen 214 [83] Anti IL 6 antibody Decreases Ucp1 in WAT Petruzelli 214 [77] IL 6 / Reduces energy expenditure Cold sensitive Knudsen 214 [83] Adipocyte specific CXCR4 / Severe obesity upon HFD Increased inflammation in WAT Cold sensitive Yao 214 [84] BAT: brown adipose tissue, HFD: high fat diet, IL: interleukin, NST: non shivering thermogenesis, ob/ob: leptin deficient mice, Treg: regulatory T cells, UCP1: uncoupling protein 1, UCP DTA: expression of a diphtheria toxin A chain on regulatory elements of the Ucp1 gene, USF1: upstream stimulatory factor 1, WAT: white adipose tissue. 25

28 Chapter 2 biogenesis and differentiation of brown adipocytes. This results in BAT hyperplasia and an increased thermogenic capacity, ultimately leading to a reduction in body weight [34, 43, 63, 66, 85]. BAT activation also increases the influx of nutrients into brown adipocytes, including glucose and fatty acids. Glucose is taken up by GLUT1/4 receptors [86, 87] and (triglyceride derived) fatty acids by CD36 and LPL [64], which is regulated by Angptl4 [88]. Whether BAT takes up TRL derived fatty acids by holoparticle uptake of TRL or after lipolysis of TRLs is under debate [89], although current evidence points to uptake of fatty acids derived from triglycerides after lipolysis [9]. By taking up substrates, BAT activation leads to lowering of plasma glucose and triglyceride levels and can even attenuate hypercholesterolemia by indirectly enhancing hepatic clearance of TRL remnants [35, 89]. Whereas UCP1 knockout mice are more prone to develop obesity, constitutive overexpression of UCP1 in adipocytes prevents obesity in mice [67] (Table 1). Transgenic mice with adipocyte specific overexpression of winged helix/forkhead transcription factor (Foxc2), which plays a regulatory role in adipocyte metabolism, have an enhanced sensitivity of the β adrenergic/camp/pka pathway and an increased expression of Ucp1. These mice develop adiposity and display improved glucose tolerance following a HFD compared to wild type mice [68, 69]. BAT mitochondria express high levels of celldeath activator CIDE A (Cidea), that also plays a role in energy homeostasis. Mice deficient for CIDE A have an increased amount of BAT and are protected from diet induced obesity (DIO) [7]. Loss of Cidea stabilizes the AMP activated protein kinase (AMPK) complex, which enhances basal AMPK activity leading to elevated fatty acid oxidation and energy expenditure [91]. In line with this, treatment of mice with the AMPK activator metformin enhances BAT activity [92]. Fatty acid binding proteins (s) are also largely involved in energy metabolism. Mice with mutations in two related adipocyte FABPs, ap2 and mal1, are protected from DIO and exhibit increased energy expenditure [93]. Moreover, FABP4/5 knockout mice have a severely impaired thermogenesis during fasting due to the depletion of energy storage and reduced energy supply in BAT and skeletal muscle [61]. Three research groups have been able to increase the amount of BAT by successfully transplanting BAT in diabetic or obese mice and observed a reversal of their obese phenotype [71 73]. These studies thus suggest that increasing BAT activity has a great potential as treatment to reverse obesity. BAT activation and atherosclerosis An interesting therapeutic effect of activating BAT includes the reduction of dyslipidemia associated atherosclerosis. In APOE*3 Leiden.CETP mice, which have a human like lipoprotein metabolism through expression of a mutated variant of the human APOE*3 gene combined with human cholesteryl ester transfer protein (CETP), BAT activation by β 3 adrenergic receptor stimulation was shown to be protective against diet induced atherosclerosis [35]. β 3 adrenergic receptor agonist treatment decreased plasma triglyceride and cholesterol levels [35]. Notably, using ApoE knockout or LDLR knockout mice, β 3 adrenergic receptor agonist treatment similarly reduced plasma triglycerides but did not affect plasma cholesterol levels or atherosclerosis [35]. In fact, cold exposure of ApoE knockout or LDLR knockout mice increased plasma levels of small low density lipoprotein (LDL) remnants, leading to unfavourable plasma lipid levels and even accelerated development of atherosclerotic lesions [65]. These findings show that in the presence of an intact ApoE LDLR clearance pathway, as is present in APOE*3 Leiden.CETP mice, BAT mediated local lipolysis of triglyceride rich lipoproteins stimulates the hepatic clearance of lipoprotein remnants via ApoE and 26

29 Immune regulation in adipose tissue the LDLR. As a result, BAT activation reduces atherosclerotic lesion size and severity in APOE*3 Leiden.CETP mice but not in mice lacking ApoE or the LDL receptor [35]. BAT and obesity in humans Presence of active BAT in human adults In humans, BAT was initially considered to be only present in new borns, but was recently detected to be still present in adults as well [9, 94, 95]. Simultaneous examinations of positron emission tomographic (PET) and X ray computed tomography (CT) revealed sites in human adult adipose tissue with increased 18 F fluorodeoxy glucose (FDG) uptake, indicating metabolically active tissue. Tissue biopsies confirmed that the FDG intense tissue was indeed composed of brown adipocytes, and thus could be classified as BAT [95]. In humans, cold exposure increases glucose uptake rate in BAT by fold compared to thermoneutral conditions [95, 96]. This was reflected by an increased FDG uptake, as well as by an increased fatty acid uptake and a higher activation of oxidative metabolism, which was demonstrated by subsequent studies in which humans were subjected to cold [95, 97]. The β adrenergic blocker propranolol reduces uptake of FDG in BAT depots in humans [98] and treatment with mirabegron, a β 3 adrenergic receptor agonist approved to treat the overactive bladder, increases the metabolic activity of BAT evidenced by increased FDG uptake in healthy male subjects [99]. Studies in humans comparing gene expression profiles of white, beige and brown adipocytes claim that BAT in adult humans is mainly composed of beige adipocytes [21 23, 1, 11] while others have found markers for classical BAT in humans [12]. Controversy exists on brown and beige markers and further studies will be needed to unequivocally characterize BAT in humans. Other discrepancies can be caused by sampling bias (i.e. variations in location where the biopsies were taken) and differences in age or body mass index of the studied population. It is likely that beige adipocytes remain present in the adult state when hypothermia is a less frequent threat than in newborns or rodents. The capability of beige cells to switch between a state of energy storage and energy dissipation is intriguing and further studies on how this switch works are needed. BAT activity and obesity Along with the discovery of BAT in adult humans by PET/CT studies, it also became clear that the presence of BAT, visualized through FDG uptake, showed a negative correlation with BMI and body fat percentage [9, 94]. A study of 2 randomly selected PET/CT scans showed a higher frequency of functionally active BAT in females than in males and a negative correlation with age, BMI, but also beta blocker use, outdoor temperature at the time of the scan and season [13, 14]. Interestingly, recent data show that cold induced fatty acid uptake by BAT is similar in individuals with type 2 diabetes, age matched controls and in healthy young controls [15], suggesting that previous conclusions based on FDG uptake may also reflect altered insulin sensitivity of BAT rather than BAT activity. Since stimulating BAT results in enhanced energy expenditure, it is an intriguing target for the control of whole body energy balance, adiposity and obesity. The first indication that BAT activity could be increased in humans was shown in a PET/CT study, where ten morbidly obese patients with an average BMI of 42 were scanned before and 1 year after bariatric surgery. Before surgery, FDG 27

30 Chapter 2 uptake by BAT was seen in only two patients, whereas 1 year after surgery, when the average BMI was 3, active BAT was present in five patients [16]. Interestingly, also repeated cold exposure can increase BAT activity in humans. Successful cold acclimation protocols include 6 hours of cold exposure for 1 consecutive days or daily exposure (17 C) for 2 hours during 6 weeks [2, 17]. Using these protocols, it was possible to increase BAT volume by 37% or increase FDG uptake by 6% and decrease fat mass in humans [2]. Although still scarce, more and more studies on human brown or beige adipocyte development, function and how they are modulated are published. Min et al. reported that pro angiogenic factors drive the proliferation of human beige adipocyte progenitors and activate beige adipocytes which, when transplanted in mice, improves systemic glucose turnover [18]. Accordingly, human individuals with a higher abundance of BAT have lower blood glucose levels and glucose uptake in BAT is associated with improvements in systemic glucose homeostasis and insulin sensitivity [19, 11]. Increased supraclavicular BAT activity in humans is inversely associated with arterial inflammation and reduced risk of cardiovascular events [111]. Interestingly, South Asians have a higher prevalence of hyperglycemia, dyslipidemia and cardiovascular disease and a lower amount of BAT [112]. Glucose uptake by BAT in humans is associated with heat production following a circadian rhythm in which BAT may buffer glucose fluctuations and maintain whole body glucose homeostasis over time [113]. In another study, 48 6 hrs of fasting induced insulin resistance resulted in a considerable decrease in BAT glucose uptake and non shivering thermogenesis during cold stimulation [114]. The effects of anti obesity treatments on human BAT have not been studied in detail but mouse studies suggest a mechanism mediated by BAT for some treatments. Compounds used in humans to improve dyslipidemia, hyperglycemia and/or lower plasma triglyceride levels such as metformin, rimonabant, salsalate and activation of the glucagon like peptide 1 receptor all activate BAT in mice [92, ]. Chronic low grade inflammation in obesity Dysfunctional adipocytes in WAT attract immune cells In obesity, excessive energy intake is accompanied by an increased storage of lipids in adipose tissues leading to hypertrophy of adipocytes, hypoxia, and cell death, causing WAT dysfunction and fibrosis. Dysfunctional adipocytes change the local microenvironment with leakage of fatty acids and other products resulting from adipocyte cell death. This causes a release of adipokines, chemokines and cytokines by WAT and subsequent recruitment of inflammatory cells [118, 119]. The chronic nature of obesity affects steady state homeostasis and leads to continuous activation of the immune system [118]. Due to the extensive communication between adipocytes and immune cells, chronic inflammation disturbs the homeostatic regulation of insulin signalling and adipogenesis in white adipocytes, leading to reduced insulin sensitivity and development of type 2 diabetes [118, 12]. Thus, by increasing circulating cytokines and attracting immune cells, WAT contributes to metabolic dysfunction. 28

31 Immune regulation in adipose tissue Inflammation in obese white adipose tissue Myeloid cell recruitment into WAT The first proof that inflammation is important in the pathogenesis of obesity and the resulting metabolic dysfunction was provided by Hotamisligil et al [121]. The pro inflammatory cytokine tumor necrosis factor α (TNF) was found to be present in WAT and correlated with insulin resistance in humans and mice [121]. It became evident that the infiltration of pro inflammatory macrophages in WAT plays a central role in the inflammatory response as a dominant source of TNF [122]. In WAT of lean mice, 1 15% of the cells are macrophages, whereas obese WAT contains 45 6% macrophages [122]. Resident macrophages in lean WAT have a predominant anti inflammatory phenotype, whereas in obesity, inflammatory Ly6C high CCR2 + monocytes are recruited to WAT, where they differentiate, acquire a pro inflammatory or M1 phenotype and form the majority of macrophages [12, ]. Anti inflammatory or M2 macrophages depend on the cytokines IL 4 and IL 13 and require STAT6 to maintain their alternative activation state [6]. Other myeloid cells that play a role in WAT include neutrophils and eosinophils. Neutrophils are very short lived cells that are already present in WAT within 3 days of HFD [128, 129]. In contrast, the amount of eosinophils is inversely correlated with adiposity, and exhaustion of eosinophils in mice results in increased body weight, glucose intolerance and insulin resistance [13]. Both type 2 innate lymphoid cells (ILC2s) and eosinophils have only recently been shown to be an important cell population in WAT and are a predominant source of IL 4 and IL 13, the cytokines required for the induction of M2 macrophage polarization [13, 131]. Lymphoid cell infiltration in WAT T cells are also a component of the repertoire of immune cells found in WAT. Ten percent of the stromal vascular fraction of lean WAT consists of T cells. A large part of these are CD4 + T helper cells, of which approximately 5% are regulatory T cells (Tregs). In humans, the number of T cells in WAT correlates with BMI [132]. In mice, the amount of T cells in WAT increases within 2 weeks of HFD. There are only a few CD8 + cytotoxic T cells and CD4 + effector T cells in lean WAT, but both populations increase drastically in an obese state, whereas CD4 + Tregs decrease [133]. Similarly, as the ratio between M1 and M2 macrophages increases in obese WAT, the Th1 and Th2 T cell ratio does as well. This results in a decrease in Th2 derived cytokines such as IL 4 and IL 13, thereby reducing M2 macrophage polarization. An increase in Th1 T cells and cytotoxic T cells results in excessive secretion of TNF and IFNγ, which polarizes macrophages to a pro inflammatory state, resulting in increased inflammation in obese WAT [13, ]. The chronic low grade inflammation in obese WAT also includes the recruitment of B cells, natural killer (NK) cells and mast cells [137, 138]. NK cells are activated by recognition of lipid antigens and mast cells contain granules that can release a variety of mediators, including histamine, serotonin and cytokines, which also promote recruitment of inflammatory cell types [139]. Overall, a variety of immune cells infiltrate WAT in DIO, inducing a switch from a homeostatic antiinflammatory environment to a state of chronic low grade inflammation. 29

32 Chapter 2 The immune system in brown and beige adipose tissue In contrast to the established role of the different immune cells in WAT, the contribution of the immune system to the development, function and activity of BAT is still largely unknown. However, we do know that obese individuals have a decreased amount of active BAT, based on FDG uptake, which is related to their low grade inflammatory state. Moreover, an inactive brown adipocyte accumulates lipids, similar to a white adipocyte and ablation of noradrenergic input by selective sympathetic denervation of BAT indeed results in a whitened appearance of brown adipocytes with large intracellular vacuoles [53]. Since the recruitment of macrophages into WAT is correlated with lipolysis of stored triglycerides [12], it is likely that release of fatty acids also induces recruitment of immune cells in BAT. However, whether this is indeed the case is still unknown. In diet induced obesity (DIO), thermoneutral housing leads to an additive increase in inflammation in white adipose tissue and in the vasculature compared to normal housing conditions. Although not causing increased insulin resistance, the increase in vascular inflammation does cause enhanced progression of atherosclerosis [14], indicating that BAT protects against obesity induced atherosclerosis. In another study, a similar phenomenon was observed. Immune compromised C57Bl/6 nude mice experience cold stress when housed at 23 C which modulates energy and body weight homeostasis and are therefore protected from DIO. However, at thermoneutrality (33 C), they do develop DIO with increased adiposity, hepatic triglyceride accumulation, increased inflammatory markers and glucose intolerance [141], showing that BAT activity protects against metabolic disarray and adipose tissue inflammation. Besides environmental temperature, other incentives such as the biological clock [53, 142], hormones [143, 144] and food intake not only modulate energy expenditure via the hypothalamus but also affect inflammation. For instance, time restricted feeding of a HFD for 8 hours per day increases BAT activity and reduces adipocyte hypertrophy and inflammation in WAT compared to ad libitum HFD fed mice [145]. Gut hormones such as GLP 1 mediate effects on food intake, energy expenditure and inflammation. GLP 1 receptor signalling activates BAT and promotes beiging of WAT [117, 146] while GLP 1 also reduces macrophage infiltration and inflammatory signalling in white adipocytes and macrophages [147, 148]. Other hormonal changes such as menopause also affect energy metabolism. Estradiol inhibits AMPK in the hypothalamus, which activates thermogenesis in BAT [143]. Indeed, ovariectomised mice with reduced estradiol levels gain more weight than sham operated mice and have reduced energy expenditure and increased WAT inflammation [144]. Currently, many other factors that affect both BAT activity and inflammation, such as p53 [74] and USF1 [75] are being investigated (Table 1). As described below, reports have also provided evidence that immune cells are directly involved in regulating the activity of brown as well as beige adipocytes. Inflammatory mediators in BAT Remarkably, Roberts Toler et al. [76] discovered that feeding mice a HFD for 16 weeks not only reduced insulin signalling, but also increased mrna levels of markers of inflammation in both WAT and BAT, including Tnf and the macrophage marker F4/8. Microarrays of interscapular BAT showed an upregulation of immune response gene networks after 24 weeks of HFD, including genes that indicate infiltration of leukocytes, monocytes and macrophages (CD44, CD52, CD68 and CD84). Furthermore, after 2, 4 and 8 weeks of HFD, immune gene networks were upregulated, including 3

33 Immune regulation in adipose tissue genes responsive to IFNγ; immunity related GTPase family M member 2 (Irgm2), guanylate binding protein 4 (Gbp4) and interferon gamma induced GTPase (Igtp) [149]. These gene expression profiles hint towards the presence of immune cell trafficking, leukocyte activation and lymphocyte activation in BAT. In another study, gene expression analysis in BAT of obese mice showed an upregulation of genes encoding the inflammatory cytokines TNF, IL 6, CCL2 and CCL5 [15]. Furthermore, activation of the pattern recognition receptors (PRRs), nucleotide oligomerization domain containing protein (NOD) 1, Toll like receptor (TLR) 2 and TLR4 in brown adipocytes induces a pro inflammatory response through NF κb and MAPK signalling pathways. PRRs are receptors responsible for the sensing of invading pathogens and can activate specific signalling pathways leading to distinct inflammatory responses. Enhanced PRRs expression decreased UCP1 expression as well as mitochondrial respiration in brown adipocytes [15]. Another factor affecting BAT function and BAT inflammation is p62, a protein involved in cell growth and differentiation, energy metabolism and inflammation. Global ablation, as well as adipocyte specific ablation of p62 results in obesity and insulin resistance. Moreover, these mice have decreased non shivering thermogenesis and a lower metabolic rate, which is caused by impaired mitochondrial function in brown adipocytes, due to decreased activation of p38α MAPK and its downstream regulators of thermogenesis, including UCP1, PGC 1α and CREB. Furthermore, gene expression of obese BAT in adipocyte specific p62 deficient mice reveals induction of pathways of inflammation and increased macrophage infiltration in BAT [62] (Table 1). BAT activation and beiging of WAT: Role of the immune system Macrophages have been shown to play a role in adaptive thermogenesis. Nguyen et al. [3] demonstrated that short term (6 h) cold exposure increases M2 macrophage markers in both WAT and interscapular BAT in mice. Furthermore, absence of M2 macrophages blunts BAT activity and induces cold intolerance in DIO mice. Acute cold exposure stimulates anti inflammatory M2 macrophages to produce tyrosine hydroxylase (TH), a catecholamine synthesizing enzyme, via IL 4 STAT6 signalling. This results in the release of noradrenaline, which binds to the β 3 adrenergic receptor on the brown adipocyte membrane. Consequently, BAT is activated, evidenced by the induction of Ucp1 and other thermogenic gene expression, such as Pgc 1α and acyl CoA synthase long chain family member 1 (Acsl1). An acute cold induced increase of M2 macrophages in BAT and WAT was not observed in Il 4/Il 13 and Stat6 knockout mice, indicating that the IL 4/IL 13 STAT6 pathway is crucial to increase the amount of M2 macrophages in interscapular BAT after a cold challenge [3]. However, other reports show that upon prolonged, 72h cold exposure, depletion of macrophages by clodronate or deficiency of CCR2 does not affect adaptive thermogenesis in classical BAT, but induces beiging of WAT. This suggests a less pronounced role for macrophages in adaptive thermogenesis in BAT, but rather an involvement in beiging of WAT. Especially M2 macrophages have been shown to induce beiging of subcutaneous WAT. Impaired monocyte recruitment in CCR2 knockout mice results in decreased biogenesis of beige adipocytes upon cold exposure, proving that CCR2 is responsible for cold induced monocyte recruitment [5]. Moreover, beiging of WAT is also observed when M1 to M2 polarization is enhanced through lowering of Receptor Interacting Protein 14 (RIP14) [151]. 31

34 Chapter 2 Not only can macrophages produce TH, they are also capable of producing the catecholamine noradrenaline [152, 153] and express β 2 adrenergic receptors. Binding of noradrenaline to β 2 adrenergic receptors promotes M2 polarization, enabling a paracrine loop and suggesting autoregulation of catecholamine levels [154]. Protein expression of the catecholamine producing enzyme TH increases upon cold exposure in BAT as well as subcutaneous WAT and epididymal WAT [5]. Interestingly, myeloid specific TH deficient mice have impaired cold induced biogenesis of beige adipocytes, but exhibit normal BAT activity. BAT is highly innervated by the sympathetic nervous system, along with a constitutively high expression of catecholamine synthesizing enzymes. The contribution of noradrenaline production by macrophages may therefore be less crucial in BAT then in the far less innervated WAT, which has a low basal TH expression [5]. Cold induced remodelling of subcutaneous WAT into thermogenic beige fat by noradrenaline producing macrophages was shown to be induced by eosinophils, via IL 4, IL 13 and STAT6 [5]. Mouse models with a genetic loss of Il 4/Il 13, Il 4Rα, Stat6 or IL 4 producing eosinophils have impaired cold induced biogenesis of beige adipocytes. In addition, administration of IL 4 into DIO mice at thermoneutrality reduces body weight, improves insulin sensitivity and increases protein expression of UCP1 in subcutaneous WAT, but not in classical interscapular BAT [5]. Meteorin like (Metrnl) hormone was found to be responsible for inducing IL 4 secretion by eosinophils [37]. Metrnl is a target of PGC 1α4 and is released by skeletal muscle during exercise and by adipose tissue upon cold exposure. It does not have a direct effect on brown adipocytes, but stimulates eosinophils to produce IL 4 which promotes activation of M2 macrophages and causes the increased expression of thermogenic and anti inflammatory gene programs [37]. Besides macrophages, ILC2s are important mediators of beiging of WAT (Figure 4). This subtype of innate lymphoid cells controls eosinophil and M2 macrophage responses by secreting IL 5 and IL 13, initiates type 2 immune responses and is designed to protect against helminth infections but also promote allergies [155]. In humans and mice suffering from obesity, decreased numbers of ILC2s have been detected in subcutaneous WAT [6]. The cytokine IL 33 is critical for the maintenance of ILC2s and experimental mouse studies have shown that IL 33 limits the development of spontaneous obesity. IL 33 deficient mice gain more weight on a HFD, have a reduced number of ILC2s and exhibit decreased numbers of beige adipocytes in WAT. Administration of IL 33 increases numbers of ILC2s and eosinophils in subcutaneous WAT, leading to increased energy expenditure in mice by inducing beiging of WAT [4, 6]. IL 33 induced beiging is a result of proliferation and differentiation of PDGFRα + adipocyte precursor cells. PDGFRα + pre adipocytes do not express receptors required for signalling by IL 33 (IL1RL1) or IL 5 (IL 5Rα), but they do express IL 4Rα, which is required for both IL 4 and IL 13 signalling [4]. Therefore, beiging of WAT is also dependent on IL 4/IL 13 signalling in adipocyte precursor cells, and not necessarily solely on signalling of these cytokines in myeloid cells [4]. Furthermore, IL 33 sensing ILC2s could be dysregulated in the setting of increased adiposity [156]. The presence of ILC2s and eosinophils in BAT and WAT in mice and the decrease in ILC2s and IL 33 upon high fat feeding and leptin deficiency induced obesity was confirmed by Ding et al [157]. Again, administration of IL 33 increased the number of ILC2s, eosinophils and induced Ucp1 and TH in WAT, but not BAT, of HFD fed mice, resulting in beiging of subcutaneous WAT. In addition, cold exposure increased IL 33 levels as well as the numbers of ILC2s and eosinophils in subcutaneous WAT. Denervation of subcutaneous WAT by injection of 6 hydroxydopamine (6 OHDA) suppressed basal and cold induced IL 33 levels and lowered the ILC2 and eosinophil fraction. Sympathetic denervation did not alter CD26 + M2 macrophages, showing the importance of catecholaminergic regulation in the IL 33 ILC2 eosinophil axis. 32

35 Immune regulation in adipose tissue Treg IL-33 ILC2 MetEnk Metrnl IL-5 IL-13 IL-4 NA Eosinophil M2 CCR2+ monocyte Macrophage UCP1 Figure 4. Cold induced remodelling of subcutaneous white adipose tissue into thermogenic beige adipocytes via immune cells. ILC2s are important players in beiging of white adipocytes. ILC2s are maintained by IL 33 and produce MetEnk peptide, which upregulates UCP1 in white adipose tissue. IL 33 also increases the percentage of regulatory T cells in WAT, the T cell fraction contributing to glucose homeostasis. ILC2s secrete IL 5 and IL 13, which stimulate eosinophils to produce IL 4. IL 4 can directly cause beiging or act via M2 macrophages that induce beiging by the production of noradrenaline. The increased number of macrophages upon cold exposure is dependent on monocyte recruitment via CCR2. Eosinophil dependent IL 4 secretion can also occur via Metrnl hormone, released by the adipose tissue upon cold exposure. IL; interleukin, ILC2s; group 2 innate lymphoid cells, MetEnk; methionine enkephalin, Metrnl; meteorinlike, NA; noradrenaline Treg; regulatory T cells, UCP1; uncoupling protein 1. Interestingly, Brestoff et al. [6] demonstrated that beige adipocytes can also develop via an alternative mechanism, independently of IL 4, IL 13, eosinophils or macrophages. DblGata1 knockout mice, which lack eosinophils, IL 4Rα knockout mice, which do not have IL 4 and IL 13 signalling, or ILC2 sufficient Rag2 knockout mice, which lack mature lymphocytes but have reconstituted ILC2s, still exhibit beiging of WAT, although all factors combined most likely contribute to optimal beiging. ILC2s produce an opioid like peptide, methionine enkephalin (MetEnk) peptide, which upregulates UCP1 in subcutaneous WAT and induces beiging of adipocytes without affecting IL 4 or IL 13 levels or the number of eosinophils or macrophages in WAT [6]. This mechanism was only restricted to subcutaneous WAT. MetEnk receptors MetEnk receptor δ1 opioid receptor (Oprd1) and MetEtnk receptor Opioid growth factor receptor (Ogfr) were differentially expressed in WAT and BAT, explaining these tissue specific effects [6]. These two different mechanisms as to how ILC2s control beiging of WAT; PDGFRα + adipocyte precursor cell proliferation and differentiation and MetEnk production, might synergize in the generation of beige adipocytes. The previously unrecognized feature of ILC2s to produce MetEnk would need to be confirmed in future studies, as well as the role for beige adipocyte precursors. Further studies are needed to determine the relative contribution of IL 4, IL 13 and MetEnk to beige 33

36 Chapter 2 adipogenesis. It would also be interesting to determine the source of IL 33 production, the cytokine that puts the ILC2 pathway in motion. Improved metabolic health by caloric restriction in mice also indicates that type 2 cytokine signalling is important in beiging of WAT. Caloric restriction increases the amount of eosinophils, type 2 cytokines and M2 macrophages in WAT, but does not affect the number of ILC2s, and consequently promotes beiging. The development of beige adipocytes upon caloric restriction is ablated in Stat6 knockout and IL4R knockout mice, suggesting that type 2 cytokines play an important role in beiging of WAT [158]. Toll like receptors (TLRs) respond to various pathogen associated molecules, including saturated fatty acids, by inducing signal transduction and transcription of various chemokines and cytokines, and are predominantly found on cells of the innate immune system. However, adipocytes also express several TLRs (TLR3 and TLR4) and contain relevant downstream signalling elements. Members of the interferon regulatory factor (IRF) family regulate both immune cell activation as well as adipocyte differentiation, via TLR3 and TLR4. HFD fed IRF3 deficient mice have improved insulin sensitivity, reduced adipose and systemic inflammation and enhanced beiging of subcutaneous WAT [159]. TLR4 activation in obesity or by lipopolysaccharide indeed suppresses beiging of subcutaneous WAT and causes defective BAT. In human primary white adipocytes, TLR4 activation by LPS suppresses UCP1 induction [16]. Irf4 flox/flox ;UCP1 Cre mice, lacking IRF4 specifically in UCP1 positive cells, are more obese on a HFD. IRF4 is induced by cold, promotes energy expenditure and increases thermogenic gene expression, including PGC1α and PRDM16 [161]. IRF4 was also shown to be a key transcription factor in controlling M2 polarization [162]. These data indicate that TLRs in both immune cells and adipocytes play a role in adaptive thermogenesis. The circuit in which metabolic improvements are associated with beige adipogenesis and mediated by eosinophils, type 2 cytokines and M2 macrophages is also linked with the status of the gut microbiome [8]. Germ free mice, or mice with antibiotic depleted microbiota have more beige adipocytes in both lean and obese WAT, leading to improved glucose tolerance and insulin sensitivity mediated by eosinophils, type 2 cytokines and M2 macrophages [8]. Cold exposure alters the gut microbiota by remodelling of the intestine and adipose tissue, increasing the absorptive surface of the gut. Transfer of cold adapted microbiota improves insulin sensitivity and enhances beiging of WAT [81]. With the current high prevalence of human obesity and the frequent use of antibiotics, the relation between beiging and microbiome is an interesting topic of future research. Another player related to beiging of WAT by M2 macrophages is adiponectin, an adipokine regulating glucose levels and fatty acid degradation. Adiponectin can promote the polarization of macrophages towards an M2 phenotype, correcting obesity induced macrophage infiltration and inflammation. Hui et al. show that cold exposure induces elevated gene expression and protein levels of adiponectin in subcutaneous WAT, and, using adiponectin knockout mice, show that adiponectin is essential for beiging of the subcutaneous adipose tissue via M2 macrophage proliferation [79]. Adiponectin deficiency has no effect on ILC2s or type 2 cytokines, suggesting that these are not the downstream effectors of adiponectin during cold induced beiging [79]. 34

