THE ROLE OF PPARα AND GROWTH HORMONE IN HEPATIC LIPID METABOLISM AND ATHEROSCLEROSIS. Anna Ljungberg

Transcription

1 THE ROLE OF PPARα AND GROWTH HORMONE IN HEPATIC LIPID METABOLISM AND ATHEROSCLEROSIS Anna Ljungberg Department of Physiology Wallenberg Laboratory for Cardiovascular Research The Sahlgrenska Academy at Göteborg University, 2006

2 A doctoral thesis at a university in Sweden is produced either as a monograph or as a collection of papers. In the latter case, the introductory part constitutes the formal thesis, which summarizes the accompanying papers. These papers have already been published or are in manuscripts at various stages (in press, submitted or in manuscript). Printed by Intellecta DocuSys Göteborg, Sweden, 2006 ISBN

3 ABSTRACT ABSTRACT Dyslipidemia mainly results from oversecretion of apob-containing lipoproteins from the liver and is one of the most important risk factors for the development of atherosclerosis. Growth hormone (GH) plays a key role in the regulation of lipoprotein metabolism and thus, disturbances in GH secretion are associated with dyslipidemia and cardiovascular disease. GH influences the activity of the nuclear hormone receptor peroxisome proliferator-activated receptor (PPAR)α, which regulates genes involved in lipid metabolism through interaction with coactivators. PPARγ coactivator-1 (PGC-1) has been shown to coactivate several transcription factors, including PPARα. GH transgenic mice on apoe-deficient background had larger atherosclerotic lesion area in the thoracic aorta and more advanced lesions in aortic sinus compared to littermate controls. Changes in serum lipoproteins were most likely not involved in the accelerated lesion formation in GH transgenic mice since their lipoprotein profile was not worsened. Instead, higher blood pressure and an increased inflammatory response could contribute to this effect. Treatment of mice with a high dose of GH increased triglyceride secretion, serum apob and cholesterol levels, whereas liver triglycerides were reduced. Most of the studied effects of GH were similar in PPARα-deficient and wild-type mice and thus independent of PPARα. However, hepatic PPARγ2 and Cyp4a10 mrna expression were PPARα dependent, indicating that PPARα is important for the effect of GH on hepatic PPARγ signaling and ω-oxidation. Treatment of mice and incubation of mouse hepatocytes with the PPARα agonist Wy14,643 (Wy) resulted in hepatic triglyceride accumulation in parallel with increased expression of adipose differentiation-related protein (ADRP). Studies in mouse hepatocytes showed that the increased triglyceride content was associated with inhibited triglyceride secretion. ADRP overexpression also resulted in accumulation of triglycerides and decreased triglyceride secretion. The decreased secretion was not due to lack of triglycerides. Rather, PPARα activation prevents the availability of cytosolic triglycerides for VLDL assembly, in part by increasing the expression of ADRP. Hepatic overexpression of PGC-1β in mice induced a hyperlipidemic response with a marked increase in apob-containing lipoproteins. The hyperlipidemia was associated with increased diacylglycerol acyltransferase (DGAT)-1 expression and plasma free fatty acids. The potentially beneficial effects of Wy on genes controlling lipid metabolism were blunted by PGC-1β overexpression. In summary, high GH levels are associated with increased atherosclerosis and hepatic VLDL secretion. A few hepatic GH effects are dependent on PPARα. PPARα activation decreases hepatic VLDL secretion by compartmentalization of intracellular triglycerides and hepatic PGC-1β overexpression results in combined hyperlipidemia and decreased PPARα signaling. 3

4 LIST OF PUBLICATIONS LIST OF PUBLICATIONS This thesis is based on the following papers, which will be referred to in the text by their Roman numerals: I II III IV Increased atherosclerotic lesion area in apoe deficient mice overexpressing bovine growth hormone. Irene J. Andersson*, Anna Ljungberg*, Lennart Svensson, Li-Ming Gan, Jan Oscarsson and Göran Bergström. Atherosclerosis, In press Importance of PPARα for the effects of growth hormone on hepatic lipid and lipoprotein metabolism. Anna Ljungberg, Daniel Lindén, Caroline Améen, Göran Bergström and Jan Oscarsson. Manuscript PPARα activation increases triglyceride mass and adipose differentiation related protein in hepatocytes. Ulrika Edvardsson, Anna Ljungberg, Daniel Lindén, Lena William-Olsson, Helena Peilot-Sjögren, Andrea Ahnmark and Jan Oscarsson. Journal of Lipid Research 47: , 2006 Hepatic PGC-1β overexpression in mice causes combined hyperlipidemia and a blunted response to PPARα activation. Anna Ljungberg, Christopher Lelliott, Andrea Ahnmark, Lena William- Olsson, Anders Elmgren, Jan Oscarsson and Daniel Lindén. Manuscript * Both authors contributed equally to this article 4

5 ABBREVIATIONS ABBREVIATIONS AADA ACO ADRP ALAT AngPtl Apo BMI CAR CETP CPT CRP CVD CYP DGAT DR ER FAS FATP FFA GH GHBP GHD GHR GHRH HEK HBSS HDL HL Hx IDL IGF IMT JAK arylacetamide deacetylase acyl-coa oxidase adipose differentiation-related protein alanine aminotransferase angiopoietin-like protein apolipoprotein body mass index coxsackie-adenovirus receptor cholesterol ester transfer protein carnitine palmitoyl transferase c-reactive protein cardiovascular disease cytochrome P450 diacylglycerol acyltransferase direct repeat endoplasmic reticulum fatty acid synthase fatty acid transport protein free fatty acids growth hormone growth hormone binding protein growth hormone deficiency growth hormone receptor growth hormone releasing hormone human embryonic kidney Hanks balanced salt solution high-density lipoproteins hepatic lipase hypophysectomized intermediate-density lipoproteins insulin-like growth factor intima media thickness janus kinase 5

6 ABBREVIATIONS LCAD LDL LDLR LFABP LPL LRP LXR MCAD MLDP MTP NEFA PGC PLTP PPAR PPRE PRC RNAi RXR SAA SCAD SCD SMC SREBP STAT TGH VLDL long-chain acyl-coa dehydrogenase low-density lipoproteins low-density lipoprotein receptor liver fatty acid binding protein lipoprotein lipase low-density lipoprotein receptor-related protein liver X receptor medium-chain acyl-coa dehydrogenase myocardial lipid droplet protein microsomal triglyceride transfer protein non-esterified fatty acids peroxisome proliferator-activated receptor γ coactivator phospholipid transfer protein peroxisome proliferator-activated receptor peroxisome proliferator response element PGC-1 related coactivator RNA interference retinoid X receptor serum amyloid A short-chain acyl-coa dehydrogenase stearoyl-coa desaturase smooth muscle cell sterol regulatory element binding protein signal transducer and activator of transcription triacylglycerol hydrolase very low-density lipoproteins 6

7 TABLE OF CONTENTS TABLE OF CONTENTS ABSTRACT... 3 LIST OF PUBLICATIONS... 4 ABBREVIATIONS... 5 INTRODUCTION... 9 GENERAL INTRODUCTION... 9 LIPOPROTEIN METABOLISM The exogenous lipoprotein pathway The endogenous lipoprotein pathway Assembly and secretion of VLDL Regulation of VLDL secretion The link between dyslipidemia and atherosclerosis ATHEROSCLEROSIS Atherogenesis PEROXISOME PROLIFERATOR-ACTIVATED RECEPTORS PPARα and lipid metabolism The phenotype of PPARα-deficient mice PPARγ-COACTIVATOR Hepatic functions of PGC-1α and PGC-1β ADIPOSE DIFFERENTIATION-RELATED PROTEIN GROWTH HORMONE Disturbances in GH secretion GH and lipoprotein metabolism The phenotype of bgh transgenic mice GH and atherosclerosis AIMS OF THE THESIS METHODOLOGICAL CONSIDERATIONS GENETICALLY MODIFIED MICE ApoE-deficient mice overexpressing bgh PPARα-deficient mice HORMONE AND DRUG TREATMENT IN VIVO GH Wy14, HEPATOCYTE CULTURES ADENOVIRAL-MEDIATED OVEREXPRESSION HEPATIC TRIGLYCERIDE SECRETION IN VIVO ApoB MEASUREMENTS Electroimmunoassay Immunoprecipitation

8 TABLE OF CONTENTS QUANTITATIVE REAL-TIME PCR ATHEROSCLEROTIC MEASUREMENTS The en face method The cross-sectional method SUMMARY OF RESULTS PAPER I PAPER II PAPER III PAPER IV DISCUSSION GH AND ATHEROSCLEROSIS Possible mechanisms behind the accelerated atherosclerosis progression in bgh transgenic mice GH AND HEPATIC LIPID METABOLISM Effects of GH on VLDL-triglyceride secretion Effects of GH on liver triglycerides and PPARγ PPARα AND HEPATIC LIPID METABOLISM Effects of PPARα activation on apob-containing lipoproteins Effects of PPARα-deficiency on apob-containing lipoproteins Do triglycerides accumulate in human liver upon PPARα activation? PGC-1β AND HEPATIC LIPID METABOLISM Possible mechanisms behind the combined hyperlipidemia PGC-1β as a potential pharmaceutical target? SUMMARY AND CONCLUSIONS ACKNOWLEDGEMENTS REFERENCES

9 INTRODUCTION INTRODUCTION GENERAL INTRODUCTION The metabolic syndrome, also known as the insulin resistance syndrome or syndrome X, is a clustering of risk factors highly associated with development of type 2 diabetes and cardiovascular disease (CVD). Due to the complexity of the metabolic syndrome, committees worldwide have not been able to agree upon standard definitions for the syndrome. However, according to most definitions the metabolic syndrome includes at least some of the following features: abdominal obesity, impaired glucose tolerance, insulin resistance, high plasma triglycerides, low HDL cholesterol or hypertension [1]. The prevalence of the metabolic syndrome is rapidly increasing and this may be due to the global epidemics of obesity. The alarming reports about increased obesity in children and young people are therefore of major concern. Furthermore, the metabolic syndrome and its complications are no longer a health problem only in western society but also in developing countries. It is predicted that the number of diabetics will increase from 175 million in 2000 to 353 million in 2030, with the largest increase expected in China and India [2]. First-line strategies for managing the metabolic syndrome are lifestyle changes, such as physical activity, dietary modifications and weight reduction. The alternative strategy is pharmacological treatment of the individual components of the metabolic syndrome, for instance with agonists to the different peroxisome proliferator-activated receptors (PPARs). PPARα agonists improve plasma triglyceride and high-density lipoprotein (HDL) cholesterol levels, whereas PPARγ agonists reduce blood glucose and improve insulin resistance. In addition, dual agonists, affecting both PPARα and PPARγ, are under development. Dyslipidemia is not only a component of the metabolic syndrome. Disturbances in the secretion of growth hormone (GH) have several effects on lipid and lipoprotein metabolism with cardiovascular consequences. Both GH-deficiency and GH-excess (acromegaly) are conditions associated with increased mortality due to CVD [3, 4], emphasizing the importance of an accurate GH secretion. Disturbed GH secretion is also observed both in obesity and type 2 diabetes [5]. The liver is central in metabolic regulation and abnormal hepatocyte function is related to several human diseases. Accumulation of fat in the liver (steatosis) and overproduction of atherogenic lipoproteins are disturbances that are highly associated with obesity, type 2 diabetes and atherosclerosis. 9

10 INTRODUCTION LIPOPROTEIN METABOLISM Triglycerides are the major energy source in the body and cholesterol is a precursor of steroid hormones, bile acids and vitamin D. In addition, cholesterol is an essential component of cell membranes. Due to their hydrophobic nature, triglycerides and cholesterol are transported in the circulation as complex particles called lipoproteins. A typical lipoprotein particle consists of a core of triglycerides and cholesterol esters, with an outer surface monolayer of amphipathic phospholipids, unesterified cholesterol and apolipoproteins. There are five major classes of lipoproteins; chylomicrons, very low-density lipoproteins (VLDL), intermediate-density lipoproteins (IDL), low-density lipoproteins (LDL) and HDL, which all differ in particle size, density, lipid composition and apolipoprotein (apo) content (Table 1). Table 1. Density, size, lipid composition and apolipoprotein content of lipoproteins. Chylomicrons VLDL IDL LDL HDL Density (g/ml) < Size (nm) Triglycerides C + CE Phospholipids Protein Apolipoproteins AI, AII, AIV, B48, CI, CII, CIII, E B48*, B100, CI, CII, CIII, E B100, E B100 AI, AII, CI, CII, CIII, E The composition of triglycerides, cholesterol (C) + cholesterol esters (CE), phospholipids and protein is expressed as % of dry weight. * Humans secrete apob100-vldl only, whereas rodents secrete VLDL containing either apob48 or apob100. All lipoproteins are associated with at least one apolipoprotein that stabilizes the structure of the lipoprotein particle. In addition, some apolipoproteins serve as ligands for specific receptors and regulate enzymes involved in lipoprotein metabolism. ApoB, which is the largest apolipoprotein, exist in two forms: apob100 and apob48. ApoB100 is the full-length protein, whereas apob48 is the truncated form representing the N-terminal 48% of apob100 [6, 7]. In humans, apob100 is produced in the liver and is an essential part of VLDL, whereas apob48 is produced in the intestine and is necessary for chylomicron production. Several species, including rodents, produce both apob48 and apob100 in the liver and therefore secrete VLDL particles containing either form. 10

11 INTRODUCTION The exogenous lipoprotein pathway Chylomicrons are responsible for the exogenous lipoprotein pathway where dietary lipids are transported from the small intestine to other tissues (Figure 1). The large amounts of triglycerides carried in chylomicrons are hydrolyzed by the action of lipoprotein lipase (LPL), which is bound via proteoglycans to the surface of endothelial cells in the capillaries. As chylomicrons pass through the capillaries, they bind to LPL via interaction with apocii at the surface of the particle. ApoCII activates LPL and the triglycerides are hydrolyzed to free fatty acids (FFA) that are taken up in the tissues, either by passive diffusion or by facilitated transport via fatty acid transporters. Most of the fatty acids are then consumed by oxidation in muscle and heart or directed for re-esterification and storage in adipose tissue. The progressive hydrolysis of triglycerides in chylomicrons results in particles of smaller size called chylomicron remnants. During this process, excess surface material in terms of phospholipids, unesterified cholesterol and apolipoproteins are passed on to other particles, mainly HDL, by phospholipid transfer protein (PLTP). Furthermore, triglycerides are transferred to HDL in exchange for cholesterol esters by the action of cholesterol ester transfer protein (CETP). This does not occur in mice since they lack CETP. The chylomicron remnants are cleared from the circulation by hepatic uptake via the LDL-receptor (LDLR) or LDL-receptor related protein (LRP) that recognizes apoe at the surface of the remnant. Mice deficient in apoe therefore accumulate chylomicron remnants in the circulation. INTESTINE Exogenous pathway SR-BI Endogenous pathway HDL Cholesterol removal LDL Cholesterol delivery LDLR HL PERIPHERAL TISSUE Chylomicron LDLR/LRP VLDL Chylomicron remnant IDL LPL LPL FFA Surface material to HDL FFA Surface material to HDL ADIPOSE TISSUE MUSCLE ADIPOSE MUSCLE TISSUE Figure 1. The exogenous and endogenous lipoprotein pathways. 11

12 INTRODUCTION The endogenous lipoprotein pathway The distribution of lipids from the liver to peripheral tissues is referred to as the endogenous lipoprotein pathway (Figure 1). Triglyceride-rich VLDL particles are secreted from hepatocytes in the liver and, similar to chylomicrons, they are subjected to LPL stimulated hydrolysis in the circulation. The resulting IDL particles are either taken up in the liver by LDL receptors or further metabolized to LDL by hepatic lipase (HL). LDL transport cholesterol to peripheral tissues. In contrast, HDL is involved in the reverse cholesterol transport, where excess cholesterol in peripheral tissues is transported back to the liver for excretion into the bile. HDL has also been shown to possess other beneficial functions, like antioxidative and anti-inflammatory action [8]. Assembly and secretion of VLDL The assembly of VLDL involves a two-step process that takes place in the endoplasmic reticulum (ER) and the Golgi apparatus [9-11]. ApoB is synthesized on ribosomes attached to the rough ER membrane and the polypeptide chain is translocated across the membrane into the lumen. In the first step, a small amount of triglycerides are transferred to apob during translocation, producing a pre- VLDL particle. Simultaneously, a VLDL-sized lipid droplet lacking apob is produced in the smooth ER. The formation of these two precursors is dependent on microsomal triglyceride transfer protein (MTP) activity [12, 13]. The second step involves fusion of the pre-vldl particle and the lipid droplet, producing mature VLDL, which is then transported through the Golgi apparatus and finally secreted from the cell membrane of the hepatocyte. Although less studied, the assembly and secretion of chylomicrons from the enterocytes is believed to involve similar mechanisms. Regulation of VLDL secretion Despite extensive research, regulation of VLDL secretion is still not completely understood. However, MTP activity, availability of lipids and insulin signaling are important factors controlling VLDL secretion. The crucial role of MTP for the assembly and secretion of apob-containing lipoproteins has been confirmed in several studies. Perhaps the most striking evidence is the genetic disease abetalipoproteinemia in which lack of MTP results in severely deficient secretion of apob-containing lipoproteins, both from the liver and intestine [14-16]. Furthermore, liver-specific disruption [12] and heterozygous knockout of MTP [17] as well as the use of MTP inhibitors [18] results in decreased lipoprotein secretion, whereas adenoviral overexpression of MTP has the opposite effect [19, 20]. The function of MTP in transferring lipids to the forming VLDL particle is highly dependent on the availability of lipids in the ER. Poor lipidation of apob results in misfolding of the protein and termination of the VLDL assembly. This is usually accomplished by retrograde translocation of apob back to the cytosol, followed by ubiquitination and proteasomal degradation [21, 22]. 12

13 INTRODUCTION Triglycerides used for VLDL assembly originate from three different sources; de novo lipogenesis, remnant lipoproteins and non-esterified fatty acids (NEFA) mainly derived from lipolysis in adipose tissue [23]. It has been shown that plasma NEFA is the major source for triglyceride production for VLDL secretion, whereas the contribution of de novo lipogenesis is limited under normal conditions [24]. However, nutritional and hormonal status, such as carbohydrate-rich diet and hyperinsulinemia, may change the importance of de novo lipogenesis. Irrespective of the origin, it is believed that triglycerides incorporated into VLDL are recruited from a cytosolic storage pool [25, 26]. Recruitment from this triglyceride pool is accomplished by lipolytic mobilization followed by re-esterification in the ER membrane and subsequent lipidation of apob via MTP. The mobilization of triglycerides for VLDL secretion involves triacylglycerol hydrolase (TGH) [27] and possibly also arylacetamide deacetylase (AADA) [28]. Another key enzyme in this process of triglyceride storage and mobilization is diacylglycerol acyltransferase (DGAT). DGAT is a membrane-bound enzyme that catalyzes the last step in triglyceride synthesis by converting diacylglycerol and fatty acyl-coa into triglycerides. DGAT was first thought to be located exclusively on the cytosolic side of the ER membrane, but was later found also on the luminal side [29]. The lumen-facing form (DGAT1) and the cytosol-facing form (DGAT2) are products from different genes. Several studies have provided evidence that DGAT2 catalyzes the production of triglycerides that are destined for the cytosolic storage pool, whereas DGAT1 catalyzes the production of triglycerides that are incorporated into VLDL [29, 30]. Thus, adenoviral overexpression of DGAT1 in mice mainly results in increased VLDL secretion and particle size, whereas overexpression of DGAT2 mainly results in accumulation of triglycerides in the liver [30]. Furthermore, overexpression of DGAT1 in rat hepatoma cells has been shown to increase the secretion of VLDL [31]. The relative activity of DGAT1 and DGAT2 may therefore be an important determinant for the degree of hypertriglyceridemia and hepatic steatosis, respectively. Overproduction of VLDL is closely associated with insulin resistance and appears to result from decreased sensitivity to the inhibitory effects of insulin on VLDL secretion. For instance, MTP gene expression is normally suppressed by insulin [32], but failure of this inhibition in the insulin-resistant state is thought to facilitate VLDL assembly. Furthermore, increased lipolysis in adipose tissue is an important consequence of insulin-resistance, leading to increased flux of NEFA to the liver. Increased NEFA flux has been suggested to increase the size of the cytosolic triglyceride pool, which in turn influences the secretion rate and size of VLDL. The size of VLDL varies between nm (Table 1), due to differences in triglyceride content. Large triglyceride-rich VLDL are termed VLDL1, whereas those containing less triglycerides are termed VLDL2. 13

14 INTRODUCTION The link between dyslipidemia and atherosclerosis Although apob-containing lipoproteins are necessary for lipid absorption and triglyceride homeostasis, their accumulation in plasma is deleterious. The incidence of CVD is highly associated with a specific risk profile, characterized by elevated plasma triglycerides, increased levels of small dense LDL and decreased levels and size of HDL. The pathophysiology behind this atherogenic dyslipidemia appears to be overproduction of VLDL, in particular VLDL1 [33]. Triglycerides in VLDL1 are preferably transferred to LDL and HDL in exchange for cholesterol esters by the action of CETP. Depletion of cholesterol esters and enrichment of triglycerides in LDL and HDL promote hydrolysis by HL, resulting in small dense LDL and HDL. Low HDL levels are associated with increased risk of CVD. There are several factors contributing to the particular atherogenicity of small dense LDL [34]. One important factor is conformational changes of apob100 that result in the poor affinity for the LDL receptor and thereby prolonged residence time in the circulation. In combination with the small particle size and the high binding affinity to proteoglycans in vessel wall, this promotes the penetration of small dense LDL into the arterial intima. In addition, small dense LDL are more easily oxidized which may further trigger the initiation of atherogenesis. ATHEROSCLEROSIS CVD is the leading cause of death in the world. The major underlying cause is atherosclerosis, which is characterized by successive thickening of the arterial wall and subsequent reduction of the arterial lumen. The development of atherosclerosis is a slow process, which begins early during childhood but does not become apparent until middle age or later. The primary risk factors for atherosclerosis are hyperlipidemia, hypertension and cigarette smoking, while secondary risk factors are diabetes, obesity, lack of exercise, stress and male sex [35]. Furthermore, the apob/apoai ratio is highly predictive of cardiovascular risk [36]. LDL Monocyte Lumen ENDOTHELIUM Modified Macrophage Foamcell LDL Figure 2. The development of atherosclerosis. Smooth muscle cells Intima ELASTIC LAMINAE Media 14

15 INTRODUCTION Atherogenesis The mechanism for initiation of atherogenesis has been an issue of debate. However, most evidence support the response-to-retention hypothesis proposing that retention of apob-containing LDL or IDL in the arterial wall is the initiating event in atherogenesis [37, 38]. Small dense LDL, in particular, can easily penetrate the endothelium of arterial walls (Figure 2). This event is followed by retention of LDL in the intima through interaction with proteoglycans and other components of the extracellular matrix. The retained LDL particles undergo oxidative and enzymatic modifications, which result in release of phospholipids that stimulate endothelial cells to secrete adhesion molecules attracting monocytes. These cells migrate into the intima where they proliferate and differentiate into macrophages. The modified LDL particles are taken up by macrophages via scavenger-receptors and the macrophages are transformed into lipid-filled foam cells [35, 39]. Lesions that mainly consist of foam cells are referred to as fatty streaks and such lesions are in fact found already in infants and young children [40]. The fatty streaks are quite harmless but they are considered to be precursors of the more advanced and clinically significant fibrous plaques. In the progression to fibrous plaques, smooth muscle cells (SMCs) migrate from the media to the intima where they proliferate and secrete extracellular matrix proteins. Subsequently, foam cells die and release their contents. The fibrous plaque is therefore characterized by an acellular necrotic core with a lot of extracellular cholesterol, covered by a fibrous cap containing SMCs and collagen. Initially, the lesions grow towards the adventitia but when a critical point is reached they begin to expand outwards and the lumen may be occluded. PEROXISOME PROLIFERATOR-ACTIVATED RECEPTORS PPARs are ligand-activated transcription factors belonging to the nuclear hormone receptor superfamily. Three subtypes, encoded by separate genes, have been identified: PPARα, PPARγ and PPARδ (also called PPARβ). PPARs regulate gene transcription by forming heterodimers with retinoid X receptor (RXR) and binding to specific peroxisome proliferator response elements (PPRE) in the promoter of target genes [41]. The PPRE contains two direct repeats (DR) of the hexameric recognition sequence AGGTCA separated by one nucleotide, thus called DR1 elements. To initiate transcriptional activity, the PPAR:RXR complex has to be activated by ligand binding. Natural ligands for PPARs are unsaturated fatty acids and eicosanoids [42], whereas 9-cis-retinoic acid is the ligand for RXR [41]. Upon ligand binding, the conformation of the receptor is altered, which promote high affinity interactions with coactivators that remodel chromatin structure and activate the transcriptional machinery [43]. PPARγ-coactivator 1 (PGC-1), which is one of several coactivators involved, is described more in detail below. 15

16 INTRODUCTION Although the PPAR subtypes have distinct tissue distribution and functions, all of them are modulators of lipid metabolism [43]. PPARα is highly expressed in tissues with high rates of fatty acid metabolism, such as liver, brown adipose tissue, heart and skeletal muscle. The expression of PPARα follows a diurnal rhythm, which coincides with circulating levels of glucocorticoids [44]. Activation of PPARα induces transcription of genes involved in fatty acid uptake and fatty acid oxidation, described more in detail below. PPARγ exists in two isoforms, PPARγ1 and PPARγ2, which are generated by alternative splicing [45]. PPARγ is primarily expressed in adipose tissue where it promotes adipogenesis and lipid storage. However, PPARγ is also expressed in smaller amounts in the liver and has been shown to correlate with accumulation of triglycerides, i.e. hepatic steatosis [46, 47]. PPARδ is ubiquitously expressed, although its expression in the liver is very low. This is the least characterized PPAR subtype but in addition to its importance for wound healing [48], it has been shown to play a role also in lipoprotein metabolism since PPARδ-deficient mice display hypertriglyceridemia due to increased VLDL secretion [49]. PPARα and lipid metabolism PPARα plays a central role in lipid metabolism and one major function of PPARα is to regulate mitochondrial, peroxisomal and microsomal fatty acid oxidation in the liver [43, 50]. PPARα activation upregulates fatty acid transport protein (FATP) that facilitates the uptake of fatty acids from the circulation into hepatocytes [51, 52]. Further metabolism in the β-oxidation pathways requires conversion of fatty acids into activated fatty acyl-coa. This reaction is mediated by acyl-coa synthetases, which are also induced by PPARα activation [53]. Mitochondrial β-oxidation is responsible for oxidation of short, medium and longchain fatty acids and contributes to energy production via oxidative phosphorylation. Activated fatty acids are transported through the mitochondrial membrane by carnitine palmitoyltransferase I (CPT-I) and β-oxidation is initiated by the action of short-, medium- or long-chain acyl-coa dehydrogenase (SCAD, MCAD or LCAD), depending on the length of the fatty acid. PPARα activation promotes mitochondrial β-oxidation by stimulating expression of these enzymes [54, 55]. Similarly, PPARα activation increases expression of acyl-coa oxidase (ACO), which is the rate-limiting enzyme in peroxisomal β-oxidation [43, 56]. This pathway is uncoupled from oxidative phosphorylation, which means that the energy is lost as heat. However, peroxisomal β-oxidation serves an important function in shortening very long-chain fatty acids since these are prevented from entering the mitochondria directly. After some rounds of peroxisomal β-oxidation the fatty acids are prepared for complete oxidation in the mitochondria. Microsomal ω-oxidation represents a minor pathway of fatty acid oxidation under normal circumstances. However, it becomes more important during conditions associated with an increased delivery of fatty acids to the liver, such as fasting or diabetes [57]. The ω-oxidation is governed by the cytochrome P450 4A (CYP4A) 16

17 INTRODUCTION enzymes, which are known to be upregulated by PPARα agonists [58]. Thus, PPARα activation regulates both transport of fatty acids and oxidation in the mitochondrial, peroxisomal and microsomal compartments. Fibrates are synthetic ligands for PPARα and these compounds are used clinically in the treatment of dyslipidemia. The major beneficial effects of fibrates are decreased plasma triglycerides, decreased levels of atherogenic small dense LDL and increased levels of HDL cholesterol [59-62]. The triglyceride-lowering effect of fibrates is partly explained by the decreased secretion of VLDL-triglycerides from the liver [63-65]. Although the mechanisms are not completely understood, decreased triglyceride secretion is thought to result from enhanced fatty acid oxidation that consequently reduces the availability of triglycerides for VLDL assembly. Beside the beneficial effect on plasma triglycerides, decreased triglyceride secretion is also believed to inhibit the formation of small dense LDL, since these particles are products of triglyceride-rich VLDL. The triglyceridelowering effect of fibrates also results from enhanced hydrolysis of triglycerides in VLDL and stimulated receptor-mediated clearance of apob-containing particles, due to increased expression of LPL [66] and decreased expression of apociii [67]. The increase in HDL cholesterol, is in humans related to transcriptional induction of the major HDL apolipoproteins, apoai [68] and apoaii [69]. This effect may in turn result in a more efficient reverse cholesterol transport. The phenotype of PPARα-deficient mice Fasting involves major changes in hepatic metabolism, including increased rates of fatty acid oxidation, gluconeogenesis and ketogenesis. The particular importance of PPARα in the fasting response is obvious from studies in PPARα-deficient mice. PPARα-deficient mice display no obvious phenotype during normal conditions, while fasting results in hypoglycemia, hypoketonemia, hypothermia, elevated plasma free fatty acids and hepatic lipid accumulation, secondary to reduced fatty acid uptake and severely impaired fatty acid oxidation [70-72]. Aged PPARαdeficient mice develop a sexually dimorphic phenotype, characterized by a more pronounced hepatic steatosis in males and a more pronounced hypertriglyceridemia and obesity in females [73]. Although the obesity phenotype has been questioned and the effect of PPARα-deficiency on plasma triglycerides and cholesterol seems to be dependent on the genetic background [74], it has been shown that female PPARα-deficient mice, in contrast to males, display higher serum apob levels associated with VLDL and increased hepatic triglyceride secretion [75]. Furthermore, female PPARα-deficient mice are less affected than males by inhibition of mitochondrial β-oxidation [76]. Thus, PPARα-deficiency results in sex-specific effects on fatty acid and lipoprotein metabolism. 17

18 INTRODUCTION PPARγ-COACTIVATOR 1 The PGC-1 family of coactivators consists of three members; PGC-1α, PGC-1β and PGC-1 related coactivator (PRC). PGC-1 was discovered in 1998 as a coactivator of PPARγ and thyroid hormone receptor. It was found to be markedly induced in brown adipose tissue and skeletal muscle upon cold exposure, suggesting an important role in adaptive thermogenesis [77]. Three years later, Andersson et al. found sequence similarities with another coactivator that was designated PRC [78]. This coactivator is ubiquitously expressed in murine and human tissues but the function of PRC is still largely unknown. When the third member of the family was identified in 2002, PGC-1 was renamed to PGC-1α and the novel homologue was termed PGC-1β [79]. The PGC-1 coactivators interact with many different transcription factors, including PPARα, PPARγ, sterol regulatory element binding protein (SREBP) and liver X receptor (LXR) [80]. Their broad range of targets makes them important regulators of glucose, lipid and energy homeostasis [80]. PGC-1α and PGC-1β are highly expressed in brown adipose tissue, heart and skeletal muscle [80], whereas the hepatic expression is relatively low during normal conditions. Fasting has been shown to induce PGC-1α in the liver, while the fasting response on PGC-1β is uncertain [81-83]. Hepatic functions of PGC-1α and PGC-1β PGC-1α has an important role in the regulation of hepatic gluconeogenesis. Adenoviral-mediated overexpression of PGC-1α has been shown to stimulate glucose production and increase the expression of genes encoding gluconeogenic enzymes, both in vivo in rats and in vitro in primary hepatocytes [81]. Furthermore, reduced glucose production and decreased expression of gluconeogenic genes was observed after hepatic knockdown of PGC-1α in mice using RNA interference (RNAi) adenovirus [84]. Hepatic PGC-1α mrna levels are elevated in streptozotocin-treated mice as well as in ob/ob mice, representing animal models of type 1 and type 2 diabetes, respectively [81, 82]. Since increased gluconeogenesis and hepatic glucose output are important pathological components of diabetes, increased PGC-1α activity in the liver has been suggested to contribute to the hyperglycemia observed in diabetes. In contrast to PGC-1α, the gluconeogenic pathway is poorly activated by PGC-1β [82]. Instead, PGC-1β is rather involved in hepatic lipid handling. A recent study showed that administration of a diet enriched in saturated fats stimulated the expression of PGC-1β in the liver, whereas PGC-1α was unaffected [85]. Furthermore, it was observed that hepatic overexpression of PGC-1β in fat-fed rats resulted in elevated triglycerides and cholesterol in the VLDL fraction, indicating increased secretion of VLDL particles [85]. Thus, PGC-1β signaling was suggested to be the link between dietary saturated fats and hyperlipidemia. In addition to these distinct functions of PGC- 1α and PGC-1β, they also share some functions in the liver, including induction of genes involved in mitochondrial energy production and fatty acid oxidation [82]. 18

19 INTRODUCTION Hepatic PGC-1α knockdown was accompanied with markedly reduced expression of genes in the mitochondrial and peroxisomal β-oxidation [84]. In addition, hepatocytes isolated from PGC-1α-deficient mice had reduced capacity to oxidize fatty acids, which partly may explain the hepatic steatosis seen in these mice [86]. Furthermore, increased expression of fatty acid oxidation genes was observed after either PGC-1α or PGC-1β overexpression in vitro [82]. ADIPOSE DIFFERENTIATION-RELATED PROTEIN Most cells contain cytosolic lipid droplets of various sizes that are used as a source of energy or in membrane biogenesis. Lipid droplets are characterized by a core of neutral lipids surrounded by a phospholipid monolayer and proteins associated to the surface. Some of these proteins share a homologous sequence in the N- terminal called the PAT-domain and they are therefore referred to as members of the PAT-family [87]. Perilipin, adipose differentiation-related protein (ADRP), TIP47 and S3-12 are included in this family and a new member termed myocardial lipid droplet protein (MLDP) was recently identified [88]. ADRP was initially recognized as an early marker of adipocyte differentiation [89], but its expression is essentially ubiquitous [90, 91]. The protein levels of ADRP has been shown to reflect the total amount of neutral lipids within the cells [90, 92]. Overexpression of ADRP in fibroblasts [93] and macrophages [94] resulted in lipid accumulation, without induction of genes involved in lipogenesis [93] or lipid efflux [94]. Furthermore, overexpression of ADRP in COS-7 cells has been shown to promote uptake of long-chain fatty acids [95]. Together, these observations clearly demonstrate that ADRP promote cellular lipid storage, although the mechanisms remain to be elucidated. It has been reported that long-chain fatty acids have the capacity to stimulate ADRP expression at the transcriptional level [96]. In addition, ADRP transcription was induced by agonists to PPARα, γ and δ in macrophages and colorectal cells [97-99] and by PPARα agonists in kidney cells [100]. Accordingly, PPREs has been identified both in the human and murine promoter of ADRP [101, 102]. GROWTH HORMONE Growth hormone (GH) is a peptide hormone produced by somatotrophic cells in the anterior pituitary. The secretion of GH is primarily regulated by two hypothalamic peptide hormones: GH releasing hormone (GHRH) and somatostatin, which stimulate and inhibit the secretion, respectively. In most species, GH is secreted in a pulsatile sexually dimorphic manner [ ] due to action of sex hormones on the hypothalamus [106]. The male pattern is characterized by large regular GH pulses with very low levels in between, whereas the female pattern is characterized by continuous GH secretion with lower pulses and higher baseline levels. GH secretion is also regulated by other hormones (e.g. glucocorticoids and thyroid hormone) and metabolic situations 19

20 INTRODUCTION (e.g. starvation, diabetes and obesity), emphasizing the complex regulation of GH secretion [107]. After secretion into the circulation, about 60% of plasma GH is bound to GH-binding proteins (GHBP) [108]. Since GH is only active in its unbound form, GHBP is thought to serve as a reservoir of GH in the circulation. GH stimulates the production of insulin-like growth factor (IGF)-I in the liver and some effects of GH are mediated via IGF-I [109]. The actions of GH are mediated via binding to transmembrane GH receptors that are present on the cell surface in most tissues, including liver, adipose tissue and skeletal muscle [110]. The signaling cascade is initiated when one GH molecule binds two GH receptors (GHR), resulting in receptor dimerization. This event increases the affinity of each GHR for a tyrosine kinase called janus kinas 2 (JAK2). JAK2 phosphorylates several tyrosine residues at the GHR, providing docking sites for other signaling molecules. The signal transducers and activators of transcription 5 (STAT5) are such molecules, which upon tyrosine phosphorylation dimerize, translocate into the nucleus, bind to specific DNA response elements and stimulate target gene transcription [110]. STAT5 exists in two isoforms termed STAT5a and STAT5b, of which STAT5b seems to be more abundant in the liver and more important in GH signaling. Interestingly, GH-activated STAT5b can inhibit PPARα and PPARγ regulated transcription [111, 112] and conversely, PPARα and PPARγ agonists can inhibit STAT5b regulated transcription [113], indicating a mutual inhibitory cross-talk between PPARs and GH. Beside the crucial growth-promoting effect of GH, this hormone also plays a central role in different aspects of carbohydrate and lipid metabolism [114]. Altered GH secretion is therefore closely related to metabolic disturbances. Disturbances in GH secretion The most dramatic conditions associated with altered GH secretion are acromegaly and GH-deficiency (GHD). The chronical oversecretion of GH that characterizes acromegaly often results from a benign pituitary adenoma. Beside effects on the skeletal system and soft tissue, acromegalic patients often display several features of the metabolic syndrome, including hypertension, type 2 diabetes, insulin resistance and dyslipidemia [115]. GHD most commonly results from a pituitary tumor. The main features of GHD include reduced lean body mass, increased abdominal adiposity, reduced muscle strength and exercise performance, impaired psychological well-being and dyslipidemia [116]. In contrast to the absolute GH-deficiency in GHD patients, there are conditions where the GH secretion is present but decreased. For instance, it has been shown that obese subjects have diminished pulsatile GH secretion and accelerated plasma clearance, which together results in markedly reduced mean plasma GH concentrations [117]. In contrast, type 1 diabetic patients have increased GH secretion due to increased frequency and higher amplitudes of the GH bursts [118]. In a study where type 2 diabetic patients were compared with 20

21 INTRODUCTION obese and lean controls, it was shown that also type 2 diabetics had an increased frequency of GH bursts [5]. However, total GH secretion rate was not increased since the amplitudes of the GH bursts were lower. The fact that GH burst mass decreases in proportion to the degree of obesity [119], suggest that the higher body mass index (BMI) in the type 2 diabetic group compared to the lean controls may explain these results. In addition to the altered GH secretion in diabetic and obese subjects, it has been shown that GH secretion decreases with age [119]. GH and lipoprotein metabolism An accurate GH secretion is important for lipoprotein metabolism, as evidenced by the dyslipidemia seen both in acromegaly and GHD. Several studies in GHD patients have demonstrated major effects on the lipoprotein profile [ ]. Although there are some differences between these studies, most of them show that GHD subjects have elevated levels of apob, triglycerides, total cholesterol and LDL cholesterol in combination with reduced levels of HDL cholesterol [ ]. Furthermore, the VLDL particles of GHD patients are large and triglyceride-rich, which may result in a larger production of small dense LDL [123]. This atherogenic lipoprotein profile may at least in part explain the high incidence of cardiovascular disease in GHD patients [4]. The most clear effects of GH replacement therapy in these subjects are decreased LDL cholesterol and increased HDL cholesterol, whereas triglyceride levels are largely unaffected [ ]. Some studies have also demonstrated decreased plasma levels of apob [125, 126], despite the fact that GH treatment has been reported to stimulate apob secretion [127]. A possible explanation may be enhanced clearance due to upregulation of hepatic LDLR expression [128]. Although GH therapy improves most components of the dyslipidemia, it has been shown to elevate plasma levels of atherogenic lipoprotein (a) (Lp(a)) [124, 126]. The dyslipidemia of acromegaly is characterized by increased plasma levels of triglycerides [ ], whereas plasma cholesterol has been shown to be either increased [131] or decreased [129]. Furthermore, Lp(a) levels are higher in acromegalic patients. Surgical removal of the adenoma has been reported to decrease triglycerides and increase HDL cholesterol, whereas total cholesterol and LDL cholesterol was unaffected [132, 133]. In addition, plasma levels of Lp(a) was reduced after adenomectomy [133]. GH has profound effects on lipid metabolism also in rodents. GH treatment of hypophysectomized (Hx) rats results in stimulated hepatic triglyceride synthesis and VLDL secretion [ ] as well as increased hepatic expression of lipogenic enzymes [137] and MTP [138]. However, the stimulated VLDL secretion is not associated with increased plasma levels [139], indicating that GH affects both secretion and clearance of VLDL. Several effects of GH may contribute to the enhanced clearance rate. GH increases apob mrna editing, which result in a larger proportion of apob48-containing VLDL with a faster turnover [140]. In addition, GH treatment increases LPL activity [141], HL activity [142], plasma 21

22 INTRODUCTION apoe [139] and hepatic LDLR expression [128], which together may stimulate both hydrolysis of VLDL-triglycerides and hepatic uptake of IDL and LDL. The phenotype of bgh transgenic mice Another model used to investigate effects of GH is transgenic mice overexpressing GH. Compared to the Hx model where the importance of normalized GH levels is studied, this model describes long-term effects of abnormally high GH levels. GH transgenic mice are characterized by increased body size, organomegaly, increased lean body mass and decreased body fat mass [143]. Furthermore, they are hyperinsulinemic and have marked alterations in hepatic lipoprotein metabolism [144, 145]. On a normal chow diet, bovine GH (bgh) transgenic mice have elevated cholesterol levels in the LDL and HDL fractions. In addition, they have reduced triglyceride secretion rate and plasma triglycerides as well as reduced cholesterol, triglycerides and apob in the VLDL fraction [144]. When fed a high fat diet, bgh transgenic mice display a markedly different phenotype, characterized by increased levels of cholesterol, triglycerides and apob in the VLDL fraction [145]. In addition to the effects on lipoprotein metabolism, bgh transgenic mice are resistant to diet-induced obesity despite hyperphagia but develop diabetes [145]. Furthermore, they show alterations in cardiovascular functions, including increased blood pressure [146] and endothelial dysfunction [147]. GH and atherosclerosis Atherosclerosis affects the intimal layer of the arterial walls and increased intima media thickness (IMT) is an early indicator of atherosclerosis [148]. In line with the increased incidence of cardiovascular disease in GHD patients [4], several studies have reported increased IMT in these subjects [ ]. However, GH replacement therapy reversed these changes, indicating that normalized GH levels have positive effects on atherogenesis [150, 151]. In contrast, excess GH seems to be deleterious, as indicated by the fact that 60% of acromegalic patients die from cardiovascular disease [115]. Vascular consequences of high circulating GH levels are in general poorly investigated. Only a few studies have examined the presence of atherosclerosis in acromegalic patients and the results are conflicting. Increased IMT was observed in the carotid arteries of acromegalic patients compared to age-, sex- and BMI-matched controls [ ], whereas the prevalence of well-defined atherosclerotic plaques was similar between the groups [152]. Treatment of acromegalic patients with a somatostatin analog, which reduced GH and IGF-I levels, resulted in decreased IMT [154]. In contrast, another study reported that IMT was actually lower in acromegalic patients compared to controls matched for age, sex, hypertension, dyslipidemia, diabetes and smoking [155]. The contradictory results may reflect small patient materials but is most likely due to difficulties to find relevant control groups matched for the independent risk factors commonly found in acromegalic patients. Indeed, age, male sex, hypertension, dyslipidemia, diabetes and smoking has been reported to predict 22

23 INTRODUCTION IMT of the carotid arteries [155]. Accordingly, increased IMT was observed in acromegalic patients compared to healthy controls, whereas there was no difference between patients and controls matched for associated risk factors [156]. In contrast, acromegalic patients had impaired vascular flow-mediated dilatation compared to both healthy and matched controls, indicating endothelial dysfunction [156]. Thus, GH excess may have a role in generating endothelial dysfunction, which is involved in early stages of atherosclerosis. 23

24 AIMS OF THE THESIS AIMS OF THE THESIS The general aim of this thesis was to investigate the effects of GH, PPARα and PGC-1 on hepatic lipid and lipoprotein metabolism in mice and to explore interactions between PPARα and GH and between PPARα and PGC-1. The specific aims of Paper I-IV were: To study if high plasma concentrations of GH in mice affect atherosclerotic lesion formation and associated risk factors behind development of atherosclerosis. To study the importance of PPARα for the effects of GH on hepatic lipid and lipoprotein metabolism in mice. To study if PPARα activation affects ADRP expression in mouse liver and to explore the importance of changed ADRP expression for the effects of PPARα activation on triglyceride accumulation and secretion from primary mouse hepatocytes. To study the effects of PGC-1α and PGC-1β overexpression on hepatic lipid and lipoprotein metabolism in mice and to explore the influence of changed expression of these coactivators for the effects of PPARα activation. 24

25 METHODOLOGICAL CONSIDERATIONS METHODOLOGICAL CONSIDERATIONS This thesis is based on studies in vivo and in vitro. Detailed descriptions of the methods are given in each individual paper. Additional methodological considerations are discussed in this section. GENETICALLY MODIFIED MICE Mice are very useful as experimental animals due to their easy maintenance and short breeding time. Furthermore, the possibility to generate genetically modified mice has resulted in great advances in medical research. The first transgenic mouse, described in 1982 by Palmiter et al [157], was in fact a GH transgenic mouse, overexpressing rat GH. Since then, several transgenic and knockout models have been generated. ApoE-deficient mice overexpressing bgh Mice are highly resistant to atherosclerosis partly due to their favourable lipoprotein profile with low levels of apob-containing lipoproteins and high levels of HDL. This problem was solved in early 90 s when two genetically modified mouse models were generated; apoe-deficient [158, 159] and LDLR-deficient [160] mice. Both models show hypercholesterolemia, although less pronounced in LDLR-deficient mice. We preferred to use apoe-deficient mice since they develop atherosclerotic lesions, similar to those in humans, on normal chow diet [161]. The bgh transgenic mice used for breeding were generated at the Department of Physiology at Göteborg University [162] using a construct generously donated by Dr Palmiter (University of Washington). The construct, in which the metallothionein promoter is linked to the sequence encoding bovine GH, was injected into the pronucleus of fertilized eggs, followed by transplantation into the oviducts of a pseudopregnant mouse. Offspring carrying the transgene were identified by PCR of DNA from tail biopsies. To be able to study how overexpression of GH affects the development of atherosclerosis, we developed a new mouse model that was used in Paper I. Heterozygous bgh transgenic mice were crossed with apoe-deficient mice on C57BL/6 background. Mice carrying the bgh gene were further bred with apoedeficient mice and offspring were genotyped to identify apoe -/- /bgh -/- and apoe -/- /bgh +/- mice. These were interbred to establish the desired apoe -/- /bgh +/- mice and apoe -/- littermate controls. Although this is the first and only model to study effects of GH excess on atherosclerosis, it should be kept in mind that it is not an optimal model of human acromegaly. One important aspect is that bgh transgenic mice have a general and constant overexpression of GH in all cells, which means that they have high levels of GH in all tissues. In addition, the GH overexpression begins around day 13 of the foetal development in contrast to acromegaly that arises after puberty. 25

26 METHODOLOGICAL CONSIDERATIONS PPARα-deficient mice Homozygous PPARα-deficient mice on pure Sv/129 background and corresponding wild-type Sv/129 controls were kindly provided by Prof. Frank Gonzalez (NIH, Bethesda, MD) [163]. In brief, the ligand-binding domain of PPARα was disrupted and the construct was electroporated into cultured embryonic stem cells for homologous recombination. Positive cells were injected into blastocysts, followed by transplantation into a pseudopregnant mouse. The PPARα-deficient mice were kept on the Sv/129 background and used in Paper II and III. HORMONE AND DRUG TREATMENT IN VIVO GH Recombinant bgh was given in a dose of 5 mg/kg/day for 7 days by means of osmotic mini pumps implanted subcutaneously on the back of the mice. Administration by osmotic pumps results in a slow and continuous release of GH that mimics the secretion pattern in females. In contrast, this type of administration to males disturbs the normal pulsatile GH pattern and may result in feminization of hepatic functions. Thus, we preferred to present the results from female mice in Paper II. GH administration to intact animals would result in decreased endogenous GH secretion because of negative feedback regulation of GH secretion. Therefore, it is likely that substitution doses of GH (about mg/kg/day), calculated from total diurnal GH secretion and plasma clearance of GH, would result in small changes in total GH exposure. However, treatment with 5 mg/kg/day would result in supraphysiological GH levels. The reason for using bovine GH instead of human GH is that the latter has been shown to induce antibody formation in the rat [164]. Furthermore, human GH also binds to the prolactin receptor, which can induce lactogenic effects. Rat and mouse GH does not bind to the prolactin receptor but is on the other hand less stable than bovine GH and also less available at the market. Wy14,643 The fibrate Wy14,643 (Chemsyn Science Laboratories) was chosen to study PPARα activation since this substance specifically activate the PPARα subtype very potently in mice. In contrast, Wy has poor affinity to human PPARα and is therefore not used in clinic [165]. In order to study the effects of Wy in vivo, mice were given 30 µmol/kg/day for one week in Paper IV and for two weeks in Paper III. 26

27 METHODOLOGICAL CONSIDERATIONS HEPATOCYTE CULTURES Hepatocytes, which comprise 80-85% of liver cells, were isolated by nonrecirculating collagenase perfusion through the vena porta of anaesthetized mice. During the first 6 minutes, the liver was perfused with Hanks balanced salt solution (HBSS) without calcium and magnesium, and supplemented with EGTA. EGTA is a cation-chelating agent which binds Ca 2+ and thereby facilitates the disruption of Ca 2+ -dependent cell-to-cell interactions. For the next 7-9 minutes, the liver was perfused with Williams E medium with Glutamax, supplemented with collagenase IV. Collagenase degrades collagen, which will result in a single-cell suspension within the liver capsule. After perfusion, the liver capsule was excised and cells were filtered to remove undigested material. The cell suspension was washed and centrifuged to remove collagenase and enriche hepatocytes since other liver cells, such as Kuppfer cells and fat-storing Ito cells, have lower density. The cells were seeded at a density of ~ cells/cm 2 in petri dishes coated with matrigel. The matrigel contains large amounts of laminin, proteoglycans and collagen, which support the hepatocytes in the subendothelial space of normal liver. Hepatocytes cultured on matrigel have been shown to maintain their phenotype better compared to hepatocytes cultured on collagen or plastic [ ]. The cells were cultured in medium containing a high dose of insulin (16nM) during the first hours, since this has been shown to improve plating efficiency and formation of dispersed monolayers [169]. After plating, the cells were cultured for up to 72 hours in a medium with a lower insulin concentration (3nM). Since the matrigel is rich in proteins, hepatocytes used for Western blot were collected in PBS supplemented with 5mM EDTA to remove the matrigel. The effects in the hepatocytes cultures were related to the DNA content in each culture dish to correct for differences in cell number. ADENOVIRAL-MEDIATED OVEREXPRESSION Adenoviruses overexpressing proteins represent a very useful tool for functional studies both in vivo and in vitro. The principle for adenovirus production is that an expression plasmid containing the gene of interest is homologously recombined with an adenovirus plasmid. The adenovirus plasmid is deficient in the E1 and E3 genes, which are important for replication and to counteract host defense mechanisms, respectively. However, transfection of the recombined plasmid into a packaging cell line results in replication and virus production since these human embryonic kidney (HEK) cells have been stably transfected with the E1 gene. Adenoviruses can easily transfect a broad range of both dividing and nondividing cells. In contrast to retroviruses, adenoviruses do not integrate into the genome and are therefore in combination with the replication-defectiveness safe to work with. However, one major drawback with the first generation of adenoviruses that are used in our studies is the short-lived effect of approximately 27

28 METHODOLOGICAL CONSIDERATIONS days in vivo. This is due to an immune response that drives infected cells to undergo apoptosis. However, a new generation of adenoviruses, called helperdependent or gutless adenoviruses, show reduced immune response and the transduction effects of these viruses are therefore much more stable in vivo [170]. Also, adenoviruses are dependent on using the coxsackie-adenovirus receptor (CAR) present on their target cell for having good infection capability. Thus, selection of appropriate cell types in culture is necessary. Tail-vein injection of adenoviruses results in hepatic overexpression of the gene of interest. This liver-specific overexpression is due to the high hepatic expression of CAR, which mediates adenoviral uptake [171]. HEPATIC TRIGLYCERIDE SECRETION IN VIVO The hepatic triglyceride secretion rate in vivo was measured by intravenous administration of Triton WR-1339 via the jugular vein (Paper II). This substance is a detergent that is thought to block the access of LPL by coating the VLDL particles. The peripheral hydrolysis of triglycerides is thereby inhibited, which results in accumulation of triglycerides in the blood. Since we wanted to measure hepatic triglycerides only, the animals were fasted 4 hours before injection to avoid measurement of chylomicrons. Serum triglycerides increased linearly after Triton WR-1339 injection and the triglyceride secretion rate could be calculated from the slope of the curve. It is important to note that increased hepatic triglyceride secretion does not necessarily mean an increased number of VLDL particles. Since each particle contains only one apob molecule, increased triglyceride secretion could also reflect a constant number of particles that contains more triglycerides. It is therefore important to measure apob in plasma as well to monitor the number of particles. ApoB MEASUREMENTS Electroimmunoassay Plasma apob was measured with an electroimmunoassay. The principle for the method is that antigens (apob) are migrating in an electric field through an agarose gel containing apob antibodies. Precipitation occurs on both sides of the central stream where the ratio between antigen and antibody is optimal. During consumption of antigens, the complex precipitates closer to the central stream and when all antigens are consumed, the precipitation lines converge. The area and height of the peak, which are proportional to the amount of antigens loaded, are compared to a standard curve with known amounts of apob and the plasma concentrations can be calculated. 28

29 METHODOLOGICAL CONSIDERATIONS Immunoprecipitation The secretion of apob from primary mouse hepatocytes into the cell culture medium was determined by immunoprecipitation. The hepatocytes were incubated with 35 S-labeled cystein/methionine mix for 2 hours followed by incubation with excess unlabeled methionine for 4 hours. Previous time course experiments have shown that 2 hours labeling is sufficient to reach intracellular steady-state levels of 35 S-labeled apob and that 4 hours chase is sufficient to complete the secretion of the labeled intracellular apob pool into the medium [75]. 35 S-labeled apob was immunoprecipitated using a rabbit anti-human apob antibody that cross-react with apob-48 and apob-100 in mouse, followed by SDS- PAGE and quantification of band density by phosphorimager. QUANTITATIVE REAL-TIME PCR For measurements of mrna expression levels, we used quantitative real-time PCR since this method has several advantages compared to other techniques such as Northern blot or ribonuclease protection assay. The method is fast and easy with high reproducibility, very low expression levels can be detected and only small quantities of RNA is required. The general principle for PCR is that double stranded DNA is denaturated into two single-stranded DNA templates, followed by annealing of sequence-specific primers to the DNA template. DNA polymerase extends the primers by incorporation of complementary nucleotides yielding a double-stranded DNA complex. The denaturation, annealing and extension cycles are repeated approximately 40 times and the amount of cdna is doubled for each cycle. Quantitative real-time PCR is based on the detection of fluorescence and there are two alternative chemistries: SYBR Green (Paper I, II and III) or TaqMan (Paper IV). SYBR Green is a fluorescent dye that binds double-stranded DNA. As more double-stranded DNA is produced during the PCR, the fluorescence increases and the intensity of fluorescence above background level is used to quantitate the amount of newly generated DNA. However, the SYBR Green dye is incorporated into all double-stranded DNA generated during the PCR, which means that both specific and non-specific products will generate signal. To assure amplification of the desired sequence only, the PCR product was analyzed by electrophoresis to check that a single amplicon of the expected size was obtained. In addition, all RNA samples were DNase-treated before cdna synthesis to remove contaminating DNA. In the TaqMan assay, sequence-specific primers and a sequence-specific probe labeled with both a reporter molecule and a quencher molecule anneal to the single-stranded DNA template. When the DNA polymerase reaches the probe, the reporter is cleaved off and starts to fluoresce. The fluorescence accumulates as cycling of PCR continues and is measured at the end of each PCR cycle. The use of 29

30 METHODOLOGICAL CONSIDERATIONS a target-specific probe results in increased specificity. In addition, a variety of fluorescent dyes are available so that multiple primers can be used simultaneously to amplify many sequences, known as multiplex RT-PCR. ATHEROSCLEROTIC MEASUREMENTS The en face method The presence of atherosclerosis in the aorta was measured with the en face method. The thoracic aorta and the aortic arch cut open longitudinally and pinned onto a silicone-coated dish so that the lumen was displayed (Figure 3). The aorta was then stained with Sudan IV, which is a red staining for neutral lipids. The surface of the lesions and the vessel was measured and the lesion area was calculated by dividing the surface of the lesions by the surface of the vessel. Figure 3. The thoracic aorta stained with Sudan IV. The cross-sectional method In addition to the size of the lesions, there are different developmental stages ranging from fatty streaks to severe lesions that could actually be more informative than the size of the lesions. In contrast to the en face method, crosssections also give the opportunity to characterize the lesion components with several types of stainings. Sections for this purpose are often taken from aortic sinus, which is the site where the aortic valves separate the left ventricle from the aorta. We used Russel-Movats Pentachrome staining to quantify and characterize lesions [172]. This method stains normal heart tissue red whereas proteoglycans, which are major constituents of atherosclerotic lesions, are stained blue. To characterize the developmental stage of the lesions, they were classified into five categories according to Van Vlijmen et al. [173]: (1) Early fatty streak: up to 10 foam cells present in the intima, (2) Regular fatty streak: more than 10 foam cells present in the intima, (3) Mild plaque: extension of foam cells into the media and mild fibrosis of the media without loss of architecture, (4) Moderate plaque: foam cells in the media, fibrosis, cholesterol clefts, mineralization and/or necrosis of the media, (5) Severe plaque: as 4 but more extensive and deeper into the media. Lesions classified as category 1-3 were grouped into mild alterations, whereas 30

31 METHODOLOGICAL CONSIDERATIONS those classified as category 4-5 were grouped into severe alterations. Since these types of judgments are subjective, the classification of the lesions was blinded and performed by two independent persons. 31

32 SUMMARY OF RESULTS SUMMARY OF RESULTS PAPER I Increased atherosclerotic lesion area in apoe-deficient mice overexpressing bovine growth hormone Oversecretion of GH in humans (acromegaly) is associated with increased cardiovascular morbidity and mortality [174]. However, there are few reports regarding effects of GH on vascular pathology and atherosclerosis. In this study, we crossed bgh transgenic mice with apoe -/- mice that are prone to develop atherosclerosis, and evaluated the presence of atherosclerotic lesions as well as effects on some well-known risk factors for the development of atherosclerosis. Systolic blood pressure was higher in 20-week-old apoe -/- /bgh mice compared to littermate apoe -/- mice. When terminated at 22 weeks of age, atherosclerotic lesion area in the thoracic aorta was larger in apoe -/- /bgh mice compared to apoe -/- mice, both on standard diet and Western diet. In contrast, there was no difference between the genotypes when analyzing lesion area in the aortic arch or in the aortic sinus. However, aortic sinus lesions were more severe in apoe -/- /bgh mice fed standard diet compared to controls, whereas Western diet resulted in generally more severe lesions with no difference between the genotypes. Serum cholesterol and triglycerides were measured to examine if a more marked dyslipidemia could contribute to the accelerated lesion formation in apoe -/- /bgh mice. However, apoe -/- /bgh mice had lower levels of cholesterol in the VLDL+LDL fractions, and higher levels of HDL cholesterol compared to controls. Furthermore, serum triglyceride levels were lower in apoe -/- /bgh mice, whereas apob levels were not different between the genotypes. Serum amyloid A (SAA) and hepatic C-reactive protein (CRP) mrna levels were measured to evaluate the contribution of inflammation to the accelerated atherosclerosis. SAA was increased in apoe -/- /bgh mice fed standard diet compared to controls. This difference was not seen in mice fed Western diet; possibly due to upregulation of SAA by Western diet per se. In addition, hepatic CRP mrna was higher in apoe -/- /bgh mice than in apoe -/- mice, indicating an ongoing inflammatory process. In conclusion, overexpression of GH accelerates the development of atherosclerotic lesions in apoe -/- mice. The larger area and more marked severity of lesions in apoe -/- /bgh mice is accompanied by higher blood pressure and signs of an inflammatory response, whereas the plasma lipoprotein profile is not worsened by GH overproduction. 32

33 SUMMARY OF RESULTS PAPER II Importance of PPARα for the effects of growth hormone on hepatic lipid and lipoprotein metabolism In addition to its crucial effects on growth, GH has several effects on lipid and lipoprotein metabolism. Studies in rats have shown that GH is lipogenic in the liver, thereby increasing triglyceride synthesis and VLDL secretion [134, 136, 137]. In contrast, GH is lipolytic and anti-lipogenic in adipose tissue, resulting in increased flux of fatty acids to the liver [175]. These fatty acids may in turn act as ligands for PPARα [42]. Opposite to GH, activation of PPARα results in decreased VLDL secretion in rat hepatocytes [65]. GH has been shown to inhibit PPARα expression [ ] and signaling [111, 112]. Thus, GH and PPARα may interact in a complex manner in the regulation of hepatic lipid metabolism. To study the importance of PPARα for the effects of GH on lipid and lipoprotein metabolism, GH was given as a continuous infusion to PPARα-deficient mice and corresponding wild-type mice. Serum and liver lipids, hepatic triglyceride secretion and hepatic gene expression were analyzed. One week of GH treatment (5mg/kg/day) increased body weight and liver weight, indicating treatment response. GH increased serum cholesterol, mainly in the HDL fraction. Furthermore, serum apob was elevated in GH treated animals, indicating an increased number of VLDL and LDL particles. A separate experiment showed that the triglyceride secretion rate was enhanced by GH also in mice. However, serum triglycerides were not elevated, which reflect an increased clearance of triglycerides. Moreover, hepatic triglyceride content was lower in GH treated mice. The effects of GH on these parameters were similar in PPARα-deficient and wild-type mice and thus independent of PPARα. To further investigate the putative interaction between GH and PPARα, we measured the expression of genes involved in hepatic lipid metabolism, with focus on targets for both GH and PPARα. Increased levels of IGF-I mrna indicated expected treatment response. However, GH had no effect on PPARα mrna expression. The increased triglyceride secretion rate following GH treatment was not associated with increased mrna levels of MTP. GH treatment resulted in decreased expression of ACO and MCAD mrna, indicating decreased peroxisomal and mitochondrial β-oxidation, respectively. The lipogenic transcription factors SREBP-1 and LXRα were not affected by GH treatment. The effect of GH on these genes was similar between the genotypes and thus PPARαindependent. In contrast, there was an interaction between GH and PPARα in the regulation of Cyp4a10, which is involved in microsomal ω-oxidation of fatty acids. GH decreased Cyp4a10 mrna levels in wt mice, while the expression was increased by GH in PPARα-deficient mice. Thus, the effect of GH on Cyp4a10 mrna expression is PPARα-dependent. Furthermore, GH treatment resulted in 33

34 SUMMARY OF RESULTS decreased expression of PPARγ1 and PPARγ2, which correlated well with the decreased triglyceride content in the liver. Interestingly, statistical analysis of PPARγ2 mrna expression showed an interaction between GH treatment and genotype. Thus, presence of PPARα blunted an effect of GH on PPARγ2 mrna expression. This study also demonstrated some interesting effects of PPARα-deficiency. Serum triglyceride levels were higher in PPARα-deficient mice due to increased triglyceride secretion rate and unaffected clearance. Serum apob levels were also higher in PPARα-deficient mice compared to controls and the combination of high GH levels and low PPARα expression gave an additive effect. Furthermore, the elevated triglyceride secretion was correlated to higher levels of MTP, SREBP-1 and LXRα mrna in the livers of PPARα-deficient mice. In conclusion, most of the effects of GH on lipid and lipoprotein metabolism were similar in PPARα-deficient and wild-type mice. However, the effects of GH on PPARγ2 and Cyp4a10 were dependent on PPARα, indicating that PPARα may be important for the effects of GH on PPARγ signaling and ω-oxidation. PAPER III PPARα activation increases triglyceride mass and adipose differentiationrelated protein in hepatocytes Treatment with PPARα agonists (i.e. fibrates) has several beneficial effects on the lipoprotein profile, including decreased plasma triglycerides, increased HDL cholesterol and decreased levels of small dense LDL [59-62]. These effects of fibrates result partly from the decreased VLDL triglyceride secretion that is seen in vivo in humans and rats [63, 64] and in vitro in primary rat hepatocytes [65]. The well-known stimulatory effect of PPARα agonists on fatty acid oxidation [55] is thought to contribute to the reduced VLDL triglyceride secretion. Furthermore, decreased triglyceride biosynthesis in rat hepatocytes has been shown to correlate with reduced triglyceride secretion [65, 179]. In contrast, in vivo studies in rats showed decreased triglyceride secretion without affecting triglyceride biosynthesis [64], indicating that other mechanisms might be involved. ADRP is a lipid-droplet associated protein that has been shown to stimulate lipid accumulation [93, 94]. PPARs are able to induce ADRP expression in macrophages and colorectal cancer cells [97-99] and PPREs were recently found both in the human and murine promoter of ADRP [101, 102]. The aims of the study were to investigate the effects of the specific PPARα agonist Wy14,643 (Wy) on the expression of ADRP in mouse liver in vivo and in primary mouse hepatocytes in vitro and to determine the importance of changed ADRP expression for the effects of PPARα activation on triglyceride secretion and intracellular triglyceride accumulation. 34

35 SUMMARY OF RESULTS C57BL/6 mice fed ordinary chow or high-fat diet were treated with Wy (30 µmol/kg/day) for 2 weeks. Irrespective of diet, Wy treatment resulted in increased ADRP mrna expression in the liver. ADRP protein increased in chow fed mice treated with Wy, whereas this effect was not significant in mice fed the high-fat diet. Interestingly, the higher ADRP expression after Wy treatment was paralleled by increased hepatic triglyceride content. This effect was seen irrespective of diet, although less pronounced in the animals fed high-fat diet. These effects were mediated by PPARα activation as shown by the absence of response in PPARαdeficient mice. Furthermore, the stimulatory effect on ADRP expression and triglyceride accumulation was a direct effect on hepatocytes, since 3 days Wy incubation (10µM) resulted in increased ADRP mrna and protein levels with a concomitant increase in intracellular triglycerides. To elucidate the mechanisms behind the triglyceride accumulation, we measured fatty acid oxidation, triglyceride biosynthesis and triglyceride secretion after Wy incubation. Fatty acid oxidation was enhanced 4-fold in parallel with increased mrna expression of genes involved in fatty acid oxidation. Triglyceride biosynthesis was not affected by Wy, whereas triglyceride secretion was reduced by 50%. Therefore, we concluded that although the triglyceride mass increases by PPARα activation, the availability of these triglycerides for VLDL secretion are prevented. Since ADRP overexpression has been shown to stimulate triglyceride accumulation [93, 94], we hypothesized that increased ADRP expression after Wy incubation could explain the intracellular triglyceride accumulation and reduced triglyceride secretion. In fact, adenovirus-mediated overexpression of ADRP decreased triglyceride secretion by 50% in parallel with increased triglyceride accumulation. However, Wy incubation of ADRP overexpressing cells resulted in a further decrease in triglyceride secretion even though ADRP protein levels were unchanged, indicating that the increased ADRP expression was not the sole mechanism for the reduced triglyceride secretion. Wy incubation of ADRP overexpressing cells enhanced the fatty acid oxidation, but this could not explain the lack of triglycerides for VLDL secretion since the triglyceride content was unchanged in these cells. In conclusion, PPARα activation does not primarily decrease the triglyceride secretion via enhanced fatty acid oxidation since the intracellular triglyceride content of hepatocytes is not limiting. Rather, PPARα activation prevents the use of cytosolic triglycerides for VLDL assembly, in part by increasing the expression of ADRP. 35

36 SUMMARY OF RESULTS PAPER IV Hepatic PGC-1β overexpression in mice causes combined hyperlipidemia and a blunted response to PPARα activation The PGC-1 coactivators play a central role in the maintenance of glucose, lipid and energy homeostasis [80]. Since dysregulation of these coactivators are thought to be involved in metabolic conditions such as obesity and diabetes, they are putative pharmaceutical targets. PGC-1α is important in the fasting response since it regulates gluconeogenesis, fatty acid oxidation and ketogenesis, whereas PGC-1β is more involved in lipid handling. It was recently shown that hepatic overexpression of PGC-1β in high fat-fed rats increased serum triglycerides and cholesterol in concert with elevated expression of fatty acid synthesis genes [85]. Moreover, cell transfection studies have shown that PGC-1 enhances the transcriptional response of PPARα on target genes [180]. In the current study, we investigated the effect of hepatic PGC-1α and PGC-1β overexpression on lipoprotein levels and profiles, hepatic lipid content and gene expression in normal C57BL/6 mice. In addition, we studied if PGC-1 overexpression influenced the response to Wy treatment. An initial dose-response study showed that PGC-1β overexpression markedly increased plasma triglyceride, cholesterol and apob levels, while PGC-1α had no or very small effects. The dose 1.2 x 10 9 ifu was chosen for the subsequent PPARα interaction study, since this dose resulted in clear effects without causing liver damage as judged by normal plasma alanine aminotransferase (ALAT) levels. As in the initial study, only PGC-1β resulted in elevated plasma triglyceride, cholesterol, NEFA and apob levels. The cholesterol was predominantly associated with non-hdl fractions. PGC-1β overexpression increased liver weights but liver triglyceride levels were not affected. Most of the PGC-1β induced effects on plasma parameters were not affected by treatment with Wy. However, the increase in plasma cholesterol in Wy treated mice was blunted in combination with PGC-1β overexpression. Quantitative realtime PCR showed that transduction with PGC-1α and PGC-1β resulted in 6-fold and 14-fold induction of PGC-1α and PGC-1β mrna levels, respectively. PGC-1α mrna was downregulated by PGC-1β overexpression, but not vice versa. MCAD, ACO-I and Cyp4a10, representing the mitochondrial, peroxisomal and microsomal fatty acid oxidation pathway, respectively, were measured since they are PPARα targets. MCAD was upregulated by both PGC-1α and PGC-1β, whereas ACO-I was not affected and Cyp4a10 was downregulated by PGC-1β only. Wy upregulated all these genes, irrespective of virus treatment, but the stimulatory effect of Wy was less pronounced in combination with PGC-1β overexpression. Similarly, the stimulatory effect of Wy on ADRP and the inhibiting effect on apociii mrna were blunted by PGC-1β overexpression. We 36

37 SUMMARY OF RESULTS suggest that downregulation of PPARα mrna in PGC-1β overexpressing mice may account for this effect. In line with the hypertriglyceridemia, DGAT1 was markedly upregulated by PGC-1β at the mrna and protein level. In contrast, DGAT2 mrna was downregulated by PGC-1β. Furthermore, the de novo lipogenic enzymes fatty acid synthase (FAS) and stearoyl-coa desaturase (SCD)-1 were not upregulated by PGC-1β. In fact SCD-1 mrna was downregulated by both PGC-1s. Despite clear indications of increased VLDL secretion, MTP mrna and protein was not affected by PGC-1β overexpression. In conclusion, hepatic overexpression of PGC-1β, but not PGC-1α, results in combined hyperlipidemia with elevated plasma triglycerides, cholesterol and apob. This effect may partly be explained by upregulation of DGAT1 in combination with elevated plasma FFA levels. PPARα activation does not potentiate the effect of PGC-1α or PGC-1β. Rather, PGC-1β overexpression in vivo blunts the effect of Wy; an effect that may be explained by downregulation of PPARα in PGC-1β overexpressing mice. 37

38 DISCUSSION DISCUSSION GH AND ATHEROSCLEROSIS Although some inbred mouse strains develop small fatty streaks when fed a highly atherogenic diet, wild-type mice are generally resistant to the development of atherosclerosis [181]. GH transgenic mice on C57BL/6 background have no lesions in the aorta (Andersson et al., unpublished results), indicating that GH overexpression alone is not sufficient to induce atherogenesis. In contrast, GH overexpression in apoe-deficient mice, which are susceptible to atherosclerosis, resulted in accelerated lesion development and more severe lesions compared to littermate controls. Thus, high levels of GH may regulate mechanisms that aggravate rather than initiate the atherogenic process. We found that bgh transgenic mice had larger lesion area in the thoracic part of the aorta, whereas no difference was observed in the aortic arch. These regional differences may be explained by previous observations that atherosclerosis appear early in the aortic arch whereas the descending aorta is affected later [182, 183]. Furthermore, atherosclerosis develops in a site-specific fashion with high prevalence of lesions in curvatures and branches, where the blood flow is turbulent. The larger lesion area in the aortic arch compared to the thoracic aorta in our study may therefore reflect both site-specific and time-dependent differences and it is possible that quantification in younger mice would have revealed a difference also in the aortic arch. Although the size of the lesions is of great importance, the composition could actually be more informative in terms of cardiovascular risk. Atherosclerotic lesions mainly composed of foam cells are in general quite harmless, whereas fibrous lesions are more susceptible to disruption. It would be interesting to study the composition of the lesions in more detail, for instance evaluating the relative collagen content since loss of collagen reduces the stability of the lesions. Possible mechanisms behind the accelerated atherosclerosis progression in bgh transgenic mice To find a possible explanation for the accelerated development of atherosclerosis in bgh transgenic mice, we studied some well-known risk factors. Although we concluded that lipids were likely not involved, it should be kept in mind that apoe-deficient mice already have a severe hypercholesterolemia, due to inhibited hepatic uptake of cholesterol-rich apob-containing lipoproteins [158, 159]. The systolic blood pressure was slightly higher in bgh transgenic mice compared to apoe-deficient controls. Since the blood pressure was measured with the tailcuff technique in conscious animals, it could be argued that this difference may reflect different sensitivity to stress between the genotypes. However, it has been shown that bgh transgenic mice have reduced sympathetic responsiveness and decreased plasma levels of noradrenaline [184], indicating that the difference in 38

39 DISCUSSION blood pressure is not due to increased stress response in bgh transgenic mice. In addition, blood pressure measurements with the stress-insensitive telemetry technique have demonstrated that bgh transgenic mice on C57BL/6 background have higher mean arterial blood pressure compared to wild-type controls [146]. Hypertension in acromegalic patients is likely related to sodium retention and increased plasma volume [115]. In contrast, bgh transgenic mice seems to have a salt-independent form of hypertension that instead may be related to increased vascular resistance [146]. Nevertheless, hypertension is highly correlated with atherosclerosis and a study of acromegalic patients showed that the prevalence of hypertension was higher in patients with atherosclerosis compared to those without atherosclerosis [155]. However, it is unlikely that the small difference seen in our animals could be the only explanation for the accelerated process. In order to elucidate the relative importance of hypertension, it could be interesting to examine if the difference in lesion size between the genotypes persists after chronic anti-hypertensive treatment. Inflammation plays an important role in the pathogenesis of atherosclerosis [185], suggesting that an increased inflammatory response could explain the accelerated atherosclerotic process in bgh transgenic animals. The inflammatory proteins SAA and CRP are clearly associated with cardiovascular disease and accumulating evidence suggests that they are not only inflammatory markers but also mediators in atherogenesis [186]. For instance, CRP promotes monocyte recruitment, uptake of oxidized LDL in macrophages and SMC proliferation, whereas SAA impairs HDL-mediated reverse cholesterol transport and attract inflammatory cells [186]. It could be speculated that the increase in HDL cholesterol in bgh transgenic mice is due to impaired SR-BI mediated uptake of HDL cholesterol as a consequence of increased plasma levels of SAA. Studies in mice have shown accelerated progression of atherosclerosis when SAA [187] or CRP [188] is elevated. It is therefore possible that the elevated levels of SAA and the increased hepatic expression of CRP in bgh transgenic mice could contribute to the accelerated lesion development. Another possibility is that GH and/or IGF-I have direct effects in the arterial wall. IGF-I has been shown to stimulate proliferation and migration of vascular SMCs [189] and in vitro studies have suggested that GH/IGF-I promotes uptake of LDL in macrophages [190]. In contrast, IGF-I has been suggested to improve plaque stability by inhibiting apoptosis and necrosis of vascular SMCs and macrophages [191]. The association between IGF-I levels and atherosclerosis is contradictory, since some studies have shown that high IGF-I levels are associated with carotid atherosclerosis [192, 193], whereas others have shown that low IGF-I levels are associated with the occurrence of cardiovascular disease [191]. To investigate the effects of excess GH/IGF-I on early processes in the vascular wall, it would be interesting to study the infiltration of inflammatory cells with different immunohistological methods. Moreover, direct effects of GH on the inflammatory 39