37 Immune regulation in adipose tissue Increased numbers of M1 macrophages in obese WAT suppress cold induced beiging of subcutaneous WAT in mice. Depletion of macrophages, using clodronate liposomes, eliminates the suppressive effects of M1 macrophages on UCP1 induction and reduces the level of TNF in obese WAT. Cold exposure in clodronate treated obese mice has metabolic benefits, such as lower body weight and decreased plasma glucose levels. Mitochondrial biogenesis is not the underlying mechanism by which M1 macrophages suppress UCP1, since this was not affected by clodronate nor intraperitoneal TNF injections [82]. Interestingly, IL 33 also increases the percentage of Tregs in WAT, the T cell fraction contributing to glucose homeostasis [6]. Medrikova et al. [78] performed a microarray on sorted Tregs from BAT and identified a BAT specific subset of Tregs, which contribute to metabolic control of BAT. Upon depletion of Tregs, the tissue exerted an increased inflammatory state with an increase in macrophages, resulting in mice that were more sensitive to cold [78]. However, nude mice on a C57Bl/6 background, which lack mature T cells, experience cold stress at 23 C and are protected from DIO by increased thermogenesis and energy expenditure. This suggests that T cells are not required for adaptive thermogenesis in C57Bl/6 nude mice. At thermoneutrality, C57Bl/6 nude mice are susceptible to DIO and associated increased inflammatory markers in adipose tissues [141]. Cytokines Besides the effects of IL 4 and IL 33, which induce beiging as described above, other cytokines have also been shown to play a role in thermogenesis. Adipose tissue becomes hypoxic in the development of obesity, which contributes to the inflammatory state and the induction of TNF. This cytokine further affects adipocyte homeostasis by reducing expression of Pgc 1α and Pparγ in both white and brown adipocytes [163]. The induction of beige adipocytes in vitro by a β adrenergic receptor agonist is blunted upon stimulation with TNF or pro inflammatory cytokines derived from LPS activated macrophages. TNF is also involved in apoptosis of brown adipocytes as shown in TNF receptor deficient ob/ob mice [164]. Furthermore, TNF impairs mitochondrial biogenesis in WAT, BAT and muscle of ob/ob and DIO mice, and leptin receptor deficient Zucker fa/fa rats [165]. Extracellular signal related kinase (ERK), which mediates the induction of inflammatory cytokines, has been shown to be an important factor in the suppression of UCP1 transcriptional activation in white adipocytes and activated macrophages [166]. Together, these data indicate that cytokines can also directly influence BAT activity. Moisan et al. performed a small molecule screen to find pathways that induce a brown like thermogenic program in human adipocytes in vitro [167]. They identified tofacitinib and R46 which both target the JAK STAT1/3 pathway in adipocytes, with R46 having additional targets including AKT and ERK1/2. TNF can signal through JAK STAT and ERK pathways and although TNF represses Ucp1 expression, simultaneous stimulation with TNF and the compounds (partially) restored Ucp1 levels [167]. Furthermore, IFNγ treatment reduces Ucp1 expression in human brown adipocytes. Interferons (α/β/γ) bind to JAKs and activate STAT1/2/3 and repression of IFN signalling by JAK inhibition contributes to the upregulation of Ucp1 and promotes the metabolic beiging of adipocytes [167]. Another major cytokine in regulating inflammatory responses in obesity is IL 6, although it is unclear yet whether IL 6 serves a harmful or a protective role. IL 6 has a homeostatic role in limiting 35

38 Chapter 2 inflammation but can also have pro inflammatory effects [168]. This leads to controversial results in studies on IL 6 in obesity. Plasma IL 6 levels increase in obesity and correlate with C reactive protein levels [169] and infusion of IL 6 immediately impairs insulin sensitivity in mice [17]. However, IL 6 deficiency results in more obese mice [171], but not in all models [172]. Myeloid specific IL 6R inactivation leads to a greater propensity to develop obesity induced inflammation and glucose intolerance. IL 6 induces the IL 4 receptor and is therefore an important determinant in M2 macrophage polarization [168]. With respect to BAT, IL 6 deficient mice have reduced energy expenditure and fatty acid oxidation at room temperature. Housed at 4 C, IL 6 deficient mice have a lower core body temperature than WT mice [173]. Knudsen et al. [83] gave daily i.p. injections of IL 6 for 7 days which increased Ucp1 mrna but not UCP1 protein levels in inguinal WAT (iwat). Cold exposure increases iwat Ucp1 mrna content similarly in IL 6 deficient as WT mice, while exercise training (which induces the release of IL 6 from contracting muscle) increases iwat Ucp1 mrna in WT but not in IL 6 deficient mice. Taken together, the appropriate cytokine stimulation can induce beiging of WAT, which could be a potential mechanism to increase energy expenditure in obesity. Beiging of WAT is also found in cancer associated cachexia (wasting syndrome), where β adrenergic activation and/or inflammatory cytokine induced lipolysis cause WAT atrophy. These beige adipocytes are present within WAT and cause increased lipid mobilization and energy expenditure [77]. The cytokine IL 6 increases Ucp1 expression in WAT, whereas silencing IL 6, an anti IL 6 antibody, the nonsteroidal anti inflammatory drug sulindac or β 3 adrenergic blockade in mice reduces WAT beiging and ameliorates the severity of cachexia. This is also observed in patients [77]. The involvement of the sympathetic nervous system in BAT also contributes to a homeostatic cytokine environment. Surgical denervation of BAT or β adrenergic antagonist propranolol treatment upregulates gene expression of Tnf, ll 6 and Ccl2, without a change in F4/8 expression, indicating that the SNS and β adrenergic signalling are necessary to maintain an anti inflammatory state in BAT [174]. Chemokines Chemokines and chemokine receptors regulate leukocyte influx into obese WAT. Chemokine receptors on immune cells in WAT and serum chemokine levels are increased in obese versus lean individuals [175]. Chemokines secreted by the stromal vascular fraction (SVF) and/or white adipocytes include CCL5, CCL2, CXCL1 and CXCL12, which promote the accumulation of leukocytes into WAT through binding to their receptors CCR5, CCR6, CXCR3 and CXCR4, respectively [119, 176, 177]. The regulation of BAT homeostasis and leukocyte influx by chemokines is scarcely studied. DIO adipocyte specific CXCR4 knockout mice develop severe obesity upon HFD feeding, have an increased pro inflammatory leukocyte content in WAT and exhibit reduced thermogenic capacity when exposed to cold. This reveals that CXCR4 prevents inflammatory leukocyte influx in WAT and that adipocyte Cxcr4 expression is required for the thermogenic capacity of BAT, thereby increasing overall energy expenditure and decreasing susceptibility to DIO [84]. Lipokines Besides metabolic effects, lipids have inflammatory signalling properties. While saturated fatty acids promote adipose tissue inflammation and insulin resistance through TLR signalling [178], n 3 fatty 36

39 Immune regulation in adipose tissue acids [179] and palmitic acid esters of hydroxyl stearic acids (PAHSAs) [18] have anti inflammatory and anti diabetic effects. The lipokine palmitoleate supresses cytokine expression in adipocytes and reverses high fat induced proinflammatory macrophage polarization [181]. Palmitoleate and PAHSAs are synthesized endogenously in the adipose tissue from fatty acids [18, 182, 183] and n 3 fatty acids are derived from diet, although palmitoleate is also present in food [183]. Interestingly, adipocyte specific LPL knockout mice have increased de novo lipogenesis, particularly of palmitoleate. However, this did not increase BAT activation, WAT beiging or amelioration of inflammation in these mice [184]. PAHSAs and n 3 fatty acids were shown to exert their beneficial metabolic and anti inflammatory effects via G protein coupled receptor 12 (GPR12), a receptor for long chain fatty acids that is present in the gut, adipose tissue, pancreas and immune cells [18, 185]. Intriguingly, PAHSA levels in mice are the highest in BAT and WAT [18] and GPR12 expression is abundant in BAT and WAT which increases upon β 3 adrenergic receptor stimulation or cold [186]. Given the abundance of these metabolically advantageous and anti inflammatory lipokines in BAT and WAT, future research is needed to unravel the thermogenic effects of lipokines and lipidactivated GPRs. Concluding remarks and future perspectives Translational challenges from mice to men Extensive studies in mouse models have shown that there is a relation between BAT and obesity, and that the immune system plays an important role in regulating BAT activity and beiging of WAT, but whether this is the case in humans still needs to be investigated. Although increasing energy expenditure by activating BAT or inducing beiging of WAT prevents diet induced or genetic obesity in mice, results from different mouse models are not always relevant to the human situation. Interscapular classical BAT in mice does not regress with age whereas adult humans no longer possess this classical BAT depot. Another difficulty in extrapolating mouse data to the human is the difference in thermoneutral temperature and the activity of BAT at room temperature (18 22 C). Housing mice in their thermoneutral condition (3 C) would be comparable to the human situation, which is not always acknowledged. Many studies have shown that BAT activity in humans is thermoresponsive, and have suggested that obesity is associated with a decrease in [ 18 F]FDG uptake in BAT, which may be related to inflammation. As described, BAT activity in humans is mostly measured by [ 18 F]FDG uptake. However, the signal generated by [ 18 F]FDG in PET CT studies likely underestimates the activity of BAT as it utilizes both glucose and fatty acids as fuel [15, 187]. Furthermore, insulin resistance reduces glucose uptake by BAT leaving the uptake of fatty acids and oxidative capacity unaffected [15]. Therefore, it has been postulated that the use of the fatty acid tracer [ 18 F]FTHA is preferred over the use of [ 18 F]FDG [187]. It would even be better to develop a triglyceride with labelled fatty acids as tracer, since BAT mostly takes up fatty acids derived from TRLs in an LPL dependent manner [89]. Besides alternative methods to quantify BAT activity independent of insulin sensitivity, identification of specific circulating markers for activated human BAT is highly desired. Interestingly, mechanisms of beiging of WAT display more similarities between both species. Mice have a large beiging capacity, especially of the subcutaneous WAT. Although beiging capacity and its 37

40 Chapter 2 contribution to energy expenditure in humans still needs to be determined in detail, studies support a similar mechanism. A 6 weeks 2 hour daily cold exposure procedure results in the reduction of body fat [2] and beiging of WAT in patients with pheochromocytoma, a catecholamine secreting tumor, leads to increased metabolism [188]. Hormones, immune cells or cytokines that induce beige adipocytes in mice and are also present in humans, suggesting that beiging via those mechanisms can also occur in humans. For example, Brestoff et al. confirm presence of ILC2s in human adipose tissue, suggesting that the circuit of ILC2s, eosinophils, type 2 cytokines and M2 macrophages that induce beiging in mice, is also operational in human beiging [6]. Speculation on human BAT from an evolutionary perspective BAT evolved as a natural defence system against hypothermia in mammals, increasing the adaptability to explore colder environments [189]. Major threats besides cold in the course of evolution include limited food supply and infections. Cold challenges were accompanied by a limited food supply, making BAT a possible survival organ, enabling efficient food hunting in a subthermal environment [113]. Formerly common helminth and parasite infections elicit an M2 immune response which appear to be important regulators of beige adipocytes [6]. Along with the migration to colder areas, both the helminth and host could potentially benefit from increased WAT beiging by promoting host survival in cold climates [19]. Immunity and cold adaptation are therefore connected. Furthermore, as starvation and infection co occurred, chemokine secretion by adipose tissue could be an evolutionary advantage in which a beneficial genotype to combat infection and higher amounts of adipose tissue promotes survival. Another adaptive conservative trait is infection induced insulin resistance, enabling nutrient supply to immune cells [191]. Although BAT seems to have lost part of its function for evolutionary reasons, young people and possibly also adults may still benefit from it. BAT participates in glucose and fatty acid clearance and may still serve as a nutrient buffering system [15, 113]. The role of thermogenesis in handling excessive energy was already shown in a study in 1979, during which diet induced thermogenesis limited weight gain after a high caloric meal [192]. Therapeutic potential Potential human drugs that induce beiging of WAT are currently being tested in the clinic, as reviewed by Giordano et al [193]. A possible target would be the β 3 adrenergic receptor, although it is not specific for brown adipocytes and is expressed in a variety of organs. Unfortunately, β adrenergic receptor agonists have so far not been shown to have major effects on energy balance. A latest generation β adrenergic receptor agonist, mirabegron (approved to treat the overactive bladder), increases the metabolic activity of BAT evidenced by increased [ 18 F]FDG uptake in healthy male subjects [99]. In mice, mirabegron also has anti inflammatory effects [76]. Aside from directly activating brown adipocytes or inducing beiging of WAT with β 3 adrenergic receptor agonists, antiinflammatory drugs may induce expansion of BAT and beiging of WAT and promote energy expenditure as an anti obesity treatment, including salsalate or amlexanox [116, 193]. A few drugs with high potential for human drug development target factors downstream of the β 3 adrenergic receptor, such as C/EPBβ and PRDM16 and have extensively been studied in mice, and proven to induce differentiation of beige and brown adipocytes in mice [193]. 38

41 Immune regulation in adipose tissue Besides receptor targeted drugs, diet and nutritional components as ways to modulate thermogenesis and inflammation can be considered as an alternative strategy. Although it is known that brown, beige and white adipocytes are fuelled by glucose and fatty acids, large knowledge gaps exist. We still do not know whether different types of dietary fatty acids or carbohydrates elicit distinct effects on thermogenesis and inflammation. Whereas saturated fatty acids promote inflammation and are detrimental for metabolic health [178], n 3 fatty acids are anti inflammatory and act beneficially [179]. Whether dietary n 3 fatty acids activate BAT or promote beiging remains to be investigated. An interesting target would be PPARγ, which can be activated by these polyunsaturated fatty acids and plays a role in both beige and brown adipogenesis as in M2 antiinflammatory macrophage activation [194]. As for specific carbohydrates, hardly anything is known about their effects on BAT activity or beiging. Almost 3 years ago, Walgren et al. showed that dietary carbohydrates increased noradrenaline turnover in heart and/or BAT in rats, unrelated to the type of carbohydrate (i.e. fructose, sucrose, dextrose, corn starch) [195]. More recent investigations have not looked into the effects of specific carbohydrates on BAT activation or beiging. Future research might shed light on this underexposed aspect of modulating thermogenesis. Conclusion Excessive energy intake results in increased storage of lipids in both white and brown adipocytes, which challenges the function of these cells. Immune cells and signals in WAT and BAT are indispensable for the homeostasis of the tissue and contribute to the efflux of lipids stored in white adipocytes and to high rates of oxidation in brown and beige adipocytes. Immune cells, including eosinophils and alternatively activated macrophages, have regulatory roles in metabolic homeostasis of both WAT and BAT and research to identify immunological players is ongoing. If the mechanism is unravelled in detail, immune regulation is an intriguing therapeutic target in increasing energy expenditure to reduce weight gain. Notably, the numbers of immune cells in lean as well as obese BAT are much lower than in WAT, indicating that BAT is relatively more resistant to diet induced inflammation, but increased tissue inflammation in BAT does occur upon a positive energy balance. The challenge is to identify the metabolic crosstalk between immune cells and brown, beige and white adipocytes and the order of events that occur during obesity development. BAT regulation by the immune system will not come down to an individual immune cell type and will involve significant crosstalk between different cell types. Important questions to further address include: What are the immune regulatory effector molecules that are secreted by brown adipocytes (or a pre beige adipocyte or a white wanting to become beige adipocyte) to attract or regulate immune cells? How is BAT activity regulated in obesity? What is the role of the sympathetic nerve system? And how does BAT activity change during aging? In conclusion, the presence of BAT in humans and the potential activation of resident BAT or induction of beige adipocytes in WAT is an interesting target to treat or even prevent obesity related disorders. Cold is still by far the strongest sympathetic signal to activate BAT, but the quest for identifying biochemical and immunological pathways that are responsible for BAT activation, and thereby can bypass prolonged cold exposure, is ongoing. The recent finding on the role for immune cells in brown and beige adipocyte development and physiology harbours a great potential to increase BAT activity and beneficially alter energy metabolism by interfering in immune responses. 39

42 Chapter 2 Corresponding author Esther Lutgens, Department of Medical Biochemistry, Experimental Vascular Biology, Academic Medical Centre, University of Amsterdam, Meibergdreef 15, 115 AZ, Amsterdam, The Netherlands. Tel. +31 () e.lutgens@amc.uva.nl Disclosure statement The authors have nothing to disclose. Acknowledgements We acknowledge the support from the Netherlands Organization for Scientific Research (NWO)(VICI grant to E.L.), the Rembrandt Institute of Cardiovascular Science (RICS), the Netherlands CardioVascular Research Initiative (CVON211 19), the Deutsche Forschungsgemeinschaft (DFG) (SFB1123 A5 to E.L) and the European Research Council (ERC consolidator grant to E.L.). 4

43 Immune regulation in adipose tissue References [1] Berrington de Gonzalez A, Hartge P, Cerhan JR, Flint AJ, Hannan L, MacInnis RJ et al. 21. Body Mass Index and Mortality among 1.46 Million White Adults. New England Journal of Medicine 363(23): [2] Yoneshiro T, Aita S, Matsushita M, Kayahara T, Kameya T, Kawai Y et al Recruited brown adipose tissue as an antiobesity agent in humans. The Journal of Clinical Investigation 123(8): [3] Nguyen KD, Qiu Y, Cui X, Goh YPS, Mwangi J, David T et al Alternatively activated macrophages produce catecholamines to sustain adaptive thermogenesis. Nature 48(7375): [4] Lee M W, Odegaard JI, Mukundan L, Qiu Y, Molofsky AB, Nussbaum JC et al Activated Type 2 Innate Lymphoid Cells Regulate Beige Fat Biogenesis. Cell 16(1 2): [5] Qiu Y, Nguyen Khoa D, Odegaard Justin I, Cui X, Tian X, Locksley Richard M et al Eosinophils and Type 2 Cytokine Signaling in Macrophages Orchestrate Development of Functional Beige Fat. Cell 157(6): [6] Brestoff JR, Kim BS, Saenz SA, Stine RR, Monticelli LA, Sonnenberg GF et al Group 2 innate lymphoid cells promote beiging of white adipose tissue and limit obesity. Nature 519(7542): [7] Warwick PM, Busby R Influence of mild cold on 24 h energy expenditure in normally clothed adults. British Journal of Nutrition 63(3): [8] Ivanov KP. 2. Physiological blocking of the mechanisms of cold death: theoretical and experimental considerations. Journal of Thermal Biology 25(6): [9] van Marken Lichtenbelt WD, Vanhommerig JW, Smulders NM, Drossaerts JMAFL, Kemerink GJ, Bouvy ND et al. 29. Cold Activated Brown Adipose Tissue in Healthy Men. New England Journal of Medicine 36(15): [1] Smith RE Thermoregulatory and Adaptive Behavior of Brown Adipose Tissue. Science 146(3652):1686. [11] de Jong JMA, Larsson O, Cannon B, Nedergaard J A stringent validation of mouse adipose tissue identity markers. American Journal of Physiology Endocrinology And Metabolism 38(12):E185. [12] Nakamura K, Morrison SF. 28. A thermosensory pathway that controls body temperature. Nature Neuroscience 11(1): [13] Morrison SF, Nakamura K, Madden CJ. 28. Central control of thermogenesis in mammals. Experimental Physiology 93(7): [14] Lowell BB, Spiegelman BM. 2. Towards a molecular understanding of adaptive thermogenesis. Nature 44(6778): [15] Cannon B, Nedergaard J. 24. Brown Adipose Tissue: Function and Physiological Significance. Physiological Reviews 84(1): [16] van Marken Lichtenbelt WD Brown adipose tissue and the regulation of nonshivering thermogenesis. Current Opinion in Clinical Nutrition & Metabolic Care 15(6): [17] Cao W, Daniel KW, Robidoux J, Puigserver P, Medvedev AV, Bai X et al. 24. p38 Mitogen Activated Protein Kinase Is the Central Regulator of Cyclic AMP Dependent Transcription of the Brown Fat Uncoupling Protein 1 Gene. Molecular and Cellular Biology 24(7): [18] Puigserver P, Spiegelman BM. 23. Peroxisome Proliferator Activated Receptor γ Coactivator 1α (PGC 1α): Transcriptional Coactivator and Metabolic Regulator. Endocrine Reviews 24(1):78 9. [19] Lidell ME, Enerback S. 21. Brown adipose tissue a new role in humans? Nature Reviews Endocrinology 6(6): [2] Timmons JA, Wennmalm K, Larsson O, Walden TB, Lassmann T, Petrovic N et al. 27. Myogenic gene expression signature establishes that brown and white adipocytes originate from distinct cell lineages. Proceedings of the National Academy of Sciences 14(11): [21] Liu W, Shan T, Yang X, Liang S, Zhang P, Liu Y et al A heterogeneous lineage origin underlies the phenotypic and molecular differences of white and beige adipocytes. Journal of Cell Science 126(16): [22] Sharp LZ, Shinoda K, Ohno H, Scheel DW, Tomoda E, Ruiz L et al Human BAT Possesses Molecular Signatures That Resemble Beige/Brite Cells. PLoS ONE 7(11):e [23] Wu J, Boström P, Sparks Lauren M, Ye L, Choi Jang H, Giang A H et al Beige Adipocytes Are a Distinct Type of Thermogenic Fat Cell in Mouse and Human. Cell 15(2): [24] Seale P, Bjork B, Yang W, Kajimura S, Chin S, Kuang S et al. 28. PRDM16 controls a brown fat/skeletal muscle switch. Nature 454(727): [25] Kajimura S, Seale P, Kubota K, Lunsford E, Frangioni JV, Gygi SP et al. 29. Initiation of myoblast to brown fat switch by a PRDM16 C/EBP β transcriptional complex. Nature 46(7259): [26] Petrovic N, Shabalina IG, Timmons JA, Cannon B, Nedergaard J. 28. Thermogenically competent nonadrenergic recruitment in brown preadipocytes by a PPARγ agonist. American Journal of Physiology Endocrinology and Metabolism 295(2):E287 E296. [27] Sanchez Gurmaches J, Hung C M, Sparks Cynthia A, Tang Y, Li H, Guertin David A PTEN Loss in the Myf5 Lineage Redistributes Body Fat and Reveals Subsets of White Adipocytes that Arise from Myf5 Precursors. Cell Metabolism 16(3): [28] Sanchez Gurmaches J, Guertin DA Adipocytes arise from multiple lineages that are heterogeneously and dynamically distributed. Nature Communications 5. 41

44 Chapter 2 [29] Shan T, Liang X, Bi P, Zhang P, Liu W, Kuang S Distinct populations of adipogenic and myogenic Myf5 lineage progenitors in white adipose tissues. Journal of Lipid Research 54(8): [3] Wang W, Kissig M, Rajakumari S, Huang L, Lim H w, Won K J et al Ebf2 is a selective marker of brown and beige adipogenic precursor cells. Proceedings of the National Academy of Sciences 111(4): [31] Cousin B, Cinti S, Morroni M, Raimbault S, Ricquier D, Penicaud L et al Occurrence of brown adipocytes in rat white adipose tissue: molecular and morphological characterization. Journal of Cell Science 13(4): [32] Himms Hagen J, Melnyk A, Zingaretti MC, Ceresi E, Barbatelli G, Cinti S. 2. Multilocular fat cells in WAT of CL treated rats derive directly from white adipocytes. The American Journal of Physiology Cell Physiology 279(3):C67 C681. [33] Granneman JG, Li P, Zhu Z, Lu Y. 25. Metabolic and cellular plasticity in white adipose tissue I: effects of β3 adrenergic receptor activation. American Journal of Physiology Endocrinology and Metabolism 289(4):E68 E616. [34] Yoneshiro T, Saito M Activation and recruitment of brown adipose tissue as anti obesity regimens in humans. Annals of Medicine:1 9. [35] Berbée JFP, Boon MR, Khedoe PPSJ, Bartelt A, Schlein C, Worthmann A et al Brown fat activation reduces hypercholesterolaemia and protects from atherosclerosis development. Nature Communications 6. [36] Seale P, Conroe HM, Estall J, Kajimura S, Frontini A, Ishibashi J et al Prdm16 determines the thermogenic program of subcutaneous white adipose tissue in mice. The Journal of Clinical Investigation 121(1): [37] Rao RR, Long JZ, White JP, Svensson KJ, Lou J, Lokurkar I et al Meteorin like Is a Hormone that Regulates Immune Adipose Interactions to Increase Beige Fat Thermogenesis. Cell 157(6): [38] Carrière A, Jeanson Y, Berger Müller S, André M, Chenouard V, Arnaud E et al Browning of White Adipose Cells by Intermediate Metabolites: An Adaptive Mechanism to Alleviate Redox Pressure. Diabetes 63(1): [39] Bostrom P, Wu J, Jedrychowski MP, Korde A, Ye L, Lo JC et al A PGC1 a dependent myokine that drives brown fat like development of white fat and thermogenesis. Nature 481(7382): [4] Fisher fm, Kleiner S, Douris N, Fox EC, Mepani RJ, Verdeguer F et al FGF21 regulates PGC 1α and browning of white adipose tissues in adaptive thermogenesis. Genes & Development 26(3): [41] Qian S W, Tang Y, Li X, Liu Y, Zhang Y Y, Huang H Y et al BMP4 mediated brown fat like changes in white adipose tissue alter glucose and energy homeostasis. Proceedings of the National Academy of Sciences 11(9):E798 E87. [42] Boon MR, van den Berg SAA, Wang Y, van den Bossche J, Karkampouna S, Bauwens M et al BMP7 Activates Brown Adipose Tissue and Reduces Diet Induced Obesity Only at Subthermoneutrality. PLoS ONE 8(9):e7483. [43] Barbatelli G, Murano I, Madsen L, Hao Q, Jimenez M, Kristiansen K et al. 21. The emergence of cold induced brown adipocytes in mouse white fat depots is determined predominantly by white to brown adipocyte transdifferentiation. American Journal of Physiology Endocrinology and Metabolism 298(6):E1244 E1253. [44] Cinti S. 22. Adipocyte differentiation and transdifferentiation: Plasticity of the adipose organ. Journal of Endocrinological Investigation 25(1): [45] Lee Y H, Petkova Anelia P, Mottillo Emilio P, Granneman James G In Vivo Identification of Bipotential Adipocyte Progenitors Recruited by β3 Adrenoceptor Activation and High Fat Feeding. Cell Metabolism 15(4): [46] Lee Y H, Petkova AP, Konkar AA, Granneman JG Cellular origins of cold induced brown adipocytes in adult mice. The FASEB Journal 29(1): [47] Rosenwald M, Perdikari A, Rülicke T, Wolfrum C Bi directional interconversion of brite and white adipocytes. Nature Cell Biology 15(6): [48] Wang QA, Tao C, Gupta RK, Scherer PE Tracking adipogenesis during white adipose tissue development, expansion and regeneration. Nature Medicine 19(1): [49] Ye R, Wang QA, Tao C, Vishvanath L, Shao M, McDonald JG et al Impact of tamoxifen on adipocyte lineage tracing: Inducer of adipogenesis and prolonged nuclear translocation of Cre recombinase. Molecular Metabolism 4(11): [5] Shimizu I, Aprahamian T, Kikuchi R, Shimizu A, Papanicolaou KN, MacLauchlan S et al Vascular rarefaction mediates whitening of brown fat in obesity. The Journal of Clinical Investigation 124(5): [51] During MJ, Liu X, Huang W, Magee D, Slater A, McMurphy T et al Adipose VEGF Links the White to Brown Fat Switch With Environmental, Genetic, and Pharmacological Stimuli in Male Mice. Endocrinology 156(6): [52] Hung C M, Calejman Camila M, Sanchez Gurmaches J, Li H, Clish Clary B, Hettmer S et al Rictor/mTORC2 Loss in the Myf5 Lineage Reprograms Brown Fat Metabolism and Protects Mice against Obesity and Metabolic Disease. Cell Reports 8(1): [53] Kooijman S, van den Berg R, Ramkisoensing A, Boon MR, Kuipers EN, Loef M et al Prolonged daily light exposure increases body fat mass through attenuation of brown adipose tissue activity. Proceedings of the National Academy of Sciences 112(21): [54] Enerback S, Jacobsson A, Simpson EM, Guerra C, Yamashita H, Harper M E et al Mice lacking mitochondrial uncoupling protein are cold sensitive but not obese. Nature 387(6628):