40 DISCUSSION response in macrophages have not been studied in detail. The relative importance of inflammation for the accelerated lesion progression in GH transgenic mice could also be studied by chronic anti-inflammatory treatment. GH AND HEPATIC LIPID METABOLISM Effects of GH on VLDL-triglyceride secretion The effects of GH on lipid and lipoprotein metabolism in animals have almost exclusively been studied in Hx rats treated with GH. A continuous GH infusion to Hx rats has been shown to stimulate VLDL secretion ex vivo [135, 136] and hepatic triglyceride secretion in vivo [137]. In contrast, bgh transgenic mice on a normal diet have lower hepatic triglyceride secretion compared to controls [144], which could reflect model- or species differences. In paper II, we show that these discrepant results likely reflect model differences, since continuous GH infusion to intact mice also resulted in stimulated hepatic triglyceride secretion. The mechanisms by which continuous GH regulates VLDL secretion in Hx rats are not completely understood but has been suggested to involve increased triglyceride synthesis [134, 136] and increased expression of MTP [138]. In addition, apoe is clearly important for VLDL secretion [194] and increased apoe secretion following GH treatment [195] could also be a possible mechanism. From Paper II, we could conclude that increased expression of MTP is not necessary for the increased hepatic triglyceride secretion following continuous GH treatment of intact mice. Thus, it would be interesting to study if apoe levels and triglyceride synthesis are affected in this model. In fact, we observed that GH treatment increased the hepatic expression of DGAT1 and DGAT2 mrna (+55 and +165%, respectively; Ljungberg et al., unpublished results), indicating that increased triglyceride synthesis could be involved. In the Hx model, GH-induced stimulation of VLDL secretion occurs in parallel with increased apob mrna editing [140], LPL and HL activity [141, 142], plasma apoe [139] and hepatic LDLR expression [128]. Together, these effects results in enhanced clearance rate, which may explain that plasma apob levels are reduced and plasma triglycerides are not elevated despite increased secretion [137]. We found that plasma triglycerides were unaffected by GH treatment (Paper II), indicating that triglyceride hydrolysis is enhanced also in our model. In GH treated Hx rats, LPL activity is increased in skeletal muscle [141], indicating distribution of fatty acids derived from triglyceride-rich lipoproteins mainly to that tissue. It is therefore possible that increased skeletal muscle LPL activity accounts for the increased triglyceride turnover also in the model used in Paper II. However, since plasma apob levels were elevated, it could be speculated that apob mrna editing, apoe production or LDLR expression is not affected following short-term GH treatment of intact mice, resulting in slower clearance of apob-containing particles. 40

41 DISCUSSION The effects of GH on lipid and lipoprotein metabolism clearly involve complex regulatory mechanisms and several questions remain to be answered. However, the overall effect of GH seems to involve redistribution of triglycerides from adipose tissue to skeletal muscle. Effects of GH on liver triglycerides and PPARγ The observation that GH treated mice had decreased levels of hepatic triglycerides may result from three potential but non-exclusive mechanisms; decreased fatty acid uptake, increased fatty acid oxidation or increased secretion of VLDL triglycerides. Liver fatty acid binding protein (LFABP), which mediates cellular uptake of fatty acids [196], is upregulated by GH both in vivo and in vitro in rats [197], indicating an increased fatty acid uptake following GH treatment. These results, together with our observation that LFABP mrna expression was about 50% higher in GH-treated mice (Ljungberg et al., unpublished results), indicate that decreased fatty acid uptake is probably not the mechanism explaining the lower triglyceride content. In addition, the inhibitory effect of GH on ACO and MCAD expression, which is in line with previous studies [ ], strongly suggest that both mitochondrial and peroxisomal β-oxidation rather decrease by GH. Although neither changed fatty acid uptake nor changed oxidation of fatty acids can be excluded, the reduced amount of liver triglycerides in GH treated mice most likely results from the increased triglyceride secretion. An increased prevalence of hepatic steatosis has been observed in GH-deficient subjects [201]. Thus, normal levels of GH seem to be important not only for plasma triglycerides but also for liver triglycerides. The reduced levels of hepatic triglycerides in GH-treated mice were accompanied by decreased expression of both isoforms of PPARγ (Paper II). This was not very surprising since PPARγ expression has been shown to correlate with hepatic steatosis. Liver-specific disruption of PPARγ resulted in markedly reduced levels of liver triglycerides in ob/ob mice [202] and in A-ZIP/F1 mice [203], which normally display hepatic steatosis. Furthermore, overexpression of hepatic PPARγ1 induced steatosis both in PPARα-deficient and wild-type mice [46]. Also PPARγ2 has been shown to induce triglyceride accumulation when overexpressed in the hepatic cell line AML-12 [47]. Thus, it could be speculated that GH reduces hepatic triglycerides by downregulation of PPARγ expression. 41

42 DISCUSSION PPARα AND HEPATIC LIPID METABOLISM Effects of PPARα activation on apob-containing lipoproteins Hepatic overproduction of triglyceride-rich VLDL1 is though to be the most important determinant for development of atherogenic dyslipidemia in humans [33]. This event is largely counteracted by fibrates, partly due to decreased secretion of VLDL-triglycerides from the liver. However, the mechanisms whereby fibrates regulate the amount of triglycerides incorporated into VLDL are not fully understood and have therefore been studied in Paper III. We have previously shown that Wy treatment of rat hepatocytes decreased triglyceride synthesis and secretion [65]. Surprisingly, the secretion of apob100 was increased, due to decreased co-translational degradation. Since each VLDL particle only consists of one apob, Wy treatment results in a shift from secretion of large triglyceride-rich particles to secretion of an increased number of smaller and triglyceride-poor particles. In line with the concept that MTP-dependent lipidation of apob is crucial to avoid degradation of apob, we studied the effect of Wy on MTP expression and found that increased apob100 secretion was paralleled by increased MTP expression in rat hepatocytes [204]. Thus, we believe that MTP mediates the stimulatory effect of Wy on apob100 secretion via decreased cotranslational degradation. We also showed that Wy treatment in vivo increased MTP mrna and protein levels as well as activity in the liver of both rats and mice [204]. Furthermore, Wy treatment of mice resulted in decreased levels of plasma triglycerides and apob, strengthening the hypothesis that Wy treatment in vivo actually stimulates the secretion of triglyceride-poor apob-containing particles that are rapidly catabolized. Our previous studies in rat hepatocytes indicated that decreased triglyceride synthesis could contribute to the reduced triglyceride secretion after Wy treatment [65]. However, we showed in Paper III that Wy decreased triglyceride secretion without affecting triglyceride synthesis in mouse hepatocytes, indicating that other mechanisms must be important. In contrast to the general believe, we could also conclude that limited availability of intracellular lipids due to increased fatty acid oxidation cannot explain the decreased triglyceride secretion, since Wy treatment resulted in triglyceride accumulation. Instead, our current hypothesis is that PPARα agonists prevent the use of cytosolic triglycerides for VLDL assembly, at least partly via increased ADRP expression. In Paper III, we suggested DGAT1 and DGAT2 as potential mediators of the balance between triglyceride storage and secretion, since rats treated with fenofibrate had decreased DGAT activity inside the ER (DGAT1) and increased DGAT activity in ER facing the cytoplasm (DGAT2), in parallel with increased hepatic triglyceride content [205]. However, mice treated with Wy had unchanged mrna levels of DGAT1 and decreased mrna levels of DGAT2 (Ljungberg et al., unpublished results). The discrepant findings may reflect post-transcriptional effects but could also result from species- 42

43 DISCUSSION differences. Thus, other mechanisms besides increased ADRP expression that promote storage of triglycerides at the expense of triglyceride secretion await further studies. Effects of PPARα-deficiency on apob-containing lipoproteins The importance of PPARα for the regulation of apob-containing lipoprotein is further strengthened by studies in PPARα-deficient mice. In line with our previous study [75], we observed that female PPARα-deficient mice had higher hepatic triglyceride secretion rate as well as elevated serum levels of apob compared to wild-type mice (Paper II), whereas apob were unaffected and hepatic triglyceride secretion rate were less affected in males (Ljungberg et al., unpublished results). Like an earlier report [70], we found no difference in MTP mrna expression between the male genotypes (Ljungberg et al., unpublished results). In contrast, female PPARα-deficient mice expressed higher levels of MTP compared to controls (Paper II). This sex-difference in response to PPARαdeficiency may help to explain the higher triglyceride secretion rate and serum apob levels in female PPARα-deficient mice compared to wild-type controls. It is also possible that triglyceride synthesis is enhanced in female but not in male PPARα-deficient livers. To the best of our knowledge, this has not been investigated. In contrast to females, the hepatic triglycerides are not directed to VLDL secretion to the same extent in male PPARα-deficient mice, but rather accumulate in the liver. Male PPARα-deficient mice expressed higher levels of PPARγ1 and PPARγ2 (+155% and 70%, respectively; Ljungberg et al., unpublished results), which correlated well with hepatic triglyceride accumulation (+600%; Ljungberg et al., unpublished results). In line with the suggested function of ADRP to promote lipid storage, we observed in Paper III that PPARα-deficient males had increased hepatic protein levels of ADRP. It is interesting to note that both PPARα activation and PPARα-deficiency results in hepatic triglyceride accumulation. However, this effect seems to result from different mechanisms. Wy treatment of mice increases both ADRP mrna and protein levels in parallel with increased triglyceride content (Paper III), indicating that triglycerides accumulate as a result of increased ADRP expression. This is in line with the observation that overexpression of ADRP promotes lipid accumulation [93, 94]. In contrast, PPARα-deficient mice have higher ADRP protein levels than littermate controls, although mrna levels are lower. This may be explained by post-translational mechanisms that involves stabilization of ADRP by association with lipid droplets, since in vitro studies have shown that ADRP is stable in the presence of oleic acid, whereas removal of the fatty acid source results in proteasomal degradation of ADRP [206]. This mechanism may also help to explain our result that hepatic ADRP protein levels were higher in livers from mice fed a high-fat diet. 43

44 DISCUSSION Do triglycerides accumulate in human liver upon PPARα activation? In Paper III, we observed that Wy treatment resulted in hepatic accumulation of triglycerides both in vivo and in vitro. This finding was surprising since there are no published reports indicating similar effects in humans, although fibrates have been used in clinic for several decades. It is possible that PPARα agonists have similar effects on ADRP expression and triglycerides in human liver, since the human ADRP promoter contains a PPRE [102]. However, since ADRP is stabilized in a lipid-rich environment [206] it could be expected that most patients on fibrates already have increased ADRP expression as a consequence of liver steatosis [207]. Thus, a further transcriptional effect would only result in a minor change in ADRP protein and thus liver lipids. This reasoning is in line with our observation of a smaller change in liver triglycerides after administration of Wy to mice on high fat diet as compared to Wy treatment of mice on low fat diet (Paper III). Accordingly, lean healthy subjects would have a more marked increase in liver triglycerides by PPARα agonist treatment. The hypothesis that Wy may have different impact depending on the lipid load could be tested by comparing the effect of Wy on normal hepatocytes and hepatocytes that have been loaded with lipids, for instance by incubation with oleic acid. Another aspect is that triglyceride accumulation can not increase continuously. It is possible that the hepatic triglyceride levels soon reach a steady state where storage is balanced by increased fatty acid oxidation. It is also possible that species-differences exists, similar to the observation that rodents respond to PPARα agonists with peroxisome proliferation, hepatomegaly and tumor development, whereas this is not the case in humans [165]. The reason for such differences is unknown, but could result from the 10 times higher basal expression of PPARα in rodents compared to humans. Furthermore, a human splice variant that lacks the ligand binding domain has been identified and shown to represent 20-50% of total hepatic PPARα mrna [165]. To examine if the triglyceride accumulation is restricted to rodents, it would be interesting to study the effect of Wy on primary hepatocytes derived from humans. PGC-1β AND HEPATIC LIPID METABOLISM Possible mechanisms behind the combined hyperlipidemia PGC-1β clearly plays an important role in the regulation of hepatic lipid metabolism, as shown by markedly elevated plasma levels of cholesterol, triglycerides and apob in Paper IV. However, the mechanisms behind this effect are less clear. Elevated plasma triglycerides could result from increased triglyceride secretion and/or decreased triglyceride clearance. To evaluate the possible importance of clearance, we measured the hepatic gene expression of angiopoietin-like protein 3 (AngPtl3) and apociii, which are both known to inhibit LPL-mediated triglyceride clearance. However, neither of them was affected by PGC-1β overexpression, indicating that clearance may not be the major 44

45 DISCUSSION mechanism. Instead, increased DGAT1 levels suggest enhanced triglyceride secretion, since adenoviral overexpression of DGAT1 in mice has been shown to increase VLDL secretion [30]. Although we cannot exclude other possibilities until we have investigated hepatic triglyceride secretion in PGC-1β overexpressing mice, these results suggest that increased triglyceride secretion is the major event leading to the hyperlipidemia. Since the amount of liver triglycerides were unaffected by PGC-1β overexpression, a stimulated triglyceride secretion would in that case result from increased fatty acid synthesis, decreased fatty acid oxidation or increased flux of FFA to the liver. Increased de novo lipogenesis is less likely since the expression of lipogenic genes were unaffected or rather decreased in PGC-1β overexpressing mice. Furthermore, in vitro studies using primary mouse hepatocytes showed that triglyceride synthesis was decreased in cells transduced with PGC-1β (-40%; Ljungberg et al., unpublished results). In contrast, fatty acid oxidation was markedly increased in these cells (+130%; Ljungberg et al., unpublished results), demonstrating that decreased fatty acid oxidation is unlikely as well. Thus, the most probable explanation is an increased FFA flux to the liver due to increased lipolysis in adipose tissue. This hypothesis is strengthened by the elevation of FFA in plasma. In addition, the adipose tissue weight was 27% lower in PGC-1β overexpressing mice, which could indicate increased lipolysis (Ljungberg et al., unpublished results). Since overexpression of DGAT1 in the liver resulted in increased gonadal fat mass [30], the lower gonadal fat weight in our study indicates an additional effect of PGC-1β overexpression on adipose tissue. Our in vitro studies showed that triglyceride secretion actually was decreased in cells transduced with PGC-1β, although DGAT1 was upregulated. We believe that these cells are running out of substrate, due to markedly elevated fatty acid oxidation. Thus, it would be interesting to try the hypothesis that addition of fatty acids to the culture media may stimulate triglyceride secretion. Since PGC-1β is overexpressed only in the liver, it could be anticipated that increased lipolysis in adipose tissue in that case would be mediated by a liver-derived factor secreted into the circulation. To examine this possibility, it would be interesting to incubate cultured adipocytes with serum collected from mice overexpressing PGC-1β and then measure lipolysis. Moreover, a DNA array on livers from mice overexpressing PGC-1β could suggest potential candidates. PGC-1β as a potential pharmaceutical target? The fact that overexpression of PGC-1β results in a combined hyperlipidemia makes it tempting to speculate that modulation of PCG-1β expression and activity in the liver could be an important pharmaceutical target. Since PGC-1β coactivates a broad range of transcription factors [80], it is however likely that such modulation may affect important functions of other pathways. It is therefore necessary to find out which transcription factor(s) is mediating these effects. Lin et 45

46 DISCUSSION al. suggested SREBP and LXR as mediators of the hyperlipidemic response of PGC-1β [85], whereas we observed a similar response without increasing wellknown targets for SREBP (e.g. SCD-1 and FAS) and LXR (e.g. Cyp7a1). Thus, extensive research remains in order to fully understand the mechanisms behind the PGC-1β induced hyperlipidemia. We observed that PGC-1β overexpression blunted the effect of Wy on several genes involved in lipid metabolism, possibly due to downregulation of PPARα. Furthermore, dietary saturated fats were shown to induce PGC-1β expression [85]. Taken together, these results could imply that treatment with a PPARα agonist may be less efficient in subjects eating a diet rich in saturated fats. 46

47 SUMMARY AND CONCLUSIONS SUMMARY AND CONCLUSIONS High GH levels accelerated the development of atherosclerotic lesions in apoedeficient mice and resulted in larger lesion area in the thoracic aorta and more severe lesions in the aortic sinus. The lipid profile was not worsened by GH overproduction, which implies that the accelerated progression of atherosclerosis is mediated by other factors than lipids. Possible mechanisms behind this effect could be direct effects of GH in the vascular wall or the result of concomitant processes such as hypertension or a general inflammatory process. The apoe -/- /bgh mouse provides a good model to further study the effects of GH on atherosclerosis. GH treatment of mice increased hepatic triglyceride secretion, plasma triglyceride clearance and HDL cholesterol levels. In contrast, the amount of liver triglycerides was reduced in parallel with decreased expression of PPARγ. The effect of GH on PPARγ2 and Cyp4a10 mrna expression was PPARα-dependent, suggesting that PPARα may be important for the effects of GH on PPARγ signaling and ω-oxidation. However, most of the studied effects of GH were similar in wild-type mice and PPARα-deficient mice, indicating that PPARα plays a minor role for the effects of GH on hepatic lipid and lipoprotein metabolism. PPARα activation resulted in increased hepatic triglyceride content in parallel with increased ADRP protein expression, both in mouse liver and primary mouse hepatocytes. Accumulation of hepatic triglycerides was accompanied by decreased triglyceride secretion, both after PPARα activation and ADRP overexpression. However, concomitant PPARα activation and ADRP overexpression resulted in further decreased triglyceride secretion despite unaffected ADRP protein levels, indicating that increased ADRP expression is not the only mechanism explaining the decreased triglyceride secretion. We can conclude that PPARα activation does not primarily decrease triglyceride secretion via enhanced fatty acid oxidation, since the intracellular triglyceride content of hepatocytes is not limiting. Instead, PPARα activation prevents the use of cytosolic triglycerides for secretion, at least partly via increased ADRP expression. Hepatic overexpression of PGC-1β, but not PGC-1α, resulted in a combined hyperlipidemia with elevated plasma triglycerides, cholesterol and apob. This effect was associated with increased DGAT1 expression and plasma FFA. The potentially beneficial effects of PPARα activation on genes controlling lipid metabolism were blunted by PGC-1β overexpression, which may be explained by downregulation of PPARα. 47

48 ACKNOWLEDGEMENTS ACKNOWLEDGEMENTS Jag är tacksam för välbehövlig hjälp och stort stöd från kollegor, vänner och familj under min doktorandperiod. Framför allt skulle jag vilja tacka: Min handledare Jan Oscarsson för ovärderlig hjälp med allt tänkbart under mina år som doktorand. Din breda kunskap, stora entusiasm och fantastiska förmåga att vända alla resultat till något positivt gör dig till en mycket inspirerande och bra handledare! Min biträdande handledare Daniel Lindén för stort engagemang i alla mina projekt och för värdefull hjälp med såväl försök som artiklar och avhandling. Du har alltid kloka synpunkter och det har varit mycket lärorikt och trevligt att samarbeta med dig! Ulrika Edvardsson för mycket trevligt samarbete och för givande vetenskapliga och ovetenskapliga diskussioner. Du är en härlig kollega som får åtskilliga timmar i cellodlingen att kännas betydligt mycket kortare! Ditt partysinne går inte heller av för hackor! Caroline Améen för vänskap, gott samarbete, många skratt och mycket trevligt sällskap under åren, inte minst på alla konferenser. Vår Road Trip längs Highway One är ett minne för livet! Linda Carlsson och Fredrik Frick för att ni under mitt första år som doktorand bidrog till en trivsam stämning i gruppen. Ett extra tack till Fredrik för samarbete med hepatocytförsöken på AstraZeneca och för roligt sällskap i HET-labbet under de sista hektiska månaderna. Nu förstår jag vad stress-konen innebär! Birgitta Odén för trevligt sällskap på Lab 7 och för att du alltid har ställt upp när det behövts en extra hand. Att du till och med tog dig an Gårda-uppdraget säger en del om din hjälpsamhet! Irene Andersson för trevligt och lärorikt samarbete på alla sätt. Jag beundrar ditt enorma tålamod med att pinna upp aortor! Göran Bergström för att du delat med dig av dina kunskaper om ateroskleros och för värdefulla synpunkter på arbete I och II. Alla trevliga och hjälpsamma människor jag har mött under min sista tid på AstraZeneca men framför allt ni som har varit inblandade i PGC-1 projektet: Chris Lelliott for excellent experimental help and scientific support, for nice company and for proof-reading my thesis. Andrea Ahnmark för utmärkt assistans vid leverperfusioner, för hjälp i cellodlingen och djurhuset samt för att ditt glada humör och din härliga humor gör det så mycket trevligare att samarbeta. Lena William-Olsson för all hjälp med virusförsöken och för mycket trevligt sällskap på lab. Jag har verkligen känt mig som en i gruppen! 48

49 ACKNOWLEDGEMENTS Övriga medförfattare för laborativa insatser och värdefulla kommentarer på manuskripten. Heimir Snorrason för datorexpertis, Mujtaba Siddiqui för artikelservice, Merja Meuronen-Österholm och Claudia Abid för att ni alltid har sett till att hålla perfusionslådan redo för nästa omgång samt Agneta Ladström, Ewa Landegren och Lena Olofsson för att ni har haft stenkoll på alla papper och pengar. Vänner och kollegor från Endokrin och Wallenberglab för att ni har skapat en trevlig och stimulerande arbetsmiljö. Goda arbetskamrater är enormt viktigt för mig och jag kunde inte ha fått bättre! Speciellt vill jag tacka Lisa Buvall och Malin Lorentzon för allt roligt vi har upplevt tillsammans, både på lab och privat. Ni är två helt fantastiska vänner! Karin, Sofia, Ann, Carl-Fredrik, Fredrik och Thomas för alla roliga fester och mini-semestrar från våra tidiga år på universitetet och för fortsatt god vänskap därefter. Tänk att det blev doktorer av oss alla till sist! Min barndomsvän Johanna för alla oförglömliga minnen sedan första skoldagen. Du kommer alltid att betyda mycket för mig. Lycka till i Wien! Åsa, Hedvig, Ida och Cissi för all vänskap och för att ni alltid finns där. Mina föräldrar, Gun-Britt och Bosse, min bror Johan, Jenny och lilla Amanda för att ni är världens bästa familj. Ni har verkligen stöttat och uppmuntrat mig under åren och intresserat er för vad jag gör, trots att det verkar obegripligt. Nu har ni en hel bok att läsa! Ronny för all kärlek, omtänksamhet och förståelse. Du gör mig lycklig! 49

50 REFERENCES REFERENCES 1. Reisin, E. and Alpert, M.A., Definition of the metabolic syndrome: current proposals and controversies. Am J Med Sci, (6): p Yach, D., Stuckler, D., and Brownell, K.D., Epidemiologic and economic consequences of the global epidemics of obesity and diabetes. Nat Med, (1): p Wright, A.D., Hill, D.M., Lowy, C., and Fraser, T.R., Mortality in acromegaly. Q J Med, (153): p Rosén, T. and Bengtsson, B.A., Premature mortality due to cardiovascular disease in hypopituitarism. Lancet, (8710): p Franek, E., Schaefer, F., Bergis, K., Feneberg, R., and Ritz, E., Abnormal pulsatile secretion of growth hormone in non-insulin-dependent diabetes mellitus. Clin Endocrinol (Oxf), (4): p Chen, S.H., Habib, G., Yang, C.Y., Gu, Z.W., Lee, B.R., Weng, S.A., Silberman, S.R., Cai, S.J., Deslypere, J.P., Rosseneu, M., and et al., Apolipoprotein B-48 is the product of a messenger RNA with an organ- specific in-frame stop codon. Science, (4825): p Powell, L.M., Wallis, S.C., Pease, R.J., Edwards, Y.H., Knott, T.J., and Scott, J., A novel form of tissue-specific RNA processing produces apolipoprotein- B48 in intestine. Cell, (6): p von Eckardstein, A., Hersberger, M., and Rohrer, L., Current understanding of the metabolism and biological actions of HDL. Curr Opin Clin Nutr Metab Care, (2): p Borén, J., Graham, L., Wettesten, M., Scott, J., White, A., and Olofsson, S.O., The assembly and secretion of ApoB 100-containing lipoproteins in Hep G2 cells. ApoB 100 is cotranslationally integrated into lipoproteins. J Biol Chem, (14): p Borén, J., Rustaeus, S., and Olofsson, S.O., Studies on the assembly of apolipoprotein B and B-48-containing very low density lipoproteins in McA-RH7777 cells. J Biol Chem, (41): p Olofsson, S.O., Asp, L., and Boren, J., The assembly and secretion of apolipoprotein B- containing lipoproteins. Curr Opin Lipidol, (4): p Raabe, M., Veniant, M.M., Sullivan, M.A., Zlot, C.H., Bjorkegren, J., Nielsen, L.B., Wong, J.S., Hamilton, R.L., and Young, S.G., Analysis of the role of microsomal triglyceride transfer protein in the liver of tissue-specific knockout mice. J Clin Invest, (9): p Gordon, D.A. and Jamil, H., Progress towards understanding the role of microsomal triglyceride transfer protein in apolipoprotein-b lipoprotein assembly. Biochim Biophys Acta, (1): p Wetterau, J.R., Aggerbeck, L.P., Bouma, M.E., Eisenberg, C., Munck, A., Hermier, M., Schmitz, J., Gay, G., Rader, D.J., and Gregg, R.E., Absence of microsomal triglyceride transfer protein in individuals with abetalipoproteinemia. Science, (5084): p

51 REFERENCES 15. Sharp, D., Blinderman, L., Combs, K.A., Kienzle, B., Ricci, B., Wager-Smith, K., Gil, C.M., Turck, C.W., Bouma, M.E., Rader, D.J., and et al., Cloning and gene defects in microsomal triglyceride transfer protein associated with abetalipoproteinaemia. Nature, (6441): p Berriot-Varoqueaux, N., Aggerbeck, L.P., Samson-Bouma, M., and Wetterau, J.R., The role of the microsomal triglygeride transfer protein in abetalipoproteinemia. Annu Rev Nutr, : p Raabe, M., Flynn, L.M., Zlot, C.H., Wong, J.S., Veniant, M.M., Hamilton, R.L., and Young, S.G., Knockout of the abetalipoproteinemia gene in mice: reduced lipoprotein secretion in heterozygotes and embryonic lethality in homozygotes. Proc Natl Acad Sci U S A, (15): p Wetterau, J.R., Gregg, R.E., Harrity, T.W., Arbeeny, C., Cap, M., Connolly, F., Chu, C.H., George, R.J., Gordon, D.A., Jamil, H., Jolibois, K.G., Kunselman, L.K., Lan, S.J., Maccagnan, T.J., Ricci, B., Yan, M., Young, D., Chen, Y., Fryszman, O.M., Logan, J.V., Musial, C.L., Poss, M.A., Robl, J.A., Simpkins, L.M., Slusarchyk, W.A., Sulsky, R., Taunk, P., Magnin, D.R., Tino, J.A., Lawrence, R.M., Dickson, J.K., Jr., and Biller, S.A., An MTP inhibitor that normalizes atherogenic lipoprotein levels in WHHL rabbits. Science, (5389): p Tietge, U.J., Bakillah, A., Maugeais, C., Tsukamoto, K., Hussain, M., and Rader, D.J., Hepatic overexpression of microsomal triglyceride transfer protein (MTP) results in increased in vivo secretion of VLDL triglycerides and apolipoprotein B. J Lipid Res, (11): p Liao, W., Kobayashi, K., and Chan, L., Adenovirus-mediated overexpression of microsomal triglyceride transfer protein (MTP): mechanistic studies on the role of MTP in apolipoprotein B-100 biogenesis, by. Biochemistry, (31): p Fisher, E.A., Zhou, M., Mitchell, D.M., Wu, X., Omura, S., Wang, H., Goldberg, A.L., and Ginsberg, H.N., The degradation of apolipoprotein B100 is mediated by the ubiquitinproteasome pathway and involves heat shock protein 70. J Biol Chem, (33): p Yeung, S.J., Chen, S.H., and Chan, L., Ubiquitin-proteasome pathway mediates intracellular degradation of apolipoprotein B. Biochemistry, (43): p Gibbons, G.F., Assembly and secretion of hepatic very-low-density lipoprotein. Biochem J, (1): p Julius, U., Influence of plasma free fatty acids on lipoprotein synthesis and diabetic dyslipidemia. Exp Clin Endocrinol Diabetes, (5): p Gibbons, G.F., Bartlett, S.M., Sparks, C.E., and Sparks, J.D., Extracellular fatty acids are not utilized directly for the synthesis of very-low-density lipoprotein in primary cultures of rat hepatocytes. Biochem J, (Pt 3): p Gibbons, G.F., Islam, K., and Pease, R.J., Mobilisation of triacylglycerol stores. Biochim Biophys Acta, (1): p Dolinsky, V.W., Gilham, D., Alam, M., Vance, D.E., and Lehner, R., Triacylglycerol hydrolase: role in intracellular lipid metabolism. Cell Mol Life Sci, (13): p Trickett, J.I., Patel, D.D., Knight, B.L., Saggerson, E.D., Gibbons, G.F., and Pease, R.J., Characterization of the rodent genes for arylacetamide deacetylase, a putative microsomal 51

52 REFERENCES lipase, and evidence for transcriptional regulation. J Biol Chem, (43): p Owen, M.R., Corstorphine, C.C., and Zammit, V.A., Overt and latent activities of diacylglycerol acytransferase in rat liver microsomes: possible roles in very-low-density lipoprotein triacylglycerol secretion. Biochem J, ( Pt 1): p Yamazaki, T., Sasaki, E., Kakinuma, C., Yano, T., Miura, S., and Ezaki, O., Increased very low density lipoprotein secretion and gonadal fat mass in mice overexpressing liver DGAT1. J Biol Chem, (22): p Liang, J.J., Oelkers, P., Guo, C., Chu, P.C., Dixon, J.L., Ginsberg, H.N., and Sturley, S.L., Overexpression of human diacylglycerol acyltransferase 1, acyl-coa:cholesterol acyltransferase 1, or acyl-coa:cholesterol acyltransferase 2 stimulates secretion of apolipoprotein B-containing lipoproteins in McA-RH7777 cells. J Biol Chem, (43): p Lin, M.C., Gordon, D., and Wetterau, J.R., Microsomal triglyceride transfer protein (MTP) regulation in HepG2 cells: insulin negatively regulates MTP gene expression. J Lipid Res, (5): p Taskinen, M.R., Diabetic dyslipidaemia: from basic research to clinical practice. Diabetologia, (6): p Packard, C.J., Triacylglycerol-rich lipoproteins and the generation of small, dense lowdensity lipoprotein. Biochem Soc Trans, (Pt 5): p Lusis, A.J., Atherosclerosis. Nature, (6801): p Walldius, G., Jungner, I., Holme, I., Aastveit, A.H., Kolar, W., and Steiner, E., High apolipoprotein B, low apolipoprotein A-I, and improvement in the prediction of fatal myocardial infarction (AMORIS study): a prospective study. Lancet, (9298): p Williams, K.J. and Tabas, I., The response-to-retention hypothesis of early atherogenesis. Arterioscler Thromb Vasc Biol, (5): p Skålen, K., Gustafsson, M., Rydberg, E.K., Hulten, L.M., Wiklund, O., Innerarity, T.L., and Borén, J., Subendothelial retention of atherogenic lipoproteins in early atherosclerosis. Nature, (6890): p Hansson, G.K., Inflammation, atherosclerosis, and coronary artery disease. N Engl J Med, (16): p Napoli, C., D'Armiento, F.P., Mancini, F.P., Postiglione, A., Witztum, J.L., Palumbo, G., and Palinski, W., Fatty streak formation occurs in human fetal aortas and is greatly enhanced by maternal hypercholesterolemia. Intimal accumulation of low density lipoprotein and its oxidation precede monocyte recruitment into early atherosclerotic lesions. J Clin Invest, (11): p Kliewer, S.A., Umesono, K., Noonan, D.J., Heyman, R.A., and Evans, R.M., Convergence of 9-cis retinoic acid and peroxisome proliferator signalling pathways through heterodimer formation of their receptors. Nature, (6389): p Kliewer, S.A., Sundseth, S.S., Jones, S.A., Brown, P.J., Wisely, G.B., Koble, C.S., Devchand, P., Wahli, W., Willson, T.M., Lenhard, J.M., and Lehmann, J.M., Fatty acids and eicosanoids regulate gene expression through direct interactions with peroxisome 52

53 REFERENCES proliferator-activated receptors alpha and gamma. Proc Natl Acad Sci U S A, (9): p Desvergne, B. and Wahli, W., Peroxisome proliferator-activated receptors: nuclear control of metabolism. Endocr Rev, (5): p Lemberger, T., Saladin, R., Vazquez, M., Assimacopoulos, F., Staels, B., Desvergne, B., Wahli, W., and Auwerx, J., Expression of the peroxisome proliferator-activated receptor alpha gene is stimulated by stress and follows a diurnal rhythm. J Biol Chem, (3): p Zhu, Y., Qi, C., Korenberg, J.R., Chen, X.N., Noya, D., Rao, M.S., and Reddy, J.K., Structural organization of mouse peroxisome proliferator-activated receptor gamma (mppar gamma) gene: alternative promoter use and different splicing yield two mppar gamma isoforms. Proc Natl Acad Sci U S A, (17): p Yu, S., Matsusue, K., Kashireddy, P., Cao, W.Q., Yeldandi, V., Yeldandi, A.V., Rao, M.S., Gonzalez, F.J., and Reddy, J.K., Adipocyte-specific gene expression and adipogenic steatosis in the mouse liver due to peroxisome proliferator-activated receptor gamma1 (PPARgamma1) overexpression. J Biol Chem, (1): p Schadinger, S.E., Bucher, N.L., Schreiber, B.M., and Farmer, S.R., PPARgamma2 regulates lipogenesis and lipid accumulation in steatotic hepatocytes. Am J Physiol Endocrinol Metab, (6): p. E Michalik, L., Desvergne, B., Tan, N.S., Basu-Modak, S., Escher, P., Rieusset, J., Peters, J.M., Kaya, G., Gonzalez, F.J., Zakany, J., Metzger, D., Chambon, P., Duboule, D., and Wahli, W., Impaired skin wound healing in peroxisome proliferator-activated receptor (PPAR)alpha and PPARbeta mutant mice. J Cell Biol, (4): p Akiyama, T.E., Lambert, G., Nicol, C.J., Matsusue, K., Peters, J.M., Brewer, H.B., Jr., and Gonzalez, F.J., Peroxisome proliferator-activated receptor beta/delta regulates very low density lipoprotein production and catabolism in mice on a Western diet. J Biol Chem, (20): p Mandard, S., Muller, M., and Kersten, S., Peroxisome proliferator-activated receptor alpha target genes. Cell Mol Life Sci, (4): p Martin, G., Schoonjans, K., Lefebvre, A.M., Staels, B., and Auwerx, J., Coordinate regulation of the expression of the fatty acid transport protein and acyl-coa synthetase genes by PPARalpha and PPARgamma activators. J Biol Chem, (45): p Motojima, K., Passilly, P., Peters, J.M., Gonzalez, F.J., and Latruffe, N., Expression of putative fatty acid transporter genes are regulated by peroxisome proliferator-activated receptor alpha and gamma activators in a tissue- and inducer-specific manner. J Biol Chem, (27): p Schoonjans, K., Watanabe, M., Suzuki, H., Mahfoudi, A., Krey, G., Wahli, W., Grimaldi, P., Staels, B., Yamamoto, T., and Auwerx, J., Induction of the acyl-coenzyme A synthetase gene by fibrates and fatty acids is mediated by a peroxisome proliferator response element in the C promoter. J Biol Chem, (33): p Brady, P.S., Marine, K.A., Brady, L.J., and Ramsay, R.R., Co-ordinate induction of hepatic mitochondrial and peroxisomal carnitine acyltransferase synthesis by diet and drugs. Biochem J, (1): p

54 REFERENCES 55. Gulick, T., Cresci, S., Caira, T., Moore, D.D., and Kelly, D.P., The peroxisome proliferator-activated receptor regulates mitochondrial fatty acid oxidative enzyme gene expression. Proc Natl Acad Sci U S A, (23): p Tugwood, J.D., Issemann, I., Anderson, R.G., Bundell, K.R., McPheat, W.L., and Green, S., The mouse peroxisome proliferator activated receptor recognizes a response element in the 5' flanking sequence of the rat acyl CoA oxidase gene. Embo J, (2): p Kroetz, D.L., Yook, P., Costet, P., Bianchi, P., and Pineau, T., Peroxisome proliferatoractivated receptor alpha controls the hepatic CYP4A induction adaptive response to starvation and diabetes. J Biol Chem, (47): p Aldridge, T.C., Tugwood, J.D., and Green, S., Identification and characterization of DNA elements implicated in the regulation of CYP4A1 transcription. Biochem J, ( Pt 2): p Sirtori, C.R. and Franceschini, G., Effects of fibrates on serum lipids and atherosclerosis. Pharmacol Ther, (2): p Investigators, D.A.I.S., Effect of fenofibrate on progression of coronary-artery disease in type 2 diabetes: the Diabetes Atherosclerosis Intervention Study, a randomised study. Lancet, (9260): p Caslake, M.J., Packard, C.J., Gaw, A., Murray, E., Griffin, B.A., Vallance, B.D., and Shepherd, J., Fenofibrate and LDL metabolic heterogeneity in hypercholesterolemia. Arterioscler Thromb, (5): p Guerin, M., Le Goff, W., Frisdal, E., Schneider, S., Milosavljevic, D., Bruckert, E., and Chapman, M.J., Action of ciprofibrate in type IIb hyperlipoproteinemia: modulation of the atherogenic lipoprotein phenotype and stimulation of high-density lipoprotein-mediated cellular cholesterol efflux. J Clin Endocrinol Metab, (8): p Kesäniemi, Y.A. and Grundy, S.M., Influence of gemfibrozil and clofibrate on metabolism of cholesterol and plasma triglycerides in man. Jama, (17): p Petit, D., Bonnefis, M.T., Rey, C., and Infante, R., Effects of ciprofibrate and fenofibrate on liver lipids and lipoprotein synthesis in normo- and hyperlipidemic rats. Atherosclerosis, (3): p Lindén, D., Lindberg, K., Oscarsson, J., Claesson, C., Asp, L., Li, L., Gustafsson, M., Borén, J., and Olofsson, S.O., Influence of peroxisome proliferator-activated receptor alpha agonists on the intracellular turnover and secretion of apolipoprotein (Apo) B-100 and ApoB-48. J Biol Chem, (25): p Schoonjans, K., Peinado-Onsurbe, J., Lefebvre, A.M., Heyman, R.A., Briggs, M., Deeb, S., Staels, B., and Auwerx, J., PPARalpha and PPARgamma activators direct a distinct tissue-specific transcriptional response via a PPRE in the lipoprotein lipase gene. Embo J, (19): p Hertz, R., Bishara-Shieban, J., and Bar-Tana, J., Mode of action of peroxisome proliferators as hypolipidemic drugs. Suppression of apolipoprotein C-III. J Biol Chem, (22): p Berthou, L., Duverger, N., Emmanuel, F., Langouet, S., Auwerx, J., Guillouzo, A., Fruchart, J.C., Rubin, E., Denefle, P., Staels, B., and Branellec, D., Opposite regulation of human versus mouse apolipoprotein A-I by fibrates in human apolipoprotein A-I transgenic mice. J Clin Invest, (11): p

55 REFERENCES 69. Vu-Dac, N., Schoonjans, K., Kosykh, V., Dallongeville, J., Fruchart, J.C., Staels, B., and Auwerx, J., Fibrates increase human apolipoprotein A-II expression through activation of the peroxisome proliferator-activated receptor. J Clin Invest, (2): p Kersten, S., Seydoux, J., Peters, J.M., Gonzalez, F.J., Desvergne, B., and Wahli, W., Peroxisome proliferator-activated receptor alpha mediates the adaptive response to fasting. J Clin Invest, (11): p Leone, T.C., Weinheimer, C.J., and Kelly, D.P., A critical role for the peroxisome proliferator-activated receptor alpha (PPARalpha) in the cellular fasting response: the PPARalpha-null mouse as a model of fatty acid oxidation disorders. Proc Natl Acad Sci U S A, (13): p Le May, C., Pineau, T., Bigot, K., Kohl, C., Girard, J., and Pegorier, J.P., Reduced hepatic fatty acid oxidation in fasting PPARalpha null mice is due to impaired mitochondrial hydroxymethylglutaryl-coa synthase gene expression. FEBS Lett, (3): p Costet, P., Legendre, C., More, J., Edgar, A., Galtier, P., and Pineau, T., Peroxisome proliferator-activated receptor alpha-isoform deficiency leads to progressive dyslipidemia with sexually dimorphic obesity and steatosis. J Biol Chem, (45): p Akiyama, T.E., Nicol, C.J., Fievet, C., Staels, B., Ward, J.M., Auwerx, J., Lee, S.S., Gonzalez, F.J., and Peters, J.M., Peroxisome proliferator-activated receptor-alpha regulates lipid homeostasis, but is not associated with obesity: studies with congenic mouse lines. J Biol Chem, (42): p Lindén, D., Alsterholm, M., Wennbo, H., and Oscarsson, J., PPARalpha deficiency increases secretion and serum levels of apolipoprotein B-containing lipoproteins. Journal of Lipid Research, (11): p Djouadi, F., Weinheimer, C.J., Saffitz, J.E., Pitchford, C., Bastin, J., Gonzalez, F.J., and Kelly, D.P., A gender-related defect in lipid metabolism and glucose homeostasis in peroxisome proliferator- activated receptor alpha- deficient mice. J Clin Invest, (6): p Puigserver, P., Wu, Z., Park, C.W., Graves, R., Wright, M., and Spiegelman, B.M., A cold-inducible coactivator of nuclear receptors linked to adaptive thermogenesis. Cell, (6): p Andersson, U. and Scarpulla, R.C., Pgc-1-related coactivator, a novel, serum-inducible coactivator of nuclear respiratory factor 1-dependent transcription in mammalian cells. Molecular & Cellular Biology, (11): p Lin, J., Puigserver, P., Donovan, J., Tarr, P., and Spiegelman, B.M., Peroxisome proliferator-activated receptor gamma coactivator 1beta (PGC-1beta ), a novel PGC-1-related transcription coactivator associated with host cell factor. Journal of Biological Chemistry, (3): p Lin, J., Handschin, C., and Spiegelman, B.M., Metabolic control through the PGC-1 family of transcription coactivators. Cell Metab, (6): p Yoon, J.C., Puigserver, P., Chen, G., Donovan, J., Wu, Z., Rhee, J., Adelmant, G., Stafford, J., Kahn, C.R., Granner, D.K., Newgard, C.B., and Spiegelman, B.M., Control of hepatic gluconeogenesis through the transcriptional coactivator PGC-1. Nature, (6852): p

56 REFERENCES 82. Lin, J., Tarr, P.T., Yang, R., Rhee, J., Puigserver, P., Newgard, C.B., and Spiegelman, B.M., PGC-1beta in the regulation of hepatic glucose and energy metabolism. Journal of Biological Chemistry, (33): p Meirhaeghe, A., Crowley, V., Lenaghan, C., Lelliott, C., Green, K., Stewart, A., Hart, K., Schinner, S., Sethi, J.K., Yeo, G., Brand, M.D., Cortright, R.N., O'Rahilly, S., Montague, C., and Vidal-Puig, A.J., Characterization of the human, mouse and rat PGC1 beta (peroxisome-proliferator-activated receptor-gamma co-activator 1 beta) gene in vitro and in vivo. Biochem J, (Pt 1): p Koo, S.H., Satoh, H., Herzig, S., Lee, C.H., Hedrick, S., Kulkarni, R., Evans, R.M., Olefsky, J., and Montminy, M., PGC-1 promotes insulin resistance in liver through PPAR-alpha-dependent induction of TRB-3. Nature Medicine, (5): p Lin, J., Yang, R., Tarr, P.T., Wu, P.H., Handschin, C., Li, S., Yang, W., Pei, L., Uldry, M., Tontonoz, P., Newgard, C.B., and Spiegelman, B.M., Hyperlipidemic effects of dietary saturated fats mediated through PGC-1beta coactivation of SREBP. Cell, (2): p Leone, T.C., Lehman, J.J., Finck, B.N., Schaeffer, P.J., Wende, A.R., Boudina, S., Courtois, M., Wozniak, D.F., Sambandam, N., Bernal-Mizrachi, C., Chen, Z., Holloszy, J.O., Medeiros, D.M., Schmidt, R.E., Saffitz, J.E., Abel, E.D., Semenkovich, C.F., and Kelly, D.P., PGC-1alpha deficiency causes multi-system energy metabolic derangements: muscle dysfunction, abnormal weight control and hepatic steatosis. PLoS Biol, (4): p. e Londos, C., Brasaemle, D.L., Schultz, C.J., Segrest, J.P., and Kimmel, A.R., Perilipins, ADRP, and other proteins that associate with intracellular neutral lipid droplets in animal cells. Semin Cell Dev Biol, (1): p Yamaguchi, T., Matsushita, S., Motojima, K., Hirose, F., and Osumi, T., MLDP, a novel PAT family protein localized to lipid droplets and enriched in the heart, is regulated by peroxisome proliferator-activated receptor alpha. J Biol Chem, Jiang, H.P. and Serrero, G., Isolation and characterization of a full-length cdna coding for an adipose differentiation-related protein. Proc Natl Acad Sci U S A, (17): p Brasaemle, D.L., Barber, T., Wolins, N.E., Serrero, G., Blanchette-Mackie, E.J., and Londos, C., Adipose differentiation-related protein is an ubiquitously expressed lipid storage droplet-associated protein. J Lipid Res, (11): p Heid, H.W., Moll, R., Schwetlick, I., Rackwitz, H.R., and Keenan, T.W., Adipophilin is a specific marker of lipid accumulation in diverse cell types and diseases. Cell Tissue Res, (2): p Steiner, S., Wahl, D., Mangold, B.L., Robison, R., Raymackers, J., Meheus, L., Anderson, N.L., and Cordier, A., Induction of the adipose differentiation-related protein in liver of etomoxir-treated rats. Biochem Biophys Res Commun, (3): p Imamura, M., Inoguchi, T., Ikuyama, S., Taniguchi, S., Kobayashi, K., Nakashima, N., and Nawata, H., ADRP stimulates lipid accumulation and lipid droplet formation in murine fibroblasts. Am J Physiol Endocrinol Metab, (4): p. E Larigauderie, G., Furman, C., Jaye, M., Lasselin, C., Copin, C., Fruchart, J.C., Castro, G., and Rouis, M., Adipophilin enhances lipid accumulation and prevents lipid efflux from 56

57 REFERENCES THP-1 macrophages: potential role in atherogenesis. Arterioscler Thromb Vasc Biol, (3): p Gao, J. and Serrero, G., Adipose differentiation related protein (ADRP) expressed in transfected COS-7 cells selectively stimulates long chain fatty acid uptake. J Biol Chem, (24): p Gao, J., Ye, H., and Serrero, G., Stimulation of adipose differentiation related protein (ADRP) expression in adipocyte precursors by long-chain fatty acids. J Cell Physiol, (2): p Vosper, H., Patel, L., Graham, T.L., Khoudoli, G.A., Hill, A., Macphee, C.H., Pinto, I., Smith, S.A., Suckling, K.E., Wolf, C.R., and Palmer, C.N., The peroxisome proliferatoractivated receptor delta promotes lipid accumulation in human macrophages. J Biol Chem, (47): p Hodgkinson, C.P. and Ye, S., Microarray analysis of peroxisome proliferator-activated receptor-gamma induced changes in gene expression in macrophages. Biochem Biophys Res Commun, (3): p Gupta, R.A., Brockman, J.A., Sarraf, P., Willson, T.M., and DuBois, R.N., Target genes of peroxisome proliferator-activated receptor gamma in colorectal cancer cells. J Biol Chem, (32): p Liu, P.C., Huber, R., Stow, M.D., Schlingmann, K.L., Collier, P., Liao, B., Link, J., Burn, T.C., Hollis, G., Young, P.R., and Mukherjee, R., Induction of endogenous genes by peroxisome proliferator activated receptor alpha ligands in a human kidney cell line and in vivo. J Steroid Biochem Mol Biol, (1): p Chawla, A., Lee, C.H., Barak, Y., He, W., Rosenfeld, J., Liao, D., Han, J., Kang, H., and Evans, R.M., PPARdelta is a very low-density lipoprotein sensor in macrophages. Proc Natl Acad Sci U S A, (3): p Targett-Adams, P., McElwee, M.J., Ehrenborg, E., Gustafsson, M.C., Palmer, C.N., and McLauchlan, J., A PPAR response element regulates transcription of the gene for human adipose differentiation-related protein. Biochim Biophys Acta, (1-2): p Tannenbaum, G.S. and Martin, J.B., Evidence for an endogenous ultradian rhythm governing growth hormone secretion in the rat. Endocrinology, (3): p Edén, S., Age- and sex-related differences in episodic growth hormone secretion in the rat. Endocrinology, (2): p Winer, L.M., Shaw, M.A., and Baumann, G., Basal plasma growth hormone levels in man: new evidence for rhythmicity of growth hormone secretion. J Clin Endocrinol Metab, (6): p Jansson, J.O., Eden, S., and Isaksson, O., Sexual dimorphism in the control of growth hormone secretion. Endocr Rev, (2): p Giustina, A. and Veldhuis, J.D., Pathophysiology of the neuroregulation of growth hormone secretion in experimental animals and the human. Endocr Rev, (6): p Baumann, G., Amburn, K., and Shaw, M.A., The circulating growth hormone (GH)- binding protein complex: a major constituent of plasma GH in man. Endocrinology, (3): p

58 REFERENCES 109. Le Roith, D., Bondy, C., Yakar, S., Liu, J.L., and Butler, A., The somatomedin hypothesis: Endocr Rev, (1): p Kopchick, J.J. and Andry, J.M., Growth hormone (GH), GH receptor, and signal transduction. Mol Genet Metab, (1-2): p Zhou, Y.C. and Waxman, D.J., Cross-talk between janus kinase-signal transducer and activator of transcription (JAK-STAT) and peroxisome proliferator-activated receptor- alpha (PPARalpha) signaling pathways. Growth hormone inhibition of pparalpha transcriptional activity mediated by stat5b. J Biol Chem, (5): p Zhou, Y.C. and Waxman, D.J., STAT5b down-regulates peroxisome proliferator-activated receptor alpha transcription by inhibition of ligand-independent activation function region-1 trans-activation domain. J Biol Chem, (42): p Shipley, J.M. and Waxman, D.J., Down-regulation of STAT5b transcriptional activity by ligand-activated peroxisome proliferator-activated receptor (PPAR) alpha and PPARgamma. Mol Pharmacol, (2): p Davidson, M.B., Effect of growth hormone on carbohydrate and lipid metabolism. Endocr Rev, (2): p Colao, A., Ferone, D., Marzullo, P., and Lombardi, G., Systemic complications of acromegaly: epidemiology, pathogenesis, and management. Endocr Rev, (1): p Carroll, P.V., Christ, E.R., Bengtsson, B.A., Carlsson, L., Christiansen, J.S., Clemmons, D., Hintz, R., Ho, K., Laron, Z., Sizonenko, P., Sonksen, P.H., Tanaka, T., and Thorne, M., Growth hormone deficiency in adulthood and the effects of growth hormone replacement: a review. Growth Hormone Research Society Scientific Committee. J Clin Endocrinol Metab, (2): p Veldhuis, J.D., Iranmanesh, A., Ho, K.K., Waters, M.J., Johnson, M.L., and Lizarralde, G., Dual defects in pulsatile growth hormone secretion and clearance subserve the hyposomatotropism of obesity in man. J Clin Endocrinol Metab, (1): p Asplin, C.M., Faria, A.C., Carlsen, E.C., Vaccaro, V.A., Barr, R.E., Iranmanesh, A., Lee, M.M., Veldhuis, J.D., and Evans, W.S., Alterations in the pulsatile mode of growth hormone release in men and women with insulin-dependent diabetes mellitus. J Clin Endocrinol Metab, (2): p Veldhuis, J.D., Liem, A.Y., South, S., Weltman, A., Weltman, J., Clemmons, D.A., Abbott, R., Mulligan, T., Johnson, M.L., Pincus, S., and et al., Differential impact of age, sex steroid hormones, and obesity on basal versus pulsatile growth hormone secretion in men as assessed in an ultrasensitive chemiluminescence assay. J Clin Endocrinol Metab, (11): p Cuneo, R.C., Salomon, F., Watts, G.F., Hesp, R., and Sonksen, P.H., Growth hormone treatment improves serum lipids and lipoproteins in adults with growth hormone deficiency. Metabolism, (12): p Rosén, T., Edén, S., Larson, G., Wilhelmsen, L., and Bengtsson, B.A., Cardiovascular risk factors in adult patients with growth hormone deficiency. Acta Endocrinol (Copenh), (3): p de Boer, H., Blok, G.J., Voerman, H.J., Phillips, M., and Schouten, J.A., Serum lipid levels in growth hormone-deficient men. Metabolism, (2): p

59 REFERENCES 123. Kearney, T., Navas de Gallegos, C., Chrisoulidou, A., Gray, R., Bannister, P., Venkatesan, S., and Johnston, D.G., Hypopituitarsim is associated with triglyceride enrichment of very low-density lipoprotein. J Clin Endocrinol Metab, (8): p Edén, S., Wiklund, O., Oscarsson, J., Rosén, T., and Bengtsson, B.A., Growth hormone treatment of growth hormone-deficient adults results in a marked increase in Lp(a) and HDL cholesterol concentrations. Arterioscler Thromb, (2): p Russell-Jones, D.L., Watts, G.F., Weissberger, A., Naoumova, R., Myers, J., Thompson, G.R., and Sonksen, P.H., The effect of growth hormone replacement on serum lipids, lipoproteins, apolipoproteins and cholesterol precursors in adult growth hormone deficient patients. Clin Endocrinol (Oxf), (3): p Johannsson, G., Oscarsson, J., Rosén, T., Wiklund, O., Olsson, G., Wilhelmsen, L., and Bengtsson, B.A., Effects of 1 year of growth hormone therapy on serum lipoprotein levels in growth hormone-deficient adults. Influence of gender and Apo(a) and ApoE phenotypes. Arterioscler Thromb Vasc Biol, (12): p Christ, E.R., Cummings, M.H., Albany, E., Umpleby, A.M., Lumb, P.J., Wierzbicki, A.S., Naoumova, R.P., Boroujerdi, M.A., Sonksen, P.H., and Russell-Jones, D.L., Effects of growth hormone (GH) replacement therapy on very low density lipoprotein apolipoprotein B100 kinetics in patients with adult GH deficiency: a stable isotope study. J Clin Endocrinol Metab, (1): p Rudling, M., Norstedt, G., Olivecrona, H., Reihner, E., Gustafsson, J.A., and Angelin, B., Importance of growth hormone for the induction of hepatic low density lipoprotein receptors. Proc Natl Acad Sci U S A, (15): p Nikkilä, E.A. and Pelkonen, R., Serum lipids in acromegaly. Metabolism, (7): p Murase, T., Yamada, N., Ohsawa, N., Kosaka, K., Morita, S., and Yoshida, S., Decline of postheparin plasma lipoprotein lipase in acromegalic patients. Metabolism, (7): p Takeda, R., Tatami, R., Ueda, K., Sagara, H., Nakabayashi, H., and Mabuchi, H., The incidence and pathogenesis of hyperlipidaemia in 16 consecutive acromegalic patients. Acta Endocrinol (Copenh), (3): p Tsuchiya, H., Onishi, T., Mogami, H., and Iida, M., Lipid metabolism in acromegalic patients before and after selective pituitary adenomectomy. Endocrinol Jpn, (6): p Oscarsson, J., Wiklund, O., Jakobsson, K.E., Petruson, B., and Bengtsson, B.A., Serum lipoproteins in acromegaly before and 6-15 months after transsphenoidal adenomectomy. Clin Endocrinol (Oxf), (5): p Elam, M.B., Simkevich, C.P., Solomon, S.S., Wilcox, H.G., and Heimberg, M., Stimulation of in vitro triglyceride synthesis in the rat hepatocyte by growth hormone treatment in vivo. Endocrinology, (4): p Elam, M.B., Wilcox, H.G., Solomon, S.S., and Heimberg, M., In vivo growth hormone treatment stimulates secretion of very low density lipoprotein by the isolated perfused rat liver. Endocrinology, (6): p

60 REFERENCES 136. Sjöberg, A., Oscarsson, J., Borén, J., Edén, S., and Olofsson, S.O., Mode of growth hormone administration influences triacylglycerol synthesis and assembly of apolipoprotein B-containing lipoproteins in cultured rat hepatocytes. J Lipid Res, (2): p Frick, F., Lindén, D., Améen, C., Edén, S., Mode, A., and Oscarsson, J., Interaction between growth hormone and insulin in the regulation of lipoprotein metabolism in the rat. Am J Physiol Endocrinol Metab, (5): p. E Améen, C. and Oscarsson, J., Sex difference in hepatic microsomal triglyceride transfer protein expression is determined by the growth hormone secretory pattern in the rat. Endocrinology, (9): p Oscarsson, J., Olofsson, S.O., Bondjers, G., and Edén, S., Differential effects of continuous versus intermittent administration of growth hormone to hypophysectomized female rats on serum lipoproteins and their apoproteins. Endocrinology, (3): p Sjöberg, A., Oscarsson, J., Boström, K., Innerarity, T.L., Edén, S., and Olofsson, S.O., Effects of growth hormone on apolipoprotein-b (apob) messenger ribonucleic acid editing, and apob 48 and apob 100 synthesis and secretion in the rat liver. Endocrinology, (6): p Oscarsson, J., Ottosson, M., Vikman-Adolfsson, K., Frick, F., Enerbäck, S., Lithell, H., and Edén, S., GH but not IGF-I or insulin increases lipoprotein lipase activity in muscle tissues of hypophysectomised rats. J Endocrinol, (2): p Vikman-Adolfsson, K., Oscarsson, J., Nilsson-Ehle, P., and Edén, S., Growth hormone but not gonadal steroids influence lipoprotein lipase and hepatic lipase activity in hypophysectomized rats. J Endocrinol, (2): p Kopchick, J.J., Bellush, L.L., and Coschigano, K.T., Transgenic models of growth hormone action. Annu Rev Nutr, : p Frick, F., Bohlooly, M., Lindén, D., B., O., Törnell, J., Edén, S., and Oscarsson, J., Long-term growth hormone excess induces marked alterations in lipoprotein metabolism in mice. Am J Physiol Endocrinol Metab, : p. E Olsson, B., Bohlooly, Y.M., Fitzgerald, S.M., Frick, F., Ljungberg, A., Ahrén, B., Törnell, J., Bergström, G., and Oscarsson, J., Bovine growth hormone transgenic mice are resistant to diet-induced obesity but develop hyperphagia, dyslipidemia, and diabetes on a high-fat diet. Endocrinology, (2): p Bohlooly, Y.M., Carlson, L., Olsson, B., Gustafsson, H., Andersson, I.J., Törnell, J., and Bergström, G., Vascular function and blood pressure in GH transgenic mice. Endocrinology, (8): p Andersson, I.J., Johansson, M.E., Wickman, A., Bohlooly, Y.M., Klintland, N., Caidahl, K., Gustafsson, M., Borén, J., Gan, L.M., and Bergström, G., Endothelial dysfunction in growth hormone transgenic mice. Clin Sci (Lond), (2): p Grobbee, D.E. and Bots, M.L., Carotid artery intima-media thickness as an indicator of generalized atherosclerosis. J Intern Med, (5): p Markussis, V., Beshyah, S.A., Fisher, C., Sharp, P., Nicolaides, A.N., and Johnston, D.G., Detection of premature atherosclerosis by high-resolution ultrasonography in symptom-free hypopituitary adults. Lancet, (8829): p

61 REFERENCES 150. Borson-Chazot, F., Serusclat, A., Kalfallah, Y., Ducottet, X., Sassolas, G., Bernard, S., Labrousse, F., Pastene, J., Sassolas, A., Roux, Y., and Berthezene, F., Decrease in carotid intima-media thickness after one year growth hormone (GH) treatment in adults with GH deficiency. J Clin Endocrinol Metab, (4): p Pfeifer, M., Verhovec, R., Zizek, B., Prezelj, J., Poredos, P., and Clayton, R.N., Growth hormone (GH) treatment reverses early atherosclerotic changes in GH-deficient adults. J Clin Endocrinol Metab, (2): p Colao, A., Spiezia, S., Cerbone, G., Pivonello, R., Marzullo, P., Ferone, D., Di Somma, C., Assanti, A.P., and Lombardi, G., Increased arterial intima-media thickness by B-M mode echodoppler ultrasonography in acromegaly. Clin Endocrinol (Oxf), (4): p Colao, A., Marzullo, P., and Lombardi, G., Effect of a six-month treatment with lanreotide on cardiovascular risk factors and arterial intima-media thickness in patients with acromegaly. Eur J Endocrinol, (3): p Colao, A., Spinelli, L., Cuocolo, A., Spiezia, S., Pivonello, R., di Somma, C., Bonaduce, D., Salvatore, M., and Lombardi, G., Cardiovascular consequences of earlyonset growth hormone excess. J Clin Endocrinol Metab, (7): p Otsuki, M., Kasayama, S., Yamamoto, H., Saito, H., Sumitani, S., Kouhara, H., Saitoh, Y., Ohnishi, T., and Arita, N., Characterization of premature atherosclerosis of carotid arteries in acromegalic patients. Clin Endocrinol (Oxf), (6): p Brevetti, G., Marzullo, P., Silvestro, A., Pivonello, R., Oliva, G., di Somma, C., Lombardi, G., and Colao, A., Early vascular alterations in acromegaly. J Clin Endocrinol Metab, (7): p Palmiter, R.D., Brinster, R.L., Hammer, R.E., Trumbauer, M.E., Rosenfeld, M.G., Birnberg, N.C., and Evans, R.M., Dramatic growth of mice that develop from eggs microinjected with metallothionein-growth hormone fusion genes. Nature, (5893): p Plump, A.S., Smith, J.D., Hayek, T., Aalto-Setala, K., Walsh, A., Verstuyft, J.G., Rubin, E.M., and Breslow, J.L., Severe hypercholesterolemia and atherosclerosis in apolipoprotein E-deficient mice created by homologous recombination in ES cells. Cell, (2): p Zhang, S.H., Reddick, R.L., Piedrahita, J.A., and Maeda, N., Spontaneous hypercholesterolemia and arterial lesions in mice lacking apolipoprotein E. Science, (5081): p Ishibashi, S., Brown, M.S., Goldstein, J.L., Gerard, R.D., Hammer, R.E., and Herz, J., Hypercholesterolemia in low density lipoprotein receptor knockout mice and its reversal by adenovirus-mediated gene delivery. J Clin Invest, (2): p Breslow, J.L., Mouse models of atherosclerosis. Science, (5262): p Sandstedt, J., Ohlsson, C., Norjavaara, E., Nilsson, J., and Törnell, J., Disproportional bone growth and reduced weight gain in gonadectomized male bovine growth hormone transgenic and normal mice. Endocrinology, (6): p Lee, S.S., Pineau, T., Drago, J., Lee, E.J., Owens, J.W., Kroetz, D.L., Fernandez- Salguero, P.M., Westphal, H., and Gonzalez, F.J., Targeted disruption of the alpha isoform of the peroxisome proliferator- activated receptor gene in mice results in abolishment of the pleiotropic effects of peroxisome proliferators. Mol Cell Biol, (6): p

62 REFERENCES 164. Groesbeck, M.D. and Parlow, A.F., Highly improved precision of the hypophysectomized female rat body weight gain bioassay for growth hormone by increased frequency of injections, avoidance of antibody formation, and other simple modifications. Endocrinology, (6): p Willson, T.M., Brown, P.J., Sternbach, D.D., and Henke, B.R., The PPARs: from orphan receptors to drug discovery. J Med Chem, (4): p Sawada, N., Tomomura, A., Sattler, C.A., Sattler, G.L., Kleinman, H.K., and Pitot, H.C., Effects of extracellular matrix components on the growth and differentiation of cultured rat hepatocytes. In Vitro Cell Dev Biol, (4): p Bissell, D.M., Arenson, D.M., Maher, J.J., and Roll, F.J., Support of cultured hepatocytes by a laminin-rich gel. Evidence for a functionally significant subendothelial matrix in normal rat liver. J Clin Invest, (3): p Runge, D., Runge, D.M., Bowen, W.C., Locker, J., and Michalopoulos, G.K., Matrix induced re-differentiation of cultured rat hepatocytes and changes of CCAAT/enhancer binding proteins. Biol Chem, (8): p Michalopoulos, G. and Pitot, H.C., Primary culture of parenchymal liver cells on collagen membranes. Morphological and biochemical observations. Exp Cell Res, (1): p Alba, R., Bosch, A., and Chillon, M., Gutless adenovirus: last-generation adenovirus for gene therapy. Gene Ther, Suppl 1: p. S Fechner, H., Haack, A., Wang, H., Wang, X., Eizema, K., Pauschinger, M., Schoemaker, R., Veghel, R., Houtsmuller, A., Schultheiss, H.P., Lamers, J., and Poller, W., Expression of coxsackie adenovirus receptor and alphav-integrin does not correlate with adenovector targeting in vivo indicating anatomical vector barriers. Gene Ther, (9): p Russell, H.K., Jr., A modification of Movat's pentachrome stain. Arch Pathol, (2): p van Vlijmen, B.J., van den Maagdenberg, A.M., Gijbels, M.J., van der Boom, H., HogenEsch, H., Frants, R.R., Hofker, M.H., and Havekes, L.M., Diet-induced hyperlipoproteinemia and atherosclerosis in apolipoprotein E3-Leiden transgenic mice. J Clin Invest, (4): p Bengtsson, B.A., Edén, S., Ernest, I., Oden, A., and Sjögren, B., Epidemiology and longterm survival in acromegaly. A study of 166 cases diagnosed between 1955 and Acta Med Scand, (4): p Kanaley, J.A., Dall, R., Moller, N., Nielsen, S.C., Christiansen, J.S., Jensen, M.D., and Jorgensen, J.O., Acute exposure to GH during exercise stimulates the turnover of free fatty acids in GH-deficient men. J Appl Physiol, (2): p Yamada, J., Sugiyama, H., Watanabe, T., and Suga, T., Suppressive effect of growth hormone on the expression of peroxisome proliferator-activated receptor in cultured rat hepatocytes. Res Commun Mol Pathol Pharmacol, (1): p Carlsson, L., Lindén, D., Jalouli, M., and Oscarsson, J., Effects of fatty acids and growth hormone on liver fatty acid binding protein and PPARalpha in rat liver. Am J Physiol Endocrinol Metab, (4): p. E

63 REFERENCES 178. Jalouli, M., Carlsson, L., Améen, C., Lindén, D., Ljungberg, A., Michalik, L., Edén, S., Wahli, W., and Oscarsson, J., Sex difference in hepatic peroxisome proliferator-activated receptor alpha expression: influence of pituitary and gonadal hormones. Endocrinology, (1): p Lamb, R.G., Koch, J.C., and Bush, S.R., An enzymatic explanation of the differential effects of oleate and gemfibrozil on cultured hepatocyte triacylglycerol and phosphatidylcholine biosynthesis and secretion. Biochim Biophys Acta, (3): p Vega, R.B., Huss, J.M., and Kelly, D.P., The Coactivator PGC-1 Cooperates with Peroxisome Proliferator-Activated Receptor alpha in Transcriptional Control of Nuclear Genes Encoding Mitochondrial Fatty Acid Oxidation Enzymes. Mol. Cell. Biol., (5): p Paigen, B., Morrow, A., Brandon, C., Mitchell, D., and Holmes, P., Variation in susceptibility to atherosclerosis among inbred strains of mice. Atherosclerosis, (1): p Nakashima, Y., Plump, A.S., Raines, E.W., Breslow, J.L., and Ross, R., ApoE-deficient mice develop lesions of all phases of atherosclerosis throughout the arterial tree. Arterioscler Thromb, (1): p Reddick, R.L., Zhang, S.H., and Maeda, N., Atherosclerosis in mice lacking apo E. Evaluation of lesional development and progression. Arterioscler Thromb, (1): p Andersson, I.J., Barlind, A., Nyström, H.C., Olsson, B., Skott, O., Mobini, R., Johansson, M., and Bergström, G., Reduced sympathetic responsiveness as well as plasma and tissue noradrenaline concentration in growth hormone transgenic mice. Acta Physiol Scand, (4): p Ross, R., Atherosclerosis--an inflammatory disease. N Engl J Med, (2): p Chait, A., Han, C.Y., Oram, J.F., and Heinecke, J.W., Thematic review series: The immune system and atherogenesis. Lipoprotein-associated inflammatory proteins: markers or mediators of cardiovascular disease? J Lipid Res, (3): p Lewis, K.E., Kirk, E.A., McDonald, T.O., Wang, S., Wight, T.N., O'Brien, K.D., and Chait, A., Increase in serum amyloid a evoked by dietary cholesterol is associated with increased atherosclerosis in mice. Circulation, (5): p Paul, A., Ko, K.W., Li, L., Yechoor, V., McCrory, M.A., Szalai, A.J., and Chan, L., C- reactive protein accelerates the progression of atherosclerosis in apolipoprotein E-deficient mice. Circulation, (5): p Delafontaine, P., Song, Y.H., and Li, Y., Expression, regulation, and function of IGF-1, IGF-1R, and IGF-1 binding proteins in blood vessels. Arterioscler Thromb Vasc Biol, (3): p Hochberg, Z., Hertz, P., Maor, G., Oiknine, J., and Aviram, M., Growth hormone and insulin-like growth factor-i increase macrophage uptake and degradation of low density lipoprotein. Endocrinology, (1): p Kaplan, R.C., Strickler, H.D., Rohan, T.E., Muzumdar, R., and Brown, D.L., Insulinlike growth factors and coronary heart disease. Cardiol Rev, (1): p

64 REFERENCES 192. Hietaniemi, M., Poykko, S.M., Ukkola, O., Paivansalo, M., and Antero Kesaniemi, Y., IGF-I concentrations are positively associated with carotid artery atherosclerosis in women. Ann Med, (5): p Kawachi, S., Takeda, N., Sasaki, A., Kokubo, Y., Takami, K., Sarui, H., Hayashi, M., Yamakita, N., and Yasuda, K., Circulating insulin-like growth factor-1 and insulin-like growth factor binding protein-3 are associated with early carotid atherosclerosis. Arterioscler Thromb Vasc Biol, (3): p Mensenkamp, A.R., Jong, M.C., van Goor, H., van Luyn, M.J., Bloks, V., Havinga, R., Voshol, P.J., Hofker, M.H., van Dijk, K.W., Havekes, L.M., and Kuipers, F., Apolipoprotein E participates in the regulation of very low density lipoprotein-triglyceride secretion by the liver. J Biol Chem, (50): p Sjöberg, A., Oscarsson, J., Edén, S., and Olofsson, S.O., Continuous but not intermittent administration of growth hormone to hypophysectomized rats increases apolipoprotein-e secretion from cultured hepatocytes. Endocrinology, (2): p Schroeder, F., Jefferson, J.R., Powell, D., Incerpi, S., Woodford, J.K., Colles, S.M., Myers-Payne, S., Emge, T., Hubbell, T., Moncecchi, D., and et al., Expression of rat L- FABP in mouse fibroblasts: role in fat absorption. Mol Cell Biochem, (1-2): p Carlsson, L., Nilsson, I., and Oscarsson, J., Hormonal regulation of liver fatty acidbinding protein in vivo and in vitro: effects of growth hormone and insulin. Endocrinology, (6): p Sugiyama, H., Yamada, J., and Suga, T., Effects of testosterone, hypophysectomy and growth hormone treatment on clofibrate induction of peroxisomal beta-oxidation in female rat liver. Biochem Pharmacol, (5): p Sato, T., Murayama, N., Yamazoe, Y., and Kato, R., Suppression of clofibrate-induction of peroxisomal and microsomal fatty acid-oxidizing enzymes by growth hormone and thyroid hormone in primary cultures of rat hepatocytes. Biochim Biophys Acta, (3): p Yamada, J., Sugiyama, H., Tamura, H., and Suga, T., Hormonal modulation of peroxisomal enzyme induction caused by peroxisome proliferators: suppression by growth and thyroid hormones in cultured rat hepatocytes. Arch Biochem Biophys, (2): p Ichikawa, T., Hamasaki, K., Ishikawa, H., Ejima, E., Eguchi, K., and Nakao, K., Nonalcoholic steatohepatitis and hepatic steatosis in patients with adult onset growth hormone deficiency. Gut, (6): p Matsusue, K., Haluzik, M., Lambert, G., Yim, S.H., Gavrilova, O., Ward, J.M., Brewer, B., Jr., Reitman, M.L., and Gonzalez, F.J., Liver-specific disruption of PPARgamma in leptin-deficient mice improves fatty liver but aggravates diabetic phenotypes. J Clin Invest, (5): p Gavrilova, O., Haluzik, M., Matsusue, K., Cutson, J.J., Johnson, L., Dietz, K.R., Nicol, C.J., Vinson, C., Gonzalez, F.J., and Reitman, M.L., Liver peroxisome proliferatoractivated receptor gamma contributes to hepatic steatosis, triglyceride clearance, and regulation of body fat mass. J Biol Chem, (36): p Améen, C., Edvardsson, U., Ljungberg, A., Asp, L., Åkerblad, P., Tuneld, A., Olofsson, S.O., Lindén, D., and Oscarsson, J., Activation of peroxisome proliferator- 64

65 REFERENCES activated receptor alpha increases the expression and activity of microsomal triglyceride transfer protein in the liver. J Biol Chem, (2): p Waterman, I.J. and Zammit, V.A., Differential effects of fenofibrate or simvastatin treatment of rats on hepatic microsomal overt and latent diacylglycerol acyltransferase activities. Diabetes, (6): p Xu, G., Sztalryd, C., Lu, X., Tansey, J.T., Gan, J., Dorward, H., Kimmel, A.R., and Londos, C., Post-translational regulation of adipose differentiation-related protein by the ubiquitin/proteasome pathway. J Biol Chem, (52): p Motomura, W., Inoue, M., Ohtake, T., Takahashi, N., Nagamine, M., Tanno, S., Kohgo, Y., and Okumura, T., Up-regulation of ADRP in fatty liver in human and liver steatosis in mice fed with high fat diet. Biochem Biophys Res Commun, (4): p

66

67 I

68

69 Atherosclerosis xxx (2005) xxx xxx Increased atherosclerotic lesion area in apoe deficient mice overexpressing bovine growth hormone Irene J. Andersson a,1, Anna Ljungberg a,b,1, Lennart Svensson c, Li-Ming Gan a,c,d, Jan Oscarsson a,b,c,göran Bergström d, a Department of Physiology, Göteborg University, Göteborg, Sweden b Wallenberg Laboratory for Cardiovascular Research, Göteborg University, Göteborg, Sweden c AstraZeneca R&D, Mölndal, Sweden d Department of Clinical Physiology, Sahlgrenska Academy, Göteborg University, Göteborg, Sweden Received 3 July 2005; received in revised form 2 November 2005; accepted 7 November 2005 Abstract Human growth hormone (GH) excess is linked to increased cardiovascular morbidity and mortality. However, little is known about the effect of GH excess on atherosclerosis. We developed a new mouse model to assess the hypothesis that GH overexpression accelerates atherosclerotic lesion formation. apoe / mice were crossed with bovine GH (bgh) transgenic mice to yield apoe / mice overexpressing bgh (apoe / /bgh). The mice were fed either standard or Western diet. At 22 weeks, atherosclerotic lesion area of thoracic aorta was larger in apoe / /bgh mice compared with littermate apoe / mice fed either diet (standard: +161 ± 50%, Western: +430 ± 134%). Aortic sinus lesions were more severe in apoe / /bgh mice fed standard diet compared with littermate apoe / mice. apoe / /bgh mice had lower (VLDL + LDL)/HDL ratios compared with littermate apoe / mice, while systolic blood pressure was higher in apoe / /bgh mice, irrespective of diet. The levels of serum amyloid A and hepatic CRP mrna were higher in apoe / /bgh mice than in littermate apoe / mice. In conclusion, this study shows that excess GH augments the development of atherosclerosis in apoe / mice. The mechanisms could be direct effects of GH on cellular processes in the vessel wall or the result of concomitant processes such as hypertension or a general inflammatory state Elsevier Ireland Ltd. All rights reserved. Keywords: Apolipoprotein E; Atherosclerosis; Growth hormone; Diet; Atherogenic 1. Introduction Growth hormone (GH) and insulin-like growth factor-i (IGF-I) exert important effects on the heart and vasculature [1]. Overexpression of GH (acromegaly) in humans is associated with increased cardiovascular morbidity and mortality [2]. The high cardiovascular risk in acromegalic patients has generally been attributed to cardiomyopathy and Corresponding author at: Department of Clinical Physiology, Cardiovascular Institute, Göteborg University, Box 432, SE Göteborg, Sweden. Tel.: ; fax: address: g.bergstrom@fysiologi.gu.se (G. Bergström). 1 These authors contributed equally to this work. the increased incidence of associated disorders such as hypertension and dyslipidemia [3]; risk factors that are associated with cardiovascular disease [4]. However, there are currently few data on vascular pathology in acromegaly, and further, the reports regarding early markers of atherosclerosis such as intima media thickness (IMT) and endothelial dysfunction, are not in agreement [3,5]. The aim of this study was to determine to what extent high concentrations of GH/IGF-I affects atherosclerotic lesion formation and the risk factors behind development of atherosclerosis. During the past years, we have used the transgenic bovine GH (bgh) overexpressing mouse as an animal model for cardiovascular disease associated with high circulating levels of GH/IGF-1. Similar to acromegalic patients, these /$ see front matter 2005 Elsevier Ireland Ltd. All rights reserved. doi: /j.atherosclerosis ATH-9287; No. of Pages 10