45 Immune regulation in adipose tissue [55] Kontani Y, Wang Y, Kimura K, Inokuma KI, Saito M, Suzuki Miura T et al. 25. UCP1 deficiency increases susceptibility to diet induced obesity with age. Aging Cell 4(3): [56] Feldmann HM, Golozoubova V, Cannon B, Nedergaard J. 29. UCP1 Ablation Induces Obesity and Abolishes Diet Induced Thermogenesis in Mice Exempt from Thermal Stress by Living at Thermoneutrality. Cell Metabolism 9(2): [57] Lowell BB, S Susulic V, Hamann A, Lawitts JA, Himms Hagen J, Boyer BB et al Development of obesity in transgenic mice after genetic ablation of brown adipose tissue. Nature 366(6457): [58] Thomas SA, Palmiter RD Thermoregulatory and metabolic phenotypes of mice lacking noradrenaline and adrenaline. Nature 387(6628): [59] Bachman ES, Dhillon H, Zhang C Y, Cinti S, Bianco AC, Kobilka BK et al. 22. β AR Signaling Required for Diet Induced Thermogenesis and Obesity Resistance. Science 297(5582): [6] Ukropec J, Anunciado R, Ravussin Y, Kozak L. 26. Leptin Is Required for Uncoupling Protein 1 Independent Thermogenesis during Cold Stress. Endocrinology 147(5): [61] Syamsunarno MRAA, Iso T, Yamaguchi A, Hanaoka H, Putri M, Obokata M et al Fatty Acid Binding Protein 4 and 5 Play a Crucial Role in Thermogenesis under the Conditions of Fasting and Cold Stress. PLoS ONE 9(3):e9825. [62] Muller TD, Lee SJ, Jastroch M, Kabra D, Stemmer K, Aichler M et al p62 Links β adrenergic input to mitochondrial function and thermogenesis. The Journal of Clinical Investigation 123(1): [63] Ravussin Y, Xiao C, Gavrilova O, Reitman ML Effect of Intermittent Cold Exposure on Brown Fat Activation, Obesity, and Energy Homeostasis in Mice. PLoS ONE 9(1):e [64] Bartelt A, Bruns OT, Reimer R, Hohenberg H, Ittrich H, Peldschus K et al Brown adipose tissue activity controls triglyceride clearance. Nature Medicine 17(2):2 25. [65] Dong M, Yang X, Lim S, Cao Z, Honek J, Lu H et al Cold Exposure Promotes Atherosclerotic Plaque Growth and Instability via UCP1 Dependent Lipolysis. Cell Metabolism 18(1): [66] Himms Hagen J, Cui J, Danforth EJ, Taatjes D, Lang S, Waters B et al Effect of CL 316,243, a thermogenic beta 3 agonist, on energy balance and brown and white adipose tissues in rats. The American Journal of Physiology 266(4 pt 2): [67] Kopecky J, Clarke G, Enerback S, Spiegelman B, Kozak LP Expression of the mitochondrial uncoupling protein gene from the ap2 gene promoter prevents genetic obesity. The Journal of Clinical Investigation 96(6): [68] Cederberg A, Grønning LM, Ahrén B, Taskén K, Carlsson P, Enerbäck S. 21. FOXC2 Is a Winged Helix Gene that Counteracts Obesity, Hypertriglyceridemia, and Diet Induced Insulin Resistance. Cell 16(5): [69] Kim JK, Kim H J, Park S Y, Cederberg A, Westergren R, Nilsson D et al. 25. Adipocyte Specific Overexpression of FOXC2 Prevents Diet Induced Increases in Intramuscular Fatty Acyl CoA and Insulin Resistance. Diabetes 54(6): [7] Zhou Z, Yon Toh S, Chen Z, Guo K, Peng Ng C, Ponniah S et al. 23. Cidea deficient mice have lean phenotype and are resistant to obesity. Nature Genetics 35(1): [71] Gunawardana SC, Piston DW Reversal of Type 1 Diabetes in Mice by Brown Adipose Tissue Transplant. Diabetes 61(3): [72] Stanford KI, Middelbeek RJW, Townsend KL, An D, Nygaard EB, Hitchcox KM et al Brown adipose tissue regulates glucose homeostasis and insulin sensitivity. The Journal of Clinical Investigation 123(1): [73] Zhu Z, Spicer EG, Gavini CK, Goudjo Ako AJ, Novak CM, Shi H Enhanced sympathetic activity in mice with brown adipose tissue transplantation (transbatation). Physiology & Behavior 125(): [74] Al Massadi O, Porteiro B, Kuhlow D, Köhler M, Gonzalez Rellan MJ, Garcia Lavandeira M et al Pharmacological and Genetic Manipulation of p53 in Brown Fat at Adult But Not Embryonic Stages Regulates Thermogenesis and Body Weight in Male Mice. Endocrinology 157(7): [75] Laurila P P, Soronen J, Kooijman S, Forsström S, Boon MR, Surakka I et al USF1 deficiency activates brown adipose tissue and improves cardiometabolic health. Science Translational Medicine 8(323):323ra13 323ra13. [76] Roberts Toler C, O'Neill BT, Cypess AM Diet induced obesity causes insulin resistance in mouse brown adipose tissue. Obesity 23(9): [77] Petruzzelli M, Schweiger M, Schreiber R, Campos Olivas R, Tsoli M, Allen J et al A Switch from White to Brown Fat Increases Energy Expenditure in Cancer Associated Cachexia. Cell Metabolism 2(3): [78] Medrikova D, Sijmonsma TP, Sowodniok K, Richards DM, Delacher M, Sticht C et al Brown Adipose Tissue Harbors a Distinct Sub Population of Regulatory T Cells. PLoS ONE 1(2):e [79] Hui X, Gu P, Zhang J, Nie T, Pan Y, Wu D et al Adiponectin Enhances Cold Induced Browning of Subcutaneous Adipose Tissue via Promoting M2 Macrophage Proliferation. Cell Metabolism 22(2): [8] Suarez Zamorano N, Fabbiano S, Chevalier C, Stojanovic O, Colin DJ, Stevanovic A et al Microbiota depletion promotes browning of white adipose tissue and reduces obesity. Nature Medicine 21(12): [81] Chevalier C, Stojanović O, Colin Didier J, Suarez Zamorano N, Tarallo V, Veyrat Durebex C et al Gut Microbiota Orchestrates Energy Homeostasis during Cold. Cell 163(6): [82] Sakamoto T, Nitta T, Maruno K, Yeh YS, Kuwata H, Tomita K et al Macrophage infiltration into obese adipose tissues suppresses the induction of UCP1 expression in mice. American Journal of Physiology Endocrinology and Metabolism 31:E676 E

46 Chapter 2 [83] Knudsen JG, Murholm M, Carey AL, Biensø RS, Basse AL, Allen TL et al Role of IL 6 in Exercise Training and Cold Induced UCP1 Expression in Subcutaneous White Adipose Tissue. PLoS ONE 9(1):e8491. [84] Yao L, Heuser Baker J, Herlea Pana O, Zhang N, Szweda LI, Griffin TM et al Deficiency in adipocyte chemokine receptor CXCR4 exacerbates obesity and compromises thermoregulatory responses of brown adipose tissue in a mouse model of diet induced obesity. The FASEB Journal. [85] Ursino MG, Vasina V, Raschi E, Crema F, De Ponti F. 29. The β3 adrenoceptor as a therapeutic target: Current perspectives. Pharmacological Research 59(4): [86] Labbé SM, Caron A, Bakan I, Laplante M, Carpentier AC, Lecomte R et al In vivo measurement of energy substrate contribution to cold induced brown adipose tissue thermogenesis. The FASEB Journal 29(5): [87] Nikami H, Shimizu Y, Endoh D, Yano H, Saito M Cold exposure increases glucose utilization and glucose transporter expression in brown adipose tissue. Biochemical and Biophysical Research Communications 185(3): [88] Dijk W, Heine M, Vergnes L, Boon MR, Schaart G, Hesselink MKC et al ANGPTL4 mediates shuttling of lipid fuel to brown adipose tissue during sustained cold exposure. elife 4:e8428. [89] Hoeke G, Kooijman S, Boon MR, Rensen PCN, Berbée JFP Role of brown fat in lipoprotein metabolism and atherosclerosis. Circulation Research 118(1): [9] Khedoe PPSJ, Hoeke G, Kooijman S, Dijk W, Buijs JT, Kersten S et al Brown adipose tissue takes up plasma triglycerides mostly after lipolysis. Journal of Lipid Research 56(1): [91] Qi J, Gong J, Zhao T, Zhao J, Lam P, Ye J et al. 28. Downregulation of AMP activated protein kinase by Cideamediated ubiquitination and degradation in brown adipose tissue. The EMBO Journal 27(11): [92] Geerling JJ, Boon MR, van der Zon GC, van den Berg SAA, van den Hoek AM, Lombès M et al Metformin Lowers Plasma Triglycerides by Promoting VLDL Triglyceride Clearance by Brown Adipose Tissue in Mice. Diabetes 63(3): [93] Maeda K, Cao H, Kono K, Gorgun CZ, Furuhashi M, Uysal KT et al. 25. Adipocyte/macrophage fatty acid binding proteins control integrated metabolic responses in obesity and diabetes. Cell Metabolism 1(2): [94] Saito M, Okamatsu Ogura Y, Matsushita M, Watanabe K, Yoneshiro T, Nio Kobayashi J et al. 29. High Incidence of Metabolically Active Brown Adipose Tissue in Healthy Adult Humans: Effects of Cold Exposure and Adiposity. Diabetes 58(7): [95] Virtanen KA, Lidell ME, Orava J, Heglind M, Westergren R, Niemi T et al. 29. Functional Brown Adipose Tissue in Healthy Adults. New England Journal of Medicine 36(15): [96] Orava J, Nuutila P, Lidell Martin E, Oikonen V, Noponen T, Viljanen T et al Different Metabolic Responses of Human Brown Adipose Tissue to Activation by Cold and Insulin. Cell Metabolism 14(2): [97] Ouellet V, Labbé SM, Blondin DP, Phoenix S, Guérin B, Haman F et al Brown adipose tissue oxidative metabolism contributes to energy expenditure during acute cold exposure in humans. The Journal of Clinical Investigation 122(2): [98] Parysow O, Mollerach AM, Jager V, Racioppi S, Roman JS, Gerbaudo VH. 27. Low Dose Oral Propranolol Could Reduce Brown Adipose Tissue F 18 FDG Uptake in Patients Undergoing PET Scans. Clinical Nuclear Medicine 32(5): [99] Cypess AM, Weiner LS, Roberts Toler C, Elía EF, Kessler SH, Kahn PA et al Activation of Human Brown Adipose Tissue by a β3 Adrenergic Receptor Agonist. Cell Metabolism 21(1): [1] Xue R, Lynes MD, Dreyfuss JM, Shamsi F, Schulz TJ, Zhang H et al Clonal analyses and gene profiling identify genetic biomarkers of the thermogenic potential of human brown and white preadipocytes. Nature Medicine 21(7): [11] Shinoda K, Luijten IHN, Hasegawa Y, Hong H, Sonne SB, Kim M et al Genetic and functional characterization of clonally derived adult human brown adipocytes. Nature Medicine 21(4): [12] 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. Nature Medicine 19(5): [13] Cypess AM, Lehman S, Williams G, Tal I, Rodman D, Goldfine AB et al. 29. Identification and Importance of Brown Adipose Tissue in Adult Humans. New England Journal of Medicine 36(15): [14] Au Yong ITH, Thorn N, Ganatra R, Perkins AC, Symonds ME. 29. Brown Adipose Tissue and Seasonal Variation in Humans. Diabetes 58(11): [15] Blondin DP, Labbé SM, Noll C, Kunach M, Phoenix S, Guérin B et al Selective Impairment of Glucose but Not Fatty Acid or Oxidative Metabolism in Brown Adipose Tissue of Subjects With Type 2 Diabetes. Diabetes 64(7):2388. [16] Vijgen G, Bouvy N, Teule G, Brans B, Hoeks J, Schrauwen P et al Increase in Brown Adipose Tissue Activity after Weight Loss in Morbidly Obese Subjects. The Journal of Clinical Endocrinology & Metabolism 97(7):E1229 E1233. [17] van der Lans AAJJ, Hoeks J, Brans B, Vijgen GHEJ, Visser MGW, Vosselman MJ et al Cold acclimation recruits human brown fat and increases nonshivering thermogenesis. The Journal of Clinical Investigation 123(8): [18] Min SY, Kady J, Nam M, Rojas Rodriguez R, Berkenwald A, Kim JH et al Human 'brite/beige' adipocytes develop from capillary networks, and their implantation improves metabolic homeostasis in mice. Nature Medicine 22(3):

47 Immune regulation in adipose tissue [19] Chondronikola M, Volpi E, Børsheim E, Porter C, Annamalai P, Enerbäck S et al Brown Adipose Tissue Improves Whole Body Glucose Homeostasis and Insulin Sensitivity in Humans. Diabetes 63(12): [11] Lee P, Smith S, Linderman J, Courville AB, Brychta RJ, Dieckmann W et al Temperature Acclimated Brown Adipose Tissue Modulates Insulin Sensitivity in Humans. Diabetes 63(11): [111] Takx R, Ishai A, Truong QA, MacNabb MH, Scherrer Crosbie M, Tawakol A Supraclavicular Brown adipose tissue FDG uptake and cardiovascular disease. Journal of Nuclear Medicine 57(8): [112] Boon MR, Bakker LEH, van der Linden RAD, van Ouwerkerk AF, de Goeje PL, Counotte J et al High prevalence of cardiovascular disease in South Asians: Central role for brown adipose tissue? Critical Reviews in Clinical Laboratory Sciences 52(3): [113] Lee P, Bova R, Schofield L, Bryant W, Dieckmann W, Slattery A et al Brown Adipose Tissue Exhibits a Glucose Responsive Thermogenic Biorhythm in Humans. Cell Metabolism 23(4): [114] Hanssen MJW, Wierts R, Hoeks J, Gemmink A, Brans B, Mottaghy FM et al Glucose uptake in human brown adipose tissue is impaired upon fasting induced insulin resistance. Diabetologia 58(3): [115] Boon MR, Kooijman S, van Dam AD, Pelgrom LR, Berbée JFP, Visseren CAR et al Peripheral cannabinoid 1 receptor blockade activates brown adipose tissue and diminishes dyslipidemia and obesity. The FASEB Journal 28(12): [116] van Dam AD, Nahon KJ, Kooijman S, van den Berg SM, Kanhai AA, Kikuchi T et al Salsalate Activates Brown Adipose Tissue in Mice. Diabetes 64(5): [117] Kooijman S, Wang Y, Parlevliet ET, Boon MR, Edelschaap D, Snaterse G et al Central GLP 1 receptor signalling accelerates plasma clearance of triacylglycerol and glucose by activating brown adipose tissue in mice. Diabetologia 58(11): [118] Lumeng CN, Saltiel AR Inflammatory links between obesity and metabolic disease. Journal of Clinical Investigation 121(6): [119] Ouchi N, Parker JL, Lugus JJ, Walsh K Adipokines in inflammation and metabolic disease. Nature Reviews Immunology 11(2): [12] Chawla A, Nguyen KD, Goh YPS Macrophage mediated inflammation in metabolic disease. Nature Reviews Immunology 11(11): [121] Hotamisligil G, Shargill N, Spiegelman B Adipose expression of tumor necrosis factor alpha: direct role in obesity linked insulin resistance. Science 259(591): [122] Weisberg SP, McCann D, Desai M, Rosenbaum M, Leibel RL, Ferrante AW, Jr. 23. Obesity is associated with macrophage accumulation in adipose tissue. The Journal of Clinical Investigation 112(12): [123] Mathis D Immunological Goings on in Visceral Adipose Tissue. Cell Metabolism 17(6): [124] Lumeng CN, Bodzin JL, Saltiel AR. 27. Obesity induces a phenotypic switch in adipose tissue macrophage polarization. The Journal of Clinical Investigation 117(1): [125] Gregor MF, Hotamisligil GS Inflammatory Mechanisms in Obesity. Annual Review of Immunology 29(1): [126] Perreault M, Marette A. 21. Targeted disruption of inducible nitric oxide synthase protects against obesitylinked insulin resistance in muscle. Nature Medicine 7(1): [127] Patsouris D, Li P P, Thapar D, Chapman J, Olefsky JM, Neels JG. 28. Ablation of CD11c Positive Cells Normalizes Insulin Sensitivity in Obese Insulin Resistant Animals. Cell Metabolism 8(4): [128] Elgazar Carmon V, Rudich A, Hadad N, Levy R. 28. Neutrophils transiently infiltrate intra abdominal fat early in the course of high fat feeding. Journal of Lipid Research 49(9): [129] Talukdar S, Oh DY, Bandyopadhyay G, Li D, Xu J, McNelis J et al Neutrophils mediate insulin resistance in mice fed a high fat diet through secreted elastase. Nature Medicine 18(9): [13] Wu D, Molofsky AB, Liang H E, Ricardo Gonzalez RR, Jouihan HA, Bando JK et al Eosinophils Sustain Adipose Alternatively Activated Macrophages Associated with Glucose Homeostasis. Science 332(626): [131] Odegaard JI, Ricardo Gonzalez RR, Goforth MH, Morel CR, Subramanian V, Mukundan L et al. 27. Macrophagespecific PPARγ controls alternative activation and improves insulin resistance. Nature 447(7148): [132] Duffaut C, Zakaroff Girard A, Bourlier V, Decaunes P, Maumus M, Chiotasso P et al. 29. Interplay Between Human Adipocytes and T Lymphocytes in Obesity: CCL2 as an Adipochemokine and T Lymphocytes as Lipogenic Modulators. Arteriosclerosis, Thrombosis, and Vascular Biology 29(1): [133] Feuerer M, Herrero L, Cipolletta D, Naaz A, Wong J, Nayer A et al. 29. Lean, but not obese, fat is enriched for a unique population of regulatory T cells that affect metabolic parameters. Nature Medicine 15(8): [134] Nishimura S, Manabe I, Nagasaki M, Eto K, Yamashita H, Ohsugi M et al. 29. CD8+ effector T cells contribute to macrophage recruitment and adipose tissue inflammation in obesity. Nature Medicine 15(8): [135] Chatzigeorgiou A, Karalis KP, Bornstein SR, Chavakis T Lymphocytes in obesity related adipose tissue inflammation. Diabetologia 55(1): [136] Winer S, Chan Y, Paltser G, Truong D, Tsui H, Bahrami J et al. 29. Normalization of obesity associated insulin resistance through immunotherapy. Nature Medicine 15(8): [137] Winer S, Winer DA The adaptive immune system as a fundamental regulator of adipose tissue inflammation and insulin resistance. Immunology & Cell Biology 9(8): [138] 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. Nature Medicine 17(5):

48 Chapter 2 [139] Liu J, Divoux A, Sun J, Zhang J, Clement K, Glickman JN et al. 29. Genetic deficiency and pharmacological stabilization of mast cells reduce diet induced obesity and diabetes in mice. Nature Medicine 15(8): [14] Tian Xiao Y, Ganeshan K, Hong C, Nguyen Khoa D, Qiu Y, Kim J et al Thermoneutral Housing Accelerates Metabolic Inflammation to Potentiate Atherosclerosis but Not Insulin Resistance. Cell Metabolism 23(1): [141] Stemmer K, Kotzbeck P, Zani F, Bauer M, Neff C, Muller TD et al Thermoneutral housing is a critical factor for immune function and diet induced obesity in C57BL/6 nude mice. International Journal of Obesity 39(5): [142] Orozco Solis R, Aguilar Arnal L, Murakami M, Peruquetti R, Ramadori G, Coppari R et al The Circadian Clock in the Ventromedial Hypothalamus Controls Cyclic Energy Expenditure. Cell Metabolism 23(3): [143] Martínez de Morentin Pablo B, González García I, Martins L, Lage R, Fernández Mallo D, Martínez Sánchez N et al Estradiol Regulates Brown Adipose Tissue Thermogenesis via Hypothalamic AMPK. Cell Metabolism 2(1): [144] Rogers NH, Perfield JW, Strissel KJ, Obin MS, Greenberg AS. 29. Reduced Energy Expenditure and Increased Inflammation Are Early Events in the Development of Ovariectomy Induced Obesity. Endocrinology 15(5): [145] Hatori M, Vollmers C, Zarrinpar A, DiTacchio L, Bushong Eric A, Gill S et al Time Restricted Feeding without Reducing Caloric Intake Prevents Metabolic Diseases in Mice Fed a High Fat Diet. Cell Metabolism 15(6): [146] Beiroa D, Imbernon M, Gallego R, Senra A, Herranz D, Villarroya F et al GLP 1 Agonism Stimulates Brown Adipose Tissue Thermogenesis and Browning Through Hypothalamic AMPK. Diabetes 63(1):3346. [147] Lee YS, Park MS, Choung JS, Kim SS, Oh HH, Choi CS et al Glucagon like peptide 1 inhibits adipose tissue macrophage infiltration and inflammation in an obese mouse model of diabetes. Diabetologia 55(9): [148] Wang Y, Parlevliet ET, Geerling JJ, van der Tuin SJL, Zhang H, Bieghs V et al Exendin 4 decreases liver inflammation and atherosclerosis development simultaneously by reducing macrophage infiltration. British Journal of Pharmacology 171(3): [149] McGregor RA, Kwon EY, Shin SK, Jung UJ, Kim E, Park JHY et al Time course microarrays reveal modulation of developmental, lipid metabolism and immune gene networks in intrascapular brown adipose tissue during the development of diet induced obesity. International Journal of Obesity 37(12): [15] Bae J, Ricciardi CJ, Esposito D, Komarnytsky S, Hu P, Curry BJ et al Activation of pattern recognition receptors in brown adipocytes induces inflammation and suppresses uncoupling protein 1 expression and mitochondrial respiration. American journal of physiology cell physiology 36(1):C918 C93. [151] Liu P S, Lin Y W, Burton FH, Wei L N M1 M2 balancing act in white adipose tissue browning a new role for RIP14. Adipocyte 4(2): [152] Brown SW, Meyers RT, Brennan KM, Rumble JM, Narasimhachari N, Perozzi EF et al. 23. Catecholamines in a macrophage cell line. Journal of Neuroimmunology 135(1 2): [153] Stanojević S, Dimitrijević M, Kuštrimović N, Mitić K, Vujić V, Leposavić G Adrenal hormone deprivation affects macrophage catecholamine metabolism and β2 adrenoceptor density, but not propranolol stimulation of tumour necrosis factor α production. Experimental Physiology 98(3): [154] Grailer JJ, Haggadone MD, Sarma JV, Zetoune FS, Ward PA Induction of M2 Regulatory Macrophages through the β2 Adrenergic Receptor with Protection during Endotoxemia and Acute Lung Injury. Journal of Innate Immunity 6(5): [155] Brestoff JR, Artis D Immune Regulation of Metabolic Homeostasis in Health and Disease. Cell 161(1): [156] Flach M, Diefenbach A Adipose Tissue: ILC2 Crank Up the Heat. Cell Metabolism 21(2): [157] Ding X, Luo Y, Zhang X, Zheng H, Yang X, Yang X et al IL 33 driven ILC2/eosinophil axis in fat is induced by sympathetic tone and suppressed by obesity. Journal of Endocrinology 231(1): [158] Fabbiano S, Suárez Zamorano N, Rigo D, Veyrat Durebex C, Stevanovic Dokic A, Colin Didier J et al Caloric Restriction Leads to Browning of White Adipose Tissue through Type 2 Immune Signaling. Cell Metabolism 24(3): [159] Kumari M, Wang X, Lantier L, Lyubetskaya A, Eguchi J, Kang S et al IRF3 promotes adipose inflammation and insulin resistance and represses browning. The Journal of Clinical Investigation 126(8): [16] Okla M, Wang W, Kang I, Pashaj A, Carr T, Chung S Activation of Toll like Receptor 4 (TLR4) Attenuates Adaptive Thermogenesis via Endoplasmic Reticulum Stress. Journal of Biological Chemistry 29(44): [161] Kong X, Banks A, Liu T, Kazak L, Rao Rajesh R, Cohen P et al IRF4 Is a Key Thermogenic Transcriptional Partner of PGC 1α. Cell 158(1): [162] Satoh T, Takeuchi O, Vandenbon A, Yasuda K, Tanaka Y, Kumagai Y et al. 21. The Jmjd3 Irf4 axis regulates M2 macrophage polarization and host responses against helminth infection. Nature Immunology 11(1): [163] Bhattacharya I, Domínguez AP, Drägert K, Humar R, Haas E, Battegay EJ Hypoxia potentiates tumor necrosis factor α induced expression of inducible nitric oxide synthase and cyclooxygenase 2 in white and brown adipocytes. Biochemical and Biophysical Research Communications 461(2): [164] Nisoli E, Briscini L, Giordano A, Tonello C, Wiesbrock SM, Uysal KT et al. 2. Tumor necrosis factor α mediates apoptosis of brown adipocytes and defective brown adipocyte function in obesity. Proceedings of the National Academy of Sciences 97(14):

49 Immune regulation in adipose tissue [165] Valerio A, Cardile A, Cozzi V, Bracale R, Tedesco L, Pisconti A et al. 26. TNF α downregulates enos expression and mitochondrial biogenesis in fat and muscle of obese rodents. The Journal of Clinical Investigation 116(1): [166] Sakamoto T, Takahashi N, Sawaragi Y, Naknukool S, Yu R, Goto T et al Inflammation induced by RAW macrophages suppresses UCP1 mrna induction via ERK activation in 1T1/2 adipocytes. The American Journal of Physiology Cell Physiology 34(8):C729 C738. [167] Moisan A, Lee Y K, Zhang JD, Hudak CS, Meyer CA, Prummer M et al White to brown metabolic conversion of human adipocytes by JAK inhibition. Nature Cell Biology 17(1): [168] Mauer J, Chaurasia B, Goldau J, Vogt MC, Ruud J, Nguyen KD et al Signaling by IL 6 promotes alternative activation of macrophages to limit endotoxemia and obesity associated resistance to insulin. Nature Immunology 15(5): [169] Weiss R, Dziura J, Burgert TS, Tamborlane WV, Taksali SE, Yeckel CW et al. 24. Obesity and the Metabolic Syndrome in Children and Adolescents. New England Journal of Medicine 35(23): [17] Kim H J, Higashimori T, Park S Y, Choi H, Dong J, Kim Y J et al. 24. Differential Effects of Interleukin 6 and 1 on Skeletal Muscle and Liver Insulin Action In Vivo. Diabetes 53(4): [171] Wallenius V, Wallenius K, Ahren B, Rudling M, Carlsten H, Dickson SL et al. 22. Interleukin 6 deficient mice develop mature onset obesity. Nature Medicine 8(1): [172] Di Gregorio GB, Hensley L, Lu T, Ranganathan G, Kern PA. 24. Lipid and carbohydrate metabolism in mice with a targeted mutation in the IL 6 gene: absence of development of age related obesity. American Journal of Physiology Endocrinology and Metabolism 287(1):E182 E187. [173] Wernstedt I, Edgley A, Berndtsson A, Fäldt J, Bergström G, Wallenius V et al. 26. Reduced stress and coldinduced increase in energy expenditure in interleukin 6 deficient mice. American Journal of Physiology Regulatory, Integrative and Comparative Physiology 291(3):R551 R557. [174] Tang L, Okamoto S, Shiuchi T, Toda C, Takagi K, Sato T et al Sympathetic Nerve Activity Maintains an Anti Inflammatory State in Adipose Tissue in Male Mice by Inhibiting TNF α Gene Expression in Macrophages. Endocrinology 156(1): [175] Anderson EK, Gutierrez DA, Hasty AH. 21. Adipose tissue recruitment of leukocytes. Current Opinion in Lipidology 21(3): [176] Sengenès C, Miranville A, Maumus M, de Barros S, Busse R, Bouloumié A. 27. Chemotaxis and Differentiation of Human Adipose Tissue CD34+/CD31 Progenitor Cells: Role of Stromal Derived Factor 1 Released by Adipose Tissue Capillary Endothelial Cells. Stem Cells 25(9): [177] Herder C, Hauner H, Kempf K, Kolb H, Skurk T. 26. Constitutive and regulated expression and secretion of interferon [gamma] inducible protein 1 (IP 1//CXCL1) in human adipocytes. International Journal of Obesity 31(3): [178] Shi H, Kokoeva MV, Inouye K, Tzameli I, Yin H, Flier JS. 26. TLR4 links innate immunity and fatty acid induced insulin resistance. The Journal of Clinical Investigation 116(11): [179] Lee JY, Plakidas A, Lee WH, Heikkinen A, Chanmugam P, Bray G et al. 23. Differential modulation of Toll like receptors by fatty acids: preferential inhibition by n 3 polyunsaturated fatty acids. Journal of Lipid Research 44(3): [18] Yore Mark M, Syed I, Moraes Vieira Pedro M, Zhang T, Herman Mark A, Homan Edwin A et al Discovery of a Class of Endogenous Mammalian Lipids with Anti Diabetic and Anti inflammatory Effects. Cell 159(2): [181] Cao H, Gerhold K, Mayers JR, Wiest MM, Watkins SM, Hotamisligil GS. 28. Identification of a Lipokine, a Lipid Hormone Linking Adipose Tissue to Systemic Metabolism. Cell 134(6): [182] Kuda O, Brezinova M, Rombaldova M, Slavikova B, Posta M, Beier P et al Docosahexaenoic Acid Derived Fatty Acid Esters of Hydroxy Fatty Acids (FAHFAs) With Anti inflammatory Properties. Diabetes 65(9):258. [183] Hodson L, Karpe F Is there something special about palmitoleate? Current Opinion in Clinical Nutrition & Metabolic Care 16(2): [184] Bartelt A, Weigelt C, Cherradi ML, Niemeier A, Tödter K, Heeren J et al Effects of adipocyte lipoprotein lipase on de novo lipogenesis and white adipose tissue browning. Biochimica et Biophysica Acta (BBA) Molecular and Cell Biology of Lipids 1831(5): [185] Oh DY, Talukdar S, Bae EJ, Imamura T, Morinaga H, Fan W et al. 21. GPR12 Is an Omega 3 Fatty Acid Receptor Mediating Potent Anti inflammatory and Insulin Sensitizing Effects. Cell 142(5): [186] Rosell M, Kaforou M, Frontini A, Okolo A, Chan Y W, Nikolopoulou E et al Brown and white adipose tissues: intrinsic differences in gene expression and response to cold exposure in mice. American Journal of Physiology Endocrinology And Metabolism 36(8):E945. [187] Schilperoort M, Hoeke G, Kooijman S, Rensen PCN Relevance of lipid metabolism for brown fat visualization and quantification. Current Opinion in Lipidology 27(3): [188] Frontini A, Vitali A, Perugini J, Murano I, Romiti C, Ricquier D et al White to brown transdifferentiation of omental adipocytes in patients affected by pheochromocytoma. Biochimica et Biophysica Acta (BBA) Molecular and Cell Biology of Lipids 1831(5): [189] Sidossis L, Kajimura S Brown and beige fat in humans: thermogenic adipocytes that control energy and glucose homeostasis. The Journal of Clinical Investigation 125(2):

50 Chapter 2 [19] Guigas B, Molofsky AB A worm of one's own: how helminths modulate host adipose tissue function and metabolism. Trends in Parasitology 31(9): [191] Odegaard JI, Chawla A Pleiotropic Actions of Insulin Resistance and Inflammation in Metabolic Homeostasis. Science 339(6116): [192] Rothwell NJ, Stock MJ A role for brown adipose tissue in diet induced thermogenesis. Nature 281(5726):31 5. [193] Giordano A, Frontini A, Cinti S Convertible visceral fat as a therapeutic target to curb obesity. Nature Reviews Drug Discovery advance online publication. [194] Varga T, Czimmerer Z, Nagy L PPARs are a unique set of fatty acid regulated transcription factors controlling both lipid metabolism and inflammation. Biochimica et Biophysica Acta 1812(8): [195] Walgren MC, Young JB, Kaufman LN, Landsberg L The effects of various carbohydrates on sympathetic activity in heart and interscapular brown adipose tissue of the rat. Metabolism 36(6):

51 3 Diet induced obesity induces rapid inflammatory changes in brown adipose tissue in mice Susan M. van den Berg 1, Andrea D. van Dam 2,3, Tom T.P. Seijkens 1, Pascal J.H. Kusters 1, Linda Beckers 1, Myrthe den Toom 1, Ntsiki M. Held 4, Saskia van der Velden 1, Jan Van den Bossche 1, Mariëtte R. Boon 2,3, Patrick C.N. Rensen 2,3, Esther Lutgens *,1,5, Menno P.J. de Winther *,1,5 1 Department of Medical Biochemistry, Experimental Vascular Biology, Academic Medical Center, University of Amsterdam, the Netherlands 2 Department of Medicine, Division Endocrinology, Leiden University Medical Center, Leiden, the Netherlands 3 Einthoven Laboratory for Experimental Vascular Medicine, Leiden, the Netherlands 4 Laboratory Genetic Metabolic Diseases, Academic Medical Center, University of Amsterdam, The Netherlands 5 Institute for Cardiovascular Prevention (IPEK), Ludwig Maximilian s University, Munich, Germany *these authors contributed equally to this work. Submitted

52 Chapter 3 Abstract Brown adipose tissue (BAT) contributes to non shivering thermogenesis by burning glucose and fat to produce heat. Obesity is associated with a loss of BAT activity. In obesity, white adipose tissue inflammation contributes to metabolic disorders. We aimed to investigate if diet induced obesity also induces inflammation in BAT and if this affects brown adipocyte activity. We performed a time course of diet induced obesity ranging from 1 day to 18 weeks by feeding mice a 45% high fat diet and subsequently analysed BAT biology. Furthermore, we stimulated a brown adipocyte cell line with the cytokines TNF, IFNγ, and LPS to mimic a pro inflammatory environment. BAT rapidly accumulated lipids after short term high fat diet (3 days). Macrophage numbers as well as cytokine and chemokine gene expression in BAT increased along with this increased lipid storage. Surprisingly, brown adipocytes in vitro were very capable of producing a variety of chemokines in response to inflammatory stimuli. BAT is a very plastic tissue, which is rapidly remodelled upon diet induced obesity. This remodelling is characterized by a swift recruitment of macrophages and induction of inflammatory mediators. Moreover, brown adipocytes themselves may be cytokine producers in the inflammatory profile of BAT. 5

53 Brown adipose tissue in obesity Introduction Brown adipose tissue (BAT) contributes to adaptive thermogenesis by burning glucose and fat to produce heat [1]. Brown adipocytes have many small lipid droplets and numerous mitochondria that contain large amounts of uncoupling protein 1 (UCP1). This protein mediates the thermogenic function by increasing the permeability of the inner mitochondrial membrane, which causes protons to flux back into the mitochondrial matrix, bypassing ATP synthesis, ultimately resulting in generation of heat [2]. Cold activates the hypothalamus in the brain, which stimulates the sympathetic nervous system (SNS) to produce noradrenaline. In turn, noradrenaline activates brown adipocytes by binding to β 3 adrenergic receptors on their surface [3]. BAT was previously known to be present in human new borns and rodents, but interest in BAT has rapidly increased since active BAT depots were visualized in human adults. Cold activated metabolically active adipose tissue was shown by 18 F fluorodeodeoxy glucose positron emission tomography computed tomography ([ 18 F]FDG PET/CT) imaging, mostly in the supraclavicular region [4 6]. Furthermore, repeated cold exposure as well as a single oral dose of mirabegron, a β 3 adrenergic receptor agonist, increases the uptake of [ 18 F]FDG in the supraclavicular region [4, 7 9]. Mouse models with genetically decreased BAT activity often show a cold sensitive but also an obesogenic phenotype [1]. Examples include UCP1 knockout mice, leptin deficient ob/ob mice and mice that lack all three β adrenergic receptor knockout mice [11 13]. White adipose tissue (WAT) stores energy, whereas active BAT increases energy expenditure. Intriguingly, BAT activity, measured by [ 18 F]FDG uptake, is inversely correlated with body mass index [4]. Since obesity is one of the most prevalent chronic disorders worldwide and increases the risk of a variety of comorbidities including insulin resistance and cardiovascular disease, it could be therapeutically beneficial to increase the amount and activity of BAT, which would result in increased energy expenditure and reduce body fat. Adipose tissue contains a variety of immune cells, including macrophages, eosinophils and T cells, which contribute to tissue homeostasis [14]. In obesity, hypertrophy of white adipocytes leads to hypoxia, cell death and tissue dysfunction. This causes recruitment of a variety of immune cells and interferes with tissue homeostasis [15]. Inflammation is a key feature of obesity and insulin resistance [16]. The recruitment of pro inflammatory macrophages is a critical event in obese WAT inflammation [14, 17, 18]. Macrophages in lean healthy WAT mainly exhibit an anti inflammatory phenotype, whereas in obese WAT, the number of pro inflammatory classically activated macrophages increases [14]. Adipocytes as well as immune cells in obese WAT produce proinflammatory cytokines, including TNF, IL 6, CCL2 (Mcp 1), CCL4 (Mip 1b) and CXCL12 (SDF 1), which further propagate the ongoing inflammatory response, both within the tissue and systemically [19 23]. Whether and how immune cells play a role in (obese) BAT is largely unknown [1]. Similar to WAT, obese BAT shows decreased insulin signalling and increased gene expression of the macrophage marker Adgre1 (encoding F4/8), and gene expression of cytokines Tnf, Il6 and Ccl2 increases in obese BAT in mice [21, 24, 25]. The pro inflammatory cytokine TNF suppresses UCP1 in brown 51

54 Chapter 3 adipocytes, suggesting a decreased activity of BAT in an obese environment [25 28]. Furthermore, anti inflammatory macrophages were shown to contribute to BAT activation as cold exposure increased macrophage markers in both WAT and BAT in mice, and absence of anti inflammatory macrophages blunted BAT activity and induced cold intolerance [29]. However, subsequent studies showed no effect on adaptive thermogenesis upon depletion of macrophages in BAT [3]. As BAT activation can be beneficial in combatting obesity, it is critical to understand what underlies BAT function in an obese state. Here, we determined whether and when the immune system in BAT is affected by a high fat diet (HFD), by studying a time course of HFD in which 8 groups of mice received a HFD ranging from 1 day to 18 weeks. We studied short term and long term immunological adipose tissue changes upon a HFD and compared WAT and BAT. Additionally, we tested brown adipocyte functioning in vitro under inflammatory conditions. Materials and methods Mice Age matched male wild type C57Bl/6 mice (Charles River) received different durations of HFD (45% kcal fat, 35% kcal carbohydrate, 2% kcal protein, Special Diets Services, Witham, United Kingdom). All mice were included in the experiments at the age of 7 weeks and were given a HFD in a time course ranging from 1 day up to 18 weeks before being sacrificed at the age of 25 weeks (8 groups of n=1 11). Mice had ad libitum access to food and water and were maintained under a 12 hour lightdark cycle. Mice were fasted overnight, weighed and subsequently euthanized using.25 mg/g ketamine and.5 mg/g xylazine. Glucose levels were measured from whole blood using a glucometer (Bayercontour, Basel, Switzerland). Blood was obtained by cardiac puncture using EDTAfilled syringes and liver, spleen, interscapular BAT, epididymal WAT (EpAT) and subcutaneous WAT (ScAT) were dissected and weighed. All experimental procedures were approved by the Animal Experimentation Ethics Committee of the University of Amsterdam. Histology Tissues were collected, fixed in 4% paraformaldehyde and embedded in paraffin. The area of intracellular lipid vacuoles in BAT was quantified on a haematoxylin and eosin staining using ImageJ software (NIH, USA). Immunohistochemistry on EpAT and ScAT was performed for CD45 (clone 3F11, BD Biosciences, Breda, the Netherlands). Frozen BAT sections were stained for CD68 (clone FA 11, Bio Rad, Veenendaal, the Netherlands) and quantified in ImageJ (NIH, USA). Real time PCR Total RNA was extracted using TRIzol (Invitrogen, Carlsbad, CA, USA) and reverse transcribed with an iscript cdna synthesis kit (Bio Rad, Veenendaal, the Netherlands). The quantitative PCR was performed using a SYBR green PCR kit and a ViiA7 RT PCR system (Applied Biosystems, Leusden, the Netherlands). Expression was normalized to the housekeeping genes Rplp, β Actin and cycloa. The results are expressed as relative to the control group, which was assigned a value of 1. Culture and differentiation of brown adipocytes T37i cells were kindly provided by Marc Lombès [31] and cultured in DMEM F12 Glutamax, 1% FCS, penicillin (1 U/ml) and streptomycin (1 µg/ml) (ThermoFisher Scientific, Waltham, MA, USA). 52

55 Brown adipose tissue in obesity After growing confluent, cells were differentiated by adding 2 nm Triiodothyronine (T3) (Sigma Aldrich, Zwijndrecht, the Netherlands) and 112 ng/ml insulin (Sigma Aldrich) to the media. After 9 days of differentiation, the cells were stimulated with 1 ng/ml TNF (Invitrogen), 1 U/ml IFNγ (Peprotech, Rocky Hill, NJ, USA) or 2 ng/ml LPS (Sigma) for 24 hours. Supernatant was collected for ELISA and cells were harvested for RNA analysis. Transwell migration assay RAW264.7 macrophages (5x1 4 ) were seeded in a 24 well transwell migration assay with 8 µm pores (Sigma Aldrich) and incubated for 3 hours with supernatant from differentiated T37i brown adipocytes. Macrophages were fixed with 4% formaldehyde, washed and stained with 5% toluidine blue. Non migrated cells were swiped with a cotton tip and migrated cells counted in ImageJ (NIH, USA). ELISA IL 6 (Invitrogen) and CXCL1 (R&D systems, Minneapolis, MN, USA) were quantified by ELISA in accordance to the suppliers protocols. Statistics Results are presented as mean ± SEM. Analysis between more than two groups in the time course experiments were done by a one way ANOVA with Tukey post test analysis and the change per group was expressed relative to the control group. Analysis between two groups was done by a Student s t test. Statistics were calculated in GraphPad Prism 5. (GraphPad Software, Inc., La Jolla, CA, USA). P values <.5 were considered significant. Results Adipose tissue changes in a HFD time course Eight groups of mice were fed a HFD for different durations in time (Figure 1A). Body weight increased after two weeks of HFD and continued to increase throughout the experiment, without major changes in liver or spleen weight (Table 1). Blood glucose levels already increased after 1 day of HFD and remained high in all HFD groups (Table 1). Table 1. General characteristics of mice in the course of DIO Body weight (g) Glucose (mg/dl) Weeks of HFD Mean ± SEM (p) a 1d 3d 1w 2w 4w 1w 18w 27.3 ±.6 69 ± 3 Liver (g) 1.21 ± ±.6 13 ± 15 (***) 1.3 ± ± ± 7 (**) 1.32 ± ± ± 4 (***) 1.24 ± ±.9 (***) 14 ± 5 (*) 1.22 ± ± 1.2 (***) 91 ± ± ±.9 (***) 112 ± 9 (**) 1.11 ± ± 2. (***) 124 ± 7 (***) 1.39 ±.9 Spleen (mg) 83.6 ± ± ± ± ± ± ± ± 5.1 a Data in this table is analysed by a one way ANOVA with Tukey post test analysis and the change per group expressed relative to the weeks of HFD group. n=1 11, *=P<.5, **=P<.1, ***=P<.1. 53

56 Chapter 3 A Chow diet High-fat diet 18w 12w 4w 2w 1w 3d 1d B BAT C BAT gram ** *** % lipid droplets *** *** * * **. 1d 3d 1w 2w 4w 1w 18w HFD 1d 3d 1w 2w 4w 1w 18w HFD D EpAT E ScAT gram *** *** *** *** gram ** *** *** ***.5 1d 3d 1w 2w 4w 1w 18w HFD. 1d 3d 1w 2w 4w 1w 18w HFD Figure 1. Lipid uptake in BAT occurs early in time after initiation of high fat diet feeding. A. Study outline. Eight different groups of mice were subjected to a 45% HFD for different times (1 day (d) to 18 weeks (w)) and sacrificed at the same age. B. Interscapular BAT was removed after sacrifice and its weight was determined. C. The percentage of lipid as measured on an H&E stained BAT section. D. EpAT and (E) ScAT were removed after sacrifice and weighed. n=1 11, *=P<.5, **=P<.1, ***=P<.1 (1 way ANOVA with Tukey post test analysis). Upon HFD feeding, an increase in BAT weight was already present after 3 days of HFD (Figure 1B). Histological analysis showed that the lipid content in brown adipocytes rapidly increased and plateaued after 3 days of HFD and remained high until 18 weeks of HFD (Figure 1C), indicating that BAT accumulates lipids very rapidly. In contrast, the increase in EpAT and ScAT weight reached significance after 2 weeks of HFD and kept rising during the entire duration of the experiment (Figure 1D and E). 54

57 Brown adipose tissue in obesity Obesity associated inflammation in BAT Obese WAT is associated with leukocyte infiltration with a prominent role for macrophages, as the number of AT macrophages correlates with the extent of metabolic dysfunction [14, 18]. In BAT, we observed rapid changes in lipid uptake and increased brown adipocyte size after the start of HFD, and we wondered whether macrophages also infiltrate BAT in the course of diet induced obesity (DIO). We histologically analysed BAT macrophages in situ by CD68 staining and observed increased macrophage counts again already after three days upon induction of DIO (Figure 2A,B). Gene expression of the macrophage specific marker F4/8 also increased rapidly and reached significance after 1 week of HFD (Figure 2C). Additional markers for classical pro inflammatory (Nos2) vs antiinflammatory (CD31, Arg1) macrophages hinted towards an increase in pro inflammatory macrophages rather than in anti inflammatory macrophages (Figure 2C). A 1 μm B.3 BAT macrophages *** # CD68 + /adipocyte.2.1 * ** * *. 1d 3d 1w 2w 4w 1w 18w HFD C Relative expression *** * ** * * 1d 3d 1w 2w 4w 1w 18w. Adgre1 Nos2 Clec1a Arg1 D + # CD45 /adipocyte EpAT leukocytes *** E + # CD45 /adipocyte * ScAT leukocytes. 1d 3d 1w 2w 4w 1w 18w HFD. 1d 3d 1w 2w 4w 1w 18w HFD Figure 2. Increased macrophage numbers in obese BAT. A. Representative image of a CD68 + macrophage immunohistochemical staining on frozen BAT sections. B. Quantified immunohistochemical staining of relative to the number of adipocytes. C. Gene expression analysis of BAT showing the macrophage gene Adgre1 (encoding F4/8) and in the pro inflammatory macrophage marker Nos2 (inos) and anti inflammatory macrophage marker genes Clec1a (CD31/Mgl1) and Arg1. D. CD45 + leukocyte counts relative to the number of adipocytes in EpAT after 18 weeks of HFD counted on paraffin immunohistochemical stainings and (E) leukocyte count upon a HFD in ScAT. n=1 11, *=P<.5, **=P<.1, ***=P<.1 (1 way ANOVA with Tukey post test analysis). 55

58 Chapter 3 Total leukocytes (CD45 + cells) in EpAT increased with more established obesity (Figure 2D). Furthermore, we observed that EpAT was more prone to obesity associated inflammation than ScAT; EpAT displays an increase in leukocytes after 18 weeks of HFD (Figure 2D), whereas ScAT leukocyte counts did not increase within this time frame (Figure 2E). Increased cytokine expression in BAT upon HFD In obese WAT, adipocytes and immune cells secrete a variety of cytokines that recruit inflammatory cells and propagate the inflammatory profile of the tissue [15]. To study whether factors secreted by brown adipocytes can affect macrophage migration, we performed an in vitro macrophage migration assay in which RAW264.7 macrophages were seeded on a transwell membrane and incubated with supernatant from a differentiated brown adipocyte cell line (T37i). Interestingly, macrophage migration massively increased in the presence of T37i supernatant compared to the spontaneous migration using medium only (Figure 3A), suggesting that chemoattractants are secreted by brown adipocytes. We then studied alterations in cytokine and chemokine expression in BAT in vivo and found that the expression of cytokines occurs simultaneously with the increase in lipid uptake and macrophage numbers in BAT. Gene expression of Tgfb1, Il6, Tnf, Ccl2, Ccl4, Ccl7, Cxcl1 and Cx3cl1 was increased after 18 weeks of HFD, with a rapid increase already starting after 3 days of HFD for most mediators (Figure 3B I). The increase in pro inflammatory cytokines, including Tnf, was relatively higher than the increase in macrophage F4/8 expression, again suggesting a shift to a more pro inflammatory phenotype of the macrophages. After 18 weeks of HFD we found large increases in gene expression of the cytokines Tnf, Ccl2, Ccl4 and Ccl5 in EpAT. Interestingly, relative to the gene expression in EpAT of chow fed mice, obese BAT also displayed an inflammatory cytokine environment in which Tnf, Ccl2 and Ccl4 increased to the same extent as in obese EpAT, whereas Il6 and Cx3cl1 specifically increased in BAT and Ccl5 was specifically induced in EpAT (Figure 3J O). Brown adipocytes are regulated by pro inflammatory cytokines in vitro We next studied the effects of inflammatory stimuli on brown adipocyte function in vitro. A brown adipocyte cell line (T37i) was stimulated for 24 hours with pro inflammatory cytokines TNF and IFNγ, and lipopolysaccharide (LPS), which is a surface membrane component of gram negative bacteria and allows cells to recognize bacterial invasion via Toll like receptor 4 (TLR4) (Figure 4A). All stimuli lowered Ucp1 gene expression, which suggests inhibition of thermogenic function (Figure 4B). Furthermore, the expression of Ppargc1a, a gene regulating mitochondrial biogenesis, was decreased after TNF and LPS stimulation (Figure 4B). The cytokine stimulations TNF and IFNγ also lowered the expression of the fatty acid receptor Cd36 and all stimulations decreased gene expression of lipoprotein lipase (Lpl), suggesting that pro inflammatory stimuli also inhibit fatty acid uptake by the cells (Figure 4B). The HFD rapidly increased cytokine and chemokine gene expression in BAT (Figure 3) but it is unclear which cell type is responsible for this production. Although macrophages as immune cells are key candidates to produce these inflammatory mediators we also examined expression of these genes in brown adipocytes in vitro. In the brown adipocytes that were stimulated with pro inflammatory cytokines TNF and IFNγ, and LPS, we analysed gene expression of a set of relevant chemokines. Interestingly, all three conditions were able to induce a variety of chemokines, including Il6, Ccl2, Ccl3, Ccl4, Ccl5, Ccl7, Cxcl1, Cxcl12 and Cx3cl1 (Figure 4C,DI). Even more intriguing, IL 6 and CXCL1 were also secreted in the supernatant in nanogram quantities, confirming secretion of relevant levels 56

59 Brown adipose tissue in obesity of these inflammatory mediators by brown adipocytes (Figure 4E). Altogether, the brown adipocytes themselves likely also contribute to the inflammatory signals within BAT upon a HFD challenge. A B C # migrated RAW cells Control **** T37i Supernatant Relative expression Tgfb1 * ** * 1d 3d 1w 2w 4w 1w 18w Relative expression Tnf ** 1d 3d 1w 2w 4w 1w 18w D Il6 E Ccl2 F Ccl4 Relative expression d 3d 1w 2w 4w 1w 18w * Relative expression *** ** 1d 3d 1w 2w 4w 1w 18w Relative expression * ** 1d 3d 1w 2w 4w 1w 18w G Ccl7 H Cxcl1 I Cx3cl1 Relative expression ** ** 1d 3d 1w 2w 4w 1w 18w Relative expression *** * 1d 3d 1w 2w 4w 1w 18w Relative expression *** *** *** *** *** 1d 3d 1w 2w 4w 1w 18w J Relative expression Tnf *** ** Chow 18w HFD Chow 18w HFD K Relative expression Il6 * * Chow 18w HFD Chow 18w HFD Relative expression L Ccl2 * ** *** ** Chow 18w HFD Chow 18w HFD EpAT BAT M Relative expression Ccl4 ** ** Chow 18w HFD Chow 18w HFD N Relative expression Ccl5 ** ** ** Chow 18w HFD Chow 18w HFD Relative expression O Cx3cl1 * *** Chow 18w HFD Chow 18w HFD Figure 3. Increased cytokine expression in obese BAT. A. The migration of RAW264.7 macrophages in vitro when incubated with supernatant from a brown adipocyte cell line (T37i). B. BAT gene expression of Tgfb1, (C) Il6, (D) Tnf, (E) Ccl2, (F) Ccl4, (G) Ccl7, (H) Cxcl1 and (I) Cx3cl1. J O. Comparison of gene expression in EpAT and BAT, expression was calculated relative to the chow EpAT group for (J) Tnf, (K) Il 6, (L) Ccl2, (M) Ccl4, (N) Ccl5 and (O) Cx3cl1. n=1 11, *=P<.5, **=P<.1, ***=P<.1 (A, J O Student t test, B I 1 way ANOVA with Tukey post test analysis). 57

60 Chapter 3 A TNF 24h B IFNγ Differentiated T37i brown adipocytes LPS Relative expression TNF IFN 1. LPS * ** ** ** ** *** ** ** *.5 ***. Ucp1 Ppargc1a Cd36 Lpl C Relative expression *** *** *** * * Ccl2 Ccl3 Ccl4 Ccl5 Ccl7 *** *** *** *** *** *** *** *** Relative expression 4 - TNF *** IFN 3 LPS *** 2 1 Cxcl12 * Cx3cl1 ** D Relative expression *** TNF 6 *** IFN 12 *** *** LPS *** *** Il6 Cxcl1 E ng/ml *** IL-6 *** *** *** * CXCL1 - TNF IFN LPS Figure 4. Inflammatory stimuli reduce Ucp1 expression and increase chemokine secretion by brown adipocytes in vitro A. In vitro model of brown adipocyte cell line (T37i) stimulations. B. Gene expression of Ucp1, Ppargc1a, Cd36 and Lpl after 24 hours of stimulation with TNF, IFNγ or LPS. C, D. Gene expression of Ccl2, Ccl3, Ccl4, Ccl5, Ccl7, Cxcl12, Cx3cl1, Il6 and Cxcl1 upon pro inflammatory stimuli. E. ELISA measurements of the supernatant showing secretion of IL 6 and CXCL1 upon pro inflammatory stimuli. n=4, *=P<.5, **=P<.1, ***=P<.1 (Students t test). 58

61 Brown adipose tissue in obesity Discussion Obesity is associated with a chronic state of inflammation in WAT and a reduction in BAT function. Our study shows that BAT of mice subjected to diet induced obesity undergoes tissue expansion within 3 days of HFD. A rapid increase in BAT weight and lipid droplet size after the start of a HFD suggests that BAT is an extremely plastic tissue that adapts to changes in dietary conditions. These increases in BAT weight and lipid content plateaued after 3 days of dietary challenge and was only further increased in the 18 weeks HFD group. Within the first week of HFD, the expanding BAT attracted immune cells associated with increased cytokine and chemokine gene expression, although the highest levels of inflammation were measured in the 18 weeks of HFD group. In vitro data using a brown adipocyte cell line showed that a pro inflammatory environment decreases Ucp1 expression in brown adipocytes and that brown adipocytes themselves are capable of producing and secreting inflammatory cytokines and chemokines, and increases macrophage migration. Within the expanding BAT we observed an increase in macrophages, with a more pro inflammatory phenotype as evidenced by higher expression of Nos2 and increased inflammatory cytokines including Tnf and Il6. Inflammatory stimuli also enhance the release of TNF and IL 6 from proinflammatory macrophages. However, in adipose tissue, IL 6 is not necessarily pro inflammatory and may mediate anti inflammatory macrophage polarization by inducing the IL 4 receptor [32]. We also observed increased expression of Ccl2, Ccl4, and Ccl7, which are chemotactic signals for monocytes and macrophages. Cxcl1, a neutrophil chemokine, and especially Cx3cl1 also rapidly increased in BAT after the start of HFD. CX3CL1, also known as fractalkine, recruits monocytes and macrophages and loss of CX3CL1 CX3CR1 signalling was previously shown to reduce macrophage accumulation in WAT as well as BAT and to reduces development of DIO [33]. We show that pro inflammatory signals associated with obesity affects brown adipocytes. By mimicking a pro inflammatory environment in an in vitro model in which brown adipocytes were stimulated with TNF, IFNγ and LPS we found decreased Ucp1 and Ppargc1a expression. This is in line with reports by others who show that TNF suppresses Ucp1 expression and impairs mitochondrial biogenesis in brown adipocytes [25 28]. These data thus suggest decreased brown adipocyte activity upon inflammatory signals, which could be a mechanism behind the well established obesityassociated decreased BAT activity. White adipocytes secrete many adipokines, including leptin and adiponectin but also (adipo ) cytokines, including IL 6 and CCL2 [34]. Furthermore, under inflammatory (i.e. LPS stimulated) conditions, white adipocytes secrete TNF, IL 6, CCL3, CCL4 and CXCL12 [35]. Brown adipocytes secrete BATokines as well, which can act in a paracrine or autocrine manner [36]. For example, thermogenic active brown and/or beige adipocytes secrete factors that promote hypertrophy, hyperplasia, vascularization and innervation including irisin, metrnl, FGF21, neuregulin 4 and also IL 6 [36]. Interestingly, we observed that, especially upon pro inflammatory stimuli in vitro, brown adipocytes themselves express a variety of cytokines and chemokines including Il 6, Ccl2, Ccl3, Ccl4, Ccl5, Ccl7, Cxcl1, Cxcl12 and Cx3cl1 and probably others. These data indicate that the brown adipocytes themselves may be an important source of inflammatory signals within BAT upon a HFD challenge. 59

62 Chapter 3 Our study further shows that BAT stores excessive lipids, in which the accumulation of lipids gives the tissue a whitened phenotype. Others have shown that whitening of BAT in mice is characterized by hypoxia, mitochondrial dysfunction and decreased SNS innervation [37]. One mechanism contributing to immune regulation in BAT could be the SNS. Surgical denervation of BAT and treatment with the β adrenergic antagonist propranolol upregulate gene expression of Tnf, ll6 and Ccl2, without a change in F4/8 expression. This shows that the SNS is involved in maintaining the anti inflammatory state in BAT and decreased SNS signalling disturbs the anti inflammatory environment [38]. Future research should further examine the contribution of the SNS to immune regulation of obese BAT. In mice, cold exposure increases markers for anti inflammatory macrophages in BAT and beiging in ScAT, and macrophage depletion decreases BAT activity. Anti inflammatory macrophages are thus important stimuli of adaptive thermogenesis. They release noradrenaline, via the production of tyrosine hydroxylase, a catecholamine synthesizing enzyme, and IL 4 STAT6 signalling [29, 3]. Not only brown adipocytes express β adrenergic receptors; macrophages also express β 2 adrenergic receptors, enabling a complex auto and paracrine loop [39, 4]. Although literature states an important link between anti inflammatory macrophages and BAT functioning, our data does not support a loss of anti inflammatory macrophages in obese BAT but rather an increase in proinflammatory macrophages. This would imply that a disturbed mechanism of BAT activation by antiinflammatory macrophages is not the major cause of obesity associated dysfunctional BAT. Insulin resistance that accompanies obesity is partly attributable to changes in adipokine secretion by WAT. BAT is highly vascularized to supply the adipocytes with substrate and oxygen for oxidation and enable efficient distribution of heat [1]. Therefore, the cytokines and chemokines produced by brown adipocytes in obesity might be transported out of the tissue and thereby also contribute to systemic inflammation. However, it will be challenging to determine the relative contribution of BAT to increased circulating cytokine levels and insulin resistance. We found many inflammatory signals and a rapid increase in macrophages in BAT of HFD fed mice, although we cannot exclude the possibility that other immune cell types play a role as well. Our study shows that brown adipocytes produce inflammatory mediators, which likely mediate crosstalk between adipocytes and macrophages. Future studies will be needed to elucidate on the complete spectrum of BAT regulation by immune cells. Furthermore, the contribution of immune cells to differentiation, proliferation and activation of beige adipocytes within WAT and possible overlapping mechanisms with classical BAT will be of great interest. Although dysfunctional BAT in obesity likely results from combined physiological alterations at multiple sites, we have shown that the immunological signals are one important regulatory factor. 6