70 2 I.J. Andersson et al. / Atherosclerosis xxx (2005) xxx xxx mice develop several risk factors for cardiovascular disease, such as dyslipidemia, hypertension and insulin resistance [6 8]. Furthermore, we have shown that bgh mice have impaired endothelial function of conduit arteries, possibly due to oxidative stress [9]. To study the effect of GH on atherosclerosis development, we crossed bgh transgenic mice with apolipoprotein E deficient (apoe / ) mice, which spontaneously develop hypercholesterolemia and atherosclerotic lesions similar to those observed in humans [10]. In this new mouse model (apoe / /bgh), we measured the aortic lesion area as well as several potential risk factors for atherosclerosis. Our main findings were that apoe / /bgh mice had larger atherosclerotic lesion area, hypertension and signs of a general inflammation, but less severe dyslipidemia than littermate apoe / mice. 2. Materials and methods 2.1. Animals and diets bgh transgenic mice on a C57BL/6JxCBA background were originally generated by microinjection of a gene construct containing a metallothionein promoter linked to a sequence encoding bgh [11]. In this strain of bgh transgenic mice, which previously has been used in several studies at our laboratory [7,9,12], bgh is constitutively and ectopically expressed in a head-to-tail fashion. bgh transgenic mice were crossed with apoe / mice on a Sv129xC57BL/6 background (Taconic, M&B Breeding and Research Centre, Ry, Denmark). apoe heterozygous mice carrying one bgh allele were further bred with apoe / mice. Tail biopsies were taken from offspring and genotyped to identify all homozygous apoe / mice. Mice from this pool were further bred, using one mouse with and one without the bgh allele. This generated apoe / /bgh +/ mice, from now on denoted as apoe / /bgh, and apoe / littermate controls, which were used in the experiments. In the first experiment, male apoe / /bgh and control mice were maintained on standard diet (Harlan Teklad Global Diet 2016, Harlan Blackthorn, England) and water ad libitum for weeks. In addition, a group of female apoe / /bgh and littermate controls were maintained on standard diet and used for quantification of lesion area. In a second experiment, male mice were maintained on the standard diet until 13 weeks of age, and were then fed a Western diet containing 40% energy from triglycerides and 0.15% (w/w) cholesterol (R638, Lactamin, Kimstad, Sweden) ad libitum for 8 weeks Tail cuff blood pressure Systolic blood pressure (SBP) was measured at 20 weeks of age using a computerized non-invasive tail-cuff system (RTBP Monitor; Harvard Apparatus Inc., South Natick, MA). Conscious animals were kept in a restrainer, with a standard acclimatization time of 10 min and gentle heating of the tail before every recording session. Recordings were performed during 2 3 consecutive days, collecting 6 satisfactory measurements each day Tissue preparation At weeks of age, all mice were anaesthetized with pentobarbital sodium (Apoteks-bolaget, Sweden, ml/g body weight intraperitoneally). The chest and abdomen were opened, blood was drawn from the right chamber of the heart, kept on ice and centrifuged at g for 5 min. From mice fed Western diet, a piece of the liver was removed, snapfrozen in liquid nitrogen and stored at 80 C for subsequent analysis. The heart and aorta were perfused at 100 mmhg with 0.9% NaCl followed by 4% paraformaldehyde for 5 min and then dissected. The aorta was separated from the heart 1.3 mm proximal to the brachiocephalic artery. At the distal end the aorta was cut transversely above the right renal artery and fixed in 4% paraformaldehyde until processing Quantification of aorta lesion area The fixed aorta was dissected and excised free of connective and adipose tissue. It was cut open longitudinally and pinned onto a silicone-coated dish. The aortas were rinsed with 70% ethanol for 5 min, stained with Sudan IV (Sigma Aldrich, 5 g/l in 1:1 70% ethanol and 100% acetone) for 6 min, rinsed with 80% ethanol for 3 min and then kept in 0.9% NaCl during quantification. Images were collected with Canon Utilities Remote Capture 2.2 using a digital camera connected to the dissection microscope. Adobe Photoshop 7.0 was used to manually trace the outline of the intimal surface and lesions to subsequently calculate the lesion area as percentage of the intimal area. Two different regions were examined; the arch and the thoracic portion, the point of separation being immediately distal to the left subclavian artery Quantification of aortic sinus cross-sectional lesion area The hearts were embedded in paraffin and 5 m sections were taken from the aortic sinus, at the first appearance of all three aortic valves [13]. Russell Movat pentachrome staining [14] with some modifications were used for quantification of atherosclerotic lesions. Before staining, the slides were deparaffinized and rehydrated by repeated washing in xylene and alcohol with decreasing concentrations (99.5%, 95% and 70%). The slides were placed in Verhoeffs elastic stain working solution for 20 min, washed in running tap water for 20 min and in distilled water for 1 min. The preparations were differentiated in 2% ferric chloride solution until elastic fibres were sharply defined and the background was grey (about 5 min). Then the reaction was terminated by washing in distilled water. The slides were placed in 5% sodium

71 I.J. Andersson et al. / Atherosclerosis xxx (2005) xxx xxx 3 thiosulfate solution for 1 min, washed in tap water for 5 min and placed in 3% glacial acetic acid solution for 3 min. In the next step, the slides were put in 2% alcian blue solution for 20 min, washed in running tap water for 10 min and in distilled water for 1 min. Then the preparations were put in crocein scarlet-acid fuchsin solution for 2 min, washed several times in distilled water and quickly in 1% acetic acid solution. Next, the slides were placed in 5% phosphotungstic acid solution two times 5 min and washed in 1% acetic acid solution. After dehydration in three changes of 99.5% alcohol, the slides were put in alcoholic saffron solution for 5 min, washed in absolute alcohol, placed in xylene for a few minutes and finally mounted in Mountex (Histolab Products AB, Göteborg, Sweden). Images were collected using a computerized system. Micro Image 4.0 for Windows (Olympus Optical Co., Hamburg, Germany) was used to manually outline the internal elastica lamina (IEL) circumference, and thereby calculate the vessel area to avoid errors due to folding artifacts. The lesions were outlined from the internal elastica lamina to the luminal edge. The relative lesion area was calculated by dividing the lesion area by the total cross-sectional area. The lesions were classified into five different categories according to van Vlijmen et al. [15] to examine if there were any differences in lesion severity Serum analyses Serum triglycerides and cholesterol were determined with enzymatic colorimetric assays (Roche, Mannheim, Germany). The size distribution profiles of serum lipoproteins were measured using a high performance liquid chromatography system and Superose 6 PC 3.2/30 column (Amersham Pharmacia Biotech, Uppsala, Sweden) as described before [16]. Serum apolipoprotein B (apob) concentrations were determined with an electroimmunoassay as previously described [16,17]. Serum amyloid A (S-amyloid A) was measured by a solid phase sandwich ELISA, using a commercial reagent kit (Cat. no.: TP802-M, Tridelta Development Ltd., Wicklow, Ireland). After acid ethanol extraction, serum IGF- I was measured with a mouse/rat IGF-I RIA (DSL-2900, DSL, Diagnostic Systems Laboratories Inc., Webster, TX, USA) Hepatic CRP mrna expression Total liver RNA was isolated from frozen liver with TriReagent TM (Sigma) according to the manufacturer s protocol and RNA concentration was determined spectrophotometrically at 260 nm. DNA-free TM (Ambion, Austin, TX, USA) was used to remove contaminating DNA from the RNA preparations. First strand cdna was synthesized from 0.4 g of total RNA with TaqMan Reverse Transcription Reagents (Applied Biosystems, Foster City, CA, USA). Quantitative real-time PCR was performed with the ABI Prism 7700 Sequence Detection system, using the SYBR Green labeling system (Applied Biosystems). All samples were analysed in triplicate and the expression data were normalized to the endogenous control acidic ribosomal phosphoprotein P0 (36B4). The relative expression levels were calculated according to the formula 2 Ct, where C t is the difference in threshold cycle (C t ) values between the target and the endogenous control. Specific primers for C reactive protein (CRP) and 36B4 were designed with Primer Express TM software (Applied Biosystems). The forward and reverse primers for CRP were 5 -AAG CTA CTC TGG TGC CTT CTG ATC-3 and 5 -GAA GTA TCT GAC TCC TTG GGA AAT ACA-3 and the forward and reverse primers for 36B4 were 5 -GAG GAA TCA GAT GAG GAT ATG GGA-3 and 5 -AAG CAG GCT GAC TTG GTT GC-3. To avoid amplification of genomic DNA, primers were positioned to span exon junctions. The correct sizes of the amplicons were verified by gel electrophoresis Statistics Data are presented as mean ± S.E.M. Equality of means between two groups was tested using Student s t-test for all variables except lesion quantifications and classifications, which were tested by Mann Whitney U-test and Fisher s exact test, respectively. The significance level is P < Results 3.1. Body weights and IGF-I levels The apoe / /bgh groups had higher body weights and serum IGF-I concentrations compared to controls (Table 1). Epididymal fat mass was not different between apoe / /bgh mice and littermate apoe / mice fed Western diet (832 ± 52 mg compared with 1038 ± 121 mg), but the proportion of epididymal fat relative to body weight was lower in apoe / /bgh mice compared with littermate apoe / (1.8 ± 0.08% compared with 3.0 ± 0.3%, P < 0.05). Table 1 Body weights and IGF-I levels in male apoe / /bgh and apoe / littermate mice fed standard diet or Western diet Standard diet apoe / (n = 11) apoe / /bgh (n =8) Body weight at 12 weeks (g) 30 ± ± 1.0 *** Final weight (g) 34 ± ± 1.4 *** IGF-I (nmol/l) 39 ± ± 3.9 *** Western diet apoe / (n = 8) apoe / /bgh (n =6) Body weight at 14 weeks (g) 29 ± ± 0.5 *** Final weight (g) 34 ± ± 1.1 *** IGF-I (nmol/l) 55 ± ± 11 ** Values are means ± S.E.M. P < 0.01, Student s t-test. P < 0.001, Student s t-test.

72 4 I.J. Andersson et al. / Atherosclerosis xxx (2005) xxx xxx 3.4. Aortic sinus cross-sectional lesion area Lesion area in the aortic sinus was larger in apoe / /bgh mice fed standard diet, compared to littermate apoe / mice (Fig. 3A). However, when the lesion area was normalized to the calculated total cross-sectional area of the aortic sinus, this difference was abolished (Fig. 3B). When fed a Western diet, the same pattern of response as in the mice fed standard diet was observed (Fig. 3C and D). A gross morphological examination revealed the presence of lesions in different developmental stages, from fatty streaks with welldefined foam cells to more advanced necrotic lesions. The histological techniques used did not allow for exact quantification of plaque composition. However, the lesions were classified into different categories according to van Vlijmen et al. [15] as shown in Table 2. In order to do statistical analysis, the changes belonging to categories 1 3 (fatty streaks and mild plaques) were grouped into mild alterations, whereas plaques belonging to categories 4 and 5 (moderate to severe plaques) were grouped into severe alterations. When fed the standard diet, 7 out of 8 apoe / /bgh mice had moderate or severe plaques (categories 4 and 5) whereas 4 out of 11 apoe / mice were grouped into these categories (P = 0.04, one-tailed, Fisher s exact test). When fed the Western diet, the lesions were generally more advanced and all mice had moderate or severe plaques with no difference between the groups (Table 2). Fig. 4A D shows representative photographs of Fig. 1. Systolic blood pressure (SBP) in male apoe / /bgh and apoe / littermate mice fed standard diet (A, n = 10 apoe / /bgh, n = 14 apoe / ) or Western diet (B, n = 6 apoe / /bgh, n = 8 apoe / ). Systolic blood pressure was measured using a non-invasive tail-cuff system. Values are means ± S.E.M. * P < 0.05, ** P < 0.01, Student s t-test Systolic blood pressure As previously shown in bgh transgenic mice [18], SBP was significantly higher in 20-week old apoe / /bgh mice compared to littermate apoe / mice, irrespective of diet (Fig. 1A and B) Aorta lesion area Lesion area in the thoracic aorta and the aortic arch was determined after Sudan IV staining. Lesion area in the thoracic aorta was larger in male apoe / /bgh mice compared to littermate apoe / mice fed either diet (Fig. 2A and C), whereas there was no difference in lesion area in the aortic arch (Fig. 2B and D) nor when the thoracic and aortic areas were combined (data not shown). Neither of the aortic lesion areas differed between female apoe / /bgh mice and littermate apoe / mice fed standard diet (thoracic aorta: 0.86 ± 0.23% versus 0.53 ± 0.09%). Table 2 Classification of lesions in aortic sinus of male apoe / /bgh and apoe / littermate mice fed standard diet or Western diet Standard diet apoe / (n = 11) apoe / /bgh (n =8) Category 1 Mild alterations Category 2 2 Mild alterations Category Mild alterations Category Severe alterations Category 5 4 Severe alterations Western diet apoe / (n =9) apoe / /bgh (n =6) Category 1 Mild alterations Category 2 Mild alterations Category 3 Mild alterations Category Severe alterations Category Severe alterations Lesions in the aortic sinus were classified into 5 different categories according to van Vlijmen et al. [15]. (1) Early fatty streak: per section up to 10 foam cells present in the intima, (2) regular fatty streak: more than 10 foam cells present in the intima, (3) mild plaque: extension of foam cells into the media and mild fibrosis of the media without loss of architecture, (4) moderate plaque: foam cells in the media, fibrosis, cholesterol clefts, mineralization and/or necrosis of the media, and (5) severe plaque: as category 4 but more extensive and deeper into the media. When grouping categories 1 3 and categories 4 and 5 into mild and severe alterations, respectively, statistical analysis showed significant differences between the genotypes fed standard diet (p = 0.04, one-tailed Fisher s exact test) but not in those fed Western diet.

73 I.J. Andersson et al. / Atherosclerosis xxx (2005) xxx xxx 5 Fig. 2. Lesion area in the thoracic aorta (A, C) and the arch (B, D) of male apoe / /bgh and apoe / littermate mice fed standard diet (A, B) or Western diet (C, D). Lesion areas were quantified after staining with Sudan IV and are expressed as percentage of intimal area. ** P < 0.01, Mann Whitney U-test. sections from aortic sinus in apoe / /bgh and littermate apoe / mice fed standard diet Serum lipid and lipoprotein levels Serum cholesterol and triglycerides were measured to evaluate if a more marked dyslipidemia in the apoe / /bgh mice than in littermate apoe / mice could contribute to the accelerated atherosclerosis. On a standard diet, apoe / /bgh mice had lower serum levels of total cholesterol due to lower levels of cholesterol in the VLDL + LDL fractions (Fig. 5A). Moreover, HDL cholesterol levels were higher in the apoe / /bgh mice than in the littermate apoe / mice. When fed Western diet, both groups had generally higher serum levels of cholesterol (Fig. 5B). No differences in total or VLDL + LDL cholesterol levels between the groups were observed, while HDL cholesterol levels were higher in apoe / /bgh mice than in littermate apoe / mice. On both diets, apoe / /bgh mice had lower (VLDL + LDL)/HDL ratios than littermate apoe / mice (standard diet: 8.5 ± 0.3 versus 28.6 ± 1.9, P < 0.001, Western diet: 7.9 ± 1.5 versus 20.5 ± 2.2, P < 0.001). Serum triglyceride concentrations were significantly lower in apoe / /bgh mice compared to littermate apoe / mice (Fig. 6A and C), whereas apob levels were not different between the genotypes (Fig. 6B and D). However, apob levels were markedly higher in the mice given Western diet Inflammatory markers Inflammatory markers were measured to assess the contribution of inflammation to the accelerated atherosclerosis. S-amyloid A was increased in apoe / /bgh mice fed standard diet compared to littermate apoe / mice (Fig. 7A). However, S-amyloid A levels in apoe / /bgh mice fed Western diet did not differ from littermate apoe / mice (Fig. 7B). To exclude the possibility that S-amyloid A concentrations only reflected a specific effect of GH overexpression on S-amyloid A production and not a general inflammatory response, expression of CRP mrna was also determined in the livers of mice fed a Western diet. The hepatic CRP mrna expression in apoe / /bgh mice were 140 ± 37% higher than in littermate apoe / mice (Fig. 7C).

74 6 I.J. Andersson et al. / Atherosclerosis xxx (2005) xxx xxx Fig. 3. Aortic sinus lesion area expressed as mm 2 (A, C) and percentage (B, D) of total vessel cross-sectional area in male apoe / /bgh and apoe / littermate mice fed standard diet (A, B) or Western diet (C, D). Cross-sections from the aortic sinus were stained with the modified Russel Movat pentachrome staining followed by quantification of lesion area. *** P < 0.001, Mann Whitney U-test. 4. Discussion This study on cross-bred apoe / and bgh transgenic mice showed that atherosclerotic lesion area in the thoracic aorta was larger in male apoe / /bgh mice compared to littermate apoe / mice, both on a low fat diet and a Western diet. The larger lesion area in apoe / /bgh mice was accompanied by higher blood pressure and an inflammatory response. However, the markedly disturbed serum lipoprotein pattern in apoe / mice that is characterized by accumulation of remnants particles was not worsened by GH overproduction. Thus, in this first model of acromegalic atherosclerosis, the increased lesion area was not associated with worsened dyslipidemia, but with high blood pressure and inflammation. Oversecretion of GH in humans, leading to gigantism or acromegaly, is associated with increased cardiovascular morbidity and mortality [2]. Indeed, cardiovascular disease is the most common cause of death in these patients [2]. The prevalence of hypertension is higher in acromegalic patients [3], but there is little information regarding vascular pathology in general and atherosclerosis in particular. For example, data on early markers associated with atherosclerotic changes, such as increased IMT and endothelial dysfunction, is scant and conflicting [3,5], probably because of difficulties to find a relevant control group. In the current study, we directly addressed this issue using apoe / mice, which is a well known animal model of spontaneous atherosclerosis [10]. apoe / mice develop atherosclerotic lesions throughout the arterial tree, similar to those in humans [10]. These mice show severe hypercholesterolemia and develop atherosclerotic lesions even on a standard diet [10]. In the present study, we crossed apoe / mice with bgh transgenic mice, to produce heterozygous bgh overexpressing mice on a homozygous apoe null background and littermate apoe null mice as controls. The IGF- I levels corresponded well with previously published data from bgh transgenic mice [6]. The pro-atherogenic phenotype of the cross-bred animals was also well maintained as evidenced by a 0.6% extent of atherosclerosis in the thoracic aorta of littermate apoe / control animals at 22 weeks of age. This finding is comparable to a report by

75 I.J. Andersson et al. / Atherosclerosis xxx (2005) xxx xxx 7 Fig. 4. Representative aortic sinus lesions in male apoe / /bgh (C, D) and apoe / littermate mice (A, B) fed standard diet. The aortic sinus cross-section is shown in (A) and (C) (10, scale bar = 300 m). Lesions in apoe / littermates ranged from regular fatty streaks to moderate plaques, whereas lesions in apoe / /bgh mice almost exclusively were moderate or severe plaques. The lesions shown in (B) and (D) (40, scale bar = 100 m) are classified as mild and severe alterations, respectively. Daugherty et al. [19] and data from our own laboratory [20]. We found a larger lesion area in the thoracic aorta of male apoe / /bgh mice than in littermate apoe / mice on a standard diet. When fed Western diet, the difference in lesion area between the groups was even larger. We also quantified lesion area in female apoe / /bgh mice fed standard diet but found no differences in the extent of lesions between the genotypes. Important differences regarding lesion area between the sexes has been reported previously in several different models [21]. The reason for the different sensitivity between females and males can only be speculated upon. However, it is tempting to speculate that estrogen might protect against the pro-atherogenic effect of GH. Our experiments in mice on both standard and Western diet convincingly showed that thoracic aorta lesion area was larger in bgh overexpressing mice. However, when analysing the extent of atherosclerosis in other sections of the aorta, it was evident that the extent of atherosclerosis in the aortic arch was not influenced by GH overproduction. The different response to GH between the aortic regions is most likely explained by the fact that atherosclerosis in apoe / mice first appear in the proximal aorta and later in the more distal segments [10,22]. Thus, an earlier quantification of the atherosclerosis in the aortic arch could have revealed a difference between the genotypes. However, although there was no numeric increase in lesion area in the aortic arch, the frequency of moderate and severe plaque in the aortic sinus was higher in apoe / /bgh mice fed standard diet compared with littermate apoe / mice. When fed the Western diet, the lesions were generally more advanced and all mice had moderate or severe plaques with no difference between the groups. To elucidate the mechanisms behind accelerated lesion formation, we measured known risk factors associated with atherosclerosis. High plasma concentrations of cholesterol, and especially high ratio between LDL/HDL, have been considered a principal risk factor for atherosclerosis [23]. However, apoe / /bgh mice on both standard and Western diet had substantially reduced (VLDL + LDL)/HDL ratios compared with littermate apoe / mice. Furthermore, triglycerides, which are also considered a cardiovascular risk factor [24], were reduced in apoe / /bgh mice. However, serum concentrations of apob, which is a sensitive risk index of

76 8 I.J. Andersson et al. / Atherosclerosis xxx (2005) xxx xxx Fig. 5. Total cholesterol, HDL cholesterol, and VLDL + LDL cholesterol in male apoe / /bgh and apoe / littermate mice fed standard diet (A, n = 8 apoe / /bgh, n = 7 apoe / ) or Western diet (B, n = 6 apoe / /bgh, n = 8 apoe / ). The cholesterol profiles were measured using a high performance liquid chromatography system. Values are means ± S.E.M. * P < 0.05, ** P < 0.01, *** P < 0.001, Student s t-test. vascular disease [25], were not changed in apoe / /bgh mice compared with littermate apoe / mice, showing that the number of atherogenic particles in the circulation did not differ between the genotypes. Altogether these results point towards a relative improvement of the lipid profile in the apoe / /bgh mice compared with littermate apoe / mice. The lipid-profile followed the same pattern in mice fed Western diet for the last 8 weeks. However, the serum levels of cholesterol and apob were roughly 50% higher than in mice without dietary challenge. Taken together, dyslipidemia was not worsened in apoe / /bgh mice compared with littermate apoe / mice and thus, a more severe plasma lipoprotein profile could not explain the increased atherogenesis in apoe / /bgh mice. Another well established risk factor for atherosclerosis is hypertension [26]. We have previously reported that bgh transgenic mice have increased blood pressure [7], a finding which is confirmed in the present study, both in animals fed standard diet and Western diet. Thus, high blood pressure might be one mechanism underlying aggravated atherosclerosis in this mouse model. Although the initiation step in the atherosclerotic process is still an issue of debate, it is suggested that inflammation play an important role in atherosclerosis [26]. To evaluate the involvement of inflammation, we measured S-amyloid A levels. There is a strong relationship between S-amyloid A and cardiovascular events [27]. Interestingly, this protein Fig. 6. Serum triglycerides (A, C) and apob (B, D) in male apoe / /bgh and apoe / littermate mice fed standard diet (A, n = 8 apoe / /bgh, n = 11 apoe / ; B, n = 7 apoe / /bgh, n = 11 apoe / ) or Western diet (C, n = 6 apoe / /bgh, n = 8 apoe / ;D,n = 6 apoe / /bgh, n = 8 apoe / ). Serum triglyceride and apob concentrations were determined with an enzymatic colorimetric assay and an electroimmunoassay, respectively. Values are means ± S.E.M. ** P < 0.01, Student s t-test.

77 I.J. Andersson et al. / Atherosclerosis xxx (2005) xxx xxx 9 Fig. 7. S-amyloid A in male apoe / /bgh and apoe / littermate mice fed standard diet (A, n = 7 apoe / /bgh, n = 5 apoe / ) or Western diet (B, n = 4 apoe / /bgh, n = 6 apoe / ) and hepatic CRP mrna in apoe / /bgh and apoe / littermate mice fed Western diet (C, n =6 apoe / /bgh, n = 8 apoe / ). S-amyloid A was measured by a solid phase sandwich ELISA and hepatic CRP mrna expression was determined with quantitative real-time PCR and normalized to the endogenous control 36B4. Values are means ± S.E.M. ** P < 0.01, Student s t-test. was about 100% higher in apoe / /bgh mice on a standard diet, compared to the control group. This finding may suggest that there indeed is an ongoing inflammation, but the possibility that GH specifically regulates hepatic S-amyloid A 3 gene expression cannot be excluded [28]. The difference in S-amyloid A between the genotypes was not seen when they were fed Western diet. A possible explanation for this might be that S-amyloid A levels are up-regulated by Western diet [29]. We therefore also measured hepatic CRP mrna levels in the mice fed Western diet. Indeed, hepatic CRP levels were higher in apoe / /bgh mice compared with littermate apoe / mice. This finding is a strong indication of an ongoing general inflammatory process in the apoe / /bgh mice that is likely to contribute to the accelerated atherosclerotic process. In vitro data suggest that CRP is not only a risk marker but also a mediator in atherosclerosis due to its capacity to enhance monocyte recruitment [30] and facilitate the uptake of LDL by macrophages [31]. Indeed, overexpression of human CRP was reported to augment atherosclerosis in apoe / mice [32]. In the same study, CRP mrna was detected in the liver but not in the aorta, indicating that the CRP deposited in the lesions was derived from the liver. Thus, it is possible that high hepatic CRP expression may have accelerated the atherosclerotic process. In contrast to the findings in mice, acromegalic patients had lower plasma CRP levels than controls [33]. Moreover, treatment with a GH receptor antagonist resulted in markedly higher CRP levels than in the control population, indicating that treatment induced an inflammatory response [33]. The reason for the discrepant effects of GH overproduction on CRP expression in mice and humans is unclear. First, CRP expression could be differently regulated between the species as exemplified by the induction of hepatic CRP mrna by IL-6 in humans but not in mice [34]. Second, increased body fat is associated with increased CRP levels. apoe / and apoe / /bgh mice did not differ in absolute epididymal fat mass when fed Western diet but apoe / /bgh mice had a lower proportion fat relative to body weight, compared with apoe / mice. Acromegalic patients also have a decreased fat mass and the increased fat mass after treatment [35] might contribute to the increased CRP levels [33]. While it is possible that high blood pressure and inflammatory factors are involved in the atherosclerotic process in apoe / /bgh mice, it cannot be excluded that GH or IGF-I have direct effects in the cells of the vessel wall that contributes to the accelerated lesion formation. One possible mechanism of action is that the well known mitogenic effect of IGF-I could stimulate vascular smooth muscle cell proliferation, a process intimately involved in atherogenesis [36]. Moreover, GH has been shown to stimulate the growth of arterial medial cells in vitro [37]. In addition, cell culture studies have suggested that GH/IGF-I promotes macrophage uptake of LDL [38]. Thus, formation of foam cells that is an early event in atherosclerosis [26] might be accelerated in the presence of high GH and IGF-I levels, although apoe / /bgh mice had no worsened plasma lipoprotein profile. In conclusion, this study shows that excess GH accelerates the development of atherosclerosis in apoe / mice. The mechanism behind this effect could be direct effects of GH on cellular processes in the vessel wall or the result of concomitant processes such as hypertension or a general inflammatory process.

78 10 I.J. Andersson et al. / Atherosclerosis xxx (2005) xxx xxx Acknowledgements This study was supported by the Swedish Medical Research Council (12 580), the Swedish Heart Lung foundation, the Lundberg foundation and funds at Sahlgrenska University Hospital (LUA/ALF). The authors wish to thank Mrs. Jia Jing, for excellent help with blood pressure measurements. References [1] Sowers JR. Insulin and insulin-like growth factor in normal and pathological cardiovascular physiology. Hypertension 1997;29(3): [2] Bengtsson BA, et al. Epidemiology and long-term survival in acromegaly. A study of 166 cases diagnosed between 1955 and Acta Med Scand 1988;223(4): [3] Colao A, et al. Systemic complications of acromegaly: epidemiology, pathogenesis, and management. Endocr Rev 2004;25(1): [4] Poulter N. Global risk of cardiovascular disease. Heart 2003; 89(90002):2 5. [5] Otsuki M, et al. Characterization of premature atherosclerosis of carotid arteries in acromegalic patients. Clin Endocrinol (Oxf) 2001;54(6): [6] Frick F, et al. Long-term growth hormone excess induces marked alterations in lipoprotein metabolism in mice. Am J Physiol Endocrinol Metabol 2001;281(6):E [7] Bohlooly YM, et al. Vascular function and blood pressure in GH transgenic mice. Endocrinology 2001;142(8): [8] Olsson B, et al. Bovine growth hormone transgenic mice are resistant to diet-induced obesity but develop hyperphagia, dyslipidemia, and diabetes on a high-fat diet. Endocrinology 2005;146(2): [9] Andersson IJ, et al. Endothelial dysfunction in growth hormone transgenic mice. Clin Sci (Lond), in press (Epub ahead of print). [10] Nakashima Y, et al. apoe-deficient mice develop lesions of all phases of atherosclerosis throughout the arterial tree. Arterioscler Thromb 1994;14(1): [11] Sandstedt J, et al. Disproportional bone growth and reduced weight gain in gonadectomized male bovine growth hormone transgenic and normal mice. Endocrinology 1994;135(6): [12] Andersson IJ, et al. Reduced sympathetic responsiveness as well as plasma and tissue noradrenaline concentration in growth hormone transgenic mice. Acta Physiol Scand 2004;182(4): [13] Paigen B, et al. Quantitative assessment of atherosclerotic lesions in mice. Atherosclerosis 1987;68(3): [14] Russell Jr HK. A modification of Movat s pentachrome stain. Arch Pathol 1972;94(2): [15] van Vlijmen BJ, et al. Diet-induced hyperlipoproteinemia and atherosclerosis in apolipoprotein E3-Leiden transgenic mice. J Clin Invest 1994;93(4): [16] Linden D, et al. PPARalpha deficiency increases secretion and serum levels of apolipoprotein B-containing lipoproteins. J Lipid Res 2001;42(11): [17] Oscarsson J, et al. Differential effects of continuous versus intermittent administration of growth hormone to hypophysectomized female rats on serum lipoproteins and their apoproteins. Endocrinology 1989;125(3): [18] Bohlooly YM, et al. Enhanced spontaneous locomotor activity in bovine GH transgenic mice involves peripheral mechanisms. Endocrinology 2001;142(10): [19] Daugherty A, Manning MW, Cassis LA. Angiotensin II promotes atherosclerotic lesions and aneurysms in apolipoprotein E-deficient mice. J Clin Invest 2000;105(11): [20] Johansson ME, et al. Angiotensin II, type 2 receptor is not involved in the angiotensin II-mediated pro-atherogenic process in apoe / mice. J Hypertens 2005;23(8): [21] Caligiuri G, et al. Effects of sex and age on atherosclerosis and autoimmunity in apoe-deficient mice. Atherosclerosis 1999;145(2): [22] Reddick RL, Zhang SH, Maeda N. Atherosclerosis in mice lacking apoe. Evaluation of lesional development and progression. Arterioscler Thromb 1994;14(1): [23] Tulenko TN, Sumner AE. The physiology of lipoproteins. J Nucl Cardiol 2002;9(6): [24] Austin MA. Epidemiology of hypertriglyceridemia and cardiovascular disease. Am J Cardiol 1999;83(9B):13F 6F. [25] Sniderman AD, et al. Apolipoproteins versus lipids as indices of coronary risk and as targets for statin treatment. Lancet 2003;361 (9359): [26] Ross R. Atherosclerosis an inflammatory disease. N Engl J Med 1999;340(2): [27] Johnson BD, et al. Serum amyloid A as a predictor of coronary artery disease and cardiovascular outcome in women: the national heart, lung, and blood institute-sponsored women s ischemia syndrome evaluation (WISE). Circulation 2004;109(6): [28] Olsson B, et al. Bovine growth hormone-transgenic mice have major alterations in hepatic expression of metabolic genes. Am J Physiol Endocrinol Metabol 2003;285(3):E [29] Lewis KE, et al. Increase in serum amyloid A evoked by dietary cholesterol is associated with increased atherosclerosis in mice. Circulation 2004;110(5): [30] Torzewski M, et al. C-reactive protein in the arterial intima: role of C-reactive protein receptor-dependent monocyte recruitment in atherogenesis. Arterioscler Thromb Vasc Biol 2000;20(9): [31] Zwaka TP, Hombach V, Torzewski J. C-reactive protein-mediated low density lipoprotein uptake by macrophages: implications for atherosclerosis. Circulation 2001;103(9): [32] Paul A, et al. C-reactive protein accelerates the progression of atherosclerosis in apolipoprotein E-deficient mice. Circulation 2004;109(5): [33] Sesmilo G, et al. Cardiovascular risk factors in acromegaly before and after normalization of serum IGF-I levels with the GH antagonist pegvisomant. J Clin Endocrinol Metabol 2002;87(4): [34] Ku NO, Mortensen RF. The mouse C-reactive protein (CRP) gene is expressed in response to IL-1 but not IL-6. Cytokine 1993;5(4): [35] Bengtsson BA, Brummer RJ, Bosaeus I. Growth hormone and body composition. Horm Res 1990;33(Suppl. 4): [36] Delafontaine P, Song Y-H, Li Y. Expression regulation, and function of IGF-1, IGF-1R, and IGF-1 binding proteins in blood vessels. Arterioscler Thromb Vasc Biol 2004;24(3): [37] Ledet T. Growth hormone stimulating the growth of arterial medial cells in vitro. Absence of effect of insulin. Diabetes 1976;25(11): [38] Hochberg Z, et al. Growth hormone and insulin-like growth factor-i increase macrophage uptake and degradation of low density lipoprotein. Endocrinology 1992;131(1):430 5.