63 Brown adipose tissue in obesity Corresponding author Prof. Menno P.J. de Winther. Department of Medical Biochemistry, Experimental Vascular Biology, Academic Medical Center, University of Amsterdam, Meibergdreef 9, 115 AZ, Amsterdam, The Netherlands. +31 () m.dewinther@amc.uva.nl Conflict of interest The authors declare no conflict of interest. Acknowledgements We acknowledge the support from the Rembrandt Institute of Cardiovascular Science (PR, MdW, EL) and the Netherlands CardioVascular Research Initiative: the Dutch Heart Foundation, Dutch Federation of University Medical Centres, the Netherlands Organisation for Health Research and Development and the Royal Netherlands Academy of Sciences" for the GENIUS project Generating the best evidence based pharmaceutical targets for atherosclerosis (CVON211 19). This work was supported by the Netherlands Organization for Scientific Research (NWO) (VENI to JVdV, VICI grant to EL), the Netherlands Heart Foundation (Dr E. Dekker grant to TS and junior postdoc grant to JVdB), the European Research Council (ERC con grant to EL) and Horizon 22 (REPROGRAM to EL and MdW). 61

64 Chapter 3 References [1] Kajimura S, Spiegelman Bruce M, Seale P Brown and Beige Fat: Physiological Roles beyond Heat Generation. Cell Metabolism 22(4): [2] Cannon B, Nedergaard J. 24. Brown Adipose Tissue: Function and Physiological Significance. Physiological Reviews 84(1): [3] Kooijman S, van den Heuvel JK, Rensen PCN Neuronal Control of Brown Fat Activity. Trends in Endocrinology & Metabolism 26(11): [4] Cypess AM, Lehman S, Williams G, Tal I, Rodman D, Goldfine AB et al. 29. Identification and Importance of Brown Adipose Tissue in Adult Humans. New England Journal of Medicine 36(15): [5] Saito M, Okamatsu Ogura Y, Matsushita M, Watanabe K, Yoneshiro T, Nio Kobayashi J et al. 29. High Incidence of Metabolically Active Brown Adipose Tissue in Healthy Adult Humans: Effects of Cold Exposure and Adiposity. Diabetes 58(7): [6] Nedergaard J, Bengtsson T, Cannon B. 27. Unexpected evidence for active brown adipose tissue in adult humans. American Journal of Physiology Endocrinology and Metabolism 293(2):E444 E452. [7] van Marken Lichtenbelt WD, Vanhommerig JW, Smulders NM, Drossaerts JMAFL, Kemerink GJ, Bouvy ND et al. 29. Cold Activated Brown Adipose Tissue in Healthy Men. New England Journal of Medicine 36(15): [8] Virtanen KA, Lidell ME, Orava J, Heglind M, Westergren R, Niemi T et al. 29. Functional Brown Adipose Tissue in Healthy Adults. New England Journal of Medicine 36(15): [9] Cypess AM, Weiner LS, Roberts Toler C, Elía EF, Kessler SH, Kahn PA et al Activation of Human Brown Adipose Tissue by a β3 Adrenergic Receptor Agonist. Cell Metabolism 21(1): [1] van den Berg SM, van Dam AD, Rensen PCN, de Winther MPJ, Lutgens E Immune modulation of brown(ing) adipose tissue in obesity. Endocrine Reviews:er [11] Feldmann HM, Golozoubova V, Cannon B, Nedergaard J. 29. UCP1 Ablation Induces Obesity and Abolishes Diet Induced Thermogenesis in Mice Exempt from Thermal Stress by Living at Thermoneutrality. Cell Metabolism 9(2): [12] Bachman ES, Dhillon H, Zhang C Y, Cinti S, Bianco AC, Kobilka BK et al. 22. β AR Signaling Required for Diet Induced Thermogenesis and Obesity Resistance. Science 297(5582): [13] Ukropec J, Anunciado R, Ravussin Y, Kozak L. 26. Leptin Is Required for Uncoupling Protein 1 Independent Thermogenesis during Cold Stress. Endocrinology 147(5): [14] Chawla A, Nguyen KD, Goh YPS Macrophage mediated inflammation in metabolic disease. Nature Reviews Immunology 11(11): [15] Lumeng CN, Saltiel AR Inflammatory links between obesity and metabolic disease. Journal of Clinical Investigation 121(6): [16] Hotamisligil GS. 26. Inflammation and metabolic disorders. Nature 444(7121): [17] Weisberg SP, McCann D, Desai M, Rosenbaum M, Leibel RL, Ferrante AW, Jr. 23. Obesity is associated with macrophage accumulation in adipose tissue. The Journal of Clinical Investigation 112(12): [18] Osborn O, Olefsky JM The cellular and signaling networks linking the immune system and metabolism in disease. Nat Med 18(3): [19] Hotamisligil G, Shargill N, Spiegelman B Adipose expression of tumor necrosis factor alpha: direct role in obesity linked insulin resistance. Science 259(591): [2] Pradhan AD, Manson JE, Rifai N, Buring JE, Ridker PM. 21. C reactive protein, interleukin 6, and risk of developing type 2 diabetes mellitus. JAMA 286(3): [21] Kanda H, Tateya S, Tamori Y, Kotani K, Hiasa K i, Kitazawa R et al. 26. MCP 1 contributes to macrophage infiltration into adipose tissue, insulin resistance, and hepatic steatosis in obesity. The Journal of Clinical Investigation 116(6): [22] Ognjanovic S, Jacobs DR, Steinberger J, Moran A, Sinaiko AR Relation of chemokines to BMI and insulin resistance at ages Int J Obes 37(3): [23] Kim D, Kim J, Yoon JH, Ghim J, Yea K, Song P et al CXCL12 secreted from adipose tissue recruits macrophages and induces insulin resistance in mice. Diabetologia 57(7): [24] Roberts Toler C, O'Neill BT, Cypess AM Diet induced obesity causes insulin resistance in mouse brown adipose tissue. Obesity 23(9): [25] Sakamoto T, Nitta T, Maruno K, Yeh Y S, Kuwata H, Tomita K et al Macrophage infiltration into obese adipose tissues suppresses the induction of UCP1 level in mice. American Journal of Physiology Endocrinology And Metabolism 31(8):E676. [26] Nisoli E, Briscini L, Giordano A, Tonello C, Wiesbrock SM, Uysal KT et al. 2. Tumor necrosis factor α mediates apoptosis of brown adipocytes and defective brown adipocyte function in obesity. Proceedings of the National Academy of Sciences 97(14): [27] Valladares A, Roncero C, Benito M, Porras A. 21. TNF α inhibits UCP 1 expression in brown adipocytes via ERKs: Opposite effect of p38mapk. FEBS Letters 493(1):

65 Brown adipose tissue in obesity [28] Sakamoto T, Takahashi N, Sawaragi Y, Naknukool S, Yu R, Goto T et al Inflammation induced by RAW macrophages suppresses UCP1 mrna induction via ERK activation in 1T1/2 adipocytes. The American Journal of Physiology Cell Physiology 34(8):C729 C738. [29] Nguyen KD, Qiu Y, Cui X, Goh YPS, Mwangi J, David T et al Alternatively activated macrophages produce catecholamines to sustain adaptive thermogenesis. Nature 48(7375): [3] Qiu Y, Nguyen Khoa D, Odegaard Justin I, Cui X, Tian X, Locksley Richard M et al Eosinophils and Type 2 Cytokine Signaling in Macrophages Orchestrate Development of Functional Beige Fat. Cell 157(6): [31] Zennaro MC, Le Menuet D, Viengchareun S, Walker F, Ricquier D, Lombès M Hibernoma development in transgenic mice identifies brown adipose tissue as a novel target of aldosterone action. The Journal of Clinical Investigation 11(6): [32] Mauer J, Chaurasia B, Goldau J, Vogt MC, Ruud J, Nguyen KD et al Signaling by IL 6 promotes alternative activation of macrophages to limit endotoxemia and obesity associated resistance to insulin. Nature Immunology 15(5): [33] Polyák Á, Winkler Z, Kuti D, Ferenczi S, Kovács KJ Brown adipose tissue in obesity: Fractalkine receptor dependent immune cell recruitment affects metabolic related gene expression. Biochimica et Biophysica Acta (BBA) Molecular and Cell Biology of Lipids 1861(11): [34] Ouchi N, Parker JL, Lugus JJ, Walsh K Adipokines in inflammation and metabolic disease. Nature Reviews Immunology 11(2): [35] Chirumbolo S, Franceschetti G, Zoico E, Bambace C, Cominacini L, Zamboni M LPS response pattern of inflammatory adipokines in an in vitro 3T3 L1 murine adipocyte model. Inflammation Research 63(6): [36] Villarroya F, Cereijo R, Villarroya J, Giralt M Brown adipose tissue as a secretory organ. Nat Rev Endocrinol advance online publication. [37] Shimizu I, Aprahamian T, Kikuchi R, Shimizu A, Papanicolaou KN, MacLauchlan S et al Vascular rarefaction mediates whitening of brown fat in obesity. The Journal of Clinical Investigation 124(5): [38] Tang L, Okamoto S, Shiuchi T, Toda C, Takagi K, Sato T et al Sympathetic Nerve Activity Maintains an Anti Inflammatory State in Adipose Tissue in Male Mice by Inhibiting TNF α Gene Expression in Macrophages. Endocrinology 156(1): [39] Stanojević S, Dimitrijević M, Kuštrimović N, Mitić K, Vujić V, Leposavić G Adrenal hormone deprivation affects macrophage catecholamine metabolism and β2 adrenoceptor density, but not propranolol stimulation of tumour necrosis factor α production. Experimental Physiology 98(3): [4] Grailer JJ, Haggadone MD, Sarma JV, Zetoune FS, Ward PA Induction of M2 Regulatory Macrophages through the β2 Adrenergic Receptor with Protection during Endotoxemia and Acute Lung Injury. Journal of Innate Immunity 6(5):

66 Chapter 3 64

67 4 Helminth antigens counteract a rapid high fat diet induced drop in eosinophils in a depot specific manner in mice Susan M. van den Berg 1, Andrea D. van Dam 2,3, Pascal J.H. Kusters 1, Linda Beckers 1, Myrthe den Toom 1, Saskia van der Velden 1, Jan Van den Bossche 1, Irma van Die 4, Mariëtte R. Boon 2,3, Patrick C.N. Rensen 2,3, Esther Lutgens *,1,5, Menno P.J. de Winther *,1,5 1 Department of Medical Biochemistry, Experimental Vascular Biology, Academic Medical Center, University of Amsterdam, the Netherlands 2 Department of Medicine, Division Endocrinology, Leiden University Medical Center, Leiden, the Netherlands 3 Einthoven Laboratory for Experimental Vascular Medicine, Leiden, the Netherlands 4 Department of Molecular Cell Biology and Immunology, Neuroscience Campus Amsterdam, Vrije Universiteit Medical Center, Amsterdam, the Netherlands 5 Institute for Cardiovascular Prevention (IPEK), Ludwig Maximilian s University, Munich, Germany *these authors contributed equally to this work. Submitted

68 Chapter 4 Abstract Brown adipose tissue (BAT) activation or white adipose tissue (WAT) beiging can increase energy expenditure and reduce obesity associated diseases. The immune system is a potential target in mediating brown and beige adipocyte activation. Homeostasis of lean WAT is maintained by type 2 and anti inflammatory immune cells, in which eosinophils produce interleukin (IL) 4 and sustain antiinflammatory macrophage activation. In obesity, WAT shows a decreased number of eosinophils and an increased infiltration of pro inflammatory immune cells. We studied eosinophil numbers in BAT, epididymal WAT (EpAT) and subcutaneous WAT (ScAT) after 1 day, 3 days or 1 week of high fat diet (HFD) in C57Bl/6 mice. Helminth antigens induce a type 2 immune response and can be protective against metabolic defects in obesity. We hypothesized that this phenotype involves activation of BAT or beige adipogenesis. Therefore, 1 week HFD fed mice were injected with helminth antigens from Schistosoma mansoni and Trichuris suis. After 1 day of HFD, eosinophils started to decline in EpAT and BAT, and after 3 days in ScAT, along with a decrease in the eosinophil chemoattractant Ccl3 in EpAT and ScAT and Ccl5 and Ccl11 in BAT. Furthermore, macrophages increased after 3 days of HFD in EpAT. Interestingly, IL 4 stimulation of brown adipocytes in vitro resulted in increased Ucp1 expression and the production of CCL11 (26 fold increase, p<.1). Treatment with helminth antigens resulted in high numbers of eosinophils, macrophages and T cells in EpAT. However, there were almost no effects of the helminth antigens on ScAT and BAT immune cells and no activation of BAT or beiging of WAT. Adipose tissue eosinophils decline rapidly after starting a HFD in mice. Shortterm treatment with helminth antigens results in an adipose depot specific type 2 immune response which does not affect BAT activation or beiging of WAT. 66

69 Type 2 immunological effects of HFD and helminth antigens on adipose tissue Introduction Brown adipose tissue (BAT) contributes to the control of body temperature, by the production of heat in response to cold. Brown adipocytes have numerous mitochondria that express uncoupling protein 1 (UCP1). Activation of UCP1 by noradrenaline results in uncoupling of the respiratory chain, bypassing ATP synthesis and the dissipation of energy into heat [1]. Rodents as well as new borns from larger organisms, including humans, have interscapular BAT depots. In humans, BAT regresses with age but can still be found supraclavicular and in the neck region, but also periaortic, paravertebral and perirenal [2]. BAT activity can be detected using a positron emission tomographic and X ray computed tomography (PET/CT) scan and 18 F fluorodeoxy glucose ([ 18 F]FDG) uptake, in which cold exposure increases the uptake of glucose [3]. Interestingly, [ 18 F]FDG uptake is negatively correlated with BMI, suggestive of decreased BAT activity in obesity [3]. Brown like or beige/brite adipocytes can also be present within white adipose tissue (WAT) [4]. These beige adipocytes display similar characteristics as classical brown adipocytes, including a multilocular appearance and Ucp1 expression. They are present at low quantities but can be induced by cold or β 3 adrenergic receptor agonists. The inguinal subcutaneous adipose tissue (ScAT) depot has the highest beiging potential after cold exposure [5], although the thermogenic capacity is only 1% of that of interscapular BAT [6]. Activation of BAT and the induction of beige adipocytes, beneficially increases energy expenditure and therefore has the potential to reduce excessive energy stores in obesity [7]. Homeostasis of adipose tissue is maintained by immune cells, in which lean WAT and BAT are characterized by an anti inflammatory immune cell composition [8]. The stromal vascular fraction (SVF) has high numbers of anti inflammatory macrophages, eosinophils, CD4 + T helper cells and regulatory T cells. Anti inflammatory macrophages are sustained by Th2 type cytokines, such as IL 4. In adipose tissue, the majority of IL 4 secreting cells are eosinophils [9]. In obesity, increased lipid uptake, hypertrophic adipocytes and leakage of fatty acids causes recruitment of pro inflammatory immune cells. Obese WAT and BAT to a lower extent, is infiltrated by pro inflammatory macrophages [1, 11]. Interestingly, eosinophil numbers in WAT are inversely correlated with body weight in mice and absence of eosinophils results in increased body weight, impaired glucose tolerance and less cold induced beige adipogenesis [9, 12]. Anti inflammatory macrophages have been implicated in BAT activation and beiging of WAT, in which they release noradrenaline and activate brown and beige adipocytes [12, 13]. Furthermore, administration of IL 4 in mice reduces body weight, improves insulin sensitivity and promotes beiging of WAT [12]. Thus, BAT activation and WAT beiging is promoted by eosinophils that produce IL 4, which stimulates anti inflammatory macrophages to release noradrenaline, resulting in activation and increased expression of UCP1. Decreased BAT activity in obesity and a decrease of both eosinophils and anti inflammatory macrophages in obese WAT suggests that this immunological circuitry is disturbed in an obese state. Helminths are parasitic worms that induce a type 2 immune response in their host, including a massive increase in eosinophils. Interestingly, this type 2 response induced by Schistosoma mansonisoluble egg antigens (SEA) or Nippostrongylus brasiliensis can be protective against metabolic disorders [9, 14, 15]. Furthermore, soluble products of the whipworm Trichuris suis (TsSP) also induce 67

70 Chapter 4 a type 2 response and are proven to be beneficial in autoimmune diseases, including multiple sclerosis [16]. How TsSP affects adipose tissue immune cells is unknown. The type 2 immune response induced by helminth antigens has already been shown to be metabolically beneficial and we hypothesized that this phenotype could be related to activation of BAT and/or beiging of WAT. Therefore, we first determined how a high fat diet (HFD) affects eosinophils in BAT, epididymal WAT (EpAT) and ScAT using a short time course of HFD. We then studied how 1 week of SEA and TsSP alters the immune composition of these three adipose tissues and whether this resulted in BAT activation and beiging of EpAT and ScAT. Materials and methods Mice Male wild type C57Bl/6 mice (Charles River) were given a HFD (45% kcal fat, 35% kcal carbohydrate, 2% kcal protein, Special Diets Services, Witham, United Kingdom) for 1 day, 3 days or 1 week at the age of 16 weeks (4 groups of n=11). For the SEA and TsSP studies, male wild type C57Bl/6 mice (Charles River) were given a HFD for 1 week and received 4 intraperitoneal injections every 3 days with PBS, 5 µg SEA or 5 µg TsSP starting 3 days before the diet (3 groups of n=12). Mice had ad libitum access to food and water and were maintained under a 12h light dark cycle. Mice were euthanized using.25 mg/g ketamine and.5 mg/g xylazine. Epididymal adipose tissue (EpAT), subcutaneous inguinal adipose tissue (ScAT) and interscapular BAT were dissected and weighed. All experimental procedures were approved by the Animal Experimentation Ethics Committee of the University of Amsterdam. SEA and TsSP preparation Schistosoma mansoni soluble egg antigens (SEA) was kindly provided by Fred Lewis (Biomedical Research Institute, Rockville, MD, USA) and prepared as described previously [17]. Soluble products of Trichuris suis (TsSP) was kindly provided by Irma van Die and prepared as described in [18]. Flow cytometry Adipose tissues were collected in PBS, transferred to DMEM 2 mm HEPES (ThermoFisher Scientific, Waltham, MA, USA) and minced into small pieces. Digestion was done using.25 mg/ml liberase (Roche, Basal, Switzerland) in DMEM 2 mm HEPES for 45 minutes at 37 C. Digested tissue was passed through a 7 µm nylon mesh (BD Biosciences) and centrifuged at 125 rpm for 6 minutes. The adipocyte fraction was removed and the pellet containing the SVF was resuspended in FACS buffer (.5% BSA in PBS). Cell suspensions were incubated with an Fc receptor blocking antibody to prevent non specific binding and a biotin PE dump antibody mix (BioLegend and ebioscience, San Diego, CA, USA) followed by a staining using the antibodies: CD45, CD68 (BioLegend), CD11b, SiglecF, Ly6G (BD Biosciences, Breda, the Netherlands), CD3, MHCII, F4/8 (ebioscience). Stainings were analysed by FACS (LSR Fortessa, BD Biosciences) and FlowJo software (Tree star). Histology Tissues were collected, fixed in 4% paraformaldehyde and embedded in paraffin. Immunohistochemistry on EpAT and ScAT was performed for UCP1 (Sigma Aldrich, Zwijndrecht, the Netherlands). The presence of beige adipocytes, defined by a multilocular appearance and positive 68

71 Type 2 immunological effects of HFD and helminth antigens on adipose tissue for UCP1 was scored by M.T., who was blinded for the experimental conditions. When present, the quantity was expressed as + being a few cells, ++ a few areas and +++ when the section was full of beige adipocytes. Real time PCR Total RNA was extracted using TRIzol (Invitrogen, Carlsbad, CA, USA) and reverse transcribed with an iscript cdna synthesis kit (Bio Rad, Veenendaal, the Netherlands). The quantitative PCR was performed using a SYBR green PCR kit and a ViiA7 RT PCR system (Applied Biosystems, Leusden, the Netherlands). The results are expressed as relative to the control group, which was assigned a value of 1. Culture and differentiation of brown adipocytes T37i cells were kindly provided by Marc Lombès [19] and cultured in DMEM F12 Glutamax, 1% FCS, penicillin (1 U/ml) and streptomycin (1 µg/ml) (ThermoFisher Scientific). After growing confluent, cells were differentiated by adding 2 nm Triiodothyronine (T3) (Sigma Aldrich) and 112 ng/ml insulin (Sigma Aldrich) to the media. After 9 days of differentiation, the cells were stimulated with 2 ng/ml IL 4 (Peprotech, Rocky Hill, NJ, USA) for 24 hours. Supernatant was collected for the quantification of CCL11 (R&D systems, Minneapolis, MN, USA) by ELISA in accordance to the suppliers protocols. Statistics Results are presented as mean ± SEM. Analysis between more than two groups in the time course experiments were done by a one way ANOVA with Tukey post test analysis and the change per group expressed relative to the control group. Analysis between two groups was done by a Student s t test or a chi square test. Statistics were calculated in GraphPad Prism 5. (GraphPad Software, Inc., La Jolla, CA, USA). P values <.5 were considered significant. Results Eosinophils rapidly decrease in adipose tissue upon HFD We were interested in how a HFD affects eosinophil numbers in adipose tissue. Because we previously observed how only one week of HFD induces rapid inflammatory changes in BAT related to cytokine and chemokine expression, we included very short time points of HFD; 1 day, 3 days and 1 week. Furthermore, others have described an important role for eosinophils in adipose tissue in which they are involved in the regulation of brown adipocyte activity and beige adipogenesis. We determined eosinophil numbers in age matched HFD fed C57Bl/6 wild type mice and compared them to mice on a chow diet (Fig.1A). Analysis of adipose tissue immune cells by flow cytometry revealed that already after 1 day of HFD the numbers of eosinophils decline in EpAT and BAT and after 3 days in ScAT as well (Fig.1B). Particularly in BAT, eosinophils dramatically dropped by 65%. In EpAT, this was 35%. In the EpAT, we also observed an increase in macrophages after 3 days of HFD (Fig.1C). These data show us that before the onset of obesity, a HFD has rapid effects on the immune cell composition of brown and white adipose tissues. 69

72 Chapter 4 A B C 45% HFD 1w 3d 1d 4 groups of mice - No HFD - 1 day - 3 days - 1 week EpAT ScAT SiglecF + (% of CD11b + ) SiglecF + (% of CD11b + ) Eosinophils ** *** *** No HFD 1d 3d 1w Eosinophils ** *** F4/8 + (% of CD11b + ) F4/8 + (% of CD11b + ) Macrophages ** ** No HFD 1d 3d 1w Macrophages No HFD 1d 3d 1w No HFD 1d 3d 1w Eosinophils Macrophages BAT SiglecF + (% of CD11b + ) *** ** * F4/8 + (% of CD11b + ) No HFD 1d 3d 1w No HFD 1d 3d 1w Figure 1. A rapid decline in adipose tissue eosinophils upon a short time high fat diet in mice A. Schematic study outline. B. Flow cytometry analysis of the stromal vascular fraction of EpAT, ScAT and BAT showing a decrease in the percentage of eosinophils (SiglecF + leukocytes) after 1 day of HFD in EpAT, 3 days of HFD in ScAT and 1 day of HFD in BAT. C. The percentage of macrophages (F4/8 + leukocytes), which start to increase of 3 days of HFD in EpAT and not in ScAT and BAT. n=11, *=P<.5, **=P<.1, ***=P<.1 (1 way ANOVA with Tukey post test analysis and the change per group expressed relative to the weeks of HFD group). To determine how a HFD can cause such rapid changes in eosinophil numbers, we started by looking at the expression of key chemokines that are known to attract eosinophils, Ccl3, Ccl5 and Ccl11. Although the expression of these chemokines was indeed downregulated, this differed per adipose tissue depot. In EpAT and ScAT, especially Ccl3 was downregulated in the groups of mice that received a HFD for 1 day or more (Fig.2A, B), and in BAT the expression of Ccl5 and Ccl11 was lower in the HFD groups (Fig.2B). IL 4 induced brown adipocyte activation and CCL11 production in vitro The immunological circuit of BAT activation and beiging of WAT is via IL 4 producing eosinophils that stimulate anti inflammatory macrophages to produce noradrenaline which activates brown adipocytes and induces beiging of WAT [9, 2]. To study the effects of such a type 2 antiinflammatory cytokine environment in vitro, we stimulated a brown adipocyte cell line (T37i) with the cytokine IL 4 for 24 hours. Interestingly, IL 4 upregulated Ucp1 expression, suggesting that IL 4 alone already has a direct effect on the activation of brown adipocytes, without the necessity of noradrenalin secreting macrophages (Fig.2C). Furthermore, to have an indication on how CCL11, the most important chemoattractant for eosinophils, is regulated in BAT, we found that stimulation of the brown adipocytes with IL 4 massively increased the expression and secretion of CCL11 (Fig.2D). Altogether, these data suggest a mechanism in which a HFD suppresses the activation of BAT and possibly the beiging potential of WAT by a decrease in IL 4 producing eosinophils. 7

73 Type 2 immunological effects of HFD and helminth antigens on adipose tissue A mrna expression (AU) * * EpAT ** 1d 3d 1w Ccl3 Ccl5 Ccl11 B ScAT mrna expression (AU) *** ** ** 1d 3d 1w. Ccl3 Ccl5 Ccl11 C BAT mrna expression (AU) * * * *** * 1d 3d 1w. Ccl3 Ccl5 Ccl11 D E mrna expression (AU) Ucp1 * - IL-4 mrna expression (AU) Ccl11 *** - IL-4 pg/ml *** CCL11 - IL-4 Figure 2. A HFD decreases gene expression of eosinophil chemoattractants in adipose tissues A. In ScAT, the chemoattractant Ccl3 decreased after 1 day of HFD, whereas in (B) BAT the chemoattractant Ccl5 and Ccl11 were downregulated upon a HFD. C. A brown adipocyte cell line (T37i) was stimulated with the cytokine IL 4 for 24 hours. This upregulated Ucp1, indicating that IL 4 has a direct effect 4 on the activation of brown adipocytes. D. Both gene expression of Ccl11 and the presence of CCL11 in the supernatant as assessed by ELISA was increased by IL 4 stimulation. n=11, *=P<.5, **=P<.1, ***=P<.1 (A,B: 1 way ANOVA with Tukey post test analysis and the change per group expressed relative to the weeks of HFD group. C,D: Student s t test). 71

74 Chapter 4 Adipose tissue depot specific effects of helminth antigen induced type 2 immune responses We wondered if we could rescue the HFD induced decrease in adipose tissue eosinophils in mice by the administration of the helminth antigens SEA and TsSP, which are known to induce a type 2 immune response. We subjected C57Bl/6 wild type mice (n=12) to a HFD for 1 week and injected them with the helminth antigens SEA and TsSP (Fig.3A). These treatments did not cause changes in body weight or BAT, EpAT and ScAT weight (Fig.3B). Flow cytometry of the adipose tissues revealed major effects on immune cell composition of the EpAT in the SEA and TsSP groups (Fig.4A). Total leukocyte numbers (CD45 + cells) were 6.4 fold and 4. fold increased in SEA and TsSP groups, respectively. High counts of adipose tissue eosinophils (SiglecF + cells) were observed by SEA, +34 fold difference and +23 fold difference by TsSP. The number of macrophages (CD68 + MHCII + cells) increased by 5.8 fold and 5.2 fold in the SEA and TsSP groups respectively. Furthermore, SEA increased T cell numbers (CD3 + cells) 12.3 fold, and TsSP 7.6 fold. Neutrophil counts (Ly6G + cells) were unaffected by the treatments. However, the effects of the helminth antigens on ScAT and BAT were minor compared to EpAT with slightly more T cells (+3.3 fold difference) and macrophages (+2.4 fold difference) by TsSP in the ScAT and no changes in BAT (Fig.4B,C). Gene expression data confirmed increased macrophages and eosinophils in EpAT by Cd68 and Ccr3 expression respectively (Fig.4D). Furthermore, the chemokines responsible for attracting eosinophils Ccl3, Ccl5 and Ccl11 were indeed upregulated in EpAT (Fig.4D). In ScAT and BAT, no alterations in gene expression were observed (Fig.4E,F). A B Day -3 Day Day 3 Day 6 1 week 45% HFD Body weight (g) 2 1 EpAT (mg) 4 2 ScAT (mg) BAT (mg) Control SEA TsSP Control SEA TsSP Control SEA TsSP Control SEA TsSP Figure 3. No changes in body weight after short term treatment with SEA and TsSP A. Schematic study design. B. Body weight and the weight of three adipose tissue depots, EpAT, ScAT and BAT, did not differ between the groups. n=12 Helminth antigen treatment does not cause beiging in WAT As especially ScAT is prone to beiging, we wondered whether the small change in leukocyte composition in ScAT altered the presence of beige adipocytes. We performed immunohistochemistry for UCP1 on ScAT and scored the occurrence of UCP1 positive areas in a blinded fashion. This quantification showed a small trend towards a higher presence of beige adipocytes in ScAT of mice treated with the helminth antigens SEA, and not in the TsSP group (Table 1, Fig.5A). Furthermore, we did not observe any beige adipocytes in EpAT and thus the major increase in eosinophils in EpAT did not result in beige adipogenesis in this adipose tissue depot (data not shown). In BAT, Ucp1 expression was unaffected by either SEA or TsSP treatment (Fig.5B). 72