79 II

80

81 Importance of PPARα for the effects of growth hormone on hepatic lipid and lipoprotein metabolism Anna Ljungberg 1,2, Daniel Lindén 1,4, Caroline Améen 1,2, Göran Bergström 3 and Jan Oscarsson 1,2,4 1 Wallenberg Laboratory for Cardiovascular Research, 2 Department of Physiology, and 3 Department of Clinical Physiology, Cardiovascular Institute, The Sahlgrenska Academy at Göteborg University, SE Göteborg, Sweden; 4 AstraZeneca R&D, SE Mölndal, Sweden. ABSTRACT Growth hormone (GH) enhances lipolysis in adipose tissue, thereby increasing the flux of fatty acids to other tissues. Moreover, GH increases hepatic triglyceride synthesis and secretion in rats and decreases the action of peroxisome proliferator-activated receptor (PPAR)α. PPARα is activated by fatty acids and regulates hepatic lipid metabolism in rodents. The aim of this study was to investigate the importance of PPARα for the effects of GH on hepatic gene expression and lipoprotein metabolism. GH was given as a continuous infusion (5 mg/kg/day) for 7 days to female PPARα-null and wild-type (wt) mice. GH treatment decreased hepatic triglyceride content and increased hepatic triglyceride secretion rate and serum cholesterol levels. Furthermore, GH decreased the hepatic mrna expression of acyl-coa oxidase, medium-chain acyl-coa dehydrogenase and PPARγ1. All these GH effects were independent of PPARα. GH treatment decreased Cyp4a10 mrna expression in wt mice, but increased Cyp4a10 mrna expression in PPARα-null mice. Thus, the decrease in Cyp4a10 expression by GH is dependent on PPARα. Furthermore, GH decreased hepatic PPARγ2 mrna expression in PPARα-null mice, whereas this response was not observed in wt controls. In summary, PPARα plays a minor role for the effects of GH on lipid and lipoprotein metabolism. However, the effect of GH on Cyp4a10 and PPARγ2 gene expression is PPARα-dependent. Key words: apolipoprotein B, triglycerides, Cyp4a10, MTP, PPARγ, liver INTRODUCTION In addition to its well-known effect on longitudinal bone growth, growth hormone (GH) plays an important role in the regulation of lipid metabolism. GH influences body composition in terms of a moderate increase in lean body mass and a more marked decrease in body fat mass [1]. This is due to the lipolytic and antilipogenic action of GH in adipose tissue [2] that results in increased flux of fatty acids to other tissues [3]. In contrast to the effect of GH in adipose tissue, GH increases lipid synthesis in the liver. GH treatment in vivo increased hepatic triglyceride synthesis and very low-density lipoprotein (VLDL) secretion in rats [4-6] and VLDL secretion in man [7]. These effects of GH could be direct on hepatocytes or indirect via increased flux of fatty acids to the liver, since GH and oleic acid incubation of primary rat hepatocytes had similar effects on triglyceride synthesis and VLDL secretion [8]. Moreover, continuous GH infusion to hypophysectomized rats increased hepatic expression of sterol regulatory element binding protein (SREBP)-1c and most of its downstream target genes [9], indicating that GH also increases hepatic de novo lipogenesis. Thus, continuous GH administration could result in both increased flux of fatty acids to the liver and increased hepatic lipogenesis. Unsaturated long-chain fatty acids and their derivatives are potent endogenous activators of the nuclear receptor peroxisome proliferator-activated receptor 1

82 (PPAR)α [10]. PPARα is mainly expressed in tissues with a high degree of fatty acid metabolism, such as liver, heart, brown adipose tissue, kidney and skeletal muscle [11]. Although PPARα expression is highest in liver in rodents, this is not the case for human [12] or non-human [13] primates. PPARα agonists, i.e. fibrates, are used in the treatment of hypertriglyceridemia [14]. Activation of PPARα in the liver increases transcription of genes involved in fatty acid uptake as well as mitochondrial, peroxisomal and microsomal fatty acid oxidation [15]. In addition, the hepatic expression of apolipoprotein (apo) A-I/A-II, apoc-iii [15] and microsomal triglyceride transfer protein (MTP) [16] is regulated by PPARα. Furthermore, PPARα activation by fibrates leads to increased hepatic fatty acid oxidation and decreased VLDLtriglyceride secretion from primary rat hepatocytes [17]. There are a few studies showing that GH and PPARα interact in the regulation of hepatic metabolism. A continuous infusion of GH has been found to suppress the peroxisome proliferator induction of hepatic peroxisomal β-oxidation [18, 19] and acyl-coa oxidase (ACO) mrna [20] as well as cytochrome P450 (CYP) 4Amediated ω oxidation [19] and CYP4A mrna [21]. Moreover, GH decreased PPARα mrna in cultured rat hepatocytes [22, 23] and hepatic PPARα mrna and protein expression in hypophysectomized rats [24]. GH has also been shown to decrease PPARα transcriptional activity [25] at the ligand-independent N-terminal activation function region-1 (AF-1 region) domain of PPARα [26]. Thus, GH may counteract PPARα signalling by decreasing the expression level of PPARα or interfering with PPARα signalling by other means. On the other hand, GH may increase the supply of ligands for PPARα, either via increased lipolysis in adipose tissue or via de novo synthesized fatty acids in the liver. Thus, GH and PPARα may interact in a complex manner in the regulation of hepatic lipid metabolism. The aim of this study was therefore to determine the importance of PPARα for the effects of GH on lipid and lipoprotein metabolism. GH was administered as a continuous infusion (5 mg/kg/day) to both PPARα-null and wild-type (wt) mice for 7 days. The effect of GH on hepatic triglyceride secretion rate, serum and liver lipid levels, as well as hepatic expression of several genes of importance for lipid and lipoprotein metabolism, was investigated. MATERIALS AND METHODS Animals and hormonal treatment Homozygous PPARα-null mice on pure Sv/129 genetic background and corresponding wt Sv/129 control mice were kindly provided by Dr. F. J. Gonzalez (NIH, Bethesda, MD, USA) [27] and kept on the Sv/129 background. The mice were maintained under standardized conditions of temperature (24-26 o C) and humidity (50-60%), with lights on between h. The animals had free access to water and standard laboratory chow containing (w/w) 4% fat, 58% carbohydrates, 16.5% protein and 6% ashes with a total energy content of 12.6 kj/g (R- 34, Lactamin AB, Kimstad, Sweden). Female mice were given recombinant bovine GH (2.5 or 5 mg/kg/day) as a continuous infusion by means of Alzet osmotic minipumps (model 2001, Alza, Palo Alto, CA) implanted subcutaneously between the scapulae on the back. The mice were anesthetized with a combination of ketamine hydrochloride (77 mg/kg; Ketalar, Parke-Davis, Detroit, MI) and xylazine (9 mg/kg; Rompun, Bayer, Lever- Kusen, Germany) during implantation of the osmotic minipumps. The recombinant bovine GH was a generous gift from Dr. Parlow (NIH, Torrance, CA). The hormone was diluted in 0.05 M phosphate buffer (ph 8.6) with 1.6% glycerol and 0.02% sodium azide. The hormonal treatment continued for 7 days. At the end of the experiments, mice were anesthetized with isoflurane (Forene, Abbot Scandinavia AB, Sweden) and killed between 0900 and 1100 h. Blood was collected by cardiac puncture and tissues were rapidly removed, frozen in liquid nitrogen and stored at - 2

83 80 o C until analysis. The Ethics Committee of Göteborg University approved this study. All animal experimentation was conducted in accordance with accepted standards of humane animal care. Serum lipids and hepatic triglyceride content Serum apob concentrations were determined with an electroimmunoassay as previously described [28]. Serum triglyceride and cholesterol concentrations were determined with enzymatic colorimetric assays (TG and chol; Roche, Mannheim, Germany). The size distribution profiles of serum lipoproteins were measured using a high performance liquid chromatography system, SMART, as described before [28]. In brief, 10 µl pooled serum from 8 mice in each group was loaded on a Superose 6 PC 3.2/30 column (Amersham Pharmacia Biotech, Uppsala, Sweden). The chromatographic system was linked to an air segmented continuous flow system for on-line postderivatization analysis of total cholesterol. Frozen livers were homogenized in isopropanol (1 ml/50 mg tissue) and incubated at 4ºC for 1 h. The samples were centrifuged in 4ºC for 5 min at 2500 rpm and triglyceride concentrations in the supernatants were measured as described above. In vivo hepatic triglyceride secretion rate Triglyceride secretion rate in vivo was measured by intravenous administration of Triton WR-1339 (Sigma, St Louis, MO, USA) [29] that blocks the peripheral hydrolysis of triglycerides. The animals were fasted for 4 h ( h) to avoid the influence of chylomicrons from the intestine. Thereafter, the mice were anesthetized with a combination of ketamine hydrochloride and xylazine, and injected intravenously with Triton WR diluted in saline (200 mg/ml) via the jugular vein (500 mg/kg body weight). Blood samples were taken before the injection (baseline fasting triglyceride concentration) and 30, 60 and 90 min after Triton WR-1339 administration. The triglyceride accumulation was linear during this time period. Plasma triglyceride levels were analyzed as described above, and accumulation of triglycerides was calculated using published plasma volume in normal female mice (0.09 ml/g body weight) [29]. Hepatic triglyceride secretion rate, expressed as µmol/min/kg body weight, was calculated from the slope of the curve. The triglyceride clearance rate (ml/min) was calculated as the ratio of hepatic triglyceride secretion rate (µmol/min) to baseline fasting triglyceride concentration (µmol/ml). cdna synthesis and real-time PCR Total liver RNA was isolated from frozen liver with TriReagent (Sigma). DNAfree (Ambion, Austin, Tx, USA) was used to remove contaminating DNA from the RNA preparations. First strand cdna was synthesized from 0.4 µg of total RNA with TaqMan Reverse Transcription Reagents (Applied Biosystems, Foster City, CA, USA). Quantitative real-time PCR was performed with the ABI Prism 7700 Sequence Detection system, using the SYBR Green labelling system (Applied Biosystems). All samples were analyzed in triplicate and the expression data were normalized to the endogenous control acidic ribosomal phosphoprotein P0 (36B4). Specific primers for each gene (Table 1) were designed with Primer Express software (Applied Biosystems) and gene sequences available from GenBank database. To avoid amplification of genomic DNA, primers were positioned to span exon junctions when possible. The correct sizes of the amplicons were verified by gel electrophoresis. Statistics Values are expressed as means ± SEM. Comparisons between groups were made by 2-way ANOVA using genotype and GH treatment as factors. An interaction term was used to investigate if the interaction between the factors was significant. If significant, the interaction term was kept in the model. A significant interaction describes that the GH response was significantly different between genotypes. Values were transformed to logarithms when appropriate. P<0.05 was considered significant. 3

84 Table 1. Primers used for real-time PCR Gene Acc. No Forward primer (5 3 ) Reverse primer (5 3 ) ACO AF CAGCAGGAGAAATGGATGCA GGGCGTAGGTGCCAATTATCT Cyp4a10 XM_ CAACACATCTCCTTAATGACCCTAGAC CCTGTAATTTCCATCTACCTGAACACT IGF-I NM_ GCTGGTGGATGCTCTTCAGTT CGAATGCTGGAGCCATAGC LXRα AF CAATGCCTGATGTTTCTCCTGAT CCTGCATCTTGAGGTTCTGTCTT MCAD BC GCCCTCCGCAGGCTCT ACCCTTCTTCTCTGCTTTGGTCT MTP L47970 GCTCCCTCAGCTGGTGGAT CAGGATGGCTTCTAGCGAGTCT PPARα NM_ CACGATGCTGCTCTCCTTGA GTGTGATAAAGCCATTGCCGTA PPARγ U01841 TGACAGGAAAGACAACGGACAA ATCTTCTCCCATCATTAAGGAATTCAT PPARγ1 U01841 GCGGCTGAGAAATCACGTTC GAATATCAGTGGTTCACCGCTTC PPARγ2 U09138 AACTCTGGGAGATTCTCCTGTTGA GAAGTGCTCATAGGCAGTGCAT SREBP-1 AB GCAGACCCTGGTGAGTGGA GTCGGTGGATGGGCAGTTT 36B4 NM_ GAGGAATCAGATGAGGATATGGGA AAGCAGGCTGACTTGGTTGC RESULTS In initial studies, 20 weeks old female mice (Sv/129) were given either vehicle or two different doses of bovine GH (2.5 mg/kg/day or 5 mg/kg/day) for 7 days via osmotic minipumps. Neither dose of GH significantly influenced PPARα mrna expression (data not shown). The higher dose was chosen for the subsequent experiments since this dose had an evident effect on plasma cholesterol levels (data not shown). Furthermore, this dose of human GH has been shown to increase body weight gain and hepatic mrna expression of insulin-like growth factor I (IGF-I) in intact mice [30], indicating that the given dose is clearly higher than the endogenous GH secretion. Effects on body and organ weights In the next experiments, weeks old female PPARα-null and wt mice were given either vehicle or GH (5 mg/kg/day) via osmotic minipumps. Before the start of the experiments, body weights were comparable between the genotypes (wt: 27.1 ± 1.40 g, PPARα-null: 26.8 ± 0.83 g). One week of GH treatment increased body weight gain and liver weights, indicating a GH response. Gonadal fat weights were not significantly influenced by GH treatment. The response to GH treatment was not different between the genotypes (Table 2). Table 2. Effects of GH on body weight and tissue weights in PPARα-null mice and wt mice Group Body weight gain (g) Liver weight (% bw) Gonadal fat weight (% bw) Wt vehicle ± ± ± 0.45 Wt GH 1.91 ± 0.22 * 4.10 ± 0.22 * 3.29 ± 0.27 KO vehicle ± ± ± 0.28 KO GH 1.28 ± 0.12 * 4.32 ± 0.10 * 3.91 ± 0.97 PPARα-null mice (KO) and wt mice were administered bovine GH (5 mg/kg/day) for 7 days as a continuous infusion. Values are means ± SEM (n=4-8). Comparisons between groups were made by 2-way ANOVA using GH treatment and genotype as factors. Since no significant interaction between the factors was observed, no interaction term was used in the model. *p<0.05 GH treated mice vs control mice in both genotypes. 4

85 Effects on serum lipids and hepatic triglycerides GH treatment had no effect on serum triglycerides, whereas serum cholesterol levels increased (Table 3). PPARα-null mice had higher serum triglycerides than wt mice, while cholesterol levels were similar between the genotypes (Table 3). The lipoprotein size distribution profiles were measured in pooled serum from 8 mice in each group (Fig. 1A). In both genotypes, GH treatment increased total serum cholesterol and shifted the large peak, representing LDL and HDL, to the left, i.e. towards less dense particles. Moreover, the height of the main peak increased by GH, indicating increased HDL cholesterol levels. Serum apob levels increased by GH treatment (Fig. 1B). In addition, serum apob levels were higher in PPARα-null mice compared to wt mice, as previously observed [28]. Thus, PPARαdeficiency and high GH levels have additive effects on serum apob levels. GH treatment decreased hepatic triglyceride content, especially in PPARα-null mice, while PPARα-deficiency had no significant effect (Fig. 1C). Table 3. Effects of GH on serum lipids in PPARα-null mice and wt mice. Figure 1. Serum lipoproteins (A), serum apob (B) and hepatic triglyceride content (C) in PPARα-null and wt mice treated with GH. PPARα-null mice (KO) and wt mice were administered bovine GH (5 mg/kg/day) for 7 days as a continuous infusion. The cholesterol profiles were determined in pooled serum from 8 mice in each group using a size exclusion high performance liquid chromatography system, SMART. The yellow line denotes wt mice given vehicle, the blue line denotes wt mice given GH, the black line denotes PPARα-null given vehicle and the red line denotes PPARα-null mice given GH. Serum apob was determined with an electroimmunoassay (n=4-5). Hepatic TG content was measured after homogenization in isopropanol (n=4-7). Values are means ± SEM. Comparisons between groups were made by 2-way ANOVA using GH treatment and genotype as factors. Since no significant interaction between the factors was observed, no interaction term was used in the model. *p<0.05 GH treated mice vs control mice including both genotypes. #p<0.05 PPARα-null mice vs wt mice including both GH treated and vehicle treated mice. Group Triglycerides (mm) Cholesterol (mm) Wt vehicle 1.13 ± ± 0.21 Wt GH 0.74 ± ± 0.42 * KO vehicle 1.46 ± 0.29 # 2.83 ± 0.13 KO GH 1.33 ± 0.06 # 4.36 ± 0.35 * PPARα-null mice (KO) and wt mice were administered bovine GH (5 mg/kg/day) for 7 days as a continuous infusion. Values are means ± SEM (n=4-8). Comparisons between groups were made by 2-way ANOVA using GH treatment and genotype as factors. Since no significant interaction between the factors was observed, no interaction term was used in the model. *p<0.05 GH treated mice vs control mice including both genotypes. #p<0.05 PPARα-null mice vs wt mice including both GH treated and vehicle treated mice. 5

86 Effects on hepatic triglyceride secretion To study the effects of GH on hepatic triglyceride secretion and triglyceride clearance rates in vivo, female PPARα-null and wt mice of weeks of age were given either vehicle or GH (5 mg/kg/day) via osmotic minipumps for 7 days in a separate experiment. GH treatment as well as PPARα-deficiency increased triglyceride secretion (Fig. 2A). Thus, PPARα-deficiency and GH treatment have additive effects on hepatic triglyceride secretion. Since the serum triglyceride levels were unchanged by GH treatment, the calculated triglyceride clearance rate was increased by GH treatment (Fig. 2B). However, PPARα-deficiency did not influence calculated triglyceride clearance. Effects on IGF-I and genes involved in VLDL secretion and fatty acid oxidation To further study the importance of PPARα for the effects of GH, known GH or PPARα-sensitive genes, representing different aspects of hepatic lipid metabolism, were measured with quantitative real-time PCR. GH treatment had no significant effect on PPARα mrna expression (data not shown), while the hepatic expression of IGF-I mrna increased after GH treatment (Fig. 3A). MTP is essential and rate-limiting for the assembly and secretion of apob-containing lipoproteins [31, 32]. The increased triglyceride secretion rate following GH treatment was not associated with increased MTP gene expression (Fig. 3B). However, the higher triglyceride secretion in PPARα-null mice than their wt controls correlated to higher MTP mrna levels. GH treatment resulted in decreased mrna expression of ACO (Fig. 3C) and mediumchain acyl-coa dehydrogenase (MCAD; Fig. 3D), indicating decreased peroxisomal and mitochondrial β-oxidation, respectively. However, the expression levels were not influenced by PPARαdeficiency. Cyp4a10, which is involved in microsomal ω-oxidation of fatty acids, was expressed at very low levels in PPARα-null mice (Fig. 3E). Interestingly, there was an interaction between GH and PPARα in the regulation of Cyp4a10. GH decreased Cyp4a10 mrna levels in wt mice (-59%), while the expression was increased by GH in PPARα-null mice (+210%). Thus, the effect of GH on Cyp4a10 mrna expression is PPARα-dependent. Figure 2. In vivo hepatic triglyceride secretion rate (A) and triglyceride clearance rate (B) in PPARα-null mice and wt mice treated with GH. PPARα-null mice and wt mice were administered bovine GH (5 mg/kg/day) for 7 days as a continuous infusion. Serum triglyceride (TG) levels were determined before (0 min) and 30, 60, and 90 min after Triton WR-1339 injection (500 mg/kg). Hepatic TG secretion rate was then calculated from the slope of the curve and TG clearance was determined from the hepatic TG secretion and the baseline fasting TG concentration (n=6-10). Values are means ± SEM. Comparisons between groups were made by 2-way ANOVA using GH treatment and genotype as factors. Since no significant interaction between the factors was observed, no interaction term was used in the model. *p<0.05 GH treated mice vs control mice including both genotypes. #p<0.05 PPARα-null mice vs wt mice including both GH treated and vehicle treated mice. 6

87 Figure. 3. Hepatic mrna expression of IGF-I (A), MTP (B), ACO (C), MCAD (D) and Cyp4a10 (E) in PPARα-null mice and wt mice treated with GH. PPARα-null mice and wt mice were administered bovine GH (5 mg/kg/day) for 7 days as a continuous infusion. Hepatic mrna expression was determined with quantitative real-time PCR. Values are means ± SEM (n=4-7). Comparisons between groups were made by 2-way ANOVA using GH treatment and genotype as factors. An interaction term was used to investigate if the interaction between the factors was significant. If significant, the interaction term was kept in the model. *p<0.05 GH treated mice vs control mice including both genotypes. #p<0.05 PPARα-null mice vs wt mice including both GH treated and vehicle treated mice. p<0.05 different responses in wild-type and PPARα-null mice. IGF-1: insulin-like growth factor-1, MTP: microsomal triglyceride transfer protein, ACO: acyl-coa oxidase, MCAD: medium-chain acyl-coa dehydrogenase. Effects on genes involved in lipogenesis To examine whether the stimulated triglyceride secretion was associated with an increased hepatic lipogenesis, we measured the lipogenic transcription factors SREBP-1 and liver X receptor (LXR)α. GH treatment had no effect on SREBP-1 (Fig. 4A) or LXRα (Fig. 4B) mrna, while PPARα-deficiency induced the expression of both genes. Furthermore, hepatic PPARγ was measured since it has been shown to be important for liver triglycerides [33], plasma triglyceride clearance and insulin sensitivity [34]. In 7 line with the decreased hepatic triglyceride content, GH treatment resulted in decreased expression of PPARγ1 (Fig. 4C) and PPARγ2 (Fig. 4D). The regulation of total PPARγ (data not shown) was similar to PPARγ1 as expected since the hepatic expression level of PPARγ1 mrna is higher than that of PPARγ2. Interestingly, statistical analysis of PPARγ2 mrna expression showed an interaction between GH treatment and genotype. Thus, presence of PPARα blunted an effect of GH on PPARγ2 mrna expression (Fig. 4D).

88 Figure 4. Hepatic mrna expression of SREBP-1 (A), LXRα (B), PPARγ1 (C) and PPARγ2 (D) in PPARα-null mice and wt mice treated with GH. PPARα-null mice and wt mice were administered bovine GH (5 mg/kg/day) for 7 days as a continuous infusion. Hepatic mrna expression was determined with quantitative real-time PCR. Values are means ± SEM (n=4-7). Comparisons between groups were made by 2-way ANOVA using GH treatment and genotype as factors. An interaction term was used to investigate if the interaction between the factors was significant. If significant, the interaction term was kept in the model. *p<0.05 GH treated mice vs control mice including both genotypes. #p<0.05 PPARα-null mice vs wild type mice including both GH treated and vehicle treated mice. p<0.05 different responses in wild-type and PPARα-null mice. SREBP-1: sterol regulatory element binding protein-1, LXRα: liver X receptor α, PPARγ: peroxisome proliferator-activated receptor γ. DISCUSSION The main finding in this study was that PPARα plays a minor role for the overall effect of GH on hepatic lipid and lipoprotein metabolism, including hepatic expression of genes involved in different aspects of fatty acid metabolism. However, an interaction was observed between PPARα and GH in the regulation of hepatic Cyp4a10 and PPARγ2 gene expression. Hepatic expression of PPARα was important for GH-mediated downregulation of Cyp4a10 and for maintenance of PPARγ2 expression in the liver following GH treatment. The reason for the few observed interactions between GH and PPARα is unclear. It has recently been shown that adipose tissue-derived fatty acids cannot replace de novo synthesized fatty acids as PPARα ligands in a situation of abolished hepatic fatty acid synthesis [35]. Thus, distinct pools of PPARα could be activated in various metabolic situations. It could be speculated that PPARα is more important for GH effects than observed in this study in a stress situation like fasting when fatty acid synthesis is inhibited. GH treatment decreased Cyp4a10 mrna in wt mice, whereas the expression was upregulated in response to GH in PPARα-null mice. Cyp4a10 is involved in ω-oxidation of fatty acids that produces dicarboxylic fatty acids. Compared to the β-oxidation pathway, ω-oxidation is a minor pathway for fatty acid oxidation. However, induction of Cyp4a10 mrna in streptozotocin-induced diabetic mice [36] and obese (ob/ob) mice [37], suggest that this pathway may be of increased importance when fatty acid flux to the liver is increased. Furthermore, microsomal lipid peroxidation in Cyp2e1-null mice was markedly inhibited by anti-mouse Cyp4a10 antibodies [38]. Thus, presence of PPARα might be of importance for down- 8

89 regulation of hepatic ω-oxidation and lipid peroxidation by GH. To the best of our knowledge, this study shows for the first time that GH regulates hepatic PPARγ mrna levels. It has previously been shown that incubation of primary rat preadipocytes with GH markedly reduced triglyceride content in parallel with decreased expression of PPARγ [39]. GH treatment decreased the hepatic expression of both PPARγ1 and PPARγ2. Interestingly, there was an interaction between GH and PPARα in the regulation of PPARγ2. PPARα prevented down-regulation of PPARγ2 expression by GH since the down-regulation was only observed in GH treated PPARα-null mice. Hepatic overexpression of PPARγ1 in vivo [33] and PPARγ2 in vitro [40] showed that PPARγ increases hepatic triglyceride synthesis and accumulation. In contrast to PPARα, PPARγ is most likely expressed at a higher level in human liver as compared to rat and mouse liver [41]. Subjects with liver steatosis have increased hepatic expression of PPARγ mrna [42], indicating a possible role of PPARγ also in development of fatty liver in humans. An increased prevalence of hepatic steatosis has been observed in patients with GHdeficiency [43]. It is therefore tempting to speculate that the effect of GH on PPARγ has a role for liver triglyceride content also in man. Based on the present findings, it could be speculated that low hepatic expression of PPARα would facilitate GH induced down-regulation of PPARγ2 and subsequently the liver triglyceride content. Interestingly, the effect of GH on hepatic triglyceride content tended to be more pronounced in PPARα-null mice, suggesting that the decreased PPARγ2 expression could contribute to the effect of GH on hepatic triglyceride content [34, 44]. The decrease in hepatic triglyceride stores after GH treatment is probably not explained by increased peroxisomal and mitochondrial fatty acid oxidation, since ACO and MCAD mrna expression decreased by GH, in line with previous studies [18-20]. Another possibility is that the use of the hepatic triglyceride stores for VLDL assembly and secretion is facilitated by GH. The present study, shows for the first time that GH treatment results in stimulated hepatic triglyceride secretion in mice. In contrast to rats [45], this effect was not associated with elevated expression of MTP. Thus, increased expression of MTP is not required for increased hepatic triglyceride secretion following continuous infusion of GH to mice. Also PPARα-null mice had stimulated hepatic triglyceride secretion along with increased serum triglycerides and apob compared to wt controls, as described before [28]. However, in this case, the increased triglyceride secretion was accompanied with increased levels of MTP mrna, suggesting that increased MTP expression might contribute to the increased hepatic triglyceride secretion in PPARα-null female mice. MTP mrna expression in PPARα-null males was not different from wt males (data not shown), in line with a previous study [46]. Thus, the MTP expression is differently regulated by lack of PPARα in males and females. In line with findings in bgh transgenic mice [47], GH treatment increased total serum cholesterol levels, mainly as a result of increased HDL cholesterol levels. GH increased serum apob levels, indicating an increased number of LDL-VLDL particles. This finding is in line with the increased hepatic triglyceride secretion observed after GH treatment, indicating increased secretion of apob. Moreover, the increased serum level of apob following GH treatment is in line with the observation of markedly decreased serum levels of apob in GH receptor-deficient mice [48]. GH treatment increased body weight gain and liver weights, demonstrating that the given dose was sufficient to induce effects of GH. However, in contrast to studies in GH treated hypophysectomized rats [24] and bgh transgenic mice [49], no effect on PPARα mrna expression was observed. Thus, GH treatment resulted in an expected increase in body weight gain and IGF-I mrna [30] but was not sufficient to decrease PPARα mrna expression. The 9

90 reason for this is unclear, but may be due to the duration of GH treatment as compared to bgh transgenic mice [49] or to the fact that the control mice in this study have a physiological GH secretion in contrast to hypophysectomized rats [24]. In summary, GH treatment decreased hepatic triglyceride content and expression of PPARγ as well as ACO and MCAD mrna in the liver. In addition, GH increased hepatic triglyceride secretion, plasma triglyceride clearance and HDL cholesterol levels. These effects were not different between PPARα-null and wt mice, demonstrating that most of the studied effects of GH on hepatic lipid and lipoprotein metabolism were PPARαindependent. However, the effects of GH on Cyp4a10 and PPARγ2 expression were dependent on PPARα indicating that PPARα is important for some aspects of GH regulation of hepatic lipid metabolism, e.g. PPARγ signalling and ω-oxidation. ACKNOWLEDGEMENTS We thank Lennart Svensson and coworkers at AstraZeneca R&D, Mölndal for the analyses of lipoprotein fractions. We also thank Jing Jia for excellent technical assistance and Karin Nelander for valuable help with the statistical analyses. FUNDING This work was supported by grant from the Swedish Medical Research Council, King Gustav V:s and Queen Victorias Foundation, AstraZeneca R&D and the Swedish Heart and Lung Foundation. REFERENCES 1. Bengtsson, B.A., et al., Treatment of adults with growth hormone (GH) deficiency with recombinant human GH. J Clin Endocrinol Metab, (2): p Davidson, M.B., Effect of growth hormone on carbohydrate and lipid metabolism. Endocr Rev, (2): p Kanaley, J.A., et al., Acute exposure to GH during exercise stimulates the turnover of free fatty acids in GH-deficient men. J Appl Physiol, (2): p Elam, M.B., et al., Stimulation of in vitro triglyceride synthesis in the rat hepatocyte by growth hormone treatment in vivo. Endocrinology, (4): p Sjöberg, A., et al., Mode of growth hormone administration influences triacylglycerol synthesis and assembly of apolipoprotein B- containing lipoproteins in cultured rat hepatocytes. J Lipid Res, (2): p Frick, F., et al., Interaction between growth hormone and insulin in the regulation of lipoprotein metabolism in the rat. Am J Physiol Endocrinol Metab, (5): p. E Christ, E.R., et al., Effects of growth hormone (GH) replacement therapy on very low density lipoprotein apolipoprotein B100 kinetics in patients with adult GH deficiency: a stable isotope study. J Clin Endocrinol Metab, (1): p Lindén, D., et al., Direct effects of growth hormone on production and secretion of apolipoprotein B from rat hepatocytes [In Process Citation]. Am J Physiol Endocrinol Metab, (6): p. E Améen, C., et al., Effects of gender and growth hormone secretory pattern on sterol regulatory element binding protein-1c and its target genes in rat liver. Am J Physiol Endocrinol Metab, Kliewer, S.A., et al., Fatty acids and eicosanoids regulate gene expression through direct interactions with peroxisome proliferator-activated receptors alpha and gamma. Proc Natl Acad Sci U S A, (9): p Braissant, O., et al., Differential expression of peroxisome proliferator-activated receptors (PPARs): tissue distribution of PPAR-alpha, - beta, and -gamma in the adult rat. Endocrinology, (1): p Mukherjee, R., et al., Human and rat peroxisome proliferator activated receptors (PPARs) demonstrate similar tissue distribution but different responsiveness to PPAR activators. J Steroid Biochem Mol Biol, (3-4): p

91 13. Winegar, D.A., et al., Effects of fenofibrate on lipid parameters in obese rhesus monkeys. J Lipid Res, (10): p Colagiuri, S. and J. Best, Lipid-lowering therapy in people with type 2 diabetes. Curr Opin Lipidol, (6): p Schoonjans, K., B. Staels, and J. Auwerx, Role of the peroxisome proliferator-activated receptor (PPAR) in mediating the effects of fibrates and fatty acids on gene expression. J Lipid Res, (5): p Améen, C., et al., Activation of peroxisome proliferator-activated receptor alpha increases the expression and activity of microsomal triglyceride transfer protein in the liver. J Biol Chem, (2): p Lindén, D., et al., Influence of peroxisome proliferator-activated receptor alpha agonists on the intracellular turnover and secretion of apolipoprotein (Apo) B-100 and ApoB-48. J Biol Chem, (25): p Sugiyama, H., J. Yamada, and T. Suga, Effects of testosterone, hypophysectomy and growth hormone treatment on clofibrate induction of peroxisomal beta-oxidation in female rat liver. Biochem Pharmacol, (5): p Sato, T., et al., Suppression of clofibrateinduction of peroxisomal and microsomal fatty acid-oxidizing enzymes by growth hormone and thyroid hormone in primary cultures of rat hepatocytes. Biochim Biophys Acta, (3): p Yamada, J., et al., Hormonal modulation of peroxisomal enzyme induction caused by peroxisome proliferators: suppression by growth and thyroid hormones in cultured rat hepatocytes. Arch Biochem Biophys, (2): p Sundseth, S.S. and D.J. Waxman, Sexdependent expression and clofibrate inducibility of cytochrome P450 4A fatty acid omega-hydroxylases. Male specificity of liver and kidney CYP4A2 mrna and tissue-specific regulation by growth hormone and testosterone. J Biol Chem, (6): p Yamada, J., et al., Suppressive effect of growth hormone on the expression of peroxisome proliferator-activated receptor in cultured rat hepatocytes. Res Commun Mol Pathol Pharmacol, (1): p Carlsson, L., et al., Effects of fatty acids and growth hormone on liver fatty acid binding protein and PPARalpha in rat liver. Am J Physiol Endocrinol Metab, (4): p. E Jalouli, M., et al., Sex difference in hepatic peroxisome proliferator-activated receptor alpha expression: influence of pituitary and gonadal hormones. Endocrinology, (1): p Zhou, Y.C. and D.J. Waxman, Cross-talk between janus kinase-signal transducer and activator of transcription (JAK-STAT) and peroxisome proliferator-activated receptoralpha (PPARalpha) signaling pathways. Growth hormone inhibition of pparalpha transcriptional activity mediated by stat5b. J Biol Chem, (5): p Zhou, Y.C. and D.J. Waxman, STAT5b downregulates peroxisome proliferator-activated receptor alpha transcription by inhibition of ligand-independent activation function region- 1 trans-activation domain. J Biol Chem, (42): p Lee, S.S., et al., Targeted disruption of the alpha isoform of the peroxisome proliferatoractivated receptor gene in mice results in abolishment of the pleiotropic effects of peroxisome proliferators. Mol Cell Biol, (6): p Lindén, D., et al., PPARalpha deficiency increases secretion and serum levels of apolipoprotein B-containing lipoproteins. J Lipid Res, (11): p Li, X., et al., Method to measure apolipoprotein B-48 and B-100 secretion rates in an individual mouse: evidence for a very rapid turnover of VLDL and preferential removal of B-48- relative to B-100-containing lipoproteins. J Lipid Res, (1): p Tao, R., et al., Human growth hormone increases apo(a) expression in transgenic mice. Arterioscler Thromb Vasc Biol, (10): p Raabe, M., et al., Knockout of the abetalipoproteinemia gene in mice: reduced 11

92 lipoprotein secretion in heterozygotes and embryonic lethality in homozygotes. Proc Natl Acad Sci U S A, (15): p Tietge, U.J., et al., Hepatic overexpression of microsomal triglyceride transfer protein (MTP) results in increased in vivo secretion of VLDL triglycerides and apolipoprotein B. J Lipid Res, (11): p Yu, S., et al., Adipocyte-specific gene expression and adipogenic steatosis in the mouse liver due to peroxisome proliferatoractivated receptor gamma1 (PPARgamma1) overexpression. J Biol Chem, (1): p Gavrilova, O., et al., Liver peroxisome proliferator-activated receptor gamma contributes to hepatic steatosis, triglyceride clearance, and regulation of body fat mass. J Biol Chem, (36): p Chakravarthy, M.V., et al., "New" hepatic fat activates PPARalpha to maintain glucose, lipid, and cholesterol homeostasis. Cell Metab, (5): p Sakuma, T., et al., Different expression of hepatic and renal cytochrome P450s between the streptozotocin-induced diabetic mouse and rat. Xenobiotica, (4): p Enriquez, A., et al., Altered expression of hepatic CYP2E1 and CYP4A in obese, diabetic ob/ob mice, and fa/fa Zucker rats. Biochem Biophys Res Commun, (2): p Leclercq, I.A., et al., CYP2E1 and CYP4A as microsomal catalysts of lipid peroxides in murine nonalcoholic steatohepatitis. J Clin Invest, (8): p Hansen, L.H., et al., Characterization of the inhibitory effect of growth hormone on primary preadipocyte differentiation. Mol Endocrinol, (8): p glucocorticoids. J Clin Invest, (10): p Nakamuta, M., et al., Evaluation of fatty acid metabolism-related gene expression in nonalcoholic fatty liver disease. Int J Mol Med, (4): p Ichikawa, T., et al., Non-alcoholic steatohepatitis and hepatic steatosis in patients with adult onset growth hormone deficiency. Gut, (6): p Matsusue, K., et al., Liver-specific disruption of PPARgamma in leptin-deficient mice improves fatty liver but aggravates diabetic phenotypes. J Clin Invest, (5): p Améen, C. and J. Oscarsson, Sex difference in hepatic microsomal triglyceride transfer protein expression is determined by the growth hormone secretory pattern in the rat. Endocrinology, (9): p Kersten, S., et al., Peroxisome proliferatoractivated receptor alpha mediates the adaptive response to fasting. J Clin Invest, (11): p Frick, F., et al., Long-term growth hormone excess induces marked alterations in lipoprotein metabolism in mice. Am J Physiol Endocrinol Metab, : p. E Egecioglu, E., et al., Growth hormone receptor deficiency results in blunted ghrelin feeding response, obesity, and hypolipidemia in mice. Am J Physiol Endocrinol Metab, (2): p. E Olsson, B., et al., Bovine growth hormonetransgenic mice have major alterations in hepatic expression of metabolic genes. Am J Physiol Endocrinol Metab, (3): p. E Schadinger, S.E., et al., PPARgamma2 regulates lipogenesis and lipid accumulation in steatotic hepatocytes. Am J Physiol Endocrinol Metab, (6): p. E Vidal-Puig, A.J., et al., Peroxisome proliferator-activated receptor gene expression in human tissues. Effects of obesity, weight loss, and regulation by insulin and 12

93 III

94

95 PPARa activation increases triglyceride mass and adipose differentiation-related protein in hepatocytes Ulrika Edvardsson, 1, *,, Anna Ljungberg,*, Daniel Lindén,*, Lena William-Olsson, Helena Peilot-Sjögren, Andrea Ahnmark, and Jan Oscarsson*,, Wallenberg Laboratory for Cardiovascular Research* and Department of Physiology, Sahlgrenska University Hospital, Göteborg, Sweden; and AstraZeneca Research and Development, Mölndal, Sweden Abstract Adipose differentiation-related protein (ADRP) is a lipid droplet-associated protein that is expressed in various tissues. In mice treated with the peroxisome proliferatoractivated receptor a (PPARa) agonist Wy14,643 (Wy), hepatic mrna and protein levels of ADRP as well as hepatic triglyceride content increased. Also in primary mouse hepatocytes, Wy increased ADRP expression and intracellular triglyceride mass. The triglyceride mass increased in spite of unchanged triglyceride biosynthesis and increased palmitic acid oxidation. However, Wy incubation decreased the secretion of newly synthesized triglycerides, whereas apolipoprotein B secretion increased. Thus, decreased availability of triglycerides for VLDL assembly could help to explain the cellular accumulation of triglycerides after Wy treatment. We hypothesized that this effect could be mediated by increased ADRP expression. Similar to PPARa activation, adenovirusmediated ADRP overexpression in mouse hepatocytes enhanced cellular triglyceride mass and decreased the secretion of newly synthesized triglycerides. In ADRP-overexpressing cells, Wy incubation resulted in a further decrease in triglyceride secretion. This effect of Wy was not attributable to decreased cellular triglycerides after increased fatty acid oxidation because the triglyceride mass in Wy-treated ADRPoverexpressing cells was unchanged. In summary, PPARa activation prevents the availability of triglycerides for VLDL assembly and increases hepatic triglyceride content in part by increasing the expression of ADRP. Edvardsson, U., A. Ljungberg, D. Lindén, L. William-Olsson, H. Peilot-Sjögren, A. Ahnmark, and J. Oscarsson. PPARa activation increases triglyceride mass and adipose differentiation-related protein in hepatocytes. J. Lipid Res : Supplementary key words Wy14,643. primary hepatocytes. triglyceride synthesis. fatty acid oxidation. triglyceride secretion. apolipoprotein B-100. apolipoprotein B-48. peroxisome proliferator-activated receptor a Peroxisome proliferator-activated receptor a (PPARa) is a ligand-activated transcription factor that plays a key role in the regulation of genes involved in carbohydrate, lipid, and lipoprotein metabolism (for review, see 1). Manuscript received 19 May 2005 and in revised form 11 October 2005 and in re-revised form 10 November Published, JLR Papers in Press, November 10, DOI /jlr.M JLR200 PPARa is highly expressed in tissues with high mitochondrial and peroxisomal b-oxidation activities, such as liver, heart, kidney, and skeletal muscle (2 5). In humans, treatment with PPARa agonists (i.e., fibrates) results in decreased plasma levels of triglycerides and increased plasma HDL cholesterol levels (6, 7). The triglyceride-lowering effect of fibrates is partly explained by increased lipoprotein lipase expression (8) and downregulation of hepatic apolipoprotein C-III expression (9, 10), which results in increased turnover of VLDL. In addition, fibrates have been shown to decrease the plasma concentration of atherogenic small dense LDL particles (11, 12), indicating decreased hepatic triglyceride secretion, because small dense LDLs are products of triglyceride-rich VLDL (VLDL 1 ) particles (13). Indeed, fibrates have been shown to decrease VLDL triglyceride secretion in both humans and rats (14, 15). In primary rat hepatocytes, fibrates reduced triglyceride secretion and decreased the size of the secreted apolipoprotein B (apob)-containing lipoproteins (16). The reduced triglyceride secretion may be explained by decreased triglyceride biosynthesis in rat hepatocytes (16, 17). However, in vivo in rats, incorporation of palmitate into liver triglycerides was unchanged, whereas triglyceride secretion decreased (15), indicating that other effects of PPARa activation than decreased triglyceride biosynthesis gave rise to the decreased triglyceride secretion. One suggested possibility is that PPARa activation increases diacylglycerol acyltransferase (DGAT) activity in the cytoplasm while inhibiting DGAT activity in the microsomal compartment, thereby diverting triglycerides away from the secretory pathway without influencing total cellular triglyceride synthesis (18). Adipose differentiation-related protein (ADRP) is a lipid storage droplet-associated protein belonging to the PAT (Perilipin, ADRP, and TIP 47) family (19). ADRP was first Abbreviations: ADRP, adipose differentiation-related protein; apob, apolipoprotein B; DGAT, diacylglycerol acyltransferase; MTP, microsomal triglyceride transfer protein; PPARa, peroxisome proliferatoractivated receptor a; PPRE, peroxisome proliferator-activated receptor response element; Wy, Wy14,643; 36B4, acidic ribosomal phosphoprotein P0. 1 To whom correspondence should be addressed. ulrika.edvardsson@astrazeneca.com Copyright D 2006 by the American Society for Biochemistry and Molecular Biology, Inc. This article is available online at Journal of Lipid Research Volume 47,

96 identified as an early marker of adipocyte differentiation (20). However, studies have shown that ADRP is expressed in a variety of tissues and cultured cells (21, 22). It has been suggested to be a marker of lipid accumulation, as the cellular protein level of ADRP is related to the total mass of neutral lipids within the cell (21, 23). A few studies have explored the function of ADRP. Overexpression of ADRP in fibroblasts resulted in lipid accumulation and lipid droplet formation without induction of adipogenic genes (24). Also in macrophages, ADRP overexpression gave rise to lipid accumulation without changed expression of the genes involved in lipid efflux (25). Furthermore, transfection of COS-7 cells with ADRP was shown to promote the uptake of long-chain fatty acids (26). Together, these results indicate that ADRP increases intracellular lipids without changing the expression of genes involved in lipid metabolism, but the mechanisms of this effect are still unclear. ADRP expression has been reported to be regulated by long-chain fatty acids at the transcriptional level (27). In macrophages and in colorectal cancer cells, ADRP expression was induced by agonists of PPAR subtypes a, g, and y (28 30). Peroxisome proliferator-activated receptor response elements (PPREs) were recently identified both in the murine and human ADRP promoters (31, 32), showing that PPAR agonists regulate ADRP expression by increasing transcription of the gene. However, the regulation of ADRP by PPARa in liver and hepatocytes has not been investigated. The aims of this study were to investigate the effects of the specific PPARa agonist Wy14,643 (Wy) (33) on the expression of ADRP in mouse liver in vivo and in primary mouse hepatocytes in vitro and to determine the importance of changed ADRP expression for the effects of PPARa activation on triglyceride secretion and intracellular triglyceride accumulation. MATERIALS AND METHODS Animals and treatment C57BL/6 mice were from Taconic Europe (Ry, Denmark). Homozygous PPARa null mice and corresponding wild-type control mice, on a pure Sv/129 genetic background, were used for in vitro experiments (kindly provided by Dr. F. J. Gonzalez, National Institutes of Health, Bethesda, MD) (34). PPARa null mice and littermate controls, backcrossed for two generations with C57BL/ 6, were used for in vivo experiments. The animals were housed individually and maintained under standardized conditions of temperature (21 22jC) and humidity (40 60%), with light from 6:00 AM to 6:00 PM for at least 1 week before the experiments. The mice were given either standard laboratory chow containing (energy %) 12% fat, 62% carbohydrates, and 26% protein, with a total energy content of 12.6 kj/g (R3; Lactamin AB, Kimstad, Sweden), or a high-fat diet containing 48% fat (mainly saturated), 15% protein, and 37% carbohydrates, with a total energy content of 21.4 kj/g (Lactamin). The mice were fed laboratory chow or high-fat diet for 3 weeks and treated with Wy (30 Amol/kg/day; Chemsyn Science Laboratories, Lenaxa, KS) in 0.5% (w/v) methyl cellulose by gavage once daily for the last 2 weeks. Age-matched control mice received only vehicle. Food intake was estimated by weighing the food twice weekly. The mice were anesthetized with isoflurane (Forene; Abbot Scandinavia AB), and the livers were removed, immediately frozen in liquid nitrogen, and stored at 2 70jC. The study protocol was approved by the Ethics Committee of Göteborg University. All experiments were conducted in accordance with accepted standards of humane animal care. Liver triglycerides Frozen livers were homogenized in isopropanol (1 ml/50 mg tissue) and incubated at 4jC for 1 h. The samples were centrifuged at 4jC for 5 min at 2,500 rpm, and triglyceride concentrations in the supernatants were measured with an enzymatic colorimetric assay (Roche, Mannheim, Germany). Production of recombinant adenoviruses A pbluescript SK(+) vector containing full-length ADRP was kindly provided by Björn Magnusson (Wallenberg Laboratory for Cardiovascular Research, Sahlgrenska University Hospital, Göteborg, Sweden). The ADRP construct was then transferred to a pentr vector (Invitrogen, Carlsbad, CA) and recombined into pad/cmv/v5-dest (Invitrogen) according to the manufacturer s manual. The packaging cell line Ad-293 (Stratagene, La Jolla, CA) was grown in Dulbecco s modified Eagle s medium with 10% FBS supplemented with 100,000 IU/l penicillin and 100 mg/l streptomycin. Cells were seeded in 25 cm 2 culture flasks, cultured to 90 95% confluence, and transfected with adenoviral constructs for ADRP or the control zsgreen (35) digested with Pac-1 using Lipofectamine 2000 in Opti-MEM according to the manufacturer s manual. After 4 h of incubation, DMEM containing 20% FBS was added, resulting in a concentration of 10% FBS. Cells were cultured for days until cytopathic effects were z80%. Viruses were harvested by repeated freeze/ thaw cycles in 10 mm Tris-HCl, ph 8.0. After large-scale amplification using Cell Factories (Nunc, Rochester, NY), recombinant adenoviruses were purified by two rounds of CsCl density gradient ultracentrifugation. The purified virus stocks were desalted over 10DG columns (Bio-Rad, Hercules, CA) and eluted in sterile PBS. Glycerol (65%) was added (1:5 dilution with virus suspension) before the virus stocks were divided into aliquots and stored at 280jC until use. Infectious viral titers were determined using the Adeno-X Rapid Titer Kit (Clontech, Palo Alto, CA). All purified virus stocks were screened for possible wild-type virus contamination according to Zhang, Koch, and Roth (36) before use. Primary hepatocyte cultures Mouse hepatocytes were obtained by nonrecirculating collagenase perfusion through the portal vein of the mice (10 16 weeks of age) as described (16, 37, 38). In brief, the cells were seeded at 100,000 cells/cm 2 in dishes (Falcon, Plymouth, UK) coated with laminin-rich matrigel (BD Biosciences, Bedford, MA). The cells were cultured during the first h in Williams E medium with Glutamax (Invitrogen) supplemented as described (37). The cells were then treated for up to 3 days with 1 or 10 AM Wy (Chemsyn Science Laboratories) dissolved in DMSO [final concentration, 0.15% (v/v)] in medium supplemented as described above plus 1 nm dexamethasone (Sigma, St. Louis, MO) and 3 nm insulin (Actrapid; Novo Nordisk A/S). In experiments with adenoviral overexpression of ADRP or the control zsgreen, cells were infected with virus (500 infectious units/cell) in 0.75 ml of culture medium (10 cm 2 dish) starting 4 h after seeding. Two hours later, medium was added to a final volume of 2 ml and infection was continued overnight. After 17 h of infection with 330 Journal of Lipid Research Volume 47, 2006

97 virus, the medium was replaced with virus-free medium. Analyses were performed 3 days after infection. Quantitative real-time PCR analysis Total RNA of cultured primary hepatocytes and mouse liver was isolated with TriReagentk (Sigma) according to the manufacturer s protocol, and the concentration of RNA was determined spectrophotometrically at 260 nm. To remove contaminating DNA, total RNA was treated with DNA free (Ambion, Austin, TX) before being retrotranscribed using TaqMan Reverse Transcription Reagents (Applied Biosystems, Foster City, CA). Quantitative real-time PCR analysis was performed on the ABI Prism 7700 Sequence Detection System (Applied Biosystems) using SYBR Green detection chemistry. All samples were analyzed in triplicate. To exclude that the amplification-associated fluorescence was associated with residual genomic DNA and/or from the formation of primer dimers, controls without reverse transcriptase or DNA template were analyzed. RT-PCR products were also analyzed by electrophoresis in ethidium bromide-stained agarose gels to check that a single amplicon of the expected size was obtained. The expression data were normalized to the endogenous control acidic ribosomal phosphoprotein P0 (36B4). The expression of 36B4 was not influenced by the various treatments in this study. The relative expression levels were calculated according to the formula 2 2DCt, where DCt is the difference in threshold cycle (Ct) values between the target and the 36B4 endogenous control. Specific primers for each gene (Table 1) were designed using Primer Express software (Applied Biosystems). Protein preparation and Western blot Matrigel was removed from cultured primary hepatocytes by incubation on ice for 60 min in PBS containing 5 mm EDTA followed by washings in PBS. Total protein extracts from frozen livers and cultured hepatocytes were prepared as described previously (39), and protein concentrations were determined with the RC/DC Protein Assay Kit II (Bio-Rad). Proteins were separated on 10 20% Tris-glycine gels (Invitrogen) and transferred to Hybond-P polyvinylidene difluoride transfer membrane (Amersham Biosciences, Bucks, UK). Equal loading was confirmed by staining the membranes with 0.2% Ponceau S (Serva, Heidelberg, Germany). Immunoblotting was performed using a guinea pig polyclonal anti-adrp antibody at 1:2,000 (Research Diagnostics, Inc., Flanders, NJ) and a horseradish peroxidase-conjugated antiguinea pig antibody at 1:20,000 (Dako, Glostrup, Denmark), followed by detection using the enhanced chemiluminescence plus detection system (Amersham Biosciences). Band intensity was quantified with ImageQuant software (Molecular Dynamics, Sunnyvale, CA). Palmitic acid oxidation Williams E medium with Glutamax, supplemented as described above, containing [9,10(n)- 3 H]palmitic acid was prepared as described by Leung and Ho (40). To each culture dish (10 cm 2 ), 1 ml of medium containing 110 AM unlabeled palmitic acid and 8.3 ACi of [9,10(n)- 3 H]palmitic acid (specific activity, 54 Ci/mmol) was added. The fatty acid oxidation was shown to be linear between 30 and 120 min of incubation at 37jC (data not shown). Fatty acid oxidation was thereafter determined after 60 or 90 min of incubation with labeled palmitic acid. Labeled water-soluble products were isolated and analyzed as described previously (35) but with one additional precipitation step. Background radioactivity was determined by precipitation of fatty acids in medium that had not been in contact with cells. The fatty acid oxidation was related to the DNA content in each culture dish, which was determined according to Labarca and Paigen (41). Triglyceride biosynthesis and accumulation of triglycerides in cell medium Triglyceride biosynthesis in cultured hepatocytes was estimated by measurement of incorporated [9,10(n)- 3 H]palmitic acid (concentration as described above) in cellular triglycerides after min of incubation at 37jC. Accumulation of newly synthesized triglycerides in the medium was determined after 2 6 h of incubation with [9,10(n)- 3 H]palmitic acid at 37jC. Cells and medium were then collected and lipid extraction was performed according to Bligh and Dyer (42). Lipids were separated by thinlayer chromatography (silica gel 60 on plastic sheets; Merck, Darmstadt, Germany) with chloroform-acetic acid (96:4). The bands corresponding to triglycerides were recovered and extracted from the silica gel with 1 ml of cyclohexane followed by the addition of 10 ml of scintillation solution (Ready Safek ; Beckman Coulter, Fullerton, CA) before the radioactivity was measured. Triglyceride synthesis and the accumulation of newly synthesized triglycerides in the cell culture medium were related to the DNA content in each culture dish as described above. Estimation of apob secretion The secretion of apob-48 and apob-100 from primary mouse hepatocyte cultures was estimated by labeling the cells with a [ 35 S]methionine-cysteine mix (Amersham Biosciences) for 2 h followed by a 4 h chase in culture medium supplemented with 10 mm methionine, as described (38, 43, 44). Labeled apob-48 and apob-100 were isolated by immunoprecipitation with 10 Al of polyclonal rabbit anti-human apob antibodies (DakoCytomation, Glostrup, Denmark), followed by 5% polyacrylamide gel electrophoresis containing SDS. The bands corresponding to apob-48 and apob-100 were quantified using a FLA-3000 phosphorimager (Fujifilm). The densities were related to the total amount of DNA in each culture dish, as described above. Triglyceride mass in hepatocytes Intracellular triglyceride content in hepatocytes was determined by HPLC separation of neutral lipids as described (45). Briefly, after extraction with Folch reagent (46) and evaporation, TABLE 1. Primers used for quantitative real-time PCR Gene Forward Primer (5V 3V) Reverse Primer (5V 3V) ADRP TGGCAGCAGCAGTAGTGGAT CAGGTTGGCCACTCTCATCA CPT-I TGAGTGGCGTCCTCTTTGG CAGCGAGTAGCGCATAGTCATG LCAD GCGAAATACTGGGCATCTGAA TCCGTGGAGTTGCACACATT ACO CAGCAGGAGAAATGGATGCA GGGCGTAGGTGCCAATTATCT 36B4 GAGGAATCAGATGAGGATATGGGA AAGCAGGCTGACTTGGTTGC ACO, acyl-coenzyme A oxidase; ADRP, adipose differentiation-related protein; CPT-I, carnitine palmitoyltransferase I; LCAD, long-chain acyl-coenzyme A dehydrogenase; 36B4, acidic ribosomal phosphoprotein P0. PPARa and ADRP in mouse liver 331

98 lipids were dissolved in hexane-isopropanol-acetic acid (98.7:1.2:0.1) and separated by HPLC in an isocratic system with 40% hexane (0.6 ml/min). Lipids were detected using a light-scattering detector (PL-ELS 1000; Polymer Laboratories, Shropshire, UK), and the amount of triglycerides was quantified using the standard HPLC Mix 42 (Larodan, Malmö, Sweden). The amount of triglycerides was related to the total amount of DNA in each culture dish as described above. Statistics Values are expressed as means 6 SEM. Comparisons between groups were made by Kruskal-Wallis test and Mann-Whitney U-test. P, 0.05 was considered statistically significant. RESULTS Effects of Wy on ADRP expression and liver triglycerides in vivo C57BL/6 mice were fed ordinary chow or high-fat diet for 3 weeks and treated with Wy (30 Amol/kg/day) by gavage during the last 2 weeks. Body weight gain during the treatment did not differ between the groups. However, the food consumption decreased, whereas the energy intake increased in the groups fed the high-fat diet (Table 2). Irrespective of diet, Wy treatment resulted in increased hepatic ADRP mrna expression (Fig. 1A). ADRP protein expression increased by 2-fold with Wy treatment in mice on the chow diet, whereas Wy had no significant effect on ADRP protein expression in mice fed the high-fat diet (Fig. 1B). The high-fat diet per se increased ADRP protein expression, whereas mrna expression was unaffected. Wy treatment increased hepatic triglyceride concentration by 33% in chow-fed mice, and to a lesser degree (15%) in mice fed the high-fat diet (Fig. 1C). When taking into account the increased liver weight as a result of Wy treatment, the total triglyceride content increased significantly in the chow-fed group by 42%, and by 64% in the high-fat diet group (data not shown). The experiment in which the mice were on the ordinary chow diet was repeated. Also in this experiment, Wy significantly increased ADRP mrna (500%) and protein (90%) expression as well as hepatic triglyceride concentration (54%) (data not shown). TABLE 2. Treatment and Diet Food consumption and body weight gain Food Consumption Energy Intake Body Weight Gain g/day kj/day g Vehicle chow Wy chow Vehicle high-fat diet a a Wy high-fat diet a a Wy, Wy14,643. C57BL/6 mice were fed the chow or high-fat diet for 3 weeks and were treated with vehicle or Wy (30 Amol/kg/day) for the last 2 weeks. Values are means 6 SEM (n 5 7). a P, 0.05, vehicle chow versus vehicle high-fat diet or Wy chow versus Wy high-fat diet (Kruskal-Wallis test followed by Mann-Whitney U test). Fig. 1. Effects of Wy14,643 (Wy) treatment on adipose differentiation-related protein (ADRP) mrna expression (A), protein expression (B), and hepatic triglyceride concentration (C) in vivo. C57BL/6 mice were fed the chow or high-fat diet for 3 weeks and treated with Wy (30 Amol/kg/day) or vehicle (Veh) for the last 2 weeks. ADRP mrna and protein levels were estimated by quantitative real-time PCR and Western blotting, respectively. Liver triglyceride concentration was estimated as described in Materials and Methods. The Western blot in B shows two representative individuals per group. 36B4, acidic ribosomal phosphoprotein P0. Values are means 6 SEM based on six (A), four (B), or seven (C) observations. *P, 0.05, vehicle chow versus Wy chow or vehicle high-fat diet versus Wy high-fat diet; # P, 0.05, vehicle chow versus vehicle high-fat diet (Kruskal-Wallis test followed by Mann- Whitney U-test). 332 Journal of Lipid Research Volume 47, 2006

99 Effects of Wy on ADRP expression in PPARa null mice To investigate whether the effect of Wy was mediated via PPARa activation, littermate wild-type and PPARa null mice were fed a high-fat diet for 3 weeks and treated the last 2 weeks with Wy (as described above). Food consumption and body weight gain during the treatment did not differ between the groups (Table 3). The effect of Wy on ADRP mrna and protein expression was dependent on PPARa, as shown in Fig. 2. ADRP mrna expression was lower, whereas ADRP protein levels were higher in the PPARa null mice than in littermate controls. The hepatic triglyceride concentration was not significantly influenced by Wy treatment or PPARa deficiency in this experiment. Effects of Wy on ADRP expression in primary mouse hepatocytes To determine whether the stimulatory effect on ADRP expression was a direct effect on hepatocytes, we incubated primary mouse hepatocytes from C57BL/6 mice with 10 AM Wy for 3 days. Wy increased the ADRP mrna expression (Fig. 3A) and protein expression (Fig. 3B) in parallel with increased cellular triglyceride mass (Fig. 3C). Experiments were also performed on hepatocytes derived from PPARa null mice and wild-type mice on a pure Sv/ 129 genetic background. Incubation with 10 AM Wy for 3 days increased ADRP mrna expression 200% in wildtype cells but had no effect in cells from PPARa null mice (data not shown). TABLE 3. Food consumption and body weight gain Treatment and Genotype Food Consumption Body Weight Gain g/day g Vehicle wild type Wy wild type Vehicle PPARa null Wy PPARa null PPARa, peroxisome proliferator-activated receptor a. Littermate wild-type and PPARa null mice were fed a high-fat diet for 3 weeks and treated with vehicle or Wy (30 Amol/kg/day) for the last 2 weeks. Values are means 6 SEM (n 5 7). No significant differences were found between the groups (Kruskal-Wallis test). Fig. 2. Effects of Wy treatment on ADRP mrna expression (A) and protein expression (B) in peroxisome proliferator-activated receptor a (PPARa) null mice in vivo. PPARa null mice and wildtype littermates were fed a high-fat diet for 3 weeks and treated with Wy (30 Amol/kg/day) or vehicle (Veh) for the last 2 weeks. ADRP mrna and protein levels were estimated by quantitative realtime PCR and Western blotting, respectively. Values are means 6 SEM based on seven (A) or three (B) observations. *P, 0.05, vehicle wild type versus Wy wild type; # P, 0.05, vehicle wild type versus vehicle PPARa null (Kruskal-Wallis test followed by Mann- Whitney U-test). The statistical analysis of protein expression (B) showed a significant difference between the groups, P, 0.05 (Kruskal-Wallis test; but the Mann-Whitney U-test showed no significant differences between individual groups). Effects of Wy on palmitic acid oxidation The increased triglyceride mass in hepatocytes after Wy treatment was unexpected, as PPARa activation is known to increase the oxidation of fatty acids. Therefore, we characterized the effect of Wy on metabolic events that influence the amount of triglycerides in hepatocytes. First, we measured the oxidation of palmitic acid in mouse hepatocytes after 3 days of exposure to 1 or 10 AM Wy. As expected, the fatty acid oxidation was enhanced by Wy, resulting in a 4-fold increase when cells were incubated with 10 AM Wy(Table 4). The enhanced fatty acid oxidation was paralleled by increased mrna levels of carnitine palmitoyltransferase I and long-chain acylcoenzyme A dehydrogenase (Table 4), enzymes involved in the mitochondrial b-oxidation of palmitic acid. We also measured the mrna expression of acyl-coenzyme A oxidase, which participates in the peroxisomal fatty acid oxidation and is known to be PPAR-responsive (47). Wy (10 AM) increased the expression of acyl-coenzyme A oxidase by 14-fold, as shown in Table 4. Thus, these cells responded well to Wy incubation in terms of fatty acid oxidation and gene expression. Effects of Wy on triglyceride biosynthesis and the accumulation of triglycerides in the medium To further characterize the effects of Wy on lipid metabolism in primary mouse hepatocytes, we estimated triglyceride biosynthesis using [ 3 H]palmitic acid as a tracer. Three days of exposure to Wy did not affect the PPARa and ADRP in mouse liver 333

100 TABLE 4. Effect of Wy on fatty acid oxidation and gene expression Fatty Acid or Gene Control Wy (1 AM) Wy (10 AM) Palmitic acid oxidation (n 5 4) CPT-I expression (n 5 4) LCAD expression (n 5 5) ACO expression (n 5 3) a a a a a a a a Mouse hepatocytes were isolated and cultured with Wy for 72 h before a 90 min incubation in the presence of [9,10(n)- 3 H]palmitic acid. mrna expression of CPT-I, LCAD, and ACO was estimated by quantitative real-time PCR, and the expression data were normalized to 36B4. Values are means 6 SEM (% of control) of three to five independent liver perfusions (as indicated) with one or two culture dishes per group. a P, 0.05, control versus Wy (Kruskal-Wallis test, followed by Mann- Whitney U test). triglyceride biosynthesis, estimated as palmitic acid incorporation into cellular triglycerides during 90 min (data not shown). Because Wy increased the intracellular triglyceride content despite an increased fatty acid oxidation and unchanged triglyceride biosynthesis, we also investigated whether PPARa activation influenced the partitioning of newly synthesized triglycerides between the cellular and extracellular compartments. Figure 4A shows data from a representative experiment, which illustrates the distribution of newly synthesized triglycerides in cells and in the medium after 2 6 h of incubation with [ 3 H]palmitic acid. In control cells, 30% of the newly synthesized triglycerides were recovered in the medium after 6 h of incubation, compared with 12% in the Wy-incubated cells. As shown in Fig. 4B, Wy decreased the accumulation of newly synthesized triglycerides in the medium by 50%, whereas no change in intracellular [ 3 H]triglycerides was detected after 6 h of incubation. Therefore, we conclude that PPARa activation results in decreased availability of newly synthesized triglycerides for VLDL assembly, although the cellular triglyceride mass increases. Fig. 3. Effects of Wy on ADRP mrna expression (A), protein expression (B), and intracellular triglyceride mass (C) in primary mouse hepatocytes. Hepatocytes were isolated from C57BL/6 mice by liver perfusion and incubated with 10 AM Wy for 72 h. ADRP mrna and protein expressions were quantified by quantitative real-time PCR and Western blotting, respectively. Triglyceride mass was determined by HPLC separation of neutral lipids as described in Materials and Methods. Data are based on measurements of 2 2DCt (A) and Ag triglyceride/ag DNA (C) and presented as percentages of control (Ctrl). Values are means 6 SEM based on five (A) and three (B, C) independent liver perfusions with one to two (A, B) or three (C) culture dishes in each group. *P, 0.05, Mann- Whitney U-test. Effects of Wy on the secretion of apob-containing lipoprotein particles Wy incubation of primary rat hepatocytes has been shown to result in increased apob-100 secretion, whereas apob-48 secretion was unaffected (16, 48). Because it is not known whether PPARa activation also influences apob secretion from mouse hepatocytes, the effect of Wy on the secretion of apob-100 and apob-48 from primary mouse hepatocytes was investigated. As shown in Fig. 5, Wy incubation increased apob-100 and apob-48 secretion by 2.5-fold and 2-fold, respectively. Because triglyceride secretion was reduced, it can be concluded that Wy incubation of mouse hepatocytes gives to the secretion of an increased number of triglyceride-poor apob-containing lipoprotein particles. Effects of ADRP overexpression in mouse hepatocytes Because ADRP overexpression in fibroblasts and in macrophages has been shown to stimulate lipid accumulation (24, 25), we hypothesized that the increase in ADRP ex- 334 Journal of Lipid Research Volume 47, 2006

101 Effects of ADRP overexpression in the presence of Wy To further explore the importance of ADRP for the effects of Wy, ADRP-overexpressing mouse hepatocytes were incubated with Wy. As shown in Fig. 7A, Wy did not further increase the ADRP protein expression in ADRPoverexpressing cells. Nevertheless, Wy decreased the accumulation of triglycerides in the medium (Fig. 7B). ADRP overexpression resulted in a slightly decreased palmitic acid oxidation (210%) (Fig. 7C). However, also in ADRPoverexpressing cells, Wy increased the oxidation of palmitic acid (Fig. 7C). This result indicated that Wy might decrease the secretion of newly synthesized triglycerides by increasing the oxidation of fatty acids in ADRPoverexpressing cells. Therefore, we also measured the intracellular triglyceride mass (Fig. 7D). In spite of increased fatty acid oxidation, Wy incubation tended to further increase the triglyceride mass in ADRP-overexpressing cells (P ). This finding showed that the increased fatty acid oxidation after Wy incubation did not result in a lack of cellular triglycerides for VLDL secretion. Thus, increased ADRP expression in combination with other effects of PPARa activation are responsible for the decreased availability of the cytosolic triglycerides for VLDL assembly. DISCUSSION Fig. 4. Effect of Wy on the distribution of newly synthesized triglycerides (TG) between the cells and culture medium. Mouse hepatocytes were isolated from C57BL/6 mice by liver perfusion and treated with Wy for 72 h. Triglycerides were labeled using [ 3 H]palmitic acid as described in Materials and Methods. A: Data from one representative experiment show the distribution of newly synthesized triglycerides between the cell culture medium (dashed lines) and the cells (solid lines) after 2, 4, and 6 h of incubation with [ 3 H]palmitic acid. B, C: Recovery of newly synthesized triglycerides in the cell culture medium (B) and intracellularly (C) after 6 h of incubation with [ 3 H]palmitic acid. Data are based on measurements of dpm/ag DNA and presented as percentages of control (Ctrl). Values are means 6 SEM based on three independent liver perfusions with two culture dishes in each group. *P 0.05, Mann-Whitney U-test. pression after Wy treatment might be responsible for the increased cellular triglyceride content and decreased triglyceride secretion. Adenovirus-mediated overexpression of ADRP in primary hepatocytes resulted in a 2-fold increase in ADRP protein expression, as shown in Fig. 6A.To estimate the effect of ADRP overexpression on triglyceride biosynthesis, cells were incubated with [ 3 H]palmitic acid for 60 min. ADRP overexpression resulted in a slight increase in triglyceride biosynthesis (Fig. 6B). However, ADRP overexpression reduced the secretion of newly synthesized triglycerides by 50% (Fig. 6C), an effect that was paralleled by increased intracellular accumulation of newly synthesized triglycerides (Fig. 6D). Thus, these findings support the hypothesis that the increased ADRP expression by Wy may contribute to increased cellular triglyceride mass and decreased triglyceride secretion. In this study, we extend previous findings of a regulation of ADRP by PPARa agonists (29, 30, 49) by showing that ADRP is regulated by a PPARa agonist also in the liver and in hepatocytes. The regulation of ADRP was at the level of mrna, with a similar change in mrna and protein levels both in vitro and in vivo. Interestingly, this regulation was shown to be accompanied by an increased cellular accumulation of triglycerides in vivo and in vitro. Because no change in triglyceride synthesis was observed, and as expected, the fatty acid oxidation was markedly enhanced by PPARa activation, we concluded that the decreased availability of triglycerides for VLDL assembly contributed to the increased intracellular content of triglycerides, an effect that could be mediated by increased ADRP expression. Therefore, we investigated the effect of ADRP overexpression and found decreased secretion of newly synthesized triglycerides from ADRP-overexpressing cells, although triglyceride synthesis increased. Thus, increased ADRP expression could contribute to the decreased secretion of newly synthesized triglycerides after Wy incubation. However, in cells overexpressing ADRP, incubation with Wy further decreased triglyceride secretion, indicating that increased ADRP expression is not the sole mechanism responsible for the decreased hepatic triglyceride secretion. Few studies have addressed the effect of PPARa activation on the triglyceride content of hepatocytes. Ten days of fenofibrate treatment of rats resulted in a 50% increase in liver triglycerides (18). Moreover, it has been shown that incubation of primary rat hepatocytes with bezafibrate for 48 h resulted in increased cellular triglyceride mass (50). Because bezafibrate has been shown to activate the murine PPARa and PPARg promoters with similar potency (33), PPARa and ADRP in mouse liver 335

102 Fig. 5. Effect of Wy on apolipoprotein B-100 (apob-100) and apob-48 secretion. Mouse hepatocytes were isolated from C57BL/6 mice by liver perfusion and treated with Wy for 72 h. ApoB-100 and apob-48 secretion was determined by labeling the cells with [ 35 S]methionine-cysteine mix for 2 h followed by a 4 h chase with an excess of cold methionine and immunoprecipitation of the culture medium as described in Materials and Methods. A: Autoradiogram from one representative experiment demonstrating the effect of Wy on radiolabeled apob-48 and apob-100 in the medium from three culture dishes in each group. B, C: Data are based on band densities/ag DNA, and apob-100 secretion (B) and apob-48 secretion (C) are presented as percentages of controls (Ctrl). Values are means 6 SEM based on four independent liver perfusions with three culture dishes in each group. *P, 0.05, Mann-Whitney U-test. the relative importance of PPARa and PPARg activation for the increased triglyceride mass cannot be determined. In contrast to these findings, there are studies in lipoatrophic mice and mice fed a methionine- and cholinedeficient diet demonstrating that treatment with fibrates alleviates hepatic steatosis (51, 52). However, these animals have a severe liver steatosis that may explain the different effect of Wy treatment. In addition, the mice were given a markedly higher dose of Wy than the dose used in our study (400 vs. 30 Amol/kg/day). This high dose of Wy resulted in a marked decrease in body weight gain in one of the studies (52), probably because of decreased food intake. Thus, the decreased body weight gain may explain the decreased triglyceride content of the liver. The lower dose of Wy used in the present study results in decreased plasma triglycerides and apob levels without influencing food intake or body weight gain (48). Thus, the increased liver content of triglycerides occurs when a therapeutic dose of a PPARa agonist is given. The hepatic expression of ADRP might also increase in human subjects upon PPARa activation, because a PPRE has been demonstrated in the human ADRP (adipophilin) promoter (32). To the best of our knowledge, Wy has not been used in published clinical trials. Therefore, it is difficult to relate the present findings to the effects of fibrates used for the treatment of patients. However, the published maximal plasma concentrations for the commonly used fibrates are in the same range (fenofibrate, z40 Amol/l; gemfibrozil, z80 Amol/l) as the concentrations of Wy used in the in vitro experiments (53). The clinical importance of increased ADRP expression for the liver concentration of triglycerides in human subjects after PPARa activation awaits further studies. From our results, it can be concluded that the hepatic triglyceride concentration increased less by Wy treatment when the animals were on a high-fat diet. The reason for this is probably that the high-fat diet induced ADRP protein expression. Interestingly, increased mrna levels did not parallel the increased ADRP protein expression induced by a high-fat diet. It has been shown that trans-10, cis-12-conjugated linoleic acid induces ADRP protein to a much greater extent than ADRP mrna in adipocytes (54). Trans-10, cis-12-conjugated linoleic acid was suggested to increase the translation of ADRP via increased mtor/ p70s6k/s6 signaling. In addition, oleic acid has been shown to stabilize the ADRP protein in Chinese hamster ovary cells by inhibition of proteasomal degradation (55). Thus, PPARa activation and dietary fat seem to regulate ADRP expression via different mechanisms. Increased ADRP expression also results in increased cellular accumulation of triglycerides in cells that do not secrete lipoproteins (24, 25). Therefore, the decreased secretion of triglyceride-rich lipoproteins is not a prerequisite for the intracellular accumulation of triglycerides as a result of overexpression of ADRP. We showed that increased ADRP expression decreases fatty acid oxidation and increases net incorporation of palmitic acid into triglycerides. Thus, ADRP seems to compartmentalize fatty acids toward glycerolipid synthesis and away from oxidation, which could contribute to the increased accumulation of triglycerides in the cells. Therefore, the finding that Wy did not affect triglyceride synthesis in isolated hepatocytes might be the result of other effects of PPARa activation, such as increased fatty acid oxidation, that may counteract the effect of increased ADRP expression. 336 Journal of Lipid Research Volume 47, 2006

103 Fig. 6. Effect of ADRP overexpression on ADRP protein expression (A), triglyceride (TG) biosynthesis (B), and distribution of newly synthesized triglycerides to the cell culture medium (C) and intracellularly (D). Mouse hepatocytes were isolated from C57BL/6 mice by liver perfusion and transduced with adenoviruses expressing zsgreen or ADRP 4 h after seeding. Analyses were performed after 3 days of culture in control medium. A: ADRP protein expression was estimated by Western blotting. B: Triglyceride biosynthesis was determined by incubation with [ 3 H]palmitic acid for 60 min. C, D: Distribution of newly synthesized triglycerides in cell culture medium (C) and intracellularly (D) after 3 and 6 h of incubation with [ 3 H]palmitic acid. Data are based on measurements of dpm/ag DNA (B D) and presented as percentages of zsgreen. Values are means 6 SEM based on three independent liver perfusions with two culture dishes in each group. *P, 0.05, zsgreen versus ADRP at each time point [Mann-Whitney U-test (A, B) and Kruskal- Wallis test followed by Mann-Whitney U-test (C, D)]. ADRP overexpression has also been shown to increase fatty acid uptake in COS-7 cells (26). Interestingly, PPARa activation also results in changed expression of other genes that may take part in the increased uptake of fatty acids, such as acyl-coa synthase, fatty acid transport protein (56), and liver fatty acid binding protein (57, 58). Thus, ADRP may increase cellular triglyceride accumulation in hepatocytes by increasing fatty acid uptake (26), diverting fatty acids to triglyceride formation, and preventing the use of triglycerides for VLDL assembly. In a previous study using primary rat hepatocytes, we observed that Wy decreased both triglyceride synthesis and secretion (16). Thus, PPARa agonists seem to have different effects on triglyceride synthesis in cultured mouse and rat hepatocytes, although secretion of newly synthesized triglycerides decreases as a result of PPARa activation in both species. In this study, we showed that decreased triglyceride synthesis is not a prerequisite for the decreased availability of triglycerides for VLDL assembly. PPARa activation increased the expression of microsomal triglyceride transfer protein (MTP) in both mouse and rat hepatocytes (48). MTP catalyzes the transfer of neutral lipids to apob in the endoplasmic reticulum (59), and MTP expression has been shown to determine the rate of apob secretion (60 62). In agreement with increased MTP expression, PPARa activation increased the secretion of both apob-100 and apob-48 from mouse hepatocytes. These data are partially in agreement with earlier results from rat hepatocytes. ApoB-100 secretion increased, whereas apob- 48 secretion was unaffected by Wy incubation of rat hepatocytes (16, 48). In cultured rat hepatocytes, the effect of Wy on apob-100 secretion was not explained by the changed editing of apob mrna (16). Moreover, it is unlikely that the increased apob-100 secretion from mouse hepatocytes is the result of decreased editing of apob mrna (63), because both apob-48 and apob-100 secretion increased to a similar degree. Thus, the reason for the different effects of Wy on apob-48 secretion in mouse and rat hepatocytes is unclear. PPARa and ADRP in mouse liver 337

104 Fig. 7. Effect of Wy in cells overexpressing ADRP on ADRP protein expression (A), secretion of newly synthesized triglycerides (TG; B), palmitic acid oxidation (C), and intracellular triglyceride mass (D). Hepatocytes from C57BL/6 mice were isolated by liver perfusion and transduced with adenoviruses expressing zsgreen or ADRP 4 h after seeding. After 17 h of infection, medium was replaced with medium with or without 10 AM Wy. Analyses were performed 3 days after infection. A: ADRP protein expression was determined by Western blot. B: Accumulation of newly synthesized triglycerides in the cell culture medium was determined after 6 h of incubation with [ 3 H]palmitic acid. C: Fatty oxidation was determined after 1 h of incubation with [ 3 H]palmitic acid. D: Total triglyceride mass in hepatocytes was determined by HPLC separation of neutral lipids. Data are based on measurements of dpm/ag DNA (B, C) and Ag triglyceride/ag DNA (D) and presented as percentages of zsgreen. Values are means 6 SEM based on one to two (A), two (B, C), or three (D) culture dishes of three (A C) or two (D) independent liver perfusions. *P, 0.05, zsgreen versus ADRP and zsgreen versus ADRP 1 Wy; # P, 0.05, ADRP versus ADRP + Wy (Kruskal- Wallis test followed by Mann-Whitney U-test). It has been shown that the fatty acids taken up by liver cells are not immediately available for VLDL assembly. A large part of the fatty acids are esterified in the cytosol, and cytosolic triglycerides need to be hydrolyzed into diacylglycerol and fatty acids to be available for the triglyceride synthesis for VLDL assembly (64, 65). Because Wy increases MTP activity (48) and cytosolic triglyceride levels, the availability of cytosolic triglycerides for the MTP-mediated lipidation of apob in the VLDL assembly process must be prevented by PPARa activation. Moreover, PPARa activation must result in other changes than increased ADRP expression that prevents the flux of substrates for triglyceride synthesis to endoplasmic reticulum for VLDL assembly, because treatment with Wy had an effect also in ADRPoverexpressing cells. Fenofibrate treatment of rats increased DGAT activity in the cytoplasm and decreased DGAT activity in the endoplasmic reticulum (18). These changes would, like increased ADRP expression, result in the selective accumulation of triglycerides in the cytoplasm and decrease the availability of triglycerides for the MTP-dependent lipidation of apob. It is less likely that the reduced activity of triglyceride hydrolase takes part in the decreased triglyceride secretion, because clofibrate treatment of mice did not influence the lipolytic activity of microsomes (66). In summary, PPARa activation does not primarily decrease triglyceride secretion via enhanced fatty acid oxidation, because the intracellular triglyceride content of hepatocytes is not limiting for the availability of triglycerides. Rather, PPARa activation prevents the use of cytosolic triglycerides for VLDL assembly, in part by increasing the expression of ADRP. The authors thank Björn Magnusson (Wallenberg Laboratory for Cardiovascular Research) for the ADRP construct, Lennart Svensson and coworkers (Department of Molecular Pharmacology, AstraZeneca R&D) for the analyses of hepatic triglyceride content, and Anna-Karin Asztely and Carina Hallberg (Department of Molecular Pharmacology, AstraZeneca R&D) for analyses of triglyceride content in mouse hepatocytes. The authors also thank Dr. Frank Gonzalez for providing the PPARa null mice. This work was supported by Grant from the Swedish Medical Research Council, AstraZeneca, and the Swedish Heart and Lung Foundation. REFERENCES 1. Desvergne, B., and W. Wahli Peroxisome proliferatoractivated receptors: nuclear control of metabolism. Endocr. Rev. 20: Braissant, O., F. Foufelle, C. Scotto, M. Dauça, and W. Wahli Differential expression of peroxisome proliferator-activated receptors (PPARs): tissue distribution of PPAR-alpha, -beta, and -gamma in the adult rat. Endocrinology. 137: Kliewer, S. A., B. M. Forman, B. Blumberg, E. S. Ong, U. Borgmeyer, D. J. Mangelsdorf, K. Umesono, and R. M. Evans. 338 Journal of Lipid Research Volume 47, 2006