75 Type 2 immunological effects of HFD and helminth antigens on adipose tissue A EpAT Leukocytes T cells Macrophages Eosinophils Neutrophils CD45 + (#/g) **** **** CD3 + (#/g) *** **** CD68 + MHCII + (#/g) *** ** SiglecF + (#/g) **** **** Ly6G + (#/g) Control SEA TsSP Control SEA TsSP Control SEA TsSP Control SEA TsSP Control SEA TsSP B ScAT Leukocytes T cells Macrophages Eosinophils Neutrophils CD45 + (#/g) CD3 + (#/g) * CD68 + MHCII + (#/g) * SiglecF + (#/g) Ly6G + (#/g) Control SEA TsSP Control SEA TsSP Control SEA TsSP Control SEA TsSP Control SEA TsSP C BAT Leukocytes T cells Macrophages Eosinophils Neutrophils CD45 + (#/g) CD3 + (#/g) CD68 + MHCII + (#/g) 4 2 SiglecF + (#/g) 1 5 Ly6G + (#/g) Control SEA TsSP Control SEA TsSP Control SEA TsSP Control SEA TsSP Control SEA TsSP D EpAT E ScAT mrna expression (AU) *** ** ** * ** * Cd68 Ccr3 Ccl3 Ccl5 Ccl11 *** *** Ctrl SEA TsSP *** ** mrna expression (AU) Cd68 Ccr3 Ccl5 Ccl11 Ctrl SEA TsSP F mrna expression (AU) BAT Cd68 Ccr3 Ccl3 Ccl5 Ccl11 Ctrl SEA TsSP Figure 4. Helminth antigens induce an adipose depot specific type 2 immune response A. Injections with SEA and TsSP resulted in a large increase in the number of total EpAT leukocytes (CD45 + ), T cells (CD3 + ), Macrophages (CD68 + MHCII + ), Eosinophils (SiglecF + ) and not in neutrophils (Ly6G + ). B. In ScAT, only TsSP showed a significant induction of T cells and macrophages. C. No changes in immune cell composition in BAT were observed after SEA and TsSP treatment. D. Gene expression analysis showing that macrophage marker Cd68, eosinophil marker Ccr3 and chemokines Ccl3, Ccl5 and Ccl11 were all increased in EpAT of SEA and TsSP treated mice. E,F. No effects of the treatments were observed in ScAT and BAT. n=12, *=P<.5, **=P<.1, ***=P<.1 (Student s t test between the control group and either SEA or TsSP) 73

76 Chapter 4 Table 1. Presence of beiging in ScAT after helminth treatment p value Ctrl SEA TsSP Values represent the number of mice within the groups that were scored on no ( ) or a few beige adipocytes (+), a few beige areas (++) or many beige areas (+++) in ScAT. P value as determined by a Chi square test. n=1 12 A ScAT Control SEA TsSP B BAT Ucp1 mrna expression (AU) Ctrl SEA TsSP Figure 5. No effects on thermogenic capacity by helminth antigen treatment A. Scoring of presence of beige adipocytes in ScAT. Beige adipocytes were defined by a multilocular appearance and positive for UCP1. When present, the quantity was expressed as +: a few cells, ++: a few areas and +++ when the section was full of beige adipocytes. B. Gene expression of Ucp1 in BAT was not different upon SEA or TsSP treatment. n=12 74

77 Type 2 immunological effects of HFD and helminth antigens on adipose tissue Discussion Immune cells contribute to activation of BAT and beiging of WAT, via IL 4 producing eosinophils that stimulate anti inflammatory macrophages to secrete noradrenaline [9, 13]. In the current study, we show that short term HFD causes a rapid decline in eosinophils in EpAT, ScAT and BAT, along with a decrease in chemotactic signals for eosinophils. We also show that the anti inflammatory cytokine IL 4 can directly activate brown adipocytes in vitro and that brown adipocytes themselves are extremely capable of producing and secreting CCL11, the chemokine that recruits eosinophils. Altogether, these findings indicate that a type 2 and anti inflammatory environment in adipose tissues and the associated beige/brown adipocyte activation or beige adipogenesis could beneficially alter metabolism in obesity. Therefore, we administrated helminth antigens from SEA and TsSP to mice on a HFD for 1 week to induce a type 2 immune response. We observed very depot specific effects of SEA and TsSP in which EpAT was massively infiltrated with eosinophils, macrophages and T cells whereas almost no differences were found in ScAT and BAT. Furthermore, SEA and TsSP treatments did not result in increased thermogenic energy expenditure, as we did not observe alterations in BAT activity or beiging of the WAT depots. Eosinophil numbers depend on signals that affect recruitment, survival and/or cell death. Our data suggests that the recruitment of eosinophils is indeed altered, as we observe a decrease in gene expression of the chemotactic signals Ccl3, Ccl5 and Ccl11. Eosinophils greatly depend on survival signals, including IL 3, IL 5 and granulocyte macrophage colony stimulating factor (GM CSF), of which IL 5 is the most important [21]. IL 5 in turn is produced by group 2 innate lymphoid cells (ILC2s) [22]. Eosinophils have a short half life, reports vary between 3 to 18 hours in blood and up to 6 days in tissues, with a great tissue dependent variety [21 23]. In lean adipose tissue, eosinophils have a low turnover (1% in 3 days) [22], suggesting that a lean state contains high survival signals. Factors related to obesity and potentially involved in reducing eosinophil numbers include the proinflammatory cytokines TNF, which can induce eosinophil apoptosis [21] and IFNγ, which can inhibit differentiation and migration of eosinophils [24]. Furthermore, TGFβ has been suggested to block the survival effects of IL 5 on eosinophils [25], although TGFβ expression also correlates with an increase in eosinophils in airways of asthma patients [26]. TGFβ is a multifunctional cytokine that possibly has diverse effects in different tissues. Eosinophil degranulation can also result in cell death [27]. As eosinophils have a short half life, a withdrawal of survival signals and/or increased apoptotic factors while chemoattractants are decreased can result in a rapid decline. To unravel a genuine mechanism behind this fast drop in adipose tissue eosinophils upon a HFD, future studies should address which signals cause chemokines to decrease, determine whether IL 5 and ILC2s are affected by short term HFD and find correlations between apoptotic factors and eosinophil numbers in adipose tissues. Eosinophils migrate via vascular cell adhesion molecule 1 (VCAM 1) and intercellular cell adhesion molecule 1 (ICAM 1) through α4 and αl integrins expressed on eosinophils, as integrin antibodies block eosinophil accumulation in adipose tissue [9]. This makes VCAM 1 and ICAM 1 also interesting targets to study if eosinophil migration is affected in obese adipose tissues. Our results are in line with others who found that WAT of obese mice has a reduction in eosinophils [9]. We further state that eosinophils immediately decline shortly after the start of a HFD. The next 75

78 Chapter 4 important step would be to confirm these immunological changes in human adipose tissue, as the physiological resemblance for a shift from chow food to 1 day of HFD in mice is difficult to conceptualize. To further improve the translational relevance of these findings, it would be very interesting to determine whether increased lipid uptake of adipocytes or a specific dietary component is responsible for the rapid immune response in adipose tissue. Others have shown that chronic SEA treatment and the associated increase in type 2 immunity can be protective in the pathogenesis of metabolic disorders [14]. We hypothesized that these metabolic improvements were related to BAT activity and beiging of WAT. In our study, the type 2 immune response induced by helminth antigens from both SEA as well as TsSP was very depot specific and was only full blown present in EpAT. Surprisingly, an extreme increase in EpAT eosinophils (+34 fold by SEA and +23 fold by TsSP) did not cause beige adipogenesis. This might be explained by a low beiging capacity of EpAT [5]. Therefore, we speculate that BAT activation or WAT beiging is not the mechanism that is causing the metabolic effects that others have seen upon a chronic infection with SEA. Beiging predominantly occurs in ScAT, possibly due to differences in a precursor population for beige adipocytes [28], which is an interesting topic for future research. Administration of IL 4 for 1 days in vivo induces beiging of ScAT [12], indicating that this mechanism could take place in this time frame. In our study, although 1 days of treatment with SEA and TsSP was enough to induce a massive immune response in EpAT, we did not observe effects on body weight or adipose tissue weight. Therefore, this short term study might not have been long enough to observe either metabolic effects or BAT activation and beige adipogenesis. With the use of helminth antigens, we wanted to reverse the effects of a drop in eosinophils in adipose tissue caused by a HFD. The mice in this experiment were therefore given a HFD during the treatment. In ScAT and BAT, the effect of HFD might have caused a larger decline in eosinophils then the SEA and TsSP could restore. Therefore, with the HFD, we might have decreased the capacity to activate BAT and the beiging potential of ScAT and we could have seen effects when performing the study under chow conditions. Helminths express a wide variety of protein and lipid linked glycans and it is unknown which molecular mechanisms are responsible for the immune modulating effects [29]. In our study, we observe similar effects from SEA and TsSP, indicating that they have a common parasitic signal that exerts the effects in adipose tissue. Before helminth antigens could be a therapeutically applied, it will be needed to identify which parasitic signal causes the metabolically beneficial type 2 immune response. In healthy adipose tissue, a type 2 immunological circuitry contributes to tissue homeostasis, by IL 4 producing eosinophils that sustain anti inflammatory macrophages [9]. Our data shows that Il 4 probably has a bigger role in adipose tissue than only the polarization of macrophages with also a direct effect on brown adipocytes. A HFD disturbs the homeostatic type 2 immunological circuitry, and cannot be reversed by a helminth induced type 2 response. The crosstalk between immune cells and adipocytes and how they regulate tissue homeostasis is more complex than previously thought with many questions that remain. 76

79 Type 2 immunological effects of HFD and helminth antigens on adipose tissue Corresponding author Prof. Menno P.J. de Winther. Department of Medical Biochemistry, Experimental Vascular Biology, Academic Medical Center, University of Amsterdam, Meibergdreef 9, 115 AZ, Amsterdam, The Netherlands. +31 () m.dewinther@amc.uva.nl Conflict of interest The authors declare no conflict of interest. Acknowledgements We acknowledge the support from the Rembrandt Institute of Cardiovascular Science (PR, MW, EL) and the Netherlands CardioVascular Research Initiative: the Dutch Heart Foundation, Dutch Federation of University Medical Centres, the Netherlands Organisation for Health Research and Development and the Royal Netherlands Academy of Sciences" for the GENIUS project Generating the best evidence based pharmaceutical targets for atherosclerosis (CVON211 19). This work was supported by the Netherlands Organization for Scientific Research (NWO) (VICI grant to EL), the Netherlands Heart (Dr E. Dekker grant to TS) and the European Research Council (ERC con grant to EL). 77

80 Chapter 4 References [1] Cannon B, Nedergaard J. 24. Brown Adipose Tissue: Function and Physiological Significance. Physiological Reviews 84(1): [2] van Marken Lichtenbelt WD, Vanhommerig JW, Smulders NM, Drossaerts JMAFL, Kemerink GJ, Bouvy ND et al. 29. Cold Activated Brown Adipose Tissue in Healthy Men. New England Journal of Medicine 36(15): [3] Saito M, Okamatsu Ogura Y, Matsushita M, Watanabe K, Yoneshiro T, Nio Kobayashi J et al. 29. High Incidence of Metabolically Active Brown Adipose Tissue in Healthy Adult Humans: Effects of Cold Exposure and Adiposity. Diabetes 58(7): [4] Sanchez Gurmaches J, Hung C M, Guertin DA Emerging Complexities in Adipocyte Origins and Identity. Trends in Cell Biology 26(5): [5] Waldén TB, Hansen IR, Timmons JA, Cannon B, Nedergaard J Recruited vs. nonrecruited molecular signatures of brown, brite, and white adipose tissues. American Journal of Physiology Endocrinology And Metabolism 32(1):E19. [6] Shabalina Irina G, Petrovic N, de Jong Jasper MA, Kalinovich Anastasia V, Cannon B, Nedergaard J UCP1 in Brite/Beige Adipose Tissue Mitochondria Is Functionally Thermogenic. Cell Reports 5(5): [7] Yoneshiro T, Saito M Activation and recruitment of brown adipose tissue as anti obesity regimens in humans. Annals of Medicine 47(2): [8] Lumeng CN, Bodzin JL, Saltiel AR. 27. Obesity induces a phenotypic switch in adipose tissue macrophage polarization. The Journal of Clinical Investigation 117(1): [9] Wu D, Molofsky AB, Liang H E, Ricardo Gonzalez RR, Jouihan HA, Bando JK et al Eosinophils Sustain Adipose Alternatively Activated Macrophages Associated with Glucose Homeostasis. Science 332(626): [1] Weisberg SP, McCann D, Desai M, Rosenbaum M, Leibel RL, Ferrante AW, Jr. 23. Obesity is associated with macrophage accumulation in adipose tissue. The Journal of Clinical Investigation 112(12): [11] Roberts Toler C, O'Neill BT, Cypess AM Diet induced obesity causes insulin resistance in mouse brown adipose tissue. Obesity 23(9): [12] Qiu Y, Nguyen Khoa D, Odegaard Justin I, Cui X, Tian X, Locksley Richard M et al Eosinophils and Type 2 Cytokine Signaling in Macrophages Orchestrate Development of Functional Beige Fat. Cell 157(6): [13] Nguyen KD, Qiu Y, Cui X, Goh YPS, Mwangi J, David T et al Alternatively activated macrophages produce catecholamines to sustain adaptive thermogenesis. Nature 48(7375): [14] Hussaarts L, García Tardón N, van Beek L, Heemskerk MM, Haeberlein S, van der Zon GC et al Chronic helminth infection and helminth derived egg antigens promote adipose tissue M2 macrophages and improve insulin sensitivity in obese mice. The FASEB Journal 29(7): [15] Wolfs IMJ, Stöger JL, Goossens P, Pöttgens C, Gijbels MJJ, Wijnands E et al Reprogramming macrophages to an anti inflammatory phenotype by helminth antigens reduces murine atherosclerosis. The FASEB Journal 28(1): [16] Kooij G, Braster R, Koning JJ, Laan LC, van Vliet SJ, Los T et al Trichuris suis induces human non classical patrolling monocytes via the mannose receptor and PKC: implications for multiple sclerosis. Acta Neuropathologica Communications 3(1):45. [17] Boros DL, Warren KS Delayed hypersensitivity type granuloma formation and dermal reaction induced and elicited by a soluble factor isolated from Schistosoma mansoni eggs. The Journal of Experimental Medicine 132(3):488. [18] Klaver EJ, Kuijk LM, Laan LC, Kringel H, van Vliet SJ, Bouma G et al Trichuris suis induced modulation of human dendritic cell function is glycan mediated. International Journal for Parasitology 43(3 4): [19] Zennaro MC, Le Menuet D, Viengchareun S, Walker F, Ricquier D, Lombès M Hibernoma development in transgenic mice identifies brown adipose tissue as a novel target of aldosterone action. The Journal of Clinical Investigation 11(6): [2] Lee M W, Odegaard JI, Mukundan L, Qiu Y, Molofsky AB, Nussbaum JC et al Activated Type 2 Innate Lymphoid Cells Regulate Beige Fat Biogenesis. Cell 16(1 2): [21] Park YM, Bochner BS. 21. Eosinophil Survival and Apoptosis in Health and Disease. Allergy, Asthma & Immunology Research 2(2): [22] Molofsky AB, Nussbaum JC, Liang H E, Van Dyken SJ, Cheng LE, Mohapatra A et al Innate lymphoid type 2 cells sustain visceral adipose tissue eosinophils and alternatively activated macrophages. The Journal of Experimental Medicine 21(3):535. [23] Wen T, Besse JA, Mingler MK, Fulkerson PC, Rothenberg ME Eosinophil adoptive transfer system to directly evaluate pulmonary eosinophil trafficking in vivo. Proceedings of the National Academy of Sciences 11(15): [24] Ochiai K, Iwamoto I, Takahashi H, Yoshida S, Tomioka H, Yoshida S Effect of IL 4 and interferon gamma (IFN gamma) on IL 3 and IL 5 induced eosinophil differentiation from human cord blood mononuclear cells. Clinical and Experimental Immunology 99(1):

81 Type 2 immunological effects of HFD and helminth antigens on adipose tissue [25] Alam R, Forsythe P, Stafford S, Fukuda Y Transforming growth factor beta abrogates the effects of hematopoietins on eosinophils and induces their apoptosis. The Journal of Experimental Medicine 179(3):141. [26] Tirado Rodriguez B, Ortega E, Segura Medina P, Huerta Yepez S TGF β: An Important Mediator of Allergic Disease and a Molecule with Dual Activity in Cancer Development. Journal of Immunology Research 214: [27] Rosenberg HF, Dyer KD, Foster PS Eosinophils: changing perspectives in health and disease. Nature reviews. Immunology 13(1):9 22. [28] Sanchez Gurmaches J, Guertin DA Adipocytes arise from multiple lineages that are heterogeneously and dynamically distributed. Nature Communications 5. [29] Kuijk LM, van Die I. 21. Worms to the rescue: Can worm glycans protect from autoimmune diseases? IUBMB Life 62(4):

82 Chapter 4 8

83 5 Diet induced obesity in mice diminishes hematopoietic stem and progenitor cells in the bone marrow Susan M. van den Berg 1,*, Tom T.P. Seijkens 1,*, Pascal J.H. Kusters 1, Linda Beckers 1, Myrthe den Toom 1, Esther Smeets 1, Johannes Levels 2, Menno P.J. de Winther 1 and Esther Lutgens 1,3 1 Department of Medical Biochemistry, Experimental Vascular Biology, Academic Medical Centre, University of Amsterdam, the Netherlands 2 Department of Experimental Vascular Medicine, Academic Medical Centre, University of Amsterdam, the Netherlands 3 Institute for Cardiovascular Prevention, Ludwig Maximilians University, Munich, Germany * these authors contributed equally to this work. The FASEB Journal 216; 3(5):

84 Chapter 5 Abstract Obesity is associated with chronic low grade inflammation, characterized by leukocytosis and inflammation in the adipose tissue. Continuous activation of the immune system is a stressor for hematopoietic stem and progenitor cells (HSPCs) in the bone marrow (BM). Here we studied how diet induced obesity (DIO) affects HSPC population dynamics in the BM. Eight groups of age matched C57Bl/6 mice received a high fat diet (HFD; 45% kcal from fat) ranging from 1 day up to 18 weeks. The obesogenic diet caused decreased proliferation of lineage Sca 1 + c Kit + (LSK) cells in the BM and a general suppression of progenitor cell populations including common lymphoid progenitors and common myeloid progenitors. Within the LSK population, DIO induced a shift in stem cells that are capable of self renewal towards maturing multipotent progenitor cells. The higher differentiation potential resulted in increased lymphoid and myeloid ex vivo colony forming capacity. In a competitive BM transplantation, BM from obese animals showed impaired multilineage reconstitution when transplanted into chow fed mice. Our data demonstrate that obesity stimulates the differentiation and reduces proliferation of HSPCs in the BM, leading to a decreased HSPC population. This implies that the effects of obesity on HSPCs hampers proper functioning of the immune system. 82

85 DIO in mice decreases HSPCs in the BM Introduction Obesity and obesity associated complications such as cardiovascular diseases and type 2 diabetes are responsible for high morbidity and mortality rates worldwide. Obesity is characterized by a proinflammatory activation status of immune cells within the adipose tissue, but also in liver, pancreas, muscle and hypothalamus [1]. Cells from both the innate and adaptive immune system, especially adipose tissue macrophages, but also CD8 +, CD4 + effector and regulatory T cells, are crucial in the pathogenesis of obesity and its metabolic complications [1, 2]. Interestingly, in experimental as well as human obesity, an increase in the circulating pool of these (activated) immune cells has been observed [3 5]. The primary site of immune cell production is the bone marrow (BM), where the most primitive precursors, the hematopoietic stem and progenitor cells (HSPCs), reside. These cells, defined as lineage Sca 1 + c Kit + (LSK) cells, are located in a specialized BM microenvironment, which critically regulates self renewal, quiescence, differentiation and mobilization of the HSPCs via growth factors (including stem cell factor; SCF, macrophage, granulocyte macrophage, granulocyte, colony stimulating factor; M CSF, GM CSF, G CSF), chemokines, cytokines (including TNF, IFNγ, IL 6), cell cycle regulators (cyclins and tumor suppressor genes) and adhesion signals [6, 7]. Hence, HSPC quiescence and proliferation is tightly regulated to prevent stem cell exhaustion, thereby ensuring lifelong hematopoiesis [8]. The most primitive hematopoietic stem cells are defined by the cell surface marker CD15 and can give rise to an active self renewing HSPC population [9]. The acquisition of CD34 is one of the earliest events during activation of HSPCs from their dormant state. Further differentiation is marked by loss of CD15 and gain of CD135, resulting in differentiating progenitor cells [9, 1]. Loss of CD15 expressing cells is therefore functionally associated with less self renewal capacity. Hematopoietic stressors, such as inflammation, affect HSPC homeostasis. HSPCs directly respond to cytokines, such as IFNγ, TNF, IL 1 and IL 6 [11 16]. For example, Escherichia coli infection leads to an expansion of the HSPC population via LPS induced TNF and NF κb signaling [17, 18]. Furthermore, pseudomonas aeruginosa and mycobacterium avium infections are associated with defective stem cell activity and BM from infected mice does not engraft well after transplantation [14, 19]. These studies suggest that immune mediated BM exhaustion can result from persistent activation of quiescent HSPCs, thus depleting the total HSPC population over time. Besides acute inflammation during infection, chronic inflammatory conditions also alter HSPC biology. We and others have reported that atherosclerosis, a chronic inflammatory disease of the arteries, is associated with hypercholesterolemia driven expansion of the HSPC population in BM of LDLr / mice, especially of the myeloid population [2, 21]. Mouse models of extreme obesity, including leptin deficient Ob/Ob mice, as well as C57Bl/6 mice on a 6% high fat diet (HFD), also present with an increase in progenitor cells in the BM [5, 22]. However, the precise effects of obesity on the HSPC population remain elusive. We therefore investigated how an obesogenic diet affects the HSPC population in the BM over the course of 18 weeks, using a relatively mild model (45% kcal from fat) of diet induced obesity (DIO) in mice. 83

86 Chapter 5 Materials and methods Animals Age matched male C57Bl/6 mice (Charles River) were included in the experiments at the age of 7 weeks, at which time the first group of mice received a HFD (45% kcal fat, 35% kcal carbohydrate, 2% kcal protein, Special Diets Services, Witham, United Kingdom) and all other groups were kept on a chow diet. All mice were given a HFD in a time course ranging from 1 day up to 18 weeks before sacrifice. All mice were sacrificed at the age of 25 weeks, regardless of the length of obesogenic diet. Mice had ad libitum access to food and water and were maintained under a 12h light dark cycle. Mice were fasted overnight and subsequently euthanized. Glucose levels were measured from whole blood using a glucometer (Bayercontour, Basel, Switzerland). Blood was obtained by cardiac puncture using EDTA filled syringes and organs (BM and epididymal adipose tissue; EpAT) were dissected and processed for flow cytometric analysis. All experimental procedures were approved by the Animal Experimentation Ethics Committee of the University of Amsterdam. Hematological and lipoprotein measurements Hematological analysis was performed on a ScilVet abc plus+ (ScilVet, Oostelbeers, The Netherlands). Plasma triglyceride and cholesterol levels were measured by fast performance liquid chromatography, as described previously [2]. Flow cytometry BM was harvested in cold PBS. BM cell suspension was passed through a 7 µm nylon mesh (BD Biosciences, Breda, the Netherlands). Lineage depletion was performed for the HSPC analysis by magnetic bead isolation according to the manufacturer s instructions (Lineage Cell Depletion Kit; Miltenyi Biotec, Teterow, Germany). EpAT was rinsed in PBS, minced into small pieces, digested in a collagenase mixture (DMEM 2 mm HEPES, Collagenase I and XI, Sigma Aldrich, Zwijndrecht, the Netherlands) for 45 minutes at 37 C, passed through a 7 µm nylon mesh (BD Biosciences) and centrifuged at 1,25 rpm for 6 minutes. The pelleted SVF was resuspended in FACS buffer (.5% BSA in PBS). Blood and BM cell suspensions were incubated with hypotonic lysis buffer (8.4 g NH 4 Cl and.84 g NaHCO 3 per liter distilled water) to remove erythrocytes. Cell suspensions were incubated with an Fc receptor blocking antibody to prevent non specific binding. CD3, CD8, CD25, FoxP3, F4/8, CD11b, CD11c, Gr 1, CD45.1, B22 (ebioscience, San Diego, CA, USA), CD4 (BD Biosciences) and CD45, Ly6G, CD26 (Biolegend, San Diego, CA, USA) antibodies were incubated with the indicated tissues. BM cells were characterized using antibodies for CD5, CD11b, Ter119, Sca 1, CD34, CD16/32, CD127 (ebioscience), B22 (BD Biosciences) and Ly6G, c Kit, CD15, CD135 (Biolegend). Staining was analyzed by FACS (FACSCanto II, BD Biosciences) and FlowJo software version (Tree star). FACS analysis on mature hematopoietic cells in blood started by first gating on the total CD45 + leukocyte population. Myeloid cells were selected on CD11b +. Within the CD11b + population, monocytes were selected by excluding Ly6G + neutrophils and pro inflammatory monocytes were then defined by the remaining Gr1 + population, as Gr1 binds both Ly6C as Ly6G. Lymphoid cells were selected on CD3 +, then gated for CD8 + CD4 (cytotoxic T cells) and CD8 CD4 + (T helper cells). Within this T helper cell population we characterized regulatory T cells by CD25 + FoxP

87 DIO in mice decreases HSPCs in the BM BrdU labelling Mice were injected intraperitoneally with.2 mg/g BrdU. BM was collected 16 hours later, and Lin cells were isolated. BrdU incorporation was determined by intracellular staining with anti BrdU antibodies, using the FITC BrdU Flow Kit (BD Biosciences). Propidium iodide (PI) staining For cell cycle analysis, Lin BM cells were stained with the primary antibodies, fixed in 7% ethanol for 48 hours, and treated with PI RNase buffer (BD Biosciences). Colony forming unit (CFU) assays BM was isolated and 1x1 4 BM cells were cultured in 2 ml semisolid methylcellulose medium supplemented with growth factors (MethoCult; Stem Cell Technologies, Grenoble, France) at 37 C in 98% humidity and 5% CO 2 for 7 days. The amount of lymphoid, myeloid, granulocyte and monocyte colonies was counted by T.S., who was blinded for the experimental conditions. Competitive Bone Marrow Transplantation (cbmt) C57Bl/6 CD45.2 recipient mice were housed in filter top cages and received antibiotics in their drinking water (6 U/ml polymyxin B sulfate, Invitrogen, Carlsbad, CA, USA and 1 g/ml neomycin, Sigma) 1 week pre BMT and 5 weeks post BMT. The mice received 2x6 Gy total body irradiation on two consecutive days. BM was isolated from age matched donor C57Bl/6 CD45.1 mice fed either chow or 45% HFD for 1 weeks, and mixed 2:1 with BM from donor C57Bl/6 CD45.2 chow fed mice. Donor BM was injected intravenously. Starting 4 weeks post BMT, blood samples were taken every 3 weeks to analyze the reconstitution of peripheral blood leukocytes by flow cytometry. Quantitative PCR Total RNA was extracted using TRIzol (Invitrogen) and reverse transcribed with an iscript cdna synthesis kit (Bio Rad, Veenendaal, the Netherlands). The quantitative PCR was performed using a SYBR green PCR kit and a ViiA7 RT PCR system (Applied Biosystems, Leusden, the Netherlands). The result is expressed as relative to the control group, which was assigned a value of 1. Statistics Results are presented as mean ± SEM. Analysis between two groups was done by a Student s t test, more than two groups by a one way ANOVA with Tukey post test analysis and the change per group expressed relative to the control group. The repeated measures on blood leukocytes in the cbmt experiments were analyzed by a two way ANOVA with matched values and a bonferroni multiple comparison test. Statistics were calculated in GraphPad Prism 5. (GraphPad Software, Inc., La Jolla, CA, USA). P values <.5 were considered significant. Results A 45% obesogenic diet induces a pro inflammatory immune cell profile in the peripheral blood Eight groups of age matched C57Bl/6 mice were subjected to different durations of obesogenic diet (Fig. 1A for a graphic study design). Body weight increased after 2 weeks of the obesogenic diet (Table 1). Blood glucose levels increased in the HFD groups (Table 1). Total plasma cholesterol levels (in VLDL, LDL and HDL) increased after 3 days, and plasma triglyceride levels (in VLDL) were elevated after 2 weeks of HFD feeding (Table 1) compared to the days HFD group. 85

88 Chapter 5 Table 1. General characteristics of mice in the course of DIO Body weight, g 27.3 ±.6 Glucose, mg/dl 69 ± 3 Triglycerides, µm 23 ± 19 VLDL 1 ± 15 LDL 9 ± 1 HDL 13 ± 2 Cholesterol, mm 1.75 ±.13 VLDL.2 ±.4 LDL.27 ±.3 HDL 1.46 ±.13 Weeks of HFD Mean ± SEM (p) a 1/7 3/ ±.6 13 ± 15 (***) 253 ± ± ± 3 1 ± ±.17.2 ±.3.42 ± ± ± ± 7 (**) 45 ± ± ± 27 2 ± ±.18 (***).3 ±.9.81 ±.11 (**) 2.7 ±.16 (*) 31.5 ± ± 4 (***) 343 ± ± ± 3 15 ± ±.26 (***).3 ±.8.68 ±.8 (*) 2.17 ±.19 (**) 34.4 ±.9(***) 14 ± 5 (*) 58 ± 71 (**) 48 ± 55 (**) 88 ± ± ±.13 (**).5 ±.7.74 ±.13 (**) 2.5 ±.8 (*) 36.5 ± 1.2(***) 91 ± 4 81 ± 82 (***) 668 ± 9 (***) 128 ± ± ±.27 (***).9 ±.2 (***).76 ±.14 (**) 2.3 ±.15 (*) 38.1 ±.9(***) 112 ± 9 (**) 535 ± 28 (**) 423 ± 24 (***) 95 ± 1 18 ± ±.8 (***).6 ±.5 (*).89 ±.9 (***) 2.32 ±.8 (***) 41.3 ± 2. (***) 124 ± 7 (***) 63 ± 47 (***) 473 ± 46 (***) 112 ± 6 18 ± ±.8 (***).6 ±.6 (*) 1. ±.6 (***) 2.26 ±.4 (***) a Data in this table is analyzed by a one way ANOVA with Tukey post test analysis and the change per group expressed relative to the weeks of HFD group. n=1 11, *=P<.5, **=P<.1, ***=P<.1. Total blood leukocyte counts did not differ between the groups (Fig. 1B). The percentage of CD11b + Ly6G monocytes as well as the percentage of CD3 + T cells was not affected by the obesogenic diet (Fig. 1C,D). However, both the myeloid and lymphoid fraction showed a pro inflammatory profile. Starting at 4 weeks of HFD, the number of pro inflammatory Gr1 high monocytes (Fig. 1E) and number of cytotoxic CD8 + T cells, the T cell fraction considered to induce AT inflammation and insulin resistance [23], had increased (Fig. 1F). This was accompanied by a relative decrease in CD4 + T helper cells (Fig. 1G). The regulatory T cell fraction showed a slight increase during the first few weeks of HFD, but decreased again with more advanced obesity (Fig. 1H). Similar results were obtained in spleen (Supplemental Fig. 1). An obesogenic diet causes a switch from quiescent to differentiating LSK cells Since the obesogenic diet affects the mature immune cell populations, we investigated the effects of an obesogenic diet on HSPC biology. HSPCs were defined as Lin Sca 1 + c Kit + (LSK) cells and within this LSK population we studied different subsets; long term hematopoietic stem cells (LT HSCs; Lin Sca 1 + c Kit + CD15 + CD34 ), short term hematopoietic stem cells (ST HSCs; Lin Sca 1 + c Kit + CD15 + CD34 + ), as well as the early multipotent progenitors (E MPPs; Lin Sca 1 + c Kit + CD15 CD135 ), and late multipotent progenitors (L MPPs; Lin Sca 1 + c Kit + CD15 CD135 + )(Fig. 2A for an overview) [2]. Interestingly, within the LSK population, the obesogenic diet relatively decreased the ST HSCs after 4 weeks whereas the percentage of LT HSC fraction was not affected (Fig. 2B,C). Remarkably, the percentage of E MPPs was increased already after 3 days of HFD feeding and remained increased 86

89 DIO in mice decreases HSPCs in the BM A Chow HFD 7 weeks Age 25 weeks Weeks of HFD 1/7 3/ B WBC 1 9 /l Blood leukocyte count 1/7 3/ Weeks of HFD C % CD11b + Ly6G - (of CD45 + ) Monocytes 1/7 3/ Weeks of HFD D % CD3 + (of CD45 + ) T cells 1/7 3/ Weeks of HFD E % Gr1 high (of CD45 + ) Pro-inflammatory monocytes * * 1/7 3/ Weeks of HFD F % CD8 + (of CD45 + CD3 + ) Cytotoxic T cells 4 * * /7 3/ Weeks of HFD G % CD4 + (of CD45 + CD3 + ) Helper T cells * * ** ** 1/7 3/ Weeks of HFD H % CD25 + FoxP3 + (of CD4 + ) Regulatory T cells ** 1/7 3/ Weeks of HFD Figure 1. Pro inflammatory immune cell profile in peripheral blood by FACS in a time course of DIO. (a) Study design. (b) No changes in total blood leukocyte counts or (c) percentage of CD11b + Ly6G monocytes and (d) CD3 + T cells of the total leukocyte population (CD45 + ). (e) After 4 weeks of the obesogenic diet, pro inflammatory Gr1 high monocytes increased, as well as (f) CD8 + cytotoxic T cells. (g) CD4 + T helper cells decreased, whereas (h) the regulatory CD4 + CD25 + FoxP3 + T cells increased after 1 week of HFD but stabilized again in the course of DIO. n=1 11, *=P<.5, **=P<.1 (1 way ANOVA with Tukey post test analysis and the change per group expressed relative to the weeks of HFD group). over the course of obesity (Fig. 2D). We also observed a small decrease in percentage of L MPP after 1 weeks of HFD (Fig. 2E). LT and ST HSCs are capable of self renewal, however, this characteristic is lost when they mature and differentiate to multipotent progenitors. Our data thus indicate that an obesogenic diet shifts the characteristics of early HSPCs from a quiescent population capable of selfrenewal, towards a more mature HSPC population with increased differentiation potential. 87

90 Chapter 5 A B C LT- HSC CD15 + CD34 - ST-HSC CD15 + CD34 + E-MPP Primitivity % CD15 + CD34 - (of LSK) D LT-HSCs in BM 1/7 3/ Weeks of HFD E-MMPs in BM 8 4 E % CD15 + CD34 + (of LSK) ST-HSCs in BM *** *** ** 1/7 3/ Weeks of HFD L-MMPs in BM CD15 - CD135 - L-MPP % CD15 - CD135 - (of LSK) * * *** *** *** % CD15 - CD135 + (of LSK) * CD15 - CD /7 3/ Weeks of HFD 1/7 3/ Weeks of HFD Colonies (n) F G H I Lymphoid 15 *** ** 1 * * 5 Colonies (n) Myeloid 1 * Colonies (n) Granulocytes Colonies (n) Monocytes 6 * 4 2 1/7 3/ Weeks of HFD 1/7 3/ Weeks of HFD 1/7 3/ Weeks of HFD 1/7 3/ Weeks of HFD Figure 2. DIO shifts early HSPCs towards a more mature HSPC population with higher differentiation potential. (a) Schematic overview of HSPC maturation in the BM. (b) FACS data of LT HSC, (c) ST HSC, (d) E MMP, (e) L MMP of all LSK cells showing a decrease of ST HSC after 4, 1 and 18 weeks of the obesogenic diet, together with an increase in E MMP cells after 3 days, and 2 to 18 weeks of HFD. (f i) CFU assays of BM from mice on an obesogenic diet showing increased potential to form lymphoid and myeloid colonies. n=1 11, *=P<.5, **=P<.1, ***=P<.1 (1 way ANOVA with Tukey posttest analysis and the change per group expressed relative to the weeks of HFD group). When culturing BM from the obesogenic fed mice in a CFU assay ex vivo, the increase in differentiation potential further pushed the cells towards differentiation, as indicated by an increase in lymphoid colony forming potential of the BM of mice receiving an obesogenic diet which was already increased after 3 days and continued to increase over the course of obesity (Fig. 2F). The myeloid colony forming potential of BM cells increased after a more prolonged duration of the obesogenic diet, resulting in an increase in myeloid and monocyte forming colonies after 18 weeks of HFD feeding, when obesity was established (Fig. 2G I). HSPCs of obesogenic mice exhibit decreased proliferation Differentiating LSK cells loose Sca 1 (LK cells) and/or c Kit (LS cells) and start to express markers for common myeloid progenitors (CMPs; Lin Sca 1 c Kit + CD34 + CD16/32 low ) or common lymphoid progenitors (CLP; Lin Sca 1 + c Kit CD127 + ) (Fig. 3A for an overview) [1, 2]. 88

91 DIO in mice decreases HSPCs in the BM A B LSK LSK cells in BM Primitivity Lin - Sca-1 + c-kit + LK LS Lin - c-kit + Lin - Sca-1 + % Sca-1 + c-kit + (of lin - ) CMP CLP. 1/7 3/ Weeks of HFD CD34 + CD16/32 + CD127 + C D LK cells in BM LS cells in BM 5 4 % c-kit + (of lin - ) ** * *** % Sca-1 + (of lin - ) * 1/7 3/ /7 3/ Weeks of HFD Weeks of HFD E F CMP cells in BM CLP cells in BM % CD34 + CD16/32 + (of LK cells) /7 3/ * % CD127 + (of LSK cells) *** 1/7 3/ Weeks of HFD Weeks of HFD Figure 3. LSK cell maturation in BM of obesogenic mice. (a) Schematic overview of LSK cell maturation. (b) The percentage LSK cells of all Lin cells in the BM did not differ in the course of DIO. (c) Decreased myeloid progenitors (LK cells) after 4, 1 and 18 weeks and (d) a decrease in lymphoid progenitors (LS cells) after 18 weeks of obesogenic diet. (e) A decrease in CMP was seen after 18 weeks and (f) a decrease in CLP after 1 weeks. n=1 11, *=P<.5, **=P<.1, ***=P<.1 (1 way ANOVA followed by Tukey post test analysis and the change per group expressed relative to the weeks of HFD group). 89

92 Chapter 5 Although we observed a shift within the total LSK population from early HSPC populations towards a more differentiation capable progenitor cell, which we confirmed by the increased lymphoid and myeloid CFU potential, the fraction of LSK, LK and LS cells, as well as the fraction of CLP and CMP remained stable or even decreased over the course of obesity (Fig. 3B F). This observation made us hypothesize that although the HSPCs show an increased differentiation into mature immune cells, they may suffer from exhaustion in an obesogenic environment. We therefore analyzed the proliferative and reconstitution capacity of HSPCs primed by an obesogenic diet. BrdU uptake in LSK cells in the BM was measured in mice that received the 45% obesogenic diet or a chow diet for 18 weeks. Interestingly, Lin, LSK as well as LK cells exhibited a lower proliferation rate in the HFD group compared to chow diet (Fig. 4A D). Furthermore, PI staining of Lin cells in the BM showed that in the obesogenic diet group, a higher percentage of cells was in the G/G1 phase of the cell cycle (Fig. 4e), and a lower percentage of cells was in the M phase (Fig. 4F,G). A B C % BrdU+ cells (of Lin - ) % proliferating lin - cells 8 6 ** 4 2 % BrdU + cells (of LSK) % proliferating LSK cells 25 2 * Gap 2 G2 Mitosis M D 5 Chow 18w HFD % proliferating LK cells E 1 Chow 18w HFD % proliferating LS cells DNA synthesis S G1 Gap 1 % BrdU + cells (of LK cells) ** % BrdU + cells (of LS cells) F 8 6 Chow * 18w HFD Chow Chow 18w HFD 18w HFD G Chow 18w HFD % of cells 4 2 * G/G1 S/G2 M Figure 4. Decreased proliferation of HSPCs in BM of obese mice after 18 weeks of obesogenic diet. FACS data showing a decrease in BrdU labelled (a) Lin cells, (b) LSK cells and (c) LK cells, but no differences in (d) LS cells between obesogenic and chow fed mice. (e) Schematic overview of different phases in cell division. (f) Cell cycle analysis of the BM by DNA staining showing a higher percentage of cells in the G/G1 resting phase and a lower percentage of cells in the G2/M dividing phase of the cell cycle in the HFD group compared to a chow diet. (g) A representative PI staining of the two experimental groups. n=9 1, *=P<.5, **=P<.1 (Student s t tests). 9

93 DIO in mice decreases HSPCs in the BM Gene expression of cell cycle regulators, Cyclin D2, E1 and A2, which regulate G1 progression, G1 S transition and S phase progression respectively [24 27], were decreased in obesogenic BM cells (Fig. 5A C). In line with this, tumor suppressor genes p15 and p27, which prevent cdk4 and cdk2 activation via cyclin D and cyclin E interactions, were upregulated in the obesogenic BM, all reflecting suppression of proliferation (Fig. 5D,E). Furthermore, cytokines Tnf, Il 1β and Il 6 (but not Ifnγ), factors that have been shown to suppress HSPC proliferation [16], were increased in the HFD BM, indicating alterations in the BM microenvironment that affect the proliferation of the HSPCs (Fig. 5F I). We did not observe differences in Scf, M csf, Gm csf or G csf or their receptors (Fig. 5J,K), although our data does suggest a slight increase in these growth factors upon a HFD, which may be sufficient to stimulate differentiation of BM cells. relative gene expression relative gene expression relative gene expression A B C D Chow Chow Chow Cyclin D2 p15 TNF 18w HFD * ** 18w HFD p=.7 18w HFD relative gene expression relative gene expression relative gene expression Chow Chow Cyclin E1 p27 IL-1b 18w HFD 1.5 * Chow 1.5 * ** 18w HFD 18w HFD relative gene expression relative gene expression Chow Chow Cyclin A2 IL-6 ** 18w HFD F G H I J relative gene expression E SCF M-CSF M-CSFr K relative gene expression 8 Chow 18w HFD * 18w HFD p=.5 GM-CSF GM-CSFr G-CSF G-CSFr relative gene expression Chow Chow 18w HFD IFNy 18w HFD Figure 5. An obesogenic diet for 18 weeks induced alterations in cell cycle regulator and cytokine gene expression in the BM. (a c) Gene expression of cyclins was increased in obesogenic BM compared to chow fed mice. (d,e) Gene expression of tumor suppressor genes p15 and p27 was decreased in obesogenic BM. (f i) Cytokine gene expression of TNF, IL 1β and IL 6, but not IFNγ, was increased in BM from HFD mice. (j,k) Gene expression of SCF, M CSF, GM CSF or G CSF or their receptors were not significantly increased in HFD BM. n=16 19 *=P<.5, **=P<.1 (Student s t tests). 91

94 Chapter 5 Obesogenic primed HSPCs have decreased reconstitution capacity The increase in differentiation potential of HSPCs in vivo in mice fed an obesogenic diet and a decreased proliferation capacity led to a decrease in LK and LS cells. Interestingly, the decrease in proliferation was lower than the decrease in cells. As the percentage of LSK cells decreased by 29%, the proliferation in this population was decreased by 25%. For the LK and LS cells, the decrease in cells was 52% and 23%, whereas the decrease in proliferation was 37% and 2.5% respectively. This indicates that the decrease in these progenitor cell populations is not solely due to a decrease in proliferation. Therefore, we hypothesized that an obesogenic environment can prime HSPCs affecting their potential to initiate multilineage reconstitution. To test this hypothesis, a cbmt was performed with BM obtained from either HFD or chow fed donor CD45.1 mice, which was mixed with BM obtained from chow fed donor CD45.2 mice, and then transplanted in chow fed CD45.2 recipients (Fig. 6A). Reconstitution of immune cells was analyzed every three weeks starting 4 weeks post BMT. We determined the percentage of CD45.1 cells within different leukocyte subsets, which had reconstituted from the BM of mice fed either an obesogenic or chow diet. We observed that the percentage of CD45.1 T cells, B cells, neutrophils, monocytes, Gr1 high and Gr1 low monocytes on all time points was lower in mice that received BM from an obesogenic donor than mice that received BM from a chow fed donor (Fig. 6B G). Furthermore, upon sacrifice, the percentage of CD8 + cytotoxic T cells, CD4 + CD25 + FoxP3 + regulatory T cells, monocytes and both M1 and M2 like macrophages originating from the obesogenic donor BM was lower in EpAT compared to the chow donor BM (Fig. 6H,I). This indicates that obesogenic primed BM is less capable of immune cell reconstitution in blood and tissues. As the recipient mice did not receive a HFD, these results show that HSPCs subjected to an obesogenic diet are primed and have long term central effects on HSPC biology. The shift we observed from an early quiescent self renewing HSPC population towards mature differentiating HSPCs, and a decreased potential to initiate multilineage reconstitution implies impaired self renewal capacity. 92

95 DIO in mice decreases HSPCs in the BM A Donors Bone marrow mixed 2:1 Lean CD Lean CD45.2 Recipients Lean CD45.2 B % CD45.1 (of CD3 + B22 - ) D % CD45.1 (of CD11b + Gr1 low ) F * ** *** ** 4w 6w 9w 12w Time ** ** ** * 4w 6w 9w 12w Time % CD45.1 (of B22 + CD3 - ) % CD45.1 (of CD11b + Ly6G - ) C E 8 Chow HFD G **** **** **** **** 4w 6w 9w 12w Time **** ** ** * 4w 6w 9w 12w Time Chow HFD Obese CD45.1 Lean CD45.2 Lean CD45.2 % CD45.1 (of CD11b + Ly6G + ) **** *** *** ** 4w 6w 9w 12w Time % CD45.1 (of CD11b + Gr1 + ) 8 Chow HFD *** ** ** ** 4w 6w 9w 12w Time H EpAT I EpAT % CD ** * % CD * * * * Chow HFD CD4 + CD8 + CD25 + FoxP3 + Mo Ma M1 M2 Figure 6. Less immune cell reconstitution of obesogenic primed BM in a cbmt. (a) Schematic overview of a cbmt in which CD45.1 BM from age matched 1 weeks HFD or 1 weeks chow fed donor mice was mixed with CD45.2 BM from chow donor mice and transplanted into chow CD45.2 recipients. (b) Blood FACS data showing that transplanting obesogenic primed BM decreased the reconstitution of CD T cells, (c) CD B cells, (d) CD monocytes, (e) CD neutrophils, (f) CD inflammatory Gr1 high monocytes and (g) CD Gr1 low monocytes compared to transplantation of chow BM. (h,i) Reconstitution of CD45.1 lymphoid and myeloid cells in EpAT. Mo=monocytes (CD11b + Ly6G ), Ma=macrophages (CD11b + F4/8 + ), M1=M1 like macrophages (CD11b + F4/8 + CD11c + CD26 ), M2=M2 like macrophages (CD11b + F4/8 + CD11c CD26 + ). n =14 *=P<.5, **=P<.1, ***=P<.1, ****=P<.1 (b g: two way ANOVA with matched values followed by bonferroni multiple comparison test. h,i: Student s t tests). 93

96 Chapter 5 Discussion Mild chronic inflammation associated with obesity is a stressor for HSPCs in the BM. Here we show that a 45% obesogenic diet causes a switch from quiescent to differentiating HSPCs. The loss of selfrenewing characteristics and a decrease in proliferation results in impaired multilineage reconstitution, revealing that an obesogenic environment has large effects on HSPC dynamics. In addition to the traditional comorbidities of obesity, studies in humans as well as animal models have demonstrated that obesity causes impaired immune function, leading to an increased susceptibility to infectious diseases [28 32]. Signals in the BM microenvironment regulate HSPC dynamics. The composition of the BM is therefore essential for a correct functioning of the hematopoietic system. An increase in visceral fat in obesity is positively correlated with BM adiposity [33]. Consequently, adipocyte infiltration and accumulation in the BM can affect hematopoietic maintenance and differentiation [34]. Mesenchymal stem cells (MSC) in the BM are also involved in hematopoietic niches and microenvironments. A HFD induces alterations in MSCs, including increased production of cytokines IL 1, IL 6, and TNF. These signals may influence the BM microenvironment and modulate hematopoiesis, and consequently hamper proper functioning of the immune system [35]. Besides dyslipidemia, cholesterol and glucose metabolism appear to be critical for proper maintenance of the HSPC population. Mice with defective cholesterol efflux pathways due to a deficiency of apolipoprotein E or abca1/abcg1 have increased levels of growth factors in the BM, which causes the HSPCs to proliferate and expand, leading to leukocytosis [36 38]. Interestingly, in a previous study by our group using a model of hypercholesterolemia in LDLR / mice, we also observed an increased expression of inflammatory cytokines in the BM, including Tnf, Il 1β and Il 6 [2]. However, here we show alterations in the HSPC population in DIO that opposite to our findings in hypercholesterolemic LDLR / mice. Where the 45% HFD caused decreased proliferation of the LSK cells and decreased reconstitution in a cbmt, hypercholesterolemia resulted in increased proliferation and increased reconstitution in a cbmt [2]. Using a highly similar experimental setup in the same lab, the major differences are the diet and mouse model with 45% kcal from fat in C57Bl/6 mice vs a hypercholesterolemic diet, containing.15% cholesterol, in LDLR / mice. Full knock out models for defective cholesterol efflux pathways have a different lipoprotein metabolism which causes systemic hypercholesterolemia but can also directly affect the HSPCs. Our data show that an obesogenic diet results in signals that causes HSPCs to differentiate; factors that are independent from the genotype of the HSPCs and that are dominant over signals that could stimulate proliferation. Furthermore, hypercholesterolemia alone is potentially a less severe stressor compared to the systemic inflammation seen in obesity, which does stimulate leukocytosis but does not deplete the HSPC pool. Two models of obesity, leptin deficient Ob/Ob mice, as well as C57Bl/6 mice on a 6% HFD for 2 weeks, show a relative increase in CMP and GMPs in the BM, with Ob/Ob mice also having an expanded HSPC population [5]. The 45% HFD in our milder obesity model resulted in a relative decrease in CLP and CMP cells and their lymphoid and myeloid progenitors (LK and LS cells). One week of 6% HFD has been shown to trigger a transient depletion of LT HSCs followed by a later increase in LT HSCs and MPPs [22]. The 45% HFD may not have been as robust as a 6% HFD to 94

97 DIO in mice decreases HSPCs in the BM deplete the LT HSC pool early in the course of HFD [22]. Furthermore, leukocytosis in our study is less severe than reported in Ob/Ob mice or in C57Bl/6 on a 6% HFD. However, in our model, after 4 weeks of HFD, we did observe a decrease in hematopoietic stem cells, possibly due to a gradual depletion of the HSPC pool. Nagareddy et al also show that lowering glucose with a sodium glucose cotransporter 2 inhibitor in type 2 diabetic obese mice does not correct leukocytosis, indicating that hyperglycemia by itself is not the major driver of leukocyte production [21]. Another difference in these diets is the carbohydrate content. As more kcal are coming from fat, the relative carbohydrate content is less (35% in our 45% diet versus 2% in the 6% HFD). To distinguish to contribution of different nutrients on HSPC dynamics would be interesting for future studies. When we transplanted BM from either lean or obese mice, competitively mixed with BM from lean mice, into lean recipient mice, the HFD primed BM was less capable of systemic as well as tissue leukocyte reconstitution. This indicates that priming of the HSPCs by a HFD cannot be reversed by a lean BM niche. In weight loss experiments, returning mice from a HFD to a chow diet results in normalization of body weight and glucose tolerance and normalizes the number of LT HSPCs [22]. However, the quantity of granulocyte, macrophages progenitors and Lin cells remained elevated in the BM, further suggesting long term, persisting, dietary effects on the HSPCs [22]. Within the LSK population, DIO induces a switch in cells that are capable of self renewal (the LT and ST HSCs) towards immune progenitor cells with a higher differentiation potential (E MPP and L MPPs), as demonstrated by a loss of CD15 expression. Our data on a relative increase in LSK cells with a higher differentiation potential is in line with data from Singer et al [22] who transplanted 4 sorted LSK cells from either lean or obese donors in lean recipient mice, which increased myelopoiesis in mice that received obese donor LSK cells. This shows that when transplanting an absolute number of LSK cells, the population also includes relatively more cells with maturation potential leading to an increased mature leukocytes in the circulation [22]. We observed an increase in ex vivo lymphoid and myeloid colony forming potential in the course of DIO. Taking cells out of their obesogenic BM environment possibly removes the suppressive effects on proliferation leaving the increased potential to differentiation, resulting in an increased number of colonies in our CFU assays. This increased differentiation combined with decreased proliferation in vivo, leads to a decreased supply and a reduction in progenitor cell populations. Interestingly, transplantation of obesogenic BM into a normolipidemic environment revealed a decrease in multilineage reconstitution of obesogenic primed HSPCs, most likely resulting from the increased quiescence of the HSPC. However, as we observed that the decrease in progenitor cell populations was not solely due to a decrease in proliferation, the additional effects observed after obesogenic diet priming may still result in an increased differentiation potential of the HSPCs. Our results and that of others show that the BM pool is altered in the low grade inflammatory state associated with obesity. Different nutritional and/or genetic models can induce a variety and even opposing effects on the HSPCs. Signals from different tissues act on the hematopoietic system, which are influenced by hypercholesterolemia, dyslipidemia and hyperglycemia. Which model or dietary component contains the most dominant signals remains to be determined and should be taken into account when speculating on a potential therapy in humans. 95

98 Chapter 5 The present study demonstrates that an obesogenic diet induces long term alterations in the hematopoietic system including loss of stemcellness and loss of self renewal, which may result in a depletion of the most primitive HSPCs. This may consequently cause disturbed immunological responses to infections and contribute to the persistence and/or progression of chronic inflammatory diseases. Furthermore, as an obesogenic diet induces long term, cell intrinsic alterations in HSPCs, obesity may hamper proper functioning of the immune system, even after successful weight loss. 96

99 DIO in mice decreases HSPCs in the BM Corresponding author Prof. Esther Lutgens, Department of Medical Biochemistry, Experimental Vascular Biology, Academic Medical Centre, University of Amsterdam, Meibergdreef 9, 115 AZ, Amsterdam, The Netherlands. +31 () e.lutgens@amc.uva.nl Conflict of interest The authors declare no conflict of interest. Acknowledgements We acknowledge the support from the Netherlands Organization for Scientific Research (NWO)(VICI grant to E.L.), the Rembrandt foundation (S.B., M.W., E.L.), the Dutch Heart Foundation (Dr. E. Dekker MD grant to T.S.), the Netherlands CardioVascular Research Initiative (CVON211 19) and the Deutsche Forschungsgemeinschaft (DFG) (SFB154 B4 to E.L. and SFB1123 A5 to E.L.). 97

100 Chapter 5 References [1] Lumeng CN, Saltiel AR Inflammatory links between obesity and metabolic disease. Journal of Clinical Investigation 121(6): [2] Ferrante AW The immune cells in adipose tissue. Diabetes Obes Metab 15(s3): [3] Ouchi N, Parker JL, Lugus JJ, Walsh K Adipokines in inflammation and metabolic disease. Nature Reviews Immunology 11(2): [4] Odegaard JI, Chawla A Alternative Macrophage Activation and Metabolism. Annu Rev Pathol 6(1): [5] Nagareddy Prabhakara R, Kraakman M, Masters Seth L, Stirzaker Roslynn A, Gorman Darren J, Grant Ryan W et al Adipose Tissue Macrophages Promote Myelopoiesis and Monocytosis in Obesity. Cell Metab 19(5): [6] King KY, Goodell MA Inflammatory modulation of HSCs: viewing the HSC as a foundation for the immune response. Nat Rev Immunol 11(1): [7] Mendelson A, Frenette PS Hematopoietic stem cell niche maintenance during homeostasis and regeneration. Nat Med 2(8): [8] Wilson A, Laurenti E, Oser G, van der Wath RC, Blanco Bose W, Jaworski M et al. 28. Hematopoietic Stem Cells Reversibly Switch from Dormancy to Self Renewal during Homeostasis and Repair. Cell 135(6): [9] Wilson A, Oser GM, Jaworski M, Blanco Bose WE, Laurenti E, Adolphe C et al. 27. Dormant and Self Renewing Hematopoietic Stem Cells and Their Niches. Ann N Y Acad Sci 116(1): [1] King KYG, Margaret A Inflammatory modulation of HSCs: viewing the HSC as a foundation for the immune response. Nat Rev Immunol 11(1):8. [11] Essers MAG, Offner S, Blanco Bose WE, Waibler Z, Kalinke U, Duchosal M et al. 29. IFNα activates dormant haematopoietic stem cells in vivo. Nature 458(724):5. [12] Zeng W, Miyazato A, Chen G, Kajigaya S, Young NS, Maciejewski JP. 26. Interferon γ induced gene expression in CD34 cells: identification of pathologic cytokine specific signature profiles. Blood 17(1): [13] Selleri C, Sato T, Anderson S, Young N, Maciejewski J Interferon γ and tumor necrosis factor α suppress both early and late stages of hematopoiesis and induce programmed cell death. J Cell Physiol 165(3): [14] Baldridge MT, King KY, Boles NC, Weksberg DC, Goodell MA. 21. Quiescent haematopoietic stem cells are activated by IFN γ in response to chronic infection. Nature 465(7299):5. [15] Moldenhauer A, Genter G, Lun A, Bal G, Kiesewetter H, Salama A. 28. Hematopoietic progenitor cells and interleukin stimulated endothelium: expansion and differentiation of myeloid precursors. BMC Immunol 9(1):56. [16] Mirantes C, Passegué E, Pietras EM Pro inflammatory cytokines: Emerging players regulating HSC function in normal and diseased hematopoiesis. Exp Cell Res 329(2): [17] Shahbazian LM, Quinton LJ, Bagby GJ, Nelson S, Wang G, Zhang P. 24. Escherichia coli pneumonia enhances granulopoiesis and the mobilization of myeloid progenitor cells into the systemic circulation. Crit Care Med 32(8): [18] Quinton LJ, Nelson S, Boé DM, Zhang P, Zhong Q, Kolls JK et al. 22. The Granulocyte Colony Stimulating Factor Response after Intrapulmonary and Systemic Bacterial Challenges. J Infect Dis 185(1): [19] Rodriguez S, Chora A, Goumnerov B, Mumaw C, Goebel WS, Fernandez L et al. 29. Dysfunctional expansion of hematopoietic stem cells and block of myeloid differentiation in lethal sepsis. Blood 114(19): [2] Seijkens T, Hoeksema MA, Beckers L, Smeets E, Meiler S, Levels J et al Hypercholesterolemia induced priming of hematopoietic stem and progenitor cells aggravates atherosclerosis. FASEB J 28(5): [21] Nagareddy PR, Murphy AJ, Stirzaker RA, Hu Y, Yu S, Miller RG et al Hyperglycemia promotes myelopoiesis and impairs the resolution of atherosclerosis. Cell metab 17(5): [22] Singer K, DelProposto J, Lee Morris D, Zamarron B, Mergian T, Maley N et al Diet induced obesity promotes myelopoiesis in hematopoietic stem cells. Mol Metab 3(6): [23] Kanneganti T D, Dixit VD Immunological complications of obesity. Nat Immunol 13(8): [24] Sherr CJ, Roberts JM Inhibitors of mammalian G1 cyclin dependent kinases. Genes Dev 9(1): [25] Ando K, Ajchenbaum Cymbalista F, Griffin JD Regulation of G1/S transition by cyclins D2 and D3 in hematopoietic cells. Proc Natl Acad Sci U S A 9(2): [26] Burger C, Wick M, Muller R Lineage specific regulation of cell cycle gene expression in differentiating myeloid cells. J Cell Sci 17(7): [27] Sherr CJ, Roberts JM CDK inhibitors: positive and negative regulators of G1 phase progression. Genes Dev 13(12): [28] Dixit VD. 28. Adipose immune interactions during obesity and caloric restriction: reciprocal mechanisms regulating immunity and health span. J Leukoc Biol 84(4): [29] Strandberg L, Verdrengh M, Enge M, Andersson N, Amu S, Önnheim K et al. 29. Mice Chronically Fed High Fat Diet Have Increased Mortality and Disturbed Immune Response in Sepsis. PLoS ONE 4(1):e765. [3] Falagas ME, Kompoti M. 26. Obesity and infection. Lancet Infect Dis 6(7): [31] Huttunen R, Syrjanen J Obesity and the risk and outcome of infection. Int J Obes (Lond) 37(3): [32] Adler BJ, Kaushansky K, Rubin CT Obesity driven disruption of haematopoiesis and the bone marrow niche. Nat Rev Endocrinol 1(12):