105 1994. Differential expression and activation of a family of murine peroxisome proliferator-activated receptors. Proc. Natl. Acad. Sci. USA. 91: Mukherjee, R., L. Jow, G. E. Croston, and J. R. Paterniti, Jr Identification, characterization, and tissue distribution of human peroxisome proliferator-activated receptor (PPAR) isoforms PPARgamma2 versus PPARgamma1 and activation with retinoid X receptor agonists and antagonists. J. Biol. Chem. 272: Auboeuf, D., J. Rieusset, L. Fajas, P. Vallier, V. Frering, J. P. Riou, B. Staels, J. Auwerx, M. Laville, and H. Vidal Tissue distribution and quantification of the expression of mrnas of peroxisome proliferator-activated receptors and liver X receptor-alpha in humans: no alteration in adipose tissue of obese and NIDDM patients. Diabetes. 46: Sirtori, C. R., and G. Franceschini Effects of fibrates on serum lipids and atherosclerosis. Pharmacol. Ther. 37: Diabetes Atherosclerosis Intervention Study Investigators Effect of fenofibrate on progression of coronary-artery disease in type 2 diabetes: the Diabetes Atherosclerosis Intervention Study, a randomised study. Lancet. 357: Schoonjans, K., J. Peinado-Onsurbe, A. M. Lefebvre, R. A. Heyman, M. Briggs, S. Deeb, B. Staels, and J. Auwerx PPARalpha and PPARgamma activators direct a distinct tissue-specific transcriptional response via a PPRE in the lipoprotein lipase gene. EMBO J. 15: Hertz, R., J. Bishara-Shieban, and J. Bar-Tana Mode of action of peroxisome proliferators as hypolipidemic drugs. Suppression of apolipoprotein C-III. J. Biol. Chem. 270: Staels, B., N. Vu-Dac, V. A. Kosykh, R. Saladin, J. C. Fruchart, J. Dallongeville, and J. Auwerx Fibrates downregulate apolipoprotein C-III expression independent of induction of peroxisomal acyl coenzyme A oxidase. A potential mechanism for the hypolipidemic action of fibrates. J. Clin. Invest. 95: Caslake, M. J., C. J. Packard, A. Gaw, E. Murray, B. A. Griffin, B. D. Vallance, and J. Shepherd Fenofibrate and LDL metabolic heterogeneity in hypercholesterolemia. Arterioscler. Thromb. 13: Guerin, M., W. Le Goff, E. Frisdal, S. Schneider, D. Milosavljevic, E. Bruckert, and M. J. Chapman Action of ciprofibrate in type IIb hyperlipoproteinemia: modulation of the atherogenic lipoprotein phenotype and stimulation of high-density lipoproteinmediated cellular cholesterol efflux. J. Clin. Endocrinol. Metab. 88: Packard, C. J Triacylglycerol-rich lipoproteins and the generation of small, dense low-density lipoprotein. Biochem. Soc. Trans. 31: Kesäniemi, Y. A., and S. M. Grundy Influence of gemfibrozil and clofibrate on metabolism of cholesterol and plasma triglycerides in man. J. Am. Med. Assoc. 251: Petit, D., M. T. Bonnefis, C. Rey, and R. Infante Effects of ciprofibrate and fenofibrate on liver lipids and lipoprotein synthesis in normo- and hyperlipidemic rats. Atherosclerosis. 74: Lindén, D., K. Lindberg, J. Oscarsson, C. Claesson, L. Asp, L. Li, M. Gustafsson, J. Borén, and S. O. Olofsson Influence of peroxisome proliferator-activated receptor alpha agonists on the intracellular turnover and secretion of apolipoprotein (Apo) B- 100 and ApoB-48. J. Biol. Chem. 277: Lamb, R. G., J. C. Koch, and S. R. Bush An enzymatic explanation of the differential effects of oleate and gemfibrozil on cultured hepatocyte triacylglycerol and phosphatidylcholine biosynthesis and secretion. Biochim. Biophys. Acta. 1165: Waterman, I. J., and V. A. Zammit Differential effects of fenofibrate or simvastatin treatment of rats on hepatic microsomal overt and latent diacylglycerol acyltransferase activities. Diabetes. 51: Londos, C., D. L. Brasaemle, C. J. Schultz, J. P. Segrest, and A. R. Kimmel Perilipins, ADRP, and other proteins that associate with intracellular neutral lipid droplets in animal cells. Semin. Cell Dev. Biol. 10: Jiang, H. P., and G. Serrero Isolation and characterization of a full-length cdna coding for an adipose differentiation-related protein. Proc. Natl. Acad. Sci. USA. 89: Brasaemle, D. L., T. Barber, N. E. Wolins, G. Serrero, E. J. Blanchette-Mackie, and C. Londos Adipose differentiationrelated protein is an ubiquitously expressed lipid storage dropletassociated protein. J. Lipid Res. 38: Heid, H. W., R. Moll, I. Schwetlick, H. R. Rackwitz, and T. W. Keenan Adipophilin is a specific marker of lipid accumulation in diverse cell types and diseases. Cell Tissue Res. 294: Steiner, S., D. Wahl, B. L. K. Mangold, R. Robison, J. Raymackers, L. Meheus, N. L. Anderson, and A. Cordier Induction of the adipose differentiation-related protein in liver of etomoxir-treated rats. Biochem. Biophys. Res. Commun. 218: Imamura, M., T. Inoguchi, S. Ikuyama, S. Taniguchi, K. Kobayashi, N. Nakashima, and H. Nawata ADRP stimulates lipid accumulation and lipid droplet formation in murine fibroblasts. Am. J. Physiol. Endocrinol. Metab. 283: E775 E Larigauderie, G., C. Furman, M. Jaye, C. Lasselin, C. Copin, J. C. Fruchart, G. Castro, and M. Rouis Adipophilin enhances lipid accumulation and prevents lipid efflux from THP-1 macrophages: potential role in atherogenesis. Arterioscler. Thromb. Vasc. Biol. 24: Gao, J., and G. Serrero Adipose differentiation related protein (ADRP) expressed in transfected COS-7 cells selectively stimulates long chain fatty acid uptake. J. Biol. Chem. 274: Gao, J., H. Ye, and G. Serrero Stimulation of adipose differentiation related protein (ADRP) expression in adipocyte precursors by long-chain fatty acids. J. Cell. Physiol. 182: Vosper, H., L. Patel, T. L. Graham, G. A. Khoudoli, A. Hill, C. H. Macphee, I. Pinto, S. A. Smith, K. E. Suckling, C. R. Wolf, et al The peroxisome proliferator-activated receptor delta promotes lipid accumulation in human macrophages. J. Biol. Chem. 276: Hodgkinson, C. P., and S. Ye Microarray analysis of peroxisome proliferator-activated receptor-gamma induced changes in gene expression in macrophages. Biochem. Biophys. Res. Commun. 308: Gupta, R. A., J. A. Brockman, P. Sarraf, T. M. Willson, and R. N. DuBois Target genes of peroxisome proliferator-activated receptor gamma in colorectal cancer cells. J. Biol. Chem. 276: Chawla, A., C. H. Lee, Y. Barak, W. He, J. Rosenfeld, D. Liao, J. Han, H. Kang, and R. M. Evans PPARdelta is a very low-density lipoprotein sensor in macrophages. Proc. Natl. Acad. Sci. USA. 100: Targett-Adams, P., M. J. McElwee, E. Ehrenborg, M. C. Gustafsson, C. N. Palmer, and J. McLauchlan A PPAR response element regulates transcription of the gene for human adipose differentiation-related protein. Biochim. Biophys. Acta. 1728: Ljung, B., K. Bamberg, B. Dahllöf, A. Kjellstedt, N. D. Oakes, J. Östling, L. Svensson, and G. Camejo AZ 242, a novel PPARalpha/gamma agonist with beneficial effects on insulin resistance and carbohydrate and lipid metabolism in ob/ob mice and obese Zucker rats. J. Lipid Res. 43: Lee, S. S., T. Pineau, J. Drago, E. J. Lee, J. W. Owens, D. L. Kroetz, P. M. Fernandez-Salguero, H. Westphal, and F. J. Gonzalez Targeted disruption of the alpha isoform of the peroxisome proliferator-activated receptor gene in mice results in abolishment of the pleiotropic effects of peroxisome proliferators. Mol. Cell. Biol. 15: Lindén, D., L. William-Olsson, M. Rhedin, A. K. Asztely, J. C. Clapham, and S. Schreyer Overexpression of mitochondrial GPAT in rat hepatocytes leads to decreased fatty acid oxidation and increased glycerolipid biosynthesis. J. Lipid Res. 45: Zhang, W. W., P. E. Koch, and J. A. Roth Detection of wild-type contamination in a recombinant adenoviral preparation by PCR. Biotechniques. 18: Carlsson, L., I. Nilsson, and J. Oscarsson Hormonal regulation of liver fatty acid-binding protein in vivo and in vitro: effects of growth hormone and insulin. Endocrinology. 139: Lindén, D., M. Alsterholm, H. Wennbo, and J. Oscarsson PPARalpha deficiency increases secretion and serum levels of apolipoprotein B-containing lipoproteins. J. Lipid Res. 42: Améen, C., and J. Oscarsson Sex difference in hepatic microsomal triglyceride transfer protein expression is determined by the growth hormone secretory pattern in the rat. Endocrinology. 144: Leung, K. C., and K. K. Ho Stimulation of mitochondrial fatty acid oxidation by growth hormone in human fibroblasts. J. Clin. Endocrinol. Metab. 82: Labarca, C., and K. Paigen A simple, rapid, and sensitive DNA assay procedure. Anal. Biochem. 102: Bligh, E. G., and W. J. Dyer A rapid method of total lipid extraction and purification. Can. J. Med. Sci. 37: PPARa and ADRP in mouse liver 339

106 43. Sjöberg, A., J. Oscarsson, K. Boström, T. L. Innerarity, S. Edén, and S. O. Olofsson Effects of growth hormone on apolipoprotein-b (apob) messenger ribonucleic acid editing, and apob 48 and apob 100 synthesis and secretion in the rat liver. Endocrinology. 130: Lindén, D., A. Sjöberg, L. Asp, L. Carlsson, and J. Oscarsson Direct effects of growth hormone on production and secretion of apolipoprotein B from rat hepatocytes. Am. J. Physiol. Endocrinol. Metab. 279: E1335 E Murphy, E. J., T. A. Rosenberger, and L. A. Horrocks Separation of neutral lipids by high-performance liquid chromatography: quantification by ultraviolet, light scattering and fluorescence detection. J. Chromatogr. B Biomed. Appl. 685: Folch, J., M. Lees, and G. H. Sloane Stanley A simple method for the isolation and purification of total lipides from animal tissues. J. Biol. Chem. 226: Tugwood, J. D., I. Issemann, R. G. Anderson, K. R. Bundell, W. L. McPheat, and S. Green The mouse peroxisome proliferator activated receptor recognizes a response element in the 5V flanking sequence of the rat acyl CoA oxidase gene. EMBO J. 11: Améen, C., U. Edvardsson, A. Ljungberg, L. Asp, P. Åkerblad, A. Tuneld, S. O. Olofsson, D. Lindén, and J. Oscarsson Activation of peroxisome proliferator-activated receptor alpha increases the expression and activity of microsomal triglyceride transfer protein in the liver. J. Biol. Chem. 280: Liu, P. C., R. Huber, M. D. Stow, K. L. Schlingmann, P. Collier, B. Liao, J. Link, T. C. Burn, G. Hollis, P. R. Young, et al Induction of endogenous genes by peroxisome proliferator activated receptor alpha ligands in a human kidney cell line and in vivo. J. Steroid Biochem. Mol. Biol. 85: Hertz, R., J. Arnon, and J. Bar-Tana The effect of bezafibrate and long-chain fatty acids on peroxisomal activities in cultured rat hepatocytes. Biochim. Biophys. Acta. 836: Chou, C. J., M. Haluzik, C. Gregory, K. R. Dietz, C. Vinson, O. Gavrilova, and M. L. Reitman WY14,643, a peroxisome proliferator-activated receptor alpha (PPARalpha) agonist, improves hepatic and muscle steatosis and reverses insulin resistance in lipoatrophic A-ZIP/F-1 mice. J. Biol. Chem. 277: Ip, E., G. C. Farrell, G. Robertson, P. Hall, R. Kirsch, and I. Leclercq Central role of PPARalpha-dependent hepatic lipid turnover in dietary steatohepatitis in mice. Hepatology. 38: Miller, D. B., and J. D. Spence Clinical pharmacokinetics of fibric acid derivatives (fibrates). Clin. Pharmacokinet. 34: Chung, S., J. M. Brown, M. B. Sandberg, and M. McIntosh Trans-10,cis-12 CLA increases adipocyte lipolysis and alters lipid droplet-associated proteins: role of mtor and ERK signaling. J. Lipid Res. 46: Xu, G., C. Sztalryd, X. Lu, J. T. Tansey, J. Gan, H. Dorward, A. R. Kimmel, and C. Londos. Post-translational regulation of ADRP by the ubiquitin/proteosome pathway. J. Biol. Chem. Epub ahead of print. August 22, 2005, doi: /jbc.m Martin, G., K. Schoonjans, A. M. Lefebvre, B. Staels, and J. Auwerx Coordinate regulation of the expression of the fatty acid transport protein and acyl-coa synthetase genes by PPARalpha and PPARgamma activators. J. Biol. Chem. 272: Issemann, I., R. Prince, J. Tugwood, and S. Green A role for fatty acids and liver fatty acid binding protein in peroxisome proliferation? Biochem. Soc. Trans. 20: Storch, J., and A. E. Thumser The fatty acid transport functionof fatty acid-binding proteins. Biochim. Biophys. Acta. 1486: Jamil, H., J. K. Dickson, Jr., C. H. Chu, M. W. Lago, J. K. Rinehart, S. A. Biller, R. E. Gregg, and J. R. Wetterau Microsomal triglyceride transfer protein. Specificity of lipid binding and transport. J. Biol. Chem. 270: Liao, W., K. Kobayashi, and L. Chan Adenovirus-mediated overexpression of microsomal triglyceride transfer protein (MTP): mechanistic studies on the role of MTP in apolipoprotein B-100 biogenesis. Biochemistry. 38: Tietge, U. J., A. Bakillah, C. Maugeais, K. Tsukamoto, M. Hussain, and D. J. Rader Hepatic overexpression of microsomal triglyceride transfer protein (MTP) results in increased in vivo secretion of VLDL triglycerides and apolipoprotein B. J. Lipid Res. 40: Raabe, M., L. M. Flynn, C. H. Zlot, J. S. Wong, M. M. Veniant, R. L. Hamilton, and S. G. Young Knockout of the abetalipoproteinemia gene in mice: reduced lipoprotein secretion in heterozygotes and embryonic lethality in homozygotes. Proc. Natl. Acad. Sci. USA. 95: Fu, T., D. Mukhopadhyay, N. O. Davidson, and J. Borensztajn The peroxisome proliferator-activated receptor alpha (PPARalpha) agonist ciprofibrate inhibits apolipoprotein B mrna editing in low density lipoprotein receptor-deficient mice: effects on plasma lipoproteins and the development of atherosclerotic lesions. J. Biol. Chem. 279: Gibbons, G. F., D. Wiggins, A. M. Brown, and A. M. Hebbachi Synthesis and function of hepatic very-low-density lipoprotein. Biochem. Soc. Trans. 32: Dolinsky, V. W., D. Gilham, M. Alam, D. E. Vance, and R. Lehner Triacylglycerol hydrolase: role in intracellular lipid metabolism. Cell. Mol. Life Sci. 61: Dolinsky, V. W., D. Gilham, G. M. Hatch, L. B. Agellon, R. Lehner, and D. E. Vance Regulation of triacylglycerol hydrolase expression by dietary fatty acids and peroxisomal proliferatoractivated receptors. Biochim. Biophys. Acta. 1635: Journal of Lipid Research Volume 47, 2006

107 IV

108

109 Hepatic PGC-1β overexpression in mice causes combined hyperlipidemia and a blunted response to PPARα activation Anna Ljungberg 1,2, Christopher Lelliott 3, Andrea Ahnmark 3, Lena William-Olsson 3, Anders Elmgren 3, Jan Oscarsson 1,2,3, Daniel Lindén 1,2,3. 1 Wallenberg Laboratory for Cardiovascular Research, 2 Department of Physiology, Sahlgrenska Academy, Göteborg University, SE Göteborg, Sweden. 3 AstraZeneca R&D, SE Mölndal, Sweden. ABSTRACT Objective: Members of the peroxisome proliferator-activated receptor γ coactivator-1 (PGC-1) family are important regulators of hepatic metabolism participating in the function of several transcription factors. We studied the effect of hepatic overexpression of PGC-1α and PGC-1β on plasma lipids, genes controlling hepatic lipid metabolism and PPARα activation. Methods and Results: C57BL/6 mice transduced with PGC-1α or PGC-1β adenoviruses were treated with the PPARα agonist Wy14,643 (Wy). PGC-1β overexpression suppressed PGC-1α mrna expression but not vice versa. Only PGC-1β increased plasma triglycerides, free fatty acids (FFA), VLDL-LDL cholesterol and apob levels. PGC-1β markedly upregulated DGAT1, which may contribute to the combined hyperlipidemia. On the other hand, DGAT2, FAS, SCD-1 or MTP expression were not increased by PGC-1α or PGC-1β overexpression. Wy did not decrease the hyperlipidemia caused by PGC-1β. This may be due to the decreased effect of Wy treatment on MCAD, ACO, Cyp4a10 and ADRP mrna expression in PGC-1β overexpressing mice. Conclusion: Hepatic PGC-1β overexpression induced a combined hyperlipidemia associated with increased DGAT1 expression and plasma FFA. The potentially beneficial effects of PPARα activation on genes controlling lipid metabolism were blunted by PGC- 1β overexpression. Key words: DGAT, apolipoprotein B, hypertriglyceridemia, hypercholesterolemia, adenovirus INTRODUCTION Alterations of liver lipid homeostasis have profound effects on circulating lipoproteins and peripheral tissue lipid metabolism. This can be seen in conditions such as familial combined hyperlipidemia and the progression of fatty liver in the insulin resistance syndrome; conditions associated with a marked increase in hepatic secretion of very low-density lipoproteins (VLDL) 1-3. The PGC family of transcriptional coactivators consists of three members; PGC-1α, PGC-1β and PRC 4-8. All share a basic structure; an N-terminal domain containing nuclear hormone receptorinteracting motifs (LXXLL motifs), a C- terminal region containing an RNA binding motif (RMM) and arginine/serine rich (RS) domains 9. Despite the name of this family, it is clear that the PGC family coactivate a range of transcription factors involved in general metabolic control, placing them at the centre of the homeostatic mechanism 1. Both PGC-1α and β are known to regulate the expression of genes involved in fatty acid oxidation and nuclear-encoded subunits of the 7, 10, mitochondrial electron transport chain 11. PGC-1α is a key regulator of the hepatic fasting response with regards to gluconeogenesis 11-14, whereas PGC-1β is more involved in the handling of acute lipid loads 15. Hepatic overexpression of 1

110 PGC-1β in fat-fed rats increased serum triglycerides and VLDL-cholesterol levels, with a concomitant reduction of liver lipids 15. This response was associated with elevated expression of fatty acid synthesis genes. These functional divergences suggest that PGC family members have specific functions in hepatic carbohydrate and lipid metabolism. Based on cell transfection studies, both PGC-1α and β have been shown to coactivate the nuclear receptor peroxisome proliferator-activated receptor α (PPARα) 9, 16. Agonists to PPARα, such as fibrates, are in clinical use for treatment of hypertriglyceridemia, a property that may be related to effects on hepatic fatty acid oxidation and lipoprotein production and catabolism 17, 18. However, the molecular basis of these actions of PPARα agonists is far from clear. Thus, it is possible that PPARα-mediated effects on hepatic lipid homeostasis may be mediated in part by an interaction of PPARα with the PGC-1s. In the current study we used adenoviralmediated gene transfer to study the effects of hepatic PGC-1α and β overexpression on lipoprotein levels, hepatic lipid content and expression of genes involved in lipid metabolism. In addition, we studied the effects of the specific PPARα agonist Wy14,643 on the changes induced by hepatic PGC-1α and β overexpression. METHODS Materials AD-293 cells and Top10 E. coli were obtained from Stratagene (La Jolla, CA). Dulbeccos modified Eagle s medium (DMEM), fetal bovine serum (FBS), penicillin, streptomycin, lipofectamine, Trizol, pentr1a vector, pad/pl-dest Gateway vector and LR ClonaseII enzyme mix were purchased from Invitrogen (Paisley, Scotland). Cell factories were obtained from Nunc (Rochester, NY), 10DG columns and polyvinyldifluoride (PVDF) membranes from BioRad (Hercules, CA) and BCA protein determination kit and horseradish peroxidase-conjugated goat α-rabbit secondary antibody from Pierce Biotechnology (Rockford, IL). ZsGreen and Adeno-X Rapid Titer Kit were from Clontech (Palo Alto, CA), GemT-Easy cloning vector from Promega (Madison, WI) and PacI from Qiagen (Hilden, Germany). Standard chow diet (R3) was purchased from Lactamin AB (Kimstad, Sweden), Isoflurane (Forene) from Abbot Scandinavia AB (Sweden) and Wy14,643 (Wy) from Chemsyn Science Laboratories (Lenaxa, KS). Triglyceride and cholesterol assays and Mini Complete protease inhibitor were obtained from Roche (Mannheim, Germany) and alanine aminotransferase (ALAT) kit from ABX Diagnostics (Montpellier, France). Superose 6 PC 3.2/30 column and enhanced chemiluminescence kit were from Amersham Pharmacia Biotech (Uppsala, Sweden) and the NEFA kit was from Wako Chemicals (Neuss, Germany). Turbo DNA-free was purchased from Ambion (Austin, TX) and High-Capacity cdna Archive Kit, ABI Prism 7900HT Sequence Detection System and Primer Express software from Applied Biosystems (Foster City, CA). Other chemicals were obtained from Sigma (St Louis, MO). Production of recombinant adenoviruses Full-length cdna sequences encoding mouse PGC-1α and β were amplified by PCR from C57BL/6 mouse liver cdna using sequences from the NCBI Genbank website ( /index.html). After ligating into GemT- Easy cloning vector and sequencing for PCR fidelity, the cdnas were ligated into a pentr1a vector containing a CMV promoter. These vectors were then used for homologous recombination of the expression cassette into pad/pl-dest Gateway vector using LR ClonaseII enzyme mix. After isolation of plasmids from Top10 E. coli and sequencing to ensure the accuracy of the recombination, isolated plasmids were linearized with PacI and purified by gel extraction. The packaging cell line AD-293 was grown in DMEM supplemented with 10% FBS, penicillin ( IU/L) and 2

111 streptomycin (100 mg/l). Cells were seeded in 25 cm 2 culture flasks and transfected with adenoviral constructs for PGC-1α, PGC-1β or the control ZsGreen, using lipofectamine in DMEM without serum. After 4 hours incubation, 10% FBS was added and cells were cultured for days. Viruses were harvested by repeated freeze/thaw cycles in 10 mm Tris- HCl (ph 8.0). Following large-scale amplification using Cell Factories, recombinant adenoviruses were purified by two rounds of CsCl density gradient ultracentrifugation. The purified virus stocks were desalted over 10DG columns and eluted in sterile PBS. Glycerol (13 %, final concentration) was added before the virus stocks were aliquoted and stored at - 80 C until use. Infectious viral titers were determined using the Adeno-X Rapid Titer Kit. All purified virus stocks were screened for possible wild type virus contamination according to Zhang et al. 19 before use. Animals and treatment Male C57BL/6 mice (7 weeks old) from Harlan (Netherlands) were maintained under standardized conditions of temperature (21-22 C) and humidity (40-60%), with light from 0600 to 1800 h. The animals had free access to water and standard chow diet containing (energy %) 12% fat, 62% carbohydrates, and 26% protein, with a total energy content of 12.6 kj/g. Weight-matched mice were transduced with either Ad-PGC-1α, Ad- PGC-1β or the control virus Ad-ZsGreen via tail vein injections alone or in combination with Wy treatment. The virus doses used were 0.5, 1.0 and 2.0 x 10 9 infectious forming units (ifu)/mouse in 250 µl PBS for the dose-response study and 1.2 x 10 9 ifu for the interaction study. Wy was given by gavage (30 µmol/kg/day in 0.5% (w/v) methyl cellulose) once daily for 7 days. Weight-matched control mice received only vehicle. The viruses were injected 2 days after Wy treatment commenced. Five days after virus transduction, mice were anesthetized with isoflurane, blood was collected by cardiac puncture and the livers were snap-frozen in liquid nitrogen and stored at -80 o C until analysis. Experimental procedures were approved by the Ethics Review Committee on Animal Experiments (Gothenburg region). Lipid analyses Plasma levels of triglycerides, cholesterol, NEFA and ALAT were measured using commercially available kits. The size distribution profiles of serum lipoproteins were measured using a high performance liquid chromatography system, SMART, as described before 20. In brief, 10 µl pooled serum from 6-7 mice in each group was loaded on a Superose 6 PC 3.2/30 column and the chromatographic system was linked to an air segmented continuous flow system for on-line post-derivatisation analysis of total cholesterol. Serum apob concentrations were determined with an electroimmunoassay as previously described 20. Frozen livers were homogenized in isopropanol (1 ml/50 mg tissue) and incubated at 4ºC for 1 h. The samples were centrifuged at 4ºC for 5 min at 2500 rpm and triglyceride concentrations in the supernatants were measured as described above. cdna synthesis and real-time PCR Total RNA was isolated from mouse liver with Trizol. Turbo DNA-free was used to remove contaminating DNA from the RNA preparations. First strand cdna was synthesized from 2.5 µg of total RNA with High-Capacity cdna Archive Kit. Quantitative real-time PCR was performed in 384-well plates with the ABI Prism 7900HT Sequence Detection System. All samples were analyzed in triplicate and the expression data were normalized to the endogenous control, acidic ribosomal phosphoprotein P0 (36B4). Specific primers and probes for each gene (Table I, available online at org) were designed with Primer Express software (version 2.0). Protein extractions and Western blots Frozen liver pieces were homogenized in ice-cold homogenization buffer (10 mm Hepes-KOH, ph 7.4, 125 mm mannitol, 40 mm sucrose, 2 mm EDTA with 1 tablet/10 3

112 ml Mini Complete protease inhibitor) and centrifuged at 800 x g for 5 minutes at 4ºC. The supernatant was centrifuged at 200,000 x g for 30 minutes at 4ºC to pellet the membrane fraction which was resuspended (10 mm Hepes ph 7.2, 1 mm EDTA, 1 tablet/10 ml Mini Complete protease inhibitor). Protein concentration was measured with BCA protein determination kit. 10 µg protein was separated on 10% SDS-PAGE gel and subsequently transferred onto polyvinyldifluoride (PVDF) membranes. Membranes were blocked with 5% milk powder in PBS with 0.1% Tween-20 (PBS- T). A polyclonal rabbit α-mouse DGAT1 antibody (1:800) was kindly provided by Xiao-Rong Peng (AstraZeneca, Mölndal, Sweden) and a polyclonal rabbit α-mouse MTP antibody (1:2000) was kindly provided by Carol Shoulders (Imperial College, London). The MTP antibody is known to cross-react with protein disulphide isomerase (PDI) 21. The corresponding horseradish peroxidaseconjugated goat α-rabbit secondary antibody was used at 1:5000 and 1:50000, respectively. Bands were visualized by enhanced chemiluminescence kit, blots were scanned and then analyzed using NIH ImageJ 1.34s. Statistics Values are expressed as means ± SEM. Comparisons between groups were made by Kruskal-Wallis test and Mann-Whitney U test. P < 0.05 was considered statistically significant. RESULTS Liver-directed overexpression of PGC-1β results in combined hyperlipidemia Male C57BL/6 mice were transduced via the tail vein with either Ad-PGC-1α, Ad- PGC-1β or the control virus Ad-ZsGreen. The mice were given either 0.5 x 10 9, 1 x 10 9 or 2 x 10 9 ifu. Neither dose resulted in any difference in body weight gain or food consumption (data not shown). PGC-1α mrna was induced up to 16-fold and PGC-1β was induced up to 75-fold with the highest virus dose given (Figure 1A and B). Interestingly, PGC-1β overexpression suppressed PGC-1α mrna levels, while Ad-PGC-1α had no effect on PGC-1β gene expression. Plasma triglyceride levels were not affected by Ad-PGC-1α, irrespective of virus dose, whereas Ad-PGC-1β markedly increased plasma triglyceride levels (Figure 1C). In addition, PGC-1β increased plasma apob (Figure 1D) and cholesterol levels (Figure 1E), while PGC-1α had no or very small effects. The PGC-1β induced hypertriglyceridemia was not associated with depletion of liver triglycerides (data not shown). Plasma ALAT levels were not increased in mice transduced with 0.5 x 10 9 or 1 x 10 9 ifu, whereas the groups given the highest dose showed elevated ALAT levels (Figure 1F). Taken together, hepatic overexpression of PGC-1β, but not PGC- 1α, results in combined hyperlipidemia. 4

113 Figure 1. Effects of hepatic PGC-1α or PGC-1β overexpression on PGC-1α mrna (A), PGC-1β mrna (B), plasma triglyceride (C), plasma apob (D), plasma cholesterol (E) and plasma ALAT (F) levels. Mice were injected via the tail vein with 3 different doses (0.5, 1.0 or 2.0 x 10 9 ifu/mouse) of Ad-PGC-1α (1α), Ad-PGC-1β (1β) or Ad-ZsGreen (ZsG) and killed 5 days later. Values are means ± SEM (n=2-3). PPARα activation does not potentiate the effect of PGC-1β in vivo Both PGC-1α and β have been shown in vitro to coactivate PPARα 9, 16. In order to study this in vivo, male C57BL/6 mice were treated with the PPARα agonist Wy for 7 days, injected with Ad-PGC-1α, Ad- PGC-1β or Ad-ZsGreen 2 days after treatment start, and killed 5 days later. Based on the dose-response experiment, we chose to give 1.2 x 10 9 ifu/mouse. Although PGC-1β overexpression was more efficient than that of PGC-1α, the same dose was chosen for both viruses to avoid possible differences in non-specific virus effects. In addition, we know that this dose of Ad-ZsGreen is well within the dose-range that results in a liver-specific overexpression 22. As before, PGC-1β overexpression resulted in increased plasma triglyceride, apob and cholesterol levels, whereas Ad-PGC-1α had no effect (Figure 2A-C). The PGC-1β induced cholesterol levels were associated with non-hdl fractions; i.e fractions with exclusion volume (size) in the VLDL-LDL range (Figure 2D). Wy treatment increased cholesterol levels in Ad-ZsGreen and Ad- PGC-1α transduced mice, but not in mice transduced with Ad-PGC-1β (Figure 2C). Plasma levels of non-esterified fatty acids (NEFA) were increased in Ad-PGC-1β transduced mice (Figure 2E), whereas neither virus affected liver triglyceride levels significantly (Figure 2F). Wy 5

114 treatment had no effect on NEFA or liver triglycerides. PGC-1β overexpression increased the liver weight and Wy increased liver weights irrespective of virus treatment (ZsGreen vehicle: 6.0 ± 0.3 % of body weight (% bw), ZsGreen Wy: 7.6 ± 0.1 % bw, PGC-1α vehicle: 6.2 ± 0.3 % bw, PGC-1α Wy: 8.2 ± 0.3 % bw, PGC-1β vehicle: 8.3 ± 0.2 % bw, PGC-1β Wy: 9.6 ± 0.3 % bw). Food consumption, body weight gain or ALAT levels were not different between the groups given PGC-1 viruses and Ad-ZsGreen (data not shown). Figure 2. Effects of hepatic PGC-1α or PGC-1β overexpression and Wy treatment on plasma triglyceride (A), plasma apob (B), plasma cholesterol (C), plasma lipoproteins (D), plasma NEFA (E) and liver triglyceride (F) levels. Mice were treated with Wy (30 µmol/kg/day) for 7 days. Ad-PGC-1α, Ad-PGC-1β or Ad-ZsGreen were injected via the tail vein (1.2 x 10 9 ifu) after 2 days Wy treatment. Values are means ± SEM (n=6-7). * p < 0.05; PGC-1α or PGC-1β vs ZsGreen, # p < 0.05; Wy vs vehicle treatment in respective virus group, Kruskal-Wallis test followed by Mann-Whitney U test. 6

115 Figure 3. Effects of hepatic PGC-1α or PGC-1β overexpression and Wy treatment on PGC-1α (A), PGC-1β (B), PPARα (C), MCAD (D), ACO-I (E), Cyp4a10 (F), ADRP (G), ApoC-III (H) and DGAT1 (I) mrna expression. Mice were treated with Wy (30 µmol/kg/day) for 7 days. Ad-PGC-1α, Ad-PGC-1β or Ad- ZsGreen were injected via the tail vein (1.2 x 10 9 ifu) after 2 days Wy treatment. Values are means ± SEM (n=6-7). * p < 0.05; PGC-1α or PGC-1β vs ZsGreen, # p < 0.05; Wy vs vehicle treatment in respective virus group, p < 0.05; Wy treated ZsGreen vs Wy treated PGC-1β, Kruskal-Wallis test followed by Mann- Whitney U test. PGC-1β overexpression induces DGAT1 gene expression and blunts the effect of Wy Transduction with Ad-PGC-1α and Ad- PGC-1β resulted in 6-fold and 14-fold induction of PGC-1α (Figure 3A) and PGC-1β (Figure 3B) mrna levels, respectively. Also in this experiment, PGC- 1α mrna was downregulated by PGC-1β overexpression, but not vice versa. Furthermore, Wy treatment decreased PGC-1α mrna levels. Interestingly, PPARα mrna was downregulated by PGC-1β (Figure 3C). Medium-chain acyl- CoA dehydrogenase (MCAD), acyl-coa 7 oxidase-i (ACO-I) and cytochrome P450 4a10 (Cyp4a10), representing the mitochondrial, peroxisomal and microsomal fatty acid oxidation pathway, respectively, were measured since these genes are PPARα responsive. MCAD was upregulated by both Ad-PGC-1α and Ad- PGC-1β (Figure 3D), whereas ACO-I was not affected (Figure 3E) and Cyp4a10 was downregulated by Ad-PGC-1β (Figure 3F). Wy upregulated all these genes, irrespective of virus treatment. However, the effect of Wy was less pronounced when combined with PGC-1β overexpression. Adipose differentiation-related protein (ADRP) mrna (Figure 3G) and apoc-iii

116 mrna levels (Figure 3H) were not transcriptionally regulated by the PGC-1s, but the effect of Wy on these genes was blunted by PGC-1β overexpression. In line with the hypertriglyceridemia, acyl-coa: diacylglycerol acyltransferase (DGAT)1 mrna was strongly upregulated by Ad- PGC-1β and slightly upregulated by Ad- PGC-1α (Figure 3I). In contrast, DGAT2 was downregulated by Ad-PGC-1β (Table II, available online at atvb.ahajournals.org). In spite of the hyperlipidemia, the lipogenic genes fatty acid synthase (FAS) and stearoyl-coa desaturase (SCD)-1 were not upregulated by PGC-1β (Table II, available online at In fact, SCD-1 mrna was downregulated by both PGC- 1s. Furthermore, microsomal triglyceride transfer protein (MTP) and angiopoietinlike protein 3 (AngPtl3) gene expression was measured due to their known effects on VLDL secretion and clearance, respectively, but none of them were regulated by the PGC-1s (Table II, available online at org). Thus, PPARα activation does not potentiate the effect of PGC-1α or β. Rather, PGC-1β overexpression in vivo blunts the effect of Wy. PGC-1β overexpression induces DGAT1 protein expression The strong induction of DGAT1 mrna by Ad-PGC-1β was accompanied with a 2.5- fold increase in DGAT1 protein (Figure 4A). In contrast, PGC-1α overexpression had no effect on DGAT1. As indicated by the mrna data, MTP was not affected by PGC-1 overexpression at the protein level (Figure 4B; ZsG: 1.0 ± 0.02 fold change (f.c.), PGC-1α: 0.97 ± 0.04 f.c., PGC-1β: 0.94 ± 0.02 f.c., n = 6). Protein disulfide isomerase (PDI), which forms a heterodimer with MTP, was also unaffected by the treatment (Figure 4B). Figure 4. Effects of hepatic PGC-1α or PGC-1β overexpression on DGAT1 (A) and MTP/PDI (B) protein expression. Ad-PGC-1α, Ad-PGC-1β or Ad- ZsGreen were injected via the tail vein (1.2 x 10 9 ifu) and the mice were killed 5 days later. Protein expression was measured on membrane proteins from liver tissue. Values are means ± SEM (n=6). * p < 0.05; PGC-1α or PGC-1β vs ZsGreen, Kruskal- Wallis test followed by Mann-Whitney U test. DISCUSSION Both PGC-1α and β are induced in the liver upon fasting and promote mitochondrial biogenesis 11, 23. In addition, PGC-1α overexpression in the liver increases gluconeogenesis 13. On the other hand, hepatic PGC-1β overexpression in fat-fed rats lead to elevated plasma triglyceride and cholesterol levels and it was postulated that PGC-1β could be the link between dietary saturated fats and hyperlipidemia 15. In this study, we extended these findings by showing that hepatic PGC-1β overexpression also increased plasma apob levels, thus leading to an atherogenic lipoprotein profile with elevated triglycerides, apob and cholesterol in VLDL-LDL-like particles. Moreover, we showed that the hyperlipidemia was associated with increased DGAT-1 expression and plasma FFA, and not with increased expression of genes in de novo lipogenesis or changed liver 8