101 DIO in mice decreases HSPCs in the BM [33] Bredella MA, Torriani M, Ghomi RH, Thomas BJ, Brick DJ, Gerweck AV et al Vertebral Bone Marrow Fat Is Positively Associated With Visceral Fat and Inversely Associated With IGF 1 in Obese Women. Obesity (Silver Spring) 19(1): [34] Naveiras O, Nardi V, Wenzel PL, Hauschka PV, Fahey F, Daley GQ. 29. Bone marrow adipocytes as negative regulators of the haematopoietic microenvironment. Nature 46(7252): [35] Cortez M, Carmo L, Rogero M, Borelli P, Fock R A High Fat Diet Increases IL 1, IL 6, and TNF α Production by Increasing NF κb and Attenuating PPAR γ Expression in Bone Marrow Mesenchymal Stem Cells. Inflammation 36(2): [36] Yvan Charvet L, Pagler T, Gautier EL, Avagyan S, Siry RL, Han S et al. 21. ATP Binding Cassette Transporters and HDL Suppress Hematopoietic Stem Cell Proliferation. Science 328(5986): [37] Murphy AJ, Akhtari M, Tolani S, Pagler T, Bijl N, Kuo C L et al ApoE regulates hematopoietic stem cell proliferation, monocytosis, and monocyte accumulation in atherosclerotic lesions in mice. J Clin Invest 121(1): [38] Feng Y, Schouteden S, Geenens R, Van Duppen V, Herijgers P, Holvoet P et al Hematopoietic Stem/Progenitor Cell Proliferation and Differentiation Is Differentially Regulated by High Density and Low Density Lipoproteins in Mice. PLoS ONE 7(11):e

102 Chapter 5 Supplementary data A Monocytes B T cells % CD11b + Ly6G - (of CD45 + ) C /7 3/ Weeks of HFD Pro-inflammatory monocytes D % CD3 + (of CD45 + ) /7 3/ Weeks of HFD Cytotoxic T cells % Gr1 high (of CD45 + ) E /7 3/ Weeks of HFD Helper T cells F % CD8 + (of CD45 + CD3 + ) ** 1/7 3/ Weeks of HFD Regulatory T cells % CD4 + (of CD45 + CD3 + ) /7 3/ Weeks of HFD % CD25 + FoxP3 + (of CD4 + ) * * * 1/7 3/ Weeks of HFD Supplementary Figure 1. Spleen immune cell profile by FACS in a time course of DIO. (a) No changes in percentage of CD11b + Ly6G monocytes and (b) CD3 + T cells of the total leukocyte population (CD45 + ). (c) The pro inflammatory Gr1 high monocytes were unchanges in the spleen after an obesogenic diet. (d) After 18 weeks of the obesogenic diet CD8 + cytotoxic T cells were increased. (e) CD4 + T helper cells were unaltered, whereas (f) the regulatory CD4 + CD25 + FoxP3 + T cells decreased only after short term HFD. n=1 11, *=P<.5 (1 way ANOVA with Tukey post test analysis and the change per group expressed relative to the weeks of HFD group). 1

103 6 Blocking CD4 TRAF6 interactions by small molecule inhibitor ameliorates the complications of dietinduced obesity in mice Susan M. van den Berg*,1, Tom T.P. Seijkens*,1, Pascal J.H. Kusters 1, Barbara Zarzycka 2, Linda Beckers 1, Myrthe den Toom 1, Marion J.J. Gijbels 1,3,4, Antonios Chatzigeorgiou 5, Christian Weber 2,6, Menno P.J. de Winther 1, Triantafyllos Chavakis 5, Gerry A.F. Nicolaes 2, Esther Lutgens 1,6 1 Department of Medical Biochemistry, Experimental Vascular Biology, Academic Medical Centre, University of Amsterdam, The Netherlands. e.lutgens@amc.uva.nl 2 Department of Biochemistry, University of Maastricht, Maastricht, The Netherlands 3 Department of Pathology, Maastricht University, Maastricht, The Netherlands 4 Department of Molecular Genetics, Maastricht University, The Netherlands 5 Department of Clinical Pathobiochemistry and Institute for Clinical Chemistry and Laboratory Medicine, Medical Faculty, Technische Universität Dresden, Dresden, Germany 6 Institute for Cardiovascular Prevention (IPEK), Ludwig Maximilian s University, Munich, Germany * these authors contributed equally to this work. International Journal of Obesity 215; 39(5):782 79

104 Chapter 6 Abstract Immune processes contribute to the development of obesity and its complications, such as insulin resistance, type 2 diabetes mellitus and cardiovascular disease. Approaches that target the inflammatory response are promising therapeutic strategies for obesity. In this context, we recently demonstrated that the interaction between the co stimulatory protein CD4 and its downstream adaptor protein Tumor necrosis factor Receptor Associated Factor (TRAF) 6 promotes adipose tissue inflammation, insulin resistance and hepatic steatosis in mice in the course of diet induced obesity. Here, we evaluated the effects of a small molecule inhibitor (SMI) of the CD4 TRAF6 interaction, SMI , on the development of obesity and its complications in mice that were subjected to diet induced obesity (DIO). Treatment with SMI did not result in differences in weight gain, but improved glucose tolerance. Moreover, SMI treatment reduced the amount of CD45 + leukocytes in the epididymal adipose tissue by 69%. Especially the number of adipose tissue CD4 + and CD8 + T cells and macrophages were significantly decreased. Our results indicate that smallmolecule mediated inhibition of the CD4 TRAF6 interaction is a promising therapeutic strategy for the treatment of metabolic complications of obesity by improving glucose tolerance, by reducing the accumulation of immune cells to the adipose tissue as well as by skewing of the immune response towards a more anti inflammatory profile. 12

105 SMI mediated CD4 TRAF6 inhibition in obesity Introduction Obesity and its associated conditions, including insulin resistance, type 2 diabetes mellitus, and cardiovascular diseases (CVD), affect more than 1 million people worldwide, a number that is increasing and is expected to do so for the next decades [1, 2]. Chronic caloric excess may shorten healthy lifespan of humans by 5 2 years, which results in a tremendous socio economic burden [1]. Hence the development of novel therapeutic strategies for obesity and its related disorders is a public health priority. Over the past decades, chronic inflammation has been identified as the pathological substrate of obesity and its complications [1, 3, 4]. Elevated plasma levels of cytokines, such as TNF and IL 6, characterize obese subjects and show a positive correlation with the extent of metabolic dysfunction [5 9]. Obese adipose tissue (AT) exhibits hallmarks of an ongoing inflammatory response, such as immune cell infiltration and activation, as well as the presence of pro inflammatory cytokines and adipokines including TNF, IL 6, and leptin [5 9]. Moreover, chronic inflammation has a pivotal role in the development of obesity associated diseases, such as hepatic steatosis, type 2 diabetes mellitus and CVD [5 1]. Strategies that modulate the inflammatory response are therefore promising therapeutic modalities for obesity and its complications. The co stimulatory dyad CD4 and its ligand CD4L (CD154) have a well known role in immune cell activation and inflammation [11]. After binding of CD4L, CD4 recruits adaptor proteins, the Tumor necrosis factor Receptor Associated Factors (TRAFs) [11]. The intracellular domain of CD4 contains two binding sites for these proteins: a proximal site that binds TRAF2/3/5 and a distal site, which binds TRAF6 [11]. After binding of the TRAF proteins to CD4, signal transduction is initiated, which eventually results in the expression of inflammatory mediators [11]. Mice with a genetic deficiency in the CD4 TRAF6 interaction, but not the CD4 TRAF2/3/5 interaction, are protected against the development of neointima formation and atherosclerosis, both exponents of an ongoing inflammatory condition of the arterial wall [12, 13]. Accumulating evidence indicates that the CD4 CD4L dyad has an important role in obesity [6, 14 17]. For example, plasma levels of soluble CD4L correlate with the body mass index in obese subjects and decrease after bariatric surgery [14]. CD4 CD4L mediated interactions between adipocytes and AT immune cells, including T cells and macrophages, promote the expression of proinflammatory cytokines and chemokines that increase the recruitment of other inflammatory cells to the AT [15, 16]. Ligation of CD4 on adipocytes promotes insulin resistance by decreasing IRS 1 and GLUT4 expression [15, 16]. Genetic deficiency or antibody mediated inhibition of CD4L reduced the inflammatory and metabolic complications of diet induced obesity (DIO) [15, 16]. In contrast, deficiency of CD4 aggravates the complications of DIO [18 2]. To understand these opposing effects of CD4L and CD4, we previously investigated the role of different downstream adaptor proteins of the CD4 CD4L pathway. We found that deficiency of CD4 TRAF2/3/5 interactions increased metabolic and inflammatory complications of DIO, whereas deficiency of CD4 TRAF6 interactions rather improved AT inflammation and insulin resistance in DIO mice. Based on these data, we developed a small molecule inhibitor (SMI) against CD4 TRAF6 interactions, and we could show that treatment with SMI improved insulin resistance, hepatic steatosis and M1 13

106 Chapter 6 macrophage accumulation into the epididymal AT (EpAT), when treatment was initiated 6 weeks after the beginning of the diet [19]. Thus, inhibition of the CD4 TRAF6 interaction is a promising therapeutic strategy for obesity. Here we performed a more thorough analysis of the effects of SMI mediated CD4 TRAF6 inhibition in DIO. We investigated the effects of SMI , an analogue of with a lower IC5 [19], in DIO. In addition, we here administered the CD4 TRAF6 inhibitor after 12 weeks of high fat diet (HFD), when obesity was much more progressed and continued diet and treatment until week 18. Our data suggest that SMI mediated CD4 TRAF6 inhibition is a promising novel therapeutic approach in DIO related metabolic dysregulation. Materials and methods TRAF 6 C domain expression, purification and binding analysis His tagged TRAF6 C domain (residues ) was expressed in E. Coli using the pet21d expression vector (Novagen). Protein was purified by affinity chromatography, followed by gel filtration in running buffer (25 mm TRIS, 2mM NaCl and.5 mm TCEP). The direct binding between the TRAF6 C domain and SMI was measured by Surface Plasmon Resonance (SPR) (Biacore T2, GE Healthcare). TRAF6 C domain was immobilized on Sensor Chip CM5 using the amine coupling method. This reached a density of approximately 12, and 7,5 RU. SMI was dissolved in PBS buffer containing 5% v/v DMSO. All measurements were carried out at 25 C and with a flow rate of 5 ul min 1 in SPR running buffer (PBS,.5% Tween2, 5% DMSO, ph=7.4). Sensorgrams were corrected by subtracting the initial level of SPR signal before injection of the SMI or the TRAF6 C domain. Data were analysed using the BIAevaluation software. Equilibrium dissociation constants (Kd) were determined from a model of the steady state affinity (3 independent runs were averaged). In vitro macrophage culture Bone marrow (BM) cells were isolated from CD4 +/+, CD4 / mice (C57Bl/6 background)[21] as well as mice containing site directed mutagenesis for the respective CD4 TRAF binding domains (CD4 Twt, CD4 TRAF2/3/5 / and CD4 TRAF6 / ) [22] and cultured in RPMI supplemented with 15% L929 conditioned medium to generate BM derived macrophages. BM derived macrophages were activated with an agonistic CD4 antibody cocktail FGK45 and 5D12 (both 25 µg/ml, Bioceros BV) overnight, incubated with SMI for 1 hour and frozen for real time PCR analysis. Mice Male C57Bl/6 mice (n=24) were purchased from Charles River and maintained at the animal facility of the University of Amsterdam. Mice received a HFD (35% kcal carbohydrate, 45% kcal fat, 2% kcal protein, Special Diets Services, Witham, United Kingdom) for 18 weeks from the age of 7 weeks. After 12 weeks of HFD, mice were treated daily with SMI (1µmol kg 1)(n=12), or vehicle (.5% tween8, 5% DMSO in PBS) [19] (n=12) for 6 weeks. Mice had ad libitum access to food and water and were maintained under a 12 hour light dark cycle. Food intake and body weight were measured weekly. After the experimental procedure, mice were fasted overnight and subsequently euthanized. Blood was collected and organs were dissected and processed for flow cytometric and histological analysis. All experimental procedures were approved by the Animal Experimentation Ethics Committee of the University of Amsterdam. 14

107 SMI mediated CD4 TRAF6 inhibition in obesity Haematology and biochemical measurements Blood was obtained by cardiac puncture with EDTA filled syringes. Haematological analysis was performed on a ScilVet abc plus+ (ScilVet, Oostelbeers, The Netherlands). Insulin assay and calculation of HOMA IR Fasting insulin was measured in plasma by enzyme linked immunoabsorbent assay (Mercodia, Uppsala, Sweden) following manufacturers' protocol. Glucose levels were measured from whole blood upon sacrifice using a glucometer (Bayercontour, Basel, Switzerland). The homeostasis model assessment of insulin resistance (HOMA IR) was calculated using the following formula: HOMA IR = fasting glucose (mmol/l) fasting insulin (mu/l)/22.5. Glucose tolerance test One week before and 3 and 6 weeks after the initiation of SMI or vehicle treatment, a glucose tolerance test (GTT) was performed. 12h fasted mice were injected i.p. with glucose (1mg g 1, Sigma Aldrich, Zwijndrecht, the Netherlands). Glucose levels were measured from whole blood using a glucometer (Bayercontour, Basel, Switzerland) at times indicated in the figures. Real time PCR Total RNA was extracted using TRIzol (Invitrogen, Carlsbad, CA, USA) and reverse transcribed with an iscript cdna synthesis kit (Bio Rad, Veenendaal, the Netherlands). qpcr was performed using a SYBR green PCR kit (Applied Biosystems, Leusden, the Netherlands) on a ViiA7 real time PCR system (Applied Biosystems). The result is expressed as relative to the control group, which was assigned a value of 1. Flow cytometry EpAT and subcutaneous AT (ScAT) were removed, rinsed in PBS and minced into small pieces. Tissues were digested in a collagenase mixture (DMEM 2mM HEPES, Collagenase I and Collagenase XI, Sigma Aldrich, Zwijndrecht, the Netherlands) for 45 minutes at 37 C. The digested samples were passed through a 7 µm nylon mesh (BD Biosciences, Breda, the Netherlands). The suspension was centrifuged at 125 rpm for 6 minutes and the pelleted SVF was resuspended in FACS buffer. Erythrocytes in blood were removed by incubation with hypotonic lysis buffer (8.4 g of NH 4 Cl and.84 g of NaHCO 3 per liter of distilled water). To prevent non specific binding of antibodies to the Fc receptor, all cell suspensions were incubated with a CD16/32 antibody prior to labelling. CD3, CD8, CD25, FoxP3, F4/8, CD11b, CD11c, Gr 1 (ebioscience, San Diego, CA, USA) CD4, Ly6G (BD, Breda, the Netherlands) and CD26 (Biolegend, San Diego, CA, USA) antibodies were incubated with the indicated tissues. Staining was analysed by flow cytometry (FACSCanto II, BD Biosciences, Breda, The Netherlands) and FlowJo software version (Tree star). Histology Tissues were collected, fixed in 4% paraformaldehyde and embedded in paraffin. Liver steatosis and inflammation were graded on 4 μm thick haematoxylin eosin (H&E) stained sections. Immunohistochemistry on liver, EpAT and ScAT was performed for CD45 (BD, Breda, the Netherlands) and on EpAT for CD68 (Bio rad). Five μm frozen sections of the liver were stained with Oil red O (Sigma Aldrich, Zwijndrecht, the Netherlands). Organs were analysed by H&E staining. Morphometric analyses were performed using the Las4.1 software (Leica, Rijswijk, the Netherlands) 15

108 Chapter 6 and ImageJ software. Analyses were performed by an observer who was blinded for the experimental conditions. Statistical analysis Results are presented as mean ± SEM. Data were analysed by a Student s t test or Chi squared test using GraphPad Prism 5. software (GraphPad Software, Inc., La Jolla, CA, USA). P values <.5 were considered significant. 16

109 SMI mediated CD4 TRAF6 inhibition in obesity Results Characterization of SMI The CD4 TRAF6 SMI is specifically modelled for the CD4 TRAF6 binding site on the TRAF6 molecule, the molecular structure of SMI is shown in Figure 1a. Biacore analysis shows that SMI binds well to the part of the TRAF6 molecule in which the CD4 binding site is present (Figure 1b). In vitro, SMI dose dependently suppressed CD4 induced gene expression of IL 1β and IL 6 cytokines in BM derived macrophages (Figure 1c). Further functional specificity can be deducted from the following experiment: BM derived macrophages from CD4 +/+, CD4 /, CD4 Twt, CD4 TRAF2/3/5 / and CD4 TRAF6 / mice (mice that contain a site directed mutagenesis in the CD4 gene for the respective TRAF binding domain on CD4) were activated with the CD4 agonistic antibodies 5D12 and FGK45. When stimulating these macrophages overnight, CD4 +/+, CD4 Twt and CD4 TRAF2/3/5 / macrophages display a high expression of CCL2, whereas CCL2 was significantly reduced in CD4 / and CD4 TRAF6 / macrophages, proving that CD4 TRAF6 interactions are responsible for the decrease in CCL2 levels. When adding SMI , together with the CD4 activating antibody, CCL2 levels were also reduced in CD4 +/+, CD4 Twt and CD4 TRAF2/3/5 / macrophages, suggesting that the CD4 TRAF6 inhibiting SMI is specific for CD4 TRAF6 interactions (Figure 1d). A B C relative gene expression IL Concentration (μm) relative gene expression IL Concentration (μm) D relative gene expression 15 ** 1 5 C57Bl6 CCL2 CD4-Twt *** *** Ctrl CD4-/- CD4-T2/2/5-/- CD4-T6-/- Figure 1. Characterization of SMI (a) Molecular structure of SMI (b) Surface plasmon resonance sensogram of SMI which confirmed the direct binding of SMI to immobilized TRAF6 C domain. Data represent the average of three independent experiments. (c) Dose dependent inhibition of IL 1β and IL 6 gene expression in FGK45 (agonistic CD4 antibody) stimulated bone marrow derived macrophages. (d) CCL2 gene expression in bone marrow derived macrophages of the respective genotypes stimulated with the CD4 agonistic antibody FGK45 and 5D12. n =6 per group, **=p<.1, ***=p<.1. 17

110 Chapter 6 Small molecule treatment did not result in side effects We previously developed small molecule inhibitors of the CD4 TRAF6 interaction [19], and found no side effects for SMI In the present study, we administered SMI , an analogue of with a lower IC5, in a more progressed model of DIO where treatment with the CD4 TRAF6 inhibitor was started after 12 weeks of HFD and continued until week 18 of HFD. We did not observe any side effects. Daily treatment with SMI for the last 6 weeks of 18 weeks of HFD did not result in any changes in peripheral blood leukocyte counts or cell composition (Supplementary figure 1a f). Macroscopic and histopathological analysis revealed no abnormalities of the SMI treatment in more than 2 organs analysed, which included spleen, colon, small intestine, stomach, kidney, lung, heart and muscle. Alterations in adipocyte size in the adipose tissue During the course of DIO, both groups showed a significant gain in body weight (Figure 2a). However, there were no differences in body weight gain or food intake between the experimental groups (Figure 2a, Table 1). SMI treatment did not affect EpAT, ScAT or BAT weights (Table 1). Because SMI binds to CD4 TRAF6 we confirmed presence of CD4 in the adipose tissues and found an 11 fold higher expression of CD4 in EpAT compared to ScAT in the control group, indicating that the SMI can exert larger effects in EpAT (Figure 2b). Interestingly, although the weight of adipose depots did not differ, the number of adipocytes per field in EpAT of mice treated with SMI was increased by 15.3% revealing that SMI treatment decreased adipocyte size, suggesting less lipid storage and improved metabolic function (Figure 2c,d) [23]. A slight increase in number of adipocytes per field in ScAT did not reach statistical significance (Figure 2c). SMI improved glucose tolerance in DIO Next, we evaluated the effect of SMI treatment on glucose sensitivity. Basal levels of glucose and plasma insulin did not differ between the experimental groups and the HOMA IR indices were not different (Table 1). Glucose tolerance tests (GTT) were performed before and after 3 and 6 weeks of treatment. No differences between the groups were observed before treatment was initiated (data not shown). However, three weeks of treatment with SMI resulted in improved glucose sensitivity, compared with vehicle treated mice and an even more pronounced effect on glucose sensitivity was seen after 6 weeks of treatment (Figure 2e). These data indicate that SMI improves glucose sensitivity in DIO. Table 1. Characteristics of mice who received a HFD for 18 weeks and were treated with vehicle or SMI after 12 weeks of HFD and continued treatment until week 18. CTRL Food intake (g/wk) 2.9 ± ± 1.1 EpAT (g) 1.8 ± ±.2 ScAT (g) 1.1 ±.2.8 ±.1 BAT (mg) 142 ± ± 14 Glucose (mg/dl) 153 ± ± 5.4 Insulin (mu/l) 11. ± ± 1. HOMA IR 4.2 ± ±.4 n=12, *=p<.5. 18

111 SMI mediated CD4 TRAF6 inhibition in obesity A Body weight (g) Weeks of HFD CTRL B relative gene expression CD4 *** CTRL CTRL ScAT EpAT C # adipocytes /field EpAT * # adipocytes /field ScAT CTRL D Ctrl μm 5 μm E 3 weeks of SMI treatment 6 weeks of SMI treatment % of initial blood glucose * time (min) % of initial blood glucose ** *** *** * ** time (min) CTRL Figure 2. Daily treatment with SMI during the last 6 weeks of 18 weeks HFD reduced adipocyte size and improved glucose sensitivity in mice. (a) Treatment with SMI did not result in differences in body weight gain. (b) Gene expression of CD4 showing a higher expression of CD4 in EpAT compared to ScAT. (c) Adipocyte size of EpAT and ScAT showing a 15.3% decrease in epididymal adipocyte size of SMI treated mice as indicated by an increase in adipocyte numbers per field. n=12 per group. (d) H&E staining illustrating that epididymal adipocyte size was decreased in SMI treated mice compared to vehicle. (e) GTT of mice fed a HFD for 15 weeks with 3 weeks of SMI treatment and GTT of mice fed a HFD for 18 weeks with 6 weeks of treatment showing improved glucose sensitivity in mice treated with SMI compared to vehicle. n =6 per group, *=p<.5, **=p<.1, ***=p<.1. 19

112 Chapter 6 SMI does not induce differences in hepatosteatosis As liver steatosis is an important metabolic complication, we assessed liver weights, hepatic inflammation and degree of steatosis. However, in this relatively mild model of DIO (45% calories from fat) we did not observe differences in liver weight between the groups (Figure 3a) or severe inflammation in either group (Figure 3b,c). Furthermore, there were no differences in hepatic lipid content and the degree of steatosis (Figure 3d f). A Liver weight B Leukocytes C Inflammation weight (mg) # CD45+ cells /field % of mice / CTRL CTRL CTRL D E F Lipid content Macro steatosis Micro steatosis Oil red O-stained area (%) CTRL % of mice 1 5 CTRL / % of mice 1 5 CTRL / Figure 3. No liver abnormalities were seen in vehicle or SMI treated groups. (a) Liver weights did not differ. (b) Immunohistochemical leukocyte (CD45) staining and (c) scoring of hepatic inflammation by H&E staining did not show significant differences between the two experimental groups. (d) Oil red O staining of liver sections and (e,f) scoring of hepatosteatosis did not show differences in degree of steatosis or lipid content. n=12 per group. Amelioration of adipose tissue inflammation after compound treatment As CD4 is important in inflammation and AT inflammation has a key role in the development of insulin resistance, we analysed immune cell composition by flow cytometry. We observed a 68.5% decrease in number of total leukocytes (CD45 + cells) in the EpAT of mice treated with SMI (Figure 4a). Leukocyte subset analysis revealed that T helper cell numbers (CD3 + CD4 + cells) were decreased by 5.4% and cytotoxic T cell numbers (CD3 + CD8 + cells) were decreased by 61.6%, whereas the numbers of regulatory T cells (CD3 + CD4 + CD25 + FoxP3 + cells) were not affected (Figure 4b d). No differences in ScAT T cell numbers were observed (Supplemental Figure 2a d). Additionally, the number of EpAT macrophages (CD11b + Ly6G F4/8 + cells) was decreased by 81.7% by SMI (Figure 4e). Macrophage subset analysis showed that pro inflammatory M1 macrophages (CD11b + Ly6G F4/8 + CD11c + CD26 cells) were decreased by 8.8% and antiinflammatory M2 macrophages (CD11b + Ly6G F4/8 + CD11c CD26 + cells) by 7.5% (Figure 4f,g). The pro inflammatory M1 macrophages showed a larger decrease than the M2 macrophages, therefore, the ratio between M1 and M2 macrophages was 2.4 in the control group whereas the treatment resulted in a ratio 1.5, suggesting a more anti inflammatory profile of the epididymal adipose tissue depot (Figure 4h). No aberrations in ScAT macrophage numbers were observed (Supplemental Figure 2e g). 11

113 SMI mediated CD4 TRAF6 inhibition in obesity A EpAT Leukocytes B T helper cells # of CD45+ cells * # of CD3+CD4+ cells ** CTRL C Cytotoxic T cells D Regulatory T cells # of CD3+CD8+ cells * # of CD4+CD25+FoxP3+ cells # of CD11b+F4/8+ cells E Macrophages ** # of F4/8+CD11c+CD26- cells F M1 macrophages * # of F4/8+CD11c-CD26+ cells G M2 macrophages * H M1:M2 ratio Ratio M1:M2 macrophages Figure 4. Flow cytometric analysis demonstrates that SMI treatment reduces AT leukocyte count. (a) Flow cytometry showed decreased leukocyte numbers (CD45 + cells) in the EpAT of SMI treated mice. (b) Lymphocyte subset analysis revealed that especially T helper (CD3 + CD4 + ) cells and (c) cytotoxic (CD3 + CD4 CD8 + ) T cells were decreased, whereas (d) regulatory (CD4 + CD25 + FoxP3 + ) T cells were not affected. (e) The number of total macrophages (CD11b + F4/8 + ), as well as (f) pro inflammatory M1 (CD11b + F4/8 + CD11c + CD26 ) and (g) antiinflammatory M2 (CD11b + F4/8 + CD11c CD26 + ) macrophages was decreased by treatment with SMI (h) The ratio between M1 and M2 macrophages. n=12 per group, *=p<.5, **=p<.1. Furthermore, immunohistochemical analysis of the immune cells in EpAT showed decreased amounts of leukocytes (CD45 + ) in the SMI treated group, which was not seen in ScAT (Figure 5a,b). Additionally, we observed a reduction in macrophage number (CD68 + ) relative to the number of adipocytes in EpAT of the SMI treated group (Figure 5c,d). Together our data demonstrate that treatment with SMI ameliorated diet induced AT inflammation. 111

114 Chapter 6 Figure 5. Immunohistochemical assessment of adipose tissue leukocytes demonstrating reduced infiltration of total leukocytes and macrophages into AT of SMI treated mice. (a) Immunohistochemical analysis of EpAT and ScAT by leukocyte (CD45) staining showing a decreased amount of leukocytes in the EpAT of SMI treated mice. No differences were seen in leukocyte counts in ScAT of SMI treated mice. (b) CD45 staining of EpAT + illustrating the decreased amount of CD45 cells in SMI treated mice compared to vehicle. (c) Immunohistochemical analysis by macrophage (CD68) staining showing a decreased amount of macrophages in the + EpAT of SMI treated mice. (d) Pictures of a CD68 staining of EPAT to illustrate the decrease in CD68 cells in SMI treated mice. n=12 per group,**=p<